diff --git a/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/all_results.json b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..4af28ce500762056ea2a1fbd62547a02124e3ea5
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/all_results.json
@@ -0,0 +1,24 @@
+{
+  "predict_exact_match": 0.0,
+  "predict_rouge1": 50.3879,
+  "predict_eval_rougeL": 41.3728,
+  "predict_gen_len": 27.35,
+  "predict_runtime": 42.18466920289211,
+  "predict_samples": 20,
+  "predict_samples_per_second": 0.474,
+  "predict_steps_per_second": 0.474,
+  "epoch": 15.0,
+  "predict_global_step": 75,
+  "predict_exact_match_for_CL": 0.0,
+  "predict_rouge1_for_CL": 50.3879,
+  "predict_eval_rougeL_for_CL": 41.3728,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "train_runtime": 375.3924,
+  "train_samples_per_second": 6.393,
+  "train_steps_per_second": 0.2,
+  "total_flos": 2.419733437612032e+16,
+  "train_loss": 1.0816661707560222,
+  "train_samples": 160
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2d528ec3177aab7e1c53513613958d6447d81ff8
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:000e7deb567646a9c3bcf8d1fe8dc691cded4d0b943e75ea349a5ad1888ec0a0
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8ddaf3f3e5487477220a96302bfa542fe7151167
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b0fe14befd0519ccc989f0b1c5548fbfd9c61e7ce9841f19458d72cf5e84398
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..72fff74554ad2a5dcd17e4d7bd6316a706afe271
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e1e978739f5987279b42a1d8f9f3c4aeb884fdf3dfc77afc82d087f903ac32b
+size 266927761
diff --git a/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/train.log b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..496710ef5b96b220d9ae302c16425682d273aab2
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/train.log
@@ -0,0 +1,176 @@
+[2026-04-29 19:26:05,126] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 19:26:05,127] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29500 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1572_samsum_summary --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary --cur_task_id 0 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42
+[2026-04-29 19:26:08,340] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 19:26:08,340] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 19:26:08,340] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 19:26:08,340] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 19:26:08,340] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 19:26:08,340] [INFO] [launch.py:272:main] process 471369 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1572_samsum_summary', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary', '--cur_task_id', '0', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42']
+
+============================================================
+Task 0: task1572_samsum_summary
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: None
+srt_load_path:      None
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 160/20/20
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 11/291 [00:00<00:03, 93.00it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:02, 128.50it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 146.91it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 156.48it/s]
Loading weights:  29%|██▉       | 85/291 [00:00<00:01, 153.18it/s]
Loading weights:  35%|███▌      | 103/291 [00:00<00:01, 161.37it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:01, 160.56it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 156.93it/s]
Loading weights:  54%|█████▍    | 158/291 [00:01<00:00, 165.74it/s]
Loading weights:  60%|██████    | 175/291 [00:01<00:00, 163.61it/s]
Loading weights:  67%|██████▋   | 194/291 [00:01<00:00, 168.45it/s]
Loading weights:  73%|███████▎  | 211/291 [00:01<00:00, 166.91it/s]
Loading weights:  79%|███████▊  | 229/291 [00:01<00:00, 168.69it/s]
Loading weights:  85%|████████▌ | 248/291 [00:01<00:00, 173.50it/s]
Loading weights:  92%|█████████▏| 267/291 [00:01<00:00, 176.49it/s]
Loading weights: 100%|█████████▉| 290/291 [00:01<00:00, 191.09it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 166.79it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=163)
+[FIX] Re-initialized lora_A/B in 64 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6741.336M (0.04%)
+[TRAIN] step_per_epoch=5, eval_steps=25
+[SGWI] Task 0: no prior adapters, skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 19:27:21,534] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/75 [00:00<?, ?it/s]
  1%|▏         | 1/75 [00:04<05:49,  4.72s/it]
  3%|▎         | 2/75 [00:08<05:01,  4.13s/it]
  4%|▍         | 3/75 [00:12<04:48,  4.01s/it]
  5%|▌         | 4/75 [00:16<04:45,  4.02s/it]
  7%|▋         | 5/75 [00:20<04:35,  3.93s/it]
  8%|▊         | 6/75 [00:24<04:43,  4.10s/it]
  9%|▉         | 7/75 [00:28<04:42,  4.16s/it]
 11%|█         | 8/75 [00:32<04:28,  4.00s/it]
 12%|█▏        | 9/75 [00:36<04:29,  4.08s/it]
 13%|█▎        | 10/75 [00:40<04:24,  4.07s/it]
                                               
{'loss': '1.468', 'grad_norm': '0.1311', 'learning_rate': '5e-05', 'epoch': '2'}
+
 13%|█▎        | 10/75 [00:40<04:24,  4.07s/it]
 15%|█▍        | 11/75 [00:44<04:20,  4.07s/it]
 16%|█▌        | 12/75 [00:48<04:15,  4.05s/it]
 17%|█▋        | 13/75 [00:52<04:05,  3.96s/it]
 19%|█▊        | 14/75 [00:56<04:01,  3.96s/it]
 20%|██        | 15/75 [01:00<03:53,  3.89s/it]
 21%|██▏       | 16/75 [01:04<03:49,  3.90s/it]
 23%|██▎       | 17/75 [01:08<03:51,  3.99s/it]
 24%|██▍       | 18/75 [01:12<03:44,  3.94s/it]
 25%|██▌       | 19/75 [01:15<03:33,  3.81s/it]
 27%|██▋       | 20/75 [01:20<03:38,  3.97s/it]
                                               
{'loss': '1.282', 'grad_norm': '0.1732', 'learning_rate': '5e-05', 'epoch': '4'}
+
 27%|██▋       | 20/75 [01:20<03:38,  3.97s/it]
 28%|██▊       | 21/75 [01:23<03:29,  3.88s/it]
 29%|██▉       | 22/75 [01:27<03:26,  3.89s/it]
 31%|███       | 23/75 [01:31<03:22,  3.90s/it]
 32%|███▏      | 24/75 [01:35<03:21,  3.95s/it]
 33%|███▎      | 25/75 [01:39<03:17,  3.94s/it]
+
Eval @ step 25:   0%|          | 0/20 [00:00<?, ?it/s]
+
Eval @ step 25:   5%|▌         | 1/20 [00:01<00:21,  1.15s/it]
+
Eval @ step 25:  10%|█         | 2/20 [00:02<00:23,  1.31s/it]
+
Eval @ step 25:  15%|█▌        | 3/20 [00:03<00:22,  1.33s/it]
+
Eval @ step 25:  20%|██        | 4/20 [00:04<00:14,  1.13it/s]
+
Eval @ step 25:  25%|██▌       | 5/20 [00:04<00:11,  1.27it/s]
+
Eval @ step 25:  30%|███       | 6/20 [00:07<00:18,  1.31s/it]
+
Eval @ step 25:  35%|███▌      | 7/20 [00:08<00:18,  1.43s/it]
+
Eval @ step 25:  40%|████      | 8/20 [00:09<00:14,  1.17s/it]
+
Eval @ step 25:  45%|████▌     | 9/20 [00:09<00:10,  1.01it/s]
+
Eval @ step 25:  50%|█████     | 10/20 [00:10<00:08,  1.12it/s]
+
Eval @ step 25:  55%|█████▌    | 11/20 [00:13<00:12,  1.43s/it]
+
Eval @ step 25:  60%|██████    | 12/20 [00:15<00:12,  1.57s/it]
+
Eval @ step 25:  65%|██████▌   | 13/20 [00:17<00:12,  1.72s/it]
+
Eval @ step 25:  70%|███████   | 14/20 [00:18<00:09,  1.52s/it]
+
Eval @ step 25:  75%|███████▌  | 15/20 [00:20<00:08,  1.77s/it]
+
Eval @ step 25:  80%|████████  | 16/20 [00:23<00:07,  1.96s/it]
+
Eval @ step 25:  85%|████████▌ | 17/20 [00:26<00:07,  2.50s/it]
+
Eval @ step 25:  90%|█████████ | 18/20 [00:27<00:03,  1.99s/it]
+
Eval @ step 25:  95%|█████████▌| 19/20 [00:29<00:02,  2.05s/it]
+
Eval @ step 25: 100%|██████████| 20/20 [00:30<00:00,  1.67s/it]
+
                                                                 [EVAL] ★ New best eval_rougeL=37.1337 (step=25) — snapshot saved in memory
+
                                               
{'eval_rougeL': '37.13', 'epoch': '5'}
+
 33%|███▎      | 25/75 [02:10<03:17,  3.94s/it]
 35%|███▍      | 26/75 [02:14<10:41, 13.09s/it]
 36%|███▌      | 27/75 [02:17<08:13, 10.28s/it]
 37%|███▋      | 28/75 [02:22<06:40,  8.53s/it]
 39%|███▊      | 29/75 [02:26<05:30,  7.19s/it]
 40%|████      | 30/75 [02:30<04:37,  6.16s/it]
                                               
{'loss': '1.103', 'grad_norm': '0.2044', 'learning_rate': '5e-05', 'epoch': '6'}
+
 40%|████      | 30/75 [02:30<04:37,  6.16s/it]
 41%|████▏     | 31/75 [02:33<03:57,  5.40s/it]
 43%|████▎     | 32/75 [02:37<03:36,  5.03s/it]
 44%|████▍     | 33/75 [02:41<03:17,  4.70s/it]
 45%|████▌     | 34/75 [02:45<03:01,  4.42s/it]
 47%|████▋     | 35/75 [02:49<02:48,  4.22s/it]
 48%|████▊     | 36/75 [02:53<02:41,  4.13s/it]
 49%|████▉     | 37/75 [02:57<02:34,  4.06s/it]
 51%|█████     | 38/75 [03:00<02:27,  3.98s/it]
 52%|█████▏    | 39/75 [03:04<02:22,  3.95s/it]
 53%|█████▎    | 40/75 [03:08<02:17,  3.94s/it]
                                               
{'loss': '1.022', 'grad_norm': '0.2338', 'learning_rate': '5e-05', 'epoch': '8'}
+
 53%|█████▎    | 40/75 [03:08<02:17,  3.94s/it]
 55%|█████▍    | 41/75 [03:12<02:15,  4.00s/it]
 56%|█████▌    | 42/75 [03:16<02:09,  3.93s/it]
 57%|█████▋    | 43/75 [03:20<02:03,  3.87s/it]
 59%|█████▊    | 44/75 [03:24<01:58,  3.82s/it]
 60%|██████    | 45/75 [03:28<01:59,  3.98s/it]
 61%|██████▏   | 46/75 [03:32<01:59,  4.13s/it]
 63%|██████▎   | 47/75 [03:36<01:55,  4.13s/it]
 64%|██████▍   | 48/75 [03:40<01:45,  3.91s/it]
 65%|██████▌   | 49/75 [03:44<01:43,  3.98s/it]
 67%|██████▋   | 50/75 [03:48<01:40,  4.03s/it]
                                               
{'loss': '0.9652', 'grad_norm': '0.2146', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 50/75 [03:48<01:40,  4.03s/it]
+
Eval @ step 50:   0%|          | 0/20 [00:00<?, ?it/s]
+
Eval @ step 50:   5%|▌         | 1/20 [00:00<00:15,  1.22it/s]
+
Eval @ step 50:  10%|█         | 2/20 [00:01<00:18,  1.02s/it]
+
Eval @ step 50:  15%|█▌        | 3/20 [00:02<00:12,  1.40it/s]
+
Eval @ step 50:  20%|██        | 4/20 [00:02<00:08,  1.91it/s]
+
Eval @ step 50:  25%|██▌       | 5/20 [00:03<00:09,  1.57it/s]
+
Eval @ step 50:  30%|███       | 6/20 [00:05<00:16,  1.21s/it]
+
Eval @ step 50:  35%|███▌      | 7/20 [00:07<00:17,  1.33s/it]
+
Eval @ step 50:  40%|████      | 8/20 [00:08<00:14,  1.21s/it]
+
Eval @ step 50:  45%|████▌     | 9/20 [00:09<00:12,  1.10s/it]
+
Eval @ step 50:  50%|█████     | 10/20 [00:09<00:08,  1.15it/s]
+
Eval @ step 50:  55%|█████▌    | 11/20 [00:10<00:08,  1.02it/s]
+
Eval @ step 50:  60%|██████    | 12/20 [00:12<00:09,  1.16s/it]
+
Eval @ step 50:  65%|██████▌   | 13/20 [00:12<00:06,  1.02it/s]
+
Eval @ step 50:  70%|███████   | 14/20 [00:13<00:06,  1.02s/it]
+
Eval @ step 50:  75%|███████▌  | 15/20 [00:15<00:05,  1.03s/it]
+
Eval @ step 50:  80%|████████  | 16/20 [00:16<00:04,  1.10s/it]
+
Eval @ step 50:  85%|████████▌ | 17/20 [00:20<00:05,  1.90s/it]
+
Eval @ step 50:  90%|█████████ | 18/20 [00:20<00:03,  1.62s/it]
+
Eval @ step 50:  95%|█████████▌| 19/20 [00:22<00:01,  1.49s/it]
+
Eval @ step 50: 100%|██████████| 20/20 [00:22<00:00,  1.23s/it]
+
                                                                 [EVAL] ★ New best eval_rougeL=41.3728 (step=50) — snapshot saved in memory
+
                                               
{'eval_rougeL': '41.37', 'epoch': '10'}
+
 67%|██████▋   | 50/75 [04:11<01:40,  4.03s/it]
 68%|██████▊   | 51/75 [04:15<04:21, 10.89s/it]
 69%|██████▉   | 52/75 [04:19<03:23,  8.83s/it]
 71%|███████   | 53/75 [04:23<02:41,  7.34s/it]
 72%|███████▏  | 54/75 [04:27<02:12,  6.29s/it]
 73%|███████▎  | 55/75 [04:31<01:50,  5.54s/it]
 75%|███████▍  | 56/75 [04:34<01:35,  5.00s/it]
 76%|███████▌  | 57/75 [04:38<01:24,  4.72s/it]
 77%|███████▋  | 58/75 [04:42<01:16,  4.49s/it]
 79%|███████▊  | 59/75 [04:46<01:08,  4.29s/it]
 80%|████████  | 60/75 [04:50<01:02,  4.17s/it]
                                               
{'loss': '0.9301', 'grad_norm': '0.194', 'learning_rate': '5e-05', 'epoch': '12'}
+
 80%|████████  | 60/75 [04:50<01:02,  4.17s/it]
 81%|████████▏ | 61/75 [04:54<00:57,  4.09s/it]
 83%|████████▎ | 62/75 [04:58<00:51,  3.97s/it]
 84%|████████▍ | 63/75 [05:01<00:46,  3.90s/it]
 85%|████████▌ | 64/75 [05:05<00:43,  3.93s/it]
 87%|████████▋ | 65/75 [05:10<00:40,  4.05s/it]
 88%|████████▊ | 66/75 [05:14<00:36,  4.07s/it]
 89%|████████▉ | 67/75 [05:18<00:32,  4.10s/it]
 91%|█████████ | 68/75 [05:22<00:28,  4.13s/it]
 92%|█████████▏| 69/75 [05:26<00:24,  4.15s/it]
 93%|█████████▎| 70/75 [05:30<00:20,  4.08s/it]
                                               
{'loss': '0.9015', 'grad_norm': '0.3102', 'learning_rate': '5e-05', 'epoch': '14'}
+
 93%|█████████▎| 70/75 [05:30<00:20,  4.08s/it]
 95%|█████████▍| 71/75 [05:34<00:16,  4.05s/it]
 96%|█████████▌| 72/75 [05:38<00:11,  3.95s/it]
 97%|█████████▋| 73/75 [05:42<00:08,  4.00s/it]
 99%|█████████▊| 74/75 [05:46<00:04,  4.02s/it]
100%|██████████| 75/75 [05:50<00:00,  4.09s/it]
+
Eval @ step 75:   0%|          | 0/20 [00:00<?, ?it/s]
+
Eval @ step 75:   5%|▌         | 1/20 [00:01<00:22,  1.17s/it]
+
Eval @ step 75:  10%|█         | 2/20 [00:01<00:17,  1.05it/s]
+
Eval @ step 75:  15%|█▌        | 3/20 [00:02<00:11,  1.49it/s]
+
Eval @ step 75:  20%|██        | 4/20 [00:02<00:07,  2.00it/s]
+
Eval @ step 75:  25%|██▌       | 5/20 [00:03<00:11,  1.34it/s]
+
Eval @ step 75:  30%|███       | 6/20 [00:06<00:18,  1.29s/it]
+
Eval @ step 75:  35%|███▌      | 7/20 [00:06<00:13,  1.07s/it]
+
Eval @ step 75:  40%|████      | 8/20 [00:07<00:12,  1.03s/it]
+
Eval @ step 75:  45%|████▌     | 9/20 [00:08<00:10,  1.08it/s]
+
Eval @ step 75:  50%|█████     | 10/20 [00:08<00:07,  1.33it/s]
+
Eval @ step 75:  55%|█████▌    | 11/20 [00:09<00:07,  1.13it/s]
+
Eval @ step 75:  60%|██████    | 12/20 [00:11<00:08,  1.10s/it]
+
Eval @ step 75:  65%|██████▌   | 13/20 [00:12<00:07,  1.01s/it]
+
Eval @ step 75:  70%|███████   | 14/20 [00:12<00:05,  1.14it/s]
+
Eval @ step 75:  75%|███████▌  | 15/20 [00:14<00:05,  1.03s/it]
+
Eval @ step 75:  80%|████████  | 16/20 [00:16<00:05,  1.44s/it]
+
Eval @ step 75:  85%|████████▌ | 17/20 [00:20<00:06,  2.14s/it]
+
Eval @ step 75:  90%|█████████ | 18/20 [00:21<00:03,  1.79s/it]
+
Eval @ step 75:  95%|█████████▌| 19/20 [00:23<00:01,  1.90s/it]
+
Eval @ step 75: 100%|██████████| 20/20 [00:24<00:00,  1.61s/it]
+
                                                                 [EVAL] eval_rougeL=40.6301 (best=41.3728)
+
                                               
{'eval_rougeL': '40.63', 'epoch': '15'}
+
100%|██████████| 75/75 [06:15<00:00,  4.09s/it]
                                               
{'train_runtime': '375.4', 'train_samples_per_second': '6.393', 'train_steps_per_second': '0.2', 'train_loss': '1.082', 'epoch': '15'}
+
100%|██████████| 75/75 [06:15<00:00,  4.09s/it]
100%|██████████| 75/75 [06:15<00:00,  5.01s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 22535523GF
+  train_loss               =     1.0817
+  train_runtime            = 0:06:15.39
+  train_samples            =        160
+  train_samples_per_second =      6.393
+  train_steps_per_second   =        0.2
+  [EVAL] Restored best model (eval_rougeL=41.3728, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   7%|▋         | 19/291 [00:00<00:01, 173.15it/s]
Loading weights:  13%|█▎        | 37/291 [00:00<00:01, 175.35it/s]
Loading weights:  20%|█▉        | 58/291 [00:00<00:01, 167.75it/s]
Loading weights:  26%|██▋       | 77/291 [00:00<00:01, 174.40it/s]
Loading weights:  33%|███▎      | 96/291 [00:00<00:01, 178.59it/s]
Loading weights:  40%|████      | 117/291 [00:00<00:00, 184.03it/s]
Loading weights:  47%|████▋     | 137/291 [00:00<00:00, 181.83it/s]
Loading weights:  54%|█████▎    | 156/291 [00:00<00:00, 181.23it/s]
Loading weights:  60%|██████    | 175/291 [00:00<00:00, 179.36it/s]
Loading weights:  66%|██████▋   | 193/291 [00:01<00:00, 179.55it/s]
Loading weights:  73%|███████▎  | 212/291 [00:01<00:00, 182.19it/s]
Loading weights:  79%|███████▉  | 231/291 [00:01<00:00, 176.58it/s]
Loading weights:  87%|████████▋ | 252/291 [00:01<00:00, 185.54it/s]
Loading weights:  93%|█████████▎| 271/291 [00:01<00:00, 184.82it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 183.06it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (160 samples, batch_size=1)
+  [SRT] Task 'task1572_samsum_summary': PaR=3376.9, shrinkage=ridge, n=160
+  [SRT] Wired router: 1 sigs, cur=task1572_samsum_summary, mapping={'task1572_samsum_summary': 0}
+[SRT] Router: 1 tasks, avg_PaR=3376.9, metrics=['ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary']
+  [SRT-INIT] task_id_to_idx={'task1572_samsum_summary': 0}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/20 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=1
+
Continual predict:   5%|▌         | 1/20 [00:01<00:28,  1.52s/it]
Continual predict:  10%|█         | 2/20 [00:03<00:34,  1.89s/it]
Continual predict:  15%|█▌        | 3/20 [00:04<00:22,  1.32s/it]
Continual predict:  20%|██        | 4/20 [00:04<00:15,  1.03it/s]
Continual predict:  25%|██▌       | 5/20 [00:06<00:17,  1.18s/it]
Continual predict:  30%|███       | 6/20 [00:10<00:30,  2.20s/it]
Continual predict:  35%|███▌      | 7/20 [00:13<00:31,  2.44s/it]
Continual predict:  40%|████      | 8/20 [00:15<00:26,  2.24s/it]
Continual predict:  45%|████▌     | 9/20 [00:16<00:21,  1.99s/it]
Continual predict:  50%|█████     | 10/20 [00:17<00:15,  1.58s/it]
Continual predict:  55%|█████▌    | 11/20 [00:19<00:16,  1.82s/it]
Continual predict:  60%|██████    | 12/20 [00:22<00:16,  2.10s/it]
Continual predict:  65%|██████▌   | 13/20 [00:23<00:12,  1.79s/it]
Continual predict:  70%|███████   | 14/20 [00:25<00:11,  1.87s/it]
Continual predict:  75%|███████▌  | 15/20 [00:27<00:09,  1.91s/it]
Continual predict:  80%|████████  | 16/20 [00:29<00:08,  2.02s/it]
Continual predict:  85%|████████▌ | 17/20 [00:36<00:10,  3.53s/it]
Continual predict:  90%|█████████ | 18/20 [00:38<00:06,  3.03s/it]
Continual predict:  95%|█████████▌| 19/20 [00:40<00:02,  2.80s/it]
Continual predict: 100%|██████████| 20/20 [00:42<00:00,  2.31s/it]
Continual predict: 100%|██████████| 20/20 [00:42<00:00,  2.11s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=20]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% ( 20/ 20)  label_acc=100.0% ( 20/ 20)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                           =       15.0
+  predict_eval_rougeL                             =    41.3728
+  predict_eval_rougeL_for_CL                      =    41.3728
+  predict_eval_rougeL_for_task1572_samsum_summary =    41.3728
+  predict_exact_match                             =        0.0
+  predict_exact_match_for_CL                      =        0.0
+  predict_exact_match_for_task1572_samsum_summary =        0.0
+  predict_gen_len                                 =      27.35
+  predict_global_step                             =         75
+  predict_rouge1                                  =    50.3879
+  predict_rouge1_for_CL                           =    50.3879
+  predict_rouge1_for_task1572_samsum_summary      =    50.3879
+  predict_runtime                                 = 0:00:42.18
+  predict_samples                                 =         20
+  predict_samples_per_second                      =      0.474
+  predict_steps_per_second                        =      0.474
+[2026-04-29 19:34:52,405] [INFO] [launch.py:367:main] Process 471369 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/all_results.json b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..48296455d3191d96158d29308476d969bf43814a
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/all_results.json
@@ -0,0 +1,51 @@
+{
+  "predict_exact_match": 49.6413,
+  "predict_rouge1": 61.2565,
+  "predict_eval_rougeL": 59.7787,
+  "predict_gen_len": 8.1894,
+  "predict_runtime": 664.7405987819657,
+  "predict_samples": 697,
+  "predict_samples_per_second": 1.049,
+  "predict_steps_per_second": 1.049,
+  "epoch": 15.0,
+  "predict_global_step": 60,
+  "predict_exact_match_for_CL": 49.6413,
+  "predict_rouge1_for_CL": 61.2565,
+  "predict_eval_rougeL_for_CL": 59.7787,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "train_runtime": 385.9215,
+  "train_samples_per_second": 4.897,
+  "train_steps_per_second": 0.155,
+  "total_flos": 2.264373813628109e+16,
+  "train_loss": 2.9621963183085125,
+  "train_samples": 126
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f7723f033403ffae5c22e1e214b7445fb16d466b
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1dd81db47388972d50f87f7e6bac06b3f2f6817e70b484ad0cd45e7b335f407
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2d9a7c30de4bd45bf966f4fda0619c2ec7ae7880
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db385665eab6b66c41a794e1254714d218fc73aaa2f458994801044dcc6b788c
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..55ed62bce839b71978d176e7953b3cf403f6c261
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73e1e5734116c6c51499132c6c7f6cb4699e452ca79dd5669c5f33dae42bdd1b
+size 1573363363
diff --git a/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/train.log b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..e1f1d749df5c11c77c4263f26702e51752647db6
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/train.log
@@ -0,0 +1,340 @@
+[2026-04-30 01:38:16,499] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 01:38:16,499] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29509 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1590_diplomacy_text_generation --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation --cur_task_id 9 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[2026-04-30 01:38:19,742] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 01:38:19,742] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 01:38:19,742] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 01:38:19,742] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 01:38:19,742] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 01:38:19,743] [INFO] [launch.py:272:main] process 1351998 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1590_diplomacy_text_generation', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation', '--cur_task_id', '9', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights']
+
+============================================================
+Task 9: task1590_diplomacy_text_generation
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 126/16/697
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   0%|          | 1/291 [00:00<00:30,  9.55it/s]
Loading weights:   8%|▊         | 23/291 [00:00<00:02, 129.12it/s]
Loading weights:  14%|█▍        | 41/291 [00:00<00:01, 150.40it/s]
Loading weights:  22%|██▏       | 63/291 [00:00<00:01, 177.22it/s]
Loading weights:  28%|██▊       | 81/291 [00:00<00:01, 173.27it/s]
Loading weights:  34%|███▍      | 100/291 [00:00<00:01, 168.60it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:01, 168.63it/s]
Loading weights:  50%|████▉     | 145/291 [00:00<00:00, 181.85it/s]
Loading weights:  56%|█████▋    | 164/291 [00:00<00:00, 181.41it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 182.11it/s]
Loading weights:  70%|██████▉   | 203/291 [00:01<00:00, 183.33it/s]
Loading weights:  76%|███████▋  | 222/291 [00:01<00:00, 184.77it/s]
Loading weights:  83%|████████▎ | 241/291 [00:01<00:00, 182.82it/s]
Loading weights:  89%|████████▉ | 260/291 [00:01<00:00, 181.48it/s]
Loading weights:  96%|█████████▌| 280/291 [00:01<00:00, 180.12it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 175.44it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1315)
+[FIX] Re-initialized lora_A/B in 640 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(9, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 9 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6760.247M (0.04%)
+[TRAIN] step_per_epoch=4, eval_steps=20
+  [SRT] Loaded 9 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 9 sigs, cur=task1590_diplomacy_text_generation, mapping={'task1590_diplomacy_text_generation': 0, 'task1572_samsum_summary': 9, 'task363_sst2_polarity_classification': 8, 'task1290_xsum_summarization': 7, 'task181_outcome_extraction': 6, 'task002_quoref_answer_generation': 5, 'task1510_evalution_relation_extraction': 4, 'task639_multi_woz_user_utterance_generation': 3, 'task1729_personachat_generate_next': 2, 'task073_commonsenseqa_answer_generation': 1}
+[SGWI] Task 9: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 01:39:23,886] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/60 [00:00<?, ?it/s]
  2%|▏         | 1/60 [00:06<06:12,  6.31s/it]
  3%|▎         | 2/60 [00:11<05:29,  5.67s/it]
  5%|▌         | 3/60 [00:17<05:19,  5.60s/it]
  7%|▋         | 4/60 [00:21<04:58,  5.33s/it]
  8%|▊         | 5/60 [00:27<04:53,  5.33s/it]
 10%|█         | 6/60 [00:32<04:49,  5.36s/it]
 12%|█▏        | 7/60 [00:38<04:45,  5.38s/it]
 13%|█▎        | 8/60 [00:43<04:33,  5.26s/it]
 15%|█▌        | 9/60 [00:48<04:28,  5.26s/it]
 17%|█▋        | 10/60 [00:53<04:25,  5.30s/it]
                                               
{'loss': '3.525', 'grad_norm': '0.3089', 'learning_rate': '5e-05', 'epoch': '2.508'}
+
 17%|█▋        | 10/60 [00:53<04:25,  5.30s/it]
 18%|█▊        | 11/60 [00:59<04:18,  5.28s/it]
 20%|██        | 12/60 [01:03<04:08,  5.18s/it]
 22%|██▏       | 13/60 [01:09<04:03,  5.18s/it]
 23%|██▎       | 14/60 [01:14<04:01,  5.25s/it]
 25%|██▌       | 15/60 [01:19<03:58,  5.29s/it]
 27%|██▋       | 16/60 [01:24<03:49,  5.21s/it]
 28%|██▊       | 17/60 [01:29<03:40,  5.14s/it]
 30%|███       | 18/60 [01:35<03:36,  5.16s/it]
 32%|███▏      | 19/60 [01:40<03:31,  5.17s/it]
 33%|███▎      | 20/60 [01:45<03:25,  5.14s/it]
                                               
{'loss': '3.125', 'grad_norm': '0.2845', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 20/60 [01:45<03:25,  5.14s/it]
+
Eval @ step 20:   0%|          | 0/16 [00:00<?, ?it/s]
+
Eval @ step 20:   6%|▋         | 1/16 [00:02<00:37,  2.50s/it]
+
Eval @ step 20:  12%|█▎        | 2/16 [00:05<00:35,  2.57s/it]
+
Eval @ step 20:  19%|█▉        | 3/16 [00:08<00:37,  2.87s/it]
+
Eval @ step 20:  25%|██▌       | 4/16 [00:10<00:31,  2.59s/it]
+
Eval @ step 20:  31%|███▏      | 5/16 [00:13<00:30,  2.78s/it]
+
Eval @ step 20:  38%|███▊      | 6/16 [00:15<00:25,  2.56s/it]
+
Eval @ step 20:  44%|████▍     | 7/16 [00:18<00:24,  2.78s/it]
+
Eval @ step 20:  50%|█████     | 8/16 [00:21<00:21,  2.70s/it]
+
Eval @ step 20:  56%|█████▋    | 9/16 [00:23<00:17,  2.54s/it]
+
Eval @ step 20:  62%|██████▎   | 10/16 [00:26<00:16,  2.77s/it]
+
Eval @ step 20:  69%|██████▉   | 11/16 [00:29<00:13,  2.67s/it]
+
Eval @ step 20:  75%|███████▌  | 12/16 [00:33<00:12,  3.03s/it]
+
Eval @ step 20:  81%|████████▏ | 13/16 [00:35<00:08,  2.90s/it]
+
Eval @ step 20:  88%|████████▊ | 14/16 [00:39<00:06,  3.08s/it]
+
Eval @ step 20:  94%|█████████▍| 15/16 [00:41<00:02,  2.86s/it]
+
Eval @ step 20: 100%|██████████| 16/16 [00:43<00:00,  2.65s/it]
+
                                                                 [EVAL] ★ New best eval_rougeL=9.3567 (step=20) — snapshot saved in memory
+
                                               
{'eval_rougeL': '9.357', 'epoch': '5'}
+
 33%|███▎      | 20/60 [02:29<03:25,  5.14s/it]
 35%|███▌      | 21/60 [02:34<11:55, 18.35s/it]
 37%|███▋      | 22/60 [02:39<09:08, 14.42s/it]
 38%|███▊      | 23/60 [02:45<07:11, 11.65s/it]
 40%|████      | 24/60 [02:49<05:47,  9.65s/it]
 42%|████▏     | 25/60 [02:55<04:52,  8.35s/it]
 43%|████▎     | 26/60 [03:00<04:13,  7.47s/it]
 45%|████▌     | 27/60 [03:06<03:48,  6.93s/it]
 47%|████▋     | 28/60 [03:11<03:22,  6.32s/it]
 48%|████▊     | 29/60 [03:16<03:06,  6.02s/it]
 50%|█████     | 30/60 [03:21<02:54,  5.80s/it]
                                               
{'loss': '2.938', 'grad_norm': '0.3535', 'learning_rate': '5e-05', 'epoch': '7.508'}
+
 50%|█████     | 30/60 [03:21<02:54,  5.80s/it]
 52%|█████▏    | 31/60 [03:26<02:38,  5.48s/it]
 53%|█████▎    | 32/60 [03:31<02:28,  5.30s/it]
 55%|█████▌    | 33/60 [03:36<02:20,  5.22s/it]
 57%|█████▋    | 34/60 [03:41<02:13,  5.14s/it]
 58%|█████▊    | 35/60 [03:46<02:08,  5.13s/it]
 60%|██████    | 36/60 [03:51<02:04,  5.17s/it]
 62%|██████▏   | 37/60 [03:56<01:57,  5.09s/it]
 63%|██████▎   | 38/60 [04:02<01:53,  5.14s/it]
 65%|██████▌   | 39/60 [04:07<01:48,  5.18s/it]
 67%|██████▋   | 40/60 [04:12<01:41,  5.09s/it]
                                               
{'loss': '2.798', 'grad_norm': '0.4182', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 40/60 [04:12<01:41,  5.09s/it]
+
Eval @ step 40:   0%|          | 0/16 [00:00<?, ?it/s]
+
Eval @ step 40:   6%|▋         | 1/16 [00:00<00:11,  1.30it/s]
+
Eval @ step 40:  12%|█▎        | 2/16 [00:01<00:06,  2.12it/s]
+
Eval @ step 40:  19%|█▉        | 3/16 [00:01<00:08,  1.49it/s]
+
Eval @ step 40:  25%|██▌       | 4/16 [00:02<00:06,  1.73it/s]
+
Eval @ step 40:  31%|███▏      | 5/16 [00:03<00:09,  1.18it/s]
+
Eval @ step 40:  38%|███▊      | 6/16 [00:04<00:08,  1.15it/s]
+
Eval @ step 40:  44%|████▍     | 7/16 [00:07<00:12,  1.43s/it]
+
Eval @ step 40:  50%|█████     | 8/16 [00:07<00:09,  1.13s/it]
+
Eval @ step 40:  56%|█████▋    | 9/16 [00:09<00:10,  1.46s/it]
+
Eval @ step 40:  62%|██████▎   | 10/16 [00:10<00:07,  1.29s/it]
+
Eval @ step 40:  69%|██████▉   | 11/16 [00:11<00:04,  1.03it/s]
+
Eval @ step 40:  75%|███████▌  | 12/16 [00:11<00:03,  1.07it/s]
+
Eval @ step 40:  81%|████████▏ | 13/16 [00:12<00:02,  1.19it/s]
+
Eval @ step 40:  88%|████████▊ | 14/16 [00:12<00:01,  1.46it/s]
+
Eval @ step 40:  94%|█████████▍| 15/16 [00:13<00:00,  1.41it/s]
+
Eval @ step 40: 100%|██████████| 16/16 [00:13<00:00,  1.78it/s]
+
                                                                 [EVAL] eval_rougeL=8.3955 (best=9.3567)
+
                                               
{'eval_rougeL': '8.396', 'epoch': '10'}
+
 67%|██████▋   | 40/60 [04:25<01:41,  5.09s/it]
 68%|██████▊   | 41/60 [04:31<02:57,  9.33s/it]
 70%|███████   | 42/60 [04:36<02:26,  8.14s/it]
 72%|███████▏  | 43/60 [04:41<02:02,  7.23s/it]
 73%|███████▎  | 44/60 [04:46<01:45,  6.59s/it]
 75%|███████▌  | 45/60 [04:52<01:33,  6.22s/it]
 77%|███████▋  | 46/60 [04:57<01:21,  5.84s/it]
 78%|███████▊  | 47/60 [05:02<01:12,  5.58s/it]
 80%|████████  | 48/60 [05:07<01:05,  5.46s/it]
 82%|████████▏ | 49/60 [05:12<00:59,  5.39s/it]
 83%|████████▎ | 50/60 [05:18<00:54,  5.42s/it]
                                               
{'loss': '2.723', 'grad_norm': '0.372', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 50/60 [05:18<00:54,  5.42s/it]
 85%|████████▌ | 51/60 [05:23<00:47,  5.27s/it]
 87%|████████▋ | 52/60 [05:27<00:41,  5.16s/it]
 88%|████████▊ | 53/60 [05:33<00:35,  5.12s/it]
 90%|█████████ | 54/60 [05:37<00:30,  5.06s/it]
 92%|█████████▏| 55/60 [05:43<00:26,  5.22s/it]
 93%|█████████▎| 56/60 [05:48<00:20,  5.13s/it]
 95%|█████████▌| 57/60 [05:53<00:15,  5.15s/it]
 97%|█████████▋| 58/60 [05:59<00:10,  5.29s/it]
 98%|█████████▊| 59/60 [06:04<00:05,  5.21s/it]
100%|██████████| 60/60 [06:09<00:00,  5.15s/it]
                                               
{'loss': '2.663', 'grad_norm': '0.4802', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 60/60 [06:09<00:00,  5.15s/it]
+
Eval @ step 60:   0%|          | 0/16 [00:00<?, ?it/s]
+
Eval @ step 60:   6%|▋         | 1/16 [00:00<00:09,  1.51it/s]
+
Eval @ step 60:  12%|█▎        | 2/16 [00:02<00:15,  1.09s/it]
+
Eval @ step 60:  19%|█▉        | 3/16 [00:03<00:13,  1.06s/it]
+
Eval @ step 60:  25%|██▌       | 4/16 [00:03<00:09,  1.23it/s]
+
Eval @ step 60:  31%|███▏      | 5/16 [00:04<00:10,  1.03it/s]
+
Eval @ step 60:  38%|███▊      | 6/16 [00:05<00:09,  1.06it/s]
+
Eval @ step 60:  44%|████▍     | 7/16 [00:08<00:13,  1.48s/it]
+
Eval @ step 60:  50%|█████     | 8/16 [00:08<00:09,  1.16s/it]
+
Eval @ step 60:  56%|█████▋    | 9/16 [00:10<00:10,  1.49s/it]
+
Eval @ step 60:  62%|██████▎   | 10/16 [00:11<00:07,  1.21s/it]
+
Eval @ step 60:  69%|██████▉   | 11/16 [00:11<00:04,  1.05it/s]
+
Eval @ step 60:  75%|███████▌  | 12/16 [00:13<00:04,  1.17s/it]
+
Eval @ step 60:  81%|████████▏ | 13/16 [00:14<00:02,  1.04it/s]
+
Eval @ step 60:  88%|████████▊ | 14/16 [00:15<00:02,  1.11s/it]
+
Eval @ step 60:  94%|█████████▍| 15/16 [00:16<00:00,  1.00it/s]
+
Eval @ step 60: 100%|██████████| 16/16 [00:16<00:00,  1.23it/s]
+
                                                                 [EVAL] ★ New best eval_rougeL=11.1335 (step=60) — snapshot saved in memory
+
                                               
{'eval_rougeL': '11.13', 'epoch': '15'}
+
100%|██████████| 60/60 [06:25<00:00,  5.15s/it]
                                               
{'train_runtime': '385.9', 'train_samples_per_second': '4.897', 'train_steps_per_second': '0.155', 'train_loss': '2.962', 'epoch': '15'}
+
100%|██████████| 60/60 [06:25<00:00,  5.15s/it]
100%|██████████| 60/60 [06:25<00:00,  6.43s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 21088624GF
+  train_loss               =     2.9622
+  train_runtime            = 0:06:25.92
+  train_samples            =        126
+  train_samples_per_second =      4.897
+  train_steps_per_second   =      0.155
+  [EVAL] Restored best model (eval_rougeL=11.1335, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:14, 19.96it/s]
Loading weights:  10%|▉         | 28/291 [00:00<00:01, 156.49it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 157.22it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 161.28it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 166.98it/s]
Loading weights:  36%|███▌      | 105/291 [00:00<00:01, 169.81it/s]
Loading weights:  45%|████▍     | 130/291 [00:00<00:00, 178.66it/s]
Loading weights:  51%|█████     | 149/291 [00:00<00:00, 181.34it/s]
Loading weights:  58%|█████▊    | 168/291 [00:00<00:00, 182.59it/s]
Loading weights:  64%|██████▍   | 187/291 [00:01<00:00, 184.31it/s]
Loading weights:  71%|███████   | 206/291 [00:01<00:00, 179.17it/s]
Loading weights:  78%|███████▊  | 227/291 [00:01<00:00, 186.38it/s]
Loading weights:  85%|████████▍ | 246/291 [00:01<00:00, 185.75it/s]
Loading weights:  91%|█████████ | 265/291 [00:01<00:00, 177.00it/s]
Loading weights:  98%|█████████▊| 285/291 [00:01<00:00, 183.37it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 177.27it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (126 samples, batch_size=1)
+  [SRT] Task 'task1590_diplomacy_text_generation': PaR=65.2, shrinkage=ridge, n=126
+  [SRT] Wired router: 10 sigs, cur=task1590_diplomacy_text_generation, mapping={'task1590_diplomacy_text_generation': 0, 'task1572_samsum_summary': 9, 'task363_sst2_polarity_classification': 8, 'task1290_xsum_summarization': 7, 'task181_outcome_extraction': 6, 'task002_quoref_answer_generation': 5, 'task1510_evalution_relation_extraction': 4, 'task639_multi_woz_user_utterance_generation': 3, 'task1729_personachat_generate_next': 2, 'task073_commonsenseqa_answer_generation': 1}
+[SRT] Router: 10 tasks, avg_PaR=65.2, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation']
+  [SRT-INIT] task_id_to_idx={'task1590_diplomacy_text_generation': 0, 'task1572_samsum_summary': 9, 'task363_sst2_polarity_classification': 8, 'task1290_xsum_summarization': 7, 'task181_outcome_extraction': 6, 'task002_quoref_answer_generation': 5, 'task1510_evalution_relation_extraction': 4, 'task639_multi_woz_user_utterance_generation': 3, 'task1729_personachat_generate_next': 2, 'task073_commonsenseqa_answer_generation': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/697 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=10
+
Continual predict:   0%|          | 1/697 [00:01<21:22,  1.84s/it]
Continual predict:   0%|          | 2/697 [00:04<26:32,  2.29s/it]
Continual predict:   0%|          | 3/697 [00:05<18:56,  1.64s/it]
Continual predict:   1%|          | 4/697 [00:05<14:15,  1.23s/it]
Continual predict:   1%|          | 5/697 [00:07<17:00,  1.47s/it]
Continual predict:   1%|          | 6/697 [00:12<30:02,  2.61s/it]
Continual predict:   1%|          | 7/697 [00:16<33:00,  2.87s/it]
Continual predict:   1%|          | 8/697 [00:18<30:24,  2.65s/it]
Continual predict:   1%|▏         | 9/697 [00:20<27:18,  2.38s/it]
Continual predict:   1%|▏         | 10/697 [00:20<21:47,  1.90s/it]
Continual predict:   2%|▏         | 11/697 [00:23<24:42,  2.16s/it]
Continual predict:   2%|▏         | 12/697 [00:26<28:29,  2.50s/it]
Continual predict:   2%|▏         | 13/697 [00:28<24:45,  2.17s/it]
Continual predict:   2%|▏         | 14/697 [00:30<25:50,  2.27s/it]
Continual predict:   2%|▏         | 15/697 [00:33<25:57,  2.28s/it]
Continual predict:   2%|▏         | 16/697 [00:35<27:08,  2.39s/it]
Continual predict:   2%|▏         | 17/697 [00:43<45:49,  4.04s/it]
Continual predict:   3%|▎         | 18/697 [00:45<39:22,  3.48s/it]
Continual predict:   3%|▎         | 19/697 [00:48<36:45,  3.25s/it]
Continual predict:   3%|▎         | 20/697 [00:49<30:33,  2.71s/it]
Continual predict:   3%|▎         | 21/697 [00:50<21:56,  1.95s/it]
Continual predict:   3%|▎         | 22/697 [00:50<15:53,  1.41s/it]
Continual predict:   3%|▎         | 23/697 [00:50<11:41,  1.04s/it]
Continual predict:   3%|▎         | 24/697 [00:50<08:42,  1.29it/s]
Continual predict:   4%|▎         | 25/697 [00:50<06:38,  1.69it/s]
Continual predict:   4%|▎         | 26/697 [00:50<05:11,  2.16it/s]
Continual predict:   4%|▍         | 27/697 [00:51<04:10,  2.68it/s]
Continual predict:   4%|▍         | 28/697 [00:51<03:27,  3.22it/s]
Continual predict:   4%|▍         | 29/697 [00:51<02:57,  3.76it/s]
Continual predict:   4%|▍         | 30/697 [00:51<02:38,  4.20it/s]
Continual predict:   4%|▍         | 31/697 [00:51<02:23,  4.64it/s]
Continual predict:   5%|▍         | 32/697 [00:51<02:12,  5.01it/s]
Continual predict:   5%|▍         | 33/697 [00:52<02:07,  5.21it/s]
Continual predict:   5%|▍         | 34/697 [00:52<02:03,  5.38it/s]
Continual predict:   5%|▌         | 35/697 [00:52<02:00,  5.50it/s]
Continual predict:   5%|▌         | 36/697 [00:52<01:56,  5.68it/s]
Continual predict:   5%|▌         | 37/697 [00:52<01:55,  5.72it/s]
Continual predict:   5%|▌         | 38/697 [00:52<01:53,  5.83it/s]
Continual predict:   6%|▌         | 39/697 [00:53<01:50,  5.93it/s]
Continual predict:   6%|▌         | 40/697 [00:53<01:49,  6.01it/s]
Continual predict:   6%|▌         | 41/697 [00:53<01:50,  5.96it/s]
Continual predict:   6%|▌         | 42/697 [00:53<01:49,  6.01it/s]
Continual predict:   6%|▌         | 43/697 [00:53<01:49,  5.96it/s]
Continual predict:   6%|▋         | 44/697 [00:53<01:50,  5.93it/s]
Continual predict:   6%|▋         | 45/697 [00:54<01:48,  5.99it/s]
Continual predict:   7%|▋         | 46/697 [00:54<01:47,  6.04it/s]
Continual predict:   7%|▋         | 47/697 [00:54<01:47,  6.05it/s]
Continual predict:   7%|▋         | 48/697 [00:54<01:49,  5.95it/s]
Continual predict:   7%|▋         | 49/697 [00:54<01:49,  5.92it/s]
Continual predict:   7%|▋         | 50/697 [00:54<01:53,  5.69it/s]
Continual predict:   7%|▋         | 51/697 [00:55<01:51,  5.80it/s]
Continual predict:   7%|▋         | 52/697 [00:55<01:49,  5.89it/s]
Continual predict:   8%|▊         | 53/697 [00:55<01:47,  5.97it/s]
Continual predict:   8%|▊         | 54/697 [00:55<01:48,  5.92it/s]
Continual predict:   8%|▊         | 55/697 [00:55<01:46,  6.00it/s]
Continual predict:   8%|▊         | 56/697 [00:55<01:47,  5.94it/s]
Continual predict:   8%|▊         | 57/697 [00:56<01:46,  6.00it/s]
Continual predict:   8%|▊         | 58/697 [00:56<01:45,  6.04it/s]
Continual predict:   8%|▊         | 59/697 [00:56<01:44,  6.08it/s]
Continual predict:   9%|▊         | 60/697 [00:56<01:46,  5.98it/s]
Continual predict:   9%|▉         | 61/697 [00:56<01:45,  6.03it/s]
Continual predict:   9%|▉         | 62/697 [00:56<01:44,  6.08it/s]
Continual predict:   9%|▉         | 63/697 [00:57<01:45,  6.01it/s]
Continual predict:   9%|▉         | 64/697 [00:57<01:44,  6.06it/s]
Continual predict:   9%|▉         | 65/697 [00:57<01:43,  6.10it/s]
Continual predict:   9%|▉         | 66/697 [00:57<01:43,  6.12it/s]
Continual predict:  10%|▉         | 67/697 [00:57<01:43,  6.12it/s]
Continual predict:  10%|▉         | 68/697 [00:57<01:42,  6.14it/s]
Continual predict:  10%|▉         | 69/697 [00:58<01:42,  6.15it/s]
Continual predict:  10%|█         | 70/697 [00:58<01:41,  6.16it/s]
Continual predict:  10%|█         | 71/697 [00:58<01:43,  6.06it/s]
Continual predict:  10%|█         | 72/697 [00:58<01:48,  5.76it/s]
Continual predict:  10%|█         | 73/697 [00:58<01:50,  5.64it/s]
Continual predict:  11%|█         | 74/697 [00:58<01:49,  5.69it/s]
Continual predict:  11%|█         | 75/697 [00:59<01:47,  5.80it/s]
Continual predict:  11%|█         | 76/697 [00:59<01:45,  5.91it/s]
Continual predict:  11%|█         | 77/697 [00:59<01:43,  5.98it/s]
Continual predict:  11%|█         | 78/697 [00:59<01:42,  6.04it/s]
Continual predict:  11%|█▏        | 79/697 [00:59<01:42,  6.06it/s]
Continual predict:  11%|█▏        | 80/697 [00:59<01:41,  6.10it/s]
Continual predict:  12%|█▏        | 81/697 [01:00<01:40,  6.11it/s]
Continual predict:  12%|█▏        | 82/697 [01:00<01:40,  6.12it/s]
Continual predict:  12%|█▏        | 83/697 [01:00<01:40,  6.13it/s]
Continual predict:  12%|█▏        | 84/697 [01:00<01:41,  6.03it/s]
Continual predict:  12%|█▏        | 85/697 [01:00<01:41,  6.05it/s]
Continual predict:  12%|█▏        | 86/697 [01:00<01:40,  6.06it/s]
Continual predict:  12%|█▏        | 87/697 [01:01<01:40,  6.06it/s]
Continual predict:  13%|█▎        | 88/697 [01:01<01:39,  6.10it/s]
Continual predict:  13%|█▎        | 89/697 [01:01<01:39,  6.12it/s]
Continual predict:  13%|█▎        | 90/697 [01:01<01:38,  6.15it/s]
Continual predict:  13%|█▎        | 91/697 [01:01<01:38,  6.15it/s]
Continual predict:  13%|█▎        | 92/697 [01:01<01:38,  6.12it/s]
Continual predict:  13%|█▎        | 93/697 [01:02<01:40,  6.02it/s]
Continual predict:  13%|█▎        | 94/697 [01:02<01:40,  5.97it/s]
Continual predict:  14%|█▎        | 95/697 [01:02<01:39,  6.02it/s]
Continual predict:  14%|█▍        | 96/697 [01:02<01:39,  6.05it/s]
Continual predict:  14%|█▍        | 97/697 [01:02<01:38,  6.08it/s]
Continual predict:  14%|█▍        | 98/697 [01:02<01:39,  6.00it/s]
Continual predict:  14%|█▍        | 99/697 [01:03<01:40,  5.95it/s]
Continual predict:  14%|█▍        | 100/697 [01:03<01:40,  5.92it/s]
Continual predict:  14%|█▍        | 101/697 [01:03<01:39,  5.99it/s]
Continual predict:  15%|█▍        | 102/697 [01:03<01:38,  6.05it/s]
Continual predict:  15%|█▍        | 103/697 [01:03<01:37,  6.09it/s]
Continual predict:  15%|█▍        | 104/697 [01:03<01:37,  6.10it/s]
Continual predict:  15%|█▌        | 105/697 [01:04<01:38,  6.02it/s]
Continual predict:  15%|█▌        | 106/697 [01:04<01:37,  6.07it/s]
Continual predict:  15%|█▌        | 107/697 [01:04<01:36,  6.11it/s]
Continual predict:  15%|█▌        | 108/697 [01:04<01:36,  6.13it/s]
Continual predict:  16%|█▌        | 109/697 [01:04<01:35,  6.15it/s]
Continual predict:  16%|█▌        | 110/697 [01:04<01:35,  6.14it/s]
Continual predict:  16%|█▌        | 111/697 [01:05<01:35,  6.13it/s]
Continual predict:  16%|█▌        | 112/697 [01:05<01:35,  6.13it/s]
Continual predict:  16%|█▌        | 113/697 [01:05<01:35,  6.13it/s]
Continual predict:  16%|█▋        | 114/697 [01:05<01:35,  6.13it/s]
Continual predict:  16%|█▋        | 115/697 [01:05<01:36,  6.04it/s]
Continual predict:  17%|█▋        | 116/697 [01:05<01:35,  6.08it/s]
Continual predict:  17%|█▋        | 117/697 [01:06<01:36,  6.00it/s]
Continual predict:  17%|█▋        | 118/697 [01:06<01:37,  5.94it/s]
Continual predict:  17%|█▋        | 119/697 [01:06<01:36,  6.00it/s]
Continual predict:  17%|█▋        | 120/697 [01:06<01:35,  6.05it/s]
Continual predict:  17%|█▋        | 121/697 [01:11<14:38,  1.53s/it]
Continual predict:  18%|█▊        | 122/697 [01:15<22:45,  2.38s/it]
Continual predict:  18%|█▊        | 123/697 [01:21<31:40,  3.31s/it]
Continual predict:  18%|█▊        | 124/697 [01:23<29:12,  3.06s/it]
Continual predict:  18%|█▊        | 125/697 [01:28<34:34,  3.63s/it]
Continual predict:  18%|█▊        | 126/697 [01:33<38:16,  4.02s/it]
Continual predict:  18%|█▊        | 127/697 [01:35<31:13,  3.29s/it]
Continual predict:  18%|█▊        | 128/697 [01:37<30:00,  3.16s/it]
Continual predict:  19%|█▊        | 129/697 [01:41<30:59,  3.27s/it]
Continual predict:  19%|█▊        | 130/697 [01:44<30:37,  3.24s/it]
Continual predict:  19%|█▉        | 131/697 [01:46<28:05,  2.98s/it]
Continual predict:  19%|█▉        | 132/697 [01:49<27:31,  2.92s/it]
Continual predict:  19%|█▉        | 133/697 [01:53<28:31,  3.03s/it]
Continual predict:  19%|█▉        | 134/697 [01:59<37:31,  4.00s/it]
Continual predict:  19%|█▉        | 135/697 [02:04<40:15,  4.30s/it]
Continual predict:  20%|█▉        | 136/697 [02:07<37:39,  4.03s/it]
Continual predict:  20%|█▉        | 137/697 [02:10<32:52,  3.52s/it]
Continual predict:  20%|█▉        | 138/697 [02:10<25:13,  2.71s/it]
Continual predict:  20%|█▉        | 139/697 [02:12<21:05,  2.27s/it]
Continual predict:  20%|██        | 140/697 [02:14<22:32,  2.43s/it]
Continual predict:  20%|██        | 141/697 [02:21<34:37,  3.74s/it]
Continual predict:  20%|██        | 142/697 [02:26<37:19,  4.03s/it]
Continual predict:  21%|██        | 143/697 [02:29<35:29,  3.84s/it]
Continual predict:  21%|██        | 144/697 [02:39<51:40,  5.61s/it]
Continual predict:  21%|██        | 145/697 [02:44<49:04,  5.33s/it]
Continual predict:  21%|██        | 146/697 [02:46<39:26,  4.29s/it]
Continual predict:  21%|██        | 147/697 [02:50<40:02,  4.37s/it]
Continual predict:  21%|██        | 148/697 [02:52<34:09,  3.73s/it]
Continual predict:  21%|██▏       | 149/697 [02:56<32:56,  3.61s/it]
Continual predict:  22%|██▏       | 150/697 [02:58<28:29,  3.13s/it]
Continual predict:  22%|██▏       | 151/697 [03:01<28:31,  3.13s/it]
Continual predict:  22%|██▏       | 152/697 [03:07<35:57,  3.96s/it]
Continual predict:  22%|██▏       | 153/697 [03:09<32:07,  3.54s/it]
Continual predict:  22%|██▏       | 154/697 [03:14<35:57,  3.97s/it]
Continual predict:  22%|██▏       | 155/697 [03:18<35:18,  3.91s/it]
Continual predict:  22%|██▏       | 156/697 [03:22<34:13,  3.80s/it]
Continual predict:  23%|██▎       | 157/697 [03:26<34:44,  3.86s/it]
Continual predict:  23%|██▎       | 158/697 [03:27<27:07,  3.02s/it]
Continual predict:  23%|██▎       | 159/697 [03:34<38:13,  4.26s/it]
Continual predict:  23%|██▎       | 160/697 [03:36<33:05,  3.70s/it]
Continual predict:  23%|██▎       | 161/697 [03:39<30:02,  3.36s/it]
Continual predict:  23%|██▎       | 162/697 [03:44<34:16,  3.84s/it]
Continual predict:  23%|██▎       | 163/697 [03:46<29:44,  3.34s/it]
Continual predict:  24%|██▎       | 164/697 [03:48<27:07,  3.05s/it]
Continual predict:  24%|██▎       | 165/697 [03:52<27:38,  3.12s/it]
Continual predict:  24%|██▍       | 166/697 [03:54<25:17,  2.86s/it]
Continual predict:  24%|██▍       | 167/697 [03:55<21:15,  2.41s/it]
Continual predict:  24%|██▍       | 168/697 [03:58<21:04,  2.39s/it]
Continual predict:  24%|██▍       | 169/697 [03:59<19:29,  2.21s/it]
Continual predict:  24%|██▍       | 170/697 [04:02<21:57,  2.50s/it]
Continual predict:  25%|██▍       | 171/697 [04:04<19:37,  2.24s/it]
Continual predict:  25%|██▍       | 172/697 [04:07<20:48,  2.38s/it]
Continual predict:  25%|██▍       | 173/697 [04:11<25:10,  2.88s/it]
Continual predict:  25%|██▍       | 174/697 [04:12<20:48,  2.39s/it]
Continual predict:  25%|██▌       | 175/697 [04:16<23:57,  2.75s/it]
Continual predict:  25%|██▌       | 176/697 [04:19<24:32,  2.83s/it]
Continual predict:  25%|██▌       | 177/697 [04:22<24:31,  2.83s/it]
Continual predict:  26%|██▌       | 178/697 [04:23<20:35,  2.38s/it]
Continual predict:  26%|██▌       | 179/697 [04:24<17:27,  2.02s/it]
Continual predict:  26%|██▌       | 180/697 [04:25<15:41,  1.82s/it]
Continual predict:  26%|██▌       | 181/697 [04:27<14:27,  1.68s/it]
Continual predict:  26%|██▌       | 182/697 [04:32<22:36,  2.63s/it]
Continual predict:  26%|██▋       | 183/697 [04:32<16:22,  1.91s/it]
Continual predict:  26%|██▋       | 184/697 [04:33<14:19,  1.68s/it]
Continual predict:  27%|██▋       | 185/697 [04:36<18:59,  2.22s/it]
Continual predict:  27%|██▋       | 186/697 [04:38<15:58,  1.88s/it]
Continual predict:  27%|██▋       | 187/697 [04:41<19:45,  2.33s/it]
Continual predict:  27%|██▋       | 188/697 [04:44<20:54,  2.47s/it]
Continual predict:  27%|██▋       | 189/697 [04:49<27:41,  3.27s/it]
Continual predict:  27%|██▋       | 190/697 [04:52<28:00,  3.31s/it]
Continual predict:  27%|██▋       | 191/697 [04:54<23:02,  2.73s/it]
Continual predict:  28%|██▊       | 192/697 [04:57<24:47,  2.94s/it]
Continual predict:  28%|██▊       | 193/697 [05:02<29:10,  3.47s/it]
Continual predict:  28%|██▊       | 194/697 [05:07<32:52,  3.92s/it]
Continual predict:  28%|██▊       | 195/697 [05:12<36:47,  4.40s/it]
Continual predict:  28%|██▊       | 196/697 [05:16<34:19,  4.11s/it]
Continual predict:  28%|██▊       | 197/697 [05:19<32:27,  3.90s/it]
Continual predict:  28%|██▊       | 198/697 [05:27<41:42,  5.01s/it]
Continual predict:  29%|██▊       | 199/697 [05:28<31:56,  3.85s/it]
Continual predict:  29%|██▊       | 200/697 [05:29<26:06,  3.15s/it]
Continual predict:  29%|██▉       | 201/697 [05:32<24:46,  3.00s/it]
Continual predict:  29%|██▉       | 202/697 [05:36<25:57,  3.15s/it]
Continual predict:  29%|██▉       | 203/697 [05:41<31:12,  3.79s/it]
Continual predict:  29%|██▉       | 204/697 [05:44<29:42,  3.61s/it]
Continual predict:  29%|██▉       | 205/697 [05:51<37:26,  4.57s/it]
Continual predict:  30%|██▉       | 206/697 [05:54<33:07,  4.05s/it]
Continual predict:  30%|██▉       | 207/697 [05:57<32:32,  3.98s/it]
Continual predict:  30%|██▉       | 208/697 [06:04<38:13,  4.69s/it]
Continual predict:  30%|██▉       | 209/697 [06:07<33:29,  4.12s/it]
Continual predict:  30%|███       | 210/697 [06:08<27:33,  3.40s/it]
Continual predict:  30%|███       | 211/697 [06:11<25:49,  3.19s/it]
Continual predict:  30%|███       | 212/697 [06:14<25:52,  3.20s/it]
Continual predict:  31%|███       | 213/697 [06:18<27:31,  3.41s/it]
Continual predict:  31%|███       | 214/697 [06:22<28:22,  3.53s/it]
Continual predict:  31%|███       | 215/697 [06:23<21:50,  2.72s/it]
Continual predict:  31%|███       | 216/697 [06:26<22:57,  2.86s/it]
Continual predict:  31%|███       | 217/697 [06:28<21:17,  2.66s/it]
Continual predict:  31%|███▏      | 218/697 [06:32<24:50,  3.11s/it]
Continual predict:  31%|███▏      | 219/697 [06:36<25:23,  3.19s/it]
Continual predict:  32%|███▏      | 220/697 [06:40<27:51,  3.51s/it]
Continual predict:  32%|███▏      | 221/697 [06:40<19:59,  2.52s/it]
Continual predict:  32%|███▏      | 222/697 [06:40<14:39,  1.85s/it]
Continual predict:  32%|███▏      | 223/697 [06:41<10:45,  1.36s/it]
Continual predict:  32%|███▏      | 224/697 [06:41<08:02,  1.02s/it]
Continual predict:  32%|███▏      | 225/697 [06:41<06:08,  1.28it/s]
Continual predict:  32%|███▏      | 226/697 [06:41<04:48,  1.64it/s]
Continual predict:  33%|███▎      | 227/697 [06:42<03:59,  1.96it/s]
Continual predict:  33%|███▎      | 228/697 [06:42<03:25,  2.28it/s]
Continual predict:  33%|███▎      | 229/697 [06:42<02:57,  2.64it/s]
Continual predict:  33%|███▎      | 230/697 [06:43<03:05,  2.52it/s]
Continual predict:  33%|███▎      | 231/697 [06:43<02:40,  2.91it/s]
Continual predict:  33%|███▎      | 232/697 [06:43<02:22,  3.26it/s]
Continual predict:  33%|███▎      | 233/697 [06:43<02:09,  3.57it/s]
Continual predict:  34%|███▎      | 234/697 [06:43<02:01,  3.82it/s]
Continual predict:  34%|███▎      | 235/697 [06:44<01:54,  4.02it/s]
Continual predict:  34%|███▍      | 236/697 [06:44<01:50,  4.17it/s]
Continual predict:  34%|███▍      | 237/697 [06:44<01:47,  4.29it/s]
Continual predict:  34%|███▍      | 238/697 [06:44<01:45,  4.37it/s]
Continual predict:  34%|███▍      | 239/697 [06:45<01:43,  4.43it/s]
Continual predict:  34%|███▍      | 240/697 [06:45<02:02,  3.73it/s]
Continual predict:  35%|███▍      | 241/697 [06:45<01:55,  3.95it/s]
Continual predict:  35%|███▍      | 242/697 [06:45<01:50,  4.12it/s]
Continual predict:  35%|███▍      | 243/697 [06:46<01:46,  4.25it/s]
Continual predict:  35%|███▌      | 244/697 [06:46<01:44,  4.34it/s]
Continual predict:  35%|███▌      | 245/697 [06:46<01:42,  4.41it/s]
Continual predict:  35%|███▌      | 246/697 [06:46<01:41,  4.45it/s]
Continual predict:  35%|███▌      | 247/697 [06:46<01:47,  4.18it/s]
Continual predict:  36%|███▌      | 248/697 [06:47<01:44,  4.28it/s]
Continual predict:  36%|███▌      | 249/697 [06:47<01:49,  4.08it/s]
Continual predict:  36%|███▌      | 250/697 [06:47<01:45,  4.22it/s]
Continual predict:  36%|███▌      | 251/697 [06:47<01:50,  4.04it/s]
Continual predict:  36%|███▌      | 252/697 [06:49<04:07,  1.80it/s]
Continual predict:  36%|███▋      | 253/697 [06:49<03:29,  2.12it/s]
Continual predict:  36%|███▋      | 254/697 [06:49<02:55,  2.53it/s]
Continual predict:  37%|███▋      | 255/697 [06:49<02:31,  2.91it/s]
Continual predict:  37%|███▋      | 256/697 [06:50<02:15,  3.26it/s]
Continual predict:  37%|███▋      | 257/697 [06:50<03:14,  2.26it/s]
Continual predict:  37%|███▋      | 258/697 [06:51<02:44,  2.67it/s]
Continual predict:  37%|███▋      | 259/697 [06:51<02:23,  3.05it/s]
Continual predict:  37%|███▋      | 260/697 [06:51<02:09,  3.38it/s]
Continual predict:  37%|███▋      | 261/697 [06:52<03:21,  2.17it/s]
Continual predict:  38%|███▊      | 262/697 [06:52<02:51,  2.53it/s]
Continual predict:  38%|███▊      | 263/697 [06:52<02:35,  2.80it/s]
Continual predict:  38%|███▊      | 264/697 [06:54<04:32,  1.59it/s]
Continual predict:  38%|███▊      | 265/697 [06:54<04:22,  1.64it/s]
Continual predict:  38%|███▊      | 266/697 [06:55<04:15,  1.69it/s]
Continual predict:  38%|███▊      | 267/697 [06:55<04:18,  1.66it/s]
Continual predict:  38%|███▊      | 268/697 [06:56<04:50,  1.48it/s]
Continual predict:  39%|███▊      | 269/697 [06:57<04:42,  1.51it/s]
Continual predict:  39%|███▊      | 270/697 [06:58<04:42,  1.51it/s]
Continual predict:  39%|███▉      | 271/697 [06:58<04:58,  1.43it/s]
Continual predict:  39%|███▉      | 272/697 [06:59<05:27,  1.30it/s]
Continual predict:  39%|███▉      | 273/697 [07:00<05:12,  1.36it/s]
Continual predict:  39%|███▉      | 274/697 [07:01<05:03,  1.39it/s]
Continual predict:  39%|███▉      | 275/697 [07:01<04:50,  1.45it/s]
Continual predict:  40%|███▉      | 276/697 [07:02<05:40,  1.24it/s]
Continual predict:  40%|███▉      | 277/697 [07:04<06:43,  1.04it/s]
Continual predict:  40%|███▉      | 278/697 [07:05<06:28,  1.08it/s]
Continual predict:  40%|████      | 279/697 [07:05<06:03,  1.15it/s]
Continual predict:  40%|████      | 280/697 [07:06<05:11,  1.34it/s]
Continual predict:  40%|████      | 281/697 [07:07<05:40,  1.22it/s]
Continual predict:  40%|████      | 282/697 [07:08<05:41,  1.22it/s]
Continual predict:  41%|████      | 283/697 [07:08<05:24,  1.28it/s]
Continual predict:  41%|████      | 284/697 [07:09<04:44,  1.45it/s]
Continual predict:  41%|████      | 285/697 [07:10<04:59,  1.38it/s]
Continual predict:  41%|████      | 286/697 [07:10<05:05,  1.34it/s]
Continual predict:  41%|████      | 287/697 [07:11<05:09,  1.32it/s]
Continual predict:  41%|████▏     | 288/697 [07:12<04:56,  1.38it/s]
Continual predict:  41%|████▏     | 289/697 [07:13<05:25,  1.25it/s]
Continual predict:  42%|████▏     | 290/697 [07:13<05:23,  1.26it/s]
Continual predict:  42%|████▏     | 291/697 [07:14<05:01,  1.34it/s]
Continual predict:  42%|████▏     | 292/697 [07:14<04:10,  1.62it/s]
Continual predict:  42%|████▏     | 293/697 [07:15<04:38,  1.45it/s]
Continual predict:  42%|████▏     | 294/697 [07:16<04:21,  1.54it/s]
Continual predict:  42%|████▏     | 295/697 [07:16<03:57,  1.69it/s]
Continual predict:  42%|████▏     | 296/697 [07:17<04:46,  1.40it/s]
Continual predict:  43%|████▎     | 297/697 [07:18<04:24,  1.51it/s]
Continual predict:  43%|████▎     | 298/697 [07:18<04:01,  1.65it/s]
Continual predict:  43%|████▎     | 299/697 [07:19<04:40,  1.42it/s]
Continual predict:  43%|████▎     | 300/697 [07:20<03:55,  1.69it/s]
Continual predict:  43%|████▎     | 301/697 [07:21<04:35,  1.44it/s]
Continual predict:  43%|████▎     | 302/697 [07:21<04:04,  1.61it/s]
Continual predict:  43%|████▎     | 303/697 [07:23<05:55,  1.11it/s]
Continual predict:  44%|████▎     | 304/697 [07:23<04:47,  1.37it/s]
Continual predict:  44%|████▍     | 305/697 [07:24<05:10,  1.26it/s]
Continual predict:  44%|████▍     | 306/697 [07:25<05:14,  1.24it/s]
Continual predict:  44%|████▍     | 307/697 [07:25<05:15,  1.24it/s]
Continual predict:  44%|████▍     | 308/697 [07:26<04:32,  1.43it/s]
Continual predict:  44%|████▍     | 309/697 [07:26<04:04,  1.59it/s]
Continual predict:  44%|████▍     | 310/697 [07:27<04:03,  1.59it/s]
Continual predict:  45%|████▍     | 311/697 [07:28<03:54,  1.65it/s]
Continual predict:  45%|████▍     | 312/697 [07:28<04:08,  1.55it/s]
Continual predict:  45%|████▍     | 313/697 [07:29<04:38,  1.38it/s]
Continual predict:  45%|████▌     | 314/697 [07:30<04:08,  1.54it/s]
Continual predict:  45%|████▌     | 315/697 [07:31<04:39,  1.37it/s]
Continual predict:  45%|████▌     | 316/697 [07:31<04:49,  1.31it/s]
Continual predict:  45%|████▌     | 317/697 [07:32<04:00,  1.58it/s]
Continual predict:  46%|████▌     | 318/697 [07:32<03:40,  1.72it/s]
Continual predict:  46%|████▌     | 319/697 [07:34<05:42,  1.11it/s]
Continual predict:  46%|████▌     | 320/697 [07:35<05:20,  1.17it/s]
Continual predict:  46%|████▌     | 321/697 [07:35<05:05,  1.23it/s]
Continual predict:  46%|████▌     | 322/697 [07:36<04:22,  1.43it/s]
Continual predict:  46%|████▋     | 323/697 [07:37<04:30,  1.38it/s]
Continual predict:  46%|████▋     | 324/697 [07:37<04:51,  1.28it/s]
Continual predict:  47%|████▋     | 325/697 [07:38<04:51,  1.28it/s]
Continual predict:  47%|████▋     | 326/697 [07:39<04:35,  1.35it/s]
Continual predict:  47%|████▋     | 327/697 [07:40<05:20,  1.15it/s]
Continual predict:  47%|████▋     | 328/697 [07:41<05:44,  1.07it/s]
Continual predict:  47%|████▋     | 329/697 [07:42<05:00,  1.22it/s]
Continual predict:  47%|████▋     | 330/697 [07:42<04:06,  1.49it/s]
Continual predict:  47%|████▋     | 331/697 [07:43<04:00,  1.52it/s]
Continual predict:  48%|████▊     | 332/697 [07:43<03:49,  1.59it/s]
Continual predict:  48%|████▊     | 333/697 [07:44<04:56,  1.23it/s]
Continual predict:  48%|████▊     | 334/697 [07:45<05:07,  1.18it/s]
Continual predict:  48%|████▊     | 335/697 [07:46<04:23,  1.38it/s]
Continual predict:  48%|████▊     | 336/697 [07:46<03:54,  1.54it/s]
Continual predict:  48%|████▊     | 337/697 [07:47<03:46,  1.59it/s]
Continual predict:  48%|████▊     | 338/697 [07:48<03:49,  1.56it/s]
Continual predict:  49%|████▊     | 339/697 [07:48<03:39,  1.63it/s]
Continual predict:  49%|████▉     | 340/697 [07:49<04:31,  1.31it/s]
Continual predict:  49%|████▉     | 341/697 [07:50<04:32,  1.30it/s]
Continual predict:  49%|████▉     | 342/697 [07:50<04:00,  1.48it/s]
Continual predict:  49%|████▉     | 343/697 [07:51<04:10,  1.41it/s]
Continual predict:  49%|████▉     | 344/697 [07:52<04:38,  1.27it/s]
Continual predict:  49%|████▉     | 345/697 [07:53<04:03,  1.44it/s]
Continual predict:  50%|████▉     | 346/697 [07:53<04:17,  1.36it/s]
Continual predict:  50%|████▉     | 347/697 [07:55<05:36,  1.04it/s]
Continual predict:  50%|████▉     | 348/697 [07:55<04:31,  1.28it/s]
Continual predict:  50%|█████     | 349/697 [07:56<03:56,  1.47it/s]
Continual predict:  50%|█████     | 350/697 [07:56<03:54,  1.48it/s]
Continual predict:  50%|█████     | 351/697 [07:58<04:42,  1.22it/s]
Continual predict:  51%|█████     | 352/697 [07:58<04:37,  1.24it/s]
Continual predict:  51%|█████     | 353/697 [08:00<05:30,  1.04it/s]
Continual predict:  51%|█████     | 354/697 [08:01<06:03,  1.06s/it]
Continual predict:  51%|█████     | 355/697 [08:02<05:18,  1.08it/s]
Continual predict:  51%|█████     | 356/697 [08:03<05:17,  1.07it/s]
Continual predict:  51%|█████     | 357/697 [08:04<05:33,  1.02it/s]
Continual predict:  51%|█████▏    | 358/697 [08:04<05:11,  1.09it/s]
Continual predict:  52%|█████▏    | 359/697 [08:05<04:43,  1.19it/s]
Continual predict:  52%|█████▏    | 360/697 [08:06<05:28,  1.03it/s]
Continual predict:  52%|█████▏    | 361/697 [08:08<06:32,  1.17s/it]
Continual predict:  52%|█████▏    | 362/697 [08:09<05:35,  1.00s/it]
Continual predict:  52%|█████▏    | 363/697 [08:09<04:41,  1.19it/s]
Continual predict:  52%|█████▏    | 364/697 [08:09<03:50,  1.44it/s]
Continual predict:  52%|█████▏    | 365/697 [08:10<03:09,  1.75it/s]
Continual predict:  53%|█████▎    | 366/697 [08:10<02:40,  2.06it/s]
Continual predict:  53%|█████▎    | 367/697 [08:10<02:20,  2.35it/s]
Continual predict:  53%|█████▎    | 368/697 [08:11<02:06,  2.60it/s]
Continual predict:  53%|█████▎    | 369/697 [08:11<02:04,  2.64it/s]
Continual predict:  53%|█████▎    | 370/697 [08:11<02:00,  2.72it/s]
Continual predict:  53%|█████▎    | 371/697 [08:12<01:58,  2.76it/s]
Continual predict:  53%|█████▎    | 372/697 [08:12<02:02,  2.65it/s]
Continual predict:  54%|█████▎    | 373/697 [08:12<02:04,  2.60it/s]
Continual predict:  54%|█████▎    | 374/697 [08:13<01:54,  2.82it/s]
Continual predict:  54%|█████▍    | 375/697 [08:13<01:59,  2.70it/s]
Continual predict:  54%|█████▍    | 376/697 [08:13<01:56,  2.76it/s]
Continual predict:  54%|█████▍    | 377/697 [08:14<01:48,  2.95it/s]
Continual predict:  54%|█████▍    | 378/697 [08:14<01:43,  3.08it/s]
Continual predict:  54%|█████▍    | 379/697 [08:14<01:39,  3.19it/s]
Continual predict:  55%|█████▍    | 380/697 [08:15<01:36,  3.28it/s]
Continual predict:  55%|█████▍    | 381/697 [08:15<01:40,  3.16it/s]
Continual predict:  55%|█████▍    | 382/697 [08:15<01:36,  3.25it/s]
Continual predict:  55%|█████▍    | 383/697 [08:16<01:34,  3.31it/s]
Continual predict:  55%|█████▌    | 384/697 [08:16<01:33,  3.36it/s]
Continual predict:  55%|█████▌    | 385/697 [08:16<01:31,  3.39it/s]
Continual predict:  55%|█████▌    | 386/697 [08:16<01:30,  3.42it/s]
Continual predict:  56%|█████▌    | 387/697 [08:17<01:30,  3.44it/s]
Continual predict:  56%|█████▌    | 388/697 [08:17<01:29,  3.45it/s]
Continual predict:  56%|█████▌    | 389/697 [08:17<01:34,  3.26it/s]
Continual predict:  56%|█████▌    | 390/697 [08:18<01:32,  3.32it/s]
Continual predict:  56%|█████▌    | 391/697 [08:18<01:30,  3.37it/s]
Continual predict:  56%|█████▌    | 392/697 [08:18<01:37,  3.12it/s]
Continual predict:  56%|█████▋    | 393/697 [08:19<01:34,  3.22it/s]
Continual predict:  57%|█████▋    | 394/697 [08:19<01:31,  3.30it/s]
Continual predict:  57%|█████▋    | 395/697 [08:19<01:24,  3.56it/s]
Continual predict:  57%|█████▋    | 396/697 [08:19<01:35,  3.15it/s]
Continual predict:  57%|█████▋    | 397/697 [08:20<01:32,  3.25it/s]
Continual predict:  57%|█████▋    | 398/697 [08:20<01:30,  3.31it/s]
Continual predict:  57%|█████▋    | 399/697 [08:20<01:28,  3.36it/s]
Continual predict:  57%|█████▋    | 400/697 [08:21<01:27,  3.39it/s]
Continual predict:  58%|█████▊    | 401/697 [08:21<01:26,  3.42it/s]
Continual predict:  58%|█████▊    | 402/697 [08:21<01:25,  3.44it/s]
Continual predict:  58%|█████▊    | 403/697 [08:22<01:44,  2.82it/s]
Continual predict:  58%|█████▊    | 404/697 [08:22<01:38,  2.99it/s]
Continual predict:  58%|█████▊    | 405/697 [08:22<01:43,  2.82it/s]
Continual predict:  58%|█████▊    | 406/697 [08:23<01:37,  2.99it/s]
Continual predict:  58%|█████▊    | 407/697 [08:23<01:33,  3.12it/s]
Continual predict:  59%|█████▊    | 408/697 [08:23<01:29,  3.22it/s]
Continual predict:  59%|█████▊    | 409/697 [08:24<01:27,  3.29it/s]
Continual predict:  59%|█████▉    | 410/697 [08:24<01:35,  3.00it/s]
Continual predict:  59%|█████▉    | 411/697 [08:24<01:41,  2.83it/s]
Continual predict:  59%|█████▉    | 412/697 [08:25<01:35,  2.99it/s]
Continual predict:  59%|█████▉    | 413/697 [08:25<01:31,  3.11it/s]
Continual predict:  59%|█████▉    | 414/697 [08:25<01:28,  3.21it/s]
Continual predict:  60%|█████▉    | 415/697 [08:25<01:25,  3.28it/s]
Continual predict:  60%|█████▉    | 416/697 [08:26<01:24,  3.34it/s]
Continual predict:  60%|█████▉    | 417/697 [08:26<01:22,  3.39it/s]
Continual predict:  60%|█████▉    | 418/697 [08:26<01:26,  3.22it/s]
Continual predict:  60%|██████    | 419/697 [08:27<01:33,  2.96it/s]
Continual predict:  60%|██████    | 420/697 [08:27<01:39,  2.80it/s]
Continual predict:  60%|██████    | 421/697 [08:28<01:42,  2.70it/s]
Continual predict:  61%|██████    | 422/697 [08:28<01:39,  2.75it/s]
Continual predict:  61%|██████    | 423/697 [08:28<01:40,  2.73it/s]
Continual predict:  61%|██████    | 424/697 [08:29<01:43,  2.65it/s]
Continual predict:  61%|██████    | 425/697 [08:29<01:44,  2.60it/s]
Continual predict:  61%|██████    | 426/697 [08:30<01:45,  2.56it/s]
Continual predict:  61%|██████▏   | 427/697 [08:30<01:41,  2.65it/s]
Continual predict:  61%|██████▏   | 428/697 [08:30<01:34,  2.86it/s]
Continual predict:  62%|██████▏   | 429/697 [08:30<01:28,  3.02it/s]
Continual predict:  62%|██████▏   | 430/697 [08:31<01:24,  3.14it/s]
Continual predict:  62%|██████▏   | 431/697 [08:31<01:26,  3.06it/s]
Continual predict:  62%|██████▏   | 432/697 [08:31<01:23,  3.18it/s]
Continual predict:  62%|██████▏   | 433/697 [08:32<01:20,  3.27it/s]
Continual predict:  62%|██████▏   | 434/697 [08:32<01:25,  3.07it/s]
Continual predict:  62%|██████▏   | 435/697 [08:32<01:23,  3.15it/s]
Continual predict:  63%|██████▎   | 436/697 [08:33<01:29,  2.91it/s]
Continual predict:  63%|██████▎   | 437/697 [08:33<01:38,  2.64it/s]
Continual predict:  63%|██████▎   | 438/697 [08:34<01:30,  2.85it/s]
Continual predict:  63%|██████▎   | 439/697 [08:34<01:30,  2.86it/s]
Continual predict:  63%|██████▎   | 440/697 [08:34<01:29,  2.87it/s]
Continual predict:  63%|██████▎   | 441/697 [08:35<01:28,  2.88it/s]
Continual predict:  63%|██████▎   | 442/697 [08:35<01:23,  3.04it/s]
Continual predict:  64%|██████▎   | 443/697 [08:35<01:20,  3.16it/s]
Continual predict:  64%|██████▎   | 444/697 [08:35<01:17,  3.25it/s]
Continual predict:  64%|██████▍   | 445/697 [08:36<01:15,  3.32it/s]
Continual predict:  64%|██████▍   | 446/697 [08:36<01:18,  3.18it/s]
Continual predict:  64%|██████▍   | 447/697 [08:36<01:20,  3.09it/s]
Continual predict:  64%|██████▍   | 448/697 [08:37<01:22,  3.03it/s]
Continual predict:  64%|██████▍   | 449/697 [08:37<01:27,  2.85it/s]
Continual predict:  65%|██████▍   | 450/697 [08:37<01:22,  3.01it/s]
Continual predict:  65%|██████▍   | 451/697 [08:38<01:18,  3.14it/s]
Continual predict:  65%|██████▍   | 452/697 [08:38<01:15,  3.23it/s]
Continual predict:  65%|██████▍   | 453/697 [08:38<01:13,  3.30it/s]
Continual predict:  65%|██████▌   | 454/697 [08:39<01:12,  3.35it/s]
Continual predict:  65%|██████▌   | 455/697 [08:39<01:11,  3.39it/s]
Continual predict:  65%|██████▌   | 456/697 [08:39<01:14,  3.22it/s]
Continual predict:  66%|██████▌   | 457/697 [08:40<01:16,  3.12it/s]
Continual predict:  66%|██████▌   | 458/697 [08:40<01:14,  3.22it/s]
Continual predict:  66%|██████▌   | 459/697 [08:40<01:12,  3.29it/s]
Continual predict:  66%|██████▌   | 460/697 [08:40<01:10,  3.34it/s]
Continual predict:  66%|██████▌   | 461/697 [08:41<01:09,  3.38it/s]
Continual predict:  66%|██████▋   | 462/697 [08:41<01:08,  3.41it/s]
Continual predict:  66%|██████▋   | 463/697 [08:41<01:08,  3.43it/s]
Continual predict:  67%|██████▋   | 464/697 [08:42<01:13,  3.16it/s]
Continual predict:  67%|██████▋   | 465/697 [08:42<01:06,  3.49it/s]
Continual predict:  67%|██████▋   | 466/697 [08:42<01:08,  3.37it/s]
Continual predict:  67%|██████▋   | 467/697 [08:42<01:06,  3.44it/s]
Continual predict:  67%|██████▋   | 468/697 [08:43<01:11,  3.20it/s]
Continual predict:  67%|██████▋   | 469/697 [08:43<01:11,  3.17it/s]
Continual predict:  67%|██████▋   | 470/697 [08:43<01:12,  3.11it/s]
Continual predict:  68%|██████▊   | 471/697 [08:44<01:12,  3.12it/s]
Continual predict:  68%|██████▊   | 472/697 [08:45<01:52,  2.01it/s]
Continual predict:  68%|██████▊   | 473/697 [08:45<01:54,  1.96it/s]
Continual predict:  68%|██████▊   | 474/697 [08:45<01:34,  2.37it/s]
Continual predict:  68%|██████▊   | 475/697 [08:46<01:27,  2.55it/s]
Continual predict:  68%|██████▊   | 476/697 [08:46<01:25,  2.59it/s]
Continual predict:  68%|██████▊   | 477/697 [08:46<01:20,  2.72it/s]
Continual predict:  69%|██████▊   | 478/697 [08:47<01:20,  2.71it/s]
Continual predict:  69%|██████▊   | 479/697 [08:48<01:46,  2.06it/s]
Continual predict:  69%|██████▉   | 480/697 [08:48<01:35,  2.28it/s]
Continual predict:  69%|██████▉   | 481/697 [08:48<01:30,  2.38it/s]
Continual predict:  69%|██████▉   | 482/697 [08:49<01:35,  2.25it/s]
Continual predict:  69%|██████▉   | 483/697 [08:50<02:02,  1.75it/s]
Continual predict:  69%|██████▉   | 484/697 [08:51<02:34,  1.38it/s]
Continual predict:  70%|██████▉   | 485/697 [08:52<03:00,  1.18it/s]
Continual predict:  70%|██████▉   | 486/697 [08:53<03:16,  1.07it/s]
Continual predict:  70%|██████▉   | 487/697 [08:55<03:53,  1.11s/it]
Continual predict:  70%|███████   | 488/697 [08:55<03:09,  1.10it/s]
Continual predict:  70%|███████   | 489/697 [08:56<02:57,  1.17it/s]
Continual predict:  70%|███████   | 490/697 [08:57<02:52,  1.20it/s]
Continual predict:  70%|███████   | 491/697 [08:57<02:50,  1.21it/s]
Continual predict:  71%|███████   | 492/697 [08:58<02:56,  1.16it/s]
Continual predict:  71%|███████   | 493/697 [08:59<02:48,  1.21it/s]
Continual predict:  71%|███████   | 494/697 [09:00<02:52,  1.18it/s]
Continual predict:  71%|███████   | 495/697 [09:02<03:36,  1.07s/it]
Continual predict:  71%|███████   | 496/697 [09:02<03:20,  1.00it/s]
Continual predict:  71%|███████▏  | 497/697 [09:03<02:53,  1.15it/s]
Continual predict:  71%|███████▏  | 498/697 [09:04<02:52,  1.15it/s]
Continual predict:  72%|███████▏  | 499/697 [09:05<02:51,  1.15it/s]
Continual predict:  72%|███████▏  | 500/697 [09:05<02:26,  1.35it/s]
Continual predict:  72%|███████▏  | 501/697 [09:06<02:14,  1.45it/s]
Continual predict:  72%|███████▏  | 502/697 [09:06<02:12,  1.47it/s]
Continual predict:  72%|███████▏  | 503/697 [09:07<02:19,  1.39it/s]
Continual predict:  72%|███████▏  | 504/697 [09:08<02:10,  1.48it/s]
Continual predict:  72%|███████▏  | 505/697 [09:09<02:36,  1.23it/s]
Continual predict:  73%|███████▎  | 506/697 [09:10<02:40,  1.19it/s]
Continual predict:  73%|███████▎  | 507/697 [09:11<02:45,  1.15it/s]
Continual predict:  73%|███████▎  | 508/697 [09:11<02:36,  1.21it/s]
Continual predict:  73%|███████▎  | 509/697 [09:12<02:37,  1.20it/s]
Continual predict:  73%|███████▎  | 510/697 [09:14<03:30,  1.13s/it]
Continual predict:  73%|███████▎  | 511/697 [09:15<03:38,  1.18s/it]
Continual predict:  73%|███████▎  | 512/697 [09:16<03:09,  1.02s/it]
Continual predict:  74%|███████▎  | 513/697 [09:18<03:41,  1.20s/it]
Continual predict:  74%|███████▎  | 514/697 [09:19<03:28,  1.14s/it]
Continual predict:  74%|███████▍  | 515/697 [09:19<03:05,  1.02s/it]
Continual predict:  74%|███████▍  | 516/697 [09:20<02:48,  1.08it/s]
Continual predict:  74%|███████▍  | 517/697 [09:21<02:50,  1.05it/s]
Continual predict:  74%|███████▍  | 518/697 [09:22<02:49,  1.06it/s]
Continual predict:  74%|███████▍  | 519/697 [09:24<03:21,  1.13s/it]
Continual predict:  75%|███████▍  | 520/697 [09:24<02:55,  1.01it/s]
Continual predict:  75%|███████▍  | 521/697 [09:25<02:58,  1.01s/it]
Continual predict:  75%|███████▍  | 522/697 [09:26<02:37,  1.11it/s]
Continual predict:  75%|███████▌  | 523/697 [09:27<02:31,  1.15it/s]
Continual predict:  75%|███████▌  | 524/697 [09:27<02:20,  1.24it/s]
Continual predict:  75%|███████▌  | 525/697 [09:28<02:21,  1.21it/s]
Continual predict:  75%|███████▌  | 526/697 [09:29<02:26,  1.16it/s]
Continual predict:  76%|███████▌  | 527/697 [09:30<02:23,  1.18it/s]
Continual predict:  76%|███████▌  | 528/697 [09:31<02:53,  1.03s/it]
Continual predict:  76%|███████▌  | 529/697 [09:32<02:36,  1.07it/s]
Continual predict:  76%|███████▌  | 530/697 [09:33<02:48,  1.01s/it]
Continual predict:  76%|███████▌  | 531/697 [09:35<03:00,  1.09s/it]
Continual predict:  76%|███████▋  | 532/697 [09:36<03:18,  1.20s/it]
Continual predict:  76%|███████▋  | 533/697 [09:37<02:39,  1.03it/s]
Continual predict:  77%|███████▋  | 534/697 [09:37<02:30,  1.08it/s]
Continual predict:  77%|███████▋  | 535/697 [09:38<02:29,  1.09it/s]
Continual predict:  77%|███████▋  | 536/697 [09:39<02:01,  1.33it/s]
Continual predict:  77%|███████▋  | 537/697 [09:39<02:05,  1.27it/s]
Continual predict:  77%|███████▋  | 538/697 [09:40<02:05,  1.27it/s]
Continual predict:  77%|███████▋  | 539/697 [09:41<02:15,  1.17it/s]
Continual predict:  77%|███████▋  | 540/697 [09:43<02:33,  1.02it/s]
Continual predict:  78%|███████▊  | 541/697 [09:43<02:13,  1.17it/s]
Continual predict:  78%|███████▊  | 542/697 [09:44<01:58,  1.31it/s]
Continual predict:  78%|███████▊  | 543/697 [09:44<01:59,  1.29it/s]
Continual predict:  78%|███████▊  | 544/697 [09:46<02:16,  1.12it/s]
Continual predict:  78%|███████▊  | 545/697 [09:47<02:51,  1.13s/it]
Continual predict:  78%|███████▊  | 546/697 [09:48<02:35,  1.03s/it]
Continual predict:  78%|███████▊  | 547/697 [09:49<02:29,  1.00it/s]
Continual predict:  79%|███████▊  | 548/697 [09:50<02:24,  1.03it/s]
Continual predict:  79%|███████▉  | 549/697 [09:51<02:25,  1.01it/s]
Continual predict:  79%|███████▉  | 550/697 [09:52<02:29,  1.02s/it]
Continual predict:  79%|███████▉  | 551/697 [09:53<02:05,  1.17it/s]
Continual predict:  79%|███████▉  | 552/697 [09:53<01:54,  1.27it/s]
Continual predict:  79%|███████▉  | 553/697 [09:55<02:19,  1.03it/s]
Continual predict:  79%|███████▉  | 554/697 [09:56<02:27,  1.03s/it]
Continual predict:  80%|███████▉  | 555/697 [09:57<02:28,  1.05s/it]
Continual predict:  80%|███████▉  | 556/697 [09:58<02:11,  1.07it/s]
Continual predict:  80%|███████▉  | 557/697 [09:59<02:36,  1.12s/it]
Continual predict:  80%|████████  | 558/697 [10:00<02:11,  1.06it/s]
Continual predict:  80%|████████  | 559/697 [10:01<02:19,  1.01s/it]
Continual predict:  80%|████████  | 560/697 [10:02<02:24,  1.05s/it]
Continual predict:  80%|████████  | 561/697 [10:03<02:37,  1.15s/it]
Continual predict:  81%|████████  | 562/697 [10:05<02:43,  1.21s/it]
Continual predict:  81%|████████  | 563/697 [10:06<02:52,  1.29s/it]
Continual predict:  81%|████████  | 564/697 [10:07<02:55,  1.32s/it]
Continual predict:  81%|████████  | 565/697 [10:09<02:52,  1.31s/it]
Continual predict:  81%|████████  | 566/697 [10:10<02:31,  1.15s/it]
Continual predict:  81%|████████▏ | 567/697 [10:11<02:32,  1.17s/it]
Continual predict:  81%|████████▏ | 568/697 [10:12<02:16,  1.06s/it]
Continual predict:  82%|████████▏ | 569/697 [10:13<02:13,  1.04s/it]
Continual predict:  82%|████████▏ | 570/697 [10:13<01:48,  1.17it/s]
Continual predict:  82%|████████▏ | 571/697 [10:14<01:49,  1.15it/s]
Continual predict:  82%|████████▏ | 572/697 [10:14<01:36,  1.29it/s]
Continual predict:  82%|████████▏ | 573/697 [10:16<01:51,  1.11it/s]
Continual predict:  82%|████████▏ | 574/697 [10:16<01:41,  1.21it/s]
Continual predict:  82%|████████▏ | 575/697 [10:17<01:48,  1.12it/s]
Continual predict:  83%|████████▎ | 576/697 [10:18<01:36,  1.25it/s]
Continual predict:  83%|████████▎ | 577/697 [10:19<01:39,  1.20it/s]
Continual predict:  83%|████████▎ | 578/697 [10:19<01:27,  1.37it/s]
Continual predict:  83%|████████▎ | 579/697 [10:20<01:33,  1.26it/s]
Continual predict:  83%|████████▎ | 580/697 [10:22<01:52,  1.04it/s]
Continual predict:  83%|████████▎ | 581/697 [10:22<01:40,  1.16it/s]
Continual predict:  84%|████████▎ | 582/697 [10:22<01:14,  1.54it/s]
Continual predict:  84%|████████▎ | 583/697 [10:23<00:56,  2.01it/s]
Continual predict:  84%|████████▍ | 584/697 [10:23<00:44,  2.56it/s]
Continual predict:  84%|████████▍ | 585/697 [10:23<00:35,  3.15it/s]
Continual predict:  84%|████████▍ | 586/697 [10:23<00:29,  3.78it/s]
Continual predict:  84%|████████▍ | 587/697 [10:23<00:25,  4.38it/s]
Continual predict:  84%|████████▍ | 588/697 [10:23<00:22,  4.92it/s]
Continual predict:  85%|████████▍ | 589/697 [10:23<00:20,  5.39it/s]
Continual predict:  85%|████████▍ | 590/697 [10:24<00:18,  5.81it/s]
Continual predict:  85%|████████▍ | 591/697 [10:24<00:17,  6.12it/s]
Continual predict:  85%|████████▍ | 592/697 [10:24<00:16,  6.35it/s]
Continual predict:  85%|████████▌ | 593/697 [10:24<00:15,  6.56it/s]
Continual predict:  85%|████████▌ | 594/697 [10:24<00:15,  6.70it/s]
Continual predict:  85%|████████▌ | 595/697 [10:24<00:14,  6.82it/s]
Continual predict:  86%|████████▌ | 596/697 [10:24<00:14,  6.90it/s]
Continual predict:  86%|████████▌ | 597/697 [10:25<00:14,  6.96it/s]
Continual predict:  86%|████████▌ | 598/697 [10:25<00:14,  6.96it/s]
Continual predict:  86%|████████▌ | 599/697 [10:25<00:14,  6.96it/s]
Continual predict:  86%|████████▌ | 600/697 [10:25<00:13,  7.00it/s]
Continual predict:  86%|████████▌ | 601/697 [10:25<00:13,  6.99it/s]
Continual predict:  86%|████████▋ | 602/697 [10:25<00:13,  6.97it/s]
Continual predict:  87%|████████▋ | 603/697 [10:25<00:13,  6.96it/s]
Continual predict:  87%|████████▋ | 604/697 [10:26<00:13,  6.95it/s]
Continual predict:  87%|████████▋ | 605/697 [10:26<00:13,  6.95it/s]
Continual predict:  87%|████████▋ | 606/697 [10:26<00:13,  6.94it/s]
Continual predict:  87%|████████▋ | 607/697 [10:26<00:12,  6.99it/s]
Continual predict:  87%|████████▋ | 608/697 [10:26<00:12,  7.02it/s]
Continual predict:  87%|████████▋ | 609/697 [10:26<00:12,  7.05it/s]
Continual predict:  88%|████████▊ | 610/697 [10:26<00:12,  7.05it/s]
Continual predict:  88%|████████▊ | 611/697 [10:27<00:12,  7.06it/s]
Continual predict:  88%|████████▊ | 612/697 [10:27<00:12,  7.03it/s]
Continual predict:  88%|████████▊ | 613/697 [10:27<00:11,  7.01it/s]
Continual predict:  88%|████████▊ | 614/697 [10:27<00:11,  7.04it/s]
Continual predict:  88%|████████▊ | 615/697 [10:27<00:11,  7.02it/s]
Continual predict:  88%|████████▊ | 616/697 [10:27<00:11,  7.00it/s]
Continual predict:  89%|████████▊ | 617/697 [10:27<00:11,  7.03it/s]
Continual predict:  89%|████████▊ | 618/697 [10:28<00:11,  7.01it/s]
Continual predict:  89%|████████▉ | 619/697 [10:28<00:11,  6.99it/s]
Continual predict:  89%|████████▉ | 620/697 [10:28<00:11,  6.97it/s]
Continual predict:  89%|████████▉ | 621/697 [10:28<00:10,  6.97it/s]
Continual predict:  89%|████████▉ | 622/697 [10:28<00:10,  6.96it/s]
Continual predict:  89%|████████▉ | 623/697 [10:28<00:10,  6.95it/s]
Continual predict:  90%|████████▉ | 624/697 [10:28<00:10,  6.97it/s]
Continual predict:  90%|████████▉ | 625/697 [10:29<00:10,  6.79it/s]
Continual predict:  90%|████████▉ | 626/697 [10:29<00:10,  6.87it/s]
Continual predict:  90%|████████▉ | 627/697 [10:29<00:10,  6.89it/s]
Continual predict:  90%|█████████ | 628/697 [10:29<00:09,  6.94it/s]
Continual predict:  90%|█████████ | 629/697 [10:29<00:09,  6.99it/s]
Continual predict:  90%|█████████ | 630/697 [10:29<00:09,  6.99it/s]
Continual predict:  91%|█████████ | 631/697 [10:29<00:09,  6.97it/s]
Continual predict:  91%|█████████ | 632/697 [10:30<00:09,  6.97it/s]
Continual predict:  91%|█████████ | 633/697 [10:30<00:09,  7.00it/s]
Continual predict:  91%|█████████ | 634/697 [10:30<00:09,  6.98it/s]
Continual predict:  91%|█████████ | 635/697 [10:30<00:08,  7.02it/s]
Continual predict:  91%|█████████ | 636/697 [10:30<00:08,  6.99it/s]
Continual predict:  91%|█████████▏| 637/697 [10:30<00:08,  6.99it/s]
Continual predict:  92%|█████████▏| 638/697 [10:30<00:08,  6.96it/s]
Continual predict:  92%|█████████▏| 639/697 [10:31<00:08,  6.95it/s]
Continual predict:  92%|█████████▏| 640/697 [10:31<00:08,  6.99it/s]
Continual predict:  92%|█████████▏| 641/697 [10:31<00:07,  7.02it/s]
Continual predict:  92%|█████████▏| 642/697 [10:31<00:07,  7.01it/s]
Continual predict:  92%|█████████▏| 643/697 [10:31<00:07,  6.99it/s]
Continual predict:  92%|█████████▏| 644/697 [10:31<00:07,  7.03it/s]
Continual predict:  93%|█████████▎| 645/697 [10:31<00:07,  7.01it/s]
Continual predict:  93%|█████████▎| 646/697 [10:32<00:07,  7.00it/s]
Continual predict:  93%|█████████▎| 647/697 [10:32<00:07,  6.98it/s]
Continual predict:  93%|█████████▎| 648/697 [10:32<00:06,  7.02it/s]
Continual predict:  93%|█████████▎| 649/697 [10:32<00:06,  7.01it/s]
Continual predict:  93%|█████████▎| 650/697 [10:32<00:06,  7.04it/s]
Continual predict:  93%|█████████▎| 651/697 [10:32<00:06,  7.06it/s]
Continual predict:  94%|█████████▎| 652/697 [10:32<00:06,  7.06it/s]
Continual predict:  94%|█████████▎| 653/697 [10:33<00:06,  7.08it/s]
Continual predict:  94%|█████████▍| 654/697 [10:33<00:06,  7.09it/s]
Continual predict:  94%|█████████▍| 655/697 [10:33<00:05,  7.09it/s]
Continual predict:  94%|█████████▍| 656/697 [10:33<00:05,  7.04it/s]
Continual predict:  94%|█████████▍| 657/697 [10:33<00:05,  7.02it/s]
Continual predict:  94%|█████████▍| 658/697 [10:33<00:05,  7.00it/s]
Continual predict:  95%|█████████▍| 659/697 [10:33<00:05,  7.03it/s]
Continual predict:  95%|█████████▍| 660/697 [10:34<00:05,  7.02it/s]
Continual predict:  95%|█████████▍| 661/697 [10:34<00:05,  7.00it/s]
Continual predict:  95%|█████████▍| 662/697 [10:34<00:05,  6.98it/s]
Continual predict:  95%|█████████▌| 663/697 [10:34<00:04,  6.97it/s]
Continual predict:  95%|█████████▌| 664/697 [10:34<00:04,  6.96it/s]
Continual predict:  95%|█████████▌| 665/697 [10:34<00:04,  6.97it/s]
Continual predict:  96%|█████████▌| 666/697 [10:34<00:04,  7.01it/s]
Continual predict:  96%|█████████▌| 667/697 [10:35<00:04,  7.00it/s]
Continual predict:  96%|█████████▌| 668/697 [10:35<00:04,  6.99it/s]
Continual predict:  96%|█████████▌| 669/697 [10:35<00:03,  7.02it/s]
Continual predict:  96%|█████████▌| 670/697 [10:35<00:03,  7.00it/s]
Continual predict:  96%|█████████▋| 671/697 [10:35<00:03,  6.99it/s]
Continual predict:  96%|█████████▋| 672/697 [10:35<00:03,  6.99it/s]
Continual predict:  97%|█████████▋| 673/697 [10:35<00:03,  7.01it/s]
Continual predict:  97%|█████████▋| 674/697 [10:36<00:03,  6.99it/s]
Continual predict:  97%|█████████▋| 675/697 [10:36<00:03,  7.03it/s]
Continual predict:  97%|█████████▋| 676/697 [10:36<00:02,  7.05it/s]
Continual predict:  97%|█████████▋| 677/697 [10:36<00:02,  7.07it/s]
Continual predict:  97%|█████████▋| 678/697 [10:36<00:02,  7.04it/s]
Continual predict:  97%|█████████▋| 679/697 [10:36<00:02,  7.02it/s]
Continual predict:  98%|█████████▊| 680/697 [10:36<00:02,  7.00it/s]
Continual predict:  98%|█████████▊| 681/697 [10:37<00:02,  6.98it/s]
Continual predict:  98%|█████████▊| 682/697 [10:38<00:06,  2.33it/s]
Continual predict:  98%|█████████▊| 683/697 [10:40<00:14,  1.02s/it]
Continual predict:  98%|█████████▊| 684/697 [10:42<00:15,  1.20s/it]
Continual predict:  98%|█████████▊| 685/697 [10:42<00:12,  1.05s/it]
Continual predict:  98%|█████████▊| 686/697 [10:45<00:15,  1.42s/it]
Continual predict:  99%|█████████▊| 687/697 [10:46<00:14,  1.44s/it]
Continual predict:  99%|█████████▊| 688/697 [10:51<00:21,  2.34s/it]
Continual predict:  99%|█████████▉| 689/697 [10:51<00:15,  1.88s/it]
Continual predict:  99%|█████████▉| 690/697 [10:55<00:16,  2.42s/it]
Continual predict:  99%|█████████▉| 691/697 [10:56<00:11,  2.00s/it]
Continual predict:  99%|█████████▉| 692/697 [10:57<00:07,  1.59s/it]
Continual predict:  99%|█████████▉| 693/697 [11:00<00:07,  1.96s/it]
Continual predict: 100%|█████████▉| 694/697 [11:00<00:04,  1.62s/it]
Continual predict: 100%|█████████▉| 695/697 [11:02<00:03,  1.73s/it]
Continual predict: 100%|█████████▉| 696/697 [11:04<00:01,  1.58s/it]
Continual predict: 100%|██████████| 697/697 [11:04<00:00,  1.30s/it]
Continual predict: 100%|██████████| 697/697 [11:04<00:00,  1.05it/s]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=16]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=18]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (697/697)  label_acc=100.0% (697/697)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                               =       15.0
+  predict_eval_rougeL                                                 =    59.7787
+  predict_eval_rougeL_for_CL                                          =    59.7787
+  predict_eval_rougeL_for_task002_quoref_answer_generation            =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation     =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                 =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction      =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                     =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation          =    11.1335
+  predict_eval_rougeL_for_task1729_personachat_generate_next          =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                  =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification        =       91.0
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation =    15.2912
+  predict_exact_match                                                 =    49.6413
+  predict_exact_match_for_CL                                          =    49.6413
+  predict_exact_match_for_task002_quoref_answer_generation            =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation     =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                 =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction      =       98.0
+  predict_exact_match_for_task1572_samsum_summary                     =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation          =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next          =        0.0
+  predict_exact_match_for_task181_outcome_extraction                  =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification        =       91.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation =        0.0
+  predict_gen_len                                                     =     8.1894
+  predict_global_step                                                 =         60
+  predict_rouge1                                                      =    61.2565
+  predict_rouge1_for_CL                                               =    61.2565
+  predict_rouge1_for_task002_quoref_answer_generation                 =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation          =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                      =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction           =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                          =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation               =    13.7971
+  predict_rouge1_for_task1729_personachat_generate_next               =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                       =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification             =       91.0
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation      =    15.2912
+  predict_runtime                                                     = 0:11:04.74
+  predict_samples                                                     =        697
+  predict_samples_per_second                                          =      1.049
+  predict_steps_per_second                                            =      1.049
+[2026-04-30 02:00:14,890] [INFO] [launch.py:367:main] Process 1351998 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/all_results.json b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..755e9f2310e001e31a3e8250966fbef4ad182e4c
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/all_results.json
@@ -0,0 +1,54 @@
+{
+  "predict_exact_match": 43.7892,
+  "predict_rouge1": 62.2164,
+  "predict_eval_rougeL": 60.6727,
+  "predict_gen_len": 9.808,
+  "predict_runtime": 835.6393646709621,
+  "predict_samples": 797,
+  "predict_samples_per_second": 0.954,
+  "predict_steps_per_second": 0.954,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 43.7892,
+  "predict_rouge1_for_CL": 62.2164,
+  "predict_eval_rougeL_for_CL": 60.6727,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "predict_exact_match_for_task748_glucose_reverse_cause_event_detection": 3.0,
+  "predict_rouge1_for_task748_glucose_reverse_cause_event_detection": 68.907,
+  "predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection": 66.9039,
+  "train_runtime": 2375.7951,
+  "train_samples_per_second": 6.314,
+  "train_steps_per_second": 0.202,
+  "total_flos": 1.2176157685946778e+17,
+  "train_loss": 0.5081466535727183,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a18a955a77c3f0667948f7292656b58205953b6e
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:847fa1dff0c1aa7eebef6955a524c064e5eccec1294d5e3a6c60ff489dc9da7a
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..6b2b8fd265819fa26b0c26250722e412aa9468f7
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6df5c4b5ac7cc7353efe7e2722cb07f483c5814823dc48558372f3d0efdb71ad
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..e8cfe8e7e2e3922f25acf9f022b629e0af0f3b99
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7952e207d7c410c9b6c38bfb14a7939fedf132b3f2ce199d1fccd36ad83b4728
+size 1718692693
diff --git a/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/train.log b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..14e3977ce106627e765093dd382f72d03d52a3ef
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/train.log
@@ -0,0 +1,663 @@
+[2026-04-30 02:00:20,803] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 02:00:20,803] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29510 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task748_glucose_reverse_cause_event_detection --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection --cur_task_id 10 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[2026-04-30 02:00:23,983] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 02:00:23,983] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 02:00:23,983] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 02:00:23,983] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 02:00:23,983] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 02:00:23,984] [INFO] [launch.py:272:main] process 1403486 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task748_glucose_reverse_cause_event_detection', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection', '--cur_task_id', '10', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights']
+
+============================================================
+Task 10: task748_glucose_reverse_cause_event_detection
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/797
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   3%|▎         | 10/291 [00:00<00:02, 96.00it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 141.60it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 146.49it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 156.09it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 165.25it/s]
Loading weights:  36%|███▌      | 105/291 [00:00<00:01, 164.13it/s]
Loading weights:  44%|████▎     | 127/291 [00:00<00:00, 177.52it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 180.07it/s]
Loading weights:  57%|█████▋    | 166/291 [00:01<00:00, 175.53it/s]
Loading weights:  64%|██████▍   | 186/291 [00:01<00:00, 180.58it/s]
Loading weights:  70%|███████   | 205/291 [00:01<00:00, 173.20it/s]
Loading weights:  79%|███████▊  | 229/291 [00:01<00:00, 184.28it/s]
Loading weights:  85%|████████▌ | 248/291 [00:01<00:00, 181.61it/s]
Loading weights:  92%|█████████▏| 267/291 [00:01<00:00, 175.51it/s]
Loading weights:  99%|█████████▉| 288/291 [00:01<00:00, 184.77it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 174.11it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1443)
+[FIX] Re-initialized lora_A/B in 704 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(10, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[9] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 10 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6762.349M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 10 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 10 sigs, cur=task748_glucose_reverse_cause_event_detection, mapping={'task748_glucose_reverse_cause_event_detection': 0, 'task1572_samsum_summary': 10, 'task363_sst2_polarity_classification': 9, 'task1290_xsum_summarization': 8, 'task181_outcome_extraction': 7, 'task002_quoref_answer_generation': 6, 'task1510_evalution_relation_extraction': 5, 'task639_multi_woz_user_utterance_generation': 4, 'task1729_personachat_generate_next': 3, 'task073_commonsenseqa_answer_generation': 2, 'task1590_diplomacy_text_generation': 1}
+[SGWI] Task 10: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 02:01:29,388] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:04<39:40,  4.97s/it]
  0%|          | 2/480 [00:09<36:02,  4.52s/it]
  1%|          | 3/480 [00:13<35:29,  4.47s/it]
  1%|          | 4/480 [00:17<35:11,  4.44s/it]
  1%|          | 5/480 [00:22<36:18,  4.59s/it]
  1%|▏         | 6/480 [00:27<36:51,  4.66s/it]
  1%|▏         | 7/480 [00:32<36:02,  4.57s/it]
  2%|▏         | 8/480 [00:36<35:22,  4.50s/it]
  2%|▏         | 9/480 [00:40<35:06,  4.47s/it]
  2%|▏         | 10/480 [00:45<34:54,  4.46s/it]
                                                
{'loss': '1.678', 'grad_norm': '0.1772', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:45<34:54,  4.46s/it]
  2%|▏         | 11/480 [00:49<35:13,  4.51s/it]
  2%|▎         | 12/480 [00:54<34:53,  4.47s/it]
  3%|▎         | 13/480 [00:58<34:26,  4.43s/it]
  3%|▎         | 14/480 [01:02<33:54,  4.37s/it]
  3%|▎         | 15/480 [01:07<33:34,  4.33s/it]
  3%|▎         | 16/480 [01:11<33:26,  4.32s/it]
  4%|▎         | 17/480 [01:15<33:08,  4.29s/it]
  4%|▍         | 18/480 [01:19<33:13,  4.32s/it]
  4%|▍         | 19/480 [01:24<33:07,  4.31s/it]
  4%|▍         | 20/480 [01:28<33:57,  4.43s/it]
                                                
{'loss': '1.251', 'grad_norm': '0.2153', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:28<33:57,  4.43s/it]
  4%|▍         | 21/480 [01:33<33:26,  4.37s/it]
  5%|▍         | 22/480 [01:37<33:04,  4.33s/it]
  5%|▍         | 23/480 [01:41<33:02,  4.34s/it]
  5%|▌         | 24/480 [01:45<32:44,  4.31s/it]
  5%|▌         | 25/480 [01:50<33:01,  4.35s/it]
  5%|▌         | 26/480 [01:54<32:44,  4.33s/it]
  6%|▌         | 27/480 [01:59<32:35,  4.32s/it]
  6%|▌         | 28/480 [02:03<32:57,  4.38s/it]
  6%|▌         | 29/480 [02:07<32:36,  4.34s/it]
  6%|▋         | 30/480 [02:12<32:54,  4.39s/it]
                                                
{'loss': '0.7891', 'grad_norm': '0.3311', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [02:12<32:54,  4.39s/it]
  6%|▋         | 31/480 [02:16<32:47,  4.38s/it]
  7%|▋         | 32/480 [02:17<25:41,  3.44s/it]
  7%|▋         | 33/480 [02:22<27:43,  3.72s/it]
  7%|▋         | 34/480 [02:26<29:15,  3.94s/it]
  7%|▋         | 35/480 [02:31<30:16,  4.08s/it]
  8%|▊         | 36/480 [02:35<31:08,  4.21s/it]
  8%|▊         | 37/480 [02:39<31:09,  4.22s/it]
  8%|▊         | 38/480 [02:44<32:03,  4.35s/it]
  8%|▊         | 39/480 [02:49<32:31,  4.43s/it]
  8%|▊         | 40/480 [02:53<32:35,  4.45s/it]
                                                
{'loss': '0.6304', 'grad_norm': '0.3208', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [02:53<32:35,  4.45s/it]
  9%|▊         | 41/480 [02:57<32:06,  4.39s/it]
  9%|▉         | 42/480 [03:02<31:43,  4.35s/it]
  9%|▉         | 43/480 [03:06<31:39,  4.35s/it]
  9%|▉         | 44/480 [03:10<31:24,  4.32s/it]
  9%|▉         | 45/480 [03:15<31:28,  4.34s/it]
 10%|▉         | 46/480 [03:19<31:14,  4.32s/it]
 10%|▉         | 47/480 [03:23<31:08,  4.32s/it]
 10%|█         | 48/480 [03:28<31:39,  4.40s/it]
 10%|█         | 49/480 [03:32<31:16,  4.35s/it]
 10%|█         | 50/480 [03:36<31:14,  4.36s/it]
                                                
{'loss': '0.5714', 'grad_norm': '0.2726', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [03:36<31:14,  4.36s/it]
 11%|█         | 51/480 [03:41<31:03,  4.34s/it]
 11%|█         | 52/480 [03:45<31:30,  4.42s/it]
 11%|█         | 53/480 [03:50<31:08,  4.37s/it]
 11%|█▏        | 54/480 [03:54<30:48,  4.34s/it]
 11%|█▏        | 55/480 [03:59<31:58,  4.51s/it]
 12%|█▏        | 56/480 [04:03<32:23,  4.58s/it]
 12%|█▏        | 57/480 [04:08<32:00,  4.54s/it]
 12%|█▏        | 58/480 [04:13<32:47,  4.66s/it]
 12%|█▏        | 59/480 [04:17<32:30,  4.63s/it]
 12%|█▎        | 60/480 [04:22<32:00,  4.57s/it]
                                                
{'loss': '0.5173', 'grad_norm': '0.2076', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [04:22<32:00,  4.57s/it]
 13%|█▎        | 61/480 [04:26<31:40,  4.54s/it]
 13%|█▎        | 62/480 [04:31<31:37,  4.54s/it]
 13%|█▎        | 63/480 [04:35<30:58,  4.46s/it]
 13%|█▎        | 64/480 [04:36<24:14,  3.50s/it]
 14%|█▎        | 65/480 [04:41<26:21,  3.81s/it]
 14%|█▍        | 66/480 [04:45<27:42,  4.01s/it]
 14%|█▍        | 67/480 [04:50<28:42,  4.17s/it]
 14%|█▍        | 68/480 [04:54<29:11,  4.25s/it]
 14%|█▍        | 69/480 [04:59<29:19,  4.28s/it]
 15%|█▍        | 70/480 [05:03<29:14,  4.28s/it]
                                                
{'loss': '0.561', 'grad_norm': '0.4556', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [05:03<29:14,  4.28s/it]
 15%|█▍        | 71/480 [05:07<29:12,  4.29s/it]
 15%|█▌        | 72/480 [05:12<29:23,  4.32s/it]
 15%|█▌        | 73/480 [05:16<29:13,  4.31s/it]
 15%|█▌        | 74/480 [05:20<29:15,  4.32s/it]
 16%|█▌        | 75/480 [05:25<29:04,  4.31s/it]
 16%|█▌        | 76/480 [05:29<29:15,  4.34s/it]
 16%|█▌        | 77/480 [05:33<29:16,  4.36s/it]
 16%|█▋        | 78/480 [05:38<29:01,  4.33s/it]
 16%|█▋        | 79/480 [05:42<28:54,  4.33s/it]
 17%|█▋        | 80/480 [05:47<29:06,  4.37s/it]
                                                
{'loss': '0.4985', 'grad_norm': '0.2548', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [05:47<29:06,  4.37s/it]
 17%|█▋        | 81/480 [05:51<28:52,  4.34s/it]
 17%|█▋        | 82/480 [05:55<28:58,  4.37s/it]
 17%|█▋        | 83/480 [06:00<29:03,  4.39s/it]
 18%|█▊        | 84/480 [06:04<29:17,  4.44s/it]
 18%|█▊        | 85/480 [06:09<28:54,  4.39s/it]
 18%|█▊        | 86/480 [06:13<29:55,  4.56s/it]
 18%|█▊        | 87/480 [06:18<30:30,  4.66s/it]
 18%|█▊        | 88/480 [06:23<29:40,  4.54s/it]
 19%|█▊        | 89/480 [06:27<29:40,  4.55s/it]
 19%|█▉        | 90/480 [06:32<29:26,  4.53s/it]
                                                
{'loss': '0.5169', 'grad_norm': '0.3006', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [06:32<29:26,  4.53s/it]
 19%|█▉        | 91/480 [06:36<29:22,  4.53s/it]
 19%|█▉        | 92/480 [06:40<28:50,  4.46s/it]
 19%|█▉        | 93/480 [06:45<29:20,  4.55s/it]
 20%|█▉        | 94/480 [06:50<29:05,  4.52s/it]
 20%|█▉        | 95/480 [06:54<28:52,  4.50s/it]
 20%|██        | 96/480 [06:55<22:23,  3.50s/it]
 20%|██        | 97/480 [07:00<23:48,  3.73s/it]
 20%|██        | 98/480 [07:04<24:52,  3.91s/it]
 21%|██        | 99/480 [07:08<25:48,  4.06s/it]
 21%|██        | 100/480 [07:13<26:10,  4.13s/it]
                                                 
{'loss': '0.5159', 'grad_norm': '0.2203', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [07:13<26:10,  4.13s/it]
 21%|██        | 101/480 [07:17<26:46,  4.24s/it]
 21%|██▏       | 102/480 [07:22<27:32,  4.37s/it]
 21%|██▏       | 103/480 [07:26<27:55,  4.44s/it]
 22%|██▏       | 104/480 [07:31<27:59,  4.47s/it]
 22%|██▏       | 105/480 [07:35<27:32,  4.41s/it]
 22%|██▏       | 106/480 [07:40<27:24,  4.40s/it]
 22%|██▏       | 107/480 [07:44<27:05,  4.36s/it]
 22%|██▎       | 108/480 [07:48<27:28,  4.43s/it]
 23%|██▎       | 109/480 [07:53<27:09,  4.39s/it]
 23%|██▎       | 110/480 [07:57<26:51,  4.36s/it]
                                                 
{'loss': '0.515', 'grad_norm': '0.2687', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [07:57<26:51,  4.36s/it]
 23%|██▎       | 111/480 [08:02<27:09,  4.42s/it]
 23%|██▎       | 112/480 [08:06<27:36,  4.50s/it]
 24%|██▎       | 113/480 [08:11<27:42,  4.53s/it]
 24%|██▍       | 114/480 [08:15<27:14,  4.47s/it]
 24%|██▍       | 115/480 [08:20<27:30,  4.52s/it]
 24%|██▍       | 116/480 [08:24<27:31,  4.54s/it]
 24%|██▍       | 117/480 [08:29<28:08,  4.65s/it]
 25%|██▍       | 118/480 [08:34<28:03,  4.65s/it]
 25%|██▍       | 119/480 [08:38<27:36,  4.59s/it]
 25%|██▌       | 120/480 [08:43<27:40,  4.61s/it]
                                                 
{'loss': '0.5011', 'grad_norm': '0.2807', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [08:43<27:40,  4.61s/it]
 25%|██▌       | 121/480 [08:48<27:20,  4.57s/it]
 25%|██▌       | 122/480 [08:53<27:57,  4.69s/it]
 26%|██▌       | 123/480 [08:57<27:27,  4.61s/it]
 26%|██▌       | 124/480 [09:01<26:48,  4.52s/it]
 26%|██▌       | 125/480 [09:06<26:48,  4.53s/it]
 26%|██▋       | 126/480 [09:10<26:41,  4.52s/it]
 26%|██▋       | 127/480 [09:15<26:48,  4.56s/it]
 27%|██▋       | 128/480 [09:16<20:42,  3.53s/it]
 27%|██▋       | 129/480 [09:20<21:56,  3.75s/it]
 27%|██▋       | 130/480 [09:25<22:56,  3.93s/it]
                                                 
{'loss': '0.4751', 'grad_norm': '0.4279', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [09:25<22:56,  3.93s/it]
 27%|██▋       | 131/480 [09:29<23:27,  4.03s/it]
 28%|██▊       | 132/480 [09:34<24:20,  4.20s/it]
 28%|██▊       | 133/480 [09:38<24:31,  4.24s/it]
 28%|██▊       | 134/480 [09:42<24:30,  4.25s/it]
 28%|██▊       | 135/480 [09:47<24:49,  4.32s/it]
 28%|██▊       | 136/480 [09:51<25:24,  4.43s/it]
 29%|██▊       | 137/480 [09:56<25:15,  4.42s/it]
 29%|██▉       | 138/480 [10:00<24:55,  4.37s/it]
 29%|██▉       | 139/480 [10:04<24:59,  4.40s/it]
 29%|██▉       | 140/480 [10:09<25:08,  4.44s/it]
                                                 
{'loss': '0.4949', 'grad_norm': '0.3025', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [10:09<25:08,  4.44s/it]
 29%|██▉       | 141/480 [10:13<25:04,  4.44s/it]
 30%|██▉       | 142/480 [10:18<24:58,  4.43s/it]
 30%|██▉       | 143/480 [10:22<24:55,  4.44s/it]
 30%|███       | 144/480 [10:27<24:42,  4.41s/it]
 30%|███       | 145/480 [10:31<24:40,  4.42s/it]
 30%|███       | 146/480 [10:36<24:38,  4.43s/it]
 31%|███       | 147/480 [10:40<24:47,  4.47s/it]
 31%|███       | 148/480 [10:45<25:13,  4.56s/it]
 31%|███       | 149/480 [10:50<25:43,  4.66s/it]
 31%|███▏      | 150/480 [10:54<25:19,  4.60s/it]
                                                 
{'loss': '0.4625', 'grad_norm': '0.191', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [10:54<25:19,  4.60s/it]
 31%|███▏      | 151/480 [10:59<25:12,  4.60s/it]
 32%|███▏      | 152/480 [11:03<24:35,  4.50s/it]
 32%|███▏      | 153/480 [11:07<24:09,  4.43s/it]
 32%|███▏      | 154/480 [11:12<24:08,  4.44s/it]
 32%|███▏      | 155/480 [11:16<23:46,  4.39s/it]
 32%|███▎      | 156/480 [11:20<23:39,  4.38s/it]
 33%|███▎      | 157/480 [11:25<23:24,  4.35s/it]
 33%|███▎      | 158/480 [11:29<23:15,  4.34s/it]
 33%|███▎      | 159/480 [11:34<23:31,  4.40s/it]
 33%|███▎      | 160/480 [11:35<18:08,  3.40s/it]
                                                 
{'loss': '0.4766', 'grad_norm': '1.339', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [11:35<18:08,  3.40s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:01<01:49,  1.11s/it]
+
Eval @ step 160:   2%|▏         | 2/100 [00:01<01:30,  1.09it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:02<01:25,  1.14it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:03<01:24,  1.14it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:04<01:15,  1.26it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:05<01:22,  1.14it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:05<01:14,  1.24it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:07<01:43,  1.13s/it]
+
Eval @ step 160:   9%|▉         | 9/100 [00:08<01:37,  1.08s/it]
+
Eval @ step 160:  10%|█         | 10/100 [00:09<01:30,  1.00s/it]
+
Eval @ step 160:  11%|█         | 11/100 [00:11<01:48,  1.22s/it]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:12<01:39,  1.13s/it]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:13<01:33,  1.08s/it]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:14<01:27,  1.02s/it]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:14<01:22,  1.03it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:15<01:17,  1.09it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:16<01:22,  1.00it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:18<01:27,  1.07s/it]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:18<01:21,  1.01s/it]
+
Eval @ step 160:  20%|██        | 20/100 [00:19<01:15,  1.05it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:20<01:14,  1.07it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:21<01:16,  1.01it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:22<01:09,  1.11it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:23<01:09,  1.09it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:24<01:06,  1.12it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:25<01:04,  1.15it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:26<01:05,  1.12it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:26<01:05,  1.11it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:27<01:02,  1.14it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:28<01:00,  1.16it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:29<01:04,  1.07it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:30<01:04,  1.05it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:31<01:05,  1.02it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:32<01:04,  1.02it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:33<01:00,  1.07it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:34<00:54,  1.18it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:35<01:01,  1.02it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:36<01:06,  1.07s/it]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:37<01:06,  1.09s/it]
+
Eval @ step 160:  40%|████      | 40/100 [00:38<00:59,  1.02it/s]
+
Eval @ step 160:  41%|████      | 41/100 [00:39<00:58,  1.01it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:40<01:00,  1.05s/it]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:41<00:58,  1.03s/it]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:42<00:58,  1.04s/it]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:43<00:55,  1.02s/it]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:44<00:53,  1.00it/s]
+
Eval @ step 160:  47%|████▋     | 47/100 [00:45<00:53,  1.01s/it]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:46<00:53,  1.04s/it]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:47<00:52,  1.04s/it]
+
Eval @ step 160:  50%|█████     | 50/100 [00:48<00:51,  1.03s/it]
+
Eval @ step 160:  51%|█████     | 51/100 [00:50<00:51,  1.06s/it]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:51<00:52,  1.10s/it]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:52<00:47,  1.01s/it]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:53<00:48,  1.06s/it]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:54<00:43,  1.03it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:54<00:40,  1.08it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:55<00:36,  1.16it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:56<00:39,  1.06it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:57<00:37,  1.09it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:58<00:35,  1.13it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:59<00:33,  1.15it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [01:00<00:33,  1.13it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [01:00<00:32,  1.16it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [01:02<00:37,  1.05s/it]
+
Eval @ step 160:  65%|██████▌   | 65/100 [01:03<00:40,  1.14s/it]
+
Eval @ step 160:  66%|██████▌   | 66/100 [01:04<00:38,  1.14s/it]
+
Eval @ step 160:  67%|██████▋   | 67/100 [01:05<00:36,  1.11s/it]
+
Eval @ step 160:  68%|██████▊   | 68/100 [01:06<00:31,  1.01it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [01:07<00:30,  1.01it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [01:08<00:29,  1.03it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [01:09<00:27,  1.06it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [01:10<00:25,  1.09it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [01:11<00:24,  1.09it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [01:11<00:22,  1.17it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [01:12<00:20,  1.24it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [01:13<00:22,  1.06it/s]
+
Eval @ step 160:  77%|███████▋  | 77/100 [01:14<00:21,  1.07it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [01:15<00:19,  1.11it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [01:16<00:20,  1.04it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [01:17<00:18,  1.07it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [01:18<00:18,  1.04it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [01:19<00:16,  1.07it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [01:20<00:15,  1.09it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [01:21<00:15,  1.06it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [01:22<00:13,  1.12it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [01:23<00:13,  1.06it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [01:24<00:12,  1.05it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [01:25<00:10,  1.11it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [01:25<00:10,  1.09it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [01:27<00:09,  1.04it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [01:27<00:07,  1.14it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [01:28<00:07,  1.09it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [01:30<00:07,  1.05s/it]
+
Eval @ step 160:  94%|█████████▍| 94/100 [01:31<00:06,  1.04s/it]
+
Eval @ step 160:  95%|█████████▌| 95/100 [01:31<00:04,  1.04it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [01:32<00:03,  1.04it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [01:34<00:03,  1.07s/it]
+
Eval @ step 160:  98%|█████████▊| 98/100 [01:34<00:01,  1.02it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [01:35<00:00,  1.03it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [01:36<00:00,  1.03it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=69.0710 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '69.07', 'epoch': '5'}
+
 33%|███▎      | 160/480 [13:12<18:08,  3.40s/it]
 34%|███▎      | 161/480 [13:16<2:54:10, 32.76s/it]
 34%|███▍      | 162/480 [13:20<2:08:35, 24.26s/it]
 34%|███▍      | 163/480 [13:25<1:36:45, 18.31s/it]
 34%|███▍      | 164/480 [13:29<1:14:24, 14.13s/it]
 34%|███▍      | 165/480 [13:34<58:56, 11.23s/it]  
 35%|███▍      | 166/480 [13:38<48:15,  9.22s/it]
 35%|███▍      | 167/480 [13:43<40:59,  7.86s/it]
 35%|███▌      | 168/480 [13:47<35:53,  6.90s/it]
 35%|███▌      | 169/480 [13:52<31:56,  6.16s/it]
 35%|███▌      | 170/480 [13:56<29:11,  5.65s/it]
                                                 
{'loss': '0.4671', 'grad_norm': '0.355', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [13:56<29:11,  5.65s/it]
 36%|███▌      | 171/480 [14:01<27:54,  5.42s/it]
 36%|███▌      | 172/480 [14:06<26:20,  5.13s/it]
 36%|███▌      | 173/480 [14:10<25:16,  4.94s/it]
 36%|███▋      | 174/480 [14:15<24:15,  4.76s/it]
 36%|███▋      | 175/480 [14:19<23:33,  4.63s/it]
 37%|███▋      | 176/480 [14:23<22:54,  4.52s/it]
 37%|███▋      | 177/480 [14:27<22:26,  4.44s/it]
 37%|███▋      | 178/480 [14:32<22:26,  4.46s/it]
 37%|███▋      | 179/480 [14:36<22:04,  4.40s/it]
 38%|███▊      | 180/480 [14:41<22:00,  4.40s/it]
                                                 
{'loss': '0.4525', 'grad_norm': '0.2864', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [14:41<22:00,  4.40s/it]
 38%|███▊      | 181/480 [14:45<21:44,  4.36s/it]
 38%|███▊      | 182/480 [14:49<21:46,  4.38s/it]
 38%|███▊      | 183/480 [14:54<21:44,  4.39s/it]
 38%|███▊      | 184/480 [14:59<22:22,  4.54s/it]
 39%|███▊      | 185/480 [15:03<22:20,  4.54s/it]
 39%|███▉      | 186/480 [15:07<21:54,  4.47s/it]
 39%|███▉      | 187/480 [15:12<22:11,  4.54s/it]
 39%|███▉      | 188/480 [15:17<22:16,  4.58s/it]
 39%|███▉      | 189/480 [15:21<22:09,  4.57s/it]
 40%|███▉      | 190/480 [15:26<21:44,  4.50s/it]
                                                 
{'loss': '0.4656', 'grad_norm': '0.2774', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [15:26<21:44,  4.50s/it]
 40%|███▉      | 191/480 [15:30<21:36,  4.49s/it]
 40%|████      | 192/480 [15:31<16:37,  3.46s/it]
 40%|████      | 193/480 [15:36<18:04,  3.78s/it]
 40%|████      | 194/480 [15:40<18:58,  3.98s/it]
 41%|████      | 195/480 [15:45<19:40,  4.14s/it]
 41%|████      | 196/480 [15:49<20:02,  4.24s/it]
 41%|████      | 197/480 [15:54<20:24,  4.33s/it]
 41%|████▏     | 198/480 [15:58<20:33,  4.37s/it]
 41%|████▏     | 199/480 [16:02<20:23,  4.36s/it]
 42%|████▏     | 200/480 [16:07<20:17,  4.35s/it]
                                                 
{'loss': '0.4629', 'grad_norm': '0.1408', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [16:07<20:17,  4.35s/it]
 42%|████▏     | 201/480 [16:11<20:07,  4.33s/it]
 42%|████▏     | 202/480 [16:16<20:12,  4.36s/it]
 42%|████▏     | 203/480 [16:20<20:04,  4.35s/it]
 42%|████▎     | 204/480 [16:24<19:53,  4.32s/it]
 43%|████▎     | 205/480 [16:29<20:08,  4.39s/it]
 43%|████▎     | 206/480 [16:33<19:54,  4.36s/it]
 43%|████▎     | 207/480 [16:37<19:49,  4.36s/it]
 43%|████▎     | 208/480 [16:42<19:40,  4.34s/it]
 44%|████▎     | 209/480 [16:46<19:34,  4.34s/it]
 44%|████▍     | 210/480 [16:50<19:42,  4.38s/it]
                                                 
{'loss': '0.4993', 'grad_norm': '0.4232', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [16:50<19:42,  4.38s/it]
 44%|████▍     | 211/480 [16:55<19:36,  4.37s/it]
 44%|████▍     | 212/480 [17:00<20:01,  4.48s/it]
 44%|████▍     | 213/480 [17:04<19:53,  4.47s/it]
 45%|████▍     | 214/480 [17:08<19:47,  4.46s/it]
 45%|████▍     | 215/480 [17:13<19:51,  4.50s/it]
 45%|████▌     | 216/480 [17:17<19:29,  4.43s/it]
 45%|████▌     | 217/480 [17:22<19:39,  4.48s/it]
 45%|████▌     | 218/480 [17:26<19:37,  4.49s/it]
 46%|████▌     | 219/480 [17:31<19:32,  4.49s/it]
 46%|████▌     | 220/480 [17:35<19:36,  4.52s/it]
                                                 
{'loss': '0.4353', 'grad_norm': '0.2265', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [17:35<19:36,  4.52s/it]
 46%|████▌     | 221/480 [17:40<19:41,  4.56s/it]
 46%|████▋     | 222/480 [17:45<19:49,  4.61s/it]
 46%|████▋     | 223/480 [17:50<20:07,  4.70s/it]
 47%|████▋     | 224/480 [17:51<15:30,  3.63s/it]
 47%|████▋     | 225/480 [17:56<16:59,  4.00s/it]
 47%|████▋     | 226/480 [18:00<17:22,  4.11s/it]
 47%|████▋     | 227/480 [18:04<17:34,  4.17s/it]
 48%|████▊     | 228/480 [18:09<17:39,  4.20s/it]
 48%|████▊     | 229/480 [18:13<17:55,  4.29s/it]
 48%|████▊     | 230/480 [18:17<17:50,  4.28s/it]
                                                 
{'loss': '0.4491', 'grad_norm': '0.2242', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [18:17<17:50,  4.28s/it]
 48%|████▊     | 231/480 [18:22<17:55,  4.32s/it]
 48%|████▊     | 232/480 [18:26<17:53,  4.33s/it]
 49%|████▊     | 233/480 [18:31<17:58,  4.36s/it]
 49%|████▉     | 234/480 [18:35<17:51,  4.35s/it]
 49%|████▉     | 235/480 [18:39<17:42,  4.34s/it]
 49%|████▉     | 236/480 [18:44<17:39,  4.34s/it]
 49%|████▉     | 237/480 [18:48<17:30,  4.32s/it]
 50%|████▉     | 238/480 [18:52<17:35,  4.36s/it]
 50%|████▉     | 239/480 [18:57<17:30,  4.36s/it]
 50%|█████     | 240/480 [19:01<17:21,  4.34s/it]
                                                 
{'loss': '0.4396', 'grad_norm': '0.2615', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [19:01<17:21,  4.34s/it]
 50%|█████     | 241/480 [19:06<17:41,  4.44s/it]
 50%|█████     | 242/480 [19:10<17:27,  4.40s/it]
 51%|█████     | 243/480 [19:15<17:34,  4.45s/it]
 51%|█████     | 244/480 [19:19<17:17,  4.40s/it]
 51%|█████     | 245/480 [19:23<17:18,  4.42s/it]
 51%|█████▏    | 246/480 [19:28<17:31,  4.49s/it]
 51%|█████▏    | 247/480 [19:33<17:34,  4.53s/it]
 52%|█████▏    | 248/480 [19:37<17:43,  4.59s/it]
 52%|█████▏    | 249/480 [19:42<17:32,  4.55s/it]
 52%|█████▏    | 250/480 [19:46<17:33,  4.58s/it]
                                                 
{'loss': '0.4299', 'grad_norm': '0.1987', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [19:46<17:33,  4.58s/it]
 52%|█████▏    | 251/480 [19:51<17:20,  4.54s/it]
 52%|█████▎    | 252/480 [19:55<17:10,  4.52s/it]
 53%|█████▎    | 253/480 [20:00<17:17,  4.57s/it]
 53%|█████▎    | 254/480 [20:04<16:54,  4.49s/it]
 53%|█████▎    | 255/480 [20:09<16:46,  4.47s/it]
 53%|█████▎    | 256/480 [20:10<12:55,  3.46s/it]
 54%|█████▎    | 257/480 [20:14<13:46,  3.70s/it]
 54%|█████▍    | 258/480 [20:18<14:25,  3.90s/it]
 54%|█████▍    | 259/480 [20:23<14:47,  4.02s/it]
 54%|█████▍    | 260/480 [20:27<15:00,  4.09s/it]
                                                 
{'loss': '0.4697', 'grad_norm': '0.3662', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [20:27<15:00,  4.09s/it]
 54%|█████▍    | 261/480 [20:32<15:28,  4.24s/it]
 55%|█████▍    | 262/480 [20:36<15:27,  4.25s/it]
 55%|█████▍    | 263/480 [20:40<15:31,  4.29s/it]
 55%|█████▌    | 264/480 [20:45<15:27,  4.29s/it]
 55%|█████▌    | 265/480 [20:49<15:33,  4.34s/it]
 55%|█████▌    | 266/480 [20:54<15:35,  4.37s/it]
 56%|█████▌    | 267/480 [20:58<15:25,  4.34s/it]
 56%|█████▌    | 268/480 [21:02<15:29,  4.38s/it]
 56%|█████▌    | 269/480 [21:07<15:30,  4.41s/it]
 56%|█████▋    | 270/480 [21:11<15:23,  4.40s/it]
                                                 
{'loss': '0.4239', 'grad_norm': '0.3069', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [21:11<15:23,  4.40s/it]
 56%|█████▋    | 271/480 [21:16<15:22,  4.41s/it]
 57%|█████▋    | 272/480 [21:20<15:12,  4.39s/it]
 57%|█████▋    | 273/480 [21:24<15:20,  4.45s/it]
 57%|█████▋    | 274/480 [21:29<15:15,  4.45s/it]
 57%|█████▋    | 275/480 [21:33<15:09,  4.44s/it]
 57%|█████▊    | 276/480 [21:38<15:06,  4.44s/it]
 58%|█████▊    | 277/480 [21:42<15:07,  4.47s/it]
 58%|█████▊    | 278/480 [21:47<15:02,  4.47s/it]
 58%|█████▊    | 279/480 [21:51<14:57,  4.46s/it]
 58%|█████▊    | 280/480 [21:56<15:08,  4.54s/it]
                                                 
{'loss': '0.4254', 'grad_norm': '0.3108', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [21:56<15:08,  4.54s/it]
 59%|█████▊    | 281/480 [22:00<14:48,  4.46s/it]
 59%|█████▉    | 282/480 [22:05<14:40,  4.44s/it]
 59%|█████▉    | 283/480 [22:09<14:26,  4.40s/it]
 59%|█████▉    | 284/480 [22:13<14:20,  4.39s/it]
 59%|█████▉    | 285/480 [22:18<14:09,  4.36s/it]
 60%|█████▉    | 286/480 [22:22<14:03,  4.35s/it]
 60%|█████▉    | 287/480 [22:26<14:08,  4.40s/it]
 60%|██████    | 288/480 [22:28<10:53,  3.41s/it]
 60%|██████    | 289/480 [22:32<11:50,  3.72s/it]
 60%|██████    | 290/480 [22:36<12:31,  3.95s/it]
                                                 
{'loss': '0.478', 'grad_norm': '0.2905', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [22:36<12:31,  3.95s/it]
 61%|██████    | 291/480 [22:41<12:45,  4.05s/it]
 61%|██████    | 292/480 [22:45<13:00,  4.15s/it]
 61%|██████    | 293/480 [22:50<13:18,  4.27s/it]
 61%|██████▏   | 294/480 [22:54<13:23,  4.32s/it]
 61%|██████▏   | 295/480 [22:58<13:21,  4.33s/it]
 62%|██████▏   | 296/480 [23:03<13:23,  4.37s/it]
 62%|██████▏   | 297/480 [23:07<13:21,  4.38s/it]
 62%|██████▏   | 298/480 [23:12<13:12,  4.36s/it]
 62%|██████▏   | 299/480 [23:16<13:14,  4.39s/it]
 62%|██████▎   | 300/480 [23:20<13:11,  4.39s/it]
                                                 
{'loss': '0.4215', 'grad_norm': '0.3269', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [23:20<13:11,  4.39s/it]
 63%|██████▎   | 301/480 [23:25<13:09,  4.41s/it]
 63%|██████▎   | 302/480 [23:29<13:11,  4.44s/it]
 63%|██████▎   | 303/480 [23:34<13:06,  4.45s/it]
 63%|██████▎   | 304/480 [23:38<12:57,  4.42s/it]
 64%|██████▎   | 305/480 [23:43<12:57,  4.44s/it]
 64%|██████▍   | 306/480 [23:47<12:56,  4.46s/it]
 64%|██████▍   | 307/480 [23:52<13:11,  4.57s/it]
 64%|██████▍   | 308/480 [23:57<13:00,  4.54s/it]
 64%|██████▍   | 309/480 [24:01<12:49,  4.50s/it]
 65%|██████▍   | 310/480 [24:05<12:36,  4.45s/it]
                                                 
{'loss': '0.4328', 'grad_norm': '0.2605', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [24:05<12:36,  4.45s/it]
 65%|██████▍   | 311/480 [24:10<12:30,  4.44s/it]
 65%|██████▌   | 312/480 [24:14<12:19,  4.40s/it]
 65%|██████▌   | 313/480 [24:18<12:15,  4.41s/it]
 65%|██████▌   | 314/480 [24:23<12:28,  4.51s/it]
 66%|██████▌   | 315/480 [24:28<12:18,  4.48s/it]
 66%|██████▌   | 316/480 [24:32<12:20,  4.51s/it]
 66%|██████▌   | 317/480 [24:36<12:03,  4.44s/it]
 66%|██████▋   | 318/480 [24:41<11:53,  4.41s/it]
 66%|██████▋   | 319/480 [24:45<11:49,  4.40s/it]
 67%|██████▋   | 320/480 [24:46<09:07,  3.42s/it]
                                                 
{'loss': '0.448', 'grad_norm': '0.7384', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [24:46<09:07,  3.42s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:00<01:33,  1.06it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:01<01:24,  1.16it/s]
+
Eval @ step 320:   3%|▎         | 3/100 [00:02<01:23,  1.17it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:03<01:23,  1.15it/s]
+
Eval @ step 320:   5%|▌         | 5/100 [00:04<01:17,  1.22it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:05<01:34,  1.01s/it]
+
Eval @ step 320:   7%|▋         | 7/100 [00:06<01:23,  1.12it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:07<01:28,  1.04it/s]
+
Eval @ step 320:   9%|▉         | 9/100 [00:08<01:24,  1.07it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:09<01:24,  1.06it/s]
+
Eval @ step 320:  11%|█         | 11/100 [00:10<01:30,  1.01s/it]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:11<01:28,  1.01s/it]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:12<01:26,  1.00it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:13<01:22,  1.04it/s]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:14<01:20,  1.06it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:14<01:15,  1.11it/s]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:15<01:13,  1.13it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:16<01:20,  1.02it/s]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:17<01:16,  1.05it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:18<01:12,  1.10it/s]
+
Eval @ step 320:  21%|██        | 21/100 [00:19<01:11,  1.10it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:20<01:15,  1.03it/s]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:21<01:09,  1.11it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:22<01:10,  1.08it/s]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:23<01:15,  1.01s/it]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:24<01:08,  1.07it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:25<01:09,  1.05it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:26<01:13,  1.03s/it]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:27<01:08,  1.03it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:28<01:05,  1.06it/s]
+
Eval @ step 320:  31%|███       | 31/100 [00:29<01:08,  1.00it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:30<01:08,  1.00s/it]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:31<01:07,  1.01s/it]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:32<01:05,  1.00it/s]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:33<00:58,  1.10it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:34<00:58,  1.10it/s]
+
Eval @ step 320:  37%|███▋      | 37/100 [00:34<00:57,  1.10it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:36<01:03,  1.02s/it]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:37<01:01,  1.00s/it]
+
Eval @ step 320:  40%|████      | 40/100 [00:38<00:57,  1.05it/s]
+
Eval @ step 320:  41%|████      | 41/100 [00:39<00:57,  1.02it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:40<01:01,  1.07s/it]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:41<00:59,  1.05s/it]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:42<00:57,  1.02s/it]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:43<00:54,  1.01it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:44<00:53,  1.02it/s]
+
Eval @ step 320:  47%|████▋     | 47/100 [00:45<00:53,  1.00s/it]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:46<00:50,  1.03it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:47<00:57,  1.12s/it]
+
Eval @ step 320:  50%|█████     | 50/100 [00:48<00:54,  1.09s/it]
+
Eval @ step 320:  51%|█████     | 51/100 [00:49<00:52,  1.08s/it]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:50<00:53,  1.11s/it]
+
Eval @ step 320:  53%|█████▎    | 53/100 [00:51<00:48,  1.03s/it]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:52<00:49,  1.07s/it]
+
Eval @ step 320:  55%|█████▌    | 55/100 [00:53<00:45,  1.01s/it]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:54<00:42,  1.04it/s]
+
Eval @ step 320:  57%|█████▋    | 57/100 [00:55<00:37,  1.13it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:56<00:37,  1.12it/s]
+
Eval @ step 320:  59%|█████▉    | 59/100 [00:56<00:35,  1.15it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:57<00:33,  1.18it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:58<00:36,  1.06it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:59<00:35,  1.07it/s]
+
Eval @ step 320:  63%|██████▎   | 63/100 [01:00<00:32,  1.12it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [01:02<00:38,  1.07s/it]
+
Eval @ step 320:  65%|██████▌   | 65/100 [01:03<00:41,  1.18s/it]
+
Eval @ step 320:  66%|██████▌   | 66/100 [01:04<00:35,  1.05s/it]
+
Eval @ step 320:  67%|██████▋   | 67/100 [01:05<00:32,  1.02it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [01:05<00:29,  1.08it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [01:07<00:30,  1.01it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [01:08<00:29,  1.01it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [01:08<00:27,  1.04it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [01:09<00:25,  1.09it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [01:10<00:24,  1.09it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [01:11<00:22,  1.17it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [01:12<00:21,  1.17it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [01:13<00:23,  1.02it/s]
+
Eval @ step 320:  77%|███████▋  | 77/100 [01:14<00:23,  1.04s/it]
+
Eval @ step 320:  78%|███████▊  | 78/100 [01:15<00:21,  1.02it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [01:16<00:20,  1.03it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [01:17<00:18,  1.06it/s]
+
Eval @ step 320:  81%|████████  | 81/100 [01:18<00:18,  1.04it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [01:19<00:17,  1.02it/s]
+
Eval @ step 320:  83%|████████▎ | 83/100 [01:20<00:16,  1.05it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [01:21<00:15,  1.03it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [01:22<00:13,  1.10it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [01:23<00:13,  1.05it/s]
+
Eval @ step 320:  87%|████████▋ | 87/100 [01:24<00:12,  1.05it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [01:24<00:10,  1.10it/s]
+
Eval @ step 320:  89%|████████▉ | 89/100 [01:26<00:10,  1.03it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [01:27<00:10,  1.09s/it]
+
Eval @ step 320:  91%|█████████ | 91/100 [01:28<00:08,  1.02it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [01:29<00:07,  1.02it/s]
+
Eval @ step 320:  93%|█████████▎| 93/100 [01:30<00:07,  1.09s/it]
+
Eval @ step 320:  94%|█████████▍| 94/100 [01:31<00:06,  1.05s/it]
+
Eval @ step 320:  95%|█████████▌| 95/100 [01:32<00:04,  1.03it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [01:33<00:03,  1.08it/s]
+
Eval @ step 320:  97%|█████████▋| 97/100 [01:34<00:03,  1.07s/it]
+
Eval @ step 320:  98%|█████████▊| 98/100 [01:35<00:01,  1.02it/s]
+
Eval @ step 320:  99%|█████████▉| 99/100 [01:36<00:00,  1.04it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [01:37<00:00,  1.05it/s]
+
                                                                    [EVAL] eval_rougeL=68.3550 (best=69.0710)
+
                                                 
{'eval_rougeL': '68.36', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [26:23<09:07,  3.42s/it]
 67%|██████▋   | 321/480 [26:28<1:27:10, 32.90s/it]
 67%|██████▋   | 322/480 [26:33<1:04:23, 24.45s/it]
 67%|██████▋   | 323/480 [26:38<48:34, 18.56s/it]  
 68%|██████▊   | 324/480 [26:42<37:15, 14.33s/it]
 68%|██████▊   | 325/480 [26:47<29:30, 11.42s/it]
 68%|██████▊   | 326/480 [26:51<24:06,  9.39s/it]
 68%|██████▊   | 327/480 [26:56<20:01,  7.86s/it]
 68%|██████▊   | 328/480 [27:00<17:14,  6.80s/it]
 69%|██████▊   | 329/480 [27:04<15:12,  6.04s/it]
 69%|██████▉   | 330/480 [27:09<13:50,  5.54s/it]
                                                 
{'loss': '0.4076', 'grad_norm': '0.3089', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [27:09<13:50,  5.54s/it]
 69%|██████▉   | 331/480 [27:13<12:53,  5.19s/it]
 69%|██████▉   | 332/480 [27:17<12:08,  4.92s/it]
 69%|██████▉   | 333/480 [27:22<11:50,  4.83s/it]
 70%|██████▉   | 334/480 [27:26<11:22,  4.68s/it]
 70%|██████▉   | 335/480 [27:31<11:07,  4.60s/it]
 70%|███████   | 336/480 [27:35<10:48,  4.51s/it]
 70%|███████   | 337/480 [27:39<10:34,  4.44s/it]
 70%|███████   | 338/480 [27:44<10:34,  4.47s/it]
 71%|███████   | 339/480 [27:48<10:22,  4.41s/it]
 71%|███████   | 340/480 [27:52<10:15,  4.40s/it]
                                                 
{'loss': '0.434', 'grad_norm': '0.2074', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [27:52<10:15,  4.40s/it]
 71%|███████   | 341/480 [27:57<10:14,  4.42s/it]
 71%|███████▏  | 342/480 [28:01<10:10,  4.42s/it]
 71%|███████▏  | 343/480 [28:06<10:12,  4.47s/it]
 72%|███████▏  | 344/480 [28:10<10:00,  4.42s/it]
 72%|███████▏  | 345/480 [28:15<10:02,  4.46s/it]
 72%|███████▏  | 346/480 [28:19<09:50,  4.41s/it]
 72%|███████▏  | 347/480 [28:24<09:56,  4.49s/it]
 72%|███████▎  | 348/480 [28:28<09:52,  4.49s/it]
 73%|███████▎  | 349/480 [28:33<10:02,  4.60s/it]
 73%|███████▎  | 350/480 [28:37<09:53,  4.57s/it]
                                                 
{'loss': '0.4468', 'grad_norm': '0.3294', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [28:37<09:53,  4.57s/it]
 73%|███████▎  | 351/480 [28:42<09:44,  4.53s/it]
 73%|███████▎  | 352/480 [28:43<07:27,  3.50s/it]
 74%|███████▎  | 353/480 [28:48<08:05,  3.82s/it]
 74%|███████▍  | 354/480 [28:52<08:25,  4.01s/it]
 74%|███████▍  | 355/480 [28:56<08:33,  4.11s/it]
 74%|███████▍  | 356/480 [29:01<08:35,  4.16s/it]
 74%|███████▍  | 357/480 [29:05<08:38,  4.21s/it]
 75%|███████▍  | 358/480 [29:09<08:38,  4.25s/it]
 75%|███████▍  | 359/480 [29:14<08:36,  4.27s/it]
 75%|███████▌  | 360/480 [29:18<08:35,  4.30s/it]
                                                 
{'loss': '0.4234', 'grad_norm': '0.2426', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [29:18<08:35,  4.30s/it]
 75%|███████▌  | 361/480 [29:22<08:36,  4.34s/it]
 75%|███████▌  | 362/480 [29:27<08:32,  4.34s/it]
 76%|███████▌  | 363/480 [29:31<08:25,  4.32s/it]
 76%|███████▌  | 364/480 [29:35<08:20,  4.31s/it]
 76%|███████▌  | 365/480 [29:40<08:18,  4.33s/it]
 76%|███████▋  | 366/480 [29:44<08:17,  4.37s/it]
 76%|███████▋  | 367/480 [29:49<08:13,  4.37s/it]
 77%|███████▋  | 368/480 [29:53<08:06,  4.34s/it]
 77%|███████▋  | 369/480 [29:57<08:06,  4.38s/it]
 77%|███████▋  | 370/480 [30:02<08:00,  4.37s/it]
                                                 
{'loss': '0.4058', 'grad_norm': '0.4602', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [30:02<08:00,  4.37s/it]
 77%|███████▋  | 371/480 [30:06<07:59,  4.40s/it]
 78%|███████▊  | 372/480 [30:11<07:54,  4.39s/it]
 78%|███████▊  | 373/480 [30:15<07:52,  4.42s/it]
 78%|███████▊  | 374/480 [30:20<07:51,  4.45s/it]
 78%|███████▊  | 375/480 [30:24<07:44,  4.42s/it]
 78%|███████▊  | 376/480 [30:28<07:41,  4.44s/it]
 79%|███████▊  | 377/480 [30:33<07:43,  4.50s/it]
 79%|███████▉  | 378/480 [30:37<07:38,  4.49s/it]
 79%|███████▉  | 379/480 [30:42<07:36,  4.52s/it]
 79%|███████▉  | 380/480 [30:46<07:29,  4.50s/it]
                                                 
{'loss': '0.4333', 'grad_norm': '0.2111', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [30:46<07:29,  4.50s/it]
 79%|███████▉  | 381/480 [30:51<07:26,  4.51s/it]
 80%|███████▉  | 382/480 [30:55<07:15,  4.44s/it]
 80%|███████▉  | 383/480 [31:00<07:05,  4.39s/it]
 80%|████████  | 384/480 [31:01<05:26,  3.40s/it]
 80%|████████  | 385/480 [31:05<05:51,  3.70s/it]
 80%|████████  | 386/480 [31:09<06:05,  3.88s/it]
 81%|████████  | 387/480 [31:14<06:20,  4.09s/it]
 81%|████████  | 388/480 [31:18<06:22,  4.15s/it]
 81%|████████  | 389/480 [31:23<06:27,  4.26s/it]
 81%|████████▏ | 390/480 [31:27<06:24,  4.28s/it]
                                                 
{'loss': '0.4568', 'grad_norm': '0.1967', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [31:27<06:24,  4.28s/it]
 81%|████████▏ | 391/480 [31:31<06:21,  4.29s/it]
 82%|████████▏ | 392/480 [31:36<06:18,  4.30s/it]
 82%|████████▏ | 393/480 [31:40<06:13,  4.30s/it]
 82%|████████▏ | 394/480 [31:45<06:18,  4.41s/it]
 82%|████████▏ | 395/480 [31:49<06:12,  4.38s/it]
 82%|████████▎ | 396/480 [31:53<06:05,  4.35s/it]
 83%|████████▎ | 397/480 [31:58<06:05,  4.40s/it]
 83%|████████▎ | 398/480 [32:02<05:57,  4.36s/it]
 83%|████████▎ | 399/480 [32:07<06:02,  4.47s/it]
 83%|████████▎ | 400/480 [32:11<05:53,  4.42s/it]
                                                 
{'loss': '0.3853', 'grad_norm': '0.1973', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [32:11<05:53,  4.42s/it]
 84%|████████▎ | 401/480 [32:16<05:55,  4.50s/it]
 84%|████████▍ | 402/480 [32:20<05:49,  4.49s/it]
 84%|████████▍ | 403/480 [32:25<05:41,  4.43s/it]
 84%|████████▍ | 404/480 [32:29<05:40,  4.48s/it]
 84%|████████▍ | 405/480 [32:34<05:35,  4.47s/it]
 85%|████████▍ | 406/480 [32:38<05:32,  4.49s/it]
 85%|████████▍ | 407/480 [32:43<05:27,  4.48s/it]
 85%|████████▌ | 408/480 [32:47<05:24,  4.50s/it]
 85%|████████▌ | 409/480 [32:52<05:18,  4.49s/it]
 85%|████████▌ | 410/480 [32:56<05:09,  4.42s/it]
                                                 
{'loss': '0.4325', 'grad_norm': '0.2846', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [32:56<05:09,  4.42s/it]
 86%|████████▌ | 411/480 [33:00<05:05,  4.43s/it]
 86%|████████▌ | 412/480 [33:05<04:57,  4.38s/it]
 86%|████████▌ | 413/480 [33:09<04:54,  4.40s/it]
 86%|████████▋ | 414/480 [33:13<04:48,  4.38s/it]
 86%|████████▋ | 415/480 [33:18<04:49,  4.45s/it]
 87%|████████▋ | 416/480 [33:19<03:41,  3.46s/it]
 87%|████████▋ | 417/480 [33:24<03:57,  3.77s/it]
 87%|████████▋ | 418/480 [33:28<04:02,  3.92s/it]
 87%|████████▋ | 419/480 [33:32<04:07,  4.05s/it]
 88%|████████▊ | 420/480 [33:36<04:07,  4.12s/it]
                                                 
{'loss': '0.4305', 'grad_norm': '0.3128', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [33:36<04:07,  4.12s/it]
 88%|████████▊ | 421/480 [33:41<04:12,  4.27s/it]
 88%|████████▊ | 422/480 [33:45<04:07,  4.27s/it]
 88%|████████▊ | 423/480 [33:50<04:06,  4.32s/it]
 88%|████████▊ | 424/480 [33:54<04:04,  4.37s/it]
 89%|████████▊ | 425/480 [33:59<03:59,  4.35s/it]
 89%|████████▉ | 426/480 [34:03<03:54,  4.35s/it]
 89%|████████▉ | 427/480 [34:07<03:52,  4.38s/it]
 89%|████████▉ | 428/480 [34:12<03:47,  4.38s/it]
 89%|████████▉ | 429/480 [34:16<03:45,  4.42s/it]
 90%|████████▉ | 430/480 [34:21<03:48,  4.57s/it]
                                                 
{'loss': '0.3986', 'grad_norm': '0.2096', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [34:21<03:48,  4.57s/it]
 90%|████████▉ | 431/480 [34:26<03:44,  4.58s/it]
 90%|█████████ | 432/480 [34:30<03:35,  4.49s/it]
 90%|█████████ | 433/480 [34:35<03:33,  4.54s/it]
 90%|█████████ | 434/480 [34:39<03:27,  4.51s/it]
 91%|█████████ | 435/480 [34:44<03:23,  4.52s/it]
 91%|█████████ | 436/480 [34:48<03:20,  4.56s/it]
 91%|█████████ | 437/480 [34:53<03:12,  4.47s/it]
 91%|█████████▏| 438/480 [34:57<03:07,  4.47s/it]
 91%|█████████▏| 439/480 [35:01<03:01,  4.41s/it]
 92%|█████████▏| 440/480 [35:06<02:57,  4.43s/it]
                                                 
{'loss': '0.4359', 'grad_norm': '0.3264', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [35:06<02:57,  4.43s/it]
 92%|█████████▏| 441/480 [35:10<02:51,  4.40s/it]
 92%|█████████▏| 442/480 [35:15<02:46,  4.39s/it]
 92%|█████████▏| 443/480 [35:19<02:41,  4.36s/it]
 92%|█████████▎| 444/480 [35:23<02:37,  4.39s/it]
 93%|█████████▎| 445/480 [35:28<02:33,  4.39s/it]
 93%|█████████▎| 446/480 [35:32<02:28,  4.37s/it]
 93%|█████████▎| 447/480 [35:37<02:25,  4.40s/it]
 93%|█████████▎| 448/480 [35:38<01:48,  3.40s/it]
 94%|█████████▎| 449/480 [35:42<01:55,  3.73s/it]
 94%|█████████▍| 450/480 [35:46<01:57,  3.93s/it]
                                                 
{'loss': '0.4207', 'grad_norm': '0.2992', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [35:46<01:57,  3.93s/it]
 94%|█████████▍| 451/480 [35:51<01:57,  4.04s/it]
 94%|█████████▍| 452/480 [35:55<01:56,  4.16s/it]
 94%|█████████▍| 453/480 [36:00<01:54,  4.23s/it]
 95%|█████████▍| 454/480 [36:04<01:50,  4.25s/it]
 95%|█████████▍| 455/480 [36:08<01:48,  4.32s/it]
 95%|█████████▌| 456/480 [36:13<01:45,  4.38s/it]
 95%|█████████▌| 457/480 [36:17<01:40,  4.35s/it]
 95%|█████████▌| 458/480 [36:22<01:37,  4.42s/it]
 96%|█████████▌| 459/480 [36:26<01:33,  4.45s/it]
 96%|█████████▌| 460/480 [36:31<01:28,  4.44s/it]
                                                 
{'loss': '0.3995', 'grad_norm': '0.2974', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [36:31<01:28,  4.44s/it]
 96%|█████████▌| 461/480 [36:35<01:24,  4.44s/it]
 96%|█████████▋| 462/480 [36:40<01:21,  4.54s/it]
 96%|█████████▋| 463/480 [36:45<01:17,  4.57s/it]
 97%|█████████▋| 464/480 [36:49<01:12,  4.56s/it]
 97%|█████████▋| 465/480 [36:53<01:07,  4.50s/it]
 97%|█████████▋| 466/480 [36:58<01:02,  4.43s/it]
 97%|█████████▋| 467/480 [37:02<00:57,  4.41s/it]
 98%|█████████▊| 468/480 [37:06<00:52,  4.37s/it]
 98%|█████████▊| 469/480 [37:11<00:48,  4.38s/it]
 98%|█████████▊| 470/480 [37:15<00:43,  4.35s/it]
                                                 
{'loss': '0.4266', 'grad_norm': '0.3847', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [37:15<00:43,  4.35s/it]
 98%|█████████▊| 471/480 [37:19<00:38,  4.33s/it]
 98%|█████████▊| 472/480 [37:24<00:35,  4.40s/it]
 99%|█████████▊| 473/480 [37:28<00:30,  4.36s/it]
 99%|█████████▉| 474/480 [37:33<00:26,  4.37s/it]
 99%|█████████▉| 475/480 [37:37<00:22,  4.41s/it]
 99%|█████████▉| 476/480 [37:42<00:17,  4.42s/it]
 99%|█████████▉| 477/480 [37:46<00:13,  4.44s/it]
100%|█████████▉| 478/480 [37:50<00:08,  4.39s/it]
100%|█████████▉| 479/480 [37:55<00:04,  4.45s/it]
100%|██████████| 480/480 [37:56<00:00,  3.44s/it]
                                                 
{'loss': '0.3976', 'grad_norm': '1.808', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [37:56<00:00,  3.44s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:01<01:50,  1.12s/it]
+
Eval @ step 480:   2%|▏         | 2/100 [00:01<01:31,  1.07it/s]
+
Eval @ step 480:   3%|▎         | 3/100 [00:02<01:27,  1.11it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:03<01:23,  1.15it/s]
+
Eval @ step 480:   5%|▌         | 5/100 [00:04<01:16,  1.24it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:05<01:31,  1.02it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:06<01:21,  1.15it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:07<01:41,  1.11s/it]
+
Eval @ step 480:   9%|▉         | 9/100 [00:08<01:40,  1.10s/it]
+
Eval @ step 480:  10%|█         | 10/100 [00:09<01:35,  1.06s/it]
+
Eval @ step 480:  11%|█         | 11/100 [00:11<01:53,  1.27s/it]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:12<01:44,  1.19s/it]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:13<01:41,  1.16s/it]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:14<01:32,  1.07s/it]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:15<01:27,  1.03s/it]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:16<01:20,  1.05it/s]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:17<01:18,  1.05it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:18<01:25,  1.04s/it]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:19<01:18,  1.04it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:20<01:13,  1.09it/s]
+
Eval @ step 480:  21%|██        | 21/100 [00:21<01:12,  1.09it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:22<01:15,  1.03it/s]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:22<01:09,  1.11it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:23<01:09,  1.09it/s]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:24<01:12,  1.03it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:25<01:08,  1.08it/s]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:26<01:09,  1.05it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:27<01:13,  1.02s/it]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:29<01:14,  1.06s/it]
+
Eval @ step 480:  30%|███       | 30/100 [00:29<01:09,  1.01it/s]
+
Eval @ step 480:  31%|███       | 31/100 [00:31<01:16,  1.11s/it]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:32<01:13,  1.08s/it]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:33<01:11,  1.07s/it]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:34<01:08,  1.04s/it]
+
Eval @ step 480:  35%|███▌      | 35/100 [00:35<01:02,  1.04it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:36<00:59,  1.07it/s]
+
Eval @ step 480:  37%|███▋      | 37/100 [00:37<01:05,  1.03s/it]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:38<01:08,  1.11s/it]
+
Eval @ step 480:  39%|███▉      | 39/100 [00:39<01:06,  1.09s/it]
+
Eval @ step 480:  40%|████      | 40/100 [00:40<01:00,  1.01s/it]
+
Eval @ step 480:  41%|████      | 41/100 [00:41<00:59,  1.01s/it]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:42<01:03,  1.09s/it]
+
Eval @ step 480:  43%|████▎     | 43/100 [00:43<01:00,  1.06s/it]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:44<00:57,  1.03s/it]
+
Eval @ step 480:  45%|████▌     | 45/100 [00:45<00:54,  1.00it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:46<00:53,  1.01it/s]
+
Eval @ step 480:  47%|████▋     | 47/100 [00:47<00:53,  1.00s/it]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:48<00:53,  1.03s/it]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:49<00:53,  1.04s/it]
+
Eval @ step 480:  50%|█████     | 50/100 [00:50<00:51,  1.03s/it]
+
Eval @ step 480:  51%|█████     | 51/100 [00:51<00:51,  1.05s/it]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:53<00:52,  1.09s/it]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:53<00:48,  1.04s/it]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:55<00:49,  1.08s/it]
+
Eval @ step 480:  55%|█████▌    | 55/100 [00:55<00:44,  1.02it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:56<00:41,  1.07it/s]
+
Eval @ step 480:  57%|█████▋    | 57/100 [00:57<00:37,  1.16it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:58<00:37,  1.12it/s]
+
Eval @ step 480:  59%|█████▉    | 59/100 [00:59<00:36,  1.13it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [01:00<00:36,  1.10it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [01:01<00:36,  1.06it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [01:02<00:35,  1.07it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [01:02<00:32,  1.13it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [01:04<00:39,  1.11s/it]
+
Eval @ step 480:  65%|██████▌   | 65/100 [01:05<00:41,  1.20s/it]
+
Eval @ step 480:  66%|██████▌   | 66/100 [01:06<00:36,  1.06s/it]
+
Eval @ step 480:  67%|██████▋   | 67/100 [01:07<00:33,  1.00s/it]
+
Eval @ step 480:  68%|██████▊   | 68/100 [01:08<00:30,  1.07it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [01:09<00:30,  1.00it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [01:10<00:29,  1.03it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [01:11<00:27,  1.06it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [01:12<00:25,  1.09it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [01:12<00:23,  1.13it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [01:13<00:21,  1.21it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [01:14<00:21,  1.17it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [01:15<00:23,  1.02it/s]
+
Eval @ step 480:  77%|███████▋  | 77/100 [01:17<00:24,  1.05s/it]
+
Eval @ step 480:  78%|███████▊  | 78/100 [01:17<00:21,  1.01it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [01:18<00:20,  1.02it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [01:19<00:18,  1.05it/s]
+
Eval @ step 480:  81%|████████  | 81/100 [01:20<00:18,  1.04it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [01:21<00:16,  1.06it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [01:22<00:15,  1.08it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [01:23<00:15,  1.05it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [01:24<00:13,  1.11it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [01:25<00:13,  1.06it/s]
+
Eval @ step 480:  87%|████████▋ | 87/100 [01:26<00:12,  1.05it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [01:27<00:10,  1.11it/s]
+
Eval @ step 480:  89%|████████▉ | 89/100 [01:28<00:10,  1.04it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [01:29<00:10,  1.07s/it]
+
Eval @ step 480:  91%|█████████ | 91/100 [01:30<00:08,  1.03it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [01:31<00:07,  1.02it/s]
+
Eval @ step 480:  93%|█████████▎| 93/100 [01:32<00:07,  1.07s/it]
+
Eval @ step 480:  94%|█████████▍| 94/100 [01:33<00:06,  1.05s/it]
+
Eval @ step 480:  95%|█████████▌| 95/100 [01:34<00:04,  1.01it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [01:35<00:03,  1.02it/s]
+
Eval @ step 480:  97%|█████████▋| 97/100 [01:36<00:03,  1.06s/it]
+
Eval @ step 480:  98%|█████████▊| 98/100 [01:37<00:01,  1.01it/s]
+
Eval @ step 480:  99%|█████████▉| 99/100 [01:38<00:00,  1.03it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [01:39<00:00,  1.04it/s]
+
                                                                    [EVAL] eval_rougeL=69.0091 (best=69.0710)
+
                                                 
{'eval_rougeL': '69.01', 'epoch': '15'}
+
100%|██████████| 480/480 [39:35<00:00,  3.44s/it]
                                                 
{'train_runtime': '2376', 'train_samples_per_second': '6.314', 'train_steps_per_second': '0.202', 'train_loss': '0.5081', 'epoch': '15'}
+
100%|██████████| 480/480 [39:35<00:00,  3.44s/it]
100%|██████████| 480/480 [39:35<00:00,  4.95s/it]
+***** train metrics *****
+  epoch                    =        15.0
+  total_flos               = 113399305GF
+  train_loss               =      0.5081
+  train_runtime            =  0:39:35.79
+  train_samples            =        1000
+  train_samples_per_second =       6.314
+  train_steps_per_second   =       0.202
+  [EVAL] Restored best model (eval_rougeL=69.0710, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:15, 19.15it/s]
Loading weights:   8%|▊         | 23/291 [00:00<00:02, 129.41it/s]
Loading weights:  15%|█▌        | 45/291 [00:00<00:01, 162.97it/s]
Loading weights:  22%|██▏       | 63/291 [00:00<00:01, 168.10it/s]
Loading weights:  28%|██▊       | 81/291 [00:00<00:01, 170.15it/s]
Loading weights:  34%|███▍      | 99/291 [00:00<00:01, 164.43it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:00, 173.14it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 172.86it/s]
Loading weights:  54%|█████▍    | 158/291 [00:00<00:00, 174.12it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 182.61it/s]
Loading weights:  70%|███████   | 204/291 [00:01<00:00, 186.90it/s]
Loading weights:  77%|███████▋  | 223/291 [00:01<00:00, 186.69it/s]
Loading weights:  83%|████████▎ | 242/291 [00:01<00:00, 183.43it/s]
Loading weights:  90%|████████▉ | 261/291 [00:01<00:00, 180.16it/s]
Loading weights:  97%|█████████▋| 282/291 [00:01<00:00, 185.61it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 175.79it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task748_glucose_reverse_cause_event_detection': PaR=66.6, shrinkage=ridge, n=1000
+  [SRT] Wired router: 11 sigs, cur=task748_glucose_reverse_cause_event_detection, mapping={'task748_glucose_reverse_cause_event_detection': 0, 'task1572_samsum_summary': 10, 'task363_sst2_polarity_classification': 9, 'task1290_xsum_summarization': 8, 'task181_outcome_extraction': 7, 'task002_quoref_answer_generation': 6, 'task1510_evalution_relation_extraction': 5, 'task639_multi_woz_user_utterance_generation': 4, 'task1729_personachat_generate_next': 3, 'task073_commonsenseqa_answer_generation': 2, 'task1590_diplomacy_text_generation': 1}
+[SRT] Router: 11 tasks, avg_PaR=66.6, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection']
+  [SRT-INIT] task_id_to_idx={'task748_glucose_reverse_cause_event_detection': 0, 'task1572_samsum_summary': 10, 'task363_sst2_polarity_classification': 9, 'task1290_xsum_summarization': 8, 'task181_outcome_extraction': 7, 'task002_quoref_answer_generation': 6, 'task1510_evalution_relation_extraction': 5, 'task639_multi_woz_user_utterance_generation': 4, 'task1729_personachat_generate_next': 3, 'task073_commonsenseqa_answer_generation': 2, 'task1590_diplomacy_text_generation': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/797 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=11
+
Continual predict:   0%|          | 1/797 [00:01<24:48,  1.87s/it]
Continual predict:   0%|          | 2/797 [00:04<31:10,  2.35s/it]
Continual predict:   0%|          | 3/797 [00:05<22:40,  1.71s/it]
Continual predict:   1%|          | 4/797 [00:06<16:48,  1.27s/it]
Continual predict:   1%|          | 5/797 [00:08<19:55,  1.51s/it]
Continual predict:   1%|          | 6/797 [00:12<34:57,  2.65s/it]
Continual predict:   1%|          | 7/797 [00:16<38:59,  2.96s/it]
Continual predict:   1%|          | 8/797 [00:18<35:40,  2.71s/it]
Continual predict:   1%|          | 9/797 [00:20<31:59,  2.44s/it]
Continual predict:   1%|▏         | 10/797 [00:21<25:31,  1.95s/it]
Continual predict:   1%|▏         | 11/797 [00:24<28:52,  2.20s/it]
Continual predict:   2%|▏         | 12/797 [00:27<33:44,  2.58s/it]
Continual predict:   2%|▏         | 13/797 [00:28<28:31,  2.18s/it]
Continual predict:   2%|▏         | 14/797 [00:31<29:52,  2.29s/it]
Continual predict:   2%|▏         | 15/797 [00:33<30:00,  2.30s/it]
Continual predict:   2%|▏         | 16/797 [00:36<31:26,  2.42s/it]
Continual predict:   2%|▏         | 17/797 [00:44<53:06,  4.09s/it]
Continual predict:   2%|▏         | 18/797 [00:46<45:38,  3.52s/it]
Continual predict:   2%|▏         | 19/797 [00:49<42:16,  3.26s/it]
Continual predict:   3%|▎         | 20/797 [00:50<35:13,  2.72s/it]
Continual predict:   3%|▎         | 21/797 [00:50<25:18,  1.96s/it]
Continual predict:   3%|▎         | 22/797 [00:51<18:20,  1.42s/it]
Continual predict:   3%|▎         | 23/797 [00:51<13:30,  1.05s/it]
Continual predict:   3%|▎         | 24/797 [00:51<10:05,  1.28it/s]
Continual predict:   3%|▎         | 25/797 [00:51<07:41,  1.67it/s]
Continual predict:   3%|▎         | 26/797 [00:51<06:01,  2.13it/s]
Continual predict:   3%|▎         | 27/797 [00:51<04:51,  2.64it/s]
Continual predict:   4%|▎         | 28/797 [00:52<04:02,  3.17it/s]
Continual predict:   4%|▎         | 29/797 [00:52<03:27,  3.69it/s]
Continual predict:   4%|▍         | 30/797 [00:52<03:05,  4.13it/s]
Continual predict:   4%|▍         | 31/797 [00:52<02:48,  4.55it/s]
Continual predict:   4%|▍         | 32/797 [00:52<02:36,  4.90it/s]
Continual predict:   4%|▍         | 33/797 [00:52<02:30,  5.09it/s]
Continual predict:   4%|▍         | 34/797 [00:53<02:25,  5.25it/s]
Continual predict:   4%|▍         | 35/797 [00:53<02:21,  5.37it/s]
Continual predict:   5%|▍         | 36/797 [00:53<02:17,  5.55it/s]
Continual predict:   5%|▍         | 37/797 [00:53<02:15,  5.59it/s]
Continual predict:   5%|▍         | 38/797 [00:53<02:13,  5.70it/s]
Continual predict:   5%|▍         | 39/797 [00:53<02:11,  5.78it/s]
Continual predict:   5%|▌         | 40/797 [00:54<02:09,  5.84it/s]
Continual predict:   5%|▌         | 41/797 [00:54<02:12,  5.73it/s]
Continual predict:   5%|▌         | 42/797 [00:54<02:10,  5.78it/s]
Continual predict:   5%|▌         | 43/797 [00:54<02:11,  5.75it/s]
Continual predict:   6%|▌         | 44/797 [00:54<02:11,  5.72it/s]
Continual predict:   6%|▌         | 45/797 [00:54<02:10,  5.76it/s]
Continual predict:   6%|▌         | 46/797 [00:55<02:09,  5.80it/s]
Continual predict:   6%|▌         | 47/797 [00:55<02:09,  5.78it/s]
Continual predict:   6%|▌         | 48/797 [00:55<02:11,  5.71it/s]
Continual predict:   6%|▌         | 49/797 [00:55<02:11,  5.71it/s]
Continual predict:   6%|▋         | 50/797 [00:55<02:09,  5.79it/s]
Continual predict:   6%|▋         | 51/797 [00:56<02:08,  5.81it/s]
Continual predict:   7%|▋         | 52/797 [00:56<02:07,  5.86it/s]
Continual predict:   7%|▋         | 53/797 [00:56<02:06,  5.90it/s]
Continual predict:   7%|▋         | 54/797 [00:56<02:07,  5.84it/s]
Continual predict:   7%|▋         | 55/797 [00:56<02:05,  5.89it/s]
Continual predict:   7%|▋         | 56/797 [00:56<02:07,  5.82it/s]
Continual predict:   7%|▋         | 57/797 [00:57<02:06,  5.86it/s]
Continual predict:   7%|▋         | 58/797 [00:57<02:06,  5.83it/s]
Continual predict:   7%|▋         | 59/797 [00:57<02:08,  5.73it/s]
Continual predict:   8%|▊         | 60/797 [00:57<02:16,  5.41it/s]
Continual predict:   8%|▊         | 61/797 [00:57<02:12,  5.55it/s]
Continual predict:   8%|▊         | 62/797 [00:57<02:10,  5.65it/s]
Continual predict:   8%|▊         | 63/797 [00:58<02:10,  5.64it/s]
Continual predict:   8%|▊         | 64/797 [00:58<02:08,  5.71it/s]
Continual predict:   8%|▊         | 65/797 [00:58<02:06,  5.76it/s]
Continual predict:   8%|▊         | 66/797 [00:58<02:05,  5.80it/s]
Continual predict:   8%|▊         | 67/797 [00:58<02:05,  5.84it/s]
Continual predict:   9%|▊         | 68/797 [00:58<02:03,  5.88it/s]
Continual predict:   9%|▊         | 69/797 [00:59<02:03,  5.87it/s]
Continual predict:   9%|▉         | 70/797 [00:59<02:02,  5.92it/s]
Continual predict:   9%|▉         | 71/797 [00:59<02:02,  5.94it/s]
Continual predict:   9%|▉         | 72/797 [00:59<02:03,  5.87it/s]
Continual predict:   9%|▉         | 73/797 [00:59<02:02,  5.91it/s]
Continual predict:   9%|▉         | 74/797 [00:59<02:03,  5.84it/s]
Continual predict:   9%|▉         | 75/797 [01:00<02:03,  5.85it/s]
Continual predict:  10%|▉         | 76/797 [01:00<02:02,  5.89it/s]
Continual predict:  10%|▉         | 77/797 [01:00<02:01,  5.91it/s]
Continual predict:  10%|▉         | 78/797 [01:00<02:01,  5.94it/s]
Continual predict:  10%|▉         | 79/797 [01:00<02:00,  5.95it/s]
Continual predict:  10%|█         | 80/797 [01:00<02:00,  5.97it/s]
Continual predict:  10%|█         | 81/797 [01:01<02:00,  5.96it/s]
Continual predict:  10%|█         | 82/797 [01:01<01:59,  5.97it/s]
Continual predict:  10%|█         | 83/797 [01:01<01:59,  5.98it/s]
Continual predict:  11%|█         | 84/797 [01:01<02:01,  5.88it/s]
Continual predict:  11%|█         | 85/797 [01:01<02:00,  5.91it/s]
Continual predict:  11%|█         | 86/797 [01:02<01:59,  5.93it/s]
Continual predict:  11%|█         | 87/797 [01:02<01:59,  5.93it/s]
Continual predict:  11%|█         | 88/797 [01:02<01:59,  5.94it/s]
Continual predict:  11%|█         | 89/797 [01:02<01:58,  5.96it/s]
Continual predict:  11%|█▏        | 90/797 [01:02<01:58,  5.98it/s]
Continual predict:  11%|█▏        | 91/797 [01:02<01:58,  5.98it/s]
Continual predict:  12%|█▏        | 92/797 [01:03<01:58,  5.95it/s]
Continual predict:  12%|█▏        | 93/797 [01:03<01:59,  5.87it/s]
Continual predict:  12%|█▏        | 94/797 [01:03<02:00,  5.82it/s]
Continual predict:  12%|█▏        | 95/797 [01:03<01:59,  5.86it/s]
Continual predict:  12%|█▏        | 96/797 [01:03<01:58,  5.90it/s]
Continual predict:  12%|█▏        | 97/797 [01:03<01:58,  5.92it/s]
Continual predict:  12%|█▏        | 98/797 [01:04<01:59,  5.86it/s]
Continual predict:  12%|█▏        | 99/797 [01:04<02:00,  5.81it/s]
Continual predict:  13%|█▎        | 100/797 [01:04<02:00,  5.77it/s]
Continual predict:  13%|█▎        | 101/797 [01:04<01:59,  5.84it/s]
Continual predict:  13%|█▎        | 102/797 [01:04<01:58,  5.89it/s]
Continual predict:  13%|█▎        | 103/797 [01:04<01:57,  5.92it/s]
Continual predict:  13%|█▎        | 104/797 [01:05<01:56,  5.94it/s]
Continual predict:  13%|█▎        | 105/797 [01:05<01:57,  5.87it/s]
Continual predict:  13%|█▎        | 106/797 [01:05<01:57,  5.89it/s]
Continual predict:  13%|█▎        | 107/797 [01:05<01:56,  5.92it/s]
Continual predict:  14%|█▎        | 108/797 [01:05<01:55,  5.95it/s]
Continual predict:  14%|█▎        | 109/797 [01:05<01:55,  5.97it/s]
Continual predict:  14%|█▍        | 110/797 [01:06<01:55,  5.96it/s]
Continual predict:  14%|█▍        | 111/797 [01:06<01:54,  5.97it/s]
Continual predict:  14%|█▍        | 112/797 [01:06<01:54,  5.97it/s]
Continual predict:  14%|█▍        | 113/797 [01:06<01:54,  5.97it/s]
Continual predict:  14%|█▍        | 114/797 [01:06<01:54,  5.98it/s]
Continual predict:  14%|█▍        | 115/797 [01:06<01:55,  5.90it/s]
Continual predict:  15%|█▍        | 116/797 [01:07<01:55,  5.91it/s]
Continual predict:  15%|█▍        | 117/797 [01:07<01:56,  5.85it/s]
Continual predict:  15%|█▍        | 118/797 [01:07<01:57,  5.80it/s]
Continual predict:  15%|█▍        | 119/797 [01:07<01:56,  5.84it/s]
Continual predict:  15%|█▌        | 120/797 [01:07<01:56,  5.82it/s]
Continual predict:  15%|█▌        | 121/797 [01:12<17:27,  1.55s/it]
Continual predict:  15%|█▌        | 122/797 [01:16<27:03,  2.41s/it]
Continual predict:  15%|█▌        | 123/797 [01:22<37:36,  3.35s/it]
Continual predict:  16%|█▌        | 124/797 [01:24<34:43,  3.10s/it]
Continual predict:  16%|█▌        | 125/797 [01:29<41:04,  3.67s/it]
Continual predict:  16%|█▌        | 126/797 [01:34<45:26,  4.06s/it]
Continual predict:  16%|█▌        | 127/797 [01:36<37:04,  3.32s/it]
Continual predict:  16%|█▌        | 128/797 [01:39<35:40,  3.20s/it]
Continual predict:  16%|█▌        | 129/797 [01:43<36:50,  3.31s/it]
Continual predict:  16%|█▋        | 130/797 [01:46<36:30,  3.28s/it]
Continual predict:  16%|█▋        | 131/797 [01:48<33:31,  3.02s/it]
Continual predict:  17%|█▋        | 132/797 [01:51<32:57,  2.97s/it]
Continual predict:  17%|█▋        | 133/797 [01:54<34:07,  3.08s/it]
Continual predict:  17%|█▋        | 134/797 [02:01<44:56,  4.07s/it]
Continual predict:  17%|█▋        | 135/797 [02:06<48:08,  4.36s/it]
Continual predict:  17%|█▋        | 136/797 [02:09<45:04,  4.09s/it]
Continual predict:  17%|█▋        | 137/797 [02:12<39:23,  3.58s/it]
Continual predict:  17%|█▋        | 138/797 [02:12<30:15,  2.75s/it]
Continual predict:  17%|█▋        | 139/797 [02:14<25:17,  2.31s/it]
Continual predict:  18%|█▊        | 140/797 [02:17<27:00,  2.47s/it]
Continual predict:  18%|█▊        | 141/797 [02:23<41:21,  3.78s/it]
Continual predict:  18%|█▊        | 142/797 [02:28<44:32,  4.08s/it]
Continual predict:  18%|█▊        | 143/797 [02:32<42:23,  3.89s/it]
Continual predict:  18%|█▊        | 144/797 [02:41<1:01:39,  5.67s/it]
Continual predict:  18%|█▊        | 145/797 [02:46<58:18,  5.37s/it]  
Continual predict:  18%|█▊        | 146/797 [02:48<46:56,  4.33s/it]
Continual predict:  18%|█▊        | 147/797 [02:53<47:43,  4.40s/it]
Continual predict:  19%|█▊        | 148/797 [02:55<40:49,  3.77s/it]
Continual predict:  19%|█▊        | 149/797 [02:58<39:20,  3.64s/it]
Continual predict:  19%|█▉        | 150/797 [03:00<34:05,  3.16s/it]
Continual predict:  19%|█▉        | 151/797 [03:03<34:09,  3.17s/it]
Continual predict:  19%|█▉        | 152/797 [03:09<43:04,  4.01s/it]
Continual predict:  19%|█▉        | 153/797 [03:12<38:33,  3.59s/it]
Continual predict:  19%|█▉        | 154/797 [03:17<43:04,  4.02s/it]
Continual predict:  19%|█▉        | 155/797 [03:21<42:16,  3.95s/it]
Continual predict:  20%|█▉        | 156/797 [03:24<41:03,  3.84s/it]
Continual predict:  20%|█▉        | 157/797 [03:29<41:44,  3.91s/it]
Continual predict:  20%|█▉        | 158/797 [03:30<32:32,  3.06s/it]
Continual predict:  20%|█▉        | 159/797 [03:37<45:48,  4.31s/it]
Continual predict:  20%|██        | 160/797 [03:39<39:42,  3.74s/it]
Continual predict:  20%|██        | 161/797 [03:42<36:06,  3.41s/it]
Continual predict:  20%|██        | 162/797 [03:47<41:09,  3.89s/it]
Continual predict:  20%|██        | 163/797 [03:49<35:43,  3.38s/it]
Continual predict:  21%|██        | 164/797 [03:51<32:36,  3.09s/it]
Continual predict:  21%|██        | 165/797 [03:55<33:16,  3.16s/it]
Continual predict:  21%|██        | 166/797 [03:57<30:32,  2.90s/it]
Continual predict:  21%|██        | 167/797 [03:58<25:38,  2.44s/it]
Continual predict:  21%|██        | 168/797 [04:01<25:24,  2.42s/it]
Continual predict:  21%|██        | 169/797 [04:03<23:29,  2.24s/it]
Continual predict:  21%|██▏       | 170/797 [04:06<26:30,  2.54s/it]
Continual predict:  21%|██▏       | 171/797 [04:08<23:47,  2.28s/it]
Continual predict:  22%|██▏       | 172/797 [04:10<25:11,  2.42s/it]
Continual predict:  22%|██▏       | 173/797 [04:14<30:23,  2.92s/it]
Continual predict:  22%|██▏       | 174/797 [04:16<25:07,  2.42s/it]
Continual predict:  22%|██▏       | 175/797 [04:19<29:00,  2.80s/it]
Continual predict:  22%|██▏       | 176/797 [04:22<29:41,  2.87s/it]
Continual predict:  22%|██▏       | 177/797 [04:25<29:43,  2.88s/it]
Continual predict:  22%|██▏       | 178/797 [04:27<24:57,  2.42s/it]
Continual predict:  22%|██▏       | 179/797 [04:28<21:10,  2.06s/it]
Continual predict:  23%|██▎       | 180/797 [04:29<19:00,  1.85s/it]
Continual predict:  23%|██▎       | 181/797 [04:31<17:29,  1.70s/it]
Continual predict:  23%|██▎       | 182/797 [04:35<27:18,  2.66s/it]
Continual predict:  23%|██▎       | 183/797 [04:36<19:47,  1.93s/it]
Continual predict:  23%|██▎       | 184/797 [04:37<17:19,  1.69s/it]
Continual predict:  23%|██▎       | 185/797 [04:40<23:08,  2.27s/it]
Continual predict:  23%|██▎       | 186/797 [04:42<19:29,  1.91s/it]
Continual predict:  23%|██▎       | 187/797 [04:45<24:00,  2.36s/it]
Continual predict:  24%|██▎       | 188/797 [04:48<25:22,  2.50s/it]
Continual predict:  24%|██▎       | 189/797 [04:53<33:36,  3.32s/it]
Continual predict:  24%|██▍       | 190/797 [04:56<34:02,  3.37s/it]
Continual predict:  24%|██▍       | 191/797 [04:58<28:01,  2.77s/it]
Continual predict:  24%|██▍       | 192/797 [05:01<30:31,  3.03s/it]
Continual predict:  24%|██▍       | 193/797 [05:06<35:43,  3.55s/it]
Continual predict:  24%|██▍       | 194/797 [05:11<40:16,  4.01s/it]
Continual predict:  24%|██▍       | 195/797 [05:17<44:52,  4.47s/it]
Continual predict:  25%|██▍       | 196/797 [05:20<41:48,  4.17s/it]
Continual predict:  25%|██▍       | 197/797 [05:24<39:39,  3.97s/it]
Continual predict:  25%|██▍       | 198/797 [05:32<50:48,  5.09s/it]
Continual predict:  25%|██▍       | 199/797 [05:33<38:53,  3.90s/it]
Continual predict:  25%|██▌       | 200/797 [05:34<31:49,  3.20s/it]
Continual predict:  25%|██▌       | 201/797 [05:37<30:15,  3.05s/it]
Continual predict:  25%|██▌       | 202/797 [05:40<31:19,  3.16s/it]
Continual predict:  25%|██▌       | 203/797 [05:46<37:46,  3.82s/it]
Continual predict:  26%|██▌       | 204/797 [05:49<35:52,  3.63s/it]
Continual predict:  26%|██▌       | 205/797 [05:56<45:20,  4.59s/it]
Continual predict:  26%|██▌       | 206/797 [05:59<40:10,  4.08s/it]
Continual predict:  26%|██▌       | 207/797 [06:03<39:33,  4.02s/it]
Continual predict:  26%|██▌       | 208/797 [06:09<46:30,  4.74s/it]
Continual predict:  26%|██▌       | 209/797 [06:12<40:46,  4.16s/it]
Continual predict:  26%|██▋       | 210/797 [06:13<33:33,  3.43s/it]
Continual predict:  26%|██▋       | 211/797 [06:16<31:28,  3.22s/it]
Continual predict:  27%|██▋       | 212/797 [06:19<31:37,  3.24s/it]
Continual predict:  27%|██▋       | 213/797 [06:23<33:35,  3.45s/it]
Continual predict:  27%|██▋       | 214/797 [06:27<34:37,  3.56s/it]
Continual predict:  27%|██▋       | 215/797 [06:28<26:39,  2.75s/it]
Continual predict:  27%|██▋       | 216/797 [06:31<28:01,  2.89s/it]
Continual predict:  27%|██▋       | 217/797 [06:34<25:53,  2.68s/it]
Continual predict:  27%|██▋       | 218/797 [06:38<30:14,  3.13s/it]
Continual predict:  27%|██▋       | 219/797 [06:41<31:00,  3.22s/it]
Continual predict:  28%|██▊       | 220/797 [06:45<34:08,  3.55s/it]
Continual predict:  28%|██▊       | 221/797 [06:46<24:29,  2.55s/it]
Continual predict:  28%|██▊       | 222/797 [06:46<17:58,  1.87s/it]
Continual predict:  28%|██▊       | 223/797 [06:46<13:11,  1.38s/it]
Continual predict:  28%|██▊       | 224/797 [06:46<09:51,  1.03s/it]
Continual predict:  28%|██▊       | 225/797 [06:47<07:31,  1.27it/s]
Continual predict:  28%|██▊       | 226/797 [06:47<05:53,  1.61it/s]
Continual predict:  28%|██▊       | 227/797 [06:47<04:54,  1.94it/s]
Continual predict:  29%|██▊       | 228/797 [06:47<04:12,  2.25it/s]
Continual predict:  29%|██▊       | 229/797 [06:48<03:37,  2.61it/s]
Continual predict:  29%|██▉       | 230/797 [06:48<03:47,  2.49it/s]
Continual predict:  29%|██▉       | 231/797 [06:48<03:17,  2.87it/s]
Continual predict:  29%|██▉       | 232/797 [06:49<02:55,  3.21it/s]
Continual predict:  29%|██▉       | 233/797 [06:49<02:40,  3.52it/s]
Continual predict:  29%|██▉       | 234/797 [06:49<02:29,  3.76it/s]
Continual predict:  29%|██▉       | 235/797 [06:49<02:21,  3.96it/s]
Continual predict:  30%|██▉       | 236/797 [06:49<02:16,  4.11it/s]
Continual predict:  30%|██▉       | 237/797 [06:50<02:12,  4.22it/s]
Continual predict:  30%|██▉       | 238/797 [06:50<02:09,  4.30it/s]
Continual predict:  30%|██▉       | 239/797 [06:50<02:07,  4.36it/s]
Continual predict:  30%|███       | 240/797 [06:50<02:31,  3.68it/s]
Continual predict:  30%|███       | 241/797 [06:51<02:22,  3.89it/s]
Continual predict:  30%|███       | 242/797 [06:51<02:16,  4.06it/s]
Continual predict:  30%|███       | 243/797 [06:51<02:12,  4.19it/s]
Continual predict:  31%|███       | 244/797 [06:51<02:09,  4.28it/s]
Continual predict:  31%|███       | 245/797 [06:52<02:07,  4.34it/s]
Continual predict:  31%|███       | 246/797 [06:52<02:05,  4.39it/s]
Continual predict:  31%|███       | 247/797 [06:52<02:13,  4.13it/s]
Continual predict:  31%|███       | 248/797 [06:52<02:09,  4.24it/s]
Continual predict:  31%|███       | 249/797 [06:53<02:16,  4.02it/s]
Continual predict:  31%|███▏      | 250/797 [06:53<02:11,  4.16it/s]
Continual predict:  31%|███▏      | 251/797 [06:53<02:16,  3.99it/s]
Continual predict:  32%|███▏      | 252/797 [06:54<05:09,  1.76it/s]
Continual predict:  32%|███▏      | 253/797 [06:55<04:20,  2.09it/s]
Continual predict:  32%|███▏      | 254/797 [06:55<03:38,  2.49it/s]
Continual predict:  32%|███▏      | 255/797 [06:55<03:08,  2.87it/s]
Continual predict:  32%|███▏      | 256/797 [06:55<02:48,  3.21it/s]
Continual predict:  32%|███▏      | 257/797 [06:56<04:05,  2.20it/s]
Continual predict:  32%|███▏      | 258/797 [06:56<03:27,  2.60it/s]
Continual predict:  32%|███▏      | 259/797 [06:57<03:00,  2.98it/s]
Continual predict:  33%|███▎      | 260/797 [06:57<02:42,  3.31it/s]
Continual predict:  33%|███▎      | 261/797 [06:57<03:28,  2.57it/s]
Continual predict:  33%|███▎      | 262/797 [06:58<03:04,  2.90it/s]
Continual predict:  33%|███▎      | 263/797 [06:58<02:52,  3.09it/s]
Continual predict:  33%|███▎      | 264/797 [06:59<05:25,  1.64it/s]
Continual predict:  33%|███▎      | 265/797 [07:00<05:17,  1.67it/s]
Continual predict:  33%|███▎      | 266/797 [07:00<05:11,  1.70it/s]
Continual predict:  34%|███▎      | 267/797 [07:01<05:18,  1.67it/s]
Continual predict:  34%|███▎      | 268/797 [07:02<05:58,  1.47it/s]
Continual predict:  34%|███▍      | 269/797 [07:02<05:50,  1.51it/s]
Continual predict:  34%|███▍      | 270/797 [07:03<05:49,  1.51it/s]
Continual predict:  34%|███▍      | 271/797 [07:04<06:08,  1.43it/s]
Continual predict:  34%|███▍      | 272/797 [07:05<06:45,  1.29it/s]
Continual predict:  34%|███▍      | 273/797 [07:05<06:29,  1.35it/s]
Continual predict:  34%|███▍      | 274/797 [07:06<06:20,  1.38it/s]
Continual predict:  35%|███▍      | 275/797 [07:07<06:04,  1.43it/s]
Continual predict:  35%|███▍      | 276/797 [07:08<07:06,  1.22it/s]
Continual predict:  35%|███▍      | 277/797 [07:09<08:25,  1.03it/s]
Continual predict:  35%|███▍      | 278/797 [07:10<08:06,  1.07it/s]
Continual predict:  35%|███▌      | 279/797 [07:11<07:35,  1.14it/s]
Continual predict:  35%|███▌      | 280/797 [07:11<06:30,  1.32it/s]
Continual predict:  35%|███▌      | 281/797 [07:12<07:07,  1.21it/s]
Continual predict:  35%|███▌      | 282/797 [07:13<07:08,  1.20it/s]
Continual predict:  36%|███▌      | 283/797 [07:14<06:47,  1.26it/s]
Continual predict:  36%|███▌      | 284/797 [07:14<05:58,  1.43it/s]
Continual predict:  36%|███▌      | 285/797 [07:15<06:16,  1.36it/s]
Continual predict:  36%|███▌      | 286/797 [07:16<06:24,  1.33it/s]
Continual predict:  36%|███▌      | 287/797 [07:17<06:29,  1.31it/s]
Continual predict:  36%|███▌      | 288/797 [07:17<06:15,  1.36it/s]
Continual predict:  36%|███▋      | 289/797 [07:18<06:51,  1.23it/s]
Continual predict:  36%|███▋      | 290/797 [07:19<06:46,  1.25it/s]
Continual predict:  37%|███▋      | 291/797 [07:20<06:19,  1.33it/s]
Continual predict:  37%|███▋      | 292/797 [07:20<05:15,  1.60it/s]
Continual predict:  37%|███▋      | 293/797 [07:21<05:50,  1.44it/s]
Continual predict:  37%|███▋      | 294/797 [07:22<05:30,  1.52it/s]
Continual predict:  37%|███▋      | 295/797 [07:22<05:00,  1.67it/s]
Continual predict:  37%|███▋      | 296/797 [07:23<06:02,  1.38it/s]
Continual predict:  37%|███▋      | 297/797 [07:24<05:35,  1.49it/s]
Continual predict:  37%|███▋      | 298/797 [07:24<05:05,  1.63it/s]
Continual predict:  38%|███▊      | 299/797 [07:25<05:53,  1.41it/s]
Continual predict:  38%|███▊      | 300/797 [07:25<04:57,  1.67it/s]
Continual predict:  38%|███▊      | 301/797 [07:26<05:49,  1.42it/s]
Continual predict:  38%|███▊      | 302/797 [07:27<05:10,  1.59it/s]
Continual predict:  38%|███▊      | 303/797 [07:28<07:33,  1.09it/s]
Continual predict:  38%|███▊      | 304/797 [07:29<06:07,  1.34it/s]
Continual predict:  38%|███▊      | 305/797 [07:30<06:35,  1.24it/s]
Continual predict:  38%|███▊      | 306/797 [07:30<06:40,  1.23it/s]
Continual predict:  39%|███▊      | 307/797 [07:31<06:41,  1.22it/s]
Continual predict:  39%|███▊      | 308/797 [07:32<05:47,  1.41it/s]
Continual predict:  39%|███▉      | 309/797 [07:32<05:11,  1.57it/s]
Continual predict:  39%|███▉      | 310/797 [07:33<05:10,  1.57it/s]
Continual predict:  39%|███▉      | 311/797 [07:33<04:59,  1.62it/s]
Continual predict:  39%|███▉      | 312/797 [07:34<05:16,  1.53it/s]
Continual predict:  39%|███▉      | 313/797 [07:35<05:52,  1.37it/s]
Continual predict:  39%|███▉      | 314/797 [07:36<05:15,  1.53it/s]
Continual predict:  40%|███▉      | 315/797 [07:36<05:55,  1.35it/s]
Continual predict:  40%|███▉      | 316/797 [07:37<06:09,  1.30it/s]
Continual predict:  40%|███▉      | 317/797 [07:38<05:06,  1.56it/s]
Continual predict:  40%|███▉      | 318/797 [07:38<04:42,  1.70it/s]
Continual predict:  40%|████      | 319/797 [07:40<07:18,  1.09it/s]
Continual predict:  40%|████      | 320/797 [07:41<06:51,  1.16it/s]
Continual predict:  40%|████      | 321/797 [07:41<06:31,  1.22it/s]
Continual predict:  40%|████      | 322/797 [07:42<05:37,  1.41it/s]
Continual predict:  41%|████      | 323/797 [07:42<05:47,  1.36it/s]
Continual predict:  41%|████      | 324/797 [07:43<06:14,  1.26it/s]
Continual predict:  41%|████      | 325/797 [07:44<06:13,  1.26it/s]
Continual predict:  41%|████      | 326/797 [07:45<05:54,  1.33it/s]
Continual predict:  41%|████      | 327/797 [07:46<06:52,  1.14it/s]
Continual predict:  41%|████      | 328/797 [07:47<07:22,  1.06it/s]
Continual predict:  41%|████▏     | 329/797 [07:48<06:27,  1.21it/s]
Continual predict:  41%|████▏     | 330/797 [07:48<05:18,  1.47it/s]
Continual predict:  42%|████▏     | 331/797 [07:49<05:10,  1.50it/s]
Continual predict:  42%|████▏     | 332/797 [07:49<04:55,  1.57it/s]
Continual predict:  42%|████▏     | 333/797 [07:50<06:22,  1.21it/s]
Continual predict:  42%|████▏     | 334/797 [07:51<06:37,  1.17it/s]
Continual predict:  42%|████▏     | 335/797 [07:52<05:40,  1.36it/s]
Continual predict:  42%|████▏     | 336/797 [07:52<05:02,  1.52it/s]
Continual predict:  42%|████▏     | 337/797 [07:53<04:53,  1.57it/s]
Continual predict:  42%|████▏     | 338/797 [07:54<04:57,  1.54it/s]
Continual predict:  43%|████▎     | 339/797 [07:54<04:44,  1.61it/s]
Continual predict:  43%|████▎     | 340/797 [07:55<05:52,  1.30it/s]
Continual predict:  43%|████▎     | 341/797 [07:56<05:52,  1.29it/s]
Continual predict:  43%|████▎     | 342/797 [07:57<05:10,  1.46it/s]
Continual predict:  43%|████▎     | 343/797 [07:57<05:24,  1.40it/s]
Continual predict:  43%|████▎     | 344/797 [07:58<06:00,  1.26it/s]
Continual predict:  43%|████▎     | 345/797 [07:59<05:16,  1.43it/s]
Continual predict:  43%|████▎     | 346/797 [08:00<05:34,  1.35it/s]
Continual predict:  44%|████▎     | 347/797 [08:01<07:18,  1.03it/s]
Continual predict:  44%|████▎     | 348/797 [08:02<05:54,  1.27it/s]
Continual predict:  44%|████▍     | 349/797 [08:02<05:08,  1.45it/s]
Continual predict:  44%|████▍     | 350/797 [08:03<05:05,  1.46it/s]
Continual predict:  44%|████▍     | 351/797 [08:04<06:11,  1.20it/s]
Continual predict:  44%|████▍     | 352/797 [08:05<06:03,  1.22it/s]
Continual predict:  44%|████▍     | 353/797 [08:06<07:12,  1.03it/s]
Continual predict:  44%|████▍     | 354/797 [08:07<07:55,  1.07s/it]
Continual predict:  45%|████▍     | 355/797 [08:08<06:56,  1.06it/s]
Continual predict:  45%|████▍     | 356/797 [08:09<06:55,  1.06it/s]
Continual predict:  45%|████▍     | 357/797 [08:10<07:19,  1.00it/s]
Continual predict:  45%|████▍     | 358/797 [08:11<06:50,  1.07it/s]
Continual predict:  45%|████▌     | 359/797 [08:11<06:12,  1.18it/s]
Continual predict:  45%|████▌     | 360/797 [08:13<07:11,  1.01it/s]
Continual predict:  45%|████▌     | 361/797 [08:14<08:35,  1.18s/it]
Continual predict:  45%|████▌     | 362/797 [08:15<07:21,  1.01s/it]
Continual predict:  46%|████▌     | 363/797 [08:15<06:09,  1.17it/s]
Continual predict:  46%|████▌     | 364/797 [08:16<05:04,  1.42it/s]
Continual predict:  46%|████▌     | 365/797 [08:16<04:10,  1.72it/s]
Continual predict:  46%|████▌     | 366/797 [08:16<03:32,  2.02it/s]
Continual predict:  46%|████▌     | 367/797 [08:17<03:06,  2.30it/s]
Continual predict:  46%|████▌     | 368/797 [08:17<02:48,  2.55it/s]
Continual predict:  46%|████▋     | 369/797 [08:17<02:34,  2.76it/s]
Continual predict:  46%|████▋     | 370/797 [08:18<02:25,  2.93it/s]
Continual predict:  47%|████▋     | 371/797 [08:18<02:30,  2.84it/s]
Continual predict:  47%|████▋     | 372/797 [08:18<02:37,  2.70it/s]
Continual predict:  47%|████▋     | 373/797 [08:19<02:42,  2.61it/s]
Continual predict:  47%|████▋     | 374/797 [08:19<02:30,  2.81it/s]
Continual predict:  47%|████▋     | 375/797 [08:19<02:37,  2.69it/s]
Continual predict:  47%|████▋     | 376/797 [08:20<02:34,  2.73it/s]
Continual predict:  47%|████▋     | 377/797 [08:20<02:25,  2.88it/s]
Continual predict:  47%|████▋     | 378/797 [08:20<02:22,  2.95it/s]
Continual predict:  48%|████▊     | 379/797 [08:21<02:17,  3.03it/s]
Continual predict:  48%|████▊     | 380/797 [08:21<02:12,  3.14it/s]
Continual predict:  48%|████▊     | 381/797 [08:21<02:16,  3.04it/s]
Continual predict:  48%|████▊     | 382/797 [08:22<02:11,  3.15it/s]
Continual predict:  48%|████▊     | 383/797 [08:22<02:08,  3.22it/s]
Continual predict:  48%|████▊     | 384/797 [08:22<02:06,  3.27it/s]
Continual predict:  48%|████▊     | 385/797 [08:23<02:04,  3.31it/s]
Continual predict:  48%|████▊     | 386/797 [08:23<02:03,  3.34it/s]
Continual predict:  49%|████▊     | 387/797 [08:23<02:02,  3.35it/s]
Continual predict:  49%|████▊     | 388/797 [08:23<02:01,  3.36it/s]
Continual predict:  49%|████▉     | 389/797 [08:24<02:08,  3.18it/s]
Continual predict:  49%|████▉     | 390/797 [08:24<02:05,  3.25it/s]
Continual predict:  49%|████▉     | 391/797 [08:24<02:03,  3.29it/s]
Continual predict:  49%|████▉     | 392/797 [08:25<02:08,  3.14it/s]
Continual predict:  49%|████▉     | 393/797 [08:25<02:05,  3.22it/s]
Continual predict:  49%|████▉     | 394/797 [08:25<02:03,  3.27it/s]
Continual predict:  50%|████▉     | 395/797 [08:26<01:54,  3.52it/s]
Continual predict:  50%|████▉     | 396/797 [08:26<02:09,  3.10it/s]
Continual predict:  50%|████▉     | 397/797 [08:26<02:05,  3.19it/s]
Continual predict:  50%|████▉     | 398/797 [08:27<02:02,  3.25it/s]
Continual predict:  50%|█████     | 399/797 [08:27<02:00,  3.29it/s]
Continual predict:  50%|█████     | 400/797 [08:27<01:59,  3.32it/s]
Continual predict:  50%|█████     | 401/797 [08:27<01:58,  3.35it/s]
Continual predict:  50%|█████     | 402/797 [08:28<01:57,  3.37it/s]
Continual predict:  51%|█████     | 403/797 [08:28<02:06,  3.11it/s]
Continual predict:  51%|█████     | 404/797 [08:28<02:03,  3.19it/s]
Continual predict:  51%|█████     | 405/797 [08:29<02:14,  2.92it/s]
Continual predict:  51%|█████     | 406/797 [08:29<02:08,  3.04it/s]
Continual predict:  51%|█████     | 407/797 [08:29<02:04,  3.14it/s]
Continual predict:  51%|█████     | 408/797 [08:30<02:00,  3.22it/s]
Continual predict:  51%|█████▏    | 409/797 [08:30<01:58,  3.27it/s]
Continual predict:  51%|█████▏    | 410/797 [08:30<02:10,  2.96it/s]
Continual predict:  52%|█████▏    | 411/797 [08:31<02:21,  2.73it/s]
Continual predict:  52%|█████▏    | 412/797 [08:31<02:14,  2.86it/s]
Continual predict:  52%|█████▏    | 413/797 [08:31<02:07,  3.01it/s]
Continual predict:  52%|█████▏    | 414/797 [08:32<02:02,  3.11it/s]
Continual predict:  52%|█████▏    | 415/797 [08:32<01:59,  3.20it/s]
Continual predict:  52%|█████▏    | 416/797 [08:32<01:57,  3.26it/s]
Continual predict:  52%|█████▏    | 417/797 [08:33<01:55,  3.30it/s]
Continual predict:  52%|█████▏    | 418/797 [08:33<02:00,  3.15it/s]
Continual predict:  53%|█████▎    | 419/797 [08:33<02:10,  2.89it/s]
Continual predict:  53%|█████▎    | 420/797 [08:34<02:18,  2.73it/s]
Continual predict:  53%|█████▎    | 421/797 [08:34<02:22,  2.63it/s]
Continual predict:  53%|█████▎    | 422/797 [08:35<02:19,  2.69it/s]
Continual predict:  53%|█████▎    | 423/797 [08:35<02:16,  2.73it/s]
Continual predict:  53%|█████▎    | 424/797 [08:35<02:21,  2.63it/s]
Continual predict:  53%|█████▎    | 425/797 [08:36<02:24,  2.57it/s]
Continual predict:  53%|█████▎    | 426/797 [08:36<02:26,  2.53it/s]
Continual predict:  54%|█████▎    | 427/797 [08:36<02:21,  2.61it/s]
Continual predict:  54%|█████▎    | 428/797 [08:37<02:11,  2.81it/s]
Continual predict:  54%|█████▍    | 429/797 [08:37<02:04,  2.96it/s]
Continual predict:  54%|█████▍    | 430/797 [08:37<01:59,  3.08it/s]
Continual predict:  54%|█████▍    | 431/797 [08:38<02:01,  3.00it/s]
Continual predict:  54%|█████▍    | 432/797 [08:38<01:57,  3.11it/s]
Continual predict:  54%|█████▍    | 433/797 [08:38<01:54,  3.19it/s]
Continual predict:  54%|█████▍    | 434/797 [08:39<01:58,  3.07it/s]
Continual predict:  55%|█████▍    | 435/797 [08:39<01:54,  3.17it/s]
Continual predict:  55%|█████▍    | 436/797 [08:39<02:04,  2.90it/s]
Continual predict:  55%|█████▍    | 437/797 [08:40<02:17,  2.62it/s]
Continual predict:  55%|█████▍    | 438/797 [08:40<02:07,  2.81it/s]
Continual predict:  55%|█████▌    | 439/797 [08:40<02:06,  2.82it/s]
Continual predict:  55%|█████▌    | 440/797 [08:41<02:06,  2.82it/s]
Continual predict:  55%|█████▌    | 441/797 [08:41<02:23,  2.48it/s]
Continual predict:  55%|█████▌    | 442/797 [08:42<02:12,  2.68it/s]
Continual predict:  56%|█████▌    | 443/797 [08:42<02:03,  2.87it/s]
Continual predict:  56%|█████▌    | 444/797 [08:42<01:57,  3.01it/s]
Continual predict:  56%|█████▌    | 445/797 [08:43<01:52,  3.12it/s]
Continual predict:  56%|█████▌    | 446/797 [08:43<01:55,  3.03it/s]
Continual predict:  56%|█████▌    | 447/797 [08:43<01:57,  2.97it/s]
Continual predict:  56%|█████▌    | 448/797 [08:44<01:59,  2.93it/s]
Continual predict:  56%|█████▋    | 449/797 [08:44<02:07,  2.73it/s]
Continual predict:  56%|█████▋    | 450/797 [08:44<02:00,  2.89it/s]
Continual predict:  57%|█████▋    | 451/797 [08:45<01:55,  3.00it/s]
Continual predict:  57%|█████▋    | 452/797 [08:45<01:50,  3.11it/s]
Continual predict:  57%|█████▋    | 453/797 [08:45<01:47,  3.19it/s]
Continual predict:  57%|█████▋    | 454/797 [08:46<01:45,  3.25it/s]
Continual predict:  57%|█████▋    | 455/797 [08:46<01:43,  3.29it/s]
Continual predict:  57%|█████▋    | 456/797 [08:46<01:48,  3.14it/s]
Continual predict:  57%|█████▋    | 457/797 [08:47<01:51,  3.04it/s]
Continual predict:  57%|█████▋    | 458/797 [08:47<01:48,  3.14it/s]
Continual predict:  58%|█████▊    | 459/797 [08:47<01:45,  3.21it/s]
Continual predict:  58%|█████▊    | 460/797 [08:47<01:43,  3.27it/s]
Continual predict:  58%|█████▊    | 461/797 [08:48<01:41,  3.31it/s]
Continual predict:  58%|█████▊    | 462/797 [08:48<01:40,  3.33it/s]
Continual predict:  58%|█████▊    | 463/797 [08:48<01:39,  3.35it/s]
Continual predict:  58%|█████▊    | 464/797 [08:49<01:48,  3.08it/s]
Continual predict:  58%|█████▊    | 465/797 [08:49<01:37,  3.40it/s]
Continual predict:  58%|█████▊    | 466/797 [08:49<01:40,  3.29it/s]
Continual predict:  59%|█████▊    | 467/797 [08:49<01:37,  3.39it/s]
Continual predict:  59%|█████▊    | 468/797 [08:50<01:40,  3.28it/s]
Continual predict:  59%|█████▉    | 469/797 [08:50<01:42,  3.20it/s]
Continual predict:  59%|█████▉    | 470/797 [08:50<01:44,  3.14it/s]
Continual predict:  59%|█████▉    | 471/797 [08:51<01:44,  3.11it/s]
Continual predict:  59%|█████▉    | 472/797 [08:52<02:48,  1.93it/s]
Continual predict:  59%|█████▉    | 473/797 [08:52<02:53,  1.87it/s]
Continual predict:  59%|█████▉    | 474/797 [08:53<02:22,  2.27it/s]
Continual predict:  60%|█████▉    | 475/797 [08:53<02:11,  2.45it/s]
Continual predict:  60%|█████▉    | 476/797 [08:53<02:09,  2.49it/s]
Continual predict:  60%|█████▉    | 477/797 [08:54<02:01,  2.63it/s]
Continual predict:  60%|█████▉    | 478/797 [08:54<02:02,  2.61it/s]
Continual predict:  60%|██████    | 479/797 [08:55<02:40,  1.98it/s]
Continual predict:  60%|██████    | 480/797 [08:55<02:23,  2.20it/s]
Continual predict:  60%|██████    | 481/797 [08:56<02:17,  2.30it/s]
Continual predict:  60%|██████    | 482/797 [08:56<02:24,  2.18it/s]
Continual predict:  61%|██████    | 483/797 [08:57<03:03,  1.71it/s]
Continual predict:  61%|██████    | 484/797 [08:58<03:51,  1.35it/s]
Continual predict:  61%|██████    | 485/797 [08:59<04:30,  1.15it/s]
Continual predict:  61%|██████    | 486/797 [09:00<04:51,  1.07it/s]
Continual predict:  61%|██████    | 487/797 [09:02<05:46,  1.12s/it]
Continual predict:  61%|██████    | 488/797 [09:02<04:43,  1.09it/s]
Continual predict:  61%|██████▏   | 489/797 [09:03<04:26,  1.16it/s]
Continual predict:  61%|██████▏   | 490/797 [09:04<04:19,  1.18it/s]
Continual predict:  62%|██████▏   | 491/797 [09:05<04:17,  1.19it/s]
Continual predict:  62%|██████▏   | 492/797 [09:06<04:26,  1.15it/s]
Continual predict:  62%|██████▏   | 493/797 [09:06<04:15,  1.19it/s]
Continual predict:  62%|██████▏   | 494/797 [09:07<04:21,  1.16it/s]
Continual predict:  62%|██████▏   | 495/797 [09:09<05:29,  1.09s/it]
Continual predict:  62%|██████▏   | 496/797 [09:10<05:04,  1.01s/it]
Continual predict:  62%|██████▏   | 497/797 [09:10<04:24,  1.13it/s]
Continual predict:  62%|██████▏   | 498/797 [09:11<04:22,  1.14it/s]
Continual predict:  63%|██████▎   | 499/797 [09:12<04:21,  1.14it/s]
Continual predict:  63%|██████▎   | 500/797 [09:13<03:43,  1.33it/s]
Continual predict:  63%|██████▎   | 501/797 [09:13<03:27,  1.43it/s]
Continual predict:  63%|██████▎   | 502/797 [09:14<03:24,  1.44it/s]
Continual predict:  63%|██████▎   | 503/797 [09:15<03:34,  1.37it/s]
Continual predict:  63%|██████▎   | 504/797 [09:15<03:21,  1.46it/s]
Continual predict:  63%|██████▎   | 505/797 [09:16<04:02,  1.21it/s]
Continual predict:  63%|██████▎   | 506/797 [09:17<04:08,  1.17it/s]
Continual predict:  64%|██████▎   | 507/797 [09:18<04:16,  1.13it/s]
Continual predict:  64%|██████▎   | 508/797 [09:19<04:02,  1.19it/s]
Continual predict:  64%|██████▍   | 509/797 [09:20<04:04,  1.18it/s]
Continual predict:  64%|██████▍   | 510/797 [09:22<05:28,  1.15s/it]
Continual predict:  64%|██████▍   | 511/797 [09:23<05:24,  1.13s/it]
Continual predict:  64%|██████▍   | 512/797 [09:23<04:44,  1.00it/s]
Continual predict:  64%|██████▍   | 513/797 [09:25<05:39,  1.20s/it]
Continual predict:  64%|██████▍   | 514/797 [09:26<05:21,  1.14s/it]
Continual predict:  65%|██████▍   | 515/797 [09:27<04:47,  1.02s/it]
Continual predict:  65%|██████▍   | 516/797 [09:28<04:22,  1.07it/s]
Continual predict:  65%|██████▍   | 517/797 [09:29<04:27,  1.05it/s]
Continual predict:  65%|██████▍   | 518/797 [09:30<04:26,  1.05it/s]
Continual predict:  65%|██████▌   | 519/797 [09:31<05:18,  1.15s/it]
Continual predict:  65%|██████▌   | 520/797 [09:32<04:37,  1.00s/it]
Continual predict:  65%|██████▌   | 521/797 [09:33<04:37,  1.00s/it]
Continual predict:  65%|██████▌   | 522/797 [09:33<04:06,  1.12it/s]
Continual predict:  66%|██████▌   | 523/797 [09:34<03:59,  1.15it/s]
Continual predict:  66%|██████▌   | 524/797 [09:35<03:42,  1.22it/s]
Continual predict:  66%|██████▌   | 525/797 [09:36<03:47,  1.20it/s]
Continual predict:  66%|██████▌   | 526/797 [09:37<03:55,  1.15it/s]
Continual predict:  66%|██████▌   | 527/797 [09:38<03:51,  1.17it/s]
Continual predict:  66%|██████▌   | 528/797 [09:39<04:39,  1.04s/it]
Continual predict:  66%|██████▋   | 529/797 [09:40<04:13,  1.06it/s]
Continual predict:  66%|██████▋   | 530/797 [09:41<04:32,  1.02s/it]
Continual predict:  67%|██████▋   | 531/797 [09:42<04:52,  1.10s/it]
Continual predict:  67%|██████▋   | 532/797 [09:44<05:22,  1.22s/it]
Continual predict:  67%|██████▋   | 533/797 [09:44<04:19,  1.02it/s]
Continual predict:  67%|██████▋   | 534/797 [09:45<04:06,  1.07it/s]
Continual predict:  67%|██████▋   | 535/797 [09:46<04:05,  1.07it/s]
Continual predict:  67%|██████▋   | 536/797 [09:46<03:19,  1.31it/s]
Continual predict:  67%|██████▋   | 537/797 [09:47<03:27,  1.25it/s]
Continual predict:  68%|██████▊   | 538/797 [09:48<03:27,  1.25it/s]
Continual predict:  68%|██████▊   | 539/797 [09:49<03:43,  1.15it/s]
Continual predict:  68%|██████▊   | 540/797 [09:50<04:14,  1.01it/s]
Continual predict:  68%|██████▊   | 541/797 [09:51<03:42,  1.15it/s]
Continual predict:  68%|██████▊   | 542/797 [09:51<03:17,  1.29it/s]
Continual predict:  68%|██████▊   | 543/797 [09:52<03:20,  1.27it/s]
Continual predict:  68%|██████▊   | 544/797 [09:53<03:49,  1.10it/s]
Continual predict:  68%|██████▊   | 545/797 [09:55<04:48,  1.14s/it]
Continual predict:  69%|██████▊   | 546/797 [09:56<04:22,  1.04s/it]
Continual predict:  69%|██████▊   | 547/797 [09:57<04:14,  1.02s/it]
Continual predict:  69%|██████▉   | 548/797 [09:58<04:05,  1.01it/s]
Continual predict:  69%|██████▉   | 549/797 [09:59<04:08,  1.00s/it]
Continual predict:  69%|██████▉   | 550/797 [10:00<04:15,  1.03s/it]
Continual predict:  69%|██████▉   | 551/797 [10:00<03:35,  1.14it/s]
Continual predict:  69%|██████▉   | 552/797 [10:01<03:17,  1.24it/s]
Continual predict:  69%|██████▉   | 553/797 [10:03<04:00,  1.01it/s]
Continual predict:  70%|██████▉   | 554/797 [10:04<04:14,  1.05s/it]
Continual predict:  70%|██████▉   | 555/797 [10:05<04:17,  1.06s/it]
Continual predict:  70%|██████▉   | 556/797 [10:05<03:47,  1.06it/s]
Continual predict:  70%|██████▉   | 557/797 [10:07<04:30,  1.13s/it]
Continual predict:  70%|███████   | 558/797 [10:08<03:48,  1.05it/s]
Continual predict:  70%|███████   | 559/797 [10:09<04:03,  1.03s/it]
Continual predict:  70%|███████   | 560/797 [10:10<04:12,  1.07s/it]
Continual predict:  70%|███████   | 561/797 [10:11<04:37,  1.17s/it]
Continual predict:  71%|███████   | 562/797 [10:13<04:49,  1.23s/it]
Continual predict:  71%|███████   | 563/797 [10:14<05:06,  1.31s/it]
Continual predict:  71%|███████   | 564/797 [10:16<05:09,  1.33s/it]
Continual predict:  71%|███████   | 565/797 [10:17<05:08,  1.33s/it]
Continual predict:  71%|███████   | 566/797 [10:18<04:31,  1.18s/it]
Continual predict:  71%|███████   | 567/797 [10:19<04:34,  1.20s/it]
Continual predict:  71%|███████▏  | 568/797 [10:20<04:07,  1.08s/it]
Continual predict:  71%|███████▏  | 569/797 [10:21<04:01,  1.06s/it]
Continual predict:  72%|███████▏  | 570/797 [10:21<03:17,  1.15it/s]
Continual predict:  72%|███████▏  | 571/797 [10:22<03:20,  1.13it/s]
Continual predict:  72%|███████▏  | 572/797 [10:23<02:56,  1.27it/s]
Continual predict:  72%|███████▏  | 573/797 [10:24<03:24,  1.10it/s]
Continual predict:  72%|███████▏  | 574/797 [10:25<03:06,  1.20it/s]
Continual predict:  72%|███████▏  | 575/797 [10:26<03:20,  1.11it/s]
Continual predict:  72%|███████▏  | 576/797 [10:26<02:58,  1.24it/s]
Continual predict:  72%|███████▏  | 577/797 [10:27<03:06,  1.18it/s]
Continual predict:  73%|███████▎  | 578/797 [10:28<02:43,  1.34it/s]
Continual predict:  73%|███████▎  | 579/797 [10:29<02:56,  1.24it/s]
Continual predict:  73%|███████▎  | 580/797 [10:30<03:32,  1.02it/s]
Continual predict:  73%|███████▎  | 581/797 [10:31<03:09,  1.14it/s]
Continual predict:  73%|███████▎  | 582/797 [10:31<02:21,  1.52it/s]
Continual predict:  73%|███████▎  | 583/797 [10:31<01:47,  1.98it/s]
Continual predict:  73%|███████▎  | 584/797 [10:31<01:24,  2.52it/s]
Continual predict:  73%|███████▎  | 585/797 [10:31<01:08,  3.11it/s]
Continual predict:  74%|███████▎  | 586/797 [10:31<00:56,  3.73it/s]
Continual predict:  74%|███████▎  | 587/797 [10:32<00:48,  4.32it/s]
Continual predict:  74%|███████▍  | 588/797 [10:32<00:43,  4.85it/s]
Continual predict:  74%|███████▍  | 589/797 [10:32<00:39,  5.31it/s]
Continual predict:  74%|███████▍  | 590/797 [10:32<00:36,  5.73it/s]
Continual predict:  74%|███████▍  | 591/797 [10:32<00:34,  6.03it/s]
Continual predict:  74%|███████▍  | 592/797 [10:32<00:32,  6.25it/s]
Continual predict:  74%|███████▍  | 593/797 [10:32<00:31,  6.46it/s]
Continual predict:  75%|███████▍  | 594/797 [10:33<00:30,  6.60it/s]
Continual predict:  75%|███████▍  | 595/797 [10:33<00:30,  6.71it/s]
Continual predict:  75%|███████▍  | 596/797 [10:33<00:29,  6.80it/s]
Continual predict:  75%|███████▍  | 597/797 [10:33<00:29,  6.85it/s]
Continual predict:  75%|███████▌  | 598/797 [10:33<00:29,  6.86it/s]
Continual predict:  75%|███████▌  | 599/797 [10:33<00:28,  6.85it/s]
Continual predict:  75%|███████▌  | 600/797 [10:33<00:28,  6.89it/s]
Continual predict:  75%|███████▌  | 601/797 [10:34<00:28,  6.88it/s]
Continual predict:  76%|███████▌  | 602/797 [10:34<00:28,  6.86it/s]
Continual predict:  76%|███████▌  | 603/797 [10:34<00:28,  6.86it/s]
Continual predict:  76%|███████▌  | 604/797 [10:34<00:28,  6.85it/s]
Continual predict:  76%|███████▌  | 605/797 [10:34<00:28,  6.85it/s]
Continual predict:  76%|███████▌  | 606/797 [10:34<00:27,  6.84it/s]
Continual predict:  76%|███████▌  | 607/797 [10:34<00:27,  6.88it/s]
Continual predict:  76%|███████▋  | 608/797 [10:35<00:27,  6.92it/s]
Continual predict:  76%|███████▋  | 609/797 [10:35<00:27,  6.94it/s]
Continual predict:  77%|███████▋  | 610/797 [10:35<00:26,  6.95it/s]
Continual predict:  77%|███████▋  | 611/797 [10:35<00:26,  6.96it/s]
Continual predict:  77%|███████▋  | 612/797 [10:35<00:26,  6.92it/s]
Continual predict:  77%|███████▋  | 613/797 [10:35<00:26,  6.90it/s]
Continual predict:  77%|███████▋  | 614/797 [10:35<00:26,  6.93it/s]
Continual predict:  77%|███████▋  | 615/797 [10:36<00:26,  6.91it/s]
Continual predict:  77%|███████▋  | 616/797 [10:36<00:26,  6.89it/s]
Continual predict:  77%|███████▋  | 617/797 [10:36<00:26,  6.92it/s]
Continual predict:  78%|███████▊  | 618/797 [10:36<00:25,  6.90it/s]
Continual predict:  78%|███████▊  | 619/797 [10:36<00:25,  6.89it/s]
Continual predict:  78%|███████▊  | 620/797 [10:36<00:25,  6.86it/s]
Continual predict:  78%|███████▊  | 621/797 [10:36<00:25,  6.86it/s]
Continual predict:  78%|███████▊  | 622/797 [10:37<00:25,  6.85it/s]
Continual predict:  78%|███████▊  | 623/797 [10:37<00:25,  6.89it/s]
Continual predict:  78%|███████▊  | 624/797 [10:37<00:24,  6.92it/s]
Continual predict:  78%|███████▊  | 625/797 [10:37<00:24,  6.90it/s]
Continual predict:  79%|███████▊  | 626/797 [10:37<00:24,  6.85it/s]
Continual predict:  79%|███████▊  | 627/797 [10:37<00:24,  6.80it/s]
Continual predict:  79%|███████▉  | 628/797 [10:37<00:24,  6.82it/s]
Continual predict:  79%|███████▉  | 629/797 [10:38<00:24,  6.79it/s]
Continual predict:  79%|███████▉  | 630/797 [10:38<00:24,  6.78it/s]
Continual predict:  79%|███████▉  | 631/797 [10:38<00:24,  6.80it/s]
Continual predict:  79%|███████▉  | 632/797 [10:38<00:24,  6.82it/s]
Continual predict:  79%|███████▉  | 633/797 [10:38<00:23,  6.87it/s]
Continual predict:  80%|███████▉  | 634/797 [10:38<00:23,  6.86it/s]
Continual predict:  80%|███████▉  | 635/797 [10:38<00:23,  6.90it/s]
Continual predict:  80%|███████▉  | 636/797 [10:39<00:23,  6.88it/s]
Continual predict:  80%|███████▉  | 637/797 [10:39<00:23,  6.88it/s]
Continual predict:  80%|████████  | 638/797 [10:39<00:23,  6.86it/s]
Continual predict:  80%|████████  | 639/797 [10:39<00:23,  6.86it/s]
Continual predict:  80%|████████  | 640/797 [10:39<00:22,  6.90it/s]
Continual predict:  80%|████████  | 641/797 [10:39<00:22,  6.93it/s]
Continual predict:  81%|████████  | 642/797 [10:40<00:22,  6.91it/s]
Continual predict:  81%|████████  | 643/797 [10:40<00:22,  6.89it/s]
Continual predict:  81%|████████  | 644/797 [10:40<00:22,  6.92it/s]
Continual predict:  81%|████████  | 645/797 [10:40<00:22,  6.91it/s]
Continual predict:  81%|████████  | 646/797 [10:40<00:21,  6.89it/s]
Continual predict:  81%|████████  | 647/797 [10:40<00:21,  6.88it/s]
Continual predict:  81%|████████▏ | 648/797 [10:40<00:21,  6.91it/s]
Continual predict:  81%|████████▏ | 649/797 [10:41<00:21,  6.90it/s]
Continual predict:  82%|████████▏ | 650/797 [10:41<00:21,  6.91it/s]
Continual predict:  82%|████████▏ | 651/797 [10:41<00:21,  6.93it/s]
Continual predict:  82%|████████▏ | 652/797 [10:41<00:20,  6.94it/s]
Continual predict:  82%|████████▏ | 653/797 [10:41<00:20,  6.96it/s]
Continual predict:  82%|████████▏ | 654/797 [10:41<00:20,  6.97it/s]
Continual predict:  82%|████████▏ | 655/797 [10:41<00:20,  6.98it/s]
Continual predict:  82%|████████▏ | 656/797 [10:42<00:20,  6.94it/s]
Continual predict:  82%|████████▏ | 657/797 [10:42<00:20,  6.92it/s]
Continual predict:  83%|████████▎ | 658/797 [10:42<00:20,  6.90it/s]
Continual predict:  83%|████████▎ | 659/797 [10:42<00:19,  6.93it/s]
Continual predict:  83%|████████▎ | 660/797 [10:42<00:19,  6.91it/s]
Continual predict:  83%|████████▎ | 661/797 [10:42<00:19,  6.89it/s]
Continual predict:  83%|████████▎ | 662/797 [10:42<00:19,  6.88it/s]
Continual predict:  83%|████████▎ | 663/797 [10:43<00:19,  6.87it/s]
Continual predict:  83%|████████▎ | 664/797 [10:43<00:19,  6.86it/s]
Continual predict:  83%|████████▎ | 665/797 [10:43<00:19,  6.86it/s]
Continual predict:  84%|████████▎ | 666/797 [10:43<00:18,  6.90it/s]
Continual predict:  84%|████████▎ | 667/797 [10:43<00:18,  6.89it/s]
Continual predict:  84%|████████▍ | 668/797 [10:43<00:18,  6.88it/s]
Continual predict:  84%|████████▍ | 669/797 [10:43<00:18,  6.91it/s]
Continual predict:  84%|████████▍ | 670/797 [10:44<00:18,  6.89it/s]
Continual predict:  84%|████████▍ | 671/797 [10:44<00:18,  6.88it/s]
Continual predict:  84%|████████▍ | 672/797 [10:44<00:18,  6.88it/s]
Continual predict:  84%|████████▍ | 673/797 [10:44<00:17,  6.90it/s]
Continual predict:  85%|████████▍ | 674/797 [10:44<00:17,  6.88it/s]
Continual predict:  85%|████████▍ | 675/797 [10:44<00:17,  6.91it/s]
Continual predict:  85%|████████▍ | 676/797 [10:44<00:17,  6.94it/s]
Continual predict:  85%|████████▍ | 677/797 [10:45<00:17,  6.95it/s]
Continual predict:  85%|████████▌ | 678/797 [10:45<00:17,  6.92it/s]
Continual predict:  85%|████████▌ | 679/797 [10:45<00:17,  6.90it/s]
Continual predict:  85%|████████▌ | 680/797 [10:45<00:16,  6.89it/s]
Continual predict:  85%|████████▌ | 681/797 [10:45<00:16,  6.87it/s]
Continual predict:  86%|████████▌ | 682/797 [10:46<00:50,  2.29it/s]
Continual predict:  86%|████████▌ | 683/797 [10:49<01:56,  1.02s/it]
Continual predict:  86%|████████▌ | 684/797 [10:50<02:16,  1.21s/it]
Continual predict:  86%|████████▌ | 685/797 [10:51<01:59,  1.07s/it]
Continual predict:  86%|████████▌ | 686/797 [10:53<02:39,  1.44s/it]
Continual predict:  86%|████████▌ | 687/797 [10:55<02:41,  1.47s/it]
Continual predict:  86%|████████▋ | 688/797 [10:59<04:19,  2.38s/it]
Continual predict:  86%|████████▋ | 689/797 [11:00<03:26,  1.91s/it]
Continual predict:  87%|████████▋ | 690/797 [11:04<04:22,  2.45s/it]
Continual predict:  87%|████████▋ | 691/797 [11:05<03:35,  2.03s/it]
Continual predict:  87%|████████▋ | 692/797 [11:06<02:50,  1.62s/it]
Continual predict:  87%|████████▋ | 693/797 [11:08<03:27,  1.99s/it]
Continual predict:  87%|████████▋ | 694/797 [11:09<02:49,  1.64s/it]
Continual predict:  87%|████████▋ | 695/797 [11:11<02:59,  1.76s/it]
Continual predict:  87%|████████▋ | 696/797 [11:13<02:42,  1.61s/it]
Continual predict:  87%|████████▋ | 697/797 [11:13<02:12,  1.32s/it]
Continual predict:  88%|████████▊ | 698/797 [11:15<02:25,  1.47s/it]
Continual predict:  88%|████████▊ | 699/797 [11:16<02:19,  1.43s/it]
Continual predict:  88%|████████▊ | 700/797 [11:18<02:17,  1.42s/it]
Continual predict:  88%|████████▊ | 701/797 [11:19<02:17,  1.43s/it]
Continual predict:  88%|████████▊ | 702/797 [11:20<02:06,  1.33s/it]
Continual predict:  88%|████████▊ | 703/797 [11:22<02:15,  1.44s/it]
Continual predict:  88%|████████▊ | 704/797 [11:23<02:04,  1.33s/it]
Continual predict:  88%|████████▊ | 705/797 [11:26<02:46,  1.81s/it]
Continual predict:  89%|████████▊ | 706/797 [11:28<02:39,  1.75s/it]
Continual predict:  89%|████████▊ | 707/797 [11:29<02:27,  1.64s/it]
Continual predict:  89%|████████▉ | 708/797 [11:32<02:59,  2.02s/it]
Continual predict:  89%|████████▉ | 709/797 [11:34<02:45,  1.88s/it]
Continual predict:  89%|████████▉ | 710/797 [11:35<02:36,  1.80s/it]
Continual predict:  89%|████████▉ | 711/797 [11:37<02:26,  1.70s/it]
Continual predict:  89%|████████▉ | 712/797 [11:38<02:18,  1.63s/it]
Continual predict:  89%|████████▉ | 713/797 [11:39<02:08,  1.53s/it]
Continual predict:  90%|████████▉ | 714/797 [11:41<02:16,  1.64s/it]
Continual predict:  90%|████████▉ | 715/797 [11:43<02:24,  1.76s/it]
Continual predict:  90%|████████▉ | 716/797 [11:45<02:14,  1.67s/it]
Continual predict:  90%|████████▉ | 717/797 [11:46<02:05,  1.57s/it]
Continual predict:  90%|█████████ | 718/797 [11:48<02:03,  1.56s/it]
Continual predict:  90%|█████████ | 719/797 [11:49<02:08,  1.64s/it]
Continual predict:  90%|█████████ | 720/797 [11:51<01:55,  1.50s/it]
Continual predict:  90%|█████████ | 721/797 [11:52<01:56,  1.53s/it]
Continual predict:  91%|█████████ | 722/797 [11:54<01:51,  1.49s/it]
Continual predict:  91%|█████████ | 723/797 [11:55<01:47,  1.46s/it]
Continual predict:  91%|█████████ | 724/797 [11:57<01:47,  1.48s/it]
Continual predict:  91%|█████████ | 725/797 [11:58<01:47,  1.50s/it]
Continual predict:  91%|█████████ | 726/797 [11:59<01:43,  1.46s/it]
Continual predict:  91%|█████████ | 727/797 [12:01<01:40,  1.44s/it]
Continual predict:  91%|█████████▏| 728/797 [12:03<01:47,  1.56s/it]
Continual predict:  91%|█████████▏| 729/797 [12:04<01:48,  1.59s/it]
Continual predict:  92%|█████████▏| 730/797 [12:06<01:49,  1.64s/it]
Continual predict:  92%|█████████▏| 731/797 [12:08<01:49,  1.67s/it]
Continual predict:  92%|█████████▏| 732/797 [12:09<01:42,  1.58s/it]
Continual predict:  92%|█████████▏| 733/797 [12:10<01:31,  1.43s/it]
Continual predict:  92%|█████████▏| 734/797 [12:12<01:43,  1.64s/it]
Continual predict:  92%|█████████▏| 735/797 [12:15<01:50,  1.79s/it]
Continual predict:  92%|█████████▏| 736/797 [12:16<01:50,  1.82s/it]
Continual predict:  92%|█████████▏| 737/797 [12:18<01:38,  1.65s/it]
Continual predict:  93%|█████████▎| 738/797 [12:19<01:37,  1.66s/it]
Continual predict:  93%|█████████▎| 739/797 [12:21<01:41,  1.75s/it]
Continual predict:  93%|█████████▎| 740/797 [12:23<01:38,  1.73s/it]
Continual predict:  93%|█████████▎| 741/797 [12:25<01:37,  1.74s/it]
Continual predict:  93%|█████████▎| 742/797 [12:26<01:33,  1.70s/it]
Continual predict:  93%|█████████▎| 743/797 [12:28<01:30,  1.67s/it]
Continual predict:  93%|█████████▎| 744/797 [12:30<01:29,  1.69s/it]
Continual predict:  93%|█████████▎| 745/797 [12:32<01:29,  1.73s/it]
Continual predict:  94%|█████████▎| 746/797 [12:33<01:27,  1.71s/it]
Continual predict:  94%|█████████▎| 747/797 [12:35<01:27,  1.75s/it]
Continual predict:  94%|█████████▍| 748/797 [12:37<01:27,  1.79s/it]
Continual predict:  94%|█████████▍| 749/797 [12:39<01:28,  1.85s/it]
Continual predict:  94%|█████████▍| 750/797 [12:40<01:20,  1.71s/it]
Continual predict:  94%|█████████▍| 751/797 [12:42<01:22,  1.79s/it]
Continual predict:  94%|█████████▍| 752/797 [12:44<01:13,  1.62s/it]
Continual predict:  94%|█████████▍| 753/797 [12:45<01:09,  1.57s/it]
Continual predict:  95%|█████████▍| 754/797 [12:46<01:02,  1.45s/it]
Continual predict:  95%|█████████▍| 755/797 [12:48<01:06,  1.59s/it]
Continual predict:  95%|█████████▍| 756/797 [12:49<01:03,  1.55s/it]
Continual predict:  95%|█████████▍| 757/797 [12:51<00:59,  1.48s/it]
Continual predict:  95%|█████████▌| 758/797 [12:52<00:56,  1.45s/it]
Continual predict:  95%|█████████▌| 759/797 [12:54<00:55,  1.47s/it]
Continual predict:  95%|█████████▌| 760/797 [12:55<00:53,  1.45s/it]
Continual predict:  95%|█████████▌| 761/797 [12:58<01:03,  1.76s/it]
Continual predict:  96%|█████████▌| 762/797 [13:00<01:07,  1.92s/it]
Continual predict:  96%|█████████▌| 763/797 [13:02<01:04,  1.89s/it]
Continual predict:  96%|█████████▌| 764/797 [13:03<01:00,  1.85s/it]
Continual predict:  96%|█████████▌| 765/797 [13:05<00:52,  1.65s/it]
Continual predict:  96%|█████████▌| 766/797 [13:06<00:51,  1.66s/it]
Continual predict:  96%|█████████▌| 767/797 [13:08<00:48,  1.62s/it]
Continual predict:  96%|█████████▋| 768/797 [13:09<00:45,  1.57s/it]
Continual predict:  96%|█████████▋| 769/797 [13:11<00:42,  1.53s/it]
Continual predict:  97%|█████████▋| 770/797 [13:12<00:41,  1.53s/it]
Continual predict:  97%|█████████▋| 771/797 [13:13<00:36,  1.42s/it]
Continual predict:  97%|█████████▋| 772/797 [13:15<00:33,  1.34s/it]
Continual predict:  97%|█████████▋| 773/797 [13:17<00:39,  1.63s/it]
Continual predict:  97%|█████████▋| 774/797 [13:18<00:36,  1.61s/it]
Continual predict:  97%|█████████▋| 775/797 [13:20<00:33,  1.54s/it]
Continual predict:  97%|█████████▋| 776/797 [13:22<00:34,  1.63s/it]
Continual predict:  97%|█████████▋| 777/797 [13:23<00:31,  1.58s/it]
Continual predict:  98%|█████████▊| 778/797 [13:25<00:30,  1.61s/it]
Continual predict:  98%|█████████▊| 779/797 [13:26<00:28,  1.57s/it]
Continual predict:  98%|█████████▊| 780/797 [13:28<00:26,  1.54s/it]
Continual predict:  98%|█████████▊| 781/797 [13:29<00:25,  1.58s/it]
Continual predict:  98%|█████████▊| 782/797 [13:31<00:22,  1.50s/it]
Continual predict:  98%|█████████▊| 783/797 [13:32<00:22,  1.57s/it]
Continual predict:  98%|█████████▊| 784/797 [13:34<00:20,  1.58s/it]
Continual predict:  98%|█████████▊| 785/797 [13:35<00:18,  1.50s/it]
Continual predict:  99%|█████████▊| 786/797 [13:37<00:16,  1.54s/it]
Continual predict:  99%|█████████▊| 787/797 [13:39<00:16,  1.61s/it]
Continual predict:  99%|█████████▉| 788/797 [13:40<00:13,  1.47s/it]
Continual predict:  99%|█████████▉| 789/797 [13:42<00:12,  1.53s/it]
Continual predict:  99%|█████████▉| 790/797 [13:44<00:12,  1.75s/it]
Continual predict:  99%|█████████▉| 791/797 [13:46<00:10,  1.73s/it]
Continual predict:  99%|█████████▉| 792/797 [13:47<00:08,  1.60s/it]
Continual predict:  99%|█████████▉| 793/797 [13:48<00:06,  1.61s/it]
Continual predict: 100%|█████████▉| 794/797 [13:51<00:05,  1.79s/it]
Continual predict: 100%|█████████▉| 795/797 [13:52<00:03,  1.64s/it]
Continual predict: 100%|█████████▉| 796/797 [13:54<00:01,  1.61s/it]
Continual predict: 100%|██████████| 797/797 [13:55<00:00,  1.61s/it]
Continual predict: 100%|██████████| 797/797 [13:55<00:00,  1.05s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot10=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=16]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=18]│
+  │  GT=task748_glucose_reverse_cause_ev  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (797/797)  label_acc=100.0% (797/797)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                                 =       15.0
+  predict_eval_rougeL                                                   =    60.6727
+  predict_eval_rougeL_for_CL                                            =    60.6727
+  predict_eval_rougeL_for_task002_quoref_answer_generation              =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                   =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction        =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                       =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation            =    11.1335
+  predict_eval_rougeL_for_task1729_personachat_generate_next            =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                    =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification          =       91.0
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation   =    15.2912
+  predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection =    66.9039
+  predict_exact_match                                                   =    43.7892
+  predict_exact_match_for_CL                                            =    43.7892
+  predict_exact_match_for_task002_quoref_answer_generation              =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                   =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction        =       98.0
+  predict_exact_match_for_task1572_samsum_summary                       =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation            =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next            =        0.0
+  predict_exact_match_for_task181_outcome_extraction                    =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification          =       91.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation   =        0.0
+  predict_exact_match_for_task748_glucose_reverse_cause_event_detection =        3.0
+  predict_gen_len                                                       =      9.808
+  predict_global_step                                                   =        480
+  predict_rouge1                                                        =    62.2164
+  predict_rouge1_for_CL                                                 =    62.2164
+  predict_rouge1_for_task002_quoref_answer_generation                   =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation            =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                        =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction             =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                            =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation                 =    13.7971
+  predict_rouge1_for_task1729_personachat_generate_next                 =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                         =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification               =       91.0
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation        =    15.2912
+  predict_rouge1_for_task748_glucose_reverse_cause_event_detection      =     68.907
+  predict_runtime                                                       = 0:13:55.63
+  predict_samples                                                       =        797
+  predict_samples_per_second                                            =      0.954
+  predict_steps_per_second                                              =      0.954
+[2026-04-30 02:59:03,365] [INFO] [launch.py:367:main] Process 1403486 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/all_results.json b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..c735f9d8396f0e8420114884fa7cf685bbe6244d
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/all_results.json
@@ -0,0 +1,57 @@
+{
+  "predict_exact_match": 39.019,
+  "predict_rouge1": 58.134,
+  "predict_eval_rougeL": 56.2843,
+  "predict_gen_len": 10.8517,
+  "predict_runtime": 1144.777478321921,
+  "predict_samples": 897,
+  "predict_samples_per_second": 0.784,
+  "predict_steps_per_second": 0.784,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 39.019,
+  "predict_rouge1_for_CL": 58.134,
+  "predict_eval_rougeL_for_CL": 56.2843,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "predict_exact_match_for_task748_glucose_reverse_cause_event_detection": 3.0,
+  "predict_rouge1_for_task748_glucose_reverse_cause_event_detection": 68.907,
+  "predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection": 66.9039,
+  "predict_exact_match_for_task511_reddit_tifu_long_text_summarization": 1.0,
+  "predict_rouge1_for_task511_reddit_tifu_long_text_summarization": 25.5967,
+  "predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization": 21.3091,
+  "train_runtime": 4617.6579,
+  "train_samples_per_second": 3.248,
+  "train_steps_per_second": 0.104,
+  "total_flos": 3.477434868726497e+17,
+  "train_loss": 2.010433729489644,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2916469676e6fa4f6385915390bffb72d69d90aa
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6c76e27af422c30337bb2a410726f747f6940556441453548d7cec89bd702ca
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0fecd705d4c7b403a67e2c713fb3292556c33843
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9578ecc953a28bd0f70fc7b73a9505d5464d9b217f7ab9cf8827cb4a2481e35
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..77d13232739c28f096e1bad282be0c2606892fac
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bf0a95045b70ca6bdd6fe1d884836bd71abb1aee64971e475906f3cbfbbe8cf
+size 1864040766
diff --git a/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/train.log b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..0486a4a268af04b092348791f6e0ae47c8ff2acb
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/train.log
@@ -0,0 +1,694 @@
+[2026-04-30 02:59:09,426] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 02:59:09,426] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29511 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task511_reddit_tifu_long_text_summarization --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization --cur_task_id 11 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[2026-04-30 02:59:12,629] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 02:59:12,629] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 02:59:12,629] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 02:59:12,629] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 02:59:12,629] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 02:59:12,630] [INFO] [launch.py:272:main] process 1540635 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task511_reddit_tifu_long_text_summarization', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization', '--cur_task_id', '11', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights']
+
+============================================================
+Task 11: task511_reddit_tifu_long_text_summarization
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/897
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:17, 16.73it/s]
Loading weights:   8%|▊         | 24/291 [00:00<00:02, 119.12it/s]
Loading weights:  16%|█▌        | 46/291 [00:00<00:01, 156.55it/s]
Loading weights:  22%|██▏       | 65/291 [00:00<00:01, 155.09it/s]
Loading weights:  29%|██▉       | 84/291 [00:00<00:01, 162.87it/s]
Loading weights:  35%|███▌      | 102/291 [00:00<00:01, 164.43it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:01, 166.45it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 162.36it/s]
Loading weights:  56%|█████▌    | 163/291 [00:01<00:00, 179.08it/s]
Loading weights:  62%|██████▏   | 181/291 [00:01<00:00, 176.13it/s]
Loading weights:  68%|██████▊   | 199/291 [00:01<00:00, 175.64it/s]
Loading weights:  75%|███████▍  | 217/291 [00:01<00:00, 167.98it/s]
Loading weights:  82%|████████▏ | 238/291 [00:01<00:00, 165.73it/s]
Loading weights:  88%|████████▊ | 257/291 [00:01<00:00, 169.05it/s]
Loading weights:  94%|█████████▍| 274/291 [00:01<00:00, 167.84it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 166.34it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1571)
+[FIX] Re-initialized lora_A/B in 768 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(11, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[9] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[10] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 11 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6764.450M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 11 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 11 sigs, cur=task511_reddit_tifu_long_text_summarization, mapping={'task511_reddit_tifu_long_text_summarization': 0, 'task1572_samsum_summary': 11, 'task363_sst2_polarity_classification': 10, 'task1290_xsum_summarization': 9, 'task181_outcome_extraction': 8, 'task002_quoref_answer_generation': 7, 'task1510_evalution_relation_extraction': 6, 'task639_multi_woz_user_utterance_generation': 5, 'task1729_personachat_generate_next': 4, 'task073_commonsenseqa_answer_generation': 3, 'task1590_diplomacy_text_generation': 2, 'task748_glucose_reverse_cause_event_detection': 1}
+[SGWI] Task 11: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 03:00:19,501] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:09<1:12:34,  9.09s/it]
  0%|          | 2/480 [00:17<1:07:28,  8.47s/it]
  1%|          | 3/480 [00:27<1:13:18,  9.22s/it]
  1%|          | 4/480 [00:35<1:09:45,  8.79s/it]
  1%|          | 5/480 [00:44<1:09:21,  8.76s/it]
  1%|▏         | 6/480 [00:51<1:06:35,  8.43s/it]
  1%|▏         | 7/480 [01:00<1:08:08,  8.64s/it]
  2%|▏         | 8/480 [01:09<1:08:38,  8.73s/it]
  2%|▏         | 9/480 [01:18<1:07:58,  8.66s/it]
  2%|▏         | 10/480 [01:27<1:08:53,  8.79s/it]
                                                  
{'loss': '2.688', 'grad_norm': '0.1709', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [01:27<1:08:53,  8.79s/it]
  2%|▏         | 11/480 [01:35<1:07:59,  8.70s/it]
  2%|▎         | 12/480 [01:44<1:07:39,  8.67s/it]
  3%|▎         | 13/480 [01:51<1:04:07,  8.24s/it]
  3%|▎         | 14/480 [01:59<1:03:49,  8.22s/it]
  3%|▎         | 15/480 [02:08<1:05:06,  8.40s/it]
  3%|▎         | 16/480 [02:17<1:05:55,  8.53s/it]
  4%|▎         | 17/480 [02:25<1:04:10,  8.32s/it]
  4%|▍         | 18/480 [02:35<1:08:25,  8.89s/it]
  4%|▍         | 19/480 [02:44<1:07:27,  8.78s/it]
  4%|▍         | 20/480 [02:52<1:05:58,  8.61s/it]
                                                  
{'loss': '2.491', 'grad_norm': '0.2341', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [02:52<1:05:58,  8.61s/it]
  4%|▍         | 21/480 [03:01<1:06:30,  8.69s/it]
  5%|▍         | 22/480 [03:09<1:05:41,  8.60s/it]
  5%|▍         | 23/480 [03:17<1:03:38,  8.36s/it]
  5%|▌         | 24/480 [03:26<1:04:32,  8.49s/it]
  5%|▌         | 25/480 [03:35<1:05:27,  8.63s/it]
  5%|▌         | 26/480 [03:43<1:05:04,  8.60s/it]
  6%|▌         | 27/480 [03:52<1:04:26,  8.54s/it]
  6%|▌         | 28/480 [04:01<1:07:05,  8.91s/it]
  6%|▌         | 29/480 [04:11<1:07:50,  9.03s/it]
  6%|▋         | 30/480 [04:19<1:06:53,  8.92s/it]
                                                  
{'loss': '2.357', 'grad_norm': '0.3867', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [04:19<1:06:53,  8.92s/it]
  6%|▋         | 31/480 [04:27<1:04:51,  8.67s/it]
  7%|▋         | 32/480 [04:29<49:45,  6.66s/it]  
  7%|▋         | 33/480 [04:38<53:30,  7.18s/it]
  7%|▋         | 34/480 [04:46<56:12,  7.56s/it]
  7%|▋         | 35/480 [04:55<58:40,  7.91s/it]
  8%|▊         | 36/480 [05:04<1:00:57,  8.24s/it]
  8%|▊         | 37/480 [05:13<1:02:36,  8.48s/it]
  8%|▊         | 38/480 [05:23<1:05:02,  8.83s/it]
  8%|▊         | 39/480 [05:32<1:05:32,  8.92s/it]
  8%|▊         | 40/480 [05:42<1:07:20,  9.18s/it]
                                                  
{'loss': '2.173', 'grad_norm': '0.2176', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [05:42<1:07:20,  9.18s/it]
  9%|▊         | 41/480 [05:50<1:04:33,  8.82s/it]
  9%|▉         | 42/480 [05:58<1:02:32,  8.57s/it]
  9%|▉         | 43/480 [06:05<1:00:29,  8.31s/it]
  9%|▉         | 44/480 [06:14<1:01:30,  8.46s/it]
  9%|▉         | 45/480 [06:24<1:03:25,  8.75s/it]
 10%|▉         | 46/480 [06:32<1:03:38,  8.80s/it]
 10%|▉         | 47/480 [06:41<1:02:20,  8.64s/it]
 10%|█         | 48/480 [06:50<1:02:44,  8.71s/it]
 10%|█         | 49/480 [06:58<1:02:50,  8.75s/it]
 10%|█         | 50/480 [07:07<1:02:18,  8.69s/it]
                                                  
{'loss': '2.066', 'grad_norm': '0.4644', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [07:07<1:02:18,  8.69s/it]
 11%|█         | 51/480 [07:16<1:03:39,  8.90s/it]
 11%|█         | 52/480 [07:25<1:02:13,  8.72s/it]
 11%|█         | 53/480 [07:34<1:02:25,  8.77s/it]
 11%|█▏        | 54/480 [07:43<1:03:28,  8.94s/it]
 11%|█▏        | 55/480 [07:52<1:03:37,  8.98s/it]
 12%|█▏        | 56/480 [08:00<1:02:03,  8.78s/it]
 12%|█▏        | 57/480 [08:09<1:01:32,  8.73s/it]
 12%|█▏        | 58/480 [08:18<1:01:25,  8.73s/it]
 12%|█▏        | 59/480 [08:26<59:35,  8.49s/it]  
 12%|█▎        | 60/480 [08:35<1:02:14,  8.89s/it]
                                                  
{'loss': '2.205', 'grad_norm': '0.3518', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [08:35<1:02:14,  8.89s/it]
 13%|█▎        | 61/480 [08:43<59:17,  8.49s/it]  
 13%|█▎        | 62/480 [08:51<58:43,  8.43s/it]
 13%|█▎        | 63/480 [08:58<56:04,  8.07s/it]
 13%|█▎        | 64/480 [09:00<43:18,  6.25s/it]
 14%|█▎        | 65/480 [09:09<48:03,  6.95s/it]
 14%|█▍        | 66/480 [09:18<51:06,  7.41s/it]
 14%|█▍        | 67/480 [09:27<55:03,  8.00s/it]
 14%|█▍        | 68/480 [09:35<54:24,  7.92s/it]
 14%|█▍        | 69/480 [09:44<58:05,  8.48s/it]
 15%|█▍        | 70/480 [09:53<58:50,  8.61s/it]
                                                
{'loss': '2.094', 'grad_norm': '0.2625', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [09:53<58:50,  8.61s/it]
 15%|█▍        | 71/480 [10:02<58:13,  8.54s/it]
 15%|█▌        | 72/480 [10:10<57:51,  8.51s/it]
 15%|█▌        | 73/480 [10:18<55:53,  8.24s/it]
 15%|█▌        | 74/480 [10:26<55:21,  8.18s/it]
 16%|█▌        | 75/480 [10:34<55:33,  8.23s/it]
 16%|█▌        | 76/480 [10:43<56:00,  8.32s/it]
 16%|█▌        | 77/480 [10:51<55:39,  8.29s/it]
 16%|█▋        | 78/480 [11:00<56:33,  8.44s/it]
 16%|█▋        | 79/480 [11:08<56:59,  8.53s/it]
 17%|█▋        | 80/480 [11:17<57:02,  8.56s/it]
                                                
{'loss': '2.155', 'grad_norm': '0.3225', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [11:17<57:02,  8.56s/it]
 17%|█▋        | 81/480 [11:26<57:36,  8.66s/it]
 17%|█▋        | 82/480 [11:35<57:59,  8.74s/it]
 17%|█▋        | 83/480 [11:45<1:00:14,  9.10s/it]
 18%|█▊        | 84/480 [11:53<57:24,  8.70s/it]  
 18%|█▊        | 85/480 [12:02<58:36,  8.90s/it]
 18%|█▊        | 86/480 [12:10<56:32,  8.61s/it]
 18%|█▊        | 87/480 [12:18<55:10,  8.42s/it]
 18%|█▊        | 88/480 [12:26<55:16,  8.46s/it]
 19%|█▊        | 89/480 [12:35<55:08,  8.46s/it]
 19%|█▉        | 90/480 [12:44<56:18,  8.66s/it]
                                                
{'loss': '2.021', 'grad_norm': '0.4036', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [12:44<56:18,  8.66s/it]
 19%|█▉        | 91/480 [12:53<57:32,  8.88s/it]
 19%|█▉        | 92/480 [13:03<58:23,  9.03s/it]
 19%|█▉        | 93/480 [13:11<57:32,  8.92s/it]
 20%|█▉        | 94/480 [13:21<58:15,  9.06s/it]
 20%|█▉        | 95/480 [13:29<57:04,  8.90s/it]
 20%|██        | 96/480 [13:32<44:05,  6.89s/it]
 20%|██        | 97/480 [13:40<47:02,  7.37s/it]
 20%|██        | 98/480 [13:49<50:20,  7.91s/it]
 21%|██        | 99/480 [13:57<50:39,  7.98s/it]
 21%|██        | 100/480 [14:06<51:02,  8.06s/it]
                                                 
{'loss': '2.081', 'grad_norm': '0.493', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [14:06<51:02,  8.06s/it]
 21%|██        | 101/480 [14:14<52:08,  8.25s/it]
 21%|██▏       | 102/480 [14:23<51:56,  8.24s/it]
 21%|██▏       | 103/480 [14:30<50:47,  8.08s/it]
 22%|██▏       | 104/480 [14:39<52:15,  8.34s/it]
 22%|██▏       | 105/480 [14:49<54:01,  8.64s/it]
 22%|██▏       | 106/480 [14:57<53:01,  8.51s/it]
 22%|██▏       | 107/480 [15:05<51:45,  8.33s/it]
 22%|██▎       | 108/480 [15:13<52:15,  8.43s/it]
 23%|██▎       | 109/480 [15:23<53:42,  8.69s/it]
 23%|██▎       | 110/480 [15:32<54:11,  8.79s/it]
                                                 
{'loss': '2.06', 'grad_norm': '0.4236', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [15:32<54:11,  8.79s/it]
 23%|██▎       | 111/480 [15:41<55:34,  9.04s/it]
 23%|██▎       | 112/480 [15:50<54:35,  8.90s/it]
 24%|██▎       | 113/480 [15:58<52:44,  8.62s/it]
 24%|██▍       | 114/480 [16:06<51:51,  8.50s/it]
 24%|██▍       | 115/480 [16:14<51:08,  8.41s/it]
 24%|██▍       | 116/480 [16:23<52:12,  8.60s/it]
 24%|██▍       | 117/480 [16:32<51:54,  8.58s/it]
 25%|██▍       | 118/480 [16:41<52:28,  8.70s/it]
 25%|██▍       | 119/480 [16:48<50:36,  8.41s/it]
 25%|██▌       | 120/480 [16:58<51:43,  8.62s/it]
                                                 
{'loss': '2.111', 'grad_norm': '0.266', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [16:58<51:43,  8.62s/it]
 25%|██▌       | 121/480 [17:06<51:55,  8.68s/it]
 25%|██▌       | 122/480 [17:15<52:20,  8.77s/it]
 26%|██▌       | 123/480 [17:24<51:33,  8.67s/it]
 26%|██▌       | 124/480 [17:33<52:19,  8.82s/it]
 26%|██▌       | 125/480 [17:42<51:50,  8.76s/it]
 26%|██▋       | 126/480 [17:51<52:44,  8.94s/it]
 26%|██▋       | 127/480 [18:00<53:00,  9.01s/it]
 27%|██▋       | 128/480 [18:02<40:14,  6.86s/it]
 27%|██▋       | 129/480 [18:12<45:20,  7.75s/it]
 27%|██▋       | 130/480 [18:21<47:47,  8.19s/it]
                                                 
{'loss': '1.984', 'grad_norm': '0.2733', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [18:21<47:47,  8.19s/it]
 27%|██▋       | 131/480 [18:29<46:58,  8.08s/it]
 28%|██▊       | 132/480 [18:38<48:32,  8.37s/it]
 28%|██▊       | 133/480 [18:46<48:13,  8.34s/it]
 28%|██▊       | 134/480 [18:56<50:19,  8.73s/it]
 28%|██▊       | 135/480 [19:04<49:18,  8.58s/it]
 28%|██▊       | 136/480 [19:13<49:48,  8.69s/it]
 29%|██▊       | 137/480 [19:22<49:46,  8.71s/it]
 29%|██▉       | 138/480 [19:30<49:07,  8.62s/it]
 29%|██▉       | 139/480 [19:38<48:21,  8.51s/it]
 29%|██▉       | 140/480 [19:48<49:53,  8.80s/it]
                                                 
{'loss': '2.044', 'grad_norm': '0.3612', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [19:48<49:53,  8.80s/it]
 29%|██▉       | 141/480 [19:57<49:48,  8.82s/it]
 30%|██▉       | 142/480 [20:05<48:54,  8.68s/it]
 30%|██▉       | 143/480 [20:14<48:46,  8.68s/it]
 30%|███       | 144/480 [20:22<47:37,  8.50s/it]
 30%|███       | 145/480 [20:30<47:05,  8.43s/it]
 30%|███       | 146/480 [20:39<47:18,  8.50s/it]
 31%|███       | 147/480 [20:48<48:39,  8.77s/it]
 31%|███       | 148/480 [20:57<49:07,  8.88s/it]
 31%|███       | 149/480 [21:06<47:55,  8.69s/it]
 31%|███▏      | 150/480 [21:15<48:50,  8.88s/it]
                                                 
{'loss': '2.012', 'grad_norm': '0.565', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [21:15<48:50,  8.88s/it]
 31%|███▏      | 151/480 [21:23<48:01,  8.76s/it]
 32%|███▏      | 152/480 [21:32<47:21,  8.66s/it]
 32%|███▏      | 153/480 [21:40<46:52,  8.60s/it]
 32%|███▏      | 154/480 [21:49<46:44,  8.60s/it]
 32%|███▏      | 155/480 [21:57<45:23,  8.38s/it]
 32%|███▎      | 156/480 [22:06<46:26,  8.60s/it]
 33%|███▎      | 157/480 [22:14<45:17,  8.41s/it]
 33%|███▎      | 158/480 [22:23<45:40,  8.51s/it]
 33%|███▎      | 159/480 [22:31<45:17,  8.47s/it]
 33%|███▎      | 160/480 [22:33<34:56,  6.55s/it]
                                                 
{'loss': '2.054', 'grad_norm': '1.301', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [22:33<34:56,  6.55s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:01<02:00,  1.22s/it]
+
Eval @ step 160:   2%|▏         | 2/100 [00:05<04:49,  2.95s/it]
+
Eval @ step 160:   3%|▎         | 3/100 [00:06<03:04,  1.90s/it]
+
Eval @ step 160:   4%|▍         | 4/100 [00:06<02:19,  1.45s/it]
+
Eval @ step 160:   5%|▌         | 5/100 [00:07<02:08,  1.35s/it]
+
Eval @ step 160:   6%|▌         | 6/100 [00:08<01:37,  1.04s/it]
+
Eval @ step 160:   7%|▋         | 7/100 [00:10<02:07,  1.37s/it]
+
Eval @ step 160:   8%|▊         | 8/100 [00:12<02:12,  1.44s/it]
+
Eval @ step 160:   9%|▉         | 9/100 [00:13<02:06,  1.39s/it]
+
Eval @ step 160:  10%|█         | 10/100 [00:14<01:48,  1.21s/it]
+
Eval @ step 160:  11%|█         | 11/100 [00:15<01:57,  1.32s/it]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:16<01:51,  1.27s/it]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:17<01:43,  1.19s/it]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:18<01:39,  1.15s/it]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:26<04:11,  2.96s/it]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:26<03:13,  2.30s/it]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:28<02:45,  2.00s/it]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:31<03:06,  2.27s/it]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:31<02:28,  1.84s/it]
+
Eval @ step 160:  20%|██        | 20/100 [00:32<01:54,  1.43s/it]
+
Eval @ step 160:  21%|██        | 21/100 [00:33<01:43,  1.31s/it]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:35<02:03,  1.58s/it]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:37<01:57,  1.53s/it]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:38<01:46,  1.41s/it]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:39<01:51,  1.49s/it]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:41<02:02,  1.65s/it]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:48<03:44,  3.07s/it]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:49<03:10,  2.65s/it]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:57<04:43,  4.00s/it]
+
Eval @ step 160:  30%|███       | 30/100 [00:57<03:28,  2.98s/it]
+
Eval @ step 160:  31%|███       | 31/100 [01:01<03:40,  3.19s/it]
+
Eval @ step 160:  32%|███▏      | 32/100 [01:02<03:03,  2.70s/it]
+
Eval @ step 160:  33%|███▎      | 33/100 [01:04<02:38,  2.37s/it]
+
Eval @ step 160:  34%|███▍      | 34/100 [01:11<04:10,  3.80s/it]
+
Eval @ step 160:  35%|███▌      | 35/100 [01:12<03:10,  2.93s/it]
+
Eval @ step 160:  36%|███▌      | 36/100 [01:13<02:22,  2.22s/it]
+
Eval @ step 160:  37%|███▋      | 37/100 [01:14<02:05,  1.99s/it]
+
Eval @ step 160:  38%|███▊      | 38/100 [01:15<01:42,  1.65s/it]
+
Eval @ step 160:  39%|███▉      | 39/100 [01:17<01:45,  1.73s/it]
+
Eval @ step 160:  40%|████      | 40/100 [01:18<01:42,  1.71s/it]
+
Eval @ step 160:  41%|████      | 41/100 [01:20<01:43,  1.75s/it]
+
Eval @ step 160:  42%|████▏     | 42/100 [01:27<03:15,  3.37s/it]
+
Eval @ step 160:  43%|████▎     | 43/100 [01:30<03:01,  3.19s/it]
+
Eval @ step 160:  44%|████▍     | 44/100 [01:31<02:24,  2.58s/it]
+
Eval @ step 160:  45%|████▌     | 45/100 [01:33<01:59,  2.17s/it]
+
Eval @ step 160:  46%|████▌     | 46/100 [01:34<01:40,  1.86s/it]
+
Eval @ step 160:  47%|████▋     | 47/100 [01:37<01:53,  2.14s/it]
+
Eval @ step 160:  48%|████▊     | 48/100 [01:38<01:46,  2.06s/it]
+
Eval @ step 160:  49%|████▉     | 49/100 [01:40<01:40,  1.97s/it]
+
Eval @ step 160:  50%|█████     | 50/100 [01:47<02:56,  3.52s/it]
+
Eval @ step 160:  51%|█████     | 51/100 [01:54<03:44,  4.57s/it]
+
Eval @ step 160:  52%|█████▏    | 52/100 [01:55<02:43,  3.41s/it]
+
Eval @ step 160:  53%|█████▎    | 53/100 [01:57<02:16,  2.91s/it]
+
Eval @ step 160:  54%|█████▍    | 54/100 [01:58<01:47,  2.33s/it]
+
Eval @ step 160:  55%|█████▌    | 55/100 [01:59<01:26,  1.93s/it]
+
Eval @ step 160:  56%|█████▌    | 56/100 [02:06<02:33,  3.49s/it]
+
Eval @ step 160:  57%|█████▋    | 57/100 [02:13<03:17,  4.59s/it]
+
Eval @ step 160:  58%|█████▊    | 58/100 [02:14<02:25,  3.47s/it]
+
Eval @ step 160:  59%|█████▉    | 59/100 [02:15<01:47,  2.61s/it]
+
Eval @ step 160:  60%|██████    | 60/100 [02:15<01:24,  2.10s/it]
+
Eval @ step 160:  61%|██████    | 61/100 [02:17<01:10,  1.80s/it]
+
Eval @ step 160:  62%|██████▏   | 62/100 [02:19<01:16,  2.02s/it]
+
Eval @ step 160:  63%|██████▎   | 63/100 [02:21<01:11,  1.93s/it]
+
Eval @ step 160:  64%|██████▍   | 64/100 [02:28<02:05,  3.50s/it]
+
Eval @ step 160:  65%|██████▌   | 65/100 [02:28<01:29,  2.56s/it]
+
Eval @ step 160:  66%|██████▌   | 66/100 [02:31<01:25,  2.52s/it]
+
Eval @ step 160:  67%|██████▋   | 67/100 [02:32<01:06,  2.01s/it]
+
Eval @ step 160:  68%|██████▊   | 68/100 [02:33<00:56,  1.77s/it]
+
Eval @ step 160:  69%|██████▉   | 69/100 [02:34<00:47,  1.52s/it]
+
Eval @ step 160:  70%|███████   | 70/100 [02:35<00:44,  1.49s/it]
+
Eval @ step 160:  71%|███████   | 71/100 [02:36<00:39,  1.37s/it]
+
Eval @ step 160:  72%|███████▏  | 72/100 [02:39<00:50,  1.80s/it]
+
Eval @ step 160:  73%|███████▎  | 73/100 [02:40<00:39,  1.48s/it]
+
Eval @ step 160:  74%|███████▍  | 74/100 [02:41<00:34,  1.33s/it]
+
Eval @ step 160:  75%|███████▌  | 75/100 [02:43<00:39,  1.57s/it]
+
Eval @ step 160:  76%|███████▌  | 76/100 [02:45<00:42,  1.76s/it]
+
Eval @ step 160:  77%|███████▋  | 77/100 [02:49<00:56,  2.47s/it]
+
Eval @ step 160:  78%|███████▊  | 78/100 [02:50<00:46,  2.11s/it]
+
Eval @ step 160:  79%|███████▉  | 79/100 [02:51<00:35,  1.70s/it]
+
Eval @ step 160:  80%|████████  | 80/100 [02:52<00:31,  1.58s/it]
+
Eval @ step 160:  81%|████████  | 81/100 [02:57<00:46,  2.42s/it]
+
Eval @ step 160:  82%|████████▏ | 82/100 [02:58<00:39,  2.18s/it]
+
Eval @ step 160:  83%|████████▎ | 83/100 [03:01<00:40,  2.41s/it]
+
Eval @ step 160:  84%|████████▍ | 84/100 [03:03<00:36,  2.30s/it]
+
Eval @ step 160:  85%|████████▌ | 85/100 [03:05<00:29,  1.96s/it]
+
Eval @ step 160:  86%|████████▌ | 86/100 [03:05<00:22,  1.63s/it]
+
Eval @ step 160:  87%|████████▋ | 87/100 [03:06<00:18,  1.41s/it]
+
Eval @ step 160:  88%|████████▊ | 88/100 [03:07<00:15,  1.30s/it]
+
Eval @ step 160:  89%|████████▉ | 89/100 [03:09<00:14,  1.29s/it]
+
Eval @ step 160:  90%|█████████ | 90/100 [03:10<00:13,  1.34s/it]
+
Eval @ step 160:  91%|█████████ | 91/100 [03:11<00:11,  1.31s/it]
+
Eval @ step 160:  92%|█████████▏| 92/100 [03:13<00:10,  1.34s/it]
+
Eval @ step 160:  93%|█████████▎| 93/100 [03:14<00:09,  1.37s/it]
+
Eval @ step 160:  94%|█████████▍| 94/100 [03:16<00:08,  1.45s/it]
+
Eval @ step 160:  95%|█████████▌| 95/100 [03:18<00:08,  1.62s/it]
+
Eval @ step 160:  96%|█████████▌| 96/100 [03:19<00:05,  1.48s/it]
+
Eval @ step 160:  97%|█████████▋| 97/100 [03:20<00:04,  1.35s/it]
+
Eval @ step 160:  98%|█████████▊| 98/100 [03:23<00:03,  1.73s/it]
+
Eval @ step 160:  99%|█████████▉| 99/100 [03:24<00:01,  1.55s/it]
+
Eval @ step 160: 100%|██████████| 100/100 [03:27<00:00,  1.98s/it]
+
                                                                    [EVAL] ★ New best eval_rougeL=20.2337 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '20.23', 'epoch': '5'}
+
 33%|███▎      | 160/480 [26:00<34:56,  6.55s/it]
 34%|███▎      | 161/480 [26:09<6:09:11, 69.44s/it]
 34%|███▍      | 162/480 [26:18<4:31:18, 51.19s/it]
 34%|███▍      | 163/480 [26:26<3:22:22, 38.30s/it]
 34%|███▍      | 164/480 [26:35<2:36:00, 29.62s/it]
 34%|███▍      | 165/480 [26:44<2:02:24, 23.32s/it]
 35%|███▍      | 166/480 [26:52<1:38:17, 18.78s/it]
 35%|███▍      | 167/480 [27:01<1:22:22, 15.79s/it]
 35%|███▌      | 168/480 [27:10<1:11:05, 13.67s/it]
 35%|███▌      | 169/480 [27:19<1:03:21, 12.22s/it]
 35%|███▌      | 170/480 [27:27<57:25, 11.12s/it]  
                                                 
{'loss': '2.04', 'grad_norm': '0.6919', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [27:27<57:25, 11.12s/it]
 36%|███▌      | 171/480 [27:35<52:45, 10.24s/it]
 36%|███▌      | 172/480 [27:44<50:24,  9.82s/it]
 36%|███▌      | 173/480 [27:53<49:01,  9.58s/it]
 36%|███▋      | 174/480 [28:02<47:59,  9.41s/it]
 36%|███▋      | 175/480 [28:10<45:41,  8.99s/it]
 37%|███▋      | 176/480 [28:19<45:06,  8.90s/it]
 37%|███▋      | 177/480 [28:27<43:46,  8.67s/it]
 37%|███▋      | 178/480 [28:36<43:44,  8.69s/it]
 37%|███▋      | 179/480 [28:44<42:58,  8.57s/it]
 38%|███▊      | 180/480 [28:53<44:06,  8.82s/it]
                                                 
{'loss': '2.046', 'grad_norm': '0.4122', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [28:53<44:06,  8.82s/it]
 38%|███▊      | 181/480 [29:03<44:36,  8.95s/it]
 38%|███▊      | 182/480 [29:11<44:02,  8.87s/it]
 38%|███▊      | 183/480 [29:20<43:43,  8.83s/it]
 38%|███▊      | 184/480 [29:29<43:50,  8.89s/it]
 39%|███▊      | 185/480 [29:38<43:27,  8.84s/it]
 39%|███▉      | 186/480 [29:46<41:55,  8.56s/it]
 39%|███▉      | 187/480 [29:54<41:20,  8.47s/it]
 39%|███▉      | 188/480 [30:03<42:15,  8.68s/it]
 39%|███▉      | 189/480 [30:12<42:13,  8.71s/it]
 40%|███▉      | 190/480 [30:20<41:17,  8.54s/it]
                                                 
{'loss': '1.978', 'grad_norm': '0.3392', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [30:20<41:17,  8.54s/it]
 40%|███▉      | 191/480 [30:29<41:21,  8.59s/it]
 40%|████      | 192/480 [30:30<31:14,  6.51s/it]
 40%|████      | 193/480 [30:39<33:42,  7.05s/it]
 40%|████      | 194/480 [30:48<36:54,  7.74s/it]
 41%|████      | 195/480 [30:57<37:54,  7.98s/it]
 41%|████      | 196/480 [31:06<39:51,  8.42s/it]
 41%|████      | 197/480 [31:14<39:27,  8.37s/it]
 41%|████▏     | 198/480 [31:23<39:18,  8.36s/it]
 41%|████▏     | 199/480 [31:31<38:32,  8.23s/it]
 42%|████▏     | 200/480 [31:39<37:56,  8.13s/it]
                                                 
{'loss': '1.894', 'grad_norm': '0.4022', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [31:39<37:56,  8.13s/it]
 42%|████▏     | 201/480 [31:47<38:24,  8.26s/it]
 42%|████▏     | 202/480 [31:56<38:55,  8.40s/it]
 42%|████▏     | 203/480 [32:05<40:08,  8.70s/it]
 42%|████▎     | 204/480 [32:14<39:33,  8.60s/it]
 43%|████▎     | 205/480 [32:22<38:59,  8.51s/it]
 43%|████▎     | 206/480 [32:31<39:58,  8.75s/it]
 43%|████▎     | 207/480 [32:40<39:11,  8.61s/it]
 43%|████▎     | 208/480 [32:48<38:33,  8.50s/it]
 44%|████▎     | 209/480 [32:57<40:02,  8.86s/it]
 44%|████▍     | 210/480 [33:07<40:17,  8.96s/it]
                                                 
{'loss': '2.005', 'grad_norm': '0.6256', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [33:07<40:17,  8.96s/it]
 44%|████▍     | 211/480 [33:15<39:09,  8.73s/it]
 44%|████▍     | 212/480 [33:24<39:08,  8.76s/it]
 44%|████▍     | 213/480 [33:32<38:36,  8.68s/it]
 45%|████▍     | 214/480 [33:41<39:02,  8.81s/it]
 45%|████▍     | 215/480 [33:50<38:28,  8.71s/it]
 45%|████▌     | 216/480 [33:58<38:20,  8.72s/it]
 45%|████▌     | 217/480 [34:07<38:16,  8.73s/it]
 45%|████▌     | 218/480 [34:16<37:55,  8.69s/it]
 46%|████▌     | 219/480 [34:25<37:59,  8.73s/it]
 46%|████▌     | 220/480 [34:33<37:45,  8.71s/it]
                                                 
{'loss': '2', 'grad_norm': '0.6014', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [34:33<37:45,  8.71s/it]
 46%|████▌     | 221/480 [34:42<37:20,  8.65s/it]
 46%|████▋     | 222/480 [34:51<38:23,  8.93s/it]
 46%|████▋     | 223/480 [34:59<37:04,  8.66s/it]
 47%|████▋     | 224/480 [35:02<28:55,  6.78s/it]
 47%|████▋     | 225/480 [35:10<30:23,  7.15s/it]
 47%|████▋     | 226/480 [35:18<31:59,  7.56s/it]
 47%|████▋     | 227/480 [35:27<33:14,  7.88s/it]
 48%|████▊     | 228/480 [35:35<33:19,  7.93s/it]
 48%|████▊     | 229/480 [35:44<34:36,  8.27s/it]
 48%|████▊     | 230/480 [35:53<35:46,  8.59s/it]
                                                 
{'loss': '2.076', 'grad_norm': '0.4138', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [35:53<35:46,  8.59s/it]
 48%|████▊     | 231/480 [36:02<35:35,  8.57s/it]
 48%|████▊     | 232/480 [36:11<35:25,  8.57s/it]
 49%|████▊     | 233/480 [36:19<35:38,  8.66s/it]
 49%|████▉     | 234/480 [36:29<36:04,  8.80s/it]
 49%|████▉     | 235/480 [36:37<35:50,  8.78s/it]
 49%|████▉     | 236/480 [36:46<35:17,  8.68s/it]
 49%|████▉     | 237/480 [36:53<33:45,  8.33s/it]
 50%|████▉     | 238/480 [37:01<33:27,  8.30s/it]
 50%|████▉     | 239/480 [37:11<34:27,  8.58s/it]
 50%|█████     | 240/480 [37:19<34:06,  8.53s/it]
                                                 
{'loss': '1.949', 'grad_norm': '0.4195', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [37:19<34:06,  8.53s/it]
 50%|█████     | 241/480 [37:28<34:21,  8.62s/it]
 50%|█████     | 242/480 [37:36<33:00,  8.32s/it]
 51%|█████     | 243/480 [37:43<32:18,  8.18s/it]
 51%|█████     | 244/480 [37:52<33:09,  8.43s/it]
 51%|█████     | 245/480 [38:03<35:09,  8.98s/it]
 51%|█████▏    | 246/480 [38:11<34:21,  8.81s/it]
 51%|█████▏    | 247/480 [38:19<33:21,  8.59s/it]
 52%|█████▏    | 248/480 [38:29<34:21,  8.88s/it]
 52%|█████▏    | 249/480 [38:37<33:45,  8.77s/it]
 52%|█████▏    | 250/480 [38:45<32:43,  8.54s/it]
                                                 
{'loss': '1.937', 'grad_norm': '0.8001', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [38:45<32:43,  8.54s/it]
 52%|█████▏    | 251/480 [38:54<33:09,  8.69s/it]
 52%|█████▎    | 252/480 [39:02<32:01,  8.43s/it]
 53%|█████▎    | 253/480 [39:12<33:10,  8.77s/it]
 53%|█████▎    | 254/480 [39:22<34:31,  9.17s/it]
 53%|█████▎    | 255/480 [39:30<33:17,  8.88s/it]
 53%|█████▎    | 256/480 [39:32<25:39,  6.87s/it]
 54%|█████▎    | 257/480 [39:41<27:20,  7.36s/it]
 54%|█████▍    | 258/480 [39:50<29:07,  7.87s/it]
 54%|█████▍    | 259/480 [39:58<29:13,  7.94s/it]
 54%|█████▍    | 260/480 [40:06<29:19,  8.00s/it]
                                                 
{'loss': '1.94', 'grad_norm': '0.3867', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [40:06<29:19,  8.00s/it]
 54%|█████▍    | 261/480 [40:14<29:39,  8.13s/it]
 55%|█████▍    | 262/480 [40:22<28:50,  7.94s/it]
 55%|█████▍    | 263/480 [40:30<28:56,  8.00s/it]
 55%|█████▌    | 264/480 [40:38<28:58,  8.05s/it]
 55%|█████▌    | 265/480 [40:47<29:39,  8.28s/it]
 55%|█████▌    | 266/480 [40:56<30:24,  8.52s/it]
 56%|█████▌    | 267/480 [41:05<30:17,  8.53s/it]
 56%|█████▌    | 268/480 [41:14<30:39,  8.68s/it]
 56%|█████▌    | 269/480 [41:22<30:06,  8.56s/it]
 56%|█████▋    | 270/480 [41:30<29:39,  8.47s/it]
                                                 
{'loss': '1.922', 'grad_norm': '0.3647', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [41:30<29:39,  8.47s/it]
 56%|█████▋    | 271/480 [41:39<29:56,  8.60s/it]
 57%|█████▋    | 272/480 [41:48<29:59,  8.65s/it]
 57%|█████▋    | 273/480 [41:57<30:07,  8.73s/it]
 57%|█████▋    | 274/480 [42:06<30:21,  8.84s/it]
 57%|█████▋    | 275/480 [42:15<30:18,  8.87s/it]
 57%|█████▊    | 276/480 [42:23<29:48,  8.77s/it]
 58%|█████▊    | 277/480 [42:32<29:16,  8.65s/it]
 58%|█████▊    | 278/480 [42:42<30:16,  8.99s/it]
 58%|█████▊    | 279/480 [42:50<29:56,  8.94s/it]
 58%|█████▊    | 280/480 [43:00<30:07,  9.04s/it]
                                                 
{'loss': '1.937', 'grad_norm': '0.4517', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [43:00<30:07,  9.04s/it]
 59%|█████▊    | 281/480 [43:08<29:23,  8.86s/it]
 59%|█████▉    | 282/480 [43:17<28:58,  8.78s/it]
 59%|█████▉    | 283/480 [43:26<28:53,  8.80s/it]
 59%|█████▉    | 284/480 [43:34<28:36,  8.76s/it]
 59%|█████▉    | 285/480 [43:43<28:14,  8.69s/it]
 60%|█████▉    | 286/480 [43:52<28:13,  8.73s/it]
 60%|█████▉    | 287/480 [44:01<28:34,  8.88s/it]
 60%|██████    | 288/480 [44:03<22:04,  6.90s/it]
 60%|██████    | 289/480 [44:13<24:27,  7.69s/it]
 60%|██████    | 290/480 [44:21<25:28,  8.05s/it]
                                                 
{'loss': '2.009', 'grad_norm': '0.406', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [44:21<25:28,  8.05s/it]
 61%|██████    | 291/480 [44:30<25:55,  8.23s/it]
 61%|██████    | 292/480 [44:39<25:57,  8.28s/it]
 61%|██████    | 293/480 [44:47<25:51,  8.30s/it]
 61%|██████▏   | 294/480 [44:55<25:29,  8.22s/it]
 61%|██████▏   | 295/480 [45:04<25:43,  8.35s/it]
 62%|██████▏   | 296/480 [45:12<25:49,  8.42s/it]
 62%|██████▏   | 297/480 [45:21<25:53,  8.49s/it]
 62%|██████▏   | 298/480 [45:30<26:17,  8.67s/it]
 62%|██████▏   | 299/480 [45:39<26:14,  8.70s/it]
 62%|██████▎   | 300/480 [45:47<25:54,  8.63s/it]
                                                 
{'loss': '1.891', 'grad_norm': '0.3832', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [45:47<25:54,  8.63s/it]
 63%|██████▎   | 301/480 [45:56<25:57,  8.70s/it]
 63%|██████▎   | 302/480 [46:05<26:26,  8.91s/it]
 63%|██████▎   | 303/480 [46:14<26:27,  8.97s/it]
 63%|██████▎   | 304/480 [46:23<25:52,  8.82s/it]
 64%|██████▎   | 305/480 [46:32<25:34,  8.77s/it]
 64%|██████▍   | 306/480 [46:41<25:47,  8.89s/it]
 64%|██████▍   | 307/480 [46:49<25:12,  8.74s/it]
 64%|██████▍   | 308/480 [46:57<24:08,  8.42s/it]
 64%|██████▍   | 309/480 [47:05<23:43,  8.33s/it]
 65%|██████▍   | 310/480 [47:13<23:37,  8.34s/it]
                                                 
{'loss': '1.918', 'grad_norm': '0.4119', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [47:13<23:37,  8.34s/it]
 65%|██████▍   | 311/480 [47:22<24:08,  8.57s/it]
 65%|██████▌   | 312/480 [47:31<23:54,  8.54s/it]
 65%|██████▌   | 313/480 [47:40<24:20,  8.75s/it]
 65%|██████▌   | 314/480 [47:49<24:17,  8.78s/it]
 66%|██████▌   | 315/480 [47:58<24:26,  8.89s/it]
 66%|██████▌   | 316/480 [48:06<23:39,  8.66s/it]
 66%|██████▌   | 317/480 [48:15<23:24,  8.62s/it]
 66%|██████▋   | 318/480 [48:23<22:56,  8.50s/it]
 66%|██████▋   | 319/480 [48:31<22:32,  8.40s/it]
 67%|██████▋   | 320/480 [48:34<17:34,  6.59s/it]
                                                 
{'loss': '2.069', 'grad_norm': '2.574', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [48:34<17:34,  6.59s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:01<01:59,  1.21s/it]
+
Eval @ step 320:   2%|▏         | 2/100 [00:05<04:48,  2.95s/it]
+
Eval @ step 320:   3%|▎         | 3/100 [00:06<03:35,  2.22s/it]
+
Eval @ step 320:   4%|▍         | 4/100 [00:07<02:38,  1.65s/it]
+
Eval @ step 320:   5%|▌         | 5/100 [00:08<02:08,  1.35s/it]
+
Eval @ step 320:   6%|▌         | 6/100 [00:08<01:43,  1.11s/it]
+
Eval @ step 320:   7%|▋         | 7/100 [00:11<02:19,  1.49s/it]
+
Eval @ step 320:   8%|▊         | 8/100 [00:13<02:36,  1.70s/it]
+
Eval @ step 320:   9%|▉         | 9/100 [00:14<02:21,  1.55s/it]
+
Eval @ step 320:  10%|█         | 10/100 [00:15<01:59,  1.32s/it]
+
Eval @ step 320:  11%|█         | 11/100 [00:16<01:48,  1.22s/it]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:17<01:37,  1.11s/it]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:18<01:44,  1.21s/it]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:19<01:37,  1.14s/it]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:22<02:13,  1.57s/it]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:26<03:28,  2.48s/it]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:28<02:56,  2.12s/it]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:29<02:35,  1.90s/it]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:31<02:26,  1.81s/it]
+
Eval @ step 320:  20%|██        | 20/100 [00:31<02:02,  1.53s/it]
+
Eval @ step 320:  21%|██        | 21/100 [00:32<01:40,  1.27s/it]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:33<01:39,  1.28s/it]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:34<01:31,  1.19s/it]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:36<01:28,  1.17s/it]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:37<01:38,  1.31s/it]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:39<01:52,  1.52s/it]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:46<03:37,  2.99s/it]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:47<02:52,  2.40s/it]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:49<02:50,  2.40s/it]
+
Eval @ step 320:  30%|███       | 30/100 [00:50<02:11,  1.88s/it]
+
Eval @ step 320:  31%|███       | 31/100 [00:52<02:23,  2.08s/it]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:55<02:42,  2.38s/it]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:57<02:15,  2.02s/it]
+
Eval @ step 320:  34%|███▍      | 34/100 [01:04<03:54,  3.56s/it]
+
Eval @ step 320:  35%|███▌      | 35/100 [01:05<02:59,  2.76s/it]
+
Eval @ step 320:  36%|███▌      | 36/100 [01:05<02:14,  2.11s/it]
+
Eval @ step 320:  37%|███▋      | 37/100 [01:06<01:55,  1.83s/it]
+
Eval @ step 320:  38%|███▊      | 38/100 [01:07<01:35,  1.55s/it]
+
Eval @ step 320:  39%|███▉      | 39/100 [01:09<01:33,  1.53s/it]
+
Eval @ step 320:  40%|████      | 40/100 [01:12<01:57,  1.96s/it]
+
Eval @ step 320:  41%|████      | 41/100 [01:13<01:43,  1.76s/it]
+
Eval @ step 320:  42%|████▏     | 42/100 [01:17<02:23,  2.48s/it]
+
Eval @ step 320:  43%|████▎     | 43/100 [01:20<02:26,  2.56s/it]
+
Eval @ step 320:  44%|████▍     | 44/100 [01:21<01:55,  2.07s/it]
+
Eval @ step 320:  45%|████▌     | 45/100 [01:22<01:46,  1.94s/it]
+
Eval @ step 320:  46%|████▌     | 46/100 [01:24<01:31,  1.70s/it]
+
Eval @ step 320:  47%|████▋     | 47/100 [01:27<01:49,  2.06s/it]
+
Eval @ step 320:  48%|████▊     | 48/100 [01:29<01:49,  2.11s/it]
+
Eval @ step 320:  49%|████▉     | 49/100 [01:30<01:28,  1.75s/it]
+
Eval @ step 320:  50%|█████     | 50/100 [01:30<01:07,  1.35s/it]
+
Eval @ step 320:  51%|█████     | 51/100 [01:31<01:00,  1.24s/it]
+
Eval @ step 320:  52%|█████▏    | 52/100 [01:32<00:51,  1.08s/it]
+
Eval @ step 320:  53%|█████▎    | 53/100 [01:34<01:01,  1.31s/it]
+
Eval @ step 320:  54%|█████▍    | 54/100 [01:35<00:58,  1.27s/it]
+
Eval @ step 320:  55%|█████▌    | 55/100 [01:36<00:50,  1.13s/it]
+
Eval @ step 320:  56%|█████▌    | 56/100 [01:41<01:51,  2.54s/it]
+
Eval @ step 320:  57%|█████▋    | 57/100 [01:49<02:48,  3.92s/it]
+
Eval @ step 320:  58%|█████▊    | 58/100 [01:50<02:11,  3.14s/it]
+
Eval @ step 320:  59%|█████▉    | 59/100 [01:51<01:45,  2.58s/it]
+
Eval @ step 320:  60%|██████    | 60/100 [01:52<01:24,  2.12s/it]
+
Eval @ step 320:  61%|██████    | 61/100 [01:53<01:11,  1.83s/it]
+
Eval @ step 320:  62%|██████▏   | 62/100 [01:56<01:17,  2.04s/it]
+
Eval @ step 320:  63%|██████▎   | 63/100 [01:58<01:19,  2.15s/it]
+
Eval @ step 320:  64%|██████▍   | 64/100 [02:05<02:11,  3.65s/it]
+
Eval @ step 320:  65%|██████▌   | 65/100 [02:06<01:33,  2.66s/it]
+
Eval @ step 320:  66%|██████▌   | 66/100 [02:07<01:15,  2.22s/it]
+
Eval @ step 320:  67%|██████▋   | 67/100 [02:08<01:00,  1.83s/it]
+
Eval @ step 320:  68%|██████▊   | 68/100 [02:09<00:53,  1.66s/it]
+
Eval @ step 320:  69%|██████▉   | 69/100 [02:10<00:44,  1.45s/it]
+
Eval @ step 320:  70%|███████   | 70/100 [02:12<00:43,  1.44s/it]
+
Eval @ step 320:  71%|███████   | 71/100 [02:13<00:45,  1.58s/it]
+
Eval @ step 320:  72%|███████▏  | 72/100 [02:16<00:48,  1.73s/it]
+
Eval @ step 320:  73%|███████▎  | 73/100 [02:16<00:38,  1.43s/it]
+
Eval @ step 320:  74%|███████▍  | 74/100 [02:17<00:33,  1.28s/it]
+
Eval @ step 320:  75%|███████▌  | 75/100 [02:20<00:45,  1.81s/it]
+
Eval @ step 320:  76%|███████▌  | 76/100 [02:22<00:39,  1.66s/it]
+
Eval @ step 320:  77%|███████▋  | 77/100 [02:22<00:32,  1.42s/it]
+
Eval @ step 320:  78%|███████▊  | 78/100 [02:24<00:29,  1.33s/it]
+
Eval @ step 320:  79%|███████▉  | 79/100 [02:24<00:24,  1.16s/it]
+
Eval @ step 320:  80%|████████  | 80/100 [02:26<00:24,  1.22s/it]
+
Eval @ step 320:  81%|████████  | 81/100 [02:26<00:16,  1.12it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [02:28<00:21,  1.21s/it]
+
Eval @ step 320:  83%|████████▎ | 83/100 [02:29<00:19,  1.14s/it]
+
Eval @ step 320:  84%|████████▍ | 84/100 [02:31<00:22,  1.41s/it]
+
Eval @ step 320:  85%|████████▌ | 85/100 [02:32<00:20,  1.38s/it]
+
Eval @ step 320:  86%|████████▌ | 86/100 [02:33<00:17,  1.24s/it]
+
Eval @ step 320:  87%|████████▋ | 87/100 [02:35<00:19,  1.46s/it]
+
Eval @ step 320:  88%|████████▊ | 88/100 [02:38<00:23,  1.93s/it]
+
Eval @ step 320:  89%|████████▉ | 89/100 [02:39<00:19,  1.74s/it]
+
Eval @ step 320:  90%|█████████ | 90/100 [02:41<00:17,  1.74s/it]
+
Eval @ step 320:  91%|█████████ | 91/100 [02:42<00:14,  1.58s/it]
+
Eval @ step 320:  92%|█████████▏| 92/100 [02:43<00:11,  1.48s/it]
+
Eval @ step 320:  93%|█████████▎| 93/100 [02:45<00:10,  1.53s/it]
+
Eval @ step 320:  94%|█████████▍| 94/100 [02:47<00:09,  1.57s/it]
+
Eval @ step 320:  95%|█████████▌| 95/100 [02:48<00:07,  1.54s/it]
+
Eval @ step 320:  96%|█████████▌| 96/100 [02:50<00:06,  1.57s/it]
+
Eval @ step 320:  97%|█████████▋| 97/100 [02:51<00:04,  1.41s/it]
+
Eval @ step 320:  98%|█████████▊| 98/100 [02:53<00:03,  1.77s/it]
+
Eval @ step 320:  99%|█████████▉| 99/100 [02:55<00:01,  1.56s/it]
+
Eval @ step 320: 100%|██████████| 100/100 [02:57<00:00,  1.91s/it]
+
                                                                    [EVAL] ★ New best eval_rougeL=21.3091 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '21.31', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [51:31<17:34,  6.59s/it]
 67%|██████▋   | 321/480 [51:41<2:41:00, 60.76s/it]
 67%|██████▋   | 322/480 [51:50<1:59:08, 45.25s/it]
 67%|██████▋   | 323/480 [52:00<1:30:45, 34.69s/it]
 68%|██████▊   | 324/480 [52:08<1:09:50, 26.86s/it]
 68%|██████▊   | 325/480 [52:17<55:13, 21.38s/it]  
 68%|██████▊   | 326/480 [52:25<44:45, 17.44s/it]
 68%|██████▊   | 327/480 [52:32<36:26, 14.29s/it]
 68%|██████▊   | 328/480 [52:40<31:15, 12.34s/it]
 69%|██████▊   | 329/480 [52:48<27:55, 11.10s/it]
 69%|██████▉   | 330/480 [52:57<25:45, 10.30s/it]
                                                 
{'loss': '1.917', 'grad_norm': '0.4604', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [52:57<25:45, 10.30s/it]
 69%|██████▉   | 331/480 [53:04<23:35,  9.50s/it]
 69%|██████▉   | 332/480 [53:14<23:18,  9.45s/it]
 69%|██████▉   | 333/480 [53:23<22:47,  9.30s/it]
 70%|██████▉   | 334/480 [53:31<21:56,  9.02s/it]
 70%|██████▉   | 335/480 [53:40<22:03,  9.13s/it]
 70%|███████   | 336/480 [53:49<21:56,  9.14s/it]
 70%|███████   | 337/480 [53:58<21:02,  8.83s/it]
 70%|███████   | 338/480 [54:06<20:46,  8.78s/it]
 71%|███████   | 339/480 [54:15<20:31,  8.73s/it]
 71%|███████   | 340/480 [54:24<20:33,  8.81s/it]
                                                 
{'loss': '1.938', 'grad_norm': '0.5513', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [54:24<20:33,  8.81s/it]
 71%|███████   | 341/480 [54:33<20:28,  8.84s/it]
 71%|███████▏  | 342/480 [54:41<20:11,  8.78s/it]
 71%|███████▏  | 343/480 [54:50<20:08,  8.82s/it]
 72%|███████▏  | 344/480 [54:59<19:48,  8.74s/it]
 72%|███████▏  | 345/480 [55:08<19:48,  8.80s/it]
 72%|███████▏  | 346/480 [55:17<19:52,  8.90s/it]
 72%|███████▏  | 347/480 [55:26<19:34,  8.83s/it]
 72%|███████▎  | 348/480 [55:34<18:55,  8.60s/it]
 73%|███████▎  | 349/480 [55:43<19:16,  8.83s/it]
 73%|███████▎  | 350/480 [55:51<18:40,  8.62s/it]
                                                 
{'loss': '1.939', 'grad_norm': '0.344', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [55:51<18:40,  8.62s/it]
 73%|███████▎  | 351/480 [56:00<19:00,  8.84s/it]
 73%|███████▎  | 352/480 [56:02<14:18,  6.71s/it]
 74%|███████▎  | 353/480 [56:11<15:42,  7.42s/it]
 74%|███████▍  | 354/480 [56:20<16:08,  7.69s/it]
 74%|███████▍  | 355/480 [56:28<16:38,  7.99s/it]
 74%|███████▍  | 356/480 [56:37<16:49,  8.14s/it]
 74%|███████▍  | 357/480 [56:46<17:23,  8.49s/it]
 75%|███████▍  | 358/480 [56:55<17:31,  8.62s/it]
 75%|███████▍  | 359/480 [57:03<16:50,  8.35s/it]
 75%|███████▌  | 360/480 [57:11<16:44,  8.37s/it]
                                                 
{'loss': '1.894', 'grad_norm': '0.535', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [57:11<16:44,  8.37s/it]
 75%|███████▌  | 361/480 [57:19<16:11,  8.16s/it]
 75%|███████▌  | 362/480 [57:28<16:27,  8.37s/it]
 76%|███████▌  | 363/480 [57:37<16:37,  8.53s/it]
 76%|███████▌  | 364/480 [57:45<16:36,  8.59s/it]
 76%|███████▌  | 365/480 [57:53<15:55,  8.31s/it]
 76%|███████▋  | 366/480 [58:02<15:58,  8.41s/it]
 76%|███████▋  | 367/480 [58:11<16:09,  8.58s/it]
 77%|███████▋  | 368/480 [58:19<16:01,  8.58s/it]
 77%|███████▋  | 369/480 [58:27<15:38,  8.45s/it]
 77%|███████▋  | 370/480 [58:35<15:16,  8.33s/it]
                                                 
{'loss': '1.914', 'grad_norm': '0.7717', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [58:35<15:16,  8.33s/it]
 77%|███████▋  | 371/480 [58:45<15:38,  8.61s/it]
 78%|███████▊  | 372/480 [58:54<16:03,  8.92s/it]
 78%|███████▊  | 373/480 [59:02<15:27,  8.67s/it]
 78%|███████▊  | 374/480 [59:12<15:49,  8.96s/it]
 78%|███████▊  | 375/480 [59:20<15:24,  8.80s/it]
 78%|███████▊  | 376/480 [59:29<15:03,  8.69s/it]
 79%|███████▊  | 377/480 [59:38<15:22,  8.96s/it]
 79%|███████▉  | 378/480 [59:48<15:26,  9.08s/it]
 79%|███████▉  | 379/480 [59:57<15:13,  9.04s/it]
 79%|███████▉  | 380/480 [1:00:06<15:09,  9.10s/it]
                                                   
{'loss': '1.909', 'grad_norm': '0.3086', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [1:00:06<15:09,  9.10s/it]
 79%|███████▉  | 381/480 [1:00:15<15:04,  9.14s/it]
 80%|███████▉  | 382/480 [1:00:23<14:29,  8.88s/it]
 80%|███████▉  | 383/480 [1:00:32<13:55,  8.62s/it]
 80%|████████  | 384/480 [1:00:33<10:29,  6.56s/it]
 80%|████████  | 385/480 [1:00:42<11:15,  7.11s/it]
 80%|████████  | 386/480 [1:00:51<12:01,  7.68s/it]
 81%|████████  | 387/480 [1:00:58<11:58,  7.72s/it]
 81%|████████  | 388/480 [1:01:07<12:14,  7.99s/it]
 81%|████████  | 389/480 [1:01:16<12:27,  8.21s/it]
 81%|████████▏ | 390/480 [1:01:25<12:37,  8.42s/it]
                                                   
{'loss': '1.886', 'grad_norm': '0.4165', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [1:01:25<12:37,  8.42s/it]
 81%|████████▏ | 391/480 [1:01:34<12:49,  8.65s/it]
 82%|████████▏ | 392/480 [1:01:42<12:37,  8.61s/it]
 82%|████████▏ | 393/480 [1:01:51<12:36,  8.70s/it]
 82%|████████▏ | 394/480 [1:01:59<12:12,  8.52s/it]
 82%|████████▏ | 395/480 [1:02:08<12:10,  8.59s/it]
 82%|████████▎ | 396/480 [1:02:16<11:42,  8.36s/it]
 83%|████████▎ | 397/480 [1:02:25<11:54,  8.61s/it]
 83%|████████▎ | 398/480 [1:02:33<11:22,  8.32s/it]
 83%|████████▎ | 399/480 [1:02:41<11:15,  8.34s/it]
 83%|████████▎ | 400/480 [1:02:50<11:26,  8.58s/it]
                                                   
{'loss': '1.891', 'grad_norm': '0.5121', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [1:02:50<11:26,  8.58s/it]
 84%|████████▎ | 401/480 [1:02:59<11:13,  8.52s/it]
 84%|████████▍ | 402/480 [1:03:06<10:41,  8.23s/it]
 84%|████████▍ | 403/480 [1:03:15<10:48,  8.42s/it]
 84%|████████▍ | 404/480 [1:03:24<10:53,  8.60s/it]
 84%|████████▍ | 405/480 [1:03:34<11:08,  8.91s/it]
 85%|████████▍ | 406/480 [1:03:43<10:58,  8.91s/it]
 85%|████████▍ | 407/480 [1:03:51<10:31,  8.64s/it]
 85%|████████▌ | 408/480 [1:04:00<10:35,  8.82s/it]
 85%|████████▌ | 409/480 [1:04:09<10:22,  8.77s/it]
 85%|████████▌ | 410/480 [1:04:17<09:56,  8.51s/it]
                                                   
{'loss': '1.898', 'grad_norm': '0.4284', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [1:04:17<09:56,  8.51s/it]
 86%|████████▌ | 411/480 [1:04:25<09:56,  8.64s/it]
 86%|████████▌ | 412/480 [1:04:35<09:55,  8.75s/it]
 86%|████████▌ | 413/480 [1:04:43<09:46,  8.75s/it]
 86%|████████▋ | 414/480 [1:04:52<09:39,  8.78s/it]
 86%|████████▋ | 415/480 [1:05:01<09:24,  8.68s/it]
 87%|████████▋ | 416/480 [1:05:03<07:20,  6.89s/it]
 87%|████████▋ | 417/480 [1:05:12<07:47,  7.43s/it]
 87%|████████▋ | 418/480 [1:05:21<08:03,  7.80s/it]
 87%|████████▋ | 419/480 [1:05:28<07:55,  7.79s/it]
 88%|████████▊ | 420/480 [1:05:37<08:09,  8.15s/it]
                                                   
{'loss': '1.86', 'grad_norm': '0.5474', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [1:05:37<08:09,  8.15s/it]
 88%|████████▊ | 421/480 [1:05:45<07:59,  8.12s/it]
 88%|████████▊ | 422/480 [1:05:54<08:01,  8.30s/it]
 88%|████████▊ | 423/480 [1:06:02<07:49,  8.24s/it]
 88%|████████▊ | 424/480 [1:06:11<07:48,  8.36s/it]
 89%|████████▊ | 425/480 [1:06:20<07:46,  8.48s/it]
 89%|████████▉ | 426/480 [1:06:28<07:41,  8.55s/it]
 89%|████████▉ | 427/480 [1:06:37<07:36,  8.62s/it]
 89%|████████▉ | 428/480 [1:06:46<07:34,  8.75s/it]
 89%|████████▉ | 429/480 [1:06:54<07:17,  8.58s/it]
 90%|████████▉ | 430/480 [1:07:03<07:15,  8.70s/it]
                                                   
{'loss': '1.874', 'grad_norm': '0.4409', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [1:07:03<07:15,  8.70s/it]
 90%|████████▉ | 431/480 [1:07:12<07:08,  8.74s/it]
 90%|█████████ | 432/480 [1:07:21<07:00,  8.75s/it]
 90%|█████████ | 433/480 [1:07:30<06:58,  8.90s/it]
 90%|█████████ | 434/480 [1:07:38<06:36,  8.61s/it]
 91%|█████████ | 435/480 [1:07:47<06:33,  8.76s/it]
 91%|█████████ | 436/480 [1:07:56<06:31,  8.90s/it]
 91%|█████████ | 437/480 [1:08:06<06:32,  9.12s/it]
 91%|█████████▏| 438/480 [1:08:15<06:25,  9.18s/it]
 91%|█████████▏| 439/480 [1:08:24<06:07,  8.96s/it]
 92%|█████████▏| 440/480 [1:08:32<05:46,  8.67s/it]
                                                   
{'loss': '1.939', 'grad_norm': '0.3765', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [1:08:32<05:46,  8.67s/it]
 92%|█████████▏| 441/480 [1:08:41<05:40,  8.72s/it]
 92%|█████████▏| 442/480 [1:08:49<05:22,  8.50s/it]
 92%|█████████▏| 443/480 [1:08:58<05:20,  8.65s/it]
 92%|█████████▎| 444/480 [1:09:08<05:32,  9.25s/it]
 93%|█████████▎| 445/480 [1:09:17<05:14,  8.99s/it]
 93%|█████████▎| 446/480 [1:09:25<05:01,  8.88s/it]
 93%|█████████▎| 447/480 [1:09:34<04:47,  8.71s/it]
 93%|█████████▎| 448/480 [1:09:36<03:33,  6.66s/it]
 94%|█████████▎| 449/480 [1:09:45<03:48,  7.37s/it]
 94%|█████████▍| 450/480 [1:09:54<03:58,  7.95s/it]
                                                   
{'loss': '1.903', 'grad_norm': '0.481', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [1:09:54<03:58,  7.95s/it]
 94%|█████████▍| 451/480 [1:10:03<03:56,  8.17s/it]
 94%|█████████▍| 452/480 [1:10:12<03:56,  8.46s/it]
 94%|█████████▍| 453/480 [1:10:20<03:50,  8.56s/it]
 95%|█████████▍| 454/480 [1:10:30<03:46,  8.73s/it]
 95%|█████████▍| 455/480 [1:10:39<03:40,  8.83s/it]
 95%|█████████▌| 456/480 [1:10:47<03:31,  8.81s/it]
 95%|█████████▌| 457/480 [1:10:56<03:23,  8.86s/it]
 95%|█████████▌| 458/480 [1:11:05<03:15,  8.89s/it]
 96%|█████████▌| 459/480 [1:11:15<03:08,  8.99s/it]
 96%|█████████▌| 460/480 [1:11:23<02:54,  8.75s/it]
                                                   
{'loss': '1.89', 'grad_norm': '0.5342', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [1:11:23<02:54,  8.75s/it]
 96%|█████████▌| 461/480 [1:11:31<02:42,  8.56s/it]
 96%|█████████▋| 462/480 [1:11:39<02:32,  8.50s/it]
 96%|█████████▋| 463/480 [1:11:48<02:26,  8.60s/it]
 97%|█████████▋| 464/480 [1:11:58<02:21,  8.85s/it]
 97%|█████████▋| 465/480 [1:12:06<02:11,  8.78s/it]
 97%|█████████▋| 466/480 [1:12:14<02:00,  8.58s/it]
 97%|█████████▋| 467/480 [1:12:22<01:47,  8.28s/it]
 98%|█████████▊| 468/480 [1:12:30<01:40,  8.39s/it]
 98%|█████████▊| 469/480 [1:12:39<01:33,  8.49s/it]
 98%|█████████▊| 470/480 [1:12:47<01:22,  8.21s/it]
                                                   
{'loss': '1.825', 'grad_norm': '0.3859', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [1:12:47<01:22,  8.21s/it]
 98%|█████████▊| 471/480 [1:12:55<01:13,  8.16s/it]
 98%|█████████▊| 472/480 [1:13:04<01:07,  8.43s/it]
 99%|█████████▊| 473/480 [1:13:12<00:58,  8.29s/it]
 99%|█████████▉| 474/480 [1:13:21<00:51,  8.55s/it]
 99%|█████████▉| 475/480 [1:13:30<00:43,  8.62s/it]
 99%|█████████▉| 476/480 [1:13:38<00:33,  8.42s/it]
 99%|█████████▉| 477/480 [1:13:47<00:26,  8.75s/it]
100%|█████████▉| 478/480 [1:13:55<00:16,  8.50s/it]
100%|█████████▉| 479/480 [1:14:04<00:08,  8.64s/it]
100%|██████████| 480/480 [1:14:07<00:00,  6.79s/it]
                                                   
{'loss': '1.818', 'grad_norm': '2.08', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [1:14:07<00:00,  6.79s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:01<03:04,  1.86s/it]
+
Eval @ step 480:   2%|▏         | 2/100 [00:05<04:57,  3.04s/it]
+
Eval @ step 480:   3%|▎         | 3/100 [00:06<03:10,  1.97s/it]
+
Eval @ step 480:   4%|▍         | 4/100 [00:06<02:15,  1.41s/it]
+
Eval @ step 480:   5%|▌         | 5/100 [00:07<01:44,  1.10s/it]
+
Eval @ step 480:   6%|▌         | 6/100 [00:08<01:28,  1.07it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:13<03:28,  2.24s/it]
+
Eval @ step 480:   8%|▊         | 8/100 [00:13<02:42,  1.76s/it]
+
Eval @ step 480:   9%|▉         | 9/100 [00:15<02:25,  1.59s/it]
+
Eval @ step 480:  10%|█         | 10/100 [00:16<02:11,  1.46s/it]
+
Eval @ step 480:  11%|█         | 11/100 [00:16<01:41,  1.14s/it]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:17<01:26,  1.02it/s]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:18<01:28,  1.01s/it]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:19<01:26,  1.00s/it]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:23<02:53,  2.04s/it]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:29<04:15,  3.04s/it]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:30<03:28,  2.52s/it]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:31<02:58,  2.17s/it]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:32<02:19,  1.72s/it]
+
Eval @ step 480:  20%|██        | 20/100 [00:32<01:48,  1.35s/it]
+
Eval @ step 480:  21%|██        | 21/100 [00:33<01:31,  1.15s/it]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:35<01:39,  1.28s/it]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:36<01:38,  1.29s/it]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:37<01:33,  1.23s/it]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:39<01:47,  1.44s/it]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:41<01:59,  1.61s/it]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:45<02:48,  2.30s/it]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:46<02:23,  2.00s/it]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:51<03:10,  2.69s/it]
+
Eval @ step 480:  30%|███       | 30/100 [00:51<02:28,  2.12s/it]
+
Eval @ step 480:  31%|███       | 31/100 [00:54<02:35,  2.25s/it]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:55<02:14,  1.98s/it]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:56<01:56,  1.74s/it]
+
Eval @ step 480:  34%|███▍      | 34/100 [01:04<03:41,  3.36s/it]
+
Eval @ step 480:  35%|███▌      | 35/100 [01:06<03:24,  3.15s/it]
+
Eval @ step 480:  36%|███▌      | 36/100 [01:07<02:31,  2.37s/it]
+
Eval @ step 480:  37%|███▋      | 37/100 [01:08<02:08,  2.04s/it]
+
Eval @ step 480:  38%|███▊      | 38/100 [01:09<01:37,  1.58s/it]
+
Eval @ step 480:  39%|███▉      | 39/100 [01:10<01:36,  1.58s/it]
+
Eval @ step 480:  40%|████      | 40/100 [01:14<02:15,  2.26s/it]
+
Eval @ step 480:  41%|████      | 41/100 [01:16<02:03,  2.09s/it]
+
Eval @ step 480:  42%|████▏     | 42/100 [01:17<01:49,  1.89s/it]
+
Eval @ step 480:  43%|████▎     | 43/100 [01:20<02:02,  2.15s/it]
+
Eval @ step 480:  44%|████▍     | 44/100 [01:21<01:41,  1.81s/it]
+
Eval @ step 480:  45%|████▌     | 45/100 [01:23<01:51,  2.02s/it]
+
Eval @ step 480:  46%|████▌     | 46/100 [01:25<01:37,  1.81s/it]
+
Eval @ step 480:  47%|████▋     | 47/100 [01:27<01:44,  1.96s/it]
+
Eval @ step 480:  48%|████▊     | 48/100 [01:29<01:37,  1.88s/it]
+
Eval @ step 480:  49%|████▉     | 49/100 [01:32<01:52,  2.20s/it]
+
Eval @ step 480:  50%|█████     | 50/100 [01:32<01:23,  1.67s/it]
+
Eval @ step 480:  51%|█████     | 51/100 [01:32<01:01,  1.25s/it]
+
Eval @ step 480:  52%|█████▏    | 52/100 [01:33<00:49,  1.02s/it]
+
Eval @ step 480:  53%|█████▎    | 53/100 [01:35<00:59,  1.27s/it]
+
Eval @ step 480:  54%|█████▍    | 54/100 [01:36<00:58,  1.26s/it]
+
Eval @ step 480:  55%|█████▌    | 55/100 [01:37<00:50,  1.12s/it]
+
Eval @ step 480:  56%|█████▌    | 56/100 [01:39<01:06,  1.51s/it]
+
Eval @ step 480:  57%|█████▋    | 57/100 [01:43<01:37,  2.26s/it]
+
Eval @ step 480:  58%|█████▊    | 58/100 [01:45<01:32,  2.21s/it]
+
Eval @ step 480:  59%|█████▉    | 59/100 [01:46<01:12,  1.77s/it]
+
Eval @ step 480:  60%|██████    | 60/100 [01:47<01:02,  1.55s/it]
+
Eval @ step 480:  61%|██████    | 61/100 [01:48<00:57,  1.48s/it]
+
Eval @ step 480:  62%|██████▏   | 62/100 [01:51<01:05,  1.73s/it]
+
Eval @ step 480:  63%|██████▎   | 63/100 [01:53<01:11,  1.93s/it]
+
Eval @ step 480:  64%|██████▍   | 64/100 [01:58<01:44,  2.89s/it]
+
Eval @ step 480:  65%|██████▌   | 65/100 [01:59<01:14,  2.13s/it]
+
Eval @ step 480:  66%|██████▌   | 66/100 [02:00<01:00,  1.79s/it]
+
Eval @ step 480:  67%|██████▋   | 67/100 [02:01<00:52,  1.58s/it]
+
Eval @ step 480:  68%|██████▊   | 68/100 [02:02<00:45,  1.42s/it]
+
Eval @ step 480:  69%|██████▉   | 69/100 [02:02<00:34,  1.10s/it]
+
Eval @ step 480:  70%|███████   | 70/100 [02:03<00:33,  1.13s/it]
+
Eval @ step 480:  71%|███████   | 71/100 [02:05<00:38,  1.33s/it]
+
Eval @ step 480:  72%|███████▏  | 72/100 [02:09<00:55,  1.98s/it]
+
Eval @ step 480:  73%|███████▎  | 73/100 [02:09<00:43,  1.62s/it]
+
Eval @ step 480:  74%|███████▍  | 74/100 [02:10<00:37,  1.42s/it]
+
Eval @ step 480:  75%|███████▌  | 75/100 [02:13<00:47,  1.90s/it]
+
Eval @ step 480:  76%|███████▌  | 76/100 [02:14<00:35,  1.48s/it]
+
Eval @ step 480:  77%|███████▋  | 77/100 [02:15<00:29,  1.30s/it]
+
Eval @ step 480:  78%|███████▊  | 78/100 [02:16<00:27,  1.24s/it]
+
Eval @ step 480:  79%|███████▉  | 79/100 [02:17<00:23,  1.10s/it]
+
Eval @ step 480:  80%|████████  | 80/100 [02:18<00:23,  1.15s/it]
+
Eval @ step 480:  81%|████████  | 81/100 [02:18<00:16,  1.18it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [02:19<00:17,  1.04it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [02:21<00:22,  1.31s/it]
+
Eval @ step 480:  84%|████████▍ | 84/100 [02:24<00:26,  1.66s/it]
+
Eval @ step 480:  85%|████████▌ | 85/100 [02:25<00:21,  1.41s/it]
+
Eval @ step 480:  86%|████████▌ | 86/100 [02:26<00:17,  1.26s/it]
+
Eval @ step 480:  87%|████████▋ | 87/100 [02:26<00:14,  1.13s/it]
+
Eval @ step 480:  88%|████████▊ | 88/100 [02:32<00:29,  2.42s/it]
+
Eval @ step 480:  89%|████████▉ | 89/100 [02:34<00:25,  2.29s/it]
+
Eval @ step 480:  90%|█████████ | 90/100 [02:36<00:22,  2.24s/it]
+
Eval @ step 480:  91%|█████████ | 91/100 [02:37<00:16,  1.80s/it]
+
Eval @ step 480:  92%|█████████▏| 92/100 [02:38<00:12,  1.62s/it]
+
Eval @ step 480:  93%|█████████▎| 93/100 [02:39<00:11,  1.63s/it]
+
Eval @ step 480:  94%|█████████▍| 94/100 [02:41<00:09,  1.65s/it]
+
Eval @ step 480:  95%|█████████▌| 95/100 [02:43<00:07,  1.57s/it]
+
Eval @ step 480:  96%|█████████▌| 96/100 [02:44<00:05,  1.42s/it]
+
Eval @ step 480:  97%|█████████▋| 97/100 [02:44<00:03,  1.21s/it]
+
Eval @ step 480:  98%|█████████▊| 98/100 [02:46<00:02,  1.42s/it]
+
Eval @ step 480:  99%|█████████▉| 99/100 [02:47<00:01,  1.31s/it]
+
Eval @ step 480: 100%|██████████| 100/100 [02:50<00:00,  1.73s/it]
+
                                                                    [EVAL] eval_rougeL=19.1817 (best=21.3091)
+
                                                   
{'eval_rougeL': '19.18', 'epoch': '15'}
+
100%|██████████| 480/480 [1:16:57<00:00,  6.79s/it]
                                                   
{'train_runtime': '4618', 'train_samples_per_second': '3.248', 'train_steps_per_second': '0.104', 'train_loss': '2.01', 'epoch': '15'}
+
100%|██████████| 480/480 [1:16:57<00:00,  6.79s/it]
100%|██████████| 480/480 [1:16:57<00:00,  9.62s/it]
+***** train metrics *****
+  epoch                    =        15.0
+  total_flos               = 323861359GF
+  train_loss               =      2.0104
+  train_runtime            =  1:16:57.65
+  train_samples            =        1000
+  train_samples_per_second =       3.248
+  train_steps_per_second   =       0.104
+  [EVAL] Restored best model (eval_rougeL=21.3091, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   5%|▍         | 14/291 [00:00<00:02, 132.85it/s]
Loading weights:  13%|█▎        | 38/291 [00:00<00:01, 184.76it/s]
Loading weights:  20%|█▉        | 57/291 [00:00<00:01, 181.62it/s]
Loading weights:  26%|██▌       | 76/291 [00:00<00:01, 169.69it/s]
Loading weights:  33%|███▎      | 95/291 [00:00<00:01, 173.55it/s]
Loading weights:  41%|████      | 118/291 [00:00<00:00, 186.26it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 176.14it/s]
Loading weights:  54%|█████▍    | 158/291 [00:00<00:00, 176.68it/s]
Loading weights:  62%|██████▏   | 181/291 [00:01<00:00, 189.77it/s]
Loading weights:  69%|██████▉   | 201/291 [00:01<00:00, 189.55it/s]
Loading weights:  76%|███████▌  | 221/291 [00:01<00:00, 181.57it/s]
Loading weights:  82%|████████▏ | 240/291 [00:01<00:00, 180.89it/s]
Loading weights:  89%|████████▉ | 259/291 [00:01<00:00, 183.45it/s]
Loading weights:  96%|█████████▌| 279/291 [00:01<00:00, 186.64it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 182.49it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task511_reddit_tifu_long_text_summarization': PaR=44.5, shrinkage=ridge, n=1000
+  [SRT] Wired router: 12 sigs, cur=task511_reddit_tifu_long_text_summarization, mapping={'task511_reddit_tifu_long_text_summarization': 0, 'task1572_samsum_summary': 11, 'task363_sst2_polarity_classification': 10, 'task1290_xsum_summarization': 9, 'task181_outcome_extraction': 8, 'task002_quoref_answer_generation': 7, 'task1510_evalution_relation_extraction': 6, 'task639_multi_woz_user_utterance_generation': 5, 'task1729_personachat_generate_next': 4, 'task073_commonsenseqa_answer_generation': 3, 'task1590_diplomacy_text_generation': 2, 'task748_glucose_reverse_cause_event_detection': 1}
+[SRT] Router: 12 tasks, avg_PaR=44.5, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization']
+  [SRT-INIT] task_id_to_idx={'task511_reddit_tifu_long_text_summarization': 0, 'task1572_samsum_summary': 11, 'task363_sst2_polarity_classification': 10, 'task1290_xsum_summarization': 9, 'task181_outcome_extraction': 8, 'task002_quoref_answer_generation': 7, 'task1510_evalution_relation_extraction': 6, 'task639_multi_woz_user_utterance_generation': 5, 'task1729_personachat_generate_next': 4, 'task073_commonsenseqa_answer_generation': 3, 'task1590_diplomacy_text_generation': 2, 'task748_glucose_reverse_cause_event_detection': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/897 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=12
+
Continual predict:   0%|          | 1/897 [00:01<28:28,  1.91s/it]
Continual predict:   0%|          | 2/897 [00:04<35:46,  2.40s/it]
Continual predict:   0%|          | 3/897 [00:05<25:34,  1.72s/it]
Continual predict:   0%|          | 4/897 [00:06<19:02,  1.28s/it]
Continual predict:   1%|          | 5/897 [00:08<22:42,  1.53s/it]
Continual predict:   1%|          | 6/897 [00:13<40:43,  2.74s/it]
Continual predict:   1%|          | 7/897 [00:16<44:22,  2.99s/it]
Continual predict:   1%|          | 8/897 [00:18<40:40,  2.75s/it]
Continual predict:   1%|          | 9/897 [00:20<36:35,  2.47s/it]
Continual predict:   1%|          | 10/897 [00:21<29:12,  1.98s/it]
Continual predict:   1%|          | 11/897 [00:24<33:21,  2.26s/it]
Continual predict:   1%|▏         | 12/897 [00:27<38:20,  2.60s/it]
Continual predict:   1%|▏         | 13/897 [00:29<32:28,  2.20s/it]
Continual predict:   2%|▏         | 14/897 [00:31<34:07,  2.32s/it]
Continual predict:   2%|▏         | 15/897 [00:34<34:18,  2.33s/it]
Continual predict:   2%|▏         | 16/897 [00:36<35:56,  2.45s/it]
Continual predict:   2%|▏         | 17/897 [00:45<1:00:42,  4.14s/it]
Continual predict:   2%|▏         | 18/897 [00:47<52:14,  3.57s/it]  
Continual predict:   2%|▏         | 19/897 [00:49<48:19,  3.30s/it]
Continual predict:   2%|▏         | 20/897 [00:51<40:23,  2.76s/it]
Continual predict:   2%|▏         | 21/897 [00:51<29:02,  1.99s/it]
Continual predict:   2%|▏         | 22/897 [00:51<21:04,  1.45s/it]
Continual predict:   3%|▎         | 23/897 [00:51<15:35,  1.07s/it]
Continual predict:   3%|▎         | 24/897 [00:52<11:40,  1.25it/s]
Continual predict:   3%|▎         | 25/897 [00:52<08:55,  1.63it/s]
Continual predict:   3%|▎         | 26/897 [00:52<07:00,  2.07it/s]
Continual predict:   3%|▎         | 27/897 [00:52<05:39,  2.56it/s]
Continual predict:   3%|▎         | 28/897 [00:52<04:43,  3.07it/s]
Continual predict:   3%|▎         | 29/897 [00:53<04:04,  3.56it/s]
Continual predict:   3%|▎         | 30/897 [00:53<03:38,  3.97it/s]
Continual predict:   3%|▎         | 31/897 [00:53<03:19,  4.34it/s]
Continual predict:   4%|▎         | 32/897 [00:53<03:04,  4.68it/s]
Continual predict:   4%|▎         | 33/897 [00:53<02:56,  4.89it/s]
Continual predict:   4%|▍         | 34/897 [00:53<02:50,  5.06it/s]
Continual predict:   4%|▍         | 35/897 [00:54<02:45,  5.19it/s]
Continual predict:   4%|▍         | 36/897 [00:54<02:40,  5.35it/s]
Continual predict:   4%|▍         | 37/897 [00:54<02:39,  5.38it/s]
Continual predict:   4%|▍         | 38/897 [00:54<02:37,  5.46it/s]
Continual predict:   4%|▍         | 39/897 [00:54<02:38,  5.41it/s]
Continual predict:   4%|▍         | 40/897 [00:55<02:40,  5.33it/s]
Continual predict:   5%|▍         | 41/897 [00:55<02:39,  5.36it/s]
Continual predict:   5%|▍         | 42/897 [00:55<02:36,  5.48it/s]
Continual predict:   5%|▍         | 43/897 [00:55<02:35,  5.49it/s]
Continual predict:   5%|▍         | 44/897 [00:55<02:34,  5.51it/s]
Continual predict:   5%|▌         | 45/897 [00:55<02:32,  5.57it/s]
Continual predict:   5%|▌         | 46/897 [00:56<02:31,  5.62it/s]
Continual predict:   5%|▌         | 47/897 [00:56<02:30,  5.66it/s]
Continual predict:   5%|▌         | 48/897 [00:56<02:32,  5.57it/s]
Continual predict:   5%|▌         | 49/897 [00:56<02:32,  5.56it/s]
Continual predict:   6%|▌         | 50/897 [00:56<02:31,  5.58it/s]
Continual predict:   6%|▌         | 51/897 [00:57<02:30,  5.62it/s]
Continual predict:   6%|▌         | 52/897 [00:57<02:29,  5.65it/s]
Continual predict:   6%|▌         | 53/897 [00:57<02:28,  5.69it/s]
Continual predict:   6%|▌         | 54/897 [00:57<02:29,  5.63it/s]
Continual predict:   6%|▌         | 55/897 [00:57<02:28,  5.68it/s]
Continual predict:   6%|▌         | 56/897 [00:57<02:29,  5.63it/s]
Continual predict:   6%|▋         | 57/897 [00:58<02:28,  5.67it/s]
Continual predict:   6%|▋         | 58/897 [00:58<02:27,  5.70it/s]
Continual predict:   7%|▋         | 59/897 [00:58<02:26,  5.72it/s]
Continual predict:   7%|▋         | 60/897 [00:58<02:28,  5.63it/s]
Continual predict:   7%|▋         | 61/897 [00:58<02:27,  5.68it/s]
Continual predict:   7%|▋         | 62/897 [00:58<02:26,  5.71it/s]
Continual predict:   7%|▋         | 63/897 [00:59<02:27,  5.65it/s]
Continual predict:   7%|▋         | 64/897 [00:59<02:26,  5.68it/s]
Continual predict:   7%|▋         | 65/897 [00:59<02:26,  5.69it/s]
Continual predict:   7%|▋         | 66/897 [00:59<02:25,  5.70it/s]
Continual predict:   7%|▋         | 67/897 [00:59<02:24,  5.73it/s]
Continual predict:   8%|▊         | 68/897 [00:59<02:24,  5.75it/s]
Continual predict:   8%|▊         | 69/897 [01:00<02:23,  5.76it/s]
Continual predict:   8%|▊         | 70/897 [01:00<02:23,  5.77it/s]
Continual predict:   8%|▊         | 71/897 [01:00<02:27,  5.59it/s]
Continual predict:   8%|▊         | 72/897 [01:00<02:28,  5.56it/s]
Continual predict:   8%|▊         | 73/897 [01:00<02:26,  5.63it/s]
Continual predict:   8%|▊         | 74/897 [01:01<02:27,  5.59it/s]
Continual predict:   8%|▊         | 75/897 [01:01<02:26,  5.63it/s]
Continual predict:   8%|▊         | 76/897 [01:01<02:24,  5.68it/s]
Continual predict:   9%|▊         | 77/897 [01:01<02:23,  5.70it/s]
Continual predict:   9%|▊         | 78/897 [01:01<02:23,  5.71it/s]
Continual predict:   9%|▉         | 79/897 [01:01<02:23,  5.71it/s]
Continual predict:   9%|▉         | 80/897 [01:02<02:22,  5.73it/s]
Continual predict:   9%|▉         | 81/897 [01:02<02:22,  5.75it/s]
Continual predict:   9%|▉         | 82/897 [01:02<02:21,  5.76it/s]
Continual predict:   9%|▉         | 83/897 [01:02<02:21,  5.77it/s]
Continual predict:   9%|▉         | 84/897 [01:02<02:23,  5.68it/s]
Continual predict:   9%|▉         | 85/897 [01:02<02:22,  5.70it/s]
Continual predict:  10%|▉         | 86/897 [01:03<02:21,  5.72it/s]
Continual predict:  10%|▉         | 87/897 [01:03<02:21,  5.72it/s]
Continual predict:  10%|▉         | 88/897 [01:03<02:20,  5.75it/s]
Continual predict:  10%|▉         | 89/897 [01:03<02:20,  5.76it/s]
Continual predict:  10%|█         | 90/897 [01:03<02:20,  5.76it/s]
Continual predict:  10%|█         | 91/897 [01:04<02:19,  5.76it/s]
Continual predict:  10%|█         | 92/897 [01:04<02:19,  5.75it/s]
Continual predict:  10%|█         | 93/897 [01:04<02:21,  5.68it/s]
Continual predict:  10%|█         | 94/897 [01:04<02:22,  5.64it/s]
Continual predict:  11%|█         | 95/897 [01:04<02:21,  5.67it/s]
Continual predict:  11%|█         | 96/897 [01:04<02:21,  5.68it/s]
Continual predict:  11%|█         | 97/897 [01:05<02:20,  5.70it/s]
Continual predict:  11%|█         | 98/897 [01:05<02:38,  5.04it/s]
Continual predict:  11%|█         | 99/897 [01:05<03:01,  4.40it/s]
Continual predict:  11%|█         | 100/897 [01:05<02:49,  4.69it/s]
Continual predict:  11%|█▏        | 101/897 [01:05<02:40,  4.95it/s]
Continual predict:  11%|█▏        | 102/897 [01:06<02:33,  5.17it/s]
Continual predict:  11%|█▏        | 103/897 [01:06<02:28,  5.35it/s]
Continual predict:  12%|█▏        | 104/897 [01:06<02:25,  5.46it/s]
Continual predict:  12%|█▏        | 105/897 [01:06<02:24,  5.48it/s]
Continual predict:  12%|█▏        | 106/897 [01:06<02:22,  5.57it/s]
Continual predict:  12%|█▏        | 107/897 [01:07<02:20,  5.61it/s]
Continual predict:  12%|█▏        | 108/897 [01:07<02:19,  5.66it/s]
Continual predict:  12%|█▏        | 109/897 [01:07<02:18,  5.69it/s]
Continual predict:  12%|█▏        | 110/897 [01:07<02:17,  5.71it/s]
Continual predict:  12%|█▏        | 111/897 [01:07<02:17,  5.73it/s]
Continual predict:  12%|█▏        | 112/897 [01:07<02:16,  5.73it/s]
Continual predict:  13%|█▎        | 113/897 [01:08<02:16,  5.73it/s]
Continual predict:  13%|█▎        | 114/897 [01:08<02:16,  5.73it/s]
Continual predict:  13%|█▎        | 115/897 [01:08<02:17,  5.67it/s]
Continual predict:  13%|█▎        | 116/897 [01:08<02:17,  5.69it/s]
Continual predict:  13%|█▎        | 117/897 [01:08<02:18,  5.65it/s]
Continual predict:  13%|█▎        | 118/897 [01:08<02:18,  5.61it/s]
Continual predict:  13%|█▎        | 119/897 [01:09<02:18,  5.63it/s]
Continual predict:  13%|█▎        | 120/897 [01:09<02:17,  5.66it/s]
Continual predict:  13%|█▎        | 121/897 [01:14<20:16,  1.57s/it]
Continual predict:  14%|█▎        | 122/897 [01:18<31:25,  2.43s/it]
Continual predict:  14%|█▎        | 123/897 [01:24<43:40,  3.39s/it]
Continual predict:  14%|█▍        | 124/897 [01:26<40:21,  3.13s/it]
Continual predict:  14%|█▍        | 125/897 [01:31<47:45,  3.71s/it]
Continual predict:  14%|█▍        | 126/897 [01:36<52:50,  4.11s/it]
Continual predict:  14%|█▍        | 127/897 [01:38<43:07,  3.36s/it]
Continual predict:  14%|█▍        | 128/897 [01:41<41:31,  3.24s/it]
Continual predict:  14%|█▍        | 129/897 [01:45<42:53,  3.35s/it]
Continual predict:  14%|█▍        | 130/897 [01:48<42:33,  3.33s/it]
Continual predict:  15%|█▍        | 131/897 [01:50<39:06,  3.06s/it]
Continual predict:  15%|█▍        | 132/897 [01:53<38:20,  3.01s/it]
Continual predict:  15%|█▍        | 133/897 [01:56<39:42,  3.12s/it]
Continual predict:  15%|█▍        | 134/897 [02:03<52:14,  4.11s/it]
Continual predict:  15%|█▌        | 135/897 [02:08<56:03,  4.41s/it]
Continual predict:  15%|█▌        | 136/897 [02:12<52:32,  4.14s/it]
Continual predict:  15%|█▌        | 137/897 [02:14<45:53,  3.62s/it]
Continual predict:  15%|█▌        | 138/897 [02:15<35:18,  2.79s/it]
Continual predict:  15%|█▌        | 139/897 [02:16<29:32,  2.34s/it]
Continual predict:  16%|█▌        | 140/897 [02:19<31:33,  2.50s/it]
Continual predict:  16%|█▌        | 141/897 [02:26<48:17,  3.83s/it]
Continual predict:  16%|█▌        | 142/897 [02:31<51:59,  4.13s/it]
Continual predict:  16%|█▌        | 143/897 [02:34<49:34,  3.94s/it]
Continual predict:  16%|█▌        | 144/897 [02:44<1:11:59,  5.74s/it]
Continual predict:  16%|█▌        | 145/897 [02:49<1:08:00,  5.43s/it]
Continual predict:  16%|█▋        | 146/897 [02:51<54:53,  4.39s/it]  
Continual predict:  16%|█▋        | 147/897 [02:55<55:48,  4.46s/it]
Continual predict:  16%|█▋        | 148/897 [02:58<47:42,  3.82s/it]
Continual predict:  17%|█▋        | 149/897 [03:01<45:58,  3.69s/it]
Continual predict:  17%|█▋        | 150/897 [03:03<39:57,  3.21s/it]
Continual predict:  17%|█▋        | 151/897 [03:06<39:57,  3.21s/it]
Continual predict:  17%|█▋        | 152/897 [03:13<50:22,  4.06s/it]
Continual predict:  17%|█▋        | 153/897 [03:15<45:04,  3.64s/it]
Continual predict:  17%|█▋        | 154/897 [03:20<50:23,  4.07s/it]
Continual predict:  17%|█▋        | 155/897 [03:24<49:31,  4.01s/it]
Continual predict:  17%|█▋        | 156/897 [03:28<48:04,  3.89s/it]
Continual predict:  18%|█▊        | 157/897 [03:32<48:49,  3.96s/it]
Continual predict:  18%|█▊        | 158/897 [03:33<38:17,  3.11s/it]
Continual predict:  18%|█▊        | 159/897 [03:40<53:46,  4.37s/it]
Continual predict:  18%|█▊        | 160/897 [03:43<46:35,  3.79s/it]
Continual predict:  18%|█▊        | 161/897 [03:45<42:20,  3.45s/it]
Continual predict:  18%|█▊        | 162/897 [03:50<48:14,  3.94s/it]
Continual predict:  18%|█▊        | 163/897 [03:53<41:55,  3.43s/it]
Continual predict:  18%|█▊        | 164/897 [03:55<38:15,  3.13s/it]
Continual predict:  18%|█▊        | 165/897 [03:58<38:58,  3.19s/it]
Continual predict:  19%|█▊        | 166/897 [04:01<35:44,  2.93s/it]
Continual predict:  19%|█▊        | 167/897 [04:02<30:01,  2.47s/it]
Continual predict:  19%|█▊        | 168/897 [04:05<29:47,  2.45s/it]
Continual predict:  19%|█▉        | 169/897 [04:06<27:38,  2.28s/it]
Continual predict:  19%|█▉        | 170/897 [04:10<31:08,  2.57s/it]
Continual predict:  19%|█▉        | 171/897 [04:11<27:54,  2.31s/it]
Continual predict:  19%|█▉        | 172/897 [04:14<29:35,  2.45s/it]
Continual predict:  19%|█▉        | 173/897 [04:18<35:42,  2.96s/it]
Continual predict:  19%|█▉        | 174/897 [04:20<29:33,  2.45s/it]
Continual predict:  20%|█▉        | 175/897 [04:23<34:02,  2.83s/it]
Continual predict:  20%|█▉        | 176/897 [04:26<34:51,  2.90s/it]
Continual predict:  20%|█▉        | 177/897 [04:29<34:53,  2.91s/it]
Continual predict:  20%|█▉        | 178/897 [04:31<29:21,  2.45s/it]
Continual predict:  20%|█▉        | 179/897 [04:32<24:55,  2.08s/it]
Continual predict:  20%|██        | 180/897 [04:33<22:22,  1.87s/it]
Continual predict:  20%|██        | 181/897 [04:35<20:34,  1.72s/it]
Continual predict:  20%|██        | 182/897 [04:40<32:07,  2.70s/it]
Continual predict:  20%|██        | 183/897 [04:40<23:17,  1.96s/it]
Continual predict:  21%|██        | 184/897 [04:41<20:22,  1.71s/it]
Continual predict:  21%|██        | 185/897 [04:45<27:06,  2.29s/it]
Continual predict:  21%|██        | 186/897 [04:46<22:54,  1.93s/it]
Continual predict:  21%|██        | 187/897 [04:49<28:16,  2.39s/it]
Continual predict:  21%|██        | 188/897 [04:52<29:52,  2.53s/it]
Continual predict:  21%|██        | 189/897 [04:57<39:46,  3.37s/it]
Continual predict:  21%|██        | 190/897 [05:01<40:20,  3.42s/it]
Continual predict:  21%|██▏       | 191/897 [05:02<33:13,  2.82s/it]
Continual predict:  21%|██▏       | 192/897 [05:06<35:42,  3.04s/it]
Continual predict:  22%|██▏       | 193/897 [05:11<42:01,  3.58s/it]
Continual predict:  22%|██▏       | 194/897 [05:16<47:18,  4.04s/it]
Continual predict:  22%|██▏       | 195/897 [05:22<53:00,  4.53s/it]
Continual predict:  22%|██▏       | 196/897 [05:25<49:22,  4.23s/it]
Continual predict:  22%|██▏       | 197/897 [05:29<46:39,  4.00s/it]
Continual predict:  22%|██▏       | 198/897 [05:36<59:49,  5.14s/it]
Continual predict:  22%|██▏       | 199/897 [05:37<45:50,  3.94s/it]
Continual predict:  22%|██▏       | 200/897 [05:39<37:33,  3.23s/it]
Continual predict:  22%|██▏       | 201/897 [05:42<35:42,  3.08s/it]
Continual predict:  23%|██▎       | 202/897 [05:45<36:57,  3.19s/it]
Continual predict:  23%|██▎       | 203/897 [05:51<44:36,  3.86s/it]
Continual predict:  23%|██▎       | 204/897 [05:54<42:25,  3.67s/it]
Continual predict:  23%|██▎       | 205/897 [06:01<53:38,  4.65s/it]
Continual predict:  23%|██▎       | 206/897 [06:04<47:33,  4.13s/it]
Continual predict:  23%|██▎       | 207/897 [06:08<46:45,  4.07s/it]
Continual predict:  23%|██▎       | 208/897 [06:14<54:57,  4.79s/it]
Continual predict:  23%|██▎       | 209/897 [06:17<48:12,  4.20s/it]
Continual predict:  23%|██▎       | 210/897 [06:19<39:41,  3.47s/it]
Continual predict:  24%|██▎       | 211/897 [06:21<37:16,  3.26s/it]
Continual predict:  24%|██▎       | 212/897 [06:25<37:27,  3.28s/it]
Continual predict:  24%|██▎       | 213/897 [06:29<39:48,  3.49s/it]
Continual predict:  24%|██▍       | 214/897 [06:33<41:03,  3.61s/it]
Continual predict:  24%|██▍       | 215/897 [06:34<31:40,  2.79s/it]
Continual predict:  24%|██▍       | 216/897 [06:37<33:14,  2.93s/it]
Continual predict:  24%|██▍       | 217/897 [06:39<30:46,  2.72s/it]
Continual predict:  24%|██▍       | 218/897 [06:43<35:55,  3.17s/it]
Continual predict:  24%|██▍       | 219/897 [06:47<36:56,  3.27s/it]
Continual predict:  25%|██▍       | 220/897 [06:51<40:31,  3.59s/it]
Continual predict:  25%|██▍       | 221/897 [06:51<29:06,  2.58s/it]
Continual predict:  25%|██▍       | 222/897 [06:52<21:21,  1.90s/it]
Continual predict:  25%|██▍       | 223/897 [06:52<15:41,  1.40s/it]
Continual predict:  25%|██▍       | 224/897 [06:52<11:43,  1.05s/it]
Continual predict:  25%|██▌       | 225/897 [06:52<08:57,  1.25it/s]
Continual predict:  25%|██▌       | 226/897 [06:53<07:01,  1.59it/s]
Continual predict:  25%|██▌       | 227/897 [06:53<05:50,  1.91it/s]
Continual predict:  25%|██▌       | 228/897 [06:53<05:01,  2.22it/s]
Continual predict:  26%|██▌       | 229/897 [06:53<04:19,  2.57it/s]
Continual predict:  26%|██▌       | 230/897 [06:54<05:00,  2.22it/s]
Continual predict:  26%|██▌       | 231/897 [06:54<05:05,  2.18it/s]
Continual predict:  26%|██▌       | 232/897 [06:55<04:18,  2.57it/s]
Continual predict:  26%|██▌       | 233/897 [06:55<03:45,  2.94it/s]
Continual predict:  26%|██▌       | 234/897 [06:55<03:23,  3.26it/s]
Continual predict:  26%|██▌       | 235/897 [06:55<03:06,  3.54it/s]
Continual predict:  26%|██▋       | 236/897 [06:56<02:55,  3.77it/s]
Continual predict:  26%|██▋       | 237/897 [06:56<02:47,  3.94it/s]
Continual predict:  27%|██▋       | 238/897 [06:56<02:41,  4.08it/s]
Continual predict:  27%|██▋       | 239/897 [06:56<02:37,  4.18it/s]
Continual predict:  27%|██▋       | 240/897 [06:57<03:04,  3.56it/s]
Continual predict:  27%|██▋       | 241/897 [06:57<02:53,  3.78it/s]
Continual predict:  27%|██▋       | 242/897 [06:57<02:45,  3.95it/s]
Continual predict:  27%|██▋       | 243/897 [06:57<02:40,  4.09it/s]
Continual predict:  27%|██▋       | 244/897 [06:57<02:36,  4.18it/s]
Continual predict:  27%|██▋       | 245/897 [06:58<02:33,  4.25it/s]
Continual predict:  27%|██▋       | 246/897 [06:58<02:31,  4.30it/s]
Continual predict:  28%|██▊       | 247/897 [06:58<02:40,  4.05it/s]
Continual predict:  28%|██▊       | 248/897 [06:58<02:36,  4.16it/s]
Continual predict:  28%|██▊       | 249/897 [06:59<02:43,  3.97it/s]
Continual predict:  28%|██▊       | 250/897 [06:59<02:38,  4.08it/s]
Continual predict:  28%|██▊       | 251/897 [06:59<02:44,  3.92it/s]
Continual predict:  28%|██▊       | 252/897 [07:01<06:10,  1.74it/s]
Continual predict:  28%|██▊       | 253/897 [07:01<05:12,  2.06it/s]
Continual predict:  28%|██▊       | 254/897 [07:01<04:22,  2.45it/s]
Continual predict:  28%|██▊       | 255/897 [07:01<03:47,  2.83it/s]
Continual predict:  29%|██▊       | 256/897 [07:02<03:22,  3.17it/s]
Continual predict:  29%|██▊       | 257/897 [07:02<04:51,  2.19it/s]
Continual predict:  29%|██▉       | 258/897 [07:03<04:07,  2.58it/s]
Continual predict:  29%|██▉       | 259/897 [07:03<03:36,  2.95it/s]
Continual predict:  29%|██▉       | 260/897 [07:03<03:14,  3.27it/s]
Continual predict:  29%|██▉       | 261/897 [07:04<04:11,  2.53it/s]
Continual predict:  29%|██▉       | 262/897 [07:04<03:42,  2.85it/s]
Continual predict:  29%|██▉       | 263/897 [07:04<03:28,  3.03it/s]
Continual predict:  29%|██▉       | 264/897 [07:05<06:32,  1.61it/s]
Continual predict:  30%|██▉       | 265/897 [07:06<06:22,  1.65it/s]
Continual predict:  30%|██▉       | 266/897 [07:07<06:19,  1.66it/s]
Continual predict:  30%|██▉       | 267/897 [07:07<06:25,  1.63it/s]
Continual predict:  30%|██▉       | 268/897 [07:08<07:14,  1.45it/s]
Continual predict:  30%|██▉       | 269/897 [07:09<07:03,  1.48it/s]
Continual predict:  30%|███       | 270/897 [07:09<07:01,  1.49it/s]
Continual predict:  30%|███       | 271/897 [07:10<07:24,  1.41it/s]
Continual predict:  30%|███       | 272/897 [07:11<08:08,  1.28it/s]
Continual predict:  30%|███       | 273/897 [07:12<07:49,  1.33it/s]
Continual predict:  31%|███       | 274/897 [07:13<07:36,  1.36it/s]
Continual predict:  31%|███       | 275/897 [07:13<07:17,  1.42it/s]
Continual predict:  31%|███       | 276/897 [07:14<08:33,  1.21it/s]
Continual predict:  31%|███       | 277/897 [07:16<10:08,  1.02it/s]
Continual predict:  31%|███       | 278/897 [07:16<09:46,  1.06it/s]
Continual predict:  31%|███       | 279/897 [07:17<09:08,  1.13it/s]
Continual predict:  31%|███       | 280/897 [07:18<07:51,  1.31it/s]
Continual predict:  31%|███▏      | 281/897 [07:19<08:35,  1.19it/s]
Continual predict:  31%|███▏      | 282/897 [07:20<08:37,  1.19it/s]
Continual predict:  32%|███▏      | 283/897 [07:20<08:12,  1.25it/s]
Continual predict:  32%|███▏      | 284/897 [07:21<07:12,  1.42it/s]
Continual predict:  32%|███▏      | 285/897 [07:22<07:34,  1.35it/s]
Continual predict:  32%|███▏      | 286/897 [07:22<07:44,  1.31it/s]
Continual predict:  32%|███▏      | 287/897 [07:23<07:51,  1.29it/s]
Continual predict:  32%|███▏      | 288/897 [07:24<07:33,  1.34it/s]
Continual predict:  32%|███▏      | 289/897 [07:25<08:17,  1.22it/s]
Continual predict:  32%|███▏      | 290/897 [07:26<08:12,  1.23it/s]
Continual predict:  32%|███▏      | 291/897 [07:26<07:40,  1.32it/s]
Continual predict:  33%|███▎      | 292/897 [07:27<06:21,  1.58it/s]
Continual predict:  33%|███▎      | 293/897 [07:27<07:04,  1.42it/s]
Continual predict:  33%|███▎      | 294/897 [07:28<06:39,  1.51it/s]
Continual predict:  33%|███▎      | 295/897 [07:29<06:04,  1.65it/s]
Continual predict:  33%|███▎      | 296/897 [07:30<07:19,  1.37it/s]
Continual predict:  33%|███▎      | 297/897 [07:30<06:47,  1.47it/s]
Continual predict:  33%|███▎      | 298/897 [07:31<06:10,  1.62it/s]
Continual predict:  33%|███▎      | 299/897 [07:32<07:14,  1.38it/s]
Continual predict:  33%|███▎      | 300/897 [07:32<06:05,  1.63it/s]
Continual predict:  34%|███▎      | 301/897 [07:33<07:06,  1.40it/s]
Continual predict:  34%|███▎      | 302/897 [07:33<06:19,  1.57it/s]
Continual predict:  34%|███▍      | 303/897 [07:35<09:09,  1.08it/s]
Continual predict:  34%|███▍      | 304/897 [07:35<07:25,  1.33it/s]
Continual predict:  34%|███▍      | 305/897 [07:36<08:00,  1.23it/s]
Continual predict:  34%|███▍      | 306/897 [07:37<08:07,  1.21it/s]
Continual predict:  34%|███▍      | 307/897 [07:38<08:09,  1.21it/s]
Continual predict:  34%|███▍      | 308/897 [07:38<07:02,  1.39it/s]
Continual predict:  34%|███▍      | 309/897 [07:39<06:19,  1.55it/s]
Continual predict:  35%|███▍      | 310/897 [07:39<06:18,  1.55it/s]
Continual predict:  35%|███▍      | 311/897 [07:40<06:05,  1.61it/s]
Continual predict:  35%|███▍      | 312/897 [07:41<06:26,  1.52it/s]
Continual predict:  35%|███▍      | 313/897 [07:42<07:10,  1.36it/s]
Continual predict:  35%|███▌      | 314/897 [07:42<06:25,  1.51it/s]
Continual predict:  35%|███▌      | 315/897 [07:43<07:14,  1.34it/s]
Continual predict:  35%|███▌      | 316/897 [07:44<07:31,  1.29it/s]
Continual predict:  35%|███▌      | 317/897 [07:44<06:15,  1.55it/s]
Continual predict:  35%|███▌      | 318/897 [07:45<05:45,  1.68it/s]
Continual predict:  36%|███▌      | 319/897 [07:47<08:56,  1.08it/s]
Continual predict:  36%|███▌      | 320/897 [07:47<08:24,  1.14it/s]
Continual predict:  36%|███▌      | 321/897 [07:48<07:59,  1.20it/s]
Continual predict:  36%|███▌      | 322/897 [07:48<06:53,  1.39it/s]
Continual predict:  36%|███▌      | 323/897 [07:49<07:06,  1.35it/s]
Continual predict:  36%|███▌      | 324/897 [07:50<07:39,  1.25it/s]
Continual predict:  36%|███▌      | 325/897 [07:51<07:38,  1.25it/s]
Continual predict:  36%|███▋      | 326/897 [07:52<07:14,  1.31it/s]
Continual predict:  36%|███▋      | 327/897 [07:53<08:27,  1.12it/s]
Continual predict:  37%|███▋      | 328/897 [07:54<09:03,  1.05it/s]
Continual predict:  37%|███▋      | 329/897 [07:55<07:56,  1.19it/s]
Continual predict:  37%|███▋      | 330/897 [07:55<06:31,  1.45it/s]
Continual predict:  37%|███▋      | 331/897 [07:56<06:20,  1.49it/s]
Continual predict:  37%|███▋      | 332/897 [07:56<06:02,  1.56it/s]
Continual predict:  37%|███▋      | 333/897 [07:57<07:50,  1.20it/s]
Continual predict:  37%|███▋      | 334/897 [07:58<08:08,  1.15it/s]
Continual predict:  37%|███▋      | 335/897 [07:59<06:58,  1.34it/s]
Continual predict:  37%|███▋      | 336/897 [07:59<06:12,  1.51it/s]
Continual predict:  38%|███▊      | 337/897 [08:00<06:01,  1.55it/s]
Continual predict:  38%|███▊      | 338/897 [08:01<06:06,  1.52it/s]
Continual predict:  38%|███▊      | 339/897 [08:01<05:50,  1.59it/s]
Continual predict:  38%|███▊      | 340/897 [08:02<07:14,  1.28it/s]
Continual predict:  38%|███▊      | 341/897 [08:03<07:15,  1.28it/s]
Continual predict:  38%|███▊      | 342/897 [08:03<06:23,  1.45it/s]
Continual predict:  38%|███▊      | 343/897 [08:04<06:39,  1.39it/s]
Continual predict:  38%|███▊      | 344/897 [08:05<07:25,  1.24it/s]
Continual predict:  38%|███▊      | 345/897 [08:06<06:30,  1.41it/s]
Continual predict:  39%|███▊      | 346/897 [08:07<06:53,  1.33it/s]
Continual predict:  39%|███▊      | 347/897 [08:08<09:04,  1.01it/s]
Continual predict:  39%|███▉      | 348/897 [08:09<07:19,  1.25it/s]
Continual predict:  39%|███▉      | 349/897 [08:09<06:22,  1.43it/s]
Continual predict:  39%|███▉      | 350/897 [08:10<06:19,  1.44it/s]
Continual predict:  39%|███▉      | 351/897 [08:11<07:38,  1.19it/s]
Continual predict:  39%|███▉      | 352/897 [08:12<07:29,  1.21it/s]
Continual predict:  39%|███▉      | 353/897 [08:13<08:55,  1.02it/s]
Continual predict:  39%|███▉      | 354/897 [08:14<09:49,  1.09s/it]
Continual predict:  40%|███▉      | 355/897 [08:15<08:35,  1.05it/s]
Continual predict:  40%|███▉      | 356/897 [08:16<08:35,  1.05it/s]
Continual predict:  40%|███▉      | 357/897 [08:17<09:02,  1.01s/it]
Continual predict:  40%|███▉      | 358/897 [08:18<08:35,  1.05it/s]
Continual predict:  40%|████      | 359/897 [08:19<07:48,  1.15it/s]
Continual predict:  40%|████      | 360/897 [08:20<08:59,  1.00s/it]
Continual predict:  40%|████      | 361/897 [08:22<10:43,  1.20s/it]
Continual predict:  40%|████      | 362/897 [08:22<09:11,  1.03s/it]
Continual predict:  40%|████      | 363/897 [08:23<07:41,  1.16it/s]
Continual predict:  41%|████      | 364/897 [08:23<06:20,  1.40it/s]
Continual predict:  41%|████      | 365/897 [08:23<05:14,  1.69it/s]
Continual predict:  41%|████      | 366/897 [08:24<04:27,  1.98it/s]
Continual predict:  41%|████      | 367/897 [08:24<03:54,  2.26it/s]
Continual predict:  41%|████      | 368/897 [08:24<03:32,  2.49it/s]
Continual predict:  41%|████      | 369/897 [08:25<03:16,  2.69it/s]
Continual predict:  41%|████      | 370/897 [08:25<03:04,  2.85it/s]
Continual predict:  41%|████▏     | 371/897 [08:25<03:06,  2.82it/s]
Continual predict:  41%|████▏     | 372/897 [08:26<03:16,  2.67it/s]
Continual predict:  42%|████▏     | 373/897 [08:26<03:23,  2.57it/s]
Continual predict:  42%|████▏     | 374/897 [08:26<03:09,  2.76it/s]
Continual predict:  42%|████▏     | 375/897 [08:27<03:18,  2.62it/s]
Continual predict:  42%|████▏     | 376/897 [08:27<03:15,  2.66it/s]
Continual predict:  42%|████▏     | 377/897 [08:27<03:03,  2.83it/s]
Continual predict:  42%|████▏     | 378/897 [08:28<02:55,  2.96it/s]
Continual predict:  42%|████▏     | 379/897 [08:28<02:49,  3.06it/s]
Continual predict:  42%|████▏     | 380/897 [08:28<02:59,  2.88it/s]
Continual predict:  42%|████▏     | 381/897 [08:29<03:07,  2.76it/s]
Continual predict:  43%|████▎     | 382/897 [08:29<02:57,  2.90it/s]
Continual predict:  43%|████▎     | 383/897 [08:29<02:50,  3.01it/s]
Continual predict:  43%|████▎     | 384/897 [08:30<02:45,  3.09it/s]
Continual predict:  43%|████▎     | 385/897 [08:30<02:42,  3.15it/s]
Continual predict:  43%|████▎     | 386/897 [08:30<02:39,  3.20it/s]
Continual predict:  43%|████▎     | 387/897 [08:31<02:38,  3.23it/s]
Continual predict:  43%|████▎     | 388/897 [08:31<02:36,  3.25it/s]
Continual predict:  43%|████▎     | 389/897 [08:31<02:44,  3.08it/s]
Continual predict:  43%|████▎     | 390/897 [08:32<02:41,  3.15it/s]
Continual predict:  44%|████▎     | 391/897 [08:32<02:38,  3.19it/s]
Continual predict:  44%|████▎     | 392/897 [08:32<02:45,  3.05it/s]
Continual predict:  44%|████▍     | 393/897 [08:33<02:41,  3.12it/s]
Continual predict:  44%|████▍     | 394/897 [08:33<02:38,  3.17it/s]
Continual predict:  44%|████▍     | 395/897 [08:33<02:27,  3.41it/s]
Continual predict:  44%|████▍     | 396/897 [08:34<02:46,  3.01it/s]
Continual predict:  44%|████▍     | 397/897 [08:34<02:41,  3.09it/s]
Continual predict:  44%|████▍     | 398/897 [08:34<02:38,  3.15it/s]
Continual predict:  44%|████▍     | 399/897 [08:34<02:35,  3.19it/s]
Continual predict:  45%|████▍     | 400/897 [08:35<02:34,  3.22it/s]
Continual predict:  45%|████▍     | 401/897 [08:35<02:32,  3.25it/s]
Continual predict:  45%|████▍     | 402/897 [08:35<02:31,  3.27it/s]
Continual predict:  45%|████▍     | 403/897 [08:36<02:39,  3.10it/s]
Continual predict:  45%|████▌     | 404/897 [08:36<02:36,  3.16it/s]
Continual predict:  45%|████▌     | 405/897 [08:36<02:51,  2.87it/s]
Continual predict:  45%|████▌     | 406/897 [08:37<02:44,  2.99it/s]
Continual predict:  45%|████▌     | 407/897 [08:37<02:39,  3.07it/s]
Continual predict:  45%|████▌     | 408/897 [08:37<02:35,  3.14it/s]
Continual predict:  46%|████▌     | 409/897 [08:38<02:32,  3.19it/s]
Continual predict:  46%|████▌     | 410/897 [08:38<02:48,  2.89it/s]
Continual predict:  46%|████▌     | 411/897 [08:38<02:59,  2.71it/s]
Continual predict:  46%|████▌     | 412/897 [08:39<02:49,  2.85it/s]
Continual predict:  46%|████▌     | 413/897 [08:39<02:47,  2.90it/s]
Continual predict:  46%|████▌     | 414/897 [08:39<02:41,  2.99it/s]
Continual predict:  46%|████▋     | 415/897 [08:40<02:36,  3.08it/s]
Continual predict:  46%|████▋     | 416/897 [08:40<02:33,  3.14it/s]
Continual predict:  46%|████▋     | 417/897 [08:40<02:30,  3.19it/s]
Continual predict:  47%|████▋     | 418/897 [08:41<02:37,  3.05it/s]
Continual predict:  47%|████▋     | 419/897 [08:41<02:50,  2.80it/s]
Continual predict:  47%|████▋     | 420/897 [08:42<02:59,  2.65it/s]
Continual predict:  47%|████▋     | 421/897 [08:42<03:06,  2.56it/s]
Continual predict:  47%|████▋     | 422/897 [08:42<03:01,  2.61it/s]
Continual predict:  47%|████▋     | 423/897 [08:43<02:58,  2.65it/s]
Continual predict:  47%|████▋     | 424/897 [08:43<03:04,  2.56it/s]
Continual predict:  47%|████▋     | 425/897 [08:44<03:08,  2.50it/s]
Continual predict:  47%|████▋     | 426/897 [08:44<03:11,  2.46it/s]
Continual predict:  48%|████▊     | 427/897 [08:44<03:05,  2.54it/s]
Continual predict:  48%|████▊     | 428/897 [08:45<02:51,  2.73it/s]
Continual predict:  48%|████▊     | 429/897 [08:45<02:42,  2.89it/s]
Continual predict:  48%|████▊     | 430/897 [08:45<02:35,  3.00it/s]
Continual predict:  48%|████▊     | 431/897 [08:46<02:39,  2.92it/s]
Continual predict:  48%|████▊     | 432/897 [08:46<02:33,  3.03it/s]
Continual predict:  48%|████▊     | 433/897 [08:46<02:29,  3.11it/s]
Continual predict:  48%|████▊     | 434/897 [08:47<02:34,  2.99it/s]
Continual predict:  48%|████▊     | 435/897 [08:47<02:29,  3.08it/s]
Continual predict:  49%|████▊     | 436/897 [08:47<02:43,  2.82it/s]
Continual predict:  49%|████▊     | 437/897 [08:48<03:00,  2.54it/s]
Continual predict:  49%|████▉     | 438/897 [08:48<02:47,  2.73it/s]
Continual predict:  49%|████▉     | 439/897 [08:48<02:47,  2.74it/s]
Continual predict:  49%|████▉     | 440/897 [08:49<02:46,  2.74it/s]
Continual predict:  49%|████▉     | 441/897 [08:49<02:45,  2.75it/s]
Continual predict:  49%|████▉     | 442/897 [08:49<02:38,  2.87it/s]
Continual predict:  49%|████▉     | 443/897 [08:50<02:35,  2.91it/s]
Continual predict:  49%|████▉     | 444/897 [08:50<02:29,  3.02it/s]
Continual predict:  50%|████▉     | 445/897 [08:50<02:25,  3.10it/s]
Continual predict:  50%|████▉     | 446/897 [08:51<02:30,  2.99it/s]
Continual predict:  50%|████▉     | 447/897 [08:51<02:33,  2.92it/s]
Continual predict:  50%|████▉     | 448/897 [08:51<02:36,  2.87it/s]
Continual predict:  50%|█████     | 449/897 [08:52<02:46,  2.70it/s]
Continual predict:  50%|█████     | 450/897 [08:52<02:36,  2.85it/s]
Continual predict:  50%|█████     | 451/897 [08:53<02:30,  2.97it/s]
Continual predict:  50%|█████     | 452/897 [08:53<02:25,  3.06it/s]
Continual predict:  51%|█████     | 453/897 [08:53<02:21,  3.13it/s]
Continual predict:  51%|█████     | 454/897 [08:53<02:19,  3.19it/s]
Continual predict:  51%|█████     | 455/897 [08:54<02:17,  3.22it/s]
Continual predict:  51%|█████     | 456/897 [08:54<02:24,  3.06it/s]
Continual predict:  51%|█████     | 457/897 [08:54<02:29,  2.94it/s]
Continual predict:  51%|█████     | 458/897 [08:55<02:24,  3.03it/s]
Continual predict:  51%|█████     | 459/897 [08:55<02:21,  3.10it/s]
Continual predict:  51%|█████▏    | 460/897 [08:55<02:18,  3.15it/s]
Continual predict:  51%|█████▏    | 461/897 [08:56<02:16,  3.19it/s]
Continual predict:  52%|█████▏    | 462/897 [08:56<02:14,  3.23it/s]
Continual predict:  52%|█████▏    | 463/897 [08:56<02:13,  3.25it/s]
Continual predict:  52%|█████▏    | 464/897 [08:57<02:25,  2.97it/s]
Continual predict:  52%|█████▏    | 465/897 [08:57<02:11,  3.28it/s]
Continual predict:  52%|█████▏    | 466/897 [08:57<02:16,  3.16it/s]
Continual predict:  52%|█████▏    | 467/897 [08:58<02:12,  3.26it/s]
Continual predict:  52%|█████▏    | 468/897 [08:58<02:16,  3.15it/s]
Continual predict:  52%|█████▏    | 469/897 [08:58<02:19,  3.08it/s]
Continual predict:  52%|█████▏    | 470/897 [08:59<02:21,  3.02it/s]
Continual predict:  53%|█████▎    | 471/897 [08:59<02:22,  2.99it/s]
Continual predict:  53%|█████▎    | 472/897 [09:00<03:44,  1.89it/s]
Continual predict:  53%|█████▎    | 473/897 [09:01<04:00,  1.76it/s]
Continual predict:  53%|█████▎    | 474/897 [09:01<03:16,  2.15it/s]
Continual predict:  53%|█████▎    | 475/897 [09:01<03:01,  2.33it/s]
Continual predict:  53%|█████▎    | 476/897 [09:02<02:57,  2.37it/s]
Continual predict:  53%|█████▎    | 477/897 [09:02<02:47,  2.50it/s]
Continual predict:  53%|█████▎    | 478/897 [09:02<02:47,  2.50it/s]
Continual predict:  53%|█████▎    | 479/897 [09:03<03:39,  1.90it/s]
Continual predict:  54%|█████▎    | 480/897 [09:03<03:16,  2.12it/s]
Continual predict:  54%|█████▎    | 481/897 [09:04<03:07,  2.22it/s]
Continual predict:  54%|█████▎    | 482/897 [09:04<03:16,  2.12it/s]
Continual predict:  54%|█████▍    | 483/897 [09:05<04:08,  1.67it/s]
Continual predict:  54%|█████▍    | 484/897 [09:06<05:12,  1.32it/s]
Continual predict:  54%|█████▍    | 485/897 [09:08<06:04,  1.13it/s]
Continual predict:  54%|█████▍    | 486/897 [09:09<06:32,  1.05it/s]
Continual predict:  54%|█████▍    | 487/897 [09:10<07:47,  1.14s/it]
Continual predict:  54%|█████▍    | 488/897 [09:11<06:22,  1.07it/s]
Continual predict:  55%|█████▍    | 489/897 [09:11<05:58,  1.14it/s]
Continual predict:  55%|█████▍    | 490/897 [09:12<05:50,  1.16it/s]
Continual predict:  55%|█████▍    | 491/897 [09:13<05:46,  1.17it/s]
Continual predict:  55%|█████▍    | 492/897 [09:14<05:59,  1.13it/s]
Continual predict:  55%|█████▍    | 493/897 [09:15<05:44,  1.17it/s]
Continual predict:  55%|█████▌    | 494/897 [09:16<05:51,  1.15it/s]
Continual predict:  55%|█████▌    | 495/897 [09:17<07:22,  1.10s/it]
Continual predict:  55%|█████▌    | 496/897 [09:18<06:49,  1.02s/it]
Continual predict:  55%|█████▌    | 497/897 [09:19<05:57,  1.12it/s]
Continual predict:  56%|█████▌    | 498/897 [09:20<05:55,  1.12it/s]
Continual predict:  56%|█████▌    | 499/897 [09:21<05:55,  1.12it/s]
Continual predict:  56%|█████▌    | 500/897 [09:21<05:03,  1.31it/s]
Continual predict:  56%|█████▌    | 501/897 [09:22<04:44,  1.39it/s]
Continual predict:  56%|█████▌    | 502/897 [09:22<04:38,  1.42it/s]
Continual predict:  56%|█████▌    | 503/897 [09:23<04:53,  1.34it/s]
Continual predict:  56%|█████▌    | 504/897 [09:24<04:34,  1.43it/s]
Continual predict:  56%|█████▋    | 505/897 [09:25<05:29,  1.19it/s]
Continual predict:  56%|█████▋    | 506/897 [09:26<05:39,  1.15it/s]
Continual predict:  57%|█████▋    | 507/897 [09:27<05:50,  1.11it/s]
Continual predict:  57%|█████▋    | 508/897 [09:28<05:31,  1.17it/s]
Continual predict:  57%|█████▋    | 509/897 [09:29<05:35,  1.16it/s]
Continual predict:  57%|█████▋    | 510/897 [09:30<07:29,  1.16s/it]
Continual predict:  57%|█████▋    | 511/897 [09:32<07:23,  1.15s/it]
Continual predict:  57%|█████▋    | 512/897 [09:32<06:29,  1.01s/it]
Continual predict:  57%|█████▋    | 513/897 [09:34<07:44,  1.21s/it]
Continual predict:  57%|█████▋    | 514/897 [09:35<07:21,  1.15s/it]
Continual predict:  57%|█████▋    | 515/897 [09:36<06:34,  1.03s/it]
Continual predict:  58%|█████▊    | 516/897 [09:36<06:00,  1.06it/s]
Continual predict:  58%|█████▊    | 517/897 [09:37<06:08,  1.03it/s]
Continual predict:  58%|█████▊    | 518/897 [09:38<06:06,  1.03it/s]
Continual predict:  58%|█████▊    | 519/897 [09:40<07:18,  1.16s/it]
Continual predict:  58%|█████▊    | 520/897 [09:41<06:23,  1.02s/it]
Continual predict:  58%|█████▊    | 521/897 [09:42<06:22,  1.02s/it]
Continual predict:  58%|█████▊    | 522/897 [09:42<05:41,  1.10it/s]
Continual predict:  58%|█████▊    | 523/897 [09:43<05:31,  1.13it/s]
Continual predict:  58%|█████▊    | 524/897 [09:44<05:08,  1.21it/s]
Continual predict:  59%|█████▊    | 525/897 [09:45<05:15,  1.18it/s]
Continual predict:  59%|█████▊    | 526/897 [09:46<05:27,  1.13it/s]
Continual predict:  59%|█████▉    | 527/897 [09:47<05:21,  1.15it/s]
Continual predict:  59%|█████▉    | 528/897 [09:48<06:29,  1.06s/it]
Continual predict:  59%|█████▉    | 529/897 [09:49<05:53,  1.04it/s]
Continual predict:  59%|█████▉    | 530/897 [09:50<06:19,  1.04s/it]
Continual predict:  59%|█████▉    | 531/897 [09:51<06:48,  1.12s/it]
Continual predict:  59%|█████▉    | 532/897 [09:53<07:29,  1.23s/it]
Continual predict:  59%|█████▉    | 533/897 [09:53<06:06,  1.01s/it]
Continual predict:  60%|█████▉    | 534/897 [09:54<05:46,  1.05it/s]
Continual predict:  60%|█████▉    | 535/897 [09:55<05:44,  1.05it/s]
Continual predict:  60%|█████▉    | 536/897 [09:55<04:40,  1.28it/s]
Continual predict:  60%|█████▉    | 537/897 [09:56<04:51,  1.23it/s]
Continual predict:  60%|█████▉    | 538/897 [09:57<04:52,  1.23it/s]
Continual predict:  60%|██████    | 539/897 [09:58<05:16,  1.13it/s]
Continual predict:  60%|██████    | 540/897 [10:00<06:00,  1.01s/it]
Continual predict:  60%|██████    | 541/897 [10:00<05:15,  1.13it/s]
Continual predict:  60%|██████    | 542/897 [10:01<04:39,  1.27it/s]
Continual predict:  61%|██████    | 543/897 [10:02<04:43,  1.25it/s]
Continual predict:  61%|██████    | 544/897 [10:03<05:24,  1.09it/s]
Continual predict:  61%|██████    | 545/897 [10:04<06:50,  1.17s/it]
Continual predict:  61%|██████    | 546/897 [10:05<06:13,  1.06s/it]
Continual predict:  61%|██████    | 547/897 [10:06<06:02,  1.04s/it]
Continual predict:  61%|██████    | 548/897 [10:07<05:49,  1.00s/it]
Continual predict:  61%|██████    | 549/897 [10:08<05:54,  1.02s/it]
Continual predict:  61%|██████▏   | 550/897 [10:09<06:04,  1.05s/it]
Continual predict:  61%|██████▏   | 551/897 [10:10<05:06,  1.13it/s]
Continual predict:  62%|██████▏   | 552/897 [10:10<04:41,  1.23it/s]
Continual predict:  62%|██████▏   | 553/897 [10:12<05:46,  1.01s/it]
Continual predict:  62%|██████▏   | 554/897 [10:13<06:06,  1.07s/it]
Continual predict:  62%|██████▏   | 555/897 [10:14<06:12,  1.09s/it]
Continual predict:  62%|██████▏   | 556/897 [10:15<05:29,  1.03it/s]
Continual predict:  62%|██████▏   | 557/897 [10:17<06:31,  1.15s/it]
Continual predict:  62%|██████▏   | 558/897 [10:17<05:29,  1.03it/s]
Continual predict:  62%|██████▏   | 559/897 [10:18<05:51,  1.04s/it]
Continual predict:  62%|██████▏   | 560/897 [10:20<06:05,  1.08s/it]
Continual predict:  63%|██████▎   | 561/897 [10:21<06:39,  1.19s/it]
Continual predict:  63%|██████▎   | 562/897 [10:22<06:57,  1.25s/it]
Continual predict:  63%|██████▎   | 563/897 [10:24<07:22,  1.32s/it]
Continual predict:  63%|██████▎   | 564/897 [10:25<07:28,  1.35s/it]
Continual predict:  63%|██████▎   | 565/897 [10:27<07:25,  1.34s/it]
Continual predict:  63%|██████▎   | 566/897 [10:27<06:32,  1.18s/it]
Continual predict:  63%|██████▎   | 567/897 [10:29<06:38,  1.21s/it]
Continual predict:  63%|██████▎   | 568/897 [10:29<05:59,  1.09s/it]
Continual predict:  63%|██████▎   | 569/897 [10:30<05:51,  1.07s/it]
Continual predict:  64%|██████▎   | 570/897 [10:31<04:49,  1.13it/s]
Continual predict:  64%|██████▎   | 571/897 [10:32<04:52,  1.11it/s]
Continual predict:  64%|██████▍   | 572/897 [10:32<04:18,  1.26it/s]
Continual predict:  64%|██████▍   | 573/897 [10:34<04:59,  1.08it/s]
Continual predict:  64%|██████▍   | 574/897 [10:34<04:34,  1.18it/s]
Continual predict:  64%|██████▍   | 575/897 [10:35<04:57,  1.08it/s]
Continual predict:  64%|██████▍   | 576/897 [10:36<04:25,  1.21it/s]
Continual predict:  64%|██████▍   | 577/897 [10:37<04:34,  1.16it/s]
Continual predict:  64%|██████▍   | 578/897 [10:37<04:01,  1.32it/s]
Continual predict:  65%|██████▍   | 579/897 [10:38<04:20,  1.22it/s]
Continual predict:  65%|██████▍   | 580/897 [10:40<05:14,  1.01it/s]
Continual predict:  65%|██████▍   | 581/897 [10:40<04:41,  1.12it/s]
Continual predict:  65%|██████▍   | 582/897 [10:41<03:30,  1.50it/s]
Continual predict:  65%|██████▍   | 583/897 [10:41<02:40,  1.95it/s]
Continual predict:  65%|██████▌   | 584/897 [10:41<02:06,  2.48it/s]
Continual predict:  65%|██████▌   | 585/897 [10:41<01:42,  3.06it/s]
Continual predict:  65%|██████▌   | 586/897 [10:41<01:24,  3.67it/s]
Continual predict:  65%|██████▌   | 587/897 [10:41<01:13,  4.24it/s]
Continual predict:  66%|██████▌   | 588/897 [10:42<01:04,  4.77it/s]
Continual predict:  66%|██████▌   | 589/897 [10:42<00:58,  5.22it/s]
Continual predict:  66%|██████▌   | 590/897 [10:42<00:54,  5.63it/s]
Continual predict:  66%|██████▌   | 591/897 [10:42<00:51,  5.92it/s]
Continual predict:  66%|██████▌   | 592/897 [10:42<00:49,  6.14it/s]
Continual predict:  66%|██████▌   | 593/897 [10:42<00:47,  6.34it/s]
Continual predict:  66%|██████▌   | 594/897 [10:42<00:46,  6.48it/s]
Continual predict:  66%|██████▋   | 595/897 [10:43<00:45,  6.59it/s]
Continual predict:  66%|██████▋   | 596/897 [10:43<00:45,  6.67it/s]
Continual predict:  67%|██████▋   | 597/897 [10:43<00:44,  6.73it/s]
Continual predict:  67%|██████▋   | 598/897 [10:43<00:44,  6.73it/s]
Continual predict:  67%|██████▋   | 599/897 [10:43<00:44,  6.72it/s]
Continual predict:  67%|██████▋   | 600/897 [10:43<00:43,  6.77it/s]
Continual predict:  67%|██████▋   | 601/897 [10:43<00:43,  6.76it/s]
Continual predict:  67%|██████▋   | 602/897 [10:44<00:43,  6.74it/s]
Continual predict:  67%|██████▋   | 603/897 [10:44<00:43,  6.74it/s]
Continual predict:  67%|██████▋   | 604/897 [10:44<00:43,  6.73it/s]
Continual predict:  67%|██████▋   | 605/897 [10:44<00:43,  6.72it/s]
Continual predict:  68%|██████▊   | 606/897 [10:44<00:43,  6.71it/s]
Continual predict:  68%|██████▊   | 607/897 [10:44<00:42,  6.76it/s]
Continual predict:  68%|██████▊   | 608/897 [10:44<00:42,  6.79it/s]
Continual predict:  68%|██████▊   | 609/897 [10:45<00:42,  6.81it/s]
Continual predict:  68%|██████▊   | 610/897 [10:45<00:42,  6.82it/s]
Continual predict:  68%|██████▊   | 611/897 [10:45<00:41,  6.83it/s]
Continual predict:  68%|██████▊   | 612/897 [10:45<00:41,  6.80it/s]
Continual predict:  68%|██████▊   | 613/897 [10:45<00:41,  6.78it/s]
Continual predict:  68%|██████▊   | 614/897 [10:45<00:41,  6.81it/s]
Continual predict:  69%|██████▊   | 615/897 [10:45<00:41,  6.79it/s]
Continual predict:  69%|██████▊   | 616/897 [10:46<00:41,  6.76it/s]
Continual predict:  69%|██████▉   | 617/897 [10:46<00:41,  6.78it/s]
Continual predict:  69%|██████▉   | 618/897 [10:46<00:41,  6.66it/s]
Continual predict:  69%|██████▉   | 619/897 [10:46<00:42,  6.55it/s]
Continual predict:  69%|██████▉   | 620/897 [10:46<00:42,  6.59it/s]
Continual predict:  69%|██████▉   | 621/897 [10:46<00:41,  6.63it/s]
Continual predict:  69%|██████▉   | 622/897 [10:47<00:41,  6.66it/s]
Continual predict:  69%|██████▉   | 623/897 [10:47<00:40,  6.72it/s]
Continual predict:  70%|██████▉   | 624/897 [10:47<00:40,  6.76it/s]
Continual predict:  70%|██████▉   | 625/897 [10:47<00:40,  6.75it/s]
Continual predict:  70%|██████▉   | 626/897 [10:47<00:40,  6.77it/s]
Continual predict:  70%|██████▉   | 627/897 [10:47<00:39,  6.76it/s]
Continual predict:  70%|███████   | 628/897 [10:47<00:39,  6.78it/s]
Continual predict:  70%|███████   | 629/897 [10:48<00:39,  6.81it/s]
Continual predict:  70%|███████   | 630/897 [10:48<00:39,  6.79it/s]
Continual predict:  70%|███████   | 631/897 [10:48<00:39,  6.77it/s]
Continual predict:  70%|███████   | 632/897 [10:48<00:39,  6.75it/s]
Continual predict:  71%|███████   | 633/897 [10:48<00:38,  6.79it/s]
Continual predict:  71%|███████   | 634/897 [10:48<00:38,  6.76it/s]
Continual predict:  71%|███████   | 635/897 [10:48<00:38,  6.79it/s]
Continual predict:  71%|███████   | 636/897 [10:49<00:38,  6.77it/s]
Continual predict:  71%|███████   | 637/897 [10:49<00:38,  6.76it/s]
Continual predict:  71%|███████   | 638/897 [10:49<00:38,  6.74it/s]
Continual predict:  71%|███████   | 639/897 [10:49<00:38,  6.74it/s]
Continual predict:  71%|███████▏  | 640/897 [10:49<00:37,  6.77it/s]
Continual predict:  71%|███████▏  | 641/897 [10:49<00:37,  6.80it/s]
Continual predict:  72%|███████▏  | 642/897 [10:50<00:37,  6.78it/s]
Continual predict:  72%|███████▏  | 643/897 [10:50<00:37,  6.76it/s]
Continual predict:  72%|███████▏  | 644/897 [10:50<00:37,  6.80it/s]
Continual predict:  72%|███████▏  | 645/897 [10:50<00:37,  6.78it/s]
Continual predict:  72%|███████▏  | 646/897 [10:50<00:37,  6.76it/s]
Continual predict:  72%|███████▏  | 647/897 [10:50<00:37,  6.75it/s]
Continual predict:  72%|███████▏  | 648/897 [10:50<00:36,  6.78it/s]
Continual predict:  72%|███████▏  | 649/897 [10:51<00:36,  6.77it/s]
Continual predict:  72%|███████▏  | 650/897 [10:51<00:36,  6.80it/s]
Continual predict:  73%|███████▎  | 651/897 [10:51<00:36,  6.82it/s]
Continual predict:  73%|███████▎  | 652/897 [10:51<00:35,  6.82it/s]
Continual predict:  73%|███████▎  | 653/897 [10:51<00:35,  6.84it/s]
Continual predict:  73%|███████▎  | 654/897 [10:51<00:35,  6.85it/s]
Continual predict:  73%|███████▎  | 655/897 [10:51<00:35,  6.85it/s]
Continual predict:  73%|███████▎  | 656/897 [10:52<00:35,  6.81it/s]
Continual predict:  73%|███████▎  | 657/897 [10:52<00:35,  6.79it/s]
Continual predict:  73%|███████▎  | 658/897 [10:52<00:35,  6.76it/s]
Continual predict:  73%|███████▎  | 659/897 [10:52<00:35,  6.80it/s]
Continual predict:  74%|███████▎  | 660/897 [10:52<00:34,  6.78it/s]
Continual predict:  74%|███████▎  | 661/897 [10:52<00:34,  6.76it/s]
Continual predict:  74%|███████▍  | 662/897 [10:52<00:34,  6.74it/s]
Continual predict:  74%|███████▍  | 663/897 [10:53<00:34,  6.74it/s]
Continual predict:  74%|███████▍  | 664/897 [10:53<00:34,  6.73it/s]
Continual predict:  74%|███████▍  | 665/897 [10:53<00:34,  6.73it/s]
Continual predict:  74%|███████▍  | 666/897 [10:53<00:34,  6.77it/s]
Continual predict:  74%|███████▍  | 667/897 [10:53<00:34,  6.76it/s]
Continual predict:  74%|███████▍  | 668/897 [10:53<00:33,  6.75it/s]
Continual predict:  75%|███████▍  | 669/897 [10:53<00:33,  6.78it/s]
Continual predict:  75%|███████▍  | 670/897 [10:54<00:33,  6.76it/s]
Continual predict:  75%|███████▍  | 671/897 [10:54<00:33,  6.75it/s]
Continual predict:  75%|███████▍  | 672/897 [10:54<00:33,  6.75it/s]
Continual predict:  75%|███████▌  | 673/897 [10:54<00:33,  6.77it/s]
Continual predict:  75%|███████▌  | 674/897 [10:54<00:33,  6.75it/s]
Continual predict:  75%|███████▌  | 675/897 [10:54<00:32,  6.78it/s]
Continual predict:  75%|███████▌  | 676/897 [10:55<00:32,  6.80it/s]
Continual predict:  75%|███████▌  | 677/897 [10:55<00:32,  6.82it/s]
Continual predict:  76%|███████▌  | 678/897 [10:55<00:32,  6.79it/s]
Continual predict:  76%|███████▌  | 679/897 [10:55<00:32,  6.78it/s]
Continual predict:  76%|███████▌  | 680/897 [10:55<00:32,  6.76it/s]
Continual predict:  76%|███████▌  | 681/897 [10:55<00:32,  6.74it/s]
Continual predict:  76%|███████▌  | 682/897 [10:56<01:35,  2.25it/s]
Continual predict:  76%|███████▌  | 683/897 [10:59<03:42,  1.04s/it]
Continual predict:  76%|███████▋  | 684/897 [11:00<04:20,  1.22s/it]
Continual predict:  76%|███████▋  | 685/897 [11:01<03:49,  1.08s/it]
Continual predict:  76%|███████▋  | 686/897 [11:04<05:07,  1.46s/it]
Continual predict:  77%|███████▋  | 687/897 [11:05<05:12,  1.49s/it]
Continual predict:  77%|███████▋  | 688/897 [11:10<08:23,  2.41s/it]
Continual predict:  77%|███████▋  | 689/897 [11:11<06:43,  1.94s/it]
Continual predict:  77%|███████▋  | 690/897 [11:14<08:35,  2.49s/it]
Continual predict:  77%|███████▋  | 691/897 [11:15<07:04,  2.06s/it]
Continual predict:  77%|███████▋  | 692/897 [11:16<05:37,  1.64s/it]
Continual predict:  77%|███████▋  | 693/897 [11:19<06:57,  2.05s/it]
Continual predict:  77%|███████▋  | 694/897 [11:20<05:41,  1.68s/it]
Continual predict:  77%|███████▋  | 695/897 [11:22<06:03,  1.80s/it]
Continual predict:  78%|███████▊  | 696/897 [11:23<05:30,  1.64s/it]
Continual predict:  78%|███████▊  | 697/897 [11:24<04:30,  1.35s/it]
Continual predict:  78%|███████▊  | 698/897 [11:26<04:59,  1.50s/it]
Continual predict:  78%|███████▊  | 699/897 [11:27<04:47,  1.45s/it]
Continual predict:  78%|███████▊  | 700/897 [11:28<04:45,  1.45s/it]
Continual predict:  78%|███████▊  | 701/897 [11:30<04:46,  1.46s/it]
Continual predict:  78%|███████▊  | 702/897 [11:31<04:23,  1.35s/it]
Continual predict:  78%|███████▊  | 703/897 [11:33<04:43,  1.46s/it]
Continual predict:  78%|███████▊  | 704/897 [11:34<04:21,  1.36s/it]
Continual predict:  79%|███████▊  | 705/897 [11:37<05:54,  1.85s/it]
Continual predict:  79%|███████▊  | 706/897 [11:39<05:40,  1.78s/it]
Continual predict:  79%|███████▉  | 707/897 [11:40<05:18,  1.67s/it]
Continual predict:  79%|███████▉  | 708/897 [11:43<06:27,  2.05s/it]
Continual predict:  79%|███████▉  | 709/897 [11:44<05:57,  1.90s/it]
Continual predict:  79%|███████▉  | 710/897 [11:46<05:40,  1.82s/it]
Continual predict:  79%|███████▉  | 711/897 [11:48<05:20,  1.72s/it]
Continual predict:  79%|███████▉  | 712/897 [11:49<05:05,  1.65s/it]
Continual predict:  79%|███████▉  | 713/897 [11:50<04:47,  1.56s/it]
Continual predict:  80%|███████▉  | 714/897 [11:52<05:05,  1.67s/it]
Continual predict:  80%|███████▉  | 715/897 [11:54<05:26,  1.79s/it]
Continual predict:  80%|███████▉  | 716/897 [11:56<05:07,  1.70s/it]
Continual predict:  80%|███████▉  | 717/897 [11:57<04:48,  1.60s/it]
Continual predict:  80%|████████  | 718/897 [11:59<04:44,  1.59s/it]
Continual predict:  80%|████████  | 719/897 [12:01<04:58,  1.67s/it]
Continual predict:  80%|████████  | 720/897 [12:02<04:30,  1.53s/it]
Continual predict:  80%|████████  | 721/897 [12:04<04:34,  1.56s/it]
Continual predict:  80%|████████  | 722/897 [12:05<04:25,  1.52s/it]
Continual predict:  81%|████████  | 723/897 [12:06<04:18,  1.48s/it]
Continual predict:  81%|████████  | 724/897 [12:08<04:20,  1.51s/it]
Continual predict:  81%|████████  | 725/897 [12:09<04:22,  1.52s/it]
Continual predict:  81%|████████  | 726/897 [12:11<04:23,  1.54s/it]
Continual predict:  81%|████████  | 727/897 [12:12<04:15,  1.50s/it]
Continual predict:  81%|████████  | 728/897 [12:14<04:32,  1.61s/it]
Continual predict:  81%|████████▏ | 729/897 [12:16<04:35,  1.64s/it]
Continual predict:  81%|████████▏ | 730/897 [12:18<04:40,  1.68s/it]
Continual predict:  81%|████████▏ | 731/897 [12:19<04:36,  1.67s/it]
Continual predict:  82%|████████▏ | 732/897 [12:21<04:22,  1.59s/it]
Continual predict:  82%|████████▏ | 733/897 [12:22<03:57,  1.45s/it]
Continual predict:  82%|████████▏ | 734/897 [12:24<04:31,  1.66s/it]
Continual predict:  82%|████████▏ | 735/897 [12:26<04:53,  1.81s/it]
Continual predict:  82%|████████▏ | 736/897 [12:28<04:57,  1.85s/it]
Continual predict:  82%|████████▏ | 737/897 [12:29<04:27,  1.67s/it]
Continual predict:  82%|████████▏ | 738/897 [12:31<04:28,  1.69s/it]
Continual predict:  82%|████████▏ | 739/897 [12:33<04:41,  1.78s/it]
Continual predict:  82%|████████▏ | 740/897 [12:35<04:36,  1.76s/it]
Continual predict:  83%|████████▎ | 741/897 [12:37<04:35,  1.77s/it]
Continual predict:  83%|████████▎ | 742/897 [12:38<04:28,  1.73s/it]
Continual predict:  83%|████████▎ | 743/897 [12:40<04:21,  1.70s/it]
Continual predict:  83%|████████▎ | 744/897 [12:42<04:23,  1.72s/it]
Continual predict:  83%|████████▎ | 745/897 [12:44<04:28,  1.77s/it]
Continual predict:  83%|████████▎ | 746/897 [12:45<04:24,  1.75s/it]
Continual predict:  83%|████████▎ | 747/897 [12:47<04:20,  1.74s/it]
Continual predict:  83%|████████▎ | 748/897 [12:49<04:27,  1.80s/it]
Continual predict:  84%|████████▎ | 749/897 [12:51<04:35,  1.86s/it]
Continual predict:  84%|████████▎ | 750/897 [12:52<04:15,  1.73s/it]
Continual predict:  84%|████████▎ | 751/897 [12:54<04:26,  1.83s/it]
Continual predict:  84%|████████▍ | 752/897 [12:56<04:00,  1.66s/it]
Continual predict:  84%|████████▍ | 753/897 [12:57<03:48,  1.58s/it]
Continual predict:  84%|████████▍ | 754/897 [12:58<03:29,  1.47s/it]
Continual predict:  84%|████████▍ | 755/897 [13:00<03:47,  1.61s/it]
Continual predict:  84%|████████▍ | 756/897 [13:02<03:41,  1.57s/it]
Continual predict:  84%|████████▍ | 757/897 [13:03<03:30,  1.50s/it]
Continual predict:  85%|████████▍ | 758/897 [13:05<03:25,  1.48s/it]
Continual predict:  85%|████████▍ | 759/897 [13:06<03:27,  1.50s/it]
Continual predict:  85%|████████▍ | 760/897 [13:07<03:22,  1.47s/it]
Continual predict:  85%|████████▍ | 761/897 [13:10<04:04,  1.79s/it]
Continual predict:  85%|████████▍ | 762/897 [13:12<04:23,  1.95s/it]
Continual predict:  85%|████████▌ | 763/897 [13:14<04:18,  1.93s/it]
Continual predict:  85%|████████▌ | 764/897 [13:16<04:10,  1.89s/it]
Continual predict:  85%|████████▌ | 765/897 [13:17<03:41,  1.68s/it]
Continual predict:  85%|████████▌ | 766/897 [13:19<03:40,  1.69s/it]
Continual predict:  86%|████████▌ | 767/897 [13:20<03:34,  1.65s/it]
Continual predict:  86%|████████▌ | 768/897 [13:22<03:25,  1.60s/it]
Continual predict:  86%|████████▌ | 769/897 [13:23<03:19,  1.56s/it]
Continual predict:  86%|████████▌ | 770/897 [13:25<03:18,  1.56s/it]
Continual predict:  86%|████████▌ | 771/897 [13:26<03:02,  1.45s/it]
Continual predict:  86%|████████▌ | 772/897 [13:27<02:51,  1.37s/it]
Continual predict:  86%|████████▌ | 773/897 [13:30<03:19,  1.61s/it]
Continual predict:  86%|████████▋ | 774/897 [13:31<03:15,  1.59s/it]
Continual predict:  86%|████████▋ | 775/897 [13:32<03:07,  1.54s/it]
Continual predict:  87%|████████▋ | 776/897 [13:34<03:18,  1.64s/it]
Continual predict:  87%|████████▋ | 777/897 [13:36<03:14,  1.62s/it]
Continual predict:  87%|████████▋ | 778/897 [13:38<03:15,  1.65s/it]
Continual predict:  87%|████████▋ | 779/897 [13:39<03:08,  1.60s/it]
Continual predict:  87%|████████▋ | 780/897 [13:41<03:03,  1.57s/it]
Continual predict:  87%|████████▋ | 781/897 [13:42<03:07,  1.62s/it]
Continual predict:  87%|████████▋ | 782/897 [13:44<02:56,  1.53s/it]
Continual predict:  87%|████████▋ | 783/897 [13:45<03:03,  1.61s/it]
Continual predict:  87%|████████▋ | 784/897 [13:47<03:02,  1.62s/it]
Continual predict:  88%|████████▊ | 785/897 [13:48<02:52,  1.54s/it]
Continual predict:  88%|████████▊ | 786/897 [13:50<02:54,  1.57s/it]
Continual predict:  88%|████████▊ | 787/897 [13:52<02:59,  1.63s/it]
Continual predict:  88%|████████▊ | 788/897 [13:53<02:42,  1.49s/it]
Continual predict:  88%|████████▊ | 789/897 [13:55<02:48,  1.56s/it]
Continual predict:  88%|████████▊ | 790/897 [13:57<03:11,  1.79s/it]
Continual predict:  88%|████████▊ | 791/897 [13:59<03:07,  1.76s/it]
Continual predict:  88%|████████▊ | 792/897 [14:00<02:51,  1.64s/it]
Continual predict:  88%|████████▊ | 793/897 [14:02<02:49,  1.63s/it]
Continual predict:  89%|████████▊ | 794/897 [14:04<03:06,  1.81s/it]
Continual predict:  89%|████████▊ | 795/897 [14:05<02:50,  1.67s/it]
Continual predict:  89%|████████▊ | 796/897 [14:07<02:46,  1.65s/it]
Continual predict:  89%|████████▉ | 797/897 [14:09<02:45,  1.65s/it]
Continual predict:  89%|████████▉ | 798/897 [14:11<02:54,  1.76s/it]
Continual predict:  89%|████████▉ | 799/897 [14:17<05:17,  3.24s/it]
Continual predict:  89%|████████▉ | 800/897 [14:19<04:44,  2.93s/it]
Continual predict:  89%|████████▉ | 801/897 [14:21<03:53,  2.43s/it]
Continual predict:  89%|████████▉ | 802/897 [14:22<03:20,  2.11s/it]
Continual predict:  90%|████████▉ | 803/897 [14:23<02:48,  1.79s/it]
Continual predict:  90%|████████▉ | 804/897 [14:27<03:42,  2.39s/it]
Continual predict:  90%|████████▉ | 805/897 [14:31<04:11,  2.73s/it]
Continual predict:  90%|████████▉ | 806/897 [14:33<03:50,  2.53s/it]
Continual predict:  90%|████████▉ | 807/897 [14:34<03:15,  2.18s/it]
Continual predict:  90%|█████████ | 808/897 [14:36<03:00,  2.02s/it]
Continual predict:  90%|█████████ | 809/897 [14:37<02:43,  1.86s/it]
Continual predict:  90%|█████████ | 810/897 [14:39<02:52,  1.99s/it]
Continual predict:  90%|█████████ | 811/897 [14:41<02:42,  1.89s/it]
Continual predict:  91%|█████████ | 812/897 [14:45<03:38,  2.57s/it]
Continual predict:  91%|█████████ | 813/897 [14:53<05:43,  4.09s/it]
Continual predict:  91%|█████████ | 814/897 [14:55<04:47,  3.47s/it]
Continual predict:  91%|█████████ | 815/897 [14:57<04:17,  3.14s/it]
Continual predict:  91%|█████████ | 816/897 [15:00<04:01,  2.99s/it]
Continual predict:  91%|█████████ | 817/897 [15:01<03:22,  2.54s/it]
Continual predict:  91%|█████████ | 818/897 [15:02<02:48,  2.13s/it]
Continual predict:  91%|█████████▏| 819/897 [15:05<02:47,  2.15s/it]
Continual predict:  91%|█████████▏| 820/897 [15:06<02:34,  2.01s/it]
Continual predict:  92%|█████████▏| 821/897 [15:08<02:30,  1.97s/it]
Continual predict:  92%|█████████▏| 822/897 [15:11<02:52,  2.30s/it]
Continual predict:  92%|█████████▏| 823/897 [15:15<03:17,  2.67s/it]
Continual predict:  92%|█████████▏| 824/897 [15:25<05:58,  4.91s/it]
Continual predict:  92%|█████████▏| 825/897 [15:27<04:44,  3.95s/it]
Continual predict:  92%|█████████▏| 826/897 [15:31<04:39,  3.94s/it]
Continual predict:  92%|█████████▏| 827/897 [15:32<03:36,  3.10s/it]
Continual predict:  92%|█████████▏| 828/897 [15:36<03:56,  3.43s/it]
Continual predict:  92%|█████████▏| 829/897 [15:41<04:33,  4.02s/it]
Continual predict:  93%|█████████▎| 830/897 [15:43<03:48,  3.41s/it]
Continual predict:  93%|█████████▎| 831/897 [15:55<06:26,  5.85s/it]
Continual predict:  93%|█████████▎| 832/897 [15:56<04:54,  4.53s/it]
Continual predict:  93%|█████████▎| 833/897 [15:57<03:40,  3.45s/it]
Continual predict:  93%|█████████▎| 834/897 [15:59<03:11,  3.03s/it]
Continual predict:  93%|█████████▎| 835/897 [16:01<02:36,  2.53s/it]
Continual predict:  93%|█████████▎| 836/897 [16:03<02:35,  2.55s/it]
Continual predict:  93%|█████████▎| 837/897 [16:08<03:14,  3.25s/it]
Continual predict:  93%|█████████▎| 838/897 [16:10<02:52,  2.92s/it]
Continual predict:  94%|█████████▎| 839/897 [16:17<03:55,  4.05s/it]
Continual predict:  94%|█████████▎| 840/897 [16:22<04:00,  4.22s/it]
Continual predict:  94%|█████████▍| 841/897 [16:23<03:11,  3.42s/it]
Continual predict:  94%|█████████▍| 842/897 [16:26<02:57,  3.23s/it]
Continual predict:  94%|█████████▍| 843/897 [16:28<02:32,  2.82s/it]
Continual predict:  94%|█████████▍| 844/897 [16:33<03:03,  3.46s/it]
Continual predict:  94%|█████████▍| 845/897 [16:37<03:06,  3.58s/it]
Continual predict:  94%|█████████▍| 846/897 [16:38<02:30,  2.95s/it]
Continual predict:  94%|█████████▍| 847/897 [16:39<01:53,  2.27s/it]
Continual predict:  95%|█████████▍| 848/897 [16:40<01:41,  2.07s/it]
Continual predict:  95%|█████████▍| 849/897 [16:42<01:26,  1.81s/it]
Continual predict:  95%|█████████▍| 850/897 [16:45<01:43,  2.21s/it]
Continual predict:  95%|█████████▍| 851/897 [16:47<01:38,  2.15s/it]
Continual predict:  95%|█████████▍| 852/897 [16:48<01:24,  1.88s/it]
Continual predict:  95%|█████████▌| 853/897 [16:57<03:03,  4.16s/it]
Continual predict:  95%|█████████▌| 854/897 [17:09<04:34,  6.38s/it]
Continual predict:  95%|█████████▌| 855/897 [17:11<03:36,  5.16s/it]
Continual predict:  95%|█████████▌| 856/897 [17:14<02:56,  4.30s/it]
Continual predict:  96%|█████████▌| 857/897 [17:15<02:21,  3.54s/it]
Continual predict:  96%|█████████▌| 858/897 [17:17<01:59,  3.07s/it]
Continual predict:  96%|█████████▌| 859/897 [17:22<02:09,  3.42s/it]
Continual predict:  96%|█████████▌| 860/897 [17:26<02:12,  3.59s/it]
Continual predict:  96%|█████████▌| 861/897 [17:37<03:35,  5.98s/it]
Continual predict:  96%|█████████▌| 862/897 [17:38<02:32,  4.37s/it]
Continual predict:  96%|█████████▌| 863/897 [17:40<02:04,  3.67s/it]
Continual predict:  96%|█████████▋| 864/897 [17:41<01:40,  3.04s/it]
Continual predict:  96%|█████████▋| 865/897 [17:44<01:29,  2.79s/it]
Continual predict:  97%|█████████▋| 866/897 [17:45<01:13,  2.38s/it]
Continual predict:  97%|█████████▋| 867/897 [17:47<01:11,  2.38s/it]
Continual predict:  97%|█████████▋| 868/897 [17:51<01:16,  2.63s/it]
Continual predict:  97%|█████████▋| 869/897 [17:54<01:21,  2.90s/it]
Continual predict:  97%|█████████▋| 870/897 [17:55<01:04,  2.39s/it]
Continual predict:  97%|█████████▋| 871/897 [17:57<00:55,  2.15s/it]
Continual predict:  97%|█████████▋| 872/897 [18:02<01:16,  3.04s/it]
Continual predict:  97%|█████████▋| 873/897 [18:04<01:07,  2.80s/it]
Continual predict:  97%|█████████▋| 874/897 [18:06<00:54,  2.37s/it]
Continual predict:  98%|█████████▊| 875/897 [18:08<00:49,  2.23s/it]
Continual predict:  98%|█████████▊| 876/897 [18:09<00:41,  1.95s/it]
Continual predict:  98%|█████████▊| 877/897 [18:11<00:42,  2.10s/it]
Continual predict:  98%|█████████▊| 878/897 [18:12<00:29,  1.54s/it]
Continual predict:  98%|█████████▊| 879/897 [18:15<00:35,  1.99s/it]
Continual predict:  98%|█████████▊| 880/897 [18:16<00:31,  1.88s/it]
Continual predict:  98%|█████████▊| 881/897 [18:20<00:37,  2.33s/it]
Continual predict:  98%|█████████▊| 882/897 [18:22<00:34,  2.32s/it]
Continual predict:  98%|█████████▊| 883/897 [18:23<00:29,  2.09s/it]
Continual predict:  99%|█████████▊| 884/897 [18:27<00:32,  2.51s/it]
Continual predict:  99%|█████████▊| 885/897 [18:32<00:38,  3.24s/it]
Continual predict:  99%|█████████▉| 886/897 [18:34<00:31,  2.89s/it]
Continual predict:  99%|█████████▉| 887/897 [18:37<00:29,  2.93s/it]
Continual predict:  99%|█████████▉| 888/897 [18:39<00:23,  2.64s/it]
Continual predict:  99%|█████████▉| 889/897 [18:41<00:19,  2.50s/it]
Continual predict:  99%|█████████▉| 890/897 [18:44<00:17,  2.53s/it]
Continual predict:  99%|█████████▉| 891/897 [18:46<00:15,  2.62s/it]
Continual predict:  99%|█████████▉| 892/897 [18:49<00:12,  2.60s/it]
Continual predict: 100%|█████████▉| 893/897 [18:52<00:10,  2.66s/it]
Continual predict: 100%|█████████▉| 894/897 [18:54<00:07,  2.39s/it]
Continual predict: 100%|█████████▉| 895/897 [18:58<00:05,  3.00s/it]
Continual predict: 100%|█████████▉| 896/897 [19:00<00:02,  2.66s/it]
Continual predict: 100%|██████████| 897/897 [19:04<00:00,  3.18s/it]
Continual predict: 100%|██████████| 897/897 [19:04<00:00,  1.28s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot11=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=16]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot10=100]│
+  │  GT=task511_reddit_tifu_long_text_su  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=18]│
+  │  GT=task748_glucose_reverse_cause_ev  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (897/897)  label_acc=100.0% (897/897)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                                 =       15.0
+  predict_eval_rougeL                                                   =    56.2843
+  predict_eval_rougeL_for_CL                                            =    56.2843
+  predict_eval_rougeL_for_task002_quoref_answer_generation              =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                   =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction        =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                       =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation            =    11.1335
+  predict_eval_rougeL_for_task1729_personachat_generate_next            =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                    =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification          =       91.0
+  predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization   =    21.3091
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation   =    15.2912
+  predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection =    66.9039
+  predict_exact_match                                                   =     39.019
+  predict_exact_match_for_CL                                            =     39.019
+  predict_exact_match_for_task002_quoref_answer_generation              =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                   =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction        =       98.0
+  predict_exact_match_for_task1572_samsum_summary                       =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation            =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next            =        0.0
+  predict_exact_match_for_task181_outcome_extraction                    =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification          =       91.0
+  predict_exact_match_for_task511_reddit_tifu_long_text_summarization   =        1.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation   =        0.0
+  predict_exact_match_for_task748_glucose_reverse_cause_event_detection =        3.0
+  predict_gen_len                                                       =    10.8517
+  predict_global_step                                                   =        480
+  predict_rouge1                                                        =     58.134
+  predict_rouge1_for_CL                                                 =     58.134
+  predict_rouge1_for_task002_quoref_answer_generation                   =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation            =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                        =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction             =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                            =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation                 =    13.7971
+  predict_rouge1_for_task1729_personachat_generate_next                 =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                         =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification               =       91.0
+  predict_rouge1_for_task511_reddit_tifu_long_text_summarization        =    25.5967
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation        =    15.2912
+  predict_rouge1_for_task748_glucose_reverse_cause_event_detection      =     68.907
+  predict_runtime                                                       = 0:19:04.77
+  predict_samples                                                       =        897
+  predict_samples_per_second                                            =      0.784
+  predict_steps_per_second                                              =      0.784
+[2026-04-30 04:41:12,282] [INFO] [launch.py:367:main] Process 1540635 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/all_results.json b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..9646a30693ba3bb511d3262f35b61241822a70ef
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/all_results.json
@@ -0,0 +1,60 @@
+{
+  "predict_exact_match": 41.1234,
+  "predict_rouge1": 59.6657,
+  "predict_eval_rougeL": 57.9347,
+  "predict_gen_len": 10.01,
+  "predict_runtime": 1184.3730858399067,
+  "predict_samples": 997,
+  "predict_samples_per_second": 0.842,
+  "predict_steps_per_second": 0.842,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 41.1234,
+  "predict_rouge1_for_CL": 59.6657,
+  "predict_eval_rougeL_for_CL": 57.9347,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "predict_exact_match_for_task748_glucose_reverse_cause_event_detection": 3.0,
+  "predict_rouge1_for_task748_glucose_reverse_cause_event_detection": 68.907,
+  "predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection": 66.9039,
+  "predict_exact_match_for_task511_reddit_tifu_long_text_summarization": 1.0,
+  "predict_rouge1_for_task511_reddit_tifu_long_text_summarization": 25.5967,
+  "predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization": 21.3091,
+  "predict_exact_match_for_task591_sciq_answer_generation": 60.0,
+  "predict_rouge1_for_task591_sciq_answer_generation": 73.4048,
+  "predict_eval_rougeL_for_task591_sciq_answer_generation": 72.7381,
+  "train_runtime": 2277.1504,
+  "train_samples_per_second": 6.587,
+  "train_steps_per_second": 0.211,
+  "total_flos": 2.99874549497856e+16,
+  "train_loss": 0.7412465025981267,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..dbd07797a0475d2843a4d699d4eb57aa202cd8b9
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a61f722a875867ce1ae5fef4d7a16e72d7effb30ca91d99857e583c5f5b606a
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..93809b405cc008481f1396fda90c9e9718020438
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:642f53cb126391025ad66319850c2bea519349433958f087a86f634cd46f1074
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..08537f9baecfbc3a1baf67c2fd76bf5a4cee1fa0
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11a209faef91a2c0f440b58847d0b0f68d6977d6efeeda9f0df784618d9c13b0
+size 2009159940
diff --git a/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/train.log b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..1e86fb9256ed6f18fbd37e72270ae451693de88c
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/train.log
@@ -0,0 +1,658 @@
+[2026-04-30 04:41:18,253] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 04:41:18,253] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29512 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task591_sciq_answer_generation --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation --cur_task_id 12 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+[2026-04-30 04:41:21,464] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 04:41:21,464] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 04:41:21,464] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 04:41:21,464] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 04:41:21,464] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 04:41:21,465] [INFO] [launch.py:272:main] process 1775507 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task591_sciq_answer_generation', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation', '--cur_task_id', '12', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights']
+
+============================================================
+Task 12: task591_sciq_answer_generation
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/997
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 121.37it/s]
Loading weights:  11%|█▏        | 33/291 [00:00<00:01, 157.00it/s]
Loading weights:  17%|█▋        | 50/291 [00:00<00:01, 161.90it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 170.94it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 172.18it/s]
Loading weights:  37%|███▋      | 108/291 [00:00<00:00, 184.27it/s]
Loading weights:  44%|████▎     | 127/291 [00:00<00:00, 180.14it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 180.06it/s]
Loading weights:  57%|█████▋    | 166/291 [00:00<00:00, 176.84it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 177.39it/s]
Loading weights:  70%|██████▉   | 203/291 [00:01<00:00, 180.83it/s]
Loading weights:  77%|███████▋  | 225/291 [00:01<00:00, 187.14it/s]
Loading weights:  84%|████████▍ | 244/291 [00:01<00:00, 182.11it/s]
Loading weights:  90%|█████████ | 263/291 [00:01<00:00, 180.34it/s]
Loading weights:  97%|█████████▋| 282/291 [00:01<00:00, 175.31it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 178.74it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1699)
+[FIX] Re-initialized lora_A/B in 832 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(12, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[9] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[10] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[11] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 12 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6766.551M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 12 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 12 sigs, cur=task591_sciq_answer_generation, mapping={'task591_sciq_answer_generation': 0, 'task1572_samsum_summary': 12, 'task363_sst2_polarity_classification': 11, 'task1290_xsum_summarization': 10, 'task181_outcome_extraction': 9, 'task002_quoref_answer_generation': 8, 'task1510_evalution_relation_extraction': 7, 'task639_multi_woz_user_utterance_generation': 6, 'task1729_personachat_generate_next': 5, 'task073_commonsenseqa_answer_generation': 4, 'task1590_diplomacy_text_generation': 3, 'task748_glucose_reverse_cause_event_detection': 2, 'task511_reddit_tifu_long_text_summarization': 1}
+[SGWI] Task 12: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 04:42:28,449] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:05<40:08,  5.03s/it]
  0%|          | 2/480 [00:09<37:46,  4.74s/it]
  1%|          | 3/480 [00:14<38:15,  4.81s/it]
  1%|          | 4/480 [00:18<37:16,  4.70s/it]
  1%|          | 5/480 [00:23<37:16,  4.71s/it]
  1%|▏         | 6/480 [00:28<37:26,  4.74s/it]
  1%|▏         | 7/480 [00:33<37:56,  4.81s/it]
  2%|▏         | 8/480 [00:38<37:28,  4.76s/it]
  2%|▏         | 9/480 [00:43<38:33,  4.91s/it]
  2%|▏         | 10/480 [00:48<38:19,  4.89s/it]
                                                
{'loss': '3.469', 'grad_norm': '1.088', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:48<38:19,  4.89s/it]
  2%|▏         | 11/480 [00:52<37:57,  4.86s/it]
  2%|▎         | 12/480 [00:58<38:26,  4.93s/it]
  3%|▎         | 13/480 [01:02<38:08,  4.90s/it]
  3%|▎         | 14/480 [01:07<38:25,  4.95s/it]
  3%|▎         | 15/480 [01:12<37:26,  4.83s/it]
  3%|▎         | 16/480 [01:17<36:45,  4.75s/it]
  4%|▎         | 17/480 [01:21<36:16,  4.70s/it]
  4%|▍         | 18/480 [01:26<35:50,  4.65s/it]
  4%|▍         | 19/480 [01:30<35:49,  4.66s/it]
  4%|▍         | 20/480 [01:35<36:12,  4.72s/it]
                                                
{'loss': '1.909', 'grad_norm': '1.01', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:35<36:12,  4.72s/it]
  4%|▍         | 21/480 [01:40<36:43,  4.80s/it]
  5%|▍         | 22/480 [01:45<36:27,  4.78s/it]
  5%|▍         | 23/480 [01:50<36:48,  4.83s/it]
  5%|▌         | 24/480 [01:55<36:34,  4.81s/it]
  5%|▌         | 25/480 [02:00<36:43,  4.84s/it]
  5%|▌         | 26/480 [02:04<36:14,  4.79s/it]
  6%|▌         | 27/480 [02:09<35:53,  4.75s/it]
  6%|▌         | 28/480 [02:14<35:51,  4.76s/it]
  6%|▌         | 29/480 [02:18<35:19,  4.70s/it]
  6%|▋         | 30/480 [02:23<35:41,  4.76s/it]
                                                
{'loss': '0.9977', 'grad_norm': '1.104', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [02:23<35:41,  4.76s/it]
  6%|▋         | 31/480 [02:28<35:10,  4.70s/it]
  7%|▋         | 32/480 [02:29<27:31,  3.69s/it]
  7%|▋         | 33/480 [02:34<29:57,  4.02s/it]
  7%|▋         | 34/480 [02:39<32:48,  4.41s/it]
  7%|▋         | 35/480 [02:44<34:09,  4.61s/it]
  8%|▊         | 36/480 [02:49<33:59,  4.59s/it]
  8%|▊         | 37/480 [02:54<34:33,  4.68s/it]
  8%|▊         | 38/480 [02:58<34:40,  4.71s/it]
  8%|▊         | 39/480 [03:03<35:14,  4.79s/it]
  8%|▊         | 40/480 [03:08<35:02,  4.78s/it]
                                                
{'loss': '0.8277', 'grad_norm': '0.9738', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [03:08<35:02,  4.78s/it]
  9%|▊         | 41/480 [03:13<34:38,  4.73s/it]
  9%|▉         | 42/480 [03:17<34:11,  4.68s/it]
  9%|▉         | 43/480 [03:22<33:53,  4.65s/it]
  9%|▉         | 44/480 [03:27<33:57,  4.67s/it]
  9%|▉         | 45/480 [03:31<33:39,  4.64s/it]
 10%|▉         | 46/480 [03:36<33:59,  4.70s/it]
 10%|▉         | 47/480 [03:41<33:37,  4.66s/it]
 10%|█         | 48/480 [03:46<34:08,  4.74s/it]
 10%|█         | 49/480 [03:50<33:40,  4.69s/it]
 10%|█         | 50/480 [03:55<33:26,  4.67s/it]
                                                
{'loss': '0.8098', 'grad_norm': '0.4847', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [03:55<33:26,  4.67s/it]
 11%|█         | 51/480 [04:00<33:51,  4.74s/it]
 11%|█         | 52/480 [04:04<33:27,  4.69s/it]
 11%|█         | 53/480 [04:09<33:54,  4.77s/it]
 11%|█▏        | 54/480 [04:14<33:25,  4.71s/it]
 11%|█▏        | 55/480 [04:19<33:32,  4.74s/it]
 12%|█▏        | 56/480 [04:23<33:32,  4.75s/it]
 12%|█▏        | 57/480 [04:28<33:12,  4.71s/it]
 12%|█▏        | 58/480 [04:33<33:12,  4.72s/it]
 12%|█▏        | 59/480 [04:38<33:11,  4.73s/it]
 12%|█▎        | 60/480 [04:42<33:12,  4.74s/it]
                                                
{'loss': '0.8664', 'grad_norm': '0.8862', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [04:42<33:12,  4.74s/it]
 13%|█▎        | 61/480 [04:47<33:11,  4.75s/it]
 13%|█▎        | 62/480 [04:52<33:06,  4.75s/it]
 13%|█▎        | 63/480 [04:57<33:02,  4.76s/it]
 13%|█▎        | 64/480 [04:58<25:54,  3.74s/it]
 14%|█▎        | 65/480 [05:03<27:55,  4.04s/it]
 14%|█▍        | 66/480 [05:07<29:20,  4.25s/it]
 14%|█▍        | 67/480 [05:12<30:54,  4.49s/it]
 14%|█▍        | 68/480 [05:17<31:24,  4.57s/it]
 14%|█▍        | 69/480 [05:22<31:36,  4.61s/it]
 15%|█▍        | 70/480 [05:27<31:47,  4.65s/it]
                                                
{'loss': '0.7825', 'grad_norm': '0.7467', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [05:27<31:47,  4.65s/it]
 15%|█▍        | 71/480 [05:31<31:47,  4.66s/it]
 15%|█▌        | 72/480 [05:36<31:48,  4.68s/it]
 15%|█▌        | 73/480 [05:41<32:01,  4.72s/it]
 15%|█▌        | 74/480 [05:45<31:41,  4.68s/it]
 16%|█▌        | 75/480 [05:50<31:24,  4.65s/it]
 16%|█▌        | 76/480 [05:55<31:47,  4.72s/it]
 16%|█▌        | 77/480 [06:00<31:29,  4.69s/it]
 16%|█▋        | 78/480 [06:04<31:34,  4.71s/it]
 16%|█▋        | 79/480 [06:09<31:14,  4.67s/it]
 17%|█▋        | 80/480 [06:14<31:40,  4.75s/it]
                                                
{'loss': '0.7889', 'grad_norm': '0.6817', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [06:14<31:40,  4.75s/it]
 17%|█▋        | 81/480 [06:18<31:17,  4.70s/it]
 17%|█▋        | 82/480 [06:23<31:21,  4.73s/it]
 17%|█▋        | 83/480 [06:28<31:18,  4.73s/it]
 18%|█▊        | 84/480 [06:33<31:04,  4.71s/it]
 18%|█▊        | 85/480 [06:37<30:57,  4.70s/it]
 18%|█▊        | 86/480 [06:42<31:04,  4.73s/it]
 18%|█▊        | 87/480 [06:47<30:49,  4.71s/it]
 18%|█▊        | 88/480 [06:52<30:50,  4.72s/it]
 19%|█▊        | 89/480 [06:56<31:05,  4.77s/it]
 19%|█▉        | 90/480 [07:01<30:42,  4.72s/it]
                                                
{'loss': '0.8173', 'grad_norm': '0.2828', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [07:01<30:42,  4.72s/it]
 19%|█▉        | 91/480 [07:06<31:05,  4.80s/it]
 19%|█▉        | 92/480 [07:11<30:56,  4.79s/it]
 19%|█▉        | 93/480 [07:16<31:08,  4.83s/it]
 20%|█▉        | 94/480 [07:21<31:22,  4.88s/it]
 20%|█▉        | 95/480 [07:26<31:23,  4.89s/it]
 20%|██        | 96/480 [07:27<24:20,  3.80s/it]
 20%|██        | 97/480 [07:31<25:46,  4.04s/it]
 20%|██        | 98/480 [07:36<26:46,  4.20s/it]
 21%|██        | 99/480 [07:41<27:41,  4.36s/it]
 21%|██        | 100/480 [07:45<28:03,  4.43s/it]
                                                 
{'loss': '0.7158', 'grad_norm': '0.4838', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [07:45<28:03,  4.43s/it]
 21%|██        | 101/480 [07:50<28:24,  4.50s/it]
 21%|██▏       | 102/480 [07:55<28:29,  4.52s/it]
 21%|██▏       | 103/480 [07:59<28:54,  4.60s/it]
 22%|██▏       | 104/480 [08:04<29:27,  4.70s/it]
 22%|██▏       | 105/480 [08:09<29:41,  4.75s/it]
 22%|██▏       | 106/480 [08:14<29:18,  4.70s/it]
 22%|██▏       | 107/480 [08:18<29:08,  4.69s/it]
 22%|██▎       | 108/480 [08:23<29:19,  4.73s/it]
 23%|██▎       | 109/480 [08:28<28:58,  4.69s/it]
 23%|██▎       | 110/480 [08:33<28:55,  4.69s/it]
                                                 
{'loss': '0.8055', 'grad_norm': '0.762', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [08:33<28:55,  4.69s/it]
 23%|██▎       | 111/480 [08:37<29:13,  4.75s/it]
 23%|██▎       | 112/480 [08:42<29:05,  4.74s/it]
 24%|██▎       | 113/480 [08:47<28:45,  4.70s/it]
 24%|██▍       | 114/480 [08:52<29:03,  4.76s/it]
 24%|██▍       | 115/480 [08:56<28:43,  4.72s/it]
 24%|██▍       | 116/480 [09:01<29:31,  4.87s/it]
 24%|██▍       | 117/480 [09:07<30:00,  4.96s/it]
 25%|██▍       | 118/480 [09:12<29:50,  4.95s/it]
 25%|██▍       | 119/480 [09:16<29:09,  4.85s/it]
 25%|██▌       | 120/480 [09:21<28:56,  4.82s/it]
                                                 
{'loss': '0.7325', 'grad_norm': '0.7996', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [09:21<28:56,  4.82s/it]
 25%|██▌       | 121/480 [09:26<29:00,  4.85s/it]
 25%|██▌       | 122/480 [09:31<28:46,  4.82s/it]
 26%|██▌       | 123/480 [09:36<29:21,  4.94s/it]
 26%|██▌       | 124/480 [09:40<28:46,  4.85s/it]
 26%|██▌       | 125/480 [09:45<28:38,  4.84s/it]
 26%|██▋       | 126/480 [09:50<28:06,  4.76s/it]
 26%|██▋       | 127/480 [09:55<27:54,  4.74s/it]
 27%|██▋       | 128/480 [09:56<21:30,  3.67s/it]
 27%|██▋       | 129/480 [10:00<23:03,  3.94s/it]
 27%|██▋       | 130/480 [10:05<24:22,  4.18s/it]
                                                 
{'loss': '0.7436', 'grad_norm': '2.43', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [10:05<24:22,  4.18s/it]
 27%|██▋       | 131/480 [10:10<25:01,  4.30s/it]
 28%|██▊       | 132/480 [10:14<25:51,  4.46s/it]
 28%|██▊       | 133/480 [10:19<26:08,  4.52s/it]
 28%|██▊       | 134/480 [10:24<26:12,  4.54s/it]
 28%|██▊       | 135/480 [10:28<26:21,  4.58s/it]
 28%|██▊       | 136/480 [10:33<26:35,  4.64s/it]
 29%|██▊       | 137/480 [10:38<26:37,  4.66s/it]
 29%|██▉       | 138/480 [10:42<26:26,  4.64s/it]
 29%|██▉       | 139/480 [10:47<27:00,  4.75s/it]
 29%|██▉       | 140/480 [10:52<26:40,  4.71s/it]
                                                 
{'loss': '0.7171', 'grad_norm': '1.062', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [10:52<26:40,  4.71s/it]
 29%|██▉       | 141/480 [10:57<26:33,  4.70s/it]
 30%|██▉       | 142/480 [11:02<26:39,  4.73s/it]
 30%|██▉       | 143/480 [11:06<26:46,  4.77s/it]
 30%|███       | 144/480 [11:11<26:57,  4.81s/it]
 30%|███       | 145/480 [11:16<26:49,  4.81s/it]
 30%|███       | 146/480 [11:21<27:08,  4.87s/it]
 31%|███       | 147/480 [11:26<26:35,  4.79s/it]
 31%|███       | 148/480 [11:31<26:33,  4.80s/it]
 31%|███       | 149/480 [11:35<26:25,  4.79s/it]
 31%|███▏      | 150/480 [11:40<26:24,  4.80s/it]
                                                 
{'loss': '0.7236', 'grad_norm': '0.2366', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [11:40<26:24,  4.80s/it]
 31%|███▏      | 151/480 [11:45<26:32,  4.84s/it]
 32%|███▏      | 152/480 [11:50<26:19,  4.82s/it]
 32%|███▏      | 153/480 [11:55<25:58,  4.77s/it]
 32%|███▏      | 154/480 [11:59<26:08,  4.81s/it]
 32%|███▏      | 155/480 [12:04<25:48,  4.77s/it]
 32%|███▎      | 156/480 [12:09<25:28,  4.72s/it]
 33%|███▎      | 157/480 [12:13<25:26,  4.73s/it]
 33%|███▎      | 158/480 [12:18<25:08,  4.68s/it]
 33%|███▎      | 159/480 [12:23<25:35,  4.78s/it]
 33%|███▎      | 160/480 [12:24<19:42,  3.70s/it]
                                                 
{'loss': '0.6651', 'grad_norm': '2.084', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [12:24<19:42,  3.70s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<00:13,  7.42it/s]
+
Eval @ step 160:   2%|▏         | 2/100 [00:00<00:14,  6.57it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:00<00:18,  5.13it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:00<00:15,  6.00it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:00<00:11,  8.19it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:00<00:12,  7.45it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:01<00:14,  6.47it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:01<00:12,  7.10it/s]
+
Eval @ step 160:  11%|█         | 11/100 [00:01<00:12,  7.33it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:01<00:12,  6.98it/s]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:01<00:11,  7.27it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:01<00:11,  7.48it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:02<00:10,  8.34it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:02<00:11,  7.13it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:02<00:11,  7.37it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:02<00:10,  7.57it/s]
+
Eval @ step 160:  20%|██        | 20/100 [00:02<00:11,  7.12it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:03<00:10,  7.54it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:03<00:10,  7.14it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:03<00:11,  6.86it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:03<00:10,  6.93it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:03<00:12,  5.96it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:04<00:12,  5.99it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:04<00:10,  6.38it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:04<00:10,  6.72it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:04<00:10,  6.55it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:04<00:10,  6.44it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:05<00:11,  5.94it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:05<00:10,  6.41it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:05<00:10,  5.90it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:05<00:09,  6.34it/s]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:05<00:10,  5.60it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:06<00:11,  5.12it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:06<00:09,  6.05it/s]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:06<00:08,  6.44it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:06<00:08,  6.36it/s]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:06<00:08,  6.30it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:06<00:08,  6.72it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:07<00:06,  8.06it/s]
+
Eval @ step 160:  50%|█████     | 50/100 [00:07<00:07,  6.49it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:07<00:07,  6.19it/s]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:07<00:08,  5.79it/s]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:08<00:06,  6.90it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:08<00:05,  7.71it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:08<00:05,  7.80it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:08<00:06,  6.88it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:08<00:05,  7.76it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:08<00:04,  7.84it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [00:09<00:04,  7.92it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [00:09<00:04,  8.55it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:09<00:04,  7.87it/s]
+
Eval @ step 160:  66%|██████▌   | 66/100 [00:09<00:04,  7.95it/s]
+
Eval @ step 160:  68%|██████▊   | 68/100 [00:09<00:03,  8.05it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:09<00:03,  8.08it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [00:10<00:04,  7.00it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:10<00:04,  5.91it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [00:10<00:05,  5.56it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:10<00:04,  5.69it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [00:10<00:04,  5.45it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:11<00:04,  5.63it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:11<00:04,  5.75it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:11<00:03,  7.07it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [00:11<00:03,  6.82it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:11<00:03,  6.63it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:11<00:02,  6.99it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [00:12<00:02,  6.73it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [00:12<00:02,  7.30it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:12<00:02,  7.47it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [00:12<00:01,  7.63it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [00:12<00:01,  8.41it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [00:12<00:01,  9.52it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [00:13<00:00,  9.00it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:13<00:00,  7.71it/s]
+
Eval @ step 160:  94%|█████████▍| 94/100 [00:13<00:00,  7.28it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:13<00:00,  7.47it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [00:13<00:00,  7.61it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:13<00:00,  7.75it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [00:14<00:00,  9.19it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [00:14<00:00,  8.50it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=67.2381 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '67.24', 'epoch': '5'}
+
 33%|███▎      | 160/480 [12:39<19:42,  3.70s/it]
 34%|███▎      | 161/480 [12:43<44:01,  8.28s/it]
 34%|███▍      | 162/480 [12:48<38:19,  7.23s/it]
 34%|███▍      | 163/480 [12:53<34:20,  6.50s/it]
 34%|███▍      | 164/480 [12:58<31:32,  5.99s/it]
 34%|███▍      | 165/480 [13:02<29:32,  5.63s/it]
 35%|███▍      | 166/480 [13:07<28:10,  5.39s/it]
 35%|███▍      | 167/480 [13:12<27:07,  5.20s/it]
 35%|███▌      | 168/480 [13:17<26:09,  5.03s/it]
 35%|███▌      | 169/480 [13:21<25:45,  4.97s/it]
 35%|███▌      | 170/480 [13:26<25:41,  4.97s/it]
                                                 
{'loss': '0.6204', 'grad_norm': '0.4879', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [13:26<25:41,  4.97s/it]
 36%|███▌      | 171/480 [13:31<25:00,  4.86s/it]
 36%|███▌      | 172/480 [13:36<24:46,  4.83s/it]
 36%|███▌      | 173/480 [13:40<24:34,  4.80s/it]
 36%|███▋      | 174/480 [13:45<24:34,  4.82s/it]
 36%|███▋      | 175/480 [13:50<24:39,  4.85s/it]
 37%|███▋      | 176/480 [13:55<24:21,  4.81s/it]
 37%|███▋      | 177/480 [14:00<24:17,  4.81s/it]
 37%|███▋      | 178/480 [14:04<23:52,  4.74s/it]
 37%|███▋      | 179/480 [14:09<24:06,  4.80s/it]
 38%|███▊      | 180/480 [14:14<23:41,  4.74s/it]
                                                 
{'loss': '0.7199', 'grad_norm': '0.5122', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [14:14<23:41,  4.74s/it]
 38%|███▊      | 181/480 [14:19<23:37,  4.74s/it]
 38%|███▊      | 182/480 [14:23<23:33,  4.74s/it]
 38%|███▊      | 183/480 [14:28<23:21,  4.72s/it]
 38%|███▊      | 184/480 [14:33<23:19,  4.73s/it]
 39%|███▊      | 185/480 [14:37<23:04,  4.69s/it]
 39%|███▉      | 186/480 [14:42<23:12,  4.73s/it]
 39%|███▉      | 187/480 [14:47<22:56,  4.70s/it]
 39%|███▉      | 188/480 [14:52<23:08,  4.76s/it]
 39%|███▉      | 189/480 [14:57<23:05,  4.76s/it]
 40%|███▉      | 190/480 [15:01<23:13,  4.81s/it]
                                                 
{'loss': '0.691', 'grad_norm': '0.6242', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [15:01<23:13,  4.81s/it]
 40%|███▉      | 191/480 [15:06<23:05,  4.79s/it]
 40%|████      | 192/480 [15:07<17:46,  3.70s/it]
 40%|████      | 193/480 [15:12<19:41,  4.12s/it]
 40%|████      | 194/480 [15:17<20:33,  4.31s/it]
 41%|████      | 195/480 [15:22<21:21,  4.50s/it]
 41%|████      | 196/480 [15:27<21:25,  4.53s/it]
 41%|████      | 197/480 [15:31<21:31,  4.57s/it]
 41%|████▏     | 198/480 [15:36<21:30,  4.58s/it]
 41%|████▏     | 199/480 [15:41<21:28,  4.58s/it]
 42%|████▏     | 200/480 [15:45<21:37,  4.64s/it]
                                                 
{'loss': '0.6376', 'grad_norm': '0.7815', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [15:45<21:37,  4.64s/it]
 42%|████▏     | 201/480 [15:50<21:29,  4.62s/it]
 42%|████▏     | 202/480 [15:55<21:43,  4.69s/it]
 42%|████▏     | 203/480 [15:59<21:30,  4.66s/it]
 42%|████▎     | 204/480 [16:04<21:42,  4.72s/it]
 43%|████▎     | 205/480 [16:09<21:26,  4.68s/it]
 43%|████▎     | 206/480 [16:14<21:22,  4.68s/it]
 43%|████▎     | 207/480 [16:18<21:35,  4.75s/it]
 43%|████▎     | 208/480 [16:23<21:18,  4.70s/it]
 44%|████▎     | 209/480 [16:28<21:14,  4.70s/it]
 44%|████▍     | 210/480 [16:32<21:14,  4.72s/it]
                                                 
{'loss': '0.6618', 'grad_norm': '0.5657', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [16:32<21:14,  4.72s/it]
 44%|████▍     | 211/480 [16:37<21:10,  4.72s/it]
 44%|████▍     | 212/480 [16:42<21:09,  4.74s/it]
 44%|████▍     | 213/480 [16:47<21:16,  4.78s/it]
 45%|████▍     | 214/480 [16:52<21:15,  4.79s/it]
 45%|████▍     | 215/480 [16:56<21:12,  4.80s/it]
 45%|████▌     | 216/480 [17:01<20:52,  4.74s/it]
 45%|████▌     | 217/480 [17:06<21:19,  4.87s/it]
 45%|████▌     | 218/480 [17:11<21:20,  4.89s/it]
 46%|████▌     | 219/480 [17:16<21:09,  4.86s/it]
 46%|████▌     | 220/480 [17:21<21:02,  4.86s/it]
                                                 
{'loss': '0.6278', 'grad_norm': '0.5837', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [17:21<21:02,  4.86s/it]
 46%|████▌     | 221/480 [17:26<21:06,  4.89s/it]
 46%|████▋     | 222/480 [17:31<21:06,  4.91s/it]
 46%|████▋     | 223/480 [17:35<20:37,  4.81s/it]
 47%|████▋     | 224/480 [17:37<15:52,  3.72s/it]
 47%|████▋     | 225/480 [17:41<17:03,  4.01s/it]
 47%|████▋     | 226/480 [17:46<17:46,  4.20s/it]
 47%|████▋     | 227/480 [17:51<18:25,  4.37s/it]
 48%|████▊     | 228/480 [17:55<18:37,  4.44s/it]
 48%|████▊     | 229/480 [18:00<19:08,  4.57s/it]
 48%|████▊     | 230/480 [18:05<19:21,  4.65s/it]
                                                 
{'loss': '0.6338', 'grad_norm': '0.5177', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [18:05<19:21,  4.65s/it]
 48%|████▊     | 231/480 [18:10<19:17,  4.65s/it]
 48%|████▊     | 232/480 [18:14<19:11,  4.64s/it]
 49%|████▊     | 233/480 [18:19<19:04,  4.63s/it]
 49%|████▉     | 234/480 [18:23<19:02,  4.64s/it]
 49%|████▉     | 235/480 [18:28<19:09,  4.69s/it]
 49%|████▉     | 236/480 [18:33<19:07,  4.70s/it]
 49%|████▉     | 237/480 [18:38<18:55,  4.67s/it]
 50%|████▉     | 238/480 [18:43<19:07,  4.74s/it]
 50%|████▉     | 239/480 [18:47<18:53,  4.70s/it]
 50%|█████     | 240/480 [18:52<19:02,  4.76s/it]
                                                 
{'loss': '0.6453', 'grad_norm': '0.7858', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [18:52<19:02,  4.76s/it]
 50%|█████     | 241/480 [18:57<18:46,  4.71s/it]
 50%|█████     | 242/480 [19:01<18:45,  4.73s/it]
 51%|█████     | 243/480 [19:06<18:53,  4.78s/it]
 51%|█████     | 244/480 [19:11<18:35,  4.73s/it]
 51%|█████     | 245/480 [19:16<18:53,  4.82s/it]
 51%|█████▏    | 246/480 [19:21<18:44,  4.81s/it]
 51%|█████▏    | 247/480 [19:26<18:41,  4.81s/it]
 52%|█████▏    | 248/480 [19:30<18:33,  4.80s/it]
 52%|█████▏    | 249/480 [19:35<18:40,  4.85s/it]
 52%|█████▏    | 250/480 [19:40<18:29,  4.83s/it]
                                                 
{'loss': '0.6061', 'grad_norm': '0.2554', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [19:40<18:29,  4.83s/it]
 52%|█████▏    | 251/480 [19:45<18:19,  4.80s/it]
 52%|█████▎    | 252/480 [19:49<18:00,  4.74s/it]
 53%|█████▎    | 253/480 [19:54<17:46,  4.70s/it]
 53%|█████▎    | 254/480 [19:59<17:46,  4.72s/it]
 53%|█████▎    | 255/480 [20:03<17:33,  4.68s/it]
 53%|█████▎    | 256/480 [20:05<13:31,  3.62s/it]
 54%|█████▎    | 257/480 [20:09<14:45,  3.97s/it]
 54%|█████▍    | 258/480 [20:14<15:35,  4.21s/it]
 54%|█████▍    | 259/480 [20:19<16:15,  4.41s/it]
 54%|█████▍    | 260/480 [20:24<16:26,  4.48s/it]
                                                 
{'loss': '0.6535', 'grad_norm': '0.2433', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [20:24<16:26,  4.48s/it]
 54%|█████▍    | 261/480 [20:28<16:38,  4.56s/it]
 55%|█████▍    | 262/480 [20:33<16:35,  4.57s/it]
 55%|█████▍    | 263/480 [20:38<17:11,  4.75s/it]
 55%|█████▌    | 264/480 [20:43<16:59,  4.72s/it]
 55%|█████▌    | 265/480 [20:47<16:50,  4.70s/it]
 55%|█████▌    | 266/480 [20:53<17:16,  4.84s/it]
 56%|█████▌    | 267/480 [20:57<16:55,  4.77s/it]
 56%|█████▌    | 268/480 [21:02<16:58,  4.81s/it]
 56%|█████▌    | 269/480 [21:07<16:40,  4.74s/it]
 56%|█████▋    | 270/480 [21:12<16:45,  4.79s/it]
                                                 
{'loss': '0.6347', 'grad_norm': '0.5501', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [21:12<16:45,  4.79s/it]
 56%|█████▋    | 271/480 [21:16<16:39,  4.78s/it]
 57%|█████▋    | 272/480 [21:21<16:42,  4.82s/it]
 57%|█████▋    | 273/480 [21:26<16:34,  4.81s/it]
 57%|█████▋    | 274/480 [21:31<16:33,  4.82s/it]
 57%|█████▋    | 275/480 [21:36<16:24,  4.80s/it]
 57%|█████▊    | 276/480 [21:41<16:31,  4.86s/it]
 58%|█████▊    | 277/480 [21:45<16:23,  4.84s/it]
 58%|█████▊    | 278/480 [21:50<16:13,  4.82s/it]
 58%|█████▊    | 279/480 [21:55<16:03,  4.79s/it]
 58%|█████▊    | 280/480 [22:00<15:46,  4.73s/it]
                                                 
{'loss': '0.6071', 'grad_norm': '0.5469', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [22:00<15:46,  4.73s/it]
 59%|█████▊    | 281/480 [22:04<15:39,  4.72s/it]
 59%|█████▉    | 282/480 [22:09<15:28,  4.69s/it]
 59%|█████▉    | 283/480 [22:14<15:24,  4.69s/it]
 59%|█████▉    | 284/480 [22:18<15:14,  4.66s/it]
 59%|█████▉    | 285/480 [22:23<15:26,  4.75s/it]
 60%|█████▉    | 286/480 [22:28<15:12,  4.70s/it]
 60%|█████▉    | 287/480 [22:32<15:03,  4.68s/it]
 60%|██████    | 288/480 [22:34<11:42,  3.66s/it]
 60%|██████    | 289/480 [22:38<12:33,  3.94s/it]
 60%|██████    | 290/480 [22:43<13:09,  4.16s/it]
                                                 
{'loss': '0.6353', 'grad_norm': '0.1383', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [22:43<13:09,  4.16s/it]
 61%|██████    | 291/480 [22:48<13:45,  4.37s/it]
 61%|██████    | 292/480 [22:52<13:54,  4.44s/it]
 61%|██████    | 293/480 [22:57<14:06,  4.52s/it]
 61%|██████▏   | 294/480 [23:02<14:15,  4.60s/it]
 61%|██████▏   | 295/480 [23:06<14:14,  4.62s/it]
 62%|██████▏   | 296/480 [23:11<14:17,  4.66s/it]
 62%|██████▏   | 297/480 [23:16<14:17,  4.69s/it]
 62%|██████▏   | 298/480 [23:21<14:39,  4.84s/it]
 62%|██████▏   | 299/480 [23:26<14:54,  4.94s/it]
 62%|██████▎   | 300/480 [23:31<14:31,  4.84s/it]
                                                 
{'loss': '0.5715', 'grad_norm': '1.394', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [23:31<14:31,  4.84s/it]
 63%|██████▎   | 301/480 [23:36<14:23,  4.82s/it]
 63%|██████▎   | 302/480 [23:41<14:20,  4.84s/it]
 63%|██████▎   | 303/480 [23:45<14:13,  4.82s/it]
 63%|██████▎   | 304/480 [23:50<14:12,  4.85s/it]
 64%|██████▎   | 305/480 [23:55<14:04,  4.82s/it]
 64%|██████▍   | 306/480 [24:00<13:59,  4.82s/it]
 64%|██████▍   | 307/480 [24:04<13:43,  4.76s/it]
 64%|██████▍   | 308/480 [24:09<13:33,  4.73s/it]
 64%|██████▍   | 309/480 [24:14<13:24,  4.70s/it]
 65%|██████▍   | 310/480 [24:19<13:20,  4.71s/it]
                                                 
{'loss': '0.582', 'grad_norm': '0.1823', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [24:19<13:20,  4.71s/it]
 65%|██████▍   | 311/480 [24:23<13:15,  4.71s/it]
 65%|██████▌   | 312/480 [24:28<13:05,  4.68s/it]
 65%|██████▌   | 313/480 [24:33<13:14,  4.76s/it]
 65%|██████▌   | 314/480 [24:37<13:03,  4.72s/it]
 66%|██████▌   | 315/480 [24:42<13:02,  4.74s/it]
 66%|██████▌   | 316/480 [24:47<12:52,  4.71s/it]
 66%|██████▌   | 317/480 [24:52<12:47,  4.71s/it]
 66%|██████▋   | 318/480 [24:56<12:47,  4.74s/it]
 66%|██████▋   | 319/480 [25:01<12:46,  4.76s/it]
 67%|██████▋   | 320/480 [25:02<09:49,  3.68s/it]
                                                 
{'loss': '0.589', 'grad_norm': '1.519', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [25:02<09:49,  3.68s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:00<00:12,  8.07it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:14,  6.76it/s]
+
Eval @ step 320:   3%|▎         | 3/100 [00:00<00:16,  5.73it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:00<00:17,  5.35it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:00<00:12,  7.59it/s]
+
Eval @ step 320:   7%|▋         | 7/100 [00:01<00:13,  7.10it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:01<00:13,  6.76it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:01<00:12,  7.29it/s]
+
Eval @ step 320:  11%|█         | 11/100 [00:01<00:12,  7.33it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:01<00:13,  6.69it/s]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:01<00:12,  7.04it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:02<00:11,  7.31it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:02<00:10,  8.24it/s]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:02<00:11,  7.08it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:02<00:11,  7.34it/s]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:02<00:11,  6.93it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:02<00:11,  6.70it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:03<00:10,  7.27it/s]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:03<00:11,  6.52it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:03<00:11,  6.42it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:03<00:10,  7.02it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:03<00:12,  6.02it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:04<00:11,  6.05it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:04<00:09,  7.20it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:04<00:09,  7.53it/s]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:04<00:09,  7.17it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:04<00:10,  6.48it/s]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:05<00:09,  6.84it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:05<00:10,  6.21it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:05<00:09,  6.53it/s]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:05<00:10,  6.05it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:05<00:10,  5.72it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:06<00:08,  6.53it/s]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:06<00:08,  6.84it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:06<00:09,  5.87it/s]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:06<00:09,  5.93it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:06<00:08,  6.40it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:07<00:06,  8.08it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:07<00:06,  8.11it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:07<00:07,  6.50it/s]
+
Eval @ step 320:  51%|█████     | 51/100 [00:07<00:07,  6.41it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:07<00:08,  5.53it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:08<00:06,  6.84it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:08<00:05,  7.74it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:08<00:05,  7.47it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:08<00:05,  7.40it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:08<00:05,  7.29it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:09<00:05,  6.46it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:09<00:04,  7.43it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [00:09<00:04,  7.11it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:09<00:04,  7.34it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:09<00:04,  7.66it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [00:10<00:04,  7.24it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:10<00:04,  6.51it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [00:10<00:05,  5.67it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:10<00:05,  5.45it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [00:10<00:04,  5.63it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:11<00:04,  5.40it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [00:11<00:04,  5.59it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:11<00:04,  5.73it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:11<00:03,  7.07it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [00:11<00:03,  6.38it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:11<00:02,  6.77it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:12<00:02,  7.28it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:12<00:02,  7.60it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [00:12<00:02,  7.23it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:12<00:01,  7.42it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:13<00:01,  6.80it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:13<00:01,  8.10it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:13<00:00,  8.12it/s]
+
Eval @ step 320:  93%|█████████▎| 93/100 [00:13<00:00,  7.21it/s]
+
Eval @ step 320:  95%|█████████▌| 95/100 [00:13<00:00,  7.94it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:13<00:00,  7.99it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:14<00:00,  9.11it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:14<00:00,  9.34it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=72.7381 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '72.74', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [25:17<09:49,  3.68s/it]
 67%|██████▋   | 321/480 [25:21<22:04,  8.33s/it]
 67%|██████▋   | 322/480 [25:26<19:15,  7.31s/it]
 67%|██████▋   | 323/480 [25:31<17:08,  6.55s/it]
 68%|██████▊   | 324/480 [25:36<15:45,  6.06s/it]
 68%|██████▊   | 325/480 [25:41<14:32,  5.63s/it]
 68%|██████▊   | 326/480 [25:45<13:45,  5.36s/it]
 68%|██████▊   | 327/480 [25:50<13:05,  5.13s/it]
 68%|██████▊   | 328/480 [25:55<12:39,  5.00s/it]
 69%|██████▊   | 329/480 [26:00<12:25,  4.94s/it]
 69%|██████▉   | 330/480 [26:04<12:12,  4.88s/it]
                                                 
{'loss': '0.63', 'grad_norm': '0.214', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [26:04<12:12,  4.88s/it]
 69%|██████▉   | 331/480 [26:09<12:02,  4.85s/it]
 69%|██████▉   | 332/480 [26:14<11:46,  4.77s/it]
 69%|██████▉   | 333/480 [26:18<11:40,  4.77s/it]
 70%|██████▉   | 334/480 [26:23<11:36,  4.77s/it]
 70%|██████▉   | 335/480 [26:28<11:32,  4.78s/it]
 70%|███████   | 336/480 [26:33<11:27,  4.77s/it]
 70%|███████   | 337/480 [26:38<11:26,  4.80s/it]
 70%|███████   | 338/480 [26:42<11:12,  4.74s/it]
 71%|███████   | 339/480 [26:47<11:17,  4.80s/it]
 71%|███████   | 340/480 [26:52<11:04,  4.74s/it]
                                                 
{'loss': '0.5655', 'grad_norm': '1.048', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [26:52<11:04,  4.74s/it]
 71%|███████   | 341/480 [26:57<11:04,  4.78s/it]
 71%|███████▏  | 342/480 [27:01<10:57,  4.76s/it]
 71%|███████▏  | 343/480 [27:06<10:52,  4.76s/it]
 72%|███████▏  | 344/480 [27:11<10:52,  4.80s/it]
 72%|███████▏  | 345/480 [27:16<10:47,  4.80s/it]
 72%|███████▏  | 346/480 [27:21<10:49,  4.84s/it]
 72%|███████▏  | 347/480 [27:26<10:50,  4.89s/it]
 72%|███████▎  | 348/480 [27:31<10:48,  4.91s/it]
 73%|███████▎  | 349/480 [27:35<10:38,  4.87s/it]
 73%|███████▎  | 350/480 [27:40<10:35,  4.89s/it]
                                                 
{'loss': '0.5094', 'grad_norm': '0.1806', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [27:40<10:35,  4.89s/it]
 73%|███████▎  | 351/480 [27:45<10:26,  4.86s/it]
 73%|███████▎  | 352/480 [27:46<07:59,  3.75s/it]
 74%|███████▎  | 353/480 [27:51<08:41,  4.10s/it]
 74%|███████▍  | 354/480 [27:56<08:55,  4.25s/it]
 74%|███████▍  | 355/480 [28:01<09:06,  4.37s/it]
 74%|███████▍  | 356/480 [28:05<09:10,  4.44s/it]
 74%|███████▍  | 357/480 [28:10<09:12,  4.49s/it]
 75%|███████▍  | 358/480 [28:14<09:14,  4.55s/it]
 75%|███████▍  | 359/480 [28:19<09:17,  4.61s/it]
 75%|███████▌  | 360/480 [28:24<09:27,  4.73s/it]
                                                 
{'loss': '0.5753', 'grad_norm': '0.4284', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [28:24<09:27,  4.73s/it]
 75%|███████▌  | 361/480 [28:29<09:21,  4.72s/it]
 75%|███████▌  | 362/480 [28:34<09:26,  4.80s/it]
 76%|███████▌  | 363/480 [28:39<09:15,  4.74s/it]
 76%|███████▌  | 364/480 [28:43<09:14,  4.78s/it]
 76%|███████▌  | 365/480 [28:48<09:03,  4.72s/it]
 76%|███████▋  | 366/480 [28:53<08:55,  4.70s/it]
 76%|███████▋  | 367/480 [28:58<09:03,  4.81s/it]
 77%|███████▋  | 368/480 [29:02<08:51,  4.75s/it]
 77%|███████▋  | 369/480 [29:07<08:50,  4.78s/it]
 77%|███████▋  | 370/480 [29:12<08:39,  4.72s/it]
                                                 
{'loss': '0.6023', 'grad_norm': '0.6993', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [29:12<08:39,  4.72s/it]
 77%|███████▋  | 371/480 [29:17<08:41,  4.78s/it]
 78%|███████▊  | 372/480 [29:21<08:35,  4.78s/it]
 78%|███████▊  | 373/480 [29:26<08:31,  4.78s/it]
 78%|███████▊  | 374/480 [29:31<08:27,  4.78s/it]
 78%|███████▊  | 375/480 [29:36<08:27,  4.83s/it]
 78%|███████▊  | 376/480 [29:41<08:21,  4.82s/it]
 79%|███████▊  | 377/480 [29:46<08:21,  4.87s/it]
 79%|███████▉  | 378/480 [29:50<08:14,  4.84s/it]
 79%|███████▉  | 379/480 [29:55<08:06,  4.82s/it]
 79%|███████▉  | 380/480 [30:00<07:57,  4.77s/it]
                                                 
{'loss': '0.6073', 'grad_norm': '0.732', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [30:00<07:57,  4.77s/it]
 79%|███████▉  | 381/480 [30:05<07:47,  4.72s/it]
 80%|███████▉  | 382/480 [30:09<07:43,  4.73s/it]
 80%|███████▉  | 383/480 [30:14<07:34,  4.69s/it]
 80%|████████  | 384/480 [30:15<05:48,  3.63s/it]
 80%|████████  | 385/480 [30:20<06:15,  3.96s/it]
 80%|████████  | 386/480 [30:24<06:29,  4.15s/it]
 81%|████████  | 387/480 [30:29<06:45,  4.36s/it]
 81%|████████  | 388/480 [30:34<06:48,  4.44s/it]
 81%|████████  | 389/480 [30:39<06:51,  4.52s/it]
 81%|████████▏ | 390/480 [30:43<06:49,  4.54s/it]
                                                 
{'loss': '0.5714', 'grad_norm': '1.364', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [30:43<06:49,  4.54s/it]
 81%|████████▏ | 391/480 [30:48<06:45,  4.56s/it]
 82%|████████▏ | 392/480 [30:53<06:49,  4.65s/it]
 82%|████████▏ | 393/480 [30:57<06:43,  4.64s/it]
 82%|████████▏ | 394/480 [31:02<06:40,  4.65s/it]
 82%|████████▏ | 395/480 [31:07<06:41,  4.72s/it]
 82%|████████▎ | 396/480 [31:12<06:43,  4.81s/it]
 83%|████████▎ | 397/480 [31:17<06:42,  4.84s/it]
 83%|████████▎ | 398/480 [31:22<06:38,  4.86s/it]
 83%|████████▎ | 399/480 [31:26<06:31,  4.83s/it]
 83%|████████▎ | 400/480 [31:31<06:31,  4.90s/it]
                                                 
{'loss': '0.5897', 'grad_norm': '0.5642', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [31:31<06:31,  4.90s/it]
 84%|████████▎ | 401/480 [31:36<06:21,  4.82s/it]
 84%|████████▍ | 402/480 [31:41<06:15,  4.81s/it]
 84%|████████▍ | 403/480 [31:46<06:12,  4.83s/it]
 84%|████████▍ | 404/480 [31:50<06:05,  4.81s/it]
 84%|████████▍ | 405/480 [31:55<06:05,  4.87s/it]
 85%|████████▍ | 406/480 [32:00<05:58,  4.84s/it]
 85%|████████▍ | 407/480 [32:05<05:54,  4.86s/it]
 85%|████████▌ | 408/480 [32:10<05:44,  4.78s/it]
 85%|████████▌ | 409/480 [32:14<05:37,  4.76s/it]
 85%|████████▌ | 410/480 [32:19<05:29,  4.71s/it]
                                                 
{'loss': '0.5529', 'grad_norm': '1.175', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [32:19<05:29,  4.71s/it]
 86%|████████▌ | 411/480 [32:24<05:22,  4.68s/it]
 86%|████████▌ | 412/480 [32:28<05:17,  4.68s/it]
 86%|████████▌ | 413/480 [32:33<05:12,  4.66s/it]
 86%|████████▋ | 414/480 [32:38<05:13,  4.75s/it]
 86%|████████▋ | 415/480 [32:43<05:05,  4.70s/it]
 87%|████████▋ | 416/480 [32:44<03:52,  3.64s/it]
 87%|████████▋ | 417/480 [32:48<04:09,  3.96s/it]
 87%|████████▋ | 418/480 [32:53<04:17,  4.15s/it]
 87%|████████▋ | 419/480 [32:58<04:28,  4.40s/it]
 88%|████████▊ | 420/480 [33:03<04:30,  4.51s/it]
                                                 
{'loss': '0.6617', 'grad_norm': '1.542', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [33:03<04:30,  4.51s/it]
 88%|████████▊ | 421/480 [33:07<04:29,  4.57s/it]
 88%|████████▊ | 422/480 [33:12<04:25,  4.58s/it]
 88%|████████▊ | 423/480 [33:17<04:25,  4.66s/it]
 88%|████████▊ | 424/480 [33:21<04:20,  4.64s/it]
 89%|████████▊ | 425/480 [33:26<04:17,  4.68s/it]
 89%|████████▉ | 426/480 [33:31<04:13,  4.70s/it]
 89%|████████▉ | 427/480 [33:36<04:13,  4.78s/it]
 89%|████████▉ | 428/480 [33:41<04:11,  4.83s/it]
 89%|████████▉ | 429/480 [33:45<04:02,  4.76s/it]
 90%|████████▉ | 430/480 [33:50<04:00,  4.80s/it]
                                                 
{'loss': '0.5766', 'grad_norm': '0.3929', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [33:50<04:00,  4.80s/it]
 90%|████████▉ | 431/480 [33:55<03:55,  4.80s/it]
 90%|█████████ | 432/480 [34:00<03:52,  4.83s/it]
 90%|█████████ | 433/480 [34:05<03:48,  4.85s/it]
 90%|█████████ | 434/480 [34:10<03:48,  4.96s/it]
 91%|█████████ | 435/480 [34:15<03:40,  4.91s/it]
 91%|█████████ | 436/480 [34:20<03:32,  4.82s/it]
 91%|█████████ | 437/480 [34:24<03:25,  4.77s/it]
 91%|█████████▏| 438/480 [34:29<03:18,  4.72s/it]
 91%|█████████▏| 439/480 [34:34<03:14,  4.73s/it]
 92%|█████████▏| 440/480 [34:38<03:07,  4.69s/it]
                                                 
{'loss': '0.5296', 'grad_norm': '1.633', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [34:38<03:07,  4.69s/it]
 92%|█████████▏| 441/480 [34:43<03:03,  4.71s/it]
 92%|█████████▏| 442/480 [34:48<02:59,  4.72s/it]
 92%|█████████▏| 443/480 [34:52<02:54,  4.71s/it]
 92%|█████████▎| 444/480 [34:57<02:48,  4.68s/it]
 93%|█████████▎| 445/480 [35:02<02:43,  4.66s/it]
 93%|█████████▎| 446/480 [35:06<02:39,  4.68s/it]
 93%|█████████▎| 447/480 [35:11<02:36,  4.73s/it]
 93%|█████████▎| 448/480 [35:12<01:57,  3.66s/it]
 94%|█████████▎| 449/480 [35:17<02:03,  3.98s/it]
 94%|█████████▍| 450/480 [35:22<02:04,  4.17s/it]
                                                 
{'loss': '0.5448', 'grad_norm': '0.4373', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [35:22<02:04,  4.17s/it]
 94%|█████████▍| 451/480 [35:27<02:08,  4.43s/it]
 94%|█████████▍| 452/480 [35:31<02:06,  4.51s/it]
 94%|█████████▍| 453/480 [35:37<02:08,  4.75s/it]
 95%|█████████▍| 454/480 [35:41<02:02,  4.72s/it]
 95%|█████████▍| 455/480 [35:46<01:59,  4.78s/it]
 95%|█████████▌| 456/480 [35:51<01:54,  4.78s/it]
 95%|█████████▌| 457/480 [35:56<01:49,  4.75s/it]
 95%|█████████▌| 458/480 [36:01<01:44,  4.76s/it]
 96%|█████████▌| 459/480 [36:05<01:40,  4.78s/it]
 96%|█████████▌| 460/480 [36:10<01:35,  4.78s/it]
                                                 
{'loss': '0.5624', 'grad_norm': '0.7646', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [36:10<01:35,  4.78s/it]
 96%|█████████▌| 461/480 [36:15<01:30,  4.78s/it]
 96%|█████████▋| 462/480 [36:20<01:26,  4.81s/it]
 96%|█████████▋| 463/480 [36:25<01:21,  4.80s/it]
 97%|█████████▋| 464/480 [36:29<01:16,  4.78s/it]
 97%|█████████▋| 465/480 [36:34<01:10,  4.73s/it]
 97%|█████████▋| 466/480 [36:39<01:08,  4.87s/it]
 97%|█████████▋| 467/480 [36:44<01:03,  4.88s/it]
 98%|█████████▊| 468/480 [36:49<00:58,  4.85s/it]
 98%|█████████▊| 469/480 [36:53<00:52,  4.78s/it]
 98%|█████████▊| 470/480 [36:59<00:49,  4.96s/it]
                                                 
{'loss': '0.4922', 'grad_norm': '0.3151', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [36:59<00:49,  4.96s/it]
 98%|█████████▊| 471/480 [37:03<00:43,  4.86s/it]
 98%|█████████▊| 472/480 [37:08<00:38,  4.78s/it]
 99%|█████████▊| 473/480 [37:13<00:33,  4.75s/it]
 99%|█████████▉| 474/480 [37:17<00:28,  4.71s/it]
 99%|█████████▉| 475/480 [37:22<00:23,  4.78s/it]
 99%|█████████▉| 476/480 [37:27<00:18,  4.72s/it]
 99%|█████████▉| 477/480 [37:32<00:14,  4.71s/it]
100%|█████████▉| 478/480 [37:36<00:09,  4.74s/it]
100%|█████████▉| 479/480 [37:41<00:04,  4.74s/it]
100%|██████████| 480/480 [37:42<00:00,  3.67s/it]
                                                 
{'loss': '0.5206', 'grad_norm': '11.55', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [37:42<00:00,  3.67s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:00<00:12,  7.79it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:15,  6.50it/s]
+
Eval @ step 480:   3%|▎         | 3/100 [00:00<00:17,  5.58it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:00<00:18,  5.28it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:00<00:12,  7.53it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:01<00:13,  7.07it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:01<00:14,  6.27it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:01<00:12,  6.96it/s]
+
Eval @ step 480:  11%|█         | 11/100 [00:01<00:12,  7.20it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:01<00:12,  6.87it/s]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:01<00:12,  7.16it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:02<00:11,  7.39it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:02<00:10,  8.29it/s]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:02<00:11,  7.08it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:02<00:11,  7.32it/s]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:02<00:11,  6.94it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:02<00:11,  6.69it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:03<00:10,  7.27it/s]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:03<00:11,  6.53it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:03<00:11,  6.40it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:03<00:10,  7.05it/s]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:03<00:12,  6.05it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:04<00:11,  6.07it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:04<00:09,  7.21it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:04<00:09,  7.50it/s]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:04<00:09,  7.16it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:04<00:10,  6.48it/s]
+
Eval @ step 480:  35%|███▌      | 35/100 [00:05<00:09,  6.82it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:05<00:09,  6.62it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:05<00:09,  6.75it/s]
+
Eval @ step 480:  39%|███▉      | 39/100 [00:05<00:09,  6.20it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:05<00:10,  5.83it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:06<00:08,  6.62it/s]
+
Eval @ step 480:  43%|████▎     | 43/100 [00:06<00:09,  5.80it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:06<00:09,  5.87it/s]
+
Eval @ step 480:  45%|████▌     | 45/100 [00:06<00:09,  5.93it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:06<00:08,  6.40it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:07<00:06,  7.54it/s]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:07<00:06,  7.68it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:07<00:07,  6.29it/s]
+
Eval @ step 480:  51%|█████     | 51/100 [00:07<00:07,  6.24it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:07<00:08,  5.44it/s]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:08<00:07,  6.01it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:08<00:07,  6.51it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:08<00:05,  7.64it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:08<00:05,  7.03it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:08<00:05,  7.65it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [00:08<00:05,  7.74it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [00:09<00:04,  8.93it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:09<00:04,  8.75it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [00:09<00:04,  7.97it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:09<00:04,  8.02it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:09<00:03,  8.03it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [00:09<00:04,  7.50it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:10<00:04,  6.63it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [00:10<00:05,  5.71it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:10<00:05,  5.47it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [00:10<00:04,  5.62it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:10<00:04,  5.40it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [00:11<00:04,  5.60it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:11<00:04,  5.75it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:11<00:03,  7.08it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [00:11<00:03,  6.38it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:11<00:02,  6.74it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:12<00:02,  7.28it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:12<00:02,  7.60it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [00:12<00:02,  7.23it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:12<00:01,  7.42it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:12<00:01,  6.80it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:13<00:01,  8.10it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:13<00:00,  8.12it/s]
+
Eval @ step 480:  93%|█████████▎| 93/100 [00:13<00:00,  7.21it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:13<00:00,  6.94it/s]
+
Eval @ step 480:  95%|█████████▌| 95/100 [00:13<00:00,  7.20it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:13<00:00,  7.40it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:14<00:00,  8.86it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:14<00:00,  9.19it/s]
+
                                                                    [EVAL] eval_rougeL=67.4048 (best=72.7381)
+
                                                 
{'eval_rougeL': '67.4', 'epoch': '15'}
+
100%|██████████| 480/480 [37:57<00:00,  3.67s/it]
                                                 
{'train_runtime': '2277', 'train_samples_per_second': '6.587', 'train_steps_per_second': '0.211', 'train_loss': '0.7412', 'epoch': '15'}
+
100%|██████████| 480/480 [37:57<00:00,  3.67s/it]
100%|██████████| 480/480 [37:57<00:00,  4.74s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 27927993GF
+  train_loss               =     0.7412
+  train_runtime            = 0:37:57.15
+  train_samples            =       1000
+  train_samples_per_second =      6.587
+  train_steps_per_second   =      0.211
+  [EVAL] Restored best model (eval_rougeL=72.7381, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 11/291 [00:00<00:02, 100.37it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 141.77it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 159.34it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 165.47it/s]
Loading weights:  29%|██▉       | 85/291 [00:00<00:01, 169.13it/s]
Loading weights:  36%|███▌      | 104/291 [00:00<00:01, 171.30it/s]
Loading weights:  42%|████▏     | 122/291 [00:00<00:01, 167.75it/s]
Loading weights:  48%|████▊     | 140/291 [00:00<00:00, 166.81it/s]
Loading weights:  57%|█████▋    | 165/291 [00:00<00:00, 188.17it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 175.58it/s]
Loading weights:  70%|███████   | 204/291 [00:01<00:00, 180.71it/s]
Loading weights:  77%|███████▋  | 225/291 [00:01<00:00, 186.47it/s]
Loading weights:  84%|████████▍ | 244/291 [00:01<00:00, 181.35it/s]
Loading weights:  90%|█████████ | 263/291 [00:01<00:00, 182.77it/s]
Loading weights:  97%|█████████▋| 283/291 [00:01<00:00, 181.36it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 177.43it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task591_sciq_answer_generation': PaR=53.9, shrinkage=ridge, n=1000
+  [SRT] Wired router: 13 sigs, cur=task591_sciq_answer_generation, mapping={'task591_sciq_answer_generation': 0, 'task1572_samsum_summary': 12, 'task363_sst2_polarity_classification': 11, 'task1290_xsum_summarization': 10, 'task181_outcome_extraction': 9, 'task002_quoref_answer_generation': 8, 'task1510_evalution_relation_extraction': 7, 'task639_multi_woz_user_utterance_generation': 6, 'task1729_personachat_generate_next': 5, 'task073_commonsenseqa_answer_generation': 4, 'task1590_diplomacy_text_generation': 3, 'task748_glucose_reverse_cause_event_detection': 2, 'task511_reddit_tifu_long_text_summarization': 1}
+[SRT] Router: 13 tasks, avg_PaR=53.9, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation']
+  [SRT-INIT] task_id_to_idx={'task591_sciq_answer_generation': 0, 'task1572_samsum_summary': 12, 'task363_sst2_polarity_classification': 11, 'task1290_xsum_summarization': 10, 'task181_outcome_extraction': 9, 'task002_quoref_answer_generation': 8, 'task1510_evalution_relation_extraction': 7, 'task639_multi_woz_user_utterance_generation': 6, 'task1729_personachat_generate_next': 5, 'task073_commonsenseqa_answer_generation': 4, 'task1590_diplomacy_text_generation': 3, 'task748_glucose_reverse_cause_event_detection': 2, 'task511_reddit_tifu_long_text_summarization': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/997 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=13
+
Continual predict:   0%|          | 1/997 [00:01<32:15,  1.94s/it]
Continual predict:   0%|          | 2/997 [00:04<40:25,  2.44s/it]
Continual predict:   0%|          | 3/997 [00:05<29:00,  1.75s/it]
Continual predict:   0%|          | 4/997 [00:06<21:37,  1.31s/it]
Continual predict:   1%|          | 5/997 [00:08<25:45,  1.56s/it]
Continual predict:   1%|          | 6/997 [00:13<45:08,  2.73s/it]
Continual predict:   1%|          | 7/997 [00:16<49:31,  3.00s/it]
Continual predict:   1%|          | 8/997 [00:19<47:29,  2.88s/it]
Continual predict:   1%|          | 9/997 [00:21<42:28,  2.58s/it]
Continual predict:   1%|          | 10/997 [00:22<33:48,  2.05s/it]
Continual predict:   1%|          | 11/997 [00:25<37:51,  2.30s/it]
Continual predict:   1%|          | 12/997 [00:28<43:23,  2.64s/it]
Continual predict:   1%|▏         | 13/997 [00:29<36:43,  2.24s/it]
Continual predict:   1%|▏         | 14/997 [00:32<38:37,  2.36s/it]
Continual predict:   2%|▏         | 15/997 [00:34<38:53,  2.38s/it]
Continual predict:   2%|▏         | 16/997 [00:37<40:41,  2.49s/it]
Continual predict:   2%|▏         | 17/997 [00:45<1:08:31,  4.20s/it]
Continual predict:   2%|▏         | 18/997 [00:48<58:56,  3.61s/it]  
Continual predict:   2%|▏         | 19/997 [00:50<54:31,  3.35s/it]
Continual predict:   2%|▏         | 20/997 [00:52<45:30,  2.79s/it]
Continual predict:   2%|▏         | 21/997 [00:52<32:44,  2.01s/it]
Continual predict:   2%|▏         | 22/997 [00:52<23:46,  1.46s/it]
Continual predict:   2%|▏         | 23/997 [00:52<17:32,  1.08s/it]
Continual predict:   2%|▏         | 24/997 [00:53<13:08,  1.23it/s]
Continual predict:   3%|▎         | 25/997 [00:53<10:04,  1.61it/s]
Continual predict:   3%|▎         | 26/997 [00:53<07:55,  2.04it/s]
Continual predict:   3%|▎         | 27/997 [00:53<06:24,  2.52it/s]
Continual predict:   3%|▎         | 28/997 [00:53<05:21,  3.02it/s]
Continual predict:   3%|▎         | 29/997 [00:53<04:36,  3.50it/s]
Continual predict:   3%|▎         | 30/997 [00:54<04:07,  3.90it/s]
Continual predict:   3%|▎         | 31/997 [00:54<03:45,  4.28it/s]
Continual predict:   3%|▎         | 32/997 [00:54<03:30,  4.59it/s]
Continual predict:   3%|▎         | 33/997 [00:54<03:21,  4.78it/s]
Continual predict:   3%|▎         | 34/997 [00:54<03:15,  4.93it/s]
Continual predict:   4%|▎         | 35/997 [00:55<03:16,  4.89it/s]
Continual predict:   4%|▎         | 36/997 [00:55<03:15,  4.92it/s]
Continual predict:   4%|▎         | 37/997 [00:55<03:26,  4.64it/s]
Continual predict:   4%|▍         | 38/997 [00:55<03:17,  4.86it/s]
Continual predict:   4%|▍         | 39/997 [00:55<03:09,  5.04it/s]
Continual predict:   4%|▍         | 40/997 [00:56<03:04,  5.18it/s]
Continual predict:   4%|▍         | 41/997 [00:56<03:03,  5.22it/s]
Continual predict:   4%|▍         | 42/997 [00:56<03:03,  5.19it/s]
Continual predict:   4%|▍         | 43/997 [00:56<03:02,  5.23it/s]
Continual predict:   4%|▍         | 44/997 [00:56<03:00,  5.27it/s]
Continual predict:   5%|▍         | 45/997 [00:57<02:57,  5.35it/s]
Continual predict:   5%|▍         | 46/997 [00:57<02:56,  5.39it/s]
Continual predict:   5%|▍         | 47/997 [00:57<02:55,  5.43it/s]
Continual predict:   5%|▍         | 48/997 [00:57<02:56,  5.38it/s]
Continual predict:   5%|▍         | 49/997 [00:57<02:56,  5.37it/s]
Continual predict:   5%|▌         | 50/997 [00:57<02:54,  5.42it/s]
Continual predict:   5%|▌         | 51/997 [00:58<02:53,  5.45it/s]
Continual predict:   5%|▌         | 52/997 [00:58<02:52,  5.48it/s]
Continual predict:   5%|▌         | 53/997 [00:58<02:51,  5.49it/s]
Continual predict:   5%|▌         | 54/997 [00:58<02:53,  5.44it/s]
Continual predict:   6%|▌         | 55/997 [00:58<02:52,  5.47it/s]
Continual predict:   6%|▌         | 56/997 [00:59<02:53,  5.43it/s]
Continual predict:   6%|▌         | 57/997 [00:59<02:52,  5.46it/s]
Continual predict:   6%|▌         | 58/997 [00:59<02:51,  5.48it/s]
Continual predict:   6%|▌         | 59/997 [00:59<02:50,  5.50it/s]
Continual predict:   6%|▌         | 60/997 [00:59<02:53,  5.41it/s]
Continual predict:   6%|▌         | 61/997 [00:59<02:51,  5.45it/s]
Continual predict:   6%|▌         | 62/997 [01:00<02:50,  5.49it/s]
Continual predict:   6%|▋         | 63/997 [01:00<02:51,  5.45it/s]
Continual predict:   6%|▋         | 64/997 [01:00<02:50,  5.47it/s]
Continual predict:   7%|▋         | 65/997 [01:00<02:49,  5.49it/s]
Continual predict:   7%|▋         | 66/997 [01:00<02:49,  5.49it/s]
Continual predict:   7%|▋         | 67/997 [01:01<02:48,  5.51it/s]
Continual predict:   7%|▋         | 68/997 [01:01<02:48,  5.52it/s]
Continual predict:   7%|▋         | 69/997 [01:01<02:47,  5.53it/s]
Continual predict:   7%|▋         | 70/997 [01:01<02:47,  5.54it/s]
Continual predict:   7%|▋         | 71/997 [01:01<02:47,  5.53it/s]
Continual predict:   7%|▋         | 72/997 [01:01<02:49,  5.47it/s]
Continual predict:   7%|▋         | 73/997 [01:02<02:48,  5.49it/s]
Continual predict:   7%|▋         | 74/997 [01:02<02:50,  5.43it/s]
Continual predict:   8%|▊         | 75/997 [01:02<02:49,  5.45it/s]
Continual predict:   8%|▊         | 76/997 [01:02<02:47,  5.48it/s]
Continual predict:   8%|▊         | 77/997 [01:02<02:47,  5.48it/s]
Continual predict:   8%|▊         | 78/997 [01:03<02:46,  5.50it/s]
Continual predict:   8%|▊         | 79/997 [01:03<02:46,  5.51it/s]
Continual predict:   8%|▊         | 80/997 [01:03<02:45,  5.53it/s]
Continual predict:   8%|▊         | 81/997 [01:03<02:45,  5.53it/s]
Continual predict:   8%|▊         | 82/997 [01:03<02:45,  5.53it/s]
Continual predict:   8%|▊         | 83/997 [01:03<02:45,  5.53it/s]
Continual predict:   8%|▊         | 84/997 [01:04<02:47,  5.47it/s]
Continual predict:   9%|▊         | 85/997 [01:04<02:46,  5.48it/s]
Continual predict:   9%|▊         | 86/997 [01:04<02:45,  5.49it/s]
Continual predict:   9%|▊         | 87/997 [01:04<02:45,  5.49it/s]
Continual predict:   9%|▉         | 88/997 [01:04<02:45,  5.50it/s]
Continual predict:   9%|▉         | 89/997 [01:05<02:44,  5.51it/s]
Continual predict:   9%|▉         | 90/997 [01:05<02:44,  5.53it/s]
Continual predict:   9%|▉         | 91/997 [01:05<02:43,  5.53it/s]
Continual predict:   9%|▉         | 92/997 [01:05<02:47,  5.39it/s]
Continual predict:   9%|▉         | 93/997 [01:05<03:36,  4.17it/s]
Continual predict:   9%|▉         | 94/997 [01:06<03:22,  4.46it/s]
Continual predict:  10%|▉         | 95/997 [01:06<03:10,  4.73it/s]
Continual predict:  10%|▉         | 96/997 [01:06<03:02,  4.94it/s]
Continual predict:  10%|▉         | 97/997 [01:06<02:56,  5.10it/s]
Continual predict:  10%|▉         | 98/997 [01:06<02:53,  5.17it/s]
Continual predict:  10%|▉         | 99/997 [01:07<02:52,  5.21it/s]
Continual predict:  10%|█         | 100/997 [01:07<02:51,  5.23it/s]
Continual predict:  10%|█         | 101/997 [01:07<02:48,  5.32it/s]
Continual predict:  10%|█         | 102/997 [01:07<02:46,  5.39it/s]
Continual predict:  10%|█         | 103/997 [01:07<02:44,  5.43it/s]
Continual predict:  10%|█         | 104/997 [01:07<02:44,  5.44it/s]
Continual predict:  11%|█         | 105/997 [01:08<02:44,  5.41it/s]
Continual predict:  11%|█         | 106/997 [01:08<02:43,  5.45it/s]
Continual predict:  11%|█         | 107/997 [01:08<02:42,  5.48it/s]
Continual predict:  11%|█         | 108/997 [01:08<02:41,  5.50it/s]
Continual predict:  11%|█         | 109/997 [01:08<02:41,  5.51it/s]
Continual predict:  11%|█         | 110/997 [01:09<02:41,  5.50it/s]
Continual predict:  11%|█         | 111/997 [01:09<02:40,  5.51it/s]
Continual predict:  11%|█         | 112/997 [01:09<02:40,  5.52it/s]
Continual predict:  11%|█▏        | 113/997 [01:09<02:40,  5.52it/s]
Continual predict:  11%|█▏        | 114/997 [01:09<02:40,  5.51it/s]
Continual predict:  12%|█▏        | 115/997 [01:10<02:41,  5.46it/s]
Continual predict:  12%|█▏        | 116/997 [01:10<02:41,  5.46it/s]
Continual predict:  12%|█▏        | 117/997 [01:10<02:45,  5.32it/s]
Continual predict:  12%|█▏        | 118/997 [01:10<02:45,  5.32it/s]
Continual predict:  12%|█▏        | 119/997 [01:10<02:43,  5.38it/s]
Continual predict:  12%|█▏        | 120/997 [01:10<02:41,  5.43it/s]
Continual predict:  12%|█▏        | 121/997 [01:15<23:14,  1.59s/it]
Continual predict:  12%|█▏        | 122/997 [01:20<35:57,  2.47s/it]
Continual predict:  12%|█▏        | 123/997 [01:25<49:55,  3.43s/it]
Continual predict:  12%|█▏        | 124/997 [01:28<46:11,  3.17s/it]
Continual predict:  13%|█▎        | 125/997 [01:33<54:35,  3.76s/it]
Continual predict:  13%|█▎        | 126/997 [01:38<1:00:21,  4.16s/it]
Continual predict:  13%|█▎        | 127/997 [01:40<49:15,  3.40s/it]  
Continual predict:  13%|█▎        | 128/997 [01:43<47:28,  3.28s/it]
Continual predict:  13%|█▎        | 129/997 [01:47<49:02,  3.39s/it]
Continual predict:  13%|█▎        | 130/997 [01:50<48:41,  3.37s/it]
Continual predict:  13%|█▎        | 131/997 [01:52<44:48,  3.10s/it]
Continual predict:  13%|█▎        | 132/997 [01:55<43:55,  3.05s/it]
Continual predict:  13%|█▎        | 133/997 [01:59<46:02,  3.20s/it]
Continual predict:  13%|█▎        | 134/997 [02:05<1:00:16,  4.19s/it]
Continual predict:  14%|█▎        | 135/997 [02:11<1:04:33,  4.49s/it]
Continual predict:  14%|█▎        | 136/997 [02:14<1:00:23,  4.21s/it]
Continual predict:  14%|█▎        | 137/997 [02:17<52:43,  3.68s/it]  
Continual predict:  14%|█▍        | 138/997 [02:17<40:39,  2.84s/it]
Continual predict:  14%|█▍        | 139/997 [02:19<34:00,  2.38s/it]
Continual predict:  14%|█▍        | 140/997 [02:22<36:33,  2.56s/it]
Continual predict:  14%|█▍        | 141/997 [02:29<55:32,  3.89s/it]
Continual predict:  14%|█▍        | 142/997 [02:34<59:42,  4.19s/it]
Continual predict:  14%|█▍        | 143/997 [02:37<56:46,  3.99s/it]
Continual predict:  14%|█▍        | 144/997 [02:47<1:22:26,  5.80s/it]
Continual predict:  15%|█▍        | 145/997 [02:52<1:18:07,  5.50s/it]
Continual predict:  15%|█▍        | 146/997 [02:54<1:02:53,  4.43s/it]
Continual predict:  15%|█▍        | 147/997 [02:59<1:03:54,  4.51s/it]
Continual predict:  15%|█▍        | 148/997 [03:01<54:38,  3.86s/it]  
Continual predict:  15%|█▍        | 149/997 [03:04<52:42,  3.73s/it]
Continual predict:  15%|█▌        | 150/997 [03:06<45:43,  3.24s/it]
Continual predict:  15%|█▌        | 151/997 [03:10<45:45,  3.25s/it]
Continual predict:  15%|█▌        | 152/997 [03:16<57:48,  4.10s/it]
Continual predict:  15%|█▌        | 153/997 [03:18<51:43,  3.68s/it]
Continual predict:  15%|█▌        | 154/997 [03:24<57:44,  4.11s/it]
Continual predict:  16%|█▌        | 155/997 [03:27<56:46,  4.05s/it]
Continual predict:  16%|█▌        | 156/997 [03:31<55:07,  3.93s/it]
Continual predict:  16%|█▌        | 157/997 [03:35<56:08,  4.01s/it]
Continual predict:  16%|█▌        | 158/997 [03:36<43:58,  3.14s/it]
Continual predict:  16%|█▌        | 159/997 [03:44<1:01:38,  4.41s/it]
Continual predict:  16%|█▌        | 160/997 [03:46<53:25,  3.83s/it]  
Continual predict:  16%|█▌        | 161/997 [03:49<48:33,  3.49s/it]
Continual predict:  16%|█▌        | 162/997 [03:54<55:22,  3.98s/it]
Continual predict:  16%|█▋        | 163/997 [03:57<49:01,  3.53s/it]
Continual predict:  16%|█▋        | 164/997 [03:59<44:36,  3.21s/it]
Continual predict:  17%|█▋        | 165/997 [04:02<45:15,  3.26s/it]
Continual predict:  17%|█▋        | 166/997 [04:05<41:30,  3.00s/it]
Continual predict:  17%|█▋        | 167/997 [04:06<34:48,  2.52s/it]
Continual predict:  17%|█▋        | 168/997 [04:09<34:28,  2.50s/it]
Continual predict:  17%|█▋        | 169/997 [04:11<31:53,  2.31s/it]
Continual predict:  17%|█▋        | 170/997 [04:14<35:54,  2.60s/it]
Continual predict:  17%|█▋        | 171/997 [04:16<32:16,  2.34s/it]
Continual predict:  17%|█▋        | 172/997 [04:18<34:11,  2.49s/it]
Continual predict:  17%|█▋        | 173/997 [04:23<41:12,  3.00s/it]
Continual predict:  17%|█▋        | 174/997 [04:24<34:08,  2.49s/it]
Continual predict:  18%|█▊        | 175/997 [04:28<39:24,  2.88s/it]
Continual predict:  18%|█▊        | 176/997 [04:31<40:16,  2.94s/it]
Continual predict:  18%|█▊        | 177/997 [04:34<40:19,  2.95s/it]
Continual predict:  18%|█▊        | 178/997 [04:35<33:56,  2.49s/it]
Continual predict:  18%|█▊        | 179/997 [04:36<28:52,  2.12s/it]
Continual predict:  18%|█▊        | 180/997 [04:38<25:53,  1.90s/it]
Continual predict:  18%|█▊        | 181/997 [04:39<23:47,  1.75s/it]
Continual predict:  18%|█▊        | 182/997 [04:44<37:03,  2.73s/it]
Continual predict:  18%|█▊        | 183/997 [04:44<26:51,  1.98s/it]
Continual predict:  18%|█▊        | 184/997 [04:46<23:29,  1.73s/it]
Continual predict:  19%|█▊        | 185/997 [04:49<31:48,  2.35s/it]
Continual predict:  19%|█▊        | 186/997 [04:51<26:52,  1.99s/it]
Continual predict:  19%|█▉        | 187/997 [04:54<32:55,  2.44s/it]
Continual predict:  19%|█▉        | 188/997 [04:57<34:40,  2.57s/it]
Continual predict:  19%|█▉        | 189/997 [05:02<46:06,  3.42s/it]
Continual predict:  19%|█▉        | 190/997 [05:06<46:36,  3.47s/it]
Continual predict:  19%|█▉        | 191/997 [05:07<38:25,  2.86s/it]
Continual predict:  19%|█▉        | 192/997 [05:11<41:22,  3.08s/it]
Continual predict:  19%|█▉        | 193/997 [05:16<48:34,  3.62s/it]
Continual predict:  19%|█▉        | 194/997 [05:21<54:36,  4.08s/it]
Continual predict:  20%|█▉        | 195/997 [05:27<1:01:00,  4.56s/it]
Continual predict:  20%|█▉        | 196/997 [05:30<56:53,  4.26s/it]  
Continual predict:  20%|█▉        | 197/997 [05:34<53:47,  4.03s/it]
Continual predict:  20%|█▉        | 198/997 [05:42<1:09:06,  5.19s/it]
Continual predict:  20%|█▉        | 199/997 [05:43<52:56,  3.98s/it]  
Continual predict:  20%|██        | 200/997 [05:44<43:26,  3.27s/it]
Continual predict:  20%|██        | 201/997 [05:47<41:20,  3.12s/it]
Continual predict:  20%|██        | 202/997 [05:51<42:50,  3.23s/it]
Continual predict:  20%|██        | 203/997 [05:56<51:39,  3.90s/it]
Continual predict:  20%|██        | 204/997 [05:59<49:09,  3.72s/it]
Continual predict:  21%|██        | 205/997 [06:06<1:02:05,  4.70s/it]
Continual predict:  21%|██        | 206/997 [06:09<55:03,  4.18s/it]  
Continual predict:  21%|██        | 207/997 [06:13<54:07,  4.11s/it]
Continual predict:  21%|██        | 208/997 [06:20<1:03:37,  4.84s/it]
Continual predict:  21%|██        | 209/997 [06:23<56:04,  4.27s/it]  
Continual predict:  21%|██        | 210/997 [06:25<46:07,  3.52s/it]
Continual predict:  21%|██        | 211/997 [06:27<43:20,  3.31s/it]
Continual predict:  21%|██▏       | 212/997 [06:31<43:36,  3.33s/it]
Continual predict:  21%|██▏       | 213/997 [06:35<46:19,  3.55s/it]
Continual predict:  21%|██▏       | 214/997 [06:39<47:46,  3.66s/it]
Continual predict:  22%|██▏       | 215/997 [06:40<36:58,  2.84s/it]
Continual predict:  22%|██▏       | 216/997 [06:43<39:11,  3.01s/it]
Continual predict:  22%|██▏       | 217/997 [06:45<36:14,  2.79s/it]
Continual predict:  22%|██▏       | 218/997 [06:50<41:53,  3.23s/it]
Continual predict:  22%|██▏       | 219/997 [06:53<43:13,  3.33s/it]
Continual predict:  22%|██▏       | 220/997 [06:58<47:44,  3.69s/it]
Continual predict:  22%|██▏       | 221/997 [06:58<34:16,  2.65s/it]
Continual predict:  22%|██▏       | 222/997 [06:58<25:08,  1.95s/it]
Continual predict:  22%|██▏       | 223/997 [06:58<18:27,  1.43s/it]
Continual predict:  22%|██▏       | 224/997 [06:59<13:48,  1.07s/it]
Continual predict:  23%|██▎       | 225/997 [06:59<10:32,  1.22it/s]
Continual predict:  23%|██▎       | 226/997 [06:59<08:15,  1.56it/s]
Continual predict:  23%|██▎       | 227/997 [06:59<06:51,  1.87it/s]
Continual predict:  23%|██▎       | 228/997 [07:00<05:53,  2.18it/s]
Continual predict:  23%|██▎       | 229/997 [07:00<05:04,  2.52it/s]
Continual predict:  23%|██▎       | 230/997 [07:00<05:18,  2.40it/s]
Continual predict:  23%|██▎       | 231/997 [07:01<04:36,  2.77it/s]
Continual predict:  23%|██▎       | 232/997 [07:01<04:06,  3.11it/s]
Continual predict:  23%|██▎       | 233/997 [07:01<03:44,  3.40it/s]
Continual predict:  23%|██▎       | 234/997 [07:01<03:30,  3.63it/s]
Continual predict:  24%|██▎       | 235/997 [07:02<03:19,  3.82it/s]
Continual predict:  24%|██▎       | 236/997 [07:02<03:11,  3.97it/s]
Continual predict:  24%|██▍       | 237/997 [07:02<03:06,  4.07it/s]
Continual predict:  24%|██▍       | 238/997 [07:02<03:02,  4.15it/s]
Continual predict:  24%|██▍       | 239/997 [07:02<02:59,  4.21it/s]
Continual predict:  24%|██▍       | 240/997 [07:03<03:33,  3.55it/s]
Continual predict:  24%|██▍       | 241/997 [07:03<03:21,  3.76it/s]
Continual predict:  24%|██▍       | 242/997 [07:03<03:12,  3.92it/s]
Continual predict:  24%|██▍       | 243/997 [07:04<03:06,  4.04it/s]
Continual predict:  24%|██▍       | 244/997 [07:04<03:18,  3.79it/s]
Continual predict:  25%|██▍       | 245/997 [07:04<03:17,  3.80it/s]
Continual predict:  25%|██▍       | 246/997 [07:04<03:10,  3.95it/s]
Continual predict:  25%|██▍       | 247/997 [07:05<03:16,  3.81it/s]
Continual predict:  25%|██▍       | 248/997 [07:05<03:09,  3.96it/s]
Continual predict:  25%|██▍       | 249/997 [07:05<03:15,  3.82it/s]
Continual predict:  25%|██▌       | 250/997 [07:05<03:08,  3.97it/s]
Continual predict:  25%|██▌       | 251/997 [07:06<03:14,  3.83it/s]
Continual predict:  25%|██▌       | 252/997 [07:07<07:13,  1.72it/s]
Continual predict:  25%|██▌       | 253/997 [07:07<06:06,  2.03it/s]
Continual predict:  25%|██▌       | 254/997 [07:07<05:07,  2.42it/s]
Continual predict:  26%|██▌       | 255/997 [07:08<04:26,  2.79it/s]
Continual predict:  26%|██▌       | 256/997 [07:08<03:57,  3.12it/s]
Continual predict:  26%|██▌       | 257/997 [07:09<05:42,  2.16it/s]
Continual predict:  26%|██▌       | 258/997 [07:09<04:50,  2.55it/s]
Continual predict:  26%|██▌       | 259/997 [07:09<04:13,  2.91it/s]
Continual predict:  26%|██▌       | 260/997 [07:09<03:48,  3.22it/s]
Continual predict:  26%|██▌       | 261/997 [07:10<04:55,  2.49it/s]
Continual predict:  26%|██▋       | 262/997 [07:10<04:21,  2.81it/s]
Continual predict:  26%|██▋       | 263/997 [07:11<04:05,  2.99it/s]
Continual predict:  26%|██▋       | 264/997 [07:12<07:38,  1.60it/s]
Continual predict:  27%|██▋       | 265/997 [07:12<07:28,  1.63it/s]
Continual predict:  27%|██▋       | 266/997 [07:13<07:20,  1.66it/s]
Continual predict:  27%|██▋       | 267/997 [07:14<07:28,  1.63it/s]
Continual predict:  27%|██▋       | 268/997 [07:15<08:27,  1.44it/s]
Continual predict:  27%|██▋       | 269/997 [07:15<08:15,  1.47it/s]
Continual predict:  27%|██▋       | 270/997 [07:16<08:13,  1.47it/s]
Continual predict:  27%|██▋       | 271/997 [07:17<08:40,  1.39it/s]
Continual predict:  27%|██▋       | 272/997 [07:18<09:32,  1.27it/s]
Continual predict:  27%|██▋       | 273/997 [07:18<09:11,  1.31it/s]
Continual predict:  27%|██▋       | 274/997 [07:19<08:56,  1.35it/s]
Continual predict:  28%|██▊       | 275/997 [07:20<08:34,  1.40it/s]
Continual predict:  28%|██▊       | 276/997 [07:21<10:02,  1.20it/s]
Continual predict:  28%|██▊       | 277/997 [07:22<11:55,  1.01it/s]
Continual predict:  28%|██▊       | 278/997 [07:23<11:28,  1.04it/s]
Continual predict:  28%|██▊       | 279/997 [07:24<10:44,  1.11it/s]
Continual predict:  28%|██▊       | 280/997 [07:24<09:13,  1.30it/s]
Continual predict:  28%|██▊       | 281/997 [07:25<10:08,  1.18it/s]
Continual predict:  28%|██▊       | 282/997 [07:26<10:10,  1.17it/s]
Continual predict:  28%|██▊       | 283/997 [07:27<09:40,  1.23it/s]
Continual predict:  28%|██▊       | 284/997 [07:27<08:29,  1.40it/s]
Continual predict:  29%|██▊       | 285/997 [07:28<08:56,  1.33it/s]
Continual predict:  29%|██▊       | 286/997 [07:29<09:07,  1.30it/s]
Continual predict:  29%|██▉       | 287/997 [07:30<09:15,  1.28it/s]
Continual predict:  29%|██▉       | 288/997 [07:31<08:53,  1.33it/s]
Continual predict:  29%|██▉       | 289/997 [07:32<09:48,  1.20it/s]
Continual predict:  29%|██▉       | 290/997 [07:32<09:41,  1.22it/s]
Continual predict:  29%|██▉       | 291/997 [07:33<09:02,  1.30it/s]
Continual predict:  29%|██▉       | 292/997 [07:33<07:30,  1.56it/s]
Continual predict:  29%|██▉       | 293/997 [07:34<08:20,  1.41it/s]
Continual predict:  29%|██▉       | 294/997 [07:35<07:51,  1.49it/s]
Continual predict:  30%|██▉       | 295/997 [07:35<07:10,  1.63it/s]
Continual predict:  30%|██▉       | 296/997 [07:36<08:39,  1.35it/s]
Continual predict:  30%|██▉       | 297/997 [07:37<08:00,  1.46it/s]
Continual predict:  30%|██▉       | 298/997 [07:37<07:17,  1.60it/s]
Continual predict:  30%|██▉       | 299/997 [07:38<08:26,  1.38it/s]
Continual predict:  30%|███       | 300/997 [07:39<07:07,  1.63it/s]
Continual predict:  30%|███       | 301/997 [07:40<08:21,  1.39it/s]
Continual predict:  30%|███       | 302/997 [07:40<07:25,  1.56it/s]
Continual predict:  30%|███       | 303/997 [07:42<10:48,  1.07it/s]
Continual predict:  30%|███       | 304/997 [07:42<08:45,  1.32it/s]
Continual predict:  31%|███       | 305/997 [07:43<09:26,  1.22it/s]
Continual predict:  31%|███       | 306/997 [07:44<09:35,  1.20it/s]
Continual predict:  31%|███       | 307/997 [07:45<09:38,  1.19it/s]
Continual predict:  31%|███       | 308/997 [07:45<08:20,  1.38it/s]
Continual predict:  31%|███       | 309/997 [07:46<07:29,  1.53it/s]
Continual predict:  31%|███       | 310/997 [07:46<07:27,  1.53it/s]
Continual predict:  31%|███       | 311/997 [07:47<07:12,  1.59it/s]
Continual predict:  31%|███▏      | 312/997 [07:48<07:37,  1.50it/s]
Continual predict:  31%|███▏      | 313/997 [07:49<08:30,  1.34it/s]
Continual predict:  31%|███▏      | 314/997 [07:49<07:36,  1.50it/s]
Continual predict:  32%|███▏      | 315/997 [07:50<08:35,  1.32it/s]
Continual predict:  32%|███▏      | 316/997 [07:51<08:55,  1.27it/s]
Continual predict:  32%|███▏      | 317/997 [07:51<07:25,  1.53it/s]
Continual predict:  32%|███▏      | 318/997 [07:52<06:49,  1.66it/s]
Continual predict:  32%|███▏      | 319/997 [07:53<10:36,  1.07it/s]
Continual predict:  32%|███▏      | 320/997 [07:54<09:58,  1.13it/s]
Continual predict:  32%|███▏      | 321/997 [07:55<09:29,  1.19it/s]
Continual predict:  32%|███▏      | 322/997 [07:55<08:11,  1.37it/s]
Continual predict:  32%|███▏      | 323/997 [07:56<08:27,  1.33it/s]
Continual predict:  32%|███▏      | 324/997 [07:57<09:06,  1.23it/s]
Continual predict:  33%|███▎      | 325/997 [07:58<09:05,  1.23it/s]
Continual predict:  33%|███▎      | 326/997 [07:59<08:37,  1.30it/s]
Continual predict:  33%|███▎      | 327/997 [08:00<10:04,  1.11it/s]
Continual predict:  33%|███▎      | 328/997 [08:01<10:47,  1.03it/s]
Continual predict:  33%|███▎      | 329/997 [08:02<09:35,  1.16it/s]
Continual predict:  33%|███▎      | 330/997 [08:02<07:52,  1.41it/s]
Continual predict:  33%|███▎      | 331/997 [08:03<07:37,  1.46it/s]
Continual predict:  33%|███▎      | 332/997 [08:03<07:15,  1.53it/s]
Continual predict:  33%|███▎      | 333/997 [08:04<09:21,  1.18it/s]
Continual predict:  34%|███▎      | 334/997 [08:05<09:42,  1.14it/s]
Continual predict:  34%|███▎      | 335/997 [08:06<08:19,  1.33it/s]
Continual predict:  34%|███▎      | 336/997 [08:06<07:24,  1.49it/s]
Continual predict:  34%|███▍      | 337/997 [08:07<07:11,  1.53it/s]
Continual predict:  34%|███▍      | 338/997 [08:08<07:18,  1.50it/s]
Continual predict:  34%|███▍      | 339/997 [08:08<06:59,  1.57it/s]
Continual predict:  34%|███▍      | 340/997 [08:09<08:38,  1.27it/s]
Continual predict:  34%|███▍      | 341/997 [08:10<08:39,  1.26it/s]
Continual predict:  34%|███▍      | 342/997 [08:11<07:37,  1.43it/s]
Continual predict:  34%|███▍      | 343/997 [08:11<07:58,  1.37it/s]
Continual predict:  35%|███▍      | 344/997 [08:12<08:53,  1.22it/s]
Continual predict:  35%|███▍      | 345/997 [08:13<07:47,  1.39it/s]
Continual predict:  35%|███▍      | 346/997 [08:14<08:15,  1.31it/s]
Continual predict:  35%|███▍      | 347/997 [08:15<10:48,  1.00it/s]
Continual predict:  35%|███▍      | 348/997 [08:16<08:44,  1.24it/s]
Continual predict:  35%|███▌      | 349/997 [08:16<07:37,  1.42it/s]
Continual predict:  35%|███▌      | 350/997 [08:17<07:33,  1.43it/s]
Continual predict:  35%|███▌      | 351/997 [08:18<09:25,  1.14it/s]
Continual predict:  35%|███▌      | 352/997 [08:19<09:10,  1.17it/s]
Continual predict:  35%|███▌      | 353/997 [08:20<10:49,  1.01s/it]
Continual predict:  36%|███▌      | 354/997 [08:22<11:51,  1.11s/it]
Continual predict:  36%|███▌      | 355/997 [08:22<10:22,  1.03it/s]
Continual predict:  36%|███▌      | 356/997 [08:23<10:20,  1.03it/s]
Continual predict:  36%|███▌      | 357/997 [08:24<10:51,  1.02s/it]
Continual predict:  36%|███▌      | 358/997 [08:25<10:08,  1.05it/s]
Continual predict:  36%|███▌      | 359/997 [08:26<09:14,  1.15it/s]
Continual predict:  36%|███▌      | 360/997 [08:27<10:42,  1.01s/it]
Continual predict:  36%|███▌      | 361/997 [08:29<12:51,  1.21s/it]
Continual predict:  36%|███▋      | 362/997 [08:30<11:00,  1.04s/it]
Continual predict:  36%|███▋      | 363/997 [08:30<09:13,  1.15it/s]
Continual predict:  37%|███▋      | 364/997 [08:30<07:38,  1.38it/s]
Continual predict:  37%|███▋      | 365/997 [08:31<06:19,  1.67it/s]
Continual predict:  37%|███▋      | 366/997 [08:31<05:24,  1.95it/s]
Continual predict:  37%|███▋      | 367/997 [08:31<04:45,  2.21it/s]
Continual predict:  37%|███▋      | 368/997 [08:32<04:19,  2.43it/s]
Continual predict:  37%|███▋      | 369/997 [08:32<04:00,  2.62it/s]
Continual predict:  37%|███▋      | 370/997 [08:32<03:46,  2.76it/s]
Continual predict:  37%|███▋      | 371/997 [08:33<03:49,  2.73it/s]
Continual predict:  37%|███▋      | 372/997 [08:33<04:02,  2.58it/s]
Continual predict:  37%|███▋      | 373/997 [08:34<04:11,  2.48it/s]
Continual predict:  38%|███▊      | 374/997 [08:34<03:54,  2.66it/s]
Continual predict:  38%|███▊      | 375/997 [08:34<04:05,  2.54it/s]
Continual predict:  38%|███▊      | 376/997 [08:35<04:01,  2.57it/s]
Continual predict:  38%|███▊      | 377/997 [08:35<03:46,  2.73it/s]
Continual predict:  38%|███▊      | 378/997 [08:35<03:36,  2.86it/s]
Continual predict:  38%|███▊      | 379/997 [08:36<03:29,  2.95it/s]
Continual predict:  38%|███▊      | 380/997 [08:36<03:24,  3.02it/s]
Continual predict:  38%|███▊      | 381/997 [08:36<03:32,  2.91it/s]
Continual predict:  38%|███▊      | 382/997 [08:37<03:25,  2.99it/s]
Continual predict:  38%|███▊      | 383/997 [08:37<03:21,  3.04it/s]
Continual predict:  39%|███▊      | 384/997 [08:37<03:18,  3.09it/s]
Continual predict:  39%|███▊      | 385/997 [08:38<03:16,  3.12it/s]
Continual predict:  39%|███▊      | 386/997 [08:38<03:14,  3.14it/s]
Continual predict:  39%|███▉      | 387/997 [08:38<03:14,  3.14it/s]
Continual predict:  39%|███▉      | 388/997 [08:39<03:13,  3.15it/s]
Continual predict:  39%|███▉      | 389/997 [08:39<03:30,  2.89it/s]
Continual predict:  39%|███▉      | 390/997 [08:39<03:28,  2.91it/s]
Continual predict:  39%|███▉      | 391/997 [08:40<03:22,  2.99it/s]
Continual predict:  39%|███▉      | 392/997 [08:40<03:30,  2.88it/s]
Continual predict:  39%|███▉      | 393/997 [08:40<03:23,  2.97it/s]
Continual predict:  40%|███▉      | 394/997 [08:41<03:19,  3.03it/s]
Continual predict:  40%|███▉      | 395/997 [08:41<03:04,  3.25it/s]
Continual predict:  40%|███▉      | 396/997 [08:41<03:29,  2.87it/s]
Continual predict:  40%|███▉      | 397/997 [08:42<03:37,  2.76it/s]
Continual predict:  40%|███▉      | 398/997 [08:42<03:28,  2.88it/s]
Continual predict:  40%|████      | 399/997 [08:42<03:22,  2.96it/s]
Continual predict:  40%|████      | 400/997 [08:43<03:17,  3.02it/s]
Continual predict:  40%|████      | 401/997 [08:43<03:13,  3.07it/s]
Continual predict:  40%|████      | 402/997 [08:43<03:11,  3.11it/s]
Continual predict:  40%|████      | 403/997 [08:44<03:20,  2.96it/s]
Continual predict:  41%|████      | 404/997 [08:44<03:15,  3.03it/s]
Continual predict:  41%|████      | 405/997 [08:44<03:34,  2.76it/s]
Continual predict:  41%|████      | 406/997 [08:45<03:25,  2.88it/s]
Continual predict:  41%|████      | 407/997 [08:45<03:19,  2.96it/s]
Continual predict:  41%|████      | 408/997 [08:45<03:14,  3.03it/s]
Continual predict:  41%|████      | 409/997 [08:46<03:11,  3.08it/s]
Continual predict:  41%|████      | 410/997 [08:46<03:30,  2.79it/s]
Continual predict:  41%|████      | 411/997 [08:47<03:44,  2.61it/s]
Continual predict:  41%|████▏     | 412/997 [08:47<03:31,  2.76it/s]
Continual predict:  41%|████▏     | 413/997 [08:47<03:23,  2.87it/s]
Continual predict:  42%|████▏     | 414/997 [08:47<03:17,  2.95it/s]
Continual predict:  42%|████▏     | 415/997 [08:48<03:12,  3.02it/s]
Continual predict:  42%|████▏     | 416/997 [08:48<03:09,  3.07it/s]
Continual predict:  42%|████▏     | 417/997 [08:48<03:06,  3.10it/s]
Continual predict:  42%|████▏     | 418/997 [08:49<03:15,  2.96it/s]
Continual predict:  42%|████▏     | 419/997 [08:49<04:07,  2.34it/s]
Continual predict:  42%|████▏     | 420/997 [08:50<04:14,  2.27it/s]
Continual predict:  42%|████▏     | 421/997 [08:50<04:13,  2.27it/s]
Continual predict:  42%|████▏     | 422/997 [08:51<04:02,  2.37it/s]
Continual predict:  42%|████▏     | 423/997 [08:51<03:54,  2.45it/s]
Continual predict:  43%|████▎     | 424/997 [08:52<03:59,  2.39it/s]
Continual predict:  43%|████▎     | 425/997 [08:52<04:02,  2.36it/s]
Continual predict:  43%|████▎     | 426/997 [08:52<04:04,  2.33it/s]
Continual predict:  43%|████▎     | 427/997 [08:53<04:03,  2.34it/s]
Continual predict:  43%|████▎     | 428/997 [08:53<03:43,  2.55it/s]
Continual predict:  43%|████▎     | 429/997 [08:53<03:29,  2.71it/s]
Continual predict:  43%|████▎     | 430/997 [08:54<03:19,  2.84it/s]
Continual predict:  43%|████▎     | 431/997 [08:54<03:23,  2.78it/s]
Continual predict:  43%|████▎     | 432/997 [08:54<03:15,  2.90it/s]
Continual predict:  43%|████▎     | 433/997 [08:55<03:09,  2.97it/s]
Continual predict:  44%|████▎     | 434/997 [08:55<03:16,  2.87it/s]
Continual predict:  44%|████▎     | 435/997 [08:55<03:09,  2.96it/s]
Continual predict:  44%|████▎     | 436/997 [08:56<03:26,  2.71it/s]
Continual predict:  44%|████▍     | 437/997 [08:56<03:48,  2.45it/s]
Continual predict:  44%|████▍     | 438/997 [08:57<03:32,  2.63it/s]
Continual predict:  44%|████▍     | 439/997 [08:57<03:31,  2.64it/s]
Continual predict:  44%|████▍     | 440/997 [08:57<03:30,  2.64it/s]
Continual predict:  44%|████▍     | 441/997 [08:58<03:30,  2.65it/s]
Continual predict:  44%|████▍     | 442/997 [08:58<03:18,  2.79it/s]
Continual predict:  44%|████▍     | 443/997 [08:58<03:10,  2.90it/s]
Continual predict:  45%|████▍     | 444/997 [08:59<03:05,  2.98it/s]
Continual predict:  45%|████▍     | 445/997 [08:59<03:01,  3.04it/s]
Continual predict:  45%|████▍     | 446/997 [08:59<03:08,  2.92it/s]
Continual predict:  45%|████▍     | 447/997 [09:00<03:44,  2.45it/s]
Continual predict:  45%|████▍     | 448/997 [09:00<03:38,  2.51it/s]
Continual predict:  45%|████▌     | 449/997 [09:01<03:44,  2.44it/s]
Continual predict:  45%|████▌     | 450/997 [09:01<03:28,  2.62it/s]
Continual predict:  45%|████▌     | 451/997 [09:01<03:17,  2.77it/s]
Continual predict:  45%|████▌     | 452/997 [09:02<03:09,  2.88it/s]
Continual predict:  45%|████▌     | 453/997 [09:02<03:03,  2.96it/s]
Continual predict:  46%|████▌     | 454/997 [09:02<02:59,  3.03it/s]
Continual predict:  46%|████▌     | 455/997 [09:03<02:58,  3.03it/s]
Continual predict:  46%|████▌     | 456/997 [09:03<03:06,  2.91it/s]
Continual predict:  46%|████▌     | 457/997 [09:03<03:10,  2.83it/s]
Continual predict:  46%|████▌     | 458/997 [09:04<03:03,  2.93it/s]
Continual predict:  46%|████▌     | 459/997 [09:04<02:59,  3.00it/s]
Continual predict:  46%|████▌     | 460/997 [09:04<02:55,  3.06it/s]
Continual predict:  46%|████▌     | 461/997 [09:05<02:53,  3.09it/s]
Continual predict:  46%|████▋     | 462/997 [09:05<02:51,  3.12it/s]
Continual predict:  46%|████▋     | 463/997 [09:05<02:49,  3.15it/s]
Continual predict:  47%|████▋     | 464/997 [09:06<03:05,  2.87it/s]
Continual predict:  47%|████▋     | 465/997 [09:06<02:48,  3.17it/s]
Continual predict:  47%|████▋     | 466/997 [09:06<02:54,  3.05it/s]
Continual predict:  47%|████▋     | 467/997 [09:07<02:49,  3.14it/s]
Continual predict:  47%|████▋     | 468/997 [09:07<02:54,  3.03it/s]
Continual predict:  47%|████▋     | 469/997 [09:07<02:58,  2.95it/s]
Continual predict:  47%|████▋     | 470/997 [09:08<03:01,  2.90it/s]
Continual predict:  47%|████▋     | 471/997 [09:08<03:03,  2.87it/s]
Continual predict:  47%|████▋     | 472/997 [09:09<04:48,  1.82it/s]
Continual predict:  47%|████▋     | 473/997 [09:10<04:56,  1.77it/s]
Continual predict:  48%|████▊     | 474/997 [09:10<04:04,  2.14it/s]
Continual predict:  48%|████▊     | 475/997 [09:11<04:16,  2.04it/s]
Continual predict:  48%|████▊     | 476/997 [09:11<04:04,  2.13it/s]
Continual predict:  48%|████▊     | 477/997 [09:11<03:46,  2.30it/s]
Continual predict:  48%|████▊     | 478/997 [09:12<03:43,  2.32it/s]
Continual predict:  48%|████▊     | 479/997 [09:13<04:47,  1.80it/s]
Continual predict:  48%|████▊     | 480/997 [09:13<04:16,  2.01it/s]
Continual predict:  48%|████▊     | 481/997 [09:13<04:04,  2.11it/s]
Continual predict:  48%|████▊     | 482/997 [09:14<04:13,  2.04it/s]
Continual predict:  48%|████▊     | 483/997 [09:15<05:17,  1.62it/s]
Continual predict:  49%|████▊     | 484/997 [09:16<06:38,  1.29it/s]
Continual predict:  49%|████▊     | 485/997 [09:17<07:42,  1.11it/s]
Continual predict:  49%|████▊     | 486/997 [09:18<08:17,  1.03it/s]
Continual predict:  49%|████▉     | 487/997 [09:20<09:52,  1.16s/it]
Continual predict:  49%|████▉     | 488/997 [09:20<08:04,  1.05it/s]
Continual predict:  49%|████▉     | 489/997 [09:21<07:35,  1.11it/s]
Continual predict:  49%|████▉     | 490/997 [09:22<07:25,  1.14it/s]
Continual predict:  49%|████▉     | 491/997 [09:23<07:19,  1.15it/s]
Continual predict:  49%|████▉     | 492/997 [09:24<07:36,  1.11it/s]
Continual predict:  49%|████▉     | 493/997 [09:25<07:16,  1.15it/s]
Continual predict:  50%|████▉     | 494/997 [09:25<07:25,  1.13it/s]
Continual predict:  50%|████▉     | 495/997 [09:27<09:20,  1.12s/it]
Continual predict:  50%|████▉     | 496/997 [09:28<08:38,  1.04s/it]
Continual predict:  50%|████▉     | 497/997 [09:29<07:33,  1.10it/s]
Continual predict:  50%|████▉     | 498/997 [09:30<08:01,  1.04it/s]
Continual predict:  50%|█████     | 499/997 [09:31<07:52,  1.05it/s]
Continual predict:  50%|█████     | 500/997 [09:31<06:40,  1.24it/s]
Continual predict:  50%|█████     | 501/997 [09:32<06:09,  1.34it/s]
Continual predict:  50%|█████     | 502/997 [09:32<06:00,  1.37it/s]
Continual predict:  50%|█████     | 503/997 [09:33<06:17,  1.31it/s]
Continual predict:  51%|█████     | 504/997 [09:34<05:52,  1.40it/s]
Continual predict:  51%|█████     | 505/997 [09:35<07:01,  1.17it/s]
Continual predict:  51%|█████     | 506/997 [09:36<07:12,  1.13it/s]
Continual predict:  51%|█████     | 507/997 [09:37<07:27,  1.10it/s]
Continual predict:  51%|█████     | 508/997 [09:38<07:04,  1.15it/s]
Continual predict:  51%|█████     | 509/997 [09:39<07:08,  1.14it/s]
Continual predict:  51%|█████     | 510/997 [09:40<09:33,  1.18s/it]
Continual predict:  51%|█████▏    | 511/997 [09:42<09:27,  1.17s/it]
Continual predict:  51%|█████▏    | 512/997 [09:42<08:17,  1.03s/it]
Continual predict:  51%|█████▏    | 513/997 [09:44<09:53,  1.23s/it]
Continual predict:  52%|█████▏    | 514/997 [09:45<09:25,  1.17s/it]
Continual predict:  52%|█████▏    | 515/997 [09:46<08:25,  1.05s/it]
Continual predict:  52%|█████▏    | 516/997 [09:47<07:41,  1.04it/s]
Continual predict:  52%|█████▏    | 517/997 [09:48<07:51,  1.02it/s]
Continual predict:  52%|█████▏    | 518/997 [09:49<07:50,  1.02it/s]
Continual predict:  52%|█████▏    | 519/997 [09:50<09:21,  1.18s/it]
Continual predict:  52%|█████▏    | 520/997 [09:51<08:11,  1.03s/it]
Continual predict:  52%|█████▏    | 521/997 [09:52<08:11,  1.03s/it]
Continual predict:  52%|█████▏    | 522/997 [09:53<07:17,  1.09it/s]
Continual predict:  52%|█████▏    | 523/997 [09:53<07:04,  1.12it/s]
Continual predict:  53%|█████▎    | 524/997 [09:54<06:35,  1.20it/s]
Continual predict:  53%|█████▎    | 525/997 [09:55<06:44,  1.17it/s]
Continual predict:  53%|█████▎    | 526/997 [09:56<07:02,  1.12it/s]
Continual predict:  53%|█████▎    | 527/997 [09:57<06:56,  1.13it/s]
Continual predict:  53%|█████▎    | 528/997 [09:58<08:22,  1.07s/it]
Continual predict:  53%|█████▎    | 529/997 [09:59<07:37,  1.02it/s]
Continual predict:  53%|█████▎    | 530/997 [10:00<08:10,  1.05s/it]
Continual predict:  53%|█████▎    | 531/997 [10:02<08:47,  1.13s/it]
Continual predict:  53%|█████▎    | 532/997 [10:03<09:41,  1.25s/it]
Continual predict:  53%|█████▎    | 533/997 [10:04<07:50,  1.01s/it]
Continual predict:  54%|█████▎    | 534/997 [10:05<07:26,  1.04it/s]
Continual predict:  54%|█████▎    | 535/997 [10:05<07:23,  1.04it/s]
Continual predict:  54%|█████▍    | 536/997 [10:06<06:02,  1.27it/s]
Continual predict:  54%|█████▍    | 537/997 [10:07<06:17,  1.22it/s]
Continual predict:  54%|█████▍    | 538/997 [10:08<06:19,  1.21it/s]
Continual predict:  54%|█████▍    | 539/997 [10:09<06:50,  1.11it/s]
Continual predict:  54%|█████▍    | 540/997 [10:10<07:47,  1.02s/it]
Continual predict:  54%|█████▍    | 541/997 [10:11<06:49,  1.11it/s]
Continual predict:  54%|█████▍    | 542/997 [10:11<06:02,  1.25it/s]
Continual predict:  54%|█████▍    | 543/997 [10:12<06:08,  1.23it/s]
Continual predict:  55%|█████▍    | 544/997 [10:13<07:05,  1.07it/s]
Continual predict:  55%|█████▍    | 545/997 [10:15<08:59,  1.19s/it]
Continual predict:  55%|█████▍    | 546/997 [10:16<08:10,  1.09s/it]
Continual predict:  55%|█████▍    | 547/997 [10:17<07:55,  1.06s/it]
Continual predict:  55%|█████▍    | 548/997 [10:18<07:38,  1.02s/it]
Continual predict:  55%|█████▌    | 549/997 [10:19<07:45,  1.04s/it]
Continual predict:  55%|█████▌    | 550/997 [10:20<07:57,  1.07s/it]
Continual predict:  55%|█████▌    | 551/997 [10:20<06:42,  1.11it/s]
Continual predict:  55%|█████▌    | 552/997 [10:21<06:09,  1.21it/s]
Continual predict:  55%|█████▌    | 553/997 [10:23<07:29,  1.01s/it]
Continual predict:  56%|█████▌    | 554/997 [10:24<07:56,  1.08s/it]
Continual predict:  56%|█████▌    | 555/997 [10:25<08:03,  1.09s/it]
Continual predict:  56%|█████▌    | 556/997 [10:26<07:08,  1.03it/s]
Continual predict:  56%|█████▌    | 557/997 [10:27<08:30,  1.16s/it]
Continual predict:  56%|█████▌    | 558/997 [10:28<07:10,  1.02it/s]
Continual predict:  56%|█████▌    | 559/997 [10:29<07:40,  1.05s/it]
Continual predict:  56%|█████▌    | 560/997 [10:30<07:59,  1.10s/it]
Continual predict:  56%|█████▋    | 561/997 [10:32<08:44,  1.20s/it]
Continual predict:  56%|█████▋    | 562/997 [10:33<09:08,  1.26s/it]
Continual predict:  56%|█████▋    | 563/997 [10:35<09:43,  1.34s/it]
Continual predict:  57%|█████▋    | 564/997 [10:36<09:51,  1.37s/it]
Continual predict:  57%|█████▋    | 565/997 [10:37<09:48,  1.36s/it]
Continual predict:  57%|█████▋    | 566/997 [10:38<08:38,  1.20s/it]
Continual predict:  57%|█████▋    | 567/997 [10:39<08:47,  1.23s/it]
Continual predict:  57%|█████▋    | 568/997 [10:40<07:56,  1.11s/it]
Continual predict:  57%|█████▋    | 569/997 [10:41<07:45,  1.09s/it]
Continual predict:  57%|█████▋    | 570/997 [10:42<06:23,  1.11it/s]
Continual predict:  57%|█████▋    | 571/997 [10:43<06:28,  1.10it/s]
Continual predict:  57%|█████▋    | 572/997 [10:43<05:43,  1.24it/s]
Continual predict:  57%|█████▋    | 573/997 [10:45<06:38,  1.06it/s]
Continual predict:  58%|█████▊    | 574/997 [10:45<06:05,  1.16it/s]
Continual predict:  58%|█████▊    | 575/997 [10:46<06:35,  1.07it/s]
Continual predict:  58%|█████▊    | 576/997 [10:47<05:53,  1.19it/s]
Continual predict:  58%|█████▊    | 577/997 [10:48<06:05,  1.15it/s]
Continual predict:  58%|█████▊    | 578/997 [10:48<05:21,  1.30it/s]
Continual predict:  58%|█████▊    | 579/997 [10:49<05:48,  1.20it/s]
Continual predict:  58%|█████▊    | 580/997 [10:51<06:59,  1.01s/it]
Continual predict:  58%|█████▊    | 581/997 [10:52<06:15,  1.11it/s]
Continual predict:  58%|█████▊    | 582/997 [10:52<04:40,  1.48it/s]
Continual predict:  58%|█████▊    | 583/997 [10:52<03:34,  1.93it/s]
Continual predict:  59%|█████▊    | 584/997 [10:52<02:49,  2.44it/s]
Continual predict:  59%|█████▊    | 585/997 [10:52<02:16,  3.01it/s]
Continual predict:  59%|█████▉    | 586/997 [10:52<01:53,  3.61it/s]
Continual predict:  59%|█████▉    | 587/997 [10:52<01:38,  4.18it/s]
Continual predict:  59%|█████▉    | 588/997 [10:53<01:27,  4.69it/s]
Continual predict:  59%|█████▉    | 589/997 [10:53<01:19,  5.14it/s]
Continual predict:  59%|█████▉    | 590/997 [10:53<01:13,  5.53it/s]
Continual predict:  59%|█████▉    | 591/997 [10:53<01:09,  5.82it/s]
Continual predict:  59%|█████▉    | 592/997 [10:53<01:07,  6.04it/s]
Continual predict:  59%|█████▉    | 593/997 [10:53<01:04,  6.23it/s]
Continual predict:  60%|█████▉    | 594/997 [10:53<01:03,  6.37it/s]
Continual predict:  60%|█████▉    | 595/997 [10:54<01:02,  6.47it/s]
Continual predict:  60%|█████▉    | 596/997 [10:54<01:01,  6.55it/s]
Continual predict:  60%|█████▉    | 597/997 [10:54<01:00,  6.61it/s]
Continual predict:  60%|█████▉    | 598/997 [10:54<01:00,  6.62it/s]
Continual predict:  60%|██████    | 599/997 [10:54<01:00,  6.61it/s]
Continual predict:  60%|██████    | 600/997 [10:54<00:59,  6.65it/s]
Continual predict:  60%|██████    | 601/997 [10:55<00:59,  6.64it/s]
Continual predict:  60%|██████    | 602/997 [10:55<00:59,  6.63it/s]
Continual predict:  60%|██████    | 603/997 [10:55<00:59,  6.62it/s]
Continual predict:  61%|██████    | 604/997 [10:55<00:59,  6.62it/s]
Continual predict:  61%|██████    | 605/997 [10:55<00:59,  6.61it/s]
Continual predict:  61%|██████    | 606/997 [10:55<00:59,  6.60it/s]
Continual predict:  61%|██████    | 607/997 [10:55<00:58,  6.65it/s]
Continual predict:  61%|██████    | 608/997 [10:56<00:58,  6.68it/s]
Continual predict:  61%|██████    | 609/997 [10:56<00:58,  6.67it/s]
Continual predict:  61%|██████    | 610/997 [10:56<00:58,  6.62it/s]
Continual predict:  61%|██████▏   | 611/997 [10:56<00:58,  6.59it/s]
Continual predict:  61%|██████▏   | 612/997 [10:56<00:59,  6.42it/s]
Continual predict:  61%|██████▏   | 613/997 [10:56<00:59,  6.48it/s]
Continual predict:  62%|██████▏   | 614/997 [10:56<00:58,  6.55it/s]
Continual predict:  62%|██████▏   | 615/997 [10:57<00:58,  6.57it/s]
Continual predict:  62%|██████▏   | 616/997 [10:57<00:57,  6.58it/s]
Continual predict:  62%|██████▏   | 617/997 [10:57<00:57,  6.63it/s]
Continual predict:  62%|██████▏   | 618/997 [10:57<00:57,  6.63it/s]
Continual predict:  62%|██████▏   | 619/997 [10:57<00:57,  6.62it/s]
Continual predict:  62%|██████▏   | 620/997 [10:57<00:57,  6.61it/s]
Continual predict:  62%|██████▏   | 621/997 [10:58<00:56,  6.61it/s]
Continual predict:  62%|██████▏   | 622/997 [10:58<00:56,  6.61it/s]
Continual predict:  62%|██████▏   | 623/997 [10:58<00:56,  6.65it/s]
Continual predict:  63%|██████▎   | 624/997 [10:58<00:55,  6.68it/s]
Continual predict:  63%|██████▎   | 625/997 [10:58<00:55,  6.65it/s]
Continual predict:  63%|██████▎   | 626/997 [10:58<00:55,  6.67it/s]
Continual predict:  63%|██████▎   | 627/997 [10:58<00:55,  6.65it/s]
Continual predict:  63%|██████▎   | 628/997 [10:59<00:55,  6.67it/s]
Continual predict:  63%|██████▎   | 629/997 [10:59<00:55,  6.69it/s]
Continual predict:  63%|██████▎   | 630/997 [10:59<00:55,  6.67it/s]
Continual predict:  63%|██████▎   | 631/997 [10:59<00:55,  6.65it/s]
Continual predict:  63%|██████▎   | 632/997 [10:59<00:54,  6.64it/s]
Continual predict:  63%|██████▎   | 633/997 [10:59<00:54,  6.67it/s]
Continual predict:  64%|██████▎   | 634/997 [10:59<00:54,  6.65it/s]
Continual predict:  64%|██████▎   | 635/997 [11:00<00:54,  6.68it/s]
Continual predict:  64%|██████▍   | 636/997 [11:00<00:54,  6.66it/s]
Continual predict:  64%|██████▍   | 637/997 [11:00<00:54,  6.65it/s]
Continual predict:  64%|██████▍   | 638/997 [11:00<00:54,  6.63it/s]
Continual predict:  64%|██████▍   | 639/997 [11:00<00:54,  6.62it/s]
Continual predict:  64%|██████▍   | 640/997 [11:00<00:53,  6.66it/s]
Continual predict:  64%|██████▍   | 641/997 [11:01<00:53,  6.68it/s]
Continual predict:  64%|██████▍   | 642/997 [11:01<00:53,  6.67it/s]
Continual predict:  64%|██████▍   | 643/997 [11:01<00:53,  6.65it/s]
Continual predict:  65%|██████▍   | 644/997 [11:01<00:52,  6.67it/s]
Continual predict:  65%|██████▍   | 645/997 [11:01<00:52,  6.66it/s]
Continual predict:  65%|██████▍   | 646/997 [11:01<00:52,  6.64it/s]
Continual predict:  65%|██████▍   | 647/997 [11:01<00:52,  6.63it/s]
Continual predict:  65%|██████▍   | 648/997 [11:02<00:52,  6.67it/s]
Continual predict:  65%|██████▌   | 649/997 [11:02<00:52,  6.65it/s]
Continual predict:  65%|██████▌   | 650/997 [11:02<00:51,  6.68it/s]
Continual predict:  65%|██████▌   | 651/997 [11:02<00:51,  6.70it/s]
Continual predict:  65%|██████▌   | 652/997 [11:02<00:51,  6.70it/s]
Continual predict:  65%|██████▌   | 653/997 [11:02<00:51,  6.71it/s]
Continual predict:  66%|██████▌   | 654/997 [11:02<00:51,  6.72it/s]
Continual predict:  66%|██████▌   | 655/997 [11:03<00:50,  6.73it/s]
Continual predict:  66%|██████▌   | 656/997 [11:03<00:50,  6.69it/s]
Continual predict:  66%|██████▌   | 657/997 [11:03<00:51,  6.66it/s]
Continual predict:  66%|██████▌   | 658/997 [11:03<00:51,  6.65it/s]
Continual predict:  66%|██████▌   | 659/997 [11:03<00:50,  6.67it/s]
Continual predict:  66%|██████▌   | 660/997 [11:03<00:50,  6.66it/s]
Continual predict:  66%|██████▋   | 661/997 [11:04<00:50,  6.64it/s]
Continual predict:  66%|██████▋   | 662/997 [11:04<00:50,  6.63it/s]
Continual predict:  66%|██████▋   | 663/997 [11:04<00:50,  6.62it/s]
Continual predict:  67%|██████▋   | 664/997 [11:04<00:50,  6.61it/s]
Continual predict:  67%|██████▋   | 665/997 [11:04<00:50,  6.62it/s]
Continual predict:  67%|██████▋   | 666/997 [11:04<00:49,  6.65it/s]
Continual predict:  67%|██████▋   | 667/997 [11:04<00:49,  6.64it/s]
Continual predict:  67%|██████▋   | 668/997 [11:05<00:49,  6.63it/s]
Continual predict:  67%|██████▋   | 669/997 [11:05<00:49,  6.66it/s]
Continual predict:  67%|██████▋   | 670/997 [11:05<00:49,  6.64it/s]
Continual predict:  67%|██████▋   | 671/997 [11:05<00:49,  6.63it/s]
Continual predict:  67%|██████▋   | 672/997 [11:05<00:49,  6.63it/s]
Continual predict:  68%|██████▊   | 673/997 [11:05<00:48,  6.65it/s]
Continual predict:  68%|██████▊   | 674/997 [11:06<00:48,  6.64it/s]
Continual predict:  68%|██████▊   | 675/997 [11:06<00:48,  6.67it/s]
Continual predict:  68%|██████▊   | 676/997 [11:06<00:47,  6.69it/s]
Continual predict:  68%|██████▊   | 677/997 [11:06<00:47,  6.71it/s]
Continual predict:  68%|██████▊   | 678/997 [11:06<00:47,  6.68it/s]
Continual predict:  68%|██████▊   | 679/997 [11:06<00:47,  6.64it/s]
Continual predict:  68%|██████▊   | 680/997 [11:06<00:48,  6.59it/s]
Continual predict:  68%|██████▊   | 681/997 [11:07<00:48,  6.53it/s]
Continual predict:  68%|██████▊   | 682/997 [11:08<02:22,  2.21it/s]
Continual predict:  69%|██████▊   | 683/997 [11:10<05:29,  1.05s/it]
Continual predict:  69%|██████▊   | 684/997 [11:12<06:26,  1.24s/it]
Continual predict:  69%|██████▊   | 685/997 [11:13<05:40,  1.09s/it]
Continual predict:  69%|██████▉   | 686/997 [11:15<07:38,  1.47s/it]
Continual predict:  69%|██████▉   | 687/997 [11:17<07:47,  1.51s/it]
Continual predict:  69%|██████▉   | 688/997 [11:21<12:44,  2.47s/it]
Continual predict:  69%|██████▉   | 689/997 [11:22<10:12,  1.99s/it]
Continual predict:  69%|██████▉   | 690/997 [11:26<13:01,  2.55s/it]
Continual predict:  69%|██████▉   | 691/997 [11:27<10:42,  2.10s/it]
Continual predict:  69%|██████▉   | 692/997 [11:28<08:34,  1.69s/it]
Continual predict:  70%|██████▉   | 693/997 [11:31<10:27,  2.06s/it]
Continual predict:  70%|██████▉   | 694/997 [11:32<08:54,  1.76s/it]
Continual predict:  70%|██████▉   | 695/997 [11:34<09:21,  1.86s/it]
Continual predict:  70%|██████▉   | 696/997 [11:35<08:28,  1.69s/it]
Continual predict:  70%|██████▉   | 697/997 [11:36<06:56,  1.39s/it]
Continual predict:  70%|███████   | 698/997 [11:38<07:40,  1.54s/it]
Continual predict:  70%|███████   | 699/997 [11:39<07:25,  1.50s/it]
Continual predict:  70%|███████   | 700/997 [11:41<07:19,  1.48s/it]
Continual predict:  70%|███████   | 701/997 [11:42<07:20,  1.49s/it]
Continual predict:  70%|███████   | 702/997 [11:43<06:47,  1.38s/it]
Continual predict:  71%|███████   | 703/997 [11:45<07:16,  1.49s/it]
Continual predict:  71%|███████   | 704/997 [11:46<06:44,  1.38s/it]
Continual predict:  71%|███████   | 705/997 [11:49<09:11,  1.89s/it]
Continual predict:  71%|███████   | 706/997 [11:51<08:50,  1.82s/it]
Continual predict:  71%|███████   | 707/997 [11:52<08:14,  1.71s/it]
Continual predict:  71%|███████   | 708/997 [11:55<10:02,  2.09s/it]
Continual predict:  71%|███████   | 709/997 [11:57<09:16,  1.93s/it]
Continual predict:  71%|███████   | 710/997 [11:58<08:52,  1.85s/it]
Continual predict:  71%|███████▏  | 711/997 [12:00<08:24,  1.77s/it]
Continual predict:  71%|███████▏  | 712/997 [12:02<08:01,  1.69s/it]
Continual predict:  72%|███████▏  | 713/997 [12:03<07:31,  1.59s/it]
Continual predict:  72%|███████▏  | 714/997 [12:05<08:01,  1.70s/it]
Continual predict:  72%|███████▏  | 715/997 [12:07<08:33,  1.82s/it]
Continual predict:  72%|███████▏  | 716/997 [12:08<08:05,  1.73s/it]
Continual predict:  72%|███████▏  | 717/997 [12:10<07:53,  1.69s/it]
Continual predict:  72%|███████▏  | 718/997 [12:12<07:42,  1.66s/it]
Continual predict:  72%|███████▏  | 719/997 [12:14<08:01,  1.73s/it]
Continual predict:  72%|███████▏  | 720/997 [12:15<07:16,  1.57s/it]
Continual predict:  72%|███████▏  | 721/997 [12:16<07:21,  1.60s/it]
Continual predict:  72%|███████▏  | 722/997 [12:18<07:06,  1.55s/it]
Continual predict:  73%|███████▎  | 723/997 [12:19<06:55,  1.52s/it]
Continual predict:  73%|███████▎  | 724/997 [12:21<07:05,  1.56s/it]
Continual predict:  73%|███████▎  | 725/997 [12:23<07:06,  1.57s/it]
Continual predict:  73%|███████▎  | 726/997 [12:24<06:54,  1.53s/it]
Continual predict:  73%|███████▎  | 727/997 [12:25<06:45,  1.50s/it]
Continual predict:  73%|███████▎  | 728/997 [12:27<07:16,  1.62s/it]
Continual predict:  73%|███████▎  | 729/997 [12:29<07:23,  1.66s/it]
Continual predict:  73%|███████▎  | 730/997 [12:31<07:35,  1.70s/it]
Continual predict:  73%|███████▎  | 731/997 [12:33<07:56,  1.79s/it]
Continual predict:  73%|███████▎  | 732/997 [12:34<07:26,  1.68s/it]
Continual predict:  74%|███████▎  | 733/997 [12:35<06:40,  1.52s/it]
Continual predict:  74%|███████▎  | 734/997 [12:38<07:32,  1.72s/it]
Continual predict:  74%|███████▎  | 735/997 [12:40<08:08,  1.87s/it]
Continual predict:  74%|███████▍  | 736/997 [12:42<08:17,  1.90s/it]
Continual predict:  74%|███████▍  | 737/997 [12:43<07:27,  1.72s/it]
Continual predict:  74%|███████▍  | 738/997 [12:45<07:27,  1.73s/it]
Continual predict:  74%|███████▍  | 739/997 [12:47<07:52,  1.83s/it]
Continual predict:  74%|███████▍  | 740/997 [12:49<07:42,  1.80s/it]
Continual predict:  74%|███████▍  | 741/997 [12:50<07:42,  1.80s/it]
Continual predict:  74%|███████▍  | 742/997 [12:52<07:31,  1.77s/it]
Continual predict:  75%|███████▍  | 743/997 [12:54<07:21,  1.74s/it]
Continual predict:  75%|███████▍  | 744/997 [12:56<07:24,  1.76s/it]
Continual predict:  75%|███████▍  | 745/997 [12:57<07:32,  1.80s/it]
Continual predict:  75%|███████▍  | 746/997 [12:59<07:26,  1.78s/it]
Continual predict:  75%|███████▍  | 747/997 [13:01<07:21,  1.76s/it]
Continual predict:  75%|███████▌  | 748/997 [13:03<07:36,  1.83s/it]
Continual predict:  75%|███████▌  | 749/997 [13:05<07:56,  1.92s/it]
Continual predict:  75%|███████▌  | 750/997 [13:07<07:20,  1.78s/it]
Continual predict:  75%|███████▌  | 751/997 [13:09<07:37,  1.86s/it]
Continual predict:  75%|███████▌  | 752/997 [13:10<06:53,  1.69s/it]
Continual predict:  76%|███████▌  | 753/997 [13:11<06:33,  1.61s/it]
Continual predict:  76%|███████▌  | 754/997 [13:13<06:02,  1.49s/it]
Continual predict:  76%|███████▌  | 755/997 [13:15<06:39,  1.65s/it]
Continual predict:  76%|███████▌  | 756/997 [13:16<06:26,  1.61s/it]
Continual predict:  76%|███████▌  | 757/997 [13:17<06:08,  1.53s/it]
Continual predict:  76%|███████▌  | 758/997 [13:19<05:59,  1.50s/it]
Continual predict:  76%|███████▌  | 759/997 [13:20<06:03,  1.53s/it]
Continual predict:  76%|███████▌  | 760/997 [13:22<05:55,  1.50s/it]
Continual predict:  76%|███████▋  | 761/997 [13:24<07:12,  1.83s/it]
Continual predict:  76%|███████▋  | 762/997 [13:27<07:47,  1.99s/it]
Continual predict:  77%|███████▋  | 763/997 [13:29<07:39,  1.96s/it]
Continual predict:  77%|███████▋  | 764/997 [13:31<07:26,  1.92s/it]
Continual predict:  77%|███████▋  | 765/997 [13:32<06:35,  1.70s/it]
Continual predict:  77%|███████▋  | 766/997 [13:33<06:36,  1.72s/it]
Continual predict:  77%|███████▋  | 767/997 [13:35<06:27,  1.69s/it]
Continual predict:  77%|███████▋  | 768/997 [13:37<06:13,  1.63s/it]
Continual predict:  77%|███████▋  | 769/997 [13:38<06:02,  1.59s/it]
Continual predict:  77%|███████▋  | 770/997 [13:40<06:00,  1.59s/it]
Continual predict:  77%|███████▋  | 771/997 [13:41<05:33,  1.48s/it]
Continual predict:  77%|███████▋  | 772/997 [13:42<05:13,  1.39s/it]
Continual predict:  78%|███████▊  | 773/997 [13:44<06:06,  1.63s/it]
Continual predict:  78%|███████▊  | 774/997 [13:46<06:03,  1.63s/it]
Continual predict:  78%|███████▊  | 775/997 [13:47<05:48,  1.57s/it]
Continual predict:  78%|███████▊  | 776/997 [13:49<06:09,  1.67s/it]
Continual predict:  78%|███████▊  | 777/997 [13:51<05:57,  1.62s/it]
Continual predict:  78%|███████▊  | 778/997 [13:53<06:03,  1.66s/it]
Continual predict:  78%|███████▊  | 779/997 [13:54<05:56,  1.63s/it]
Continual predict:  78%|███████▊  | 780/997 [13:56<06:02,  1.67s/it]
Continual predict:  78%|███████▊  | 781/997 [13:58<06:05,  1.69s/it]
Continual predict:  78%|███████▊  | 782/997 [13:59<05:41,  1.59s/it]
Continual predict:  79%|███████▊  | 783/997 [14:01<05:54,  1.66s/it]
Continual predict:  79%|███████▊  | 784/997 [14:02<05:53,  1.66s/it]
Continual predict:  79%|███████▊  | 785/997 [14:04<05:33,  1.57s/it]
Continual predict:  79%|███████▉  | 786/997 [14:05<05:37,  1.60s/it]
Continual predict:  79%|███████▉  | 787/997 [14:07<05:51,  1.67s/it]
Continual predict:  79%|███████▉  | 788/997 [14:08<05:19,  1.53s/it]
Continual predict:  79%|███████▉  | 789/997 [14:10<05:31,  1.59s/it]
Continual predict:  79%|███████▉  | 790/997 [14:13<06:16,  1.82s/it]
Continual predict:  79%|███████▉  | 791/997 [14:14<06:09,  1.80s/it]
Continual predict:  79%|███████▉  | 792/997 [14:16<05:41,  1.67s/it]
Continual predict:  80%|███████▉  | 793/997 [14:17<05:39,  1.66s/it]
Continual predict:  80%|███████▉  | 794/997 [14:20<06:14,  1.85s/it]
Continual predict:  80%|███████▉  | 795/997 [14:21<05:43,  1.70s/it]
Continual predict:  80%|███████▉  | 796/997 [14:23<05:35,  1.67s/it]
Continual predict:  80%|███████▉  | 797/997 [14:24<05:33,  1.67s/it]
Continual predict:  80%|████████  | 798/997 [14:26<05:54,  1.78s/it]
Continual predict:  80%|████████  | 799/997 [14:33<10:48,  3.28s/it]
Continual predict:  80%|████████  | 800/997 [14:35<09:43,  2.96s/it]
Continual predict:  80%|████████  | 801/997 [14:37<08:02,  2.46s/it]
Continual predict:  80%|████████  | 802/997 [14:38<06:58,  2.15s/it]
Continual predict:  81%|████████  | 803/997 [14:39<05:52,  1.82s/it]
Continual predict:  81%|████████  | 804/997 [14:43<07:47,  2.42s/it]
Continual predict:  81%|████████  | 805/997 [14:46<08:51,  2.77s/it]
Continual predict:  81%|████████  | 806/997 [14:49<08:09,  2.56s/it]
Continual predict:  81%|████████  | 807/997 [14:50<06:59,  2.21s/it]
Continual predict:  81%|████████  | 808/997 [14:52<06:27,  2.05s/it]
Continual predict:  81%|████████  | 809/997 [14:53<05:54,  1.89s/it]
Continual predict:  81%|████████  | 810/997 [14:55<06:15,  2.01s/it]
Continual predict:  81%|████████▏ | 811/997 [14:57<05:56,  1.91s/it]
Continual predict:  81%|████████▏ | 812/997 [15:01<08:00,  2.60s/it]
Continual predict:  82%|████████▏ | 813/997 [15:09<12:40,  4.13s/it]
Continual predict:  82%|████████▏ | 814/997 [15:11<10:41,  3.51s/it]
Continual predict:  82%|████████▏ | 815/997 [15:13<09:37,  3.18s/it]
Continual predict:  82%|████████▏ | 816/997 [15:16<09:02,  3.00s/it]
Continual predict:  82%|████████▏ | 817/997 [15:18<07:38,  2.55s/it]
Continual predict:  82%|████████▏ | 818/997 [15:19<06:23,  2.14s/it]
Continual predict:  82%|████████▏ | 819/997 [15:21<06:26,  2.17s/it]
Continual predict:  82%|████████▏ | 820/997 [15:23<05:58,  2.03s/it]
Continual predict:  82%|████████▏ | 821/997 [15:25<05:51,  2.00s/it]
Continual predict:  82%|████████▏ | 822/997 [15:27<06:30,  2.23s/it]
Continual predict:  83%|████████▎ | 823/997 [15:31<07:40,  2.65s/it]
Continual predict:  83%|████████▎ | 824/997 [15:41<14:11,  4.92s/it]
Continual predict:  83%|████████▎ | 825/997 [15:43<11:22,  3.97s/it]
Continual predict:  83%|████████▎ | 826/997 [15:47<11:17,  3.96s/it]
Continual predict:  83%|████████▎ | 827/997 [15:48<08:50,  3.12s/it]
Continual predict:  83%|████████▎ | 828/997 [15:52<09:45,  3.46s/it]
Continual predict:  83%|████████▎ | 829/997 [15:58<11:21,  4.06s/it]
Continual predict:  83%|████████▎ | 830/997 [16:00<09:35,  3.45s/it]
Continual predict:  83%|████████▎ | 831/997 [16:11<16:21,  5.91s/it]
Continual predict:  83%|████████▎ | 832/997 [16:13<12:37,  4.59s/it]
Continual predict:  84%|████████▎ | 833/997 [16:14<09:32,  3.49s/it]
Continual predict:  84%|████████▎ | 834/997 [16:16<08:20,  3.07s/it]
Continual predict:  84%|████████▍ | 835/997 [16:17<06:54,  2.56s/it]
Continual predict:  84%|████████▍ | 836/997 [16:20<06:55,  2.58s/it]
Continual predict:  84%|████████▍ | 837/997 [16:25<08:46,  3.29s/it]
Continual predict:  84%|████████▍ | 838/997 [16:27<07:50,  2.96s/it]
Continual predict:  84%|████████▍ | 839/997 [16:34<10:48,  4.10s/it]
Continual predict:  84%|████████▍ | 840/997 [16:39<11:10,  4.27s/it]
Continual predict:  84%|████████▍ | 841/997 [16:40<09:00,  3.46s/it]
Continual predict:  84%|████████▍ | 842/997 [16:43<08:26,  3.27s/it]
Continual predict:  85%|████████▍ | 843/997 [16:45<07:20,  2.86s/it]
Continual predict:  85%|████████▍ | 844/997 [16:50<08:56,  3.51s/it]
Continual predict:  85%|████████▍ | 845/997 [16:54<09:10,  3.62s/it]
Continual predict:  85%|████████▍ | 846/997 [16:55<07:31,  2.99s/it]
Continual predict:  85%|████████▍ | 847/997 [16:56<05:45,  2.30s/it]
Continual predict:  85%|████████▌ | 848/997 [16:58<05:12,  2.10s/it]
Continual predict:  85%|████████▌ | 849/997 [16:59<04:31,  1.84s/it]
Continual predict:  85%|████████▌ | 850/997 [17:02<05:29,  2.24s/it]
Continual predict:  85%|████████▌ | 851/997 [17:04<05:16,  2.17s/it]
Continual predict:  85%|████████▌ | 852/997 [17:05<04:35,  1.90s/it]
Continual predict:  86%|████████▌ | 853/997 [17:15<10:04,  4.20s/it]
Continual predict:  86%|████████▌ | 854/997 [17:26<15:20,  6.44s/it]
Continual predict:  86%|████████▌ | 855/997 [17:29<12:20,  5.21s/it]
Continual predict:  86%|████████▌ | 856/997 [17:31<10:07,  4.31s/it]
Continual predict:  86%|████████▌ | 857/997 [17:33<08:17,  3.55s/it]
Continual predict:  86%|████████▌ | 858/997 [17:35<07:07,  3.08s/it]
Continual predict:  86%|████████▌ | 859/997 [17:39<07:58,  3.47s/it]
Continual predict:  86%|████████▋ | 860/997 [17:43<08:18,  3.64s/it]
Continual predict:  86%|████████▋ | 861/997 [17:55<13:42,  6.04s/it]
Continual predict:  86%|████████▋ | 862/997 [17:55<09:55,  4.41s/it]
Continual predict:  87%|████████▋ | 863/997 [17:58<08:17,  3.71s/it]
Continual predict:  87%|████████▋ | 864/997 [17:59<06:50,  3.09s/it]
Continual predict:  87%|████████▋ | 865/997 [18:01<06:13,  2.83s/it]
Continual predict:  87%|████████▋ | 866/997 [18:03<05:16,  2.42s/it]
Continual predict:  87%|████████▋ | 867/997 [18:05<05:13,  2.41s/it]
Continual predict:  87%|████████▋ | 868/997 [18:08<05:43,  2.66s/it]
Continual predict:  87%|████████▋ | 869/997 [18:12<06:16,  2.94s/it]
Continual predict:  87%|████████▋ | 870/997 [18:13<05:07,  2.42s/it]
Continual predict:  87%|████████▋ | 871/997 [18:15<04:34,  2.18s/it]
Continual predict:  87%|████████▋ | 872/997 [18:20<06:27,  3.10s/it]
Continual predict:  88%|████████▊ | 873/997 [18:22<05:52,  2.85s/it]
Continual predict:  88%|████████▊ | 874/997 [18:24<04:56,  2.41s/it]
Continual predict:  88%|████████▊ | 875/997 [18:26<04:36,  2.26s/it]
Continual predict:  88%|████████▊ | 876/997 [18:27<04:00,  1.99s/it]
Continual predict:  88%|████████▊ | 877/997 [18:29<04:11,  2.10s/it]
Continual predict:  88%|████████▊ | 878/997 [18:30<03:03,  1.54s/it]
Continual predict:  88%|████████▊ | 879/997 [18:33<03:57,  2.01s/it]
Continual predict:  88%|████████▊ | 880/997 [18:34<03:41,  1.90s/it]
Continual predict:  88%|████████▊ | 881/997 [18:38<04:32,  2.35s/it]
Continual predict:  88%|████████▊ | 882/997 [18:40<04:25,  2.31s/it]
Continual predict:  89%|████████▊ | 883/997 [18:42<03:58,  2.09s/it]
Continual predict:  89%|████████▊ | 884/997 [18:45<04:44,  2.51s/it]
Continual predict:  89%|████████▉ | 885/997 [18:50<06:05,  3.26s/it]
Continual predict:  89%|████████▉ | 886/997 [18:52<05:22,  2.91s/it]
Continual predict:  89%|████████▉ | 887/997 [18:55<05:24,  2.95s/it]
Continual predict:  89%|████████▉ | 888/997 [18:57<04:50,  2.67s/it]
Continual predict:  89%|████████▉ | 889/997 [18:59<04:32,  2.52s/it]
Continual predict:  89%|████████▉ | 890/997 [19:02<04:34,  2.56s/it]
Continual predict:  89%|████████▉ | 891/997 [19:05<04:41,  2.65s/it]
Continual predict:  89%|████████▉ | 892/997 [19:08<04:35,  2.63s/it]
Continual predict:  90%|████████▉ | 893/997 [19:10<04:37,  2.67s/it]
Continual predict:  90%|████████▉ | 894/997 [19:12<04:07,  2.41s/it]
Continual predict:  90%|████████▉ | 895/997 [19:17<05:08,  3.02s/it]
Continual predict:  90%|████████▉ | 896/997 [19:18<04:31,  2.69s/it]
Continual predict:  90%|████████▉ | 897/997 [19:23<05:20,  3.21s/it]
Continual predict:  90%|█████████ | 898/997 [19:23<03:47,  2.30s/it]
Continual predict:  90%|█████████ | 899/997 [19:23<02:44,  1.68s/it]
Continual predict:  90%|█████████ | 900/997 [19:24<02:02,  1.27s/it]
Continual predict:  90%|█████████ | 901/997 [19:24<01:33,  1.02it/s]
Continual predict:  90%|█████████ | 902/997 [19:24<01:08,  1.39it/s]
Continual predict:  91%|█████████ | 903/997 [19:24<00:50,  1.85it/s]
Continual predict:  91%|█████████ | 904/997 [19:24<00:41,  2.23it/s]
Continual predict:  91%|█████████ | 905/997 [19:25<00:35,  2.59it/s]
Continual predict:  91%|█████████ | 906/997 [19:25<00:27,  3.27it/s]
Continual predict:  91%|█████████ | 907/997 [19:25<00:25,  3.50it/s]
Continual predict:  91%|█████████ | 908/997 [19:25<00:22,  3.94it/s]
Continual predict:  91%|█████████ | 909/997 [19:25<00:21,  4.02it/s]
Continual predict:  91%|█████████▏| 910/997 [19:26<00:19,  4.38it/s]
Continual predict:  91%|█████████▏| 911/997 [19:26<00:18,  4.67it/s]
Continual predict:  91%|█████████▏| 912/997 [19:26<00:15,  5.39it/s]
Continual predict:  92%|█████████▏| 913/997 [19:26<00:15,  5.44it/s]
Continual predict:  92%|█████████▏| 914/997 [19:26<00:18,  4.57it/s]
Continual predict:  92%|█████████▏| 915/997 [19:27<00:16,  4.83it/s]
Continual predict:  92%|█████████▏| 916/997 [19:27<00:17,  4.62it/s]
Continual predict:  92%|█████████▏| 917/997 [19:27<00:17,  4.49it/s]
Continual predict:  92%|█████████▏| 918/997 [19:27<00:15,  5.22it/s]
Continual predict:  92%|█████████▏| 919/997 [19:27<00:16,  4.87it/s]
Continual predict:  92%|█████████▏| 920/997 [19:28<00:17,  4.28it/s]
Continual predict:  92%|█████████▏| 921/997 [19:28<00:17,  4.26it/s]
Continual predict:  92%|█████████▏| 922/997 [19:28<00:14,  5.00it/s]
Continual predict:  93%|█████████▎| 923/997 [19:28<00:15,  4.72it/s]
Continual predict:  93%|█████████▎| 924/997 [19:29<00:18,  3.92it/s]
Continual predict:  93%|█████████▎| 925/997 [19:29<00:18,  4.00it/s]
Continual predict:  93%|█████████▎| 926/997 [19:29<00:14,  4.75it/s]
Continual predict:  93%|█████████▎| 927/997 [19:29<00:14,  4.98it/s]
Continual predict:  93%|█████████▎| 928/997 [19:29<00:12,  5.67it/s]
Continual predict:  93%|█████████▎| 929/997 [19:29<00:13,  5.11it/s]
Continual predict:  93%|█████████▎| 930/997 [19:30<00:13,  4.79it/s]
Continual predict:  93%|█████████▎| 931/997 [19:30<00:15,  4.25it/s]
Continual predict:  93%|█████████▎| 932/997 [19:30<00:14,  4.58it/s]
Continual predict:  94%|█████████▎| 933/997 [19:31<00:15,  4.14it/s]
Continual predict:  94%|█████████▎| 934/997 [19:31<00:12,  4.86it/s]
Continual predict:  94%|█████████▍| 935/997 [19:31<00:14,  4.26it/s]
Continual predict:  94%|█████████▍| 936/997 [19:31<00:15,  3.95it/s]
Continual predict:  94%|█████████▍| 937/997 [19:32<00:15,  3.76it/s]
Continual predict:  94%|█████████▍| 938/997 [19:32<00:13,  4.50it/s]
Continual predict:  94%|█████████▍| 939/997 [19:32<00:13,  4.38it/s]
Continual predict:  94%|█████████▍| 940/997 [19:32<00:12,  4.68it/s]
Continual predict:  94%|█████████▍| 941/997 [19:32<00:14,  3.90it/s]
Continual predict:  94%|█████████▍| 942/997 [19:33<00:13,  3.99it/s]
Continual predict:  95%|█████████▍| 943/997 [19:33<00:12,  4.35it/s]
Continual predict:  95%|█████████▍| 944/997 [19:33<00:12,  4.32it/s]
Continual predict:  95%|█████████▍| 945/997 [19:33<00:10,  4.83it/s]
Continual predict:  95%|█████████▍| 946/997 [19:33<00:10,  4.77it/s]
Continual predict:  95%|█████████▍| 947/997 [19:34<00:14,  3.52it/s]
Continual predict:  95%|█████████▌| 948/997 [19:34<00:13,  3.70it/s]
Continual predict:  95%|█████████▌| 949/997 [19:34<00:14,  3.39it/s]
Continual predict:  95%|█████████▌| 950/997 [19:35<00:11,  4.12it/s]
Continual predict:  95%|█████████▌| 951/997 [19:35<00:10,  4.47it/s]
Continual predict:  95%|█████████▌| 952/997 [19:35<00:08,  5.19it/s]
Continual predict:  96%|█████████▌| 953/997 [19:35<00:08,  5.31it/s]
Continual predict:  96%|█████████▌| 954/997 [19:35<00:07,  5.97it/s]
Continual predict:  96%|█████████▌| 955/997 [19:36<00:08,  4.84it/s]
Continual predict:  96%|█████████▌| 956/997 [19:36<00:07,  5.54it/s]
Continual predict:  96%|█████████▌| 957/997 [19:36<00:07,  5.54it/s]
Continual predict:  96%|█████████▌| 958/997 [19:36<00:07,  5.54it/s]
Continual predict:  96%|█████████▌| 959/997 [19:36<00:07,  5.06it/s]
Continual predict:  96%|█████████▋| 960/997 [19:36<00:06,  5.74it/s]
Continual predict:  96%|█████████▋| 961/997 [19:37<00:06,  5.70it/s]
Continual predict:  96%|█████████▋| 962/997 [19:37<00:06,  5.15it/s]
Continual predict:  97%|█████████▋| 963/997 [19:37<00:06,  5.26it/s]
Continual predict:  97%|█████████▋| 964/997 [19:37<00:05,  5.92it/s]
Continual predict:  97%|█████████▋| 965/997 [19:37<00:06,  5.27it/s]
Continual predict:  97%|█████████▋| 966/997 [19:38<00:06,  4.90it/s]
Continual predict:  97%|█████████▋| 967/997 [19:38<00:06,  4.32it/s]
Continual predict:  97%|█████████▋| 968/997 [19:38<00:07,  3.71it/s]
Continual predict:  97%|█████████▋| 969/997 [19:38<00:07,  3.60it/s]
Continual predict:  97%|█████████▋| 970/997 [19:39<00:07,  3.77it/s]
Continual predict:  97%|█████████▋| 971/997 [19:39<00:07,  3.63it/s]
Continual predict:  97%|█████████▋| 972/997 [19:39<00:06,  3.76it/s]
Continual predict:  98%|█████████▊| 973/997 [19:40<00:06,  3.87it/s]
Continual predict:  98%|█████████▊| 974/997 [19:40<00:04,  4.61it/s]
Continual predict:  98%|█████████▊| 975/997 [19:40<00:04,  4.85it/s]
Continual predict:  98%|█████████▊| 976/997 [19:40<00:04,  4.27it/s]
Continual predict:  98%|█████████▊| 977/997 [19:40<00:04,  4.56it/s]
Continual predict:  98%|█████████▊| 978/997 [19:40<00:03,  5.28it/s]
Continual predict:  98%|█████████▊| 979/997 [19:41<00:03,  4.89it/s]
Continual predict:  98%|█████████▊| 980/997 [19:41<00:03,  5.58it/s]
Continual predict:  98%|█████████▊| 981/997 [19:41<00:03,  5.08it/s]
Continual predict:  98%|█████████▊| 982/997 [19:41<00:03,  4.78it/s]
Continual predict:  99%|█████████▊| 983/997 [19:41<00:02,  5.00it/s]
Continual predict:  99%|█████████▊| 984/997 [19:42<00:02,  5.69it/s]
Continual predict:  99%|█████████▉| 985/997 [19:42<00:02,  4.33it/s]
Continual predict:  99%|█████████▉| 986/997 [19:42<00:02,  5.06it/s]
Continual predict:  99%|█████████▉| 987/997 [19:42<00:01,  5.57it/s]
Continual predict:  99%|█████████▉| 988/997 [19:42<00:01,  6.19it/s]
Continual predict:  99%|█████████▉| 989/997 [19:43<00:01,  5.43it/s]
Continual predict:  99%|█████████▉| 990/997 [19:43<00:01,  4.59it/s]
Continual predict:  99%|█████████▉| 991/997 [19:43<00:01,  5.31it/s]
Continual predict:  99%|█████████▉| 992/997 [19:43<00:00,  5.40it/s]
Continual predict: 100%|█████████▉| 993/997 [19:43<00:00,  5.44it/s]
Continual predict: 100%|█████████▉| 994/997 [19:43<00:00,  6.07it/s]
Continual predict: 100%|█████████▉| 995/997 [19:44<00:00,  6.51it/s]
Continual predict: 100%|█████████▉| 996/997 [19:44<00:00,  6.85it/s]
Continual predict: 100%|██████████| 997/997 [19:44<00:00,  6.12it/s]
Continual predict: 100%|██████████| 997/997 [19:44<00:00,  1.19s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot10=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot12=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=16]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot11=100]│
+  │  GT=task511_reddit_tifu_long_text_su  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task591_sciq_answer_generation    slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=18]│
+  │  GT=task748_glucose_reverse_cause_ev  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (997/997)  label_acc=100.0% (997/997)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                                 =       15.0
+  predict_eval_rougeL                                                   =    57.9347
+  predict_eval_rougeL_for_CL                                            =    57.9347
+  predict_eval_rougeL_for_task002_quoref_answer_generation              =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                   =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction        =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                       =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation            =    11.1335
+  predict_eval_rougeL_for_task1729_personachat_generate_next            =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                    =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification          =       91.0
+  predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization   =    21.3091
+  predict_eval_rougeL_for_task591_sciq_answer_generation                =    72.7381
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation   =    15.2912
+  predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection =    66.9039
+  predict_exact_match                                                   =    41.1234
+  predict_exact_match_for_CL                                            =    41.1234
+  predict_exact_match_for_task002_quoref_answer_generation              =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                   =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction        =       98.0
+  predict_exact_match_for_task1572_samsum_summary                       =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation            =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next            =        0.0
+  predict_exact_match_for_task181_outcome_extraction                    =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification          =       91.0
+  predict_exact_match_for_task511_reddit_tifu_long_text_summarization   =        1.0
+  predict_exact_match_for_task591_sciq_answer_generation                =       60.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation   =        0.0
+  predict_exact_match_for_task748_glucose_reverse_cause_event_detection =        3.0
+  predict_gen_len                                                       =      10.01
+  predict_global_step                                                   =        480
+  predict_rouge1                                                        =    59.6657
+  predict_rouge1_for_CL                                                 =    59.6657
+  predict_rouge1_for_task002_quoref_answer_generation                   =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation            =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                        =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction             =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                            =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation                 =    13.7971
+  predict_rouge1_for_task1729_personachat_generate_next                 =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                         =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification               =       91.0
+  predict_rouge1_for_task511_reddit_tifu_long_text_summarization        =    25.5967
+  predict_rouge1_for_task591_sciq_answer_generation                     =    73.4048
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation        =    15.2912
+  predict_rouge1_for_task748_glucose_reverse_cause_event_detection      =     68.907
+  predict_runtime                                                       = 0:19:44.37
+  predict_samples                                                       =        997
+  predict_samples_per_second                                            =      0.842
+  predict_steps_per_second                                              =      0.842
+[2026-04-30 05:44:37,881] [INFO] [launch.py:367:main] Process 1775507 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/all_results.json b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..f89731bf0fad86afadf7a9a6b24d0582cfc1e58f
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/all_results.json
@@ -0,0 +1,63 @@
+{
+  "predict_exact_match": 45.0319,
+  "predict_rouge1": 61.8839,
+  "predict_eval_rougeL": 60.3107,
+  "predict_gen_len": 9.1887,
+  "predict_runtime": 1215.3494530417956,
+  "predict_samples": 1097,
+  "predict_samples_per_second": 0.903,
+  "predict_steps_per_second": 0.903,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 45.0319,
+  "predict_rouge1_for_CL": 61.8839,
+  "predict_eval_rougeL_for_CL": 60.3107,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "predict_exact_match_for_task748_glucose_reverse_cause_event_detection": 3.0,
+  "predict_rouge1_for_task748_glucose_reverse_cause_event_detection": 68.907,
+  "predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection": 66.9039,
+  "predict_exact_match_for_task511_reddit_tifu_long_text_summarization": 1.0,
+  "predict_rouge1_for_task511_reddit_tifu_long_text_summarization": 25.5967,
+  "predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization": 21.3091,
+  "predict_exact_match_for_task591_sciq_answer_generation": 60.0,
+  "predict_rouge1_for_task591_sciq_answer_generation": 73.4048,
+  "predict_eval_rougeL_for_task591_sciq_answer_generation": 72.7381,
+  "predict_exact_match_for_task1687_sentiment140_classification": 84.0,
+  "predict_rouge1_for_task1687_sentiment140_classification": 84.0,
+  "predict_eval_rougeL_for_task1687_sentiment140_classification": 84.0,
+  "train_runtime": 2373.6732,
+  "train_samples_per_second": 6.319,
+  "train_steps_per_second": 0.202,
+  "total_flos": 5.019504126355046e+16,
+  "train_loss": 0.3057432758311431,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ecab001e99460ebf83303a50d607d0e33c681214
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fd00c3d29d44e3dfa9157f8d2cf3e751b41f4a2354c3072d31782c3b60cb400
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..351877eb50198da380acd8f71a4e55bc95226e74
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5db1ac1b9e34103b18c2f0a73686c759e4c3f3629287080156a9566cdddeb1
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..cf9938dd756cad65e028e93668e3a87e9e87ca8d
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d4171884b0486d7f141fff7533eff3324f6abde52cb8584150d164ed10acad7
+size 2154720573
diff --git a/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/train.log b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..7d4070bc465ac572e060190c966c614c98cd7141
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/train.log
@@ -0,0 +1,612 @@
+[2026-04-30 05:44:43,811] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 05:44:43,811] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29513 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1687_sentiment140_classification --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification --cur_task_id 13 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+[2026-04-30 05:44:47,063] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 05:44:47,063] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 05:44:47,063] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 05:44:47,063] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 05:44:47,063] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 05:44:47,063] [INFO] [launch.py:272:main] process 1923254 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1687_sentiment140_classification', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification', '--cur_task_id', '13', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights']
+
+============================================================
+Task 13: task1687_sentiment140_classification
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/train.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/dev.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+Train/dev/test: 1000/100/1097
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 126.65it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 150.10it/s]
Loading weights:  18%|█▊        | 51/291 [00:00<00:01, 163.79it/s]
Loading weights:  25%|██▍       | 72/291 [00:00<00:01, 175.29it/s]
Loading weights:  31%|███▏      | 91/291 [00:00<00:01, 173.69it/s]
Loading weights:  38%|███▊      | 112/291 [00:00<00:00, 179.67it/s]
Loading weights:  45%|████▍     | 130/291 [00:00<00:00, 174.89it/s]
Loading weights:  51%|█████     | 148/291 [00:00<00:00, 174.15it/s]
Loading weights:  57%|█████▋    | 167/291 [00:00<00:00, 175.79it/s]
Loading weights:  64%|██████▎   | 185/291 [00:01<00:00, 173.50it/s]
Loading weights:  71%|███████   | 207/291 [00:01<00:00, 184.46it/s]
Loading weights:  78%|███████▊  | 226/291 [00:01<00:00, 185.19it/s]
Loading weights:  85%|████████▍ | 247/291 [00:01<00:00, 177.50it/s]
Loading weights:  92%|█████████▏| 267/291 [00:01<00:00, 179.95it/s]
Loading weights:  99%|█████████▉| 288/291 [00:01<00:00, 188.25it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 178.88it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1827)
+[FIX] Re-initialized lora_A/B in 896 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(13, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[9] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[10] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[11] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[12] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 13 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6768.652M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 13 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 13 sigs, cur=task1687_sentiment140_classification, mapping={'task1687_sentiment140_classification': 0, 'task1572_samsum_summary': 13, 'task363_sst2_polarity_classification': 12, 'task1290_xsum_summarization': 11, 'task181_outcome_extraction': 10, 'task002_quoref_answer_generation': 9, 'task1510_evalution_relation_extraction': 8, 'task639_multi_woz_user_utterance_generation': 7, 'task1729_personachat_generate_next': 6, 'task073_commonsenseqa_answer_generation': 5, 'task1590_diplomacy_text_generation': 4, 'task748_glucose_reverse_cause_event_detection': 3, 'task511_reddit_tifu_long_text_summarization': 2, 'task591_sciq_answer_generation': 1}
+[SGWI] Task 13: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 05:45:57,315] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:05<43:38,  5.47s/it]
  0%|          | 2/480 [00:10<40:07,  5.04s/it]
  1%|          | 3/480 [00:15<39:52,  5.02s/it]
  1%|          | 4/480 [00:20<39:09,  4.94s/it]
  1%|          | 5/480 [00:24<39:00,  4.93s/it]
  1%|▏         | 6/480 [00:29<39:05,  4.95s/it]
  1%|▏         | 7/480 [00:34<38:55,  4.94s/it]
  2%|▏         | 8/480 [00:39<38:40,  4.92s/it]
  2%|▏         | 9/480 [00:44<38:43,  4.93s/it]
  2%|▏         | 10/480 [00:49<38:47,  4.95s/it]
                                                
{'loss': '4.343', 'grad_norm': '1.248', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:49<38:47,  4.95s/it]
  2%|▏         | 11/480 [00:54<38:38,  4.94s/it]
  2%|▎         | 12/480 [01:00<39:49,  5.11s/it]
  3%|▎         | 13/480 [01:04<38:53,  5.00s/it]
  3%|▎         | 14/480 [01:09<38:30,  4.96s/it]
  3%|▎         | 15/480 [01:14<38:03,  4.91s/it]
  3%|▎         | 16/480 [01:19<37:47,  4.89s/it]
  4%|▎         | 17/480 [01:24<37:55,  4.92s/it]
  4%|▍         | 18/480 [01:29<37:58,  4.93s/it]
  4%|▍         | 19/480 [01:34<37:45,  4.91s/it]
  4%|▍         | 20/480 [01:39<39:46,  5.19s/it]
                                                
{'loss': '1.008', 'grad_norm': '0.639', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:39<39:46,  5.19s/it]
  4%|▍         | 21/480 [01:44<38:48,  5.07s/it]
  5%|▍         | 22/480 [01:49<38:00,  4.98s/it]
  5%|▍         | 23/480 [01:54<37:48,  4.96s/it]
  5%|▌         | 24/480 [01:59<37:24,  4.92s/it]
  5%|▌         | 25/480 [02:04<37:29,  4.94s/it]
  5%|▌         | 26/480 [02:09<36:59,  4.89s/it]
  6%|▌         | 27/480 [02:13<36:50,  4.88s/it]
  6%|▌         | 28/480 [02:18<36:57,  4.91s/it]
  6%|▌         | 29/480 [02:23<37:05,  4.93s/it]
  6%|▋         | 30/480 [02:28<37:02,  4.94s/it]
                                                
{'loss': '0.3123', 'grad_norm': '0.6623', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [02:28<37:02,  4.94s/it]
  6%|▋         | 31/480 [02:33<37:11,  4.97s/it]
  7%|▋         | 32/480 [02:35<28:42,  3.84s/it]
  7%|▋         | 33/480 [02:40<31:23,  4.21s/it]
  7%|▋         | 34/480 [02:45<33:24,  4.49s/it]
  7%|▋         | 35/480 [02:50<34:59,  4.72s/it]
  8%|▊         | 36/480 [02:55<35:29,  4.80s/it]
  8%|▊         | 37/480 [03:00<35:45,  4.84s/it]
  8%|▊         | 38/480 [03:05<36:01,  4.89s/it]
  8%|▊         | 39/480 [03:10<36:42,  4.99s/it]
  8%|▊         | 40/480 [03:15<37:15,  5.08s/it]
                                                
{'loss': '0.3241', 'grad_norm': '0.5784', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [03:15<37:15,  5.08s/it]
  9%|▊         | 41/480 [03:20<36:32,  4.99s/it]
  9%|▉         | 42/480 [03:25<36:24,  4.99s/it]
  9%|▉         | 43/480 [03:30<35:52,  4.93s/it]
  9%|▉         | 44/480 [03:35<35:34,  4.90s/it]
  9%|▉         | 45/480 [03:40<35:22,  4.88s/it]
 10%|▉         | 46/480 [03:45<35:15,  4.87s/it]
 10%|▉         | 47/480 [03:49<35:15,  4.89s/it]
 10%|█         | 48/480 [03:54<35:19,  4.91s/it]
 10%|█         | 49/480 [03:59<35:09,  4.89s/it]
 10%|█         | 50/480 [04:04<34:55,  4.87s/it]
                                                
{'loss': '0.2582', 'grad_norm': '0.5471', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [04:04<34:55,  4.87s/it]
 11%|█         | 51/480 [04:09<34:48,  4.87s/it]
 11%|█         | 52/480 [04:14<34:33,  4.84s/it]
 11%|█         | 53/480 [04:19<35:00,  4.92s/it]
 11%|█▏        | 54/480 [04:24<34:42,  4.89s/it]
 11%|█▏        | 55/480 [04:29<34:36,  4.89s/it]
 12%|█▏        | 56/480 [04:34<34:40,  4.91s/it]
 12%|█▏        | 57/480 [04:38<34:28,  4.89s/it]
 12%|█▏        | 58/480 [04:44<35:46,  5.09s/it]
 12%|█▏        | 59/480 [04:49<35:30,  5.06s/it]
 12%|█▎        | 60/480 [04:54<35:14,  5.03s/it]
                                                
{'loss': '0.2578', 'grad_norm': '0.9603', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [04:54<35:14,  5.03s/it]
 13%|█▎        | 61/480 [04:59<35:05,  5.02s/it]
 13%|█▎        | 62/480 [05:04<34:50,  5.00s/it]
 13%|█▎        | 63/480 [05:09<34:47,  5.01s/it]
 13%|█▎        | 64/480 [05:10<27:12,  3.93s/it]
 14%|█▎        | 65/480 [05:15<29:19,  4.24s/it]
 14%|█▍        | 66/480 [05:21<31:27,  4.56s/it]
 14%|█▍        | 67/480 [05:26<32:36,  4.74s/it]
 14%|█▍        | 68/480 [05:31<33:12,  4.84s/it]
 14%|█▍        | 69/480 [05:36<33:07,  4.84s/it]
 15%|█▍        | 70/480 [05:40<32:57,  4.82s/it]
                                                
{'loss': '0.2507', 'grad_norm': '0.2763', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [05:40<32:57,  4.82s/it]
 15%|█▍        | 71/480 [05:45<33:13,  4.87s/it]
 15%|█▌        | 72/480 [05:51<33:50,  4.98s/it]
 15%|█▌        | 73/480 [05:56<33:38,  4.96s/it]
 15%|█▌        | 74/480 [06:00<33:14,  4.91s/it]
 16%|█▌        | 75/480 [06:05<33:10,  4.92s/it]
 16%|█▌        | 76/480 [06:10<33:13,  4.93s/it]
 16%|█▌        | 77/480 [06:15<33:02,  4.92s/it]
 16%|█▋        | 78/480 [06:20<32:44,  4.89s/it]
 16%|█▋        | 79/480 [06:25<32:40,  4.89s/it]
 17%|█▋        | 80/480 [06:30<32:22,  4.86s/it]
                                                
{'loss': '0.2298', 'grad_norm': '0.1354', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [06:30<32:22,  4.86s/it]
 17%|█▋        | 81/480 [06:35<32:41,  4.92s/it]
 17%|█▋        | 82/480 [06:39<32:23,  4.88s/it]
 17%|█▋        | 83/480 [06:44<32:11,  4.87s/it]
 18%|█▊        | 84/480 [06:49<32:40,  4.95s/it]
 18%|█▊        | 85/480 [06:54<32:17,  4.90s/it]
 18%|█▊        | 86/480 [06:59<32:28,  4.94s/it]
 18%|█▊        | 87/480 [07:04<32:05,  4.90s/it]
 18%|█▊        | 88/480 [07:09<32:24,  4.96s/it]
 19%|█▊        | 89/480 [07:14<32:20,  4.96s/it]
 19%|█▉        | 90/480 [07:19<32:18,  4.97s/it]
                                                
{'loss': '0.2551', 'grad_norm': '0.2605', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [07:19<32:18,  4.97s/it]
 19%|█▉        | 91/480 [07:24<32:13,  4.97s/it]
 19%|█▉        | 92/480 [07:29<32:31,  5.03s/it]
 19%|█▉        | 93/480 [07:34<32:25,  5.03s/it]
 20%|█▉        | 94/480 [07:39<32:38,  5.07s/it]
 20%|█▉        | 95/480 [07:44<32:24,  5.05s/it]
 20%|██        | 96/480 [07:46<25:17,  3.95s/it]
 20%|██        | 97/480 [07:51<27:04,  4.24s/it]
 20%|██        | 98/480 [07:56<28:03,  4.41s/it]
 21%|██        | 99/480 [08:00<28:56,  4.56s/it]
 21%|██        | 100/480 [08:05<29:18,  4.63s/it]
                                                 
{'loss': '0.2422', 'grad_norm': '0.1269', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [08:05<29:18,  4.63s/it]
 21%|██        | 101/480 [08:10<29:46,  4.71s/it]
 21%|██▏       | 102/480 [08:15<30:22,  4.82s/it]
 21%|██▏       | 103/480 [08:20<30:44,  4.89s/it]
 22%|██▏       | 104/480 [08:25<30:54,  4.93s/it]
 22%|██▏       | 105/480 [08:30<30:38,  4.90s/it]
 22%|██▏       | 106/480 [08:35<30:24,  4.88s/it]
 22%|██▏       | 107/480 [08:40<31:14,  5.02s/it]
 22%|██▎       | 108/480 [08:46<31:54,  5.15s/it]
 23%|██▎       | 109/480 [08:51<31:31,  5.10s/it]
 23%|██▎       | 110/480 [08:56<31:00,  5.03s/it]
                                                 
{'loss': '0.2775', 'grad_norm': '1.459', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [08:56<31:00,  5.03s/it]
 23%|██▎       | 111/480 [09:00<30:30,  4.96s/it]
 23%|██▎       | 112/480 [09:05<30:32,  4.98s/it]
 24%|██▎       | 113/480 [09:11<31:03,  5.08s/it]
 24%|██▍       | 114/480 [09:16<31:14,  5.12s/it]
 24%|██▍       | 115/480 [09:21<30:34,  5.03s/it]
 24%|██▍       | 116/480 [09:26<30:38,  5.05s/it]
 24%|██▍       | 117/480 [09:31<30:24,  5.03s/it]
 25%|██▍       | 118/480 [09:36<30:02,  4.98s/it]
 25%|██▍       | 119/480 [09:41<29:56,  4.98s/it]
 25%|██▌       | 120/480 [09:46<30:07,  5.02s/it]
                                                 
{'loss': '0.2189', 'grad_norm': '0.4113', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [09:46<30:07,  5.02s/it]
 25%|██▌       | 121/480 [09:52<31:29,  5.26s/it]
 25%|██▌       | 122/480 [09:57<31:13,  5.23s/it]
 26%|██▌       | 123/480 [10:02<30:40,  5.16s/it]
 26%|██▌       | 124/480 [10:07<30:28,  5.14s/it]
 26%|██▌       | 125/480 [10:12<29:46,  5.03s/it]
 26%|██▋       | 126/480 [10:17<29:23,  4.98s/it]
 26%|██▋       | 127/480 [10:21<28:59,  4.93s/it]
 27%|██▋       | 128/480 [10:23<22:22,  3.81s/it]
 27%|██▋       | 129/480 [10:27<24:10,  4.13s/it]
 27%|██▋       | 130/480 [10:32<25:15,  4.33s/it]
                                                 
{'loss': '0.2259', 'grad_norm': '0.2343', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [10:32<25:15,  4.33s/it]
 27%|██▋       | 131/480 [10:37<26:20,  4.53s/it]
 28%|██▊       | 132/480 [10:42<27:11,  4.69s/it]
 28%|██▊       | 133/480 [10:47<27:54,  4.83s/it]
 28%|██▊       | 134/480 [10:52<28:09,  4.88s/it]
 28%|██▊       | 135/480 [10:57<27:59,  4.87s/it]
 28%|██▊       | 136/480 [11:02<28:03,  4.89s/it]
 29%|██▊       | 137/480 [11:07<28:07,  4.92s/it]
 29%|██▉       | 138/480 [11:12<28:12,  4.95s/it]
 29%|██▉       | 139/480 [11:17<28:00,  4.93s/it]
 29%|██▉       | 140/480 [11:22<28:17,  4.99s/it]
                                                 
{'loss': '0.186', 'grad_norm': '0.2397', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [11:22<28:17,  4.99s/it]
 29%|██▉       | 141/480 [11:27<27:52,  4.93s/it]
 30%|██▉       | 142/480 [11:32<28:03,  4.98s/it]
 30%|██▉       | 143/480 [11:37<27:40,  4.93s/it]
 30%|███       | 144/480 [11:42<27:45,  4.96s/it]
 30%|███       | 145/480 [11:47<27:44,  4.97s/it]
 30%|███       | 146/480 [11:52<27:44,  4.98s/it]
 31%|███       | 147/480 [11:57<27:38,  4.98s/it]
 31%|███       | 148/480 [12:02<27:44,  5.01s/it]
 31%|███       | 149/480 [12:07<27:40,  5.02s/it]
 31%|███▏      | 150/480 [12:12<27:38,  5.03s/it]
                                                 
{'loss': '0.2309', 'grad_norm': '0.07342', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [12:12<27:38,  5.03s/it]
 31%|███▏      | 151/480 [12:17<27:33,  5.03s/it]
 32%|███▏      | 152/480 [12:22<27:23,  5.01s/it]
 32%|███▏      | 153/480 [12:27<27:13,  4.99s/it]
 32%|███▏      | 154/480 [12:32<26:48,  4.93s/it]
 32%|███▏      | 155/480 [12:37<26:37,  4.92s/it]
 32%|███▎      | 156/480 [12:42<26:23,  4.89s/it]
 33%|███▎      | 157/480 [12:47<26:25,  4.91s/it]
 33%|███▎      | 158/480 [12:51<26:13,  4.89s/it]
 33%|███▎      | 159/480 [12:57<26:44,  5.00s/it]
 33%|███▎      | 160/480 [12:58<20:36,  3.86s/it]
                                                 
{'loss': '0.2468', 'grad_norm': '0.9904', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [12:58<20:36,  3.86s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<00:09,  9.91it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:00<00:08, 10.87it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:00<00:08, 11.04it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:00<00:08, 11.08it/s]
+
Eval @ step 160:   9%|▉         | 9/100 [00:00<00:08, 11.14it/s]
+
Eval @ step 160:  11%|█         | 11/100 [00:00<00:07, 11.15it/s]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:01<00:07, 11.18it/s]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:01<00:07, 11.17it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:01<00:07, 11.20it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:01<00:07, 11.15it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:01<00:07, 11.18it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:02<00:06, 11.19it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:02<00:06, 11.22it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:02<00:06, 11.20it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:02<00:06, 11.20it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:02<00:06, 11.16it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:02<00:05, 11.18it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:03<00:05, 11.20it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:03<00:05, 11.20it/s]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:03<00:05, 11.22it/s]
+
Eval @ step 160:  41%|████      | 41/100 [00:03<00:05, 11.24it/s]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:03<00:05, 11.15it/s]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:04<00:04, 11.17it/s]
+
Eval @ step 160:  47%|████▋     | 47/100 [00:04<00:04, 11.19it/s]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:04<00:04, 11.23it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:04<00:04, 11.24it/s]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:04<00:04, 11.23it/s]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:04<00:04, 11.20it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:05<00:03, 11.21it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:05<00:03, 11.24it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:05<00:03, 11.24it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [00:05<00:03, 11.24it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:05<00:03, 11.21it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [00:05<00:02, 11.20it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:06<00:02, 11.21it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:06<00:02, 11.18it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:06<00:02, 10.98it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:06<00:02, 10.54it/s]
+
Eval @ step 160:  77%|███████▋  | 77/100 [00:06<00:02,  9.99it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [00:07<00:02, 10.34it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:07<00:01, 10.61it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [00:07<00:01, 10.82it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:07<00:01, 10.96it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [00:07<00:01, 11.07it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [00:08<00:00, 11.11it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [00:08<00:00, 11.17it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:08<00:00, 11.22it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:08<00:00, 11.26it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:08<00:00, 11.27it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [00:08<00:00, 11.25it/s]
+
                                                                   [EVAL] ★ New best eval_rougeL=80.0000 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '80', 'epoch': '5'}
+
 33%|███▎      | 160/480 [13:07<20:36,  3.86s/it]
 34%|███▎      | 161/480 [13:12<37:08,  6.98s/it]
 34%|███▍      | 162/480 [13:17<33:57,  6.41s/it]
 34%|███▍      | 163/480 [13:22<31:52,  6.03s/it]
 34%|███▍      | 164/480 [13:27<30:07,  5.72s/it]
 34%|███▍      | 165/480 [13:32<28:34,  5.44s/it]
 35%|███▍      | 166/480 [13:37<27:48,  5.31s/it]
 35%|███▍      | 167/480 [13:42<26:56,  5.16s/it]
 35%|███▌      | 168/480 [13:47<26:30,  5.10s/it]
 35%|███▌      | 169/480 [13:52<26:19,  5.08s/it]
 35%|███▌      | 170/480 [13:57<25:51,  5.01s/it]
                                                 
{'loss': '0.263', 'grad_norm': '0.007407', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [13:57<25:51,  5.01s/it]
 36%|███▌      | 171/480 [14:02<25:56,  5.04s/it]
 36%|███▌      | 172/480 [14:07<26:16,  5.12s/it]
 36%|███▌      | 173/480 [14:12<25:51,  5.05s/it]
 36%|███▋      | 174/480 [14:17<25:22,  4.98s/it]
 36%|███▋      | 175/480 [14:22<25:37,  5.04s/it]
 37%|███▋      | 176/480 [14:27<25:09,  4.97s/it]
 37%|███▋      | 177/480 [14:32<25:02,  4.96s/it]
 37%|███▋      | 178/480 [14:37<24:58,  4.96s/it]
 37%|███▋      | 179/480 [14:42<24:55,  4.97s/it]
 38%|███▊      | 180/480 [14:47<24:37,  4.93s/it]
                                                 
{'loss': '0.2534', 'grad_norm': '0.9632', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [14:47<24:37,  4.93s/it]
 38%|███▊      | 181/480 [14:52<24:53,  5.00s/it]
 38%|███▊      | 182/480 [14:57<25:28,  5.13s/it]
 38%|███▊      | 183/480 [15:02<25:39,  5.18s/it]
 38%|███▊      | 184/480 [15:08<25:23,  5.15s/it]
 39%|███▊      | 185/480 [15:13<25:17,  5.14s/it]
 39%|███▉      | 186/480 [15:17<24:42,  5.04s/it]
 39%|███▉      | 187/480 [15:23<24:42,  5.06s/it]
 39%|███▉      | 188/480 [15:28<24:29,  5.03s/it]
 39%|███▉      | 189/480 [15:33<24:36,  5.08s/it]
 40%|███▉      | 190/480 [15:38<24:39,  5.10s/it]
                                                 
{'loss': '0.185', 'grad_norm': '0.4963', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [15:38<24:39,  5.10s/it]
 40%|███▉      | 191/480 [15:43<24:22,  5.06s/it]
 40%|████      | 192/480 [15:44<18:50,  3.93s/it]
 40%|████      | 193/480 [15:49<20:01,  4.19s/it]
 40%|████      | 194/480 [15:54<21:18,  4.47s/it]
 41%|████      | 195/480 [15:59<21:53,  4.61s/it]
 41%|████      | 196/480 [16:04<22:36,  4.78s/it]
 41%|████      | 197/480 [16:09<22:59,  4.87s/it]
 41%|████▏     | 198/480 [16:14<23:01,  4.90s/it]
 41%|████▏     | 199/480 [16:19<23:14,  4.96s/it]
 42%|████▏     | 200/480 [16:24<22:58,  4.92s/it]
                                                 
{'loss': '0.2178', 'grad_norm': '0.005392', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [16:24<22:58,  4.92s/it]
 42%|████▏     | 201/480 [16:29<23:01,  4.95s/it]
 42%|████▏     | 202/480 [16:34<22:44,  4.91s/it]
 42%|████▏     | 203/480 [16:39<22:57,  4.97s/it]
 42%|████▎     | 204/480 [16:44<22:46,  4.95s/it]
 43%|████▎     | 205/480 [16:49<22:47,  4.97s/it]
 43%|████▎     | 206/480 [16:54<22:55,  5.02s/it]
 43%|████▎     | 207/480 [16:59<23:03,  5.07s/it]
 43%|████▎     | 208/480 [17:04<22:38,  5.00s/it]
 44%|████▎     | 209/480 [17:09<22:40,  5.02s/it]
 44%|████▍     | 210/480 [17:14<22:20,  4.97s/it]
                                                 
{'loss': '0.2599', 'grad_norm': '0.9147', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [17:14<22:20,  4.97s/it]
 44%|████▍     | 211/480 [17:19<22:26,  5.01s/it]
 44%|████▍     | 212/480 [17:24<22:20,  5.00s/it]
 44%|████▍     | 213/480 [17:29<22:19,  5.02s/it]
 45%|████▍     | 214/480 [17:34<21:57,  4.95s/it]
 45%|████▍     | 215/480 [17:39<22:05,  5.00s/it]
 45%|████▌     | 216/480 [17:44<21:58,  4.99s/it]
 45%|████▌     | 217/480 [17:49<22:06,  5.05s/it]
 45%|████▌     | 218/480 [17:55<23:28,  5.37s/it]
 46%|████▌     | 219/480 [18:00<22:51,  5.25s/it]
 46%|████▌     | 220/480 [18:05<22:11,  5.12s/it]
                                                 
{'loss': '0.1832', 'grad_norm': '0.001773', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [18:05<22:11,  5.12s/it]
 46%|████▌     | 221/480 [18:10<21:41,  5.02s/it]
 46%|████▋     | 222/480 [18:15<21:36,  5.02s/it]
 46%|████▋     | 223/480 [18:20<21:14,  4.96s/it]
 47%|████▋     | 224/480 [18:21<16:29,  3.87s/it]
 47%|████▋     | 225/480 [18:26<17:40,  4.16s/it]
 47%|████▋     | 226/480 [18:31<18:48,  4.44s/it]
 47%|████▋     | 227/480 [18:36<19:31,  4.63s/it]
 48%|████▊     | 228/480 [18:41<19:41,  4.69s/it]
 48%|████▊     | 229/480 [18:46<20:02,  4.79s/it]
 48%|████▊     | 230/480 [18:51<19:58,  4.79s/it]
                                                 
{'loss': '0.3169', 'grad_norm': '0.003505', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [18:51<19:58,  4.79s/it]
 48%|████▊     | 231/480 [18:56<20:24,  4.92s/it]
 48%|████▊     | 232/480 [19:01<20:17,  4.91s/it]
 49%|████▊     | 233/480 [19:06<20:12,  4.91s/it]
 49%|████▉     | 234/480 [19:11<20:12,  4.93s/it]
 49%|████▉     | 235/480 [19:16<20:05,  4.92s/it]
 49%|████▉     | 236/480 [19:21<19:52,  4.89s/it]
 49%|████▉     | 237/480 [19:26<20:07,  4.97s/it]
 50%|████▉     | 238/480 [19:30<19:50,  4.92s/it]
 50%|████▉     | 239/480 [19:36<19:56,  4.96s/it]
 50%|█████     | 240/480 [19:41<19:53,  4.97s/it]
                                                 
{'loss': '0.18', 'grad_norm': '0.1111', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [19:41<19:53,  4.97s/it]
 50%|█████     | 241/480 [19:46<20:07,  5.05s/it]
 50%|█████     | 242/480 [19:51<19:46,  4.98s/it]
 51%|█████     | 243/480 [19:56<19:58,  5.06s/it]
 51%|█████     | 244/480 [20:01<19:57,  5.07s/it]
 51%|█████     | 245/480 [20:06<19:51,  5.07s/it]
 51%|█████▏    | 246/480 [20:11<20:05,  5.15s/it]
 51%|█████▏    | 247/480 [20:16<19:35,  5.05s/it]
 52%|█████▏    | 248/480 [20:21<19:27,  5.03s/it]
 52%|█████▏    | 249/480 [20:26<19:07,  4.97s/it]
 52%|█████▏    | 250/480 [20:31<19:00,  4.96s/it]
                                                 
{'loss': '0.159', 'grad_norm': '1.377', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [20:31<19:00,  4.96s/it]
 52%|█████▏    | 251/480 [20:36<18:45,  4.91s/it]
 52%|█████▎    | 252/480 [20:41<18:48,  4.95s/it]
 53%|█████▎    | 253/480 [20:46<18:37,  4.92s/it]
 53%|█████▎    | 254/480 [20:51<18:45,  4.98s/it]
 53%|█████▎    | 255/480 [20:56<18:28,  4.93s/it]
 53%|█████▎    | 256/480 [20:57<14:13,  3.81s/it]
 54%|█████▎    | 257/480 [21:02<15:25,  4.15s/it]
 54%|█████▍    | 258/480 [21:06<16:06,  4.35s/it]
 54%|█████▍    | 259/480 [21:12<16:54,  4.59s/it]
 54%|█████▍    | 260/480 [21:16<17:04,  4.65s/it]
                                                 
{'loss': '0.1467', 'grad_norm': '0.02723', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [21:16<17:04,  4.65s/it]
 54%|█████▍    | 261/480 [21:21<17:21,  4.75s/it]
 55%|█████▍    | 262/480 [21:26<17:31,  4.82s/it]
 55%|█████▍    | 263/480 [21:31<17:36,  4.87s/it]
 55%|█████▌    | 264/480 [21:36<17:28,  4.85s/it]
 55%|█████▌    | 265/480 [21:41<17:41,  4.94s/it]
 55%|█████▌    | 266/480 [21:46<17:28,  4.90s/it]
 56%|█████▌    | 267/480 [21:51<17:29,  4.93s/it]
 56%|█████▌    | 268/480 [21:56<17:37,  4.99s/it]
 56%|█████▌    | 269/480 [22:01<17:33,  4.99s/it]
 56%|█████▋    | 270/480 [22:06<17:26,  4.99s/it]
                                                 
{'loss': '0.2197', 'grad_norm': '0.01325', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [22:06<17:26,  4.99s/it]
 56%|█████▋    | 271/480 [22:11<17:21,  4.98s/it]
 57%|█████▋    | 272/480 [22:16<17:22,  5.01s/it]
 57%|█████▋    | 273/480 [22:21<17:16,  5.01s/it]
 57%|█████▋    | 274/480 [22:27<17:30,  5.10s/it]
 57%|█████▋    | 275/480 [22:31<17:07,  5.01s/it]
 57%|█████▊    | 276/480 [22:36<16:53,  4.97s/it]
 58%|█████▊    | 277/480 [22:41<16:38,  4.92s/it]
 58%|█████▊    | 278/480 [22:46<16:36,  4.93s/it]
 58%|█████▊    | 279/480 [22:51<16:23,  4.89s/it]
 58%|█████▊    | 280/480 [22:56<16:31,  4.96s/it]
                                                 
{'loss': '0.1933', 'grad_norm': '0.2657', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [22:56<16:31,  4.96s/it]
 59%|█████▊    | 281/480 [23:01<16:32,  4.99s/it]
 59%|█████▉    | 282/480 [23:06<16:36,  5.03s/it]
 59%|█████▉    | 283/480 [23:11<16:18,  4.97s/it]
 59%|█████▉    | 284/480 [23:16<16:18,  4.99s/it]
 59%|█████▉    | 285/480 [23:21<16:06,  4.96s/it]
 60%|█████▉    | 286/480 [23:26<15:53,  4.91s/it]
 60%|█████▉    | 287/480 [23:31<15:57,  4.96s/it]
 60%|██████    | 288/480 [23:32<12:16,  3.84s/it]
 60%|██████    | 289/480 [23:37<13:10,  4.14s/it]
 60%|██████    | 290/480 [23:42<14:02,  4.43s/it]
                                                 
{'loss': '0.1268', 'grad_norm': '0.5112', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [23:42<14:02,  4.43s/it]
 61%|██████    | 291/480 [23:47<14:50,  4.71s/it]
 61%|██████    | 292/480 [23:52<15:01,  4.80s/it]
 61%|██████    | 293/480 [23:58<16:13,  5.21s/it]
 61%|██████▏   | 294/480 [24:03<15:48,  5.10s/it]
 61%|██████▏   | 295/480 [24:08<15:39,  5.08s/it]
 62%|██████▏   | 296/480 [24:13<15:33,  5.07s/it]
 62%|██████▏   | 297/480 [24:18<15:23,  5.05s/it]
 62%|██████▏   | 298/480 [24:23<15:17,  5.04s/it]
 62%|██████▏   | 299/480 [24:28<15:09,  5.03s/it]
 62%|██████▎   | 300/480 [24:34<15:07,  5.04s/it]
                                                 
{'loss': '0.1533', 'grad_norm': '0.0901', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [24:34<15:07,  5.04s/it]
 63%|██████▎   | 301/480 [24:38<14:59,  5.03s/it]
 63%|██████▎   | 302/480 [24:44<15:10,  5.11s/it]
 63%|██████▎   | 303/480 [24:49<14:50,  5.03s/it]
 63%|██████▎   | 304/480 [24:54<14:44,  5.02s/it]
 64%|██████▎   | 305/480 [24:58<14:27,  4.96s/it]
 64%|██████▍   | 306/480 [25:03<14:22,  4.96s/it]
 64%|██████▍   | 307/480 [25:08<14:11,  4.92s/it]
 64%|██████▍   | 308/480 [25:13<14:11,  4.95s/it]
 64%|██████▍   | 309/480 [25:18<13:59,  4.91s/it]
 65%|██████▍   | 310/480 [25:23<14:06,  4.98s/it]
                                                 
{'loss': '0.1833', 'grad_norm': '1.488', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [25:23<14:06,  4.98s/it]
 65%|██████▍   | 311/480 [25:28<13:58,  4.96s/it]
 65%|██████▌   | 312/480 [25:33<14:05,  5.03s/it]
 65%|██████▌   | 313/480 [25:38<13:51,  4.98s/it]
 65%|██████▌   | 314/480 [25:43<13:37,  4.93s/it]
 66%|██████▌   | 315/480 [25:48<13:43,  4.99s/it]
 66%|██████▌   | 316/480 [25:53<13:29,  4.94s/it]
 66%|██████▌   | 317/480 [25:58<13:22,  4.92s/it]
 66%|██████▋   | 318/480 [26:03<13:20,  4.94s/it]
 66%|██████▋   | 319/480 [26:08<13:21,  4.98s/it]
 67%|██████▋   | 320/480 [26:09<10:16,  3.85s/it]
                                                 
{'loss': '0.2823', 'grad_norm': '0.4046', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [26:09<10:16,  3.85s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:08, 11.07it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:00<00:08, 11.10it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:00<00:08, 11.12it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:00<00:08, 11.08it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:00<00:08, 11.02it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:01<00:07, 11.01it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:01<00:07, 11.05it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:01<00:07, 11.16it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:01<00:07, 11.22it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:01<00:07, 11.22it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:01<00:06, 11.27it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:02<00:06, 11.33it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:02<00:06, 11.34it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:02<00:06, 11.35it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:02<00:06, 11.33it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:02<00:06, 11.30it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:03<00:05, 11.08it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:03<00:05, 11.16it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:03<00:05, 11.20it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:03<00:05, 11.25it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:03<00:05, 11.25it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:03<00:04, 11.28it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:04<00:04, 11.31it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:04<00:04, 11.36it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:04<00:04, 11.37it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:04<00:04, 11.37it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:04<00:04, 11.33it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:04<00:03, 11.33it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:05<00:03, 11.34it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:05<00:03, 11.38it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:05<00:03, 11.38it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:05<00:03, 11.37it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:05<00:03, 10.94it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:06<00:03, 10.56it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:06<00:02, 10.30it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:06<00:02, 10.61it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:06<00:02, 10.83it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:06<00:02, 10.97it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:06<00:01, 11.06it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:07<00:01, 11.14it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:07<00:01, 11.22it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:07<00:01, 11.26it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:07<00:01, 11.28it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:07<00:01, 11.28it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:08<00:00, 11.28it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:08<00:00, 11.31it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:08<00:00, 11.30it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:08<00:00, 11.29it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:08<00:00, 11.33it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:08<00:00, 11.32it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=81.0000 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '81', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [26:18<10:16,  3.85s/it]
 67%|██████▋   | 321/480 [26:23<18:14,  6.89s/it]
 67%|██████▋   | 322/480 [26:28<16:39,  6.32s/it]
 67%|██████▋   | 323/480 [26:33<15:20,  5.87s/it]
 68%|██████▊   | 324/480 [26:39<15:32,  5.97s/it]
 68%|██████▊   | 325/480 [26:44<14:38,  5.67s/it]
 68%|██████▊   | 326/480 [26:54<17:35,  6.85s/it]
 68%|██████▊   | 327/480 [27:01<17:26,  6.84s/it]
 68%|██████▊   | 328/480 [27:05<15:54,  6.28s/it]
 69%|██████▊   | 329/480 [27:11<15:08,  6.02s/it]
 69%|██████▉   | 330/480 [27:16<14:11,  5.68s/it]
                                                 
{'loss': '0.1739', 'grad_norm': '0.7478', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [27:16<14:11,  5.68s/it]
 69%|██████▉   | 331/480 [27:22<14:09,  5.70s/it]
 69%|██████▉   | 332/480 [27:26<13:24,  5.44s/it]
 69%|██████▉   | 333/480 [27:31<13:06,  5.35s/it]
 70%|██████▉   | 334/480 [27:36<12:37,  5.19s/it]
 70%|██████▉   | 335/480 [27:41<12:21,  5.12s/it]
 70%|███████   | 336/480 [27:46<12:11,  5.08s/it]
 70%|███████   | 337/480 [27:51<12:05,  5.07s/it]
 70%|███████   | 338/480 [27:56<11:56,  5.04s/it]
 71%|███████   | 339/480 [28:01<11:44,  4.99s/it]
 71%|███████   | 340/480 [28:06<11:38,  4.99s/it]
                                                 
{'loss': '0.1812', 'grad_norm': '0.04918', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [28:06<11:38,  4.99s/it]
 71%|███████   | 341/480 [28:11<11:28,  4.95s/it]
 71%|███████▏  | 342/480 [28:16<11:29,  4.99s/it]
 71%|███████▏  | 343/480 [28:21<11:25,  5.00s/it]
 72%|███████▏  | 344/480 [28:26<11:28,  5.06s/it]
 72%|███████▏  | 345/480 [28:31<11:20,  5.04s/it]
 72%|███████▏  | 346/480 [28:36<11:18,  5.06s/it]
 72%|███████▏  | 347/480 [28:41<11:10,  5.04s/it]
 72%|███████▎  | 348/480 [28:47<11:08,  5.07s/it]
 73%|███████▎  | 349/480 [28:51<10:53,  4.99s/it]
 73%|███████▎  | 350/480 [28:56<10:46,  4.97s/it]
                                                 
{'loss': '0.1508', 'grad_norm': '0.2986', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [28:56<10:46,  4.97s/it]
 73%|███████▎  | 351/480 [29:01<10:35,  4.93s/it]
 73%|███████▎  | 352/480 [29:02<08:07,  3.81s/it]
 74%|███████▎  | 353/480 [29:07<08:50,  4.18s/it]
 74%|███████▍  | 354/480 [29:12<09:11,  4.38s/it]
 74%|███████▍  | 355/480 [29:17<09:25,  4.52s/it]
 74%|███████▍  | 356/480 [29:22<09:31,  4.61s/it]
 74%|███████▍  | 357/480 [29:27<09:49,  4.79s/it]
 75%|███████▍  | 358/480 [29:32<09:45,  4.80s/it]
 75%|███████▍  | 359/480 [29:37<09:44,  4.83s/it]
 75%|███████▌  | 360/480 [29:42<09:38,  4.82s/it]
                                                 
{'loss': '0.1576', 'grad_norm': '0.05489', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [29:42<09:38,  4.82s/it]
 75%|███████▌  | 361/480 [29:47<09:41,  4.89s/it]
 75%|███████▌  | 362/480 [29:51<09:35,  4.88s/it]
 76%|███████▌  | 363/480 [29:56<09:29,  4.87s/it]
 76%|███████▌  | 364/480 [30:01<09:31,  4.93s/it]
 76%|███████▌  | 365/480 [30:06<09:22,  4.89s/it]
 76%|███████▋  | 366/480 [30:11<09:25,  4.96s/it]
 76%|███████▋  | 367/480 [30:16<09:15,  4.92s/it]
 77%|███████▋  | 368/480 [30:21<09:20,  5.00s/it]
 77%|███████▋  | 369/480 [30:26<09:08,  4.95s/it]
 77%|███████▋  | 370/480 [30:31<09:08,  4.98s/it]
                                                 
{'loss': '0.2102', 'grad_norm': '0.01203', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [30:31<09:08,  4.98s/it]
 77%|███████▋  | 371/480 [30:36<09:03,  4.98s/it]
 78%|███████▊  | 372/480 [30:41<09:00,  5.00s/it]
 78%|███████▊  | 373/480 [30:46<08:55,  5.00s/it]
 78%|███████▊  | 374/480 [30:51<08:52,  5.03s/it]
 78%|███████▊  | 375/480 [30:56<08:46,  5.01s/it]
 78%|███████▊  | 376/480 [31:01<08:43,  5.03s/it]
 79%|███████▊  | 377/480 [31:06<08:31,  4.96s/it]
 79%|███████▉  | 378/480 [31:11<08:23,  4.94s/it]
 79%|███████▉  | 379/480 [31:16<08:21,  4.96s/it]
 79%|███████▉  | 380/480 [31:21<08:12,  4.92s/it]
                                                 
{'loss': '0.1855', 'grad_norm': '1.07', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [31:21<08:12,  4.92s/it]
 79%|███████▉  | 381/480 [31:26<08:07,  4.93s/it]
 80%|███████▉  | 382/480 [31:31<07:59,  4.89s/it]
 80%|███████▉  | 383/480 [31:36<07:54,  4.89s/it]
 80%|████████  | 384/480 [31:37<06:03,  3.79s/it]
 80%|████████  | 385/480 [31:42<06:33,  4.15s/it]
 80%|████████  | 386/480 [31:47<06:50,  4.37s/it]
 81%|████████  | 387/480 [31:51<06:58,  4.50s/it]
 81%|████████  | 388/480 [31:56<07:08,  4.66s/it]
 81%|████████  | 389/480 [32:01<07:13,  4.76s/it]
 81%|████████▏ | 390/480 [32:06<07:12,  4.81s/it]
                                                 
{'loss': '0.1529', 'grad_norm': '0.537', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [32:06<07:12,  4.81s/it]
 81%|████████▏ | 391/480 [32:11<07:08,  4.81s/it]
 82%|████████▏ | 392/480 [32:16<07:13,  4.93s/it]
 82%|████████▏ | 393/480 [32:21<07:06,  4.90s/it]
 82%|████████▏ | 394/480 [32:26<07:07,  4.97s/it]
 82%|████████▏ | 395/480 [32:31<06:58,  4.92s/it]
 82%|████████▎ | 396/480 [32:37<07:03,  5.04s/it]
 83%|████████▎ | 397/480 [32:42<06:57,  5.03s/it]
 83%|████████▎ | 398/480 [32:47<06:54,  5.06s/it]
 83%|████████▎ | 399/480 [32:52<06:48,  5.04s/it]
 83%|████████▎ | 400/480 [32:57<06:44,  5.06s/it]
                                                 
{'loss': '0.1404', 'grad_norm': '0.2659', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [32:57<06:44,  5.06s/it]
 84%|████████▎ | 401/480 [33:02<06:37,  5.04s/it]
 84%|████████▍ | 402/480 [33:07<06:33,  5.04s/it]
 84%|████████▍ | 403/480 [33:12<06:27,  5.03s/it]
 84%|████████▍ | 404/480 [33:17<06:22,  5.03s/it]
 84%|████████▍ | 405/480 [33:22<06:14,  4.99s/it]
 85%|████████▍ | 406/480 [33:27<06:06,  4.95s/it]
 85%|████████▍ | 407/480 [33:32<06:01,  4.95s/it]
 85%|████████▌ | 408/480 [33:36<05:53,  4.91s/it]
 85%|████████▌ | 409/480 [33:41<05:48,  4.91s/it]
 85%|████████▌ | 410/480 [33:46<05:42,  4.89s/it]
                                                 
{'loss': '0.1222', 'grad_norm': '0.104', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [33:46<05:42,  4.89s/it]
 86%|████████▌ | 411/480 [33:51<05:38,  4.91s/it]
 86%|████████▌ | 412/480 [33:56<05:35,  4.93s/it]
 86%|████████▌ | 413/480 [34:01<05:36,  5.03s/it]
 86%|████████▋ | 414/480 [34:06<05:31,  5.02s/it]
 86%|████████▋ | 415/480 [34:11<05:24,  4.99s/it]
 87%|████████▋ | 416/480 [34:12<04:07,  3.86s/it]
 87%|████████▋ | 417/480 [34:18<04:31,  4.32s/it]
 87%|████████▋ | 418/480 [34:23<04:40,  4.52s/it]
 87%|████████▋ | 419/480 [34:28<04:41,  4.62s/it]
 88%|████████▊ | 420/480 [34:33<04:46,  4.77s/it]
                                                 
{'loss': '0.1565', 'grad_norm': '1.072', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [34:33<04:46,  4.77s/it]
 88%|████████▊ | 421/480 [34:38<04:41,  4.78s/it]
 88%|████████▊ | 422/480 [34:43<04:42,  4.87s/it]
 88%|████████▊ | 423/480 [34:47<04:36,  4.86s/it]
 88%|████████▊ | 424/480 [34:53<04:37,  4.95s/it]
 89%|████████▊ | 425/480 [34:57<04:30,  4.92s/it]
 89%|████████▉ | 426/480 [35:03<04:27,  4.95s/it]
 89%|████████▉ | 427/480 [35:08<04:24,  4.98s/it]
 89%|████████▉ | 428/480 [35:13<04:19,  4.98s/it]
 89%|████████▉ | 429/480 [35:18<04:15,  5.01s/it]
 90%|████████▉ | 430/480 [35:23<04:10,  5.00s/it]
                                                 
{'loss': '0.1442', 'grad_norm': '0.03429', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [35:23<04:10,  5.00s/it]
 90%|████████▉ | 431/480 [35:28<04:07,  5.06s/it]
 90%|█████████ | 432/480 [35:33<04:10,  5.22s/it]
 90%|█████████ | 433/480 [35:38<04:02,  5.16s/it]
 90%|█████████ | 434/480 [35:43<03:52,  5.06s/it]
 91%|█████████ | 435/480 [35:48<03:45,  5.02s/it]
 91%|█████████ | 436/480 [35:53<03:38,  4.96s/it]
 91%|█████████ | 437/480 [35:58<03:31,  4.93s/it]
 91%|█████████▏| 438/480 [36:03<03:25,  4.90s/it]
 91%|█████████▏| 439/480 [36:08<03:21,  4.92s/it]
 92%|█████████▏| 440/480 [36:13<03:17,  4.94s/it]
                                                 
{'loss': '0.1094', 'grad_norm': '0.007579', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [36:13<03:17,  4.94s/it]
 92%|█████████▏| 441/480 [36:18<03:19,  5.13s/it]
 92%|█████████▏| 442/480 [36:23<03:13,  5.09s/it]
 92%|█████████▏| 443/480 [36:28<03:05,  5.02s/it]
 92%|█████████▎| 444/480 [36:33<03:00,  5.01s/it]
 93%|█████████▎| 445/480 [36:38<02:54,  4.98s/it]
 93%|█████████▎| 446/480 [36:43<02:50,  5.01s/it]
 93%|█████████▎| 447/480 [36:48<02:45,  5.01s/it]
 93%|█████████▎| 448/480 [36:49<02:04,  3.89s/it]
 94%|█████████▎| 449/480 [36:54<02:09,  4.17s/it]
 94%|█████████▍| 450/480 [36:59<02:13,  4.45s/it]
                                                 
{'loss': '0.1715', 'grad_norm': '1.479', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [36:59<02:13,  4.45s/it]
 94%|█████████▍| 451/480 [37:04<02:12,  4.56s/it]
 94%|█████████▍| 452/480 [37:09<02:11,  4.70s/it]
 94%|█████████▍| 453/480 [37:14<02:08,  4.76s/it]
 95%|█████████▍| 454/480 [37:19<02:05,  4.83s/it]
 95%|█████████▍| 455/480 [37:24<02:03,  4.93s/it]
 95%|█████████▌| 456/480 [37:29<01:58,  4.94s/it]
 95%|█████████▌| 457/480 [37:34<01:54,  4.99s/it]
 95%|█████████▌| 458/480 [37:39<01:49,  4.99s/it]
 96%|█████████▌| 459/480 [37:44<01:45,  5.02s/it]
 96%|█████████▌| 460/480 [37:49<01:40,  5.01s/it]
                                                 
{'loss': '0.1524', 'grad_norm': '0.6808', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [37:49<01:40,  5.01s/it]
 96%|█████████▌| 461/480 [37:54<01:34,  4.97s/it]
 96%|█████████▋| 462/480 [37:59<01:28,  4.92s/it]
 96%|█████████▋| 463/480 [38:04<01:24,  4.96s/it]
 97%|█████████▋| 464/480 [38:09<01:18,  4.91s/it]
 97%|█████████▋| 465/480 [38:14<01:13,  4.90s/it]
 97%|█████████▋| 466/480 [38:19<01:08,  4.88s/it]
 97%|█████████▋| 467/480 [38:23<01:03,  4.89s/it]
 98%|█████████▊| 468/480 [38:28<00:58,  4.91s/it]
 98%|█████████▊| 469/480 [38:33<00:53,  4.88s/it]
 98%|█████████▊| 470/480 [38:38<00:49,  4.91s/it]
                                                 
{'loss': '0.1091', 'grad_norm': '0.3537', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [38:38<00:49,  4.91s/it]
 98%|█████████▊| 471/480 [38:43<00:43,  4.89s/it]
 98%|█████████▊| 472/480 [38:48<00:39,  4.96s/it]
 99%|█████████▊| 473/480 [38:53<00:34,  4.91s/it]
 99%|█████████▉| 474/480 [38:58<00:29,  4.93s/it]
 99%|█████████▉| 475/480 [39:03<00:24,  4.94s/it]
 99%|█████████▉| 476/480 [39:08<00:19,  4.98s/it]
 99%|█████████▉| 477/480 [39:13<00:15,  5.01s/it]
100%|█████████▉| 478/480 [39:18<00:10,  5.01s/it]
100%|█████████▉| 479/480 [39:23<00:05,  5.01s/it]
100%|██████████| 480/480 [39:24<00:00,  3.87s/it]
                                                 
{'loss': '0.1673', 'grad_norm': '1.816', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [39:24<00:00,  3.87s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:08, 11.11it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:00<00:08, 11.20it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:00<00:08, 11.22it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:00<00:08, 11.24it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:00<00:08, 11.22it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:01<00:07, 11.20it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:01<00:07, 11.23it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:01<00:07, 11.25it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:01<00:07, 11.21it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:01<00:07, 11.19it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:01<00:07, 10.93it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:02<00:07, 10.56it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:02<00:07, 10.51it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:02<00:06, 10.72it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:02<00:06, 10.90it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:02<00:06, 11.05it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:03<00:05, 11.10it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:03<00:05, 11.19it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:03<00:05, 11.25it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:03<00:05, 11.28it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:03<00:05, 11.31it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:03<00:04, 11.32it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:04<00:04, 11.30it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:04<00:04, 11.35it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:04<00:04, 11.36it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:04<00:04, 11.37it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:04<00:04, 11.37it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:05<00:03, 11.34it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:05<00:03, 11.26it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:05<00:03, 11.32it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [00:05<00:03, 11.33it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:05<00:03, 11.33it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:05<00:03, 11.33it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:06<00:02, 11.33it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:06<00:02, 11.30it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:06<00:02, 11.32it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:06<00:02, 11.34it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:06<00:02, 11.34it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:06<00:01, 11.35it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:07<00:01, 11.34it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:07<00:01, 11.33it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:07<00:01, 11.32it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:07<00:01, 11.30it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:07<00:01, 11.33it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:08<00:00, 11.34it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:08<00:00, 11.33it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:08<00:00, 11.32it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:08<00:00, 11.33it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:08<00:00, 11.33it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:08<00:00, 11.35it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=84.0000 (step=480) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '84', 'epoch': '15'}
+
100%|██████████| 480/480 [39:33<00:00,  3.87s/it]
                                                 
{'train_runtime': '2374', 'train_samples_per_second': '6.319', 'train_steps_per_second': '0.202', 'train_loss': '0.3057', 'epoch': '15'}
+
100%|██████████| 480/480 [39:33<00:00,  3.87s/it]
100%|██████████| 480/480 [39:33<00:00,  4.95s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 46747775GF
+  train_loss               =     0.3057
+  train_runtime            = 0:39:33.67
+  train_samples            =       1000
+  train_samples_per_second =      6.319
+  train_steps_per_second   =      0.202
+  [EVAL] Restored best model (eval_rougeL=84.0000, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 114.83it/s]
Loading weights:  11%|█▏        | 33/291 [00:00<00:01, 156.45it/s]
Loading weights:  18%|█▊        | 51/291 [00:00<00:01, 166.58it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 166.14it/s]
Loading weights:  32%|███▏      | 92/291 [00:00<00:01, 186.18it/s]
Loading weights:  38%|███▊      | 111/291 [00:00<00:00, 186.94it/s]
Loading weights:  45%|████▍     | 130/291 [00:00<00:00, 184.51it/s]
Loading weights:  51%|█████     | 149/291 [00:00<00:00, 176.36it/s]
Loading weights:  58%|█████▊    | 168/291 [00:00<00:00, 180.17it/s]
Loading weights:  65%|██████▌   | 190/291 [00:01<00:00, 190.49it/s]
Loading weights:  72%|███████▏  | 210/291 [00:01<00:00, 185.24it/s]
Loading weights:  79%|███████▊  | 229/291 [00:01<00:00, 172.05it/s]
Loading weights:  85%|████████▍ | 247/291 [00:01<00:00, 169.97it/s]
Loading weights:  93%|█████████▎| 270/291 [00:01<00:00, 180.19it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 180.87it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task1687_sentiment140_classification': PaR=51.8, shrinkage=ridge, n=1000
+  [SRT] Wired router: 14 sigs, cur=task1687_sentiment140_classification, mapping={'task1687_sentiment140_classification': 0, 'task1572_samsum_summary': 13, 'task363_sst2_polarity_classification': 12, 'task1290_xsum_summarization': 11, 'task181_outcome_extraction': 10, 'task002_quoref_answer_generation': 9, 'task1510_evalution_relation_extraction': 8, 'task639_multi_woz_user_utterance_generation': 7, 'task1729_personachat_generate_next': 6, 'task073_commonsenseqa_answer_generation': 5, 'task1590_diplomacy_text_generation': 4, 'task748_glucose_reverse_cause_event_detection': 3, 'task511_reddit_tifu_long_text_summarization': 2, 'task591_sciq_answer_generation': 1}
+[SRT] Router: 14 tasks, avg_PaR=51.8, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification']
+  [SRT-INIT] task_id_to_idx={'task1687_sentiment140_classification': 0, 'task1572_samsum_summary': 13, 'task363_sst2_polarity_classification': 12, 'task1290_xsum_summarization': 11, 'task181_outcome_extraction': 10, 'task002_quoref_answer_generation': 9, 'task1510_evalution_relation_extraction': 8, 'task639_multi_woz_user_utterance_generation': 7, 'task1729_personachat_generate_next': 6, 'task073_commonsenseqa_answer_generation': 5, 'task1590_diplomacy_text_generation': 4, 'task748_glucose_reverse_cause_event_detection': 3, 'task511_reddit_tifu_long_text_summarization': 2, 'task591_sciq_answer_generation': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/1097 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=14
+
Continual predict:   0%|          | 1/1097 [00:01<36:05,  1.98s/it]
Continual predict:   0%|          | 2/1097 [00:04<45:05,  2.47s/it]
Continual predict:   0%|          | 3/1097 [00:05<32:37,  1.79s/it]
Continual predict:   0%|          | 4/1097 [00:06<24:26,  1.34s/it]
Continual predict:   0%|          | 5/1097 [00:08<29:00,  1.59s/it]
Continual predict:   1%|          | 6/1097 [00:13<50:52,  2.80s/it]
Continual predict:   1%|          | 7/1097 [00:17<55:39,  3.06s/it]
Continual predict:   1%|          | 8/1097 [00:19<51:24,  2.83s/it]
Continual predict:   1%|          | 9/1097 [00:21<46:22,  2.56s/it]
Continual predict:   1%|          | 10/1097 [00:22<37:02,  2.04s/it]
Continual predict:   1%|          | 11/1097 [00:25<41:49,  2.31s/it]
Continual predict:   1%|          | 12/1097 [00:28<48:11,  2.67s/it]
Continual predict:   1%|          | 13/1097 [00:30<40:49,  2.26s/it]
Continual predict:   1%|▏         | 14/1097 [00:32<43:03,  2.39s/it]
Continual predict:   1%|▏         | 15/1097 [00:35<43:17,  2.40s/it]
Continual predict:   1%|▏         | 16/1097 [00:38<45:22,  2.52s/it]
Continual predict:   2%|▏         | 17/1097 [00:46<1:16:20,  4.24s/it]
Continual predict:   2%|▏         | 18/1097 [00:48<1:05:41,  3.65s/it]
Continual predict:   2%|▏         | 19/1097 [00:51<1:00:49,  3.39s/it]
Continual predict:   2%|▏         | 20/1097 [00:52<50:49,  2.83s/it]  
Continual predict:   2%|▏         | 21/1097 [00:53<36:35,  2.04s/it]
Continual predict:   2%|▏         | 22/1097 [00:53<26:35,  1.48s/it]
Continual predict:   2%|▏         | 23/1097 [00:53<19:39,  1.10s/it]
Continual predict:   2%|▏         | 24/1097 [00:53<14:45,  1.21it/s]
Continual predict:   2%|▏         | 25/1097 [00:53<11:19,  1.58it/s]
Continual predict:   2%|▏         | 26/1097 [00:53<08:55,  2.00it/s]
Continual predict:   2%|▏         | 27/1097 [00:54<07:15,  2.46it/s]
Continual predict:   3%|▎         | 28/1097 [00:54<06:04,  2.93it/s]
Continual predict:   3%|▎         | 29/1097 [00:54<05:15,  3.39it/s]
Continual predict:   3%|▎         | 30/1097 [00:54<04:42,  3.78it/s]
Continual predict:   3%|▎         | 31/1097 [00:54<04:18,  4.13it/s]
Continual predict:   3%|▎         | 32/1097 [00:55<04:00,  4.43it/s]
Continual predict:   3%|▎         | 33/1097 [00:55<03:50,  4.62it/s]
Continual predict:   3%|▎         | 34/1097 [00:55<03:43,  4.75it/s]
Continual predict:   3%|▎         | 35/1097 [00:55<03:39,  4.83it/s]
Continual predict:   3%|▎         | 36/1097 [00:55<03:39,  4.84it/s]
Continual predict:   3%|▎         | 37/1097 [00:56<03:57,  4.47it/s]
Continual predict:   3%|▎         | 38/1097 [00:56<04:03,  4.35it/s]
Continual predict:   4%|▎         | 39/1097 [00:56<03:50,  4.60it/s]
Continual predict:   4%|▎         | 40/1097 [00:56<03:40,  4.80it/s]
Continual predict:   4%|▎         | 41/1097 [00:57<03:35,  4.90it/s]
Continual predict:   4%|▍         | 42/1097 [00:57<03:29,  5.03it/s]
Continual predict:   4%|▍         | 43/1097 [00:57<03:28,  5.06it/s]
Continual predict:   4%|▍         | 44/1097 [00:57<03:27,  5.07it/s]
Continual predict:   4%|▍         | 45/1097 [00:57<03:24,  5.14it/s]
Continual predict:   4%|▍         | 46/1097 [00:57<03:22,  5.19it/s]
Continual predict:   4%|▍         | 47/1097 [00:58<03:20,  5.22it/s]
Continual predict:   4%|▍         | 48/1097 [00:58<03:22,  5.19it/s]
Continual predict:   4%|▍         | 49/1097 [00:58<03:22,  5.18it/s]
Continual predict:   5%|▍         | 50/1097 [00:58<03:21,  5.19it/s]
Continual predict:   5%|▍         | 51/1097 [00:58<03:20,  5.22it/s]
Continual predict:   5%|▍         | 52/1097 [00:59<03:19,  5.25it/s]
Continual predict:   5%|▍         | 53/1097 [00:59<03:17,  5.27it/s]
Continual predict:   5%|▍         | 54/1097 [00:59<03:19,  5.24it/s]
Continual predict:   5%|▌         | 55/1097 [00:59<03:18,  5.26it/s]
Continual predict:   5%|▌         | 56/1097 [00:59<03:19,  5.22it/s]
Continual predict:   5%|▌         | 57/1097 [01:00<03:17,  5.26it/s]
Continual predict:   5%|▌         | 58/1097 [01:00<03:16,  5.28it/s]
Continual predict:   5%|▌         | 59/1097 [01:00<03:16,  5.29it/s]
Continual predict:   5%|▌         | 60/1097 [01:00<03:17,  5.24it/s]
Continual predict:   6%|▌         | 61/1097 [01:00<03:16,  5.26it/s]
Continual predict:   6%|▌         | 62/1097 [01:01<03:15,  5.29it/s]
Continual predict:   6%|▌         | 63/1097 [01:01<03:17,  5.23it/s]
Continual predict:   6%|▌         | 64/1097 [01:01<03:16,  5.25it/s]
Continual predict:   6%|▌         | 65/1097 [01:01<03:15,  5.27it/s]
Continual predict:   6%|▌         | 66/1097 [01:01<03:15,  5.28it/s]
Continual predict:   6%|▌         | 67/1097 [01:01<03:14,  5.30it/s]
Continual predict:   6%|▌         | 68/1097 [01:02<03:13,  5.31it/s]
Continual predict:   6%|▋         | 69/1097 [01:02<03:13,  5.32it/s]
Continual predict:   6%|▋         | 70/1097 [01:02<03:12,  5.33it/s]
Continual predict:   6%|▋         | 71/1097 [01:02<03:12,  5.33it/s]
Continual predict:   7%|▋         | 72/1097 [01:02<03:15,  5.25it/s]
Continual predict:   7%|▋         | 73/1097 [01:03<03:13,  5.28it/s]
Continual predict:   7%|▋         | 74/1097 [01:03<03:15,  5.24it/s]
Continual predict:   7%|▋         | 75/1097 [01:03<03:14,  5.26it/s]
Continual predict:   7%|▋         | 76/1097 [01:03<03:13,  5.27it/s]
Continual predict:   7%|▋         | 77/1097 [01:03<03:13,  5.27it/s]
Continual predict:   7%|▋         | 78/1097 [01:04<03:14,  5.23it/s]
Continual predict:   7%|▋         | 79/1097 [01:04<03:25,  4.96it/s]
Continual predict:   7%|▋         | 80/1097 [01:04<03:20,  5.06it/s]
Continual predict:   7%|▋         | 81/1097 [01:04<03:18,  5.13it/s]
Continual predict:   7%|▋         | 82/1097 [01:04<03:15,  5.18it/s]
Continual predict:   8%|▊         | 83/1097 [01:05<03:14,  5.20it/s]
Continual predict:   8%|▊         | 84/1097 [01:05<03:15,  5.17it/s]
Continual predict:   8%|▊         | 85/1097 [01:05<03:14,  5.20it/s]
Continual predict:   8%|▊         | 86/1097 [01:05<03:13,  5.23it/s]
Continual predict:   8%|▊         | 87/1097 [01:05<03:12,  5.24it/s]
Continual predict:   8%|▊         | 88/1097 [01:05<03:12,  5.25it/s]
Continual predict:   8%|▊         | 89/1097 [01:06<03:11,  5.25it/s]
Continual predict:   8%|▊         | 90/1097 [01:06<03:40,  4.56it/s]
Continual predict:   8%|▊         | 91/1097 [01:06<03:41,  4.54it/s]
Continual predict:   8%|▊         | 92/1097 [01:06<03:32,  4.74it/s]
Continual predict:   8%|▊         | 93/1097 [01:07<03:27,  4.84it/s]
Continual predict:   9%|▊         | 94/1097 [01:07<03:23,  4.93it/s]
Continual predict:   9%|▊         | 95/1097 [01:07<03:19,  5.03it/s]
Continual predict:   9%|▉         | 96/1097 [01:07<03:16,  5.10it/s]
Continual predict:   9%|▉         | 97/1097 [01:07<03:13,  5.16it/s]
Continual predict:   9%|▉         | 98/1097 [01:08<03:14,  5.13it/s]
Continual predict:   9%|▉         | 99/1097 [01:08<03:14,  5.13it/s]
Continual predict:   9%|▉         | 100/1097 [01:08<03:14,  5.13it/s]
Continual predict:   9%|▉         | 101/1097 [01:08<03:12,  5.18it/s]
Continual predict:   9%|▉         | 102/1097 [01:08<03:10,  5.21it/s]
Continual predict:   9%|▉         | 103/1097 [01:08<03:09,  5.25it/s]
Continual predict:   9%|▉         | 104/1097 [01:09<03:09,  5.23it/s]
Continual predict:  10%|▉         | 105/1097 [01:09<03:11,  5.17it/s]
Continual predict:  10%|▉         | 106/1097 [01:09<03:10,  5.21it/s]
Continual predict:  10%|▉         | 107/1097 [01:09<03:09,  5.23it/s]
Continual predict:  10%|▉         | 108/1097 [01:09<03:08,  5.24it/s]
Continual predict:  10%|▉         | 109/1097 [01:10<03:08,  5.25it/s]
Continual predict:  10%|█         | 110/1097 [01:10<03:09,  5.22it/s]
Continual predict:  10%|█         | 111/1097 [01:10<03:08,  5.23it/s]
Continual predict:  10%|█         | 112/1097 [01:10<03:07,  5.25it/s]
Continual predict:  10%|█         | 113/1097 [01:10<03:06,  5.27it/s]
Continual predict:  10%|█         | 114/1097 [01:11<03:06,  5.27it/s]
Continual predict:  10%|█         | 115/1097 [01:11<03:09,  5.17it/s]
Continual predict:  11%|█         | 116/1097 [01:11<03:08,  5.21it/s]
Continual predict:  11%|█         | 117/1097 [01:11<03:08,  5.20it/s]
Continual predict:  11%|█         | 118/1097 [01:11<03:09,  5.18it/s]
Continual predict:  11%|█         | 119/1097 [01:12<03:07,  5.22it/s]
Continual predict:  11%|█         | 120/1097 [01:12<03:06,  5.23it/s]
Continual predict:  11%|█         | 121/1097 [01:17<26:15,  1.61s/it]
Continual predict:  11%|█         | 122/1097 [01:21<40:31,  2.49s/it]
Continual predict:  11%|█         | 123/1097 [01:27<56:22,  3.47s/it]
Continual predict:  11%|█▏        | 124/1097 [01:30<52:40,  3.25s/it]
Continual predict:  11%|█▏        | 125/1097 [01:35<1:01:54,  3.82s/it]
Continual predict:  11%|█▏        | 126/1097 [01:40<1:08:21,  4.22s/it]
Continual predict:  12%|█▏        | 127/1097 [01:42<55:43,  3.45s/it]  
Continual predict:  12%|█▏        | 128/1097 [01:45<53:43,  3.33s/it]
Continual predict:  12%|█▏        | 129/1097 [01:48<55:28,  3.44s/it]
Continual predict:  12%|█▏        | 130/1097 [01:52<55:02,  3.42s/it]
Continual predict:  12%|█▏        | 131/1097 [01:54<50:42,  3.15s/it]
Continual predict:  12%|█▏        | 132/1097 [01:57<49:43,  3.09s/it]
Continual predict:  12%|█▏        | 133/1097 [02:01<51:34,  3.21s/it]
Continual predict:  12%|█▏        | 134/1097 [02:07<1:07:42,  4.22s/it]
Continual predict:  12%|█▏        | 135/1097 [02:13<1:12:44,  4.54s/it]
Continual predict:  12%|█▏        | 136/1097 [02:16<1:08:09,  4.26s/it]
Continual predict:  12%|█▏        | 137/1097 [02:19<59:34,  3.72s/it]  
Continual predict:  13%|█▎        | 138/1097 [02:20<46:02,  2.88s/it]
Continual predict:  13%|█▎        | 139/1097 [02:21<38:37,  2.42s/it]
Continual predict:  13%|█▎        | 140/1097 [02:24<41:10,  2.58s/it]
Continual predict:  13%|█▎        | 141/1097 [02:31<1:02:39,  3.93s/it]
Continual predict:  13%|█▎        | 142/1097 [02:36<1:07:28,  4.24s/it]
Continual predict:  13%|█▎        | 143/1097 [02:39<1:04:11,  4.04s/it]
Continual predict:  13%|█▎        | 144/1097 [02:50<1:33:06,  5.86s/it]
Continual predict:  13%|█▎        | 145/1097 [02:55<1:28:39,  5.59s/it]
Continual predict:  13%|█▎        | 146/1097 [02:57<1:11:21,  4.50s/it]
Continual predict:  13%|█▎        | 147/1097 [03:01<1:12:25,  4.57s/it]
Continual predict:  13%|█▎        | 148/1097 [03:04<1:02:01,  3.92s/it]
Continual predict:  14%|█▎        | 149/1097 [03:07<59:47,  3.78s/it]  
Continual predict:  14%|█▎        | 150/1097 [03:09<51:55,  3.29s/it]
Continual predict:  14%|█▍        | 151/1097 [03:13<51:52,  3.29s/it]
Continual predict:  14%|█▍        | 152/1097 [03:19<1:05:31,  4.16s/it]
Continual predict:  14%|█▍        | 153/1097 [03:21<58:42,  3.73s/it]  
Continual predict:  14%|█▍        | 154/1097 [03:27<1:05:31,  4.17s/it]
Continual predict:  14%|█▍        | 155/1097 [03:31<1:04:26,  4.10s/it]
Continual predict:  14%|█▍        | 156/1097 [03:34<1:02:44,  4.00s/it]
Continual predict:  14%|█▍        | 157/1097 [03:39<1:03:44,  4.07s/it]
Continual predict:  14%|█▍        | 158/1097 [03:40<50:04,  3.20s/it]  
Continual predict:  14%|█▍        | 159/1097 [03:47<1:10:01,  4.48s/it]
Continual predict:  15%|█▍        | 160/1097 [03:50<1:00:45,  3.89s/it]
Continual predict:  15%|█▍        | 161/1097 [03:52<55:17,  3.54s/it]  
Continual predict:  15%|█▍        | 162/1097 [03:58<1:02:51,  4.03s/it]
Continual predict:  15%|█▍        | 163/1097 [04:00<54:42,  3.51s/it]  
Continual predict:  15%|█▍        | 164/1097 [04:02<49:56,  3.21s/it]
Continual predict:  15%|█▌        | 165/1097 [04:06<51:04,  3.29s/it]
Continual predict:  15%|█▌        | 166/1097 [04:08<46:50,  3.02s/it]
Continual predict:  15%|█▌        | 167/1097 [04:10<39:18,  2.54s/it]
Continual predict:  15%|█▌        | 168/1097 [04:12<39:01,  2.52s/it]
Continual predict:  15%|█▌        | 169/1097 [04:14<36:13,  2.34s/it]
Continual predict:  15%|█▌        | 170/1097 [04:18<41:31,  2.69s/it]
Continual predict:  16%|█▌        | 171/1097 [04:19<37:11,  2.41s/it]
Continual predict:  16%|█▌        | 172/1097 [04:22<39:13,  2.54s/it]
Continual predict:  16%|█▌        | 173/1097 [04:27<46:58,  3.05s/it]
Continual predict:  16%|█▌        | 174/1097 [04:28<39:10,  2.55s/it]
Continual predict:  16%|█▌        | 175/1097 [04:32<44:58,  2.93s/it]
Continual predict:  16%|█▌        | 176/1097 [04:35<45:57,  2.99s/it]
Continual predict:  16%|█▌        | 177/1097 [04:38<46:19,  3.02s/it]
Continual predict:  16%|█▌        | 178/1097 [04:39<38:59,  2.55s/it]
Continual predict:  16%|█▋        | 179/1097 [04:41<33:06,  2.16s/it]
Continual predict:  16%|█▋        | 180/1097 [04:42<29:36,  1.94s/it]
Continual predict:  16%|█▋        | 181/1097 [04:43<27:08,  1.78s/it]
Continual predict:  17%|█▋        | 182/1097 [04:49<42:10,  2.77s/it]
Continual predict:  17%|█▋        | 183/1097 [04:49<30:34,  2.01s/it]
Continual predict:  17%|█▋        | 184/1097 [04:50<26:43,  1.76s/it]
Continual predict:  17%|█▋        | 185/1097 [04:54<35:41,  2.35s/it]
Continual predict:  17%|█▋        | 186/1097 [04:55<30:20,  2.00s/it]
Continual predict:  17%|█▋        | 187/1097 [04:58<37:26,  2.47s/it]
Continual predict:  17%|█▋        | 188/1097 [05:01<39:26,  2.60s/it]
Continual predict:  17%|█▋        | 189/1097 [05:07<52:11,  3.45s/it]
Continual predict:  17%|█▋        | 190/1097 [05:10<52:55,  3.50s/it]
Continual predict:  17%|█▋        | 191/1097 [05:12<43:41,  2.89s/it]
Continual predict:  18%|█▊        | 192/1097 [05:15<47:03,  3.12s/it]
Continual predict:  18%|█▊        | 193/1097 [05:20<55:17,  3.67s/it]
Continual predict:  18%|█▊        | 194/1097 [05:26<1:02:08,  4.13s/it]
Continual predict:  18%|█▊        | 195/1097 [05:31<1:09:28,  4.62s/it]
Continual predict:  18%|█▊        | 196/1097 [05:35<1:04:45,  4.31s/it]
Continual predict:  18%|█▊        | 197/1097 [05:39<1:01:13,  4.08s/it]
Continual predict:  18%|█▊        | 198/1097 [05:47<1:18:43,  5.25s/it]
Continual predict:  18%|█▊        | 199/1097 [05:48<1:00:18,  4.03s/it]
Continual predict:  18%|█▊        | 200/1097 [05:49<49:32,  3.31s/it]  
Continual predict:  18%|█▊        | 201/1097 [05:52<47:11,  3.16s/it]
Continual predict:  18%|█▊        | 202/1097 [05:56<48:52,  3.28s/it]
Continual predict:  19%|█▊        | 203/1097 [06:01<58:52,  3.95s/it]
Continual predict:  19%|█▊        | 204/1097 [06:05<56:07,  3.77s/it]
Continual predict:  19%|█▊        | 205/1097 [06:12<1:10:45,  4.76s/it]
Continual predict:  19%|█▉        | 206/1097 [06:15<1:02:49,  4.23s/it]
Continual predict:  19%|█▉        | 207/1097 [06:19<1:01:42,  4.16s/it]
Continual predict:  19%|█▉        | 208/1097 [06:25<1:12:28,  4.89s/it]
Continual predict:  19%|█▉        | 209/1097 [06:28<1:03:38,  4.30s/it]
Continual predict:  19%|█▉        | 210/1097 [06:30<52:23,  3.54s/it]  
Continual predict:  19%|█▉        | 211/1097 [06:33<49:17,  3.34s/it]
Continual predict:  19%|█▉        | 212/1097 [06:36<49:42,  3.37s/it]
Continual predict:  19%|█▉        | 213/1097 [06:40<52:48,  3.58s/it]
Continual predict:  20%|█▉        | 214/1097 [06:44<54:35,  3.71s/it]
Continual predict:  20%|█▉        | 215/1097 [06:45<42:20,  2.88s/it]
Continual predict:  20%|█▉        | 216/1097 [06:49<44:25,  3.03s/it]
Continual predict:  20%|█▉        | 217/1097 [06:51<41:10,  2.81s/it]
Continual predict:  20%|█▉        | 218/1097 [06:55<47:50,  3.27s/it]
Continual predict:  20%|█▉        | 219/1097 [06:59<49:16,  3.37s/it]
Continual predict:  20%|██        | 220/1097 [07:03<54:04,  3.70s/it]
Continual predict:  20%|██        | 221/1097 [07:04<38:50,  2.66s/it]
Continual predict:  20%|██        | 222/1097 [07:04<28:31,  1.96s/it]
Continual predict:  20%|██        | 223/1097 [07:04<20:57,  1.44s/it]
Continual predict:  20%|██        | 224/1097 [07:04<15:41,  1.08s/it]
Continual predict:  21%|██        | 225/1097 [07:05<12:00,  1.21it/s]
Continual predict:  21%|██        | 226/1097 [07:05<09:31,  1.52it/s]
Continual predict:  21%|██        | 227/1097 [07:05<07:56,  1.83it/s]
Continual predict:  21%|██        | 228/1097 [07:05<06:53,  2.10it/s]
Continual predict:  21%|██        | 229/1097 [07:06<05:55,  2.44it/s]
Continual predict:  21%|██        | 230/1097 [07:06<06:10,  2.34it/s]
Continual predict:  21%|██        | 231/1097 [07:06<05:20,  2.70it/s]
Continual predict:  21%|██        | 232/1097 [07:07<04:45,  3.03it/s]
Continual predict:  21%|██        | 233/1097 [07:07<04:20,  3.32it/s]
Continual predict:  21%|██▏       | 234/1097 [07:07<04:02,  3.55it/s]
Continual predict:  21%|██▏       | 235/1097 [07:07<03:50,  3.75it/s]
Continual predict:  22%|██▏       | 236/1097 [07:08<03:41,  3.89it/s]
Continual predict:  22%|██▏       | 237/1097 [07:08<03:35,  4.00it/s]
Continual predict:  22%|██▏       | 238/1097 [07:08<03:30,  4.08it/s]
Continual predict:  22%|██▏       | 239/1097 [07:08<03:27,  4.14it/s]
Continual predict:  22%|██▏       | 240/1097 [07:09<04:06,  3.48it/s]
Continual predict:  22%|██▏       | 241/1097 [07:09<03:52,  3.69it/s]
Continual predict:  22%|██▏       | 242/1097 [07:09<03:42,  3.85it/s]
Continual predict:  22%|██▏       | 243/1097 [07:09<03:35,  3.97it/s]
Continual predict:  22%|██▏       | 244/1097 [07:10<03:30,  4.05it/s]
Continual predict:  22%|██▏       | 245/1097 [07:10<03:26,  4.12it/s]
Continual predict:  22%|██▏       | 246/1097 [07:10<03:24,  4.16it/s]
Continual predict:  23%|██▎       | 247/1097 [07:10<03:36,  3.93it/s]
Continual predict:  23%|██▎       | 248/1097 [07:11<03:30,  4.03it/s]
Continual predict:  23%|██▎       | 249/1097 [07:11<03:40,  3.85it/s]
Continual predict:  23%|██▎       | 250/1097 [07:11<03:33,  3.97it/s]
Continual predict:  23%|██▎       | 251/1097 [07:11<03:42,  3.81it/s]
Continual predict:  23%|██▎       | 252/1097 [07:13<08:17,  1.70it/s]
Continual predict:  23%|██▎       | 253/1097 [07:13<07:00,  2.01it/s]
Continual predict:  23%|██▎       | 254/1097 [07:13<05:53,  2.39it/s]
Continual predict:  23%|██▎       | 255/1097 [07:14<05:06,  2.74it/s]
Continual predict:  23%|██▎       | 256/1097 [07:14<04:33,  3.07it/s]
Continual predict:  23%|██▎       | 257/1097 [07:15<06:35,  2.12it/s]
Continual predict:  24%|██▎       | 258/1097 [07:15<05:35,  2.50it/s]
Continual predict:  24%|██▎       | 259/1097 [07:15<04:53,  2.86it/s]
Continual predict:  24%|██▎       | 260/1097 [07:15<04:24,  3.17it/s]
Continual predict:  24%|██▍       | 261/1097 [07:16<05:55,  2.35it/s]
Continual predict:  24%|██▍       | 262/1097 [07:16<05:12,  2.67it/s]
Continual predict:  24%|██▍       | 263/1097 [07:17<04:50,  2.87it/s]
Continual predict:  24%|██▍       | 264/1097 [07:18<08:53,  1.56it/s]
Continual predict:  24%|██▍       | 265/1097 [07:18<08:40,  1.60it/s]
Continual predict:  24%|██▍       | 266/1097 [07:19<08:30,  1.63it/s]
Continual predict:  24%|██▍       | 267/1097 [07:20<08:38,  1.60it/s]
Continual predict:  24%|██▍       | 268/1097 [07:21<09:44,  1.42it/s]
Continual predict:  25%|██▍       | 269/1097 [07:21<09:30,  1.45it/s]
Continual predict:  25%|██▍       | 270/1097 [07:22<09:29,  1.45it/s]
Continual predict:  25%|██▍       | 271/1097 [07:23<10:00,  1.38it/s]
Continual predict:  25%|██▍       | 272/1097 [07:24<10:59,  1.25it/s]
Continual predict:  25%|██▍       | 273/1097 [07:24<10:35,  1.30it/s]
Continual predict:  25%|██▍       | 274/1097 [07:25<10:18,  1.33it/s]
Continual predict:  25%|██▌       | 275/1097 [07:26<09:52,  1.39it/s]
Continual predict:  25%|██▌       | 276/1097 [07:27<11:36,  1.18it/s]
Continual predict:  25%|██▌       | 277/1097 [07:28<13:45,  1.01s/it]
Continual predict:  25%|██▌       | 278/1097 [07:29<13:14,  1.03it/s]
Continual predict:  25%|██▌       | 279/1097 [07:30<12:23,  1.10it/s]
Continual predict:  26%|██▌       | 280/1097 [07:30<10:38,  1.28it/s]
Continual predict:  26%|██▌       | 281/1097 [07:31<11:40,  1.17it/s]
Continual predict:  26%|██▌       | 282/1097 [07:32<11:42,  1.16it/s]
Continual predict:  26%|██▌       | 283/1097 [07:33<11:09,  1.22it/s]
Continual predict:  26%|██▌       | 284/1097 [07:34<09:47,  1.38it/s]
Continual predict:  26%|██▌       | 285/1097 [07:34<10:18,  1.31it/s]
Continual predict:  26%|██▌       | 286/1097 [07:35<10:31,  1.28it/s]
Continual predict:  26%|██▌       | 287/1097 [07:36<10:40,  1.27it/s]
Continual predict:  26%|██▋       | 288/1097 [07:37<10:17,  1.31it/s]
Continual predict:  26%|██▋       | 289/1097 [07:38<11:20,  1.19it/s]
Continual predict:  26%|██▋       | 290/1097 [07:39<11:11,  1.20it/s]
Continual predict:  27%|██▋       | 291/1097 [07:39<10:27,  1.28it/s]
Continual predict:  27%|██▋       | 292/1097 [07:40<08:41,  1.54it/s]
Continual predict:  27%|██▋       | 293/1097 [07:40<09:38,  1.39it/s]
Continual predict:  27%|██▋       | 294/1097 [07:41<09:06,  1.47it/s]
Continual predict:  27%|██▋       | 295/1097 [07:42<08:18,  1.61it/s]
Continual predict:  27%|██▋       | 296/1097 [07:43<10:01,  1.33it/s]
Continual predict:  27%|██▋       | 297/1097 [07:43<09:17,  1.44it/s]
Continual predict:  27%|██▋       | 298/1097 [07:44<08:26,  1.58it/s]
Continual predict:  27%|██▋       | 299/1097 [07:45<09:46,  1.36it/s]
Continual predict:  27%|██▋       | 300/1097 [07:45<08:15,  1.61it/s]
Continual predict:  27%|██▋       | 301/1097 [07:46<09:40,  1.37it/s]
Continual predict:  28%|██▊       | 302/1097 [07:46<08:36,  1.54it/s]
Continual predict:  28%|██▊       | 303/1097 [07:48<12:32,  1.06it/s]
Continual predict:  28%|██▊       | 304/1097 [07:48<10:10,  1.30it/s]
Continual predict:  28%|██▊       | 305/1097 [07:49<10:57,  1.21it/s]
Continual predict:  28%|██▊       | 306/1097 [07:50<11:07,  1.19it/s]
Continual predict:  28%|██▊       | 307/1097 [07:51<11:11,  1.18it/s]
Continual predict:  28%|██▊       | 308/1097 [07:52<09:40,  1.36it/s]
Continual predict:  28%|██▊       | 309/1097 [07:52<08:41,  1.51it/s]
Continual predict:  28%|██▊       | 310/1097 [07:53<08:39,  1.52it/s]
Continual predict:  28%|██▊       | 311/1097 [07:53<08:22,  1.56it/s]
Continual predict:  28%|██▊       | 312/1097 [07:54<08:50,  1.48it/s]
Continual predict:  29%|██▊       | 313/1097 [07:55<09:51,  1.32it/s]
Continual predict:  29%|██▊       | 314/1097 [07:55<08:50,  1.48it/s]
Continual predict:  29%|██▊       | 315/1097 [07:56<09:57,  1.31it/s]
Continual predict:  29%|██▉       | 316/1097 [07:57<10:21,  1.26it/s]
Continual predict:  29%|██▉       | 317/1097 [07:58<08:37,  1.51it/s]
Continual predict:  29%|██▉       | 318/1097 [07:58<07:55,  1.64it/s]
Continual predict:  29%|██▉       | 319/1097 [08:00<12:19,  1.05it/s]
Continual predict:  29%|██▉       | 320/1097 [08:01<11:35,  1.12it/s]
Continual predict:  29%|██▉       | 321/1097 [08:01<11:01,  1.17it/s]
Continual predict:  29%|██▉       | 322/1097 [08:02<09:31,  1.36it/s]
Continual predict:  29%|██▉       | 323/1097 [08:03<09:50,  1.31it/s]
Continual predict:  30%|██▉       | 324/1097 [08:04<10:35,  1.22it/s]
Continual predict:  30%|██▉       | 325/1097 [08:05<10:33,  1.22it/s]
Continual predict:  30%|██▉       | 326/1097 [08:05<10:01,  1.28it/s]
Continual predict:  30%|██▉       | 327/1097 [08:06<11:43,  1.09it/s]
Continual predict:  30%|██▉       | 328/1097 [08:08<12:33,  1.02it/s]
Continual predict:  30%|██▉       | 329/1097 [08:08<11:00,  1.16it/s]
Continual predict:  30%|███       | 330/1097 [08:08<09:06,  1.40it/s]
Continual predict:  30%|███       | 331/1097 [08:09<08:50,  1.44it/s]
Continual predict:  30%|███       | 332/1097 [08:10<08:25,  1.51it/s]
Continual predict:  30%|███       | 333/1097 [08:11<10:52,  1.17it/s]
Continual predict:  30%|███       | 334/1097 [08:12<11:17,  1.13it/s]
Continual predict:  31%|███       | 335/1097 [08:12<09:41,  1.31it/s]
Continual predict:  31%|███       | 336/1097 [08:13<08:37,  1.47it/s]
Continual predict:  31%|███       | 337/1097 [08:14<08:22,  1.51it/s]
Continual predict:  31%|███       | 338/1097 [08:14<08:30,  1.49it/s]
Continual predict:  31%|███       | 339/1097 [08:15<08:08,  1.55it/s]
Continual predict:  31%|███       | 340/1097 [08:16<10:04,  1.25it/s]
Continual predict:  31%|███       | 341/1097 [08:17<10:05,  1.25it/s]
Continual predict:  31%|███       | 342/1097 [08:17<08:53,  1.41it/s]
Continual predict:  31%|███▏      | 343/1097 [08:18<09:16,  1.35it/s]
Continual predict:  31%|███▏      | 344/1097 [08:19<10:23,  1.21it/s]
Continual predict:  31%|███▏      | 345/1097 [08:20<09:06,  1.38it/s]
Continual predict:  32%|███▏      | 346/1097 [08:21<09:38,  1.30it/s]
Continual predict:  32%|███▏      | 347/1097 [08:22<12:37,  1.01s/it]
Continual predict:  32%|███▏      | 348/1097 [08:22<10:13,  1.22it/s]
Continual predict:  32%|███▏      | 349/1097 [08:23<08:54,  1.40it/s]
Continual predict:  32%|███▏      | 350/1097 [08:24<08:50,  1.41it/s]
Continual predict:  32%|███▏      | 351/1097 [08:25<10:44,  1.16it/s]
Continual predict:  32%|███▏      | 352/1097 [08:26<10:31,  1.18it/s]
Continual predict:  32%|███▏      | 353/1097 [08:27<12:31,  1.01s/it]
Continual predict:  32%|███▏      | 354/1097 [08:28<13:46,  1.11s/it]
Continual predict:  32%|███▏      | 355/1097 [08:29<12:03,  1.03it/s]
Continual predict:  32%|███▏      | 356/1097 [08:30<12:06,  1.02it/s]
Continual predict:  33%|███▎      | 357/1097 [08:31<12:42,  1.03s/it]
Continual predict:  33%|███▎      | 358/1097 [08:32<11:52,  1.04it/s]
Continual predict:  33%|███▎      | 359/1097 [08:33<10:48,  1.14it/s]
Continual predict:  33%|███▎      | 360/1097 [08:34<12:31,  1.02s/it]
Continual predict:  33%|███▎      | 361/1097 [08:36<15:00,  1.22s/it]
Continual predict:  33%|███▎      | 362/1097 [08:36<12:51,  1.05s/it]
Continual predict:  33%|███▎      | 363/1097 [08:37<10:47,  1.13it/s]
Continual predict:  33%|███▎      | 364/1097 [08:37<08:58,  1.36it/s]
Continual predict:  33%|███▎      | 365/1097 [08:38<07:28,  1.63it/s]
Continual predict:  33%|███▎      | 366/1097 [08:38<06:25,  1.90it/s]
Continual predict:  33%|███▎      | 367/1097 [08:38<05:40,  2.14it/s]
Continual predict:  34%|███▎      | 368/1097 [08:39<05:09,  2.36it/s]
Continual predict:  34%|███▎      | 369/1097 [08:39<04:47,  2.53it/s]
Continual predict:  34%|███▎      | 370/1097 [08:39<04:32,  2.67it/s]
Continual predict:  34%|███▍      | 371/1097 [08:40<04:35,  2.63it/s]
Continual predict:  34%|███▍      | 372/1097 [08:40<04:54,  2.46it/s]
Continual predict:  34%|███▍      | 373/1097 [08:41<05:16,  2.28it/s]
Continual predict:  34%|███▍      | 374/1097 [08:41<04:52,  2.47it/s]
Continual predict:  34%|███▍      | 375/1097 [08:41<05:03,  2.38it/s]
Continual predict:  34%|███▍      | 376/1097 [08:42<04:57,  2.42it/s]
Continual predict:  34%|███▍      | 377/1097 [08:42<04:38,  2.58it/s]
Continual predict:  34%|███▍      | 378/1097 [08:42<04:25,  2.71it/s]
Continual predict:  35%|███▍      | 379/1097 [08:43<04:15,  2.80it/s]
Continual predict:  35%|███▍      | 380/1097 [08:43<04:09,  2.87it/s]
Continual predict:  35%|███▍      | 381/1097 [08:43<04:18,  2.77it/s]
Continual predict:  35%|███▍      | 382/1097 [08:44<04:10,  2.85it/s]
Continual predict:  35%|███▍      | 383/1097 [08:44<04:05,  2.91it/s]
Continual predict:  35%|███▌      | 384/1097 [08:44<04:01,  2.96it/s]
Continual predict:  35%|███▌      | 385/1097 [08:45<03:58,  2.99it/s]
Continual predict:  35%|███▌      | 386/1097 [08:45<03:55,  3.01it/s]
Continual predict:  35%|███▌      | 387/1097 [08:45<03:54,  3.03it/s]
Continual predict:  35%|███▌      | 388/1097 [08:46<03:53,  3.04it/s]
Continual predict:  35%|███▌      | 389/1097 [08:46<04:06,  2.87it/s]
Continual predict:  36%|███▌      | 390/1097 [08:46<04:01,  2.93it/s]
Continual predict:  36%|███▌      | 391/1097 [08:47<03:58,  2.97it/s]
Continual predict:  36%|███▌      | 392/1097 [08:47<04:09,  2.83it/s]
Continual predict:  36%|███▌      | 393/1097 [08:48<04:03,  2.89it/s]
Continual predict:  36%|███▌      | 394/1097 [08:48<03:58,  2.94it/s]
Continual predict:  36%|███▌      | 395/1097 [08:48<03:41,  3.17it/s]
Continual predict:  36%|███▌      | 396/1097 [08:49<04:10,  2.79it/s]
Continual predict:  36%|███▌      | 397/1097 [08:49<04:03,  2.87it/s]
Continual predict:  36%|███▋      | 398/1097 [08:49<03:59,  2.92it/s]
Continual predict:  36%|███▋      | 399/1097 [08:50<03:55,  2.96it/s]
Continual predict:  36%|███▋      | 400/1097 [08:50<03:53,  2.99it/s]
Continual predict:  37%|███▋      | 401/1097 [08:50<03:51,  3.01it/s]
Continual predict:  37%|███▋      | 402/1097 [08:51<03:54,  2.97it/s]
Continual predict:  37%|███▋      | 403/1097 [08:51<04:13,  2.74it/s]
Continual predict:  37%|███▋      | 404/1097 [08:51<04:04,  2.83it/s]
Continual predict:  37%|███▋      | 405/1097 [08:52<04:26,  2.60it/s]
Continual predict:  37%|███▋      | 406/1097 [08:52<04:13,  2.72it/s]
Continual predict:  37%|███▋      | 407/1097 [08:52<04:05,  2.81it/s]
Continual predict:  37%|███▋      | 408/1097 [08:53<03:59,  2.88it/s]
Continual predict:  37%|███▋      | 409/1097 [08:53<03:54,  2.93it/s]
Continual predict:  37%|███▋      | 410/1097 [08:54<04:18,  2.66it/s]
Continual predict:  37%|███▋      | 411/1097 [08:54<04:34,  2.50it/s]
Continual predict:  38%|███▊      | 412/1097 [08:54<04:20,  2.63it/s]
Continual predict:  38%|███▊      | 413/1097 [08:55<04:10,  2.73it/s]
Continual predict:  38%|███▊      | 414/1097 [08:55<04:03,  2.80it/s]
Continual predict:  38%|███▊      | 415/1097 [08:55<03:56,  2.88it/s]
Continual predict:  38%|███▊      | 416/1097 [08:56<03:52,  2.93it/s]
Continual predict:  38%|███▊      | 417/1097 [08:56<03:48,  2.97it/s]
Continual predict:  38%|███▊      | 418/1097 [08:56<03:59,  2.83it/s]
Continual predict:  38%|███▊      | 419/1097 [08:57<04:20,  2.60it/s]
Continual predict:  38%|███▊      | 420/1097 [08:57<04:34,  2.46it/s]
Continual predict:  38%|███▊      | 421/1097 [08:58<04:44,  2.37it/s]
Continual predict:  38%|███▊      | 422/1097 [08:58<04:38,  2.43it/s]
Continual predict:  39%|███▊      | 423/1097 [08:59<04:33,  2.46it/s]
Continual predict:  39%|███▊      | 424/1097 [08:59<04:43,  2.37it/s]
Continual predict:  39%|███▊      | 425/1097 [08:59<04:50,  2.32it/s]
Continual predict:  39%|███▉      | 426/1097 [09:00<04:54,  2.28it/s]
Continual predict:  39%|███▉      | 427/1097 [09:00<04:44,  2.36it/s]
Continual predict:  39%|███▉      | 428/1097 [09:01<04:24,  2.53it/s]
Continual predict:  39%|███▉      | 429/1097 [09:01<04:10,  2.67it/s]
Continual predict:  39%|███▉      | 430/1097 [09:01<04:07,  2.70it/s]
Continual predict:  39%|███▉      | 431/1097 [09:02<04:15,  2.61it/s]
Continual predict:  39%|███▉      | 432/1097 [09:02<04:04,  2.72it/s]
Continual predict:  39%|███▉      | 433/1097 [09:02<03:55,  2.82it/s]
Continual predict:  40%|███▉      | 434/1097 [09:03<04:02,  2.73it/s]
Continual predict:  40%|███▉      | 435/1097 [09:03<03:54,  2.82it/s]
Continual predict:  40%|███▉      | 436/1097 [09:04<04:14,  2.60it/s]
Continual predict:  40%|███▉      | 437/1097 [09:04<04:41,  2.35it/s]
Continual predict:  40%|███▉      | 438/1097 [09:04<04:20,  2.52it/s]
Continual predict:  40%|████      | 439/1097 [09:05<04:19,  2.53it/s]
Continual predict:  40%|████      | 440/1097 [09:05<04:20,  2.52it/s]
Continual predict:  40%|████      | 441/1097 [09:06<04:18,  2.53it/s]
Continual predict:  40%|████      | 442/1097 [09:06<04:05,  2.67it/s]
Continual predict:  40%|████      | 443/1097 [09:06<04:04,  2.68it/s]
Continual predict:  40%|████      | 444/1097 [09:07<03:55,  2.78it/s]
Continual predict:  41%|████      | 445/1097 [09:07<03:48,  2.85it/s]
Continual predict:  41%|████      | 446/1097 [09:07<03:56,  2.75it/s]
Continual predict:  41%|████      | 447/1097 [09:08<04:02,  2.68it/s]
Continual predict:  41%|████      | 448/1097 [09:08<04:06,  2.63it/s]
Continual predict:  41%|████      | 449/1097 [09:09<04:21,  2.48it/s]
Continual predict:  41%|████      | 450/1097 [09:09<04:06,  2.63it/s]
Continual predict:  41%|████      | 451/1097 [09:09<03:56,  2.73it/s]
Continual predict:  41%|████      | 452/1097 [09:10<03:48,  2.82it/s]
Continual predict:  41%|████▏     | 453/1097 [09:10<03:42,  2.89it/s]
Continual predict:  41%|████▏     | 454/1097 [09:10<03:38,  2.94it/s]
Continual predict:  41%|████▏     | 455/1097 [09:11<03:35,  2.98it/s]
Continual predict:  42%|████▏     | 456/1097 [09:11<03:46,  2.83it/s]
Continual predict:  42%|████▏     | 457/1097 [09:11<03:53,  2.74it/s]
Continual predict:  42%|████▏     | 458/1097 [09:12<03:51,  2.76it/s]
Continual predict:  42%|████▏     | 459/1097 [09:12<03:51,  2.76it/s]
Continual predict:  42%|████▏     | 460/1097 [09:12<03:44,  2.84it/s]
Continual predict:  42%|████▏     | 461/1097 [09:13<03:39,  2.90it/s]
Continual predict:  42%|████▏     | 462/1097 [09:13<03:35,  2.95it/s]
Continual predict:  42%|████▏     | 463/1097 [09:13<03:32,  2.98it/s]
Continual predict:  42%|████▏     | 464/1097 [09:14<03:51,  2.73it/s]
Continual predict:  42%|████▏     | 465/1097 [09:14<03:29,  3.02it/s]
Continual predict:  42%|████▏     | 466/1097 [09:14<03:37,  2.90it/s]
Continual predict:  43%|████▎     | 467/1097 [09:15<03:30,  2.99it/s]
Continual predict:  43%|████▎     | 468/1097 [09:15<03:37,  2.90it/s]
Continual predict:  43%|████▎     | 469/1097 [09:15<03:41,  2.83it/s]
Continual predict:  43%|████▎     | 470/1097 [09:16<03:45,  2.78it/s]
Continual predict:  43%|████▎     | 471/1097 [09:16<03:47,  2.75it/s]
Continual predict:  43%|████▎     | 472/1097 [09:17<05:59,  1.74it/s]
Continual predict:  43%|████▎     | 473/1097 [09:18<06:08,  1.69it/s]
Continual predict:  43%|████▎     | 474/1097 [09:18<05:04,  2.05it/s]
Continual predict:  43%|████▎     | 475/1097 [09:19<04:42,  2.20it/s]
Continual predict:  43%|████▎     | 476/1097 [09:19<04:40,  2.22it/s]
Continual predict:  43%|████▎     | 477/1097 [09:19<04:25,  2.33it/s]
Continual predict:  44%|████▎     | 478/1097 [09:20<04:28,  2.31it/s]
Continual predict:  44%|████▎     | 479/1097 [09:21<05:53,  1.75it/s]
Continual predict:  44%|████▍     | 480/1097 [09:21<05:16,  1.95it/s]
Continual predict:  44%|████▍     | 481/1097 [09:21<05:02,  2.04it/s]
Continual predict:  44%|████▍     | 482/1097 [09:22<05:11,  1.98it/s]
Continual predict:  44%|████▍     | 483/1097 [09:23<06:31,  1.57it/s]
Continual predict:  44%|████▍     | 484/1097 [09:24<08:08,  1.25it/s]
Continual predict:  44%|████▍     | 485/1097 [09:25<09:26,  1.08it/s]
Continual predict:  44%|████▍     | 486/1097 [09:27<10:08,  1.00it/s]
Continual predict:  44%|████▍     | 487/1097 [09:28<12:02,  1.18s/it]
Continual predict:  44%|████▍     | 488/1097 [09:29<09:50,  1.03it/s]
Continual predict:  45%|████▍     | 489/1097 [09:29<09:14,  1.10it/s]
Continual predict:  45%|████▍     | 490/1097 [09:30<09:01,  1.12it/s]
Continual predict:  45%|████▍     | 491/1097 [09:31<08:55,  1.13it/s]
Continual predict:  45%|████▍     | 492/1097 [09:32<09:15,  1.09it/s]
Continual predict:  45%|████▍     | 493/1097 [09:33<08:51,  1.14it/s]
Continual predict:  45%|████▌     | 494/1097 [09:34<09:07,  1.10it/s]
Continual predict:  45%|████▌     | 495/1097 [09:36<11:27,  1.14s/it]
Continual predict:  45%|████▌     | 496/1097 [09:36<10:35,  1.06s/it]
Continual predict:  45%|████▌     | 497/1097 [09:37<09:14,  1.08it/s]
Continual predict:  45%|████▌     | 498/1097 [09:38<09:12,  1.08it/s]
Continual predict:  45%|████▌     | 499/1097 [09:39<09:12,  1.08it/s]
Continual predict:  46%|████▌     | 500/1097 [09:39<07:51,  1.26it/s]
Continual predict:  46%|████▌     | 501/1097 [09:40<07:17,  1.36it/s]
Continual predict:  46%|████▌     | 502/1097 [09:41<07:09,  1.38it/s]
Continual predict:  46%|████▌     | 503/1097 [09:42<07:33,  1.31it/s]
Continual predict:  46%|████▌     | 504/1097 [09:42<07:06,  1.39it/s]
Continual predict:  46%|████▌     | 505/1097 [09:43<08:33,  1.15it/s]
Continual predict:  46%|████▌     | 506/1097 [09:44<08:48,  1.12it/s]
Continual predict:  46%|████▌     | 507/1097 [09:45<09:07,  1.08it/s]
Continual predict:  46%|████▋     | 508/1097 [09:46<08:39,  1.13it/s]
Continual predict:  46%|████▋     | 509/1097 [09:47<08:45,  1.12it/s]
Continual predict:  46%|████▋     | 510/1097 [09:49<11:44,  1.20s/it]
Continual predict:  47%|████▋     | 511/1097 [09:50<11:37,  1.19s/it]
Continual predict:  47%|████▋     | 512/1097 [09:51<10:09,  1.04s/it]
Continual predict:  47%|████▋     | 513/1097 [09:53<12:06,  1.24s/it]
Continual predict:  47%|████▋     | 514/1097 [09:54<11:32,  1.19s/it]
Continual predict:  47%|████▋     | 515/1097 [09:54<10:20,  1.07s/it]
Continual predict:  47%|████▋     | 516/1097 [09:55<09:27,  1.02it/s]
Continual predict:  47%|████▋     | 517/1097 [09:56<09:39,  1.00it/s]
Continual predict:  47%|████▋     | 518/1097 [09:57<09:38,  1.00it/s]
Continual predict:  47%|████▋     | 519/1097 [09:59<11:31,  1.20s/it]
Continual predict:  47%|████▋     | 520/1097 [10:00<10:05,  1.05s/it]
Continual predict:  47%|████▋     | 521/1097 [10:01<10:04,  1.05s/it]
Continual predict:  48%|████▊     | 522/1097 [10:01<08:58,  1.07it/s]
Continual predict:  48%|████▊     | 523/1097 [10:02<08:42,  1.10it/s]
Continual predict:  48%|████▊     | 524/1097 [10:03<08:07,  1.18it/s]
Continual predict:  48%|████▊     | 525/1097 [10:04<08:19,  1.15it/s]
Continual predict:  48%|████▊     | 526/1097 [10:05<09:00,  1.06it/s]
Continual predict:  48%|████▊     | 527/1097 [10:06<08:45,  1.09it/s]
Continual predict:  48%|████▊     | 528/1097 [10:07<10:28,  1.10s/it]
Continual predict:  48%|████▊     | 529/1097 [10:08<09:30,  1.00s/it]
Continual predict:  48%|████▊     | 530/1097 [10:09<10:09,  1.08s/it]
Continual predict:  48%|████▊     | 531/1097 [10:11<10:53,  1.16s/it]
Continual predict:  48%|████▊     | 532/1097 [10:12<12:01,  1.28s/it]
Continual predict:  49%|████▊     | 533/1097 [10:13<09:42,  1.03s/it]
Continual predict:  49%|████▊     | 534/1097 [10:13<09:12,  1.02it/s]
Continual predict:  49%|████▉     | 535/1097 [10:14<09:09,  1.02it/s]
Continual predict:  49%|████▉     | 536/1097 [10:15<07:29,  1.25it/s]
Continual predict:  49%|████▉     | 537/1097 [10:16<07:55,  1.18it/s]
Continual predict:  49%|████▉     | 538/1097 [10:17<07:55,  1.18it/s]
Continual predict:  49%|████▉     | 539/1097 [10:18<08:33,  1.09it/s]
Continual predict:  49%|████▉     | 540/1097 [10:19<09:42,  1.05s/it]
Continual predict:  49%|████▉     | 541/1097 [10:20<08:29,  1.09it/s]
Continual predict:  49%|████▉     | 542/1097 [10:20<07:31,  1.23it/s]
Continual predict:  49%|████▉     | 543/1097 [10:21<07:38,  1.21it/s]
Continual predict:  50%|████▉     | 544/1097 [10:22<08:43,  1.06it/s]
Continual predict:  50%|████▉     | 545/1097 [10:24<11:01,  1.20s/it]
Continual predict:  50%|████▉     | 546/1097 [10:25<10:02,  1.09s/it]
Continual predict:  50%|████▉     | 547/1097 [10:26<09:57,  1.09s/it]
Continual predict:  50%|████▉     | 548/1097 [10:27<09:34,  1.05s/it]
Continual predict:  50%|█████     | 549/1097 [10:28<09:41,  1.06s/it]
Continual predict:  50%|█████     | 550/1097 [10:29<09:56,  1.09s/it]
Continual predict:  50%|█████     | 551/1097 [10:30<08:22,  1.09it/s]
Continual predict:  50%|█████     | 552/1097 [10:30<07:40,  1.18it/s]
Continual predict:  50%|█████     | 553/1097 [10:32<09:19,  1.03s/it]
Continual predict:  51%|█████     | 554/1097 [10:33<09:54,  1.09s/it]
Continual predict:  51%|█████     | 555/1097 [10:34<10:01,  1.11s/it]
Continual predict:  51%|█████     | 556/1097 [10:35<08:53,  1.01it/s]
Continual predict:  51%|█████     | 557/1097 [10:37<10:36,  1.18s/it]
Continual predict:  51%|█████     | 558/1097 [10:37<08:57,  1.00it/s]
Continual predict:  51%|█████     | 559/1097 [10:38<09:33,  1.07s/it]
Continual predict:  51%|█████     | 560/1097 [10:40<09:58,  1.12s/it]
Continual predict:  51%|█████     | 561/1097 [10:41<10:54,  1.22s/it]
Continual predict:  51%|█████     | 562/1097 [10:43<11:25,  1.28s/it]
Continual predict:  51%|█████▏    | 563/1097 [10:44<12:08,  1.36s/it]
Continual predict:  51%|█████▏    | 564/1097 [10:46<12:18,  1.39s/it]
Continual predict:  52%|█████▏    | 565/1097 [10:47<12:17,  1.39s/it]
Continual predict:  52%|█████▏    | 566/1097 [10:48<10:51,  1.23s/it]
Continual predict:  52%|█████▏    | 567/1097 [10:49<11:02,  1.25s/it]
Continual predict:  52%|█████▏    | 568/1097 [10:50<09:58,  1.13s/it]
Continual predict:  52%|█████▏    | 569/1097 [10:51<09:45,  1.11s/it]
Continual predict:  52%|█████▏    | 570/1097 [10:51<08:01,  1.09it/s]
Continual predict:  52%|█████▏    | 571/1097 [10:52<08:07,  1.08it/s]
Continual predict:  52%|█████▏    | 572/1097 [10:53<07:11,  1.22it/s]
Continual predict:  52%|█████▏    | 573/1097 [10:54<08:20,  1.05it/s]
Continual predict:  52%|█████▏    | 574/1097 [10:55<07:39,  1.14it/s]
Continual predict:  52%|█████▏    | 575/1097 [10:56<08:16,  1.05it/s]
Continual predict:  53%|█████▎    | 576/1097 [10:57<07:24,  1.17it/s]
Continual predict:  53%|█████▎    | 577/1097 [10:58<07:41,  1.13it/s]
Continual predict:  53%|█████▎    | 578/1097 [10:58<06:46,  1.28it/s]
Continual predict:  53%|█████▎    | 579/1097 [10:59<07:19,  1.18it/s]
Continual predict:  53%|█████▎    | 580/1097 [11:01<08:48,  1.02s/it]
Continual predict:  53%|█████▎    | 581/1097 [11:01<07:53,  1.09it/s]
Continual predict:  53%|█████▎    | 582/1097 [11:01<05:54,  1.45it/s]
Continual predict:  53%|█████▎    | 583/1097 [11:02<04:31,  1.89it/s]
Continual predict:  53%|█████▎    | 584/1097 [11:02<03:33,  2.40it/s]
Continual predict:  53%|█████▎    | 585/1097 [11:02<02:52,  2.96it/s]
Continual predict:  53%|█████▎    | 586/1097 [11:02<02:24,  3.55it/s]
Continual predict:  54%|█████▎    | 587/1097 [11:02<02:04,  4.10it/s]
Continual predict:  54%|█████▎    | 588/1097 [11:02<01:50,  4.61it/s]
Continual predict:  54%|█████▎    | 589/1097 [11:03<01:40,  5.05it/s]
Continual predict:  54%|█████▍    | 590/1097 [11:03<01:33,  5.43it/s]
Continual predict:  54%|█████▍    | 591/1097 [11:03<01:28,  5.71it/s]
Continual predict:  54%|█████▍    | 592/1097 [11:03<01:25,  5.93it/s]
Continual predict:  54%|█████▍    | 593/1097 [11:03<01:22,  6.12it/s]
Continual predict:  54%|█████▍    | 594/1097 [11:03<01:20,  6.25it/s]
Continual predict:  54%|█████▍    | 595/1097 [11:03<01:18,  6.36it/s]
Continual predict:  54%|█████▍    | 596/1097 [11:04<01:17,  6.43it/s]
Continual predict:  54%|█████▍    | 597/1097 [11:04<01:17,  6.49it/s]
Continual predict:  55%|█████▍    | 598/1097 [11:04<01:16,  6.49it/s]
Continual predict:  55%|█████▍    | 599/1097 [11:04<01:16,  6.49it/s]
Continual predict:  55%|█████▍    | 600/1097 [11:04<01:16,  6.52it/s]
Continual predict:  55%|█████▍    | 601/1097 [11:04<01:16,  6.52it/s]
Continual predict:  55%|█████▍    | 602/1097 [11:05<01:16,  6.50it/s]
Continual predict:  55%|█████▍    | 603/1097 [11:05<01:15,  6.50it/s]
Continual predict:  55%|█████▌    | 604/1097 [11:05<01:15,  6.49it/s]
Continual predict:  55%|█████▌    | 605/1097 [11:05<01:15,  6.49it/s]
Continual predict:  55%|█████▌    | 606/1097 [11:05<01:16,  6.44it/s]
Continual predict:  55%|█████▌    | 607/1097 [11:05<01:15,  6.49it/s]
Continual predict:  55%|█████▌    | 608/1097 [11:05<01:15,  6.48it/s]
Continual predict:  56%|█████▌    | 609/1097 [11:06<01:14,  6.52it/s]
Continual predict:  56%|█████▌    | 610/1097 [11:06<01:14,  6.54it/s]
Continual predict:  56%|█████▌    | 611/1097 [11:06<01:14,  6.56it/s]
Continual predict:  56%|█████▌    | 612/1097 [11:06<01:14,  6.53it/s]
Continual predict:  56%|█████▌    | 613/1097 [11:06<01:14,  6.52it/s]
Continual predict:  56%|█████▌    | 614/1097 [11:06<01:13,  6.55it/s]
Continual predict:  56%|█████▌    | 615/1097 [11:07<01:13,  6.54it/s]
Continual predict:  56%|█████▌    | 616/1097 [11:07<01:13,  6.52it/s]
Continual predict:  56%|█████▌    | 617/1097 [11:07<01:13,  6.54it/s]
Continual predict:  56%|█████▋    | 618/1097 [11:07<01:13,  6.53it/s]
Continual predict:  56%|█████▋    | 619/1097 [11:07<01:13,  6.51it/s]
Continual predict:  57%|█████▋    | 620/1097 [11:07<01:13,  6.50it/s]
Continual predict:  57%|█████▋    | 621/1097 [11:07<01:13,  6.50it/s]
Continual predict:  57%|█████▋    | 622/1097 [11:08<01:13,  6.49it/s]
Continual predict:  57%|█████▋    | 623/1097 [11:08<01:12,  6.50it/s]
Continual predict:  57%|█████▋    | 624/1097 [11:08<01:13,  6.46it/s]
Continual predict:  57%|█████▋    | 625/1097 [11:08<01:14,  6.35it/s]
Continual predict:  57%|█████▋    | 626/1097 [11:08<01:14,  6.32it/s]
Continual predict:  57%|█████▋    | 627/1097 [11:08<01:13,  6.37it/s]
Continual predict:  57%|█████▋    | 628/1097 [11:09<01:13,  6.42it/s]
Continual predict:  57%|█████▋    | 629/1097 [11:09<01:12,  6.48it/s]
Continual predict:  57%|█████▋    | 630/1097 [11:09<01:12,  6.48it/s]
Continual predict:  58%|█████▊    | 631/1097 [11:09<01:11,  6.48it/s]
Continual predict:  58%|█████▊    | 632/1097 [11:09<01:11,  6.48it/s]
Continual predict:  58%|█████▊    | 633/1097 [11:09<01:11,  6.52it/s]
Continual predict:  58%|█████▊    | 634/1097 [11:09<01:11,  6.51it/s]
Continual predict:  58%|█████▊    | 635/1097 [11:10<01:10,  6.55it/s]
Continual predict:  58%|█████▊    | 636/1097 [11:10<01:10,  6.52it/s]
Continual predict:  58%|█████▊    | 637/1097 [11:10<01:10,  6.52it/s]
Continual predict:  58%|█████▊    | 638/1097 [11:10<01:10,  6.50it/s]
Continual predict:  58%|█████▊    | 639/1097 [11:10<01:10,  6.49it/s]
Continual predict:  58%|█████▊    | 640/1097 [11:10<01:09,  6.53it/s]
Continual predict:  58%|█████▊    | 641/1097 [11:11<01:09,  6.56it/s]
Continual predict:  59%|█████▊    | 642/1097 [11:11<01:09,  6.54it/s]
Continual predict:  59%|█████▊    | 643/1097 [11:11<01:09,  6.53it/s]
Continual predict:  59%|█████▊    | 644/1097 [11:11<01:09,  6.56it/s]
Continual predict:  59%|█████▉    | 645/1097 [11:11<01:09,  6.54it/s]
Continual predict:  59%|█████▉    | 646/1097 [11:11<01:09,  6.52it/s]
Continual predict:  59%|█████▉    | 647/1097 [11:11<01:09,  6.51it/s]
Continual predict:  59%|█████▉    | 648/1097 [11:12<01:08,  6.54it/s]
Continual predict:  59%|█████▉    | 649/1097 [11:12<01:08,  6.53it/s]
Continual predict:  59%|█████▉    | 650/1097 [11:12<01:08,  6.56it/s]
Continual predict:  59%|█████▉    | 651/1097 [11:12<01:07,  6.58it/s]
Continual predict:  59%|█████▉    | 652/1097 [11:12<01:07,  6.58it/s]
Continual predict:  60%|█████▉    | 653/1097 [11:12<01:07,  6.59it/s]
Continual predict:  60%|█████▉    | 654/1097 [11:13<01:07,  6.61it/s]
Continual predict:  60%|█████▉    | 655/1097 [11:13<01:06,  6.61it/s]
Continual predict:  60%|█████▉    | 656/1097 [11:13<01:07,  6.57it/s]
Continual predict:  60%|█████▉    | 657/1097 [11:13<01:07,  6.55it/s]
Continual predict:  60%|█████▉    | 658/1097 [11:13<01:07,  6.53it/s]
Continual predict:  60%|██████    | 659/1097 [11:13<01:06,  6.56it/s]
Continual predict:  60%|██████    | 660/1097 [11:13<01:06,  6.54it/s]
Continual predict:  60%|██████    | 661/1097 [11:14<01:06,  6.52it/s]
Continual predict:  60%|██████    | 662/1097 [11:14<01:06,  6.51it/s]
Continual predict:  60%|██████    | 663/1097 [11:14<01:06,  6.51it/s]
Continual predict:  61%|██████    | 664/1097 [11:14<01:06,  6.50it/s]
Continual predict:  61%|██████    | 665/1097 [11:14<01:06,  6.50it/s]
Continual predict:  61%|██████    | 666/1097 [11:14<01:06,  6.53it/s]
Continual predict:  61%|██████    | 667/1097 [11:15<01:05,  6.52it/s]
Continual predict:  61%|██████    | 668/1097 [11:15<01:05,  6.51it/s]
Continual predict:  61%|██████    | 669/1097 [11:15<01:05,  6.54it/s]
Continual predict:  61%|██████    | 670/1097 [11:15<01:05,  6.53it/s]
Continual predict:  61%|██████    | 671/1097 [11:15<01:05,  6.51it/s]
Continual predict:  61%|██████▏   | 672/1097 [11:15<01:05,  6.51it/s]
Continual predict:  61%|██████▏   | 673/1097 [11:15<01:04,  6.53it/s]
Continual predict:  61%|██████▏   | 674/1097 [11:16<01:04,  6.52it/s]
Continual predict:  62%|██████▏   | 675/1097 [11:16<01:04,  6.55it/s]
Continual predict:  62%|██████▏   | 676/1097 [11:16<01:04,  6.57it/s]
Continual predict:  62%|██████▏   | 677/1097 [11:16<01:03,  6.59it/s]
Continual predict:  62%|██████▏   | 678/1097 [11:16<01:03,  6.56it/s]
Continual predict:  62%|██████▏   | 679/1097 [11:16<01:03,  6.54it/s]
Continual predict:  62%|██████▏   | 680/1097 [11:16<01:03,  6.52it/s]
Continual predict:  62%|██████▏   | 681/1097 [11:17<01:03,  6.50it/s]
Continual predict:  62%|██████▏   | 682/1097 [11:18<03:10,  2.17it/s]
Continual predict:  62%|██████▏   | 683/1097 [11:20<07:21,  1.07s/it]
Continual predict:  62%|██████▏   | 684/1097 [11:22<08:40,  1.26s/it]
Continual predict:  62%|██████▏   | 685/1097 [11:23<07:40,  1.12s/it]
Continual predict:  63%|██████▎   | 686/1097 [11:25<10:19,  1.51s/it]
Continual predict:  63%|██████▎   | 687/1097 [11:27<10:31,  1.54s/it]
Continual predict:  63%|██████▎   | 688/1097 [11:32<16:54,  2.48s/it]
Continual predict:  63%|██████▎   | 689/1097 [11:32<13:35,  2.00s/it]
Continual predict:  63%|██████▎   | 690/1097 [11:36<17:27,  2.57s/it]
Continual predict:  63%|██████▎   | 691/1097 [11:37<14:23,  2.13s/it]
Continual predict:  63%|██████▎   | 692/1097 [11:38<11:28,  1.70s/it]
Continual predict:  63%|██████▎   | 693/1097 [11:41<14:15,  2.12s/it]
Continual predict:  63%|██████▎   | 694/1097 [11:42<11:41,  1.74s/it]
Continual predict:  63%|██████▎   | 695/1097 [11:44<12:24,  1.85s/it]
Continual predict:  63%|██████▎   | 696/1097 [11:45<11:18,  1.69s/it]
Continual predict:  64%|██████▎   | 697/1097 [11:46<09:17,  1.39s/it]
Continual predict:  64%|██████▎   | 698/1097 [11:48<10:19,  1.55s/it]
Continual predict:  64%|██████▎   | 699/1097 [11:49<09:58,  1.50s/it]
Continual predict:  64%|██████▍   | 700/1097 [11:51<09:54,  1.50s/it]
Continual predict:  64%|██████▍   | 701/1097 [11:52<09:57,  1.51s/it]
Continual predict:  64%|██████▍   | 702/1097 [11:54<09:14,  1.40s/it]
Continual predict:  64%|██████▍   | 703/1097 [11:55<09:55,  1.51s/it]
Continual predict:  64%|██████▍   | 704/1097 [11:57<09:11,  1.40s/it]
Continual predict:  64%|██████▍   | 705/1097 [12:00<12:29,  1.91s/it]
Continual predict:  64%|██████▍   | 706/1097 [12:01<12:08,  1.86s/it]
Continual predict:  64%|██████▍   | 707/1097 [12:03<11:19,  1.74s/it]
Continual predict:  65%|██████▍   | 708/1097 [12:06<13:47,  2.13s/it]
Continual predict:  65%|██████▍   | 709/1097 [12:08<12:45,  1.97s/it]
Continual predict:  65%|██████▍   | 710/1097 [12:09<12:11,  1.89s/it]
Continual predict:  65%|██████▍   | 711/1097 [12:11<11:33,  1.80s/it]
Continual predict:  65%|██████▍   | 712/1097 [12:12<11:03,  1.72s/it]
Continual predict:  65%|██████▍   | 713/1097 [12:14<10:22,  1.62s/it]
Continual predict:  65%|██████▌   | 714/1097 [12:16<11:03,  1.73s/it]
Continual predict:  65%|██████▌   | 715/1097 [12:18<11:49,  1.86s/it]
Continual predict:  65%|██████▌   | 716/1097 [12:19<11:10,  1.76s/it]
Continual predict:  65%|██████▌   | 717/1097 [12:21<10:30,  1.66s/it]
Continual predict:  65%|██████▌   | 718/1097 [12:22<10:26,  1.65s/it]
Continual predict:  66%|██████▌   | 719/1097 [12:24<10:56,  1.74s/it]
Continual predict:  66%|██████▌   | 720/1097 [12:26<09:57,  1.59s/it]
Continual predict:  66%|██████▌   | 721/1097 [12:27<10:07,  1.62s/it]
Continual predict:  66%|██████▌   | 722/1097 [12:29<09:49,  1.57s/it]
Continual predict:  66%|██████▌   | 723/1097 [12:30<09:35,  1.54s/it]
Continual predict:  66%|██████▌   | 724/1097 [12:32<09:41,  1.56s/it]
Continual predict:  66%|██████▌   | 725/1097 [12:34<09:50,  1.59s/it]
Continual predict:  66%|██████▌   | 726/1097 [12:35<09:34,  1.55s/it]
Continual predict:  66%|██████▋   | 727/1097 [12:36<09:24,  1.52s/it]
Continual predict:  66%|██████▋   | 728/1097 [12:38<10:07,  1.65s/it]
Continual predict:  66%|██████▋   | 729/1097 [12:40<10:19,  1.68s/it]
Continual predict:  67%|██████▋   | 730/1097 [12:42<10:35,  1.73s/it]
Continual predict:  67%|██████▋   | 731/1097 [12:44<10:30,  1.72s/it]
Continual predict:  67%|██████▋   | 732/1097 [12:45<09:59,  1.64s/it]
Continual predict:  67%|██████▋   | 733/1097 [12:46<09:04,  1.49s/it]
Continual predict:  67%|██████▋   | 734/1097 [12:49<10:23,  1.72s/it]
Continual predict:  67%|██████▋   | 735/1097 [12:51<11:18,  1.88s/it]
Continual predict:  67%|██████▋   | 736/1097 [12:53<11:29,  1.91s/it]
Continual predict:  67%|██████▋   | 737/1097 [12:54<10:25,  1.74s/it]
Continual predict:  67%|██████▋   | 738/1097 [12:56<10:28,  1.75s/it]
Continual predict:  67%|██████▋   | 739/1097 [12:58<11:01,  1.85s/it]
Continual predict:  67%|██████▋   | 740/1097 [13:00<10:52,  1.83s/it]
Continual predict:  68%|██████▊   | 741/1097 [13:02<10:56,  1.84s/it]
Continual predict:  68%|██████▊   | 742/1097 [13:03<10:39,  1.80s/it]
Continual predict:  68%|██████▊   | 743/1097 [13:05<10:35,  1.79s/it]
Continual predict:  68%|██████▊   | 744/1097 [13:07<10:38,  1.81s/it]
Continual predict:  68%|██████▊   | 745/1097 [13:09<10:48,  1.84s/it]
Continual predict:  68%|██████▊   | 746/1097 [13:11<10:39,  1.82s/it]
Continual predict:  68%|██████▊   | 747/1097 [13:12<10:31,  1.80s/it]
Continual predict:  68%|██████▊   | 748/1097 [13:14<10:51,  1.87s/it]
Continual predict:  68%|██████▊   | 749/1097 [13:17<11:16,  1.94s/it]
Continual predict:  68%|██████▊   | 750/1097 [13:18<10:23,  1.80s/it]
Continual predict:  68%|██████▊   | 751/1097 [13:20<10:51,  1.88s/it]
Continual predict:  69%|██████▊   | 752/1097 [13:21<09:50,  1.71s/it]
Continual predict:  69%|██████▊   | 753/1097 [13:23<09:22,  1.64s/it]
Continual predict:  69%|██████▊   | 754/1097 [13:24<08:40,  1.52s/it]
Continual predict:  69%|██████▉   | 755/1097 [13:26<09:31,  1.67s/it]
Continual predict:  69%|██████▉   | 756/1097 [13:28<09:15,  1.63s/it]
Continual predict:  69%|██████▉   | 757/1097 [13:29<08:49,  1.56s/it]
Continual predict:  69%|██████▉   | 758/1097 [13:30<08:38,  1.53s/it]
Continual predict:  69%|██████▉   | 759/1097 [13:32<08:45,  1.55s/it]
Continual predict:  69%|██████▉   | 760/1097 [13:34<08:34,  1.53s/it]
Continual predict:  69%|██████▉   | 761/1097 [13:36<10:25,  1.86s/it]
Continual predict:  69%|██████▉   | 762/1097 [13:39<11:19,  2.03s/it]
Continual predict:  70%|██████▉   | 763/1097 [13:41<11:07,  2.00s/it]
Continual predict:  70%|██████▉   | 764/1097 [13:42<10:50,  1.95s/it]
Continual predict:  70%|██████▉   | 765/1097 [13:44<09:36,  1.74s/it]
Continual predict:  70%|██████▉   | 766/1097 [13:45<09:38,  1.75s/it]
Continual predict:  70%|██████▉   | 767/1097 [13:47<09:24,  1.71s/it]
Continual predict:  70%|███████   | 768/1097 [13:49<09:05,  1.66s/it]
Continual predict:  70%|███████   | 769/1097 [13:50<08:49,  1.61s/it]
Continual predict:  70%|███████   | 770/1097 [13:52<08:48,  1.62s/it]
Continual predict:  70%|███████   | 771/1097 [13:53<08:09,  1.50s/it]
Continual predict:  70%|███████   | 772/1097 [13:54<07:48,  1.44s/it]
Continual predict:  70%|███████   | 773/1097 [13:56<09:04,  1.68s/it]
Continual predict:  71%|███████   | 774/1097 [13:58<08:57,  1.66s/it]
Continual predict:  71%|███████   | 775/1097 [14:00<08:36,  1.60s/it]
Continual predict:  71%|███████   | 776/1097 [14:01<09:06,  1.70s/it]
Continual predict:  71%|███████   | 777/1097 [14:03<08:49,  1.65s/it]
Continual predict:  71%|███████   | 778/1097 [14:05<08:58,  1.69s/it]
Continual predict:  71%|███████   | 779/1097 [14:06<08:43,  1.65s/it]
Continual predict:  71%|███████   | 780/1097 [14:08<08:36,  1.63s/it]
Continual predict:  71%|███████   | 781/1097 [14:10<08:48,  1.67s/it]
Continual predict:  71%|███████▏  | 782/1097 [14:11<08:18,  1.58s/it]
Continual predict:  71%|███████▏  | 783/1097 [14:13<08:42,  1.66s/it]
Continual predict:  71%|███████▏  | 784/1097 [14:15<08:42,  1.67s/it]
Continual predict:  72%|███████▏  | 785/1097 [14:16<08:15,  1.59s/it]
Continual predict:  72%|███████▏  | 786/1097 [14:18<08:24,  1.62s/it]
Continual predict:  72%|███████▏  | 787/1097 [14:20<08:45,  1.69s/it]
Continual predict:  72%|███████▏  | 788/1097 [14:21<07:59,  1.55s/it]
Continual predict:  72%|███████▏  | 789/1097 [14:23<08:17,  1.62s/it]
Continual predict:  72%|███████▏  | 790/1097 [14:25<09:26,  1.85s/it]
Continual predict:  72%|███████▏  | 791/1097 [14:27<09:18,  1.82s/it]
Continual predict:  72%|███████▏  | 792/1097 [14:28<08:36,  1.69s/it]
Continual predict:  72%|███████▏  | 793/1097 [14:30<08:34,  1.69s/it]
Continual predict:  72%|███████▏  | 794/1097 [14:32<09:30,  1.88s/it]
Continual predict:  72%|███████▏  | 795/1097 [14:34<08:43,  1.73s/it]
Continual predict:  73%|███████▎  | 796/1097 [14:35<08:31,  1.70s/it]
Continual predict:  73%|███████▎  | 797/1097 [14:37<08:32,  1.71s/it]
Continual predict:  73%|███████▎  | 798/1097 [14:39<09:02,  1.81s/it]
Continual predict:  73%|███████▎  | 799/1097 [14:46<16:29,  3.32s/it]
Continual predict:  73%|███████▎  | 800/1097 [14:48<14:52,  3.01s/it]
Continual predict:  73%|███████▎  | 801/1097 [14:49<12:20,  2.50s/it]
Continual predict:  73%|███████▎  | 802/1097 [14:51<10:41,  2.17s/it]
Continual predict:  73%|███████▎  | 803/1097 [14:52<09:01,  1.84s/it]
Continual predict:  73%|███████▎  | 804/1097 [14:56<11:58,  2.45s/it]
Continual predict:  73%|███████▎  | 805/1097 [14:59<13:39,  2.81s/it]
Continual predict:  73%|███████▎  | 806/1097 [15:01<12:39,  2.61s/it]
Continual predict:  74%|███████▎  | 807/1097 [15:03<10:52,  2.25s/it]
Continual predict:  74%|███████▎  | 808/1097 [15:05<10:03,  2.09s/it]
Continual predict:  74%|███████▎  | 809/1097 [15:06<09:12,  1.92s/it]
Continual predict:  74%|███████▍  | 810/1097 [15:08<09:47,  2.05s/it]
Continual predict:  74%|███████▍  | 811/1097 [15:10<09:17,  1.95s/it]
Continual predict:  74%|███████▍  | 812/1097 [15:14<12:30,  2.63s/it]
Continual predict:  74%|███████▍  | 813/1097 [15:22<19:49,  4.19s/it]
Continual predict:  74%|███████▍  | 814/1097 [15:24<16:45,  3.55s/it]
Continual predict:  74%|███████▍  | 815/1097 [15:27<15:24,  3.28s/it]
Continual predict:  74%|███████▍  | 816/1097 [15:30<14:26,  3.08s/it]
Continual predict:  74%|███████▍  | 817/1097 [15:31<12:13,  2.62s/it]
Continual predict:  75%|███████▍  | 818/1097 [15:32<10:24,  2.24s/it]
Continual predict:  75%|███████▍  | 819/1097 [15:35<10:25,  2.25s/it]
Continual predict:  75%|███████▍  | 820/1097 [15:36<09:38,  2.09s/it]
Continual predict:  75%|███████▍  | 821/1097 [15:38<09:25,  2.05s/it]
Continual predict:  75%|███████▍  | 822/1097 [15:41<10:26,  2.28s/it]
Continual predict:  75%|███████▌  | 823/1097 [15:45<12:13,  2.68s/it]
Continual predict:  75%|███████▌  | 824/1097 [15:55<22:39,  4.98s/it]
Continual predict:  75%|███████▌  | 825/1097 [15:57<18:11,  4.01s/it]
Continual predict:  75%|███████▌  | 826/1097 [16:01<18:06,  4.01s/it]
Continual predict:  75%|███████▌  | 827/1097 [16:02<14:12,  3.16s/it]
Continual predict:  75%|███████▌  | 828/1097 [16:06<15:41,  3.50s/it]
Continual predict:  76%|███████▌  | 829/1097 [16:12<18:18,  4.10s/it]
Continual predict:  76%|███████▌  | 830/1097 [16:14<15:30,  3.49s/it]
Continual predict:  76%|███████▌  | 831/1097 [16:26<26:29,  5.97s/it]
Continual predict:  76%|███████▌  | 832/1097 [16:27<20:27,  4.63s/it]
Continual predict:  76%|███████▌  | 833/1097 [16:28<15:30,  3.53s/it]
Continual predict:  76%|███████▌  | 834/1097 [16:30<13:35,  3.10s/it]
Continual predict:  76%|███████▌  | 835/1097 [16:32<11:19,  2.59s/it]
Continual predict:  76%|███████▌  | 836/1097 [16:34<11:21,  2.61s/it]
Continual predict:  76%|███████▋  | 837/1097 [16:39<14:27,  3.34s/it]
Continual predict:  76%|███████▋  | 838/1097 [16:42<12:56,  3.00s/it]
Continual predict:  76%|███████▋  | 839/1097 [16:48<17:51,  4.15s/it]
Continual predict:  77%|███████▋  | 840/1097 [16:53<18:30,  4.32s/it]
Continual predict:  77%|███████▋  | 841/1097 [16:55<14:57,  3.51s/it]
Continual predict:  77%|███████▋  | 842/1097 [16:58<14:07,  3.32s/it]
Continual predict:  77%|███████▋  | 843/1097 [17:00<12:18,  2.91s/it]
Continual predict:  77%|███████▋  | 844/1097 [17:05<14:59,  3.56s/it]
Continual predict:  77%|███████▋  | 845/1097 [17:09<15:25,  3.67s/it]
Continual predict:  77%|███████▋  | 846/1097 [17:10<12:40,  3.03s/it]
Continual predict:  77%|███████▋  | 847/1097 [17:11<09:43,  2.33s/it]
Continual predict:  77%|███████▋  | 848/1097 [17:12<08:47,  2.12s/it]
Continual predict:  77%|███████▋  | 849/1097 [17:14<07:40,  1.86s/it]
Continual predict:  77%|███████▋  | 850/1097 [17:17<09:19,  2.27s/it]
Continual predict:  78%|███████▊  | 851/1097 [17:19<09:02,  2.20s/it]
Continual predict:  78%|███████▊  | 852/1097 [17:20<07:52,  1.93s/it]
Continual predict:  78%|███████▊  | 853/1097 [17:30<17:20,  4.26s/it]
Continual predict:  78%|███████▊  | 854/1097 [17:42<26:23,  6.52s/it]
Continual predict:  78%|███████▊  | 855/1097 [17:44<21:17,  5.28s/it]
Continual predict:  78%|███████▊  | 856/1097 [17:46<17:30,  4.36s/it]
Continual predict:  78%|███████▊  | 857/1097 [17:48<14:23,  3.60s/it]
Continual predict:  78%|███████▊  | 858/1097 [17:50<12:26,  3.12s/it]
Continual predict:  78%|███████▊  | 859/1097 [17:55<13:49,  3.49s/it]
Continual predict:  78%|███████▊  | 860/1097 [17:59<14:27,  3.66s/it]
Continual predict:  78%|███████▊  | 861/1097 [18:10<23:58,  6.10s/it]
Continual predict:  79%|███████▊  | 862/1097 [18:11<17:26,  4.45s/it]
Continual predict:  79%|███████▊  | 863/1097 [18:13<14:37,  3.75s/it]
Continual predict:  79%|███████▉  | 864/1097 [18:15<12:05,  3.11s/it]
Continual predict:  79%|███████▉  | 865/1097 [18:17<11:02,  2.86s/it]
Continual predict:  79%|███████▉  | 866/1097 [18:19<09:26,  2.45s/it]
Continual predict:  79%|███████▉  | 867/1097 [18:21<09:23,  2.45s/it]
Continual predict:  79%|███████▉  | 868/1097 [18:24<10:18,  2.70s/it]
Continual predict:  79%|███████▉  | 869/1097 [18:28<11:17,  2.97s/it]
Continual predict:  79%|███████▉  | 870/1097 [18:29<09:17,  2.46s/it]
Continual predict:  79%|███████▉  | 871/1097 [18:31<08:18,  2.21s/it]
Continual predict:  79%|███████▉  | 872/1097 [18:36<11:44,  3.13s/it]
Continual predict:  80%|███████▉  | 873/1097 [18:38<10:44,  2.88s/it]
Continual predict:  80%|███████▉  | 874/1097 [18:40<09:03,  2.44s/it]
Continual predict:  80%|███████▉  | 875/1097 [18:42<08:28,  2.29s/it]
Continual predict:  80%|███████▉  | 876/1097 [18:43<07:33,  2.05s/it]
Continual predict:  80%|███████▉  | 877/1097 [18:46<07:56,  2.16s/it]
Continual predict:  80%|████████  | 878/1097 [18:46<05:47,  1.59s/it]
Continual predict:  80%|████████  | 879/1097 [18:49<07:27,  2.05s/it]
Continual predict:  80%|████████  | 880/1097 [18:51<06:59,  1.93s/it]
Continual predict:  80%|████████  | 881/1097 [18:54<08:39,  2.40s/it]
Continual predict:  80%|████████  | 882/1097 [18:56<08:27,  2.36s/it]
Continual predict:  80%|████████  | 883/1097 [18:58<07:35,  2.13s/it]
Continual predict:  81%|████████  | 884/1097 [19:02<09:04,  2.56s/it]
Continual predict:  81%|████████  | 885/1097 [19:07<11:41,  3.31s/it]
Continual predict:  81%|████████  | 886/1097 [19:09<10:22,  2.95s/it]
Continual predict:  81%|████████  | 887/1097 [19:12<10:27,  2.99s/it]
Continual predict:  81%|████████  | 888/1097 [19:14<09:25,  2.70s/it]
Continual predict:  81%|████████  | 889/1097 [19:16<08:51,  2.56s/it]
Continual predict:  81%|████████  | 890/1097 [19:19<08:58,  2.60s/it]
Continual predict:  81%|████████  | 891/1097 [19:22<09:15,  2.69s/it]
Continual predict:  81%|████████▏ | 892/1097 [19:24<09:07,  2.67s/it]
Continual predict:  81%|████████▏ | 893/1097 [19:27<09:13,  2.71s/it]
Continual predict:  81%|████████▏ | 894/1097 [19:29<08:16,  2.45s/it]
Continual predict:  82%|████████▏ | 895/1097 [19:33<10:19,  3.07s/it]
Continual predict:  82%|████████▏ | 896/1097 [19:35<09:07,  2.73s/it]
Continual predict:  82%|████████▏ | 897/1097 [19:40<10:49,  3.25s/it]
Continual predict:  82%|████████▏ | 898/1097 [19:40<07:43,  2.33s/it]
Continual predict:  82%|████████▏ | 899/1097 [19:40<05:37,  1.71s/it]
Continual predict:  82%|████████▏ | 900/1097 [19:41<04:13,  1.29s/it]
Continual predict:  82%|████████▏ | 901/1097 [19:41<03:14,  1.01it/s]
Continual predict:  82%|████████▏ | 902/1097 [19:41<02:23,  1.36it/s]
Continual predict:  82%|████████▏ | 903/1097 [19:41<01:46,  1.81it/s]
Continual predict:  82%|████████▏ | 904/1097 [19:41<01:28,  2.17it/s]
Continual predict:  82%|████████▏ | 905/1097 [19:42<01:16,  2.52it/s]
Continual predict:  83%|████████▎ | 906/1097 [19:42<01:00,  3.17it/s]
Continual predict:  83%|████████▎ | 907/1097 [19:42<00:56,  3.38it/s]
Continual predict:  83%|████████▎ | 908/1097 [19:42<00:49,  3.80it/s]
Continual predict:  83%|████████▎ | 909/1097 [19:42<00:48,  3.86it/s]
Continual predict:  83%|████████▎ | 910/1097 [19:43<00:44,  4.22it/s]
Continual predict:  83%|████████▎ | 911/1097 [19:43<00:41,  4.50it/s]
Continual predict:  83%|████████▎ | 912/1097 [19:43<00:35,  5.17it/s]
Continual predict:  83%|████████▎ | 913/1097 [19:43<00:35,  5.21it/s]
Continual predict:  83%|████████▎ | 914/1097 [19:43<00:41,  4.37it/s]
Continual predict:  83%|████████▎ | 915/1097 [19:44<00:39,  4.63it/s]
Continual predict:  84%|████████▎ | 916/1097 [19:44<00:40,  4.43it/s]
Continual predict:  84%|████████▎ | 917/1097 [19:44<00:41,  4.29it/s]
Continual predict:  84%|████████▎ | 918/1097 [19:44<00:35,  4.99it/s]
Continual predict:  84%|████████▍ | 919/1097 [19:44<00:38,  4.65it/s]
Continual predict:  84%|████████▍ | 920/1097 [19:45<00:43,  4.10it/s]
Continual predict:  84%|████████▍ | 921/1097 [19:45<00:45,  3.91it/s]
Continual predict:  84%|████████▍ | 922/1097 [19:45<00:38,  4.51it/s]
Continual predict:  84%|████████▍ | 923/1097 [19:46<00:42,  4.13it/s]
Continual predict:  84%|████████▍ | 924/1097 [19:46<00:48,  3.54it/s]
Continual predict:  84%|████████▍ | 925/1097 [19:46<00:46,  3.66it/s]
Continual predict:  84%|████████▍ | 926/1097 [19:46<00:39,  4.37it/s]
Continual predict:  85%|████████▍ | 927/1097 [19:46<00:36,  4.61it/s]
Continual predict:  85%|████████▍ | 928/1097 [19:47<00:32,  5.27it/s]
Continual predict:  85%|████████▍ | 929/1097 [19:47<00:35,  4.79it/s]
Continual predict:  85%|████████▍ | 930/1097 [19:47<00:36,  4.52it/s]
Continual predict:  85%|████████▍ | 931/1097 [19:47<00:41,  4.01it/s]
Continual predict:  85%|████████▍ | 932/1097 [19:48<00:38,  4.33it/s]
Continual predict:  85%|████████▌ | 933/1097 [19:48<00:42,  3.90it/s]
Continual predict:  85%|████████▌ | 934/1097 [19:48<00:35,  4.60it/s]
Continual predict:  85%|████████▌ | 935/1097 [19:48<00:40,  4.05it/s]
Continual predict:  85%|████████▌ | 936/1097 [19:49<00:42,  3.75it/s]
Continual predict:  85%|████████▌ | 937/1097 [19:49<00:44,  3.57it/s]
Continual predict:  86%|████████▌ | 938/1097 [19:49<00:37,  4.28it/s]
Continual predict:  86%|████████▌ | 939/1097 [19:49<00:37,  4.19it/s]
Continual predict:  86%|████████▌ | 940/1097 [19:50<00:35,  4.48it/s]
Continual predict:  86%|████████▌ | 941/1097 [19:50<00:41,  3.74it/s]
Continual predict:  86%|████████▌ | 942/1097 [19:50<00:40,  3.82it/s]
Continual predict:  86%|████████▌ | 943/1097 [19:50<00:37,  4.16it/s]
Continual predict:  86%|████████▌ | 944/1097 [19:50<00:31,  4.86it/s]
Continual predict:  86%|████████▌ | 945/1097 [19:51<00:27,  5.51it/s]
Continual predict:  86%|████████▌ | 946/1097 [19:51<00:27,  5.46it/s]
Continual predict:  86%|████████▋ | 947/1097 [19:51<00:35,  4.18it/s]
Continual predict:  86%|████████▋ | 948/1097 [19:51<00:36,  4.12it/s]
Continual predict:  87%|████████▋ | 949/1097 [19:52<00:41,  3.56it/s]
Continual predict:  87%|████████▋ | 950/1097 [19:52<00:34,  4.27it/s]
Continual predict:  87%|████████▋ | 951/1097 [19:52<00:32,  4.55it/s]
Continual predict:  87%|████████▋ | 952/1097 [19:52<00:27,  5.23it/s]
Continual predict:  87%|████████▋ | 953/1097 [19:52<00:27,  5.26it/s]
Continual predict:  87%|████████▋ | 954/1097 [19:53<00:24,  5.87it/s]
Continual predict:  87%|████████▋ | 955/1097 [19:53<00:30,  4.69it/s]
Continual predict:  87%|████████▋ | 956/1097 [19:53<00:26,  5.36it/s]
Continual predict:  87%|████████▋ | 957/1097 [19:53<00:26,  5.36it/s]
Continual predict:  87%|████████▋ | 958/1097 [19:53<00:25,  5.37it/s]
Continual predict:  87%|████████▋ | 959/1097 [19:54<00:28,  4.87it/s]
Continual predict:  88%|████████▊ | 960/1097 [19:54<00:24,  5.51it/s]
Continual predict:  88%|████████▊ | 961/1097 [19:54<00:24,  5.46it/s]
Continual predict:  88%|████████▊ | 962/1097 [19:54<00:27,  4.93it/s]
Continual predict:  88%|████████▊ | 963/1097 [19:54<00:26,  5.05it/s]
Continual predict:  88%|████████▊ | 964/1097 [19:54<00:23,  5.66it/s]
Continual predict:  88%|████████▊ | 965/1097 [19:55<00:26,  5.04it/s]
Continual predict:  88%|████████▊ | 966/1097 [19:55<00:27,  4.68it/s]
Continual predict:  88%|████████▊ | 967/1097 [19:55<00:31,  4.11it/s]
Continual predict:  88%|████████▊ | 968/1097 [19:56<00:37,  3.44it/s]
Continual predict:  88%|████████▊ | 969/1097 [19:56<00:52,  2.43it/s]
Continual predict:  88%|████████▊ | 970/1097 [19:57<00:50,  2.50it/s]
Continual predict:  89%|████████▊ | 971/1097 [19:57<00:47,  2.68it/s]
Continual predict:  89%|████████▊ | 972/1097 [19:57<00:41,  2.98it/s]
Continual predict:  89%|████████▊ | 973/1097 [19:58<00:38,  3.22it/s]
Continual predict:  89%|████████▉ | 974/1097 [19:58<00:31,  3.92it/s]
Continual predict:  89%|████████▉ | 975/1097 [19:58<00:28,  4.25it/s]
Continual predict:  89%|████████▉ | 976/1097 [19:58<00:31,  3.88it/s]
Continual predict:  89%|████████▉ | 977/1097 [19:58<00:28,  4.23it/s]
Continual predict:  89%|████████▉ | 978/1097 [19:58<00:24,  4.93it/s]
Continual predict:  89%|████████▉ | 979/1097 [19:59<00:25,  4.61it/s]
Continual predict:  89%|████████▉ | 980/1097 [19:59<00:22,  5.27it/s]
Continual predict:  89%|████████▉ | 981/1097 [19:59<00:24,  4.83it/s]
Continual predict:  90%|████████▉ | 982/1097 [19:59<00:25,  4.56it/s]
Continual predict:  90%|████████▉ | 983/1097 [20:00<00:23,  4.77it/s]
Continual predict:  90%|████████▉ | 984/1097 [20:00<00:20,  5.42it/s]
Continual predict:  90%|████████▉ | 985/1097 [20:00<00:27,  4.14it/s]
Continual predict:  90%|████████▉ | 986/1097 [20:00<00:22,  4.85it/s]
Continual predict:  90%|████████▉ | 987/1097 [20:00<00:20,  5.49it/s]
Continual predict:  90%|█████████ | 988/1097 [20:00<00:17,  6.06it/s]
Continual predict:  90%|█████████ | 989/1097 [20:01<00:20,  5.26it/s]
Continual predict:  90%|█████████ | 990/1097 [20:01<00:24,  4.42it/s]
Continual predict:  90%|█████████ | 991/1097 [20:01<00:20,  5.11it/s]
Continual predict:  90%|█████████ | 992/1097 [20:01<00:20,  5.19it/s]
Continual predict:  91%|█████████ | 993/1097 [20:01<00:19,  5.24it/s]
Continual predict:  91%|█████████ | 994/1097 [20:02<00:17,  5.85it/s]
Continual predict:  91%|█████████ | 995/1097 [20:02<00:16,  6.34it/s]
Continual predict:  91%|█████████ | 996/1097 [20:02<00:14,  6.75it/s]
Continual predict:  91%|█████████ | 997/1097 [20:02<00:15,  6.27it/s]
Continual predict:  91%|█████████ | 998/1097 [20:02<00:14,  6.70it/s]
Continual predict:  91%|█████████ | 999/1097 [20:02<00:13,  7.04it/s]
Continual predict:  91%|█████████ | 1000/1097 [20:02<00:13,  7.27it/s]
Continual predict:  91%|█████████ | 1001/1097 [20:03<00:12,  7.47it/s]
Continual predict:  91%|█████████▏| 1002/1097 [20:03<00:12,  7.61it/s]
Continual predict:  91%|█████████▏| 1003/1097 [20:03<00:12,  7.66it/s]
Continual predict:  92%|█████████▏| 1004/1097 [20:03<00:12,  7.74it/s]
Continual predict:  92%|█████████▏| 1005/1097 [20:03<00:11,  7.80it/s]
Continual predict:  92%|█████████▏| 1006/1097 [20:03<00:11,  7.84it/s]
Continual predict:  92%|█████████▏| 1007/1097 [20:03<00:11,  7.79it/s]
Continual predict:  92%|█████████▏| 1008/1097 [20:03<00:11,  7.82it/s]
Continual predict:  92%|█████████▏| 1009/1097 [20:04<00:11,  7.87it/s]
Continual predict:  92%|█████████▏| 1010/1097 [20:04<00:11,  7.89it/s]
Continual predict:  92%|█████████▏| 1011/1097 [20:04<00:10,  7.87it/s]
Continual predict:  92%|█████████▏| 1012/1097 [20:04<00:11,  7.39it/s]
Continual predict:  92%|█████████▏| 1013/1097 [20:04<00:11,  7.55it/s]
Continual predict:  92%|█████████▏| 1014/1097 [20:04<00:10,  7.65it/s]
Continual predict:  93%|█████████▎| 1015/1097 [20:04<00:10,  7.74it/s]
Continual predict:  93%|█████████▎| 1016/1097 [20:04<00:10,  7.79it/s]
Continual predict:  93%|█████████▎| 1017/1097 [20:05<00:10,  7.82it/s]
Continual predict:  93%|█████████▎| 1018/1097 [20:05<00:10,  7.87it/s]
Continual predict:  93%|█████████▎| 1019/1097 [20:05<00:10,  7.78it/s]
Continual predict:  93%|█████████▎| 1020/1097 [20:05<00:09,  7.82it/s]
Continual predict:  93%|█████████▎| 1021/1097 [20:05<00:09,  7.88it/s]
Continual predict:  93%|█████████▎| 1022/1097 [20:05<00:09,  7.89it/s]
Continual predict:  93%|█████████▎| 1023/1097 [20:05<00:09,  7.82it/s]
Continual predict:  93%|█████████▎| 1024/1097 [20:05<00:09,  7.78it/s]
Continual predict:  93%|█████████▎| 1025/1097 [20:06<00:09,  7.74it/s]
Continual predict:  94%|█████████▎| 1026/1097 [20:06<00:09,  7.79it/s]
Continual predict:  94%|█████████▎| 1027/1097 [20:06<00:08,  7.84it/s]
Continual predict:  94%|█████████▎| 1028/1097 [20:06<00:08,  7.79it/s]
Continual predict:  94%|█████████▍| 1029/1097 [20:06<00:09,  7.54it/s]
Continual predict:  94%|█████████▍| 1030/1097 [20:06<00:09,  7.40it/s]
Continual predict:  94%|█████████▍| 1031/1097 [20:06<00:09,  7.18it/s]
Continual predict:  94%|█████████▍| 1032/1097 [20:07<00:08,  7.34it/s]
Continual predict:  94%|█████████▍| 1033/1097 [20:07<00:08,  7.42it/s]
Continual predict:  94%|█████████▍| 1034/1097 [20:07<00:08,  7.45it/s]
Continual predict:  94%|█████████▍| 1035/1097 [20:07<00:08,  7.55it/s]
Continual predict:  94%|█████████▍| 1036/1097 [20:07<00:07,  7.64it/s]
Continual predict:  95%|█████████▍| 1037/1097 [20:07<00:07,  7.71it/s]
Continual predict:  95%|█████████▍| 1038/1097 [20:07<00:07,  7.77it/s]
Continual predict:  95%|█████████▍| 1039/1097 [20:07<00:07,  7.78it/s]
Continual predict:  95%|█████████▍| 1040/1097 [20:08<00:07,  7.82it/s]
Continual predict:  95%|█████████▍| 1041/1097 [20:08<00:07,  7.82it/s]
Continual predict:  95%|█████████▍| 1042/1097 [20:08<00:07,  7.82it/s]
Continual predict:  95%|█████████▌| 1043/1097 [20:08<00:06,  7.72it/s]
Continual predict:  95%|█████████▌| 1044/1097 [20:08<00:06,  7.76it/s]
Continual predict:  95%|█████████▌| 1045/1097 [20:08<00:06,  7.81it/s]
Continual predict:  95%|█████████▌| 1046/1097 [20:08<00:06,  7.84it/s]
Continual predict:  95%|█████████▌| 1047/1097 [20:08<00:06,  7.86it/s]
Continual predict:  96%|█████████▌| 1048/1097 [20:09<00:06,  7.87it/s]
Continual predict:  96%|█████████▌| 1049/1097 [20:09<00:06,  7.88it/s]
Continual predict:  96%|█████████▌| 1050/1097 [20:09<00:05,  7.85it/s]
Continual predict:  96%|█████████▌| 1051/1097 [20:09<00:05,  7.88it/s]
Continual predict:  96%|█████████▌| 1052/1097 [20:09<00:05,  7.78it/s]
Continual predict:  96%|█████████▌| 1053/1097 [20:09<00:05,  7.82it/s]
Continual predict:  96%|█████████▌| 1054/1097 [20:09<00:05,  7.77it/s]
Continual predict:  96%|█████████▌| 1055/1097 [20:10<00:05,  7.82it/s]
Continual predict:  96%|█████████▋| 1056/1097 [20:10<00:05,  7.88it/s]
Continual predict:  96%|█████████▋| 1057/1097 [20:10<00:05,  7.90it/s]
Continual predict:  96%|█████████▋| 1058/1097 [20:10<00:04,  7.92it/s]
Continual predict:  97%|█████████▋| 1059/1097 [20:10<00:04,  7.93it/s]
Continual predict:  97%|█████████▋| 1060/1097 [20:10<00:04,  7.89it/s]
Continual predict:  97%|█████████▋| 1061/1097 [20:10<00:04,  7.88it/s]
Continual predict:  97%|█████████▋| 1062/1097 [20:10<00:04,  7.90it/s]
Continual predict:  97%|█████████▋| 1063/1097 [20:11<00:04,  7.78it/s]
Continual predict:  97%|█████████▋| 1064/1097 [20:11<00:04,  7.83it/s]
Continual predict:  97%|█████████▋| 1065/1097 [20:11<00:04,  7.85it/s]
Continual predict:  97%|█████████▋| 1066/1097 [20:11<00:03,  7.87it/s]
Continual predict:  97%|█████████▋| 1067/1097 [20:11<00:03,  7.81it/s]
Continual predict:  97%|█████████▋| 1068/1097 [20:11<00:03,  7.84it/s]
Continual predict:  97%|█████████▋| 1069/1097 [20:11<00:03,  7.85it/s]
Continual predict:  98%|█████████▊| 1070/1097 [20:11<00:03,  7.87it/s]
Continual predict:  98%|█████████▊| 1071/1097 [20:12<00:03,  7.89it/s]
Continual predict:  98%|█████████▊| 1072/1097 [20:12<00:03,  7.89it/s]
Continual predict:  98%|█████████▊| 1073/1097 [20:12<00:03,  7.89it/s]
Continual predict:  98%|█████████▊| 1074/1097 [20:12<00:02,  7.91it/s]
Continual predict:  98%|█████████▊| 1075/1097 [20:12<00:02,  7.92it/s]
Continual predict:  98%|█████████▊| 1076/1097 [20:12<00:02,  7.90it/s]
Continual predict:  98%|█████████▊| 1077/1097 [20:12<00:02,  7.89it/s]
Continual predict:  98%|█████████▊| 1078/1097 [20:12<00:02,  7.91it/s]
Continual predict:  98%|█████████▊| 1079/1097 [20:13<00:02,  7.92it/s]
Continual predict:  98%|█████████▊| 1080/1097 [20:13<00:02,  7.94it/s]
Continual predict:  99%|█████████▊| 1081/1097 [20:13<00:02,  7.82it/s]
Continual predict:  99%|█████████▊| 1082/1097 [20:13<00:01,  7.85it/s]
Continual predict:  99%|█████████▊| 1083/1097 [20:13<00:01,  7.88it/s]
Continual predict:  99%|█████████▉| 1084/1097 [20:13<00:01,  7.89it/s]
Continual predict:  99%|█████████▉| 1085/1097 [20:13<00:01,  7.90it/s]
Continual predict:  99%|█████████▉| 1086/1097 [20:13<00:01,  7.91it/s]
Continual predict:  99%|█████████▉| 1087/1097 [20:14<00:01,  7.93it/s]
Continual predict:  99%|█████████▉| 1088/1097 [20:14<00:01,  7.94it/s]
Continual predict:  99%|█████████▉| 1089/1097 [20:14<00:01,  7.95it/s]
Continual predict:  99%|█████████▉| 1090/1097 [20:14<00:00,  7.63it/s]
Continual predict:  99%|█████████▉| 1091/1097 [20:14<00:00,  7.69it/s]
Continual predict: 100%|█████████▉| 1092/1097 [20:14<00:00,  7.73it/s]
Continual predict: 100%|█████████▉| 1093/1097 [20:14<00:00,  7.75it/s]
Continual predict: 100%|█████████▉| 1094/1097 [20:14<00:00,  7.79it/s]
Continual predict: 100%|█████████▉| 1095/1097 [20:15<00:00,  7.82it/s]
Continual predict: 100%|█████████▉| 1096/1097 [20:15<00:00,  7.84it/s]
Continual predict: 100%|██████████| 1097/1097 [20:15<00:00,  7.86it/s]
Continual predict: 100%|██████████| 1097/1097 [20:15<00:00,  1.11s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot11=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot13=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=16]│
+  │  GT=task1687_sentiment140_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot10=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot12=100]│
+  │  GT=task511_reddit_tifu_long_text_su  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task591_sciq_answer_generation    slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=18]│
+  │  GT=task748_glucose_reverse_cause_ev  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (1097/1097)  label_acc=100.0% (1097/1097)  disagree=  0                │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                                 =       15.0
+  predict_eval_rougeL                                                   =    60.3107
+  predict_eval_rougeL_for_CL                                            =    60.3107
+  predict_eval_rougeL_for_task002_quoref_answer_generation              =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                   =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction        =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                       =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation            =    11.1335
+  predict_eval_rougeL_for_task1687_sentiment140_classification          =       84.0
+  predict_eval_rougeL_for_task1729_personachat_generate_next            =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                    =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification          =       91.0
+  predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization   =    21.3091
+  predict_eval_rougeL_for_task591_sciq_answer_generation                =    72.7381
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation   =    15.2912
+  predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection =    66.9039
+  predict_exact_match                                                   =    45.0319
+  predict_exact_match_for_CL                                            =    45.0319
+  predict_exact_match_for_task002_quoref_answer_generation              =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                   =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction        =       98.0
+  predict_exact_match_for_task1572_samsum_summary                       =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation            =        0.0
+  predict_exact_match_for_task1687_sentiment140_classification          =       84.0
+  predict_exact_match_for_task1729_personachat_generate_next            =        0.0
+  predict_exact_match_for_task181_outcome_extraction                    =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification          =       91.0
+  predict_exact_match_for_task511_reddit_tifu_long_text_summarization   =        1.0
+  predict_exact_match_for_task591_sciq_answer_generation                =       60.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation   =        0.0
+  predict_exact_match_for_task748_glucose_reverse_cause_event_detection =        3.0
+  predict_gen_len                                                       =     9.1887
+  predict_global_step                                                   =        480
+  predict_rouge1                                                        =    61.8839
+  predict_rouge1_for_CL                                                 =    61.8839
+  predict_rouge1_for_task002_quoref_answer_generation                   =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation            =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                        =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction             =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                            =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation                 =    13.7971
+  predict_rouge1_for_task1687_sentiment140_classification               =       84.0
+  predict_rouge1_for_task1729_personachat_generate_next                 =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                         =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification               =       91.0
+  predict_rouge1_for_task511_reddit_tifu_long_text_summarization        =    25.5967
+  predict_rouge1_for_task591_sciq_answer_generation                     =    73.4048
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation        =    15.2912
+  predict_rouge1_for_task748_glucose_reverse_cause_event_detection      =     68.907
+  predict_runtime                                                       = 0:20:15.34
+  predict_samples                                                       =       1097
+  predict_samples_per_second                                            =      0.903
+  predict_steps_per_second                                              =      0.903
+[2026-04-30 06:50:34,498] [INFO] [launch.py:367:main] Process 1923254 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/15-task875_emotion_classification/all_results.json b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..0a213cfcc4d15ef8ce95c87c2832d2aa3f7d4c81
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/all_results.json
@@ -0,0 +1,66 @@
+{
+  "predict_exact_match": 47.3684,
+  "predict_rouge1": 62.8126,
+  "predict_eval_rougeL": 61.3708,
+  "predict_gen_len": 8.523,
+  "predict_runtime": 1249.5441576598678,
+  "predict_samples": 1197,
+  "predict_samples_per_second": 0.958,
+  "predict_steps_per_second": 0.958,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 47.3684,
+  "predict_rouge1_for_CL": 62.8126,
+  "predict_eval_rougeL_for_CL": 61.3708,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_exact_match_for_task1590_diplomacy_text_generation": 0.0,
+  "predict_rouge1_for_task1590_diplomacy_text_generation": 13.7971,
+  "predict_eval_rougeL_for_task1590_diplomacy_text_generation": 11.1335,
+  "predict_exact_match_for_task748_glucose_reverse_cause_event_detection": 3.0,
+  "predict_rouge1_for_task748_glucose_reverse_cause_event_detection": 68.907,
+  "predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection": 66.9039,
+  "predict_exact_match_for_task511_reddit_tifu_long_text_summarization": 1.0,
+  "predict_rouge1_for_task511_reddit_tifu_long_text_summarization": 25.5967,
+  "predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization": 21.3091,
+  "predict_exact_match_for_task591_sciq_answer_generation": 60.0,
+  "predict_rouge1_for_task591_sciq_answer_generation": 73.4048,
+  "predict_eval_rougeL_for_task591_sciq_answer_generation": 72.7381,
+  "predict_exact_match_for_task1687_sentiment140_classification": 84.0,
+  "predict_rouge1_for_task1687_sentiment140_classification": 84.0,
+  "predict_eval_rougeL_for_task1687_sentiment140_classification": 84.0,
+  "predict_exact_match_for_task875_emotion_classification": 73.0,
+  "predict_rouge1_for_task875_emotion_classification": 73.0,
+  "predict_eval_rougeL_for_task875_emotion_classification": 73.0,
+  "train_runtime": 2457.492,
+  "train_samples_per_second": 6.104,
+  "train_steps_per_second": 0.195,
+  "total_flos": 5.441245561015501e+16,
+  "train_loss": 0.5571656028429667,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0021848d2c34567701db9c50a3eb07605a11716c
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df3502005949993d999273a6597f58ff388ea0f42ecad1ed50b2c97b9bdcb50c
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7b1a994a10ed956bc884308c2f17f4b96d77b22f
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:236a554f4f03dd8e0bf8337cf3a5dcfe00f5f1d3736226935de5e57d36d60f8d
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..ed21fc3e6333fe9150ec187bf384abf023b4688f
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1479d7426a57786411a46d3f3fee046ef61eacf05a5049acdda40f7820805c97
+size 2300106287
diff --git a/superni_order1_llama_srt/outputs/15-task875_emotion_classification/train.log b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..1b4ced2fbca456f492ca15876669b00bf4b8479c
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/15-task875_emotion_classification/train.log
@@ -0,0 +1,661 @@
+[2026-04-30 06:50:40,390] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 06:50:40,390] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29514 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task875_emotion_classification --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/15-task875_emotion_classification --cur_task_id 14 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights
+[2026-04-30 06:50:43,636] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 06:50:43,636] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 06:50:43,636] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 06:50:43,636] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 06:50:43,636] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 06:50:43,637] [INFO] [launch.py:272:main] process 2076791 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task875_emotion_classification', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/15-task875_emotion_classification', '--cur_task_id', '14', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights']
+
+============================================================
+Task 14: task875_emotion_classification
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task875_emotion_classification/train.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task875_emotion_classification/dev.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task875_emotion_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+Train/dev/test: 1000/100/1197
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 120.69it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 152.91it/s]
Loading weights:  17%|█▋        | 50/291 [00:00<00:01, 166.72it/s]
Loading weights:  25%|██▌       | 73/291 [00:00<00:01, 186.70it/s]
Loading weights:  32%|███▏      | 92/291 [00:00<00:01, 174.50it/s]
Loading weights:  38%|███▊      | 110/291 [00:00<00:01, 175.62it/s]
Loading weights:  44%|████▍     | 128/291 [00:00<00:00, 176.03it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 176.63it/s]
Loading weights:  56%|█████▋    | 164/291 [00:00<00:00, 171.01it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 177.30it/s]
Loading weights:  69%|██████▉   | 202/291 [00:01<00:00, 174.70it/s]
Loading weights:  76%|███████▌  | 220/291 [00:01<00:00, 175.18it/s]
Loading weights:  82%|████████▏ | 240/291 [00:01<00:00, 178.79it/s]
Loading weights:  90%|████████▉ | 261/291 [00:01<00:00, 186.53it/s]
Loading weights:  96%|█████████▌| 280/291 [00:01<00:00, 181.26it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 177.78it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1955)
+[FIX] Re-initialized lora_A/B in 960 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(14, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/13-task591_sciq_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/12-task511_reddit_tifu_long_text_summarization/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/11-task748_glucose_reverse_cause_event_detection/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/10-task1590_diplomacy_text_generation/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[8] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[9] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[10] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[11] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[12] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[13] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 14 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6770.754M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 14 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/14-task1687_sentiment140_classification/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 14 sigs, cur=task875_emotion_classification, mapping={'task875_emotion_classification': 0, 'task1572_samsum_summary': 14, 'task363_sst2_polarity_classification': 13, 'task1290_xsum_summarization': 12, 'task181_outcome_extraction': 11, 'task002_quoref_answer_generation': 10, 'task1510_evalution_relation_extraction': 9, 'task639_multi_woz_user_utterance_generation': 8, 'task1729_personachat_generate_next': 7, 'task073_commonsenseqa_answer_generation': 6, 'task1590_diplomacy_text_generation': 5, 'task748_glucose_reverse_cause_event_detection': 4, 'task511_reddit_tifu_long_text_summarization': 3, 'task591_sciq_answer_generation': 2, 'task1687_sentiment140_classification': 1}
+[SGWI] Task 14: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 06:51:53,695] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:05<45:03,  5.64s/it]
  0%|          | 2/480 [00:10<42:33,  5.34s/it]
  1%|          | 3/480 [00:15<41:33,  5.23s/it]
  1%|          | 4/480 [00:20<40:45,  5.14s/it]
  1%|          | 5/480 [00:25<40:33,  5.12s/it]
  1%|▏         | 6/480 [00:31<40:50,  5.17s/it]
  1%|▏         | 7/480 [00:36<40:34,  5.15s/it]
  2%|▏         | 8/480 [00:41<40:12,  5.11s/it]
  2%|▏         | 9/480 [00:46<40:08,  5.11s/it]
  2%|▏         | 10/480 [00:51<40:14,  5.14s/it]
                                                
{'loss': '4.215', 'grad_norm': '0.8986', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:51<40:14,  5.14s/it]
  2%|▏         | 11/480 [00:57<40:39,  5.20s/it]
  2%|▎         | 12/480 [01:02<41:20,  5.30s/it]
  3%|▎         | 13/480 [01:07<40:22,  5.19s/it]
  3%|▎         | 14/480 [01:12<39:59,  5.15s/it]
  3%|▎         | 15/480 [01:17<39:23,  5.08s/it]
  3%|▎         | 16/480 [01:22<39:08,  5.06s/it]
  4%|▎         | 17/480 [01:27<38:44,  5.02s/it]
  4%|▍         | 18/480 [01:32<38:44,  5.03s/it]
  4%|▍         | 19/480 [01:37<38:26,  5.00s/it]
  4%|▍         | 20/480 [01:42<39:02,  5.09s/it]
                                                
{'loss': '1.712', 'grad_norm': '1.052', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:42<39:02,  5.09s/it]
  4%|▍         | 21/480 [01:47<38:37,  5.05s/it]
  5%|▍         | 22/480 [01:52<38:40,  5.07s/it]
  5%|▍         | 23/480 [01:57<38:43,  5.09s/it]
  5%|▌         | 24/480 [02:03<38:49,  5.11s/it]
  5%|▌         | 25/480 [02:08<39:04,  5.15s/it]
  5%|▌         | 26/480 [02:13<38:50,  5.13s/it]
  6%|▌         | 27/480 [02:18<38:21,  5.08s/it]
  6%|▌         | 28/480 [02:23<38:48,  5.15s/it]
  6%|▌         | 29/480 [02:28<38:15,  5.09s/it]
  6%|▋         | 30/480 [02:33<38:24,  5.12s/it]
                                                
{'loss': '0.7713', 'grad_norm': '0.6689', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [02:33<38:24,  5.12s/it]
  6%|▋         | 31/480 [02:38<38:20,  5.12s/it]
  7%|▋         | 32/480 [02:40<29:45,  3.99s/it]
  7%|▋         | 33/480 [02:45<32:58,  4.43s/it]
  7%|▋         | 34/480 [02:51<35:09,  4.73s/it]
  7%|▋         | 35/480 [02:56<35:48,  4.83s/it]
  8%|▊         | 36/480 [03:01<36:26,  4.92s/it]
  8%|▊         | 37/480 [03:06<37:04,  5.02s/it]
  8%|▊         | 38/480 [03:11<37:25,  5.08s/it]
  8%|▊         | 39/480 [03:17<38:13,  5.20s/it]
  8%|▊         | 40/480 [03:22<38:28,  5.25s/it]
                                                
{'loss': '0.6954', 'grad_norm': '0.6623', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [03:22<38:28,  5.25s/it]
  9%|▊         | 41/480 [03:27<37:53,  5.18s/it]
  9%|▉         | 42/480 [03:32<37:25,  5.13s/it]
  9%|▉         | 43/480 [03:37<37:14,  5.11s/it]
  9%|▉         | 44/480 [03:42<36:49,  5.07s/it]
  9%|▉         | 45/480 [03:47<36:39,  5.06s/it]
 10%|▉         | 46/480 [03:52<36:34,  5.06s/it]
 10%|▉         | 47/480 [03:57<36:18,  5.03s/it]
 10%|█         | 48/480 [04:03<37:01,  5.14s/it]
 10%|█         | 49/480 [04:08<36:38,  5.10s/it]
 10%|█         | 50/480 [04:13<36:33,  5.10s/it]
                                                
{'loss': '0.6751', 'grad_norm': '0.6985', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [04:13<36:33,  5.10s/it]
 11%|█         | 51/480 [04:18<36:11,  5.06s/it]
 11%|█         | 52/480 [04:23<36:08,  5.07s/it]
 11%|█         | 53/480 [04:28<36:13,  5.09s/it]
 11%|█▏        | 54/480 [04:33<36:14,  5.10s/it]
 11%|█▏        | 55/480 [04:38<35:56,  5.08s/it]
 12%|█▏        | 56/480 [04:43<36:10,  5.12s/it]
 12%|█▏        | 57/480 [04:49<36:20,  5.15s/it]
 12%|█▏        | 58/480 [04:54<37:03,  5.27s/it]
 12%|█▏        | 59/480 [04:59<36:20,  5.18s/it]
 12%|█▎        | 60/480 [05:04<36:42,  5.24s/it]
                                                
{'loss': '0.6155', 'grad_norm': '0.5937', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [05:04<36:42,  5.24s/it]
 13%|█▎        | 61/480 [05:10<36:31,  5.23s/it]
 13%|█▎        | 62/480 [05:15<36:07,  5.18s/it]
 13%|█▎        | 63/480 [05:20<36:01,  5.18s/it]
 13%|█▎        | 64/480 [05:21<27:46,  4.01s/it]
 14%|█▎        | 65/480 [05:26<30:21,  4.39s/it]
 14%|█▍        | 66/480 [05:32<31:53,  4.62s/it]
 14%|█▍        | 67/480 [05:37<33:10,  4.82s/it]
 14%|█▍        | 68/480 [05:42<34:08,  4.97s/it]
 14%|█▍        | 69/480 [05:47<34:18,  5.01s/it]
 15%|█▍        | 70/480 [05:52<34:08,  5.00s/it]
                                                
{'loss': '0.6237', 'grad_norm': '0.7379', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [05:52<34:08,  5.00s/it]
 15%|█▍        | 71/480 [05:57<34:09,  5.01s/it]
 15%|█▌        | 72/480 [06:02<33:59,  5.00s/it]
 15%|█▌        | 73/480 [06:08<34:19,  5.06s/it]
 15%|█▌        | 74/480 [06:13<34:18,  5.07s/it]
 16%|█▌        | 75/480 [06:18<34:25,  5.10s/it]
 16%|█▌        | 76/480 [06:23<34:25,  5.11s/it]
 16%|█▌        | 77/480 [06:28<35:05,  5.22s/it]
 16%|█▋        | 78/480 [06:34<35:04,  5.24s/it]
 16%|█▋        | 79/480 [06:39<35:36,  5.33s/it]
 17%|█▋        | 80/480 [06:44<35:15,  5.29s/it]
                                                
{'loss': '0.7138', 'grad_norm': '1.042', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [06:44<35:15,  5.29s/it]
 17%|█▋        | 81/480 [06:50<35:17,  5.31s/it]
 17%|█▋        | 82/480 [06:55<34:31,  5.21s/it]
 17%|█▋        | 83/480 [07:00<34:12,  5.17s/it]
 18%|█▊        | 84/480 [07:05<34:10,  5.18s/it]
 18%|█▊        | 85/480 [07:10<33:52,  5.15s/it]
 18%|█▊        | 86/480 [07:15<33:47,  5.15s/it]
 18%|█▊        | 87/480 [07:20<33:37,  5.13s/it]
 18%|█▊        | 88/480 [07:26<33:35,  5.14s/it]
 19%|█▊        | 89/480 [07:31<33:44,  5.18s/it]
 19%|█▉        | 90/480 [07:36<33:17,  5.12s/it]
                                                
{'loss': '0.8628', 'grad_norm': '2.552', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [07:36<33:17,  5.12s/it]
 19%|█▉        | 91/480 [07:41<33:31,  5.17s/it]
 19%|█▉        | 92/480 [07:46<33:23,  5.16s/it]
 19%|█▉        | 93/480 [07:52<33:36,  5.21s/it]
 20%|█▉        | 94/480 [07:57<33:44,  5.24s/it]
 20%|█▉        | 95/480 [08:02<34:21,  5.36s/it]
 20%|██        | 96/480 [08:04<26:24,  4.13s/it]
 20%|██        | 97/480 [08:09<28:20,  4.44s/it]
 20%|██        | 98/480 [08:14<30:06,  4.73s/it]
 21%|██        | 99/480 [08:19<30:33,  4.81s/it]
 21%|██        | 100/480 [08:24<30:56,  4.89s/it]
                                                 
{'loss': '0.6338', 'grad_norm': '0.7179', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [08:24<30:56,  4.89s/it]
 21%|██        | 101/480 [08:29<31:03,  4.92s/it]
 21%|██▏       | 102/480 [08:34<31:22,  4.98s/it]
 21%|██▏       | 103/480 [08:39<31:16,  4.98s/it]
 22%|██▏       | 104/480 [08:45<31:40,  5.05s/it]
 22%|██▏       | 105/480 [08:50<31:27,  5.03s/it]
 22%|██▏       | 106/480 [08:55<31:35,  5.07s/it]
 22%|██▏       | 107/480 [09:00<31:24,  5.05s/it]
 22%|██▎       | 108/480 [09:05<31:29,  5.08s/it]
 23%|██▎       | 109/480 [09:10<31:46,  5.14s/it]
 23%|██▎       | 110/480 [09:16<31:55,  5.18s/it]
                                                 
{'loss': '0.5725', 'grad_norm': '0.6177', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [09:16<31:55,  5.18s/it]
 23%|██▎       | 111/480 [09:20<31:28,  5.12s/it]
 23%|██▎       | 112/480 [09:26<31:43,  5.17s/it]
 24%|██▎       | 113/480 [09:31<31:19,  5.12s/it]
 24%|██▍       | 114/480 [09:36<31:44,  5.20s/it]
 24%|██▍       | 115/480 [09:41<31:15,  5.14s/it]
 24%|██▍       | 116/480 [09:46<31:27,  5.19s/it]
 24%|██▍       | 117/480 [09:51<31:02,  5.13s/it]
 25%|██▍       | 118/480 [09:57<31:16,  5.19s/it]
 25%|██▍       | 119/480 [10:02<31:10,  5.18s/it]
 25%|██▌       | 120/480 [10:07<31:13,  5.21s/it]
                                                 
{'loss': '0.5254', 'grad_norm': '0.2333', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [10:07<31:13,  5.21s/it]
 25%|██▌       | 121/480 [10:12<31:07,  5.20s/it]
 25%|██▌       | 122/480 [10:18<31:31,  5.28s/it]
 26%|██▌       | 123/480 [10:23<31:14,  5.25s/it]
 26%|██▌       | 124/480 [10:28<31:12,  5.26s/it]
 26%|██▌       | 125/480 [10:33<30:39,  5.18s/it]
 26%|██▋       | 126/480 [10:38<30:25,  5.16s/it]
 26%|██▋       | 127/480 [10:43<30:00,  5.10s/it]
 27%|██▋       | 128/480 [10:45<23:09,  3.95s/it]
 27%|██▋       | 129/480 [10:50<25:02,  4.28s/it]
 27%|██▋       | 130/480 [10:55<26:10,  4.49s/it]
                                                 
{'loss': '0.5', 'grad_norm': '2.139', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [10:55<26:10,  4.49s/it]
 27%|██▋       | 131/480 [11:00<27:07,  4.66s/it]
 28%|██▊       | 132/480 [11:05<27:57,  4.82s/it]
 28%|██▊       | 133/480 [11:10<28:45,  4.97s/it]
 28%|██▊       | 134/480 [11:15<28:43,  4.98s/it]
 28%|██▊       | 135/480 [11:20<28:39,  4.98s/it]
 28%|██▊       | 136/480 [11:25<28:52,  5.04s/it]
 29%|██▊       | 137/480 [11:31<28:58,  5.07s/it]
 29%|██▉       | 138/480 [11:36<29:05,  5.10s/it]
 29%|██▉       | 139/480 [11:41<28:52,  5.08s/it]
 29%|██▉       | 140/480 [11:46<29:10,  5.15s/it]
                                                 
{'loss': '0.4795', 'grad_norm': '1.164', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [11:46<29:10,  5.15s/it]
 29%|██▉       | 141/480 [11:51<28:48,  5.10s/it]
 30%|██▉       | 142/480 [11:56<28:53,  5.13s/it]
 30%|██▉       | 143/480 [12:01<28:33,  5.09s/it]
 30%|███       | 144/480 [12:07<28:44,  5.13s/it]
 30%|███       | 145/480 [12:12<28:27,  5.10s/it]
 30%|███       | 146/480 [12:17<28:41,  5.15s/it]
 31%|███       | 147/480 [12:22<29:25,  5.30s/it]
 31%|███       | 148/480 [12:28<29:14,  5.28s/it]
 31%|███       | 149/480 [12:33<28:59,  5.25s/it]
 31%|███▏      | 150/480 [12:38<28:55,  5.26s/it]
                                                 
{'loss': '0.4783', 'grad_norm': '0.9805', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [12:38<28:55,  5.26s/it]
 31%|███▏      | 151/480 [12:43<28:48,  5.25s/it]
 32%|███▏      | 152/480 [12:49<28:41,  5.25s/it]
 32%|███▏      | 153/480 [12:54<28:09,  5.17s/it]
 32%|███▏      | 154/480 [12:59<27:57,  5.15s/it]
 32%|███▏      | 155/480 [13:04<27:39,  5.11s/it]
 32%|███▎      | 156/480 [13:09<27:28,  5.09s/it]
 33%|███▎      | 157/480 [13:14<27:35,  5.12s/it]
 33%|███▎      | 158/480 [13:19<27:24,  5.11s/it]
 33%|███▎      | 159/480 [13:24<27:07,  5.07s/it]
 33%|███▎      | 160/480 [13:25<21:12,  3.98s/it]
                                                 
{'loss': '0.4736', 'grad_norm': '0.8207', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [13:25<21:12,  3.98s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<00:10,  9.45it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:00<00:09, 10.30it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:00<00:10,  9.33it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:00<00:10,  8.66it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:00<00:10,  8.64it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:01<00:10,  8.66it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:01<00:09,  9.29it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:01<00:09,  9.07it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:01<00:09,  9.32it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:01<00:08,  9.32it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:01<00:08,  9.29it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:02<00:08,  9.03it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:02<00:08,  9.61it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:02<00:08,  8.98it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:02<00:07,  9.63it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:02<00:08,  9.01it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:02<00:08,  8.88it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:03<00:07,  9.48it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:03<00:07,  8.90it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:03<00:08,  8.45it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:03<00:07,  9.25it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:03<00:06,  9.78it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:03<00:06, 10.12it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:04<00:06,  9.39it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:04<00:06,  9.85it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:04<00:05, 10.15it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:04<00:05, 10.33it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:04<00:05, 10.48it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:05<00:05,  9.83it/s]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:05<00:05,  9.23it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:05<00:05,  9.69it/s]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:05<00:04, 10.02it/s]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:05<00:04, 10.27it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:06<00:04,  9.03it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:06<00:04,  9.53it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:06<00:04,  9.23it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [00:06<00:03,  9.68it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:06<00:03, 10.01it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [00:07<00:03, 10.13it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:07<00:03,  9.62it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:07<00:02,  9.93it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:07<00:02, 10.17it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:07<00:02,  9.67it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:08<00:02,  9.10it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:08<00:02,  9.62it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:08<00:02,  9.28it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:08<00:02,  8.79it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [00:08<00:01,  9.41it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:08<00:01,  9.83it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [00:09<00:01, 10.15it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [00:09<00:01,  9.63it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [00:09<00:00,  9.27it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:09<00:00,  9.69it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:09<00:00,  9.98it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:10<00:00, 10.23it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [00:10<00:00,  9.73it/s]
+
                                                                   [EVAL] ★ New best eval_rougeL=61.0000 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '61', 'epoch': '5'}
+
 33%|███▎      | 160/480 [13:36<21:12,  3.98s/it]
 34%|███▎      | 161/480 [13:41<39:41,  7.47s/it]
 34%|███▍      | 162/480 [13:46<35:53,  6.77s/it]
 34%|███▍      | 163/480 [13:52<33:49,  6.40s/it]
 34%|███▍      | 164/480 [13:57<31:44,  6.03s/it]
 34%|███▍      | 165/480 [14:02<30:06,  5.74s/it]
 35%|███▍      | 166/480 [14:07<28:49,  5.51s/it]
 35%|███▍      | 167/480 [14:12<27:59,  5.37s/it]
 35%|███▌      | 168/480 [14:17<27:20,  5.26s/it]
 35%|███▌      | 169/480 [14:22<26:56,  5.20s/it]
 35%|███▌      | 170/480 [14:27<26:31,  5.13s/it]
                                                 
{'loss': '0.4664', 'grad_norm': '0.5807', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [14:27<26:31,  5.13s/it]
 36%|███▌      | 171/480 [14:32<26:34,  5.16s/it]
 36%|███▌      | 172/480 [14:37<26:28,  5.16s/it]
 36%|███▌      | 173/480 [14:43<26:28,  5.17s/it]
 36%|███▋      | 174/480 [14:48<26:19,  5.16s/it]
 36%|███▋      | 175/480 [14:53<26:07,  5.14s/it]
 37%|███▋      | 176/480 [14:58<25:47,  5.09s/it]
 37%|███▋      | 177/480 [15:03<25:54,  5.13s/it]
 37%|███▋      | 178/480 [15:08<25:45,  5.12s/it]
 37%|███▋      | 179/480 [15:13<25:48,  5.14s/it]
 38%|███▊      | 180/480 [15:19<25:47,  5.16s/it]
                                                 
{'loss': '0.4798', 'grad_norm': '0.6801', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [15:19<25:47,  5.16s/it]
 38%|███▊      | 181/480 [15:24<25:33,  5.13s/it]
 38%|███▊      | 182/480 [15:29<25:30,  5.14s/it]
 38%|███▊      | 183/480 [15:34<25:28,  5.15s/it]
 38%|███▊      | 184/480 [15:39<25:31,  5.17s/it]
 39%|███▊      | 185/480 [15:44<25:28,  5.18s/it]
 39%|███▉      | 186/480 [15:49<25:19,  5.17s/it]
 39%|███▉      | 187/480 [15:55<25:12,  5.16s/it]
 39%|███▉      | 188/480 [16:00<25:14,  5.19s/it]
 39%|███▉      | 189/480 [16:05<25:06,  5.18s/it]
 40%|███▉      | 190/480 [16:10<25:09,  5.21s/it]
                                                 
{'loss': '0.436', 'grad_norm': '0.8203', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [16:10<25:09,  5.21s/it]
 40%|███▉      | 191/480 [16:15<25:01,  5.19s/it]
 40%|████      | 192/480 [16:17<19:52,  4.14s/it]
 40%|████      | 193/480 [16:22<21:04,  4.41s/it]
 40%|████      | 194/480 [16:27<21:57,  4.61s/it]
 41%|████      | 195/480 [16:32<22:28,  4.73s/it]
 41%|████      | 196/480 [16:37<22:52,  4.83s/it]
 41%|████      | 197/480 [16:42<23:04,  4.89s/it]
 41%|████▏     | 198/480 [16:47<23:07,  4.92s/it]
 41%|████▏     | 199/480 [16:52<23:19,  4.98s/it]
 42%|████▏     | 200/480 [16:58<23:45,  5.09s/it]
                                                 
{'loss': '0.3928', 'grad_norm': '0.3833', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [16:58<23:45,  5.09s/it]
 42%|████▏     | 201/480 [17:03<23:41,  5.09s/it]
 42%|████▏     | 202/480 [17:08<23:26,  5.06s/it]
 42%|████▏     | 203/480 [17:13<23:29,  5.09s/it]
 42%|████▎     | 204/480 [17:18<23:17,  5.06s/it]
 43%|████▎     | 205/480 [17:23<23:32,  5.13s/it]
 43%|████▎     | 206/480 [17:28<23:19,  5.11s/it]
 43%|████▎     | 207/480 [17:33<23:10,  5.09s/it]
 43%|████▎     | 208/480 [17:39<23:13,  5.12s/it]
 44%|████▎     | 209/480 [17:44<23:22,  5.18s/it]
 44%|████▍     | 210/480 [17:49<23:15,  5.17s/it]
                                                 
{'loss': '0.4499', 'grad_norm': '0.03264', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [17:49<23:15,  5.17s/it]
 44%|████▍     | 211/480 [17:54<23:18,  5.20s/it]
 44%|████▍     | 212/480 [18:00<23:16,  5.21s/it]
 44%|████▍     | 213/480 [18:05<23:59,  5.39s/it]
 45%|████▍     | 214/480 [18:11<23:47,  5.37s/it]
 45%|████▍     | 215/480 [18:16<23:43,  5.37s/it]
 45%|████▌     | 216/480 [18:21<23:24,  5.32s/it]
 45%|████▌     | 217/480 [18:27<23:16,  5.31s/it]
 45%|████▌     | 218/480 [18:32<23:01,  5.27s/it]
 46%|████▌     | 219/480 [18:37<22:54,  5.27s/it]
 46%|████▌     | 220/480 [18:42<22:52,  5.28s/it]
                                                 
{'loss': '0.3983', 'grad_norm': '0.9543', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [18:42<22:52,  5.28s/it]
 46%|████▌     | 221/480 [18:48<22:39,  5.25s/it]
 46%|████▋     | 222/480 [18:53<22:14,  5.17s/it]
 46%|████▋     | 223/480 [18:58<22:07,  5.16s/it]
 47%|████▋     | 224/480 [18:59<17:02,  3.99s/it]
 47%|████▋     | 225/480 [19:04<18:14,  4.29s/it]
 47%|████▋     | 226/480 [19:09<19:09,  4.53s/it]
 47%|████▋     | 227/480 [19:14<19:40,  4.66s/it]
 48%|████▊     | 228/480 [19:19<20:16,  4.83s/it]
 48%|████▊     | 229/480 [19:24<20:24,  4.88s/it]
 48%|████▊     | 230/480 [19:29<20:42,  4.97s/it]
                                                 
{'loss': '0.416', 'grad_norm': '1.391', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [19:29<20:42,  4.97s/it]
 48%|████▊     | 231/480 [19:34<20:41,  4.99s/it]
 48%|████▊     | 232/480 [19:40<20:49,  5.04s/it]
 49%|████▊     | 233/480 [19:45<21:06,  5.13s/it]
 49%|████▉     | 234/480 [19:50<21:17,  5.19s/it]
 49%|████▉     | 235/480 [19:56<21:23,  5.24s/it]
 49%|████▉     | 236/480 [20:01<21:36,  5.31s/it]
 49%|████▉     | 237/480 [20:06<21:26,  5.29s/it]
 50%|████▉     | 238/480 [20:12<21:15,  5.27s/it]
 50%|████▉     | 239/480 [20:17<21:01,  5.23s/it]
 50%|█████     | 240/480 [20:22<21:08,  5.28s/it]
                                                 
{'loss': '0.4075', 'grad_norm': '0.3273', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [20:22<21:08,  5.28s/it]
 50%|█████     | 241/480 [20:27<20:55,  5.25s/it]
 50%|█████     | 242/480 [20:32<20:45,  5.23s/it]
 51%|█████     | 243/480 [20:38<20:36,  5.22s/it]
 51%|█████     | 244/480 [20:43<20:36,  5.24s/it]
 51%|█████     | 245/480 [20:48<20:27,  5.22s/it]
 51%|█████▏    | 246/480 [20:53<20:28,  5.25s/it]
 51%|█████▏    | 247/480 [20:59<20:20,  5.24s/it]
 52%|█████▏    | 248/480 [21:04<20:21,  5.26s/it]
 52%|█████▏    | 249/480 [21:09<19:56,  5.18s/it]
 52%|█████▏    | 250/480 [21:14<19:41,  5.14s/it]
                                                 
{'loss': '0.4052', 'grad_norm': '0.695', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [21:14<19:41,  5.14s/it]
 52%|█████▏    | 251/480 [21:19<19:25,  5.09s/it]
 52%|█████▎    | 252/480 [21:24<19:34,  5.15s/it]
 53%|█████▎    | 253/480 [21:29<19:21,  5.11s/it]
 53%|█████▎    | 254/480 [21:34<19:13,  5.11s/it]
 53%|█████▎    | 255/480 [21:40<19:14,  5.13s/it]
 53%|█████▎    | 256/480 [21:41<14:48,  3.97s/it]
 54%|█████▎    | 257/480 [21:46<16:06,  4.34s/it]
 54%|█████▍    | 258/480 [21:51<16:46,  4.53s/it]
 54%|█████▍    | 259/480 [21:56<17:25,  4.73s/it]
 54%|█████▍    | 260/480 [22:01<17:38,  4.81s/it]
                                                 
{'loss': '0.4238', 'grad_norm': '1.364', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [22:01<17:38,  4.81s/it]
 54%|█████▍    | 261/480 [22:07<18:11,  4.98s/it]
 55%|█████▍    | 262/480 [22:12<18:06,  4.98s/it]
 55%|█████▍    | 263/480 [22:17<18:10,  5.03s/it]
 55%|█████▌    | 264/480 [22:22<18:15,  5.07s/it]
 55%|█████▌    | 265/480 [22:28<18:50,  5.26s/it]
 55%|█████▌    | 266/480 [22:33<18:41,  5.24s/it]
 56%|█████▌    | 267/480 [22:38<18:37,  5.25s/it]
 56%|█████▌    | 268/480 [22:43<18:27,  5.22s/it]
 56%|█████▌    | 269/480 [22:49<18:38,  5.30s/it]
 56%|█████▋    | 270/480 [22:54<18:13,  5.21s/it]
                                                 
{'loss': '0.4064', 'grad_norm': '0.7525', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [22:54<18:13,  5.21s/it]
 56%|█████▋    | 271/480 [22:59<18:16,  5.24s/it]
 57%|█████▋    | 272/480 [23:04<18:05,  5.22s/it]
 57%|█████▋    | 273/480 [23:09<18:00,  5.22s/it]
 57%|█████▋    | 274/480 [23:15<17:51,  5.20s/it]
 57%|█████▋    | 275/480 [23:20<17:47,  5.21s/it]
 57%|█████▊    | 276/480 [23:25<17:40,  5.20s/it]
 58%|█████▊    | 277/480 [23:30<17:32,  5.19s/it]
 58%|█████▊    | 278/480 [23:35<17:18,  5.14s/it]
 58%|█████▊    | 279/480 [23:40<17:14,  5.15s/it]
 58%|█████▊    | 280/480 [23:45<17:00,  5.10s/it]
                                                 
{'loss': '0.4977', 'grad_norm': '0.1777', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [23:45<17:00,  5.10s/it]
 59%|█████▊    | 281/480 [23:50<16:53,  5.09s/it]
 59%|█████▉    | 282/480 [23:55<16:41,  5.06s/it]
 59%|█████▉    | 283/480 [24:00<16:35,  5.06s/it]
 59%|█████▉    | 284/480 [24:06<16:38,  5.10s/it]
 59%|█████▉    | 285/480 [24:11<16:36,  5.11s/it]
 60%|█████▉    | 286/480 [24:16<16:27,  5.09s/it]
 60%|█████▉    | 287/480 [24:21<16:16,  5.06s/it]
 60%|██████    | 288/480 [24:22<12:37,  3.94s/it]
 60%|██████    | 289/480 [24:27<13:44,  4.32s/it]
 60%|██████    | 290/480 [24:33<14:33,  4.60s/it]
                                                 
{'loss': '0.4312', 'grad_norm': '0.4609', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [24:33<14:33,  4.60s/it]
 61%|██████    | 291/480 [24:38<15:05,  4.79s/it]
 61%|██████    | 292/480 [24:43<15:32,  4.96s/it]
 61%|██████    | 293/480 [24:48<15:31,  4.98s/it]
 61%|██████▏   | 294/480 [24:53<15:40,  5.06s/it]
 61%|██████▏   | 295/480 [24:58<15:31,  5.04s/it]
 62%|██████▏   | 296/480 [25:04<15:35,  5.08s/it]
 62%|██████▏   | 297/480 [25:09<15:40,  5.14s/it]
 62%|██████▏   | 298/480 [25:14<15:27,  5.10s/it]
 62%|██████▏   | 299/480 [25:19<15:30,  5.14s/it]
 62%|██████▎   | 300/480 [25:24<15:26,  5.15s/it]
                                                 
{'loss': '0.3288', 'grad_norm': '0.4164', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [25:24<15:26,  5.15s/it]
 63%|██████▎   | 301/480 [25:30<15:32,  5.21s/it]
 63%|██████▎   | 302/480 [25:35<15:31,  5.23s/it]
 63%|██████▎   | 303/480 [25:40<15:37,  5.29s/it]
 63%|██████▎   | 304/480 [25:46<15:27,  5.27s/it]
 64%|██████▎   | 305/480 [25:51<15:12,  5.22s/it]
 64%|██████▍   | 306/480 [25:56<14:55,  5.15s/it]
 64%|██████▍   | 307/480 [26:01<15:04,  5.23s/it]
 64%|██████▍   | 308/480 [26:06<15:04,  5.26s/it]
 64%|██████▍   | 309/480 [26:12<15:07,  5.31s/it]
 65%|██████▍   | 310/480 [26:17<15:17,  5.40s/it]
                                                 
{'loss': '0.3767', 'grad_norm': '1.074', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [26:17<15:17,  5.40s/it]
 65%|██████▍   | 311/480 [26:22<14:55,  5.30s/it]
 65%|██████▌   | 312/480 [26:28<14:45,  5.27s/it]
 65%|██████▌   | 313/480 [26:33<14:51,  5.34s/it]
 65%|██████▌   | 314/480 [26:39<15:08,  5.48s/it]
 66%|██████▌   | 315/480 [26:44<14:40,  5.34s/it]
 66%|██████▌   | 316/480 [26:49<14:35,  5.34s/it]
 66%|██████▌   | 317/480 [26:54<14:13,  5.23s/it]
 66%|██████▋   | 318/480 [26:59<13:57,  5.17s/it]
 66%|██████▋   | 319/480 [27:05<14:04,  5.25s/it]
 67%|██████▋   | 320/480 [27:06<10:48,  4.05s/it]
                                                 
{'loss': '0.361', 'grad_norm': '1.402', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [27:06<10:48,  4.05s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:09, 10.81it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:00<00:08, 10.73it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:00<00:09,  9.69it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:00<00:09,  9.25it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:01<00:09,  9.00it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:01<00:09,  9.54it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:01<00:09,  9.23it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:01<00:08,  9.69it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:01<00:08,  9.99it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:02<00:07, 10.22it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:02<00:08,  9.66it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:02<00:07,  9.98it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:02<00:07,  9.55it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:02<00:08,  9.03it/s]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:03<00:07,  9.56it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:03<00:07,  8.97it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:03<00:07,  9.55it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:03<00:06,  9.94it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:03<00:06, 10.21it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:03<00:06,  9.51it/s]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:04<00:06,  9.58it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:04<00:06,  9.46it/s]
+
Eval @ step 320:  41%|████      | 41/100 [00:04<00:06,  9.29it/s]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:04<00:05,  9.75it/s]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:04<00:05, 10.10it/s]
+
Eval @ step 320:  47%|████▋     | 47/100 [00:04<00:05,  9.57it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:05<00:05,  9.27it/s]
+
Eval @ step 320:  51%|█████     | 51/100 [00:05<00:05,  9.69it/s]
+
Eval @ step 320:  53%|█████▎    | 53/100 [00:05<00:04, 10.02it/s]
+
Eval @ step 320:  55%|█████▌    | 55/100 [00:05<00:04, 10.26it/s]
+
Eval @ step 320:  57%|█████▋    | 57/100 [00:05<00:04,  9.05it/s]
+
Eval @ step 320:  59%|█████▉    | 59/100 [00:06<00:04,  9.54it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:06<00:04,  9.22it/s]
+
Eval @ step 320:  63%|██████▎   | 63/100 [00:06<00:03,  9.65it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [00:06<00:03,  9.98it/s]
+
Eval @ step 320:  67%|██████▋   | 67/100 [00:06<00:03, 10.22it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [00:07<00:03,  9.71it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [00:07<00:02,  9.99it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [00:07<00:02, 10.22it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [00:07<00:02,  9.70it/s]
+
Eval @ step 320:  77%|███████▋  | 77/100 [00:07<00:02, 10.00it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [00:08<00:02, 10.24it/s]
+
Eval @ step 320:  81%|████████  | 81/100 [00:08<00:02,  9.08it/s]
+
Eval @ step 320:  83%|████████▎ | 83/100 [00:08<00:01,  9.54it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [00:08<00:01,  9.89it/s]
+
Eval @ step 320:  87%|████████▋ | 87/100 [00:08<00:01,  9.99it/s]
+
Eval @ step 320:  89%|████████▉ | 89/100 [00:09<00:01,  9.45it/s]
+
Eval @ step 320:  91%|█████████ | 91/100 [00:09<00:01,  9.00it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:09<00:00,  8.64it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:09<00:00,  9.24it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:09<00:00,  9.70it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:10<00:00, 10.03it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:10<00:00,  9.56it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=73.0000 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '73', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [27:16<10:48,  4.05s/it]
 67%|██████▋   | 321/480 [27:22<19:59,  7.54s/it]
 67%|██████▋   | 322/480 [27:27<17:58,  6.83s/it]
 67%|██████▋   | 323/480 [27:32<16:29,  6.30s/it]
 68%|██████▊   | 324/480 [27:37<15:22,  5.91s/it]
 68%|██████▊   | 325/480 [27:42<14:36,  5.66s/it]
 68%|██████▊   | 326/480 [27:47<14:01,  5.47s/it]
 68%|██████▊   | 327/480 [27:52<13:34,  5.32s/it]
 68%|██████▊   | 328/480 [27:57<13:18,  5.25s/it]
 69%|██████▊   | 329/480 [28:02<13:01,  5.17s/it]
 69%|██████▉   | 330/480 [28:07<13:01,  5.21s/it]
                                                 
{'loss': '0.3855', 'grad_norm': '1.178', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [28:07<13:01,  5.21s/it]
 69%|██████▉   | 331/480 [28:12<12:46,  5.14s/it]
 69%|██████▉   | 332/480 [28:17<12:39,  5.13s/it]
 69%|██████▉   | 333/480 [28:23<12:40,  5.17s/it]
 70%|██████▉   | 334/480 [28:28<12:53,  5.30s/it]
 70%|██████▉   | 335/480 [28:34<12:58,  5.37s/it]
 70%|███████   | 336/480 [28:39<13:00,  5.42s/it]
 70%|███████   | 337/480 [28:45<12:43,  5.34s/it]
 70%|███████   | 338/480 [28:50<12:28,  5.27s/it]
 71%|███████   | 339/480 [28:55<12:19,  5.24s/it]
 71%|███████   | 340/480 [29:00<12:08,  5.21s/it]
                                                 
{'loss': '0.4117', 'grad_norm': '3.213', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [29:00<12:08,  5.21s/it]
 71%|███████   | 341/480 [29:05<12:02,  5.20s/it]
 71%|███████▏  | 342/480 [29:10<11:55,  5.19s/it]
 71%|███████▏  | 343/480 [29:15<11:49,  5.18s/it]
 72%|███████▏  | 344/480 [29:21<11:51,  5.23s/it]
 72%|███████▏  | 345/480 [29:26<11:44,  5.22s/it]
 72%|███████▏  | 346/480 [29:31<11:44,  5.26s/it]
 72%|███████▏  | 347/480 [29:37<11:37,  5.24s/it]
 72%|███████▎  | 348/480 [29:42<11:31,  5.24s/it]
 73%|███████▎  | 349/480 [29:47<11:28,  5.25s/it]
 73%|███████▎  | 350/480 [29:52<11:19,  5.22s/it]
                                                 
{'loss': '0.3546', 'grad_norm': '0.8461', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [29:52<11:19,  5.22s/it]
 73%|███████▎  | 351/480 [29:57<11:04,  5.15s/it]
 73%|███████▎  | 352/480 [29:58<08:30,  3.98s/it]
 74%|███████▎  | 353/480 [30:03<09:05,  4.30s/it]
 74%|███████▍  | 354/480 [30:08<09:27,  4.50s/it]
 74%|███████▍  | 355/480 [30:14<09:43,  4.67s/it]
 74%|███████▍  | 356/480 [30:19<09:54,  4.80s/it]
 74%|███████▍  | 357/480 [30:24<10:04,  4.91s/it]
 75%|███████▍  | 358/480 [30:29<10:07,  4.98s/it]
 75%|███████▍  | 359/480 [30:34<10:10,  5.04s/it]
 75%|███████▌  | 360/480 [30:39<10:04,  5.04s/it]
                                                 
{'loss': '0.3368', 'grad_norm': '1.953', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [30:39<10:04,  5.04s/it]
 75%|███████▌  | 361/480 [30:44<10:10,  5.13s/it]
 75%|███████▌  | 362/480 [30:50<10:06,  5.14s/it]
 76%|███████▌  | 363/480 [30:55<10:11,  5.23s/it]
 76%|███████▌  | 364/480 [31:00<09:58,  5.16s/it]
 76%|███████▌  | 365/480 [31:05<09:59,  5.22s/it]
 76%|███████▋  | 366/480 [31:10<09:46,  5.15s/it]
 76%|███████▋  | 367/480 [31:16<09:59,  5.30s/it]
 77%|███████▋  | 368/480 [31:22<10:12,  5.47s/it]
 77%|███████▋  | 369/480 [31:27<10:05,  5.46s/it]
 77%|███████▋  | 370/480 [31:32<09:45,  5.32s/it]
                                                 
{'loss': '0.3985', 'grad_norm': '6.808', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [31:32<09:45,  5.32s/it]
 77%|███████▋  | 371/480 [31:38<09:45,  5.37s/it]
 78%|███████▊  | 372/480 [31:43<09:40,  5.38s/it]
 78%|███████▊  | 373/480 [31:49<09:32,  5.35s/it]
 78%|███████▊  | 374/480 [31:54<09:17,  5.26s/it]
 78%|███████▊  | 375/480 [31:59<09:16,  5.30s/it]
 78%|███████▊  | 376/480 [32:04<09:08,  5.27s/it]
 79%|███████▊  | 377/480 [32:10<09:03,  5.28s/it]
 79%|███████▉  | 378/480 [32:15<08:58,  5.28s/it]
 79%|███████▉  | 379/480 [32:20<08:56,  5.31s/it]
 79%|███████▉  | 380/480 [32:25<08:50,  5.31s/it]
                                                 
{'loss': '0.3602', 'grad_norm': '0.238', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [32:25<08:50,  5.31s/it]
 79%|███████▉  | 381/480 [32:31<08:37,  5.23s/it]
 80%|███████▉  | 382/480 [32:36<08:26,  5.17s/it]
 80%|███████▉  | 383/480 [32:41<08:23,  5.19s/it]
 80%|████████  | 384/480 [32:42<06:24,  4.01s/it]
 80%|████████  | 385/480 [32:47<06:49,  4.31s/it]
 80%|████████  | 386/480 [32:52<07:10,  4.58s/it]
 81%|████████  | 387/480 [32:57<07:17,  4.70s/it]
 81%|████████  | 388/480 [33:03<07:30,  4.89s/it]
 81%|████████  | 389/480 [33:08<07:42,  5.08s/it]
 81%|████████▏ | 390/480 [33:14<07:47,  5.19s/it]
                                                 
{'loss': '0.3364', 'grad_norm': '0.5713', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [33:14<07:47,  5.19s/it]
 81%|████████▏ | 391/480 [33:19<07:36,  5.13s/it]
 82%|████████▏ | 392/480 [33:24<07:33,  5.15s/it]
 82%|████████▏ | 393/480 [33:29<07:28,  5.15s/it]
 82%|████████▏ | 394/480 [33:34<07:20,  5.12s/it]
 82%|████████▏ | 395/480 [33:39<07:23,  5.21s/it]
 82%|████████▎ | 396/480 [33:45<07:19,  5.23s/it]
 83%|████████▎ | 397/480 [33:50<07:12,  5.21s/it]
 83%|████████▎ | 398/480 [33:55<07:10,  5.24s/it]
 83%|████████▎ | 399/480 [34:00<07:06,  5.26s/it]
 83%|████████▎ | 400/480 [34:06<07:00,  5.26s/it]
                                                 
{'loss': '0.2641', 'grad_norm': '0.1572', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [34:06<07:00,  5.26s/it]
 84%|████████▎ | 401/480 [34:11<06:53,  5.24s/it]
 84%|████████▍ | 402/480 [34:16<06:45,  5.20s/it]
 84%|████████▍ | 403/480 [34:21<06:41,  5.21s/it]
 84%|████████▍ | 404/480 [34:27<06:44,  5.32s/it]
 84%|████████▍ | 405/480 [34:32<06:35,  5.28s/it]
 85%|████████▍ | 406/480 [34:37<06:26,  5.22s/it]
 85%|████████▍ | 407/480 [34:42<06:19,  5.20s/it]
 85%|████████▌ | 408/480 [34:47<06:11,  5.16s/it]
 85%|████████▌ | 409/480 [34:52<06:02,  5.10s/it]
 85%|████████▌ | 410/480 [34:57<05:58,  5.12s/it]
                                                 
{'loss': '0.3306', 'grad_norm': '2.052', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [34:57<05:58,  5.12s/it]
 86%|████████▌ | 411/480 [35:02<05:50,  5.08s/it]
 86%|████████▌ | 412/480 [35:07<05:45,  5.09s/it]
 86%|████████▌ | 413/480 [35:13<05:42,  5.11s/it]
 86%|████████▋ | 414/480 [35:18<05:39,  5.14s/it]
 86%|████████▋ | 415/480 [35:23<05:31,  5.10s/it]
 87%|████████▋ | 416/480 [35:24<04:12,  3.95s/it]
 87%|████████▋ | 417/480 [35:29<04:30,  4.30s/it]
 87%|████████▋ | 418/480 [35:34<04:44,  4.59s/it]
 87%|████████▋ | 419/480 [35:40<04:56,  4.86s/it]
 88%|████████▊ | 420/480 [35:45<04:55,  4.92s/it]
                                                 
{'loss': '0.3099', 'grad_norm': '0.02614', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [35:45<04:55,  4.92s/it]
 88%|████████▊ | 421/480 [35:51<05:03,  5.14s/it]
 88%|████████▊ | 422/480 [35:56<04:55,  5.09s/it]
 88%|████████▊ | 423/480 [36:01<04:50,  5.10s/it]
 88%|████████▊ | 424/480 [36:06<04:46,  5.12s/it]
 89%|████████▊ | 425/480 [36:11<04:44,  5.17s/it]
 89%|████████▉ | 426/480 [36:16<04:36,  5.12s/it]
 89%|████████▉ | 427/480 [36:21<04:33,  5.15s/it]
 89%|████████▉ | 428/480 [36:27<04:29,  5.18s/it]
 89%|████████▉ | 429/480 [36:32<04:22,  5.16s/it]
 90%|████████▉ | 430/480 [36:37<04:18,  5.16s/it]
                                                 
{'loss': '0.3858', 'grad_norm': '0.03839', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [36:37<04:18,  5.16s/it]
 90%|████████▉ | 431/480 [36:43<04:22,  5.36s/it]
 90%|█████████ | 432/480 [36:48<04:16,  5.35s/it]
 90%|█████████ | 433/480 [36:53<04:10,  5.34s/it]
 90%|█████████ | 434/480 [36:58<04:00,  5.23s/it]
 91%|█████████ | 435/480 [37:04<03:53,  5.18s/it]
 91%|█████████ | 436/480 [37:08<03:45,  5.12s/it]
 91%|█████████ | 437/480 [37:14<03:41,  5.15s/it]
 91%|█████████▏| 438/480 [37:19<03:34,  5.10s/it]
 91%|█████████▏| 439/480 [37:24<03:29,  5.12s/it]
 92%|█████████▏| 440/480 [37:29<03:23,  5.09s/it]
                                                 
{'loss': '0.3684', 'grad_norm': '0.5952', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [37:29<03:23,  5.09s/it]
 92%|█████████▏| 441/480 [37:34<03:20,  5.15s/it]
 92%|█████████▏| 442/480 [37:39<03:14,  5.11s/it]
 92%|█████████▏| 443/480 [37:44<03:09,  5.12s/it]
 92%|█████████▎| 444/480 [37:49<03:03,  5.09s/it]
 93%|█████████▎| 445/480 [37:54<02:57,  5.08s/it]
 93%|█████████▎| 446/480 [38:00<02:53,  5.10s/it]
 93%|█████████▎| 447/480 [38:05<02:48,  5.09s/it]
 93%|█████████▎| 448/480 [38:06<02:06,  3.95s/it]
 94%|█████████▎| 449/480 [38:11<02:13,  4.32s/it]
 94%|█████████▍| 450/480 [38:16<02:16,  4.54s/it]
                                                 
{'loss': '0.3009', 'grad_norm': '0.03085', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [38:16<02:16,  4.54s/it]
 94%|█████████▍| 451/480 [38:21<02:15,  4.67s/it]
 94%|█████████▍| 452/480 [38:26<02:15,  4.85s/it]
 94%|█████████▍| 453/480 [38:32<02:13,  4.94s/it]
 95%|█████████▍| 454/480 [38:37<02:10,  5.02s/it]
 95%|█████████▍| 455/480 [38:42<02:07,  5.08s/it]
 95%|█████████▌| 456/480 [38:47<02:03,  5.15s/it]
 95%|█████████▌| 457/480 [38:52<01:57,  5.10s/it]
 95%|█████████▌| 458/480 [38:58<01:54,  5.19s/it]
 96%|█████████▌| 459/480 [39:03<01:48,  5.18s/it]
 96%|█████████▌| 460/480 [39:08<01:44,  5.24s/it]
                                                 
{'loss': '0.3941', 'grad_norm': '1.397', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [39:08<01:44,  5.24s/it]
 96%|█████████▌| 461/480 [39:14<01:40,  5.27s/it]
 96%|█████████▋| 462/480 [39:19<01:35,  5.29s/it]
 96%|█████████▋| 463/480 [39:24<01:28,  5.20s/it]
 97%|█████████▋| 464/480 [39:29<01:22,  5.16s/it]
 97%|█████████▋| 465/480 [39:34<01:16,  5.11s/it]
 97%|█████████▋| 466/480 [39:39<01:11,  5.08s/it]
 97%|█████████▋| 467/480 [39:44<01:06,  5.11s/it]
 98%|█████████▊| 468/480 [39:49<01:00,  5.07s/it]
 98%|█████████▊| 469/480 [39:54<00:56,  5.13s/it]
 98%|█████████▊| 470/480 [39:59<00:50,  5.09s/it]
                                                 
{'loss': '0.3294', 'grad_norm': '0.03499', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [39:59<00:50,  5.09s/it]
 98%|█████████▊| 471/480 [40:05<00:46,  5.11s/it]
 98%|█████████▊| 472/480 [40:10<00:40,  5.07s/it]
 99%|█████████▊| 473/480 [40:15<00:35,  5.08s/it]
 99%|█████████▉| 474/480 [40:20<00:30,  5.10s/it]
 99%|█████████▉| 475/480 [40:25<00:25,  5.12s/it]
 99%|█████████▉| 476/480 [40:30<00:20,  5.08s/it]
 99%|█████████▉| 477/480 [40:35<00:15,  5.13s/it]
100%|█████████▉| 478/480 [40:40<00:10,  5.09s/it]
100%|█████████▉| 479/480 [40:46<00:05,  5.17s/it]
100%|██████████| 480/480 [40:47<00:00,  4.00s/it]
                                                 
{'loss': '0.2522', 'grad_norm': '0.7646', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [40:47<00:00,  4.00s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:09, 10.62it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:00<00:09, 10.66it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:00<00:09,  9.57it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:00<00:10,  9.11it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:01<00:10,  8.95it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:01<00:09,  9.51it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:01<00:08,  9.91it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:01<00:08, 10.18it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:01<00:07, 10.35it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:02<00:07, 10.48it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:02<00:07, 10.57it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:02<00:07, 10.64it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:02<00:07,  9.95it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:02<00:07,  9.51it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:03<00:07,  9.24it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:03<00:07,  9.67it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:03<00:06,  9.99it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:03<00:06, 10.20it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:03<00:06,  9.66it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:04<00:06,  9.99it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:04<00:05, 10.23it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:04<00:05, 10.38it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:04<00:05, 10.50it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:04<00:05,  9.84it/s]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:04<00:05,  9.22it/s]
+
Eval @ step 480:  51%|█████     | 51/100 [00:05<00:05,  9.69it/s]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:05<00:04, 10.03it/s]
+
Eval @ step 480:  55%|█████▌    | 55/100 [00:05<00:04, 10.26it/s]
+
Eval @ step 480:  57%|█████▋    | 57/100 [00:05<00:04,  9.65it/s]
+
Eval @ step 480:  59%|█████▉    | 59/100 [00:05<00:04,  9.96it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [00:06<00:03, 10.22it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [00:06<00:03, 10.39it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [00:06<00:03,  9.79it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:06<00:03,  9.65it/s]
+
Eval @ step 480:  67%|██████▋   | 67/100 [00:06<00:03,  9.58it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:06<00:03,  9.31it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [00:07<00:03,  8.68it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [00:07<00:03,  9.41it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [00:07<00:02,  9.83it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [00:07<00:02, 10.09it/s]
+
Eval @ step 480:  77%|███████▋  | 77/100 [00:07<00:02, 10.29it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [00:07<00:02, 10.39it/s]
+
Eval @ step 480:  81%|████████  | 81/100 [00:08<00:02,  9.12it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [00:08<00:01,  9.56it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [00:08<00:01,  9.87it/s]
+
Eval @ step 480:  87%|████████▋ | 87/100 [00:08<00:01, 10.12it/s]
+
Eval @ step 480:  89%|████████▉ | 89/100 [00:09<00:01,  9.57it/s]
+
Eval @ step 480:  91%|█████████ | 91/100 [00:09<00:00,  9.25it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:09<00:00,  8.81it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:09<00:00,  9.37it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:09<00:00,  9.77it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:09<00:00, 10.02it/s]
+
Eval @ step 480:  99%|█████████▉| 99/100 [00:10<00:00,  9.33it/s]
+
                                                                   [EVAL] eval_rougeL=72.0000 (best=73.0000)
+
                                                 
{'eval_rougeL': '72', 'epoch': '15'}
+
100%|██████████| 480/480 [40:57<00:00,  4.00s/it]
                                                 
{'train_runtime': '2457', 'train_samples_per_second': '6.104', 'train_steps_per_second': '0.195', 'train_loss': '0.5572', 'epoch': '15'}
+
100%|██████████| 480/480 [40:57<00:00,  4.00s/it]
100%|██████████| 480/480 [40:57<00:00,  5.12s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 50675548GF
+  train_loss               =     0.5572
+  train_runtime            = 0:40:57.49
+  train_samples            =       1000
+  train_samples_per_second =      6.104
+  train_steps_per_second   =      0.195
+  [EVAL] Restored best model (eval_rougeL=73.0000, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   5%|▌         | 15/291 [00:00<00:01, 147.27it/s]
Loading weights:  13%|█▎        | 37/291 [00:00<00:01, 175.96it/s]
Loading weights:  20%|█▉        | 58/291 [00:00<00:01, 172.72it/s]
Loading weights:  26%|██▌       | 76/291 [00:00<00:01, 174.75it/s]
Loading weights:  33%|███▎      | 95/291 [00:00<00:01, 176.05it/s]
Loading weights:  41%|████      | 119/291 [00:00<00:00, 192.84it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 177.66it/s]
Loading weights:  54%|█████▍    | 157/291 [00:00<00:00, 177.28it/s]
Loading weights:  61%|██████    | 177/291 [00:00<00:00, 181.09it/s]
Loading weights:  69%|██████▊   | 200/291 [00:01<00:00, 188.77it/s]
Loading weights:  76%|███████▌  | 220/291 [00:01<00:00, 183.66it/s]
Loading weights:  82%|████████▏ | 239/291 [00:01<00:00, 177.80it/s]
Loading weights:  89%|████████▉ | 259/291 [00:01<00:00, 182.12it/s]
Loading weights:  97%|█████████▋| 283/291 [00:01<00:00, 187.58it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 184.78it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task875_emotion_classification': PaR=50.1, shrinkage=ridge, n=1000
+  [SRT] Wired router: 15 sigs, cur=task875_emotion_classification, mapping={'task875_emotion_classification': 0, 'task1572_samsum_summary': 14, 'task363_sst2_polarity_classification': 13, 'task1290_xsum_summarization': 12, 'task181_outcome_extraction': 11, 'task002_quoref_answer_generation': 10, 'task1510_evalution_relation_extraction': 9, 'task639_multi_woz_user_utterance_generation': 8, 'task1729_personachat_generate_next': 7, 'task073_commonsenseqa_answer_generation': 6, 'task1590_diplomacy_text_generation': 5, 'task748_glucose_reverse_cause_event_detection': 4, 'task511_reddit_tifu_long_text_summarization': 3, 'task591_sciq_answer_generation': 2, 'task1687_sentiment140_classification': 1}
+[SRT] Router: 15 tasks, avg_PaR=50.1, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/15-task875_emotion_classification/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1590_diplomacy_text_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task748_glucose_reverse_cause_event_detection/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task511_reddit_tifu_long_text_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task591_sciq_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1687_sentiment140_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task875_emotion_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+  [SRT-INIT] task_id_to_idx={'task875_emotion_classification': 0, 'task1572_samsum_summary': 14, 'task363_sst2_polarity_classification': 13, 'task1290_xsum_summarization': 12, 'task181_outcome_extraction': 11, 'task002_quoref_answer_generation': 10, 'task1510_evalution_relation_extraction': 9, 'task639_multi_woz_user_utterance_generation': 8, 'task1729_personachat_generate_next': 7, 'task073_commonsenseqa_answer_generation': 6, 'task1590_diplomacy_text_generation': 5, 'task748_glucose_reverse_cause_event_detection': 4, 'task511_reddit_tifu_long_text_summarization': 3, 'task591_sciq_answer_generation': 2, 'task1687_sentiment140_classification': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/1197 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=15
+
Continual predict:   0%|          | 1/1197 [00:02<40:07,  2.01s/it]
Continual predict:   0%|          | 2/1197 [00:04<50:05,  2.51s/it]
Continual predict:   0%|          | 3/1197 [00:05<36:23,  1.83s/it]
Continual predict:   0%|          | 4/1197 [00:06<27:18,  1.37s/it]
Continual predict:   0%|          | 5/1197 [00:08<32:51,  1.65s/it]
Continual predict:   1%|          | 6/1197 [00:13<56:30,  2.85s/it]
Continual predict:   1%|          | 7/1197 [00:17<1:01:44,  3.11s/it]
Continual predict:   1%|          | 8/1197 [00:19<56:56,  2.87s/it]  
Continual predict:   1%|          | 9/1197 [00:21<51:29,  2.60s/it]
Continual predict:   1%|          | 10/1197 [00:22<41:08,  2.08s/it]
Continual predict:   1%|          | 11/1197 [00:25<46:23,  2.35s/it]
Continual predict:   1%|          | 12/1197 [00:29<53:33,  2.71s/it]
Continual predict:   1%|          | 13/1197 [00:30<45:21,  2.30s/it]
Continual predict:   1%|          | 14/1197 [00:33<47:51,  2.43s/it]
Continual predict:   1%|▏         | 15/1197 [00:35<48:01,  2.44s/it]
Continual predict:   1%|▏         | 16/1197 [00:38<50:16,  2.55s/it]
Continual predict:   1%|▏         | 17/1197 [00:47<1:24:27,  4.29s/it]
Continual predict:   2%|▏         | 18/1197 [00:49<1:12:45,  3.70s/it]
Continual predict:   2%|▏         | 19/1197 [00:52<1:07:32,  3.44s/it]
Continual predict:   2%|▏         | 20/1197 [00:53<56:25,  2.88s/it]  
Continual predict:   2%|▏         | 21/1197 [00:53<40:39,  2.07s/it]
Continual predict:   2%|▏         | 22/1197 [00:54<29:34,  1.51s/it]
Continual predict:   2%|▏         | 23/1197 [00:54<21:52,  1.12s/it]
Continual predict:   2%|▏         | 24/1197 [00:54<16:27,  1.19it/s]
Continual predict:   2%|▏         | 25/1197 [00:54<12:39,  1.54it/s]
Continual predict:   2%|▏         | 26/1197 [00:54<09:59,  1.95it/s]
Continual predict:   2%|▏         | 27/1197 [00:55<08:08,  2.39it/s]
Continual predict:   2%|▏         | 28/1197 [00:55<06:49,  2.85it/s]
Continual predict:   2%|▏         | 29/1197 [00:55<05:54,  3.29it/s]
Continual predict:   3%|▎         | 30/1197 [00:55<05:19,  3.65it/s]
Continual predict:   3%|▎         | 31/1197 [00:55<04:52,  3.99it/s]
Continual predict:   3%|▎         | 32/1197 [00:56<04:32,  4.28it/s]
Continual predict:   3%|▎         | 33/1197 [00:56<04:21,  4.45it/s]
Continual predict:   3%|▎         | 34/1197 [00:56<04:13,  4.58it/s]
Continual predict:   3%|▎         | 35/1197 [00:56<04:08,  4.68it/s]
Continual predict:   3%|▎         | 36/1197 [00:56<04:01,  4.81it/s]
Continual predict:   3%|▎         | 37/1197 [00:57<03:59,  4.85it/s]
Continual predict:   3%|▎         | 38/1197 [00:57<03:55,  4.92it/s]
Continual predict:   3%|▎         | 39/1197 [00:57<03:52,  4.98it/s]
Continual predict:   3%|▎         | 40/1197 [00:57<03:50,  5.01it/s]
Continual predict:   3%|▎         | 41/1197 [00:57<03:51,  4.99it/s]
Continual predict:   4%|▎         | 42/1197 [00:58<03:49,  5.03it/s]
Continual predict:   4%|▎         | 43/1197 [00:58<03:50,  5.00it/s]
Continual predict:   4%|▎         | 44/1197 [00:58<03:51,  4.99it/s]
Continual predict:   4%|▍         | 45/1197 [00:58<03:49,  5.01it/s]
Continual predict:   4%|▍         | 46/1197 [00:58<03:48,  5.04it/s]
Continual predict:   4%|▍         | 47/1197 [00:59<03:47,  5.06it/s]
Continual predict:   4%|▍         | 48/1197 [00:59<03:49,  5.02it/s]
Continual predict:   4%|▍         | 49/1197 [00:59<03:49,  5.00it/s]
Continual predict:   4%|▍         | 50/1197 [00:59<03:48,  5.03it/s]
Continual predict:   4%|▍         | 51/1197 [00:59<03:47,  5.04it/s]
Continual predict:   4%|▍         | 52/1197 [01:00<03:48,  5.02it/s]
Continual predict:   4%|▍         | 53/1197 [01:00<03:53,  4.91it/s]
Continual predict:   5%|▍         | 54/1197 [01:00<04:07,  4.63it/s]
Continual predict:   5%|▍         | 55/1197 [01:00<03:59,  4.77it/s]
Continual predict:   5%|▍         | 56/1197 [01:00<03:57,  4.81it/s]
Continual predict:   5%|▍         | 57/1197 [01:01<03:52,  4.90it/s]
Continual predict:   5%|▍         | 58/1197 [01:01<03:49,  4.96it/s]
Continual predict:   5%|▍         | 59/1197 [01:01<03:47,  5.00it/s]
Continual predict:   5%|▌         | 60/1197 [01:01<03:48,  4.98it/s]
Continual predict:   5%|▌         | 61/1197 [01:01<03:46,  5.02it/s]
Continual predict:   5%|▌         | 62/1197 [01:02<03:44,  5.06it/s]
Continual predict:   5%|▌         | 63/1197 [01:02<03:45,  5.02it/s]
Continual predict:   5%|▌         | 64/1197 [01:02<03:44,  5.04it/s]
Continual predict:   5%|▌         | 65/1197 [01:02<03:43,  5.07it/s]
Continual predict:   6%|▌         | 66/1197 [01:02<03:43,  5.06it/s]
Continual predict:   6%|▌         | 67/1197 [01:03<03:42,  5.09it/s]
Continual predict:   6%|▌         | 68/1197 [01:03<03:41,  5.10it/s]
Continual predict:   6%|▌         | 69/1197 [01:03<03:40,  5.11it/s]
Continual predict:   6%|▌         | 70/1197 [01:03<03:39,  5.13it/s]
Continual predict:   6%|▌         | 71/1197 [01:03<03:39,  5.13it/s]
Continual predict:   6%|▌         | 72/1197 [01:04<03:42,  5.07it/s]
Continual predict:   6%|▌         | 73/1197 [01:04<03:40,  5.09it/s]
Continual predict:   6%|▌         | 74/1197 [01:04<03:42,  5.04it/s]
Continual predict:   6%|▋         | 75/1197 [01:04<03:41,  5.05it/s]
Continual predict:   6%|▋         | 76/1197 [01:04<03:40,  5.08it/s]
Continual predict:   6%|▋         | 77/1197 [01:05<03:40,  5.07it/s]
Continual predict:   7%|▋         | 78/1197 [01:05<03:39,  5.09it/s]
Continual predict:   7%|▋         | 79/1197 [01:05<03:39,  5.10it/s]
Continual predict:   7%|▋         | 80/1197 [01:05<03:38,  5.12it/s]
Continual predict:   7%|▋         | 81/1197 [01:05<03:38,  5.11it/s]
Continual predict:   7%|▋         | 82/1197 [01:06<03:38,  5.11it/s]
Continual predict:   7%|▋         | 83/1197 [01:06<03:38,  5.11it/s]
Continual predict:   7%|▋         | 84/1197 [01:06<03:40,  5.05it/s]
Continual predict:   7%|▋         | 85/1197 [01:06<03:39,  5.07it/s]
Continual predict:   7%|▋         | 86/1197 [01:06<03:38,  5.08it/s]
Continual predict:   7%|▋         | 87/1197 [01:07<03:38,  5.07it/s]
Continual predict:   7%|▋         | 88/1197 [01:07<03:37,  5.09it/s]
Continual predict:   7%|▋         | 89/1197 [01:07<03:37,  5.10it/s]
Continual predict:   8%|▊         | 90/1197 [01:07<03:37,  5.10it/s]
Continual predict:   8%|▊         | 91/1197 [01:07<03:36,  5.10it/s]
Continual predict:   8%|▊         | 92/1197 [01:08<03:36,  5.09it/s]
Continual predict:   8%|▊         | 93/1197 [01:08<03:39,  5.04it/s]
Continual predict:   8%|▊         | 94/1197 [01:08<03:40,  5.01it/s]
Continual predict:   8%|▊         | 95/1197 [01:08<03:38,  5.03it/s]
Continual predict:   8%|▊         | 96/1197 [01:08<03:38,  5.05it/s]
Continual predict:   8%|▊         | 97/1197 [01:09<03:37,  5.07it/s]
Continual predict:   8%|▊         | 98/1197 [01:09<03:38,  5.02it/s]
Continual predict:   8%|▊         | 99/1197 [01:09<03:39,  4.99it/s]
Continual predict:   8%|▊         | 100/1197 [01:09<03:40,  4.98it/s]
Continual predict:   8%|▊         | 101/1197 [01:09<03:38,  5.02it/s]
Continual predict:   9%|▊         | 102/1197 [01:10<03:36,  5.05it/s]
Continual predict:   9%|▊         | 103/1197 [01:10<03:36,  5.06it/s]
Continual predict:   9%|▊         | 104/1197 [01:10<03:35,  5.07it/s]
Continual predict:   9%|▉         | 105/1197 [01:10<03:38,  5.00it/s]
Continual predict:   9%|▉         | 106/1197 [01:10<03:42,  4.91it/s]
Continual predict:   9%|▉         | 107/1197 [01:11<03:55,  4.62it/s]
Continual predict:   9%|▉         | 108/1197 [01:11<03:49,  4.75it/s]
Continual predict:   9%|▉         | 109/1197 [01:11<03:44,  4.85it/s]
Continual predict:   9%|▉         | 110/1197 [01:11<03:40,  4.92it/s]
Continual predict:   9%|▉         | 111/1197 [01:11<03:38,  4.98it/s]
Continual predict:   9%|▉         | 112/1197 [01:12<03:36,  5.02it/s]
Continual predict:   9%|▉         | 113/1197 [01:12<03:34,  5.05it/s]
Continual predict:  10%|▉         | 114/1197 [01:12<03:34,  5.06it/s]
Continual predict:  10%|▉         | 115/1197 [01:12<03:35,  5.02it/s]
Continual predict:  10%|▉         | 116/1197 [01:12<03:34,  5.05it/s]
Continual predict:  10%|▉         | 117/1197 [01:13<03:35,  5.01it/s]
Continual predict:  10%|▉         | 118/1197 [01:13<03:36,  4.98it/s]
Continual predict:  10%|▉         | 119/1197 [01:13<03:35,  5.01it/s]
Continual predict:  10%|█         | 120/1197 [01:13<03:33,  5.03it/s]
Continual predict:  10%|█         | 121/1197 [01:18<29:20,  1.64s/it]
Continual predict:  10%|█         | 122/1197 [01:23<45:14,  2.53s/it]
Continual predict:  10%|█         | 123/1197 [01:29<1:02:46,  3.51s/it]
Continual predict:  10%|█         | 124/1197 [01:31<58:08,  3.25s/it]  
Continual predict:  10%|█         | 125/1197 [01:36<1:08:35,  3.84s/it]
Continual predict:  11%|█         | 126/1197 [01:42<1:15:51,  4.25s/it]
Continual predict:  11%|█         | 127/1197 [01:43<1:01:54,  3.47s/it]
Continual predict:  11%|█         | 128/1197 [01:46<59:48,  3.36s/it]  
Continual predict:  11%|█         | 129/1197 [01:50<1:01:48,  3.47s/it]
Continual predict:  11%|█         | 130/1197 [01:54<1:01:32,  3.46s/it]
Continual predict:  11%|█         | 131/1197 [01:56<56:50,  3.20s/it]  
Continual predict:  11%|█         | 132/1197 [01:59<55:44,  3.14s/it]
Continual predict:  11%|█         | 133/1197 [02:03<57:40,  3.25s/it]
Continual predict:  11%|█         | 134/1197 [02:09<1:16:04,  4.29s/it]
Continual predict:  11%|█▏        | 135/1197 [02:15<1:21:42,  4.62s/it]
Continual predict:  11%|█▏        | 136/1197 [02:18<1:16:30,  4.33s/it]
Continual predict:  11%|█▏        | 137/1197 [02:21<1:06:50,  3.78s/it]
Continual predict:  12%|█▏        | 138/1197 [02:22<51:45,  2.93s/it]  
Continual predict:  12%|█▏        | 139/1197 [02:23<43:18,  2.46s/it]
Continual predict:  12%|█▏        | 140/1197 [02:26<46:13,  2.62s/it]
Continual predict:  12%|█▏        | 141/1197 [02:33<1:10:07,  3.98s/it]
Continual predict:  12%|█▏        | 142/1197 [02:38<1:15:22,  4.29s/it]
Continual predict:  12%|█▏        | 143/1197 [02:42<1:11:41,  4.08s/it]
Continual predict:  12%|█▏        | 144/1197 [02:52<1:43:59,  5.93s/it]
Continual predict:  12%|█▏        | 145/1197 [02:57<1:38:33,  5.62s/it]
Continual predict:  12%|█▏        | 146/1197 [02:59<1:19:24,  4.53s/it]
Continual predict:  12%|█▏        | 147/1197 [03:04<1:20:42,  4.61s/it]
Continual predict:  12%|█▏        | 148/1197 [03:06<1:09:06,  3.95s/it]
Continual predict:  12%|█▏        | 149/1197 [03:10<1:06:40,  3.82s/it]
Continual predict:  13%|█▎        | 150/1197 [03:12<57:58,  3.32s/it]  
Continual predict:  13%|█▎        | 151/1197 [03:15<57:55,  3.32s/it]
Continual predict:  13%|█▎        | 152/1197 [03:22<1:13:11,  4.20s/it]
Continual predict:  13%|█▎        | 153/1197 [03:24<1:05:36,  3.77s/it]
Continual predict:  13%|█▎        | 154/1197 [03:30<1:13:22,  4.22s/it]
Continual predict:  13%|█▎        | 155/1197 [03:34<1:12:11,  4.16s/it]
Continual predict:  13%|█▎        | 156/1197 [03:37<1:10:11,  4.05s/it]
Continual predict:  13%|█▎        | 157/1197 [03:42<1:11:35,  4.13s/it]
Continual predict:  13%|█▎        | 158/1197 [03:43<56:22,  3.26s/it]  
Continual predict:  13%|█▎        | 159/1197 [03:50<1:18:31,  4.54s/it]
Continual predict:  13%|█▎        | 160/1197 [03:53<1:08:10,  3.94s/it]
Continual predict:  13%|█▎        | 161/1197 [03:56<1:01:58,  3.59s/it]
Continual predict:  14%|█▎        | 162/1197 [04:01<1:10:22,  4.08s/it]
Continual predict:  14%|█▎        | 163/1197 [04:03<1:01:19,  3.56s/it]
Continual predict:  14%|█▎        | 164/1197 [04:06<55:57,  3.25s/it]  
Continual predict:  14%|█▍        | 165/1197 [04:09<57:04,  3.32s/it]
Continual predict:  14%|█▍        | 166/1197 [04:12<52:22,  3.05s/it]
Continual predict:  14%|█▍        | 167/1197 [04:13<43:57,  2.56s/it]
Continual predict:  14%|█▍        | 168/1197 [04:16<43:42,  2.55s/it]
Continual predict:  14%|█▍        | 169/1197 [04:18<40:30,  2.36s/it]
Continual predict:  14%|█▍        | 170/1197 [04:21<45:44,  2.67s/it]
Continual predict:  14%|█▍        | 171/1197 [04:23<41:13,  2.41s/it]
Continual predict:  14%|█▍        | 172/1197 [04:26<43:41,  2.56s/it]
Continual predict:  14%|█▍        | 173/1197 [04:30<52:28,  3.07s/it]
Continual predict:  15%|█▍        | 174/1197 [04:31<43:47,  2.57s/it]
Continual predict:  15%|█▍        | 175/1197 [04:35<50:21,  2.96s/it]
Continual predict:  15%|█▍        | 176/1197 [04:38<51:30,  3.03s/it]
Continual predict:  15%|█▍        | 177/1197 [04:42<51:38,  3.04s/it]
Continual predict:  15%|█▍        | 178/1197 [04:43<43:35,  2.57s/it]
Continual predict:  15%|█▍        | 179/1197 [04:44<38:07,  2.25s/it]
Continual predict:  15%|█▌        | 180/1197 [04:46<33:52,  2.00s/it]
Continual predict:  15%|█▌        | 181/1197 [04:47<30:55,  1.83s/it]
Continual predict:  15%|█▌        | 182/1197 [04:53<48:33,  2.87s/it]
Continual predict:  15%|█▌        | 183/1197 [04:53<35:10,  2.08s/it]
Continual predict:  15%|█▌        | 184/1197 [04:54<30:36,  1.81s/it]
Continual predict:  15%|█▌        | 185/1197 [04:58<40:32,  2.40s/it]
Continual predict:  16%|█▌        | 186/1197 [04:59<34:32,  2.05s/it]
Continual predict:  16%|█▌        | 187/1197 [05:03<42:12,  2.51s/it]
Continual predict:  16%|█▌        | 188/1197 [05:06<44:22,  2.64s/it]
Continual predict:  16%|█▌        | 189/1197 [05:11<58:39,  3.49s/it]
Continual predict:  16%|█▌        | 190/1197 [05:15<59:31,  3.55s/it]
Continual predict:  16%|█▌        | 191/1197 [05:16<49:10,  2.93s/it]
Continual predict:  16%|█▌        | 192/1197 [05:20<52:57,  3.16s/it]
Continual predict:  16%|█▌        | 193/1197 [05:25<1:02:12,  3.72s/it]
Continual predict:  16%|█▌        | 194/1197 [05:30<1:09:53,  4.18s/it]
Continual predict:  16%|█▋        | 195/1197 [05:36<1:18:10,  4.68s/it]
Continual predict:  16%|█▋        | 196/1197 [05:40<1:12:50,  4.37s/it]
Continual predict:  16%|█▋        | 197/1197 [05:43<1:08:50,  4.13s/it]
Continual predict:  17%|█▋        | 198/1197 [05:51<1:28:19,  5.31s/it]
Continual predict:  17%|█▋        | 199/1197 [05:53<1:07:40,  4.07s/it]
Continual predict:  17%|█▋        | 200/1197 [05:54<55:40,  3.35s/it]  
Continual predict:  17%|█▋        | 201/1197 [05:57<53:07,  3.20s/it]
Continual predict:  17%|█▋        | 202/1197 [06:01<55:02,  3.32s/it]
Continual predict:  17%|█▋        | 203/1197 [06:06<1:06:27,  4.01s/it]
Continual predict:  17%|█▋        | 204/1197 [06:10<1:03:17,  3.82s/it]
Continual predict:  17%|█▋        | 205/1197 [06:17<1:19:43,  4.82s/it]
Continual predict:  17%|█▋        | 206/1197 [06:20<1:10:42,  4.28s/it]
Continual predict:  17%|█▋        | 207/1197 [06:24<1:09:24,  4.21s/it]
Continual predict:  17%|█▋        | 208/1197 [06:31<1:21:28,  4.94s/it]
Continual predict:  17%|█▋        | 209/1197 [06:33<1:11:35,  4.35s/it]
Continual predict:  18%|█▊        | 210/1197 [06:35<58:54,  3.58s/it]  
Continual predict:  18%|█▊        | 211/1197 [06:38<55:32,  3.38s/it]
Continual predict:  18%|█▊        | 212/1197 [06:42<55:57,  3.41s/it]
Continual predict:  18%|█▊        | 213/1197 [06:46<59:29,  3.63s/it]
Continual predict:  18%|█▊        | 214/1197 [06:50<1:01:30,  3.75s/it]
Continual predict:  18%|█▊        | 215/1197 [06:51<47:49,  2.92s/it]  
Continual predict:  18%|█▊        | 216/1197 [06:54<50:08,  3.07s/it]
Continual predict:  18%|█▊        | 217/1197 [06:57<46:32,  2.85s/it]
Continual predict:  18%|█▊        | 218/1197 [07:01<54:06,  3.32s/it]
Continual predict:  18%|█▊        | 219/1197 [07:05<55:46,  3.42s/it]
Continual predict:  18%|█▊        | 220/1197 [07:09<1:01:07,  3.75s/it]
Continual predict:  18%|█▊        | 221/1197 [07:09<43:57,  2.70s/it]  
Continual predict:  19%|█▊        | 222/1197 [07:10<32:20,  1.99s/it]
Continual predict:  19%|█▊        | 223/1197 [07:10<23:46,  1.46s/it]
Continual predict:  19%|█▊        | 224/1197 [07:10<17:47,  1.10s/it]
Continual predict:  19%|█▉        | 225/1197 [07:10<13:36,  1.19it/s]
Continual predict:  19%|█▉        | 226/1197 [07:11<10:40,  1.52it/s]
Continual predict:  19%|█▉        | 227/1197 [07:11<08:52,  1.82it/s]
Continual predict:  19%|█▉        | 228/1197 [07:11<07:37,  2.12it/s]
Continual predict:  19%|█▉        | 229/1197 [07:12<06:34,  2.45it/s]
Continual predict:  19%|█▉        | 230/1197 [07:12<06:54,  2.33it/s]
Continual predict:  19%|█▉        | 231/1197 [07:12<05:59,  2.69it/s]
Continual predict:  19%|█▉        | 232/1197 [07:12<05:20,  3.01it/s]
Continual predict:  19%|█▉        | 233/1197 [07:13<04:53,  3.29it/s]
Continual predict:  20%|█▉        | 234/1197 [07:13<04:34,  3.51it/s]
Continual predict:  20%|█▉        | 235/1197 [07:13<04:20,  3.70it/s]
Continual predict:  20%|█▉        | 236/1197 [07:13<04:10,  3.83it/s]
Continual predict:  20%|█▉        | 237/1197 [07:14<04:03,  3.94it/s]
Continual predict:  20%|█▉        | 238/1197 [07:14<03:58,  4.01it/s]
Continual predict:  20%|█▉        | 239/1197 [07:14<03:55,  4.07it/s]
Continual predict:  20%|██        | 240/1197 [07:15<04:39,  3.43it/s]
Continual predict:  20%|██        | 241/1197 [07:15<04:23,  3.63it/s]
Continual predict:  20%|██        | 242/1197 [07:15<04:12,  3.78it/s]
Continual predict:  20%|██        | 243/1197 [07:15<04:04,  3.90it/s]
Continual predict:  20%|██        | 244/1197 [07:16<03:59,  3.99it/s]
Continual predict:  20%|██        | 245/1197 [07:16<03:55,  4.05it/s]
Continual predict:  21%|██        | 246/1197 [07:16<03:52,  4.09it/s]
Continual predict:  21%|██        | 247/1197 [07:16<04:05,  3.87it/s]
Continual predict:  21%|██        | 248/1197 [07:17<03:59,  3.96it/s]
Continual predict:  21%|██        | 249/1197 [07:17<04:10,  3.79it/s]
Continual predict:  21%|██        | 250/1197 [07:17<04:02,  3.90it/s]
Continual predict:  21%|██        | 251/1197 [07:17<04:12,  3.75it/s]
Continual predict:  21%|██        | 252/1197 [07:19<09:23,  1.68it/s]
Continual predict:  21%|██        | 253/1197 [07:19<07:56,  1.98it/s]
Continual predict:  21%|██        | 254/1197 [07:19<06:40,  2.35it/s]
Continual predict:  21%|██▏       | 255/1197 [07:19<05:47,  2.71it/s]
Continual predict:  21%|██▏       | 256/1197 [07:20<05:11,  3.02it/s]
Continual predict:  21%|██▏       | 257/1197 [07:21<07:35,  2.06it/s]
Continual predict:  22%|██▏       | 258/1197 [07:21<06:25,  2.43it/s]
Continual predict:  22%|██▏       | 259/1197 [07:21<05:36,  2.78it/s]
Continual predict:  22%|██▏       | 260/1197 [07:21<05:02,  3.09it/s]
Continual predict:  22%|██▏       | 261/1197 [07:22<06:31,  2.39it/s]
Continual predict:  22%|██▏       | 262/1197 [07:22<05:45,  2.71it/s]
Continual predict:  22%|██▏       | 263/1197 [07:22<05:23,  2.89it/s]
Continual predict:  22%|██▏       | 264/1197 [07:24<10:00,  1.55it/s]
Continual predict:  22%|██▏       | 265/1197 [07:24<09:47,  1.59it/s]
Continual predict:  22%|██▏       | 266/1197 [07:25<09:37,  1.61it/s]
Continual predict:  22%|██▏       | 267/1197 [07:26<09:47,  1.58it/s]
Continual predict:  22%|██▏       | 268/1197 [07:27<11:02,  1.40it/s]
Continual predict:  22%|██▏       | 269/1197 [07:27<10:46,  1.44it/s]
Continual predict:  23%|██▎       | 270/1197 [07:28<10:45,  1.44it/s]
Continual predict:  23%|██▎       | 271/1197 [07:29<11:20,  1.36it/s]
Continual predict:  23%|██▎       | 272/1197 [07:30<12:28,  1.24it/s]
Continual predict:  23%|██▎       | 273/1197 [07:30<12:01,  1.28it/s]
Continual predict:  23%|██▎       | 274/1197 [07:31<11:42,  1.31it/s]
Continual predict:  23%|██▎       | 275/1197 [07:32<11:12,  1.37it/s]
Continual predict:  23%|██▎       | 276/1197 [07:33<13:08,  1.17it/s]
Continual predict:  23%|██▎       | 277/1197 [07:34<15:35,  1.02s/it]
Continual predict:  23%|██▎       | 278/1197 [07:35<15:00,  1.02it/s]
Continual predict:  23%|██▎       | 279/1197 [07:36<14:02,  1.09it/s]
Continual predict:  23%|██▎       | 280/1197 [07:36<12:04,  1.27it/s]
Continual predict:  23%|██▎       | 281/1197 [07:38<13:14,  1.15it/s]
Continual predict:  24%|██▎       | 282/1197 [07:38<13:17,  1.15it/s]
Continual predict:  24%|██▎       | 283/1197 [07:39<12:40,  1.20it/s]
Continual predict:  24%|██▎       | 284/1197 [07:40<11:07,  1.37it/s]
Continual predict:  24%|██▍       | 285/1197 [07:41<11:42,  1.30it/s]
Continual predict:  24%|██▍       | 286/1197 [07:41<11:57,  1.27it/s]
Continual predict:  24%|██▍       | 287/1197 [07:42<12:07,  1.25it/s]
Continual predict:  24%|██▍       | 288/1197 [07:43<11:42,  1.29it/s]
Continual predict:  24%|██▍       | 289/1197 [07:44<12:51,  1.18it/s]
Continual predict:  24%|██▍       | 290/1197 [07:45<12:41,  1.19it/s]
Continual predict:  24%|██▍       | 291/1197 [07:45<11:51,  1.27it/s]
Continual predict:  24%|██▍       | 292/1197 [07:46<09:51,  1.53it/s]
Continual predict:  24%|██▍       | 293/1197 [07:47<10:56,  1.38it/s]
Continual predict:  25%|██▍       | 294/1197 [07:47<10:21,  1.45it/s]
Continual predict:  25%|██▍       | 295/1197 [07:48<09:27,  1.59it/s]
Continual predict:  25%|██▍       | 296/1197 [07:49<11:26,  1.31it/s]
Continual predict:  25%|██▍       | 297/1197 [07:50<11:10,  1.34it/s]
Continual predict:  25%|██▍       | 298/1197 [07:50<10:01,  1.49it/s]
Continual predict:  25%|██▍       | 299/1197 [07:51<11:25,  1.31it/s]
Continual predict:  25%|██▌       | 300/1197 [07:51<09:35,  1.56it/s]
Continual predict:  25%|██▌       | 301/1197 [07:52<11:10,  1.34it/s]
Continual predict:  25%|██▌       | 302/1197 [07:53<09:55,  1.50it/s]
Continual predict:  25%|██▌       | 303/1197 [07:54<14:19,  1.04it/s]
Continual predict:  25%|██▌       | 304/1197 [07:55<11:37,  1.28it/s]
Continual predict:  25%|██▌       | 305/1197 [07:56<12:30,  1.19it/s]
Continual predict:  26%|██▌       | 306/1197 [07:57<12:41,  1.17it/s]
Continual predict:  26%|██▌       | 307/1197 [07:58<12:45,  1.16it/s]
Continual predict:  26%|██▌       | 308/1197 [07:58<11:03,  1.34it/s]
Continual predict:  26%|██▌       | 309/1197 [07:59<09:55,  1.49it/s]
Continual predict:  26%|██▌       | 310/1197 [07:59<09:51,  1.50it/s]
Continual predict:  26%|██▌       | 311/1197 [08:00<09:33,  1.55it/s]
Continual predict:  26%|██▌       | 312/1197 [08:01<10:06,  1.46it/s]
Continual predict:  26%|██▌       | 313/1197 [08:02<11:15,  1.31it/s]
Continual predict:  26%|██▌       | 314/1197 [08:02<10:05,  1.46it/s]
Continual predict:  26%|██▋       | 315/1197 [08:03<11:22,  1.29it/s]
Continual predict:  26%|██▋       | 316/1197 [08:04<11:49,  1.24it/s]
Continual predict:  26%|██▋       | 317/1197 [08:04<09:51,  1.49it/s]
Continual predict:  27%|██▋       | 318/1197 [08:05<09:03,  1.62it/s]
Continual predict:  27%|██▋       | 319/1197 [08:06<14:04,  1.04it/s]
Continual predict:  27%|██▋       | 320/1197 [08:07<13:15,  1.10it/s]
Continual predict:  27%|██▋       | 321/1197 [08:08<12:36,  1.16it/s]
Continual predict:  27%|██▋       | 322/1197 [08:09<10:53,  1.34it/s]
Continual predict:  27%|██▋       | 323/1197 [08:09<11:15,  1.29it/s]
Continual predict:  27%|██▋       | 324/1197 [08:10<12:07,  1.20it/s]
Continual predict:  27%|██▋       | 325/1197 [08:11<12:05,  1.20it/s]
Continual predict:  27%|██▋       | 326/1197 [08:12<11:27,  1.27it/s]
Continual predict:  27%|██▋       | 327/1197 [08:13<13:30,  1.07it/s]
Continual predict:  27%|██▋       | 328/1197 [08:14<14:25,  1.00it/s]
Continual predict:  27%|██▋       | 329/1197 [08:15<12:39,  1.14it/s]
Continual predict:  28%|██▊       | 330/1197 [08:15<10:24,  1.39it/s]
Continual predict:  28%|██▊       | 331/1197 [08:16<10:05,  1.43it/s]
Continual predict:  28%|██▊       | 332/1197 [08:16<09:38,  1.50it/s]
Continual predict:  28%|██▊       | 333/1197 [08:18<12:25,  1.16it/s]
Continual predict:  28%|██▊       | 334/1197 [08:19<12:54,  1.11it/s]
Continual predict:  28%|██▊       | 335/1197 [08:19<11:05,  1.30it/s]
Continual predict:  28%|██▊       | 336/1197 [08:20<09:52,  1.45it/s]
Continual predict:  28%|██▊       | 337/1197 [08:20<09:35,  1.49it/s]
Continual predict:  28%|██▊       | 338/1197 [08:21<09:44,  1.47it/s]
Continual predict:  28%|██▊       | 339/1197 [08:22<09:20,  1.53it/s]
Continual predict:  28%|██▊       | 340/1197 [08:23<11:32,  1.24it/s]
Continual predict:  28%|██▊       | 341/1197 [08:24<11:34,  1.23it/s]
Continual predict:  29%|██▊       | 342/1197 [08:24<10:11,  1.40it/s]
Continual predict:  29%|██▊       | 343/1197 [08:25<10:37,  1.34it/s]
Continual predict:  29%|██▊       | 344/1197 [08:26<11:51,  1.20it/s]
Continual predict:  29%|██▉       | 345/1197 [08:26<10:24,  1.36it/s]
Continual predict:  29%|██▉       | 346/1197 [08:27<11:01,  1.29it/s]
Continual predict:  29%|██▉       | 347/1197 [08:29<14:27,  1.02s/it]
Continual predict:  29%|██▉       | 348/1197 [08:29<11:42,  1.21it/s]
Continual predict:  29%|██▉       | 349/1197 [08:30<10:13,  1.38it/s]
Continual predict:  29%|██▉       | 350/1197 [08:30<10:08,  1.39it/s]
Continual predict:  29%|██▉       | 351/1197 [08:32<12:15,  1.15it/s]
Continual predict:  29%|██▉       | 352/1197 [08:33<12:02,  1.17it/s]
Continual predict:  29%|██▉       | 353/1197 [08:34<14:23,  1.02s/it]
Continual predict:  30%|██▉       | 354/1197 [08:35<15:49,  1.13s/it]
Continual predict:  30%|██▉       | 355/1197 [08:36<13:51,  1.01it/s]
Continual predict:  30%|██▉       | 356/1197 [08:37<13:50,  1.01it/s]
Continual predict:  30%|██▉       | 357/1197 [08:38<14:31,  1.04s/it]
Continual predict:  30%|██▉       | 358/1197 [08:39<13:35,  1.03it/s]
Continual predict:  30%|██▉       | 359/1197 [08:40<12:23,  1.13it/s]
Continual predict:  30%|███       | 360/1197 [08:41<14:21,  1.03s/it]
Continual predict:  30%|███       | 361/1197 [08:43<17:11,  1.23s/it]
Continual predict:  30%|███       | 362/1197 [08:43<14:44,  1.06s/it]
Continual predict:  30%|███       | 363/1197 [08:44<12:22,  1.12it/s]
Continual predict:  30%|███       | 364/1197 [08:44<10:25,  1.33it/s]
Continual predict:  30%|███       | 365/1197 [08:45<09:11,  1.51it/s]
Continual predict:  31%|███       | 366/1197 [08:45<07:50,  1.77it/s]
Continual predict:  31%|███       | 367/1197 [08:45<06:53,  2.01it/s]
Continual predict:  31%|███       | 368/1197 [08:46<06:14,  2.22it/s]
Continual predict:  31%|███       | 369/1197 [08:46<05:46,  2.39it/s]
Continual predict:  31%|███       | 370/1197 [08:46<05:27,  2.52it/s]
Continual predict:  31%|███       | 371/1197 [08:47<05:30,  2.50it/s]
Continual predict:  31%|███       | 372/1197 [08:47<05:48,  2.37it/s]
Continual predict:  31%|███       | 373/1197 [08:48<06:00,  2.28it/s]
Continual predict:  31%|███       | 374/1197 [08:48<05:35,  2.45it/s]
Continual predict:  31%|███▏      | 375/1197 [08:49<05:51,  2.34it/s]
Continual predict:  31%|███▏      | 376/1197 [08:49<05:45,  2.37it/s]
Continual predict:  31%|███▏      | 377/1197 [08:49<05:25,  2.52it/s]
Continual predict:  32%|███▏      | 378/1197 [08:50<05:10,  2.64it/s]
Continual predict:  32%|███▏      | 379/1197 [08:50<05:00,  2.73it/s]
Continual predict:  32%|███▏      | 380/1197 [08:50<04:52,  2.79it/s]
Continual predict:  32%|███▏      | 381/1197 [08:51<05:03,  2.68it/s]
Continual predict:  32%|███▏      | 382/1197 [08:51<04:54,  2.76it/s]
Continual predict:  32%|███▏      | 383/1197 [08:51<04:48,  2.82it/s]
Continual predict:  32%|███▏      | 384/1197 [08:52<04:44,  2.86it/s]
Continual predict:  32%|███▏      | 385/1197 [08:52<04:41,  2.89it/s]
Continual predict:  32%|███▏      | 386/1197 [08:52<04:39,  2.90it/s]
Continual predict:  32%|███▏      | 387/1197 [08:53<04:37,  2.92it/s]
Continual predict:  32%|███▏      | 388/1197 [08:53<04:36,  2.93it/s]
Continual predict:  32%|███▏      | 389/1197 [08:54<04:52,  2.77it/s]
Continual predict:  33%|███▎      | 390/1197 [08:54<04:46,  2.82it/s]
Continual predict:  33%|███▎      | 391/1197 [08:54<04:42,  2.86it/s]
Continual predict:  33%|███▎      | 392/1197 [08:55<04:55,  2.72it/s]
Continual predict:  33%|███▎      | 393/1197 [08:55<04:55,  2.72it/s]
Continual predict:  33%|███▎      | 394/1197 [08:55<04:52,  2.74it/s]
Continual predict:  33%|███▎      | 395/1197 [08:56<04:30,  2.97it/s]
Continual predict:  33%|███▎      | 396/1197 [08:56<05:04,  2.63it/s]
Continual predict:  33%|███▎      | 397/1197 [08:56<04:54,  2.71it/s]
Continual predict:  33%|███▎      | 398/1197 [08:57<04:48,  2.77it/s]
Continual predict:  33%|███▎      | 399/1197 [08:57<04:43,  2.82it/s]
Continual predict:  33%|███▎      | 400/1197 [08:57<04:40,  2.84it/s]
Continual predict:  34%|███▎      | 401/1197 [08:58<04:37,  2.87it/s]
Continual predict:  34%|███▎      | 402/1197 [08:58<04:34,  2.90it/s]
Continual predict:  34%|███▎      | 403/1197 [08:59<04:48,  2.75it/s]
Continual predict:  34%|███▍      | 404/1197 [08:59<04:42,  2.81it/s]
Continual predict:  34%|███▍      | 405/1197 [08:59<05:09,  2.56it/s]
Continual predict:  34%|███▍      | 406/1197 [09:00<04:56,  2.67it/s]
Continual predict:  34%|███▍      | 407/1197 [09:00<04:47,  2.75it/s]
Continual predict:  34%|███▍      | 408/1197 [09:00<04:40,  2.81it/s]
Continual predict:  34%|███▍      | 409/1197 [09:01<04:35,  2.86it/s]
Continual predict:  34%|███▍      | 410/1197 [09:01<05:04,  2.58it/s]
Continual predict:  34%|███▍      | 411/1197 [09:02<05:24,  2.42it/s]
Continual predict:  34%|███▍      | 412/1197 [09:02<05:06,  2.56it/s]
Continual predict:  35%|███▍      | 413/1197 [09:02<04:53,  2.67it/s]
Continual predict:  35%|███▍      | 414/1197 [09:03<04:44,  2.75it/s]
Continual predict:  35%|███▍      | 415/1197 [09:03<04:38,  2.81it/s]
Continual predict:  35%|███▍      | 416/1197 [09:03<04:33,  2.85it/s]
Continual predict:  35%|███▍      | 417/1197 [09:04<04:30,  2.89it/s]
Continual predict:  35%|███▍      | 418/1197 [09:04<04:43,  2.74it/s]
Continual predict:  35%|███▌      | 419/1197 [09:05<05:09,  2.52it/s]
Continual predict:  35%|███▌      | 420/1197 [09:05<05:26,  2.38it/s]
Continual predict:  35%|███▌      | 421/1197 [09:06<05:44,  2.25it/s]
Continual predict:  35%|███▌      | 422/1197 [09:06<05:39,  2.28it/s]
Continual predict:  35%|███▌      | 423/1197 [09:06<05:32,  2.33it/s]
Continual predict:  35%|███▌      | 424/1197 [09:07<05:42,  2.26it/s]
Continual predict:  36%|███▌      | 425/1197 [09:07<05:49,  2.21it/s]
Continual predict:  36%|███▌      | 426/1197 [09:08<05:54,  2.18it/s]
Continual predict:  36%|███▌      | 427/1197 [09:08<05:41,  2.25it/s]
Continual predict:  36%|███▌      | 428/1197 [09:09<05:17,  2.43it/s]
Continual predict:  36%|███▌      | 429/1197 [09:09<04:59,  2.56it/s]
Continual predict:  36%|███▌      | 430/1197 [09:09<04:47,  2.66it/s]
Continual predict:  36%|███▌      | 431/1197 [09:10<04:54,  2.60it/s]
Continual predict:  36%|███▌      | 432/1197 [09:10<04:43,  2.70it/s]
Continual predict:  36%|███▌      | 433/1197 [09:10<04:36,  2.76it/s]
Continual predict:  36%|███▋      | 434/1197 [09:11<04:46,  2.66it/s]
Continual predict:  36%|███▋      | 435/1197 [09:11<04:37,  2.75it/s]
Continual predict:  36%|███▋      | 436/1197 [09:12<05:02,  2.52it/s]
Continual predict:  37%|███▋      | 437/1197 [09:12<05:34,  2.27it/s]
Continual predict:  37%|███▋      | 438/1197 [09:12<05:11,  2.44it/s]
Continual predict:  37%|███▋      | 439/1197 [09:13<05:09,  2.45it/s]
Continual predict:  37%|███▋      | 440/1197 [09:13<05:08,  2.45it/s]
Continual predict:  37%|███▋      | 441/1197 [09:14<05:07,  2.46it/s]
Continual predict:  37%|███▋      | 442/1197 [09:14<04:51,  2.59it/s]
Continual predict:  37%|███▋      | 443/1197 [09:14<04:40,  2.68it/s]
Continual predict:  37%|███▋      | 444/1197 [09:15<04:32,  2.76it/s]
Continual predict:  37%|███▋      | 445/1197 [09:15<04:26,  2.82it/s]
Continual predict:  37%|███▋      | 446/1197 [09:15<04:38,  2.70it/s]
Continual predict:  37%|███▋      | 447/1197 [09:16<04:49,  2.59it/s]
Continual predict:  37%|███▋      | 448/1197 [09:16<05:03,  2.47it/s]
Continual predict:  38%|███▊      | 449/1197 [09:17<05:19,  2.34it/s]
Continual predict:  38%|███▊      | 450/1197 [09:17<04:59,  2.49it/s]
Continual predict:  38%|███▊      | 451/1197 [09:17<04:45,  2.62it/s]
Continual predict:  38%|███▊      | 452/1197 [09:18<04:35,  2.71it/s]
Continual predict:  38%|███▊      | 453/1197 [09:18<04:27,  2.78it/s]
Continual predict:  38%|███▊      | 454/1197 [09:18<04:22,  2.83it/s]
Continual predict:  38%|███▊      | 455/1197 [09:19<04:19,  2.86it/s]
Continual predict:  38%|███▊      | 456/1197 [09:19<04:31,  2.73it/s]
Continual predict:  38%|███▊      | 457/1197 [09:20<04:40,  2.64it/s]
Continual predict:  38%|███▊      | 458/1197 [09:20<04:31,  2.72it/s]
Continual predict:  38%|███▊      | 459/1197 [09:20<04:24,  2.79it/s]
Continual predict:  38%|███▊      | 460/1197 [09:21<04:20,  2.83it/s]
Continual predict:  39%|███▊      | 461/1197 [09:21<04:16,  2.87it/s]
Continual predict:  39%|███▊      | 462/1197 [09:21<04:13,  2.89it/s]
Continual predict:  39%|███▊      | 463/1197 [09:22<04:12,  2.91it/s]
Continual predict:  39%|███▉      | 464/1197 [09:22<04:37,  2.65it/s]
Continual predict:  39%|███▉      | 465/1197 [09:22<04:10,  2.92it/s]
Continual predict:  39%|███▉      | 466/1197 [09:23<04:20,  2.80it/s]
Continual predict:  39%|███▉      | 467/1197 [09:23<04:13,  2.88it/s]
Continual predict:  39%|███▉      | 468/1197 [09:23<04:21,  2.79it/s]
Continual predict:  39%|███▉      | 469/1197 [09:24<04:27,  2.72it/s]
Continual predict:  39%|███▉      | 470/1197 [09:24<04:32,  2.67it/s]
Continual predict:  39%|███▉      | 471/1197 [09:25<04:34,  2.64it/s]
Continual predict:  39%|███▉      | 472/1197 [09:26<07:12,  1.68it/s]
Continual predict:  40%|███▉      | 473/1197 [09:26<07:23,  1.63it/s]
Continual predict:  40%|███▉      | 474/1197 [09:27<06:51,  1.76it/s]
Continual predict:  40%|███▉      | 475/1197 [09:27<06:12,  1.94it/s]
Continual predict:  40%|███▉      | 476/1197 [09:28<05:58,  2.01it/s]
Continual predict:  40%|███▉      | 477/1197 [09:28<05:35,  2.15it/s]
Continual predict:  40%|███▉      | 478/1197 [09:29<05:32,  2.16it/s]
Continual predict:  40%|████      | 479/1197 [09:29<07:10,  1.67it/s]
Continual predict:  40%|████      | 480/1197 [09:30<06:25,  1.86it/s]
Continual predict:  40%|████      | 481/1197 [09:30<06:07,  1.95it/s]
Continual predict:  40%|████      | 482/1197 [09:31<06:15,  1.90it/s]
Continual predict:  40%|████      | 483/1197 [09:32<07:44,  1.54it/s]
Continual predict:  40%|████      | 484/1197 [09:33<09:38,  1.23it/s]
Continual predict:  41%|████      | 485/1197 [09:34<11:11,  1.06it/s]
Continual predict:  41%|████      | 486/1197 [09:35<12:00,  1.01s/it]
Continual predict:  41%|████      | 487/1197 [09:37<14:19,  1.21s/it]
Continual predict:  41%|████      | 488/1197 [09:38<11:45,  1.00it/s]
Continual predict:  41%|████      | 489/1197 [09:38<11:00,  1.07it/s]
Continual predict:  41%|████      | 490/1197 [09:39<10:44,  1.10it/s]
Continual predict:  41%|████      | 491/1197 [09:40<10:35,  1.11it/s]
Continual predict:  41%|████      | 492/1197 [09:41<10:59,  1.07it/s]
Continual predict:  41%|████      | 493/1197 [09:42<10:30,  1.12it/s]
Continual predict:  41%|████▏     | 494/1197 [09:43<10:42,  1.09it/s]
Continual predict:  41%|████▏     | 495/1197 [09:45<13:26,  1.15s/it]
Continual predict:  41%|████▏     | 496/1197 [09:45<12:27,  1.07s/it]
Continual predict:  42%|████▏     | 497/1197 [09:46<10:53,  1.07it/s]
Continual predict:  42%|████▏     | 498/1197 [09:47<10:53,  1.07it/s]
Continual predict:  42%|████▏     | 499/1197 [09:48<11:04,  1.05it/s]
Continual predict:  42%|████▏     | 500/1197 [09:49<09:25,  1.23it/s]
Continual predict:  42%|████▏     | 501/1197 [09:49<08:42,  1.33it/s]
Continual predict:  42%|████▏     | 502/1197 [09:50<08:32,  1.36it/s]
Continual predict:  42%|████▏     | 503/1197 [09:51<08:59,  1.29it/s]
Continual predict:  42%|████▏     | 504/1197 [09:51<08:27,  1.37it/s]
Continual predict:  42%|████▏     | 505/1197 [09:53<10:07,  1.14it/s]
Continual predict:  42%|████▏     | 506/1197 [09:54<10:25,  1.10it/s]
Continual predict:  42%|████▏     | 507/1197 [09:55<10:49,  1.06it/s]
Continual predict:  42%|████▏     | 508/1197 [09:55<10:16,  1.12it/s]
Continual predict:  43%|████▎     | 509/1197 [09:56<10:24,  1.10it/s]
Continual predict:  43%|████▎     | 510/1197 [09:58<14:19,  1.25s/it]
Continual predict:  43%|████▎     | 511/1197 [10:00<14:11,  1.24s/it]
Continual predict:  43%|████▎     | 512/1197 [10:00<12:20,  1.08s/it]
Continual predict:  43%|████▎     | 513/1197 [10:02<14:35,  1.28s/it]
Continual predict:  43%|████▎     | 514/1197 [10:03<13:52,  1.22s/it]
Continual predict:  43%|████▎     | 515/1197 [10:04<12:23,  1.09s/it]
Continual predict:  43%|████▎     | 516/1197 [10:05<11:18,  1.00it/s]
Continual predict:  43%|████▎     | 517/1197 [10:06<11:31,  1.02s/it]
Continual predict:  43%|████▎     | 518/1197 [10:07<11:30,  1.02s/it]
Continual predict:  43%|████▎     | 519/1197 [10:08<13:43,  1.21s/it]
Continual predict:  43%|████▎     | 520/1197 [10:09<12:03,  1.07s/it]
Continual predict:  44%|████▎     | 521/1197 [10:10<12:02,  1.07s/it]
Continual predict:  44%|████▎     | 522/1197 [10:11<10:42,  1.05it/s]
Continual predict:  44%|████▎     | 523/1197 [10:12<10:24,  1.08it/s]
Continual predict:  44%|████▍     | 524/1197 [10:12<09:40,  1.16it/s]
Continual predict:  44%|████▍     | 525/1197 [10:13<09:55,  1.13it/s]
Continual predict:  44%|████▍     | 526/1197 [10:14<10:22,  1.08it/s]
Continual predict:  44%|████▍     | 527/1197 [10:15<10:10,  1.10it/s]
Continual predict:  44%|████▍     | 528/1197 [10:17<12:18,  1.10s/it]
Continual predict:  44%|████▍     | 529/1197 [10:18<11:12,  1.01s/it]
Continual predict:  44%|████▍     | 530/1197 [10:19<12:02,  1.08s/it]
Continual predict:  44%|████▍     | 531/1197 [10:20<12:58,  1.17s/it]
Continual predict:  44%|████▍     | 532/1197 [10:22<14:17,  1.29s/it]
Continual predict:  45%|████▍     | 533/1197 [10:22<11:33,  1.04s/it]
Continual predict:  45%|████▍     | 534/1197 [10:23<10:58,  1.01it/s]
Continual predict:  45%|████▍     | 535/1197 [10:24<10:53,  1.01it/s]
Continual predict:  45%|████▍     | 536/1197 [10:25<08:55,  1.23it/s]
Continual predict:  45%|████▍     | 537/1197 [10:25<09:19,  1.18it/s]
Continual predict:  45%|████▍     | 538/1197 [10:26<09:22,  1.17it/s]
Continual predict:  45%|████▌     | 539/1197 [10:27<10:10,  1.08it/s]
Continual predict:  45%|████▌     | 540/1197 [10:29<11:34,  1.06s/it]
Continual predict:  45%|████▌     | 541/1197 [10:29<10:08,  1.08it/s]
Continual predict:  45%|████▌     | 542/1197 [10:30<08:59,  1.21it/s]
Continual predict:  45%|████▌     | 543/1197 [10:31<09:13,  1.18it/s]
Continual predict:  45%|████▌     | 544/1197 [10:32<10:29,  1.04it/s]
Continual predict:  46%|████▌     | 545/1197 [10:34<13:14,  1.22s/it]
Continual predict:  46%|████▌     | 546/1197 [10:35<12:04,  1.11s/it]
Continual predict:  46%|████▌     | 547/1197 [10:36<11:45,  1.09s/it]
Continual predict:  46%|████▌     | 548/1197 [10:37<11:22,  1.05s/it]
Continual predict:  46%|████▌     | 549/1197 [10:38<11:33,  1.07s/it]
Continual predict:  46%|████▌     | 550/1197 [10:39<11:53,  1.10s/it]
Continual predict:  46%|████▌     | 551/1197 [10:40<10:02,  1.07it/s]
Continual predict:  46%|████▌     | 552/1197 [10:40<09:12,  1.17it/s]
Continual predict:  46%|████▌     | 553/1197 [10:42<11:15,  1.05s/it]
Continual predict:  46%|████▋     | 554/1197 [10:43<11:56,  1.11s/it]
Continual predict:  46%|████▋     | 555/1197 [10:44<12:04,  1.13s/it]
Continual predict:  46%|████▋     | 556/1197 [10:45<10:43,  1.00s/it]
Continual predict:  47%|████▋     | 557/1197 [10:47<12:45,  1.20s/it]
Continual predict:  47%|████▋     | 558/1197 [10:47<10:47,  1.01s/it]
Continual predict:  47%|████▋     | 559/1197 [10:48<11:30,  1.08s/it]
Continual predict:  47%|████▋     | 560/1197 [10:50<12:01,  1.13s/it]
Continual predict:  47%|████▋     | 561/1197 [10:51<13:08,  1.24s/it]
Continual predict:  47%|████▋     | 562/1197 [10:53<13:45,  1.30s/it]
Continual predict:  47%|████▋     | 563/1197 [10:54<14:36,  1.38s/it]
Continual predict:  47%|████▋     | 564/1197 [10:56<14:48,  1.40s/it]
Continual predict:  47%|████▋     | 565/1197 [10:57<14:47,  1.40s/it]
Continual predict:  47%|████▋     | 566/1197 [10:58<13:03,  1.24s/it]
Continual predict:  47%|████▋     | 567/1197 [10:59<13:19,  1.27s/it]
Continual predict:  47%|████▋     | 568/1197 [11:00<12:03,  1.15s/it]
Continual predict:  48%|████▊     | 569/1197 [11:01<11:46,  1.13s/it]
Continual predict:  48%|████▊     | 570/1197 [11:02<09:45,  1.07it/s]
Continual predict:  48%|████▊     | 571/1197 [11:03<09:52,  1.06it/s]
Continual predict:  48%|████▊     | 572/1197 [11:03<08:43,  1.19it/s]
Continual predict:  48%|████▊     | 573/1197 [11:05<10:07,  1.03it/s]
Continual predict:  48%|████▊     | 574/1197 [11:05<09:17,  1.12it/s]
Continual predict:  48%|████▊     | 575/1197 [11:06<10:02,  1.03it/s]
Continual predict:  48%|████▊     | 576/1197 [11:07<08:59,  1.15it/s]
Continual predict:  48%|████▊     | 577/1197 [11:08<09:17,  1.11it/s]
Continual predict:  48%|████▊     | 578/1197 [11:09<08:10,  1.26it/s]
Continual predict:  48%|████▊     | 579/1197 [11:10<08:51,  1.16it/s]
Continual predict:  48%|████▊     | 580/1197 [11:11<10:39,  1.04s/it]
Continual predict:  49%|████▊     | 581/1197 [11:12<09:32,  1.08it/s]
Continual predict:  49%|████▊     | 582/1197 [11:12<07:09,  1.43it/s]
Continual predict:  49%|████▊     | 583/1197 [11:12<05:29,  1.87it/s]
Continual predict:  49%|████▉     | 584/1197 [11:12<04:19,  2.36it/s]
Continual predict:  49%|████▉     | 585/1197 [11:12<03:31,  2.90it/s]
Continual predict:  49%|████▉     | 586/1197 [11:12<02:58,  3.43it/s]
Continual predict:  49%|████▉     | 587/1197 [11:13<02:41,  3.78it/s]
Continual predict:  49%|████▉     | 588/1197 [11:13<02:21,  4.31it/s]
Continual predict:  49%|████▉     | 589/1197 [11:13<02:07,  4.77it/s]
Continual predict:  49%|████▉     | 590/1197 [11:13<01:57,  5.18it/s]
Continual predict:  49%|████▉     | 591/1197 [11:13<01:50,  5.49it/s]
Continual predict:  49%|████▉     | 592/1197 [11:13<01:45,  5.73it/s]
Continual predict:  50%|████▉     | 593/1197 [11:14<01:41,  5.94it/s]
Continual predict:  50%|████▉     | 594/1197 [11:14<01:39,  6.09it/s]
Continual predict:  50%|████▉     | 595/1197 [11:14<01:36,  6.21it/s]
Continual predict:  50%|████▉     | 596/1197 [11:14<01:35,  6.29it/s]
Continual predict:  50%|████▉     | 597/1197 [11:14<01:34,  6.35it/s]
Continual predict:  50%|████▉     | 598/1197 [11:14<01:34,  6.36it/s]
Continual predict:  50%|█████     | 599/1197 [11:15<01:34,  6.35it/s]
Continual predict:  50%|█████     | 600/1197 [11:15<01:33,  6.39it/s]
Continual predict:  50%|█████     | 601/1197 [11:15<01:33,  6.38it/s]
Continual predict:  50%|█████     | 602/1197 [11:15<01:33,  6.36it/s]
Continual predict:  50%|█████     | 603/1197 [11:15<01:33,  6.37it/s]
Continual predict:  50%|█████     | 604/1197 [11:15<01:33,  6.36it/s]
Continual predict:  51%|█████     | 605/1197 [11:15<01:33,  6.36it/s]
Continual predict:  51%|█████     | 606/1197 [11:16<01:33,  6.35it/s]
Continual predict:  51%|█████     | 607/1197 [11:16<01:32,  6.40it/s]
Continual predict:  51%|█████     | 608/1197 [11:16<01:31,  6.43it/s]
Continual predict:  51%|█████     | 609/1197 [11:16<01:31,  6.45it/s]
Continual predict:  51%|█████     | 610/1197 [11:16<01:30,  6.46it/s]
Continual predict:  51%|█████     | 611/1197 [11:16<01:30,  6.45it/s]
Continual predict:  51%|█████     | 612/1197 [11:17<01:31,  6.43it/s]
Continual predict:  51%|█████     | 613/1197 [11:17<01:31,  6.42it/s]
Continual predict:  51%|█████▏    | 614/1197 [11:17<01:30,  6.45it/s]
Continual predict:  51%|█████▏    | 615/1197 [11:17<01:30,  6.43it/s]
Continual predict:  51%|█████▏    | 616/1197 [11:17<01:30,  6.41it/s]
Continual predict:  52%|█████▏    | 617/1197 [11:17<01:30,  6.44it/s]
Continual predict:  52%|█████▏    | 618/1197 [11:18<01:30,  6.42it/s]
Continual predict:  52%|█████▏    | 619/1197 [11:18<01:30,  6.40it/s]
Continual predict:  52%|█████▏    | 620/1197 [11:18<01:30,  6.39it/s]
Continual predict:  52%|█████▏    | 621/1197 [11:18<01:30,  6.38it/s]
Continual predict:  52%|█████▏    | 622/1197 [11:18<01:30,  6.38it/s]
Continual predict:  52%|█████▏    | 623/1197 [11:18<01:29,  6.41it/s]
Continual predict:  52%|█████▏    | 624/1197 [11:18<01:29,  6.43it/s]
Continual predict:  52%|█████▏    | 625/1197 [11:19<01:29,  6.41it/s]
Continual predict:  52%|█████▏    | 626/1197 [11:19<01:28,  6.43it/s]
Continual predict:  52%|█████▏    | 627/1197 [11:19<01:28,  6.41it/s]
Continual predict:  52%|█████▏    | 628/1197 [11:19<01:28,  6.43it/s]
Continual predict:  53%|█████▎    | 629/1197 [11:19<01:28,  6.45it/s]
Continual predict:  53%|█████▎    | 630/1197 [11:19<01:28,  6.43it/s]
Continual predict:  53%|█████▎    | 631/1197 [11:20<01:28,  6.41it/s]
Continual predict:  53%|█████▎    | 632/1197 [11:20<01:28,  6.40it/s]
Continual predict:  53%|█████▎    | 633/1197 [11:20<01:27,  6.43it/s]
Continual predict:  53%|█████▎    | 634/1197 [11:20<01:27,  6.41it/s]
Continual predict:  53%|█████▎    | 635/1197 [11:20<01:27,  6.44it/s]
Continual predict:  53%|█████▎    | 636/1197 [11:20<01:27,  6.42it/s]
Continual predict:  53%|█████▎    | 637/1197 [11:20<01:27,  6.41it/s]
Continual predict:  53%|█████▎    | 638/1197 [11:21<01:27,  6.39it/s]
Continual predict:  53%|█████▎    | 639/1197 [11:21<01:27,  6.38it/s]
Continual predict:  53%|█████▎    | 640/1197 [11:21<01:26,  6.42it/s]
Continual predict:  54%|█████▎    | 641/1197 [11:21<01:26,  6.44it/s]
Continual predict:  54%|█████▎    | 642/1197 [11:21<01:26,  6.43it/s]
Continual predict:  54%|█████▎    | 643/1197 [11:21<01:26,  6.41it/s]
Continual predict:  54%|█████▍    | 644/1197 [11:22<01:25,  6.44it/s]
Continual predict:  54%|█████▍    | 645/1197 [11:22<01:25,  6.42it/s]
Continual predict:  54%|█████▍    | 646/1197 [11:22<01:25,  6.41it/s]
Continual predict:  54%|█████▍    | 647/1197 [11:22<01:26,  6.39it/s]
Continual predict:  54%|█████▍    | 648/1197 [11:22<01:25,  6.43it/s]
Continual predict:  54%|█████▍    | 649/1197 [11:22<01:25,  6.42it/s]
Continual predict:  54%|█████▍    | 650/1197 [11:22<01:24,  6.44it/s]
Continual predict:  54%|█████▍    | 651/1197 [11:23<01:24,  6.45it/s]
Continual predict:  54%|█████▍    | 652/1197 [11:23<01:41,  5.35it/s]
Continual predict:  55%|█████▍    | 653/1197 [11:23<01:39,  5.46it/s]
Continual predict:  55%|█████▍    | 654/1197 [11:23<01:39,  5.44it/s]
Continual predict:  55%|█████▍    | 655/1197 [11:23<01:39,  5.43it/s]
Continual predict:  55%|█████▍    | 656/1197 [11:24<01:40,  5.41it/s]
Continual predict:  55%|█████▍    | 657/1197 [11:24<01:40,  5.40it/s]
Continual predict:  55%|█████▍    | 658/1197 [11:24<01:42,  5.28it/s]
Continual predict:  55%|█████▌    | 659/1197 [11:24<01:36,  5.60it/s]
Continual predict:  55%|█████▌    | 660/1197 [11:24<01:32,  5.81it/s]
Continual predict:  55%|█████▌    | 661/1197 [11:24<01:29,  5.97it/s]
Continual predict:  55%|█████▌    | 662/1197 [11:25<01:27,  6.08it/s]
Continual predict:  55%|█████▌    | 663/1197 [11:25<01:26,  6.17it/s]
Continual predict:  55%|█████▌    | 664/1197 [11:25<01:25,  6.23it/s]
Continual predict:  56%|█████▌    | 665/1197 [11:25<01:24,  6.27it/s]
Continual predict:  56%|█████▌    | 666/1197 [11:25<01:23,  6.34it/s]
Continual predict:  56%|█████▌    | 667/1197 [11:25<01:23,  6.35it/s]
Continual predict:  56%|█████▌    | 668/1197 [11:26<01:23,  6.36it/s]
Continual predict:  56%|█████▌    | 669/1197 [11:26<01:22,  6.39it/s]
Continual predict:  56%|█████▌    | 670/1197 [11:26<01:22,  6.38it/s]
Continual predict:  56%|█████▌    | 671/1197 [11:26<01:22,  6.38it/s]
Continual predict:  56%|█████▌    | 672/1197 [11:26<01:22,  6.38it/s]
Continual predict:  56%|█████▌    | 673/1197 [11:26<01:21,  6.41it/s]
Continual predict:  56%|█████▋    | 674/1197 [11:27<01:21,  6.40it/s]
Continual predict:  56%|█████▋    | 675/1197 [11:27<01:21,  6.43it/s]
Continual predict:  56%|█████▋    | 676/1197 [11:27<01:20,  6.45it/s]
Continual predict:  57%|█████▋    | 677/1197 [11:27<01:20,  6.46it/s]
Continual predict:  57%|█████▋    | 678/1197 [11:27<01:20,  6.44it/s]
Continual predict:  57%|█████▋    | 679/1197 [11:27<01:20,  6.42it/s]
Continual predict:  57%|█████▋    | 680/1197 [11:27<01:20,  6.41it/s]
Continual predict:  57%|█████▋    | 681/1197 [11:28<01:20,  6.39it/s]
Continual predict:  57%|█████▋    | 682/1197 [11:29<04:00,  2.14it/s]
Continual predict:  57%|█████▋    | 683/1197 [11:31<09:15,  1.08s/it]
Continual predict:  57%|█████▋    | 684/1197 [11:33<10:54,  1.28s/it]
Continual predict:  57%|█████▋    | 685/1197 [11:34<09:50,  1.15s/it]
Continual predict:  57%|█████▋    | 686/1197 [11:36<13:04,  1.53s/it]
Continual predict:  57%|█████▋    | 687/1197 [11:38<13:24,  1.58s/it]
Continual predict:  57%|█████▋    | 688/1197 [11:43<21:25,  2.52s/it]
Continual predict:  58%|█████▊    | 689/1197 [11:44<17:12,  2.03s/it]
Continual predict:  58%|█████▊    | 690/1197 [11:48<22:24,  2.65s/it]
Continual predict:  58%|█████▊    | 691/1197 [11:49<18:25,  2.18s/it]
Continual predict:  58%|█████▊    | 692/1197 [11:50<14:40,  1.74s/it]
Continual predict:  58%|█████▊    | 693/1197 [11:53<17:52,  2.13s/it]
Continual predict:  58%|█████▊    | 694/1197 [11:53<14:40,  1.75s/it]
Continual predict:  58%|█████▊    | 695/1197 [11:56<15:38,  1.87s/it]
Continual predict:  58%|█████▊    | 696/1197 [11:57<14:17,  1.71s/it]
Continual predict:  58%|█████▊    | 697/1197 [11:58<11:45,  1.41s/it]
Continual predict:  58%|█████▊    | 698/1197 [12:00<13:06,  1.58s/it]
Continual predict:  58%|█████▊    | 699/1197 [12:01<12:48,  1.54s/it]
Continual predict:  58%|█████▊    | 700/1197 [12:03<12:38,  1.53s/it]
Continual predict:  59%|█████▊    | 701/1197 [12:04<12:43,  1.54s/it]
Continual predict:  59%|█████▊    | 702/1197 [12:05<11:52,  1.44s/it]
Continual predict:  59%|█████▊    | 703/1197 [12:07<12:44,  1.55s/it]
Continual predict:  59%|█████▉    | 704/1197 [12:08<11:47,  1.43s/it]
Continual predict:  59%|█████▉    | 705/1197 [12:11<15:58,  1.95s/it]
Continual predict:  59%|█████▉    | 706/1197 [12:13<15:24,  1.88s/it]
Continual predict:  59%|█████▉    | 707/1197 [12:15<14:23,  1.76s/it]
Continual predict:  59%|█████▉    | 708/1197 [12:18<17:38,  2.17s/it]
Continual predict:  59%|█████▉    | 709/1197 [12:19<16:19,  2.01s/it]
Continual predict:  59%|█████▉    | 710/1197 [12:21<15:37,  1.92s/it]
Continual predict:  59%|█████▉    | 711/1197 [12:23<14:44,  1.82s/it]
Continual predict:  59%|█████▉    | 712/1197 [12:24<14:05,  1.74s/it]
Continual predict:  60%|█████▉    | 713/1197 [12:26<13:15,  1.64s/it]
Continual predict:  60%|█████▉    | 714/1197 [12:28<14:12,  1.76s/it]
Continual predict:  60%|█████▉    | 715/1197 [12:30<15:11,  1.89s/it]
Continual predict:  60%|█████▉    | 716/1197 [12:31<14:21,  1.79s/it]
Continual predict:  60%|█████▉    | 717/1197 [12:33<13:31,  1.69s/it]
Continual predict:  60%|█████▉    | 718/1197 [12:35<13:22,  1.68s/it]
Continual predict:  60%|██████    | 719/1197 [12:37<14:03,  1.76s/it]
Continual predict:  60%|██████    | 720/1197 [12:38<12:53,  1.62s/it]
Continual predict:  60%|██████    | 721/1197 [12:40<13:05,  1.65s/it]
Continual predict:  60%|██████    | 722/1197 [12:41<12:46,  1.61s/it]
Continual predict:  60%|██████    | 723/1197 [12:43<12:26,  1.57s/it]
Continual predict:  60%|██████    | 724/1197 [12:44<12:34,  1.59s/it]
Continual predict:  61%|██████    | 725/1197 [12:46<12:40,  1.61s/it]
Continual predict:  61%|██████    | 726/1197 [12:47<12:45,  1.63s/it]
Continual predict:  61%|██████    | 727/1197 [12:49<12:26,  1.59s/it]
Continual predict:  61%|██████    | 728/1197 [12:51<13:18,  1.70s/it]
Continual predict:  61%|██████    | 729/1197 [12:53<13:30,  1.73s/it]
Continual predict:  61%|██████    | 730/1197 [12:55<13:48,  1.77s/it]
Continual predict:  61%|██████    | 731/1197 [12:56<13:39,  1.76s/it]
Continual predict:  61%|██████    | 732/1197 [12:58<12:59,  1.68s/it]
Continual predict:  61%|██████    | 733/1197 [12:59<11:49,  1.53s/it]
Continual predict:  61%|██████▏   | 734/1197 [13:01<13:31,  1.75s/it]
Continual predict:  61%|██████▏   | 735/1197 [13:04<14:46,  1.92s/it]
Continual predict:  61%|██████▏   | 736/1197 [13:06<15:02,  1.96s/it]
Continual predict:  62%|██████▏   | 737/1197 [13:07<13:35,  1.77s/it]
Continual predict:  62%|██████▏   | 738/1197 [13:09<13:41,  1.79s/it]
Continual predict:  62%|██████▏   | 739/1197 [13:11<14:24,  1.89s/it]
Continual predict:  62%|██████▏   | 740/1197 [13:13<14:09,  1.86s/it]
Continual predict:  62%|██████▏   | 741/1197 [13:15<14:10,  1.87s/it]
Continual predict:  62%|██████▏   | 742/1197 [13:16<13:50,  1.83s/it]
Continual predict:  62%|██████▏   | 743/1197 [13:18<13:34,  1.79s/it]
Continual predict:  62%|██████▏   | 744/1197 [13:20<13:45,  1.82s/it]
Continual predict:  62%|██████▏   | 745/1197 [13:22<14:01,  1.86s/it]
Continual predict:  62%|██████▏   | 746/1197 [13:24<13:56,  1.85s/it]
Continual predict:  62%|██████▏   | 747/1197 [13:26<13:46,  1.84s/it]
Continual predict:  62%|██████▏   | 748/1197 [13:28<14:12,  1.90s/it]
Continual predict:  63%|██████▎   | 749/1197 [13:30<14:45,  1.98s/it]
Continual predict:  63%|██████▎   | 750/1197 [13:31<13:38,  1.83s/it]
Continual predict:  63%|██████▎   | 751/1197 [13:33<14:15,  1.92s/it]
Continual predict:  63%|██████▎   | 752/1197 [13:35<12:55,  1.74s/it]
Continual predict:  63%|██████▎   | 753/1197 [13:36<12:19,  1.67s/it]
Continual predict:  63%|██████▎   | 754/1197 [13:37<11:23,  1.54s/it]
Continual predict:  63%|██████▎   | 755/1197 [13:39<12:27,  1.69s/it]
Continual predict:  63%|██████▎   | 756/1197 [13:41<12:14,  1.67s/it]
Continual predict:  63%|██████▎   | 757/1197 [13:43<11:40,  1.59s/it]
Continual predict:  63%|██████▎   | 758/1197 [13:44<11:25,  1.56s/it]
Continual predict:  63%|██████▎   | 759/1197 [13:46<11:33,  1.58s/it]
Continual predict:  63%|██████▎   | 760/1197 [13:47<11:19,  1.56s/it]
Continual predict:  64%|██████▎   | 761/1197 [13:50<13:45,  1.89s/it]
Continual predict:  64%|██████▎   | 762/1197 [13:52<15:10,  2.09s/it]
Continual predict:  64%|██████▎   | 763/1197 [13:54<14:51,  2.06s/it]
Continual predict:  64%|██████▍   | 764/1197 [13:56<14:27,  2.00s/it]
Continual predict:  64%|██████▍   | 765/1197 [13:57<12:47,  1.78s/it]
Continual predict:  64%|██████▍   | 766/1197 [13:59<12:49,  1.78s/it]
Continual predict:  64%|██████▍   | 767/1197 [14:01<12:28,  1.74s/it]
Continual predict:  64%|██████▍   | 768/1197 [14:02<12:07,  1.69s/it]
Continual predict:  64%|██████▍   | 769/1197 [14:04<11:45,  1.65s/it]
Continual predict:  64%|██████▍   | 770/1197 [14:06<11:43,  1.65s/it]
Continual predict:  64%|██████▍   | 771/1197 [14:07<10:52,  1.53s/it]
Continual predict:  64%|██████▍   | 772/1197 [14:08<10:14,  1.45s/it]
Continual predict:  65%|██████▍   | 773/1197 [14:10<11:58,  1.69s/it]
Continual predict:  65%|██████▍   | 774/1197 [14:12<11:52,  1.68s/it]
Continual predict:  65%|██████▍   | 775/1197 [14:14<11:25,  1.63s/it]
Continual predict:  65%|██████▍   | 776/1197 [14:16<12:07,  1.73s/it]
Continual predict:  65%|██████▍   | 777/1197 [14:17<11:45,  1.68s/it]
Continual predict:  65%|██████▍   | 778/1197 [14:19<11:59,  1.72s/it]
Continual predict:  65%|██████▌   | 779/1197 [14:21<11:39,  1.67s/it]
Continual predict:  65%|██████▌   | 780/1197 [14:22<11:24,  1.64s/it]
Continual predict:  65%|██████▌   | 781/1197 [14:24<11:45,  1.70s/it]
Continual predict:  65%|██████▌   | 782/1197 [14:25<11:07,  1.61s/it]
Continual predict:  65%|██████▌   | 783/1197 [14:27<11:39,  1.69s/it]
Continual predict:  65%|██████▌   | 784/1197 [14:29<11:41,  1.70s/it]
Continual predict:  66%|██████▌   | 785/1197 [14:30<11:05,  1.61s/it]
Continual predict:  66%|██████▌   | 786/1197 [14:32<11:18,  1.65s/it]
Continual predict:  66%|██████▌   | 787/1197 [14:34<12:03,  1.76s/it]
Continual predict:  66%|██████▌   | 788/1197 [14:35<10:56,  1.61s/it]
Continual predict:  66%|██████▌   | 789/1197 [14:37<11:18,  1.66s/it]
Continual predict:  66%|██████▌   | 790/1197 [14:40<12:50,  1.89s/it]
Continual predict:  66%|██████▌   | 791/1197 [14:41<12:37,  1.87s/it]
Continual predict:  66%|██████▌   | 792/1197 [14:43<11:40,  1.73s/it]
Continual predict:  66%|██████▌   | 793/1197 [14:45<11:39,  1.73s/it]
Continual predict:  66%|██████▋   | 794/1197 [14:47<12:53,  1.92s/it]
Continual predict:  66%|██████▋   | 795/1197 [14:48<11:50,  1.77s/it]
Continual predict:  66%|██████▋   | 796/1197 [14:50<11:34,  1.73s/it]
Continual predict:  67%|██████▋   | 797/1197 [14:52<11:33,  1.73s/it]
Continual predict:  67%|██████▋   | 798/1197 [14:54<12:13,  1.84s/it]
Continual predict:  67%|██████▋   | 799/1197 [15:01<22:16,  3.36s/it]
Continual predict:  67%|██████▋   | 800/1197 [15:03<20:06,  3.04s/it]
Continual predict:  67%|██████▋   | 801/1197 [15:04<16:42,  2.53s/it]
Continual predict:  67%|██████▋   | 802/1197 [15:06<14:29,  2.20s/it]
Continual predict:  67%|██████▋   | 803/1197 [15:07<12:16,  1.87s/it]
Continual predict:  67%|██████▋   | 804/1197 [15:11<16:15,  2.48s/it]
Continual predict:  67%|██████▋   | 805/1197 [15:14<18:34,  2.84s/it]
Continual predict:  67%|██████▋   | 806/1197 [15:17<17:10,  2.64s/it]
Continual predict:  67%|██████▋   | 807/1197 [15:18<14:45,  2.27s/it]
Continual predict:  68%|██████▊   | 808/1197 [15:20<13:41,  2.11s/it]
Continual predict:  68%|██████▊   | 809/1197 [15:21<12:33,  1.94s/it]
Continual predict:  68%|██████▊   | 810/1197 [15:24<13:20,  2.07s/it]
Continual predict:  68%|██████▊   | 811/1197 [15:25<12:40,  1.97s/it]
Continual predict:  68%|██████▊   | 812/1197 [15:30<17:08,  2.67s/it]
Continual predict:  68%|██████▊   | 813/1197 [15:38<27:06,  4.23s/it]
Continual predict:  68%|██████▊   | 814/1197 [15:40<22:59,  3.60s/it]
Continual predict:  68%|██████▊   | 815/1197 [15:42<20:46,  3.26s/it]
Continual predict:  68%|██████▊   | 816/1197 [15:45<19:34,  3.08s/it]
Continual predict:  68%|██████▊   | 817/1197 [15:46<16:40,  2.63s/it]
Continual predict:  68%|██████▊   | 818/1197 [15:48<13:58,  2.21s/it]
Continual predict:  68%|██████▊   | 819/1197 [15:50<14:07,  2.24s/it]
Continual predict:  69%|██████▊   | 820/1197 [15:52<13:10,  2.10s/it]
Continual predict:  69%|██████▊   | 821/1197 [15:54<12:54,  2.06s/it]
Continual predict:  69%|██████▊   | 822/1197 [15:57<14:22,  2.30s/it]
Continual predict:  69%|██████▉   | 823/1197 [16:00<16:50,  2.70s/it]
Continual predict:  69%|██████▉   | 824/1197 [16:11<31:15,  5.03s/it]
Continual predict:  69%|██████▉   | 825/1197 [16:12<25:08,  4.05s/it]
Continual predict:  69%|██████▉   | 826/1197 [16:16<25:02,  4.05s/it]
Continual predict:  69%|██████▉   | 827/1197 [16:18<19:39,  3.19s/it]
Continual predict:  69%|██████▉   | 828/1197 [16:22<21:46,  3.54s/it]
Continual predict:  69%|██████▉   | 829/1197 [16:28<25:24,  4.14s/it]
Continual predict:  69%|██████▉   | 830/1197 [16:30<21:34,  3.53s/it]
Continual predict:  69%|██████▉   | 831/1197 [16:42<36:49,  6.04s/it]
Continual predict:  70%|██████▉   | 832/1197 [16:43<28:29,  4.68s/it]
Continual predict:  70%|██████▉   | 833/1197 [16:44<21:36,  3.56s/it]
Continual predict:  70%|██████▉   | 834/1197 [16:46<18:56,  3.13s/it]
Continual predict:  70%|██████▉   | 835/1197 [16:48<15:48,  2.62s/it]
Continual predict:  70%|██████▉   | 836/1197 [16:50<15:54,  2.64s/it]
Continual predict:  70%|██████▉   | 837/1197 [16:55<20:14,  3.37s/it]
Continual predict:  70%|███████   | 838/1197 [16:58<18:09,  3.03s/it]
Continual predict:  70%|███████   | 839/1197 [17:04<25:02,  4.20s/it]
Continual predict:  70%|███████   | 840/1197 [17:09<26:00,  4.37s/it]
Continual predict:  70%|███████   | 841/1197 [17:11<21:02,  3.55s/it]
Continual predict:  70%|███████   | 842/1197 [17:14<19:53,  3.36s/it]
Continual predict:  70%|███████   | 843/1197 [17:16<17:21,  2.94s/it]
Continual predict:  71%|███████   | 844/1197 [17:21<21:10,  3.60s/it]
Continual predict:  71%|███████   | 845/1197 [17:25<21:48,  3.72s/it]
Continual predict:  71%|███████   | 846/1197 [17:26<17:57,  3.07s/it]
Continual predict:  71%|███████   | 847/1197 [17:27<13:46,  2.36s/it]
Continual predict:  71%|███████   | 848/1197 [17:29<12:28,  2.14s/it]
Continual predict:  71%|███████   | 849/1197 [17:30<10:54,  1.88s/it]
Continual predict:  71%|███████   | 850/1197 [17:33<13:16,  2.29s/it]
Continual predict:  71%|███████   | 851/1197 [17:35<12:50,  2.23s/it]
Continual predict:  71%|███████   | 852/1197 [17:37<11:13,  1.95s/it]
Continual predict:  71%|███████▏  | 853/1197 [17:46<24:38,  4.30s/it]
Continual predict:  71%|███████▏  | 854/1197 [17:58<37:35,  6.58s/it]
Continual predict:  71%|███████▏  | 855/1197 [18:01<30:23,  5.33s/it]
Continual predict:  72%|███████▏  | 856/1197 [18:03<25:02,  4.41s/it]
Continual predict:  72%|███████▏  | 857/1197 [18:05<20:37,  3.64s/it]
Continual predict:  72%|███████▏  | 858/1197 [18:07<17:50,  3.16s/it]
Continual predict:  72%|███████▏  | 859/1197 [18:11<19:51,  3.52s/it]
Continual predict:  72%|███████▏  | 860/1197 [18:15<20:48,  3.70s/it]
Continual predict:  72%|███████▏  | 861/1197 [18:27<34:31,  6.16s/it]
Continual predict:  72%|███████▏  | 862/1197 [18:28<25:09,  4.50s/it]
Continual predict:  72%|███████▏  | 863/1197 [18:30<21:07,  3.79s/it]
Continual predict:  72%|███████▏  | 864/1197 [18:32<17:29,  3.15s/it]
Continual predict:  72%|███████▏  | 865/1197 [18:34<15:59,  2.89s/it]
Continual predict:  72%|███████▏  | 866/1197 [18:36<14:10,  2.57s/it]
Continual predict:  72%|███████▏  | 867/1197 [18:38<13:57,  2.54s/it]
Continual predict:  73%|███████▎  | 868/1197 [18:42<15:12,  2.77s/it]
Continual predict:  73%|███████▎  | 869/1197 [18:45<16:36,  3.04s/it]
Continual predict:  73%|███████▎  | 870/1197 [18:47<14:04,  2.58s/it]
Continual predict:  73%|███████▎  | 871/1197 [18:48<12:30,  2.30s/it]
Continual predict:  73%|███████▎  | 872/1197 [18:54<17:22,  3.21s/it]
Continual predict:  73%|███████▎  | 873/1197 [18:56<15:54,  2.94s/it]
Continual predict:  73%|███████▎  | 874/1197 [18:58<13:23,  2.49s/it]
Continual predict:  73%|███████▎  | 875/1197 [19:00<12:31,  2.33s/it]
Continual predict:  73%|███████▎  | 876/1197 [19:01<10:55,  2.04s/it]
Continual predict:  73%|███████▎  | 877/1197 [19:03<11:30,  2.16s/it]
Continual predict:  73%|███████▎  | 878/1197 [19:04<08:24,  1.58s/it]
Continual predict:  73%|███████▎  | 879/1197 [19:07<10:54,  2.06s/it]
Continual predict:  74%|███████▎  | 880/1197 [19:08<10:17,  1.95s/it]
Continual predict:  74%|███████▎  | 881/1197 [19:12<12:41,  2.41s/it]
Continual predict:  74%|███████▎  | 882/1197 [19:14<12:28,  2.38s/it]
Continual predict:  74%|███████▍  | 883/1197 [19:16<11:14,  2.15s/it]
Continual predict:  74%|███████▍  | 884/1197 [19:19<13:26,  2.58s/it]
Continual predict:  74%|███████▍  | 885/1197 [19:25<17:21,  3.34s/it]
Continual predict:  74%|███████▍  | 886/1197 [19:27<15:25,  2.98s/it]
Continual predict:  74%|███████▍  | 887/1197 [19:30<15:36,  3.02s/it]
Continual predict:  74%|███████▍  | 888/1197 [19:32<14:05,  2.74s/it]
Continual predict:  74%|███████▍  | 889/1197 [19:34<13:17,  2.59s/it]
Continual predict:  74%|███████▍  | 890/1197 [19:37<13:28,  2.63s/it]
Continual predict:  74%|███████▍  | 891/1197 [19:40<13:55,  2.73s/it]
Continual predict:  75%|███████▍  | 892/1197 [19:42<13:42,  2.70s/it]
Continual predict:  75%|███████▍  | 893/1197 [19:45<13:54,  2.74s/it]
Continual predict:  75%|███████▍  | 894/1197 [19:47<12:30,  2.48s/it]
Continual predict:  75%|███████▍  | 895/1197 [19:52<15:36,  3.10s/it]
Continual predict:  75%|███████▍  | 896/1197 [19:54<13:50,  2.76s/it]
Continual predict:  75%|███████▍  | 897/1197 [19:58<16:25,  3.29s/it]
Continual predict:  75%|███████▌  | 898/1197 [19:58<11:45,  2.36s/it]
Continual predict:  75%|███████▌  | 899/1197 [19:59<08:35,  1.73s/it]
Continual predict:  75%|███████▌  | 900/1197 [19:59<06:28,  1.31s/it]
Continual predict:  75%|███████▌  | 901/1197 [19:59<04:59,  1.01s/it]
Continual predict:  75%|███████▌  | 902/1197 [19:59<03:40,  1.34it/s]
Continual predict:  75%|███████▌  | 903/1197 [19:59<02:45,  1.78it/s]
Continual predict:  76%|███████▌  | 904/1197 [20:00<02:18,  2.12it/s]
Continual predict:  76%|███████▌  | 905/1197 [20:00<02:01,  2.40it/s]
Continual predict:  76%|███████▌  | 906/1197 [20:00<01:37,  3.00it/s]
Continual predict:  76%|███████▌  | 907/1197 [20:00<01:30,  3.21it/s]
Continual predict:  76%|███████▌  | 908/1197 [20:01<01:19,  3.62it/s]
Continual predict:  76%|███████▌  | 909/1197 [20:01<01:18,  3.68it/s]
Continual predict:  76%|███████▌  | 910/1197 [20:01<01:11,  4.02it/s]
Continual predict:  76%|███████▌  | 911/1197 [20:01<01:06,  4.30it/s]
Continual predict:  76%|███████▌  | 912/1197 [20:01<00:57,  4.95it/s]
Continual predict:  76%|███████▋  | 913/1197 [20:02<00:56,  5.01it/s]
Continual predict:  76%|███████▋  | 914/1197 [20:02<01:07,  4.20it/s]
Continual predict:  76%|███████▋  | 915/1197 [20:02<01:03,  4.45it/s]
Continual predict:  77%|███████▋  | 916/1197 [20:02<01:06,  4.25it/s]
Continual predict:  77%|███████▋  | 917/1197 [20:03<01:08,  4.12it/s]
Continual predict:  77%|███████▋  | 918/1197 [20:03<00:58,  4.78it/s]
Continual predict:  77%|███████▋  | 919/1197 [20:03<01:02,  4.47it/s]
Continual predict:  77%|███████▋  | 920/1197 [20:03<01:10,  3.94it/s]
Continual predict:  77%|███████▋  | 921/1197 [20:04<01:10,  3.92it/s]
Continual predict:  77%|███████▋  | 922/1197 [20:04<00:59,  4.60it/s]
Continual predict:  77%|███████▋  | 923/1197 [20:04<01:03,  4.35it/s]
Continual predict:  77%|███████▋  | 924/1197 [20:04<01:15,  3.61it/s]
Continual predict:  77%|███████▋  | 925/1197 [20:05<01:13,  3.69it/s]
Continual predict:  77%|███████▋  | 926/1197 [20:05<01:01,  4.38it/s]
Continual predict:  77%|███████▋  | 927/1197 [20:05<00:59,  4.57it/s]
Continual predict:  78%|███████▊  | 928/1197 [20:05<00:51,  5.21it/s]
Continual predict:  78%|███████▊  | 929/1197 [20:05<00:56,  4.71it/s]
Continual predict:  78%|███████▊  | 930/1197 [20:06<01:00,  4.43it/s]
Continual predict:  78%|███████▊  | 931/1197 [20:06<01:08,  3.91it/s]
Continual predict:  78%|███████▊  | 932/1197 [20:06<01:02,  4.21it/s]
Continual predict:  78%|███████▊  | 933/1197 [20:06<01:09,  3.80it/s]
Continual predict:  78%|███████▊  | 934/1197 [20:07<00:58,  4.48it/s]
Continual predict:  78%|███████▊  | 935/1197 [20:07<01:06,  3.94it/s]
Continual predict:  78%|███████▊  | 936/1197 [20:07<01:11,  3.64it/s]
Continual predict:  78%|███████▊  | 937/1197 [20:08<01:15,  3.46it/s]
Continual predict:  78%|███████▊  | 938/1197 [20:08<01:02,  4.14it/s]
Continual predict:  78%|███████▊  | 939/1197 [20:08<01:03,  4.06it/s]
Continual predict:  79%|███████▊  | 940/1197 [20:08<00:59,  4.33it/s]
Continual predict:  79%|███████▊  | 941/1197 [20:09<01:11,  3.59it/s]
Continual predict:  79%|███████▊  | 942/1197 [20:09<01:09,  3.67it/s]
Continual predict:  79%|███████▉  | 943/1197 [20:09<01:03,  4.02it/s]
Continual predict:  79%|███████▉  | 944/1197 [20:09<00:54,  4.68it/s]
Continual predict:  79%|███████▉  | 945/1197 [20:09<00:47,  5.31it/s]
Continual predict:  79%|███████▉  | 946/1197 [20:09<00:47,  5.26it/s]
Continual predict:  79%|███████▉  | 947/1197 [20:10<01:02,  3.99it/s]
Continual predict:  79%|███████▉  | 948/1197 [20:10<01:03,  3.93it/s]
Continual predict:  79%|███████▉  | 949/1197 [20:11<01:14,  3.32it/s]
Continual predict:  79%|███████▉  | 950/1197 [20:11<01:03,  3.91it/s]
Continual predict:  79%|███████▉  | 951/1197 [20:11<01:01,  4.02it/s]
Continual predict:  80%|███████▉  | 952/1197 [20:11<00:52,  4.69it/s]
Continual predict:  80%|███████▉  | 953/1197 [20:11<00:50,  4.80it/s]
Continual predict:  80%|███████▉  | 954/1197 [20:11<00:44,  5.42it/s]
Continual predict:  80%|███████▉  | 955/1197 [20:12<00:54,  4.43it/s]
Continual predict:  80%|███████▉  | 956/1197 [20:12<00:47,  5.07it/s]
Continual predict:  80%|███████▉  | 957/1197 [20:12<00:47,  5.10it/s]
Continual predict:  80%|████████  | 958/1197 [20:12<00:46,  5.12it/s]
Continual predict:  80%|████████  | 959/1197 [20:12<00:50,  4.67it/s]
Continual predict:  80%|████████  | 960/1197 [20:13<00:44,  5.30it/s]
Continual predict:  80%|████████  | 961/1197 [20:13<00:44,  5.26it/s]
Continual predict:  80%|████████  | 962/1197 [20:13<00:49,  4.75it/s]
Continual predict:  80%|████████  | 963/1197 [20:13<00:48,  4.86it/s]
Continual predict:  81%|████████  | 964/1197 [20:13<00:42,  5.46it/s]
Continual predict:  81%|████████  | 965/1197 [20:14<00:47,  4.86it/s]
Continual predict:  81%|████████  | 966/1197 [20:14<00:51,  4.52it/s]
Continual predict:  81%|████████  | 967/1197 [20:14<00:57,  3.98it/s]
Continual predict:  81%|████████  | 968/1197 [20:15<01:07,  3.41it/s]
Continual predict:  81%|████████  | 969/1197 [20:15<01:08,  3.32it/s]
Continual predict:  81%|████████  | 970/1197 [20:15<01:05,  3.47it/s]
Continual predict:  81%|████████  | 971/1197 [20:15<01:07,  3.35it/s]
Continual predict:  81%|████████  | 972/1197 [20:16<01:04,  3.49it/s]
Continual predict:  81%|████████▏ | 973/1197 [20:16<01:02,  3.59it/s]
Continual predict:  81%|████████▏ | 974/1197 [20:16<00:52,  4.28it/s]
Continual predict:  81%|████████▏ | 975/1197 [20:16<00:49,  4.51it/s]
Continual predict:  82%|████████▏ | 976/1197 [20:17<00:55,  3.96it/s]
Continual predict:  82%|████████▏ | 977/1197 [20:17<00:51,  4.26it/s]
Continual predict:  82%|████████▏ | 978/1197 [20:17<00:44,  4.92it/s]
Continual predict:  82%|████████▏ | 979/1197 [20:17<00:47,  4.56it/s]
Continual predict:  82%|████████▏ | 980/1197 [20:17<00:41,  5.19it/s]
Continual predict:  82%|████████▏ | 981/1197 [20:18<00:45,  4.70it/s]
Continual predict:  82%|████████▏ | 982/1197 [20:18<00:48,  4.43it/s]
Continual predict:  82%|████████▏ | 983/1197 [20:18<00:46,  4.62it/s]
Continual predict:  82%|████████▏ | 984/1197 [20:18<00:40,  5.26it/s]
Continual predict:  82%|████████▏ | 985/1197 [20:19<00:53,  3.99it/s]
Continual predict:  82%|████████▏ | 986/1197 [20:19<00:45,  4.67it/s]
Continual predict:  82%|████████▏ | 987/1197 [20:19<00:39,  5.30it/s]
Continual predict:  83%|████████▎ | 988/1197 [20:19<00:35,  5.85it/s]
Continual predict:  83%|████████▎ | 989/1197 [20:19<00:40,  5.08it/s]
Continual predict:  83%|████████▎ | 990/1197 [20:20<00:48,  4.27it/s]
Continual predict:  83%|████████▎ | 991/1197 [20:20<00:41,  4.93it/s]
Continual predict:  83%|████████▎ | 992/1197 [20:20<00:40,  5.00it/s]
Continual predict:  83%|████████▎ | 993/1197 [20:20<00:40,  5.06it/s]
Continual predict:  83%|████████▎ | 994/1197 [20:20<00:35,  5.65it/s]
Continual predict:  83%|████████▎ | 995/1197 [20:20<00:32,  6.12it/s]
Continual predict:  83%|████████▎ | 996/1197 [20:20<00:30,  6.53it/s]
Continual predict:  83%|████████▎ | 997/1197 [20:21<00:33,  6.04it/s]
Continual predict:  83%|████████▎ | 998/1197 [20:21<00:30,  6.45it/s]
Continual predict:  83%|████████▎ | 999/1197 [20:21<00:32,  6.16it/s]
Continual predict:  84%|████████▎ | 1000/1197 [20:21<00:31,  6.24it/s]
Continual predict:  84%|████████▎ | 1001/1197 [20:21<00:30,  6.37it/s]
Continual predict:  84%|████████▎ | 1002/1197 [20:21<00:30,  6.34it/s]
Continual predict:  84%|████████▍ | 1003/1197 [20:22<00:29,  6.65it/s]
Continual predict:  84%|████████▍ | 1004/1197 [20:22<00:27,  6.91it/s]
Continual predict:  84%|████████▍ | 1005/1197 [20:22<00:27,  7.10it/s]
Continual predict:  84%|████████▍ | 1006/1197 [20:22<00:26,  7.26it/s]
Continual predict:  84%|████████▍ | 1007/1197 [20:22<00:26,  7.29it/s]
Continual predict:  84%|████████▍ | 1008/1197 [20:22<00:25,  7.39it/s]
Continual predict:  84%|████████▍ | 1009/1197 [20:22<00:25,  7.46it/s]
Continual predict:  84%|████████▍ | 1010/1197 [20:23<00:24,  7.50it/s]
Continual predict:  84%|████████▍ | 1011/1197 [20:23<00:24,  7.53it/s]
Continual predict:  85%|████████▍ | 1012/1197 [20:23<00:24,  7.56it/s]
Continual predict:  85%|████████▍ | 1013/1197 [20:23<00:24,  7.58it/s]
Continual predict:  85%|████████▍ | 1014/1197 [20:23<00:24,  7.58it/s]
Continual predict:  85%|████████▍ | 1015/1197 [20:23<00:23,  7.60it/s]
Continual predict:  85%|████████▍ | 1016/1197 [20:23<00:23,  7.59it/s]
Continual predict:  85%|████████▍ | 1017/1197 [20:23<00:23,  7.56it/s]
Continual predict:  85%|████████▌ | 1018/1197 [20:24<00:23,  7.57it/s]
Continual predict:  85%|████████▌ | 1019/1197 [20:24<00:23,  7.47it/s]
Continual predict:  85%|████████▌ | 1020/1197 [20:24<00:23,  7.51it/s]
Continual predict:  85%|████████▌ | 1021/1197 [20:24<00:23,  7.57it/s]
Continual predict:  85%|████████▌ | 1022/1197 [20:24<00:23,  7.59it/s]
Continual predict:  85%|████████▌ | 1023/1197 [20:24<00:23,  7.52it/s]
Continual predict:  86%|████████▌ | 1024/1197 [20:24<00:23,  7.47it/s]
Continual predict:  86%|████████▌ | 1025/1197 [20:24<00:23,  7.43it/s]
Continual predict:  86%|████████▌ | 1026/1197 [20:25<00:22,  7.48it/s]
Continual predict:  86%|████████▌ | 1027/1197 [20:25<00:22,  7.50it/s]
Continual predict:  86%|████████▌ | 1028/1197 [20:25<00:22,  7.52it/s]
Continual predict:  86%|████████▌ | 1029/1197 [20:25<00:22,  7.56it/s]
Continual predict:  86%|████████▌ | 1030/1197 [20:25<00:22,  7.58it/s]
Continual predict:  86%|████████▌ | 1031/1197 [20:25<00:21,  7.59it/s]
Continual predict:  86%|████████▌ | 1032/1197 [20:25<00:21,  7.60it/s]
Continual predict:  86%|████████▋ | 1033/1197 [20:26<00:21,  7.60it/s]
Continual predict:  86%|████████▋ | 1034/1197 [20:26<00:21,  7.53it/s]
Continual predict:  86%|████████▋ | 1035/1197 [20:26<00:21,  7.52it/s]
Continual predict:  87%|████████▋ | 1036/1197 [20:26<00:21,  7.54it/s]
Continual predict:  87%|████████▋ | 1037/1197 [20:26<00:21,  7.55it/s]
Continual predict:  87%|████████▋ | 1038/1197 [20:26<00:20,  7.58it/s]
Continual predict:  87%|████████▋ | 1039/1197 [20:26<00:20,  7.58it/s]
Continual predict:  87%|████████▋ | 1040/1197 [20:26<00:20,  7.60it/s]
Continual predict:  87%|████████▋ | 1041/1197 [20:27<00:20,  7.60it/s]
Continual predict:  87%|████████▋ | 1042/1197 [20:27<00:20,  7.61it/s]
Continual predict:  87%|████████▋ | 1043/1197 [20:27<00:20,  7.51it/s]
Continual predict:  87%|████████▋ | 1044/1197 [20:27<00:20,  7.55it/s]
Continual predict:  87%|████████▋ | 1045/1197 [20:27<00:20,  7.60it/s]
Continual predict:  87%|████████▋ | 1046/1197 [20:27<00:19,  7.62it/s]
Continual predict:  87%|████████▋ | 1047/1197 [20:27<00:19,  7.62it/s]
Continual predict:  88%|████████▊ | 1048/1197 [20:28<00:19,  7.62it/s]
Continual predict:  88%|████████▊ | 1049/1197 [20:28<00:19,  7.63it/s]
Continual predict:  88%|████████▊ | 1050/1197 [20:28<00:19,  7.62it/s]
Continual predict:  88%|████████▊ | 1051/1197 [20:28<00:19,  7.60it/s]
Continual predict:  88%|████████▊ | 1052/1197 [20:28<00:19,  7.53it/s]
Continual predict:  88%|████████▊ | 1053/1197 [20:28<00:19,  7.54it/s]
Continual predict:  88%|████████▊ | 1054/1197 [20:28<00:19,  7.48it/s]
Continual predict:  88%|████████▊ | 1055/1197 [20:28<00:18,  7.51it/s]
Continual predict:  88%|████████▊ | 1056/1197 [20:29<00:18,  7.55it/s]
Continual predict:  88%|████████▊ | 1057/1197 [20:29<00:18,  7.58it/s]
Continual predict:  88%|████████▊ | 1058/1197 [20:29<00:18,  7.59it/s]
Continual predict:  88%|████████▊ | 1059/1197 [20:29<00:18,  7.58it/s]
Continual predict:  89%|████████▊ | 1060/1197 [20:29<00:18,  7.58it/s]
Continual predict:  89%|████████▊ | 1061/1197 [20:29<00:17,  7.60it/s]
Continual predict:  89%|████████▊ | 1062/1197 [20:29<00:17,  7.62it/s]
Continual predict:  89%|████████▉ | 1063/1197 [20:30<00:17,  7.54it/s]
Continual predict:  89%|████████▉ | 1064/1197 [20:30<00:17,  7.58it/s]
Continual predict:  89%|████████▉ | 1065/1197 [20:30<00:17,  7.59it/s]
Continual predict:  89%|████████▉ | 1066/1197 [20:30<00:17,  7.59it/s]
Continual predict:  89%|████████▉ | 1067/1197 [20:30<00:17,  7.51it/s]
Continual predict:  89%|████████▉ | 1068/1197 [20:30<00:17,  7.55it/s]
Continual predict:  89%|████████▉ | 1069/1197 [20:30<00:16,  7.55it/s]
Continual predict:  89%|████████▉ | 1070/1197 [20:30<00:16,  7.58it/s]
Continual predict:  89%|████████▉ | 1071/1197 [20:31<00:16,  7.60it/s]
Continual predict:  90%|████████▉ | 1072/1197 [20:31<00:16,  7.60it/s]
Continual predict:  90%|████████▉ | 1073/1197 [20:31<00:16,  7.58it/s]
Continual predict:  90%|████████▉ | 1074/1197 [20:31<00:16,  7.54it/s]
Continual predict:  90%|████████▉ | 1075/1197 [20:31<00:16,  7.55it/s]
Continual predict:  90%|████████▉ | 1076/1197 [20:31<00:15,  7.56it/s]
Continual predict:  90%|████████▉ | 1077/1197 [20:31<00:15,  7.57it/s]
Continual predict:  90%|█████████ | 1078/1197 [20:31<00:15,  7.55it/s]
Continual predict:  90%|█████████ | 1079/1197 [20:32<00:17,  6.93it/s]
Continual predict:  90%|█████████ | 1080/1197 [20:32<00:17,  6.83it/s]
Continual predict:  90%|█████████ | 1081/1197 [20:32<00:16,  6.85it/s]
Continual predict:  90%|█████████ | 1082/1197 [20:32<00:16,  7.03it/s]
Continual predict:  90%|█████████ | 1083/1197 [20:32<00:15,  7.20it/s]
Continual predict:  91%|█████████ | 1084/1197 [20:32<00:15,  7.32it/s]
Continual predict:  91%|█████████ | 1085/1197 [20:32<00:15,  7.39it/s]
Continual predict:  91%|█████████ | 1086/1197 [20:33<00:14,  7.45it/s]
Continual predict:  91%|█████████ | 1087/1197 [20:33<00:14,  7.50it/s]
Continual predict:  91%|█████████ | 1088/1197 [20:33<00:14,  7.53it/s]
Continual predict:  91%|█████████ | 1089/1197 [20:33<00:14,  7.54it/s]
Continual predict:  91%|█████████ | 1090/1197 [20:33<00:14,  7.47it/s]
Continual predict:  91%|█████████ | 1091/1197 [20:33<00:14,  7.49it/s]
Continual predict:  91%|█████████ | 1092/1197 [20:33<00:14,  7.50it/s]
Continual predict:  91%|█████████▏| 1093/1197 [20:34<00:13,  7.50it/s]
Continual predict:  91%|█████████▏| 1094/1197 [20:34<00:13,  7.54it/s]
Continual predict:  91%|█████████▏| 1095/1197 [20:34<00:13,  7.56it/s]
Continual predict:  92%|█████████▏| 1096/1197 [20:34<00:13,  7.58it/s]
Continual predict:  92%|█████████▏| 1097/1197 [20:34<00:13,  7.58it/s]
Continual predict:  92%|█████████▏| 1098/1197 [20:34<00:13,  7.59it/s]
Continual predict:  92%|█████████▏| 1099/1197 [20:34<00:12,  7.59it/s]
Continual predict:  92%|█████████▏| 1100/1197 [20:34<00:12,  7.61it/s]
Continual predict:  92%|█████████▏| 1101/1197 [20:35<00:12,  7.52it/s]
Continual predict:  92%|█████████▏| 1102/1197 [20:35<00:12,  7.54it/s]
Continual predict:  92%|█████████▏| 1103/1197 [20:35<00:14,  6.60it/s]
Continual predict:  92%|█████████▏| 1104/1197 [20:35<00:13,  6.80it/s]
Continual predict:  92%|█████████▏| 1105/1197 [20:35<00:14,  6.15it/s]
Continual predict:  92%|█████████▏| 1106/1197 [20:35<00:13,  6.51it/s]
Continual predict:  92%|█████████▏| 1107/1197 [20:36<00:15,  5.87it/s]
Continual predict:  93%|█████████▎| 1108/1197 [20:36<00:14,  6.19it/s]
Continual predict:  93%|█████████▎| 1109/1197 [20:36<00:13,  6.51it/s]
Continual predict:  93%|█████████▎| 1110/1197 [20:36<00:12,  6.78it/s]
Continual predict:  93%|█████████▎| 1111/1197 [20:36<00:14,  6.04it/s]
Continual predict:  93%|█████████▎| 1112/1197 [20:36<00:13,  6.44it/s]
Continual predict:  93%|█████████▎| 1113/1197 [20:36<00:12,  6.76it/s]
Continual predict:  93%|█████████▎| 1114/1197 [20:37<00:11,  6.93it/s]
Continual predict:  93%|█████████▎| 1115/1197 [20:37<00:11,  7.12it/s]
Continual predict:  93%|█████████▎| 1116/1197 [20:37<00:11,  7.27it/s]
Continual predict:  93%|█████████▎| 1117/1197 [20:37<00:10,  7.29it/s]
Continual predict:  93%|█████████▎| 1118/1197 [20:37<00:10,  7.31it/s]
Continual predict:  93%|█████████▎| 1119/1197 [20:37<00:12,  6.45it/s]
Continual predict:  94%|█████████▎| 1120/1197 [20:37<00:11,  6.75it/s]
Continual predict:  94%|█████████▎| 1121/1197 [20:38<00:10,  6.99it/s]
Continual predict:  94%|█████████▎| 1122/1197 [20:38<00:11,  6.29it/s]
Continual predict:  94%|█████████▍| 1123/1197 [20:38<00:11,  6.63it/s]
Continual predict:  94%|█████████▍| 1124/1197 [20:38<00:12,  6.08it/s]
Continual predict:  94%|█████████▍| 1125/1197 [20:38<00:11,  6.45it/s]
Continual predict:  94%|█████████▍| 1126/1197 [20:38<00:10,  6.76it/s]
Continual predict:  94%|█████████▍| 1127/1197 [20:39<00:11,  6.06it/s]
Continual predict:  94%|█████████▍| 1128/1197 [20:39<00:11,  5.91it/s]
Continual predict:  94%|█████████▍| 1129/1197 [20:39<00:10,  6.34it/s]
Continual predict:  94%|█████████▍| 1130/1197 [20:39<00:10,  6.67it/s]
Continual predict:  94%|█████████▍| 1131/1197 [20:39<00:09,  6.94it/s]
Continual predict:  95%|█████████▍| 1132/1197 [20:39<00:09,  7.10it/s]
Continual predict:  95%|█████████▍| 1133/1197 [20:39<00:08,  7.24it/s]
Continual predict:  95%|█████████▍| 1134/1197 [20:40<00:08,  7.36it/s]
Continual predict:  95%|█████████▍| 1135/1197 [20:40<00:09,  6.49it/s]
Continual predict:  95%|█████████▍| 1136/1197 [20:40<00:08,  6.79it/s]
Continual predict:  95%|█████████▍| 1137/1197 [20:40<00:08,  7.03it/s]
Continual predict:  95%|█████████▌| 1138/1197 [20:40<00:08,  7.21it/s]
Continual predict:  95%|█████████▌| 1139/1197 [20:40<00:08,  7.24it/s]
Continual predict:  95%|█████████▌| 1140/1197 [20:40<00:07,  7.32it/s]
Continual predict:  95%|█████████▌| 1141/1197 [20:41<00:07,  7.33it/s]
Continual predict:  95%|█████████▌| 1142/1197 [20:41<00:07,  7.35it/s]
Continual predict:  95%|█████████▌| 1143/1197 [20:41<00:07,  7.34it/s]
Continual predict:  96%|█████████▌| 1144/1197 [20:41<00:08,  6.39it/s]
Continual predict:  96%|█████████▌| 1145/1197 [20:41<00:07,  6.63it/s]
Continual predict:  96%|█████████▌| 1146/1197 [20:41<00:08,  5.98it/s]
Continual predict:  96%|█████████▌| 1147/1197 [20:42<00:07,  6.33it/s]
Continual predict:  96%|█████████▌| 1148/1197 [20:42<00:07,  6.67it/s]
Continual predict:  96%|█████████▌| 1149/1197 [20:42<00:06,  6.93it/s]
Continual predict:  96%|█████████▌| 1150/1197 [20:42<00:06,  7.05it/s]
Continual predict:  96%|█████████▌| 1151/1197 [20:42<00:06,  7.16it/s]
Continual predict:  96%|█████████▌| 1152/1197 [20:42<00:06,  6.44it/s]
Continual predict:  96%|█████████▋| 1153/1197 [20:43<00:08,  5.25it/s]
Continual predict:  96%|█████████▋| 1154/1197 [20:43<00:08,  5.14it/s]
Continual predict:  96%|█████████▋| 1155/1197 [20:43<00:07,  5.70it/s]
Continual predict:  97%|█████████▋| 1156/1197 [20:43<00:06,  6.16it/s]
Continual predict:  97%|█████████▋| 1157/1197 [20:43<00:06,  6.54it/s]
Continual predict:  97%|█████████▋| 1158/1197 [20:43<00:06,  5.94it/s]
Continual predict:  97%|█████████▋| 1159/1197 [20:43<00:06,  6.29it/s]
Continual predict:  97%|█████████▋| 1160/1197 [20:44<00:05,  6.63it/s]
Continual predict:  97%|█████████▋| 1161/1197 [20:44<00:05,  6.89it/s]
Continual predict:  97%|█████████▋| 1162/1197 [20:44<00:04,  7.08it/s]
Continual predict:  97%|█████████▋| 1163/1197 [20:44<00:04,  7.25it/s]
Continual predict:  97%|█████████▋| 1164/1197 [20:44<00:04,  7.37it/s]
Continual predict:  97%|█████████▋| 1165/1197 [20:44<00:04,  7.44it/s]
Continual predict:  97%|█████████▋| 1166/1197 [20:44<00:04,  6.44it/s]
Continual predict:  97%|█████████▋| 1167/1197 [20:45<00:04,  6.73it/s]
Continual predict:  98%|█████████▊| 1168/1197 [20:45<00:04,  6.97it/s]
Continual predict:  98%|█████████▊| 1169/1197 [20:45<00:03,  7.16it/s]
Continual predict:  98%|█████████▊| 1170/1197 [20:45<00:03,  7.28it/s]
Continual predict:  98%|█████████▊| 1171/1197 [20:45<00:04,  6.37it/s]
Continual predict:  98%|█████████▊| 1172/1197 [20:45<00:03,  6.63it/s]
Continual predict:  98%|█████████▊| 1173/1197 [20:45<00:03,  6.84it/s]
Continual predict:  98%|█████████▊| 1174/1197 [20:46<00:03,  7.03it/s]
Continual predict:  98%|█████████▊| 1175/1197 [20:46<00:03,  7.21it/s]
Continual predict:  98%|█████████▊| 1176/1197 [20:46<00:02,  7.31it/s]
Continual predict:  98%|█████████▊| 1177/1197 [20:46<00:03,  6.47it/s]
Continual predict:  98%|█████████▊| 1178/1197 [20:46<00:03,  5.97it/s]
Continual predict:  98%|█████████▊| 1179/1197 [20:46<00:02,  6.38it/s]
Continual predict:  99%|█████████▊| 1180/1197 [20:47<00:02,  6.71it/s]
Continual predict:  99%|█████████▊| 1181/1197 [20:47<00:02,  6.94it/s]
Continual predict:  99%|█████████▊| 1182/1197 [20:47<00:02,  7.11it/s]
Continual predict:  99%|█████████▉| 1183/1197 [20:47<00:01,  7.26it/s]
Continual predict:  99%|█████████▉| 1184/1197 [20:47<00:01,  7.38it/s]
Continual predict:  99%|█████████▉| 1185/1197 [20:47<00:01,  7.36it/s]
Continual predict:  99%|█████████▉| 1186/1197 [20:47<00:01,  6.39it/s]
Continual predict:  99%|█████████▉| 1187/1197 [20:48<00:01,  6.70it/s]
Continual predict:  99%|█████████▉| 1188/1197 [20:48<00:01,  6.10it/s]
Continual predict:  99%|█████████▉| 1189/1197 [20:48<00:01,  5.76it/s]
Continual predict:  99%|█████████▉| 1190/1197 [20:48<00:01,  6.17it/s]
Continual predict:  99%|█████████▉| 1191/1197 [20:48<00:00,  6.46it/s]
Continual predict: 100%|█████████▉| 1192/1197 [20:48<00:00,  6.71it/s]
Continual predict: 100%|█████████▉| 1193/1197 [20:48<00:00,  6.95it/s]
Continual predict: 100%|█████████▉| 1194/1197 [20:49<00:00,  7.16it/s]
Continual predict: 100%|█████████▉| 1195/1197 [20:49<00:00,  7.26it/s]
Continual predict: 100%|█████████▉| 1196/1197 [20:49<00:00,  6.44it/s]
Continual predict: 100%|██████████| 1197/1197 [20:49<00:00,  6.70it/s]
Continual predict: 100%|██████████| 1197/1197 [20:49<00:00,  1.04s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot10=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot12=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot9=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot14=20]│
+  │  GT=task1590_diplomacy_text_generati  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=16]│
+  │  GT=task1687_sentiment140_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot11=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot13=100]│
+  │  GT=task511_reddit_tifu_long_text_su  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task591_sciq_answer_generation    slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=18]│
+  │  GT=task748_glucose_reverse_cause_ev  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task875_emotion_classification    slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (1197/1197)  label_acc=100.0% (1197/1197)  disagree=  0                │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                                 =       15.0
+  predict_eval_rougeL                                                   =    61.3708
+  predict_eval_rougeL_for_CL                                            =    61.3708
+  predict_eval_rougeL_for_task002_quoref_answer_generation              =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                   =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction        =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                       =    41.3728
+  predict_eval_rougeL_for_task1590_diplomacy_text_generation            =    11.1335
+  predict_eval_rougeL_for_task1687_sentiment140_classification          =       84.0
+  predict_eval_rougeL_for_task1729_personachat_generate_next            =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                    =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification          =       91.0
+  predict_eval_rougeL_for_task511_reddit_tifu_long_text_summarization   =    21.3091
+  predict_eval_rougeL_for_task591_sciq_answer_generation                =    72.7381
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation   =    15.2912
+  predict_eval_rougeL_for_task748_glucose_reverse_cause_event_detection =    66.9039
+  predict_eval_rougeL_for_task875_emotion_classification                =       73.0
+  predict_exact_match                                                   =    47.3684
+  predict_exact_match_for_CL                                            =    47.3684
+  predict_exact_match_for_task002_quoref_answer_generation              =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation       =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                   =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction        =       98.0
+  predict_exact_match_for_task1572_samsum_summary                       =        0.0
+  predict_exact_match_for_task1590_diplomacy_text_generation            =        0.0
+  predict_exact_match_for_task1687_sentiment140_classification          =       84.0
+  predict_exact_match_for_task1729_personachat_generate_next            =        0.0
+  predict_exact_match_for_task181_outcome_extraction                    =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification          =       91.0
+  predict_exact_match_for_task511_reddit_tifu_long_text_summarization   =        1.0
+  predict_exact_match_for_task591_sciq_answer_generation                =       60.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation   =        0.0
+  predict_exact_match_for_task748_glucose_reverse_cause_event_detection =        3.0
+  predict_exact_match_for_task875_emotion_classification                =       73.0
+  predict_gen_len                                                       =      8.523
+  predict_global_step                                                   =        480
+  predict_rouge1                                                        =    62.8126
+  predict_rouge1_for_CL                                                 =    62.8126
+  predict_rouge1_for_task002_quoref_answer_generation                   =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation            =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                        =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction             =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                            =    50.3879
+  predict_rouge1_for_task1590_diplomacy_text_generation                 =    13.7971
+  predict_rouge1_for_task1687_sentiment140_classification               =       84.0
+  predict_rouge1_for_task1729_personachat_generate_next                 =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                         =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification               =       91.0
+  predict_rouge1_for_task511_reddit_tifu_long_text_summarization        =    25.5967
+  predict_rouge1_for_task591_sciq_answer_generation                     =    73.4048
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation        =    15.2912
+  predict_rouge1_for_task748_glucose_reverse_cause_event_detection      =     68.907
+  predict_rouge1_for_task875_emotion_classification                     =       73.0
+  predict_runtime                                                       = 0:20:49.54
+  predict_samples                                                       =       1197
+  predict_samples_per_second                                            =      0.958
+  predict_steps_per_second                                              =      0.958
+[2026-04-30 07:58:48,079] [INFO] [launch.py:367:main] Process 2076791 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/all_results.json b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..53bf8c9b8a195cc28bac409e9256daa17720501a
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/all_results.json
@@ -0,0 +1,27 @@
+{
+  "predict_exact_match": 75.8333,
+  "predict_rouge1": 84.2313,
+  "predict_eval_rougeL": 82.7288,
+  "predict_gen_len": 6.225,
+  "predict_runtime": 55.69839199515991,
+  "predict_samples": 120,
+  "predict_samples_per_second": 2.154,
+  "predict_steps_per_second": 2.154,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 75.8333,
+  "predict_rouge1_for_CL": 84.2313,
+  "predict_eval_rougeL_for_CL": 82.7288,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "train_runtime": 1512.249,
+  "train_samples_per_second": 9.919,
+  "train_steps_per_second": 0.317,
+  "total_flos": 5.460993820642509e+16,
+  "train_loss": 0.16749076483150324,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8294641be00922898565c40265b71b4e7d6a5623
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10ec2f6aaa981e910feaf66627161ae13b40169c542da1f7473ff0c62664b578
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..99357919080fceaa823dacb9565ca8dce45eae70
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f339760dcc45e9aee2f5f4dad2fccfa2e17b9cf9c22dec33da278ab273f9f960
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..a91126b2ca6df32784659d0fe62b625a01cbc16a
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94246c512940ebc63215a05bfc1daedd420ee9de3b1aece20184e0bdf5165048
+size 412018738
diff --git a/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/train.log b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..1d0eda07af9e3697e4835f40ce33a205e9b8e2a3
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/train.log
@@ -0,0 +1,324 @@
+[2026-04-29 19:34:58,332] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 19:34:58,332] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29501 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task363_sst2_polarity_classification --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification --cur_task_id 1 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[2026-04-29 19:35:01,645] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 19:35:01,646] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 19:35:01,646] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 19:35:01,646] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 19:35:01,646] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 19:35:01,646] [INFO] [launch.py:272:main] process 492775 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task363_sst2_polarity_classification', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification', '--cur_task_id', '1', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights']
+
+============================================================
+Task 1: task363_sst2_polarity_classification
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/train.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/dev.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+Train/dev/test: 1000/100/120
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   0%|          | 1/291 [00:00<00:34,  8.41it/s]
Loading weights:   9%|▉         | 27/291 [00:00<00:01, 140.49it/s]
Loading weights:  15%|█▌        | 44/291 [00:00<00:01, 153.13it/s]
Loading weights:  22%|██▏       | 63/291 [00:00<00:01, 161.39it/s]
Loading weights:  28%|██▊       | 82/291 [00:00<00:01, 169.83it/s]
Loading weights:  35%|███▌      | 103/291 [00:00<00:01, 169.42it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:01, 168.01it/s]
Loading weights:  49%|████▉     | 144/291 [00:00<00:00, 178.60it/s]
Loading weights:  56%|█████▋    | 164/291 [00:00<00:00, 184.20it/s]
Loading weights:  63%|██████▎   | 183/291 [00:01<00:00, 181.80it/s]
Loading weights:  69%|██████▉   | 202/291 [00:01<00:00, 170.42it/s]
Loading weights:  76%|███████▌  | 220/291 [00:01<00:00, 169.31it/s]
Loading weights:  82%|████████▏ | 239/291 [00:01<00:00, 173.99it/s]
Loading weights:  89%|████████▊ | 258/291 [00:01<00:00, 176.64it/s]
Loading weights:  96%|█████████▌| 279/291 [00:01<00:00, 174.56it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 171.74it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=291)
+[FIX] Re-initialized lora_A/B in 128 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(1, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 1 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6743.437M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 1 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 1 sigs, cur=task363_sst2_polarity_classification, mapping={'task363_sst2_polarity_classification': 0, 'task1572_samsum_summary': 1}
+[SGWI] Task 1: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 19:35:54,482] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:03<28:02,  3.51s/it]
  0%|          | 2/480 [00:06<26:14,  3.29s/it]
  1%|          | 3/480 [00:09<26:21,  3.32s/it]
  1%|          | 4/480 [00:13<25:49,  3.26s/it]
  1%|          | 5/480 [00:16<25:01,  3.16s/it]
  1%|▏         | 6/480 [00:19<25:13,  3.19s/it]
  1%|▏         | 7/480 [00:22<25:07,  3.19s/it]
  2%|▏         | 8/480 [00:25<25:06,  3.19s/it]
  2%|▏         | 9/480 [00:29<25:13,  3.21s/it]
  2%|▏         | 10/480 [00:32<25:05,  3.20s/it]
                                                
{'loss': '2.197', 'grad_norm': '0.4982', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:32<25:05,  3.20s/it]
  2%|▏         | 11/480 [00:35<25:03,  3.21s/it]
  2%|▎         | 12/480 [00:38<25:11,  3.23s/it]
  3%|▎         | 13/480 [00:41<25:06,  3.23s/it]
  3%|▎         | 14/480 [00:45<25:36,  3.30s/it]
  3%|▎         | 15/480 [00:48<24:51,  3.21s/it]
  3%|▎         | 16/480 [00:51<24:43,  3.20s/it]
  4%|▎         | 17/480 [00:54<25:00,  3.24s/it]
  4%|▍         | 18/480 [00:58<25:18,  3.29s/it]
  4%|▍         | 19/480 [01:01<24:58,  3.25s/it]
  4%|▍         | 20/480 [01:04<24:43,  3.23s/it]
                                                
{'loss': '0.7379', 'grad_norm': '0.3134', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:04<24:43,  3.23s/it]
  4%|▍         | 21/480 [01:07<24:51,  3.25s/it]
  5%|▍         | 22/480 [01:11<24:53,  3.26s/it]
  5%|▍         | 23/480 [01:14<24:14,  3.18s/it]
  5%|▌         | 24/480 [01:17<24:09,  3.18s/it]
  5%|▌         | 25/480 [01:20<23:41,  3.12s/it]
  5%|▌         | 26/480 [01:23<23:38,  3.12s/it]
  6%|▌         | 27/480 [01:26<23:40,  3.14s/it]
  6%|▌         | 28/480 [01:29<23:19,  3.10s/it]
  6%|▌         | 29/480 [01:32<23:42,  3.15s/it]
  6%|▋         | 30/480 [01:35<23:19,  3.11s/it]
                                                
{'loss': '0.1668', 'grad_norm': '0.797', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [01:35<23:19,  3.11s/it]
  6%|▋         | 31/480 [01:39<23:25,  3.13s/it]
  7%|▋         | 32/480 [01:39<18:04,  2.42s/it]
  7%|▋         | 33/480 [01:43<19:51,  2.67s/it]
  7%|▋         | 34/480 [01:46<20:37,  2.77s/it]
  7%|▋         | 35/480 [01:49<21:29,  2.90s/it]
  8%|▊         | 36/480 [01:52<22:07,  2.99s/it]
  8%|▊         | 37/480 [01:55<22:44,  3.08s/it]
  8%|▊         | 38/480 [01:59<22:55,  3.11s/it]
  8%|▊         | 39/480 [02:02<23:24,  3.18s/it]
  8%|▊         | 40/480 [02:05<23:10,  3.16s/it]
                                                
{'loss': '0.1485', 'grad_norm': '0.3333', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [02:05<23:10,  3.16s/it]
  9%|▊         | 41/480 [02:08<22:47,  3.11s/it]
  9%|▉         | 42/480 [02:11<22:36,  3.10s/it]
  9%|▉         | 43/480 [02:14<22:38,  3.11s/it]
  9%|▉         | 44/480 [02:17<22:23,  3.08s/it]
  9%|▉         | 45/480 [02:20<22:12,  3.06s/it]
 10%|▉         | 46/480 [02:23<22:05,  3.05s/it]
 10%|▉         | 47/480 [02:27<22:47,  3.16s/it]
 10%|█         | 48/480 [02:30<22:25,  3.11s/it]
 10%|█         | 49/480 [02:33<22:10,  3.09s/it]
 10%|█         | 50/480 [02:36<22:10,  3.10s/it]
                                                
{'loss': '0.3501', 'grad_norm': '0.05297', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [02:36<22:10,  3.10s/it]
 11%|█         | 51/480 [02:39<22:19,  3.12s/it]
 11%|█         | 52/480 [02:42<22:07,  3.10s/it]
 11%|█         | 53/480 [02:45<22:10,  3.12s/it]
 11%|█▏        | 54/480 [02:49<23:04,  3.25s/it]
 11%|█▏        | 55/480 [02:52<23:03,  3.26s/it]
 12%|█▏        | 56/480 [02:55<22:33,  3.19s/it]
 12%|█▏        | 57/480 [02:58<22:42,  3.22s/it]
 12%|█▏        | 58/480 [03:02<22:38,  3.22s/it]
 12%|█▏        | 59/480 [03:05<22:10,  3.16s/it]
 12%|█▎        | 60/480 [03:08<22:22,  3.20s/it]
                                                
{'loss': '0.1438', 'grad_norm': '0.03049', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [03:08<22:22,  3.20s/it]
 13%|█▎        | 61/480 [03:11<22:19,  3.20s/it]
 13%|█▎        | 62/480 [03:14<22:15,  3.19s/it]
 13%|█▎        | 63/480 [03:18<22:40,  3.26s/it]
 13%|█▎        | 64/480 [03:19<17:28,  2.52s/it]
 14%|█▎        | 65/480 [03:22<18:53,  2.73s/it]
 14%|█▍        | 66/480 [03:25<19:48,  2.87s/it]
 14%|█▍        | 67/480 [03:28<20:44,  3.01s/it]
 14%|█▍        | 68/480 [03:31<20:48,  3.03s/it]
 14%|█▍        | 69/480 [03:34<20:46,  3.03s/it]
 15%|█▍        | 70/480 [03:37<20:47,  3.04s/it]
                                                
{'loss': '0.1764', 'grad_norm': '0.02136', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [03:37<20:47,  3.04s/it]
 15%|█▍        | 71/480 [03:41<20:52,  3.06s/it]
 15%|█▌        | 72/480 [03:44<20:45,  3.05s/it]
 15%|█▌        | 73/480 [03:47<20:41,  3.05s/it]
 15%|█▌        | 74/480 [03:50<21:10,  3.13s/it]
 16%|█▌        | 75/480 [03:54<22:23,  3.32s/it]
 16%|█▌        | 76/480 [03:57<21:47,  3.24s/it]
 16%|█▌        | 77/480 [04:00<21:36,  3.22s/it]
 16%|█▋        | 78/480 [04:03<21:10,  3.16s/it]
 16%|█▋        | 79/480 [04:06<21:11,  3.17s/it]
 17%|█▋        | 80/480 [04:09<20:52,  3.13s/it]
                                                
{'loss': '0.1454', 'grad_norm': '0.01191', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [04:09<20:52,  3.13s/it]
 17%|█▋        | 81/480 [04:12<21:07,  3.18s/it]
 17%|█▋        | 82/480 [04:16<21:06,  3.18s/it]
 17%|█▋        | 83/480 [04:19<20:46,  3.14s/it]
 18%|█▊        | 84/480 [04:22<21:04,  3.19s/it]
 18%|█▊        | 85/480 [04:25<20:48,  3.16s/it]
 18%|█▊        | 86/480 [04:28<20:49,  3.17s/it]
 18%|█▊        | 87/480 [04:32<21:05,  3.22s/it]
 18%|█▊        | 88/480 [04:35<20:45,  3.18s/it]
 19%|█▊        | 89/480 [04:38<20:45,  3.18s/it]
 19%|█▉        | 90/480 [04:41<20:47,  3.20s/it]
                                                
{'loss': '0.1225', 'grad_norm': '0.02498', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [04:41<20:47,  3.20s/it]
 19%|█▉        | 91/480 [04:45<21:49,  3.37s/it]
 19%|█▉        | 92/480 [04:48<21:29,  3.32s/it]
 19%|█▉        | 93/480 [04:52<22:03,  3.42s/it]
 20%|█▉        | 94/480 [04:55<21:15,  3.31s/it]
 20%|█▉        | 95/480 [04:58<20:41,  3.22s/it]
 20%|██        | 96/480 [04:59<15:55,  2.49s/it]
 20%|██        | 97/480 [05:02<16:54,  2.65s/it]
 20%|██        | 98/480 [05:05<17:55,  2.82s/it]
 21%|██        | 99/480 [05:08<18:17,  2.88s/it]
 21%|██        | 100/480 [05:11<18:31,  2.93s/it]
                                                 
{'loss': '0.1044', 'grad_norm': '0.1795', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [05:11<18:31,  2.93s/it]
 21%|██        | 101/480 [05:14<18:49,  2.98s/it]
 21%|██▏       | 102/480 [05:17<19:11,  3.05s/it]
 21%|██▏       | 103/480 [05:20<19:08,  3.05s/it]
 22%|██▏       | 104/480 [05:23<19:09,  3.06s/it]
 22%|██▏       | 105/480 [05:26<19:07,  3.06s/it]
 22%|██▏       | 106/480 [05:30<19:19,  3.10s/it]
 22%|██▏       | 107/480 [05:33<19:08,  3.08s/it]
 22%|██▎       | 108/480 [05:36<19:10,  3.09s/it]
 23%|██▎       | 109/480 [05:39<19:19,  3.12s/it]
 23%|██▎       | 110/480 [05:42<19:05,  3.10s/it]
                                                 
{'loss': '0.1304', 'grad_norm': '1.241', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [05:42<19:05,  3.10s/it]
 23%|██▎       | 111/480 [05:45<19:28,  3.17s/it]
 23%|██▎       | 112/480 [05:48<19:10,  3.13s/it]
 24%|██▎       | 113/480 [05:52<19:48,  3.24s/it]
 24%|██▍       | 114/480 [05:55<19:28,  3.19s/it]
 24%|██▍       | 115/480 [05:58<19:33,  3.22s/it]
 24%|██▍       | 116/480 [06:01<19:28,  3.21s/it]
 24%|██▍       | 117/480 [06:05<19:24,  3.21s/it]
 25%|██▍       | 118/480 [06:08<19:31,  3.24s/it]
 25%|██▍       | 119/480 [06:11<19:26,  3.23s/it]
 25%|██▌       | 120/480 [06:14<19:20,  3.22s/it]
                                                 
{'loss': '0.1313', 'grad_norm': '0.2927', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [06:14<19:20,  3.22s/it]
 25%|██▌       | 121/480 [06:18<19:41,  3.29s/it]
 25%|██▌       | 122/480 [06:21<19:40,  3.30s/it]
 26%|██▌       | 123/480 [06:24<19:08,  3.22s/it]
 26%|██▌       | 124/480 [06:28<19:26,  3.28s/it]
 26%|██▌       | 125/480 [06:31<18:58,  3.21s/it]
 26%|██▋       | 126/480 [06:34<18:36,  3.15s/it]
 26%|██▋       | 127/480 [06:37<18:21,  3.12s/it]
 27%|██▋       | 128/480 [06:38<14:29,  2.47s/it]
 27%|██▋       | 129/480 [06:41<15:49,  2.70s/it]
 27%|██▋       | 130/480 [06:44<16:21,  2.80s/it]
                                                 
{'loss': '0.125', 'grad_norm': '1.133', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [06:44<16:21,  2.80s/it]
 27%|██▋       | 131/480 [06:47<16:42,  2.87s/it]
 28%|██▊       | 132/480 [06:50<17:14,  2.97s/it]
 28%|██▊       | 133/480 [06:53<17:30,  3.03s/it]
 28%|██▊       | 134/480 [06:57<17:48,  3.09s/it]
 28%|██▊       | 135/480 [07:00<17:52,  3.11s/it]
 28%|██▊       | 136/480 [07:03<17:41,  3.09s/it]
 29%|██▊       | 137/480 [07:06<17:49,  3.12s/it]
 29%|██▉       | 138/480 [07:09<17:54,  3.14s/it]
 29%|██▉       | 139/480 [07:12<17:58,  3.16s/it]
 29%|██▉       | 140/480 [07:15<17:53,  3.16s/it]
                                                 
{'loss': '0.104', 'grad_norm': '0.01027', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [07:15<17:53,  3.16s/it]
 29%|██▉       | 141/480 [07:19<18:52,  3.34s/it]
 30%|██▉       | 142/480 [07:23<18:50,  3.35s/it]
 30%|██▉       | 143/480 [07:26<18:32,  3.30s/it]
 30%|███       | 144/480 [07:29<18:03,  3.23s/it]
 30%|███       | 145/480 [07:32<18:07,  3.25s/it]
 30%|███       | 146/480 [07:36<18:27,  3.31s/it]
 31%|███       | 147/480 [07:39<19:09,  3.45s/it]
 31%|███       | 148/480 [07:43<18:57,  3.43s/it]
 31%|███       | 149/480 [07:46<18:15,  3.31s/it]
 31%|███▏      | 150/480 [07:49<17:54,  3.26s/it]
                                                 
{'loss': '0.09225', 'grad_norm': '0.06465', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [07:49<17:54,  3.26s/it]
 31%|███▏      | 151/480 [07:52<18:05,  3.30s/it]
 32%|███▏      | 152/480 [07:55<17:36,  3.22s/it]
 32%|███▏      | 153/480 [07:58<17:14,  3.16s/it]
 32%|███▏      | 154/480 [08:02<17:21,  3.20s/it]
 32%|███▏      | 155/480 [08:05<17:03,  3.15s/it]
 32%|███▎      | 156/480 [08:08<17:04,  3.16s/it]
 33%|███▎      | 157/480 [08:11<16:48,  3.12s/it]
 33%|███▎      | 158/480 [08:14<16:44,  3.12s/it]
 33%|███▎      | 159/480 [08:17<16:33,  3.09s/it]
 33%|███▎      | 160/480 [08:18<12:50,  2.41s/it]
                                                 
{'loss': '0.1288', 'grad_norm': '0.763', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [08:18<12:50,  2.41s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   2%|▏         | 2/100 [00:00<00:07, 12.61it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:00<00:07, 13.15it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:00<00:06, 13.67it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:00<00:06, 13.94it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:00<00:06, 13.89it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:00<00:06, 14.05it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:01<00:06, 13.70it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:01<00:06, 13.73it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:01<00:05, 13.70it/s]
+
Eval @ step 160:  20%|██        | 20/100 [00:01<00:05, 13.91it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:01<00:05, 13.87it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:01<00:05, 13.67it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:01<00:05, 13.87it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:02<00:05, 13.70it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:02<00:05, 13.72it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:02<00:04, 13.77it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:02<00:04, 13.77it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:02<00:04, 13.76it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:02<00:04, 13.94it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:02<00:04, 13.85it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:03<00:04, 14.00it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:03<00:04, 13.93it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:03<00:03, 14.05it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:03<00:03, 14.15it/s]
+
Eval @ step 160:  50%|█████     | 50/100 [00:03<00:03, 14.23it/s]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:03<00:03, 14.09it/s]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:03<00:03, 13.97it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:04<00:03, 13.99it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:04<00:02, 14.11it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:04<00:02, 14.19it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [00:04<00:02, 14.21it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [00:04<00:02, 13.96it/s]
+
Eval @ step 160:  66%|██████▌   | 66/100 [00:04<00:02, 13.97it/s]
+
Eval @ step 160:  68%|██████▊   | 68/100 [00:04<00:02, 13.52it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [00:05<00:02, 13.63it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [00:05<00:02, 13.78it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [00:05<00:01, 13.61it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:05<00:01, 13.80it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:05<00:01, 13.80it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:05<00:01, 13.63it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [00:05<00:01, 13.85it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [00:06<00:01, 14.00it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [00:06<00:01, 13.95it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [00:06<00:00, 14.09it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [00:06<00:00, 14.15it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [00:06<00:00, 14.20it/s]
+
Eval @ step 160:  94%|█████████▍| 94/100 [00:06<00:00, 14.24it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [00:06<00:00, 14.10it/s]
+
Eval @ step 160:  98%|█████████▊| 98/100 [00:07<00:00, 13.83it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [00:07<00:00, 14.03it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=86.0000 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '86', 'epoch': '5'}
+
 33%|███▎      | 160/480 [08:25<12:50,  2.41s/it]
 34%|███▎      | 161/480 [08:28<25:33,  4.81s/it]
 34%|███▍      | 162/480 [08:32<23:11,  4.38s/it]
 34%|███▍      | 163/480 [08:35<21:28,  4.06s/it]
 34%|███▍      | 164/480 [08:38<19:45,  3.75s/it]
 34%|███▍      | 165/480 [08:41<18:34,  3.54s/it]
 35%|███▍      | 166/480 [08:44<17:51,  3.41s/it]
 35%|███▍      | 167/480 [08:47<17:17,  3.31s/it]
 35%|███▌      | 168/480 [08:50<16:46,  3.23s/it]
 35%|███▌      | 169/480 [08:53<16:29,  3.18s/it]
 35%|███▌      | 170/480 [08:57<16:23,  3.17s/it]
                                                 
{'loss': '0.1414', 'grad_norm': '0.3912', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [08:57<16:23,  3.17s/it]
 36%|███▌      | 171/480 [09:00<16:22,  3.18s/it]
 36%|███▌      | 172/480 [09:03<16:04,  3.13s/it]
 36%|███▌      | 173/480 [09:06<15:58,  3.12s/it]
 36%|███▋      | 174/480 [09:09<15:53,  3.12s/it]
 36%|███▋      | 175/480 [09:12<15:56,  3.14s/it]
 37%|███▋      | 176/480 [09:15<15:47,  3.12s/it]
 37%|███▋      | 177/480 [09:18<15:48,  3.13s/it]
 37%|███▋      | 178/480 [09:22<15:51,  3.15s/it]
 37%|███▋      | 179/480 [09:25<15:37,  3.11s/it]
 38%|███▊      | 180/480 [09:28<15:51,  3.17s/it]
                                                 
{'loss': '0.1394', 'grad_norm': '0.01334', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [09:28<15:51,  3.17s/it]
 38%|███▊      | 181/480 [09:31<15:37,  3.14s/it]
 38%|███▊      | 182/480 [09:34<15:39,  3.15s/it]
 38%|███▊      | 183/480 [09:37<15:34,  3.15s/it]
 38%|███▊      | 184/480 [09:40<15:38,  3.17s/it]
 39%|███▊      | 185/480 [09:44<15:37,  3.18s/it]
 39%|███▉      | 186/480 [09:47<15:39,  3.19s/it]
 39%|███▉      | 187/480 [09:50<15:46,  3.23s/it]
 39%|███▉      | 188/480 [09:53<15:42,  3.23s/it]
 39%|███▉      | 189/480 [09:57<15:40,  3.23s/it]
 40%|███▉      | 190/480 [10:00<15:40,  3.24s/it]
                                                 
{'loss': '0.1035', 'grad_norm': '0.08601', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [10:00<15:40,  3.24s/it]
 40%|███▉      | 191/480 [10:03<15:18,  3.18s/it]
 40%|████      | 192/480 [10:04<11:47,  2.46s/it]
 40%|████      | 193/480 [10:07<12:33,  2.63s/it]
 40%|████      | 194/480 [10:10<13:16,  2.79s/it]
 41%|████      | 195/480 [10:13<13:58,  2.94s/it]
 41%|████      | 196/480 [10:16<14:19,  3.03s/it]
 41%|████      | 197/480 [10:20<14:27,  3.07s/it]
 41%|████▏     | 198/480 [10:23<14:41,  3.13s/it]
 41%|████▏     | 199/480 [10:26<14:29,  3.10s/it]
 42%|████▏     | 200/480 [10:29<14:22,  3.08s/it]
                                                 
{'loss': '0.1037', 'grad_norm': '0.1243', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [10:29<14:22,  3.08s/it]
 42%|████▏     | 201/480 [10:32<14:24,  3.10s/it]
 42%|████▏     | 202/480 [10:35<14:15,  3.08s/it]
 42%|████▏     | 203/480 [10:38<14:23,  3.12s/it]
 42%|████▎     | 204/480 [10:42<14:26,  3.14s/it]
 43%|████▎     | 205/480 [10:45<14:15,  3.11s/it]
 43%|████▎     | 206/480 [10:48<14:19,  3.14s/it]
 43%|████▎     | 207/480 [10:51<14:17,  3.14s/it]
 43%|████▎     | 208/480 [10:54<14:21,  3.17s/it]
 44%|████▎     | 209/480 [10:57<14:11,  3.14s/it]
 44%|████▍     | 210/480 [11:00<14:13,  3.16s/it]
                                                 
{'loss': '0.09386', 'grad_norm': '0.06139', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [11:00<14:13,  3.16s/it]
 44%|████▍     | 211/480 [11:04<14:21,  3.20s/it]
 44%|████▍     | 212/480 [11:07<14:18,  3.20s/it]
 44%|████▍     | 213/480 [11:10<14:02,  3.16s/it]
 45%|████▍     | 214/480 [11:13<14:17,  3.22s/it]
 45%|████▍     | 215/480 [11:17<14:14,  3.22s/it]
 45%|████▌     | 216/480 [11:20<14:23,  3.27s/it]
 45%|████▌     | 217/480 [11:23<14:30,  3.31s/it]
 45%|████▌     | 218/480 [11:26<14:05,  3.23s/it]
 46%|████▌     | 219/480 [11:29<13:47,  3.17s/it]
 46%|████▌     | 220/480 [11:33<13:54,  3.21s/it]
                                                 
{'loss': '0.09571', 'grad_norm': '0.1055', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [11:33<13:54,  3.21s/it]
 46%|████▌     | 221/480 [11:36<13:38,  3.16s/it]
 46%|████▋     | 222/480 [11:39<13:25,  3.12s/it]
 46%|████▋     | 223/480 [11:42<13:20,  3.11s/it]
 47%|████▋     | 224/480 [11:43<10:17,  2.41s/it]
 47%|████▋     | 225/480 [11:46<11:06,  2.61s/it]
 47%|████▋     | 226/480 [11:49<11:49,  2.79s/it]
 47%|████▋     | 227/480 [11:52<12:06,  2.87s/it]
 48%|████▊     | 228/480 [11:55<12:23,  2.95s/it]
 48%|████▊     | 229/480 [11:58<12:31,  3.00s/it]
 48%|████▊     | 230/480 [12:01<12:44,  3.06s/it]
                                                 
{'loss': '0.082', 'grad_norm': '0.249', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [12:01<12:44,  3.06s/it]
 48%|████▊     | 231/480 [12:05<12:51,  3.10s/it]
 48%|████▊     | 232/480 [12:08<12:45,  3.09s/it]
 49%|████▊     | 233/480 [12:11<12:51,  3.12s/it]
 49%|████▉     | 234/480 [12:14<12:42,  3.10s/it]
 49%|████▉     | 235/480 [12:17<13:01,  3.19s/it]
 49%|████▉     | 236/480 [12:20<12:47,  3.15s/it]
 49%|████▉     | 237/480 [12:24<12:48,  3.16s/it]
 50%|████▉     | 238/480 [12:27<12:56,  3.21s/it]
 50%|████▉     | 239/480 [12:30<12:42,  3.16s/it]
 50%|█████     | 240/480 [12:33<12:42,  3.18s/it]
                                                 
{'loss': '0.1237', 'grad_norm': '0.163', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [12:33<12:42,  3.18s/it]
 50%|█████     | 241/480 [12:37<12:56,  3.25s/it]
 50%|█████     | 242/480 [12:40<12:50,  3.24s/it]
 51%|█████     | 243/480 [12:43<12:45,  3.23s/it]
 51%|█████     | 244/480 [12:46<12:53,  3.28s/it]
 51%|█████     | 245/480 [12:50<12:45,  3.26s/it]
 51%|█████▏    | 246/480 [12:53<12:25,  3.19s/it]
 51%|█████▏    | 247/480 [12:56<12:12,  3.15s/it]
 52%|█████▏    | 248/480 [12:59<12:29,  3.23s/it]
 52%|█████▏    | 249/480 [13:03<12:40,  3.29s/it]
 52%|█████▏    | 250/480 [13:06<12:27,  3.25s/it]
                                                 
{'loss': '0.09795', 'grad_norm': '0.2716', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [13:06<12:27,  3.25s/it]
 52%|█████▏    | 251/480 [13:09<12:18,  3.23s/it]
 52%|█████▎    | 252/480 [13:12<12:13,  3.22s/it]
 53%|█████▎    | 253/480 [13:15<11:59,  3.17s/it]
 53%|█████▎    | 254/480 [13:18<11:55,  3.16s/it]
 53%|█████▎    | 255/480 [13:21<11:43,  3.13s/it]
 53%|█████▎    | 256/480 [13:22<09:01,  2.42s/it]
 54%|█████▎    | 257/480 [13:25<09:51,  2.65s/it]
 54%|█████▍    | 258/480 [13:28<10:15,  2.77s/it]
 54%|█████▍    | 259/480 [13:31<10:33,  2.87s/it]
 54%|█████▍    | 260/480 [13:35<10:52,  2.97s/it]
                                                 
{'loss': '0.09785', 'grad_norm': '0.2389', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [13:35<10:52,  2.97s/it]
 54%|█████▍    | 261/480 [13:38<10:54,  2.99s/it]
 55%|█████▍    | 262/480 [13:41<11:03,  3.04s/it]
 55%|█████▍    | 263/480 [13:44<11:10,  3.09s/it]
 55%|█████▌    | 264/480 [13:47<11:14,  3.12s/it]
 55%|█████▌    | 265/480 [13:51<11:18,  3.16s/it]
 55%|█████▌    | 266/480 [13:54<11:18,  3.17s/it]
 56%|█████▌    | 267/480 [13:57<11:18,  3.19s/it]
 56%|█████▌    | 268/480 [14:00<11:31,  3.26s/it]
 56%|█████▌    | 269/480 [14:04<11:29,  3.27s/it]
 56%|█████▋    | 270/480 [14:07<11:24,  3.26s/it]
                                                 
{'loss': '0.1023', 'grad_norm': '0.007243', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [14:07<11:24,  3.26s/it]
 56%|█████▋    | 271/480 [14:10<11:20,  3.26s/it]
 57%|█████▋    | 272/480 [14:14<11:26,  3.30s/it]
 57%|█████▋    | 273/480 [14:17<11:07,  3.22s/it]
 57%|█████▋    | 274/480 [14:20<10:52,  3.17s/it]
 57%|█████▋    | 275/480 [14:23<10:50,  3.17s/it]
 57%|█████▊    | 276/480 [14:26<10:38,  3.13s/it]
 58%|█████▊    | 277/480 [14:29<10:29,  3.10s/it]
 58%|█████▊    | 278/480 [14:32<10:27,  3.11s/it]
 58%|█████▊    | 279/480 [14:35<10:20,  3.09s/it]
 58%|█████▊    | 280/480 [14:38<10:25,  3.13s/it]
                                                 
{'loss': '0.08775', 'grad_norm': '0.05645', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [14:38<10:25,  3.13s/it]
 59%|█████▊    | 281/480 [14:41<10:16,  3.10s/it]
 59%|█████▉    | 282/480 [14:45<10:22,  3.14s/it]
 59%|█████▉    | 283/480 [14:48<10:12,  3.11s/it]
 59%|█████▉    | 284/480 [14:51<10:15,  3.14s/it]
 59%|█████▉    | 285/480 [14:54<10:11,  3.13s/it]
 60%|█████▉    | 286/480 [14:57<10:03,  3.11s/it]
 60%|█████▉    | 287/480 [15:00<10:08,  3.15s/it]
 60%|██████    | 288/480 [15:01<07:48,  2.44s/it]
 60%|██████    | 289/480 [15:04<08:22,  2.63s/it]
 60%|██████    | 290/480 [15:07<08:53,  2.81s/it]
                                                 
{'loss': '0.07292', 'grad_norm': '1.81', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [15:07<08:53,  2.81s/it]
 61%|██████    | 291/480 [15:10<09:03,  2.88s/it]
 61%|██████    | 292/480 [15:14<09:25,  3.01s/it]
 61%|██████    | 293/480 [15:17<09:40,  3.11s/it]
 61%|██████▏   | 294/480 [15:20<09:36,  3.10s/it]
 61%|██████▏   | 295/480 [15:23<09:48,  3.18s/it]
 62%|██████▏   | 296/480 [15:27<09:55,  3.24s/it]
 62%|██████▏   | 297/480 [15:30<09:51,  3.23s/it]
 62%|██████▏   | 298/480 [15:33<09:46,  3.22s/it]
 62%|██████▏   | 299/480 [15:37<10:02,  3.33s/it]
 62%|██████▎   | 300/480 [15:40<09:42,  3.24s/it]
                                                 
{'loss': '0.08683', 'grad_norm': '0.3861', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [15:40<09:42,  3.24s/it]
 63%|██████▎   | 301/480 [15:43<09:28,  3.18s/it]
 63%|██████▎   | 302/480 [15:46<09:24,  3.17s/it]
 63%|██████▎   | 303/480 [15:49<09:13,  3.13s/it]
 63%|██████▎   | 304/480 [15:52<09:05,  3.10s/it]
 64%|██████▎   | 305/480 [15:55<08:58,  3.08s/it]
 64%|██████▍   | 306/480 [15:58<09:03,  3.12s/it]
 64%|██████▍   | 307/480 [16:02<09:07,  3.16s/it]
 64%|██████▍   | 308/480 [16:05<08:57,  3.12s/it]
 64%|██████▍   | 309/480 [16:08<08:54,  3.12s/it]
 65%|██████▍   | 310/480 [16:11<08:46,  3.10s/it]
                                                 
{'loss': '0.08248', 'grad_norm': '0.2285', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [16:11<08:46,  3.10s/it]
 65%|██████▍   | 311/480 [16:14<08:48,  3.13s/it]
 65%|██████▌   | 312/480 [16:17<08:41,  3.10s/it]
 65%|██████▌   | 313/480 [16:20<08:44,  3.14s/it]
 65%|██████▌   | 314/480 [16:23<08:44,  3.16s/it]
 66%|██████▌   | 315/480 [16:26<08:34,  3.12s/it]
 66%|██████▌   | 316/480 [16:30<08:34,  3.14s/it]
 66%|██████▌   | 317/480 [16:33<08:39,  3.19s/it]
 66%|██████▋   | 318/480 [16:36<08:36,  3.19s/it]
 66%|██████▋   | 319/480 [16:39<08:29,  3.17s/it]
 67%|██████▋   | 320/480 [16:40<06:31,  2.45s/it]
                                                 
{'loss': '0.1142', 'grad_norm': '0.9874', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [16:40<06:31,  2.45s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:07, 12.60it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:00<00:07, 13.16it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:00<00:06, 13.60it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:00<00:06, 13.87it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:00<00:06, 13.82it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:00<00:06, 13.96it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:01<00:06, 13.63it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:01<00:06, 13.65it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:01<00:06, 12.84it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:01<00:06, 13.29it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:01<00:05, 13.42it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:01<00:05, 13.37it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:01<00:05, 13.65it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:02<00:05, 13.54it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:02<00:05, 13.60it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:02<00:04, 13.70it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:02<00:04, 13.71it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:02<00:04, 13.71it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:02<00:04, 13.90it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:02<00:04, 13.81it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:03<00:04, 13.98it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:03<00:04, 13.92it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:03<00:03, 14.03it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:03<00:03, 14.15it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:03<00:03, 14.21it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:03<00:03, 14.07it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:03<00:03, 13.97it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:04<00:03, 13.98it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:04<00:02, 14.11it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:04<00:02, 14.19it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:04<00:02, 14.24it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:04<00:02, 14.07it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:04<00:02, 14.13it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:04<00:02, 14.14it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:05<00:02, 14.24it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:05<00:01, 14.21it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:05<00:01, 13.88it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:05<00:01, 14.00it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:05<00:01, 13.93it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:05<00:01, 13.70it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:05<00:01, 13.91it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:06<00:01, 14.05it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:06<00:01, 13.98it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:06<00:00, 14.11it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:06<00:00, 14.18it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:06<00:00, 14.23it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:06<00:00, 14.29it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:06<00:00, 14.24it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:07<00:00, 14.01it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:07<00:00, 14.22it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=91.0000 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '91', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [16:47<06:31,  2.45s/it]
 67%|██████▋   | 321/480 [16:51<12:58,  4.89s/it]
 67%|██████▋   | 322/480 [16:54<11:24,  4.33s/it]
 67%|██████▋   | 323/480 [16:57<10:20,  3.95s/it]
 68%|██████▊   | 324/480 [17:00<09:35,  3.69s/it]
 68%|██████▊   | 325/480 [17:03<09:20,  3.62s/it]
 68%|██████▊   | 326/480 [17:06<08:49,  3.44s/it]
 68%|██████▊   | 327/480 [17:10<08:40,  3.40s/it]
 68%|██████▊   | 328/480 [17:13<08:46,  3.46s/it]
 69%|██████▊   | 329/480 [17:16<08:35,  3.42s/it]
 69%|██████▉   | 330/480 [17:20<08:14,  3.30s/it]
                                                 
{'loss': '0.09567', 'grad_norm': '0.1212', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [17:20<08:14,  3.30s/it]
 69%|██████▉   | 331/480 [17:23<08:06,  3.27s/it]
 69%|██████▉   | 332/480 [17:26<07:53,  3.20s/it]
 69%|██████▉   | 333/480 [17:29<07:49,  3.20s/it]
 70%|██████▉   | 334/480 [17:32<07:43,  3.18s/it]
 70%|██████▉   | 335/480 [17:35<07:33,  3.13s/it]
 70%|███████   | 336/480 [17:38<07:34,  3.16s/it]
 70%|███████   | 337/480 [17:41<07:26,  3.12s/it]
 70%|███████   | 338/480 [17:45<07:35,  3.21s/it]
 71%|███████   | 339/480 [17:48<07:25,  3.16s/it]
 71%|███████   | 340/480 [17:51<07:25,  3.18s/it]
                                                 
{'loss': '0.1098', 'grad_norm': '0.01813', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [17:51<07:25,  3.18s/it]
 71%|███████   | 341/480 [17:54<07:30,  3.24s/it]
 71%|███████▏  | 342/480 [17:58<07:27,  3.24s/it]
 71%|███████▏  | 343/480 [18:01<07:15,  3.18s/it]
 72%|███████▏  | 344/480 [18:04<07:21,  3.25s/it]
 72%|███████▏  | 345/480 [18:07<07:22,  3.28s/it]
 72%|███████▏  | 346/480 [18:11<07:33,  3.38s/it]
 72%|███████▏  | 347/480 [18:15<07:36,  3.44s/it]
 72%|███████▎  | 348/480 [18:18<07:17,  3.32s/it]
 73%|███████▎  | 349/480 [18:21<07:03,  3.23s/it]
 73%|███████▎  | 350/480 [18:24<06:55,  3.20s/it]
                                                 
{'loss': '0.08322', 'grad_norm': '0.01171', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [18:24<06:55,  3.20s/it]
 73%|███████▎  | 351/480 [18:27<06:46,  3.15s/it]
 73%|███████▎  | 352/480 [18:28<05:12,  2.44s/it]
 74%|███████▎  | 353/480 [18:31<05:32,  2.62s/it]
 74%|███████▍  | 354/480 [18:34<05:46,  2.75s/it]
 74%|███████▍  | 355/480 [18:37<05:59,  2.87s/it]
 74%|███████▍  | 356/480 [18:40<06:08,  2.97s/it]
 74%|███████▍  | 357/480 [18:43<06:08,  3.00s/it]
 75%|███████▍  | 358/480 [18:46<06:10,  3.04s/it]
 75%|███████▍  | 359/480 [18:49<06:07,  3.04s/it]
 75%|███████▌  | 360/480 [18:53<06:16,  3.14s/it]
                                                 
{'loss': '0.08614', 'grad_norm': '1.305', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [18:53<06:16,  3.14s/it]
 75%|███████▌  | 361/480 [18:56<06:15,  3.16s/it]
 75%|███████▌  | 362/480 [18:59<06:08,  3.12s/it]
 76%|███████▌  | 363/480 [19:02<06:09,  3.16s/it]
 76%|███████▌  | 364/480 [19:05<06:04,  3.15s/it]
 76%|███████▌  | 365/480 [19:09<06:07,  3.20s/it]
 76%|███████▋  | 366/480 [19:12<06:00,  3.17s/it]
 76%|███████▋  | 367/480 [19:15<05:58,  3.17s/it]
 77%|███████▋  | 368/480 [19:18<06:02,  3.24s/it]
 77%|███████▋  | 369/480 [19:21<05:53,  3.19s/it]
 77%|███████▋  | 370/480 [19:25<05:50,  3.19s/it]
                                                 
{'loss': '0.08452', 'grad_norm': '1.262', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [19:25<05:50,  3.19s/it]
 77%|███████▋  | 371/480 [19:28<05:50,  3.22s/it]
 78%|███████▊  | 372/480 [19:31<05:47,  3.22s/it]
 78%|███████▊  | 373/480 [19:34<05:43,  3.21s/it]
 78%|███████▊  | 374/480 [19:38<05:45,  3.26s/it]
 78%|███████▊  | 375/480 [19:41<05:42,  3.26s/it]
 78%|███████▊  | 376/480 [19:44<05:31,  3.19s/it]
 79%|███████▊  | 377/480 [19:47<05:31,  3.21s/it]
 79%|███████▉  | 378/480 [19:50<05:26,  3.20s/it]
 79%|███████▉  | 379/480 [19:53<05:18,  3.15s/it]
 79%|███████▉  | 380/480 [19:57<05:16,  3.17s/it]
                                                 
{'loss': '0.08566', 'grad_norm': '0.07563', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [19:57<05:16,  3.17s/it]
 79%|███████▉  | 381/480 [20:00<05:13,  3.17s/it]
 80%|███████▉  | 382/480 [20:03<05:12,  3.19s/it]
 80%|███████▉  | 383/480 [20:06<05:10,  3.20s/it]
 80%|████████  | 384/480 [20:07<03:57,  2.48s/it]
 80%|████████  | 385/480 [20:10<04:15,  2.69s/it]
 80%|████████  | 386/480 [20:13<04:29,  2.87s/it]
 81%|████████  | 387/480 [20:17<04:38,  2.99s/it]
 81%|████████  | 388/480 [20:20<04:40,  3.04s/it]
 81%|████████  | 389/480 [20:23<04:42,  3.10s/it]
 81%|████████▏ | 390/480 [20:27<04:46,  3.19s/it]
                                                 
{'loss': '0.07954', 'grad_norm': '0.03071', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [20:27<04:46,  3.19s/it]
 81%|████████▏ | 391/480 [20:30<04:52,  3.29s/it]
 82%|████████▏ | 392/480 [20:33<04:45,  3.24s/it]
 82%|████████▏ | 393/480 [20:36<04:41,  3.24s/it]
 82%|████████▏ | 394/480 [20:40<04:37,  3.23s/it]
 82%|████████▏ | 395/480 [20:43<04:31,  3.20s/it]
 82%|████████▎ | 396/480 [20:46<04:29,  3.21s/it]
 83%|████████▎ | 397/480 [20:49<04:26,  3.21s/it]
 83%|████████▎ | 398/480 [20:52<04:24,  3.23s/it]
 83%|████████▎ | 399/480 [20:56<04:20,  3.22s/it]
 83%|████████▎ | 400/480 [20:59<04:17,  3.22s/it]
                                                 
{'loss': '0.08848', 'grad_norm': '0.01373', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [20:59<04:17,  3.22s/it]
 84%|████████▎ | 401/480 [21:02<04:15,  3.23s/it]
 84%|████████▍ | 402/480 [21:05<04:12,  3.23s/it]
 84%|████████▍ | 403/480 [21:09<04:05,  3.19s/it]
 84%|████████▍ | 404/480 [21:12<04:00,  3.16s/it]
 84%|████████▍ | 405/480 [21:15<03:56,  3.15s/it]
 85%|████████▍ | 406/480 [21:18<03:50,  3.11s/it]
 85%|████████▍ | 407/480 [21:21<03:45,  3.09s/it]
 85%|████████▌ | 408/480 [21:24<03:43,  3.10s/it]
 85%|████████▌ | 409/480 [21:27<03:38,  3.08s/it]
 85%|████████▌ | 410/480 [21:30<03:38,  3.12s/it]
                                                 
{'loss': '0.07916', 'grad_norm': '0.5014', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [21:30<03:38,  3.12s/it]
 86%|████████▌ | 411/480 [21:33<03:33,  3.09s/it]
 86%|████████▌ | 412/480 [21:36<03:30,  3.09s/it]
 86%|████████▌ | 413/480 [21:39<03:26,  3.08s/it]
 86%|████████▋ | 414/480 [21:42<03:25,  3.12s/it]
 86%|████████▋ | 415/480 [21:46<03:23,  3.14s/it]
 87%|████████▋ | 416/480 [21:46<02:35,  2.43s/it]
 87%|████████▋ | 417/480 [21:49<02:44,  2.61s/it]
 87%|████████▋ | 418/480 [21:53<02:52,  2.79s/it]
 87%|████████▋ | 419/480 [21:56<02:56,  2.89s/it]
 88%|████████▊ | 420/480 [21:59<02:58,  2.98s/it]
                                                 
{'loss': '0.0744', 'grad_norm': '0.01786', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [21:59<02:58,  2.98s/it]
 88%|████████▊ | 421/480 [22:02<02:56,  3.00s/it]
 88%|████████▊ | 422/480 [22:05<02:59,  3.10s/it]
 88%|████████▊ | 423/480 [22:09<02:58,  3.14s/it]
 88%|████████▊ | 424/480 [22:12<02:53,  3.11s/it]
 89%|████████▊ | 425/480 [22:15<02:52,  3.14s/it]
 89%|████████▉ | 426/480 [22:18<02:52,  3.20s/it]
 89%|████████▉ | 427/480 [22:21<02:49,  3.20s/it]
 89%|████████▉ | 428/480 [22:25<02:46,  3.20s/it]
 89%|████████▉ | 429/480 [22:28<02:48,  3.30s/it]
 90%|████████▉ | 430/480 [22:31<02:40,  3.22s/it]
                                                 
{'loss': '0.08216', 'grad_norm': '0.07298', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [22:31<02:40,  3.22s/it]
 90%|████████▉ | 431/480 [22:34<02:35,  3.17s/it]
 90%|█████████ | 432/480 [22:37<02:32,  3.17s/it]
 90%|█████████ | 433/480 [22:40<02:27,  3.13s/it]
 90%|█████████ | 434/480 [22:43<02:22,  3.11s/it]
 91%|█████████ | 435/480 [22:47<02:18,  3.08s/it]
 91%|█████████ | 436/480 [22:50<02:15,  3.09s/it]
 91%|█████████ | 437/480 [22:53<02:16,  3.17s/it]
 91%|█████████▏| 438/480 [22:56<02:11,  3.13s/it]
 91%|█████████▏| 439/480 [22:59<02:08,  3.13s/it]
 92%|█████████▏| 440/480 [23:02<02:03,  3.10s/it]
                                                 
{'loss': '0.1062', 'grad_norm': '0.03291', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [23:02<02:03,  3.10s/it]
 92%|█████████▏| 441/480 [23:05<02:02,  3.13s/it]
 92%|█████████▏| 442/480 [23:08<01:59,  3.13s/it]
 92%|█████████▏| 443/480 [23:12<01:59,  3.22s/it]
 92%|█████████▎| 444/480 [23:15<01:55,  3.21s/it]
 93%|█████████▎| 445/480 [23:18<01:50,  3.16s/it]
 93%|█████████▎| 446/480 [23:21<01:47,  3.16s/it]
 93%|█████████▎| 447/480 [23:25<01:44,  3.17s/it]
 93%|█████████▎| 448/480 [23:25<01:18,  2.45s/it]
 94%|█████████▎| 449/480 [23:29<01:23,  2.68s/it]
 94%|█████████▍| 450/480 [23:32<01:24,  2.81s/it]
                                                 
{'loss': '0.07665', 'grad_norm': '0.02329', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [23:32<01:24,  2.81s/it]
 94%|█████████▍| 451/480 [23:35<01:24,  2.93s/it]
 94%|█████████▍| 452/480 [23:38<01:25,  3.04s/it]
 94%|█████████▍| 453/480 [23:41<01:23,  3.11s/it]
 95%|█████████▍| 454/480 [23:45<01:21,  3.14s/it]
 95%|█████████▍| 455/480 [23:48<01:19,  3.17s/it]
 95%|█████████▌| 456/480 [23:51<01:17,  3.22s/it]
 95%|█████████▌| 457/480 [23:54<01:13,  3.21s/it]
 95%|█████████▌| 458/480 [23:57<01:09,  3.17s/it]
 96%|█████████▌| 459/480 [24:00<01:05,  3.13s/it]
 96%|█████████▌| 460/480 [24:04<01:03,  3.17s/it]
                                                 
{'loss': '0.07732', 'grad_norm': '0.006714', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [24:04<01:03,  3.17s/it]
 96%|█████████▌| 461/480 [24:07<00:59,  3.14s/it]
 96%|█████████▋| 462/480 [24:10<00:56,  3.13s/it]
 96%|█████████▋| 463/480 [24:13<00:53,  3.15s/it]
 97%|█████████▋| 464/480 [24:16<00:50,  3.17s/it]
 97%|█████████▋| 465/480 [24:19<00:46,  3.12s/it]
 97%|█████████▋| 466/480 [24:22<00:43,  3.13s/it]
 97%|█████████▋| 467/480 [24:26<00:40,  3.10s/it]
 98%|█████████▊| 468/480 [24:29<00:37,  3.09s/it]
 98%|█████████▊| 469/480 [24:32<00:34,  3.13s/it]
 98%|█████████▊| 470/480 [24:35<00:31,  3.15s/it]
                                                 
{'loss': '0.08697', 'grad_norm': '0.003645', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [24:35<00:31,  3.15s/it]
 98%|█████████▊| 471/480 [24:38<00:28,  3.13s/it]
 98%|█████████▊| 472/480 [24:41<00:25,  3.15s/it]
 99%|█████████▊| 473/480 [24:44<00:22,  3.15s/it]
 99%|█████████▉| 474/480 [24:48<00:18,  3.17s/it]
 99%|█████████▉| 475/480 [24:51<00:15,  3.14s/it]
 99%|█████████▉| 476/480 [24:54<00:12,  3.22s/it]
 99%|█████████▉| 477/480 [24:57<00:09,  3.22s/it]
100%|█████████▉| 478/480 [25:00<00:06,  3.17s/it]
100%|█████████▉| 479/480 [25:04<00:03,  3.19s/it]
100%|██████████| 480/480 [25:05<00:00,  2.55s/it]
                                                 
{'loss': '0.114', 'grad_norm': '0.09729', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [25:05<00:00,  2.55s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:07, 13.60it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:00<00:07, 12.44it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:00<00:07, 13.23it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:00<00:06, 13.67it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:00<00:06, 13.73it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:00<00:06, 13.94it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:01<00:06, 13.64it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:01<00:06, 13.69it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:01<00:05, 13.69it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:01<00:05, 13.91it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:01<00:05, 13.87it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:01<00:05, 13.68it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:01<00:05, 13.87it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:02<00:05, 13.70it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:02<00:05, 13.72it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:02<00:04, 13.80it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:02<00:04, 13.78it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:02<00:04, 13.75it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:02<00:04, 13.93it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:02<00:04, 13.85it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:03<00:04, 14.00it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:03<00:04, 13.94it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:03<00:03, 14.06it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:03<00:03, 14.16it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:03<00:03, 14.23it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:03<00:03, 14.09it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:03<00:03, 14.02it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:04<00:03, 14.15it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:04<00:02, 14.36it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:04<00:02, 14.50it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [00:04<00:02, 14.58it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:04<00:02, 14.40it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:04<00:02, 14.51it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:04<00:02, 14.56it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:04<00:02, 14.67it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:05<00:01, 14.62it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:05<00:01, 14.26it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:05<00:01, 14.41it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:05<00:01, 14.32it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:05<00:01, 14.06it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:05<00:01, 14.19it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:05<00:01, 14.34it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:06<00:00, 14.30it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:06<00:00, 14.46it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:06<00:00, 14.53it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:06<00:00, 14.58it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:06<00:00, 14.61it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:06<00:00, 14.44it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:06<00:00, 14.09it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:07<00:00, 14.26it/s]
+
                                                                    [EVAL] eval_rougeL=88.0000 (best=91.0000)
+
                                                 
{'eval_rougeL': '88', 'epoch': '15'}
+
100%|██████████| 480/480 [25:12<00:00,  2.55s/it]
                                                 
{'train_runtime': '1512', 'train_samples_per_second': '9.919', 'train_steps_per_second': '0.317', 'train_loss': '0.1675', 'epoch': '15'}
+
100%|██████████| 480/480 [25:12<00:00,  2.55s/it]
100%|██████████| 480/480 [25:12<00:00,  3.15s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 50859468GF
+  train_loss               =     0.1675
+  train_runtime            = 0:25:12.24
+  train_samples            =       1000
+  train_samples_per_second =      9.919
+  train_steps_per_second   =      0.317
+  [EVAL] Restored best model (eval_rougeL=91.0000, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:16, 17.74it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 156.09it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 164.83it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 168.21it/s]
Loading weights:  29%|██▉       | 85/291 [00:00<00:01, 167.50it/s]
Loading weights:  35%|███▌      | 103/291 [00:00<00:01, 169.92it/s]
Loading weights:  42%|████▏     | 123/291 [00:00<00:00, 176.26it/s]
Loading weights:  49%|████▉     | 144/291 [00:00<00:00, 182.87it/s]
Loading weights:  56%|█████▋    | 164/291 [00:00<00:00, 180.71it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 179.82it/s]
Loading weights:  70%|██████▉   | 203/291 [00:01<00:00, 180.60it/s]
Loading weights:  76%|███████▋  | 222/291 [00:01<00:00, 180.04it/s]
Loading weights:  84%|████████▍ | 244/291 [00:01<00:00, 185.72it/s]
Loading weights:  91%|█████████ | 264/291 [00:01<00:00, 187.36it/s]
Loading weights:  97%|█████████▋| 283/291 [00:01<00:00, 177.14it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 176.39it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task363_sst2_polarity_classification': PaR=873.0, shrinkage=ridge, n=1000
+  [SRT] Wired router: 2 sigs, cur=task363_sst2_polarity_classification, mapping={'task363_sst2_polarity_classification': 0, 'task1572_samsum_summary': 1}
+[SRT] Router: 2 tasks, avg_PaR=873.0, metrics=['ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification']
+  [SRT-INIT] task_id_to_idx={'task363_sst2_polarity_classification': 0, 'task1572_samsum_summary': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/120 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=2
+
Continual predict:   1%|          | 1/120 [00:01<03:05,  1.56s/it]
Continual predict:   2%|▏         | 2/120 [00:03<03:53,  1.98s/it]
Continual predict:   2%|▎         | 3/120 [00:04<02:46,  1.42s/it]
Continual predict:   3%|▎         | 4/120 [00:05<02:00,  1.04s/it]
Continual predict:   4%|▍         | 5/120 [00:06<02:22,  1.24s/it]
Continual predict:   5%|▌         | 6/120 [00:10<04:17,  2.26s/it]
Continual predict:   6%|▌         | 7/120 [00:13<04:43,  2.51s/it]
Continual predict:   7%|▋         | 8/120 [00:15<04:17,  2.30s/it]
Continual predict:   8%|▊         | 9/120 [00:17<03:46,  2.04s/it]
Continual predict:   8%|▊         | 10/120 [00:17<02:59,  1.63s/it]
Continual predict:   9%|▉         | 11/120 [00:20<03:23,  1.87s/it]
Continual predict:  10%|█         | 12/120 [00:23<03:52,  2.15s/it]
Continual predict:  11%|█         | 13/120 [00:24<03:16,  1.84s/it]
Continual predict:  12%|█▏        | 14/120 [00:26<03:24,  1.93s/it]
Continual predict:  12%|█▎        | 15/120 [00:28<03:25,  1.96s/it]
Continual predict:  13%|█▎        | 16/120 [00:30<03:34,  2.07s/it]
Continual predict:  14%|█▍        | 17/120 [00:37<06:10,  3.60s/it]
Continual predict:  15%|█▌        | 18/120 [00:39<05:14,  3.09s/it]
Continual predict:  16%|█▌        | 19/120 [00:42<04:47,  2.85s/it]
Continual predict:  17%|█▋        | 20/120 [00:43<03:55,  2.36s/it]
Continual predict:  18%|█▊        | 21/120 [00:43<02:47,  1.69s/it]
Continual predict:  18%|█▊        | 22/120 [00:43<01:59,  1.22s/it]
Continual predict:  19%|█▉        | 23/120 [00:43<01:26,  1.12it/s]
Continual predict:  20%|██        | 24/120 [00:43<01:03,  1.51it/s]
Continual predict:  21%|██        | 25/120 [00:43<00:47,  2.00it/s]
Continual predict:  22%|██▏       | 26/120 [00:44<00:36,  2.60it/s]
Continual predict:  22%|██▎       | 27/120 [00:44<00:28,  3.27it/s]
Continual predict:  23%|██▎       | 28/120 [00:44<00:22,  4.01it/s]
Continual predict:  24%|██▍       | 29/120 [00:44<00:19,  4.77it/s]
Continual predict:  25%|██▌       | 30/120 [00:44<00:16,  5.36it/s]
Continual predict:  26%|██▌       | 31/120 [00:44<00:14,  6.00it/s]
Continual predict:  27%|██▋       | 32/120 [00:44<00:13,  6.57it/s]
Continual predict:  28%|██▊       | 33/120 [00:44<00:12,  6.80it/s]
Continual predict:  28%|██▊       | 34/120 [00:45<00:12,  7.00it/s]
Continual predict:  29%|██▉       | 35/120 [00:45<00:11,  7.15it/s]
Continual predict:  30%|███       | 36/120 [00:45<00:11,  7.51it/s]
Continual predict:  31%|███       | 37/120 [00:45<00:11,  7.52it/s]
Continual predict:  32%|███▏      | 38/120 [00:45<00:10,  7.74it/s]
Continual predict:  32%|███▎      | 39/120 [00:45<00:10,  7.94it/s]
Continual predict:  33%|███▎      | 40/120 [00:45<00:10,  7.35it/s]
Continual predict:  34%|███▍      | 41/120 [00:45<00:10,  7.37it/s]
Continual predict:  35%|███▌      | 42/120 [00:46<00:10,  7.51it/s]
Continual predict:  36%|███▌      | 43/120 [00:46<00:10,  7.51it/s]
Continual predict:  37%|███▋      | 44/120 [00:46<00:10,  7.54it/s]
Continual predict:  38%|███▊      | 45/120 [00:46<00:09,  7.76it/s]
Continual predict:  38%|███▊      | 46/120 [00:46<00:09,  7.92it/s]
Continual predict:  39%|███▉      | 47/120 [00:46<00:09,  7.96it/s]
Continual predict:  40%|████      | 48/120 [00:46<00:09,  7.80it/s]
Continual predict:  41%|████      | 49/120 [00:46<00:09,  7.74it/s]
Continual predict:  42%|████▏     | 50/120 [00:47<00:08,  7.91it/s]
Continual predict:  42%|████▎     | 51/120 [00:47<00:08,  7.95it/s]
Continual predict:  43%|████▎     | 52/120 [00:47<00:08,  8.06it/s]
Continual predict:  44%|████▍     | 53/120 [00:47<00:08,  8.15it/s]
Continual predict:  45%|████▌     | 54/120 [00:47<00:08,  7.96it/s]
Continual predict:  46%|████▌     | 55/120 [00:47<00:08,  8.11it/s]
Continual predict:  47%|████▋     | 56/120 [00:47<00:08,  7.91it/s]
Continual predict:  48%|████▊     | 57/120 [00:47<00:07,  8.04it/s]
Continual predict:  48%|████▊     | 58/120 [00:48<00:07,  8.14it/s]
Continual predict:  49%|████▉     | 59/120 [00:48<00:07,  8.20it/s]
Continual predict:  50%|█████     | 60/120 [00:48<00:07,  7.96it/s]
Continual predict:  51%|█████     | 61/120 [00:48<00:07,  8.09it/s]
Continual predict:  52%|█████▏    | 62/120 [00:48<00:07,  8.17it/s]
Continual predict:  52%|█████▎    | 63/120 [00:48<00:07,  7.99it/s]
Continual predict:  53%|█████▎    | 64/120 [00:48<00:06,  8.09it/s]
Continual predict:  54%|█████▍    | 65/120 [00:48<00:06,  8.16it/s]
Continual predict:  55%|█████▌    | 66/120 [00:49<00:06,  8.22it/s]
Continual predict:  56%|█████▌    | 67/120 [00:49<00:06,  8.29it/s]
Continual predict:  57%|█████▋    | 68/120 [00:49<00:06,  8.34it/s]
Continual predict:  57%|█████▊    | 69/120 [00:49<00:06,  8.36it/s]
Continual predict:  58%|█████▊    | 70/120 [00:49<00:05,  8.39it/s]
Continual predict:  59%|█████▉    | 71/120 [00:49<00:05,  8.38it/s]
Continual predict:  60%|██████    | 72/120 [00:49<00:05,  8.11it/s]
Continual predict:  61%|██████    | 73/120 [00:49<00:05,  8.21it/s]
Continual predict:  62%|██████▏   | 74/120 [00:50<00:05,  7.99it/s]
Continual predict:  62%|██████▎   | 75/120 [00:50<00:05,  7.98it/s]
Continual predict:  63%|██████▎   | 76/120 [00:50<00:05,  8.08it/s]
Continual predict:  64%|██████▍   | 77/120 [00:50<00:05,  8.12it/s]
Continual predict:  65%|██████▌   | 78/120 [00:50<00:05,  8.21it/s]
Continual predict:  66%|██████▌   | 79/120 [00:50<00:04,  8.24it/s]
Continual predict:  67%|██████▋   | 80/120 [00:50<00:04,  8.26it/s]
Continual predict:  68%|██████▊   | 81/120 [00:50<00:04,  8.27it/s]
Continual predict:  68%|██████▊   | 82/120 [00:51<00:04,  8.30it/s]
Continual predict:  69%|██████▉   | 83/120 [00:51<00:04,  8.32it/s]
Continual predict:  70%|███████   | 84/120 [00:51<00:04,  8.07it/s]
Continual predict:  71%|███████   | 85/120 [00:51<00:04,  8.08it/s]
Continual predict:  72%|███████▏  | 86/120 [00:51<00:04,  8.16it/s]
Continual predict:  72%|███████▎  | 87/120 [00:51<00:04,  8.17it/s]
Continual predict:  73%|███████▎  | 88/120 [00:51<00:03,  8.26it/s]
Continual predict:  74%|███████▍  | 89/120 [00:51<00:03,  8.31it/s]
Continual predict:  75%|███████▌  | 90/120 [00:51<00:03,  8.36it/s]
Continual predict:  76%|███████▌  | 91/120 [00:52<00:03,  8.36it/s]
Continual predict:  77%|███████▋  | 92/120 [00:52<00:03,  8.30it/s]
Continual predict:  78%|███████▊  | 93/120 [00:52<00:03,  8.05it/s]
Continual predict:  78%|███████▊  | 94/120 [00:52<00:03,  7.92it/s]
Continual predict:  79%|███████▉  | 95/120 [00:52<00:03,  8.04it/s]
Continual predict:  80%|████████  | 96/120 [00:52<00:02,  8.12it/s]
Continual predict:  81%|████████  | 97/120 [00:52<00:02,  8.19it/s]
Continual predict:  82%|████████▏ | 98/120 [00:52<00:02,  8.00it/s]
Continual predict:  82%|████████▎ | 99/120 [00:53<00:02,  7.86it/s]
Continual predict:  83%|████████▎ | 100/120 [00:53<00:02,  7.77it/s]
Continual predict:  84%|████████▍ | 101/120 [00:53<00:02,  7.94it/s]
Continual predict:  85%|████████▌ | 102/120 [00:53<00:02,  8.10it/s]
Continual predict:  86%|████████▌ | 103/120 [00:53<00:02,  8.19it/s]
Continual predict:  87%|████████▋ | 104/120 [00:53<00:01,  8.24it/s]
Continual predict:  88%|████████▊ | 105/120 [00:53<00:01,  8.03it/s]
Continual predict:  88%|████████▊ | 106/120 [00:53<00:01,  8.16it/s]
Continual predict:  89%|████████▉ | 107/120 [00:54<00:01,  8.25it/s]
Continual predict:  90%|█████████ | 108/120 [00:54<00:01,  8.29it/s]
Continual predict:  91%|█████████ | 109/120 [00:54<00:01,  8.31it/s]
Continual predict:  92%|█████████▏| 110/120 [00:54<00:01,  8.32it/s]
Continual predict:  92%|█████████▎| 111/120 [00:54<00:01,  8.33it/s]
Continual predict:  93%|█████████▎| 112/120 [00:54<00:00,  8.34it/s]
Continual predict:  94%|█████████▍| 113/120 [00:54<00:00,  8.33it/s]
Continual predict:  95%|█████████▌| 114/120 [00:54<00:00,  8.33it/s]
Continual predict:  96%|█████████▌| 115/120 [00:55<00:00,  8.09it/s]
Continual predict:  97%|█████████▋| 116/120 [00:55<00:00,  8.16it/s]
Continual predict:  98%|█████████▊| 117/120 [00:55<00:00,  7.98it/s]
Continual predict:  98%|█████████▊| 118/120 [00:55<00:00,  7.84it/s]
Continual predict:  99%|█████████▉| 119/120 [00:55<00:00,  7.99it/s]
Continual predict: 100%|██████████| 120/120 [00:55<00:00,  8.09it/s]
Continual predict: 100%|██████████| 120/120 [00:55<00:00,  2.15it/s]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=20]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (120/120)  label_acc=100.0% (120/120)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                        =       15.0
+  predict_eval_rougeL                                          =    82.7288
+  predict_eval_rougeL_for_CL                                   =    82.7288
+  predict_eval_rougeL_for_task1572_samsum_summary              =    41.3728
+  predict_eval_rougeL_for_task363_sst2_polarity_classification =       91.0
+  predict_exact_match                                          =    75.8333
+  predict_exact_match_for_CL                                   =    75.8333
+  predict_exact_match_for_task1572_samsum_summary              =        0.0
+  predict_exact_match_for_task363_sst2_polarity_classification =       91.0
+  predict_gen_len                                              =      6.225
+  predict_global_step                                          =        480
+  predict_rouge1                                               =    84.2313
+  predict_rouge1_for_CL                                        =    84.2313
+  predict_rouge1_for_task1572_samsum_summary                   =    50.3879
+  predict_rouge1_for_task363_sst2_polarity_classification      =       91.0
+  predict_runtime                                              = 0:00:55.69
+  predict_samples                                              =        120
+  predict_samples_per_second                                   =      2.154
+  predict_steps_per_second                                     =      2.154
+[2026-04-29 20:03:06,821] [INFO] [launch.py:367:main] Process 492775 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/all_results.json b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..c621c2f866ee97b7d78f0a71c558a3b782f4b776
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/all_results.json
@@ -0,0 +1,30 @@
+{
+  "predict_exact_match": 41.3636,
+  "predict_rouge1": 62.6489,
+  "predict_eval_rougeL": 58.5759,
+  "predict_gen_len": 14.6182,
+  "predict_runtime": 359.74325460591353,
+  "predict_samples": 220,
+  "predict_samples_per_second": 0.612,
+  "predict_steps_per_second": 0.612,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 41.3636,
+  "predict_rouge1_for_CL": 62.6489,
+  "predict_eval_rougeL_for_CL": 58.5759,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "train_runtime": 4023.2929,
+  "train_samples_per_second": 3.728,
+  "train_steps_per_second": 0.119,
+  "total_flos": 3.4023235089176986e+17,
+  "train_loss": 1.1839401006698609,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..26ef59297aeafa3d0682cf8a057dcab5a4cdf9db
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cebfd575e3c018a3a487b9b7502460122e8ac4256efd61c4120faa1d39c26282
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8bdb70b5454b2a43793bafd21ad86615de3a97cb
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd3a8c00b469368f12ef3998c058fd9dbb18a3ed86f704c4baf5f2f9b1c5d08e
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..420e796bf740f2683cd46f85e87cfeb2b254253f
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b15635b35368a1221a1444bce137aa0752d010db026ee457b416405f76886c2c
+size 557326126
diff --git a/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/train.log b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..479023d94d5aa41531623cc04b6ce77048091c3d
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/train.log
@@ -0,0 +1,487 @@
+[2026-04-29 20:03:12,845] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 20:03:12,845] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29502 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1290_xsum_summarization --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization --cur_task_id 2 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[2026-04-29 20:03:16,042] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 20:03:16,043] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 20:03:16,043] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 20:03:16,043] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 20:03:16,043] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 20:03:16,043] [INFO] [launch.py:272:main] process 560438 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1290_xsum_summarization', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization', '--cur_task_id', '2', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights']
+
+============================================================
+Task 2: task1290_xsum_summarization
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/220
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 114.63it/s]
Loading weights:  11%|█▏        | 33/291 [00:00<00:01, 159.40it/s]
Loading weights:  19%|█▊        | 54/291 [00:00<00:01, 180.31it/s]
Loading weights:  25%|██▌       | 74/291 [00:00<00:01, 185.01it/s]
Loading weights:  32%|███▏      | 93/291 [00:00<00:01, 177.99it/s]
Loading weights:  38%|███▊      | 112/291 [00:00<00:01, 174.11it/s]
Loading weights:  45%|████▍     | 130/291 [00:00<00:00, 171.58it/s]
Loading weights:  51%|█████     | 149/291 [00:00<00:00, 170.50it/s]
Loading weights:  59%|█████▉    | 171/291 [00:00<00:00, 184.88it/s]
Loading weights:  65%|██████▌   | 190/291 [00:01<00:00, 180.59it/s]
Loading weights:  72%|███████▏  | 209/291 [00:01<00:00, 181.45it/s]
Loading weights:  78%|███████▊  | 228/291 [00:01<00:00, 179.36it/s]
Loading weights:  85%|████████▍ | 246/291 [00:01<00:00, 179.47it/s]
Loading weights:  91%|█████████ | 264/291 [00:01<00:00, 178.98it/s]
Loading weights:  97%|█████████▋| 283/291 [00:01<00:00, 177.61it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 179.41it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=419)
+[FIX] Re-initialized lora_A/B in 192 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(2, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 2 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6745.539M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 2 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 2 sigs, cur=task1290_xsum_summarization, mapping={'task1290_xsum_summarization': 0, 'task1572_samsum_summary': 2, 'task363_sst2_polarity_classification': 1}
+[SGWI] Task 2: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 20:04:11,047] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:07<1:03:34,  7.96s/it]
  0%|          | 2/480 [00:15<1:00:31,  7.60s/it]
  1%|          | 3/480 [00:22<1:00:03,  7.55s/it]
  1%|          | 4/480 [00:31<1:02:39,  7.90s/it]
  1%|          | 5/480 [00:38<59:21,  7.50s/it]  
  1%|▏         | 6/480 [00:46<1:02:06,  7.86s/it]
  1%|▏         | 7/480 [00:54<1:01:00,  7.74s/it]
  2%|▏         | 8/480 [01:00<58:20,  7.42s/it]  
  2%|▏         | 9/480 [01:09<1:01:44,  7.86s/it]
  2%|▏         | 10/480 [01:17<1:02:07,  7.93s/it]
                                                  
{'loss': '1.638', 'grad_norm': '0.1432', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [01:17<1:02:07,  7.93s/it]
  2%|▏         | 11/480 [01:25<1:01:14,  7.83s/it]
  2%|▎         | 12/480 [01:33<1:01:57,  7.94s/it]
  3%|▎         | 13/480 [01:40<58:50,  7.56s/it]  
  3%|▎         | 14/480 [01:48<59:50,  7.71s/it]
  3%|▎         | 15/480 [01:55<58:47,  7.59s/it]
  3%|▎         | 16/480 [02:03<59:21,  7.68s/it]
  4%|▎         | 17/480 [02:10<58:49,  7.62s/it]
  4%|▍         | 18/480 [02:17<57:05,  7.41s/it]
  4%|▍         | 19/480 [02:25<56:19,  7.33s/it]
  4%|▍         | 20/480 [02:31<54:31,  7.11s/it]
                                                
{'loss': '1.504', 'grad_norm': '0.2431', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [02:31<54:31,  7.11s/it]
  4%|▍         | 21/480 [02:38<53:29,  6.99s/it]
  5%|▍         | 22/480 [02:46<56:52,  7.45s/it]
  5%|▍         | 23/480 [02:53<55:55,  7.34s/it]
  5%|▌         | 24/480 [03:01<56:24,  7.42s/it]
  5%|▌         | 25/480 [03:09<58:32,  7.72s/it]
  5%|▌         | 26/480 [03:17<57:59,  7.66s/it]
  6%|▌         | 27/480 [03:25<57:43,  7.65s/it]
  6%|▌         | 28/480 [03:31<55:26,  7.36s/it]
  6%|▌         | 29/480 [03:39<55:28,  7.38s/it]
  6%|▋         | 30/480 [03:46<55:24,  7.39s/it]
                                                
{'loss': '1.417', 'grad_norm': '0.1845', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [03:46<55:24,  7.39s/it]
  6%|▋         | 31/480 [03:53<53:30,  7.15s/it]
  7%|▋         | 32/480 [03:54<40:07,  5.37s/it]
  7%|▋         | 33/480 [04:01<44:04,  5.92s/it]
  7%|▋         | 34/480 [04:08<45:30,  6.12s/it]
  7%|▋         | 35/480 [04:15<48:28,  6.53s/it]
  8%|▊         | 36/480 [04:24<52:24,  7.08s/it]
  8%|▊         | 37/480 [04:31<53:13,  7.21s/it]
  8%|▊         | 38/480 [04:38<52:53,  7.18s/it]
  8%|▊         | 39/480 [04:47<55:51,  7.60s/it]
  8%|▊         | 40/480 [04:54<54:57,  7.49s/it]
                                                
{'loss': '1.288', 'grad_norm': '0.1836', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [04:54<54:57,  7.49s/it]
  9%|▊         | 41/480 [05:01<54:36,  7.46s/it]
  9%|▉         | 42/480 [05:09<54:37,  7.48s/it]
  9%|▉         | 43/480 [05:17<55:41,  7.65s/it]
  9%|▉         | 44/480 [05:23<52:24,  7.21s/it]
  9%|▉         | 45/480 [05:31<53:17,  7.35s/it]
 10%|▉         | 46/480 [05:38<53:47,  7.44s/it]
 10%|▉         | 47/480 [05:46<54:04,  7.49s/it]
 10%|█         | 48/480 [05:53<52:56,  7.35s/it]
 10%|█         | 49/480 [06:00<52:20,  7.29s/it]
 10%|█         | 50/480 [06:09<55:02,  7.68s/it]
                                                
{'loss': '1.313', 'grad_norm': '0.3724', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [06:09<55:02,  7.68s/it]
 11%|█         | 51/480 [06:17<54:59,  7.69s/it]
 11%|█         | 52/480 [06:24<53:50,  7.55s/it]
 11%|█         | 53/480 [06:32<55:11,  7.76s/it]
 11%|█▏        | 54/480 [06:40<54:35,  7.69s/it]
 11%|█▏        | 55/480 [06:47<53:16,  7.52s/it]
 12%|█▏        | 56/480 [06:55<54:54,  7.77s/it]
 12%|█▏        | 57/480 [07:03<54:48,  7.78s/it]
 12%|█▏        | 58/480 [07:10<53:18,  7.58s/it]
 12%|█▏        | 59/480 [07:17<51:45,  7.38s/it]
 12%|█▎        | 60/480 [07:25<53:37,  7.66s/it]
                                                
{'loss': '1.265', 'grad_norm': '0.3128', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [07:25<53:37,  7.66s/it]
 13%|█▎        | 61/480 [07:33<53:25,  7.65s/it]
 13%|█▎        | 62/480 [07:40<52:41,  7.56s/it]
 13%|█▎        | 63/480 [07:47<51:59,  7.48s/it]
 13%|█▎        | 64/480 [07:49<39:45,  5.73s/it]
 14%|█▎        | 65/480 [07:56<41:14,  5.96s/it]
 14%|█▍        | 66/480 [08:03<44:08,  6.40s/it]
 14%|█▍        | 67/480 [08:10<45:34,  6.62s/it]
 14%|█▍        | 68/480 [08:17<46:42,  6.80s/it]
 14%|█▍        | 69/480 [08:25<48:50,  7.13s/it]
 15%|█▍        | 70/480 [08:33<50:16,  7.36s/it]
                                                
{'loss': '1.246', 'grad_norm': '0.2669', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [08:33<50:16,  7.36s/it]
 15%|█▍        | 71/480 [08:41<50:17,  7.38s/it]
 15%|█▌        | 72/480 [08:48<49:50,  7.33s/it]
 15%|█▌        | 73/480 [08:54<48:07,  7.10s/it]
 15%|█▌        | 74/480 [09:02<48:31,  7.17s/it]
 16%|█▌        | 75/480 [09:10<51:01,  7.56s/it]
 16%|█▌        | 76/480 [09:18<51:58,  7.72s/it]
 16%|█▌        | 77/480 [09:26<50:59,  7.59s/it]
 16%|█▋        | 78/480 [09:32<47:51,  7.14s/it]
 16%|█▋        | 79/480 [09:39<49:00,  7.33s/it]
 17%|█▋        | 80/480 [09:47<48:57,  7.34s/it]
                                                
{'loss': '1.226', 'grad_norm': '0.211', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [09:47<48:57,  7.34s/it]
 17%|█▋        | 81/480 [09:55<50:52,  7.65s/it]
 17%|█▋        | 82/480 [10:02<48:48,  7.36s/it]
 17%|█▋        | 83/480 [10:10<50:15,  7.59s/it]
 18%|█▊        | 84/480 [10:17<49:01,  7.43s/it]
 18%|█▊        | 85/480 [10:25<49:45,  7.56s/it]
 18%|█▊        | 86/480 [10:33<50:05,  7.63s/it]
 18%|█▊        | 87/480 [10:41<51:24,  7.85s/it]
 18%|█▊        | 88/480 [10:49<51:10,  7.83s/it]
 19%|█▊        | 89/480 [10:57<51:28,  7.90s/it]
 19%|█▉        | 90/480 [11:05<52:16,  8.04s/it]
                                                
{'loss': '1.238', 'grad_norm': '0.2478', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [11:05<52:16,  8.04s/it]
 19%|█▉        | 91/480 [11:12<50:05,  7.73s/it]
 19%|█▉        | 92/480 [11:20<50:37,  7.83s/it]
 19%|█▉        | 93/480 [11:29<51:09,  7.93s/it]
 20%|█▉        | 94/480 [11:36<49:35,  7.71s/it]
 20%|█▉        | 95/480 [11:43<48:28,  7.55s/it]
 20%|██        | 96/480 [11:45<36:58,  5.78s/it]
 20%|██        | 97/480 [11:53<42:45,  6.70s/it]
 20%|██        | 98/480 [12:00<42:55,  6.74s/it]
 21%|██        | 99/480 [12:07<43:48,  6.90s/it]
 21%|██        | 100/480 [12:16<45:49,  7.23s/it]
                                                 
{'loss': '1.249', 'grad_norm': '0.1779', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [12:16<45:49,  7.23s/it]
 21%|██        | 101/480 [12:22<44:29,  7.04s/it]
 21%|██▏       | 102/480 [12:30<46:06,  7.32s/it]
 21%|██▏       | 103/480 [12:38<46:39,  7.43s/it]
 22%|██▏       | 104/480 [12:45<45:41,  7.29s/it]
 22%|██▏       | 105/480 [12:52<46:00,  7.36s/it]
 22%|██▏       | 106/480 [12:59<45:02,  7.23s/it]
 22%|██▏       | 107/480 [13:06<44:33,  7.17s/it]
 22%|██▎       | 108/480 [13:14<44:49,  7.23s/it]
 23%|██▎       | 109/480 [13:22<46:50,  7.57s/it]
 23%|██▎       | 110/480 [13:30<48:08,  7.81s/it]
                                                 
{'loss': '1.23', 'grad_norm': '0.1446', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [13:30<48:08,  7.81s/it]
 23%|██▎       | 111/480 [13:39<48:47,  7.93s/it]
 23%|██▎       | 112/480 [13:47<49:17,  8.04s/it]
 24%|██▎       | 113/480 [13:55<48:42,  7.96s/it]
 24%|██▍       | 114/480 [14:03<48:36,  7.97s/it]
 24%|██▍       | 115/480 [14:10<48:01,  7.89s/it]
 24%|██▍       | 116/480 [14:17<46:26,  7.65s/it]
 24%|██▍       | 117/480 [14:24<43:43,  7.23s/it]
 25%|██▍       | 118/480 [14:31<43:20,  7.18s/it]
 25%|██▍       | 119/480 [14:38<44:02,  7.32s/it]
 25%|██▌       | 120/480 [14:47<45:44,  7.62s/it]
                                                 
{'loss': '1.2', 'grad_norm': '0.3601', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [14:47<45:44,  7.62s/it]
 25%|██▌       | 121/480 [14:54<44:51,  7.50s/it]
 25%|██▌       | 122/480 [15:00<43:10,  7.23s/it]
 26%|██▌       | 123/480 [15:08<43:20,  7.28s/it]
 26%|██▌       | 124/480 [15:15<43:16,  7.29s/it]
 26%|██▌       | 125/480 [15:23<43:49,  7.41s/it]
 26%|██▋       | 126/480 [15:30<43:56,  7.45s/it]
 26%|██▋       | 127/480 [15:37<42:58,  7.31s/it]
 27%|██▋       | 128/480 [15:40<33:58,  5.79s/it]
 27%|██▋       | 129/480 [15:47<35:59,  6.15s/it]
 27%|██▋       | 130/480 [15:54<37:57,  6.51s/it]
                                                 
{'loss': '1.224', 'grad_norm': '0.3379', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [15:54<37:57,  6.51s/it]
 27%|██▋       | 131/480 [16:01<38:17,  6.58s/it]
 28%|██▊       | 132/480 [16:08<39:12,  6.76s/it]
 28%|██▊       | 133/480 [16:15<39:55,  6.90s/it]
 28%|██▊       | 134/480 [16:23<40:37,  7.04s/it]
 28%|██▊       | 135/480 [16:30<41:42,  7.25s/it]
 28%|██▊       | 136/480 [16:38<42:34,  7.43s/it]
 29%|██▊       | 137/480 [16:46<43:40,  7.64s/it]
 29%|██▉       | 138/480 [16:53<41:55,  7.36s/it]
 29%|██▉       | 139/480 [17:00<41:01,  7.22s/it]
 29%|██▉       | 140/480 [17:07<41:29,  7.32s/it]
                                                 
{'loss': '1.23', 'grad_norm': '0.3268', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [17:07<41:29,  7.32s/it]
 29%|██▉       | 141/480 [17:14<39:31,  7.00s/it]
 30%|██▉       | 142/480 [17:21<40:09,  7.13s/it]
 30%|██▉       | 143/480 [17:29<40:44,  7.25s/it]
 30%|███       | 144/480 [17:36<41:39,  7.44s/it]
 30%|███       | 145/480 [17:44<42:25,  7.60s/it]
 30%|███       | 146/480 [17:52<42:18,  7.60s/it]
 31%|███       | 147/480 [18:00<42:47,  7.71s/it]
 31%|███       | 148/480 [18:08<42:32,  7.69s/it]
 31%|███       | 149/480 [18:15<41:39,  7.55s/it]
 31%|███▏      | 150/480 [18:23<42:43,  7.77s/it]
                                                 
{'loss': '1.195', 'grad_norm': '0.1836', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [18:23<42:43,  7.77s/it]
 31%|███▏      | 151/480 [18:31<43:07,  7.86s/it]
 32%|███▏      | 152/480 [18:38<41:29,  7.59s/it]
 32%|███▏      | 153/480 [18:47<42:47,  7.85s/it]
 32%|███▏      | 154/480 [18:54<41:29,  7.64s/it]
 32%|███▏      | 155/480 [19:01<40:33,  7.49s/it]
 32%|███▎      | 156/480 [19:09<41:37,  7.71s/it]
 33%|███▎      | 157/480 [19:18<42:35,  7.91s/it]
 33%|███▎      | 158/480 [19:25<42:10,  7.86s/it]
 33%|███▎      | 159/480 [19:33<42:09,  7.88s/it]
 33%|███▎      | 160/480 [19:35<31:58,  6.00s/it]
                                                 
{'loss': '1.213', 'grad_norm': '1.306', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [19:35<31:58,  6.00s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:02<03:47,  2.30s/it]
+
Eval @ step 160:   2%|▏         | 2/100 [00:04<03:44,  2.29s/it]
+
Eval @ step 160:   3%|▎         | 3/100 [00:07<03:59,  2.47s/it]
+
Eval @ step 160:   4%|▍         | 4/100 [00:08<03:24,  2.13s/it]
+
Eval @ step 160:   5%|▌         | 5/100 [00:12<04:07,  2.60s/it]
+
Eval @ step 160:   6%|▌         | 6/100 [00:14<03:59,  2.55s/it]
+
Eval @ step 160:   7%|▋         | 7/100 [00:18<04:42,  3.04s/it]
+
Eval @ step 160:   8%|▊         | 8/100 [00:20<03:52,  2.53s/it]
+
Eval @ step 160:   9%|▉         | 9/100 [00:22<03:29,  2.30s/it]
+
Eval @ step 160:  10%|█         | 10/100 [00:23<03:06,  2.07s/it]
+
Eval @ step 160:  11%|█         | 11/100 [00:24<02:36,  1.76s/it]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:26<02:32,  1.73s/it]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:28<02:29,  1.72s/it]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:31<03:07,  2.18s/it]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:33<03:18,  2.33s/it]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:35<02:47,  2.00s/it]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:36<02:24,  1.74s/it]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:36<01:52,  1.37s/it]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:37<01:33,  1.15s/it]
+
Eval @ step 160:  20%|██        | 20/100 [00:38<01:30,  1.13s/it]
+
Eval @ step 160:  21%|██        | 21/100 [00:41<02:17,  1.74s/it]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:44<02:29,  1.92s/it]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:45<02:22,  1.85s/it]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:49<03:05,  2.44s/it]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:51<02:43,  2.17s/it]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:52<02:12,  1.80s/it]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:54<02:22,  1.96s/it]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:55<01:57,  1.64s/it]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:59<02:44,  2.32s/it]
+
Eval @ step 160:  30%|███       | 30/100 [01:00<02:17,  1.96s/it]
+
Eval @ step 160:  31%|███       | 31/100 [01:03<02:38,  2.29s/it]
+
Eval @ step 160:  32%|███▏      | 32/100 [01:04<02:17,  2.02s/it]
+
Eval @ step 160:  33%|███▎      | 33/100 [01:06<02:05,  1.87s/it]
+
Eval @ step 160:  34%|███▍      | 34/100 [01:07<01:58,  1.79s/it]
+
Eval @ step 160:  35%|███▌      | 35/100 [01:09<02:00,  1.85s/it]
+
Eval @ step 160:  36%|███▌      | 36/100 [01:11<01:53,  1.78s/it]
+
Eval @ step 160:  37%|███▋      | 37/100 [01:13<01:50,  1.75s/it]
+
Eval @ step 160:  38%|███▊      | 38/100 [01:13<01:25,  1.38s/it]
+
Eval @ step 160:  39%|███▉      | 39/100 [01:16<01:43,  1.70s/it]
+
Eval @ step 160:  40%|████      | 40/100 [01:17<01:35,  1.59s/it]
+
Eval @ step 160:  41%|████      | 41/100 [01:18<01:30,  1.54s/it]
+
Eval @ step 160:  42%|████▏     | 42/100 [01:21<01:49,  1.88s/it]
+
Eval @ step 160:  43%|████▎     | 43/100 [01:22<01:33,  1.63s/it]
+
Eval @ step 160:  44%|████▍     | 44/100 [01:23<01:25,  1.52s/it]
+
Eval @ step 160:  45%|████▌     | 45/100 [01:25<01:24,  1.54s/it]
+
Eval @ step 160:  46%|████▌     | 46/100 [01:26<01:22,  1.53s/it]
+
Eval @ step 160:  47%|████▋     | 47/100 [01:27<01:08,  1.29s/it]
+
Eval @ step 160:  48%|████▊     | 48/100 [01:28<01:05,  1.26s/it]
+
Eval @ step 160:  49%|████▉     | 49/100 [01:29<00:56,  1.12s/it]
+
Eval @ step 160:  50%|█████     | 50/100 [01:31<01:03,  1.26s/it]
+
Eval @ step 160:  51%|█████     | 51/100 [01:31<00:54,  1.11s/it]
+
Eval @ step 160:  52%|█████▏    | 52/100 [01:32<00:49,  1.03s/it]
+
Eval @ step 160:  53%|█████▎    | 53/100 [01:35<01:13,  1.56s/it]
+
Eval @ step 160:  54%|█████▍    | 54/100 [01:36<01:03,  1.38s/it]
+
Eval @ step 160:  55%|█████▌    | 55/100 [01:37<01:01,  1.36s/it]
+
Eval @ step 160:  56%|█████▌    | 56/100 [01:39<01:01,  1.39s/it]
+
Eval @ step 160:  57%|█████▋    | 57/100 [01:41<01:04,  1.49s/it]
+
Eval @ step 160:  58%|█████▊    | 58/100 [01:41<00:53,  1.28s/it]
+
Eval @ step 160:  59%|█████▉    | 59/100 [01:42<00:45,  1.11s/it]
+
Eval @ step 160:  60%|██████    | 60/100 [01:43<00:39,  1.01it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [01:44<00:35,  1.09it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [01:46<00:52,  1.39s/it]
+
Eval @ step 160:  63%|██████▎   | 63/100 [01:46<00:37,  1.01s/it]
+
Eval @ step 160:  64%|██████▍   | 64/100 [01:47<00:32,  1.12it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [01:48<00:39,  1.13s/it]
+
Eval @ step 160:  66%|██████▌   | 66/100 [01:49<00:32,  1.05it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [01:51<00:43,  1.32s/it]
+
Eval @ step 160:  68%|██████▊   | 68/100 [01:53<00:42,  1.34s/it]
+
Eval @ step 160:  69%|██████▉   | 69/100 [01:55<00:54,  1.74s/it]
+
Eval @ step 160:  70%|███████   | 70/100 [01:57<00:51,  1.72s/it]
+
Eval @ step 160:  71%|███████   | 71/100 [01:58<00:43,  1.49s/it]
+
Eval @ step 160:  72%|███████▏  | 72/100 [02:00<00:43,  1.54s/it]
+
Eval @ step 160:  73%|███████▎  | 73/100 [02:02<00:46,  1.72s/it]
+
Eval @ step 160:  74%|███████▍  | 74/100 [02:04<00:49,  1.92s/it]
+
Eval @ step 160:  75%|███████▌  | 75/100 [02:07<00:56,  2.25s/it]
+
Eval @ step 160:  76%|███████▌  | 76/100 [02:09<00:51,  2.14s/it]
+
Eval @ step 160:  77%|███████▋  | 77/100 [02:10<00:43,  1.91s/it]
+
Eval @ step 160:  78%|███████▊  | 78/100 [02:13<00:44,  2.03s/it]
+
Eval @ step 160:  79%|███████▉  | 79/100 [02:13<00:33,  1.61s/it]
+
Eval @ step 160:  80%|████████  | 80/100 [02:14<00:26,  1.31s/it]
+
Eval @ step 160:  81%|████████  | 81/100 [02:15<00:24,  1.30s/it]
+
Eval @ step 160:  82%|████████▏ | 82/100 [02:17<00:24,  1.37s/it]
+
Eval @ step 160:  83%|████████▎ | 83/100 [02:19<00:28,  1.68s/it]
+
Eval @ step 160:  84%|████████▍ | 84/100 [02:20<00:25,  1.59s/it]
+
Eval @ step 160:  85%|████████▌ | 85/100 [02:21<00:19,  1.30s/it]
+
Eval @ step 160:  86%|████████▌ | 86/100 [02:23<00:21,  1.55s/it]
+
Eval @ step 160:  87%|████████▋ | 87/100 [02:25<00:22,  1.70s/it]
+
Eval @ step 160:  88%|████████▊ | 88/100 [02:29<00:26,  2.22s/it]
+
Eval @ step 160:  89%|████████▉ | 89/100 [02:31<00:24,  2.27s/it]
+
Eval @ step 160:  90%|█████████ | 90/100 [02:34<00:23,  2.34s/it]
+
Eval @ step 160:  91%|█████████ | 91/100 [02:35<00:18,  2.03s/it]
+
Eval @ step 160:  92%|█████████▏| 92/100 [02:36<00:15,  1.89s/it]
+
Eval @ step 160:  93%|█████████▎| 93/100 [02:39<00:14,  2.12s/it]
+
Eval @ step 160:  94%|█████████▍| 94/100 [02:41<00:12,  2.05s/it]
+
Eval @ step 160:  95%|█████████▌| 95/100 [02:42<00:07,  1.59s/it]
+
Eval @ step 160:  96%|█████████▌| 96/100 [02:43<00:05,  1.44s/it]
+
Eval @ step 160:  97%|█████████▋| 97/100 [02:43<00:03,  1.25s/it]
+
Eval @ step 160:  98%|█████████▊| 98/100 [02:45<00:02,  1.36s/it]
+
Eval @ step 160:  99%|█████████▉| 99/100 [02:46<00:01,  1.37s/it]
+
Eval @ step 160: 100%|██████████| 100/100 [02:48<00:00,  1.41s/it]
+
                                                                    [EVAL] ★ New best eval_rougeL=28.3894 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '28.39', 'epoch': '5'}
+
 33%|███▎      | 160/480 [22:23<31:58,  6.00s/it]
 34%|███▎      | 161/480 [22:31<5:03:05, 57.01s/it]
 34%|███▍      | 162/480 [22:39<3:45:04, 42.47s/it]
 34%|███▍      | 163/480 [22:48<2:50:08, 32.20s/it]
 34%|███▍      | 164/480 [22:54<2:08:47, 24.46s/it]
 34%|███▍      | 165/480 [23:03<1:44:21, 19.88s/it]
 35%|███▍      | 166/480 [23:11<1:25:06, 16.26s/it]
 35%|███▍      | 167/480 [23:18<1:10:25, 13.50s/it]
 35%|███▌      | 168/480 [23:26<1:01:15, 11.78s/it]
 35%|███▌      | 169/480 [23:34<54:49, 10.58s/it]  
 35%|███▌      | 170/480 [23:40<47:47,  9.25s/it]
                                                 
{'loss': '1.196', 'grad_norm': '0.3569', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [23:40<47:47,  9.25s/it]
 36%|███▌      | 171/480 [23:48<45:26,  8.82s/it]
 36%|███▌      | 172/480 [23:55<43:10,  8.41s/it]
 36%|███▌      | 173/480 [24:03<42:38,  8.33s/it]
 36%|███▋      | 174/480 [24:10<39:29,  7.74s/it]
 36%|███▋      | 175/480 [24:17<38:50,  7.64s/it]
 37%|███▋      | 176/480 [24:24<37:28,  7.40s/it]
 37%|███▋      | 177/480 [24:31<37:29,  7.43s/it]
 37%|███▋      | 178/480 [24:39<37:13,  7.39s/it]
 37%|███▋      | 179/480 [24:47<37:50,  7.54s/it]
 38%|███▊      | 180/480 [24:54<37:30,  7.50s/it]
                                                 
{'loss': '1.172', 'grad_norm': '0.3158', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [24:54<37:30,  7.50s/it]
 38%|███▊      | 181/480 [25:02<38:04,  7.64s/it]
 38%|███▊      | 182/480 [25:09<37:18,  7.51s/it]
 38%|███▊      | 183/480 [25:18<38:57,  7.87s/it]
 38%|███▊      | 184/480 [25:25<37:48,  7.66s/it]
 39%|███▊      | 185/480 [25:32<36:53,  7.50s/it]
 39%|███▉      | 186/480 [25:39<35:56,  7.33s/it]
 39%|███▉      | 187/480 [25:48<37:54,  7.76s/it]
 39%|███▉      | 188/480 [25:54<35:19,  7.26s/it]
 39%|███▉      | 189/480 [26:01<35:27,  7.31s/it]
 40%|███▉      | 190/480 [26:09<35:38,  7.37s/it]
                                                 
{'loss': '1.154', 'grad_norm': '0.2492', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [26:09<35:38,  7.37s/it]
 40%|███▉      | 191/480 [26:17<36:41,  7.62s/it]
 40%|████      | 192/480 [26:19<28:00,  5.83s/it]
 40%|████      | 193/480 [26:27<31:08,  6.51s/it]
 40%|████      | 194/480 [26:35<33:13,  6.97s/it]
 41%|████      | 195/480 [26:43<34:16,  7.22s/it]
 41%|████      | 196/480 [26:49<33:28,  7.07s/it]
 41%|████      | 197/480 [26:57<34:10,  7.24s/it]
 41%|████▏     | 198/480 [27:04<34:07,  7.26s/it]
 41%|████▏     | 199/480 [27:13<35:20,  7.55s/it]
 42%|████▏     | 200/480 [27:19<34:09,  7.32s/it]
                                                 
{'loss': '1.22', 'grad_norm': '0.5723', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [27:19<34:09,  7.32s/it]
 42%|████▏     | 201/480 [27:27<34:49,  7.49s/it]
 42%|████▏     | 202/480 [27:35<35:12,  7.60s/it]
 42%|████▏     | 203/480 [27:43<35:35,  7.71s/it]
 42%|████▎     | 204/480 [27:51<35:49,  7.79s/it]
 43%|████▎     | 205/480 [27:59<35:43,  7.79s/it]
 43%|████▎     | 206/480 [28:06<34:53,  7.64s/it]
 43%|████▎     | 207/480 [28:14<34:35,  7.60s/it]
 43%|████▎     | 208/480 [28:21<34:45,  7.67s/it]
 44%|████▎     | 209/480 [28:28<33:23,  7.39s/it]
 44%|████▍     | 210/480 [28:35<33:10,  7.37s/it]
                                                 
{'loss': '1.177', 'grad_norm': '0.227', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [28:35<33:10,  7.37s/it]
 44%|████▍     | 211/480 [28:43<33:27,  7.46s/it]
 44%|████▍     | 212/480 [28:51<34:05,  7.63s/it]
 44%|████▍     | 213/480 [28:59<33:48,  7.60s/it]
 45%|████▍     | 214/480 [29:06<33:37,  7.58s/it]
 45%|████▍     | 215/480 [29:14<33:23,  7.56s/it]
 45%|████▌     | 216/480 [29:21<32:45,  7.45s/it]
 45%|████▌     | 217/480 [29:28<31:51,  7.27s/it]
 45%|████▌     | 218/480 [29:35<31:41,  7.26s/it]
 46%|████▌     | 219/480 [29:41<30:27,  7.00s/it]
 46%|████▌     | 220/480 [29:49<31:10,  7.19s/it]
                                                 
{'loss': '1.164', 'grad_norm': '0.2133', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [29:49<31:10,  7.19s/it]
 46%|████▌     | 221/480 [29:56<30:46,  7.13s/it]
 46%|████▋     | 222/480 [30:05<32:52,  7.64s/it]
 46%|████▋     | 223/480 [30:12<32:10,  7.51s/it]
 47%|████▋     | 224/480 [30:14<24:48,  5.81s/it]
 47%|████▋     | 225/480 [30:22<27:22,  6.44s/it]
 47%|████▋     | 226/480 [30:30<29:52,  7.06s/it]
 47%|████▋     | 227/480 [30:39<31:16,  7.42s/it]
 48%|████▊     | 228/480 [30:47<31:48,  7.57s/it]
 48%|████▊     | 229/480 [30:54<31:53,  7.62s/it]
 48%|████▊     | 230/480 [31:01<30:18,  7.27s/it]
                                                 
{'loss': '1.166', 'grad_norm': '0.1929', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [31:01<30:18,  7.27s/it]
 48%|████▊     | 231/480 [31:09<31:28,  7.59s/it]
 48%|████▊     | 232/480 [31:18<32:28,  7.86s/it]
 49%|████▊     | 233/480 [31:25<31:48,  7.73s/it]
 49%|████▉     | 234/480 [31:33<31:30,  7.68s/it]
 49%|████▉     | 235/480 [31:40<31:34,  7.73s/it]
 49%|████▉     | 236/480 [31:47<29:33,  7.27s/it]
 49%|████▉     | 237/480 [31:53<28:30,  7.04s/it]
 50%|████▉     | 238/480 [32:01<29:22,  7.28s/it]
 50%|████▉     | 239/480 [32:08<28:57,  7.21s/it]
 50%|█████     | 240/480 [32:15<28:57,  7.24s/it]
                                                 
{'loss': '1.18', 'grad_norm': '0.2323', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [32:15<28:57,  7.24s/it]
 50%|█████     | 241/480 [32:23<29:05,  7.31s/it]
 50%|█████     | 242/480 [32:31<30:00,  7.57s/it]
 51%|█████     | 243/480 [32:38<29:49,  7.55s/it]
 51%|█████     | 244/480 [32:46<29:35,  7.52s/it]
 51%|█████     | 245/480 [32:54<29:40,  7.58s/it]
 51%|█████▏    | 246/480 [33:01<29:30,  7.57s/it]
 51%|█████▏    | 247/480 [33:08<28:39,  7.38s/it]
 52%|█████▏    | 248/480 [33:15<28:14,  7.30s/it]
 52%|█████▏    | 249/480 [33:22<27:09,  7.06s/it]
 52%|█████▏    | 250/480 [33:29<27:38,  7.21s/it]
                                                 
{'loss': '1.134', 'grad_norm': '0.2471', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [33:29<27:38,  7.21s/it]
 52%|█████▏    | 251/480 [33:37<28:08,  7.37s/it]
 52%|█████▎    | 252/480 [33:44<27:35,  7.26s/it]
 53%|█████▎    | 253/480 [33:52<27:51,  7.36s/it]
 53%|█████▎    | 254/480 [34:00<28:28,  7.56s/it]
 53%|█████▎    | 255/480 [34:07<28:11,  7.52s/it]
 53%|█████▎    | 256/480 [34:09<21:25,  5.74s/it]
 54%|█████▎    | 257/480 [34:16<23:38,  6.36s/it]
 54%|█████▍    | 258/480 [34:25<25:48,  6.97s/it]
 54%|█████▍    | 259/480 [34:32<25:59,  7.05s/it]
 54%|█████▍    | 260/480 [34:40<26:59,  7.36s/it]
                                                 
{'loss': '1.158', 'grad_norm': '0.2826', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [34:40<26:59,  7.36s/it]
 54%|█████▍    | 261/480 [34:48<26:56,  7.38s/it]
 55%|█████▍    | 262/480 [34:56<27:45,  7.64s/it]
 55%|█████▍    | 263/480 [35:05<28:47,  7.96s/it]
 55%|█████▌    | 264/480 [35:12<28:16,  7.85s/it]
 55%|█████▌    | 265/480 [35:19<27:22,  7.64s/it]
 55%|█████▌    | 266/480 [35:27<27:37,  7.75s/it]
 56%|█████▌    | 267/480 [35:35<27:10,  7.65s/it]
 56%|█████▌    | 268/480 [35:42<26:31,  7.51s/it]
 56%|█████▌    | 269/480 [35:49<26:06,  7.43s/it]
 56%|█████▋    | 270/480 [35:56<24:59,  7.14s/it]
                                                 
{'loss': '1.189', 'grad_norm': '0.4903', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [35:56<24:59,  7.14s/it]
 56%|█████▋    | 271/480 [36:03<25:14,  7.25s/it]
 57%|█████▋    | 272/480 [36:11<25:30,  7.36s/it]
 57%|█████▋    | 273/480 [36:18<25:12,  7.31s/it]
 57%|█████▋    | 274/480 [36:26<25:35,  7.45s/it]
 57%|█████▋    | 275/480 [36:33<25:34,  7.48s/it]
 57%|█████▊    | 276/480 [36:41<25:41,  7.56s/it]
 58%|█████▊    | 277/480 [36:48<24:44,  7.31s/it]
 58%|█████▊    | 278/480 [36:55<24:42,  7.34s/it]
 58%|█████▊    | 279/480 [37:02<24:15,  7.24s/it]
 58%|█████▊    | 280/480 [37:10<24:49,  7.45s/it]
                                                 
{'loss': '1.119', 'grad_norm': '0.4958', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [37:10<24:49,  7.45s/it]
 59%|█████▊    | 281/480 [37:17<24:08,  7.28s/it]
 59%|█████▉    | 282/480 [37:24<23:51,  7.23s/it]
 59%|█████▉    | 283/480 [37:31<23:38,  7.20s/it]
 59%|█████▉    | 284/480 [37:39<24:05,  7.37s/it]
 59%|█████▉    | 285/480 [37:46<24:05,  7.41s/it]
 60%|█████▉    | 286/480 [37:53<23:32,  7.28s/it]
 60%|█████▉    | 287/480 [38:02<24:20,  7.57s/it]
 60%|██████    | 288/480 [38:03<18:36,  5.82s/it]
 60%|██████    | 289/480 [38:11<20:03,  6.30s/it]
 60%|██████    | 290/480 [38:19<22:00,  6.95s/it]
                                                 
{'loss': '1.162', 'grad_norm': '0.3802', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [38:19<22:00,  6.95s/it]
 61%|██████    | 291/480 [38:27<22:51,  7.26s/it]
 61%|██████    | 292/480 [38:34<22:16,  7.11s/it]
 61%|██████    | 293/480 [38:42<22:42,  7.29s/it]
 61%|██████▏   | 294/480 [38:51<24:04,  7.76s/it]
 61%|██████▏   | 295/480 [38:58<23:15,  7.54s/it]
 62%|██████▏   | 296/480 [39:05<23:17,  7.60s/it]
 62%|██████▏   | 297/480 [39:13<23:16,  7.63s/it]
 62%|██████▏   | 298/480 [39:21<23:41,  7.81s/it]
 62%|██████▏   | 299/480 [39:29<23:37,  7.83s/it]
 62%|██████▎   | 300/480 [39:36<22:51,  7.62s/it]
                                                 
{'loss': '1.177', 'grad_norm': '0.4883', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [39:36<22:51,  7.62s/it]
 63%|██████▎   | 301/480 [39:44<22:22,  7.50s/it]
 63%|██████▎   | 302/480 [39:52<22:54,  7.72s/it]
 63%|██████▎   | 303/480 [39:59<22:10,  7.52s/it]
 63%|██████▎   | 304/480 [40:06<21:48,  7.44s/it]
 64%|██████▎   | 305/480 [40:14<21:54,  7.51s/it]
 64%|██████▍   | 306/480 [40:21<21:47,  7.51s/it]
 64%|██████▍   | 307/480 [40:29<22:13,  7.71s/it]
 64%|██████▍   | 308/480 [40:36<21:19,  7.44s/it]
 64%|██████▍   | 309/480 [40:43<20:23,  7.15s/it]
 65%|██████▍   | 310/480 [40:51<20:59,  7.41s/it]
                                                 
{'loss': '1.113', 'grad_norm': '0.2724', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [40:51<20:59,  7.41s/it]
 65%|██████▍   | 311/480 [40:59<21:23,  7.60s/it]
 65%|██████▌   | 312/480 [41:06<21:04,  7.53s/it]
 65%|██████▌   | 313/480 [41:13<20:36,  7.41s/it]
 65%|██████▌   | 314/480 [41:21<20:25,  7.38s/it]
 66%|██████▌   | 315/480 [41:27<19:29,  7.09s/it]
 66%|██████▌   | 316/480 [41:35<19:52,  7.27s/it]
 66%|██████▌   | 317/480 [41:42<19:28,  7.17s/it]
 66%|██████▋   | 318/480 [41:49<19:24,  7.19s/it]
 66%|██████▋   | 319/480 [41:56<19:08,  7.13s/it]
 67%|██████▋   | 320/480 [41:58<15:11,  5.70s/it]
                                                 
{'loss': '1.139', 'grad_norm': '1.224', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [41:58<15:11,  5.70s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:02<04:03,  2.46s/it]
+
Eval @ step 320:   2%|▏         | 2/100 [00:04<03:31,  2.16s/it]
+
Eval @ step 320:   3%|▎         | 3/100 [00:06<03:43,  2.30s/it]
+
Eval @ step 320:   4%|▍         | 4/100 [00:07<02:47,  1.74s/it]
+
Eval @ step 320:   5%|▌         | 5/100 [00:07<01:50,  1.16s/it]
+
Eval @ step 320:   6%|▌         | 6/100 [00:09<02:15,  1.44s/it]
+
Eval @ step 320:   7%|▋         | 7/100 [00:10<01:56,  1.25s/it]
+
Eval @ step 320:   8%|▊         | 8/100 [00:12<01:59,  1.30s/it]
+
Eval @ step 320:   9%|▉         | 9/100 [00:14<02:30,  1.65s/it]
+
Eval @ step 320:  10%|█         | 10/100 [00:16<02:22,  1.59s/it]
+
Eval @ step 320:  11%|█         | 11/100 [00:17<02:12,  1.49s/it]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:18<01:59,  1.35s/it]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:19<02:02,  1.41s/it]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:22<02:29,  1.74s/it]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:25<02:51,  2.02s/it]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:26<02:40,  1.91s/it]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:27<02:18,  1.67s/it]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:28<01:44,  1.28s/it]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:29<01:35,  1.17s/it]
+
Eval @ step 320:  20%|██        | 20/100 [00:30<01:42,  1.28s/it]
+
Eval @ step 320:  21%|██        | 21/100 [00:32<01:57,  1.49s/it]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:35<02:29,  1.92s/it]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:37<02:22,  1.85s/it]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:40<03:02,  2.40s/it]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:42<02:35,  2.08s/it]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:43<02:06,  1.71s/it]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:45<02:22,  1.96s/it]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:46<01:57,  1.64s/it]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:48<01:54,  1.62s/it]
+
Eval @ step 320:  30%|███       | 30/100 [00:49<01:47,  1.54s/it]
+
Eval @ step 320:  31%|███       | 31/100 [00:51<01:49,  1.59s/it]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:53<02:10,  1.92s/it]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:55<02:04,  1.86s/it]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:57<01:54,  1.74s/it]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:59<02:03,  1.91s/it]
+
Eval @ step 320:  36%|███▌      | 36/100 [01:00<01:56,  1.82s/it]
+
Eval @ step 320:  37%|███▋      | 37/100 [01:02<01:55,  1.83s/it]
+
Eval @ step 320:  38%|███▊      | 38/100 [01:03<01:30,  1.46s/it]
+
Eval @ step 320:  39%|███▉      | 39/100 [01:05<01:43,  1.70s/it]
+
Eval @ step 320:  40%|████      | 40/100 [01:06<01:32,  1.55s/it]
+
Eval @ step 320:  41%|████      | 41/100 [01:07<01:21,  1.38s/it]
+
Eval @ step 320:  42%|████▏     | 42/100 [01:09<01:25,  1.48s/it]
+
Eval @ step 320:  43%|████▎     | 43/100 [01:10<01:17,  1.36s/it]
+
Eval @ step 320:  44%|████▍     | 44/100 [01:12<01:19,  1.42s/it]
+
Eval @ step 320:  45%|████▌     | 45/100 [01:13<01:19,  1.44s/it]
+
Eval @ step 320:  46%|████▌     | 46/100 [01:15<01:17,  1.43s/it]
+
Eval @ step 320:  47%|████▋     | 47/100 [01:15<01:04,  1.22s/it]
+
Eval @ step 320:  48%|████▊     | 48/100 [01:16<01:03,  1.22s/it]
+
Eval @ step 320:  49%|████▉     | 49/100 [01:17<00:58,  1.14s/it]
+
Eval @ step 320:  50%|█████     | 50/100 [01:19<01:02,  1.25s/it]
+
Eval @ step 320:  51%|█████     | 51/100 [01:20<00:54,  1.11s/it]
+
Eval @ step 320:  52%|█████▏    | 52/100 [01:21<00:49,  1.04s/it]
+
Eval @ step 320:  53%|█████▎    | 53/100 [01:22<00:58,  1.24s/it]
+
Eval @ step 320:  54%|█████▍    | 54/100 [01:23<00:49,  1.07s/it]
+
Eval @ step 320:  55%|█████▌    | 55/100 [01:24<00:51,  1.14s/it]
+
Eval @ step 320:  56%|█████▌    | 56/100 [01:26<01:01,  1.39s/it]
+
Eval @ step 320:  57%|█████▋    | 57/100 [01:28<01:03,  1.47s/it]
+
Eval @ step 320:  58%|█████▊    | 58/100 [01:29<00:51,  1.24s/it]
+
Eval @ step 320:  59%|█████▉    | 59/100 [01:29<00:43,  1.06s/it]
+
Eval @ step 320:  60%|██████    | 60/100 [01:30<00:38,  1.04it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [01:31<00:33,  1.17it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [01:34<00:56,  1.49s/it]
+
Eval @ step 320:  63%|██████▎   | 63/100 [01:34<00:39,  1.08s/it]
+
Eval @ step 320:  64%|██████▍   | 64/100 [01:34<00:32,  1.11it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [01:36<00:40,  1.14s/it]
+
Eval @ step 320:  66%|██████▌   | 66/100 [01:37<00:35,  1.03s/it]
+
Eval @ step 320:  67%|██████▋   | 67/100 [01:39<00:44,  1.35s/it]
+
Eval @ step 320:  68%|██████▊   | 68/100 [01:40<00:43,  1.36s/it]
+
Eval @ step 320:  69%|██████▉   | 69/100 [01:43<00:54,  1.76s/it]
+
Eval @ step 320:  70%|███████   | 70/100 [01:44<00:50,  1.69s/it]
+
Eval @ step 320:  71%|███████   | 71/100 [01:45<00:40,  1.41s/it]
+
Eval @ step 320:  72%|███████▏  | 72/100 [01:47<00:43,  1.57s/it]
+
Eval @ step 320:  73%|███████▎  | 73/100 [01:50<00:51,  1.90s/it]
+
Eval @ step 320:  74%|███████▍  | 74/100 [01:52<00:50,  1.93s/it]
+
Eval @ step 320:  75%|███████▌  | 75/100 [01:56<01:03,  2.52s/it]
+
Eval @ step 320:  76%|███████▌  | 76/100 [01:57<00:55,  2.33s/it]
+
Eval @ step 320:  77%|███████▋  | 77/100 [01:59<00:50,  2.20s/it]
+
Eval @ step 320:  78%|███████▊  | 78/100 [02:03<00:58,  2.66s/it]
+
Eval @ step 320:  79%|███████▉  | 79/100 [02:04<00:42,  2.04s/it]
+
Eval @ step 320:  80%|████████  | 80/100 [02:05<00:33,  1.69s/it]
+
Eval @ step 320:  81%|████████  | 81/100 [02:06<00:30,  1.58s/it]
+
Eval @ step 320:  82%|████████▏ | 82/100 [02:08<00:30,  1.67s/it]
+
Eval @ step 320:  83%|████████▎ | 83/100 [02:10<00:30,  1.82s/it]
+
Eval @ step 320:  84%|████████▍ | 84/100 [02:11<00:27,  1.73s/it]
+
Eval @ step 320:  85%|████████▌ | 85/100 [02:14<00:28,  1.91s/it]
+
Eval @ step 320:  86%|████████▌ | 86/100 [02:16<00:26,  1.89s/it]
+
Eval @ step 320:  87%|████████▋ | 87/100 [02:18<00:26,  2.06s/it]
+
Eval @ step 320:  88%|████████▊ | 88/100 [02:21<00:28,  2.40s/it]
+
Eval @ step 320:  89%|████████▉ | 89/100 [02:23<00:22,  2.07s/it]
+
Eval @ step 320:  90%|█████████ | 90/100 [02:27<00:27,  2.75s/it]
+
Eval @ step 320:  91%|█████████ | 91/100 [02:28<00:20,  2.30s/it]
+
Eval @ step 320:  92%|█████████▏| 92/100 [02:30<00:16,  2.07s/it]
+
Eval @ step 320:  93%|█████████▎| 93/100 [02:32<00:15,  2.15s/it]
+
Eval @ step 320:  94%|█████████▍| 94/100 [02:34<00:12,  2.00s/it]
+
Eval @ step 320:  95%|█████████▌| 95/100 [02:34<00:07,  1.52s/it]
+
Eval @ step 320:  96%|█████████▌| 96/100 [02:36<00:06,  1.51s/it]
+
Eval @ step 320:  97%|█████████▋| 97/100 [02:36<00:03,  1.33s/it]
+
Eval @ step 320:  98%|█████████▊| 98/100 [02:38<00:02,  1.28s/it]
+
Eval @ step 320:  99%|█████████▉| 99/100 [02:39<00:01,  1.33s/it]
+
Eval @ step 320: 100%|██████████| 100/100 [02:41<00:00,  1.37s/it]
+
                                                                    [EVAL] ★ New best eval_rougeL=29.2370 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '29.24', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [44:39<15:11,  5.70s/it]
 67%|██████▋   | 321/480 [44:47<2:24:47, 54.64s/it]
 67%|██████▋   | 322/480 [44:56<1:47:29, 40.82s/it]
 67%|██████▋   | 323/480 [45:03<1:20:49, 30.89s/it]
 68%|██████▊   | 324/480 [45:11<1:02:26, 24.02s/it]
 68%|██████▊   | 325/480 [45:19<49:03, 18.99s/it]  
 68%|██████▊   | 326/480 [45:26<39:45, 15.49s/it]
 68%|██████▊   | 327/480 [45:34<33:52, 13.28s/it]
 68%|██████▊   | 328/480 [45:42<29:38, 11.70s/it]
 69%|██████▊   | 329/480 [45:49<25:49, 10.26s/it]
 69%|██████▉   | 330/480 [45:56<23:11,  9.28s/it]
                                                 
{'loss': '1.134', 'grad_norm': '0.3169', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [45:56<23:11,  9.28s/it]
 69%|██████▉   | 331/480 [46:03<21:22,  8.61s/it]
 69%|██████▉   | 332/480 [46:11<20:40,  8.38s/it]
 69%|██████▉   | 333/480 [46:18<19:43,  8.05s/it]
 70%|██████▉   | 334/480 [46:25<18:34,  7.63s/it]
 70%|██████▉   | 335/480 [46:33<18:39,  7.72s/it]
 70%|███████   | 336/480 [46:40<18:17,  7.62s/it]
 70%|███████   | 337/480 [46:48<18:15,  7.66s/it]
 70%|███████   | 338/480 [46:55<17:51,  7.54s/it]
 71%|███████   | 339/480 [47:02<17:20,  7.38s/it]
 71%|███████   | 340/480 [47:10<17:40,  7.57s/it]
                                                 
{'loss': '1.105', 'grad_norm': '0.4347', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [47:10<17:40,  7.57s/it]
 71%|███████   | 341/480 [47:17<16:44,  7.22s/it]
 71%|███████▏  | 342/480 [47:24<16:31,  7.19s/it]
 71%|███████▏  | 343/480 [47:31<16:28,  7.22s/it]
 72%|███████▏  | 344/480 [47:39<17:00,  7.51s/it]
 72%|███████▏  | 345/480 [47:46<16:42,  7.42s/it]
 72%|███████▏  | 346/480 [47:53<16:20,  7.32s/it]
 72%|███████▏  | 347/480 [48:02<17:12,  7.76s/it]
 72%|███████▎  | 348/480 [48:10<17:02,  7.75s/it]
 73%|███████▎  | 349/480 [48:18<16:55,  7.75s/it]
 73%|███████▎  | 350/480 [48:26<17:18,  7.99s/it]
                                                 
{'loss': '1.134', 'grad_norm': '0.3389', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [48:26<17:18,  7.99s/it]
 73%|███████▎  | 351/480 [48:34<16:54,  7.86s/it]
 73%|███████▎  | 352/480 [48:36<12:52,  6.04s/it]
 74%|███████▎  | 353/480 [48:44<14:03,  6.64s/it]
 74%|███████▍  | 354/480 [48:52<14:51,  7.08s/it]
 74%|███████▍  | 355/480 [48:59<14:47,  7.10s/it]
 74%|███████▍  | 356/480 [49:06<14:50,  7.18s/it]
 74%|███████▍  | 357/480 [49:14<15:05,  7.36s/it]
 75%|███████▍  | 358/480 [49:21<14:37,  7.19s/it]
 75%|███████▍  | 359/480 [49:29<14:49,  7.35s/it]
 75%|███████▌  | 360/480 [49:35<14:06,  7.06s/it]
                                                 
{'loss': '1.079', 'grad_norm': '0.232', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [49:35<14:06,  7.06s/it]
 75%|███████▌  | 361/480 [49:42<14:09,  7.14s/it]
 75%|███████▌  | 362/480 [49:50<14:34,  7.41s/it]
 76%|███████▌  | 363/480 [49:58<14:33,  7.47s/it]
 76%|███████▌  | 364/480 [50:06<14:43,  7.62s/it]
 76%|███████▌  | 365/480 [50:13<14:20,  7.48s/it]
 76%|███████▋  | 366/480 [50:21<14:21,  7.56s/it]
 76%|███████▋  | 367/480 [50:29<14:28,  7.68s/it]
 77%|███████▋  | 368/480 [50:35<13:45,  7.37s/it]
 77%|███████▋  | 369/480 [50:43<13:51,  7.49s/it]
 77%|███████▋  | 370/480 [50:51<13:45,  7.50s/it]
                                                 
{'loss': '1.15', 'grad_norm': '0.3142', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [50:51<13:45,  7.50s/it]
 77%|███████▋  | 371/480 [50:58<13:30,  7.44s/it]
 78%|███████▊  | 372/480 [51:06<13:32,  7.52s/it]
 78%|███████▊  | 373/480 [51:14<13:57,  7.82s/it]
 78%|███████▊  | 374/480 [51:23<14:05,  7.98s/it]
 78%|███████▊  | 375/480 [51:29<13:11,  7.54s/it]
 78%|███████▊  | 376/480 [51:36<12:47,  7.38s/it]
 79%|███████▊  | 377/480 [51:43<12:40,  7.38s/it]
 79%|███████▉  | 378/480 [51:51<12:40,  7.45s/it]
 79%|███████▉  | 379/480 [51:59<12:45,  7.58s/it]
 79%|███████▉  | 380/480 [52:06<12:35,  7.56s/it]
                                                 
{'loss': '1.077', 'grad_norm': '0.4809', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [52:06<12:35,  7.56s/it]
 79%|███████▉  | 381/480 [52:13<12:08,  7.36s/it]
 80%|███████▉  | 382/480 [52:21<12:11,  7.46s/it]
 80%|███████▉  | 383/480 [52:28<12:02,  7.45s/it]
 80%|████████  | 384/480 [52:30<09:09,  5.72s/it]
 80%|████████  | 385/480 [52:39<10:33,  6.67s/it]
 80%|████████  | 386/480 [52:47<11:00,  7.03s/it]
 81%|████████  | 387/480 [52:55<11:24,  7.36s/it]
 81%|████████  | 388/480 [53:03<11:42,  7.64s/it]
 81%|████████  | 389/480 [53:10<11:09,  7.36s/it]
 81%|████████▏ | 390/480 [53:17<10:57,  7.31s/it]
                                                 
{'loss': '1.134', 'grad_norm': '0.3419', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [53:17<10:57,  7.31s/it]
 81%|████████▏ | 391/480 [53:24<10:45,  7.25s/it]
 82%|████████▏ | 392/480 [53:32<10:40,  7.28s/it]
 82%|████████▏ | 393/480 [53:39<10:39,  7.35s/it]
 82%|████████▏ | 394/480 [53:46<10:18,  7.19s/it]
 82%|████████▏ | 395/480 [53:53<10:14,  7.23s/it]
 82%|████████▎ | 396/480 [54:00<10:04,  7.20s/it]
 83%|████████▎ | 397/480 [54:07<09:52,  7.14s/it]
 83%|████████▎ | 398/480 [54:15<10:04,  7.37s/it]
 83%|████████▎ | 399/480 [54:22<09:47,  7.25s/it]
 83%|████████▎ | 400/480 [54:30<09:38,  7.23s/it]
                                                 
{'loss': '1.091', 'grad_norm': '0.3714', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [54:30<09:38,  7.23s/it]
 84%|████████▎ | 401/480 [54:38<09:51,  7.49s/it]
 84%|████████▍ | 402/480 [54:46<10:10,  7.83s/it]
 84%|████████▍ | 403/480 [54:54<09:59,  7.78s/it]
 84%|████████▍ | 404/480 [55:00<09:20,  7.38s/it]
 84%|████████▍ | 405/480 [55:08<09:18,  7.45s/it]
 85%|████████▍ | 406/480 [55:16<09:18,  7.55s/it]
 85%|████████▍ | 407/480 [55:23<08:54,  7.32s/it]
 85%|████████▌ | 408/480 [55:30<08:43,  7.28s/it]
 85%|████████▌ | 409/480 [55:37<08:27,  7.15s/it]
 85%|████████▌ | 410/480 [55:44<08:24,  7.21s/it]
                                                 
{'loss': '1.086', 'grad_norm': '0.3688', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [55:44<08:24,  7.21s/it]
 86%|████████▌ | 411/480 [55:52<08:34,  7.46s/it]
 86%|████████▌ | 412/480 [55:59<08:25,  7.43s/it]
 86%|████████▌ | 413/480 [56:07<08:19,  7.46s/it]
 86%|████████▋ | 414/480 [56:14<08:15,  7.51s/it]
 86%|████████▋ | 415/480 [56:23<08:22,  7.73s/it]
 87%|████████▋ | 416/480 [56:25<06:32,  6.13s/it]
 87%|████████▋ | 417/480 [56:33<06:55,  6.59s/it]
 87%|████████▋ | 418/480 [56:40<06:54,  6.69s/it]
 87%|████████▋ | 419/480 [56:47<06:58,  6.86s/it]
 88%|████████▊ | 420/480 [56:56<07:23,  7.39s/it]
                                                 
{'loss': '1.087', 'grad_norm': '0.3823', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [56:56<07:23,  7.39s/it]
 88%|████████▊ | 421/480 [57:04<07:30,  7.64s/it]
 88%|████████▊ | 422/480 [57:12<07:33,  7.81s/it]
 88%|████████▊ | 423/480 [57:19<07:15,  7.64s/it]
 88%|████████▊ | 424/480 [57:27<07:07,  7.64s/it]
 89%|████████▊ | 425/480 [57:34<06:49,  7.45s/it]
 89%|████████▉ | 426/480 [57:43<07:02,  7.82s/it]
 89%|████████▉ | 427/480 [57:50<06:54,  7.82s/it]
 89%|████████▉ | 428/480 [57:58<06:49,  7.87s/it]
 89%|████████▉ | 429/480 [58:05<06:27,  7.61s/it]
 90%|████████▉ | 430/480 [58:13<06:17,  7.55s/it]
                                                 
{'loss': '1.099', 'grad_norm': '0.6109', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [58:13<06:17,  7.55s/it]
 90%|████████▉ | 431/480 [58:20<06:11,  7.57s/it]
 90%|█████████ | 432/480 [58:27<05:53,  7.36s/it]
 90%|█████████ | 433/480 [58:35<05:52,  7.50s/it]
 90%|█████████ | 434/480 [58:42<05:33,  7.25s/it]
 91%|█████████ | 435/480 [58:49<05:28,  7.29s/it]
 91%|█████████ | 436/480 [58:56<05:17,  7.22s/it]
 91%|█████████ | 437/480 [59:04<05:15,  7.34s/it]
 91%|█████████▏| 438/480 [59:11<05:06,  7.29s/it]
 91%|█████████▏| 439/480 [59:19<05:10,  7.58s/it]
 92%|█████████▏| 440/480 [59:27<05:05,  7.63s/it]
                                                 
{'loss': '1.068', 'grad_norm': '0.2996', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [59:27<05:05,  7.63s/it]
 92%|█████████▏| 441/480 [59:34<04:47,  7.38s/it]
 92%|█████████▏| 442/480 [59:40<04:29,  7.08s/it]
 92%|█████████▏| 443/480 [59:47<04:21,  7.08s/it]
 92%|█████████▎| 444/480 [59:55<04:20,  7.22s/it]
 93%|█████████▎| 445/480 [1:00:03<04:18,  7.39s/it]
 93%|█████████▎| 446/480 [1:00:10<04:12,  7.43s/it]
 93%|█████████▎| 447/480 [1:00:18<04:05,  7.43s/it]
 93%|█████████▎| 448/480 [1:00:20<03:06,  5.83s/it]
 94%|█████████▎| 449/480 [1:00:28<03:21,  6.49s/it]
 94%|█████████▍| 450/480 [1:00:35<03:21,  6.71s/it]
                                                   
{'loss': '1.071', 'grad_norm': '0.3149', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [1:00:35<03:21,  6.71s/it]
 94%|█████████▍| 451/480 [1:00:43<03:25,  7.08s/it]
 94%|█████████▍| 452/480 [1:00:49<03:12,  6.88s/it]
 94%|█████████▍| 453/480 [1:00:57<03:10,  7.06s/it]
 95%|█████████▍| 454/480 [1:01:03<03:00,  6.95s/it]
 95%|█████████▍| 455/480 [1:01:12<03:04,  7.40s/it]
 95%|█████████▌| 456/480 [1:01:19<02:55,  7.32s/it]
 95%|█████████▌| 457/480 [1:01:27<02:50,  7.43s/it]
 95%|█████████▌| 458/480 [1:01:34<02:42,  7.38s/it]
 96%|█████████▌| 459/480 [1:01:42<02:35,  7.41s/it]
 96%|█████████▌| 460/480 [1:01:50<02:33,  7.68s/it]
                                                   
{'loss': '1.097', 'grad_norm': '0.3728', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [1:01:50<02:33,  7.68s/it]
 96%|█████████▌| 461/480 [1:01:58<02:27,  7.79s/it]
 96%|█████████▋| 462/480 [1:02:05<02:15,  7.53s/it]
 96%|█████████▋| 463/480 [1:02:12<02:04,  7.35s/it]
 97%|█████████▋| 464/480 [1:02:19<01:57,  7.32s/it]
 97%|█████████▋| 465/480 [1:02:25<01:43,  6.89s/it]
 97%|█████████▋| 466/480 [1:02:32<01:36,  6.89s/it]
 97%|█████████▋| 467/480 [1:02:39<01:30,  6.97s/it]
 98%|█████████▊| 468/480 [1:02:46<01:24,  7.01s/it]
 98%|█████████▊| 469/480 [1:02:54<01:21,  7.41s/it]
 98%|█████████▊| 470/480 [1:03:02<01:14,  7.48s/it]
                                                   
{'loss': '1.039', 'grad_norm': '0.3827', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [1:03:02<01:14,  7.48s/it]
 98%|█████████▊| 471/480 [1:03:10<01:07,  7.49s/it]
 98%|█████████▊| 472/480 [1:03:17<01:00,  7.61s/it]
 99%|█████████▊| 473/480 [1:03:25<00:52,  7.56s/it]
 99%|█████████▉| 474/480 [1:03:32<00:45,  7.53s/it]
 99%|█████████▉| 475/480 [1:03:40<00:37,  7.57s/it]
 99%|█████████▉| 476/480 [1:03:47<00:29,  7.37s/it]
 99%|█████████▉| 477/480 [1:03:55<00:22,  7.66s/it]
100%|█████████▉| 478/480 [1:04:04<00:16,  8.10s/it]
100%|█████████▉| 479/480 [1:04:13<00:08,  8.15s/it]
100%|██████████| 480/480 [1:04:15<00:00,  6.31s/it]
                                                   
{'loss': '1.085', 'grad_norm': '1.477', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [1:04:15<00:00,  6.31s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:02<03:53,  2.36s/it]
+
Eval @ step 480:   2%|▏         | 2/100 [00:04<03:40,  2.25s/it]
+
Eval @ step 480:   3%|▎         | 3/100 [00:07<04:05,  2.53s/it]
+
Eval @ step 480:   4%|▍         | 4/100 [00:08<03:12,  2.01s/it]
+
Eval @ step 480:   5%|▌         | 5/100 [00:11<03:33,  2.25s/it]
+
Eval @ step 480:   6%|▌         | 6/100 [00:13<03:38,  2.32s/it]
+
Eval @ step 480:   7%|▋         | 7/100 [00:14<02:51,  1.84s/it]
+
Eval @ step 480:   8%|▊         | 8/100 [00:15<02:36,  1.70s/it]
+
Eval @ step 480:   9%|▉         | 9/100 [00:17<02:37,  1.73s/it]
+
Eval @ step 480:  10%|█         | 10/100 [00:19<02:30,  1.68s/it]
+
Eval @ step 480:  11%|█         | 11/100 [00:20<02:14,  1.51s/it]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:21<02:08,  1.46s/it]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:23<02:09,  1.49s/it]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:26<02:51,  1.99s/it]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:29<03:07,  2.20s/it]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:30<02:50,  2.04s/it]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:32<02:26,  1.77s/it]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:32<01:50,  1.35s/it]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:33<01:31,  1.13s/it]
+
Eval @ step 480:  20%|██        | 20/100 [00:34<01:35,  1.20s/it]
+
Eval @ step 480:  21%|██        | 21/100 [00:37<02:26,  1.85s/it]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:40<02:38,  2.03s/it]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:41<02:28,  1.93s/it]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:47<03:43,  2.94s/it]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:49<03:24,  2.73s/it]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:50<02:41,  2.19s/it]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:52<02:42,  2.23s/it]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:53<02:16,  1.89s/it]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:55<02:09,  1.82s/it]
+
Eval @ step 480:  30%|███       | 30/100 [00:56<01:51,  1.59s/it]
+
Eval @ step 480:  31%|███       | 31/100 [00:58<01:51,  1.62s/it]
+
Eval @ step 480:  32%|███▏      | 32/100 [01:01<02:17,  2.02s/it]
+
Eval @ step 480:  33%|███▎      | 33/100 [01:02<02:00,  1.79s/it]
+
Eval @ step 480:  34%|███▍      | 34/100 [01:04<02:11,  2.00s/it]
+
Eval @ step 480:  35%|███▌      | 35/100 [01:06<02:09,  1.99s/it]
+
Eval @ step 480:  36%|███▌      | 36/100 [01:08<02:04,  1.94s/it]
+
Eval @ step 480:  37%|███▋      | 37/100 [01:10<02:01,  1.93s/it]
+
Eval @ step 480:  38%|███▊      | 38/100 [01:11<01:33,  1.50s/it]
+
Eval @ step 480:  39%|███▉      | 39/100 [01:14<02:09,  2.12s/it]
+
Eval @ step 480:  40%|████      | 40/100 [01:15<01:49,  1.83s/it]
+
Eval @ step 480:  41%|████      | 41/100 [01:17<01:37,  1.66s/it]
+
Eval @ step 480:  42%|████▏     | 42/100 [01:19<01:54,  1.97s/it]
+
Eval @ step 480:  43%|████▎     | 43/100 [01:20<01:37,  1.72s/it]
+
Eval @ step 480:  44%|████▍     | 44/100 [01:22<01:26,  1.55s/it]
+
Eval @ step 480:  45%|████▌     | 45/100 [01:23<01:25,  1.56s/it]
+
Eval @ step 480:  46%|████▌     | 46/100 [01:24<01:16,  1.42s/it]
+
Eval @ step 480:  47%|████▋     | 47/100 [01:25<01:04,  1.21s/it]
+
Eval @ step 480:  48%|████▊     | 48/100 [01:26<01:01,  1.19s/it]
+
Eval @ step 480:  49%|████▉     | 49/100 [01:27<00:55,  1.09s/it]
+
Eval @ step 480:  50%|█████     | 50/100 [01:28<01:01,  1.23s/it]
+
Eval @ step 480:  51%|█████     | 51/100 [01:29<00:54,  1.12s/it]
+
Eval @ step 480:  52%|█████▏    | 52/100 [01:31<00:56,  1.17s/it]
+
Eval @ step 480:  53%|█████▎    | 53/100 [01:33<01:09,  1.48s/it]
+
Eval @ step 480:  54%|█████▍    | 54/100 [01:33<00:55,  1.21s/it]
+
Eval @ step 480:  55%|█████▌    | 55/100 [01:35<01:00,  1.35s/it]
+
Eval @ step 480:  56%|█████▌    | 56/100 [01:37<01:00,  1.38s/it]
+
Eval @ step 480:  57%|█████▋    | 57/100 [01:38<00:59,  1.38s/it]
+
Eval @ step 480:  58%|█████▊    | 58/100 [01:39<00:48,  1.16s/it]
+
Eval @ step 480:  59%|█████▉    | 59/100 [01:39<00:40,  1.02it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [01:40<00:36,  1.11it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [01:41<00:33,  1.17it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [01:43<00:50,  1.32s/it]
+
Eval @ step 480:  63%|██████▎   | 63/100 [01:43<00:35,  1.04it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [01:44<00:30,  1.17it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [01:45<00:38,  1.11s/it]
+
Eval @ step 480:  66%|██████▌   | 66/100 [01:46<00:31,  1.08it/s]
+
Eval @ step 480:  67%|██████▋   | 67/100 [01:48<00:38,  1.15s/it]
+
Eval @ step 480:  68%|██████▊   | 68/100 [01:49<00:39,  1.22s/it]
+
Eval @ step 480:  69%|██████▉   | 69/100 [01:52<00:50,  1.64s/it]
+
Eval @ step 480:  70%|███████   | 70/100 [01:53<00:49,  1.64s/it]
+
Eval @ step 480:  71%|███████   | 71/100 [01:54<00:39,  1.35s/it]
+
Eval @ step 480:  72%|███████▏  | 72/100 [01:56<00:40,  1.44s/it]
+
Eval @ step 480:  73%|███████▎  | 73/100 [01:58<00:48,  1.81s/it]
+
Eval @ step 480:  74%|███████▍  | 74/100 [02:01<00:52,  2.01s/it]
+
Eval @ step 480:  75%|███████▌  | 75/100 [02:04<00:55,  2.24s/it]
+
Eval @ step 480:  76%|███████▌  | 76/100 [02:05<00:49,  2.08s/it]
+
Eval @ step 480:  77%|███████▋  | 77/100 [02:07<00:45,  1.96s/it]
+
Eval @ step 480:  78%|███████▊  | 78/100 [02:11<00:56,  2.55s/it]
+
Eval @ step 480:  79%|███████▉  | 79/100 [02:11<00:41,  1.97s/it]
+
Eval @ step 480:  80%|████████  | 80/100 [02:12<00:31,  1.59s/it]
+
Eval @ step 480:  81%|████████  | 81/100 [02:13<00:28,  1.50s/it]
+
Eval @ step 480:  82%|████████▏ | 82/100 [02:15<00:27,  1.54s/it]
+
Eval @ step 480:  83%|████████▎ | 83/100 [02:18<00:31,  1.87s/it]
+
Eval @ step 480:  84%|████████▍ | 84/100 [02:19<00:28,  1.77s/it]
+
Eval @ step 480:  85%|████████▌ | 85/100 [02:23<00:35,  2.34s/it]
+
Eval @ step 480:  86%|████████▌ | 86/100 [02:24<00:28,  2.07s/it]
+
Eval @ step 480:  87%|████████▋ | 87/100 [02:26<00:26,  2.07s/it]
+
Eval @ step 480:  88%|████████▊ | 88/100 [02:30<00:29,  2.48s/it]
+
Eval @ step 480:  89%|████████▉ | 89/100 [02:31<00:23,  2.15s/it]
+
Eval @ step 480:  90%|█████████ | 90/100 [02:32<00:17,  1.78s/it]
+
Eval @ step 480:  91%|█████████ | 91/100 [02:33<00:14,  1.64s/it]
+
Eval @ step 480:  92%|█████████▏| 92/100 [02:35<00:12,  1.62s/it]
+
Eval @ step 480:  93%|█████████▎| 93/100 [02:37<00:12,  1.79s/it]
+
Eval @ step 480:  94%|█████████▍| 94/100 [02:39<00:10,  1.80s/it]
+
Eval @ step 480:  95%|█████████▌| 95/100 [02:39<00:06,  1.37s/it]
+
Eval @ step 480:  96%|█████████▌| 96/100 [02:41<00:05,  1.43s/it]
+
Eval @ step 480:  97%|█████████▋| 97/100 [02:42<00:03,  1.31s/it]
+
Eval @ step 480:  98%|█████████▊| 98/100 [02:44<00:03,  1.51s/it]
+
Eval @ step 480:  99%|█████████▉| 99/100 [02:46<00:01,  1.54s/it]
+
Eval @ step 480: 100%|██████████| 100/100 [02:48<00:00,  1.70s/it]
+
                                                                    [EVAL] ★ New best eval_rougeL=29.5925 (step=480) — snapshot saved in memory
+
                                                   
{'eval_rougeL': '29.59', 'epoch': '15'}
+
100%|██████████| 480/480 [1:07:03<00:00,  6.31s/it]
                                                   
{'train_runtime': '4023', 'train_samples_per_second': '3.728', 'train_steps_per_second': '0.119', 'train_loss': '1.184', 'epoch': '15'}
+
100%|██████████| 480/480 [1:07:03<00:00,  6.31s/it]
100%|██████████| 480/480 [1:07:03<00:00,  8.38s/it]
+***** train metrics *****
+  epoch                    =        15.0
+  total_flos               = 316866069GF
+  train_loss               =      1.1839
+  train_runtime            =  1:07:03.29
+  train_samples            =        1000
+  train_samples_per_second =       3.728
+  train_steps_per_second   =       0.119
+  [EVAL] Restored best model (eval_rougeL=29.5925, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   0%|          | 1/291 [00:00<00:35,  8.13it/s]
Loading weights:   9%|▉         | 27/291 [00:00<00:01, 142.99it/s]
Loading weights:  16%|█▌        | 47/291 [00:00<00:01, 163.75it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 161.66it/s]
Loading weights:  30%|██▉       | 86/291 [00:00<00:01, 168.77it/s]
Loading weights:  37%|███▋      | 109/291 [00:00<00:00, 184.41it/s]
Loading weights:  44%|████▍     | 128/291 [00:00<00:00, 177.99it/s]
Loading weights:  51%|█████     | 148/291 [00:00<00:00, 177.44it/s]
Loading weights:  57%|█████▋    | 167/291 [00:00<00:00, 179.08it/s]
Loading weights:  64%|██████▍   | 186/291 [00:01<00:00, 180.88it/s]
Loading weights:  72%|███████▏  | 209/291 [00:01<00:00, 185.51it/s]
Loading weights:  79%|███████▊  | 229/291 [00:01<00:00, 183.61it/s]
Loading weights:  85%|████████▌ | 248/291 [00:01<00:00, 184.31it/s]
Loading weights:  92%|█████████▏| 267/291 [00:01<00:00, 175.88it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 176.86it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task1290_xsum_summarization': PaR=106.5, shrinkage=ridge, n=1000
+  [SRT] Wired router: 3 sigs, cur=task1290_xsum_summarization, mapping={'task1290_xsum_summarization': 0, 'task1572_samsum_summary': 2, 'task363_sst2_polarity_classification': 1}
+[SRT] Router: 3 tasks, avg_PaR=106.5, metrics=['ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization']
+  [SRT-INIT] task_id_to_idx={'task1290_xsum_summarization': 0, 'task1572_samsum_summary': 2, 'task363_sst2_polarity_classification': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/220 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=3
+
Continual predict:   0%|          | 1/220 [00:01<05:48,  1.59s/it]
Continual predict:   1%|          | 2/220 [00:03<07:14,  1.99s/it]
Continual predict:   1%|▏         | 3/220 [00:04<05:02,  1.40s/it]
Continual predict:   2%|▏         | 4/220 [00:05<03:42,  1.03s/it]
Continual predict:   2%|▏         | 5/220 [00:06<04:31,  1.26s/it]
Continual predict:   3%|▎         | 6/220 [00:11<08:12,  2.30s/it]
Continual predict:   3%|▎         | 7/220 [00:14<09:03,  2.55s/it]
Continual predict:   4%|▎         | 8/220 [00:15<08:14,  2.33s/it]
Continual predict:   4%|▍         | 9/220 [00:17<07:40,  2.18s/it]
Continual predict:   5%|▍         | 10/220 [00:18<06:03,  1.73s/it]
Continual predict:   5%|▌         | 11/220 [00:20<06:47,  1.95s/it]
Continual predict:   5%|▌         | 12/220 [00:23<07:42,  2.23s/it]
Continual predict:   6%|▌         | 13/220 [00:24<06:31,  1.89s/it]
Continual predict:   6%|▋         | 14/220 [00:27<06:47,  1.98s/it]
Continual predict:   7%|▋         | 15/220 [00:29<06:51,  2.01s/it]
Continual predict:   7%|▋         | 16/220 [00:31<07:10,  2.11s/it]
Continual predict:   8%|▊         | 17/220 [00:38<12:22,  3.66s/it]
Continual predict:   8%|▊         | 18/220 [00:40<10:34,  3.14s/it]
Continual predict:   9%|▊         | 19/220 [00:43<09:42,  2.90s/it]
Continual predict:   9%|▉         | 20/220 [00:44<08:00,  2.40s/it]
Continual predict:  10%|▉         | 21/220 [00:44<05:42,  1.72s/it]
Continual predict:  10%|█         | 22/220 [00:44<04:06,  1.24s/it]
Continual predict:  10%|█         | 23/220 [00:44<02:59,  1.10it/s]
Continual predict:  11%|█         | 24/220 [00:44<02:12,  1.48it/s]
Continual predict:  11%|█▏        | 25/220 [00:44<01:39,  1.96it/s]
Continual predict:  12%|█▏        | 26/220 [00:45<01:16,  2.53it/s]
Continual predict:  12%|█▏        | 27/220 [00:45<01:00,  3.19it/s]
Continual predict:  13%|█▎        | 28/220 [00:45<00:49,  3.90it/s]
Continual predict:  13%|█▎        | 29/220 [00:45<00:41,  4.62it/s]
Continual predict:  14%|█▎        | 30/220 [00:45<00:36,  5.19it/s]
Continual predict:  14%|█▍        | 31/220 [00:45<00:32,  5.80it/s]
Continual predict:  15%|█▍        | 32/220 [00:45<00:29,  6.34it/s]
Continual predict:  15%|█▌        | 33/220 [00:45<00:28,  6.57it/s]
Continual predict:  15%|█▌        | 34/220 [00:46<00:27,  6.76it/s]
Continual predict:  16%|█▌        | 35/220 [00:46<00:26,  6.91it/s]
Continual predict:  16%|█▋        | 36/220 [00:46<00:25,  7.25it/s]
Continual predict:  17%|█▋        | 37/220 [00:46<00:25,  7.26it/s]
Continual predict:  17%|█▋        | 38/220 [00:46<00:24,  7.46it/s]
Continual predict:  18%|█▊        | 39/220 [00:46<00:23,  7.64it/s]
Continual predict:  18%|█▊        | 40/220 [00:46<00:23,  7.79it/s]
Continual predict:  19%|█▊        | 41/220 [00:47<00:23,  7.64it/s]
Continual predict:  19%|█▉        | 42/220 [00:47<00:22,  7.76it/s]
Continual predict:  20%|█▉        | 43/220 [00:47<00:23,  7.62it/s]
Continual predict:  20%|██        | 44/220 [00:47<00:23,  7.53it/s]
Continual predict:  20%|██        | 45/220 [00:47<00:22,  7.67it/s]
Continual predict:  21%|██        | 46/220 [00:47<00:22,  7.77it/s]
Continual predict:  21%|██▏       | 47/220 [00:47<00:22,  7.76it/s]
Continual predict:  22%|██▏       | 48/220 [00:47<00:22,  7.58it/s]
Continual predict:  22%|██▏       | 49/220 [00:48<00:24,  7.09it/s]
Continual predict:  23%|██▎       | 50/220 [00:48<00:23,  7.30it/s]
Continual predict:  23%|██▎       | 51/220 [00:48<00:22,  7.38it/s]
Continual predict:  24%|██▎       | 52/220 [00:48<00:23,  7.25it/s]
Continual predict:  24%|██▍       | 53/220 [00:48<00:22,  7.47it/s]
Continual predict:  25%|██▍       | 54/220 [00:48<00:22,  7.41it/s]
Continual predict:  25%|██▌       | 55/220 [00:48<00:21,  7.62it/s]
Continual predict:  25%|██▌       | 56/220 [00:49<00:21,  7.50it/s]
Continual predict:  26%|██▌       | 57/220 [00:49<00:21,  7.65it/s]
Continual predict:  26%|██▋       | 58/220 [00:49<00:20,  7.76it/s]
Continual predict:  27%|██▋       | 59/220 [00:49<00:20,  7.84it/s]
Continual predict:  27%|██▋       | 60/220 [00:49<00:20,  7.63it/s]
Continual predict:  28%|██▊       | 61/220 [00:49<00:20,  7.76it/s]
Continual predict:  28%|██▊       | 62/220 [00:49<00:20,  7.84it/s]
Continual predict:  29%|██▊       | 63/220 [00:49<00:22,  7.09it/s]
Continual predict:  29%|██▉       | 64/220 [00:50<00:21,  7.35it/s]
Continual predict:  30%|██▉       | 65/220 [00:50<00:20,  7.54it/s]
Continual predict:  30%|███       | 66/220 [00:50<00:20,  7.67it/s]
Continual predict:  30%|███       | 67/220 [00:50<00:19,  7.81it/s]
Continual predict:  31%|███       | 68/220 [00:50<00:19,  7.89it/s]
Continual predict:  31%|███▏      | 69/220 [00:50<00:19,  7.95it/s]
Continual predict:  32%|███▏      | 70/220 [00:50<00:18,  7.99it/s]
Continual predict:  32%|███▏      | 71/220 [00:50<00:18,  8.00it/s]
Continual predict:  33%|███▎      | 72/220 [00:51<00:19,  7.77it/s]
Continual predict:  33%|███▎      | 73/220 [00:51<00:18,  7.88it/s]
Continual predict:  34%|███▎      | 74/220 [00:51<00:19,  7.68it/s]
Continual predict:  34%|███▍      | 75/220 [00:51<00:18,  7.68it/s]
Continual predict:  35%|███▍      | 76/220 [00:51<00:18,  7.78it/s]
Continual predict:  35%|███▌      | 77/220 [00:51<00:18,  7.84it/s]
Continual predict:  35%|███▌      | 78/220 [00:51<00:17,  7.93it/s]
Continual predict:  36%|███▌      | 79/220 [00:51<00:17,  7.95it/s]
Continual predict:  36%|███▋      | 80/220 [00:52<00:17,  8.00it/s]
Continual predict:  37%|███▋      | 81/220 [00:52<00:17,  8.01it/s]
Continual predict:  37%|███▋      | 82/220 [00:52<00:17,  8.03it/s]
Continual predict:  38%|███▊      | 83/220 [00:52<00:17,  8.01it/s]
Continual predict:  38%|███▊      | 84/220 [00:52<00:18,  7.35it/s]
Continual predict:  39%|███▊      | 85/220 [00:52<00:17,  7.50it/s]
Continual predict:  39%|███▉      | 86/220 [00:52<00:17,  7.66it/s]
Continual predict:  40%|███▉      | 87/220 [00:53<00:17,  7.72it/s]
Continual predict:  40%|████      | 88/220 [00:53<00:17,  7.43it/s]
Continual predict:  40%|████      | 89/220 [00:53<00:17,  7.61it/s]
Continual predict:  41%|████      | 90/220 [00:53<00:16,  7.77it/s]
Continual predict:  41%|████▏     | 91/220 [00:53<00:16,  7.84it/s]
Continual predict:  42%|████▏     | 92/220 [00:53<00:16,  7.86it/s]
Continual predict:  42%|████▏     | 93/220 [00:53<00:16,  7.68it/s]
Continual predict:  43%|████▎     | 94/220 [00:53<00:16,  7.57it/s]
Continual predict:  43%|████▎     | 95/220 [00:54<00:16,  7.69it/s]
Continual predict:  44%|████▎     | 96/220 [00:54<00:15,  7.78it/s]
Continual predict:  44%|████▍     | 97/220 [00:54<00:15,  7.86it/s]
Continual predict:  45%|████▍     | 98/220 [00:54<00:15,  7.69it/s]
Continual predict:  45%|████▌     | 99/220 [00:54<00:15,  7.57it/s]
Continual predict:  45%|████▌     | 100/220 [00:54<00:16,  7.49it/s]
Continual predict:  46%|████▌     | 101/220 [00:54<00:15,  7.64it/s]
Continual predict:  46%|████▋     | 102/220 [00:54<00:15,  7.79it/s]
Continual predict:  47%|████▋     | 103/220 [00:55<00:14,  7.86it/s]
Continual predict:  47%|████▋     | 104/220 [00:55<00:14,  7.90it/s]
Continual predict:  48%|████▊     | 105/220 [00:55<00:14,  7.72it/s]
Continual predict:  48%|████▊     | 106/220 [00:55<00:14,  7.84it/s]
Continual predict:  49%|████▊     | 107/220 [00:55<00:14,  7.94it/s]
Continual predict:  49%|████▉     | 108/220 [00:55<00:14,  7.60it/s]
Continual predict:  50%|████▉     | 109/220 [00:55<00:14,  7.72it/s]
Continual predict:  50%|█████     | 110/220 [00:55<00:14,  7.80it/s]
Continual predict:  50%|█████     | 111/220 [00:56<00:13,  7.87it/s]
Continual predict:  51%|█████     | 112/220 [00:56<00:13,  7.91it/s]
Continual predict:  51%|█████▏    | 113/220 [00:56<00:13,  7.95it/s]
Continual predict:  52%|█████▏    | 114/220 [00:56<00:13,  7.97it/s]
Continual predict:  52%|█████▏    | 115/220 [00:56<00:13,  7.76it/s]
Continual predict:  53%|█████▎    | 116/220 [00:56<00:13,  7.84it/s]
Continual predict:  53%|█████▎    | 117/220 [00:56<00:13,  7.68it/s]
Continual predict:  54%|█████▎    | 118/220 [00:57<00:13,  7.55it/s]
Continual predict:  54%|█████▍    | 119/220 [00:57<00:13,  7.68it/s]
Continual predict:  55%|█████▍    | 120/220 [00:57<00:12,  7.77it/s]
Continual predict:  55%|█████▌    | 121/220 [01:01<02:16,  1.38s/it]
Continual predict:  55%|█████▌    | 122/220 [01:05<03:32,  2.17s/it]
Continual predict:  56%|█████▌    | 123/220 [01:10<04:54,  3.04s/it]
Continual predict:  56%|█████▋    | 124/220 [01:12<04:27,  2.79s/it]
Continual predict:  57%|█████▋    | 125/220 [01:17<05:16,  3.34s/it]
Continual predict:  57%|█████▋    | 126/220 [01:22<05:48,  3.71s/it]
Continual predict:  58%|█████▊    | 127/220 [01:23<04:41,  3.03s/it]
Continual predict:  58%|█████▊    | 128/220 [01:26<04:26,  2.89s/it]
Continual predict:  59%|█████▊    | 129/220 [01:29<04:32,  2.99s/it]
Continual predict:  59%|█████▉    | 130/220 [01:32<04:25,  2.94s/it]
Continual predict:  60%|█████▉    | 131/220 [01:34<03:57,  2.67s/it]
Continual predict:  60%|██████    | 132/220 [01:36<03:50,  2.62s/it]
Continual predict:  60%|██████    | 133/220 [01:39<03:57,  2.72s/it]
Continual predict:  61%|██████    | 134/220 [01:45<05:11,  3.62s/it]
Continual predict:  61%|██████▏   | 135/220 [01:49<05:30,  3.89s/it]
Continual predict:  62%|██████▏   | 136/220 [01:52<05:05,  3.63s/it]
Continual predict:  62%|██████▏   | 137/220 [01:54<04:23,  3.17s/it]
Continual predict:  63%|██████▎   | 138/220 [01:55<03:17,  2.41s/it]
Continual predict:  63%|██████▎   | 139/220 [01:56<02:43,  2.02s/it]
Continual predict:  64%|██████▎   | 140/220 [01:59<02:53,  2.16s/it]
Continual predict:  64%|██████▍   | 141/220 [02:05<04:28,  3.40s/it]
Continual predict:  65%|██████▍   | 142/220 [02:09<04:47,  3.69s/it]
Continual predict:  65%|██████▌   | 143/220 [02:12<04:30,  3.51s/it]
Continual predict:  65%|██████▌   | 144/220 [02:22<06:33,  5.18s/it]
Continual predict:  66%|██████▌   | 145/220 [02:26<06:04,  4.86s/it]
Continual predict:  66%|██████▋   | 146/220 [02:27<04:49,  3.91s/it]
Continual predict:  67%|██████▋   | 147/220 [02:32<04:51,  4.00s/it]
Continual predict:  67%|██████▋   | 148/220 [02:34<04:05,  3.40s/it]
Continual predict:  68%|██████▊   | 149/220 [02:37<03:53,  3.28s/it]
Continual predict:  68%|██████▊   | 150/220 [02:38<03:18,  2.83s/it]
Continual predict:  69%|██████▊   | 151/220 [02:41<03:17,  2.86s/it]
Continual predict:  69%|██████▉   | 152/220 [02:47<04:05,  3.61s/it]
Continual predict:  70%|██████▉   | 153/220 [02:49<03:35,  3.22s/it]
Continual predict:  70%|███████   | 154/220 [02:54<03:59,  3.63s/it]
Continual predict:  70%|███████   | 155/220 [02:57<03:51,  3.55s/it]
Continual predict:  71%|███████   | 156/220 [03:00<03:39,  3.44s/it]
Continual predict:  71%|███████▏  | 157/220 [03:04<03:39,  3.49s/it]
Continual predict:  72%|███████▏  | 158/220 [03:05<02:46,  2.69s/it]
Continual predict:  72%|███████▏  | 159/220 [03:11<03:55,  3.86s/it]
Continual predict:  73%|███████▎  | 160/220 [03:13<03:20,  3.34s/it]
Continual predict:  73%|███████▎  | 161/220 [03:16<02:59,  3.04s/it]
Continual predict:  74%|███████▎  | 162/220 [03:20<03:23,  3.51s/it]
Continual predict:  74%|███████▍  | 163/220 [03:22<02:52,  3.03s/it]
Continual predict:  75%|███████▍  | 164/220 [03:24<02:34,  2.76s/it]
Continual predict:  75%|███████▌  | 165/220 [03:27<02:34,  2.81s/it]
Continual predict:  75%|███████▌  | 166/220 [03:29<02:18,  2.57s/it]
Continual predict:  76%|███████▌  | 167/220 [03:30<01:55,  2.18s/it]
Continual predict:  76%|███████▋  | 168/220 [03:33<01:52,  2.16s/it]
Continual predict:  77%|███████▋  | 169/220 [03:34<01:41,  1.99s/it]
Continual predict:  77%|███████▋  | 170/220 [03:37<01:52,  2.26s/it]
Continual predict:  78%|███████▊  | 171/220 [03:38<01:37,  1.99s/it]
Continual predict:  78%|███████▊  | 172/220 [03:41<01:41,  2.12s/it]
Continual predict:  79%|███████▊  | 173/220 [03:45<02:02,  2.62s/it]
Continual predict:  79%|███████▉  | 174/220 [03:46<01:38,  2.15s/it]
Continual predict:  80%|███████▉  | 175/220 [03:49<01:51,  2.48s/it]
Continual predict:  80%|████████  | 176/220 [03:52<01:51,  2.54s/it]
Continual predict:  80%|████████  | 177/220 [03:54<01:49,  2.54s/it]
Continual predict:  81%|████████  | 178/220 [03:55<01:29,  2.13s/it]
Continual predict:  81%|████████▏ | 179/220 [03:56<01:13,  1.80s/it]
Continual predict:  82%|████████▏ | 180/220 [03:58<01:05,  1.63s/it]
Continual predict:  82%|████████▏ | 181/220 [03:59<00:59,  1.52s/it]
Continual predict:  83%|████████▎ | 182/220 [04:03<01:31,  2.41s/it]
Continual predict:  83%|████████▎ | 183/220 [04:04<01:04,  1.75s/it]
Continual predict:  84%|████████▎ | 184/220 [04:05<00:55,  1.54s/it]
Continual predict:  84%|████████▍ | 185/220 [04:08<01:10,  2.01s/it]
Continual predict:  85%|████████▍ | 186/220 [04:09<00:56,  1.67s/it]
Continual predict:  85%|████████▌ | 187/220 [04:12<01:09,  2.09s/it]
Continual predict:  85%|████████▌ | 188/220 [04:14<01:11,  2.24s/it]
Continual predict:  86%|████████▌ | 189/220 [04:19<01:32,  2.98s/it]
Continual predict:  86%|████████▋ | 190/220 [04:22<01:29,  2.99s/it]
Continual predict:  87%|████████▋ | 191/220 [04:23<01:11,  2.45s/it]
Continual predict:  87%|████████▋ | 192/220 [04:26<01:13,  2.64s/it]
Continual predict:  88%|████████▊ | 193/220 [04:30<01:24,  3.13s/it]
Continual predict:  88%|████████▊ | 194/220 [04:35<01:32,  3.55s/it]
Continual predict:  89%|████████▊ | 195/220 [04:40<01:39,  3.99s/it]
Continual predict:  89%|████████▉ | 196/220 [04:43<01:29,  3.75s/it]
Continual predict:  90%|████████▉ | 197/220 [04:46<01:21,  3.56s/it]
Continual predict:  90%|█████████ | 198/220 [04:53<01:41,  4.60s/it]
Continual predict:  90%|█████████ | 199/220 [04:54<01:14,  3.53s/it]
Continual predict:  91%|█████████ | 200/220 [04:56<00:57,  2.87s/it]
Continual predict:  91%|█████████▏| 201/220 [04:58<00:51,  2.71s/it]
Continual predict:  92%|█████████▏| 202/220 [05:01<00:50,  2.82s/it]
Continual predict:  92%|█████████▏| 203/220 [05:06<00:58,  3.44s/it]
Continual predict:  93%|█████████▎| 204/220 [05:09<00:52,  3.25s/it]
Continual predict:  93%|█████████▎| 205/220 [05:15<01:02,  4.16s/it]
Continual predict:  94%|█████████▎| 206/220 [05:18<00:51,  3.69s/it]
Continual predict:  94%|█████████▍| 207/220 [05:21<00:47,  3.65s/it]
Continual predict:  95%|█████████▍| 208/220 [05:27<00:51,  4.32s/it]
Continual predict:  95%|█████████▌| 209/220 [05:30<00:41,  3.78s/it]
Continual predict:  95%|█████████▌| 210/220 [05:31<00:31,  3.12s/it]
Continual predict:  96%|█████████▌| 211/220 [05:34<00:26,  2.91s/it]
Continual predict:  96%|█████████▋| 212/220 [05:37<00:23,  2.90s/it]
Continual predict:  97%|█████████▋| 213/220 [05:40<00:21,  3.09s/it]
Continual predict:  97%|█████████▋| 214/220 [05:43<00:18,  3.16s/it]
Continual predict:  98%|█████████▊| 215/220 [05:44<00:12,  2.42s/it]
Continual predict:  98%|█████████▊| 216/220 [05:47<00:10,  2.55s/it]
Continual predict:  99%|█████████▊| 217/220 [05:49<00:07,  2.35s/it]
Continual predict:  99%|█████████▉| 218/220 [05:52<00:05,  2.74s/it]
Continual predict: 100%|█████████▉| 219/220 [05:55<00:02,  2.80s/it]
Continual predict: 100%|██████████| 220/220 [05:59<00:00,  3.11s/it]
Continual predict: 100%|██████████| 220/220 [05:59<00:00,  1.64s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=20]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (220/220)  label_acc=100.0% (220/220)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                        =       15.0
+  predict_eval_rougeL                                          =    58.5759
+  predict_eval_rougeL_for_CL                                   =    58.5759
+  predict_eval_rougeL_for_task1290_xsum_summarization          =    29.5925
+  predict_eval_rougeL_for_task1572_samsum_summary              =    41.3728
+  predict_eval_rougeL_for_task363_sst2_polarity_classification =       91.0
+  predict_exact_match                                          =    41.3636
+  predict_exact_match_for_CL                                   =    41.3636
+  predict_exact_match_for_task1290_xsum_summarization          =        0.0
+  predict_exact_match_for_task1572_samsum_summary              =        0.0
+  predict_exact_match_for_task363_sst2_polarity_classification =       91.0
+  predict_gen_len                                              =    14.6182
+  predict_global_step                                          =        480
+  predict_rouge1                                               =    62.6489
+  predict_rouge1_for_CL                                        =    62.6489
+  predict_rouge1_for_task1290_xsum_summarization               =      36.75
+  predict_rouge1_for_task1572_samsum_summary                   =    50.3879
+  predict_rouge1_for_task363_sst2_polarity_classification      =       91.0
+  predict_runtime                                              = 0:05:59.74
+  predict_samples                                              =        220
+  predict_samples_per_second                                   =      0.612
+  predict_steps_per_second                                     =      0.612
+[2026-04-29 21:19:14,516] [INFO] [launch.py:367:main] Process 560438 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/all_results.json b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..d3a92abacd32d2359156d235f6da5f05275e9cda
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/all_results.json
@@ -0,0 +1,33 @@
+{
+  "predict_exact_match": 43.346,
+  "predict_rouge1": 62.1102,
+  "predict_eval_rougeL": 58.7031,
+  "predict_gen_len": 12.654,
+  "predict_runtime": 376.28015851508826,
+  "predict_samples": 263,
+  "predict_samples_per_second": 0.699,
+  "predict_steps_per_second": 0.699,
+  "epoch": 15.0,
+  "predict_global_step": 165,
+  "predict_exact_match_for_CL": 43.346,
+  "predict_rouge1_for_CL": 62.1102,
+  "predict_eval_rougeL_for_CL": 58.7031,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "train_runtime": 653.4649,
+  "train_samples_per_second": 7.759,
+  "train_steps_per_second": 0.253,
+  "total_flos": 4.996349011427328e+16,
+  "train_loss": 0.5650005846312551,
+  "train_samples": 338
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f143ed49ad6f05efa270a00ba2841ee3668ad458
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7448fca47c2c3aab002d120ac46e1dbc7b105185ef197db1e8e48aea1d0bc80a
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..13750d9d25781ef726f2a805eba9c819a562a36d
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8fbaf6b26963c6e9bae5f948f9f7a16398b9e0b1263da03cb99fc9e22e434ad
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..e51f5d88802c695f5882673a201cb4f5bcc84a2a
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c297c272524940c7f8abf1bf3abcb8c96e3fc94c0f965ffc29a616ef9e9c56e4
+size 702402799
diff --git a/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/train.log b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..285ff3c9f655b234d58d0451d628b66e4dfd7736
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/train.log
@@ -0,0 +1,297 @@
+[2026-04-29 21:19:20,537] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 21:19:20,537] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29503 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task181_outcome_extraction --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction --cur_task_id 3 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[2026-04-29 21:19:23,701] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 21:19:23,701] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 21:19:23,701] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 21:19:23,701] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 21:19:23,701] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 21:19:23,702] [INFO] [launch.py:272:main] process 739565 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task181_outcome_extraction', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction', '--cur_task_id', '3', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights']
+
+============================================================
+Task 3: task181_outcome_extraction
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 338/43/263
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   3%|▎         | 8/291 [00:00<00:04, 69.71it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 142.85it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 155.26it/s]
Loading weights:  23%|██▎       | 68/291 [00:00<00:01, 162.39it/s]
Loading weights:  30%|██▉       | 86/291 [00:00<00:01, 167.42it/s]
Loading weights:  35%|███▌      | 103/291 [00:00<00:01, 164.79it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:01, 168.82it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 171.72it/s]
Loading weights:  56%|█████▌    | 162/291 [00:00<00:00, 182.64it/s]
Loading weights:  62%|██████▏   | 181/291 [00:01<00:00, 184.30it/s]
Loading weights:  69%|██████▊   | 200/291 [00:01<00:00, 179.19it/s]
Loading weights:  75%|███████▍  | 218/291 [00:01<00:00, 175.80it/s]
Loading weights:  82%|████████▏ | 238/291 [00:01<00:00, 176.76it/s]
Loading weights:  88%|████████▊ | 256/291 [00:01<00:00, 172.05it/s]
Loading weights:  94%|█████████▍| 274/291 [00:01<00:00, 170.10it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 172.83it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=547)
+[FIX] Re-initialized lora_A/B in 256 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(3, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 3 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6747.640M (0.04%)
+[TRAIN] step_per_epoch=11, eval_steps=55
+  [SRT] Loaded 3 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 3 sigs, cur=task181_outcome_extraction, mapping={'task181_outcome_extraction': 0, 'task1572_samsum_summary': 3, 'task363_sst2_polarity_classification': 2, 'task1290_xsum_summarization': 1}
+[SGWI] Task 3: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 21:20:18,802] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/165 [00:00<?, ?it/s]
  1%|          | 1/165 [00:04<12:14,  4.48s/it]
  1%|          | 2/165 [00:08<11:09,  4.11s/it]
  2%|▏         | 3/165 [00:12<10:55,  4.04s/it]
  2%|▏         | 4/165 [00:16<10:52,  4.05s/it]
  3%|▎         | 5/165 [00:20<10:47,  4.04s/it]
  4%|▎         | 6/165 [00:24<10:47,  4.08s/it]
  4%|▍         | 7/165 [00:28<10:36,  4.03s/it]
  5%|▍         | 8/165 [00:32<10:25,  3.99s/it]
  5%|▌         | 9/165 [00:36<10:29,  4.04s/it]
  6%|▌         | 10/165 [00:40<10:20,  4.01s/it]
                                                
{'loss': '2.676', 'grad_norm': '0.5823', 'learning_rate': '5e-05', 'epoch': '0.9467'}
+
  6%|▌         | 10/165 [00:40<10:20,  4.01s/it]
  7%|▋         | 11/165 [00:42<08:45,  3.41s/it]
  7%|▋         | 12/165 [00:46<09:21,  3.67s/it]
  8%|▊         | 13/165 [00:50<09:36,  3.79s/it]
  8%|▊         | 14/165 [00:54<09:49,  3.90s/it]
  9%|▉         | 15/165 [00:59<09:54,  3.96s/it]
 10%|▉         | 16/165 [01:03<09:49,  3.96s/it]
 10%|█         | 17/165 [01:06<09:39,  3.92s/it]
 11%|█         | 18/165 [01:10<09:37,  3.93s/it]
 12%|█▏        | 19/165 [01:14<09:33,  3.93s/it]
 12%|█▏        | 20/165 [01:18<09:31,  3.94s/it]
                                                
{'loss': '1.337', 'grad_norm': '2.767', 'learning_rate': '5e-05', 'epoch': '1.852'}
+
 12%|█▏        | 20/165 [01:18<09:31,  3.94s/it]
 13%|█▎        | 21/165 [01:22<09:20,  3.89s/it]
 13%|█▎        | 22/165 [01:24<08:05,  3.39s/it]
 14%|█▍        | 23/165 [01:28<08:23,  3.55s/it]
 15%|█▍        | 24/165 [01:32<08:33,  3.64s/it]
 15%|█▌        | 25/165 [01:36<08:42,  3.73s/it]
 16%|█▌        | 26/165 [01:40<08:46,  3.79s/it]
 16%|█▋        | 27/165 [01:44<08:43,  3.79s/it]
 17%|█▋        | 28/165 [01:48<08:47,  3.85s/it]
 18%|█▊        | 29/165 [01:52<08:47,  3.88s/it]
 18%|█▊        | 30/165 [01:56<08:47,  3.90s/it]
                                                
{'loss': '0.5873', 'grad_norm': '0.3948', 'learning_rate': '5e-05', 'epoch': '2.757'}
+
 18%|█▊        | 30/165 [01:56<08:47,  3.90s/it]
 19%|█▉        | 31/165 [01:59<08:38,  3.87s/it]
 19%|█▉        | 32/165 [02:03<08:38,  3.90s/it]
 20%|██        | 33/165 [02:06<07:28,  3.40s/it]
 21%|██        | 34/165 [02:10<07:54,  3.62s/it]
 21%|██        | 35/165 [02:14<08:01,  3.71s/it]
 22%|██▏       | 36/165 [02:18<08:13,  3.83s/it]
 22%|██▏       | 37/165 [02:22<08:27,  3.96s/it]
 23%|██▎       | 38/165 [02:26<08:23,  3.97s/it]
 24%|██▎       | 39/165 [02:30<08:21,  3.98s/it]
 24%|██▍       | 40/165 [02:34<08:19,  3.99s/it]
                                                
{'loss': '0.4541', 'grad_norm': '1.047', 'learning_rate': '5e-05', 'epoch': '3.663'}
+
 24%|██▍       | 40/165 [02:34<08:19,  3.99s/it]
 25%|██▍       | 41/165 [02:38<08:09,  3.95s/it]
 25%|██▌       | 42/165 [02:42<08:03,  3.93s/it]
 26%|██▌       | 43/165 [02:46<07:54,  3.89s/it]
 27%|██▋       | 44/165 [02:48<06:47,  3.37s/it]
 27%|██▋       | 45/165 [02:52<07:10,  3.58s/it]
 28%|██▊       | 46/165 [02:56<07:20,  3.70s/it]
 28%|██▊       | 47/165 [02:59<07:18,  3.72s/it]
 29%|██▉       | 48/165 [03:04<07:27,  3.82s/it]
 30%|██▉       | 49/165 [03:07<07:22,  3.81s/it]
 30%|███       | 50/165 [03:11<07:19,  3.82s/it]
                                                
{'loss': '0.4299', 'grad_norm': '0.8056', 'learning_rate': '5e-05', 'epoch': '4.568'}
+
 30%|███       | 50/165 [03:11<07:19,  3.82s/it]
 31%|███       | 51/165 [03:15<07:20,  3.86s/it]
 32%|███▏      | 52/165 [03:19<07:16,  3.86s/it]
 32%|███▏      | 53/165 [03:23<07:28,  4.00s/it]
 33%|███▎      | 54/165 [03:27<07:20,  3.96s/it]
 33%|███▎      | 55/165 [03:29<06:19,  3.45s/it]
+
Eval @ step 55:   0%|          | 0/43 [00:00<?, ?it/s]
+
Eval @ step 55:   2%|▏         | 1/43 [00:00<00:05,  8.14it/s]
+
Eval @ step 55:   5%|▍         | 2/43 [00:00<00:05,  7.33it/s]
+
Eval @ step 55:   7%|▋         | 3/43 [00:00<00:04,  8.03it/s]
+
Eval @ step 55:   9%|▉         | 4/43 [00:00<00:05,  7.48it/s]
+
Eval @ step 55:  12%|█▏        | 5/43 [00:00<00:04,  7.96it/s]
+
Eval @ step 55:  14%|█▍        | 6/43 [00:00<00:06,  5.48it/s]
+
Eval @ step 55:  16%|█▋        | 7/43 [00:01<00:06,  5.91it/s]
+
Eval @ step 55:  19%|█▊        | 8/43 [00:01<00:05,  6.29it/s]
+
Eval @ step 55:  21%|██        | 9/43 [00:01<00:05,  6.69it/s]
+
Eval @ step 55:  23%|██▎       | 10/43 [00:01<00:04,  6.70it/s]
+
Eval @ step 55:  26%|██▌       | 11/43 [00:01<00:04,  7.27it/s]
+
Eval @ step 55:  28%|██▊       | 12/43 [00:01<00:04,  7.72it/s]
+
Eval @ step 55:  30%|███       | 13/43 [00:01<00:03,  8.08it/s]
+
Eval @ step 55:  33%|███▎      | 14/43 [00:01<00:03,  8.34it/s]
+
Eval @ step 55:  35%|███▍      | 15/43 [00:02<00:03,  8.54it/s]
+
Eval @ step 55:  37%|███▋      | 16/43 [00:02<00:03,  7.93it/s]
+
Eval @ step 55:  40%|███▉      | 17/43 [00:02<00:05,  4.59it/s]
+
Eval @ step 55:  42%|████▏     | 18/43 [00:03<00:09,  2.64it/s]
+
Eval @ step 55:  44%|████▍     | 19/43 [00:03<00:07,  3.36it/s]
+
Eval @ step 55:  47%|████▋     | 20/43 [00:03<00:06,  3.80it/s]
+
Eval @ step 55:  49%|████▉     | 21/43 [00:03<00:04,  4.61it/s]
+
Eval @ step 55:  51%|█████     | 22/43 [00:03<00:03,  5.41it/s]
+
Eval @ step 55:  53%|█████▎    | 23/43 [00:03<00:03,  6.18it/s]
+
Eval @ step 55:  56%|█████▌    | 24/43 [00:04<00:03,  5.63it/s]
+
Eval @ step 55:  58%|█████▊    | 25/43 [00:04<00:03,  4.51it/s]
+
Eval @ step 55:  60%|██████    | 26/43 [00:04<00:03,  5.33it/s]
+
Eval @ step 55:  63%|██████▎   | 27/43 [00:04<00:02,  5.81it/s]
+
Eval @ step 55:  65%|██████▌   | 28/43 [00:04<00:02,  6.54it/s]
+
Eval @ step 55:  67%|██████▋   | 29/43 [00:05<00:03,  3.54it/s]
+
Eval @ step 55:  72%|███████▏  | 31/43 [00:06<00:03,  3.30it/s]
+
Eval @ step 55:  74%|███████▍  | 32/43 [00:06<00:03,  3.18it/s]
+
Eval @ step 55:  77%|███████▋  | 33/43 [00:06<00:02,  3.73it/s]
+
Eval @ step 55:  79%|███████▉  | 34/43 [00:06<00:02,  4.45it/s]
+
Eval @ step 55:  81%|████████▏ | 35/43 [00:06<00:01,  5.20it/s]
+
Eval @ step 55:  84%|████████▎ | 36/43 [00:06<00:01,  5.93it/s]
+
Eval @ step 55:  86%|████████▌ | 37/43 [00:07<00:00,  6.61it/s]
+
Eval @ step 55:  88%|████████▊ | 38/43 [00:07<00:00,  6.27it/s]
+
Eval @ step 55:  91%|█████████ | 39/43 [00:07<00:00,  6.92it/s]
+
Eval @ step 55:  93%|█████████▎| 40/43 [00:07<00:00,  6.91it/s]
+
Eval @ step 55:  95%|█████████▌| 41/43 [00:07<00:00,  5.33it/s]
+
Eval @ step 55:  98%|█████████▊| 42/43 [00:08<00:00,  4.77it/s]
+
Eval @ step 55: 100%|██████████| 43/43 [00:08<00:00,  4.98it/s]
+
                                                                 [EVAL] ★ New best eval_rougeL=50.8114 (step=55) — snapshot saved in memory
+
                                                
{'eval_rougeL': '50.81', 'epoch': '5'}
+
 33%|███▎      | 55/165 [03:38<06:19,  3.45s/it]
 34%|███▍      | 56/165 [03:42<11:09,  6.15s/it]
 35%|███▍      | 57/165 [03:46<09:52,  5.49s/it]
 35%|███▌      | 58/165 [03:50<08:55,  5.01s/it]
 36%|███▌      | 59/165 [03:53<08:10,  4.63s/it]
 36%|███▋      | 60/165 [03:57<07:38,  4.37s/it]
                                                
{'loss': '0.5207', 'grad_norm': '0.5056', 'learning_rate': '5e-05', 'epoch': '5.473'}
+
 36%|███▋      | 60/165 [03:57<07:38,  4.37s/it]
 37%|███▋      | 61/165 [04:01<07:18,  4.22s/it]
 38%|███▊      | 62/165 [04:05<07:01,  4.09s/it]
 38%|███▊      | 63/165 [04:09<07:03,  4.15s/it]
 39%|███▉      | 64/165 [04:13<07:02,  4.18s/it]
 39%|███▉      | 65/165 [04:17<06:51,  4.11s/it]
 40%|████      | 66/165 [04:20<05:48,  3.52s/it]
 41%|████      | 67/165 [04:23<05:54,  3.62s/it]
 41%|████      | 68/165 [04:27<05:57,  3.69s/it]
 42%|████▏     | 69/165 [04:31<06:07,  3.82s/it]
 42%|████▏     | 70/165 [04:35<06:02,  3.81s/it]
                                                
{'loss': '0.3659', 'grad_norm': '0.08554', 'learning_rate': '5e-05', 'epoch': '6.379'}
+
 42%|████▏     | 70/165 [04:35<06:02,  3.81s/it]
 43%|████▎     | 71/165 [04:39<06:00,  3.83s/it]
 44%|████▎     | 72/165 [04:43<06:00,  3.88s/it]
 44%|████▍     | 73/165 [04:47<05:57,  3.88s/it]
 45%|████▍     | 74/165 [04:51<05:59,  3.95s/it]
 45%|████▌     | 75/165 [04:55<05:53,  3.93s/it]
 46%|████▌     | 76/165 [04:59<05:53,  3.97s/it]
 47%|████▋     | 77/165 [05:01<05:02,  3.43s/it]
 47%|████▋     | 78/165 [05:05<05:10,  3.57s/it]
 48%|████▊     | 79/165 [05:09<05:17,  3.69s/it]
 48%|████▊     | 80/165 [05:13<05:23,  3.80s/it]
                                                
{'loss': '0.4115', 'grad_norm': '0.1508', 'learning_rate': '5e-05', 'epoch': '7.284'}
+
 48%|████▊     | 80/165 [05:13<05:23,  3.80s/it]
 49%|████▉     | 81/165 [05:17<05:24,  3.86s/it]
 50%|████▉     | 82/165 [05:21<05:24,  3.92s/it]
 50%|█████     | 83/165 [05:25<05:24,  3.96s/it]
 51%|█████     | 84/165 [05:29<05:25,  4.01s/it]
 52%|█████▏    | 85/165 [05:33<05:20,  4.01s/it]
 52%|█████▏    | 86/165 [05:37<05:11,  3.95s/it]
 53%|█████▎    | 87/165 [05:41<05:10,  3.98s/it]
 53%|█████▎    | 88/165 [05:43<04:25,  3.45s/it]
 54%|█████▍    | 89/165 [05:48<04:37,  3.66s/it]
 55%|█████▍    | 90/165 [05:52<04:46,  3.81s/it]
                                                
{'loss': '0.357', 'grad_norm': '0.2645', 'learning_rate': '5e-05', 'epoch': '8.189'}
+
 55%|█████▍    | 90/165 [05:52<04:46,  3.81s/it]
 55%|█████▌    | 91/165 [05:56<04:42,  3.82s/it]
 56%|█████▌    | 92/165 [06:00<04:43,  3.89s/it]
 56%|█████▋    | 93/165 [06:03<04:38,  3.87s/it]
 57%|█████▋    | 94/165 [06:07<04:32,  3.84s/it]
 58%|█████▊    | 95/165 [06:11<04:27,  3.82s/it]
 58%|█████▊    | 96/165 [06:15<04:31,  3.93s/it]
 59%|█████▉    | 97/165 [06:19<04:26,  3.92s/it]
 59%|█████▉    | 98/165 [06:23<04:19,  3.88s/it]
 60%|██████    | 99/165 [06:25<03:41,  3.36s/it]
 61%|██████    | 100/165 [06:29<03:53,  3.59s/it]
                                                 
{'loss': '0.3302', 'grad_norm': '0.3256', 'learning_rate': '5e-05', 'epoch': '9.095'}
+
 61%|██████    | 100/165 [06:29<03:53,  3.59s/it]
 61%|██████    | 101/165 [06:33<03:56,  3.69s/it]
 62%|██████▏   | 102/165 [06:37<03:57,  3.77s/it]
 62%|██████▏   | 103/165 [06:41<03:54,  3.79s/it]
 63%|██████▎   | 104/165 [06:45<03:57,  3.89s/it]
 64%|██████▎   | 105/165 [06:49<03:54,  3.91s/it]
 64%|██████▍   | 106/165 [06:53<03:49,  3.89s/it]
 65%|██████▍   | 107/165 [06:57<03:51,  3.99s/it]
 65%|██████▌   | 108/165 [07:02<04:01,  4.24s/it]
 66%|██████▌   | 109/165 [07:06<03:59,  4.28s/it]
 67%|██████▋   | 110/165 [07:09<03:24,  3.71s/it]
                                                 
{'loss': '0.3149', 'grad_norm': '0.6276', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 110/165 [07:09<03:24,  3.71s/it]
+
Eval @ step 110:   0%|          | 0/43 [00:00<?, ?it/s]
+
Eval @ step 110:   2%|▏         | 1/43 [00:00<00:04,  8.96it/s]
+
Eval @ step 110:   5%|▍         | 2/43 [00:00<00:04,  9.04it/s]
+
Eval @ step 110:   7%|▋         | 3/43 [00:00<00:04,  9.04it/s]
+
Eval @ step 110:   9%|▉         | 4/43 [00:00<00:04,  9.02it/s]
+
Eval @ step 110:  12%|█▏        | 5/43 [00:00<00:04,  8.99it/s]
+
Eval @ step 110:  14%|█▍        | 6/43 [00:00<00:04,  8.94it/s]
+
Eval @ step 110:  16%|█▋        | 7/43 [00:00<00:04,  8.20it/s]
+
Eval @ step 110:  19%|█▊        | 8/43 [00:00<00:04,  7.83it/s]
+
Eval @ step 110:  21%|██        | 9/43 [00:01<00:04,  7.77it/s]
+
Eval @ step 110:  23%|██▎       | 10/43 [00:01<00:05,  5.95it/s]
+
Eval @ step 110:  26%|██▌       | 11/43 [00:01<00:04,  6.64it/s]
+
Eval @ step 110:  28%|██▊       | 12/43 [00:01<00:04,  7.21it/s]
+
Eval @ step 110:  30%|███       | 13/43 [00:01<00:03,  7.69it/s]
+
Eval @ step 110:  33%|███▎      | 14/43 [00:01<00:03,  8.03it/s]
+
Eval @ step 110:  35%|███▍      | 15/43 [00:01<00:03,  8.31it/s]
+
Eval @ step 110:  37%|███▋      | 16/43 [00:02<00:03,  8.52it/s]
+
Eval @ step 110:  40%|███▉      | 17/43 [00:02<00:02,  8.67it/s]
+
Eval @ step 110:  42%|████▏     | 18/43 [00:02<00:02,  8.78it/s]
+
Eval @ step 110:  44%|████▍     | 19/43 [00:02<00:02,  8.87it/s]
+
Eval @ step 110:  47%|████▋     | 20/43 [00:02<00:03,  7.44it/s]
+
Eval @ step 110:  49%|████▉     | 21/43 [00:02<00:02,  7.87it/s]
+
Eval @ step 110:  51%|█████     | 22/43 [00:02<00:02,  8.19it/s]
+
Eval @ step 110:  53%|█████▎    | 23/43 [00:02<00:02,  8.43it/s]
+
Eval @ step 110:  56%|█████▌    | 24/43 [00:02<00:02,  8.61it/s]
+
Eval @ step 110:  58%|█████▊    | 25/43 [00:03<00:02,  8.72it/s]
+
Eval @ step 110:  60%|██████    | 26/43 [00:03<00:01,  8.83it/s]
+
Eval @ step 110:  63%|██████▎   | 27/43 [00:03<00:01,  8.27it/s]
+
Eval @ step 110:  65%|██████▌   | 28/43 [00:03<00:01,  8.49it/s]
+
Eval @ step 110:  67%|██████▋   | 29/43 [00:03<00:01,  8.07it/s]
+
Eval @ step 110:  72%|███████▏  | 31/43 [00:03<00:01,  8.65it/s]
+
Eval @ step 110:  74%|███████▍  | 32/43 [00:03<00:01,  7.65it/s]
+
Eval @ step 110:  77%|███████▋  | 33/43 [00:04<00:01,  7.54it/s]
+
Eval @ step 110:  79%|███████▉  | 34/43 [00:04<00:01,  7.90it/s]
+
Eval @ step 110:  81%|████████▏ | 35/43 [00:04<00:00,  8.17it/s]
+
Eval @ step 110:  84%|████████▎ | 36/43 [00:04<00:00,  8.38it/s]
+
Eval @ step 110:  86%|████████▌ | 37/43 [00:04<00:01,  5.03it/s]
+
Eval @ step 110:  88%|████████▊ | 38/43 [00:04<00:00,  5.77it/s]
+
Eval @ step 110:  91%|█████████ | 39/43 [00:05<00:00,  6.46it/s]
+
Eval @ step 110:  93%|█████████▎| 40/43 [00:05<00:00,  7.05it/s]
+
Eval @ step 110:  95%|█████████▌| 41/43 [00:05<00:00,  7.53it/s]
+
Eval @ step 110:  98%|█████████▊| 42/43 [00:05<00:00,  7.64it/s]
+
Eval @ step 110: 100%|██████████| 43/43 [00:05<00:00,  7.55it/s]
+
                                                                  [EVAL] ★ New best eval_rougeL=56.3307 (step=110) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '56.33', 'epoch': '10'}
+
 67%|██████▋   | 110/165 [07:14<03:24,  3.71s/it]
 67%|██████▋   | 111/165 [07:19<05:01,  5.59s/it]
 68%|██████▊   | 112/165 [07:23<04:34,  5.18s/it]
 68%|██████▊   | 113/165 [07:27<04:10,  4.82s/it]
 69%|██████▉   | 114/165 [07:31<03:53,  4.57s/it]
 70%|██████▉   | 115/165 [07:35<03:40,  4.41s/it]
 70%|███████   | 116/165 [07:39<03:29,  4.28s/it]
 71%|███████   | 117/165 [07:43<03:20,  4.18s/it]
 72%|███████▏  | 118/165 [07:47<03:13,  4.11s/it]
 72%|███████▏  | 119/165 [07:50<03:04,  4.01s/it]
 73%|███████▎  | 120/165 [07:54<02:57,  3.94s/it]
                                                 
{'loss': '0.2968', 'grad_norm': '0.5448', 'learning_rate': '5e-05', 'epoch': '10.95'}
+
 73%|███████▎  | 120/165 [07:54<02:57,  3.94s/it]
 73%|███████▎  | 121/165 [07:56<02:29,  3.41s/it]
 74%|███████▍  | 122/165 [08:00<02:32,  3.56s/it]
 75%|███████▍  | 123/165 [08:04<02:34,  3.68s/it]
 75%|███████▌  | 124/165 [08:08<02:33,  3.75s/it]
 76%|███████▌  | 125/165 [08:12<02:34,  3.86s/it]
 76%|███████▋  | 126/165 [08:16<02:33,  3.93s/it]
 77%|███████▋  | 127/165 [08:21<02:33,  4.05s/it]
 78%|███████▊  | 128/165 [08:25<02:28,  4.01s/it]
 78%|███████▊  | 129/165 [08:29<02:28,  4.13s/it]
 79%|███████▉  | 130/165 [08:33<02:24,  4.13s/it]
                                                 
{'loss': '0.269', 'grad_norm': '0.3309', 'learning_rate': '5e-05', 'epoch': '11.85'}
+
 79%|███████▉  | 130/165 [08:33<02:24,  4.13s/it]
 79%|███████▉  | 131/165 [08:37<02:17,  4.03s/it]
 80%|████████  | 132/165 [08:39<01:56,  3.52s/it]
 81%|████████  | 133/165 [08:44<01:59,  3.74s/it]
 81%|████████  | 134/165 [08:48<01:59,  3.86s/it]
 82%|████████▏ | 135/165 [08:52<02:00,  4.00s/it]
 82%|████████▏ | 136/165 [08:56<01:55,  3.99s/it]
 83%|████████▎ | 137/165 [09:00<01:50,  3.96s/it]
 84%|████████▎ | 138/165 [09:04<01:49,  4.04s/it]
 84%|████████▍ | 139/165 [09:08<01:46,  4.10s/it]
 85%|████████▍ | 140/165 [09:12<01:41,  4.06s/it]
                                                 
{'loss': '0.3046', 'grad_norm': '0.7305', 'learning_rate': '5e-05', 'epoch': '12.76'}
+
 85%|████████▍ | 140/165 [09:12<01:41,  4.06s/it]
 85%|████████▌ | 141/165 [09:16<01:37,  4.04s/it]
 86%|████████▌ | 142/165 [09:20<01:31,  3.99s/it]
 87%|████████▋ | 143/165 [09:23<01:18,  3.55s/it]
 87%|████████▋ | 144/165 [09:27<01:16,  3.64s/it]
 88%|████████▊ | 145/165 [09:31<01:15,  3.77s/it]
 88%|████████▊ | 146/165 [09:35<01:13,  3.86s/it]
 89%|████████▉ | 147/165 [09:39<01:09,  3.86s/it]
 90%|████████▉ | 148/165 [09:42<01:05,  3.84s/it]
 90%|█████████ | 149/165 [09:47<01:04,  4.01s/it]
 91%|█████████ | 150/165 [09:51<01:00,  4.02s/it]
                                                 
{'loss': '0.2557', 'grad_norm': '0.3667', 'learning_rate': '5e-05', 'epoch': '13.66'}
+
 91%|█████████ | 150/165 [09:51<01:00,  4.02s/it]
 92%|█████████▏| 151/165 [09:55<00:57,  4.10s/it]
 92%|█████████▏| 152/165 [09:59<00:52,  4.05s/it]
 93%|█████████▎| 153/165 [10:03<00:47,  3.98s/it]
 93%|█████████▎| 154/165 [10:05<00:37,  3.45s/it]
 94%|█████████▍| 155/165 [10:09<00:36,  3.61s/it]
 95%|█████████▍| 156/165 [10:13<00:33,  3.67s/it]
 95%|█████████▌| 157/165 [10:17<00:30,  3.76s/it]
 96%|█████████▌| 158/165 [10:21<00:27,  3.88s/it]
 96%|█████████▋| 159/165 [10:25<00:23,  3.86s/it]
 97%|█████████▋| 160/165 [10:29<00:19,  3.85s/it]
                                                 
{'loss': '0.2622', 'grad_norm': '0.3567', 'learning_rate': '5e-05', 'epoch': '14.57'}
+
 97%|█████████▋| 160/165 [10:29<00:19,  3.85s/it]
 98%|█████████▊| 161/165 [10:33<00:15,  3.88s/it]
 98%|█████████▊| 162/165 [10:37<00:11,  3.95s/it]
 99%|█████████▉| 163/165 [10:40<00:07,  3.90s/it]
 99%|█████████▉| 164/165 [10:45<00:03,  3.98s/it]
100%|██████████| 165/165 [10:47<00:00,  3.45s/it]
+
Eval @ step 165:   0%|          | 0/43 [00:00<?, ?it/s]
+
Eval @ step 165:   2%|▏         | 1/43 [00:00<00:04,  8.93it/s]
+
Eval @ step 165:   5%|▍         | 2/43 [00:00<00:05,  7.54it/s]
+
Eval @ step 165:   7%|▋         | 3/43 [00:00<00:04,  8.15it/s]
+
Eval @ step 165:   9%|▉         | 4/43 [00:00<00:04,  8.46it/s]
+
Eval @ step 165:  12%|█▏        | 5/43 [00:00<00:04,  8.37it/s]
+
Eval @ step 165:  14%|█▍        | 6/43 [00:00<00:04,  8.58it/s]
+
Eval @ step 165:  16%|█▋        | 7/43 [00:00<00:04,  8.04it/s]
+
Eval @ step 165:  19%|█▊        | 8/43 [00:00<00:04,  7.78it/s]
+
Eval @ step 165:  21%|██        | 9/43 [00:01<00:04,  7.77it/s]
+
Eval @ step 165:  23%|██▎       | 10/43 [00:01<00:05,  6.35it/s]
+
Eval @ step 165:  26%|██▌       | 11/43 [00:01<00:04,  6.97it/s]
+
Eval @ step 165:  28%|██▊       | 12/43 [00:01<00:04,  7.49it/s]
+
Eval @ step 165:  30%|███       | 13/43 [00:01<00:03,  7.90it/s]
+
Eval @ step 165:  33%|███▎      | 14/43 [00:01<00:03,  8.20it/s]
+
Eval @ step 165:  35%|███▍      | 15/43 [00:01<00:03,  8.44it/s]
+
Eval @ step 165:  37%|███▋      | 16/43 [00:02<00:03,  8.62it/s]
+
Eval @ step 165:  40%|███▉      | 17/43 [00:02<00:02,  8.76it/s]
+
Eval @ step 165:  42%|████▏     | 18/43 [00:02<00:02,  8.90it/s]
+
Eval @ step 165:  44%|████▍     | 19/43 [00:02<00:02,  9.01it/s]
+
Eval @ step 165:  47%|████▋     | 20/43 [00:02<00:03,  7.57it/s]
+
Eval @ step 165:  49%|████▉     | 21/43 [00:02<00:02,  8.01it/s]
+
Eval @ step 165:  51%|█████     | 22/43 [00:02<00:02,  8.35it/s]
+
Eval @ step 165:  53%|█████▎    | 23/43 [00:02<00:02,  8.60it/s]
+
Eval @ step 165:  56%|█████▌    | 24/43 [00:02<00:02,  8.79it/s]
+
Eval @ step 165:  58%|█████▊    | 25/43 [00:03<00:02,  8.91it/s]
+
Eval @ step 165:  60%|██████    | 26/43 [00:03<00:01,  9.01it/s]
+
Eval @ step 165:  63%|██████▎   | 27/43 [00:03<00:01,  8.44it/s]
+
Eval @ step 165:  65%|██████▌   | 28/43 [00:03<00:01,  8.66it/s]
+
Eval @ step 165:  67%|██████▋   | 29/43 [00:03<00:01,  8.23it/s]
+
Eval @ step 165:  70%|██████▉   | 30/43 [00:03<00:01,  8.51it/s]
+
Eval @ step 165:  72%|███████▏  | 31/43 [00:03<00:01,  8.14it/s]
+
Eval @ step 165:  74%|███████▍  | 32/43 [00:04<00:03,  3.60it/s]
+
Eval @ step 165:  77%|███████▋  | 33/43 [00:04<00:02,  4.25it/s]
+
Eval @ step 165:  79%|███████▉  | 34/43 [00:04<00:01,  5.08it/s]
+
Eval @ step 165:  81%|████████▏ | 35/43 [00:04<00:01,  5.86it/s]
+
Eval @ step 165:  84%|████████▎ | 36/43 [00:04<00:01,  6.57it/s]
+
Eval @ step 165:  86%|████████▌ | 37/43 [00:05<00:01,  4.49it/s]
+
Eval @ step 165:  88%|████████▊ | 38/43 [00:05<00:00,  5.31it/s]
+
Eval @ step 165:  91%|█████████ | 39/43 [00:05<00:00,  6.08it/s]
+
Eval @ step 165:  93%|█████████▎| 40/43 [00:05<00:00,  6.76it/s]
+
Eval @ step 165:  95%|█████████▌| 41/43 [00:05<00:00,  5.26it/s]
+
Eval @ step 165:  98%|█████████▊| 42/43 [00:06<00:00,  5.84it/s]
+
Eval @ step 165: 100%|██████████| 43/43 [00:06<00:00,  6.23it/s]
+
                                                                  [EVAL] ★ New best eval_rougeL=59.3540 (step=165) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '59.35', 'epoch': '15'}
+
100%|██████████| 165/165 [10:53<00:00,  3.45s/it]
                                                 
{'train_runtime': '653.5', 'train_samples_per_second': '7.759', 'train_steps_per_second': '0.253', 'train_loss': '0.565', 'epoch': '15'}
+
100%|██████████| 165/165 [10:53<00:00,  3.45s/it]
100%|██████████| 165/165 [10:53<00:00,  3.96s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 46532126GF
+  train_loss               =      0.565
+  train_runtime            = 0:10:53.46
+  train_samples            =        338
+  train_samples_per_second =      7.759
+  train_steps_per_second   =      0.253
+  [EVAL] Restored best model (eval_rougeL=59.3540, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 120.39it/s]
Loading weights:  11%|█         | 32/291 [00:00<00:01, 156.32it/s]
Loading weights:  18%|█▊        | 51/291 [00:00<00:01, 168.96it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 167.88it/s]
Loading weights:  31%|███       | 89/291 [00:00<00:01, 179.01it/s]
Loading weights:  37%|███▋      | 109/291 [00:00<00:00, 182.12it/s]
Loading weights:  44%|████▍     | 128/291 [00:00<00:00, 184.34it/s]
Loading weights:  51%|█████     | 148/291 [00:00<00:00, 174.48it/s]
Loading weights:  57%|█████▋    | 167/291 [00:00<00:00, 177.15it/s]
Loading weights:  64%|██████▍   | 186/291 [00:01<00:00, 179.68it/s]
Loading weights:  71%|███████▏  | 208/291 [00:01<00:00, 185.81it/s]
Loading weights:  78%|███████▊  | 227/291 [00:01<00:00, 182.09it/s]
Loading weights:  85%|████████▍ | 246/291 [00:01<00:00, 182.78it/s]
Loading weights:  91%|█████████ | 265/291 [00:01<00:00, 173.65it/s]
Loading weights:  98%|█████████▊| 284/291 [00:01<00:00, 177.68it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 179.73it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (338 samples, batch_size=1)
+  [SRT] Task 'task181_outcome_extraction': PaR=100.0, shrinkage=ridge, n=338
+  [SRT] Wired router: 4 sigs, cur=task181_outcome_extraction, mapping={'task181_outcome_extraction': 0, 'task1572_samsum_summary': 3, 'task363_sst2_polarity_classification': 2, 'task1290_xsum_summarization': 1}
+[SRT] Router: 4 tasks, avg_PaR=100.0, metrics=['ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction']
+  [SRT-INIT] task_id_to_idx={'task181_outcome_extraction': 0, 'task1572_samsum_summary': 3, 'task363_sst2_polarity_classification': 2, 'task1290_xsum_summarization': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/263 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=4
+
Continual predict:   0%|          | 1/263 [00:01<07:11,  1.65s/it]
Continual predict:   1%|          | 2/263 [00:03<08:51,  2.04s/it]
Continual predict:   1%|          | 3/263 [00:04<06:12,  1.43s/it]
Continual predict:   2%|▏         | 4/263 [00:05<04:33,  1.06s/it]
Continual predict:   2%|▏         | 5/263 [00:06<05:32,  1.29s/it]
Continual predict:   2%|▏         | 6/263 [00:11<10:04,  2.35s/it]
Continual predict:   3%|▎         | 7/263 [00:14<11:05,  2.60s/it]
Continual predict:   3%|▎         | 8/263 [00:16<10:06,  2.38s/it]
Continual predict:   3%|▎         | 9/263 [00:17<09:00,  2.13s/it]
Continual predict:   4%|▍         | 10/263 [00:18<07:09,  1.70s/it]
Continual predict:   4%|▍         | 11/263 [00:21<08:09,  1.94s/it]
Continual predict:   5%|▍         | 12/263 [00:24<09:21,  2.24s/it]
Continual predict:   5%|▍         | 13/263 [00:25<07:56,  1.91s/it]
Continual predict:   5%|▌         | 14/263 [00:27<08:19,  2.01s/it]
Continual predict:   6%|▌         | 15/263 [00:29<08:24,  2.03s/it]
Continual predict:   6%|▌         | 16/263 [00:31<08:50,  2.15s/it]
Continual predict:   6%|▋         | 17/263 [00:39<15:11,  3.71s/it]
Continual predict:   7%|▋         | 18/263 [00:41<12:59,  3.18s/it]
Continual predict:   7%|▋         | 19/263 [00:43<11:57,  2.94s/it]
Continual predict:   8%|▊         | 20/263 [00:44<09:52,  2.44s/it]
Continual predict:   8%|▊         | 21/263 [00:44<07:03,  1.75s/it]
Continual predict:   8%|▊         | 22/263 [00:45<05:04,  1.26s/it]
Continual predict:   9%|▊         | 23/263 [00:45<03:42,  1.08it/s]
Continual predict:   9%|▉         | 24/263 [00:45<02:44,  1.45it/s]
Continual predict:  10%|▉         | 25/263 [00:45<02:03,  1.92it/s]
Continual predict:  10%|▉         | 26/263 [00:45<01:35,  2.48it/s]
Continual predict:  10%|█         | 27/263 [00:45<01:15,  3.11it/s]
Continual predict:  11%|█         | 28/263 [00:45<01:01,  3.80it/s]
Continual predict:  11%|█         | 29/263 [00:46<00:52,  4.49it/s]
Continual predict:  11%|█▏        | 30/263 [00:46<00:46,  5.03it/s]
Continual predict:  12%|█▏        | 31/263 [00:46<00:41,  5.61it/s]
Continual predict:  12%|█▏        | 32/263 [00:46<00:37,  6.12it/s]
Continual predict:  13%|█▎        | 33/263 [00:46<00:36,  6.34it/s]
Continual predict:  13%|█▎        | 34/263 [00:46<00:35,  6.52it/s]
Continual predict:  13%|█▎        | 35/263 [00:46<00:34,  6.66it/s]
Continual predict:  14%|█▎        | 36/263 [00:46<00:32,  6.97it/s]
Continual predict:  14%|█▍        | 37/263 [00:47<00:32,  6.98it/s]
Continual predict:  14%|█▍        | 38/263 [00:47<00:31,  7.17it/s]
Continual predict:  15%|█▍        | 39/263 [00:47<00:30,  7.35it/s]
Continual predict:  15%|█▌        | 40/263 [00:47<00:29,  7.49it/s]
Continual predict:  16%|█▌        | 41/263 [00:47<00:30,  7.34it/s]
Continual predict:  16%|█▌        | 42/263 [00:47<00:29,  7.45it/s]
Continual predict:  16%|█▋        | 43/263 [00:47<00:30,  7.31it/s]
Continual predict:  17%|█▋        | 44/263 [00:48<00:30,  7.20it/s]
Continual predict:  17%|█▋        | 45/263 [00:48<00:30,  7.25it/s]
Continual predict:  17%|█▋        | 46/263 [00:48<00:30,  7.07it/s]
Continual predict:  18%|█▊        | 47/263 [00:48<00:30,  7.18it/s]
Continual predict:  18%|█▊        | 48/263 [00:48<00:30,  7.10it/s]
Continual predict:  19%|█▊        | 49/263 [00:48<00:30,  7.09it/s]
Continual predict:  19%|█▉        | 50/263 [00:48<00:29,  7.26it/s]
Continual predict:  19%|█▉        | 51/263 [00:49<00:28,  7.32it/s]
Continual predict:  20%|█▉        | 52/263 [00:49<00:28,  7.43it/s]
Continual predict:  20%|██        | 53/263 [00:49<00:27,  7.51it/s]
Continual predict:  21%|██        | 54/263 [00:49<00:28,  7.35it/s]
Continual predict:  21%|██        | 55/263 [00:49<00:27,  7.49it/s]
Continual predict:  21%|██▏       | 56/263 [00:49<00:28,  7.33it/s]
Continual predict:  22%|██▏       | 57/263 [00:49<00:27,  7.44it/s]
Continual predict:  22%|██▏       | 58/263 [00:49<00:27,  7.52it/s]
Continual predict:  22%|██▏       | 59/263 [00:50<00:26,  7.57it/s]
Continual predict:  23%|██▎       | 60/263 [00:50<00:27,  7.37it/s]
Continual predict:  23%|██▎       | 61/263 [00:50<00:27,  7.47it/s]
Continual predict:  24%|██▎       | 62/263 [00:50<00:26,  7.54it/s]
Continual predict:  24%|██▍       | 63/263 [00:50<00:27,  7.39it/s]
Continual predict:  24%|██▍       | 64/263 [00:50<00:26,  7.48it/s]
Continual predict:  25%|██▍       | 65/263 [00:50<00:26,  7.55it/s]
Continual predict:  25%|██▌       | 66/263 [00:51<00:25,  7.59it/s]
Continual predict:  25%|██▌       | 67/263 [00:51<00:25,  7.66it/s]
Continual predict:  26%|██▌       | 68/263 [00:51<00:25,  7.68it/s]
Continual predict:  26%|██▌       | 69/263 [00:51<00:25,  7.70it/s]
Continual predict:  27%|██▋       | 70/263 [00:51<00:24,  7.73it/s]
Continual predict:  27%|██▋       | 71/263 [00:51<00:24,  7.73it/s]
Continual predict:  27%|██▋       | 72/263 [00:51<00:25,  7.50it/s]
Continual predict:  28%|██▊       | 73/263 [00:51<00:25,  7.59it/s]
Continual predict:  28%|██▊       | 74/263 [00:52<00:25,  7.40it/s]
Continual predict:  29%|██▊       | 75/263 [00:52<00:25,  7.41it/s]
Continual predict:  29%|██▉       | 76/263 [00:52<00:24,  7.53it/s]
Continual predict:  29%|██▉       | 77/263 [00:52<00:24,  7.57it/s]
Continual predict:  30%|██▉       | 78/263 [00:52<00:24,  7.64it/s]
Continual predict:  30%|███       | 79/263 [00:52<00:24,  7.65it/s]
Continual predict:  30%|███       | 80/263 [00:52<00:23,  7.70it/s]
Continual predict:  31%|███       | 81/263 [00:53<00:23,  7.69it/s]
Continual predict:  31%|███       | 82/263 [00:53<00:23,  7.71it/s]
Continual predict:  32%|███▏      | 83/263 [00:53<00:23,  7.71it/s]
Continual predict:  32%|███▏      | 84/263 [00:53<00:23,  7.48it/s]
Continual predict:  32%|███▏      | 85/263 [00:53<00:23,  7.53it/s]
Continual predict:  33%|███▎      | 86/263 [00:53<00:23,  7.59it/s]
Continual predict:  33%|███▎      | 87/263 [00:53<00:23,  7.58it/s]
Continual predict:  33%|███▎      | 88/263 [00:53<00:22,  7.65it/s]
Continual predict:  34%|███▍      | 89/263 [00:54<00:22,  7.68it/s]
Continual predict:  34%|███▍      | 90/263 [00:54<00:22,  7.72it/s]
Continual predict:  35%|███▍      | 91/263 [00:54<00:22,  7.71it/s]
Continual predict:  35%|███▍      | 92/263 [00:54<00:22,  7.67it/s]
Continual predict:  35%|███▌      | 93/263 [00:54<00:22,  7.46it/s]
Continual predict:  36%|███▌      | 94/263 [00:54<00:23,  7.34it/s]
Continual predict:  36%|███▌      | 95/263 [00:54<00:22,  7.44it/s]
Continual predict:  37%|███▋      | 96/263 [00:55<00:22,  7.51it/s]
Continual predict:  37%|███▋      | 97/263 [00:55<00:21,  7.57it/s]
Continual predict:  37%|███▋      | 98/263 [00:55<00:22,  7.42it/s]
Continual predict:  38%|███▊      | 99/263 [00:55<00:22,  7.29it/s]
Continual predict:  38%|███▊      | 100/263 [00:55<00:22,  7.21it/s]
Continual predict:  38%|███▊      | 101/263 [00:55<00:22,  7.35it/s]
Continual predict:  39%|███▉      | 102/263 [00:55<00:21,  7.48it/s]
Continual predict:  39%|███▉      | 103/263 [00:55<00:21,  7.56it/s]
Continual predict:  40%|███▉      | 104/263 [00:56<00:20,  7.60it/s]
Continual predict:  40%|███▉      | 105/263 [00:56<00:21,  7.43it/s]
Continual predict:  40%|████      | 106/263 [00:56<00:20,  7.54it/s]
Continual predict:  41%|████      | 107/263 [00:56<00:20,  7.62it/s]
Continual predict:  41%|████      | 108/263 [00:56<00:20,  7.61it/s]
Continual predict:  41%|████▏     | 109/263 [00:56<00:20,  7.64it/s]
Continual predict:  42%|████▏     | 110/263 [00:56<00:19,  7.66it/s]
Continual predict:  42%|████▏     | 111/263 [00:56<00:19,  7.68it/s]
Continual predict:  43%|████▎     | 112/263 [00:57<00:19,  7.68it/s]
Continual predict:  43%|████▎     | 113/263 [00:57<00:19,  7.70it/s]
Continual predict:  43%|████▎     | 114/263 [00:57<00:19,  7.69it/s]
Continual predict:  44%|████▎     | 115/263 [00:57<00:19,  7.49it/s]
Continual predict:  44%|████▍     | 116/263 [00:57<00:19,  7.55it/s]
Continual predict:  44%|████▍     | 117/263 [00:57<00:19,  7.39it/s]
Continual predict:  45%|████▍     | 118/263 [00:57<00:19,  7.27it/s]
Continual predict:  45%|████▌     | 119/263 [00:58<00:19,  7.39it/s]
Continual predict:  46%|████▌     | 120/263 [00:58<00:19,  7.48it/s]
Continual predict:  46%|████▌     | 121/263 [01:02<03:18,  1.40s/it]
Continual predict:  46%|████▋     | 122/263 [01:06<05:09,  2.20s/it]
Continual predict:  47%|████▋     | 123/263 [01:11<07:10,  3.08s/it]
Continual predict:  47%|████▋     | 124/263 [01:13<06:32,  2.83s/it]
Continual predict:  48%|████▊     | 125/263 [01:18<07:45,  3.38s/it]
Continual predict:  48%|████▊     | 126/263 [01:23<08:33,  3.75s/it]
Continual predict:  48%|████▊     | 127/263 [01:24<06:56,  3.06s/it]
Continual predict:  49%|████▊     | 128/263 [01:27<06:37,  2.94s/it]
Continual predict:  49%|████▉     | 129/263 [01:30<06:47,  3.04s/it]
Continual predict:  49%|████▉     | 130/263 [01:33<06:37,  2.99s/it]
Continual predict:  50%|████▉     | 131/263 [01:35<05:59,  2.72s/it]
Continual predict:  50%|█████     | 132/263 [01:38<05:49,  2.67s/it]
Continual predict:  51%|█████     | 133/263 [01:41<06:01,  2.78s/it]
Continual predict:  51%|█████     | 134/263 [01:46<07:54,  3.68s/it]
Continual predict:  51%|█████▏    | 135/263 [01:51<08:25,  3.95s/it]
Continual predict:  52%|█████▏    | 136/263 [01:54<07:48,  3.69s/it]
Continual predict:  52%|█████▏    | 137/263 [01:56<06:46,  3.22s/it]
Continual predict:  52%|█████▏    | 138/263 [01:57<05:07,  2.46s/it]
Continual predict:  53%|█████▎    | 139/263 [01:58<04:14,  2.05s/it]
Continual predict:  53%|█████▎    | 140/263 [02:01<04:31,  2.20s/it]
Continual predict:  54%|█████▎    | 141/263 [02:07<07:01,  3.45s/it]
Continual predict:  54%|█████▍    | 142/263 [02:11<07:31,  3.73s/it]
Continual predict:  54%|█████▍    | 143/263 [02:15<07:06,  3.56s/it]
Continual predict:  55%|█████▍    | 144/263 [02:24<10:22,  5.23s/it]
Continual predict:  55%|█████▌    | 145/263 [02:28<09:40,  4.92s/it]
Continual predict:  56%|█████▌    | 146/263 [02:30<07:43,  3.96s/it]
Continual predict:  56%|█████▌    | 147/263 [02:34<07:48,  4.04s/it]
Continual predict:  56%|█████▋    | 148/263 [02:36<06:35,  3.44s/it]
Continual predict:  57%|█████▋    | 149/263 [02:39<06:18,  3.32s/it]
Continual predict:  57%|█████▋    | 150/263 [02:41<05:23,  2.87s/it]
Continual predict:  57%|█████▋    | 151/263 [02:44<05:23,  2.89s/it]
Continual predict:  58%|█████▊    | 152/263 [02:49<06:45,  3.65s/it]
Continual predict:  58%|█████▊    | 153/263 [02:51<05:58,  3.26s/it]
Continual predict:  59%|█████▊    | 154/263 [02:56<06:40,  3.68s/it]
Continual predict:  59%|█████▉    | 155/263 [02:59<06:28,  3.60s/it]
Continual predict:  59%|█████▉    | 156/263 [03:03<06:13,  3.49s/it]
Continual predict:  60%|█████▉    | 157/263 [03:06<06:15,  3.54s/it]
Continual predict:  60%|██████    | 158/263 [03:07<04:47,  2.74s/it]
Continual predict:  60%|██████    | 159/263 [03:14<06:48,  3.93s/it]
Continual predict:  61%|██████    | 160/263 [03:16<05:50,  3.41s/it]
Continual predict:  61%|██████    | 161/263 [03:19<05:16,  3.10s/it]
Continual predict:  62%|██████▏   | 162/263 [03:23<06:00,  3.56s/it]
Continual predict:  62%|██████▏   | 163/263 [03:25<05:07,  3.08s/it]
Continual predict:  62%|██████▏   | 164/263 [03:27<04:37,  2.80s/it]
Continual predict:  63%|██████▎   | 165/263 [03:30<04:40,  2.86s/it]
Continual predict:  63%|██████▎   | 166/263 [03:32<04:13,  2.62s/it]
Continual predict:  63%|██████▎   | 167/263 [03:34<03:32,  2.21s/it]
Continual predict:  64%|██████▍   | 168/263 [03:36<03:28,  2.19s/it]
Continual predict:  64%|██████▍   | 169/263 [03:37<03:13,  2.06s/it]
Continual predict:  65%|██████▍   | 170/263 [03:40<03:35,  2.32s/it]
Continual predict:  65%|██████▌   | 171/263 [03:42<03:08,  2.05s/it]
Continual predict:  65%|██████▌   | 172/263 [03:44<03:17,  2.17s/it]
Continual predict:  66%|██████▌   | 173/263 [03:48<03:59,  2.66s/it]
Continual predict:  66%|██████▌   | 174/263 [03:49<03:15,  2.20s/it]
Continual predict:  67%|██████▋   | 175/263 [03:53<03:42,  2.53s/it]
Continual predict:  67%|██████▋   | 176/263 [03:55<03:45,  2.59s/it]
Continual predict:  67%|██████▋   | 177/263 [03:58<03:42,  2.59s/it]
Continual predict:  68%|██████▊   | 178/263 [03:59<03:04,  2.17s/it]
Continual predict:  68%|██████▊   | 179/263 [04:00<02:33,  1.83s/it]
Continual predict:  68%|██████▊   | 180/263 [04:01<02:18,  1.66s/it]
Continual predict:  69%|██████▉   | 181/263 [04:03<02:06,  1.54s/it]
Continual predict:  69%|██████▉   | 182/263 [04:07<03:17,  2.44s/it]
Continual predict:  70%|██████▉   | 183/263 [04:07<02:21,  1.77s/it]
Continual predict:  70%|██████▉   | 184/263 [04:08<02:03,  1.56s/it]
Continual predict:  70%|███████   | 185/263 [04:12<02:39,  2.04s/it]
Continual predict:  71%|███████   | 186/263 [04:12<02:10,  1.70s/it]
Continual predict:  71%|███████   | 187/263 [04:16<02:41,  2.13s/it]
Continual predict:  71%|███████▏  | 188/263 [04:18<02:50,  2.27s/it]
Continual predict:  72%|███████▏  | 189/263 [04:23<03:43,  3.02s/it]
Continual predict:  72%|███████▏  | 190/263 [04:26<03:41,  3.04s/it]
Continual predict:  73%|███████▎  | 191/263 [04:27<02:59,  2.49s/it]
Continual predict:  73%|███████▎  | 192/263 [04:30<03:10,  2.69s/it]
Continual predict:  73%|███████▎  | 193/263 [04:35<03:42,  3.18s/it]
Continual predict:  74%|███████▍  | 194/263 [04:39<04:08,  3.61s/it]
Continual predict:  74%|███████▍  | 195/263 [04:44<04:35,  4.05s/it]
Continual predict:  75%|███████▍  | 196/263 [04:48<04:14,  3.80s/it]
Continual predict:  75%|███████▍  | 197/263 [04:51<03:58,  3.61s/it]
Continual predict:  75%|███████▌  | 198/263 [04:58<05:02,  4.65s/it]
Continual predict:  76%|███████▌  | 199/263 [04:59<03:48,  3.57s/it]
Continual predict:  76%|███████▌  | 200/263 [05:00<03:03,  2.91s/it]
Continual predict:  76%|███████▋  | 201/263 [05:03<02:50,  2.75s/it]
Continual predict:  77%|███████▋  | 202/263 [05:06<02:54,  2.86s/it]
Continual predict:  77%|███████▋  | 203/263 [05:11<03:28,  3.48s/it]
Continual predict:  78%|███████▊  | 204/263 [05:14<03:14,  3.30s/it]
Continual predict:  78%|███████▊  | 205/263 [05:20<04:04,  4.22s/it]
Continual predict:  78%|███████▊  | 206/263 [05:23<03:33,  3.74s/it]
Continual predict:  79%|███████▊  | 207/263 [05:26<03:26,  3.70s/it]
Continual predict:  79%|███████▉  | 208/263 [05:32<04:00,  4.37s/it]
Continual predict:  79%|███████▉  | 209/263 [05:35<03:26,  3.83s/it]
Continual predict:  80%|███████▉  | 210/263 [05:36<02:47,  3.16s/it]
Continual predict:  80%|████████  | 211/263 [05:39<02:33,  2.95s/it]
Continual predict:  81%|████████  | 212/263 [05:42<02:30,  2.95s/it]
Continual predict:  81%|████████  | 213/263 [05:45<02:36,  3.14s/it]
Continual predict:  81%|████████▏ | 214/263 [05:49<02:37,  3.22s/it]
Continual predict:  82%|████████▏ | 215/263 [05:49<01:57,  2.45s/it]
Continual predict:  82%|████████▏ | 216/263 [05:52<02:01,  2.59s/it]
Continual predict:  83%|████████▎ | 217/263 [05:54<01:50,  2.39s/it]
Continual predict:  83%|████████▎ | 218/263 [05:58<02:05,  2.79s/it]
Continual predict:  83%|████████▎ | 219/263 [06:01<02:05,  2.85s/it]
Continual predict:  84%|████████▎ | 220/263 [06:05<02:16,  3.17s/it]
Continual predict:  84%|████████▍ | 221/263 [06:05<01:35,  2.27s/it]
Continual predict:  84%|████████▍ | 222/263 [06:05<01:08,  1.67s/it]
Continual predict:  85%|████████▍ | 223/263 [06:05<00:49,  1.23s/it]
Continual predict:  85%|████████▌ | 224/263 [06:06<00:35,  1.09it/s]
Continual predict:  86%|████████▌ | 225/263 [06:06<00:26,  1.43it/s]
Continual predict:  86%|████████▌ | 226/263 [06:06<00:20,  1.82it/s]
Continual predict:  86%|████████▋ | 227/263 [06:06<00:16,  2.18it/s]
Continual predict:  87%|████████▋ | 228/263 [06:07<00:13,  2.53it/s]
Continual predict:  87%|████████▋ | 229/263 [06:07<00:11,  2.93it/s]
Continual predict:  87%|████████▋ | 230/263 [06:07<00:11,  2.81it/s]
Continual predict:  88%|████████▊ | 231/263 [06:07<00:09,  3.24it/s]
Continual predict:  88%|████████▊ | 232/263 [06:08<00:08,  3.64it/s]
Continual predict:  89%|████████▊ | 233/263 [06:08<00:07,  3.99it/s]
Continual predict:  89%|████████▉ | 234/263 [06:08<00:06,  4.26it/s]
Continual predict:  89%|████████▉ | 235/263 [06:08<00:06,  4.50it/s]
Continual predict:  90%|████████▉ | 236/263 [06:08<00:05,  4.67it/s]
Continual predict:  90%|█████████ | 237/263 [06:08<00:05,  4.80it/s]
Continual predict:  90%|█████████ | 238/263 [06:09<00:05,  4.90it/s]
Continual predict:  91%|█████████ | 239/263 [06:09<00:04,  4.97it/s]
Continual predict:  91%|█████████▏| 240/263 [06:09<00:05,  4.19it/s]
Continual predict:  92%|█████████▏| 241/263 [06:09<00:04,  4.44it/s]
Continual predict:  92%|█████████▏| 242/263 [06:10<00:04,  4.61it/s]
Continual predict:  92%|█████████▏| 243/263 [06:10<00:04,  4.75it/s]
Continual predict:  93%|█████████▎| 244/263 [06:10<00:03,  4.85it/s]
Continual predict:  93%|█████████▎| 245/263 [06:10<00:03,  4.93it/s]
Continual predict:  94%|█████████▎| 246/263 [06:10<00:03,  4.99it/s]
Continual predict:  94%|█████████▍| 247/263 [06:11<00:03,  4.66it/s]
Continual predict:  94%|█████████▍| 248/263 [06:11<00:03,  4.80it/s]
Continual predict:  95%|█████████▍| 249/263 [06:11<00:03,  4.54it/s]
Continual predict:  95%|█████████▌| 250/263 [06:11<00:02,  4.71it/s]
Continual predict:  95%|█████████▌| 251/263 [06:12<00:02,  4.49it/s]
Continual predict:  96%|█████████▌| 252/263 [06:13<00:05,  1.96it/s]
Continual predict:  96%|█████████▌| 253/263 [06:13<00:04,  2.32it/s]
Continual predict:  97%|█████████▋| 254/263 [06:13<00:03,  2.77it/s]
Continual predict:  97%|█████████▋| 255/263 [06:13<00:02,  3.21it/s]
Continual predict:  97%|█████████▋| 256/263 [06:14<00:01,  3.61it/s]
Continual predict:  98%|█████████▊| 257/263 [06:14<00:02,  2.52it/s]
Continual predict:  98%|█████████▊| 258/263 [06:14<00:01,  2.97it/s]
Continual predict:  98%|█████████▊| 259/263 [06:15<00:01,  3.40it/s]
Continual predict:  99%|█████████▉| 260/263 [06:15<00:00,  3.77it/s]
Continual predict:  99%|█████████▉| 261/263 [06:15<00:00,  2.91it/s]
Continual predict: 100%|█████████▉| 262/263 [06:16<00:00,  3.28it/s]
Continual predict: 100%|██████████| 263/263 [06:16<00:00,  3.48it/s]
Continual predict: 100%|██████████| 263/263 [06:16<00:00,  1.43s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=20]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (263/263)  label_acc=100.0% (263/263)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                        =       15.0
+  predict_eval_rougeL                                          =    58.7031
+  predict_eval_rougeL_for_CL                                   =    58.7031
+  predict_eval_rougeL_for_task1290_xsum_summarization          =    29.5925
+  predict_eval_rougeL_for_task1572_samsum_summary              =    41.3728
+  predict_eval_rougeL_for_task181_outcome_extraction           =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification =       91.0
+  predict_exact_match                                          =     43.346
+  predict_exact_match_for_CL                                   =     43.346
+  predict_exact_match_for_task1290_xsum_summarization          =        0.0
+  predict_exact_match_for_task1572_samsum_summary              =        0.0
+  predict_exact_match_for_task181_outcome_extraction           =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification =       91.0
+  predict_gen_len                                              =     12.654
+  predict_global_step                                          =        165
+  predict_rouge1                                               =    62.1102
+  predict_rouge1_for_CL                                        =    62.1102
+  predict_rouge1_for_task1290_xsum_summarization               =      36.75
+  predict_rouge1_for_task1572_samsum_summary                   =    50.3879
+  predict_rouge1_for_task181_outcome_extraction                =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification      =       91.0
+  predict_runtime                                              = 0:06:16.28
+  predict_samples                                              =        263
+  predict_samples_per_second                                   =      0.699
+  predict_steps_per_second                                     =      0.699
+[2026-04-29 21:39:02,840] [INFO] [launch.py:367:main] Process 739565 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/all_results.json b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..2cf0b573bbcf7588f6be8b6e37f750227764ce1e
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/all_results.json
@@ -0,0 +1,36 @@
+{
+  "predict_exact_match": 50.6887,
+  "predict_rouge1": 66.0415,
+  "predict_eval_rougeL": 63.5731,
+  "predict_gen_len": 10.3003,
+  "predict_runtime": 453.93795572197996,
+  "predict_samples": 363,
+  "predict_samples_per_second": 0.8,
+  "predict_steps_per_second": 0.8,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 50.6887,
+  "predict_rouge1_for_CL": 66.0415,
+  "predict_eval_rougeL_for_CL": 63.5731,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "train_runtime": 3911.8844,
+  "train_samples_per_second": 3.834,
+  "train_steps_per_second": 0.123,
+  "total_flos": 3.6802493662468506e+17,
+  "train_loss": 0.33569894383351007,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f2e126d9f06ffddfaf5aac1f625632109b98b62f
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00405881f7654b1ad6419166874b684654334be4aac55973eb97e0f1b213f3af
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f6d1358320bdedfc3d9930e37ef2f95619c080c0
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46ec6da1f2e0f02cf346aa39925b52cf44d430a43a8de1ef84d606b01f30298f
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..297ed68ca4332f876d5c0df42e3df7c07a2a353b
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f578eb59f54e3c06e679fd7b451f110705cff18146c8b2a0aeb9ec2a4b9388b3
+size 847522913
diff --git a/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/train.log b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..3ab6e1fe6b4f6a858687de94942bf33e447e8500
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/train.log
@@ -0,0 +1,519 @@
+[2026-04-29 21:39:08,727] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 21:39:08,727] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29504 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task002_quoref_answer_generation --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation --cur_task_id 4 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[2026-04-29 21:39:11,983] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 21:39:11,983] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 21:39:11,983] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 21:39:11,983] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 21:39:11,983] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 21:39:11,983] [INFO] [launch.py:272:main] process 786881 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task002_quoref_answer_generation', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation', '--cur_task_id', '4', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights']
+
+============================================================
+Task 4: task002_quoref_answer_generation
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/363
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 11/291 [00:00<00:02, 102.70it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 146.05it/s]
Loading weights:  17%|█▋        | 50/291 [00:00<00:01, 160.73it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 168.72it/s]
Loading weights:  32%|███▏      | 92/291 [00:00<00:01, 186.29it/s]
Loading weights:  38%|███▊      | 111/291 [00:00<00:00, 182.70it/s]
Loading weights:  45%|████▍     | 130/291 [00:00<00:00, 177.55it/s]
Loading weights:  51%|█████     | 148/291 [00:00<00:00, 175.66it/s]
Loading weights:  57%|█████▋    | 166/291 [00:00<00:00, 173.71it/s]
Loading weights:  65%|██████▍   | 189/291 [00:01<00:00, 187.70it/s]
Loading weights:  71%|███████▏  | 208/291 [00:01<00:00, 178.12it/s]
Loading weights:  78%|███████▊  | 227/291 [00:01<00:00, 177.78it/s]
Loading weights:  85%|████████▍ | 247/291 [00:01<00:00, 177.94it/s]
Loading weights:  93%|█████████▎| 270/291 [00:01<00:00, 190.16it/s]
Loading weights: 100%|█████████▉| 290/291 [00:01<00:00, 187.54it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 178.57it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=675)
+[FIX] Re-initialized lora_A/B in 320 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(4, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 4 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6749.741M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 4 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 4 sigs, cur=task002_quoref_answer_generation, mapping={'task002_quoref_answer_generation': 0, 'task1572_samsum_summary': 4, 'task363_sst2_polarity_classification': 3, 'task1290_xsum_summarization': 2, 'task181_outcome_extraction': 1}
+[SGWI] Task 4: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 21:40:09,703] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:08<1:05:18,  8.18s/it]
  0%|          | 2/480 [00:16<1:04:34,  8.10s/it]
  1%|          | 3/480 [00:23<1:02:44,  7.89s/it]
  1%|          | 4/480 [00:32<1:03:39,  8.02s/it]
  1%|          | 5/480 [00:40<1:03:36,  8.03s/it]
  1%|▏         | 6/480 [00:48<1:04:12,  8.13s/it]
  1%|▏         | 7/480 [00:56<1:04:01,  8.12s/it]
  2%|▏         | 8/480 [01:04<1:04:09,  8.16s/it]
  2%|▏         | 9/480 [01:13<1:04:38,  8.24s/it]
  2%|▏         | 10/480 [01:21<1:04:33,  8.24s/it]
                                                  
{'loss': '2.214', 'grad_norm': '0.7408', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [01:21<1:04:33,  8.24s/it]
  2%|▏         | 11/480 [01:29<1:04:09,  8.21s/it]
  2%|▎         | 12/480 [01:37<1:03:34,  8.15s/it]
  3%|▎         | 13/480 [01:45<1:02:22,  8.01s/it]
  3%|▎         | 14/480 [01:53<1:02:56,  8.11s/it]
  3%|▎         | 15/480 [02:01<1:03:12,  8.16s/it]
  3%|▎         | 16/480 [02:09<1:01:55,  8.01s/it]
  4%|▎         | 17/480 [02:17<1:01:54,  8.02s/it]
  4%|▍         | 18/480 [02:25<1:01:14,  7.95s/it]
  4%|▍         | 19/480 [02:33<1:01:21,  7.99s/it]
  4%|▍         | 20/480 [02:41<1:02:13,  8.12s/it]
                                                  
{'loss': '1.322', 'grad_norm': '0.7275', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [02:41<1:02:13,  8.12s/it]
  4%|▍         | 21/480 [02:50<1:02:15,  8.14s/it]
  5%|▍         | 22/480 [02:58<1:03:03,  8.26s/it]
  5%|▍         | 23/480 [03:06<1:01:31,  8.08s/it]
  5%|▌         | 24/480 [03:14<1:01:11,  8.05s/it]
  5%|▌         | 25/480 [03:22<1:01:08,  8.06s/it]
  5%|▌         | 26/480 [03:30<1:00:38,  8.01s/it]
  6%|▌         | 27/480 [03:38<1:00:54,  8.07s/it]
  6%|▌         | 28/480 [03:46<1:00:52,  8.08s/it]
  6%|▌         | 29/480 [03:54<1:00:20,  8.03s/it]
  6%|▋         | 30/480 [04:02<59:05,  7.88s/it]  
                                                
{'loss': '0.7611', 'grad_norm': '0.7143', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [04:02<59:05,  7.88s/it]
  6%|▋         | 31/480 [04:10<1:01:01,  8.16s/it]
  7%|▋         | 32/480 [04:12<46:46,  6.27s/it]  
  7%|▋         | 33/480 [04:20<49:44,  6.68s/it]
  7%|▋         | 34/480 [04:28<53:30,  7.20s/it]
  7%|▋         | 35/480 [04:36<55:35,  7.49s/it]
  8%|▊         | 36/480 [04:45<57:22,  7.75s/it]
  8%|▊         | 37/480 [04:53<58:36,  7.94s/it]
  8%|▊         | 38/480 [05:02<1:00:22,  8.20s/it]
  8%|▊         | 39/480 [05:10<59:10,  8.05s/it]  
  8%|▊         | 40/480 [05:18<59:21,  8.09s/it]
                                                
{'loss': '0.4422', 'grad_norm': '0.4731', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [05:18<59:21,  8.09s/it]
  9%|▊         | 41/480 [05:26<59:37,  8.15s/it]
  9%|▉         | 42/480 [05:34<59:13,  8.11s/it]
  9%|▉         | 43/480 [05:43<59:40,  8.19s/it]
  9%|▉         | 44/480 [05:51<59:24,  8.17s/it]
  9%|▉         | 45/480 [05:58<58:28,  8.07s/it]
 10%|▉         | 46/480 [06:07<58:18,  8.06s/it]
 10%|▉         | 47/480 [06:14<57:56,  8.03s/it]
 10%|█         | 48/480 [06:22<57:44,  8.02s/it]
 10%|█         | 49/480 [06:31<57:50,  8.05s/it]
 10%|█         | 50/480 [06:39<57:56,  8.09s/it]
                                                
{'loss': '0.4235', 'grad_norm': '0.5132', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [06:39<57:56,  8.09s/it]
 11%|█         | 51/480 [06:47<57:16,  8.01s/it]
 11%|█         | 52/480 [06:55<57:49,  8.11s/it]
 11%|█         | 53/480 [07:03<58:23,  8.21s/it]
 11%|█▏        | 54/480 [07:12<58:16,  8.21s/it]
 11%|█▏        | 55/480 [07:20<57:39,  8.14s/it]
 12%|█▏        | 56/480 [07:27<55:43,  7.89s/it]
 12%|█▏        | 57/480 [07:35<56:17,  7.99s/it]
 12%|█▏        | 58/480 [07:43<56:02,  7.97s/it]
 12%|█▏        | 59/480 [07:51<55:57,  7.98s/it]
 12%|█▎        | 60/480 [07:59<54:55,  7.85s/it]
                                                
{'loss': '0.3831', 'grad_norm': '0.1196', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [07:59<54:55,  7.85s/it]
 13%|█▎        | 61/480 [08:06<54:33,  7.81s/it]
 13%|█▎        | 62/480 [08:14<55:04,  7.90s/it]
 13%|█▎        | 63/480 [08:23<55:28,  7.98s/it]
 13%|█▎        | 64/480 [08:25<43:08,  6.22s/it]
 14%|█▎        | 65/480 [08:33<46:45,  6.76s/it]
 14%|█▍        | 66/480 [08:41<48:57,  7.10s/it]
 14%|█▍        | 67/480 [08:49<51:45,  7.52s/it]
 14%|█▍        | 68/480 [08:57<53:18,  7.76s/it]
 14%|█▍        | 69/480 [09:05<53:13,  7.77s/it]
 15%|█▍        | 70/480 [09:13<53:33,  7.84s/it]
                                                
{'loss': '0.3979', 'grad_norm': '0.6971', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [09:13<53:33,  7.84s/it]
 15%|█▍        | 71/480 [09:21<53:59,  7.92s/it]
 15%|█▌        | 72/480 [09:29<53:41,  7.89s/it]
 15%|█▌        | 73/480 [09:37<53:58,  7.96s/it]
 15%|█▌        | 74/480 [09:45<53:04,  7.84s/it]
 16%|█▌        | 75/480 [09:53<53:52,  7.98s/it]
 16%|█▌        | 76/480 [10:01<52:53,  7.86s/it]
 16%|█▌        | 77/480 [10:09<52:51,  7.87s/it]
 16%|█▋        | 78/480 [10:17<53:30,  7.99s/it]
 16%|█▋        | 79/480 [10:25<53:26,  8.00s/it]
 17%|█▋        | 80/480 [10:33<54:05,  8.11s/it]
                                                
{'loss': '0.3478', 'grad_norm': '0.4103', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [10:33<54:05,  8.11s/it]
 17%|█▋        | 81/480 [10:41<53:34,  8.06s/it]
 17%|█▋        | 82/480 [10:50<54:26,  8.21s/it]
 17%|█▋        | 83/480 [10:58<54:22,  8.22s/it]
 18%|█▊        | 84/480 [11:06<54:02,  8.19s/it]
 18%|█▊        | 85/480 [11:14<52:27,  7.97s/it]
 18%|█▊        | 86/480 [11:22<52:18,  7.97s/it]
 18%|█▊        | 87/480 [11:29<51:32,  7.87s/it]
 18%|█▊        | 88/480 [11:38<52:25,  8.03s/it]
 19%|█▊        | 89/480 [11:46<52:53,  8.12s/it]
 19%|█▉        | 90/480 [11:54<53:32,  8.24s/it]
                                                
{'loss': '0.3347', 'grad_norm': '0.9899', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [11:54<53:32,  8.24s/it]
 19%|█▉        | 91/480 [12:03<53:59,  8.33s/it]
 19%|█▉        | 92/480 [12:11<54:01,  8.35s/it]
 19%|█▉        | 93/480 [12:20<53:31,  8.30s/it]
 20%|█▉        | 94/480 [12:28<53:29,  8.32s/it]
 20%|█▉        | 95/480 [12:36<52:34,  8.19s/it]
 20%|██        | 96/480 [12:38<40:12,  6.28s/it]
 20%|██        | 97/480 [12:45<43:08,  6.76s/it]
 20%|██        | 98/480 [12:54<46:34,  7.32s/it]
 21%|██        | 99/480 [13:02<47:08,  7.42s/it]
 21%|██        | 100/480 [13:10<48:07,  7.60s/it]
                                                 
{'loss': '0.3397', 'grad_norm': '0.209', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [13:10<48:07,  7.60s/it]
 21%|██        | 101/480 [13:18<49:37,  7.86s/it]
 21%|██▏       | 102/480 [13:26<50:07,  7.96s/it]
 21%|██▏       | 103/480 [13:34<50:11,  7.99s/it]
 22%|██▏       | 104/480 [13:43<50:18,  8.03s/it]
 22%|██▏       | 105/480 [13:50<49:46,  7.96s/it]
 22%|██▏       | 106/480 [13:59<50:57,  8.17s/it]
 22%|██▏       | 107/480 [14:08<51:29,  8.28s/it]
 22%|██▎       | 108/480 [14:15<49:26,  7.97s/it]
 23%|██▎       | 109/480 [14:23<48:46,  7.89s/it]
 23%|██▎       | 110/480 [14:30<48:28,  7.86s/it]
                                                 
{'loss': '0.355', 'grad_norm': '0.3824', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [14:30<48:28,  7.86s/it]
 23%|██▎       | 111/480 [14:39<48:55,  7.96s/it]
 23%|██▎       | 112/480 [14:47<49:59,  8.15s/it]
 24%|██▎       | 113/480 [14:55<48:55,  8.00s/it]
 24%|██▍       | 114/480 [15:02<47:51,  7.84s/it]
 24%|██▍       | 115/480 [15:11<48:43,  8.01s/it]
 24%|██▍       | 116/480 [15:19<48:51,  8.05s/it]
 24%|██▍       | 117/480 [15:27<48:45,  8.06s/it]
 25%|██▍       | 118/480 [15:35<49:20,  8.18s/it]
 25%|██▍       | 119/480 [15:43<48:52,  8.12s/it]
 25%|██▌       | 120/480 [15:52<49:46,  8.30s/it]
                                                 
{'loss': '0.3434', 'grad_norm': '0.5115', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [15:52<49:46,  8.30s/it]
 25%|██▌       | 121/480 [16:01<50:19,  8.41s/it]
 25%|██▌       | 122/480 [16:09<49:49,  8.35s/it]
 26%|██▌       | 123/480 [16:17<49:05,  8.25s/it]
 26%|██▌       | 124/480 [16:25<48:56,  8.25s/it]
 26%|██▌       | 125/480 [16:33<48:25,  8.19s/it]
 26%|██▋       | 126/480 [16:41<47:44,  8.09s/it]
 26%|██▋       | 127/480 [16:48<45:57,  7.81s/it]
 27%|██▋       | 128/480 [16:50<35:28,  6.05s/it]
 27%|██▋       | 129/480 [16:58<38:54,  6.65s/it]
 27%|██▋       | 130/480 [17:06<40:41,  6.98s/it]
                                                 
{'loss': '0.3591', 'grad_norm': '0.3004', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [17:06<40:41,  6.98s/it]
 27%|██▋       | 131/480 [17:13<41:29,  7.13s/it]
 28%|██▊       | 132/480 [17:22<42:55,  7.40s/it]
 28%|██▊       | 133/480 [17:29<43:31,  7.53s/it]
 28%|██▊       | 134/480 [17:37<43:48,  7.60s/it]
 28%|██▊       | 135/480 [17:45<43:29,  7.56s/it]
 28%|██▊       | 136/480 [17:53<44:01,  7.68s/it]
 29%|██▊       | 137/480 [18:01<45:42,  8.00s/it]
 29%|██▉       | 138/480 [18:09<45:54,  8.06s/it]
 29%|██▉       | 139/480 [18:18<46:21,  8.16s/it]
 29%|██▉       | 140/480 [18:26<46:41,  8.24s/it]
                                                 
{'loss': '0.3115', 'grad_norm': '0.5728', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [18:26<46:41,  8.24s/it]
 29%|██▉       | 141/480 [18:35<47:05,  8.33s/it]
 30%|██▉       | 142/480 [18:43<46:51,  8.32s/it]
 30%|██▉       | 143/480 [18:51<46:03,  8.20s/it]
 30%|███       | 144/480 [18:59<45:31,  8.13s/it]
 30%|███       | 145/480 [19:07<45:14,  8.10s/it]
 30%|███       | 146/480 [19:15<45:25,  8.16s/it]
 31%|███       | 147/480 [19:23<45:06,  8.13s/it]
 31%|███       | 148/480 [19:32<45:47,  8.27s/it]
 31%|███       | 149/480 [19:40<45:20,  8.22s/it]
 31%|███▏      | 150/480 [19:48<45:09,  8.21s/it]
                                                 
{'loss': '0.2916', 'grad_norm': '0.6867', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [19:48<45:09,  8.21s/it]
 31%|███▏      | 151/480 [19:56<44:01,  8.03s/it]
 32%|███▏      | 152/480 [20:04<43:35,  7.97s/it]
 32%|███▏      | 153/480 [20:12<43:39,  8.01s/it]
 32%|███▏      | 154/480 [20:20<44:06,  8.12s/it]
 32%|███▏      | 155/480 [20:28<44:14,  8.17s/it]
 32%|███▎      | 156/480 [20:37<44:24,  8.22s/it]
 33%|███▎      | 157/480 [20:46<44:58,  8.35s/it]
 33%|███▎      | 158/480 [20:53<44:06,  8.22s/it]
 33%|███▎      | 159/480 [21:02<44:26,  8.31s/it]
 33%|███▎      | 160/480 [21:04<33:55,  6.36s/it]
                                                 
{'loss': '0.3599', 'grad_norm': '3.484', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [21:04<33:55,  6.36s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<01:08,  1.45it/s]
+
Eval @ step 160:   2%|▏         | 2/100 [00:00<00:44,  2.20it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:01<00:36,  2.65it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:01<00:34,  2.81it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:02<00:37,  2.55it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:02<00:34,  2.71it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:02<00:33,  2.80it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:03<00:34,  2.68it/s]
+
Eval @ step 160:   9%|▉         | 9/100 [00:03<00:36,  2.46it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:03<00:34,  2.59it/s]
+
Eval @ step 160:  11%|█         | 11/100 [00:04<00:36,  2.47it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:04<00:31,  2.82it/s]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:05<00:36,  2.36it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:05<00:43,  1.99it/s]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:06<00:41,  2.03it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:06<00:38,  2.16it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:06<00:32,  2.53it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:07<00:35,  2.31it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:07<00:35,  2.30it/s]
+
Eval @ step 160:  20%|██        | 20/100 [00:08<00:32,  2.44it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:08<00:28,  2.77it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:08<00:29,  2.63it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:09<00:39,  1.96it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:10<00:36,  2.08it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:10<00:31,  2.38it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:10<00:33,  2.24it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:11<00:31,  2.31it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:11<00:27,  2.66it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:11<00:22,  3.20it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:12<00:25,  2.79it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:12<00:23,  2.97it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:12<00:20,  3.27it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:13<00:25,  2.68it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:13<00:22,  2.91it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:13<00:20,  3.19it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:14<00:23,  2.75it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:14<00:19,  3.26it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:14<00:22,  2.77it/s]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:15<00:19,  3.12it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:15<00:19,  3.11it/s]
+
Eval @ step 160:  41%|████      | 41/100 [00:15<00:16,  3.61it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:16<00:19,  2.96it/s]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:16<00:21,  2.71it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:17<00:21,  2.58it/s]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:17<00:18,  2.94it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:17<00:16,  3.22it/s]
+
Eval @ step 160:  47%|████▋     | 47/100 [00:17<00:16,  3.17it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:18<00:15,  3.26it/s]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:18<00:17,  3.00it/s]
+
Eval @ step 160:  50%|█████     | 50/100 [00:18<00:18,  2.65it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:19<00:16,  2.97it/s]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:19<00:18,  2.64it/s]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:19<00:16,  2.91it/s]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:20<00:13,  3.42it/s]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:20<00:13,  3.32it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:21<00:20,  2.11it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:21<00:17,  2.49it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:21<00:16,  2.52it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:22<00:14,  2.89it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:22<00:14,  2.76it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:23<00:15,  2.49it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [00:23<00:15,  2.48it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [00:23<00:14,  2.60it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [00:24<00:18,  1.99it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:25<00:16,  2.12it/s]
+
Eval @ step 160:  66%|██████▌   | 66/100 [00:25<00:14,  2.41it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [00:25<00:11,  2.93it/s]
+
Eval @ step 160:  68%|██████▊   | 68/100 [00:25<00:10,  2.99it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:26<00:09,  3.12it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [00:26<00:12,  2.39it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:27<00:13,  2.17it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [00:27<00:10,  2.56it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:27<00:09,  2.89it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [00:28<00:08,  2.92it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:28<00:08,  2.90it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:28<00:07,  3.07it/s]
+
Eval @ step 160:  77%|███████▋  | 77/100 [00:29<00:09,  2.45it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:29<00:09,  2.43it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [00:29<00:07,  2.77it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:30<00:07,  2.68it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:30<00:07,  2.41it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [00:31<00:06,  2.75it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [00:31<00:06,  2.59it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [00:32<00:08,  1.97it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:32<00:06,  2.44it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [00:32<00:04,  2.81it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [00:33<00:04,  2.82it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [00:33<00:05,  2.33it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [00:34<00:04,  2.38it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [00:34<00:04,  2.09it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [00:35<00:04,  1.86it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [00:35<00:03,  2.23it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:36<00:03,  2.29it/s]
+
Eval @ step 160:  94%|█████████▍| 94/100 [00:36<00:02,  2.10it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:37<00:02,  2.20it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [00:37<00:01,  2.02it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:38<00:01,  1.80it/s]
+
Eval @ step 160:  98%|█████████▊| 98/100 [00:39<00:01,  1.54it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [00:39<00:00,  1.82it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [00:39<00:00,  2.19it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=71.3905 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '71.39', 'epoch': '5'}
+
 33%|███▎      | 160/480 [21:44<33:55,  6.36s/it]
 34%|███▎      | 161/480 [21:52<1:40:36, 18.92s/it]
 34%|███▍      | 162/480 [22:00<1:22:46, 15.62s/it]
 34%|███▍      | 163/480 [22:08<1:10:47, 13.40s/it]
 34%|███▍      | 164/480 [22:17<1:03:24, 12.04s/it]
 34%|███▍      | 165/480 [22:25<56:35, 10.78s/it]  
 35%|███▍      | 166/480 [22:33<52:27, 10.02s/it]
 35%|███▍      | 167/480 [22:41<49:00,  9.39s/it]
 35%|███▌      | 168/480 [22:49<46:02,  8.85s/it]
 35%|███▌      | 169/480 [22:56<44:09,  8.52s/it]
 35%|███▌      | 170/480 [23:04<42:59,  8.32s/it]
                                                 
{'loss': '0.2935', 'grad_norm': '0.2572', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [23:04<42:59,  8.32s/it]
 36%|███▌      | 171/480 [23:12<42:33,  8.26s/it]
 36%|███▌      | 172/480 [23:20<41:59,  8.18s/it]
 36%|███▌      | 173/480 [23:28<41:32,  8.12s/it]
 36%|███▋      | 174/480 [23:36<40:52,  8.01s/it]
 36%|███▋      | 175/480 [23:45<41:52,  8.24s/it]
 37%|███▋      | 176/480 [23:53<42:02,  8.30s/it]
 37%|███▋      | 177/480 [24:01<40:55,  8.11s/it]
 37%|███▋      | 178/480 [24:09<40:54,  8.13s/it]
 37%|███▋      | 179/480 [24:18<41:25,  8.26s/it]
 38%|███▊      | 180/480 [24:26<41:21,  8.27s/it]
                                                 
{'loss': '0.3218', 'grad_norm': '0.3177', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [24:26<41:21,  8.27s/it]
 38%|███▊      | 181/480 [24:34<40:38,  8.16s/it]
 38%|███▊      | 182/480 [24:42<40:16,  8.11s/it]
 38%|███▊      | 183/480 [24:50<40:35,  8.20s/it]
 38%|███▊      | 184/480 [24:59<40:47,  8.27s/it]
 39%|███▊      | 185/480 [25:07<40:22,  8.21s/it]
 39%|███▉      | 186/480 [25:15<39:57,  8.16s/it]
 39%|███▉      | 187/480 [25:22<39:10,  8.02s/it]
 39%|███▉      | 188/480 [25:31<39:31,  8.12s/it]
 39%|███▉      | 189/480 [25:39<39:30,  8.15s/it]
 40%|███▉      | 190/480 [25:47<39:38,  8.20s/it]
                                                 
{'loss': '0.2735', 'grad_norm': '0.03591', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [25:47<39:38,  8.20s/it]
 40%|███▉      | 191/480 [25:56<39:27,  8.19s/it]
 40%|████      | 192/480 [25:57<30:07,  6.28s/it]
 40%|████      | 193/480 [26:06<32:43,  6.84s/it]
 40%|████      | 194/480 [26:13<34:02,  7.14s/it]
 41%|████      | 195/480 [26:21<34:57,  7.36s/it]
 41%|████      | 196/480 [26:29<35:51,  7.58s/it]
 41%|████      | 197/480 [26:38<36:44,  7.79s/it]
 41%|████▏     | 198/480 [26:45<35:50,  7.63s/it]
 41%|████▏     | 199/480 [26:53<36:18,  7.75s/it]
 42%|████▏     | 200/480 [27:01<36:52,  7.90s/it]
                                                 
{'loss': '0.2757', 'grad_norm': '0.2616', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [27:01<36:52,  7.90s/it]
 42%|████▏     | 201/480 [27:09<37:23,  8.04s/it]
 42%|████▏     | 202/480 [27:17<37:11,  8.03s/it]
 42%|████▏     | 203/480 [27:25<36:39,  7.94s/it]
 42%|████▎     | 204/480 [27:33<36:38,  7.96s/it]
 43%|████▎     | 205/480 [27:42<37:07,  8.10s/it]
 43%|████▎     | 206/480 [27:50<36:43,  8.04s/it]
 43%|████▎     | 207/480 [27:58<37:18,  8.20s/it]
 43%|████▎     | 208/480 [28:06<36:48,  8.12s/it]
 44%|████▎     | 209/480 [28:14<36:19,  8.04s/it]
 44%|████▍     | 210/480 [28:22<35:52,  7.97s/it]
                                                 
{'loss': '0.2088', 'grad_norm': '0.3423', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [28:22<35:52,  7.97s/it]
 44%|████▍     | 211/480 [28:30<35:42,  7.96s/it]
 44%|████▍     | 212/480 [28:38<35:45,  8.01s/it]
 44%|████▍     | 213/480 [28:46<35:55,  8.07s/it]
 45%|████▍     | 214/480 [28:54<35:20,  7.97s/it]
 45%|████▍     | 215/480 [29:02<35:41,  8.08s/it]
 45%|████▌     | 216/480 [29:10<35:52,  8.15s/it]
 45%|████▌     | 217/480 [29:19<35:59,  8.21s/it]
 45%|████▌     | 218/480 [29:27<35:24,  8.11s/it]
 46%|████▌     | 219/480 [29:35<35:34,  8.18s/it]
 46%|████▌     | 220/480 [29:43<35:13,  8.13s/it]
                                                 
{'loss': '0.2938', 'grad_norm': '0.06912', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [29:43<35:13,  8.13s/it]
 46%|████▌     | 221/480 [29:52<35:54,  8.32s/it]
 46%|████▋     | 222/480 [30:00<35:21,  8.22s/it]
 46%|████▋     | 223/480 [30:07<34:30,  8.06s/it]
 47%|████▋     | 224/480 [30:09<26:42,  6.26s/it]
 47%|████▋     | 225/480 [30:17<28:42,  6.75s/it]
 47%|████▋     | 226/480 [30:26<30:35,  7.23s/it]
 47%|████▋     | 227/480 [30:34<31:48,  7.54s/it]
 48%|████▊     | 228/480 [30:42<32:22,  7.71s/it]
 48%|████▊     | 229/480 [30:51<33:21,  7.97s/it]
 48%|████▊     | 230/480 [30:58<32:56,  7.91s/it]
                                                 
{'loss': '0.2687', 'grad_norm': '0.1468', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [30:58<32:56,  7.91s/it]
 48%|████▊     | 231/480 [31:06<32:58,  7.94s/it]
 48%|████▊     | 232/480 [31:14<32:55,  7.97s/it]
 49%|████▊     | 233/480 [31:22<32:35,  7.92s/it]
 49%|████▉     | 234/480 [31:30<32:34,  7.95s/it]
 49%|████▉     | 235/480 [31:39<33:02,  8.09s/it]
 49%|████▉     | 236/480 [31:47<32:56,  8.10s/it]
 49%|████▉     | 237/480 [31:55<32:40,  8.07s/it]
 50%|████▉     | 238/480 [32:03<32:29,  8.06s/it]
 50%|████▉     | 239/480 [32:11<32:39,  8.13s/it]
 50%|█████     | 240/480 [32:20<33:24,  8.35s/it]
                                                 
{'loss': '0.3114', 'grad_norm': '1.912', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [32:20<33:24,  8.35s/it]
 50%|█████     | 241/480 [32:28<33:00,  8.29s/it]
 50%|█████     | 242/480 [32:36<32:12,  8.12s/it]
 51%|█████     | 243/480 [32:44<32:28,  8.22s/it]
 51%|█████     | 244/480 [32:52<32:11,  8.19s/it]
 51%|█████     | 245/480 [33:01<32:10,  8.22s/it]
 51%|█████▏    | 246/480 [33:09<31:54,  8.18s/it]
 51%|█████▏    | 247/480 [33:17<31:11,  8.03s/it]
 52%|█████▏    | 248/480 [33:24<30:48,  7.97s/it]
 52%|█████▏    | 249/480 [33:33<30:56,  8.04s/it]
 52%|█████▏    | 250/480 [33:41<30:56,  8.07s/it]
                                                 
{'loss': '0.1836', 'grad_norm': '0.4687', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [33:41<30:56,  8.07s/it]
 52%|█████▏    | 251/480 [33:48<30:12,  7.92s/it]
 52%|█████▎    | 252/480 [33:57<30:34,  8.04s/it]
 53%|█████▎    | 253/480 [34:04<30:12,  7.99s/it]
 53%|█████▎    | 254/480 [34:12<30:02,  7.97s/it]
 53%|█████▎    | 255/480 [34:20<29:52,  7.97s/it]
 53%|█████▎    | 256/480 [34:22<22:54,  6.14s/it]
 54%|█████▎    | 257/480 [34:30<24:45,  6.66s/it]
 54%|█████▍    | 258/480 [34:38<26:03,  7.04s/it]
 54%|█████▍    | 259/480 [34:46<27:13,  7.39s/it]
 54%|█████▍    | 260/480 [34:54<27:39,  7.54s/it]
                                                 
{'loss': '0.2038', 'grad_norm': '2.082', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [34:54<27:39,  7.54s/it]
 54%|█████▍    | 261/480 [35:02<27:46,  7.61s/it]
 55%|█████▍    | 262/480 [35:10<28:32,  7.85s/it]
 55%|█████▍    | 263/480 [35:18<28:16,  7.82s/it]
 55%|█████▌    | 264/480 [35:26<28:45,  7.99s/it]
 55%|█████▌    | 265/480 [35:34<28:34,  7.97s/it]
 55%|█████▌    | 266/480 [35:42<28:13,  7.91s/it]
 56%|█████▌    | 267/480 [35:50<27:31,  7.76s/it]
 56%|█████▌    | 268/480 [35:58<28:20,  8.02s/it]
 56%|█████▌    | 269/480 [36:06<28:22,  8.07s/it]
 56%|█████▋    | 270/480 [36:15<28:19,  8.09s/it]
                                                 
{'loss': '0.2262', 'grad_norm': '0.05799', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [36:15<28:19,  8.09s/it]
 56%|█████▋    | 271/480 [36:23<28:26,  8.17s/it]
 57%|█████▋    | 272/480 [36:31<27:53,  8.05s/it]
 57%|█████▋    | 273/480 [36:39<28:03,  8.13s/it]
 57%|█████▋    | 274/480 [36:47<28:17,  8.24s/it]
 57%|█████▋    | 275/480 [36:56<28:10,  8.25s/it]
 57%|█████▊    | 276/480 [37:04<27:57,  8.22s/it]
 58%|█████▊    | 277/480 [37:12<27:43,  8.20s/it]
 58%|█████▊    | 278/480 [37:21<28:16,  8.40s/it]
 58%|█████▊    | 279/480 [37:29<27:49,  8.31s/it]
 58%|█████▊    | 280/480 [37:37<27:27,  8.24s/it]
                                                 
{'loss': '0.2633', 'grad_norm': '0.0215', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [37:37<27:27,  8.24s/it]
 59%|█████▊    | 281/480 [37:45<27:10,  8.20s/it]
 59%|█████▉    | 282/480 [37:53<26:22,  7.99s/it]
 59%|█████▉    | 283/480 [38:00<25:31,  7.78s/it]
 59%|█████▉    | 284/480 [38:08<25:46,  7.89s/it]
 59%|█████▉    | 285/480 [38:16<25:47,  7.94s/it]
 60%|█████▉    | 286/480 [38:25<26:25,  8.17s/it]
 60%|█████▉    | 287/480 [38:33<26:13,  8.15s/it]
 60%|██████    | 288/480 [38:35<19:56,  6.23s/it]
 60%|██████    | 289/480 [38:43<21:50,  6.86s/it]
 60%|██████    | 290/480 [38:51<22:34,  7.13s/it]
                                                 
{'loss': '0.1731', 'grad_norm': '0.1703', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [38:51<22:34,  7.13s/it]
 61%|██████    | 291/480 [38:59<23:35,  7.49s/it]
 61%|██████    | 292/480 [39:07<24:09,  7.71s/it]
 61%|██████    | 293/480 [39:16<24:34,  7.89s/it]
 61%|██████▏   | 294/480 [39:24<24:25,  7.88s/it]
 61%|██████▏   | 295/480 [39:32<24:27,  7.93s/it]
 62%|██████▏   | 296/480 [39:40<24:36,  8.03s/it]
 62%|██████▏   | 297/480 [39:47<23:58,  7.86s/it]
 62%|██████▏   | 298/480 [39:55<23:53,  7.88s/it]
 62%|██████▏   | 299/480 [40:03<23:55,  7.93s/it]
 62%|██████▎   | 300/480 [40:12<24:11,  8.07s/it]
                                                 
{'loss': '0.1868', 'grad_norm': '0.8308', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [40:12<24:11,  8.07s/it]
 63%|██████▎   | 301/480 [40:19<23:50,  7.99s/it]
 63%|██████▎   | 302/480 [40:27<23:27,  7.91s/it]
 63%|██████▎   | 303/480 [40:36<23:46,  8.06s/it]
 63%|██████▎   | 304/480 [40:44<23:38,  8.06s/it]
 64%|██████▎   | 305/480 [40:52<23:23,  8.02s/it]
 64%|██████▍   | 306/480 [41:00<23:18,  8.04s/it]
 64%|██████▍   | 307/480 [41:08<23:16,  8.07s/it]
 64%|██████▍   | 308/480 [41:16<23:17,  8.13s/it]
 64%|██████▍   | 309/480 [41:25<23:35,  8.28s/it]
 65%|██████▍   | 310/480 [41:33<23:33,  8.32s/it]
                                                 
{'loss': '0.2794', 'grad_norm': '2.653', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [41:33<23:33,  8.32s/it]
 65%|██████▍   | 311/480 [41:41<23:00,  8.17s/it]
 65%|██████▌   | 312/480 [41:49<22:52,  8.17s/it]
 65%|██████▌   | 313/480 [41:57<22:33,  8.10s/it]
 65%|██████▌   | 314/480 [42:05<22:12,  8.03s/it]
 66%|██████▌   | 315/480 [42:13<22:16,  8.10s/it]
 66%|██████▌   | 316/480 [42:21<21:44,  7.95s/it]
 66%|██████▌   | 317/480 [42:29<21:53,  8.06s/it]
 66%|██████▋   | 318/480 [42:37<21:37,  8.01s/it]
 66%|██████▋   | 319/480 [42:45<21:46,  8.12s/it]
 67%|██████▋   | 320/480 [42:47<16:35,  6.22s/it]
                                                 
{'loss': '0.2114', 'grad_norm': '0.2609', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [42:47<16:35,  6.22s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:00<01:07,  1.47it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:44,  2.21it/s]
+
Eval @ step 320:   3%|▎         | 3/100 [00:01<00:36,  2.65it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:01<00:34,  2.80it/s]
+
Eval @ step 320:   5%|▌         | 5/100 [00:02<00:37,  2.54it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:02<00:34,  2.70it/s]
+
Eval @ step 320:   7%|▋         | 7/100 [00:02<00:33,  2.79it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:03<00:34,  2.68it/s]
+
Eval @ step 320:   9%|▉         | 9/100 [00:03<00:36,  2.46it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:03<00:34,  2.59it/s]
+
Eval @ step 320:  11%|█         | 11/100 [00:04<00:33,  2.65it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:04<00:31,  2.77it/s]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:05<00:37,  2.33it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:05<00:43,  1.98it/s]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:06<00:41,  2.03it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:06<00:38,  2.16it/s]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:06<00:32,  2.52it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:07<00:35,  2.31it/s]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:07<00:35,  2.30it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:08<00:32,  2.44it/s]
+
Eval @ step 320:  21%|██        | 21/100 [00:08<00:28,  2.77it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:08<00:29,  2.63it/s]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:09<00:29,  2.58it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:09<00:29,  2.54it/s]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:10<00:28,  2.65it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:10<00:30,  2.41it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:11<00:29,  2.43it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:11<00:27,  2.60it/s]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:11<00:22,  3.13it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:11<00:25,  2.75it/s]
+
Eval @ step 320:  31%|███       | 31/100 [00:12<00:23,  2.94it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:12<00:20,  3.24it/s]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:13<00:25,  2.67it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:13<00:22,  2.90it/s]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:13<00:20,  3.18it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:14<00:23,  2.75it/s]
+
Eval @ step 320:  37%|███▋      | 37/100 [00:14<00:19,  3.26it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:14<00:22,  2.77it/s]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:14<00:19,  3.10it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:15<00:28,  2.13it/s]
+
Eval @ step 320:  41%|████      | 41/100 [00:15<00:22,  2.63it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:16<00:23,  2.44it/s]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:16<00:23,  2.38it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:17<00:23,  2.37it/s]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:17<00:20,  2.74it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:17<00:17,  3.05it/s]
+
Eval @ step 320:  47%|████▋     | 47/100 [00:18<00:17,  3.05it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:18<00:16,  3.17it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:18<00:17,  2.95it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:19<00:18,  2.63it/s]
+
Eval @ step 320:  51%|█████     | 51/100 [00:19<00:16,  2.96it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:19<00:18,  2.63it/s]
+
Eval @ step 320:  53%|█████▎    | 53/100 [00:20<00:18,  2.52it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:20<00:15,  3.04it/s]
+
Eval @ step 320:  55%|█████▌    | 55/100 [00:20<00:13,  3.30it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:21<00:20,  2.10it/s]
+
Eval @ step 320:  57%|█████▋    | 57/100 [00:22<00:19,  2.25it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:22<00:17,  2.35it/s]
+
Eval @ step 320:  59%|█████▉    | 59/100 [00:22<00:15,  2.73it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:23<00:15,  2.64it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:23<00:16,  2.42it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:23<00:15,  2.43it/s]
+
Eval @ step 320:  63%|██████▎   | 63/100 [00:24<00:14,  2.56it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:24<00:16,  2.18it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [00:25<00:17,  2.04it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:25<00:14,  2.34it/s]
+
Eval @ step 320:  67%|██████▋   | 67/100 [00:25<00:11,  2.85it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:26<00:10,  2.93it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [00:26<00:10,  3.08it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:27<00:12,  2.37it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [00:27<00:12,  2.28it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:27<00:10,  2.66it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [00:28<00:09,  2.99it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:28<00:08,  3.08it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [00:28<00:08,  3.01it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:29<00:07,  3.16it/s]
+
Eval @ step 320:  77%|███████▋  | 77/100 [00:29<00:09,  2.48it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:30<00:08,  2.49it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [00:30<00:07,  2.83it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:30<00:07,  2.72it/s]
+
Eval @ step 320:  81%|████████  | 81/100 [00:31<00:07,  2.43it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:31<00:06,  2.77it/s]
+
Eval @ step 320:  83%|████████▎ | 83/100 [00:31<00:06,  2.60it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:32<00:08,  1.98it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [00:32<00:06,  2.45it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:33<00:04,  2.82it/s]
+
Eval @ step 320:  87%|████████▋ | 87/100 [00:33<00:04,  2.83it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:34<00:05,  2.33it/s]
+
Eval @ step 320:  89%|████████▉ | 89/100 [00:34<00:04,  2.38it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:35<00:05,  1.99it/s]
+
Eval @ step 320:  91%|█████████ | 91/100 [00:35<00:04,  1.80it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:36<00:03,  2.05it/s]
+
Eval @ step 320:  93%|█████████▎| 93/100 [00:36<00:03,  2.02it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:37<00:03,  1.93it/s]
+
Eval @ step 320:  95%|█████████▌| 95/100 [00:37<00:02,  2.07it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:37<00:01,  2.28it/s]
+
Eval @ step 320:  97%|█████████▋| 97/100 [00:38<00:01,  1.93it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:39<00:01,  1.61it/s]
+
Eval @ step 320:  99%|█████████▉| 99/100 [00:39<00:00,  1.89it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:40<00:00,  2.25it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=73.7143 (step=320) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '73.71', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [43:27<16:35,  6.22s/it]
 67%|██████▋   | 321/480 [43:36<50:02, 18.88s/it]
 67%|██████▋   | 322/480 [43:44<41:07, 15.61s/it]
 67%|██████▋   | 323/480 [43:52<35:03, 13.40s/it]
 68%|██████▊   | 324/480 [44:00<30:37, 11.78s/it]
 68%|██████▊   | 325/480 [44:08<27:18, 10.57s/it]
 68%|██████▊   | 326/480 [44:16<25:17,  9.86s/it]
 68%|██████▊   | 327/480 [44:24<23:50,  9.35s/it]
 68%|██████▊   | 328/480 [44:32<22:47,  9.00s/it]
 69%|██████▊   | 329/480 [44:40<22:11,  8.82s/it]
 69%|██████▉   | 330/480 [44:49<21:36,  8.64s/it]
                                                 
{'loss': '0.2392', 'grad_norm': '0.421', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [44:49<21:36,  8.64s/it]
 69%|██████▉   | 331/480 [44:57<21:13,  8.55s/it]
 69%|██████▉   | 332/480 [45:05<20:44,  8.41s/it]
 69%|██████▉   | 333/480 [45:14<20:35,  8.41s/it]
 70%|██████▉   | 334/480 [45:21<20:02,  8.23s/it]
 70%|██████▉   | 335/480 [45:29<19:43,  8.16s/it]
 70%|███████   | 336/480 [45:38<19:42,  8.21s/it]
 70%|███████   | 337/480 [45:46<19:37,  8.23s/it]
 70%|███████   | 338/480 [45:54<19:15,  8.14s/it]
 71%|███████   | 339/480 [46:02<19:17,  8.21s/it]
 71%|███████   | 340/480 [46:10<19:00,  8.15s/it]
                                                 
{'loss': '0.2109', 'grad_norm': '0.06233', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [46:10<19:00,  8.15s/it]
 71%|███████   | 341/480 [46:19<18:57,  8.18s/it]
 71%|███████▏  | 342/480 [46:26<18:37,  8.10s/it]
 71%|███████▏  | 343/480 [46:34<18:14,  7.99s/it]
 72%|███████▏  | 344/480 [46:42<18:01,  7.95s/it]
 72%|███████▏  | 345/480 [46:50<17:52,  7.94s/it]
 72%|███████▏  | 346/480 [46:58<17:49,  7.98s/it]
 72%|███████▏  | 347/480 [47:06<17:45,  8.01s/it]
 72%|███████▎  | 348/480 [47:15<18:00,  8.18s/it]
 73%|███████▎  | 349/480 [47:23<17:41,  8.11s/it]
 73%|███████▎  | 350/480 [47:30<17:06,  7.90s/it]
                                                 
{'loss': '0.1869', 'grad_norm': '0.2626', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [47:30<17:06,  7.90s/it]
 73%|███████▎  | 351/480 [47:38<17:14,  8.02s/it]
 73%|███████▎  | 352/480 [47:40<13:08,  6.16s/it]
 74%|███████▎  | 353/480 [47:48<13:58,  6.61s/it]
 74%|███████▍  | 354/480 [47:56<14:41,  6.99s/it]
 74%|███████▍  | 355/480 [48:04<15:29,  7.44s/it]
 74%|███████▍  | 356/480 [48:12<15:43,  7.61s/it]
 74%|███████▍  | 357/480 [48:20<15:37,  7.63s/it]
 75%|███████▍  | 358/480 [48:28<15:55,  7.83s/it]
 75%|███████▍  | 359/480 [48:36<16:00,  7.94s/it]
 75%|███████▌  | 360/480 [48:44<15:44,  7.87s/it]
                                                 
{'loss': '0.2518', 'grad_norm': '0.7815', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [48:44<15:44,  7.87s/it]
 75%|███████▌  | 361/480 [48:52<15:40,  7.91s/it]
 75%|███████▌  | 362/480 [49:00<15:32,  7.90s/it]
 76%|███████▌  | 363/480 [49:08<15:23,  7.90s/it]
 76%|███████▌  | 364/480 [49:16<15:36,  8.08s/it]
 76%|███████▌  | 365/480 [49:24<15:27,  8.06s/it]
 76%|███████▋  | 366/480 [49:32<15:15,  8.03s/it]
 76%|███████▋  | 367/480 [49:41<15:14,  8.10s/it]
 77%|███████▋  | 368/480 [49:48<14:52,  7.97s/it]
 77%|███████▋  | 369/480 [49:57<14:56,  8.08s/it]
 77%|███████▋  | 370/480 [50:05<14:50,  8.09s/it]
                                                 
{'loss': '0.1395', 'grad_norm': '0.1379', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [50:05<14:50,  8.09s/it]
 77%|███████▋  | 371/480 [50:12<14:33,  8.01s/it]
 78%|███████▊  | 372/480 [50:20<14:22,  7.99s/it]
 78%|███████▊  | 373/480 [50:29<14:30,  8.13s/it]
 78%|███████▊  | 374/480 [50:37<14:32,  8.23s/it]
 78%|███████▊  | 375/480 [50:45<14:19,  8.19s/it]
 78%|███████▊  | 376/480 [50:54<14:16,  8.23s/it]
 79%|███████▊  | 377/480 [51:02<14:08,  8.24s/it]
 79%|███████▉  | 378/480 [51:10<13:47,  8.12s/it]
 79%|███████▉  | 379/480 [51:18<13:41,  8.13s/it]
 79%|███████▉  | 380/480 [51:26<13:15,  7.95s/it]
                                                 
{'loss': '0.2367', 'grad_norm': '0.1339', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [51:26<13:15,  7.95s/it]
 79%|███████▉  | 381/480 [51:34<13:18,  8.06s/it]
 80%|███████▉  | 382/480 [51:42<13:10,  8.07s/it]
 80%|███████▉  | 383/480 [51:50<13:05,  8.10s/it]
 80%|████████  | 384/480 [51:52<10:08,  6.34s/it]
 80%|████████  | 385/480 [52:01<10:58,  6.93s/it]
 80%|████████  | 386/480 [52:09<11:18,  7.22s/it]
 81%|████████  | 387/480 [52:16<11:28,  7.40s/it]
 81%|████████  | 388/480 [52:24<11:39,  7.61s/it]
 81%|████████  | 389/480 [52:32<11:38,  7.68s/it]
 81%|████████▏ | 390/480 [52:41<11:57,  7.98s/it]
                                                 
{'loss': '0.2317', 'grad_norm': '1.047', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [52:41<11:57,  7.98s/it]
 81%|████████▏ | 391/480 [52:49<11:59,  8.09s/it]
 82%|████████▏ | 392/480 [52:57<11:52,  8.09s/it]
 82%|████████▏ | 393/480 [53:06<11:45,  8.10s/it]
 82%|████████▏ | 394/480 [53:13<11:27,  7.99s/it]
 82%|████████▏ | 395/480 [53:21<11:01,  7.79s/it]
 82%|████████▎ | 396/480 [53:29<10:57,  7.83s/it]
 83%|████████▎ | 397/480 [53:36<10:49,  7.83s/it]
 83%|████████▎ | 398/480 [53:44<10:49,  7.91s/it]
 83%|████████▎ | 399/480 [53:53<10:51,  8.05s/it]
 83%|████████▎ | 400/480 [54:01<10:50,  8.14s/it]
                                                 
{'loss': '0.1833', 'grad_norm': '0.9034', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [54:01<10:50,  8.14s/it]
 84%|████████▎ | 401/480 [54:09<10:38,  8.08s/it]
 84%|████████▍ | 402/480 [54:18<10:39,  8.20s/it]
 84%|████████▍ | 403/480 [54:26<10:30,  8.19s/it]
 84%|████████▍ | 404/480 [54:34<10:26,  8.25s/it]
 84%|████████▍ | 405/480 [54:42<10:07,  8.10s/it]
 85%|████████▍ | 406/480 [54:51<10:14,  8.30s/it]
 85%|████████▍ | 407/480 [54:59<10:00,  8.22s/it]
 85%|████████▌ | 408/480 [55:07<09:53,  8.25s/it]
 85%|████████▌ | 409/480 [55:15<09:41,  8.19s/it]
 85%|████████▌ | 410/480 [55:23<09:23,  8.06s/it]
                                                 
{'loss': '0.1455', 'grad_norm': '0.6927', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [55:23<09:23,  8.06s/it]
 86%|████████▌ | 411/480 [55:30<09:06,  7.92s/it]
 86%|████████▌ | 412/480 [55:39<09:07,  8.05s/it]
 86%|████████▌ | 413/480 [55:47<08:53,  7.96s/it]
 86%|████████▋ | 414/480 [55:55<08:48,  8.01s/it]
 86%|████████▋ | 415/480 [56:03<08:41,  8.02s/it]
 87%|████████▋ | 416/480 [56:05<06:39,  6.24s/it]
 87%|████████▋ | 417/480 [56:13<07:13,  6.88s/it]
 87%|████████▋ | 418/480 [56:21<07:26,  7.20s/it]
 87%|████████▋ | 419/480 [56:30<07:42,  7.58s/it]
 88%|████████▊ | 420/480 [56:38<07:42,  7.71s/it]
                                                 
{'loss': '0.2028', 'grad_norm': '0.7214', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [56:38<07:42,  7.71s/it]
 88%|████████▊ | 421/480 [56:46<07:39,  7.78s/it]
 88%|████████▊ | 422/480 [56:54<07:42,  7.97s/it]
 88%|████████▊ | 423/480 [57:02<07:38,  8.04s/it]
 88%|████████▊ | 424/480 [57:11<07:36,  8.16s/it]
 89%|████████▊ | 425/480 [57:18<07:23,  8.06s/it]
 89%|████████▉ | 426/480 [57:26<07:06,  7.90s/it]
 89%|████████▉ | 427/480 [57:34<06:59,  7.92s/it]
 89%|████████▉ | 428/480 [57:42<06:58,  8.04s/it]
 89%|████████▉ | 429/480 [57:50<06:50,  8.05s/it]
 90%|████████▉ | 430/480 [57:58<06:40,  8.02s/it]
                                                 
{'loss': '0.1784', 'grad_norm': '0.1253', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [57:58<06:40,  8.02s/it]
 90%|████████▉ | 431/480 [58:07<06:37,  8.10s/it]
 90%|█████████ | 432/480 [58:15<06:27,  8.08s/it]
 90%|█████████ | 433/480 [58:23<06:18,  8.06s/it]
 90%|█████████ | 434/480 [58:30<06:06,  7.97s/it]
 91%|█████████ | 435/480 [58:39<06:04,  8.11s/it]
 91%|█████████ | 436/480 [58:47<05:57,  8.12s/it]
 91%|█████████ | 437/480 [58:54<05:40,  7.92s/it]
 91%|█████████▏| 438/480 [59:03<05:36,  8.01s/it]
 91%|█████████▏| 439/480 [59:11<05:31,  8.10s/it]
 92%|█████████▏| 440/480 [59:19<05:23,  8.08s/it]
                                                 
{'loss': '0.1392', 'grad_norm': '0.07952', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [59:19<05:23,  8.08s/it]
 92%|█████████▏| 441/480 [59:27<05:14,  8.06s/it]
 92%|█████████▏| 442/480 [59:35<05:03,  7.98s/it]
 92%|█████████▏| 443/480 [59:43<04:56,  8.03s/it]
 92%|█████████▎| 444/480 [59:51<04:47,  8.00s/it]
 93%|█████████▎| 445/480 [59:59<04:39,  7.97s/it]
 93%|█████████▎| 446/480 [1:00:07<04:35,  8.11s/it]
 93%|█████████▎| 447/480 [1:00:15<04:26,  8.06s/it]
 93%|█████████▎| 448/480 [1:00:17<03:21,  6.29s/it]
 94%|█████████▎| 449/480 [1:00:26<03:35,  6.94s/it]
 94%|█████████▍| 450/480 [1:00:34<03:36,  7.21s/it]
                                                   
{'loss': '0.3236', 'grad_norm': '0.1843', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [1:00:34<03:36,  7.21s/it]
 94%|█████████▍| 451/480 [1:00:42<03:36,  7.47s/it]
 94%|█████████▍| 452/480 [1:00:50<03:33,  7.64s/it]
 94%|█████████▍| 453/480 [1:00:58<03:28,  7.73s/it]
 95%|█████████▍| 454/480 [1:01:06<03:29,  8.07s/it]
 95%|█████████▍| 455/480 [1:01:14<03:20,  8.01s/it]
 95%|█████████▌| 456/480 [1:01:23<03:14,  8.10s/it]
 95%|█████████▌| 457/480 [1:01:31<03:07,  8.14s/it]
 95%|█████████▌| 458/480 [1:01:39<03:00,  8.21s/it]
 96%|█████████▌| 459/480 [1:01:48<02:54,  8.31s/it]
 96%|█████████▌| 460/480 [1:01:55<02:41,  8.09s/it]
                                                   
{'loss': '0.1554', 'grad_norm': '0.01593', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [1:01:55<02:41,  8.09s/it]
 96%|█████████▌| 461/480 [1:02:04<02:35,  8.16s/it]
 96%|█████████▋| 462/480 [1:02:12<02:27,  8.19s/it]
 96%|█████████▋| 463/480 [1:02:20<02:18,  8.16s/it]
 97%|█████████▋| 464/480 [1:02:28<02:10,  8.15s/it]
 97%|█████████▋| 465/480 [1:02:36<02:02,  8.14s/it]
 97%|█████████▋| 466/480 [1:02:45<01:54,  8.16s/it]
 97%|█████████▋| 467/480 [1:02:53<01:46,  8.19s/it]
 98%|█████████▊| 468/480 [1:03:01<01:37,  8.15s/it]
 98%|█████████▊| 469/480 [1:03:08<01:26,  7.91s/it]
 98%|█████████▊| 470/480 [1:03:16<01:19,  7.92s/it]
                                                   
{'loss': '0.1879', 'grad_norm': '0.09104', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [1:03:16<01:19,  7.92s/it]
 98%|█████████▊| 471/480 [1:03:24<01:11,  7.91s/it]
 98%|█████████▊| 472/480 [1:03:32<01:03,  7.99s/it]
 99%|█████████▊| 473/480 [1:03:40<00:55,  7.98s/it]
 99%|█████████▉| 474/480 [1:03:48<00:47,  7.95s/it]
 99%|█████████▉| 475/480 [1:03:56<00:40,  8.06s/it]
 99%|█████████▉| 476/480 [1:04:04<00:31,  7.95s/it]
 99%|█████████▉| 477/480 [1:04:12<00:23,  7.91s/it]
100%|█████████▉| 478/480 [1:04:20<00:16,  8.03s/it]
100%|█████████▉| 479/480 [1:04:28<00:08,  8.05s/it]
100%|██████████| 480/480 [1:04:30<00:00,  6.27s/it]
                                                   
{'loss': '0.1392', 'grad_norm': '2.909', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [1:04:30<00:00,  6.27s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:01<02:34,  1.56s/it]
+
Eval @ step 480:   2%|▏         | 2/100 [00:01<01:19,  1.23it/s]
+
Eval @ step 480:   3%|▎         | 3/100 [00:02<00:55,  1.74it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:02<00:45,  2.10it/s]
+
Eval @ step 480:   5%|▌         | 5/100 [00:02<00:44,  2.12it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:03<00:39,  2.37it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:03<00:36,  2.55it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:03<00:36,  2.52it/s]
+
Eval @ step 480:   9%|▉         | 9/100 [00:04<00:38,  2.37it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:04<00:35,  2.51it/s]
+
Eval @ step 480:  11%|█         | 11/100 [00:05<00:34,  2.60it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:05<00:30,  2.93it/s]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:05<00:36,  2.41it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:06<00:42,  2.01it/s]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:07<00:41,  2.05it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:07<00:38,  2.17it/s]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:07<00:32,  2.54it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:08<00:35,  2.32it/s]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:08<00:35,  2.31it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:09<00:32,  2.44it/s]
+
Eval @ step 480:  21%|██        | 21/100 [00:09<00:28,  2.78it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:09<00:29,  2.64it/s]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:10<00:39,  1.96it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:10<00:36,  2.08it/s]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:11<00:31,  2.38it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:11<00:33,  2.24it/s]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:12<00:31,  2.31it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:12<00:28,  2.50it/s]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:12<00:23,  3.03it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:13<00:28,  2.47it/s]
+
Eval @ step 480:  31%|███       | 31/100 [00:13<00:25,  2.71it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:13<00:22,  3.04it/s]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:14<00:26,  2.57it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:14<00:23,  2.81it/s]
+
Eval @ step 480:  35%|███▌      | 35/100 [00:14<00:22,  2.89it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:15<00:24,  2.59it/s]
+
Eval @ step 480:  37%|███▋      | 37/100 [00:15<00:20,  3.11it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:16<00:23,  2.69it/s]
+
Eval @ step 480:  39%|███▉      | 39/100 [00:16<00:20,  3.05it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:16<00:25,  2.36it/s]
+
Eval @ step 480:  41%|████      | 41/100 [00:17<00:20,  2.87it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:17<00:22,  2.58it/s]
+
Eval @ step 480:  43%|████▎     | 43/100 [00:18<00:23,  2.48it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:18<00:23,  2.43it/s]
+
Eval @ step 480:  45%|████▌     | 45/100 [00:18<00:19,  2.80it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:18<00:17,  3.10it/s]
+
Eval @ step 480:  47%|████▋     | 47/100 [00:19<00:17,  3.09it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:19<00:15,  3.38it/s]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:19<00:16,  3.07it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:20<00:18,  2.70it/s]
+
Eval @ step 480:  51%|█████     | 51/100 [00:20<00:16,  3.01it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:21<00:18,  2.66it/s]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:21<00:18,  2.54it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:21<00:15,  3.05it/s]
+
Eval @ step 480:  55%|█████▌    | 55/100 [00:21<00:13,  3.31it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:22<00:20,  2.11it/s]
+
Eval @ step 480:  57%|█████▋    | 57/100 [00:23<00:19,  2.25it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:23<00:17,  2.35it/s]
+
Eval @ step 480:  59%|█████▉    | 59/100 [00:23<00:15,  2.73it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:24<00:15,  2.66it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [00:24<00:16,  2.42it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [00:25<00:15,  2.43it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [00:25<00:14,  2.56it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:26<00:18,  1.98it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [00:26<00:16,  2.11it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:26<00:14,  2.31it/s]
+
Eval @ step 480:  67%|██████▋   | 67/100 [00:27<00:11,  2.82it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:27<00:11,  2.91it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [00:27<00:10,  3.06it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:28<00:12,  2.37it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [00:28<00:12,  2.28it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:29<00:10,  2.66it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [00:29<00:10,  2.61it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:29<00:09,  2.79it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [00:30<00:08,  2.81it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:30<00:08,  2.72it/s]
+
Eval @ step 480:  77%|███████▋  | 77/100 [00:31<00:10,  2.14it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:31<00:09,  2.24it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [00:31<00:08,  2.60it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:32<00:07,  2.56it/s]
+
Eval @ step 480:  81%|████████  | 81/100 [00:32<00:07,  2.49it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:32<00:06,  2.83it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [00:33<00:06,  2.64it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:34<00:08,  1.99it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [00:34<00:06,  2.46it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:34<00:04,  2.83it/s]
+
Eval @ step 480:  87%|████████▋ | 87/100 [00:34<00:04,  2.84it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:35<00:05,  2.33it/s]
+
Eval @ step 480:  89%|████████▉ | 89/100 [00:35<00:04,  2.38it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:36<00:05,  1.99it/s]
+
Eval @ step 480:  91%|█████████ | 91/100 [00:37<00:04,  1.80it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:37<00:03,  2.06it/s]
+
Eval @ step 480:  93%|█████████▎| 93/100 [00:38<00:03,  2.16it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:38<00:02,  2.02it/s]
+
Eval @ step 480:  95%|█████████▌| 95/100 [00:38<00:02,  2.14it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:39<00:01,  2.33it/s]
+
Eval @ step 480:  97%|█████████▋| 97/100 [00:40<00:01,  1.96it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:40<00:00,  2.12it/s]
+
Eval @ step 480:  99%|█████████▉| 99/100 [00:40<00:00,  2.35it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:40<00:00,  2.70it/s]
+
                                                                    [EVAL] eval_rougeL=72.7508 (best=73.7143)
+
                                                   
{'eval_rougeL': '72.75', 'epoch': '15'}
+
100%|██████████| 480/480 [1:05:11<00:00,  6.27s/it]
                                                   
{'train_runtime': '3912', 'train_samples_per_second': '3.834', 'train_steps_per_second': '0.123', 'train_loss': '0.3357', 'epoch': '15'}
+
100%|██████████| 480/480 [1:05:11<00:00,  6.27s/it]
100%|██████████| 480/480 [1:05:11<00:00,  8.15s/it]
+***** train metrics *****
+  epoch                    =        15.0
+  total_flos               = 342749931GF
+  train_loss               =      0.3357
+  train_runtime            =  1:05:11.88
+  train_samples            =        1000
+  train_samples_per_second =       3.834
+  train_steps_per_second   =       0.123
+  [EVAL] Restored best model (eval_rougeL=73.7143, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   2%|▏         | 6/291 [00:00<00:05, 50.55it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 152.27it/s]
Loading weights:  17%|█▋        | 50/291 [00:00<00:01, 156.03it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 162.64it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 167.00it/s]
Loading weights:  36%|███▌      | 104/291 [00:00<00:01, 167.04it/s]
Loading weights:  42%|████▏     | 122/291 [00:00<00:01, 165.61it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 160.81it/s]
Loading weights:  54%|█████▍    | 157/291 [00:00<00:00, 161.98it/s]
Loading weights:  60%|█████▉    | 174/291 [00:01<00:00, 157.94it/s]
Loading weights:  66%|██████▌   | 191/291 [00:01<00:00, 158.87it/s]
Loading weights:  71%|███████▏  | 208/291 [00:01<00:00, 158.66it/s]
Loading weights:  77%|███████▋  | 224/291 [00:01<00:00, 156.58it/s]
Loading weights:  82%|████████▏ | 240/291 [00:01<00:00, 155.76it/s]
Loading weights:  88%|████████▊ | 256/291 [00:01<00:00, 152.96it/s]
Loading weights:  94%|█████████▍| 274/291 [00:01<00:00, 153.85it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 160.16it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task002_quoref_answer_generation': PaR=77.3, shrinkage=ridge, n=1000
+  [SRT] Wired router: 5 sigs, cur=task002_quoref_answer_generation, mapping={'task002_quoref_answer_generation': 0, 'task1572_samsum_summary': 4, 'task363_sst2_polarity_classification': 3, 'task1290_xsum_summarization': 2, 'task181_outcome_extraction': 1}
+[SRT] Router: 5 tasks, avg_PaR=77.3, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation']
+  [SRT-INIT] task_id_to_idx={'task002_quoref_answer_generation': 0, 'task1572_samsum_summary': 4, 'task363_sst2_polarity_classification': 3, 'task1290_xsum_summarization': 2, 'task181_outcome_extraction': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/363 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=5
+
Continual predict:   0%|          | 1/363 [00:01<10:10,  1.69s/it]
Continual predict:   1%|          | 2/363 [00:04<12:31,  2.08s/it]
Continual predict:   1%|          | 3/363 [00:04<08:48,  1.47s/it]
Continual predict:   1%|          | 4/363 [00:05<06:29,  1.08s/it]
Continual predict:   1%|▏         | 5/363 [00:07<07:52,  1.32s/it]
Continual predict:   2%|▏         | 6/363 [00:11<14:11,  2.39s/it]
Continual predict:   2%|▏         | 7/363 [00:14<15:39,  2.64s/it]
Continual predict:   2%|▏         | 8/363 [00:16<14:18,  2.42s/it]
Continual predict:   2%|▏         | 9/363 [00:18<12:44,  2.16s/it]
Continual predict:   3%|▎         | 10/363 [00:18<10:09,  1.73s/it]
Continual predict:   3%|▎         | 11/363 [00:21<11:33,  1.97s/it]
Continual predict:   3%|▎         | 12/363 [00:24<13:18,  2.28s/it]
Continual predict:   4%|▎         | 13/363 [00:25<11:18,  1.94s/it]
Continual predict:   4%|▍         | 14/363 [00:27<11:51,  2.04s/it]
Continual predict:   4%|▍         | 15/363 [00:29<12:00,  2.07s/it]
Continual predict:   4%|▍         | 16/363 [00:32<12:37,  2.18s/it]
Continual predict:   5%|▍         | 17/363 [00:39<21:40,  3.76s/it]
Continual predict:   5%|▍         | 18/363 [00:41<18:34,  3.23s/it]
Continual predict:   5%|▌         | 19/363 [00:44<17:07,  2.99s/it]
Continual predict:   6%|▌         | 20/363 [00:45<14:10,  2.48s/it]
Continual predict:   6%|▌         | 21/363 [00:45<10:08,  1.78s/it]
Continual predict:   6%|▌         | 22/363 [00:45<07:18,  1.29s/it]
Continual predict:   6%|▋         | 23/363 [00:46<05:21,  1.06it/s]
Continual predict:   7%|▋         | 24/363 [00:46<03:57,  1.43it/s]
Continual predict:   7%|▋         | 25/363 [00:46<02:59,  1.88it/s]
Continual predict:   7%|▋         | 26/363 [00:46<02:18,  2.43it/s]
Continual predict:   7%|▋         | 27/363 [00:46<01:50,  3.04it/s]
Continual predict:   8%|▊         | 28/363 [00:46<01:30,  3.70it/s]
Continual predict:   8%|▊         | 29/363 [00:46<01:16,  4.37it/s]
Continual predict:   8%|▊         | 30/363 [00:46<01:08,  4.89it/s]
Continual predict:   9%|▊         | 31/363 [00:47<01:00,  5.45it/s]
Continual predict:   9%|▉         | 32/363 [00:47<00:55,  5.94it/s]
Continual predict:   9%|▉         | 33/363 [00:47<00:53,  6.15it/s]
Continual predict:   9%|▉         | 34/363 [00:47<00:51,  6.33it/s]
Continual predict:  10%|▉         | 35/363 [00:47<00:50,  6.46it/s]
Continual predict:  10%|▉         | 36/363 [00:47<00:48,  6.76it/s]
Continual predict:  10%|█         | 37/363 [00:47<00:48,  6.77it/s]
Continual predict:  10%|█         | 38/363 [00:48<00:46,  6.95it/s]
Continual predict:  11%|█         | 39/363 [00:48<00:45,  7.11it/s]
Continual predict:  11%|█         | 40/363 [00:48<00:44,  7.24it/s]
Continual predict:  11%|█▏        | 41/363 [00:48<00:45,  7.11it/s]
Continual predict:  12%|█▏        | 42/363 [00:48<00:44,  7.21it/s]
Continual predict:  12%|█▏        | 43/363 [00:48<00:45,  7.09it/s]
Continual predict:  12%|█▏        | 44/363 [00:48<00:45,  6.99it/s]
Continual predict:  12%|█▏        | 45/363 [00:49<00:45,  7.05it/s]
Continual predict:  13%|█▎        | 46/363 [00:49<00:44,  7.06it/s]
Continual predict:  13%|█▎        | 47/363 [00:49<00:44,  7.06it/s]
Continual predict:  13%|█▎        | 48/363 [00:49<00:45,  6.96it/s]
Continual predict:  13%|█▎        | 49/363 [00:49<00:45,  6.93it/s]
Continual predict:  14%|█▍        | 50/363 [00:49<00:44,  7.07it/s]
Continual predict:  14%|█▍        | 51/363 [00:49<00:43,  7.11it/s]
Continual predict:  14%|█▍        | 52/363 [00:50<00:43,  7.21it/s]
Continual predict:  15%|█▍        | 53/363 [00:50<00:42,  7.28it/s]
Continual predict:  15%|█▍        | 54/363 [00:50<00:43,  7.13it/s]
Continual predict:  15%|█▌        | 55/363 [00:50<00:42,  7.25it/s]
Continual predict:  15%|█▌        | 56/363 [00:50<00:43,  7.10it/s]
Continual predict:  16%|█▌        | 57/363 [00:50<00:42,  7.21it/s]
Continual predict:  16%|█▌        | 58/363 [00:50<00:41,  7.28it/s]
Continual predict:  16%|█▋        | 59/363 [00:51<00:41,  7.33it/s]
Continual predict:  17%|█▋        | 60/363 [00:51<00:42,  7.12it/s]
Continual predict:  17%|█▋        | 61/363 [00:51<00:41,  7.23it/s]
Continual predict:  17%|█▋        | 62/363 [00:51<00:41,  7.30it/s]
Continual predict:  17%|█▋        | 63/363 [00:51<00:41,  7.16it/s]
Continual predict:  18%|█▊        | 64/363 [00:51<00:41,  7.24it/s]
Continual predict:  18%|█▊        | 65/363 [00:51<00:40,  7.30it/s]
Continual predict:  18%|█▊        | 66/363 [00:51<00:40,  7.34it/s]
Continual predict:  18%|█▊        | 67/363 [00:52<00:39,  7.40it/s]
Continual predict:  19%|█▊        | 68/363 [00:52<00:39,  7.43it/s]
Continual predict:  19%|█▉        | 69/363 [00:52<00:39,  7.45it/s]
Continual predict:  19%|█▉        | 70/363 [00:52<00:39,  7.47it/s]
Continual predict:  20%|█▉        | 71/363 [00:52<00:39,  7.47it/s]
Continual predict:  20%|█▉        | 72/363 [00:52<00:40,  7.26it/s]
Continual predict:  20%|██        | 73/363 [00:52<00:39,  7.34it/s]
Continual predict:  20%|██        | 74/363 [00:53<00:40,  7.16it/s]
Continual predict:  21%|██        | 75/363 [00:53<00:40,  7.17it/s]
Continual predict:  21%|██        | 76/363 [00:53<00:39,  7.28it/s]
Continual predict:  21%|██        | 77/363 [00:53<00:39,  7.33it/s]
Continual predict:  21%|██▏       | 78/363 [00:53<00:38,  7.39it/s]
Continual predict:  22%|██▏       | 79/363 [00:53<00:38,  7.40it/s]
Continual predict:  22%|██▏       | 80/363 [00:53<00:38,  7.44it/s]
Continual predict:  22%|██▏       | 81/363 [00:54<00:37,  7.44it/s]
Continual predict:  23%|██▎       | 82/363 [00:54<00:37,  7.46it/s]
Continual predict:  23%|██▎       | 83/363 [00:54<00:37,  7.46it/s]
Continual predict:  23%|██▎       | 84/363 [00:54<00:38,  7.24it/s]
Continual predict:  23%|██▎       | 85/363 [00:54<00:38,  7.29it/s]
Continual predict:  24%|██▎       | 86/363 [00:54<00:37,  7.34it/s]
Continual predict:  24%|██▍       | 87/363 [00:54<00:37,  7.34it/s]
Continual predict:  24%|██▍       | 88/363 [00:54<00:37,  7.40it/s]
Continual predict:  25%|██▍       | 89/363 [00:55<00:36,  7.43it/s]
Continual predict:  25%|██▍       | 90/363 [00:55<00:36,  7.47it/s]
Continual predict:  25%|██▌       | 91/363 [00:55<00:36,  7.46it/s]
Continual predict:  25%|██▌       | 92/363 [00:55<00:36,  7.42it/s]
Continual predict:  26%|██▌       | 93/363 [00:55<00:37,  7.22it/s]
Continual predict:  26%|██▌       | 94/363 [00:55<00:38,  7.08it/s]
Continual predict:  26%|██▌       | 95/363 [00:55<00:37,  7.18it/s]
Continual predict:  26%|██▋       | 96/363 [00:56<00:39,  6.74it/s]
Continual predict:  27%|██▋       | 97/363 [00:56<00:38,  6.94it/s]
Continual predict:  27%|██▋       | 98/363 [00:56<00:38,  6.91it/s]
Continual predict:  27%|██▋       | 99/363 [00:56<00:38,  6.88it/s]
Continual predict:  28%|██▊       | 100/363 [00:56<00:38,  6.86it/s]
Continual predict:  28%|██▊       | 101/363 [00:56<00:37,  7.03it/s]
Continual predict:  28%|██▊       | 102/363 [00:56<00:36,  7.18it/s]
Continual predict:  28%|██▊       | 103/363 [00:57<00:35,  7.27it/s]
Continual predict:  29%|██▊       | 104/363 [00:57<00:35,  7.32it/s]
Continual predict:  29%|██▉       | 105/363 [00:57<00:36,  7.17it/s]
Continual predict:  29%|██▉       | 106/363 [00:57<00:35,  7.27it/s]
Continual predict:  29%|██▉       | 107/363 [00:57<00:34,  7.36it/s]
Continual predict:  30%|██▉       | 108/363 [00:57<00:34,  7.39it/s]
Continual predict:  30%|███       | 109/363 [00:57<00:34,  7.41it/s]
Continual predict:  30%|███       | 110/363 [00:58<00:34,  7.42it/s]
Continual predict:  31%|███       | 111/363 [00:58<00:33,  7.43it/s]
Continual predict:  31%|███       | 112/363 [00:58<00:33,  7.44it/s]
Continual predict:  31%|███       | 113/363 [00:58<00:33,  7.43it/s]
Continual predict:  31%|███▏      | 114/363 [00:58<00:33,  7.42it/s]
Continual predict:  32%|███▏      | 115/363 [00:58<00:34,  7.24it/s]
Continual predict:  32%|███▏      | 116/363 [00:58<00:33,  7.30it/s]
Continual predict:  32%|███▏      | 117/363 [00:58<00:34,  7.16it/s]
Continual predict:  33%|███▎      | 118/363 [00:59<00:34,  7.05it/s]
Continual predict:  33%|███▎      | 119/363 [00:59<00:34,  7.16it/s]
Continual predict:  33%|███▎      | 120/363 [00:59<00:33,  7.24it/s]
Continual predict:  33%|███▎      | 121/363 [01:03<05:42,  1.42s/it]
Continual predict:  34%|███▎      | 122/363 [01:07<08:56,  2.23s/it]
Continual predict:  34%|███▍      | 123/363 [01:13<12:26,  3.11s/it]
Continual predict:  34%|███▍      | 124/363 [01:15<11:23,  2.86s/it]
Continual predict:  34%|███▍      | 125/363 [01:20<13:32,  3.41s/it]
Continual predict:  35%|███▍      | 126/363 [01:24<14:57,  3.79s/it]
Continual predict:  35%|███▍      | 127/363 [01:26<12:11,  3.10s/it]
Continual predict:  35%|███▌      | 128/363 [01:28<11:37,  2.97s/it]
Continual predict:  36%|███▌      | 129/363 [01:32<11:57,  3.07s/it]
Continual predict:  36%|███▌      | 130/363 [01:35<11:44,  3.02s/it]
Continual predict:  36%|███▌      | 131/363 [01:37<10:39,  2.76s/it]
Continual predict:  36%|███▋      | 132/363 [01:39<10:24,  2.70s/it]
Continual predict:  37%|███▋      | 133/363 [01:42<10:48,  2.82s/it]
Continual predict:  37%|███▋      | 134/363 [01:48<14:14,  3.73s/it]
Continual predict:  37%|███▋      | 135/363 [01:53<15:13,  4.01s/it]
Continual predict:  37%|███▋      | 136/363 [01:56<14:10,  3.75s/it]
Continual predict:  38%|███▊      | 137/363 [01:58<12:19,  3.27s/it]
Continual predict:  38%|███▊      | 138/363 [01:59<09:22,  2.50s/it]
Continual predict:  38%|███▊      | 139/363 [02:00<07:47,  2.09s/it]
Continual predict:  39%|███▊      | 140/363 [02:03<08:19,  2.24s/it]
Continual predict:  39%|███▉      | 141/363 [02:09<12:56,  3.50s/it]
Continual predict:  39%|███▉      | 142/363 [02:14<13:55,  3.78s/it]
Continual predict:  39%|███▉      | 143/363 [02:17<13:12,  3.60s/it]
Continual predict:  40%|███▉      | 144/363 [02:26<19:19,  5.29s/it]
Continual predict:  40%|███▉      | 145/363 [02:30<18:06,  4.98s/it]
Continual predict:  40%|████      | 146/363 [02:32<14:30,  4.01s/it]
Continual predict:  40%|████      | 147/363 [02:36<14:43,  4.09s/it]
Continual predict:  41%|████      | 148/363 [02:38<12:29,  3.49s/it]
Continual predict:  41%|████      | 149/363 [02:41<12:00,  3.37s/it]
Continual predict:  41%|████▏     | 150/363 [02:43<10:18,  2.91s/it]
Continual predict:  42%|████▏     | 151/363 [02:46<10:21,  2.93s/it]
Continual predict:  42%|████▏     | 152/363 [02:52<13:01,  3.70s/it]
Continual predict:  42%|████▏     | 153/363 [02:54<11:34,  3.31s/it]
Continual predict:  42%|████▏     | 154/363 [02:59<12:58,  3.72s/it]
Continual predict:  43%|████▎     | 155/363 [03:02<12:38,  3.65s/it]
Continual predict:  43%|████▎     | 156/363 [03:06<12:11,  3.53s/it]
Continual predict:  43%|████▎     | 157/363 [03:09<12:19,  3.59s/it]
Continual predict:  44%|████▎     | 158/363 [03:10<09:31,  2.79s/it]
Continual predict:  44%|████▍     | 159/363 [03:17<13:32,  3.98s/it]
Continual predict:  44%|████▍     | 160/363 [03:19<11:45,  3.47s/it]
Continual predict:  44%|████▍     | 161/363 [03:22<10:37,  3.15s/it]
Continual predict:  45%|████▍     | 162/363 [03:26<12:07,  3.62s/it]
Continual predict:  45%|████▍     | 163/363 [03:28<10:31,  3.16s/it]
Continual predict:  45%|████▌     | 164/363 [03:31<09:30,  2.87s/it]
Continual predict:  45%|████▌     | 165/363 [03:34<09:37,  2.92s/it]
Continual predict:  46%|████▌     | 166/363 [03:36<08:45,  2.67s/it]
Continual predict:  46%|████▌     | 167/363 [03:37<07:21,  2.25s/it]
Continual predict:  46%|████▋     | 168/363 [03:39<07:17,  2.24s/it]
Continual predict:  47%|████▋     | 169/363 [03:41<06:41,  2.07s/it]
Continual predict:  47%|████▋     | 170/363 [03:44<07:30,  2.34s/it]
Continual predict:  47%|████▋     | 171/363 [03:45<06:37,  2.07s/it]
Continual predict:  47%|████▋     | 172/363 [03:48<06:59,  2.20s/it]
Continual predict:  48%|████▊     | 173/363 [03:52<08:31,  2.69s/it]
Continual predict:  48%|████▊     | 174/363 [03:53<06:59,  2.22s/it]
Continual predict:  48%|████▊     | 175/363 [03:56<08:01,  2.56s/it]
Continual predict:  48%|████▊     | 176/363 [03:59<08:10,  2.63s/it]
Continual predict:  49%|████▉     | 177/363 [04:02<08:07,  2.62s/it]
Continual predict:  49%|████▉     | 178/363 [04:03<06:46,  2.19s/it]
Continual predict:  49%|████▉     | 179/363 [04:04<05:41,  1.86s/it]
Continual predict:  50%|████▉     | 180/363 [04:05<05:08,  1.68s/it]
Continual predict:  50%|████▉     | 181/363 [04:06<04:44,  1.56s/it]
Continual predict:  50%|█████     | 182/363 [04:11<07:27,  2.47s/it]
Continual predict:  50%|█████     | 183/363 [04:11<05:22,  1.79s/it]
Continual predict:  51%|█████     | 184/363 [04:12<04:42,  1.58s/it]
Continual predict:  51%|█████     | 185/363 [04:15<06:08,  2.07s/it]
Continual predict:  51%|█████     | 186/363 [04:17<05:13,  1.77s/it]
Continual predict:  52%|█████▏    | 187/363 [04:20<06:25,  2.19s/it]
Continual predict:  52%|█████▏    | 188/363 [04:22<06:46,  2.32s/it]
Continual predict:  52%|█████▏    | 189/363 [04:27<08:54,  3.07s/it]
Continual predict:  52%|█████▏    | 190/363 [04:30<08:54,  3.09s/it]
Continual predict:  53%|█████▎    | 191/363 [04:32<07:16,  2.54s/it]
Continual predict:  53%|█████▎    | 192/363 [04:35<07:47,  2.73s/it]
Continual predict:  53%|█████▎    | 193/363 [04:39<09:09,  3.23s/it]
Continual predict:  53%|█████▎    | 194/363 [04:44<10:18,  3.66s/it]
Continual predict:  54%|█████▎    | 195/363 [04:49<11:29,  4.10s/it]
Continual predict:  54%|█████▍    | 196/363 [04:52<10:42,  3.85s/it]
Continual predict:  54%|█████▍    | 197/363 [04:55<10:06,  3.65s/it]
Continual predict:  55%|█████▍    | 198/363 [05:03<12:56,  4.71s/it]
Continual predict:  55%|█████▍    | 199/363 [05:04<09:52,  3.62s/it]
Continual predict:  55%|█████▌    | 200/363 [05:05<08:00,  2.94s/it]
Continual predict:  55%|█████▌    | 201/363 [05:07<07:31,  2.79s/it]
Continual predict:  56%|█████▌    | 202/363 [05:11<07:45,  2.89s/it]
Continual predict:  56%|█████▌    | 203/363 [05:16<09:23,  3.52s/it]
Continual predict:  56%|█████▌    | 204/363 [05:18<08:51,  3.34s/it]
Continual predict:  56%|█████▋    | 205/363 [05:25<11:14,  4.27s/it]
Continual predict:  57%|█████▋    | 206/363 [05:28<09:54,  3.79s/it]
Continual predict:  57%|█████▋    | 207/363 [05:31<09:43,  3.74s/it]
Continual predict:  57%|█████▋    | 208/363 [05:37<11:24,  4.42s/it]
Continual predict:  58%|█████▊    | 209/363 [05:40<09:56,  3.87s/it]
Continual predict:  58%|█████▊    | 210/363 [05:41<08:09,  3.20s/it]
Continual predict:  58%|█████▊    | 211/363 [05:44<07:34,  2.99s/it]
Continual predict:  58%|█████▊    | 212/363 [05:47<07:31,  2.99s/it]
Continual predict:  59%|█████▊    | 213/363 [05:50<07:56,  3.18s/it]
Continual predict:  59%|█████▉    | 214/363 [05:54<08:06,  3.26s/it]
Continual predict:  59%|█████▉    | 215/363 [05:55<06:09,  2.49s/it]
Continual predict:  60%|█████▉    | 216/363 [05:58<06:27,  2.64s/it]
Continual predict:  60%|█████▉    | 217/363 [06:00<05:54,  2.43s/it]
Continual predict:  60%|██████    | 218/363 [06:03<06:50,  2.83s/it]
Continual predict:  60%|██████    | 219/363 [06:06<06:56,  2.90s/it]
Continual predict:  61%|██████    | 220/363 [06:10<07:38,  3.21s/it]
Continual predict:  61%|██████    | 221/363 [06:10<05:27,  2.31s/it]
Continual predict:  61%|██████    | 222/363 [06:11<03:58,  1.69s/it]
Continual predict:  61%|██████▏   | 223/363 [06:11<02:54,  1.24s/it]
Continual predict:  62%|██████▏   | 224/363 [06:11<02:09,  1.07it/s]
Continual predict:  62%|██████▏   | 225/363 [06:11<01:38,  1.40it/s]
Continual predict:  62%|██████▏   | 226/363 [06:12<01:16,  1.79it/s]
Continual predict:  63%|██████▎   | 227/363 [06:12<01:03,  2.14it/s]
Continual predict:  63%|██████▎   | 228/363 [06:12<00:54,  2.49it/s]
Continual predict:  63%|██████▎   | 229/363 [06:12<00:46,  2.89it/s]
Continual predict:  63%|██████▎   | 230/363 [06:13<00:48,  2.76it/s]
Continual predict:  64%|██████▎   | 231/363 [06:13<00:41,  3.19it/s]
Continual predict:  64%|██████▍   | 232/363 [06:13<00:36,  3.58it/s]
Continual predict:  64%|██████▍   | 233/363 [06:13<00:33,  3.92it/s]
Continual predict:  64%|██████▍   | 234/363 [06:13<00:30,  4.19it/s]
Continual predict:  65%|██████▍   | 235/363 [06:14<00:28,  4.41it/s]
Continual predict:  65%|██████▌   | 236/363 [06:14<00:27,  4.58it/s]
Continual predict:  65%|██████▌   | 237/363 [06:14<00:26,  4.71it/s]
Continual predict:  66%|██████▌   | 238/363 [06:14<00:26,  4.81it/s]
Continual predict:  66%|██████▌   | 239/363 [06:14<00:25,  4.88it/s]
Continual predict:  66%|██████▌   | 240/363 [06:15<00:29,  4.11it/s]
Continual predict:  66%|██████▋   | 241/363 [06:15<00:28,  4.35it/s]
Continual predict:  67%|██████▋   | 242/363 [06:15<00:26,  4.54it/s]
Continual predict:  67%|██████▋   | 243/363 [06:15<00:25,  4.68it/s]
Continual predict:  67%|██████▋   | 244/363 [06:16<00:24,  4.78it/s]
Continual predict:  67%|██████▋   | 245/363 [06:16<00:28,  4.18it/s]
Continual predict:  68%|██████▊   | 246/363 [06:16<00:27,  4.30it/s]
Continual predict:  68%|██████▊   | 247/363 [06:16<00:27,  4.20it/s]
Continual predict:  68%|██████▊   | 248/363 [06:17<00:26,  4.42it/s]
Continual predict:  69%|██████▊   | 249/363 [06:17<00:26,  4.28it/s]
Continual predict:  69%|██████▉   | 250/363 [06:17<00:25,  4.48it/s]
Continual predict:  69%|██████▉   | 251/363 [06:17<00:25,  4.32it/s]
Continual predict:  69%|██████▉   | 252/363 [06:18<00:57,  1.93it/s]
Continual predict:  70%|██████▉   | 253/363 [06:19<00:48,  2.28it/s]
Continual predict:  70%|██████▉   | 254/363 [06:19<00:39,  2.73it/s]
Continual predict:  70%|███████   | 255/363 [06:19<00:34,  3.16it/s]
Continual predict:  71%|███████   | 256/363 [06:19<00:30,  3.55it/s]
Continual predict:  71%|███████   | 257/363 [06:20<00:42,  2.47it/s]
Continual predict:  71%|███████   | 258/363 [06:20<00:35,  2.92it/s]
Continual predict:  71%|███████▏  | 259/363 [06:20<00:31,  3.34it/s]
Continual predict:  72%|███████▏  | 260/363 [06:21<00:27,  3.71it/s]
Continual predict:  72%|███████▏  | 261/363 [06:21<00:35,  2.87it/s]
Continual predict:  72%|███████▏  | 262/363 [06:21<00:31,  3.23it/s]
Continual predict:  72%|███████▏  | 263/363 [06:22<00:29,  3.43it/s]
Continual predict:  73%|███████▎  | 264/363 [06:23<00:55,  1.78it/s]
Continual predict:  73%|███████▎  | 265/363 [06:23<00:53,  1.83it/s]
Continual predict:  73%|███████▎  | 266/363 [06:24<00:52,  1.86it/s]
Continual predict:  74%|███████▎  | 267/363 [06:24<00:53,  1.81it/s]
Continual predict:  74%|███████▍  | 268/363 [06:25<00:59,  1.60it/s]
Continual predict:  74%|███████▍  | 269/363 [06:26<00:57,  1.63it/s]
Continual predict:  74%|███████▍  | 270/363 [06:26<00:58,  1.58it/s]
Continual predict:  75%|███████▍  | 271/363 [06:27<01:01,  1.50it/s]
Continual predict:  75%|███████▍  | 272/363 [06:28<01:06,  1.37it/s]
Continual predict:  75%|███████▌  | 273/363 [06:29<01:02,  1.44it/s]
Continual predict:  75%|███████▌  | 274/363 [06:29<01:00,  1.48it/s]
Continual predict:  76%|███████▌  | 275/363 [06:30<00:57,  1.54it/s]
Continual predict:  76%|███████▌  | 276/363 [06:31<01:06,  1.31it/s]
Continual predict:  76%|███████▋  | 277/363 [06:32<01:17,  1.11it/s]
Continual predict:  77%|███████▋  | 278/363 [06:33<01:14,  1.14it/s]
Continual predict:  77%|███████▋  | 279/363 [06:34<01:08,  1.22it/s]
Continual predict:  77%|███████▋  | 280/363 [06:34<00:58,  1.42it/s]
Continual predict:  77%|███████▋  | 281/363 [06:35<01:03,  1.30it/s]
Continual predict:  78%|███████▊  | 282/363 [06:36<01:02,  1.30it/s]
Continual predict:  78%|███████▊  | 283/363 [06:36<00:58,  1.37it/s]
Continual predict:  78%|███████▊  | 284/363 [06:37<00:51,  1.54it/s]
Continual predict:  79%|███████▊  | 285/363 [06:38<00:53,  1.46it/s]
Continual predict:  79%|███████▉  | 286/363 [06:38<00:53,  1.43it/s]
Continual predict:  79%|███████▉  | 287/363 [06:39<00:54,  1.40it/s]
Continual predict:  79%|███████▉  | 288/363 [06:40<00:50,  1.47it/s]
Continual predict:  80%|███████▉  | 289/363 [06:41<00:55,  1.34it/s]
Continual predict:  80%|███████▉  | 290/363 [06:41<00:54,  1.34it/s]
Continual predict:  80%|████████  | 291/363 [06:42<00:50,  1.43it/s]
Continual predict:  80%|████████  | 292/363 [06:42<00:41,  1.73it/s]
Continual predict:  81%|████████  | 293/363 [06:43<00:45,  1.54it/s]
Continual predict:  81%|████████  | 294/363 [06:44<00:41,  1.65it/s]
Continual predict:  81%|████████▏ | 295/363 [06:44<00:37,  1.81it/s]
Continual predict:  82%|████████▏ | 296/363 [06:45<00:44,  1.50it/s]
Continual predict:  82%|████████▏ | 297/363 [06:45<00:40,  1.62it/s]
Continual predict:  82%|████████▏ | 298/363 [06:46<00:36,  1.77it/s]
Continual predict:  82%|████████▏ | 299/363 [06:47<00:42,  1.52it/s]
Continual predict:  83%|████████▎ | 300/363 [06:47<00:34,  1.81it/s]
Continual predict:  83%|████████▎ | 301/363 [06:48<00:40,  1.53it/s]
Continual predict:  83%|████████▎ | 302/363 [06:48<00:35,  1.73it/s]
Continual predict:  83%|████████▎ | 303/363 [06:50<00:50,  1.18it/s]
Continual predict:  84%|████████▎ | 304/363 [06:50<00:40,  1.46it/s]
Continual predict:  84%|████████▍ | 305/363 [06:51<00:43,  1.34it/s]
Continual predict:  84%|████████▍ | 306/363 [06:52<00:42,  1.33it/s]
Continual predict:  85%|████████▍ | 307/363 [06:53<00:42,  1.32it/s]
Continual predict:  85%|████████▍ | 308/363 [06:53<00:35,  1.53it/s]
Continual predict:  85%|████████▌ | 309/363 [06:53<00:31,  1.69it/s]
Continual predict:  85%|████████▌ | 310/363 [06:54<00:31,  1.69it/s]
Continual predict:  86%|████████▌ | 311/363 [06:55<00:29,  1.76it/s]
Continual predict:  86%|████████▌ | 312/363 [06:55<00:30,  1.66it/s]
Continual predict:  86%|████████▌ | 313/363 [06:56<00:33,  1.48it/s]
Continual predict:  87%|████████▋ | 314/363 [06:57<00:29,  1.66it/s]
Continual predict:  87%|████████▋ | 315/363 [06:57<00:32,  1.46it/s]
Continual predict:  87%|████████▋ | 316/363 [06:58<00:33,  1.40it/s]
Continual predict:  87%|████████▋ | 317/363 [06:58<00:27,  1.69it/s]
Continual predict:  88%|████████▊ | 318/363 [06:59<00:24,  1.83it/s]
Continual predict:  88%|████████▊ | 319/363 [07:00<00:37,  1.18it/s]
Continual predict:  88%|████████▊ | 320/363 [07:01<00:34,  1.26it/s]
Continual predict:  88%|████████▊ | 321/363 [07:02<00:31,  1.32it/s]
Continual predict:  89%|████████▊ | 322/363 [07:02<00:26,  1.53it/s]
Continual predict:  89%|████████▉ | 323/363 [07:03<00:26,  1.49it/s]
Continual predict:  89%|████████▉ | 324/363 [07:04<00:28,  1.37it/s]
Continual predict:  90%|████████▉ | 325/363 [07:05<00:27,  1.36it/s]
Continual predict:  90%|████████▉ | 326/363 [07:05<00:25,  1.44it/s]
Continual predict:  90%|█████████ | 327/363 [07:06<00:29,  1.24it/s]
Continual predict:  90%|█████████ | 328/363 [07:07<00:30,  1.15it/s]
Continual predict:  91%|█████████ | 329/363 [07:08<00:25,  1.31it/s]
Continual predict:  91%|█████████ | 330/363 [07:08<00:20,  1.60it/s]
Continual predict:  91%|█████████ | 331/363 [07:09<00:19,  1.63it/s]
Continual predict:  91%|█████████▏| 332/363 [07:09<00:18,  1.72it/s]
Continual predict:  92%|█████████▏| 333/363 [07:10<00:22,  1.31it/s]
Continual predict:  92%|█████████▏| 334/363 [07:11<00:23,  1.26it/s]
Continual predict:  92%|█████████▏| 335/363 [07:12<00:19,  1.47it/s]
Continual predict:  93%|█████████▎| 336/363 [07:12<00:16,  1.65it/s]
Continual predict:  93%|█████████▎| 337/363 [07:13<00:15,  1.70it/s]
Continual predict:  93%|█████████▎| 338/363 [07:13<00:14,  1.67it/s]
Continual predict:  93%|█████████▎| 339/363 [07:14<00:13,  1.76it/s]
Continual predict:  94%|█████████▎| 340/363 [07:15<00:16,  1.40it/s]
Continual predict:  94%|█████████▍| 341/363 [07:15<00:15,  1.39it/s]
Continual predict:  94%|█████████▍| 342/363 [07:16<00:13,  1.58it/s]
Continual predict:  94%|█████████▍| 343/363 [07:17<00:13,  1.51it/s]
Continual predict:  95%|█████████▍| 344/363 [07:18<00:14,  1.35it/s]
Continual predict:  95%|█████████▌| 345/363 [07:18<00:11,  1.54it/s]
Continual predict:  95%|█████████▌| 346/363 [07:19<00:11,  1.45it/s]
Continual predict:  96%|█████████▌| 347/363 [07:20<00:14,  1.11it/s]
Continual predict:  96%|█████████▌| 348/363 [07:21<00:10,  1.37it/s]
Continual predict:  96%|█████████▌| 349/363 [07:21<00:08,  1.58it/s]
Continual predict:  96%|█████████▋| 350/363 [07:22<00:08,  1.59it/s]
Continual predict:  97%|█████████▋| 351/363 [07:23<00:09,  1.31it/s]
Continual predict:  97%|█████████▋| 352/363 [07:23<00:08,  1.32it/s]
Continual predict:  97%|█████████▋| 353/363 [07:25<00:08,  1.11it/s]
Continual predict:  98%|█████████▊| 354/363 [07:26<00:08,  1.01it/s]
Continual predict:  98%|█████████▊| 355/363 [07:26<00:06,  1.15it/s]
Continual predict:  98%|█████████▊| 356/363 [07:27<00:06,  1.14it/s]
Continual predict:  98%|█████████▊| 357/363 [07:28<00:05,  1.08it/s]
Continual predict:  99%|█████████▊| 358/363 [07:29<00:04,  1.15it/s]
Continual predict:  99%|█████████▉| 359/363 [07:30<00:03,  1.27it/s]
Continual predict:  99%|█████████▉| 360/363 [07:31<00:02,  1.09it/s]
Continual predict:  99%|█████████▉| 361/363 [07:32<00:02,  1.10s/it]
Continual predict: 100%|█████████▉| 362/363 [07:33<00:00,  1.06it/s]
Continual predict: 100%|██████████| 363/363 [07:33<00:00,  1.26it/s]
Continual predict: 100%|██████████| 363/363 [07:33<00:00,  1.25s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=20]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (363/363)  label_acc=100.0% (363/363)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                        =       15.0
+  predict_eval_rougeL                                          =    63.5731
+  predict_eval_rougeL_for_CL                                   =    63.5731
+  predict_eval_rougeL_for_task002_quoref_answer_generation     =     76.381
+  predict_eval_rougeL_for_task1290_xsum_summarization          =    29.5925
+  predict_eval_rougeL_for_task1572_samsum_summary              =    41.3728
+  predict_eval_rougeL_for_task181_outcome_extraction           =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification =       91.0
+  predict_exact_match                                          =    50.6887
+  predict_exact_match_for_CL                                   =    50.6887
+  predict_exact_match_for_task002_quoref_answer_generation     =       70.0
+  predict_exact_match_for_task1290_xsum_summarization          =        0.0
+  predict_exact_match_for_task1572_samsum_summary              =        0.0
+  predict_exact_match_for_task181_outcome_extraction           =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification =       91.0
+  predict_gen_len                                              =    10.3003
+  predict_global_step                                          =        480
+  predict_rouge1                                               =    66.0415
+  predict_rouge1_for_CL                                        =    66.0415
+  predict_rouge1_for_task002_quoref_answer_generation          =     76.381
+  predict_rouge1_for_task1290_xsum_summarization               =      36.75
+  predict_rouge1_for_task1572_samsum_summary                   =    50.3879
+  predict_rouge1_for_task181_outcome_extraction                =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification      =       91.0
+  predict_runtime                                              = 0:07:33.93
+  predict_samples                                              =        363
+  predict_samples_per_second                                   =        0.8
+  predict_steps_per_second                                     =        0.8
+[2026-04-29 22:55:40,464] [INFO] [launch.py:367:main] Process 786881 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/all_results.json b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..6cd9ce30605486510fb2e638edce92e08c01a362
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/all_results.json
@@ -0,0 +1,39 @@
+{
+  "predict_exact_match": 60.9071,
+  "predict_rouge1": 73.1888,
+  "predict_eval_rougeL": 71.2535,
+  "predict_gen_len": 9.054,
+  "predict_runtime": 489.0271430411376,
+  "predict_samples": 463,
+  "predict_samples_per_second": 0.947,
+  "predict_steps_per_second": 0.947,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 60.9071,
+  "predict_rouge1_for_CL": 73.1888,
+  "predict_eval_rougeL_for_CL": 71.2535,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "train_runtime": 1824.6796,
+  "train_samples_per_second": 8.221,
+  "train_steps_per_second": 0.263,
+  "total_flos": 6.583819752282522e+16,
+  "train_loss": 0.06608551328523997,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..320bab7b03e163cd62514b39a68683a028b43c3d
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4dee7d527c0edf4cae44f9db8219f8ea6b8f3bab5a6dd09f747f8233704860c
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..def902b3106457179a30182ddeb0421106c64c5a
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee3e425d5c9e8604a4ce7f0893907bc18b4f0459eafffb4d15278d07c51f0738
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..0c905e2d3c4bdae0553b89d5a73378919bf24328
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ddf6b81eddee2624f0cf6f87a19d20a14d6292813daf352112fd68da3149fa2
+size 992519515
diff --git a/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/train.log b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..4103753c3876b36857afeb5a4b912c77d3253669
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/train.log
@@ -0,0 +1,538 @@
+[2026-04-29 22:55:46,459] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 22:55:46,459] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29505 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1510_evalution_relation_extraction --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction --cur_task_id 5 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[2026-04-29 22:55:49,655] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 22:55:49,655] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 22:55:49,655] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 22:55:49,655] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 22:55:49,655] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 22:55:49,656] [INFO] [launch.py:272:main] process 966508 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1510_evalution_relation_extraction', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction', '--cur_task_id', '5', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights']
+
+============================================================
+Task 5: task1510_evalution_relation_extraction
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/463
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 109.91it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 147.55it/s]
Loading weights:  18%|█▊        | 51/291 [00:00<00:01, 166.78it/s]
Loading weights:  24%|██▎       | 69/291 [00:00<00:01, 171.54it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 171.80it/s]
Loading weights:  36%|███▌      | 105/291 [00:00<00:01, 168.92it/s]
Loading weights:  44%|████▎     | 127/291 [00:00<00:00, 182.46it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 183.09it/s]
Loading weights:  57%|█████▋    | 165/291 [00:00<00:00, 179.27it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 174.99it/s]
Loading weights:  70%|███████   | 204/291 [00:01<00:00, 181.83it/s]
Loading weights:  77%|███████▋  | 225/291 [00:01<00:00, 182.94it/s]
Loading weights:  84%|████████▍ | 244/291 [00:01<00:00, 181.48it/s]
Loading weights:  91%|█████████ | 265/291 [00:01<00:00, 176.69it/s]
Loading weights:  98%|█████████▊| 284/291 [00:01<00:00, 180.29it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 178.14it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=803)
+[FIX] Re-initialized lora_A/B in 384 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(5, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 5 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6751.842M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 5 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 5 sigs, cur=task1510_evalution_relation_extraction, mapping={'task1510_evalution_relation_extraction': 0, 'task1572_samsum_summary': 5, 'task363_sst2_polarity_classification': 4, 'task1290_xsum_summarization': 3, 'task181_outcome_extraction': 2, 'task002_quoref_answer_generation': 1}
+[SGWI] Task 5: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 22:56:47,932] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:04<32:46,  4.10s/it]
  0%|          | 2/480 [00:07<30:07,  3.78s/it]
  1%|          | 3/480 [00:11<29:57,  3.77s/it]
  1%|          | 4/480 [00:15<30:12,  3.81s/it]
  1%|          | 5/480 [00:18<29:41,  3.75s/it]
  1%|▏         | 6/480 [00:23<30:29,  3.86s/it]
  1%|▏         | 7/480 [00:26<30:07,  3.82s/it]
  2%|▏         | 8/480 [00:30<29:51,  3.80s/it]
  2%|▏         | 9/480 [00:34<29:45,  3.79s/it]
  2%|▏         | 10/480 [00:38<29:59,  3.83s/it]
                                                
{'loss': '1.971', 'grad_norm': '0.4898', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:38<29:59,  3.83s/it]
  2%|▏         | 11/480 [00:41<29:44,  3.81s/it]
  2%|▎         | 12/480 [00:45<30:07,  3.86s/it]
  3%|▎         | 13/480 [00:49<29:25,  3.78s/it]
  3%|▎         | 14/480 [00:53<28:52,  3.72s/it]
  3%|▎         | 15/480 [00:56<29:00,  3.74s/it]
  3%|▎         | 16/480 [01:00<29:11,  3.77s/it]
  4%|▎         | 17/480 [01:04<29:24,  3.81s/it]
  4%|▍         | 18/480 [01:08<29:03,  3.77s/it]
  4%|▍         | 19/480 [01:11<28:32,  3.71s/it]
  4%|▍         | 20/480 [01:15<28:54,  3.77s/it]
                                                
{'loss': '0.7093', 'grad_norm': '0.3607', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:15<28:54,  3.77s/it]
  4%|▍         | 21/480 [01:19<28:30,  3.73s/it]
  5%|▍         | 22/480 [01:23<28:06,  3.68s/it]
  5%|▍         | 23/480 [01:26<28:02,  3.68s/it]
  5%|▌         | 24/480 [01:30<28:07,  3.70s/it]
  5%|▌         | 25/480 [01:34<27:46,  3.66s/it]
  5%|▌         | 26/480 [01:37<27:52,  3.68s/it]
  6%|▌         | 27/480 [01:41<27:58,  3.70s/it]
  6%|▌         | 28/480 [01:45<27:52,  3.70s/it]
  6%|▌         | 29/480 [01:49<28:12,  3.75s/it]
  6%|▋         | 30/480 [01:52<27:46,  3.70s/it]
                                                
{'loss': '0.1739', 'grad_norm': '0.1719', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [01:52<27:46,  3.70s/it]
  6%|▋         | 31/480 [01:56<28:05,  3.75s/it]
  7%|▋         | 32/480 [01:57<21:53,  2.93s/it]
  7%|▋         | 33/480 [02:01<23:20,  3.13s/it]
  7%|▋         | 34/480 [02:04<24:40,  3.32s/it]
  7%|▋         | 35/480 [02:08<25:51,  3.49s/it]
  8%|▊         | 36/480 [02:12<26:27,  3.57s/it]
  8%|▊         | 37/480 [02:16<26:50,  3.64s/it]
  8%|▊         | 38/480 [02:20<27:18,  3.71s/it]
  8%|▊         | 39/480 [02:24<27:31,  3.74s/it]
  8%|▊         | 40/480 [02:27<27:29,  3.75s/it]
                                                
{'loss': '0.07992', 'grad_norm': '0.1991', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [02:27<27:29,  3.75s/it]
  9%|▊         | 41/480 [02:31<27:33,  3.77s/it]
  9%|▉         | 42/480 [02:35<27:06,  3.71s/it]
  9%|▉         | 43/480 [02:38<26:47,  3.68s/it]
  9%|▉         | 44/480 [02:42<26:55,  3.70s/it]
  9%|▉         | 45/480 [02:46<26:39,  3.68s/it]
 10%|▉         | 46/480 [02:49<26:47,  3.70s/it]
 10%|▉         | 47/480 [02:53<26:53,  3.73s/it]
 10%|█         | 48/480 [02:57<26:38,  3.70s/it]
 10%|█         | 49/480 [03:01<26:43,  3.72s/it]
 10%|█         | 50/480 [03:04<26:24,  3.68s/it]
                                                
{'loss': '0.06766', 'grad_norm': '0.1903', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [03:04<26:24,  3.68s/it]
 11%|█         | 51/480 [03:08<26:13,  3.67s/it]
 11%|█         | 52/480 [03:12<26:38,  3.73s/it]
 11%|█         | 53/480 [03:15<26:38,  3.74s/it]
 11%|█▏        | 54/480 [03:19<26:24,  3.72s/it]
 11%|█▏        | 55/480 [03:23<26:50,  3.79s/it]
 12%|█▏        | 56/480 [03:27<26:25,  3.74s/it]
 12%|█▏        | 57/480 [03:31<26:25,  3.75s/it]
 12%|█▏        | 58/480 [03:34<26:17,  3.74s/it]
 12%|█▏        | 59/480 [03:38<26:18,  3.75s/it]
 12%|█▎        | 60/480 [03:42<26:25,  3.77s/it]
                                                
{'loss': '0.06506', 'grad_norm': '0.03311', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [03:42<26:25,  3.77s/it]
 13%|█▎        | 61/480 [03:46<26:14,  3.76s/it]
 13%|█▎        | 62/480 [03:49<26:15,  3.77s/it]
 13%|█▎        | 63/480 [03:53<26:30,  3.81s/it]
 13%|█▎        | 64/480 [03:54<20:24,  2.94s/it]
 14%|█▎        | 65/480 [03:58<22:10,  3.21s/it]
 14%|█▍        | 66/480 [04:02<23:40,  3.43s/it]
 14%|█▍        | 67/480 [04:06<24:39,  3.58s/it]
 14%|█▍        | 68/480 [04:09<24:39,  3.59s/it]
 14%|█▍        | 69/480 [04:13<24:40,  3.60s/it]
 15%|█▍        | 70/480 [04:17<24:58,  3.66s/it]
                                                
{'loss': '0.02861', 'grad_norm': '0.01786', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [04:17<24:58,  3.66s/it]
 15%|█▍        | 71/480 [04:21<24:49,  3.64s/it]
 15%|█▌        | 72/480 [04:24<24:46,  3.64s/it]
 15%|█▌        | 73/480 [04:28<25:16,  3.73s/it]
 15%|█▌        | 74/480 [04:32<25:35,  3.78s/it]
 16%|█▌        | 75/480 [04:36<25:55,  3.84s/it]
 16%|█▌        | 76/480 [04:40<25:23,  3.77s/it]
 16%|█▌        | 77/480 [04:43<24:59,  3.72s/it]
 16%|█▋        | 78/480 [04:47<25:15,  3.77s/it]
 16%|█▋        | 79/480 [04:51<25:14,  3.78s/it]
 17%|█▋        | 80/480 [04:54<24:52,  3.73s/it]
                                                
{'loss': '0.0162', 'grad_norm': '0.02142', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [04:54<24:52,  3.73s/it]
 17%|█▋        | 81/480 [04:58<24:48,  3.73s/it]
 17%|█▋        | 82/480 [05:02<24:59,  3.77s/it]
 17%|█▋        | 83/480 [05:06<24:37,  3.72s/it]
 18%|█▊        | 84/480 [05:10<25:17,  3.83s/it]
 18%|█▊        | 85/480 [05:14<25:53,  3.93s/it]
 18%|█▊        | 86/480 [05:18<26:34,  4.05s/it]
 18%|█▊        | 87/480 [05:22<26:21,  4.02s/it]
 18%|█▊        | 88/480 [05:26<25:48,  3.95s/it]
 19%|█▊        | 89/480 [05:30<26:10,  4.02s/it]
 19%|█▉        | 90/480 [05:34<25:55,  3.99s/it]
                                                
{'loss': '0.01722', 'grad_norm': '0.006293', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [05:34<25:55,  3.99s/it]
 19%|█▉        | 91/480 [05:38<25:28,  3.93s/it]
 19%|█▉        | 92/480 [05:42<25:21,  3.92s/it]
 19%|█▉        | 93/480 [05:46<25:01,  3.88s/it]
 20%|█▉        | 94/480 [05:50<25:09,  3.91s/it]
 20%|█▉        | 95/480 [05:53<24:31,  3.82s/it]
 20%|██        | 96/480 [05:54<18:52,  2.95s/it]
 20%|██        | 97/480 [05:58<20:06,  3.15s/it]
 20%|██        | 98/480 [06:01<21:16,  3.34s/it]
 21%|██        | 99/480 [06:05<21:45,  3.43s/it]
 21%|██        | 100/480 [06:09<22:03,  3.48s/it]
                                                 
{'loss': '0.01213', 'grad_norm': '1.301', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [06:09<22:03,  3.48s/it]
 21%|██        | 101/480 [06:13<22:38,  3.59s/it]
 21%|██▏       | 102/480 [06:16<22:45,  3.61s/it]
 21%|██▏       | 103/480 [06:20<23:02,  3.67s/it]
 22%|██▏       | 104/480 [06:24<23:08,  3.69s/it]
 22%|██▏       | 105/480 [06:27<22:55,  3.67s/it]
 22%|██▏       | 106/480 [06:31<22:53,  3.67s/it]
 22%|██▏       | 107/480 [06:35<23:07,  3.72s/it]
 22%|██▎       | 108/480 [06:39<22:51,  3.69s/it]
 23%|██▎       | 109/480 [06:42<22:57,  3.71s/it]
 23%|██▎       | 110/480 [06:46<23:01,  3.73s/it]
                                                 
{'loss': '0.006193', 'grad_norm': '0.006389', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [06:46<23:01,  3.73s/it]
 23%|██▎       | 111/480 [06:50<22:52,  3.72s/it]
 23%|██▎       | 112/480 [06:54<23:02,  3.76s/it]
 24%|██▎       | 113/480 [06:57<22:43,  3.71s/it]
 24%|██▍       | 114/480 [07:01<22:46,  3.73s/it]
 24%|██▍       | 115/480 [07:05<22:41,  3.73s/it]
 24%|██▍       | 116/480 [07:08<22:43,  3.75s/it]
 24%|██▍       | 117/480 [07:12<22:44,  3.76s/it]
 25%|██▍       | 118/480 [07:16<23:13,  3.85s/it]
 25%|██▍       | 119/480 [07:20<23:05,  3.84s/it]
 25%|██▌       | 120/480 [07:24<23:08,  3.86s/it]
                                                 
{'loss': '0.002899', 'grad_norm': '0.001452', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [07:24<23:08,  3.86s/it]
 25%|██▌       | 121/480 [07:28<22:57,  3.84s/it]
 25%|██▌       | 122/480 [07:32<22:47,  3.82s/it]
 26%|██▌       | 123/480 [07:36<22:51,  3.84s/it]
 26%|██▌       | 124/480 [07:39<22:22,  3.77s/it]
 26%|██▌       | 125/480 [07:43<22:01,  3.72s/it]
 26%|██▋       | 126/480 [07:46<22:01,  3.73s/it]
 26%|██▋       | 127/480 [07:50<21:45,  3.70s/it]
 27%|██▋       | 128/480 [07:51<16:50,  2.87s/it]
 27%|██▋       | 129/480 [07:55<18:08,  3.10s/it]
 27%|██▋       | 130/480 [07:59<19:30,  3.34s/it]
                                                 
{'loss': '0.002263', 'grad_norm': '0.0008609', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [07:59<19:30,  3.34s/it]
 27%|██▋       | 131/480 [08:02<19:54,  3.42s/it]
 28%|██▊       | 132/480 [08:06<20:20,  3.51s/it]
 28%|██▊       | 133/480 [08:10<20:36,  3.56s/it]
 28%|██▊       | 134/480 [08:13<20:55,  3.63s/it]
 28%|██▊       | 135/480 [08:17<21:16,  3.70s/it]
 28%|██▊       | 136/480 [08:21<21:03,  3.67s/it]
 29%|██▊       | 137/480 [08:25<21:11,  3.71s/it]
 29%|██▉       | 138/480 [08:28<21:04,  3.70s/it]
 29%|██▉       | 139/480 [08:32<20:52,  3.67s/it]
 29%|██▉       | 140/480 [08:36<21:02,  3.71s/it]
                                                 
{'loss': '0.002072', 'grad_norm': '0.03412', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [08:36<21:02,  3.71s/it]
 29%|██▉       | 141/480 [08:39<21:00,  3.72s/it]
 30%|██▉       | 142/480 [08:43<21:04,  3.74s/it]
 30%|██▉       | 143/480 [08:47<21:04,  3.75s/it]
 30%|███       | 144/480 [08:51<21:20,  3.81s/it]
 30%|███       | 145/480 [08:55<21:05,  3.78s/it]
 30%|███       | 146/480 [08:59<21:26,  3.85s/it]
 31%|███       | 147/480 [09:03<21:17,  3.84s/it]
 31%|███       | 148/480 [09:07<21:37,  3.91s/it]
 31%|███       | 149/480 [09:11<22:04,  4.00s/it]
 31%|███▏      | 150/480 [09:14<21:24,  3.89s/it]
                                                 
{'loss': '0.004371', 'grad_norm': '0.009656', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [09:14<21:24,  3.89s/it]
 31%|███▏      | 151/480 [09:18<20:57,  3.82s/it]
 32%|███▏      | 152/480 [09:22<21:37,  3.96s/it]
 32%|███▏      | 153/480 [09:26<21:01,  3.86s/it]
 32%|███▏      | 154/480 [09:30<20:35,  3.79s/it]
 32%|███▏      | 155/480 [09:33<20:35,  3.80s/it]
 32%|███▎      | 156/480 [09:37<20:13,  3.75s/it]
 33%|███▎      | 157/480 [09:41<20:28,  3.80s/it]
 33%|███▎      | 158/480 [09:45<20:15,  3.78s/it]
 33%|███▎      | 159/480 [09:48<19:55,  3.73s/it]
 33%|███▎      | 160/480 [09:49<15:22,  2.88s/it]
                                                 
{'loss': '0.001157', 'grad_norm': '0.003031', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [09:49<15:22,  2.88s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<00:19,  4.95it/s]
+
Eval @ step 160:   2%|▏         | 2/100 [00:00<00:17,  5.68it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:00<00:16,  5.96it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:00<00:15,  6.10it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:00<00:15,  6.18it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:00<00:15,  6.23it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:01<00:14,  6.27it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:01<00:15,  5.91it/s]
+
Eval @ step 160:   9%|▉         | 9/100 [00:01<00:16,  5.39it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:01<00:17,  5.08it/s]
+
Eval @ step 160:  11%|█         | 11/100 [00:01<00:16,  5.32it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:02<00:17,  5.00it/s]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:02<00:18,  4.81it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:02<00:16,  5.19it/s]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:02<00:15,  5.48it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:02<00:15,  5.45it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:03<00:14,  5.69it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:03<00:14,  5.57it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:03<00:14,  5.78it/s]
+
Eval @ step 160:  20%|██        | 20/100 [00:03<00:13,  5.93it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:03<00:13,  6.01it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:03<00:13,  5.85it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:04<00:19,  3.91it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:04<00:17,  4.41it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:04<00:15,  4.85it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:04<00:14,  4.96it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:05<00:13,  5.34it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:05<00:12,  5.66it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:05<00:12,  5.60it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:05<00:11,  5.86it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:05<00:11,  6.05it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:05<00:10,  6.56it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:06<00:11,  5.86it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:06<00:10,  6.05it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:06<00:10,  6.19it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:06<00:10,  6.29it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:06<00:09,  6.36it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:06<00:09,  6.42it/s]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:06<00:09,  6.45it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:07<00:09,  6.12it/s]
+
Eval @ step 160:  41%|████      | 41/100 [00:07<00:09,  6.25it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:07<00:10,  5.68it/s]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:07<00:09,  5.92it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:07<00:09,  6.09it/s]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:07<00:08,  6.22it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:08<00:08,  6.30it/s]
+
Eval @ step 160:  47%|████▋     | 47/100 [00:08<00:09,  5.71it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:08<00:09,  5.36it/s]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:08<00:09,  5.65it/s]
+
Eval @ step 160:  50%|█████     | 50/100 [00:08<00:08,  5.85it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:08<00:08,  6.01it/s]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:09<00:07,  6.14it/s]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:09<00:07,  6.24it/s]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:09<00:07,  6.31it/s]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:09<00:07,  6.03it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:09<00:07,  5.55it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:10<00:08,  5.25it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:10<00:08,  5.06it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:10<00:07,  5.18it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:10<00:07,  5.26it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:10<00:07,  5.07it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [00:11<00:07,  4.95it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [00:11<00:07,  4.87it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [00:11<00:07,  5.03it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:11<00:06,  5.41it/s]
+
Eval @ step 160:  66%|██████▌   | 66/100 [00:11<00:05,  5.71it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [00:11<00:05,  5.94it/s]
+
Eval @ step 160:  68%|██████▊   | 68/100 [00:12<00:05,  5.78it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:12<00:05,  5.99it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [00:12<00:04,  6.14it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:12<00:04,  5.86it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [00:12<00:04,  5.94it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:12<00:04,  5.41it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [00:13<00:05,  4.93it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:13<00:04,  5.33it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:13<00:04,  5.36it/s]
+
Eval @ step 160:  77%|███████▋  | 77/100 [00:13<00:04,  5.39it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:13<00:04,  5.42it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [00:14<00:03,  5.71it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:14<00:03,  5.94it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:14<00:03,  6.10it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [00:14<00:02,  6.23it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [00:14<00:02,  5.98it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [00:14<00:02,  5.82it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:15<00:02,  5.70it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [00:15<00:02,  5.35it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [00:15<00:02,  5.66it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [00:15<00:02,  5.90it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [00:15<00:01,  6.08it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [00:15<00:01,  6.21it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [00:16<00:01,  6.31it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [00:16<00:01,  6.38it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:16<00:01,  6.06it/s]
+
Eval @ step 160:  94%|█████████▍| 94/100 [00:16<00:01,  5.87it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:16<00:00,  6.06it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [00:16<00:00,  6.19it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:17<00:00,  6.29it/s]
+
Eval @ step 160:  98%|█████████▊| 98/100 [00:17<00:00,  6.36it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [00:17<00:00,  6.42it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [00:17<00:00,  6.45it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=99.1333 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '99.13', 'epoch': '5'}
+
 33%|███▎      | 160/480 [10:07<15:22,  2.88s/it]
 34%|███▎      | 161/480 [10:11<44:40,  8.40s/it]
 34%|███▍      | 162/480 [10:15<37:56,  7.16s/it]
 34%|███▍      | 163/480 [10:19<32:54,  6.23s/it]
 34%|███▍      | 164/480 [10:23<29:01,  5.51s/it]
 34%|███▍      | 165/480 [10:27<26:16,  5.01s/it]
 35%|███▍      | 166/480 [10:30<23:59,  4.58s/it]
 35%|███▍      | 167/480 [10:34<22:40,  4.35s/it]
 35%|███▌      | 168/480 [10:38<21:28,  4.13s/it]
 35%|███▌      | 169/480 [10:41<20:35,  3.97s/it]
 35%|███▌      | 170/480 [10:45<20:04,  3.88s/it]
                                                 
{'loss': '0.002523', 'grad_norm': '0.00247', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [10:45<20:04,  3.88s/it]
 36%|███▌      | 171/480 [10:48<19:35,  3.80s/it]
 36%|███▌      | 172/480 [10:52<19:29,  3.80s/it]
 36%|███▌      | 173/480 [10:56<19:27,  3.80s/it]
 36%|███▋      | 174/480 [11:00<19:07,  3.75s/it]
 36%|███▋      | 175/480 [11:03<18:57,  3.73s/it]
 37%|███▋      | 176/480 [11:07<18:48,  3.71s/it]
 37%|███▋      | 177/480 [11:11<18:50,  3.73s/it]
 37%|███▋      | 178/480 [11:14<18:34,  3.69s/it]
 37%|███▋      | 179/480 [11:18<18:29,  3.69s/it]
 38%|███▊      | 180/480 [11:22<18:36,  3.72s/it]
                                                 
{'loss': '0.001929', 'grad_norm': '0.0005607', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [11:22<18:36,  3.72s/it]
 38%|███▊      | 181/480 [11:26<18:47,  3.77s/it]
 38%|███▊      | 182/480 [11:30<18:58,  3.82s/it]
 38%|███▊      | 183/480 [11:33<18:35,  3.76s/it]
 38%|███▊      | 184/480 [11:37<18:43,  3.80s/it]
 39%|███▊      | 185/480 [11:41<18:38,  3.79s/it]
 39%|███▉      | 186/480 [11:45<18:22,  3.75s/it]
 39%|███▉      | 187/480 [11:49<18:35,  3.81s/it]
 39%|███▉      | 188/480 [11:52<18:35,  3.82s/it]
 39%|███▉      | 189/480 [11:56<18:30,  3.81s/it]
 40%|███▉      | 190/480 [12:00<18:43,  3.87s/it]
                                                 
{'loss': '0.0005096', 'grad_norm': '0.007904', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [12:00<18:43,  3.87s/it]
 40%|███▉      | 191/480 [12:04<18:48,  3.90s/it]
 40%|████      | 192/480 [12:05<14:25,  3.01s/it]
 40%|████      | 193/480 [12:09<15:19,  3.20s/it]
 40%|████      | 194/480 [12:12<15:50,  3.33s/it]
 41%|████      | 195/480 [12:16<16:12,  3.41s/it]
 41%|████      | 196/480 [12:20<16:39,  3.52s/it]
 41%|████      | 197/480 [12:23<16:43,  3.55s/it]
 41%|████▏     | 198/480 [12:27<16:46,  3.57s/it]
 41%|████▏     | 199/480 [12:31<17:07,  3.66s/it]
 42%|████▏     | 200/480 [12:35<17:14,  3.69s/it]
                                                 
{'loss': '0.0004088', 'grad_norm': '0.004252', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [12:35<17:14,  3.69s/it]
 42%|████▏     | 201/480 [12:38<17:03,  3.67s/it]
 42%|████▏     | 202/480 [12:42<17:09,  3.70s/it]
 42%|████▏     | 203/480 [12:46<16:57,  3.67s/it]
 42%|████▎     | 204/480 [12:49<17:05,  3.72s/it]
 43%|████▎     | 205/480 [12:53<17:09,  3.74s/it]
 43%|████▎     | 206/480 [12:57<16:57,  3.71s/it]
 43%|████▎     | 207/480 [13:01<17:09,  3.77s/it]
 43%|████▎     | 208/480 [13:05<17:01,  3.75s/it]
 44%|████▎     | 209/480 [13:08<16:59,  3.76s/it]
 44%|████▍     | 210/480 [13:12<17:00,  3.78s/it]
                                                 
{'loss': '0.0005122', 'grad_norm': '0.0004133', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [13:12<17:00,  3.78s/it]
 44%|████▍     | 211/480 [13:16<16:57,  3.78s/it]
 44%|████▍     | 212/480 [13:20<16:42,  3.74s/it]
 44%|████▍     | 213/480 [13:23<16:55,  3.80s/it]
 45%|████▍     | 214/480 [13:27<16:50,  3.80s/it]
 45%|████▍     | 215/480 [13:31<16:45,  3.80s/it]
 45%|████▌     | 216/480 [13:35<16:49,  3.83s/it]
 45%|████▌     | 217/480 [13:39<16:43,  3.81s/it]
 45%|████▌     | 218/480 [13:43<16:37,  3.81s/it]
 46%|████▌     | 219/480 [13:46<16:42,  3.84s/it]
 46%|████▌     | 220/480 [13:50<16:20,  3.77s/it]
                                                 
{'loss': '0.0002486', 'grad_norm': '0.0008414', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [13:50<16:20,  3.77s/it]
 46%|████▌     | 221/480 [13:54<16:07,  3.73s/it]
 46%|████▋     | 222/480 [13:58<16:27,  3.83s/it]
 46%|████▋     | 223/480 [14:01<16:08,  3.77s/it]
 47%|████▋     | 224/480 [14:02<12:26,  2.91s/it]
 47%|████▋     | 225/480 [14:06<13:27,  3.17s/it]
 47%|████▋     | 226/480 [14:10<13:58,  3.30s/it]
 47%|████▋     | 227/480 [14:13<14:33,  3.45s/it]
 48%|████▊     | 228/480 [14:17<15:04,  3.59s/it]
 48%|████▊     | 229/480 [14:21<15:19,  3.66s/it]
 48%|████▊     | 230/480 [14:25<15:22,  3.69s/it]
                                                 
{'loss': '0.0002565', 'grad_norm': '0.0005208', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [14:25<15:22,  3.69s/it]
 48%|████▊     | 231/480 [14:29<15:27,  3.73s/it]
 48%|████▊     | 232/480 [14:33<15:28,  3.74s/it]
 49%|████▊     | 233/480 [14:36<15:20,  3.73s/it]
 49%|████▉     | 234/480 [14:40<15:11,  3.70s/it]
 49%|████▉     | 235/480 [14:44<15:14,  3.73s/it]
 49%|████▉     | 236/480 [14:47<15:11,  3.74s/it]
 49%|████▉     | 237/480 [14:51<15:16,  3.77s/it]
 50%|████▉     | 238/480 [14:55<15:01,  3.72s/it]
 50%|████▉     | 239/480 [14:59<15:24,  3.84s/it]
 50%|█████     | 240/480 [15:03<15:20,  3.84s/it]
                                                 
{'loss': '0.000352', 'grad_norm': '0.000292', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [15:03<15:20,  3.84s/it]
 50%|█████     | 241/480 [15:07<15:05,  3.79s/it]
 50%|█████     | 242/480 [15:10<15:11,  3.83s/it]
 51%|█████     | 243/480 [15:14<15:05,  3.82s/it]
 51%|█████     | 244/480 [15:18<15:01,  3.82s/it]
 51%|█████     | 245/480 [15:22<15:30,  3.96s/it]
 51%|█████▏    | 246/480 [15:26<15:16,  3.92s/it]
 51%|█████▏    | 247/480 [15:30<15:11,  3.91s/it]
 52%|█████▏    | 248/480 [15:34<14:46,  3.82s/it]
 52%|█████▏    | 249/480 [15:37<14:28,  3.76s/it]
 52%|█████▏    | 250/480 [15:41<14:29,  3.78s/it]
                                                 
{'loss': '0.000247', 'grad_norm': '0.0007481', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [15:41<14:29,  3.78s/it]
 52%|█████▏    | 251/480 [15:45<14:14,  3.73s/it]
 52%|█████▎    | 252/480 [15:48<14:02,  3.70s/it]
 53%|█████▎    | 253/480 [15:52<14:01,  3.71s/it]
 53%|█████▎    | 254/480 [15:56<14:02,  3.73s/it]
 53%|█████▎    | 255/480 [16:00<13:51,  3.70s/it]
 53%|█████▎    | 256/480 [16:00<10:43,  2.87s/it]
 54%|█████▎    | 257/480 [16:04<11:49,  3.18s/it]
 54%|█████▍    | 258/480 [16:08<12:15,  3.31s/it]
 54%|█████▍    | 259/480 [16:12<12:48,  3.48s/it]
 54%|█████▍    | 260/480 [16:15<12:54,  3.52s/it]
                                                 
{'loss': '0.0001672', 'grad_norm': '0.000464', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [16:15<12:54,  3.52s/it]
 54%|█████▍    | 261/480 [16:19<12:58,  3.55s/it]
 55%|█████▍    | 262/480 [16:23<13:18,  3.66s/it]
 55%|█████▍    | 263/480 [16:27<13:12,  3.65s/it]
 55%|█████▌    | 264/480 [16:30<13:06,  3.64s/it]
 55%|█████▌    | 265/480 [16:34<13:24,  3.74s/it]
 55%|█████▌    | 266/480 [16:38<13:23,  3.75s/it]
 56%|█████▌    | 267/480 [16:42<13:11,  3.72s/it]
 56%|█████▌    | 268/480 [16:46<13:20,  3.78s/it]
 56%|█████▌    | 269/480 [16:49<13:17,  3.78s/it]
 56%|█████▋    | 270/480 [16:53<13:18,  3.80s/it]
                                                 
{'loss': '0.0002188', 'grad_norm': '0.0007793', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [16:53<13:18,  3.80s/it]
 56%|█████▋    | 271/480 [16:57<13:39,  3.92s/it]
 57%|█████▋    | 272/480 [17:01<13:33,  3.91s/it]
 57%|█████▋    | 273/480 [17:06<13:55,  4.03s/it]
 57%|█████▋    | 274/480 [17:09<13:37,  3.97s/it]
 57%|█████▋    | 275/480 [17:13<13:13,  3.87s/it]
 57%|█████▊    | 276/480 [17:17<13:01,  3.83s/it]
 58%|█████▊    | 277/480 [17:20<12:45,  3.77s/it]
 58%|█████▊    | 278/480 [17:24<12:32,  3.73s/it]
 58%|█████▊    | 279/480 [17:28<12:35,  3.76s/it]
 58%|█████▊    | 280/480 [17:31<12:22,  3.71s/it]
                                                 
{'loss': '0.000517', 'grad_norm': '0.002678', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [17:31<12:22,  3.71s/it]
 59%|█████▊    | 281/480 [17:35<12:29,  3.77s/it]
 59%|█████▉    | 282/480 [17:39<12:36,  3.82s/it]
 59%|█████▉    | 283/480 [17:43<12:20,  3.76s/it]
 59%|█████▉    | 284/480 [17:47<12:18,  3.77s/it]
 59%|█████▉    | 285/480 [17:50<12:05,  3.72s/it]
 60%|█████▉    | 286/480 [17:54<12:15,  3.79s/it]
 60%|█████▉    | 287/480 [17:58<12:20,  3.84s/it]
 60%|██████    | 288/480 [17:59<09:28,  2.96s/it]
 60%|██████    | 289/480 [18:03<10:03,  3.16s/it]
 60%|██████    | 290/480 [18:07<10:41,  3.38s/it]
                                                 
{'loss': '0.0001625', 'grad_norm': '0.002777', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [18:07<10:41,  3.38s/it]
 61%|██████    | 291/480 [18:10<11:02,  3.50s/it]
 61%|██████    | 292/480 [18:14<11:14,  3.59s/it]
 61%|██████    | 293/480 [18:18<11:22,  3.65s/it]
 61%|██████▏   | 294/480 [18:22<11:28,  3.70s/it]
 61%|██████▏   | 295/480 [18:26<11:21,  3.68s/it]
 62%|██████▏   | 296/480 [18:29<11:28,  3.74s/it]
 62%|██████▏   | 297/480 [18:33<11:27,  3.76s/it]
 62%|██████▏   | 298/480 [18:37<11:45,  3.87s/it]
 62%|██████▏   | 299/480 [18:41<11:42,  3.88s/it]
 62%|██████▎   | 300/480 [18:45<11:35,  3.86s/it]
                                                 
{'loss': '0.0002096', 'grad_norm': '0.0002562', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [18:45<11:35,  3.86s/it]
 63%|██████▎   | 301/480 [18:49<11:27,  3.84s/it]
 63%|██████▎   | 302/480 [18:53<11:30,  3.88s/it]
 63%|██████▎   | 303/480 [18:56<11:12,  3.80s/it]
 63%|██████▎   | 304/480 [19:00<11:03,  3.77s/it]
 64%|██████▎   | 305/480 [19:04<10:52,  3.73s/it]
 64%|██████▍   | 306/480 [19:07<10:47,  3.72s/it]
 64%|██████▍   | 307/480 [19:11<10:46,  3.74s/it]
 64%|██████▍   | 308/480 [19:15<10:36,  3.70s/it]
 64%|██████▍   | 309/480 [19:19<10:37,  3.73s/it]
 65%|██████▍   | 310/480 [19:22<10:34,  3.73s/it]
                                                 
{'loss': '0.0001513', 'grad_norm': '0.001094', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [19:22<10:34,  3.73s/it]
 65%|██████▍   | 311/480 [19:26<10:27,  3.71s/it]
 65%|██████▌   | 312/480 [19:30<10:18,  3.68s/it]
 65%|██████▌   | 313/480 [19:33<10:14,  3.68s/it]
 65%|██████▌   | 314/480 [19:37<10:16,  3.71s/it]
 66%|██████▌   | 315/480 [19:41<10:11,  3.71s/it]
 66%|██████▌   | 316/480 [19:45<10:08,  3.71s/it]
 66%|██████▌   | 317/480 [19:48<10:08,  3.73s/it]
 66%|██████▋   | 318/480 [19:52<09:59,  3.70s/it]
 66%|██████▋   | 319/480 [19:56<10:02,  3.74s/it]
 67%|██████▋   | 320/480 [19:57<07:43,  2.89s/it]
                                                 
{'loss': '0.0003888', 'grad_norm': '0.001183', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [19:57<07:43,  2.89s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:00<00:19,  5.15it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:16,  5.79it/s]
+
Eval @ step 320:   3%|▎         | 3/100 [00:00<00:16,  6.02it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:00<00:15,  6.14it/s]
+
Eval @ step 320:   5%|▌         | 5/100 [00:00<00:15,  6.20it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:00<00:15,  6.24it/s]
+
Eval @ step 320:   7%|▋         | 7/100 [00:01<00:14,  6.27it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:01<00:15,  5.92it/s]
+
Eval @ step 320:   9%|▉         | 9/100 [00:01<00:16,  5.40it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:01<00:17,  5.08it/s]
+
Eval @ step 320:  11%|█         | 11/100 [00:01<00:16,  5.41it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:02<00:17,  5.10it/s]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:02<00:16,  5.16it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:02<00:15,  5.46it/s]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:02<00:14,  5.70it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:02<00:14,  5.87it/s]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:02<00:13,  6.01it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:03<00:14,  5.77it/s]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:03<00:13,  5.93it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:03<00:13,  6.05it/s]
+
Eval @ step 320:  21%|██        | 21/100 [00:03<00:12,  6.13it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:03<00:12,  6.18it/s]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:03<00:12,  6.22it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:04<00:12,  6.25it/s]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:04<00:11,  6.27it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:04<00:12,  5.92it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:04<00:12,  6.04it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:04<00:11,  6.12it/s]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:04<00:12,  5.84it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:05<00:11,  5.98it/s]
+
Eval @ step 320:  31%|███       | 31/100 [00:05<00:11,  6.08it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:05<00:10,  6.53it/s]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:05<00:11,  5.77it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:05<00:11,  5.93it/s]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:05<00:10,  6.04it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:06<00:10,  6.12it/s]
+
Eval @ step 320:  37%|███▋      | 37/100 [00:06<00:10,  6.17it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:06<00:09,  6.21it/s]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:06<00:09,  6.20it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:06<00:10,  5.82it/s]
+
Eval @ step 320:  41%|████      | 41/100 [00:06<00:10,  5.50it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:07<00:11,  5.23it/s]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:07<00:10,  5.56it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:07<00:09,  5.83it/s]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:07<00:09,  6.02it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:07<00:08,  6.18it/s]
+
Eval @ step 320:  47%|████▋     | 47/100 [00:08<00:09,  5.63it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:08<00:09,  5.28it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:08<00:09,  5.60it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:08<00:08,  5.85it/s]
+
Eval @ step 320:  51%|█████     | 51/100 [00:08<00:08,  6.03it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:08<00:07,  6.18it/s]
+
Eval @ step 320:  53%|█████▎    | 53/100 [00:09<00:07,  6.27it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:09<00:07,  6.34it/s]
+
Eval @ step 320:  55%|█████▌    | 55/100 [00:09<00:07,  6.03it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:09<00:07,  5.54it/s]
+
Eval @ step 320:  57%|█████▋    | 57/100 [00:09<00:08,  5.20it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:09<00:08,  5.01it/s]
+
Eval @ step 320:  59%|█████▉    | 59/100 [00:10<00:08,  5.12it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:10<00:07,  5.21it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:10<00:07,  5.04it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:10<00:07,  4.93it/s]
+
Eval @ step 320:  63%|██████▎   | 63/100 [00:11<00:07,  4.85it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:11<00:07,  4.88it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [00:11<00:06,  5.20it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:11<00:06,  5.54it/s]
+
Eval @ step 320:  67%|██████▋   | 67/100 [00:11<00:05,  5.81it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:11<00:05,  5.69it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [00:12<00:05,  5.92it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:12<00:05,  5.99it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [00:12<00:05,  5.78it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:12<00:04,  5.99it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [00:12<00:04,  5.51it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:12<00:05,  5.00it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [00:13<00:04,  5.38it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:13<00:04,  5.40it/s]
+
Eval @ step 320:  77%|███████▋  | 77/100 [00:13<00:04,  5.42it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:13<00:04,  5.43it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [00:13<00:03,  5.72it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:13<00:03,  5.95it/s]
+
Eval @ step 320:  81%|████████  | 81/100 [00:14<00:03,  6.11it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:14<00:02,  6.23it/s]
+
Eval @ step 320:  83%|████████▎ | 83/100 [00:14<00:02,  5.98it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:14<00:02,  5.82it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [00:14<00:02,  5.70it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:15<00:02,  5.35it/s]
+
Eval @ step 320:  87%|████████▋ | 87/100 [00:15<00:02,  5.66it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:15<00:02,  5.89it/s]
+
Eval @ step 320:  89%|████████▉ | 89/100 [00:15<00:01,  6.07it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:15<00:01,  6.21it/s]
+
Eval @ step 320:  91%|█████████ | 91/100 [00:15<00:01,  6.31it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:15<00:01,  6.38it/s]
+
Eval @ step 320:  93%|█████████▎| 93/100 [00:16<00:01,  6.07it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:16<00:01,  5.87it/s]
+
Eval @ step 320:  95%|█████████▌| 95/100 [00:16<00:00,  6.05it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:16<00:00,  6.19it/s]
+
Eval @ step 320:  97%|█████████▋| 97/100 [00:16<00:00,  5.84it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:16<00:00,  6.02it/s]
+
Eval @ step 320:  99%|█████████▉| 99/100 [00:17<00:00,  6.11it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:17<00:00,  6.13it/s]
+
                                                                    [EVAL] eval_rougeL=99.1333 (best=99.1333)
+
                                                 
{'eval_rougeL': '99.13', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [20:14<07:43,  2.89s/it]
 67%|██████▋   | 321/480 [20:18<22:11,  8.38s/it]
 67%|██████▋   | 322/480 [20:22<18:26,  7.00s/it]
 67%|██████▋   | 323/480 [20:26<16:05,  6.15s/it]
 68%|██████▊   | 324/480 [20:29<14:00,  5.39s/it]
 68%|██████▊   | 325/480 [20:33<12:32,  4.86s/it]
 68%|██████▊   | 326/480 [20:37<11:33,  4.51s/it]
 68%|██████▊   | 327/480 [20:40<10:48,  4.24s/it]
 68%|██████▊   | 328/480 [20:44<10:15,  4.05s/it]
 69%|██████▊   | 329/480 [20:48<09:56,  3.95s/it]
 69%|██████▉   | 330/480 [20:51<09:45,  3.90s/it]
                                                 
{'loss': '0.0004011', 'grad_norm': '0.0001113', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [20:51<09:45,  3.90s/it]
 69%|██████▉   | 331/480 [20:55<09:28,  3.82s/it]
 69%|██████▉   | 332/480 [20:59<09:20,  3.79s/it]
 69%|██████▉   | 333/480 [21:02<09:09,  3.74s/it]
 70%|██████▉   | 334/480 [21:07<09:21,  3.85s/it]
 70%|██████▉   | 335/480 [21:11<09:23,  3.89s/it]
 70%|███████   | 336/480 [21:14<09:09,  3.81s/it]
 70%|███████   | 337/480 [21:18<09:05,  3.82s/it]
 70%|███████   | 338/480 [21:22<09:00,  3.81s/it]
 71%|███████   | 339/480 [21:25<08:48,  3.75s/it]
 71%|███████   | 340/480 [21:29<08:54,  3.82s/it]
                                                 
{'loss': '0.000145', 'grad_norm': '0.001131', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [21:29<08:54,  3.82s/it]
 71%|███████   | 341/480 [21:33<08:42,  3.76s/it]
 71%|███████▏  | 342/480 [21:37<08:41,  3.78s/it]
 71%|███████▏  | 343/480 [21:41<08:46,  3.84s/it]
 72%|███████▏  | 344/480 [21:45<08:37,  3.80s/it]
 72%|███████▏  | 345/480 [21:48<08:33,  3.80s/it]
 72%|███████▏  | 346/480 [21:52<08:33,  3.83s/it]
 72%|███████▏  | 347/480 [21:56<08:30,  3.84s/it]
 72%|███████▎  | 348/480 [22:00<08:29,  3.86s/it]
 73%|███████▎  | 349/480 [22:04<08:24,  3.85s/it]
 73%|███████▎  | 350/480 [22:08<08:30,  3.93s/it]
                                                 
{'loss': '0.0001892', 'grad_norm': '0.0006113', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [22:08<08:30,  3.93s/it]
 73%|███████▎  | 351/480 [22:12<08:32,  3.98s/it]
 73%|███████▎  | 352/480 [22:13<06:32,  3.07s/it]
 74%|███████▎  | 353/480 [22:17<07:02,  3.33s/it]
 74%|███████▍  | 354/480 [22:21<07:16,  3.47s/it]
 74%|███████▍  | 355/480 [22:24<07:21,  3.53s/it]
 74%|███████▍  | 356/480 [22:28<07:20,  3.56s/it]
 74%|███████▍  | 357/480 [22:32<07:23,  3.60s/it]
 75%|███████▍  | 358/480 [22:35<07:26,  3.66s/it]
 75%|███████▍  | 359/480 [22:39<07:21,  3.65s/it]
 75%|███████▌  | 360/480 [22:43<07:23,  3.70s/it]
                                                 
{'loss': '0.0001029', 'grad_norm': '0.0008723', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [22:43<07:23,  3.70s/it]
 75%|███████▌  | 361/480 [22:47<07:17,  3.68s/it]
 75%|███████▌  | 362/480 [22:50<07:17,  3.71s/it]
 76%|███████▌  | 363/480 [22:54<07:14,  3.71s/it]
 76%|███████▌  | 364/480 [22:58<07:07,  3.68s/it]
 76%|███████▌  | 365/480 [23:02<07:09,  3.73s/it]
 76%|███████▋  | 366/480 [23:05<07:09,  3.76s/it]
 76%|███████▋  | 367/480 [23:09<07:15,  3.86s/it]
 77%|███████▋  | 368/480 [23:13<07:10,  3.84s/it]
 77%|███████▋  | 369/480 [23:17<07:11,  3.89s/it]
 77%|███████▋  | 370/480 [23:21<06:59,  3.81s/it]
                                                 
{'loss': '0.0002423', 'grad_norm': '0.0005574', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [23:21<06:59,  3.81s/it]
 77%|███████▋  | 371/480 [23:25<07:06,  3.91s/it]
 78%|███████▊  | 372/480 [23:29<06:59,  3.89s/it]
 78%|███████▊  | 373/480 [23:33<06:52,  3.86s/it]
 78%|███████▊  | 374/480 [23:37<06:50,  3.88s/it]
 78%|███████▊  | 375/480 [23:40<06:44,  3.85s/it]
 78%|███████▊  | 376/480 [23:44<06:38,  3.83s/it]
 79%|███████▊  | 377/480 [23:48<06:37,  3.86s/it]
 79%|███████▉  | 378/480 [23:52<06:26,  3.79s/it]
 79%|███████▉  | 379/480 [23:55<06:17,  3.74s/it]
 79%|███████▉  | 380/480 [23:59<06:12,  3.73s/it]
                                                 
{'loss': '0.0001614', 'grad_norm': '0.0002046', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [23:59<06:12,  3.73s/it]
 79%|███████▉  | 381/480 [24:03<06:05,  3.69s/it]
 80%|███████▉  | 382/480 [24:06<06:04,  3.72s/it]
 80%|███████▉  | 383/480 [24:10<06:11,  3.83s/it]
 80%|████████  | 384/480 [24:11<04:44,  2.96s/it]
 80%|████████  | 385/480 [24:15<04:59,  3.16s/it]
 80%|████████  | 386/480 [24:19<05:19,  3.40s/it]
 81%|████████  | 387/480 [24:23<05:22,  3.47s/it]
 81%|████████  | 388/480 [24:26<05:24,  3.52s/it]
 81%|████████  | 389/480 [24:30<05:28,  3.61s/it]
 81%|████████▏ | 390/480 [24:34<05:29,  3.66s/it]
                                                 
{'loss': '0.0001402', 'grad_norm': '0.0001314', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [24:34<05:29,  3.66s/it]
 81%|████████▏ | 391/480 [24:37<05:24,  3.65s/it]
 82%|████████▏ | 392/480 [24:41<05:25,  3.69s/it]
 82%|████████▏ | 393/480 [24:45<05:24,  3.72s/it]
 82%|████████▏ | 394/480 [24:49<05:18,  3.71s/it]
 82%|████████▏ | 395/480 [24:53<05:17,  3.73s/it]
 82%|████████▎ | 396/480 [24:56<05:10,  3.70s/it]
 83%|████████▎ | 397/480 [25:00<05:11,  3.75s/it]
 83%|████████▎ | 398/480 [25:04<05:04,  3.71s/it]
 83%|████████▎ | 399/480 [25:07<05:02,  3.73s/it]
 83%|████████▎ | 400/480 [25:11<05:03,  3.79s/it]
                                                 
{'loss': '0.0001547', 'grad_norm': '0.0006027', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [25:11<05:03,  3.79s/it]
 84%|████████▎ | 401/480 [25:15<05:00,  3.80s/it]
 84%|████████▍ | 402/480 [25:19<04:56,  3.80s/it]
 84%|████████▍ | 403/480 [25:23<04:56,  3.85s/it]
 84%|████████▍ | 404/480 [25:27<04:52,  3.85s/it]
 84%|████████▍ | 405/480 [25:31<04:59,  4.00s/it]
 85%|████████▍ | 406/480 [25:35<04:47,  3.88s/it]
 85%|████████▍ | 407/480 [25:38<04:37,  3.80s/it]
 85%|████████▌ | 408/480 [25:42<04:32,  3.79s/it]
 85%|████████▌ | 409/480 [25:46<04:25,  3.74s/it]
 85%|████████▌ | 410/480 [25:49<04:19,  3.70s/it]
                                                 
{'loss': '9.495e-05', 'grad_norm': '0.0002309', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [25:49<04:19,  3.70s/it]
 86%|████████▌ | 411/480 [25:53<04:16,  3.72s/it]
 86%|████████▌ | 412/480 [25:57<04:18,  3.80s/it]
 86%|████████▌ | 413/480 [26:01<04:13,  3.78s/it]
 86%|████████▋ | 414/480 [26:05<04:09,  3.79s/it]
 86%|████████▋ | 415/480 [26:09<04:08,  3.82s/it]
 87%|████████▋ | 416/480 [26:09<03:09,  2.96s/it]
 87%|████████▋ | 417/480 [26:13<03:24,  3.25s/it]
 87%|████████▋ | 418/480 [26:17<03:28,  3.36s/it]
 87%|████████▋ | 419/480 [26:21<03:32,  3.48s/it]
 88%|████████▊ | 420/480 [26:25<03:42,  3.70s/it]
                                                 
{'loss': '0.0001774', 'grad_norm': '0.0005778', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [26:25<03:42,  3.70s/it]
 88%|████████▊ | 421/480 [26:29<03:36,  3.68s/it]
 88%|████████▊ | 422/480 [26:32<03:32,  3.66s/it]
 88%|████████▊ | 423/480 [26:36<03:32,  3.73s/it]
 88%|████████▊ | 424/480 [26:40<03:26,  3.70s/it]
 89%|████████▊ | 425/480 [26:44<03:24,  3.72s/it]
 89%|████████▉ | 426/480 [26:47<03:24,  3.79s/it]
 89%|████████▉ | 427/480 [26:51<03:20,  3.79s/it]
 89%|████████▉ | 428/480 [26:55<03:18,  3.81s/it]
 89%|████████▉ | 429/480 [26:59<03:14,  3.80s/it]
 90%|████████▉ | 430/480 [27:03<03:14,  3.90s/it]
                                                 
{'loss': '0.000115', 'grad_norm': '0.0004607', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [27:03<03:14,  3.90s/it]
 90%|████████▉ | 431/480 [27:07<03:14,  3.96s/it]
 90%|█████████ | 432/480 [27:11<03:08,  3.92s/it]
 90%|█████████ | 433/480 [27:15<03:00,  3.84s/it]
 90%|█████████ | 434/480 [27:18<02:56,  3.83s/it]
 91%|█████████ | 435/480 [27:22<02:49,  3.77s/it]
 91%|█████████ | 436/480 [27:26<02:44,  3.73s/it]
 91%|█████████ | 437/480 [27:30<02:42,  3.77s/it]
 91%|█████████▏| 438/480 [27:33<02:36,  3.73s/it]
 91%|█████████▏| 439/480 [27:37<02:32,  3.72s/it]
 92%|█████████▏| 440/480 [27:41<02:30,  3.76s/it]
                                                 
{'loss': '0.0006975', 'grad_norm': '9.197e-05', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [27:41<02:30,  3.76s/it]
 92%|█████████▏| 441/480 [27:44<02:24,  3.71s/it]
 92%|█████████▏| 442/480 [27:48<02:21,  3.73s/it]
 92%|█████████▏| 443/480 [27:52<02:16,  3.70s/it]
 92%|█████████▎| 444/480 [27:56<02:14,  3.72s/it]
 93%|█████████▎| 445/480 [27:59<02:09,  3.71s/it]
 93%|█████████▎| 446/480 [28:03<02:05,  3.68s/it]
 93%|█████████▎| 447/480 [28:07<02:02,  3.71s/it]
 93%|█████████▎| 448/480 [28:08<01:31,  2.87s/it]
 94%|█████████▎| 449/480 [28:11<01:38,  3.18s/it]
 94%|█████████▍| 450/480 [28:15<01:41,  3.38s/it]
                                                 
{'loss': '8.958e-05', 'grad_norm': '0.004529', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [28:15<01:41,  3.38s/it]
 94%|█████████▍| 451/480 [28:19<01:40,  3.48s/it]
 94%|█████████▍| 452/480 [28:23<01:40,  3.58s/it]
 94%|█████████▍| 453/480 [28:27<01:39,  3.69s/it]
 95%|█████████▍| 454/480 [28:31<01:37,  3.76s/it]
 95%|█████████▍| 455/480 [28:35<01:36,  3.86s/it]
 95%|█████████▌| 456/480 [28:39<01:34,  3.95s/it]
 95%|█████████▌| 457/480 [28:43<01:30,  3.92s/it]
 95%|█████████▌| 458/480 [28:47<01:27,  4.00s/it]
 96%|█████████▌| 459/480 [28:51<01:24,  4.00s/it]
 96%|█████████▌| 460/480 [28:55<01:18,  3.94s/it]
                                                 
{'loss': '8.509e-05', 'grad_norm': '0.0002905', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [28:55<01:18,  3.94s/it]
 96%|█████████▌| 461/480 [28:58<01:12,  3.84s/it]
 96%|█████████▋| 462/480 [29:02<01:08,  3.80s/it]
 96%|█████████▋| 463/480 [29:06<01:03,  3.74s/it]
 97%|█████████▋| 464/480 [29:09<00:59,  3.71s/it]
 97%|█████████▋| 465/480 [29:14<00:58,  3.87s/it]
 97%|█████████▋| 466/480 [29:17<00:53,  3.81s/it]
 97%|█████████▋| 467/480 [29:21<00:48,  3.76s/it]
 98%|█████████▊| 468/480 [29:25<00:45,  3.81s/it]
 98%|█████████▊| 469/480 [29:28<00:41,  3.76s/it]
 98%|█████████▊| 470/480 [29:32<00:37,  3.72s/it]
                                                 
{'loss': '0.0001046', 'grad_norm': '0.0002523', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [29:32<00:37,  3.72s/it]
 98%|█████████▊| 471/480 [29:36<00:33,  3.73s/it]
 98%|█████████▊| 472/480 [29:40<00:30,  3.75s/it]
 99%|█████████▊| 473/480 [29:43<00:26,  3.72s/it]
 99%|█████████▉| 474/480 [29:47<00:22,  3.71s/it]
 99%|█████████▉| 475/480 [29:51<00:18,  3.73s/it]
 99%|█████████▉| 476/480 [29:54<00:14,  3.72s/it]
 99%|█████████▉| 477/480 [29:59<00:11,  3.83s/it]
100%|█████████▉| 478/480 [30:02<00:07,  3.77s/it]
100%|█████████▉| 479/480 [30:06<00:03,  3.81s/it]
100%|██████████| 480/480 [30:07<00:00,  2.94s/it]
                                                 
{'loss': '0.0007711', 'grad_norm': '0.0003221', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [30:07<00:00,  2.94s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:00<00:18,  5.26it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:16,  5.84it/s]
+
Eval @ step 480:   3%|▎         | 3/100 [00:00<00:16,  6.05it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:00<00:15,  6.16it/s]
+
Eval @ step 480:   5%|▌         | 5/100 [00:00<00:15,  6.22it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:00<00:15,  6.25it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:01<00:14,  6.27it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:01<00:15,  5.92it/s]
+
Eval @ step 480:   9%|▉         | 9/100 [00:01<00:16,  5.40it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:01<00:17,  5.09it/s]
+
Eval @ step 480:  11%|█         | 11/100 [00:01<00:16,  5.41it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:02<00:17,  5.11it/s]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:02<00:16,  5.16it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:02<00:15,  5.46it/s]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:02<00:14,  5.70it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:02<00:14,  5.88it/s]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:02<00:13,  6.01it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:03<00:14,  5.77it/s]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:03<00:13,  5.93it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:03<00:13,  6.05it/s]
+
Eval @ step 480:  21%|██        | 21/100 [00:03<00:12,  6.13it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:03<00:12,  6.18it/s]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:03<00:12,  6.23it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:04<00:12,  6.25it/s]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:04<00:11,  6.27it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:04<00:12,  5.93it/s]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:04<00:12,  6.05it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:04<00:11,  6.12it/s]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:04<00:12,  5.80it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:05<00:11,  5.93it/s]
+
Eval @ step 480:  31%|███       | 31/100 [00:05<00:11,  6.03it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:05<00:10,  6.49it/s]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:05<00:11,  5.73it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:05<00:11,  5.92it/s]
+
Eval @ step 480:  35%|███▌      | 35/100 [00:05<00:10,  6.09it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:06<00:10,  6.22it/s]
+
Eval @ step 480:  37%|███▋      | 37/100 [00:06<00:09,  6.31it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:06<00:09,  6.38it/s]
+
Eval @ step 480:  39%|███▉      | 39/100 [00:06<00:09,  6.43it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:06<00:09,  6.11it/s]
+
Eval @ step 480:  41%|████      | 41/100 [00:06<00:09,  6.24it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:07<00:10,  5.67it/s]
+
Eval @ step 480:  43%|████▎     | 43/100 [00:07<00:09,  5.91it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:07<00:09,  6.08it/s]
+
Eval @ step 480:  45%|████▌     | 45/100 [00:07<00:08,  6.21it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:07<00:08,  6.32it/s]
+
Eval @ step 480:  47%|████▋     | 47/100 [00:07<00:09,  5.71it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:08<00:09,  5.31it/s]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:08<00:09,  5.56it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:08<00:08,  5.72it/s]
+
Eval @ step 480:  51%|█████     | 51/100 [00:08<00:08,  5.94it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:08<00:07,  6.11it/s]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:08<00:07,  6.24it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:09<00:07,  6.34it/s]
+
Eval @ step 480:  55%|█████▌    | 55/100 [00:09<00:07,  6.04it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:09<00:07,  5.53it/s]
+
Eval @ step 480:  57%|█████▋    | 57/100 [00:09<00:08,  5.23it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:09<00:08,  5.05it/s]
+
Eval @ step 480:  59%|█████▉    | 59/100 [00:10<00:07,  5.16it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:10<00:07,  5.24it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [00:10<00:07,  5.02it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [00:10<00:07,  4.90it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [00:10<00:07,  4.83it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:11<00:07,  5.00it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [00:11<00:06,  5.38it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:11<00:05,  5.68it/s]
+
Eval @ step 480:  67%|██████▋   | 67/100 [00:11<00:05,  5.92it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:11<00:05,  5.77it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [00:11<00:05,  5.68it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:12<00:05,  5.63it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [00:12<00:05,  5.58it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:12<00:04,  5.83it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [00:12<00:04,  5.43it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:12<00:05,  4.94it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [00:13<00:04,  5.34it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:13<00:04,  5.37it/s]
+
Eval @ step 480:  77%|███████▋  | 77/100 [00:13<00:04,  5.39it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:13<00:04,  5.41it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [00:13<00:03,  5.70it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:13<00:03,  5.94it/s]
+
Eval @ step 480:  81%|████████  | 81/100 [00:14<00:03,  6.11it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:14<00:02,  6.24it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [00:14<00:02,  5.98it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:14<00:02,  5.81it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [00:14<00:02,  5.70it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:14<00:02,  5.35it/s]
+
Eval @ step 480:  87%|████████▋ | 87/100 [00:15<00:02,  5.66it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:15<00:02,  5.88it/s]
+
Eval @ step 480:  89%|████████▉ | 89/100 [00:15<00:01,  6.06it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:15<00:01,  6.19it/s]
+
Eval @ step 480:  91%|█████████ | 91/100 [00:15<00:01,  6.30it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:15<00:01,  6.37it/s]
+
Eval @ step 480:  93%|█████████▎| 93/100 [00:16<00:01,  6.06it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:16<00:01,  5.86it/s]
+
Eval @ step 480:  95%|█████████▌| 95/100 [00:16<00:00,  6.05it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:16<00:00,  6.18it/s]
+
Eval @ step 480:  97%|█████████▋| 97/100 [00:16<00:00,  6.28it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:16<00:00,  6.35it/s]
+
Eval @ step 480:  99%|█████████▉| 99/100 [00:17<00:00,  6.41it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:17<00:00,  6.45it/s]
+
                                                                    [EVAL] eval_rougeL=99.1333 (best=99.1333)
+
                                                 
{'eval_rougeL': '99.13', 'epoch': '15'}
+
100%|██████████| 480/480 [30:24<00:00,  2.94s/it]
                                                 
{'train_runtime': '1825', 'train_samples_per_second': '8.221', 'train_steps_per_second': '0.263', 'train_loss': '0.06609', 'epoch': '15'}
+
100%|██████████| 480/480 [30:24<00:00,  2.94s/it]
100%|██████████| 480/480 [30:24<00:00,  3.80s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 61316599GF
+  train_loss               =     0.0661
+  train_runtime            = 0:30:24.67
+  train_samples            =       1000
+  train_samples_per_second =      8.221
+  train_steps_per_second   =      0.263
+  [EVAL] Restored best model (eval_rougeL=99.1333, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   4%|▍         | 13/291 [00:00<00:02, 117.74it/s]
Loading weights:  11%|█         | 31/291 [00:00<00:01, 149.72it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 161.81it/s]
Loading weights:  23%|██▎       | 68/291 [00:00<00:01, 165.06it/s]
Loading weights:  30%|██▉       | 87/291 [00:00<00:01, 165.87it/s]
Loading weights:  38%|███▊      | 112/291 [00:00<00:01, 178.94it/s]
Loading weights:  45%|████▌     | 131/291 [00:00<00:00, 181.90it/s]
Loading weights:  53%|█████▎    | 154/291 [00:00<00:00, 187.21it/s]
Loading weights:  59%|█████▉    | 173/291 [00:00<00:00, 181.75it/s]
Loading weights:  66%|██████▌   | 192/291 [00:01<00:00, 176.61it/s]
Loading weights:  72%|███████▏  | 210/291 [00:01<00:00, 162.09it/s]
Loading weights:  78%|███████▊  | 227/291 [00:01<00:00, 149.57it/s]
Loading weights:  84%|████████▎ | 243/291 [00:01<00:00, 142.29it/s]
Loading weights:  89%|████████▊ | 258/291 [00:01<00:00, 135.30it/s]
Loading weights:  95%|█████████▍| 275/291 [00:01<00:00, 140.08it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 160.89it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task1510_evalution_relation_extraction': PaR=77.1, shrinkage=ridge, n=1000
+  [SRT] Wired router: 6 sigs, cur=task1510_evalution_relation_extraction, mapping={'task1510_evalution_relation_extraction': 0, 'task1572_samsum_summary': 5, 'task363_sst2_polarity_classification': 4, 'task1290_xsum_summarization': 3, 'task181_outcome_extraction': 2, 'task002_quoref_answer_generation': 1}
+[SRT] Router: 6 tasks, avg_PaR=77.1, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction']
+  [SRT-INIT] task_id_to_idx={'task1510_evalution_relation_extraction': 0, 'task1572_samsum_summary': 5, 'task363_sst2_polarity_classification': 4, 'task1290_xsum_summarization': 3, 'task181_outcome_extraction': 2, 'task002_quoref_answer_generation': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/463 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=6
+
Continual predict:   0%|          | 1/463 [00:01<13:06,  1.70s/it]
Continual predict:   0%|          | 2/463 [00:04<16:15,  2.12s/it]
Continual predict:   1%|          | 3/463 [00:04<11:28,  1.50s/it]
Continual predict:   1%|          | 4/463 [00:05<08:28,  1.11s/it]
Continual predict:   1%|          | 5/463 [00:07<10:13,  1.34s/it]
Continual predict:   1%|▏         | 6/463 [00:11<18:27,  2.42s/it]
Continual predict:   2%|▏         | 7/463 [00:14<20:20,  2.68s/it]
Continual predict:   2%|▏         | 8/463 [00:16<18:36,  2.45s/it]
Continual predict:   2%|▏         | 9/463 [00:18<16:37,  2.20s/it]
Continual predict:   2%|▏         | 10/463 [00:19<13:14,  1.75s/it]
Continual predict:   2%|▏         | 11/463 [00:21<15:05,  2.00s/it]
Continual predict:   3%|▎         | 12/463 [00:24<17:23,  2.31s/it]
Continual predict:   3%|▎         | 13/463 [00:25<14:46,  1.97s/it]
Continual predict:   3%|▎         | 14/463 [00:28<15:33,  2.08s/it]
Continual predict:   3%|▎         | 15/463 [00:30<15:42,  2.10s/it]
Continual predict:   3%|▎         | 16/463 [00:32<16:34,  2.22s/it]
Continual predict:   4%|▎         | 17/463 [00:40<28:21,  3.82s/it]
Continual predict:   4%|▍         | 18/463 [00:42<24:18,  3.28s/it]
Continual predict:   4%|▍         | 19/463 [00:44<22:25,  3.03s/it]
Continual predict:   4%|▍         | 20/463 [00:46<18:35,  2.52s/it]
Continual predict:   5%|▍         | 21/463 [00:46<13:19,  1.81s/it]
Continual predict:   5%|▍         | 22/463 [00:46<09:36,  1.31s/it]
Continual predict:   5%|▍         | 23/463 [00:46<07:06,  1.03it/s]
Continual predict:   5%|▌         | 24/463 [00:46<05:16,  1.39it/s]
Continual predict:   5%|▌         | 25/463 [00:47<03:59,  1.83it/s]
Continual predict:   6%|▌         | 26/463 [00:47<03:05,  2.36it/s]
Continual predict:   6%|▌         | 27/463 [00:47<02:27,  2.95it/s]
Continual predict:   6%|▌         | 28/463 [00:47<02:00,  3.60it/s]
Continual predict:   6%|▋         | 29/463 [00:47<01:42,  4.24it/s]
Continual predict:   6%|▋         | 30/463 [00:47<01:31,  4.75it/s]
Continual predict:   7%|▋         | 31/463 [00:47<01:21,  5.28it/s]
Continual predict:   7%|▋         | 32/463 [00:48<01:14,  5.76it/s]
Continual predict:   7%|▋         | 33/463 [00:48<01:12,  5.96it/s]
Continual predict:   7%|▋         | 34/463 [00:48<01:09,  6.13it/s]
Continual predict:   8%|▊         | 35/463 [00:48<01:08,  6.26it/s]
Continual predict:   8%|▊         | 36/463 [00:48<01:05,  6.54it/s]
Continual predict:   8%|▊         | 37/463 [00:48<01:05,  6.54it/s]
Continual predict:   8%|▊         | 38/463 [00:48<01:03,  6.72it/s]
Continual predict:   8%|▊         | 39/463 [00:49<01:01,  6.88it/s]
Continual predict:   9%|▊         | 40/463 [00:49<01:00,  7.00it/s]
Continual predict:   9%|▉         | 41/463 [00:49<01:01,  6.88it/s]
Continual predict:   9%|▉         | 42/463 [00:49<01:00,  6.97it/s]
Continual predict:   9%|▉         | 43/463 [00:49<01:01,  6.86it/s]
Continual predict:  10%|▉         | 44/463 [00:49<01:01,  6.78it/s]
Continual predict:  10%|▉         | 45/463 [00:49<01:00,  6.88it/s]
Continual predict:  10%|▉         | 46/463 [00:50<00:59,  6.97it/s]
Continual predict:  10%|█         | 47/463 [00:50<00:59,  6.97it/s]
Continual predict:  10%|█         | 48/463 [00:50<01:00,  6.82it/s]
Continual predict:  11%|█         | 49/463 [00:50<01:01,  6.75it/s]
Continual predict:  11%|█         | 50/463 [00:50<01:00,  6.86it/s]
Continual predict:  11%|█         | 51/463 [00:50<01:03,  6.52it/s]
Continual predict:  11%|█         | 52/463 [00:50<01:01,  6.70it/s]
Continual predict:  11%|█▏        | 53/463 [00:51<00:59,  6.84it/s]
Continual predict:  12%|█▏        | 54/463 [00:51<01:00,  6.77it/s]
Continual predict:  12%|█▏        | 55/463 [00:51<00:58,  6.92it/s]
Continual predict:  12%|█▏        | 56/463 [00:51<00:59,  6.80it/s]
Continual predict:  12%|█▏        | 57/463 [00:51<00:58,  6.92it/s]
Continual predict:  13%|█▎        | 58/463 [00:51<00:57,  7.00it/s]
Continual predict:  13%|█▎        | 59/463 [00:51<00:57,  7.06it/s]
Continual predict:  13%|█▎        | 60/463 [00:52<00:58,  6.88it/s]
Continual predict:  13%|█▎        | 61/463 [00:52<00:57,  6.98it/s]
Continual predict:  13%|█▎        | 62/463 [00:52<00:56,  7.04it/s]
Continual predict:  14%|█▎        | 63/463 [00:52<00:57,  6.91it/s]
Continual predict:  14%|█▍        | 64/463 [00:52<00:57,  6.99it/s]
Continual predict:  14%|█▍        | 65/463 [00:52<00:56,  7.05it/s]
Continual predict:  14%|█▍        | 66/463 [00:52<00:56,  7.06it/s]
Continual predict:  14%|█▍        | 67/463 [00:53<00:56,  7.07it/s]
Continual predict:  15%|█▍        | 68/463 [00:53<00:56,  6.95it/s]
Continual predict:  15%|█▍        | 69/463 [00:53<00:56,  6.91it/s]
Continual predict:  15%|█▌        | 70/463 [00:53<00:55,  7.02it/s]
Continual predict:  15%|█▌        | 71/463 [00:53<00:55,  7.08it/s]
Continual predict:  16%|█▌        | 72/463 [00:53<00:56,  6.93it/s]
Continual predict:  16%|█▌        | 73/463 [00:53<00:55,  7.03it/s]
Continual predict:  16%|█▌        | 74/463 [00:54<00:56,  6.88it/s]
Continual predict:  16%|█▌        | 75/463 [00:54<00:56,  6.91it/s]
Continual predict:  16%|█▋        | 76/463 [00:54<00:55,  7.02it/s]
Continual predict:  17%|█▋        | 77/463 [00:54<00:54,  7.08it/s]
Continual predict:  17%|█▋        | 78/463 [00:54<00:53,  7.14it/s]
Continual predict:  17%|█▋        | 79/463 [00:54<00:53,  7.15it/s]
Continual predict:  17%|█▋        | 80/463 [00:54<00:53,  7.19it/s]
Continual predict:  17%|█▋        | 81/463 [00:55<00:53,  7.19it/s]
Continual predict:  18%|█▊        | 82/463 [00:55<00:53,  7.15it/s]
Continual predict:  18%|█▊        | 83/463 [00:55<00:53,  7.17it/s]
Continual predict:  18%|█▊        | 84/463 [00:55<00:54,  6.98it/s]
Continual predict:  18%|█▊        | 85/463 [00:55<00:53,  7.03it/s]
Continual predict:  19%|█▊        | 86/463 [00:55<00:53,  7.08it/s]
Continual predict:  19%|█▉        | 87/463 [00:55<00:53,  7.08it/s]
Continual predict:  19%|█▉        | 88/463 [00:56<00:52,  7.14it/s]
Continual predict:  19%|█▉        | 89/463 [00:56<00:52,  7.17it/s]
Continual predict:  19%|█▉        | 90/463 [00:56<00:51,  7.21it/s]
Continual predict:  20%|█▉        | 91/463 [00:56<00:51,  7.21it/s]
Continual predict:  20%|█▉        | 92/463 [00:56<00:51,  7.17it/s]
Continual predict:  20%|██        | 93/463 [00:56<00:52,  6.98it/s]
Continual predict:  20%|██        | 94/463 [00:56<00:53,  6.88it/s]
Continual predict:  21%|██        | 95/463 [00:57<00:52,  6.96it/s]
Continual predict:  21%|██        | 96/463 [00:57<00:52,  7.03it/s]
Continual predict:  21%|██        | 97/463 [00:57<00:51,  7.07it/s]
Continual predict:  21%|██        | 98/463 [00:57<00:52,  6.94it/s]
Continual predict:  21%|██▏       | 99/463 [00:57<00:53,  6.84it/s]
Continual predict:  22%|██▏       | 100/463 [00:57<00:53,  6.76it/s]
Continual predict:  22%|██▏       | 101/463 [00:57<00:52,  6.89it/s]
Continual predict:  22%|██▏       | 102/463 [00:58<00:51,  7.00it/s]
Continual predict:  22%|██▏       | 103/463 [00:58<00:50,  7.07it/s]
Continual predict:  22%|██▏       | 104/463 [00:58<00:50,  7.11it/s]
Continual predict:  23%|██▎       | 105/463 [00:58<00:51,  6.96it/s]
Continual predict:  23%|██▎       | 106/463 [00:58<00:50,  7.05it/s]
Continual predict:  23%|██▎       | 107/463 [00:58<00:49,  7.13it/s]
Continual predict:  23%|██▎       | 108/463 [00:58<00:49,  7.15it/s]
Continual predict:  24%|██▎       | 109/463 [00:59<00:49,  7.17it/s]
Continual predict:  24%|██▍       | 110/463 [00:59<00:49,  7.17it/s]
Continual predict:  24%|██▍       | 111/463 [00:59<00:49,  7.18it/s]
Continual predict:  24%|██▍       | 112/463 [00:59<00:48,  7.18it/s]
Continual predict:  24%|██▍       | 113/463 [00:59<00:48,  7.19it/s]
Continual predict:  25%|██▍       | 114/463 [00:59<00:48,  7.19it/s]
Continual predict:  25%|██▍       | 115/463 [00:59<00:49,  7.00it/s]
Continual predict:  25%|██▌       | 116/463 [01:00<00:49,  7.06it/s]
Continual predict:  25%|██▌       | 117/463 [01:00<00:50,  6.92it/s]
Continual predict:  25%|██▌       | 118/463 [01:00<00:50,  6.81it/s]
Continual predict:  26%|██▌       | 119/463 [01:00<00:49,  6.92it/s]
Continual predict:  26%|██▌       | 120/463 [01:00<00:49,  6.99it/s]
Continual predict:  26%|██▌       | 121/463 [01:05<08:11,  1.44s/it]
Continual predict:  26%|██▋       | 122/463 [01:09<12:49,  2.26s/it]
Continual predict:  27%|██▋       | 123/463 [01:14<17:50,  3.15s/it]
Continual predict:  27%|██▋       | 124/463 [01:16<16:24,  2.90s/it]
Continual predict:  27%|██▋       | 125/463 [01:21<19:27,  3.45s/it]
Continual predict:  27%|██▋       | 126/463 [01:26<21:33,  3.84s/it]
Continual predict:  27%|██▋       | 127/463 [01:27<17:34,  3.14s/it]
Continual predict:  28%|██▊       | 128/463 [01:30<16:47,  3.01s/it]
Continual predict:  28%|██▊       | 129/463 [01:33<17:18,  3.11s/it]
Continual predict:  28%|██▊       | 130/463 [01:36<17:00,  3.07s/it]
Continual predict:  28%|██▊       | 131/463 [01:38<15:29,  2.80s/it]
Continual predict:  29%|██▊       | 132/463 [01:41<15:09,  2.75s/it]
Continual predict:  29%|██▊       | 133/463 [01:44<15:41,  2.85s/it]
Continual predict:  29%|██▉       | 134/463 [01:50<20:42,  3.78s/it]
Continual predict:  29%|██▉       | 135/463 [01:55<22:09,  4.05s/it]
Continual predict:  29%|██▉       | 136/463 [01:58<20:41,  3.80s/it]
Continual predict:  30%|██▉       | 137/463 [02:00<18:01,  3.32s/it]
Continual predict:  30%|██▉       | 138/463 [02:01<13:44,  2.54s/it]
Continual predict:  30%|███       | 139/463 [02:02<11:27,  2.12s/it]
Continual predict:  30%|███       | 140/463 [02:05<12:13,  2.27s/it]
Continual predict:  30%|███       | 141/463 [02:11<18:56,  3.53s/it]
Continual predict:  31%|███       | 142/463 [02:16<20:25,  3.82s/it]
Continual predict:  31%|███       | 143/463 [02:19<19:27,  3.65s/it]
Continual predict:  31%|███       | 144/463 [02:28<28:26,  5.35s/it]
Continual predict:  31%|███▏      | 145/463 [02:33<26:42,  5.04s/it]
Continual predict:  32%|███▏      | 146/463 [02:34<21:26,  4.06s/it]
Continual predict:  32%|███▏      | 147/463 [02:39<21:46,  4.14s/it]
Continual predict:  32%|███▏      | 148/463 [02:41<18:31,  3.53s/it]
Continual predict:  32%|███▏      | 149/463 [02:44<17:50,  3.41s/it]
Continual predict:  32%|███▏      | 150/463 [02:46<15:21,  2.94s/it]
Continual predict:  33%|███▎      | 151/463 [02:49<15:26,  2.97s/it]
Continual predict:  33%|███▎      | 152/463 [02:54<19:30,  3.76s/it]
Continual predict:  33%|███▎      | 153/463 [02:57<17:21,  3.36s/it]
Continual predict:  33%|███▎      | 154/463 [03:02<19:26,  3.77s/it]
Continual predict:  33%|███▎      | 155/463 [03:05<18:59,  3.70s/it]
Continual predict:  34%|███▎      | 156/463 [03:08<18:20,  3.59s/it]
Continual predict:  34%|███▍      | 157/463 [03:12<18:35,  3.64s/it]
Continual predict:  34%|███▍      | 158/463 [03:13<14:21,  2.82s/it]
Continual predict:  34%|███▍      | 159/463 [03:20<20:24,  4.03s/it]
Continual predict:  35%|███▍      | 160/463 [03:22<17:36,  3.49s/it]
Continual predict:  35%|███▍      | 161/463 [03:25<15:58,  3.17s/it]
Continual predict:  35%|███▍      | 162/463 [03:29<18:18,  3.65s/it]
Continual predict:  35%|███▌      | 163/463 [03:31<15:48,  3.16s/it]
Continual predict:  35%|███▌      | 164/463 [03:34<14:21,  2.88s/it]
Continual predict:  36%|███▌      | 165/463 [03:37<14:35,  2.94s/it]
Continual predict:  36%|███▌      | 166/463 [03:39<13:18,  2.69s/it]
Continual predict:  36%|███▌      | 167/463 [03:40<11:12,  2.27s/it]
Continual predict:  36%|███▋      | 168/463 [03:42<11:06,  2.26s/it]
Continual predict:  37%|███▋      | 169/463 [03:44<10:14,  2.09s/it]
Continual predict:  37%|███▋      | 170/463 [03:47<11:31,  2.36s/it]
Continual predict:  37%|███▋      | 171/463 [03:49<10:12,  2.10s/it]
Continual predict:  37%|███▋      | 172/463 [03:51<10:48,  2.23s/it]
Continual predict:  37%|███▋      | 173/463 [03:55<13:11,  2.73s/it]
Continual predict:  38%|███▊      | 174/463 [03:56<10:50,  2.25s/it]
Continual predict:  38%|███▊      | 175/463 [03:59<12:27,  2.59s/it]
Continual predict:  38%|███▊      | 176/463 [04:02<12:43,  2.66s/it]
Continual predict:  38%|███▊      | 177/463 [04:05<12:40,  2.66s/it]
Continual predict:  38%|███▊      | 178/463 [04:06<10:35,  2.23s/it]
Continual predict:  39%|███▊      | 179/463 [04:07<08:56,  1.89s/it]
Continual predict:  39%|███▉      | 180/463 [04:09<08:04,  1.71s/it]
Continual predict:  39%|███▉      | 181/463 [04:10<07:27,  1.59s/it]
Continual predict:  39%|███▉      | 182/463 [04:14<11:43,  2.50s/it]
Continual predict:  40%|███▉      | 183/463 [04:15<08:28,  1.82s/it]
Continual predict:  40%|███▉      | 184/463 [04:16<07:25,  1.60s/it]
Continual predict:  40%|███▉      | 185/463 [04:19<09:44,  2.10s/it]
Continual predict:  40%|████      | 186/463 [04:20<08:05,  1.75s/it]
Continual predict:  40%|████      | 187/463 [04:23<10:04,  2.19s/it]
Continual predict:  41%|████      | 188/463 [04:26<10:41,  2.33s/it]
Continual predict:  41%|████      | 189/463 [04:31<14:09,  3.10s/it]
Continual predict:  41%|████      | 190/463 [04:34<14:13,  3.13s/it]
Continual predict:  41%|████▏     | 191/463 [04:35<11:41,  2.58s/it]
Continual predict:  41%|████▏     | 192/463 [04:39<12:32,  2.78s/it]
Continual predict:  42%|████▏     | 193/463 [04:43<14:45,  3.28s/it]
Continual predict:  42%|████▏     | 194/463 [04:48<16:39,  3.71s/it]
Continual predict:  42%|████▏     | 195/463 [04:53<18:35,  4.16s/it]
Continual predict:  42%|████▏     | 196/463 [04:56<17:21,  3.90s/it]
Continual predict:  43%|████▎     | 197/463 [04:59<16:24,  3.70s/it]
Continual predict:  43%|████▎     | 198/463 [05:07<21:02,  4.76s/it]
Continual predict:  43%|████▎     | 199/463 [05:08<16:05,  3.66s/it]
Continual predict:  43%|████▎     | 200/463 [05:09<13:04,  2.98s/it]
Continual predict:  43%|████▎     | 201/463 [05:12<12:20,  2.83s/it]
Continual predict:  44%|████▎     | 202/463 [05:15<12:44,  2.93s/it]
Continual predict:  44%|████▍     | 203/463 [05:20<15:27,  3.57s/it]
Continual predict:  44%|████▍     | 204/463 [05:23<14:36,  3.38s/it]
Continual predict:  44%|████▍     | 205/463 [05:29<18:34,  4.32s/it]
Continual predict:  44%|████▍     | 206/463 [05:32<16:24,  3.83s/it]
Continual predict:  45%|████▍     | 207/463 [05:36<16:08,  3.78s/it]
Continual predict:  45%|████▍     | 208/463 [05:42<18:59,  4.47s/it]
Continual predict:  45%|████▌     | 209/463 [05:44<16:35,  3.92s/it]
Continual predict:  45%|████▌     | 210/463 [05:46<13:38,  3.23s/it]
Continual predict:  46%|████▌     | 211/463 [05:49<12:42,  3.03s/it]
Continual predict:  46%|████▌     | 212/463 [05:52<12:39,  3.03s/it]
Continual predict:  46%|████▌     | 213/463 [05:55<13:25,  3.22s/it]
Continual predict:  46%|████▌     | 214/463 [05:59<13:44,  3.31s/it]
Continual predict:  46%|████▋     | 215/463 [06:00<10:29,  2.54s/it]
Continual predict:  47%|████▋     | 216/463 [06:03<11:01,  2.68s/it]
Continual predict:  47%|████▋     | 217/463 [06:04<10:06,  2.47s/it]
Continual predict:  47%|████▋     | 218/463 [06:08<11:44,  2.87s/it]
Continual predict:  47%|████▋     | 219/463 [06:12<12:06,  2.98s/it]
Continual predict:  48%|████▊     | 220/463 [06:16<13:17,  3.28s/it]
Continual predict:  48%|████▊     | 221/463 [06:16<09:31,  2.36s/it]
Continual predict:  48%|████▊     | 222/463 [06:16<06:57,  1.73s/it]
Continual predict:  48%|████▊     | 223/463 [06:16<05:05,  1.27s/it]
Continual predict:  48%|████▊     | 224/463 [06:16<03:47,  1.05it/s]
Continual predict:  49%|████▊     | 225/463 [06:17<02:53,  1.37it/s]
Continual predict:  49%|████▉     | 226/463 [06:17<02:15,  1.75it/s]
Continual predict:  49%|████▉     | 227/463 [06:17<01:52,  2.10it/s]
Continual predict:  49%|████▉     | 228/463 [06:17<01:36,  2.44it/s]
Continual predict:  49%|████▉     | 229/463 [06:18<01:22,  2.83it/s]
Continual predict:  50%|████▉     | 230/463 [06:18<01:25,  2.71it/s]
Continual predict:  50%|████▉     | 231/463 [06:18<01:14,  3.13it/s]
Continual predict:  50%|█████     | 232/463 [06:18<01:05,  3.52it/s]
Continual predict:  50%|█████     | 233/463 [06:19<00:59,  3.85it/s]
Continual predict:  51%|█████     | 234/463 [06:19<00:55,  4.12it/s]
Continual predict:  51%|█████     | 235/463 [06:19<00:52,  4.34it/s]
Continual predict:  51%|█████     | 236/463 [06:19<00:50,  4.51it/s]
Continual predict:  51%|█████     | 237/463 [06:19<00:48,  4.64it/s]
Continual predict:  51%|█████▏    | 238/463 [06:20<00:47,  4.73it/s]
Continual predict:  52%|█████▏    | 239/463 [06:20<00:46,  4.79it/s]
Continual predict:  52%|█████▏    | 240/463 [06:20<00:55,  4.04it/s]
Continual predict:  52%|█████▏    | 241/463 [06:20<00:51,  4.27it/s]
Continual predict:  52%|█████▏    | 242/463 [06:21<00:49,  4.44it/s]
Continual predict:  52%|█████▏    | 243/463 [06:21<00:48,  4.58it/s]
Continual predict:  53%|█████▎    | 244/463 [06:21<00:46,  4.69it/s]
Continual predict:  53%|█████▎    | 245/463 [06:21<00:45,  4.76it/s]
Continual predict:  53%|█████▎    | 246/463 [06:21<00:45,  4.82it/s]
Continual predict:  53%|█████▎    | 247/463 [06:22<00:47,  4.51it/s]
Continual predict:  54%|█████▎    | 248/463 [06:22<00:46,  4.63it/s]
Continual predict:  54%|█████▍    | 249/463 [06:22<00:48,  4.39it/s]
Continual predict:  54%|█████▍    | 250/463 [06:22<00:46,  4.55it/s]
Continual predict:  54%|█████▍    | 251/463 [06:22<00:48,  4.34it/s]
Continual predict:  54%|█████▍    | 252/463 [06:24<01:50,  1.91it/s]
Continual predict:  55%|█████▍    | 253/463 [06:24<01:32,  2.26it/s]
Continual predict:  55%|█████▍    | 254/463 [06:24<01:17,  2.70it/s]
Continual predict:  55%|█████▌    | 255/463 [06:24<01:06,  3.12it/s]
Continual predict:  55%|█████▌    | 256/463 [06:25<00:59,  3.51it/s]
Continual predict:  56%|█████▌    | 257/463 [06:25<01:24,  2.44it/s]
Continual predict:  56%|█████▌    | 258/463 [06:25<01:11,  2.88it/s]
Continual predict:  56%|█████▌    | 259/463 [06:26<01:02,  3.29it/s]
Continual predict:  56%|█████▌    | 260/463 [06:26<00:55,  3.65it/s]
Continual predict:  56%|█████▋    | 261/463 [06:26<01:11,  2.83it/s]
Continual predict:  57%|█████▋    | 262/463 [06:27<01:03,  3.19it/s]
Continual predict:  57%|█████▋    | 263/463 [06:27<00:59,  3.38it/s]
Continual predict:  57%|█████▋    | 264/463 [06:28<01:53,  1.76it/s]
Continual predict:  57%|█████▋    | 265/463 [06:29<01:49,  1.81it/s]
Continual predict:  57%|█████▋    | 266/463 [06:29<01:47,  1.84it/s]
Continual predict:  58%|█████▊    | 267/463 [06:30<01:49,  1.79it/s]
Continual predict:  58%|█████▊    | 268/463 [06:31<02:03,  1.58it/s]
Continual predict:  58%|█████▊    | 269/463 [06:31<02:00,  1.61it/s]
Continual predict:  58%|█████▊    | 270/463 [06:32<01:59,  1.62it/s]
Continual predict:  59%|█████▊    | 271/463 [06:32<02:06,  1.52it/s]
Continual predict:  59%|█████▊    | 272/463 [06:33<02:18,  1.38it/s]
Continual predict:  59%|█████▉    | 273/463 [06:34<02:12,  1.44it/s]
Continual predict:  59%|█████▉    | 274/463 [06:35<02:08,  1.47it/s]
Continual predict:  59%|█████▉    | 275/463 [06:35<02:02,  1.53it/s]
Continual predict:  60%|█████▉    | 276/463 [06:36<02:23,  1.30it/s]
Continual predict:  60%|█████▉    | 277/463 [06:37<02:49,  1.10it/s]
Continual predict:  60%|██████    | 278/463 [06:38<02:43,  1.13it/s]
Continual predict:  60%|██████    | 279/463 [06:39<02:31,  1.21it/s]
Continual predict:  60%|██████    | 280/463 [06:39<02:10,  1.41it/s]
Continual predict:  61%|██████    | 281/463 [06:40<02:21,  1.29it/s]
Continual predict:  61%|██████    | 282/463 [06:41<02:21,  1.28it/s]
Continual predict:  61%|██████    | 283/463 [06:42<02:13,  1.34it/s]
Continual predict:  61%|██████▏   | 284/463 [06:42<01:57,  1.53it/s]
Continual predict:  62%|██████▏   | 285/463 [06:43<02:02,  1.45it/s]
Continual predict:  62%|██████▏   | 286/463 [06:44<02:05,  1.41it/s]
Continual predict:  62%|██████▏   | 287/463 [06:45<02:06,  1.39it/s]
Continual predict:  62%|██████▏   | 288/463 [06:45<02:00,  1.45it/s]
Continual predict:  62%|██████▏   | 289/463 [06:46<02:11,  1.32it/s]
Continual predict:  63%|██████▎   | 290/463 [06:47<02:10,  1.33it/s]
Continual predict:  63%|██████▎   | 291/463 [06:47<02:01,  1.42it/s]
Continual predict:  63%|██████▎   | 292/463 [06:48<01:40,  1.71it/s]
Continual predict:  63%|██████▎   | 293/463 [06:49<01:51,  1.53it/s]
Continual predict:  63%|██████▎   | 294/463 [06:49<01:43,  1.63it/s]
Continual predict:  64%|██████▎   | 295/463 [06:49<01:34,  1.79it/s]
Continual predict:  64%|██████▍   | 296/463 [06:50<01:53,  1.48it/s]
Continual predict:  64%|██████▍   | 297/463 [06:51<01:43,  1.60it/s]
Continual predict:  64%|██████▍   | 298/463 [06:51<01:34,  1.75it/s]
Continual predict:  65%|██████▍   | 299/463 [06:52<01:50,  1.48it/s]
Continual predict:  65%|██████▍   | 300/463 [06:53<01:38,  1.66it/s]
Continual predict:  65%|██████▌   | 301/463 [06:54<01:52,  1.44it/s]
Continual predict:  65%|██████▌   | 302/463 [06:54<01:37,  1.65it/s]
Continual predict:  65%|██████▌   | 303/463 [06:56<02:20,  1.14it/s]
Continual predict:  66%|██████▌   | 304/463 [06:56<01:52,  1.42it/s]
Continual predict:  66%|██████▌   | 305/463 [06:57<02:00,  1.31it/s]
Continual predict:  66%|██████▌   | 306/463 [06:58<02:00,  1.30it/s]
Continual predict:  66%|██████▋   | 307/463 [06:58<02:00,  1.30it/s]
Continual predict:  67%|██████▋   | 308/463 [06:59<01:43,  1.50it/s]
Continual predict:  67%|██████▋   | 309/463 [06:59<01:32,  1.67it/s]
Continual predict:  67%|██████▋   | 310/463 [07:00<01:31,  1.67it/s]
Continual predict:  67%|██████▋   | 311/463 [07:00<01:27,  1.74it/s]
Continual predict:  67%|██████▋   | 312/463 [07:01<01:32,  1.63it/s]
Continual predict:  68%|██████▊   | 313/463 [07:02<01:42,  1.46it/s]
Continual predict:  68%|██████▊   | 314/463 [07:02<01:31,  1.63it/s]
Continual predict:  68%|██████▊   | 315/463 [07:03<01:43,  1.44it/s]
Continual predict:  68%|██████▊   | 316/463 [07:04<01:46,  1.38it/s]
Continual predict:  68%|██████▊   | 317/463 [07:04<01:27,  1.67it/s]
Continual predict:  69%|██████▊   | 318/463 [07:05<01:20,  1.81it/s]
Continual predict:  69%|██████▉   | 319/463 [07:06<02:03,  1.16it/s]
Continual predict:  69%|██████▉   | 320/463 [07:07<01:55,  1.24it/s]
Continual predict:  69%|██████▉   | 321/463 [07:08<01:49,  1.30it/s]
Continual predict:  70%|██████▉   | 322/463 [07:08<01:33,  1.51it/s]
Continual predict:  70%|██████▉   | 323/463 [07:09<01:35,  1.47it/s]
Continual predict:  70%|██████▉   | 324/463 [07:10<01:42,  1.35it/s]
Continual predict:  70%|███████   | 325/463 [07:10<01:42,  1.35it/s]
Continual predict:  70%|███████   | 326/463 [07:11<01:36,  1.42it/s]
Continual predict:  71%|███████   | 327/463 [07:12<01:51,  1.22it/s]
Continual predict:  71%|███████   | 328/463 [07:13<01:59,  1.13it/s]
Continual predict:  71%|███████   | 329/463 [07:14<01:45,  1.28it/s]
Continual predict:  71%|███████▏  | 330/463 [07:14<01:25,  1.56it/s]
Continual predict:  71%|███████▏  | 331/463 [07:15<01:22,  1.60it/s]
Continual predict:  72%|███████▏  | 332/463 [07:15<01:17,  1.68it/s]
Continual predict:  72%|███████▏  | 333/463 [07:16<01:40,  1.29it/s]
Continual predict:  72%|███████▏  | 334/463 [07:17<01:44,  1.24it/s]
Continual predict:  72%|███████▏  | 335/463 [07:18<01:28,  1.45it/s]
Continual predict:  73%|███████▎  | 336/463 [07:18<01:18,  1.62it/s]
Continual predict:  73%|███████▎  | 337/463 [07:19<01:15,  1.67it/s]
Continual predict:  73%|███████▎  | 338/463 [07:19<01:15,  1.65it/s]
Continual predict:  73%|███████▎  | 339/463 [07:20<01:11,  1.73it/s]
Continual predict:  73%|███████▎  | 340/463 [07:21<01:28,  1.39it/s]
Continual predict:  74%|███████▎  | 341/463 [07:22<01:28,  1.38it/s]
Continual predict:  74%|███████▍  | 342/463 [07:22<01:17,  1.56it/s]
Continual predict:  74%|███████▍  | 343/463 [07:23<01:20,  1.49it/s]
Continual predict:  74%|███████▍  | 344/463 [07:24<01:29,  1.34it/s]
Continual predict:  75%|███████▍  | 345/463 [07:24<01:17,  1.52it/s]
Continual predict:  75%|███████▍  | 346/463 [07:25<01:21,  1.43it/s]
Continual predict:  75%|███████▍  | 347/463 [07:26<01:45,  1.09it/s]
Continual predict:  75%|███████▌  | 348/463 [07:27<01:25,  1.35it/s]
Continual predict:  75%|███████▌  | 349/463 [07:27<01:13,  1.56it/s]
Continual predict:  76%|███████▌  | 350/463 [07:28<01:12,  1.56it/s]
Continual predict:  76%|███████▌  | 351/463 [07:29<01:26,  1.29it/s]
Continual predict:  76%|███████▌  | 352/463 [07:30<01:24,  1.31it/s]
Continual predict:  76%|███████▌  | 353/463 [07:31<01:40,  1.10it/s]
Continual predict:  76%|███████▋  | 354/463 [07:32<01:49,  1.01s/it]
Continual predict:  77%|███████▋  | 355/463 [07:33<01:35,  1.13it/s]
Continual predict:  77%|███████▋  | 356/463 [07:34<01:34,  1.13it/s]
Continual predict:  77%|███████▋  | 357/463 [07:35<01:39,  1.07it/s]
Continual predict:  77%|███████▋  | 358/463 [07:35<01:32,  1.14it/s]
Continual predict:  78%|███████▊  | 359/463 [07:36<01:23,  1.25it/s]
Continual predict:  78%|███████▊  | 360/463 [07:37<01:35,  1.08it/s]
Continual predict:  78%|███████▊  | 361/463 [07:39<01:53,  1.11s/it]
Continual predict:  78%|███████▊  | 362/463 [07:39<01:36,  1.05it/s]
Continual predict:  78%|███████▊  | 363/463 [07:40<01:20,  1.25it/s]
Continual predict:  79%|███████▊  | 364/463 [07:40<01:04,  1.53it/s]
Continual predict:  79%|███████▉  | 365/463 [07:40<00:52,  1.88it/s]
Continual predict:  79%|███████▉  | 366/463 [07:41<00:43,  2.23it/s]
Continual predict:  79%|███████▉  | 367/463 [07:41<00:37,  2.56it/s]
Continual predict:  79%|███████▉  | 368/463 [07:41<00:33,  2.86it/s]
Continual predict:  80%|███████▉  | 369/463 [07:41<00:30,  3.12it/s]
Continual predict:  80%|███████▉  | 370/463 [07:42<00:27,  3.33it/s]
Continual predict:  80%|████████  | 371/463 [07:42<00:27,  3.32it/s]
Continual predict:  80%|████████  | 372/463 [07:42<00:28,  3.15it/s]
Continual predict:  81%|████████  | 373/463 [07:43<00:29,  3.04it/s]
Continual predict:  81%|████████  | 374/463 [07:43<00:27,  3.27it/s]
Continual predict:  81%|████████  | 375/463 [07:43<00:28,  3.12it/s]
Continual predict:  81%|████████  | 376/463 [07:43<00:27,  3.17it/s]
Continual predict:  81%|████████▏ | 377/463 [07:44<00:25,  3.37it/s]
Continual predict:  82%|████████▏ | 378/463 [07:44<00:24,  3.52it/s]
Continual predict:  82%|████████▏ | 379/463 [07:44<00:23,  3.64it/s]
Continual predict:  82%|████████▏ | 380/463 [07:45<00:22,  3.73it/s]
Continual predict:  82%|████████▏ | 381/463 [07:45<00:23,  3.52it/s]
Continual predict:  83%|████████▎ | 382/463 [07:45<00:22,  3.61it/s]
Continual predict:  83%|████████▎ | 383/463 [07:45<00:21,  3.68it/s]
Continual predict:  83%|████████▎ | 384/463 [07:46<00:21,  3.74it/s]
Continual predict:  83%|████████▎ | 385/463 [07:46<00:26,  2.93it/s]
Continual predict:  83%|████████▎ | 386/463 [07:46<00:24,  3.17it/s]
Continual predict:  84%|████████▎ | 387/463 [07:47<00:22,  3.37it/s]
Continual predict:  84%|████████▍ | 388/463 [07:47<00:21,  3.52it/s]
Continual predict:  84%|████████▍ | 389/463 [07:47<00:21,  3.44it/s]
Continual predict:  84%|████████▍ | 390/463 [07:47<00:20,  3.58it/s]
Continual predict:  84%|████████▍ | 391/463 [07:48<00:19,  3.68it/s]
Continual predict:  85%|████████▍ | 392/463 [07:48<00:20,  3.55it/s]
Continual predict:  85%|████████▍ | 393/463 [07:48<00:19,  3.58it/s]
Continual predict:  85%|████████▌ | 394/463 [07:49<00:18,  3.68it/s]
Continual predict:  85%|████████▌ | 395/463 [07:49<00:17,  3.98it/s]
Continual predict:  86%|████████▌ | 396/463 [07:49<00:18,  3.55it/s]
Continual predict:  86%|████████▌ | 397/463 [07:49<00:18,  3.66it/s]
Continual predict:  86%|████████▌ | 398/463 [07:50<00:17,  3.74it/s]
Continual predict:  86%|████████▌ | 399/463 [07:50<00:16,  3.80it/s]
Continual predict:  86%|████████▋ | 400/463 [07:50<00:16,  3.84it/s]
Continual predict:  87%|████████▋ | 401/463 [07:50<00:16,  3.87it/s]
Continual predict:  87%|████████▋ | 402/463 [07:51<00:15,  3.89it/s]
Continual predict:  87%|████████▋ | 403/463 [07:51<00:16,  3.69it/s]
Continual predict:  87%|████████▋ | 404/463 [07:51<00:15,  3.77it/s]
Continual predict:  87%|████████▋ | 405/463 [07:52<00:16,  3.42it/s]
Continual predict:  88%|████████▊ | 406/463 [07:52<00:15,  3.56it/s]
Continual predict:  88%|████████▊ | 407/463 [07:52<00:15,  3.67it/s]
Continual predict:  88%|████████▊ | 408/463 [07:52<00:14,  3.74it/s]
Continual predict:  88%|████████▊ | 409/463 [07:53<00:14,  3.80it/s]
Continual predict:  89%|████████▊ | 410/463 [07:53<00:15,  3.44it/s]
Continual predict:  89%|████████▉ | 411/463 [07:53<00:16,  3.23it/s]
Continual predict:  89%|████████▉ | 412/463 [07:53<00:14,  3.41it/s]
Continual predict:  89%|████████▉ | 413/463 [07:54<00:14,  3.55it/s]
Continual predict:  89%|████████▉ | 414/463 [07:54<00:13,  3.66it/s]
Continual predict:  90%|████████▉ | 415/463 [07:54<00:13,  3.63it/s]
Continual predict:  90%|████████▉ | 416/463 [07:55<00:12,  3.72it/s]
Continual predict:  90%|█████████ | 417/463 [07:55<00:12,  3.78it/s]
Continual predict:  90%|█████████ | 418/463 [07:55<00:12,  3.62it/s]
Continual predict:  90%|█████████ | 419/463 [07:55<00:13,  3.33it/s]
Continual predict:  91%|█████████ | 420/463 [07:56<00:15,  2.76it/s]
Continual predict:  91%|█████████ | 421/463 [07:56<00:15,  2.71it/s]
Continual predict:  91%|█████████ | 422/463 [07:57<00:14,  2.86it/s]
Continual predict:  91%|█████████▏| 423/463 [07:57<00:13,  2.98it/s]
Continual predict:  92%|█████████▏| 424/463 [07:57<00:13,  2.93it/s]
Continual predict:  92%|█████████▏| 425/463 [07:58<00:13,  2.90it/s]
Continual predict:  92%|█████████▏| 426/463 [07:58<00:13,  2.84it/s]
Continual predict:  92%|█████████▏| 427/463 [07:58<00:12,  2.96it/s]
Continual predict:  92%|█████████▏| 428/463 [07:59<00:10,  3.20it/s]
Continual predict:  93%|█████████▎| 429/463 [07:59<00:10,  3.39it/s]
Continual predict:  93%|█████████▎| 430/463 [07:59<00:09,  3.54it/s]
Continual predict:  93%|█████████▎| 431/463 [07:59<00:09,  3.46it/s]
Continual predict:  93%|█████████▎| 432/463 [08:00<00:08,  3.59it/s]
Continual predict:  94%|█████████▎| 433/463 [08:00<00:08,  3.69it/s]
Continual predict:  94%|█████████▎| 434/463 [08:00<00:08,  3.56it/s]
Continual predict:  94%|█████████▍| 435/463 [08:00<00:07,  3.67it/s]
Continual predict:  94%|█████████▍| 436/463 [08:01<00:08,  3.36it/s]
Continual predict:  94%|█████████▍| 437/463 [08:01<00:08,  2.98it/s]
Continual predict:  95%|█████████▍| 438/463 [08:02<00:08,  3.12it/s]
Continual predict:  95%|█████████▍| 439/463 [08:02<00:08,  2.80it/s]
Continual predict:  95%|█████████▌| 440/463 [08:02<00:07,  2.93it/s]
Continual predict:  95%|█████████▌| 441/463 [08:03<00:07,  3.03it/s]
Continual predict:  95%|█████████▌| 442/463 [08:03<00:06,  3.26it/s]
Continual predict:  96%|█████████▌| 443/463 [08:03<00:05,  3.44it/s]
Continual predict:  96%|█████████▌| 444/463 [08:03<00:05,  3.58it/s]
Continual predict:  96%|█████████▌| 445/463 [08:04<00:04,  3.68it/s]
Continual predict:  96%|█████████▋| 446/463 [08:04<00:04,  3.56it/s]
Continual predict:  97%|█████████▋| 447/463 [08:04<00:04,  3.47it/s]
Continual predict:  97%|█████████▋| 448/463 [08:05<00:04,  3.42it/s]
Continual predict:  97%|█████████▋| 449/463 [08:05<00:04,  3.21it/s]
Continual predict:  97%|█████████▋| 450/463 [08:05<00:03,  3.40it/s]
Continual predict:  97%|█████████▋| 451/463 [08:05<00:03,  3.54it/s]
Continual predict:  98%|█████████▊| 452/463 [08:06<00:03,  3.65it/s]
Continual predict:  98%|█████████▊| 453/463 [08:06<00:02,  3.73it/s]
Continual predict:  98%|█████████▊| 454/463 [08:06<00:02,  3.75it/s]
Continual predict:  98%|█████████▊| 455/463 [08:06<00:02,  3.79it/s]
Continual predict:  98%|█████████▊| 456/463 [08:07<00:01,  3.62it/s]
Continual predict:  99%|█████████▊| 457/463 [08:07<00:01,  3.51it/s]
Continual predict:  99%|█████████▉| 458/463 [08:07<00:01,  3.63it/s]
Continual predict:  99%|█████████▉| 459/463 [08:08<00:01,  3.72it/s]
Continual predict:  99%|█████████▉| 460/463 [08:08<00:00,  3.78it/s]
Continual predict: 100%|█████████▉| 461/463 [08:08<00:00,  3.83it/s]
Continual predict: 100%|█████████▉| 462/463 [08:08<00:00,  3.86it/s]
Continual predict: 100%|██████████| 463/463 [08:09<00:00,  3.89it/s]
Continual predict: 100%|██████████| 463/463 [08:09<00:00,  1.06s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=20]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (463/463)  label_acc=100.0% (463/463)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                          =       15.0
+  predict_eval_rougeL                                            =    71.2535
+  predict_eval_rougeL_for_CL                                     =    71.2535
+  predict_eval_rougeL_for_task002_quoref_answer_generation       =     76.381
+  predict_eval_rougeL_for_task1290_xsum_summarization            =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                =    41.3728
+  predict_eval_rougeL_for_task181_outcome_extraction             =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification   =       91.0
+  predict_exact_match                                            =    60.9071
+  predict_exact_match_for_CL                                     =    60.9071
+  predict_exact_match_for_task002_quoref_answer_generation       =       70.0
+  predict_exact_match_for_task1290_xsum_summarization            =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction =       98.0
+  predict_exact_match_for_task1572_samsum_summary                =        0.0
+  predict_exact_match_for_task181_outcome_extraction             =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification   =       91.0
+  predict_gen_len                                                =      9.054
+  predict_global_step                                            =        480
+  predict_rouge1                                                 =    73.1888
+  predict_rouge1_for_CL                                          =    73.1888
+  predict_rouge1_for_task002_quoref_answer_generation            =     76.381
+  predict_rouge1_for_task1290_xsum_summarization                 =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction      =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                     =    50.3879
+  predict_rouge1_for_task181_outcome_extraction                  =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification        =       91.0
+  predict_runtime                                                = 0:08:09.02
+  predict_samples                                                =        463
+  predict_samples_per_second                                     =      0.947
+  predict_steps_per_second                                       =      0.947
+[2026-04-29 23:37:41,934] [INFO] [launch.py:367:main] Process 966508 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/all_results.json b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..24df5e21be4159ad488369fe669f8dc2727ed749
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/all_results.json
@@ -0,0 +1,42 @@
+{
+  "predict_exact_match": 58.6279,
+  "predict_rouge1": 71.0221,
+  "predict_eval_rougeL": 69.1592,
+  "predict_gen_len": 8.948,
+  "predict_runtime": 503.45249552000314,
+  "predict_samples": 481,
+  "predict_samples_per_second": 0.955,
+  "predict_steps_per_second": 0.955,
+  "epoch": 15.0,
+  "predict_global_step": 75,
+  "predict_exact_match_for_CL": 58.6279,
+  "predict_rouge1_for_CL": 71.0221,
+  "predict_eval_rougeL_for_CL": 69.1592,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "train_runtime": 294.6273,
+  "train_samples_per_second": 7.229,
+  "train_steps_per_second": 0.255,
+  "total_flos": 6923218845696000.0,
+  "train_loss": 2.5188296127319334,
+  "train_samples": 142
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..bd3cb4484c223951c88588478bfc1cd5defbea37
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb5f8bb056d41403b49d4b83ea360b9db64c25b1f207a53a4a10171cb767e449
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..3b392e38e5c716cae5f73859c21bbc3c0e36938e
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5393fbbdf615d7076d6db55111e2518557436d234abd6728472824a394208d7
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..91833a736140b509ca1c4074ed079467ba7aee82
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59386cfd59616e76395a312fbae03aed4105053ac6e13b7b7d9bfb4d3ce28f6e
+size 1137678470
diff --git a/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/train.log b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..a66138cf6074f4ca442a1cf9933f0b02cd996cf1
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/train.log
@@ -0,0 +1,272 @@
+[2026-04-29 23:37:47,850] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 23:37:47,850] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29506 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task639_multi_woz_user_utterance_generation --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation --cur_task_id 6 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[2026-04-29 23:37:51,061] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 23:37:51,061] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 23:37:51,061] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 23:37:51,061] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 23:37:51,061] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 23:37:51,061] [INFO] [launch.py:272:main] process 1066468 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task639_multi_woz_user_utterance_generation', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation', '--cur_task_id', '6', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights']
+
+============================================================
+Task 6: task639_multi_woz_user_utterance_generation
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 142/18/481
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:14, 19.86it/s]
Loading weights:  10%|▉         | 28/291 [00:00<00:01, 160.22it/s]
Loading weights:  16%|█▌        | 47/291 [00:00<00:01, 171.36it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 173.35it/s]
Loading weights:  29%|██▉       | 85/291 [00:00<00:01, 174.51it/s]
Loading weights:  35%|███▌      | 103/291 [00:00<00:01, 167.21it/s]
Loading weights:  42%|████▏     | 122/291 [00:00<00:00, 172.33it/s]
Loading weights:  48%|████▊     | 141/291 [00:00<00:00, 172.34it/s]
Loading weights:  56%|█████▌    | 162/291 [00:00<00:00, 181.76it/s]
Loading weights:  62%|██████▏   | 181/291 [00:01<00:00, 178.41it/s]
Loading weights:  69%|██████▊   | 200/291 [00:01<00:00, 177.73it/s]
Loading weights:  76%|███████▌  | 220/291 [00:01<00:00, 176.95it/s]
Loading weights:  82%|████████▏ | 238/291 [00:01<00:00, 174.30it/s]
Loading weights:  88%|████████▊ | 257/291 [00:01<00:00, 177.60it/s]
Loading weights:  95%|█████████▍| 275/291 [00:01<00:00, 177.60it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 175.76it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=931)
+[FIX] Re-initialized lora_A/B in 448 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(6, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 6 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6753.944M (0.04%)
+[TRAIN] step_per_epoch=5, eval_steps=25
+  [SRT] Loaded 6 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 6 sigs, cur=task639_multi_woz_user_utterance_generation, mapping={'task639_multi_woz_user_utterance_generation': 0, 'task1572_samsum_summary': 6, 'task363_sst2_polarity_classification': 5, 'task1290_xsum_summarization': 4, 'task181_outcome_extraction': 3, 'task002_quoref_answer_generation': 2, 'task1510_evalution_relation_extraction': 1}
+[SGWI] Task 6: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 23:38:50,958] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/75 [00:00<?, ?it/s]
  1%|▏         | 1/75 [00:04<05:11,  4.21s/it]
  3%|▎         | 2/75 [00:07<04:45,  3.91s/it]
  4%|▍         | 3/75 [00:11<04:45,  3.97s/it]
  5%|▌         | 4/75 [00:15<04:39,  3.94s/it]
  7%|▋         | 5/75 [00:17<03:43,  3.19s/it]
  8%|▊         | 6/75 [00:21<03:52,  3.36s/it]
  9%|▉         | 7/75 [00:25<04:05,  3.61s/it]
 11%|█         | 8/75 [00:29<04:07,  3.69s/it]
 12%|█▏        | 9/75 [00:33<04:14,  3.85s/it]
 13%|█▎        | 10/75 [00:35<03:25,  3.16s/it]
                                               
{'loss': '3.099', 'grad_norm': '0.5011', 'learning_rate': '5e-05', 'epoch': '2'}
+
 13%|█▎        | 10/75 [00:35<03:25,  3.16s/it]
 15%|█▍        | 11/75 [00:39<03:37,  3.39s/it]
 16%|█▌        | 12/75 [00:43<03:47,  3.61s/it]
 17%|█▋        | 13/75 [00:47<03:50,  3.72s/it]
 19%|█▊        | 14/75 [00:50<03:46,  3.72s/it]
 20%|██        | 15/75 [00:52<03:05,  3.09s/it]
 21%|██▏       | 16/75 [00:56<03:15,  3.31s/it]
 23%|██▎       | 17/75 [01:00<03:19,  3.43s/it]
 24%|██▍       | 18/75 [01:03<03:20,  3.51s/it]
 25%|██▌       | 19/75 [01:07<03:23,  3.64s/it]
 27%|██▋       | 20/75 [01:09<02:47,  3.04s/it]
                                               
{'loss': '2.843', 'grad_norm': '0.9966', 'learning_rate': '5e-05', 'epoch': '4'}
+
 27%|██▋       | 20/75 [01:09<02:47,  3.04s/it]
 28%|██▊       | 21/75 [01:13<02:55,  3.24s/it]
 29%|██▉       | 22/75 [01:17<03:04,  3.48s/it]
 31%|███       | 23/75 [01:20<03:05,  3.56s/it]
 32%|███▏      | 24/75 [01:24<03:08,  3.69s/it]
 33%|███▎      | 25/75 [01:26<02:36,  3.12s/it]
+
Eval @ step 25:   0%|          | 0/18 [00:00<?, ?it/s]
+
Eval @ step 25:   6%|▌         | 1/18 [00:01<00:27,  1.63s/it]
+
Eval @ step 25:  11%|█         | 2/18 [00:03<00:25,  1.60s/it]
+
Eval @ step 25:  17%|█▋        | 3/18 [00:04<00:23,  1.58s/it]
+
Eval @ step 25:  22%|██▏       | 4/18 [00:05<00:15,  1.14s/it]
+
Eval @ step 25:  28%|██▊       | 5/18 [00:05<00:11,  1.17it/s]
+
Eval @ step 25:  33%|███▎      | 6/18 [00:07<00:12,  1.08s/it]
+
Eval @ step 25:  39%|███▉      | 7/18 [00:08<00:13,  1.23s/it]
+
Eval @ step 25:  44%|████▍     | 8/18 [00:10<00:14,  1.40s/it]
+
Eval @ step 25:  50%|█████     | 9/18 [00:10<00:09,  1.07s/it]
+
Eval @ step 25:  56%|█████▌    | 10/18 [00:12<00:09,  1.21s/it]
+
Eval @ step 25:  61%|██████    | 11/18 [00:13<00:09,  1.29s/it]
+
Eval @ step 25:  67%|██████▋   | 12/18 [00:15<00:08,  1.37s/it]
+
Eval @ step 25:  72%|███████▏  | 13/18 [00:16<00:07,  1.40s/it]
+
Eval @ step 25:  78%|███████▊  | 14/18 [00:18<00:05,  1.43s/it]
+
Eval @ step 25:  83%|████████▎ | 15/18 [00:19<00:04,  1.45s/it]
+
Eval @ step 25:  89%|████████▉ | 16/18 [00:20<00:02,  1.16s/it]
+
Eval @ step 25:  94%|█████████▍| 17/18 [00:21<00:01,  1.28s/it]
+
Eval @ step 25: 100%|██████████| 18/18 [00:23<00:00,  1.36s/it]
+
                                                                 [EVAL] ★ New best eval_rougeL=8.7919 (step=25) — snapshot saved in memory
+
                                               
{'eval_rougeL': '8.792', 'epoch': '5'}
+
 33%|███▎      | 25/75 [01:50<02:36,  3.12s/it]
 35%|███▍      | 26/75 [01:54<08:29, 10.41s/it]
 36%|███▌      | 27/75 [01:58<06:48,  8.50s/it]
 37%|███▋      | 28/75 [02:02<05:35,  7.13s/it]
 39%|███▊      | 29/75 [02:05<04:43,  6.15s/it]
 40%|████      | 30/75 [02:07<03:39,  4.88s/it]
                                               
{'loss': '2.589', 'grad_norm': '1.599', 'learning_rate': '5e-05', 'epoch': '6'}
+
 40%|████      | 30/75 [02:07<03:39,  4.88s/it]
 41%|████▏     | 31/75 [02:11<03:19,  4.53s/it]
 43%|████▎     | 32/75 [02:15<03:04,  4.28s/it]
 44%|████▍     | 33/75 [02:19<02:57,  4.23s/it]
 45%|████▌     | 34/75 [02:23<02:46,  4.07s/it]
 47%|████▋     | 35/75 [02:24<02:13,  3.34s/it]
 48%|████▊     | 36/75 [02:28<02:14,  3.45s/it]
 49%|████▉     | 37/75 [02:32<02:18,  3.64s/it]
 51%|█████     | 38/75 [02:36<02:17,  3.72s/it]
 52%|█████▏    | 39/75 [02:40<02:15,  3.75s/it]
 53%|█████▎    | 40/75 [02:41<01:49,  3.13s/it]
                                               
{'loss': '2.464', 'grad_norm': '0.9267', 'learning_rate': '5e-05', 'epoch': '8'}
+
 53%|█████▎    | 40/75 [02:41<01:49,  3.13s/it]
 55%|█████▍    | 41/75 [02:45<01:53,  3.34s/it]
 56%|█████▌    | 42/75 [02:49<01:55,  3.50s/it]
 57%|█████▋    | 43/75 [02:53<01:58,  3.70s/it]
 59%|█████▊    | 44/75 [02:57<01:54,  3.70s/it]
 60%|██████    | 45/75 [02:59<01:32,  3.08s/it]
 61%|██████▏   | 46/75 [03:03<01:36,  3.34s/it]
 63%|██████▎   | 47/75 [03:06<01:38,  3.50s/it]
 64%|██████▍   | 48/75 [03:10<01:36,  3.57s/it]
 65%|██████▌   | 49/75 [03:14<01:36,  3.70s/it]
 67%|██████▋   | 50/75 [03:16<01:17,  3.08s/it]
                                               
{'loss': '2.344', 'grad_norm': '1.196', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 50/75 [03:16<01:17,  3.08s/it]
+
Eval @ step 50:   0%|          | 0/18 [00:00<?, ?it/s]
+
Eval @ step 50:   6%|▌         | 1/18 [00:00<00:03,  4.72it/s]
+
Eval @ step 50:  11%|█         | 2/18 [00:00<00:02,  6.32it/s]
+
Eval @ step 50:  17%|█▋        | 3/18 [00:00<00:02,  5.93it/s]
+
Eval @ step 50:  22%|██▏       | 4/18 [00:00<00:02,  6.19it/s]
+
Eval @ step 50:  28%|██▊       | 5/18 [00:00<00:02,  5.94it/s]
+
Eval @ step 50:  33%|███▎      | 6/18 [00:01<00:02,  5.79it/s]
+
Eval @ step 50:  39%|███▉      | 7/18 [00:01<00:01,  5.70it/s]
+
Eval @ step 50:  44%|████▍     | 8/18 [00:01<00:01,  5.65it/s]
+
Eval @ step 50:  50%|█████     | 9/18 [00:01<00:02,  3.55it/s]
+
Eval @ step 50:  56%|█████▌    | 10/18 [00:02<00:02,  3.48it/s]
+
Eval @ step 50:  61%|██████    | 11/18 [00:02<00:01,  4.23it/s]
+
Eval @ step 50:  67%|██████▋   | 12/18 [00:02<00:01,  4.54it/s]
+
Eval @ step 50:  72%|███████▏  | 13/18 [00:02<00:01,  4.53it/s]
+
Eval @ step 50:  78%|███████▊  | 14/18 [00:02<00:00,  4.61it/s]
+
Eval @ step 50:  83%|████████▎ | 15/18 [00:03<00:00,  4.63it/s]
+
Eval @ step 50:  89%|████████▉ | 16/18 [00:03<00:00,  3.55it/s]
+
Eval @ step 50:  94%|█████████▍| 17/18 [00:03<00:00,  3.97it/s]
+
Eval @ step 50: 100%|██████████| 18/18 [00:03<00:00,  4.16it/s]
+
                                                                 [EVAL] ★ New best eval_rougeL=15.2912 (step=50) — snapshot saved in memory
+
                                               
{'eval_rougeL': '15.29', 'epoch': '10'}
+
 67%|██████▋   | 50/75 [03:20<01:17,  3.08s/it]
 68%|██████▊   | 51/75 [03:24<01:49,  4.54s/it]
 69%|██████▉   | 52/75 [03:28<01:40,  4.35s/it]
 71%|███████   | 53/75 [03:32<01:32,  4.21s/it]
 72%|███████▏  | 54/75 [03:36<01:29,  4.28s/it]
 73%|███████▎  | 55/75 [03:38<01:09,  3.49s/it]
 75%|███████▍  | 56/75 [03:41<01:07,  3.57s/it]
 76%|███████▌  | 57/75 [03:45<01:07,  3.73s/it]
 77%|███████▋  | 58/75 [03:49<01:04,  3.78s/it]
 79%|███████▊  | 59/75 [03:53<01:01,  3.82s/it]
 80%|████████  | 60/75 [03:55<00:48,  3.22s/it]
                                               
{'loss': '2.286', 'grad_norm': '1.395', 'learning_rate': '5e-05', 'epoch': '12'}
+
 80%|████████  | 60/75 [03:55<00:48,  3.22s/it]
 81%|████████▏ | 61/75 [03:59<00:48,  3.47s/it]
 83%|████████▎ | 62/75 [04:03<00:47,  3.69s/it]
 84%|████████▍ | 63/75 [04:07<00:45,  3.79s/it]
 85%|████████▌ | 64/75 [04:11<00:42,  3.83s/it]
 87%|████████▋ | 65/75 [04:13<00:32,  3.22s/it]
 88%|████████▊ | 66/75 [04:17<00:31,  3.46s/it]
 89%|████████▉ | 67/75 [04:21<00:28,  3.55s/it]
 91%|█████████ | 68/75 [04:25<00:25,  3.67s/it]
 92%|█████████▏| 69/75 [04:29<00:22,  3.75s/it]
 93%|█████████▎| 70/75 [04:30<00:15,  3.12s/it]
                                               
{'loss': '2.184', 'grad_norm': '1.845', 'learning_rate': '5e-05', 'epoch': '14'}
+
 93%|█████████▎| 70/75 [04:30<00:15,  3.12s/it]
 95%|█████████▍| 71/75 [04:34<00:13,  3.32s/it]
 96%|█████████▌| 72/75 [04:38<00:10,  3.48s/it]
 97%|█████████▋| 73/75 [04:42<00:07,  3.57s/it]
 99%|█████████▊| 74/75 [04:46<00:03,  3.70s/it]
100%|██████████| 75/75 [04:47<00:00,  3.08s/it]
+
Eval @ step 75:   0%|          | 0/18 [00:00<?, ?it/s]
+
Eval @ step 75:   6%|▌         | 1/18 [00:00<00:08,  1.95it/s]
+
Eval @ step 75:  11%|█         | 2/18 [00:00<00:04,  3.55it/s]
+
Eval @ step 75:  17%|█▋        | 3/18 [00:01<00:05,  2.79it/s]
+
Eval @ step 75:  22%|██▏       | 4/18 [00:02<00:08,  1.71it/s]
+
Eval @ step 75:  28%|██▊       | 5/18 [00:02<00:05,  2.39it/s]
+
Eval @ step 75:  33%|███▎      | 6/18 [00:02<00:05,  2.39it/s]
+
Eval @ step 75:  39%|███▉      | 7/18 [00:02<00:03,  3.11it/s]
+
Eval @ step 75:  44%|████▍     | 8/18 [00:03<00:03,  2.83it/s]
+
Eval @ step 75:  50%|█████     | 9/18 [00:03<00:03,  2.54it/s]
+
Eval @ step 75:  56%|█████▌    | 10/18 [00:03<00:03,  2.55it/s]
+
Eval @ step 75:  61%|██████    | 11/18 [00:04<00:02,  2.61it/s]
+
Eval @ step 75:  67%|██████▋   | 12/18 [00:04<00:02,  2.52it/s]
+
Eval @ step 75:  72%|███████▏  | 13/18 [00:04<00:01,  3.19it/s]
+
Eval @ step 75:  78%|███████▊  | 14/18 [00:05<00:01,  3.92it/s]
+
Eval @ step 75:  83%|████████▎ | 15/18 [00:05<00:01,  2.93it/s]
+
Eval @ step 75:  89%|████████▉ | 16/18 [00:06<00:00,  2.42it/s]
+
Eval @ step 75:  94%|█████████▍| 17/18 [00:06<00:00,  3.08it/s]
+
Eval @ step 75: 100%|██████████| 18/18 [00:06<00:00,  2.85it/s]
+
                                                                 [EVAL] eval_rougeL=13.5595 (best=15.2912)
+
                                               
{'eval_rougeL': '13.56', 'epoch': '15'}
+
100%|██████████| 75/75 [04:54<00:00,  3.08s/it]
                                               
{'train_runtime': '294.6', 'train_samples_per_second': '7.229', 'train_steps_per_second': '0.255', 'train_loss': '2.519', 'epoch': '15'}
+
100%|██████████| 75/75 [04:54<00:00,  3.08s/it]
100%|██████████| 75/75 [04:54<00:00,  3.93s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               =  6447750GF
+  train_loss               =     2.5188
+  train_runtime            = 0:04:54.62
+  train_samples            =        142
+  train_samples_per_second =      7.229
+  train_steps_per_second   =      0.255
+  [EVAL] Restored best model (eval_rougeL=15.2912, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   7%|▋         | 20/291 [00:00<00:01, 184.48it/s]
Loading weights:  14%|█▎        | 40/291 [00:00<00:01, 169.27it/s]
Loading weights:  22%|██▏       | 63/291 [00:00<00:01, 194.47it/s]
Loading weights:  29%|██▊       | 83/291 [00:00<00:01, 185.86it/s]
Loading weights:  35%|███▌      | 102/291 [00:00<00:01, 180.21it/s]
Loading weights:  42%|████▏     | 121/291 [00:00<00:00, 176.10it/s]
Loading weights:  48%|████▊     | 139/291 [00:00<00:00, 175.97it/s]
Loading weights:  55%|█████▍    | 159/291 [00:00<00:00, 182.31it/s]
Loading weights:  62%|██████▏   | 180/291 [00:00<00:00, 189.90it/s]
Loading weights:  69%|██████▊   | 200/291 [00:01<00:00, 187.46it/s]
Loading weights:  75%|███████▌  | 219/291 [00:01<00:00, 182.84it/s]
Loading weights:  82%|████████▏ | 238/291 [00:01<00:00, 181.73it/s]
Loading weights:  88%|████████▊ | 257/291 [00:01<00:00, 175.95it/s]
Loading weights:  95%|█████████▍| 275/291 [00:01<00:00, 169.99it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 180.20it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (142 samples, batch_size=1)
+  [SRT] Task 'task639_multi_woz_user_utterance_generation': PaR=76.6, shrinkage=ridge, n=142
+  [SRT] Wired router: 7 sigs, cur=task639_multi_woz_user_utterance_generation, mapping={'task639_multi_woz_user_utterance_generation': 0, 'task1572_samsum_summary': 6, 'task363_sst2_polarity_classification': 5, 'task1290_xsum_summarization': 4, 'task181_outcome_extraction': 3, 'task002_quoref_answer_generation': 2, 'task1510_evalution_relation_extraction': 1}
+[SRT] Router: 7 tasks, avg_PaR=76.6, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation']
+  [SRT-INIT] task_id_to_idx={'task639_multi_woz_user_utterance_generation': 0, 'task1572_samsum_summary': 6, 'task363_sst2_polarity_classification': 5, 'task1290_xsum_summarization': 4, 'task181_outcome_extraction': 3, 'task002_quoref_answer_generation': 2, 'task1510_evalution_relation_extraction': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/481 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=7
+
Continual predict:   0%|          | 1/481 [00:01<13:53,  1.74s/it]
Continual predict:   0%|          | 2/481 [00:04<17:14,  2.16s/it]
Continual predict:   1%|          | 3/481 [00:04<12:11,  1.53s/it]
Continual predict:   1%|          | 4/481 [00:05<09:01,  1.14s/it]
Continual predict:   1%|          | 5/481 [00:07<10:58,  1.38s/it]
Continual predict:   1%|          | 6/481 [00:12<19:49,  2.50s/it]
Continual predict:   1%|▏         | 7/481 [00:15<21:44,  2.75s/it]
Continual predict:   2%|▏         | 8/481 [00:17<19:51,  2.52s/it]
Continual predict:   2%|▏         | 9/481 [00:18<17:47,  2.26s/it]
Continual predict:   2%|▏         | 10/481 [00:19<14:23,  1.83s/it]
Continual predict:   2%|▏         | 11/481 [00:22<16:27,  2.10s/it]
Continual predict:   2%|▏         | 12/481 [00:25<18:45,  2.40s/it]
Continual predict:   3%|▎         | 13/481 [00:26<15:52,  2.04s/it]
Continual predict:   3%|▎         | 14/481 [00:29<16:34,  2.13s/it]
Continual predict:   3%|▎         | 15/481 [00:31<16:41,  2.15s/it]
Continual predict:   3%|▎         | 16/481 [00:33<17:32,  2.26s/it]
Continual predict:   4%|▎         | 17/481 [00:41<29:56,  3.87s/it]
Continual predict:   4%|▎         | 18/481 [00:43<25:40,  3.33s/it]
Continual predict:   4%|▍         | 19/481 [00:46<23:40,  3.08s/it]
Continual predict:   4%|▍         | 20/481 [00:47<19:39,  2.56s/it]
Continual predict:   4%|▍         | 21/481 [00:47<14:06,  1.84s/it]
Continual predict:   5%|▍         | 22/481 [00:47<10:10,  1.33s/it]
Continual predict:   5%|▍         | 23/481 [00:47<07:28,  1.02it/s]
Continual predict:   5%|▍         | 24/481 [00:48<05:32,  1.37it/s]
Continual predict:   5%|▌         | 25/481 [00:48<04:12,  1.81it/s]
Continual predict:   5%|▌         | 26/481 [00:48<03:15,  2.33it/s]
Continual predict:   6%|▌         | 27/481 [00:48<02:36,  2.91it/s]
Continual predict:   6%|▌         | 28/481 [00:48<02:08,  3.53it/s]
Continual predict:   6%|▌         | 29/481 [00:48<01:48,  4.15it/s]
Continual predict:   6%|▌         | 30/481 [00:48<01:37,  4.64it/s]
Continual predict:   6%|▋         | 31/481 [00:49<01:27,  5.15it/s]
Continual predict:   7%|▋         | 32/481 [00:49<01:20,  5.60it/s]
Continual predict:   7%|▋         | 33/481 [00:49<01:17,  5.79it/s]
Continual predict:   7%|▋         | 34/481 [00:49<01:15,  5.96it/s]
Continual predict:   7%|▋         | 35/481 [00:49<01:13,  6.08it/s]
Continual predict:   7%|▋         | 36/481 [00:49<01:10,  6.35it/s]
Continual predict:   8%|▊         | 37/481 [00:49<01:09,  6.36it/s]
Continual predict:   8%|▊         | 38/481 [00:50<01:07,  6.52it/s]
Continual predict:   8%|▊         | 39/481 [00:50<01:06,  6.67it/s]
Continual predict:   8%|▊         | 40/481 [00:50<01:05,  6.78it/s]
Continual predict:   9%|▊         | 41/481 [00:50<01:06,  6.67it/s]
Continual predict:   9%|▊         | 42/481 [00:50<01:05,  6.73it/s]
Continual predict:   9%|▉         | 43/481 [00:50<01:09,  6.26it/s]
Continual predict:   9%|▉         | 44/481 [00:51<01:10,  6.23it/s]
Continual predict:   9%|▉         | 45/481 [00:51<01:15,  5.78it/s]
Continual predict:  10%|▉         | 46/481 [00:51<01:11,  6.08it/s]
Continual predict:  10%|▉         | 47/481 [00:51<01:09,  6.26it/s]
Continual predict:  10%|▉         | 48/481 [00:51<01:11,  6.09it/s]
Continual predict:  10%|█         | 49/481 [00:51<01:09,  6.19it/s]
Continual predict:  10%|█         | 50/481 [00:52<01:07,  6.40it/s]
Continual predict:  11%|█         | 51/481 [00:52<01:06,  6.50it/s]
Continual predict:  11%|█         | 52/481 [00:52<01:04,  6.62it/s]
Continual predict:  11%|█         | 53/481 [00:52<01:03,  6.72it/s]
Continual predict:  11%|█         | 54/481 [00:52<01:04,  6.62it/s]
Continual predict:  11%|█▏        | 55/481 [00:52<01:03,  6.75it/s]
Continual predict:  12%|█▏        | 56/481 [00:52<01:04,  6.62it/s]
Continual predict:  12%|█▏        | 57/481 [00:53<01:03,  6.73it/s]
Continual predict:  12%|█▏        | 58/481 [00:53<01:02,  6.80it/s]
Continual predict:  12%|█▏        | 59/481 [00:53<01:01,  6.85it/s]
Continual predict:  12%|█▏        | 60/481 [00:53<01:03,  6.68it/s]
Continual predict:  13%|█▎        | 61/481 [00:53<01:01,  6.78it/s]
Continual predict:  13%|█▎        | 62/481 [00:53<01:01,  6.84it/s]
Continual predict:  13%|█▎        | 63/481 [00:53<01:02,  6.71it/s]
Continual predict:  13%|█▎        | 64/481 [00:54<01:01,  6.78it/s]
Continual predict:  14%|█▎        | 65/481 [00:54<01:00,  6.84it/s]
Continual predict:  14%|█▎        | 66/481 [00:54<01:00,  6.88it/s]
Continual predict:  14%|█▍        | 67/481 [00:54<00:59,  6.92it/s]
Continual predict:  14%|█▍        | 68/481 [00:54<00:59,  6.95it/s]
Continual predict:  14%|█▍        | 69/481 [00:54<00:59,  6.98it/s]
Continual predict:  15%|█▍        | 70/481 [00:54<00:58,  6.99it/s]
Continual predict:  15%|█▍        | 71/481 [00:55<00:58,  6.99it/s]
Continual predict:  15%|█▍        | 72/481 [00:55<01:00,  6.80it/s]
Continual predict:  15%|█▌        | 73/481 [00:55<00:59,  6.88it/s]
Continual predict:  15%|█▌        | 74/481 [00:55<01:00,  6.71it/s]
Continual predict:  16%|█▌        | 75/481 [00:55<01:00,  6.72it/s]
Continual predict:  16%|█▌        | 76/481 [00:55<00:59,  6.82it/s]
Continual predict:  16%|█▌        | 77/481 [00:55<00:58,  6.86it/s]
Continual predict:  16%|█▌        | 78/481 [00:56<00:58,  6.92it/s]
Continual predict:  16%|█▋        | 79/481 [00:56<00:58,  6.93it/s]
Continual predict:  17%|█▋        | 80/481 [00:56<00:57,  6.96it/s]
Continual predict:  17%|█▋        | 81/481 [00:56<00:57,  6.95it/s]
Continual predict:  17%|█▋        | 82/481 [00:56<00:57,  6.97it/s]
Continual predict:  17%|█▋        | 83/481 [00:56<00:57,  6.97it/s]
Continual predict:  17%|█▋        | 84/481 [00:56<00:58,  6.78it/s]
Continual predict:  18%|█▊        | 85/481 [00:57<00:58,  6.82it/s]
Continual predict:  18%|█▊        | 86/481 [00:57<00:57,  6.87it/s]
Continual predict:  18%|█▊        | 87/481 [00:57<00:57,  6.86it/s]
Continual predict:  18%|█▊        | 88/481 [00:57<00:56,  6.90it/s]
Continual predict:  19%|█▊        | 89/481 [00:57<00:56,  6.94it/s]
Continual predict:  19%|█▊        | 90/481 [00:57<00:56,  6.97it/s]
Continual predict:  19%|█▉        | 91/481 [00:57<00:55,  6.97it/s]
Continual predict:  19%|█▉        | 92/481 [00:58<00:56,  6.93it/s]
Continual predict:  19%|█▉        | 93/481 [00:58<00:57,  6.76it/s]
Continual predict:  20%|█▉        | 94/481 [00:58<00:58,  6.66it/s]
Continual predict:  20%|█▉        | 95/481 [00:58<00:57,  6.74it/s]
Continual predict:  20%|█▉        | 96/481 [00:58<00:56,  6.79it/s]
Continual predict:  20%|██        | 97/481 [00:58<00:56,  6.84it/s]
Continual predict:  20%|██        | 98/481 [00:59<00:57,  6.71it/s]
Continual predict:  21%|██        | 99/481 [00:59<00:57,  6.62it/s]
Continual predict:  21%|██        | 100/481 [00:59<00:58,  6.55it/s]
Continual predict:  21%|██        | 101/481 [00:59<00:56,  6.68it/s]
Continual predict:  21%|██        | 102/481 [00:59<00:55,  6.78it/s]
Continual predict:  21%|██▏       | 103/481 [00:59<00:55,  6.85it/s]
Continual predict:  22%|██▏       | 104/481 [00:59<00:54,  6.88it/s]
Continual predict:  22%|██▏       | 105/481 [01:00<00:55,  6.74it/s]
Continual predict:  22%|██▏       | 106/481 [01:00<00:54,  6.83it/s]
Continual predict:  22%|██▏       | 107/481 [01:00<00:54,  6.90it/s]
Continual predict:  22%|██▏       | 108/481 [01:00<00:53,  6.92it/s]
Continual predict:  23%|██▎       | 109/481 [01:00<00:53,  6.92it/s]
Continual predict:  23%|██▎       | 110/481 [01:00<00:53,  6.91it/s]
Continual predict:  23%|██▎       | 111/481 [01:00<00:53,  6.93it/s]
Continual predict:  23%|██▎       | 112/481 [01:01<00:53,  6.94it/s]
Continual predict:  23%|██▎       | 113/481 [01:01<00:53,  6.94it/s]
Continual predict:  24%|██▎       | 114/481 [01:01<00:53,  6.85it/s]
Continual predict:  24%|██▍       | 115/481 [01:01<00:55,  6.61it/s]
Continual predict:  24%|██▍       | 116/481 [01:01<00:55,  6.60it/s]
Continual predict:  24%|██▍       | 117/481 [01:01<00:55,  6.54it/s]
Continual predict:  25%|██▍       | 118/481 [01:01<00:55,  6.49it/s]
Continual predict:  25%|██▍       | 119/481 [01:02<00:54,  6.62it/s]
Continual predict:  25%|██▍       | 120/481 [01:02<00:53,  6.72it/s]
Continual predict:  25%|██▌       | 121/481 [01:06<08:46,  1.46s/it]
Continual predict:  25%|██▌       | 122/481 [01:11<13:42,  2.29s/it]
Continual predict:  26%|██▌       | 123/481 [01:16<19:02,  3.19s/it]
Continual predict:  26%|██▌       | 124/481 [01:18<17:29,  2.94s/it]
Continual predict:  26%|██▌       | 125/481 [01:23<20:44,  3.50s/it]
Continual predict:  26%|██▌       | 126/481 [01:28<22:56,  3.88s/it]
Continual predict:  26%|██▋       | 127/481 [01:29<18:41,  3.17s/it]
Continual predict:  27%|██▋       | 128/481 [01:32<17:53,  3.04s/it]
Continual predict:  27%|██▋       | 129/481 [01:35<18:26,  3.14s/it]
Continual predict:  27%|██▋       | 130/481 [01:38<18:10,  3.11s/it]
Continual predict:  27%|██▋       | 131/481 [01:41<16:34,  2.84s/it]
Continual predict:  27%|██▋       | 132/481 [01:43<16:13,  2.79s/it]
Continual predict:  28%|██▊       | 133/481 [01:46<16:48,  2.90s/it]
Continual predict:  28%|██▊       | 134/481 [01:52<22:09,  3.83s/it]
Continual predict:  28%|██▊       | 135/481 [01:57<23:43,  4.11s/it]
Continual predict:  28%|██▊       | 136/481 [02:00<22:07,  3.85s/it]
Continual predict:  28%|██▊       | 137/481 [02:03<19:16,  3.36s/it]
Continual predict:  29%|██▊       | 138/481 [02:03<14:43,  2.58s/it]
Continual predict:  29%|██▉       | 139/481 [02:05<12:25,  2.18s/it]
Continual predict:  29%|██▉       | 140/481 [02:07<13:17,  2.34s/it]
Continual predict:  29%|██▉       | 141/481 [02:14<20:26,  3.61s/it]
Continual predict:  30%|██▉       | 142/481 [02:19<21:58,  3.89s/it]
Continual predict:  30%|██▉       | 143/481 [02:22<20:52,  3.70s/it]
Continual predict:  30%|██▉       | 144/481 [02:31<30:26,  5.42s/it]
Continual predict:  30%|███       | 145/481 [02:36<28:36,  5.11s/it]
Continual predict:  30%|███       | 146/481 [02:37<23:06,  4.14s/it]
Continual predict:  31%|███       | 147/481 [02:42<23:25,  4.21s/it]
Continual predict:  31%|███       | 148/481 [02:44<19:55,  3.59s/it]
Continual predict:  31%|███       | 149/481 [02:47<19:12,  3.47s/it]
Continual predict:  31%|███       | 150/481 [02:49<16:32,  3.00s/it]
Continual predict:  31%|███▏      | 151/481 [02:52<16:35,  3.02s/it]
Continual predict:  32%|███▏      | 152/481 [02:58<20:52,  3.81s/it]
Continual predict:  32%|███▏      | 153/481 [03:00<18:35,  3.40s/it]
Continual predict:  32%|███▏      | 154/481 [03:05<20:49,  3.82s/it]
Continual predict:  32%|███▏      | 155/481 [03:09<20:23,  3.75s/it]
Continual predict:  32%|███▏      | 156/481 [03:12<19:42,  3.64s/it]
Continual predict:  33%|███▎      | 157/481 [03:16<19:57,  3.70s/it]
Continual predict:  33%|███▎      | 158/481 [03:17<15:26,  2.87s/it]
Continual predict:  33%|███▎      | 159/481 [03:24<21:58,  4.10s/it]
Continual predict:  33%|███▎      | 160/481 [03:26<18:58,  3.55s/it]
Continual predict:  33%|███▎      | 161/481 [03:28<17:12,  3.23s/it]
Continual predict:  34%|███▎      | 162/481 [03:33<19:41,  3.70s/it]
Continual predict:  34%|███▍      | 163/481 [03:35<17:00,  3.21s/it]
Continual predict:  34%|███▍      | 164/481 [03:38<15:26,  2.92s/it]
Continual predict:  34%|███▍      | 165/481 [03:41<15:43,  2.98s/it]
Continual predict:  35%|███▍      | 166/481 [03:43<14:21,  2.73s/it]
Continual predict:  35%|███▍      | 167/481 [03:44<12:04,  2.31s/it]
Continual predict:  35%|███▍      | 168/481 [03:46<11:55,  2.29s/it]
Continual predict:  35%|███▌      | 169/481 [03:48<11:00,  2.12s/it]
Continual predict:  35%|███▌      | 170/481 [03:51<12:24,  2.40s/it]
Continual predict:  36%|███▌      | 171/481 [03:53<11:00,  2.13s/it]
Continual predict:  36%|███▌      | 172/481 [03:55<11:39,  2.26s/it]
Continual predict:  36%|███▌      | 173/481 [03:59<14:11,  2.77s/it]
Continual predict:  36%|███▌      | 174/481 [04:00<11:41,  2.28s/it]
Continual predict:  36%|███▋      | 175/481 [04:04<13:26,  2.64s/it]
Continual predict:  37%|███▋      | 176/481 [04:07<13:43,  2.70s/it]
Continual predict:  37%|███▋      | 177/481 [04:09<13:41,  2.70s/it]
Continual predict:  37%|███▋      | 178/481 [04:11<11:26,  2.27s/it]
Continual predict:  37%|███▋      | 179/481 [04:12<09:43,  1.93s/it]
Continual predict:  37%|███▋      | 180/481 [04:13<08:45,  1.74s/it]
Continual predict:  38%|███▊      | 181/481 [04:14<08:04,  1.61s/it]
Continual predict:  38%|███▊      | 182/481 [04:19<12:38,  2.54s/it]
Continual predict:  38%|███▊      | 183/481 [04:19<09:09,  1.84s/it]
Continual predict:  38%|███▊      | 184/481 [04:20<08:00,  1.62s/it]
Continual predict:  38%|███▊      | 185/481 [04:24<10:34,  2.14s/it]
Continual predict:  39%|███▊      | 186/481 [04:25<08:50,  1.80s/it]
Continual predict:  39%|███▉      | 187/481 [04:28<10:58,  2.24s/it]
Continual predict:  39%|███▉      | 188/481 [04:31<11:36,  2.38s/it]
Continual predict:  39%|███▉      | 189/481 [04:36<15:19,  3.15s/it]
Continual predict:  40%|███▉      | 190/481 [04:39<15:24,  3.18s/it]
Continual predict:  40%|███▉      | 191/481 [04:40<12:37,  2.61s/it]
Continual predict:  40%|███▉      | 192/481 [04:44<13:34,  2.82s/it]
Continual predict:  40%|████      | 193/481 [04:48<15:58,  3.33s/it]
Continual predict:  40%|████      | 194/481 [04:53<17:59,  3.76s/it]
Continual predict:  41%|████      | 195/481 [04:58<20:06,  4.22s/it]
Continual predict:  41%|████      | 196/481 [05:01<18:46,  3.95s/it]
Continual predict:  41%|████      | 197/481 [05:05<17:46,  3.76s/it]
Continual predict:  41%|████      | 198/481 [05:12<22:46,  4.83s/it]
Continual predict:  41%|████▏     | 199/481 [05:13<17:25,  3.71s/it]
Continual predict:  42%|████▏     | 200/481 [05:15<14:11,  3.03s/it]
Continual predict:  42%|████▏     | 201/481 [05:17<13:26,  2.88s/it]
Continual predict:  42%|████▏     | 202/481 [05:20<13:52,  2.98s/it]
Continual predict:  42%|████▏     | 203/481 [05:25<16:48,  3.63s/it]
Continual predict:  42%|████▏     | 204/481 [05:29<15:55,  3.45s/it]
Continual predict:  43%|████▎     | 205/481 [05:35<20:10,  4.38s/it]
Continual predict:  43%|████▎     | 206/481 [05:38<17:49,  3.89s/it]
Continual predict:  43%|████▎     | 207/481 [05:42<17:31,  3.84s/it]
Continual predict:  43%|████▎     | 208/481 [05:48<20:35,  4.52s/it]
Continual predict:  43%|████▎     | 209/481 [05:50<17:59,  3.97s/it]
Continual predict:  44%|████▎     | 210/481 [05:52<14:47,  3.28s/it]
Continual predict:  44%|████▍     | 211/481 [05:55<13:47,  3.07s/it]
Continual predict:  44%|████▍     | 212/481 [05:58<13:46,  3.07s/it]
Continual predict:  44%|████▍     | 213/481 [06:01<14:36,  3.27s/it]
Continual predict:  44%|████▍     | 214/481 [06:05<14:58,  3.36s/it]
Continual predict:  45%|████▍     | 215/481 [06:06<11:25,  2.58s/it]
Continual predict:  45%|████▍     | 216/481 [06:09<12:02,  2.73s/it]
Continual predict:  45%|████▌     | 217/481 [06:11<11:03,  2.51s/it]
Continual predict:  45%|████▌     | 218/481 [06:15<12:49,  2.93s/it]
Continual predict:  46%|████▌     | 219/481 [06:18<13:05,  3.00s/it]
Continual predict:  46%|████▌     | 220/481 [06:22<14:25,  3.31s/it]
Continual predict:  46%|████▌     | 221/481 [06:22<10:19,  2.38s/it]
Continual predict:  46%|████▌     | 222/481 [06:22<07:33,  1.75s/it]
Continual predict:  46%|████▋     | 223/481 [06:23<05:31,  1.29s/it]
Continual predict:  47%|████▋     | 224/481 [06:23<04:07,  1.04it/s]
Continual predict:  47%|████▋     | 225/481 [06:23<03:08,  1.36it/s]
Continual predict:  47%|████▋     | 226/481 [06:23<02:27,  1.73it/s]
Continual predict:  47%|████▋     | 227/481 [06:23<02:02,  2.07it/s]
Continual predict:  47%|████▋     | 228/481 [06:24<01:44,  2.41it/s]
Continual predict:  48%|████▊     | 229/481 [06:24<01:30,  2.79it/s]
Continual predict:  48%|████▊     | 230/481 [06:24<01:34,  2.67it/s]
Continual predict:  48%|████▊     | 231/481 [06:25<01:21,  3.08it/s]
Continual predict:  48%|████▊     | 232/481 [06:25<01:12,  3.46it/s]
Continual predict:  48%|████▊     | 233/481 [06:25<01:05,  3.78it/s]
Continual predict:  49%|████▊     | 234/481 [06:25<01:01,  4.05it/s]
Continual predict:  49%|████▉     | 235/481 [06:25<00:57,  4.26it/s]
Continual predict:  49%|████▉     | 236/481 [06:26<00:55,  4.42it/s]
Continual predict:  49%|████▉     | 237/481 [06:26<00:53,  4.54it/s]
Continual predict:  49%|████▉     | 238/481 [06:26<01:07,  3.61it/s]
Continual predict:  50%|████▉     | 239/481 [06:26<01:02,  3.90it/s]
Continual predict:  50%|████▉     | 240/481 [06:27<01:08,  3.53it/s]
Continual predict:  50%|█████     | 241/481 [06:27<01:02,  3.85it/s]
Continual predict:  50%|█████     | 242/481 [06:27<00:58,  4.10it/s]
Continual predict:  51%|█████     | 243/481 [06:28<01:03,  3.75it/s]
Continual predict:  51%|█████     | 244/481 [06:28<01:10,  3.38it/s]
Continual predict:  51%|█████     | 245/481 [06:28<01:03,  3.71it/s]
Continual predict:  51%|█████     | 246/481 [06:28<01:00,  3.88it/s]
Continual predict:  51%|█████▏    | 247/481 [06:29<01:00,  3.85it/s]
Continual predict:  52%|█████▏    | 248/481 [06:29<00:56,  4.11it/s]
Continual predict:  52%|█████▏    | 249/481 [06:29<00:57,  4.03it/s]
Continual predict:  52%|█████▏    | 250/481 [06:29<00:54,  4.25it/s]
Continual predict:  52%|█████▏    | 251/481 [06:30<00:55,  4.13it/s]
Continual predict:  52%|█████▏    | 252/481 [06:31<02:03,  1.86it/s]
Continual predict:  53%|█████▎    | 253/481 [06:31<01:43,  2.20it/s]
Continual predict:  53%|█████▎    | 254/481 [06:31<01:26,  2.64it/s]
Continual predict:  53%|█████▎    | 255/481 [06:31<01:14,  3.05it/s]
Continual predict:  53%|█████▎    | 256/481 [06:32<01:05,  3.43it/s]
Continual predict:  53%|█████▎    | 257/481 [06:32<01:33,  2.39it/s]
Continual predict:  54%|█████▎    | 258/481 [06:33<01:19,  2.82it/s]
Continual predict:  54%|█████▍    | 259/481 [06:33<01:08,  3.22it/s]
Continual predict:  54%|█████▍    | 260/481 [06:33<01:01,  3.58it/s]
Continual predict:  54%|█████▍    | 261/481 [06:34<01:19,  2.77it/s]
Continual predict:  54%|█████▍    | 262/481 [06:34<01:10,  3.12it/s]
Continual predict:  55%|█████▍    | 263/481 [06:34<01:05,  3.31it/s]
Continual predict:  55%|█████▍    | 264/481 [06:35<02:05,  1.74it/s]
Continual predict:  55%|█████▌    | 265/481 [06:36<02:01,  1.78it/s]
Continual predict:  55%|█████▌    | 266/481 [06:36<01:58,  1.81it/s]
Continual predict:  56%|█████▌    | 267/481 [06:37<02:01,  1.76it/s]
Continual predict:  56%|█████▌    | 268/481 [06:38<02:16,  1.56it/s]
Continual predict:  56%|█████▌    | 269/481 [06:38<02:13,  1.59it/s]
Continual predict:  56%|█████▌    | 270/481 [06:39<02:12,  1.59it/s]
Continual predict:  56%|█████▋    | 271/481 [06:40<02:19,  1.50it/s]
Continual predict:  57%|█████▋    | 272/481 [06:41<02:33,  1.36it/s]
Continual predict:  57%|█████▋    | 273/481 [06:41<02:26,  1.42it/s]
Continual predict:  57%|█████▋    | 274/481 [06:42<02:22,  1.46it/s]
Continual predict:  57%|█████▋    | 275/481 [06:42<02:16,  1.51it/s]
Continual predict:  57%|█████▋    | 276/481 [06:43<02:39,  1.29it/s]
Continual predict:  58%|█████▊    | 277/481 [06:45<03:08,  1.08it/s]
Continual predict:  58%|█████▊    | 278/481 [06:46<03:01,  1.12it/s]
Continual predict:  58%|█████▊    | 279/481 [06:46<02:48,  1.20it/s]
Continual predict:  58%|█████▊    | 280/481 [06:47<02:24,  1.39it/s]
Continual predict:  58%|█████▊    | 281/481 [06:48<02:37,  1.27it/s]
Continual predict:  59%|█████▊    | 282/481 [06:48<02:37,  1.26it/s]
Continual predict:  59%|█████▉    | 283/481 [06:49<02:28,  1.33it/s]
Continual predict:  59%|█████▉    | 284/481 [06:50<02:10,  1.51it/s]
Continual predict:  59%|█████▉    | 285/481 [06:50<02:16,  1.43it/s]
Continual predict:  59%|█████▉    | 286/481 [06:51<02:19,  1.40it/s]
Continual predict:  60%|█████▉    | 287/481 [06:52<02:21,  1.37it/s]
Continual predict:  60%|█████▉    | 288/481 [06:53<02:14,  1.43it/s]
Continual predict:  60%|██████    | 289/481 [06:53<02:27,  1.30it/s]
Continual predict:  60%|██████    | 290/481 [06:54<02:25,  1.31it/s]
Continual predict:  60%|██████    | 291/481 [06:55<02:15,  1.40it/s]
Continual predict:  61%|██████    | 292/481 [06:55<01:51,  1.69it/s]
Continual predict:  61%|██████    | 293/481 [06:56<02:04,  1.51it/s]
Continual predict:  61%|██████    | 294/481 [06:56<01:56,  1.61it/s]
Continual predict:  61%|██████▏   | 295/481 [06:57<01:45,  1.76it/s]
Continual predict:  62%|██████▏   | 296/481 [06:58<02:06,  1.46it/s]
Continual predict:  62%|██████▏   | 297/481 [06:58<01:56,  1.58it/s]
Continual predict:  62%|██████▏   | 298/481 [06:59<01:45,  1.73it/s]
Continual predict:  62%|██████▏   | 299/481 [07:00<02:02,  1.48it/s]
Continual predict:  62%|██████▏   | 300/481 [07:00<01:44,  1.74it/s]
Continual predict:  63%|██████▎   | 301/481 [07:01<02:01,  1.48it/s]
Continual predict:  63%|██████▎   | 302/481 [07:01<01:46,  1.67it/s]
Continual predict:  63%|██████▎   | 303/481 [07:03<02:35,  1.15it/s]
Continual predict:  63%|██████▎   | 304/481 [07:03<02:04,  1.42it/s]
Continual predict:  63%|██████▎   | 305/481 [07:04<02:14,  1.31it/s]
Continual predict:  64%|██████▎   | 306/481 [07:05<02:15,  1.29it/s]
Continual predict:  64%|██████▍   | 307/481 [07:06<02:15,  1.28it/s]
Continual predict:  64%|██████▍   | 308/481 [07:06<01:56,  1.49it/s]
Continual predict:  64%|██████▍   | 309/481 [07:07<01:44,  1.65it/s]
Continual predict:  64%|██████▍   | 310/481 [07:07<01:43,  1.65it/s]
Continual predict:  65%|██████▍   | 311/481 [07:08<01:39,  1.71it/s]
Continual predict:  65%|██████▍   | 312/481 [07:08<01:44,  1.61it/s]
Continual predict:  65%|██████▌   | 313/481 [07:09<01:56,  1.44it/s]
Continual predict:  65%|██████▌   | 314/481 [07:10<01:43,  1.61it/s]
Continual predict:  65%|██████▌   | 315/481 [07:11<01:56,  1.42it/s]
Continual predict:  66%|██████▌   | 316/481 [07:11<02:00,  1.37it/s]
Continual predict:  66%|██████▌   | 317/481 [07:12<01:39,  1.65it/s]
Continual predict:  66%|██████▌   | 318/481 [07:12<01:31,  1.79it/s]
Continual predict:  66%|██████▋   | 319/481 [07:14<02:20,  1.15it/s]
Continual predict:  67%|██████▋   | 320/481 [07:14<02:11,  1.22it/s]
Continual predict:  67%|██████▋   | 321/481 [07:15<02:04,  1.28it/s]
Continual predict:  67%|██████▋   | 322/481 [07:16<01:46,  1.49it/s]
Continual predict:  67%|██████▋   | 323/481 [07:16<01:49,  1.45it/s]
Continual predict:  67%|██████▋   | 324/481 [07:17<01:57,  1.33it/s]
Continual predict:  68%|██████▊   | 325/481 [07:18<01:57,  1.33it/s]
Continual predict:  68%|██████▊   | 326/481 [07:19<01:50,  1.40it/s]
Continual predict:  68%|██████▊   | 327/481 [07:20<02:07,  1.20it/s]
Continual predict:  68%|██████▊   | 328/481 [07:21<02:17,  1.12it/s]
Continual predict:  68%|██████▊   | 329/481 [07:21<01:59,  1.27it/s]
Continual predict:  69%|██████▊   | 330/481 [07:22<01:37,  1.55it/s]
Continual predict:  69%|██████▉   | 331/481 [07:22<01:34,  1.59it/s]
Continual predict:  69%|██████▉   | 332/481 [07:23<01:29,  1.67it/s]
Continual predict:  69%|██████▉   | 333/481 [07:24<01:55,  1.28it/s]
Continual predict:  69%|██████▉   | 334/481 [07:25<01:59,  1.23it/s]
Continual predict:  70%|██████▉   | 335/481 [07:25<01:41,  1.43it/s]
Continual predict:  70%|██████▉   | 336/481 [07:26<01:30,  1.60it/s]
Continual predict:  70%|███████   | 337/481 [07:26<01:27,  1.65it/s]
Continual predict:  70%|███████   | 338/481 [07:27<01:28,  1.62it/s]
Continual predict:  70%|███████   | 339/481 [07:27<01:23,  1.70it/s]
Continual predict:  71%|███████   | 340/481 [07:28<01:43,  1.37it/s]
Continual predict:  71%|███████   | 341/481 [07:29<01:42,  1.36it/s]
Continual predict:  71%|███████   | 342/481 [07:30<01:30,  1.54it/s]
Continual predict:  71%|███████▏  | 343/481 [07:30<01:33,  1.47it/s]
Continual predict:  72%|███████▏  | 344/481 [07:31<01:43,  1.32it/s]
Continual predict:  72%|███████▏  | 345/481 [07:32<01:30,  1.50it/s]
Continual predict:  72%|███████▏  | 346/481 [07:33<01:35,  1.41it/s]
Continual predict:  72%|███████▏  | 347/481 [07:34<02:04,  1.08it/s]
Continual predict:  72%|███████▏  | 348/481 [07:34<01:39,  1.33it/s]
Continual predict:  73%|███████▎  | 349/481 [07:35<01:26,  1.53it/s]
Continual predict:  73%|███████▎  | 350/481 [07:35<01:24,  1.54it/s]
Continual predict:  73%|███████▎  | 351/481 [07:37<01:42,  1.27it/s]
Continual predict:  73%|███████▎  | 352/481 [07:37<01:40,  1.29it/s]
Continual predict:  73%|███████▎  | 353/481 [07:39<01:58,  1.08it/s]
Continual predict:  74%|███████▎  | 354/481 [07:40<02:09,  1.02s/it]
Continual predict:  74%|███████▍  | 355/481 [07:40<01:52,  1.12it/s]
Continual predict:  74%|███████▍  | 356/481 [07:41<01:52,  1.11it/s]
Continual predict:  74%|███████▍  | 357/481 [07:42<01:57,  1.06it/s]
Continual predict:  74%|███████▍  | 358/481 [07:43<01:49,  1.13it/s]
Continual predict:  75%|███████▍  | 359/481 [07:44<01:38,  1.24it/s]
Continual predict:  75%|███████▍  | 360/481 [07:45<01:53,  1.07it/s]
Continual predict:  75%|███████▌  | 361/481 [07:47<02:14,  1.12s/it]
Continual predict:  75%|███████▌  | 362/481 [07:47<01:54,  1.03it/s]
Continual predict:  75%|███████▌  | 363/481 [07:48<01:35,  1.23it/s]
Continual predict:  76%|███████▌  | 364/481 [07:48<01:17,  1.51it/s]
Continual predict:  76%|███████▌  | 365/481 [07:48<01:02,  1.84it/s]
Continual predict:  76%|███████▌  | 366/481 [07:48<00:52,  2.18it/s]
Continual predict:  76%|███████▋  | 367/481 [07:49<00:45,  2.51it/s]
Continual predict:  77%|███████▋  | 368/481 [07:49<00:40,  2.80it/s]
Continual predict:  77%|███████▋  | 369/481 [07:49<00:36,  3.04it/s]
Continual predict:  77%|███████▋  | 370/481 [07:50<00:34,  3.24it/s]
Continual predict:  77%|███████▋  | 371/481 [07:50<00:34,  3.22it/s]
Continual predict:  77%|███████▋  | 372/481 [07:50<00:35,  3.06it/s]
Continual predict:  78%|███████▊  | 373/481 [07:51<00:36,  2.95it/s]
Continual predict:  78%|███████▊  | 374/481 [07:51<00:33,  3.17it/s]
Continual predict:  78%|███████▊  | 375/481 [07:51<00:35,  3.02it/s]
Continual predict:  78%|███████▊  | 376/481 [07:52<00:34,  3.03it/s]
Continual predict:  78%|███████▊  | 377/481 [07:52<00:32,  3.23it/s]
Continual predict:  79%|███████▊  | 378/481 [07:52<00:30,  3.39it/s]
Continual predict:  79%|███████▉  | 379/481 [07:52<00:29,  3.51it/s]
Continual predict:  79%|███████▉  | 380/481 [07:53<00:28,  3.60it/s]
Continual predict:  79%|███████▉  | 381/481 [07:53<00:29,  3.42it/s]
Continual predict:  79%|███████▉  | 382/481 [07:53<00:28,  3.53it/s]
Continual predict:  80%|███████▉  | 383/481 [07:53<00:27,  3.61it/s]
Continual predict:  80%|███████▉  | 384/481 [07:54<00:26,  3.68it/s]
Continual predict:  80%|████████  | 385/481 [07:54<00:25,  3.72it/s]
Continual predict:  80%|████████  | 386/481 [07:54<00:25,  3.75it/s]
Continual predict:  80%|████████  | 387/481 [07:54<00:24,  3.77it/s]
Continual predict:  81%|████████  | 388/481 [07:55<00:24,  3.78it/s]
Continual predict:  81%|████████  | 389/481 [07:55<00:25,  3.58it/s]
Continual predict:  81%|████████  | 390/481 [07:55<00:24,  3.65it/s]
Continual predict:  81%|████████▏ | 391/481 [07:56<00:24,  3.70it/s]
Continual predict:  81%|████████▏ | 392/481 [07:56<00:25,  3.53it/s]
Continual predict:  82%|████████▏ | 393/481 [07:56<00:24,  3.61it/s]
Continual predict:  82%|████████▏ | 394/481 [07:56<00:23,  3.67it/s]
Continual predict:  82%|████████▏ | 395/481 [07:57<00:21,  3.95it/s]
Continual predict:  82%|████████▏ | 396/481 [07:57<00:24,  3.49it/s]
Continual predict:  83%|████████▎ | 397/481 [07:57<00:23,  3.58it/s]
Continual predict:  83%|████████▎ | 398/481 [07:57<00:22,  3.65it/s]
Continual predict:  83%|████████▎ | 399/481 [07:58<00:22,  3.70it/s]
Continual predict:  83%|████████▎ | 400/481 [07:58<00:21,  3.73it/s]
Continual predict:  83%|████████▎ | 401/481 [07:58<00:21,  3.76it/s]
Continual predict:  84%|████████▎ | 402/481 [07:59<00:20,  3.78it/s]
Continual predict:  84%|████████▍ | 403/481 [07:59<00:21,  3.58it/s]
Continual predict:  84%|████████▍ | 404/481 [07:59<00:21,  3.65it/s]
Continual predict:  84%|████████▍ | 405/481 [07:59<00:22,  3.32it/s]
Continual predict:  84%|████████▍ | 406/481 [08:00<00:21,  3.46it/s]
Continual predict:  85%|████████▍ | 407/481 [08:00<00:20,  3.56it/s]
Continual predict:  85%|████████▍ | 408/481 [08:00<00:20,  3.63it/s]
Continual predict:  85%|████████▌ | 409/481 [08:01<00:19,  3.69it/s]
Continual predict:  85%|████████▌ | 410/481 [08:01<00:21,  3.34it/s]
Continual predict:  85%|████████▌ | 411/481 [08:01<00:22,  3.13it/s]
Continual predict:  86%|████████▌ | 412/481 [08:02<00:20,  3.31it/s]
Continual predict:  86%|████████▌ | 413/481 [08:02<00:19,  3.45it/s]
Continual predict:  86%|████████▌ | 414/481 [08:02<00:18,  3.55it/s]
Continual predict:  86%|████████▋ | 415/481 [08:02<00:18,  3.63it/s]
Continual predict:  86%|████████▋ | 416/481 [08:03<00:17,  3.69it/s]
Continual predict:  87%|████████▋ | 417/481 [08:03<00:17,  3.73it/s]
Continual predict:  87%|████████▋ | 418/481 [08:03<00:17,  3.54it/s]
Continual predict:  87%|████████▋ | 419/481 [08:04<00:19,  3.21it/s]
Continual predict:  87%|████████▋ | 420/481 [08:04<00:20,  3.05it/s]
Continual predict:  88%|████████▊ | 421/481 [08:04<00:20,  2.94it/s]
Continual predict:  88%|████████▊ | 422/481 [08:05<00:19,  3.01it/s]
Continual predict:  88%|████████▊ | 423/481 [08:05<00:19,  3.05it/s]
Continual predict:  88%|████████▊ | 424/481 [08:05<00:19,  2.95it/s]
Continual predict:  88%|████████▊ | 425/481 [08:06<00:19,  2.88it/s]
Continual predict:  89%|████████▊ | 426/481 [08:06<00:19,  2.83it/s]
Continual predict:  89%|████████▉ | 427/481 [08:06<00:18,  2.93it/s]
Continual predict:  89%|████████▉ | 428/481 [08:07<00:16,  3.15it/s]
Continual predict:  89%|████████▉ | 429/481 [08:07<00:15,  3.33it/s]
Continual predict:  89%|████████▉ | 430/481 [08:07<00:14,  3.46it/s]
Continual predict:  90%|████████▉ | 431/481 [08:07<00:14,  3.37it/s]
Continual predict:  90%|████████▉ | 432/481 [08:08<00:14,  3.50it/s]
Continual predict:  90%|█████████ | 433/481 [08:08<00:13,  3.59it/s]
Continual predict:  90%|█████████ | 434/481 [08:08<00:13,  3.46it/s]
Continual predict:  90%|█████████ | 435/481 [08:09<00:12,  3.56it/s]
Continual predict:  91%|█████████ | 436/481 [08:09<00:13,  3.26it/s]
Continual predict:  91%|█████████ | 437/481 [08:09<00:15,  2.92it/s]
Continual predict:  91%|█████████ | 438/481 [08:10<00:14,  3.05it/s]
Continual predict:  91%|█████████▏| 439/481 [08:10<00:13,  3.08it/s]
Continual predict:  91%|█████████▏| 440/481 [08:10<00:13,  3.11it/s]
Continual predict:  92%|█████████▏| 441/481 [08:11<00:12,  3.13it/s]
Continual predict:  92%|█████████▏| 442/481 [08:11<00:11,  3.31it/s]
Continual predict:  92%|█████████▏| 443/481 [08:11<00:11,  3.45it/s]
Continual predict:  92%|█████████▏| 444/481 [08:11<00:10,  3.56it/s]
Continual predict:  93%|█████████▎| 445/481 [08:12<00:09,  3.63it/s]
Continual predict:  93%|█████████▎| 446/481 [08:12<00:10,  3.49it/s]
Continual predict:  93%|█████████▎| 447/481 [08:12<00:10,  3.39it/s]
Continual predict:  93%|█████████▎| 448/481 [08:13<00:09,  3.33it/s]
Continual predict:  93%|█████████▎| 449/481 [08:13<00:10,  3.12it/s]
Continual predict:  94%|█████████▎| 450/481 [08:13<00:09,  3.30it/s]
Continual predict:  94%|█████████▍| 451/481 [08:13<00:08,  3.44it/s]
Continual predict:  94%|█████████▍| 452/481 [08:14<00:08,  3.53it/s]
Continual predict:  94%|█████████▍| 453/481 [08:14<00:09,  3.04it/s]
Continual predict:  94%|█████████▍| 454/481 [08:14<00:08,  3.24it/s]
Continual predict:  95%|█████████▍| 455/481 [08:15<00:07,  3.39it/s]
Continual predict:  95%|█████████▍| 456/481 [08:15<00:07,  3.32it/s]
Continual predict:  95%|█████████▌| 457/481 [08:15<00:07,  3.28it/s]
Continual predict:  95%|█████████▌| 458/481 [08:16<00:06,  3.42it/s]
Continual predict:  95%|█████████▌| 459/481 [08:16<00:06,  3.53it/s]
Continual predict:  96%|█████████▌| 460/481 [08:16<00:05,  3.61it/s]
Continual predict:  96%|█████████▌| 461/481 [08:16<00:05,  3.67it/s]
Continual predict:  96%|█████████▌| 462/481 [08:17<00:05,  3.72it/s]
Continual predict:  96%|█████████▋| 463/481 [08:17<00:04,  3.75it/s]
Continual predict:  96%|█████████▋| 464/481 [08:17<00:04,  3.50it/s]
Continual predict:  97%|█████████▋| 465/481 [08:17<00:04,  3.89it/s]
Continual predict:  97%|█████████▋| 466/481 [08:18<00:03,  3.80it/s]
Continual predict:  97%|█████████▋| 467/481 [08:18<00:03,  3.94it/s]
Continual predict:  97%|█████████▋| 468/481 [08:18<00:03,  3.83it/s]
Continual predict:  98%|█████████▊| 469/481 [08:18<00:03,  3.75it/s]
Continual predict:  98%|█████████▊| 470/481 [08:19<00:02,  3.68it/s]
Continual predict:  98%|█████████▊| 471/481 [08:19<00:02,  3.66it/s]
Continual predict:  98%|█████████▊| 472/481 [08:20<00:03,  2.31it/s]
Continual predict:  98%|█████████▊| 473/481 [08:20<00:03,  2.25it/s]
Continual predict:  99%|█████████▊| 474/481 [08:20<00:02,  2.73it/s]
Continual predict:  99%|█████████▉| 475/481 [08:21<00:02,  2.92it/s]
Continual predict:  99%|█████████▉| 476/481 [08:21<00:01,  2.97it/s]
Continual predict:  99%|█████████▉| 477/481 [08:21<00:01,  3.13it/s]
Continual predict:  99%|█████████▉| 478/481 [08:22<00:00,  3.11it/s]
Continual predict: 100%|█████████▉| 479/481 [08:22<00:00,  2.35it/s]
Continual predict: 100%|█████████▉| 480/481 [08:23<00:00,  2.60it/s]
Continual predict: 100%|██████████| 481/481 [08:23<00:00,  2.71it/s]
Continual predict: 100%|██████████| 481/481 [08:23<00:00,  1.05s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=20]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=18]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (481/481)  label_acc=100.0% (481/481)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                               =       15.0
+  predict_eval_rougeL                                                 =    69.1592
+  predict_eval_rougeL_for_CL                                          =    69.1592
+  predict_eval_rougeL_for_task002_quoref_answer_generation            =     76.381
+  predict_eval_rougeL_for_task1290_xsum_summarization                 =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction      =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                     =    41.3728
+  predict_eval_rougeL_for_task181_outcome_extraction                  =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification        =       91.0
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation =    15.2912
+  predict_exact_match                                                 =    58.6279
+  predict_exact_match_for_CL                                          =    58.6279
+  predict_exact_match_for_task002_quoref_answer_generation            =       70.0
+  predict_exact_match_for_task1290_xsum_summarization                 =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction      =       98.0
+  predict_exact_match_for_task1572_samsum_summary                     =        0.0
+  predict_exact_match_for_task181_outcome_extraction                  =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification        =       91.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation =        0.0
+  predict_gen_len                                                     =      8.948
+  predict_global_step                                                 =         75
+  predict_rouge1                                                      =    71.0221
+  predict_rouge1_for_CL                                               =    71.0221
+  predict_rouge1_for_task002_quoref_answer_generation                 =     76.381
+  predict_rouge1_for_task1290_xsum_summarization                      =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction           =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                          =    50.3879
+  predict_rouge1_for_task181_outcome_extraction                       =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification             =       91.0
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation      =    15.2912
+  predict_runtime                                                     = 0:08:23.45
+  predict_samples                                                     =        481
+  predict_samples_per_second                                          =      0.955
+  predict_steps_per_second                                            =      0.955
+[2026-04-29 23:54:32,172] [INFO] [launch.py:367:main] Process 1066468 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/all_results.json b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..b0c21b3e4360366e24062a0578a25796d5409dd5
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/all_results.json
@@ -0,0 +1,45 @@
+{
+  "predict_exact_match": 48.537,
+  "predict_rouge1": 62.0913,
+  "predict_eval_rougeL": 60.3918,
+  "predict_gen_len": 9.1532,
+  "predict_runtime": 603.3824323599692,
+  "predict_samples": 581,
+  "predict_samples_per_second": 0.963,
+  "predict_steps_per_second": 0.963,
+  "epoch": 15.0,
+  "predict_global_step": 480,
+  "predict_exact_match_for_CL": 48.537,
+  "predict_rouge1_for_CL": 62.0913,
+  "predict_eval_rougeL_for_CL": 60.3918,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "train_runtime": 2457.5693,
+  "train_samples_per_second": 6.104,
+  "train_steps_per_second": 0.195,
+  "total_flos": 1.5907288413057843e+17,
+  "train_loss": 1.9921946803728738,
+  "train_samples": 1000
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e6166947bc53bb045a8cd450066d51f46d0f4408
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f248c687b3cdca143fe60ad5566bb96df11f6add43bf61300327e24fdcccdc5
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a2f058ef2f62e7e53d79a4d2cb1123b9c01543d5
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dd02e647de1401bcdbb320da01f6f24db25b86448201d8c7729f7505d7f5131
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..09a9f8db80fe73484fa05312a7cb06c97debddb8
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2dd95ba2f28a76c17572c92d324bc8c0f940999c79055addfbd69a0c3ddcf794
+size 1282674463
diff --git a/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/train.log b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..d3acdd9c1df3131c515e8325f388e4e87e0cc1fa
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/train.log
@@ -0,0 +1,582 @@
+[2026-04-29 23:54:38,226] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-29 23:54:38,226] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29507 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1729_personachat_generate_next --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next --cur_task_id 7 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[2026-04-29 23:54:41,440] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-29 23:54:41,440] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-29 23:54:41,440] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-29 23:54:41,440] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-29 23:54:41,440] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-29 23:54:41,440] [INFO] [launch.py:272:main] process 1106737 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task1729_personachat_generate_next', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next', '--cur_task_id', '7', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights']
+
+============================================================
+Task 7: task1729_personachat_generate_next
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 1000/100/581
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   3%|▎         | 10/291 [00:00<00:02, 96.24it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 144.44it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 158.25it/s]
Loading weights:  23%|██▎       | 66/291 [00:00<00:01, 159.17it/s]
Loading weights:  29%|██▉       | 85/291 [00:00<00:01, 163.82it/s]
Loading weights:  36%|███▌      | 104/291 [00:00<00:01, 168.96it/s]
Loading weights:  43%|████▎     | 126/291 [00:00<00:00, 179.25it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 183.89it/s]
Loading weights:  57%|█████▋    | 165/291 [00:00<00:00, 181.25it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 176.83it/s]
Loading weights:  69%|██████▉   | 202/291 [00:01<00:00, 174.01it/s]
Loading weights:  76%|███████▋  | 222/291 [00:01<00:00, 178.33it/s]
Loading weights:  82%|████████▏ | 240/291 [00:01<00:00, 178.31it/s]
Loading weights:  90%|█████████ | 262/291 [00:01<00:00, 188.71it/s]
Loading weights:  97%|█████████▋| 281/291 [00:01<00:00, 180.90it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 175.82it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1059)
+[FIX] Re-initialized lora_A/B in 512 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(7, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 7 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6756.045M (0.04%)
+[TRAIN] step_per_epoch=32, eval_steps=160
+  [SRT] Loaded 7 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 7 sigs, cur=task1729_personachat_generate_next, mapping={'task1729_personachat_generate_next': 0, 'task1572_samsum_summary': 7, 'task363_sst2_polarity_classification': 6, 'task1290_xsum_summarization': 5, 'task181_outcome_extraction': 4, 'task002_quoref_answer_generation': 3, 'task1510_evalution_relation_extraction': 2, 'task639_multi_woz_user_utterance_generation': 1}
+[SGWI] Task 7: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-29 23:55:42,392] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/480 [00:00<?, ?it/s]
  0%|          | 1/480 [00:05<41:22,  5.18s/it]
  0%|          | 2/480 [00:09<38:19,  4.81s/it]
  1%|          | 3/480 [00:14<37:48,  4.75s/it]
  1%|          | 4/480 [00:19<38:38,  4.87s/it]
  1%|          | 5/480 [00:24<38:21,  4.85s/it]
  1%|▏         | 6/480 [00:29<38:03,  4.82s/it]
  1%|▏         | 7/480 [00:34<39:02,  4.95s/it]
  2%|▏         | 8/480 [00:39<38:50,  4.94s/it]
  2%|▏         | 9/480 [00:44<38:47,  4.94s/it]
  2%|▏         | 10/480 [00:48<38:09,  4.87s/it]
                                                
{'loss': '2.868', 'grad_norm': '0.2307', 'learning_rate': '5e-05', 'epoch': '0.32'}
+
  2%|▏         | 10/480 [00:48<38:09,  4.87s/it]
  2%|▏         | 11/480 [00:54<38:58,  4.99s/it]
  2%|▎         | 12/480 [00:59<39:36,  5.08s/it]
  3%|▎         | 13/480 [01:03<37:57,  4.88s/it]
  3%|▎         | 14/480 [01:08<37:31,  4.83s/it]
  3%|▎         | 15/480 [01:13<37:33,  4.85s/it]
  3%|▎         | 16/480 [01:18<37:14,  4.81s/it]
  4%|▎         | 17/480 [01:22<36:53,  4.78s/it]
  4%|▍         | 18/480 [01:27<37:31,  4.87s/it]
  4%|▍         | 19/480 [01:32<37:33,  4.89s/it]
  4%|▍         | 20/480 [01:37<37:41,  4.92s/it]
                                                
{'loss': '2.514', 'grad_norm': '0.3071', 'learning_rate': '5e-05', 'epoch': '0.64'}
+
  4%|▍         | 20/480 [01:37<37:41,  4.92s/it]
  4%|▍         | 21/480 [01:42<37:08,  4.86s/it]
  5%|▍         | 22/480 [01:47<36:43,  4.81s/it]
  5%|▍         | 23/480 [01:52<37:01,  4.86s/it]
  5%|▌         | 24/480 [01:57<37:32,  4.94s/it]
  5%|▌         | 25/480 [02:02<38:12,  5.04s/it]
  5%|▌         | 26/480 [02:07<37:40,  4.98s/it]
  6%|▌         | 27/480 [02:12<37:19,  4.94s/it]
  6%|▌         | 28/480 [02:17<36:52,  4.89s/it]
  6%|▌         | 29/480 [02:22<36:56,  4.91s/it]
  6%|▋         | 30/480 [02:26<36:46,  4.90s/it]
                                                
{'loss': '2.294', 'grad_norm': '0.265', 'learning_rate': '5e-05', 'epoch': '0.96'}
+
  6%|▋         | 30/480 [02:26<36:46,  4.90s/it]
  6%|▋         | 31/480 [02:32<37:18,  4.99s/it]
  7%|▋         | 32/480 [02:33<29:03,  3.89s/it]
  7%|▋         | 33/480 [02:38<31:26,  4.22s/it]
  7%|▋         | 34/480 [02:43<33:14,  4.47s/it]
  7%|▋         | 35/480 [02:48<33:41,  4.54s/it]
  8%|▊         | 36/480 [02:53<34:47,  4.70s/it]
  8%|▊         | 37/480 [02:58<36:22,  4.93s/it]
  8%|▊         | 38/480 [03:03<36:58,  5.02s/it]
  8%|▊         | 39/480 [03:09<37:04,  5.04s/it]
  8%|▊         | 40/480 [03:14<37:10,  5.07s/it]
                                                
{'loss': '2.15', 'grad_norm': '0.3886', 'learning_rate': '5e-05', 'epoch': '1.256'}
+
  8%|▊         | 40/480 [03:14<37:10,  5.07s/it]
  9%|▊         | 41/480 [03:18<36:10,  4.94s/it]
  9%|▉         | 42/480 [03:23<35:25,  4.85s/it]
  9%|▉         | 43/480 [03:28<35:38,  4.89s/it]
  9%|▉         | 44/480 [03:33<35:24,  4.87s/it]
  9%|▉         | 45/480 [03:37<34:33,  4.77s/it]
 10%|▉         | 46/480 [03:42<34:50,  4.82s/it]
 10%|▉         | 47/480 [03:47<34:49,  4.83s/it]
 10%|█         | 48/480 [03:52<34:52,  4.84s/it]
 10%|█         | 49/480 [03:56<33:45,  4.70s/it]
 10%|█         | 50/480 [04:01<34:19,  4.79s/it]
                                                
{'loss': '2.127', 'grad_norm': '0.5696', 'learning_rate': '5e-05', 'epoch': '1.576'}
+
 10%|█         | 50/480 [04:01<34:19,  4.79s/it]
 11%|█         | 51/480 [04:06<34:18,  4.80s/it]
 11%|█         | 52/480 [04:11<34:10,  4.79s/it]
 11%|█         | 53/480 [04:16<34:13,  4.81s/it]
 11%|█▏        | 54/480 [04:21<34:14,  4.82s/it]
 11%|█▏        | 55/480 [04:26<34:13,  4.83s/it]
 12%|█▏        | 56/480 [04:30<34:27,  4.88s/it]
 12%|█▏        | 57/480 [04:35<33:59,  4.82s/it]
 12%|█▏        | 58/480 [04:40<34:05,  4.85s/it]
 12%|█▏        | 59/480 [04:45<34:15,  4.88s/it]
 12%|█▎        | 60/480 [04:50<34:26,  4.92s/it]
                                                
{'loss': '2.075', 'grad_norm': '0.5494', 'learning_rate': '5e-05', 'epoch': '1.896'}
+
 12%|█▎        | 60/480 [04:50<34:26,  4.92s/it]
 13%|█▎        | 61/480 [04:55<34:11,  4.90s/it]
 13%|█▎        | 62/480 [05:00<33:45,  4.85s/it]
 13%|█▎        | 63/480 [05:04<33:44,  4.85s/it]
 13%|█▎        | 64/480 [05:06<25:57,  3.74s/it]
 14%|█▎        | 65/480 [05:11<28:47,  4.16s/it]
 14%|█▍        | 66/480 [05:16<30:15,  4.39s/it]
 14%|█▍        | 67/480 [05:21<31:40,  4.60s/it]
 14%|█▍        | 68/480 [05:26<32:51,  4.79s/it]
 14%|█▍        | 69/480 [05:31<33:10,  4.84s/it]
 15%|█▍        | 70/480 [05:36<32:36,  4.77s/it]
                                                
{'loss': '2.052', 'grad_norm': '0.6321', 'learning_rate': '5e-05', 'epoch': '2.192'}
+
 15%|█▍        | 70/480 [05:36<32:36,  4.77s/it]
 15%|█▍        | 71/480 [05:40<31:51,  4.67s/it]
 15%|█▌        | 72/480 [05:45<31:48,  4.68s/it]
 15%|█▌        | 73/480 [05:50<32:03,  4.73s/it]
 15%|█▌        | 74/480 [05:55<33:00,  4.88s/it]
 16%|█▌        | 75/480 [06:00<33:20,  4.94s/it]
 16%|█▌        | 76/480 [06:05<32:47,  4.87s/it]
 16%|█▌        | 77/480 [06:09<32:26,  4.83s/it]
 16%|█▋        | 78/480 [06:14<32:26,  4.84s/it]
 16%|█▋        | 79/480 [06:19<31:59,  4.79s/it]
 17%|█▋        | 80/480 [06:24<32:34,  4.89s/it]
                                                
{'loss': '2.079', 'grad_norm': '0.493', 'learning_rate': '5e-05', 'epoch': '2.512'}
+
 17%|█▋        | 80/480 [06:24<32:34,  4.89s/it]
 17%|█▋        | 81/480 [06:29<32:18,  4.86s/it]
 17%|█▋        | 82/480 [06:34<32:10,  4.85s/it]
 17%|█▋        | 83/480 [06:39<32:36,  4.93s/it]
 18%|█▊        | 84/480 [06:44<32:44,  4.96s/it]
 18%|█▊        | 85/480 [06:49<33:19,  5.06s/it]
 18%|█▊        | 86/480 [06:54<32:45,  4.99s/it]
 18%|█▊        | 87/480 [06:59<32:48,  5.01s/it]
 18%|█▊        | 88/480 [07:04<32:29,  4.97s/it]
 19%|█▊        | 89/480 [07:08<31:29,  4.83s/it]
 19%|█▉        | 90/480 [07:13<31:55,  4.91s/it]
                                                
{'loss': '2.002', 'grad_norm': '0.3994', 'learning_rate': '5e-05', 'epoch': '2.832'}
+
 19%|█▉        | 90/480 [07:13<31:55,  4.91s/it]
 19%|█▉        | 91/480 [07:18<31:46,  4.90s/it]
 19%|█▉        | 92/480 [07:23<32:09,  4.97s/it]
 19%|█▉        | 93/480 [07:29<32:20,  5.01s/it]
 20%|█▉        | 94/480 [07:33<31:39,  4.92s/it]
 20%|█▉        | 95/480 [07:38<31:30,  4.91s/it]
 20%|██        | 96/480 [07:39<24:11,  3.78s/it]
 20%|██        | 97/480 [07:44<26:28,  4.15s/it]
 20%|██        | 98/480 [07:49<27:49,  4.37s/it]
 21%|██        | 99/480 [07:54<28:22,  4.47s/it]
 21%|██        | 100/480 [07:59<29:08,  4.60s/it]
                                                 
{'loss': '2.033', 'grad_norm': '0.5144', 'learning_rate': '5e-05', 'epoch': '3.128'}
+
 21%|██        | 100/480 [07:59<29:08,  4.60s/it]
 21%|██        | 101/480 [08:03<29:00,  4.59s/it]
 21%|██▏       | 102/480 [08:08<29:17,  4.65s/it]
 21%|██▏       | 103/480 [08:13<29:53,  4.76s/it]
 22%|██▏       | 104/480 [08:18<29:53,  4.77s/it]
 22%|██▏       | 105/480 [08:22<29:22,  4.70s/it]
 22%|██▏       | 106/480 [08:27<29:38,  4.75s/it]
 22%|██▏       | 107/480 [08:33<30:34,  4.92s/it]
 22%|██▎       | 108/480 [08:37<29:51,  4.82s/it]
 23%|██▎       | 109/480 [08:42<29:43,  4.81s/it]
 23%|██▎       | 110/480 [08:47<30:13,  4.90s/it]
                                                 
{'loss': '2.001', 'grad_norm': '0.3317', 'learning_rate': '5e-05', 'epoch': '3.448'}
+
 23%|██▎       | 110/480 [08:47<30:13,  4.90s/it]
 23%|██▎       | 111/480 [08:52<29:51,  4.85s/it]
 23%|██▎       | 112/480 [08:57<29:48,  4.86s/it]
 24%|██▎       | 113/480 [09:02<30:20,  4.96s/it]
 24%|██▍       | 114/480 [09:07<30:09,  4.94s/it]
 24%|██▍       | 115/480 [09:12<29:37,  4.87s/it]
 24%|██▍       | 116/480 [09:17<29:44,  4.90s/it]
 24%|██▍       | 117/480 [09:22<30:02,  4.96s/it]
 25%|██▍       | 118/480 [09:27<29:51,  4.95s/it]
 25%|██▍       | 119/480 [09:31<29:17,  4.87s/it]
 25%|██▌       | 120/480 [09:36<29:29,  4.92s/it]
                                                 
{'loss': '2.054', 'grad_norm': '0.4306', 'learning_rate': '5e-05', 'epoch': '3.768'}
+
 25%|██▌       | 120/480 [09:36<29:29,  4.92s/it]
 25%|██▌       | 121/480 [09:41<29:27,  4.92s/it]
 25%|██▌       | 122/480 [09:46<28:58,  4.85s/it]
 26%|██▌       | 123/480 [09:51<28:40,  4.82s/it]
 26%|██▌       | 124/480 [09:55<28:32,  4.81s/it]
 26%|██▌       | 125/480 [10:01<29:09,  4.93s/it]
 26%|██▋       | 126/480 [10:06<29:20,  4.97s/it]
 26%|██▋       | 127/480 [10:11<29:10,  4.96s/it]
 27%|██▋       | 128/480 [10:12<22:27,  3.83s/it]
 27%|██▋       | 129/480 [10:17<24:06,  4.12s/it]
 27%|██▋       | 130/480 [10:21<25:16,  4.33s/it]
                                                 
{'loss': '2.004', 'grad_norm': '0.3808', 'learning_rate': '5e-05', 'epoch': '4.064'}
+
 27%|██▋       | 130/480 [10:21<25:16,  4.33s/it]
 27%|██▋       | 131/480 [10:26<26:22,  4.53s/it]
 28%|██▊       | 132/480 [10:31<26:45,  4.61s/it]
 28%|██▊       | 133/480 [10:36<27:05,  4.68s/it]
 28%|██▊       | 134/480 [10:41<27:06,  4.70s/it]
 28%|██▊       | 135/480 [10:46<27:38,  4.81s/it]
 28%|██▊       | 136/480 [10:51<27:32,  4.80s/it]
 29%|██▊       | 137/480 [10:55<27:09,  4.75s/it]
 29%|██▉       | 138/480 [11:00<27:26,  4.81s/it]
 29%|██▉       | 139/480 [11:05<27:38,  4.86s/it]
 29%|██▉       | 140/480 [11:10<27:49,  4.91s/it]
                                                 
{'loss': '2.013', 'grad_norm': '0.5087', 'learning_rate': '5e-05', 'epoch': '4.384'}
+
 29%|██▉       | 140/480 [11:10<27:49,  4.91s/it]
 29%|██▉       | 141/480 [11:15<27:36,  4.89s/it]
 30%|██▉       | 142/480 [11:20<27:39,  4.91s/it]
 30%|██▉       | 143/480 [11:25<26:59,  4.80s/it]
 30%|███       | 144/480 [11:30<27:20,  4.88s/it]
 30%|███       | 145/480 [11:34<27:07,  4.86s/it]
 30%|███       | 146/480 [11:39<26:45,  4.81s/it]
 31%|███       | 147/480 [11:44<26:53,  4.84s/it]
 31%|███       | 148/480 [11:49<27:08,  4.90s/it]
 31%|███       | 149/480 [11:55<27:52,  5.05s/it]
 31%|███▏      | 150/480 [11:59<27:14,  4.95s/it]
                                                 
{'loss': '1.992', 'grad_norm': '0.3734', 'learning_rate': '5e-05', 'epoch': '4.704'}
+
 31%|███▏      | 150/480 [11:59<27:14,  4.95s/it]
 31%|███▏      | 151/480 [12:04<26:39,  4.86s/it]
 32%|███▏      | 152/480 [12:09<26:30,  4.85s/it]
 32%|███▏      | 153/480 [12:14<26:26,  4.85s/it]
 32%|███▏      | 154/480 [12:18<25:47,  4.75s/it]
 32%|███▏      | 155/480 [12:23<26:03,  4.81s/it]
 32%|███▎      | 156/480 [12:28<25:55,  4.80s/it]
 33%|███▎      | 157/480 [12:33<25:43,  4.78s/it]
 33%|███▎      | 158/480 [12:37<25:51,  4.82s/it]
 33%|███▎      | 159/480 [12:42<25:39,  4.79s/it]
 33%|███▎      | 160/480 [12:44<20:00,  3.75s/it]
                                                 
{'loss': '1.985', 'grad_norm': '2.004', 'learning_rate': '5e-05', 'epoch': '5'}
+
 33%|███▎      | 160/480 [12:44<20:00,  3.75s/it]
+
Eval @ step 160:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 160:   1%|          | 1/100 [00:00<00:33,  2.92it/s]
+
Eval @ step 160:   2%|▏         | 2/100 [00:00<00:28,  3.49it/s]
+
Eval @ step 160:   3%|▎         | 3/100 [00:01<00:55,  1.74it/s]
+
Eval @ step 160:   4%|▍         | 4/100 [00:01<00:49,  1.95it/s]
+
Eval @ step 160:   5%|▌         | 5/100 [00:02<00:56,  1.69it/s]
+
Eval @ step 160:   6%|▌         | 6/100 [00:03<00:56,  1.68it/s]
+
Eval @ step 160:   7%|▋         | 7/100 [00:03<00:55,  1.68it/s]
+
Eval @ step 160:   8%|▊         | 8/100 [00:04<00:50,  1.83it/s]
+
Eval @ step 160:   9%|▉         | 9/100 [00:04<00:42,  2.15it/s]
+
Eval @ step 160:  10%|█         | 10/100 [00:05<00:41,  2.17it/s]
+
Eval @ step 160:  11%|█         | 11/100 [00:05<00:42,  2.09it/s]
+
Eval @ step 160:  12%|█▏        | 12/100 [00:06<00:42,  2.09it/s]
+
Eval @ step 160:  13%|█▎        | 13/100 [00:06<00:42,  2.07it/s]
+
Eval @ step 160:  14%|█▍        | 14/100 [00:07<00:47,  1.81it/s]
+
Eval @ step 160:  15%|█▌        | 15/100 [00:07<00:45,  1.87it/s]
+
Eval @ step 160:  16%|█▌        | 16/100 [00:08<00:42,  1.97it/s]
+
Eval @ step 160:  17%|█▋        | 17/100 [00:08<00:41,  2.00it/s]
+
Eval @ step 160:  18%|█▊        | 18/100 [00:09<00:45,  1.80it/s]
+
Eval @ step 160:  19%|█▉        | 19/100 [00:09<00:39,  2.03it/s]
+
Eval @ step 160:  20%|██        | 20/100 [00:10<00:38,  2.09it/s]
+
Eval @ step 160:  21%|██        | 21/100 [00:10<00:36,  2.14it/s]
+
Eval @ step 160:  22%|██▏       | 22/100 [00:11<00:36,  2.17it/s]
+
Eval @ step 160:  23%|██▎       | 23/100 [00:11<00:31,  2.45it/s]
+
Eval @ step 160:  24%|██▍       | 24/100 [00:12<00:41,  1.82it/s]
+
Eval @ step 160:  25%|██▌       | 25/100 [00:12<00:45,  1.64it/s]
+
Eval @ step 160:  26%|██▌       | 26/100 [00:13<00:42,  1.75it/s]
+
Eval @ step 160:  27%|██▋       | 27/100 [00:13<00:39,  1.83it/s]
+
Eval @ step 160:  28%|██▊       | 28/100 [00:14<00:37,  1.94it/s]
+
Eval @ step 160:  29%|██▉       | 29/100 [00:15<00:46,  1.52it/s]
+
Eval @ step 160:  30%|███       | 30/100 [00:16<00:50,  1.39it/s]
+
Eval @ step 160:  31%|███       | 31/100 [00:16<00:42,  1.63it/s]
+
Eval @ step 160:  32%|███▏      | 32/100 [00:17<00:47,  1.43it/s]
+
Eval @ step 160:  33%|███▎      | 33/100 [00:17<00:39,  1.68it/s]
+
Eval @ step 160:  34%|███▍      | 34/100 [00:18<00:38,  1.73it/s]
+
Eval @ step 160:  35%|███▌      | 35/100 [00:18<00:34,  1.91it/s]
+
Eval @ step 160:  36%|███▌      | 36/100 [00:19<00:35,  1.78it/s]
+
Eval @ step 160:  37%|███▋      | 37/100 [00:19<00:30,  2.09it/s]
+
Eval @ step 160:  38%|███▊      | 38/100 [00:20<00:41,  1.49it/s]
+
Eval @ step 160:  39%|███▉      | 39/100 [00:21<00:38,  1.58it/s]
+
Eval @ step 160:  40%|████      | 40/100 [00:21<00:34,  1.73it/s]
+
Eval @ step 160:  41%|████      | 41/100 [00:22<00:30,  1.96it/s]
+
Eval @ step 160:  42%|████▏     | 42/100 [00:22<00:29,  1.94it/s]
+
Eval @ step 160:  43%|████▎     | 43/100 [00:23<00:37,  1.51it/s]
+
Eval @ step 160:  44%|████▍     | 44/100 [00:24<00:33,  1.68it/s]
+
Eval @ step 160:  45%|████▌     | 45/100 [00:24<00:30,  1.78it/s]
+
Eval @ step 160:  46%|████▌     | 46/100 [00:25<00:35,  1.51it/s]
+
Eval @ step 160:  47%|████▋     | 47/100 [00:26<00:37,  1.42it/s]
+
Eval @ step 160:  48%|████▊     | 48/100 [00:28<00:58,  1.13s/it]
+
Eval @ step 160:  49%|████▉     | 49/100 [00:29<00:49,  1.03it/s]
+
Eval @ step 160:  50%|█████     | 50/100 [00:29<00:44,  1.13it/s]
+
Eval @ step 160:  51%|█████     | 51/100 [00:30<00:45,  1.09it/s]
+
Eval @ step 160:  52%|█████▏    | 52/100 [00:31<00:40,  1.17it/s]
+
Eval @ step 160:  53%|█████▎    | 53/100 [00:32<00:40,  1.17it/s]
+
Eval @ step 160:  54%|█████▍    | 54/100 [00:32<00:35,  1.30it/s]
+
Eval @ step 160:  55%|█████▌    | 55/100 [00:33<00:28,  1.57it/s]
+
Eval @ step 160:  56%|█████▌    | 56/100 [00:33<00:22,  1.93it/s]
+
Eval @ step 160:  57%|█████▋    | 57/100 [00:34<00:23,  1.84it/s]
+
Eval @ step 160:  58%|█████▊    | 58/100 [00:34<00:24,  1.74it/s]
+
Eval @ step 160:  59%|█████▉    | 59/100 [00:35<00:25,  1.59it/s]
+
Eval @ step 160:  60%|██████    | 60/100 [00:36<00:25,  1.58it/s]
+
Eval @ step 160:  61%|██████    | 61/100 [00:36<00:24,  1.61it/s]
+
Eval @ step 160:  62%|██████▏   | 62/100 [00:37<00:22,  1.67it/s]
+
Eval @ step 160:  63%|██████▎   | 63/100 [00:37<00:23,  1.57it/s]
+
Eval @ step 160:  64%|██████▍   | 64/100 [00:39<00:29,  1.21it/s]
+
Eval @ step 160:  65%|██████▌   | 65/100 [00:39<00:25,  1.36it/s]
+
Eval @ step 160:  66%|██████▌   | 66/100 [00:40<00:21,  1.57it/s]
+
Eval @ step 160:  67%|██████▋   | 67/100 [00:40<00:19,  1.68it/s]
+
Eval @ step 160:  68%|██████▊   | 68/100 [00:41<00:19,  1.67it/s]
+
Eval @ step 160:  69%|██████▉   | 69/100 [00:42<00:23,  1.32it/s]
+
Eval @ step 160:  70%|███████   | 70/100 [00:42<00:19,  1.57it/s]
+
Eval @ step 160:  71%|███████   | 71/100 [00:43<00:17,  1.64it/s]
+
Eval @ step 160:  72%|███████▏  | 72/100 [00:43<00:17,  1.61it/s]
+
Eval @ step 160:  73%|███████▎  | 73/100 [00:44<00:18,  1.48it/s]
+
Eval @ step 160:  74%|███████▍  | 74/100 [00:45<00:17,  1.46it/s]
+
Eval @ step 160:  75%|███████▌  | 75/100 [00:45<00:13,  1.81it/s]
+
Eval @ step 160:  76%|███████▌  | 76/100 [00:46<00:15,  1.56it/s]
+
Eval @ step 160:  77%|███████▋  | 77/100 [00:47<00:14,  1.63it/s]
+
Eval @ step 160:  78%|███████▊  | 78/100 [00:47<00:14,  1.47it/s]
+
Eval @ step 160:  79%|███████▉  | 79/100 [00:48<00:13,  1.57it/s]
+
Eval @ step 160:  80%|████████  | 80/100 [00:49<00:13,  1.44it/s]
+
Eval @ step 160:  81%|████████  | 81/100 [00:50<00:14,  1.29it/s]
+
Eval @ step 160:  82%|████████▏ | 82/100 [00:50<00:12,  1.41it/s]
+
Eval @ step 160:  83%|████████▎ | 83/100 [00:51<00:12,  1.39it/s]
+
Eval @ step 160:  84%|████████▍ | 84/100 [00:51<00:10,  1.57it/s]
+
Eval @ step 160:  85%|████████▌ | 85/100 [00:52<00:08,  1.72it/s]
+
Eval @ step 160:  86%|████████▌ | 86/100 [00:53<00:08,  1.62it/s]
+
Eval @ step 160:  87%|████████▋ | 87/100 [00:53<00:08,  1.59it/s]
+
Eval @ step 160:  88%|████████▊ | 88/100 [00:54<00:08,  1.43it/s]
+
Eval @ step 160:  89%|████████▉ | 89/100 [00:54<00:06,  1.78it/s]
+
Eval @ step 160:  90%|█████████ | 90/100 [00:55<00:06,  1.66it/s]
+
Eval @ step 160:  91%|█████████ | 91/100 [00:56<00:05,  1.75it/s]
+
Eval @ step 160:  92%|█████████▏| 92/100 [00:56<00:04,  1.91it/s]
+
Eval @ step 160:  93%|█████████▎| 93/100 [00:57<00:03,  1.91it/s]
+
Eval @ step 160:  94%|█████████▍| 94/100 [00:57<00:03,  1.84it/s]
+
Eval @ step 160:  95%|█████████▌| 95/100 [00:58<00:02,  1.82it/s]
+
Eval @ step 160:  96%|█████████▌| 96/100 [00:58<00:02,  1.87it/s]
+
Eval @ step 160:  97%|█████████▋| 97/100 [00:58<00:01,  2.24it/s]
+
Eval @ step 160:  98%|█████████▊| 98/100 [00:59<00:00,  2.03it/s]
+
Eval @ step 160:  99%|█████████▉| 99/100 [01:00<00:00,  1.71it/s]
+
Eval @ step 160: 100%|██████████| 100/100 [01:00<00:00,  1.95it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=18.0212 (step=160) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '18.02', 'epoch': '5'}
+
 33%|███▎      | 160/480 [13:44<20:00,  3.75s/it]
 34%|███▎      | 161/480 [13:49<1:58:29, 22.29s/it]
 34%|███▍      | 162/480 [13:53<1:29:42, 16.92s/it]
 34%|███▍      | 163/480 [13:58<1:10:28, 13.34s/it]
 34%|███▍      | 164/480 [14:04<57:11, 10.86s/it]  
 34%|███▍      | 165/480 [14:09<47:53,  9.12s/it]
 35%|███▍      | 166/480 [14:14<41:54,  8.01s/it]
 35%|███▍      | 167/480 [14:19<37:08,  7.12s/it]
 35%|███▌      | 168/480 [14:24<33:19,  6.41s/it]
 35%|███▌      | 169/480 [14:28<30:30,  5.89s/it]
 35%|███▌      | 170/480 [14:33<28:49,  5.58s/it]
                                                 
{'loss': '1.969', 'grad_norm': '0.6532', 'learning_rate': '5e-05', 'epoch': '5.32'}
+
 35%|███▌      | 170/480 [14:33<28:49,  5.58s/it]
 36%|███▌      | 171/480 [14:38<27:57,  5.43s/it]
 36%|███▌      | 172/480 [14:43<26:46,  5.22s/it]
 36%|███▌      | 173/480 [14:48<25:32,  4.99s/it]
 36%|███▋      | 174/480 [14:52<25:01,  4.91s/it]
 36%|███▋      | 175/480 [14:57<25:07,  4.94s/it]
 37%|███▋      | 176/480 [15:02<24:46,  4.89s/it]
 37%|███▋      | 177/480 [15:07<24:27,  4.84s/it]
 37%|███▋      | 178/480 [15:12<24:17,  4.83s/it]
 37%|███▋      | 179/480 [15:16<23:59,  4.78s/it]
 38%|███▊      | 180/480 [15:22<24:33,  4.91s/it]
                                                 
{'loss': '1.956', 'grad_norm': '0.4201', 'learning_rate': '5e-05', 'epoch': '5.64'}
+
 38%|███▊      | 180/480 [15:22<24:33,  4.91s/it]
 38%|███▊      | 181/480 [15:26<24:10,  4.85s/it]
 38%|███▊      | 182/480 [15:31<24:03,  4.84s/it]
 38%|███▊      | 183/480 [15:36<24:37,  4.98s/it]
 38%|███▊      | 184/480 [15:41<24:00,  4.87s/it]
 39%|███▊      | 185/480 [15:46<24:38,  5.01s/it]
 39%|███▉      | 186/480 [15:51<24:42,  5.04s/it]
 39%|███▉      | 187/480 [15:57<25:04,  5.14s/it]
 39%|███▉      | 188/480 [16:02<24:52,  5.11s/it]
 39%|███▉      | 189/480 [16:06<24:02,  4.96s/it]
 40%|███▉      | 190/480 [16:11<23:40,  4.90s/it]
                                                 
{'loss': '1.993', 'grad_norm': '0.4078', 'learning_rate': '5e-05', 'epoch': '5.96'}
+
 40%|███▉      | 190/480 [16:11<23:40,  4.90s/it]
 40%|███▉      | 191/480 [16:16<23:28,  4.87s/it]
 40%|████      | 192/480 [16:17<18:10,  3.79s/it]
 40%|████      | 193/480 [16:22<19:25,  4.06s/it]
 40%|████      | 194/480 [16:27<21:25,  4.49s/it]
 41%|████      | 195/480 [16:32<21:51,  4.60s/it]
 41%|████      | 196/480 [16:37<22:07,  4.68s/it]
 41%|████      | 197/480 [16:42<22:23,  4.75s/it]
 41%|████▏     | 198/480 [16:47<22:14,  4.73s/it]
 41%|████▏     | 199/480 [16:51<22:06,  4.72s/it]
 42%|████▏     | 200/480 [16:56<22:24,  4.80s/it]
                                                 
{'loss': '1.972', 'grad_norm': '0.6589', 'learning_rate': '5e-05', 'epoch': '6.256'}
+
 42%|████▏     | 200/480 [16:56<22:24,  4.80s/it]
 42%|████▏     | 201/480 [17:02<22:44,  4.89s/it]
 42%|████▏     | 202/480 [17:06<22:30,  4.86s/it]
 42%|████▏     | 203/480 [17:11<22:23,  4.85s/it]
 42%|████▎     | 204/480 [17:16<22:10,  4.82s/it]
 43%|████▎     | 205/480 [17:21<22:01,  4.81s/it]
 43%|████▎     | 206/480 [17:26<22:17,  4.88s/it]
 43%|████▎     | 207/480 [17:31<22:43,  5.00s/it]
 43%|████▎     | 208/480 [17:36<22:22,  4.94s/it]
 44%|████▎     | 209/480 [17:41<22:09,  4.90s/it]
 44%|████▍     | 210/480 [17:45<21:53,  4.86s/it]
                                                 
{'loss': '1.955', 'grad_norm': '0.8584', 'learning_rate': '5e-05', 'epoch': '6.576'}
+
 44%|████▍     | 210/480 [17:45<21:53,  4.86s/it]
 44%|████▍     | 211/480 [17:50<21:30,  4.80s/it]
 44%|████▍     | 212/480 [17:55<21:24,  4.79s/it]
 44%|████▍     | 213/480 [18:00<22:03,  4.96s/it]
 45%|████▍     | 214/480 [18:05<21:42,  4.90s/it]
 45%|████▍     | 215/480 [18:10<21:36,  4.89s/it]
 45%|████▌     | 216/480 [18:15<21:54,  4.98s/it]
 45%|████▌     | 217/480 [18:20<22:10,  5.06s/it]
 45%|████▌     | 218/480 [18:25<21:50,  5.00s/it]
 46%|████▌     | 219/480 [18:30<21:16,  4.89s/it]
 46%|████▌     | 220/480 [18:35<21:49,  5.04s/it]
                                                 
{'loss': '1.989', 'grad_norm': '0.3912', 'learning_rate': '5e-05', 'epoch': '6.896'}
+
 46%|████▌     | 220/480 [18:35<21:49,  5.04s/it]
 46%|████▌     | 221/480 [18:40<21:17,  4.93s/it]
 46%|████▋     | 222/480 [18:45<21:31,  5.01s/it]
 46%|████▋     | 223/480 [18:50<20:55,  4.89s/it]
 47%|████▋     | 224/480 [18:51<16:05,  3.77s/it]
 47%|████▋     | 225/480 [18:55<16:56,  3.99s/it]
 47%|████▋     | 226/480 [19:00<17:45,  4.20s/it]
 47%|████▋     | 227/480 [19:05<19:00,  4.51s/it]
 48%|████▊     | 228/480 [19:10<18:57,  4.51s/it]
 48%|████▊     | 229/480 [19:15<19:31,  4.67s/it]
 48%|████▊     | 230/480 [19:20<19:49,  4.76s/it]
                                                 
{'loss': '1.958', 'grad_norm': '0.379', 'learning_rate': '5e-05', 'epoch': '7.192'}
+
 48%|████▊     | 230/480 [19:20<19:49,  4.76s/it]
 48%|████▊     | 231/480 [19:25<20:11,  4.87s/it]
 48%|████▊     | 232/480 [19:29<19:46,  4.78s/it]
 49%|████▊     | 233/480 [19:34<19:56,  4.85s/it]
 49%|████▉     | 234/480 [19:39<19:43,  4.81s/it]
 49%|████▉     | 235/480 [19:44<19:41,  4.82s/it]
 49%|████▉     | 236/480 [19:49<20:02,  4.93s/it]
 49%|████▉     | 237/480 [19:54<19:51,  4.90s/it]
 50%|████▉     | 238/480 [19:59<19:45,  4.90s/it]
 50%|████▉     | 239/480 [20:04<19:28,  4.85s/it]
 50%|█████     | 240/480 [20:09<19:29,  4.87s/it]
                                                 
{'loss': '1.974', 'grad_norm': '0.6317', 'learning_rate': '5e-05', 'epoch': '7.512'}
+
 50%|█████     | 240/480 [20:09<19:29,  4.87s/it]
 50%|█████     | 241/480 [20:14<19:34,  4.91s/it]
 50%|█████     | 242/480 [20:18<19:27,  4.90s/it]
 51%|█████     | 243/480 [20:23<19:06,  4.84s/it]
 51%|█████     | 244/480 [20:28<19:14,  4.89s/it]
 51%|█████     | 245/480 [20:33<19:05,  4.88s/it]
 51%|█████▏    | 246/480 [20:38<19:03,  4.89s/it]
 51%|█████▏    | 247/480 [20:43<19:14,  4.96s/it]
 52%|█████▏    | 248/480 [20:48<19:03,  4.93s/it]
 52%|█████▏    | 249/480 [20:53<18:54,  4.91s/it]
 52%|█████▏    | 250/480 [20:58<18:57,  4.95s/it]
                                                 
{'loss': '1.967', 'grad_norm': '0.4234', 'learning_rate': '5e-05', 'epoch': '7.832'}
+
 52%|█████▏    | 250/480 [20:58<18:57,  4.95s/it]
 52%|█████▏    | 251/480 [21:02<18:38,  4.88s/it]
 52%|█████▎    | 252/480 [21:07<18:06,  4.77s/it]
 53%|█████▎    | 253/480 [21:11<17:42,  4.68s/it]
 53%|█████▎    | 254/480 [21:16<17:44,  4.71s/it]
 53%|█████▎    | 255/480 [21:22<18:29,  4.93s/it]
 53%|█████▎    | 256/480 [21:23<14:16,  3.82s/it]
 54%|█████▎    | 257/480 [21:28<15:34,  4.19s/it]
 54%|█████▍    | 258/480 [21:33<16:40,  4.51s/it]
 54%|█████▍    | 259/480 [21:38<16:42,  4.54s/it]
 54%|█████▍    | 260/480 [21:43<16:51,  4.60s/it]
                                                 
{'loss': '1.938', 'grad_norm': '0.4942', 'learning_rate': '5e-05', 'epoch': '8.128'}
+
 54%|█████▍    | 260/480 [21:43<16:51,  4.60s/it]
 54%|█████▍    | 261/480 [21:47<17:02,  4.67s/it]
 55%|█████▍    | 262/480 [21:52<16:53,  4.65s/it]
 55%|█████▍    | 263/480 [21:57<17:10,  4.75s/it]
 55%|█████▌    | 264/480 [22:02<17:13,  4.79s/it]
 55%|█████▌    | 265/480 [22:06<16:48,  4.69s/it]
 55%|█████▌    | 266/480 [22:11<16:54,  4.74s/it]
 56%|█████▌    | 267/480 [22:16<16:53,  4.76s/it]
 56%|█████▌    | 268/480 [22:21<16:44,  4.74s/it]
 56%|█████▌    | 269/480 [22:25<16:40,  4.74s/it]
 56%|█████▋    | 270/480 [22:30<16:38,  4.76s/it]
                                                 
{'loss': '1.969', 'grad_norm': '0.681', 'learning_rate': '5e-05', 'epoch': '8.448'}
+
 56%|█████▋    | 270/480 [22:30<16:38,  4.76s/it]
 56%|█████▋    | 271/480 [22:35<16:57,  4.87s/it]
 57%|█████▋    | 272/480 [22:40<16:43,  4.83s/it]
 57%|█████▋    | 273/480 [22:45<16:55,  4.91s/it]
 57%|█████▋    | 274/480 [22:50<16:54,  4.92s/it]
 57%|█████▋    | 275/480 [22:55<16:54,  4.95s/it]
 57%|█████▊    | 276/480 [23:00<16:52,  4.96s/it]
 58%|█████▊    | 277/480 [23:05<17:05,  5.05s/it]
 58%|█████▊    | 278/480 [23:11<17:14,  5.12s/it]
 58%|█████▊    | 279/480 [23:16<17:07,  5.11s/it]
 58%|█████▊    | 280/480 [23:21<16:58,  5.09s/it]
                                                 
{'loss': '1.883', 'grad_norm': '0.4303', 'learning_rate': '5e-05', 'epoch': '8.768'}
+
 58%|█████▊    | 280/480 [23:21<16:58,  5.09s/it]
 59%|█████▊    | 281/480 [23:25<16:20,  4.93s/it]
 59%|█████▉    | 282/480 [23:30<16:05,  4.87s/it]
 59%|█████▉    | 283/480 [23:35<15:54,  4.85s/it]
 59%|█████▉    | 284/480 [23:40<15:49,  4.84s/it]
 59%|█████▉    | 285/480 [23:44<15:34,  4.79s/it]
 60%|█████▉    | 286/480 [23:49<15:28,  4.79s/it]
 60%|█████▉    | 287/480 [23:53<14:56,  4.65s/it]
 60%|██████    | 288/480 [23:55<11:36,  3.63s/it]
 60%|██████    | 289/480 [24:00<12:54,  4.05s/it]
 60%|██████    | 290/480 [24:05<13:40,  4.32s/it]
                                                 
{'loss': '1.977', 'grad_norm': '0.4055', 'learning_rate': '5e-05', 'epoch': '9.064'}
+
 60%|██████    | 290/480 [24:05<13:40,  4.32s/it]
 61%|██████    | 291/480 [24:10<14:06,  4.48s/it]
 61%|██████    | 292/480 [24:15<14:28,  4.62s/it]
 61%|██████    | 293/480 [24:20<15:03,  4.83s/it]
 61%|██████▏   | 294/480 [24:24<14:33,  4.69s/it]
 61%|██████▏   | 295/480 [24:29<14:23,  4.67s/it]
 62%|██████▏   | 296/480 [24:34<14:37,  4.77s/it]
 62%|██████▏   | 297/480 [24:39<14:54,  4.89s/it]
 62%|██████▏   | 298/480 [24:44<14:51,  4.90s/it]
 62%|██████▏   | 299/480 [24:49<14:48,  4.91s/it]
 62%|██████▎   | 300/480 [24:54<14:46,  4.93s/it]
                                                 
{'loss': '1.893', 'grad_norm': '0.5018', 'learning_rate': '5e-05', 'epoch': '9.384'}
+
 62%|██████▎   | 300/480 [24:54<14:46,  4.93s/it]
 63%|██████▎   | 301/480 [24:58<14:27,  4.85s/it]
 63%|██████▎   | 302/480 [25:04<14:37,  4.93s/it]
 63%|██████▎   | 303/480 [25:09<14:58,  5.07s/it]
 63%|██████▎   | 304/480 [25:14<14:56,  5.09s/it]
 64%|██████▎   | 305/480 [25:19<14:51,  5.09s/it]
 64%|██████▍   | 306/480 [25:24<14:26,  4.98s/it]
 64%|██████▍   | 307/480 [25:29<14:10,  4.91s/it]
 64%|██████▍   | 308/480 [25:34<14:10,  4.95s/it]
 64%|██████▍   | 309/480 [25:38<13:53,  4.87s/it]
 65%|██████▍   | 310/480 [25:43<13:37,  4.81s/it]
                                                 
{'loss': '1.95', 'grad_norm': '0.5064', 'learning_rate': '5e-05', 'epoch': '9.704'}
+
 65%|██████▍   | 310/480 [25:43<13:37,  4.81s/it]
 65%|██████▍   | 311/480 [25:48<13:32,  4.81s/it]
 65%|██████▌   | 312/480 [25:53<13:19,  4.76s/it]
 65%|██████▌   | 313/480 [25:58<13:30,  4.86s/it]
 65%|██████▌   | 314/480 [26:02<13:08,  4.75s/it]
 66%|██████▌   | 315/480 [26:07<13:16,  4.83s/it]
 66%|██████▌   | 316/480 [26:12<12:49,  4.69s/it]
 66%|██████▌   | 317/480 [26:17<13:15,  4.88s/it]
 66%|██████▋   | 318/480 [26:22<13:07,  4.86s/it]
 66%|██████▋   | 319/480 [26:27<13:18,  4.96s/it]
 67%|██████▋   | 320/480 [26:28<10:12,  3.83s/it]
                                                 
{'loss': '1.873', 'grad_norm': '1.818', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [26:28<10:12,  3.83s/it]
+
Eval @ step 320:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 320:   1%|          | 1/100 [00:00<00:53,  1.86it/s]
+
Eval @ step 320:   2%|▏         | 2/100 [00:00<00:43,  2.27it/s]
+
Eval @ step 320:   3%|▎         | 3/100 [00:01<00:58,  1.66it/s]
+
Eval @ step 320:   4%|▍         | 4/100 [00:02<00:50,  1.91it/s]
+
Eval @ step 320:   5%|▌         | 5/100 [00:02<00:42,  2.21it/s]
+
Eval @ step 320:   6%|▌         | 6/100 [00:02<00:37,  2.54it/s]
+
Eval @ step 320:   7%|▋         | 7/100 [00:02<00:32,  2.89it/s]
+
Eval @ step 320:   8%|▊         | 8/100 [00:03<00:33,  2.74it/s]
+
Eval @ step 320:   9%|▉         | 9/100 [00:03<00:33,  2.75it/s]
+
Eval @ step 320:  10%|█         | 10/100 [00:04<00:35,  2.56it/s]
+
Eval @ step 320:  11%|█         | 11/100 [00:04<00:37,  2.39it/s]
+
Eval @ step 320:  12%|█▏        | 12/100 [00:05<00:41,  2.12it/s]
+
Eval @ step 320:  13%|█▎        | 13/100 [00:05<00:41,  2.09it/s]
+
Eval @ step 320:  14%|█▍        | 14/100 [00:06<00:46,  1.84it/s]
+
Eval @ step 320:  15%|█▌        | 15/100 [00:06<00:43,  1.94it/s]
+
Eval @ step 320:  16%|█▌        | 16/100 [00:07<00:39,  2.13it/s]
+
Eval @ step 320:  17%|█▋        | 17/100 [00:07<00:39,  2.12it/s]
+
Eval @ step 320:  18%|█▊        | 18/100 [00:08<00:46,  1.76it/s]
+
Eval @ step 320:  19%|█▉        | 19/100 [00:08<00:41,  1.95it/s]
+
Eval @ step 320:  20%|██        | 20/100 [00:09<00:36,  2.21it/s]
+
Eval @ step 320:  21%|██        | 21/100 [00:09<00:36,  2.15it/s]
+
Eval @ step 320:  22%|██▏       | 22/100 [00:10<00:35,  2.17it/s]
+
Eval @ step 320:  23%|██▎       | 23/100 [00:10<00:33,  2.32it/s]
+
Eval @ step 320:  24%|██▍       | 24/100 [00:11<00:44,  1.70it/s]
+
Eval @ step 320:  25%|██▌       | 25/100 [00:12<00:45,  1.65it/s]
+
Eval @ step 320:  26%|██▌       | 26/100 [00:12<00:42,  1.76it/s]
+
Eval @ step 320:  27%|██▋       | 27/100 [00:12<00:35,  2.04it/s]
+
Eval @ step 320:  28%|██▊       | 28/100 [00:13<00:38,  1.86it/s]
+
Eval @ step 320:  29%|██▉       | 29/100 [00:13<00:34,  2.05it/s]
+
Eval @ step 320:  30%|███       | 30/100 [00:14<00:36,  1.91it/s]
+
Eval @ step 320:  31%|███       | 31/100 [00:14<00:34,  2.01it/s]
+
Eval @ step 320:  32%|███▏      | 32/100 [00:15<00:41,  1.62it/s]
+
Eval @ step 320:  33%|███▎      | 33/100 [00:16<00:40,  1.67it/s]
+
Eval @ step 320:  34%|███▍      | 34/100 [00:17<00:46,  1.41it/s]
+
Eval @ step 320:  35%|███▌      | 35/100 [00:17<00:37,  1.75it/s]
+
Eval @ step 320:  36%|███▌      | 36/100 [00:18<00:36,  1.77it/s]
+
Eval @ step 320:  37%|███▋      | 37/100 [00:18<00:30,  2.08it/s]
+
Eval @ step 320:  38%|███▊      | 38/100 [00:19<00:37,  1.68it/s]
+
Eval @ step 320:  39%|███▉      | 39/100 [00:19<00:34,  1.78it/s]
+
Eval @ step 320:  40%|████      | 40/100 [00:20<00:31,  1.90it/s]
+
Eval @ step 320:  41%|████      | 41/100 [00:20<00:27,  2.11it/s]
+
Eval @ step 320:  42%|████▏     | 42/100 [00:21<00:26,  2.15it/s]
+
Eval @ step 320:  43%|████▎     | 43/100 [00:22<00:36,  1.56it/s]
+
Eval @ step 320:  44%|████▍     | 44/100 [00:22<00:33,  1.68it/s]
+
Eval @ step 320:  45%|████▌     | 45/100 [00:23<00:29,  1.86it/s]
+
Eval @ step 320:  46%|████▌     | 46/100 [00:23<00:34,  1.55it/s]
+
Eval @ step 320:  47%|████▋     | 47/100 [00:24<00:36,  1.44it/s]
+
Eval @ step 320:  48%|████▊     | 48/100 [00:25<00:30,  1.72it/s]
+
Eval @ step 320:  49%|████▉     | 49/100 [00:25<00:31,  1.61it/s]
+
Eval @ step 320:  50%|█████     | 50/100 [00:26<00:32,  1.55it/s]
+
Eval @ step 320:  51%|█████     | 51/100 [00:27<00:33,  1.47it/s]
+
Eval @ step 320:  52%|█████▏    | 52/100 [00:27<00:27,  1.74it/s]
+
Eval @ step 320:  53%|█████▎    | 53/100 [00:27<00:25,  1.86it/s]
+
Eval @ step 320:  54%|█████▍    | 54/100 [00:28<00:25,  1.84it/s]
+
Eval @ step 320:  55%|█████▌    | 55/100 [00:28<00:21,  2.14it/s]
+
Eval @ step 320:  56%|█████▌    | 56/100 [00:29<00:21,  2.07it/s]
+
Eval @ step 320:  57%|█████▋    | 57/100 [00:29<00:18,  2.36it/s]
+
Eval @ step 320:  58%|█████▊    | 58/100 [00:29<00:15,  2.70it/s]
+
Eval @ step 320:  59%|█████▉    | 59/100 [00:30<00:17,  2.34it/s]
+
Eval @ step 320:  60%|██████    | 60/100 [00:31<00:21,  1.89it/s]
+
Eval @ step 320:  61%|██████    | 61/100 [00:31<00:17,  2.18it/s]
+
Eval @ step 320:  62%|██████▏   | 62/100 [00:32<00:18,  2.06it/s]
+
Eval @ step 320:  63%|██████▎   | 63/100 [00:32<00:19,  1.87it/s]
+
Eval @ step 320:  64%|██████▍   | 64/100 [00:33<00:22,  1.62it/s]
+
Eval @ step 320:  65%|██████▌   | 65/100 [00:34<00:20,  1.70it/s]
+
Eval @ step 320:  66%|██████▌   | 66/100 [00:34<00:18,  1.79it/s]
+
Eval @ step 320:  67%|██████▋   | 67/100 [00:34<00:17,  1.90it/s]
+
Eval @ step 320:  68%|██████▊   | 68/100 [00:35<00:17,  1.85it/s]
+
Eval @ step 320:  69%|██████▉   | 69/100 [00:36<00:17,  1.76it/s]
+
Eval @ step 320:  70%|███████   | 70/100 [00:36<00:15,  1.99it/s]
+
Eval @ step 320:  71%|███████   | 71/100 [00:37<00:15,  1.88it/s]
+
Eval @ step 320:  72%|███████▏  | 72/100 [00:37<00:16,  1.65it/s]
+
Eval @ step 320:  73%|███████▎  | 73/100 [00:38<00:20,  1.35it/s]
+
Eval @ step 320:  74%|███████▍  | 74/100 [00:39<00:17,  1.46it/s]
+
Eval @ step 320:  75%|███████▌  | 75/100 [00:39<00:13,  1.81it/s]
+
Eval @ step 320:  76%|███████▌  | 76/100 [00:40<00:15,  1.51it/s]
+
Eval @ step 320:  77%|███████▋  | 77/100 [00:41<00:13,  1.67it/s]
+
Eval @ step 320:  78%|███████▊  | 78/100 [00:41<00:13,  1.63it/s]
+
Eval @ step 320:  79%|███████▉  | 79/100 [00:42<00:11,  1.90it/s]
+
Eval @ step 320:  80%|████████  | 80/100 [00:42<00:11,  1.72it/s]
+
Eval @ step 320:  81%|████████  | 81/100 [00:43<00:13,  1.37it/s]
+
Eval @ step 320:  82%|████████▏ | 82/100 [00:44<00:14,  1.23it/s]
+
Eval @ step 320:  83%|████████▎ | 83/100 [00:45<00:13,  1.26it/s]
+
Eval @ step 320:  84%|████████▍ | 84/100 [00:46<00:10,  1.46it/s]
+
Eval @ step 320:  85%|████████▌ | 85/100 [00:46<00:09,  1.63it/s]
+
Eval @ step 320:  86%|████████▌ | 86/100 [00:47<00:09,  1.52it/s]
+
Eval @ step 320:  87%|████████▋ | 87/100 [00:47<00:07,  1.69it/s]
+
Eval @ step 320:  88%|████████▊ | 88/100 [00:48<00:07,  1.70it/s]
+
Eval @ step 320:  89%|████████▉ | 89/100 [00:48<00:05,  2.07it/s]
+
Eval @ step 320:  90%|█████████ | 90/100 [00:49<00:04,  2.05it/s]
+
Eval @ step 320:  91%|█████████ | 91/100 [00:49<00:04,  2.10it/s]
+
Eval @ step 320:  92%|█████████▏| 92/100 [00:49<00:03,  2.38it/s]
+
Eval @ step 320:  93%|█████████▎| 93/100 [00:50<00:02,  2.56it/s]
+
Eval @ step 320:  94%|█████████▍| 94/100 [00:50<00:02,  2.56it/s]
+
Eval @ step 320:  95%|█████████▌| 95/100 [00:51<00:02,  2.14it/s]
+
Eval @ step 320:  96%|█████████▌| 96/100 [00:51<00:01,  2.09it/s]
+
Eval @ step 320:  97%|█████████▋| 97/100 [00:51<00:01,  2.46it/s]
+
Eval @ step 320:  98%|█████████▊| 98/100 [00:52<00:01,  1.96it/s]
+
Eval @ step 320:  99%|█████████▉| 99/100 [00:53<00:00,  1.68it/s]
+
Eval @ step 320: 100%|██████████| 100/100 [00:53<00:00,  1.91it/s]
+
                                                                    [EVAL] eval_rougeL=17.5688 (best=18.0212)
+
                                                 
{'eval_rougeL': '17.57', 'epoch': '10'}
+
 67%|██████▋   | 320/480 [27:22<10:12,  3.83s/it]
 67%|██████▋   | 321/480 [27:27<54:00, 20.38s/it]
 67%|██████▋   | 322/480 [27:32<41:20, 15.70s/it]
 67%|██████▋   | 323/480 [27:37<32:32, 12.43s/it]
 68%|██████▊   | 324/480 [27:42<26:34, 10.22s/it]
 68%|██████▊   | 325/480 [27:46<22:11,  8.59s/it]
 68%|██████▊   | 326/480 [27:51<19:04,  7.43s/it]
 68%|██████▊   | 327/480 [27:56<17:01,  6.68s/it]
 68%|██████▊   | 328/480 [28:01<15:51,  6.26s/it]
 69%|██████▊   | 329/480 [28:06<14:32,  5.78s/it]
 69%|██████▉   | 330/480 [28:11<14:00,  5.60s/it]
                                                 
{'loss': '1.888', 'grad_norm': '0.6228', 'learning_rate': '5e-05', 'epoch': '10.32'}
+
 69%|██████▉   | 330/480 [28:11<14:00,  5.60s/it]
 69%|██████▉   | 331/480 [28:16<13:19,  5.36s/it]
 69%|██████▉   | 332/480 [28:21<12:52,  5.22s/it]
 69%|██████▉   | 333/480 [28:26<12:26,  5.08s/it]
 70%|██████▉   | 334/480 [28:31<12:15,  5.04s/it]
 70%|██████▉   | 335/480 [28:36<12:03,  4.99s/it]
 70%|███████   | 336/480 [28:40<11:37,  4.85s/it]
 70%|███████   | 337/480 [28:45<11:55,  5.01s/it]
 70%|███████   | 338/480 [28:50<11:45,  4.97s/it]
 71%|███████   | 339/480 [28:55<11:42,  4.98s/it]
 71%|███████   | 340/480 [29:00<11:24,  4.89s/it]
                                                 
{'loss': '1.92', 'grad_norm': '0.7758', 'learning_rate': '5e-05', 'epoch': '10.64'}
+
 71%|███████   | 340/480 [29:00<11:24,  4.89s/it]
 71%|███████   | 341/480 [29:05<11:17,  4.87s/it]
 71%|███████▏  | 342/480 [29:10<11:19,  4.92s/it]
 71%|███████▏  | 343/480 [29:15<11:03,  4.85s/it]
 72%|███████▏  | 344/480 [29:19<10:56,  4.83s/it]
 72%|███████▏  | 345/480 [29:24<10:50,  4.82s/it]
 72%|███████▏  | 346/480 [29:29<10:50,  4.85s/it]
 72%|███████▏  | 347/480 [29:34<10:47,  4.87s/it]
 72%|███████▎  | 348/480 [29:39<10:54,  4.96s/it]
 73%|███████▎  | 349/480 [29:44<10:43,  4.91s/it]
 73%|███████▎  | 350/480 [29:49<10:43,  4.95s/it]
                                                 
{'loss': '1.902', 'grad_norm': '1.004', 'learning_rate': '5e-05', 'epoch': '10.96'}
+
 73%|███████▎  | 350/480 [29:49<10:43,  4.95s/it]
 73%|███████▎  | 351/480 [29:54<10:37,  4.94s/it]
 73%|███████▎  | 352/480 [29:55<08:12,  3.85s/it]
 74%|███████▎  | 353/480 [30:00<08:52,  4.19s/it]
 74%|███████▍  | 354/480 [30:05<09:03,  4.31s/it]
 74%|███████▍  | 355/480 [30:09<09:11,  4.42s/it]
 74%|███████▍  | 356/480 [30:14<09:26,  4.57s/it]
 74%|███████▍  | 357/480 [30:19<09:41,  4.73s/it]
 75%|███████▍  | 358/480 [30:24<09:42,  4.78s/it]
 75%|███████▍  | 359/480 [30:30<09:56,  4.93s/it]
 75%|███████▌  | 360/480 [30:35<09:54,  4.95s/it]
                                                 
{'loss': '1.943', 'grad_norm': '0.5017', 'learning_rate': '5e-05', 'epoch': '11.26'}
+
 75%|███████▌  | 360/480 [30:35<09:54,  4.95s/it]
 75%|███████▌  | 361/480 [30:39<09:30,  4.79s/it]
 75%|███████▌  | 362/480 [30:44<09:32,  4.85s/it]
 76%|███████▌  | 363/480 [30:49<09:45,  5.01s/it]
 76%|███████▌  | 364/480 [30:54<09:42,  5.02s/it]
 76%|███████▌  | 365/480 [30:59<09:22,  4.89s/it]
 76%|███████▋  | 366/480 [31:04<09:13,  4.86s/it]
 76%|███████▋  | 367/480 [31:09<09:17,  4.94s/it]
 77%|███████▋  | 368/480 [31:14<09:01,  4.84s/it]
 77%|███████▋  | 369/480 [31:18<08:50,  4.78s/it]
 77%|███████▋  | 370/480 [31:23<08:53,  4.85s/it]
                                                 
{'loss': '1.882', 'grad_norm': '0.5755', 'learning_rate': '5e-05', 'epoch': '11.58'}
+
 77%|███████▋  | 370/480 [31:23<08:53,  4.85s/it]
 77%|███████▋  | 371/480 [31:28<08:50,  4.87s/it]
 78%|███████▊  | 372/480 [31:33<08:54,  4.95s/it]
 78%|███████▊  | 373/480 [31:38<08:48,  4.93s/it]
 78%|███████▊  | 374/480 [31:43<08:40,  4.91s/it]
 78%|███████▊  | 375/480 [31:48<08:34,  4.90s/it]
 78%|███████▊  | 376/480 [31:53<08:30,  4.91s/it]
 79%|███████▊  | 377/480 [31:58<08:32,  4.97s/it]
 79%|███████▉  | 378/480 [32:03<08:25,  4.96s/it]
 79%|███████▉  | 379/480 [32:08<08:24,  4.99s/it]
 79%|███████▉  | 380/480 [32:13<08:18,  4.98s/it]
                                                 
{'loss': '1.9', 'grad_norm': '0.5196', 'learning_rate': '5e-05', 'epoch': '11.9'}
+
 79%|███████▉  | 380/480 [32:13<08:18,  4.98s/it]
 79%|███████▉  | 381/480 [32:18<08:05,  4.91s/it]
 80%|███████▉  | 382/480 [32:22<07:54,  4.84s/it]
 80%|███████▉  | 383/480 [32:27<07:49,  4.84s/it]
 80%|████████  | 384/480 [32:28<05:57,  3.73s/it]
 80%|████████  | 385/480 [32:33<06:23,  4.03s/it]
 80%|████████  | 386/480 [32:38<06:35,  4.21s/it]
 81%|████████  | 387/480 [32:42<06:47,  4.38s/it]
 81%|████████  | 388/480 [32:47<06:51,  4.48s/it]
 81%|████████  | 389/480 [32:52<06:53,  4.54s/it]
 81%|████████▏ | 390/480 [32:56<06:52,  4.58s/it]
                                                 
{'loss': '1.911', 'grad_norm': '0.5263', 'learning_rate': '5e-05', 'epoch': '12.19'}
+
 81%|████████▏ | 390/480 [32:56<06:52,  4.58s/it]
 81%|████████▏ | 391/480 [33:01<06:57,  4.69s/it]
 82%|████████▏ | 392/480 [33:06<06:59,  4.77s/it]
 82%|████████▏ | 393/480 [33:12<07:10,  4.95s/it]
 82%|████████▏ | 394/480 [33:16<06:53,  4.81s/it]
 82%|████████▏ | 395/480 [33:21<06:55,  4.89s/it]
 82%|████████▎ | 396/480 [33:26<06:47,  4.85s/it]
 83%|████████▎ | 397/480 [33:31<06:39,  4.81s/it]
 83%|████████▎ | 398/480 [33:36<06:36,  4.83s/it]
 83%|████████▎ | 399/480 [33:41<06:33,  4.86s/it]
 83%|████████▎ | 400/480 [33:45<06:24,  4.81s/it]
                                                 
{'loss': '1.897', 'grad_norm': '0.5021', 'learning_rate': '5e-05', 'epoch': '12.51'}
+
 83%|████████▎ | 400/480 [33:45<06:24,  4.81s/it]
 84%|████████▎ | 401/480 [33:50<06:21,  4.83s/it]
 84%|████████▍ | 402/480 [33:55<06:16,  4.83s/it]
 84%|████████▍ | 403/480 [34:00<06:13,  4.86s/it]
 84%|████████▍ | 404/480 [34:05<06:05,  4.81s/it]
 84%|████████▍ | 405/480 [34:10<06:06,  4.89s/it]
 85%|████████▍ | 406/480 [34:15<06:04,  4.92s/it]
 85%|████████▍ | 407/480 [34:20<06:00,  4.94s/it]
 85%|████████▌ | 408/480 [34:25<05:57,  4.96s/it]
 85%|████████▌ | 409/480 [34:29<05:48,  4.91s/it]
 85%|████████▌ | 410/480 [34:34<05:36,  4.81s/it]
                                                 
{'loss': '1.861', 'grad_norm': '0.5635', 'learning_rate': '5e-05', 'epoch': '12.83'}
+
 85%|████████▌ | 410/480 [34:34<05:36,  4.81s/it]
 86%|████████▌ | 411/480 [34:39<05:29,  4.77s/it]
 86%|████████▌ | 412/480 [34:44<05:28,  4.82s/it]
 86%|████████▌ | 413/480 [34:48<05:19,  4.78s/it]
 86%|████████▋ | 414/480 [34:54<05:24,  4.91s/it]
 86%|████████▋ | 415/480 [34:58<05:18,  4.90s/it]
 87%|████████▋ | 416/480 [35:00<04:02,  3.79s/it]
 87%|████████▋ | 417/480 [35:04<04:15,  4.05s/it]
 87%|████████▋ | 418/480 [35:09<04:20,  4.20s/it]
 87%|████████▋ | 419/480 [35:14<04:25,  4.36s/it]
 88%|████████▊ | 420/480 [35:18<04:29,  4.48s/it]
                                                 
{'loss': '1.867', 'grad_norm': '0.5138', 'learning_rate': '5e-05', 'epoch': '13.13'}
+
 88%|████████▊ | 420/480 [35:18<04:29,  4.48s/it]
 88%|████████▊ | 421/480 [35:23<04:35,  4.67s/it]
 88%|████████▊ | 422/480 [35:28<04:32,  4.69s/it]
 88%|████████▊ | 423/480 [35:33<04:30,  4.75s/it]
 88%|████████▊ | 424/480 [35:38<04:29,  4.82s/it]
 89%|████████▊ | 425/480 [35:43<04:23,  4.79s/it]
 89%|████████▉ | 426/480 [35:48<04:21,  4.84s/it]
 89%|████████▉ | 427/480 [35:53<04:19,  4.89s/it]
 89%|████████▉ | 428/480 [35:58<04:16,  4.93s/it]
 89%|████████▉ | 429/480 [36:03<04:08,  4.88s/it]
 90%|████████▉ | 430/480 [36:08<04:06,  4.92s/it]
                                                 
{'loss': '1.859', 'grad_norm': '0.7345', 'learning_rate': '5e-05', 'epoch': '13.45'}
+
 90%|████████▉ | 430/480 [36:08<04:06,  4.92s/it]
 90%|████████▉ | 431/480 [36:13<04:03,  4.97s/it]
 90%|█████████ | 432/480 [36:18<04:06,  5.13s/it]
 90%|█████████ | 433/480 [36:23<04:00,  5.11s/it]
 90%|█████████ | 434/480 [36:28<03:57,  5.17s/it]
 91%|█████████ | 435/480 [36:33<03:46,  5.04s/it]
 91%|█████████ | 436/480 [36:38<03:40,  5.02s/it]
 91%|█████████ | 437/480 [36:43<03:31,  4.91s/it]
 91%|█████████▏| 438/480 [36:48<03:22,  4.83s/it]
 91%|█████████▏| 439/480 [36:52<03:16,  4.80s/it]
 92%|█████████▏| 440/480 [36:57<03:10,  4.77s/it]
                                                 
{'loss': '1.875', 'grad_norm': '0.5194', 'learning_rate': '5e-05', 'epoch': '13.77'}
+
 92%|█████████▏| 440/480 [36:57<03:10,  4.77s/it]
 92%|█████████▏| 441/480 [37:02<03:04,  4.73s/it]
 92%|█████████▏| 442/480 [37:06<02:59,  4.72s/it]
 92%|█████████▏| 443/480 [37:11<02:55,  4.74s/it]
 92%|█████████▎| 444/480 [37:16<02:50,  4.74s/it]
 93%|█████████▎| 445/480 [37:21<02:46,  4.74s/it]
 93%|█████████▎| 446/480 [37:25<02:41,  4.76s/it]
 93%|█████████▎| 447/480 [37:30<02:37,  4.79s/it]
 93%|█████████▎| 448/480 [37:31<01:58,  3.72s/it]
 94%|█████████▎| 449/480 [37:36<02:03,  4.00s/it]
 94%|█████████▍| 450/480 [37:42<02:14,  4.49s/it]
                                                 
{'loss': '1.865', 'grad_norm': '0.3632', 'learning_rate': '5e-05', 'epoch': '14.06'}
+
 94%|█████████▍| 450/480 [37:42<02:14,  4.49s/it]
 94%|█████████▍| 451/480 [37:47<02:12,  4.59s/it]
 94%|█████████▍| 452/480 [37:51<02:07,  4.56s/it]
 94%|█████████▍| 453/480 [37:56<02:08,  4.74s/it]
 95%|█████████▍| 454/480 [38:01<02:05,  4.82s/it]
 95%|█████████▍| 455/480 [38:06<02:00,  4.83s/it]
 95%|█████████▌| 456/480 [38:11<01:56,  4.87s/it]
 95%|█████████▌| 457/480 [38:16<01:51,  4.85s/it]
 95%|█████████▌| 458/480 [38:20<01:44,  4.74s/it]
 96%|█████████▌| 459/480 [38:25<01:41,  4.83s/it]
 96%|█████████▌| 460/480 [38:30<01:36,  4.81s/it]
                                                 
{'loss': '1.864', 'grad_norm': '0.6945', 'learning_rate': '5e-05', 'epoch': '14.38'}
+
 96%|█████████▌| 460/480 [38:30<01:36,  4.81s/it]
 96%|█████████▌| 461/480 [38:35<01:32,  4.88s/it]
 96%|█████████▋| 462/480 [38:40<01:26,  4.83s/it]
 96%|█████████▋| 463/480 [38:45<01:21,  4.81s/it]
 97%|█████████▋| 464/480 [38:49<01:16,  4.76s/it]
 97%|█████████▋| 465/480 [38:54<01:12,  4.81s/it]
 97%|█████████▋| 466/480 [38:59<01:06,  4.78s/it]
 97%|█████████▋| 467/480 [39:04<01:01,  4.75s/it]
 98%|█████████▊| 468/480 [39:08<00:57,  4.77s/it]
 98%|█████████▊| 469/480 [39:13<00:52,  4.80s/it]
 98%|█████████▊| 470/480 [39:18<00:48,  4.83s/it]
                                                 
{'loss': '1.857', 'grad_norm': '0.4279', 'learning_rate': '5e-05', 'epoch': '14.7'}
+
 98%|█████████▊| 470/480 [39:18<00:48,  4.83s/it]
 98%|█████████▊| 471/480 [39:23<00:42,  4.74s/it]
 98%|█████████▊| 472/480 [39:27<00:37,  4.73s/it]
 99%|█████████▊| 473/480 [39:32<00:33,  4.78s/it]
 99%|█████████▉| 474/480 [39:38<00:29,  4.99s/it]
 99%|█████████▉| 475/480 [39:43<00:25,  5.03s/it]
 99%|█████████▉| 476/480 [39:48<00:19,  4.98s/it]
 99%|█████████▉| 477/480 [39:53<00:15,  5.05s/it]
100%|█████████▉| 478/480 [39:58<00:09,  4.99s/it]
100%|█████████▉| 479/480 [40:03<00:05,  5.04s/it]
100%|██████████| 480/480 [40:05<00:00,  3.98s/it]
                                                 
{'loss': '1.878', 'grad_norm': '1.928', 'learning_rate': '5e-05', 'epoch': '15'}
+
100%|██████████| 480/480 [40:05<00:00,  3.98s/it]
+
Eval @ step 480:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 480:   1%|          | 1/100 [00:00<00:28,  3.45it/s]
+
Eval @ step 480:   2%|▏         | 2/100 [00:00<00:40,  2.43it/s]
+
Eval @ step 480:   3%|▎         | 3/100 [00:01<00:49,  1.96it/s]
+
Eval @ step 480:   4%|▍         | 4/100 [00:02<00:54,  1.75it/s]
+
Eval @ step 480:   5%|▌         | 5/100 [00:02<00:55,  1.71it/s]
+
Eval @ step 480:   6%|▌         | 6/100 [00:03<01:03,  1.48it/s]
+
Eval @ step 480:   7%|▋         | 7/100 [00:03<00:49,  1.87it/s]
+
Eval @ step 480:   8%|▊         | 8/100 [00:04<00:45,  2.02it/s]
+
Eval @ step 480:   9%|▉         | 9/100 [00:04<00:44,  2.06it/s]
+
Eval @ step 480:  10%|█         | 10/100 [00:05<00:42,  2.11it/s]
+
Eval @ step 480:  11%|█         | 11/100 [00:05<00:43,  2.04it/s]
+
Eval @ step 480:  12%|█▏        | 12/100 [00:06<00:41,  2.14it/s]
+
Eval @ step 480:  13%|█▎        | 13/100 [00:06<00:41,  2.10it/s]
+
Eval @ step 480:  14%|█▍        | 14/100 [00:07<00:51,  1.68it/s]
+
Eval @ step 480:  15%|█▌        | 15/100 [00:07<00:46,  1.81it/s]
+
Eval @ step 480:  16%|█▌        | 16/100 [00:08<00:40,  2.07it/s]
+
Eval @ step 480:  17%|█▋        | 17/100 [00:08<00:40,  2.07it/s]
+
Eval @ step 480:  18%|█▊        | 18/100 [00:09<00:39,  2.07it/s]
+
Eval @ step 480:  19%|█▉        | 19/100 [00:09<00:33,  2.42it/s]
+
Eval @ step 480:  20%|██        | 20/100 [00:09<00:30,  2.61it/s]
+
Eval @ step 480:  21%|██        | 21/100 [00:10<00:30,  2.63it/s]
+
Eval @ step 480:  22%|██▏       | 22/100 [00:10<00:31,  2.50it/s]
+
Eval @ step 480:  23%|██▎       | 23/100 [00:10<00:29,  2.65it/s]
+
Eval @ step 480:  24%|██▍       | 24/100 [00:11<00:34,  2.22it/s]
+
Eval @ step 480:  25%|██▌       | 25/100 [00:11<00:34,  2.15it/s]
+
Eval @ step 480:  26%|██▌       | 26/100 [00:12<00:35,  2.07it/s]
+
Eval @ step 480:  27%|██▋       | 27/100 [00:12<00:33,  2.18it/s]
+
Eval @ step 480:  28%|██▊       | 28/100 [00:13<00:33,  2.15it/s]
+
Eval @ step 480:  29%|██▉       | 29/100 [00:14<00:44,  1.60it/s]
+
Eval @ step 480:  30%|███       | 30/100 [00:15<00:44,  1.56it/s]
+
Eval @ step 480:  31%|███       | 31/100 [00:15<00:38,  1.78it/s]
+
Eval @ step 480:  32%|███▏      | 32/100 [00:16<00:44,  1.51it/s]
+
Eval @ step 480:  33%|███▎      | 33/100 [00:16<00:42,  1.59it/s]
+
Eval @ step 480:  34%|███▍      | 34/100 [00:17<00:37,  1.77it/s]
+
Eval @ step 480:  35%|███▌      | 35/100 [00:17<00:33,  1.94it/s]
+
Eval @ step 480:  36%|███▌      | 36/100 [00:18<00:33,  1.90it/s]
+
Eval @ step 480:  37%|███▋      | 37/100 [00:18<00:33,  1.91it/s]
+
Eval @ step 480:  38%|███▊      | 38/100 [00:19<00:39,  1.58it/s]
+
Eval @ step 480:  39%|███▉      | 39/100 [00:20<00:33,  1.80it/s]
+
Eval @ step 480:  40%|████      | 40/100 [00:20<00:33,  1.81it/s]
+
Eval @ step 480:  41%|████      | 41/100 [00:20<00:29,  2.03it/s]
+
Eval @ step 480:  42%|████▏     | 42/100 [00:21<00:27,  2.09it/s]
+
Eval @ step 480:  43%|████▎     | 43/100 [00:21<00:25,  2.24it/s]
+
Eval @ step 480:  44%|████▍     | 44/100 [00:22<00:25,  2.19it/s]
+
Eval @ step 480:  45%|████▌     | 45/100 [00:22<00:26,  2.09it/s]
+
Eval @ step 480:  46%|████▌     | 46/100 [00:23<00:25,  2.13it/s]
+
Eval @ step 480:  47%|████▋     | 47/100 [00:24<00:30,  1.76it/s]
+
Eval @ step 480:  48%|████▊     | 48/100 [00:24<00:26,  1.93it/s]
+
Eval @ step 480:  49%|████▉     | 49/100 [00:25<00:28,  1.79it/s]
+
Eval @ step 480:  50%|█████     | 50/100 [00:25<00:30,  1.66it/s]
+
Eval @ step 480:  51%|█████     | 51/100 [00:26<00:32,  1.51it/s]
+
Eval @ step 480:  52%|█████▏    | 52/100 [00:26<00:25,  1.86it/s]
+
Eval @ step 480:  53%|█████▎    | 53/100 [00:27<00:24,  1.96it/s]
+
Eval @ step 480:  54%|█████▍    | 54/100 [00:27<00:23,  1.97it/s]
+
Eval @ step 480:  55%|█████▌    | 55/100 [00:27<00:18,  2.40it/s]
+
Eval @ step 480:  56%|█████▌    | 56/100 [00:28<00:19,  2.29it/s]
+
Eval @ step 480:  57%|█████▋    | 57/100 [00:28<00:18,  2.28it/s]
+
Eval @ step 480:  58%|█████▊    | 58/100 [00:29<00:20,  2.10it/s]
+
Eval @ step 480:  59%|█████▉    | 59/100 [00:30<00:22,  1.84it/s]
+
Eval @ step 480:  60%|██████    | 60/100 [00:30<00:19,  2.10it/s]
+
Eval @ step 480:  61%|██████    | 61/100 [00:30<00:16,  2.36it/s]
+
Eval @ step 480:  62%|██████▏   | 62/100 [00:31<00:16,  2.32it/s]
+
Eval @ step 480:  63%|██████▎   | 63/100 [00:31<00:18,  2.02it/s]
+
Eval @ step 480:  64%|██████▍   | 64/100 [00:32<00:22,  1.57it/s]
+
Eval @ step 480:  65%|██████▌   | 65/100 [00:33<00:20,  1.73it/s]
+
Eval @ step 480:  66%|██████▌   | 66/100 [00:33<00:19,  1.78it/s]
+
Eval @ step 480:  67%|██████▋   | 67/100 [00:34<00:17,  1.84it/s]
+
Eval @ step 480:  68%|██████▊   | 68/100 [00:34<00:17,  1.81it/s]
+
Eval @ step 480:  69%|██████▉   | 69/100 [00:35<00:17,  1.73it/s]
+
Eval @ step 480:  70%|███████   | 70/100 [00:35<00:14,  2.06it/s]
+
Eval @ step 480:  71%|███████   | 71/100 [00:36<00:12,  2.24it/s]
+
Eval @ step 480:  72%|███████▏  | 72/100 [00:36<00:15,  1.83it/s]
+
Eval @ step 480:  73%|███████▎  | 73/100 [00:37<00:15,  1.73it/s]
+
Eval @ step 480:  74%|███████▍  | 74/100 [00:38<00:15,  1.71it/s]
+
Eval @ step 480:  75%|███████▌  | 75/100 [00:38<00:12,  1.93it/s]
+
Eval @ step 480:  76%|███████▌  | 76/100 [00:39<00:14,  1.62it/s]
+
Eval @ step 480:  77%|███████▋  | 77/100 [00:39<00:12,  1.91it/s]
+
Eval @ step 480:  78%|███████▊  | 78/100 [00:40<00:12,  1.79it/s]
+
Eval @ step 480:  79%|███████▉  | 79/100 [00:40<00:12,  1.71it/s]
+
Eval @ step 480:  80%|████████  | 80/100 [00:41<00:12,  1.56it/s]
+
Eval @ step 480:  81%|████████  | 81/100 [00:42<00:13,  1.46it/s]
+
Eval @ step 480:  82%|████████▏ | 82/100 [00:43<00:12,  1.39it/s]
+
Eval @ step 480:  83%|████████▎ | 83/100 [00:44<00:12,  1.37it/s]
+
Eval @ step 480:  84%|████████▍ | 84/100 [00:44<00:11,  1.38it/s]
+
Eval @ step 480:  85%|████████▌ | 85/100 [00:45<00:09,  1.56it/s]
+
Eval @ step 480:  86%|████████▌ | 86/100 [00:45<00:09,  1.51it/s]
+
Eval @ step 480:  87%|████████▋ | 87/100 [00:46<00:07,  1.67it/s]
+
Eval @ step 480:  88%|████████▊ | 88/100 [00:46<00:07,  1.70it/s]
+
Eval @ step 480:  89%|████████▉ | 89/100 [00:47<00:05,  2.07it/s]
+
Eval @ step 480:  90%|█████████ | 90/100 [00:47<00:05,  1.98it/s]
+
Eval @ step 480:  91%|█████████ | 91/100 [00:48<00:04,  2.19it/s]
+
Eval @ step 480:  92%|█████████▏| 92/100 [00:48<00:04,  1.90it/s]
+
Eval @ step 480:  93%|█████████▎| 93/100 [00:49<00:03,  2.10it/s]
+
Eval @ step 480:  94%|█████████▍| 94/100 [00:49<00:03,  1.96it/s]
+
Eval @ step 480:  95%|█████████▌| 95/100 [00:50<00:02,  2.20it/s]
+
Eval @ step 480:  96%|█████████▌| 96/100 [00:50<00:01,  2.14it/s]
+
Eval @ step 480:  97%|█████████▋| 97/100 [00:50<00:01,  2.43it/s]
+
Eval @ step 480:  98%|█████████▊| 98/100 [00:51<00:00,  2.20it/s]
+
Eval @ step 480:  99%|█████████▉| 99/100 [00:52<00:00,  1.85it/s]
+
Eval @ step 480: 100%|██████████| 100/100 [00:52<00:00,  2.07it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=18.2202 (step=480) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '18.22', 'epoch': '15'}
+
100%|██████████| 480/480 [40:57<00:00,  3.98s/it]
                                                 
{'train_runtime': '2458', 'train_samples_per_second': '6.104', 'train_steps_per_second': '0.195', 'train_loss': '1.992', 'epoch': '15'}
+
100%|██████████| 480/480 [40:57<00:00,  3.98s/it]
100%|██████████| 480/480 [40:57<00:00,  5.12s/it]
+***** train metrics *****
+  epoch                    =        15.0
+  total_flos               = 148148168GF
+  train_loss               =      1.9922
+  train_runtime            =  0:40:57.56
+  train_samples            =        1000
+  train_samples_per_second =       6.104
+  train_steps_per_second   =       0.195
+  [EVAL] Restored best model (eval_rougeL=18.2202, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   1%|          | 2/291 [00:00<00:15, 18.59it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 160.07it/s]
Loading weights:  16%|█▌        | 47/291 [00:00<00:01, 164.87it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 167.23it/s]
Loading weights:  30%|██▉       | 86/291 [00:00<00:01, 172.15it/s]
Loading weights:  36%|███▌      | 105/291 [00:00<00:01, 176.71it/s]
Loading weights:  43%|████▎     | 126/291 [00:00<00:00, 182.27it/s]
Loading weights:  50%|████▉     | 145/291 [00:00<00:00, 182.50it/s]
Loading weights:  56%|█████▋    | 164/291 [00:00<00:00, 182.93it/s]
Loading weights:  63%|██████▎   | 183/291 [00:01<00:00, 180.53it/s]
Loading weights:  69%|██████▉   | 202/291 [00:01<00:00, 180.28it/s]
Loading weights:  76%|███████▌  | 221/291 [00:01<00:00, 180.81it/s]
Loading weights:  82%|████████▏ | 240/291 [00:01<00:00, 177.71it/s]
Loading weights:  89%|████████▊ | 258/291 [00:01<00:00, 171.02it/s]
Loading weights:  96%|█████████▌| 280/291 [00:01<00:00, 177.23it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 176.38it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (1000 samples, batch_size=1)
+  [SRT] Task 'task1729_personachat_generate_next': PaR=65.6, shrinkage=ridge, n=1000
+  [SRT] Wired router: 8 sigs, cur=task1729_personachat_generate_next, mapping={'task1729_personachat_generate_next': 0, 'task1572_samsum_summary': 7, 'task363_sst2_polarity_classification': 6, 'task1290_xsum_summarization': 5, 'task181_outcome_extraction': 4, 'task002_quoref_answer_generation': 3, 'task1510_evalution_relation_extraction': 2, 'task639_multi_woz_user_utterance_generation': 1}
+[SRT] Router: 8 tasks, avg_PaR=65.6, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next']
+  [SRT-INIT] task_id_to_idx={'task1729_personachat_generate_next': 0, 'task1572_samsum_summary': 7, 'task363_sst2_polarity_classification': 6, 'task1290_xsum_summarization': 5, 'task181_outcome_extraction': 4, 'task002_quoref_answer_generation': 3, 'task1510_evalution_relation_extraction': 2, 'task639_multi_woz_user_utterance_generation': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/581 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=8
+
Continual predict:   0%|          | 1/581 [00:01<17:08,  1.77s/it]
Continual predict:   0%|          | 2/581 [00:04<21:21,  2.21s/it]
Continual predict:   1%|          | 3/581 [00:05<15:27,  1.60s/it]
Continual predict:   1%|          | 4/581 [00:05<11:25,  1.19s/it]
Continual predict:   1%|          | 5/581 [00:07<13:36,  1.42s/it]
Continual predict:   1%|          | 6/581 [00:12<24:11,  2.52s/it]
Continual predict:   1%|          | 7/581 [00:15<26:36,  2.78s/it]
Continual predict:   1%|▏         | 8/581 [00:17<24:23,  2.55s/it]
Continual predict:   2%|▏         | 9/581 [00:19<21:51,  2.29s/it]
Continual predict:   2%|▏         | 10/581 [00:20<17:25,  1.83s/it]
Continual predict:   2%|▏         | 11/581 [00:22<19:48,  2.08s/it]
Continual predict:   2%|▏         | 12/581 [00:25<22:53,  2.41s/it]
Continual predict:   2%|▏         | 13/581 [00:27<19:25,  2.05s/it]
Continual predict:   2%|▏         | 14/581 [00:29<20:23,  2.16s/it]
Continual predict:   3%|▎         | 15/581 [00:31<20:34,  2.18s/it]
Continual predict:   3%|▎         | 16/581 [00:34<21:38,  2.30s/it]
Continual predict:   3%|▎         | 17/581 [00:42<36:59,  3.93s/it]
Continual predict:   3%|▎         | 18/581 [00:44<31:47,  3.39s/it]
Continual predict:   3%|▎         | 19/581 [00:46<29:21,  3.13s/it]
Continual predict:   3%|▎         | 20/581 [00:48<24:23,  2.61s/it]
Continual predict:   4%|▎         | 21/581 [00:48<17:30,  1.88s/it]
Continual predict:   4%|▍         | 22/581 [00:48<12:39,  1.36s/it]
Continual predict:   4%|▍         | 23/581 [00:48<09:18,  1.00s/it]
Continual predict:   4%|▍         | 24/581 [00:48<06:56,  1.34it/s]
Continual predict:   4%|▍         | 25/581 [00:48<05:16,  1.76it/s]
Continual predict:   4%|▍         | 26/581 [00:49<04:06,  2.26it/s]
Continual predict:   5%|▍         | 27/581 [00:49<03:16,  2.81it/s]
Continual predict:   5%|▍         | 28/581 [00:49<02:42,  3.41it/s]
Continual predict:   5%|▍         | 29/581 [00:49<02:17,  4.00it/s]
Continual predict:   5%|▌         | 30/581 [00:49<02:03,  4.47it/s]
Continual predict:   5%|▌         | 31/581 [00:49<01:50,  4.96it/s]
Continual predict:   6%|▌         | 32/581 [00:50<01:41,  5.39it/s]
Continual predict:   6%|▌         | 33/581 [00:50<01:38,  5.58it/s]
Continual predict:   6%|▌         | 34/581 [00:50<01:35,  5.74it/s]
Continual predict:   6%|▌         | 35/581 [00:50<01:33,  5.85it/s]
Continual predict:   6%|▌         | 36/581 [00:50<01:29,  6.10it/s]
Continual predict:   6%|▋         | 37/581 [00:50<01:29,  6.11it/s]
Continual predict:   7%|▋         | 38/581 [00:50<01:26,  6.27it/s]
Continual predict:   7%|▋         | 39/581 [00:51<01:24,  6.40it/s]
Continual predict:   7%|▋         | 40/581 [00:51<01:23,  6.50it/s]
Continual predict:   7%|▋         | 41/581 [00:51<01:24,  6.40it/s]
Continual predict:   7%|▋         | 42/581 [00:51<01:23,  6.48it/s]
Continual predict:   7%|▋         | 43/581 [00:51<01:24,  6.39it/s]
Continual predict:   8%|▊         | 44/581 [00:51<01:24,  6.33it/s]
Continual predict:   8%|▊         | 45/581 [00:52<01:23,  6.43it/s]
Continual predict:   8%|▊         | 46/581 [00:52<01:22,  6.49it/s]
Continual predict:   8%|▊         | 47/581 [00:52<01:22,  6.49it/s]
Continual predict:   8%|▊         | 48/581 [00:52<01:23,  6.36it/s]
Continual predict:   8%|▊         | 49/581 [00:52<01:24,  6.32it/s]
Continual predict:   9%|▊         | 50/581 [00:52<01:22,  6.42it/s]
Continual predict:   9%|▉         | 51/581 [00:52<01:22,  6.44it/s]
Continual predict:   9%|▉         | 52/581 [00:53<01:21,  6.51it/s]
Continual predict:   9%|▉         | 53/581 [00:53<01:20,  6.56it/s]
Continual predict:   9%|▉         | 54/581 [00:53<01:21,  6.44it/s]
Continual predict:   9%|▉         | 55/581 [00:53<01:20,  6.52it/s]
Continual predict:  10%|▉         | 56/581 [00:53<01:22,  6.39it/s]
Continual predict:  10%|▉         | 57/581 [00:53<01:20,  6.47it/s]
Continual predict:  10%|▉         | 58/581 [00:54<01:20,  6.53it/s]
Continual predict:  10%|█         | 59/581 [00:54<01:19,  6.57it/s]
Continual predict:  10%|█         | 60/581 [00:54<01:21,  6.42it/s]
Continual predict:  10%|█         | 61/581 [00:54<01:19,  6.51it/s]
Continual predict:  11%|█         | 62/581 [00:54<01:19,  6.54it/s]
Continual predict:  11%|█         | 63/581 [00:54<01:20,  6.44it/s]
Continual predict:  11%|█         | 64/581 [00:54<01:19,  6.51it/s]
Continual predict:  11%|█         | 65/581 [00:55<01:18,  6.56it/s]
Continual predict:  11%|█▏        | 66/581 [00:55<01:18,  6.59it/s]
Continual predict:  12%|█▏        | 67/581 [00:55<01:17,  6.64it/s]
Continual predict:  12%|█▏        | 68/581 [00:55<01:16,  6.66it/s]
Continual predict:  12%|█▏        | 69/581 [00:55<01:16,  6.66it/s]
Continual predict:  12%|█▏        | 70/581 [00:55<01:16,  6.69it/s]
Continual predict:  12%|█▏        | 71/581 [00:56<01:16,  6.69it/s]
Continual predict:  12%|█▏        | 72/581 [00:56<01:18,  6.52it/s]
Continual predict:  13%|█▎        | 73/581 [00:56<01:17,  6.58it/s]
Continual predict:  13%|█▎        | 74/581 [00:56<01:18,  6.44it/s]
Continual predict:  13%|█▎        | 75/581 [00:56<01:18,  6.45it/s]
Continual predict:  13%|█▎        | 76/581 [00:56<01:17,  6.51it/s]
Continual predict:  13%|█▎        | 77/581 [00:56<01:17,  6.47it/s]
Continual predict:  13%|█▎        | 78/581 [00:57<01:19,  6.36it/s]
Continual predict:  14%|█▎        | 79/581 [00:57<01:19,  6.31it/s]
Continual predict:  14%|█▍        | 80/581 [00:57<01:17,  6.43it/s]
Continual predict:  14%|█▍        | 81/581 [00:57<01:16,  6.50it/s]
Continual predict:  14%|█▍        | 82/581 [00:57<01:16,  6.56it/s]
Continual predict:  14%|█▍        | 83/581 [00:57<01:15,  6.60it/s]
Continual predict:  14%|█▍        | 84/581 [00:58<01:16,  6.46it/s]
Continual predict:  15%|█▍        | 85/581 [00:58<01:16,  6.51it/s]
Continual predict:  15%|█▍        | 86/581 [00:58<01:15,  6.56it/s]
Continual predict:  15%|█▍        | 87/581 [00:58<01:15,  6.57it/s]
Continual predict:  15%|█▌        | 88/581 [00:58<01:14,  6.63it/s]
Continual predict:  15%|█▌        | 89/581 [00:58<01:14,  6.62it/s]
Continual predict:  15%|█▌        | 90/581 [00:58<01:13,  6.66it/s]
Continual predict:  16%|█▌        | 91/581 [00:59<01:13,  6.66it/s]
Continual predict:  16%|█▌        | 92/581 [00:59<01:13,  6.64it/s]
Continual predict:  16%|█▌        | 93/581 [00:59<01:15,  6.49it/s]
Continual predict:  16%|█▌        | 94/581 [00:59<01:16,  6.40it/s]
Continual predict:  16%|█▋        | 95/581 [00:59<01:14,  6.48it/s]
Continual predict:  17%|█▋        | 96/581 [00:59<01:14,  6.53it/s]
Continual predict:  17%|█▋        | 97/581 [01:00<01:13,  6.57it/s]
Continual predict:  17%|█▋        | 98/581 [01:00<01:14,  6.46it/s]
Continual predict:  17%|█▋        | 99/581 [01:00<01:15,  6.37it/s]
Continual predict:  17%|█▋        | 100/581 [01:00<01:16,  6.32it/s]
Continual predict:  17%|█▋        | 101/581 [01:00<01:14,  6.42it/s]
Continual predict:  18%|█▊        | 102/581 [01:00<01:13,  6.51it/s]
Continual predict:  18%|█▊        | 103/581 [01:00<01:12,  6.56it/s]
Continual predict:  18%|█▊        | 104/581 [01:01<01:12,  6.59it/s]
Continual predict:  18%|█▊        | 105/581 [01:01<01:13,  6.46it/s]
Continual predict:  18%|█▊        | 106/581 [01:01<01:12,  6.54it/s]
Continual predict:  18%|█▊        | 107/581 [01:01<01:11,  6.60it/s]
Continual predict:  19%|█▊        | 108/581 [01:01<01:11,  6.62it/s]
Continual predict:  19%|█▉        | 109/581 [01:01<01:11,  6.64it/s]
Continual predict:  19%|█▉        | 110/581 [01:02<01:10,  6.64it/s]
Continual predict:  19%|█▉        | 111/581 [01:02<01:10,  6.65it/s]
Continual predict:  19%|█▉        | 112/581 [01:02<01:10,  6.66it/s]
Continual predict:  19%|█▉        | 113/581 [01:02<01:10,  6.66it/s]
Continual predict:  20%|█▉        | 114/581 [01:02<01:10,  6.67it/s]
Continual predict:  20%|█▉        | 115/581 [01:02<01:11,  6.52it/s]
Continual predict:  20%|█▉        | 116/581 [01:02<01:10,  6.55it/s]
Continual predict:  20%|██        | 117/581 [01:03<01:12,  6.44it/s]
Continual predict:  20%|██        | 118/581 [01:03<01:12,  6.35it/s]
Continual predict:  20%|██        | 119/581 [01:03<01:11,  6.44it/s]
Continual predict:  21%|██        | 120/581 [01:03<01:10,  6.51it/s]
Continual predict:  21%|██        | 121/581 [01:08<11:22,  1.48s/it]
Continual predict:  21%|██        | 122/581 [01:12<17:43,  2.32s/it]
Continual predict:  21%|██        | 123/581 [01:17<24:38,  3.23s/it]
Continual predict:  21%|██▏       | 124/581 [01:20<22:41,  2.98s/it]
Continual predict:  22%|██▏       | 125/581 [01:24<26:53,  3.54s/it]
Continual predict:  22%|██▏       | 126/581 [01:29<29:45,  3.92s/it]
Continual predict:  22%|██▏       | 127/581 [01:31<24:17,  3.21s/it]
Continual predict:  22%|██▏       | 128/581 [01:34<23:17,  3.09s/it]
Continual predict:  22%|██▏       | 129/581 [01:37<24:01,  3.19s/it]
Continual predict:  22%|██▏       | 130/581 [01:40<23:44,  3.16s/it]
Continual predict:  23%|██▎       | 131/581 [01:42<21:41,  2.89s/it]
Continual predict:  23%|██▎       | 132/581 [01:45<21:14,  2.84s/it]
Continual predict:  23%|██▎       | 133/581 [01:48<22:01,  2.95s/it]
Continual predict:  23%|██▎       | 134/581 [01:54<29:03,  3.90s/it]
Continual predict:  23%|██▎       | 135/581 [01:59<31:08,  4.19s/it]
Continual predict:  23%|██▎       | 136/581 [02:03<29:06,  3.93s/it]
Continual predict:  24%|██▎       | 137/581 [02:05<25:23,  3.43s/it]
Continual predict:  24%|██▍       | 138/581 [02:06<19:25,  2.63s/it]
Continual predict:  24%|██▍       | 139/581 [02:07<16:12,  2.20s/it]
Continual predict:  24%|██▍       | 140/581 [02:10<17:20,  2.36s/it]
Continual predict:  24%|██▍       | 141/581 [02:16<26:46,  3.65s/it]
Continual predict:  24%|██▍       | 142/581 [02:21<28:49,  3.94s/it]
Continual predict:  25%|██▍       | 143/581 [02:24<27:24,  3.75s/it]
Continual predict:  25%|██▍       | 144/581 [02:34<39:57,  5.49s/it]
Continual predict:  25%|██▍       | 145/581 [02:38<37:41,  5.19s/it]
Continual predict:  25%|██▌       | 146/581 [02:40<30:19,  4.18s/it]
Continual predict:  25%|██▌       | 147/581 [02:45<30:49,  4.26s/it]
Continual predict:  25%|██▌       | 148/581 [02:47<26:18,  3.64s/it]
Continual predict:  26%|██▌       | 149/581 [02:50<25:18,  3.52s/it]
Continual predict:  26%|██▌       | 150/581 [02:52<21:51,  3.04s/it]
Continual predict:  26%|██▌       | 151/581 [02:55<21:54,  3.06s/it]
Continual predict:  26%|██▌       | 152/581 [03:01<27:35,  3.86s/it]
Continual predict:  26%|██▋       | 153/581 [03:03<24:39,  3.46s/it]
Continual predict:  27%|██▋       | 154/581 [03:08<27:36,  3.88s/it]
Continual predict:  27%|██▋       | 155/581 [03:12<27:01,  3.81s/it]
Continual predict:  27%|██▋       | 156/581 [03:15<26:13,  3.70s/it]
Continual predict:  27%|██▋       | 157/581 [03:19<26:39,  3.77s/it]
Continual predict:  27%|██▋       | 158/581 [03:20<20:40,  2.93s/it]
Continual predict:  27%|██▋       | 159/581 [03:27<29:13,  4.16s/it]
Continual predict:  28%|██▊       | 160/581 [03:29<25:16,  3.60s/it]
Continual predict:  28%|██▊       | 161/581 [03:32<22:58,  3.28s/it]
Continual predict:  28%|██▊       | 162/581 [03:37<26:14,  3.76s/it]
Continual predict:  28%|██▊       | 163/581 [03:39<22:42,  3.26s/it]
Continual predict:  28%|██▊       | 164/581 [03:41<20:39,  2.97s/it]
Continual predict:  28%|██▊       | 165/581 [03:44<21:04,  3.04s/it]
Continual predict:  29%|██▊       | 166/581 [03:47<19:20,  2.80s/it]
Continual predict:  29%|██▊       | 167/581 [03:48<16:14,  2.35s/it]
Continual predict:  29%|██▉       | 168/581 [03:50<16:05,  2.34s/it]
Continual predict:  29%|██▉       | 169/581 [03:52<14:50,  2.16s/it]
Continual predict:  29%|██▉       | 170/581 [03:55<16:46,  2.45s/it]
Continual predict:  29%|██▉       | 171/581 [03:57<14:57,  2.19s/it]
Continual predict:  30%|██▉       | 172/581 [03:59<15:48,  2.32s/it]
Continual predict:  30%|██▉       | 173/581 [04:03<19:09,  2.82s/it]
Continual predict:  30%|██▉       | 174/581 [04:05<15:47,  2.33s/it]
Continual predict:  30%|███       | 175/581 [04:08<18:12,  2.69s/it]
Continual predict:  30%|███       | 176/581 [04:11<18:36,  2.76s/it]
Continual predict:  30%|███       | 177/581 [04:14<18:33,  2.76s/it]
Continual predict:  31%|███       | 178/581 [04:15<15:32,  2.31s/it]
Continual predict:  31%|███       | 179/581 [04:16<13:08,  1.96s/it]
Continual predict:  31%|███       | 180/581 [04:17<11:50,  1.77s/it]
Continual predict:  31%|███       | 181/581 [04:19<10:54,  1.64s/it]
Continual predict:  31%|███▏      | 182/581 [04:24<17:05,  2.57s/it]
Continual predict:  31%|███▏      | 183/581 [04:24<12:22,  1.87s/it]
Continual predict:  32%|███▏      | 184/581 [04:25<10:50,  1.64s/it]
Continual predict:  32%|███▏      | 185/581 [04:28<14:19,  2.17s/it]
Continual predict:  32%|███▏      | 186/581 [04:29<11:59,  1.82s/it]
Continual predict:  32%|███▏      | 187/581 [04:33<14:51,  2.26s/it]
Continual predict:  32%|███▏      | 188/581 [04:35<15:46,  2.41s/it]
Continual predict:  33%|███▎      | 189/581 [04:40<20:52,  3.19s/it]
Continual predict:  33%|███▎      | 190/581 [04:44<21:03,  3.23s/it]
Continual predict:  33%|███▎      | 191/581 [04:45<17:17,  2.66s/it]
Continual predict:  33%|███▎      | 192/581 [04:48<18:35,  2.87s/it]
Continual predict:  33%|███▎      | 193/581 [04:53<21:52,  3.38s/it]
Continual predict:  33%|███▎      | 194/581 [04:58<24:38,  3.82s/it]
Continual predict:  34%|███▎      | 195/581 [05:03<27:32,  4.28s/it]
Continual predict:  34%|███▎      | 196/581 [05:06<25:42,  4.01s/it]
Continual predict:  34%|███▍      | 197/581 [05:10<24:33,  3.84s/it]
Continual predict:  34%|███▍      | 198/581 [05:17<31:21,  4.91s/it]
Continual predict:  34%|███▍      | 199/581 [05:18<24:00,  3.77s/it]
Continual predict:  34%|███▍      | 200/581 [05:20<19:33,  3.08s/it]
Continual predict:  35%|███▍      | 201/581 [05:23<18:40,  2.95s/it]
Continual predict:  35%|███▍      | 202/581 [05:26<19:16,  3.05s/it]
Continual predict:  35%|███▍      | 203/581 [05:31<23:14,  3.69s/it]
Continual predict:  35%|███▌      | 204/581 [05:34<22:00,  3.50s/it]
Continual predict:  35%|███▌      | 205/581 [05:41<27:51,  4.44s/it]
Continual predict:  35%|███▌      | 206/581 [05:44<24:38,  3.94s/it]
Continual predict:  36%|███▌      | 207/581 [05:47<24:14,  3.89s/it]
Continual predict:  36%|███▌      | 208/581 [05:53<28:29,  4.58s/it]
Continual predict:  36%|███▌      | 209/581 [05:56<24:56,  4.02s/it]
Continual predict:  36%|███▌      | 210/581 [05:58<20:30,  3.32s/it]
Continual predict:  36%|███▋      | 211/581 [06:01<19:10,  3.11s/it]
Continual predict:  36%|███▋      | 212/581 [06:04<19:13,  3.13s/it]
Continual predict:  37%|███▋      | 213/581 [06:07<20:22,  3.32s/it]
Continual predict:  37%|███▋      | 214/581 [06:11<20:54,  3.42s/it]
Continual predict:  37%|███▋      | 215/581 [06:12<16:00,  2.62s/it]
Continual predict:  37%|███▋      | 216/581 [06:15<16:51,  2.77s/it]
Continual predict:  37%|███▋      | 217/581 [06:17<15:31,  2.56s/it]
Continual predict:  38%|███▊      | 218/581 [06:21<18:02,  2.98s/it]
Continual predict:  38%|███▊      | 219/581 [06:24<18:29,  3.07s/it]
Continual predict:  38%|███▊      | 220/581 [06:28<20:22,  3.39s/it]
Continual predict:  38%|███▊      | 221/581 [06:29<14:36,  2.44s/it]
Continual predict:  38%|███▊      | 222/581 [06:29<10:42,  1.79s/it]
Continual predict:  38%|███▊      | 223/581 [06:29<07:50,  1.32s/it]
Continual predict:  39%|███▊      | 224/581 [06:29<05:51,  1.02it/s]
Continual predict:  39%|███▊      | 225/581 [06:30<04:28,  1.33it/s]
Continual predict:  39%|███▉      | 226/581 [06:30<03:29,  1.69it/s]
Continual predict:  39%|███▉      | 227/581 [06:30<02:54,  2.03it/s]
Continual predict:  39%|███▉      | 228/581 [06:30<02:29,  2.36it/s]
Continual predict:  39%|███▉      | 229/581 [06:31<02:09,  2.73it/s]
Continual predict:  40%|███▉      | 230/581 [06:31<02:14,  2.61it/s]
Continual predict:  40%|███▉      | 231/581 [06:31<01:56,  3.01it/s]
Continual predict:  40%|███▉      | 232/581 [06:31<01:43,  3.37it/s]
Continual predict:  40%|████      | 233/581 [06:32<01:34,  3.69it/s]
Continual predict:  40%|████      | 234/581 [06:32<01:27,  3.95it/s]
Continual predict:  40%|████      | 235/581 [06:32<01:23,  4.16it/s]
Continual predict:  41%|████      | 236/581 [06:32<01:19,  4.32it/s]
Continual predict:  41%|████      | 237/581 [06:32<01:17,  4.44it/s]
Continual predict:  41%|████      | 238/581 [06:33<01:15,  4.53it/s]
Continual predict:  41%|████      | 239/581 [06:33<01:14,  4.60it/s]
Continual predict:  41%|████▏     | 240/581 [06:33<01:28,  3.87it/s]
Continual predict:  41%|████▏     | 241/581 [06:33<01:22,  4.10it/s]
Continual predict:  42%|████▏     | 242/581 [06:34<01:19,  4.28it/s]
Continual predict:  42%|████▏     | 243/581 [06:34<01:16,  4.41it/s]
Continual predict:  42%|████▏     | 244/581 [06:34<01:14,  4.51it/s]
Continual predict:  42%|████▏     | 245/581 [06:34<01:13,  4.57it/s]
Continual predict:  42%|████▏     | 246/581 [06:34<01:12,  4.61it/s]
Continual predict:  43%|████▎     | 247/581 [06:35<01:17,  4.32it/s]
Continual predict:  43%|████▎     | 248/581 [06:35<01:14,  4.44it/s]
Continual predict:  43%|████▎     | 249/581 [06:35<01:18,  4.22it/s]
Continual predict:  43%|████▎     | 250/581 [06:35<01:15,  4.37it/s]
Continual predict:  43%|████▎     | 251/581 [06:36<01:18,  4.18it/s]
Continual predict:  43%|████▎     | 252/581 [06:37<02:58,  1.84it/s]
Continual predict:  44%|████▎     | 253/581 [06:37<02:30,  2.18it/s]
Continual predict:  44%|████▎     | 254/581 [06:37<02:05,  2.60it/s]
Continual predict:  44%|████▍     | 255/581 [06:38<01:48,  3.00it/s]
Continual predict:  44%|████▍     | 256/581 [06:38<01:36,  3.37it/s]
Continual predict:  44%|████▍     | 257/581 [06:39<02:18,  2.34it/s]
Continual predict:  44%|████▍     | 258/581 [06:39<01:57,  2.76it/s]
Continual predict:  45%|████▍     | 259/581 [06:39<01:42,  3.16it/s]
Continual predict:  45%|████▍     | 260/581 [06:39<01:31,  3.50it/s]
Continual predict:  45%|████▍     | 261/581 [06:40<01:58,  2.71it/s]
Continual predict:  45%|████▌     | 262/581 [06:40<01:44,  3.06it/s]
Continual predict:  45%|████▌     | 263/581 [06:40<01:37,  3.25it/s]
Continual predict:  45%|████▌     | 264/581 [06:41<03:05,  1.71it/s]
Continual predict:  46%|████▌     | 265/581 [06:42<03:00,  1.75it/s]
Continual predict:  46%|████▌     | 266/581 [06:43<02:56,  1.78it/s]
Continual predict:  46%|████▌     | 267/581 [06:43<03:00,  1.74it/s]
Continual predict:  46%|████▌     | 268/581 [06:44<03:24,  1.53it/s]
Continual predict:  46%|████▋     | 269/581 [06:45<03:19,  1.56it/s]
Continual predict:  46%|████▋     | 270/581 [06:45<03:18,  1.56it/s]
Continual predict:  47%|████▋     | 271/581 [06:46<03:29,  1.48it/s]
Continual predict:  47%|████▋     | 272/581 [06:47<03:50,  1.34it/s]
Continual predict:  47%|████▋     | 273/581 [06:48<03:40,  1.40it/s]
Continual predict:  47%|████▋     | 274/581 [06:48<03:34,  1.43it/s]
Continual predict:  47%|████▋     | 275/581 [06:49<03:25,  1.49it/s]
Continual predict:  48%|████▊     | 276/581 [06:50<04:00,  1.27it/s]
Continual predict:  48%|████▊     | 277/581 [06:51<04:44,  1.07it/s]
Continual predict:  48%|████▊     | 278/581 [06:52<04:34,  1.11it/s]
Continual predict:  48%|████▊     | 279/581 [06:53<04:15,  1.18it/s]
Continual predict:  48%|████▊     | 280/581 [06:53<03:39,  1.37it/s]
Continual predict:  48%|████▊     | 281/581 [06:54<03:59,  1.25it/s]
Continual predict:  49%|████▊     | 282/581 [06:55<03:59,  1.25it/s]
Continual predict:  49%|████▊     | 283/581 [06:56<03:47,  1.31it/s]
Continual predict:  49%|████▉     | 284/581 [06:56<03:20,  1.48it/s]
Continual predict:  49%|████▉     | 285/581 [06:57<03:30,  1.41it/s]
Continual predict:  49%|████▉     | 286/581 [06:58<03:34,  1.37it/s]
Continual predict:  49%|████▉     | 287/581 [06:58<03:37,  1.35it/s]
Continual predict:  50%|████▉     | 288/581 [06:59<03:28,  1.41it/s]
Continual predict:  50%|████▉     | 289/581 [07:00<03:48,  1.28it/s]
Continual predict:  50%|████▉     | 290/581 [07:01<03:45,  1.29it/s]
Continual predict:  50%|█████     | 291/581 [07:01<03:30,  1.38it/s]
Continual predict:  50%|█████     | 292/581 [07:02<02:54,  1.66it/s]
Continual predict:  50%|█████     | 293/581 [07:03<03:13,  1.49it/s]
Continual predict:  51%|█████     | 294/581 [07:03<03:01,  1.58it/s]
Continual predict:  51%|█████     | 295/581 [07:03<02:45,  1.73it/s]
Continual predict:  51%|█████     | 296/581 [07:04<03:18,  1.43it/s]
Continual predict:  51%|█████     | 297/581 [07:05<03:03,  1.55it/s]
Continual predict:  51%|█████▏    | 298/581 [07:05<02:46,  1.70it/s]
Continual predict:  51%|█████▏    | 299/581 [07:06<03:17,  1.42it/s]
Continual predict:  52%|█████▏    | 300/581 [07:07<02:45,  1.70it/s]
Continual predict:  52%|█████▏    | 301/581 [07:08<03:12,  1.45it/s]
Continual predict:  52%|█████▏    | 302/581 [07:08<02:49,  1.64it/s]
Continual predict:  52%|█████▏    | 303/581 [07:10<04:06,  1.13it/s]
Continual predict:  52%|█████▏    | 304/581 [07:10<03:18,  1.39it/s]
Continual predict:  52%|█████▏    | 305/581 [07:11<03:33,  1.29it/s]
Continual predict:  53%|█████▎    | 306/581 [07:12<03:36,  1.27it/s]
Continual predict:  53%|█████▎    | 307/581 [07:12<03:36,  1.27it/s]
Continual predict:  53%|█████▎    | 308/581 [07:13<03:06,  1.46it/s]
Continual predict:  53%|█████▎    | 309/581 [07:13<02:47,  1.63it/s]
Continual predict:  53%|█████▎    | 310/581 [07:14<02:46,  1.62it/s]
Continual predict:  54%|█████▎    | 311/581 [07:15<02:39,  1.69it/s]
Continual predict:  54%|█████▎    | 312/581 [07:15<02:49,  1.59it/s]
Continual predict:  54%|█████▍    | 313/581 [07:16<03:08,  1.42it/s]
Continual predict:  54%|█████▍    | 314/581 [07:17<02:55,  1.52it/s]
Continual predict:  54%|█████▍    | 315/581 [07:18<03:14,  1.36it/s]
Continual predict:  54%|█████▍    | 316/581 [07:18<03:19,  1.33it/s]
Continual predict:  55%|█████▍    | 317/581 [07:19<02:45,  1.60it/s]
Continual predict:  55%|█████▍    | 318/581 [07:19<02:31,  1.74it/s]
Continual predict:  55%|█████▍    | 319/581 [07:21<03:52,  1.13it/s]
Continual predict:  55%|█████▌    | 320/581 [07:21<03:37,  1.20it/s]
Continual predict:  55%|█████▌    | 321/581 [07:22<03:26,  1.26it/s]
Continual predict:  55%|█████▌    | 322/581 [07:23<02:57,  1.46it/s]
Continual predict:  56%|█████▌    | 323/581 [07:23<03:01,  1.42it/s]
Continual predict:  56%|█████▌    | 324/581 [07:24<03:15,  1.31it/s]
Continual predict:  56%|█████▌    | 325/581 [07:25<03:15,  1.31it/s]
Continual predict:  56%|█████▌    | 326/581 [07:26<03:05,  1.38it/s]
Continual predict:  56%|█████▋    | 327/581 [07:27<03:34,  1.18it/s]
Continual predict:  56%|█████▋    | 328/581 [07:28<03:50,  1.10it/s]
Continual predict:  57%|█████▋    | 329/581 [07:28<03:20,  1.26it/s]
Continual predict:  57%|█████▋    | 330/581 [07:29<02:44,  1.53it/s]
Continual predict:  57%|█████▋    | 331/581 [07:29<02:39,  1.56it/s]
Continual predict:  57%|█████▋    | 332/581 [07:30<02:31,  1.64it/s]
Continual predict:  57%|█████▋    | 333/581 [07:31<03:16,  1.26it/s]
Continual predict:  57%|█████▋    | 334/581 [07:32<03:24,  1.21it/s]
Continual predict:  58%|█████▊    | 335/581 [07:32<02:54,  1.41it/s]
Continual predict:  58%|█████▊    | 336/581 [07:33<02:34,  1.58it/s]
Continual predict:  58%|█████▊    | 337/581 [07:33<02:30,  1.63it/s]
Continual predict:  58%|█████▊    | 338/581 [07:34<02:31,  1.60it/s]
Continual predict:  58%|█████▊    | 339/581 [07:35<02:24,  1.68it/s]
Continual predict:  59%|█████▊    | 340/581 [07:36<02:58,  1.35it/s]
Continual predict:  59%|█████▊    | 341/581 [07:36<02:59,  1.34it/s]
Continual predict:  59%|█████▉    | 342/581 [07:37<02:37,  1.52it/s]
Continual predict:  59%|█████▉    | 343/581 [07:38<02:47,  1.42it/s]
Continual predict:  59%|█████▉    | 344/581 [07:39<03:05,  1.28it/s]
Continual predict:  59%|█████▉    | 345/581 [07:39<02:41,  1.46it/s]
Continual predict:  60%|█████▉    | 346/581 [07:40<02:49,  1.38it/s]
Continual predict:  60%|█████▉    | 347/581 [07:41<03:40,  1.06it/s]
Continual predict:  60%|█████▉    | 348/581 [07:42<02:57,  1.31it/s]
Continual predict:  60%|██████    | 349/581 [07:42<02:33,  1.51it/s]
Continual predict:  60%|██████    | 350/581 [07:43<02:32,  1.52it/s]
Continual predict:  60%|██████    | 351/581 [07:44<03:03,  1.25it/s]
Continual predict:  61%|██████    | 352/581 [07:45<03:00,  1.27it/s]
Continual predict:  61%|██████    | 353/581 [07:46<03:33,  1.07it/s]
Continual predict:  61%|██████    | 354/581 [07:47<03:54,  1.03s/it]
Continual predict:  61%|██████    | 355/581 [07:48<03:25,  1.10it/s]
Continual predict:  61%|██████▏   | 356/581 [07:49<03:25,  1.10it/s]
Continual predict:  61%|██████▏   | 357/581 [07:50<03:35,  1.04it/s]
Continual predict:  62%|██████▏   | 358/581 [07:51<03:20,  1.11it/s]
Continual predict:  62%|██████▏   | 359/581 [07:51<03:02,  1.22it/s]
Continual predict:  62%|██████▏   | 360/581 [07:53<03:30,  1.05it/s]
Continual predict:  62%|██████▏   | 361/581 [07:54<04:11,  1.14s/it]
Continual predict:  62%|██████▏   | 362/581 [07:55<03:34,  1.02it/s]
Continual predict:  62%|██████▏   | 363/581 [07:55<02:59,  1.21it/s]
Continual predict:  63%|██████▎   | 364/581 [07:55<02:26,  1.48it/s]
Continual predict:  63%|██████▎   | 365/581 [07:56<01:59,  1.81it/s]
Continual predict:  63%|██████▎   | 366/581 [07:56<01:40,  2.13it/s]
Continual predict:  63%|██████▎   | 367/581 [07:56<01:27,  2.44it/s]
Continual predict:  63%|██████▎   | 368/581 [07:57<01:18,  2.72it/s]
Continual predict:  64%|██████▎   | 369/581 [07:57<01:11,  2.95it/s]
Continual predict:  64%|██████▎   | 370/581 [07:57<01:07,  3.14it/s]
Continual predict:  64%|██████▍   | 371/581 [07:57<01:07,  3.12it/s]
Continual predict:  64%|██████▍   | 372/581 [07:58<01:10,  2.96it/s]
Continual predict:  64%|██████▍   | 373/581 [07:58<01:12,  2.85it/s]
Continual predict:  64%|██████▍   | 374/581 [07:59<01:09,  2.96it/s]
Continual predict:  65%|██████▍   | 375/581 [07:59<01:21,  2.54it/s]
Continual predict:  65%|██████▍   | 376/581 [07:59<01:16,  2.68it/s]
Continual predict:  65%|██████▍   | 377/581 [08:00<01:09,  2.92it/s]
Continual predict:  65%|██████▌   | 378/581 [08:00<01:05,  3.12it/s]
Continual predict:  65%|██████▌   | 379/581 [08:00<01:01,  3.27it/s]
Continual predict:  65%|██████▌   | 380/581 [08:00<00:59,  3.39it/s]
Continual predict:  66%|██████▌   | 381/581 [08:01<01:00,  3.29it/s]
Continual predict:  66%|██████▌   | 382/581 [08:01<00:58,  3.40it/s]
Continual predict:  66%|██████▌   | 383/581 [08:01<00:56,  3.48it/s]
Continual predict:  66%|██████▌   | 384/581 [08:02<00:55,  3.54it/s]
Continual predict:  66%|██████▋   | 385/581 [08:02<00:54,  3.58it/s]
Continual predict:  66%|██████▋   | 386/581 [08:02<00:53,  3.62it/s]
Continual predict:  67%|██████▋   | 387/581 [08:02<00:53,  3.64it/s]
Continual predict:  67%|██████▋   | 388/581 [08:03<00:52,  3.65it/s]
Continual predict:  67%|██████▋   | 389/581 [08:03<00:55,  3.45it/s]
Continual predict:  67%|██████▋   | 390/581 [08:03<00:54,  3.52it/s]
Continual predict:  67%|██████▋   | 391/581 [08:04<00:53,  3.57it/s]
Continual predict:  67%|██████▋   | 392/581 [08:04<00:55,  3.41it/s]
Continual predict:  68%|██████▊   | 393/581 [08:04<00:53,  3.49it/s]
Continual predict:  68%|██████▊   | 394/581 [08:04<00:53,  3.53it/s]
Continual predict:  68%|██████▊   | 395/581 [08:05<00:49,  3.79it/s]
Continual predict:  68%|██████▊   | 396/581 [08:05<00:55,  3.35it/s]
Continual predict:  68%|██████▊   | 397/581 [08:05<00:53,  3.45it/s]
Continual predict:  69%|██████▊   | 398/581 [08:06<00:52,  3.52it/s]
Continual predict:  69%|██████▊   | 399/581 [08:06<00:51,  3.56it/s]
Continual predict:  69%|██████▉   | 400/581 [08:06<00:50,  3.60it/s]
Continual predict:  69%|██████▉   | 401/581 [08:06<00:50,  3.53it/s]
Continual predict:  69%|██████▉   | 402/581 [08:07<00:49,  3.58it/s]
Continual predict:  69%|██████▉   | 403/581 [08:07<00:52,  3.41it/s]
Continual predict:  70%|██████▉   | 404/581 [08:07<00:50,  3.49it/s]
Continual predict:  70%|██████▉   | 405/581 [08:08<00:55,  3.18it/s]
Continual predict:  70%|██████▉   | 406/581 [08:08<00:52,  3.32it/s]
Continual predict:  70%|███████   | 407/581 [08:08<00:50,  3.42it/s]
Continual predict:  70%|███████   | 408/581 [08:08<00:49,  3.50it/s]
Continual predict:  70%|███████   | 409/581 [08:09<00:48,  3.55it/s]
Continual predict:  71%|███████   | 410/581 [08:09<00:57,  2.99it/s]
Continual predict:  71%|███████   | 411/581 [08:10<00:59,  2.85it/s]
Continual predict:  71%|███████   | 412/581 [08:10<00:55,  3.06it/s]
Continual predict:  71%|███████   | 413/581 [08:10<00:52,  3.22it/s]
Continual predict:  71%|███████▏  | 414/581 [08:10<00:49,  3.35it/s]
Continual predict:  71%|███████▏  | 415/581 [08:11<00:48,  3.45it/s]
Continual predict:  72%|███████▏  | 416/581 [08:11<00:46,  3.52it/s]
Continual predict:  72%|███████▏  | 417/581 [08:11<00:45,  3.57it/s]
Continual predict:  72%|███████▏  | 418/581 [08:12<00:47,  3.40it/s]
Continual predict:  72%|███████▏  | 419/581 [08:12<00:51,  3.13it/s]
Continual predict:  72%|███████▏  | 420/581 [08:12<00:54,  2.96it/s]
Continual predict:  72%|███████▏  | 421/581 [08:13<00:56,  2.85it/s]
Continual predict:  73%|███████▎  | 422/581 [08:13<00:54,  2.91it/s]
Continual predict:  73%|███████▎  | 423/581 [08:13<00:53,  2.96it/s]
Continual predict:  73%|███████▎  | 424/581 [08:14<00:55,  2.85it/s]
Continual predict:  73%|███████▎  | 425/581 [08:14<00:56,  2.78it/s]
Continual predict:  73%|███████▎  | 426/581 [08:14<00:56,  2.74it/s]
Continual predict:  73%|███████▎  | 427/581 [08:15<00:54,  2.83it/s]
Continual predict:  74%|███████▎  | 428/581 [08:15<00:50,  3.05it/s]
Continual predict:  74%|███████▍  | 429/581 [08:15<00:47,  3.21it/s]
Continual predict:  74%|███████▍  | 430/581 [08:16<00:45,  3.34it/s]
Continual predict:  74%|███████▍  | 431/581 [08:16<00:46,  3.26it/s]
Continual predict:  74%|███████▍  | 432/581 [08:16<00:45,  3.30it/s]
Continual predict:  75%|███████▍  | 433/581 [08:16<00:43,  3.41it/s]
Continual predict:  75%|███████▍  | 434/581 [08:17<00:44,  3.30it/s]
Continual predict:  75%|███████▍  | 435/581 [08:17<00:42,  3.41it/s]
Continual predict:  75%|███████▌  | 436/581 [08:17<00:46,  3.13it/s]
Continual predict:  75%|███████▌  | 437/581 [08:18<00:50,  2.82it/s]
Continual predict:  75%|███████▌  | 438/581 [08:18<00:47,  3.04it/s]
Continual predict:  76%|███████▌  | 439/581 [08:18<00:46,  3.05it/s]
Continual predict:  76%|███████▌  | 440/581 [08:19<00:46,  3.06it/s]
Continual predict:  76%|███████▌  | 441/581 [08:19<00:45,  3.06it/s]
Continual predict:  76%|███████▌  | 442/581 [08:19<00:43,  3.23it/s]
Continual predict:  76%|███████▌  | 443/581 [08:20<00:41,  3.31it/s]
Continual predict:  76%|███████▋  | 444/581 [08:20<00:40,  3.37it/s]
Continual predict:  77%|███████▋  | 445/581 [08:20<00:39,  3.46it/s]
Continual predict:  77%|███████▋  | 446/581 [08:21<00:40,  3.33it/s]
Continual predict:  77%|███████▋  | 447/581 [08:21<00:41,  3.25it/s]
Continual predict:  77%|███████▋  | 448/581 [08:21<00:41,  3.20it/s]
Continual predict:  77%|███████▋  | 449/581 [08:22<00:43,  3.00it/s]
Continual predict:  77%|███████▋  | 450/581 [08:22<00:41,  3.18it/s]
Continual predict:  78%|███████▊  | 451/581 [08:22<00:39,  3.32it/s]
Continual predict:  78%|███████▊  | 452/581 [08:22<00:37,  3.42it/s]
Continual predict:  78%|███████▊  | 453/581 [08:23<00:37,  3.42it/s]
Continual predict:  78%|███████▊  | 454/581 [08:23<00:36,  3.49it/s]
Continual predict:  78%|███████▊  | 455/581 [08:23<00:35,  3.55it/s]
Continual predict:  78%|███████▊  | 456/581 [08:24<00:36,  3.39it/s]
Continual predict:  79%|███████▊  | 457/581 [08:24<00:37,  3.29it/s]
Continual predict:  79%|███████▉  | 458/581 [08:24<00:36,  3.40it/s]
Continual predict:  79%|███████▉  | 459/581 [08:24<00:35,  3.48it/s]
Continual predict:  79%|███████▉  | 460/581 [08:25<00:34,  3.54it/s]
Continual predict:  79%|███████▉  | 461/581 [08:25<00:33,  3.58it/s]
Continual predict:  80%|███████▉  | 462/581 [08:25<00:32,  3.61it/s]
Continual predict:  80%|███████▉  | 463/581 [08:26<00:32,  3.63it/s]
Continual predict:  80%|███████▉  | 464/581 [08:26<00:34,  3.38it/s]
Continual predict:  80%|████████  | 465/581 [08:26<00:30,  3.76it/s]
Continual predict:  80%|████████  | 466/581 [08:26<00:31,  3.65it/s]
Continual predict:  80%|████████  | 467/581 [08:27<00:30,  3.77it/s]
Continual predict:  81%|████████  | 468/581 [08:27<00:30,  3.66it/s]
Continual predict:  81%|████████  | 469/581 [08:27<00:31,  3.59it/s]
Continual predict:  81%|████████  | 470/581 [08:27<00:31,  3.53it/s]
Continual predict:  81%|████████  | 471/581 [08:28<00:31,  3.49it/s]
Continual predict:  81%|████████  | 472/581 [08:29<00:49,  2.21it/s]
Continual predict:  81%|████████▏ | 473/581 [08:29<00:50,  2.15it/s]
Continual predict:  82%|████████▏ | 474/581 [08:29<00:41,  2.61it/s]
Continual predict:  82%|████████▏ | 475/581 [08:30<00:37,  2.80it/s]
Continual predict:  82%|████████▏ | 476/581 [08:30<00:37,  2.84it/s]
Continual predict:  82%|████████▏ | 477/581 [08:30<00:35,  2.93it/s]
Continual predict:  82%|████████▏ | 478/581 [08:31<00:35,  2.86it/s]
Continual predict:  82%|████████▏ | 479/581 [08:31<00:46,  2.17it/s]
Continual predict:  83%|████████▎ | 480/581 [08:32<00:41,  2.42it/s]
Continual predict:  83%|████████▎ | 481/581 [08:32<00:39,  2.54it/s]
Continual predict:  83%|████████▎ | 482/581 [08:32<00:41,  2.37it/s]
Continual predict:  83%|████████▎ | 483/581 [08:33<00:53,  1.83it/s]
Continual predict:  83%|████████▎ | 484/581 [08:34<01:07,  1.44it/s]
Continual predict:  83%|████████▎ | 485/581 [08:35<01:18,  1.22it/s]
Continual predict:  84%|████████▎ | 486/581 [08:36<01:24,  1.13it/s]
Continual predict:  84%|████████▍ | 487/581 [08:38<01:39,  1.06s/it]
Continual predict:  84%|████████▍ | 488/581 [08:38<01:20,  1.15it/s]
Continual predict:  84%|████████▍ | 489/581 [08:39<01:15,  1.23it/s]
Continual predict:  84%|████████▍ | 490/581 [08:40<01:12,  1.25it/s]
Continual predict:  85%|████████▍ | 491/581 [08:41<01:11,  1.25it/s]
Continual predict:  85%|████████▍ | 492/581 [08:42<01:14,  1.20it/s]
Continual predict:  85%|████████▍ | 493/581 [08:42<01:10,  1.25it/s]
Continual predict:  85%|████████▌ | 494/581 [08:43<01:11,  1.21it/s]
Continual predict:  85%|████████▌ | 495/581 [08:45<01:29,  1.04s/it]
Continual predict:  85%|████████▌ | 496/581 [08:45<01:22,  1.03it/s]
Continual predict:  86%|████████▌ | 497/581 [08:46<01:10,  1.19it/s]
Continual predict:  86%|████████▌ | 498/581 [08:47<01:09,  1.19it/s]
Continual predict:  86%|████████▌ | 499/581 [08:48<01:08,  1.20it/s]
Continual predict:  86%|████████▌ | 500/581 [08:48<00:58,  1.39it/s]
Continual predict:  86%|████████▌ | 501/581 [08:49<00:53,  1.51it/s]
Continual predict:  86%|████████▋ | 502/581 [08:49<00:55,  1.43it/s]
Continual predict:  87%|████████▋ | 503/581 [08:50<00:56,  1.38it/s]
Continual predict:  87%|████████▋ | 504/581 [08:51<00:51,  1.49it/s]
Continual predict:  87%|████████▋ | 505/581 [08:52<01:01,  1.24it/s]
Continual predict:  87%|████████▋ | 506/581 [08:53<01:02,  1.21it/s]
Continual predict:  87%|████████▋ | 507/581 [08:54<01:03,  1.17it/s]
Continual predict:  87%|████████▋ | 508/581 [08:54<00:58,  1.24it/s]
Continual predict:  88%|████████▊ | 509/581 [08:55<00:58,  1.23it/s]
Continual predict:  88%|████████▊ | 510/581 [08:57<01:17,  1.10s/it]
Continual predict:  88%|████████▊ | 511/581 [08:58<01:15,  1.08s/it]
Continual predict:  88%|████████▊ | 512/581 [08:59<01:05,  1.06it/s]
Continual predict:  88%|████████▊ | 513/581 [09:00<01:17,  1.14s/it]
Continual predict:  88%|████████▊ | 514/581 [09:01<01:12,  1.08s/it]
Continual predict:  89%|████████▊ | 515/581 [09:02<01:04,  1.03it/s]
Continual predict:  89%|████████▉ | 516/581 [09:03<00:57,  1.12it/s]
Continual predict:  89%|████████▉ | 517/581 [09:04<00:58,  1.10it/s]
Continual predict:  89%|████████▉ | 518/581 [09:04<00:57,  1.10it/s]
Continual predict:  89%|████████▉ | 519/581 [09:06<01:08,  1.10s/it]
Continual predict:  90%|████████▉ | 520/581 [09:07<00:58,  1.04it/s]
Continual predict:  90%|████████▉ | 521/581 [09:08<00:58,  1.03it/s]
Continual predict:  90%|████████▉ | 522/581 [09:08<00:51,  1.15it/s]
Continual predict:  90%|█████████ | 523/581 [09:09<00:49,  1.18it/s]
Continual predict:  90%|█████████ | 524/581 [09:10<00:44,  1.27it/s]
Continual predict:  90%|█████████ | 525/581 [09:11<00:44,  1.25it/s]
Continual predict:  91%|█████████ | 526/581 [09:11<00:45,  1.20it/s]
Continual predict:  91%|█████████ | 527/581 [09:12<00:44,  1.22it/s]
Continual predict:  91%|█████████ | 528/581 [09:14<00:52,  1.00it/s]
Continual predict:  91%|█████████ | 529/581 [09:14<00:47,  1.11it/s]
Continual predict:  91%|█████████ | 530/581 [09:16<00:49,  1.02it/s]
Continual predict:  91%|█████████▏| 531/581 [09:17<00:52,  1.05s/it]
Continual predict:  92%|█████████▏| 532/581 [09:18<00:57,  1.17s/it]
Continual predict:  92%|█████████▏| 533/581 [09:19<00:45,  1.05it/s]
Continual predict:  92%|█████████▏| 534/581 [09:19<00:42,  1.11it/s]
Continual predict:  92%|█████████▏| 535/581 [09:20<00:41,  1.11it/s]
Continual predict:  92%|█████████▏| 536/581 [09:21<00:32,  1.36it/s]
Continual predict:  92%|█████████▏| 537/581 [09:21<00:33,  1.31it/s]
Continual predict:  93%|█████████▎| 538/581 [09:22<00:32,  1.31it/s]
Continual predict:  93%|█████████▎| 539/581 [09:23<00:34,  1.21it/s]
Continual predict:  93%|█████████▎| 540/581 [09:24<00:38,  1.06it/s]
Continual predict:  93%|█████████▎| 541/581 [09:25<00:33,  1.19it/s]
Continual predict:  93%|█████████▎| 542/581 [09:26<00:29,  1.34it/s]
Continual predict:  93%|█████████▎| 543/581 [09:26<00:28,  1.32it/s]
Continual predict:  94%|█████████▎| 544/581 [09:27<00:32,  1.15it/s]
Continual predict:  94%|█████████▍| 545/581 [09:29<00:39,  1.09s/it]
Continual predict:  94%|█████████▍| 546/581 [09:30<00:34,  1.01it/s]
Continual predict:  94%|█████████▍| 547/581 [09:31<00:32,  1.03it/s]
Continual predict:  94%|█████████▍| 548/581 [09:32<00:31,  1.06it/s]
Continual predict:  94%|█████████▍| 549/581 [09:33<00:30,  1.05it/s]
Continual predict:  95%|█████████▍| 550/581 [09:34<00:30,  1.02it/s]
Continual predict:  95%|█████████▍| 551/581 [09:34<00:24,  1.21it/s]
Continual predict:  95%|█████████▌| 552/581 [09:35<00:22,  1.31it/s]
Continual predict:  95%|█████████▌| 553/581 [09:36<00:26,  1.06it/s]
Continual predict:  95%|█████████▌| 554/581 [09:37<00:27,  1.00s/it]
Continual predict:  96%|█████████▌| 555/581 [09:38<00:26,  1.02s/it]
Continual predict:  96%|█████████▌| 556/581 [09:39<00:22,  1.11it/s]
Continual predict:  96%|█████████▌| 557/581 [09:40<00:25,  1.08s/it]
Continual predict:  96%|█████████▌| 558/581 [09:41<00:20,  1.10it/s]
Continual predict:  96%|█████████▌| 559/581 [09:42<00:21,  1.02it/s]
Continual predict:  96%|█████████▋| 560/581 [09:43<00:21,  1.02s/it]
Continual predict:  97%|█████████▋| 561/581 [09:45<00:22,  1.12s/it]
Continual predict:  97%|█████████▋| 562/581 [09:46<00:22,  1.18s/it]
Continual predict:  97%|█████████▋| 563/581 [09:47<00:22,  1.25s/it]
Continual predict:  97%|█████████▋| 564/581 [09:49<00:21,  1.27s/it]
Continual predict:  97%|█████████▋| 565/581 [09:50<00:20,  1.26s/it]
Continual predict:  97%|█████████▋| 566/581 [09:51<00:16,  1.11s/it]
Continual predict:  98%|█████████▊| 567/581 [09:52<00:15,  1.13s/it]
Continual predict:  98%|█████████▊| 568/581 [09:53<00:13,  1.02s/it]
Continual predict:  98%|█████████▊| 569/581 [09:54<00:12,  1.01s/it]
Continual predict:  98%|█████████▊| 570/581 [09:54<00:09,  1.20it/s]
Continual predict:  98%|█████████▊| 571/581 [09:55<00:08,  1.18it/s]
Continual predict:  98%|█████████▊| 572/581 [09:55<00:06,  1.33it/s]
Continual predict:  99%|█████████▊| 573/581 [09:56<00:06,  1.15it/s]
Continual predict:  99%|█████████▉| 574/581 [09:57<00:05,  1.26it/s]
Continual predict:  99%|█████████▉| 575/581 [09:58<00:05,  1.16it/s]
Continual predict:  99%|█████████▉| 576/581 [09:59<00:03,  1.30it/s]
Continual predict:  99%|█████████▉| 577/581 [10:00<00:03,  1.25it/s]
Continual predict:  99%|█████████▉| 578/581 [10:00<00:02,  1.42it/s]
Continual predict: 100%|█████████▉| 579/581 [10:01<00:01,  1.31it/s]
Continual predict: 100%|█████████▉| 580/581 [10:02<00:00,  1.08it/s]
Continual predict: 100%|██████████| 581/581 [10:03<00:00,  1.20it/s]
Continual predict: 100%|██████████| 581/581 [10:03<00:00,  1.04s/it]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=20]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=18]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (581/581)  label_acc=100.0% (581/581)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                               =       15.0
+  predict_eval_rougeL                                                 =    60.3918
+  predict_eval_rougeL_for_CL                                          =    60.3918
+  predict_eval_rougeL_for_task002_quoref_answer_generation            =     76.381
+  predict_eval_rougeL_for_task1290_xsum_summarization                 =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction      =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                     =    41.3728
+  predict_eval_rougeL_for_task1729_personachat_generate_next          =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                  =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification        =       91.0
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation =    15.2912
+  predict_exact_match                                                 =     48.537
+  predict_exact_match_for_CL                                          =     48.537
+  predict_exact_match_for_task002_quoref_answer_generation            =       70.0
+  predict_exact_match_for_task1290_xsum_summarization                 =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction      =       98.0
+  predict_exact_match_for_task1572_samsum_summary                     =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next          =        0.0
+  predict_exact_match_for_task181_outcome_extraction                  =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification        =       91.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation =        0.0
+  predict_gen_len                                                     =     9.1532
+  predict_global_step                                                 =        480
+  predict_rouge1                                                      =    62.0913
+  predict_rouge1_for_CL                                               =    62.0913
+  predict_rouge1_for_task002_quoref_answer_generation                 =     76.381
+  predict_rouge1_for_task1290_xsum_summarization                      =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction           =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                          =    50.3879
+  predict_rouge1_for_task1729_personachat_generate_next               =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                       =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification             =       91.0
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation      =    15.2912
+  predict_runtime                                                     = 0:10:03.38
+  predict_samples                                                     =        581
+  predict_samples_per_second                                          =      0.963
+  predict_steps_per_second                                            =      0.963
+[2026-04-30 00:49:55,798] [INFO] [launch.py:367:main] Process 1106737 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/all_results.json b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/all_results.json
new file mode 100644
index 0000000000000000000000000000000000000000..624d0dc895adea7e67a13cfafd048776ca4b1b4c
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/all_results.json
@@ -0,0 +1,48 @@
+{
+  "predict_exact_match": 50.8076,
+  "predict_rouge1": 62.3716,
+  "predict_eval_rougeL": 60.9216,
+  "predict_gen_len": 7.9559,
+  "predict_runtime": 626.9854890499264,
+  "predict_samples": 681,
+  "predict_samples_per_second": 1.086,
+  "predict_steps_per_second": 1.086,
+  "epoch": 15.0,
+  "predict_global_step": 465,
+  "predict_exact_match_for_CL": 50.8076,
+  "predict_rouge1_for_CL": 62.3716,
+  "predict_eval_rougeL_for_CL": 60.9216,
+  "predict_exact_match_for_task1572_samsum_summary": 0.0,
+  "predict_rouge1_for_task1572_samsum_summary": 50.3879,
+  "predict_eval_rougeL_for_task1572_samsum_summary": 41.3728,
+  "predict_exact_match_for_task363_sst2_polarity_classification": 91.0,
+  "predict_rouge1_for_task363_sst2_polarity_classification": 91.0,
+  "predict_eval_rougeL_for_task363_sst2_polarity_classification": 91.0,
+  "predict_exact_match_for_task1290_xsum_summarization": 0.0,
+  "predict_rouge1_for_task1290_xsum_summarization": 36.75,
+  "predict_eval_rougeL_for_task1290_xsum_summarization": 29.5925,
+  "predict_exact_match_for_task181_outcome_extraction": 53.4884,
+  "predict_rouge1_for_task181_outcome_extraction": 59.354,
+  "predict_eval_rougeL_for_task181_outcome_extraction": 59.354,
+  "predict_exact_match_for_task002_quoref_answer_generation": 70.0,
+  "predict_rouge1_for_task002_quoref_answer_generation": 76.381,
+  "predict_eval_rougeL_for_task002_quoref_answer_generation": 76.381,
+  "predict_exact_match_for_task1510_evalution_relation_extraction": 98.0,
+  "predict_rouge1_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_eval_rougeL_for_task1510_evalution_relation_extraction": 99.1333,
+  "predict_exact_match_for_task639_multi_woz_user_utterance_generation": 0.0,
+  "predict_rouge1_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation": 15.2912,
+  "predict_exact_match_for_task1729_personachat_generate_next": 0.0,
+  "predict_rouge1_for_task1729_personachat_generate_next": 19.1341,
+  "predict_eval_rougeL_for_task1729_personachat_generate_next": 18.2202,
+  "predict_exact_match_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_rouge1_for_task073_commonsenseqa_answer_generation": 64.0,
+  "predict_eval_rougeL_for_task073_commonsenseqa_answer_generation": 64.0,
+  "train_runtime": 1991.7116,
+  "train_samples_per_second": 7.343,
+  "train_steps_per_second": 0.233,
+  "total_flos": 9.421432838265242e+16,
+  "train_loss": 0.5980332377136395,
+  "train_samples": 975
+}
\ No newline at end of file
diff --git a/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_A.pt b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_A.pt
new file mode 100644
index 0000000000000000000000000000000000000000..55d3a0eb217eb8163f938c0588738a9e53d85a1b
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_A.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e681477893eb45664ecb1b66af130aa13f0535801cac81b5e425cb541107fc0b
+size 2118413
diff --git a/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_B.pt b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_B.pt
new file mode 100644
index 0000000000000000000000000000000000000000..165b242d6871b716fe80f284b10ead3ea92405c4
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/lora_weights_B.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1ece74ba9168142ecdcd600a24df61b4c33e032969ada85a071c894e3d94e65
+size 2118349
diff --git a/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/srt_signatures.npz b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/srt_signatures.npz
new file mode 100644
index 0000000000000000000000000000000000000000..b5d6d652c8dba2f195beb91bad71b7f164992736
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/srt_signatures.npz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c5236c538cb5f7d69b380867c5693f1313a2588cdbd581da5fa94689d3441ee
+size 1428116234
diff --git a/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/train.log b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/train.log
new file mode 100644
index 0000000000000000000000000000000000000000..dc383209fd56b64fc9c5a2c58f963a7046730b46
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/train.log
@@ -0,0 +1,455 @@
+[2026-04-30 00:50:01,752] [WARNING] [runner.py:232:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2026-04-30 00:50:01,753] [INFO] [runner.py:630:main] cmd = /home/dev-52/miniconda3/envs/myenv/bin/python3.12 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29508 --enable_each_rank_log=None --log_level=info /home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py --model_name_or_path /home/dev-52/minhnguyen/Continual/Llama --data_dir /home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark --task_config_dir /home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task073_commonsenseqa_answer_generation --output_dir /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation --cur_task_id 8 --task_order task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification --lora_r 4 --lora_alpha 32 --lora_dropout 0.0 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 32 --learning_rate 5e-5 --num_train_epochs 15 --max_source_length 1024 --max_target_length 50 --max_new_tokens 50 --srt_shrinkage ridge --srt_max_emb_samples 2000 --use_srt_router --gradient_checkpointing --bf16 --deepspeed /home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config --logging_steps 10 --seed 42 --previous_lora_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights --srt_load_path /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[2026-04-30 00:50:05,009] [INFO] [launch.py:162:main] WORLD INFO DICT: {'localhost': [0]}
+[2026-04-30 00:50:05,009] [INFO] [launch.py:168:main] nnodes=1, num_local_procs=1, node_rank=0
+[2026-04-30 00:50:05,009] [INFO] [launch.py:179:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
+[2026-04-30 00:50:05,009] [INFO] [launch.py:180:main] dist_world_size=1
+[2026-04-30 00:50:05,009] [INFO] [launch.py:184:main] Setting CUDA_VISIBLE_DEVICES=0
+[2026-04-30 00:50:05,009] [INFO] [launch.py:272:main] process 1238647 spawned with command: ['/home/dev-52/miniconda3/envs/myenv/bin/python3.12', '-u', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/run_llama_gainlora_cl.py', '--local_rank=0', '--model_name_or_path', '/home/dev-52/minhnguyen/Continual/Llama', '--data_dir', '/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark', '--task_config_dir', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/gen_script_superni_order1_llama_configs/task073_commonsenseqa_answer_generation', '--output_dir', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation', '--cur_task_id', '8', '--task_order', 'task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification', '--lora_r', '4', '--lora_alpha', '32', '--lora_dropout', '0.0', '--per_device_train_batch_size', '1', '--per_device_eval_batch_size', '1', '--gradient_accumulation_steps', '32', '--learning_rate', '5e-5', '--num_train_epochs', '15', '--max_source_length', '1024', '--max_target_length', '50', '--max_new_tokens', '50', '--srt_shrinkage', 'ridge', '--srt_max_emb_samples', '2000', '--use_srt_router', '--gradient_checkpointing', '--bf16', '--deepspeed', '/home/dev-52/minhnguyen/Continual/new_gainlora/configs/ds_configs/stage2_cpu_offload.config', '--logging_steps', '10', '--seed', '42', '--previous_lora_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights', '--srt_load_path', '/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights']
+
+============================================================
+Task 8: task073_commonsenseqa_answer_generation
+Task order (15 tasks): ['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation', 'task1590_diplomacy_text_generation', 'task748_glucose_reverse_cause_event_detection', 'task511_reddit_tifu_long_text_summarization', 'task591_sciq_answer_generation', 'task1687_sentiment140_classification', 'task875_emotion_classification']
+previous_lora_path: /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights,/home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+srt_load_path:      /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+sgwi:               True
+============================================================
+
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/train.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/dev.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+Train/dev/test: 975/100/681
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   3%|▎         | 10/291 [00:00<00:02, 96.84it/s]
Loading weights:  10%|▉         | 29/291 [00:00<00:01, 144.57it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 161.24it/s]
Loading weights:  23%|██▎       | 67/291 [00:00<00:01, 167.10it/s]
Loading weights:  30%|██▉       | 86/291 [00:00<00:01, 172.22it/s]
Loading weights:  37%|███▋      | 108/291 [00:00<00:00, 183.99it/s]
Loading weights:  44%|████▎     | 127/291 [00:00<00:00, 178.95it/s]
Loading weights:  50%|█████     | 146/291 [00:00<00:00, 181.71it/s]
Loading weights:  57%|█████▋    | 166/291 [00:00<00:00, 167.80it/s]
Loading weights:  63%|██████▎   | 184/291 [00:01<00:00, 170.55it/s]
Loading weights:  70%|██████▉   | 203/291 [00:01<00:00, 172.69it/s]
Loading weights:  78%|███████▊  | 227/291 [00:01<00:00, 186.25it/s]
Loading weights:  85%|████████▍ | 246/291 [00:01<00:00, 178.11it/s]
Loading weights:  91%|█████████ | 265/291 [00:01<00:00, 178.24it/s]
Loading weights:  97%|█████████▋| 283/291 [00:01<00:00, 178.55it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 176.71it/s]
+LlamaForCausalLM has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+LlamaModel has generative capabilities, as `prepare_inputs_for_generation` is explicitly defined. However, it doesn't directly inherit from `GenerationMixin`. From 👉v4.50👈 onwards, `PreTrainedModel` will NOT inherit from `GenerationMixin`, and this model will lose the ability to call `generate` and other related functions.
+  - If you're using `trust_remote_code=True`, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
+  - If you are the owner of the model architecture code, please modify your model class such that it inherits from `GenerationMixin` (after `PreTrainedModel`, otherwise you'll get an exception).
+  - If you are not the owner of the model architecture class, please contact the model code owner to update it.
+[LOAD] HF checkpoint copied into custom LLaMA (allowed_missing=1187)
+[FIX] Re-initialized lora_A/B in 576 LoRA layers
+[FIX] Re-initialized prompt_key ~ U(-1,1)
+[FIX] n_slots: dummy previous_prompts_keys shape=(8, 4096)
+[PREV-LORA] Loaded slot[0] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights
+[PREV-LORA] Loaded slot[1] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/7-task639_multi_woz_user_utterance_generation/saved_weights
+[PREV-LORA] Loaded slot[2] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/6-task1510_evalution_relation_extraction/saved_weights
+[PREV-LORA] Loaded slot[3] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/5-task002_quoref_answer_generation/saved_weights
+[PREV-LORA] Loaded slot[4] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/4-task181_outcome_extraction/saved_weights
+[PREV-LORA] Loaded slot[5] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/3-task1290_xsum_summarization/saved_weights
+[PREV-LORA] Loaded slot[6] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/2-task363_sst2_polarity_classification/saved_weights
+[PREV-LORA] Loaded slot[7] ← /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/1-task1572_samsum_summary/saved_weights
+[PREV-LORA] Moved 8 past adapters to CPU
+[FIX] Upcast 131 trainable tensors to fp32
+Trainable: 2.920M / 6758.146M (0.04%)
+[TRAIN] step_per_epoch=31, eval_steps=155
+  [SRT] Loaded 8 signatures from /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/8-task1729_personachat_generate_next/saved_weights/srt_signatures.npz
+  [SRT] Wired router: 8 sigs, cur=task073_commonsenseqa_answer_generation, mapping={'task073_commonsenseqa_answer_generation': 0, 'task1572_samsum_summary': 8, 'task363_sst2_polarity_classification': 7, 'task1290_xsum_summarization': 6, 'task181_outcome_extraction': 5, 'task002_quoref_answer_generation': 4, 'task1510_evalution_relation_extraction': 3, 'task639_multi_woz_user_utterance_generation': 2, 'task1729_personachat_generate_next': 1}
+[SGWI] Task 8: computing SRT weights for warm-init…
+  [SRT] WARNING: encoder_frozen not attached. Returning empty embeddings.
+[SGWI] No SRT weights available. Skipping warm-init.
+[RANK 0] Gradient accumulation steps mismatch: GradientAccumulationPlugin has 1, DeepSpeed config has 32. Using DeepSpeed's value.
+[2026-04-30 00:51:07,548] [WARNING] [lr_schedules.py:690:get_lr] Attempting to get learning rate from scheduler before it has started
+
  0%|          | 0/465 [00:00<?, ?it/s]
  0%|          | 1/465 [00:04<37:56,  4.91s/it]
  0%|          | 2/465 [00:08<34:05,  4.42s/it]
  1%|          | 3/465 [00:13<33:24,  4.34s/it]
  1%|          | 4/465 [00:17<32:56,  4.29s/it]
  1%|          | 5/465 [00:21<32:48,  4.28s/it]
  1%|▏         | 6/465 [00:26<33:16,  4.35s/it]
  2%|▏         | 7/465 [00:30<32:57,  4.32s/it]
  2%|▏         | 8/465 [00:34<32:30,  4.27s/it]
  2%|▏         | 9/465 [00:39<32:54,  4.33s/it]
  2%|▏         | 10/465 [00:43<32:41,  4.31s/it]
                                                
{'loss': '3.426', 'grad_norm': '1.011', 'learning_rate': '5e-05', 'epoch': '0.3282'}
+
  2%|▏         | 10/465 [00:43<32:41,  4.31s/it]
  2%|▏         | 11/465 [00:47<33:05,  4.37s/it]
  3%|▎         | 12/465 [00:52<33:13,  4.40s/it]
  3%|▎         | 13/465 [00:56<32:42,  4.34s/it]
  3%|▎         | 14/465 [01:00<32:32,  4.33s/it]
  3%|▎         | 15/465 [01:04<32:01,  4.27s/it]
  3%|▎         | 16/465 [01:09<32:01,  4.28s/it]
  4%|▎         | 17/465 [01:13<31:31,  4.22s/it]
  4%|▍         | 18/465 [01:17<31:19,  4.20s/it]
  4%|▍         | 19/465 [01:21<31:07,  4.19s/it]
  4%|▍         | 20/465 [01:25<31:14,  4.21s/it]
                                                
{'loss': '1.208', 'grad_norm': '0.3496', 'learning_rate': '5e-05', 'epoch': '0.6564'}
+
  4%|▍         | 20/465 [01:25<31:14,  4.21s/it]
  5%|▍         | 21/465 [01:30<31:12,  4.22s/it]
  5%|▍         | 22/465 [01:34<30:52,  4.18s/it]
  5%|▍         | 23/465 [01:38<30:57,  4.20s/it]
  5%|▌         | 24/465 [01:42<31:01,  4.22s/it]
  5%|▌         | 25/465 [01:46<30:41,  4.19s/it]
  6%|▌         | 26/465 [01:51<30:47,  4.21s/it]
  6%|▌         | 27/465 [01:55<30:53,  4.23s/it]
  6%|▌         | 28/465 [01:59<30:45,  4.22s/it]
  6%|▌         | 29/465 [02:03<30:28,  4.19s/it]
  6%|▋         | 30/465 [02:08<30:35,  4.22s/it]
                                                
{'loss': '0.8217', 'grad_norm': '0.9899', 'learning_rate': '5e-05', 'epoch': '0.9846'}
+
  6%|▋         | 30/465 [02:08<30:35,  4.22s/it]
  7%|▋         | 31/465 [02:10<26:03,  3.60s/it]
  7%|▋         | 32/465 [02:14<27:31,  3.81s/it]
  7%|▋         | 33/465 [02:18<28:30,  3.96s/it]
  7%|▋         | 34/465 [02:23<29:51,  4.16s/it]
  8%|▊         | 35/465 [02:27<30:03,  4.19s/it]
  8%|▊         | 36/465 [02:32<30:35,  4.28s/it]
  8%|▊         | 37/465 [02:36<30:32,  4.28s/it]
  8%|▊         | 38/465 [02:41<32:45,  4.60s/it]
  8%|▊         | 39/465 [02:46<32:04,  4.52s/it]
  9%|▊         | 40/465 [02:50<31:30,  4.45s/it]
                                                
{'loss': '0.7474', 'grad_norm': '0.8704', 'learning_rate': '5e-05', 'epoch': '1.295'}
+
  9%|▊         | 40/465 [02:50<31:30,  4.45s/it]
  9%|▉         | 41/465 [02:54<31:01,  4.39s/it]
  9%|▉         | 42/465 [02:58<30:21,  4.31s/it]
  9%|▉         | 43/465 [03:03<30:10,  4.29s/it]
  9%|▉         | 44/465 [03:07<29:45,  4.24s/it]
 10%|▉         | 45/465 [03:11<29:25,  4.20s/it]
 10%|▉         | 46/465 [03:15<29:23,  4.21s/it]
 10%|█         | 47/465 [03:19<29:07,  4.18s/it]
 10%|█         | 48/465 [03:24<29:44,  4.28s/it]
 11%|█         | 49/465 [03:28<29:20,  4.23s/it]
 11%|█         | 50/465 [03:32<29:15,  4.23s/it]
                                                
{'loss': '0.7079', 'grad_norm': '1.178', 'learning_rate': '5e-05', 'epoch': '1.624'}
+
 11%|█         | 50/465 [03:32<29:15,  4.23s/it]
 11%|█         | 51/465 [03:36<29:18,  4.25s/it]
 11%|█         | 52/465 [03:41<30:20,  4.41s/it]
 11%|█▏        | 53/465 [03:45<29:48,  4.34s/it]
 12%|█▏        | 54/465 [03:49<29:19,  4.28s/it]
 12%|█▏        | 55/465 [03:54<29:28,  4.31s/it]
 12%|█▏        | 56/465 [03:58<30:11,  4.43s/it]
 12%|█▏        | 57/465 [04:03<29:50,  4.39s/it]
 12%|█▏        | 58/465 [04:07<29:58,  4.42s/it]
 13%|█▎        | 59/465 [04:12<29:39,  4.38s/it]
 13%|█▎        | 60/465 [04:16<29:25,  4.36s/it]
                                                
{'loss': '0.6605', 'grad_norm': '0.3512', 'learning_rate': '5e-05', 'epoch': '1.952'}
+
 13%|█▎        | 60/465 [04:16<29:25,  4.36s/it]
 13%|█▎        | 61/465 [04:20<29:20,  4.36s/it]
 13%|█▎        | 62/465 [04:22<24:32,  3.65s/it]
 14%|█▎        | 63/465 [04:27<26:01,  3.88s/it]
 14%|█▍        | 64/465 [04:31<26:47,  4.01s/it]
 14%|█▍        | 65/465 [04:35<27:22,  4.11s/it]
 14%|█▍        | 66/465 [04:40<29:03,  4.37s/it]
 14%|█▍        | 67/465 [04:45<30:08,  4.54s/it]
 15%|█▍        | 68/465 [04:50<29:51,  4.51s/it]
 15%|█▍        | 69/465 [04:54<29:04,  4.41s/it]
 15%|█▌        | 70/465 [04:58<28:38,  4.35s/it]
                                                
{'loss': '0.731', 'grad_norm': '7.853', 'learning_rate': '5e-05', 'epoch': '2.263'}
+
 15%|█▌        | 70/465 [04:58<28:38,  4.35s/it]
 15%|█▌        | 71/465 [05:02<28:12,  4.29s/it]
 15%|█▌        | 72/465 [05:06<27:47,  4.24s/it]
 16%|█▌        | 73/465 [05:11<27:41,  4.24s/it]
 16%|█▌        | 74/465 [05:15<27:23,  4.20s/it]
 16%|█▌        | 75/465 [05:19<27:39,  4.25s/it]
 16%|█▋        | 76/465 [05:23<27:29,  4.24s/it]
 17%|█▋        | 77/465 [05:27<27:11,  4.20s/it]
 17%|█▋        | 78/465 [05:32<27:08,  4.21s/it]
 17%|█▋        | 79/465 [05:36<26:55,  4.19s/it]
 17%|█▋        | 80/465 [05:40<27:13,  4.24s/it]
                                                
{'loss': '0.7994', 'grad_norm': '0.8498', 'learning_rate': '5e-05', 'epoch': '2.591'}
+
 17%|█▋        | 80/465 [05:40<27:13,  4.24s/it]
 17%|█▋        | 81/465 [05:44<27:27,  4.29s/it]
 18%|█▊        | 82/465 [05:49<27:28,  4.30s/it]
 18%|█▊        | 83/465 [05:53<27:56,  4.39s/it]
 18%|█▊        | 84/465 [05:58<27:22,  4.31s/it]
 18%|█▊        | 85/465 [06:02<27:35,  4.36s/it]
 18%|█▊        | 86/465 [06:06<27:06,  4.29s/it]
 19%|█▊        | 87/465 [06:10<27:05,  4.30s/it]
 19%|█▉        | 88/465 [06:15<27:18,  4.35s/it]
 19%|█▉        | 89/465 [06:19<26:53,  4.29s/it]
 19%|█▉        | 90/465 [06:24<27:10,  4.35s/it]
                                                
{'loss': '0.6636', 'grad_norm': '1.495', 'learning_rate': '5e-05', 'epoch': '2.919'}
+
 19%|█▉        | 90/465 [06:24<27:10,  4.35s/it]
 20%|█▉        | 91/465 [06:28<27:00,  4.33s/it]
 20%|█▉        | 92/465 [06:32<26:53,  4.32s/it]
 20%|██        | 93/465 [06:34<22:58,  3.71s/it]
 20%|██        | 94/465 [06:39<24:01,  3.89s/it]
 20%|██        | 95/465 [06:43<24:48,  4.02s/it]
 21%|██        | 96/465 [06:47<25:08,  4.09s/it]
 21%|██        | 97/465 [06:51<25:10,  4.10s/it]
 21%|██        | 98/465 [06:56<25:34,  4.18s/it]
 21%|██▏       | 99/465 [07:00<25:26,  4.17s/it]
 22%|██▏       | 100/465 [07:04<25:21,  4.17s/it]
                                                 
{'loss': '0.6279', 'grad_norm': '0.8302', 'learning_rate': '5e-05', 'epoch': '3.23'}
+
 22%|██▏       | 100/465 [07:04<25:21,  4.17s/it]
 22%|██▏       | 101/465 [07:08<25:20,  4.18s/it]
 22%|██▏       | 102/465 [07:12<25:11,  4.16s/it]
 22%|██▏       | 103/465 [07:17<25:34,  4.24s/it]
 22%|██▏       | 104/465 [07:21<25:20,  4.21s/it]
 23%|██▎       | 105/465 [07:25<25:09,  4.19s/it]
 23%|██▎       | 106/465 [07:29<25:10,  4.21s/it]
 23%|██▎       | 107/465 [07:34<25:16,  4.24s/it]
 23%|██▎       | 108/465 [07:38<25:19,  4.26s/it]
 23%|██▎       | 109/465 [07:42<25:06,  4.23s/it]
 24%|██▎       | 110/465 [07:46<25:10,  4.25s/it]
                                                 
{'loss': '0.6103', 'grad_norm': '0.7692', 'learning_rate': '5e-05', 'epoch': '3.558'}
+
 24%|██▎       | 110/465 [07:46<25:10,  4.25s/it]
 24%|██▍       | 111/465 [07:51<25:22,  4.30s/it]
 24%|██▍       | 112/465 [07:55<25:24,  4.32s/it]
 24%|██▍       | 113/465 [08:00<25:18,  4.31s/it]
 25%|██▍       | 114/465 [08:04<25:12,  4.31s/it]
 25%|██▍       | 115/465 [08:08<25:00,  4.29s/it]
 25%|██▍       | 116/465 [08:12<25:00,  4.30s/it]
 25%|██▌       | 117/465 [08:17<24:58,  4.30s/it]
 25%|██▌       | 118/465 [08:21<25:17,  4.37s/it]
 26%|██▌       | 119/465 [08:26<25:08,  4.36s/it]
 26%|██▌       | 120/465 [08:30<25:06,  4.37s/it]
                                                 
{'loss': '0.6197', 'grad_norm': '1.241', 'learning_rate': '5e-05', 'epoch': '3.886'}
+
 26%|██▌       | 120/465 [08:30<25:06,  4.37s/it]
 26%|██▌       | 121/465 [08:34<24:56,  4.35s/it]
 26%|██▌       | 122/465 [08:39<24:46,  4.33s/it]
 26%|██▋       | 123/465 [08:43<24:40,  4.33s/it]
 27%|██▋       | 124/465 [08:45<20:34,  3.62s/it]
 27%|██▋       | 125/465 [08:49<21:27,  3.79s/it]
 27%|██▋       | 126/465 [08:53<22:06,  3.91s/it]
 27%|██▋       | 127/465 [08:58<22:40,  4.03s/it]
 28%|██▊       | 128/465 [09:02<23:20,  4.16s/it]
 28%|██▊       | 129/465 [09:06<23:13,  4.15s/it]
 28%|██▊       | 130/465 [09:10<23:24,  4.19s/it]
                                                 
{'loss': '0.5876', 'grad_norm': '0.954', 'learning_rate': '5e-05', 'epoch': '4.197'}
+
 28%|██▊       | 130/465 [09:10<23:24,  4.19s/it]
 28%|██▊       | 131/465 [09:15<23:27,  4.21s/it]
 28%|██▊       | 132/465 [09:19<23:22,  4.21s/it]
 29%|██▊       | 133/465 [09:23<23:16,  4.21s/it]
 29%|██▉       | 134/465 [09:27<23:05,  4.18s/it]
 29%|██▉       | 135/465 [09:32<23:14,  4.22s/it]
 29%|██▉       | 136/465 [09:36<23:22,  4.26s/it]
 29%|██▉       | 137/465 [09:40<23:09,  4.24s/it]
 30%|██▉       | 138/465 [09:44<23:19,  4.28s/it]
 30%|██▉       | 139/465 [09:49<23:03,  4.24s/it]
 30%|███       | 140/465 [09:53<23:04,  4.26s/it]
                                                 
{'loss': '0.6152', 'grad_norm': '0.5598', 'learning_rate': '5e-05', 'epoch': '4.525'}
+
 30%|███       | 140/465 [09:53<23:04,  4.26s/it]
 30%|███       | 141/465 [09:57<22:53,  4.24s/it]
 31%|███       | 142/465 [10:02<23:28,  4.36s/it]
 31%|███       | 143/465 [10:06<23:30,  4.38s/it]
 31%|███       | 144/465 [10:10<23:05,  4.32s/it]
 31%|███       | 145/465 [10:15<23:07,  4.34s/it]
 31%|███▏      | 146/465 [10:19<23:04,  4.34s/it]
 32%|███▏      | 147/465 [10:23<22:56,  4.33s/it]
 32%|███▏      | 148/465 [10:28<22:58,  4.35s/it]
 32%|███▏      | 149/465 [10:32<22:50,  4.34s/it]
 32%|███▏      | 150/465 [10:36<22:53,  4.36s/it]
                                                 
{'loss': '0.5704', 'grad_norm': '1.049', 'learning_rate': '5e-05', 'epoch': '4.853'}
+
 32%|███▏      | 150/465 [10:36<22:53,  4.36s/it]
 32%|███▏      | 151/465 [10:41<22:35,  4.32s/it]
 33%|███▎      | 152/465 [10:45<22:44,  4.36s/it]
 33%|███▎      | 153/465 [10:50<22:53,  4.40s/it]
 33%|███▎      | 154/465 [10:54<22:24,  4.32s/it]
 33%|███▎      | 155/465 [10:56<18:38,  3.61s/it]
+
Eval @ step 155:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 155:   2%|▏         | 2/100 [00:00<00:09, 10.81it/s]
+
Eval @ step 155:   4%|▍         | 4/100 [00:00<00:08, 11.25it/s]
+
Eval @ step 155:   6%|▌         | 6/100 [00:00<00:08, 11.44it/s]
+
Eval @ step 155:   8%|▊         | 8/100 [00:00<00:08, 11.48it/s]
+
Eval @ step 155:  10%|█         | 10/100 [00:00<00:07, 11.57it/s]
+
Eval @ step 155:  12%|█▏        | 12/100 [00:01<00:07, 11.60it/s]
+
Eval @ step 155:  14%|█▍        | 14/100 [00:01<00:07, 11.64it/s]
+
Eval @ step 155:  16%|█▌        | 16/100 [00:01<00:07, 11.47it/s]
+
Eval @ step 155:  18%|█▊        | 18/100 [00:01<00:07, 11.30it/s]
+
Eval @ step 155:  20%|██        | 20/100 [00:01<00:06, 11.45it/s]
+
Eval @ step 155:  22%|██▏       | 22/100 [00:01<00:06, 11.51it/s]
+
Eval @ step 155:  24%|██▍       | 24/100 [00:02<00:06, 11.52it/s]
+
Eval @ step 155:  26%|██▌       | 26/100 [00:02<00:06, 11.56it/s]
+
Eval @ step 155:  28%|██▊       | 28/100 [00:02<00:06, 11.66it/s]
+
Eval @ step 155:  30%|███       | 30/100 [00:02<00:05, 11.72it/s]
+
Eval @ step 155:  32%|███▏      | 32/100 [00:02<00:05, 11.70it/s]
+
Eval @ step 155:  34%|███▍      | 34/100 [00:02<00:05, 11.73it/s]
+
Eval @ step 155:  36%|███▌      | 36/100 [00:03<00:05, 11.75it/s]
+
Eval @ step 155:  38%|███▊      | 38/100 [00:03<00:05, 11.83it/s]
+
Eval @ step 155:  40%|████      | 40/100 [00:03<00:05, 11.88it/s]
+
Eval @ step 155:  42%|████▏     | 42/100 [00:03<00:04, 11.95it/s]
+
Eval @ step 155:  44%|████▍     | 44/100 [00:03<00:04, 12.00it/s]
+
Eval @ step 155:  46%|████▌     | 46/100 [00:03<00:04, 12.02it/s]
+
Eval @ step 155:  48%|████▊     | 48/100 [00:04<00:04, 12.08it/s]
+
Eval @ step 155:  50%|█████     | 50/100 [00:04<00:04, 12.06it/s]
+
Eval @ step 155:  52%|█████▏    | 52/100 [00:04<00:03, 12.08it/s]
+
Eval @ step 155:  54%|█████▍    | 54/100 [00:04<00:03, 12.10it/s]
+
Eval @ step 155:  56%|█████▌    | 56/100 [00:04<00:03, 12.08it/s]
+
Eval @ step 155:  58%|█████▊    | 58/100 [00:04<00:03, 12.05it/s]
+
Eval @ step 155:  60%|██████    | 60/100 [00:05<00:03, 12.11it/s]
+
Eval @ step 155:  62%|██████▏   | 62/100 [00:05<00:03, 12.08it/s]
+
Eval @ step 155:  64%|██████▍   | 64/100 [00:05<00:02, 12.10it/s]
+
Eval @ step 155:  66%|██████▌   | 66/100 [00:05<00:02, 12.07it/s]
+
Eval @ step 155:  68%|██████▊   | 68/100 [00:05<00:02, 12.10it/s]
+
Eval @ step 155:  70%|███████   | 70/100 [00:05<00:02, 12.14it/s]
+
Eval @ step 155:  72%|███████▏  | 72/100 [00:06<00:02, 12.16it/s]
+
Eval @ step 155:  74%|███████▍  | 74/100 [00:06<00:02, 12.19it/s]
+
Eval @ step 155:  76%|███████▌  | 76/100 [00:06<00:01, 12.13it/s]
+
Eval @ step 155:  78%|███████▊  | 78/100 [00:06<00:01, 12.13it/s]
+
Eval @ step 155:  80%|████████  | 80/100 [00:06<00:01, 12.10it/s]
+
Eval @ step 155:  82%|████████▏ | 82/100 [00:06<00:01, 12.07it/s]
+
Eval @ step 155:  84%|████████▍ | 84/100 [00:07<00:01, 12.06it/s]
+
Eval @ step 155:  86%|████████▌ | 86/100 [00:07<00:01, 12.08it/s]
+
Eval @ step 155:  88%|████████▊ | 88/100 [00:07<00:00, 12.09it/s]
+
Eval @ step 155:  90%|█████████ | 90/100 [00:07<00:00, 12.07it/s]
+
Eval @ step 155:  92%|█████████▏| 92/100 [00:07<00:00, 12.08it/s]
+
Eval @ step 155:  94%|█████████▍| 94/100 [00:07<00:00, 12.09it/s]
+
Eval @ step 155:  96%|█████████▌| 96/100 [00:08<00:00, 12.14it/s]
+
Eval @ step 155:  98%|█████████▊| 98/100 [00:08<00:00, 12.11it/s]
+
Eval @ step 155: 100%|██████████| 100/100 [00:08<00:00, 11.24it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=56.0000 (step=155) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '56', 'epoch': '5'}
+
 33%|███▎      | 155/465 [11:04<18:38,  3.61s/it]
 34%|███▎      | 156/465 [11:09<33:01,  6.41s/it]
 34%|███▍      | 157/465 [11:13<29:39,  5.78s/it]
 34%|███▍      | 158/465 [11:18<27:41,  5.41s/it]
 34%|███▍      | 159/465 [11:22<26:06,  5.12s/it]
 34%|███▍      | 160/465 [11:26<24:29,  4.82s/it]
                                                 
{'loss': '0.5724', 'grad_norm': '1.221', 'learning_rate': '5e-05', 'epoch': '5.164'}
+
 34%|███▍      | 160/465 [11:26<24:29,  4.82s/it]
 35%|███▍      | 161/465 [11:30<23:28,  4.63s/it]
 35%|███▍      | 162/465 [11:34<22:39,  4.49s/it]
 35%|███▌      | 163/465 [11:39<22:04,  4.38s/it]
 35%|███▌      | 164/465 [11:43<22:08,  4.41s/it]
 35%|███▌      | 165/465 [11:47<21:40,  4.34s/it]
 36%|███▌      | 166/465 [11:52<21:37,  4.34s/it]
 36%|███▌      | 167/465 [11:56<21:14,  4.28s/it]
 36%|███▌      | 168/465 [12:00<21:00,  4.24s/it]
 36%|███▋      | 169/465 [12:05<21:37,  4.38s/it]
 37%|███▋      | 170/465 [12:09<21:10,  4.31s/it]
                                                 
{'loss': '0.5256', 'grad_norm': '0.3565', 'learning_rate': '5e-05', 'epoch': '5.492'}
+
 37%|███▋      | 170/465 [12:09<21:10,  4.31s/it]
 37%|███▋      | 171/465 [12:13<21:15,  4.34s/it]
 37%|███▋      | 172/465 [12:17<20:53,  4.28s/it]
 37%|███▋      | 173/465 [12:21<20:38,  4.24s/it]
 37%|███▋      | 174/465 [12:26<21:10,  4.37s/it]
 38%|███▊      | 175/465 [12:31<21:30,  4.45s/it]
 38%|███▊      | 176/465 [12:35<21:24,  4.45s/it]
 38%|███▊      | 177/465 [12:39<20:56,  4.36s/it]
 38%|███▊      | 178/465 [12:44<20:50,  4.36s/it]
 38%|███▊      | 179/465 [12:48<20:42,  4.35s/it]
 39%|███▊      | 180/465 [12:52<20:24,  4.30s/it]
                                                 
{'loss': '0.587', 'grad_norm': '1.247', 'learning_rate': '5e-05', 'epoch': '5.821'}
+
 39%|███▊      | 180/465 [12:52<20:24,  4.30s/it]
 39%|███▉      | 181/465 [12:57<20:34,  4.35s/it]
 39%|███▉      | 182/465 [13:01<20:25,  4.33s/it]
 39%|███▉      | 183/465 [13:05<20:28,  4.36s/it]
 40%|███▉      | 184/465 [13:10<20:19,  4.34s/it]
 40%|███▉      | 185/465 [13:14<20:24,  4.37s/it]
 40%|████      | 186/465 [13:16<17:16,  3.72s/it]
 40%|████      | 187/465 [13:20<17:51,  3.85s/it]
 40%|████      | 188/465 [13:25<18:53,  4.09s/it]
 41%|████      | 189/465 [13:29<18:57,  4.12s/it]
 41%|████      | 190/465 [13:33<18:53,  4.12s/it]
                                                 
{'loss': '0.5685', 'grad_norm': '0.5744', 'learning_rate': '5e-05', 'epoch': '6.131'}
+
 41%|████      | 190/465 [13:33<18:53,  4.12s/it]
 41%|████      | 191/465 [13:38<19:03,  4.17s/it]
 41%|████▏     | 192/465 [13:42<18:55,  4.16s/it]
 42%|████▏     | 193/465 [13:46<18:53,  4.17s/it]
 42%|████▏     | 194/465 [13:50<19:05,  4.23s/it]
 42%|████▏     | 195/465 [13:55<18:53,  4.20s/it]
 42%|████▏     | 196/465 [13:59<18:54,  4.22s/it]
 42%|████▏     | 197/465 [14:03<18:47,  4.21s/it]
 43%|████▎     | 198/465 [14:07<19:06,  4.29s/it]
 43%|████▎     | 199/465 [14:12<18:50,  4.25s/it]
 43%|████▎     | 200/465 [14:16<18:36,  4.21s/it]
                                                 
{'loss': '0.5677', 'grad_norm': '0.3822', 'learning_rate': '5e-05', 'epoch': '6.459'}
+
 43%|████▎     | 200/465 [14:16<18:36,  4.21s/it]
 43%|████▎     | 201/465 [14:20<18:55,  4.30s/it]
 43%|████▎     | 202/465 [14:24<18:37,  4.25s/it]
 44%|████▎     | 203/465 [14:29<19:06,  4.38s/it]
 44%|████▍     | 204/465 [14:33<18:43,  4.30s/it]
 44%|████▍     | 205/465 [14:37<18:39,  4.30s/it]
 44%|████▍     | 206/465 [14:42<18:30,  4.29s/it]
 45%|████▍     | 207/465 [14:46<18:30,  4.30s/it]
 45%|████▍     | 208/465 [14:50<18:31,  4.33s/it]
 45%|████▍     | 209/465 [14:55<18:28,  4.33s/it]
 45%|████▌     | 210/465 [14:59<18:23,  4.33s/it]
                                                 
{'loss': '0.5249', 'grad_norm': '1.139', 'learning_rate': '5e-05', 'epoch': '6.788'}
+
 45%|████▌     | 210/465 [14:59<18:23,  4.33s/it]
 45%|████▌     | 211/465 [15:04<18:35,  4.39s/it]
 46%|████▌     | 212/465 [15:08<18:25,  4.37s/it]
 46%|████▌     | 213/465 [15:12<18:25,  4.39s/it]
 46%|████▌     | 214/465 [15:17<18:02,  4.31s/it]
 46%|████▌     | 215/465 [15:21<17:47,  4.27s/it]
 46%|████▋     | 216/465 [15:25<17:59,  4.33s/it]
 47%|████▋     | 217/465 [15:27<14:57,  3.62s/it]
 47%|████▋     | 218/465 [15:32<16:15,  3.95s/it]
 47%|████▋     | 219/465 [15:36<16:25,  4.01s/it]
 47%|████▋     | 220/465 [15:40<16:44,  4.10s/it]
                                                 
{'loss': '0.4867', 'grad_norm': '1.266', 'learning_rate': '5e-05', 'epoch': '7.098'}
+
 47%|████▋     | 220/465 [15:40<16:44,  4.10s/it]
 48%|████▊     | 221/465 [15:45<17:16,  4.25s/it]
 48%|████▊     | 222/465 [15:49<17:08,  4.23s/it]
 48%|████▊     | 223/465 [15:53<17:03,  4.23s/it]
 48%|████▊     | 224/465 [15:57<16:53,  4.20s/it]
 48%|████▊     | 225/465 [16:02<16:44,  4.19s/it]
 49%|████▊     | 226/465 [16:06<17:06,  4.29s/it]
 49%|████▉     | 227/465 [16:10<16:50,  4.24s/it]
 49%|████▉     | 228/465 [16:15<16:47,  4.25s/it]
 49%|████▉     | 229/465 [16:19<16:47,  4.27s/it]
 49%|████▉     | 230/465 [16:23<16:34,  4.23s/it]
                                                 
{'loss': '0.5534', 'grad_norm': '0.8952', 'learning_rate': '5e-05', 'epoch': '7.427'}
+
 49%|████▉     | 230/465 [16:23<16:34,  4.23s/it]
 50%|████▉     | 231/465 [16:27<16:45,  4.30s/it]
 50%|████▉     | 232/465 [16:32<16:29,  4.25s/it]
 50%|█████     | 233/465 [16:36<16:38,  4.30s/it]
 50%|█████     | 234/465 [16:40<16:34,  4.31s/it]
 51%|█████     | 235/465 [16:45<16:22,  4.27s/it]
 51%|█████     | 236/465 [16:49<16:25,  4.30s/it]
 51%|█████     | 237/465 [16:53<16:22,  4.31s/it]
 51%|█████     | 238/465 [16:58<16:47,  4.44s/it]
 51%|█████▏    | 239/465 [17:02<16:34,  4.40s/it]
 52%|█████▏    | 240/465 [17:07<16:30,  4.40s/it]
                                                 
{'loss': '0.501', 'grad_norm': '1.388', 'learning_rate': '5e-05', 'epoch': '7.755'}
+
 52%|█████▏    | 240/465 [17:07<16:30,  4.40s/it]
 52%|█████▏    | 241/465 [17:11<16:18,  4.37s/it]
 52%|█████▏    | 242/465 [17:15<15:58,  4.30s/it]
 52%|█████▏    | 243/465 [17:19<15:50,  4.28s/it]
 52%|█████▏    | 244/465 [17:23<15:36,  4.24s/it]
 53%|█████▎    | 245/465 [17:28<15:41,  4.28s/it]
 53%|█████▎    | 246/465 [17:32<15:29,  4.25s/it]
 53%|█████▎    | 247/465 [17:36<15:17,  4.21s/it]
 53%|█████▎    | 248/465 [17:38<12:51,  3.56s/it]
 54%|█████▎    | 249/465 [17:43<13:40,  3.80s/it]
 54%|█████▍    | 250/465 [17:47<13:57,  3.90s/it]
                                                 
{'loss': '0.5155', 'grad_norm': '0.8428', 'learning_rate': '5e-05', 'epoch': '8.066'}
+
 54%|█████▍    | 250/465 [17:47<13:57,  3.90s/it]
 54%|█████▍    | 251/465 [17:51<14:11,  3.98s/it]
 54%|█████▍    | 252/465 [17:55<14:17,  4.03s/it]
 54%|█████▍    | 253/465 [17:59<14:35,  4.13s/it]
 55%|█████▍    | 254/465 [18:04<14:34,  4.14s/it]
 55%|█████▍    | 255/465 [18:08<14:30,  4.15s/it]
 55%|█████▌    | 256/465 [18:12<14:39,  4.21s/it]
 55%|█████▌    | 257/465 [18:16<14:30,  4.19s/it]
 55%|█████▌    | 258/465 [18:20<14:31,  4.21s/it]
 56%|█████▌    | 259/465 [18:25<14:37,  4.26s/it]
 56%|█████▌    | 260/465 [18:29<14:35,  4.27s/it]
                                                 
{'loss': '0.5066', 'grad_norm': '0.3254', 'learning_rate': '5e-05', 'epoch': '8.394'}
+
 56%|█████▌    | 260/465 [18:29<14:35,  4.27s/it]
 56%|█████▌    | 261/465 [18:33<14:30,  4.27s/it]
 56%|█████▋    | 262/465 [18:38<14:29,  4.28s/it]
 57%|█████▋    | 263/465 [18:42<14:36,  4.34s/it]
 57%|█████▋    | 264/465 [18:47<14:38,  4.37s/it]
 57%|█████▋    | 265/465 [18:51<14:30,  4.35s/it]
 57%|█████▋    | 266/465 [18:56<14:42,  4.43s/it]
 57%|█████▋    | 267/465 [19:00<14:31,  4.40s/it]
 58%|█████▊    | 268/465 [19:04<14:28,  4.41s/it]
 58%|█████▊    | 269/465 [19:08<14:07,  4.32s/it]
 58%|█████▊    | 270/465 [19:13<13:52,  4.27s/it]
                                                 
{'loss': '0.4864', 'grad_norm': '0.7875', 'learning_rate': '5e-05', 'epoch': '8.722'}
+
 58%|█████▊    | 270/465 [19:13<13:52,  4.27s/it]
 58%|█████▊    | 271/465 [19:17<13:55,  4.31s/it]
 58%|█████▊    | 272/465 [19:21<13:41,  4.26s/it]
 59%|█████▊    | 273/465 [19:25<13:37,  4.26s/it]
 59%|█████▉    | 274/465 [19:30<13:27,  4.23s/it]
 59%|█████▉    | 275/465 [19:34<13:18,  4.20s/it]
 59%|█████▉    | 276/465 [19:38<13:24,  4.26s/it]
 60%|█████▉    | 277/465 [19:42<13:14,  4.22s/it]
 60%|█████▉    | 278/465 [19:47<13:15,  4.25s/it]
 60%|██████    | 279/465 [19:48<11:02,  3.56s/it]
 60%|██████    | 280/465 [19:53<11:32,  3.74s/it]
                                                 
{'loss': '0.448', 'grad_norm': '0.4647', 'learning_rate': '5e-05', 'epoch': '9.033'}
+
 60%|██████    | 280/465 [19:53<11:32,  3.74s/it]
 60%|██████    | 281/465 [19:57<12:04,  3.94s/it]
 61%|██████    | 282/465 [20:01<12:13,  4.01s/it]
 61%|██████    | 283/465 [20:05<12:18,  4.06s/it]
 61%|██████    | 284/465 [20:10<12:43,  4.22s/it]
 61%|██████▏   | 285/465 [20:14<12:38,  4.22s/it]
 62%|██████▏   | 286/465 [20:19<12:46,  4.28s/it]
 62%|██████▏   | 287/465 [20:23<12:34,  4.24s/it]
 62%|██████▏   | 288/465 [20:27<12:38,  4.28s/it]
 62%|██████▏   | 289/465 [20:31<12:26,  4.24s/it]
 62%|██████▏   | 290/465 [20:36<12:25,  4.26s/it]
                                                 
{'loss': '0.4506', 'grad_norm': '2.271', 'learning_rate': '5e-05', 'epoch': '9.361'}
+
 62%|██████▏   | 290/465 [20:36<12:25,  4.26s/it]
 63%|██████▎   | 291/465 [20:40<12:29,  4.31s/it]
 63%|██████▎   | 292/465 [20:44<12:25,  4.31s/it]
 63%|██████▎   | 293/465 [20:49<12:24,  4.33s/it]
 63%|██████▎   | 294/465 [20:53<12:19,  4.32s/it]
 63%|██████▎   | 295/465 [20:57<12:14,  4.32s/it]
 64%|██████▎   | 296/465 [21:02<12:14,  4.34s/it]
 64%|██████▍   | 297/465 [21:06<11:59,  4.28s/it]
 64%|██████▍   | 298/465 [21:10<11:50,  4.26s/it]
 64%|██████▍   | 299/465 [21:14<11:41,  4.23s/it]
 65%|██████▍   | 300/465 [21:18<11:40,  4.24s/it]
                                                 
{'loss': '0.4365', 'grad_norm': '0.9307', 'learning_rate': '5e-05', 'epoch': '9.689'}
+
 65%|██████▍   | 300/465 [21:18<11:40,  4.24s/it]
 65%|██████▍   | 301/465 [21:23<11:34,  4.24s/it]
 65%|██████▍   | 302/465 [21:27<11:25,  4.20s/it]
 65%|██████▌   | 303/465 [21:31<11:22,  4.21s/it]
 65%|██████▌   | 304/465 [21:35<11:23,  4.25s/it]
 66%|██████▌   | 305/465 [21:40<11:13,  4.21s/it]
 66%|██████▌   | 306/465 [21:44<11:09,  4.21s/it]
 66%|██████▌   | 307/465 [21:48<11:03,  4.20s/it]
 66%|██████▌   | 308/465 [21:52<11:09,  4.26s/it]
 66%|██████▋   | 309/465 [21:57<11:04,  4.26s/it]
 67%|██████▋   | 310/465 [21:59<09:15,  3.58s/it]
                                                 
{'loss': '0.4342', 'grad_norm': '2.23', 'learning_rate': '5e-05', 'epoch': '10'}
+
 67%|██████▋   | 310/465 [21:59<09:15,  3.58s/it]
+
Eval @ step 310:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 310:   2%|▏         | 2/100 [00:00<00:08, 11.54it/s]
+
Eval @ step 310:   4%|▍         | 4/100 [00:00<00:08, 11.56it/s]
+
Eval @ step 310:   6%|▌         | 6/100 [00:00<00:08, 11.61it/s]
+
Eval @ step 310:   8%|▊         | 8/100 [00:00<00:07, 11.58it/s]
+
Eval @ step 310:  10%|█         | 10/100 [00:00<00:07, 11.64it/s]
+
Eval @ step 310:  12%|█▏        | 12/100 [00:01<00:07, 11.68it/s]
+
Eval @ step 310:  14%|█▍        | 14/100 [00:01<00:07, 11.73it/s]
+
Eval @ step 310:  16%|█▌        | 16/100 [00:01<00:07, 11.77it/s]
+
Eval @ step 310:  18%|█▊        | 18/100 [00:01<00:07, 11.67it/s]
+
Eval @ step 310:  20%|██        | 20/100 [00:01<00:06, 11.70it/s]
+
Eval @ step 310:  22%|██▏       | 22/100 [00:01<00:06, 11.67it/s]
+
Eval @ step 310:  24%|██▍       | 24/100 [00:02<00:06, 11.62it/s]
+
Eval @ step 310:  26%|██▌       | 26/100 [00:02<00:06, 11.64it/s]
+
Eval @ step 310:  28%|██▊       | 28/100 [00:02<00:06, 11.71it/s]
+
Eval @ step 310:  30%|███       | 30/100 [00:02<00:05, 11.73it/s]
+
Eval @ step 310:  32%|███▏      | 32/100 [00:02<00:05, 11.68it/s]
+
Eval @ step 310:  34%|███▍      | 34/100 [00:02<00:05, 11.71it/s]
+
Eval @ step 310:  36%|███▌      | 36/100 [00:03<00:05, 11.70it/s]
+
Eval @ step 310:  38%|███▊      | 38/100 [00:03<00:05, 11.67it/s]
+
Eval @ step 310:  40%|████      | 40/100 [00:03<00:05, 11.60it/s]
+
Eval @ step 310:  42%|████▏     | 42/100 [00:03<00:05, 11.57it/s]
+
Eval @ step 310:  44%|████▍     | 44/100 [00:03<00:04, 11.40it/s]
+
Eval @ step 310:  46%|████▌     | 46/100 [00:03<00:04, 11.47it/s]
+
Eval @ step 310:  48%|████▊     | 48/100 [00:04<00:04, 11.58it/s]
+
Eval @ step 310:  50%|█████     | 50/100 [00:04<00:04, 11.60it/s]
+
Eval @ step 310:  52%|█████▏    | 52/100 [00:04<00:04, 11.64it/s]
+
Eval @ step 310:  54%|█████▍    | 54/100 [00:04<00:03, 11.67it/s]
+
Eval @ step 310:  56%|█████▌    | 56/100 [00:04<00:03, 11.65it/s]
+
Eval @ step 310:  58%|█████▊    | 58/100 [00:04<00:03, 11.63it/s]
+
Eval @ step 310:  60%|██████    | 60/100 [00:05<00:03, 11.70it/s]
+
Eval @ step 310:  62%|██████▏   | 62/100 [00:05<00:03, 11.68it/s]
+
Eval @ step 310:  64%|██████▍   | 64/100 [00:05<00:03, 11.71it/s]
+
Eval @ step 310:  66%|██████▌   | 66/100 [00:05<00:02, 11.66it/s]
+
Eval @ step 310:  68%|██████▊   | 68/100 [00:05<00:02, 11.69it/s]
+
Eval @ step 310:  70%|███████   | 70/100 [00:06<00:02, 11.71it/s]
+
Eval @ step 310:  72%|███████▏  | 72/100 [00:06<00:02, 11.71it/s]
+
Eval @ step 310:  74%|███████▍  | 74/100 [00:06<00:02, 11.73it/s]
+
Eval @ step 310:  76%|███████▌  | 76/100 [00:06<00:02, 11.67it/s]
+
Eval @ step 310:  78%|███████▊  | 78/100 [00:06<00:01, 11.67it/s]
+
Eval @ step 310:  80%|████████  | 80/100 [00:06<00:01, 11.64it/s]
+
Eval @ step 310:  82%|████████▏ | 82/100 [00:07<00:01, 11.74it/s]
+
Eval @ step 310:  84%|████████▍ | 84/100 [00:07<00:01, 11.81it/s]
+
Eval @ step 310:  86%|████████▌ | 86/100 [00:07<00:01, 11.90it/s]
+
Eval @ step 310:  88%|████████▊ | 88/100 [00:07<00:01, 11.96it/s]
+
Eval @ step 310:  90%|█████████ | 90/100 [00:07<00:00, 11.96it/s]
+
Eval @ step 310:  92%|█████████▏| 92/100 [00:07<00:00, 11.99it/s]
+
Eval @ step 310:  94%|█████████▍| 94/100 [00:08<00:00, 12.02it/s]
+
Eval @ step 310:  96%|█████████▌| 96/100 [00:08<00:00, 12.08it/s]
+
Eval @ step 310:  98%|█████████▊| 98/100 [00:08<00:00, 12.06it/s]
+
Eval @ step 310: 100%|██████████| 100/100 [00:08<00:00, 12.03it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=60.0000 (step=310) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '60', 'epoch': '10'}
+
 67%|██████▋   | 310/465 [22:07<09:15,  3.58s/it]
 67%|██████▋   | 311/465 [22:11<16:21,  6.37s/it]
 67%|██████▋   | 312/465 [22:16<14:49,  5.81s/it]
 67%|██████▋   | 313/465 [22:20<13:27,  5.31s/it]
 68%|██████▊   | 314/465 [22:24<12:37,  5.02s/it]
 68%|██████▊   | 315/465 [22:29<11:53,  4.76s/it]
 68%|██████▊   | 316/465 [22:33<11:20,  4.57s/it]
 68%|██████▊   | 317/465 [22:37<10:59,  4.46s/it]
 68%|██████▊   | 318/465 [22:41<10:40,  4.36s/it]
 69%|██████▊   | 319/465 [22:45<10:38,  4.37s/it]
 69%|██████▉   | 320/465 [22:50<10:41,  4.42s/it]
                                                 
{'loss': '0.4617', 'grad_norm': '1.325', 'learning_rate': '5e-05', 'epoch': '10.33'}
+
 69%|██████▉   | 320/465 [22:50<10:41,  4.42s/it]
 69%|██████▉   | 321/465 [22:54<10:31,  4.39s/it]
 69%|██████▉   | 322/465 [22:59<10:22,  4.36s/it]
 69%|██████▉   | 323/465 [23:03<10:09,  4.30s/it]
 70%|██████▉   | 324/465 [23:07<10:24,  4.43s/it]
 70%|██████▉   | 325/465 [23:12<10:07,  4.34s/it]
 70%|███████   | 326/465 [23:16<09:54,  4.28s/it]
 70%|███████   | 327/465 [23:20<09:53,  4.30s/it]
 71%|███████   | 328/465 [23:24<09:42,  4.25s/it]
 71%|███████   | 329/465 [23:29<09:45,  4.31s/it]
 71%|███████   | 330/465 [23:33<09:34,  4.26s/it]
                                                 
{'loss': '0.442', 'grad_norm': '0.5633', 'learning_rate': '5e-05', 'epoch': '10.66'}
+
 71%|███████   | 330/465 [23:33<09:34,  4.26s/it]
 71%|███████   | 331/465 [23:37<09:38,  4.32s/it]
 71%|███████▏  | 332/465 [23:42<09:33,  4.31s/it]
 72%|███████▏  | 333/465 [23:46<09:22,  4.26s/it]
 72%|███████▏  | 334/465 [23:50<09:24,  4.31s/it]
 72%|███████▏  | 335/465 [23:54<09:19,  4.31s/it]
 72%|███████▏  | 336/465 [23:59<09:18,  4.33s/it]
 72%|███████▏  | 337/465 [24:03<09:14,  4.33s/it]
 73%|███████▎  | 338/465 [24:07<09:11,  4.34s/it]
 73%|███████▎  | 339/465 [24:12<09:19,  4.44s/it]
 73%|███████▎  | 340/465 [24:16<09:03,  4.35s/it]
                                                 
{'loss': '0.4836', 'grad_norm': '2.108', 'learning_rate': '5e-05', 'epoch': '10.98'}
+
 73%|███████▎  | 340/465 [24:16<09:03,  4.35s/it]
 73%|███████▎  | 341/465 [24:18<07:30,  3.63s/it]
 74%|███████▎  | 342/465 [24:22<07:48,  3.81s/it]
 74%|███████▍  | 343/465 [24:27<07:56,  3.90s/it]
 74%|███████▍  | 344/465 [24:31<08:03,  4.00s/it]
 74%|███████▍  | 345/465 [24:35<08:04,  4.04s/it]
 74%|███████▍  | 346/465 [24:39<08:04,  4.07s/it]
 75%|███████▍  | 347/465 [24:44<08:15,  4.20s/it]
 75%|███████▍  | 348/465 [24:48<08:09,  4.18s/it]
 75%|███████▌  | 349/465 [24:52<08:09,  4.22s/it]
 75%|███████▌  | 350/465 [24:56<08:02,  4.19s/it]
                                                 
{'loss': '0.4308', 'grad_norm': '0.8209', 'learning_rate': '5e-05', 'epoch': '11.3'}
+
 75%|███████▌  | 350/465 [24:56<08:02,  4.19s/it]
 75%|███████▌  | 351/465 [25:00<08:01,  4.22s/it]
 76%|███████▌  | 352/465 [25:05<07:59,  4.25s/it]
 76%|███████▌  | 353/465 [25:09<07:51,  4.21s/it]
 76%|███████▌  | 354/465 [25:13<07:58,  4.31s/it]
 76%|███████▋  | 355/465 [25:18<07:49,  4.26s/it]
 77%|███████▋  | 356/465 [25:22<07:41,  4.24s/it]
 77%|███████▋  | 357/465 [25:26<07:44,  4.30s/it]
 77%|███████▋  | 358/465 [25:31<07:40,  4.30s/it]
 77%|███████▋  | 359/465 [25:35<07:36,  4.31s/it]
 77%|███████▋  | 360/465 [25:39<07:32,  4.31s/it]
                                                 
{'loss': '0.4612', 'grad_norm': '0.1664', 'learning_rate': '5e-05', 'epoch': '11.62'}
+
 77%|███████▋  | 360/465 [25:39<07:32,  4.31s/it]
 78%|███████▊  | 361/465 [25:44<07:34,  4.37s/it]
 78%|███████▊  | 362/465 [25:48<07:29,  4.36s/it]
 78%|███████▊  | 363/465 [25:52<07:19,  4.31s/it]
 78%|███████▊  | 364/465 [25:57<07:22,  4.38s/it]
 78%|███████▊  | 365/465 [26:01<07:16,  4.36s/it]
 79%|███████▊  | 366/465 [26:05<07:13,  4.38s/it]
 79%|███████▉  | 367/465 [26:10<07:02,  4.31s/it]
 79%|███████▉  | 368/465 [26:14<06:53,  4.26s/it]
 79%|███████▉  | 369/465 [26:18<06:51,  4.28s/it]
 80%|███████▉  | 370/465 [26:22<06:42,  4.24s/it]
                                                 
{'loss': '0.3949', 'grad_norm': '1.036', 'learning_rate': '5e-05', 'epoch': '11.95'}
+
 80%|███████▉  | 370/465 [26:22<06:42,  4.24s/it]
 80%|███████▉  | 371/465 [26:27<06:40,  4.26s/it]
 80%|████████  | 372/465 [26:28<05:31,  3.56s/it]
 80%|████████  | 373/465 [26:33<05:43,  3.73s/it]
 80%|████████  | 374/465 [26:37<05:52,  3.87s/it]
 81%|████████  | 375/465 [26:41<06:00,  4.01s/it]
 81%|████████  | 376/465 [26:45<06:00,  4.05s/it]
 81%|████████  | 377/465 [26:50<06:04,  4.14s/it]
 81%|████████▏ | 378/465 [26:54<06:00,  4.14s/it]
 82%|████████▏ | 379/465 [26:58<06:02,  4.22s/it]
 82%|████████▏ | 380/465 [27:02<05:56,  4.19s/it]
                                                 
{'loss': '0.4279', 'grad_norm': '1.313', 'learning_rate': '5e-05', 'epoch': '12.26'}
+
 82%|████████▏ | 380/465 [27:02<05:56,  4.19s/it]
 82%|████████▏ | 381/465 [27:06<05:50,  4.17s/it]
 82%|████████▏ | 382/465 [27:11<05:52,  4.25s/it]
 82%|████████▏ | 383/465 [27:15<05:45,  4.21s/it]
 83%|████████▎ | 384/465 [27:19<05:45,  4.26s/it]
 83%|████████▎ | 385/465 [27:24<05:38,  4.23s/it]
 83%|████████▎ | 386/465 [27:28<05:37,  4.28s/it]
 83%|████████▎ | 387/465 [27:32<05:35,  4.30s/it]
 83%|████████▎ | 388/465 [27:37<05:31,  4.31s/it]
 84%|████████▎ | 389/465 [27:41<05:33,  4.39s/it]
 84%|████████▍ | 390/465 [27:45<05:27,  4.37s/it]
                                                 
{'loss': '0.3553', 'grad_norm': '0.5276', 'learning_rate': '5e-05', 'epoch': '12.59'}
+
 84%|████████▍ | 390/465 [27:45<05:27,  4.37s/it]
 84%|████████▍ | 391/465 [27:50<05:21,  4.35s/it]
 84%|████████▍ | 392/465 [27:54<05:19,  4.37s/it]
 85%|████████▍ | 393/465 [27:59<05:13,  4.36s/it]
 85%|████████▍ | 394/465 [28:03<05:10,  4.37s/it]
 85%|████████▍ | 395/465 [28:07<05:00,  4.30s/it]
 85%|████████▌ | 396/465 [28:12<05:06,  4.45s/it]
 85%|████████▌ | 397/465 [28:16<04:56,  4.35s/it]
 86%|████████▌ | 398/465 [28:20<04:47,  4.29s/it]
 86%|████████▌ | 399/465 [28:24<04:43,  4.29s/it]
 86%|████████▌ | 400/465 [28:29<04:35,  4.24s/it]
                                                 
{'loss': '0.3386', 'grad_norm': '2.45', 'learning_rate': '5e-05', 'epoch': '12.92'}
+
 86%|████████▌ | 400/465 [28:29<04:35,  4.24s/it]
 86%|████████▌ | 401/465 [28:33<04:34,  4.28s/it]
 86%|████████▋ | 402/465 [28:37<04:30,  4.29s/it]
 87%|████████▋ | 403/465 [28:39<03:42,  3.59s/it]
 87%|████████▋ | 404/465 [28:44<03:51,  3.80s/it]
 87%|████████▋ | 405/465 [28:48<03:54,  3.91s/it]
 87%|████████▋ | 406/465 [28:52<03:57,  4.03s/it]
 88%|████████▊ | 407/465 [28:56<03:59,  4.12s/it]
 88%|████████▊ | 408/465 [29:00<03:55,  4.13s/it]
 88%|████████▊ | 409/465 [29:05<03:57,  4.24s/it]
 88%|████████▊ | 410/465 [29:10<03:58,  4.34s/it]
                                                 
{'loss': '0.3919', 'grad_norm': '2.163', 'learning_rate': '5e-05', 'epoch': '13.23'}
+
 88%|████████▊ | 410/465 [29:10<03:58,  4.34s/it]
 88%|████████▊ | 411/465 [29:14<03:55,  4.36s/it]
 89%|████████▊ | 412/465 [29:19<03:54,  4.43s/it]
 89%|████████▉ | 413/465 [29:23<03:45,  4.35s/it]
 89%|████████▉ | 414/465 [29:27<03:42,  4.36s/it]
 89%|████████▉ | 415/465 [29:31<03:37,  4.34s/it]
 89%|████████▉ | 416/465 [29:36<03:33,  4.36s/it]
 90%|████████▉ | 417/465 [29:40<03:26,  4.30s/it]
 90%|████████▉ | 418/465 [29:44<03:22,  4.30s/it]
 90%|█████████ | 419/465 [29:49<03:20,  4.36s/it]
 90%|█████████ | 420/465 [29:53<03:16,  4.36s/it]
                                                 
{'loss': '0.3982', 'grad_norm': '0.6052', 'learning_rate': '5e-05', 'epoch': '13.56'}
+
 90%|█████████ | 420/465 [29:53<03:16,  4.36s/it]
 91%|█████████ | 421/465 [29:58<03:14,  4.42s/it]
 91%|█████████ | 422/465 [30:02<03:06,  4.33s/it]
 91%|█████████ | 423/465 [30:06<02:59,  4.28s/it]
 91%|█████████ | 424/465 [30:11<02:59,  4.39s/it]
 91%|█████████▏| 425/465 [30:15<02:52,  4.31s/it]
 92%|█████████▏| 426/465 [30:19<02:47,  4.30s/it]
 92%|█████████▏| 427/465 [30:23<02:41,  4.25s/it]
 92%|█████████▏| 428/465 [30:27<02:36,  4.22s/it]
 92%|█████████▏| 429/465 [30:32<02:33,  4.27s/it]
 92%|█████████▏| 430/465 [30:36<02:28,  4.23s/it]
                                                 
{'loss': '0.4013', 'grad_norm': '0.59', 'learning_rate': '5e-05', 'epoch': '13.89'}
+
 92%|█████████▏| 430/465 [30:36<02:28,  4.23s/it]
 93%|█████████▎| 431/465 [30:40<02:25,  4.28s/it]
 93%|█████████▎| 432/465 [30:44<02:19,  4.24s/it]
 93%|█████████▎| 433/465 [30:49<02:14,  4.21s/it]
 93%|█████████▎| 434/465 [30:51<01:51,  3.61s/it]
 94%|█████████▎| 435/465 [30:55<01:53,  3.77s/it]
 94%|█████████▍| 436/465 [30:59<01:52,  3.88s/it]
 94%|█████████▍| 437/465 [31:03<01:53,  4.05s/it]
 94%|█████████▍| 438/465 [31:08<01:50,  4.07s/it]
 94%|█████████▍| 439/465 [31:13<01:53,  4.36s/it]
 95%|█████████▍| 440/465 [31:17<01:48,  4.36s/it]
                                                 
{'loss': '0.3593', 'grad_norm': '0.1354', 'learning_rate': '5e-05', 'epoch': '14.2'}
+
 95%|█████████▍| 440/465 [31:17<01:48,  4.36s/it]
 95%|█████████▍| 441/465 [31:21<01:45,  4.41s/it]
 95%|█████████▌| 442/465 [31:26<01:39,  4.33s/it]
 95%|█████████▌| 443/465 [31:30<01:35,  4.32s/it]
 95%|█████████▌| 444/465 [31:35<01:32,  4.41s/it]
 96%|█████████▌| 445/465 [31:39<01:27,  4.37s/it]
 96%|█████████▌| 446/465 [31:43<01:23,  4.38s/it]
 96%|█████████▌| 447/465 [31:48<01:20,  4.45s/it]
 96%|█████████▋| 448/465 [31:52<01:15,  4.45s/it]
 97%|█████████▋| 449/465 [31:57<01:11,  4.48s/it]
 97%|█████████▋| 450/465 [32:01<01:05,  4.37s/it]
                                                 
{'loss': '0.3711', 'grad_norm': '0.3518', 'learning_rate': '5e-05', 'epoch': '14.53'}
+
 97%|█████████▋| 450/465 [32:01<01:05,  4.37s/it]
 97%|█████████▋| 451/465 [32:05<01:00,  4.32s/it]
 97%|█████████▋| 452/465 [32:09<00:55,  4.26s/it]
 97%|█████████▋| 453/465 [32:13<00:50,  4.24s/it]
 98%|█████████▊| 454/465 [32:18<00:47,  4.32s/it]
 98%|█████████▊| 455/465 [32:22<00:42,  4.27s/it]
 98%|█████████▊| 456/465 [32:26<00:38,  4.25s/it]
 98%|█████████▊| 457/465 [32:31<00:34,  4.36s/it]
 98%|█████████▊| 458/465 [32:35<00:30,  4.39s/it]
 99%|█████████▊| 459/465 [32:40<00:25,  4.31s/it]
 99%|█████████▉| 460/465 [32:44<00:21,  4.26s/it]
                                                 
{'loss': '0.3452', 'grad_norm': '0.4756', 'learning_rate': '5e-05', 'epoch': '14.85'}
+
 99%|█████████▉| 460/465 [32:44<00:21,  4.26s/it]
 99%|█████████▉| 461/465 [32:48<00:17,  4.32s/it]
 99%|█████████▉| 462/465 [32:52<00:12,  4.27s/it]
100%|█████████▉| 463/465 [32:57<00:08,  4.27s/it]
100%|█████████▉| 464/465 [33:01<00:04,  4.29s/it]
100%|██████████| 465/465 [33:03<00:00,  3.59s/it]
+
Eval @ step 465:   0%|          | 0/100 [00:00<?, ?it/s]
+
Eval @ step 465:   2%|▏         | 2/100 [00:00<00:08, 11.46it/s]
+
Eval @ step 465:   4%|▍         | 4/100 [00:00<00:08, 11.55it/s]
+
Eval @ step 465:   6%|▌         | 6/100 [00:00<00:08, 11.58it/s]
+
Eval @ step 465:   8%|▊         | 8/100 [00:00<00:07, 11.57it/s]
+
Eval @ step 465:  10%|█         | 10/100 [00:00<00:07, 11.64it/s]
+
Eval @ step 465:  12%|█▏        | 12/100 [00:01<00:07, 11.69it/s]
+
Eval @ step 465:  14%|█▍        | 14/100 [00:01<00:07, 11.75it/s]
+
Eval @ step 465:  16%|█▌        | 16/100 [00:01<00:07, 11.81it/s]
+
Eval @ step 465:  18%|█▊        | 18/100 [00:01<00:06, 11.76it/s]
+
Eval @ step 465:  20%|██        | 20/100 [00:01<00:06, 11.79it/s]
+
Eval @ step 465:  22%|██▏       | 22/100 [00:01<00:06, 11.75it/s]
+
Eval @ step 465:  24%|██▍       | 24/100 [00:02<00:06, 11.69it/s]
+
Eval @ step 465:  26%|██▌       | 26/100 [00:02<00:06, 11.81it/s]
+
Eval @ step 465:  28%|██▊       | 28/100 [00:02<00:06, 11.95it/s]
+
Eval @ step 465:  30%|███       | 30/100 [00:02<00:05, 12.02it/s]
+
Eval @ step 465:  32%|███▏      | 32/100 [00:02<00:05, 12.02it/s]
+
Eval @ step 465:  34%|███▍      | 34/100 [00:02<00:05, 12.06it/s]
+
Eval @ step 465:  36%|███▌      | 36/100 [00:03<00:05, 12.09it/s]
+
Eval @ step 465:  38%|███▊      | 38/100 [00:03<00:05, 12.06it/s]
+
Eval @ step 465:  40%|████      | 40/100 [00:03<00:04, 12.03it/s]
+
Eval @ step 465:  42%|████▏     | 42/100 [00:03<00:04, 12.04it/s]
+
Eval @ step 465:  44%|████▍     | 44/100 [00:03<00:04, 12.02it/s]
+
Eval @ step 465:  46%|████▌     | 46/100 [00:03<00:04, 11.96it/s]
+
Eval @ step 465:  48%|████▊     | 48/100 [00:04<00:04, 12.00it/s]
+
Eval @ step 465:  50%|█████     | 50/100 [00:04<00:04, 12.00it/s]
+
Eval @ step 465:  52%|█████▏    | 52/100 [00:04<00:03, 12.04it/s]
+
Eval @ step 465:  54%|█████▍    | 54/100 [00:04<00:03, 12.07it/s]
+
Eval @ step 465:  56%|█████▌    | 56/100 [00:04<00:03, 12.06it/s]
+
Eval @ step 465:  58%|█████▊    | 58/100 [00:04<00:03, 12.03it/s]
+
Eval @ step 465:  60%|██████    | 60/100 [00:05<00:03, 12.10it/s]
+
Eval @ step 465:  62%|██████▏   | 62/100 [00:05<00:03, 12.08it/s]
+
Eval @ step 465:  64%|██████▍   | 64/100 [00:05<00:02, 12.10it/s]
+
Eval @ step 465:  66%|██████▌   | 66/100 [00:05<00:02, 12.07it/s]
+
Eval @ step 465:  68%|██████▊   | 68/100 [00:05<00:02, 12.10it/s]
+
Eval @ step 465:  70%|███████   | 70/100 [00:05<00:02, 12.14it/s]
+
Eval @ step 465:  72%|███████▏  | 72/100 [00:06<00:02, 12.16it/s]
+
Eval @ step 465:  74%|███████▍  | 74/100 [00:06<00:02, 12.19it/s]
+
Eval @ step 465:  76%|███████▌  | 76/100 [00:06<00:01, 12.13it/s]
+
Eval @ step 465:  78%|███████▊  | 78/100 [00:06<00:01, 12.13it/s]
+
Eval @ step 465:  80%|████████  | 80/100 [00:06<00:01, 12.10it/s]
+
Eval @ step 465:  82%|████████▏ | 82/100 [00:06<00:01, 12.07it/s]
+
Eval @ step 465:  84%|████████▍ | 84/100 [00:07<00:01, 12.03it/s]
+
Eval @ step 465:  86%|████████▌ | 86/100 [00:07<00:01, 12.06it/s]
+
Eval @ step 465:  88%|████████▊ | 88/100 [00:07<00:00, 12.06it/s]
+
Eval @ step 465:  90%|█████████ | 90/100 [00:07<00:00, 12.01it/s]
+
Eval @ step 465:  92%|█████████▏| 92/100 [00:07<00:00, 12.03it/s]
+
Eval @ step 465:  94%|█████████▍| 94/100 [00:07<00:00, 12.06it/s]
+
Eval @ step 465:  96%|█████████▌| 96/100 [00:08<00:00, 12.12it/s]
+
Eval @ step 465:  98%|█████████▊| 98/100 [00:08<00:00, 12.10it/s]
+
Eval @ step 465: 100%|██████████| 100/100 [00:08<00:00, 12.02it/s]
+
                                                                    [EVAL] ★ New best eval_rougeL=64.0000 (step=465) — snapshot saved in memory
+
                                                 
{'eval_rougeL': '64', 'epoch': '15'}
+
100%|██████████| 465/465 [33:11<00:00,  3.59s/it]
                                                 
{'train_runtime': '1992', 'train_samples_per_second': '7.343', 'train_steps_per_second': '0.233', 'train_loss': '0.598', 'epoch': '15'}
+
100%|██████████| 465/465 [33:11<00:00,  3.59s/it]
100%|██████████| 465/465 [33:11<00:00,  4.28s/it]
+***** train metrics *****
+  epoch                    =       15.0
+  total_flos               = 87743930GF
+  train_loss               =      0.598
+  train_runtime            = 0:33:11.71
+  train_samples            =        975
+  train_samples_per_second =      7.343
+  train_steps_per_second   =      0.233
+  [EVAL] Restored best model (eval_rougeL=64.0000, 128 tensors)
+[SAVE] LoRA weights → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights
+[SRT] Loading frozen backbone from /home/dev-52/minhnguyen/Continual/Llama …
+
Loading weights:   0%|          | 0/291 [00:00<?, ?it/s]
Loading weights:   2%|▏         | 6/291 [00:00<00:05, 56.32it/s]
Loading weights:   9%|▉         | 27/291 [00:00<00:01, 139.13it/s]
Loading weights:  17%|█▋        | 49/291 [00:00<00:01, 157.46it/s]
Loading weights:  24%|██▍       | 70/291 [00:00<00:01, 175.85it/s]
Loading weights:  31%|███▏      | 91/291 [00:00<00:01, 184.40it/s]
Loading weights:  38%|███▊      | 112/291 [00:00<00:01, 174.99it/s]
Loading weights:  45%|████▌     | 132/291 [00:00<00:00, 179.24it/s]
Loading weights:  53%|█████▎    | 154/291 [00:00<00:00, 187.82it/s]
Loading weights:  59%|█████▉    | 173/291 [00:00<00:00, 185.23it/s]
Loading weights:  66%|██████▌   | 192/291 [00:01<00:00, 179.15it/s]
Loading weights:  72%|███████▏  | 210/291 [00:01<00:00, 177.46it/s]
Loading weights:  79%|███████▊  | 229/291 [00:01<00:00, 172.29it/s]
Loading weights:  85%|████████▍ | 247/291 [00:01<00:00, 172.24it/s]
Loading weights:  91%|█████████▏| 266/291 [00:01<00:00, 175.87it/s]
Loading weights:  98%|█████████▊| 284/291 [00:01<00:00, 176.91it/s]
Loading weights: 100%|██████████| 291/291 [00:01<00:00, 177.13it/s]
+[SRT] Frozen backbone attached (moved to cuda:0).
+  [SRT] → Forward extraction (975 samples, batch_size=1)
+  [SRT] Task 'task073_commonsenseqa_answer_generation': PaR=65.7, shrinkage=ridge, n=975
+  [SRT] Wired router: 9 sigs, cur=task073_commonsenseqa_answer_generation, mapping={'task073_commonsenseqa_answer_generation': 0, 'task1572_samsum_summary': 8, 'task363_sst2_polarity_classification': 7, 'task1290_xsum_summarization': 6, 'task181_outcome_extraction': 5, 'task002_quoref_answer_generation': 4, 'task1510_evalution_relation_extraction': 3, 'task639_multi_woz_user_utterance_generation': 2, 'task1729_personachat_generate_next': 1}
+[SRT] Router: 9 tasks, avg_PaR=65.7, metrics=['ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge', 'ridge']
+  [SRT] Saved signatures → /home/dev-52/minhnguyen/Continual/new_llama_gainlora/logs_and_outputs/superni_order1_llama_srt/outputs/9-task073_commonsenseqa_answer_generation/saved_weights/srt_signatures.npz
+*** Prediction ***
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1572_samsum_summary/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task363_sst2_polarity_classification/test.json
+['URL', 'Categories', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Definition']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1290_xsum_summarization/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task181_outcome_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task002_quoref_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1510_evalution_relation_extraction/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task639_multi_woz_user_utterance_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task1729_personachat_generate_next/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+/home/dev-52/minhnguyen/Continual/root_gainlora/CL_Benchmark/SuperNI/task073_commonsenseqa_answer_generation/test.json
+['Contributors', 'Source', 'URL', 'Categories', 'Reasoning', 'Definition', 'Input_language', 'Output_language', 'Instruction_language', 'Domains', 'Positive Examples', 'Negative Examples', 'Instances', 'Instance License']
+  [SRT-INIT] router tasks=['task1572_samsum_summary', 'task363_sst2_polarity_classification', 'task1290_xsum_summarization', 'task181_outcome_extraction', 'task002_quoref_answer_generation', 'task1510_evalution_relation_extraction', 'task639_multi_woz_user_utterance_generation', 'task1729_personachat_generate_next', 'task073_commonsenseqa_answer_generation']
+  [SRT-INIT] task_id_to_idx={'task073_commonsenseqa_answer_generation': 0, 'task1572_samsum_summary': 8, 'task363_sst2_polarity_classification': 7, 'task1290_xsum_summarization': 6, 'task181_outcome_extraction': 5, 'task002_quoref_answer_generation': 4, 'task1510_evalution_relation_extraction': 3, 'task639_multi_woz_user_utterance_generation': 2, 'task1729_personachat_generate_next': 1}
+  [SRT-INIT] use_srt_routing=True
+  [SRT-INIT] encoder_frozen=FrozenLlamaExtractor
+
Continual predict:   0%|          | 0/681 [00:00<?, ?it/s]  [SRT-GEN] use_srt=True  router=OK  encoder=OK  n_tasks=9
+
Continual predict:   0%|          | 1/681 [00:01<20:23,  1.80s/it]
Continual predict:   0%|          | 2/681 [00:04<25:23,  2.24s/it]
Continual predict:   0%|          | 3/681 [00:05<18:14,  1.61s/it]
Continual predict:   1%|          | 4/681 [00:05<13:32,  1.20s/it]
Continual predict:   1%|          | 5/681 [00:07<16:12,  1.44s/it]
Continual predict:   1%|          | 6/681 [00:12<28:47,  2.56s/it]
Continual predict:   1%|          | 7/681 [00:15<31:41,  2.82s/it]
Continual predict:   1%|          | 8/681 [00:17<29:05,  2.59s/it]
Continual predict:   1%|▏         | 9/681 [00:19<26:06,  2.33s/it]
Continual predict:   1%|▏         | 10/681 [00:20<20:50,  1.86s/it]
Continual predict:   2%|▏         | 11/681 [00:23<23:42,  2.12s/it]
Continual predict:   2%|▏         | 12/681 [00:26<27:20,  2.45s/it]
Continual predict:   2%|▏         | 13/681 [00:27<23:12,  2.09s/it]
Continual predict:   2%|▏         | 14/681 [00:30<24:23,  2.19s/it]
Continual predict:   2%|▏         | 15/681 [00:32<24:35,  2.22s/it]
Continual predict:   2%|▏         | 16/681 [00:34<25:50,  2.33s/it]
Continual predict:   2%|▏         | 17/681 [00:42<43:59,  3.98s/it]
Continual predict:   3%|▎         | 18/681 [00:44<37:47,  3.42s/it]
Continual predict:   3%|▎         | 19/681 [00:47<35:10,  3.19s/it]
Continual predict:   3%|▎         | 20/681 [00:48<29:14,  2.65s/it]
Continual predict:   3%|▎         | 21/681 [00:49<20:59,  1.91s/it]
Continual predict:   3%|▎         | 22/681 [00:49<15:11,  1.38s/it]
Continual predict:   3%|▎         | 23/681 [00:49<11:10,  1.02s/it]
Continual predict:   4%|▎         | 24/681 [00:49<08:18,  1.32it/s]
Continual predict:   4%|▎         | 25/681 [00:49<06:19,  1.73it/s]
Continual predict:   4%|▍         | 26/681 [00:49<04:55,  2.21it/s]
Continual predict:   4%|▍         | 27/681 [00:49<03:57,  2.75it/s]
Continual predict:   4%|▍         | 28/681 [00:50<03:18,  3.29it/s]
Continual predict:   4%|▍         | 29/681 [00:50<02:48,  3.86it/s]
Continual predict:   4%|▍         | 30/681 [00:50<02:30,  4.32it/s]
Continual predict:   5%|▍         | 31/681 [00:50<02:15,  4.79it/s]
Continual predict:   5%|▍         | 32/681 [00:50<02:05,  5.17it/s]
Continual predict:   5%|▍         | 33/681 [00:50<02:00,  5.37it/s]
Continual predict:   5%|▍         | 34/681 [00:51<01:56,  5.53it/s]
Continual predict:   5%|▌         | 35/681 [00:51<01:54,  5.66it/s]
Continual predict:   5%|▌         | 36/681 [00:51<01:49,  5.88it/s]
Continual predict:   5%|▌         | 37/681 [00:51<01:49,  5.91it/s]
Continual predict:   6%|▌         | 38/681 [00:51<01:46,  6.05it/s]
Continual predict:   6%|▌         | 39/681 [00:51<01:44,  6.16it/s]
Continual predict:   6%|▌         | 40/681 [00:52<01:42,  6.25it/s]
Continual predict:   6%|▌         | 41/681 [00:52<01:43,  6.17it/s]
Continual predict:   6%|▌         | 42/681 [00:52<01:42,  6.24it/s]
Continual predict:   6%|▋         | 43/681 [00:52<01:43,  6.17it/s]
Continual predict:   6%|▋         | 44/681 [00:52<01:43,  6.13it/s]
Continual predict:   7%|▋         | 45/681 [00:52<01:42,  6.21it/s]
Continual predict:   7%|▋         | 46/681 [00:53<01:41,  6.27it/s]
Continual predict:   7%|▋         | 47/681 [00:53<01:41,  6.27it/s]
Continual predict:   7%|▋         | 48/681 [00:53<01:42,  6.16it/s]
Continual predict:   7%|▋         | 49/681 [00:53<01:43,  6.12it/s]
Continual predict:   7%|▋         | 50/681 [00:53<01:41,  6.19it/s]
Continual predict:   7%|▋         | 51/681 [00:53<01:41,  6.22it/s]
Continual predict:   8%|▊         | 52/681 [00:54<01:40,  6.27it/s]
Continual predict:   8%|▊         | 53/681 [00:54<01:39,  6.31it/s]
Continual predict:   8%|▊         | 54/681 [00:54<01:41,  6.21it/s]
Continual predict:   8%|▊         | 55/681 [00:54<01:39,  6.28it/s]
Continual predict:   8%|▊         | 56/681 [00:54<01:41,  6.18it/s]
Continual predict:   8%|▊         | 57/681 [00:54<01:39,  6.25it/s]
Continual predict:   9%|▊         | 58/681 [00:54<01:38,  6.30it/s]
Continual predict:   9%|▊         | 59/681 [00:55<01:38,  6.33it/s]
Continual predict:   9%|▉         | 60/681 [00:55<01:40,  6.20it/s]
Continual predict:   9%|▉         | 61/681 [00:55<01:38,  6.27it/s]
Continual predict:   9%|▉         | 62/681 [00:55<01:37,  6.32it/s]
Continual predict:   9%|▉         | 63/681 [00:55<01:39,  6.23it/s]
Continual predict:   9%|▉         | 64/681 [00:55<01:38,  6.28it/s]
Continual predict:  10%|▉         | 65/681 [00:56<01:37,  6.31it/s]
Continual predict:  10%|▉         | 66/681 [00:56<01:36,  6.34it/s]
Continual predict:  10%|▉         | 67/681 [00:56<01:36,  6.37it/s]
Continual predict:  10%|▉         | 68/681 [00:56<01:35,  6.40it/s]
Continual predict:  10%|█         | 69/681 [00:56<01:35,  6.41it/s]
Continual predict:  10%|█         | 70/681 [00:56<01:34,  6.43it/s]
Continual predict:  10%|█         | 71/681 [00:57<01:34,  6.43it/s]
Continual predict:  11%|█         | 72/681 [00:57<01:37,  6.28it/s]
Continual predict:  11%|█         | 73/681 [00:57<01:37,  6.21it/s]
Continual predict:  11%|█         | 74/681 [00:57<01:42,  5.94it/s]
Continual predict:  11%|█         | 75/681 [00:57<01:43,  5.86it/s]
Continual predict:  11%|█         | 76/681 [00:57<01:40,  6.02it/s]
Continual predict:  11%|█▏        | 77/681 [00:58<01:38,  6.13it/s]
Continual predict:  11%|█▏        | 78/681 [00:58<01:36,  6.22it/s]
Continual predict:  12%|█▏        | 79/681 [00:58<01:36,  6.26it/s]
Continual predict:  12%|█▏        | 80/681 [00:58<01:35,  6.29it/s]
Continual predict:  12%|█▏        | 81/681 [00:58<01:34,  6.32it/s]
Continual predict:  12%|█▏        | 82/681 [00:58<01:34,  6.35it/s]
Continual predict:  12%|█▏        | 83/681 [00:58<01:33,  6.37it/s]
Continual predict:  12%|█▏        | 84/681 [00:59<01:35,  6.24it/s]
Continual predict:  12%|█▏        | 85/681 [00:59<01:34,  6.29it/s]
Continual predict:  13%|█▎        | 86/681 [00:59<01:34,  6.30it/s]
Continual predict:  13%|█▎        | 87/681 [00:59<01:34,  6.31it/s]
Continual predict:  13%|█▎        | 88/681 [00:59<01:33,  6.36it/s]
Continual predict:  13%|█▎        | 89/681 [00:59<01:32,  6.38it/s]
Continual predict:  13%|█▎        | 90/681 [01:00<01:32,  6.40it/s]
Continual predict:  13%|█▎        | 91/681 [01:00<01:32,  6.41it/s]
Continual predict:  14%|█▎        | 92/681 [01:00<01:32,  6.39it/s]
Continual predict:  14%|█▎        | 93/681 [01:00<01:33,  6.26it/s]
Continual predict:  14%|█▍        | 94/681 [01:00<01:34,  6.19it/s]
Continual predict:  14%|█▍        | 95/681 [01:00<01:33,  6.25it/s]
Continual predict:  14%|█▍        | 96/681 [01:01<01:32,  6.29it/s]
Continual predict:  14%|█▍        | 97/681 [01:01<01:32,  6.33it/s]
Continual predict:  14%|█▍        | 98/681 [01:01<01:33,  6.23it/s]
Continual predict:  15%|█▍        | 99/681 [01:01<01:34,  6.16it/s]
Continual predict:  15%|█▍        | 100/681 [01:01<01:35,  6.11it/s]
Continual predict:  15%|█▍        | 101/681 [01:01<01:33,  6.20it/s]
Continual predict:  15%|█▍        | 102/681 [01:01<01:32,  6.27it/s]
Continual predict:  15%|█▌        | 103/681 [01:02<01:31,  6.32it/s]
Continual predict:  15%|█▌        | 104/681 [01:02<01:30,  6.35it/s]
Continual predict:  15%|█▌        | 105/681 [01:02<01:32,  6.25it/s]
Continual predict:  16%|█▌        | 106/681 [01:02<01:31,  6.29it/s]
Continual predict:  16%|█▌        | 107/681 [01:02<01:30,  6.35it/s]
Continual predict:  16%|█▌        | 108/681 [01:02<01:29,  6.38it/s]
Continual predict:  16%|█▌        | 109/681 [01:03<01:29,  6.40it/s]
Continual predict:  16%|█▌        | 110/681 [01:03<01:29,  6.40it/s]
Continual predict:  16%|█▋        | 111/681 [01:03<01:28,  6.41it/s]
Continual predict:  16%|█▋        | 112/681 [01:03<01:28,  6.40it/s]
Continual predict:  17%|█▋        | 113/681 [01:03<01:28,  6.39it/s]
Continual predict:  17%|█▋        | 114/681 [01:03<01:28,  6.40it/s]
Continual predict:  17%|█▋        | 115/681 [01:04<01:30,  6.28it/s]
Continual predict:  17%|█▋        | 116/681 [01:04<01:29,  6.32it/s]
Continual predict:  17%|█▋        | 117/681 [01:04<01:30,  6.22it/s]
Continual predict:  17%|█▋        | 118/681 [01:04<01:31,  6.15it/s]
Continual predict:  17%|█▋        | 119/681 [01:04<01:30,  6.21it/s]
Continual predict:  18%|█▊        | 120/681 [01:04<01:29,  6.26it/s]
Continual predict:  18%|█▊        | 121/681 [01:09<14:05,  1.51s/it]
Continual predict:  18%|█▊        | 122/681 [01:13<21:53,  2.35s/it]
Continual predict:  18%|█▊        | 123/681 [01:19<30:25,  3.27s/it]
Continual predict:  18%|█▊        | 124/681 [01:21<28:02,  3.02s/it]
Continual predict:  18%|█▊        | 125/681 [01:26<33:12,  3.58s/it]
Continual predict:  19%|█▊        | 126/681 [01:31<36:46,  3.98s/it]
Continual predict:  19%|█▊        | 127/681 [01:33<29:59,  3.25s/it]
Continual predict:  19%|█▉        | 128/681 [01:35<28:47,  3.12s/it]
Continual predict:  19%|█▉        | 129/681 [01:39<29:42,  3.23s/it]
Continual predict:  19%|█▉        | 130/681 [01:42<29:26,  3.21s/it]
Continual predict:  19%|█▉        | 131/681 [01:44<26:56,  2.94s/it]
Continual predict:  19%|█▉        | 132/681 [01:47<26:24,  2.89s/it]
Continual predict:  20%|█▉        | 133/681 [01:50<27:29,  3.01s/it]
Continual predict:  20%|█▉        | 134/681 [01:57<36:06,  3.96s/it]
Continual predict:  20%|█▉        | 135/681 [02:01<38:40,  4.25s/it]
Continual predict:  20%|█▉        | 136/681 [02:05<36:10,  3.98s/it]
Continual predict:  20%|██        | 137/681 [02:07<31:33,  3.48s/it]
Continual predict:  20%|██        | 138/681 [02:08<24:10,  2.67s/it]
Continual predict:  20%|██        | 139/681 [02:09<20:11,  2.24s/it]
Continual predict:  21%|██        | 140/681 [02:12<21:36,  2.40s/it]
Continual predict:  21%|██        | 141/681 [02:19<33:12,  3.69s/it]
Continual predict:  21%|██        | 142/681 [02:23<35:47,  3.98s/it]
Continual predict:  21%|██        | 143/681 [02:27<34:02,  3.80s/it]
Continual predict:  21%|██        | 144/681 [02:36<49:36,  5.54s/it]
Continual predict:  21%|██▏       | 145/681 [02:41<46:56,  5.26s/it]
Continual predict:  21%|██▏       | 146/681 [02:43<37:49,  4.24s/it]
Continual predict:  22%|██▏       | 147/681 [02:47<38:24,  4.31s/it]
Continual predict:  22%|██▏       | 148/681 [02:49<32:47,  3.69s/it]
Continual predict:  22%|██▏       | 149/681 [02:53<31:35,  3.56s/it]
Continual predict:  22%|██▏       | 150/681 [02:55<27:41,  3.13s/it]
Continual predict:  22%|██▏       | 151/681 [02:58<27:36,  3.13s/it]
Continual predict:  22%|██▏       | 152/681 [03:04<34:39,  3.93s/it]
Continual predict:  22%|██▏       | 153/681 [03:06<30:57,  3.52s/it]
Continual predict:  23%|██▎       | 154/681 [03:11<34:35,  3.94s/it]
Continual predict:  23%|██▎       | 155/681 [03:15<33:52,  3.86s/it]
Continual predict:  23%|██▎       | 156/681 [03:18<32:49,  3.75s/it]
Continual predict:  23%|██▎       | 157/681 [03:22<33:19,  3.82s/it]
Continual predict:  23%|██▎       | 158/681 [03:23<25:54,  2.97s/it]
Continual predict:  23%|██▎       | 159/681 [03:30<36:40,  4.22s/it]
Continual predict:  23%|██▎       | 160/681 [03:33<31:44,  3.66s/it]
Continual predict:  24%|██▎       | 161/681 [03:35<28:50,  3.33s/it]
Continual predict:  24%|██▍       | 162/681 [03:40<32:53,  3.80s/it]
Continual predict:  24%|██▍       | 163/681 [03:42<28:29,  3.30s/it]
Continual predict:  24%|██▍       | 164/681 [03:45<25:57,  3.01s/it]
Continual predict:  24%|██▍       | 165/681 [03:48<26:48,  3.12s/it]
Continual predict:  24%|██▍       | 166/681 [03:51<24:51,  2.90s/it]
Continual predict:  25%|██▍       | 167/681 [03:52<20:48,  2.43s/it]
Continual predict:  25%|██▍       | 168/681 [03:54<20:28,  2.39s/it]
Continual predict:  25%|██▍       | 169/681 [03:56<18:50,  2.21s/it]
Continual predict:  25%|██▍       | 170/681 [03:59<21:10,  2.49s/it]
Continual predict:  25%|██▌       | 171/681 [04:01<18:51,  2.22s/it]
Continual predict:  25%|██▌       | 172/681 [04:03<19:57,  2.35s/it]
Continual predict:  25%|██▌       | 173/681 [04:07<24:08,  2.85s/it]
Continual predict:  26%|██▌       | 174/681 [04:09<19:55,  2.36s/it]
Continual predict:  26%|██▌       | 175/681 [04:12<22:58,  2.72s/it]
Continual predict:  26%|██▌       | 176/681 [04:15<23:30,  2.79s/it]
Continual predict:  26%|██▌       | 177/681 [04:18<23:41,  2.82s/it]
Continual predict:  26%|██▌       | 178/681 [04:19<19:50,  2.37s/it]
Continual predict:  26%|██▋       | 179/681 [04:20<16:46,  2.01s/it]
Continual predict:  26%|██▋       | 180/681 [04:22<15:03,  1.80s/it]
Continual predict:  27%|██▋       | 181/681 [04:23<13:51,  1.66s/it]
Continual predict:  27%|██▋       | 182/681 [04:28<21:40,  2.61s/it]
Continual predict:  27%|██▋       | 183/681 [04:28<15:41,  1.89s/it]
Continual predict:  27%|██▋       | 184/681 [04:29<13:44,  1.66s/it]
Continual predict:  27%|██▋       | 185/681 [04:33<18:12,  2.20s/it]
Continual predict:  27%|██▋       | 186/681 [04:34<15:16,  1.85s/it]
Continual predict:  27%|██▋       | 187/681 [04:37<18:53,  2.29s/it]
Continual predict:  28%|██▊       | 188/681 [04:40<20:01,  2.44s/it]
Continual predict:  28%|██▊       | 189/681 [04:45<26:33,  3.24s/it]
Continual predict:  28%|██▊       | 190/681 [04:48<26:50,  3.28s/it]
Continual predict:  28%|██▊       | 191/681 [04:50<22:04,  2.70s/it]
Continual predict:  28%|██▊       | 192/681 [04:53<23:45,  2.91s/it]
Continual predict:  28%|██▊       | 193/681 [04:58<27:55,  3.43s/it]
Continual predict:  28%|██▊       | 194/681 [05:03<31:31,  3.88s/it]
Continual predict:  29%|██▊       | 195/681 [05:08<35:11,  4.34s/it]
Continual predict:  29%|██▉       | 196/681 [05:12<32:52,  4.07s/it]
Continual predict:  29%|██▉       | 197/681 [05:15<31:05,  3.85s/it]
Continual predict:  29%|██▉       | 198/681 [05:22<39:52,  4.95s/it]
Continual predict:  29%|██▉       | 199/681 [05:24<30:31,  3.80s/it]
Continual predict:  29%|██▉       | 200/681 [05:25<24:55,  3.11s/it]
Continual predict:  30%|██▉       | 201/681 [05:28<23:40,  2.96s/it]
Continual predict:  30%|██▉       | 202/681 [05:31<24:31,  3.07s/it]
Continual predict:  30%|██▉       | 203/681 [05:36<29:38,  3.72s/it]
Continual predict:  30%|██▉       | 204/681 [05:39<28:07,  3.54s/it]
Continual predict:  30%|███       | 205/681 [05:46<35:36,  4.49s/it]
Continual predict:  30%|███       | 206/681 [05:49<31:31,  3.98s/it]
Continual predict:  30%|███       | 207/681 [05:53<31:01,  3.93s/it]
Continual predict:  31%|███       | 208/681 [05:59<36:29,  4.63s/it]
Continual predict:  31%|███       | 209/681 [06:02<31:58,  4.06s/it]
Continual predict:  31%|███       | 210/681 [06:03<26:18,  3.35s/it]
Continual predict:  31%|███       | 211/681 [06:06<24:38,  3.15s/it]
Continual predict:  31%|███       | 212/681 [06:09<24:41,  3.16s/it]
Continual predict:  31%|███▏      | 213/681 [06:13<26:13,  3.36s/it]
Continual predict:  31%|███▏      | 214/681 [06:17<27:07,  3.49s/it]
Continual predict:  32%|███▏      | 215/681 [06:18<20:48,  2.68s/it]
Continual predict:  32%|███▏      | 216/681 [06:21<21:51,  2.82s/it]
Continual predict:  32%|███▏      | 217/681 [06:23<20:08,  2.60s/it]
Continual predict:  32%|███▏      | 218/681 [06:27<23:33,  3.05s/it]
Continual predict:  32%|███▏      | 219/681 [06:30<24:05,  3.13s/it]
Continual predict:  32%|███▏      | 220/681 [06:34<26:28,  3.45s/it]
Continual predict:  32%|███▏      | 221/681 [06:35<19:00,  2.48s/it]
Continual predict:  33%|███▎      | 222/681 [06:35<13:56,  1.82s/it]
Continual predict:  33%|███▎      | 223/681 [06:35<10:13,  1.34s/it]
Continual predict:  33%|███▎      | 224/681 [06:35<07:38,  1.00s/it]
Continual predict:  33%|███▎      | 225/681 [06:36<05:49,  1.30it/s]
Continual predict:  33%|███▎      | 226/681 [06:36<04:33,  1.66it/s]
Continual predict:  33%|███▎      | 227/681 [06:36<03:47,  1.99it/s]
Continual predict:  33%|███▎      | 228/681 [06:36<03:15,  2.32it/s]
Continual predict:  34%|███▎      | 229/681 [06:37<02:48,  2.69it/s]
Continual predict:  34%|███▍      | 230/681 [06:37<02:55,  2.57it/s]
Continual predict:  34%|███▍      | 231/681 [06:37<02:31,  2.96it/s]
Continual predict:  34%|███▍      | 232/681 [06:37<02:15,  3.32it/s]
Continual predict:  34%|███▍      | 233/681 [06:38<02:03,  3.64it/s]
Continual predict:  34%|███▍      | 234/681 [06:38<01:55,  3.89it/s]
Continual predict:  35%|███▍      | 235/681 [06:38<01:49,  4.09it/s]
Continual predict:  35%|███▍      | 236/681 [06:38<01:44,  4.25it/s]
Continual predict:  35%|███▍      | 237/681 [06:38<01:41,  4.37it/s]
Continual predict:  35%|███▍      | 238/681 [06:39<01:39,  4.45it/s]
Continual predict:  35%|███▌      | 239/681 [06:39<01:37,  4.51it/s]
Continual predict:  35%|███▌      | 240/681 [06:39<01:55,  3.81it/s]
Continual predict:  35%|███▌      | 241/681 [06:39<01:49,  4.02it/s]
Continual predict:  36%|███▌      | 242/681 [06:40<01:44,  4.20it/s]
Continual predict:  36%|███▌      | 243/681 [06:40<01:41,  4.33it/s]
Continual predict:  36%|███▌      | 244/681 [06:40<01:38,  4.42it/s]
Continual predict:  36%|███▌      | 245/681 [06:40<01:37,  4.49it/s]
Continual predict:  36%|███▌      | 246/681 [06:41<01:35,  4.54it/s]
Continual predict:  36%|███▋      | 247/681 [06:41<01:41,  4.26it/s]
Continual predict:  36%|███▋      | 248/681 [06:41<01:38,  4.38it/s]
Continual predict:  37%|███▋      | 249/681 [06:41<01:43,  4.16it/s]
Continual predict:  37%|███▋      | 250/681 [06:42<01:40,  4.30it/s]
Continual predict:  37%|███▋      | 251/681 [06:42<01:44,  4.12it/s]
Continual predict:  37%|███▋      | 252/681 [06:43<03:56,  1.81it/s]
Continual predict:  37%|███▋      | 253/681 [06:43<03:19,  2.15it/s]
Continual predict:  37%|███▋      | 254/681 [06:44<02:46,  2.56it/s]
Continual predict:  37%|███▋      | 255/681 [06:44<02:26,  2.91it/s]
Continual predict:  38%|███▊      | 256/681 [06:44<02:09,  3.27it/s]
Continual predict:  38%|███▊      | 257/681 [06:45<03:05,  2.28it/s]
Continual predict:  38%|███▊      | 258/681 [06:45<02:37,  2.69it/s]
Continual predict:  38%|███▊      | 259/681 [06:45<02:17,  3.07it/s]
Continual predict:  38%|███▊      | 260/681 [06:45<02:03,  3.41it/s]
Continual predict:  38%|███▊      | 261/681 [06:46<02:38,  2.65it/s]
Continual predict:  38%|███▊      | 262/681 [06:46<02:20,  2.99it/s]
Continual predict:  39%|███▊      | 263/681 [06:46<02:11,  3.18it/s]
Continual predict:  39%|███▉      | 264/681 [06:48<04:07,  1.68it/s]
Continual predict:  39%|███▉      | 265/681 [06:48<04:02,  1.71it/s]
Continual predict:  39%|███▉      | 266/681 [06:49<03:58,  1.74it/s]
Continual predict:  39%|███▉      | 267/681 [06:49<04:02,  1.71it/s]
Continual predict:  39%|███▉      | 268/681 [06:50<04:33,  1.51it/s]
Continual predict:  40%|███▉      | 269/681 [06:51<04:27,  1.54it/s]
Continual predict:  40%|███▉      | 270/681 [06:52<04:25,  1.55it/s]
Continual predict:  40%|███▉      | 271/681 [06:52<04:40,  1.46it/s]
Continual predict:  40%|███▉      | 272/681 [06:53<05:08,  1.32it/s]
Continual predict:  40%|████      | 273/681 [06:54<04:55,  1.38it/s]
Continual predict:  40%|████      | 274/681 [06:55<04:47,  1.42it/s]
Continual predict:  40%|████      | 275/681 [06:55<04:35,  1.47it/s]
Continual predict:  41%|████      | 276/681 [06:56<05:24,  1.25it/s]
Continual predict:  41%|████      | 277/681 [06:58<06:23,  1.05it/s]
Continual predict:  41%|████      | 278/681 [06:58<06:09,  1.09it/s]
Continual predict:  41%|████      | 279/681 [06:59<05:44,  1.17it/s]
Continual predict:  41%|████      | 280/681 [07:00<04:56,  1.35it/s]
Continual predict:  41%|████▏     | 281/681 [07:01<05:23,  1.24it/s]
Continual predict:  41%|████▏     | 282/681 [07:01<05:24,  1.23it/s]
Continual predict:  42%|████▏     | 283/681 [07:02<05:07,  1.29it/s]
Continual predict:  42%|████▏     | 284/681 [07:03<04:30,  1.47it/s]
Continual predict:  42%|████▏     | 285/681 [07:03<04:44,  1.39it/s]
Continual predict:  42%|████▏     | 286/681 [07:04<04:50,  1.36it/s]
Continual predict:  42%|████▏     | 287/681 [07:05<04:54,  1.34it/s]
Continual predict:  42%|████▏     | 288/681 [07:06<04:42,  1.39it/s]
Continual predict:  42%|████▏     | 289/681 [07:06<05:09,  1.27it/s]
Continual predict:  43%|████▎     | 290/681 [07:07<05:06,  1.28it/s]
Continual predict:  43%|████▎     | 291/681 [07:08<04:46,  1.36it/s]
Continual predict:  43%|████▎     | 292/681 [07:08<03:57,  1.64it/s]
Continual predict:  43%|████▎     | 293/681 [07:09<04:23,  1.47it/s]
Continual predict:  43%|████▎     | 294/681 [07:10<04:07,  1.56it/s]
Continual predict:  43%|████▎     | 295/681 [07:10<03:45,  1.71it/s]
Continual predict:  43%|████▎     | 296/681 [07:11<04:31,  1.42it/s]
Continual predict:  44%|████▎     | 297/681 [07:12<04:11,  1.53it/s]
Continual predict:  44%|████▍     | 298/681 [07:12<03:48,  1.68it/s]
Continual predict:  44%|████▍     | 299/681 [07:13<04:24,  1.44it/s]
Continual predict:  44%|████▍     | 300/681 [07:13<03:42,  1.71it/s]
Continual predict:  44%|████▍     | 301/681 [07:14<04:21,  1.45it/s]
Continual predict:  44%|████▍     | 302/681 [07:15<03:51,  1.64it/s]
Continual predict:  44%|████▍     | 303/681 [07:16<05:37,  1.12it/s]
Continual predict:  45%|████▍     | 304/681 [07:17<04:32,  1.38it/s]
Continual predict:  45%|████▍     | 305/681 [07:17<04:54,  1.28it/s]
Continual predict:  45%|████▍     | 306/681 [07:18<04:58,  1.26it/s]
Continual predict:  45%|████▌     | 307/681 [07:19<04:58,  1.25it/s]
Continual predict:  45%|████▌     | 308/681 [07:19<04:17,  1.45it/s]
Continual predict:  45%|████▌     | 309/681 [07:20<03:51,  1.61it/s]
Continual predict:  46%|████▌     | 310/681 [07:21<03:50,  1.61it/s]
Continual predict:  46%|████▌     | 311/681 [07:21<03:41,  1.67it/s]
Continual predict:  46%|████▌     | 312/681 [07:22<03:54,  1.57it/s]
Continual predict:  46%|████▌     | 313/681 [07:23<04:21,  1.41it/s]
Continual predict:  46%|████▌     | 314/681 [07:23<03:54,  1.57it/s]
Continual predict:  46%|████▋     | 315/681 [07:24<04:24,  1.39it/s]
Continual predict:  46%|████▋     | 316/681 [07:25<04:33,  1.33it/s]
Continual predict:  47%|████▋     | 317/681 [07:25<03:46,  1.60it/s]
Continual predict:  47%|████▋     | 318/681 [07:26<03:28,  1.74it/s]
Continual predict:  47%|████▋     | 319/681 [07:27<05:24,  1.12it/s]
Continual predict:  47%|████▋     | 320/681 [07:28<05:04,  1.19it/s]
Continual predict:  47%|████▋     | 321/681 [07:29<04:49,  1.24it/s]
Continual predict:  47%|████▋     | 322/681 [07:29<04:08,  1.44it/s]
Continual predict:  47%|████▋     | 323/681 [07:30<04:15,  1.40it/s]
Continual predict:  48%|████▊     | 324/681 [07:31<04:35,  1.30it/s]
Continual predict:  48%|████▊     | 325/681 [07:32<04:35,  1.29it/s]
Continual predict:  48%|████▊     | 326/681 [07:32<04:20,  1.36it/s]
Continual predict:  48%|████▊     | 327/681 [07:33<05:02,  1.17it/s]
Continual predict:  48%|████▊     | 328/681 [07:35<05:25,  1.09it/s]
Continual predict:  48%|████▊     | 329/681 [07:35<04:44,  1.24it/s]
Continual predict:  48%|████▊     | 330/681 [07:35<03:53,  1.51it/s]
Continual predict:  49%|████▊     | 331/681 [07:36<03:46,  1.54it/s]
Continual predict:  49%|████▉     | 332/681 [07:37<03:35,  1.62it/s]
Continual predict:  49%|████▉     | 333/681 [07:38<04:40,  1.24it/s]
Continual predict:  49%|████▉     | 334/681 [07:39<04:52,  1.19it/s]
Continual predict:  49%|████▉     | 335/681 [07:39<04:09,  1.39it/s]
Continual predict:  49%|████▉     | 336/681 [07:40<03:41,  1.55it/s]
Continual predict:  49%|████▉     | 337/681 [07:40<03:34,  1.60it/s]
Continual predict:  50%|████▉     | 338/681 [07:41<03:37,  1.58it/s]
Continual predict:  50%|████▉     | 339/681 [07:41<03:27,  1.65it/s]
Continual predict:  50%|████▉     | 340/681 [07:43<04:16,  1.33it/s]
Continual predict:  50%|█████     | 341/681 [07:43<04:17,  1.32it/s]
Continual predict:  50%|█████     | 342/681 [07:44<03:46,  1.50it/s]
Continual predict:  50%|█████     | 343/681 [07:45<03:56,  1.43it/s]
Continual predict:  51%|█████     | 344/681 [07:45<04:22,  1.28it/s]
Continual predict:  51%|█████     | 345/681 [07:46<03:49,  1.46it/s]
Continual predict:  51%|█████     | 346/681 [07:47<04:02,  1.38it/s]
Continual predict:  51%|█████     | 347/681 [07:48<05:17,  1.05it/s]
Continual predict:  51%|█████     | 348/681 [07:49<04:16,  1.30it/s]
Continual predict:  51%|█████     | 349/681 [07:49<03:43,  1.49it/s]
Continual predict:  51%|█████▏    | 350/681 [07:50<03:40,  1.50it/s]
Continual predict:  52%|█████▏    | 351/681 [07:51<04:26,  1.24it/s]
Continual predict:  52%|█████▏    | 352/681 [07:52<04:21,  1.26it/s]
Continual predict:  52%|█████▏    | 353/681 [07:53<05:11,  1.05it/s]
Continual predict:  52%|█████▏    | 354/681 [07:54<05:42,  1.05s/it]
Continual predict:  52%|█████▏    | 355/681 [07:55<05:01,  1.08it/s]
Continual predict:  52%|█████▏    | 356/681 [07:56<05:00,  1.08it/s]
Continual predict:  52%|█████▏    | 357/681 [07:57<05:15,  1.03it/s]
Continual predict:  53%|█████▎    | 358/681 [07:58<04:54,  1.10it/s]
Continual predict:  53%|█████▎    | 359/681 [07:58<04:27,  1.20it/s]
Continual predict:  53%|█████▎    | 360/681 [08:00<05:09,  1.04it/s]
Continual predict:  53%|█████▎    | 361/681 [08:01<06:10,  1.16s/it]
Continual predict:  53%|█████▎    | 362/681 [08:02<05:16,  1.01it/s]
Continual predict:  53%|█████▎    | 363/681 [08:02<04:25,  1.20it/s]
Continual predict:  53%|█████▎    | 364/681 [08:03<03:37,  1.46it/s]
Continual predict:  54%|█████▎    | 365/681 [08:03<02:57,  1.78it/s]
Continual predict:  54%|█████▎    | 366/681 [08:03<02:30,  2.09it/s]
Continual predict:  54%|█████▍    | 367/681 [08:03<02:11,  2.39it/s]
Continual predict:  54%|█████▍    | 368/681 [08:04<01:57,  2.66it/s]
Continual predict:  54%|█████▍    | 369/681 [08:04<01:48,  2.88it/s]
Continual predict:  54%|█████▍    | 370/681 [08:04<01:41,  3.06it/s]
Continual predict:  54%|█████▍    | 371/681 [08:05<01:42,  3.04it/s]
Continual predict:  55%|█████▍    | 372/681 [08:05<01:47,  2.88it/s]
Continual predict:  55%|█████▍    | 373/681 [08:05<01:51,  2.77it/s]
Continual predict:  55%|█████▍    | 374/681 [08:06<01:43,  2.98it/s]
Continual predict:  55%|█████▌    | 375/681 [08:06<01:47,  2.84it/s]
Continual predict:  55%|█████▌    | 376/681 [08:06<01:45,  2.88it/s]
Continual predict:  55%|█████▌    | 377/681 [08:07<01:39,  3.06it/s]
Continual predict:  56%|█████▌    | 378/681 [08:07<01:34,  3.20it/s]
Continual predict:  56%|█████▌    | 379/681 [08:07<01:31,  3.31it/s]
Continual predict:  56%|█████▌    | 380/681 [08:07<01:28,  3.39it/s]
Continual predict:  56%|█████▌    | 381/681 [08:08<01:32,  3.26it/s]
Continual predict:  56%|█████▌    | 382/681 [08:08<01:29,  3.35it/s]
Continual predict:  56%|█████▌    | 383/681 [08:08<01:27,  3.42it/s]
Continual predict:  56%|█████▋    | 384/681 [08:09<01:25,  3.46it/s]
Continual predict:  57%|█████▋    | 385/681 [08:09<01:24,  3.50it/s]
Continual predict:  57%|█████▋    | 386/681 [08:09<01:23,  3.52it/s]
Continual predict:  57%|█████▋    | 387/681 [08:10<01:28,  3.31it/s]
Continual predict:  57%|█████▋    | 388/681 [08:10<01:39,  2.95it/s]
Continual predict:  57%|█████▋    | 389/681 [08:10<01:38,  2.96it/s]
Continual predict:  57%|█████▋    | 390/681 [08:11<01:33,  3.12it/s]
Continual predict:  57%|█████▋    | 391/681 [08:11<01:29,  3.25it/s]
Continual predict:  58%|█████▊    | 392/681 [08:11<01:31,  3.16it/s]
Continual predict:  58%|█████▊    | 393/681 [08:11<01:27,  3.28it/s]
Continual predict:  58%|█████▊    | 394/681 [08:12<01:25,  3.37it/s]
Continual predict:  58%|█████▊    | 395/681 [08:12<01:18,  3.64it/s]
Continual predict:  58%|█████▊    | 396/681 [08:12<01:28,  3.23it/s]
Continual predict:  58%|█████▊    | 397/681 [08:13<01:25,  3.33it/s]
Continual predict:  58%|█████▊    | 398/681 [08:13<01:23,  3.40it/s]
Continual predict:  59%|█████▊    | 399/681 [08:13<01:21,  3.45it/s]
Continual predict:  59%|█████▊    | 400/681 [08:13<01:20,  3.49it/s]
Continual predict:  59%|█████▉    | 401/681 [08:14<01:19,  3.51it/s]
Continual predict:  59%|█████▉    | 402/681 [08:14<01:18,  3.54it/s]
Continual predict:  59%|█████▉    | 403/681 [08:14<01:22,  3.35it/s]
Continual predict:  59%|█████▉    | 404/681 [08:15<01:21,  3.42it/s]
Continual predict:  59%|█████▉    | 405/681 [08:15<01:31,  3.03it/s]
Continual predict:  60%|█████▉    | 406/681 [08:15<01:26,  3.18it/s]
Continual predict:  60%|█████▉    | 407/681 [08:16<01:23,  3.29it/s]
Continual predict:  60%|█████▉    | 408/681 [08:16<01:22,  3.30it/s]
Continual predict:  60%|██████    | 409/681 [08:16<01:20,  3.38it/s]
Continual predict:  60%|██████    | 410/681 [08:17<01:27,  3.09it/s]
Continual predict:  60%|██████    | 411/681 [08:17<01:32,  2.91it/s]
Continual predict:  60%|██████    | 412/681 [08:17<01:27,  3.08it/s]
Continual predict:  61%|██████    | 413/681 [08:18<01:23,  3.21it/s]
Continual predict:  61%|██████    | 414/681 [08:18<01:20,  3.32it/s]
Continual predict:  61%|██████    | 415/681 [08:18<01:18,  3.39it/s]
Continual predict:  61%|██████    | 416/681 [08:18<01:16,  3.45it/s]
Continual predict:  61%|██████    | 417/681 [08:19<01:15,  3.49it/s]
Continual predict:  61%|██████▏   | 418/681 [08:19<01:19,  3.32it/s]
Continual predict:  62%|██████▏   | 419/681 [08:19<01:25,  3.05it/s]
Continual predict:  62%|██████▏   | 420/681 [08:20<01:30,  2.88it/s]
Continual predict:  62%|██████▏   | 421/681 [08:20<01:35,  2.74it/s]
Continual predict:  62%|██████▏   | 422/681 [08:21<01:34,  2.74it/s]
Continual predict:  62%|██████▏   | 423/681 [08:21<01:31,  2.81it/s]
Continual predict:  62%|██████▏   | 424/681 [08:21<01:34,  2.73it/s]
Continual predict:  62%|██████▏   | 425/681 [08:22<01:35,  2.68it/s]
Continual predict:  63%|██████▎   | 426/681 [08:22<01:36,  2.64it/s]
Continual predict:  63%|██████▎   | 427/681 [08:22<01:32,  2.74it/s]
Continual predict:  63%|██████▎   | 428/681 [08:23<01:25,  2.95it/s]
Continual predict:  63%|██████▎   | 429/681 [08:23<01:20,  3.11it/s]
Continual predict:  63%|██████▎   | 430/681 [08:23<01:17,  3.24it/s]
Continual predict:  63%|██████▎   | 431/681 [08:24<01:19,  3.15it/s]
Continual predict:  63%|██████▎   | 432/681 [08:24<01:16,  3.27it/s]
Continual predict:  64%|██████▎   | 433/681 [08:24<01:13,  3.36it/s]
Continual predict:  64%|██████▎   | 434/681 [08:24<01:16,  3.24it/s]
Continual predict:  64%|██████▍   | 435/681 [08:25<01:13,  3.33it/s]
Continual predict:  64%|██████▍   | 436/681 [08:25<01:20,  3.06it/s]
Continual predict:  64%|██████▍   | 437/681 [08:26<01:28,  2.75it/s]
Continual predict:  64%|██████▍   | 438/681 [08:26<01:22,  2.96it/s]
Continual predict:  64%|██████▍   | 439/681 [08:26<01:21,  2.96it/s]
Continual predict:  65%|██████▍   | 440/681 [08:27<01:21,  2.97it/s]
Continual predict:  65%|██████▍   | 441/681 [08:27<01:20,  2.98it/s]
Continual predict:  65%|██████▍   | 442/681 [08:27<01:16,  3.14it/s]
Continual predict:  65%|██████▌   | 443/681 [08:27<01:13,  3.26it/s]
Continual predict:  65%|██████▌   | 444/681 [08:28<01:10,  3.35it/s]
Continual predict:  65%|██████▌   | 445/681 [08:28<01:09,  3.41it/s]
Continual predict:  65%|██████▌   | 446/681 [08:28<01:11,  3.27it/s]
Continual predict:  66%|██████▌   | 447/681 [08:29<01:13,  3.18it/s]
Continual predict:  66%|██████▌   | 448/681 [08:29<01:14,  3.12it/s]
Continual predict:  66%|██████▌   | 449/681 [08:29<01:19,  2.93it/s]
Continual predict:  66%|██████▌   | 450/681 [08:30<01:14,  3.10it/s]
Continual predict:  66%|██████▌   | 451/681 [08:30<01:11,  3.23it/s]
Continual predict:  66%|██████▋   | 452/681 [08:30<01:08,  3.32it/s]
Continual predict:  67%|██████▋   | 453/681 [08:30<01:07,  3.39it/s]
Continual predict:  67%|██████▋   | 454/681 [08:31<01:09,  3.25it/s]
Continual predict:  67%|██████▋   | 455/681 [08:31<01:10,  3.19it/s]
Continual predict:  67%|██████▋   | 456/681 [08:31<01:12,  3.12it/s]
Continual predict:  67%|██████▋   | 457/681 [08:32<01:12,  3.07it/s]
Continual predict:  67%|██████▋   | 458/681 [08:32<01:09,  3.21it/s]
Continual predict:  67%|██████▋   | 459/681 [08:32<01:06,  3.31it/s]
Continual predict:  68%|██████▊   | 460/681 [08:33<01:05,  3.39it/s]
Continual predict:  68%|██████▊   | 461/681 [08:33<01:03,  3.44it/s]
Continual predict:  68%|██████▊   | 462/681 [08:33<01:02,  3.48it/s]
Continual predict:  68%|██████▊   | 463/681 [08:33<01:02,  3.51it/s]
Continual predict:  68%|██████▊   | 464/681 [08:34<01:06,  3.24it/s]
Continual predict:  68%|██████▊   | 465/681 [08:34<00:59,  3.60it/s]
Continual predict:  68%|██████▊   | 466/681 [08:34<01:01,  3.49it/s]
Continual predict:  69%|██████▊   | 467/681 [08:35<00:59,  3.60it/s]
Continual predict:  69%|██████▊   | 468/681 [08:35<01:01,  3.49it/s]
Continual predict:  69%|██████▉   | 469/681 [08:35<01:02,  3.42it/s]
Continual predict:  69%|██████▉   | 470/681 [08:36<01:02,  3.35it/s]
Continual predict:  69%|██████▉   | 471/681 [08:36<01:03,  3.32it/s]
Continual predict:  69%|██████▉   | 472/681 [08:37<01:39,  2.11it/s]
Continual predict:  69%|██████▉   | 473/681 [08:37<01:41,  2.06it/s]
Continual predict:  70%|██████▉   | 474/681 [08:37<01:23,  2.49it/s]
Continual predict:  70%|██████▉   | 475/681 [08:38<01:17,  2.67it/s]
Continual predict:  70%|██████▉   | 476/681 [08:38<01:15,  2.71it/s]
Continual predict:  70%|███████   | 477/681 [08:38<01:11,  2.85it/s]
Continual predict:  70%|███████   | 478/681 [08:39<01:11,  2.82it/s]
Continual predict:  70%|███████   | 479/681 [08:40<01:34,  2.14it/s]
Continual predict:  70%|███████   | 480/681 [08:40<01:24,  2.38it/s]
Continual predict:  71%|███████   | 481/681 [08:40<01:20,  2.48it/s]
Continual predict:  71%|███████   | 482/681 [08:41<01:25,  2.32it/s]
Continual predict:  71%|███████   | 483/681 [08:42<01:50,  1.79it/s]
Continual predict:  71%|███████   | 484/681 [08:43<02:20,  1.40it/s]
Continual predict:  71%|███████   | 485/681 [08:44<02:43,  1.20it/s]
Continual predict:  71%|███████▏  | 486/681 [08:45<02:55,  1.11it/s]
Continual predict:  72%|███████▏  | 487/681 [08:46<03:29,  1.08s/it]
Continual predict:  72%|███████▏  | 488/681 [08:47<02:50,  1.13it/s]
Continual predict:  72%|███████▏  | 489/681 [08:47<02:39,  1.20it/s]
Continual predict:  72%|███████▏  | 490/681 [08:48<02:35,  1.23it/s]
Continual predict:  72%|███████▏  | 491/681 [08:49<02:34,  1.23it/s]
Continual predict:  72%|███████▏  | 492/681 [08:50<02:39,  1.19it/s]
Continual predict:  72%|███████▏  | 493/681 [08:51<02:32,  1.23it/s]
Continual predict:  73%|███████▎  | 494/681 [08:52<02:36,  1.20it/s]
Continual predict:  73%|███████▎  | 495/681 [08:53<03:17,  1.06s/it]
Continual predict:  73%|███████▎  | 496/681 [08:54<03:02,  1.02it/s]
Continual predict:  73%|███████▎  | 497/681 [08:55<02:37,  1.17it/s]
Continual predict:  73%|███████▎  | 498/681 [08:55<02:36,  1.17it/s]
Continual predict:  73%|███████▎  | 499/681 [08:56<02:35,  1.17it/s]
Continual predict:  73%|███████▎  | 500/681 [08:57<02:12,  1.37it/s]
Continual predict:  74%|███████▎  | 501/681 [08:57<02:01,  1.48it/s]
Continual predict:  74%|███████▎  | 502/681 [08:58<01:59,  1.50it/s]
Continual predict:  74%|███████▍  | 503/681 [08:59<02:06,  1.41it/s]
Continual predict:  74%|███████▍  | 504/681 [08:59<01:57,  1.50it/s]
Continual predict:  74%|███████▍  | 505/681 [09:00<02:21,  1.24it/s]
Continual predict:  74%|███████▍  | 506/681 [09:01<02:25,  1.21it/s]
Continual predict:  74%|███████▍  | 507/681 [09:02<02:29,  1.17it/s]
Continual predict:  75%|███████▍  | 508/681 [09:03<02:21,  1.22it/s]
Continual predict:  75%|███████▍  | 509/681 [09:04<02:22,  1.21it/s]
Continual predict:  75%|███████▍  | 510/681 [09:06<03:11,  1.12s/it]
Continual predict:  75%|███████▌  | 511/681 [09:07<03:07,  1.10s/it]
Continual predict:  75%|███████▌  | 512/681 [09:07<02:42,  1.04it/s]
Continual predict:  75%|███████▌  | 513/681 [09:09<03:13,  1.15s/it]
Continual predict:  75%|███████▌  | 514/681 [09:10<03:03,  1.10s/it]
Continual predict:  76%|███████▌  | 515/681 [09:11<02:43,  1.01it/s]
Continual predict:  76%|███████▌  | 516/681 [09:11<02:29,  1.11it/s]
Continual predict:  76%|███████▌  | 517/681 [09:12<02:31,  1.08it/s]
Continual predict:  76%|███████▌  | 518/681 [09:13<02:32,  1.07it/s]
Continual predict:  76%|███████▌  | 519/681 [09:15<03:01,  1.12s/it]
Continual predict:  76%|███████▋  | 520/681 [09:15<02:37,  1.02it/s]
Continual predict:  77%|███████▋  | 521/681 [09:16<02:36,  1.02it/s]
Continual predict:  77%|███████▋  | 522/681 [09:17<02:18,  1.15it/s]
Continual predict:  77%|███████▋  | 523/681 [09:18<02:13,  1.18it/s]
Continual predict:  77%|███████▋  | 524/681 [09:18<02:04,  1.26it/s]
Continual predict:  77%|███████▋  | 525/681 [09:19<02:05,  1.24it/s]
Continual predict:  77%|███████▋  | 526/681 [09:20<02:10,  1.19it/s]
Continual predict:  77%|███████▋  | 527/681 [09:21<02:07,  1.21it/s]
Continual predict:  78%|███████▊  | 528/681 [09:22<02:34,  1.01s/it]
Continual predict:  78%|███████▊  | 529/681 [09:23<02:19,  1.09it/s]
Continual predict:  78%|███████▊  | 530/681 [09:24<02:29,  1.01it/s]
Continual predict:  78%|███████▊  | 531/681 [09:26<02:40,  1.07s/it]
Continual predict:  78%|███████▊  | 532/681 [09:27<02:56,  1.18s/it]
Continual predict:  78%|███████▊  | 533/681 [09:27<02:21,  1.05it/s]
Continual predict:  78%|███████▊  | 534/681 [09:28<02:13,  1.10it/s]
Continual predict:  79%|███████▊  | 535/681 [09:29<02:12,  1.10it/s]
Continual predict:  79%|███████▊  | 536/681 [09:29<01:47,  1.35it/s]
Continual predict:  79%|███████▉  | 537/681 [09:30<01:51,  1.29it/s]
Continual predict:  79%|███████▉  | 538/681 [09:31<01:50,  1.29it/s]
Continual predict:  79%|███████▉  | 539/681 [09:32<01:59,  1.19it/s]
Continual predict:  79%|███████▉  | 540/681 [09:33<02:15,  1.04it/s]
Continual predict:  79%|███████▉  | 541/681 [09:34<01:58,  1.19it/s]
Continual predict:  80%|███████▉  | 542/681 [09:34<01:44,  1.33it/s]
Continual predict:  80%|███████▉  | 543/681 [09:35<01:45,  1.30it/s]
Continual predict:  80%|███████▉  | 544/681 [09:36<02:00,  1.13it/s]
Continual predict:  80%|████████  | 545/681 [09:38<02:30,  1.11s/it]
Continual predict:  80%|████████  | 546/681 [09:39<02:16,  1.01s/it]
Continual predict:  80%|████████  | 547/681 [09:40<02:11,  1.02it/s]
Continual predict:  80%|████████  | 548/681 [09:41<02:06,  1.05it/s]
Continual predict:  81%|████████  | 549/681 [09:42<02:07,  1.03it/s]
Continual predict:  81%|████████  | 550/681 [09:43<02:10,  1.00it/s]
Continual predict:  81%|████████  | 551/681 [09:43<01:49,  1.19it/s]
Continual predict:  81%|████████  | 552/681 [09:44<01:40,  1.29it/s]
Continual predict:  81%|████████  | 553/681 [09:45<02:02,  1.05it/s]
Continual predict:  81%|████████▏ | 554/681 [09:46<02:09,  1.02s/it]
Continual predict:  81%|████████▏ | 555/681 [09:47<02:10,  1.03s/it]
Continual predict:  82%|████████▏ | 556/681 [09:48<01:54,  1.09it/s]
Continual predict:  82%|████████▏ | 557/681 [09:50<02:15,  1.10s/it]
Continual predict:  82%|████████▏ | 558/681 [09:50<01:54,  1.08it/s]
Continual predict:  82%|████████▏ | 559/681 [09:51<02:01,  1.00it/s]
Continual predict:  82%|████████▏ | 560/681 [09:52<02:05,  1.04s/it]
Continual predict:  82%|████████▏ | 561/681 [09:54<02:17,  1.14s/it]
Continual predict:  83%|████████▎ | 562/681 [09:55<02:22,  1.20s/it]
Continual predict:  83%|████████▎ | 563/681 [09:57<02:32,  1.29s/it]
Continual predict:  83%|████████▎ | 564/681 [09:58<02:32,  1.31s/it]
Continual predict:  83%|████████▎ | 565/681 [09:59<02:29,  1.29s/it]
Continual predict:  83%|████████▎ | 566/681 [10:00<02:10,  1.14s/it]
Continual predict:  83%|████████▎ | 567/681 [10:01<02:11,  1.16s/it]
Continual predict:  83%|████████▎ | 568/681 [10:02<01:58,  1.05s/it]
Continual predict:  84%|████████▎ | 569/681 [10:03<01:55,  1.03s/it]
Continual predict:  84%|████████▎ | 570/681 [10:03<01:33,  1.18it/s]
Continual predict:  84%|████████▍ | 571/681 [10:04<01:34,  1.16it/s]
Continual predict:  84%|████████▍ | 572/681 [10:05<01:23,  1.31it/s]
Continual predict:  84%|████████▍ | 573/681 [10:06<01:35,  1.13it/s]
Continual predict:  84%|████████▍ | 574/681 [10:07<01:26,  1.23it/s]
Continual predict:  84%|████████▍ | 575/681 [10:08<01:32,  1.14it/s]
Continual predict:  85%|████████▍ | 576/681 [10:08<01:22,  1.28it/s]
Continual predict:  85%|████████▍ | 577/681 [10:09<01:24,  1.23it/s]
Continual predict:  85%|████████▍ | 578/681 [10:10<01:14,  1.39it/s]
Continual predict:  85%|████████▌ | 579/681 [10:11<01:19,  1.28it/s]
Continual predict:  85%|████████▌ | 580/681 [10:12<01:35,  1.06it/s]
Continual predict:  85%|████████▌ | 581/681 [10:12<01:24,  1.18it/s]
Continual predict:  85%|████████▌ | 582/681 [10:13<01:03,  1.57it/s]
Continual predict:  86%|████████▌ | 583/681 [10:13<00:47,  2.05it/s]
Continual predict:  86%|████████▌ | 584/681 [10:13<00:37,  2.60it/s]
Continual predict:  86%|████████▌ | 585/681 [10:13<00:29,  3.21it/s]
Continual predict:  86%|████████▌ | 586/681 [10:13<00:24,  3.85it/s]
Continual predict:  86%|████████▌ | 587/681 [10:13<00:21,  4.46it/s]
Continual predict:  86%|████████▋ | 588/681 [10:13<00:18,  5.02it/s]
Continual predict:  86%|████████▋ | 589/681 [10:14<00:16,  5.50it/s]
Continual predict:  87%|████████▋ | 590/681 [10:14<00:15,  5.93it/s]
Continual predict:  87%|████████▋ | 591/681 [10:14<00:14,  6.24it/s]
Continual predict:  87%|████████▋ | 592/681 [10:14<00:13,  6.47it/s]
Continual predict:  87%|████████▋ | 593/681 [10:14<00:13,  6.69it/s]
Continual predict:  87%|████████▋ | 594/681 [10:14<00:12,  6.83it/s]
Continual predict:  87%|████████▋ | 595/681 [10:14<00:12,  6.95it/s]
Continual predict:  88%|████████▊ | 596/681 [10:15<00:12,  7.04it/s]
Continual predict:  88%|████████▊ | 597/681 [10:15<00:11,  7.10it/s]
Continual predict:  88%|████████▊ | 598/681 [10:15<00:11,  7.10it/s]
Continual predict:  88%|████████▊ | 599/681 [10:15<00:11,  7.10it/s]
Continual predict:  88%|████████▊ | 600/681 [10:15<00:11,  7.14it/s]
Continual predict:  88%|████████▊ | 601/681 [10:15<00:11,  7.13it/s]
Continual predict:  88%|████████▊ | 602/681 [10:15<00:11,  7.11it/s]
Continual predict:  89%|████████▊ | 603/681 [10:16<00:10,  7.11it/s]
Continual predict:  89%|████████▊ | 604/681 [10:16<00:10,  7.10it/s]
Continual predict:  89%|████████▉ | 605/681 [10:16<00:10,  7.09it/s]
Continual predict:  89%|████████▉ | 606/681 [10:16<00:10,  7.07it/s]
Continual predict:  89%|████████▉ | 607/681 [10:16<00:10,  7.11it/s]
Continual predict:  89%|████████▉ | 608/681 [10:16<00:10,  7.12it/s]
Continual predict:  89%|████████▉ | 609/681 [10:16<00:10,  7.12it/s]
Continual predict:  90%|████████▉ | 610/681 [10:17<00:09,  7.14it/s]
Continual predict:  90%|████████▉ | 611/681 [10:17<00:09,  7.17it/s]
Continual predict:  90%|████████▉ | 612/681 [10:17<00:09,  7.13it/s]
Continual predict:  90%|█████████ | 613/681 [10:17<00:09,  7.13it/s]
Continual predict:  90%|█████████ | 614/681 [10:17<00:09,  7.16it/s]
Continual predict:  90%|█████████ | 615/681 [10:17<00:09,  7.15it/s]
Continual predict:  90%|█████████ | 616/681 [10:17<00:09,  7.13it/s]
Continual predict:  91%|█████████ | 617/681 [10:18<00:08,  7.17it/s]
Continual predict:  91%|█████████ | 618/681 [10:18<00:08,  7.15it/s]
Continual predict:  91%|█████████ | 619/681 [10:18<00:08,  7.13it/s]
Continual predict:  91%|█████████ | 620/681 [10:18<00:08,  7.11it/s]
Continual predict:  91%|█████████ | 621/681 [10:18<00:08,  7.11it/s]
Continual predict:  91%|█████████▏| 622/681 [10:18<00:08,  7.10it/s]
Continual predict:  91%|█████████▏| 623/681 [10:18<00:08,  7.14it/s]
Continual predict:  92%|█████████▏| 624/681 [10:19<00:07,  7.17it/s]
Continual predict:  92%|█████████▏| 625/681 [10:19<00:07,  7.14it/s]
Continual predict:  92%|█████████▏| 626/681 [10:19<00:07,  7.16it/s]
Continual predict:  92%|█████████▏| 627/681 [10:19<00:07,  7.14it/s]
Continual predict:  92%|█████████▏| 628/681 [10:19<00:07,  7.17it/s]
Continual predict:  92%|█████████▏| 629/681 [10:19<00:07,  7.19it/s]
Continual predict:  93%|█████████▎| 630/681 [10:19<00:07,  7.17it/s]
Continual predict:  93%|█████████▎| 631/681 [10:19<00:07,  7.14it/s]
Continual predict:  93%|█████████▎| 632/681 [10:20<00:06,  7.13it/s]
Continual predict:  93%|█████████▎| 633/681 [10:20<00:06,  7.17it/s]
Continual predict:  93%|█████████▎| 634/681 [10:20<00:06,  7.14it/s]
Continual predict:  93%|█████████▎| 635/681 [10:20<00:06,  7.17it/s]
Continual predict:  93%|█████████▎| 636/681 [10:20<00:06,  7.15it/s]
Continual predict:  94%|█████████▎| 637/681 [10:20<00:06,  7.14it/s]
Continual predict:  94%|█████████▎| 638/681 [10:20<00:06,  7.11it/s]
Continual predict:  94%|█████████▍| 639/681 [10:21<00:05,  7.11it/s]
Continual predict:  94%|█████████▍| 640/681 [10:21<00:05,  7.15it/s]
Continual predict:  94%|█████████▍| 641/681 [10:21<00:05,  7.18it/s]
Continual predict:  94%|█████████▍| 642/681 [10:21<00:05,  7.16it/s]
Continual predict:  94%|█████████▍| 643/681 [10:21<00:05,  7.14it/s]
Continual predict:  95%|█████████▍| 644/681 [10:21<00:05,  7.17it/s]
Continual predict:  95%|█████████▍| 645/681 [10:21<00:05,  7.16it/s]
Continual predict:  95%|█████████▍| 646/681 [10:22<00:04,  7.14it/s]
Continual predict:  95%|█████████▌| 647/681 [10:22<00:04,  7.12it/s]
Continual predict:  95%|█████████▌| 648/681 [10:22<00:04,  7.16it/s]
Continual predict:  95%|█████████▌| 649/681 [10:22<00:04,  7.15it/s]
Continual predict:  95%|█████████▌| 650/681 [10:22<00:04,  7.18it/s]
Continual predict:  96%|█████████▌| 651/681 [10:22<00:04,  7.20it/s]
Continual predict:  96%|█████████▌| 652/681 [10:22<00:04,  7.20it/s]
Continual predict:  96%|█████████▌| 653/681 [10:23<00:03,  7.22it/s]
Continual predict:  96%|█████████▌| 654/681 [10:23<00:03,  7.23it/s]
Continual predict:  96%|█████████▌| 655/681 [10:23<00:03,  7.23it/s]
Continual predict:  96%|█████████▋| 656/681 [10:23<00:03,  7.19it/s]
Continual predict:  96%|█████████▋| 657/681 [10:23<00:03,  7.16it/s]
Continual predict:  97%|█████████▋| 658/681 [10:23<00:03,  7.14it/s]
Continual predict:  97%|█████████▋| 659/681 [10:23<00:03,  7.17it/s]
Continual predict:  97%|█████████▋| 660/681 [10:24<00:02,  7.15it/s]
Continual predict:  97%|█████████▋| 661/681 [10:24<00:02,  7.14it/s]
Continual predict:  97%|█████████▋| 662/681 [10:24<00:02,  7.12it/s]
Continual predict:  97%|█████████▋| 663/681 [10:24<00:02,  7.11it/s]
Continual predict:  98%|█████████▊| 664/681 [10:24<00:02,  7.10it/s]
Continual predict:  98%|█████████▊| 665/681 [10:24<00:02,  7.10it/s]
Continual predict:  98%|█████████▊| 666/681 [10:24<00:02,  7.14it/s]
Continual predict:  98%|█████████▊| 667/681 [10:25<00:01,  7.13it/s]
Continual predict:  98%|█████████▊| 668/681 [10:25<00:01,  7.12it/s]
Continual predict:  98%|█████████▊| 669/681 [10:25<00:01,  7.16it/s]
Continual predict:  98%|█████████▊| 670/681 [10:25<00:01,  7.14it/s]
Continual predict:  99%|█████████▊| 671/681 [10:25<00:01,  7.12it/s]
Continual predict:  99%|█████████▊| 672/681 [10:25<00:01,  7.12it/s]
Continual predict:  99%|█████████▉| 673/681 [10:25<00:01,  7.15it/s]
Continual predict:  99%|█████████▉| 674/681 [10:26<00:00,  7.13it/s]
Continual predict:  99%|█████████▉| 675/681 [10:26<00:00,  7.17it/s]
Continual predict:  99%|█████████▉| 676/681 [10:26<00:00,  7.19it/s]
Continual predict:  99%|█████████▉| 677/681 [10:26<00:00,  7.21it/s]
Continual predict: 100%|█████████▉| 678/681 [10:26<00:00,  7.18it/s]
Continual predict: 100%|█████████▉| 679/681 [10:26<00:00,  7.16it/s]
Continual predict: 100%|█████████▉| 680/681 [10:26<00:00,  7.14it/s]
Continual predict: 100%|██████████| 681/681 [10:26<00:00,  7.10it/s]
Continual predict: 100%|██████████| 681/681 [10:26<00:00,  1.09it/s]
+
+  ┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
+  │           SRT ROUTING ACCURACY SUMMARY                                                      │
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  GT=task002_quoref_answer_generation  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot4=100]│
+  │  GT=task073_commonsenseqa_answer_gen  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot0=100]│
+  │  GT=task1290_xsum_summarization       slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot6=100]│
+  │  GT=task1510_evalution_relation_extr  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot3=100]│
+  │  GT=task1572_samsum_summary           slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot8=20]│
+  │  GT=task1729_personachat_generate_ne  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot1=100]│
+  │  GT=task181_outcome_extraction        slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot5=43]│
+  │  GT=task363_sst2_polarity_classifica  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot7=100]│
+  │  GT=task639_multi_woz_user_utterance  slot_acc=100.0%  label_acc=100.0%  disagree=  0  [slot2=18]│
+  ├──────────────────────────────────────────────────────────────────────────────────────────────────┤
+  │  OVERALL: slot_acc=100.0% (681/681)  label_acc=100.0% (681/681)  disagree=  0                    │
+  └──────────────────────────────────────────────────────────────────────────────────────────────────┘
+***** predict metrics *****
+  epoch                                                               =       15.0
+  predict_eval_rougeL                                                 =    60.9216
+  predict_eval_rougeL_for_CL                                          =    60.9216
+  predict_eval_rougeL_for_task002_quoref_answer_generation            =     76.381
+  predict_eval_rougeL_for_task073_commonsenseqa_answer_generation     =       64.0
+  predict_eval_rougeL_for_task1290_xsum_summarization                 =    29.5925
+  predict_eval_rougeL_for_task1510_evalution_relation_extraction      =    99.1333
+  predict_eval_rougeL_for_task1572_samsum_summary                     =    41.3728
+  predict_eval_rougeL_for_task1729_personachat_generate_next          =    18.2202
+  predict_eval_rougeL_for_task181_outcome_extraction                  =     59.354
+  predict_eval_rougeL_for_task363_sst2_polarity_classification        =       91.0
+  predict_eval_rougeL_for_task639_multi_woz_user_utterance_generation =    15.2912
+  predict_exact_match                                                 =    50.8076
+  predict_exact_match_for_CL                                          =    50.8076
+  predict_exact_match_for_task002_quoref_answer_generation            =       70.0
+  predict_exact_match_for_task073_commonsenseqa_answer_generation     =       64.0
+  predict_exact_match_for_task1290_xsum_summarization                 =        0.0
+  predict_exact_match_for_task1510_evalution_relation_extraction      =       98.0
+  predict_exact_match_for_task1572_samsum_summary                     =        0.0
+  predict_exact_match_for_task1729_personachat_generate_next          =        0.0
+  predict_exact_match_for_task181_outcome_extraction                  =    53.4884
+  predict_exact_match_for_task363_sst2_polarity_classification        =       91.0
+  predict_exact_match_for_task639_multi_woz_user_utterance_generation =        0.0
+  predict_gen_len                                                     =     7.9559
+  predict_global_step                                                 =        465
+  predict_rouge1                                                      =    62.3716
+  predict_rouge1_for_CL                                               =    62.3716
+  predict_rouge1_for_task002_quoref_answer_generation                 =     76.381
+  predict_rouge1_for_task073_commonsenseqa_answer_generation          =       64.0
+  predict_rouge1_for_task1290_xsum_summarization                      =      36.75
+  predict_rouge1_for_task1510_evalution_relation_extraction           =    99.1333
+  predict_rouge1_for_task1572_samsum_summary                          =    50.3879
+  predict_rouge1_for_task1729_personachat_generate_next               =    19.1341
+  predict_rouge1_for_task181_outcome_extraction                       =     59.354
+  predict_rouge1_for_task363_sst2_polarity_classification             =       91.0
+  predict_rouge1_for_task639_multi_woz_user_utterance_generation      =    15.2912
+  predict_runtime                                                     = 0:10:26.98
+  predict_samples                                                     =        681
+  predict_samples_per_second                                          =      1.086
+  predict_steps_per_second                                            =      1.086
+[2026-04-30 01:38:10,318] [INFO] [launch.py:367:main] Process 1238647 exits successfully.
diff --git a/superni_order1_llama_srt/outputs/task_order.txt b/superni_order1_llama_srt/outputs/task_order.txt
new file mode 100644
index 0000000000000000000000000000000000000000..29bfeb74c045d82e91086a18197e6f0614fdbb63
--- /dev/null
+++ b/superni_order1_llama_srt/outputs/task_order.txt
@@ -0,0 +1 @@
+task1572_samsum_summary,task363_sst2_polarity_classification,task1290_xsum_summarization,task181_outcome_extraction,task002_quoref_answer_generation,task1510_evalution_relation_extraction,task639_multi_woz_user_utterance_generation,task1729_personachat_generate_next,task073_commonsenseqa_answer_generation,task1590_diplomacy_text_generation,task748_glucose_reverse_cause_event_detection,task511_reddit_tifu_long_text_summarization,task591_sciq_answer_generation,task1687_sentiment140_classification,task875_emotion_classification
\ No newline at end of file