Training in progress, step 10

adapter_config.json

@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "gate_proj",
-    "k_proj",
     "down_proj",
+    "k_proj",
     "q_proj",
+    "gate_proj",
     "up_proj",
-    "v_proj",
-    "o_proj"
+    "o_proj",
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e02a732a97109085620824021d0ee4dc74fa0457f2941ea2250852cb25388606
+oid sha256:4543dc3325c39698fbbddf9356ca7f09abd2079c5e4495fdccbbae8efd448994
 size 1656903768

args.json

@@ -70,7 +70,7 @@
   "custom_register_path": [],
   "ignore_args_error": false,
   "use_swift_lora": false,
-  "output_dir": "/root/dataDisk/hakob/output/v2-20250126-184559",
+  "output_dir": "/root/dataDisk/hakob/output/v4-20250126-221202",
   "overwrite_output_dir": false,
   "do_train": false,
   "do_eval": false,
@@ -81,7 +81,7 @@
   "per_device_eval_batch_size": 1,
   "per_gpu_train_batch_size": null,
   "per_gpu_eval_batch_size": null,
-  "gradient_accumulation_steps": 32,
+  "gradient_accumulation_steps": 1,
   "eval_accumulation_steps": null,
   "eval_delay": 0,
   "torch_empty_cache_steps": null,
@@ -100,7 +100,7 @@
   "log_level": "passive",
   "log_level_replica": "warning",
   "log_on_each_node": true,
-  "logging_dir": "/root/dataDisk/hakob/output/v2-20250126-184559/runs",
+  "logging_dir": "/root/dataDisk/hakob/output/v4-20250126-221202/runs",
   "logging_strategy": "steps",
   "logging_first_step": true,
   "logging_steps": 1,
@@ -335,8 +335,8 @@
   "local_world_size": 8,
   "model_suffix": "agentic_v5",
   "model_info": "ModelInfo(model_type='llama3_2', model_dir='/root/highspeedstorage/ag-model-sg/agentic_v5', torch_dtype=torch.bfloat16, max_model_len=131072, quant_method=None, quant_bits=None, config=None, task_type=None)",
-  "model_meta": "ModelMeta(model_type='llama3_2', model_groups=[ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.2-1B', hf_model_id='meta-llama/Llama-3.2-1B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B', hf_model_id='meta-llama/Llama-3.2-3B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-1B-Instruct', hf_model_id='meta-llama/Llama-3.2-1B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B-Instruct', hf_model_id='meta-llama/Llama-3.2-3B-Instruct', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.3-70B-Instruct', hf_model_id='meta-llama/Llama-3.3-70B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', hf_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='llama3_2', get_function=<function get_model_tokenizer_with_flash_attn at 0x7ff5789bc7c0>, model_arch='llama', architectures=['LlamaForCausalLM'], is_multimodal=False, additional_saved_files=[], torch_dtype=None, ignore_patterns=[], requires=['transformers>=4.45'], tags=[])",
+  "model_meta": "ModelMeta(model_type='llama3_2', model_groups=[ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.2-1B', hf_model_id='meta-llama/Llama-3.2-1B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B', hf_model_id='meta-llama/Llama-3.2-3B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-1B-Instruct', hf_model_id='meta-llama/Llama-3.2-1B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B-Instruct', hf_model_id='meta-llama/Llama-3.2-3B-Instruct', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.3-70B-Instruct', hf_model_id='meta-llama/Llama-3.3-70B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', hf_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='llama3_2', get_function=<function get_model_tokenizer_with_flash_attn at 0x7f5ce8ded6c0>, model_arch='llama', architectures=['LlamaForCausalLM'], is_multimodal=False, additional_saved_files=[], torch_dtype=None, ignore_patterns=[], requires=['transformers>=4.45'], tags=[])",
   "model_dir": "/root/highspeedstorage/ag-model-sg/agentic_v5",
   "hub": "<class 'swift.hub.hub.HFHub'>",
-  "training_args": "Seq2SeqTrainingArguments(output_dir='/root/dataDisk/hakob/output/v2-20250126-184559', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=1, per_device_eval_batch_size=1, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=32, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=7e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, max_grad_norm=1.0, num_train_epochs=3.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.05, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/root/dataDisk/hakob/output/v2-20250126-184559/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=5, save_total_limit=2, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=5, dataloader_num_workers=0, dataloader_prefetch_factor=None, past_index=-1, run_name='/root/dataDisk/hakob/output/v2-20250126-184559', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=False, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed={'fp16': {'enabled': 'auto', 'loss_scale': 0, 'loss_scale_window': 1000, 'initial_scale_power': 16, 'hysteresis': 2, 'min_loss_scale': 1}, 'bf16': {'enabled': 'auto'}, 'zero_optimization': {'stage': 3, 'offload_optimizer': {'device': 'none', 'pin_memory': True}, 'offload_param': {'device': 'cpu', 'pin_memory': True}, 'overlap_comm': True, 'contiguous_gradients': True, 'sub_group_size': 1000000000.0, 'reduce_bucket_size': 'auto', 'stage3_prefetch_bucket_size': 'auto', 'stage3_param_persistence_threshold': 'auto', 'stage3_max_live_parameters': 1000000000.0, 'stage3_max_reuse_distance': 1000000000.0, 'stage3_gather_16bit_weights_on_model_save': True}, 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'steps_per_print': 2000, 'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'wall_clock_breakdown': False}, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['comet_ml'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=True, resume_from_checkpoint=None, hub_model_id='TheAgenticAI/Agentic-V7', hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=True, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': True}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=1800, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, acc_strategy='token', sequence_parallel_size=1, check_model=True, train_sampler_random=True, is_encoder_decoder=False, metric_warmup_step=0, train_dataset_sample=-1, fsdp_num=1, acc_steps=1, train_type='lora')"
+  "training_args": "Seq2SeqTrainingArguments(output_dir='/root/dataDisk/hakob/output/v4-20250126-221202', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=1, per_device_eval_batch_size=1, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=1, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=7e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, max_grad_norm=1.0, num_train_epochs=3.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.05, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/root/dataDisk/hakob/output/v4-20250126-221202/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=5, save_total_limit=2, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=5, dataloader_num_workers=0, dataloader_prefetch_factor=None, past_index=-1, run_name='/root/dataDisk/hakob/output/v4-20250126-221202', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=False, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed={'fp16': {'enabled': 'auto', 'loss_scale': 0, 'loss_scale_window': 1000, 'initial_scale_power': 16, 'hysteresis': 2, 'min_loss_scale': 1}, 'bf16': {'enabled': 'auto'}, 'zero_optimization': {'stage': 3, 'offload_optimizer': {'device': 'none', 'pin_memory': True}, 'offload_param': {'device': 'cpu', 'pin_memory': True}, 'overlap_comm': True, 'contiguous_gradients': True, 'sub_group_size': 1000000000.0, 'reduce_bucket_size': 'auto', 'stage3_prefetch_bucket_size': 'auto', 'stage3_param_persistence_threshold': 'auto', 'stage3_max_live_parameters': 1000000000.0, 'stage3_max_reuse_distance': 1000000000.0, 'stage3_gather_16bit_weights_on_model_save': True}, 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'steps_per_print': 2000, 'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'wall_clock_breakdown': False}, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['comet_ml'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=True, resume_from_checkpoint=None, hub_model_id='TheAgenticAI/Agentic-V7', hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=True, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': True}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=1800, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, acc_strategy='token', sequence_parallel_size=1, check_model=True, train_sampler_random=True, is_encoder_decoder=False, metric_warmup_step=0, train_dataset_sample=-1, fsdp_num=1, acc_steps=1, train_type='lora')"
 }

logging.jsonl

@@ -1,12 +1,14 @@
-{"loss": 30.97451973, "token_acc": 0.74518379, "grad_norm": 3.53021257, "learning_rate": 3.5e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001131, "epoch": 0.1221374, "global_step/max_steps": "1/24", "percentage": "4.17%", "elapsed_time": "14m 37s", "remaining_time": "5h 36m 11s"}
-{"loss": 32.12932587, "token_acc": 0.72883569, "grad_norm": 2.1962259, "learning_rate": 7e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001128, "epoch": 0.24427481, "global_step/max_steps": "2/24", "percentage": "8.33%", "elapsed_time": "29m 25s", "remaining_time": "5h 23m 40s"}
-{"loss": 30.97593689, "token_acc": 0.73727767, "grad_norm": 5.23222928, "learning_rate": 6.96e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001122, "epoch": 0.36641221, "global_step/max_steps": "3/24", "percentage": "12.50%", "elapsed_time": "44m 25s", "remaining_time": "5h 10m 56s"}
-{"loss": 30.84427452, "token_acc": 0.76022193, "grad_norm": 3.02257953, "learning_rate": 6.86e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001128, "epoch": 0.48854962, "global_step/max_steps": "4/24", "percentage": "16.67%", "elapsed_time": "58m 58s", "remaining_time": "4h 54m 54s"}
-{"loss": 32.8370285, "token_acc": 0.75285632, "grad_norm": 2.24804417, "learning_rate": 6.68e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001133, "epoch": 0.61068702, "global_step/max_steps": "5/24", "percentage": "20.83%", "elapsed_time": "1h 13m 26s", "remaining_time": "4h 39m 3s"}
-{"eval_loss": 0.91817999, "eval_token_acc": 0.77373173, "eval_runtime": 22.0971, "eval_samples_per_second": 0.95, "eval_steps_per_second": 0.136, "epoch": 0.61068702, "global_step/max_steps": "5/24", "percentage": "20.83%", "elapsed_time": "1h 13m 48s", "remaining_time": "4h 40m 27s"}
-{"loss": 33.37625504, "token_acc": 0.74381874, "grad_norm": 16.70630121, "learning_rate": 6.44e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001118, "epoch": 0.73282443, "global_step/max_steps": "6/24", "percentage": "25.00%", "elapsed_time": "1h 29m 18s", "remaining_time": "4h 27m 56s"}
-{"loss": 31.02261353, "token_acc": 0.75488503, "grad_norm": 9.27220664, "learning_rate": 6.15e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001122, "epoch": 0.85496183, "global_step/max_steps": "7/24", "percentage": "29.17%", "elapsed_time": "1h 43m 49s", "remaining_time": "4h 12m 8s"}
-{"loss": 33.63227844, "token_acc": 0.75749061, "grad_norm": 1.92525593, "learning_rate": 5.79e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001122, "epoch": 0.97709924, "global_step/max_steps": "8/24", "percentage": "33.33%", "elapsed_time": "1h 58m 45s", "remaining_time": "3h 57m 30s"}
-{"loss": 6.81848335, "token_acc": 0.73555901, "grad_norm": 1.92525593, "learning_rate": 5.39e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001232, "epoch": 1.0, "global_step/max_steps": "9/24", "percentage": "37.50%", "elapsed_time": "2h 1m 40s", "remaining_time": "3h 22m 46s"}
-{"loss": 31.74690628, "token_acc": 0.74373912, "grad_norm": 50.72140627, "learning_rate": 4.95e-06, "memory(GiB)": 91.05, "train_speed(iter/s)": 0.001224, "epoch": 1.1221374, "global_step/max_steps": "10/24", "percentage": "41.67%", "elapsed_time": "2h 16m 2s", "remaining_time": "3h 10m 27s"}
-{"eval_loss": 0.91494149, "eval_token_acc": 0.77429834, "eval_runtime": 22.0647, "eval_samples_per_second": 0.952, "eval_steps_per_second": 0.136, "epoch": 1.1221374, "global_step/max_steps": "10/24", "percentage": "41.67%", "elapsed_time": "2h 16m 24s", "remaining_time": "3h 10m 58s"}
+{"loss": 1.2135762, "token_acc": 0.78263989, "grad_norm": 3.97283773, "learning_rate": 1.8e-07, "memory(GiB)": 48.55, "train_speed(iter/s)": 0.020719, "epoch": 0.00381679, "global_step/max_steps": "1/786", "percentage": "0.13%", "elapsed_time": "41s", "remaining_time": "9h 0m 20s"}
+{"loss": 1.06550384, "token_acc": 0.59249677, "grad_norm": 2.18691791, "learning_rate": 3.5e-07, "memory(GiB)": 86.66, "train_speed(iter/s)": 0.025632, "epoch": 0.00763359, "global_step/max_steps": "2/786", "percentage": "0.25%", "elapsed_time": "1m 11s", "remaining_time": "7h 44m 16s"}
+{"loss": 1.11164117, "token_acc": 0.69653179, "grad_norm": 2.83606289, "learning_rate": 5.2e-07, "memory(GiB)": 86.66, "train_speed(iter/s)": 0.027824, "epoch": 0.01145038, "global_step/max_steps": "3/786", "percentage": "0.38%", "elapsed_time": "1m 40s", "remaining_time": "7h 18m 43s"}
+{"loss": 0.76855314, "token_acc": 0.76425368, "grad_norm": 0.41384293, "learning_rate": 7e-07, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.031311, "epoch": 0.01526718, "global_step/max_steps": "4/786", "percentage": "0.51%", "elapsed_time": "2m 0s", "remaining_time": "6h 33m 33s"}
+{"loss": 0.85836798, "token_acc": 0.75914362, "grad_norm": 3.80895579, "learning_rate": 8.7e-07, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.031763, "epoch": 0.01908397, "global_step/max_steps": "5/786", "percentage": "0.64%", "elapsed_time": "2m 30s", "remaining_time": "6h 31m 40s"}
+{"eval_loss": 0.87787485, "eval_token_acc": 0.7750916, "eval_runtime": 22.1205, "eval_samples_per_second": 0.949, "eval_steps_per_second": 0.136, "epoch": 0.01908397, "global_step/max_steps": "5/786", "percentage": "0.64%", "elapsed_time": "2m 52s", "remaining_time": "7h 29m 16s"}
+{"loss": 1.34543896, "token_acc": 0.72805184, "grad_norm": 8.86843324, "learning_rate": 1.05e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.027512, "epoch": 0.02290076, "global_step/max_steps": "6/786", "percentage": "0.76%", "elapsed_time": "3m 31s", "remaining_time": "7h 37m 25s"}
+{"loss": 0.9533174, "token_acc": 0.79079903, "grad_norm": 2.2993539, "learning_rate": 1.22e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.028252, "epoch": 0.02671756, "global_step/max_steps": "7/786", "percentage": "0.89%", "elapsed_time": "4m 0s", "remaining_time": "7h 26m 38s"}
+{"loss": 0.85449177, "token_acc": 0.72237569, "grad_norm": 1.17990306, "learning_rate": 1.4e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029069, "epoch": 0.03053435, "global_step/max_steps": "8/786", "percentage": "1.02%", "elapsed_time": "4m 28s", "remaining_time": "7h 14m 46s"}
+{"loss": 0.89482594, "token_acc": 0.72823219, "grad_norm": 4.32352436, "learning_rate": 1.57e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029524, "epoch": 0.03435115, "global_step/max_steps": "9/786", "percentage": "1.15%", "elapsed_time": "4m 57s", "remaining_time": "7h 8m 36s"}
+{"loss": 0.87332559, "token_acc": 0.74550129, "grad_norm": 0.63810492, "learning_rate": 1.75e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.0311, "epoch": 0.03816794, "global_step/max_steps": "10/786", "percentage": "1.27%", "elapsed_time": "5m 14s", "remaining_time": "6h 46m 51s"}
+{"eval_loss": 0.92826158, "eval_token_acc": 0.77293847, "eval_runtime": 22.0765, "eval_samples_per_second": 0.951, "eval_steps_per_second": 0.136, "epoch": 0.03816794, "global_step/max_steps": "10/786", "percentage": "1.27%", "elapsed_time": "5m 36s", "remaining_time": "7h 15m 24s"}
+{"loss": 0.78038925, "token_acc": 0.75861374, "grad_norm": 0.60470217, "learning_rate": 1.93e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.028719, "epoch": 0.04198473, "global_step/max_steps": "11/786", "percentage": "1.40%", "elapsed_time": "6m 16s", "remaining_time": "7h 21m 34s"}
+{"loss": 1.00450289, "token_acc": 0.66912442, "grad_norm": 7.0186057, "learning_rate": 2.1e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.028992, "epoch": 0.04580153, "global_step/max_steps": "12/786", "percentage": "1.53%", "elapsed_time": "6m 46s", "remaining_time": "7h 17m 27s"}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aeccf90f0cf016eeaa0434c60da1524c9bda422b99ccb5977d409c47ad9a87b1
+oid sha256:75663f4fdbefe8e0f9e23be9d80f6a5e67f2bc2f1427ceccc1edd3618b020c33
 size 8120