Training in progress, step 10

adapter_config.json

@@ -14,7 +14,7 @@
   "loftq_config": {},
   "lora_alpha": 64,
   "lora_bias": false,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.2,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": [],
@@ -24,12 +24,12 @@
   "revision": null,
   "target_modules": [
     "down_proj",
+    "v_proj",
+    "q_proj",
     "gate_proj",
-    "k_proj",
     "o_proj",
-    "q_proj",
-    "v_proj",
-    "up_proj"
+    "up_proj",
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9094dd1114c70436de0a0e319fd0588b1cd44566c485099bcac708952c3f37ce
+oid sha256:e617914e9f0f26258ec0f468c79d22b3729ad7630e08635f5c27565958fec453
 size 6627156248

args.json

@@ -21,10 +21,10 @@
   "use_chat_template": true,
   "template_backend": "swift",
   "dataset": [
-    "tool-merged.jsonl"
+    "clean.jsonl"
   ],
   "val_dataset": [],
-  "split_dataset_ratio": 0.01,
+  "split_dataset_ratio": 0.03,
   "data_seed": 42,
   "dataset_num_proc": 1,
   "streaming": false,
@@ -72,7 +72,7 @@
   "custom_register_path": [],
   "ignore_args_error": false,
   "use_swift_lora": false,
-  "output_dir": "/root/dataDisk/llama-v1/output/v21-20250202-125059",
+  "output_dir": "/root/dataDisk/llama-v1/output/v24-20250203-164956",
   "overwrite_output_dir": false,
   "do_train": false,
   "do_eval": false,
@@ -87,7 +87,7 @@
   "eval_accumulation_steps": null,
   "eval_delay": 0,
   "torch_empty_cache_steps": null,
-  "learning_rate": 3e-05,
+  "learning_rate": 5e-05,
   "weight_decay": 0.01,
   "adam_beta1": 0.9,
   "adam_beta2": 0.999,
@@ -97,12 +97,12 @@
   "max_steps": -1,
   "lr_scheduler_type": "linear",
   "lr_scheduler_kwargs": null,
-  "warmup_ratio": 0.01,
+  "warmup_ratio": 0.1,
   "warmup_steps": 0,
   "log_level": "passive",
   "log_level_replica": "warning",
   "log_on_each_node": true,
-  "logging_dir": "/root/dataDisk/llama-v1/output/v21-20250202-125059/runs",
+  "logging_dir": "/root/dataDisk/llama-v1/output/v24-20250203-164956/runs",
   "logging_strategy": "steps",
   "logging_first_step": true,
   "logging_steps": 1,
@@ -262,7 +262,7 @@
   "modules_to_save": [],
   "lora_rank": 256,
   "lora_alpha": 64,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.2,
   "lora_bias": "none",
   "lora_dtype": null,
   "lorap_lr_ratio": null,
@@ -366,8 +366,8 @@
   "local_world_size": 8,
   "model_suffix": "llama-model",
   "model_info": "ModelInfo(model_type='llama3_2', model_dir='/root/dataDisk/llama-model', torch_dtype=torch.bfloat16, max_model_len=128000, quant_method=None, quant_bits=None, config=None, task_type='causal_lm', num_labels=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 0x7f716f69eb00>, model_arch='llama', architectures=['LlamaForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=False, is_reward=False, task_type=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 0x7f29fda56b00>, model_arch='llama', architectures=['LlamaForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=False, is_reward=False, task_type=None, ignore_patterns=[], requires=['transformers>=4.45'], tags=[])",
   "model_dir": "/root/dataDisk/llama-model",
   "hub": "<class 'swift.hub.hub.HFHub'>",
-  "training_args": "ORPOConfig(output_dir='/root/dataDisk/llama-v1/output/v21-20250202-125059', 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=4, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=3e-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.LINEAR: 'linear'>, lr_scheduler_kwargs=None, warmup_ratio=0.01, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/root/dataDisk/llama-v1/output/v21-20250202-125059/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/llama-v1/output/v21-20250202-125059', 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': 'cpu', '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=['wandb'], 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/LLAMA-3.3-70B-Reasoning', 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, max_length=22000, max_prompt_length=None, max_completion_length=None, beta=0.1, disable_dropout=True, label_pad_token_id=None, padding_value=None, truncation_mode='keep_end', generate_during_eval=False, is_encoder_decoder=False, model_init_kwargs=None, dataset_num_proc=1, acc_strategy='token', sequence_parallel_size=1, check_model=True, train_sampler_random=True, metric_warmup_step=0, train_dataset_sample=-1, fsdp_num=1, acc_steps=1, train_type='lora', optimizer=None, galore_config=None)"
+  "training_args": "ORPOConfig(output_dir='/root/dataDisk/llama-v1/output/v24-20250203-164956', 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=4, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=5e-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.LINEAR: 'linear'>, lr_scheduler_kwargs=None, warmup_ratio=0.1, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/root/dataDisk/llama-v1/output/v24-20250203-164956/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/llama-v1/output/v24-20250203-164956', 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': 'cpu', '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=['wandb'], 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/LLAMA-3.3-70B-Reasoning', 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, max_length=22000, max_prompt_length=None, max_completion_length=None, beta=0.1, disable_dropout=True, label_pad_token_id=None, padding_value=None, truncation_mode='keep_end', generate_during_eval=False, is_encoder_decoder=False, model_init_kwargs=None, dataset_num_proc=1, acc_strategy='token', sequence_parallel_size=1, check_model=True, train_sampler_random=True, metric_warmup_step=0, train_dataset_sample=-1, fsdp_num=1, acc_steps=1, train_type='lora', optimizer=None, galore_config=None)"
 }

logging.jsonl

@@ -1,33 +1,12 @@
-{"loss": 1.70019531, "grad_norm": 1.46354806, "learning_rate": 5e-06, "memory(GiB)": 89.75, "train_speed(iter/s)": 0.002594, "rewards/chosen": -0.11590576, "rewards/rejected": -0.12602234, "rewards/accuracies": 0.75, "rewards/margins": 0.01013184, "logps/rejected": -1.25976562, "logps/chosen": -1.15869141, "logits/rejected": -0.2199707, "logits/chosen": -0.29058838, "nll_loss": 1.15478516, "log_odds_ratio": -0.78955078, "log_odds_chosen": -0.07446289, "epoch": 0.00566572, "global_step/max_steps": "1/528", "percentage": "0.19%", "elapsed_time": "6m 7s", "remaining_time": "2d 5h 48m 59s"}
-{"loss": 1.8729248, "grad_norm": 3.37244034, "learning_rate": 1e-05, "memory(GiB)": 123.97, "train_speed(iter/s)": 0.00289, "rewards/chosen": -0.18859863, "rewards/rejected": -0.1776123, "rewards/accuracies": 0.5, "rewards/margins": -0.01098633, "logps/rejected": -1.77734375, "logps/chosen": -1.8828125, "logits/rejected": -0.31152344, "logits/chosen": -0.31970215, "nll_loss": 1.88183594, "log_odds_ratio": -0.7578125, "log_odds_chosen": -0.10546875, "epoch": 0.01133144, "global_step/max_steps": "2/528", "percentage": "0.38%", "elapsed_time": "11m 14s", "remaining_time": "2d 1h 15m 34s"}
-{"loss": 1.54760742, "grad_norm": 0.71019006, "learning_rate": 1.5e-05, "memory(GiB)": 123.97, "train_speed(iter/s)": 0.002959, "rewards/chosen": -0.11352539, "rewards/rejected": -0.11279297, "rewards/accuracies": 0.75, "rewards/margins": -0.00073242, "logps/rejected": -1.12695312, "logps/chosen": -1.13574219, "logits/rejected": -0.20922852, "logits/chosen": -0.14807129, "nll_loss": 1.13476562, "log_odds_ratio": -0.6953125, "log_odds_chosen": 0.03125, "epoch": 0.01699717, "global_step/max_steps": "3/528", "percentage": "0.57%", "elapsed_time": "16m 35s", "remaining_time": "2d 0h 24m 54s"}
-{"loss": 1.49133301, "grad_norm": 0.48966399, "learning_rate": 2e-05, "memory(GiB)": 132.83, "train_speed(iter/s)": 0.003052, "rewards/chosen": -0.15979004, "rewards/rejected": -0.14248657, "rewards/accuracies": 0.5, "rewards/margins": -0.01733398, "logps/rejected": -1.42358398, "logps/chosen": -1.59960938, "logits/rejected": -0.46899414, "logits/chosen": -0.51147461, "nll_loss": 1.59082031, "log_odds_ratio": -1.04589844, "log_odds_chosen": -0.48974609, "epoch": 0.02266289, "global_step/max_steps": "4/528", "percentage": "0.76%", "elapsed_time": "21m 32s", "remaining_time": "1d 23h 2m 44s"}
-{"loss": 1.50598145, "grad_norm": 0.45412952, "learning_rate": 2.5e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002891, "rewards/chosen": -0.23962402, "rewards/rejected": -0.2175293, "rewards/accuracies": 0.5, "rewards/margins": -0.02209473, "logps/rejected": -2.17773438, "logps/chosen": -2.39941406, "logits/rejected": -0.80786133, "logits/chosen": -0.75708008, "nll_loss": 2.39941406, "log_odds_ratio": -0.80078125, "log_odds_chosen": -0.14013672, "epoch": 0.02832861, "global_step/max_steps": "5/528", "percentage": "0.95%", "elapsed_time": "28m 31s", "remaining_time": "2d 1h 44m 3s"}
-{"loss": 1.4644165, "grad_norm": 0.53225052, "learning_rate": 3e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002913, "rewards/chosen": -0.11218262, "rewards/rejected": -0.09570312, "rewards/accuracies": 0.5, "rewards/margins": -0.01647949, "logps/rejected": -0.95800781, "logps/chosen": -1.12011719, "logits/rejected": -0.35595703, "logits/chosen": -0.33984375, "nll_loss": 1.12109375, "log_odds_ratio": -0.79882812, "log_odds_chosen": -0.16699219, "epoch": 0.03399433, "global_step/max_steps": "6/528", "percentage": "1.14%", "elapsed_time": "34m 1s", "remaining_time": "2d 1h 20m 31s"}
-{"loss": 0.96185303, "grad_norm": 0.50674963, "learning_rate": 2.994e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.00296, "rewards/chosen": -0.06665039, "rewards/rejected": -0.05834961, "rewards/accuracies": 0.75, "rewards/margins": -0.00830078, "logps/rejected": -0.58398438, "logps/chosen": -0.66748047, "logits/rejected": -0.40087891, "logits/chosen": -0.35864258, "nll_loss": 0.66455078, "log_odds_ratio": -0.74023438, "log_odds_chosen": -0.06787109, "epoch": 0.03966006, "global_step/max_steps": "7/528", "percentage": "1.33%", "elapsed_time": "39m 6s", "remaining_time": "2d 0h 31m 19s"}
-{"loss": 0.63311768, "grad_norm": 0.28535429, "learning_rate": 2.989e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.0031, "rewards/chosen": -0.15863037, "rewards/rejected": -0.15472412, "rewards/accuracies": 0.5, "rewards/margins": -0.00390625, "logps/rejected": -1.54541016, "logps/chosen": -1.59008789, "logits/rejected": -0.64941406, "logits/chosen": -0.63623047, "nll_loss": 1.58959961, "log_odds_ratio": -0.70996094, "log_odds_chosen": -0.01196289, "epoch": 0.04532578, "global_step/max_steps": "8/528", "percentage": "1.52%", "elapsed_time": "42m 42s", "remaining_time": "1d 22h 16m 16s"}
-{"loss": 0.64984131, "grad_norm": 0.63884169, "learning_rate": 2.983e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.003143, "rewards/chosen": -0.04238892, "rewards/rejected": -0.04241943, "rewards/accuracies": 0.5, "rewards/margins": 3.052e-05, "logps/rejected": -0.42382812, "logps/chosen": -0.42382812, "logits/rejected": -0.47265625, "logits/chosen": -0.44482422, "nll_loss": 0.42382812, "log_odds_ratio": -0.70898438, "log_odds_chosen": -0.00585938, "epoch": 0.0509915, "global_step/max_steps": "9/528", "percentage": "1.70%", "elapsed_time": "47m 25s", "remaining_time": "1d 21h 34m 32s"}
-{"loss": 0.83450317, "grad_norm": 0.63739359, "learning_rate": 2.977e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002998, "rewards/chosen": -0.12145996, "rewards/rejected": -0.09274292, "rewards/accuracies": 0.0, "rewards/margins": -0.02871704, "logps/rejected": -0.92749023, "logps/chosen": -1.21533203, "logits/rejected": -0.75292969, "logits/chosen": -0.74707031, "nll_loss": 1.21142578, "log_odds_ratio": -0.91699219, "log_odds_chosen": -0.38891602, "epoch": 0.05665722, "global_step/max_steps": "10/528", "percentage": "1.89%", "elapsed_time": "55m 17s", "remaining_time": "1d 23h 44m 6s"}
-{"eval_loss": 0.68448466, "eval_runtime": 214.9852, "eval_samples_per_second": 0.265, "eval_steps_per_second": 0.037, "eval_rewards/chosen": -0.05220795, "eval_rewards/rejected": -0.05303955, "eval_rewards/accuracies": 0.625, "eval_rewards/margins": 0.0008316, "eval_logps/rejected": -0.53039551, "eval_logps/chosen": -0.52233887, "eval_logits/rejected": -0.55395508, "eval_logits/chosen": -0.52734375, "eval_nll_loss": 0.52227783, "eval_log_odds_ratio": -0.65014648, "eval_log_odds_chosen": 0.10693359, "epoch": 0.05665722, "global_step/max_steps": "10/528", "percentage": "1.89%", "elapsed_time": "58m 52s", "remaining_time": "2d 2h 49m 42s"}
-{"loss": 0.65496826, "grad_norm": 0.65517849, "learning_rate": 2.971e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002752, "rewards/chosen": -0.02752686, "rewards/rejected": -0.03305054, "rewards/accuracies": 1.0, "rewards/margins": 0.00552368, "logps/rejected": -0.33007812, "logps/chosen": -0.27539062, "logits/rejected": -0.50341797, "logits/chosen": -0.44140625, "nll_loss": 0.27587891, "log_odds_ratio": -0.59179688, "log_odds_chosen": 0.21484375, "epoch": 0.06232295, "global_step/max_steps": "11/528", "percentage": "2.08%", "elapsed_time": "1h 6m 18s", "remaining_time": "2d 3h 56m 49s"}
-{"loss": 0.90829468, "grad_norm": 0.96871412, "learning_rate": 2.966e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002783, "rewards/chosen": -0.05822754, "rewards/rejected": -0.05548096, "rewards/accuracies": 0.5, "rewards/margins": -0.00274658, "logps/rejected": -0.55517578, "logps/chosen": -0.58276367, "logits/rejected": -0.49023438, "logits/chosen": -0.46630859, "nll_loss": 0.58325195, "log_odds_ratio": -0.65429688, "log_odds_chosen": 0.11669922, "epoch": 0.06798867, "global_step/max_steps": "12/528", "percentage": "2.27%", "elapsed_time": "1h 11m 33s", "remaining_time": "2d 3h 17m 11s"}
-{"loss": 0.82220459, "grad_norm": 1.00971079, "learning_rate": 2.96e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002834, "rewards/chosen": -0.04037476, "rewards/rejected": -0.04879761, "rewards/accuracies": 0.75, "rewards/margins": 0.00842285, "logps/rejected": -0.48730469, "logps/chosen": -0.40380859, "logits/rejected": -0.41308594, "logits/chosen": -0.3828125, "nll_loss": 0.40356445, "log_odds_ratio": -0.59277344, "log_odds_chosen": 0.21899414, "epoch": 0.07365439, "global_step/max_steps": "13/528", "percentage": "2.46%", "elapsed_time": "1h 16m 8s", "remaining_time": "2d 2h 16m 26s"}
-{"loss": 0.87780762, "grad_norm": 0.95921201, "learning_rate": 2.954e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002892, "rewards/chosen": -0.06549072, "rewards/rejected": -0.07128906, "rewards/accuracies": 0.75, "rewards/margins": 0.00579834, "logps/rejected": -0.71289062, "logps/chosen": -0.65551758, "logits/rejected": -0.53320312, "logits/chosen": -0.49316406, "nll_loss": 0.65649414, "log_odds_ratio": -0.64550781, "log_odds_chosen": 0.1027832, "epoch": 0.07932011, "global_step/max_steps": "14/528", "percentage": "2.65%", "elapsed_time": "1h 20m 22s", "remaining_time": "2d 1h 11m 6s"}
-{"loss": 0.47705078, "grad_norm": 0.24082753, "learning_rate": 2.948e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.00292, "rewards/chosen": -0.02780151, "rewards/rejected": -0.03814697, "rewards/accuracies": 0.75, "rewards/margins": 0.01034546, "logps/rejected": -0.38232422, "logps/chosen": -0.27832031, "logits/rejected": -0.51171875, "logits/chosen": -0.51464844, "nll_loss": 0.27832031, "log_odds_ratio": -0.54785156, "log_odds_chosen": 0.34594727, "epoch": 0.08498584, "global_step/max_steps": "15/528", "percentage": "2.84%", "elapsed_time": "1h 25m 19s", "remaining_time": "2d 0h 37m 56s"}
-{"loss": 0.75949097, "grad_norm": 0.40383622, "learning_rate": 2.943e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002884, "rewards/chosen": -0.02700806, "rewards/rejected": -0.03167725, "rewards/accuracies": 1.0, "rewards/margins": 0.00466919, "logps/rejected": -0.31713867, "logps/chosen": -0.27001953, "logits/rejected": -0.47021484, "logits/chosen": -0.42871094, "nll_loss": 0.26953125, "log_odds_ratio": -0.61035156, "log_odds_chosen": 0.18188477, "epoch": 0.09065156, "global_step/max_steps": "16/528", "percentage": "3.03%", "elapsed_time": "1h 32m 10s", "remaining_time": "2d 1h 9m 29s"}
-{"loss": 0.6574707, "grad_norm": 0.29921553, "learning_rate": 2.937e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002857, "rewards/chosen": -0.02571106, "rewards/rejected": -0.03115845, "rewards/accuracies": 0.75, "rewards/margins": 0.00544739, "logps/rejected": -0.31176758, "logps/chosen": -0.25683594, "logits/rejected": -0.53759766, "logits/chosen": -0.49902344, "nll_loss": 0.25683594, "log_odds_ratio": -0.58984375, "log_odds_chosen": 0.23999023, "epoch": 0.09631728, "global_step/max_steps": "17/528", "percentage": "3.22%", "elapsed_time": "1h 38m 51s", "remaining_time": "2d 1h 31m 41s"}
-{"loss": 0.61532593, "grad_norm": 0.58009589, "learning_rate": 2.931e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002855, "rewards/chosen": -0.07330322, "rewards/rejected": -0.06774902, "rewards/accuracies": 0.5, "rewards/margins": -0.0055542, "logps/rejected": -0.67700195, "logps/chosen": -0.73291016, "logits/rejected": -0.67382812, "logits/chosen": -0.66601562, "nll_loss": 0.7331543, "log_odds_ratio": -0.70019531, "log_odds_chosen": -0.00219727, "epoch": 0.101983, "global_step/max_steps": "18/528", "percentage": "3.41%", "elapsed_time": "1h 44m 47s", "remaining_time": "2d 1h 29m 2s"}
-{"loss": 0.87954712, "grad_norm": 1.18141782, "learning_rate": 2.925e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002877, "rewards/chosen": -0.0887146, "rewards/rejected": -0.08203125, "rewards/accuracies": 0.75, "rewards/margins": -0.00668335, "logps/rejected": -0.8203125, "logps/chosen": -0.88525391, "logits/rejected": -0.74121094, "logits/chosen": -0.6796875, "nll_loss": 0.88623047, "log_odds_ratio": -0.65039062, "log_odds_chosen": 0.14306641, "epoch": 0.10764873, "global_step/max_steps": "19/528", "percentage": "3.60%", "elapsed_time": "1h 49m 46s", "remaining_time": "2d 1h 0m 41s"}
-{"loss": 0.4979248, "grad_norm": 0.49843666, "learning_rate": 2.92e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002855, "rewards/chosen": -0.05352783, "rewards/rejected": -0.04724121, "rewards/accuracies": 0.75, "rewards/margins": -0.00628662, "logps/rejected": -0.47216797, "logps/chosen": -0.53515625, "logits/rejected": -0.65527344, "logits/chosen": -0.63183594, "nll_loss": 0.53515625, "log_odds_ratio": -0.74023438, "log_odds_chosen": -0.04248047, "epoch": 0.11331445, "global_step/max_steps": "20/528", "percentage": "3.79%", "elapsed_time": "1h 56m 28s", "remaining_time": "2d 1h 18m 18s"}
-{"eval_loss": 0.50023127, "eval_runtime": 214.6405, "eval_samples_per_second": 0.266, "eval_steps_per_second": 0.037, "eval_rewards/chosen": -0.03556061, "eval_rewards/rejected": -0.03941345, "eval_rewards/accuracies": 0.75, "eval_rewards/margins": 0.00385284, "eval_logps/rejected": -0.39416504, "eval_logps/chosen": -0.35577393, "eval_logits/rejected": -0.78125, "eval_logits/chosen": -0.73828125, "eval_nll_loss": 0.35552979, "eval_log_odds_ratio": -0.59912109, "eval_log_odds_chosen": 0.22839355, "epoch": 0.11331445, "global_step/max_steps": "20/528", "percentage": "3.79%", "elapsed_time": "2h 0m 2s", "remaining_time": "2d 2h 49m 10s"}
-{"loss": 0.51702881, "grad_norm": 0.73438716, "learning_rate": 2.914e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002801, "rewards/chosen": -0.03884888, "rewards/rejected": -0.03894043, "rewards/accuracies": 0.5, "rewards/margins": 9.155e-05, "logps/rejected": -0.38989258, "logps/chosen": -0.38867188, "logits/rejected": -0.74121094, "logits/chosen": -0.72851562, "nll_loss": 0.3894043, "log_odds_ratio": -0.64941406, "log_odds_chosen": 0.1472168, "epoch": 0.11898017, "global_step/max_steps": "21/528", "percentage": "3.98%", "elapsed_time": "2h 4m 40s", "remaining_time": "2d 2h 10m 0s"}
-{"loss": 0.81253052, "grad_norm": 1.53866589, "learning_rate": 2.908e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002805, "rewards/chosen": -0.04421997, "rewards/rejected": -0.04016113, "rewards/accuracies": 0.75, "rewards/margins": -0.00405884, "logps/rejected": -0.40136719, "logps/chosen": -0.44238281, "logits/rejected": -0.75976562, "logits/chosen": -0.69824219, "nll_loss": 0.44238281, "log_odds_ratio": -0.68554688, "log_odds_chosen": 0.10449219, "epoch": 0.12464589, "global_step/max_steps": "22/528", "percentage": "4.17%", "elapsed_time": "2h 10m 23s", "remaining_time": "2d 1h 59m 10s"}
-{"loss": 0.39785767, "grad_norm": 4.08225965, "learning_rate": 2.902e-05, "memory(GiB)": 133.21, "train_speed(iter/s)": 0.002831, "rewards/chosen": -0.02941895, "rewards/rejected": -0.03512573, "rewards/accuracies": 0.75, "rewards/margins": 0.00570679, "logps/rejected": -0.35107422, "logps/chosen": -0.29370117, "logits/rejected": -0.48242188, "logits/chosen": -0.44238281, "nll_loss": 0.29345703, "log_odds_ratio": -0.58300781, "log_odds_chosen": 0.24072266, "epoch": 0.13031161, "global_step/max_steps": "23/528", "percentage": "4.36%", "elapsed_time": "2h 15m 5s", "remaining_time": "2d 1h 26m 7s"}
-{"loss": 0.50842285, "grad_norm": 0.75203818, "learning_rate": 2.897e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002829, "rewards/chosen": -0.04470825, "rewards/rejected": -0.04074097, "rewards/accuracies": 0.25, "rewards/margins": -0.00396729, "logps/rejected": -0.4074707, "logps/chosen": -0.44677734, "logits/rejected": -0.41992188, "logits/chosen": -0.47167969, "nll_loss": 0.44775391, "log_odds_ratio": -0.80273438, "log_odds_chosen": -0.15820312, "epoch": 0.13597734, "global_step/max_steps": "24/528", "percentage": "4.55%", "elapsed_time": "2h 21m 7s", "remaining_time": "2d 1h 23m 29s"}
-{"loss": 0.66778564, "grad_norm": 2.3309772, "learning_rate": 2.891e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002853, "rewards/chosen": -0.12059021, "rewards/rejected": -0.14254761, "rewards/accuracies": 0.75, "rewards/margins": 0.02194214, "logps/rejected": -1.42749023, "logps/chosen": -1.20690918, "logits/rejected": -0.37792969, "logits/chosen": -0.37036133, "nll_loss": 1.20678711, "log_odds_ratio": -0.5480957, "log_odds_chosen": 0.44335938, "epoch": 0.14164306, "global_step/max_steps": "25/528", "percentage": "4.73%", "elapsed_time": "2h 25m 43s", "remaining_time": "2d 0h 52m 5s"}
-{"loss": 0.54031372, "grad_norm": 2.24770308, "learning_rate": 2.885e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002837, "rewards/chosen": -0.02548218, "rewards/rejected": -0.02676392, "rewards/accuracies": 0.5, "rewards/margins": 0.00128174, "logps/rejected": -0.26733398, "logps/chosen": -0.25488281, "logits/rejected": -0.53955078, "logits/chosen": -0.43457031, "nll_loss": 0.25463867, "log_odds_ratio": -0.6796875, "log_odds_chosen": 0.05322266, "epoch": 0.14730878, "global_step/max_steps": "26/528", "percentage": "4.92%", "elapsed_time": "2h 32m 28s", "remaining_time": "2d 1h 3m 47s"}
-{"loss": 0.29516602, "grad_norm": 0.13307834, "learning_rate": 2.879e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002869, "rewards/chosen": -0.01672363, "rewards/rejected": -0.02449036, "rewards/accuracies": 1.0, "rewards/margins": 0.00776672, "logps/rejected": -0.24511719, "logps/chosen": -0.16723633, "logits/rejected": -0.44970703, "logits/chosen": -0.40771484, "nll_loss": 0.16699219, "log_odds_ratio": -0.52685547, "log_odds_chosen": 0.38671875, "epoch": 0.1529745, "global_step/max_steps": "27/528", "percentage": "5.11%", "elapsed_time": "2h 36m 31s", "remaining_time": "2d 0h 24m 24s"}
-{"loss": 0.67800903, "grad_norm": 1.45895541, "learning_rate": 2.874e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002851, "rewards/chosen": -0.1439209, "rewards/rejected": -0.1578064, "rewards/accuracies": 1.0, "rewards/margins": 0.0138855, "logps/rejected": -1.57885742, "logps/chosen": -1.43896484, "logits/rejected": -0.57958984, "logits/chosen": -0.55810547, "nll_loss": 1.4387207, "log_odds_ratio": -0.56494141, "log_odds_chosen": 0.28515625, "epoch": 0.15864023, "global_step/max_steps": "28/528", "percentage": "5.30%", "elapsed_time": "2h 43m 24s", "remaining_time": "2d 0h 38m 5s"}
-{"loss": 0.5168457, "grad_norm": 0.36635926, "learning_rate": 2.868e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002861, "rewards/chosen": -0.03540039, "rewards/rejected": -0.04214478, "rewards/accuracies": 1.0, "rewards/margins": 0.00674438, "logps/rejected": -0.421875, "logps/chosen": -0.35351562, "logits/rejected": -0.41064453, "logits/chosen": -0.38623047, "nll_loss": 0.35302734, "log_odds_ratio": -0.58349609, "log_odds_chosen": 0.24853516, "epoch": 0.16430595, "global_step/max_steps": "29/528", "percentage": "5.49%", "elapsed_time": "2h 48m 39s", "remaining_time": "2d 0h 21m 59s"}
-{"loss": 0.54055786, "grad_norm": 2.31381583, "learning_rate": 2.862e-05, "memory(GiB)": 133.31, "train_speed(iter/s)": 0.002872, "rewards/chosen": -0.04043579, "rewards/rejected": -0.04840088, "rewards/accuracies": 1.0, "rewards/margins": 0.00796509, "logps/rejected": -0.484375, "logps/chosen": -0.40429688, "logits/rejected": -0.41455078, "logits/chosen": -0.36621094, "nll_loss": 0.40527344, "log_odds_ratio": -0.57568359, "log_odds_chosen": 0.25878906, "epoch": 0.16997167, "global_step/max_steps": "30/528", "percentage": "5.68%", "elapsed_time": "2h 53m 48s", "remaining_time": "2d 0h 5m 10s"}
-{"eval_loss": 0.42711759, "eval_runtime": 214.3814, "eval_samples_per_second": 0.266, "eval_steps_per_second": 0.037, "eval_rewards/chosen": -0.02880859, "eval_rewards/rejected": -0.03607178, "eval_rewards/accuracies": 1.0, "eval_rewards/margins": 0.00726318, "eval_logps/rejected": -0.36108398, "eval_logps/chosen": -0.28771973, "eval_logits/rejected": -0.53515625, "eval_logits/chosen": -0.4855957, "eval_nll_loss": 0.28778076, "eval_log_odds_ratio": -0.54736328, "eval_log_odds_chosen": 0.34350586, "epoch": 0.16997167, "global_step/max_steps": "30/528", "percentage": "5.68%", "elapsed_time": "2h 57m 22s", "remaining_time": "2d 1h 4m 29s"}
+{"loss": 1.13067627, "grad_norm": 0.64583141, "learning_rate": 5.9e-07, "memory(GiB)": 71.89, "train_speed(iter/s)": 0.003615, "rewards/chosen": -0.07641602, "rewards/rejected": -0.07684326, "rewards/accuracies": 0.75, "rewards/margins": 0.00042725, "logps/rejected": -0.76855469, "logps/chosen": -0.76245117, "logits/rejected": -0.13487244, "logits/chosen": -0.01712036, "nll_loss": 0.76245117, "log_odds_ratio": -0.67871094, "log_odds_chosen": 0.03588867, "epoch": 0.0035461, "global_step/max_steps": "1/846", "percentage": "0.12%", "elapsed_time": "4m 19s", "remaining_time": "2d 13h 0m 27s"}
+{"loss": 1.12945557, "grad_norm": 0.40638471, "learning_rate": 1.18e-06, "memory(GiB)": 132.59, "train_speed(iter/s)": 0.003565, "rewards/chosen": -0.20227051, "rewards/rejected": -0.15795898, "rewards/accuracies": 0.5, "rewards/margins": -0.04431152, "logps/rejected": -1.57910156, "logps/chosen": -2.0234375, "logits/rejected": -0.42211914, "logits/chosen": -0.40869141, "nll_loss": 2.02148438, "log_odds_ratio": -0.98632812, "log_odds_chosen": -0.35058594, "epoch": 0.0070922, "global_step/max_steps": "2/846", "percentage": "0.24%", "elapsed_time": "9m 4s", "remaining_time": "2d 15h 47m 48s"}
+{"loss": 1.15829468, "grad_norm": 0.81067997, "learning_rate": 1.76e-06, "memory(GiB)": 132.59, "train_speed(iter/s)": 0.003583, "rewards/chosen": -0.03308105, "rewards/rejected": -0.02667236, "rewards/accuracies": 0.25, "rewards/margins": -0.00640869, "logps/rejected": -0.26708984, "logps/chosen": -0.33056641, "logits/rejected": -0.19258118, "logits/chosen": -0.13775635, "nll_loss": 0.33056641, "log_odds_ratio": -0.87207031, "log_odds_chosen": -0.28295898, "epoch": 0.0106383, "global_step/max_steps": "3/846", "percentage": "0.35%", "elapsed_time": "13m 40s", "remaining_time": "2d 16h 2m 54s"}
+{"loss": 1.11914062, "grad_norm": 0.91426826, "learning_rate": 2.35e-06, "memory(GiB)": 132.59, "train_speed(iter/s)": 0.003807, "rewards/chosen": -0.04940796, "rewards/rejected": -0.05230713, "rewards/accuracies": 0.5, "rewards/margins": 0.00289917, "logps/rejected": -0.5234375, "logps/chosen": -0.49365234, "logits/rejected": -0.22619629, "logits/chosen": -0.07446289, "nll_loss": 0.49414062, "log_odds_ratio": -0.70898438, "log_odds_chosen": -0.01660156, "epoch": 0.0141844, "global_step/max_steps": "4/846", "percentage": "0.47%", "elapsed_time": "17m 13s", "remaining_time": "2d 12h 27m 13s"}
+{"loss": 1.1842041, "grad_norm": 0.51992393, "learning_rate": 2.94e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.003961, "rewards/chosen": -0.44195557, "rewards/rejected": -0.4315033, "rewards/accuracies": 0.5, "rewards/margins": -0.01043701, "logps/rejected": -4.30773926, "logps/chosen": -4.41992188, "logits/rejected": -0.80297852, "logits/chosen": -0.92907715, "nll_loss": 4.41894531, "log_odds_ratio": -1.01074219, "log_odds_chosen": -0.42773438, "epoch": 0.0177305, "global_step/max_steps": "5/846", "percentage": "0.59%", "elapsed_time": "20m 45s", "remaining_time": "2d 10h 11m 25s"}
+{"loss": 0.94470215, "grad_norm": 0.43907565, "learning_rate": 3.53e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.004064, "rewards/chosen": -0.04534912, "rewards/rejected": -0.03717041, "rewards/accuracies": 0.0, "rewards/margins": -0.00817871, "logps/rejected": -0.37121582, "logps/chosen": -0.45361328, "logits/rejected": -0.02148438, "logits/chosen": -0.22029114, "nll_loss": 0.45458984, "log_odds_ratio": -0.94628906, "log_odds_chosen": -0.40332031, "epoch": 0.0212766, "global_step/max_steps": "6/846", "percentage": "0.71%", "elapsed_time": "24m 19s", "remaining_time": "2d 8h 45m 40s"}
+{"loss": 0.69799805, "grad_norm": 0.33003554, "learning_rate": 4.12e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.00413, "rewards/chosen": -0.04724121, "rewards/rejected": -0.04620361, "rewards/accuracies": 0.5, "rewards/margins": -0.0010376, "logps/rejected": -0.46240234, "logps/chosen": -0.47265625, "logits/rejected": -0.20344543, "logits/chosen": -0.24220276, "nll_loss": 0.47216797, "log_odds_ratio": -0.81787109, "log_odds_chosen": -0.16503906, "epoch": 0.0248227, "global_step/max_steps": "7/846", "percentage": "0.83%", "elapsed_time": "27m 58s", "remaining_time": "2d 7h 52m 30s"}
+{"loss": 1.16265869, "grad_norm": 0.52832752, "learning_rate": 4.71e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.003879, "rewards/chosen": -0.0425415, "rewards/rejected": -0.06494141, "rewards/accuracies": 0.75, "rewards/margins": 0.0223999, "logps/rejected": -0.64916992, "logps/chosen": -0.42578125, "logits/rejected": -0.28625488, "logits/chosen": -0.14660645, "nll_loss": 0.42529297, "log_odds_ratio": -0.56982422, "log_odds_chosen": 0.37695312, "epoch": 0.02836879, "global_step/max_steps": "8/846", "percentage": "0.95%", "elapsed_time": "34m 5s", "remaining_time": "2d 11h 31m 22s"}
+{"loss": 1.17169189, "grad_norm": 0.5640173, "learning_rate": 5.29e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.003907, "rewards/chosen": -0.09313965, "rewards/rejected": -0.10186768, "rewards/accuracies": 0.5, "rewards/margins": 0.00872803, "logps/rejected": -1.01782227, "logps/chosen": -0.93261719, "logits/rejected": -0.22460938, "logits/chosen": -0.27026367, "nll_loss": 0.93261719, "log_odds_ratio": -0.71630859, "log_odds_chosen": 0.06787109, "epoch": 0.03191489, "global_step/max_steps": "9/846", "percentage": "1.06%", "elapsed_time": "38m 6s", "remaining_time": "2d 11h 4m 40s"}
+{"loss": 1.09545898, "grad_norm": 0.77669293, "learning_rate": 5.88e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.003952, "rewards/chosen": -0.05889893, "rewards/rejected": -0.05953979, "rewards/accuracies": 0.75, "rewards/margins": 0.00067139, "logps/rejected": -0.59533691, "logps/chosen": -0.58886719, "logits/rejected": -0.19677734, "logits/chosen": -0.28369141, "nll_loss": 0.58886719, "log_odds_ratio": -0.95654297, "log_odds_chosen": -0.1953125, "epoch": 0.03546099, "global_step/max_steps": "10/846", "percentage": "1.18%", "elapsed_time": "41m 53s", "remaining_time": "2d 10h 22m 4s"}
+{"eval_loss": 1.01127422, "eval_runtime": 568.4331, "eval_samples_per_second": 0.491, "eval_steps_per_second": 0.062, "eval_rewards/chosen": -0.08544225, "eval_rewards/rejected": -0.07977121, "eval_rewards/accuracies": 0.51428574, "eval_rewards/margins": -0.00566755, "eval_logps/rejected": -0.79744697, "eval_logps/chosen": -0.85415739, "eval_logits/rejected": -0.20883882, "eval_logits/chosen": -0.30125907, "eval_nll_loss": 0.85499442, "eval_log_odds_ratio": -0.8639788, "eval_log_odds_chosen": -0.20164621, "epoch": 0.03546099, "global_step/max_steps": "10/846", "percentage": "1.18%", "elapsed_time": "51m 21s", "remaining_time": "2d 23h 34m 5s"}
+{"loss": 1.0057373, "grad_norm": 0.39458957, "learning_rate": 6.47e-06, "memory(GiB)": 132.63, "train_speed(iter/s)": 0.00328, "rewards/chosen": -0.04196167, "rewards/rejected": -0.01541138, "rewards/accuracies": 0.0, "rewards/margins": -0.02655029, "logps/rejected": -0.15393066, "logps/chosen": -0.41918945, "logits/rejected": -0.23266602, "logits/chosen": -0.4309082, "nll_loss": 0.41918945, "log_odds_ratio": -1.38867188, "log_odds_chosen": -1.06347656, "epoch": 0.03900709, "global_step/max_steps": "11/846", "percentage": "1.30%", "elapsed_time": "55m 37s", "remaining_time": "2d 22h 22m 2s"}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec080daa55ccaa16dce98722dd3f4110f3c0f05ae6dd304295227bb5e3867e7b
+oid sha256:810584ab90190436ff1fa6ef63a701dab6345e3367ef183fe4023d14ad422c93
 size 8312