diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/args.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/args.json new file mode 100644 index 0000000000000000000000000000000000000000..ae770be8587ed458df62f9adb5eb838f7060e6a0 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/args.json @@ -0,0 +1,377 @@ +{ + "output_dir": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825/v0-20260606-155855", + "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": false, + "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": "qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825", + "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/qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825/v0-20260606-155855/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": true, + "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": true, + "packing_length": 12288, + "packing_num_proc": 1, + "lazy_tokenize": false, + "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/qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825/v0-20260606-155855', 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=False, 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='qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825', 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/qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825/v0-20260606-155855/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/chat_template.jinja b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/config.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/config.json new file mode 100644 index 0000000000000000000000000000000000000000..314297e9ae3a6f7ca90b01d29ee0a46cfbf53c2a --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/generation_config.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1f1ab970fed81cb788c2945bed33833d247b9091 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..73e458e7dad3f14604cb282b997bca8e75b52af2 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86a48b0ff60bc5330251af690ce82a8f59b2a78b1ee1c5837eb2a499b1fa068f +size 6337637425 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..2d9d4b0d886bd120c11a699a8809afc60addcbfd --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0d7df1bb3b670620c959742624e25fa894465282b010165472676d4020b59435 +size 6337640689 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..2f1b87126f9dddb8bf5cd71545816fb0e9a3677e --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6372d468602853156250bf28794524341e0a94917f747f79752a58a633033085 +size 6337641521 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..ddbc8767343fa888a2e0b0ca6dc413e4a616def3 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/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:c2a76a81cbb4df77edf6f10f581a2ce7938f21db372bd7e97eaf5ec406434e5d +size 6337641521 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..3bd2986fc27913b81230e378f43ec96233ebf71d --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9cbd43342a442c313024f9627179b88b74f5a99d412dc69436cd20b5d20077de +size 6337641457 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..e9a168115e5b799c606536249c1304b527af4ff4 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4652a317772a79ed9f108b8999036c4875339a4556860128c7e9d54c6fa35d82 +size 6337641521 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..b7182ac2f1bc364b35fb665b797efeddb1db1be9 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/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:cba4712abaeeabd0b5db9731f699bc2900cc5e3b50f7a93e8c7d864b57854c7b +size 6337641201 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..0bedb2b1bd580bf5d4ea41fa6c3b579ec8019c0a --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:17c4bd3a96918a6369f0c262f498e22b65e8a976c4e614221f25821ca93a1896 +size 6337640241 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/mp_rank_00_model_states.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/mp_rank_00_model_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..cd6f1d9d73d3b16346251a5049e95d771d25d35a --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/global_step833/mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf1729ea04f77bda62098416e88799fa485d0dfee4597cbb68cde735890c8dc7 +size 9784608719 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/latest b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/latest new file mode 100644 index 0000000000000000000000000000000000000000..131c2cb4974a72e6ba851a7046374db119a24431 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/latest @@ -0,0 +1 @@ +global_step833 \ No newline at end of file diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00001-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..860825b410a5af914eabd98a76af66363fed367e --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ef6599251213a0e5f3d4018a05cb5a28aa71f24c0a7a690e1097a3e83acce3bd +size 4995635952 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00002-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b64909bae41fb2b3288d6d03fdfc0b6c315b08f0 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d3c0fd248187ec0314209b58a46df884c0d8e9b79424a6c4addd7380d50e7a1 +size 4993498424 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00003-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4ab8023a740649342f1bc7e224320898de73054c --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model.safetensors.index.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..0bdeac0c13fb0bc1bd472391718513cd39619d7a --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/preprocessor_config.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..2ea84a437d448ff71b08df68fdd949d5cc4ebb64 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/processor_config.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_0.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..86eaf5e40dd94669ca76b101cf9f8739e75c9294 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_1.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..38dbff679bfbf2f3aade238dfdfc43c385f23334 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_2.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..31978c299901c622b39507e1513d26f45e6e4a79 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_3.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_3.pth new file mode 100644 index 0000000000000000000000000000000000000000..55b8da5b9e39cd017c04f97abe46db037ebf16f8 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_4.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_4.pth new file mode 100644 index 0000000000000000000000000000000000000000..0022f2989901cd3a52f9032b32a09841521ef45e --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_5.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_5.pth new file mode 100644 index 0000000000000000000000000000000000000000..4613c4f8de3fef534415098e475909b8f7ab97c2 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_6.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_6.pth new file mode 100644 index 0000000000000000000000000000000000000000..5e799753068722228ae33a268259fe20661b5d2d --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_7.pth b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/rng_state_7.pth new file mode 100644 index 0000000000000000000000000000000000000000..2239a6b21428f351d266a85ab12af52e47a6e3a7 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/scheduler.pt b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..f6aab70c2ab697bbfe8835b14ec938e940d25453 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:975e960570c78a9936d4a0ec4e2fe738533f4c8e03b49e049d0e48ebf67f311b +size 1465 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/tokenizer.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..9e8ac8634b7c90a05a65182d43516ddf7371c521 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/tokenizer_config.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/trainer_state.json b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3dd3c5802868ce21e4a4f143c5ebb84a37c0f112 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/trainer_state.json @@ -0,0 +1,9323 @@ +{ + "best_global_step": 833, + "best_metric": 0.58488131, + "best_model_checkpoint": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825/v0-20260606-155855/checkpoint-833", + "epoch": 1.0, + "eval_steps": 100, + "global_step": 833, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012008405884118883, + "grad_norm": 273.5415954589844, + "learning_rate": 5.0000000000000004e-08, + "loss": 23.425777435302734, + "loss_alignment": 0.982421875, + "loss_alignment_w": 0.4912109375, + "loss_sft": 0.9714352041482925, + "loss_total_v6": 1.464111015200615, + "step": 1 + }, + { + "epoch": 0.0024016811768237767, + "grad_norm": 261.60888671875, + "learning_rate": 1.0000000000000001e-07, + "loss": 23.169954299926758, + "loss_alignment": 0.9580078125, + "loss_alignment_w": 0.47900390625, + "loss_sft": 0.9675770774483681, + "loss_total_v6": 1.4481221139431, + "step": 2 + }, + { + "epoch": 0.003602521765235665, + "grad_norm": 261.2533874511719, + "learning_rate": 1.5000000000000002e-07, + "loss": 23.213726043701172, + "loss_alignment": 0.94287109375, + "loss_alignment_w": 0.471435546875, + "loss_sft": 0.9778813198208809, + "loss_total_v6": 1.4508579522371292, + "step": 3 + }, + { + "epoch": 0.004803362353647553, + "grad_norm": 261.6865234375, + "learning_rate": 2.0000000000000002e-07, + "loss": 23.325180053710938, + "loss_alignment": 0.94873046875, + "loss_alignment_w": 0.474365234375, + "loss_sft": 0.9812002778053284, + "loss_total_v6": 1.457823783159256, + "step": 4 + }, + { + "epoch": 0.006004202942059442, + "grad_norm": 265.5802917480469, + "learning_rate": 2.5000000000000004e-07, + "loss": 23.26262664794922, + "loss_alignment": 0.92626953125, + "loss_alignment_w": 0.463134765625, + "loss_sft": 0.9894671514630318, + "loss_total_v6": 1.4539141952991486, + "step": 5 + }, + { + "epoch": 0.00720504353047133, + "grad_norm": 250.35877990722656, + "learning_rate": 3.0000000000000004e-07, + "loss": 23.254159927368164, + "loss_alignment": 0.95703125, + "loss_alignment_w": 0.478515625, + "loss_sft": 0.9727179706096649, + "loss_total_v6": 1.4533850699663162, + "step": 6 + }, + { + "epoch": 0.008405884118883218, + "grad_norm": 251.25498962402344, + "learning_rate": 3.5000000000000004e-07, + "loss": 23.79116439819336, + "loss_alignment": 0.95703125, + "loss_alignment_w": 0.478515625, + "loss_sft": 1.0061738938093185, + "loss_total_v6": 1.4869478195905685, + "step": 7 + }, + { + "epoch": 0.009606724707295107, + "grad_norm": 221.600830078125, + "learning_rate": 4.0000000000000003e-07, + "loss": 23.29102325439453, + "loss_alignment": 0.94091796875, + "loss_alignment_w": 0.470458984375, + "loss_sft": 0.9837040901184082, + "loss_total_v6": 1.4556889683008194, + "step": 8 + }, + { + "epoch": 0.010807565295706995, + "grad_norm": 220.65504455566406, + "learning_rate": 4.5000000000000003e-07, + "loss": 23.02918243408203, + "loss_alignment": 0.94921875, + "loss_alignment_w": 0.474609375, + "loss_sft": 0.9629140719771385, + "loss_total_v6": 1.4393239319324493, + "step": 9 + }, + { + "epoch": 0.012008405884118884, + "grad_norm": 216.0348663330078, + "learning_rate": 5.000000000000001e-07, + "loss": 22.897785186767578, + "loss_alignment": 0.9248046875, + "loss_alignment_w": 0.46240234375, + "loss_sft": 0.9665576592087746, + "loss_total_v6": 1.4311115145683289, + "step": 10 + }, + { + "epoch": 0.013209246472530772, + "grad_norm": 154.06028747558594, + "learning_rate": 5.5e-07, + "loss": 22.121400833129883, + "loss_alignment": 0.92236328125, + "loss_alignment_w": 0.461181640625, + "loss_sft": 0.9185220077633858, + "loss_total_v6": 1.38258758187294, + "step": 11 + }, + { + "epoch": 0.01441008706094266, + "grad_norm": 152.06919860839844, + "learning_rate": 6.000000000000001e-07, + "loss": 22.554420471191406, + "loss_alignment": 0.9619140625, + "loss_alignment_w": 0.48095703125, + "loss_sft": 0.926359660923481, + "loss_total_v6": 1.409651279449463, + "step": 12 + }, + { + "epoch": 0.015610927649354548, + "grad_norm": 144.66363525390625, + "learning_rate": 6.5e-07, + "loss": 22.0379581451416, + "loss_alignment": 0.95458984375, + "loss_alignment_w": 0.477294921875, + "loss_sft": 0.8979565054178238, + "loss_total_v6": 1.3773724138736725, + "step": 13 + }, + { + "epoch": 0.016811768237766436, + "grad_norm": 99.39318084716797, + "learning_rate": 7.000000000000001e-07, + "loss": 20.396865844726562, + "loss_alignment": 0.890625, + "loss_alignment_w": 0.4453125, + "loss_sft": 0.8284081965684891, + "loss_total_v6": 1.2748040854930878, + "step": 14 + }, + { + "epoch": 0.018012608826178327, + "grad_norm": 92.45083618164062, + "learning_rate": 7.5e-07, + "loss": 21.163644790649414, + "loss_alignment": 0.953125, + "loss_alignment_w": 0.4765625, + "loss_sft": 0.8454328402876854, + "loss_total_v6": 1.3227278143167496, + "step": 15 + }, + { + "epoch": 0.019213449414590213, + "grad_norm": 90.05431365966797, + "learning_rate": 8.000000000000001e-07, + "loss": 20.793277740478516, + "loss_alignment": 0.96142578125, + "loss_alignment_w": 0.480712890625, + "loss_sft": 0.8189126998186111, + "loss_total_v6": 1.299579843878746, + "step": 16 + }, + { + "epoch": 0.0204142900030021, + "grad_norm": 76.92969512939453, + "learning_rate": 8.500000000000001e-07, + "loss": 20.303958892822266, + "loss_alignment": 0.9296875, + "loss_alignment_w": 0.46484375, + "loss_sft": 0.8049319237470627, + "loss_total_v6": 1.2689974755048752, + "step": 17 + }, + { + "epoch": 0.02161513059141399, + "grad_norm": 75.47506713867188, + "learning_rate": 9.000000000000001e-07, + "loss": 20.686416625976562, + "loss_alignment": 0.953125, + "loss_alignment_w": 0.4765625, + "loss_sft": 0.8174983263015747, + "loss_total_v6": 1.2929010689258575, + "step": 18 + }, + { + "epoch": 0.022815971179825877, + "grad_norm": 64.40116882324219, + "learning_rate": 9.500000000000001e-07, + "loss": 20.026031494140625, + "loss_alignment": 0.91064453125, + "loss_alignment_w": 0.455322265625, + "loss_sft": 0.7977999970316887, + "loss_total_v6": 1.251626878976822, + "step": 19 + }, + { + "epoch": 0.024016811768237768, + "grad_norm": 76.72650146484375, + "learning_rate": 1.0000000000000002e-06, + "loss": 19.87127685546875, + "loss_alignment": 0.921875, + "loss_alignment_w": 0.4609375, + "loss_sft": 0.7835350036621094, + "loss_total_v6": 1.2419548481702805, + "step": 20 + }, + { + "epoch": 0.025217652356649654, + "grad_norm": 60.478336334228516, + "learning_rate": 1.0500000000000001e-06, + "loss": 19.906423568725586, + "loss_alignment": 0.94775390625, + "loss_alignment_w": 0.473876953125, + "loss_sft": 0.7722734361886978, + "loss_total_v6": 1.2441515028476715, + "step": 21 + }, + { + "epoch": 0.026418492945061545, + "grad_norm": 50.64890670776367, + "learning_rate": 1.1e-06, + "loss": 19.06339454650879, + "loss_alignment": 0.86962890625, + "loss_alignment_w": 0.434814453125, + "loss_sft": 0.75934848934412, + "loss_total_v6": 1.191462129354477, + "step": 22 + }, + { + "epoch": 0.02761933353347343, + "grad_norm": 48.15632629394531, + "learning_rate": 1.1500000000000002e-06, + "loss": 19.33170509338379, + "loss_alignment": 0.91162109375, + "loss_alignment_w": 0.455810546875, + "loss_sft": 0.7518564984202385, + "loss_total_v6": 1.2082316130399704, + "step": 23 + }, + { + "epoch": 0.02882017412188532, + "grad_norm": 44.23044967651367, + "learning_rate": 1.2000000000000002e-06, + "loss": 18.774185180664062, + "loss_alignment": 0.91455078125, + "loss_alignment_w": 0.457275390625, + "loss_sft": 0.7124642953276634, + "loss_total_v6": 1.1733865141868591, + "step": 24 + }, + { + "epoch": 0.03002101471029721, + "grad_norm": 44.27703094482422, + "learning_rate": 1.25e-06, + "loss": 19.3323974609375, + "loss_alignment": 0.91015625, + "loss_alignment_w": 0.455078125, + "loss_sft": 0.7495499476790428, + "loss_total_v6": 1.2082748860120773, + "step": 25 + }, + { + "epoch": 0.031221855298709095, + "grad_norm": 38.145599365234375, + "learning_rate": 1.3e-06, + "loss": 18.56531524658203, + "loss_alignment": 0.86865234375, + "loss_alignment_w": 0.434326171875, + "loss_sft": 0.7241596579551697, + "loss_total_v6": 1.1603321731090546, + "step": 26 + }, + { + "epoch": 0.03242269588712098, + "grad_norm": 40.08750915527344, + "learning_rate": 1.3500000000000002e-06, + "loss": 18.652450561523438, + "loss_alignment": 0.87158203125, + "loss_alignment_w": 0.435791015625, + "loss_sft": 0.7329320982098579, + "loss_total_v6": 1.1657781302928925, + "step": 27 + }, + { + "epoch": 0.03362353647553287, + "grad_norm": 44.167179107666016, + "learning_rate": 1.4000000000000001e-06, + "loss": 17.974018096923828, + "loss_alignment": 0.84228515625, + "loss_alignment_w": 0.421142578125, + "loss_sft": 0.7005550563335419, + "loss_total_v6": 1.1233760714530945, + "step": 28 + }, + { + "epoch": 0.03482437706394476, + "grad_norm": 40.96624755859375, + "learning_rate": 1.45e-06, + "loss": 18.16710662841797, + "loss_alignment": 0.85205078125, + "loss_alignment_w": 0.426025390625, + "loss_sft": 0.7069162577390671, + "loss_total_v6": 1.1354440748691559, + "step": 29 + }, + { + "epoch": 0.03602521765235665, + "grad_norm": 39.131893157958984, + "learning_rate": 1.5e-06, + "loss": 18.174560546875, + "loss_alignment": 0.84326171875, + "loss_alignment_w": 0.421630859375, + "loss_sft": 0.7133485078811646, + "loss_total_v6": 1.1359100937843323, + "step": 30 + }, + { + "epoch": 0.037226058240768536, + "grad_norm": 32.85837936401367, + "learning_rate": 1.5500000000000002e-06, + "loss": 17.662914276123047, + "loss_alignment": 0.8046875, + "loss_alignment_w": 0.40234375, + "loss_sft": 0.7029311954975128, + "loss_total_v6": 1.1039321720600128, + "step": 31 + }, + { + "epoch": 0.03842689882918043, + "grad_norm": 32.66069412231445, + "learning_rate": 1.6000000000000001e-06, + "loss": 17.747953414916992, + "loss_alignment": 0.8095703125, + "loss_alignment_w": 0.40478515625, + "loss_sft": 0.7065676152706146, + "loss_total_v6": 1.1092470735311508, + "step": 32 + }, + { + "epoch": 0.03962773941759232, + "grad_norm": 28.89281463623047, + "learning_rate": 1.6500000000000003e-06, + "loss": 17.129390716552734, + "loss_alignment": 0.78466796875, + "loss_alignment_w": 0.392333984375, + "loss_sft": 0.6792753487825394, + "loss_total_v6": 1.0705869793891907, + "step": 33 + }, + { + "epoch": 0.0408285800060042, + "grad_norm": 29.122234344482422, + "learning_rate": 1.7000000000000002e-06, + "loss": 17.103179931640625, + "loss_alignment": 0.7900390625, + "loss_alignment_w": 0.39501953125, + "loss_sft": 0.672662653028965, + "loss_total_v6": 1.068948671221733, + "step": 34 + }, + { + "epoch": 0.04202942059441609, + "grad_norm": 26.229719161987305, + "learning_rate": 1.75e-06, + "loss": 17.085567474365234, + "loss_alignment": 0.7822265625, + "loss_alignment_w": 0.39111328125, + "loss_sft": 0.6736523509025574, + "loss_total_v6": 1.0678479373455048, + "step": 35 + }, + { + "epoch": 0.04323026118282798, + "grad_norm": 27.475351333618164, + "learning_rate": 1.8000000000000001e-06, + "loss": 16.733558654785156, + "loss_alignment": 0.7685546875, + "loss_alignment_w": 0.38427734375, + "loss_sft": 0.6615091562271118, + "loss_total_v6": 1.0458474904298782, + "step": 36 + }, + { + "epoch": 0.04443110177123987, + "grad_norm": 27.9060115814209, + "learning_rate": 1.85e-06, + "loss": 16.606050491333008, + "loss_alignment": 0.7490234375, + "loss_alignment_w": 0.37451171875, + "loss_sft": 0.6649075374007225, + "loss_total_v6": 1.0378781259059906, + "step": 37 + }, + { + "epoch": 0.045631942359651755, + "grad_norm": 28.860977172851562, + "learning_rate": 1.9000000000000002e-06, + "loss": 16.364057540893555, + "loss_alignment": 0.73876953125, + "loss_alignment_w": 0.369384765625, + "loss_sft": 0.6527889966964722, + "loss_total_v6": 1.022753544151783, + "step": 38 + }, + { + "epoch": 0.046832782948063645, + "grad_norm": 25.084413528442383, + "learning_rate": 1.9500000000000004e-06, + "loss": 16.211116790771484, + "loss_alignment": 0.7255859375, + "loss_alignment_w": 0.36279296875, + "loss_sft": 0.6488910987973213, + "loss_total_v6": 1.0131946578621864, + "step": 39 + }, + { + "epoch": 0.048033623536475535, + "grad_norm": 22.567272186279297, + "learning_rate": 2.0000000000000003e-06, + "loss": 15.983833312988281, + "loss_alignment": 0.72412109375, + "loss_alignment_w": 0.362060546875, + "loss_sft": 0.6356320381164551, + "loss_total_v6": 0.9989895969629288, + "step": 40 + }, + { + "epoch": 0.04923446412488742, + "grad_norm": 23.858991622924805, + "learning_rate": 2.05e-06, + "loss": 15.74779987335205, + "loss_alignment": 0.6708984375, + "loss_alignment_w": 0.33544921875, + "loss_sft": 0.6479490324854851, + "loss_total_v6": 0.9842375293374062, + "step": 41 + }, + { + "epoch": 0.05043530471329931, + "grad_norm": 22.86210823059082, + "learning_rate": 2.1000000000000002e-06, + "loss": 15.872639656066895, + "loss_alignment": 0.70166015625, + "loss_alignment_w": 0.350830078125, + "loss_sft": 0.6419422999024391, + "loss_total_v6": 0.9920399412512779, + "step": 42 + }, + { + "epoch": 0.0516361453017112, + "grad_norm": 22.467544555664062, + "learning_rate": 2.15e-06, + "loss": 15.803930282592773, + "loss_alignment": 0.69091796875, + "loss_alignment_w": 0.345458984375, + "loss_sft": 0.6435683593153954, + "loss_total_v6": 0.9877456054091454, + "step": 43 + }, + { + "epoch": 0.05283698589012309, + "grad_norm": 21.971820831298828, + "learning_rate": 2.2e-06, + "loss": 15.966170310974121, + "loss_alignment": 0.69482421875, + "loss_alignment_w": 0.347412109375, + "loss_sft": 0.6523503661155701, + "loss_total_v6": 0.9978856444358826, + "step": 44 + }, + { + "epoch": 0.05403782647853497, + "grad_norm": 19.46234893798828, + "learning_rate": 2.25e-06, + "loss": 15.608603477478027, + "loss_alignment": 0.68505859375, + "loss_alignment_w": 0.342529296875, + "loss_sft": 0.6333746388554573, + "loss_total_v6": 0.9755377098917961, + "step": 45 + }, + { + "epoch": 0.05523866706694686, + "grad_norm": 19.503480911254883, + "learning_rate": 2.3000000000000004e-06, + "loss": 15.681380271911621, + "loss_alignment": 0.673828125, + "loss_alignment_w": 0.3369140625, + "loss_sft": 0.6423176899552345, + "loss_total_v6": 0.9800862744450569, + "step": 46 + }, + { + "epoch": 0.05643950765535875, + "grad_norm": 20.760461807250977, + "learning_rate": 2.35e-06, + "loss": 15.540517807006836, + "loss_alignment": 0.67041015625, + "loss_alignment_w": 0.335205078125, + "loss_sft": 0.6338800489902496, + "loss_total_v6": 0.9712824076414108, + "step": 47 + }, + { + "epoch": 0.05764034824377064, + "grad_norm": 20.50266456604004, + "learning_rate": 2.4000000000000003e-06, + "loss": 15.329458236694336, + "loss_alignment": 0.66845703125, + "loss_alignment_w": 0.334228515625, + "loss_sft": 0.6225350648164749, + "loss_total_v6": 0.9580910950899124, + "step": 48 + }, + { + "epoch": 0.05884118883218253, + "grad_norm": 18.917444229125977, + "learning_rate": 2.4500000000000003e-06, + "loss": 15.368103981018066, + "loss_alignment": 0.650390625, + "loss_alignment_w": 0.3251953125, + "loss_sft": 0.634075254201889, + "loss_total_v6": 0.9605065137147903, + "step": 49 + }, + { + "epoch": 0.06004202942059442, + "grad_norm": 19.14118766784668, + "learning_rate": 2.5e-06, + "loss": 15.318925857543945, + "loss_alignment": 0.6552734375, + "loss_alignment_w": 0.32763671875, + "loss_sft": 0.6307573914527893, + "loss_total_v6": 0.9574328437447548, + "step": 50 + }, + { + "epoch": 0.06124287000900631, + "grad_norm": 20.24860191345215, + "learning_rate": 2.55e-06, + "loss": 14.829022407531738, + "loss_alignment": 0.5966796875, + "loss_alignment_w": 0.29833984375, + "loss_sft": 0.6296032518148422, + "loss_total_v6": 0.9268139377236366, + "step": 51 + }, + { + "epoch": 0.06244371059741819, + "grad_norm": 19.014110565185547, + "learning_rate": 2.6e-06, + "loss": 15.246732711791992, + "loss_alignment": 0.625, + "loss_alignment_w": 0.3125, + "loss_sft": 0.6404055282473564, + "loss_total_v6": 0.9529207944869995, + "step": 52 + }, + { + "epoch": 0.06364455118583008, + "grad_norm": 17.819082260131836, + "learning_rate": 2.6500000000000005e-06, + "loss": 15.061542510986328, + "loss_alignment": 0.638671875, + "loss_alignment_w": 0.3193359375, + "loss_sft": 0.6214916855096817, + "loss_total_v6": 0.9413464292883873, + "step": 53 + }, + { + "epoch": 0.06484539177424196, + "grad_norm": 18.729915618896484, + "learning_rate": 2.7000000000000004e-06, + "loss": 14.905588150024414, + "loss_alignment": 0.61181640625, + "loss_alignment_w": 0.305908203125, + "loss_sft": 0.6248823180794716, + "loss_total_v6": 0.9315992072224617, + "step": 54 + }, + { + "epoch": 0.06604623236265386, + "grad_norm": 19.099245071411133, + "learning_rate": 2.7500000000000004e-06, + "loss": 14.79786491394043, + "loss_alignment": 0.6083984375, + "loss_alignment_w": 0.30419921875, + "loss_sft": 0.6197671070694923, + "loss_total_v6": 0.9248665869235992, + "step": 55 + }, + { + "epoch": 0.06724707295106575, + "grad_norm": 18.123111724853516, + "learning_rate": 2.8000000000000003e-06, + "loss": 14.59544563293457, + "loss_alignment": 0.58154296875, + "loss_alignment_w": 0.290771484375, + "loss_sft": 0.6203910186886787, + "loss_total_v6": 0.9122153669595718, + "step": 56 + }, + { + "epoch": 0.06844791353947763, + "grad_norm": 17.95774269104004, + "learning_rate": 2.85e-06, + "loss": 14.49292278289795, + "loss_alignment": 0.5888671875, + "loss_alignment_w": 0.29443359375, + "loss_sft": 0.6108400151133537, + "loss_total_v6": 0.9058076590299606, + "step": 57 + }, + { + "epoch": 0.06964875412788953, + "grad_norm": 21.067527770996094, + "learning_rate": 2.9e-06, + "loss": 14.430557250976562, + "loss_alignment": 0.59228515625, + "loss_alignment_w": 0.296142578125, + "loss_sft": 0.6048516780138016, + "loss_total_v6": 0.9019098058342934, + "step": 58 + }, + { + "epoch": 0.07084959471630141, + "grad_norm": 18.833768844604492, + "learning_rate": 2.95e-06, + "loss": 14.756272315979004, + "loss_alignment": 0.60009765625, + "loss_alignment_w": 0.300048828125, + "loss_sft": 0.6224928498268127, + "loss_total_v6": 0.9222670570015907, + "step": 59 + }, + { + "epoch": 0.0720504353047133, + "grad_norm": 19.55579376220703, + "learning_rate": 3e-06, + "loss": 14.284360885620117, + "loss_alignment": 0.59716796875, + "loss_alignment_w": 0.298583984375, + "loss_sft": 0.5948751866817474, + "loss_total_v6": 0.8927725329995155, + "step": 60 + }, + { + "epoch": 0.07325127589312519, + "grad_norm": 17.28466796875, + "learning_rate": 3.05e-06, + "loss": 14.286556243896484, + "loss_alignment": 0.5595703125, + "loss_alignment_w": 0.27978515625, + "loss_sft": 0.6135518625378609, + "loss_total_v6": 0.8929097503423691, + "step": 61 + }, + { + "epoch": 0.07445211648153707, + "grad_norm": 17.653114318847656, + "learning_rate": 3.1000000000000004e-06, + "loss": 14.50197696685791, + "loss_alignment": 0.561279296875, + "loss_alignment_w": 0.2806396484375, + "loss_sft": 0.6259322687983513, + "loss_total_v6": 0.9063735380768776, + "step": 62 + }, + { + "epoch": 0.07565295706994897, + "grad_norm": 18.906156539916992, + "learning_rate": 3.1500000000000003e-06, + "loss": 14.476066589355469, + "loss_alignment": 0.58349609375, + "loss_alignment_w": 0.291748046875, + "loss_sft": 0.6139216125011444, + "loss_total_v6": 0.9047541171312332, + "step": 63 + }, + { + "epoch": 0.07685379765836085, + "grad_norm": 17.438262939453125, + "learning_rate": 3.2000000000000003e-06, + "loss": 14.299430847167969, + "loss_alignment": 0.56982421875, + "loss_alignment_w": 0.284912109375, + "loss_sft": 0.6088938489556313, + "loss_total_v6": 0.8937143981456757, + "step": 64 + }, + { + "epoch": 0.07805463824677274, + "grad_norm": 19.33206558227539, + "learning_rate": 3.2500000000000002e-06, + "loss": 14.323710441589355, + "loss_alignment": 0.56689453125, + "loss_alignment_w": 0.283447265625, + "loss_sft": 0.612105056643486, + "loss_total_v6": 0.8952319025993347, + "step": 65 + }, + { + "epoch": 0.07925547883518463, + "grad_norm": 18.30605697631836, + "learning_rate": 3.3000000000000006e-06, + "loss": 14.330591201782227, + "loss_alignment": 0.564453125, + "loss_alignment_w": 0.2822265625, + "loss_sft": 0.6138931959867477, + "loss_total_v6": 0.8956619948148727, + "step": 66 + }, + { + "epoch": 0.08045631942359652, + "grad_norm": 18.95276641845703, + "learning_rate": 3.3500000000000005e-06, + "loss": 13.929073333740234, + "loss_alignment": 0.54296875, + "loss_alignment_w": 0.271484375, + "loss_sft": 0.5990674793720245, + "loss_total_v6": 0.8705670982599258, + "step": 67 + }, + { + "epoch": 0.0816571600120084, + "grad_norm": 18.362070083618164, + "learning_rate": 3.4000000000000005e-06, + "loss": 14.193041801452637, + "loss_alignment": 0.565185546875, + "loss_alignment_w": 0.2825927734375, + "loss_sft": 0.6033126637339592, + "loss_total_v6": 0.8870651200413704, + "step": 68 + }, + { + "epoch": 0.0828580006004203, + "grad_norm": 18.54249382019043, + "learning_rate": 3.45e-06, + "loss": 13.814424514770508, + "loss_alignment": 0.531494140625, + "loss_alignment_w": 0.2657470703125, + "loss_sft": 0.596845731139183, + "loss_total_v6": 0.8634014949202538, + "step": 69 + }, + { + "epoch": 0.08405884118883218, + "grad_norm": 18.532800674438477, + "learning_rate": 3.5e-06, + "loss": 13.886805534362793, + "loss_alignment": 0.5380859375, + "loss_alignment_w": 0.26904296875, + "loss_sft": 0.5989891812205315, + "loss_total_v6": 0.867925338447094, + "step": 70 + }, + { + "epoch": 0.08525968177724406, + "grad_norm": 18.512670516967773, + "learning_rate": 3.5500000000000003e-06, + "loss": 13.94318962097168, + "loss_alignment": 0.538818359375, + "loss_alignment_w": 0.2694091796875, + "loss_sft": 0.6025590002536774, + "loss_total_v6": 0.8714493736624718, + "step": 71 + }, + { + "epoch": 0.08646052236565596, + "grad_norm": 17.929311752319336, + "learning_rate": 3.6000000000000003e-06, + "loss": 13.941381454467773, + "loss_alignment": 0.5185546875, + "loss_alignment_w": 0.25927734375, + "loss_sft": 0.6127150803804398, + "loss_total_v6": 0.8713363260030746, + "step": 72 + }, + { + "epoch": 0.08766136295406785, + "grad_norm": 18.109718322753906, + "learning_rate": 3.65e-06, + "loss": 13.68699836730957, + "loss_alignment": 0.516357421875, + "loss_alignment_w": 0.2581787109375, + "loss_sft": 0.5965262278914452, + "loss_total_v6": 0.8554373532533646, + "step": 73 + }, + { + "epoch": 0.08886220354247974, + "grad_norm": 18.32699203491211, + "learning_rate": 3.7e-06, + "loss": 13.949085235595703, + "loss_alignment": 0.5263671875, + "loss_alignment_w": 0.26318359375, + "loss_sft": 0.6087410822510719, + "loss_total_v6": 0.8718178570270538, + "step": 74 + }, + { + "epoch": 0.09006304413089163, + "grad_norm": 17.40155029296875, + "learning_rate": 3.7500000000000005e-06, + "loss": 13.735994338989258, + "loss_alignment": 0.504638671875, + "loss_alignment_w": 0.2523193359375, + "loss_sft": 0.6060582920908928, + "loss_total_v6": 0.8584997057914734, + "step": 75 + }, + { + "epoch": 0.09126388471930351, + "grad_norm": 20.965850830078125, + "learning_rate": 3.8000000000000005e-06, + "loss": 13.365169525146484, + "loss_alignment": 0.5078125, + "loss_alignment_w": 0.25390625, + "loss_sft": 0.581310048699379, + "loss_total_v6": 0.8353231027722359, + "step": 76 + }, + { + "epoch": 0.0924647253077154, + "grad_norm": 18.82003402709961, + "learning_rate": 3.85e-06, + "loss": 13.718546867370605, + "loss_alignment": 0.502197265625, + "loss_alignment_w": 0.2510986328125, + "loss_sft": 0.6065241992473602, + "loss_total_v6": 0.857409194111824, + "step": 77 + }, + { + "epoch": 0.09366556589612729, + "grad_norm": 18.153112411499023, + "learning_rate": 3.900000000000001e-06, + "loss": 13.429386138916016, + "loss_alignment": 0.47607421875, + "loss_alignment_w": 0.238037109375, + "loss_sft": 0.6015589386224747, + "loss_total_v6": 0.8393366560339928, + "step": 78 + }, + { + "epoch": 0.09486640648453917, + "grad_norm": 17.37431526184082, + "learning_rate": 3.95e-06, + "loss": 13.550000190734863, + "loss_alignment": 0.499267578125, + "loss_alignment_w": 0.2496337890625, + "loss_sft": 0.5975616350769997, + "loss_total_v6": 0.8468750268220901, + "step": 79 + }, + { + "epoch": 0.09606724707295107, + "grad_norm": 18.0377140045166, + "learning_rate": 4.000000000000001e-06, + "loss": 13.352758407592773, + "loss_alignment": 0.4814453125, + "loss_alignment_w": 0.24072265625, + "loss_sft": 0.5935348644852638, + "loss_total_v6": 0.8345474302768707, + "step": 80 + }, + { + "epoch": 0.09726808766136295, + "grad_norm": 18.646120071411133, + "learning_rate": 4.05e-06, + "loss": 13.562028884887695, + "loss_alignment": 0.465087890625, + "loss_alignment_w": 0.2325439453125, + "loss_sft": 0.6147929355502129, + "loss_total_v6": 0.8476267829537392, + "step": 81 + }, + { + "epoch": 0.09846892824977484, + "grad_norm": 19.19743537902832, + "learning_rate": 4.1e-06, + "loss": 13.633824348449707, + "loss_alignment": 0.486572265625, + "loss_alignment_w": 0.2432861328125, + "loss_sft": 0.6092551499605179, + "loss_total_v6": 0.8521140292286873, + "step": 82 + }, + { + "epoch": 0.09966976883818673, + "grad_norm": 20.46786880493164, + "learning_rate": 4.15e-06, + "loss": 13.225265502929688, + "loss_alignment": 0.475830078125, + "loss_alignment_w": 0.2379150390625, + "loss_sft": 0.5883588865399361, + "loss_total_v6": 0.8265791013836861, + "step": 83 + }, + { + "epoch": 0.10087060942659862, + "grad_norm": 30.74140167236328, + "learning_rate": 4.2000000000000004e-06, + "loss": 13.286161422729492, + "loss_alignment": 0.47216796875, + "loss_alignment_w": 0.236083984375, + "loss_sft": 0.5944994240999222, + "loss_total_v6": 0.8303850889205933, + "step": 84 + }, + { + "epoch": 0.1020714500150105, + "grad_norm": 98.67473602294922, + "learning_rate": 4.25e-06, + "loss": 13.320695877075195, + "loss_alignment": 0.466552734375, + "loss_alignment_w": 0.2332763671875, + "loss_sft": 0.5994502007961273, + "loss_total_v6": 0.8325434848666191, + "step": 85 + }, + { + "epoch": 0.1032722906034224, + "grad_norm": 19.395931243896484, + "learning_rate": 4.3e-06, + "loss": 13.043434143066406, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.594175823032856, + "loss_total_v6": 0.8152146339416504, + "step": 86 + }, + { + "epoch": 0.10447313119183428, + "grad_norm": 20.196395874023438, + "learning_rate": 4.350000000000001e-06, + "loss": 13.042743682861328, + "loss_alignment": 0.4697265625, + "loss_alignment_w": 0.23486328125, + "loss_sft": 0.5798809677362442, + "loss_total_v6": 0.8151715025305748, + "step": 87 + }, + { + "epoch": 0.10567397178024618, + "grad_norm": 20.368724822998047, + "learning_rate": 4.4e-06, + "loss": 13.10164737701416, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.5920005217194557, + "loss_total_v6": 0.8188529536128044, + "step": 88 + }, + { + "epoch": 0.10687481236865806, + "grad_norm": 18.192522048950195, + "learning_rate": 4.450000000000001e-06, + "loss": 13.12777328491211, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.6034448444843292, + "loss_total_v6": 0.8204858675599098, + "step": 89 + }, + { + "epoch": 0.10807565295706995, + "grad_norm": 19.095006942749023, + "learning_rate": 4.5e-06, + "loss": 12.913753509521484, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.5862996131181717, + "loss_total_v6": 0.8071095570921898, + "step": 90 + }, + { + "epoch": 0.10927649354548184, + "grad_norm": 19.450748443603516, + "learning_rate": 4.5500000000000005e-06, + "loss": 13.02344799041748, + "loss_alignment": 0.4404296875, + "loss_alignment_w": 0.22021484375, + "loss_sft": 0.5941016525030136, + "loss_total_v6": 0.8139655292034149, + "step": 91 + }, + { + "epoch": 0.11047733413389373, + "grad_norm": 20.14071273803711, + "learning_rate": 4.600000000000001e-06, + "loss": 12.743757247924805, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.5788182094693184, + "loss_total_v6": 0.7964848056435585, + "step": 92 + }, + { + "epoch": 0.11167817472230561, + "grad_norm": 19.384336471557617, + "learning_rate": 4.65e-06, + "loss": 12.95509147644043, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.5914620384573936, + "loss_total_v6": 0.8096932396292686, + "step": 93 + }, + { + "epoch": 0.1128790153107175, + "grad_norm": 19.376102447509766, + "learning_rate": 4.7e-06, + "loss": 12.751161575317383, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.5844232216477394, + "loss_total_v6": 0.79694764316082, + "step": 94 + }, + { + "epoch": 0.11407985589912939, + "grad_norm": 19.051918029785156, + "learning_rate": 4.75e-06, + "loss": 12.704109191894531, + "loss_alignment": 0.403564453125, + "loss_alignment_w": 0.2017822265625, + "loss_sft": 0.5923314392566681, + "loss_total_v6": 0.7940068766474724, + "step": 95 + }, + { + "epoch": 0.11528069648754127, + "grad_norm": 18.82223892211914, + "learning_rate": 4.800000000000001e-06, + "loss": 12.706315040588379, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.5873422995209694, + "loss_total_v6": 0.7941446825861931, + "step": 96 + }, + { + "epoch": 0.11648153707595317, + "grad_norm": 19.380216598510742, + "learning_rate": 4.85e-06, + "loss": 12.623498916625977, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.581525482237339, + "loss_total_v6": 0.7889686971902847, + "step": 97 + }, + { + "epoch": 0.11768237766436505, + "grad_norm": 19.686569213867188, + "learning_rate": 4.9000000000000005e-06, + "loss": 12.418291091918945, + "loss_alignment": 0.390625, + "loss_alignment_w": 0.1953125, + "loss_sft": 0.5804797112941742, + "loss_total_v6": 0.7761431634426117, + "step": 98 + }, + { + "epoch": 0.11888321825277694, + "grad_norm": 19.675025939941406, + "learning_rate": 4.95e-06, + "loss": 12.611610412597656, + "loss_alignment": 0.39794921875, + "loss_alignment_w": 0.198974609375, + "loss_sft": 0.5893120467662811, + "loss_total_v6": 0.7882256135344505, + "step": 99 + }, + { + "epoch": 0.12008405884118883, + "grad_norm": 18.4698543548584, + "learning_rate": 5e-06, + "loss": 12.835301399230957, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.5935729295015335, + "loss_total_v6": 0.802206352353096, + "step": 100 + }, + { + "epoch": 0.12008405884118883, + "eval_loss": 0.6967750191688538, + "eval_loss_alignment": 0.39363685021033656, + "eval_loss_alignment_w": 0.19681842510516828, + "eval_loss_sft": 0.49995661510799366, + "eval_loss_total_v6": 0.696775040589273, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 71.0535, + "eval_samples_per_second": 23.349, + "eval_steps_per_second": 2.927, + "step": 100 + }, + { + "epoch": 0.12128489942960072, + "grad_norm": 18.373960494995117, + "learning_rate": 5.050000000000001e-06, + "loss": 12.515024185180664, + "loss_alignment": 0.404296875, + "loss_alignment_w": 0.2021484375, + "loss_sft": 0.5801931396126747, + "loss_total_v6": 0.7821889892220497, + "step": 101 + }, + { + "epoch": 0.12248574001801261, + "grad_norm": 18.914031982421875, + "learning_rate": 5.1e-06, + "loss": 12.305254936218262, + "loss_alignment": 0.3916015625, + "loss_alignment_w": 0.19580078125, + "loss_sft": 0.5727894157171249, + "loss_total_v6": 0.7690784335136414, + "step": 102 + }, + { + "epoch": 0.1236865806064245, + "grad_norm": 18.641374588012695, + "learning_rate": 5.150000000000001e-06, + "loss": 12.429571151733398, + "loss_alignment": 0.39990234375, + "loss_alignment_w": 0.199951171875, + "loss_sft": 0.5767596885561943, + "loss_total_v6": 0.7768482118844986, + "step": 103 + }, + { + "epoch": 0.12488742119483638, + "grad_norm": 18.50574493408203, + "learning_rate": 5.2e-06, + "loss": 12.333516120910645, + "loss_alignment": 0.393310546875, + "loss_alignment_w": 0.1966552734375, + "loss_sft": 0.5743268057703972, + "loss_total_v6": 0.7708447501063347, + "step": 104 + }, + { + "epoch": 0.12608826178324828, + "grad_norm": 18.61127471923828, + "learning_rate": 5.2500000000000006e-06, + "loss": 12.330256462097168, + "loss_alignment": 0.397216796875, + "loss_alignment_w": 0.1986083984375, + "loss_sft": 0.5719563141465187, + "loss_total_v6": 0.7706410139799118, + "step": 105 + }, + { + "epoch": 0.12728910237166016, + "grad_norm": 19.19919776916504, + "learning_rate": 5.300000000000001e-06, + "loss": 12.194719314575195, + "loss_alignment": 0.390625, + "loss_alignment_w": 0.1953125, + "loss_sft": 0.566933773458004, + "loss_total_v6": 0.7621699869632721, + "step": 106 + }, + { + "epoch": 0.12848994296007205, + "grad_norm": 19.35420799255371, + "learning_rate": 5.3500000000000004e-06, + "loss": 12.320907592773438, + "loss_alignment": 0.38671875, + "loss_alignment_w": 0.193359375, + "loss_sft": 0.5770482495427132, + "loss_total_v6": 0.7700566723942757, + "step": 107 + }, + { + "epoch": 0.12969078354848393, + "grad_norm": 21.768569946289062, + "learning_rate": 5.400000000000001e-06, + "loss": 12.326772689819336, + "loss_alignment": 0.384521484375, + "loss_alignment_w": 0.1922607421875, + "loss_sft": 0.5784829556941986, + "loss_total_v6": 0.7704232633113861, + "step": 108 + }, + { + "epoch": 0.13089162413689584, + "grad_norm": 20.858535766601562, + "learning_rate": 5.450000000000001e-06, + "loss": 12.154984474182129, + "loss_alignment": 0.370361328125, + "loss_alignment_w": 0.1851806640625, + "loss_sft": 0.5743532627820969, + "loss_total_v6": 0.7596865370869637, + "step": 109 + }, + { + "epoch": 0.13209246472530772, + "grad_norm": 18.89188003540039, + "learning_rate": 5.500000000000001e-06, + "loss": 12.08928394317627, + "loss_alignment": 0.38037109375, + "loss_alignment_w": 0.190185546875, + "loss_sft": 0.5655930489301682, + "loss_total_v6": 0.7555802538990974, + "step": 110 + }, + { + "epoch": 0.1332933053137196, + "grad_norm": 20.387290954589844, + "learning_rate": 5.550000000000001e-06, + "loss": 12.18515396118164, + "loss_alignment": 0.36865234375, + "loss_alignment_w": 0.184326171875, + "loss_sft": 0.5772001445293427, + "loss_total_v6": 0.7615720853209496, + "step": 111 + }, + { + "epoch": 0.1344941459021315, + "grad_norm": 19.69748306274414, + "learning_rate": 5.600000000000001e-06, + "loss": 11.931169509887695, + "loss_alignment": 0.355712890625, + "loss_alignment_w": 0.1778564453125, + "loss_sft": 0.5681468695402145, + "loss_total_v6": 0.7456981465220451, + "step": 112 + }, + { + "epoch": 0.13569498649054337, + "grad_norm": 20.778657913208008, + "learning_rate": 5.65e-06, + "loss": 12.010578155517578, + "loss_alignment": 0.361083984375, + "loss_alignment_w": 0.1805419921875, + "loss_sft": 0.5699513182044029, + "loss_total_v6": 0.7506611496210098, + "step": 113 + }, + { + "epoch": 0.13689582707895526, + "grad_norm": 19.390981674194336, + "learning_rate": 5.7e-06, + "loss": 11.9735107421875, + "loss_alignment": 0.355224609375, + "loss_alignment_w": 0.1776123046875, + "loss_sft": 0.5707931369543076, + "loss_total_v6": 0.748344399034977, + "step": 114 + }, + { + "epoch": 0.13809666766736717, + "grad_norm": 19.619083404541016, + "learning_rate": 5.75e-06, + "loss": 12.094594955444336, + "loss_alignment": 0.3603515625, + "loss_alignment_w": 0.18017578125, + "loss_sft": 0.5756905972957611, + "loss_total_v6": 0.7559121549129486, + "step": 115 + }, + { + "epoch": 0.13929750825577905, + "grad_norm": 19.489187240600586, + "learning_rate": 5.8e-06, + "loss": 12.218469619750977, + "loss_alignment": 0.367431640625, + "loss_alignment_w": 0.1837158203125, + "loss_sft": 0.579969048500061, + "loss_total_v6": 0.7636543586850166, + "step": 116 + }, + { + "epoch": 0.14049834884419093, + "grad_norm": 20.191062927246094, + "learning_rate": 5.85e-06, + "loss": 11.738313674926758, + "loss_alignment": 0.338623046875, + "loss_alignment_w": 0.1693115234375, + "loss_sft": 0.5647450312972069, + "loss_total_v6": 0.7336445897817612, + "step": 117 + }, + { + "epoch": 0.14169918943260282, + "grad_norm": 20.968852996826172, + "learning_rate": 5.9e-06, + "loss": 12.024965286254883, + "loss_alignment": 0.354736328125, + "loss_alignment_w": 0.1773681640625, + "loss_sft": 0.5745431035757065, + "loss_total_v6": 0.751560315489769, + "step": 118 + }, + { + "epoch": 0.1429000300210147, + "grad_norm": 20.830049514770508, + "learning_rate": 5.950000000000001e-06, + "loss": 11.981098175048828, + "loss_alignment": 0.340087890625, + "loss_alignment_w": 0.1700439453125, + "loss_sft": 0.579079881310463, + "loss_total_v6": 0.7488186582922935, + "step": 119 + }, + { + "epoch": 0.1441008706094266, + "grad_norm": 21.28738784790039, + "learning_rate": 6e-06, + "loss": 11.994365692138672, + "loss_alignment": 0.3486328125, + "loss_alignment_w": 0.17431640625, + "loss_sft": 0.5750262960791588, + "loss_total_v6": 0.7496478632092476, + "step": 120 + }, + { + "epoch": 0.1453017111978385, + "grad_norm": 20.167396545410156, + "learning_rate": 6.0500000000000005e-06, + "loss": 11.98311996459961, + "loss_alignment": 0.346435546875, + "loss_alignment_w": 0.1732177734375, + "loss_sft": 0.5756356790661812, + "loss_total_v6": 0.7489450052380562, + "step": 121 + }, + { + "epoch": 0.14650255178625038, + "grad_norm": 20.404165267944336, + "learning_rate": 6.1e-06, + "loss": 11.710367202758789, + "loss_alignment": 0.332275390625, + "loss_alignment_w": 0.1661376953125, + "loss_sft": 0.5650583878159523, + "loss_total_v6": 0.7318979799747467, + "step": 122 + }, + { + "epoch": 0.14770339237466226, + "grad_norm": 21.81605339050293, + "learning_rate": 6.15e-06, + "loss": 11.724924087524414, + "loss_alignment": 0.344970703125, + "loss_alignment_w": 0.1724853515625, + "loss_sft": 0.5601240172982216, + "loss_total_v6": 0.7328077331185341, + "step": 123 + }, + { + "epoch": 0.14890423296307415, + "grad_norm": 19.203542709350586, + "learning_rate": 6.200000000000001e-06, + "loss": 11.976912498474121, + "loss_alignment": 0.333251953125, + "loss_alignment_w": 0.1666259765625, + "loss_sft": 0.5814122632145882, + "loss_total_v6": 0.7485570311546326, + "step": 124 + }, + { + "epoch": 0.15010507355148603, + "grad_norm": 20.981592178344727, + "learning_rate": 6.25e-06, + "loss": 11.830583572387695, + "loss_alignment": 0.333740234375, + "loss_alignment_w": 0.1668701171875, + "loss_sft": 0.5726634636521339, + "loss_total_v6": 0.7394115254282951, + "step": 125 + }, + { + "epoch": 0.15130591413989794, + "grad_norm": 20.379554748535156, + "learning_rate": 6.300000000000001e-06, + "loss": 11.692924499511719, + "loss_alignment": 0.3369140625, + "loss_alignment_w": 0.16845703125, + "loss_sft": 0.5630678683519363, + "loss_total_v6": 0.73080775141716, + "step": 126 + }, + { + "epoch": 0.15250675472830982, + "grad_norm": 19.734130859375, + "learning_rate": 6.35e-06, + "loss": 11.690948486328125, + "loss_alignment": 0.333740234375, + "loss_alignment_w": 0.1668701171875, + "loss_sft": 0.564149871468544, + "loss_total_v6": 0.7306842803955078, + "step": 127 + }, + { + "epoch": 0.1537075953167217, + "grad_norm": 19.35181999206543, + "learning_rate": 6.4000000000000006e-06, + "loss": 11.750207901000977, + "loss_alignment": 0.326904296875, + "loss_alignment_w": 0.1634521484375, + "loss_sft": 0.57113416492939, + "loss_total_v6": 0.7343879789113998, + "step": 128 + }, + { + "epoch": 0.1549084359051336, + "grad_norm": 18.380294799804688, + "learning_rate": 6.450000000000001e-06, + "loss": 11.599071502685547, + "loss_alignment": 0.31005859375, + "loss_alignment_w": 0.155029296875, + "loss_sft": 0.5702025815844536, + "loss_total_v6": 0.7249419838190079, + "step": 129 + }, + { + "epoch": 0.15610927649354547, + "grad_norm": 18.100481033325195, + "learning_rate": 6.5000000000000004e-06, + "loss": 11.667004585266113, + "loss_alignment": 0.328857421875, + "loss_alignment_w": 0.1644287109375, + "loss_sft": 0.5648811459541321, + "loss_total_v6": 0.7291877865791321, + "step": 130 + }, + { + "epoch": 0.15731011708195736, + "grad_norm": 19.291879653930664, + "learning_rate": 6.550000000000001e-06, + "loss": 11.681163787841797, + "loss_alignment": 0.31982421875, + "loss_alignment_w": 0.159912109375, + "loss_sft": 0.5703132376074791, + "loss_total_v6": 0.7300727590918541, + "step": 131 + }, + { + "epoch": 0.15851095767036927, + "grad_norm": 19.256879806518555, + "learning_rate": 6.600000000000001e-06, + "loss": 11.587900161743164, + "loss_alignment": 0.322021484375, + "loss_alignment_w": 0.1610107421875, + "loss_sft": 0.563202515244484, + "loss_total_v6": 0.7242437824606895, + "step": 132 + }, + { + "epoch": 0.15971179825878115, + "grad_norm": 19.271259307861328, + "learning_rate": 6.650000000000001e-06, + "loss": 11.698934555053711, + "loss_alignment": 0.314453125, + "loss_alignment_w": 0.1572265625, + "loss_sft": 0.5735295861959457, + "loss_total_v6": 0.7311834171414375, + "step": 133 + }, + { + "epoch": 0.16091263884719303, + "grad_norm": 20.704248428344727, + "learning_rate": 6.700000000000001e-06, + "loss": 11.553850173950195, + "loss_alignment": 0.314208984375, + "loss_alignment_w": 0.1571044921875, + "loss_sft": 0.5645228624343872, + "loss_total_v6": 0.7221156284213066, + "step": 134 + }, + { + "epoch": 0.16211347943560492, + "grad_norm": 21.075891494750977, + "learning_rate": 6.750000000000001e-06, + "loss": 11.724282264709473, + "loss_alignment": 0.3134765625, + "loss_alignment_w": 0.15673828125, + "loss_sft": 0.5760141462087631, + "loss_total_v6": 0.7327676713466644, + "step": 135 + }, + { + "epoch": 0.1633143200240168, + "grad_norm": 21.681089401245117, + "learning_rate": 6.800000000000001e-06, + "loss": 11.838336944580078, + "loss_alignment": 0.320068359375, + "loss_alignment_w": 0.1600341796875, + "loss_sft": 0.5800145044922829, + "loss_total_v6": 0.7398960962891579, + "step": 136 + }, + { + "epoch": 0.1645151606124287, + "grad_norm": 19.26380157470703, + "learning_rate": 6.850000000000001e-06, + "loss": 11.528253555297852, + "loss_alignment": 0.306640625, + "loss_alignment_w": 0.1533203125, + "loss_sft": 0.5665851905941963, + "loss_total_v6": 0.7205158621072769, + "step": 137 + }, + { + "epoch": 0.1657160012008406, + "grad_norm": 18.236190795898438, + "learning_rate": 6.9e-06, + "loss": 11.429460525512695, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5624553188681602, + "loss_total_v6": 0.714341290295124, + "step": 138 + }, + { + "epoch": 0.16691684178925248, + "grad_norm": 18.359649658203125, + "learning_rate": 6.95e-06, + "loss": 11.408217430114746, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5614937916398048, + "loss_total_v6": 0.713013581931591, + "step": 139 + }, + { + "epoch": 0.16811768237766436, + "grad_norm": 19.36452293395996, + "learning_rate": 7e-06, + "loss": 11.334491729736328, + "loss_alignment": 0.307373046875, + "loss_alignment_w": 0.1536865234375, + "loss_sft": 0.5545208230614662, + "loss_total_v6": 0.7084057107567787, + "step": 140 + }, + { + "epoch": 0.16931852296607625, + "grad_norm": 19.57249641418457, + "learning_rate": 7.05e-06, + "loss": 11.46609878540039, + "loss_alignment": 0.31396484375, + "loss_alignment_w": 0.156982421875, + "loss_sft": 0.5591757670044899, + "loss_total_v6": 0.7166312038898468, + "step": 141 + }, + { + "epoch": 0.17051936355448813, + "grad_norm": 18.53314781188965, + "learning_rate": 7.100000000000001e-06, + "loss": 11.657434463500977, + "loss_alignment": 0.322265625, + "loss_alignment_w": 0.1611328125, + "loss_sft": 0.5671363398432732, + "loss_total_v6": 0.7285896018147469, + "step": 142 + }, + { + "epoch": 0.17172020414290004, + "grad_norm": 20.676111221313477, + "learning_rate": 7.15e-06, + "loss": 11.592945098876953, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.5733140110969543, + "loss_total_v6": 0.7245590910315514, + "step": 143 + }, + { + "epoch": 0.17292104473131192, + "grad_norm": 18.182674407958984, + "learning_rate": 7.2000000000000005e-06, + "loss": 11.485458374023438, + "loss_alignment": 0.302978515625, + "loss_alignment_w": 0.1514892578125, + "loss_sft": 0.5663366168737411, + "loss_total_v6": 0.7178411185741425, + "step": 144 + }, + { + "epoch": 0.1741218853197238, + "grad_norm": 19.59800910949707, + "learning_rate": 7.25e-06, + "loss": 11.46536922454834, + "loss_alignment": 0.307861328125, + "loss_alignment_w": 0.1539306640625, + "loss_sft": 0.5627922564744949, + "loss_total_v6": 0.7165856063365936, + "step": 145 + }, + { + "epoch": 0.1753227259081357, + "grad_norm": 18.738489151000977, + "learning_rate": 7.3e-06, + "loss": 11.317211151123047, + "loss_alignment": 0.301513671875, + "loss_alignment_w": 0.1507568359375, + "loss_sft": 0.5565230548381805, + "loss_total_v6": 0.7073256522417068, + "step": 146 + }, + { + "epoch": 0.17652356649654757, + "grad_norm": 21.002687454223633, + "learning_rate": 7.350000000000001e-06, + "loss": 11.473021507263184, + "loss_alignment": 0.304931640625, + "loss_alignment_w": 0.1524658203125, + "loss_sft": 0.5642165616154671, + "loss_total_v6": 0.7170638516545296, + "step": 147 + }, + { + "epoch": 0.17772440708495948, + "grad_norm": 19.753053665161133, + "learning_rate": 7.4e-06, + "loss": 11.431808471679688, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.5660810843110085, + "loss_total_v6": 0.7144880443811417, + "step": 148 + }, + { + "epoch": 0.17892524767337137, + "grad_norm": 18.953718185424805, + "learning_rate": 7.450000000000001e-06, + "loss": 11.375835418701172, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.5621097385883331, + "loss_total_v6": 0.710989736020565, + "step": 149 + }, + { + "epoch": 0.18012608826178325, + "grad_norm": 18.29124641418457, + "learning_rate": 7.500000000000001e-06, + "loss": 11.341362953186035, + "loss_alignment": 0.308837890625, + "loss_alignment_w": 0.1544189453125, + "loss_sft": 0.5544925704598427, + "loss_total_v6": 0.7088352292776108, + "step": 150 + }, + { + "epoch": 0.18132692885019513, + "grad_norm": 20.239791870117188, + "learning_rate": 7.5500000000000006e-06, + "loss": 11.27140998840332, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.5560409128665924, + "loss_total_v6": 0.7044631689786911, + "step": 151 + }, + { + "epoch": 0.18252776943860702, + "grad_norm": 18.233776092529297, + "learning_rate": 7.600000000000001e-06, + "loss": 11.404203414916992, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5664156004786491, + "loss_total_v6": 0.7127626687288284, + "step": 152 + }, + { + "epoch": 0.1837286100270189, + "grad_norm": 17.825902938842773, + "learning_rate": 7.650000000000001e-06, + "loss": 11.520368576049805, + "loss_alignment": 0.321533203125, + "loss_alignment_w": 0.1607666015625, + "loss_sft": 0.5591038167476654, + "loss_total_v6": 0.7200229838490486, + "step": 153 + }, + { + "epoch": 0.1849294506154308, + "grad_norm": 17.787363052368164, + "learning_rate": 7.7e-06, + "loss": 11.19700813293457, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.5477286726236343, + "loss_total_v6": 0.6998130083084106, + "step": 154 + }, + { + "epoch": 0.1861302912038427, + "grad_norm": 17.686506271362305, + "learning_rate": 7.75e-06, + "loss": 11.353361129760742, + "loss_alignment": 0.304443359375, + "loss_alignment_w": 0.1522216796875, + "loss_sft": 0.5574701875448227, + "loss_total_v6": 0.7095850706100464, + "step": 155 + }, + { + "epoch": 0.18733113179225458, + "grad_norm": 18.108062744140625, + "learning_rate": 7.800000000000002e-06, + "loss": 11.232809066772461, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.5512479767203331, + "loss_total_v6": 0.7020506113767624, + "step": 156 + }, + { + "epoch": 0.18853197238066646, + "grad_norm": 17.37985610961914, + "learning_rate": 7.850000000000001e-06, + "loss": 11.372188568115234, + "loss_alignment": 0.310791015625, + "loss_alignment_w": 0.1553955078125, + "loss_sft": 0.5553662553429604, + "loss_total_v6": 0.710761770606041, + "step": 157 + }, + { + "epoch": 0.18973281296907835, + "grad_norm": 18.024606704711914, + "learning_rate": 7.9e-06, + "loss": 11.095492362976074, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5469075813889503, + "loss_total_v6": 0.6934682428836823, + "step": 158 + }, + { + "epoch": 0.19093365355749023, + "grad_norm": 17.355154037475586, + "learning_rate": 7.950000000000002e-06, + "loss": 11.428421974182129, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5649233162403107, + "loss_total_v6": 0.7142763510346413, + "step": 159 + }, + { + "epoch": 0.19213449414590214, + "grad_norm": 18.201440811157227, + "learning_rate": 8.000000000000001e-06, + "loss": 11.299263000488281, + "loss_alignment": 0.30517578125, + "loss_alignment_w": 0.152587890625, + "loss_sft": 0.5537839084863663, + "loss_total_v6": 0.7062039449810982, + "step": 160 + }, + { + "epoch": 0.19333533473431402, + "grad_norm": 17.452972412109375, + "learning_rate": 8.050000000000001e-06, + "loss": 11.621477127075195, + "loss_alignment": 0.3076171875, + "loss_alignment_w": 0.15380859375, + "loss_sft": 0.5724879205226898, + "loss_total_v6": 0.7263423055410385, + "step": 161 + }, + { + "epoch": 0.1945361753227259, + "grad_norm": 17.6041202545166, + "learning_rate": 8.1e-06, + "loss": 11.472217559814453, + "loss_alignment": 0.300048828125, + "loss_alignment_w": 0.1500244140625, + "loss_sft": 0.5671570897102356, + "loss_total_v6": 0.7170136496424675, + "step": 162 + }, + { + "epoch": 0.1957370159111378, + "grad_norm": 16.704477310180664, + "learning_rate": 8.15e-06, + "loss": 11.365734100341797, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.5607459470629692, + "loss_total_v6": 0.7103584185242653, + "step": 163 + }, + { + "epoch": 0.19693785649954967, + "grad_norm": 17.80534553527832, + "learning_rate": 8.2e-06, + "loss": 11.1963472366333, + "loss_alignment": 0.316650390625, + "loss_alignment_w": 0.1583251953125, + "loss_sft": 0.5411108061671257, + "loss_total_v6": 0.6997716724872589, + "step": 164 + }, + { + "epoch": 0.19813869708796158, + "grad_norm": 16.307361602783203, + "learning_rate": 8.25e-06, + "loss": 11.554378509521484, + "loss_alignment": 0.309326171875, + "loss_alignment_w": 0.1546630859375, + "loss_sft": 0.5675923749804497, + "loss_total_v6": 0.7221486568450928, + "step": 165 + }, + { + "epoch": 0.19933953767637347, + "grad_norm": 16.695676803588867, + "learning_rate": 8.3e-06, + "loss": 11.269201278686523, + "loss_alignment": 0.302978515625, + "loss_alignment_w": 0.1514892578125, + "loss_sft": 0.5527289733290672, + "loss_total_v6": 0.7043250501155853, + "step": 166 + }, + { + "epoch": 0.20054037826478535, + "grad_norm": 18.714075088500977, + "learning_rate": 8.35e-06, + "loss": 11.2262544631958, + "loss_alignment": 0.306640625, + "loss_alignment_w": 0.1533203125, + "loss_sft": 0.5482900887727737, + "loss_total_v6": 0.7016409263014793, + "step": 167 + }, + { + "epoch": 0.20174121885319724, + "grad_norm": 17.490022659301758, + "learning_rate": 8.400000000000001e-06, + "loss": 11.412918090820312, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.5635118409991264, + "loss_total_v6": 0.7133073955774307, + "step": 168 + }, + { + "epoch": 0.20294205944160912, + "grad_norm": 17.46786880493164, + "learning_rate": 8.45e-06, + "loss": 11.169425964355469, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.5540155693888664, + "loss_total_v6": 0.6980890780687332, + "step": 169 + }, + { + "epoch": 0.204142900030021, + "grad_norm": 16.937700271606445, + "learning_rate": 8.5e-06, + "loss": 11.44674301147461, + "loss_alignment": 0.30517578125, + "loss_alignment_w": 0.152587890625, + "loss_sft": 0.5627877786755562, + "loss_total_v6": 0.7154214382171631, + "step": 170 + }, + { + "epoch": 0.2053437406184329, + "grad_norm": 16.34071922302246, + "learning_rate": 8.550000000000001e-06, + "loss": 11.395407676696777, + "loss_alignment": 0.3056640625, + "loss_alignment_w": 0.15283203125, + "loss_sft": 0.5597471222281456, + "loss_total_v6": 0.7122129499912262, + "step": 171 + }, + { + "epoch": 0.2065445812068448, + "grad_norm": 18.063827514648438, + "learning_rate": 8.6e-06, + "loss": 11.579654693603516, + "loss_alignment": 0.306396484375, + "loss_alignment_w": 0.1531982421875, + "loss_sft": 0.5709574148058891, + "loss_total_v6": 0.7237284183502197, + "step": 172 + }, + { + "epoch": 0.20774542179525668, + "grad_norm": 16.664209365844727, + "learning_rate": 8.65e-06, + "loss": 11.48653793334961, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5664956718683243, + "loss_total_v6": 0.7179086208343506, + "step": 173 + }, + { + "epoch": 0.20894626238366856, + "grad_norm": 16.80164909362793, + "learning_rate": 8.700000000000001e-06, + "loss": 11.236305236816406, + "loss_alignment": 0.30615234375, + "loss_alignment_w": 0.153076171875, + "loss_sft": 0.5492692440748215, + "loss_total_v6": 0.7022691071033478, + "step": 174 + }, + { + "epoch": 0.21014710297208045, + "grad_norm": 17.60702133178711, + "learning_rate": 8.750000000000001e-06, + "loss": 11.403547286987305, + "loss_alignment": 0.301025390625, + "loss_alignment_w": 0.1505126953125, + "loss_sft": 0.5619801506400108, + "loss_total_v6": 0.7127217054367065, + "step": 175 + }, + { + "epoch": 0.21134794356049236, + "grad_norm": 17.47794532775879, + "learning_rate": 8.8e-06, + "loss": 11.161784172058105, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5459086447954178, + "loss_total_v6": 0.6976115256547928, + "step": 176 + }, + { + "epoch": 0.21254878414890424, + "grad_norm": 16.756771087646484, + "learning_rate": 8.85e-06, + "loss": 11.444368362426758, + "loss_alignment": 0.304931640625, + "loss_alignment_w": 0.1524658203125, + "loss_sft": 0.5626699328422546, + "loss_total_v6": 0.7152730450034142, + "step": 177 + }, + { + "epoch": 0.21374962473731612, + "grad_norm": 16.543827056884766, + "learning_rate": 8.900000000000001e-06, + "loss": 10.953163146972656, + "loss_alignment": 0.281982421875, + "loss_alignment_w": 0.1409912109375, + "loss_sft": 0.5435509532690048, + "loss_total_v6": 0.6845726743340492, + "step": 178 + }, + { + "epoch": 0.214950465325728, + "grad_norm": 18.157094955444336, + "learning_rate": 8.95e-06, + "loss": 11.339534759521484, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5582998245954514, + "loss_total_v6": 0.7087209671735764, + "step": 179 + }, + { + "epoch": 0.2161513059141399, + "grad_norm": 15.793365478515625, + "learning_rate": 9e-06, + "loss": 11.497632026672363, + "loss_alignment": 0.306884765625, + "loss_alignment_w": 0.1534423828125, + "loss_sft": 0.5649307072162628, + "loss_total_v6": 0.7186020016670227, + "step": 180 + }, + { + "epoch": 0.21735214650255177, + "grad_norm": 16.348413467407227, + "learning_rate": 9.050000000000001e-06, + "loss": 11.380693435668945, + "loss_alignment": 0.306396484375, + "loss_alignment_w": 0.1531982421875, + "loss_sft": 0.5576678216457367, + "loss_total_v6": 0.7112933024764061, + "step": 181 + }, + { + "epoch": 0.21855298709096369, + "grad_norm": 16.03524398803711, + "learning_rate": 9.100000000000001e-06, + "loss": 11.586191177368164, + "loss_alignment": 0.30224609375, + "loss_alignment_w": 0.151123046875, + "loss_sft": 0.573013961315155, + "loss_total_v6": 0.7241369932889938, + "step": 182 + }, + { + "epoch": 0.21975382767937557, + "grad_norm": 17.58668327331543, + "learning_rate": 9.15e-06, + "loss": 11.178691864013672, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5511310324072838, + "loss_total_v6": 0.6986682564020157, + "step": 183 + }, + { + "epoch": 0.22095466826778745, + "grad_norm": 16.031145095825195, + "learning_rate": 9.200000000000002e-06, + "loss": 11.247462272644043, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5506226643919945, + "loss_total_v6": 0.7029663845896721, + "step": 184 + }, + { + "epoch": 0.22215550885619934, + "grad_norm": 17.36590003967285, + "learning_rate": 9.250000000000001e-06, + "loss": 11.32850170135498, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.559944786131382, + "loss_total_v6": 0.7080313488841057, + "step": 185 + }, + { + "epoch": 0.22335634944461122, + "grad_norm": 15.562406539916992, + "learning_rate": 9.3e-06, + "loss": 11.29940414428711, + "loss_alignment": 0.306396484375, + "loss_alignment_w": 0.1531982421875, + "loss_sft": 0.5526025220751762, + "loss_total_v6": 0.7062127739191055, + "step": 186 + }, + { + "epoch": 0.22455719003302313, + "grad_norm": 17.152189254760742, + "learning_rate": 9.350000000000002e-06, + "loss": 11.12330150604248, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.5479895621538162, + "loss_total_v6": 0.6952063590288162, + "step": 187 + }, + { + "epoch": 0.225758030621435, + "grad_norm": 16.013532638549805, + "learning_rate": 9.4e-06, + "loss": 11.243585586547852, + "loss_alignment": 0.309326171875, + "loss_alignment_w": 0.1546630859375, + "loss_sft": 0.5482898727059364, + "loss_total_v6": 0.7027240842580795, + "step": 188 + }, + { + "epoch": 0.2269588712098469, + "grad_norm": 16.23442268371582, + "learning_rate": 9.450000000000001e-06, + "loss": 11.315248489379883, + "loss_alignment": 0.316650390625, + "loss_alignment_w": 0.1583251953125, + "loss_sft": 0.5492897890508175, + "loss_total_v6": 0.7072029933333397, + "step": 189 + }, + { + "epoch": 0.22815971179825878, + "grad_norm": 16.035728454589844, + "learning_rate": 9.5e-06, + "loss": 11.352038383483887, + "loss_alignment": 0.3056640625, + "loss_alignment_w": 0.15283203125, + "loss_sft": 0.5565940514206886, + "loss_total_v6": 0.7095023840665817, + "step": 190 + }, + { + "epoch": 0.22936055238667066, + "grad_norm": 16.21926498413086, + "learning_rate": 9.55e-06, + "loss": 11.369878768920898, + "loss_alignment": 0.298828125, + "loss_alignment_w": 0.1494140625, + "loss_sft": 0.5609591901302338, + "loss_total_v6": 0.7106174156069756, + "step": 191 + }, + { + "epoch": 0.23056139297508255, + "grad_norm": 16.385438919067383, + "learning_rate": 9.600000000000001e-06, + "loss": 11.513836860656738, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.5690716207027435, + "loss_total_v6": 0.7196148335933685, + "step": 192 + }, + { + "epoch": 0.23176223356349446, + "grad_norm": 15.925395011901855, + "learning_rate": 9.65e-06, + "loss": 11.073999404907227, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5446945279836655, + "loss_total_v6": 0.6921249330043793, + "step": 193 + }, + { + "epoch": 0.23296307415190634, + "grad_norm": 16.23680305480957, + "learning_rate": 9.7e-06, + "loss": 11.146970748901367, + "loss_alignment": 0.297607421875, + "loss_alignment_w": 0.1488037109375, + "loss_sft": 0.5479735136032104, + "loss_total_v6": 0.6966856867074966, + "step": 194 + }, + { + "epoch": 0.23416391474031822, + "grad_norm": 15.569156646728516, + "learning_rate": 9.75e-06, + "loss": 11.085140228271484, + "loss_alignment": 0.291259765625, + "loss_alignment_w": 0.1456298828125, + "loss_sft": 0.5472981259226799, + "loss_total_v6": 0.6928212121129036, + "step": 195 + }, + { + "epoch": 0.2353647553287301, + "grad_norm": 15.328362464904785, + "learning_rate": 9.800000000000001e-06, + "loss": 11.06766414642334, + "loss_alignment": 0.30029296875, + "loss_alignment_w": 0.150146484375, + "loss_sft": 0.5417350716888905, + "loss_total_v6": 0.6917289793491364, + "step": 196 + }, + { + "epoch": 0.236565595917142, + "grad_norm": 16.112821578979492, + "learning_rate": 9.85e-06, + "loss": 11.384299278259277, + "loss_alignment": 0.30615234375, + "loss_alignment_w": 0.153076171875, + "loss_sft": 0.5584883317351341, + "loss_total_v6": 0.7115187346935272, + "step": 197 + }, + { + "epoch": 0.23776643650555387, + "grad_norm": 15.900322914123535, + "learning_rate": 9.9e-06, + "loss": 11.102060317993164, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5476384982466698, + "loss_total_v6": 0.6938787400722504, + "step": 198 + }, + { + "epoch": 0.23896727709396579, + "grad_norm": 17.022384643554688, + "learning_rate": 9.950000000000001e-06, + "loss": 11.139970779418945, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.5482378676533699, + "loss_total_v6": 0.6962481290102005, + "step": 199 + }, + { + "epoch": 0.24016811768237767, + "grad_norm": 16.250913619995117, + "learning_rate": 1e-05, + "loss": 11.267572402954102, + "loss_alignment": 0.30029296875, + "loss_alignment_w": 0.150146484375, + "loss_sft": 0.553985245525837, + "loss_total_v6": 0.7042232975363731, + "step": 200 + }, + { + "epoch": 0.24016811768237767, + "eval_loss": 0.632781982421875, + "eval_loss_alignment": 0.2932856633112981, + "eval_loss_alignment_w": 0.14664283165564904, + "eval_loss_sft": 0.48613908457068294, + "eval_loss_total_v6": 0.6327819166024431, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.9337, + "eval_samples_per_second": 24.421, + "eval_steps_per_second": 3.062, + "step": 200 + }, + { + "epoch": 0.24136895827078955, + "grad_norm": 15.668530464172363, + "learning_rate": 9.999938421168647e-06, + "loss": 11.24782943725586, + "loss_alignment": 0.301513671875, + "loss_alignment_w": 0.1507568359375, + "loss_sft": 0.5522020310163498, + "loss_total_v6": 0.7029893919825554, + "step": 201 + }, + { + "epoch": 0.24256979885920144, + "grad_norm": 16.006284713745117, + "learning_rate": 9.999753686191369e-06, + "loss": 11.125907897949219, + "loss_alignment": 0.290283203125, + "loss_alignment_w": 0.1451416015625, + "loss_sft": 0.549739345908165, + "loss_total_v6": 0.6953692585229874, + "step": 202 + }, + { + "epoch": 0.24377063944761332, + "grad_norm": 15.988797187805176, + "learning_rate": 9.99944579961847e-06, + "loss": 11.466083526611328, + "loss_alignment": 0.308349609375, + "loss_alignment_w": 0.1541748046875, + "loss_sft": 0.5623028203845024, + "loss_total_v6": 0.716630220413208, + "step": 203 + }, + { + "epoch": 0.24497148003602523, + "grad_norm": 16.585973739624023, + "learning_rate": 9.99901476903367e-06, + "loss": 11.250412940979004, + "loss_alignment": 0.3037109375, + "loss_alignment_w": 0.15185546875, + "loss_sft": 0.5509138777852058, + "loss_total_v6": 0.7031508013606071, + "step": 204 + }, + { + "epoch": 0.2461723206244371, + "grad_norm": 16.113616943359375, + "learning_rate": 9.998460605053911e-06, + "loss": 11.266327857971191, + "loss_alignment": 0.300537109375, + "loss_alignment_w": 0.1502685546875, + "loss_sft": 0.5540753155946732, + "loss_total_v6": 0.7041455209255219, + "step": 205 + }, + { + "epoch": 0.247373161212849, + "grad_norm": 15.76477336883545, + "learning_rate": 9.997783321329104e-06, + "loss": 11.0635986328125, + "loss_alignment": 0.30908203125, + "loss_alignment_w": 0.154541015625, + "loss_sft": 0.5365219265222549, + "loss_total_v6": 0.6914749443531036, + "step": 206 + }, + { + "epoch": 0.24857400180126088, + "grad_norm": 16.164003372192383, + "learning_rate": 9.996982934541781e-06, + "loss": 11.057146072387695, + "loss_alignment": 0.283447265625, + "loss_alignment_w": 0.1417236328125, + "loss_sft": 0.5493937656283379, + "loss_total_v6": 0.691071629524231, + "step": 207 + }, + { + "epoch": 0.24977484238967276, + "grad_norm": 15.619718551635742, + "learning_rate": 9.9960594644067e-06, + "loss": 11.208463668823242, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.5513437911868095, + "loss_total_v6": 0.7005289867520332, + "step": 208 + }, + { + "epoch": 0.25097568297808465, + "grad_norm": 15.867145538330078, + "learning_rate": 9.995012933670341e-06, + "loss": 11.249981880187988, + "loss_alignment": 0.2998046875, + "loss_alignment_w": 0.14990234375, + "loss_sft": 0.5534350797533989, + "loss_total_v6": 0.7031238526105881, + "step": 209 + }, + { + "epoch": 0.25217652356649656, + "grad_norm": 15.80473518371582, + "learning_rate": 9.993843368110363e-06, + "loss": 10.962508201599121, + "loss_alignment": 0.2978515625, + "loss_alignment_w": 0.14892578125, + "loss_sft": 0.5361394174396992, + "loss_total_v6": 0.6851567625999451, + "step": 210 + }, + { + "epoch": 0.2533773641549084, + "grad_norm": 15.36597728729248, + "learning_rate": 9.992550796534957e-06, + "loss": 11.225578308105469, + "loss_alignment": 0.2991943359375, + "loss_alignment_w": 0.14959716796875, + "loss_sft": 0.5523218661546707, + "loss_total_v6": 0.7015985921025276, + "step": 211 + }, + { + "epoch": 0.2545782047433203, + "grad_norm": 14.912618637084961, + "learning_rate": 9.991135250782143e-06, + "loss": 11.16735553741455, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.5484846085309982, + "loss_total_v6": 0.6979597061872482, + "step": 212 + }, + { + "epoch": 0.25577904533173224, + "grad_norm": 15.370009422302246, + "learning_rate": 9.989596765718981e-06, + "loss": 11.131149291992188, + "loss_alignment": 0.3125, + "loss_alignment_w": 0.15625, + "loss_sft": 0.5391874462366104, + "loss_total_v6": 0.6956968382000923, + "step": 213 + }, + { + "epoch": 0.2569798859201441, + "grad_norm": 15.626402854919434, + "learning_rate": 9.987935379240715e-06, + "loss": 11.034884452819824, + "loss_alignment": 0.291259765625, + "loss_alignment_w": 0.1456298828125, + "loss_sft": 0.5441114827990532, + "loss_total_v6": 0.6896803006529808, + "step": 214 + }, + { + "epoch": 0.258180726508556, + "grad_norm": 15.11711597442627, + "learning_rate": 9.986151132269843e-06, + "loss": 11.168190956115723, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5484910532832146, + "loss_total_v6": 0.6980119422078133, + "step": 215 + }, + { + "epoch": 0.25938156709696786, + "grad_norm": 15.138630867004395, + "learning_rate": 9.9842440687551e-06, + "loss": 11.064812660217285, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.5442119166254997, + "loss_total_v6": 0.6915507987141609, + "step": 216 + }, + { + "epoch": 0.26058240768537977, + "grad_norm": 15.752643585205078, + "learning_rate": 9.982214235670383e-06, + "loss": 11.103343963623047, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.5430800840258598, + "loss_total_v6": 0.6939589828252792, + "step": 217 + }, + { + "epoch": 0.2617832482737917, + "grad_norm": 15.390222549438477, + "learning_rate": 9.980061683013594e-06, + "loss": 11.065828323364258, + "loss_alignment": 0.30322265625, + "loss_alignment_w": 0.151611328125, + "loss_sft": 0.5402317866683006, + "loss_total_v6": 0.6916142627596855, + "step": 218 + }, + { + "epoch": 0.26298408886220354, + "grad_norm": 14.59194278717041, + "learning_rate": 9.977786463805399e-06, + "loss": 10.905159950256348, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5360341519117355, + "loss_total_v6": 0.6815724819898605, + "step": 219 + }, + { + "epoch": 0.26418492945061545, + "grad_norm": 14.941032409667969, + "learning_rate": 9.97538863408794e-06, + "loss": 11.16221809387207, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.5500099062919617, + "loss_total_v6": 0.6976386681199074, + "step": 220 + }, + { + "epoch": 0.2653857700390273, + "grad_norm": 16.580644607543945, + "learning_rate": 9.972868252923433e-06, + "loss": 11.202434539794922, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.5564754009246826, + "loss_total_v6": 0.7001521661877632, + "step": 221 + }, + { + "epoch": 0.2665866106274392, + "grad_norm": 15.270177841186523, + "learning_rate": 9.970225382392733e-06, + "loss": 10.980694770812988, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.5375354774296284, + "loss_total_v6": 0.6862934231758118, + "step": 222 + }, + { + "epoch": 0.26778745121585107, + "grad_norm": 15.575606346130371, + "learning_rate": 9.967460087593786e-06, + "loss": 11.13050365447998, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5488058552145958, + "loss_total_v6": 0.6956564486026764, + "step": 223 + }, + { + "epoch": 0.268988291804263, + "grad_norm": 15.59407901763916, + "learning_rate": 9.964572436640046e-06, + "loss": 10.972294807434082, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5390703864395618, + "loss_total_v6": 0.6857684105634689, + "step": 224 + }, + { + "epoch": 0.2701891323926749, + "grad_norm": 15.131574630737305, + "learning_rate": 9.961562500658779e-06, + "loss": 11.300939559936523, + "loss_alignment": 0.3038330078125, + "loss_alignment_w": 0.15191650390625, + "loss_sft": 0.5547278970479965, + "loss_total_v6": 0.7063086926937103, + "step": 225 + }, + { + "epoch": 0.27138997298108675, + "grad_norm": 15.048116683959961, + "learning_rate": 9.958430353789321e-06, + "loss": 10.89356803894043, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.5374001041054726, + "loss_total_v6": 0.6808479577302933, + "step": 226 + }, + { + "epoch": 0.27259081356949866, + "grad_norm": 14.99754524230957, + "learning_rate": 9.95517607318125e-06, + "loss": 11.3193359375, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5617981106042862, + "loss_total_v6": 0.7074585109949112, + "step": 227 + }, + { + "epoch": 0.2737916541579105, + "grad_norm": 14.415155410766602, + "learning_rate": 9.951799738992484e-06, + "loss": 11.204935073852539, + "loss_alignment": 0.30615234375, + "loss_alignment_w": 0.153076171875, + "loss_sft": 0.54740010201931, + "loss_total_v6": 0.7003084197640419, + "step": 228 + }, + { + "epoch": 0.2749924947463224, + "grad_norm": 14.657048225402832, + "learning_rate": 9.948301434387308e-06, + "loss": 11.108799934387207, + "loss_alignment": 0.298828125, + "loss_alignment_w": 0.1494140625, + "loss_sft": 0.5449316874146461, + "loss_total_v6": 0.6942999735474586, + "step": 229 + }, + { + "epoch": 0.27619333533473434, + "grad_norm": 14.650943756103516, + "learning_rate": 9.944681245534329e-06, + "loss": 10.950518608093262, + "loss_alignment": 0.3056640625, + "loss_alignment_w": 0.15283203125, + "loss_sft": 0.5317889750003815, + "loss_total_v6": 0.6844074204564095, + "step": 230 + }, + { + "epoch": 0.2773941759231462, + "grad_norm": 14.562070846557617, + "learning_rate": 9.940939261604344e-06, + "loss": 11.168685913085938, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.5490102097392082, + "loss_total_v6": 0.6980428174138069, + "step": 231 + }, + { + "epoch": 0.2785950165115581, + "grad_norm": 14.916683197021484, + "learning_rate": 9.937075574768152e-06, + "loss": 10.875402450561523, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.532801054418087, + "loss_total_v6": 0.6797126829624176, + "step": 232 + }, + { + "epoch": 0.27979585709996996, + "grad_norm": 14.305237770080566, + "learning_rate": 9.93309028019428e-06, + "loss": 10.963587760925293, + "loss_alignment": 0.302734375, + "loss_alignment_w": 0.1513671875, + "loss_sft": 0.533795990049839, + "loss_total_v6": 0.6852242276072502, + "step": 233 + }, + { + "epoch": 0.28099669768838187, + "grad_norm": 14.76286506652832, + "learning_rate": 9.928983476046643e-06, + "loss": 11.165079116821289, + "loss_alignment": 0.2958984375, + "loss_alignment_w": 0.14794921875, + "loss_sft": 0.5499902814626694, + "loss_total_v6": 0.6978174224495888, + "step": 234 + }, + { + "epoch": 0.2821975382767938, + "grad_norm": 14.542924880981445, + "learning_rate": 9.924755263482121e-06, + "loss": 11.324996948242188, + "loss_alignment": 0.309326171875, + "loss_alignment_w": 0.1546630859375, + "loss_sft": 0.5532102957367897, + "loss_total_v6": 0.7078123241662979, + "step": 235 + }, + { + "epoch": 0.28339837886520564, + "grad_norm": 14.990667343139648, + "learning_rate": 9.920405746648067e-06, + "loss": 11.018781661987305, + "loss_alignment": 0.289794921875, + "loss_alignment_w": 0.1448974609375, + "loss_sft": 0.5440967828035355, + "loss_total_v6": 0.688673846423626, + "step": 236 + }, + { + "epoch": 0.28459921945361755, + "grad_norm": 14.327614784240723, + "learning_rate": 9.91593503267975e-06, + "loss": 11.072295188903809, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.5496234372258186, + "loss_total_v6": 0.6920184344053268, + "step": 237 + }, + { + "epoch": 0.2858000600420294, + "grad_norm": 14.20103931427002, + "learning_rate": 9.911343231697703e-06, + "loss": 10.731241226196289, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5207696668803692, + "loss_total_v6": 0.6707025393843651, + "step": 238 + }, + { + "epoch": 0.2870009006304413, + "grad_norm": 14.553077697753906, + "learning_rate": 9.906630456805024e-06, + "loss": 11.03237533569336, + "loss_alignment": 0.307373046875, + "loss_alignment_w": 0.1536865234375, + "loss_sft": 0.5358674824237823, + "loss_total_v6": 0.6895234808325768, + "step": 239 + }, + { + "epoch": 0.2882017412188532, + "grad_norm": 14.51986312866211, + "learning_rate": 9.90179682408458e-06, + "loss": 11.253288269042969, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.5589824020862579, + "loss_total_v6": 0.7033305317163467, + "step": 240 + }, + { + "epoch": 0.2894025818072651, + "grad_norm": 14.746524810791016, + "learning_rate": 9.896842452596151e-06, + "loss": 10.93039321899414, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.534712091088295, + "loss_total_v6": 0.6831495836377144, + "step": 241 + }, + { + "epoch": 0.290603422395677, + "grad_norm": 14.733476638793945, + "learning_rate": 9.891767464373503e-06, + "loss": 11.064056396484375, + "loss_alignment": 0.299560546875, + "loss_alignment_w": 0.1497802734375, + "loss_sft": 0.5420131981372833, + "loss_total_v6": 0.6915035620331764, + "step": 242 + }, + { + "epoch": 0.29180426298408885, + "grad_norm": 14.44231128692627, + "learning_rate": 9.886571984421371e-06, + "loss": 10.775794982910156, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.5270333290100098, + "loss_total_v6": 0.6734871864318848, + "step": 243 + }, + { + "epoch": 0.29300510357250076, + "grad_norm": 14.841545104980469, + "learning_rate": 9.881256140712389e-06, + "loss": 10.955181121826172, + "loss_alignment": 0.300537109375, + "loss_alignment_w": 0.1502685546875, + "loss_sft": 0.5351932123303413, + "loss_total_v6": 0.6846988275647163, + "step": 244 + }, + { + "epoch": 0.2942059441609126, + "grad_norm": 14.675090789794922, + "learning_rate": 9.875820064183936e-06, + "loss": 11.079031944274902, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.5477098822593689, + "loss_total_v6": 0.692439503967762, + "step": 245 + }, + { + "epoch": 0.2954067847493245, + "grad_norm": 15.016912460327148, + "learning_rate": 9.870263888734905e-06, + "loss": 11.043752670288086, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5404390692710876, + "loss_total_v6": 0.690234586596489, + "step": 246 + }, + { + "epoch": 0.29660762533773644, + "grad_norm": 14.591277122497559, + "learning_rate": 9.864587751222416e-06, + "loss": 11.098499298095703, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.543189249932766, + "loss_total_v6": 0.6936561837792397, + "step": 247 + }, + { + "epoch": 0.2978084659261483, + "grad_norm": 14.697257995605469, + "learning_rate": 9.858791791458431e-06, + "loss": 11.023180961608887, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.5462028458714485, + "loss_total_v6": 0.688948817551136, + "step": 248 + }, + { + "epoch": 0.2990093065145602, + "grad_norm": 14.042425155639648, + "learning_rate": 9.852876152206325e-06, + "loss": 10.959819793701172, + "loss_alignment": 0.30126953125, + "loss_alignment_w": 0.150634765625, + "loss_sft": 0.5346896611154079, + "loss_total_v6": 0.6849887445569038, + "step": 249 + }, + { + "epoch": 0.30021014710297206, + "grad_norm": 15.556241035461426, + "learning_rate": 9.846840979177354e-06, + "loss": 10.949596405029297, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.5324332639575005, + "loss_total_v6": 0.6843497678637505, + "step": 250 + }, + { + "epoch": 0.30141098769138397, + "grad_norm": 14.894906044006348, + "learning_rate": 9.840686421027085e-06, + "loss": 11.122945785522461, + "loss_alignment": 0.2998046875, + "loss_alignment_w": 0.14990234375, + "loss_sft": 0.5454495474696159, + "loss_total_v6": 0.6951840445399284, + "step": 251 + }, + { + "epoch": 0.3026118282797959, + "grad_norm": 14.74615478515625, + "learning_rate": 9.834412629351712e-06, + "loss": 11.150127410888672, + "loss_alignment": 0.3001708984375, + "loss_alignment_w": 0.15008544921875, + "loss_sft": 0.5471332520246506, + "loss_total_v6": 0.6968829929828644, + "step": 252 + }, + { + "epoch": 0.30381266886820774, + "grad_norm": 14.449004173278809, + "learning_rate": 9.828019758684343e-06, + "loss": 11.149313926696777, + "loss_alignment": 0.30322265625, + "loss_alignment_w": 0.151611328125, + "loss_sft": 0.5452360138297081, + "loss_total_v6": 0.696832075715065, + "step": 253 + }, + { + "epoch": 0.30501350945661965, + "grad_norm": 14.435112953186035, + "learning_rate": 9.821507966491178e-06, + "loss": 11.010536193847656, + "loss_alignment": 0.296630859375, + "loss_alignment_w": 0.1483154296875, + "loss_sft": 0.5398431047797203, + "loss_total_v6": 0.6881585121154785, + "step": 254 + }, + { + "epoch": 0.3062143500450315, + "grad_norm": 14.169970512390137, + "learning_rate": 9.814877413167635e-06, + "loss": 11.169275283813477, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.5501305162906647, + "loss_total_v6": 0.6980797201395035, + "step": 255 + }, + { + "epoch": 0.3074151906334434, + "grad_norm": 14.606317520141602, + "learning_rate": 9.808128262034411e-06, + "loss": 10.901815414428711, + "loss_alignment": 0.2841796875, + "loss_alignment_w": 0.14208984375, + "loss_sft": 0.539304107427597, + "loss_total_v6": 0.6813634485006332, + "step": 256 + }, + { + "epoch": 0.3086160312218553, + "grad_norm": 15.409659385681152, + "learning_rate": 9.801260679333435e-06, + "loss": 10.77487564086914, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.5299818553030491, + "loss_total_v6": 0.6734297573566437, + "step": 257 + }, + { + "epoch": 0.3098168718102672, + "grad_norm": 14.533916473388672, + "learning_rate": 9.794274834223797e-06, + "loss": 10.813558578491211, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5306600704789162, + "loss_total_v6": 0.6758474484086037, + "step": 258 + }, + { + "epoch": 0.3110177123986791, + "grad_norm": 14.478206634521484, + "learning_rate": 9.787170898777571e-06, + "loss": 10.947996139526367, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.5382841974496841, + "loss_total_v6": 0.6842497736215591, + "step": 259 + }, + { + "epoch": 0.31221855298709095, + "grad_norm": 14.537369728088379, + "learning_rate": 9.779949047975568e-06, + "loss": 11.222685813903809, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5491962060332298, + "loss_total_v6": 0.7014178857207298, + "step": 260 + }, + { + "epoch": 0.31341939357550286, + "grad_norm": 14.813121795654297, + "learning_rate": 9.772609459703046e-06, + "loss": 10.985515594482422, + "loss_alignment": 0.304931640625, + "loss_alignment_w": 0.1524658203125, + "loss_sft": 0.5342509672045708, + "loss_total_v6": 0.6865947172045708, + "step": 261 + }, + { + "epoch": 0.3146202341639147, + "grad_norm": 14.503021240234375, + "learning_rate": 9.76515231474531e-06, + "loss": 11.009197235107422, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.535471647977829, + "loss_total_v6": 0.6880747824907303, + "step": 262 + }, + { + "epoch": 0.3158210747523266, + "grad_norm": 14.411008834838867, + "learning_rate": 9.757577796783268e-06, + "loss": 11.053728103637695, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5460978746414185, + "loss_total_v6": 0.6908579990267754, + "step": 263 + }, + { + "epoch": 0.31702191534073854, + "grad_norm": 15.193751335144043, + "learning_rate": 9.749886092388907e-06, + "loss": 10.904058456420898, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.5359806194901466, + "loss_total_v6": 0.6815036907792091, + "step": 264 + }, + { + "epoch": 0.3182227559291504, + "grad_norm": 14.537820816040039, + "learning_rate": 9.742077391020695e-06, + "loss": 11.140337944030762, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.5508243441581726, + "loss_total_v6": 0.696271114051342, + "step": 265 + }, + { + "epoch": 0.3194235965175623, + "grad_norm": 14.21373176574707, + "learning_rate": 9.73415188501891e-06, + "loss": 10.98354434967041, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5400634557008743, + "loss_total_v6": 0.6864715293049812, + "step": 266 + }, + { + "epoch": 0.32062443710597416, + "grad_norm": 14.199689865112305, + "learning_rate": 9.726109769600915e-06, + "loss": 11.089834213256836, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.5442346520721912, + "loss_total_v6": 0.6931146532297134, + "step": 267 + }, + { + "epoch": 0.32182527769438607, + "grad_norm": 14.778456687927246, + "learning_rate": 9.717951242856338e-06, + "loss": 10.795755386352539, + "loss_alignment": 0.2857666015625, + "loss_alignment_w": 0.14288330078125, + "loss_sft": 0.5316988229751587, + "loss_total_v6": 0.6747347563505173, + "step": 268 + }, + { + "epoch": 0.323026118282798, + "grad_norm": 14.3944091796875, + "learning_rate": 9.709676505742194e-06, + "loss": 11.253223419189453, + "loss_alignment": 0.30322265625, + "loss_alignment_w": 0.151611328125, + "loss_sft": 0.5517761558294296, + "loss_total_v6": 0.7033264711499214, + "step": 269 + }, + { + "epoch": 0.32422695887120984, + "grad_norm": 15.034976959228516, + "learning_rate": 9.701285762077938e-06, + "loss": 11.17374324798584, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.5504707768559456, + "loss_total_v6": 0.6983589679002762, + "step": 270 + }, + { + "epoch": 0.32542779945962175, + "grad_norm": 14.790785789489746, + "learning_rate": 9.692779218540449e-06, + "loss": 10.748385429382324, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5254575498402119, + "loss_total_v6": 0.6717740520834923, + "step": 271 + }, + { + "epoch": 0.3266286400480336, + "grad_norm": 14.070577621459961, + "learning_rate": 9.684157084658929e-06, + "loss": 10.637947082519531, + "loss_alignment": 0.291259765625, + "loss_alignment_w": 0.1456298828125, + "loss_sft": 0.5193486027419567, + "loss_total_v6": 0.6648716703057289, + "step": 272 + }, + { + "epoch": 0.3278294806364455, + "grad_norm": 13.964717864990234, + "learning_rate": 9.675419572809748e-06, + "loss": 11.052640914916992, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.5484408363699913, + "loss_total_v6": 0.6907900795340538, + "step": 273 + }, + { + "epoch": 0.3290303212248574, + "grad_norm": 15.3572416305542, + "learning_rate": 9.666566898211219e-06, + "loss": 10.881266593933105, + "loss_alignment": 0.30712890625, + "loss_alignment_w": 0.153564453125, + "loss_sft": 0.5266978107392788, + "loss_total_v6": 0.6800791621208191, + "step": 274 + }, + { + "epoch": 0.3302311618132693, + "grad_norm": 13.764782905578613, + "learning_rate": 9.657599278918278e-06, + "loss": 10.880480766296387, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5349189713597298, + "loss_total_v6": 0.6800300478935242, + "step": 275 + }, + { + "epoch": 0.3314320024016812, + "grad_norm": 15.093720436096191, + "learning_rate": 9.648516935817133e-06, + "loss": 10.65201187133789, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5194037146866322, + "loss_total_v6": 0.6657507568597794, + "step": 276 + }, + { + "epoch": 0.33263284299009305, + "grad_norm": 14.620620727539062, + "learning_rate": 9.639320092619814e-06, + "loss": 10.645565032958984, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.5223576836287975, + "loss_total_v6": 0.6653477773070335, + "step": 277 + }, + { + "epoch": 0.33383368357850496, + "grad_norm": 14.578938484191895, + "learning_rate": 9.630008975858667e-06, + "loss": 11.020940780639648, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.541820801794529, + "loss_total_v6": 0.6888087317347527, + "step": 278 + }, + { + "epoch": 0.33503452416691687, + "grad_norm": 13.876437187194824, + "learning_rate": 9.620583814880763e-06, + "loss": 10.820361137390137, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.5267363749444485, + "loss_total_v6": 0.6762725263834, + "step": 279 + }, + { + "epoch": 0.3362353647553287, + "grad_norm": 13.978527069091797, + "learning_rate": 9.611044841842264e-06, + "loss": 10.810380935668945, + "loss_alignment": 0.30517578125, + "loss_alignment_w": 0.152587890625, + "loss_sft": 0.5230914205312729, + "loss_total_v6": 0.6756487712264061, + "step": 280 + }, + { + "epoch": 0.33743620534374064, + "grad_norm": 14.472561836242676, + "learning_rate": 9.601392291702693e-06, + "loss": 11.157621383666992, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.5525149777531624, + "loss_total_v6": 0.6973513588309288, + "step": 281 + }, + { + "epoch": 0.3386370459321525, + "grad_norm": 14.25695514678955, + "learning_rate": 9.591626402219154e-06, + "loss": 11.047627449035645, + "loss_alignment": 0.2958984375, + "loss_alignment_w": 0.14794921875, + "loss_sft": 0.5425274819135666, + "loss_total_v6": 0.6904767155647278, + "step": 282 + }, + { + "epoch": 0.3398378865205644, + "grad_norm": 13.722317695617676, + "learning_rate": 9.581747413940472e-06, + "loss": 10.845264434814453, + "loss_alignment": 0.3017578125, + "loss_alignment_w": 0.15087890625, + "loss_sft": 0.5270721688866615, + "loss_total_v6": 0.6778290048241615, + "step": 283 + }, + { + "epoch": 0.34103872710897626, + "grad_norm": 14.289695739746094, + "learning_rate": 9.571755570201266e-06, + "loss": 10.931567192077637, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.5404312163591385, + "loss_total_v6": 0.6832229644060135, + "step": 284 + }, + { + "epoch": 0.34223956769738817, + "grad_norm": 13.533595085144043, + "learning_rate": 9.561651117115962e-06, + "loss": 10.877212524414062, + "loss_alignment": 0.298828125, + "loss_alignment_w": 0.1494140625, + "loss_sft": 0.5307474285364151, + "loss_total_v6": 0.6798258125782013, + "step": 285 + }, + { + "epoch": 0.3434404082858001, + "grad_norm": 14.50034236907959, + "learning_rate": 9.551434303572725e-06, + "loss": 10.984813690185547, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.5391967743635178, + "loss_total_v6": 0.6865509152412415, + "step": 286 + }, + { + "epoch": 0.34464124887421194, + "grad_norm": 14.141195297241211, + "learning_rate": 9.54110538122733e-06, + "loss": 10.841154098510742, + "loss_alignment": 0.302001953125, + "loss_alignment_w": 0.1510009765625, + "loss_sft": 0.5265711471438408, + "loss_total_v6": 0.6775721162557602, + "step": 287 + }, + { + "epoch": 0.34584208946262385, + "grad_norm": 14.851096153259277, + "learning_rate": 9.530664604496964e-06, + "loss": 10.83686637878418, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.5326966270804405, + "loss_total_v6": 0.6773041635751724, + "step": 288 + }, + { + "epoch": 0.3470429300510357, + "grad_norm": 13.86386489868164, + "learning_rate": 9.520112230553959e-06, + "loss": 11.164548873901367, + "loss_alignment": 0.301513671875, + "loss_alignment_w": 0.1507568359375, + "loss_sft": 0.547210618853569, + "loss_total_v6": 0.6977843195199966, + "step": 289 + }, + { + "epoch": 0.3482437706394476, + "grad_norm": 13.843390464782715, + "learning_rate": 9.509448519319455e-06, + "loss": 11.00540828704834, + "loss_alignment": 0.302734375, + "loss_alignment_w": 0.1513671875, + "loss_sft": 0.5362876951694489, + "loss_total_v6": 0.6878380179405212, + "step": 290 + }, + { + "epoch": 0.3494446112278595, + "grad_norm": 14.386309623718262, + "learning_rate": 9.498673733457007e-06, + "loss": 10.965662956237793, + "loss_alignment": 0.302734375, + "loss_alignment_w": 0.1513671875, + "loss_sft": 0.5343376956880093, + "loss_total_v6": 0.6853539422154427, + "step": 291 + }, + { + "epoch": 0.3506454518162714, + "grad_norm": 14.17036247253418, + "learning_rate": 9.487788138366098e-06, + "loss": 11.048591613769531, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5420384258031845, + "loss_total_v6": 0.6905369833111763, + "step": 292 + }, + { + "epoch": 0.3518462924046833, + "grad_norm": 14.850576400756836, + "learning_rate": 9.476792002175621e-06, + "loss": 10.892827987670898, + "loss_alignment": 0.309326171875, + "loss_alignment_w": 0.1546630859375, + "loss_sft": 0.5266117006540298, + "loss_total_v6": 0.6808017417788506, + "step": 293 + }, + { + "epoch": 0.35304713299309515, + "grad_norm": 14.12306022644043, + "learning_rate": 9.465685595737263e-06, + "loss": 10.786500930786133, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.5256882719695568, + "loss_total_v6": 0.6741563081741333, + "step": 294 + }, + { + "epoch": 0.35424797358150706, + "grad_norm": 13.780097961425781, + "learning_rate": 9.454469192618834e-06, + "loss": 10.939574241638184, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.5372237786650658, + "loss_total_v6": 0.6837234199047089, + "step": 295 + }, + { + "epoch": 0.35544881416991897, + "grad_norm": 14.44544792175293, + "learning_rate": 9.443143069097531e-06, + "loss": 10.764629364013672, + "loss_alignment": 0.309326171875, + "loss_alignment_w": 0.1546630859375, + "loss_sft": 0.5185229294002056, + "loss_total_v6": 0.6727892979979515, + "step": 296 + }, + { + "epoch": 0.3566496547583308, + "grad_norm": 14.544143676757812, + "learning_rate": 9.431707504153138e-06, + "loss": 10.915324211120605, + "loss_alignment": 0.283935546875, + "loss_alignment_w": 0.1419677734375, + "loss_sft": 0.5405146330595016, + "loss_total_v6": 0.682207740843296, + "step": 297 + }, + { + "epoch": 0.35785049534674274, + "grad_norm": 14.441455841064453, + "learning_rate": 9.420162779461142e-06, + "loss": 10.712688446044922, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.5193965174257755, + "loss_total_v6": 0.6695429980754852, + "step": 298 + }, + { + "epoch": 0.3590513359351546, + "grad_norm": 14.592631340026855, + "learning_rate": 9.408509179385806e-06, + "loss": 10.985031127929688, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.5430098176002502, + "loss_total_v6": 0.686564490199089, + "step": 299 + }, + { + "epoch": 0.3602521765235665, + "grad_norm": 14.523307800292969, + "learning_rate": 9.39674699097316e-06, + "loss": 10.87796688079834, + "loss_alignment": 0.29833984375, + "loss_alignment_w": 0.149169921875, + "loss_sft": 0.530962411314249, + "loss_total_v6": 0.6798729300498962, + "step": 300 + }, + { + "epoch": 0.3602521765235665, + "eval_loss": 0.620146632194519, + "eval_loss_alignment": 0.2869614821213942, + "eval_loss_alignment_w": 0.1434807410606971, + "eval_loss_sft": 0.4766658697671329, + "eval_loss_total_v6": 0.6201466107920098, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.3751, + "eval_samples_per_second": 24.623, + "eval_steps_per_second": 3.087, + "step": 300 + }, + { + "epoch": 0.36145301711197836, + "grad_norm": 14.663086891174316, + "learning_rate": 9.384876503943929e-06, + "loss": 10.877357482910156, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.5322976037859917, + "loss_total_v6": 0.6798348650336266, + "step": 301 + }, + { + "epoch": 0.36265385770039027, + "grad_norm": 13.666095733642578, + "learning_rate": 9.3728980106864e-06, + "loss": 10.84619140625, + "loss_alignment": 0.279296875, + "loss_alignment_w": 0.1396484375, + "loss_sft": 0.5380859524011612, + "loss_total_v6": 0.677886962890625, + "step": 302 + }, + { + "epoch": 0.3638546982888022, + "grad_norm": 14.652335166931152, + "learning_rate": 9.360811806249224e-06, + "loss": 10.8510160446167, + "loss_alignment": 0.287109375, + "loss_alignment_w": 0.1435546875, + "loss_sft": 0.5344049260020256, + "loss_total_v6": 0.6781885102391243, + "step": 303 + }, + { + "epoch": 0.36505553887721404, + "grad_norm": 14.252913475036621, + "learning_rate": 9.348618188334135e-06, + "loss": 10.856021881103516, + "loss_alignment": 0.2864990234375, + "loss_alignment_w": 0.14324951171875, + "loss_sft": 0.5351908057928085, + "loss_total_v6": 0.6785013303160667, + "step": 304 + }, + { + "epoch": 0.36625637946562595, + "grad_norm": 14.090014457702637, + "learning_rate": 9.33631745728863e-06, + "loss": 11.087255477905273, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.5469115674495697, + "loss_total_v6": 0.6929534301161766, + "step": 305 + }, + { + "epoch": 0.3674572200540378, + "grad_norm": 13.137598037719727, + "learning_rate": 9.323909916098566e-06, + "loss": 10.672443389892578, + "loss_alignment": 0.289794921875, + "loss_alignment_w": 0.1448974609375, + "loss_sft": 0.5224507041275501, + "loss_total_v6": 0.6670277491211891, + "step": 306 + }, + { + "epoch": 0.3686580606424497, + "grad_norm": 13.962239265441895, + "learning_rate": 9.311395870380699e-06, + "loss": 10.668598175048828, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.523125909268856, + "loss_total_v6": 0.6667874231934547, + "step": 307 + }, + { + "epoch": 0.3698589012308616, + "grad_norm": 13.38227367401123, + "learning_rate": 9.29877562837515e-06, + "loss": 10.737651824951172, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.5229403972625732, + "loss_total_v6": 0.6711032390594482, + "step": 308 + }, + { + "epoch": 0.3710597418192735, + "grad_norm": 15.197580337524414, + "learning_rate": 9.286049500937826e-06, + "loss": 10.825841903686523, + "loss_alignment": 0.302490234375, + "loss_alignment_w": 0.1512451171875, + "loss_sft": 0.5254767201840878, + "loss_total_v6": 0.6766150593757629, + "step": 309 + }, + { + "epoch": 0.3722605824076854, + "grad_norm": 13.932950973510742, + "learning_rate": 9.273217801532744e-06, + "loss": 10.928186416625977, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.5401283502578735, + "loss_total_v6": 0.6830116435885429, + "step": 310 + }, + { + "epoch": 0.37346142299609725, + "grad_norm": 13.325008392333984, + "learning_rate": 9.260280846224328e-06, + "loss": 10.799266815185547, + "loss_alignment": 0.298828125, + "loss_alignment_w": 0.1494140625, + "loss_sft": 0.5254790894687176, + "loss_total_v6": 0.6749541983008385, + "step": 311 + }, + { + "epoch": 0.37466226358450916, + "grad_norm": 13.923306465148926, + "learning_rate": 9.247238953669612e-06, + "loss": 11.007818222045898, + "loss_alignment": 0.3046875, + "loss_alignment_w": 0.15234375, + "loss_sft": 0.5357668995857239, + "loss_total_v6": 0.6879886016249657, + "step": 312 + }, + { + "epoch": 0.37586310417292107, + "grad_norm": 13.835144996643066, + "learning_rate": 9.234092445110401e-06, + "loss": 11.097851753234863, + "loss_alignment": 0.297119140625, + "loss_alignment_w": 0.1485595703125, + "loss_sft": 0.5449341163039207, + "loss_total_v6": 0.6936157643795013, + "step": 313 + }, + { + "epoch": 0.3770639447613329, + "grad_norm": 13.721220016479492, + "learning_rate": 9.220841644365343e-06, + "loss": 10.682448387145996, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5180100277066231, + "loss_total_v6": 0.6676529794931412, + "step": 314 + }, + { + "epoch": 0.37826478534974484, + "grad_norm": 14.15974235534668, + "learning_rate": 9.207486877821971e-06, + "loss": 11.014200210571289, + "loss_alignment": 0.30517578125, + "loss_alignment_w": 0.152587890625, + "loss_sft": 0.5351130738854408, + "loss_total_v6": 0.6883875653147697, + "step": 315 + }, + { + "epoch": 0.3794656259381567, + "grad_norm": 13.514354705810547, + "learning_rate": 9.194028474428651e-06, + "loss": 10.564706802368164, + "loss_alignment": 0.279052734375, + "loss_alignment_w": 0.1395263671875, + "loss_sft": 0.5207220762968063, + "loss_total_v6": 0.660294197499752, + "step": 316 + }, + { + "epoch": 0.3806664665265686, + "grad_norm": 13.864730834960938, + "learning_rate": 9.180466765686485e-06, + "loss": 10.693239212036133, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.5202867202460766, + "loss_total_v6": 0.6683274805545807, + "step": 317 + }, + { + "epoch": 0.38186730711498046, + "grad_norm": 14.48923397064209, + "learning_rate": 9.16680208564114e-06, + "loss": 10.414825439453125, + "loss_alignment": 0.279541015625, + "loss_alignment_w": 0.1397705078125, + "loss_sft": 0.5108203329145908, + "loss_total_v6": 0.6509265303611755, + "step": 318 + }, + { + "epoch": 0.38306814770339237, + "grad_norm": 13.775564193725586, + "learning_rate": 9.15303477087463e-06, + "loss": 10.753029823303223, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5271515995264053, + "loss_total_v6": 0.6720643192529678, + "step": 319 + }, + { + "epoch": 0.3842689882918043, + "grad_norm": 13.82923412322998, + "learning_rate": 9.139165160497017e-06, + "loss": 10.819082260131836, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.5296930260956287, + "loss_total_v6": 0.6761926710605621, + "step": 320 + }, + { + "epoch": 0.38546982888021614, + "grad_norm": 15.232104301452637, + "learning_rate": 9.125193596138057e-06, + "loss": 10.542519569396973, + "loss_alignment": 0.299560546875, + "loss_alignment_w": 0.1497802734375, + "loss_sft": 0.5093561075627804, + "loss_total_v6": 0.6589074805378914, + "step": 321 + }, + { + "epoch": 0.38667066946862805, + "grad_norm": 13.529447555541992, + "learning_rate": 9.111120421938796e-06, + "loss": 10.71749210357666, + "loss_alignment": 0.28857421875, + "loss_alignment_w": 0.144287109375, + "loss_sft": 0.5256934687495232, + "loss_total_v6": 0.6698432564735413, + "step": 322 + }, + { + "epoch": 0.3878715100570399, + "grad_norm": 13.86987018585205, + "learning_rate": 9.096945984543082e-06, + "loss": 10.898045539855957, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.5345366671681404, + "loss_total_v6": 0.6811278387904167, + "step": 323 + }, + { + "epoch": 0.3890723506454518, + "grad_norm": 13.67231273651123, + "learning_rate": 9.082670633089028e-06, + "loss": 10.886521339416504, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.5349150151014328, + "loss_total_v6": 0.6804075688123703, + "step": 324 + }, + { + "epoch": 0.3902731912338637, + "grad_norm": 14.0606050491333, + "learning_rate": 9.068294719200422e-06, + "loss": 10.842710494995117, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5305898487567902, + "loss_total_v6": 0.6776693388819695, + "step": 325 + }, + { + "epoch": 0.3914740318222756, + "grad_norm": 13.901698112487793, + "learning_rate": 9.053818596978051e-06, + "loss": 10.787849426269531, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.5325169749557972, + "loss_total_v6": 0.6742406263947487, + "step": 326 + }, + { + "epoch": 0.3926748724106875, + "grad_norm": 14.12863540649414, + "learning_rate": 9.039242622990991e-06, + "loss": 10.811777114868164, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.5236364379525185, + "loss_total_v6": 0.6757360696792603, + "step": 327 + }, + { + "epoch": 0.39387571299909935, + "grad_norm": 13.484013557434082, + "learning_rate": 9.02456715626782e-06, + "loss": 10.939228057861328, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5378734618425369, + "loss_total_v6": 0.68370171636343, + "step": 328 + }, + { + "epoch": 0.39507655358751126, + "grad_norm": 13.934995651245117, + "learning_rate": 9.009792558287777e-06, + "loss": 10.90304946899414, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5298140682280064, + "loss_total_v6": 0.6814406365156174, + "step": 329 + }, + { + "epoch": 0.39627739417592317, + "grad_norm": 13.046490669250488, + "learning_rate": 8.994919192971849e-06, + "loss": 10.615158081054688, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5187787674367428, + "loss_total_v6": 0.6634473651647568, + "step": 330 + }, + { + "epoch": 0.397478234764335, + "grad_norm": 13.36274242401123, + "learning_rate": 8.97994742667382e-06, + "loss": 10.7949857711792, + "loss_alignment": 0.3076171875, + "loss_alignment_w": 0.15380859375, + "loss_sft": 0.5214120969176292, + "loss_total_v6": 0.6746866032481194, + "step": 331 + }, + { + "epoch": 0.39867907535274694, + "grad_norm": 13.903371810913086, + "learning_rate": 8.96487762817124e-06, + "loss": 10.702251434326172, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.5271976143121719, + "loss_total_v6": 0.6688907071948051, + "step": 332 + }, + { + "epoch": 0.3998799159411588, + "grad_norm": 14.09846019744873, + "learning_rate": 8.949710168656338e-06, + "loss": 10.804533004760742, + "loss_alignment": 0.283935546875, + "loss_alignment_w": 0.1419677734375, + "loss_sft": 0.5335749164223671, + "loss_total_v6": 0.675283282995224, + "step": 333 + }, + { + "epoch": 0.4010807565295707, + "grad_norm": 12.898571968078613, + "learning_rate": 8.934445421726888e-06, + "loss": 10.791620254516602, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.5257030539214611, + "loss_total_v6": 0.6744762659072876, + "step": 334 + }, + { + "epoch": 0.4022815971179826, + "grad_norm": 13.447714805603027, + "learning_rate": 8.919083763377001e-06, + "loss": 10.954362869262695, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.5369272977113724, + "loss_total_v6": 0.6846476420760155, + "step": 335 + }, + { + "epoch": 0.40348243770639447, + "grad_norm": 13.755511283874512, + "learning_rate": 8.903625571987863e-06, + "loss": 10.732601165771484, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5240133032202721, + "loss_total_v6": 0.6707875728607178, + "step": 336 + }, + { + "epoch": 0.4046832782948064, + "grad_norm": 13.791305541992188, + "learning_rate": 8.888071228318422e-06, + "loss": 11.049768447875977, + "loss_alignment": 0.30615234375, + "loss_alignment_w": 0.153076171875, + "loss_sft": 0.5373818092048168, + "loss_total_v6": 0.6906105801463127, + "step": 337 + }, + { + "epoch": 0.40588411888321824, + "grad_norm": 13.338412284851074, + "learning_rate": 8.872421115495996e-06, + "loss": 10.754451751708984, + "loss_alignment": 0.301025390625, + "loss_alignment_w": 0.1505126953125, + "loss_sft": 0.5213811621069908, + "loss_total_v6": 0.6721532568335533, + "step": 338 + }, + { + "epoch": 0.40708495947163015, + "grad_norm": 14.231742858886719, + "learning_rate": 8.85667561900685e-06, + "loss": 11.007278442382812, + "loss_alignment": 0.29541015625, + "loss_alignment_w": 0.147705078125, + "loss_sft": 0.5403566211462021, + "loss_total_v6": 0.6879549100995064, + "step": 339 + }, + { + "epoch": 0.408285800060042, + "grad_norm": 13.718606948852539, + "learning_rate": 8.840835126686694e-06, + "loss": 11.012235641479492, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.5391253419220448, + "loss_total_v6": 0.6882647573947906, + "step": 340 + }, + { + "epoch": 0.4094866406484539, + "grad_norm": 13.968916893005371, + "learning_rate": 8.824900028711128e-06, + "loss": 10.997642517089844, + "loss_alignment": 0.296630859375, + "loss_alignment_w": 0.1483154296875, + "loss_sft": 0.5391746088862419, + "loss_total_v6": 0.6873527020215988, + "step": 341 + }, + { + "epoch": 0.4106874812368658, + "grad_norm": 14.205462455749512, + "learning_rate": 8.80887071758604e-06, + "loss": 10.944540023803711, + "loss_alignment": 0.3017578125, + "loss_alignment_w": 0.15087890625, + "loss_sft": 0.5333531871438026, + "loss_total_v6": 0.6840337663888931, + "step": 342 + }, + { + "epoch": 0.4118883218252777, + "grad_norm": 13.901556968688965, + "learning_rate": 8.792747588137925e-06, + "loss": 10.603385925292969, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.5165323838591576, + "loss_total_v6": 0.6627115830779076, + "step": 343 + }, + { + "epoch": 0.4130891624136896, + "grad_norm": 12.84968376159668, + "learning_rate": 8.77653103750417e-06, + "loss": 10.809629440307617, + "loss_alignment": 0.284912109375, + "loss_alignment_w": 0.1424560546875, + "loss_sft": 0.5336036011576653, + "loss_total_v6": 0.6756018698215485, + "step": 344 + }, + { + "epoch": 0.41429000300210145, + "grad_norm": 13.278827667236328, + "learning_rate": 8.760221465123273e-06, + "loss": 10.667316436767578, + "loss_alignment": 0.303955078125, + "loss_alignment_w": 0.1519775390625, + "loss_sft": 0.5149127878248692, + "loss_total_v6": 0.66670723259449, + "step": 345 + }, + { + "epoch": 0.41549084359051336, + "grad_norm": 13.545820236206055, + "learning_rate": 8.743819272724996e-06, + "loss": 10.636959075927734, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.5130765400826931, + "loss_total_v6": 0.664809912443161, + "step": 346 + }, + { + "epoch": 0.41669168417892527, + "grad_norm": 13.951521873474121, + "learning_rate": 8.727324864320472e-06, + "loss": 10.913329124450684, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.5300750210881233, + "loss_total_v6": 0.6820830777287483, + "step": 347 + }, + { + "epoch": 0.4178925247673371, + "grad_norm": 13.577295303344727, + "learning_rate": 8.710738646192262e-06, + "loss": 10.803337097167969, + "loss_alignment": 0.30224609375, + "loss_alignment_w": 0.151123046875, + "loss_sft": 0.5243296585977077, + "loss_total_v6": 0.6752085760235786, + "step": 348 + }, + { + "epoch": 0.41909336535574904, + "grad_norm": 13.746606826782227, + "learning_rate": 8.694061026884336e-06, + "loss": 10.947652816772461, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.5372098535299301, + "loss_total_v6": 0.6842282861471176, + "step": 349 + }, + { + "epoch": 0.4202942059441609, + "grad_norm": 15.260991096496582, + "learning_rate": 8.677292417192018e-06, + "loss": 10.49916934967041, + "loss_alignment": 0.280517578125, + "loss_alignment_w": 0.1402587890625, + "loss_sft": 0.5155425406992435, + "loss_total_v6": 0.65619807690382, + "step": 350 + }, + { + "epoch": 0.4214950465325728, + "grad_norm": 13.864262580871582, + "learning_rate": 8.66043323015186e-06, + "loss": 10.797057151794434, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.5298117771744728, + "loss_total_v6": 0.6748160570859909, + "step": 351 + }, + { + "epoch": 0.4226958871209847, + "grad_norm": 13.335013389587402, + "learning_rate": 8.643483881031479e-06, + "loss": 10.814506530761719, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5294069573283195, + "loss_total_v6": 0.6759066134691238, + "step": 352 + }, + { + "epoch": 0.42389672770939657, + "grad_norm": 14.836807250976562, + "learning_rate": 8.62644478731932e-06, + "loss": 10.593568801879883, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.5194283500313759, + "loss_total_v6": 0.6620980575680733, + "step": 353 + }, + { + "epoch": 0.4250975682978085, + "grad_norm": 13.76627254486084, + "learning_rate": 8.609316368714371e-06, + "loss": 10.644466400146484, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.5184437930583954, + "loss_total_v6": 0.6652791276574135, + "step": 354 + }, + { + "epoch": 0.42629840888622034, + "grad_norm": 58.31990432739258, + "learning_rate": 8.592099047115834e-06, + "loss": 10.607135772705078, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.5205203890800476, + "loss_total_v6": 0.6629459336400032, + "step": 355 + }, + { + "epoch": 0.42749924947463225, + "grad_norm": 14.897151947021484, + "learning_rate": 8.574793246612727e-06, + "loss": 10.804222106933594, + "loss_alignment": 0.291748046875, + "loss_alignment_w": 0.1458740234375, + "loss_sft": 0.5296492874622345, + "loss_total_v6": 0.6752638965845108, + "step": 356 + }, + { + "epoch": 0.4287000900630441, + "grad_norm": 14.6962308883667, + "learning_rate": 8.557399393473437e-06, + "loss": 10.587580680847168, + "loss_alignment": 0.301025390625, + "loss_alignment_w": 0.1505126953125, + "loss_sft": 0.5114704929292202, + "loss_total_v6": 0.661723792552948, + "step": 357 + }, + { + "epoch": 0.429900930651456, + "grad_norm": 13.577737808227539, + "learning_rate": 8.539917916135227e-06, + "loss": 10.59320068359375, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.5158196166157722, + "loss_total_v6": 0.6620750948786736, + "step": 358 + }, + { + "epoch": 0.4311017712398679, + "grad_norm": 13.996513366699219, + "learning_rate": 8.52234924519367e-06, + "loss": 10.58682632446289, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.5156195275485516, + "loss_total_v6": 0.6616766527295113, + "step": 359 + }, + { + "epoch": 0.4323026118282798, + "grad_norm": 13.9138765335083, + "learning_rate": 8.504693813392059e-06, + "loss": 10.779126167297363, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.526524368673563, + "loss_total_v6": 0.673695370554924, + "step": 360 + }, + { + "epoch": 0.4335034524166917, + "grad_norm": 14.310792922973633, + "learning_rate": 8.486952055610736e-06, + "loss": 10.754645347595215, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.5257267281413078, + "loss_total_v6": 0.6721653342247009, + "step": 361 + }, + { + "epoch": 0.43470429300510355, + "grad_norm": 16.376554489135742, + "learning_rate": 8.469124408856384e-06, + "loss": 10.788247108459473, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.5269418433308601, + "loss_total_v6": 0.674265444278717, + "step": 362 + }, + { + "epoch": 0.43590513359351546, + "grad_norm": 13.713703155517578, + "learning_rate": 8.451211312251266e-06, + "loss": 10.53208065032959, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.5116333216428757, + "loss_total_v6": 0.6582550406455994, + "step": 363 + }, + { + "epoch": 0.43710597418192737, + "grad_norm": 25.21196174621582, + "learning_rate": 8.433213207022404e-06, + "loss": 10.610979080200195, + "loss_alignment": 0.300537109375, + "loss_alignment_w": 0.1502685546875, + "loss_sft": 0.512917622923851, + "loss_total_v6": 0.6631861925125122, + "step": 364 + }, + { + "epoch": 0.4383068147703392, + "grad_norm": 15.220426559448242, + "learning_rate": 8.415130536490712e-06, + "loss": 10.533714294433594, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5136580690741539, + "loss_total_v6": 0.658357135951519, + "step": 365 + }, + { + "epoch": 0.43950765535875114, + "grad_norm": 13.656045913696289, + "learning_rate": 8.396963746060076e-06, + "loss": 10.686354637145996, + "loss_alignment": 0.297119140625, + "loss_alignment_w": 0.1485595703125, + "loss_sft": 0.5190019309520721, + "loss_total_v6": 0.6678971871733665, + "step": 366 + }, + { + "epoch": 0.440708495947163, + "grad_norm": 13.74402904510498, + "learning_rate": 8.37871328320639e-06, + "loss": 10.998809814453125, + "loss_alignment": 0.2978515625, + "loss_alignment_w": 0.14892578125, + "loss_sft": 0.5382862016558647, + "loss_total_v6": 0.6874256134033203, + "step": 367 + }, + { + "epoch": 0.4419093365355749, + "grad_norm": 14.51016902923584, + "learning_rate": 8.360379597466519e-06, + "loss": 10.411026954650879, + "loss_alignment": 0.2952880859375, + "loss_alignment_w": 0.14764404296875, + "loss_sft": 0.5031977333128452, + "loss_total_v6": 0.6506891697645187, + "step": 368 + }, + { + "epoch": 0.4431101771239868, + "grad_norm": 26.49835777282715, + "learning_rate": 8.341963140427242e-06, + "loss": 10.52084732055664, + "loss_alignment": 0.2802734375, + "loss_alignment_w": 0.14013671875, + "loss_sft": 0.5175383128225803, + "loss_total_v6": 0.6575529798865318, + "step": 369 + }, + { + "epoch": 0.44431101771239867, + "grad_norm": 14.554941177368164, + "learning_rate": 8.32346436571412e-06, + "loss": 10.535572052001953, + "loss_alignment": 0.2880859375, + "loss_alignment_w": 0.14404296875, + "loss_sft": 0.5146134346723557, + "loss_total_v6": 0.6584732756018639, + "step": 370 + }, + { + "epoch": 0.4455118583008106, + "grad_norm": 13.557249069213867, + "learning_rate": 8.304883728980325e-06, + "loss": 10.554488182067871, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.5118893831968307, + "loss_total_v6": 0.6596554964780807, + "step": 371 + }, + { + "epoch": 0.44671269888922244, + "grad_norm": 14.875314712524414, + "learning_rate": 8.286221687895412e-06, + "loss": 10.517334938049316, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5111237168312073, + "loss_total_v6": 0.6573334485292435, + "step": 372 + }, + { + "epoch": 0.44791353947763435, + "grad_norm": 14.369518280029297, + "learning_rate": 8.26747870213406e-06, + "loss": 10.946680068969727, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5369964651763439, + "loss_total_v6": 0.6841675117611885, + "step": 373 + }, + { + "epoch": 0.44911438006604626, + "grad_norm": 14.285539627075195, + "learning_rate": 8.248655233364724e-06, + "loss": 10.698344230651855, + "loss_alignment": 0.300048828125, + "loss_alignment_w": 0.1500244140625, + "loss_sft": 0.5185610912740231, + "loss_total_v6": 0.6686464995145798, + "step": 374 + }, + { + "epoch": 0.4503152206544581, + "grad_norm": 14.230682373046875, + "learning_rate": 8.229751745238294e-06, + "loss": 10.697490692138672, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.5281359776854515, + "loss_total_v6": 0.6685931384563446, + "step": 375 + }, + { + "epoch": 0.45151606124287, + "grad_norm": 13.569973945617676, + "learning_rate": 8.210768703376651e-06, + "loss": 10.714760780334473, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.5292154066264629, + "loss_total_v6": 0.6696725487709045, + "step": 376 + }, + { + "epoch": 0.4527169018312819, + "grad_norm": 14.432129859924316, + "learning_rate": 8.191706575361208e-06, + "loss": 10.802877426147461, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.5318692326545715, + "loss_total_v6": 0.6751797944307327, + "step": 377 + }, + { + "epoch": 0.4539177424196938, + "grad_norm": 14.713804244995117, + "learning_rate": 8.172565830721391e-06, + "loss": 10.6187105178833, + "loss_alignment": 0.2847900390625, + "loss_alignment_w": 0.14239501953125, + "loss_sft": 0.5216711238026619, + "loss_total_v6": 0.6636693924665451, + "step": 378 + }, + { + "epoch": 0.45511858300810565, + "grad_norm": 13.646299362182617, + "learning_rate": 8.153346940923076e-06, + "loss": 10.801868438720703, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.5275643020868301, + "loss_total_v6": 0.6751167923212051, + "step": 379 + }, + { + "epoch": 0.45631942359651756, + "grad_norm": 13.814051628112793, + "learning_rate": 8.13405037935697e-06, + "loss": 10.667513847351074, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.5239583738148212, + "loss_total_v6": 0.6667196154594421, + "step": 380 + }, + { + "epoch": 0.45752026418492947, + "grad_norm": 13.454007148742676, + "learning_rate": 8.114676621326961e-06, + "loss": 10.527471542358398, + "loss_alignment": 0.297119140625, + "loss_alignment_w": 0.1485595703125, + "loss_sft": 0.5095905177295208, + "loss_total_v6": 0.6579669862985611, + "step": 381 + }, + { + "epoch": 0.4587211047733413, + "grad_norm": 14.487754821777344, + "learning_rate": 8.095226144038398e-06, + "loss": 10.55661392211914, + "loss_alignment": 0.279541015625, + "loss_alignment_w": 0.1397705078125, + "loss_sft": 0.5201247110962868, + "loss_total_v6": 0.6597883924841881, + "step": 382 + }, + { + "epoch": 0.45992194536175324, + "grad_norm": 13.340232849121094, + "learning_rate": 8.075699426586345e-06, + "loss": 10.487873077392578, + "loss_alignment": 0.284423828125, + "loss_alignment_w": 0.1422119140625, + "loss_sft": 0.5133564211428165, + "loss_total_v6": 0.6554920524358749, + "step": 383 + }, + { + "epoch": 0.4611227859501651, + "grad_norm": 13.999850273132324, + "learning_rate": 8.056096949943777e-06, + "loss": 10.394500732421875, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.506773017346859, + "loss_total_v6": 0.6496563330292702, + "step": 384 + }, + { + "epoch": 0.462323626538577, + "grad_norm": 13.97973918914795, + "learning_rate": 8.036419196949738e-06, + "loss": 10.698076248168945, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5237322673201561, + "loss_total_v6": 0.6686297282576561, + "step": 385 + }, + { + "epoch": 0.4635244671269889, + "grad_norm": 13.171207427978516, + "learning_rate": 8.016666652297443e-06, + "loss": 10.844980239868164, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5316015854477882, + "loss_total_v6": 0.6778113096952438, + "step": 386 + }, + { + "epoch": 0.46472530771540077, + "grad_norm": 13.80904483795166, + "learning_rate": 7.996839802522331e-06, + "loss": 10.342613220214844, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.49875408411026, + "loss_total_v6": 0.6464133709669113, + "step": 387 + }, + { + "epoch": 0.4659261483038127, + "grad_norm": 13.482093811035156, + "learning_rate": 7.976939135990106e-06, + "loss": 10.684560775756836, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.5233758687973022, + "loss_total_v6": 0.6677850484848022, + "step": 388 + }, + { + "epoch": 0.46712698889222454, + "grad_norm": 14.472938537597656, + "learning_rate": 7.956965142884678e-06, + "loss": 10.216931343078613, + "loss_alignment": 0.275146484375, + "loss_alignment_w": 0.1375732421875, + "loss_sft": 0.5008781850337982, + "loss_total_v6": 0.6385582163929939, + "step": 389 + }, + { + "epoch": 0.46832782948063645, + "grad_norm": 13.590041160583496, + "learning_rate": 7.936918315196109e-06, + "loss": 10.390765190124512, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5028926394879818, + "loss_total_v6": 0.6494228020310402, + "step": 390 + }, + { + "epoch": 0.46952867006904836, + "grad_norm": 14.239372253417969, + "learning_rate": 7.916799146708489e-06, + "loss": 10.543566703796387, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.5152351558208466, + "loss_total_v6": 0.6589729264378548, + "step": 391 + }, + { + "epoch": 0.4707295106574602, + "grad_norm": 14.650139808654785, + "learning_rate": 7.89660813298777e-06, + "loss": 10.458236694335938, + "loss_alignment": 0.27099609375, + "loss_alignment_w": 0.135498046875, + "loss_sft": 0.5182028040289879, + "loss_total_v6": 0.6536398157477379, + "step": 392 + }, + { + "epoch": 0.4719303512458721, + "grad_norm": 14.161212921142578, + "learning_rate": 7.876345771369565e-06, + "loss": 10.555498123168945, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5137225314974785, + "loss_total_v6": 0.6597186028957367, + "step": 393 + }, + { + "epoch": 0.473131191834284, + "grad_norm": 14.0673828125, + "learning_rate": 7.85601256094689e-06, + "loss": 10.622039794921875, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.5194225572049618, + "loss_total_v6": 0.6638775169849396, + "step": 394 + }, + { + "epoch": 0.4743320324226959, + "grad_norm": 14.83708667755127, + "learning_rate": 7.835609002557883e-06, + "loss": 10.581676483154297, + "loss_alignment": 0.296630859375, + "loss_alignment_w": 0.1483154296875, + "loss_sft": 0.5129782892763615, + "loss_total_v6": 0.6613547578454018, + "step": 395 + }, + { + "epoch": 0.47553287301110775, + "grad_norm": 13.592988014221191, + "learning_rate": 7.815135598773455e-06, + "loss": 10.544696807861328, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5143749751150608, + "loss_total_v6": 0.6590435802936554, + "step": 396 + }, + { + "epoch": 0.47673371359951966, + "grad_norm": 13.725505828857422, + "learning_rate": 7.79459285388492e-06, + "loss": 10.567272186279297, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.5150840356945992, + "loss_total_v6": 0.6604545190930367, + "step": 397 + }, + { + "epoch": 0.47793455418793157, + "grad_norm": 13.753689765930176, + "learning_rate": 7.773981273891563e-06, + "loss": 10.57811164855957, + "loss_alignment": 0.284912109375, + "loss_alignment_w": 0.1424560546875, + "loss_sft": 0.5190421342849731, + "loss_total_v6": 0.6611319854855537, + "step": 398 + }, + { + "epoch": 0.4791353947763434, + "grad_norm": 15.098764419555664, + "learning_rate": 7.753301366488187e-06, + "loss": 10.700029373168945, + "loss_alignment": 0.2938232421875, + "loss_alignment_w": 0.14691162109375, + "loss_sft": 0.5219317264854908, + "loss_total_v6": 0.6687518209218979, + "step": 399 + }, + { + "epoch": 0.48033623536475534, + "grad_norm": 14.025445938110352, + "learning_rate": 7.73255364105261e-06, + "loss": 10.522448539733887, + "loss_alignment": 0.28857421875, + "loss_alignment_w": 0.144287109375, + "loss_sft": 0.5131217651069164, + "loss_total_v6": 0.6576530113816261, + "step": 400 + }, + { + "epoch": 0.48033623536475534, + "eval_loss": 0.6104982495307922, + "eval_loss_alignment": 0.28488393930288464, + "eval_loss_alignment_w": 0.14244196965144232, + "eval_loss_sft": 0.4680562876075363, + "eval_loss_total_v6": 0.6104982568380924, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.6414, + "eval_samples_per_second": 24.526, + "eval_steps_per_second": 3.075, + "step": 400 + }, + { + "epoch": 0.4815370759531672, + "grad_norm": 13.4631986618042, + "learning_rate": 7.7117386086331e-06, + "loss": 10.733118057250977, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.529111459851265, + "loss_total_v6": 0.6708198562264442, + "step": 401 + }, + { + "epoch": 0.4827379165415791, + "grad_norm": 20.81960105895996, + "learning_rate": 7.690856781935808e-06, + "loss": 10.525272369384766, + "loss_alignment": 0.2945556640625, + "loss_alignment_w": 0.14727783203125, + "loss_sft": 0.51010911911726, + "loss_total_v6": 0.6578294634819031, + "step": 402 + }, + { + "epoch": 0.483938757129991, + "grad_norm": 13.15988540649414, + "learning_rate": 7.66990867531213e-06, + "loss": 10.586530685424805, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.5190343335270882, + "loss_total_v6": 0.6616582125425339, + "step": 403 + }, + { + "epoch": 0.48513959771840287, + "grad_norm": 13.504483222961426, + "learning_rate": 7.648894804746031e-06, + "loss": 10.700987815856934, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5242499522864819, + "loss_total_v6": 0.6688117310404778, + "step": 404 + }, + { + "epoch": 0.4863404383068148, + "grad_norm": 13.883885383605957, + "learning_rate": 7.627815687841351e-06, + "loss": 10.677778244018555, + "loss_alignment": 0.2880859375, + "loss_alignment_w": 0.14404296875, + "loss_sft": 0.5232113711535931, + "loss_total_v6": 0.6673611551523209, + "step": 405 + }, + { + "epoch": 0.48754127889522664, + "grad_norm": 14.975886344909668, + "learning_rate": 7.606671843809045e-06, + "loss": 10.193424224853516, + "loss_alignment": 0.2802734375, + "loss_alignment_w": 0.14013671875, + "loss_sft": 0.49658599123358727, + "loss_total_v6": 0.6370889469981194, + "step": 406 + }, + { + "epoch": 0.48874211948363855, + "grad_norm": 14.56751537322998, + "learning_rate": 7.585463793454393e-06, + "loss": 10.717411994934082, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.5263141021132469, + "loss_total_v6": 0.6698382571339607, + "step": 407 + }, + { + "epoch": 0.48994296007205046, + "grad_norm": 13.794228553771973, + "learning_rate": 7.5641920591641775e-06, + "loss": 10.44034194946289, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.5087683536112309, + "loss_total_v6": 0.6525213941931725, + "step": 408 + }, + { + "epoch": 0.4911438006604623, + "grad_norm": 12.96858024597168, + "learning_rate": 7.542857164893816e-06, + "loss": 10.454488754272461, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5095914527773857, + "loss_total_v6": 0.6534055471420288, + "step": 409 + }, + { + "epoch": 0.4923446412488742, + "grad_norm": 13.487493515014648, + "learning_rate": 7.521459636154448e-06, + "loss": 10.563074111938477, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.5174767039716244, + "loss_total_v6": 0.6601921319961548, + "step": 410 + }, + { + "epoch": 0.4935454818372861, + "grad_norm": 13.743353843688965, + "learning_rate": 7.500000000000001e-06, + "loss": 10.285221099853516, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.493534330278635, + "loss_total_v6": 0.6428263261914253, + "step": 411 + }, + { + "epoch": 0.494746322425698, + "grad_norm": 13.420330047607422, + "learning_rate": 7.478478785014201e-06, + "loss": 10.680971145629883, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.5271493159234524, + "loss_total_v6": 0.6675606891512871, + "step": 412 + }, + { + "epoch": 0.4959471630141099, + "grad_norm": 14.884543418884277, + "learning_rate": 7.456896521297554e-06, + "loss": 10.595900535583496, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.5148133747279644, + "loss_total_v6": 0.6622437909245491, + "step": 413 + }, + { + "epoch": 0.49714800360252176, + "grad_norm": 14.395074844360352, + "learning_rate": 7.4352537404542935e-06, + "loss": 10.405900955200195, + "loss_alignment": 0.283935546875, + "loss_alignment_w": 0.1419677734375, + "loss_sft": 0.5084010548889637, + "loss_total_v6": 0.650368832051754, + "step": 414 + }, + { + "epoch": 0.49834884419093367, + "grad_norm": 13.983481407165527, + "learning_rate": 7.4135509755792776e-06, + "loss": 10.671184539794922, + "loss_alignment": 0.2841796875, + "loss_alignment_w": 0.14208984375, + "loss_sft": 0.5248896703124046, + "loss_total_v6": 0.6669489964842796, + "step": 415 + }, + { + "epoch": 0.4995496847793455, + "grad_norm": 21.229415893554688, + "learning_rate": 7.3917887612448665e-06, + "loss": 10.546099662780762, + "loss_alignment": 0.294677734375, + "loss_alignment_w": 0.1473388671875, + "loss_sft": 0.5117618665099144, + "loss_total_v6": 0.6591312512755394, + "step": 416 + }, + { + "epoch": 0.5007505253677574, + "grad_norm": 13.815094947814941, + "learning_rate": 7.369967633487752e-06, + "loss": 10.681581497192383, + "loss_alignment": 0.28271484375, + "loss_alignment_w": 0.141357421875, + "loss_sft": 0.5260583162307739, + "loss_total_v6": 0.6675988137722015, + "step": 417 + }, + { + "epoch": 0.5019513659561693, + "grad_norm": 14.489091873168945, + "learning_rate": 7.348088129795751e-06, + "loss": 10.168842315673828, + "loss_alignment": 0.2789306640625, + "loss_alignment_w": 0.13946533203125, + "loss_sft": 0.49607207998633385, + "loss_total_v6": 0.6355526596307755, + "step": 418 + }, + { + "epoch": 0.5031522065445813, + "grad_norm": 12.756722450256348, + "learning_rate": 7.326150789094571e-06, + "loss": 10.626371383666992, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.517709594219923, + "loss_total_v6": 0.664148211479187, + "step": 419 + }, + { + "epoch": 0.5043530471329931, + "grad_norm": 14.389547348022461, + "learning_rate": 7.304156151734534e-06, + "loss": 10.774971961975098, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.5275769829750061, + "loss_total_v6": 0.6734357327222824, + "step": 420 + }, + { + "epoch": 0.505553887721405, + "grad_norm": 12.988795280456543, + "learning_rate": 7.282104759477264e-06, + "loss": 10.376076698303223, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.5041108913719654, + "loss_total_v6": 0.6485048085451126, + "step": 421 + }, + { + "epoch": 0.5067547283098168, + "grad_norm": 13.901676177978516, + "learning_rate": 7.259997155482349e-06, + "loss": 10.45598316192627, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5087388157844543, + "loss_total_v6": 0.653498962521553, + "step": 422 + }, + { + "epoch": 0.5079555688982288, + "grad_norm": 13.199481964111328, + "learning_rate": 7.237833884293955e-06, + "loss": 10.657429695129395, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5198338739573956, + "loss_total_v6": 0.666089378297329, + "step": 423 + }, + { + "epoch": 0.5091564094866406, + "grad_norm": 13.131726264953613, + "learning_rate": 7.2156154918274194e-06, + "loss": 10.505651473999023, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5121025070548058, + "loss_total_v6": 0.6566032096743584, + "step": 424 + }, + { + "epoch": 0.5103572500750525, + "grad_norm": 14.93184757232666, + "learning_rate": 7.193342525355796e-06, + "loss": 10.4666109085083, + "loss_alignment": 0.303466796875, + "loss_alignment_w": 0.1517333984375, + "loss_sft": 0.5024450309574604, + "loss_total_v6": 0.654163159430027, + "step": 425 + }, + { + "epoch": 0.5115580906634645, + "grad_norm": 12.698034286499023, + "learning_rate": 7.1710155334963875e-06, + "loss": 10.542771339416504, + "loss_alignment": 0.294677734375, + "loss_alignment_w": 0.1473388671875, + "loss_sft": 0.5114012695848942, + "loss_total_v6": 0.6589232459664345, + "step": 426 + }, + { + "epoch": 0.5127589312518763, + "grad_norm": 12.525036811828613, + "learning_rate": 7.148635066197216e-06, + "loss": 10.440214157104492, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5075395926833153, + "loss_total_v6": 0.6525133773684502, + "step": 427 + }, + { + "epoch": 0.5139597718402882, + "grad_norm": 14.551371574401855, + "learning_rate": 7.126201674723493e-06, + "loss": 10.729175567626953, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5249435529112816, + "loss_total_v6": 0.670573465526104, + "step": 428 + }, + { + "epoch": 0.5151606124287, + "grad_norm": 14.45254898071289, + "learning_rate": 7.103715911644029e-06, + "loss": 10.323490142822266, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.5021975003182888, + "loss_total_v6": 0.6452181413769722, + "step": 429 + }, + { + "epoch": 0.516361453017112, + "grad_norm": 13.256187438964844, + "learning_rate": 7.081178330817631e-06, + "loss": 10.401119232177734, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.5070035718381405, + "loss_total_v6": 0.6500699818134308, + "step": 430 + }, + { + "epoch": 0.5175622936055239, + "grad_norm": 13.659975051879883, + "learning_rate": 7.0585894873794514e-06, + "loss": 10.588777542114258, + "loss_alignment": 0.301025390625, + "loss_alignment_w": 0.1505126953125, + "loss_sft": 0.5113926641643047, + "loss_total_v6": 0.6617985516786575, + "step": 431 + }, + { + "epoch": 0.5187631341939357, + "grad_norm": 13.335607528686523, + "learning_rate": 7.035949937727325e-06, + "loss": 10.679037094116211, + "loss_alignment": 0.284423828125, + "loss_alignment_w": 0.1422119140625, + "loss_sft": 0.5254872813820839, + "loss_total_v6": 0.6674397960305214, + "step": 432 + }, + { + "epoch": 0.5199639747823477, + "grad_norm": 13.647432327270508, + "learning_rate": 7.013260239508055e-06, + "loss": 10.336041450500488, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.5018070191144943, + "loss_total_v6": 0.6460025906562805, + "step": 433 + }, + { + "epoch": 0.5211648153707595, + "grad_norm": 13.112367630004883, + "learning_rate": 6.990520951603682e-06, + "loss": 10.326403617858887, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.498366542160511, + "loss_total_v6": 0.6454002410173416, + "step": 434 + }, + { + "epoch": 0.5223656559591714, + "grad_norm": 13.0883150100708, + "learning_rate": 6.9677326341177145e-06, + "loss": 10.608328819274902, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5168108344078064, + "loss_total_v6": 0.6630205810070038, + "step": 435 + }, + { + "epoch": 0.5235664965475834, + "grad_norm": 13.055717468261719, + "learning_rate": 6.944895848361338e-06, + "loss": 10.217334747314453, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.4912598878145218, + "loss_total_v6": 0.6385834589600563, + "step": 436 + }, + { + "epoch": 0.5247673371359952, + "grad_norm": 13.503323554992676, + "learning_rate": 6.9220111568395835e-06, + "loss": 10.40333366394043, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.5028542056679726, + "loss_total_v6": 0.6502083465456963, + "step": 437 + }, + { + "epoch": 0.5259681777244071, + "grad_norm": 12.64638900756836, + "learning_rate": 6.899079123237473e-06, + "loss": 10.60662841796875, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.5194969214498997, + "loss_total_v6": 0.6629142835736275, + "step": 438 + }, + { + "epoch": 0.5271690183128189, + "grad_norm": 14.427616119384766, + "learning_rate": 6.876100312406141e-06, + "loss": 10.558539390563965, + "loss_alignment": 0.296630859375, + "loss_alignment_w": 0.1483154296875, + "loss_sft": 0.5117763876914978, + "loss_total_v6": 0.659908726811409, + "step": 439 + }, + { + "epoch": 0.5283698589012309, + "grad_norm": 14.249329566955566, + "learning_rate": 6.853075290348913e-06, + "loss": 10.405811309814453, + "loss_alignment": 0.284423828125, + "loss_alignment_w": 0.1422119140625, + "loss_sft": 0.507937591522932, + "loss_total_v6": 0.6503631547093391, + "step": 440 + }, + { + "epoch": 0.5295706994896427, + "grad_norm": 13.233119010925293, + "learning_rate": 6.830004624207368e-06, + "loss": 10.658005714416504, + "loss_alignment": 0.2906494140625, + "loss_alignment_w": 0.14532470703125, + "loss_sft": 0.5210600234568119, + "loss_total_v6": 0.6661253347992897, + "step": 441 + }, + { + "epoch": 0.5307715400780546, + "grad_norm": 13.265010833740234, + "learning_rate": 6.8068888822473735e-06, + "loss": 10.227020263671875, + "loss_alignment": 0.280029296875, + "loss_alignment_w": 0.1400146484375, + "loss_sft": 0.4994030110538006, + "loss_total_v6": 0.6391887813806534, + "step": 442 + }, + { + "epoch": 0.5319723806664666, + "grad_norm": 13.33232307434082, + "learning_rate": 6.783728633845076e-06, + "loss": 10.331169128417969, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.5006479844450951, + "loss_total_v6": 0.6456980630755424, + "step": 443 + }, + { + "epoch": 0.5331732212548784, + "grad_norm": 15.658650398254395, + "learning_rate": 6.760524449472889e-06, + "loss": 10.410555839538574, + "loss_alignment": 0.2801513671875, + "loss_alignment_w": 0.14007568359375, + "loss_sft": 0.5106603503227234, + "loss_total_v6": 0.6506597474217415, + "step": 444 + }, + { + "epoch": 0.5343740618432903, + "grad_norm": 13.089006423950195, + "learning_rate": 6.7372769006854355e-06, + "loss": 10.282716751098633, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.49696364626288414, + "loss_total_v6": 0.6426698192954063, + "step": 445 + }, + { + "epoch": 0.5355749024317021, + "grad_norm": 15.59626579284668, + "learning_rate": 6.71398656010547e-06, + "loss": 10.555440902709961, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5132612623274326, + "loss_total_v6": 0.6597151011228561, + "step": 446 + }, + { + "epoch": 0.5367757430201141, + "grad_norm": 13.766706466674805, + "learning_rate": 6.690654001409773e-06, + "loss": 10.589094161987305, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.5152119621634483, + "loss_total_v6": 0.6618184074759483, + "step": 447 + }, + { + "epoch": 0.537976583608526, + "grad_norm": 14.623412132263184, + "learning_rate": 6.667279799315025e-06, + "loss": 10.502900123596191, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5100384652614594, + "loss_total_v6": 0.6564312726259232, + "step": 448 + }, + { + "epoch": 0.5391774241969378, + "grad_norm": 14.28329086303711, + "learning_rate": 6.643864529563644e-06, + "loss": 10.568023681640625, + "loss_alignment": 0.283203125, + "loss_alignment_w": 0.1416015625, + "loss_sft": 0.518640547990799, + "loss_total_v6": 0.6605014726519585, + "step": 449 + }, + { + "epoch": 0.5403782647853498, + "grad_norm": 13.22235107421875, + "learning_rate": 6.62040876890961e-06, + "loss": 10.291336059570312, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.49623577669262886, + "loss_total_v6": 0.643208459019661, + "step": 450 + }, + { + "epoch": 0.5415791053737616, + "grad_norm": 14.1896333694458, + "learning_rate": 6.596913095104254e-06, + "loss": 10.681130409240723, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5208116360008717, + "loss_total_v6": 0.6675706580281258, + "step": 451 + }, + { + "epoch": 0.5427799459621735, + "grad_norm": 13.492206573486328, + "learning_rate": 6.573378086882031e-06, + "loss": 10.436399459838867, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.5036085844039917, + "loss_total_v6": 0.652274951338768, + "step": 452 + }, + { + "epoch": 0.5439807865505855, + "grad_norm": 14.573113441467285, + "learning_rate": 6.54980432394626e-06, + "loss": 10.331741333007812, + "loss_alignment": 0.2745361328125, + "loss_alignment_w": 0.13726806640625, + "loss_sft": 0.5088167116045952, + "loss_total_v6": 0.6457338184118271, + "step": 453 + }, + { + "epoch": 0.5451816271389973, + "grad_norm": 13.210264205932617, + "learning_rate": 6.526192386954853e-06, + "loss": 10.384389877319336, + "loss_alignment": 0.29541015625, + "loss_alignment_w": 0.147705078125, + "loss_sft": 0.5014566108584404, + "loss_total_v6": 0.6490243673324585, + "step": 454 + }, + { + "epoch": 0.5463824677274092, + "grad_norm": 15.048332214355469, + "learning_rate": 6.502542857506001e-06, + "loss": 10.675209045410156, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.5246377363801003, + "loss_total_v6": 0.6672006174921989, + "step": 455 + }, + { + "epoch": 0.547583308315821, + "grad_norm": 12.957703590393066, + "learning_rate": 6.478856318123861e-06, + "loss": 10.59524154663086, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.5174271464347839, + "loss_total_v6": 0.6622025519609451, + "step": 456 + }, + { + "epoch": 0.548784148904233, + "grad_norm": 13.87967300415039, + "learning_rate": 6.455133352244195e-06, + "loss": 10.216728210449219, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.4925646632909775, + "loss_total_v6": 0.6385455057024956, + "step": 457 + }, + { + "epoch": 0.5499849894926448, + "grad_norm": 12.856234550476074, + "learning_rate": 6.431374544200013e-06, + "loss": 10.652717590332031, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.5169300958514214, + "loss_total_v6": 0.665794849395752, + "step": 458 + }, + { + "epoch": 0.5511858300810567, + "grad_norm": 13.453978538513184, + "learning_rate": 6.407580479207166e-06, + "loss": 10.335090637207031, + "loss_alignment": 0.290283203125, + "loss_alignment_w": 0.1451416015625, + "loss_sft": 0.5008015483617783, + "loss_total_v6": 0.6459431573748589, + "step": 459 + }, + { + "epoch": 0.5523866706694687, + "grad_norm": 12.725017547607422, + "learning_rate": 6.383751743349942e-06, + "loss": 10.590585708618164, + "loss_alignment": 0.2752685546875, + "loss_alignment_w": 0.13763427734375, + "loss_sft": 0.5242468416690826, + "loss_total_v6": 0.6619116440415382, + "step": 460 + }, + { + "epoch": 0.5535875112578805, + "grad_norm": 13.424452781677246, + "learning_rate": 6.359888923566621e-06, + "loss": 10.544189453125, + "loss_alignment": 0.290283203125, + "loss_alignment_w": 0.1451416015625, + "loss_sft": 0.513885460793972, + "loss_total_v6": 0.6590117961168289, + "step": 461 + }, + { + "epoch": 0.5547883518462924, + "grad_norm": 13.68714714050293, + "learning_rate": 6.335992607635027e-06, + "loss": 10.391884803771973, + "loss_alignment": 0.2891845703125, + "loss_alignment_w": 0.14459228515625, + "loss_sft": 0.5049920491874218, + "loss_total_v6": 0.6494927704334259, + "step": 462 + }, + { + "epoch": 0.5559891924347042, + "grad_norm": 13.637923240661621, + "learning_rate": 6.312063384158044e-06, + "loss": 10.589103698730469, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5143580548465252, + "loss_total_v6": 0.6618190258741379, + "step": 463 + }, + { + "epoch": 0.5571900330231162, + "grad_norm": 13.91714859008789, + "learning_rate": 6.288101842549117e-06, + "loss": 10.666608810424805, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.5223912112414837, + "loss_total_v6": 0.6666630432009697, + "step": 464 + }, + { + "epoch": 0.5583908736115281, + "grad_norm": 13.262784957885742, + "learning_rate": 6.264108573017739e-06, + "loss": 10.366815567016602, + "loss_alignment": 0.291748046875, + "loss_alignment_w": 0.1458740234375, + "loss_sft": 0.5022045224905014, + "loss_total_v6": 0.6479259580373764, + "step": 465 + }, + { + "epoch": 0.5595917141999399, + "grad_norm": 13.007487297058105, + "learning_rate": 6.240084166554909e-06, + "loss": 10.414974212646484, + "loss_alignment": 0.29833984375, + "loss_alignment_w": 0.149169921875, + "loss_sft": 0.5018880553543568, + "loss_total_v6": 0.6509359255433083, + "step": 466 + }, + { + "epoch": 0.5607925547883519, + "grad_norm": 15.262102127075195, + "learning_rate": 6.2160292149185765e-06, + "loss": 10.287354469299316, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.4990540072321892, + "loss_total_v6": 0.6429596468806267, + "step": 467 + }, + { + "epoch": 0.5619933953767637, + "grad_norm": 13.829389572143555, + "learning_rate": 6.191944310619066e-06, + "loss": 10.169233322143555, + "loss_alignment": 0.2825927734375, + "loss_alignment_w": 0.14129638671875, + "loss_sft": 0.49443330988287926, + "loss_total_v6": 0.6355770975351334, + "step": 468 + }, + { + "epoch": 0.5631942359651756, + "grad_norm": 14.299245834350586, + "learning_rate": 6.167830046904481e-06, + "loss": 10.489315032958984, + "loss_alignment": 0.2998046875, + "loss_alignment_w": 0.14990234375, + "loss_sft": 0.5058172047138214, + "loss_total_v6": 0.6555822119116783, + "step": 469 + }, + { + "epoch": 0.5643950765535876, + "grad_norm": 13.688910484313965, + "learning_rate": 6.143687017746092e-06, + "loss": 10.542692184448242, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.5135324746370316, + "loss_total_v6": 0.6589182242751122, + "step": 470 + }, + { + "epoch": 0.5655959171419994, + "grad_norm": 13.767922401428223, + "learning_rate": 6.119515817823709e-06, + "loss": 10.430647850036621, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5045766308903694, + "loss_total_v6": 0.6519154906272888, + "step": 471 + }, + { + "epoch": 0.5667967577304113, + "grad_norm": 13.68610954284668, + "learning_rate": 6.095317042511028e-06, + "loss": 10.258862495422363, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.4998062588274479, + "loss_total_v6": 0.6411789283156395, + "step": 472 + }, + { + "epoch": 0.5679975983188231, + "grad_norm": 13.023197174072266, + "learning_rate": 6.071091287860973e-06, + "loss": 10.280057907104492, + "loss_alignment": 0.28173828125, + "loss_alignment_w": 0.140869140625, + "loss_sft": 0.501756526529789, + "loss_total_v6": 0.642503596842289, + "step": 473 + }, + { + "epoch": 0.5691984389072351, + "grad_norm": 14.178404808044434, + "learning_rate": 6.0468391505910064e-06, + "loss": 10.703865051269531, + "loss_alignment": 0.294677734375, + "loss_alignment_w": 0.1473388671875, + "loss_sft": 0.5216068848967552, + "loss_total_v6": 0.6689915359020233, + "step": 474 + }, + { + "epoch": 0.570399279495647, + "grad_norm": 13.222380638122559, + "learning_rate": 6.022561228068435e-06, + "loss": 10.425679206848145, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.5085537880659103, + "loss_total_v6": 0.651604950428009, + "step": 475 + }, + { + "epoch": 0.5716001200840588, + "grad_norm": 13.5001802444458, + "learning_rate": 5.998258118295699e-06, + "loss": 10.421666145324707, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5044120028614998, + "loss_total_v6": 0.6513541415333748, + "step": 476 + }, + { + "epoch": 0.5728009606724708, + "grad_norm": 14.937047958374023, + "learning_rate": 5.973930419895638e-06, + "loss": 10.281848907470703, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.49724506586790085, + "loss_total_v6": 0.6426155641674995, + "step": 477 + }, + { + "epoch": 0.5740018012608826, + "grad_norm": 13.330464363098145, + "learning_rate": 5.9495787320967465e-06, + "loss": 10.360252380371094, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.500772051513195, + "loss_total_v6": 0.6475158110260963, + "step": 478 + }, + { + "epoch": 0.5752026418492945, + "grad_norm": 13.30711555480957, + "learning_rate": 5.925203654718416e-06, + "loss": 10.63255500793457, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.5215598195791245, + "loss_total_v6": 0.6645347028970718, + "step": 479 + }, + { + "epoch": 0.5764034824377064, + "grad_norm": 12.590935707092285, + "learning_rate": 5.900805788156155e-06, + "loss": 10.471479415893555, + "loss_alignment": 0.2901611328125, + "loss_alignment_w": 0.14508056640625, + "loss_sft": 0.5095852613449097, + "loss_total_v6": 0.6544674560427666, + "step": 480 + }, + { + "epoch": 0.5776043230261183, + "grad_norm": 12.832357406616211, + "learning_rate": 5.876385733366811e-06, + "loss": 10.189651489257812, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4922914430499077, + "loss_total_v6": 0.6368532106280327, + "step": 481 + }, + { + "epoch": 0.5788051636145302, + "grad_norm": 12.557623863220215, + "learning_rate": 5.8519440918537565e-06, + "loss": 10.469928741455078, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.5099308714270592, + "loss_total_v6": 0.6543705835938454, + "step": 482 + }, + { + "epoch": 0.580006004202942, + "grad_norm": 13.814363479614258, + "learning_rate": 5.8274814656520795e-06, + "loss": 10.255504608154297, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.49292825162410736, + "loss_total_v6": 0.6409690231084824, + "step": 483 + }, + { + "epoch": 0.581206844791354, + "grad_norm": 12.578022956848145, + "learning_rate": 5.8029984573137545e-06, + "loss": 10.449186325073242, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.5099772624671459, + "loss_total_v6": 0.6530741751194, + "step": 484 + }, + { + "epoch": 0.5824076853797658, + "grad_norm": 13.59779167175293, + "learning_rate": 5.778495669892797e-06, + "loss": 10.57998275756836, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5147645510733128, + "loss_total_v6": 0.6612489297986031, + "step": 485 + }, + { + "epoch": 0.5836085259681777, + "grad_norm": 12.41385269165039, + "learning_rate": 5.753973706930414e-06, + "loss": 10.25668716430664, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4976256228983402, + "loss_total_v6": 0.6410429775714874, + "step": 486 + }, + { + "epoch": 0.5848093665565897, + "grad_norm": 15.536314010620117, + "learning_rate": 5.729433172440133e-06, + "loss": 10.076804161071777, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.4831022433936596, + "loss_total_v6": 0.6298002526164055, + "step": 487 + }, + { + "epoch": 0.5860102071450015, + "grad_norm": 13.167682647705078, + "learning_rate": 5.7048746708929295e-06, + "loss": 10.238680839538574, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.4951574355363846, + "loss_total_v6": 0.6399175524711609, + "step": 488 + }, + { + "epoch": 0.5872110477334134, + "grad_norm": 14.59572696685791, + "learning_rate": 5.680298807202332e-06, + "loss": 10.251449584960938, + "loss_alignment": 0.2978515625, + "loss_alignment_w": 0.14892578125, + "loss_sft": 0.4917287565767765, + "loss_total_v6": 0.6407155767083168, + "step": 489 + }, + { + "epoch": 0.5884118883218252, + "grad_norm": 12.965907096862793, + "learning_rate": 5.655706186709526e-06, + "loss": 10.415971755981445, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.5082675516605377, + "loss_total_v6": 0.6509982496500015, + "step": 490 + }, + { + "epoch": 0.5896127289102372, + "grad_norm": 13.841190338134766, + "learning_rate": 5.631097415168444e-06, + "loss": 10.337448120117188, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.5025510750710964, + "loss_total_v6": 0.646090529859066, + "step": 491 + }, + { + "epoch": 0.590813569498649, + "grad_norm": 12.854229927062988, + "learning_rate": 5.606473098730844e-06, + "loss": 10.300878524780273, + "loss_alignment": 0.284423828125, + "loss_alignment_w": 0.1422119140625, + "loss_sft": 0.5015471950173378, + "loss_total_v6": 0.6438048928976059, + "step": 492 + }, + { + "epoch": 0.5920144100870609, + "grad_norm": 13.396013259887695, + "learning_rate": 5.581833843931378e-06, + "loss": 10.38999080657959, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5036682523787022, + "loss_total_v6": 0.6493744403123856, + "step": 493 + }, + { + "epoch": 0.5932152506754729, + "grad_norm": 12.630484580993652, + "learning_rate": 5.557180257672651e-06, + "loss": 10.300992965698242, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.49952495843172073, + "loss_total_v6": 0.6438120752573013, + "step": 494 + }, + { + "epoch": 0.5944160912638847, + "grad_norm": 12.729873657226562, + "learning_rate": 5.5325129472102755e-06, + "loss": 10.149979591369629, + "loss_alignment": 0.2720947265625, + "loss_alignment_w": 0.13604736328125, + "loss_sft": 0.4983568526804447, + "loss_total_v6": 0.63437370210886, + "step": 495 + }, + { + "epoch": 0.5956169318522966, + "grad_norm": 13.680142402648926, + "learning_rate": 5.507832520137913e-06, + "loss": 10.216117858886719, + "loss_alignment": 0.291015625, + "loss_alignment_w": 0.1455078125, + "loss_sft": 0.49298427626490593, + "loss_total_v6": 0.6385073438286781, + "step": 496 + }, + { + "epoch": 0.5968177724407085, + "grad_norm": 13.034774780273438, + "learning_rate": 5.483139584372305e-06, + "loss": 10.056310653686523, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.4845985025167465, + "loss_total_v6": 0.6285194084048271, + "step": 497 + }, + { + "epoch": 0.5980186130291204, + "grad_norm": 13.657236099243164, + "learning_rate": 5.458434748138302e-06, + "loss": 10.323333740234375, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.5017910525202751, + "loss_total_v6": 0.6452084109187126, + "step": 498 + }, + { + "epoch": 0.5992194536175323, + "grad_norm": 12.33836841583252, + "learning_rate": 5.433718619953883e-06, + "loss": 10.3373441696167, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.5024225413799286, + "loss_total_v6": 0.6460840404033661, + "step": 499 + }, + { + "epoch": 0.6004202942059441, + "grad_norm": 12.948677062988281, + "learning_rate": 5.408991808615163e-06, + "loss": 10.603233337402344, + "loss_alignment": 0.2796630859375, + "loss_alignment_w": 0.13983154296875, + "loss_sft": 0.5235114172101021, + "loss_total_v6": 0.6627020984888077, + "step": 500 + }, + { + "epoch": 0.6004202942059441, + "eval_loss": 0.6005465984344482, + "eval_loss_alignment": 0.2829167292668269, + "eval_loss_alignment_w": 0.14145836463341346, + "eval_loss_sft": 0.45908820966724306, + "eval_loss_total_v6": 0.600546574112601, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.485, + "eval_samples_per_second": 24.583, + "eval_steps_per_second": 3.082, + "step": 500 + }, + { + "epoch": 0.6016211347943561, + "grad_norm": 15.072790145874023, + "learning_rate": 5.384254923181399e-06, + "loss": 10.28596019744873, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.49499960616230965, + "loss_total_v6": 0.6428725123405457, + "step": 501 + }, + { + "epoch": 0.6028219753827679, + "grad_norm": 12.887722969055176, + "learning_rate": 5.359508572959992e-06, + "loss": 10.173322677612305, + "loss_alignment": 0.283203125, + "loss_alignment_w": 0.1416015625, + "loss_sft": 0.49421584233641624, + "loss_total_v6": 0.6358326599001884, + "step": 502 + }, + { + "epoch": 0.6040228159711798, + "grad_norm": 13.056795120239258, + "learning_rate": 5.33475336749147e-06, + "loss": 10.539846420288086, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.5123323388397694, + "loss_total_v6": 0.658740408718586, + "step": 503 + }, + { + "epoch": 0.6052236565595918, + "grad_norm": 13.36933422088623, + "learning_rate": 5.309989916534482e-06, + "loss": 10.483774185180664, + "loss_alignment": 0.2998046875, + "loss_alignment_w": 0.14990234375, + "loss_sft": 0.505119938403368, + "loss_total_v6": 0.6552358940243721, + "step": 504 + }, + { + "epoch": 0.6064244971480036, + "grad_norm": 12.703725814819336, + "learning_rate": 5.28521883005078e-06, + "loss": 10.474664688110352, + "loss_alignment": 0.308349609375, + "loss_alignment_w": 0.1541748046875, + "loss_sft": 0.5004154779016972, + "loss_total_v6": 0.6546665579080582, + "step": 505 + }, + { + "epoch": 0.6076253377364155, + "grad_norm": 13.047313690185547, + "learning_rate": 5.260440718190183e-06, + "loss": 10.293411254882812, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.49639610573649406, + "loss_total_v6": 0.6433382406830788, + "step": 506 + }, + { + "epoch": 0.6088261783248273, + "grad_norm": 15.672074317932129, + "learning_rate": 5.235656191275561e-06, + "loss": 10.17343521118164, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.49135422706604004, + "loss_total_v6": 0.6358396932482719, + "step": 507 + }, + { + "epoch": 0.6100270189132393, + "grad_norm": 12.881000518798828, + "learning_rate": 5.210865859787795e-06, + "loss": 10.378957748413086, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5032380372285843, + "loss_total_v6": 0.6486848518252373, + "step": 508 + }, + { + "epoch": 0.6112278595016511, + "grad_norm": 12.800606727600098, + "learning_rate": 5.1860703343507415e-06, + "loss": 10.18568229675293, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.493874479085207, + "loss_total_v6": 0.6366051807999611, + "step": 509 + }, + { + "epoch": 0.612428700090063, + "grad_norm": 12.557250022888184, + "learning_rate": 5.1612702257161925e-06, + "loss": 9.959648132324219, + "loss_alignment": 0.271728515625, + "loss_alignment_w": 0.1358642578125, + "loss_sft": 0.4867815412580967, + "loss_total_v6": 0.6224779635667801, + "step": 510 + }, + { + "epoch": 0.613629540678475, + "grad_norm": 13.167038917541504, + "learning_rate": 5.136466144748832e-06, + "loss": 10.447469711303711, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.5075353607535362, + "loss_total_v6": 0.6529668867588043, + "step": 511 + }, + { + "epoch": 0.6148303812668868, + "grad_norm": 12.57446575164795, + "learning_rate": 5.111658702411185e-06, + "loss": 10.301204681396484, + "loss_alignment": 0.2840576171875, + "loss_alignment_w": 0.14202880859375, + "loss_sft": 0.5018270239233971, + "loss_total_v6": 0.6438253000378609, + "step": 512 + }, + { + "epoch": 0.6160312218552987, + "grad_norm": 13.142411231994629, + "learning_rate": 5.086848509748576e-06, + "loss": 10.257258415222168, + "loss_alignment": 0.2764892578125, + "loss_alignment_w": 0.13824462890625, + "loss_sft": 0.5025898590683937, + "loss_total_v6": 0.6410786136984825, + "step": 513 + }, + { + "epoch": 0.6172320624437106, + "grad_norm": 12.939610481262207, + "learning_rate": 5.062036177874075e-06, + "loss": 10.503740310668945, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.5085040479898453, + "loss_total_v6": 0.6564837619662285, + "step": 514 + }, + { + "epoch": 0.6184329030321225, + "grad_norm": 14.044941902160645, + "learning_rate": 5.037222317953437e-06, + "loss": 10.176543235778809, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.49341005831956863, + "loss_total_v6": 0.6360339447855949, + "step": 515 + }, + { + "epoch": 0.6196337436205344, + "grad_norm": 13.403106689453125, + "learning_rate": 5.0124075411900655e-06, + "loss": 10.31628704071045, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.4966356083750725, + "loss_total_v6": 0.6447679325938225, + "step": 516 + }, + { + "epoch": 0.6208345842089462, + "grad_norm": 13.451895713806152, + "learning_rate": 4.987592458809937e-06, + "loss": 10.147735595703125, + "loss_alignment": 0.284423828125, + "loss_alignment_w": 0.1422119140625, + "loss_sft": 0.4918231703341007, + "loss_total_v6": 0.6342334598302841, + "step": 517 + }, + { + "epoch": 0.6220354247973582, + "grad_norm": 12.755326271057129, + "learning_rate": 4.962777682046565e-06, + "loss": 10.195171356201172, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.49237703531980515, + "loss_total_v6": 0.637198194861412, + "step": 518 + }, + { + "epoch": 0.62323626538577, + "grad_norm": 13.713449478149414, + "learning_rate": 4.937963822125928e-06, + "loss": 10.239023208618164, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.49714723974466324, + "loss_total_v6": 0.6399389877915382, + "step": 519 + }, + { + "epoch": 0.6244371059741819, + "grad_norm": 13.046950340270996, + "learning_rate": 4.913151490251426e-06, + "loss": 10.092033386230469, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4874110743403435, + "loss_total_v6": 0.6307521238923073, + "step": 520 + }, + { + "epoch": 0.6256379465625939, + "grad_norm": 12.391763687133789, + "learning_rate": 4.888341297588816e-06, + "loss": 10.659740447998047, + "loss_alignment": 0.28857421875, + "loss_alignment_w": 0.144287109375, + "loss_sft": 0.5220839902758598, + "loss_total_v6": 0.666233740746975, + "step": 521 + }, + { + "epoch": 0.6268387871510057, + "grad_norm": 12.742067337036133, + "learning_rate": 4.8635338552511705e-06, + "loss": 10.213722229003906, + "loss_alignment": 0.2850341796875, + "loss_alignment_w": 0.14251708984375, + "loss_sft": 0.49599314853549004, + "loss_total_v6": 0.6383576542139053, + "step": 522 + }, + { + "epoch": 0.6280396277394176, + "grad_norm": 13.011760711669922, + "learning_rate": 4.838729774283808e-06, + "loss": 10.401603698730469, + "loss_alignment": 0.300048828125, + "loss_alignment_w": 0.1500244140625, + "loss_sft": 0.5001368336379528, + "loss_total_v6": 0.6501002237200737, + "step": 523 + }, + { + "epoch": 0.6292404683278294, + "grad_norm": 13.444523811340332, + "learning_rate": 4.813929665649261e-06, + "loss": 10.184769630432129, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.49187955260276794, + "loss_total_v6": 0.6365481242537498, + "step": 524 + }, + { + "epoch": 0.6304413089162414, + "grad_norm": 14.027856826782227, + "learning_rate": 4.7891341402122065e-06, + "loss": 10.39614200592041, + "loss_alignment": 0.302734375, + "loss_alignment_w": 0.1513671875, + "loss_sft": 0.49834591895341873, + "loss_total_v6": 0.6497588977217674, + "step": 525 + }, + { + "epoch": 0.6316421495046532, + "grad_norm": 13.067005157470703, + "learning_rate": 4.764343808724442e-06, + "loss": 10.282273292541504, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.4980803281068802, + "loss_total_v6": 0.6426420956850052, + "step": 526 + }, + { + "epoch": 0.6328429900930651, + "grad_norm": 12.918837547302246, + "learning_rate": 4.739559281809819e-06, + "loss": 10.301896095275879, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.4995661526918411, + "loss_total_v6": 0.6438684985041618, + "step": 527 + }, + { + "epoch": 0.6340438306814771, + "grad_norm": 13.650721549987793, + "learning_rate": 4.714781169949223e-06, + "loss": 10.461540222167969, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.5096660368144512, + "loss_total_v6": 0.6538463234901428, + "step": 528 + }, + { + "epoch": 0.6352446712698889, + "grad_norm": 12.097370147705078, + "learning_rate": 4.6900100834655185e-06, + "loss": 10.332887649536133, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.4983902834355831, + "loss_total_v6": 0.6458054482936859, + "step": 529 + }, + { + "epoch": 0.6364455118583008, + "grad_norm": 18.400671005249023, + "learning_rate": 4.665246632508531e-06, + "loss": 10.314319610595703, + "loss_alignment": 0.2978515625, + "loss_alignment_w": 0.14892578125, + "loss_sft": 0.49587180837988853, + "loss_total_v6": 0.6446449980139732, + "step": 530 + }, + { + "epoch": 0.6376463524467127, + "grad_norm": 13.018949508666992, + "learning_rate": 4.640491427040011e-06, + "loss": 10.212677001953125, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4953174479305744, + "loss_total_v6": 0.6382923126220703, + "step": 531 + }, + { + "epoch": 0.6388471930351246, + "grad_norm": 15.10692024230957, + "learning_rate": 4.615745076818602e-06, + "loss": 10.137521743774414, + "loss_alignment": 0.2899169921875, + "loss_alignment_w": 0.14495849609375, + "loss_sft": 0.4885450340807438, + "loss_total_v6": 0.6335950717329979, + "step": 532 + }, + { + "epoch": 0.6400480336235365, + "grad_norm": 13.029293060302734, + "learning_rate": 4.591008191384839e-06, + "loss": 10.463785171508789, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.5051371268928051, + "loss_total_v6": 0.6539866179227829, + "step": 533 + }, + { + "epoch": 0.6412488742119483, + "grad_norm": 15.196734428405762, + "learning_rate": 4.566281380046117e-06, + "loss": 10.395520210266113, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.5063179060816765, + "loss_total_v6": 0.6497200205922127, + "step": 534 + }, + { + "epoch": 0.6424497148003603, + "grad_norm": 12.814765930175781, + "learning_rate": 4.541565251861699e-06, + "loss": 10.212923049926758, + "loss_alignment": 0.3017578125, + "loss_alignment_w": 0.15087890625, + "loss_sft": 0.48748981580138206, + "loss_total_v6": 0.6383076906204224, + "step": 535 + }, + { + "epoch": 0.6436505553887721, + "grad_norm": 13.029295921325684, + "learning_rate": 4.5168604156276954e-06, + "loss": 9.887266159057617, + "loss_alignment": 0.2763671875, + "loss_alignment_w": 0.13818359375, + "loss_sft": 0.4797399900853634, + "loss_total_v6": 0.6179541051387787, + "step": 536 + }, + { + "epoch": 0.644851395977184, + "grad_norm": 12.631991386413574, + "learning_rate": 4.492167479862088e-06, + "loss": 10.380716323852539, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.5043398588895798, + "loss_total_v6": 0.6487948074936867, + "step": 537 + }, + { + "epoch": 0.646052236565596, + "grad_norm": 12.72430419921875, + "learning_rate": 4.4674870527897244e-06, + "loss": 9.887868881225586, + "loss_alignment": 0.2841796875, + "loss_alignment_w": 0.14208984375, + "loss_sft": 0.47582563757896423, + "loss_total_v6": 0.6179917752742767, + "step": 538 + }, + { + "epoch": 0.6472530771540078, + "grad_norm": 12.51884651184082, + "learning_rate": 4.44281974232735e-06, + "loss": 10.383283615112305, + "loss_alignment": 0.2958984375, + "loss_alignment_w": 0.14794921875, + "loss_sft": 0.501173846423626, + "loss_total_v6": 0.6489552333950996, + "step": 539 + }, + { + "epoch": 0.6484539177424197, + "grad_norm": 12.981233596801758, + "learning_rate": 4.418166156068623e-06, + "loss": 10.392187118530273, + "loss_alignment": 0.2974853515625, + "loss_alignment_w": 0.14874267578125, + "loss_sft": 0.5006011091172695, + "loss_total_v6": 0.6495116502046585, + "step": 540 + }, + { + "epoch": 0.6496547583308315, + "grad_norm": 12.859923362731934, + "learning_rate": 4.393526901269158e-06, + "loss": 10.381662368774414, + "loss_alignment": 0.2835693359375, + "loss_alignment_w": 0.14178466796875, + "loss_sft": 0.5069776438176632, + "loss_total_v6": 0.6488538682460785, + "step": 541 + }, + { + "epoch": 0.6508555989192435, + "grad_norm": 12.044489860534668, + "learning_rate": 4.368902584831557e-06, + "loss": 10.125541687011719, + "loss_alignment": 0.282470703125, + "loss_alignment_w": 0.1412353515625, + "loss_sft": 0.4916110821068287, + "loss_total_v6": 0.6328464075922966, + "step": 542 + }, + { + "epoch": 0.6520564395076553, + "grad_norm": 12.60694408416748, + "learning_rate": 4.344293813290477e-06, + "loss": 10.312738418579102, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.497970275580883, + "loss_total_v6": 0.6445461958646774, + "step": 543 + }, + { + "epoch": 0.6532572800960672, + "grad_norm": 12.803033828735352, + "learning_rate": 4.319701192797671e-06, + "loss": 10.148714065551758, + "loss_alignment": 0.279296875, + "loss_alignment_w": 0.1396484375, + "loss_sft": 0.49461568892002106, + "loss_total_v6": 0.6342946216464043, + "step": 544 + }, + { + "epoch": 0.6544581206844792, + "grad_norm": 12.584644317626953, + "learning_rate": 4.295125329107074e-06, + "loss": 9.959033966064453, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.4783662036061287, + "loss_total_v6": 0.6224396675825119, + "step": 545 + }, + { + "epoch": 0.655658961272891, + "grad_norm": 12.868959426879883, + "learning_rate": 4.270566827559868e-06, + "loss": 10.44150447845459, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.5119689851999283, + "loss_total_v6": 0.6525940448045731, + "step": 546 + }, + { + "epoch": 0.6568598018613029, + "grad_norm": 12.868577003479004, + "learning_rate": 4.246026293069588e-06, + "loss": 10.07911491394043, + "loss_alignment": 0.2755126953125, + "loss_alignment_w": 0.13775634765625, + "loss_sft": 0.49234092608094215, + "loss_total_v6": 0.6299446672201157, + "step": 547 + }, + { + "epoch": 0.6580606424497148, + "grad_norm": 12.744641304016113, + "learning_rate": 4.221504330107204e-06, + "loss": 9.798276901245117, + "loss_alignment": 0.274169921875, + "loss_alignment_w": 0.1370849609375, + "loss_sft": 0.475551500916481, + "loss_total_v6": 0.612392321228981, + "step": 548 + }, + { + "epoch": 0.6592614830381267, + "grad_norm": 12.618844985961914, + "learning_rate": 4.197001542686248e-06, + "loss": 10.152666091918945, + "loss_alignment": 0.2752685546875, + "loss_alignment_w": 0.13763427734375, + "loss_sft": 0.4971514865756035, + "loss_total_v6": 0.6345416232943535, + "step": 549 + }, + { + "epoch": 0.6604623236265386, + "grad_norm": 13.055123329162598, + "learning_rate": 4.172518534347922e-06, + "loss": 10.320218086242676, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.5000551156699657, + "loss_total_v6": 0.6450136005878448, + "step": 550 + }, + { + "epoch": 0.6616631642149504, + "grad_norm": 13.165721893310547, + "learning_rate": 4.148055908146247e-06, + "loss": 10.348695755004883, + "loss_alignment": 0.298583984375, + "loss_alignment_w": 0.1492919921875, + "loss_sft": 0.4982186630368233, + "loss_total_v6": 0.6467934772372246, + "step": 551 + }, + { + "epoch": 0.6628640048033624, + "grad_norm": 13.947592735290527, + "learning_rate": 4.123614266633191e-06, + "loss": 10.080323219299316, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.4876709431409836, + "loss_total_v6": 0.6300201937556267, + "step": 552 + }, + { + "epoch": 0.6640648453917742, + "grad_norm": 12.695610046386719, + "learning_rate": 4.099194211843847e-06, + "loss": 10.420171737670898, + "loss_alignment": 0.288818359375, + "loss_alignment_w": 0.1444091796875, + "loss_sft": 0.5068973489105701, + "loss_total_v6": 0.6512607485055923, + "step": 553 + }, + { + "epoch": 0.6652656859801861, + "grad_norm": 13.364655494689941, + "learning_rate": 4.074796345281586e-06, + "loss": 9.991522789001465, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.48111382126808167, + "loss_total_v6": 0.6244701594114304, + "step": 554 + }, + { + "epoch": 0.6664665265685981, + "grad_norm": 12.41392993927002, + "learning_rate": 4.050421267903254e-06, + "loss": 10.168806076049805, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.4910649210214615, + "loss_total_v6": 0.6355503797531128, + "step": 555 + }, + { + "epoch": 0.6676673671570099, + "grad_norm": 12.291377067565918, + "learning_rate": 4.0260695801043626e-06, + "loss": 10.04928970336914, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.4877455569803715, + "loss_total_v6": 0.6280806362628937, + "step": 556 + }, + { + "epoch": 0.6688682077454218, + "grad_norm": 12.584766387939453, + "learning_rate": 4.001741881704302e-06, + "loss": 10.381609916687012, + "loss_alignment": 0.2860107421875, + "loss_alignment_w": 0.14300537109375, + "loss_sft": 0.505982581526041, + "loss_total_v6": 0.6488506123423576, + "step": 557 + }, + { + "epoch": 0.6700690483338337, + "grad_norm": 13.120285034179688, + "learning_rate": 3.977438771931566e-06, + "loss": 9.958975791931152, + "loss_alignment": 0.283935546875, + "loss_alignment_w": 0.1419677734375, + "loss_sft": 0.4803308695554733, + "loss_total_v6": 0.6224359646439552, + "step": 558 + }, + { + "epoch": 0.6712698889222456, + "grad_norm": 13.31394100189209, + "learning_rate": 3.953160849408996e-06, + "loss": 10.371875762939453, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.5068848021328449, + "loss_total_v6": 0.648242212831974, + "step": 559 + }, + { + "epoch": 0.6724707295106575, + "grad_norm": 12.817798614501953, + "learning_rate": 3.928908712139028e-06, + "loss": 10.000528335571289, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.47929631546139717, + "loss_total_v6": 0.6250330060720444, + "step": 560 + }, + { + "epoch": 0.6736715700990693, + "grad_norm": 12.821146011352539, + "learning_rate": 3.904682957488973e-06, + "loss": 10.003207206726074, + "loss_alignment": 0.2806396484375, + "loss_alignment_w": 0.14031982421875, + "loss_sft": 0.48462123051285744, + "loss_total_v6": 0.6252004578709602, + "step": 561 + }, + { + "epoch": 0.6748724106874813, + "grad_norm": 12.887779235839844, + "learning_rate": 3.880484182176292e-06, + "loss": 10.372209548950195, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.505150880664587, + "loss_total_v6": 0.6482630893588066, + "step": 562 + }, + { + "epoch": 0.6760732512758931, + "grad_norm": 12.393461227416992, + "learning_rate": 3.856312982253909e-06, + "loss": 10.281373977661133, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.4979630336165428, + "loss_total_v6": 0.642585851252079, + "step": 563 + }, + { + "epoch": 0.677274091864305, + "grad_norm": 13.377799034118652, + "learning_rate": 3.832169953095521e-06, + "loss": 10.159512519836426, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4914606064558029, + "loss_total_v6": 0.6349695026874542, + "step": 564 + }, + { + "epoch": 0.678474932452717, + "grad_norm": 13.18189525604248, + "learning_rate": 3.808055689380936e-06, + "loss": 10.286460876464844, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.5004935599863529, + "loss_total_v6": 0.6429038494825363, + "step": 565 + }, + { + "epoch": 0.6796757730411288, + "grad_norm": 12.65172290802002, + "learning_rate": 3.7839707850814243e-06, + "loss": 10.045024871826172, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.48282505199313164, + "loss_total_v6": 0.6278140693902969, + "step": 566 + }, + { + "epoch": 0.6808766136295407, + "grad_norm": 12.752508163452148, + "learning_rate": 3.7599158334450925e-06, + "loss": 10.056190490722656, + "loss_alignment": 0.285400390625, + "loss_alignment_w": 0.1427001953125, + "loss_sft": 0.4857354089617729, + "loss_total_v6": 0.6285119280219078, + "step": 567 + }, + { + "epoch": 0.6820774542179525, + "grad_norm": 12.435800552368164, + "learning_rate": 3.7358914269822623e-06, + "loss": 9.958332061767578, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.47652170807123184, + "loss_total_v6": 0.6223957315087318, + "step": 568 + }, + { + "epoch": 0.6832782948063645, + "grad_norm": 12.135154724121094, + "learning_rate": 3.7118981574508845e-06, + "loss": 10.13548469543457, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.4886160157620907, + "loss_total_v6": 0.6334677264094353, + "step": 569 + }, + { + "epoch": 0.6844791353947763, + "grad_norm": 12.61174201965332, + "learning_rate": 3.6879366158419578e-06, + "loss": 10.05044174194336, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.48468951508402824, + "loss_total_v6": 0.6281526535749435, + "step": 570 + }, + { + "epoch": 0.6856799759831882, + "grad_norm": 13.11091136932373, + "learning_rate": 3.6640073923649743e-06, + "loss": 10.07650375366211, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.48502136394381523, + "loss_total_v6": 0.6297814920544624, + "step": 571 + }, + { + "epoch": 0.6868808165716002, + "grad_norm": 12.58213996887207, + "learning_rate": 3.6401110764333802e-06, + "loss": 10.119109153747559, + "loss_alignment": 0.291259765625, + "loss_alignment_w": 0.1456298828125, + "loss_sft": 0.4868297129869461, + "loss_total_v6": 0.632444329559803, + "step": 572 + }, + { + "epoch": 0.688081657160012, + "grad_norm": 12.527066230773926, + "learning_rate": 3.61624825665006e-06, + "loss": 10.28040885925293, + "loss_alignment": 0.302001953125, + "loss_alignment_w": 0.1510009765625, + "loss_sft": 0.4920280836522579, + "loss_total_v6": 0.6425255462527275, + "step": 573 + }, + { + "epoch": 0.6892824977484239, + "grad_norm": 12.761164665222168, + "learning_rate": 3.5924195207928353e-06, + "loss": 10.136287689208984, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.489505548030138, + "loss_total_v6": 0.6335179805755615, + "step": 574 + }, + { + "epoch": 0.6904833383368358, + "grad_norm": 12.173779487609863, + "learning_rate": 3.568625455799989e-06, + "loss": 10.291715621948242, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.49503885954618454, + "loss_total_v6": 0.6432322189211845, + "step": 575 + }, + { + "epoch": 0.6916841789252477, + "grad_norm": 12.588232040405273, + "learning_rate": 3.5448666477558058e-06, + "loss": 10.094593048095703, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.48802870512008667, + "loss_total_v6": 0.6309120208024979, + "step": 576 + }, + { + "epoch": 0.6928850195136596, + "grad_norm": 12.921107292175293, + "learning_rate": 3.5211436818761412e-06, + "loss": 10.322202682495117, + "loss_alignment": 0.294677734375, + "loss_alignment_w": 0.1473388671875, + "loss_sft": 0.4981040060520172, + "loss_total_v6": 0.6451377123594284, + "step": 577 + }, + { + "epoch": 0.6940858601020714, + "grad_norm": 13.031691551208496, + "learning_rate": 3.4974571424940008e-06, + "loss": 10.14596939086914, + "loss_alignment": 0.291748046875, + "loss_alignment_w": 0.1458740234375, + "loss_sft": 0.4884169027209282, + "loss_total_v6": 0.6341230794787407, + "step": 578 + }, + { + "epoch": 0.6952867006904834, + "grad_norm": 12.799546241760254, + "learning_rate": 3.4738076130451486e-06, + "loss": 10.362672805786133, + "loss_alignment": 0.287109375, + "loss_alignment_w": 0.1435546875, + "loss_sft": 0.5039292499423027, + "loss_total_v6": 0.6476670503616333, + "step": 579 + }, + { + "epoch": 0.6964875412788952, + "grad_norm": 13.023580551147461, + "learning_rate": 3.450195676053739e-06, + "loss": 10.366745948791504, + "loss_alignment": 0.287109375, + "loss_alignment_w": 0.1435546875, + "loss_sft": 0.5041228048503399, + "loss_total_v6": 0.6479216367006302, + "step": 580 + }, + { + "epoch": 0.6976883818673071, + "grad_norm": 13.16224479675293, + "learning_rate": 3.4266219131179702e-06, + "loss": 10.321920394897461, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.4974912032485008, + "loss_total_v6": 0.6451200023293495, + "step": 581 + }, + { + "epoch": 0.698889222455719, + "grad_norm": 12.581500053405762, + "learning_rate": 3.403086904895746e-06, + "loss": 10.287764549255371, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.4960431903600693, + "loss_total_v6": 0.6429853141307831, + "step": 582 + }, + { + "epoch": 0.7000900630441309, + "grad_norm": 12.915231704711914, + "learning_rate": 3.379591231090391e-06, + "loss": 10.440594673156738, + "loss_alignment": 0.297119140625, + "loss_alignment_w": 0.1485595703125, + "loss_sft": 0.5037944950163364, + "loss_total_v6": 0.6525371745228767, + "step": 583 + }, + { + "epoch": 0.7012909036325428, + "grad_norm": 12.882129669189453, + "learning_rate": 3.3561354704363563e-06, + "loss": 10.159818649291992, + "loss_alignment": 0.2784423828125, + "loss_alignment_w": 0.13922119140625, + "loss_sft": 0.4957217611372471, + "loss_total_v6": 0.6349887102842331, + "step": 584 + }, + { + "epoch": 0.7024917442209546, + "grad_norm": 12.594104766845703, + "learning_rate": 3.332720200684977e-06, + "loss": 10.26405143737793, + "loss_alignment": 0.29638671875, + "loss_alignment_w": 0.148193359375, + "loss_sft": 0.4932335652410984, + "loss_total_v6": 0.6415032297372818, + "step": 585 + }, + { + "epoch": 0.7036925848093666, + "grad_norm": 13.303410530090332, + "learning_rate": 3.309345998590228e-06, + "loss": 10.289093017578125, + "loss_alignment": 0.30029296875, + "loss_alignment_w": 0.150146484375, + "loss_sft": 0.4930438734591007, + "loss_total_v6": 0.6430682986974716, + "step": 586 + }, + { + "epoch": 0.7048934253977784, + "grad_norm": 12.892866134643555, + "learning_rate": 3.286013439894532e-06, + "loss": 10.354276657104492, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.49984922632575035, + "loss_total_v6": 0.6471422910690308, + "step": 587 + }, + { + "epoch": 0.7060942659861903, + "grad_norm": 13.714106559753418, + "learning_rate": 3.262723099314565e-06, + "loss": 10.149900436401367, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.49267571046948433, + "loss_total_v6": 0.6343688368797302, + "step": 588 + }, + { + "epoch": 0.7072951065746023, + "grad_norm": 13.252981185913086, + "learning_rate": 3.2394755505271125e-06, + "loss": 9.816271781921387, + "loss_alignment": 0.2855224609375, + "loss_alignment_w": 0.14276123046875, + "loss_sft": 0.4707404896616936, + "loss_total_v6": 0.6135169714689255, + "step": 589 + }, + { + "epoch": 0.7084959471630141, + "grad_norm": 13.331924438476562, + "learning_rate": 3.216271366154925e-06, + "loss": 9.942400932312012, + "loss_alignment": 0.2830810546875, + "loss_alignment_w": 0.14154052734375, + "loss_sft": 0.4800731725990772, + "loss_total_v6": 0.6214000731706619, + "step": 590 + }, + { + "epoch": 0.709696787751426, + "grad_norm": 12.65157699584961, + "learning_rate": 3.1931111177526285e-06, + "loss": 10.201189041137695, + "loss_alignment": 0.281982421875, + "loss_alignment_w": 0.1409912109375, + "loss_sft": 0.49664412438869476, + "loss_total_v6": 0.6375742852687836, + "step": 591 + }, + { + "epoch": 0.7108976283398379, + "grad_norm": 12.39628791809082, + "learning_rate": 3.1699953757926322e-06, + "loss": 10.194109916687012, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.49051015079021454, + "loss_total_v6": 0.6371318623423576, + "step": 592 + }, + { + "epoch": 0.7120984689282498, + "grad_norm": 12.9371337890625, + "learning_rate": 3.1469247096510893e-06, + "loss": 9.924175262451172, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.4777286499738693, + "loss_total_v6": 0.62026097625494, + "step": 593 + }, + { + "epoch": 0.7132993095166617, + "grad_norm": 12.828069686889648, + "learning_rate": 3.1238996875938604e-06, + "loss": 10.270505905151367, + "loss_alignment": 0.2978515625, + "loss_alignment_w": 0.14892578125, + "loss_sft": 0.49336225539445877, + "loss_total_v6": 0.6419065669178963, + "step": 594 + }, + { + "epoch": 0.7145001501050735, + "grad_norm": 13.762063980102539, + "learning_rate": 3.1009208767625293e-06, + "loss": 10.384879112243652, + "loss_alignment": 0.28173828125, + "loss_alignment_w": 0.140869140625, + "loss_sft": 0.5083842054009438, + "loss_total_v6": 0.6490549519658089, + "step": 595 + }, + { + "epoch": 0.7157009906934855, + "grad_norm": 12.737523078918457, + "learning_rate": 3.0779888431604186e-06, + "loss": 10.355340957641602, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.5001445710659027, + "loss_total_v6": 0.6472087800502777, + "step": 596 + }, + { + "epoch": 0.7169018312818973, + "grad_norm": 13.11309814453125, + "learning_rate": 3.0551041516386646e-06, + "loss": 9.931150436401367, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.47689809277653694, + "loss_total_v6": 0.6206969171762466, + "step": 597 + }, + { + "epoch": 0.7181026718703092, + "grad_norm": 12.49449348449707, + "learning_rate": 3.0322673658822867e-06, + "loss": 10.15231990814209, + "loss_alignment": 0.285400390625, + "loss_alignment_w": 0.1427001953125, + "loss_sft": 0.4917587861418724, + "loss_total_v6": 0.634520024061203, + "step": 598 + }, + { + "epoch": 0.7193035124587212, + "grad_norm": 14.135873794555664, + "learning_rate": 3.009479048396321e-06, + "loss": 10.268256187438965, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.4951290711760521, + "loss_total_v6": 0.6417660266160965, + "step": 599 + }, + { + "epoch": 0.720504353047133, + "grad_norm": 12.229159355163574, + "learning_rate": 2.986739760491946e-06, + "loss": 10.326519012451172, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.5026461631059647, + "loss_total_v6": 0.6454074084758759, + "step": 600 + }, + { + "epoch": 0.720504353047133, + "eval_loss": 0.5926557183265686, + "eval_loss_alignment": 0.281768798828125, + "eval_loss_alignment_w": 0.1408843994140625, + "eval_loss_sft": 0.45177134318957823, + "eval_loss_total_v6": 0.5926557424872254, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.3993, + "eval_samples_per_second": 24.615, + "eval_steps_per_second": 3.086, + "step": 600 + }, + { + "epoch": 0.7217051936355449, + "grad_norm": 12.753277778625488, + "learning_rate": 2.9640500622726777e-06, + "loss": 10.070724487304688, + "loss_alignment": 0.2738037109375, + "loss_alignment_w": 0.13690185546875, + "loss_sft": 0.49259473010897636, + "loss_total_v6": 0.6294202953577042, + "step": 601 + }, + { + "epoch": 0.7229060342239567, + "grad_norm": 13.880197525024414, + "learning_rate": 2.94141051262055e-06, + "loss": 10.341669082641602, + "loss_alignment": 0.30224609375, + "loss_alignment_w": 0.151123046875, + "loss_sft": 0.4953685887157917, + "loss_total_v6": 0.6463543251156807, + "step": 602 + }, + { + "epoch": 0.7241068748123687, + "grad_norm": 12.71673583984375, + "learning_rate": 2.918821669182372e-06, + "loss": 9.951238632202148, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.477787334471941, + "loss_total_v6": 0.6219523623585701, + "step": 603 + }, + { + "epoch": 0.7253077154007805, + "grad_norm": 13.969202995300293, + "learning_rate": 2.8962840883559724e-06, + "loss": 10.155558586120605, + "loss_alignment": 0.29296875, + "loss_alignment_w": 0.146484375, + "loss_sft": 0.48819226026535034, + "loss_total_v6": 0.6347224116325378, + "step": 604 + }, + { + "epoch": 0.7265085559891924, + "grad_norm": 12.905563354492188, + "learning_rate": 2.87379832527651e-06, + "loss": 9.998880386352539, + "loss_alignment": 0.2706298828125, + "loss_alignment_w": 0.13531494140625, + "loss_sft": 0.4901796616613865, + "loss_total_v6": 0.6249300390481949, + "step": 605 + }, + { + "epoch": 0.7277093965776044, + "grad_norm": 13.518686294555664, + "learning_rate": 2.8513649338027853e-06, + "loss": 10.164752960205078, + "loss_alignment": 0.29541015625, + "loss_alignment_w": 0.147705078125, + "loss_sft": 0.4875919707119465, + "loss_total_v6": 0.635297067463398, + "step": 606 + }, + { + "epoch": 0.7289102371660162, + "grad_norm": 12.680583953857422, + "learning_rate": 2.8289844665036137e-06, + "loss": 10.26591682434082, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.4953184798359871, + "loss_total_v6": 0.6416197642683983, + "step": 607 + }, + { + "epoch": 0.7301110777544281, + "grad_norm": 13.457469940185547, + "learning_rate": 2.8066574746442044e-06, + "loss": 10.169246673583984, + "loss_alignment": 0.281494140625, + "loss_alignment_w": 0.1407470703125, + "loss_sft": 0.49458667635917664, + "loss_total_v6": 0.6355778574943542, + "step": 608 + }, + { + "epoch": 0.73131191834284, + "grad_norm": 12.507440567016602, + "learning_rate": 2.7843845081725814e-06, + "loss": 10.089159965515137, + "loss_alignment": 0.281494140625, + "loss_alignment_w": 0.1407470703125, + "loss_sft": 0.48947447538375854, + "loss_total_v6": 0.630572497844696, + "step": 609 + }, + { + "epoch": 0.7325127589312519, + "grad_norm": 13.810867309570312, + "learning_rate": 2.762166115706045e-06, + "loss": 10.344322204589844, + "loss_alignment": 0.300048828125, + "loss_alignment_w": 0.1500244140625, + "loss_sft": 0.49629734829068184, + "loss_total_v6": 0.6465201079845428, + "step": 610 + }, + { + "epoch": 0.7337135995196638, + "grad_norm": 13.039945602416992, + "learning_rate": 2.7400028445176512e-06, + "loss": 10.11867904663086, + "loss_alignment": 0.296630859375, + "loss_alignment_w": 0.1483154296875, + "loss_sft": 0.48422401770949364, + "loss_total_v6": 0.6324174106121063, + "step": 611 + }, + { + "epoch": 0.7349144401080756, + "grad_norm": 13.638359069824219, + "learning_rate": 2.717895240522737e-06, + "loss": 9.92306137084961, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4771859459578991, + "loss_total_v6": 0.6201913207769394, + "step": 612 + }, + { + "epoch": 0.7361152806964876, + "grad_norm": 13.091618537902832, + "learning_rate": 2.695843848265467e-06, + "loss": 10.231025695800781, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.49318360164761543, + "loss_total_v6": 0.6394390985369682, + "step": 613 + }, + { + "epoch": 0.7373161212848994, + "grad_norm": 12.795111656188965, + "learning_rate": 2.6738492109054305e-06, + "loss": 10.252845764160156, + "loss_alignment": 0.28173828125, + "loss_alignment_w": 0.140869140625, + "loss_sft": 0.49979640915989876, + "loss_total_v6": 0.6408028602600098, + "step": 614 + }, + { + "epoch": 0.7385169618733113, + "grad_norm": 12.945899963378906, + "learning_rate": 2.6519118702042498e-06, + "loss": 9.916860580444336, + "loss_alignment": 0.279541015625, + "loss_alignment_w": 0.1397705078125, + "loss_sft": 0.4796822816133499, + "loss_total_v6": 0.6198037266731262, + "step": 615 + }, + { + "epoch": 0.7397178024617233, + "grad_norm": 12.159116744995117, + "learning_rate": 2.6300323665122496e-06, + "loss": 10.224992752075195, + "loss_alignment": 0.281982421875, + "loss_alignment_w": 0.1409912109375, + "loss_sft": 0.4977046251296997, + "loss_total_v6": 0.6390620470046997, + "step": 616 + }, + { + "epoch": 0.7409186430501351, + "grad_norm": 13.498726844787598, + "learning_rate": 2.6082112387551334e-06, + "loss": 10.189699172973633, + "loss_alignment": 0.294677734375, + "loss_alignment_w": 0.1473388671875, + "loss_sft": 0.48950202763080597, + "loss_total_v6": 0.6368561536073685, + "step": 617 + }, + { + "epoch": 0.742119483638547, + "grad_norm": 13.073212623596191, + "learning_rate": 2.586449024420724e-06, + "loss": 10.203340530395508, + "loss_alignment": 0.272216796875, + "loss_alignment_w": 0.1361083984375, + "loss_sft": 0.5015241019427776, + "loss_total_v6": 0.6377087831497192, + "step": 618 + }, + { + "epoch": 0.7433203242269588, + "grad_norm": 12.918984413146973, + "learning_rate": 2.5647462595457073e-06, + "loss": 10.040000915527344, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.4802069701254368, + "loss_total_v6": 0.6275000348687172, + "step": 619 + }, + { + "epoch": 0.7445211648153708, + "grad_norm": 12.050413131713867, + "learning_rate": 2.5431034787024478e-06, + "loss": 10.191502571105957, + "loss_alignment": 0.279541015625, + "loss_alignment_w": 0.1397705078125, + "loss_sft": 0.4974425509572029, + "loss_total_v6": 0.6369689255952835, + "step": 620 + }, + { + "epoch": 0.7457220054037826, + "grad_norm": 15.005857467651367, + "learning_rate": 2.5215212149858002e-06, + "loss": 9.962251663208008, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.4789639674127102, + "loss_total_v6": 0.6226407438516617, + "step": 621 + }, + { + "epoch": 0.7469228459921945, + "grad_norm": 13.53681755065918, + "learning_rate": 2.5000000000000015e-06, + "loss": 10.31345272064209, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.4982284605503082, + "loss_total_v6": 0.64459078758955, + "step": 622 + }, + { + "epoch": 0.7481236865806065, + "grad_norm": 12.825779914855957, + "learning_rate": 2.4785403638455537e-06, + "loss": 9.814971923828125, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.4670886844396591, + "loss_total_v6": 0.613435722887516, + "step": 623 + }, + { + "epoch": 0.7493245271690183, + "grad_norm": 12.220579147338867, + "learning_rate": 2.4571428351061872e-06, + "loss": 10.228473663330078, + "loss_alignment": 0.281982421875, + "loss_alignment_w": 0.1409912109375, + "loss_sft": 0.4982578307390213, + "loss_total_v6": 0.6392795667052269, + "step": 624 + }, + { + "epoch": 0.7505253677574302, + "grad_norm": 14.498781204223633, + "learning_rate": 2.4358079408358238e-06, + "loss": 10.152268409729004, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.48842909559607506, + "loss_total_v6": 0.6345167383551598, + "step": 625 + }, + { + "epoch": 0.7517262083458421, + "grad_norm": 12.66677474975586, + "learning_rate": 2.4145362065456097e-06, + "loss": 10.307714462280273, + "loss_alignment": 0.298583984375, + "loss_alignment_w": 0.1492919921875, + "loss_sft": 0.4953215941786766, + "loss_total_v6": 0.6442321166396141, + "step": 626 + }, + { + "epoch": 0.752927048934254, + "grad_norm": 14.155133247375488, + "learning_rate": 2.3933281561909565e-06, + "loss": 10.00169563293457, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.48194804415106773, + "loss_total_v6": 0.6251059994101524, + "step": 627 + }, + { + "epoch": 0.7541278895226659, + "grad_norm": 13.291533470153809, + "learning_rate": 2.3721843121586507e-06, + "loss": 10.059614181518555, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.4810207523405552, + "loss_total_v6": 0.6287258490920067, + "step": 628 + }, + { + "epoch": 0.7553287301110777, + "grad_norm": 14.023688316345215, + "learning_rate": 2.3511051952539703e-06, + "loss": 10.036150932312012, + "loss_alignment": 0.2928466796875, + "loss_alignment_w": 0.14642333984375, + "loss_sft": 0.4805767126381397, + "loss_total_v6": 0.6272594407200813, + "step": 629 + }, + { + "epoch": 0.7565295706994897, + "grad_norm": 13.762120246887207, + "learning_rate": 2.3300913246878736e-06, + "loss": 10.241943359375, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.4987335316836834, + "loss_total_v6": 0.6401214599609375, + "step": 630 + }, + { + "epoch": 0.7577304112879015, + "grad_norm": 12.854844093322754, + "learning_rate": 2.309143218064193e-06, + "loss": 10.0477933883667, + "loss_alignment": 0.2926025390625, + "loss_alignment_w": 0.14630126953125, + "loss_sft": 0.4814569838345051, + "loss_total_v6": 0.6279871091246605, + "step": 631 + }, + { + "epoch": 0.7589312518763134, + "grad_norm": 12.611706733703613, + "learning_rate": 2.288261391366901e-06, + "loss": 10.03469467163086, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.4820268154144287, + "loss_total_v6": 0.6271684169769287, + "step": 632 + }, + { + "epoch": 0.7601320924647254, + "grad_norm": 12.59184455871582, + "learning_rate": 2.2674463589473927e-06, + "loss": 10.043264389038086, + "loss_alignment": 0.2781982421875, + "loss_alignment_w": 0.13909912109375, + "loss_sft": 0.48866596072912216, + "loss_total_v6": 0.6277040094137192, + "step": 633 + }, + { + "epoch": 0.7613329330531372, + "grad_norm": 11.965803146362305, + "learning_rate": 2.246698633511813e-06, + "loss": 10.081955909729004, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.48290547728538513, + "loss_total_v6": 0.6301222741603851, + "step": 634 + }, + { + "epoch": 0.7625337736415491, + "grad_norm": 12.129365921020508, + "learning_rate": 2.22601872610844e-06, + "loss": 10.296886444091797, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.4966896064579487, + "loss_total_v6": 0.6435554549098015, + "step": 635 + }, + { + "epoch": 0.7637346142299609, + "grad_norm": 12.38943099975586, + "learning_rate": 2.2054071461150817e-06, + "loss": 10.23069953918457, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.49555886536836624, + "loss_total_v6": 0.639418713748455, + "step": 636 + }, + { + "epoch": 0.7649354548183729, + "grad_norm": 11.845989227294922, + "learning_rate": 2.1848644012265457e-06, + "loss": 10.051528930664062, + "loss_alignment": 0.2822265625, + "loss_alignment_w": 0.14111328125, + "loss_sft": 0.4872904159128666, + "loss_total_v6": 0.6282205805182457, + "step": 637 + }, + { + "epoch": 0.7661362954067847, + "grad_norm": 12.58675479888916, + "learning_rate": 2.1643909974421167e-06, + "loss": 10.030183792114258, + "loss_alignment": 0.27685546875, + "loss_alignment_w": 0.138427734375, + "loss_sft": 0.48842816799879074, + "loss_total_v6": 0.6268864348530769, + "step": 638 + }, + { + "epoch": 0.7673371359951966, + "grad_norm": 11.929073333740234, + "learning_rate": 2.143987439053111e-06, + "loss": 10.058064460754395, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.48144276440143585, + "loss_total_v6": 0.6286290362477303, + "step": 639 + }, + { + "epoch": 0.7685379765836086, + "grad_norm": 13.176642417907715, + "learning_rate": 2.123654228630437e-06, + "loss": 10.071150779724121, + "loss_alignment": 0.2880859375, + "loss_alignment_w": 0.14404296875, + "loss_sft": 0.48590750247240067, + "loss_total_v6": 0.6294469311833382, + "step": 640 + }, + { + "epoch": 0.7697388171720204, + "grad_norm": 12.3876314163208, + "learning_rate": 2.1033918670122315e-06, + "loss": 10.048162460327148, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.483448326587677, + "loss_total_v6": 0.628010131418705, + "step": 641 + }, + { + "epoch": 0.7709396577604323, + "grad_norm": 12.359309196472168, + "learning_rate": 2.0832008532915117e-06, + "loss": 9.989266395568848, + "loss_alignment": 0.281494140625, + "loss_alignment_w": 0.1407470703125, + "loss_sft": 0.4833226948976517, + "loss_total_v6": 0.6243291646242142, + "step": 642 + }, + { + "epoch": 0.7721404983488442, + "grad_norm": 12.69482421875, + "learning_rate": 2.063081684803892e-06, + "loss": 10.281545639038086, + "loss_alignment": 0.282470703125, + "loss_alignment_w": 0.1412353515625, + "loss_sft": 0.5012849383056164, + "loss_total_v6": 0.642596572637558, + "step": 643 + }, + { + "epoch": 0.7733413389372561, + "grad_norm": 12.054780960083008, + "learning_rate": 2.043034857115323e-06, + "loss": 10.132749557495117, + "loss_alignment": 0.2825927734375, + "loss_alignment_w": 0.14129638671875, + "loss_sft": 0.49198516830801964, + "loss_total_v6": 0.6332968100905418, + "step": 644 + }, + { + "epoch": 0.774542179525668, + "grad_norm": 12.73280143737793, + "learning_rate": 2.023060864009896e-06, + "loss": 9.89870834350586, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.4781510606408119, + "loss_total_v6": 0.618669256567955, + "step": 645 + }, + { + "epoch": 0.7757430201140798, + "grad_norm": 13.081756591796875, + "learning_rate": 2.0031601974776686e-06, + "loss": 10.118379592895508, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.48962225019931793, + "loss_total_v6": 0.632398746907711, + "step": 646 + }, + { + "epoch": 0.7769438607024918, + "grad_norm": 12.672027587890625, + "learning_rate": 1.98333334770256e-06, + "loss": 10.401211738586426, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.5031488128006458, + "loss_total_v6": 0.6500757038593292, + "step": 647 + }, + { + "epoch": 0.7781447012909036, + "grad_norm": 13.040009498596191, + "learning_rate": 1.9635808030502617e-06, + "loss": 9.983322143554688, + "loss_alignment": 0.2783203125, + "loss_alignment_w": 0.13916015625, + "loss_sft": 0.48493480309844017, + "loss_total_v6": 0.6239576414227486, + "step": 648 + }, + { + "epoch": 0.7793455418793155, + "grad_norm": 12.8660888671875, + "learning_rate": 1.9439030500562243e-06, + "loss": 9.939643859863281, + "loss_alignment": 0.2650146484375, + "loss_alignment_w": 0.13250732421875, + "loss_sft": 0.4887966811656952, + "loss_total_v6": 0.6212277114391327, + "step": 649 + }, + { + "epoch": 0.7805463824677275, + "grad_norm": 13.348841667175293, + "learning_rate": 1.9243005734136572e-06, + "loss": 10.176505088806152, + "loss_alignment": 0.2869873046875, + "loss_alignment_w": 0.14349365234375, + "loss_sft": 0.49276677891612053, + "loss_total_v6": 0.6360315456986427, + "step": 650 + }, + { + "epoch": 0.7817472230561393, + "grad_norm": 12.332558631896973, + "learning_rate": 1.9047738559616046e-06, + "loss": 10.043840408325195, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4840938076376915, + "loss_total_v6": 0.6277400478720665, + "step": 651 + }, + { + "epoch": 0.7829480636445512, + "grad_norm": 13.035482406616211, + "learning_rate": 1.8853233786730395e-06, + "loss": 9.972091674804688, + "loss_alignment": 0.285400390625, + "loss_alignment_w": 0.1427001953125, + "loss_sft": 0.48031139746308327, + "loss_total_v6": 0.6232557371258736, + "step": 652 + }, + { + "epoch": 0.7841489042329631, + "grad_norm": 13.238353729248047, + "learning_rate": 1.8659496206430306e-06, + "loss": 10.243101119995117, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.4958304166793823, + "loss_total_v6": 0.6401938274502754, + "step": 653 + }, + { + "epoch": 0.785349744821375, + "grad_norm": 12.244568824768066, + "learning_rate": 1.846653059076925e-06, + "loss": 10.096550941467285, + "loss_alignment": 0.281982421875, + "loss_alignment_w": 0.1409912109375, + "loss_sft": 0.4899974390864372, + "loss_total_v6": 0.6310344263911247, + "step": 654 + }, + { + "epoch": 0.7865505854097868, + "grad_norm": 13.463228225708008, + "learning_rate": 1.8274341692786108e-06, + "loss": 9.823810577392578, + "loss_alignment": 0.280029296875, + "loss_alignment_w": 0.1400146484375, + "loss_sft": 0.47400403767824173, + "loss_total_v6": 0.6139881536364555, + "step": 655 + }, + { + "epoch": 0.7877514259981987, + "grad_norm": 13.47460651397705, + "learning_rate": 1.8082934246387935e-06, + "loss": 9.827214241027832, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.4712718203663826, + "loss_total_v6": 0.6142009124159813, + "step": 656 + }, + { + "epoch": 0.7889522665866107, + "grad_norm": 12.583433151245117, + "learning_rate": 1.7892312966233489e-06, + "loss": 10.538660049438477, + "loss_alignment": 0.29931640625, + "loss_alignment_w": 0.149658203125, + "loss_sft": 0.5089164823293686, + "loss_total_v6": 0.6586662381887436, + "step": 657 + }, + { + "epoch": 0.7901531071750225, + "grad_norm": 13.250965118408203, + "learning_rate": 1.7702482547617068e-06, + "loss": 9.977141380310059, + "loss_alignment": 0.2781982421875, + "loss_alignment_w": 0.13909912109375, + "loss_sft": 0.48444170877337456, + "loss_total_v6": 0.6235713586211205, + "step": 658 + }, + { + "epoch": 0.7913539477634344, + "grad_norm": 12.420228958129883, + "learning_rate": 1.7513447666352752e-06, + "loss": 9.879204750061035, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.4734225943684578, + "loss_total_v6": 0.6174502819776535, + "step": 659 + }, + { + "epoch": 0.7925547883518463, + "grad_norm": 12.61929988861084, + "learning_rate": 1.7325212978659423e-06, + "loss": 9.975431442260742, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4803980477154255, + "loss_total_v6": 0.6234644502401352, + "step": 660 + }, + { + "epoch": 0.7937556289402582, + "grad_norm": 12.154006958007812, + "learning_rate": 1.7137783121045864e-06, + "loss": 10.105705261230469, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.4877925328910351, + "loss_total_v6": 0.6316066235303879, + "step": 661 + }, + { + "epoch": 0.79495646952867, + "grad_norm": 12.344738006591797, + "learning_rate": 1.695116271019676e-06, + "loss": 10.235828399658203, + "loss_alignment": 0.28369140625, + "loss_alignment_w": 0.141845703125, + "loss_sft": 0.4979240857064724, + "loss_total_v6": 0.6397392824292183, + "step": 662 + }, + { + "epoch": 0.7961573101170819, + "grad_norm": 12.280138969421387, + "learning_rate": 1.6765356342858796e-06, + "loss": 10.024679183959961, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.48251475766301155, + "loss_total_v6": 0.6265424713492393, + "step": 663 + }, + { + "epoch": 0.7973581507054939, + "grad_norm": 12.432888984680176, + "learning_rate": 1.6580368595727586e-06, + "loss": 9.951620101928711, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.47654470801353455, + "loss_total_v6": 0.6219762340188026, + "step": 664 + }, + { + "epoch": 0.7985589912939057, + "grad_norm": 12.911731719970703, + "learning_rate": 1.6396204025334817e-06, + "loss": 10.099747657775879, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.4841547794640064, + "loss_total_v6": 0.6312342137098312, + "step": 665 + }, + { + "epoch": 0.7997598318823176, + "grad_norm": 12.39340877532959, + "learning_rate": 1.621286716793612e-06, + "loss": 9.913644790649414, + "loss_alignment": 0.27734375, + "loss_alignment_w": 0.138671875, + "loss_sft": 0.4810682386159897, + "loss_total_v6": 0.6196027919650078, + "step": 666 + }, + { + "epoch": 0.8009606724707296, + "grad_norm": 12.954190254211426, + "learning_rate": 1.6030362539399235e-06, + "loss": 10.26510238647461, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.4925515316426754, + "loss_total_v6": 0.6415688619017601, + "step": 667 + }, + { + "epoch": 0.8021615130591414, + "grad_norm": 12.205570220947266, + "learning_rate": 1.58486946350929e-06, + "loss": 10.09154224395752, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.48614437133073807, + "loss_total_v6": 0.6307213827967644, + "step": 668 + }, + { + "epoch": 0.8033623536475533, + "grad_norm": 12.613666534423828, + "learning_rate": 1.566786792977597e-06, + "loss": 9.943534851074219, + "loss_alignment": 0.277587890625, + "loss_alignment_w": 0.1387939453125, + "loss_sft": 0.48269228637218475, + "loss_total_v6": 0.6214709505438805, + "step": 669 + }, + { + "epoch": 0.8045631942359652, + "grad_norm": 12.436336517333984, + "learning_rate": 1.5487886877487352e-06, + "loss": 10.192336082458496, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.4904451034963131, + "loss_total_v6": 0.6370210275053978, + "step": 670 + }, + { + "epoch": 0.8057640348243771, + "grad_norm": 12.935964584350586, + "learning_rate": 1.5308755911436169e-06, + "loss": 10.0994234085083, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4883764460682869, + "loss_total_v6": 0.6312139630317688, + "step": 671 + }, + { + "epoch": 0.8069648754127889, + "grad_norm": 12.023333549499512, + "learning_rate": 1.5130479443892665e-06, + "loss": 10.381909370422363, + "loss_alignment": 0.29345703125, + "loss_alignment_w": 0.146728515625, + "loss_sft": 0.5019882284104824, + "loss_total_v6": 0.6488693282008171, + "step": 672 + }, + { + "epoch": 0.8081657160012008, + "grad_norm": 12.398541450500488, + "learning_rate": 1.4953061866079421e-06, + "loss": 10.125155448913574, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4898473992943764, + "loss_total_v6": 0.6328222379088402, + "step": 673 + }, + { + "epoch": 0.8093665565896128, + "grad_norm": 12.503318786621094, + "learning_rate": 1.4776507548063319e-06, + "loss": 10.291694641113281, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.4983029179275036, + "loss_total_v6": 0.6432309001684189, + "step": 674 + }, + { + "epoch": 0.8105673971780246, + "grad_norm": 11.784712791442871, + "learning_rate": 1.460082083864774e-06, + "loss": 10.203242301940918, + "loss_alignment": 0.298095703125, + "loss_alignment_w": 0.1490478515625, + "loss_sft": 0.4889294169843197, + "loss_total_v6": 0.6377026364207268, + "step": 675 + }, + { + "epoch": 0.8117682377664365, + "grad_norm": 13.713647842407227, + "learning_rate": 1.442600606526563e-06, + "loss": 10.097785949707031, + "loss_alignment": 0.2841796875, + "loss_alignment_w": 0.14208984375, + "loss_sft": 0.48915914073586464, + "loss_total_v6": 0.6311116367578506, + "step": 676 + }, + { + "epoch": 0.8129690783548484, + "grad_norm": 13.745400428771973, + "learning_rate": 1.4252067533872732e-06, + "loss": 10.053775787353516, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.48331092298030853, + "loss_total_v6": 0.6283609718084335, + "step": 677 + }, + { + "epoch": 0.8141699189432603, + "grad_norm": 11.861856460571289, + "learning_rate": 1.407900952884167e-06, + "loss": 9.915769577026367, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.47094718366861343, + "loss_total_v6": 0.6197356283664703, + "step": 678 + }, + { + "epoch": 0.8153707595316722, + "grad_norm": 13.27472972869873, + "learning_rate": 1.3906836312856304e-06, + "loss": 9.852130889892578, + "loss_alignment": 0.28271484375, + "loss_alignment_w": 0.141357421875, + "loss_sft": 0.47464485466480255, + "loss_total_v6": 0.6157581433653831, + "step": 679 + }, + { + "epoch": 0.816571600120084, + "grad_norm": 13.040146827697754, + "learning_rate": 1.3735552126806823e-06, + "loss": 9.95979118347168, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.47966468334198, + "loss_total_v6": 0.6224869638681412, + "step": 680 + }, + { + "epoch": 0.817772440708496, + "grad_norm": 12.765035629272461, + "learning_rate": 1.3565161189685217e-06, + "loss": 10.196415901184082, + "loss_alignment": 0.2808837890625, + "loss_alignment_w": 0.14044189453125, + "loss_sft": 0.49678830057382584, + "loss_total_v6": 0.6372760161757469, + "step": 681 + }, + { + "epoch": 0.8189732812969078, + "grad_norm": 12.966492652893066, + "learning_rate": 1.3395667698481397e-06, + "loss": 9.964202880859375, + "loss_alignment": 0.291259765625, + "loss_alignment_w": 0.1456298828125, + "loss_sft": 0.47723955288529396, + "loss_total_v6": 0.6227626353502274, + "step": 682 + }, + { + "epoch": 0.8201741218853197, + "grad_norm": 12.564489364624023, + "learning_rate": 1.3227075828079833e-06, + "loss": 10.106225967407227, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.48724517971277237, + "loss_total_v6": 0.6316390857100487, + "step": 683 + }, + { + "epoch": 0.8213749624737317, + "grad_norm": 12.146851539611816, + "learning_rate": 1.3059389731156635e-06, + "loss": 9.956363677978516, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.47554417699575424, + "loss_total_v6": 0.622272714972496, + "step": 684 + }, + { + "epoch": 0.8225758030621435, + "grad_norm": 12.518815994262695, + "learning_rate": 1.289261353807738e-06, + "loss": 10.003316879272461, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.4792417399585247, + "loss_total_v6": 0.6252073273062706, + "step": 685 + }, + { + "epoch": 0.8237766436505554, + "grad_norm": 12.445732116699219, + "learning_rate": 1.2726751356795275e-06, + "loss": 9.75027847290039, + "loss_alignment": 0.2750244140625, + "loss_alignment_w": 0.13751220703125, + "loss_sft": 0.4717581048607826, + "loss_total_v6": 0.6093923896551132, + "step": 686 + }, + { + "epoch": 0.8249774842389673, + "grad_norm": 12.650611877441406, + "learning_rate": 1.2561807272750053e-06, + "loss": 10.119905471801758, + "loss_alignment": 0.2919921875, + "loss_alignment_w": 0.14599609375, + "loss_sft": 0.48672693222761154, + "loss_total_v6": 0.632494144141674, + "step": 687 + }, + { + "epoch": 0.8261783248273792, + "grad_norm": 11.917102813720703, + "learning_rate": 1.2397785348767272e-06, + "loss": 10.197155952453613, + "loss_alignment": 0.297607421875, + "loss_alignment_w": 0.1488037109375, + "loss_sft": 0.48822860419750214, + "loss_total_v6": 0.6373222395777702, + "step": 688 + }, + { + "epoch": 0.827379165415791, + "grad_norm": 12.405603408813477, + "learning_rate": 1.2234689624958307e-06, + "loss": 9.801873207092285, + "loss_alignment": 0.2801513671875, + "loss_alignment_w": 0.14007568359375, + "loss_sft": 0.4725413806736469, + "loss_total_v6": 0.6126170754432678, + "step": 689 + }, + { + "epoch": 0.8285800060042029, + "grad_norm": 16.960424423217773, + "learning_rate": 1.2072524118620766e-06, + "loss": 9.97866439819336, + "loss_alignment": 0.2850341796875, + "loss_alignment_w": 0.14251708984375, + "loss_sft": 0.48128674179315567, + "loss_total_v6": 0.6236665174365044, + "step": 690 + }, + { + "epoch": 0.8297808465926149, + "grad_norm": 12.37277603149414, + "learning_rate": 1.191129282413962e-06, + "loss": 10.11089038848877, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.4886964447796345, + "loss_total_v6": 0.6319306716322899, + "step": 691 + }, + { + "epoch": 0.8309816871810267, + "grad_norm": 11.887831687927246, + "learning_rate": 1.1750999712888717e-06, + "loss": 10.106128692626953, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.48839881271123886, + "loss_total_v6": 0.6316330507397652, + "step": 692 + }, + { + "epoch": 0.8321825277694386, + "grad_norm": 13.696721076965332, + "learning_rate": 1.1591648733133071e-06, + "loss": 9.976297378540039, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.47575248777866364, + "loss_total_v6": 0.623518593609333, + "step": 693 + }, + { + "epoch": 0.8333833683578505, + "grad_norm": 12.029386520385742, + "learning_rate": 1.14332438099315e-06, + "loss": 10.061089515686035, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4860873706638813, + "loss_total_v6": 0.6288180872797966, + "step": 694 + }, + { + "epoch": 0.8345842089462624, + "grad_norm": 12.505904197692871, + "learning_rate": 1.1275788845040054e-06, + "loss": 10.283845901489258, + "loss_alignment": 0.299072265625, + "loss_alignment_w": 0.1495361328125, + "loss_sft": 0.4933415465056896, + "loss_total_v6": 0.6427403166890144, + "step": 695 + }, + { + "epoch": 0.8357850495346743, + "grad_norm": 12.378521919250488, + "learning_rate": 1.111928771681579e-06, + "loss": 9.85534954071045, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.4723588339984417, + "loss_total_v6": 0.6159593239426613, + "step": 696 + }, + { + "epoch": 0.8369858901230861, + "grad_norm": 12.121461868286133, + "learning_rate": 1.0963744280121375e-06, + "loss": 9.925413131713867, + "loss_alignment": 0.2860107421875, + "loss_alignment_w": 0.14300537109375, + "loss_sft": 0.4773482456803322, + "loss_total_v6": 0.6203383207321167, + "step": 697 + }, + { + "epoch": 0.8381867307114981, + "grad_norm": 12.20280647277832, + "learning_rate": 1.0809162366229996e-06, + "loss": 9.92149543762207, + "loss_alignment": 0.285400390625, + "loss_alignment_w": 0.1427001953125, + "loss_sft": 0.4769812151789665, + "loss_total_v6": 0.6200934201478958, + "step": 698 + }, + { + "epoch": 0.8393875712999099, + "grad_norm": 11.952805519104004, + "learning_rate": 1.065554578273113e-06, + "loss": 10.267139434814453, + "loss_alignment": 0.290283203125, + "loss_alignment_w": 0.1451416015625, + "loss_sft": 0.49676819518208504, + "loss_total_v6": 0.6416961997747421, + "step": 699 + }, + { + "epoch": 0.8405884118883218, + "grad_norm": 12.283554077148438, + "learning_rate": 1.0502898313436626e-06, + "loss": 10.238393783569336, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.4949716553092003, + "loss_total_v6": 0.6398996114730835, + "step": 700 + }, + { + "epoch": 0.8405884118883218, + "eval_loss": 0.58695387840271, + "eval_loss_alignment": 0.2812852125901442, + "eval_loss_alignment_w": 0.1406426062950721, + "eval_loss_sft": 0.44631129155795163, + "eval_loss_total_v6": 0.586953897602283, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.0016, + "eval_samples_per_second": 24.761, + "eval_steps_per_second": 3.104, + "step": 700 + }, + { + "epoch": 0.8417892524767338, + "grad_norm": 11.933323860168457, + "learning_rate": 1.0351223718287612e-06, + "loss": 10.116357803344727, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.4870239458978176, + "loss_total_v6": 0.6322723478078842, + "step": 701 + }, + { + "epoch": 0.8429900930651456, + "grad_norm": 12.201210975646973, + "learning_rate": 1.020052573326179e-06, + "loss": 9.569036483764648, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4545100964605808, + "loss_total_v6": 0.5980647727847099, + "step": 702 + }, + { + "epoch": 0.8441909336535575, + "grad_norm": 12.288529396057129, + "learning_rate": 1.005080807028151e-06, + "loss": 9.940888404846191, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.47408876568078995, + "loss_total_v6": 0.6213055327534676, + "step": 703 + }, + { + "epoch": 0.8453917742419694, + "grad_norm": 14.78847599029541, + "learning_rate": 9.902074417122233e-07, + "loss": 9.993820190429688, + "loss_alignment": 0.29541015625, + "loss_alignment_w": 0.147705078125, + "loss_sft": 0.47687821835279465, + "loss_total_v6": 0.6246137768030167, + "step": 704 + }, + { + "epoch": 0.8465926148303813, + "grad_norm": 12.079480171203613, + "learning_rate": 9.754328437321802e-07, + "loss": 10.162857055664062, + "loss_alignment": 0.29248046875, + "loss_alignment_w": 0.146240234375, + "loss_sft": 0.48875517770648, + "loss_total_v6": 0.6351785436272621, + "step": 705 + }, + { + "epoch": 0.8477934554187931, + "grad_norm": 12.02500057220459, + "learning_rate": 9.607573770090102e-07, + "loss": 9.975421905517578, + "loss_alignment": 0.289794921875, + "loss_alignment_w": 0.1448974609375, + "loss_sft": 0.4785206615924835, + "loss_total_v6": 0.623463898897171, + "step": 706 + }, + { + "epoch": 0.848994296007205, + "grad_norm": 12.242568016052246, + "learning_rate": 9.461814030219518e-07, + "loss": 9.982665061950684, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.4770354628562927, + "loss_total_v6": 0.6239165738224983, + "step": 707 + }, + { + "epoch": 0.850195136595617, + "grad_norm": 12.151057243347168, + "learning_rate": 9.317052807995796e-07, + "loss": 10.034599304199219, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.48188356310129166, + "loss_total_v6": 0.6271625012159348, + "step": 708 + }, + { + "epoch": 0.8513959771840288, + "grad_norm": 12.760107040405273, + "learning_rate": 9.173293669109728e-07, + "loss": 9.901786804199219, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47396423295140266, + "loss_total_v6": 0.6188616752624512, + "step": 709 + }, + { + "epoch": 0.8525968177724407, + "grad_norm": 12.311018943786621, + "learning_rate": 9.030540154569189e-07, + "loss": 9.734725952148438, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.46535400301218033, + "loss_total_v6": 0.6084204018115997, + "step": 710 + }, + { + "epoch": 0.8537976583608526, + "grad_norm": 11.90987491607666, + "learning_rate": 8.888795780612047e-07, + "loss": 10.101485252380371, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4889325313270092, + "loss_total_v6": 0.6313428282737732, + "step": 711 + }, + { + "epoch": 0.8549984989492645, + "grad_norm": 12.273737907409668, + "learning_rate": 8.748064038619436e-07, + "loss": 9.727742195129395, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.46563461795449257, + "loss_total_v6": 0.6079838648438454, + "step": 712 + }, + { + "epoch": 0.8561993395376764, + "grad_norm": 13.309226036071777, + "learning_rate": 8.608348395029859e-07, + "loss": 9.947473526000977, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.47552259638905525, + "loss_total_v6": 0.6217170730233192, + "step": 713 + }, + { + "epoch": 0.8574001801260882, + "grad_norm": 11.901158332824707, + "learning_rate": 8.46965229125371e-07, + "loss": 10.097033500671387, + "loss_alignment": 0.2900390625, + "loss_alignment_w": 0.14501953125, + "loss_sft": 0.48586196824908257, + "loss_total_v6": 0.6310646161437035, + "step": 714 + }, + { + "epoch": 0.8586010207145002, + "grad_norm": 12.188934326171875, + "learning_rate": 8.33197914358862e-07, + "loss": 10.191509246826172, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.48950834944844246, + "loss_total_v6": 0.6369692832231522, + "step": 715 + }, + { + "epoch": 0.859801861302912, + "grad_norm": 12.035649299621582, + "learning_rate": 8.195332343135171e-07, + "loss": 10.070125579833984, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.48700301721692085, + "loss_total_v6": 0.6293827965855598, + "step": 716 + }, + { + "epoch": 0.8610027018913239, + "grad_norm": 12.420633316040039, + "learning_rate": 8.059715255713508e-07, + "loss": 9.946662902832031, + "loss_alignment": 0.2861328125, + "loss_alignment_w": 0.14306640625, + "loss_sft": 0.47873736917972565, + "loss_total_v6": 0.6216664463281631, + "step": 717 + }, + { + "epoch": 0.8622035424797359, + "grad_norm": 12.154691696166992, + "learning_rate": 7.925131221780297e-07, + "loss": 9.80339241027832, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.47013383731245995, + "loss_total_v6": 0.6127119809389114, + "step": 718 + }, + { + "epoch": 0.8634043830681477, + "grad_norm": 12.00091552734375, + "learning_rate": 7.791583556346577e-07, + "loss": 9.88347339630127, + "loss_alignment": 0.283203125, + "loss_alignment_w": 0.1416015625, + "loss_sft": 0.47616131976246834, + "loss_total_v6": 0.6177170798182487, + "step": 719 + }, + { + "epoch": 0.8646052236565596, + "grad_norm": 12.45433521270752, + "learning_rate": 7.659075548896e-07, + "loss": 10.286280632019043, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.49437878280878067, + "loss_total_v6": 0.6428925693035126, + "step": 720 + }, + { + "epoch": 0.8658060642449715, + "grad_norm": 12.105317115783691, + "learning_rate": 7.52761046330388e-07, + "loss": 10.005892753601074, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.47850245982408524, + "loss_total_v6": 0.6253682971000671, + "step": 721 + }, + { + "epoch": 0.8670069048333834, + "grad_norm": 11.663182258605957, + "learning_rate": 7.397191537756727e-07, + "loss": 9.903463363647461, + "loss_alignment": 0.286376953125, + "loss_alignment_w": 0.1431884765625, + "loss_sft": 0.4758695624768734, + "loss_total_v6": 0.6189665049314499, + "step": 722 + }, + { + "epoch": 0.8682077454217952, + "grad_norm": 12.916326522827148, + "learning_rate": 7.267821984672574e-07, + "loss": 10.164314270019531, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.491714958101511, + "loss_total_v6": 0.6352696493268013, + "step": 723 + }, + { + "epoch": 0.8694085860102071, + "grad_norm": 12.092696189880371, + "learning_rate": 7.139504990621754e-07, + "loss": 10.340091705322266, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.4995425194501877, + "loss_total_v6": 0.6462557688355446, + "step": 724 + }, + { + "epoch": 0.8706094265986191, + "grad_norm": 12.064631462097168, + "learning_rate": 7.012243716248506e-07, + "loss": 10.123165130615234, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.4902874566614628, + "loss_total_v6": 0.6326977759599686, + "step": 725 + }, + { + "epoch": 0.8718102671870309, + "grad_norm": 11.965213775634766, + "learning_rate": 6.88604129619303e-07, + "loss": 10.227666854858398, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.4986041523516178, + "loss_total_v6": 0.6392291560769081, + "step": 726 + }, + { + "epoch": 0.8730111077754428, + "grad_norm": 12.674707412719727, + "learning_rate": 6.760900839014356e-07, + "loss": 10.150254249572754, + "loss_alignment": 0.2862548828125, + "loss_alignment_w": 0.14312744140625, + "loss_sft": 0.4913855232298374, + "loss_total_v6": 0.6343909054994583, + "step": 727 + }, + { + "epoch": 0.8742119483638547, + "grad_norm": 12.208616256713867, + "learning_rate": 6.636825427113719e-07, + "loss": 10.040380477905273, + "loss_alignment": 0.28759765625, + "loss_alignment_w": 0.143798828125, + "loss_sft": 0.483541876077652, + "loss_total_v6": 0.6275238171219826, + "step": 728 + }, + { + "epoch": 0.8754127889522666, + "grad_norm": 12.521137237548828, + "learning_rate": 6.513818116658671e-07, + "loss": 9.88771915435791, + "loss_alignment": 0.294921875, + "loss_alignment_w": 0.1474609375, + "loss_sft": 0.4704757444560528, + "loss_total_v6": 0.6179824396967888, + "step": 729 + }, + { + "epoch": 0.8766136295406785, + "grad_norm": 12.513144493103027, + "learning_rate": 6.391881937507771e-07, + "loss": 10.212996482849121, + "loss_alignment": 0.2841796875, + "loss_alignment_w": 0.14208984375, + "loss_sft": 0.49584098905324936, + "loss_total_v6": 0.6383122578263283, + "step": 730 + }, + { + "epoch": 0.8778144701290904, + "grad_norm": 12.953263282775879, + "learning_rate": 6.271019893136004e-07, + "loss": 9.964601516723633, + "loss_alignment": 0.287109375, + "loss_alignment_w": 0.1435546875, + "loss_sft": 0.47920234501361847, + "loss_total_v6": 0.6227875575423241, + "step": 731 + }, + { + "epoch": 0.8790153107175023, + "grad_norm": 12.38868522644043, + "learning_rate": 6.151234960560731e-07, + "loss": 10.162817001342773, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.49055328592658043, + "loss_total_v6": 0.6351760849356651, + "step": 732 + }, + { + "epoch": 0.8802161513059141, + "grad_norm": 12.983209609985352, + "learning_rate": 6.032530090268429e-07, + "loss": 9.80543327331543, + "loss_alignment": 0.279052734375, + "loss_alignment_w": 0.1395263671875, + "loss_sft": 0.47313009202480316, + "loss_total_v6": 0.6128395348787308, + "step": 733 + }, + { + "epoch": 0.881416991894326, + "grad_norm": 12.636907577514648, + "learning_rate": 5.914908206141956e-07, + "loss": 9.91826057434082, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.4717131592333317, + "loss_total_v6": 0.6198912635445595, + "step": 734 + }, + { + "epoch": 0.882617832482738, + "grad_norm": 12.076606750488281, + "learning_rate": 5.798372205388603e-07, + "loss": 10.069696426391602, + "loss_alignment": 0.292236328125, + "loss_alignment_w": 0.1461181640625, + "loss_sft": 0.4836193099617958, + "loss_total_v6": 0.6293559819459915, + "step": 735 + }, + { + "epoch": 0.8838186730711498, + "grad_norm": 12.464773178100586, + "learning_rate": 5.682924958468639e-07, + "loss": 9.98388671875, + "loss_alignment": 0.2801513671875, + "loss_alignment_w": 0.14007568359375, + "loss_sft": 0.4841308705508709, + "loss_total_v6": 0.6239929124712944, + "step": 736 + }, + { + "epoch": 0.8850195136595617, + "grad_norm": 12.576714515686035, + "learning_rate": 5.568569309024701e-07, + "loss": 10.172845840454102, + "loss_alignment": 0.285400390625, + "loss_alignment_w": 0.1427001953125, + "loss_sft": 0.49328574910759926, + "loss_total_v6": 0.6358028501272202, + "step": 737 + }, + { + "epoch": 0.8862203542479736, + "grad_norm": 12.40670108795166, + "learning_rate": 5.455308073811683e-07, + "loss": 10.05661678314209, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4839768074452877, + "loss_total_v6": 0.6285385712981224, + "step": 738 + }, + { + "epoch": 0.8874211948363855, + "grad_norm": 12.691132545471191, + "learning_rate": 5.343144042627391e-07, + "loss": 9.911696434020996, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.47537698969244957, + "loss_total_v6": 0.6194810047745705, + "step": 739 + }, + { + "epoch": 0.8886220354247973, + "grad_norm": 12.713736534118652, + "learning_rate": 5.232079978243798e-07, + "loss": 10.049959182739258, + "loss_alignment": 0.295654296875, + "loss_alignment_w": 0.1478271484375, + "loss_sft": 0.480341088026762, + "loss_total_v6": 0.6281224563717842, + "step": 740 + }, + { + "epoch": 0.8898228760132092, + "grad_norm": 13.35789680480957, + "learning_rate": 5.122118616339033e-07, + "loss": 9.885300636291504, + "loss_alignment": 0.2916259765625, + "loss_alignment_w": 0.14581298828125, + "loss_sft": 0.4717436283826828, + "loss_total_v6": 0.6178312823176384, + "step": 741 + }, + { + "epoch": 0.8910237166016212, + "grad_norm": 12.619976997375488, + "learning_rate": 5.013262665429952e-07, + "loss": 9.85289192199707, + "loss_alignment": 0.2838134765625, + "loss_alignment_w": 0.14190673828125, + "loss_sft": 0.4742957465350628, + "loss_total_v6": 0.6158057823777199, + "step": 742 + }, + { + "epoch": 0.892224557190033, + "grad_norm": 11.930037498474121, + "learning_rate": 4.905514806805456e-07, + "loss": 10.133148193359375, + "loss_alignment": 0.2916259765625, + "loss_alignment_w": 0.14581298828125, + "loss_sft": 0.48752398043870926, + "loss_total_v6": 0.6333217173814774, + "step": 743 + }, + { + "epoch": 0.8934253977784449, + "grad_norm": 13.203819274902344, + "learning_rate": 4.798877694460424e-07, + "loss": 10.18716812133789, + "loss_alignment": 0.30419921875, + "loss_alignment_w": 0.152099609375, + "loss_sft": 0.48462892323732376, + "loss_total_v6": 0.6366980075836182, + "step": 744 + }, + { + "epoch": 0.8946262383668568, + "grad_norm": 12.534483909606934, + "learning_rate": 4.6933539550303774e-07, + "loss": 9.944570541381836, + "loss_alignment": 0.30029296875, + "loss_alignment_w": 0.150146484375, + "loss_sft": 0.47138913720846176, + "loss_total_v6": 0.6215355917811394, + "step": 745 + }, + { + "epoch": 0.8958270789552687, + "grad_norm": 11.6961030960083, + "learning_rate": 4.588946187726706e-07, + "loss": 10.0322265625, + "loss_alignment": 0.29833984375, + "loss_alignment_w": 0.149169921875, + "loss_sft": 0.4779357872903347, + "loss_total_v6": 0.6270141899585724, + "step": 746 + }, + { + "epoch": 0.8970279195436806, + "grad_norm": 11.742012977600098, + "learning_rate": 4.485656964272761e-07, + "loss": 9.859698295593262, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.47255438193678856, + "loss_total_v6": 0.6162311360239983, + "step": 747 + }, + { + "epoch": 0.8982287601320925, + "grad_norm": 12.050057411193848, + "learning_rate": 4.383488828840388e-07, + "loss": 9.920145034790039, + "loss_alignment": 0.290283203125, + "loss_alignment_w": 0.1451416015625, + "loss_sft": 0.4748827628791332, + "loss_total_v6": 0.6200090795755386, + "step": 748 + }, + { + "epoch": 0.8994296007205044, + "grad_norm": 11.649482727050781, + "learning_rate": 4.282444297987359e-07, + "loss": 9.861640930175781, + "loss_alignment": 0.289794921875, + "loss_alignment_w": 0.1448974609375, + "loss_sft": 0.47195861861109734, + "loss_total_v6": 0.6163525655865669, + "step": 749 + }, + { + "epoch": 0.9006304413089162, + "grad_norm": 12.117549896240234, + "learning_rate": 4.182525860595299e-07, + "loss": 9.763935089111328, + "loss_alignment": 0.287841796875, + "loss_alignment_w": 0.1439208984375, + "loss_sft": 0.466096181422472, + "loss_total_v6": 0.6102459654211998, + "step": 750 + }, + { + "epoch": 0.9018312818973281, + "grad_norm": 12.382481575012207, + "learning_rate": 4.083735977808473e-07, + "loss": 9.973761558532715, + "loss_alignment": 0.3017578125, + "loss_alignment_w": 0.15087890625, + "loss_sft": 0.4727710895240307, + "loss_total_v6": 0.6233600601553917, + "step": 751 + }, + { + "epoch": 0.90303212248574, + "grad_norm": 11.8955078125, + "learning_rate": 3.9860770829730776e-07, + "loss": 9.936565399169922, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.4763362780213356, + "loss_total_v6": 0.6210353747010231, + "step": 752 + }, + { + "epoch": 0.9042329630741519, + "grad_norm": 12.384514808654785, + "learning_rate": 3.889551581577378e-07, + "loss": 10.023040771484375, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.4795895032584667, + "loss_total_v6": 0.6264400631189346, + "step": 753 + }, + { + "epoch": 0.9054338036625638, + "grad_norm": 12.6446533203125, + "learning_rate": 3.794161851192374e-07, + "loss": 9.783945083618164, + "loss_alignment": 0.283935546875, + "loss_alignment_w": 0.1419677734375, + "loss_sft": 0.4697118438780308, + "loss_total_v6": 0.6114965155720711, + "step": 754 + }, + { + "epoch": 0.9066346442509757, + "grad_norm": 12.137813568115234, + "learning_rate": 3.699910241413346e-07, + "loss": 10.072802543640137, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.48277588933706284, + "loss_total_v6": 0.6295501589775085, + "step": 755 + }, + { + "epoch": 0.9078354848393876, + "grad_norm": 12.871853828430176, + "learning_rate": 3.6067990738018587e-07, + "loss": 10.029464721679688, + "loss_alignment": 0.2813720703125, + "loss_alignment_w": 0.14068603515625, + "loss_sft": 0.48624705895781517, + "loss_total_v6": 0.6268415674567223, + "step": 756 + }, + { + "epoch": 0.9090363254277994, + "grad_norm": 12.269662857055664, + "learning_rate": 3.5148306418286924e-07, + "loss": 9.763185501098633, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.46670546755194664, + "loss_total_v6": 0.6101991385221481, + "step": 757 + }, + { + "epoch": 0.9102371660162113, + "grad_norm": 12.009734153747559, + "learning_rate": 3.424007210817248e-07, + "loss": 10.070610046386719, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.48129602521657944, + "loss_total_v6": 0.6294130980968475, + "step": 758 + }, + { + "epoch": 0.9114380066046233, + "grad_norm": 11.882965087890625, + "learning_rate": 3.334331017887837e-07, + "loss": 10.185598373413086, + "loss_alignment": 0.2894287109375, + "loss_alignment_w": 0.14471435546875, + "loss_sft": 0.49176348745822906, + "loss_total_v6": 0.6365998908877373, + "step": 759 + }, + { + "epoch": 0.9126388471930351, + "grad_norm": 12.259139060974121, + "learning_rate": 3.245804271902525e-07, + "loss": 10.338912963867188, + "loss_alignment": 0.30078125, + "loss_alignment_w": 0.150390625, + "loss_sft": 0.49614234268665314, + "loss_total_v6": 0.6461820155382156, + "step": 760 + }, + { + "epoch": 0.913839687781447, + "grad_norm": 12.21993350982666, + "learning_rate": 3.1584291534107247e-07, + "loss": 9.927209854125977, + "loss_alignment": 0.2760009765625, + "loss_alignment_w": 0.13800048828125, + "loss_sft": 0.48272478207945824, + "loss_total_v6": 0.6204506233334541, + "step": 761 + }, + { + "epoch": 0.9150405283698589, + "grad_norm": 13.187544822692871, + "learning_rate": 3.072207814595529e-07, + "loss": 10.111783981323242, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.48953042551875114, + "loss_total_v6": 0.6319864839315414, + "step": 762 + }, + { + "epoch": 0.9162413689582708, + "grad_norm": 11.990710258483887, + "learning_rate": 2.9871423792206257e-07, + "loss": 10.150581359863281, + "loss_alignment": 0.28369140625, + "loss_alignment_w": 0.141845703125, + "loss_sft": 0.49216895923018456, + "loss_total_v6": 0.6344113647937775, + "step": 763 + }, + { + "epoch": 0.9174422095466827, + "grad_norm": 12.116122245788574, + "learning_rate": 2.903234942578081e-07, + "loss": 9.995555877685547, + "loss_alignment": 0.29052734375, + "loss_alignment_w": 0.145263671875, + "loss_sft": 0.4790770635008812, + "loss_total_v6": 0.6247222051024437, + "step": 764 + }, + { + "epoch": 0.9186430501350946, + "grad_norm": 12.241293907165527, + "learning_rate": 2.820487571436631e-07, + "loss": 10.02401351928711, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.4835870638489723, + "loss_total_v6": 0.6265008747577667, + "step": 765 + }, + { + "epoch": 0.9198438907235065, + "grad_norm": 11.922917366027832, + "learning_rate": 2.7389023039908536e-07, + "loss": 9.932190895080566, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.4763222336769104, + "loss_total_v6": 0.620761938393116, + "step": 766 + }, + { + "epoch": 0.9210447313119183, + "grad_norm": 12.784666061401367, + "learning_rate": 2.658481149810904e-07, + "loss": 9.858722686767578, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47180673480033875, + "loss_total_v6": 0.61617012321949, + "step": 767 + }, + { + "epoch": 0.9222455719003302, + "grad_norm": 12.975274085998535, + "learning_rate": 2.579226089793074e-07, + "loss": 9.962772369384766, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.47821830213069916, + "loss_total_v6": 0.6226732581853867, + "step": 768 + }, + { + "epoch": 0.9234464124887422, + "grad_norm": 12.225621223449707, + "learning_rate": 2.5011390761109424e-07, + "loss": 9.526288986206055, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4527996964752674, + "loss_total_v6": 0.5953930765390396, + "step": 769 + }, + { + "epoch": 0.924647253077154, + "grad_norm": 11.931862831115723, + "learning_rate": 2.4242220321673394e-07, + "loss": 9.985420227050781, + "loss_alignment": 0.289306640625, + "loss_alignment_w": 0.1446533203125, + "loss_sft": 0.47964904457330704, + "loss_total_v6": 0.624088741838932, + "step": 770 + }, + { + "epoch": 0.9258480936655659, + "grad_norm": 12.366680145263672, + "learning_rate": 2.3484768525469215e-07, + "loss": 10.335588455200195, + "loss_alignment": 0.295166015625, + "loss_alignment_w": 0.1475830078125, + "loss_sft": 0.4986964948475361, + "loss_total_v6": 0.6459743231534958, + "step": 771 + }, + { + "epoch": 0.9270489342539778, + "grad_norm": 12.074766159057617, + "learning_rate": 2.2739054029695562e-07, + "loss": 9.95258617401123, + "loss_alignment": 0.282470703125, + "loss_alignment_w": 0.1412353515625, + "loss_sft": 0.4811827205121517, + "loss_total_v6": 0.6220366209745407, + "step": 772 + }, + { + "epoch": 0.9282497748423897, + "grad_norm": 12.083561897277832, + "learning_rate": 2.200509520244326e-07, + "loss": 9.864420890808105, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.4733225479722023, + "loss_total_v6": 0.6165263056755066, + "step": 773 + }, + { + "epoch": 0.9294506154308015, + "grad_norm": 11.807228088378906, + "learning_rate": 2.1282910122243038e-07, + "loss": 9.951371192932129, + "loss_alignment": 0.2823486328125, + "loss_alignment_w": 0.14117431640625, + "loss_sft": 0.48077111318707466, + "loss_total_v6": 0.6219606995582581, + "step": 774 + }, + { + "epoch": 0.9306514560192134, + "grad_norm": 12.500630378723145, + "learning_rate": 2.0572516577620216e-07, + "loss": 10.014382362365723, + "loss_alignment": 0.2806396484375, + "loss_alignment_w": 0.14031982421875, + "loss_sft": 0.48545701801776886, + "loss_total_v6": 0.6258989050984383, + "step": 775 + }, + { + "epoch": 0.9318522966076254, + "grad_norm": 12.386434555053711, + "learning_rate": 1.9873932066656555e-07, + "loss": 10.255945205688477, + "loss_alignment": 0.299560546875, + "loss_alignment_w": 0.1497802734375, + "loss_sft": 0.4913994371891022, + "loss_total_v6": 0.6409965977072716, + "step": 776 + }, + { + "epoch": 0.9330531371960372, + "grad_norm": 12.310290336608887, + "learning_rate": 1.9187173796559044e-07, + "loss": 9.908509254455566, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.4739113561809063, + "loss_total_v6": 0.6192818358540535, + "step": 777 + }, + { + "epoch": 0.9342539777844491, + "grad_norm": 12.548418998718262, + "learning_rate": 1.851225868323653e-07, + "loss": 10.15916633605957, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.49117955937981606, + "loss_total_v6": 0.6349478587508202, + "step": 778 + }, + { + "epoch": 0.935454818372861, + "grad_norm": 12.45630931854248, + "learning_rate": 1.7849203350882415e-07, + "loss": 10.114280700683594, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.4858260303735733, + "loss_total_v6": 0.6321425512433052, + "step": 779 + }, + { + "epoch": 0.9366556589612729, + "grad_norm": 12.204605102539062, + "learning_rate": 1.7198024131565848e-07, + "loss": 10.047460556030273, + "loss_alignment": 0.297119140625, + "loss_alignment_w": 0.1485595703125, + "loss_sft": 0.4789031967520714, + "loss_total_v6": 0.6279663071036339, + "step": 780 + }, + { + "epoch": 0.9378564995496848, + "grad_norm": 12.615283012390137, + "learning_rate": 1.65587370648288e-07, + "loss": 9.84217643737793, + "loss_alignment": 0.2840576171875, + "loss_alignment_w": 0.14202880859375, + "loss_sft": 0.4728936217725277, + "loss_total_v6": 0.6151360496878624, + "step": 781 + }, + { + "epoch": 0.9390573401380967, + "grad_norm": 12.853215217590332, + "learning_rate": 1.5931357897291665e-07, + "loss": 9.935422897338867, + "loss_alignment": 0.2724609375, + "loss_alignment_w": 0.13623046875, + "loss_sft": 0.4845045432448387, + "loss_total_v6": 0.6209639087319374, + "step": 782 + }, + { + "epoch": 0.9402581807265086, + "grad_norm": 11.932971954345703, + "learning_rate": 1.5315902082264578e-07, + "loss": 9.948887825012207, + "loss_alignment": 0.299560546875, + "loss_alignment_w": 0.1497802734375, + "loss_sft": 0.47194892913103104, + "loss_total_v6": 0.6218055039644241, + "step": 783 + }, + { + "epoch": 0.9414590213149204, + "grad_norm": 11.955374717712402, + "learning_rate": 1.471238477936765e-07, + "loss": 10.192296981811523, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.491953257471323, + "loss_total_v6": 0.6370185762643814, + "step": 784 + }, + { + "epoch": 0.9426598619033323, + "grad_norm": 12.232365608215332, + "learning_rate": 1.4120820854156946e-07, + "loss": 10.09596061706543, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.4840400852262974, + "loss_total_v6": 0.6309974864125252, + "step": 785 + }, + { + "epoch": 0.9438607024917443, + "grad_norm": 12.013309478759766, + "learning_rate": 1.3541224877758498e-07, + "loss": 10.129365921020508, + "loss_alignment": 0.285888671875, + "loss_alignment_w": 0.1429443359375, + "loss_sft": 0.48997312411665916, + "loss_total_v6": 0.6330853179097176, + "step": 786 + }, + { + "epoch": 0.9450615430801561, + "grad_norm": 12.656401634216309, + "learning_rate": 1.2973611126509465e-07, + "loss": 10.000003814697266, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.4821779765188694, + "loss_total_v6": 0.6250002384185791, + "step": 787 + }, + { + "epoch": 0.946262383668568, + "grad_norm": 11.848062515258789, + "learning_rate": 1.2417993581606446e-07, + "loss": 9.959399223327637, + "loss_alignment": 0.28466796875, + "loss_alignment_w": 0.142333984375, + "loss_sft": 0.4802962876856327, + "loss_total_v6": 0.6224624291062355, + "step": 788 + }, + { + "epoch": 0.9474632242569799, + "grad_norm": 12.167801856994629, + "learning_rate": 1.1874385928761112e-07, + "loss": 9.852306365966797, + "loss_alignment": 0.2723388671875, + "loss_alignment_w": 0.13616943359375, + "loss_sft": 0.4797218181192875, + "loss_total_v6": 0.6157691925764084, + "step": 789 + }, + { + "epoch": 0.9486640648453918, + "grad_norm": 11.859004020690918, + "learning_rate": 1.1342801557862937e-07, + "loss": 10.083163261413574, + "loss_alignment": 0.293212890625, + "loss_alignment_w": 0.1466064453125, + "loss_sft": 0.4834386445581913, + "loss_total_v6": 0.6301976963877678, + "step": 790 + }, + { + "epoch": 0.9498649054338036, + "grad_norm": 11.752028465270996, + "learning_rate": 1.0823253562649793e-07, + "loss": 9.974288940429688, + "loss_alignment": 0.2884521484375, + "loss_alignment_w": 0.14422607421875, + "loss_sft": 0.4789685904979706, + "loss_total_v6": 0.6233930587768555, + "step": 791 + }, + { + "epoch": 0.9510657460222155, + "grad_norm": 11.949479103088379, + "learning_rate": 1.0315754740384931e-07, + "loss": 9.866317749023438, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47193052992224693, + "loss_total_v6": 0.6166448593139648, + "step": 792 + }, + { + "epoch": 0.9522665866106275, + "grad_norm": 12.403303146362305, + "learning_rate": 9.820317591542172e-08, + "loss": 9.942386627197266, + "loss_alignment": 0.277587890625, + "loss_alignment_w": 0.1387939453125, + "loss_sft": 0.4828036017715931, + "loss_total_v6": 0.6213991865515709, + "step": 793 + }, + { + "epoch": 0.9534674271990393, + "grad_norm": 11.979961395263672, + "learning_rate": 9.336954319497716e-08, + "loss": 9.974323272705078, + "loss_alignment": 0.2725830078125, + "loss_alignment_w": 0.13629150390625, + "loss_sft": 0.4871952384710312, + "loss_total_v6": 0.6233951821923256, + "step": 794 + }, + { + "epoch": 0.9546682677874512, + "grad_norm": 11.870890617370605, + "learning_rate": 8.865676830229819e-08, + "loss": 10.114316940307617, + "loss_alignment": 0.28125, + "loss_alignment_w": 0.140625, + "loss_sft": 0.4916113615036011, + "loss_total_v6": 0.6321448236703873, + "step": 795 + }, + { + "epoch": 0.9558691083758631, + "grad_norm": 12.01485824584961, + "learning_rate": 8.406496732025149e-08, + "loss": 10.108768463134766, + "loss_alignment": 0.2939453125, + "loss_alignment_w": 0.14697265625, + "loss_sft": 0.48481014370918274, + "loss_total_v6": 0.6317980512976646, + "step": 796 + }, + { + "epoch": 0.957069948964275, + "grad_norm": 11.960858345031738, + "learning_rate": 7.959425335193338e-08, + "loss": 9.835641860961914, + "loss_alignment": 0.2890625, + "loss_alignment_w": 0.14453125, + "loss_sft": 0.47027264907956123, + "loss_total_v6": 0.6147276237607002, + "step": 797 + }, + { + "epoch": 0.9582707895526869, + "grad_norm": 14.04468059539795, + "learning_rate": 7.524473651788045e-08, + "loss": 10.11624526977539, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.4868032969534397, + "loss_total_v6": 0.6322653517127037, + "step": 798 + }, + { + "epoch": 0.9594716301410988, + "grad_norm": 11.928390502929688, + "learning_rate": 7.101652395335779e-08, + "loss": 10.165857315063477, + "loss_alignment": 0.277587890625, + "loss_alignment_w": 0.1387939453125, + "loss_sft": 0.4962669350206852, + "loss_total_v6": 0.6353660225868225, + "step": 799 + }, + { + "epoch": 0.9606724707295107, + "grad_norm": 12.137543678283691, + "learning_rate": 6.690971980572059e-08, + "loss": 10.008519172668457, + "loss_alignment": 0.293701171875, + "loss_alignment_w": 0.1468505859375, + "loss_sft": 0.47888022288680077, + "loss_total_v6": 0.6255324557423592, + "step": 800 + }, + { + "epoch": 0.9606724707295107, + "eval_loss": 0.5849262475967407, + "eval_loss_alignment": 0.2810011643629808, + "eval_loss_alignment_w": 0.1405005821814904, + "eval_loss_sft": 0.44442569095498097, + "eval_loss_total_v6": 0.5849262732618417, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 68.1083, + "eval_samples_per_second": 24.358, + "eval_steps_per_second": 3.054, + "step": 800 + }, + { + "epoch": 0.9618733113179225, + "grad_norm": 12.581014633178711, + "learning_rate": 6.29244252318495e-08, + "loss": 10.034568786621094, + "loss_alignment": 0.282958984375, + "loss_alignment_w": 0.1414794921875, + "loss_sft": 0.48572683334350586, + "loss_total_v6": 0.6271605640649796, + "step": 801 + }, + { + "epoch": 0.9630741519063344, + "grad_norm": 14.88102912902832, + "learning_rate": 5.906073839565707e-08, + "loss": 9.986679077148438, + "loss_alignment": 0.2763671875, + "loss_alignment_w": 0.13818359375, + "loss_sft": 0.48583124577999115, + "loss_total_v6": 0.6241674199700356, + "step": 802 + }, + { + "epoch": 0.9642749924947464, + "grad_norm": 13.074020385742188, + "learning_rate": 5.5318754465671366e-08, + "loss": 9.99166202545166, + "loss_alignment": 0.296142578125, + "loss_alignment_w": 0.1480712890625, + "loss_sft": 0.47660598158836365, + "loss_total_v6": 0.6244789063930511, + "step": 803 + }, + { + "epoch": 0.9654758330831582, + "grad_norm": 12.110273361206055, + "learning_rate": 5.169856561269171e-08, + "loss": 9.942655563354492, + "loss_alignment": 0.272216796875, + "loss_alignment_w": 0.1361083984375, + "loss_sft": 0.48504816740751266, + "loss_total_v6": 0.6214159578084946, + "step": 804 + }, + { + "epoch": 0.9666766736715701, + "grad_norm": 11.828030586242676, + "learning_rate": 4.820026100751718e-08, + "loss": 9.852203369140625, + "loss_alignment": 0.278076171875, + "loss_alignment_w": 0.1390380859375, + "loss_sft": 0.476633045822382, + "loss_total_v6": 0.6157626807689667, + "step": 805 + }, + { + "epoch": 0.967877514259982, + "grad_norm": 12.038790702819824, + "learning_rate": 4.482392681875114e-08, + "loss": 9.836400985717773, + "loss_alignment": 0.2845458984375, + "loss_alignment_w": 0.14227294921875, + "loss_sft": 0.47219690680503845, + "loss_total_v6": 0.6147750616073608, + "step": 806 + }, + { + "epoch": 0.9690783548483939, + "grad_norm": 11.445390701293945, + "learning_rate": 4.1569646210680156e-08, + "loss": 10.039230346679688, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.48058604449033737, + "loss_total_v6": 0.6274518743157387, + "step": 807 + }, + { + "epoch": 0.9702791954368057, + "grad_norm": 12.598727226257324, + "learning_rate": 3.843749934122232e-08, + "loss": 9.949462890625, + "loss_alignment": 0.28662109375, + "loss_alignment_w": 0.143310546875, + "loss_sft": 0.4785309173166752, + "loss_total_v6": 0.6218414753675461, + "step": 808 + }, + { + "epoch": 0.9714800360252176, + "grad_norm": 11.94807243347168, + "learning_rate": 3.54275633599549e-08, + "loss": 10.040105819702148, + "loss_alignment": 0.29443359375, + "loss_alignment_w": 0.147216796875, + "loss_sft": 0.4804271459579468, + "loss_total_v6": 0.6275065913796425, + "step": 809 + }, + { + "epoch": 0.9726808766136296, + "grad_norm": 11.753931999206543, + "learning_rate": 3.2539912406214235e-08, + "loss": 10.126696586608887, + "loss_alignment": 0.28564453125, + "loss_alignment_w": 0.142822265625, + "loss_sft": 0.4907066151499748, + "loss_total_v6": 0.6329185217618942, + "step": 810 + }, + { + "epoch": 0.9738817172020414, + "grad_norm": 11.965219497680664, + "learning_rate": 2.977461760726885e-08, + "loss": 10.274662971496582, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.4990237019956112, + "loss_total_v6": 0.6421664357185364, + "step": 811 + }, + { + "epoch": 0.9750825577904533, + "grad_norm": 12.096508026123047, + "learning_rate": 2.7131747076567517e-08, + "loss": 9.976122856140137, + "loss_alignment": 0.29736328125, + "loss_alignment_w": 0.148681640625, + "loss_sft": 0.47462768480181694, + "loss_total_v6": 0.6235077008605003, + "step": 812 + }, + { + "epoch": 0.9762833983788652, + "grad_norm": 12.359648704528809, + "learning_rate": 2.4611365912061147e-08, + "loss": 10.092123031616211, + "loss_alignment": 0.290771484375, + "loss_alignment_w": 0.1453857421875, + "loss_sft": 0.4850972853600979, + "loss_total_v6": 0.6307576969265938, + "step": 813 + }, + { + "epoch": 0.9774842389672771, + "grad_norm": 12.034058570861816, + "learning_rate": 2.2213536194601314e-08, + "loss": 10.015835762023926, + "loss_alignment": 0.281005859375, + "loss_alignment_w": 0.1405029296875, + "loss_sft": 0.48551737517118454, + "loss_total_v6": 0.6259897798299789, + "step": 814 + }, + { + "epoch": 0.978685079555689, + "grad_norm": 11.980525970458984, + "learning_rate": 1.9938316986408134e-08, + "loss": 9.909453392028809, + "loss_alignment": 0.281494140625, + "loss_alignment_w": 0.1407470703125, + "loss_sft": 0.4785632751882076, + "loss_total_v6": 0.6193408444523811, + "step": 815 + }, + { + "epoch": 0.9798859201441009, + "grad_norm": 11.844223022460938, + "learning_rate": 1.7785764329618092e-08, + "loss": 10.046817779541016, + "loss_alignment": 0.2764892578125, + "loss_alignment_w": 0.13824462890625, + "loss_sft": 0.4895441271364689, + "loss_total_v6": 0.6279260963201523, + "step": 816 + }, + { + "epoch": 0.9810867607325128, + "grad_norm": 12.536227226257324, + "learning_rate": 1.5755931244901822e-08, + "loss": 10.000528335571289, + "loss_alignment": 0.2763671875, + "loss_alignment_w": 0.13818359375, + "loss_sft": 0.48730722442269325, + "loss_total_v6": 0.6250330209732056, + "step": 817 + }, + { + "epoch": 0.9822876013209246, + "grad_norm": 12.34549617767334, + "learning_rate": 1.3848867730158477e-08, + "loss": 9.913007736206055, + "loss_alignment": 0.2747802734375, + "loss_alignment_w": 0.13739013671875, + "loss_sft": 0.4823101572692394, + "loss_total_v6": 0.6195629760622978, + "step": 818 + }, + { + "epoch": 0.9834884419093365, + "grad_norm": 11.613489151000977, + "learning_rate": 1.2064620759285605e-08, + "loss": 10.130988121032715, + "loss_alignment": 0.288330078125, + "loss_alignment_w": 0.1441650390625, + "loss_sft": 0.4893573597073555, + "loss_total_v6": 0.6331867352128029, + "step": 819 + }, + { + "epoch": 0.9846892824977485, + "grad_norm": 12.142189979553223, + "learning_rate": 1.0403234281020636e-08, + "loss": 9.908531188964844, + "loss_alignment": 0.2813720703125, + "loss_alignment_w": 0.14068603515625, + "loss_sft": 0.47882604226469994, + "loss_total_v6": 0.6192831993103027, + "step": 820 + }, + { + "epoch": 0.9858901230861603, + "grad_norm": 12.081549644470215, + "learning_rate": 8.864749217858404e-09, + "loss": 9.850061416625977, + "loss_alignment": 0.28173828125, + "loss_alignment_w": 0.140869140625, + "loss_sft": 0.47465285658836365, + "loss_total_v6": 0.6156288012862206, + "step": 821 + }, + { + "epoch": 0.9870909636745722, + "grad_norm": 11.961663246154785, + "learning_rate": 7.449203465043631e-09, + "loss": 10.015340805053711, + "loss_alignment": 0.287353515625, + "loss_alignment_w": 0.1436767578125, + "loss_sft": 0.48241934552788734, + "loss_total_v6": 0.6259587556123734, + "step": 822 + }, + { + "epoch": 0.9882918042629841, + "grad_norm": 12.38041877746582, + "learning_rate": 6.156631889637776e-09, + "loss": 9.909435272216797, + "loss_alignment": 0.28955078125, + "loss_alignment_w": 0.144775390625, + "loss_sft": 0.474823784083128, + "loss_total_v6": 0.6193397417664528, + "step": 823 + }, + { + "epoch": 0.989492644851396, + "grad_norm": 12.468403816223145, + "learning_rate": 4.987066329659729e-09, + "loss": 10.181056022644043, + "loss_alignment": 0.29150390625, + "loss_alignment_w": 0.145751953125, + "loss_sft": 0.4903961904346943, + "loss_total_v6": 0.6363159865140915, + "step": 824 + }, + { + "epoch": 0.9906934854398078, + "grad_norm": 12.210439682006836, + "learning_rate": 3.940535593301431e-09, + "loss": 9.853723526000977, + "loss_alignment": 0.2833251953125, + "loss_alignment_w": 0.14166259765625, + "loss_sft": 0.47413407266139984, + "loss_total_v6": 0.6158577278256416, + "step": 825 + }, + { + "epoch": 0.9918943260282198, + "grad_norm": 11.77171802520752, + "learning_rate": 3.017065458219559e-09, + "loss": 10.25671672821045, + "loss_alignment": 0.296875, + "loss_alignment_w": 0.1484375, + "loss_sft": 0.49291251972317696, + "loss_total_v6": 0.6410448327660561, + "step": 826 + }, + { + "epoch": 0.9930951666166317, + "grad_norm": 12.224926948547363, + "learning_rate": 2.2166786708976985e-09, + "loss": 10.212980270385742, + "loss_alignment": 0.291748046875, + "loss_alignment_w": 0.1458740234375, + "loss_sft": 0.492406751960516, + "loss_total_v6": 0.6383112892508507, + "step": 827 + }, + { + "epoch": 0.9942960072050435, + "grad_norm": 11.933417320251465, + "learning_rate": 1.5393949460895675e-09, + "loss": 9.755203247070312, + "loss_alignment": 0.286865234375, + "loss_alignment_w": 0.1434326171875, + "loss_sft": 0.46631335839629173, + "loss_total_v6": 0.6097002103924751, + "step": 828 + }, + { + "epoch": 0.9954968477934554, + "grad_norm": 11.755271911621094, + "learning_rate": 9.852309663310744e-10, + "loss": 10.11227035522461, + "loss_alignment": 0.282470703125, + "loss_alignment_w": 0.1412353515625, + "loss_sft": 0.4911935031414032, + "loss_total_v6": 0.6320168748497963, + "step": 829 + }, + { + "epoch": 0.9966976883818673, + "grad_norm": 12.263010025024414, + "learning_rate": 5.542003815306451e-10, + "loss": 9.884382247924805, + "loss_alignment": 0.294189453125, + "loss_alignment_w": 0.1470947265625, + "loss_sft": 0.4707402288913727, + "loss_total_v6": 0.6177739128470421, + "step": 830 + }, + { + "epoch": 0.9978985289702792, + "grad_norm": 12.584160804748535, + "learning_rate": 2.4631380863227026e-10, + "loss": 9.687171936035156, + "loss_alignment": 0.292724609375, + "loss_alignment_w": 0.1463623046875, + "loss_sft": 0.45908593758940697, + "loss_total_v6": 0.6054482385516167, + "step": 831 + }, + { + "epoch": 0.999099369558691, + "grad_norm": 12.609256744384766, + "learning_rate": 6.157883135349264e-11, + "loss": 9.835334777832031, + "loss_alignment": 0.28515625, + "loss_alignment_w": 0.142578125, + "loss_sft": 0.47231339663267136, + "loss_total_v6": 0.6147084161639214, + "step": 832 + }, + { + "epoch": 1.0, + "grad_norm": 13.967070579528809, + "learning_rate": 0.0, + "loss": 7.222443103790283, + "loss_alignment": 0.2784830729166667, + "loss_alignment_w": 0.13924153645833334, + "loss_sft": 0.4628728727499644, + "loss_total_v6": 0.6018702735503515, + "step": 833 + }, + { + "epoch": 1.0, + "eval_loss": 0.5848813056945801, + "eval_loss_alignment": 0.28100820688100964, + "eval_loss_alignment_w": 0.14050410344050482, + "eval_loss_sft": 0.44437721299114996, + "eval_loss_total_v6": 0.5848813164406098, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 67.0067, + "eval_samples_per_second": 24.759, + "eval_steps_per_second": 3.104, + "step": 833 + } + ], + "logging_steps": 1, + "max_steps": 833, + "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": true + }, + "attributes": {} + } + }, + "total_flos": 2.8094110387910214e+19, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/training_args.bin b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..ac4a5592013e7721dcd97083dfb20607916be312 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:12f6393c2b2034efc0ed40b547b24c2d88742c623ca65973540f99f44b9994d7 +size 9041 diff --git a/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/zero_to_fp32.py b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..5995d6e6f04e43b989587aa9022a3aef0c66d694 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_planA_wm_joint_v6_parallel_k32_packing-20260606-155825__v0-20260606-155855__checkpoint-833/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)