Training in progress, step 10

adapter_config.json

@@ -12,7 +12,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 32,
+  "lora_alpha": 64,
   "lora_bias": false,
   "lora_dropout": 0.05,
   "megatron_config": null,
@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
-    "k_proj",
     "q_proj",
     "gate_proj",
-    "up_proj",
+    "v_proj",
     "o_proj",
-    "v_proj"
+    "up_proj",
+    "k_proj",
+    "down_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d5af0cff5fc1605dcb6d70cdb035216d6e30b8a0c9b02bff722d4c59ba18c47
+oid sha256:b3b540ffd8ebcee47d535e8e3a7bf89de0c2a0c62c396e9e5f8f2ad2c22e5577
 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/v4-20250126-221202",
+  "output_dir": "/root/dataDisk/hakob/output/v6-20250127-000818",
   "overwrite_output_dir": false,
   "do_train": false,
   "do_eval": false,
@@ -81,11 +81,11 @@
   "per_device_eval_batch_size": 1,
   "per_gpu_train_batch_size": null,
   "per_gpu_eval_batch_size": null,
-  "gradient_accumulation_steps": 1,
+  "gradient_accumulation_steps": 32,
   "eval_accumulation_steps": null,
   "eval_delay": 0,
   "torch_empty_cache_steps": null,
-  "learning_rate": 7e-06,
+  "learning_rate": 7e-05,
   "weight_decay": 0.01,
   "adam_beta1": 0.9,
   "adam_beta2": 0.999,
@@ -100,13 +100,13 @@
   "log_level": "passive",
   "log_level_replica": "warning",
   "log_on_each_node": true,
-  "logging_dir": "/root/dataDisk/hakob/output/v4-20250126-221202/runs",
+  "logging_dir": "/root/dataDisk/hakob/output/v6-20250127-000818/runs",
   "logging_strategy": "steps",
   "logging_first_step": true,
   "logging_steps": 1,
   "logging_nan_inf_filter": true,
   "save_strategy": "steps",
-  "save_steps": 5.0,
+  "save_steps": 10.0,
   "save_total_limit": 2,
   "save_safetensors": true,
   "save_on_each_node": false,
@@ -130,7 +130,7 @@
   "tpu_metrics_debug": false,
   "debug": null,
   "dataloader_drop_last": false,
-  "eval_steps": 5.0,
+  "eval_steps": 10.0,
   "dataloader_num_workers": 0,
   "dataloader_prefetch_factor": null,
   "past_index": -1,
@@ -259,7 +259,7 @@
   "target_regex": null,
   "modules_to_save": [],
   "lora_rank": 64,
-  "lora_alpha": 32,
+  "lora_alpha": 64,
   "lora_dropout": 0.05,
   "lora_bias": "none",
   "lora_dtype": null,
@@ -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 0x7f5ce8ded6c0>, 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 0x7fa1333f56c0>, 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/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')"
+  "training_args": "Seq2SeqTrainingArguments(output_dir='/root/dataDisk/hakob/output/v6-20250127-000818', 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-05, 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/v6-20250127-000818/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=10, 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=10, dataloader_num_workers=0, dataloader_prefetch_factor=None, past_index=-1, run_name='/root/dataDisk/hakob/output/v6-20250127-000818', 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

@@ -1,32 +1,11 @@
-{"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"}
-{"loss": 1.2894696, "token_acc": 0.818409, "grad_norm": 15.69610859, "learning_rate": 2.28e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029284, "epoch": 0.04961832, "global_step/max_steps": "13/786", "percentage": "1.65%", "elapsed_time": "7m 16s", "remaining_time": "7h 13m 2s"}
-{"loss": 0.91177201, "token_acc": 0.76051188, "grad_norm": 19.4416792, "learning_rate": 2.45e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029532, "epoch": 0.05343511, "global_step/max_steps": "14/786", "percentage": "1.78%", "elapsed_time": "7m 47s", "remaining_time": "7h 9m 17s"}
-{"loss": 0.83453828, "token_acc": 0.76897133, "grad_norm": 0.14502233, "learning_rate": 2.62e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.030603, "epoch": 0.05725191, "global_step/max_steps": "15/786", "percentage": "1.91%", "elapsed_time": "8m 3s", "remaining_time": "6h 53m 55s"}
-{"eval_loss": 0.89954847, "eval_token_acc": 0.77380728, "eval_runtime": 22.059, "eval_samples_per_second": 0.952, "eval_steps_per_second": 0.136, "epoch": 0.05725191, "global_step/max_steps": "15/786", "percentage": "1.91%", "elapsed_time": "8m 25s", "remaining_time": "7h 12m 49s"}
-{"loss": 0.99048138, "token_acc": 0.70949721, "grad_norm": 2.69842438, "learning_rate": 2.8e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029029, "epoch": 0.0610687, "global_step/max_steps": "16/786", "percentage": "2.04%", "elapsed_time": "9m 4s", "remaining_time": "7h 16m 29s"}
-{"loss": 0.98029929, "token_acc": 0.63352827, "grad_norm": 2.74421964, "learning_rate": 2.97e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029253, "epoch": 0.0648855, "global_step/max_steps": "17/786", "percentage": "2.16%", "elapsed_time": "9m 34s", "remaining_time": "7h 12m 52s"}
-{"loss": 1.27479923, "token_acc": 0.73897059, "grad_norm": 2.36793428, "learning_rate": 3.15e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029456, "epoch": 0.06870229, "global_step/max_steps": "18/786", "percentage": "2.29%", "elapsed_time": "10m 4s", "remaining_time": "7h 9m 35s"}
-{"loss": 0.79431921, "token_acc": 0.74406332, "grad_norm": 0.40841569, "learning_rate": 3.32e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.030149, "epoch": 0.07251908, "global_step/max_steps": "19/786", "percentage": "2.42%", "elapsed_time": "10m 23s", "remaining_time": "6h 59m 18s"}
-{"loss": 0.72544003, "token_acc": 0.76136364, "grad_norm": 8.14447521, "learning_rate": 3.5e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.030299, "epoch": 0.07633588, "global_step/max_steps": "20/786", "percentage": "2.54%", "elapsed_time": "10m 53s", "remaining_time": "6h 56m 54s"}
-{"eval_loss": 0.93264651, "eval_token_acc": 0.77342953, "eval_runtime": 22.0341, "eval_samples_per_second": 0.953, "eval_steps_per_second": 0.136, "epoch": 0.07633588, "global_step/max_steps": "20/786", "percentage": "2.54%", "elapsed_time": "11m 15s", "remaining_time": "7h 10m 58s"}
-{"loss": 1.03133607, "token_acc": 0.75963771, "grad_norm": 10.5853401, "learning_rate": 3.67e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029077, "epoch": 0.08015267, "global_step/max_steps": "21/786", "percentage": "2.67%", "elapsed_time": "11m 55s", "remaining_time": "7h 14m 15s"}
-{"loss": 0.90063345, "token_acc": 0.79555237, "grad_norm": 4.18292453, "learning_rate": 3.85e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029268, "epoch": 0.08396947, "global_step/max_steps": "22/786", "percentage": "2.80%", "elapsed_time": "12m 24s", "remaining_time": "7h 11m 1s"}
-{"loss": 1.25162315, "token_acc": 0.74662669, "grad_norm": 6.2844299, "learning_rate": 4.02e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029426, "epoch": 0.08778626, "global_step/max_steps": "23/786", "percentage": "2.93%", "elapsed_time": "12m 54s", "remaining_time": "7h 8m 17s"}
-{"loss": 0.91385037, "token_acc": 0.79717587, "grad_norm": 14.13236802, "learning_rate": 4.2e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029573, "epoch": 0.09160305, "global_step/max_steps": "24/786", "percentage": "3.05%", "elapsed_time": "13m 24s", "remaining_time": "7h 5m 45s"}
-{"loss": 0.80503768, "token_acc": 0.75471698, "grad_norm": 0.57895672, "learning_rate": 4.37e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029711, "epoch": 0.09541985, "global_step/max_steps": "25/786", "percentage": "3.18%", "elapsed_time": "13m 54s", "remaining_time": "7h 3m 21s"}
-{"eval_loss": 0.92728376, "eval_token_acc": 0.77350508, "eval_runtime": 22.14, "eval_samples_per_second": 0.949, "eval_steps_per_second": 0.136, "epoch": 0.09541985, "global_step/max_steps": "25/786", "percentage": "3.18%", "elapsed_time": "14m 16s", "remaining_time": "7h 14m 35s"}
-{"loss": 0.91751623, "token_acc": 0.77953335, "grad_norm": 125.80540284, "learning_rate": 4.55e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.02879, "epoch": 0.09923664, "global_step/max_steps": "26/786", "percentage": "3.31%", "elapsed_time": "14m 56s", "remaining_time": "7h 16m 34s"}
-{"loss": 0.80560565, "token_acc": 0.76595745, "grad_norm": 0.16683542, "learning_rate": 4.73e-06, "memory(GiB)": 88.62, "train_speed(iter/s)": 0.029367, "epoch": 0.10305344, "global_step/max_steps": "27/786", "percentage": "3.44%", "elapsed_time": "15m 12s", "remaining_time": "7h 7m 29s"}
+{"loss": 30.96281052, "token_acc": 0.74524095, "grad_norm": 7.9422058, "learning_rate": 3.5e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001132, "epoch": 0.1221374, "global_step/max_steps": "1/24", "percentage": "4.17%", "elapsed_time": "14m 35s", "remaining_time": "5h 35m 38s"}
+{"loss": 32.10971832, "token_acc": 0.72850925, "grad_norm": 6.52010408, "learning_rate": 7e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001129, "epoch": 0.24427481, "global_step/max_steps": "2/24", "percentage": "8.33%", "elapsed_time": "29m 24s", "remaining_time": "5h 23m 26s"}
+{"loss": 30.22758484, "token_acc": 0.74227122, "grad_norm": 10.43869129, "learning_rate": 6.964e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001123, "epoch": 0.36641221, "global_step/max_steps": "3/24", "percentage": "12.50%", "elapsed_time": "44m 24s", "remaining_time": "5h 10m 50s"}
+{"loss": 27.7996254, "token_acc": 0.76933951, "grad_norm": 1.79717579, "learning_rate": 6.858e-05, "memory(GiB)": 91.74, "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 50s"}
+{"loss": 26.51483536, "token_acc": 0.77443033, "grad_norm": 0.42850104, "learning_rate": 6.684e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001133, "epoch": 0.61068702, "global_step/max_steps": "5/24", "percentage": "20.83%", "elapsed_time": "1h 13m 25s", "remaining_time": "4h 39m 1s"}
+{"loss": 25.06520462, "token_acc": 0.78730703, "grad_norm": 0.40429934, "learning_rate": 6.444e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001125, "epoch": 0.73282443, "global_step/max_steps": "6/24", "percentage": "25.00%", "elapsed_time": "1h 28m 47s", "remaining_time": "4h 26m 22s"}
+{"loss": 23.57491112, "token_acc": 0.8091305, "grad_norm": 0.17600205, "learning_rate": 6.145e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001128, "epoch": 0.85496183, "global_step/max_steps": "7/24", "percentage": "29.17%", "elapsed_time": "1h 43m 17s", "remaining_time": "4h 10m 51s"}
+{"loss": 23.25542068, "token_acc": 0.80147298, "grad_norm": 0.1694329, "learning_rate": 5.792e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001127, "epoch": 0.97709924, "global_step/max_steps": "8/24", "percentage": "33.33%", "elapsed_time": "1h 58m 13s", "remaining_time": "3h 56m 26s"}
+{"loss": 4.33601952, "token_acc": 0.78835404, "grad_norm": 0.1694329, "learning_rate": 5.392e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001237, "epoch": 1.0, "global_step/max_steps": "9/24", "percentage": "37.50%", "elapsed_time": "2h 1m 8s", "remaining_time": "3h 21m 53s"}
+{"loss": 21.8188076, "token_acc": 0.79576805, "grad_norm": 0.14743451, "learning_rate": 4.954e-05, "memory(GiB)": 91.74, "train_speed(iter/s)": 0.001229, "epoch": 1.1221374, "global_step/max_steps": "10/24", "percentage": "41.67%", "elapsed_time": "2h 15m 30s", "remaining_time": "3h 9m 42s"}
+{"eval_loss": 0.6696381, "eval_token_acc": 0.81154384, "eval_runtime": 21.8063, "eval_samples_per_second": 0.963, "eval_steps_per_second": 0.138, "epoch": 1.1221374, "global_step/max_steps": "10/24", "percentage": "41.67%", "elapsed_time": "2h 15m 52s", "remaining_time": "3h 10m 13s"}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:75663f4fdbefe8e0f9e23be9d80f6a5e67f2bc2f1427ceccc1edd3618b020c33
+oid sha256:e2939157cd90e2392859bfd03b1841ec3163f353049d6720f5f85d7b69c9c038
 size 8120