diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/args.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/args.json new file mode 100644 index 0000000000000000000000000000000000000000..57a03f0d5d3b16ca36cb3cdda00b2512c64a0096 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/args.json @@ -0,0 +1,377 @@ +{ + "output_dir": "/NHNHOME/data/sanghyeok/qwen-cua/runs/v6-recurtorch-img3-hw2-k16-12k-joint-trajonly/v0-20260605-205532", + "per_device_train_batch_size": 1, + "num_train_epochs": 1.0, + "max_steps": -1, + "learning_rate": 1e-05, + "lr_scheduler_type": "cosine", + "lr_scheduler_kwargs": null, + "warmup_steps": 200.0, + "optim": "adamw_torch_fused", + "optim_args": null, + "weight_decay": 0.0, + "adam_beta1": 0.9, + "adam_beta2": 0.95, + "adam_epsilon": 1e-08, + "optim_target_modules": null, + "gradient_accumulation_steps": 16, + "average_tokens_across_devices": true, + "max_grad_norm": 1.0, + "label_smoothing_factor": 0.0, + "bf16": true, + "fp16": false, + "bf16_full_eval": false, + "fp16_full_eval": false, + "tf32": true, + "gradient_checkpointing": true, + "gradient_checkpointing_kwargs": null, + "torch_compile": false, + "torch_compile_backend": null, + "torch_compile_mode": null, + "use_liger_kernel": false, + "liger_kernel_config": null, + "use_cache": false, + "neftune_noise_alpha": null, + "torch_empty_cache_steps": null, + "auto_find_batch_size": false, + "logging_strategy": "steps", + "logging_steps": 1, + "logging_first_step": true, + "log_on_each_node": true, + "logging_nan_inf_filter": true, + "include_num_input_tokens_seen": false, + "log_level": "passive", + "log_level_replica": "warning", + "disable_tqdm": null, + "report_to": [ + "wandb" + ], + "run_name": "v6-recurtorch-img3-hw2-k16-12k-joint-trajonly", + "project": "huggingface", + "trackio_space_id": null, + "trackio_bucket_id": null, + "trackio_static_space_id": null, + "eval_strategy": "steps", + "eval_steps": 100.0, + "eval_delay": 0, + "per_device_eval_batch_size": 1, + "prediction_loss_only": false, + "eval_on_start": false, + "eval_do_concat_batches": true, + "eval_use_gather_object": false, + "eval_accumulation_steps": null, + "include_for_metrics": [], + "batch_eval_metrics": false, + "save_only_model": false, + "save_strategy": "steps", + "save_steps": 100.0, + "save_on_each_node": false, + "save_total_limit": 3, + "enable_jit_checkpoint": false, + "push_to_hub": false, + "hub_token": null, + "hub_private_repo": null, + "hub_model_id": null, + "hub_strategy": "every_save", + "hub_always_push": false, + "hub_revision": null, + "load_best_model_at_end": false, + "metric_for_best_model": "loss", + "greater_is_better": false, + "ignore_data_skip": false, + "restore_callback_states_from_checkpoint": false, + "full_determinism": false, + "seed": 42, + "data_seed": 42, + "use_cpu": false, + "accelerator_config": { + "dispatch_batches": false + }, + "parallelism_config": null, + "dataloader_drop_last": false, + "dataloader_num_workers": 8, + "dataloader_pin_memory": true, + "dataloader_persistent_workers": true, + "dataloader_prefetch_factor": 4, + "remove_unused_columns": true, + "label_names": null, + "train_sampling_strategy": "random", + "length_column_name": "length", + "ddp_find_unused_parameters": null, + "ddp_bucket_cap_mb": null, + "ddp_broadcast_buffers": null, + "ddp_static_graph": null, + "ddp_backend": null, + "ddp_timeout": 18000000, + "fsdp": [], + "fsdp_config": null, + "deepspeed": { + "bf16": { + "enabled": true + }, + "fp16": { + "enabled": false + }, + "zero_optimization": { + "stage": 2, + "overlap_comm": true, + "contiguous_gradients": true, + "reduce_bucket_size": 500000000.0, + "allgather_bucket_size": 500000000.0, + "reduce_scatter": true, + "round_robin_gradients": true + }, + "gradient_clipping": "auto", + "gradient_accumulation_steps": "auto", + "train_batch_size": "auto", + "train_micro_batch_size_per_gpu": "auto", + "steps_per_print": 100, + "wall_clock_breakdown": false, + "_comment_bf16_safety": "DeepSpeed 0.19 BF16_Optimizer keeps fp32 master grad partitions (fp32_groups_flat_partition + accumulate_hp_grads_and_remove_lp hook), so gradient accumulation across many micro-batches is done in fp32. bf16 is only used for the param master copy and forward/backward activations. This is safe at grad_accum=16+; no need to force fp32 collectives." + }, + "debug": null, + "skip_memory_metrics": true, + "do_train": false, + "do_eval": false, + "do_predict": false, + "resume_from_checkpoint": null, + "warmup_ratio": null, + "logging_dir": "/NHNHOME/data/sanghyeok/qwen-cua/runs/v6-recurtorch-img3-hw2-k16-12k-joint-trajonly/v0-20260605-205532/runs", + "local_rank": 0, + "sortish_sampler": false, + "predict_with_generate": false, + "generation_max_length": null, + "generation_num_beams": null, + "generation_config": null, + "tuner_backend": "peft", + "vit_gradient_checkpointing": false, + "router_aux_loss_coef": 0.0, + "enable_dft_loss": false, + "enable_channel_loss": false, + "safe_serialization": true, + "max_shard_size": "5GB", + "check_model": true, + "acc_strategy": "token", + "train_dataloader_shuffle": true, + "group_by_length": false, + "max_epochs": null, + "aligner_lr": null, + "vit_lr": null, + "use_logits_to_keep": null, + "ds3_gather_for_generation": true, + "resume_only_model": false, + "optimizer": null, + "loss_type": null, + "eval_metric": null, + "callbacks": [], + "early_stop_interval": null, + "eval_use_evalscope": false, + "eval_dataset": [], + "eval_dataset_args": null, + "eval_limit": null, + "eval_generation_config": null, + "extra_eval_args": null, + "tuner_type": "full", + "use_galore": false, + "galore_target_modules": null, + "galore_rank": 128, + "galore_update_proj_gap": 50, + "galore_scale": 1.0, + "galore_proj_type": "std", + "galore_optim_per_parameter": false, + "galore_with_embedding": false, + "galore_quantization": false, + "galore_proj_quant": false, + "galore_proj_bits": 4, + "galore_proj_group_size": 256, + "galore_cos_threshold": 0.4, + "galore_gamma_proj": 2, + "galore_queue_size": 5, + "lisa_activated_layers": 0, + "lisa_step_interval": 20, + "use_flash_ckpt": false, + "use_ray": false, + "ray_exp_name": null, + "device_groups": null, + "model": "/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", + "model_type": "qwen3_5", + "model_revision": null, + "task_type": "causal_lm", + "torch_dtype": "bfloat16", + "attn_impl": "flash_attention_2", + "experts_impl": null, + "new_special_tokens": [], + "num_labels": null, + "problem_type": null, + "rope_scaling": null, + "device_map": null, + "max_memory": {}, + "max_model_len": null, + "local_repo_path": null, + "init_strategy": null, + "template": "qwen3_5", + "system": null, + "max_length": 12288, + "truncation_strategy": "delete", + "max_pixels": 16777216, + "agent_template": null, + "norm_bbox": null, + "use_chat_template": true, + "padding_side": "right", + "padding_free": false, + "loss_scale": "default", + "sequence_parallel_size": 1, + "template_backend": "swift", + "response_prefix": null, + "enable_thinking": null, + "add_non_thinking_prefix": true, + "dataset": [ + "/NHNHOME/data/sanghyeok/qwen-cua/_traj_expanded_cache/train_traj_a91429cd18dd0934.jsonl", + "/NHNHOME/data/sanghyeok/qwen-cua/_traj_expanded_cache/train_traj_93f9095470ba8031.jsonl", + "/NHNHOME/data/sanghyeok/qwen-cua/_traj_expanded_cache/train_traj_b30d45f5dd2f85dd.jsonl", + "/NHNHOME/data/sanghyeok/qwen-cua/_traj_expanded_cache/train_traj_9382fa43e1278ec0.jsonl" + ], + "val_dataset": [ + "/NHNHOME/data/sanghyeok/qwen-cua/_traj_expanded_cache/val_traj_b4ad0016c2d07d39.jsonl" + ], + "cached_dataset": [], + "cached_val_dataset": [], + "split_dataset_ratio": 0.0, + "dataset_num_proc": 4, + "load_from_cache_file": false, + "dataset_shuffle": true, + "val_dataset_shuffle": false, + "streaming": false, + "interleave_prob": null, + "stopping_strategy": "first_exhausted", + "shuffle_buffer_size": 1000, + "download_mode": "reuse_dataset_if_exists", + "columns": {}, + "strict": false, + "disable_auto_column_mapping": false, + "model_name": null, + "model_author": null, + "custom_dataset_info": [], + "quant_method": null, + "quant_bits": null, + "hqq_axis": null, + "bnb_4bit_compute_dtype": "bfloat16", + "bnb_4bit_quant_type": "nf4", + "bnb_4bit_use_double_quant": true, + "bnb_4bit_quant_storage": null, + "max_new_tokens": 64, + "temperature": 0.0, + "top_k": null, + "top_p": null, + "repetition_penalty": null, + "num_beams": 1, + "stream": false, + "stop_words": [], + "logprobs": false, + "top_logprobs": null, + "structured_outputs_regex": null, + "adapters": [], + "external_plugins": [], + "custom_register_path": [], + "model_kwargs": {}, + "enable_npu_model_patch": true, + "load_args": false, + "load_data_args": false, + "packing": false, + "packing_length": null, + "packing_num_proc": 1, + "lazy_tokenize": true, + "use_hf": false, + "ignore_args_error": false, + "use_swift_lora": false, + "freeze_parameters": [ + "model.visual", + "model.visual.merger" + ], + "freeze_parameters_regex": null, + "freeze_parameters_ratio": 0.0, + "trainable_parameters": [], + "trainable_parameters_regex": null, + "freeze_llm": false, + "freeze_vit": true, + "freeze_aligner": true, + "target_modules": [ + "all-linear" + ], + "target_regex": null, + "target_parameters": null, + "modules_to_save": [], + "lora_rank": 8, + "lora_alpha": 32, + "lora_dropout": 0.05, + "lora_bias": "none", + "lora_dtype": null, + "lorap_lr_ratio": null, + "use_rslora": false, + "use_dora": false, + "lora_ga_batch_size": 2, + "lora_ga_iters": 2, + "lora_ga_max_length": 1024, + "lora_ga_direction": "ArB2r", + "lora_ga_scale": "stable", + "lora_ga_stable_gamma": 16, + "init_weights": true, + "fourier_n_frequency": 2000, + "fourier_scaling": 300.0, + "boft_block_size": 4, + "boft_block_num": 0, + "boft_n_butterfly_factor": 1, + "boft_dropout": 0.0, + "vera_rank": 256, + "vera_projection_prng_key": 0, + "vera_dropout": 0.0, + "vera_d_initial": 0.1, + "adapter_act": "gelu", + "adapter_length": 128, + "adalora_target_r": 8, + "adalora_init_r": 12, + "adalora_tinit": 0, + "adalora_tfinal": 0, + "adalora_deltaT": 1, + "adalora_beta1": 0.85, + "adalora_beta2": 0.85, + "adalora_orth_reg_weight": 0.5, + "llamapro_num_new_blocks": 4, + "llamapro_num_groups": null, + "reft_layer_key": null, + "reft_layers": null, + "reft_rank": 4, + "reft_intervention_type": "LoreftIntervention", + "reft_args": null, + "swanlab_token": null, + "swanlab_project": "ms-swift", + "swanlab_workspace": null, + "swanlab_exp_name": null, + "swanlab_notification_method": null, + "swanlab_webhook_url": null, + "swanlab_secret": null, + "swanlab_sender_email": null, + "swanlab_receiver_email": null, + "swanlab_smtp_server": null, + "swanlab_smtp_port": null, + "swanlab_email_language": "zh", + "swanlab_mode": "cloud", + "add_version": true, + "create_checkpoint_symlink": false, + "zero_hpz_partition_size": null, + "deepspeed_autotp_size": null, + "swift_version": "4.2.1", + "ckpt_dir": null, + "rank": 0, + "global_world_size": 8, + "local_world_size": 8, + "model_suffix": "Qwen3.5-4B", + "model_info": "ModelInfo(model_type='qwen3_5', model_dir='/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a', torch_dtype=torch.bfloat16, max_model_len=262144, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)", + "model_meta": "ModelMeta(model_type='qwen3_5', model_groups=[ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3.5-0.8B', hf_model_id='Qwen/Qwen3.5-0.8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-2B', hf_model_id='Qwen/Qwen3.5-2B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-4B', hf_model_id='Qwen/Qwen3.5-4B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-9B', hf_model_id='Qwen/Qwen3.5-9B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-27B', hf_model_id='Qwen/Qwen3.5-27B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-27B-FP8', hf_model_id='Qwen/Qwen3.5-27B-FP8', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-0.8B-Base', hf_model_id='Qwen/Qwen3.5-0.8B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-2B-Base', hf_model_id='Qwen/Qwen3.5-2B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-4B-Base', hf_model_id='Qwen/Qwen3.5-4B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-9B-Base', hf_model_id='Qwen/Qwen3.5-9B-Base', model_path=None, ms_revision=None, hf_revision=None)], template='qwen3_5', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3.6-27B', hf_model_id='Qwen/Qwen3.6-27B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.6-27B-FP8', hf_model_id='Qwen/Qwen3.6-27B-FP8', model_path=None, ms_revision=None, hf_revision=None)], template='qwen3_5', ignore_patterns=None, requires=None, tags=[])], loader=, template='qwen3_5', model_arch=MultiModelKeys(arch_name='qwen2_vl', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model.language_model', 'lm_head'], aligner=['model.visual.merger'], vision_tower=['model.visual'], generator=[]), mcore_model_type=None, architectures=['Qwen3_5ForConditionalGeneration'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, task_type=None, ignore_patterns=None, requires=['transformers>=5.0.0.dev', 'qwen_vl_utils>=0.0.14', 'decord'], tags=[])", + "model_dir": "/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", + "template_meta": "QwenTemplateMeta(template_type='qwen3_5', prefix=[], prompt=['<|im_start|>user\\n{{QUERY}}<|im_end|>\\n<|im_start|>assistant\\n'], chat_sep=['<|im_end|>\\n'], suffix=['<|im_end|>\\n'], template_cls=, system_prefix=['<|im_start|>system\\n{{SYSTEM}}<|im_end|>\\n'], default_system=None, auto_add_bos=False, stop_words=['<|endoftext|>'], agent_template='qwen3_5', is_thinking=True, thinking_prefix='\\n', non_thinking_prefix='\\n\\n\\n\\n', history_thinking_prefix='')", + "_val_dataset_exists": true, + "hub": "", + "evaluation_strategy": "steps", + "training_args": "Seq2SeqTrainingArguments(output_dir='/NHNHOME/data/sanghyeok/qwen-cua/runs/v6-recurtorch-img3-hw2-k16-12k-joint-trajonly/v0-20260605-205532', per_device_train_batch_size=1, num_train_epochs=1.0, max_steps=-1, learning_rate=1e-05, lr_scheduler_type=, lr_scheduler_kwargs=None, warmup_steps=200.0, optim=, optim_args=None, weight_decay=0.0, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, optim_target_modules=None, gradient_accumulation_steps=16, average_tokens_across_devices=None, max_grad_norm=1.0, label_smoothing_factor=0.0, bf16=True, fp16=False, bf16_full_eval=False, fp16_full_eval=False, tf32=True, gradient_checkpointing=True, gradient_checkpointing_kwargs=None, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, use_liger_kernel=False, liger_kernel_config=None, use_cache=False, neftune_noise_alpha=None, torch_empty_cache_steps=None, auto_find_batch_size=False, logging_strategy=, logging_steps=1, logging_first_step=True, log_on_each_node=True, logging_nan_inf_filter=True, include_num_input_tokens_seen=None, log_level='passive', log_level_replica='warning', disable_tqdm=False, report_to=['wandb'], run_name='v6-recurtorch-img3-hw2-k16-12k-joint-trajonly', project='huggingface', trackio_space_id=None, trackio_bucket_id=None, trackio_static_space_id=None, eval_strategy=, eval_steps=100, eval_delay=0, per_device_eval_batch_size=1, prediction_loss_only=False, eval_on_start=False, eval_do_concat_batches=True, eval_use_gather_object=False, eval_accumulation_steps=None, include_for_metrics=[], batch_eval_metrics=False, save_only_model=False, save_strategy=, save_steps=100, save_on_each_node=False, save_total_limit=3, enable_jit_checkpoint=False, push_to_hub=False, hub_token=None, hub_private_repo=None, hub_model_id=None, hub_strategy=, hub_always_push=False, hub_revision=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=False, restore_callback_states_from_checkpoint=False, full_determinism=False, seed=42, data_seed=42, use_cpu=False, 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), parallelism_config=None, dataloader_drop_last=False, dataloader_num_workers=8, dataloader_pin_memory=True, dataloader_persistent_workers=True, dataloader_prefetch_factor=4, remove_unused_columns=False, label_names=None, train_sampling_strategy='random', length_column_name='length', ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, ddp_static_graph=None, ddp_backend=None, ddp_timeout=18000000, fsdp=[], fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, deepspeed={'bf16': {'enabled': True}, 'fp16': {'enabled': False}, 'zero_optimization': {'stage': 2, 'overlap_comm': True, 'contiguous_gradients': True, 'reduce_bucket_size': 500000000.0, 'allgather_bucket_size': 500000000.0, 'reduce_scatter': True, 'round_robin_gradients': True}, 'gradient_clipping': 'auto', 'gradient_accumulation_steps': 'auto', 'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'steps_per_print': 100, 'wall_clock_breakdown': False, '_comment_bf16_safety': 'DeepSpeed 0.19 BF16_Optimizer keeps fp32 master grad partitions (fp32_groups_flat_partition + accumulate_hp_grads_and_remove_lp hook), so gradient accumulation across many micro-batches is done in fp32. bf16 is only used for the param master copy and forward/backward activations. This is safe at grad_accum=16+; no need to force fp32 collectives.'}, debug=[], skip_memory_metrics=True, do_train=False, do_eval=True, do_predict=False, resume_from_checkpoint=None, warmup_ratio=None, logging_dir='/NHNHOME/data/sanghyeok/qwen-cua/runs/v6-recurtorch-img3-hw2-k16-12k-joint-trajonly/v0-20260605-205532/runs', local_rank=0, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=False, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, safe_serialization=True, max_shard_size='5GB', check_model=True, acc_strategy='token', train_dataloader_shuffle=True, group_by_length=False, max_epochs=None, aligner_lr=None, vit_lr=None, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=False, optimizer=None, loss_type=None, eval_metric=None, callbacks=[], early_stop_interval=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, tuner_type='full', use_galore=False, galore_target_modules=None, galore_rank=128, galore_update_proj_gap=50, galore_scale=1.0, galore_proj_type='std', galore_optim_per_parameter=False, galore_with_embedding=False, galore_quantization=False, galore_proj_quant=False, galore_proj_bits=4, galore_proj_group_size=256, galore_cos_threshold=0.4, galore_gamma_proj=2, galore_queue_size=5, lisa_activated_layers=0, lisa_step_interval=20, use_flash_ckpt=False)" +} \ No newline at end of file diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/chat_template.jinja b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/config.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/config.json new file mode 100644 index 0000000000000000000000000000000000000000..314297e9ae3a6f7ca90b01d29ee0a46cfbf53c2a --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/config.json @@ -0,0 +1,119 @@ +{ + "architectures": [ + "Qwen35VLLatentForConditionalGeneration" + ], + "dtype": "bfloat16", + "eos_token_id": 248046, + "hidden_size": 2560, + "image_token_id": 248056, + "latent_end_id": 248078, + "latent_pad_id": 248079, + "latent_start_id": 248077, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "latent_end_id": 248078, + "latent_pad_id": 248079, + "latent_start_id": 248077, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 1, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": 248044, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": false, + "vocab_size": 248080 + }, + "tie_word_embeddings": true, + "transformers_version": "5.8.1", + "use_cache": false, + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/generation_config.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1f1ab970fed81cb788c2945bed33833d247b9091 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": [ + 248044, + 248046 + ], + "transformers_version": "5.8.1", + "use_cache": true +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..fe4405f652cbbc3243387dce905174557a7d8db9 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aa0f8bd0f4fdc7b06622bf4443c87d1ced693e18d1b5fcb28d2edd0132621f68 +size 6337637425 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..c8b4e580e3b1af7b7ffabd8c1e6aed58c3338ae2 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:219117149641a590015479de731b8114723dbcd80e9f6efbdde1a535ac60a58a +size 6337640689 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..0e37d0c6c29ce690dfff178d72ffedf65b099b67 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:49147bdcd589b10f08235d77625d71e569385760d1d0b607dd2c9f4ff47f1bad +size 6337641521 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..f562e17949efd6afe4ed04efe0d3879217e02b9c --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a93c4f6949890b6eeceed32634fc3836978ed1927211df570ceef85432dda4f +size 6337641521 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..987d7056d5af207da0b380eb842d1dcf77882787 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29efcd83980df481b6d05cb2f00720fdb16250adcc25c054657a48394a60ffea +size 6337641457 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..2a4369f7b5042f8dcc82a0042e368fec42cebc0c --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca89d0aa5a5af1edf09e13d9af4a32dbd09f92855fa4ec353080f20c2f1aa2f5 +size 6337641521 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..2cad57baf520a71cffe3724a2cd0f46f6b675465 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ba272d648abc30129c181b70bfa9e5636639e7a20efb6a4b17af9e73eca0488 +size 6337641201 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..3b4d6589d7b20a0e6c8a56d05293d21c41369788 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1408126901681495a60d176f71320e0eb242123c7535b6c03f835679a1ec54c9 +size 6337640241 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/mp_rank_00_model_states.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/mp_rank_00_model_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..bba144c13f6d60c38d4c19088bdcef322ab2a085 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/global_step1500/mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62cd7489be536b5c0426beb5a2b7a3cb1d5ca0484c9ea64023601eaabbab497c +size 9784608719 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/latest b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/latest new file mode 100644 index 0000000000000000000000000000000000000000..c56ff7708f44fb7928fea2f70d6d7342ce0d5b67 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/latest @@ -0,0 +1 @@ +global_step1500 \ No newline at end of file diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00001-of-00003.safetensors b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..419c7ab2b334ec1fbb5329e6d9e09d81c1b8759b --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:647ae253f81a8ff63e7f7535915712a767fd3b62414bb7620ac176e7fe6ecc87 +size 4995635952 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00002-of-00003.safetensors b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ac3f84a0c0a620d20c54b04b37f6985aa867b732 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:700ac3670671ef2066549ebf3251c2ab0f8cede4b69f11b175a057cc159ad868 +size 4993498424 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00003-of-00003.safetensors b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4ab8023a740649342f1bc7e224320898de73054c --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9132529e66cac3079a828b70ff720888e735b62dcc8a839f7c42730ef896d42c +size 398327248 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model.safetensors.index.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..0bdeac0c13fb0bc1bd472391718513cd39619d7a --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/model.safetensors.index.json @@ -0,0 +1,747 @@ +{ + "metadata": { + "total_parameters": 4558600704, + "total_size": 10387371008 + }, + "weight_map": { + "alignment_projector.action_proj.bias": "model-00001-of-00003.safetensors", + "alignment_projector.action_proj.weight": "model-00001-of-00003.safetensors", + "alignment_projector.ln_action.bias": "model-00001-of-00003.safetensors", + "alignment_projector.ln_action.weight": "model-00001-of-00003.safetensors", + "alignment_projector.ln_state.bias": "model-00001-of-00003.safetensors", + "alignment_projector.ln_state.weight": "model-00001-of-00003.safetensors", + "alignment_projector.mlp.1.bias": "model-00001-of-00003.safetensors", + "alignment_projector.mlp.1.weight": "model-00001-of-00003.safetensors", + "alignment_projector.mlp.3.bias": "model-00001-of-00003.safetensors", + "alignment_projector.mlp.3.weight": "model-00001-of-00003.safetensors", + "alignment_projector.norm.bias": "model-00001-of-00003.safetensors", + "alignment_projector.norm.weight": "model-00001-of-00003.safetensors", + "alignment_projector.state_proj.bias": "model-00001-of-00003.safetensors", + "alignment_projector.state_proj.weight": "model-00001-of-00003.safetensors", + "latent_init_embedding": "model-00001-of-00003.safetensors", + "lm_head.weight": "model-00001-of-00003.safetensors", + "model.language_model.embed_tokens.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.10.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.22.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.23.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.24.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.25.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.26.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.27.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.28.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.29.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.k_norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.q_norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.30.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.A_log": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.conv1d.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.dt_bias": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.in_proj_a.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.in_proj_b.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.in_proj_qkv.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.in_proj_z.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.30.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.k_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.q_norm.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.31.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.language_model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.k_norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.q_norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.A_log": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.conv1d.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.dt_bias": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.in_proj_a.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.in_proj_b.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.in_proj_qkv.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.in_proj_z.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.norm.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.language_model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.language_model.norm.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.0.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.1.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.10.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.10.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.10.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.10.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.10.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.11.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.12.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.13.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.14.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.15.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.16.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.17.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.18.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.19.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.2.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.2.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.20.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.20.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.21.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.22.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.attn.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.attn.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.attn.qkv.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.attn.qkv.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.mlp.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.mlp.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.mlp.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.mlp.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.norm1.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.norm1.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.norm2.bias": "model-00003-of-00003.safetensors", + "model.visual.blocks.23.norm2.weight": "model-00003-of-00003.safetensors", + "model.visual.blocks.3.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.3.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.4.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.5.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.6.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.7.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.8.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.attn.proj.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.attn.proj.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.attn.qkv.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.attn.qkv.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.mlp.linear_fc1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.mlp.linear_fc1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.mlp.linear_fc2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.mlp.linear_fc2.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.norm1.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.norm1.weight": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.norm2.bias": "model-00002-of-00003.safetensors", + "model.visual.blocks.9.norm2.weight": "model-00002-of-00003.safetensors", + "model.visual.merger.linear_fc1.bias": "model-00003-of-00003.safetensors", + "model.visual.merger.linear_fc1.weight": "model-00003-of-00003.safetensors", + "model.visual.merger.linear_fc2.bias": "model-00003-of-00003.safetensors", + "model.visual.merger.linear_fc2.weight": "model-00003-of-00003.safetensors", + "model.visual.merger.norm.bias": "model-00003-of-00003.safetensors", + "model.visual.merger.norm.weight": "model-00003-of-00003.safetensors", + "model.visual.patch_embed.proj.bias": "model-00003-of-00003.safetensors", + "model.visual.patch_embed.proj.weight": "model-00003-of-00003.safetensors", + "model.visual.pos_embed.weight": "model-00003-of-00003.safetensors" + } +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/preprocessor_config.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..2ea84a437d448ff71b08df68fdd949d5cc4ebb64 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/preprocessor_config.json @@ -0,0 +1,21 @@ +{ + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "patch_size": 16, + "temporal_patch_size": 2, + "merge_size": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "processor_class": "Qwen3VLProcessor", + "image_processor_type": "Qwen2VLImageProcessorFast" +} \ No newline at end of file diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/processor_config.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_0.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..86eaf5e40dd94669ca76b101cf9f8739e75c9294 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:386f8cf5a371270a033a0afdb301ab96f7f1100a98c73953094cb02fc0679c0d +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_1.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..38dbff679bfbf2f3aade238dfdfc43c385f23334 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aa040f762df4f5f45d2206432293edf4deac596688c61e11ae45adf6b291ddf3 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_2.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..31978c299901c622b39507e1513d26f45e6e4a79 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f26b632812a9d2f9aa70c5eb9f2847e528ca75d8c9d2510398ffd8defb9e1057 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_3.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_3.pth new file mode 100644 index 0000000000000000000000000000000000000000..55b8da5b9e39cd017c04f97abe46db037ebf16f8 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e97740583dc518497b02e411b0dae69b0a201400dfa7820c19089ed6a6e1ba3 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_4.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_4.pth new file mode 100644 index 0000000000000000000000000000000000000000..0022f2989901cd3a52f9032b32a09841521ef45e --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_4.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53a02afdd75402e5ed1c5de1176a70e488494ba4a086375772f7e83fa05e6cc0 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_5.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_5.pth new file mode 100644 index 0000000000000000000000000000000000000000..4613c4f8de3fef534415098e475909b8f7ab97c2 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_5.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da29652c6ecadcbeae7f614b7db9bd0341f933bb9f3e0e52fa772073c97e3692 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_6.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_6.pth new file mode 100644 index 0000000000000000000000000000000000000000..5e799753068722228ae33a268259fe20661b5d2d --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_6.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5a9b5cddad27c33d03f513715fe5cdb3fc63ad355da2fdcb28fa9a89da68f06 +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_7.pth b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_7.pth new file mode 100644 index 0000000000000000000000000000000000000000..2239a6b21428f351d266a85ab12af52e47a6e3a7 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/rng_state_7.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:349a60aa899b4ead79139d01c2b7896cafd74f8aa94b0ce5385e5c2409ef0cfb +size 16389 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/scheduler.pt b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..d4b7e488bfa24a9c7680a3ba105fcade57c48cbb --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1825dc5eb0e39ce8b4f2f0bbb2c7a4aa27ccba972a812fa0ac4e17a9865bb99a +size 1465 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..9e8ac8634b7c90a05a65182d43516ddf7371c521 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4012eb888c8edaca05866241a3a8319ca02d45c367eef3da25b8ba4ef002f5b3 +size 19989885 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer_config.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/trainer_state.json b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3d331f9796c3a3aa70ea8d7d9d24f43b30f1e98a --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/trainer_state.json @@ -0,0 +1,16714 @@ +{ + "best_global_step": 1500, + "best_metric": 0.59888691, + "best_model_checkpoint": "/NHNHOME/data/sanghyeok/qwen-cua/runs/v6-recurtorch-img3-hw2-k16-12k-joint-trajonly/v0-20260605-205532/checkpoint-1500", + "epoch": 0.5590756615728661, + "eval_steps": 100, + "global_step": 1500, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0003727171077152441, + "grad_norm": 643.0632934570312, + "learning_rate": 5.0000000000000004e-08, + "loss": 25.36384391784668, + "loss_alignment": 1.0, + "loss_alignment_w": 0.5, + "loss_sft": 0.8918107151985168, + "loss_total_v6": 1.391810655593872, + "step": 1 + }, + { + "epoch": 0.0007454342154304882, + "grad_norm": 739.0370483398438, + "learning_rate": 1.0000000000000001e-07, + "loss": 25.825382232666016, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 0.9509366154670715, + "loss_total_v6": 1.4548428058624268, + "step": 2 + }, + { + "epoch": 0.0011181513231457323, + "grad_norm": 676.4771728515625, + "learning_rate": 1.5000000000000002e-07, + "loss": 25.557247161865234, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 0.9966087937355042, + "loss_total_v6": 1.5005149841308594, + "step": 3 + }, + { + "epoch": 0.0014908684308609765, + "grad_norm": 673.6318359375, + "learning_rate": 2.0000000000000002e-07, + "loss": 26.474090576171875, + "loss_alignment": 0.98828125, + "loss_alignment_w": 0.494140625, + "loss_sft": 1.1548562049865723, + "loss_total_v6": 1.6489968299865723, + "step": 4 + }, + { + "epoch": 0.0018635855385762206, + "grad_norm": 591.028076171875, + "learning_rate": 2.5000000000000004e-07, + "loss": 25.20639991760254, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 1.370214819908142, + "loss_total_v6": 1.874121069908142, + "step": 5 + }, + { + "epoch": 0.0022363026462914647, + "grad_norm": 617.078857421875, + "learning_rate": 3.0000000000000004e-07, + "loss": 26.411823272705078, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 1.389568567276001, + "loss_total_v6": 1.893474817276001, + "step": 6 + }, + { + "epoch": 0.002609019754006709, + "grad_norm": 555.0621337890625, + "learning_rate": 3.5000000000000004e-07, + "loss": 24.953838348388672, + "loss_alignment": 1.015625, + "loss_alignment_w": 0.5078125, + "loss_sft": 1.402374267578125, + "loss_total_v6": 1.910186767578125, + "step": 7 + }, + { + "epoch": 0.002981736861721953, + "grad_norm": 432.5752868652344, + "learning_rate": 4.0000000000000003e-07, + "loss": 24.18780517578125, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 0.6765084266662598, + "loss_total_v6": 1.1804146766662598, + "step": 8 + }, + { + "epoch": 0.003354453969437197, + "grad_norm": 416.21697998046875, + "learning_rate": 4.5000000000000003e-07, + "loss": 24.23617935180664, + "loss_alignment": 1.0, + "loss_alignment_w": 0.5, + "loss_sft": 1.585355281829834, + "loss_total_v6": 2.085355281829834, + "step": 9 + }, + { + "epoch": 0.003727171077152441, + "grad_norm": 398.6941223144531, + "learning_rate": 5.000000000000001e-07, + "loss": 23.68105125427246, + "loss_alignment": 0.99609375, + "loss_alignment_w": 0.498046875, + "loss_sft": 0.9940811991691589, + "loss_total_v6": 1.4921281337738037, + "step": 10 + }, + { + "epoch": 0.004099888184867686, + "grad_norm": 245.6785888671875, + "learning_rate": 5.5e-07, + "loss": 23.7061767578125, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 1.181485652923584, + "loss_total_v6": 1.685391902923584, + "step": 11 + }, + { + "epoch": 0.004472605292582929, + "grad_norm": 239.88099670410156, + "learning_rate": 6.000000000000001e-07, + "loss": 23.629940032958984, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 0.6826500296592712, + "loss_total_v6": 1.186556339263916, + "step": 12 + }, + { + "epoch": 0.004845322400298174, + "grad_norm": 199.00933837890625, + "learning_rate": 6.5e-07, + "loss": 22.333751678466797, + "loss_alignment": 0.99609375, + "loss_alignment_w": 0.498046875, + "loss_sft": 0.9474824070930481, + "loss_total_v6": 1.4455292224884033, + "step": 13 + }, + { + "epoch": 0.005218039508013418, + "grad_norm": 149.41921997070312, + "learning_rate": 7.000000000000001e-07, + "loss": 21.385147094726562, + "loss_alignment": 1.015625, + "loss_alignment_w": 0.5078125, + "loss_sft": 0.4753382205963135, + "loss_total_v6": 0.9831507205963135, + "step": 14 + }, + { + "epoch": 0.005590756615728662, + "grad_norm": 140.5125732421875, + "learning_rate": 7.5e-07, + "loss": 21.593677520751953, + "loss_alignment": 1.015625, + "loss_alignment_w": 0.5078125, + "loss_sft": 1.1277178525924683, + "loss_total_v6": 1.6355303525924683, + "step": 15 + }, + { + "epoch": 0.005963473723443906, + "grad_norm": 169.3380126953125, + "learning_rate": 8.000000000000001e-07, + "loss": 22.067859649658203, + "loss_alignment": 1.0078125, + "loss_alignment_w": 0.50390625, + "loss_sft": 0.7362151741981506, + "loss_total_v6": 1.2401213645935059, + "step": 16 + }, + { + "epoch": 0.00633619083115915, + "grad_norm": 116.58922576904297, + "learning_rate": 8.500000000000001e-07, + "loss": 21.020849227905273, + "loss_alignment": 0.98828125, + "loss_alignment_w": 0.494140625, + "loss_sft": 0.4898141920566559, + "loss_total_v6": 0.9839547872543335, + "step": 17 + }, + { + "epoch": 0.006708907938874394, + "grad_norm": 118.91185760498047, + "learning_rate": 9.000000000000001e-07, + "loss": 21.43584442138672, + "loss_alignment": 1.0, + "loss_alignment_w": 0.5, + "loss_sft": 1.0262619256973267, + "loss_total_v6": 1.5262619256973267, + "step": 18 + }, + { + "epoch": 0.007081625046589639, + "grad_norm": 109.05036926269531, + "learning_rate": 9.500000000000001e-07, + "loss": 22.060020446777344, + "loss_alignment": 1.0, + "loss_alignment_w": 0.5, + "loss_sft": 0.6619189381599426, + "loss_total_v6": 1.1619188785552979, + "step": 19 + }, + { + "epoch": 0.007454342154304882, + "grad_norm": 100.84404754638672, + "learning_rate": 1.0000000000000002e-06, + "loss": 20.71269989013672, + "loss_alignment": 0.9921875, + "loss_alignment_w": 0.49609375, + "loss_sft": 0.7731999158859253, + "loss_total_v6": 1.2692936658859253, + "step": 20 + }, + { + "epoch": 0.007827059262020127, + "grad_norm": 85.67799377441406, + "learning_rate": 1.0500000000000001e-06, + "loss": 19.962120056152344, + "loss_alignment": 0.9765625, + "loss_alignment_w": 0.48828125, + "loss_sft": 0.7308831214904785, + "loss_total_v6": 1.2191643714904785, + "step": 21 + }, + { + "epoch": 0.008199776369735371, + "grad_norm": 78.51901245117188, + "learning_rate": 1.1e-06, + "loss": 20.593730926513672, + "loss_alignment": 0.98828125, + "loss_alignment_w": 0.494140625, + "loss_sft": 0.699788510799408, + "loss_total_v6": 1.1939291954040527, + "step": 22 + }, + { + "epoch": 0.008572493477450614, + "grad_norm": 65.32942199707031, + "learning_rate": 1.1500000000000002e-06, + "loss": 20.362258911132812, + "loss_alignment": 0.984375, + "loss_alignment_w": 0.4921875, + "loss_sft": 0.7115195989608765, + "loss_total_v6": 1.2037070989608765, + "step": 23 + }, + { + "epoch": 0.008945210585165859, + "grad_norm": 58.363990783691406, + "learning_rate": 1.2000000000000002e-06, + "loss": 19.321300506591797, + "loss_alignment": 0.99609375, + "loss_alignment_w": 0.498046875, + "loss_sft": 0.7411585450172424, + "loss_total_v6": 1.2392053604125977, + "step": 24 + }, + { + "epoch": 0.009317927692881103, + "grad_norm": 65.69908905029297, + "learning_rate": 1.25e-06, + "loss": 19.714149475097656, + "loss_alignment": 0.98046875, + "loss_alignment_w": 0.490234375, + "loss_sft": 0.5112648606300354, + "loss_total_v6": 1.0014991760253906, + "step": 25 + }, + { + "epoch": 0.009690644800596348, + "grad_norm": 62.45393371582031, + "learning_rate": 1.3e-06, + "loss": 19.25128936767578, + "loss_alignment": 0.9765625, + "loss_alignment_w": 0.48828125, + "loss_sft": 0.562895655632019, + "loss_total_v6": 1.051176905632019, + "step": 26 + }, + { + "epoch": 0.01006336190831159, + "grad_norm": 58.540985107421875, + "learning_rate": 1.3500000000000002e-06, + "loss": 18.925086975097656, + "loss_alignment": 0.9765625, + "loss_alignment_w": 0.48828125, + "loss_sft": 0.5138747096061707, + "loss_total_v6": 1.0021560192108154, + "step": 27 + }, + { + "epoch": 0.010436079016026835, + "grad_norm": 54.053165435791016, + "learning_rate": 1.4000000000000001e-06, + "loss": 18.781091690063477, + "loss_alignment": 0.97265625, + "loss_alignment_w": 0.486328125, + "loss_sft": 0.3268469274044037, + "loss_total_v6": 0.8131750822067261, + "step": 28 + }, + { + "epoch": 0.01080879612374208, + "grad_norm": 57.27824401855469, + "learning_rate": 1.45e-06, + "loss": 19.341501235961914, + "loss_alignment": 0.9609375, + "loss_alignment_w": 0.48046875, + "loss_sft": 0.6492088437080383, + "loss_total_v6": 1.1296775341033936, + "step": 29 + }, + { + "epoch": 0.011181513231457324, + "grad_norm": 56.68754577636719, + "learning_rate": 1.5e-06, + "loss": 19.05861473083496, + "loss_alignment": 0.953125, + "loss_alignment_w": 0.4765625, + "loss_sft": 0.6362125277519226, + "loss_total_v6": 1.1127750873565674, + "step": 30 + }, + { + "epoch": 0.011554230339172567, + "grad_norm": 50.223628997802734, + "learning_rate": 1.5500000000000002e-06, + "loss": 18.249225616455078, + "loss_alignment": 0.94140625, + "loss_alignment_w": 0.470703125, + "loss_sft": 0.5512086749076843, + "loss_total_v6": 1.021911859512329, + "step": 31 + }, + { + "epoch": 0.011926947446887812, + "grad_norm": 45.14276885986328, + "learning_rate": 1.6000000000000001e-06, + "loss": 18.38421630859375, + "loss_alignment": 0.9375, + "loss_alignment_w": 0.46875, + "loss_sft": 0.6443678736686707, + "loss_total_v6": 1.1131179332733154, + "step": 32 + }, + { + "epoch": 0.012299664554603056, + "grad_norm": 47.39469909667969, + "learning_rate": 1.6500000000000003e-06, + "loss": 18.149253845214844, + "loss_alignment": 0.94140625, + "loss_alignment_w": 0.470703125, + "loss_sft": 0.8582465052604675, + "loss_total_v6": 1.3289496898651123, + "step": 33 + }, + { + "epoch": 0.0126723816623183, + "grad_norm": 45.93950271606445, + "learning_rate": 1.7000000000000002e-06, + "loss": 17.981571197509766, + "loss_alignment": 0.94140625, + "loss_alignment_w": 0.470703125, + "loss_sft": 0.6096228957176208, + "loss_total_v6": 1.0803260803222656, + "step": 34 + }, + { + "epoch": 0.013045098770033545, + "grad_norm": 42.14314651489258, + "learning_rate": 1.75e-06, + "loss": 17.644222259521484, + "loss_alignment": 0.921875, + "loss_alignment_w": 0.4609375, + "loss_sft": 0.5166834592819214, + "loss_total_v6": 0.9776209592819214, + "step": 35 + }, + { + "epoch": 0.013417815877748788, + "grad_norm": 40.2923583984375, + "learning_rate": 1.8000000000000001e-06, + "loss": 17.168575286865234, + "loss_alignment": 0.91015625, + "loss_alignment_w": 0.455078125, + "loss_sft": 0.6975638270378113, + "loss_total_v6": 1.152642011642456, + "step": 36 + }, + { + "epoch": 0.013790532985464033, + "grad_norm": 49.44568634033203, + "learning_rate": 1.85e-06, + "loss": 17.363134384155273, + "loss_alignment": 0.90625, + "loss_alignment_w": 0.453125, + "loss_sft": 0.5976616740226746, + "loss_total_v6": 1.0507867336273193, + "step": 37 + }, + { + "epoch": 0.014163250093179277, + "grad_norm": 46.53181457519531, + "learning_rate": 1.9000000000000002e-06, + "loss": 17.321949005126953, + "loss_alignment": 0.890625, + "loss_alignment_w": 0.4453125, + "loss_sft": 0.5999412536621094, + "loss_total_v6": 1.0452537536621094, + "step": 38 + }, + { + "epoch": 0.014535967200894522, + "grad_norm": 42.753536224365234, + "learning_rate": 1.9500000000000004e-06, + "loss": 17.153568267822266, + "loss_alignment": 0.890625, + "loss_alignment_w": 0.4453125, + "loss_sft": 0.5795389413833618, + "loss_total_v6": 1.0248514413833618, + "step": 39 + }, + { + "epoch": 0.014908684308609765, + "grad_norm": 38.422115325927734, + "learning_rate": 2.0000000000000003e-06, + "loss": 17.10763931274414, + "loss_alignment": 0.890625, + "loss_alignment_w": 0.4453125, + "loss_sft": 0.7512833476066589, + "loss_total_v6": 1.1965959072113037, + "step": 40 + }, + { + "epoch": 0.01528140141632501, + "grad_norm": 40.91168212890625, + "learning_rate": 2.05e-06, + "loss": 16.95140266418457, + "loss_alignment": 0.85546875, + "loss_alignment_w": 0.427734375, + "loss_sft": 0.6459410786628723, + "loss_total_v6": 1.0736753940582275, + "step": 41 + }, + { + "epoch": 0.015654118524040254, + "grad_norm": 39.496864318847656, + "learning_rate": 2.1000000000000002e-06, + "loss": 17.07189178466797, + "loss_alignment": 0.85546875, + "loss_alignment_w": 0.427734375, + "loss_sft": 0.7375544905662537, + "loss_total_v6": 1.1652889251708984, + "step": 42 + }, + { + "epoch": 0.0160268356317555, + "grad_norm": 38.13604736328125, + "learning_rate": 2.15e-06, + "loss": 16.322662353515625, + "loss_alignment": 0.84765625, + "loss_alignment_w": 0.423828125, + "loss_sft": 0.574190080165863, + "loss_total_v6": 0.998018205165863, + "step": 43 + }, + { + "epoch": 0.016399552739470743, + "grad_norm": 42.819740295410156, + "learning_rate": 2.2e-06, + "loss": 16.529129028320312, + "loss_alignment": 0.8203125, + "loss_alignment_w": 0.41015625, + "loss_sft": 0.6748767495155334, + "loss_total_v6": 1.0850329399108887, + "step": 44 + }, + { + "epoch": 0.016772269847185987, + "grad_norm": 37.56767654418945, + "learning_rate": 2.25e-06, + "loss": 16.4493465423584, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.47742560505867004, + "loss_total_v6": 0.47742560505867004, + "step": 45 + }, + { + "epoch": 0.01714498695490123, + "grad_norm": 40.35765075683594, + "learning_rate": 2.3000000000000004e-06, + "loss": 16.304941177368164, + "loss_alignment": 0.8046875, + "loss_alignment_w": 0.40234375, + "loss_sft": 0.5231730937957764, + "loss_total_v6": 0.9255168437957764, + "step": 46 + }, + { + "epoch": 0.017517704062616473, + "grad_norm": 37.33998489379883, + "learning_rate": 2.35e-06, + "loss": 16.724491119384766, + "loss_alignment": 0.8046875, + "loss_alignment_w": 0.40234375, + "loss_sft": 0.5488316416740417, + "loss_total_v6": 0.9511753916740417, + "step": 47 + }, + { + "epoch": 0.017890421170331718, + "grad_norm": 38.94746017456055, + "learning_rate": 2.4000000000000003e-06, + "loss": 16.450820922851562, + "loss_alignment": 0.80078125, + "loss_alignment_w": 0.400390625, + "loss_sft": 0.5042392015457153, + "loss_total_v6": 0.9046298265457153, + "step": 48 + }, + { + "epoch": 0.018263138278046962, + "grad_norm": 38.3525505065918, + "learning_rate": 2.4500000000000003e-06, + "loss": 16.540077209472656, + "loss_alignment": 0.78515625, + "loss_alignment_w": 0.392578125, + "loss_sft": 0.7153707146644592, + "loss_total_v6": 1.1079487800598145, + "step": 49 + }, + { + "epoch": 0.018635855385762207, + "grad_norm": 42.156089782714844, + "learning_rate": 2.5e-06, + "loss": 16.100284576416016, + "loss_alignment": 0.7578125, + "loss_alignment_w": 0.37890625, + "loss_sft": 0.6974010467529297, + "loss_total_v6": 1.0763072967529297, + "step": 50 + }, + { + "epoch": 0.01900857249347745, + "grad_norm": 38.806915283203125, + "learning_rate": 2.55e-06, + "loss": 15.34006404876709, + "loss_alignment": 0.76171875, + "loss_alignment_w": 0.380859375, + "loss_sft": 0.7244893312454224, + "loss_total_v6": 1.1053487062454224, + "step": 51 + }, + { + "epoch": 0.019381289601192696, + "grad_norm": 36.31513977050781, + "learning_rate": 2.6e-06, + "loss": 16.099275588989258, + "loss_alignment": 0.7734375, + "loss_alignment_w": 0.38671875, + "loss_sft": 0.650540828704834, + "loss_total_v6": 1.037259578704834, + "step": 52 + }, + { + "epoch": 0.01975400670890794, + "grad_norm": 33.53144454956055, + "learning_rate": 2.6500000000000005e-06, + "loss": 15.513650894165039, + "loss_alignment": 0.7421875, + "loss_alignment_w": 0.37109375, + "loss_sft": 0.7753499150276184, + "loss_total_v6": 1.1464436054229736, + "step": 53 + }, + { + "epoch": 0.02012672381662318, + "grad_norm": 35.95008087158203, + "learning_rate": 2.7000000000000004e-06, + "loss": 15.592812538146973, + "loss_alignment": 0.7421875, + "loss_alignment_w": 0.37109375, + "loss_sft": 0.8353683352470398, + "loss_total_v6": 1.2064621448516846, + "step": 54 + }, + { + "epoch": 0.020499440924338426, + "grad_norm": 36.34403991699219, + "learning_rate": 2.7500000000000004e-06, + "loss": 16.054763793945312, + "loss_alignment": 0.7265625, + "loss_alignment_w": 0.36328125, + "loss_sft": 0.6333206295967102, + "loss_total_v6": 0.9966018795967102, + "step": 55 + }, + { + "epoch": 0.02087215803205367, + "grad_norm": 31.066387176513672, + "learning_rate": 2.8000000000000003e-06, + "loss": 15.223623275756836, + "loss_alignment": 0.703125, + "loss_alignment_w": 0.3515625, + "loss_sft": 0.5697507858276367, + "loss_total_v6": 0.9213132858276367, + "step": 56 + }, + { + "epoch": 0.021244875139768915, + "grad_norm": 32.782875061035156, + "learning_rate": 2.85e-06, + "loss": 15.326842308044434, + "loss_alignment": 0.71875, + "loss_alignment_w": 0.359375, + "loss_sft": 0.7238661050796509, + "loss_total_v6": 1.0832411050796509, + "step": 57 + }, + { + "epoch": 0.02161759224748416, + "grad_norm": 34.250709533691406, + "learning_rate": 2.9e-06, + "loss": 15.717256546020508, + "loss_alignment": 0.70703125, + "loss_alignment_w": 0.353515625, + "loss_sft": 0.591575026512146, + "loss_total_v6": 0.945090651512146, + "step": 58 + }, + { + "epoch": 0.021990309355199404, + "grad_norm": 34.61017608642578, + "learning_rate": 2.95e-06, + "loss": 15.166189193725586, + "loss_alignment": 0.69140625, + "loss_alignment_w": 0.345703125, + "loss_sft": 0.6988440155982971, + "loss_total_v6": 1.0445470809936523, + "step": 59 + }, + { + "epoch": 0.02236302646291465, + "grad_norm": 38.98619842529297, + "learning_rate": 3e-06, + "loss": 14.897673606872559, + "loss_alignment": 0.6796875, + "loss_alignment_w": 0.33984375, + "loss_sft": 0.5975462198257446, + "loss_total_v6": 0.9373899698257446, + "step": 60 + }, + { + "epoch": 0.022735743570629893, + "grad_norm": 37.24393081665039, + "learning_rate": 3.05e-06, + "loss": 14.665550231933594, + "loss_alignment": 0.6953125, + "loss_alignment_w": 0.34765625, + "loss_sft": 0.49588656425476074, + "loss_total_v6": 0.8435428142547607, + "step": 61 + }, + { + "epoch": 0.023108460678345134, + "grad_norm": 34.912193298339844, + "learning_rate": 3.1000000000000004e-06, + "loss": 14.77892780303955, + "loss_alignment": 0.6953125, + "loss_alignment_w": 0.34765625, + "loss_sft": 0.645811140537262, + "loss_total_v6": 0.993467390537262, + "step": 62 + }, + { + "epoch": 0.02348117778606038, + "grad_norm": 32.69119644165039, + "learning_rate": 3.1500000000000003e-06, + "loss": 14.96495246887207, + "loss_alignment": 0.67578125, + "loss_alignment_w": 0.337890625, + "loss_sft": 0.49217090010643005, + "loss_total_v6": 0.8300615549087524, + "step": 63 + }, + { + "epoch": 0.023853894893775623, + "grad_norm": 33.11100769042969, + "learning_rate": 3.2000000000000003e-06, + "loss": 14.711910247802734, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5837761759757996, + "loss_total_v6": 0.5837761759757996, + "step": 64 + }, + { + "epoch": 0.024226612001490868, + "grad_norm": 34.39384841918945, + "learning_rate": 3.2500000000000002e-06, + "loss": 14.773680686950684, + "loss_alignment": 0.65625, + "loss_alignment_w": 0.328125, + "loss_sft": 0.7192052602767944, + "loss_total_v6": 1.0473302602767944, + "step": 65 + }, + { + "epoch": 0.024599329109206113, + "grad_norm": 29.624584197998047, + "learning_rate": 3.3000000000000006e-06, + "loss": 14.732731819152832, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7869117856025696, + "loss_total_v6": 0.7869117856025696, + "step": 66 + }, + { + "epoch": 0.024972046216921357, + "grad_norm": 33.57868957519531, + "learning_rate": 3.3500000000000005e-06, + "loss": 14.530835151672363, + "loss_alignment": 0.6484375, + "loss_alignment_w": 0.32421875, + "loss_sft": 0.6792049407958984, + "loss_total_v6": 1.0034236907958984, + "step": 67 + }, + { + "epoch": 0.0253447633246366, + "grad_norm": 32.31406784057617, + "learning_rate": 3.4000000000000005e-06, + "loss": 14.717488288879395, + "loss_alignment": 0.6484375, + "loss_alignment_w": 0.32421875, + "loss_sft": 0.5509804487228394, + "loss_total_v6": 0.8751991987228394, + "step": 68 + }, + { + "epoch": 0.025717480432351846, + "grad_norm": 32.3331298828125, + "learning_rate": 3.45e-06, + "loss": 14.706356048583984, + "loss_alignment": 0.640625, + "loss_alignment_w": 0.3203125, + "loss_sft": 0.5567969083786011, + "loss_total_v6": 0.8771094083786011, + "step": 69 + }, + { + "epoch": 0.02609019754006709, + "grad_norm": 38.12213134765625, + "learning_rate": 3.5e-06, + "loss": 14.228458404541016, + "loss_alignment": 0.640625, + "loss_alignment_w": 0.3203125, + "loss_sft": 0.7827085256576538, + "loss_total_v6": 1.1030210256576538, + "step": 70 + }, + { + "epoch": 0.026462914647782332, + "grad_norm": 29.69866180419922, + "learning_rate": 3.5500000000000003e-06, + "loss": 14.315841674804688, + "loss_alignment": 0.6484375, + "loss_alignment_w": 0.32421875, + "loss_sft": 0.5380566120147705, + "loss_total_v6": 0.8622753620147705, + "step": 71 + }, + { + "epoch": 0.026835631755497576, + "grad_norm": 30.201005935668945, + "learning_rate": 3.6000000000000003e-06, + "loss": 14.61131477355957, + "loss_alignment": 0.640625, + "loss_alignment_w": 0.3203125, + "loss_sft": 0.5236844420433044, + "loss_total_v6": 0.8439969420433044, + "step": 72 + }, + { + "epoch": 0.02720834886321282, + "grad_norm": 31.410850524902344, + "learning_rate": 3.65e-06, + "loss": 14.552396774291992, + "loss_alignment": 0.609375, + "loss_alignment_w": 0.3046875, + "loss_sft": 0.7219250798225403, + "loss_total_v6": 1.0266125202178955, + "step": 73 + }, + { + "epoch": 0.027581065970928065, + "grad_norm": 33.20163345336914, + "learning_rate": 3.7e-06, + "loss": 14.319782257080078, + "loss_alignment": 0.6328125, + "loss_alignment_w": 0.31640625, + "loss_sft": 0.7619980573654175, + "loss_total_v6": 1.0784043073654175, + "step": 74 + }, + { + "epoch": 0.02795378307864331, + "grad_norm": 31.179431915283203, + "learning_rate": 3.7500000000000005e-06, + "loss": 14.337654113769531, + "loss_alignment": 0.640625, + "loss_alignment_w": 0.3203125, + "loss_sft": 0.9927076697349548, + "loss_total_v6": 1.3130202293395996, + "step": 75 + }, + { + "epoch": 0.028326500186358555, + "grad_norm": 31.489347457885742, + "learning_rate": 3.8000000000000005e-06, + "loss": 14.611846923828125, + "loss_alignment": 0.6015625, + "loss_alignment_w": 0.30078125, + "loss_sft": 0.8174423575401306, + "loss_total_v6": 1.1182236671447754, + "step": 76 + }, + { + "epoch": 0.0286992172940738, + "grad_norm": 31.611051559448242, + "learning_rate": 3.85e-06, + "loss": 14.280086517333984, + "loss_alignment": 0.60546875, + "loss_alignment_w": 0.302734375, + "loss_sft": 0.49403929710388184, + "loss_total_v6": 0.7967736721038818, + "step": 77 + }, + { + "epoch": 0.029071934401789044, + "grad_norm": 30.13182258605957, + "learning_rate": 3.900000000000001e-06, + "loss": 14.259308815002441, + "loss_alignment": 0.6171875, + "loss_alignment_w": 0.30859375, + "loss_sft": 0.640002429485321, + "loss_total_v6": 0.948596179485321, + "step": 78 + }, + { + "epoch": 0.029444651509504285, + "grad_norm": 45.755714416503906, + "learning_rate": 3.95e-06, + "loss": 13.931146621704102, + "loss_alignment": 0.6328125, + "loss_alignment_w": 0.31640625, + "loss_sft": 0.6008468270301819, + "loss_total_v6": 0.9172530770301819, + "step": 79 + }, + { + "epoch": 0.02981736861721953, + "grad_norm": 29.806835174560547, + "learning_rate": 4.000000000000001e-06, + "loss": 14.22589111328125, + "loss_alignment": 0.6328125, + "loss_alignment_w": 0.31640625, + "loss_sft": 0.6780732870101929, + "loss_total_v6": 0.9944795370101929, + "step": 80 + }, + { + "epoch": 0.030190085724934774, + "grad_norm": 32.569847106933594, + "learning_rate": 4.05e-06, + "loss": 14.176970481872559, + "loss_alignment": 0.6015625, + "loss_alignment_w": 0.30078125, + "loss_sft": 0.4747808277606964, + "loss_total_v6": 0.775562047958374, + "step": 81 + }, + { + "epoch": 0.03056280283265002, + "grad_norm": 31.78152084350586, + "learning_rate": 4.1e-06, + "loss": 13.598535537719727, + "loss_alignment": 0.5859375, + "loss_alignment_w": 0.29296875, + "loss_sft": 0.4890978932380676, + "loss_total_v6": 0.7820666432380676, + "step": 82 + }, + { + "epoch": 0.030935519940365263, + "grad_norm": 30.199851989746094, + "learning_rate": 4.15e-06, + "loss": 13.957791328430176, + "loss_alignment": 0.5703125, + "loss_alignment_w": 0.28515625, + "loss_sft": 0.3347610831260681, + "loss_total_v6": 0.6199173331260681, + "step": 83 + }, + { + "epoch": 0.03130823704808051, + "grad_norm": 31.790266036987305, + "learning_rate": 4.2000000000000004e-06, + "loss": 13.885854721069336, + "loss_alignment": 0.5703125, + "loss_alignment_w": 0.28515625, + "loss_sft": 0.6924788355827332, + "loss_total_v6": 0.9776350855827332, + "step": 84 + }, + { + "epoch": 0.03168095415579575, + "grad_norm": 30.41127586364746, + "learning_rate": 4.25e-06, + "loss": 13.71098518371582, + "loss_alignment": 0.5625, + "loss_alignment_w": 0.28125, + "loss_sft": 0.26263123750686646, + "loss_total_v6": 0.5438812375068665, + "step": 85 + }, + { + "epoch": 0.032053671263511, + "grad_norm": 32.29167556762695, + "learning_rate": 4.3e-06, + "loss": 13.851821899414062, + "loss_alignment": 0.60546875, + "loss_alignment_w": 0.302734375, + "loss_sft": 0.7781162261962891, + "loss_total_v6": 1.080850601196289, + "step": 86 + }, + { + "epoch": 0.03242638837122624, + "grad_norm": 38.42442321777344, + "learning_rate": 4.350000000000001e-06, + "loss": 13.55086898803711, + "loss_alignment": 0.5625, + "loss_alignment_w": 0.28125, + "loss_sft": 0.7790597677230835, + "loss_total_v6": 1.0603097677230835, + "step": 87 + }, + { + "epoch": 0.032799105478941486, + "grad_norm": 30.731632232666016, + "learning_rate": 4.4e-06, + "loss": 13.833663940429688, + "loss_alignment": 0.546875, + "loss_alignment_w": 0.2734375, + "loss_sft": 0.47841107845306396, + "loss_total_v6": 0.751848578453064, + "step": 88 + }, + { + "epoch": 0.03317182258665673, + "grad_norm": 30.22502326965332, + "learning_rate": 4.450000000000001e-06, + "loss": 13.972530364990234, + "loss_alignment": 0.5546875, + "loss_alignment_w": 0.27734375, + "loss_sft": 0.6720701456069946, + "loss_total_v6": 0.9494138956069946, + "step": 89 + }, + { + "epoch": 0.033544539694371975, + "grad_norm": 29.825044631958008, + "learning_rate": 4.5e-06, + "loss": 13.605220794677734, + "loss_alignment": 0.53125, + "loss_alignment_w": 0.265625, + "loss_sft": 0.742428183555603, + "loss_total_v6": 1.008053183555603, + "step": 90 + }, + { + "epoch": 0.033917256802087216, + "grad_norm": 29.43014144897461, + "learning_rate": 4.5500000000000005e-06, + "loss": 13.261663436889648, + "loss_alignment": 0.53125, + "loss_alignment_w": 0.265625, + "loss_sft": 0.5261843800544739, + "loss_total_v6": 0.7918093800544739, + "step": 91 + }, + { + "epoch": 0.03428997390980246, + "grad_norm": 33.06686782836914, + "learning_rate": 4.600000000000001e-06, + "loss": 13.400988578796387, + "loss_alignment": 0.55078125, + "loss_alignment_w": 0.275390625, + "loss_sft": 0.36539435386657715, + "loss_total_v6": 0.6407849788665771, + "step": 92 + }, + { + "epoch": 0.034662691017517705, + "grad_norm": 34.67592239379883, + "learning_rate": 4.65e-06, + "loss": 13.433706283569336, + "loss_alignment": 0.515625, + "loss_alignment_w": 0.2578125, + "loss_sft": 0.3823792636394501, + "loss_total_v6": 0.6401917934417725, + "step": 93 + }, + { + "epoch": 0.035035408125232946, + "grad_norm": 33.61492156982422, + "learning_rate": 4.7e-06, + "loss": 13.366116523742676, + "loss_alignment": 0.515625, + "loss_alignment_w": 0.2578125, + "loss_sft": 0.716579258441925, + "loss_total_v6": 0.974391758441925, + "step": 94 + }, + { + "epoch": 0.035408125232948194, + "grad_norm": 32.52003860473633, + "learning_rate": 4.75e-06, + "loss": 13.789382934570312, + "loss_alignment": 0.54296875, + "loss_alignment_w": 0.271484375, + "loss_sft": 0.7586503624916077, + "loss_total_v6": 1.030134677886963, + "step": 95 + }, + { + "epoch": 0.035780842340663435, + "grad_norm": 31.057945251464844, + "learning_rate": 4.800000000000001e-06, + "loss": 13.756771087646484, + "loss_alignment": 0.53125, + "loss_alignment_w": 0.265625, + "loss_sft": 0.702137291431427, + "loss_total_v6": 0.967762291431427, + "step": 96 + }, + { + "epoch": 0.03615355944837868, + "grad_norm": 31.591094970703125, + "learning_rate": 4.85e-06, + "loss": 13.001449584960938, + "loss_alignment": 0.55859375, + "loss_alignment_w": 0.279296875, + "loss_sft": 0.6719346046447754, + "loss_total_v6": 0.9512314796447754, + "step": 97 + }, + { + "epoch": 0.036526276556093924, + "grad_norm": 29.6240234375, + "learning_rate": 4.9000000000000005e-06, + "loss": 12.929828643798828, + "loss_alignment": 0.5078125, + "loss_alignment_w": 0.25390625, + "loss_sft": 0.5389934778213501, + "loss_total_v6": 0.7928997278213501, + "step": 98 + }, + { + "epoch": 0.03689899366380917, + "grad_norm": 29.43594741821289, + "learning_rate": 4.95e-06, + "loss": 13.398805618286133, + "loss_alignment": 0.5078125, + "loss_alignment_w": 0.25390625, + "loss_sft": 0.7576403021812439, + "loss_total_v6": 1.0115466117858887, + "step": 99 + }, + { + "epoch": 0.03727171077152441, + "grad_norm": 30.13565444946289, + "learning_rate": 5e-06, + "loss": 13.506148338317871, + "loss_alignment": 0.5, + "loss_alignment_w": 0.25, + "loss_sft": 0.6251489520072937, + "loss_total_v6": 0.8751489520072937, + "step": 100 + }, + { + "epoch": 0.03727171077152441, + "eval_loss": 0.7494901418685913, + "eval_loss_alignment": 0.4709973457532051, + "eval_loss_alignment_w": 0.23549867287660256, + "eval_loss_sft": 0.5139914752804061, + "eval_loss_total_v6": 0.7494901480614884, + "eval_runtime": 206.2917, + "eval_samples_per_second": 9.045, + "eval_steps_per_second": 1.134, + "step": 100 + }, + { + "epoch": 0.037644427879239654, + "grad_norm": 27.860788345336914, + "learning_rate": 5.050000000000001e-06, + "loss": 13.680850982666016, + "loss_alignment": 0.49609375, + "loss_alignment_w": 0.248046875, + "loss_sft": 0.7272000908851624, + "loss_total_v6": 0.9752469658851624, + "step": 101 + }, + { + "epoch": 0.0380171449869549, + "grad_norm": 30.027984619140625, + "learning_rate": 5.1e-06, + "loss": 13.015209197998047, + "loss_alignment": 0.484375, + "loss_alignment_w": 0.2421875, + "loss_sft": 0.3715229630470276, + "loss_total_v6": 0.6137104630470276, + "step": 102 + }, + { + "epoch": 0.038389862094670144, + "grad_norm": 30.74447250366211, + "learning_rate": 5.150000000000001e-06, + "loss": 12.826108932495117, + "loss_alignment": 0.4765625, + "loss_alignment_w": 0.23828125, + "loss_sft": 0.6347817778587341, + "loss_total_v6": 0.8730630278587341, + "step": 103 + }, + { + "epoch": 0.03876257920238539, + "grad_norm": 34.92304229736328, + "learning_rate": 5.2e-06, + "loss": 12.928760528564453, + "loss_alignment": 0.4921875, + "loss_alignment_w": 0.24609375, + "loss_sft": 0.5967477560043335, + "loss_total_v6": 0.8428415060043335, + "step": 104 + }, + { + "epoch": 0.03913529631010063, + "grad_norm": 49.34968185424805, + "learning_rate": 5.2500000000000006e-06, + "loss": 12.856857299804688, + "loss_alignment": 0.47265625, + "loss_alignment_w": 0.236328125, + "loss_sft": 0.5897846817970276, + "loss_total_v6": 0.8261128067970276, + "step": 105 + }, + { + "epoch": 0.03950801341781588, + "grad_norm": 36.31321334838867, + "learning_rate": 5.300000000000001e-06, + "loss": 12.942293167114258, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.7296481728553772, + "loss_total_v6": 0.9562106728553772, + "step": 106 + }, + { + "epoch": 0.03988073052553112, + "grad_norm": 144.69345092773438, + "learning_rate": 5.3500000000000004e-06, + "loss": 13.220405578613281, + "loss_alignment": 0.48828125, + "loss_alignment_w": 0.244140625, + "loss_sft": 0.6385214924812317, + "loss_total_v6": 0.8826621174812317, + "step": 107 + }, + { + "epoch": 0.04025344763324636, + "grad_norm": 30.892560958862305, + "learning_rate": 5.400000000000001e-06, + "loss": 12.789255142211914, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.6118764281272888, + "loss_total_v6": 0.8364858031272888, + "step": 108 + }, + { + "epoch": 0.04062616474096161, + "grad_norm": 28.053911209106445, + "learning_rate": 5.450000000000001e-06, + "loss": 12.683679580688477, + "loss_alignment": 0.4921875, + "loss_alignment_w": 0.24609375, + "loss_sft": 0.7137305736541748, + "loss_total_v6": 0.9598243236541748, + "step": 109 + }, + { + "epoch": 0.04099888184867685, + "grad_norm": 27.57847023010254, + "learning_rate": 5.500000000000001e-06, + "loss": 12.81431770324707, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.6941633224487305, + "loss_total_v6": 0.9207258224487305, + "step": 110 + }, + { + "epoch": 0.0413715989563921, + "grad_norm": 27.08588981628418, + "learning_rate": 5.550000000000001e-06, + "loss": 12.76923942565918, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.4223923683166504, + "loss_total_v6": 0.6489548683166504, + "step": 111 + }, + { + "epoch": 0.04174431606410734, + "grad_norm": 29.203020095825195, + "learning_rate": 5.600000000000001e-06, + "loss": 12.848174095153809, + "loss_alignment": 0.484375, + "loss_alignment_w": 0.2421875, + "loss_sft": 0.4605550169944763, + "loss_total_v6": 0.7027425169944763, + "step": 112 + }, + { + "epoch": 0.04211703317182259, + "grad_norm": 28.39984703063965, + "learning_rate": 5.65e-06, + "loss": 12.971609115600586, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.6137115359306335, + "loss_total_v6": 0.8402740359306335, + "step": 113 + }, + { + "epoch": 0.04248975027953783, + "grad_norm": 28.384878158569336, + "learning_rate": 5.7e-06, + "loss": 12.5191650390625, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.34641551971435547, + "loss_total_v6": 0.5612592697143555, + "step": 114 + }, + { + "epoch": 0.04286246738725308, + "grad_norm": 34.94135284423828, + "learning_rate": 5.75e-06, + "loss": 12.586258888244629, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.839617133140564, + "loss_total_v6": 1.056414008140564, + "step": 115 + }, + { + "epoch": 0.04323518449496832, + "grad_norm": 29.2644100189209, + "learning_rate": 5.8e-06, + "loss": 12.748828887939453, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.542955756187439, + "loss_total_v6": 0.759752631187439, + "step": 116 + }, + { + "epoch": 0.04360790160268356, + "grad_norm": 29.225252151489258, + "learning_rate": 5.85e-06, + "loss": 12.569576263427734, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.4425225555896759, + "loss_total_v6": 0.6651787757873535, + "step": 117 + }, + { + "epoch": 0.04398061871039881, + "grad_norm": 29.76346206665039, + "learning_rate": 5.9e-06, + "loss": 12.79169750213623, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.5814849138259888, + "loss_total_v6": 0.8021880388259888, + "step": 118 + }, + { + "epoch": 0.04435333581811405, + "grad_norm": 30.76055145263672, + "learning_rate": 5.950000000000001e-06, + "loss": 12.674832344055176, + "loss_alignment": 0.39453125, + "loss_alignment_w": 0.197265625, + "loss_sft": 0.5986208319664001, + "loss_total_v6": 0.7958864569664001, + "step": 119 + }, + { + "epoch": 0.0447260529258293, + "grad_norm": 27.756593704223633, + "learning_rate": 6e-06, + "loss": 12.073689460754395, + "loss_alignment": 0.40625, + "loss_alignment_w": 0.203125, + "loss_sft": 0.43744969367980957, + "loss_total_v6": 0.6405746936798096, + "step": 120 + }, + { + "epoch": 0.04509877003354454, + "grad_norm": 31.499536514282227, + "learning_rate": 6.0500000000000005e-06, + "loss": 12.521804809570312, + "loss_alignment": 0.39453125, + "loss_alignment_w": 0.197265625, + "loss_sft": 0.5809541940689087, + "loss_total_v6": 0.7782198190689087, + "step": 121 + }, + { + "epoch": 0.045471487141259787, + "grad_norm": 31.931222915649414, + "learning_rate": 6.1e-06, + "loss": 12.28573226928711, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.6246975064277649, + "loss_total_v6": 0.8356350064277649, + "step": 122 + }, + { + "epoch": 0.04584420424897503, + "grad_norm": 103.48419189453125, + "learning_rate": 6.15e-06, + "loss": 12.678766250610352, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.545957088470459, + "loss_total_v6": 0.762753963470459, + "step": 123 + }, + { + "epoch": 0.04621692135669027, + "grad_norm": 27.835554122924805, + "learning_rate": 6.200000000000001e-06, + "loss": 12.786260604858398, + "loss_alignment": 0.39453125, + "loss_alignment_w": 0.197265625, + "loss_sft": 0.37255918979644775, + "loss_total_v6": 0.5698248147964478, + "step": 124 + }, + { + "epoch": 0.04658963846440552, + "grad_norm": 28.72994613647461, + "learning_rate": 6.25e-06, + "loss": 11.95068359375, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.6981203556060791, + "loss_total_v6": 0.9149172306060791, + "step": 125 + }, + { + "epoch": 0.04696235557212076, + "grad_norm": 35.61981201171875, + "learning_rate": 6.300000000000001e-06, + "loss": 12.221455574035645, + "loss_alignment": 0.390625, + "loss_alignment_w": 0.1953125, + "loss_sft": 0.5833699703216553, + "loss_total_v6": 0.7786824703216553, + "step": 126 + }, + { + "epoch": 0.047335072679836006, + "grad_norm": 28.727388381958008, + "learning_rate": 6.35e-06, + "loss": 11.854105949401855, + "loss_alignment": 0.39453125, + "loss_alignment_w": 0.197265625, + "loss_sft": 0.5008569359779358, + "loss_total_v6": 0.6981225609779358, + "step": 127 + }, + { + "epoch": 0.04770778978755125, + "grad_norm": 27.57211685180664, + "learning_rate": 6.4000000000000006e-06, + "loss": 12.252575874328613, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.497096985578537, + "loss_total_v6": 0.6709251403808594, + "step": 128 + }, + { + "epoch": 0.048080506895266495, + "grad_norm": 28.564050674438477, + "learning_rate": 6.450000000000001e-06, + "loss": 11.741657257080078, + "loss_alignment": 0.38671875, + "loss_alignment_w": 0.193359375, + "loss_sft": 0.35097891092300415, + "loss_total_v6": 0.5443382859230042, + "step": 129 + }, + { + "epoch": 0.048453224002981736, + "grad_norm": 29.248882293701172, + "learning_rate": 6.5000000000000004e-06, + "loss": 11.822461128234863, + "loss_alignment": 0.37890625, + "loss_alignment_w": 0.189453125, + "loss_sft": 0.4472099840641022, + "loss_total_v6": 0.6366630792617798, + "step": 130 + }, + { + "epoch": 0.048825941110696984, + "grad_norm": 28.068803787231445, + "learning_rate": 6.550000000000001e-06, + "loss": 11.702528953552246, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.5506353974342346, + "loss_total_v6": 0.7381353974342346, + "step": 131 + }, + { + "epoch": 0.049198658218412225, + "grad_norm": 29.390419006347656, + "learning_rate": 6.600000000000001e-06, + "loss": 11.977767944335938, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.5332278609275818, + "loss_total_v6": 0.7207278609275818, + "step": 132 + }, + { + "epoch": 0.049571375326127466, + "grad_norm": 28.372915267944336, + "learning_rate": 6.650000000000001e-06, + "loss": 11.886527061462402, + "loss_alignment": 0.3828125, + "loss_alignment_w": 0.19140625, + "loss_sft": 0.7204277515411377, + "loss_total_v6": 0.9118340015411377, + "step": 133 + }, + { + "epoch": 0.049944092433842714, + "grad_norm": 35.91057205200195, + "learning_rate": 6.700000000000001e-06, + "loss": 11.816113471984863, + "loss_alignment": 0.3671875, + "loss_alignment_w": 0.18359375, + "loss_sft": 0.6684812307357788, + "loss_total_v6": 0.8520749807357788, + "step": 134 + }, + { + "epoch": 0.050316809541557955, + "grad_norm": 28.6835880279541, + "learning_rate": 6.750000000000001e-06, + "loss": 11.838126182556152, + "loss_alignment": 0.3671875, + "loss_alignment_w": 0.18359375, + "loss_sft": 0.610861599445343, + "loss_total_v6": 0.794455349445343, + "step": 135 + }, + { + "epoch": 0.0506895266492732, + "grad_norm": 28.922822952270508, + "learning_rate": 6.800000000000001e-06, + "loss": 11.980234146118164, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.6661719679832458, + "loss_total_v6": 0.8360938429832458, + "step": 136 + }, + { + "epoch": 0.051062243756988444, + "grad_norm": 27.56918716430664, + "learning_rate": 6.850000000000001e-06, + "loss": 12.30390739440918, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.643670916557312, + "loss_total_v6": 0.829217791557312, + "step": 137 + }, + { + "epoch": 0.05143496086470369, + "grad_norm": 29.505952835083008, + "learning_rate": 6.9e-06, + "loss": 12.537721633911133, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.5682717561721802, + "loss_total_v6": 0.7479592561721802, + "step": 138 + }, + { + "epoch": 0.05180767797241893, + "grad_norm": 28.955163955688477, + "learning_rate": 6.95e-06, + "loss": 11.721149444580078, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.6527304649353027, + "loss_total_v6": 0.8343710899353027, + "step": 139 + }, + { + "epoch": 0.05218039508013418, + "grad_norm": 28.25800132751465, + "learning_rate": 7e-06, + "loss": 11.733381271362305, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.42602697014808655, + "loss_total_v6": 0.6115738153457642, + "step": 140 + }, + { + "epoch": 0.05255311218784942, + "grad_norm": 49.87938690185547, + "learning_rate": 7.05e-06, + "loss": 11.911565780639648, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.6913131475448608, + "loss_total_v6": 0.8729537725448608, + "step": 141 + }, + { + "epoch": 0.052925829295564664, + "grad_norm": 28.469282150268555, + "learning_rate": 7.100000000000001e-06, + "loss": 12.01460075378418, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.5847467184066772, + "loss_total_v6": 0.7624810934066772, + "step": 142 + }, + { + "epoch": 0.05329854640327991, + "grad_norm": 27.827333450317383, + "learning_rate": 7.15e-06, + "loss": 12.07193374633789, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6647759675979614, + "loss_total_v6": 0.8327447175979614, + "step": 143 + }, + { + "epoch": 0.05367126351099515, + "grad_norm": 27.698406219482422, + "learning_rate": 7.2000000000000005e-06, + "loss": 11.531166076660156, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.5996710062026978, + "loss_total_v6": 0.7695928812026978, + "step": 144 + }, + { + "epoch": 0.0540439806187104, + "grad_norm": 33.61573028564453, + "learning_rate": 7.25e-06, + "loss": 11.806937217712402, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6785311698913574, + "loss_total_v6": 0.8425936698913574, + "step": 145 + }, + { + "epoch": 0.05441669772642564, + "grad_norm": 28.504987716674805, + "learning_rate": 7.3e-06, + "loss": 11.779170989990234, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.47392451763153076, + "loss_total_v6": 0.6457995176315308, + "step": 146 + }, + { + "epoch": 0.05478941483414089, + "grad_norm": 25.289932250976562, + "learning_rate": 7.350000000000001e-06, + "loss": 11.836795806884766, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.7268218398094177, + "loss_total_v6": 0.8752593398094177, + "step": 147 + }, + { + "epoch": 0.05516213194185613, + "grad_norm": 25.70249366760254, + "learning_rate": 7.4e-06, + "loss": 11.754301071166992, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.2766382098197937, + "loss_total_v6": 0.4407007098197937, + "step": 148 + }, + { + "epoch": 0.05553484904957137, + "grad_norm": 26.032983779907227, + "learning_rate": 7.450000000000001e-06, + "loss": 11.873943328857422, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.6586533784866333, + "loss_total_v6": 0.8324815034866333, + "step": 149 + }, + { + "epoch": 0.05590756615728662, + "grad_norm": 27.88697052001953, + "learning_rate": 7.500000000000001e-06, + "loss": 11.607440948486328, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.38458940386772156, + "loss_total_v6": 0.38458940386772156, + "step": 150 + }, + { + "epoch": 0.05628028326500186, + "grad_norm": 26.79559898376465, + "learning_rate": 7.5500000000000006e-06, + "loss": 11.820796012878418, + "loss_alignment": 0.3515625, + "loss_alignment_w": 0.17578125, + "loss_sft": 0.6376487016677856, + "loss_total_v6": 0.8134299516677856, + "step": 151 + }, + { + "epoch": 0.05665300037271711, + "grad_norm": 32.74031066894531, + "learning_rate": 7.600000000000001e-06, + "loss": 11.847285270690918, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.2993646562099457, + "loss_total_v6": 0.4595209062099457, + "step": 152 + }, + { + "epoch": 0.05702571748043235, + "grad_norm": 25.251291275024414, + "learning_rate": 7.650000000000001e-06, + "loss": 11.215269088745117, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.6646570563316345, + "loss_total_v6": 0.8384851813316345, + "step": 153 + }, + { + "epoch": 0.0573984345881476, + "grad_norm": 25.49034309387207, + "learning_rate": 7.7e-06, + "loss": 11.289850234985352, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.2649202346801758, + "loss_total_v6": 0.4367952346801758, + "step": 154 + }, + { + "epoch": 0.05777115169586284, + "grad_norm": 30.095701217651367, + "learning_rate": 7.75e-06, + "loss": 11.624927520751953, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.4943125545978546, + "loss_total_v6": 0.6642343997955322, + "step": 155 + }, + { + "epoch": 0.05814386880357809, + "grad_norm": 40.37640380859375, + "learning_rate": 7.800000000000002e-06, + "loss": 11.813579559326172, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4870538115501404, + "loss_total_v6": 0.6491631865501404, + "step": 156 + }, + { + "epoch": 0.05851658591129333, + "grad_norm": 28.424407958984375, + "learning_rate": 7.850000000000001e-06, + "loss": 11.470141410827637, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7225965857505798, + "loss_total_v6": 0.8886122107505798, + "step": 157 + }, + { + "epoch": 0.05888930301900857, + "grad_norm": 25.606475830078125, + "learning_rate": 7.9e-06, + "loss": 11.99289321899414, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.5608013868331909, + "loss_total_v6": 0.7326763868331909, + "step": 158 + }, + { + "epoch": 0.05926202012672382, + "grad_norm": 25.20802116394043, + "learning_rate": 7.950000000000002e-06, + "loss": 11.377740859985352, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.35605838894844055, + "loss_total_v6": 0.5220739841461182, + "step": 159 + }, + { + "epoch": 0.05963473723443906, + "grad_norm": 24.68560791015625, + "learning_rate": 8.000000000000001e-06, + "loss": 11.657951354980469, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.7064711451530457, + "loss_total_v6": 0.8685805201530457, + "step": 160 + }, + { + "epoch": 0.06000745434215431, + "grad_norm": 25.21956443786621, + "learning_rate": 8.050000000000001e-06, + "loss": 11.740341186523438, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7041884660720825, + "loss_total_v6": 0.8604384660720825, + "step": 161 + }, + { + "epoch": 0.06038017144986955, + "grad_norm": 24.80145835876465, + "learning_rate": 8.1e-06, + "loss": 11.3768892288208, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.48268336057662964, + "loss_total_v6": 0.6428396105766296, + "step": 162 + }, + { + "epoch": 0.060752888557584796, + "grad_norm": 26.111114501953125, + "learning_rate": 8.15e-06, + "loss": 11.269079208374023, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.7461106181144714, + "loss_total_v6": 0.9590012431144714, + "step": 163 + }, + { + "epoch": 0.06112560566530004, + "grad_norm": 24.681116104125977, + "learning_rate": 8.2e-06, + "loss": 11.46795654296875, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3394290506839752, + "loss_total_v6": 0.4898196756839752, + "step": 164 + }, + { + "epoch": 0.061498322773015285, + "grad_norm": 24.46356964111328, + "learning_rate": 8.25e-06, + "loss": 11.460844993591309, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.366894394159317, + "loss_total_v6": 0.5250974893569946, + "step": 165 + }, + { + "epoch": 0.061871039880730526, + "grad_norm": 23.259138107299805, + "learning_rate": 8.3e-06, + "loss": 11.629441261291504, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.514498770236969, + "loss_total_v6": 0.674655020236969, + "step": 166 + }, + { + "epoch": 0.06224375698844577, + "grad_norm": 25.26540184020996, + "learning_rate": 8.35e-06, + "loss": 11.584936141967773, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.5269323587417603, + "loss_total_v6": 0.7066198587417603, + "step": 167 + }, + { + "epoch": 0.06261647409616102, + "grad_norm": 25.013046264648438, + "learning_rate": 8.400000000000001e-06, + "loss": 11.307807922363281, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5837149024009705, + "loss_total_v6": 0.7419180274009705, + "step": 168 + }, + { + "epoch": 0.06298919120387626, + "grad_norm": 29.969247817993164, + "learning_rate": 8.45e-06, + "loss": 11.388710021972656, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6387720108032227, + "loss_total_v6": 0.7989282608032227, + "step": 169 + }, + { + "epoch": 0.0633619083115915, + "grad_norm": 26.087955474853516, + "learning_rate": 8.5e-06, + "loss": 11.561959266662598, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5673043727874756, + "loss_total_v6": 0.7352731227874756, + "step": 170 + }, + { + "epoch": 0.06373462541930675, + "grad_norm": 29.59478187561035, + "learning_rate": 8.550000000000001e-06, + "loss": 11.32304859161377, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.69902503490448, + "loss_total_v6": 0.86113440990448, + "step": 171 + }, + { + "epoch": 0.064107342527022, + "grad_norm": 25.50128173828125, + "learning_rate": 8.6e-06, + "loss": 11.248335838317871, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5480064749717712, + "loss_total_v6": 0.7159752249717712, + "step": 172 + }, + { + "epoch": 0.06448005963473724, + "grad_norm": 24.937849044799805, + "learning_rate": 8.65e-06, + "loss": 11.987014770507812, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.8008258938789368, + "loss_total_v6": 0.9746540188789368, + "step": 173 + }, + { + "epoch": 0.06485277674245248, + "grad_norm": 25.220277786254883, + "learning_rate": 8.700000000000001e-06, + "loss": 10.919715881347656, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.7366300225257874, + "loss_total_v6": 0.9045987725257874, + "step": 174 + }, + { + "epoch": 0.06522549385016772, + "grad_norm": 24.569435119628906, + "learning_rate": 8.750000000000001e-06, + "loss": 11.327447891235352, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.30409935116767883, + "loss_total_v6": 0.45839622616767883, + "step": 175 + }, + { + "epoch": 0.06559821095788297, + "grad_norm": 25.39592742919922, + "learning_rate": 8.8e-06, + "loss": 11.150741577148438, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6707674264907837, + "loss_total_v6": 0.8289705514907837, + "step": 176 + }, + { + "epoch": 0.0659709280655982, + "grad_norm": 27.384084701538086, + "learning_rate": 8.85e-06, + "loss": 11.561264038085938, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5298512578010559, + "loss_total_v6": 0.6939137578010559, + "step": 177 + }, + { + "epoch": 0.06634364517331345, + "grad_norm": 25.057832717895508, + "learning_rate": 8.900000000000001e-06, + "loss": 11.412894248962402, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.3204365074634552, + "loss_total_v6": 0.4942646324634552, + "step": 178 + }, + { + "epoch": 0.0667163622810287, + "grad_norm": 24.770795822143555, + "learning_rate": 8.95e-06, + "loss": 11.285418510437012, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5028116106987, + "loss_total_v6": 0.6707803606987, + "step": 179 + }, + { + "epoch": 0.06708907938874395, + "grad_norm": 26.431163787841797, + "learning_rate": 9e-06, + "loss": 11.051408767700195, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.404610812664032, + "loss_total_v6": 0.566720187664032, + "step": 180 + }, + { + "epoch": 0.06746179649645918, + "grad_norm": 24.699811935424805, + "learning_rate": 9.050000000000001e-06, + "loss": 11.298482894897461, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.49191027879714966, + "loss_total_v6": 0.6559727787971497, + "step": 181 + }, + { + "epoch": 0.06783451360417443, + "grad_norm": 26.159475326538086, + "learning_rate": 9.100000000000001e-06, + "loss": 11.428911209106445, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6183590292930603, + "loss_total_v6": 0.7843746542930603, + "step": 182 + }, + { + "epoch": 0.06820723071188968, + "grad_norm": 26.096853256225586, + "learning_rate": 9.15e-06, + "loss": 11.72265625, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6602779030799866, + "loss_total_v6": 0.8282466530799866, + "step": 183 + }, + { + "epoch": 0.06857994781960491, + "grad_norm": 27.702289581298828, + "learning_rate": 9.200000000000002e-06, + "loss": 11.532999038696289, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.6006656885147095, + "loss_total_v6": 0.7823063135147095, + "step": 184 + }, + { + "epoch": 0.06895266492732016, + "grad_norm": 26.504289627075195, + "learning_rate": 9.250000000000001e-06, + "loss": 11.597881317138672, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.8003504276275635, + "loss_total_v6": 0.9624598026275635, + "step": 185 + }, + { + "epoch": 0.06932538203503541, + "grad_norm": 25.84273910522461, + "learning_rate": 9.3e-06, + "loss": 11.46905517578125, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6599860191345215, + "loss_total_v6": 0.8220953941345215, + "step": 186 + }, + { + "epoch": 0.06969809914275066, + "grad_norm": 30.34550666809082, + "learning_rate": 9.350000000000002e-06, + "loss": 11.420743942260742, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5115851759910583, + "loss_total_v6": 0.6756476759910583, + "step": 187 + }, + { + "epoch": 0.07007081625046589, + "grad_norm": 25.030258178710938, + "learning_rate": 9.4e-06, + "loss": 11.22921085357666, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6152749061584473, + "loss_total_v6": 0.7773842811584473, + "step": 188 + }, + { + "epoch": 0.07044353335818114, + "grad_norm": 26.156705856323242, + "learning_rate": 9.450000000000001e-06, + "loss": 11.244671821594238, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.801915168762207, + "loss_total_v6": 0.960118293762207, + "step": 189 + }, + { + "epoch": 0.07081625046589639, + "grad_norm": 28.65916633605957, + "learning_rate": 9.5e-06, + "loss": 11.344503402709961, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4717220366001129, + "loss_total_v6": 0.6299251317977905, + "step": 190 + }, + { + "epoch": 0.07118896757361162, + "grad_norm": 26.117918014526367, + "learning_rate": 9.55e-06, + "loss": 11.34701919555664, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5144450664520264, + "loss_total_v6": 0.6746013164520264, + "step": 191 + }, + { + "epoch": 0.07156168468132687, + "grad_norm": 24.867666244506836, + "learning_rate": 9.600000000000001e-06, + "loss": 11.617500305175781, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.49247071146965027, + "loss_total_v6": 0.6506738662719727, + "step": 192 + }, + { + "epoch": 0.07193440178904212, + "grad_norm": 28.156047821044922, + "learning_rate": 9.65e-06, + "loss": 10.991273880004883, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5448174476623535, + "loss_total_v6": 0.7069268226623535, + "step": 193 + }, + { + "epoch": 0.07230711889675737, + "grad_norm": 26.55216407775879, + "learning_rate": 9.7e-06, + "loss": 11.804953575134277, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6154850125312805, + "loss_total_v6": 0.7775943875312805, + "step": 194 + }, + { + "epoch": 0.0726798360044726, + "grad_norm": 25.263093948364258, + "learning_rate": 9.75e-06, + "loss": 11.782463073730469, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6184107661247253, + "loss_total_v6": 0.7863795161247253, + "step": 195 + }, + { + "epoch": 0.07305255311218785, + "grad_norm": 23.686431884765625, + "learning_rate": 9.800000000000001e-06, + "loss": 11.212419509887695, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.535455584526062, + "loss_total_v6": 0.693658709526062, + "step": 196 + }, + { + "epoch": 0.0734252702199031, + "grad_norm": 24.300321578979492, + "learning_rate": 9.85e-06, + "loss": 11.592643737792969, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.7063671946525574, + "loss_total_v6": 0.8645703196525574, + "step": 197 + }, + { + "epoch": 0.07379798732761834, + "grad_norm": 23.51018714904785, + "learning_rate": 9.9e-06, + "loss": 11.064301490783691, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.858885645866394, + "loss_total_v6": 1.015135645866394, + "step": 198 + }, + { + "epoch": 0.07417070443533358, + "grad_norm": 25.55925178527832, + "learning_rate": 9.950000000000001e-06, + "loss": 11.55112075805664, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6860422492027283, + "loss_total_v6": 0.8442453742027283, + "step": 199 + }, + { + "epoch": 0.07454342154304883, + "grad_norm": 24.758529663085938, + "learning_rate": 1e-05, + "loss": 11.171196937561035, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.3951422870159149, + "loss_total_v6": 0.5592048168182373, + "step": 200 + }, + { + "epoch": 0.07454342154304883, + "eval_loss": 0.6634879112243652, + "eval_loss_alignment": 0.298740484775641, + "eval_loss_alignment_w": 0.1493702423878205, + "eval_loss_sft": 0.5141176704285492, + "eval_loss_total_v6": 0.6634879127845296, + "eval_runtime": 205.4662, + "eval_samples_per_second": 9.082, + "eval_steps_per_second": 1.139, + "step": 200 + }, + { + "epoch": 0.07491613865076407, + "grad_norm": 82.2469711303711, + "learning_rate": 9.999995997915474e-06, + "loss": 11.471867561340332, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5297922492027283, + "loss_total_v6": 0.6938547492027283, + "step": 201 + }, + { + "epoch": 0.07528885575847931, + "grad_norm": 30.21335792541504, + "learning_rate": 9.9999839916683e-06, + "loss": 11.435491561889648, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.6761285066604614, + "loss_total_v6": 0.8460503816604614, + "step": 202 + }, + { + "epoch": 0.07566157286619456, + "grad_norm": 25.028169631958008, + "learning_rate": 9.999963981277698e-06, + "loss": 11.507705688476562, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6041499376296997, + "loss_total_v6": 0.7682124376296997, + "step": 203 + }, + { + "epoch": 0.0760342899739098, + "grad_norm": 24.61888313293457, + "learning_rate": 9.999935966775703e-06, + "loss": 11.363700866699219, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.39819905161857605, + "loss_total_v6": 0.5564022064208984, + "step": 204 + }, + { + "epoch": 0.07640700708162505, + "grad_norm": 27.175294876098633, + "learning_rate": 9.99989994820716e-06, + "loss": 11.219963073730469, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6984971761703491, + "loss_total_v6": 0.8547471761703491, + "step": 205 + }, + { + "epoch": 0.07677972418934029, + "grad_norm": 25.966838836669922, + "learning_rate": 9.999855925629728e-06, + "loss": 11.368810653686523, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6108423471450806, + "loss_total_v6": 0.7690454721450806, + "step": 206 + }, + { + "epoch": 0.07715244129705554, + "grad_norm": 24.834362030029297, + "learning_rate": 9.999803899113885e-06, + "loss": 11.34504222869873, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.27429431676864624, + "loss_total_v6": 0.43249744176864624, + "step": 207 + }, + { + "epoch": 0.07752515840477078, + "grad_norm": 24.66141128540039, + "learning_rate": 9.999743868742911e-06, + "loss": 11.895995140075684, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 1.0008925199508667, + "loss_total_v6": 1.1805800199508667, + "step": 208 + }, + { + "epoch": 0.07789787551248602, + "grad_norm": 25.498186111450195, + "learning_rate": 9.999675834612905e-06, + "loss": 11.15536117553711, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.45804136991500854, + "loss_total_v6": 0.6240569949150085, + "step": 209 + }, + { + "epoch": 0.07827059262020127, + "grad_norm": 26.67243194580078, + "learning_rate": 9.999599796832781e-06, + "loss": 11.570856094360352, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6289753913879395, + "loss_total_v6": 0.7910847663879395, + "step": 210 + }, + { + "epoch": 0.07864330972791651, + "grad_norm": 25.312877655029297, + "learning_rate": 9.999515755524262e-06, + "loss": 12.050422668457031, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3403167426586151, + "loss_total_v6": 0.4985198676586151, + "step": 211 + }, + { + "epoch": 0.07901602683563176, + "grad_norm": 23.823875427246094, + "learning_rate": 9.999423710821886e-06, + "loss": 11.569028854370117, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.40141454339027405, + "loss_total_v6": 0.5615707635879517, + "step": 212 + }, + { + "epoch": 0.079388743943347, + "grad_norm": 23.90646743774414, + "learning_rate": 9.999323662872998e-06, + "loss": 11.154228210449219, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.3794401288032532, + "loss_total_v6": 0.5337370038032532, + "step": 213 + }, + { + "epoch": 0.07976146105106224, + "grad_norm": 23.08623504638672, + "learning_rate": 9.999215611837758e-06, + "loss": 11.258766174316406, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.55023592710495, + "loss_total_v6": 0.71039217710495, + "step": 214 + }, + { + "epoch": 0.08013417815877749, + "grad_norm": 21.28853416442871, + "learning_rate": 9.999099557889141e-06, + "loss": 11.340235710144043, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7091361880302429, + "loss_total_v6": 0.8634330630302429, + "step": 215 + }, + { + "epoch": 0.08050689526649273, + "grad_norm": 23.92483901977539, + "learning_rate": 9.998975501212926e-06, + "loss": 11.262042045593262, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3483208119869232, + "loss_total_v6": 0.5045708417892456, + "step": 216 + }, + { + "epoch": 0.08087961237420797, + "grad_norm": 24.75493812561035, + "learning_rate": 9.998843442007712e-06, + "loss": 11.471992492675781, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.9658172130584717, + "loss_total_v6": 1.1259734630584717, + "step": 217 + }, + { + "epoch": 0.08125232948192322, + "grad_norm": 24.464506149291992, + "learning_rate": 9.9987033804849e-06, + "loss": 11.190614700317383, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.590316653251648, + "loss_total_v6": 0.752426028251648, + "step": 218 + }, + { + "epoch": 0.08162504658963847, + "grad_norm": 23.85846710205078, + "learning_rate": 9.998555316868706e-06, + "loss": 11.413606643676758, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.5397694110870361, + "loss_total_v6": 0.7057850360870361, + "step": 219 + }, + { + "epoch": 0.0819977636973537, + "grad_norm": 29.281993865966797, + "learning_rate": 9.998399251396156e-06, + "loss": 11.451358795166016, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.7008848190307617, + "loss_total_v6": 0.8629941940307617, + "step": 220 + }, + { + "epoch": 0.08237048080506895, + "grad_norm": 22.98732566833496, + "learning_rate": 9.998235184317086e-06, + "loss": 11.654879570007324, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5661384463310242, + "loss_total_v6": 0.7204353213310242, + "step": 221 + }, + { + "epoch": 0.0827431979127842, + "grad_norm": 23.05590057373047, + "learning_rate": 9.998063115894137e-06, + "loss": 11.62356185913086, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.37103787064552307, + "loss_total_v6": 0.5155690908432007, + "step": 222 + }, + { + "epoch": 0.08311591502049943, + "grad_norm": 25.558683395385742, + "learning_rate": 9.997883046402763e-06, + "loss": 11.570382118225098, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5908209681510925, + "loss_total_v6": 0.7470709681510925, + "step": 223 + }, + { + "epoch": 0.08348863212821468, + "grad_norm": 24.020353317260742, + "learning_rate": 9.997694976131229e-06, + "loss": 11.375284194946289, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.42364856600761414, + "loss_total_v6": 0.5877110958099365, + "step": 224 + }, + { + "epoch": 0.08386134923592993, + "grad_norm": 29.76023292541504, + "learning_rate": 9.997498905380598e-06, + "loss": 11.361135482788086, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.7145047783851624, + "loss_total_v6": 0.8727079033851624, + "step": 225 + }, + { + "epoch": 0.08423406634364518, + "grad_norm": 26.742464065551758, + "learning_rate": 9.99729483446475e-06, + "loss": 11.386390686035156, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7799468040466309, + "loss_total_v6": 0.9459624290466309, + "step": 226 + }, + { + "epoch": 0.08460678345136041, + "grad_norm": 24.72962188720703, + "learning_rate": 9.99708276371037e-06, + "loss": 10.98497200012207, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47892922163009644, + "loss_total_v6": 0.6234604716300964, + "step": 227 + }, + { + "epoch": 0.08497950055907566, + "grad_norm": 27.012969970703125, + "learning_rate": 9.996862693456948e-06, + "loss": 11.726984024047852, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6790594458580017, + "loss_total_v6": 0.8333563208580017, + "step": 228 + }, + { + "epoch": 0.08535221766679091, + "grad_norm": 24.43129539489746, + "learning_rate": 9.996634624056775e-06, + "loss": 11.703922271728516, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3103558421134949, + "loss_total_v6": 0.4666058421134949, + "step": 229 + }, + { + "epoch": 0.08572493477450616, + "grad_norm": 31.334762573242188, + "learning_rate": 9.996398555874958e-06, + "loss": 11.326579093933105, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4980235695838928, + "loss_total_v6": 0.6601329445838928, + "step": 230 + }, + { + "epoch": 0.08609765188222139, + "grad_norm": 24.67103385925293, + "learning_rate": 9.9961544892894e-06, + "loss": 11.475371360778809, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.28909996151924133, + "loss_total_v6": 0.45706871151924133, + "step": 231 + }, + { + "epoch": 0.08647036898993664, + "grad_norm": 24.023910522460938, + "learning_rate": 9.995902424690813e-06, + "loss": 11.30167007446289, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5473243594169617, + "loss_total_v6": 0.7074806094169617, + "step": 232 + }, + { + "epoch": 0.08684308609765189, + "grad_norm": 24.261066436767578, + "learning_rate": 9.995642362482708e-06, + "loss": 11.064987182617188, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5882874131202698, + "loss_total_v6": 0.7562561631202698, + "step": 233 + }, + { + "epoch": 0.08721580320536712, + "grad_norm": 21.837242126464844, + "learning_rate": 9.9953743030814e-06, + "loss": 11.340072631835938, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5674760937690735, + "loss_total_v6": 0.7295854687690735, + "step": 234 + }, + { + "epoch": 0.08758852031308237, + "grad_norm": 22.989248275756836, + "learning_rate": 9.995098246916013e-06, + "loss": 11.343286514282227, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.2643027901649475, + "loss_total_v6": 0.4322715401649475, + "step": 235 + }, + { + "epoch": 0.08796123742079762, + "grad_norm": 23.350502014160156, + "learning_rate": 9.994814194428464e-06, + "loss": 11.530061721801758, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5166326761245728, + "loss_total_v6": 0.6689764261245728, + "step": 236 + }, + { + "epoch": 0.08833395452851286, + "grad_norm": 25.125337600708008, + "learning_rate": 9.994522146073474e-06, + "loss": 11.296618461608887, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3928042948246002, + "loss_total_v6": 0.5549136400222778, + "step": 237 + }, + { + "epoch": 0.0887066716362281, + "grad_norm": 24.11391258239746, + "learning_rate": 9.994222102318563e-06, + "loss": 11.511462211608887, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5526203513145447, + "loss_total_v6": 0.7049641013145447, + "step": 238 + }, + { + "epoch": 0.08907938874394335, + "grad_norm": 26.43235969543457, + "learning_rate": 9.993914063644053e-06, + "loss": 11.218240737915039, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5012437105178833, + "loss_total_v6": 0.6633530855178833, + "step": 239 + }, + { + "epoch": 0.0894521058516586, + "grad_norm": 24.2873592376709, + "learning_rate": 9.993598030543061e-06, + "loss": 11.416260719299316, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3108470141887665, + "loss_total_v6": 0.4710032641887665, + "step": 240 + }, + { + "epoch": 0.08982482295937383, + "grad_norm": 23.45045280456543, + "learning_rate": 9.993274003521504e-06, + "loss": 11.359440803527832, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7942347526550293, + "loss_total_v6": 0.7942347526550293, + "step": 241 + }, + { + "epoch": 0.09019754006708908, + "grad_norm": 22.725770950317383, + "learning_rate": 9.992941983098097e-06, + "loss": 11.373712539672852, + "loss_alignment": 0.3515625, + "loss_alignment_w": 0.17578125, + "loss_sft": 0.6355465054512024, + "loss_total_v6": 0.8113277554512024, + "step": 242 + }, + { + "epoch": 0.09057025717480433, + "grad_norm": 22.71143913269043, + "learning_rate": 9.992601969804348e-06, + "loss": 11.407299995422363, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.9044947028160095, + "loss_total_v6": 1.0685572624206543, + "step": 243 + }, + { + "epoch": 0.09094297428251957, + "grad_norm": 22.695114135742188, + "learning_rate": 9.99225396418456e-06, + "loss": 11.218238830566406, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.49947986006736755, + "loss_total_v6": 0.6537767648696899, + "step": 244 + }, + { + "epoch": 0.09131569139023481, + "grad_norm": 23.517061233520508, + "learning_rate": 9.991897966795835e-06, + "loss": 10.995895385742188, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.717555820941925, + "loss_total_v6": 0.883571445941925, + "step": 245 + }, + { + "epoch": 0.09168840849795006, + "grad_norm": 24.045515060424805, + "learning_rate": 9.991533978208065e-06, + "loss": 11.390800476074219, + "loss_alignment": 0.265625, + "loss_alignment_w": 0.1328125, + "loss_sft": 0.3601967394351959, + "loss_total_v6": 0.4930092394351959, + "step": 246 + }, + { + "epoch": 0.0920611256056653, + "grad_norm": 26.570606231689453, + "learning_rate": 9.991161999003936e-06, + "loss": 11.59162712097168, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6572479009628296, + "loss_total_v6": 0.6572479009628296, + "step": 247 + }, + { + "epoch": 0.09243384271338054, + "grad_norm": 23.511167526245117, + "learning_rate": 9.990782029778922e-06, + "loss": 11.751054763793945, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5472851395606995, + "loss_total_v6": 0.6937695145606995, + "step": 248 + }, + { + "epoch": 0.09280655982109579, + "grad_norm": 27.236177444458008, + "learning_rate": 9.990394071141295e-06, + "loss": 11.135986328125, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.653698742389679, + "loss_total_v6": 0.807995617389679, + "step": 249 + }, + { + "epoch": 0.09317927692881103, + "grad_norm": 22.86897850036621, + "learning_rate": 9.989998123712108e-06, + "loss": 10.93806266784668, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.44356226921081543, + "loss_total_v6": 0.6134841442108154, + "step": 250 + }, + { + "epoch": 0.09355199403652628, + "grad_norm": 23.749149322509766, + "learning_rate": 9.98959418812521e-06, + "loss": 11.49368667602539, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.7334921360015869, + "loss_total_v6": 0.9034140110015869, + "step": 251 + }, + { + "epoch": 0.09392471114424152, + "grad_norm": 22.83121681213379, + "learning_rate": 9.989182265027232e-06, + "loss": 11.076770782470703, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.962500810623169, + "loss_total_v6": 1.126563310623169, + "step": 252 + }, + { + "epoch": 0.09429742825195676, + "grad_norm": 23.506052017211914, + "learning_rate": 9.988762355077597e-06, + "loss": 11.451047897338867, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6523056626319885, + "loss_total_v6": 0.8085556626319885, + "step": 253 + }, + { + "epoch": 0.09467014535967201, + "grad_norm": 22.706012725830078, + "learning_rate": 9.98833445894851e-06, + "loss": 11.871060371398926, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.6782039403915405, + "loss_total_v6": 0.8227351903915405, + "step": 254 + }, + { + "epoch": 0.09504286246738726, + "grad_norm": 22.473918914794922, + "learning_rate": 9.98789857732496e-06, + "loss": 11.268640518188477, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.44429269433021545, + "loss_total_v6": 0.6064020395278931, + "step": 255 + }, + { + "epoch": 0.0954155795751025, + "grad_norm": 22.47380256652832, + "learning_rate": 9.987454710904725e-06, + "loss": 11.41482925415039, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3266724944114685, + "loss_total_v6": 0.4770631194114685, + "step": 256 + }, + { + "epoch": 0.09578829668281774, + "grad_norm": 22.895347595214844, + "learning_rate": 9.987002860398358e-06, + "loss": 11.2366943359375, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.1952398121356964, + "loss_total_v6": 0.3436773121356964, + "step": 257 + }, + { + "epoch": 0.09616101379053299, + "grad_norm": 24.103939056396484, + "learning_rate": 9.986543026529197e-06, + "loss": 11.88573932647705, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5264126062393188, + "loss_total_v6": 0.6807094812393188, + "step": 258 + }, + { + "epoch": 0.09653373089824822, + "grad_norm": 22.33839225769043, + "learning_rate": 9.986075210033362e-06, + "loss": 11.127331733703613, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.39572176337242126, + "loss_total_v6": 0.5461124181747437, + "step": 259 + }, + { + "epoch": 0.09690644800596347, + "grad_norm": 22.876659393310547, + "learning_rate": 9.985599411659746e-06, + "loss": 11.457901954650879, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.525189220905304, + "loss_total_v6": 0.679486095905304, + "step": 260 + }, + { + "epoch": 0.09727916511367872, + "grad_norm": 22.999208450317383, + "learning_rate": 9.985115632170024e-06, + "loss": 11.712095260620117, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.41171348094940186, + "loss_total_v6": 0.5679634809494019, + "step": 261 + }, + { + "epoch": 0.09765188222139397, + "grad_norm": 22.217605590820312, + "learning_rate": 9.984623872338648e-06, + "loss": 10.825801849365234, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3889547884464264, + "loss_total_v6": 0.551064133644104, + "step": 262 + }, + { + "epoch": 0.0980245993291092, + "grad_norm": 22.1431884765625, + "learning_rate": 9.984124132952845e-06, + "loss": 10.642801284790039, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6035763621330261, + "loss_total_v6": 0.7676388621330261, + "step": 263 + }, + { + "epoch": 0.09839731643682445, + "grad_norm": 22.51676368713379, + "learning_rate": 9.983616414812611e-06, + "loss": 11.410966873168945, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.7105580568313599, + "loss_total_v6": 0.8804799318313599, + "step": 264 + }, + { + "epoch": 0.0987700335445397, + "grad_norm": 24.7966251373291, + "learning_rate": 9.98310071873072e-06, + "loss": 10.864057540893555, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.9076775908470154, + "loss_total_v6": 1.0658807754516602, + "step": 265 + }, + { + "epoch": 0.09914275065225493, + "grad_norm": 21.978235244750977, + "learning_rate": 9.982577045532717e-06, + "loss": 11.246742248535156, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.665847897529602, + "loss_total_v6": 0.827957272529602, + "step": 266 + }, + { + "epoch": 0.09951546775997018, + "grad_norm": 22.86284637451172, + "learning_rate": 9.982045396056916e-06, + "loss": 11.207635879516602, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.8580790162086487, + "loss_total_v6": 0.8580790162086487, + "step": 267 + }, + { + "epoch": 0.09988818486768543, + "grad_norm": 22.49290657043457, + "learning_rate": 9.981505771154395e-06, + "loss": 11.271628379821777, + "loss_alignment": 0.3671875, + "loss_alignment_w": 0.18359375, + "loss_sft": 0.673467755317688, + "loss_total_v6": 0.857061505317688, + "step": 268 + }, + { + "epoch": 0.10026090197540068, + "grad_norm": 22.785818099975586, + "learning_rate": 9.980958171689008e-06, + "loss": 11.588794708251953, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6235964298248291, + "loss_total_v6": 0.7778933048248291, + "step": 269 + }, + { + "epoch": 0.10063361908311591, + "grad_norm": 23.01845359802246, + "learning_rate": 9.980402598537371e-06, + "loss": 11.196735382080078, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.8757519125938416, + "loss_total_v6": 1.0320019721984863, + "step": 270 + }, + { + "epoch": 0.10100633619083116, + "grad_norm": 24.10038185119629, + "learning_rate": 9.979839052588863e-06, + "loss": 10.99837875366211, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7121546864509583, + "loss_total_v6": 0.8723109364509583, + "step": 271 + }, + { + "epoch": 0.1013790532985464, + "grad_norm": 23.360370635986328, + "learning_rate": 9.979267534745624e-06, + "loss": 11.468342781066895, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5713663101196289, + "loss_total_v6": 0.7198038101196289, + "step": 272 + }, + { + "epoch": 0.10175177040626164, + "grad_norm": 26.27503776550293, + "learning_rate": 9.978688045922566e-06, + "loss": 10.93787670135498, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5602408647537231, + "loss_total_v6": 0.7047721147537231, + "step": 273 + }, + { + "epoch": 0.10212448751397689, + "grad_norm": 22.856613159179688, + "learning_rate": 9.97810058704735e-06, + "loss": 11.49018669128418, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5420021414756775, + "loss_total_v6": 0.5420021414756775, + "step": 274 + }, + { + "epoch": 0.10249720462169214, + "grad_norm": 22.501758575439453, + "learning_rate": 9.977505159060398e-06, + "loss": 11.153072357177734, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5815086364746094, + "loss_total_v6": 0.7494773864746094, + "step": 275 + }, + { + "epoch": 0.10286992172940738, + "grad_norm": 21.44687843322754, + "learning_rate": 9.976901762914898e-06, + "loss": 11.322721481323242, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7267612218856812, + "loss_total_v6": 0.8791049718856812, + "step": 276 + }, + { + "epoch": 0.10324263883712262, + "grad_norm": 25.805028915405273, + "learning_rate": 9.97629039957678e-06, + "loss": 11.101606369018555, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4116356372833252, + "loss_total_v6": 0.5678856372833252, + "step": 277 + }, + { + "epoch": 0.10361535594483787, + "grad_norm": 23.217514038085938, + "learning_rate": 9.975671070024741e-06, + "loss": 11.013144493103027, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5051152110099792, + "loss_total_v6": 0.6672245860099792, + "step": 278 + }, + { + "epoch": 0.10398807305255311, + "grad_norm": 22.843753814697266, + "learning_rate": 9.97504377525022e-06, + "loss": 11.136174201965332, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7365843057632446, + "loss_total_v6": 0.8928343057632446, + "step": 279 + }, + { + "epoch": 0.10436079016026836, + "grad_norm": 25.54730796813965, + "learning_rate": 9.974408516257414e-06, + "loss": 10.778390884399414, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5316784977912903, + "loss_total_v6": 0.6898816227912903, + "step": 280 + }, + { + "epoch": 0.1047335072679836, + "grad_norm": 22.14451789855957, + "learning_rate": 9.973765294063265e-06, + "loss": 11.07178783416748, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3457458019256592, + "loss_total_v6": 0.5059020519256592, + "step": 281 + }, + { + "epoch": 0.10510622437569885, + "grad_norm": 21.597055435180664, + "learning_rate": 9.973114109697468e-06, + "loss": 10.77035140991211, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5606133341789246, + "loss_total_v6": 0.7246758341789246, + "step": 282 + }, + { + "epoch": 0.1054789414834141, + "grad_norm": 22.893115997314453, + "learning_rate": 9.972454964202458e-06, + "loss": 11.332418441772461, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6396092176437378, + "loss_total_v6": 0.6396092176437378, + "step": 283 + }, + { + "epoch": 0.10585165859112933, + "grad_norm": 21.540475845336914, + "learning_rate": 9.971787858633419e-06, + "loss": 10.908392906188965, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.22343453764915466, + "loss_total_v6": 0.36991891264915466, + "step": 284 + }, + { + "epoch": 0.10622437569884458, + "grad_norm": 21.877187728881836, + "learning_rate": 9.971112794058278e-06, + "loss": 10.99422550201416, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.32978102564811707, + "loss_total_v6": 0.48798415064811707, + "step": 285 + }, + { + "epoch": 0.10659709280655982, + "grad_norm": 21.980087280273438, + "learning_rate": 9.970429771557697e-06, + "loss": 10.979190826416016, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.35676687955856323, + "loss_total_v6": 0.35676687955856323, + "step": 286 + }, + { + "epoch": 0.10696980991427507, + "grad_norm": 32.07426071166992, + "learning_rate": 9.969738792225083e-06, + "loss": 11.00467300415039, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.1057976484298706, + "loss_total_v6": 0.2659538984298706, + "step": 287 + }, + { + "epoch": 0.1073425270219903, + "grad_norm": 22.82600212097168, + "learning_rate": 9.96903985716658e-06, + "loss": 11.109477996826172, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5430742502212524, + "loss_total_v6": 0.7032305002212524, + "step": 288 + }, + { + "epoch": 0.10771524412970555, + "grad_norm": 23.13313102722168, + "learning_rate": 9.968332967501068e-06, + "loss": 11.63078498840332, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.5833154320716858, + "loss_total_v6": 0.7551904320716858, + "step": 289 + }, + { + "epoch": 0.1080879612374208, + "grad_norm": 23.23664665222168, + "learning_rate": 9.967618124360156e-06, + "loss": 11.211320877075195, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.522067666053772, + "loss_total_v6": 0.686130166053772, + "step": 290 + }, + { + "epoch": 0.10846067834513604, + "grad_norm": 24.193843841552734, + "learning_rate": 9.966895328888195e-06, + "loss": 10.964338302612305, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5227862000465393, + "loss_total_v6": 0.6751299500465393, + "step": 291 + }, + { + "epoch": 0.10883339545285128, + "grad_norm": 22.814495086669922, + "learning_rate": 9.966164582242255e-06, + "loss": 11.163288116455078, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5036325454711914, + "loss_total_v6": 0.6540231704711914, + "step": 292 + }, + { + "epoch": 0.10920611256056653, + "grad_norm": 23.202857971191406, + "learning_rate": 9.96542588559214e-06, + "loss": 10.965642929077148, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.5476856827735901, + "loss_total_v6": 0.6863575577735901, + "step": 293 + }, + { + "epoch": 0.10957882966828178, + "grad_norm": 29.251680374145508, + "learning_rate": 9.964679240120385e-06, + "loss": 11.231146812438965, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.9062564373016357, + "loss_total_v6": 1.0625064373016357, + "step": 294 + }, + { + "epoch": 0.10995154677599701, + "grad_norm": 22.181774139404297, + "learning_rate": 9.963924647022241e-06, + "loss": 11.453174591064453, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.8577368855476379, + "loss_total_v6": 1.0315649509429932, + "step": 295 + }, + { + "epoch": 0.11032426388371226, + "grad_norm": 22.360090255737305, + "learning_rate": 9.963162107505688e-06, + "loss": 11.12214469909668, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4409306049346924, + "loss_total_v6": 0.5952274799346924, + "step": 296 + }, + { + "epoch": 0.11069698099142751, + "grad_norm": 23.761966705322266, + "learning_rate": 9.962391622791427e-06, + "loss": 11.254911422729492, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.6419786214828491, + "loss_total_v6": 0.8197129964828491, + "step": 297 + }, + { + "epoch": 0.11106969809914274, + "grad_norm": 22.832876205444336, + "learning_rate": 9.961613194112872e-06, + "loss": 11.193018913269043, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.9354682564735413, + "loss_total_v6": 1.1014838218688965, + "step": 298 + }, + { + "epoch": 0.11144241520685799, + "grad_norm": 22.902782440185547, + "learning_rate": 9.96082682271616e-06, + "loss": 11.113229751586914, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4383259415626526, + "loss_total_v6": 0.5926228165626526, + "step": 299 + }, + { + "epoch": 0.11181513231457324, + "grad_norm": 23.683238983154297, + "learning_rate": 9.96003250986014e-06, + "loss": 11.199775695800781, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6088377237319946, + "loss_total_v6": 0.7611814737319946, + "step": 300 + }, + { + "epoch": 0.11181513231457324, + "eval_loss": 0.657538890838623, + "eval_loss_alignment": 0.2926098424145299, + "eval_loss_alignment_w": 0.14630492120726496, + "eval_loss_sft": 0.5112339633827409, + "eval_loss_total_v6": 0.6575388849800469, + "eval_runtime": 207.2326, + "eval_samples_per_second": 9.004, + "eval_steps_per_second": 1.129, + "step": 300 + }, + { + "epoch": 0.11218784942228849, + "grad_norm": 23.283475875854492, + "learning_rate": 9.959230256816378e-06, + "loss": 11.854829788208008, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.63017737865448, + "loss_total_v6": 0.78252112865448, + "step": 301 + }, + { + "epoch": 0.11256056653000372, + "grad_norm": 21.841049194335938, + "learning_rate": 9.958420064869143e-06, + "loss": 11.822651863098145, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5912052989006042, + "loss_total_v6": 0.7533146739006042, + "step": 302 + }, + { + "epoch": 0.11293328363771897, + "grad_norm": 22.115314483642578, + "learning_rate": 9.957601935315422e-06, + "loss": 11.41370964050293, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3646821677684784, + "loss_total_v6": 0.5248384475708008, + "step": 303 + }, + { + "epoch": 0.11330600074543422, + "grad_norm": 24.09991455078125, + "learning_rate": 9.956775869464901e-06, + "loss": 11.344932556152344, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7322479486465454, + "loss_total_v6": 0.8865448236465454, + "step": 304 + }, + { + "epoch": 0.11367871785314947, + "grad_norm": 24.00020980834961, + "learning_rate": 9.955941868639976e-06, + "loss": 10.966205596923828, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7225505113601685, + "loss_total_v6": 0.8788005113601685, + "step": 305 + }, + { + "epoch": 0.1140514349608647, + "grad_norm": 23.746110916137695, + "learning_rate": 9.955099934175746e-06, + "loss": 11.198699951171875, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4640411138534546, + "loss_total_v6": 0.6183379888534546, + "step": 306 + }, + { + "epoch": 0.11442415206857995, + "grad_norm": 23.51089096069336, + "learning_rate": 9.954250067420001e-06, + "loss": 11.297557830810547, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4694594740867615, + "loss_total_v6": 0.6257094740867615, + "step": 307 + }, + { + "epoch": 0.1147968691762952, + "grad_norm": 21.35750389099121, + "learning_rate": 9.953392269733244e-06, + "loss": 11.497364044189453, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5788524150848389, + "loss_total_v6": 0.7370555400848389, + "step": 308 + }, + { + "epoch": 0.11516958628401043, + "grad_norm": 22.820951461791992, + "learning_rate": 9.952526542488663e-06, + "loss": 11.172540664672852, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.383622944355011, + "loss_total_v6": 0.383622944355011, + "step": 309 + }, + { + "epoch": 0.11554230339172568, + "grad_norm": 55.59081268310547, + "learning_rate": 9.951652887072142e-06, + "loss": 11.188230514526367, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.7123116850852966, + "loss_total_v6": 0.8744210600852966, + "step": 310 + }, + { + "epoch": 0.11591502049944093, + "grad_norm": 22.54860496520996, + "learning_rate": 9.950771304882262e-06, + "loss": 11.053382873535156, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.4116823077201843, + "loss_total_v6": 0.5991823077201843, + "step": 311 + }, + { + "epoch": 0.11628773760715617, + "grad_norm": 21.572959899902344, + "learning_rate": 9.94988179733029e-06, + "loss": 11.464082717895508, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5688577890396118, + "loss_total_v6": 0.7212015390396118, + "step": 312 + }, + { + "epoch": 0.11666045471487141, + "grad_norm": 21.37523651123047, + "learning_rate": 9.948984365840174e-06, + "loss": 10.879380226135254, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6387176513671875, + "loss_total_v6": 0.7988739013671875, + "step": 313 + }, + { + "epoch": 0.11703317182258666, + "grad_norm": 21.34028434753418, + "learning_rate": 9.948079011848557e-06, + "loss": 10.670356750488281, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6201522946357727, + "loss_total_v6": 0.7803085446357727, + "step": 314 + }, + { + "epoch": 0.1174058889303019, + "grad_norm": 22.28590965270996, + "learning_rate": 9.94716573680476e-06, + "loss": 10.870960235595703, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.549823522567749, + "loss_total_v6": 0.717792272567749, + "step": 315 + }, + { + "epoch": 0.11777860603801714, + "grad_norm": 23.065288543701172, + "learning_rate": 9.946244542170784e-06, + "loss": 10.971685409545898, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.48123225569725037, + "loss_total_v6": 0.6394354104995728, + "step": 316 + }, + { + "epoch": 0.11815132314573239, + "grad_norm": 23.669719696044922, + "learning_rate": 9.945315429421307e-06, + "loss": 11.46750259399414, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6476104259490967, + "loss_total_v6": 0.8038604259490967, + "step": 317 + }, + { + "epoch": 0.11852404025344764, + "grad_norm": 22.01180648803711, + "learning_rate": 9.944378400043685e-06, + "loss": 11.091371536254883, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6320104598999023, + "loss_total_v6": 0.7882604598999023, + "step": 318 + }, + { + "epoch": 0.11889675736116288, + "grad_norm": 111.17440795898438, + "learning_rate": 9.94343345553795e-06, + "loss": 10.604940414428711, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5784343481063843, + "loss_total_v6": 0.7385905981063843, + "step": 319 + }, + { + "epoch": 0.11926947446887812, + "grad_norm": 155.65599060058594, + "learning_rate": 9.942480597416797e-06, + "loss": 11.345887184143066, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.31963640451431274, + "loss_total_v6": 0.31963640451431274, + "step": 320 + }, + { + "epoch": 0.11964219157659337, + "grad_norm": 26.65423583984375, + "learning_rate": 9.941519827205595e-06, + "loss": 10.768564224243164, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4680943191051483, + "loss_total_v6": 0.6262974739074707, + "step": 321 + }, + { + "epoch": 0.12001490868430861, + "grad_norm": 23.608732223510742, + "learning_rate": 9.940551146442375e-06, + "loss": 11.455142974853516, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.7063157558441162, + "loss_total_v6": 0.8918626308441162, + "step": 322 + }, + { + "epoch": 0.12038762579202385, + "grad_norm": 23.332101821899414, + "learning_rate": 9.939574556677837e-06, + "loss": 10.969060897827148, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.7082730531692505, + "loss_total_v6": 0.8469449281692505, + "step": 323 + }, + { + "epoch": 0.1207603428997391, + "grad_norm": 24.58036994934082, + "learning_rate": 9.938590059475338e-06, + "loss": 10.94526481628418, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5296091437339783, + "loss_total_v6": 0.6741403937339783, + "step": 324 + }, + { + "epoch": 0.12113306000745434, + "grad_norm": 26.331689834594727, + "learning_rate": 9.937597656410893e-06, + "loss": 11.72280216217041, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5032468438148499, + "loss_total_v6": 0.6634030938148499, + "step": 325 + }, + { + "epoch": 0.12150577711516959, + "grad_norm": 22.004520416259766, + "learning_rate": 9.936597349073176e-06, + "loss": 11.310832977294922, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.45880576968193054, + "loss_total_v6": 0.6365401744842529, + "step": 326 + }, + { + "epoch": 0.12187849422288483, + "grad_norm": 21.72319984436035, + "learning_rate": 9.935589139063513e-06, + "loss": 11.939213752746582, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.42702236771583557, + "loss_total_v6": 0.5832723379135132, + "step": 327 + }, + { + "epoch": 0.12225121133060007, + "grad_norm": 21.89547348022461, + "learning_rate": 9.93457302799588e-06, + "loss": 10.731821060180664, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5654523372650146, + "loss_total_v6": 0.7197492122650146, + "step": 328 + }, + { + "epoch": 0.12262392843831532, + "grad_norm": 25.504230499267578, + "learning_rate": 9.933549017496902e-06, + "loss": 10.948042869567871, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.34107932448387146, + "loss_total_v6": 0.5031887292861938, + "step": 329 + }, + { + "epoch": 0.12299664554603057, + "grad_norm": 22.34630012512207, + "learning_rate": 9.932517109205849e-06, + "loss": 11.239425659179688, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.8384466171264648, + "loss_total_v6": 0.9907903671264648, + "step": 330 + }, + { + "epoch": 0.1233693626537458, + "grad_norm": 24.18584632873535, + "learning_rate": 9.931477304774637e-06, + "loss": 11.121136665344238, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.48916536569595337, + "loss_total_v6": 0.6493216156959534, + "step": 331 + }, + { + "epoch": 0.12374207976146105, + "grad_norm": 22.946672439575195, + "learning_rate": 9.930429605867817e-06, + "loss": 10.685807228088379, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6209961771965027, + "loss_total_v6": 0.7831055521965027, + "step": 332 + }, + { + "epoch": 0.1241147968691763, + "grad_norm": 23.78714370727539, + "learning_rate": 9.929374014162581e-06, + "loss": 10.948630332946777, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5643031597137451, + "loss_total_v6": 0.7146937847137451, + "step": 333 + }, + { + "epoch": 0.12448751397689153, + "grad_norm": 23.038591384887695, + "learning_rate": 9.928310531348759e-06, + "loss": 11.39359188079834, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5222453474998474, + "loss_total_v6": 0.6824015974998474, + "step": 334 + }, + { + "epoch": 0.12486023108460678, + "grad_norm": 23.244449615478516, + "learning_rate": 9.927239159128807e-06, + "loss": 11.395822525024414, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.19934771955013275, + "loss_total_v6": 0.34973835945129395, + "step": 335 + }, + { + "epoch": 0.12523294819232203, + "grad_norm": 20.67098617553711, + "learning_rate": 9.926159899217818e-06, + "loss": 11.140880584716797, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.7783012390136719, + "loss_total_v6": 0.9423637390136719, + "step": 336 + }, + { + "epoch": 0.12560566530003728, + "grad_norm": 21.994626998901367, + "learning_rate": 9.925072753343501e-06, + "loss": 10.410880088806152, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.3982803523540497, + "loss_total_v6": 0.5506241321563721, + "step": 337 + }, + { + "epoch": 0.12597838240775253, + "grad_norm": 22.59394645690918, + "learning_rate": 9.923977723246203e-06, + "loss": 10.672718048095703, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.745572030544281, + "loss_total_v6": 0.909634530544281, + "step": 338 + }, + { + "epoch": 0.12635109951546775, + "grad_norm": 22.8772029876709, + "learning_rate": 9.922874810678881e-06, + "loss": 11.151812553405762, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.36927327513694763, + "loss_total_v6": 0.51380455493927, + "step": 339 + }, + { + "epoch": 0.126723816623183, + "grad_norm": 21.56969451904297, + "learning_rate": 9.921764017407117e-06, + "loss": 10.979615211486816, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.4258961081504822, + "loss_total_v6": 0.5899586081504822, + "step": 340 + }, + { + "epoch": 0.12709653373089824, + "grad_norm": 20.985755920410156, + "learning_rate": 9.920645345209104e-06, + "loss": 10.891694068908691, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.622139573097229, + "loss_total_v6": 0.778389573097229, + "step": 341 + }, + { + "epoch": 0.1274692508386135, + "grad_norm": 21.045766830444336, + "learning_rate": 9.919518795875652e-06, + "loss": 11.48928165435791, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5267106890678406, + "loss_total_v6": 0.6790544390678406, + "step": 342 + }, + { + "epoch": 0.12784196794632874, + "grad_norm": 22.472808837890625, + "learning_rate": 9.918384371210178e-06, + "loss": 11.132970809936523, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.7283799052238464, + "loss_total_v6": 0.8748642802238464, + "step": 343 + }, + { + "epoch": 0.128214685054044, + "grad_norm": 21.341148376464844, + "learning_rate": 9.917242073028708e-06, + "loss": 11.09800910949707, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.48836904764175415, + "loss_total_v6": 0.6504784226417542, + "step": 344 + }, + { + "epoch": 0.12858740216175923, + "grad_norm": 22.689844131469727, + "learning_rate": 9.916091903159873e-06, + "loss": 11.384960174560547, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4437897801399231, + "loss_total_v6": 0.5902741551399231, + "step": 345 + }, + { + "epoch": 0.12896011926947448, + "grad_norm": 26.828676223754883, + "learning_rate": 9.9149338634449e-06, + "loss": 10.865995407104492, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6452163457870483, + "loss_total_v6": 0.8014663457870483, + "step": 346 + }, + { + "epoch": 0.1293328363771897, + "grad_norm": 22.00214958190918, + "learning_rate": 9.913767955737623e-06, + "loss": 10.727317810058594, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.533516526222229, + "loss_total_v6": 0.689766526222229, + "step": 347 + }, + { + "epoch": 0.12970555348490495, + "grad_norm": 22.792957305908203, + "learning_rate": 9.912594181904463e-06, + "loss": 11.341272354125977, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5672198534011841, + "loss_total_v6": 0.7117511034011841, + "step": 348 + }, + { + "epoch": 0.1300782705926202, + "grad_norm": 23.057687759399414, + "learning_rate": 9.911412543824438e-06, + "loss": 11.269739151000977, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5984808206558228, + "loss_total_v6": 0.7547308206558228, + "step": 349 + }, + { + "epoch": 0.13045098770033545, + "grad_norm": 21.484508514404297, + "learning_rate": 9.910223043389156e-06, + "loss": 11.221260070800781, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.7030081152915955, + "loss_total_v6": 0.8729299902915955, + "step": 350 + }, + { + "epoch": 0.1308237048080507, + "grad_norm": 21.20382308959961, + "learning_rate": 9.909025682502806e-06, + "loss": 11.276355743408203, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.4862704575061798, + "loss_total_v6": 0.6600985527038574, + "step": 351 + }, + { + "epoch": 0.13119642191576594, + "grad_norm": 20.714155197143555, + "learning_rate": 9.907820463082165e-06, + "loss": 10.931199073791504, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.616358757019043, + "loss_total_v6": 0.762843132019043, + "step": 352 + }, + { + "epoch": 0.1315691390234812, + "grad_norm": 21.366191864013672, + "learning_rate": 9.906607387056591e-06, + "loss": 11.433656692504883, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.33749860525131226, + "loss_total_v6": 0.49570173025131226, + "step": 353 + }, + { + "epoch": 0.1319418561311964, + "grad_norm": 24.040395736694336, + "learning_rate": 9.905386456368013e-06, + "loss": 11.226346969604492, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.8490189909934998, + "loss_total_v6": 1.0111284255981445, + "step": 354 + }, + { + "epoch": 0.13231457323891166, + "grad_norm": 23.183155059814453, + "learning_rate": 9.904157672970943e-06, + "loss": 11.28631591796875, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.33045247197151184, + "loss_total_v6": 0.49646809697151184, + "step": 355 + }, + { + "epoch": 0.1326872903466269, + "grad_norm": 97.30240631103516, + "learning_rate": 9.902921038832456e-06, + "loss": 11.303163528442383, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6257040500640869, + "loss_total_v6": 0.7897665500640869, + "step": 356 + }, + { + "epoch": 0.13306000745434216, + "grad_norm": 21.531253814697266, + "learning_rate": 9.901676555932199e-06, + "loss": 10.997499465942383, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6743394732475281, + "loss_total_v6": 0.8403550982475281, + "step": 357 + }, + { + "epoch": 0.1334327245620574, + "grad_norm": 21.145896911621094, + "learning_rate": 9.90042422626238e-06, + "loss": 11.014497756958008, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5930299162864685, + "loss_total_v6": 0.7531861662864685, + "step": 358 + }, + { + "epoch": 0.13380544166977265, + "grad_norm": 20.598005294799805, + "learning_rate": 9.899164051827774e-06, + "loss": 10.779167175292969, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.24100328981876373, + "loss_total_v6": 0.40506577491760254, + "step": 359 + }, + { + "epoch": 0.1341781587774879, + "grad_norm": 20.105478286743164, + "learning_rate": 9.897896034645709e-06, + "loss": 10.762897491455078, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4592745304107666, + "loss_total_v6": 0.6155245304107666, + "step": 360 + }, + { + "epoch": 0.13455087588520312, + "grad_norm": 22.028701782226562, + "learning_rate": 9.89662017674607e-06, + "loss": 11.102819442749023, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6906780004501343, + "loss_total_v6": 0.8508342504501343, + "step": 361 + }, + { + "epoch": 0.13492359299291837, + "grad_norm": 20.926647186279297, + "learning_rate": 9.895336480171294e-06, + "loss": 11.775465965270996, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6297129988670349, + "loss_total_v6": 0.7918223738670349, + "step": 362 + }, + { + "epoch": 0.13529631010063362, + "grad_norm": 21.083274841308594, + "learning_rate": 9.894044946976365e-06, + "loss": 11.220085144042969, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.517815113067627, + "loss_total_v6": 0.670158863067627, + "step": 363 + }, + { + "epoch": 0.13566902720834886, + "grad_norm": 23.407142639160156, + "learning_rate": 9.892745579228811e-06, + "loss": 10.820121765136719, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.37073302268981934, + "loss_total_v6": 0.5230767726898193, + "step": 364 + }, + { + "epoch": 0.1360417443160641, + "grad_norm": 20.391435623168945, + "learning_rate": 9.89143837900871e-06, + "loss": 11.012206077575684, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5671567916870117, + "loss_total_v6": 0.7214536666870117, + "step": 365 + }, + { + "epoch": 0.13641446142377936, + "grad_norm": 22.659521102905273, + "learning_rate": 9.890123348408666e-06, + "loss": 11.275385856628418, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7724807858467102, + "loss_total_v6": 0.9287307858467102, + "step": 366 + }, + { + "epoch": 0.1367871785314946, + "grad_norm": 21.63907814025879, + "learning_rate": 9.888800489533828e-06, + "loss": 10.898557662963867, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.43447768688201904, + "loss_total_v6": 0.590727686882019, + "step": 367 + }, + { + "epoch": 0.13715989563920983, + "grad_norm": 20.052181243896484, + "learning_rate": 9.887469804501873e-06, + "loss": 10.63371467590332, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.5823771357536316, + "loss_total_v6": 0.7620646357536316, + "step": 368 + }, + { + "epoch": 0.13753261274692508, + "grad_norm": 20.790945053100586, + "learning_rate": 9.886131295443003e-06, + "loss": 11.285510063171387, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6342601180076599, + "loss_total_v6": 0.7826976180076599, + "step": 369 + }, + { + "epoch": 0.13790532985464032, + "grad_norm": 24.818498611450195, + "learning_rate": 9.884784964499954e-06, + "loss": 11.286174774169922, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5490412712097168, + "loss_total_v6": 0.7013850212097168, + "step": 370 + }, + { + "epoch": 0.13827804696235557, + "grad_norm": 21.84427261352539, + "learning_rate": 9.883430813827975e-06, + "loss": 11.312088012695312, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5342064499855042, + "loss_total_v6": 0.6826439499855042, + "step": 371 + }, + { + "epoch": 0.13865076407007082, + "grad_norm": 22.4840145111084, + "learning_rate": 9.882068845594835e-06, + "loss": 10.98829460144043, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.7595707774162292, + "loss_total_v6": 0.9275395274162292, + "step": 372 + }, + { + "epoch": 0.13902348117778607, + "grad_norm": 21.326251983642578, + "learning_rate": 9.880699061980822e-06, + "loss": 11.046815872192383, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.8485291004180908, + "loss_total_v6": 1.0086853504180908, + "step": 373 + }, + { + "epoch": 0.13939619828550132, + "grad_norm": 23.122568130493164, + "learning_rate": 9.879321465178729e-06, + "loss": 10.617700576782227, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.4088861346244812, + "loss_total_v6": 0.5729486346244812, + "step": 374 + }, + { + "epoch": 0.13976891539321654, + "grad_norm": 20.39756202697754, + "learning_rate": 9.877936057393862e-06, + "loss": 11.442386627197266, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.601041316986084, + "loss_total_v6": 0.745572566986084, + "step": 375 + }, + { + "epoch": 0.14014163250093178, + "grad_norm": 21.167062759399414, + "learning_rate": 9.876542840844027e-06, + "loss": 11.03483772277832, + "loss_alignment": 0.3828125, + "loss_alignment_w": 0.19140625, + "loss_sft": 0.6692704558372498, + "loss_total_v6": 0.8606767058372498, + "step": 376 + }, + { + "epoch": 0.14051434960864703, + "grad_norm": 22.020517349243164, + "learning_rate": 9.875141817759533e-06, + "loss": 11.074594497680664, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.747592568397522, + "loss_total_v6": 0.896030068397522, + "step": 377 + }, + { + "epoch": 0.14088706671636228, + "grad_norm": 21.83411407470703, + "learning_rate": 9.873732990383185e-06, + "loss": 11.629728317260742, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5831121206283569, + "loss_total_v6": 0.7393621206283569, + "step": 378 + }, + { + "epoch": 0.14125978382407753, + "grad_norm": 21.137086868286133, + "learning_rate": 9.87231636097028e-06, + "loss": 10.973651885986328, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4740408658981323, + "loss_total_v6": 0.6302908658981323, + "step": 379 + }, + { + "epoch": 0.14163250093179278, + "grad_norm": 21.02739143371582, + "learning_rate": 9.870891931788608e-06, + "loss": 10.946105003356934, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.6696449518203735, + "loss_total_v6": 0.8141762018203735, + "step": 380 + }, + { + "epoch": 0.14200521803950802, + "grad_norm": 20.90304183959961, + "learning_rate": 9.869459705118444e-06, + "loss": 10.945596694946289, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.43893179297447205, + "loss_total_v6": 0.5912755727767944, + "step": 381 + }, + { + "epoch": 0.14237793514722324, + "grad_norm": 20.821430206298828, + "learning_rate": 9.868019683252543e-06, + "loss": 10.801993370056152, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.46102815866470337, + "loss_total_v6": 0.6153250336647034, + "step": 382 + }, + { + "epoch": 0.1427506522549385, + "grad_norm": 20.395126342773438, + "learning_rate": 9.866571868496142e-06, + "loss": 10.61677360534668, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.1889742910861969, + "loss_total_v6": 0.3432711660861969, + "step": 383 + }, + { + "epoch": 0.14312336936265374, + "grad_norm": 21.06722068786621, + "learning_rate": 9.86511626316695e-06, + "loss": 10.732183456420898, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7181051969528198, + "loss_total_v6": 0.8841208219528198, + "step": 384 + }, + { + "epoch": 0.143496086470369, + "grad_norm": 23.307188034057617, + "learning_rate": 9.863652869595151e-06, + "loss": 11.412867546081543, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.485163152217865, + "loss_total_v6": 0.635553777217865, + "step": 385 + }, + { + "epoch": 0.14386880357808424, + "grad_norm": 21.83273696899414, + "learning_rate": 9.862181690123395e-06, + "loss": 11.017049789428711, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5284392833709717, + "loss_total_v6": 0.6905486583709717, + "step": 386 + }, + { + "epoch": 0.14424152068579948, + "grad_norm": 22.637781143188477, + "learning_rate": 9.860702727106794e-06, + "loss": 11.507485389709473, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4755098819732666, + "loss_total_v6": 0.6298067569732666, + "step": 387 + }, + { + "epoch": 0.14461423779351473, + "grad_norm": 22.26451873779297, + "learning_rate": 9.859215982912927e-06, + "loss": 11.190335273742676, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4942798316478729, + "loss_total_v6": 0.4942798316478729, + "step": 388 + }, + { + "epoch": 0.14498695490122995, + "grad_norm": 22.1140079498291, + "learning_rate": 9.857721459921815e-06, + "loss": 10.612945556640625, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5919528603553772, + "loss_total_v6": 0.7462497353553772, + "step": 389 + }, + { + "epoch": 0.1453596720089452, + "grad_norm": 21.99896240234375, + "learning_rate": 9.856219160525951e-06, + "loss": 10.709577560424805, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3827468752861023, + "loss_total_v6": 0.5409500002861023, + "step": 390 + }, + { + "epoch": 0.14573238911666045, + "grad_norm": 22.054561614990234, + "learning_rate": 9.854709087130261e-06, + "loss": 11.127849578857422, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4257470667362213, + "loss_total_v6": 0.5839501619338989, + "step": 391 + }, + { + "epoch": 0.1461051062243757, + "grad_norm": 22.175193786621094, + "learning_rate": 9.853191242152123e-06, + "loss": 10.889039993286133, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5714983344078064, + "loss_total_v6": 0.7218889594078064, + "step": 392 + }, + { + "epoch": 0.14647782333209095, + "grad_norm": 21.333328247070312, + "learning_rate": 9.851665628021354e-06, + "loss": 11.145269393920898, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5025808811187744, + "loss_total_v6": 0.6529715061187744, + "step": 393 + }, + { + "epoch": 0.1468505404398062, + "grad_norm": 21.571374893188477, + "learning_rate": 9.85013224718021e-06, + "loss": 11.102363586425781, + "loss_alignment": 0.38671875, + "loss_alignment_w": 0.193359375, + "loss_sft": 0.55335932970047, + "loss_total_v6": 0.74671870470047, + "step": 394 + }, + { + "epoch": 0.14722325754752144, + "grad_norm": 21.354511260986328, + "learning_rate": 9.848591102083375e-06, + "loss": 10.681215286254883, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5445255041122437, + "loss_total_v6": 0.6988223791122437, + "step": 395 + }, + { + "epoch": 0.1475959746552367, + "grad_norm": 22.432947158813477, + "learning_rate": 9.847042195197971e-06, + "loss": 10.868369102478027, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4438914954662323, + "loss_total_v6": 0.4438914954662323, + "step": 396 + }, + { + "epoch": 0.1479686917629519, + "grad_norm": 21.131629943847656, + "learning_rate": 9.84548552900354e-06, + "loss": 10.934986114501953, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7745891809463501, + "loss_total_v6": 0.9347454309463501, + "step": 397 + }, + { + "epoch": 0.14834140887066716, + "grad_norm": 21.478342056274414, + "learning_rate": 9.843921105992043e-06, + "loss": 11.265302658081055, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4171673059463501, + "loss_total_v6": 0.5656048059463501, + "step": 398 + }, + { + "epoch": 0.1487141259783824, + "grad_norm": 19.867666244506836, + "learning_rate": 9.842348928667862e-06, + "loss": 11.214735984802246, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.2738250195980072, + "loss_total_v6": 0.4359343945980072, + "step": 399 + }, + { + "epoch": 0.14908684308609765, + "grad_norm": 20.653627395629883, + "learning_rate": 9.840768999547794e-06, + "loss": 10.91183090209961, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.8111086487770081, + "loss_total_v6": 0.9771242737770081, + "step": 400 + }, + { + "epoch": 0.14908684308609765, + "eval_loss": 0.6472082138061523, + "eval_loss_alignment": 0.29037919003739315, + "eval_loss_alignment_w": 0.14518959501869658, + "eval_loss_sft": 0.5020185596365322, + "eval_loss_total_v6": 0.6472081545358285, + "eval_runtime": 206.3912, + "eval_samples_per_second": 9.041, + "eval_steps_per_second": 1.134, + "step": 400 + }, + { + "epoch": 0.1494595601938129, + "grad_norm": 21.18906593322754, + "learning_rate": 9.83918132116104e-06, + "loss": 10.88831901550293, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.7091809511184692, + "loss_total_v6": 0.8810559511184692, + "step": 401 + }, + { + "epoch": 0.14983227730152815, + "grad_norm": 21.242555618286133, + "learning_rate": 9.83758589604921e-06, + "loss": 10.855352401733398, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.08628037571907043, + "loss_total_v6": 0.22885850071907043, + "step": 402 + }, + { + "epoch": 0.1502049944092434, + "grad_norm": 22.19816017150879, + "learning_rate": 9.835982726766318e-06, + "loss": 11.236944198608398, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 1.0136196613311768, + "loss_total_v6": 1.1737759113311768, + "step": 403 + }, + { + "epoch": 0.15057771151695862, + "grad_norm": 38.725830078125, + "learning_rate": 9.834371815878766e-06, + "loss": 10.915067672729492, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.11466269940137863, + "loss_total_v6": 0.27091270685195923, + "step": 404 + }, + { + "epoch": 0.15095042862467387, + "grad_norm": 21.09138298034668, + "learning_rate": 9.832753165965359e-06, + "loss": 10.58181381225586, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.5548688173294067, + "loss_total_v6": 0.7208844423294067, + "step": 405 + }, + { + "epoch": 0.1513231457323891, + "grad_norm": 21.977218627929688, + "learning_rate": 9.831126779617284e-06, + "loss": 11.20839786529541, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5167040228843689, + "loss_total_v6": 0.6729540228843689, + "step": 406 + }, + { + "epoch": 0.15169586284010436, + "grad_norm": 21.827041625976562, + "learning_rate": 9.829492659438115e-06, + "loss": 11.3363618850708, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.32970350980758667, + "loss_total_v6": 0.47814100980758667, + "step": 407 + }, + { + "epoch": 0.1520685799478196, + "grad_norm": 21.344161987304688, + "learning_rate": 9.82785080804381e-06, + "loss": 10.843435287475586, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6582527160644531, + "loss_total_v6": 0.8262214660644531, + "step": 408 + }, + { + "epoch": 0.15244129705553486, + "grad_norm": 21.640605926513672, + "learning_rate": 9.826201228062696e-06, + "loss": 11.182035446166992, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6450145840644836, + "loss_total_v6": 0.7954052090644836, + "step": 409 + }, + { + "epoch": 0.1528140141632501, + "grad_norm": 21.91765594482422, + "learning_rate": 9.824543922135479e-06, + "loss": 11.07080078125, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6285154223442078, + "loss_total_v6": 0.7828122973442078, + "step": 410 + }, + { + "epoch": 0.15318673127096533, + "grad_norm": 22.098031997680664, + "learning_rate": 9.822878892915231e-06, + "loss": 10.873644828796387, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.8472007513046265, + "loss_total_v6": 1.0034507513046265, + "step": 411 + }, + { + "epoch": 0.15355944837868057, + "grad_norm": 21.26818084716797, + "learning_rate": 9.821206143067385e-06, + "loss": 11.005070686340332, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.43974071741104126, + "loss_total_v6": 0.6057563424110413, + "step": 412 + }, + { + "epoch": 0.15393216548639582, + "grad_norm": 27.823421478271484, + "learning_rate": 9.819525675269736e-06, + "loss": 11.314704895019531, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6018725037574768, + "loss_total_v6": 0.7561693787574768, + "step": 413 + }, + { + "epoch": 0.15430488259411107, + "grad_norm": 20.152088165283203, + "learning_rate": 9.817837492212434e-06, + "loss": 10.913497924804688, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5228976607322693, + "loss_total_v6": 0.6752414107322693, + "step": 414 + }, + { + "epoch": 0.15467759970182632, + "grad_norm": 20.75598907470703, + "learning_rate": 9.816141596597981e-06, + "loss": 11.06777572631836, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7105897665023804, + "loss_total_v6": 0.8707460165023804, + "step": 415 + }, + { + "epoch": 0.15505031680954157, + "grad_norm": 19.78862953186035, + "learning_rate": 9.81443799114122e-06, + "loss": 10.856348991394043, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6274642944335938, + "loss_total_v6": 0.7759017944335938, + "step": 416 + }, + { + "epoch": 0.15542303391725681, + "grad_norm": 20.671297073364258, + "learning_rate": 9.812726678569347e-06, + "loss": 10.771970748901367, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4419811964035034, + "loss_total_v6": 0.5845593214035034, + "step": 417 + }, + { + "epoch": 0.15579575102497203, + "grad_norm": 19.83351707458496, + "learning_rate": 9.811007661621881e-06, + "loss": 10.715304374694824, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.42131099104881287, + "loss_total_v6": 0.5756078958511353, + "step": 418 + }, + { + "epoch": 0.15616846813268728, + "grad_norm": 20.961942672729492, + "learning_rate": 9.809280943050688e-06, + "loss": 11.09864330291748, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.636458158493042, + "loss_total_v6": 0.808333158493042, + "step": 419 + }, + { + "epoch": 0.15654118524040253, + "grad_norm": 21.26742935180664, + "learning_rate": 9.807546525619956e-06, + "loss": 10.746097564697266, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5347318053245544, + "loss_total_v6": 0.6948880553245544, + "step": 420 + }, + { + "epoch": 0.15691390234811778, + "grad_norm": 20.687761306762695, + "learning_rate": 9.805804412106197e-06, + "loss": 11.046759605407715, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5871495008468628, + "loss_total_v6": 0.7375401258468628, + "step": 421 + }, + { + "epoch": 0.15728661945583303, + "grad_norm": 20.9792423248291, + "learning_rate": 9.80405460529825e-06, + "loss": 10.706931114196777, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5710369348526001, + "loss_total_v6": 0.7272869348526001, + "step": 422 + }, + { + "epoch": 0.15765933656354827, + "grad_norm": 20.622888565063477, + "learning_rate": 9.802297107997261e-06, + "loss": 10.84571361541748, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5281345844268799, + "loss_total_v6": 0.6863377094268799, + "step": 423 + }, + { + "epoch": 0.15803205367126352, + "grad_norm": 22.602155685424805, + "learning_rate": 9.80053192301669e-06, + "loss": 11.251531600952148, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5911733508110046, + "loss_total_v6": 0.7435171008110046, + "step": 424 + }, + { + "epoch": 0.15840477077897874, + "grad_norm": 20.03777313232422, + "learning_rate": 9.79875905318231e-06, + "loss": 10.645566940307617, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.599470853805542, + "loss_total_v6": 0.749861478805542, + "step": 425 + }, + { + "epoch": 0.158777487886694, + "grad_norm": 19.849201202392578, + "learning_rate": 9.796978501332185e-06, + "loss": 10.511791229248047, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4248848855495453, + "loss_total_v6": 0.5772286653518677, + "step": 426 + }, + { + "epoch": 0.15915020499440924, + "grad_norm": 24.704301834106445, + "learning_rate": 9.795190270316687e-06, + "loss": 11.142498970031738, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.410718709230423, + "loss_total_v6": 0.5591561794281006, + "step": 427 + }, + { + "epoch": 0.1595229221021245, + "grad_norm": 21.02276611328125, + "learning_rate": 9.793394362998475e-06, + "loss": 10.484672546386719, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.33658698201179504, + "loss_total_v6": 0.5026025772094727, + "step": 428 + }, + { + "epoch": 0.15989563920983973, + "grad_norm": 21.2802734375, + "learning_rate": 9.791590782252498e-06, + "loss": 11.021810531616211, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4304656982421875, + "loss_total_v6": 0.5886688232421875, + "step": 429 + }, + { + "epoch": 0.16026835631755498, + "grad_norm": 22.532373428344727, + "learning_rate": 9.789779530965991e-06, + "loss": 11.13211727142334, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.37005284428596497, + "loss_total_v6": 0.5243496894836426, + "step": 430 + }, + { + "epoch": 0.16064107342527023, + "grad_norm": 21.341676712036133, + "learning_rate": 9.787960612038463e-06, + "loss": 11.351921081542969, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6561948657035828, + "loss_total_v6": 0.8124448657035828, + "step": 431 + }, + { + "epoch": 0.16101379053298545, + "grad_norm": 20.79878044128418, + "learning_rate": 9.786134028381701e-06, + "loss": 10.838508605957031, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6850196123123169, + "loss_total_v6": 0.8275977373123169, + "step": 432 + }, + { + "epoch": 0.1613865076407007, + "grad_norm": 20.835247039794922, + "learning_rate": 9.784299782919766e-06, + "loss": 10.845494270324707, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5073224306106567, + "loss_total_v6": 0.6577130556106567, + "step": 433 + }, + { + "epoch": 0.16175922474841595, + "grad_norm": 20.396482467651367, + "learning_rate": 9.782457878588977e-06, + "loss": 10.978655815124512, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.20110802352428436, + "loss_total_v6": 0.35149866342544556, + "step": 434 + }, + { + "epoch": 0.1621319418561312, + "grad_norm": 20.84609031677246, + "learning_rate": 9.780608318337917e-06, + "loss": 10.569086074829102, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5274694561958313, + "loss_total_v6": 0.6778600811958313, + "step": 435 + }, + { + "epoch": 0.16250465896384644, + "grad_norm": 21.144607543945312, + "learning_rate": 9.778751105127425e-06, + "loss": 10.704736709594727, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.2541338801383972, + "loss_total_v6": 0.4181963801383972, + "step": 436 + }, + { + "epoch": 0.1628773760715617, + "grad_norm": 21.23587989807129, + "learning_rate": 9.77688624193059e-06, + "loss": 10.862020492553711, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5264399647712708, + "loss_total_v6": 0.6846430897712708, + "step": 437 + }, + { + "epoch": 0.16325009317927694, + "grad_norm": 20.982524871826172, + "learning_rate": 9.77501373173275e-06, + "loss": 10.73328971862793, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.39217647910118103, + "loss_total_v6": 0.5406140089035034, + "step": 438 + }, + { + "epoch": 0.16362281028699216, + "grad_norm": 21.290246963500977, + "learning_rate": 9.77313357753148e-06, + "loss": 11.132549285888672, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.542256772518158, + "loss_total_v6": 0.688741147518158, + "step": 439 + }, + { + "epoch": 0.1639955273947074, + "grad_norm": 20.983325958251953, + "learning_rate": 9.771245782336596e-06, + "loss": 10.759343147277832, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5954297184944153, + "loss_total_v6": 0.5954297184944153, + "step": 440 + }, + { + "epoch": 0.16436824450242266, + "grad_norm": 21.42027473449707, + "learning_rate": 9.769350349170145e-06, + "loss": 11.05537223815918, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.7021195888519287, + "loss_total_v6": 0.8525102138519287, + "step": 441 + }, + { + "epoch": 0.1647409616101379, + "grad_norm": 21.52903938293457, + "learning_rate": 9.767447281066399e-06, + "loss": 11.099339485168457, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6400485038757324, + "loss_total_v6": 0.7982516288757324, + "step": 442 + }, + { + "epoch": 0.16511367871785315, + "grad_norm": 21.11085319519043, + "learning_rate": 9.765536581071854e-06, + "loss": 11.184331893920898, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4299594759941101, + "loss_total_v6": 0.5842563509941101, + "step": 443 + }, + { + "epoch": 0.1654863958255684, + "grad_norm": 19.858715057373047, + "learning_rate": 9.763618252245224e-06, + "loss": 10.537578582763672, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5617650747299194, + "loss_total_v6": 0.7199681997299194, + "step": 444 + }, + { + "epoch": 0.16585911293328365, + "grad_norm": 22.649898529052734, + "learning_rate": 9.761692297657432e-06, + "loss": 10.792835235595703, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.37453073263168335, + "loss_total_v6": 0.5268744826316833, + "step": 445 + }, + { + "epoch": 0.16623183004099887, + "grad_norm": 19.89338493347168, + "learning_rate": 9.759758720391616e-06, + "loss": 10.907858848571777, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.24361290037631989, + "loss_total_v6": 0.4018160104751587, + "step": 446 + }, + { + "epoch": 0.16660454714871412, + "grad_norm": 21.890914916992188, + "learning_rate": 9.75781752354311e-06, + "loss": 10.88514518737793, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.7250400185585022, + "loss_total_v6": 0.8656650185585022, + "step": 447 + }, + { + "epoch": 0.16697726425642936, + "grad_norm": 21.28089141845703, + "learning_rate": 9.755868710219444e-06, + "loss": 10.842275619506836, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.48660388588905334, + "loss_total_v6": 0.633088231086731, + "step": 448 + }, + { + "epoch": 0.1673499813641446, + "grad_norm": 20.950546264648438, + "learning_rate": 9.753912283540348e-06, + "loss": 10.400175094604492, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5325614809989929, + "loss_total_v6": 0.6966239809989929, + "step": 449 + }, + { + "epoch": 0.16772269847185986, + "grad_norm": 21.108966827392578, + "learning_rate": 9.751948246637734e-06, + "loss": 11.028120040893555, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.19423140585422516, + "loss_total_v6": 0.34657514095306396, + "step": 450 + }, + { + "epoch": 0.1680954155795751, + "grad_norm": 21.264034271240234, + "learning_rate": 9.749976602655701e-06, + "loss": 11.018844604492188, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7241370677947998, + "loss_total_v6": 0.7241370677947998, + "step": 451 + }, + { + "epoch": 0.16846813268729036, + "grad_norm": 20.92652702331543, + "learning_rate": 9.747997354750523e-06, + "loss": 10.420624732971191, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6034488081932068, + "loss_total_v6": 0.7694644331932068, + "step": 452 + }, + { + "epoch": 0.1688408497950056, + "grad_norm": 20.48203468322754, + "learning_rate": 9.746010506090645e-06, + "loss": 10.729405403137207, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7531223893165588, + "loss_total_v6": 0.7531223893165588, + "step": 453 + }, + { + "epoch": 0.16921356690272082, + "grad_norm": 20.80611228942871, + "learning_rate": 9.744016059856682e-06, + "loss": 10.910205841064453, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.26648038625717163, + "loss_total_v6": 0.43249601125717163, + "step": 454 + }, + { + "epoch": 0.16958628401043607, + "grad_norm": 20.92402458190918, + "learning_rate": 9.742014019241413e-06, + "loss": 10.942037582397461, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.48797544836997986, + "loss_total_v6": 0.6461786031723022, + "step": 455 + }, + { + "epoch": 0.16995900111815132, + "grad_norm": 25.675678253173828, + "learning_rate": 9.74000438744977e-06, + "loss": 10.353361129760742, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.46844273805618286, + "loss_total_v6": 0.46844273805618286, + "step": 456 + }, + { + "epoch": 0.17033171822586657, + "grad_norm": 25.195281982421875, + "learning_rate": 9.737987167698839e-06, + "loss": 11.321041107177734, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6683962941169739, + "loss_total_v6": 0.8285525441169739, + "step": 457 + }, + { + "epoch": 0.17070443533358182, + "grad_norm": 23.547679901123047, + "learning_rate": 9.735962363217854e-06, + "loss": 11.180096626281738, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.15136225521564484, + "loss_total_v6": 0.30761224031448364, + "step": 458 + }, + { + "epoch": 0.17107715244129706, + "grad_norm": 20.386096954345703, + "learning_rate": 9.733929977248194e-06, + "loss": 11.278165817260742, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.35507330298423767, + "loss_total_v6": 0.5113233327865601, + "step": 459 + }, + { + "epoch": 0.1714498695490123, + "grad_norm": 20.971948623657227, + "learning_rate": 9.731890013043367e-06, + "loss": 11.3346529006958, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3389471471309662, + "loss_total_v6": 0.4971502721309662, + "step": 460 + }, + { + "epoch": 0.17182258665672753, + "grad_norm": 21.378881454467773, + "learning_rate": 9.72984247386902e-06, + "loss": 10.720638275146484, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.674212634563446, + "loss_total_v6": 0.824603259563446, + "step": 461 + }, + { + "epoch": 0.17219530376444278, + "grad_norm": 20.93598747253418, + "learning_rate": 9.72778736300292e-06, + "loss": 10.613014221191406, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5751492977142334, + "loss_total_v6": 0.7333524227142334, + "step": 462 + }, + { + "epoch": 0.17256802087215803, + "grad_norm": 21.52539825439453, + "learning_rate": 9.72572468373496e-06, + "loss": 10.936567306518555, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.46519190073013306, + "loss_total_v6": 0.46519190073013306, + "step": 463 + }, + { + "epoch": 0.17294073797987328, + "grad_norm": 20.592905044555664, + "learning_rate": 9.723654439367146e-06, + "loss": 10.896614074707031, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 1.0001164674758911, + "loss_total_v6": 1.1817570924758911, + "step": 464 + }, + { + "epoch": 0.17331345508758852, + "grad_norm": 20.204090118408203, + "learning_rate": 9.721576633213593e-06, + "loss": 10.558393478393555, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5251951813697815, + "loss_total_v6": 0.6833983063697815, + "step": 465 + }, + { + "epoch": 0.17368617219530377, + "grad_norm": 21.43115234375, + "learning_rate": 9.719491268600527e-06, + "loss": 10.566516876220703, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.40690290927886963, + "loss_total_v6": 0.5533872842788696, + "step": 466 + }, + { + "epoch": 0.17405888930301902, + "grad_norm": 25.997177124023438, + "learning_rate": 9.717398348866266e-06, + "loss": 10.894552230834961, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5270963907241821, + "loss_total_v6": 0.6833463907241821, + "step": 467 + }, + { + "epoch": 0.17443160641073424, + "grad_norm": 20.854019165039062, + "learning_rate": 9.715297877361231e-06, + "loss": 10.647363662719727, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6525829434394836, + "loss_total_v6": 0.8049266934394836, + "step": 468 + }, + { + "epoch": 0.1748043235184495, + "grad_norm": 23.064208984375, + "learning_rate": 9.713189857447926e-06, + "loss": 10.848617553710938, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7962415814399719, + "loss_total_v6": 0.9485853314399719, + "step": 469 + }, + { + "epoch": 0.17517704062616474, + "grad_norm": 19.758684158325195, + "learning_rate": 9.71107429250094e-06, + "loss": 10.807037353515625, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.2837807834148407, + "loss_total_v6": 0.4400307834148407, + "step": 470 + }, + { + "epoch": 0.17554975773387999, + "grad_norm": 20.94491958618164, + "learning_rate": 9.708951185906943e-06, + "loss": 11.142560958862305, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6644638180732727, + "loss_total_v6": 0.8187606930732727, + "step": 471 + }, + { + "epoch": 0.17592247484159523, + "grad_norm": 20.523319244384766, + "learning_rate": 9.706820541064675e-06, + "loss": 10.799959182739258, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.4976331889629364, + "loss_total_v6": 0.6675550937652588, + "step": 472 + }, + { + "epoch": 0.17629519194931048, + "grad_norm": 35.602725982666016, + "learning_rate": 9.704682361384941e-06, + "loss": 10.578387260437012, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.25910887122154236, + "loss_total_v6": 0.41145262122154236, + "step": 473 + }, + { + "epoch": 0.17666790905702573, + "grad_norm": 20.934814453125, + "learning_rate": 9.702536650290615e-06, + "loss": 11.060168266296387, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5709978938102722, + "loss_total_v6": 0.7311541438102722, + "step": 474 + }, + { + "epoch": 0.17704062616474095, + "grad_norm": 21.14507484436035, + "learning_rate": 9.700383411216623e-06, + "loss": 10.954195022583008, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6770729422569275, + "loss_total_v6": 0.8411354422569275, + "step": 475 + }, + { + "epoch": 0.1774133432724562, + "grad_norm": 20.174524307250977, + "learning_rate": 9.698222647609941e-06, + "loss": 10.233619689941406, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5568335056304932, + "loss_total_v6": 0.7189428806304932, + "step": 476 + }, + { + "epoch": 0.17778606038017145, + "grad_norm": 20.43751335144043, + "learning_rate": 9.696054362929597e-06, + "loss": 10.766202926635742, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.7397504448890686, + "loss_total_v6": 0.9077191948890686, + "step": 477 + }, + { + "epoch": 0.1781587774878867, + "grad_norm": 21.06536102294922, + "learning_rate": 9.69387856064665e-06, + "loss": 10.893203735351562, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4504096806049347, + "loss_total_v6": 0.6047065258026123, + "step": 478 + }, + { + "epoch": 0.17853149459560194, + "grad_norm": 22.123136520385742, + "learning_rate": 9.691695244244199e-06, + "loss": 10.522293090820312, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5653221607208252, + "loss_total_v6": 0.7332909107208252, + "step": 479 + }, + { + "epoch": 0.1789042117033172, + "grad_norm": 20.289134979248047, + "learning_rate": 9.689504417217371e-06, + "loss": 10.715692520141602, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6051525473594666, + "loss_total_v6": 0.7633556723594666, + "step": 480 + }, + { + "epoch": 0.17927692881103244, + "grad_norm": 20.414398193359375, + "learning_rate": 9.687306083073316e-06, + "loss": 10.195802688598633, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7713183760643005, + "loss_total_v6": 0.9314746260643005, + "step": 481 + }, + { + "epoch": 0.17964964591874766, + "grad_norm": 20.1355037689209, + "learning_rate": 9.685100245331202e-06, + "loss": 10.698356628417969, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6002524495124817, + "loss_total_v6": 0.7506430745124817, + "step": 482 + }, + { + "epoch": 0.1800223630264629, + "grad_norm": 20.480592727661133, + "learning_rate": 9.68288690752221e-06, + "loss": 10.506152153015137, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5140690803527832, + "loss_total_v6": 0.6703190803527832, + "step": 483 + }, + { + "epoch": 0.18039508013417815, + "grad_norm": 21.21554183959961, + "learning_rate": 9.680666073189522e-06, + "loss": 10.649763107299805, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6699121594429016, + "loss_total_v6": 0.8281152844429016, + "step": 484 + }, + { + "epoch": 0.1807677972418934, + "grad_norm": 20.36192512512207, + "learning_rate": 9.678437745888328e-06, + "loss": 10.977621078491211, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5962189435958862, + "loss_total_v6": 0.7505158185958862, + "step": 485 + }, + { + "epoch": 0.18114051434960865, + "grad_norm": 20.76042366027832, + "learning_rate": 9.676201929185809e-06, + "loss": 10.748703956604004, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5686967968940735, + "loss_total_v6": 0.7249467968940735, + "step": 486 + }, + { + "epoch": 0.1815132314573239, + "grad_norm": 19.545700073242188, + "learning_rate": 9.673958626661135e-06, + "loss": 10.346196174621582, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.40671461820602417, + "loss_total_v6": 0.5551521182060242, + "step": 487 + }, + { + "epoch": 0.18188594856503915, + "grad_norm": 20.00909996032715, + "learning_rate": 9.671707841905462e-06, + "loss": 10.648526191711426, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.14122766256332397, + "loss_total_v6": 0.289665162563324, + "step": 488 + }, + { + "epoch": 0.18225866567275437, + "grad_norm": 21.402544021606445, + "learning_rate": 9.66944957852192e-06, + "loss": 10.832386016845703, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.40963509678840637, + "loss_total_v6": 0.567838191986084, + "step": 489 + }, + { + "epoch": 0.18263138278046961, + "grad_norm": 19.862417221069336, + "learning_rate": 9.667183840125619e-06, + "loss": 10.925893783569336, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.49815070629119873, + "loss_total_v6": 0.6504944562911987, + "step": 490 + }, + { + "epoch": 0.18300409988818486, + "grad_norm": 19.8511962890625, + "learning_rate": 9.664910630343622e-06, + "loss": 10.885456085205078, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6013116836547852, + "loss_total_v6": 0.6013116836547852, + "step": 491 + }, + { + "epoch": 0.1833768169959001, + "grad_norm": 21.191566467285156, + "learning_rate": 9.662629952814964e-06, + "loss": 10.895037651062012, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7372295260429382, + "loss_total_v6": 0.7372295260429382, + "step": 492 + }, + { + "epoch": 0.18374953410361536, + "grad_norm": 22.738317489624023, + "learning_rate": 9.66034181119063e-06, + "loss": 11.126462936401367, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.611508309841156, + "loss_total_v6": 0.775570809841156, + "step": 493 + }, + { + "epoch": 0.1841222512113306, + "grad_norm": 21.905902862548828, + "learning_rate": 9.658046209133557e-06, + "loss": 11.044755935668945, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.5937763452529907, + "loss_total_v6": 0.7636982202529907, + "step": 494 + }, + { + "epoch": 0.18449496831904585, + "grad_norm": 21.448034286499023, + "learning_rate": 9.655743150318618e-06, + "loss": 10.734565734863281, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3327285945415497, + "loss_total_v6": 0.4831192195415497, + "step": 495 + }, + { + "epoch": 0.18486768542676107, + "grad_norm": 21.331573486328125, + "learning_rate": 9.653432638432631e-06, + "loss": 11.400251388549805, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.31312450766563416, + "loss_total_v6": 0.46742138266563416, + "step": 496 + }, + { + "epoch": 0.18524040253447632, + "grad_norm": 20.166711807250977, + "learning_rate": 9.651114677174338e-06, + "loss": 10.795427322387695, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.8170276880264282, + "loss_total_v6": 0.9693714380264282, + "step": 497 + }, + { + "epoch": 0.18561311964219157, + "grad_norm": 20.459049224853516, + "learning_rate": 9.648789270254415e-06, + "loss": 10.67059326171875, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.81296306848526, + "loss_total_v6": 0.96921306848526, + "step": 498 + }, + { + "epoch": 0.18598583674990682, + "grad_norm": 20.283771514892578, + "learning_rate": 9.646456421395447e-06, + "loss": 10.851800918579102, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.43309450149536133, + "loss_total_v6": 0.5834851264953613, + "step": 499 + }, + { + "epoch": 0.18635855385762207, + "grad_norm": 20.99987030029297, + "learning_rate": 9.64411613433194e-06, + "loss": 11.110597610473633, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4486486315727234, + "loss_total_v6": 0.6107580065727234, + "step": 500 + }, + { + "epoch": 0.18635855385762207, + "eval_loss": 0.6401329636573792, + "eval_loss_alignment": 0.2890812800480769, + "eval_loss_alignment_w": 0.14454064002403846, + "eval_loss_sft": 0.4955922722832387, + "eval_loss_total_v6": 0.6401329120127563, + "eval_runtime": 206.4837, + "eval_samples_per_second": 9.037, + "eval_steps_per_second": 1.133, + "step": 500 + }, + { + "epoch": 0.18673127096533731, + "grad_norm": 20.363306045532227, + "learning_rate": 9.641768412810305e-06, + "loss": 10.80221939086914, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4437483251094818, + "loss_total_v6": 0.6058577299118042, + "step": 501 + }, + { + "epoch": 0.18710398807305256, + "grad_norm": 20.286020278930664, + "learning_rate": 9.63941326058885e-06, + "loss": 10.875722885131836, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5612384676933289, + "loss_total_v6": 0.7213947176933289, + "step": 502 + }, + { + "epoch": 0.1874767051807678, + "grad_norm": 19.897123336791992, + "learning_rate": 9.637050681437787e-06, + "loss": 10.910832405090332, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.8145530819892883, + "loss_total_v6": 0.9805687069892883, + "step": 503 + }, + { + "epoch": 0.18784942228848303, + "grad_norm": 21.2502498626709, + "learning_rate": 9.634680679139213e-06, + "loss": 10.698272705078125, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4775601029396057, + "loss_total_v6": 0.6279507279396057, + "step": 504 + }, + { + "epoch": 0.18822213939619828, + "grad_norm": 19.74827003479004, + "learning_rate": 9.632303257487102e-06, + "loss": 10.177326202392578, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.48835253715515137, + "loss_total_v6": 0.6426494121551514, + "step": 505 + }, + { + "epoch": 0.18859485650391353, + "grad_norm": 20.21869659423828, + "learning_rate": 9.629918420287316e-06, + "loss": 10.978227615356445, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2445695549249649, + "loss_total_v6": 0.3949601650238037, + "step": 506 + }, + { + "epoch": 0.18896757361162878, + "grad_norm": 23.06920051574707, + "learning_rate": 9.627526171357582e-06, + "loss": 10.742166519165039, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7750749588012695, + "loss_total_v6": 0.9313249588012695, + "step": 507 + }, + { + "epoch": 0.18934029071934402, + "grad_norm": 20.907459259033203, + "learning_rate": 9.625126514527492e-06, + "loss": 10.73603343963623, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.41968488693237305, + "loss_total_v6": 0.570075511932373, + "step": 508 + }, + { + "epoch": 0.18971300782705927, + "grad_norm": 21.63737678527832, + "learning_rate": 9.622719453638499e-06, + "loss": 10.776741027832031, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4938976466655731, + "loss_total_v6": 0.6442883014678955, + "step": 509 + }, + { + "epoch": 0.19008572493477452, + "grad_norm": 20.834012985229492, + "learning_rate": 9.620304992543908e-06, + "loss": 10.473770141601562, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.47598350048065186, + "loss_total_v6": 0.47598350048065186, + "step": 510 + }, + { + "epoch": 0.19045844204248974, + "grad_norm": 22.05328941345215, + "learning_rate": 9.617883135108867e-06, + "loss": 10.90552806854248, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.676301121711731, + "loss_total_v6": 0.836457371711731, + "step": 511 + }, + { + "epoch": 0.190831159150205, + "grad_norm": 20.2021541595459, + "learning_rate": 9.615453885210368e-06, + "loss": 10.18064022064209, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.14851114153862, + "loss_total_v6": 0.14851114153862, + "step": 512 + }, + { + "epoch": 0.19120387625792024, + "grad_norm": 22.037073135375977, + "learning_rate": 9.613017246737239e-06, + "loss": 10.996655464172363, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5454165935516357, + "loss_total_v6": 0.7055728435516357, + "step": 513 + }, + { + "epoch": 0.19157659336563548, + "grad_norm": 25.070825576782227, + "learning_rate": 9.610573223590132e-06, + "loss": 10.64869499206543, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5627189874649048, + "loss_total_v6": 0.7228752374649048, + "step": 514 + }, + { + "epoch": 0.19194931047335073, + "grad_norm": 21.23046112060547, + "learning_rate": 9.608121819681521e-06, + "loss": 10.706674575805664, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6898404359817505, + "loss_total_v6": 0.8402310609817505, + "step": 515 + }, + { + "epoch": 0.19232202758106598, + "grad_norm": 23.10162925720215, + "learning_rate": 9.605663038935698e-06, + "loss": 10.473968505859375, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.22890286147594452, + "loss_total_v6": 0.3871060013771057, + "step": 516 + }, + { + "epoch": 0.19269474468878123, + "grad_norm": 20.940189361572266, + "learning_rate": 9.603196885288759e-06, + "loss": 10.297701835632324, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5922445058822632, + "loss_total_v6": 0.7348226308822632, + "step": 517 + }, + { + "epoch": 0.19306746179649645, + "grad_norm": 27.474023818969727, + "learning_rate": 9.60072336268861e-06, + "loss": 10.886380195617676, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.44727668166160583, + "loss_total_v6": 0.6035267114639282, + "step": 518 + }, + { + "epoch": 0.1934401789042117, + "grad_norm": 21.628992080688477, + "learning_rate": 9.598242475094948e-06, + "loss": 10.942192077636719, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5626935362815857, + "loss_total_v6": 0.7130841612815857, + "step": 519 + }, + { + "epoch": 0.19381289601192694, + "grad_norm": 21.36297607421875, + "learning_rate": 9.59575422647926e-06, + "loss": 11.261552810668945, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6645832061767578, + "loss_total_v6": 0.8149738311767578, + "step": 520 + }, + { + "epoch": 0.1941856131196422, + "grad_norm": 20.662086486816406, + "learning_rate": 9.593258620824822e-06, + "loss": 10.568257331848145, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3782343864440918, + "loss_total_v6": 0.5403437614440918, + "step": 521 + }, + { + "epoch": 0.19455833022735744, + "grad_norm": 21.42827033996582, + "learning_rate": 9.590755662126683e-06, + "loss": 10.362619400024414, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5165591239929199, + "loss_total_v6": 0.6767153739929199, + "step": 522 + }, + { + "epoch": 0.1949310473350727, + "grad_norm": 18.95876693725586, + "learning_rate": 9.58824535439166e-06, + "loss": 10.856792449951172, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7119036912918091, + "loss_total_v6": 0.8720599412918091, + "step": 523 + }, + { + "epoch": 0.19530376444278794, + "grad_norm": 19.777467727661133, + "learning_rate": 9.585727701638343e-06, + "loss": 11.005253791809082, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.720456063747406, + "loss_total_v6": 0.720456063747406, + "step": 524 + }, + { + "epoch": 0.19567648155050316, + "grad_norm": 20.20841407775879, + "learning_rate": 9.583202707897075e-06, + "loss": 10.465526580810547, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.444733589887619, + "loss_total_v6": 0.6146554946899414, + "step": 525 + }, + { + "epoch": 0.1960491986582184, + "grad_norm": 20.11646842956543, + "learning_rate": 9.580670377209949e-06, + "loss": 10.83837604522705, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6021455526351929, + "loss_total_v6": 0.7505830526351929, + "step": 526 + }, + { + "epoch": 0.19642191576593365, + "grad_norm": 22.05068588256836, + "learning_rate": 9.578130713630809e-06, + "loss": 10.820718765258789, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.48493102192878723, + "loss_total_v6": 0.6392278671264648, + "step": 527 + }, + { + "epoch": 0.1967946328736489, + "grad_norm": 19.957548141479492, + "learning_rate": 9.57558372122523e-06, + "loss": 10.516803741455078, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.5642390251159668, + "loss_total_v6": 0.7419734001159668, + "step": 528 + }, + { + "epoch": 0.19716734998136415, + "grad_norm": 20.959264755249023, + "learning_rate": 9.573029404070529e-06, + "loss": 10.652360916137695, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.46581119298934937, + "loss_total_v6": 0.6279205679893494, + "step": 529 + }, + { + "epoch": 0.1975400670890794, + "grad_norm": 21.250593185424805, + "learning_rate": 9.570467766255737e-06, + "loss": 10.595315933227539, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.23562820255756378, + "loss_total_v6": 0.3801594376564026, + "step": 530 + }, + { + "epoch": 0.19791278419679464, + "grad_norm": 20.82377815246582, + "learning_rate": 9.567898811881615e-06, + "loss": 11.065237045288086, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.632617712020874, + "loss_total_v6": 0.784961462020874, + "step": 531 + }, + { + "epoch": 0.19828550130450986, + "grad_norm": 21.832717895507812, + "learning_rate": 9.565322545060631e-06, + "loss": 11.059711456298828, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4839165210723877, + "loss_total_v6": 0.6382133960723877, + "step": 532 + }, + { + "epoch": 0.1986582184122251, + "grad_norm": 19.364608764648438, + "learning_rate": 9.56273896991696e-06, + "loss": 10.846482276916504, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5685622692108154, + "loss_total_v6": 0.7267653942108154, + "step": 533 + }, + { + "epoch": 0.19903093551994036, + "grad_norm": 20.822019577026367, + "learning_rate": 9.560148090586475e-06, + "loss": 10.703420639038086, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.53298020362854, + "loss_total_v6": 0.67555832862854, + "step": 534 + }, + { + "epoch": 0.1994036526276556, + "grad_norm": 19.5634822845459, + "learning_rate": 9.557549911216745e-06, + "loss": 10.598161697387695, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2270423322916031, + "loss_total_v6": 0.3774329423904419, + "step": 535 + }, + { + "epoch": 0.19977636973537086, + "grad_norm": 19.919872283935547, + "learning_rate": 9.554944435967024e-06, + "loss": 10.64052963256836, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.41065120697021484, + "loss_total_v6": 0.5708074569702148, + "step": 536 + }, + { + "epoch": 0.2001490868430861, + "grad_norm": 19.867996215820312, + "learning_rate": 9.552331669008243e-06, + "loss": 11.056713104248047, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.31046292185783386, + "loss_total_v6": 0.46280667185783386, + "step": 537 + }, + { + "epoch": 0.20052180395080135, + "grad_norm": 22.495628356933594, + "learning_rate": 9.549711614523007e-06, + "loss": 10.863077163696289, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7463144063949585, + "loss_total_v6": 0.8888925313949585, + "step": 538 + }, + { + "epoch": 0.20089452105851657, + "grad_norm": 21.617267608642578, + "learning_rate": 9.547084276705591e-06, + "loss": 10.559889793395996, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.33756834268569946, + "loss_total_v6": 0.49967771768569946, + "step": 539 + }, + { + "epoch": 0.20126723816623182, + "grad_norm": 20.206201553344727, + "learning_rate": 9.544449659761923e-06, + "loss": 10.502864837646484, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6803160309791565, + "loss_total_v6": 0.8404722809791565, + "step": 540 + }, + { + "epoch": 0.20163995527394707, + "grad_norm": 21.67242431640625, + "learning_rate": 9.541807767909589e-06, + "loss": 10.787635803222656, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.40741950273513794, + "loss_total_v6": 0.5558570027351379, + "step": 541 + }, + { + "epoch": 0.20201267238166232, + "grad_norm": 19.745126724243164, + "learning_rate": 9.539158605377817e-06, + "loss": 10.687127113342285, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.35520297288894653, + "loss_total_v6": 0.5094998478889465, + "step": 542 + }, + { + "epoch": 0.20238538948937757, + "grad_norm": 19.959796905517578, + "learning_rate": 9.536502176407479e-06, + "loss": 10.617915153503418, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.31359127163887024, + "loss_total_v6": 0.47960689663887024, + "step": 543 + }, + { + "epoch": 0.2027581065970928, + "grad_norm": 22.905960083007812, + "learning_rate": 9.533838485251073e-06, + "loss": 10.304606437683105, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6143130660057068, + "loss_total_v6": 0.7783755660057068, + "step": 544 + }, + { + "epoch": 0.20313082370480806, + "grad_norm": 20.213275909423828, + "learning_rate": 9.531167536172726e-06, + "loss": 10.588693618774414, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7990896105766296, + "loss_total_v6": 0.9651052355766296, + "step": 545 + }, + { + "epoch": 0.20350354081252328, + "grad_norm": 21.4104061126709, + "learning_rate": 9.528489333448186e-06, + "loss": 10.796764373779297, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5377131700515747, + "loss_total_v6": 0.6939631700515747, + "step": 546 + }, + { + "epoch": 0.20387625792023853, + "grad_norm": 19.378257751464844, + "learning_rate": 9.52580388136481e-06, + "loss": 10.747916221618652, + "loss_alignment": 0.26953125, + "loss_alignment_w": 0.134765625, + "loss_sft": 0.5141235589981079, + "loss_total_v6": 0.6488891839981079, + "step": 547 + }, + { + "epoch": 0.20424897502795378, + "grad_norm": 20.22576332092285, + "learning_rate": 9.523111184221559e-06, + "loss": 10.357135772705078, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.8126034140586853, + "loss_total_v6": 0.9766659140586853, + "step": 548 + }, + { + "epoch": 0.20462169213566903, + "grad_norm": 20.623441696166992, + "learning_rate": 9.520411246328995e-06, + "loss": 10.656484603881836, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6042689085006714, + "loss_total_v6": 0.7585657835006714, + "step": 549 + }, + { + "epoch": 0.20499440924338427, + "grad_norm": 20.70395278930664, + "learning_rate": 9.517704072009269e-06, + "loss": 11.105523109436035, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.0883333757519722, + "loss_total_v6": 0.2465364933013916, + "step": 550 + }, + { + "epoch": 0.20536712635109952, + "grad_norm": 21.029626846313477, + "learning_rate": 9.514989665596114e-06, + "loss": 10.323655128479004, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.4896044433116913, + "loss_total_v6": 0.6497607231140137, + "step": 551 + }, + { + "epoch": 0.20573984345881477, + "grad_norm": 20.638248443603516, + "learning_rate": 9.512268031434852e-06, + "loss": 10.8798828125, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.6793403029441833, + "loss_total_v6": 0.8492621779441833, + "step": 552 + }, + { + "epoch": 0.20611256056653002, + "grad_norm": 33.003231048583984, + "learning_rate": 9.50953917388236e-06, + "loss": 11.190340042114258, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7418990731239319, + "loss_total_v6": 0.8844771981239319, + "step": 553 + }, + { + "epoch": 0.20648527767424524, + "grad_norm": 20.498384475708008, + "learning_rate": 9.506803097307087e-06, + "loss": 10.819112777709961, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.8163158893585205, + "loss_total_v6": 0.9823315143585205, + "step": 554 + }, + { + "epoch": 0.20685799478196049, + "grad_norm": 19.982084274291992, + "learning_rate": 9.504059806089037e-06, + "loss": 10.435559272766113, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.35594338178634644, + "loss_total_v6": 0.49852150678634644, + "step": 555 + }, + { + "epoch": 0.20723071188967573, + "grad_norm": 20.319580078125, + "learning_rate": 9.501309304619765e-06, + "loss": 10.595762252807617, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2201283872127533, + "loss_total_v6": 0.3724721372127533, + "step": 556 + }, + { + "epoch": 0.20760342899739098, + "grad_norm": 20.153427124023438, + "learning_rate": 9.498551597302364e-06, + "loss": 10.498897552490234, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5659708976745605, + "loss_total_v6": 0.7300333976745605, + "step": 557 + }, + { + "epoch": 0.20797614610510623, + "grad_norm": 20.043256759643555, + "learning_rate": 9.495786688551467e-06, + "loss": 10.759626388549805, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.8772667050361633, + "loss_total_v6": 1.0569541454315186, + "step": 558 + }, + { + "epoch": 0.20834886321282148, + "grad_norm": 20.298198699951172, + "learning_rate": 9.493014582793234e-06, + "loss": 10.574451446533203, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.1323319375514984, + "loss_total_v6": 0.2807694375514984, + "step": 559 + }, + { + "epoch": 0.20872158032053673, + "grad_norm": 19.68936538696289, + "learning_rate": 9.490235284465343e-06, + "loss": 10.402654647827148, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.33140361309051514, + "loss_total_v6": 0.47984111309051514, + "step": 560 + }, + { + "epoch": 0.20909429742825195, + "grad_norm": 19.599014282226562, + "learning_rate": 9.48744879801699e-06, + "loss": 10.947955131530762, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4722130596637726, + "loss_total_v6": 0.6186974048614502, + "step": 561 + }, + { + "epoch": 0.2094670145359672, + "grad_norm": 19.495763778686523, + "learning_rate": 9.484655127908878e-06, + "loss": 10.902610778808594, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6683700680732727, + "loss_total_v6": 0.8304794430732727, + "step": 562 + }, + { + "epoch": 0.20983973164368244, + "grad_norm": 19.958221435546875, + "learning_rate": 9.481854278613208e-06, + "loss": 10.845463752746582, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.34389516711235046, + "loss_total_v6": 0.5060045719146729, + "step": 563 + }, + { + "epoch": 0.2102124487513977, + "grad_norm": 20.713712692260742, + "learning_rate": 9.479046254613673e-06, + "loss": 10.684883117675781, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.38709619641304016, + "loss_total_v6": 0.5413930416107178, + "step": 564 + }, + { + "epoch": 0.21058516585911294, + "grad_norm": 19.053415298461914, + "learning_rate": 9.476231060405452e-06, + "loss": 10.953866958618164, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.40171414613723755, + "loss_total_v6": 0.5540578961372375, + "step": 565 + }, + { + "epoch": 0.2109578829668282, + "grad_norm": 20.181102752685547, + "learning_rate": 9.473408700495207e-06, + "loss": 10.33537769317627, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5589158535003662, + "loss_total_v6": 0.7112596035003662, + "step": 566 + }, + { + "epoch": 0.21133060007454343, + "grad_norm": 19.959558486938477, + "learning_rate": 9.470579179401065e-06, + "loss": 10.208800315856934, + "loss_alignment": 0.3671875, + "loss_alignment_w": 0.18359375, + "loss_sft": 0.7342295050621033, + "loss_total_v6": 0.9178232550621033, + "step": 567 + }, + { + "epoch": 0.21170331718225865, + "grad_norm": 20.063697814941406, + "learning_rate": 9.467742501652618e-06, + "loss": 10.505708694458008, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6275531053543091, + "loss_total_v6": 0.7701312303543091, + "step": 568 + }, + { + "epoch": 0.2120760342899739, + "grad_norm": 19.61211585998535, + "learning_rate": 9.464898671790917e-06, + "loss": 10.618064880371094, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5242488384246826, + "loss_total_v6": 0.6863582134246826, + "step": 569 + }, + { + "epoch": 0.21244875139768915, + "grad_norm": 19.899011611938477, + "learning_rate": 9.462047694368459e-06, + "loss": 10.833480834960938, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3354488015174866, + "loss_total_v6": 0.4975581765174866, + "step": 570 + }, + { + "epoch": 0.2128214685054044, + "grad_norm": 22.54633903503418, + "learning_rate": 9.459189573949187e-06, + "loss": 11.255403518676758, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.896774172782898, + "loss_total_v6": 1.060836672782898, + "step": 571 + }, + { + "epoch": 0.21319418561311965, + "grad_norm": 19.849803924560547, + "learning_rate": 9.456324315108476e-06, + "loss": 10.64378547668457, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.39226314425468445, + "loss_total_v6": 0.5426537990570068, + "step": 572 + }, + { + "epoch": 0.2135669027208349, + "grad_norm": 19.64010238647461, + "learning_rate": 9.45345192243313e-06, + "loss": 10.825605392456055, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.41567927598953247, + "loss_total_v6": 0.5621636509895325, + "step": 573 + }, + { + "epoch": 0.21393961982855014, + "grad_norm": 19.850910186767578, + "learning_rate": 9.450572400521371e-06, + "loss": 10.603312492370605, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.46870654821395874, + "loss_total_v6": 0.6230034232139587, + "step": 574 + }, + { + "epoch": 0.21431233693626536, + "grad_norm": 20.407268524169922, + "learning_rate": 9.447685753982837e-06, + "loss": 10.492708206176758, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7896159887313843, + "loss_total_v6": 0.9439128637313843, + "step": 575 + }, + { + "epoch": 0.2146850540439806, + "grad_norm": 19.673818588256836, + "learning_rate": 9.44479198743857e-06, + "loss": 10.637666702270508, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2996925711631775, + "loss_total_v6": 0.4520363211631775, + "step": 576 + }, + { + "epoch": 0.21505777115169586, + "grad_norm": 20.51412010192871, + "learning_rate": 9.441891105521005e-06, + "loss": 10.349771499633789, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.291892409324646, + "loss_total_v6": 0.446189284324646, + "step": 577 + }, + { + "epoch": 0.2154304882594111, + "grad_norm": 19.42109489440918, + "learning_rate": 9.438983112873977e-06, + "loss": 10.532623291015625, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5881061553955078, + "loss_total_v6": 0.7384967803955078, + "step": 578 + }, + { + "epoch": 0.21580320536712636, + "grad_norm": 19.265705108642578, + "learning_rate": 9.436068014152692e-06, + "loss": 10.454669952392578, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3269440531730652, + "loss_total_v6": 0.4753815531730652, + "step": 579 + }, + { + "epoch": 0.2161759224748416, + "grad_norm": 19.059091567993164, + "learning_rate": 9.433145814023747e-06, + "loss": 10.553180694580078, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2569444477558136, + "loss_total_v6": 0.4092881977558136, + "step": 580 + }, + { + "epoch": 0.21654863958255685, + "grad_norm": 21.45777702331543, + "learning_rate": 9.430216517165094e-06, + "loss": 10.587305068969727, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.8640875816345215, + "loss_total_v6": 1.0301032066345215, + "step": 581 + }, + { + "epoch": 0.21692135669027207, + "grad_norm": 22.34040641784668, + "learning_rate": 9.427280128266049e-06, + "loss": 10.751550674438477, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.41940563917160034, + "loss_total_v6": 0.5873743891716003, + "step": 582 + }, + { + "epoch": 0.21729407379798732, + "grad_norm": 22.98474884033203, + "learning_rate": 9.424336652027288e-06, + "loss": 10.777807235717773, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5970267653465271, + "loss_total_v6": 0.7474173903465271, + "step": 583 + }, + { + "epoch": 0.21766679090570257, + "grad_norm": 19.905858993530273, + "learning_rate": 9.421386093160823e-06, + "loss": 10.955205917358398, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7953162789344788, + "loss_total_v6": 0.9496131539344788, + "step": 584 + }, + { + "epoch": 0.21803950801341782, + "grad_norm": 20.049442291259766, + "learning_rate": 9.418428456390008e-06, + "loss": 10.919537544250488, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6287957429885864, + "loss_total_v6": 0.7909051179885864, + "step": 585 + }, + { + "epoch": 0.21841222512113306, + "grad_norm": 18.94263458251953, + "learning_rate": 9.415463746449532e-06, + "loss": 10.668828964233398, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.20191890001296997, + "loss_total_v6": 0.35426265001296997, + "step": 586 + }, + { + "epoch": 0.2187849422288483, + "grad_norm": 20.19829559326172, + "learning_rate": 9.4124919680854e-06, + "loss": 10.761632919311523, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7078631520271301, + "loss_total_v6": 0.7078631520271301, + "step": 587 + }, + { + "epoch": 0.21915765933656356, + "grad_norm": 19.40629005432129, + "learning_rate": 9.409513126054934e-06, + "loss": 10.458111763000488, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.9633920192718506, + "loss_total_v6": 1.1235482692718506, + "step": 588 + }, + { + "epoch": 0.21953037644427878, + "grad_norm": 19.737995147705078, + "learning_rate": 9.406527225126767e-06, + "loss": 10.591230392456055, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.567674994468689, + "loss_total_v6": 0.723924994468689, + "step": 589 + }, + { + "epoch": 0.21990309355199403, + "grad_norm": 21.153406143188477, + "learning_rate": 9.40353427008083e-06, + "loss": 10.726644515991211, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7332523465156555, + "loss_total_v6": 0.8855960965156555, + "step": 590 + }, + { + "epoch": 0.22027581065970928, + "grad_norm": 19.659164428710938, + "learning_rate": 9.400534265708348e-06, + "loss": 10.773065567016602, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5877357721328735, + "loss_total_v6": 0.7400795221328735, + "step": 591 + }, + { + "epoch": 0.22064852776742452, + "grad_norm": 20.19511604309082, + "learning_rate": 9.397527216811827e-06, + "loss": 10.68071174621582, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4295092523097992, + "loss_total_v6": 0.5759936571121216, + "step": 592 + }, + { + "epoch": 0.22102124487513977, + "grad_norm": 21.935941696166992, + "learning_rate": 9.394513128205052e-06, + "loss": 10.438279151916504, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4032094478607178, + "loss_total_v6": 0.5653188228607178, + "step": 593 + }, + { + "epoch": 0.22139396198285502, + "grad_norm": 20.100568771362305, + "learning_rate": 9.39149200471308e-06, + "loss": 10.63038158416748, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.2544266879558563, + "loss_total_v6": 0.4028641879558563, + "step": 594 + }, + { + "epoch": 0.22176667909057027, + "grad_norm": 19.785795211791992, + "learning_rate": 9.38846385117223e-06, + "loss": 10.086865425109863, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7129361629486084, + "loss_total_v6": 0.8691861629486084, + "step": 595 + }, + { + "epoch": 0.2221393961982855, + "grad_norm": 20.141027450561523, + "learning_rate": 9.385428672430066e-06, + "loss": 10.834304809570312, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.35687553882598877, + "loss_total_v6": 0.49750053882598877, + "step": 596 + }, + { + "epoch": 0.22251211330600074, + "grad_norm": 18.861858367919922, + "learning_rate": 9.382386473345411e-06, + "loss": 10.585531234741211, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5874165296554565, + "loss_total_v6": 0.7456196546554565, + "step": 597 + }, + { + "epoch": 0.22288483041371598, + "grad_norm": 19.622140884399414, + "learning_rate": 9.379337258788316e-06, + "loss": 10.162105560302734, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.6724510788917542, + "loss_total_v6": 0.8091698288917542, + "step": 598 + }, + { + "epoch": 0.22325754752143123, + "grad_norm": 20.110816955566406, + "learning_rate": 9.376281033640068e-06, + "loss": 10.768350601196289, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.7315778136253357, + "loss_total_v6": 0.8897809386253357, + "step": 599 + }, + { + "epoch": 0.22363026462914648, + "grad_norm": 19.319704055786133, + "learning_rate": 9.373217802793177e-06, + "loss": 10.720623016357422, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6103020906448364, + "loss_total_v6": 0.6103020906448364, + "step": 600 + }, + { + "epoch": 0.22363026462914648, + "eval_loss": 0.6353172659873962, + "eval_loss_alignment": 0.2879210904113248, + "eval_loss_alignment_w": 0.1439605452056624, + "eval_loss_sft": 0.4913568019739583, + "eval_loss_total_v6": 0.6353173473547412, + "eval_runtime": 207.0594, + "eval_samples_per_second": 9.012, + "eval_steps_per_second": 1.13, + "step": 600 + }, + { + "epoch": 0.22400298173686173, + "grad_norm": 21.240846633911133, + "learning_rate": 9.370147571151364e-06, + "loss": 10.507665634155273, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3014085292816162, + "loss_total_v6": 0.4615647792816162, + "step": 601 + }, + { + "epoch": 0.22437569884457698, + "grad_norm": 20.22723960876465, + "learning_rate": 9.367070343629562e-06, + "loss": 10.430288314819336, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.5908236503601074, + "loss_total_v6": 0.7568392753601074, + "step": 602 + }, + { + "epoch": 0.2247484159522922, + "grad_norm": 21.189929962158203, + "learning_rate": 9.3639861251539e-06, + "loss": 11.002920150756836, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5403297543525696, + "loss_total_v6": 0.6829078793525696, + "step": 603 + }, + { + "epoch": 0.22512113306000744, + "grad_norm": 20.14267349243164, + "learning_rate": 9.3608949206617e-06, + "loss": 10.650136947631836, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3665887117385864, + "loss_total_v6": 0.5286980867385864, + "step": 604 + }, + { + "epoch": 0.2254938501677227, + "grad_norm": 20.897226333618164, + "learning_rate": 9.357796735101463e-06, + "loss": 11.029614448547363, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2951396405696869, + "loss_total_v6": 0.4455302655696869, + "step": 605 + }, + { + "epoch": 0.22586656727543794, + "grad_norm": 19.35726547241211, + "learning_rate": 9.354691573432875e-06, + "loss": 10.66611385345459, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.620603084564209, + "loss_total_v6": 0.778806209564209, + "step": 606 + }, + { + "epoch": 0.2262392843831532, + "grad_norm": 20.131450653076172, + "learning_rate": 9.35157944062678e-06, + "loss": 10.630762100219727, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.46490567922592163, + "loss_total_v6": 0.6172494292259216, + "step": 607 + }, + { + "epoch": 0.22661200149086844, + "grad_norm": 19.53032684326172, + "learning_rate": 9.348460341665184e-06, + "loss": 10.006502151489258, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.47331923246383667, + "loss_total_v6": 0.6295692324638367, + "step": 608 + }, + { + "epoch": 0.22698471859858368, + "grad_norm": 20.23382568359375, + "learning_rate": 9.34533428154125e-06, + "loss": 10.471797943115234, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.6467056274414062, + "loss_total_v6": 0.8342056274414062, + "step": 609 + }, + { + "epoch": 0.22735743570629893, + "grad_norm": 20.436904907226562, + "learning_rate": 9.342201265259277e-06, + "loss": 10.281278610229492, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.408018559217453, + "loss_total_v6": 0.5642685890197754, + "step": 610 + }, + { + "epoch": 0.22773015281401415, + "grad_norm": 18.983440399169922, + "learning_rate": 9.339061297834708e-06, + "loss": 10.528688430786133, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.8253102898597717, + "loss_total_v6": 0.9991384148597717, + "step": 611 + }, + { + "epoch": 0.2281028699217294, + "grad_norm": 19.65081787109375, + "learning_rate": 9.335914384294105e-06, + "loss": 10.532763481140137, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4564150273799896, + "loss_total_v6": 0.610711932182312, + "step": 612 + }, + { + "epoch": 0.22847558702944465, + "grad_norm": 18.29380226135254, + "learning_rate": 9.332760529675154e-06, + "loss": 10.244487762451172, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3953258991241455, + "loss_total_v6": 0.5554821491241455, + "step": 613 + }, + { + "epoch": 0.2288483041371599, + "grad_norm": 19.067487716674805, + "learning_rate": 9.329599739026655e-06, + "loss": 10.253257751464844, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6911758780479431, + "loss_total_v6": 0.8532852530479431, + "step": 614 + }, + { + "epoch": 0.22922102124487514, + "grad_norm": 20.538204193115234, + "learning_rate": 9.326432017408506e-06, + "loss": 10.2799654006958, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6612233519554138, + "loss_total_v6": 0.8233327269554138, + "step": 615 + }, + { + "epoch": 0.2295937383525904, + "grad_norm": 19.630783081054688, + "learning_rate": 9.323257369891702e-06, + "loss": 10.845731735229492, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5251355767250061, + "loss_total_v6": 0.6774793267250061, + "step": 616 + }, + { + "epoch": 0.22996645546030564, + "grad_norm": 20.545034408569336, + "learning_rate": 9.32007580155833e-06, + "loss": 10.389596939086914, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.42529746890068054, + "loss_total_v6": 0.5854537487030029, + "step": 617 + }, + { + "epoch": 0.23033917256802086, + "grad_norm": 19.21124839782715, + "learning_rate": 9.316887317501549e-06, + "loss": 10.659993171691895, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.674889087677002, + "loss_total_v6": 0.827232837677002, + "step": 618 + }, + { + "epoch": 0.2307118896757361, + "grad_norm": 22.94989585876465, + "learning_rate": 9.313691922825592e-06, + "loss": 10.779918670654297, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6392261981964111, + "loss_total_v6": 0.8071949481964111, + "step": 619 + }, + { + "epoch": 0.23108460678345136, + "grad_norm": 20.791790008544922, + "learning_rate": 9.310489622645756e-06, + "loss": 10.666974067687988, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5250275135040283, + "loss_total_v6": 0.6851837635040283, + "step": 620 + }, + { + "epoch": 0.2314573238911666, + "grad_norm": 20.240428924560547, + "learning_rate": 9.30728042208839e-06, + "loss": 11.131153106689453, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6383030414581299, + "loss_total_v6": 0.7925999164581299, + "step": 621 + }, + { + "epoch": 0.23183004099888185, + "grad_norm": 19.239543914794922, + "learning_rate": 9.304064326290894e-06, + "loss": 10.275995254516602, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6823310256004333, + "loss_total_v6": 0.8502997756004333, + "step": 622 + }, + { + "epoch": 0.2322027581065971, + "grad_norm": 20.148845672607422, + "learning_rate": 9.300841340401699e-06, + "loss": 10.709805488586426, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.8008734583854675, + "loss_total_v6": 0.9590765833854675, + "step": 623 + }, + { + "epoch": 0.23257547521431235, + "grad_norm": 19.052845001220703, + "learning_rate": 9.297611469580272e-06, + "loss": 10.371206283569336, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.45131778717041016, + "loss_total_v6": 0.5997552871704102, + "step": 624 + }, + { + "epoch": 0.23294819232202757, + "grad_norm": 19.354976654052734, + "learning_rate": 9.2943747189971e-06, + "loss": 10.432642936706543, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6746180057525635, + "loss_total_v6": 0.8347742557525635, + "step": 625 + }, + { + "epoch": 0.23332090942974282, + "grad_norm": 18.803495407104492, + "learning_rate": 9.291131093833678e-06, + "loss": 10.782578468322754, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.48696771264076233, + "loss_total_v6": 0.6510301828384399, + "step": 626 + }, + { + "epoch": 0.23369362653745807, + "grad_norm": 18.586496353149414, + "learning_rate": 9.287880599282517e-06, + "loss": 9.95150089263916, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4564753472805023, + "loss_total_v6": 0.6107722520828247, + "step": 627 + }, + { + "epoch": 0.2340663436451733, + "grad_norm": 21.120159149169922, + "learning_rate": 9.284623240547117e-06, + "loss": 10.457893371582031, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7870796322822571, + "loss_total_v6": 0.7870796322822571, + "step": 628 + }, + { + "epoch": 0.23443906075288856, + "grad_norm": 19.343856811523438, + "learning_rate": 9.281359022841966e-06, + "loss": 10.54143238067627, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.8346289992332458, + "loss_total_v6": 0.9928321242332458, + "step": 629 + }, + { + "epoch": 0.2348117778606038, + "grad_norm": 18.96506118774414, + "learning_rate": 9.278087951392536e-06, + "loss": 10.6455078125, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5708301663398743, + "loss_total_v6": 0.7192676663398743, + "step": 630 + }, + { + "epoch": 0.23518449496831906, + "grad_norm": 18.580018997192383, + "learning_rate": 9.274810031435267e-06, + "loss": 10.627593994140625, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5217687487602234, + "loss_total_v6": 0.6682531237602234, + "step": 631 + }, + { + "epoch": 0.23555721207603428, + "grad_norm": 18.443326950073242, + "learning_rate": 9.271525268217564e-06, + "loss": 10.762475967407227, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.43833959102630615, + "loss_total_v6": 0.6199802160263062, + "step": 632 + }, + { + "epoch": 0.23592992918374953, + "grad_norm": 19.89031410217285, + "learning_rate": 9.268233666997789e-06, + "loss": 10.288780212402344, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6154573559761047, + "loss_total_v6": 0.7717073559761047, + "step": 633 + }, + { + "epoch": 0.23630264629146477, + "grad_norm": 19.651222229003906, + "learning_rate": 9.264935233045248e-06, + "loss": 10.361771583557129, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6920868754386902, + "loss_total_v6": 0.8541962504386902, + "step": 634 + }, + { + "epoch": 0.23667536339918002, + "grad_norm": 20.706462860107422, + "learning_rate": 9.261629971640185e-06, + "loss": 10.353874206542969, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.46292728185653687, + "loss_total_v6": 0.6172241568565369, + "step": 635 + }, + { + "epoch": 0.23704808050689527, + "grad_norm": 21.036975860595703, + "learning_rate": 9.258317888073773e-06, + "loss": 10.787607192993164, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5010190606117249, + "loss_total_v6": 0.6611753106117249, + "step": 636 + }, + { + "epoch": 0.23742079761461052, + "grad_norm": 19.373201370239258, + "learning_rate": 9.25499898764811e-06, + "loss": 11.042759895324707, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5754315853118896, + "loss_total_v6": 0.7258222103118896, + "step": 637 + }, + { + "epoch": 0.23779351472232577, + "grad_norm": 18.94626235961914, + "learning_rate": 9.251673275676203e-06, + "loss": 10.62346076965332, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.5232194662094116, + "loss_total_v6": 0.6618913412094116, + "step": 638 + }, + { + "epoch": 0.238166231830041, + "grad_norm": 19.897489547729492, + "learning_rate": 9.248340757481965e-06, + "loss": 10.592296600341797, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3896980583667755, + "loss_total_v6": 0.5400886535644531, + "step": 639 + }, + { + "epoch": 0.23853894893775623, + "grad_norm": 19.982030868530273, + "learning_rate": 9.2450014384002e-06, + "loss": 10.160107612609863, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.1676846593618393, + "loss_total_v6": 0.3239346742630005, + "step": 640 + }, + { + "epoch": 0.23891166604547148, + "grad_norm": 19.576236724853516, + "learning_rate": 9.241655323776606e-06, + "loss": 10.446675300598145, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6739594340324402, + "loss_total_v6": 0.8282563090324402, + "step": 641 + }, + { + "epoch": 0.23928438315318673, + "grad_norm": 19.705259323120117, + "learning_rate": 9.238302418967757e-06, + "loss": 10.502681732177734, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.40674105286598206, + "loss_total_v6": 0.5590847730636597, + "step": 642 + }, + { + "epoch": 0.23965710026090198, + "grad_norm": 20.39151954650879, + "learning_rate": 9.234942729341096e-06, + "loss": 10.293344497680664, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.42224764823913574, + "loss_total_v6": 0.5726382732391357, + "step": 643 + }, + { + "epoch": 0.24002981736861723, + "grad_norm": 22.549060821533203, + "learning_rate": 9.231576260274926e-06, + "loss": 10.66135311126709, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4434126019477844, + "loss_total_v6": 0.5957563519477844, + "step": 644 + }, + { + "epoch": 0.24040253447633247, + "grad_norm": 19.760133743286133, + "learning_rate": 9.228203017158406e-06, + "loss": 11.100351333618164, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5310654640197754, + "loss_total_v6": 0.6775498390197754, + "step": 645 + }, + { + "epoch": 0.2407752515840477, + "grad_norm": 19.2553653717041, + "learning_rate": 9.224823005391537e-06, + "loss": 10.804352760314941, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5913333892822266, + "loss_total_v6": 0.7553958892822266, + "step": 646 + }, + { + "epoch": 0.24114796869176294, + "grad_norm": 20.343963623046875, + "learning_rate": 9.221436230385157e-06, + "loss": 10.323196411132812, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 1.0339596271514893, + "loss_total_v6": 1.1843502521514893, + "step": 647 + }, + { + "epoch": 0.2415206857994782, + "grad_norm": 18.922094345092773, + "learning_rate": 9.218042697560928e-06, + "loss": 10.51635456085205, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7454928159713745, + "loss_total_v6": 0.8978365659713745, + "step": 648 + }, + { + "epoch": 0.24189340290719344, + "grad_norm": 20.165876388549805, + "learning_rate": 9.214642412351334e-06, + "loss": 10.699472427368164, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.259490966796875, + "loss_total_v6": 0.425506591796875, + "step": 649 + }, + { + "epoch": 0.2422661200149087, + "grad_norm": 20.094633102416992, + "learning_rate": 9.211235380199666e-06, + "loss": 10.405843734741211, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4498676061630249, + "loss_total_v6": 0.6080707311630249, + "step": 650 + }, + { + "epoch": 0.24263883712262393, + "grad_norm": 19.947031021118164, + "learning_rate": 9.207821606560017e-06, + "loss": 10.4458589553833, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5345615148544312, + "loss_total_v6": 0.6771396398544312, + "step": 651 + }, + { + "epoch": 0.24301155423033918, + "grad_norm": 19.480052947998047, + "learning_rate": 9.204401096897271e-06, + "loss": 10.736316680908203, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6146059632301331, + "loss_total_v6": 0.7689028382301331, + "step": 652 + }, + { + "epoch": 0.2433842713380544, + "grad_norm": 20.005462646484375, + "learning_rate": 9.200973856687093e-06, + "loss": 9.994514465332031, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.3411007523536682, + "loss_total_v6": 0.5071163773536682, + "step": 653 + }, + { + "epoch": 0.24375698844576965, + "grad_norm": 21.644329071044922, + "learning_rate": 9.197539891415928e-06, + "loss": 10.768509864807129, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.20343592762947083, + "loss_total_v6": 0.20343592762947083, + "step": 654 + }, + { + "epoch": 0.2441297055534849, + "grad_norm": 22.79046630859375, + "learning_rate": 9.194099206580981e-06, + "loss": 10.38597583770752, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.8125951886177063, + "loss_total_v6": 0.8125951886177063, + "step": 655 + }, + { + "epoch": 0.24450242266120015, + "grad_norm": 19.138154983520508, + "learning_rate": 9.190651807690221e-06, + "loss": 10.320751190185547, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.22114066779613495, + "loss_total_v6": 0.38129693269729614, + "step": 656 + }, + { + "epoch": 0.2448751397689154, + "grad_norm": 20.627506256103516, + "learning_rate": 9.187197700262359e-06, + "loss": 9.742643356323242, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.49969708919525146, + "loss_total_v6": 0.6559470891952515, + "step": 657 + }, + { + "epoch": 0.24524785687663064, + "grad_norm": 20.264020919799805, + "learning_rate": 9.183736889826842e-06, + "loss": 10.636207580566406, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5310145616531372, + "loss_total_v6": 0.6774989366531372, + "step": 658 + }, + { + "epoch": 0.2456205739843459, + "grad_norm": 19.987773895263672, + "learning_rate": 9.180269381923859e-06, + "loss": 10.687931060791016, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5134900808334351, + "loss_total_v6": 0.6716932058334351, + "step": 659 + }, + { + "epoch": 0.24599329109206114, + "grad_norm": 19.64288902282715, + "learning_rate": 9.17679518210431e-06, + "loss": 10.308930397033691, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.8743200898170471, + "loss_total_v6": 1.0325231552124023, + "step": 660 + }, + { + "epoch": 0.24636600819977636, + "grad_norm": 20.729225158691406, + "learning_rate": 9.173314295929814e-06, + "loss": 10.662651062011719, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4111422300338745, + "loss_total_v6": 0.5595797300338745, + "step": 661 + }, + { + "epoch": 0.2467387253074916, + "grad_norm": 19.120134353637695, + "learning_rate": 9.169826728972688e-06, + "loss": 10.568450927734375, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.545928955078125, + "loss_total_v6": 0.700225830078125, + "step": 662 + }, + { + "epoch": 0.24711144241520686, + "grad_norm": 56.14055252075195, + "learning_rate": 9.166332486815952e-06, + "loss": 10.773563385009766, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6321728229522705, + "loss_total_v6": 0.7845165729522705, + "step": 663 + }, + { + "epoch": 0.2474841595229221, + "grad_norm": 21.183168411254883, + "learning_rate": 9.162831575053302e-06, + "loss": 10.64750862121582, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.23880775272846222, + "loss_total_v6": 0.389198362827301, + "step": 664 + }, + { + "epoch": 0.24785687663063735, + "grad_norm": 20.14591407775879, + "learning_rate": 9.159323999289119e-06, + "loss": 10.594017028808594, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5044064521789551, + "loss_total_v6": 0.6547970771789551, + "step": 665 + }, + { + "epoch": 0.2482295937383526, + "grad_norm": 19.80781364440918, + "learning_rate": 9.155809765138447e-06, + "loss": 9.826844215393066, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3003200590610504, + "loss_total_v6": 0.4468044340610504, + "step": 666 + }, + { + "epoch": 0.24860231084606785, + "grad_norm": 20.44622802734375, + "learning_rate": 9.152288878226991e-06, + "loss": 10.75230598449707, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.46147531270980835, + "loss_total_v6": 0.6079596877098083, + "step": 667 + }, + { + "epoch": 0.24897502795378307, + "grad_norm": 20.66183090209961, + "learning_rate": 9.14876134419111e-06, + "loss": 10.37263011932373, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5687063336372375, + "loss_total_v6": 0.7249563336372375, + "step": 668 + }, + { + "epoch": 0.24934774506149832, + "grad_norm": 19.285188674926758, + "learning_rate": 9.145227168677794e-06, + "loss": 10.549962997436523, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6758520603179932, + "loss_total_v6": 0.8340551853179932, + "step": 669 + }, + { + "epoch": 0.24972046216921356, + "grad_norm": 19.709461212158203, + "learning_rate": 9.141686357344672e-06, + "loss": 10.256462097167969, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.240813210606575, + "loss_total_v6": 0.3912038207054138, + "step": 670 + }, + { + "epoch": 0.2500931792769288, + "grad_norm": 20.072059631347656, + "learning_rate": 9.138138915859997e-06, + "loss": 10.979692459106445, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.733975350856781, + "loss_total_v6": 0.894131600856781, + "step": 671 + }, + { + "epoch": 0.25046589638464406, + "grad_norm": 21.27431297302246, + "learning_rate": 9.134584849902632e-06, + "loss": 10.843225479125977, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6569486260414124, + "loss_total_v6": 0.8210111260414124, + "step": 672 + }, + { + "epoch": 0.2508386134923593, + "grad_norm": 20.121143341064453, + "learning_rate": 9.131024165162047e-06, + "loss": 10.76073169708252, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.46517476439476013, + "loss_total_v6": 0.6214247941970825, + "step": 673 + }, + { + "epoch": 0.25121133060007456, + "grad_norm": 20.240358352661133, + "learning_rate": 9.127456867338302e-06, + "loss": 10.50996208190918, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.24168513715267181, + "loss_total_v6": 0.394028902053833, + "step": 674 + }, + { + "epoch": 0.2515840477077898, + "grad_norm": 18.68674659729004, + "learning_rate": 9.123882962142051e-06, + "loss": 10.371700286865234, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.646003782749176, + "loss_total_v6": 0.792488157749176, + "step": 675 + }, + { + "epoch": 0.25195676481550505, + "grad_norm": 20.69846534729004, + "learning_rate": 9.120302455294526e-06, + "loss": 11.06518840789795, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.6077491044998169, + "loss_total_v6": 0.7522803544998169, + "step": 676 + }, + { + "epoch": 0.2523294819232203, + "grad_norm": 18.959030151367188, + "learning_rate": 9.11671535252752e-06, + "loss": 10.643938064575195, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6191727519035339, + "loss_total_v6": 0.6191727519035339, + "step": 677 + }, + { + "epoch": 0.2527021990309355, + "grad_norm": 19.680055618286133, + "learning_rate": 9.113121659583385e-06, + "loss": 10.119819641113281, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7112913727760315, + "loss_total_v6": 0.8538694977760315, + "step": 678 + }, + { + "epoch": 0.25307491613865074, + "grad_norm": 19.476703643798828, + "learning_rate": 9.10952138221503e-06, + "loss": 10.745797157287598, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5070182681083679, + "loss_total_v6": 0.6574088931083679, + "step": 679 + }, + { + "epoch": 0.253447633246366, + "grad_norm": 20.4124755859375, + "learning_rate": 9.105914526185902e-06, + "loss": 10.209053039550781, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.16466620564460754, + "loss_total_v6": 0.32482245564460754, + "step": 680 + }, + { + "epoch": 0.25382035035408124, + "grad_norm": 18.911951065063477, + "learning_rate": 9.102301097269974e-06, + "loss": 10.817229270935059, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.48770925402641296, + "loss_total_v6": 0.6341936588287354, + "step": 681 + }, + { + "epoch": 0.2541930674617965, + "grad_norm": 18.22420310974121, + "learning_rate": 9.09868110125175e-06, + "loss": 10.556461334228516, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.524662435054779, + "loss_total_v6": 0.678959310054779, + "step": 682 + }, + { + "epoch": 0.25456578456951173, + "grad_norm": 21.748533248901367, + "learning_rate": 9.095054543926238e-06, + "loss": 10.450334548950195, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4226625859737396, + "loss_total_v6": 0.4226625859737396, + "step": 683 + }, + { + "epoch": 0.254938501677227, + "grad_norm": 19.55162239074707, + "learning_rate": 9.091421431098955e-06, + "loss": 10.478399276733398, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3961023986339569, + "loss_total_v6": 0.5425868034362793, + "step": 684 + }, + { + "epoch": 0.25531121878494223, + "grad_norm": 19.120540618896484, + "learning_rate": 9.087781768585913e-06, + "loss": 10.596155166625977, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6247889399528503, + "loss_total_v6": 0.7790858149528503, + "step": 685 + }, + { + "epoch": 0.2556839358926575, + "grad_norm": 19.24106216430664, + "learning_rate": 9.084135562213603e-06, + "loss": 10.230325698852539, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.47966063022613525, + "loss_total_v6": 0.6417700052261353, + "step": 686 + }, + { + "epoch": 0.2560566530003727, + "grad_norm": 19.244726181030273, + "learning_rate": 9.080482817818996e-06, + "loss": 10.424449920654297, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.34422028064727783, + "loss_total_v6": 0.5121890306472778, + "step": 687 + }, + { + "epoch": 0.256429370108088, + "grad_norm": 18.361942291259766, + "learning_rate": 9.076823541249532e-06, + "loss": 10.133331298828125, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6468648910522461, + "loss_total_v6": 0.6468648910522461, + "step": 688 + }, + { + "epoch": 0.2568020872158032, + "grad_norm": 20.022974014282227, + "learning_rate": 9.073157738363101e-06, + "loss": 10.345832824707031, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5227341651916504, + "loss_total_v6": 0.6867966651916504, + "step": 689 + }, + { + "epoch": 0.25717480432351847, + "grad_norm": 19.056072235107422, + "learning_rate": 9.069485415028047e-06, + "loss": 10.69914436340332, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.48964324593544006, + "loss_total_v6": 0.6517525911331177, + "step": 690 + }, + { + "epoch": 0.2575475214312337, + "grad_norm": 20.376310348510742, + "learning_rate": 9.065806577123148e-06, + "loss": 10.840707778930664, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.33512479066848755, + "loss_total_v6": 0.48356229066848755, + "step": 691 + }, + { + "epoch": 0.25792023853894896, + "grad_norm": 18.342422485351562, + "learning_rate": 9.062121230537611e-06, + "loss": 10.598602294921875, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6466180086135864, + "loss_total_v6": 0.7891961336135864, + "step": 692 + }, + { + "epoch": 0.25829295564666416, + "grad_norm": 19.828277587890625, + "learning_rate": 9.058429381171066e-06, + "loss": 9.964676856994629, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.36256545782089233, + "loss_total_v6": 0.5090498328208923, + "step": 693 + }, + { + "epoch": 0.2586656727543794, + "grad_norm": 18.79833984375, + "learning_rate": 9.05473103493355e-06, + "loss": 10.656782150268555, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.634590744972229, + "loss_total_v6": 0.790840744972229, + "step": 694 + }, + { + "epoch": 0.25903838986209465, + "grad_norm": 20.03409767150879, + "learning_rate": 9.051026197745498e-06, + "loss": 10.679925918579102, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5900750160217285, + "loss_total_v6": 0.7482781410217285, + "step": 695 + }, + { + "epoch": 0.2594111069698099, + "grad_norm": 18.1480712890625, + "learning_rate": 9.04731487553774e-06, + "loss": 10.252643585205078, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8501763343811035, + "loss_total_v6": 1.0044732093811035, + "step": 696 + }, + { + "epoch": 0.25978382407752515, + "grad_norm": 19.692626953125, + "learning_rate": 9.043597074251487e-06, + "loss": 10.62895393371582, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.22322534024715424, + "loss_total_v6": 0.36775660514831543, + "step": 697 + }, + { + "epoch": 0.2601565411852404, + "grad_norm": 19.260860443115234, + "learning_rate": 9.03987279983832e-06, + "loss": 10.497167587280273, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5779924392700195, + "loss_total_v6": 0.7283830642700195, + "step": 698 + }, + { + "epoch": 0.26052925829295565, + "grad_norm": 20.443450927734375, + "learning_rate": 9.036142058260183e-06, + "loss": 10.572010040283203, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4333077669143677, + "loss_total_v6": 0.5856515169143677, + "step": 699 + }, + { + "epoch": 0.2609019754006709, + "grad_norm": 20.075786590576172, + "learning_rate": 9.032404855489375e-06, + "loss": 10.69954776763916, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4806779623031616, + "loss_total_v6": 0.6369279623031616, + "step": 700 + }, + { + "epoch": 0.2609019754006709, + "eval_loss": 0.6320605278015137, + "eval_loss_alignment": 0.28671082064636755, + "eval_loss_alignment_w": 0.14335541032318377, + "eval_loss_sft": 0.48870511984245646, + "eval_loss_total_v6": 0.6320605302054403, + "eval_runtime": 206.957, + "eval_samples_per_second": 9.016, + "eval_steps_per_second": 1.131, + "step": 700 + }, + { + "epoch": 0.26127469250838614, + "grad_norm": 18.875110626220703, + "learning_rate": 9.028661197508536e-06, + "loss": 10.222972869873047, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5459979772567749, + "loss_total_v6": 0.6983417272567749, + "step": 701 + }, + { + "epoch": 0.2616474096161014, + "grad_norm": 18.350387573242188, + "learning_rate": 9.024911090310638e-06, + "loss": 10.657750129699707, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.2771456837654114, + "loss_total_v6": 0.4236300587654114, + "step": 702 + }, + { + "epoch": 0.26202012672381664, + "grad_norm": 20.088823318481445, + "learning_rate": 9.021154539898984e-06, + "loss": 10.03676986694336, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.677615225315094, + "loss_total_v6": 0.841677725315094, + "step": 703 + }, + { + "epoch": 0.2623928438315319, + "grad_norm": 18.040287017822266, + "learning_rate": 9.017391552287182e-06, + "loss": 10.653461456298828, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7615103721618652, + "loss_total_v6": 0.9177603721618652, + "step": 704 + }, + { + "epoch": 0.26276556093924713, + "grad_norm": 19.733901977539062, + "learning_rate": 9.013622133499153e-06, + "loss": 10.416589736938477, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5214716792106628, + "loss_total_v6": 0.6777216792106628, + "step": 705 + }, + { + "epoch": 0.2631382780469624, + "grad_norm": 18.37191390991211, + "learning_rate": 9.009846289569108e-06, + "loss": 10.42918586730957, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5436506867408752, + "loss_total_v6": 0.7018538117408752, + "step": 706 + }, + { + "epoch": 0.2635109951546776, + "grad_norm": 19.1311092376709, + "learning_rate": 9.006064026541549e-06, + "loss": 10.753580093383789, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4298804700374603, + "loss_total_v6": 0.5822242498397827, + "step": 707 + }, + { + "epoch": 0.2638837122623928, + "grad_norm": 18.30912971496582, + "learning_rate": 9.002275350471244e-06, + "loss": 10.667699813842773, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5084232687950134, + "loss_total_v6": 0.6549076437950134, + "step": 708 + }, + { + "epoch": 0.26425642937010807, + "grad_norm": 18.259960174560547, + "learning_rate": 8.99848026742324e-06, + "loss": 10.568180084228516, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5036157965660095, + "loss_total_v6": 0.6598657965660095, + "step": 709 + }, + { + "epoch": 0.2646291464778233, + "grad_norm": 18.20842170715332, + "learning_rate": 8.994678783472834e-06, + "loss": 10.593017578125, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4496549069881439, + "loss_total_v6": 0.5980924367904663, + "step": 710 + }, + { + "epoch": 0.26500186358553857, + "grad_norm": 18.301265716552734, + "learning_rate": 8.990870904705566e-06, + "loss": 10.406161308288574, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4309422969818115, + "loss_total_v6": 0.5891454219818115, + "step": 711 + }, + { + "epoch": 0.2653745806932538, + "grad_norm": 18.397485733032227, + "learning_rate": 8.98705663721722e-06, + "loss": 10.260213851928711, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.47933539748191833, + "loss_total_v6": 0.6277729272842407, + "step": 712 + }, + { + "epoch": 0.26574729780096906, + "grad_norm": 19.807708740234375, + "learning_rate": 8.983235987113802e-06, + "loss": 10.181990623474121, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4033496379852295, + "loss_total_v6": 0.5498340129852295, + "step": 713 + }, + { + "epoch": 0.2661200149086843, + "grad_norm": 18.70478630065918, + "learning_rate": 8.979408960511541e-06, + "loss": 10.607416152954102, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6365731954574585, + "loss_total_v6": 0.7947763204574585, + "step": 714 + }, + { + "epoch": 0.26649273201639956, + "grad_norm": 20.61451530456543, + "learning_rate": 8.97557556353687e-06, + "loss": 10.393577575683594, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.7311753630638123, + "loss_total_v6": 0.8718003630638123, + "step": 715 + }, + { + "epoch": 0.2668654491241148, + "grad_norm": 19.011337280273438, + "learning_rate": 8.971735802326418e-06, + "loss": 10.413066864013672, + "loss_alignment": 0.26953125, + "loss_alignment_w": 0.134765625, + "loss_sft": 0.32646286487579346, + "loss_total_v6": 0.46122848987579346, + "step": 716 + }, + { + "epoch": 0.26723816623183005, + "grad_norm": 19.86182403564453, + "learning_rate": 8.967889683027009e-06, + "loss": 10.21196460723877, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2932618260383606, + "loss_total_v6": 0.4456055760383606, + "step": 717 + }, + { + "epoch": 0.2676108833395453, + "grad_norm": 18.954906463623047, + "learning_rate": 8.964037211795634e-06, + "loss": 10.931074142456055, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.47468045353889465, + "loss_total_v6": 0.6348366737365723, + "step": 718 + }, + { + "epoch": 0.26798360044726055, + "grad_norm": 19.35100555419922, + "learning_rate": 8.960178394799464e-06, + "loss": 10.81981086730957, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6291767954826355, + "loss_total_v6": 0.7912861704826355, + "step": 719 + }, + { + "epoch": 0.2683563175549758, + "grad_norm": 19.290502548217773, + "learning_rate": 8.956313238215824e-06, + "loss": 10.220743179321289, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4130662977695465, + "loss_total_v6": 0.5615037679672241, + "step": 720 + }, + { + "epoch": 0.268729034662691, + "grad_norm": 21.395687103271484, + "learning_rate": 8.952441748232186e-06, + "loss": 10.488340377807617, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6873489022254944, + "loss_total_v6": 0.8416457772254944, + "step": 721 + }, + { + "epoch": 0.26910175177040624, + "grad_norm": 18.25525665283203, + "learning_rate": 8.948563931046162e-06, + "loss": 10.460630416870117, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5099238157272339, + "loss_total_v6": 0.6583613157272339, + "step": 722 + }, + { + "epoch": 0.2694744688781215, + "grad_norm": 19.863489151000977, + "learning_rate": 8.944679792865494e-06, + "loss": 10.629706382751465, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6105562448501587, + "loss_total_v6": 0.7648531198501587, + "step": 723 + }, + { + "epoch": 0.26984718598583673, + "grad_norm": 19.189823150634766, + "learning_rate": 8.940789339908041e-06, + "loss": 10.047163009643555, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.400640606880188, + "loss_total_v6": 0.556890606880188, + "step": 724 + }, + { + "epoch": 0.270219903093552, + "grad_norm": 22.107078552246094, + "learning_rate": 8.93689257840177e-06, + "loss": 10.214731216430664, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.27317070960998535, + "loss_total_v6": 0.41965508460998535, + "step": 725 + }, + { + "epoch": 0.27059262020126723, + "grad_norm": 18.74631118774414, + "learning_rate": 8.93298951458475e-06, + "loss": 10.890793800354004, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7378838062286377, + "loss_total_v6": 0.8980400562286377, + "step": 726 + }, + { + "epoch": 0.2709653373089825, + "grad_norm": 19.1406307220459, + "learning_rate": 8.92908015470514e-06, + "loss": 10.75726318359375, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.548944890499115, + "loss_total_v6": 0.714960515499115, + "step": 727 + }, + { + "epoch": 0.2713380544166977, + "grad_norm": 20.22531509399414, + "learning_rate": 8.92516450502117e-06, + "loss": 10.829019546508789, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.39061057567596436, + "loss_total_v6": 0.5390480756759644, + "step": 728 + }, + { + "epoch": 0.271710771524413, + "grad_norm": 20.09929847717285, + "learning_rate": 8.92124257180115e-06, + "loss": 10.281632423400879, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.26280635595321655, + "loss_total_v6": 0.41905635595321655, + "step": 729 + }, + { + "epoch": 0.2720834886321282, + "grad_norm": 20.021244049072266, + "learning_rate": 8.917314361323439e-06, + "loss": 10.405336380004883, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.198362335562706, + "loss_total_v6": 0.3565654754638672, + "step": 730 + }, + { + "epoch": 0.27245620573984347, + "grad_norm": 18.704315185546875, + "learning_rate": 8.913379879876451e-06, + "loss": 10.344457626342773, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5536390542984009, + "loss_total_v6": 0.7059828042984009, + "step": 731 + }, + { + "epoch": 0.2728289228475587, + "grad_norm": 21.420549392700195, + "learning_rate": 8.909439133758637e-06, + "loss": 10.30716609954834, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5097173452377319, + "loss_total_v6": 0.6698735952377319, + "step": 732 + }, + { + "epoch": 0.27320163995527397, + "grad_norm": 19.155075073242188, + "learning_rate": 8.905492129278478e-06, + "loss": 10.454119682312012, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.486240953207016, + "loss_total_v6": 0.6483503580093384, + "step": 733 + }, + { + "epoch": 0.2735743570629892, + "grad_norm": 20.479965209960938, + "learning_rate": 8.90153887275447e-06, + "loss": 10.338595390319824, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5091615915298462, + "loss_total_v6": 0.6615053415298462, + "step": 734 + }, + { + "epoch": 0.2739470741707044, + "grad_norm": 19.741069793701172, + "learning_rate": 8.897579370515119e-06, + "loss": 10.288107872009277, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5820184946060181, + "loss_total_v6": 0.7304559946060181, + "step": 735 + }, + { + "epoch": 0.27431979127841966, + "grad_norm": 19.29381561279297, + "learning_rate": 8.893613628898934e-06, + "loss": 10.282466888427734, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5874691009521484, + "loss_total_v6": 0.7398128509521484, + "step": 736 + }, + { + "epoch": 0.2746925083861349, + "grad_norm": 18.731613159179688, + "learning_rate": 8.889641654254404e-06, + "loss": 10.033449172973633, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4713357985019684, + "loss_total_v6": 0.629538893699646, + "step": 737 + }, + { + "epoch": 0.27506522549385015, + "grad_norm": 22.134431838989258, + "learning_rate": 8.88566345294e-06, + "loss": 10.243110656738281, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.606642484664917, + "loss_total_v6": 0.760939359664917, + "step": 738 + }, + { + "epoch": 0.2754379426015654, + "grad_norm": 22.045568466186523, + "learning_rate": 8.881679031324167e-06, + "loss": 10.518436431884766, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.1804424375295639, + "loss_total_v6": 0.3269268274307251, + "step": 739 + }, + { + "epoch": 0.27581065970928065, + "grad_norm": 20.5867977142334, + "learning_rate": 8.877688395785296e-06, + "loss": 10.197607040405273, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6410340070724487, + "loss_total_v6": 0.7972840070724487, + "step": 740 + }, + { + "epoch": 0.2761833768169959, + "grad_norm": 18.394622802734375, + "learning_rate": 8.873691552711734e-06, + "loss": 10.715818405151367, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2128695398569107, + "loss_total_v6": 0.3632601499557495, + "step": 741 + }, + { + "epoch": 0.27655609392471114, + "grad_norm": 21.111539840698242, + "learning_rate": 8.869688508501762e-06, + "loss": 10.683889389038086, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5050843358039856, + "loss_total_v6": 0.6632874608039856, + "step": 742 + }, + { + "epoch": 0.2769288110324264, + "grad_norm": 25.012685775756836, + "learning_rate": 8.865679269563588e-06, + "loss": 10.67688274383545, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.46949225664138794, + "loss_total_v6": 0.6140235066413879, + "step": 743 + }, + { + "epoch": 0.27730152814014164, + "grad_norm": 21.447998046875, + "learning_rate": 8.861663842315339e-06, + "loss": 10.461769104003906, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8865519165992737, + "loss_total_v6": 1.040848731994629, + "step": 744 + }, + { + "epoch": 0.2776742452478569, + "grad_norm": 18.60648536682129, + "learning_rate": 8.857642233185043e-06, + "loss": 10.33150863647461, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6079779863357544, + "loss_total_v6": 0.7544623613357544, + "step": 745 + }, + { + "epoch": 0.27804696235557214, + "grad_norm": 19.989389419555664, + "learning_rate": 8.85361444861063e-06, + "loss": 10.473348617553711, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.32316210865974426, + "loss_total_v6": 0.47355273365974426, + "step": 746 + }, + { + "epoch": 0.2784196794632874, + "grad_norm": 19.55875015258789, + "learning_rate": 8.849580495039916e-06, + "loss": 10.531733512878418, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5837666988372803, + "loss_total_v6": 0.7400166988372803, + "step": 747 + }, + { + "epoch": 0.27879239657100263, + "grad_norm": 18.913801193237305, + "learning_rate": 8.845540378930589e-06, + "loss": 10.757987976074219, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.536030113697052, + "loss_total_v6": 0.700092613697052, + "step": 748 + }, + { + "epoch": 0.2791651136787179, + "grad_norm": 18.33855628967285, + "learning_rate": 8.8414941067502e-06, + "loss": 10.240854263305664, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.4971354901790619, + "loss_total_v6": 0.6358073949813843, + "step": 749 + }, + { + "epoch": 0.2795378307864331, + "grad_norm": 19.267454147338867, + "learning_rate": 8.837441684976163e-06, + "loss": 10.701611518859863, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.4695560336112976, + "loss_total_v6": 0.6297122836112976, + "step": 750 + }, + { + "epoch": 0.2799105478941483, + "grad_norm": 19.274009704589844, + "learning_rate": 8.833383120095729e-06, + "loss": 9.768448829650879, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3742867708206177, + "loss_total_v6": 0.5344430208206177, + "step": 751 + }, + { + "epoch": 0.28028326500186357, + "grad_norm": 19.60342788696289, + "learning_rate": 8.829318418605986e-06, + "loss": 10.514190673828125, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.38499191403388977, + "loss_total_v6": 0.5392887592315674, + "step": 752 + }, + { + "epoch": 0.2806559821095788, + "grad_norm": 19.151430130004883, + "learning_rate": 8.825247587013846e-06, + "loss": 10.651098251342773, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.7635512948036194, + "loss_total_v6": 0.9334731698036194, + "step": 753 + }, + { + "epoch": 0.28102869921729406, + "grad_norm": 19.47800064086914, + "learning_rate": 8.821170631836035e-06, + "loss": 10.46371841430664, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6880778074264526, + "loss_total_v6": 0.8443278074264526, + "step": 754 + }, + { + "epoch": 0.2814014163250093, + "grad_norm": 19.318742752075195, + "learning_rate": 8.81708755959908e-06, + "loss": 10.629312515258789, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.4513950049877167, + "loss_total_v6": 0.6174106597900391, + "step": 755 + }, + { + "epoch": 0.28177413343272456, + "grad_norm": 20.70475959777832, + "learning_rate": 8.8129983768393e-06, + "loss": 10.033051490783691, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.47929662466049194, + "loss_total_v6": 0.6296872496604919, + "step": 756 + }, + { + "epoch": 0.2821468505404398, + "grad_norm": 20.598594665527344, + "learning_rate": 8.808903090102799e-06, + "loss": 10.552090644836426, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.3245709240436554, + "loss_total_v6": 0.4788677990436554, + "step": 757 + }, + { + "epoch": 0.28251956764815506, + "grad_norm": 17.74674415588379, + "learning_rate": 8.804801705945448e-06, + "loss": 10.461275100708008, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2692456841468811, + "loss_total_v6": 0.4215894341468811, + "step": 758 + }, + { + "epoch": 0.2828922847558703, + "grad_norm": 18.340423583984375, + "learning_rate": 8.800694230932885e-06, + "loss": 10.39638614654541, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6436030268669128, + "loss_total_v6": 0.7978999018669128, + "step": 759 + }, + { + "epoch": 0.28326500186358555, + "grad_norm": 18.29485511779785, + "learning_rate": 8.79658067164049e-06, + "loss": 10.748709678649902, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4965032935142517, + "loss_total_v6": 0.6468939185142517, + "step": 760 + }, + { + "epoch": 0.2836377189713008, + "grad_norm": 19.21091079711914, + "learning_rate": 8.792461034653393e-06, + "loss": 10.771799087524414, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6585721373558044, + "loss_total_v6": 0.8187283873558044, + "step": 761 + }, + { + "epoch": 0.28401043607901605, + "grad_norm": 20.22444725036621, + "learning_rate": 8.788335326566444e-06, + "loss": 10.471339225769043, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.41760680079460144, + "loss_total_v6": 0.5719037055969238, + "step": 762 + }, + { + "epoch": 0.2843831531867313, + "grad_norm": 20.07430648803711, + "learning_rate": 8.784203553984216e-06, + "loss": 10.327651023864746, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.29865074157714844, + "loss_total_v6": 0.44904136657714844, + "step": 763 + }, + { + "epoch": 0.2847558702944465, + "grad_norm": 19.68398094177246, + "learning_rate": 8.780065723520995e-06, + "loss": 10.29498291015625, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5356948971748352, + "loss_total_v6": 0.6841323971748352, + "step": 764 + }, + { + "epoch": 0.28512858740216174, + "grad_norm": 20.544578552246094, + "learning_rate": 8.775921841800756e-06, + "loss": 10.953461647033691, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4421679973602295, + "loss_total_v6": 0.5906054973602295, + "step": 765 + }, + { + "epoch": 0.285501304509877, + "grad_norm": 18.99736976623535, + "learning_rate": 8.771771915457165e-06, + "loss": 10.314193725585938, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7220743894577026, + "loss_total_v6": 0.8783243894577026, + "step": 766 + }, + { + "epoch": 0.28587402161759223, + "grad_norm": 20.421031951904297, + "learning_rate": 8.767615951133566e-06, + "loss": 10.574945449829102, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6384841203689575, + "loss_total_v6": 0.7986403703689575, + "step": 767 + }, + { + "epoch": 0.2862467387253075, + "grad_norm": 19.463449478149414, + "learning_rate": 8.763453955482966e-06, + "loss": 10.780710220336914, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.4457699954509735, + "loss_total_v6": 0.5844419002532959, + "step": 768 + }, + { + "epoch": 0.28661945583302273, + "grad_norm": 18.331295013427734, + "learning_rate": 8.75928593516803e-06, + "loss": 10.759777069091797, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.24064943194389343, + "loss_total_v6": 0.40080568194389343, + "step": 769 + }, + { + "epoch": 0.286992172940738, + "grad_norm": 19.14393424987793, + "learning_rate": 8.755111896861062e-06, + "loss": 10.50533676147461, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4701438546180725, + "loss_total_v6": 0.6205344796180725, + "step": 770 + }, + { + "epoch": 0.2873648900484532, + "grad_norm": 19.433347702026367, + "learning_rate": 8.750931847244007e-06, + "loss": 9.988710403442383, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5000196099281311, + "loss_total_v6": 0.6484571099281311, + "step": 771 + }, + { + "epoch": 0.2877376071561685, + "grad_norm": 21.002748489379883, + "learning_rate": 8.74674579300843e-06, + "loss": 10.543783187866211, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6827598214149475, + "loss_total_v6": 0.8370566964149475, + "step": 772 + }, + { + "epoch": 0.2881103242638837, + "grad_norm": 19.297279357910156, + "learning_rate": 8.742553740855507e-06, + "loss": 10.724886894226074, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.8077712655067444, + "loss_total_v6": 0.9815993905067444, + "step": 773 + }, + { + "epoch": 0.28848304137159897, + "grad_norm": 19.460060119628906, + "learning_rate": 8.738355697496015e-06, + "loss": 10.291951179504395, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5009430646896362, + "loss_total_v6": 0.6571930646896362, + "step": 774 + }, + { + "epoch": 0.2888557584793142, + "grad_norm": 19.901443481445312, + "learning_rate": 8.734151669650327e-06, + "loss": 10.54587173461914, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6604688167572021, + "loss_total_v6": 0.8284375667572021, + "step": 775 + }, + { + "epoch": 0.28922847558702947, + "grad_norm": 18.573240280151367, + "learning_rate": 8.729941664048394e-06, + "loss": 10.760456085205078, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5821225047111511, + "loss_total_v6": 0.7266537547111511, + "step": 776 + }, + { + "epoch": 0.2896011926947447, + "grad_norm": 18.841344833374023, + "learning_rate": 8.72572568742973e-06, + "loss": 10.778419494628906, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5495365858078003, + "loss_total_v6": 0.6979740858078003, + "step": 777 + }, + { + "epoch": 0.2899739098024599, + "grad_norm": 18.91331672668457, + "learning_rate": 8.721503746543417e-06, + "loss": 10.20022201538086, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.48402512073516846, + "loss_total_v6": 0.6422282457351685, + "step": 778 + }, + { + "epoch": 0.29034662691017515, + "grad_norm": 18.791118621826172, + "learning_rate": 8.717275848148078e-06, + "loss": 10.158016204833984, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.559654176235199, + "loss_total_v6": 0.706138551235199, + "step": 779 + }, + { + "epoch": 0.2907193440178904, + "grad_norm": 18.7811336517334, + "learning_rate": 8.71304199901188e-06, + "loss": 10.07348918914795, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.586184561252594, + "loss_total_v6": 0.730715811252594, + "step": 780 + }, + { + "epoch": 0.29109206112560565, + "grad_norm": 19.142107009887695, + "learning_rate": 8.708802205912508e-06, + "loss": 10.461957931518555, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.2536628544330597, + "loss_total_v6": 0.3981941044330597, + "step": 781 + }, + { + "epoch": 0.2914647782333209, + "grad_norm": 18.679946899414062, + "learning_rate": 8.704556475637166e-06, + "loss": 10.6745023727417, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6434369683265686, + "loss_total_v6": 0.7977338433265686, + "step": 782 + }, + { + "epoch": 0.29183749534103615, + "grad_norm": 19.426700592041016, + "learning_rate": 8.700304814982565e-06, + "loss": 10.711553573608398, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5564456582069397, + "loss_total_v6": 0.7087894082069397, + "step": 783 + }, + { + "epoch": 0.2922102124487514, + "grad_norm": 19.053653717041016, + "learning_rate": 8.696047230754905e-06, + "loss": 10.42132568359375, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5357387661933899, + "loss_total_v6": 0.6822231411933899, + "step": 784 + }, + { + "epoch": 0.29258292955646664, + "grad_norm": 20.257003784179688, + "learning_rate": 8.691783729769874e-06, + "loss": 10.219240188598633, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6981673240661621, + "loss_total_v6": 0.8524641990661621, + "step": 785 + }, + { + "epoch": 0.2929556466641819, + "grad_norm": 19.02440071105957, + "learning_rate": 8.687514318852623e-06, + "loss": 10.45258617401123, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4140147566795349, + "loss_total_v6": 0.5644053816795349, + "step": 786 + }, + { + "epoch": 0.29332836377189714, + "grad_norm": 20.096067428588867, + "learning_rate": 8.683239004837776e-06, + "loss": 10.395929336547852, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.7349768877029419, + "loss_total_v6": 0.9048987627029419, + "step": 787 + }, + { + "epoch": 0.2937010808796124, + "grad_norm": 20.10261344909668, + "learning_rate": 8.678957794569395e-06, + "loss": 10.166891098022461, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.474395751953125, + "loss_total_v6": 0.628692626953125, + "step": 788 + }, + { + "epoch": 0.29407379798732763, + "grad_norm": 19.712533950805664, + "learning_rate": 8.674670694900988e-06, + "loss": 10.092635154724121, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.5698098540306091, + "loss_total_v6": 0.7436379790306091, + "step": 789 + }, + { + "epoch": 0.2944465150950429, + "grad_norm": 19.203983306884766, + "learning_rate": 8.67037771269549e-06, + "loss": 10.643967628479004, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5307943224906921, + "loss_total_v6": 0.6831380724906921, + "step": 790 + }, + { + "epoch": 0.29481923220275813, + "grad_norm": 20.349992752075195, + "learning_rate": 8.66607885482525e-06, + "loss": 10.276748657226562, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.48523539304733276, + "loss_total_v6": 0.6395322680473328, + "step": 791 + }, + { + "epoch": 0.2951919493104734, + "grad_norm": 20.215425491333008, + "learning_rate": 8.661774128172027e-06, + "loss": 10.235944747924805, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4663100838661194, + "loss_total_v6": 0.6127944588661194, + "step": 792 + }, + { + "epoch": 0.29556466641818857, + "grad_norm": 19.141504287719727, + "learning_rate": 8.657463539626971e-06, + "loss": 10.11736011505127, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5000617504119873, + "loss_total_v6": 0.6582648754119873, + "step": 793 + }, + { + "epoch": 0.2959373835259038, + "grad_norm": 19.70263671875, + "learning_rate": 8.65314709609062e-06, + "loss": 10.269944190979004, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.19497449696063995, + "loss_total_v6": 0.34536510705947876, + "step": 794 + }, + { + "epoch": 0.29631010063361907, + "grad_norm": 19.148300170898438, + "learning_rate": 8.648824804472882e-06, + "loss": 10.083478927612305, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.608933687210083, + "loss_total_v6": 0.772996187210083, + "step": 795 + }, + { + "epoch": 0.2966828177413343, + "grad_norm": 19.459962844848633, + "learning_rate": 8.644496671693028e-06, + "loss": 10.134906768798828, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3553490936756134, + "loss_total_v6": 0.511599063873291, + "step": 796 + }, + { + "epoch": 0.29705553484904956, + "grad_norm": 20.262216567993164, + "learning_rate": 8.640162704679677e-06, + "loss": 10.747279167175293, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4989436864852905, + "loss_total_v6": 0.6473811864852905, + "step": 797 + }, + { + "epoch": 0.2974282519567648, + "grad_norm": 19.941665649414062, + "learning_rate": 8.635822910370793e-06, + "loss": 10.723616600036621, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5739613771438599, + "loss_total_v6": 0.7184926271438599, + "step": 798 + }, + { + "epoch": 0.29780096906448006, + "grad_norm": 20.010944366455078, + "learning_rate": 8.631477295713663e-06, + "loss": 10.169868469238281, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5912213921546936, + "loss_total_v6": 0.7533307671546936, + "step": 799 + }, + { + "epoch": 0.2981736861721953, + "grad_norm": 19.786327362060547, + "learning_rate": 8.627125867664895e-06, + "loss": 10.389223098754883, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.41656556725502014, + "loss_total_v6": 0.5728155374526978, + "step": 800 + }, + { + "epoch": 0.2981736861721953, + "eval_loss": 0.6263818740844727, + "eval_loss_alignment": 0.2853419638087607, + "eval_loss_alignment_w": 0.14267098190438035, + "eval_loss_sft": 0.4837109211784525, + "eval_loss_total_v6": 0.6263819033136735, + "eval_runtime": 206.7398, + "eval_samples_per_second": 9.026, + "eval_steps_per_second": 1.132, + "step": 800 + }, + { + "epoch": 0.29854640327991055, + "grad_norm": 19.515151977539062, + "learning_rate": 8.622768633190404e-06, + "loss": 10.255491256713867, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6115258932113647, + "loss_total_v6": 0.7580102682113647, + "step": 801 + }, + { + "epoch": 0.2989191203876258, + "grad_norm": 18.247835159301758, + "learning_rate": 8.618405599265394e-06, + "loss": 10.147729873657227, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5041210055351257, + "loss_total_v6": 0.6486522555351257, + "step": 802 + }, + { + "epoch": 0.29929183749534105, + "grad_norm": 19.60084342956543, + "learning_rate": 8.61403677287436e-06, + "loss": 10.387177467346191, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6647974252700806, + "loss_total_v6": 0.8132349252700806, + "step": 803 + }, + { + "epoch": 0.2996645546030563, + "grad_norm": 25.334524154663086, + "learning_rate": 8.609662161011068e-06, + "loss": 10.262443542480469, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.7319759130477905, + "loss_total_v6": 0.8901790380477905, + "step": 804 + }, + { + "epoch": 0.30003727171077155, + "grad_norm": 19.198532104492188, + "learning_rate": 8.605281770678541e-06, + "loss": 10.376089096069336, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5971904397010803, + "loss_total_v6": 0.7495341897010803, + "step": 805 + }, + { + "epoch": 0.3004099888184868, + "grad_norm": 18.74089813232422, + "learning_rate": 8.600895608889059e-06, + "loss": 10.633451461791992, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5097585320472717, + "loss_total_v6": 0.6640554070472717, + "step": 806 + }, + { + "epoch": 0.300782705926202, + "grad_norm": 20.357595443725586, + "learning_rate": 8.596503682664138e-06, + "loss": 10.325540542602539, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.4361581802368164, + "loss_total_v6": 0.6041269302368164, + "step": 807 + }, + { + "epoch": 0.30115542303391724, + "grad_norm": 20.349098205566406, + "learning_rate": 8.59210599903452e-06, + "loss": 10.529550552368164, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.36361488699913025, + "loss_total_v6": 0.5120524168014526, + "step": 808 + }, + { + "epoch": 0.3015281401416325, + "grad_norm": 19.545852661132812, + "learning_rate": 8.587702565040168e-06, + "loss": 10.445069313049316, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8635703921318054, + "loss_total_v6": 1.0178673267364502, + "step": 809 + }, + { + "epoch": 0.30190085724934773, + "grad_norm": 17.830509185791016, + "learning_rate": 8.583293387730245e-06, + "loss": 10.396310806274414, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.701001763343811, + "loss_total_v6": 0.857251763343811, + "step": 810 + }, + { + "epoch": 0.302273574357063, + "grad_norm": 18.525848388671875, + "learning_rate": 8.578878474163115e-06, + "loss": 10.285547256469727, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2508299648761749, + "loss_total_v6": 0.4012205898761749, + "step": 811 + }, + { + "epoch": 0.3026462914647782, + "grad_norm": 20.143070220947266, + "learning_rate": 8.574457831406317e-06, + "loss": 10.246715545654297, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6435902118682861, + "loss_total_v6": 0.7900745868682861, + "step": 812 + }, + { + "epoch": 0.3030190085724935, + "grad_norm": 19.664783477783203, + "learning_rate": 8.57003146653657e-06, + "loss": 10.309915542602539, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2397170066833496, + "loss_total_v6": 0.3901076316833496, + "step": 813 + }, + { + "epoch": 0.3033917256802087, + "grad_norm": 20.932308197021484, + "learning_rate": 8.565599386639744e-06, + "loss": 10.18804931640625, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6638144850730896, + "loss_total_v6": 0.8298301100730896, + "step": 814 + }, + { + "epoch": 0.30376444278792397, + "grad_norm": 18.895681381225586, + "learning_rate": 8.561161598810864e-06, + "loss": 9.905094146728516, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.41396546363830566, + "loss_total_v6": 0.5702154636383057, + "step": 815 + }, + { + "epoch": 0.3041371598956392, + "grad_norm": 20.384490966796875, + "learning_rate": 8.55671811015409e-06, + "loss": 10.592598915100098, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.9464711546897888, + "loss_total_v6": 1.0890493392944336, + "step": 816 + }, + { + "epoch": 0.30450987700335447, + "grad_norm": 19.78874969482422, + "learning_rate": 8.552268927782713e-06, + "loss": 10.20390796661377, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.5087872743606567, + "loss_total_v6": 0.6728497743606567, + "step": 817 + }, + { + "epoch": 0.3048825941110697, + "grad_norm": 20.3404541015625, + "learning_rate": 8.547814058819128e-06, + "loss": 10.110319137573242, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.8047623634338379, + "loss_total_v6": 0.9610123634338379, + "step": 818 + }, + { + "epoch": 0.30525531121878496, + "grad_norm": 19.091894149780273, + "learning_rate": 8.543353510394846e-06, + "loss": 10.605086326599121, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.2596352994441986, + "loss_total_v6": 0.4061196744441986, + "step": 819 + }, + { + "epoch": 0.3056280283265002, + "grad_norm": 18.660451889038086, + "learning_rate": 8.53888728965046e-06, + "loss": 10.794519424438477, + "loss_alignment": 0.26953125, + "loss_alignment_w": 0.134765625, + "loss_sft": 0.5880552530288696, + "loss_total_v6": 0.7228208780288696, + "step": 820 + }, + { + "epoch": 0.3060007454342154, + "grad_norm": 20.52309226989746, + "learning_rate": 8.534415403735647e-06, + "loss": 10.456007957458496, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.654296875, + "loss_total_v6": 0.802734375, + "step": 821 + }, + { + "epoch": 0.30637346254193065, + "grad_norm": 18.608932495117188, + "learning_rate": 8.529937859809156e-06, + "loss": 10.650293350219727, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.644860565662384, + "loss_total_v6": 0.791344940662384, + "step": 822 + }, + { + "epoch": 0.3067461796496459, + "grad_norm": 19.075342178344727, + "learning_rate": 8.525454665038788e-06, + "loss": 9.800556182861328, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6059462428092957, + "loss_total_v6": 0.7661024928092957, + "step": 823 + }, + { + "epoch": 0.30711889675736115, + "grad_norm": 20.321392059326172, + "learning_rate": 8.520965826601394e-06, + "loss": 10.347410202026367, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5633559226989746, + "loss_total_v6": 0.7196059226989746, + "step": 824 + }, + { + "epoch": 0.3074916138650764, + "grad_norm": 20.230506896972656, + "learning_rate": 8.516471351682858e-06, + "loss": 10.490832328796387, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.712181806564331, + "loss_total_v6": 0.870384931564331, + "step": 825 + }, + { + "epoch": 0.30786433097279164, + "grad_norm": 18.81288719177246, + "learning_rate": 8.511971247478089e-06, + "loss": 10.28844165802002, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.704404354095459, + "loss_total_v6": 0.870419979095459, + "step": 826 + }, + { + "epoch": 0.3082370480805069, + "grad_norm": 20.41475486755371, + "learning_rate": 8.507465521191003e-06, + "loss": 10.34769344329834, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.4276556968688965, + "loss_total_v6": 0.5956244468688965, + "step": 827 + }, + { + "epoch": 0.30860976518822214, + "grad_norm": 20.09066390991211, + "learning_rate": 8.502954180034519e-06, + "loss": 10.441192626953125, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.42252641916275024, + "loss_total_v6": 0.5826826691627502, + "step": 828 + }, + { + "epoch": 0.3089824822959374, + "grad_norm": 19.956256866455078, + "learning_rate": 8.49843723123055e-06, + "loss": 10.230354309082031, + "loss_alignment": 0.26171875, + "loss_alignment_w": 0.130859375, + "loss_sft": 0.3613034188747406, + "loss_total_v6": 0.4921627938747406, + "step": 829 + }, + { + "epoch": 0.30935519940365264, + "grad_norm": 19.37479019165039, + "learning_rate": 8.493914682009974e-06, + "loss": 10.368806838989258, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.7783516645431519, + "loss_total_v6": 0.9404610395431519, + "step": 830 + }, + { + "epoch": 0.3097279165113679, + "grad_norm": 19.792335510253906, + "learning_rate": 8.489386539612643e-06, + "loss": 10.395374298095703, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.35119888186454773, + "loss_total_v6": 0.5015895366668701, + "step": 831 + }, + { + "epoch": 0.31010063361908313, + "grad_norm": 19.221940994262695, + "learning_rate": 8.48485281128736e-06, + "loss": 10.188166618347168, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6011605262756348, + "loss_total_v6": 0.7476449012756348, + "step": 832 + }, + { + "epoch": 0.3104733507267984, + "grad_norm": 19.359085083007812, + "learning_rate": 8.480313504291871e-06, + "loss": 10.684158325195312, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.404175728559494, + "loss_total_v6": 0.5604257583618164, + "step": 833 + }, + { + "epoch": 0.31084606783451363, + "grad_norm": 20.05169105529785, + "learning_rate": 8.475768625892853e-06, + "loss": 10.436107635498047, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.7004373669624329, + "loss_total_v6": 0.8449686169624329, + "step": 834 + }, + { + "epoch": 0.3112187849422288, + "grad_norm": 19.048686981201172, + "learning_rate": 8.471218183365901e-06, + "loss": 9.882018089294434, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.6285873055458069, + "loss_total_v6": 0.7692123055458069, + "step": 835 + }, + { + "epoch": 0.31159150204994407, + "grad_norm": 19.347137451171875, + "learning_rate": 8.466662183995515e-06, + "loss": 9.869548797607422, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.2625984251499176, + "loss_total_v6": 0.4168953001499176, + "step": 836 + }, + { + "epoch": 0.3119642191576593, + "grad_norm": 19.64715003967285, + "learning_rate": 8.462100635075097e-06, + "loss": 9.905113220214844, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.030075369402766228, + "loss_total_v6": 0.18827849626541138, + "step": 837 + }, + { + "epoch": 0.31233693626537457, + "grad_norm": 18.931974411010742, + "learning_rate": 8.457533543906923e-06, + "loss": 9.710095405578613, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5290651917457581, + "loss_total_v6": 0.6872683167457581, + "step": 838 + }, + { + "epoch": 0.3127096533730898, + "grad_norm": 19.72159194946289, + "learning_rate": 8.452960917802151e-06, + "loss": 10.395320892333984, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.7440706491470337, + "loss_total_v6": 0.9178987741470337, + "step": 839 + }, + { + "epoch": 0.31308237048080506, + "grad_norm": 20.426008224487305, + "learning_rate": 8.448382764080794e-06, + "loss": 10.534547805786133, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4782771170139313, + "loss_total_v6": 0.6345270872116089, + "step": 840 + }, + { + "epoch": 0.3134550875885203, + "grad_norm": 18.594654083251953, + "learning_rate": 8.443799090071716e-06, + "loss": 9.743221282958984, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.49180126190185547, + "loss_total_v6": 0.6441450119018555, + "step": 841 + }, + { + "epoch": 0.31382780469623556, + "grad_norm": 19.474586486816406, + "learning_rate": 8.439209903112617e-06, + "loss": 10.346916198730469, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4020620882511139, + "loss_total_v6": 0.5504995584487915, + "step": 842 + }, + { + "epoch": 0.3142005218039508, + "grad_norm": 22.092988967895508, + "learning_rate": 8.434615210550024e-06, + "loss": 10.393232345581055, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.16038304567337036, + "loss_total_v6": 0.30491429567337036, + "step": 843 + }, + { + "epoch": 0.31457323891166605, + "grad_norm": 18.68972396850586, + "learning_rate": 8.430015019739274e-06, + "loss": 10.21088981628418, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.36161813139915466, + "loss_total_v6": 0.506149411201477, + "step": 844 + }, + { + "epoch": 0.3149459560193813, + "grad_norm": 20.469209671020508, + "learning_rate": 8.425409338044509e-06, + "loss": 10.602561950683594, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5100750923156738, + "loss_total_v6": 0.6663250923156738, + "step": 845 + }, + { + "epoch": 0.31531867312709655, + "grad_norm": 19.812885284423828, + "learning_rate": 8.42079817283866e-06, + "loss": 10.56320571899414, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.8134104609489441, + "loss_total_v6": 0.8134104609489441, + "step": 846 + }, + { + "epoch": 0.3156913902348118, + "grad_norm": 23.291362762451172, + "learning_rate": 8.416181531503436e-06, + "loss": 9.999361038208008, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6915258765220642, + "loss_total_v6": 0.8497290015220642, + "step": 847 + }, + { + "epoch": 0.31606410734252705, + "grad_norm": 20.853317260742188, + "learning_rate": 8.411559421429312e-06, + "loss": 10.357938766479492, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4902481138706207, + "loss_total_v6": 0.6464980840682983, + "step": 848 + }, + { + "epoch": 0.3164368244502423, + "grad_norm": 20.02155876159668, + "learning_rate": 8.406931850015519e-06, + "loss": 10.214245796203613, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.09696272015571594, + "loss_total_v6": 0.24149397015571594, + "step": 849 + }, + { + "epoch": 0.3168095415579575, + "grad_norm": 19.33135986328125, + "learning_rate": 8.40229882467003e-06, + "loss": 10.421871185302734, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.495225191116333, + "loss_total_v6": 0.655381441116333, + "step": 850 + }, + { + "epoch": 0.31718225866567273, + "grad_norm": 19.77585792541504, + "learning_rate": 8.397660352809546e-06, + "loss": 10.136238098144531, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7861964702606201, + "loss_total_v6": 0.9404933452606201, + "step": 851 + }, + { + "epoch": 0.317554975773388, + "grad_norm": 19.328624725341797, + "learning_rate": 8.393016441859493e-06, + "loss": 10.55006217956543, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.32312628626823425, + "loss_total_v6": 0.47156378626823425, + "step": 852 + }, + { + "epoch": 0.31792769288110323, + "grad_norm": 18.947338104248047, + "learning_rate": 8.388367099254002e-06, + "loss": 10.522826194763184, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47265806794166565, + "loss_total_v6": 0.6171892881393433, + "step": 853 + }, + { + "epoch": 0.3183004099888185, + "grad_norm": 20.40739631652832, + "learning_rate": 8.383712332435892e-06, + "loss": 10.275495529174805, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6870444416999817, + "loss_total_v6": 0.8354819416999817, + "step": 854 + }, + { + "epoch": 0.3186731270965337, + "grad_norm": 18.217262268066406, + "learning_rate": 8.379052148856675e-06, + "loss": 10.109725952148438, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.4150889813899994, + "loss_total_v6": 0.5830577611923218, + "step": 855 + }, + { + "epoch": 0.319045844204249, + "grad_norm": 19.290416717529297, + "learning_rate": 8.374386555976528e-06, + "loss": 10.346710205078125, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.8361350893974304, + "loss_total_v6": 0.9962913393974304, + "step": 856 + }, + { + "epoch": 0.3194185613119642, + "grad_norm": 17.606735229492188, + "learning_rate": 8.369715561264295e-06, + "loss": 10.366580963134766, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.3452901244163513, + "loss_total_v6": 0.4839619994163513, + "step": 857 + }, + { + "epoch": 0.31979127841967947, + "grad_norm": 18.21479606628418, + "learning_rate": 8.365039172197455e-06, + "loss": 10.07874870300293, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4248141050338745, + "loss_total_v6": 0.5752047300338745, + "step": 858 + }, + { + "epoch": 0.3201639955273947, + "grad_norm": 19.715654373168945, + "learning_rate": 8.360357396262136e-06, + "loss": 10.564218521118164, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6512513160705566, + "loss_total_v6": 0.8055481910705566, + "step": 859 + }, + { + "epoch": 0.32053671263510997, + "grad_norm": 20.28968620300293, + "learning_rate": 8.355670240953082e-06, + "loss": 10.591796875, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5776392221450806, + "loss_total_v6": 0.7280298471450806, + "step": 860 + }, + { + "epoch": 0.3209094297428252, + "grad_norm": 18.96575164794922, + "learning_rate": 8.350977713773645e-06, + "loss": 10.258102416992188, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.428821325302124, + "loss_total_v6": 0.573352575302124, + "step": 861 + }, + { + "epoch": 0.32128214685054046, + "grad_norm": 19.57468605041504, + "learning_rate": 8.346279822235788e-06, + "loss": 10.471841812133789, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.31461870670318604, + "loss_total_v6": 0.45914995670318604, + "step": 862 + }, + { + "epoch": 0.3216548639582557, + "grad_norm": 19.190649032592773, + "learning_rate": 8.341576573860049e-06, + "loss": 10.049988746643066, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.45434728264808655, + "loss_total_v6": 0.6086441278457642, + "step": 863 + }, + { + "epoch": 0.3220275810659709, + "grad_norm": 20.499134063720703, + "learning_rate": 8.336867976175549e-06, + "loss": 10.300240516662598, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6007087826728821, + "loss_total_v6": 0.7510994076728821, + "step": 864 + }, + { + "epoch": 0.32240029817368615, + "grad_norm": 19.872882843017578, + "learning_rate": 8.332154036719971e-06, + "loss": 9.918525695800781, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5169247388839722, + "loss_total_v6": 0.6790341138839722, + "step": 865 + }, + { + "epoch": 0.3227730152814014, + "grad_norm": 23.089982986450195, + "learning_rate": 8.32743476303955e-06, + "loss": 9.889410972595215, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.42639482021331787, + "loss_total_v6": 0.5631135702133179, + "step": 866 + }, + { + "epoch": 0.32314573238911665, + "grad_norm": 19.93797492980957, + "learning_rate": 8.322710162689056e-06, + "loss": 10.383750915527344, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.4691576063632965, + "loss_total_v6": 0.6547044515609741, + "step": 867 + }, + { + "epoch": 0.3235184494968319, + "grad_norm": 19.860198974609375, + "learning_rate": 8.317980243231788e-06, + "loss": 10.921018600463867, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3889395296573639, + "loss_total_v6": 0.5510488748550415, + "step": 868 + }, + { + "epoch": 0.32389116660454714, + "grad_norm": 18.937564849853516, + "learning_rate": 8.313245012239565e-06, + "loss": 10.168912887573242, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.48539265990257263, + "loss_total_v6": 0.639689564704895, + "step": 869 + }, + { + "epoch": 0.3242638837122624, + "grad_norm": 19.706159591674805, + "learning_rate": 8.308504477292704e-06, + "loss": 10.515310287475586, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3630555272102356, + "loss_total_v6": 0.5212586522102356, + "step": 870 + }, + { + "epoch": 0.32463660081997764, + "grad_norm": 18.756290435791016, + "learning_rate": 8.303758645980013e-06, + "loss": 10.548624038696289, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5820618867874146, + "loss_total_v6": 0.5820618867874146, + "step": 871 + }, + { + "epoch": 0.3250093179276929, + "grad_norm": 25.764251708984375, + "learning_rate": 8.299007525898776e-06, + "loss": 10.302593231201172, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.32940712571144104, + "loss_total_v6": 0.48565712571144104, + "step": 872 + }, + { + "epoch": 0.32538203503540813, + "grad_norm": 18.661237716674805, + "learning_rate": 8.294251124654753e-06, + "loss": 10.114164352416992, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6607386469841003, + "loss_total_v6": 0.8169886469841003, + "step": 873 + }, + { + "epoch": 0.3257547521431234, + "grad_norm": 20.34593391418457, + "learning_rate": 8.289489449862145e-06, + "loss": 10.248104095458984, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5806341171264648, + "loss_total_v6": 0.7368841171264648, + "step": 874 + }, + { + "epoch": 0.32612746925083863, + "grad_norm": 25.289480209350586, + "learning_rate": 8.284722509143608e-06, + "loss": 10.514486312866211, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.6744999289512634, + "loss_total_v6": 0.8424686789512634, + "step": 875 + }, + { + "epoch": 0.3265001863585539, + "grad_norm": 19.219867706298828, + "learning_rate": 8.279950310130218e-06, + "loss": 9.935791969299316, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.39675745368003845, + "loss_total_v6": 0.5451949834823608, + "step": 876 + }, + { + "epoch": 0.3268729034662691, + "grad_norm": 18.310728073120117, + "learning_rate": 8.275172860461474e-06, + "loss": 10.110109329223633, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3081715404987335, + "loss_total_v6": 0.4585621654987335, + "step": 877 + }, + { + "epoch": 0.3272456205739843, + "grad_norm": 19.351160049438477, + "learning_rate": 8.270390167785278e-06, + "loss": 10.836860656738281, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4888249933719635, + "loss_total_v6": 0.6450749635696411, + "step": 878 + }, + { + "epoch": 0.32761833768169957, + "grad_norm": 19.0300235748291, + "learning_rate": 8.265602239757929e-06, + "loss": 10.260648727416992, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.525502622127533, + "loss_total_v6": 0.685658872127533, + "step": 879 + }, + { + "epoch": 0.3279910547894148, + "grad_norm": 19.407251358032227, + "learning_rate": 8.260809084044102e-06, + "loss": 10.400870323181152, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.320166677236557, + "loss_total_v6": 0.464697927236557, + "step": 880 + }, + { + "epoch": 0.32836377189713006, + "grad_norm": 18.400447845458984, + "learning_rate": 8.256010708316839e-06, + "loss": 10.032439231872559, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4341377913951874, + "loss_total_v6": 0.5825753211975098, + "step": 881 + }, + { + "epoch": 0.3287364890048453, + "grad_norm": 22.63980484008789, + "learning_rate": 8.251207120257549e-06, + "loss": 10.820075035095215, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6421679258346558, + "loss_total_v6": 0.8003710508346558, + "step": 882 + }, + { + "epoch": 0.32910920611256056, + "grad_norm": 19.039186477661133, + "learning_rate": 8.246398327555976e-06, + "loss": 10.273321151733398, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.44706958532333374, + "loss_total_v6": 0.6013664603233337, + "step": 883 + }, + { + "epoch": 0.3294819232202758, + "grad_norm": 18.848562240600586, + "learning_rate": 8.241584337910195e-06, + "loss": 10.452336311340332, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3193075954914093, + "loss_total_v6": 0.4657919704914093, + "step": 884 + }, + { + "epoch": 0.32985464032799106, + "grad_norm": 19.508033752441406, + "learning_rate": 8.236765159026605e-06, + "loss": 10.385538101196289, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 1.141599416732788, + "loss_total_v6": 1.311521291732788, + "step": 885 + }, + { + "epoch": 0.3302273574357063, + "grad_norm": 31.374744415283203, + "learning_rate": 8.231940798619913e-06, + "loss": 10.670254707336426, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.3015936613082886, + "loss_total_v6": 0.4676092863082886, + "step": 886 + }, + { + "epoch": 0.33060007454342155, + "grad_norm": 19.214872360229492, + "learning_rate": 8.227111264413115e-06, + "loss": 10.276068687438965, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6801304221153259, + "loss_total_v6": 0.8285679221153259, + "step": 887 + }, + { + "epoch": 0.3309727916511368, + "grad_norm": 18.416723251342773, + "learning_rate": 8.222276564137492e-06, + "loss": 10.368326187133789, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.21548186242580414, + "loss_total_v6": 0.36782562732696533, + "step": 888 + }, + { + "epoch": 0.33134550875885205, + "grad_norm": 17.795333862304688, + "learning_rate": 8.2174367055326e-06, + "loss": 10.142353057861328, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.43583232164382935, + "loss_total_v6": 0.5823166966438293, + "step": 889 + }, + { + "epoch": 0.3317182258665673, + "grad_norm": 20.89551544189453, + "learning_rate": 8.212591696346244e-06, + "loss": 10.220333099365234, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47164562344551086, + "loss_total_v6": 0.6161768436431885, + "step": 890 + }, + { + "epoch": 0.33209094297428254, + "grad_norm": 20.35068130493164, + "learning_rate": 8.207741544334482e-06, + "loss": 10.396333694458008, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4727858304977417, + "loss_total_v6": 0.6231764554977417, + "step": 891 + }, + { + "epoch": 0.33246366008199774, + "grad_norm": 19.830970764160156, + "learning_rate": 8.202886257261598e-06, + "loss": 10.529850006103516, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.31945937871932983, + "loss_total_v6": 0.47180312871932983, + "step": 892 + }, + { + "epoch": 0.332836377189713, + "grad_norm": 19.927661895751953, + "learning_rate": 8.198025842900103e-06, + "loss": 10.18465805053711, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.4244807958602905, + "loss_total_v6": 0.5885432958602905, + "step": 893 + }, + { + "epoch": 0.33320909429742823, + "grad_norm": 20.104352951049805, + "learning_rate": 8.193160309030708e-06, + "loss": 10.320089340209961, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6159841418266296, + "loss_total_v6": 0.7722341418266296, + "step": 894 + }, + { + "epoch": 0.3335818114051435, + "grad_norm": 18.36610984802246, + "learning_rate": 8.18828966344233e-06, + "loss": 10.578956604003906, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3631458878517151, + "loss_total_v6": 0.5135365128517151, + "step": 895 + }, + { + "epoch": 0.33395452851285873, + "grad_norm": 19.9779109954834, + "learning_rate": 8.18341391393206e-06, + "loss": 9.830070495605469, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.16167068481445312, + "loss_total_v6": 0.3062019348144531, + "step": 896 + }, + { + "epoch": 0.334327245620574, + "grad_norm": 19.267271041870117, + "learning_rate": 8.17853306830516e-06, + "loss": 10.111917495727539, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.38436830043792725, + "loss_total_v6": 0.5367120504379272, + "step": 897 + }, + { + "epoch": 0.3346999627282892, + "grad_norm": 20.054428100585938, + "learning_rate": 8.173647134375058e-06, + "loss": 10.013364791870117, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7775718569755554, + "loss_total_v6": 0.9201499819755554, + "step": 898 + }, + { + "epoch": 0.3350726798360045, + "grad_norm": 18.757829666137695, + "learning_rate": 8.16875611996332e-06, + "loss": 10.48208999633789, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.675530195236206, + "loss_total_v6": 0.675530195236206, + "step": 899 + }, + { + "epoch": 0.3354453969437197, + "grad_norm": 19.30046272277832, + "learning_rate": 8.163860032899646e-06, + "loss": 10.402149200439453, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.5376886129379272, + "loss_total_v6": 0.7173761129379272, + "step": 900 + }, + { + "epoch": 0.3354453969437197, + "eval_loss": 0.6207536458969116, + "eval_loss_alignment": 0.2843403612446581, + "eval_loss_alignment_w": 0.14217018062232906, + "eval_loss_sft": 0.4785834435039224, + "eval_loss_total_v6": 0.620753623975011, + "eval_runtime": 206.6226, + "eval_samples_per_second": 9.031, + "eval_steps_per_second": 1.132, + "step": 900 + }, + { + "epoch": 0.33581811405143497, + "grad_norm": 19.671632766723633, + "learning_rate": 8.15895888102186e-06, + "loss": 10.37048625946045, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5738008618354797, + "loss_total_v6": 0.7261446118354797, + "step": 901 + }, + { + "epoch": 0.3361908311591502, + "grad_norm": 17.905223846435547, + "learning_rate": 8.154052672175888e-06, + "loss": 10.032369613647461, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5573602914810181, + "loss_total_v6": 0.7136102914810181, + "step": 902 + }, + { + "epoch": 0.33656354826686546, + "grad_norm": 18.99173355102539, + "learning_rate": 8.149141414215759e-06, + "loss": 10.06932258605957, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.6830957531929016, + "loss_total_v6": 0.8569238781929016, + "step": 903 + }, + { + "epoch": 0.3369362653745807, + "grad_norm": 19.343969345092773, + "learning_rate": 8.144225115003578e-06, + "loss": 9.890565872192383, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5564582347869873, + "loss_total_v6": 0.7185676097869873, + "step": 904 + }, + { + "epoch": 0.33730898248229596, + "grad_norm": 19.362083435058594, + "learning_rate": 8.139303782409527e-06, + "loss": 10.004714012145996, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.43981584906578064, + "loss_total_v6": 0.596065878868103, + "step": 905 + }, + { + "epoch": 0.3376816995900112, + "grad_norm": 18.4774169921875, + "learning_rate": 8.134377424311838e-06, + "loss": 10.751402854919434, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.393903523683548, + "loss_total_v6": 0.5364816188812256, + "step": 906 + }, + { + "epoch": 0.3380544166977264, + "grad_norm": 18.414152145385742, + "learning_rate": 8.12944604859679e-06, + "loss": 10.5197114944458, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6530413627624512, + "loss_total_v6": 0.8073382377624512, + "step": 907 + }, + { + "epoch": 0.33842713380544165, + "grad_norm": 21.919042587280273, + "learning_rate": 8.124509663158699e-06, + "loss": 10.324548721313477, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5761296153068542, + "loss_total_v6": 0.7226139903068542, + "step": 908 + }, + { + "epoch": 0.3387998509131569, + "grad_norm": 19.279455184936523, + "learning_rate": 8.119568275899897e-06, + "loss": 10.100152969360352, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6008135080337524, + "loss_total_v6": 0.7668291330337524, + "step": 909 + }, + { + "epoch": 0.33917256802087214, + "grad_norm": 21.458938598632812, + "learning_rate": 8.114621894730725e-06, + "loss": 10.06148910522461, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7864084839820862, + "loss_total_v6": 0.9407053589820862, + "step": 910 + }, + { + "epoch": 0.3395452851285874, + "grad_norm": 19.47991943359375, + "learning_rate": 8.109670527569517e-06, + "loss": 10.228113174438477, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4430772066116333, + "loss_total_v6": 0.5954209566116333, + "step": 911 + }, + { + "epoch": 0.33991800223630264, + "grad_norm": 20.00844383239746, + "learning_rate": 8.104714182342586e-06, + "loss": 10.194330215454102, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.3486599922180176, + "loss_total_v6": 0.4853787422180176, + "step": 912 + }, + { + "epoch": 0.3402907193440179, + "grad_norm": 19.326183319091797, + "learning_rate": 8.09975286698422e-06, + "loss": 10.082367897033691, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5411863923072815, + "loss_total_v6": 0.6915770173072815, + "step": 913 + }, + { + "epoch": 0.34066343645173314, + "grad_norm": 19.271347045898438, + "learning_rate": 8.09478658943666e-06, + "loss": 10.439953804016113, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4484194815158844, + "loss_total_v6": 0.602716326713562, + "step": 914 + }, + { + "epoch": 0.3410361535594484, + "grad_norm": 18.89221954345703, + "learning_rate": 8.08981535765009e-06, + "loss": 10.496641159057617, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5930313467979431, + "loss_total_v6": 0.7453750967979431, + "step": 915 + }, + { + "epoch": 0.34140887066716363, + "grad_norm": 18.636226654052734, + "learning_rate": 8.084839179582626e-06, + "loss": 10.158735275268555, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.8589374423027039, + "loss_total_v6": 1.0171406269073486, + "step": 916 + }, + { + "epoch": 0.3417815877748789, + "grad_norm": 24.369863510131836, + "learning_rate": 8.079858063200303e-06, + "loss": 10.40866470336914, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.46470922231674194, + "loss_total_v6": 0.6209592223167419, + "step": 917 + }, + { + "epoch": 0.34215430488259413, + "grad_norm": 20.006744384765625, + "learning_rate": 8.07487201647706e-06, + "loss": 10.641796112060547, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6396182179450989, + "loss_total_v6": 0.7900088429450989, + "step": 918 + }, + { + "epoch": 0.3425270219903094, + "grad_norm": 19.150257110595703, + "learning_rate": 8.06988104739473e-06, + "loss": 10.546600341796875, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5737324953079224, + "loss_total_v6": 0.7299824953079224, + "step": 919 + }, + { + "epoch": 0.3428997390980246, + "grad_norm": 19.519365310668945, + "learning_rate": 8.064885163943022e-06, + "loss": 10.204014778137207, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6896573901176453, + "loss_total_v6": 0.8439542651176453, + "step": 920 + }, + { + "epoch": 0.3432724562057398, + "grad_norm": 18.053936004638672, + "learning_rate": 8.059884374119518e-06, + "loss": 10.532630920410156, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6963410973548889, + "loss_total_v6": 0.8428254723548889, + "step": 921 + }, + { + "epoch": 0.34364517331345507, + "grad_norm": 18.60445785522461, + "learning_rate": 8.054878685929652e-06, + "loss": 10.274992942810059, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4341052770614624, + "loss_total_v6": 0.5786365270614624, + "step": 922 + }, + { + "epoch": 0.3440178904211703, + "grad_norm": 19.870737075805664, + "learning_rate": 8.049868107386696e-06, + "loss": 10.372498512268066, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5853244662284851, + "loss_total_v6": 0.7435275912284851, + "step": 923 + }, + { + "epoch": 0.34439060752888556, + "grad_norm": 18.357131958007812, + "learning_rate": 8.044852646511756e-06, + "loss": 10.437417984008789, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5576837658882141, + "loss_total_v6": 0.7100275158882141, + "step": 924 + }, + { + "epoch": 0.3447633246366008, + "grad_norm": 19.08140754699707, + "learning_rate": 8.039832311333753e-06, + "loss": 10.66491985321045, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.41330283880233765, + "loss_total_v6": 0.5636934638023376, + "step": 925 + }, + { + "epoch": 0.34513604174431606, + "grad_norm": 20.53042221069336, + "learning_rate": 8.034807109889405e-06, + "loss": 10.36145305633545, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.24394676089286804, + "loss_total_v6": 0.39433738589286804, + "step": 926 + }, + { + "epoch": 0.3455087588520313, + "grad_norm": 19.217557907104492, + "learning_rate": 8.029777050223226e-06, + "loss": 9.78750991821289, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4437503218650818, + "loss_total_v6": 0.5941409468650818, + "step": 927 + }, + { + "epoch": 0.34588147595974655, + "grad_norm": 19.0993595123291, + "learning_rate": 8.024742140387506e-06, + "loss": 10.689041137695312, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5595823526382446, + "loss_total_v6": 0.5595823526382446, + "step": 928 + }, + { + "epoch": 0.3462541930674618, + "grad_norm": 19.573158264160156, + "learning_rate": 8.019702388442299e-06, + "loss": 10.465438842773438, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.7271634340286255, + "loss_total_v6": 0.8853665590286255, + "step": 929 + }, + { + "epoch": 0.34662691017517705, + "grad_norm": 18.82100486755371, + "learning_rate": 8.01465780245541e-06, + "loss": 10.309015274047852, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.35923513770103455, + "loss_total_v6": 0.49790701270103455, + "step": 930 + }, + { + "epoch": 0.3469996272828923, + "grad_norm": 19.499204635620117, + "learning_rate": 8.00960839050238e-06, + "loss": 10.177362442016602, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5754716396331787, + "loss_total_v6": 0.7297685146331787, + "step": 931 + }, + { + "epoch": 0.34737234439060755, + "grad_norm": 18.932788848876953, + "learning_rate": 8.004554160666487e-06, + "loss": 10.216018676757812, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.7823003530502319, + "loss_total_v6": 0.9307378530502319, + "step": 932 + }, + { + "epoch": 0.3477450614983228, + "grad_norm": 18.697721481323242, + "learning_rate": 7.999495121038702e-06, + "loss": 10.131791114807129, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5042477250099182, + "loss_total_v6": 0.6585446000099182, + "step": 933 + }, + { + "epoch": 0.34811777860603804, + "grad_norm": 18.715866088867188, + "learning_rate": 7.994431279717715e-06, + "loss": 10.288818359375, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.38127243518829346, + "loss_total_v6": 0.5297099351882935, + "step": 934 + }, + { + "epoch": 0.34849049571375323, + "grad_norm": 19.792619705200195, + "learning_rate": 7.989362644809892e-06, + "loss": 10.109922409057617, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.33498862385749817, + "loss_total_v6": 0.48537924885749817, + "step": 935 + }, + { + "epoch": 0.3488632128214685, + "grad_norm": 18.454803466796875, + "learning_rate": 7.984289224429273e-06, + "loss": 10.374666213989258, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4405111074447632, + "loss_total_v6": 0.5928548574447632, + "step": 936 + }, + { + "epoch": 0.34923592992918373, + "grad_norm": 24.88942527770996, + "learning_rate": 7.979211026697562e-06, + "loss": 10.184667587280273, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.7288481593132019, + "loss_total_v6": 0.8890044093132019, + "step": 937 + }, + { + "epoch": 0.349608647036899, + "grad_norm": 18.652814865112305, + "learning_rate": 7.97412805974411e-06, + "loss": 10.31109619140625, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.49586644768714905, + "loss_total_v6": 0.6443039178848267, + "step": 938 + }, + { + "epoch": 0.3499813641446142, + "grad_norm": 18.91797637939453, + "learning_rate": 7.969040331705904e-06, + "loss": 9.922298431396484, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.32570409774780273, + "loss_total_v6": 0.47804784774780273, + "step": 939 + }, + { + "epoch": 0.3503540812523295, + "grad_norm": 19.8724308013916, + "learning_rate": 7.963947850727548e-06, + "loss": 10.246066093444824, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5200601816177368, + "loss_total_v6": 0.6645914316177368, + "step": 940 + }, + { + "epoch": 0.3507267983600447, + "grad_norm": 19.426496505737305, + "learning_rate": 7.95885062496126e-06, + "loss": 10.574071884155273, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.292135626077652, + "loss_total_v6": 0.432760626077652, + "step": 941 + }, + { + "epoch": 0.35109951546775997, + "grad_norm": 19.397598266601562, + "learning_rate": 7.95374866256685e-06, + "loss": 9.99898910522461, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5027691125869751, + "loss_total_v6": 0.6707378625869751, + "step": 942 + }, + { + "epoch": 0.3514722325754752, + "grad_norm": 18.91301155090332, + "learning_rate": 7.948641971711712e-06, + "loss": 10.400931358337402, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6041914820671082, + "loss_total_v6": 0.7506758570671082, + "step": 943 + }, + { + "epoch": 0.35184494968319047, + "grad_norm": 20.677881240844727, + "learning_rate": 7.943530560570809e-06, + "loss": 9.693474769592285, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5472509860992432, + "loss_total_v6": 0.6956884860992432, + "step": 944 + }, + { + "epoch": 0.3522176667909057, + "grad_norm": 18.42949676513672, + "learning_rate": 7.938414437326663e-06, + "loss": 10.494956970214844, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4484638273715973, + "loss_total_v6": 0.5988544225692749, + "step": 945 + }, + { + "epoch": 0.35259038389862096, + "grad_norm": 18.989009857177734, + "learning_rate": 7.933293610169337e-06, + "loss": 10.190771102905273, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.13397525250911713, + "loss_total_v6": 0.2765533924102783, + "step": 946 + }, + { + "epoch": 0.3529631010063362, + "grad_norm": 18.980777740478516, + "learning_rate": 7.92816808729642e-06, + "loss": 10.098220825195312, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.3848641812801361, + "loss_total_v6": 0.5469735860824585, + "step": 947 + }, + { + "epoch": 0.35333581811405146, + "grad_norm": 20.266029357910156, + "learning_rate": 7.923037876913027e-06, + "loss": 10.526601791381836, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5025836825370789, + "loss_total_v6": 0.5025836825370789, + "step": 948 + }, + { + "epoch": 0.3537085352217667, + "grad_norm": 19.996252059936523, + "learning_rate": 7.917902987231769e-06, + "loss": 10.303288459777832, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.346707820892334, + "loss_total_v6": 0.506864070892334, + "step": 949 + }, + { + "epoch": 0.3540812523294819, + "grad_norm": 19.37786102294922, + "learning_rate": 7.912763426472754e-06, + "loss": 10.513230323791504, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.11363078653812408, + "loss_total_v6": 0.2698807716369629, + "step": 950 + }, + { + "epoch": 0.35445396943719715, + "grad_norm": 19.100961685180664, + "learning_rate": 7.907619202863561e-06, + "loss": 10.205622673034668, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5690202713012695, + "loss_total_v6": 0.7252702713012695, + "step": 951 + }, + { + "epoch": 0.3548266865449124, + "grad_norm": 19.919361114501953, + "learning_rate": 7.902470324639242e-06, + "loss": 10.055856704711914, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.48598340153694153, + "loss_total_v6": 0.6363739967346191, + "step": 952 + }, + { + "epoch": 0.35519940365262764, + "grad_norm": 19.247785568237305, + "learning_rate": 7.897316800042292e-06, + "loss": 9.969009399414062, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.38259196281433105, + "loss_total_v6": 0.542748212814331, + "step": 953 + }, + { + "epoch": 0.3555721207603429, + "grad_norm": 19.0268611907959, + "learning_rate": 7.892158637322647e-06, + "loss": 9.976408004760742, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.510596752166748, + "loss_total_v6": 0.657081127166748, + "step": 954 + }, + { + "epoch": 0.35594483786805814, + "grad_norm": 20.505098342895508, + "learning_rate": 7.886995844737668e-06, + "loss": 9.923785209655762, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.21317440271377563, + "loss_total_v6": 0.38114315271377563, + "step": 955 + }, + { + "epoch": 0.3563175549757734, + "grad_norm": 19.790122985839844, + "learning_rate": 7.88182843055213e-06, + "loss": 10.03528118133545, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.636181116104126, + "loss_total_v6": 0.817821741104126, + "step": 956 + }, + { + "epoch": 0.35669027208348864, + "grad_norm": 18.351755142211914, + "learning_rate": 7.876656403038204e-06, + "loss": 10.071598052978516, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5039224624633789, + "loss_total_v6": 0.6543130874633789, + "step": 957 + }, + { + "epoch": 0.3570629891912039, + "grad_norm": 19.137693405151367, + "learning_rate": 7.871479770475445e-06, + "loss": 10.231790542602539, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5086461901664734, + "loss_total_v6": 0.5086461901664734, + "step": 958 + }, + { + "epoch": 0.35743570629891913, + "grad_norm": 19.12816619873047, + "learning_rate": 7.866298541150784e-06, + "loss": 10.127815246582031, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3201756775379181, + "loss_total_v6": 0.4686131775379181, + "step": 959 + }, + { + "epoch": 0.3578084234066344, + "grad_norm": 17.92137336730957, + "learning_rate": 7.861112723358506e-06, + "loss": 10.494773864746094, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.3379485309123993, + "loss_total_v6": 0.4922454059123993, + "step": 960 + }, + { + "epoch": 0.3581811405143496, + "grad_norm": 21.34272575378418, + "learning_rate": 7.855922325400245e-06, + "loss": 9.965108871459961, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.8338715434074402, + "loss_total_v6": 0.8338715434074402, + "step": 961 + }, + { + "epoch": 0.3585538576220649, + "grad_norm": 19.570772171020508, + "learning_rate": 7.850727355584963e-06, + "loss": 10.11489486694336, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.24879084527492523, + "loss_total_v6": 0.3913689851760864, + "step": 962 + }, + { + "epoch": 0.3589265747297801, + "grad_norm": 18.96324348449707, + "learning_rate": 7.845527822228945e-06, + "loss": 10.344245910644531, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5420868992805481, + "loss_total_v6": 0.6905243992805481, + "step": 963 + }, + { + "epoch": 0.3592992918374953, + "grad_norm": 18.3253173828125, + "learning_rate": 7.84032373365578e-06, + "loss": 10.02255630493164, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.649605393409729, + "loss_total_v6": 0.796089768409729, + "step": 964 + }, + { + "epoch": 0.35967200894521056, + "grad_norm": 18.33603858947754, + "learning_rate": 7.835115098196348e-06, + "loss": 10.432304382324219, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5790025591850281, + "loss_total_v6": 0.7332994341850281, + "step": 965 + }, + { + "epoch": 0.3600447260529258, + "grad_norm": 20.730897903442383, + "learning_rate": 7.82990192418881e-06, + "loss": 10.03276538848877, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5039575099945068, + "loss_total_v6": 0.6523950099945068, + "step": 966 + }, + { + "epoch": 0.36041744316064106, + "grad_norm": 20.120840072631836, + "learning_rate": 7.824684219978591e-06, + "loss": 10.060653686523438, + "loss_alignment": 0.26171875, + "loss_alignment_w": 0.130859375, + "loss_sft": 0.47984787821769714, + "loss_total_v6": 0.6107072830200195, + "step": 967 + }, + { + "epoch": 0.3607901602683563, + "grad_norm": 18.00377082824707, + "learning_rate": 7.819461993918367e-06, + "loss": 10.472599029541016, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6992887258529663, + "loss_total_v6": 0.8418668508529663, + "step": 968 + }, + { + "epoch": 0.36116287737607156, + "grad_norm": 18.786548614501953, + "learning_rate": 7.814235254368054e-06, + "loss": 10.271588325500488, + "loss_alignment": 0.38671875, + "loss_alignment_w": 0.193359375, + "loss_sft": 0.8098167181015015, + "loss_total_v6": 1.0031760931015015, + "step": 969 + }, + { + "epoch": 0.3615355944837868, + "grad_norm": 17.824487686157227, + "learning_rate": 7.809004009694795e-06, + "loss": 10.48410415649414, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.48017066717147827, + "loss_total_v6": 0.6305612921714783, + "step": 970 + }, + { + "epoch": 0.36190831159150205, + "grad_norm": 19.07480239868164, + "learning_rate": 7.803768268272945e-06, + "loss": 10.377117156982422, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4584762752056122, + "loss_total_v6": 0.6166794300079346, + "step": 971 + }, + { + "epoch": 0.3622810286992173, + "grad_norm": 18.037372589111328, + "learning_rate": 7.79852803848405e-06, + "loss": 10.047541618347168, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.3512430787086487, + "loss_total_v6": 0.5035868287086487, + "step": 972 + }, + { + "epoch": 0.36265374580693255, + "grad_norm": 19.505308151245117, + "learning_rate": 7.793283328716853e-06, + "loss": 10.382509231567383, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5628038048744202, + "loss_total_v6": 0.7092881798744202, + "step": 973 + }, + { + "epoch": 0.3630264629146478, + "grad_norm": 19.341995239257812, + "learning_rate": 7.78803414736726e-06, + "loss": 10.077085494995117, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.44404977560043335, + "loss_total_v6": 0.6022529006004333, + "step": 974 + }, + { + "epoch": 0.36339918002236304, + "grad_norm": 18.611698150634766, + "learning_rate": 7.782780502838337e-06, + "loss": 9.845462799072266, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5459567904472351, + "loss_total_v6": 0.6904880404472351, + "step": 975 + }, + { + "epoch": 0.3637718971300783, + "grad_norm": 18.79244613647461, + "learning_rate": 7.7775224035403e-06, + "loss": 10.374990463256836, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3718017339706421, + "loss_total_v6": 0.5300048589706421, + "step": 976 + }, + { + "epoch": 0.36414461423779354, + "grad_norm": 18.816709518432617, + "learning_rate": 7.772259857890485e-06, + "loss": 10.02558422088623, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4983406066894531, + "loss_total_v6": 0.6409187316894531, + "step": 977 + }, + { + "epoch": 0.36451733134550873, + "grad_norm": 19.457059860229492, + "learning_rate": 7.76699287431336e-06, + "loss": 9.749518394470215, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5488371849060059, + "loss_total_v6": 0.7011809349060059, + "step": 978 + }, + { + "epoch": 0.364890048453224, + "grad_norm": 19.20989227294922, + "learning_rate": 7.761721461240487e-06, + "loss": 10.086663246154785, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6625809669494629, + "loss_total_v6": 0.8246903419494629, + "step": 979 + }, + { + "epoch": 0.36526276556093923, + "grad_norm": 19.155302047729492, + "learning_rate": 7.756445627110523e-06, + "loss": 10.240860939025879, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3040061593055725, + "loss_total_v6": 0.4524436593055725, + "step": 980 + }, + { + "epoch": 0.3656354826686545, + "grad_norm": 19.578266143798828, + "learning_rate": 7.7511653803692e-06, + "loss": 10.30836296081543, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.43136948347091675, + "loss_total_v6": 0.5837132334709167, + "step": 981 + }, + { + "epoch": 0.3660081997763697, + "grad_norm": 18.44889259338379, + "learning_rate": 7.74588072946932e-06, + "loss": 10.168623924255371, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.35022613406181335, + "loss_total_v6": 0.49671050906181335, + "step": 982 + }, + { + "epoch": 0.366380916884085, + "grad_norm": 19.856307983398438, + "learning_rate": 7.740591682870725e-06, + "loss": 10.206247329711914, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6288147568702698, + "loss_total_v6": 0.7928772568702698, + "step": 983 + }, + { + "epoch": 0.3667536339918002, + "grad_norm": 18.188873291015625, + "learning_rate": 7.735298249040304e-06, + "loss": 9.900286674499512, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3179844617843628, + "loss_total_v6": 0.4683750867843628, + "step": 984 + }, + { + "epoch": 0.36712635109951547, + "grad_norm": 17.852949142456055, + "learning_rate": 7.730000436451963e-06, + "loss": 9.91226863861084, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5865423083305359, + "loss_total_v6": 0.7408391833305359, + "step": 985 + }, + { + "epoch": 0.3674990682072307, + "grad_norm": 18.864398956298828, + "learning_rate": 7.72469825358662e-06, + "loss": 10.464982986450195, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.8924365043640137, + "loss_total_v6": 1.0428271293640137, + "step": 986 + }, + { + "epoch": 0.36787178531494596, + "grad_norm": 19.74420928955078, + "learning_rate": 7.719391708932189e-06, + "loss": 9.831573486328125, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4716441035270691, + "loss_total_v6": 0.6298472285270691, + "step": 987 + }, + { + "epoch": 0.3682445024226612, + "grad_norm": 18.32442855834961, + "learning_rate": 7.714080810983564e-06, + "loss": 9.718988418579102, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.365475594997406, + "loss_total_v6": 0.519772469997406, + "step": 988 + }, + { + "epoch": 0.36861721953037646, + "grad_norm": 17.308513641357422, + "learning_rate": 7.708765568242613e-06, + "loss": 10.434839248657227, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5934789180755615, + "loss_total_v6": 0.5934789180755615, + "step": 989 + }, + { + "epoch": 0.3689899366380917, + "grad_norm": 19.58819007873535, + "learning_rate": 7.703445989218154e-06, + "loss": 10.158392906188965, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.7174656987190247, + "loss_total_v6": 0.8737156987190247, + "step": 990 + }, + { + "epoch": 0.36936265374580696, + "grad_norm": 18.298166275024414, + "learning_rate": 7.69812208242595e-06, + "loss": 9.765239715576172, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.8495590090751648, + "loss_total_v6": 0.9979965090751648, + "step": 991 + }, + { + "epoch": 0.36973537085352215, + "grad_norm": 18.392410278320312, + "learning_rate": 7.692793856388692e-06, + "loss": 10.0471773147583, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7368810772895813, + "loss_total_v6": 0.9028967022895813, + "step": 992 + }, + { + "epoch": 0.3701080879612374, + "grad_norm": 18.056684494018555, + "learning_rate": 7.68746131963598e-06, + "loss": 10.305004119873047, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.8237742781639099, + "loss_total_v6": 0.9819774031639099, + "step": 993 + }, + { + "epoch": 0.37048080506895265, + "grad_norm": 19.221418380737305, + "learning_rate": 7.682124480704328e-06, + "loss": 10.236331939697266, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7258742451667786, + "loss_total_v6": 0.8782179951667786, + "step": 994 + }, + { + "epoch": 0.3708535221766679, + "grad_norm": 19.802392959594727, + "learning_rate": 7.676783348137119e-06, + "loss": 10.469035148620605, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5388534069061279, + "loss_total_v6": 0.6931502819061279, + "step": 995 + }, + { + "epoch": 0.37122623928438314, + "grad_norm": 18.288528442382812, + "learning_rate": 7.671437930484622e-06, + "loss": 10.174678802490234, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.42859113216400146, + "loss_total_v6": 0.5711692571640015, + "step": 996 + }, + { + "epoch": 0.3715989563920984, + "grad_norm": 18.844600677490234, + "learning_rate": 7.66608823630396e-06, + "loss": 10.122669219970703, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.4320946931838989, + "loss_total_v6": 0.6020165681838989, + "step": 997 + }, + { + "epoch": 0.37197167349981364, + "grad_norm": 20.42418098449707, + "learning_rate": 7.660734274159112e-06, + "loss": 10.118670463562012, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4648175835609436, + "loss_total_v6": 0.4648175835609436, + "step": 998 + }, + { + "epoch": 0.3723443906075289, + "grad_norm": 18.239110946655273, + "learning_rate": 7.655376052620874e-06, + "loss": 10.068621635437012, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.28997617959976196, + "loss_total_v6": 0.43841367959976196, + "step": 999 + }, + { + "epoch": 0.37271710771524413, + "grad_norm": 18.429149627685547, + "learning_rate": 7.650013580266871e-06, + "loss": 10.432504653930664, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.30677568912506104, + "loss_total_v6": 0.45130693912506104, + "step": 1000 + }, + { + "epoch": 0.37271710771524413, + "eval_loss": 0.6176013946533203, + "eval_loss_alignment": 0.2834284855769231, + "eval_loss_alignment_w": 0.14171424278846154, + "eval_loss_sft": 0.47588716270640874, + "eval_loss_total_v6": 0.6176014060679919, + "eval_runtime": 207.0324, + "eval_samples_per_second": 9.013, + "eval_steps_per_second": 1.13, + "step": 1000 + }, + { + "epoch": 0.3730898248229594, + "grad_norm": 18.535198211669922, + "learning_rate": 7.644646865681528e-06, + "loss": 10.152164459228516, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6945264935493469, + "loss_total_v6": 0.8546827435493469, + "step": 1001 + }, + { + "epoch": 0.37346254193067463, + "grad_norm": 18.560340881347656, + "learning_rate": 7.639275917456068e-06, + "loss": 9.946110725402832, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.46661320328712463, + "loss_total_v6": 0.6306756734848022, + "step": 1002 + }, + { + "epoch": 0.3738352590383899, + "grad_norm": 19.039064407348633, + "learning_rate": 7.63390074418848e-06, + "loss": 10.63829231262207, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6224868893623352, + "loss_total_v6": 0.7826431393623352, + "step": 1003 + }, + { + "epoch": 0.3742079761461051, + "grad_norm": 18.203027725219727, + "learning_rate": 7.628521354483528e-06, + "loss": 10.30445671081543, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4702269434928894, + "loss_total_v6": 0.6206175684928894, + "step": 1004 + }, + { + "epoch": 0.3745806932538204, + "grad_norm": 20.644760131835938, + "learning_rate": 7.623137756952719e-06, + "loss": 10.219537734985352, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6199581623077393, + "loss_total_v6": 0.7683956623077393, + "step": 1005 + }, + { + "epoch": 0.3749534103615356, + "grad_norm": 18.193593978881836, + "learning_rate": 7.6177499602143e-06, + "loss": 10.534387588500977, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.8584583401679993, + "loss_total_v6": 1.0361926555633545, + "step": 1006 + }, + { + "epoch": 0.3753261274692508, + "grad_norm": 18.067346572875977, + "learning_rate": 7.612357972893234e-06, + "loss": 10.219295501708984, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4267261028289795, + "loss_total_v6": 0.5771167278289795, + "step": 1007 + }, + { + "epoch": 0.37569884457696606, + "grad_norm": 17.979610443115234, + "learning_rate": 7.606961803621201e-06, + "loss": 10.09583568572998, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.32148852944374084, + "loss_total_v6": 0.45820727944374084, + "step": 1008 + }, + { + "epoch": 0.3760715616846813, + "grad_norm": 18.199871063232422, + "learning_rate": 7.601561461036568e-06, + "loss": 10.673809051513672, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.28476354479789734, + "loss_total_v6": 0.42734166979789734, + "step": 1009 + }, + { + "epoch": 0.37644427879239656, + "grad_norm": 19.02595329284668, + "learning_rate": 7.596156953784388e-06, + "loss": 9.80797290802002, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.27641499042510986, + "loss_total_v6": 0.43071186542510986, + "step": 1010 + }, + { + "epoch": 0.3768169959001118, + "grad_norm": 17.173707962036133, + "learning_rate": 7.590748290516379e-06, + "loss": 10.117758750915527, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.17713436484336853, + "loss_total_v6": 0.32361873984336853, + "step": 1011 + }, + { + "epoch": 0.37718971300782705, + "grad_norm": 18.269060134887695, + "learning_rate": 7.58533547989091e-06, + "loss": 10.20527458190918, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.38743457198143005, + "loss_total_v6": 0.5241533517837524, + "step": 1012 + }, + { + "epoch": 0.3775624301155423, + "grad_norm": 22.198421478271484, + "learning_rate": 7.579918530572994e-06, + "loss": 10.338647842407227, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6315038204193115, + "loss_total_v6": 0.7838475704193115, + "step": 1013 + }, + { + "epoch": 0.37793514722325755, + "grad_norm": 18.862001419067383, + "learning_rate": 7.574497451234264e-06, + "loss": 9.566008567810059, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.5674557089805603, + "loss_total_v6": 0.7334713339805603, + "step": 1014 + }, + { + "epoch": 0.3783078643309728, + "grad_norm": 18.425567626953125, + "learning_rate": 7.569072250552968e-06, + "loss": 9.715806007385254, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5922657251358032, + "loss_total_v6": 0.7426563501358032, + "step": 1015 + }, + { + "epoch": 0.37868058143868805, + "grad_norm": 18.50400161743164, + "learning_rate": 7.563642937213952e-06, + "loss": 10.061700820922852, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5685957074165344, + "loss_total_v6": 0.7209394574165344, + "step": 1016 + }, + { + "epoch": 0.3790532985464033, + "grad_norm": 20.126611709594727, + "learning_rate": 7.558209519908644e-06, + "loss": 10.431958198547363, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.46826857328414917, + "loss_total_v6": 0.6049873232841492, + "step": 1017 + }, + { + "epoch": 0.37942601565411854, + "grad_norm": 18.029102325439453, + "learning_rate": 7.5527720073350385e-06, + "loss": 10.247882843017578, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.35908231139183044, + "loss_total_v6": 0.5114260911941528, + "step": 1018 + }, + { + "epoch": 0.3797987327618338, + "grad_norm": 19.11174201965332, + "learning_rate": 7.547330408197695e-06, + "loss": 10.225922584533691, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.435059130191803, + "loss_total_v6": 0.597168505191803, + "step": 1019 + }, + { + "epoch": 0.38017144986954904, + "grad_norm": 21.312036514282227, + "learning_rate": 7.541884731207707e-06, + "loss": 10.671725273132324, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3296787142753601, + "loss_total_v6": 0.4742099642753601, + "step": 1020 + }, + { + "epoch": 0.38054416697726423, + "grad_norm": 20.287574768066406, + "learning_rate": 7.536434985082696e-06, + "loss": 10.574339866638184, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.29934948682785034, + "loss_total_v6": 0.44778698682785034, + "step": 1021 + }, + { + "epoch": 0.3809168840849795, + "grad_norm": 18.85419464111328, + "learning_rate": 7.5309811785468035e-06, + "loss": 9.66445255279541, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.42831218242645264, + "loss_total_v6": 0.5943278074264526, + "step": 1022 + }, + { + "epoch": 0.3812896011926947, + "grad_norm": 18.402908325195312, + "learning_rate": 7.5255233203306665e-06, + "loss": 10.439735412597656, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.43165165185928345, + "loss_total_v6": 0.5839954018592834, + "step": 1023 + }, + { + "epoch": 0.38166231830041, + "grad_norm": 18.579744338989258, + "learning_rate": 7.520061419171407e-06, + "loss": 10.406330108642578, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5403175354003906, + "loss_total_v6": 0.6946144104003906, + "step": 1024 + }, + { + "epoch": 0.3820350354081252, + "grad_norm": 19.298267364501953, + "learning_rate": 7.514595483812624e-06, + "loss": 9.980335235595703, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.1916971355676651, + "loss_total_v6": 0.3342752456665039, + "step": 1025 + }, + { + "epoch": 0.38240775251584047, + "grad_norm": 19.19319725036621, + "learning_rate": 7.509125523004368e-06, + "loss": 10.141075134277344, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5446619391441345, + "loss_total_v6": 0.6911463141441345, + "step": 1026 + }, + { + "epoch": 0.3827804696235557, + "grad_norm": 19.813262939453125, + "learning_rate": 7.5036515455031415e-06, + "loss": 10.490936279296875, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5015360116958618, + "loss_total_v6": 0.6616922616958618, + "step": 1027 + }, + { + "epoch": 0.38315318673127097, + "grad_norm": 19.936281204223633, + "learning_rate": 7.4981735600718685e-06, + "loss": 9.952556610107422, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6652848124504089, + "loss_total_v6": 0.8156754374504089, + "step": 1028 + }, + { + "epoch": 0.3835259038389862, + "grad_norm": 17.970413208007812, + "learning_rate": 7.492691575479896e-06, + "loss": 9.833782196044922, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.20649653673171997, + "loss_total_v6": 0.36860591173171997, + "step": 1029 + }, + { + "epoch": 0.38389862094670146, + "grad_norm": 19.117265701293945, + "learning_rate": 7.48720560050297e-06, + "loss": 10.562263488769531, + "loss_alignment": 0.34375, + "loss_alignment_w": 0.171875, + "loss_sft": 0.7585457563400269, + "loss_total_v6": 0.9304207563400269, + "step": 1030 + }, + { + "epoch": 0.3842713380544167, + "grad_norm": 18.441761016845703, + "learning_rate": 7.481715643923225e-06, + "loss": 10.20582389831543, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6247785091400146, + "loss_total_v6": 0.7771222591400146, + "step": 1031 + }, + { + "epoch": 0.38464405516213196, + "grad_norm": 19.205568313598633, + "learning_rate": 7.476221714529167e-06, + "loss": 10.458718299865723, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.8052559494972229, + "loss_total_v6": 0.9654121994972229, + "step": 1032 + }, + { + "epoch": 0.3850167722698472, + "grad_norm": 19.669878005981445, + "learning_rate": 7.470723821115666e-06, + "loss": 10.43705940246582, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5031907558441162, + "loss_total_v6": 0.6516282558441162, + "step": 1033 + }, + { + "epoch": 0.38538948937756246, + "grad_norm": 18.9105224609375, + "learning_rate": 7.465221972483936e-06, + "loss": 10.134952545166016, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5411601066589355, + "loss_total_v6": 0.7091288566589355, + "step": 1034 + }, + { + "epoch": 0.38576220648527765, + "grad_norm": 18.916954040527344, + "learning_rate": 7.45971617744152e-06, + "loss": 9.985311508178711, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3808766305446625, + "loss_total_v6": 0.5293141603469849, + "step": 1035 + }, + { + "epoch": 0.3861349235929929, + "grad_norm": 18.542926788330078, + "learning_rate": 7.454206444802283e-06, + "loss": 10.044458389282227, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.29127708077430725, + "loss_total_v6": 0.43776145577430725, + "step": 1036 + }, + { + "epoch": 0.38650764070070814, + "grad_norm": 18.618070602416992, + "learning_rate": 7.448692783386391e-06, + "loss": 10.168730735778809, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.1433953046798706, + "loss_total_v6": 0.2937859296798706, + "step": 1037 + }, + { + "epoch": 0.3868803578084234, + "grad_norm": 19.884614944458008, + "learning_rate": 7.443175202020297e-06, + "loss": 10.30300235748291, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.15260322391986847, + "loss_total_v6": 0.2990875840187073, + "step": 1038 + }, + { + "epoch": 0.38725307491613864, + "grad_norm": 20.087697982788086, + "learning_rate": 7.437653709536735e-06, + "loss": 10.123648643493652, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6287940740585327, + "loss_total_v6": 0.7889503240585327, + "step": 1039 + }, + { + "epoch": 0.3876257920238539, + "grad_norm": 18.337465286254883, + "learning_rate": 7.432128314774696e-06, + "loss": 10.180351257324219, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.44781607389450073, + "loss_total_v6": 0.5962535738945007, + "step": 1040 + }, + { + "epoch": 0.38799850913156914, + "grad_norm": 19.159732818603516, + "learning_rate": 7.426599026579419e-06, + "loss": 10.306532859802246, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.46503522992134094, + "loss_total_v6": 0.6251914501190186, + "step": 1041 + }, + { + "epoch": 0.3883712262392844, + "grad_norm": 19.890195846557617, + "learning_rate": 7.421065853802375e-06, + "loss": 10.391267776489258, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.32920071482658386, + "loss_total_v6": 0.47177883982658386, + "step": 1042 + }, + { + "epoch": 0.38874394334699963, + "grad_norm": 19.37592124938965, + "learning_rate": 7.415528805301254e-06, + "loss": 10.159322738647461, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.35302841663360596, + "loss_total_v6": 0.509278416633606, + "step": 1043 + }, + { + "epoch": 0.3891166604547149, + "grad_norm": 19.865880966186523, + "learning_rate": 7.4099878899399506e-06, + "loss": 10.350104331970215, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.40838152170181274, + "loss_total_v6": 0.5626783967018127, + "step": 1044 + }, + { + "epoch": 0.38948937756243013, + "grad_norm": 20.27294158935547, + "learning_rate": 7.404443116588548e-06, + "loss": 10.001806259155273, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5777188539505005, + "loss_total_v6": 0.7281094789505005, + "step": 1045 + }, + { + "epoch": 0.3898620946701454, + "grad_norm": 19.1053524017334, + "learning_rate": 7.3988944941233095e-06, + "loss": 10.14964771270752, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.7851765751838684, + "loss_total_v6": 0.9511922001838684, + "step": 1046 + }, + { + "epoch": 0.3902348117778606, + "grad_norm": 19.072223663330078, + "learning_rate": 7.393342031426656e-06, + "loss": 9.852804183959961, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.8351857662200928, + "loss_total_v6": 0.9836232662200928, + "step": 1047 + }, + { + "epoch": 0.39060752888557587, + "grad_norm": 18.51904296875, + "learning_rate": 7.3877857373871585e-06, + "loss": 9.815811157226562, + "loss_alignment": 0.2578125, + "loss_alignment_w": 0.12890625, + "loss_sft": 0.3814171254634857, + "loss_total_v6": 0.5103234052658081, + "step": 1048 + }, + { + "epoch": 0.39098024599329106, + "grad_norm": 18.675724029541016, + "learning_rate": 7.382225620899519e-06, + "loss": 9.979406356811523, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.20317162573337555, + "loss_total_v6": 0.35746848583221436, + "step": 1049 + }, + { + "epoch": 0.3913529631010063, + "grad_norm": 18.492347717285156, + "learning_rate": 7.37666169086456e-06, + "loss": 10.363842010498047, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4169122576713562, + "loss_total_v6": 0.5712091326713562, + "step": 1050 + }, + { + "epoch": 0.39172568020872156, + "grad_norm": 19.217247009277344, + "learning_rate": 7.371093956189211e-06, + "loss": 10.163841247558594, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8253976106643677, + "loss_total_v6": 0.9796944856643677, + "step": 1051 + }, + { + "epoch": 0.3920983973164368, + "grad_norm": 18.77161979675293, + "learning_rate": 7.365522425786487e-06, + "loss": 10.322860717773438, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5404940843582153, + "loss_total_v6": 0.7006503343582153, + "step": 1052 + }, + { + "epoch": 0.39247111442415206, + "grad_norm": 19.635648727416992, + "learning_rate": 7.359947108575485e-06, + "loss": 10.301475524902344, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.06960736960172653, + "loss_total_v6": 0.21609175205230713, + "step": 1053 + }, + { + "epoch": 0.3928438315318673, + "grad_norm": 17.7391357421875, + "learning_rate": 7.354368013481359e-06, + "loss": 10.178421020507812, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.46684959530830383, + "loss_total_v6": 0.6191933155059814, + "step": 1054 + }, + { + "epoch": 0.39321654863958255, + "grad_norm": 18.4008731842041, + "learning_rate": 7.348785149435315e-06, + "loss": 9.930498123168945, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.7162768840789795, + "loss_total_v6": 0.8647143840789795, + "step": 1055 + }, + { + "epoch": 0.3935892657472978, + "grad_norm": 18.341176986694336, + "learning_rate": 7.343198525374588e-06, + "loss": 10.15871524810791, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5314085483551025, + "loss_total_v6": 0.6798460483551025, + "step": 1056 + }, + { + "epoch": 0.39396198285501305, + "grad_norm": 18.27481460571289, + "learning_rate": 7.337608150242438e-06, + "loss": 10.16711139678955, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5992483496665955, + "loss_total_v6": 0.7457327246665955, + "step": 1057 + }, + { + "epoch": 0.3943346999627283, + "grad_norm": 20.105545043945312, + "learning_rate": 7.332014032988123e-06, + "loss": 10.532638549804688, + "loss_alignment": 0.3984375, + "loss_alignment_w": 0.19921875, + "loss_sft": 0.5558859705924988, + "loss_total_v6": 0.7551047205924988, + "step": 1058 + }, + { + "epoch": 0.39470741707044354, + "grad_norm": 18.974985122680664, + "learning_rate": 7.326416182566897e-06, + "loss": 9.984166145324707, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.41301512718200684, + "loss_total_v6": 0.5653588771820068, + "step": 1059 + }, + { + "epoch": 0.3950801341781588, + "grad_norm": 18.645709991455078, + "learning_rate": 7.3208146079399885e-06, + "loss": 10.022509574890137, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.45253247022628784, + "loss_total_v6": 0.6009699702262878, + "step": 1060 + }, + { + "epoch": 0.39545285128587404, + "grad_norm": 18.781862258911133, + "learning_rate": 7.315209318074587e-06, + "loss": 10.005636215209961, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4963228404521942, + "loss_total_v6": 0.6584322452545166, + "step": 1061 + }, + { + "epoch": 0.3958255683935893, + "grad_norm": 18.074256896972656, + "learning_rate": 7.30960032194383e-06, + "loss": 9.975028991699219, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5882519483566284, + "loss_total_v6": 0.5882519483566284, + "step": 1062 + }, + { + "epoch": 0.39619828550130454, + "grad_norm": 18.868499755859375, + "learning_rate": 7.303987628526789e-06, + "loss": 9.844830513000488, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4546624422073364, + "loss_total_v6": 0.5972405672073364, + "step": 1063 + }, + { + "epoch": 0.39657100260901973, + "grad_norm": 19.00550651550293, + "learning_rate": 7.298371246808451e-06, + "loss": 10.10534954071045, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6359749436378479, + "loss_total_v6": 0.7902718186378479, + "step": 1064 + }, + { + "epoch": 0.396943719716735, + "grad_norm": 18.438798904418945, + "learning_rate": 7.292751185779713e-06, + "loss": 10.576850891113281, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.798975944519043, + "loss_total_v6": 0.980616569519043, + "step": 1065 + }, + { + "epoch": 0.3973164368244502, + "grad_norm": 18.515867233276367, + "learning_rate": 7.287127454437356e-06, + "loss": 9.952792167663574, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.1590363085269928, + "loss_total_v6": 0.3074738085269928, + "step": 1066 + }, + { + "epoch": 0.3976891539321655, + "grad_norm": 18.21072006225586, + "learning_rate": 7.2815000617840406e-06, + "loss": 10.003227233886719, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.41852831840515137, + "loss_total_v6": 0.5611064434051514, + "step": 1067 + }, + { + "epoch": 0.3980618710398807, + "grad_norm": 18.55438232421875, + "learning_rate": 7.275869016828288e-06, + "loss": 10.438294410705566, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5022625923156738, + "loss_total_v6": 0.6507000923156738, + "step": 1068 + }, + { + "epoch": 0.39843458814759597, + "grad_norm": 19.356199264526367, + "learning_rate": 7.270234328584463e-06, + "loss": 10.567673683166504, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4337080419063568, + "loss_total_v6": 0.5782393217086792, + "step": 1069 + }, + { + "epoch": 0.3988073052553112, + "grad_norm": 19.421140670776367, + "learning_rate": 7.264596006072766e-06, + "loss": 10.43764877319336, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.363209068775177, + "loss_total_v6": 0.509693443775177, + "step": 1070 + }, + { + "epoch": 0.39918002236302647, + "grad_norm": 19.69707489013672, + "learning_rate": 7.2589540583192165e-06, + "loss": 9.728747367858887, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5201717615127563, + "loss_total_v6": 0.6764217615127563, + "step": 1071 + }, + { + "epoch": 0.3995527394707417, + "grad_norm": 18.392379760742188, + "learning_rate": 7.253308494355633e-06, + "loss": 10.235330581665039, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.380368173122406, + "loss_total_v6": 0.530758798122406, + "step": 1072 + }, + { + "epoch": 0.39992545657845696, + "grad_norm": 18.25395393371582, + "learning_rate": 7.247659323219624e-06, + "loss": 10.311935424804688, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4454704225063324, + "loss_total_v6": 0.5880485773086548, + "step": 1073 + }, + { + "epoch": 0.4002981736861722, + "grad_norm": 18.430164337158203, + "learning_rate": 7.242006553954577e-06, + "loss": 9.94454574584961, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.45271673798561096, + "loss_total_v6": 0.5933417081832886, + "step": 1074 + }, + { + "epoch": 0.40067089079388746, + "grad_norm": 19.160865783691406, + "learning_rate": 7.236350195609634e-06, + "loss": 10.223634719848633, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.41645097732543945, + "loss_total_v6": 0.6019978523254395, + "step": 1075 + }, + { + "epoch": 0.4010436079016027, + "grad_norm": 18.07987403869629, + "learning_rate": 7.230690257239684e-06, + "loss": 10.015888214111328, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.5066972374916077, + "loss_total_v6": 0.6746659874916077, + "step": 1076 + }, + { + "epoch": 0.40141632500931795, + "grad_norm": 18.1564998626709, + "learning_rate": 7.22502674790535e-06, + "loss": 10.04794979095459, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6190998554229736, + "loss_total_v6": 0.7792561054229736, + "step": 1077 + }, + { + "epoch": 0.40178904211703315, + "grad_norm": 17.894447326660156, + "learning_rate": 7.219359676672969e-06, + "loss": 10.055662155151367, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.4854587912559509, + "loss_total_v6": 0.6260837912559509, + "step": 1078 + }, + { + "epoch": 0.4021617592247484, + "grad_norm": 19.115951538085938, + "learning_rate": 7.213689052614576e-06, + "loss": 10.426483154296875, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2987864017486572, + "loss_total_v6": 0.4511301517486572, + "step": 1079 + }, + { + "epoch": 0.40253447633246364, + "grad_norm": 19.10027313232422, + "learning_rate": 7.208014884807903e-06, + "loss": 10.232072830200195, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.3954317271709442, + "loss_total_v6": 0.5380098819732666, + "step": 1080 + }, + { + "epoch": 0.4029071934401789, + "grad_norm": 18.282150268554688, + "learning_rate": 7.20233718233635e-06, + "loss": 10.078231811523438, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.41607287526130676, + "loss_total_v6": 0.5723228454589844, + "step": 1081 + }, + { + "epoch": 0.40327991054789414, + "grad_norm": 18.934316635131836, + "learning_rate": 7.196655954288969e-06, + "loss": 9.920389175415039, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.26883211731910706, + "loss_total_v6": 0.42508211731910706, + "step": 1082 + }, + { + "epoch": 0.4036526276556094, + "grad_norm": 18.357797622680664, + "learning_rate": 7.190971209760469e-06, + "loss": 10.074283599853516, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6970258951187134, + "loss_total_v6": 0.8532758951187134, + "step": 1083 + }, + { + "epoch": 0.40402534476332463, + "grad_norm": 20.02956199645996, + "learning_rate": 7.185282957851175e-06, + "loss": 10.508424758911133, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.21569903194904327, + "loss_total_v6": 0.3680427670478821, + "step": 1084 + }, + { + "epoch": 0.4043980618710399, + "grad_norm": 19.6437931060791, + "learning_rate": 7.179591207667038e-06, + "loss": 10.167834281921387, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.47315555810928345, + "loss_total_v6": 0.6215930581092834, + "step": 1085 + }, + { + "epoch": 0.40477077897875513, + "grad_norm": 18.499502182006836, + "learning_rate": 7.173895968319601e-06, + "loss": 10.371898651123047, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.6225690245628357, + "loss_total_v6": 0.8042096495628357, + "step": 1086 + }, + { + "epoch": 0.4051434960864704, + "grad_norm": 18.13855743408203, + "learning_rate": 7.168197248925996e-06, + "loss": 9.887114524841309, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.7062813639640808, + "loss_total_v6": 0.8742501139640808, + "step": 1087 + }, + { + "epoch": 0.4055162131941856, + "grad_norm": 18.271848678588867, + "learning_rate": 7.162495058608928e-06, + "loss": 9.953078269958496, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2766216993331909, + "loss_total_v6": 0.4270123243331909, + "step": 1088 + }, + { + "epoch": 0.4058889303019009, + "grad_norm": 18.666563034057617, + "learning_rate": 7.156789406496655e-06, + "loss": 10.114616394042969, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.27833595871925354, + "loss_total_v6": 0.43263283371925354, + "step": 1089 + }, + { + "epoch": 0.4062616474096161, + "grad_norm": 19.967729568481445, + "learning_rate": 7.151080301722976e-06, + "loss": 9.889925003051758, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5216626524925232, + "loss_total_v6": 0.6701001524925232, + "step": 1090 + }, + { + "epoch": 0.40663436451733137, + "grad_norm": 18.55568504333496, + "learning_rate": 7.14536775342722e-06, + "loss": 10.15947151184082, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6536261439323425, + "loss_total_v6": 0.8176886439323425, + "step": 1091 + }, + { + "epoch": 0.40700708162504656, + "grad_norm": 20.11887550354004, + "learning_rate": 7.13965177075423e-06, + "loss": 9.91783332824707, + "loss_alignment": 0.25390625, + "loss_alignment_w": 0.126953125, + "loss_sft": 0.5129448771476746, + "loss_total_v6": 0.6398980021476746, + "step": 1092 + }, + { + "epoch": 0.4073797987327618, + "grad_norm": 18.695579528808594, + "learning_rate": 7.1339323628543385e-06, + "loss": 9.369267463684082, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.6541772484779358, + "loss_total_v6": 0.8065209984779358, + "step": 1093 + }, + { + "epoch": 0.40775251584047706, + "grad_norm": 18.33115577697754, + "learning_rate": 7.128209538883373e-06, + "loss": 9.656522750854492, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.45857399702072144, + "loss_total_v6": 0.6031052470207214, + "step": 1094 + }, + { + "epoch": 0.4081252329481923, + "grad_norm": 18.653392791748047, + "learning_rate": 7.1224833080026216e-06, + "loss": 9.899252891540527, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4501936733722687, + "loss_total_v6": 0.4501936733722687, + "step": 1095 + }, + { + "epoch": 0.40849795005590755, + "grad_norm": 19.27971649169922, + "learning_rate": 7.116753679378826e-06, + "loss": 10.063419342041016, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.10560417175292969, + "loss_total_v6": 0.2501354217529297, + "step": 1096 + }, + { + "epoch": 0.4088706671636228, + "grad_norm": 19.602426528930664, + "learning_rate": 7.111020662184174e-06, + "loss": 10.094999313354492, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3947664201259613, + "loss_total_v6": 0.5392976999282837, + "step": 1097 + }, + { + "epoch": 0.40924338427133805, + "grad_norm": 18.710155487060547, + "learning_rate": 7.105284265596269e-06, + "loss": 10.290708541870117, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3439589738845825, + "loss_total_v6": 0.5002089738845825, + "step": 1098 + }, + { + "epoch": 0.4096161013790533, + "grad_norm": 20.036184310913086, + "learning_rate": 7.09954449879813e-06, + "loss": 10.137693405151367, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.4563947021961212, + "loss_total_v6": 0.6204571723937988, + "step": 1099 + }, + { + "epoch": 0.40998881848676855, + "grad_norm": 19.324323654174805, + "learning_rate": 7.093801370978171e-06, + "loss": 10.329985618591309, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.19161663949489594, + "loss_total_v6": 0.34591352939605713, + "step": 1100 + }, + { + "epoch": 0.40998881848676855, + "eval_loss": 0.6141290664672852, + "eval_loss_alignment": 0.28316973824786323, + "eval_loss_alignment_w": 0.14158486912393162, + "eval_loss_sft": 0.47254419755827415, + "eval_loss_total_v6": 0.6141290663080847, + "eval_runtime": 206.8187, + "eval_samples_per_second": 9.022, + "eval_steps_per_second": 1.131, + "step": 1100 + }, + { + "epoch": 0.4103615355944838, + "grad_norm": 19.206857681274414, + "learning_rate": 7.0880548913301835e-06, + "loss": 10.305269241333008, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.2626798152923584, + "loss_total_v6": 0.4033048152923584, + "step": 1101 + }, + { + "epoch": 0.41073425270219904, + "grad_norm": 18.500625610351562, + "learning_rate": 7.082305069053328e-06, + "loss": 10.670758247375488, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5647034645080566, + "loss_total_v6": 0.7229065895080566, + "step": 1102 + }, + { + "epoch": 0.4111069698099143, + "grad_norm": 20.688016891479492, + "learning_rate": 7.076551913352113e-06, + "loss": 9.80243968963623, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3342362344264984, + "loss_total_v6": 0.4943924844264984, + "step": 1103 + }, + { + "epoch": 0.41147968691762954, + "grad_norm": 18.135223388671875, + "learning_rate": 7.070795433436386e-06, + "loss": 9.921445846557617, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4283420145511627, + "loss_total_v6": 0.5728732347488403, + "step": 1104 + }, + { + "epoch": 0.4118524040253448, + "grad_norm": 18.463947296142578, + "learning_rate": 7.065035638521312e-06, + "loss": 10.193368911743164, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.46209338307380676, + "loss_total_v6": 0.6476402282714844, + "step": 1105 + }, + { + "epoch": 0.41222512113306004, + "grad_norm": 19.27543067932129, + "learning_rate": 7.05927253782737e-06, + "loss": 10.029507637023926, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5935847163200378, + "loss_total_v6": 0.7381159663200378, + "step": 1106 + }, + { + "epoch": 0.4125978382407752, + "grad_norm": 18.77130699157715, + "learning_rate": 7.053506140580322e-06, + "loss": 9.730865478515625, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4550001323223114, + "loss_total_v6": 0.5995314121246338, + "step": 1107 + }, + { + "epoch": 0.4129705553484905, + "grad_norm": 19.646278381347656, + "learning_rate": 7.0477364560112134e-06, + "loss": 9.7592191696167, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.641903817653656, + "loss_total_v6": 0.784481942653656, + "step": 1108 + }, + { + "epoch": 0.4133432724562057, + "grad_norm": 19.229656219482422, + "learning_rate": 7.041963493356352e-06, + "loss": 10.460289001464844, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5703855752944946, + "loss_total_v6": 0.7246824502944946, + "step": 1109 + }, + { + "epoch": 0.41371598956392097, + "grad_norm": 18.082536697387695, + "learning_rate": 7.036187261857289e-06, + "loss": 10.019771575927734, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5978800058364868, + "loss_total_v6": 0.7502237558364868, + "step": 1110 + }, + { + "epoch": 0.4140887066716362, + "grad_norm": 18.52757453918457, + "learning_rate": 7.030407770760812e-06, + "loss": 10.276262283325195, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.8224483132362366, + "loss_total_v6": 0.9728389382362366, + "step": 1111 + }, + { + "epoch": 0.41446142377935147, + "grad_norm": 19.045658111572266, + "learning_rate": 7.024625029318927e-06, + "loss": 10.149389266967773, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5578130483627319, + "loss_total_v6": 0.7140630483627319, + "step": 1112 + }, + { + "epoch": 0.4148341408870667, + "grad_norm": 18.139022827148438, + "learning_rate": 7.018839046788839e-06, + "loss": 10.061399459838867, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7008710503578186, + "loss_total_v6": 0.7008710503578186, + "step": 1113 + }, + { + "epoch": 0.41520685799478196, + "grad_norm": 18.287233352661133, + "learning_rate": 7.013049832432947e-06, + "loss": 9.959737777709961, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6210196018218994, + "loss_total_v6": 0.7772696018218994, + "step": 1114 + }, + { + "epoch": 0.4155795751024972, + "grad_norm": 18.530176162719727, + "learning_rate": 7.007257395518822e-06, + "loss": 10.537445068359375, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.7224568724632263, + "loss_total_v6": 0.8728474974632263, + "step": 1115 + }, + { + "epoch": 0.41595229221021246, + "grad_norm": 19.567527770996094, + "learning_rate": 7.001461745319189e-06, + "loss": 9.72557258605957, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5074242353439331, + "loss_total_v6": 0.6558617353439331, + "step": 1116 + }, + { + "epoch": 0.4163250093179277, + "grad_norm": 17.73177146911621, + "learning_rate": 6.995662891111925e-06, + "loss": 9.675371170043945, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8479734063148499, + "loss_total_v6": 1.002270221710205, + "step": 1117 + }, + { + "epoch": 0.41669772642564296, + "grad_norm": 19.043306350708008, + "learning_rate": 6.98986084218003e-06, + "loss": 10.015917778015137, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.9474261403083801, + "loss_total_v6": 1.0900042057037354, + "step": 1118 + }, + { + "epoch": 0.4170704435333582, + "grad_norm": 20.090133666992188, + "learning_rate": 6.9840556078116195e-06, + "loss": 10.372686386108398, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3092702329158783, + "loss_total_v6": 0.4538014829158783, + "step": 1119 + }, + { + "epoch": 0.41744316064107345, + "grad_norm": 18.374666213989258, + "learning_rate": 6.97824719729991e-06, + "loss": 9.816178321838379, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5808717012405396, + "loss_total_v6": 0.7312623262405396, + "step": 1120 + }, + { + "epoch": 0.41781587774878864, + "grad_norm": 18.03748321533203, + "learning_rate": 6.972435619943198e-06, + "loss": 10.091514587402344, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5197076201438904, + "loss_total_v6": 0.6740044951438904, + "step": 1121 + }, + { + "epoch": 0.4181885948565039, + "grad_norm": 18.176706314086914, + "learning_rate": 6.9666208850448595e-06, + "loss": 9.950470924377441, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.244108185172081, + "loss_total_v6": 0.3905925750732422, + "step": 1122 + }, + { + "epoch": 0.41856131196421914, + "grad_norm": 18.537565231323242, + "learning_rate": 6.960803001913315e-06, + "loss": 10.064655303955078, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4220920503139496, + "loss_total_v6": 0.578342080116272, + "step": 1123 + }, + { + "epoch": 0.4189340290719344, + "grad_norm": 16.996028900146484, + "learning_rate": 6.954981979862027e-06, + "loss": 9.72652816772461, + "loss_alignment": 0.265625, + "loss_alignment_w": 0.1328125, + "loss_sft": 0.4176430404186249, + "loss_total_v6": 0.5504555702209473, + "step": 1124 + }, + { + "epoch": 0.41930674617964964, + "grad_norm": 17.531169891357422, + "learning_rate": 6.9491578282094876e-06, + "loss": 10.317323684692383, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4806536138057709, + "loss_total_v6": 0.6329973936080933, + "step": 1125 + }, + { + "epoch": 0.4196794632873649, + "grad_norm": 18.419763565063477, + "learning_rate": 6.9433305562791935e-06, + "loss": 9.969287872314453, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5629038214683533, + "loss_total_v6": 0.7113413214683533, + "step": 1126 + }, + { + "epoch": 0.42005218039508013, + "grad_norm": 17.99957847595215, + "learning_rate": 6.937500173399639e-06, + "loss": 10.10445499420166, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5568579435348511, + "loss_total_v6": 0.7150610685348511, + "step": 1127 + }, + { + "epoch": 0.4204248975027954, + "grad_norm": 18.666015625, + "learning_rate": 6.9316666889042995e-06, + "loss": 9.860526084899902, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.33761870861053467, + "loss_total_v6": 0.48214995861053467, + "step": 1128 + }, + { + "epoch": 0.42079761461051063, + "grad_norm": 18.953128814697266, + "learning_rate": 6.925830112131613e-06, + "loss": 10.162003517150879, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3371051549911499, + "loss_total_v6": 0.4933551549911499, + "step": 1129 + }, + { + "epoch": 0.4211703317182259, + "grad_norm": 18.34357261657715, + "learning_rate": 6.9199904524249695e-06, + "loss": 9.602071762084961, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.3748781979084015, + "loss_total_v6": 0.5291750431060791, + "step": 1130 + }, + { + "epoch": 0.4215430488259411, + "grad_norm": 18.81026840209961, + "learning_rate": 6.914147719132692e-06, + "loss": 10.029041290283203, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5154183506965637, + "loss_total_v6": 0.5154183506965637, + "step": 1131 + }, + { + "epoch": 0.4219157659336564, + "grad_norm": 18.30891227722168, + "learning_rate": 6.908301921608028e-06, + "loss": 9.470868110656738, + "loss_alignment": 0.2734375, + "loss_alignment_w": 0.13671875, + "loss_sft": 0.5442559123039246, + "loss_total_v6": 0.6809746623039246, + "step": 1132 + }, + { + "epoch": 0.4222884830413716, + "grad_norm": 19.687335968017578, + "learning_rate": 6.902453069209128e-06, + "loss": 10.268665313720703, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.26494958996772766, + "loss_total_v6": 0.41534021496772766, + "step": 1133 + }, + { + "epoch": 0.42266120014908687, + "grad_norm": 19.09560775756836, + "learning_rate": 6.8966011712990275e-06, + "loss": 10.119365692138672, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.19758619368076324, + "loss_total_v6": 0.34407055377960205, + "step": 1134 + }, + { + "epoch": 0.42303391725680206, + "grad_norm": 18.399085998535156, + "learning_rate": 6.890746237245649e-06, + "loss": 9.968439102172852, + "loss_alignment": 0.359375, + "loss_alignment_w": 0.1796875, + "loss_sft": 0.7613890767097473, + "loss_total_v6": 0.9410765767097473, + "step": 1135 + }, + { + "epoch": 0.4234066343645173, + "grad_norm": 17.200349807739258, + "learning_rate": 6.884888276421766e-06, + "loss": 9.776071548461914, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4029659032821655, + "loss_total_v6": 0.5572627782821655, + "step": 1136 + }, + { + "epoch": 0.42377935147223256, + "grad_norm": 17.957408905029297, + "learning_rate": 6.879027298205001e-06, + "loss": 10.262551307678223, + "loss_alignment": 0.265625, + "loss_alignment_w": 0.1328125, + "loss_sft": 0.4341038763523102, + "loss_total_v6": 0.5669163465499878, + "step": 1137 + }, + { + "epoch": 0.4241520685799478, + "grad_norm": 18.07813835144043, + "learning_rate": 6.873163311977805e-06, + "loss": 9.787353515625, + "loss_alignment": 0.2578125, + "loss_alignment_w": 0.12890625, + "loss_sft": 0.29023227095603943, + "loss_total_v6": 0.41913852095603943, + "step": 1138 + }, + { + "epoch": 0.42452478568766305, + "grad_norm": 17.829431533813477, + "learning_rate": 6.8672963271274465e-06, + "loss": 10.329450607299805, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4556255638599396, + "loss_total_v6": 0.6099224090576172, + "step": 1139 + }, + { + "epoch": 0.4248975027953783, + "grad_norm": 17.95781707763672, + "learning_rate": 6.8614263530459925e-06, + "loss": 9.969070434570312, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.23703841865062714, + "loss_total_v6": 0.38156968355178833, + "step": 1140 + }, + { + "epoch": 0.42527021990309355, + "grad_norm": 19.522130966186523, + "learning_rate": 6.855553399130296e-06, + "loss": 10.249927520751953, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.3980022966861725, + "loss_total_v6": 0.5366741418838501, + "step": 1141 + }, + { + "epoch": 0.4256429370108088, + "grad_norm": 18.907617568969727, + "learning_rate": 6.849677474781982e-06, + "loss": 10.15264892578125, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.25277069211006165, + "loss_total_v6": 0.39534881711006165, + "step": 1142 + }, + { + "epoch": 0.42601565411852405, + "grad_norm": 19.12146759033203, + "learning_rate": 6.843798589407424e-06, + "loss": 9.799612045288086, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6021513342857361, + "loss_total_v6": 0.7642607092857361, + "step": 1143 + }, + { + "epoch": 0.4263883712262393, + "grad_norm": 18.38926887512207, + "learning_rate": 6.837916752417748e-06, + "loss": 9.834699630737305, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6589422225952148, + "loss_total_v6": 0.8073797225952148, + "step": 1144 + }, + { + "epoch": 0.42676108833395454, + "grad_norm": 19.110248565673828, + "learning_rate": 6.83203197322879e-06, + "loss": 9.994636535644531, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.7445192933082581, + "loss_total_v6": 0.8949099183082581, + "step": 1145 + }, + { + "epoch": 0.4271338054416698, + "grad_norm": 17.614980697631836, + "learning_rate": 6.826144261261108e-06, + "loss": 10.483269691467285, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.3028664290904999, + "loss_total_v6": 0.4708351790904999, + "step": 1146 + }, + { + "epoch": 0.42750652254938504, + "grad_norm": 18.3537654876709, + "learning_rate": 6.82025362593995e-06, + "loss": 10.0421142578125, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6678138375282288, + "loss_total_v6": 0.8299232125282288, + "step": 1147 + }, + { + "epoch": 0.4278792396571003, + "grad_norm": 17.317428588867188, + "learning_rate": 6.814360076695242e-06, + "loss": 10.276944160461426, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6366278529167175, + "loss_total_v6": 0.7967841029167175, + "step": 1148 + }, + { + "epoch": 0.4282519567648155, + "grad_norm": 17.87140464782715, + "learning_rate": 6.8084636229615786e-06, + "loss": 10.223657608032227, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.3112138509750366, + "loss_total_v6": 0.4518388509750366, + "step": 1149 + }, + { + "epoch": 0.4286246738725307, + "grad_norm": 17.59190559387207, + "learning_rate": 6.802564274178203e-06, + "loss": 10.168716430664062, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5791211724281311, + "loss_total_v6": 0.7334180474281311, + "step": 1150 + }, + { + "epoch": 0.428997390980246, + "grad_norm": 17.94816780090332, + "learning_rate": 6.7966620397889906e-06, + "loss": 10.141374588012695, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.299591988325119, + "loss_total_v6": 0.451935738325119, + "step": 1151 + }, + { + "epoch": 0.4293701080879612, + "grad_norm": 18.113487243652344, + "learning_rate": 6.790756929242439e-06, + "loss": 10.078097343444824, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.35561496019363403, + "loss_total_v6": 0.515771210193634, + "step": 1152 + }, + { + "epoch": 0.42974282519567647, + "grad_norm": 18.924650192260742, + "learning_rate": 6.784848951991647e-06, + "loss": 9.899173736572266, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4426640272140503, + "loss_total_v6": 0.5891484022140503, + "step": 1153 + }, + { + "epoch": 0.4301155423033917, + "grad_norm": 19.0810604095459, + "learning_rate": 6.778938117494306e-06, + "loss": 10.478363037109375, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.518978476524353, + "loss_total_v6": 0.661556601524353, + "step": 1154 + }, + { + "epoch": 0.43048825941110697, + "grad_norm": 18.075973510742188, + "learning_rate": 6.773024435212678e-06, + "loss": 9.85010051727295, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4119311571121216, + "loss_total_v6": 0.4119311571121216, + "step": 1155 + }, + { + "epoch": 0.4308609765188222, + "grad_norm": 17.98818588256836, + "learning_rate": 6.767107914613589e-06, + "loss": 9.382771492004395, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4768129289150238, + "loss_total_v6": 0.6389223337173462, + "step": 1156 + }, + { + "epoch": 0.43123369362653746, + "grad_norm": 18.47049903869629, + "learning_rate": 6.761188565168402e-06, + "loss": 10.425439834594727, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.41691893339157104, + "loss_total_v6": 0.584887683391571, + "step": 1157 + }, + { + "epoch": 0.4316064107342527, + "grad_norm": 17.7348690032959, + "learning_rate": 6.755266396353013e-06, + "loss": 9.789487838745117, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5705189108848572, + "loss_total_v6": 0.7170032858848572, + "step": 1158 + }, + { + "epoch": 0.43197912784196796, + "grad_norm": 18.32476043701172, + "learning_rate": 6.74934141764783e-06, + "loss": 9.66920280456543, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.32373756170272827, + "loss_total_v6": 0.47217506170272827, + "step": 1159 + }, + { + "epoch": 0.4323518449496832, + "grad_norm": 18.847814559936523, + "learning_rate": 6.743413638537757e-06, + "loss": 9.648874282836914, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6249885559082031, + "loss_total_v6": 0.7753791809082031, + "step": 1160 + }, + { + "epoch": 0.43272456205739845, + "grad_norm": 20.43865203857422, + "learning_rate": 6.7374830685121876e-06, + "loss": 9.874052047729492, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4711494743824005, + "loss_total_v6": 0.6234931945800781, + "step": 1161 + }, + { + "epoch": 0.4330972791651137, + "grad_norm": 18.801311492919922, + "learning_rate": 6.731549717064975e-06, + "loss": 9.900138854980469, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.43823757767677307, + "loss_total_v6": 0.5788625478744507, + "step": 1162 + }, + { + "epoch": 0.43346999627282895, + "grad_norm": 18.668214797973633, + "learning_rate": 6.72561359369443e-06, + "loss": 10.008856773376465, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5546541213989258, + "loss_total_v6": 0.6991853713989258, + "step": 1163 + }, + { + "epoch": 0.43384271338054414, + "grad_norm": 18.906862258911133, + "learning_rate": 6.7196747079033006e-06, + "loss": 9.814335823059082, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.37716150283813477, + "loss_total_v6": 0.5236458778381348, + "step": 1164 + }, + { + "epoch": 0.4342154304882594, + "grad_norm": 18.482683181762695, + "learning_rate": 6.7137330691987565e-06, + "loss": 9.63996696472168, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.49299174547195435, + "loss_total_v6": 0.6453354954719543, + "step": 1165 + }, + { + "epoch": 0.43458814759597464, + "grad_norm": 17.46723747253418, + "learning_rate": 6.707788687092371e-06, + "loss": 10.068194389343262, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5033804774284363, + "loss_total_v6": 0.6518179774284363, + "step": 1166 + }, + { + "epoch": 0.4349608647036899, + "grad_norm": 18.68836212158203, + "learning_rate": 6.701841571100114e-06, + "loss": 9.762747764587402, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5569261312484741, + "loss_total_v6": 0.7053636312484741, + "step": 1167 + }, + { + "epoch": 0.43533358181140513, + "grad_norm": 17.420459747314453, + "learning_rate": 6.6958917307423295e-06, + "loss": 10.400045394897461, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4304240942001343, + "loss_total_v6": 0.5788615942001343, + "step": 1168 + }, + { + "epoch": 0.4357062989191204, + "grad_norm": 17.893693923950195, + "learning_rate": 6.689939175543723e-06, + "loss": 9.679702758789062, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4009280204772949, + "loss_total_v6": 0.5513186454772949, + "step": 1169 + }, + { + "epoch": 0.43607901602683563, + "grad_norm": 18.555150985717773, + "learning_rate": 6.683983915033345e-06, + "loss": 9.693046569824219, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4093095660209656, + "loss_total_v6": 0.4093095660209656, + "step": 1170 + }, + { + "epoch": 0.4364517331345509, + "grad_norm": 18.125473022460938, + "learning_rate": 6.67802595874458e-06, + "loss": 9.823509216308594, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.42997226119041443, + "loss_total_v6": 0.5901285409927368, + "step": 1171 + }, + { + "epoch": 0.4368244502422661, + "grad_norm": 18.22650909423828, + "learning_rate": 6.672065316215125e-06, + "loss": 10.25756549835205, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.7820839285850525, + "loss_total_v6": 0.9363808035850525, + "step": 1172 + }, + { + "epoch": 0.4371971673499814, + "grad_norm": 18.726423263549805, + "learning_rate": 6.666101996986978e-06, + "loss": 9.841748237609863, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.49224016070365906, + "loss_total_v6": 0.49224016070365906, + "step": 1173 + }, + { + "epoch": 0.4375698844576966, + "grad_norm": 17.68535041809082, + "learning_rate": 6.660136010606421e-06, + "loss": 9.947957992553711, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.481697142124176, + "loss_total_v6": 0.637947142124176, + "step": 1174 + }, + { + "epoch": 0.43794260156541187, + "grad_norm": 19.549556732177734, + "learning_rate": 6.654167366624009e-06, + "loss": 10.090959548950195, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5165023803710938, + "loss_total_v6": 0.6688461303710938, + "step": 1175 + }, + { + "epoch": 0.4383153186731271, + "grad_norm": 17.649507522583008, + "learning_rate": 6.648196074594547e-06, + "loss": 10.196812629699707, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5624691843986511, + "loss_total_v6": 0.7226254343986511, + "step": 1176 + }, + { + "epoch": 0.43868803578084237, + "grad_norm": 18.614999771118164, + "learning_rate": 6.642222144077082e-06, + "loss": 9.971231460571289, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.920539140701294, + "loss_total_v6": 1.076789140701294, + "step": 1177 + }, + { + "epoch": 0.43906075288855756, + "grad_norm": 18.031402587890625, + "learning_rate": 6.636245584634885e-06, + "loss": 10.27604866027832, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.596264660358429, + "loss_total_v6": 0.748608410358429, + "step": 1178 + }, + { + "epoch": 0.4394334699962728, + "grad_norm": 19.234357833862305, + "learning_rate": 6.630266405835433e-06, + "loss": 10.030893325805664, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.46134108304977417, + "loss_total_v6": 0.6156379580497742, + "step": 1179 + }, + { + "epoch": 0.43980618710398806, + "grad_norm": 19.31571388244629, + "learning_rate": 6.624284617250397e-06, + "loss": 10.18440055847168, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6026726961135864, + "loss_total_v6": 0.7511101961135864, + "step": 1180 + }, + { + "epoch": 0.4401789042117033, + "grad_norm": 18.412572860717773, + "learning_rate": 6.6183002284556285e-06, + "loss": 10.37274169921875, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5657684803009033, + "loss_total_v6": 0.7083466053009033, + "step": 1181 + }, + { + "epoch": 0.44055162131941855, + "grad_norm": 18.63165283203125, + "learning_rate": 6.612313249031137e-06, + "loss": 10.059426307678223, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5370424389839172, + "loss_total_v6": 0.6971986889839172, + "step": 1182 + }, + { + "epoch": 0.4409243384271338, + "grad_norm": 18.418903350830078, + "learning_rate": 6.606323688561082e-06, + "loss": 10.410564422607422, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.7161421775817871, + "loss_total_v6": 0.8841109275817871, + "step": 1183 + }, + { + "epoch": 0.44129705553484905, + "grad_norm": 19.561063766479492, + "learning_rate": 6.600331556633759e-06, + "loss": 10.153022766113281, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6625344157218933, + "loss_total_v6": 0.8265969157218933, + "step": 1184 + }, + { + "epoch": 0.4416697726425643, + "grad_norm": 22.56521987915039, + "learning_rate": 6.594336862841569e-06, + "loss": 10.064590454101562, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.7483383417129517, + "loss_total_v6": 0.9006820917129517, + "step": 1185 + }, + { + "epoch": 0.44204248975027954, + "grad_norm": 18.02579116821289, + "learning_rate": 6.5883396167810235e-06, + "loss": 9.773775100708008, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5056686401367188, + "loss_total_v6": 0.6482467651367188, + "step": 1186 + }, + { + "epoch": 0.4424152068579948, + "grad_norm": 18.292768478393555, + "learning_rate": 6.582339828052717e-06, + "loss": 9.969819068908691, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3435250222682953, + "loss_total_v6": 0.4919625222682953, + "step": 1187 + }, + { + "epoch": 0.44278792396571004, + "grad_norm": 18.134008407592773, + "learning_rate": 6.576337506261314e-06, + "loss": 9.864309310913086, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3778553605079651, + "loss_total_v6": 0.5243397355079651, + "step": 1188 + }, + { + "epoch": 0.4431606410734253, + "grad_norm": 17.730398178100586, + "learning_rate": 6.570332661015531e-06, + "loss": 9.521085739135742, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5964417457580566, + "loss_total_v6": 0.7390198707580566, + "step": 1189 + }, + { + "epoch": 0.44353335818114054, + "grad_norm": 18.885663986206055, + "learning_rate": 6.564325301928134e-06, + "loss": 9.984046936035156, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.24665987491607666, + "loss_total_v6": 0.39900362491607666, + "step": 1190 + }, + { + "epoch": 0.4439060752888558, + "grad_norm": 18.706871032714844, + "learning_rate": 6.558315438615901e-06, + "loss": 10.000699996948242, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.5395943522453308, + "loss_total_v6": 0.7017037272453308, + "step": 1191 + }, + { + "epoch": 0.444278792396571, + "grad_norm": 21.71302604675293, + "learning_rate": 6.552303080699625e-06, + "loss": 9.92029857635498, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.20784494280815125, + "loss_total_v6": 0.35042306780815125, + "step": 1192 + }, + { + "epoch": 0.4446515095042862, + "grad_norm": 18.85983657836914, + "learning_rate": 6.546288237804093e-06, + "loss": 9.895620346069336, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6380820870399475, + "loss_total_v6": 0.8021445870399475, + "step": 1193 + }, + { + "epoch": 0.4450242266120015, + "grad_norm": 19.865407943725586, + "learning_rate": 6.540270919558069e-06, + "loss": 10.19200325012207, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.9905762672424316, + "loss_total_v6": 1.1390137672424316, + "step": 1194 + }, + { + "epoch": 0.4453969437197167, + "grad_norm": 19.3592472076416, + "learning_rate": 6.534251135594281e-06, + "loss": 9.66955280303955, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.23521068692207336, + "loss_total_v6": 0.38755443692207336, + "step": 1195 + }, + { + "epoch": 0.44576966082743197, + "grad_norm": 19.30669403076172, + "learning_rate": 6.528228895549397e-06, + "loss": 9.840475082397461, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.19741196930408478, + "loss_total_v6": 0.3614744544029236, + "step": 1196 + }, + { + "epoch": 0.4461423779351472, + "grad_norm": 17.977489471435547, + "learning_rate": 6.52220420906403e-06, + "loss": 10.2056884765625, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.41311150789260864, + "loss_total_v6": 0.5615490078926086, + "step": 1197 + }, + { + "epoch": 0.44651509504286246, + "grad_norm": 17.953561782836914, + "learning_rate": 6.516177085782697e-06, + "loss": 10.358281135559082, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3733263611793518, + "loss_total_v6": 0.5295763611793518, + "step": 1198 + }, + { + "epoch": 0.4468878121505777, + "grad_norm": 19.285139083862305, + "learning_rate": 6.5101475353538225e-06, + "loss": 9.82713508605957, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4859023988246918, + "loss_total_v6": 0.6323868036270142, + "step": 1199 + }, + { + "epoch": 0.44726052925829296, + "grad_norm": 18.90863800048828, + "learning_rate": 6.504115567429713e-06, + "loss": 10.04555606842041, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3893708884716034, + "loss_total_v6": 0.5339021682739258, + "step": 1200 + }, + { + "epoch": 0.44726052925829296, + "eval_loss": 0.60779869556427, + "eval_loss_alignment": 0.28265433026175213, + "eval_loss_alignment_w": 0.14132716513087606, + "eval_loss_sft": 0.46647154639323807, + "eval_loss_total_v6": 0.6077987112614334, + "eval_runtime": 206.6937, + "eval_samples_per_second": 9.028, + "eval_steps_per_second": 1.132, + "step": 1200 + }, + { + "epoch": 0.4476332463660082, + "grad_norm": 17.763038635253906, + "learning_rate": 6.498081191666549e-06, + "loss": 10.167903900146484, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.22415819764137268, + "loss_total_v6": 0.22415819764137268, + "step": 1201 + }, + { + "epoch": 0.44800596347372346, + "grad_norm": 18.645071029663086, + "learning_rate": 6.492044417724362e-06, + "loss": 10.216821670532227, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.24238526821136475, + "loss_total_v6": 0.38886964321136475, + "step": 1202 + }, + { + "epoch": 0.4483786805814387, + "grad_norm": 18.348529815673828, + "learning_rate": 6.4860052552670216e-06, + "loss": 9.8485107421875, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.40205201506614685, + "loss_total_v6": 0.5465832948684692, + "step": 1203 + }, + { + "epoch": 0.44875139768915395, + "grad_norm": 18.422773361206055, + "learning_rate": 6.4799637139622275e-06, + "loss": 9.34031867980957, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.5083121061325073, + "loss_total_v6": 0.6743277311325073, + "step": 1204 + }, + { + "epoch": 0.4491241147968692, + "grad_norm": 18.917552947998047, + "learning_rate": 6.473919803481482e-06, + "loss": 10.041792869567871, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3529689610004425, + "loss_total_v6": 0.4994533360004425, + "step": 1205 + }, + { + "epoch": 0.4494968319045844, + "grad_norm": 19.941362380981445, + "learning_rate": 6.46787353350008e-06, + "loss": 9.998137474060059, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5634405612945557, + "loss_total_v6": 0.7177374362945557, + "step": 1206 + }, + { + "epoch": 0.44986954901229964, + "grad_norm": 19.082229614257812, + "learning_rate": 6.461824913697096e-06, + "loss": 10.427724838256836, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.7033023834228516, + "loss_total_v6": 0.8771305084228516, + "step": 1207 + }, + { + "epoch": 0.4502422661200149, + "grad_norm": 19.743879318237305, + "learning_rate": 6.4557739537553644e-06, + "loss": 10.058670997619629, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4211481213569641, + "loss_total_v6": 0.5734918713569641, + "step": 1208 + }, + { + "epoch": 0.45061498322773014, + "grad_norm": 18.92098045349121, + "learning_rate": 6.449720663361465e-06, + "loss": 10.303316116333008, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.2145526111125946, + "loss_total_v6": 0.3629901111125946, + "step": 1209 + }, + { + "epoch": 0.4509877003354454, + "grad_norm": 18.710771560668945, + "learning_rate": 6.4436650522057135e-06, + "loss": 10.078498840332031, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2821039855480194, + "loss_total_v6": 0.4324946105480194, + "step": 1210 + }, + { + "epoch": 0.45136041744316063, + "grad_norm": 19.3259334564209, + "learning_rate": 6.4376071299821355e-06, + "loss": 9.732528686523438, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.49654728174209595, + "loss_total_v6": 0.650844156742096, + "step": 1211 + }, + { + "epoch": 0.4517331345508759, + "grad_norm": 17.737407684326172, + "learning_rate": 6.431546906388455e-06, + "loss": 10.256330490112305, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.16737690567970276, + "loss_total_v6": 0.30800190567970276, + "step": 1212 + }, + { + "epoch": 0.45210585165859113, + "grad_norm": 19.417993545532227, + "learning_rate": 6.425484391126086e-06, + "loss": 9.547640800476074, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.17265482246875763, + "loss_total_v6": 0.32890480756759644, + "step": 1213 + }, + { + "epoch": 0.4524785687663064, + "grad_norm": 18.560768127441406, + "learning_rate": 6.419419593900109e-06, + "loss": 10.082670211791992, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.16044260561466217, + "loss_total_v6": 0.308880090713501, + "step": 1214 + }, + { + "epoch": 0.4528512858740216, + "grad_norm": 17.594484329223633, + "learning_rate": 6.413352524419252e-06, + "loss": 9.745847702026367, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.2814737856388092, + "loss_total_v6": 0.4377237856388092, + "step": 1215 + }, + { + "epoch": 0.4532240029817369, + "grad_norm": 18.031911849975586, + "learning_rate": 6.407283192395888e-06, + "loss": 9.28054141998291, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.39732012152671814, + "loss_total_v6": 0.5457576513290405, + "step": 1216 + }, + { + "epoch": 0.4535967200894521, + "grad_norm": 18.943424224853516, + "learning_rate": 6.4012116075460085e-06, + "loss": 10.184242248535156, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5949234962463379, + "loss_total_v6": 0.7492203712463379, + "step": 1217 + }, + { + "epoch": 0.45396943719716737, + "grad_norm": 18.716106414794922, + "learning_rate": 6.3951377795892104e-06, + "loss": 10.048358917236328, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.32366347312927246, + "loss_total_v6": 0.32366347312927246, + "step": 1218 + }, + { + "epoch": 0.4543421543048826, + "grad_norm": 19.487546920776367, + "learning_rate": 6.389061718248686e-06, + "loss": 10.188018798828125, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.42572712898254395, + "loss_total_v6": 0.576117753982544, + "step": 1219 + }, + { + "epoch": 0.45471487141259787, + "grad_norm": 19.43324851989746, + "learning_rate": 6.382983433251196e-06, + "loss": 9.787210464477539, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.606680691242218, + "loss_total_v6": 0.753165066242218, + "step": 1220 + }, + { + "epoch": 0.45508758852031306, + "grad_norm": 18.81880760192871, + "learning_rate": 6.376902934327067e-06, + "loss": 9.988673210144043, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.47862669825553894, + "loss_total_v6": 0.6329236030578613, + "step": 1221 + }, + { + "epoch": 0.4554603056280283, + "grad_norm": 18.83611297607422, + "learning_rate": 6.370820231210168e-06, + "loss": 9.583648681640625, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5061571002006531, + "loss_total_v6": 0.6604539752006531, + "step": 1222 + }, + { + "epoch": 0.45583302273574355, + "grad_norm": 17.772872924804688, + "learning_rate": 6.364735333637892e-06, + "loss": 10.227848052978516, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.26104795932769775, + "loss_total_v6": 0.41534483432769775, + "step": 1223 + }, + { + "epoch": 0.4562057398434588, + "grad_norm": 21.2315616607666, + "learning_rate": 6.358648251351151e-06, + "loss": 9.434331893920898, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.41907820105552673, + "loss_total_v6": 0.5675157308578491, + "step": 1224 + }, + { + "epoch": 0.45657845695117405, + "grad_norm": 17.822580337524414, + "learning_rate": 6.3525589940943545e-06, + "loss": 9.830863952636719, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.684092104434967, + "loss_total_v6": 0.848154604434967, + "step": 1225 + }, + { + "epoch": 0.4569511740588893, + "grad_norm": 19.406291961669922, + "learning_rate": 6.346467571615387e-06, + "loss": 9.846757888793945, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5474458932876587, + "loss_total_v6": 0.6900240182876587, + "step": 1226 + }, + { + "epoch": 0.45732389116660455, + "grad_norm": 18.608951568603516, + "learning_rate": 6.340373993665607e-06, + "loss": 9.71076488494873, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5644844174385071, + "loss_total_v6": 0.7148750424385071, + "step": 1227 + }, + { + "epoch": 0.4576966082743198, + "grad_norm": 18.86043357849121, + "learning_rate": 6.334278269999819e-06, + "loss": 10.34832763671875, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.3842046558856964, + "loss_total_v6": 0.542407751083374, + "step": 1228 + }, + { + "epoch": 0.45806932538203504, + "grad_norm": 18.542648315429688, + "learning_rate": 6.328180410376264e-06, + "loss": 10.084797859191895, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.44945427775382996, + "loss_total_v6": 0.6017980575561523, + "step": 1229 + }, + { + "epoch": 0.4584420424897503, + "grad_norm": 18.850120544433594, + "learning_rate": 6.322080424556598e-06, + "loss": 9.85732650756836, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.38586941361427307, + "loss_total_v6": 0.5401662588119507, + "step": 1230 + }, + { + "epoch": 0.45881475959746554, + "grad_norm": 19.35776710510254, + "learning_rate": 6.315978322305892e-06, + "loss": 10.18812370300293, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.35085734724998474, + "loss_total_v6": 0.49538859724998474, + "step": 1231 + }, + { + "epoch": 0.4591874767051808, + "grad_norm": 18.40578269958496, + "learning_rate": 6.30987411339259e-06, + "loss": 9.820175170898438, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.490987092256546, + "loss_total_v6": 0.6433308124542236, + "step": 1232 + }, + { + "epoch": 0.45956019381289603, + "grad_norm": 18.226966857910156, + "learning_rate": 6.303767807588521e-06, + "loss": 9.616035461425781, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6068428754806519, + "loss_total_v6": 0.7533272504806519, + "step": 1233 + }, + { + "epoch": 0.4599329109206113, + "grad_norm": 18.416044235229492, + "learning_rate": 6.297659414668865e-06, + "loss": 10.175918579101562, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.4313523471355438, + "loss_total_v6": 0.5719773769378662, + "step": 1234 + }, + { + "epoch": 0.4603056280283265, + "grad_norm": 17.798215866088867, + "learning_rate": 6.291548944412141e-06, + "loss": 9.655624389648438, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.3351820111274719, + "loss_total_v6": 0.4758070111274719, + "step": 1235 + }, + { + "epoch": 0.4606783451360417, + "grad_norm": 17.757240295410156, + "learning_rate": 6.285436406600201e-06, + "loss": 10.011069297790527, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.48289066553115845, + "loss_total_v6": 0.6410937905311584, + "step": 1236 + }, + { + "epoch": 0.46105106224375697, + "grad_norm": 18.693326950073242, + "learning_rate": 6.2793218110181974e-06, + "loss": 10.032690048217773, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.43022969365119934, + "loss_total_v6": 0.586479663848877, + "step": 1237 + }, + { + "epoch": 0.4614237793514722, + "grad_norm": 19.321475982666016, + "learning_rate": 6.273205167454585e-06, + "loss": 9.439492225646973, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5472821593284607, + "loss_total_v6": 0.6898602843284607, + "step": 1238 + }, + { + "epoch": 0.46179649645918747, + "grad_norm": 17.728870391845703, + "learning_rate": 6.267086485701093e-06, + "loss": 9.414215087890625, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.3574922978878021, + "loss_total_v6": 0.5137423276901245, + "step": 1239 + }, + { + "epoch": 0.4621692135669027, + "grad_norm": 19.87376594543457, + "learning_rate": 6.2609657755527135e-06, + "loss": 9.736448287963867, + "loss_alignment": 0.34765625, + "loss_alignment_w": 0.173828125, + "loss_sft": 0.40101075172424316, + "loss_total_v6": 0.5748388767242432, + "step": 1240 + }, + { + "epoch": 0.46254193067461796, + "grad_norm": 18.86068344116211, + "learning_rate": 6.2548430468076835e-06, + "loss": 10.02295970916748, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.43723273277282715, + "loss_total_v6": 0.43723273277282715, + "step": 1241 + }, + { + "epoch": 0.4629146477823332, + "grad_norm": 18.89183235168457, + "learning_rate": 6.2487183092674805e-06, + "loss": 10.018863677978516, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5592917203903198, + "loss_total_v6": 0.7038229703903198, + "step": 1242 + }, + { + "epoch": 0.46328736489004846, + "grad_norm": 18.89933204650879, + "learning_rate": 6.242591572736786e-06, + "loss": 10.152568817138672, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.5019441843032837, + "loss_total_v6": 0.6425691843032837, + "step": 1243 + }, + { + "epoch": 0.4636600819977637, + "grad_norm": 18.857933044433594, + "learning_rate": 6.236462847023488e-06, + "loss": 9.850930213928223, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6339585781097412, + "loss_total_v6": 0.7823960781097412, + "step": 1244 + }, + { + "epoch": 0.46403279910547895, + "grad_norm": 18.559354782104492, + "learning_rate": 6.230332141938661e-06, + "loss": 9.723461151123047, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.2057819664478302, + "loss_total_v6": 0.3639850914478302, + "step": 1245 + }, + { + "epoch": 0.4644055162131942, + "grad_norm": 20.36945152282715, + "learning_rate": 6.224199467296542e-06, + "loss": 9.84465503692627, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.0763193890452385, + "loss_total_v6": 0.2188975214958191, + "step": 1246 + }, + { + "epoch": 0.46477823332090945, + "grad_norm": 19.88013458251953, + "learning_rate": 6.218064832914523e-06, + "loss": 9.747693061828613, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6171227693557739, + "loss_total_v6": 0.7792321443557739, + "step": 1247 + }, + { + "epoch": 0.4651509504286247, + "grad_norm": 18.35284423828125, + "learning_rate": 6.211928248613138e-06, + "loss": 10.049890518188477, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.3174038231372833, + "loss_total_v6": 0.4599819481372833, + "step": 1248 + }, + { + "epoch": 0.4655236675363399, + "grad_norm": 17.91904640197754, + "learning_rate": 6.205789724216034e-06, + "loss": 9.772644996643066, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.3807080090045929, + "loss_total_v6": 0.5467236042022705, + "step": 1249 + }, + { + "epoch": 0.46589638464405514, + "grad_norm": 17.9078369140625, + "learning_rate": 6.199649269549972e-06, + "loss": 9.897762298583984, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.836516261100769, + "loss_total_v6": 1.002531886100769, + "step": 1250 + }, + { + "epoch": 0.4662691017517704, + "grad_norm": 19.47863006591797, + "learning_rate": 6.1935068944448e-06, + "loss": 9.821249008178711, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.10630357265472412, + "loss_total_v6": 0.2508348226547241, + "step": 1251 + }, + { + "epoch": 0.46664181885948564, + "grad_norm": 17.95462417602539, + "learning_rate": 6.187362608733436e-06, + "loss": 9.794645309448242, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6352116465568542, + "loss_total_v6": 0.7895085215568542, + "step": 1252 + }, + { + "epoch": 0.4670145359672009, + "grad_norm": 18.616348266601562, + "learning_rate": 6.1812164222518626e-06, + "loss": 9.393173217773438, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.39694204926490784, + "loss_total_v6": 0.5473326444625854, + "step": 1253 + }, + { + "epoch": 0.46738725307491613, + "grad_norm": 17.791004180908203, + "learning_rate": 6.175068344839102e-06, + "loss": 9.520263671875, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.3277364671230316, + "loss_total_v6": 0.4800802171230316, + "step": 1254 + }, + { + "epoch": 0.4677599701826314, + "grad_norm": 19.754289627075195, + "learning_rate": 6.168918386337206e-06, + "loss": 9.899789810180664, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6226240992546082, + "loss_total_v6": 0.7652022242546082, + "step": 1255 + }, + { + "epoch": 0.4681326872903466, + "grad_norm": 19.624130249023438, + "learning_rate": 6.162766556591236e-06, + "loss": 9.945652961730957, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.49190109968185425, + "loss_total_v6": 0.6422917246818542, + "step": 1256 + }, + { + "epoch": 0.4685054043980619, + "grad_norm": 21.79755210876465, + "learning_rate": 6.156612865449245e-06, + "loss": 9.615288734436035, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6571848392486572, + "loss_total_v6": 0.8114817142486572, + "step": 1257 + }, + { + "epoch": 0.4688781215057771, + "grad_norm": 18.135122299194336, + "learning_rate": 6.150457322762274e-06, + "loss": 10.085906982421875, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.25536996126174927, + "loss_total_v6": 0.40185433626174927, + "step": 1258 + }, + { + "epoch": 0.46925083861349237, + "grad_norm": 20.94281768798828, + "learning_rate": 6.144299938384325e-06, + "loss": 10.133480072021484, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.17690831422805786, + "loss_total_v6": 0.31948643922805786, + "step": 1259 + }, + { + "epoch": 0.4696235557212076, + "grad_norm": 17.961605072021484, + "learning_rate": 6.138140722172344e-06, + "loss": 9.850629806518555, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.596171498298645, + "loss_total_v6": 0.750468373298645, + "step": 1260 + }, + { + "epoch": 0.46999627282892287, + "grad_norm": 18.326553344726562, + "learning_rate": 6.131979683986213e-06, + "loss": 10.015396118164062, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6073002219200134, + "loss_total_v6": 0.7655033469200134, + "step": 1261 + }, + { + "epoch": 0.4703689899366381, + "grad_norm": 18.428436279296875, + "learning_rate": 6.125816833688731e-06, + "loss": 9.94648551940918, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5793877243995667, + "loss_total_v6": 0.7395439743995667, + "step": 1262 + }, + { + "epoch": 0.47074170704435336, + "grad_norm": 17.846284866333008, + "learning_rate": 6.1196521811455975e-06, + "loss": 9.794283866882324, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.45390555262565613, + "loss_total_v6": 0.45390555262565613, + "step": 1263 + }, + { + "epoch": 0.47111442415206856, + "grad_norm": 18.580772399902344, + "learning_rate": 6.113485736225394e-06, + "loss": 9.517759323120117, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.3619576096534729, + "loss_total_v6": 0.5143013596534729, + "step": 1264 + }, + { + "epoch": 0.4714871412597838, + "grad_norm": 18.707935333251953, + "learning_rate": 6.107317508799578e-06, + "loss": 9.811895370483398, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.679096519947052, + "loss_total_v6": 0.833393394947052, + "step": 1265 + }, + { + "epoch": 0.47185985836749905, + "grad_norm": 17.8643856048584, + "learning_rate": 6.101147508742456e-06, + "loss": 10.28312873840332, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5672187209129333, + "loss_total_v6": 0.7234687209129333, + "step": 1266 + }, + { + "epoch": 0.4722325754752143, + "grad_norm": 18.46849822998047, + "learning_rate": 6.0949757459311675e-06, + "loss": 9.741453170776367, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5515904426574707, + "loss_total_v6": 0.7078404426574707, + "step": 1267 + }, + { + "epoch": 0.47260529258292955, + "grad_norm": 17.567323684692383, + "learning_rate": 6.088802230245686e-06, + "loss": 9.721205711364746, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.38164979219436646, + "loss_total_v6": 0.5300872921943665, + "step": 1268 + }, + { + "epoch": 0.4729780096906448, + "grad_norm": 18.539230346679688, + "learning_rate": 6.082626971568781e-06, + "loss": 9.834489822387695, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.71551114320755, + "loss_total_v6": 0.86004239320755, + "step": 1269 + }, + { + "epoch": 0.47335072679836004, + "grad_norm": 19.658906936645508, + "learning_rate": 6.076449979786014e-06, + "loss": 10.37962532043457, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.428671658039093, + "loss_total_v6": 0.584921658039093, + "step": 1270 + }, + { + "epoch": 0.4737234439060753, + "grad_norm": 18.113277435302734, + "learning_rate": 6.070271264785723e-06, + "loss": 10.292442321777344, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.1472204029560089, + "loss_total_v6": 0.3112829029560089, + "step": 1271 + }, + { + "epoch": 0.47409616101379054, + "grad_norm": 18.003795623779297, + "learning_rate": 6.064090836459007e-06, + "loss": 9.821722030639648, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2487718015909195, + "loss_total_v6": 0.4011155366897583, + "step": 1272 + }, + { + "epoch": 0.4744688781215058, + "grad_norm": 18.682613372802734, + "learning_rate": 6.0579087046996995e-06, + "loss": 9.419794082641602, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.40028247237205505, + "loss_total_v6": 0.5565325021743774, + "step": 1273 + }, + { + "epoch": 0.47484159522922104, + "grad_norm": 18.91328239440918, + "learning_rate": 6.051724879404371e-06, + "loss": 10.096877098083496, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.45813998579978943, + "loss_total_v6": 0.6007181406021118, + "step": 1274 + }, + { + "epoch": 0.4752143123369363, + "grad_norm": 18.73351287841797, + "learning_rate": 6.045539370472296e-06, + "loss": 10.092939376831055, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.278556227684021, + "loss_total_v6": 0.432853102684021, + "step": 1275 + }, + { + "epoch": 0.47558702944465153, + "grad_norm": 20.289371490478516, + "learning_rate": 6.039352187805445e-06, + "loss": 9.505654335021973, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4519868791103363, + "loss_total_v6": 0.6062837839126587, + "step": 1276 + }, + { + "epoch": 0.4759597465523668, + "grad_norm": 20.5711669921875, + "learning_rate": 6.0331633413084725e-06, + "loss": 9.954030990600586, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.521842360496521, + "loss_total_v6": 0.666373610496521, + "step": 1277 + }, + { + "epoch": 0.476332463660082, + "grad_norm": 18.63289451599121, + "learning_rate": 6.026972840888688e-06, + "loss": 9.796639442443848, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.1836356222629547, + "loss_total_v6": 0.3320731222629547, + "step": 1278 + }, + { + "epoch": 0.4767051807677972, + "grad_norm": 18.221498489379883, + "learning_rate": 6.020780696456059e-06, + "loss": 9.981130599975586, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.1539374440908432, + "loss_total_v6": 0.2945624589920044, + "step": 1279 + }, + { + "epoch": 0.47707789787551247, + "grad_norm": 19.187726974487305, + "learning_rate": 6.014586917923178e-06, + "loss": 9.458593368530273, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6392130851745605, + "loss_total_v6": 0.6392130851745605, + "step": 1280 + }, + { + "epoch": 0.4774506149832277, + "grad_norm": 18.87299346923828, + "learning_rate": 6.008391515205253e-06, + "loss": 9.621209144592285, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.32549065351486206, + "loss_total_v6": 0.47197502851486206, + "step": 1281 + }, + { + "epoch": 0.47782333209094296, + "grad_norm": 18.29596710205078, + "learning_rate": 6.002194498220097e-06, + "loss": 9.873558044433594, + "loss_alignment": 0.3671875, + "loss_alignment_w": 0.18359375, + "loss_sft": 0.6377631425857544, + "loss_total_v6": 0.8213568925857544, + "step": 1282 + }, + { + "epoch": 0.4781960491986582, + "grad_norm": 17.92997169494629, + "learning_rate": 5.995995876888105e-06, + "loss": 9.56619644165039, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.40601974725723267, + "loss_total_v6": 0.5544572472572327, + "step": 1283 + }, + { + "epoch": 0.47856876630637346, + "grad_norm": 18.988075256347656, + "learning_rate": 5.989795661132236e-06, + "loss": 9.468099594116211, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.26750776171684265, + "loss_total_v6": 0.41985151171684265, + "step": 1284 + }, + { + "epoch": 0.4789414834140887, + "grad_norm": 19.436044692993164, + "learning_rate": 5.983593860878006e-06, + "loss": 10.444366455078125, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6842300295829773, + "loss_total_v6": 0.8385269045829773, + "step": 1285 + }, + { + "epoch": 0.47931420052180396, + "grad_norm": 18.465078353881836, + "learning_rate": 5.977390486053469e-06, + "loss": 10.277481079101562, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.4824267029762268, + "loss_total_v6": 0.6386767029762268, + "step": 1286 + }, + { + "epoch": 0.4796869176295192, + "grad_norm": 19.574857711791992, + "learning_rate": 5.971185546589196e-06, + "loss": 9.884164810180664, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.6607560515403748, + "loss_total_v6": 0.8248185515403748, + "step": 1287 + }, + { + "epoch": 0.48005963473723445, + "grad_norm": 18.927318572998047, + "learning_rate": 5.964979052418263e-06, + "loss": 10.021333694458008, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3773937523365021, + "loss_total_v6": 0.5258312225341797, + "step": 1288 + }, + { + "epoch": 0.4804323518449497, + "grad_norm": 19.933618545532227, + "learning_rate": 5.958771013476235e-06, + "loss": 9.676248550415039, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.15895017981529236, + "loss_total_v6": 0.31715330481529236, + "step": 1289 + }, + { + "epoch": 0.48080506895266495, + "grad_norm": 18.151765823364258, + "learning_rate": 5.9525614397011535e-06, + "loss": 9.692731857299805, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.7251452803611755, + "loss_total_v6": 0.8755359053611755, + "step": 1290 + }, + { + "epoch": 0.4811777860603802, + "grad_norm": 18.08060073852539, + "learning_rate": 5.9463503410335125e-06, + "loss": 9.775557518005371, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7793158888816833, + "loss_total_v6": 0.9218940138816833, + "step": 1291 + }, + { + "epoch": 0.4815505031680954, + "grad_norm": 17.93898582458496, + "learning_rate": 5.940137727416247e-06, + "loss": 10.278434753417969, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6243524551391602, + "loss_total_v6": 0.7825555801391602, + "step": 1292 + }, + { + "epoch": 0.48192322027581064, + "grad_norm": 18.731367111206055, + "learning_rate": 5.93392360879472e-06, + "loss": 9.749397277832031, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4283905029296875, + "loss_total_v6": 0.5807342529296875, + "step": 1293 + }, + { + "epoch": 0.4822959373835259, + "grad_norm": 18.742868423461914, + "learning_rate": 5.927707995116706e-06, + "loss": 9.51498031616211, + "loss_alignment": 0.25, + "loss_alignment_w": 0.125, + "loss_sft": 0.2629067003726959, + "loss_total_v6": 0.3879067003726959, + "step": 1294 + }, + { + "epoch": 0.48266865449124113, + "grad_norm": 19.256135940551758, + "learning_rate": 5.921490896332367e-06, + "loss": 9.464387893676758, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.15458400547504425, + "loss_total_v6": 0.30302149057388306, + "step": 1295 + }, + { + "epoch": 0.4830413715989564, + "grad_norm": 20.51107406616211, + "learning_rate": 5.915272322394243e-06, + "loss": 9.713451385498047, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.46848416328430176, + "loss_total_v6": 0.6247341632843018, + "step": 1296 + }, + { + "epoch": 0.48341408870667163, + "grad_norm": 18.13973617553711, + "learning_rate": 5.909052283257239e-06, + "loss": 9.838947296142578, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.6335064172744751, + "loss_total_v6": 0.7819439172744751, + "step": 1297 + }, + { + "epoch": 0.4837868058143869, + "grad_norm": 17.704904556274414, + "learning_rate": 5.902830788878605e-06, + "loss": 9.651153564453125, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.18671058118343353, + "loss_total_v6": 0.3331949710845947, + "step": 1298 + }, + { + "epoch": 0.4841595229221021, + "grad_norm": 18.471614837646484, + "learning_rate": 5.896607849217919e-06, + "loss": 10.151886940002441, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4060315489768982, + "loss_total_v6": 0.5564221739768982, + "step": 1299 + }, + { + "epoch": 0.4845322400298174, + "grad_norm": 17.686246871948242, + "learning_rate": 5.890383474237074e-06, + "loss": 9.620084762573242, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.584147572517395, + "loss_total_v6": 0.744303822517395, + "step": 1300 + }, + { + "epoch": 0.4845322400298174, + "eval_loss": 0.6058202385902405, + "eval_loss_alignment": 0.2822265625, + "eval_loss_alignment_w": 0.14111328125, + "eval_loss_sft": 0.46470692325542623, + "eval_loss_total_v6": 0.6058202050307877, + "eval_runtime": 206.5247, + "eval_samples_per_second": 9.035, + "eval_steps_per_second": 1.133, + "step": 1300 + }, + { + "epoch": 0.4849049571375326, + "grad_norm": 18.268264770507812, + "learning_rate": 5.884157673900257e-06, + "loss": 9.779903411865234, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.1969374120235443, + "loss_total_v6": 0.3414686620235443, + "step": 1301 + }, + { + "epoch": 0.48527767424524787, + "grad_norm": 18.194530487060547, + "learning_rate": 5.877930458173942e-06, + "loss": 9.871236801147461, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.1934986263513565, + "loss_total_v6": 0.3399829864501953, + "step": 1302 + }, + { + "epoch": 0.4856503913529631, + "grad_norm": 18.09754753112793, + "learning_rate": 5.871701837026866e-06, + "loss": 10.03207015991211, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.46660277247428894, + "loss_total_v6": 0.6326184272766113, + "step": 1303 + }, + { + "epoch": 0.48602310846067837, + "grad_norm": 18.6680850982666, + "learning_rate": 5.8654718204300164e-06, + "loss": 10.052816390991211, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.736120879650116, + "loss_total_v6": 0.884558379650116, + "step": 1304 + }, + { + "epoch": 0.4863958255683936, + "grad_norm": 18.52768325805664, + "learning_rate": 5.859240418356614e-06, + "loss": 10.10396957397461, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.28914663195610046, + "loss_total_v6": 0.43758413195610046, + "step": 1305 + }, + { + "epoch": 0.4867685426761088, + "grad_norm": 19.234025955200195, + "learning_rate": 5.8530076407821006e-06, + "loss": 9.82063102722168, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.2867024540901184, + "loss_total_v6": 0.2867024540901184, + "step": 1306 + }, + { + "epoch": 0.48714125978382405, + "grad_norm": 20.696508407592773, + "learning_rate": 5.8467734976841125e-06, + "loss": 9.493282318115234, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4673148989677429, + "loss_total_v6": 0.6216117739677429, + "step": 1307 + }, + { + "epoch": 0.4875139768915393, + "grad_norm": 18.665456771850586, + "learning_rate": 5.84053799904248e-06, + "loss": 9.725763320922852, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.47758880257606506, + "loss_total_v6": 0.6182137727737427, + "step": 1308 + }, + { + "epoch": 0.48788669399925455, + "grad_norm": 19.05797004699707, + "learning_rate": 5.8343011548392e-06, + "loss": 9.555869102478027, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2186923623085022, + "loss_total_v6": 0.3710361123085022, + "step": 1309 + }, + { + "epoch": 0.4882594111069698, + "grad_norm": 19.598217010498047, + "learning_rate": 5.828062975058422e-06, + "loss": 9.845598220825195, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4396298825740814, + "loss_total_v6": 0.4396298825740814, + "step": 1310 + }, + { + "epoch": 0.48863212821468505, + "grad_norm": 18.884923934936523, + "learning_rate": 5.8218234696864366e-06, + "loss": 9.397979736328125, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.9550549983978271, + "loss_total_v6": 1.1132581233978271, + "step": 1311 + }, + { + "epoch": 0.4890048453224003, + "grad_norm": 19.66588020324707, + "learning_rate": 5.815582648711654e-06, + "loss": 10.091890335083008, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.13462422788143158, + "loss_total_v6": 0.2791554927825928, + "step": 1312 + }, + { + "epoch": 0.48937756243011554, + "grad_norm": 17.55278968811035, + "learning_rate": 5.809340522124592e-06, + "loss": 10.17869758605957, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7536638975143433, + "loss_total_v6": 0.7536638975143433, + "step": 1313 + }, + { + "epoch": 0.4897502795378308, + "grad_norm": 18.141366958618164, + "learning_rate": 5.803097099917859e-06, + "loss": 9.939016342163086, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4428631663322449, + "loss_total_v6": 0.5873944163322449, + "step": 1314 + }, + { + "epoch": 0.49012299664554604, + "grad_norm": 18.041610717773438, + "learning_rate": 5.796852392086133e-06, + "loss": 9.513535499572754, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4635378122329712, + "loss_total_v6": 0.6119753122329712, + "step": 1315 + }, + { + "epoch": 0.4904957137532613, + "grad_norm": 18.659448623657227, + "learning_rate": 5.790606408626156e-06, + "loss": 10.24341106414795, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6031561493873596, + "loss_total_v6": 0.7633123993873596, + "step": 1316 + }, + { + "epoch": 0.49086843086097653, + "grad_norm": 19.845657348632812, + "learning_rate": 5.7843591595367095e-06, + "loss": 9.852234840393066, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.36556127667427063, + "loss_total_v6": 0.5120456218719482, + "step": 1317 + }, + { + "epoch": 0.4912411479686918, + "grad_norm": 18.624900817871094, + "learning_rate": 5.778110654818602e-06, + "loss": 9.685371398925781, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5463994145393372, + "loss_total_v6": 0.6909306645393372, + "step": 1318 + }, + { + "epoch": 0.49161386507640703, + "grad_norm": 18.458187103271484, + "learning_rate": 5.771860904474646e-06, + "loss": 9.809377670288086, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.2143302857875824, + "loss_total_v6": 0.3686271607875824, + "step": 1319 + }, + { + "epoch": 0.4919865821841223, + "grad_norm": 18.198976516723633, + "learning_rate": 5.765609918509661e-06, + "loss": 9.679168701171875, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.17419922351837158, + "loss_total_v6": 0.3167773485183716, + "step": 1320 + }, + { + "epoch": 0.49235929929183747, + "grad_norm": 18.083023071289062, + "learning_rate": 5.759357706930431e-06, + "loss": 10.13176155090332, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6838996410369873, + "loss_total_v6": 0.8342902660369873, + "step": 1321 + }, + { + "epoch": 0.4927320163995527, + "grad_norm": 18.26811408996582, + "learning_rate": 5.753104279745708e-06, + "loss": 9.658205032348633, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.20219430327415466, + "loss_total_v6": 0.20219430327415466, + "step": 1322 + }, + { + "epoch": 0.49310473350726797, + "grad_norm": 17.702808380126953, + "learning_rate": 5.746849646966192e-06, + "loss": 9.98281192779541, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6181514263153076, + "loss_total_v6": 0.7763545513153076, + "step": 1323 + }, + { + "epoch": 0.4934774506149832, + "grad_norm": 17.750778198242188, + "learning_rate": 5.7405938186045104e-06, + "loss": 10.34278678894043, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.6743651628494263, + "loss_total_v6": 0.8247557878494263, + "step": 1324 + }, + { + "epoch": 0.49385016772269846, + "grad_norm": 18.893142700195312, + "learning_rate": 5.734336804675203e-06, + "loss": 9.810176849365234, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.09450999647378922, + "loss_total_v6": 0.24685373902320862, + "step": 1325 + }, + { + "epoch": 0.4942228848304137, + "grad_norm": 18.80739974975586, + "learning_rate": 5.7280786151947095e-06, + "loss": 9.793638229370117, + "loss_alignment": 0.36328125, + "loss_alignment_w": 0.181640625, + "loss_sft": 0.5628743767738342, + "loss_total_v6": 0.7445150017738342, + "step": 1326 + }, + { + "epoch": 0.49459560193812896, + "grad_norm": 18.170984268188477, + "learning_rate": 5.721819260181354e-06, + "loss": 9.847609519958496, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.44716301560401917, + "loss_total_v6": 0.5956004858016968, + "step": 1327 + }, + { + "epoch": 0.4949683190458442, + "grad_norm": 50.1312370300293, + "learning_rate": 5.715558749655319e-06, + "loss": 9.990912437438965, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.48865967988967896, + "loss_total_v6": 0.644909679889679, + "step": 1328 + }, + { + "epoch": 0.49534103615355946, + "grad_norm": 18.675588607788086, + "learning_rate": 5.709297093638646e-06, + "loss": 10.02066421508789, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.6464923620223999, + "loss_total_v6": 0.7910236120223999, + "step": 1329 + }, + { + "epoch": 0.4957137532612747, + "grad_norm": 18.296506881713867, + "learning_rate": 5.703034302155202e-06, + "loss": 9.661064147949219, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5014680027961731, + "loss_total_v6": 0.6440461277961731, + "step": 1330 + }, + { + "epoch": 0.49608647036898995, + "grad_norm": 19.31875228881836, + "learning_rate": 5.696770385230679e-06, + "loss": 9.936034202575684, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.568183183670044, + "loss_total_v6": 0.716620683670044, + "step": 1331 + }, + { + "epoch": 0.4964591874767052, + "grad_norm": 19.252296447753906, + "learning_rate": 5.690505352892563e-06, + "loss": 9.74087905883789, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5861960053443909, + "loss_total_v6": 0.7385397553443909, + "step": 1332 + }, + { + "epoch": 0.49683190458442045, + "grad_norm": 18.965957641601562, + "learning_rate": 5.684239215170133e-06, + "loss": 10.279897689819336, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6022854447364807, + "loss_total_v6": 0.7624416947364807, + "step": 1333 + }, + { + "epoch": 0.4972046216921357, + "grad_norm": 18.83904266357422, + "learning_rate": 5.677971982094431e-06, + "loss": 10.059349060058594, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.18642978370189667, + "loss_total_v6": 0.32900792360305786, + "step": 1334 + }, + { + "epoch": 0.4975773387998509, + "grad_norm": 18.217960357666016, + "learning_rate": 5.671703663698259e-06, + "loss": 9.978147506713867, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.25944456458091736, + "loss_total_v6": 0.40788206458091736, + "step": 1335 + }, + { + "epoch": 0.49795005590756614, + "grad_norm": 18.2158260345459, + "learning_rate": 5.66543427001615e-06, + "loss": 9.70966911315918, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3991073668003082, + "loss_total_v6": 0.5475448369979858, + "step": 1336 + }, + { + "epoch": 0.4983227730152814, + "grad_norm": 18.523944854736328, + "learning_rate": 5.659163811084364e-06, + "loss": 9.788084030151367, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.2223176658153534, + "loss_total_v6": 0.3648957908153534, + "step": 1337 + }, + { + "epoch": 0.49869549012299663, + "grad_norm": 20.054874420166016, + "learning_rate": 5.652892296940862e-06, + "loss": 10.428165435791016, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5991517305374146, + "loss_total_v6": 0.7436829805374146, + "step": 1338 + }, + { + "epoch": 0.4990682072307119, + "grad_norm": 18.693864822387695, + "learning_rate": 5.646619737625295e-06, + "loss": 10.266279220581055, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.4064802825450897, + "loss_total_v6": 0.5588240623474121, + "step": 1339 + }, + { + "epoch": 0.49944092433842713, + "grad_norm": 18.57000732421875, + "learning_rate": 5.640346143178989e-06, + "loss": 9.738253593444824, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.8360828161239624, + "loss_total_v6": 0.9981921911239624, + "step": 1340 + }, + { + "epoch": 0.4998136414461424, + "grad_norm": 18.15778160095215, + "learning_rate": 5.634071523644929e-06, + "loss": 10.099638938903809, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.32443785667419434, + "loss_total_v6": 0.47092223167419434, + "step": 1341 + }, + { + "epoch": 0.5001863585538576, + "grad_norm": 17.38858985900879, + "learning_rate": 5.627795889067734e-06, + "loss": 9.619349479675293, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5649614334106445, + "loss_total_v6": 0.7231645584106445, + "step": 1342 + }, + { + "epoch": 0.5005590756615729, + "grad_norm": 18.348270416259766, + "learning_rate": 5.621519249493654e-06, + "loss": 9.561555862426758, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4609149098396301, + "loss_total_v6": 0.6073992848396301, + "step": 1343 + }, + { + "epoch": 0.5009317927692881, + "grad_norm": 17.946523666381836, + "learning_rate": 5.615241614970546e-06, + "loss": 9.99552059173584, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6551087498664856, + "loss_total_v6": 0.8172181248664856, + "step": 1344 + }, + { + "epoch": 0.5013045098770034, + "grad_norm": 18.507606506347656, + "learning_rate": 5.608962995547856e-06, + "loss": 9.58548641204834, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.7164838314056396, + "loss_total_v6": 0.8590619564056396, + "step": 1345 + }, + { + "epoch": 0.5016772269847186, + "grad_norm": 18.112003326416016, + "learning_rate": 5.6026834012766155e-06, + "loss": 10.166748046875, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.46559926867485046, + "loss_total_v6": 0.6238024234771729, + "step": 1346 + }, + { + "epoch": 0.5020499440924339, + "grad_norm": 17.090167999267578, + "learning_rate": 5.59640284220941e-06, + "loss": 9.396923065185547, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.889765202999115, + "loss_total_v6": 1.0421090126037598, + "step": 1347 + }, + { + "epoch": 0.5024226612001491, + "grad_norm": 18.133535385131836, + "learning_rate": 5.590121328400367e-06, + "loss": 10.090888977050781, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6537493467330933, + "loss_total_v6": 0.8119524717330933, + "step": 1348 + }, + { + "epoch": 0.5027953783078644, + "grad_norm": 18.021230697631836, + "learning_rate": 5.583838869905151e-06, + "loss": 9.849587440490723, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6205248236656189, + "loss_total_v6": 0.7767748236656189, + "step": 1349 + }, + { + "epoch": 0.5031680954155796, + "grad_norm": 18.5119686126709, + "learning_rate": 5.577555476780933e-06, + "loss": 9.779231071472168, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47122788429260254, + "loss_total_v6": 0.6157591342926025, + "step": 1350 + }, + { + "epoch": 0.5035408125232949, + "grad_norm": 20.31577491760254, + "learning_rate": 5.5712711590863775e-06, + "loss": 9.388309478759766, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3437569737434387, + "loss_total_v6": 0.4941475987434387, + "step": 1351 + }, + { + "epoch": 0.5039135296310101, + "grad_norm": 18.829439163208008, + "learning_rate": 5.564985926881638e-06, + "loss": 10.150659561157227, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4093782305717468, + "loss_total_v6": 0.5539094805717468, + "step": 1352 + }, + { + "epoch": 0.5042862467387254, + "grad_norm": 18.2781982421875, + "learning_rate": 5.558699790228321e-06, + "loss": 9.569859504699707, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.2487325370311737, + "loss_total_v6": 0.4362325370311737, + "step": 1353 + }, + { + "epoch": 0.5046589638464406, + "grad_norm": 17.8409481048584, + "learning_rate": 5.552412759189491e-06, + "loss": 9.731775283813477, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.3897026479244232, + "loss_total_v6": 0.5381401777267456, + "step": 1354 + }, + { + "epoch": 0.5050316809541558, + "grad_norm": 20.148923873901367, + "learning_rate": 5.54612484382964e-06, + "loss": 9.649547576904297, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.4990151524543762, + "loss_total_v6": 0.6767495274543762, + "step": 1355 + }, + { + "epoch": 0.505404398061871, + "grad_norm": 18.056297302246094, + "learning_rate": 5.539836054214673e-06, + "loss": 9.429037094116211, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.3664742112159729, + "loss_total_v6": 0.5051460862159729, + "step": 1356 + }, + { + "epoch": 0.5057771151695862, + "grad_norm": 18.399253845214844, + "learning_rate": 5.533546400411899e-06, + "loss": 9.715758323669434, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5037680268287659, + "loss_total_v6": 0.6580649018287659, + "step": 1357 + }, + { + "epoch": 0.5061498322773015, + "grad_norm": 18.577072143554688, + "learning_rate": 5.527255892490007e-06, + "loss": 9.662057876586914, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.40780818462371826, + "loss_total_v6": 0.5621050596237183, + "step": 1358 + }, + { + "epoch": 0.5065225493850167, + "grad_norm": 17.581268310546875, + "learning_rate": 5.520964540519055e-06, + "loss": 9.924256324768066, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4278821349143982, + "loss_total_v6": 0.5743665099143982, + "step": 1359 + }, + { + "epoch": 0.506895266492732, + "grad_norm": 17.99384307861328, + "learning_rate": 5.5146723545704515e-06, + "loss": 9.949012756347656, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3991435170173645, + "loss_total_v6": 0.5495341420173645, + "step": 1360 + }, + { + "epoch": 0.5072679836004472, + "grad_norm": 18.743555068969727, + "learning_rate": 5.508379344716943e-06, + "loss": 9.745753288269043, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4389273524284363, + "loss_total_v6": 0.5893179774284363, + "step": 1361 + }, + { + "epoch": 0.5076407007081625, + "grad_norm": 18.59726905822754, + "learning_rate": 5.502085521032589e-06, + "loss": 9.519843101501465, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5314156413078308, + "loss_total_v6": 0.6876656413078308, + "step": 1362 + }, + { + "epoch": 0.5080134178158777, + "grad_norm": 18.167604446411133, + "learning_rate": 5.4957908935927594e-06, + "loss": 9.736618995666504, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6264124512672424, + "loss_total_v6": 0.7865687012672424, + "step": 1363 + }, + { + "epoch": 0.508386134923593, + "grad_norm": 19.041940689086914, + "learning_rate": 5.489495472474103e-06, + "loss": 9.973005294799805, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.40678587555885315, + "loss_total_v6": 0.5552233457565308, + "step": 1364 + }, + { + "epoch": 0.5087588520313082, + "grad_norm": 17.714859008789062, + "learning_rate": 5.483199267754544e-06, + "loss": 9.622838973999023, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.41814231872558594, + "loss_total_v6": 0.5587673187255859, + "step": 1365 + }, + { + "epoch": 0.5091315691390235, + "grad_norm": 18.542320251464844, + "learning_rate": 5.476902289513261e-06, + "loss": 9.513177871704102, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.24166372418403625, + "loss_total_v6": 0.40572622418403625, + "step": 1366 + }, + { + "epoch": 0.5095042862467387, + "grad_norm": 18.746360778808594, + "learning_rate": 5.470604547830667e-06, + "loss": 9.67723274230957, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.7354816794395447, + "loss_total_v6": 0.7354816794395447, + "step": 1367 + }, + { + "epoch": 0.509877003354454, + "grad_norm": 19.84618377685547, + "learning_rate": 5.4643060527884e-06, + "loss": 9.817509651184082, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.11577158421278, + "loss_total_v6": 0.2622559666633606, + "step": 1368 + }, + { + "epoch": 0.5102497204621692, + "grad_norm": 19.36513900756836, + "learning_rate": 5.4580068144693075e-06, + "loss": 9.863138198852539, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3039785325527191, + "loss_total_v6": 0.4485097825527191, + "step": 1369 + }, + { + "epoch": 0.5106224375698845, + "grad_norm": 18.237777709960938, + "learning_rate": 5.4517068429574215e-06, + "loss": 10.050317764282227, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.5205857753753662, + "loss_total_v6": 0.5205857753753662, + "step": 1370 + }, + { + "epoch": 0.5109951546775997, + "grad_norm": 19.123767852783203, + "learning_rate": 5.445406148337947e-06, + "loss": 9.454367637634277, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.6868598461151123, + "loss_total_v6": 0.8313910961151123, + "step": 1371 + }, + { + "epoch": 0.511367871785315, + "grad_norm": 18.604053497314453, + "learning_rate": 5.439104740697252e-06, + "loss": 9.461771011352539, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5505732297897339, + "loss_total_v6": 0.7029169797897339, + "step": 1372 + }, + { + "epoch": 0.5117405888930302, + "grad_norm": 18.059490203857422, + "learning_rate": 5.432802630122841e-06, + "loss": 9.717364311218262, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5398807525634766, + "loss_total_v6": 0.6863651275634766, + "step": 1373 + }, + { + "epoch": 0.5121133060007454, + "grad_norm": 18.857357025146484, + "learning_rate": 5.4264998267033445e-06, + "loss": 9.3492431640625, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4441007077693939, + "loss_total_v6": 0.5905851125717163, + "step": 1374 + }, + { + "epoch": 0.5124860231084607, + "grad_norm": 18.064544677734375, + "learning_rate": 5.420196340528507e-06, + "loss": 9.641088485717773, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4112819731235504, + "loss_total_v6": 0.559719443321228, + "step": 1375 + }, + { + "epoch": 0.512858740216176, + "grad_norm": 19.635807037353516, + "learning_rate": 5.413892181689161e-06, + "loss": 10.014988899230957, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.4350748658180237, + "loss_total_v6": 0.5893717408180237, + "step": 1376 + }, + { + "epoch": 0.5132314573238912, + "grad_norm": 18.243892669677734, + "learning_rate": 5.407587360277215e-06, + "loss": 9.691911697387695, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.32872623205184937, + "loss_total_v6": 0.47911685705184937, + "step": 1377 + }, + { + "epoch": 0.5136041744316064, + "grad_norm": 18.951953887939453, + "learning_rate": 5.401281886385643e-06, + "loss": 9.858990669250488, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.06781752407550812, + "loss_total_v6": 0.21625502407550812, + "step": 1378 + }, + { + "epoch": 0.5139768915393217, + "grad_norm": 19.75745391845703, + "learning_rate": 5.394975770108459e-06, + "loss": 10.314376831054688, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.49148914217948914, + "loss_total_v6": 0.6418797969818115, + "step": 1379 + }, + { + "epoch": 0.5143496086470369, + "grad_norm": 20.57443618774414, + "learning_rate": 5.388669021540707e-06, + "loss": 10.037381172180176, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.09196396172046661, + "loss_total_v6": 0.2443077117204666, + "step": 1380 + }, + { + "epoch": 0.5147223257547522, + "grad_norm": 18.703536987304688, + "learning_rate": 5.382361650778445e-06, + "loss": 9.648204803466797, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.44620847702026367, + "loss_total_v6": 0.6044116020202637, + "step": 1381 + }, + { + "epoch": 0.5150950428624674, + "grad_norm": 19.043251037597656, + "learning_rate": 5.376053667918726e-06, + "loss": 9.947700500488281, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.35950613021850586, + "loss_total_v6": 0.5020842552185059, + "step": 1382 + }, + { + "epoch": 0.5154677599701827, + "grad_norm": 18.872085571289062, + "learning_rate": 5.3697450830595775e-06, + "loss": 9.868326187133789, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.4845446050167084, + "loss_total_v6": 0.6447008848190308, + "step": 1383 + }, + { + "epoch": 0.5158404770778979, + "grad_norm": 18.928129196166992, + "learning_rate": 5.363435906300001e-06, + "loss": 9.514250755310059, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.4020526111125946, + "loss_total_v6": 0.4020526111125946, + "step": 1384 + }, + { + "epoch": 0.5162131941856131, + "grad_norm": 17.659212112426758, + "learning_rate": 5.3571261477399375e-06, + "loss": 9.878704071044922, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.46859312057495117, + "loss_total_v6": 0.6228899955749512, + "step": 1385 + }, + { + "epoch": 0.5165859112933283, + "grad_norm": 19.223478317260742, + "learning_rate": 5.3508158174802615e-06, + "loss": 9.979747772216797, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4005943238735199, + "loss_total_v6": 0.5587974786758423, + "step": 1386 + }, + { + "epoch": 0.5169586284010436, + "grad_norm": 19.451045989990234, + "learning_rate": 5.344504925622763e-06, + "loss": 9.658578872680664, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5086764693260193, + "loss_total_v6": 0.6649264693260193, + "step": 1387 + }, + { + "epoch": 0.5173313455087588, + "grad_norm": 19.632837295532227, + "learning_rate": 5.338193482270132e-06, + "loss": 9.690301895141602, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5875697135925293, + "loss_total_v6": 0.7301478385925293, + "step": 1388 + }, + { + "epoch": 0.5177040626164741, + "grad_norm": 19.640155792236328, + "learning_rate": 5.33188149752594e-06, + "loss": 9.680625915527344, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5457378625869751, + "loss_total_v6": 0.6902691125869751, + "step": 1389 + }, + { + "epoch": 0.5180767797241893, + "grad_norm": 16.977100372314453, + "learning_rate": 5.325568981494627e-06, + "loss": 9.445148468017578, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.22093327343463898, + "loss_total_v6": 0.3654645085334778, + "step": 1390 + }, + { + "epoch": 0.5184494968319046, + "grad_norm": 23.766630172729492, + "learning_rate": 5.319255944281479e-06, + "loss": 10.089134216308594, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.37672027945518494, + "loss_total_v6": 0.5290640592575073, + "step": 1391 + }, + { + "epoch": 0.5188222139396198, + "grad_norm": 19.092334747314453, + "learning_rate": 5.312942395992621e-06, + "loss": 9.852352142333984, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.3270097076892853, + "loss_total_v6": 0.4715409576892853, + "step": 1392 + }, + { + "epoch": 0.519194931047335, + "grad_norm": 18.86615562438965, + "learning_rate": 5.306628346734995e-06, + "loss": 9.907169342041016, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5729973912239075, + "loss_total_v6": 0.7292473912239075, + "step": 1393 + }, + { + "epoch": 0.5195676481550503, + "grad_norm": 18.140443801879883, + "learning_rate": 5.300313806616344e-06, + "loss": 10.05192756652832, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4226943552494049, + "loss_total_v6": 0.5652724504470825, + "step": 1394 + }, + { + "epoch": 0.5199403652627655, + "grad_norm": 18.34271812438965, + "learning_rate": 5.293998785745198e-06, + "loss": 10.348482131958008, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.57366544008255, + "loss_total_v6": 0.71429044008255, + "step": 1395 + }, + { + "epoch": 0.5203130823704808, + "grad_norm": 19.275196075439453, + "learning_rate": 5.287683294230855e-06, + "loss": 9.82571029663086, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.39076346158981323, + "loss_total_v6": 0.5431072115898132, + "step": 1396 + }, + { + "epoch": 0.520685799478196, + "grad_norm": 18.851591110229492, + "learning_rate": 5.281367342183368e-06, + "loss": 9.482742309570312, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.3658107817173004, + "loss_total_v6": 0.525967001914978, + "step": 1397 + }, + { + "epoch": 0.5210585165859113, + "grad_norm": 18.343225479125977, + "learning_rate": 5.2750509397135276e-06, + "loss": 9.305611610412598, + "loss_alignment": 0.33984375, + "loss_alignment_w": 0.169921875, + "loss_sft": 0.18085364997386932, + "loss_total_v6": 0.3507755398750305, + "step": 1398 + }, + { + "epoch": 0.5214312336936265, + "grad_norm": 18.162559509277344, + "learning_rate": 5.268734096932842e-06, + "loss": 10.022436141967773, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.8056532144546509, + "loss_total_v6": 0.9697157144546509, + "step": 1399 + }, + { + "epoch": 0.5218039508013418, + "grad_norm": 19.069374084472656, + "learning_rate": 5.262416823953528e-06, + "loss": 9.676130294799805, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.3755340576171875, + "loss_total_v6": 0.5395965576171875, + "step": 1400 + }, + { + "epoch": 0.5218039508013418, + "eval_loss": 0.6022669076919556, + "eval_loss_alignment": 0.2819511217948718, + "eval_loss_alignment_w": 0.1409755608974359, + "eval_loss_sft": 0.4612913518013735, + "eval_loss_total_v6": 0.6022669122769282, + "eval_runtime": 206.6951, + "eval_samples_per_second": 9.028, + "eval_steps_per_second": 1.132, + "step": 1400 + }, + { + "epoch": 0.522176667909057, + "grad_norm": 17.692646026611328, + "learning_rate": 5.256099130888489e-06, + "loss": 9.846752166748047, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5216609835624695, + "loss_total_v6": 0.6779109835624695, + "step": 1401 + }, + { + "epoch": 0.5225493850167723, + "grad_norm": 17.798234939575195, + "learning_rate": 5.249781027851303e-06, + "loss": 9.811836242675781, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.6812924742698669, + "loss_total_v6": 0.8434018492698669, + "step": 1402 + }, + { + "epoch": 0.5229221021244875, + "grad_norm": 18.0760440826416, + "learning_rate": 5.243462524956202e-06, + "loss": 9.69305419921875, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5813484787940979, + "loss_total_v6": 0.7258797287940979, + "step": 1403 + }, + { + "epoch": 0.5232948192322028, + "grad_norm": 17.945438385009766, + "learning_rate": 5.23714363231806e-06, + "loss": 9.659000396728516, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6035118103027344, + "loss_total_v6": 0.7617149353027344, + "step": 1404 + }, + { + "epoch": 0.523667536339918, + "grad_norm": 30.88479995727539, + "learning_rate": 5.230824360052375e-06, + "loss": 10.296781539916992, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.44621753692626953, + "loss_total_v6": 0.5907487869262695, + "step": 1405 + }, + { + "epoch": 0.5240402534476333, + "grad_norm": 18.84475326538086, + "learning_rate": 5.224504718275248e-06, + "loss": 9.755327224731445, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.4693285822868347, + "loss_total_v6": 0.6275317072868347, + "step": 1406 + }, + { + "epoch": 0.5244129705553485, + "grad_norm": 24.28799057006836, + "learning_rate": 5.218184717103378e-06, + "loss": 9.590720176696777, + "loss_alignment": 0.3359375, + "loss_alignment_w": 0.16796875, + "loss_sft": 0.3886721730232239, + "loss_total_v6": 0.5566409230232239, + "step": 1407 + }, + { + "epoch": 0.5247856876630638, + "grad_norm": 19.192523956298828, + "learning_rate": 5.211864366654034e-06, + "loss": 10.256681442260742, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5798010230064392, + "loss_total_v6": 0.7380041480064392, + "step": 1408 + }, + { + "epoch": 0.525158404770779, + "grad_norm": 18.52926254272461, + "learning_rate": 5.20554367704505e-06, + "loss": 10.163744926452637, + "loss_alignment": 0.35546875, + "loss_alignment_w": 0.177734375, + "loss_sft": 0.5262024402618408, + "loss_total_v6": 0.7039368152618408, + "step": 1409 + }, + { + "epoch": 0.5255311218784943, + "grad_norm": 17.8638916015625, + "learning_rate": 5.199222658394798e-06, + "loss": 9.668426513671875, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6700360774993896, + "loss_total_v6": 0.8282392024993896, + "step": 1410 + }, + { + "epoch": 0.5259038389862095, + "grad_norm": 19.335968017578125, + "learning_rate": 5.192901320822179e-06, + "loss": 9.417526245117188, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.3461480736732483, + "loss_total_v6": 0.4848199486732483, + "step": 1411 + }, + { + "epoch": 0.5262765560939248, + "grad_norm": 18.784820556640625, + "learning_rate": 5.1865796744466026e-06, + "loss": 9.356924057006836, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.4473128914833069, + "loss_total_v6": 0.5937972664833069, + "step": 1412 + }, + { + "epoch": 0.5266492732016399, + "grad_norm": 18.620882034301758, + "learning_rate": 5.180257729387976e-06, + "loss": 9.929226875305176, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.16867157816886902, + "loss_total_v6": 0.328827828168869, + "step": 1413 + }, + { + "epoch": 0.5270219903093551, + "grad_norm": 18.65167999267578, + "learning_rate": 5.173935495766681e-06, + "loss": 9.675765037536621, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.8712068796157837, + "loss_total_v6": 1.0196443796157837, + "step": 1414 + }, + { + "epoch": 0.5273947074170704, + "grad_norm": 19.0520076751709, + "learning_rate": 5.167612983703563e-06, + "loss": 10.068851470947266, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.1364814043045044, + "loss_total_v6": 0.2810126543045044, + "step": 1415 + }, + { + "epoch": 0.5277674245247856, + "grad_norm": 17.937753677368164, + "learning_rate": 5.161290203319914e-06, + "loss": 9.7723388671875, + "loss_alignment": 0.3828125, + "loss_alignment_w": 0.19140625, + "loss_sft": 0.6380907297134399, + "loss_total_v6": 0.8294969797134399, + "step": 1416 + }, + { + "epoch": 0.5281401416325009, + "grad_norm": 18.447011947631836, + "learning_rate": 5.1549671647374545e-06, + "loss": 9.916983604431152, + "loss_alignment": 0.33203125, + "loss_alignment_w": 0.166015625, + "loss_sft": 0.6006788611412048, + "loss_total_v6": 0.7666944861412048, + "step": 1417 + }, + { + "epoch": 0.5285128587402161, + "grad_norm": 18.60335922241211, + "learning_rate": 5.148643878078318e-06, + "loss": 10.001272201538086, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5490748882293701, + "loss_total_v6": 0.6994655132293701, + "step": 1418 + }, + { + "epoch": 0.5288855758479314, + "grad_norm": 18.14855194091797, + "learning_rate": 5.142320353465037e-06, + "loss": 9.239645004272461, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.22814854979515076, + "loss_total_v6": 0.38244542479515076, + "step": 1419 + }, + { + "epoch": 0.5292582929556466, + "grad_norm": 52.536163330078125, + "learning_rate": 5.13599660102052e-06, + "loss": 9.904208183288574, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.20918752253055573, + "loss_total_v6": 0.36348438262939453, + "step": 1420 + }, + { + "epoch": 0.5296310100633619, + "grad_norm": 19.289325714111328, + "learning_rate": 5.1296726308680455e-06, + "loss": 9.97152328491211, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.12949129939079285, + "loss_total_v6": 0.28378817439079285, + "step": 1421 + }, + { + "epoch": 0.5300037271710771, + "grad_norm": 18.235036849975586, + "learning_rate": 5.123348453131242e-06, + "loss": 9.43801498413086, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6117064356803894, + "loss_total_v6": 0.6117064356803894, + "step": 1422 + }, + { + "epoch": 0.5303764442787924, + "grad_norm": 18.141990661621094, + "learning_rate": 5.117024077934062e-06, + "loss": 9.62432861328125, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.5662450790405273, + "loss_total_v6": 0.7049169540405273, + "step": 1423 + }, + { + "epoch": 0.5307491613865076, + "grad_norm": 18.451980590820312, + "learning_rate": 5.110699515400782e-06, + "loss": 9.835288047790527, + "loss_alignment": 0.37109375, + "loss_alignment_w": 0.185546875, + "loss_sft": 0.741764485836029, + "loss_total_v6": 0.927311360836029, + "step": 1424 + }, + { + "epoch": 0.5311218784942229, + "grad_norm": 18.338973999023438, + "learning_rate": 5.104374775655976e-06, + "loss": 9.474331855773926, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.34322038292884827, + "loss_total_v6": 0.49556413292884827, + "step": 1425 + }, + { + "epoch": 0.5314945956019381, + "grad_norm": 18.43532943725586, + "learning_rate": 5.098049868824499e-06, + "loss": 9.404816627502441, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.6406015753746033, + "loss_total_v6": 0.7948984503746033, + "step": 1426 + }, + { + "epoch": 0.5318673127096534, + "grad_norm": 18.75092124938965, + "learning_rate": 5.091724805031476e-06, + "loss": 9.59475326538086, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6215673089027405, + "loss_total_v6": 0.7817235589027405, + "step": 1427 + }, + { + "epoch": 0.5322400298173686, + "grad_norm": 17.888225555419922, + "learning_rate": 5.085399594402285e-06, + "loss": 9.491439819335938, + "loss_alignment": 0.328125, + "loss_alignment_w": 0.1640625, + "loss_sft": 0.48206913471221924, + "loss_total_v6": 0.6461316347122192, + "step": 1428 + }, + { + "epoch": 0.5326127469250839, + "grad_norm": 18.71849250793457, + "learning_rate": 5.079074247062534e-06, + "loss": 9.720815658569336, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4863984286785126, + "loss_total_v6": 0.628976583480835, + "step": 1429 + }, + { + "epoch": 0.5329854640327991, + "grad_norm": 18.61954116821289, + "learning_rate": 5.072748773138057e-06, + "loss": 10.014299392700195, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5528221130371094, + "loss_total_v6": 0.7129783630371094, + "step": 1430 + }, + { + "epoch": 0.5333581811405144, + "grad_norm": 17.583637237548828, + "learning_rate": 5.066423182754885e-06, + "loss": 10.399723052978516, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.6938459873199463, + "loss_total_v6": 0.8540022373199463, + "step": 1431 + }, + { + "epoch": 0.5337308982482296, + "grad_norm": 18.14898109436035, + "learning_rate": 5.060097486039234e-06, + "loss": 9.812572479248047, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.43813759088516235, + "loss_total_v6": 0.5904813408851624, + "step": 1432 + }, + { + "epoch": 0.5341036153559449, + "grad_norm": 19.21245765686035, + "learning_rate": 5.053771693117497e-06, + "loss": 10.171585083007812, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.48887479305267334, + "loss_total_v6": 0.48887479305267334, + "step": 1433 + }, + { + "epoch": 0.5344763324636601, + "grad_norm": 18.267854690551758, + "learning_rate": 5.047445814116215e-06, + "loss": 9.959306716918945, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.15249599516391754, + "loss_total_v6": 0.31069910526275635, + "step": 1434 + }, + { + "epoch": 0.5348490495713754, + "grad_norm": 18.725582122802734, + "learning_rate": 5.041119859162068e-06, + "loss": 9.868698120117188, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.4979369640350342, + "loss_total_v6": 0.6366088390350342, + "step": 1435 + }, + { + "epoch": 0.5352217666790906, + "grad_norm": 17.987682342529297, + "learning_rate": 5.0347938383818615e-06, + "loss": 9.77878475189209, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.5311190485954285, + "loss_total_v6": 0.6717440485954285, + "step": 1436 + }, + { + "epoch": 0.5355944837868059, + "grad_norm": 18.187875747680664, + "learning_rate": 5.028467761902504e-06, + "loss": 9.858424186706543, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.11939665675163269, + "loss_total_v6": 0.2658810317516327, + "step": 1437 + }, + { + "epoch": 0.5359672008945211, + "grad_norm": 18.445287704467773, + "learning_rate": 5.022141639850988e-06, + "loss": 9.649052619934082, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.35401710867881775, + "loss_total_v6": 0.5044077634811401, + "step": 1438 + }, + { + "epoch": 0.5363399180022363, + "grad_norm": 19.361602783203125, + "learning_rate": 5.015815482354387e-06, + "loss": 9.801151275634766, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.6071875095367432, + "loss_total_v6": 0.7653906345367432, + "step": 1439 + }, + { + "epoch": 0.5367126351099516, + "grad_norm": 18.58696746826172, + "learning_rate": 5.009489299539827e-06, + "loss": 9.501449584960938, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.172402024269104, + "loss_total_v6": 0.324745774269104, + "step": 1440 + }, + { + "epoch": 0.5370853522176668, + "grad_norm": 18.280380249023438, + "learning_rate": 5.003163101534476e-06, + "loss": 9.74276351928711, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.25589385628700256, + "loss_total_v6": 0.41019073128700256, + "step": 1441 + }, + { + "epoch": 0.537458069325382, + "grad_norm": 17.983776092529297, + "learning_rate": 4.996836898465525e-06, + "loss": 9.488801956176758, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.8271750211715698, + "loss_total_v6": 0.9834250211715698, + "step": 1442 + }, + { + "epoch": 0.5378307864330972, + "grad_norm": 18.427778244018555, + "learning_rate": 4.9905107004601745e-06, + "loss": 9.760231971740723, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5865771174430847, + "loss_total_v6": 0.7311083674430847, + "step": 1443 + }, + { + "epoch": 0.5382035035408125, + "grad_norm": 19.148488998413086, + "learning_rate": 4.984184517645614e-06, + "loss": 10.071535110473633, + "loss_alignment": 0.375, + "loss_alignment_w": 0.1875, + "loss_sft": 0.4920845031738281, + "loss_total_v6": 0.6795845031738281, + "step": 1444 + }, + { + "epoch": 0.5385762206485277, + "grad_norm": 17.545656204223633, + "learning_rate": 4.9778583601490125e-06, + "loss": 9.508636474609375, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.547625720500946, + "loss_total_v6": 0.705828845500946, + "step": 1445 + }, + { + "epoch": 0.538948937756243, + "grad_norm": 18.146074295043945, + "learning_rate": 4.971532238097498e-06, + "loss": 9.800627708435059, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.15083858370780945, + "loss_total_v6": 0.31099483370780945, + "step": 1446 + }, + { + "epoch": 0.5393216548639582, + "grad_norm": 18.369089126586914, + "learning_rate": 4.965206161618138e-06, + "loss": 9.37707805633545, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6235036849975586, + "loss_total_v6": 0.7797536849975586, + "step": 1447 + }, + { + "epoch": 0.5396943719716735, + "grad_norm": 18.382415771484375, + "learning_rate": 4.958880140837934e-06, + "loss": 9.587251663208008, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4495640695095062, + "loss_total_v6": 0.5940953493118286, + "step": 1448 + }, + { + "epoch": 0.5400670890793887, + "grad_norm": 18.353839874267578, + "learning_rate": 4.9525541858837865e-06, + "loss": 9.432518005371094, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.2563556432723999, + "loss_total_v6": 0.4047931432723999, + "step": 1449 + }, + { + "epoch": 0.540439806187104, + "grad_norm": 18.065288543701172, + "learning_rate": 4.946228306882506e-06, + "loss": 9.623066902160645, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.43916231393814087, + "loss_total_v6": 0.5954123139381409, + "step": 1450 + }, + { + "epoch": 0.5408125232948192, + "grad_norm": 18.77948760986328, + "learning_rate": 4.9399025139607675e-06, + "loss": 9.589278221130371, + "loss_alignment": 0.31640625, + "loss_alignment_w": 0.158203125, + "loss_sft": 0.5017170906066895, + "loss_total_v6": 0.6599202156066895, + "step": 1451 + }, + { + "epoch": 0.5411852404025345, + "grad_norm": 19.019023895263672, + "learning_rate": 4.933576817245117e-06, + "loss": 9.977897644042969, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.6301026344299316, + "loss_total_v6": 0.7863526344299316, + "step": 1452 + }, + { + "epoch": 0.5415579575102497, + "grad_norm": 17.878293991088867, + "learning_rate": 4.927251226861943e-06, + "loss": 9.461090087890625, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.36824285984039307, + "loss_total_v6": 0.5127741098403931, + "step": 1453 + }, + { + "epoch": 0.541930674617965, + "grad_norm": 19.093711853027344, + "learning_rate": 4.9209257529374656e-06, + "loss": 10.055463790893555, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.5140863060951233, + "loss_total_v6": 0.6742425560951233, + "step": 1454 + }, + { + "epoch": 0.5423033917256802, + "grad_norm": 17.828903198242188, + "learning_rate": 4.914600405597717e-06, + "loss": 9.866737365722656, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.2793950140476227, + "loss_total_v6": 0.4317387640476227, + "step": 1455 + }, + { + "epoch": 0.5426761088333955, + "grad_norm": 18.61227798461914, + "learning_rate": 4.908275194968525e-06, + "loss": 9.607654571533203, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.25558921694755554, + "loss_total_v6": 0.40597984194755554, + "step": 1456 + }, + { + "epoch": 0.5430488259411107, + "grad_norm": 18.81659507751465, + "learning_rate": 4.901950131175504e-06, + "loss": 9.616889953613281, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.42006421089172363, + "loss_total_v6": 0.5685017108917236, + "step": 1457 + }, + { + "epoch": 0.543421543048826, + "grad_norm": 18.762168884277344, + "learning_rate": 4.895625224344026e-06, + "loss": 9.664912223815918, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.45577719807624817, + "loss_total_v6": 0.6003084182739258, + "step": 1458 + }, + { + "epoch": 0.5437942601565412, + "grad_norm": 18.453201293945312, + "learning_rate": 4.889300484599218e-06, + "loss": 9.70350456237793, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.3231510519981384, + "loss_total_v6": 0.4657291769981384, + "step": 1459 + }, + { + "epoch": 0.5441669772642564, + "grad_norm": 18.23265266418457, + "learning_rate": 4.88297592206594e-06, + "loss": 9.725738525390625, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.13861538469791412, + "loss_total_v6": 0.29095911979675293, + "step": 1460 + }, + { + "epoch": 0.5445396943719717, + "grad_norm": 18.0729923248291, + "learning_rate": 4.876651546868759e-06, + "loss": 9.57188606262207, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.6616588234901428, + "loss_total_v6": 0.8042369484901428, + "step": 1461 + }, + { + "epoch": 0.5449124114796869, + "grad_norm": 17.549062728881836, + "learning_rate": 4.870327369131956e-06, + "loss": 9.793890953063965, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.09245739132165909, + "loss_total_v6": 0.2448011338710785, + "step": 1462 + }, + { + "epoch": 0.5452851285874022, + "grad_norm": 19.829814910888672, + "learning_rate": 4.864003398979482e-06, + "loss": 9.177041053771973, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.38085561990737915, + "loss_total_v6": 0.5214806199073792, + "step": 1463 + }, + { + "epoch": 0.5456578456951174, + "grad_norm": 18.246110916137695, + "learning_rate": 4.857679646534966e-06, + "loss": 9.434690475463867, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.7253232598304749, + "loss_total_v6": 0.8659482598304749, + "step": 1464 + }, + { + "epoch": 0.5460305628028327, + "grad_norm": 19.242216110229492, + "learning_rate": 4.851356121921683e-06, + "loss": 10.375299453735352, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.47785335779190063, + "loss_total_v6": 0.6341033577919006, + "step": 1465 + }, + { + "epoch": 0.5464032799105479, + "grad_norm": 18.84519386291504, + "learning_rate": 4.845032835262546e-06, + "loss": 9.796209335327148, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4983062446117401, + "loss_total_v6": 0.6486968994140625, + "step": 1466 + }, + { + "epoch": 0.5467759970182632, + "grad_norm": 18.697664260864258, + "learning_rate": 4.838709796680087e-06, + "loss": 9.611549377441406, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.30791884660720825, + "loss_total_v6": 0.45830947160720825, + "step": 1467 + }, + { + "epoch": 0.5471487141259784, + "grad_norm": 17.5909423828125, + "learning_rate": 4.832387016296438e-06, + "loss": 9.303560256958008, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.3361201882362366, + "loss_total_v6": 0.4865108132362366, + "step": 1468 + }, + { + "epoch": 0.5475214312336937, + "grad_norm": 19.036457061767578, + "learning_rate": 4.8260645042333206e-06, + "loss": 10.224522590637207, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4275863468647003, + "loss_total_v6": 0.5721175670623779, + "step": 1469 + }, + { + "epoch": 0.5478941483414088, + "grad_norm": 17.774490356445312, + "learning_rate": 4.819742270612025e-06, + "loss": 9.818746566772461, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.7008171081542969, + "loss_total_v6": 0.8414421081542969, + "step": 1470 + }, + { + "epoch": 0.5482668654491241, + "grad_norm": 18.134517669677734, + "learning_rate": 4.8134203255534e-06, + "loss": 9.880034446716309, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5277352333068848, + "loss_total_v6": 0.6800789833068848, + "step": 1471 + }, + { + "epoch": 0.5486395825568393, + "grad_norm": 18.985729217529297, + "learning_rate": 4.807098679177822e-06, + "loss": 9.806861877441406, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4487643837928772, + "loss_total_v6": 0.5991550087928772, + "step": 1472 + }, + { + "epoch": 0.5490122996645546, + "grad_norm": 18.674667358398438, + "learning_rate": 4.800777341605203e-06, + "loss": 9.434414863586426, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.19290728867053986, + "loss_total_v6": 0.33743852376937866, + "step": 1473 + }, + { + "epoch": 0.5493850167722698, + "grad_norm": 18.55124855041504, + "learning_rate": 4.794456322954953e-06, + "loss": 9.50826644897461, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.40765824913978577, + "loss_total_v6": 0.5580488443374634, + "step": 1474 + }, + { + "epoch": 0.549757733879985, + "grad_norm": 17.99994659423828, + "learning_rate": 4.788135633345968e-06, + "loss": 9.891565322875977, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.3060050904750824, + "loss_total_v6": 0.4524894654750824, + "step": 1475 + }, + { + "epoch": 0.5501304509877003, + "grad_norm": 18.42738151550293, + "learning_rate": 4.781815282896626e-06, + "loss": 9.386445045471191, + "loss_alignment": 0.32421875, + "loss_alignment_w": 0.162109375, + "loss_sft": 0.4528059661388397, + "loss_total_v6": 0.6149153709411621, + "step": 1476 + }, + { + "epoch": 0.5505031680954156, + "grad_norm": 18.60844612121582, + "learning_rate": 4.775495281724755e-06, + "loss": 9.593276977539062, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.40575098991394043, + "loss_total_v6": 0.5600478649139404, + "step": 1477 + }, + { + "epoch": 0.5508758852031308, + "grad_norm": 19.21306610107422, + "learning_rate": 4.7691756399476285e-06, + "loss": 9.77379035949707, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.2305261343717575, + "loss_total_v6": 0.3809167742729187, + "step": 1478 + }, + { + "epoch": 0.551248602310846, + "grad_norm": 18.240205764770508, + "learning_rate": 4.7628563676819415e-06, + "loss": 9.73385238647461, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5216213464736938, + "loss_total_v6": 0.6700588464736938, + "step": 1479 + }, + { + "epoch": 0.5516213194185613, + "grad_norm": 17.5374698638916, + "learning_rate": 4.756537475043798e-06, + "loss": 9.502796173095703, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4747956693172455, + "loss_total_v6": 0.6173738241195679, + "step": 1480 + }, + { + "epoch": 0.5519940365262765, + "grad_norm": 18.795236587524414, + "learning_rate": 4.750218972148698e-06, + "loss": 9.535545349121094, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.18527477979660034, + "loss_total_v6": 0.33761852979660034, + "step": 1481 + }, + { + "epoch": 0.5523667536339918, + "grad_norm": 18.845863342285156, + "learning_rate": 4.743900869111512e-06, + "loss": 9.632287979125977, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.4150211215019226, + "loss_total_v6": 0.5654117465019226, + "step": 1482 + }, + { + "epoch": 0.552739470741707, + "grad_norm": 19.47191047668457, + "learning_rate": 4.737583176046475e-06, + "loss": 9.922479629516602, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5706976652145386, + "loss_total_v6": 0.7269476652145386, + "step": 1483 + }, + { + "epoch": 0.5531121878494223, + "grad_norm": 18.399742126464844, + "learning_rate": 4.731265903067159e-06, + "loss": 10.181595802307129, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.6229554414749146, + "loss_total_v6": 0.7694398164749146, + "step": 1484 + }, + { + "epoch": 0.5534849049571375, + "grad_norm": 18.05133819580078, + "learning_rate": 4.724949060286476e-06, + "loss": 9.873199462890625, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.2121797502040863, + "loss_total_v6": 0.3664766252040863, + "step": 1485 + }, + { + "epoch": 0.5538576220648528, + "grad_norm": 18.71822738647461, + "learning_rate": 4.718632657816634e-06, + "loss": 9.691189765930176, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.1591513305902481, + "loss_total_v6": 0.3036825656890869, + "step": 1486 + }, + { + "epoch": 0.554230339172568, + "grad_norm": 17.48577880859375, + "learning_rate": 4.7123167057691446e-06, + "loss": 9.81814956665039, + "loss_alignment": 0.0, + "loss_alignment_w": 0.0, + "loss_sft": 0.6679906845092773, + "loss_total_v6": 0.6679906845092773, + "step": 1487 + }, + { + "epoch": 0.5546030562802833, + "grad_norm": 17.79203987121582, + "learning_rate": 4.706001214254804e-06, + "loss": 9.916873931884766, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.3386543393135071, + "loss_total_v6": 0.4792793393135071, + "step": 1488 + }, + { + "epoch": 0.5549757733879985, + "grad_norm": 19.44068145751953, + "learning_rate": 4.699686193383657e-06, + "loss": 9.73422622680664, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5777281522750854, + "loss_total_v6": 0.7281187772750854, + "step": 1489 + }, + { + "epoch": 0.5553484904957138, + "grad_norm": 18.061351776123047, + "learning_rate": 4.693371653265007e-06, + "loss": 9.940500259399414, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.37903276085853577, + "loss_total_v6": 0.5352827310562134, + "step": 1490 + }, + { + "epoch": 0.555721207603429, + "grad_norm": 18.437379837036133, + "learning_rate": 4.6870576040073805e-06, + "loss": 9.82089614868164, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.8578596711158752, + "loss_total_v6": 1.0121564865112305, + "step": 1491 + }, + { + "epoch": 0.5560939247111443, + "grad_norm": 18.298946380615234, + "learning_rate": 4.680744055718524e-06, + "loss": 9.911579132080078, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.44848912954330444, + "loss_total_v6": 0.5949735045433044, + "step": 1492 + }, + { + "epoch": 0.5564666418188595, + "grad_norm": 18.542369842529297, + "learning_rate": 4.674431018505376e-06, + "loss": 9.672189712524414, + "loss_alignment": 0.30859375, + "loss_alignment_w": 0.154296875, + "loss_sft": 0.5402552485466003, + "loss_total_v6": 0.6945521235466003, + "step": 1493 + }, + { + "epoch": 0.5568393589265748, + "grad_norm": 17.98310661315918, + "learning_rate": 4.6681185024740604e-06, + "loss": 9.196855545043945, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.2155986726284027, + "loss_total_v6": 0.3718486726284027, + "step": 1494 + }, + { + "epoch": 0.55721207603429, + "grad_norm": 18.062358856201172, + "learning_rate": 4.661806517729869e-06, + "loss": 9.338319778442383, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5239943861961365, + "loss_total_v6": 0.6724318861961365, + "step": 1495 + }, + { + "epoch": 0.5575847931420053, + "grad_norm": 19.528976440429688, + "learning_rate": 4.655495074377237e-06, + "loss": 9.893160820007324, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5456331372261047, + "loss_total_v6": 0.6960237622261047, + "step": 1496 + }, + { + "epoch": 0.5579575102497205, + "grad_norm": 19.597930908203125, + "learning_rate": 4.649184182519742e-06, + "loss": 9.67820930480957, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5274760127067566, + "loss_total_v6": 0.6720072627067566, + "step": 1497 + }, + { + "epoch": 0.5583302273574358, + "grad_norm": 18.255430221557617, + "learning_rate": 4.642873852260065e-06, + "loss": 9.821673393249512, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.26673972606658936, + "loss_total_v6": 0.41908347606658936, + "step": 1498 + }, + { + "epoch": 0.5587029444651509, + "grad_norm": 19.199234008789062, + "learning_rate": 4.6365640937000015e-06, + "loss": 9.711814880371094, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5661517977714539, + "loss_total_v6": 0.7126361727714539, + "step": 1499 + }, + { + "epoch": 0.5590756615728661, + "grad_norm": 18.124595642089844, + "learning_rate": 4.630254916940424e-06, + "loss": 9.798891067504883, + "loss_alignment": 0.3203125, + "loss_alignment_w": 0.16015625, + "loss_sft": 0.4901702105998993, + "loss_total_v6": 0.6503264904022217, + "step": 1500 + }, + { + "epoch": 0.5590756615728661, + "eval_loss": 0.5988869071006775, + "eval_loss_alignment": 0.28185722155448717, + "eval_loss_alignment_w": 0.14092861077724358, + "eval_loss_sft": 0.4579583600513701, + "eval_loss_total_v6": 0.5988869710514942, + "eval_runtime": 207.5928, + "eval_samples_per_second": 8.989, + "eval_steps_per_second": 1.127, + "step": 1500 + } + ], + "logging_steps": 1, + "max_steps": 2683, + "num_input_tokens_seen": 0, + "num_train_epochs": 1, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 1.563320238856156e+19, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/training_args.bin b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..db8b9f54e40f57229271a5e51fd351ecf66469f0 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0ac8e4947fe1d684c1b303d2e1494f594e507cde3997b206443f89b40b9dd51 +size 8977 diff --git a/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/zero_to_fp32.py b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..5995d6e6f04e43b989587aa9022a3aef0c66d694 --- /dev/null +++ b/qwen-cua__runs__v6-recurtorch-img3-hw2-k16-12k-joint-trajonly__v0-20260605-205532__checkpoint-1500/zero_to_fp32.py @@ -0,0 +1,760 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: +# python zero_to_fp32.py . output_dir/ +# or +# python zero_to_fp32.py . output_dir/ --safe_serialization + +import argparse +import torch +import glob +import math +import os +import re +import gc +import json +import numpy as np +from tqdm import tqdm +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device, weights_only=False) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + total_files = len(files) + state_dicts = [] + for f in tqdm(files, desc='Loading checkpoint shards'): + state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if ZERO_STAGE not in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +class GatheredTensor: + """ + A pseudo tensor that collects partitioned weights. + It is more memory efficient when there are multiple groups. + """ + + def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape): + self.flat_groups = flat_groups + self.flat_groups_offset = flat_groups_offset + self.offset = offset + self.partitioned_numel = partitioned_numel + self.shape = shape + self.dtype = self.flat_groups[0][0].dtype + + def contiguous(self): + """ + Merge partitioned weights from flat_groups into a single tensor. + """ + end_idx = self.offset + self.partitioned_numel + world_size = len(self.flat_groups) + pad_flat_param_chunks = [] + + for rank_i in range(world_size): + # for each rank, we need to collect weights from related group/groups + flat_groups_at_rank_i = self.flat_groups[rank_i] + start_group_id = None + end_group_id = None + for group_id in range(len(self.flat_groups_offset)): + if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]: + start_group_id = group_id + if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]: + end_group_id = group_id + break + # collect weights from related group/groups + for group_id in range(start_group_id, end_group_id + 1): + flat_tensor = flat_groups_at_rank_i[group_id] + start_offset = self.offset - self.flat_groups_offset[group_id] + end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id] + pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset]) + + # collect weights from all ranks + pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0) + param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous() + return param + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size + + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]])) + for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'): + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # memory efficient tensor + tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape) + state_dict[name] = tensor + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def to_torch_tensor(state_dict, return_empty_tensor=False): + """ + Convert state_dict of GatheredTensor to torch tensor + """ + torch_state_dict = {} + converted_tensors = {} + for name, tensor in state_dict.items(): + tensor_id = id(tensor) + if tensor_id in converted_tensors: # shared tensors + shared_tensor = torch_state_dict[converted_tensors[tensor_id]] + torch_state_dict[name] = shared_tensor + else: + converted_tensors[tensor_id] = name + if return_empty_tensor: + torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype) + else: + torch_state_dict[name] = tensor.contiguous() + return torch_state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, + tag=None, + exclude_frozen_parameters=False, + lazy_mode=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient. + Convert the pesduo tensor to torch tensor by ``.contiguous()`` + + Returns: + - pytorch ``state_dict`` + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + Note: the above usage may not work if your application doesn't have sufficient free CPU memory. + You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. Or you can load state_dict in lazy mode :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu + for name, lazy_tensor in state_dict.item(): + tensor = lazy_tensor.contiguous() # to cpu + print(name, tensor) + # del tensor to release memory if it no longer in use + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + if lazy_mode: + return state_dict + else: + return to_torch_tensor(state_dict) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, + output_dir, + max_shard_size="5GB", + safe_serialization=False, + tag=None, + exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_dir``: directory to the pytorch fp32 state_dict output files + - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB + - ``safe_serialization``: whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`). + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + # Dependency pre-check + if safe_serialization: + try: + from safetensors.torch import save_file + except ImportError: + print('If you want to use `safe_serialization`, please `pip install safetensors`') + raise + if max_shard_size is not None: + try: + from huggingface_hub import split_torch_state_dict_into_shards + except ImportError: + print('If you want to use `max_shard_size`, please `pip install huggingface_hub`') + raise + + # Convert zero checkpoint to state_dict + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, + tag, + exclude_frozen_parameters, + lazy_mode=True) + + # Shard the model if it is too big. + weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin" + if max_shard_size is not None: + filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors") + # an memory-efficient approach for sharding + empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True) + state_dict_split = split_torch_state_dict_into_shards(empty_state_dict, + filename_pattern=filename_pattern, + max_shard_size=max_shard_size) + else: + from collections import namedtuple + StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"]) + state_dict_split = StateDictSplit(is_sharded=False, + filename_to_tensors={weights_name: list(state_dict.keys())}) + + # Save the model by shard + os.makedirs(output_dir, exist_ok=True) + filename_to_tensors = state_dict_split.filename_to_tensors.items() + for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"): + shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors} + shard_state_dict = to_torch_tensor(shard_state_dict) + output_path = os.path.join(output_dir, shard_file) + if safe_serialization: + save_file(shard_state_dict, output_path, metadata={"format": "pt"}) + else: + torch.save(shard_state_dict, output_path) + # release the memory of current shard + for tensor_name in list(shard_state_dict.keys()): + del state_dict[tensor_name] + del shard_state_dict[tensor_name] + del shard_state_dict + gc.collect() + + # Save index if sharded + if state_dict_split.is_sharded: + index = { + "metadata": state_dict_split.metadata, + "weight_map": state_dict_split.tensor_to_filename, + } + save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json" + save_index_file = os.path.join(output_dir, save_index_file) + with open(save_index_file, "w", encoding="utf-8") as f: + content = json.dumps(index, indent=2, sort_keys=True) + "\n" + f.write(content) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info("Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info("Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument("output_dir", + type=str, + help="directory to the pytorch fp32 state_dict output files" + "(e.g. path/checkpoint-12-output/)") + parser.add_argument( + "--max_shard_size", + type=str, + default="5GB", + help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size" + "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`" + "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances" + "without CPU OOM issues.") + parser.add_argument( + "--safe_serialization", + default=False, + action='store_true', + help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_dir, + max_shard_size=args.max_shard_size, + safe_serialization=args.safe_serialization, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters)