diff --git a/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/args.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/args.json new file mode 100644 index 0000000000000000000000000000000000000000..b33a2faf9faa5aa07952d341a1da7647915018ed --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/args.json @@ -0,0 +1,382 @@ +{ + "output_dir": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v2-20260610-131619", + "per_device_train_batch_size": 1, + "num_train_epochs": 1.0, + "max_steps": -1, + "learning_rate": 1e-05, + "lr_scheduler_type": "cosine", + "lr_scheduler_kwargs": null, + "warmup_steps": 200.0, + "optim": "adamw_torch_fused", + "optim_args": null, + "weight_decay": 0.0, + "adam_beta1": 0.9, + "adam_beta2": 0.95, + "adam_epsilon": 1e-08, + "optim_target_modules": null, + "gradient_accumulation_steps": 16, + "average_tokens_across_devices": true, + "max_grad_norm": 1.0, + "label_smoothing_factor": 0.0, + "bf16": true, + "fp16": false, + "bf16_full_eval": false, + "fp16_full_eval": false, + "tf32": true, + "gradient_checkpointing": true, + "gradient_checkpointing_kwargs": null, + "torch_compile": false, + "torch_compile_backend": null, + "torch_compile_mode": null, + "use_liger_kernel": false, + "liger_kernel_config": null, + "use_cache": false, + "neftune_noise_alpha": null, + "torch_empty_cache_steps": null, + "auto_find_batch_size": false, + "logging_strategy": "steps", + "logging_steps": 1, + "logging_first_step": true, + "log_on_each_node": true, + "logging_nan_inf_filter": true, + "include_num_input_tokens_seen": false, + "log_level": "passive", + "log_level_replica": "warning", + "disable_tqdm": null, + "report_to": [ + "wandb" + ], + "run_name": "ocu_add_spar_resume2", + "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": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v0-20260609-213839/checkpoint-900", + "warmup_ratio": null, + "logging_dir": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v2-20260610-131619/runs", + "local_rank": 0, + "sortish_sampler": false, + "predict_with_generate": false, + "generation_max_length": null, + "generation_num_beams": null, + "generation_config": null, + "tuner_backend": "peft", + "vit_gradient_checkpointing": false, + "router_aux_loss_coef": 0.0, + "enable_dft_loss": false, + "enable_channel_loss": false, + "safe_serialization": true, + "max_shard_size": "5GB", + "check_model": true, + "acc_strategy": "token", + "train_dataloader_shuffle": true, + "group_by_length": false, + "max_epochs": null, + "aligner_lr": null, + "vit_lr": null, + "use_logits_to_keep": null, + "ds3_gather_for_generation": true, + "resume_only_model": false, + "optimizer": null, + "loss_type": null, + "eval_metric": null, + "callbacks": [], + "early_stop_interval": null, + "eval_use_evalscope": false, + "eval_dataset": [], + "eval_dataset_args": null, + "eval_limit": null, + "eval_generation_config": null, + "extra_eval_args": null, + "tuner_type": "full", + "use_galore": false, + "galore_target_modules": null, + "galore_rank": 128, + "galore_update_proj_gap": 50, + "galore_scale": 1.0, + "galore_proj_type": "std", + "galore_optim_per_parameter": false, + "galore_with_embedding": false, + "galore_quantization": false, + "galore_proj_quant": false, + "galore_proj_bits": 4, + "galore_proj_group_size": 256, + "galore_cos_threshold": 0.4, + "galore_gamma_proj": 2, + "galore_queue_size": 5, + "lisa_activated_layers": 0, + "lisa_step_interval": 20, + "use_flash_ckpt": false, + "use_ray": false, + "ray_exp_name": null, + "device_groups": null, + "model": "/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", + "model_type": "qwen3_5", + "model_revision": null, + "task_type": "causal_lm", + "torch_dtype": "bfloat16", + "attn_impl": "flash_attention_2", + "experts_impl": null, + "new_special_tokens": [], + "num_labels": null, + "problem_type": null, + "rope_scaling": null, + "device_map": null, + "max_memory": {}, + "max_model_len": null, + "local_repo_path": null, + "init_strategy": null, + "template": "qwen3_5", + "system": null, + "max_length": 12288, + "truncation_strategy": "delete", + "max_pixels": 16777216, + "agent_template": null, + "norm_bbox": null, + "use_chat_template": true, + "padding_side": "right", + "padding_free": false, + "loss_scale": "default", + "sequence_parallel_size": 1, + "template_backend": "swift", + "response_prefix": null, + "enable_thinking": null, + "add_non_thinking_prefix": true, + "dataset": [ + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l1/ubuntu/train_wm.jsonl#35000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l2/ubuntu/train_wm.jsonl#35000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l3/ubuntu/train_wm.jsonl#35000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l1/win_mac/train_wm.jsonl#100000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l2/win_mac/train_wm.jsonl#100000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l3/win_mac/train_wm.jsonl#100000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/GUI-360/l1/train_wm.jsonl#50000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/GUI-360/l2/train_wm.jsonl#50000", + "/NHNHOME/data/sanghyeok/qwen-cua/data/GUI-360/l3/train_wm.jsonl#50000" + ], + "val_dataset": [ + "/NHNHOME/data/sanghyeok/qwen-cua/data/AgentNet/l2/ubuntu/val_wm.jsonl" + ], + "cached_dataset": [], + "cached_val_dataset": [], + "split_dataset_ratio": 0.0, + "dataset_num_proc": 4, + "load_from_cache_file": false, + "dataset_shuffle": true, + "val_dataset_shuffle": false, + "streaming": false, + "interleave_prob": null, + "stopping_strategy": "first_exhausted", + "shuffle_buffer_size": 1000, + "download_mode": "reuse_dataset_if_exists", + "columns": {}, + "strict": false, + "disable_auto_column_mapping": false, + "model_name": null, + "model_author": null, + "custom_dataset_info": [], + "quant_method": null, + "quant_bits": null, + "hqq_axis": null, + "bnb_4bit_compute_dtype": "bfloat16", + "bnb_4bit_quant_type": "nf4", + "bnb_4bit_use_double_quant": true, + "bnb_4bit_quant_storage": null, + "max_new_tokens": 64, + "temperature": 0.0, + "top_k": null, + "top_p": null, + "repetition_penalty": null, + "num_beams": 1, + "stream": false, + "stop_words": [], + "logprobs": false, + "top_logprobs": null, + "structured_outputs_regex": null, + "adapters": [], + "external_plugins": [], + "custom_register_path": [], + "model_kwargs": {}, + "enable_npu_model_patch": true, + "load_args": false, + "load_data_args": false, + "packing": false, + "packing_length": null, + "packing_num_proc": 1, + "lazy_tokenize": true, + "use_hf": false, + "ignore_args_error": false, + "use_swift_lora": false, + "freeze_parameters": [ + "model.visual", + "model.visual.merger" + ], + "freeze_parameters_regex": null, + "freeze_parameters_ratio": 0.0, + "trainable_parameters": [], + "trainable_parameters_regex": null, + "freeze_llm": false, + "freeze_vit": true, + "freeze_aligner": true, + "target_modules": [ + "all-linear" + ], + "target_regex": null, + "target_parameters": null, + "modules_to_save": [], + "lora_rank": 8, + "lora_alpha": 32, + "lora_dropout": 0.05, + "lora_bias": "none", + "lora_dtype": null, + "lorap_lr_ratio": null, + "use_rslora": false, + "use_dora": false, + "lora_ga_batch_size": 2, + "lora_ga_iters": 2, + "lora_ga_max_length": 1024, + "lora_ga_direction": "ArB2r", + "lora_ga_scale": "stable", + "lora_ga_stable_gamma": 16, + "init_weights": true, + "fourier_n_frequency": 2000, + "fourier_scaling": 300.0, + "boft_block_size": 4, + "boft_block_num": 0, + "boft_n_butterfly_factor": 1, + "boft_dropout": 0.0, + "vera_rank": 256, + "vera_projection_prng_key": 0, + "vera_dropout": 0.0, + "vera_d_initial": 0.1, + "adapter_act": "gelu", + "adapter_length": 128, + "adalora_target_r": 8, + "adalora_init_r": 12, + "adalora_tinit": 0, + "adalora_tfinal": 0, + "adalora_deltaT": 1, + "adalora_beta1": 0.85, + "adalora_beta2": 0.85, + "adalora_orth_reg_weight": 0.5, + "llamapro_num_new_blocks": 4, + "llamapro_num_groups": null, + "reft_layer_key": null, + "reft_layers": null, + "reft_rank": 4, + "reft_intervention_type": "LoreftIntervention", + "reft_args": null, + "swanlab_token": null, + "swanlab_project": "ms-swift", + "swanlab_workspace": null, + "swanlab_exp_name": null, + "swanlab_notification_method": null, + "swanlab_webhook_url": null, + "swanlab_secret": null, + "swanlab_sender_email": null, + "swanlab_receiver_email": null, + "swanlab_smtp_server": null, + "swanlab_smtp_port": null, + "swanlab_email_language": "zh", + "swanlab_mode": "cloud", + "add_version": true, + "create_checkpoint_symlink": false, + "zero_hpz_partition_size": null, + "deepspeed_autotp_size": null, + "swift_version": "4.2.1", + "ckpt_dir": null, + "rank": 0, + "global_world_size": 8, + "local_world_size": 8, + "model_suffix": "Qwen3.5-4B", + "model_info": "ModelInfo(model_type='qwen3_5', model_dir='/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a', torch_dtype=torch.bfloat16, max_model_len=262144, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)", + "model_meta": "ModelMeta(model_type='qwen3_5', model_groups=[ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3.5-0.8B', hf_model_id='Qwen/Qwen3.5-0.8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-2B', hf_model_id='Qwen/Qwen3.5-2B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-4B', hf_model_id='Qwen/Qwen3.5-4B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-9B', hf_model_id='Qwen/Qwen3.5-9B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-27B', hf_model_id='Qwen/Qwen3.5-27B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-27B-FP8', hf_model_id='Qwen/Qwen3.5-27B-FP8', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-0.8B-Base', hf_model_id='Qwen/Qwen3.5-0.8B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-2B-Base', hf_model_id='Qwen/Qwen3.5-2B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-4B-Base', hf_model_id='Qwen/Qwen3.5-4B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.5-9B-Base', hf_model_id='Qwen/Qwen3.5-9B-Base', model_path=None, ms_revision=None, hf_revision=None)], template='qwen3_5', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='Qwen/Qwen3.6-27B', hf_model_id='Qwen/Qwen3.6-27B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Qwen/Qwen3.6-27B-FP8', hf_model_id='Qwen/Qwen3.6-27B-FP8', model_path=None, ms_revision=None, hf_revision=None)], template='qwen3_5', ignore_patterns=None, requires=None, tags=[])], loader=, template='qwen3_5', model_arch=MultiModelKeys(arch_name='qwen2_vl', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model.language_model', 'lm_head'], aligner=['model.visual.merger'], vision_tower=['model.visual'], generator=[]), mcore_model_type=None, architectures=['Qwen3_5ForConditionalGeneration'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, task_type=None, ignore_patterns=None, requires=['transformers>=5.0.0.dev', 'qwen_vl_utils>=0.0.14', 'decord'], tags=[])", + "model_dir": "/NHNHOME/data/sanghyeok/qwen-cua/hf_cache/hub/models--Qwen--Qwen3.5-4B/snapshots/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a", + "template_meta": "QwenTemplateMeta(template_type='qwen3_5', prefix=[], prompt=['<|im_start|>user\\n{{QUERY}}<|im_end|>\\n<|im_start|>assistant\\n'], chat_sep=['<|im_end|>\\n'], suffix=['<|im_end|>\\n'], template_cls=, system_prefix=['<|im_start|>system\\n{{SYSTEM}}<|im_end|>\\n'], default_system=None, auto_add_bos=False, stop_words=['<|endoftext|>'], agent_template='qwen3_5', is_thinking=True, thinking_prefix='\\n', non_thinking_prefix='\\n\\n\\n\\n', history_thinking_prefix='')", + "_val_dataset_exists": true, + "hub": "", + "evaluation_strategy": "steps", + "training_args": "Seq2SeqTrainingArguments(output_dir='/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v2-20260610-131619', per_device_train_batch_size=1, num_train_epochs=1.0, max_steps=-1, learning_rate=1e-05, lr_scheduler_type=, lr_scheduler_kwargs=None, warmup_steps=200.0, optim=, optim_args=None, weight_decay=0.0, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, optim_target_modules=None, gradient_accumulation_steps=16, average_tokens_across_devices=None, max_grad_norm=1.0, label_smoothing_factor=0.0, bf16=True, fp16=False, bf16_full_eval=False, fp16_full_eval=False, tf32=True, gradient_checkpointing=True, gradient_checkpointing_kwargs=None, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, use_liger_kernel=False, liger_kernel_config=None, use_cache=False, neftune_noise_alpha=None, torch_empty_cache_steps=None, auto_find_batch_size=False, logging_strategy=, logging_steps=1, logging_first_step=True, log_on_each_node=True, logging_nan_inf_filter=True, include_num_input_tokens_seen=None, log_level='passive', log_level_replica='warning', disable_tqdm=False, report_to=['wandb'], run_name='ocu_add_spar_resume2', 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='/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v0-20260609-213839/checkpoint-900', warmup_ratio=None, logging_dir='/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v2-20260610-131619/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/chat_template.jinja b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/config.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/config.json new file mode 100644 index 0000000000000000000000000000000000000000..314297e9ae3a6f7ca90b01d29ee0a46cfbf53c2a --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/generation_config.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1f1ab970fed81cb788c2945bed33833d247b9091 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/latest b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/latest new file mode 100644 index 0000000000000000000000000000000000000000..cfe584fa09082a54d48545d05e97aee6fd00673d --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/latest @@ -0,0 +1 @@ +global_step2600 \ No newline at end of file diff --git a/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00001-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..59e29d789f92786b67ab6ff64c9a4681c633fccd --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb999910509a6c3f6ab420fc830b22c25593711f8c21831a9fb3fe3279f2cafb +size 4989860336 diff --git a/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00002-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..15c05bd1d1f4057ef9d7737245b717082db70b34 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e10a790a0cd505886ea181faa61d2945719b156a42204b303e96d2f81fa39a1 +size 4993498424 diff --git a/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00003-of-00003.safetensors b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4ab8023a740649342f1bc7e224320898de73054c --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model.safetensors.index.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..6a37ec4c02c77592002094503fd812d32a3c3c13 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/model.safetensors.index.json @@ -0,0 +1,747 @@ +{ + "metadata": { + "total_parameters": 4555712896, + "total_size": 10381595392 + }, + "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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/preprocessor_config.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..2ea84a437d448ff71b08df68fdd949d5cc4ebb64 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/processor_config.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/tokenizer.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..9e8ac8634b7c90a05a65182d43516ddf7371c521 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/tokenizer_config.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/trainer_state.json b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..7e233ef6e8270f67a3010fb335066aea151564b3 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/trainer_state.json @@ -0,0 +1,36798 @@ +{ + "best_global_step": 2600, + "best_metric": 0.6510933, + "best_model_checkpoint": "/NHNHOME/data/sanghyeok/qwen-cua/runs/qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735/v2-20260610-131619/checkpoint-2600", + "epoch": 0.5996396396396396, + "eval_steps": 100, + "global_step": 2600, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.00023063063063063062, + "grad_norm": 979.8712158203125, + "learning_rate": 5.0000000000000004e-08, + "loss": 27.91953468322754, + "loss_alignment": 0.9677734375, + "loss_alignment_w": 0.48388671875, + "loss_sft": 1.2628998905420303, + "loss_total_v6": 1.7449708878993988, + "step": 1 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.00046126126126126124, + "grad_norm": 961.8486938476562, + "learning_rate": 1.0000000000000001e-07, + "loss": 27.783802032470703, + "loss_alignment": 0.9404296875, + "loss_alignment_w": 0.47021484375, + "loss_sft": 1.2690804302692413, + "loss_total_v6": 1.736487716436386, + "step": 2 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0006918918918918919, + "grad_norm": 1012.0112915039062, + "learning_rate": 1.5000000000000002e-07, + "loss": 28.72005844116211, + "loss_alignment": 0.94775390625, + "loss_alignment_w": 0.473876953125, + "loss_sft": 1.3240869343280792, + "loss_total_v6": 1.7950036078691483, + "step": 3 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0009225225225225225, + "grad_norm": 934.4891357421875, + "learning_rate": 2.0000000000000002e-07, + "loss": 28.20231056213379, + "loss_alignment": 0.9541015625, + "loss_alignment_w": 0.47705078125, + "loss_sft": 1.2873789370059967, + "loss_total_v6": 1.7626444697380066, + "step": 4 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0011531531531531532, + "grad_norm": 844.5490112304688, + "learning_rate": 2.5000000000000004e-07, + "loss": 27.084012985229492, + "loss_alignment": 0.9462890625, + "loss_alignment_w": 0.47314453125, + "loss_sft": 1.2225665003061295, + "loss_total_v6": 1.6927508413791656, + "step": 5 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0013837837837837837, + "grad_norm": 791.7418823242188, + "learning_rate": 3.0000000000000004e-07, + "loss": 27.254131317138672, + "loss_alignment": 0.931640625, + "loss_alignment_w": 0.4658203125, + "loss_sft": 1.2400043904781342, + "loss_total_v6": 1.703383281826973, + "step": 6 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0016144144144144145, + "grad_norm": 969.467041015625, + "learning_rate": 3.5000000000000004e-07, + "loss": 27.982948303222656, + "loss_alignment": 0.92919921875, + "loss_alignment_w": 0.464599609375, + "loss_sft": 1.2862266600131989, + "loss_total_v6": 1.7489342242479324, + "step": 7 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.001845045045045045, + "grad_norm": 627.6566162109375, + "learning_rate": 4.0000000000000003e-07, + "loss": 24.99344825744629, + "loss_alignment": 0.9462890625, + "loss_alignment_w": 0.47314453125, + "loss_sft": 1.0916772931814194, + "loss_total_v6": 1.562090516090393, + "step": 8 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0020756756756756755, + "grad_norm": 735.4047241210938, + "learning_rate": 4.5000000000000003e-07, + "loss": 25.875030517578125, + "loss_alignment": 0.96240234375, + "loss_alignment_w": 0.481201171875, + "loss_sft": 1.138399139046669, + "loss_total_v6": 1.6171894371509552, + "step": 9 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0023063063063063064, + "grad_norm": 797.4049682617188, + "learning_rate": 5.000000000000001e-07, + "loss": 26.46940803527832, + "loss_alignment": 0.947265625, + "loss_alignment_w": 0.4736328125, + "loss_sft": 1.1833907961845398, + "loss_total_v6": 1.6543380171060562, + "step": 10 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.002536936936936937, + "grad_norm": 346.0902404785156, + "learning_rate": 5.5e-07, + "loss": 24.01948356628418, + "loss_alignment": 0.95361328125, + "loss_alignment_w": 0.476806640625, + "loss_sft": 1.0269593968987465, + "loss_total_v6": 1.5012178122997284, + "step": 11 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0027675675675675675, + "grad_norm": 301.3338317871094, + "learning_rate": 6.000000000000001e-07, + "loss": 23.15773582458496, + "loss_alignment": 0.916015625, + "loss_alignment_w": 0.4580078125, + "loss_sft": 0.9917005151510239, + "loss_total_v6": 1.4473584592342377, + "step": 12 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0029981981981981984, + "grad_norm": 304.8746032714844, + "learning_rate": 6.5e-07, + "loss": 23.269031524658203, + "loss_alignment": 0.927734375, + "loss_alignment_w": 0.4638671875, + "loss_sft": 0.9922936409711838, + "loss_total_v6": 1.4543145447969437, + "step": 13 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.003228828828828829, + "grad_norm": 238.10948181152344, + "learning_rate": 7.000000000000001e-07, + "loss": 22.5914363861084, + "loss_alignment": 0.92724609375, + "loss_alignment_w": 0.463623046875, + "loss_sft": 0.9502491652965546, + "loss_total_v6": 1.4119648188352585, + "step": 14 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0034594594594594594, + "grad_norm": 213.53884887695312, + "learning_rate": 7.5e-07, + "loss": 22.380325317382812, + "loss_alignment": 0.94921875, + "loss_alignment_w": 0.474609375, + "loss_sft": 0.9261293709278107, + "loss_total_v6": 1.3987702876329422, + "step": 15 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.00369009009009009, + "grad_norm": 213.8672637939453, + "learning_rate": 8.000000000000001e-07, + "loss": 22.11905288696289, + "loss_alignment": 0.943359375, + "loss_alignment_w": 0.4716796875, + "loss_sft": 0.9134314581751823, + "loss_total_v6": 1.3824408650398254, + "step": 16 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.003920720720720721, + "grad_norm": 184.2181396484375, + "learning_rate": 8.500000000000001e-07, + "loss": 21.486988067626953, + "loss_alignment": 0.923828125, + "loss_alignment_w": 0.4619140625, + "loss_sft": 0.8823654055595398, + "loss_total_v6": 1.3429367244243622, + "step": 17 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.004151351351351351, + "grad_norm": 153.22406005859375, + "learning_rate": 9.000000000000001e-07, + "loss": 21.572601318359375, + "loss_alignment": 0.97216796875, + "loss_alignment_w": 0.486083984375, + "loss_sft": 0.8647212311625481, + "loss_total_v6": 1.3482875227928162, + "step": 18 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.004381981981981982, + "grad_norm": 137.60369873046875, + "learning_rate": 9.500000000000001e-07, + "loss": 21.863948822021484, + "loss_alignment": 0.935546875, + "loss_alignment_w": 0.4677734375, + "loss_sft": 0.9019124582409859, + "loss_total_v6": 1.366496741771698, + "step": 19 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.004612612612612613, + "grad_norm": 107.79788208007812, + "learning_rate": 1.0000000000000002e-06, + "loss": 20.279420852661133, + "loss_alignment": 0.94677734375, + "loss_alignment_w": 0.473388671875, + "loss_sft": 0.7952805683016777, + "loss_total_v6": 1.2674638032913208, + "step": 20 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.004843243243243243, + "grad_norm": 114.06724548339844, + "learning_rate": 1.0500000000000001e-06, + "loss": 20.764495849609375, + "loss_alignment": 0.93994140625, + "loss_alignment_w": 0.469970703125, + "loss_sft": 0.8268946334719658, + "loss_total_v6": 1.2977808713912964, + "step": 21 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.005073873873873874, + "grad_norm": 113.68041229248047, + "learning_rate": 1.1e-06, + "loss": 20.284202575683594, + "loss_alignment": 0.91357421875, + "loss_alignment_w": 0.456787109375, + "loss_sft": 0.8095412701368332, + "loss_total_v6": 1.2677627205848694, + "step": 22 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.005304504504504505, + "grad_norm": 109.9312973022461, + "learning_rate": 1.1500000000000002e-06, + "loss": 18.805381774902344, + "loss_alignment": 0.90478515625, + "loss_alignment_w": 0.452392578125, + "loss_sft": 0.7210821285843849, + "loss_total_v6": 1.1753363460302353, + "step": 23 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.005535135135135135, + "grad_norm": 112.7437515258789, + "learning_rate": 1.2000000000000002e-06, + "loss": 19.837434768676758, + "loss_alignment": 0.90283203125, + "loss_alignment_w": 0.451416015625, + "loss_sft": 0.7870351895689964, + "loss_total_v6": 1.2398396879434586, + "step": 24 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.005765765765765766, + "grad_norm": 105.73051452636719, + "learning_rate": 1.25e-06, + "loss": 19.116127014160156, + "loss_alignment": 0.91015625, + "loss_alignment_w": 0.455078125, + "loss_sft": 0.7360176593065262, + "loss_total_v6": 1.1947578936815262, + "step": 25 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.005996396396396397, + "grad_norm": 103.85659790039062, + "learning_rate": 1.3e-06, + "loss": 19.036968231201172, + "loss_alignment": 0.91064453125, + "loss_alignment_w": 0.455322265625, + "loss_sft": 0.7320468574762344, + "loss_total_v6": 1.1898104846477509, + "step": 26 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.006227027027027027, + "grad_norm": 91.46253967285156, + "learning_rate": 1.3500000000000002e-06, + "loss": 18.361953735351562, + "loss_alignment": 0.88818359375, + "loss_alignment_w": 0.444091796875, + "loss_sft": 0.7031793296337128, + "loss_total_v6": 1.1476220935583115, + "step": 27 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.006457657657657658, + "grad_norm": 83.3201675415039, + "learning_rate": 1.4000000000000001e-06, + "loss": 18.544090270996094, + "loss_alignment": 0.91357421875, + "loss_alignment_w": 0.456787109375, + "loss_sft": 0.7036224007606506, + "loss_total_v6": 1.159005731344223, + "step": 28 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.006688288288288288, + "grad_norm": 71.37712860107422, + "learning_rate": 1.45e-06, + "loss": 17.84052085876465, + "loss_alignment": 0.89306640625, + "loss_alignment_w": 0.446533203125, + "loss_sft": 0.6703608632087708, + "loss_total_v6": 1.115032508969307, + "step": 29 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.006918918918918919, + "grad_norm": 73.171142578125, + "learning_rate": 1.5e-06, + "loss": 17.87433624267578, + "loss_alignment": 0.87548828125, + "loss_alignment_w": 0.437744140625, + "loss_sft": 0.6796307787299156, + "loss_total_v6": 1.1171460002660751, + "step": 30 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.00714954954954955, + "grad_norm": 62.43604278564453, + "learning_rate": 1.5500000000000002e-06, + "loss": 17.606325149536133, + "loss_alignment": 0.85888671875, + "loss_alignment_w": 0.429443359375, + "loss_sft": 0.6701584458351135, + "loss_total_v6": 1.100395292043686, + "step": 31 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.00738018018018018, + "grad_norm": 64.13766479492188, + "learning_rate": 1.6000000000000001e-06, + "loss": 17.346832275390625, + "loss_alignment": 0.8515625, + "loss_alignment_w": 0.42578125, + "loss_sft": 0.6578770726919174, + "loss_total_v6": 1.0841771215200424, + "step": 32 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.007610810810810811, + "grad_norm": 65.32746124267578, + "learning_rate": 1.6500000000000003e-06, + "loss": 17.510814666748047, + "loss_alignment": 0.876953125, + "loss_alignment_w": 0.4384765625, + "loss_sft": 0.6543929278850555, + "loss_total_v6": 1.0944259092211723, + "step": 33 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.007841441441441442, + "grad_norm": 62.108455657958984, + "learning_rate": 1.7000000000000002e-06, + "loss": 17.2572021484375, + "loss_alignment": 0.85107421875, + "loss_alignment_w": 0.425537109375, + "loss_sft": 0.6527786627411842, + "loss_total_v6": 1.078575186431408, + "step": 34 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.008072072072072072, + "grad_norm": 65.7811279296875, + "learning_rate": 1.75e-06, + "loss": 17.083797454833984, + "loss_alignment": 0.82568359375, + "loss_alignment_w": 0.412841796875, + "loss_sft": 0.654773510992527, + "loss_total_v6": 1.0677373856306076, + "step": 35 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.008302702702702702, + "grad_norm": 53.271148681640625, + "learning_rate": 1.8000000000000001e-06, + "loss": 16.822509765625, + "loss_alignment": 0.8369140625, + "loss_alignment_w": 0.41845703125, + "loss_sft": 0.6324158906936646, + "loss_total_v6": 1.0514069348573685, + "step": 36 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.008533333333333334, + "grad_norm": 53.1756706237793, + "learning_rate": 1.85e-06, + "loss": 16.736557006835938, + "loss_alignment": 0.81298828125, + "loss_alignment_w": 0.406494140625, + "loss_sft": 0.6398458778858185, + "loss_total_v6": 1.0460348278284073, + "step": 37 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.008763963963963964, + "grad_norm": 57.947017669677734, + "learning_rate": 1.9000000000000002e-06, + "loss": 16.53668212890625, + "loss_alignment": 0.79833984375, + "loss_alignment_w": 0.399169921875, + "loss_sft": 0.63449477404356, + "loss_total_v6": 1.0335426181554794, + "step": 38 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.008994594594594594, + "grad_norm": 55.32988739013672, + "learning_rate": 1.9500000000000004e-06, + "loss": 16.978687286376953, + "loss_alignment": 0.81640625, + "loss_alignment_w": 0.408203125, + "loss_sft": 0.652034118771553, + "loss_total_v6": 1.0611680448055267, + "step": 39 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.009225225225225226, + "grad_norm": 59.903892517089844, + "learning_rate": 2.0000000000000003e-06, + "loss": 16.85297203063965, + "loss_alignment": 0.794921875, + "loss_alignment_w": 0.3974609375, + "loss_sft": 0.6567348092794418, + "loss_total_v6": 1.0533107221126556, + "step": 40 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.009455855855855856, + "grad_norm": 54.71234893798828, + "learning_rate": 2.05e-06, + "loss": 16.22533416748047, + "loss_alignment": 0.765625, + "loss_alignment_w": 0.3828125, + "loss_sft": 0.6314692124724388, + "loss_total_v6": 1.0140833854675293, + "step": 41 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.009686486486486486, + "grad_norm": 56.236270904541016, + "learning_rate": 2.1000000000000002e-06, + "loss": 16.08710479736328, + "loss_alignment": 0.75830078125, + "loss_alignment_w": 0.379150390625, + "loss_sft": 0.6268124431371689, + "loss_total_v6": 1.0054440349340439, + "step": 42 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.009917117117117118, + "grad_norm": 55.526702880859375, + "learning_rate": 2.15e-06, + "loss": 15.93226146697998, + "loss_alignment": 0.77099609375, + "loss_alignment_w": 0.385498046875, + "loss_sft": 0.6098563000559807, + "loss_total_v6": 0.99576635658741, + "step": 43 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.010147747747747748, + "grad_norm": 54.65618133544922, + "learning_rate": 2.2e-06, + "loss": 15.9290771484375, + "loss_alignment": 0.759765625, + "loss_alignment_w": 0.3798828125, + "loss_sft": 0.6155471056699753, + "loss_total_v6": 0.9955672696232796, + "step": 44 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.010378378378378378, + "grad_norm": 60.837730407714844, + "learning_rate": 2.25e-06, + "loss": 15.938434600830078, + "loss_alignment": 0.728515625, + "loss_alignment_w": 0.3642578125, + "loss_sft": 0.6308720335364342, + "loss_total_v6": 0.9961521774530411, + "step": 45 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01060900900900901, + "grad_norm": 59.344051361083984, + "learning_rate": 2.3000000000000004e-06, + "loss": 15.700410842895508, + "loss_alignment": 0.7548828125, + "loss_alignment_w": 0.37744140625, + "loss_sft": 0.60342226177454, + "loss_total_v6": 0.9812756255269051, + "step": 46 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01083963963963964, + "grad_norm": 59.58332824707031, + "learning_rate": 2.35e-06, + "loss": 15.634405136108398, + "loss_alignment": 0.75830078125, + "loss_alignment_w": 0.379150390625, + "loss_sft": 0.5985797941684723, + "loss_total_v6": 0.9771503284573555, + "step": 47 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01107027027027027, + "grad_norm": 57.241634368896484, + "learning_rate": 2.4000000000000003e-06, + "loss": 15.597479820251465, + "loss_alignment": 0.728515625, + "loss_alignment_w": 0.3642578125, + "loss_sft": 0.6111797392368317, + "loss_total_v6": 0.9748424738645554, + "step": 48 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.011300900900900902, + "grad_norm": 57.80820846557617, + "learning_rate": 2.4500000000000003e-06, + "loss": 15.392841339111328, + "loss_alignment": 0.71728515625, + "loss_alignment_w": 0.358642578125, + "loss_sft": 0.603471040725708, + "loss_total_v6": 0.9620525687932968, + "step": 49 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.011531531531531532, + "grad_norm": 52.43745040893555, + "learning_rate": 2.5e-06, + "loss": 15.112565040588379, + "loss_alignment": 0.70263671875, + "loss_alignment_w": 0.351318359375, + "loss_sft": 0.5937815457582474, + "loss_total_v6": 0.9445353224873543, + "step": 50 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.011762162162162162, + "grad_norm": 51.884918212890625, + "learning_rate": 2.55e-06, + "loss": 15.171095848083496, + "loss_alignment": 0.6884765625, + "loss_alignment_w": 0.34423828125, + "loss_sft": 0.6048554703593254, + "loss_total_v6": 0.9481935128569603, + "step": 51 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.011992792792792794, + "grad_norm": 50.638633728027344, + "learning_rate": 2.6e-06, + "loss": 14.845616340637207, + "loss_alignment": 0.6787109375, + "loss_alignment_w": 0.33935546875, + "loss_sft": 0.5883276984095573, + "loss_total_v6": 0.927851028740406, + "step": 52 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.012223423423423424, + "grad_norm": 54.81181335449219, + "learning_rate": 2.6500000000000005e-06, + "loss": 15.112927436828613, + "loss_alignment": 0.669921875, + "loss_alignment_w": 0.3349609375, + "loss_sft": 0.6103599816560745, + "loss_total_v6": 0.9445579722523689, + "step": 53 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.012454054054054054, + "grad_norm": 54.93523406982422, + "learning_rate": 2.7000000000000004e-06, + "loss": 14.830133438110352, + "loss_alignment": 0.6875, + "loss_alignment_w": 0.34375, + "loss_sft": 0.5836063399910927, + "loss_total_v6": 0.9268833249807358, + "step": 54 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.012684684684684684, + "grad_norm": 55.31487274169922, + "learning_rate": 2.7500000000000004e-06, + "loss": 14.723230361938477, + "loss_alignment": 0.66064453125, + "loss_alignment_w": 0.330322265625, + "loss_sft": 0.5894218273460865, + "loss_total_v6": 0.9202018976211548, + "step": 55 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.012915315315315316, + "grad_norm": 58.530052185058594, + "learning_rate": 2.8000000000000003e-06, + "loss": 14.992244720458984, + "loss_alignment": 0.67041015625, + "loss_alignment_w": 0.335205078125, + "loss_sft": 0.6017491593956947, + "loss_total_v6": 0.9370152875781059, + "step": 56 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.013145945945945946, + "grad_norm": 54.76213455200195, + "learning_rate": 2.85e-06, + "loss": 14.57115650177002, + "loss_alignment": 0.64892578125, + "loss_alignment_w": 0.324462890625, + "loss_sft": 0.5863869711756706, + "loss_total_v6": 0.9106972962617874, + "step": 57 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.013376576576576576, + "grad_norm": 53.79553985595703, + "learning_rate": 2.9e-06, + "loss": 14.402111053466797, + "loss_alignment": 0.6474609375, + "loss_alignment_w": 0.32373046875, + "loss_sft": 0.5767371729016304, + "loss_total_v6": 0.9001319333910942, + "step": 58 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.013607207207207208, + "grad_norm": 48.28541564941406, + "learning_rate": 2.95e-06, + "loss": 14.342458724975586, + "loss_alignment": 0.6337890625, + "loss_alignment_w": 0.31689453125, + "loss_sft": 0.5796922594308853, + "loss_total_v6": 0.8964036554098129, + "step": 59 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.013837837837837838, + "grad_norm": 54.82633590698242, + "learning_rate": 3e-06, + "loss": 14.663397789001465, + "loss_alignment": 0.65869140625, + "loss_alignment_w": 0.329345703125, + "loss_sft": 0.5864910706877708, + "loss_total_v6": 0.9164623618125916, + "step": 60 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.014068468468468468, + "grad_norm": 50.040985107421875, + "learning_rate": 3.05e-06, + "loss": 14.229324340820312, + "loss_alignment": 0.64111328125, + "loss_alignment_w": 0.320556640625, + "loss_sft": 0.5679216235876083, + "loss_total_v6": 0.8893327414989471, + "step": 61 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0142990990990991, + "grad_norm": 52.577449798583984, + "learning_rate": 3.1000000000000004e-06, + "loss": 14.379213333129883, + "loss_alignment": 0.63525390625, + "loss_alignment_w": 0.317626953125, + "loss_sft": 0.5809823349118233, + "loss_total_v6": 0.8987008407711983, + "step": 62 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01452972972972973, + "grad_norm": 50.401180267333984, + "learning_rate": 3.1500000000000003e-06, + "loss": 14.505417823791504, + "loss_alignment": 0.6201171875, + "loss_alignment_w": 0.31005859375, + "loss_sft": 0.5972013846039772, + "loss_total_v6": 0.9065886214375496, + "step": 63 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01476036036036036, + "grad_norm": 50.46403503417969, + "learning_rate": 3.2000000000000003e-06, + "loss": 13.956527709960938, + "loss_alignment": 0.6259765625, + "loss_alignment_w": 0.31298828125, + "loss_sft": 0.5597982704639435, + "loss_total_v6": 0.872283011674881, + "step": 64 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.014990990990990992, + "grad_norm": 51.50680923461914, + "learning_rate": 3.2500000000000002e-06, + "loss": 13.878986358642578, + "loss_alignment": 0.59375, + "loss_alignment_w": 0.296875, + "loss_sft": 0.5707752071321011, + "loss_total_v6": 0.8674366027116776, + "step": 65 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.015221621621621622, + "grad_norm": 47.66145706176758, + "learning_rate": 3.3000000000000006e-06, + "loss": 13.84383773803711, + "loss_alignment": 0.61279296875, + "loss_alignment_w": 0.306396484375, + "loss_sft": 0.5598656348884106, + "loss_total_v6": 0.8652397990226746, + "step": 66 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.015452252252252252, + "grad_norm": 50.790748596191406, + "learning_rate": 3.3500000000000005e-06, + "loss": 13.994750022888184, + "loss_alignment": 0.6123046875, + "loss_alignment_w": 0.30615234375, + "loss_sft": 0.5688247084617615, + "loss_total_v6": 0.8746718987822533, + "step": 67 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.015682882882882884, + "grad_norm": 66.06356811523438, + "learning_rate": 3.4000000000000005e-06, + "loss": 13.974298477172852, + "loss_alignment": 0.6005859375, + "loss_alignment_w": 0.30029296875, + "loss_sft": 0.5722767040133476, + "loss_total_v6": 0.873393639922142, + "step": 68 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.015913513513513514, + "grad_norm": 51.64221954345703, + "learning_rate": 3.45e-06, + "loss": 13.883466720581055, + "loss_alignment": 0.59619140625, + "loss_alignment_w": 0.298095703125, + "loss_sft": 0.5691937357187271, + "loss_total_v6": 0.8677166998386383, + "step": 69 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.016144144144144144, + "grad_norm": 48.372276306152344, + "learning_rate": 3.5e-06, + "loss": 13.633970260620117, + "loss_alignment": 0.599609375, + "loss_alignment_w": 0.2998046875, + "loss_sft": 0.5524557828903198, + "loss_total_v6": 0.8521231338381767, + "step": 70 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.016374774774774774, + "grad_norm": 47.52553939819336, + "learning_rate": 3.5500000000000003e-06, + "loss": 13.452033996582031, + "loss_alignment": 0.57470703125, + "loss_alignment_w": 0.287353515625, + "loss_sft": 0.5532002374529839, + "loss_total_v6": 0.8407521098852158, + "step": 71 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.016605405405405404, + "grad_norm": 56.31544876098633, + "learning_rate": 3.6000000000000003e-06, + "loss": 13.383362770080566, + "loss_alignment": 0.57666015625, + "loss_alignment_w": 0.288330078125, + "loss_sft": 0.5491982214152813, + "loss_total_v6": 0.8364601731300354, + "step": 72 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.016836036036036037, + "grad_norm": 47.09490966796875, + "learning_rate": 3.65e-06, + "loss": 13.170068740844727, + "loss_alignment": 0.543701171875, + "loss_alignment_w": 0.2718505859375, + "loss_sft": 0.551644891500473, + "loss_total_v6": 0.8231293112039566, + "step": 73 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.017066666666666667, + "grad_norm": 47.69663619995117, + "learning_rate": 3.7e-06, + "loss": 13.69873332977295, + "loss_alignment": 0.5888671875, + "loss_alignment_w": 0.29443359375, + "loss_sft": 0.5623781196773052, + "loss_total_v6": 0.8561708480119705, + "step": 74 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.017297297297297298, + "grad_norm": 46.242042541503906, + "learning_rate": 3.7500000000000005e-06, + "loss": 13.28950309753418, + "loss_alignment": 0.55029296875, + "loss_alignment_w": 0.275146484375, + "loss_sft": 0.5554017126560211, + "loss_total_v6": 0.8305939584970474, + "step": 75 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.017527927927927928, + "grad_norm": 45.41549301147461, + "learning_rate": 3.8000000000000005e-06, + "loss": 13.568946838378906, + "loss_alignment": 0.58349609375, + "loss_alignment_w": 0.291748046875, + "loss_sft": 0.5561890006065369, + "loss_total_v6": 0.8480591252446175, + "step": 76 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.017758558558558558, + "grad_norm": 49.87102508544922, + "learning_rate": 3.85e-06, + "loss": 13.664475440979004, + "loss_alignment": 0.5703125, + "loss_alignment_w": 0.28515625, + "loss_sft": 0.5690108127892017, + "loss_total_v6": 0.8540297001600266, + "step": 77 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.017989189189189188, + "grad_norm": 76.67233276367188, + "learning_rate": 3.900000000000001e-06, + "loss": 14.091609954833984, + "loss_alignment": 0.5703125, + "loss_alignment_w": 0.28515625, + "loss_sft": 0.5954930558800697, + "loss_total_v6": 0.8807256296277046, + "step": 78 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01821981981981982, + "grad_norm": 43.54582595825195, + "learning_rate": 3.95e-06, + "loss": 13.39872932434082, + "loss_alignment": 0.5537109375, + "loss_alignment_w": 0.27685546875, + "loss_sft": 0.560519315302372, + "loss_total_v6": 0.8374205604195595, + "step": 79 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01845045045045045, + "grad_norm": 48.182456970214844, + "learning_rate": 4.000000000000001e-06, + "loss": 13.776968002319336, + "loss_alignment": 0.53564453125, + "loss_alignment_w": 0.267822265625, + "loss_sft": 0.5933603420853615, + "loss_total_v6": 0.8610605299472809, + "step": 80 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01868108108108108, + "grad_norm": 43.823211669921875, + "learning_rate": 4.05e-06, + "loss": 13.126890182495117, + "loss_alignment": 0.55322265625, + "loss_alignment_w": 0.276611328125, + "loss_sft": 0.54389563575387, + "loss_total_v6": 0.8204306289553642, + "step": 81 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01891171171171171, + "grad_norm": 50.35432052612305, + "learning_rate": 4.1e-06, + "loss": 13.397335052490234, + "loss_alignment": 0.53515625, + "loss_alignment_w": 0.267578125, + "loss_sft": 0.569144956767559, + "loss_total_v6": 0.837333433330059, + "step": 82 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.01914234234234234, + "grad_norm": 58.17121124267578, + "learning_rate": 4.15e-06, + "loss": 12.820657730102539, + "loss_alignment": 0.55224609375, + "loss_alignment_w": 0.276123046875, + "loss_sft": 0.5251527801156044, + "loss_total_v6": 0.8012911006808281, + "step": 83 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.019372972972972972, + "grad_norm": 51.290016174316406, + "learning_rate": 4.2000000000000004e-06, + "loss": 13.01709270477295, + "loss_alignment": 0.533203125, + "loss_alignment_w": 0.2666015625, + "loss_sft": 0.5470735803246498, + "loss_total_v6": 0.8135683238506317, + "step": 84 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.019603603603603605, + "grad_norm": 53.828189849853516, + "learning_rate": 4.25e-06, + "loss": 12.994726181030273, + "loss_alignment": 0.55322265625, + "loss_alignment_w": 0.276611328125, + "loss_sft": 0.5358490496873856, + "loss_total_v6": 0.8121704384684563, + "step": 85 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.019834234234234235, + "grad_norm": 46.14897537231445, + "learning_rate": 4.3e-06, + "loss": 13.48746109008789, + "loss_alignment": 0.5341796875, + "loss_alignment_w": 0.26708984375, + "loss_sft": 0.5763647705316544, + "loss_total_v6": 0.8429663255810738, + "step": 86 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.020064864864864865, + "grad_norm": 52.87785720825195, + "learning_rate": 4.350000000000001e-06, + "loss": 13.060661315917969, + "loss_alignment": 0.51318359375, + "loss_alignment_w": 0.256591796875, + "loss_sft": 0.5592570826411247, + "loss_total_v6": 0.816291369497776, + "step": 87 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.020295495495495496, + "grad_norm": 53.25971221923828, + "learning_rate": 4.4e-06, + "loss": 12.390493392944336, + "loss_alignment": 0.53662109375, + "loss_alignment_w": 0.268310546875, + "loss_sft": 0.5066293589770794, + "loss_total_v6": 0.7744058445096016, + "step": 88 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.020526126126126126, + "grad_norm": 42.14784622192383, + "learning_rate": 4.450000000000001e-06, + "loss": 12.961808204650879, + "loss_alignment": 0.521484375, + "loss_alignment_w": 0.2607421875, + "loss_sft": 0.5492182150483131, + "loss_total_v6": 0.8101129978895187, + "step": 89 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.020756756756756756, + "grad_norm": 43.611297607421875, + "learning_rate": 4.5e-06, + "loss": 12.91300106048584, + "loss_alignment": 0.517333984375, + "loss_alignment_w": 0.2586669921875, + "loss_sft": 0.548700712621212, + "loss_total_v6": 0.8070625439286232, + "step": 90 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.020987387387387386, + "grad_norm": 51.5624885559082, + "learning_rate": 4.5500000000000005e-06, + "loss": 13.209714889526367, + "loss_alignment": 0.515869140625, + "loss_alignment_w": 0.2579345703125, + "loss_sft": 0.5679319798946381, + "loss_total_v6": 0.8256071731448174, + "step": 91 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02121801801801802, + "grad_norm": 59.60592269897461, + "learning_rate": 4.600000000000001e-06, + "loss": 13.005121231079102, + "loss_alignment": 0.515625, + "loss_alignment_w": 0.2578125, + "loss_sft": 0.5552517548203468, + "loss_total_v6": 0.8128200694918633, + "step": 92 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02144864864864865, + "grad_norm": 49.69599151611328, + "learning_rate": 4.65e-06, + "loss": 12.309629440307617, + "loss_alignment": 0.51708984375, + "loss_alignment_w": 0.258544921875, + "loss_sft": 0.510989997535944, + "loss_total_v6": 0.7693518176674843, + "step": 93 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02167927927927928, + "grad_norm": 49.74225997924805, + "learning_rate": 4.7e-06, + "loss": 12.934028625488281, + "loss_alignment": 0.5068359375, + "loss_alignment_w": 0.25341796875, + "loss_sft": 0.5541348159313202, + "loss_total_v6": 0.808376781642437, + "step": 94 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02190990990990991, + "grad_norm": 49.871036529541016, + "learning_rate": 4.75e-06, + "loss": 12.569002151489258, + "loss_alignment": 0.518310546875, + "loss_alignment_w": 0.2591552734375, + "loss_sft": 0.5267430171370506, + "loss_total_v6": 0.7855625972151756, + "step": 95 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02214054054054054, + "grad_norm": 63.07365417480469, + "learning_rate": 4.800000000000001e-06, + "loss": 12.859170913696289, + "loss_alignment": 0.505859375, + "loss_alignment_w": 0.2529296875, + "loss_sft": 0.5512568093836308, + "loss_total_v6": 0.803698219358921, + "step": 96 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02237117117117117, + "grad_norm": 43.5764045715332, + "learning_rate": 4.85e-06, + "loss": 12.707586288452148, + "loss_alignment": 0.513916015625, + "loss_alignment_w": 0.2569580078125, + "loss_sft": 0.5370829999446869, + "loss_total_v6": 0.7942241281270981, + "step": 97 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.022601801801801803, + "grad_norm": 46.449127197265625, + "learning_rate": 4.9000000000000005e-06, + "loss": 12.633597373962402, + "loss_alignment": 0.509765625, + "loss_alignment_w": 0.2548828125, + "loss_sft": 0.5340608954429626, + "loss_total_v6": 0.7895998656749725, + "step": 98 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.022832432432432433, + "grad_norm": 44.30062484741211, + "learning_rate": 4.95e-06, + "loss": 12.598287582397461, + "loss_alignment": 0.50732421875, + "loss_alignment_w": 0.253662109375, + "loss_sft": 0.5341734103858471, + "loss_total_v6": 0.7873930037021637, + "step": 99 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.023063063063063063, + "grad_norm": 44.36838912963867, + "learning_rate": 5e-06, + "loss": 12.96684455871582, + "loss_alignment": 0.49853515625, + "loss_alignment_w": 0.249267578125, + "loss_sft": 0.561205979436636, + "loss_total_v6": 0.810427762567997, + "step": 100 + }, + { + "epoch": 0.023063063063063063, + "eval_loss": 0.743026614189148, + "eval_loss_alignment": 0.4788620125534188, + "eval_loss_alignment_w": 0.2394310062767094, + "eval_loss_sft": 0.5035955500828787, + "eval_loss_total_v6": 0.7430265573545908, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 84.8255, + "eval_samples_per_second": 21.998, + "eval_steps_per_second": 2.759, + "step": 100 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.023293693693693693, + "grad_norm": 39.536006927490234, + "learning_rate": 5.050000000000001e-06, + "loss": 12.956521034240723, + "loss_alignment": 0.5009765625, + "loss_alignment_w": 0.25048828125, + "loss_sft": 0.5594163239002228, + "loss_total_v6": 0.8097825571894646, + "step": 101 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.023524324324324324, + "grad_norm": 40.17134094238281, + "learning_rate": 5.1e-06, + "loss": 12.855351448059082, + "loss_alignment": 0.50537109375, + "loss_alignment_w": 0.252685546875, + "loss_sft": 0.5510333031415939, + "loss_total_v6": 0.8034594729542732, + "step": 102 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.023754954954954954, + "grad_norm": 47.4295539855957, + "learning_rate": 5.150000000000001e-06, + "loss": 13.010074615478516, + "loss_alignment": 0.4970703125, + "loss_alignment_w": 0.24853515625, + "loss_sft": 0.5645487606525421, + "loss_total_v6": 0.8131297081708908, + "step": 103 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.023985585585585587, + "grad_norm": 45.52815628051758, + "learning_rate": 5.2e-06, + "loss": 12.109628677368164, + "loss_alignment": 0.486572265625, + "loss_alignment_w": 0.2432861328125, + "loss_sft": 0.5139472037553787, + "loss_total_v6": 0.756851851940155, + "step": 104 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.024216216216216217, + "grad_norm": 40.882694244384766, + "learning_rate": 5.2500000000000006e-06, + "loss": 12.26878547668457, + "loss_alignment": 0.4765625, + "loss_alignment_w": 0.23828125, + "loss_sft": 0.5284110568463802, + "loss_total_v6": 0.766799084842205, + "step": 105 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.024446846846846847, + "grad_norm": 42.00320053100586, + "learning_rate": 5.300000000000001e-06, + "loss": 12.346090316772461, + "loss_alignment": 0.494140625, + "loss_alignment_w": 0.2470703125, + "loss_sft": 0.5244991891086102, + "loss_total_v6": 0.7716305926442146, + "step": 106 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.024677477477477477, + "grad_norm": 42.446773529052734, + "learning_rate": 5.3500000000000004e-06, + "loss": 12.685343742370605, + "loss_alignment": 0.5029296875, + "loss_alignment_w": 0.25146484375, + "loss_sft": 0.5415369719266891, + "loss_total_v6": 0.792833961546421, + "step": 107 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.024908108108108108, + "grad_norm": 46.0847282409668, + "learning_rate": 5.400000000000001e-06, + "loss": 12.455568313598633, + "loss_alignment": 0.490966796875, + "loss_alignment_w": 0.2454833984375, + "loss_sft": 0.5323640294373035, + "loss_total_v6": 0.7784730494022369, + "step": 108 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.025138738738738738, + "grad_norm": 40.32667541503906, + "learning_rate": 5.450000000000001e-06, + "loss": 12.269501686096191, + "loss_alignment": 0.4697265625, + "loss_alignment_w": 0.23486328125, + "loss_sft": 0.5317822247743607, + "loss_total_v6": 0.766843855381012, + "step": 109 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.025369369369369368, + "grad_norm": 38.9453010559082, + "learning_rate": 5.500000000000001e-06, + "loss": 12.465185165405273, + "loss_alignment": 0.477783203125, + "loss_alignment_w": 0.2388916015625, + "loss_sft": 0.5396789461374283, + "loss_total_v6": 0.7790740579366684, + "step": 110 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0256, + "grad_norm": 36.894859313964844, + "learning_rate": 5.550000000000001e-06, + "loss": 12.679071426391602, + "loss_alignment": 0.493408203125, + "loss_alignment_w": 0.2467041015625, + "loss_sft": 0.5454174056649208, + "loss_total_v6": 0.7924419343471527, + "step": 111 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02583063063063063, + "grad_norm": 41.667972564697266, + "learning_rate": 5.600000000000001e-06, + "loss": 11.9881591796875, + "loss_alignment": 0.4697265625, + "loss_alignment_w": 0.23486328125, + "loss_sft": 0.5139083825051785, + "loss_total_v6": 0.7492599338293076, + "step": 112 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02606126126126126, + "grad_norm": 42.74580383300781, + "learning_rate": 5.65e-06, + "loss": 12.296443939208984, + "loss_alignment": 0.46533203125, + "loss_alignment_w": 0.232666015625, + "loss_sft": 0.5353276766836643, + "loss_total_v6": 0.7685277685523033, + "step": 113 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02629189189189189, + "grad_norm": 42.454627990722656, + "learning_rate": 5.7e-06, + "loss": 12.129032135009766, + "loss_alignment": 0.48193359375, + "loss_alignment_w": 0.240966796875, + "loss_sft": 0.5170061439275742, + "loss_total_v6": 0.7580644935369492, + "step": 114 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02652252252252252, + "grad_norm": 43.30904769897461, + "learning_rate": 5.75e-06, + "loss": 12.44132137298584, + "loss_alignment": 0.475830078125, + "loss_alignment_w": 0.2379150390625, + "loss_sft": 0.5397438257932663, + "loss_total_v6": 0.7775825932621956, + "step": 115 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02675315315315315, + "grad_norm": 41.56555938720703, + "learning_rate": 5.8e-06, + "loss": 11.79411506652832, + "loss_alignment": 0.46435546875, + "loss_alignment_w": 0.232177734375, + "loss_sft": 0.5047713667154312, + "loss_total_v6": 0.7371321991086006, + "step": 116 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.026983783783783785, + "grad_norm": 42.184547424316406, + "learning_rate": 5.85e-06, + "loss": 12.431295394897461, + "loss_alignment": 0.4609375, + "loss_alignment_w": 0.23046875, + "loss_sft": 0.5462582595646381, + "loss_total_v6": 0.7769559100270271, + "step": 117 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.027214414414414415, + "grad_norm": 44.23741912841797, + "learning_rate": 5.9e-06, + "loss": 12.23984432220459, + "loss_alignment": 0.480224609375, + "loss_alignment_w": 0.2401123046875, + "loss_sft": 0.524816956371069, + "loss_total_v6": 0.7649902626872063, + "step": 118 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.027445045045045045, + "grad_norm": 40.500911712646484, + "learning_rate": 5.950000000000001e-06, + "loss": 12.118008613586426, + "loss_alignment": 0.467041015625, + "loss_alignment_w": 0.2335205078125, + "loss_sft": 0.5237787142395973, + "loss_total_v6": 0.7573755159974098, + "step": 119 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.027675675675675675, + "grad_norm": 45.34736633300781, + "learning_rate": 6e-06, + "loss": 12.116660118103027, + "loss_alignment": 0.478759765625, + "loss_alignment_w": 0.2393798828125, + "loss_sft": 0.5183080919086933, + "loss_total_v6": 0.7572912499308586, + "step": 120 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.027906306306306305, + "grad_norm": 40.128292083740234, + "learning_rate": 6.0500000000000005e-06, + "loss": 12.339106559753418, + "loss_alignment": 0.464111328125, + "loss_alignment_w": 0.2320556640625, + "loss_sft": 0.5397641323506832, + "loss_total_v6": 0.7711941599845886, + "step": 121 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.028136936936936936, + "grad_norm": 42.65806579589844, + "learning_rate": 6.1e-06, + "loss": 12.611117362976074, + "loss_alignment": 0.469482421875, + "loss_alignment_w": 0.2347412109375, + "loss_sft": 0.553987666964531, + "loss_total_v6": 0.7881948351860046, + "step": 122 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02836756756756757, + "grad_norm": 38.82716369628906, + "learning_rate": 6.15e-06, + "loss": 12.297222137451172, + "loss_alignment": 0.460693359375, + "loss_alignment_w": 0.2303466796875, + "loss_sft": 0.5380007587373257, + "loss_total_v6": 0.7685763463377953, + "step": 123 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0285981981981982, + "grad_norm": 40.30464172363281, + "learning_rate": 6.200000000000001e-06, + "loss": 12.194438934326172, + "loss_alignment": 0.489990234375, + "loss_alignment_w": 0.2449951171875, + "loss_sft": 0.5174472667276859, + "loss_total_v6": 0.7621524780988693, + "step": 124 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02882882882882883, + "grad_norm": 40.28455352783203, + "learning_rate": 6.25e-06, + "loss": 12.016544342041016, + "loss_alignment": 0.479248046875, + "loss_alignment_w": 0.2396240234375, + "loss_sft": 0.511226911097765, + "loss_total_v6": 0.7510340288281441, + "step": 125 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02905945945945946, + "grad_norm": 35.46963119506836, + "learning_rate": 6.300000000000001e-06, + "loss": 12.349624633789062, + "loss_alignment": 0.480224609375, + "loss_alignment_w": 0.2401123046875, + "loss_sft": 0.5316782370209694, + "loss_total_v6": 0.771851547062397, + "step": 126 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02929009009009009, + "grad_norm": 37.68815994262695, + "learning_rate": 6.35e-06, + "loss": 11.95141315460205, + "loss_alignment": 0.47265625, + "loss_alignment_w": 0.236328125, + "loss_sft": 0.5105436630547047, + "loss_total_v6": 0.7469633519649506, + "step": 127 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.02952072072072072, + "grad_norm": 37.99477767944336, + "learning_rate": 6.4000000000000006e-06, + "loss": 12.362911224365234, + "loss_alignment": 0.472900390625, + "loss_alignment_w": 0.2364501953125, + "loss_sft": 0.5359875597059727, + "loss_total_v6": 0.7726819291710854, + "step": 128 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.029751351351351353, + "grad_norm": 38.81325149536133, + "learning_rate": 6.450000000000001e-06, + "loss": 11.645975112915039, + "loss_alignment": 0.478759765625, + "loss_alignment_w": 0.2393798828125, + "loss_sft": 0.48850879818201065, + "loss_total_v6": 0.7278734371066093, + "step": 129 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.029981981981981983, + "grad_norm": 35.971923828125, + "learning_rate": 6.5000000000000004e-06, + "loss": 11.922149658203125, + "loss_alignment": 0.463623046875, + "loss_alignment_w": 0.2318115234375, + "loss_sft": 0.5130176432430744, + "loss_total_v6": 0.7451343312859535, + "step": 130 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.030212612612612613, + "grad_norm": 36.924007415771484, + "learning_rate": 6.550000000000001e-06, + "loss": 11.513376235961914, + "loss_alignment": 0.47265625, + "loss_alignment_w": 0.236328125, + "loss_sft": 0.484097171574831, + "loss_total_v6": 0.7195860296487808, + "step": 131 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.030443243243243243, + "grad_norm": 41.05062484741211, + "learning_rate": 6.600000000000001e-06, + "loss": 11.745655059814453, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.507373072206974, + "loss_total_v6": 0.7341034114360809, + "step": 132 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.030673873873873873, + "grad_norm": 36.97251510620117, + "learning_rate": 6.650000000000001e-06, + "loss": 12.078536987304688, + "loss_alignment": 0.4716796875, + "loss_alignment_w": 0.23583984375, + "loss_sft": 0.5191754922270775, + "loss_total_v6": 0.7549085542559624, + "step": 133 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.030904504504504503, + "grad_norm": 38.44514846801758, + "learning_rate": 6.700000000000001e-06, + "loss": 11.937063217163086, + "loss_alignment": 0.46484375, + "loss_alignment_w": 0.232421875, + "loss_sft": 0.5136750563979149, + "loss_total_v6": 0.7460663989186287, + "step": 134 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.031135135135135134, + "grad_norm": 38.62439727783203, + "learning_rate": 6.750000000000001e-06, + "loss": 12.222555160522461, + "loss_alignment": 0.464111328125, + "loss_alignment_w": 0.2320556640625, + "loss_sft": 0.5320066027343273, + "loss_total_v6": 0.7639096826314926, + "step": 135 + }, + { + "action_cond_aligned": 0.8984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03136576576576577, + "grad_norm": 36.275184631347656, + "learning_rate": 6.800000000000001e-06, + "loss": 12.064521789550781, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.5330243073403835, + "loss_total_v6": 0.754032589495182, + "step": 136 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.031596396396396394, + "grad_norm": 36.634830474853516, + "learning_rate": 6.850000000000001e-06, + "loss": 11.749494552612305, + "loss_alignment": 0.470947265625, + "loss_alignment_w": 0.2354736328125, + "loss_sft": 0.49856453016400337, + "loss_total_v6": 0.7343433573842049, + "step": 137 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03182702702702703, + "grad_norm": 37.161563873291016, + "learning_rate": 6.9e-06, + "loss": 11.753988265991211, + "loss_alignment": 0.45556640625, + "loss_alignment_w": 0.227783203125, + "loss_sft": 0.5072835572063923, + "loss_total_v6": 0.7346242517232895, + "step": 138 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03205765765765766, + "grad_norm": 33.2685661315918, + "learning_rate": 6.95e-06, + "loss": 12.111741065979004, + "loss_alignment": 0.4814453125, + "loss_alignment_w": 0.24072265625, + "loss_sft": 0.5162611790001392, + "loss_total_v6": 0.7569838389754295, + "step": 139 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03228828828828829, + "grad_norm": 41.02652359008789, + "learning_rate": 7e-06, + "loss": 11.511321067810059, + "loss_alignment": 0.461181640625, + "loss_alignment_w": 0.2305908203125, + "loss_sft": 0.48894306644797325, + "loss_total_v6": 0.719457596540451, + "step": 140 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03251891891891892, + "grad_norm": 37.82137680053711, + "learning_rate": 7.05e-06, + "loss": 11.774141311645508, + "loss_alignment": 0.4814453125, + "loss_alignment_w": 0.24072265625, + "loss_sft": 0.49529851227998734, + "loss_total_v6": 0.7358838617801666, + "step": 141 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03274954954954955, + "grad_norm": 37.70842361450195, + "learning_rate": 7.100000000000001e-06, + "loss": 12.36075210571289, + "loss_alignment": 0.47607421875, + "loss_alignment_w": 0.238037109375, + "loss_sft": 0.5348455756902695, + "loss_total_v6": 0.7725469768047333, + "step": 142 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03298018018018018, + "grad_norm": 33.69681930541992, + "learning_rate": 7.15e-06, + "loss": 12.177348136901855, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.5339877158403397, + "loss_total_v6": 0.7610842660069466, + "step": 143 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03321081081081081, + "grad_norm": 35.713539123535156, + "learning_rate": 7.2000000000000005e-06, + "loss": 11.442316055297852, + "loss_alignment": 0.457275390625, + "loss_alignment_w": 0.2286376953125, + "loss_sft": 0.4862629435956478, + "loss_total_v6": 0.7151447758078575, + "step": 144 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03344144144144144, + "grad_norm": 33.82164001464844, + "learning_rate": 7.25e-06, + "loss": 12.139039993286133, + "loss_alignment": 0.468994140625, + "loss_alignment_w": 0.2344970703125, + "loss_sft": 0.5241166539490223, + "loss_total_v6": 0.7586900144815445, + "step": 145 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.033672072072072075, + "grad_norm": 38.01202392578125, + "learning_rate": 7.3e-06, + "loss": 11.613012313842773, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.5024398043751717, + "loss_total_v6": 0.725813239812851, + "step": 146 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0339027027027027, + "grad_norm": 33.07648468017578, + "learning_rate": 7.350000000000001e-06, + "loss": 11.6666259765625, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.5099410936236382, + "loss_total_v6": 0.7291641086339951, + "step": 147 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.034133333333333335, + "grad_norm": 34.56536865234375, + "learning_rate": 7.4e-06, + "loss": 11.985418319702148, + "loss_alignment": 0.464599609375, + "loss_alignment_w": 0.2322998046875, + "loss_sft": 0.5166363082826138, + "loss_total_v6": 0.7490886822342873, + "step": 148 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03436396396396396, + "grad_norm": 33.70107650756836, + "learning_rate": 7.450000000000001e-06, + "loss": 12.289119720458984, + "loss_alignment": 0.461669921875, + "loss_alignment_w": 0.2308349609375, + "loss_sft": 0.53757069632411, + "loss_total_v6": 0.7680699676275253, + "step": 149 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.034594594594594595, + "grad_norm": 36.353214263916016, + "learning_rate": 7.500000000000001e-06, + "loss": 11.395709991455078, + "loss_alignment": 0.472900390625, + "loss_alignment_w": 0.2364501953125, + "loss_sft": 0.4756748788058758, + "loss_total_v6": 0.7122318595647812, + "step": 150 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03482522522522523, + "grad_norm": 34.330772399902344, + "learning_rate": 7.5500000000000006e-06, + "loss": 12.26359748840332, + "loss_alignment": 0.478271484375, + "loss_alignment_w": 0.2391357421875, + "loss_sft": 0.5269576236605644, + "loss_total_v6": 0.7664748355746269, + "step": 151 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.035055855855855855, + "grad_norm": 37.2525749206543, + "learning_rate": 7.600000000000001e-06, + "loss": 11.756328582763672, + "loss_alignment": 0.46630859375, + "loss_alignment_w": 0.233154296875, + "loss_sft": 0.5016162879765034, + "loss_total_v6": 0.7347705513238907, + "step": 152 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03528648648648649, + "grad_norm": 34.60265350341797, + "learning_rate": 7.650000000000001e-06, + "loss": 12.250591278076172, + "loss_alignment": 0.46826171875, + "loss_alignment_w": 0.234130859375, + "loss_sft": 0.5311801843345165, + "loss_total_v6": 0.7656619772315025, + "step": 153 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.035517117117117115, + "grad_norm": 38.41831970214844, + "learning_rate": 7.7e-06, + "loss": 12.23784351348877, + "loss_alignment": 0.47021484375, + "loss_alignment_w": 0.235107421875, + "loss_sft": 0.5302918031811714, + "loss_total_v6": 0.7648652121424675, + "step": 154 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03574774774774775, + "grad_norm": 40.565162658691406, + "learning_rate": 7.75e-06, + "loss": 11.908676147460938, + "loss_alignment": 0.47119140625, + "loss_alignment_w": 0.235595703125, + "loss_sft": 0.5088644064962864, + "loss_total_v6": 0.7442922368645668, + "step": 155 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.035978378378378376, + "grad_norm": 33.9465217590332, + "learning_rate": 7.800000000000002e-06, + "loss": 11.713268280029297, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.5086753852665424, + "loss_total_v6": 0.7320792824029922, + "step": 156 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03620900900900901, + "grad_norm": 33.83683395385742, + "learning_rate": 7.850000000000001e-06, + "loss": 11.7782564163208, + "loss_alignment": 0.45556640625, + "loss_alignment_w": 0.227783203125, + "loss_sft": 0.5084952041506767, + "loss_total_v6": 0.7361410558223724, + "step": 157 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03643963963963964, + "grad_norm": 29.896337509155273, + "learning_rate": 7.9e-06, + "loss": 12.20293140411377, + "loss_alignment": 0.47412109375, + "loss_alignment_w": 0.237060546875, + "loss_sft": 0.5261719487607479, + "loss_total_v6": 0.76268320530653, + "step": 158 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03667027027027027, + "grad_norm": 32.35505294799805, + "learning_rate": 7.950000000000002e-06, + "loss": 11.776859283447266, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.5111391432583332, + "loss_total_v6": 0.7360536828637123, + "step": 159 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0369009009009009, + "grad_norm": 33.1829719543457, + "learning_rate": 8.000000000000001e-06, + "loss": 11.9539213180542, + "loss_alignment": 0.467529296875, + "loss_alignment_w": 0.2337646484375, + "loss_sft": 0.5128366835415363, + "loss_total_v6": 0.7471201047301292, + "step": 160 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03713153153153153, + "grad_norm": 30.750667572021484, + "learning_rate": 8.050000000000001e-06, + "loss": 12.059009552001953, + "loss_alignment": 0.469970703125, + "loss_alignment_w": 0.2349853515625, + "loss_sft": 0.5188858360052109, + "loss_total_v6": 0.7536880970001221, + "step": 161 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03736216216216216, + "grad_norm": 34.45330047607422, + "learning_rate": 8.1e-06, + "loss": 11.67724609375, + "loss_alignment": 0.461669921875, + "loss_alignment_w": 0.2308349609375, + "loss_sft": 0.4990997314453125, + "loss_total_v6": 0.7298278883099556, + "step": 162 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03759279279279279, + "grad_norm": 32.197418212890625, + "learning_rate": 8.15e-06, + "loss": 12.213680267333984, + "loss_alignment": 0.466064453125, + "loss_alignment_w": 0.2330322265625, + "loss_sft": 0.5303838439285755, + "loss_total_v6": 0.763355053961277, + "step": 163 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03782342342342342, + "grad_norm": 35.55005645751953, + "learning_rate": 8.2e-06, + "loss": 12.137815475463867, + "loss_alignment": 0.468505859375, + "loss_alignment_w": 0.2342529296875, + "loss_sft": 0.5240553431212902, + "loss_total_v6": 0.7586134523153305, + "step": 164 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03805405405405406, + "grad_norm": 32.83258819580078, + "learning_rate": 8.25e-06, + "loss": 11.873748779296875, + "loss_alignment": 0.4501953125, + "loss_alignment_w": 0.22509765625, + "loss_sft": 0.5173778459429741, + "loss_total_v6": 0.7421092763543129, + "step": 165 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03828468468468468, + "grad_norm": 34.56532287597656, + "learning_rate": 8.3e-06, + "loss": 11.7706298828125, + "loss_alignment": 0.466064453125, + "loss_alignment_w": 0.2330322265625, + "loss_sft": 0.5024184882640839, + "loss_total_v6": 0.7356643304228783, + "step": 166 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03851531531531532, + "grad_norm": 32.047325134277344, + "learning_rate": 8.35e-06, + "loss": 11.256280899047852, + "loss_alignment": 0.47216796875, + "loss_alignment_w": 0.236083984375, + "loss_sft": 0.46754034981131554, + "loss_total_v6": 0.7035175263881683, + "step": 167 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.038745945945945943, + "grad_norm": 40.5429573059082, + "learning_rate": 8.400000000000001e-06, + "loss": 12.265466690063477, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.5390068143606186, + "loss_total_v6": 0.7665916457772255, + "step": 168 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03897657657657658, + "grad_norm": 33.70920181274414, + "learning_rate": 8.45e-06, + "loss": 11.663644790649414, + "loss_alignment": 0.47607421875, + "loss_alignment_w": 0.238037109375, + "loss_sft": 0.49138318374753, + "loss_total_v6": 0.7289777919650078, + "step": 169 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03920720720720721, + "grad_norm": 34.16844940185547, + "learning_rate": 8.5e-06, + "loss": 12.31064224243164, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.5418913215398788, + "loss_total_v6": 0.7694151103496552, + "step": 170 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03943783783783784, + "grad_norm": 34.93946075439453, + "learning_rate": 8.550000000000001e-06, + "loss": 11.527324676513672, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4994189664721489, + "loss_total_v6": 0.7204578071832657, + "step": 171 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.03966846846846847, + "grad_norm": 35.49284362792969, + "learning_rate": 8.6e-06, + "loss": 11.746993064880371, + "loss_alignment": 0.467041015625, + "loss_alignment_w": 0.2335205078125, + "loss_sft": 0.5005749985575676, + "loss_total_v6": 0.7341870442032814, + "step": 172 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0398990990990991, + "grad_norm": 34.15147018432617, + "learning_rate": 8.65e-06, + "loss": 11.381599426269531, + "loss_alignment": 0.46484375, + "loss_alignment_w": 0.232421875, + "loss_sft": 0.4793248027563095, + "loss_total_v6": 0.7113499268889427, + "step": 173 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04012972972972973, + "grad_norm": 31.669206619262695, + "learning_rate": 8.700000000000001e-06, + "loss": 11.988602638244629, + "loss_alignment": 0.4560546875, + "loss_alignment_w": 0.22802734375, + "loss_sft": 0.5223589390516281, + "loss_total_v6": 0.7492876797914505, + "step": 174 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04036036036036036, + "grad_norm": 34.38711166381836, + "learning_rate": 8.750000000000001e-06, + "loss": 11.656688690185547, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.5055815987288952, + "loss_total_v6": 0.7285430431365967, + "step": 175 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04059099099099099, + "grad_norm": 31.036054611206055, + "learning_rate": 8.8e-06, + "loss": 11.990935325622559, + "loss_alignment": 0.46484375, + "loss_alignment_w": 0.232421875, + "loss_sft": 0.516675915569067, + "loss_total_v6": 0.7494334802031517, + "step": 176 + }, + { + "action_cond_aligned": 0.8984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.040821621621621625, + "grad_norm": 37.190528869628906, + "learning_rate": 8.85e-06, + "loss": 11.248697280883789, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4876810126006603, + "loss_total_v6": 0.7030435651540756, + "step": 177 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04105225225225225, + "grad_norm": 29.931453704833984, + "learning_rate": 8.900000000000001e-06, + "loss": 11.848845481872559, + "loss_alignment": 0.468017578125, + "loss_alignment_w": 0.2340087890625, + "loss_sft": 0.5063761733472347, + "loss_total_v6": 0.7405528426170349, + "step": 178 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.041282882882882885, + "grad_norm": 34.52415084838867, + "learning_rate": 8.95e-06, + "loss": 11.83298110961914, + "loss_alignment": 0.458984375, + "loss_alignment_w": 0.2294921875, + "loss_sft": 0.5101301483809948, + "loss_total_v6": 0.7395612969994545, + "step": 179 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04151351351351351, + "grad_norm": 33.04740524291992, + "learning_rate": 9e-06, + "loss": 12.34731674194336, + "loss_alignment": 0.4716796875, + "loss_alignment_w": 0.23583984375, + "loss_sft": 0.5358216725289822, + "loss_total_v6": 0.7717072740197182, + "step": 180 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.041744144144144145, + "grad_norm": 33.22628402709961, + "learning_rate": 9.050000000000001e-06, + "loss": 11.599200248718262, + "loss_alignment": 0.458740234375, + "loss_alignment_w": 0.2293701171875, + "loss_sft": 0.49547310546040535, + "loss_total_v6": 0.7249500304460526, + "step": 181 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04197477477477477, + "grad_norm": 32.126094818115234, + "learning_rate": 9.100000000000001e-06, + "loss": 12.093183517456055, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.5326488986611366, + "loss_total_v6": 0.7558239698410034, + "step": 182 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.042205405405405405, + "grad_norm": 36.98952865600586, + "learning_rate": 9.15e-06, + "loss": 11.47929573059082, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.49028313532471657, + "loss_total_v6": 0.7174559608101845, + "step": 183 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04243603603603604, + "grad_norm": 37.32700729370117, + "learning_rate": 9.200000000000002e-06, + "loss": 12.461755752563477, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.5516106560826302, + "loss_total_v6": 0.7788597792387009, + "step": 184 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.042666666666666665, + "grad_norm": 30.9171085357666, + "learning_rate": 9.250000000000001e-06, + "loss": 12.054076194763184, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.5265121124684811, + "loss_total_v6": 0.753379762172699, + "step": 185 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0428972972972973, + "grad_norm": 32.08089065551758, + "learning_rate": 9.3e-06, + "loss": 11.57971477508545, + "loss_alignment": 0.44677734375, + "loss_alignment_w": 0.223388671875, + "loss_sft": 0.5003435462713242, + "loss_total_v6": 0.723732203245163, + "step": 186 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.043127927927927925, + "grad_norm": 31.995634078979492, + "learning_rate": 9.350000000000002e-06, + "loss": 11.8929443359375, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.5234909430146217, + "loss_total_v6": 0.7433090507984161, + "step": 187 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04335855855855856, + "grad_norm": 30.82999038696289, + "learning_rate": 9.4e-06, + "loss": 11.747058868408203, + "loss_alignment": 0.449462890625, + "loss_alignment_w": 0.2247314453125, + "loss_sft": 0.5094749927520752, + "loss_total_v6": 0.7341911718249321, + "step": 188 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04358918918918919, + "grad_norm": 36.3910026550293, + "learning_rate": 9.450000000000001e-06, + "loss": 11.697846412658691, + "loss_alignment": 0.474365234375, + "loss_alignment_w": 0.2371826171875, + "loss_sft": 0.4938870146870613, + "loss_total_v6": 0.7311154082417488, + "step": 189 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04381981981981982, + "grad_norm": 34.27309036254883, + "learning_rate": 9.5e-06, + "loss": 11.773165702819824, + "loss_alignment": 0.463134765625, + "loss_alignment_w": 0.2315673828125, + "loss_sft": 0.5033551976084709, + "loss_total_v6": 0.735822856426239, + "step": 190 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04405045045045045, + "grad_norm": 33.006832122802734, + "learning_rate": 9.55e-06, + "loss": 11.412626266479492, + "loss_alignment": 0.455810546875, + "loss_alignment_w": 0.2279052734375, + "loss_sft": 0.48527703434228897, + "loss_total_v6": 0.7132891044020653, + "step": 191 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04428108108108108, + "grad_norm": 29.673646926879883, + "learning_rate": 9.600000000000001e-06, + "loss": 12.092256546020508, + "loss_alignment": 0.46728515625, + "loss_alignment_w": 0.233642578125, + "loss_sft": 0.5221539214253426, + "loss_total_v6": 0.7557660043239594, + "step": 192 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04451171171171171, + "grad_norm": 33.874237060546875, + "learning_rate": 9.65e-06, + "loss": 11.784994125366211, + "loss_alignment": 0.4677734375, + "loss_alignment_w": 0.23388671875, + "loss_sft": 0.5027211643755436, + "loss_total_v6": 0.7365620955824852, + "step": 193 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04474234234234234, + "grad_norm": 33.80225372314453, + "learning_rate": 9.7e-06, + "loss": 12.042694091796875, + "loss_alignment": 0.467529296875, + "loss_alignment_w": 0.2337646484375, + "loss_sft": 0.518964733928442, + "loss_total_v6": 0.7526683509349823, + "step": 194 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04497297297297297, + "grad_norm": 30.392051696777344, + "learning_rate": 9.75e-06, + "loss": 11.828643798828125, + "loss_alignment": 0.450439453125, + "loss_alignment_w": 0.2252197265625, + "loss_sft": 0.5139942392706871, + "loss_total_v6": 0.739290252327919, + "step": 195 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.045203603603603607, + "grad_norm": 35.53730010986328, + "learning_rate": 9.800000000000001e-06, + "loss": 11.98377799987793, + "loss_alignment": 0.4609375, + "loss_alignment_w": 0.23046875, + "loss_sft": 0.5183800347149372, + "loss_total_v6": 0.74898611754179, + "step": 196 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04543423423423423, + "grad_norm": 33.34260177612305, + "learning_rate": 9.85e-06, + "loss": 11.847597122192383, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.5131188258528709, + "loss_total_v6": 0.7404747754335403, + "step": 197 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04566486486486487, + "grad_norm": 31.352413177490234, + "learning_rate": 9.9e-06, + "loss": 11.54684829711914, + "loss_alignment": 0.45751953125, + "loss_alignment_w": 0.228759765625, + "loss_sft": 0.4932081736624241, + "loss_total_v6": 0.7216780111193657, + "step": 198 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04589549549549549, + "grad_norm": 33.06145095825195, + "learning_rate": 9.950000000000001e-06, + "loss": 11.562885284423828, + "loss_alignment": 0.46435546875, + "loss_alignment_w": 0.232177734375, + "loss_sft": 0.49024317413568497, + "loss_total_v6": 0.7226803153753281, + "step": 199 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04612612612612613, + "grad_norm": 36.62339782714844, + "learning_rate": 1e-05, + "loss": 11.918329238891602, + "loss_alignment": 0.46923828125, + "loss_alignment_w": 0.234619140625, + "loss_sft": 0.5106884352862835, + "loss_total_v6": 0.7448955774307251, + "step": 200 + }, + { + "epoch": 0.04612612612612613, + "eval_loss": 0.7289741635322571, + "eval_loss_alignment": 0.4404651609241453, + "eval_loss_alignment_w": 0.22023258046207264, + "eval_loss_sft": 0.5087415825289029, + "eval_loss_total_v6": 0.7289741624815341, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.2546, + "eval_samples_per_second": 22.686, + "eval_steps_per_second": 2.845, + "step": 200 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04635675675675675, + "grad_norm": 34.70073699951172, + "learning_rate": 9.999998557623385e-06, + "loss": 11.406926155090332, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.48589735850691795, + "loss_total_v6": 0.7129328474402428, + "step": 201 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04658738738738739, + "grad_norm": 31.8533992767334, + "learning_rate": 9.999994230494368e-06, + "loss": 11.18117618560791, + "loss_alignment": 0.45263671875, + "loss_alignment_w": 0.226318359375, + "loss_sft": 0.4720626398921013, + "loss_total_v6": 0.6988234966993332, + "step": 202 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04681801801801802, + "grad_norm": 34.72489929199219, + "learning_rate": 9.999987018615447e-06, + "loss": 11.987990379333496, + "loss_alignment": 0.46923828125, + "loss_alignment_w": 0.234619140625, + "loss_sft": 0.5142335332930088, + "loss_total_v6": 0.7492494061589241, + "step": 203 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04704864864864865, + "grad_norm": 30.350446701049805, + "learning_rate": 9.999976921990784e-06, + "loss": 11.771291732788086, + "loss_alignment": 0.4501953125, + "loss_alignment_w": 0.22509765625, + "loss_sft": 0.51091318577528, + "loss_total_v6": 0.7357056960463524, + "step": 204 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04727927927927928, + "grad_norm": 34.23567581176758, + "learning_rate": 9.999963940626203e-06, + "loss": 11.878779411315918, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.5190960951149464, + "loss_total_v6": 0.7424237430095673, + "step": 205 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04750990990990991, + "grad_norm": 29.209897994995117, + "learning_rate": 9.999948074529194e-06, + "loss": 11.768851280212402, + "loss_alignment": 0.44677734375, + "loss_alignment_w": 0.223388671875, + "loss_sft": 0.5125917680561543, + "loss_total_v6": 0.735553190112114, + "step": 206 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04774054054054054, + "grad_norm": 47.937747955322266, + "learning_rate": 9.999929323708913e-06, + "loss": 12.088214874267578, + "loss_alignment": 0.457275390625, + "loss_alignment_w": 0.2286376953125, + "loss_sft": 0.5271961726248264, + "loss_total_v6": 0.7555134147405624, + "step": 207 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.047971171171171174, + "grad_norm": 29.748279571533203, + "learning_rate": 9.999907688176173e-06, + "loss": 11.54991340637207, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.5023108422756195, + "loss_total_v6": 0.721869558095932, + "step": 208 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0482018018018018, + "grad_norm": 29.658599853515625, + "learning_rate": 9.999883167943461e-06, + "loss": 11.5630464553833, + "loss_alignment": 0.45947265625, + "loss_alignment_w": 0.229736328125, + "loss_sft": 0.4928319901227951, + "loss_total_v6": 0.7226904109120369, + "step": 209 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.048432432432432435, + "grad_norm": 34.094261169433594, + "learning_rate": 9.999855763024923e-06, + "loss": 11.491765975952148, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.4937632866203785, + "loss_total_v6": 0.7182353362441063, + "step": 210 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04866306306306306, + "grad_norm": 30.454448699951172, + "learning_rate": 9.99982547343637e-06, + "loss": 11.870224952697754, + "loss_alignment": 0.4560546875, + "loss_alignment_w": 0.22802734375, + "loss_sft": 0.5137243792414665, + "loss_total_v6": 0.7418890371918678, + "step": 211 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.048893693693693695, + "grad_norm": 29.56576156616211, + "learning_rate": 9.999792299195278e-06, + "loss": 11.730812072753906, + "loss_alignment": 0.46337890625, + "loss_alignment_w": 0.231689453125, + "loss_sft": 0.5007233396172523, + "loss_total_v6": 0.733175702393055, + "step": 212 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04912432432432432, + "grad_norm": 33.298946380615234, + "learning_rate": 9.999756240320786e-06, + "loss": 12.208221435546875, + "loss_alignment": 0.466796875, + "loss_alignment_w": 0.2333984375, + "loss_sft": 0.5291119031608105, + "loss_total_v6": 0.7630138397216797, + "step": 213 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.049354954954954955, + "grad_norm": 31.89348602294922, + "learning_rate": 9.9997172968337e-06, + "loss": 11.576068878173828, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.5028164088726044, + "loss_total_v6": 0.7235043048858643, + "step": 214 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.04958558558558559, + "grad_norm": 30.277254104614258, + "learning_rate": 9.999675468756485e-06, + "loss": 12.340288162231445, + "loss_alignment": 0.46875, + "loss_alignment_w": 0.234375, + "loss_sft": 0.5367251262068748, + "loss_total_v6": 0.7712679728865623, + "step": 215 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.049816216216216215, + "grad_norm": 29.645095825195312, + "learning_rate": 9.999630756113278e-06, + "loss": 11.312914848327637, + "loss_alignment": 0.46142578125, + "loss_alignment_w": 0.230712890625, + "loss_sft": 0.4757949709892273, + "loss_total_v6": 0.7070571929216385, + "step": 216 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05004684684684685, + "grad_norm": 29.06071662902832, + "learning_rate": 9.999583158929873e-06, + "loss": 11.46748161315918, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.4935120455920696, + "loss_total_v6": 0.7167176008224487, + "step": 217 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.050277477477477475, + "grad_norm": 29.021289825439453, + "learning_rate": 9.999532677233733e-06, + "loss": 11.416111946105957, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.4890502542257309, + "loss_total_v6": 0.7135069891810417, + "step": 218 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05050810810810811, + "grad_norm": 32.8226203918457, + "learning_rate": 9.999479311053982e-06, + "loss": 11.55020809173584, + "loss_alignment": 0.469482421875, + "loss_alignment_w": 0.2347412109375, + "loss_sft": 0.4874367043375969, + "loss_total_v6": 0.72188800573349, + "step": 219 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.050738738738738735, + "grad_norm": 37.23264694213867, + "learning_rate": 9.999423060421412e-06, + "loss": 12.04061508178711, + "loss_alignment": 0.46240234375, + "loss_alignment_w": 0.231201171875, + "loss_sft": 0.5214898362755775, + "loss_total_v6": 0.7525384351611137, + "step": 220 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05096936936936937, + "grad_norm": 32.45183563232422, + "learning_rate": 9.999363925368472e-06, + "loss": 12.267354011535645, + "loss_alignment": 0.471435546875, + "loss_alignment_w": 0.2357177734375, + "loss_sft": 0.530701931566, + "loss_total_v6": 0.7667096182703972, + "step": 221 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0512, + "grad_norm": 31.216938018798828, + "learning_rate": 9.999301905929286e-06, + "loss": 11.55540657043457, + "loss_alignment": 0.451416015625, + "loss_alignment_w": 0.2257080078125, + "loss_sft": 0.4963828884065151, + "loss_total_v6": 0.7222129479050636, + "step": 222 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05143063063063063, + "grad_norm": 30.690147399902344, + "learning_rate": 9.999237002139633e-06, + "loss": 11.450875282287598, + "loss_alignment": 0.471435546875, + "loss_alignment_w": 0.2357177734375, + "loss_sft": 0.48023658618330956, + "loss_total_v6": 0.7156796827912331, + "step": 223 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05166126126126126, + "grad_norm": 36.514225006103516, + "learning_rate": 9.999169214036958e-06, + "loss": 11.146854400634766, + "loss_alignment": 0.4619140625, + "loss_alignment_w": 0.23095703125, + "loss_sft": 0.46566034853458405, + "loss_total_v6": 0.6966784074902534, + "step": 224 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05189189189189189, + "grad_norm": 29.506427764892578, + "learning_rate": 9.999098541660375e-06, + "loss": 11.965679168701172, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.5243747383356094, + "loss_total_v6": 0.7478549554944038, + "step": 225 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05212252252252252, + "grad_norm": 29.613876342773438, + "learning_rate": 9.999024985050653e-06, + "loss": 11.609557151794434, + "loss_alignment": 0.44873046875, + "loss_alignment_w": 0.224365234375, + "loss_sft": 0.5010947845876217, + "loss_total_v6": 0.7255973592400551, + "step": 226 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.052353153153153156, + "grad_norm": 38.60995101928711, + "learning_rate": 9.998948544250237e-06, + "loss": 11.638998985290527, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.5007528625428677, + "loss_total_v6": 0.7274374142289162, + "step": 227 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05258378378378378, + "grad_norm": 32.84150695800781, + "learning_rate": 9.998869219303227e-06, + "loss": 11.907663345336914, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.5246550180017948, + "loss_total_v6": 0.7442290186882019, + "step": 228 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.052814414414414416, + "grad_norm": 31.941545486450195, + "learning_rate": 9.998787010255388e-06, + "loss": 11.593565940856934, + "loss_alignment": 0.452880859375, + "loss_alignment_w": 0.2264404296875, + "loss_sft": 0.49791333079338074, + "loss_total_v6": 0.7245978862047195, + "step": 229 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05304504504504504, + "grad_norm": 30.050046920776367, + "learning_rate": 9.998701917154152e-06, + "loss": 11.782629013061523, + "loss_alignment": 0.46533203125, + "loss_alignment_w": 0.232666015625, + "loss_sft": 0.5037177726626396, + "loss_total_v6": 0.7364143282175064, + "step": 230 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05327567567567568, + "grad_norm": 31.86957359313965, + "learning_rate": 9.998613940048613e-06, + "loss": 11.657645225524902, + "loss_alignment": 0.45947265625, + "loss_alignment_w": 0.229736328125, + "loss_sft": 0.4990496337413788, + "loss_total_v6": 0.7286028414964676, + "step": 231 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0535063063063063, + "grad_norm": 28.93741798400879, + "learning_rate": 9.99852307898953e-06, + "loss": 11.442533493041992, + "loss_alignment": 0.456298828125, + "loss_alignment_w": 0.2281494140625, + "loss_sft": 0.4875277318060398, + "loss_total_v6": 0.7151583284139633, + "step": 232 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05373693693693694, + "grad_norm": 30.016620635986328, + "learning_rate": 9.998429334029323e-06, + "loss": 11.640024185180664, + "loss_alignment": 0.46337890625, + "loss_alignment_w": 0.231689453125, + "loss_sft": 0.4961477629840374, + "loss_total_v6": 0.7275015413761139, + "step": 233 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05396756756756757, + "grad_norm": 30.419105529785156, + "learning_rate": 9.998332705222083e-06, + "loss": 12.157950401306152, + "loss_alignment": 0.451171875, + "loss_alignment_w": 0.2255859375, + "loss_sft": 0.5343164876103401, + "loss_total_v6": 0.7598719000816345, + "step": 234 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0541981981981982, + "grad_norm": 28.770614624023438, + "learning_rate": 9.998233192623556e-06, + "loss": 11.433634757995605, + "loss_alignment": 0.45556640625, + "loss_alignment_w": 0.227783203125, + "loss_sft": 0.48712415993213654, + "loss_total_v6": 0.714602142572403, + "step": 235 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05442882882882883, + "grad_norm": 34.35268020629883, + "learning_rate": 9.998130796291156e-06, + "loss": 11.744308471679688, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.5160627216100693, + "loss_total_v6": 0.7340192571282387, + "step": 236 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05465945945945946, + "grad_norm": 26.551176071166992, + "learning_rate": 9.998025516283964e-06, + "loss": 11.3525390625, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.4872894734144211, + "loss_total_v6": 0.709533728659153, + "step": 237 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05489009009009009, + "grad_norm": 29.867353439331055, + "learning_rate": 9.997917352662718e-06, + "loss": 11.696304321289062, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.5045480579137802, + "loss_total_v6": 0.7310190424323082, + "step": 238 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05512072072072072, + "grad_norm": 26.87836265563965, + "learning_rate": 9.997806305489826e-06, + "loss": 12.043008804321289, + "loss_alignment": 0.46728515625, + "loss_alignment_w": 0.233642578125, + "loss_sft": 0.5189843773841858, + "loss_total_v6": 0.7526880204677582, + "step": 239 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05535135135135135, + "grad_norm": 29.37978744506836, + "learning_rate": 9.997692374829352e-06, + "loss": 11.805815696716309, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.5143374502658844, + "loss_total_v6": 0.7378634735941887, + "step": 240 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.055581981981981984, + "grad_norm": 30.53933334350586, + "learning_rate": 9.997575560747033e-06, + "loss": 11.408082962036133, + "loss_alignment": 0.47314453125, + "loss_alignment_w": 0.236572265625, + "loss_sft": 0.4760362058877945, + "loss_total_v6": 0.7130052000284195, + "step": 241 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05581261261261261, + "grad_norm": 30.731420516967773, + "learning_rate": 9.997455863310263e-06, + "loss": 11.437803268432617, + "loss_alignment": 0.47265625, + "loss_alignment_w": 0.236328125, + "loss_sft": 0.47812261432409286, + "loss_total_v6": 0.7148627638816833, + "step": 242 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.056043243243243245, + "grad_norm": 33.87355422973633, + "learning_rate": 9.997333282588103e-06, + "loss": 11.470616340637207, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.49375372007489204, + "loss_total_v6": 0.7169135361909866, + "step": 243 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05627387387387387, + "grad_norm": 28.911174774169922, + "learning_rate": 9.997207818651273e-06, + "loss": 11.649812698364258, + "loss_alignment": 0.460693359375, + "loss_alignment_w": 0.2303466796875, + "loss_sft": 0.49752243235707283, + "loss_total_v6": 0.7281132489442825, + "step": 244 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.056504504504504505, + "grad_norm": 27.71599006652832, + "learning_rate": 9.997079471572163e-06, + "loss": 11.22364616394043, + "loss_alignment": 0.457275390625, + "loss_alignment_w": 0.2286376953125, + "loss_sft": 0.47288600727915764, + "loss_total_v6": 0.7014779224991798, + "step": 245 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05673513513513514, + "grad_norm": 28.450580596923828, + "learning_rate": 9.996948241424819e-06, + "loss": 11.916086196899414, + "loss_alignment": 0.45166015625, + "loss_alignment_w": 0.225830078125, + "loss_sft": 0.5190167985856533, + "loss_total_v6": 0.744755357503891, + "step": 246 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.056965765765765765, + "grad_norm": 32.166439056396484, + "learning_rate": 9.99681412828496e-06, + "loss": 11.880169868469238, + "loss_alignment": 0.455078125, + "loss_alignment_w": 0.2275390625, + "loss_sft": 0.515307255089283, + "loss_total_v6": 0.7425106093287468, + "step": 247 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0571963963963964, + "grad_norm": 27.623302459716797, + "learning_rate": 9.996677132229957e-06, + "loss": 11.851598739624023, + "loss_alignment": 0.46435546875, + "loss_alignment_w": 0.232177734375, + "loss_sft": 0.5085776709020138, + "loss_total_v6": 0.7407248988747597, + "step": 248 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.057427027027027025, + "grad_norm": 33.16157531738281, + "learning_rate": 9.996537253338852e-06, + "loss": 11.212041854858398, + "loss_alignment": 0.44677734375, + "loss_alignment_w": 0.223388671875, + "loss_sft": 0.4771198257803917, + "loss_total_v6": 0.7007526233792305, + "step": 249 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05765765765765766, + "grad_norm": 29.621408462524414, + "learning_rate": 9.99639449169235e-06, + "loss": 11.492774963378906, + "loss_alignment": 0.46044921875, + "loss_alignment_w": 0.230224609375, + "loss_sft": 0.48792125657200813, + "loss_total_v6": 0.7182984575629234, + "step": 250 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.057888288288288285, + "grad_norm": 27.76146697998047, + "learning_rate": 9.996248847372813e-06, + "loss": 11.536991119384766, + "loss_alignment": 0.457763671875, + "loss_alignment_w": 0.2288818359375, + "loss_sft": 0.49282095581293106, + "loss_total_v6": 0.7210619151592255, + "step": 251 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05811891891891892, + "grad_norm": 31.08251953125, + "learning_rate": 9.996100320464274e-06, + "loss": 11.891618728637695, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.5161296129226685, + "loss_total_v6": 0.7432261854410172, + "step": 252 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05834954954954955, + "grad_norm": 32.4248161315918, + "learning_rate": 9.995948911052427e-06, + "loss": 11.986920356750488, + "loss_alignment": 0.462890625, + "loss_alignment_w": 0.2314453125, + "loss_sft": 0.5170658454298973, + "loss_total_v6": 0.7491825222969055, + "step": 253 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05858018018018018, + "grad_norm": 33.267478942871094, + "learning_rate": 9.995794619224626e-06, + "loss": 11.851934432983398, + "loss_alignment": 0.45751953125, + "loss_alignment_w": 0.228759765625, + "loss_sft": 0.5112994685769081, + "loss_total_v6": 0.7407458573579788, + "step": 254 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05881081081081081, + "grad_norm": 30.079423904418945, + "learning_rate": 9.995637445069889e-06, + "loss": 12.001468658447266, + "loss_alignment": 0.46435546875, + "loss_alignment_w": 0.232177734375, + "loss_sft": 0.5181276723742485, + "loss_total_v6": 0.7500917986035347, + "step": 255 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05904144144144144, + "grad_norm": 33.44623947143555, + "learning_rate": 9.995477388678898e-06, + "loss": 11.608413696289062, + "loss_alignment": 0.45751953125, + "loss_alignment_w": 0.228759765625, + "loss_sft": 0.49633879214525223, + "loss_total_v6": 0.725525826215744, + "step": 256 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05927207207207207, + "grad_norm": 33.06224060058594, + "learning_rate": 9.995314450143999e-06, + "loss": 11.65203857421875, + "loss_alignment": 0.45166015625, + "loss_alignment_w": 0.225830078125, + "loss_sft": 0.5026359744369984, + "loss_total_v6": 0.7282524406909943, + "step": 257 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.059502702702702706, + "grad_norm": 44.706016540527344, + "learning_rate": 9.995148629559196e-06, + "loss": 11.030210494995117, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.46259675920009613, + "loss_total_v6": 0.6893881261348724, + "step": 258 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.05973333333333333, + "grad_norm": 34.840389251708984, + "learning_rate": 9.994979927020163e-06, + "loss": 11.430460929870605, + "loss_alignment": 0.465087890625, + "loss_alignment_w": 0.2325439453125, + "loss_sft": 0.48169199377298355, + "loss_total_v6": 0.7144038081169128, + "step": 259 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.059963963963963966, + "grad_norm": 32.81713104248047, + "learning_rate": 9.994808342624234e-06, + "loss": 11.736468315124512, + "loss_alignment": 0.462158203125, + "loss_alignment_w": 0.2310791015625, + "loss_sft": 0.5024806782603264, + "loss_total_v6": 0.733529269695282, + "step": 260 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06019459459459459, + "grad_norm": 31.72714614868164, + "learning_rate": 9.9946338764704e-06, + "loss": 11.461612701416016, + "loss_alignment": 0.453369140625, + "loss_alignment_w": 0.2266845703125, + "loss_sft": 0.48939159885048866, + "loss_total_v6": 0.716350831091404, + "step": 261 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.060425225225225226, + "grad_norm": 32.101993560791016, + "learning_rate": 9.994456528659322e-06, + "loss": 11.865608215332031, + "loss_alignment": 0.45654296875, + "loss_alignment_w": 0.228271484375, + "loss_sft": 0.5131611377000809, + "loss_total_v6": 0.7416005209088326, + "step": 262 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06065585585585585, + "grad_norm": 31.736644744873047, + "learning_rate": 9.994276299293321e-06, + "loss": 11.689223289489746, + "loss_alignment": 0.453369140625, + "loss_alignment_w": 0.2266845703125, + "loss_sft": 0.5040139481425285, + "loss_total_v6": 0.7305764555931091, + "step": 263 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06088648648648649, + "grad_norm": 31.48605728149414, + "learning_rate": 9.994093188476383e-06, + "loss": 11.725569725036621, + "loss_alignment": 0.462158203125, + "loss_alignment_w": 0.2310791015625, + "loss_sft": 0.5016468912363052, + "loss_total_v6": 0.7328481003642082, + "step": 264 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06111711711711712, + "grad_norm": 43.17271423339844, + "learning_rate": 9.993907196314148e-06, + "loss": 11.650121688842773, + "loss_alignment": 0.47265625, + "loss_alignment_w": 0.236328125, + "loss_sft": 0.492262277752161, + "loss_total_v6": 0.7281326428055763, + "step": 265 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06134774774774775, + "grad_norm": 32.97384262084961, + "learning_rate": 9.99371832291393e-06, + "loss": 11.300453186035156, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.4790596663951874, + "loss_total_v6": 0.7062783390283585, + "step": 266 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06157837837837838, + "grad_norm": 29.84583854675293, + "learning_rate": 9.993526568384694e-06, + "loss": 11.286233901977539, + "loss_alignment": 0.451171875, + "loss_alignment_w": 0.2255859375, + "loss_sft": 0.4794985093176365, + "loss_total_v6": 0.705389641225338, + "step": 267 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06180900900900901, + "grad_norm": 34.72515869140625, + "learning_rate": 9.99333193283708e-06, + "loss": 11.046730041503906, + "loss_alignment": 0.452392578125, + "loss_alignment_w": 0.2261962890625, + "loss_sft": 0.46381234377622604, + "loss_total_v6": 0.6904206201434135, + "step": 268 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06203963963963964, + "grad_norm": 29.612077713012695, + "learning_rate": 9.993134416383376e-06, + "loss": 11.538249015808105, + "loss_alignment": 0.456787109375, + "loss_alignment_w": 0.2283935546875, + "loss_sft": 0.4929453991353512, + "loss_total_v6": 0.721140556037426, + "step": 269 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06227027027027027, + "grad_norm": 28.409353256225586, + "learning_rate": 9.992934019137544e-06, + "loss": 11.154943466186523, + "loss_alignment": 0.44384765625, + "loss_alignment_w": 0.221923828125, + "loss_sft": 0.47518380358815193, + "loss_total_v6": 0.6971839442849159, + "step": 270 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06250090090090091, + "grad_norm": 34.82999038696289, + "learning_rate": 9.992730741215202e-06, + "loss": 11.46072769165039, + "loss_alignment": 0.45654296875, + "loss_alignment_w": 0.228271484375, + "loss_sft": 0.4877645783126354, + "loss_total_v6": 0.7162954583764076, + "step": 271 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06273153153153153, + "grad_norm": 35.205326080322266, + "learning_rate": 9.99252458273363e-06, + "loss": 11.811737060546875, + "loss_alignment": 0.457763671875, + "loss_alignment_w": 0.2288818359375, + "loss_sft": 0.5099315643310547, + "loss_total_v6": 0.7382335513830185, + "step": 272 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06296216216216216, + "grad_norm": 28.0919246673584, + "learning_rate": 9.992315543811773e-06, + "loss": 11.650668144226074, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.5009176433086395, + "loss_total_v6": 0.7281667664647102, + "step": 273 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06319279279279279, + "grad_norm": 30.649927139282227, + "learning_rate": 9.992103624570234e-06, + "loss": 12.184056282043457, + "loss_alignment": 0.46533203125, + "loss_alignment_w": 0.232666015625, + "loss_sft": 0.5293868482112885, + "loss_total_v6": 0.7615035623311996, + "step": 274 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06342342342342343, + "grad_norm": 29.373218536376953, + "learning_rate": 9.991888825131283e-06, + "loss": 11.508679389953613, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.49611741676926613, + "loss_total_v6": 0.719292476773262, + "step": 275 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06365405405405405, + "grad_norm": 31.78783416748047, + "learning_rate": 9.991671145618847e-06, + "loss": 11.449666976928711, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.4915288835763931, + "loss_total_v6": 0.715604193508625, + "step": 276 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06388468468468468, + "grad_norm": 28.269887924194336, + "learning_rate": 9.991450586158515e-06, + "loss": 11.56843376159668, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.4960984066128731, + "loss_total_v6": 0.7230271026492119, + "step": 277 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06411531531531532, + "grad_norm": 31.216846466064453, + "learning_rate": 9.991227146877542e-06, + "loss": 11.55601692199707, + "loss_alignment": 0.465087890625, + "loss_alignment_w": 0.2325439453125, + "loss_sft": 0.4897528365254402, + "loss_total_v6": 0.7222510501742363, + "step": 278 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06434594594594595, + "grad_norm": 30.95551300048828, + "learning_rate": 9.991000827904839e-06, + "loss": 11.360111236572266, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.4876711182296276, + "loss_total_v6": 0.7100069299340248, + "step": 279 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06457657657657657, + "grad_norm": 29.65535545349121, + "learning_rate": 9.99077162937098e-06, + "loss": 11.275829315185547, + "loss_alignment": 0.45166015625, + "loss_alignment_w": 0.225830078125, + "loss_sft": 0.4784667603671551, + "loss_total_v6": 0.7047393396496773, + "step": 280 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0648072072072072, + "grad_norm": 25.848867416381836, + "learning_rate": 9.990539551408207e-06, + "loss": 11.194833755493164, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.47193970158696175, + "loss_total_v6": 0.699677087366581, + "step": 281 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06503783783783784, + "grad_norm": 29.809242248535156, + "learning_rate": 9.990304594150411e-06, + "loss": 11.343477249145508, + "loss_alignment": 0.4501953125, + "loss_alignment_w": 0.22509765625, + "loss_sft": 0.48388486355543137, + "loss_total_v6": 0.7089672833681107, + "step": 282 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06526846846846847, + "grad_norm": 31.195209503173828, + "learning_rate": 9.990066757733152e-06, + "loss": 11.534358978271484, + "loss_alignment": 0.46142578125, + "loss_alignment_w": 0.230712890625, + "loss_sft": 0.49038296937942505, + "loss_total_v6": 0.7208974659442902, + "step": 283 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0654990990990991, + "grad_norm": 28.25478172302246, + "learning_rate": 9.989826042293653e-06, + "loss": 11.783456802368164, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.5114141628146172, + "loss_total_v6": 0.7364660277962685, + "step": 284 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06572972972972974, + "grad_norm": 32.560115814208984, + "learning_rate": 9.989582447970792e-06, + "loss": 11.985109329223633, + "loss_alignment": 0.46142578125, + "loss_alignment_w": 0.230712890625, + "loss_sft": 0.518661592155695, + "loss_total_v6": 0.7490692883729935, + "step": 285 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06596036036036036, + "grad_norm": 29.149295806884766, + "learning_rate": 9.989335974905112e-06, + "loss": 11.942413330078125, + "loss_alignment": 0.463134765625, + "loss_alignment_w": 0.2315673828125, + "loss_sft": 0.5152302123606205, + "loss_total_v6": 0.7464008256793022, + "step": 286 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06619099099099099, + "grad_norm": 31.819385528564453, + "learning_rate": 9.989086623238815e-06, + "loss": 11.64992618560791, + "loss_alignment": 0.443359375, + "loss_alignment_w": 0.2216796875, + "loss_sft": 0.506669569760561, + "loss_total_v6": 0.7281203716993332, + "step": 287 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06642162162162162, + "grad_norm": 29.710874557495117, + "learning_rate": 9.988834393115768e-06, + "loss": 11.392253875732422, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.48517870903015137, + "loss_total_v6": 0.7120158821344376, + "step": 288 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06665225225225226, + "grad_norm": 31.802400588989258, + "learning_rate": 9.98857928468149e-06, + "loss": 11.766717910766602, + "loss_alignment": 0.457275390625, + "loss_alignment_w": 0.2286376953125, + "loss_sft": 0.5068890266120434, + "loss_total_v6": 0.7354199141263962, + "step": 289 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06688288288288288, + "grad_norm": 27.264209747314453, + "learning_rate": 9.988321298083167e-06, + "loss": 11.57318115234375, + "loss_alignment": 0.44970703125, + "loss_alignment_w": 0.224853515625, + "loss_sft": 0.49868393689393997, + "loss_total_v6": 0.7233238220214844, + "step": 290 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06711351351351351, + "grad_norm": 29.110340118408203, + "learning_rate": 9.98806043346965e-06, + "loss": 11.252796173095703, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.49098901450634, + "loss_total_v6": 0.7032997980713844, + "step": 291 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06734414414414415, + "grad_norm": 31.437707901000977, + "learning_rate": 9.987796690991438e-06, + "loss": 11.445655822753906, + "loss_alignment": 0.46484375, + "loss_alignment_w": 0.232421875, + "loss_sft": 0.48287059366703033, + "loss_total_v6": 0.7153535112738609, + "step": 292 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06757477477477478, + "grad_norm": 30.707019805908203, + "learning_rate": 9.987530070800703e-06, + "loss": 11.808066368103027, + "loss_alignment": 0.45654296875, + "loss_alignment_w": 0.228271484375, + "loss_sft": 0.5102515034377575, + "loss_total_v6": 0.738004170358181, + "step": 293 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0678054054054054, + "grad_norm": 29.90939712524414, + "learning_rate": 9.987260573051268e-06, + "loss": 11.80727767944336, + "loss_alignment": 0.46142578125, + "loss_alignment_w": 0.230712890625, + "loss_sft": 0.5074097625911236, + "loss_total_v6": 0.7379548102617264, + "step": 294 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06803603603603603, + "grad_norm": 30.0562801361084, + "learning_rate": 9.986988197898621e-06, + "loss": 11.322029113769531, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.4841923117637634, + "loss_total_v6": 0.7076267823576927, + "step": 295 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06826666666666667, + "grad_norm": 29.967344284057617, + "learning_rate": 9.98671294549991e-06, + "loss": 11.085673332214355, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.4682146981358528, + "loss_total_v6": 0.6928545981645584, + "step": 296 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0684972972972973, + "grad_norm": 26.281373977661133, + "learning_rate": 9.986434816013941e-06, + "loss": 11.37337875366211, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.49144532158970833, + "loss_total_v6": 0.7108362168073654, + "step": 297 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06872792792792792, + "grad_norm": 35.11698913574219, + "learning_rate": 9.98615380960118e-06, + "loss": 11.165824890136719, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.4772372543811798, + "loss_total_v6": 0.6978640630841255, + "step": 298 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06895855855855856, + "grad_norm": 27.550884246826172, + "learning_rate": 9.985869926423757e-06, + "loss": 11.532064437866211, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.49950163438916206, + "loss_total_v6": 0.72075404971838, + "step": 299 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06918918918918919, + "grad_norm": 27.737232208251953, + "learning_rate": 9.985583166645455e-06, + "loss": 11.087084770202637, + "loss_alignment": 0.446044921875, + "loss_alignment_w": 0.2230224609375, + "loss_sft": 0.4694473147392273, + "loss_total_v6": 0.6929427981376648, + "step": 300 + }, + { + "epoch": 0.06918918918918919, + "eval_loss": 0.7235292196273804, + "eval_loss_alignment": 0.42959777310363245, + "eval_loss_alignment_w": 0.21479888655181623, + "eval_loss_sft": 0.5087303143695124, + "eval_loss_total_v6": 0.7235292009372487, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.4129, + "eval_samples_per_second": 22.642, + "eval_steps_per_second": 2.839, + "step": 300 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06941981981981982, + "grad_norm": 28.57557487487793, + "learning_rate": 9.985293530431722e-06, + "loss": 11.556394577026367, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.501266360282898, + "loss_total_v6": 0.7222746685147285, + "step": 301 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06965045045045046, + "grad_norm": 28.238719940185547, + "learning_rate": 9.985001017949664e-06, + "loss": 11.188663482666016, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4780084975063801, + "loss_total_v6": 0.6992914602160454, + "step": 302 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.06988108108108108, + "grad_norm": 27.428916931152344, + "learning_rate": 9.984705629368046e-06, + "loss": 11.552434921264648, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.5032466724514961, + "loss_total_v6": 0.7220271900296211, + "step": 303 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07011171171171171, + "grad_norm": 25.6920166015625, + "learning_rate": 9.984407364857292e-06, + "loss": 11.78223991394043, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.5089730322360992, + "loss_total_v6": 0.7363900318741798, + "step": 304 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07034234234234234, + "grad_norm": 28.30015754699707, + "learning_rate": 9.984106224589488e-06, + "loss": 11.48825454711914, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.4991896115243435, + "loss_total_v6": 0.7180158942937851, + "step": 305 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07057297297297298, + "grad_norm": 26.668306350708008, + "learning_rate": 9.983802208738373e-06, + "loss": 11.33114242553711, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.48401425406336784, + "loss_total_v6": 0.7081963494420052, + "step": 306 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0708036036036036, + "grad_norm": 29.21381378173828, + "learning_rate": 9.983495317479353e-06, + "loss": 11.265275955200195, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.48070627078413963, + "loss_total_v6": 0.7040797024965286, + "step": 307 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07103423423423423, + "grad_norm": 29.297151565551758, + "learning_rate": 9.983185550989488e-06, + "loss": 11.997926712036133, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.5275956429541111, + "loss_total_v6": 0.7498704195022583, + "step": 308 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07126486486486487, + "grad_norm": 25.41880989074707, + "learning_rate": 9.982872909447496e-06, + "loss": 11.650769233703613, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.5069053769111633, + "loss_total_v6": 0.7281730622053146, + "step": 309 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0714954954954955, + "grad_norm": 27.059541702270508, + "learning_rate": 9.982557393033758e-06, + "loss": 11.88414192199707, + "loss_alignment": 0.448974609375, + "loss_alignment_w": 0.2244873046875, + "loss_sft": 0.5186682529747486, + "loss_total_v6": 0.7427588179707527, + "step": 310 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07172612612612612, + "grad_norm": 30.44021987915039, + "learning_rate": 9.982239001930311e-06, + "loss": 11.503634452819824, + "loss_alignment": 0.459716796875, + "loss_alignment_w": 0.2298583984375, + "loss_sft": 0.48916448652744293, + "loss_total_v6": 0.7189771309494972, + "step": 311 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07195675675675675, + "grad_norm": 29.05365753173828, + "learning_rate": 9.981917736320852e-06, + "loss": 11.68640422821045, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.5108720511198044, + "loss_total_v6": 0.7304002642631531, + "step": 312 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07218738738738739, + "grad_norm": 28.67793846130371, + "learning_rate": 9.981593596390733e-06, + "loss": 11.328319549560547, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.4842345789074898, + "loss_total_v6": 0.7080199792981148, + "step": 313 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07241801801801802, + "grad_norm": 28.412620544433594, + "learning_rate": 9.98126658232697e-06, + "loss": 11.662999153137207, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.5081122852861881, + "loss_total_v6": 0.7289374843239784, + "step": 314 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07264864864864864, + "grad_norm": 29.014169692993164, + "learning_rate": 9.980936694318231e-06, + "loss": 11.571901321411133, + "loss_alignment": 0.46484375, + "loss_alignment_w": 0.232421875, + "loss_sft": 0.49048619344830513, + "loss_total_v6": 0.7232437878847122, + "step": 315 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07287927927927929, + "grad_norm": 29.935726165771484, + "learning_rate": 9.980603932554844e-06, + "loss": 11.38880443572998, + "loss_alignment": 0.44970703125, + "loss_alignment_w": 0.224853515625, + "loss_sft": 0.4869924932718277, + "loss_total_v6": 0.711800254881382, + "step": 316 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07310990990990991, + "grad_norm": 26.376354217529297, + "learning_rate": 9.9802682972288e-06, + "loss": 10.958566665649414, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.46229996532201767, + "loss_total_v6": 0.6849104389548302, + "step": 317 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07334054054054054, + "grad_norm": 29.210689544677734, + "learning_rate": 9.979929788533742e-06, + "loss": 11.354658126831055, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.4889478050172329, + "loss_total_v6": 0.7096661776304245, + "step": 318 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07357117117117117, + "grad_norm": 29.466552734375, + "learning_rate": 9.979588406664972e-06, + "loss": 11.568257331848145, + "loss_alignment": 0.457763671875, + "loss_alignment_w": 0.2288818359375, + "loss_sft": 0.49436314031481743, + "loss_total_v6": 0.7230160906910896, + "step": 319 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0738018018018018, + "grad_norm": 33.14206314086914, + "learning_rate": 9.979244151819454e-06, + "loss": 10.97862720489502, + "loss_alignment": 0.452880859375, + "loss_alignment_w": 0.2264404296875, + "loss_sft": 0.45969322323799133, + "loss_total_v6": 0.6861641630530357, + "step": 320 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07403243243243243, + "grad_norm": 31.375455856323242, + "learning_rate": 9.978897024195801e-06, + "loss": 10.805095672607422, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.46087150275707245, + "loss_total_v6": 0.675318531692028, + "step": 321 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07426306306306306, + "grad_norm": 30.104442596435547, + "learning_rate": 9.978547023994292e-06, + "loss": 11.427815437316895, + "loss_alignment": 0.452880859375, + "loss_alignment_w": 0.2264404296875, + "loss_sft": 0.4875538758933544, + "loss_total_v6": 0.7142384573817253, + "step": 322 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0744936936936937, + "grad_norm": 30.781618118286133, + "learning_rate": 9.978194151416857e-06, + "loss": 10.756420135498047, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.45898886397480965, + "loss_total_v6": 0.6722762286663055, + "step": 323 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07472432432432433, + "grad_norm": 26.66539192199707, + "learning_rate": 9.977838406667091e-06, + "loss": 11.233168601989746, + "loss_alignment": 0.463623046875, + "loss_alignment_w": 0.2318115234375, + "loss_sft": 0.47016996890306473, + "loss_total_v6": 0.7020730227231979, + "step": 324 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07495495495495495, + "grad_norm": 29.40648651123047, + "learning_rate": 9.977479789950238e-06, + "loss": 11.18327522277832, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4858504645526409, + "loss_total_v6": 0.6989547237753868, + "step": 325 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07518558558558558, + "grad_norm": 32.111778259277344, + "learning_rate": 9.977118301473199e-06, + "loss": 10.78118896484375, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.4541740380227566, + "loss_total_v6": 0.6738242954015732, + "step": 326 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07541621621621622, + "grad_norm": 29.77756118774414, + "learning_rate": 9.976753941444541e-06, + "loss": 11.444326400756836, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.5020288154482841, + "loss_total_v6": 0.7152704074978828, + "step": 327 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07564684684684685, + "grad_norm": 32.571510314941406, + "learning_rate": 9.976386710074479e-06, + "loss": 11.021586418151855, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.46980924159288406, + "loss_total_v6": 0.6888491436839104, + "step": 328 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07587747747747747, + "grad_norm": 28.591218948364258, + "learning_rate": 9.976016607574886e-06, + "loss": 11.288956642150879, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.48331553488969803, + "loss_total_v6": 0.7055598050355911, + "step": 329 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07610810810810811, + "grad_norm": 29.59139060974121, + "learning_rate": 9.975643634159294e-06, + "loss": 10.971750259399414, + "loss_alignment": 0.451416015625, + "loss_alignment_w": 0.2257080078125, + "loss_sft": 0.4597059041261673, + "loss_total_v6": 0.6857343763113022, + "step": 330 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07633873873873874, + "grad_norm": 31.33982276916504, + "learning_rate": 9.975267790042892e-06, + "loss": 11.33255386352539, + "loss_alignment": 0.460205078125, + "loss_alignment_w": 0.2301025390625, + "loss_sft": 0.477800615131855, + "loss_total_v6": 0.7082846388220787, + "step": 331 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07656936936936937, + "grad_norm": 30.093433380126953, + "learning_rate": 9.97488907544252e-06, + "loss": 11.721088409423828, + "loss_alignment": 0.446044921875, + "loss_alignment_w": 0.2230224609375, + "loss_sft": 0.509591318666935, + "loss_total_v6": 0.7325680181384087, + "step": 332 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0768, + "grad_norm": 28.462993621826172, + "learning_rate": 9.974507490576682e-06, + "loss": 11.534416198730469, + "loss_alignment": 0.464111328125, + "loss_alignment_w": 0.2320556640625, + "loss_sft": 0.48886067047715187, + "loss_total_v6": 0.7209010645747185, + "step": 333 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07703063063063063, + "grad_norm": 30.571706771850586, + "learning_rate": 9.974123035665531e-06, + "loss": 11.413601875305176, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.48986993357539177, + "loss_total_v6": 0.7133501395583153, + "step": 334 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07726126126126126, + "grad_norm": 28.629528045654297, + "learning_rate": 9.973735710930878e-06, + "loss": 11.086606979370117, + "loss_alignment": 0.450927734375, + "loss_alignment_w": 0.2254638671875, + "loss_sft": 0.46747957170009613, + "loss_total_v6": 0.6929129064083099, + "step": 335 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07749189189189189, + "grad_norm": 29.253631591796875, + "learning_rate": 9.973345516596192e-06, + "loss": 11.029577255249023, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.46600567921996117, + "loss_total_v6": 0.6893485710024834, + "step": 336 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07772252252252253, + "grad_norm": 27.040510177612305, + "learning_rate": 9.972952452886594e-06, + "loss": 11.709606170654297, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.5108573883771896, + "loss_total_v6": 0.7318504303693771, + "step": 337 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07795315315315315, + "grad_norm": 29.741857528686523, + "learning_rate": 9.972556520028863e-06, + "loss": 11.18979263305664, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.48085610568523407, + "loss_total_v6": 0.6993619874119759, + "step": 338 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07818378378378378, + "grad_norm": 28.401206970214844, + "learning_rate": 9.972157718251433e-06, + "loss": 10.63288688659668, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4499710276722908, + "loss_total_v6": 0.6645554080605507, + "step": 339 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07841441441441442, + "grad_norm": 28.848756790161133, + "learning_rate": 9.971756047784393e-06, + "loss": 11.189970016479492, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.47465696558356285, + "loss_total_v6": 0.6993731409311295, + "step": 340 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07864504504504505, + "grad_norm": 34.947994232177734, + "learning_rate": 9.971351508859488e-06, + "loss": 11.364013671875, + "loss_alignment": 0.46240234375, + "loss_alignment_w": 0.231201171875, + "loss_sft": 0.4788817949593067, + "loss_total_v6": 0.7102508097887039, + "step": 341 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07887567567567567, + "grad_norm": 31.91033172607422, + "learning_rate": 9.970944101710115e-06, + "loss": 11.064449310302734, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.4726407080888748, + "loss_total_v6": 0.6915280297398567, + "step": 342 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0791063063063063, + "grad_norm": 33.1578254699707, + "learning_rate": 9.970533826571329e-06, + "loss": 11.119545936584473, + "loss_alignment": 0.45263671875, + "loss_alignment_w": 0.226318359375, + "loss_sft": 0.46871429681777954, + "loss_total_v6": 0.6949716433882713, + "step": 343 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07933693693693694, + "grad_norm": 31.937467575073242, + "learning_rate": 9.970120683679837e-06, + "loss": 11.356756210327148, + "loss_alignment": 0.459228515625, + "loss_alignment_w": 0.2296142578125, + "loss_sft": 0.4806407503783703, + "loss_total_v6": 0.7097972333431244, + "step": 344 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.07956756756756757, + "grad_norm": 28.996990203857422, + "learning_rate": 9.969704673274006e-06, + "loss": 11.6646146774292, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.5096933133900166, + "loss_total_v6": 0.729038417339325, + "step": 345 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0797981981981982, + "grad_norm": 33.95769500732422, + "learning_rate": 9.969285795593851e-06, + "loss": 11.290488243103027, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.4858068786561489, + "loss_total_v6": 0.7056555151939392, + "step": 346 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08002882882882884, + "grad_norm": 28.853342056274414, + "learning_rate": 9.96886405088104e-06, + "loss": 11.028006553649902, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.4730180986225605, + "loss_total_v6": 0.6892503947019577, + "step": 347 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08025945945945946, + "grad_norm": 26.21298599243164, + "learning_rate": 9.968439439378905e-06, + "loss": 11.193748474121094, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4779905825853348, + "loss_total_v6": 0.6996092647314072, + "step": 348 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08049009009009009, + "grad_norm": 27.436994552612305, + "learning_rate": 9.968011961332424e-06, + "loss": 11.229185104370117, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.4810599200427532, + "loss_total_v6": 0.7018240764737129, + "step": 349 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08072072072072072, + "grad_norm": 27.1320743560791, + "learning_rate": 9.967581616988229e-06, + "loss": 11.590816497802734, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.5026395618915558, + "loss_total_v6": 0.7244260385632515, + "step": 350 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08095135135135136, + "grad_norm": 28.161888122558594, + "learning_rate": 9.967148406594608e-06, + "loss": 11.092453002929688, + "loss_alignment": 0.450927734375, + "loss_alignment_w": 0.2254638671875, + "loss_sft": 0.4680585563182831, + "loss_total_v6": 0.6932782903313637, + "step": 351 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08118198198198198, + "grad_norm": 31.009990692138672, + "learning_rate": 9.966712330401503e-06, + "loss": 11.797285079956055, + "loss_alignment": 0.4560546875, + "loss_alignment_w": 0.22802734375, + "loss_sft": 0.5099286139011383, + "loss_total_v6": 0.737330362200737, + "step": 352 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08141261261261261, + "grad_norm": 30.038999557495117, + "learning_rate": 9.966273388660508e-06, + "loss": 11.426652908325195, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.4873896352946758, + "loss_total_v6": 0.7141657993197441, + "step": 353 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08164324324324325, + "grad_norm": 29.770814895629883, + "learning_rate": 9.965831581624872e-06, + "loss": 11.663978576660156, + "loss_alignment": 0.44873046875, + "loss_alignment_w": 0.224365234375, + "loss_sft": 0.5044045262038708, + "loss_total_v6": 0.7289986312389374, + "step": 354 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08187387387387388, + "grad_norm": 27.255666732788086, + "learning_rate": 9.965386909549492e-06, + "loss": 11.005714416503906, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.47046518698334694, + "loss_total_v6": 0.6878571659326553, + "step": 355 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0821045045045045, + "grad_norm": 27.43111801147461, + "learning_rate": 9.964939372690927e-06, + "loss": 11.812007904052734, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.5156094990670681, + "loss_total_v6": 0.7382504716515541, + "step": 356 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08233513513513513, + "grad_norm": 29.026050567626953, + "learning_rate": 9.96448897130738e-06, + "loss": 11.231258392333984, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.4803044758737087, + "loss_total_v6": 0.7019536346197128, + "step": 357 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08256576576576577, + "grad_norm": 25.761032104492188, + "learning_rate": 9.96403570565871e-06, + "loss": 11.240567207336426, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.4755151942372322, + "loss_total_v6": 0.702535443007946, + "step": 358 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0827963963963964, + "grad_norm": 25.1030216217041, + "learning_rate": 9.963579576006433e-06, + "loss": 11.611893653869629, + "loss_alignment": 0.46533203125, + "loss_alignment_w": 0.232666015625, + "loss_sft": 0.4928179457783699, + "loss_total_v6": 0.7257433608174324, + "step": 359 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08302702702702702, + "grad_norm": 26.307510375976562, + "learning_rate": 9.96312058261371e-06, + "loss": 11.344156265258789, + "loss_alignment": 0.4404296875, + "loss_alignment_w": 0.22021484375, + "loss_sft": 0.48794040083885193, + "loss_total_v6": 0.7090097293257713, + "step": 360 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08325765765765766, + "grad_norm": 29.18123435974121, + "learning_rate": 9.962658725745358e-06, + "loss": 11.301422119140625, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.47928810119628906, + "loss_total_v6": 0.7063388898968697, + "step": 361 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08348828828828829, + "grad_norm": 28.149505615234375, + "learning_rate": 9.962194005667844e-06, + "loss": 11.356132507324219, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.4923510476946831, + "loss_total_v6": 0.7097582817077637, + "step": 362 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08371891891891892, + "grad_norm": 29.372400283813477, + "learning_rate": 9.96172642264929e-06, + "loss": 11.166772842407227, + "loss_alignment": 0.443603515625, + "loss_alignment_w": 0.2218017578125, + "loss_sft": 0.47563328966498375, + "loss_total_v6": 0.6979233101010323, + "step": 363 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08394954954954954, + "grad_norm": 29.335710525512695, + "learning_rate": 9.961255976959473e-06, + "loss": 10.784849166870117, + "loss_alignment": 0.4677734375, + "loss_alignment_w": 0.23388671875, + "loss_sft": 0.4403495378792286, + "loss_total_v6": 0.6740531101822853, + "step": 364 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08418018018018018, + "grad_norm": 26.10329246520996, + "learning_rate": 9.960782668869811e-06, + "loss": 11.04189395904541, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.47059018164873123, + "loss_total_v6": 0.6901183873414993, + "step": 365 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08441081081081081, + "grad_norm": 27.30992889404297, + "learning_rate": 9.96030649865338e-06, + "loss": 11.434240341186523, + "loss_alignment": 0.45361328125, + "loss_alignment_w": 0.226806640625, + "loss_sft": 0.4873298555612564, + "loss_total_v6": 0.7146400213241577, + "step": 366 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08464144144144144, + "grad_norm": 26.093050003051758, + "learning_rate": 9.959827466584907e-06, + "loss": 11.583940505981445, + "loss_alignment": 0.453369140625, + "loss_alignment_w": 0.2266845703125, + "loss_sft": 0.4976321756839752, + "loss_total_v6": 0.7239963039755821, + "step": 367 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08487207207207208, + "grad_norm": 27.824661254882812, + "learning_rate": 9.959345572940771e-06, + "loss": 11.559937477111816, + "loss_alignment": 0.456298828125, + "loss_alignment_w": 0.2281494140625, + "loss_sft": 0.4944077357649803, + "loss_total_v6": 0.7224960774183273, + "step": 368 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0851027027027027, + "grad_norm": 26.62984275817871, + "learning_rate": 9.958860817999002e-06, + "loss": 11.04394245147705, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.47203046828508377, + "loss_total_v6": 0.6902463883161545, + "step": 369 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08533333333333333, + "grad_norm": 24.995609283447266, + "learning_rate": 9.958373202039278e-06, + "loss": 11.256861686706543, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4846665598452091, + "loss_total_v6": 0.7035538703203201, + "step": 370 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08556396396396396, + "grad_norm": 27.894420623779297, + "learning_rate": 9.957882725342926e-06, + "loss": 11.4934663772583, + "loss_alignment": 0.457763671875, + "loss_alignment_w": 0.2288818359375, + "loss_sft": 0.4898108132183552, + "loss_total_v6": 0.7183416485786438, + "step": 371 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0857945945945946, + "grad_norm": 29.48624038696289, + "learning_rate": 9.957389388192935e-06, + "loss": 10.87598991394043, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.46077051013708115, + "loss_total_v6": 0.6797493994235992, + "step": 372 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08602522522522522, + "grad_norm": 26.481958389282227, + "learning_rate": 9.956893190873928e-06, + "loss": 11.713956832885742, + "loss_alignment": 0.447998046875, + "loss_alignment_w": 0.2239990234375, + "loss_sft": 0.5082148462533951, + "loss_total_v6": 0.7321223393082619, + "step": 373 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08625585585585585, + "grad_norm": 29.41984748840332, + "learning_rate": 9.956394133672193e-06, + "loss": 11.252193450927734, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4861753173172474, + "loss_total_v6": 0.7032621279358864, + "step": 374 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08648648648648649, + "grad_norm": 27.534440994262695, + "learning_rate": 9.955892216875656e-06, + "loss": 11.3914794921875, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.49160006269812584, + "loss_total_v6": 0.7119674757122993, + "step": 375 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08671711711711712, + "grad_norm": 25.709659576416016, + "learning_rate": 9.955387440773902e-06, + "loss": 10.487455368041992, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4374788776040077, + "loss_total_v6": 0.6554659456014633, + "step": 376 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08694774774774774, + "grad_norm": 27.509511947631836, + "learning_rate": 9.95487980565816e-06, + "loss": 11.801471710205078, + "loss_alignment": 0.452880859375, + "loss_alignment_w": 0.2264404296875, + "loss_sft": 0.5112125612795353, + "loss_total_v6": 0.737591952085495, + "step": 377 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08717837837837838, + "grad_norm": 28.727340698242188, + "learning_rate": 9.95436931182131e-06, + "loss": 11.563343048095703, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.5024482719600201, + "loss_total_v6": 0.7227088883519173, + "step": 378 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08740900900900901, + "grad_norm": 30.104290008544922, + "learning_rate": 9.953855959557883e-06, + "loss": 11.066493034362793, + "loss_alignment": 0.455322265625, + "loss_alignment_w": 0.2276611328125, + "loss_sft": 0.46425406262278557, + "loss_total_v6": 0.6916558220982552, + "step": 379 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08763963963963964, + "grad_norm": 25.38064956665039, + "learning_rate": 9.953339749164057e-06, + "loss": 11.216751098632812, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4826325960457325, + "loss_total_v6": 0.7010469064116478, + "step": 380 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08787027027027026, + "grad_norm": 28.550804138183594, + "learning_rate": 9.95282068093766e-06, + "loss": 11.56042766571045, + "loss_alignment": 0.45849609375, + "loss_alignment_w": 0.229248046875, + "loss_sft": 0.4930345229804516, + "loss_total_v6": 0.7225267142057419, + "step": 381 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0881009009009009, + "grad_norm": 26.116981506347656, + "learning_rate": 9.952298755178171e-06, + "loss": 11.746980667114258, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.5095769539475441, + "loss_total_v6": 0.7341862842440605, + "step": 382 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08833153153153153, + "grad_norm": 27.73813819885254, + "learning_rate": 9.951773972186712e-06, + "loss": 10.903629302978516, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.46982215344905853, + "loss_total_v6": 0.681476816534996, + "step": 383 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08856216216216216, + "grad_norm": 27.25075912475586, + "learning_rate": 9.951246332266058e-06, + "loss": 11.342487335205078, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.48455553501844406, + "loss_total_v6": 0.7089054957032204, + "step": 384 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0887927927927928, + "grad_norm": 27.051881790161133, + "learning_rate": 9.950715835720633e-06, + "loss": 11.764689445495605, + "loss_alignment": 0.458984375, + "loss_alignment_w": 0.2294921875, + "loss_sft": 0.5056177563965321, + "loss_total_v6": 0.7352930977940559, + "step": 385 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08902342342342343, + "grad_norm": 29.8158016204834, + "learning_rate": 9.950182482856506e-06, + "loss": 11.048182487487793, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.47121210023760796, + "loss_total_v6": 0.6905113905668259, + "step": 386 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08925405405405405, + "grad_norm": 25.72268295288086, + "learning_rate": 9.949646273981394e-06, + "loss": 11.475069046020508, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.49964726343750954, + "loss_total_v6": 0.7171918004751205, + "step": 387 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08948468468468468, + "grad_norm": 26.996070861816406, + "learning_rate": 9.949107209404664e-06, + "loss": 11.089040756225586, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.4819597043097019, + "loss_total_v6": 0.6930650323629379, + "step": 388 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08971531531531532, + "grad_norm": 31.46682357788086, + "learning_rate": 9.94856528943733e-06, + "loss": 11.013996124267578, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.46564218774437904, + "loss_total_v6": 0.688374750316143, + "step": 389 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.08994594594594595, + "grad_norm": 27.844886779785156, + "learning_rate": 9.948020514392055e-06, + "loss": 11.128021240234375, + "loss_alignment": 0.45654296875, + "loss_alignment_w": 0.228271484375, + "loss_sft": 0.46733666956424713, + "loss_total_v6": 0.6955013573169708, + "step": 390 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09017657657657657, + "grad_norm": 28.694536209106445, + "learning_rate": 9.947472884583145e-06, + "loss": 11.400623321533203, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.49305659905076027, + "loss_total_v6": 0.7125390097498894, + "step": 391 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09040720720720721, + "grad_norm": 30.593429565429688, + "learning_rate": 9.946922400326554e-06, + "loss": 11.091445922851562, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.46737006679177284, + "loss_total_v6": 0.6932153925299644, + "step": 392 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09063783783783784, + "grad_norm": 25.742643356323242, + "learning_rate": 9.946369061939886e-06, + "loss": 11.281946182250977, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.48502882942557335, + "loss_total_v6": 0.7051216438412666, + "step": 393 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09086846846846847, + "grad_norm": 26.9558048248291, + "learning_rate": 9.94581286974239e-06, + "loss": 11.104230880737305, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.473311360925436, + "loss_total_v6": 0.6940144449472427, + "step": 394 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09109909909909909, + "grad_norm": 27.883655548095703, + "learning_rate": 9.945253824054964e-06, + "loss": 11.244206428527832, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.4754832461476326, + "loss_total_v6": 0.7027629017829895, + "step": 395 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09132972972972973, + "grad_norm": 26.981019973754883, + "learning_rate": 9.944691925200145e-06, + "loss": 11.364527702331543, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.48323216661810875, + "loss_total_v6": 0.710282988846302, + "step": 396 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09156036036036036, + "grad_norm": 30.640958786010742, + "learning_rate": 9.944127173502123e-06, + "loss": 11.30667781829834, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.48544542863965034, + "loss_total_v6": 0.706667385995388, + "step": 397 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09179099099099099, + "grad_norm": 27.942893981933594, + "learning_rate": 9.943559569286731e-06, + "loss": 11.357218742370605, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.49173231795430183, + "loss_total_v6": 0.7098261639475822, + "step": 398 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09202162162162163, + "grad_norm": 26.533599853515625, + "learning_rate": 9.942989112881451e-06, + "loss": 11.313238143920898, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.48460428416728973, + "loss_total_v6": 0.7070774212479591, + "step": 399 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09225225225225225, + "grad_norm": 27.1458683013916, + "learning_rate": 9.942415804615407e-06, + "loss": 11.33221435546875, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.48698044568300247, + "loss_total_v6": 0.7082634344696999, + "step": 400 + }, + { + "epoch": 0.09225225225225225, + "eval_loss": 0.7166559100151062, + "eval_loss_alignment": 0.4236904046474359, + "eval_loss_alignment_w": 0.21184520232371795, + "eval_loss_sft": 0.5048107340390611, + "eval_loss_total_v6": 0.7166559365378995, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.9479, + "eval_samples_per_second": 22.771, + "eval_steps_per_second": 2.855, + "step": 400 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09248288288288288, + "grad_norm": 24.91004753112793, + "learning_rate": 9.941839644819366e-06, + "loss": 11.58151626586914, + "loss_alignment": 0.460205078125, + "loss_alignment_w": 0.2301025390625, + "loss_sft": 0.49365072697401047, + "loss_total_v6": 0.7238448038697243, + "step": 401 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0927135135135135, + "grad_norm": 26.88804054260254, + "learning_rate": 9.941260633825752e-06, + "loss": 11.682621002197266, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.5180819444358349, + "loss_total_v6": 0.7301638275384903, + "step": 402 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09294414414414415, + "grad_norm": 24.303524017333984, + "learning_rate": 9.940678771968617e-06, + "loss": 10.639737129211426, + "loss_alignment": 0.451171875, + "loss_alignment_w": 0.2255859375, + "loss_sft": 0.4397180341184139, + "loss_total_v6": 0.6649835705757141, + "step": 403 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09317477477477477, + "grad_norm": 25.812564849853516, + "learning_rate": 9.940094059583672e-06, + "loss": 11.407072067260742, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4983423687517643, + "loss_total_v6": 0.7129419893026352, + "step": 404 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0934054054054054, + "grad_norm": 28.003536224365234, + "learning_rate": 9.939506497008264e-06, + "loss": 11.542268753051758, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.49960531294345856, + "loss_total_v6": 0.7213918268680573, + "step": 405 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09363603603603604, + "grad_norm": 26.44243621826172, + "learning_rate": 9.938916084581391e-06, + "loss": 11.472530364990234, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.49909187853336334, + "loss_total_v6": 0.7170331329107285, + "step": 406 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09386666666666667, + "grad_norm": 26.810640335083008, + "learning_rate": 9.938322822643688e-06, + "loss": 11.677901268005371, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.5086621753871441, + "loss_total_v6": 0.7298688367009163, + "step": 407 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0940972972972973, + "grad_norm": 24.54094123840332, + "learning_rate": 9.937726711537441e-06, + "loss": 11.399065017700195, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4970332607626915, + "loss_total_v6": 0.712441585958004, + "step": 408 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09432792792792793, + "grad_norm": 26.584129333496094, + "learning_rate": 9.937127751606577e-06, + "loss": 11.172761917114258, + "loss_alignment": 0.45556640625, + "loss_alignment_w": 0.227783203125, + "loss_sft": 0.4700109250843525, + "loss_total_v6": 0.6982976421713829, + "step": 409 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09455855855855856, + "grad_norm": 27.463342666625977, + "learning_rate": 9.936525943196663e-06, + "loss": 11.324691772460938, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.4855947047472, + "loss_total_v6": 0.7077931761741638, + "step": 410 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09478918918918919, + "grad_norm": 25.687305450439453, + "learning_rate": 9.935921286654914e-06, + "loss": 11.433106422424316, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.491714496165514, + "loss_total_v6": 0.7145691141486168, + "step": 411 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09501981981981981, + "grad_norm": 31.573997497558594, + "learning_rate": 9.93531378233019e-06, + "loss": 11.280133247375488, + "loss_alignment": 0.462158203125, + "loss_alignment_w": 0.2310791015625, + "loss_sft": 0.4739292338490486, + "loss_total_v6": 0.705008327960968, + "step": 412 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09525045045045046, + "grad_norm": 28.66705894470215, + "learning_rate": 9.934703430572988e-06, + "loss": 11.034075736999512, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.4668361768126488, + "loss_total_v6": 0.6896297261118889, + "step": 413 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09548108108108108, + "grad_norm": 27.40375518798828, + "learning_rate": 9.93409023173545e-06, + "loss": 11.237479209899902, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4825243204832077, + "loss_total_v6": 0.7023424357175827, + "step": 414 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09571171171171171, + "grad_norm": 25.924837112426758, + "learning_rate": 9.933474186171363e-06, + "loss": 11.39125919342041, + "loss_alignment": 0.450439453125, + "loss_alignment_w": 0.2252197265625, + "loss_sft": 0.48635251820087433, + "loss_total_v6": 0.711953729391098, + "step": 415 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09594234234234235, + "grad_norm": 25.303159713745117, + "learning_rate": 9.932855294236154e-06, + "loss": 11.5601224899292, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.500583790242672, + "loss_total_v6": 0.7225076332688332, + "step": 416 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09617297297297298, + "grad_norm": 27.318199157714844, + "learning_rate": 9.932233556286897e-06, + "loss": 10.748483657836914, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.4538999944925308, + "loss_total_v6": 0.6717802360653877, + "step": 417 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0964036036036036, + "grad_norm": 25.923189163208008, + "learning_rate": 9.931608972682298e-06, + "loss": 11.667896270751953, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.5089828185737133, + "loss_total_v6": 0.7292434573173523, + "step": 418 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09663423423423423, + "grad_norm": 25.941078186035156, + "learning_rate": 9.930981543782716e-06, + "loss": 11.606263160705566, + "loss_alignment": 0.4580078125, + "loss_alignment_w": 0.22900390625, + "loss_sft": 0.4964943863451481, + "loss_total_v6": 0.7253914773464203, + "step": 419 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09686486486486487, + "grad_norm": 25.522388458251953, + "learning_rate": 9.930351269950144e-06, + "loss": 11.268019676208496, + "loss_alignment": 0.453369140625, + "loss_alignment_w": 0.2266845703125, + "loss_sft": 0.4773072637617588, + "loss_total_v6": 0.7042512446641922, + "step": 420 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0970954954954955, + "grad_norm": 28.475189208984375, + "learning_rate": 9.929718151548218e-06, + "loss": 11.023386001586914, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.46859414875507355, + "loss_total_v6": 0.6889615878462791, + "step": 421 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09732612612612612, + "grad_norm": 28.821537017822266, + "learning_rate": 9.929082188942219e-06, + "loss": 11.376764297485352, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.48875774070620537, + "loss_total_v6": 0.7110477685928345, + "step": 422 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09755675675675676, + "grad_norm": 26.405364990234375, + "learning_rate": 9.928443382499063e-06, + "loss": 10.978561401367188, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4672727435827255, + "loss_total_v6": 0.6861600652337074, + "step": 423 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09778738738738739, + "grad_norm": 28.668960571289062, + "learning_rate": 9.927801732587311e-06, + "loss": 11.464613914489746, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.49836816638708115, + "loss_total_v6": 0.7165383547544479, + "step": 424 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09801801801801802, + "grad_norm": 28.264392852783203, + "learning_rate": 9.927157239577165e-06, + "loss": 11.356938362121582, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.4906314127147198, + "loss_total_v6": 0.7098086699843407, + "step": 425 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09824864864864864, + "grad_norm": 25.217702865600586, + "learning_rate": 9.926509903840464e-06, + "loss": 11.38100814819336, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.494455061852932, + "loss_total_v6": 0.7113129794597626, + "step": 426 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09847927927927928, + "grad_norm": 25.722023010253906, + "learning_rate": 9.92585972575069e-06, + "loss": 11.55448055267334, + "loss_alignment": 0.45703125, + "loss_alignment_w": 0.228515625, + "loss_sft": 0.4941429980099201, + "loss_total_v6": 0.7221550494432449, + "step": 427 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09870990990990991, + "grad_norm": 25.776371002197266, + "learning_rate": 9.925206705682961e-06, + "loss": 11.430065155029297, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.5047538056969643, + "loss_total_v6": 0.7143790870904922, + "step": 428 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09894054054054054, + "grad_norm": 24.329133987426758, + "learning_rate": 9.924550844014039e-06, + "loss": 11.007933616638184, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.467109601944685, + "loss_total_v6": 0.6879958435893059, + "step": 429 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09917117117117118, + "grad_norm": 26.33008575439453, + "learning_rate": 9.923892141122324e-06, + "loss": 11.937064170837402, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.524371549487114, + "loss_total_v6": 0.7460664957761765, + "step": 430 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.0994018018018018, + "grad_norm": 28.04679298400879, + "learning_rate": 9.923230597387856e-06, + "loss": 11.232162475585938, + "loss_alignment": 0.45849609375, + "loss_alignment_w": 0.229248046875, + "loss_sft": 0.4724569581449032, + "loss_total_v6": 0.7020101547241211, + "step": 431 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09963243243243243, + "grad_norm": 27.77845001220703, + "learning_rate": 9.92256621319231e-06, + "loss": 11.070272445678711, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.47071588039398193, + "loss_total_v6": 0.6918920204043388, + "step": 432 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.09986306306306306, + "grad_norm": 34.13277816772461, + "learning_rate": 9.921898988919006e-06, + "loss": 11.36225414276123, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.4874235913157463, + "loss_total_v6": 0.7101408690214157, + "step": 433 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1000936936936937, + "grad_norm": 27.15593910217285, + "learning_rate": 9.9212289249529e-06, + "loss": 11.31576156616211, + "loss_alignment": 0.452880859375, + "loss_alignment_w": 0.2264404296875, + "loss_sft": 0.480901513248682, + "loss_total_v6": 0.7072350904345512, + "step": 434 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10032432432432432, + "grad_norm": 27.45899200439453, + "learning_rate": 9.92055602168058e-06, + "loss": 10.720725059509277, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.45340102165937424, + "loss_total_v6": 0.6700453162193298, + "step": 435 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10055495495495495, + "grad_norm": 28.589221954345703, + "learning_rate": 9.919880279490286e-06, + "loss": 11.118029594421387, + "loss_alignment": 0.45263671875, + "loss_alignment_w": 0.226318359375, + "loss_sft": 0.46916883811354637, + "loss_total_v6": 0.6948768347501755, + "step": 436 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10078558558558559, + "grad_norm": 29.076961517333984, + "learning_rate": 9.919201698771882e-06, + "loss": 11.073562622070312, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.46870893985033035, + "loss_total_v6": 0.6920976266264915, + "step": 437 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10101621621621622, + "grad_norm": 28.21442985534668, + "learning_rate": 9.918520279916879e-06, + "loss": 11.384393692016602, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4947276748716831, + "loss_total_v6": 0.7115245833992958, + "step": 438 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10124684684684684, + "grad_norm": 29.37422752380371, + "learning_rate": 9.91783602331842e-06, + "loss": 11.465917587280273, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.491781622171402, + "loss_total_v6": 0.7166198641061783, + "step": 439 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10147747747747747, + "grad_norm": 28.4581241607666, + "learning_rate": 9.917148929371287e-06, + "loss": 11.046344757080078, + "loss_alignment": 0.45361328125, + "loss_alignment_w": 0.226806640625, + "loss_sft": 0.46363565325737, + "loss_total_v6": 0.6903965100646019, + "step": 440 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10170810810810811, + "grad_norm": 29.321556091308594, + "learning_rate": 9.916458998471902e-06, + "loss": 11.15117359161377, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.480365090072155, + "loss_total_v6": 0.6969483494758606, + "step": 441 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10193873873873874, + "grad_norm": 28.69033432006836, + "learning_rate": 9.915766231018317e-06, + "loss": 10.682655334472656, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4484581798315048, + "loss_total_v6": 0.667665958404541, + "step": 442 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10216936936936936, + "grad_norm": 28.06540870666504, + "learning_rate": 9.915070627410229e-06, + "loss": 11.410057067871094, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.49041130766272545, + "loss_total_v6": 0.713128574192524, + "step": 443 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1024, + "grad_norm": 36.065757751464844, + "learning_rate": 9.914372188048964e-06, + "loss": 10.880657196044922, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.45501967146992683, + "loss_total_v6": 0.6800410374999046, + "step": 444 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10263063063063063, + "grad_norm": 26.746559143066406, + "learning_rate": 9.91367091333749e-06, + "loss": 10.5686674118042, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.44647616147994995, + "loss_total_v6": 0.6605417281389236, + "step": 445 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10286126126126126, + "grad_norm": 28.744384765625, + "learning_rate": 9.912966803680404e-06, + "loss": 11.242752075195312, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.48167894780635834, + "loss_total_v6": 0.7026720270514488, + "step": 446 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1030918918918919, + "grad_norm": 28.30727767944336, + "learning_rate": 9.912259859483943e-06, + "loss": 11.205696105957031, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.48279616236686707, + "loss_total_v6": 0.7003559917211533, + "step": 447 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10332252252252253, + "grad_norm": 26.289276123046875, + "learning_rate": 9.911550081155983e-06, + "loss": 11.263870239257812, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.4800996817648411, + "loss_total_v6": 0.7039918750524521, + "step": 448 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10355315315315315, + "grad_norm": 33.08310317993164, + "learning_rate": 9.910837469106028e-06, + "loss": 11.278775215148926, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.48077191039919853, + "loss_total_v6": 0.7049234881997108, + "step": 449 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10378378378378378, + "grad_norm": 25.804088592529297, + "learning_rate": 9.910122023745219e-06, + "loss": 11.032815933227539, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.4692598357796669, + "loss_total_v6": 0.6895510032773018, + "step": 450 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10401441441441442, + "grad_norm": 24.38065528869629, + "learning_rate": 9.909403745486335e-06, + "loss": 10.607091903686523, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4463905394077301, + "loss_total_v6": 0.6629432737827301, + "step": 451 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10424504504504505, + "grad_norm": 26.947189331054688, + "learning_rate": 9.908682634743785e-06, + "loss": 11.374394416809082, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.48990660160779953, + "loss_total_v6": 0.7108996734023094, + "step": 452 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10447567567567567, + "grad_norm": 28.793928146362305, + "learning_rate": 9.907958691933616e-06, + "loss": 10.903107643127441, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.46222124621272087, + "loss_total_v6": 0.6814442351460457, + "step": 453 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10470630630630631, + "grad_norm": 27.71890640258789, + "learning_rate": 9.907231917473507e-06, + "loss": 11.270236015319824, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.4846784509718418, + "loss_total_v6": 0.704389750957489, + "step": 454 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10493693693693694, + "grad_norm": 24.613983154296875, + "learning_rate": 9.90650231178277e-06, + "loss": 11.064895629882812, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.4677705653011799, + "loss_total_v6": 0.691555991768837, + "step": 455 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10516756756756757, + "grad_norm": 24.54472541809082, + "learning_rate": 9.905769875282351e-06, + "loss": 11.121420860290527, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.47192898392677307, + "loss_total_v6": 0.6950887888669968, + "step": 456 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10539819819819819, + "grad_norm": 27.47614860534668, + "learning_rate": 9.905034608394835e-06, + "loss": 10.857702255249023, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.46147380769252777, + "loss_total_v6": 0.6786063686013222, + "step": 457 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10562882882882883, + "grad_norm": 26.878366470336914, + "learning_rate": 9.904296511544427e-06, + "loss": 11.313644409179688, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.48586558550596237, + "loss_total_v6": 0.7071027308702469, + "step": 458 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10585945945945946, + "grad_norm": 27.665996551513672, + "learning_rate": 9.903555585156977e-06, + "loss": 10.89101791381836, + "loss_alignment": 0.451904296875, + "loss_alignment_w": 0.2259521484375, + "loss_sft": 0.4549347944557667, + "loss_total_v6": 0.6806885972619057, + "step": 459 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10609009009009009, + "grad_norm": 28.856172561645508, + "learning_rate": 9.902811829659962e-06, + "loss": 10.606754302978516, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.44459937885403633, + "loss_total_v6": 0.662922129034996, + "step": 460 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10632072072072073, + "grad_norm": 28.820030212402344, + "learning_rate": 9.902065245482493e-06, + "loss": 11.4293794631958, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.49769191443920135, + "loss_total_v6": 0.7143361940979958, + "step": 461 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10655135135135135, + "grad_norm": 26.87721061706543, + "learning_rate": 9.90131583305531e-06, + "loss": 10.655073165893555, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4499691091477871, + "loss_total_v6": 0.6659420356154442, + "step": 462 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10678198198198198, + "grad_norm": 27.253591537475586, + "learning_rate": 9.900563592810789e-06, + "loss": 11.23504638671875, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4826621860265732, + "loss_total_v6": 0.7021903768181801, + "step": 463 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1070126126126126, + "grad_norm": 26.786314010620117, + "learning_rate": 9.899808525182935e-06, + "loss": 10.825377464294434, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4596366509795189, + "loss_total_v6": 0.6765860915184021, + "step": 464 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10724324324324325, + "grad_norm": 24.017526626586914, + "learning_rate": 9.899050630607386e-06, + "loss": 10.8063325881958, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.45801907405257225, + "loss_total_v6": 0.6753957942128181, + "step": 465 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10747387387387387, + "grad_norm": 30.280921936035156, + "learning_rate": 9.898289909521405e-06, + "loss": 10.540740966796875, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.4405955597758293, + "loss_total_v6": 0.6587962731719017, + "step": 466 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1077045045045045, + "grad_norm": 27.084056854248047, + "learning_rate": 9.897526362363896e-06, + "loss": 11.536672592163086, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.5009187646210194, + "loss_total_v6": 0.7210420519113541, + "step": 467 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10793513513513514, + "grad_norm": 28.2257137298584, + "learning_rate": 9.896759989575386e-06, + "loss": 10.904271125793457, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4666884206235409, + "loss_total_v6": 0.6815169379115105, + "step": 468 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10816576576576577, + "grad_norm": 25.916242599487305, + "learning_rate": 9.895990791598037e-06, + "loss": 10.970686912536621, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4726552404463291, + "loss_total_v6": 0.6856679469347, + "step": 469 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1083963963963964, + "grad_norm": 24.78522300720215, + "learning_rate": 9.89521876887563e-06, + "loss": 10.97653579711914, + "loss_alignment": 0.453125, + "loss_alignment_w": 0.2265625, + "loss_sft": 0.4596235863864422, + "loss_total_v6": 0.6860335320234299, + "step": 470 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10862702702702702, + "grad_norm": 29.446426391601562, + "learning_rate": 9.894443921853594e-06, + "loss": 11.043212890625, + "loss_alignment": 0.449462890625, + "loss_alignment_w": 0.2247314453125, + "loss_sft": 0.4658202789723873, + "loss_total_v6": 0.6902007758617401, + "step": 471 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10885765765765766, + "grad_norm": 30.592803955078125, + "learning_rate": 9.893666250978971e-06, + "loss": 10.889028549194336, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4629587233066559, + "loss_total_v6": 0.680564321577549, + "step": 472 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10908828828828829, + "grad_norm": 26.426082611083984, + "learning_rate": 9.892885756700441e-06, + "loss": 11.290277481079102, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.48838772252202034, + "loss_total_v6": 0.7056423723697662, + "step": 473 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10931891891891891, + "grad_norm": 25.81511688232422, + "learning_rate": 9.89210243946831e-06, + "loss": 11.020698547363281, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.46806007251143456, + "loss_total_v6": 0.6887937039136887, + "step": 474 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10954954954954955, + "grad_norm": 41.192054748535156, + "learning_rate": 9.891316299734514e-06, + "loss": 11.172481536865234, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.48993662372231483, + "loss_total_v6": 0.6982801035046577, + "step": 475 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.10978018018018018, + "grad_norm": 27.79509162902832, + "learning_rate": 9.890527337952617e-06, + "loss": 11.415933609008789, + "loss_alignment": 0.447998046875, + "loss_alignment_w": 0.2239990234375, + "loss_sft": 0.4895120598375797, + "loss_total_v6": 0.7134958058595657, + "step": 476 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11001081081081081, + "grad_norm": 28.088666915893555, + "learning_rate": 9.88973555457781e-06, + "loss": 11.778175354003906, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.5152649879455566, + "loss_total_v6": 0.7361359670758247, + "step": 477 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11024144144144143, + "grad_norm": 28.79584312438965, + "learning_rate": 9.888940950066915e-06, + "loss": 11.096693992614746, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4767007492482662, + "loss_total_v6": 0.6935433819890022, + "step": 478 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11047207207207208, + "grad_norm": 26.64463996887207, + "learning_rate": 9.888143524878378e-06, + "loss": 10.807662010192871, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.45480625703930855, + "loss_total_v6": 0.6754788607358932, + "step": 479 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1107027027027027, + "grad_norm": 24.392343521118164, + "learning_rate": 9.887343279472272e-06, + "loss": 11.304396629333496, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.4862183481454849, + "loss_total_v6": 0.7065247520804405, + "step": 480 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11093333333333333, + "grad_norm": 37.36918640136719, + "learning_rate": 9.886540214310303e-06, + "loss": 11.134832382202148, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.4725230745971203, + "loss_total_v6": 0.6959270015358925, + "step": 481 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11116396396396397, + "grad_norm": 27.247228622436523, + "learning_rate": 9.885734329855798e-06, + "loss": 11.5765380859375, + "loss_alignment": 0.446044921875, + "loss_alignment_w": 0.2230224609375, + "loss_sft": 0.5005875043570995, + "loss_total_v6": 0.7235336825251579, + "step": 482 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1113945945945946, + "grad_norm": 26.233985900878906, + "learning_rate": 9.884925626573714e-06, + "loss": 10.899261474609375, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.46610064804553986, + "loss_total_v6": 0.6812037825584412, + "step": 483 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11162522522522522, + "grad_norm": 27.670480728149414, + "learning_rate": 9.884114104930629e-06, + "loss": 11.125368118286133, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4784776270389557, + "loss_total_v6": 0.6953355073928833, + "step": 484 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11185585585585586, + "grad_norm": 26.617414474487305, + "learning_rate": 9.883299765394756e-06, + "loss": 11.378989219665527, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.49452726542949677, + "loss_total_v6": 0.7111868336796761, + "step": 485 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11208648648648649, + "grad_norm": 32.58820343017578, + "learning_rate": 9.882482608435924e-06, + "loss": 11.419721603393555, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.49505892023444176, + "loss_total_v6": 0.7137326151132584, + "step": 486 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11231711711711712, + "grad_norm": 28.05935287475586, + "learning_rate": 9.881662634525596e-06, + "loss": 10.900257110595703, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.46398095414042473, + "loss_total_v6": 0.681266114115715, + "step": 487 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11254774774774774, + "grad_norm": 25.242948532104492, + "learning_rate": 9.880839844136854e-06, + "loss": 10.997100830078125, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.46704286709427834, + "loss_total_v6": 0.687318779528141, + "step": 488 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11277837837837838, + "grad_norm": 27.706588745117188, + "learning_rate": 9.88001423774441e-06, + "loss": 11.068282127380371, + "loss_alignment": 0.44677734375, + "loss_alignment_w": 0.223388671875, + "loss_sft": 0.4685010276734829, + "loss_total_v6": 0.6917676404118538, + "step": 489 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11300900900900901, + "grad_norm": 31.50993537902832, + "learning_rate": 9.879185815824595e-06, + "loss": 11.052852630615234, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.47241953387856483, + "loss_total_v6": 0.6908033043146133, + "step": 490 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11323963963963964, + "grad_norm": 28.694150924682617, + "learning_rate": 9.87835457885537e-06, + "loss": 11.27336597442627, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.48435527458786964, + "loss_total_v6": 0.7045853734016418, + "step": 491 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11347027027027028, + "grad_norm": 24.089975357055664, + "learning_rate": 9.877520527316317e-06, + "loss": 10.942562103271484, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.45977380126714706, + "loss_total_v6": 0.6839101612567902, + "step": 492 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1137009009009009, + "grad_norm": 23.445209503173828, + "learning_rate": 9.876683661688642e-06, + "loss": 11.180480003356934, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.47906871140003204, + "loss_total_v6": 0.6987800225615501, + "step": 493 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11393153153153153, + "grad_norm": 29.939395904541016, + "learning_rate": 9.875843982455176e-06, + "loss": 11.052606582641602, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.474174153059721, + "loss_total_v6": 0.6907879188656807, + "step": 494 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11416216216216216, + "grad_norm": 25.064191818237305, + "learning_rate": 9.875001490100372e-06, + "loss": 10.761079788208008, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.45532816648483276, + "loss_total_v6": 0.6725675091147423, + "step": 495 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1143927927927928, + "grad_norm": 27.568721771240234, + "learning_rate": 9.874156185110307e-06, + "loss": 11.233222007751465, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.4786266162991524, + "loss_total_v6": 0.7020763456821442, + "step": 496 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11462342342342342, + "grad_norm": 28.214569091796875, + "learning_rate": 9.873308067972679e-06, + "loss": 11.288830757141113, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.48356710746884346, + "loss_total_v6": 0.7055519372224808, + "step": 497 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11485405405405405, + "grad_norm": 26.93047332763672, + "learning_rate": 9.872457139176812e-06, + "loss": 11.116937637329102, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.47431905940175056, + "loss_total_v6": 0.6948085799813271, + "step": 498 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11508468468468469, + "grad_norm": 27.224868774414062, + "learning_rate": 9.871603399213647e-06, + "loss": 11.000957489013672, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.468901339918375, + "loss_total_v6": 0.6875597909092903, + "step": 499 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11531531531531532, + "grad_norm": 25.05263900756836, + "learning_rate": 9.870746848575751e-06, + "loss": 10.687437057495117, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4531821124255657, + "loss_total_v6": 0.667964830994606, + "step": 500 + }, + { + "epoch": 0.11531531531531532, + "eval_loss": 0.7102240920066833, + "eval_loss_alignment": 0.42021192240918803, + "eval_loss_alignment_w": 0.21010596120459402, + "eval_loss_sft": 0.5001181645287026, + "eval_loss_total_v6": 0.7102241262347779, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.5327, + "eval_samples_per_second": 22.887, + "eval_steps_per_second": 2.87, + "step": 500 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11554594594594594, + "grad_norm": 26.921615600585938, + "learning_rate": 9.869887487757312e-06, + "loss": 11.046420097351074, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.472154825925827, + "loss_total_v6": 0.6904012635350227, + "step": 501 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11577657657657657, + "grad_norm": 25.630756378173828, + "learning_rate": 9.869025317254139e-06, + "loss": 11.305912017822266, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.4795381873846054, + "loss_total_v6": 0.706619493663311, + "step": 502 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11600720720720721, + "grad_norm": 25.10771942138672, + "learning_rate": 9.868160337563662e-06, + "loss": 11.011528015136719, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.46997401118278503, + "loss_total_v6": 0.6882204711437225, + "step": 503 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11623783783783784, + "grad_norm": 25.297687530517578, + "learning_rate": 9.86729254918493e-06, + "loss": 11.222771644592285, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.4767375476658344, + "loss_total_v6": 0.7014232277870178, + "step": 504 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11646846846846846, + "grad_norm": 25.34276008605957, + "learning_rate": 9.866421952618616e-06, + "loss": 11.041790962219238, + "loss_alignment": 0.45166015625, + "loss_alignment_w": 0.225830078125, + "loss_sft": 0.4639613889157772, + "loss_total_v6": 0.6901119127869606, + "step": 505 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1166990990990991, + "grad_norm": 26.515172958374023, + "learning_rate": 9.86554854836701e-06, + "loss": 10.950626373291016, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.47166087478399277, + "loss_total_v6": 0.6844141855835915, + "step": 506 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11692972972972973, + "grad_norm": 25.221155166625977, + "learning_rate": 9.864672336934022e-06, + "loss": 11.409151077270508, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.49879271537065506, + "loss_total_v6": 0.7130719125270844, + "step": 507 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11716036036036036, + "grad_norm": 27.11290168762207, + "learning_rate": 9.863793318825186e-06, + "loss": 11.429054260253906, + "loss_alignment": 0.447998046875, + "loss_alignment_w": 0.2239990234375, + "loss_sft": 0.4906068444252014, + "loss_total_v6": 0.7143159136176109, + "step": 508 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11739099099099098, + "grad_norm": 25.347810745239258, + "learning_rate": 9.86291149454765e-06, + "loss": 10.872259140014648, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.46720536425709724, + "loss_total_v6": 0.6795161589980125, + "step": 509 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11762162162162162, + "grad_norm": 26.74480438232422, + "learning_rate": 9.862026864610183e-06, + "loss": 11.529205322265625, + "loss_alignment": 0.456298828125, + "loss_alignment_w": 0.2281494140625, + "loss_sft": 0.49219705536961555, + "loss_total_v6": 0.7205753177404404, + "step": 510 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11785225225225225, + "grad_norm": 24.958589553833008, + "learning_rate": 9.861139429523173e-06, + "loss": 11.556869506835938, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.5027761682868004, + "loss_total_v6": 0.7223043814301491, + "step": 511 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11808288288288288, + "grad_norm": 26.066490173339844, + "learning_rate": 9.860249189798628e-06, + "loss": 11.028762817382812, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4739198237657547, + "loss_total_v6": 0.6892976388335228, + "step": 512 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11831351351351352, + "grad_norm": 25.963958740234375, + "learning_rate": 9.859356145950169e-06, + "loss": 11.069575309753418, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4749600253999233, + "loss_total_v6": 0.6918484568595886, + "step": 513 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11854414414414415, + "grad_norm": 30.121706008911133, + "learning_rate": 9.85846029849304e-06, + "loss": 11.318191528320312, + "loss_alignment": 0.450927734375, + "loss_alignment_w": 0.2254638671875, + "loss_sft": 0.48210620135068893, + "loss_total_v6": 0.7073869407176971, + "step": 514 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11877477477477477, + "grad_norm": 25.8807430267334, + "learning_rate": 9.857561647944101e-06, + "loss": 10.916950225830078, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4675113968551159, + "loss_total_v6": 0.6823093891143799, + "step": 515 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11900540540540541, + "grad_norm": 24.34150505065918, + "learning_rate": 9.856660194821829e-06, + "loss": 11.121655464172363, + "loss_alignment": 0.44873046875, + "loss_alignment_w": 0.224365234375, + "loss_sft": 0.47105860710144043, + "loss_total_v6": 0.6951034665107727, + "step": 516 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11923603603603604, + "grad_norm": 28.021913528442383, + "learning_rate": 9.855755939646318e-06, + "loss": 11.720874786376953, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.5149185657501221, + "loss_total_v6": 0.732554703950882, + "step": 517 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11946666666666667, + "grad_norm": 28.923341751098633, + "learning_rate": 9.854848882939278e-06, + "loss": 11.066678047180176, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.4668291248381138, + "loss_total_v6": 0.691667377948761, + "step": 518 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11969729729729729, + "grad_norm": 25.625343322753906, + "learning_rate": 9.853939025224037e-06, + "loss": 11.42182445526123, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.499600175768137, + "loss_total_v6": 0.7138640508055687, + "step": 519 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.11992792792792793, + "grad_norm": 27.80303192138672, + "learning_rate": 9.853026367025535e-06, + "loss": 10.927528381347656, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.4606652073562145, + "loss_total_v6": 0.6829704865813255, + "step": 520 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12015855855855856, + "grad_norm": 25.23647117614746, + "learning_rate": 9.852110908870336e-06, + "loss": 11.504825592041016, + "loss_alignment": 0.450439453125, + "loss_alignment_w": 0.2252197265625, + "loss_sft": 0.49351146072149277, + "loss_total_v6": 0.7190516218543053, + "step": 521 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12038918918918919, + "grad_norm": 50.65693283081055, + "learning_rate": 9.85119265128661e-06, + "loss": 11.033626556396484, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4781758636236191, + "loss_total_v6": 0.6896016597747803, + "step": 522 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12061981981981983, + "grad_norm": 28.35624885559082, + "learning_rate": 9.850271594804147e-06, + "loss": 11.090773582458496, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4768342152237892, + "loss_total_v6": 0.693173311650753, + "step": 523 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12085045045045045, + "grad_norm": 31.74187660217285, + "learning_rate": 9.849347739954353e-06, + "loss": 11.062495231628418, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4757230021059513, + "loss_total_v6": 0.6914059743285179, + "step": 524 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12108108108108108, + "grad_norm": 25.41636848449707, + "learning_rate": 9.848421087270243e-06, + "loss": 10.547477722167969, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.44248148053884506, + "loss_total_v6": 0.6592173427343369, + "step": 525 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1213117117117117, + "grad_norm": 24.62541961669922, + "learning_rate": 9.847491637286455e-06, + "loss": 10.73086166381836, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.45063190907239914, + "loss_total_v6": 0.670678898692131, + "step": 526 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12154234234234235, + "grad_norm": 26.233854293823242, + "learning_rate": 9.846559390539232e-06, + "loss": 10.806599617004395, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.46348315104842186, + "loss_total_v6": 0.6754124909639359, + "step": 527 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12177297297297297, + "grad_norm": 30.272735595703125, + "learning_rate": 9.845624347566434e-06, + "loss": 11.135625839233398, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4846118539571762, + "loss_total_v6": 0.6959765926003456, + "step": 528 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1220036036036036, + "grad_norm": 26.180696487426758, + "learning_rate": 9.844686508907538e-06, + "loss": 10.920455932617188, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.45956704765558243, + "loss_total_v6": 0.682528480887413, + "step": 529 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12223423423423424, + "grad_norm": 25.26958465576172, + "learning_rate": 9.843745875103628e-06, + "loss": 10.686695098876953, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.44669653847813606, + "loss_total_v6": 0.6679184511303902, + "step": 530 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12246486486486487, + "grad_norm": 25.568145751953125, + "learning_rate": 9.842802446697403e-06, + "loss": 11.133543014526367, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.47759997844696045, + "loss_total_v6": 0.695846438407898, + "step": 531 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1226954954954955, + "grad_norm": 29.11138153076172, + "learning_rate": 9.841856224233175e-06, + "loss": 10.687569618225098, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4511915221810341, + "loss_total_v6": 0.6679731085896492, + "step": 532 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12292612612612612, + "grad_norm": 42.47186279296875, + "learning_rate": 9.840907208256869e-06, + "loss": 10.94589614868164, + "loss_alignment": 0.451171875, + "loss_alignment_w": 0.2255859375, + "loss_sft": 0.4586241543292999, + "loss_total_v6": 0.6841185241937637, + "step": 533 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12315675675675676, + "grad_norm": 25.168794631958008, + "learning_rate": 9.83995539931602e-06, + "loss": 10.333486557006836, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.42529238760471344, + "loss_total_v6": 0.6458429247140884, + "step": 534 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12338738738738739, + "grad_norm": 25.537538528442383, + "learning_rate": 9.839000797959775e-06, + "loss": 11.059288024902344, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4778723679482937, + "loss_total_v6": 0.6912054792046547, + "step": 535 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12361801801801801, + "grad_norm": 27.358427047729492, + "learning_rate": 9.838043404738889e-06, + "loss": 11.003782272338867, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4701613821089268, + "loss_total_v6": 0.687736414372921, + "step": 536 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12384864864864865, + "grad_norm": 25.934724807739258, + "learning_rate": 9.837083220205735e-06, + "loss": 10.449430465698242, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.43220318481326103, + "loss_total_v6": 0.6530894339084625, + "step": 537 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12407927927927928, + "grad_norm": 23.92965316772461, + "learning_rate": 9.83612024491429e-06, + "loss": 11.339990615844727, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4918305054306984, + "loss_total_v6": 0.708749420940876, + "step": 538 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12430990990990991, + "grad_norm": 25.33251953125, + "learning_rate": 9.835154479420141e-06, + "loss": 11.120677947998047, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.48019857704639435, + "loss_total_v6": 0.6950423419475555, + "step": 539 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12454054054054053, + "grad_norm": 30.96417808532715, + "learning_rate": 9.83418592428049e-06, + "loss": 11.328641891479492, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.49064819887280464, + "loss_total_v6": 0.7080401703715324, + "step": 540 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12477117117117117, + "grad_norm": 25.809385299682617, + "learning_rate": 9.833214580054145e-06, + "loss": 11.350399017333984, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.49090932309627533, + "loss_total_v6": 0.7093999162316322, + "step": 541 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12500180180180182, + "grad_norm": 24.196636199951172, + "learning_rate": 9.832240447301523e-06, + "loss": 10.382972717285156, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.4271797947585583, + "loss_total_v6": 0.6489357799291611, + "step": 542 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12523243243243243, + "grad_norm": 25.963359832763672, + "learning_rate": 9.831263526584651e-06, + "loss": 11.131057739257812, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4755983129143715, + "loss_total_v6": 0.6956910938024521, + "step": 543 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12546306306306307, + "grad_norm": 28.714574813842773, + "learning_rate": 9.830283818467164e-06, + "loss": 11.486273765563965, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.49322163686156273, + "loss_total_v6": 0.717892087996006, + "step": 544 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12569369369369368, + "grad_norm": 25.63113784790039, + "learning_rate": 9.829301323514307e-06, + "loss": 11.168208122253418, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.474364947527647, + "loss_total_v6": 0.6980130076408386, + "step": 545 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12592432432432432, + "grad_norm": 24.317636489868164, + "learning_rate": 9.828316042292927e-06, + "loss": 11.031980514526367, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.46609481796622276, + "loss_total_v6": 0.6894987598061562, + "step": 546 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12615495495495496, + "grad_norm": 26.216066360473633, + "learning_rate": 9.827327975371485e-06, + "loss": 10.856877326965332, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.45945389941334724, + "loss_total_v6": 0.6785548478364944, + "step": 547 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12638558558558557, + "grad_norm": 23.77033233642578, + "learning_rate": 9.826337123320046e-06, + "loss": 10.831605911254883, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.46548858284950256, + "loss_total_v6": 0.676975391805172, + "step": 548 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12661621621621622, + "grad_norm": 26.5935115814209, + "learning_rate": 9.825343486710286e-06, + "loss": 10.706244468688965, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.4505428597331047, + "loss_total_v6": 0.6691402718424797, + "step": 549 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12684684684684686, + "grad_norm": 25.162864685058594, + "learning_rate": 9.82434706611548e-06, + "loss": 10.691930770874023, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.45039599016308784, + "loss_total_v6": 0.6682457029819489, + "step": 550 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12707747747747747, + "grad_norm": 24.52174949645996, + "learning_rate": 9.823347862110515e-06, + "loss": 11.148059844970703, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.4745247922837734, + "loss_total_v6": 0.6967537775635719, + "step": 551 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1273081081081081, + "grad_norm": 24.941835403442383, + "learning_rate": 9.822345875271884e-06, + "loss": 11.254854202270508, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.495451170951128, + "loss_total_v6": 0.7034284397959709, + "step": 552 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12753873873873875, + "grad_norm": 22.695810317993164, + "learning_rate": 9.821341106177681e-06, + "loss": 11.000884056091309, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4663027934730053, + "loss_total_v6": 0.6875552535057068, + "step": 553 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12776936936936936, + "grad_norm": 25.91490936279297, + "learning_rate": 9.820333555407611e-06, + "loss": 11.43138313293457, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.49598613381385803, + "loss_total_v6": 0.714461475610733, + "step": 554 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.128, + "grad_norm": 27.763519287109375, + "learning_rate": 9.81932322354298e-06, + "loss": 11.398141860961914, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.48865949362516403, + "loss_total_v6": 0.7123838365077972, + "step": 555 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12823063063063064, + "grad_norm": 24.381853103637695, + "learning_rate": 9.818310111166699e-06, + "loss": 11.082942962646484, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.4724996015429497, + "loss_total_v6": 0.6926839575171471, + "step": 556 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12846126126126126, + "grad_norm": 23.561569213867188, + "learning_rate": 9.817294218863285e-06, + "loss": 10.485099792480469, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4425349682569504, + "loss_total_v6": 0.6553187668323517, + "step": 557 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1286918918918919, + "grad_norm": 24.838382720947266, + "learning_rate": 9.816275547218855e-06, + "loss": 10.463693618774414, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.42972246930003166, + "loss_total_v6": 0.6539808586239815, + "step": 558 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1289225225225225, + "grad_norm": 25.508251190185547, + "learning_rate": 9.815254096821136e-06, + "loss": 10.77515983581543, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.46011435240507126, + "loss_total_v6": 0.6734474748373032, + "step": 559 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12915315315315315, + "grad_norm": 26.281888961791992, + "learning_rate": 9.814229868259453e-06, + "loss": 10.971029281616211, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.46489468216896057, + "loss_total_v6": 0.6856893375515938, + "step": 560 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1293837837837838, + "grad_norm": 24.18878746032715, + "learning_rate": 9.813202862124735e-06, + "loss": 11.627304077148438, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.5064916610717773, + "loss_total_v6": 0.7267065122723579, + "step": 561 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1296144144144144, + "grad_norm": 23.316743850708008, + "learning_rate": 9.812173079009512e-06, + "loss": 11.137422561645508, + "loss_alignment": 0.45751953125, + "loss_alignment_w": 0.228759765625, + "loss_sft": 0.46746648848056793, + "loss_total_v6": 0.696088895201683, + "step": 562 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.12984504504504504, + "grad_norm": 22.614107131958008, + "learning_rate": 9.811140519507922e-06, + "loss": 11.395461082458496, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4933137036859989, + "loss_total_v6": 0.7122163102030754, + "step": 563 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13007567567567568, + "grad_norm": 24.85346031188965, + "learning_rate": 9.8101051842157e-06, + "loss": 10.871023178100586, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4584154151380062, + "loss_total_v6": 0.679438978433609, + "step": 564 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1303063063063063, + "grad_norm": 25.309188842773438, + "learning_rate": 9.809067073730178e-06, + "loss": 10.954638481140137, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.46206972375512123, + "loss_total_v6": 0.6846649274230003, + "step": 565 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13053693693693694, + "grad_norm": 25.10101890563965, + "learning_rate": 9.808026188650303e-06, + "loss": 11.618889808654785, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.5108332820236683, + "loss_total_v6": 0.7261805757880211, + "step": 566 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13076756756756758, + "grad_norm": 26.792787551879883, + "learning_rate": 9.806982529576608e-06, + "loss": 11.250219345092773, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.4807418882846832, + "loss_total_v6": 0.7031387165188789, + "step": 567 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1309981981981982, + "grad_norm": 27.172693252563477, + "learning_rate": 9.805936097111235e-06, + "loss": 10.958664894104004, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.47031694650650024, + "loss_total_v6": 0.6849165484309196, + "step": 568 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13122882882882883, + "grad_norm": 26.490598678588867, + "learning_rate": 9.804886891857924e-06, + "loss": 11.266362190246582, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.4837344251573086, + "loss_total_v6": 0.704147607088089, + "step": 569 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13145945945945947, + "grad_norm": 24.207311630249023, + "learning_rate": 9.803834914422015e-06, + "loss": 10.86634349822998, + "loss_alignment": 0.44775390625, + "loss_alignment_w": 0.223876953125, + "loss_sft": 0.45526956021785736, + "loss_total_v6": 0.6791464909911156, + "step": 570 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13169009009009008, + "grad_norm": 23.8433895111084, + "learning_rate": 9.802780165410446e-06, + "loss": 10.953775405883789, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.46624238044023514, + "loss_total_v6": 0.6846109256148338, + "step": 571 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13192072072072072, + "grad_norm": 23.31813621520996, + "learning_rate": 9.801722645431755e-06, + "loss": 11.274360656738281, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4879421517252922, + "loss_total_v6": 0.704647496342659, + "step": 572 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13215135135135136, + "grad_norm": 23.398639678955078, + "learning_rate": 9.80066235509608e-06, + "loss": 10.764522552490234, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.46028876304626465, + "loss_total_v6": 0.6727826371788979, + "step": 573 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13238198198198198, + "grad_norm": 26.351293563842773, + "learning_rate": 9.799599295015154e-06, + "loss": 10.68187141418457, + "loss_alignment": 0.4521484375, + "loss_alignment_w": 0.22607421875, + "loss_sft": 0.4413290433585644, + "loss_total_v6": 0.6676169186830521, + "step": 574 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13261261261261262, + "grad_norm": 25.96378517150879, + "learning_rate": 9.798533465802313e-06, + "loss": 10.840603828430176, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4574602134525776, + "loss_total_v6": 0.6775377467274666, + "step": 575 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13284324324324323, + "grad_norm": 22.25572967529297, + "learning_rate": 9.797464868072489e-06, + "loss": 10.5018310546875, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.43909455463290215, + "loss_total_v6": 0.6563644930720329, + "step": 576 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13307387387387387, + "grad_norm": 27.932174682617188, + "learning_rate": 9.796393502442205e-06, + "loss": 11.50114631652832, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.4934340752661228, + "loss_total_v6": 0.7188216596841812, + "step": 577 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1333045045045045, + "grad_norm": 24.40034294128418, + "learning_rate": 9.795319369529588e-06, + "loss": 11.578683853149414, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.5002790465950966, + "loss_total_v6": 0.723667748272419, + "step": 578 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13353513513513512, + "grad_norm": 24.414644241333008, + "learning_rate": 9.794242469954362e-06, + "loss": 11.22279167175293, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.4823997765779495, + "loss_total_v6": 0.7014244198799133, + "step": 579 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13376576576576577, + "grad_norm": 25.923254013061523, + "learning_rate": 9.793162804337844e-06, + "loss": 10.99160385131836, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.46575334295630455, + "loss_total_v6": 0.6869752705097198, + "step": 580 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1339963963963964, + "grad_norm": 26.955289840698242, + "learning_rate": 9.792080373302949e-06, + "loss": 10.793094635009766, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4539110101759434, + "loss_total_v6": 0.6745683625340462, + "step": 581 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13422702702702702, + "grad_norm": 25.928302764892578, + "learning_rate": 9.790995177474181e-06, + "loss": 10.76009750366211, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4533288888633251, + "loss_total_v6": 0.6725061237812042, + "step": 582 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13445765765765766, + "grad_norm": 27.38094139099121, + "learning_rate": 9.789907217477649e-06, + "loss": 11.141763687133789, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.47335299849510193, + "loss_total_v6": 0.69636020809412, + "step": 583 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1346882882882883, + "grad_norm": 26.065982818603516, + "learning_rate": 9.788816493941052e-06, + "loss": 11.383740425109863, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.48833923414349556, + "loss_total_v6": 0.711483784019947, + "step": 584 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1349189189189189, + "grad_norm": 26.688695907592773, + "learning_rate": 9.787723007493681e-06, + "loss": 11.194294929504395, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.48454026505351067, + "loss_total_v6": 0.6996434032917023, + "step": 585 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13514954954954955, + "grad_norm": 26.975141525268555, + "learning_rate": 9.786626758766427e-06, + "loss": 10.499940872192383, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.44007501751184464, + "loss_total_v6": 0.6562462970614433, + "step": 586 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1353801801801802, + "grad_norm": 26.265689849853516, + "learning_rate": 9.78552774839177e-06, + "loss": 11.21693229675293, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4875877760350704, + "loss_total_v6": 0.7010582387447357, + "step": 587 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1356108108108108, + "grad_norm": 25.184654235839844, + "learning_rate": 9.784425977003786e-06, + "loss": 11.38167667388916, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.4894614517688751, + "loss_total_v6": 0.7113547697663307, + "step": 588 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13584144144144145, + "grad_norm": 27.51082992553711, + "learning_rate": 9.783321445238138e-06, + "loss": 10.735422134399414, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4560438543558121, + "loss_total_v6": 0.6709638759493828, + "step": 589 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13607207207207206, + "grad_norm": 29.5749454498291, + "learning_rate": 9.782214153732093e-06, + "loss": 11.042062759399414, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.4758039824664593, + "loss_total_v6": 0.6901289150118828, + "step": 590 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1363027027027027, + "grad_norm": 23.62076187133789, + "learning_rate": 9.781104103124499e-06, + "loss": 11.173794746398926, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.48006999865174294, + "loss_total_v6": 0.6983622163534164, + "step": 591 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13653333333333334, + "grad_norm": 24.232397079467773, + "learning_rate": 9.779991294055802e-06, + "loss": 10.89824104309082, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.46173394098877907, + "loss_total_v6": 0.6811400502920151, + "step": 592 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13676396396396395, + "grad_norm": 23.156021118164062, + "learning_rate": 9.778875727168038e-06, + "loss": 10.738509178161621, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.4502858445048332, + "loss_total_v6": 0.6711568012833595, + "step": 593 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1369945945945946, + "grad_norm": 25.332181930541992, + "learning_rate": 9.777757403104833e-06, + "loss": 10.894134521484375, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.46652794629335403, + "loss_total_v6": 0.680883415043354, + "step": 594 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13722522522522523, + "grad_norm": 28.821250915527344, + "learning_rate": 9.776636322511405e-06, + "loss": 10.812444686889648, + "loss_alignment": 0.45458984375, + "loss_alignment_w": 0.227294921875, + "loss_sft": 0.44849811866879463, + "loss_total_v6": 0.6757777780294418, + "step": 595 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13745585585585585, + "grad_norm": 23.137754440307617, + "learning_rate": 9.775512486034564e-06, + "loss": 11.051257133483887, + "loss_alignment": 0.44677734375, + "loss_alignment_w": 0.223388671875, + "loss_sft": 0.46743694692850113, + "loss_total_v6": 0.6907035484910011, + "step": 596 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1376864864864865, + "grad_norm": 30.104324340820312, + "learning_rate": 9.774385894322705e-06, + "loss": 10.953548431396484, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.4622914344072342, + "loss_total_v6": 0.6845967397093773, + "step": 597 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13791711711711713, + "grad_norm": 22.598432540893555, + "learning_rate": 9.773256548025817e-06, + "loss": 10.634014129638672, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.449064914137125, + "loss_total_v6": 0.6646258533000946, + "step": 598 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13814774774774774, + "grad_norm": 24.925621032714844, + "learning_rate": 9.772124447795482e-06, + "loss": 11.231783866882324, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.48305339738726616, + "loss_total_v6": 0.7019865065813065, + "step": 599 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13837837837837838, + "grad_norm": 23.149938583374023, + "learning_rate": 9.770989594284858e-06, + "loss": 10.499706268310547, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.4421813078224659, + "loss_total_v6": 0.6562316045165062, + "step": 600 + }, + { + "epoch": 0.13837837837837838, + "eval_loss": 0.7051356434822083, + "eval_loss_alignment": 0.41759314903846156, + "eval_loss_alignment_w": 0.20879657451923078, + "eval_loss_sft": 0.4963390991951411, + "eval_loss_total_v6": 0.7051356744466939, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.295, + "eval_samples_per_second": 22.953, + "eval_steps_per_second": 2.878, + "step": 600 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13860900900900902, + "grad_norm": 24.98779296875, + "learning_rate": 9.769851988148704e-06, + "loss": 11.107508659362793, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.47046446055173874, + "loss_total_v6": 0.6942193061113358, + "step": 601 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13883963963963963, + "grad_norm": 24.573104858398438, + "learning_rate": 9.768711630043362e-06, + "loss": 11.22426986694336, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.49050306528806686, + "loss_total_v6": 0.7015168741345406, + "step": 602 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13907027027027027, + "grad_norm": 24.67538070678711, + "learning_rate": 9.767568520626762e-06, + "loss": 10.78637409210205, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.4503934867680073, + "loss_total_v6": 0.674148365855217, + "step": 603 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13930090090090091, + "grad_norm": 25.931480407714844, + "learning_rate": 9.76642266055842e-06, + "loss": 10.80766487121582, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4606963209807873, + "loss_total_v6": 0.675479032099247, + "step": 604 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13953153153153153, + "grad_norm": 28.56105613708496, + "learning_rate": 9.765274050499445e-06, + "loss": 11.002561569213867, + "loss_alignment": 0.448974609375, + "loss_alignment_w": 0.2244873046875, + "loss_sft": 0.46335597336292267, + "loss_total_v6": 0.6876601427793503, + "step": 605 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13976216216216217, + "grad_norm": 26.572147369384766, + "learning_rate": 9.764122691112525e-06, + "loss": 10.737751960754395, + "loss_alignment": 0.447998046875, + "loss_alignment_w": 0.2239990234375, + "loss_sft": 0.44721728190779686, + "loss_total_v6": 0.6711094975471497, + "step": 606 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.13999279279279278, + "grad_norm": 25.470537185668945, + "learning_rate": 9.762968583061938e-06, + "loss": 11.536836624145508, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.503202598541975, + "loss_total_v6": 0.721052311360836, + "step": 607 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14022342342342342, + "grad_norm": 23.860973358154297, + "learning_rate": 9.761811727013548e-06, + "loss": 11.035518646240234, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.47600531950592995, + "loss_total_v6": 0.6897198930382729, + "step": 608 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14045405405405406, + "grad_norm": 23.702199935913086, + "learning_rate": 9.760652123634804e-06, + "loss": 10.776127815246582, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.4537051171064377, + "loss_total_v6": 0.6735079884529114, + "step": 609 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14068468468468467, + "grad_norm": 23.58705711364746, + "learning_rate": 9.759489773594739e-06, + "loss": 10.83289909362793, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.4529656507074833, + "loss_total_v6": 0.677056185901165, + "step": 610 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14091531531531531, + "grad_norm": 24.258682250976562, + "learning_rate": 9.758324677563972e-06, + "loss": 10.938247680664062, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.4728403314948082, + "loss_total_v6": 0.6836404949426651, + "step": 611 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14114594594594596, + "grad_norm": 25.413434982299805, + "learning_rate": 9.757156836214707e-06, + "loss": 10.560359954833984, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.4459111802279949, + "loss_total_v6": 0.6600225046277046, + "step": 612 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14137657657657657, + "grad_norm": 24.979116439819336, + "learning_rate": 9.755986250220728e-06, + "loss": 11.513162612915039, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.5007464103400707, + "loss_total_v6": 0.7195726782083511, + "step": 613 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1416072072072072, + "grad_norm": 24.912803649902344, + "learning_rate": 9.754812920257409e-06, + "loss": 10.71670913696289, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4512731693685055, + "loss_total_v6": 0.6697942763566971, + "step": 614 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14183783783783785, + "grad_norm": 26.806520462036133, + "learning_rate": 9.753636847001701e-06, + "loss": 11.002567291259766, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.46630117669701576, + "loss_total_v6": 0.6876604184508324, + "step": 615 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14206846846846846, + "grad_norm": 24.824073791503906, + "learning_rate": 9.752458031132142e-06, + "loss": 10.775361061096191, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.45179568231105804, + "loss_total_v6": 0.6734600886702538, + "step": 616 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1422990990990991, + "grad_norm": 26.322111129760742, + "learning_rate": 9.751276473328848e-06, + "loss": 10.904541015625, + "loss_alignment": 0.445556640625, + "loss_alignment_w": 0.2227783203125, + "loss_sft": 0.4589081294834614, + "loss_total_v6": 0.6815338656306267, + "step": 617 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14252972972972974, + "grad_norm": 24.835485458374023, + "learning_rate": 9.75009217427352e-06, + "loss": 10.743158340454102, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.45120206475257874, + "loss_total_v6": 0.6714474111795425, + "step": 618 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14276036036036036, + "grad_norm": 26.77558708190918, + "learning_rate": 9.748905134649444e-06, + "loss": 11.20828628540039, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4811422899365425, + "loss_total_v6": 0.7005179077386856, + "step": 619 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.142990990990991, + "grad_norm": 24.932727813720703, + "learning_rate": 9.747715355141478e-06, + "loss": 10.732951164245605, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.44772594794631004, + "loss_total_v6": 0.6708094477653503, + "step": 620 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1432216216216216, + "grad_norm": 25.90454864501953, + "learning_rate": 9.74652283643607e-06, + "loss": 10.897183418273926, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.45925694704055786, + "loss_total_v6": 0.6810739487409592, + "step": 621 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14345225225225225, + "grad_norm": 27.010408401489258, + "learning_rate": 9.745327579221241e-06, + "loss": 10.936285018920898, + "loss_alignment": 0.454345703125, + "loss_alignment_w": 0.2271728515625, + "loss_sft": 0.45634499192237854, + "loss_total_v6": 0.6835178211331367, + "step": 622 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1436828828828829, + "grad_norm": 23.58951187133789, + "learning_rate": 9.744129584186599e-06, + "loss": 11.078348159790039, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.46992358192801476, + "loss_total_v6": 0.692396730184555, + "step": 623 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1439135135135135, + "grad_norm": 24.098535537719727, + "learning_rate": 9.742928852023326e-06, + "loss": 11.022672653198242, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.46839695051312447, + "loss_total_v6": 0.688916988670826, + "step": 624 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14414414414414414, + "grad_norm": 26.07022476196289, + "learning_rate": 9.741725383424185e-06, + "loss": 11.0472412109375, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.46578217670321465, + "loss_total_v6": 0.6904525831341743, + "step": 625 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14437477477477478, + "grad_norm": 24.345279693603516, + "learning_rate": 9.740519179083519e-06, + "loss": 11.250428199768066, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.48211298882961273, + "loss_total_v6": 0.7031517922878265, + "step": 626 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1446054054054054, + "grad_norm": 24.426435470581055, + "learning_rate": 9.739310239697248e-06, + "loss": 10.726875305175781, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4521680660545826, + "loss_total_v6": 0.6704297661781311, + "step": 627 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14483603603603604, + "grad_norm": 26.704517364501953, + "learning_rate": 9.73809856596287e-06, + "loss": 11.195884704589844, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.480596125125885, + "loss_total_v6": 0.6997428387403488, + "step": 628 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14506666666666668, + "grad_norm": 25.195356369018555, + "learning_rate": 9.73688415857946e-06, + "loss": 10.735472679138184, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.45339198410511017, + "loss_total_v6": 0.6709670498967171, + "step": 629 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1452972972972973, + "grad_norm": 28.483427047729492, + "learning_rate": 9.735667018247674e-06, + "loss": 10.87966251373291, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4632278010249138, + "loss_total_v6": 0.6799788922071457, + "step": 630 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14552792792792793, + "grad_norm": 25.770994186401367, + "learning_rate": 9.73444714566974e-06, + "loss": 10.87991714477539, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4665702022612095, + "loss_total_v6": 0.6799948513507843, + "step": 631 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14575855855855857, + "grad_norm": 26.77912139892578, + "learning_rate": 9.733224541549464e-06, + "loss": 10.832841873168945, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4600725844502449, + "loss_total_v6": 0.6770525872707367, + "step": 632 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14598918918918918, + "grad_norm": 26.127485275268555, + "learning_rate": 9.731999206592229e-06, + "loss": 10.704217910766602, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.4448620490729809, + "loss_total_v6": 0.6690136194229126, + "step": 633 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14621981981981982, + "grad_norm": 27.10730743408203, + "learning_rate": 9.730771141504993e-06, + "loss": 11.012162208557129, + "loss_alignment": 0.456298828125, + "loss_alignment_w": 0.2281494140625, + "loss_sft": 0.46040067449212074, + "loss_total_v6": 0.6882601603865623, + "step": 634 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14645045045045044, + "grad_norm": 27.17549705505371, + "learning_rate": 9.729540346996286e-06, + "loss": 11.599630355834961, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.5009626187384129, + "loss_total_v6": 0.7249768897891045, + "step": 635 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14668108108108108, + "grad_norm": 25.55482292175293, + "learning_rate": 9.72830682377622e-06, + "loss": 11.126571655273438, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4805365018546581, + "loss_total_v6": 0.6954107359051704, + "step": 636 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14691171171171172, + "grad_norm": 26.59814453125, + "learning_rate": 9.727070572556476e-06, + "loss": 10.775904655456543, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.45889442414045334, + "loss_total_v6": 0.6734940484166145, + "step": 637 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14714234234234233, + "grad_norm": 26.473236083984375, + "learning_rate": 9.725831594050307e-06, + "loss": 10.684648513793945, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.45592227950692177, + "loss_total_v6": 0.667790561914444, + "step": 638 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14737297297297297, + "grad_norm": 27.552547454833984, + "learning_rate": 9.724589888972547e-06, + "loss": 11.239826202392578, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.48482248559594154, + "loss_total_v6": 0.7024891227483749, + "step": 639 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1476036036036036, + "grad_norm": 24.61324119567871, + "learning_rate": 9.723345458039595e-06, + "loss": 11.050281524658203, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4746849611401558, + "loss_total_v6": 0.6906426101922989, + "step": 640 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14783423423423422, + "grad_norm": 25.987308502197266, + "learning_rate": 9.722098301969427e-06, + "loss": 11.032499313354492, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.47746455296874046, + "loss_total_v6": 0.6895311921834946, + "step": 641 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14806486486486486, + "grad_norm": 25.14360809326172, + "learning_rate": 9.720848421481592e-06, + "loss": 10.811304092407227, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.46476898714900017, + "loss_total_v6": 0.6757064759731293, + "step": 642 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1482954954954955, + "grad_norm": 30.223711013793945, + "learning_rate": 9.719595817297207e-06, + "loss": 10.552057266235352, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4463229924440384, + "loss_total_v6": 0.6595035642385483, + "step": 643 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14852612612612612, + "grad_norm": 24.698974609375, + "learning_rate": 9.718340490138965e-06, + "loss": 10.874759674072266, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.4626619443297386, + "loss_total_v6": 0.6796724423766136, + "step": 644 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14875675675675676, + "grad_norm": 26.18997573852539, + "learning_rate": 9.717082440731125e-06, + "loss": 10.695086479187012, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.44912830740213394, + "loss_total_v6": 0.6684428825974464, + "step": 645 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1489873873873874, + "grad_norm": 28.391651153564453, + "learning_rate": 9.715821669799521e-06, + "loss": 10.723812103271484, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4620473198592663, + "loss_total_v6": 0.6702382192015648, + "step": 646 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.149218018018018, + "grad_norm": 27.240148544311523, + "learning_rate": 9.714558178071559e-06, + "loss": 10.774825096130371, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4558667466044426, + "loss_total_v6": 0.6734265685081482, + "step": 647 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.14944864864864865, + "grad_norm": 24.22049331665039, + "learning_rate": 9.713291966276205e-06, + "loss": 11.587471008300781, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.5018353052437305, + "loss_total_v6": 0.7242169454693794, + "step": 648 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1496792792792793, + "grad_norm": 27.596532821655273, + "learning_rate": 9.712023035144004e-06, + "loss": 10.871593475341797, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4709480367600918, + "loss_total_v6": 0.6794746369123459, + "step": 649 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1499099099099099, + "grad_norm": 24.518674850463867, + "learning_rate": 9.710751385407068e-06, + "loss": 11.081573486328125, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.47255127504467964, + "loss_total_v6": 0.6925982981920242, + "step": 650 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15014054054054055, + "grad_norm": 27.885793685913086, + "learning_rate": 9.709477017799076e-06, + "loss": 10.792703628540039, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.46317923441529274, + "loss_total_v6": 0.6745439693331718, + "step": 651 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15037117117117116, + "grad_norm": 23.04557228088379, + "learning_rate": 9.708199933055272e-06, + "loss": 10.548033714294434, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.4412040002644062, + "loss_total_v6": 0.6592520996928215, + "step": 652 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1506018018018018, + "grad_norm": 23.996599197387695, + "learning_rate": 9.706920131912473e-06, + "loss": 11.111205101013184, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.48245997354388237, + "loss_total_v6": 0.6944503337144852, + "step": 653 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15083243243243244, + "grad_norm": 23.797075271606445, + "learning_rate": 9.70563761510906e-06, + "loss": 11.353161811828613, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.492623183876276, + "loss_total_v6": 0.7095726355910301, + "step": 654 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15106306306306305, + "grad_norm": 22.52061653137207, + "learning_rate": 9.704352383384984e-06, + "loss": 10.827045440673828, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.45371363312005997, + "loss_total_v6": 0.6766903027892113, + "step": 655 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1512936936936937, + "grad_norm": 22.262659072875977, + "learning_rate": 9.70306443748176e-06, + "loss": 11.259531021118164, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.48695437982678413, + "loss_total_v6": 0.7037207037210464, + "step": 656 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15152432432432433, + "grad_norm": 25.19525909423828, + "learning_rate": 9.701773778142467e-06, + "loss": 10.440370559692383, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.43679436668753624, + "loss_total_v6": 0.6525231227278709, + "step": 657 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15175495495495495, + "grad_norm": 23.385026931762695, + "learning_rate": 9.700480406111752e-06, + "loss": 11.192622184753418, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4795071966946125, + "loss_total_v6": 0.699538916349411, + "step": 658 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1519855855855856, + "grad_norm": 25.18726921081543, + "learning_rate": 9.69918432213583e-06, + "loss": 11.189081192016602, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4876476898789406, + "loss_total_v6": 0.6993176117539406, + "step": 659 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15221621621621623, + "grad_norm": 26.541534423828125, + "learning_rate": 9.697885526962474e-06, + "loss": 10.971565246582031, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.46875814348459244, + "loss_total_v6": 0.685722827911377, + "step": 660 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15244684684684684, + "grad_norm": 26.007034301757812, + "learning_rate": 9.696584021341025e-06, + "loss": 11.029805183410645, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.47473269328475, + "loss_total_v6": 0.6893628090620041, + "step": 661 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15267747747747748, + "grad_norm": 24.755931854248047, + "learning_rate": 9.695279806022391e-06, + "loss": 11.645208358764648, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.5006069205701351, + "loss_total_v6": 0.7278255373239517, + "step": 662 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15290810810810812, + "grad_norm": 22.53142547607422, + "learning_rate": 9.693972881759036e-06, + "loss": 10.682865142822266, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.453216765075922, + "loss_total_v6": 0.6676790788769722, + "step": 663 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15313873873873873, + "grad_norm": 23.656463623046875, + "learning_rate": 9.692663249304992e-06, + "loss": 10.915420532226562, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.4594201594591141, + "loss_total_v6": 0.682213768362999, + "step": 664 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15336936936936937, + "grad_norm": 23.96087074279785, + "learning_rate": 9.691350909415853e-06, + "loss": 10.931034088134766, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.46459220722317696, + "loss_total_v6": 0.6831896156072617, + "step": 665 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1536, + "grad_norm": 26.70453453063965, + "learning_rate": 9.690035862848773e-06, + "loss": 10.71902847290039, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4563925266265869, + "loss_total_v6": 0.6699392721056938, + "step": 666 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15383063063063063, + "grad_norm": 24.411462783813477, + "learning_rate": 9.688718110362472e-06, + "loss": 11.300695419311523, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4928993098437786, + "loss_total_v6": 0.7062934637069702, + "step": 667 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15406126126126127, + "grad_norm": 22.63142967224121, + "learning_rate": 9.687397652717223e-06, + "loss": 10.825883865356445, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.46279629319906235, + "loss_total_v6": 0.676617719233036, + "step": 668 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15429189189189188, + "grad_norm": 27.080415725708008, + "learning_rate": 9.686074490674871e-06, + "loss": 10.883611679077148, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4644054286181927, + "loss_total_v6": 0.68022570759058, + "step": 669 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15452252252252252, + "grad_norm": 24.349700927734375, + "learning_rate": 9.68474862499881e-06, + "loss": 10.371864318847656, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4332146793603897, + "loss_total_v6": 0.6482415273785591, + "step": 670 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15475315315315316, + "grad_norm": 24.69081687927246, + "learning_rate": 9.683420056454e-06, + "loss": 10.631478309631348, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.45066118240356445, + "loss_total_v6": 0.6644673570990562, + "step": 671 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15498378378378377, + "grad_norm": 24.150482177734375, + "learning_rate": 9.682088785806964e-06, + "loss": 10.38918399810791, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4364791437983513, + "loss_total_v6": 0.6493239924311638, + "step": 672 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15521441441441441, + "grad_norm": 28.323728561401367, + "learning_rate": 9.680754813825774e-06, + "loss": 10.849044799804688, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.45708753541111946, + "loss_total_v6": 0.678065299987793, + "step": 673 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15544504504504505, + "grad_norm": 24.703718185424805, + "learning_rate": 9.679418141280068e-06, + "loss": 10.710352897644043, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4531495049595833, + "loss_total_v6": 0.6693970710039139, + "step": 674 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15567567567567567, + "grad_norm": 23.899608612060547, + "learning_rate": 9.678078768941039e-06, + "loss": 10.389158248901367, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.44076526165008545, + "loss_total_v6": 0.649322398006916, + "step": 675 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1559063063063063, + "grad_norm": 25.86900520324707, + "learning_rate": 9.676736697581441e-06, + "loss": 10.909463882446289, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.46620434150099754, + "loss_total_v6": 0.6818415448069572, + "step": 676 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15613693693693695, + "grad_norm": 27.22681999206543, + "learning_rate": 9.675391927975584e-06, + "loss": 10.809085845947266, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4569857195019722, + "loss_total_v6": 0.6755678728222847, + "step": 677 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15636756756756756, + "grad_norm": 23.40148162841797, + "learning_rate": 9.674044460899329e-06, + "loss": 10.591865539550781, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.45018432661890984, + "loss_total_v6": 0.6619915589690208, + "step": 678 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1565981981981982, + "grad_norm": 23.469074249267578, + "learning_rate": 9.672694297130101e-06, + "loss": 10.42073917388916, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.43430095911026, + "loss_total_v6": 0.6512962207198143, + "step": 679 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15682882882882884, + "grad_norm": 24.838491439819336, + "learning_rate": 9.671341437446878e-06, + "loss": 10.517995834350586, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4437517151236534, + "loss_total_v6": 0.6573747619986534, + "step": 680 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15705945945945946, + "grad_norm": 23.241106033325195, + "learning_rate": 9.66998588263019e-06, + "loss": 10.543794631958008, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4476681500673294, + "loss_total_v6": 0.6589871197938919, + "step": 681 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1572900900900901, + "grad_norm": 23.5855655670166, + "learning_rate": 9.66862763346213e-06, + "loss": 11.039323806762695, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.47876080870628357, + "loss_total_v6": 0.6899577304720879, + "step": 682 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1575207207207207, + "grad_norm": 24.16090202331543, + "learning_rate": 9.66726669072634e-06, + "loss": 11.220783233642578, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.48213696107268333, + "loss_total_v6": 0.7012989297509193, + "step": 683 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15775135135135135, + "grad_norm": 23.751619338989258, + "learning_rate": 9.665903055208013e-06, + "loss": 10.78554916381836, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.46340350806713104, + "loss_total_v6": 0.6740968376398087, + "step": 684 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.157981981981982, + "grad_norm": 24.098241806030273, + "learning_rate": 9.664536727693903e-06, + "loss": 10.655415534973145, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.45401889830827713, + "loss_total_v6": 0.6659634709358215, + "step": 685 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1582126126126126, + "grad_norm": 23.4880428314209, + "learning_rate": 9.663167708972312e-06, + "loss": 11.159411430358887, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4766380302608013, + "loss_total_v6": 0.6974632143974304, + "step": 686 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15844324324324324, + "grad_norm": 22.286043167114258, + "learning_rate": 9.661795999833098e-06, + "loss": 10.950610160827637, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4684249609708786, + "loss_total_v6": 0.6844131350517273, + "step": 687 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15867387387387388, + "grad_norm": 23.62272071838379, + "learning_rate": 9.660421601067667e-06, + "loss": 10.655012130737305, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.44465531781315804, + "loss_total_v6": 0.6659382581710815, + "step": 688 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1589045045045045, + "grad_norm": 23.243642807006836, + "learning_rate": 9.659044513468978e-06, + "loss": 10.744512557983398, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.4521717242896557, + "loss_total_v6": 0.671532079577446, + "step": 689 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15913513513513514, + "grad_norm": 25.65534210205078, + "learning_rate": 9.657664737831548e-06, + "loss": 10.791129112243652, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.4555277079343796, + "loss_total_v6": 0.6744455620646477, + "step": 690 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15936576576576578, + "grad_norm": 27.065427780151367, + "learning_rate": 9.656282274951433e-06, + "loss": 10.980661392211914, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.46152936667203903, + "loss_total_v6": 0.6862913072109222, + "step": 691 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1595963963963964, + "grad_norm": 23.662918090820312, + "learning_rate": 9.654897125626253e-06, + "loss": 10.552959442138672, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4433123953640461, + "loss_total_v6": 0.6595599502325058, + "step": 692 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.15982702702702703, + "grad_norm": 21.49384880065918, + "learning_rate": 9.653509290655166e-06, + "loss": 10.863204956054688, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.45696547999978065, + "loss_total_v6": 0.6789503321051598, + "step": 693 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16005765765765767, + "grad_norm": 24.655380249023438, + "learning_rate": 9.652118770838884e-06, + "loss": 10.659088134765625, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.44762609153985977, + "loss_total_v6": 0.6661929860711098, + "step": 694 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16028828828828828, + "grad_norm": 24.29806137084961, + "learning_rate": 9.650725566979671e-06, + "loss": 10.583324432373047, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4473922811448574, + "loss_total_v6": 0.6614577993750572, + "step": 695 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16051891891891892, + "grad_norm": 26.20797348022461, + "learning_rate": 9.649329679881334e-06, + "loss": 10.491327285766602, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4461590312421322, + "loss_total_v6": 0.655707985162735, + "step": 696 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16074954954954954, + "grad_norm": 22.835487365722656, + "learning_rate": 9.647931110349233e-06, + "loss": 10.8250732421875, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.46286768838763237, + "loss_total_v6": 0.6765670627355576, + "step": 697 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16098018018018018, + "grad_norm": 26.478750228881836, + "learning_rate": 9.646529859190273e-06, + "loss": 11.09925651550293, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4802941530942917, + "loss_total_v6": 0.6937035322189331, + "step": 698 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16121081081081082, + "grad_norm": 23.343669891357422, + "learning_rate": 9.645125927212907e-06, + "loss": 10.923614501953125, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.4651661179959774, + "loss_total_v6": 0.6827259212732315, + "step": 699 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16144144144144143, + "grad_norm": 23.84650993347168, + "learning_rate": 9.643719315227134e-06, + "loss": 11.323736190795898, + "loss_alignment": 0.44921875, + "loss_alignment_w": 0.224609375, + "loss_sft": 0.48303263634443283, + "loss_total_v6": 0.7077335640788078, + "step": 700 + }, + { + "epoch": 0.16144144144144143, + "eval_loss": 0.7037755846977234, + "eval_loss_alignment": 0.4148742154113248, + "eval_loss_alignment_w": 0.2074371077056624, + "eval_loss_sft": 0.4963385070570641, + "eval_loss_total_v6": 0.7037756143567654, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.4528, + "eval_samples_per_second": 22.909, + "eval_steps_per_second": 2.873, + "step": 700 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16167207207207207, + "grad_norm": 25.606155395507812, + "learning_rate": 9.642310024044499e-06, + "loss": 10.888256072998047, + "loss_alignment": 0.443603515625, + "loss_alignment_w": 0.2218017578125, + "loss_sft": 0.45854633301496506, + "loss_total_v6": 0.6805159747600555, + "step": 701 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1619027027027027, + "grad_norm": 25.443452835083008, + "learning_rate": 9.640898054478095e-06, + "loss": 10.949325561523438, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.46303461119532585, + "loss_total_v6": 0.6843328550457954, + "step": 702 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16213333333333332, + "grad_norm": 22.789796829223633, + "learning_rate": 9.639483407342557e-06, + "loss": 10.43932819366455, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4376752935349941, + "loss_total_v6": 0.6524580046534538, + "step": 703 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16236396396396396, + "grad_norm": 22.99015998840332, + "learning_rate": 9.63806608345407e-06, + "loss": 10.637579917907715, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.4444660134613514, + "loss_total_v6": 0.6648487225174904, + "step": 704 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1625945945945946, + "grad_norm": 24.504329681396484, + "learning_rate": 9.636646083630356e-06, + "loss": 10.647965431213379, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4502268508076668, + "loss_total_v6": 0.6654978469014168, + "step": 705 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16282522522522522, + "grad_norm": 26.074241638183594, + "learning_rate": 9.635223408690688e-06, + "loss": 11.35641860961914, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.4899733252823353, + "loss_total_v6": 0.7097761854529381, + "step": 706 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16305585585585586, + "grad_norm": 25.318880081176758, + "learning_rate": 9.633798059455877e-06, + "loss": 10.365354537963867, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4366377405822277, + "loss_total_v6": 0.6478346511721611, + "step": 707 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1632864864864865, + "grad_norm": 24.559280395507812, + "learning_rate": 9.63237003674828e-06, + "loss": 10.562300682067871, + "loss_alignment": 0.45263671875, + "loss_alignment_w": 0.226318359375, + "loss_sft": 0.43330664560198784, + "loss_total_v6": 0.6601438075304031, + "step": 708 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1635171171171171, + "grad_norm": 22.94980239868164, + "learning_rate": 9.630939341391796e-06, + "loss": 10.493569374084473, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.43580105528235435, + "loss_total_v6": 0.6558480635285378, + "step": 709 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16374774774774775, + "grad_norm": 23.515705108642578, + "learning_rate": 9.629505974211863e-06, + "loss": 10.831907272338867, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4591292068362236, + "loss_total_v6": 0.6769941970705986, + "step": 710 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1639783783783784, + "grad_norm": 26.161670684814453, + "learning_rate": 9.628069936035467e-06, + "loss": 11.073456764221191, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.48377805203199387, + "loss_total_v6": 0.6920910477638245, + "step": 711 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.164209009009009, + "grad_norm": 25.05362319946289, + "learning_rate": 9.626631227691128e-06, + "loss": 10.650388717651367, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.45114124566316605, + "loss_total_v6": 0.665649302303791, + "step": 712 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16443963963963965, + "grad_norm": 23.456832885742188, + "learning_rate": 9.625189850008912e-06, + "loss": 10.370040893554688, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.43142225965857506, + "loss_total_v6": 0.6481275781989098, + "step": 713 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16467027027027026, + "grad_norm": 23.53427505493164, + "learning_rate": 9.623745803820422e-06, + "loss": 10.849540710449219, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.458857998251915, + "loss_total_v6": 0.6780963093042374, + "step": 714 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1649009009009009, + "grad_norm": 24.24706268310547, + "learning_rate": 9.6222990899588e-06, + "loss": 10.912517547607422, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.46692924201488495, + "loss_total_v6": 0.6820323541760445, + "step": 715 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16513153153153154, + "grad_norm": 24.090251922607422, + "learning_rate": 9.62084970925873e-06, + "loss": 10.845343589782715, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4608539901673794, + "loss_total_v6": 0.6778339594602585, + "step": 716 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16536216216216215, + "grad_norm": 23.126201629638672, + "learning_rate": 9.619397662556434e-06, + "loss": 11.109869003295898, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.47390784323215485, + "loss_total_v6": 0.694366842508316, + "step": 717 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1655927927927928, + "grad_norm": 23.49302864074707, + "learning_rate": 9.617942950689671e-06, + "loss": 10.835298538208008, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.45979899168014526, + "loss_total_v6": 0.6772062033414841, + "step": 718 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16582342342342343, + "grad_norm": 22.927513122558594, + "learning_rate": 9.616485574497734e-06, + "loss": 10.58169937133789, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.442835059016943, + "loss_total_v6": 0.6613561734557152, + "step": 719 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16605405405405405, + "grad_norm": 23.506696701049805, + "learning_rate": 9.615025534821463e-06, + "loss": 10.503608703613281, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.44051792100071907, + "loss_total_v6": 0.6564755737781525, + "step": 720 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16628468468468469, + "grad_norm": 25.249629974365234, + "learning_rate": 9.613562832503225e-06, + "loss": 10.877130508422852, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.46735721081495285, + "loss_total_v6": 0.6798206269741058, + "step": 721 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16651531531531533, + "grad_norm": 24.05678367614746, + "learning_rate": 9.612097468386927e-06, + "loss": 10.860804557800293, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.4575020372867584, + "loss_total_v6": 0.6788002625107765, + "step": 722 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16674594594594594, + "grad_norm": 23.05959701538086, + "learning_rate": 9.610629443318015e-06, + "loss": 10.702912330627441, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4541035108268261, + "loss_total_v6": 0.6689320206642151, + "step": 723 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16697657657657658, + "grad_norm": 25.702558517456055, + "learning_rate": 9.609158758143464e-06, + "loss": 10.56045913696289, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.449274268001318, + "loss_total_v6": 0.6600286588072777, + "step": 724 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16720720720720722, + "grad_norm": 21.975740432739258, + "learning_rate": 9.607685413711786e-06, + "loss": 10.428719520568848, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4381566643714905, + "loss_total_v6": 0.6517949774861336, + "step": 725 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16743783783783783, + "grad_norm": 27.3087100982666, + "learning_rate": 9.606209410873032e-06, + "loss": 10.720422744750977, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.44482192024588585, + "loss_total_v6": 0.6700264066457748, + "step": 726 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16766846846846847, + "grad_norm": 25.86197853088379, + "learning_rate": 9.60473075047878e-06, + "loss": 10.713800430297852, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.45782050862908363, + "loss_total_v6": 0.6696124896407127, + "step": 727 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1678990990990991, + "grad_norm": 24.227008819580078, + "learning_rate": 9.603249433382145e-06, + "loss": 11.014806747436523, + "loss_alignment": 0.455322265625, + "loss_alignment_w": 0.2276611328125, + "loss_sft": 0.46125249192118645, + "loss_total_v6": 0.6884253770112991, + "step": 728 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16812972972972973, + "grad_norm": 23.851606369018555, + "learning_rate": 9.601765460437774e-06, + "loss": 10.217635154724121, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.42084402218461037, + "loss_total_v6": 0.6386022120714188, + "step": 729 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16836036036036037, + "grad_norm": 26.130970001220703, + "learning_rate": 9.600278832501843e-06, + "loss": 10.520397186279297, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.43474654480814934, + "loss_total_v6": 0.6575248539447784, + "step": 730 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16859099099099098, + "grad_norm": 24.2880916595459, + "learning_rate": 9.598789550432065e-06, + "loss": 10.804357528686523, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.4560950696468353, + "loss_total_v6": 0.6752723306417465, + "step": 731 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16882162162162162, + "grad_norm": 26.37623405456543, + "learning_rate": 9.597297615087686e-06, + "loss": 11.248603820800781, + "loss_alignment": 0.443359375, + "loss_alignment_w": 0.2216796875, + "loss_sft": 0.4813123419880867, + "loss_total_v6": 0.70303775370121, + "step": 732 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16905225225225226, + "grad_norm": 26.36607551574707, + "learning_rate": 9.595803027329473e-06, + "loss": 10.987191200256348, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.4643178656697273, + "loss_total_v6": 0.6866994798183441, + "step": 733 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16928288288288287, + "grad_norm": 25.866092681884766, + "learning_rate": 9.594305788019736e-06, + "loss": 10.102237701416016, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.4216577894985676, + "loss_total_v6": 0.631389819085598, + "step": 734 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16951351351351351, + "grad_norm": 26.01093292236328, + "learning_rate": 9.5928058980223e-06, + "loss": 10.787914276123047, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.455326821655035, + "loss_total_v6": 0.674244686961174, + "step": 735 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16974414414414415, + "grad_norm": 24.497859954833984, + "learning_rate": 9.591303358202535e-06, + "loss": 10.765010833740234, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.45870184898376465, + "loss_total_v6": 0.672813206911087, + "step": 736 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.16997477477477477, + "grad_norm": 29.820446014404297, + "learning_rate": 9.589798169427327e-06, + "loss": 10.37649917602539, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4332754351198673, + "loss_total_v6": 0.6485311686992645, + "step": 737 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1702054054054054, + "grad_norm": 28.440214157104492, + "learning_rate": 9.588290332565098e-06, + "loss": 11.203716278076172, + "loss_alignment": 0.44384765625, + "loss_alignment_w": 0.221923828125, + "loss_sft": 0.4785221368074417, + "loss_total_v6": 0.7002323120832443, + "step": 738 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17043603603603605, + "grad_norm": 25.07602882385254, + "learning_rate": 9.586779848485797e-06, + "loss": 11.054758071899414, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.4718671962618828, + "loss_total_v6": 0.6909223720431328, + "step": 739 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17066666666666666, + "grad_norm": 25.98268699645996, + "learning_rate": 9.585266718060897e-06, + "loss": 10.612253189086914, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4477965272963047, + "loss_total_v6": 0.6632658615708351, + "step": 740 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1708972972972973, + "grad_norm": 24.814897537231445, + "learning_rate": 9.583750942163398e-06, + "loss": 10.714883804321289, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.44551341608166695, + "loss_total_v6": 0.6696802824735641, + "step": 741 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17112792792792791, + "grad_norm": 25.109107971191406, + "learning_rate": 9.58223252166783e-06, + "loss": 10.969030380249023, + "loss_alignment": 0.452392578125, + "loss_alignment_w": 0.2261962890625, + "loss_sft": 0.459398590028286, + "loss_total_v6": 0.6855643689632416, + "step": 742 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17135855855855855, + "grad_norm": 26.90677261352539, + "learning_rate": 9.580711457450246e-06, + "loss": 10.829643249511719, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4664034843444824, + "loss_total_v6": 0.6768527254462242, + "step": 743 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1715891891891892, + "grad_norm": 23.360872268676758, + "learning_rate": 9.579187750388227e-06, + "loss": 10.905098915100098, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4617200568318367, + "loss_total_v6": 0.6815686970949173, + "step": 744 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1718198198198198, + "grad_norm": 25.885221481323242, + "learning_rate": 9.577661401360873e-06, + "loss": 10.35033130645752, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.4395897872745991, + "loss_total_v6": 0.646895706653595, + "step": 745 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17205045045045045, + "grad_norm": 26.232881546020508, + "learning_rate": 9.576132411248815e-06, + "loss": 10.606977462768555, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.44525424018502235, + "loss_total_v6": 0.6629361510276794, + "step": 746 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1722810810810811, + "grad_norm": 27.059024810791016, + "learning_rate": 9.574600780934204e-06, + "loss": 10.361783981323242, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.42939556390047073, + "loss_total_v6": 0.647611528635025, + "step": 747 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1725117117117117, + "grad_norm": 24.927583694458008, + "learning_rate": 9.573066511300715e-06, + "loss": 10.91915512084961, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4674660749733448, + "loss_total_v6": 0.6824471652507782, + "step": 748 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17274234234234234, + "grad_norm": 23.778823852539062, + "learning_rate": 9.571529603233545e-06, + "loss": 10.958869934082031, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.46590467169880867, + "loss_total_v6": 0.6849293559789658, + "step": 749 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17297297297297298, + "grad_norm": 23.052209854125977, + "learning_rate": 9.569990057619414e-06, + "loss": 10.834781646728516, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.45618074014782906, + "loss_total_v6": 0.6771738156676292, + "step": 750 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1732036036036036, + "grad_norm": 24.01239013671875, + "learning_rate": 9.568447875346567e-06, + "loss": 10.312007904052734, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.4302518144249916, + "loss_total_v6": 0.6445004865527153, + "step": 751 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17343423423423424, + "grad_norm": 27.05109977722168, + "learning_rate": 9.566903057304764e-06, + "loss": 10.822549819946289, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.45331063494086266, + "loss_total_v6": 0.6764093786478043, + "step": 752 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17366486486486488, + "grad_norm": 25.0561580657959, + "learning_rate": 9.56535560438529e-06, + "loss": 10.489412307739258, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.4487706162035465, + "loss_total_v6": 0.6555882468819618, + "step": 753 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1738954954954955, + "grad_norm": 27.9637508392334, + "learning_rate": 9.563805517480946e-06, + "loss": 10.310807228088379, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.4269876591861248, + "loss_total_v6": 0.6444254294037819, + "step": 754 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17412612612612613, + "grad_norm": 25.324108123779297, + "learning_rate": 9.56225279748606e-06, + "loss": 11.395256996154785, + "loss_alignment": 0.4560546875, + "loss_alignment_w": 0.22802734375, + "loss_sft": 0.48428305238485336, + "loss_total_v6": 0.7122035697102547, + "step": 755 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17435675675675677, + "grad_norm": 25.71493911743164, + "learning_rate": 9.560697445296474e-06, + "loss": 10.438724517822266, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.4456331729888916, + "loss_total_v6": 0.6524202674627304, + "step": 756 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17458738738738738, + "grad_norm": 23.63957977294922, + "learning_rate": 9.559139461809547e-06, + "loss": 10.6611328125, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.4519043639302254, + "loss_total_v6": 0.6663208454847336, + "step": 757 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17481801801801802, + "grad_norm": 24.901594161987305, + "learning_rate": 9.557578847924161e-06, + "loss": 10.94060230255127, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.46850139275193214, + "loss_total_v6": 0.6837876290082932, + "step": 758 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17504864864864864, + "grad_norm": 24.52097511291504, + "learning_rate": 9.556015604540712e-06, + "loss": 10.708392143249512, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.455178439617157, + "loss_total_v6": 0.6692745238542557, + "step": 759 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17527927927927928, + "grad_norm": 29.468862533569336, + "learning_rate": 9.554449732561114e-06, + "loss": 10.578689575195312, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.44034286588430405, + "loss_total_v6": 0.6611680686473846, + "step": 760 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17550990990990992, + "grad_norm": 24.524438858032227, + "learning_rate": 9.5528812328888e-06, + "loss": 10.460918426513672, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4420916922390461, + "loss_total_v6": 0.6538074016571045, + "step": 761 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17574054054054053, + "grad_norm": 23.687728881835938, + "learning_rate": 9.551310106428713e-06, + "loss": 9.956302642822266, + "loss_alignment": 0.44970703125, + "loss_alignment_w": 0.224853515625, + "loss_sft": 0.3975679837167263, + "loss_total_v6": 0.6222689300775528, + "step": 762 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17597117117117117, + "grad_norm": 24.922868728637695, + "learning_rate": 9.549736354087317e-06, + "loss": 11.205801010131836, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.47865235432982445, + "loss_total_v6": 0.7003625556826591, + "step": 763 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1762018018018018, + "grad_norm": 21.492639541625977, + "learning_rate": 9.548159976772593e-06, + "loss": 10.685487747192383, + "loss_alignment": 0.446044921875, + "loss_alignment_w": 0.2230224609375, + "loss_sft": 0.4447442889213562, + "loss_total_v6": 0.6678430289030075, + "step": 764 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17643243243243242, + "grad_norm": 22.246944427490234, + "learning_rate": 9.546580975394029e-06, + "loss": 10.568286895751953, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4428360387682915, + "loss_total_v6": 0.6605179309844971, + "step": 765 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17666306306306306, + "grad_norm": 23.44942855834961, + "learning_rate": 9.54499935086263e-06, + "loss": 10.850445747375488, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.46935155987739563, + "loss_total_v6": 0.6781528368592262, + "step": 766 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1768936936936937, + "grad_norm": 24.664867401123047, + "learning_rate": 9.54341510409092e-06, + "loss": 11.089299201965332, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.47771867364645004, + "loss_total_v6": 0.6930812150239944, + "step": 767 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17712432432432432, + "grad_norm": 23.235628128051758, + "learning_rate": 9.541828235992927e-06, + "loss": 10.42074966430664, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.44244982302188873, + "loss_total_v6": 0.6512968391180038, + "step": 768 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17735495495495496, + "grad_norm": 27.457969665527344, + "learning_rate": 9.540238747484197e-06, + "loss": 10.406661987304688, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.44149303436279297, + "loss_total_v6": 0.6504163667559624, + "step": 769 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1775855855855856, + "grad_norm": 23.72147560119629, + "learning_rate": 9.538646639481785e-06, + "loss": 10.83967399597168, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4562119133770466, + "loss_total_v6": 0.6774796172976494, + "step": 770 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1778162162162162, + "grad_norm": 24.892969131469727, + "learning_rate": 9.537051912904263e-06, + "loss": 11.19206428527832, + "loss_alignment": 0.44970703125, + "loss_alignment_w": 0.224853515625, + "loss_sft": 0.4745894595980644, + "loss_total_v6": 0.6995040029287338, + "step": 771 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17804684684684685, + "grad_norm": 25.085248947143555, + "learning_rate": 9.535454568671705e-06, + "loss": 10.440889358520508, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.43581976369023323, + "loss_total_v6": 0.6525555849075317, + "step": 772 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17827747747747746, + "grad_norm": 23.790943145751953, + "learning_rate": 9.533854607705702e-06, + "loss": 10.518619537353516, + "loss_alignment": 0.408203125, + "loss_alignment_w": 0.2041015625, + "loss_sft": 0.4536631479859352, + "loss_total_v6": 0.6574137210845947, + "step": 773 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1785081081081081, + "grad_norm": 32.48297119140625, + "learning_rate": 9.532252030929351e-06, + "loss": 10.53277587890625, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4427833668887615, + "loss_total_v6": 0.6582985147833824, + "step": 774 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17873873873873874, + "grad_norm": 26.007226943969727, + "learning_rate": 9.530646839267262e-06, + "loss": 10.711807250976562, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4564446993172169, + "loss_total_v6": 0.6694879233837128, + "step": 775 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17896936936936936, + "grad_norm": 25.749147415161133, + "learning_rate": 9.529039033645548e-06, + "loss": 9.902859687805176, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4070756919682026, + "loss_total_v6": 0.6189287230372429, + "step": 776 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1792, + "grad_norm": 25.504711151123047, + "learning_rate": 9.527428614991835e-06, + "loss": 10.935628890991211, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4678090214729309, + "loss_total_v6": 0.6834767386317253, + "step": 777 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17943063063063064, + "grad_norm": 25.347780227661133, + "learning_rate": 9.525815584235258e-06, + "loss": 10.806153297424316, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.46751412376761436, + "loss_total_v6": 0.6753845885396004, + "step": 778 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.17966126126126125, + "grad_norm": 25.27945899963379, + "learning_rate": 9.524199942306454e-06, + "loss": 11.299967765808105, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.495768278837204, + "loss_total_v6": 0.7062480077147484, + "step": 779 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1798918918918919, + "grad_norm": 25.799257278442383, + "learning_rate": 9.522581690137567e-06, + "loss": 10.375856399536133, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4327927827835083, + "loss_total_v6": 0.6484910473227501, + "step": 780 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18012252252252253, + "grad_norm": 24.741037368774414, + "learning_rate": 9.52096082866225e-06, + "loss": 10.448657989501953, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.43586281314492226, + "loss_total_v6": 0.6530411317944527, + "step": 781 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18035315315315315, + "grad_norm": 24.040307998657227, + "learning_rate": 9.51933735881566e-06, + "loss": 10.786600112915039, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.45672475546598434, + "loss_total_v6": 0.674162469804287, + "step": 782 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18058378378378379, + "grad_norm": 24.08324432373047, + "learning_rate": 9.51771128153446e-06, + "loss": 10.796926498413086, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.45715654268860817, + "loss_total_v6": 0.6748079061508179, + "step": 783 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18081441441441443, + "grad_norm": 25.983169555664062, + "learning_rate": 9.516082597756817e-06, + "loss": 11.200939178466797, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.48115605488419533, + "loss_total_v6": 0.7000586614012718, + "step": 784 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18104504504504504, + "grad_norm": 27.75440788269043, + "learning_rate": 9.514451308422397e-06, + "loss": 11.37866497039795, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4941408261656761, + "loss_total_v6": 0.7111665457487106, + "step": 785 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18127567567567568, + "grad_norm": 24.807252883911133, + "learning_rate": 9.512817414472376e-06, + "loss": 10.680278778076172, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.44881317764520645, + "loss_total_v6": 0.6675173938274384, + "step": 786 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18150630630630632, + "grad_norm": 22.999853134155273, + "learning_rate": 9.511180916849431e-06, + "loss": 11.28093147277832, + "loss_alignment": 0.443359375, + "loss_alignment_w": 0.2216796875, + "loss_sft": 0.4839431121945381, + "loss_total_v6": 0.705058254301548, + "step": 787 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18173693693693693, + "grad_norm": 23.40178108215332, + "learning_rate": 9.509541816497738e-06, + "loss": 10.38988208770752, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.44013914093375206, + "loss_total_v6": 0.64936763048172, + "step": 788 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18196756756756757, + "grad_norm": 25.070791244506836, + "learning_rate": 9.50790011436298e-06, + "loss": 10.655574798583984, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.44652147963643074, + "loss_total_v6": 0.6659734174609184, + "step": 789 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18219819819819819, + "grad_norm": 22.798227310180664, + "learning_rate": 9.506255811392338e-06, + "loss": 10.444056510925293, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.4331795834004879, + "loss_total_v6": 0.6527535319328308, + "step": 790 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18242882882882883, + "grad_norm": 27.633832931518555, + "learning_rate": 9.504608908534489e-06, + "loss": 11.013772964477539, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.472342137247324, + "loss_total_v6": 0.6883608251810074, + "step": 791 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18265945945945947, + "grad_norm": 24.489280700683594, + "learning_rate": 9.50295940673962e-06, + "loss": 10.289623260498047, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.42351221293210983, + "loss_total_v6": 0.6431014388799667, + "step": 792 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18289009009009008, + "grad_norm": 26.319761276245117, + "learning_rate": 9.501307306959411e-06, + "loss": 10.723094940185547, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.453686460852623, + "loss_total_v6": 0.6701934188604355, + "step": 793 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18312072072072072, + "grad_norm": 26.22319221496582, + "learning_rate": 9.49965261014704e-06, + "loss": 11.009713172912598, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.46831948310136795, + "loss_total_v6": 0.6881070733070374, + "step": 794 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18335135135135136, + "grad_norm": 24.24945068359375, + "learning_rate": 9.497995317257187e-06, + "loss": 10.53466796875, + "loss_alignment": 0.443115234375, + "loss_alignment_w": 0.2215576171875, + "loss_sft": 0.4370422139763832, + "loss_total_v6": 0.6584167182445526, + "step": 795 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18358198198198197, + "grad_norm": 24.35004425048828, + "learning_rate": 9.496335429246026e-06, + "loss": 10.642669677734375, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.44914817810058594, + "loss_total_v6": 0.6651668548583984, + "step": 796 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1838126126126126, + "grad_norm": 24.177318572998047, + "learning_rate": 9.494672947071236e-06, + "loss": 10.059816360473633, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.4102631136775017, + "loss_total_v6": 0.6287384778261185, + "step": 797 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18404324324324325, + "grad_norm": 25.421951293945312, + "learning_rate": 9.49300787169198e-06, + "loss": 11.003578186035156, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4702400974929333, + "loss_total_v6": 0.6877236589789391, + "step": 798 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18427387387387387, + "grad_norm": 27.430265426635742, + "learning_rate": 9.491340204068929e-06, + "loss": 11.121980667114258, + "loss_alignment": 0.443603515625, + "loss_alignment_w": 0.2218017578125, + "loss_sft": 0.473535630851984, + "loss_total_v6": 0.6951237767934799, + "step": 799 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1845045045045045, + "grad_norm": 25.198354721069336, + "learning_rate": 9.489669945164242e-06, + "loss": 10.904255867004395, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.46421556919813156, + "loss_total_v6": 0.6815159916877747, + "step": 800 + }, + { + "epoch": 0.1845045045045045, + "eval_loss": 0.6950480937957764, + "eval_loss_alignment": 0.413841312767094, + "eval_loss_alignment_w": 0.206920656383547, + "eval_loss_sft": 0.4881274353903838, + "eval_loss_total_v6": 0.6950480917420907, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.1823, + "eval_samples_per_second": 22.985, + "eval_steps_per_second": 2.882, + "step": 800 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18473513513513515, + "grad_norm": 26.256553649902344, + "learning_rate": 9.487997095941579e-06, + "loss": 10.273143768310547, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4257171154022217, + "loss_total_v6": 0.6420715004205704, + "step": 801 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18496576576576576, + "grad_norm": 24.065034866333008, + "learning_rate": 9.48632165736609e-06, + "loss": 10.93728256225586, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.4617326110601425, + "loss_total_v6": 0.6835801303386688, + "step": 802 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1851963963963964, + "grad_norm": 24.997909545898438, + "learning_rate": 9.48464363040442e-06, + "loss": 10.426115036010742, + "loss_alignment": 0.447265625, + "loss_alignment_w": 0.2236328125, + "loss_sft": 0.4281976893544197, + "loss_total_v6": 0.6516321524977684, + "step": 803 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.185427027027027, + "grad_norm": 25.569921493530273, + "learning_rate": 9.482963016024709e-06, + "loss": 11.145281791687012, + "loss_alignment": 0.454833984375, + "loss_alignment_w": 0.2274169921875, + "loss_sft": 0.46882742643356323, + "loss_total_v6": 0.6965801045298576, + "step": 804 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18565765765765765, + "grad_norm": 23.871742248535156, + "learning_rate": 9.481279815196587e-06, + "loss": 10.894888877868652, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4633707292377949, + "loss_total_v6": 0.6809305474162102, + "step": 805 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1858882882882883, + "grad_norm": 24.14041519165039, + "learning_rate": 9.479594028891178e-06, + "loss": 10.7291898727417, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.4516564905643463, + "loss_total_v6": 0.670574352145195, + "step": 806 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1861189189189189, + "grad_norm": 26.497600555419922, + "learning_rate": 9.477905658081098e-06, + "loss": 10.889947891235352, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.4610630236566067, + "loss_total_v6": 0.6806217283010483, + "step": 807 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18634954954954955, + "grad_norm": 24.062768936157227, + "learning_rate": 9.476214703740454e-06, + "loss": 10.239547729492188, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4243650324642658, + "loss_total_v6": 0.6399717181921005, + "step": 808 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1865801801801802, + "grad_norm": 21.736181259155273, + "learning_rate": 9.474521166844843e-06, + "loss": 10.612358093261719, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.4456057772040367, + "loss_total_v6": 0.6632724106311798, + "step": 809 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1868108108108108, + "grad_norm": 27.133441925048828, + "learning_rate": 9.472825048371353e-06, + "loss": 10.507087707519531, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.4445042051374912, + "loss_total_v6": 0.6566929370164871, + "step": 810 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18704144144144144, + "grad_norm": 26.44197654724121, + "learning_rate": 9.471126349298557e-06, + "loss": 10.3043851852417, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4289056323468685, + "loss_total_v6": 0.6440240740776062, + "step": 811 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18727207207207208, + "grad_norm": 26.693756103515625, + "learning_rate": 9.469425070606525e-06, + "loss": 11.06719970703125, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.4677773006260395, + "loss_total_v6": 0.6917000114917755, + "step": 812 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1875027027027027, + "grad_norm": 23.434534072875977, + "learning_rate": 9.467721213276808e-06, + "loss": 10.977762222290039, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.46725334227085114, + "loss_total_v6": 0.6861101612448692, + "step": 813 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18773333333333334, + "grad_norm": 23.13188934326172, + "learning_rate": 9.46601477829245e-06, + "loss": 10.572742462158203, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.4415733553469181, + "loss_total_v6": 0.6607963964343071, + "step": 814 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18796396396396398, + "grad_norm": 27.503284454345703, + "learning_rate": 9.464305766637978e-06, + "loss": 10.618767738342285, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.4554668106138706, + "loss_total_v6": 0.6636729761958122, + "step": 815 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1881945945945946, + "grad_norm": 24.62271499633789, + "learning_rate": 9.462594179299408e-06, + "loss": 11.528722763061523, + "loss_alignment": 0.443115234375, + "loss_alignment_w": 0.2215576171875, + "loss_sft": 0.49878915026783943, + "loss_total_v6": 0.7205451503396034, + "step": 816 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18842522522522523, + "grad_norm": 23.979583740234375, + "learning_rate": 9.46088001726424e-06, + "loss": 10.814502716064453, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.4563172198832035, + "loss_total_v6": 0.6759064346551895, + "step": 817 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18865585585585587, + "grad_norm": 26.215394973754883, + "learning_rate": 9.459163281521462e-06, + "loss": 10.957914352416992, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4650667682290077, + "loss_total_v6": 0.6848696619272232, + "step": 818 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18888648648648648, + "grad_norm": 25.542644500732422, + "learning_rate": 9.457443973061546e-06, + "loss": 10.599197387695312, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.44479847326874733, + "loss_total_v6": 0.6624498143792152, + "step": 819 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18911711711711712, + "grad_norm": 22.954370498657227, + "learning_rate": 9.45572209287645e-06, + "loss": 10.81641960144043, + "loss_alignment": 0.4521484375, + "loss_alignment_w": 0.22607421875, + "loss_sft": 0.44952480122447014, + "loss_total_v6": 0.6760262474417686, + "step": 820 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18934774774774774, + "grad_norm": 24.621784210205078, + "learning_rate": 9.45399764195961e-06, + "loss": 10.730895042419434, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.4546164646744728, + "loss_total_v6": 0.6706809177994728, + "step": 821 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18957837837837838, + "grad_norm": 23.816967010498047, + "learning_rate": 9.452270621305952e-06, + "loss": 11.201845169067383, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4868432767689228, + "loss_total_v6": 0.7001153603196144, + "step": 822 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.18980900900900902, + "grad_norm": 23.502092361450195, + "learning_rate": 9.450541031911879e-06, + "loss": 10.886409759521484, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.4665181376039982, + "loss_total_v6": 0.6804005801677704, + "step": 823 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19003963963963963, + "grad_norm": 23.94746971130371, + "learning_rate": 9.448808874775282e-06, + "loss": 10.957717895507812, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.46572593227028847, + "loss_total_v6": 0.6848573982715607, + "step": 824 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19027027027027027, + "grad_norm": 23.34552764892578, + "learning_rate": 9.447074150895528e-06, + "loss": 11.300532341003418, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4902646131813526, + "loss_total_v6": 0.7062832862138748, + "step": 825 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1905009009009009, + "grad_norm": 25.147428512573242, + "learning_rate": 9.445336861273468e-06, + "loss": 11.147978782653809, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.48068422824144363, + "loss_total_v6": 0.6967486664652824, + "step": 826 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19073153153153152, + "grad_norm": 23.54167938232422, + "learning_rate": 9.443597006911432e-06, + "loss": 10.708419799804688, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.45160962641239166, + "loss_total_v6": 0.6692762523889542, + "step": 827 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19096216216216216, + "grad_norm": 22.533653259277344, + "learning_rate": 9.441854588813228e-06, + "loss": 10.495704650878906, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.43755199760198593, + "loss_total_v6": 0.6559815630316734, + "step": 828 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1911927927927928, + "grad_norm": 26.335954666137695, + "learning_rate": 9.440109607984148e-06, + "loss": 10.977560043334961, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.46611153706908226, + "loss_total_v6": 0.6860975176095963, + "step": 829 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19142342342342342, + "grad_norm": 24.532142639160156, + "learning_rate": 9.438362065430961e-06, + "loss": 10.470415115356445, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.4320041351020336, + "loss_total_v6": 0.6544009894132614, + "step": 830 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19165405405405406, + "grad_norm": 23.819923400878906, + "learning_rate": 9.43661196216191e-06, + "loss": 10.112564086914062, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4198617860674858, + "loss_total_v6": 0.6320352479815483, + "step": 831 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1918846846846847, + "grad_norm": 25.093347549438477, + "learning_rate": 9.434859299186719e-06, + "loss": 11.270482063293457, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.48962241783738136, + "loss_total_v6": 0.7044051364064217, + "step": 832 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1921153153153153, + "grad_norm": 22.93609046936035, + "learning_rate": 9.433104077516588e-06, + "loss": 10.780401229858398, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.45668825507164, + "loss_total_v6": 0.6737750917673111, + "step": 833 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19234594594594595, + "grad_norm": 26.557477951049805, + "learning_rate": 9.431346298164195e-06, + "loss": 11.323319435119629, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4871417097747326, + "loss_total_v6": 0.7077074944972992, + "step": 834 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19257657657657656, + "grad_norm": 25.528879165649414, + "learning_rate": 9.429585962143687e-06, + "loss": 11.068218231201172, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.47830840200185776, + "loss_total_v6": 0.6917636096477509, + "step": 835 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1928072072072072, + "grad_norm": 23.99338722229004, + "learning_rate": 9.4278230704707e-06, + "loss": 10.76942253112793, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.463143240660429, + "loss_total_v6": 0.6730889528989792, + "step": 836 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19303783783783784, + "grad_norm": 22.609642028808594, + "learning_rate": 9.426057624162324e-06, + "loss": 10.415117263793945, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4363909587264061, + "loss_total_v6": 0.650944784283638, + "step": 837 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19326846846846846, + "grad_norm": 25.370046615600586, + "learning_rate": 9.424289624237143e-06, + "loss": 10.54883861541748, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.44387881085276604, + "loss_total_v6": 0.6593023911118507, + "step": 838 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1934990990990991, + "grad_norm": 24.252439498901367, + "learning_rate": 9.422519071715204e-06, + "loss": 10.314006805419922, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.424547903239727, + "loss_total_v6": 0.6446254327893257, + "step": 839 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19372972972972974, + "grad_norm": 25.529272079467773, + "learning_rate": 9.420745967618026e-06, + "loss": 10.45068073272705, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.43405134975910187, + "loss_total_v6": 0.6531675457954407, + "step": 840 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19396036036036035, + "grad_norm": 25.57604217529297, + "learning_rate": 9.418970312968603e-06, + "loss": 10.647821426391602, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4523388333618641, + "loss_total_v6": 0.6654888540506363, + "step": 841 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.194190990990991, + "grad_norm": 24.00800132751465, + "learning_rate": 9.417192108791402e-06, + "loss": 10.185396194458008, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4257717877626419, + "loss_total_v6": 0.6365872547030449, + "step": 842 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19442162162162163, + "grad_norm": 28.709949493408203, + "learning_rate": 9.41541135611236e-06, + "loss": 10.474433898925781, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.445377916097641, + "loss_total_v6": 0.6546521782875061, + "step": 843 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19465225225225224, + "grad_norm": 23.1684513092041, + "learning_rate": 9.413628055958878e-06, + "loss": 10.421356201171875, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.430479034781456, + "loss_total_v6": 0.6513347327709198, + "step": 844 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19488288288288289, + "grad_norm": 22.773225784301758, + "learning_rate": 9.411842209359838e-06, + "loss": 10.812911987304688, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4632673151791096, + "loss_total_v6": 0.6758069843053818, + "step": 845 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19511351351351353, + "grad_norm": 22.802961349487305, + "learning_rate": 9.410053817345584e-06, + "loss": 10.473878860473633, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.44345107674598694, + "loss_total_v6": 0.6546174362301826, + "step": 846 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19534414414414414, + "grad_norm": 24.01570701599121, + "learning_rate": 9.408262880947927e-06, + "loss": 10.22673225402832, + "loss_alignment": 0.40869140625, + "loss_alignment_w": 0.204345703125, + "loss_sft": 0.43496234342455864, + "loss_total_v6": 0.6391707584261894, + "step": 847 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19557477477477478, + "grad_norm": 23.531352996826172, + "learning_rate": 9.406469401200151e-06, + "loss": 11.019872665405273, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.46919863671064377, + "loss_total_v6": 0.6887420564889908, + "step": 848 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1958054054054054, + "grad_norm": 23.449628829956055, + "learning_rate": 9.404673379137007e-06, + "loss": 10.4336519241333, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.4340704083442688, + "loss_total_v6": 0.6521032452583313, + "step": 849 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19603603603603603, + "grad_norm": 24.733898162841797, + "learning_rate": 9.402874815794708e-06, + "loss": 10.628534317016602, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4517436698079109, + "loss_total_v6": 0.664283350110054, + "step": 850 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19626666666666667, + "grad_norm": 22.699907302856445, + "learning_rate": 9.401073712210937e-06, + "loss": 10.291906356811523, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.4244178310036659, + "loss_total_v6": 0.643244095146656, + "step": 851 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19649729729729729, + "grad_norm": 23.729862213134766, + "learning_rate": 9.399270069424843e-06, + "loss": 10.408117294311523, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4383949115872383, + "loss_total_v6": 0.6505073234438896, + "step": 852 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19672792792792793, + "grad_norm": 21.612974166870117, + "learning_rate": 9.39746388847704e-06, + "loss": 10.95298957824707, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.46883315220475197, + "loss_total_v6": 0.6845619007945061, + "step": 853 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19695855855855857, + "grad_norm": 24.555965423583984, + "learning_rate": 9.3956551704096e-06, + "loss": 10.841779708862305, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4643391743302345, + "loss_total_v6": 0.6776112914085388, + "step": 854 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19718918918918918, + "grad_norm": 22.03118133544922, + "learning_rate": 9.39384391626607e-06, + "loss": 10.616863250732422, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.4440715052187443, + "loss_total_v6": 0.6635539308190346, + "step": 855 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19741981981981982, + "grad_norm": 26.15074348449707, + "learning_rate": 9.392030127091452e-06, + "loss": 10.720296859741211, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4561818353831768, + "loss_total_v6": 0.6700185015797615, + "step": 856 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19765045045045046, + "grad_norm": 24.26850128173828, + "learning_rate": 9.39021380393221e-06, + "loss": 10.440101623535156, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.43506862223148346, + "loss_total_v6": 0.6525063663721085, + "step": 857 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19788108108108107, + "grad_norm": 21.652576446533203, + "learning_rate": 9.388394947836278e-06, + "loss": 10.758559226989746, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.46533291041851044, + "loss_total_v6": 0.6724099516868591, + "step": 858 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1981117117117117, + "grad_norm": 23.433385848999023, + "learning_rate": 9.386573559853046e-06, + "loss": 10.775293350219727, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.4556366056203842, + "loss_total_v6": 0.6734558269381523, + "step": 859 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19834234234234235, + "grad_norm": 23.635404586791992, + "learning_rate": 9.384749641033358e-06, + "loss": 10.915145874023438, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.46350760385394096, + "loss_total_v6": 0.6821965649724007, + "step": 860 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19857297297297297, + "grad_norm": 25.02082061767578, + "learning_rate": 9.382923192429532e-06, + "loss": 10.302083969116211, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.4252675771713257, + "loss_total_v6": 0.643880195915699, + "step": 861 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1988036036036036, + "grad_norm": 25.525634765625, + "learning_rate": 9.381094215095336e-06, + "loss": 10.973679542541504, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.4673796035349369, + "loss_total_v6": 0.6858549863100052, + "step": 862 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19903423423423425, + "grad_norm": 26.591381072998047, + "learning_rate": 9.379262710086002e-06, + "loss": 10.299173355102539, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4252992980182171, + "loss_total_v6": 0.6436983495950699, + "step": 863 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19926486486486486, + "grad_norm": 22.81828498840332, + "learning_rate": 9.377428678458215e-06, + "loss": 11.014394760131836, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.46919194981455803, + "loss_total_v6": 0.6883997097611427, + "step": 864 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1994954954954955, + "grad_norm": 22.057064056396484, + "learning_rate": 9.375592121270122e-06, + "loss": 10.453254699707031, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.43495986610651016, + "loss_total_v6": 0.6533284112811089, + "step": 865 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.1997261261261261, + "grad_norm": 22.74774169921875, + "learning_rate": 9.373753039581326e-06, + "loss": 10.456436157226562, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.43599799647927284, + "loss_total_v6": 0.6535272896289825, + "step": 866 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.19995675675675675, + "grad_norm": 25.39815902709961, + "learning_rate": 9.371911434452885e-06, + "loss": 10.307745933532715, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.4357227794826031, + "loss_total_v6": 0.6442341059446335, + "step": 867 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2001873873873874, + "grad_norm": 21.22335433959961, + "learning_rate": 9.370067306947317e-06, + "loss": 10.768878936767578, + "loss_alignment": 0.453857421875, + "loss_alignment_w": 0.2269287109375, + "loss_sft": 0.4465077370405197, + "loss_total_v6": 0.673054963350296, + "step": 868 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.200418018018018, + "grad_norm": 24.253604888916016, + "learning_rate": 9.36822065812859e-06, + "loss": 10.697639465332031, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.45336194708943367, + "loss_total_v6": 0.6686024144291878, + "step": 869 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20064864864864865, + "grad_norm": 26.821813583374023, + "learning_rate": 9.36637148906213e-06, + "loss": 10.863160133361816, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.45803075283765793, + "loss_total_v6": 0.6789475008845329, + "step": 870 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2008792792792793, + "grad_norm": 21.343246459960938, + "learning_rate": 9.364519800814818e-06, + "loss": 10.232046127319336, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.4303353950381279, + "loss_total_v6": 0.6395028680562973, + "step": 871 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2011099099099099, + "grad_norm": 23.416126251220703, + "learning_rate": 9.362665594454985e-06, + "loss": 10.660758018493652, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4513315297663212, + "loss_total_v6": 0.6662973538041115, + "step": 872 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20134054054054054, + "grad_norm": 23.32686996459961, + "learning_rate": 9.360808871052415e-06, + "loss": 10.64870834350586, + "loss_alignment": 0.4521484375, + "loss_alignment_w": 0.22607421875, + "loss_sft": 0.4392563924193382, + "loss_total_v6": 0.6655442416667938, + "step": 873 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20157117117117118, + "grad_norm": 21.159643173217773, + "learning_rate": 9.35894963167835e-06, + "loss": 10.031982421875, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4061431996524334, + "loss_total_v6": 0.6269989088177681, + "step": 874 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2018018018018018, + "grad_norm": 24.14276695251465, + "learning_rate": 9.357087877405475e-06, + "loss": 10.41248607635498, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.439812358468771, + "loss_total_v6": 0.6507803946733475, + "step": 875 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20203243243243243, + "grad_norm": 22.306129455566406, + "learning_rate": 9.355223609307932e-06, + "loss": 11.11566162109375, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4758415147662163, + "loss_total_v6": 0.6947288513183594, + "step": 876 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20226306306306308, + "grad_norm": 23.87224769592285, + "learning_rate": 9.353356828461313e-06, + "loss": 10.395095825195312, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.43991561606526375, + "loss_total_v6": 0.6496934667229652, + "step": 877 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2024936936936937, + "grad_norm": 23.829301834106445, + "learning_rate": 9.351487535942658e-06, + "loss": 10.066583633422852, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4128529205918312, + "loss_total_v6": 0.6291615068912506, + "step": 878 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20272432432432433, + "grad_norm": 23.687564849853516, + "learning_rate": 9.349615732830455e-06, + "loss": 10.429400444030762, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4351932257413864, + "loss_total_v6": 0.6518375128507614, + "step": 879 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20295495495495494, + "grad_norm": 22.953670501708984, + "learning_rate": 9.347741420204643e-06, + "loss": 10.359125137329102, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.43139608576893806, + "loss_total_v6": 0.6474452987313271, + "step": 880 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20318558558558558, + "grad_norm": 24.993459701538086, + "learning_rate": 9.345864599146608e-06, + "loss": 10.702325820922852, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.45428046956658363, + "loss_total_v6": 0.6688953414559364, + "step": 881 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20341621621621622, + "grad_norm": 23.826629638671875, + "learning_rate": 9.343985270739184e-06, + "loss": 10.621990203857422, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4497630074620247, + "loss_total_v6": 0.6638743653893471, + "step": 882 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20364684684684684, + "grad_norm": 23.352867126464844, + "learning_rate": 9.342103436066647e-06, + "loss": 9.831937789916992, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.3965090326964855, + "loss_total_v6": 0.614496111869812, + "step": 883 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20387747747747748, + "grad_norm": 25.79502296447754, + "learning_rate": 9.340219096214728e-06, + "loss": 10.910013198852539, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4662843495607376, + "loss_total_v6": 0.6818758025765419, + "step": 884 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20410810810810812, + "grad_norm": 25.548097610473633, + "learning_rate": 9.338332252270594e-06, + "loss": 10.469614028930664, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4413687027990818, + "loss_total_v6": 0.6543508991599083, + "step": 885 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20433873873873873, + "grad_norm": 23.753765106201172, + "learning_rate": 9.336442905322862e-06, + "loss": 10.405472755432129, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.43224814534187317, + "loss_total_v6": 0.6503420248627663, + "step": 886 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20456936936936937, + "grad_norm": 22.98143196105957, + "learning_rate": 9.334551056461593e-06, + "loss": 10.511423110961914, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.44141827151179314, + "loss_total_v6": 0.6569639444351196, + "step": 887 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2048, + "grad_norm": 22.379905700683594, + "learning_rate": 9.33265670677829e-06, + "loss": 10.613679885864258, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4522648863494396, + "loss_total_v6": 0.6633549630641937, + "step": 888 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20503063063063062, + "grad_norm": 22.637033462524414, + "learning_rate": 9.330759857365899e-06, + "loss": 11.360834121704102, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.49417079612612724, + "loss_total_v6": 0.7100521549582481, + "step": 889 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20526126126126126, + "grad_norm": 22.548831939697266, + "learning_rate": 9.328860509318807e-06, + "loss": 10.950008392333984, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.46850938349962234, + "loss_total_v6": 0.6843754947185516, + "step": 890 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2054918918918919, + "grad_norm": 24.356769561767578, + "learning_rate": 9.326958663732848e-06, + "loss": 11.010415077209473, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.47872409224510193, + "loss_total_v6": 0.6881509497761726, + "step": 891 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20572252252252252, + "grad_norm": 23.2979679107666, + "learning_rate": 9.32505432170529e-06, + "loss": 9.965615272521973, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.4048943892121315, + "loss_total_v6": 0.6228509470820427, + "step": 892 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20595315315315316, + "grad_norm": 23.582700729370117, + "learning_rate": 9.323147484334843e-06, + "loss": 10.556755065917969, + "loss_alignment": 0.411376953125, + "loss_alignment_w": 0.2056884765625, + "loss_sft": 0.45369675755500793, + "loss_total_v6": 0.6597971990704536, + "step": 893 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2061837837837838, + "grad_norm": 30.583036422729492, + "learning_rate": 9.32123815272166e-06, + "loss": 10.620388984680176, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4430406652390957, + "loss_total_v6": 0.6637743040919304, + "step": 894 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2064144144144144, + "grad_norm": 23.074378967285156, + "learning_rate": 9.319326327967332e-06, + "loss": 10.161826133728027, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4186682179570198, + "loss_total_v6": 0.6351141110062599, + "step": 895 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20664504504504505, + "grad_norm": 23.70868682861328, + "learning_rate": 9.317412011174886e-06, + "loss": 10.942242622375488, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4695804789662361, + "loss_total_v6": 0.6838901564478874, + "step": 896 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20687567567567566, + "grad_norm": 24.218826293945312, + "learning_rate": 9.315495203448792e-06, + "loss": 10.979804039001465, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.47011226788163185, + "loss_total_v6": 0.6862377673387527, + "step": 897 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2071063063063063, + "grad_norm": 27.000829696655273, + "learning_rate": 9.313575905894947e-06, + "loss": 11.289941787719727, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4871002212166786, + "loss_total_v6": 0.7056213617324829, + "step": 898 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20733693693693694, + "grad_norm": 23.01691436767578, + "learning_rate": 9.311654119620697e-06, + "loss": 10.732444763183594, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.4493880271911621, + "loss_total_v6": 0.670777790248394, + "step": 899 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20756756756756756, + "grad_norm": 25.47703742980957, + "learning_rate": 9.309729845734813e-06, + "loss": 10.904678344726562, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4727563224732876, + "loss_total_v6": 0.6815423518419266, + "step": 900 + }, + { + "epoch": 0.20756756756756756, + "eval_loss": 0.6933605074882507, + "eval_loss_alignment": 0.41317983774038464, + "eval_loss_alignment_w": 0.20658991887019232, + "eval_loss_sft": 0.48677055947641784, + "eval_loss_total_v6": 0.6933604791664925, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.6952, + "eval_samples_per_second": 22.841, + "eval_steps_per_second": 2.864, + "step": 900 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2077981981981982, + "grad_norm": 23.58331298828125, + "learning_rate": 9.307803085347508e-06, + "loss": 10.52331256866455, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4359968267381191, + "loss_total_v6": 0.6577070280909538, + "step": 901 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20802882882882884, + "grad_norm": 21.96754264831543, + "learning_rate": 9.305873839570429e-06, + "loss": 10.527322769165039, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.44224417209625244, + "loss_total_v6": 0.6579576879739761, + "step": 902 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20825945945945945, + "grad_norm": 23.43715476989746, + "learning_rate": 9.303942109516655e-06, + "loss": 10.903125762939453, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4703400246798992, + "loss_total_v6": 0.6814453527331352, + "step": 903 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2084900900900901, + "grad_norm": 21.84600067138672, + "learning_rate": 9.302007896300697e-06, + "loss": 10.485212326049805, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4390476606786251, + "loss_total_v6": 0.655325748026371, + "step": 904 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20872072072072073, + "grad_norm": 24.81993293762207, + "learning_rate": 9.300071201038503e-06, + "loss": 10.84048843383789, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.4572851397097111, + "loss_total_v6": 0.6775305271148682, + "step": 905 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20895135135135134, + "grad_norm": 23.682310104370117, + "learning_rate": 9.29813202484745e-06, + "loss": 11.397357940673828, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.49427152052521706, + "loss_total_v6": 0.7123348638415337, + "step": 906 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20918198198198198, + "grad_norm": 23.88041877746582, + "learning_rate": 9.296190368846345e-06, + "loss": 10.936031341552734, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4750973992049694, + "loss_total_v6": 0.6835019513964653, + "step": 907 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20941261261261263, + "grad_norm": 25.415075302124023, + "learning_rate": 9.29424623415543e-06, + "loss": 11.053552627563477, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4744773916900158, + "loss_total_v6": 0.6908470019698143, + "step": 908 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20964324324324324, + "grad_norm": 22.880422592163086, + "learning_rate": 9.292299621896375e-06, + "loss": 10.956342697143555, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.46748630329966545, + "loss_total_v6": 0.6847714558243752, + "step": 909 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.20987387387387388, + "grad_norm": 26.550004959106445, + "learning_rate": 9.290350533192277e-06, + "loss": 10.509655952453613, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4434440955519676, + "loss_total_v6": 0.656853511929512, + "step": 910 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2101045045045045, + "grad_norm": 25.18743896484375, + "learning_rate": 9.288398969167665e-06, + "loss": 10.54200553894043, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.44227684661746025, + "loss_total_v6": 0.658875361084938, + "step": 911 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21033513513513513, + "grad_norm": 30.046995162963867, + "learning_rate": 9.286444930948496e-06, + "loss": 10.163105010986328, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4185802601277828, + "loss_total_v6": 0.6351940333843231, + "step": 912 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21056576576576577, + "grad_norm": 24.41754913330078, + "learning_rate": 9.284488419662151e-06, + "loss": 10.59661865234375, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4497642107307911, + "loss_total_v6": 0.6622886210680008, + "step": 913 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21079639639639638, + "grad_norm": 22.36368751525879, + "learning_rate": 9.282529436437445e-06, + "loss": 10.308480262756348, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.42728474363684654, + "loss_total_v6": 0.6442800015211105, + "step": 914 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21102702702702703, + "grad_norm": 24.468935012817383, + "learning_rate": 9.280567982404611e-06, + "loss": 10.58640193939209, + "loss_alignment": 0.40966796875, + "loss_alignment_w": 0.204833984375, + "loss_sft": 0.4569992423057556, + "loss_total_v6": 0.6616501286625862, + "step": 915 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21125765765765767, + "grad_norm": 23.76603126525879, + "learning_rate": 9.278604058695313e-06, + "loss": 10.247971534729004, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.43061360716819763, + "loss_total_v6": 0.6404982209205627, + "step": 916 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21148828828828828, + "grad_norm": 23.455533981323242, + "learning_rate": 9.276637666442636e-06, + "loss": 10.818172454833984, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.4503514766693115, + "loss_total_v6": 0.676135778427124, + "step": 917 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21171891891891892, + "grad_norm": 22.743593215942383, + "learning_rate": 9.274668806781093e-06, + "loss": 10.367208480834961, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.43287789076566696, + "loss_total_v6": 0.6479505002498627, + "step": 918 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21194954954954956, + "grad_norm": 24.8084659576416, + "learning_rate": 9.272697480846618e-06, + "loss": 10.762869834899902, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4618639722466469, + "loss_total_v6": 0.6726793721318245, + "step": 919 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21218018018018017, + "grad_norm": 23.70018196105957, + "learning_rate": 9.270723689776568e-06, + "loss": 10.215082168579102, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.42935147508978844, + "loss_total_v6": 0.6384426653385162, + "step": 920 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2124108108108108, + "grad_norm": 22.59269142150879, + "learning_rate": 9.268747434709726e-06, + "loss": 10.48033618927002, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.440467182546854, + "loss_total_v6": 0.6550210192799568, + "step": 921 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21264144144144145, + "grad_norm": 24.60857582092285, + "learning_rate": 9.26676871678629e-06, + "loss": 10.457891464233398, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.4357379712164402, + "loss_total_v6": 0.653618223965168, + "step": 922 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21287207207207207, + "grad_norm": 22.960128784179688, + "learning_rate": 9.264787537147886e-06, + "loss": 10.564494132995605, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.44145458936691284, + "loss_total_v6": 0.6602808684110641, + "step": 923 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2131027027027027, + "grad_norm": 24.221073150634766, + "learning_rate": 9.262803896937556e-06, + "loss": 10.9864501953125, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.47048185020685196, + "loss_total_v6": 0.6866531223058701, + "step": 924 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21333333333333335, + "grad_norm": 24.06954002380371, + "learning_rate": 9.26081779729976e-06, + "loss": 10.93710708618164, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.47188401222229004, + "loss_total_v6": 0.6835691928863525, + "step": 925 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21356396396396396, + "grad_norm": 23.858600616455078, + "learning_rate": 9.25882923938038e-06, + "loss": 10.580848693847656, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4451623372733593, + "loss_total_v6": 0.6613030880689621, + "step": 926 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2137945945945946, + "grad_norm": 24.343996047973633, + "learning_rate": 9.256838224326722e-06, + "loss": 10.74771499633789, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.45250920206308365, + "loss_total_v6": 0.6717322319746017, + "step": 927 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2140252252252252, + "grad_norm": 26.943992614746094, + "learning_rate": 9.254844753287493e-06, + "loss": 10.79844856262207, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4624091126024723, + "loss_total_v6": 0.6749029755592346, + "step": 928 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21425585585585585, + "grad_norm": 23.583951950073242, + "learning_rate": 9.252848827412835e-06, + "loss": 10.526988983154297, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.44286416843533516, + "loss_total_v6": 0.6579368188977242, + "step": 929 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2144864864864865, + "grad_norm": 23.289579391479492, + "learning_rate": 9.250850447854296e-06, + "loss": 10.393301010131836, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.43749940767884254, + "loss_total_v6": 0.6495813280344009, + "step": 930 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2147171171171171, + "grad_norm": 24.945632934570312, + "learning_rate": 9.248849615764846e-06, + "loss": 10.731608390808105, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.4562327116727829, + "loss_total_v6": 0.6707255244255066, + "step": 931 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21494774774774775, + "grad_norm": 25.679418563842773, + "learning_rate": 9.24684633229886e-06, + "loss": 10.965341567993164, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.46925416216254234, + "loss_total_v6": 0.6853338554501534, + "step": 932 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2151783783783784, + "grad_norm": 24.348522186279297, + "learning_rate": 9.244840598612137e-06, + "loss": 10.417391777038574, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.4414464756846428, + "loss_total_v6": 0.6510869860649109, + "step": 933 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.215409009009009, + "grad_norm": 24.57731056213379, + "learning_rate": 9.242832415861888e-06, + "loss": 10.59756088256836, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4497315473854542, + "loss_total_v6": 0.6623475402593613, + "step": 934 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21563963963963964, + "grad_norm": 23.643571853637695, + "learning_rate": 9.240821785206732e-06, + "loss": 10.514886856079102, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.44786033406853676, + "loss_total_v6": 0.6571803838014603, + "step": 935 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21587027027027028, + "grad_norm": 22.966251373291016, + "learning_rate": 9.238808707806706e-06, + "loss": 10.28638744354248, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.4270178787410259, + "loss_total_v6": 0.6428992301225662, + "step": 936 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2161009009009009, + "grad_norm": 25.501428604125977, + "learning_rate": 9.236793184823257e-06, + "loss": 10.67365837097168, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4563187472522259, + "loss_total_v6": 0.6671036705374718, + "step": 937 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21633153153153153, + "grad_norm": 23.57956314086914, + "learning_rate": 9.23477521741924e-06, + "loss": 10.806264877319336, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4587472975254059, + "loss_total_v6": 0.6753915771842003, + "step": 938 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21656216216216218, + "grad_norm": 22.38211441040039, + "learning_rate": 9.232754806758923e-06, + "loss": 10.58574104309082, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.44383537769317627, + "loss_total_v6": 0.6616088300943375, + "step": 939 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2167927927927928, + "grad_norm": 23.685754776000977, + "learning_rate": 9.230731954007983e-06, + "loss": 10.503877639770508, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.44231995567679405, + "loss_total_v6": 0.6564923226833344, + "step": 940 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21702342342342343, + "grad_norm": 23.183679580688477, + "learning_rate": 9.228706660333508e-06, + "loss": 11.09539794921875, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.48069382458925247, + "loss_total_v6": 0.6934624016284943, + "step": 941 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21725405405405404, + "grad_norm": 23.050039291381836, + "learning_rate": 9.226678926903991e-06, + "loss": 10.415042877197266, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4344026781618595, + "loss_total_v6": 0.6509401723742485, + "step": 942 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21748468468468468, + "grad_norm": 21.42415428161621, + "learning_rate": 9.224648754889337e-06, + "loss": 10.649019241333008, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.4415951631963253, + "loss_total_v6": 0.66556366533041, + "step": 943 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21771531531531532, + "grad_norm": 21.722015380859375, + "learning_rate": 9.22261614546085e-06, + "loss": 10.23332405090332, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4304457940161228, + "loss_total_v6": 0.6395827308297157, + "step": 944 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21794594594594593, + "grad_norm": 28.68640899658203, + "learning_rate": 9.220581099791248e-06, + "loss": 10.801280975341797, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4664314202964306, + "loss_total_v6": 0.6750800982117653, + "step": 945 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21817657657657658, + "grad_norm": 28.88886070251465, + "learning_rate": 9.218543619054652e-06, + "loss": 10.74516487121582, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.45585931837558746, + "loss_total_v6": 0.67157281935215, + "step": 946 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21840720720720722, + "grad_norm": 25.00819969177246, + "learning_rate": 9.216503704426587e-06, + "loss": 10.511470794677734, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4409329779446125, + "loss_total_v6": 0.656966932117939, + "step": 947 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21863783783783783, + "grad_norm": 23.467281341552734, + "learning_rate": 9.214461357083986e-06, + "loss": 10.864273071289062, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.464066531509161, + "loss_total_v6": 0.6790170967578888, + "step": 948 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21886846846846847, + "grad_norm": 24.109453201293945, + "learning_rate": 9.212416578205178e-06, + "loss": 10.631991386413574, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4556523859500885, + "loss_total_v6": 0.6644994467496872, + "step": 949 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2190990990990991, + "grad_norm": 23.760244369506836, + "learning_rate": 9.210369368969904e-06, + "loss": 10.380636215209961, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.43425120040774345, + "loss_total_v6": 0.6487897783517838, + "step": 950 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21932972972972972, + "grad_norm": 22.804725646972656, + "learning_rate": 9.2083197305593e-06, + "loss": 11.063642501831055, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4752911552786827, + "loss_total_v6": 0.6914776712656021, + "step": 951 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.21956036036036036, + "grad_norm": 24.89687156677246, + "learning_rate": 9.206267664155906e-06, + "loss": 10.361774444580078, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4358493834733963, + "loss_total_v6": 0.6476108580827713, + "step": 952 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.219790990990991, + "grad_norm": 20.887359619140625, + "learning_rate": 9.204213170943666e-06, + "loss": 10.298994064331055, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.43152889609336853, + "loss_total_v6": 0.6436871141195297, + "step": 953 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22002162162162162, + "grad_norm": 23.644859313964844, + "learning_rate": 9.202156252107916e-06, + "loss": 10.814452171325684, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4538573697209358, + "loss_total_v6": 0.675903283059597, + "step": 954 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22025225225225226, + "grad_norm": 24.135305404663086, + "learning_rate": 9.200096908835401e-06, + "loss": 10.376733779907227, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.43831023946404457, + "loss_total_v6": 0.6485458239912987, + "step": 955 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22048288288288287, + "grad_norm": 24.48111343383789, + "learning_rate": 9.198035142314259e-06, + "loss": 10.766793251037598, + "loss_alignment": 0.4404296875, + "loss_alignment_w": 0.22021484375, + "loss_sft": 0.45241980999708176, + "loss_total_v6": 0.6729245707392693, + "step": 956 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2207135135135135, + "grad_norm": 24.847070693969727, + "learning_rate": 9.195970953734027e-06, + "loss": 10.765689849853516, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.45807287469506264, + "loss_total_v6": 0.6728555932641029, + "step": 957 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22094414414414415, + "grad_norm": 23.821533203125, + "learning_rate": 9.193904344285643e-06, + "loss": 11.403299331665039, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4957720376551151, + "loss_total_v6": 0.7127062380313873, + "step": 958 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22117477477477476, + "grad_norm": 22.766963958740234, + "learning_rate": 9.191835315161432e-06, + "loss": 10.327230453491211, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.428426131606102, + "loss_total_v6": 0.6454519182443619, + "step": 959 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2214054054054054, + "grad_norm": 24.988853454589844, + "learning_rate": 9.18976386755513e-06, + "loss": 10.521441459655762, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4365665875375271, + "loss_total_v6": 0.6575901135802269, + "step": 960 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22163603603603604, + "grad_norm": 30.22198486328125, + "learning_rate": 9.187690002661853e-06, + "loss": 10.648406982421875, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.44671444222331047, + "loss_total_v6": 0.6655254438519478, + "step": 961 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22186666666666666, + "grad_norm": 23.492496490478516, + "learning_rate": 9.185613721678123e-06, + "loss": 11.04362678527832, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4725753180682659, + "loss_total_v6": 0.6902266666293144, + "step": 962 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2220972972972973, + "grad_norm": 22.103092193603516, + "learning_rate": 9.18353502580185e-06, + "loss": 10.335981369018555, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.42817961424589157, + "loss_total_v6": 0.6459988132119179, + "step": 963 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22232792792792794, + "grad_norm": 24.59446907043457, + "learning_rate": 9.181453916232339e-06, + "loss": 10.281876564025879, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.43921760469675064, + "loss_total_v6": 0.6426172703504562, + "step": 964 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22255855855855855, + "grad_norm": 22.581336975097656, + "learning_rate": 9.179370394170289e-06, + "loss": 10.210841178894043, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4262787774205208, + "loss_total_v6": 0.6381775736808777, + "step": 965 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2227891891891892, + "grad_norm": 22.58670425415039, + "learning_rate": 9.177284460817788e-06, + "loss": 10.27652645111084, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4225258119404316, + "loss_total_v6": 0.6422829031944275, + "step": 966 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22301981981981983, + "grad_norm": 22.967370986938477, + "learning_rate": 9.175196117378316e-06, + "loss": 10.860475540161133, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.459114171564579, + "loss_total_v6": 0.6787797063589096, + "step": 967 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22325045045045044, + "grad_norm": 22.89080810546875, + "learning_rate": 9.173105365056742e-06, + "loss": 10.8424072265625, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.46089938282966614, + "loss_total_v6": 0.6776504516601562, + "step": 968 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22348108108108108, + "grad_norm": 24.74887466430664, + "learning_rate": 9.171012205059333e-06, + "loss": 10.592647552490234, + "loss_alignment": 0.447509765625, + "loss_alignment_w": 0.2237548828125, + "loss_sft": 0.43866709247231483, + "loss_total_v6": 0.662040501832962, + "step": 969 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22371171171171172, + "grad_norm": 23.51821517944336, + "learning_rate": 9.168916638593736e-06, + "loss": 10.902400970458984, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.46910450980067253, + "loss_total_v6": 0.6814000234007835, + "step": 970 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22394234234234234, + "grad_norm": 25.266456604003906, + "learning_rate": 9.166818666868988e-06, + "loss": 10.520456314086914, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.44056377187371254, + "loss_total_v6": 0.6575285047292709, + "step": 971 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22417297297297298, + "grad_norm": 23.624969482421875, + "learning_rate": 9.164718291095515e-06, + "loss": 10.99917984008789, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.47587043419480324, + "loss_total_v6": 0.6874487847089767, + "step": 972 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2244036036036036, + "grad_norm": 23.241405487060547, + "learning_rate": 9.162615512485135e-06, + "loss": 10.521825790405273, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.4389099031686783, + "loss_total_v6": 0.6576141119003296, + "step": 973 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22463423423423423, + "grad_norm": 23.450483322143555, + "learning_rate": 9.16051033225104e-06, + "loss": 10.658418655395508, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4524366036057472, + "loss_total_v6": 0.6661511808633804, + "step": 974 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22486486486486487, + "grad_norm": 23.85755157470703, + "learning_rate": 9.158402751607822e-06, + "loss": 10.592117309570312, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.442830141633749, + "loss_total_v6": 0.6620073541998863, + "step": 975 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22509549549549548, + "grad_norm": 23.846342086791992, + "learning_rate": 9.156292771771446e-06, + "loss": 10.880182266235352, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.46554915606975555, + "loss_total_v6": 0.680011436343193, + "step": 976 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22532612612612613, + "grad_norm": 22.189485549926758, + "learning_rate": 9.15418039395927e-06, + "loss": 10.719869613647461, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.4468931034207344, + "loss_total_v6": 0.6699918583035469, + "step": 977 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22555675675675677, + "grad_norm": 23.073415756225586, + "learning_rate": 9.15206561939003e-06, + "loss": 10.519562721252441, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.4335041791200638, + "loss_total_v6": 0.6574726700782776, + "step": 978 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22578738738738738, + "grad_norm": 21.99590492248535, + "learning_rate": 9.149948449283846e-06, + "loss": 10.335408210754395, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.43455250933766365, + "loss_total_v6": 0.6459630355238914, + "step": 979 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22601801801801802, + "grad_norm": 23.36374282836914, + "learning_rate": 9.14782888486222e-06, + "loss": 11.023114204406738, + "loss_alignment": 0.44970703125, + "loss_alignment_w": 0.224853515625, + "loss_sft": 0.46467097476124763, + "loss_total_v6": 0.6889446526765823, + "step": 980 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22624864864864866, + "grad_norm": 33.11121368408203, + "learning_rate": 9.14570692734804e-06, + "loss": 11.065152168273926, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4799478240311146, + "loss_total_v6": 0.691571980714798, + "step": 981 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22647927927927927, + "grad_norm": 22.110858917236328, + "learning_rate": 9.143582577965567e-06, + "loss": 10.923511505126953, + "loss_alignment": 0.40869140625, + "loss_alignment_w": 0.204345703125, + "loss_sft": 0.4783432297408581, + "loss_total_v6": 0.6827194541692734, + "step": 982 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2267099099099099, + "grad_norm": 21.325925827026367, + "learning_rate": 9.141455837940446e-06, + "loss": 10.458945274353027, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4384588822722435, + "loss_total_v6": 0.6536840945482254, + "step": 983 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22694054054054055, + "grad_norm": 21.069049835205078, + "learning_rate": 9.139326708499701e-06, + "loss": 10.699798583984375, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4559994041919708, + "loss_total_v6": 0.6687374487519264, + "step": 984 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22717117117117117, + "grad_norm": 20.718616485595703, + "learning_rate": 9.137195190871737e-06, + "loss": 9.856213569641113, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4006507582962513, + "loss_total_v6": 0.6160133183002472, + "step": 985 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2274018018018018, + "grad_norm": 22.75216293334961, + "learning_rate": 9.135061286286333e-06, + "loss": 10.48751449584961, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4423195868730545, + "loss_total_v6": 0.6554696261882782, + "step": 986 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22763243243243242, + "grad_norm": 22.058971405029297, + "learning_rate": 9.132924995974646e-06, + "loss": 10.261417388916016, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4258692339062691, + "loss_total_v6": 0.6413385719060898, + "step": 987 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22786306306306306, + "grad_norm": 20.272184371948242, + "learning_rate": 9.130786321169209e-06, + "loss": 10.974552154541016, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.46659497916698456, + "loss_total_v6": 0.6859095245599747, + "step": 988 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2280936936936937, + "grad_norm": 24.306947708129883, + "learning_rate": 9.128645263103937e-06, + "loss": 10.509735107421875, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4483470544219017, + "loss_total_v6": 0.6568584069609642, + "step": 989 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2283243243243243, + "grad_norm": 24.481609344482422, + "learning_rate": 9.12650182301411e-06, + "loss": 10.512491226196289, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.44391122460365295, + "loss_total_v6": 0.6570307239890099, + "step": 990 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22855495495495495, + "grad_norm": 21.950037002563477, + "learning_rate": 9.124356002136386e-06, + "loss": 11.063562393188477, + "loss_alignment": 0.447021484375, + "loss_alignment_w": 0.2235107421875, + "loss_sft": 0.467412568628788, + "loss_total_v6": 0.691472627222538, + "step": 991 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2287855855855856, + "grad_norm": 24.142425537109375, + "learning_rate": 9.122207801708802e-06, + "loss": 10.74698257446289, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.45548461005091667, + "loss_total_v6": 0.6716863811016083, + "step": 992 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2290162162162162, + "grad_norm": 23.988933563232422, + "learning_rate": 9.120057222970763e-06, + "loss": 10.459854125976562, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.43982793763279915, + "loss_total_v6": 0.653740867972374, + "step": 993 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22924684684684685, + "grad_norm": 21.64521026611328, + "learning_rate": 9.117904267163044e-06, + "loss": 10.029756546020508, + "loss_alignment": 0.452392578125, + "loss_alignment_w": 0.2261962890625, + "loss_sft": 0.40072452649474144, + "loss_total_v6": 0.6268598139286041, + "step": 994 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2294774774774775, + "grad_norm": 22.631820678710938, + "learning_rate": 9.115748935527795e-06, + "loss": 10.821905136108398, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4607623554766178, + "loss_total_v6": 0.6763690635561943, + "step": 995 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2297081081081081, + "grad_norm": 28.48998260498047, + "learning_rate": 9.113591229308538e-06, + "loss": 11.059707641601562, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.47419069334864616, + "loss_total_v6": 0.6912316828966141, + "step": 996 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.22993873873873874, + "grad_norm": 23.161279678344727, + "learning_rate": 9.111431149750164e-06, + "loss": 11.006186485290527, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.46625276282429695, + "loss_total_v6": 0.6878866702318192, + "step": 997 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23016936936936938, + "grad_norm": 23.167205810546875, + "learning_rate": 9.109268698098927e-06, + "loss": 10.554201126098633, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.43817152082920074, + "loss_total_v6": 0.6596375703811646, + "step": 998 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2304, + "grad_norm": 25.083526611328125, + "learning_rate": 9.107103875602458e-06, + "loss": 10.62297248840332, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4479934237897396, + "loss_total_v6": 0.6639358177781105, + "step": 999 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23063063063063063, + "grad_norm": 23.091426849365234, + "learning_rate": 9.104936683509755e-06, + "loss": 10.775049209594727, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.4598022326827049, + "loss_total_v6": 0.673440545797348, + "step": 1000 + }, + { + "epoch": 0.23063063063063063, + "eval_loss": 0.6910439729690552, + "eval_loss_alignment": 0.41117663261217946, + "eval_loss_alignment_w": 0.20558831630608973, + "eval_loss_sft": 0.4854556365241098, + "eval_loss_total_v6": 0.6910439527824394, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 84.8494, + "eval_samples_per_second": 21.992, + "eval_steps_per_second": 2.758, + "step": 1000 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23086126126126127, + "grad_norm": 22.184429168701172, + "learning_rate": 9.102767123071177e-06, + "loss": 10.438231468200684, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.43705740571022034, + "loss_total_v6": 0.6523894295096397, + "step": 1001 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2310918918918919, + "grad_norm": 21.254196166992188, + "learning_rate": 9.100595195538456e-06, + "loss": 10.263533592224121, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.43140309303998947, + "loss_total_v6": 0.6414708569645882, + "step": 1002 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23132252252252253, + "grad_norm": 23.014856338500977, + "learning_rate": 9.098420902164684e-06, + "loss": 10.198352813720703, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.42056961730122566, + "loss_total_v6": 0.637397013604641, + "step": 1003 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23155315315315314, + "grad_norm": 21.38472557067871, + "learning_rate": 9.096244244204325e-06, + "loss": 10.215829849243164, + "loss_alignment": 0.443115234375, + "loss_alignment_w": 0.2215576171875, + "loss_sft": 0.41728271916508675, + "loss_total_v6": 0.6384893879294395, + "step": 1004 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23178378378378378, + "grad_norm": 22.824708938598633, + "learning_rate": 9.094065222913199e-06, + "loss": 10.496469497680664, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.4484792947769165, + "loss_total_v6": 0.6560293659567833, + "step": 1005 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23201441441441442, + "grad_norm": 24.788311004638672, + "learning_rate": 9.091883839548497e-06, + "loss": 10.03492259979248, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4140936844050884, + "loss_total_v6": 0.6271826922893524, + "step": 1006 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23224504504504503, + "grad_norm": 28.1843318939209, + "learning_rate": 9.089700095368768e-06, + "loss": 10.231616973876953, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.42148901894688606, + "loss_total_v6": 0.639476090669632, + "step": 1007 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23247567567567567, + "grad_norm": 21.322147369384766, + "learning_rate": 9.087513991633926e-06, + "loss": 10.144355773925781, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.41437190026044846, + "loss_total_v6": 0.6340221911668777, + "step": 1008 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23270630630630632, + "grad_norm": 23.757610321044922, + "learning_rate": 9.085325529605242e-06, + "loss": 10.849577903747559, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.45849408581852913, + "loss_total_v6": 0.6780985891819, + "step": 1009 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23293693693693693, + "grad_norm": 22.38945960998535, + "learning_rate": 9.083134710545354e-06, + "loss": 10.554515838623047, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.44118187949061394, + "loss_total_v6": 0.6596572399139404, + "step": 1010 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23316756756756757, + "grad_norm": 23.404468536376953, + "learning_rate": 9.080941535718254e-06, + "loss": 10.256508827209473, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.42840058356523514, + "loss_total_v6": 0.6410318166017532, + "step": 1011 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2333981981981982, + "grad_norm": 24.837818145751953, + "learning_rate": 9.078746006389299e-06, + "loss": 10.531338691711426, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.435170978307724, + "loss_total_v6": 0.6582086831331253, + "step": 1012 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23362882882882882, + "grad_norm": 23.52907943725586, + "learning_rate": 9.076548123825195e-06, + "loss": 10.62598991394043, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.44821251556277275, + "loss_total_v6": 0.6641243621706963, + "step": 1013 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23385945945945946, + "grad_norm": 23.862882614135742, + "learning_rate": 9.074347889294017e-06, + "loss": 10.641105651855469, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.45310933142900467, + "loss_total_v6": 0.6650691330432892, + "step": 1014 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2340900900900901, + "grad_norm": 22.423738479614258, + "learning_rate": 9.072145304065193e-06, + "loss": 10.127235412597656, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4180169217288494, + "loss_total_v6": 0.6329522132873535, + "step": 1015 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23432072072072072, + "grad_norm": 22.006505966186523, + "learning_rate": 9.0699403694095e-06, + "loss": 10.311748504638672, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.43052560091018677, + "loss_total_v6": 0.6444843336939812, + "step": 1016 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23455135135135136, + "grad_norm": 22.191171646118164, + "learning_rate": 9.06773308659908e-06, + "loss": 10.842758178710938, + "loss_alignment": 0.44873046875, + "loss_alignment_w": 0.224365234375, + "loss_sft": 0.4537343792617321, + "loss_total_v6": 0.6776724010705948, + "step": 1017 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23478198198198197, + "grad_norm": 22.445127487182617, + "learning_rate": 9.065523456907425e-06, + "loss": 10.664353370666504, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4531431496143341, + "loss_total_v6": 0.6665220782160759, + "step": 1018 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2350126126126126, + "grad_norm": 21.820781707763672, + "learning_rate": 9.063311481609387e-06, + "loss": 10.556838035583496, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4441957175731659, + "loss_total_v6": 0.6598024070262909, + "step": 1019 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23524324324324325, + "grad_norm": 21.7253475189209, + "learning_rate": 9.061097161981159e-06, + "loss": 10.265355110168457, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.432478204369545, + "loss_total_v6": 0.6415846794843674, + "step": 1020 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23547387387387386, + "grad_norm": 24.264493942260742, + "learning_rate": 9.0588804993003e-06, + "loss": 10.809661865234375, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.45906636118888855, + "loss_total_v6": 0.6756038293242455, + "step": 1021 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2357045045045045, + "grad_norm": 22.320016860961914, + "learning_rate": 9.05666149484571e-06, + "loss": 10.179882049560547, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4252135567367077, + "loss_total_v6": 0.6362426057457924, + "step": 1022 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23593513513513514, + "grad_norm": 22.505680084228516, + "learning_rate": 9.054440149897652e-06, + "loss": 10.520307540893555, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.44238557294011116, + "loss_total_v6": 0.6575192287564278, + "step": 1023 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23616576576576576, + "grad_norm": 22.469131469726562, + "learning_rate": 9.052216465737727e-06, + "loss": 10.557372093200684, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.43952931836247444, + "loss_total_v6": 0.6598357446491718, + "step": 1024 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2363963963963964, + "grad_norm": 26.154895782470703, + "learning_rate": 9.04999044364889e-06, + "loss": 10.999631881713867, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4693067967891693, + "loss_total_v6": 0.6874769628047943, + "step": 1025 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23662702702702704, + "grad_norm": 21.931119918823242, + "learning_rate": 9.04776208491545e-06, + "loss": 10.422834396362305, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.43006789684295654, + "loss_total_v6": 0.6514271646738052, + "step": 1026 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23685765765765765, + "grad_norm": 20.791913986206055, + "learning_rate": 9.045531390823058e-06, + "loss": 10.619194030761719, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4501376375555992, + "loss_total_v6": 0.6636996120214462, + "step": 1027 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2370882882882883, + "grad_norm": 23.28887939453125, + "learning_rate": 9.043298362658715e-06, + "loss": 10.474437713623047, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4387557618319988, + "loss_total_v6": 0.6546523794531822, + "step": 1028 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23731891891891893, + "grad_norm": 23.959583282470703, + "learning_rate": 9.041063001710767e-06, + "loss": 9.925415992736816, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.40978245809674263, + "loss_total_v6": 0.6203385069966316, + "step": 1029 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23754954954954954, + "grad_norm": 23.2241268157959, + "learning_rate": 9.038825309268908e-06, + "loss": 10.456789016723633, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.43539439141750336, + "loss_total_v6": 0.6535492911934853, + "step": 1030 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23778018018018018, + "grad_norm": 24.495641708374023, + "learning_rate": 9.036585286624175e-06, + "loss": 10.268438339233398, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4295581579208374, + "loss_total_v6": 0.6417773813009262, + "step": 1031 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23801081081081082, + "grad_norm": 24.158918380737305, + "learning_rate": 9.034342935068953e-06, + "loss": 10.555500030517578, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.44672132283449173, + "loss_total_v6": 0.6597187891602516, + "step": 1032 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23824144144144144, + "grad_norm": 23.49578857421875, + "learning_rate": 9.032098255896965e-06, + "loss": 10.607995986938477, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.452138464897871, + "loss_total_v6": 0.6629997193813324, + "step": 1033 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23847207207207208, + "grad_norm": 24.93792152404785, + "learning_rate": 9.029851250403282e-06, + "loss": 10.122729301452637, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.4122115783393383, + "loss_total_v6": 0.6326705738902092, + "step": 1034 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2387027027027027, + "grad_norm": 24.314502716064453, + "learning_rate": 9.027601919884314e-06, + "loss": 10.570121765136719, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4448428191244602, + "loss_total_v6": 0.6606325954198837, + "step": 1035 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23893333333333333, + "grad_norm": 23.55431365966797, + "learning_rate": 9.025350265637816e-06, + "loss": 10.388965606689453, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.43731996417045593, + "loss_total_v6": 0.6493103355169296, + "step": 1036 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23916396396396397, + "grad_norm": 21.382930755615234, + "learning_rate": 9.023096288962878e-06, + "loss": 10.066247940063477, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4204460605978966, + "loss_total_v6": 0.6291405260562897, + "step": 1037 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23939459459459458, + "grad_norm": 24.039443969726562, + "learning_rate": 9.020839991159937e-06, + "loss": 10.158254623413086, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.41707170009613037, + "loss_total_v6": 0.6348909437656403, + "step": 1038 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23962522522522522, + "grad_norm": 24.517887115478516, + "learning_rate": 9.018581373530762e-06, + "loss": 10.125785827636719, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.41482874751091003, + "loss_total_v6": 0.6328615769743919, + "step": 1039 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.23985585585585587, + "grad_norm": 24.082496643066406, + "learning_rate": 9.016320437378466e-06, + "loss": 10.006940841674805, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4141148030757904, + "loss_total_v6": 0.6254337728023529, + "step": 1040 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24008648648648648, + "grad_norm": 26.258533477783203, + "learning_rate": 9.014057184007497e-06, + "loss": 10.545848846435547, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4468047395348549, + "loss_total_v6": 0.6591155230998993, + "step": 1041 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24031711711711712, + "grad_norm": 22.004520416259766, + "learning_rate": 9.011791614723642e-06, + "loss": 10.999762535095215, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.46882668510079384, + "loss_total_v6": 0.6874851286411285, + "step": 1042 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24054774774774776, + "grad_norm": 22.221038818359375, + "learning_rate": 9.00952373083402e-06, + "loss": 10.159215927124023, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.41873394697904587, + "loss_total_v6": 0.6349510028958321, + "step": 1043 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24077837837837837, + "grad_norm": 32.69746398925781, + "learning_rate": 9.007253533647088e-06, + "loss": 10.758537292480469, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4562525525689125, + "loss_total_v6": 0.6724085658788681, + "step": 1044 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.241009009009009, + "grad_norm": 22.96283721923828, + "learning_rate": 9.004981024472641e-06, + "loss": 10.438536643981934, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.44002143293619156, + "loss_total_v6": 0.6524085178971291, + "step": 1045 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24123963963963965, + "grad_norm": 21.89017677307129, + "learning_rate": 9.002706204621802e-06, + "loss": 10.846189498901367, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.45620715990662575, + "loss_total_v6": 0.6778868436813354, + "step": 1046 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24147027027027027, + "grad_norm": 24.7288818359375, + "learning_rate": 9.00042907540703e-06, + "loss": 10.000226974487305, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.40972786024212837, + "loss_total_v6": 0.6250141337513924, + "step": 1047 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2417009009009009, + "grad_norm": 21.900535583496094, + "learning_rate": 8.99814963814212e-06, + "loss": 10.727630615234375, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.46500205993652344, + "loss_total_v6": 0.670476920902729, + "step": 1048 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24193153153153152, + "grad_norm": 22.68524932861328, + "learning_rate": 8.995867894142187e-06, + "loss": 10.304590225219727, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.4355865940451622, + "loss_total_v6": 0.6440368816256523, + "step": 1049 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24216216216216216, + "grad_norm": 21.969266891479492, + "learning_rate": 8.993583844723693e-06, + "loss": 10.561393737792969, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.44562484323978424, + "loss_total_v6": 0.6600871384143829, + "step": 1050 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2423927927927928, + "grad_norm": 23.46780014038086, + "learning_rate": 8.991297491204415e-06, + "loss": 10.76030158996582, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.45880426093935966, + "loss_total_v6": 0.6725188419222832, + "step": 1051 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2426234234234234, + "grad_norm": 23.71099281311035, + "learning_rate": 8.98900883490347e-06, + "loss": 10.907651901245117, + "loss_alignment": 0.44384765625, + "loss_alignment_w": 0.221923828125, + "loss_sft": 0.4598960019648075, + "loss_total_v6": 0.681728295981884, + "step": 1052 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24285405405405405, + "grad_norm": 22.43546485900879, + "learning_rate": 8.986717877141297e-06, + "loss": 10.49392032623291, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.441255159676075, + "loss_total_v6": 0.6558700129389763, + "step": 1053 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2430846846846847, + "grad_norm": 22.655517578125, + "learning_rate": 8.984424619239671e-06, + "loss": 10.927101135253906, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.46663523092865944, + "loss_total_v6": 0.6829438358545303, + "step": 1054 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2433153153153153, + "grad_norm": 24.136781692504883, + "learning_rate": 8.982129062521683e-06, + "loss": 10.570282936096191, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.44682130962610245, + "loss_total_v6": 0.6606426984071732, + "step": 1055 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24354594594594595, + "grad_norm": 24.357595443725586, + "learning_rate": 8.979831208311758e-06, + "loss": 10.340642929077148, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.4253886751830578, + "loss_total_v6": 0.6462901681661606, + "step": 1056 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2437765765765766, + "grad_norm": 23.203981399536133, + "learning_rate": 8.977531057935644e-06, + "loss": 10.457801818847656, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4333215430378914, + "loss_total_v6": 0.6536126509308815, + "step": 1057 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2440072072072072, + "grad_norm": 22.872209548950195, + "learning_rate": 8.975228612720415e-06, + "loss": 10.514719009399414, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4443861246109009, + "loss_total_v6": 0.6571699380874634, + "step": 1058 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24423783783783784, + "grad_norm": 22.483543395996094, + "learning_rate": 8.972923873994468e-06, + "loss": 10.585227966308594, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4405379444360733, + "loss_total_v6": 0.6615767478942871, + "step": 1059 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24446846846846848, + "grad_norm": 23.401643753051758, + "learning_rate": 8.970616843087526e-06, + "loss": 10.647919654846191, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.44979672878980637, + "loss_total_v6": 0.6654949709773064, + "step": 1060 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2446990990990991, + "grad_norm": 25.689428329467773, + "learning_rate": 8.968307521330626e-06, + "loss": 10.886295318603516, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.46277254447340965, + "loss_total_v6": 0.680393397808075, + "step": 1061 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24492972972972973, + "grad_norm": 21.490554809570312, + "learning_rate": 8.965995910056135e-06, + "loss": 9.633011817932129, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.39042383804917336, + "loss_total_v6": 0.6020632311701775, + "step": 1062 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24516036036036035, + "grad_norm": 21.340049743652344, + "learning_rate": 8.963682010597744e-06, + "loss": 10.49880599975586, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.4461533986032009, + "loss_total_v6": 0.6561753749847412, + "step": 1063 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.245390990990991, + "grad_norm": 20.09687614440918, + "learning_rate": 8.961365824290451e-06, + "loss": 10.13120174407959, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.41861579567193985, + "loss_total_v6": 0.6332001313567162, + "step": 1064 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24562162162162163, + "grad_norm": 20.842924118041992, + "learning_rate": 8.959047352470584e-06, + "loss": 10.296935081481934, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.43339915201067924, + "loss_total_v6": 0.6435584276914597, + "step": 1065 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24585225225225224, + "grad_norm": 22.11627769470215, + "learning_rate": 8.956726596475792e-06, + "loss": 10.21885871887207, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.421607106924057, + "loss_total_v6": 0.638678640127182, + "step": 1066 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24608288288288288, + "grad_norm": 23.668025970458984, + "learning_rate": 8.954403557645029e-06, + "loss": 10.503923416137695, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.4442911446094513, + "loss_total_v6": 0.6564951613545418, + "step": 1067 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24631351351351352, + "grad_norm": 22.232603073120117, + "learning_rate": 8.952078237318576e-06, + "loss": 9.776695251464844, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.40485142543911934, + "loss_total_v6": 0.6110434457659721, + "step": 1068 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24654414414414413, + "grad_norm": 22.813913345336914, + "learning_rate": 8.94975063683803e-06, + "loss": 10.177896499633789, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.4306436590850353, + "loss_total_v6": 0.6361185163259506, + "step": 1069 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24677477477477477, + "grad_norm": 22.169954299926758, + "learning_rate": 8.947420757546301e-06, + "loss": 10.631819725036621, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.4459065869450569, + "loss_total_v6": 0.6644887328147888, + "step": 1070 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24700540540540541, + "grad_norm": 20.41404151916504, + "learning_rate": 8.945088600787614e-06, + "loss": 10.47078800201416, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.4512229599058628, + "loss_total_v6": 0.6544242650270462, + "step": 1071 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24723603603603603, + "grad_norm": 22.483591079711914, + "learning_rate": 8.942754167907506e-06, + "loss": 10.26797866821289, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.42354802042245865, + "loss_total_v6": 0.6417486816644669, + "step": 1072 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24746666666666667, + "grad_norm": 25.457828521728516, + "learning_rate": 8.940417460252834e-06, + "loss": 10.68377685546875, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4555320478975773, + "loss_total_v6": 0.6677360236644745, + "step": 1073 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2476972972972973, + "grad_norm": 19.479936599731445, + "learning_rate": 8.938078479171761e-06, + "loss": 10.065925598144531, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4118809849023819, + "loss_total_v6": 0.629120372235775, + "step": 1074 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24792792792792792, + "grad_norm": 21.906373977661133, + "learning_rate": 8.935737226013763e-06, + "loss": 10.229740142822266, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4272005334496498, + "loss_total_v6": 0.639358714222908, + "step": 1075 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24815855855855856, + "grad_norm": 23.033430099487305, + "learning_rate": 8.933393702129627e-06, + "loss": 10.166460990905762, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4221011959016323, + "loss_total_v6": 0.6354037970304489, + "step": 1076 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2483891891891892, + "grad_norm": 21.73366928100586, + "learning_rate": 8.931047908871455e-06, + "loss": 10.394893646240234, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.43648501858115196, + "loss_total_v6": 0.6496808156371117, + "step": 1077 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24861981981981982, + "grad_norm": 19.422941207885742, + "learning_rate": 8.928699847592647e-06, + "loss": 10.090782165527344, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4126562885940075, + "loss_total_v6": 0.6306738927960396, + "step": 1078 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24885045045045046, + "grad_norm": 21.541372299194336, + "learning_rate": 8.926349519647924e-06, + "loss": 10.634796142578125, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4471454620361328, + "loss_total_v6": 0.6646747663617134, + "step": 1079 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24908108108108107, + "grad_norm": 22.760637283325195, + "learning_rate": 8.923996926393306e-06, + "loss": 10.995208740234375, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.46954918280243874, + "loss_total_v6": 0.6872005686163902, + "step": 1080 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2493117117117117, + "grad_norm": 23.310752868652344, + "learning_rate": 8.921642069186125e-06, + "loss": 10.54114055633545, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.44843312352895737, + "loss_total_v6": 0.6588212847709656, + "step": 1081 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24954234234234235, + "grad_norm": 22.713623046875, + "learning_rate": 8.919284949385016e-06, + "loss": 10.753225326538086, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4561799541115761, + "loss_total_v6": 0.6720765605568886, + "step": 1082 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.24977297297297296, + "grad_norm": 21.31304931640625, + "learning_rate": 8.916925568349922e-06, + "loss": 10.886575698852539, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4702821597456932, + "loss_total_v6": 0.6804109290242195, + "step": 1083 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25000360360360363, + "grad_norm": 22.813152313232422, + "learning_rate": 8.91456392744209e-06, + "loss": 10.213420867919922, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.4316127225756645, + "loss_total_v6": 0.6383388265967369, + "step": 1084 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25023423423423424, + "grad_norm": 22.675060272216797, + "learning_rate": 8.912200028024069e-06, + "loss": 10.334040641784668, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.42283984273672104, + "loss_total_v6": 0.6458775624632835, + "step": 1085 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25046486486486486, + "grad_norm": 23.55803680419922, + "learning_rate": 8.909833871459711e-06, + "loss": 10.487505912780762, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4439975880086422, + "loss_total_v6": 0.6554691344499588, + "step": 1086 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25069549549549547, + "grad_norm": 21.458457946777344, + "learning_rate": 8.907465459114175e-06, + "loss": 10.201068878173828, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4224788695573807, + "loss_total_v6": 0.6375667601823807, + "step": 1087 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25092612612612614, + "grad_norm": 20.80111312866211, + "learning_rate": 8.905094792353917e-06, + "loss": 10.398269653320312, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.4354295916855335, + "loss_total_v6": 0.6498918756842613, + "step": 1088 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25115675675675675, + "grad_norm": 21.319589614868164, + "learning_rate": 8.902721872546692e-06, + "loss": 10.786666870117188, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.4561948627233505, + "loss_total_v6": 0.6741666346788406, + "step": 1089 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25138738738738736, + "grad_norm": 23.645000457763672, + "learning_rate": 8.90034670106156e-06, + "loss": 10.400188446044922, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.434115182608366, + "loss_total_v6": 0.6500117853283882, + "step": 1090 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25161801801801803, + "grad_norm": 23.27083396911621, + "learning_rate": 8.897969279268877e-06, + "loss": 10.59597110748291, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.4400802217423916, + "loss_total_v6": 0.6622481718659401, + "step": 1091 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25184864864864864, + "grad_norm": 22.66503143310547, + "learning_rate": 8.895589608540299e-06, + "loss": 10.338602066040039, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.43356189876794815, + "loss_total_v6": 0.6461625844240189, + "step": 1092 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25207927927927926, + "grad_norm": 24.077425003051758, + "learning_rate": 8.893207690248776e-06, + "loss": 10.689538955688477, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4493309445679188, + "loss_total_v6": 0.668096199631691, + "step": 1093 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2523099099099099, + "grad_norm": 22.983415603637695, + "learning_rate": 8.890823525768562e-06, + "loss": 10.049125671386719, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4173922762274742, + "loss_total_v6": 0.6280704066157341, + "step": 1094 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25254054054054054, + "grad_norm": 22.21544647216797, + "learning_rate": 8.888437116475196e-06, + "loss": 10.431268692016602, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.438377033919096, + "loss_total_v6": 0.6519543156027794, + "step": 1095 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25277117117117115, + "grad_norm": 22.50020980834961, + "learning_rate": 8.886048463745525e-06, + "loss": 10.592019081115723, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.44300706312060356, + "loss_total_v6": 0.6620012074708939, + "step": 1096 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2530018018018018, + "grad_norm": 24.631229400634766, + "learning_rate": 8.88365756895768e-06, + "loss": 10.664290428161621, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.445341981947422, + "loss_total_v6": 0.6665181443095207, + "step": 1097 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25323243243243243, + "grad_norm": 22.831233978271484, + "learning_rate": 8.881264433491089e-06, + "loss": 10.787798881530762, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4632236137986183, + "loss_total_v6": 0.6742374151945114, + "step": 1098 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25346306306306304, + "grad_norm": 25.09992027282715, + "learning_rate": 8.878869058726474e-06, + "loss": 10.380910873413086, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4365419037640095, + "loss_total_v6": 0.6488069221377373, + "step": 1099 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2536936936936937, + "grad_norm": 22.384899139404297, + "learning_rate": 8.876471446045847e-06, + "loss": 10.315011978149414, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.43495624512434006, + "loss_total_v6": 0.6446882784366608, + "step": 1100 + }, + { + "epoch": 0.2536936936936937, + "eval_loss": 0.6892310380935669, + "eval_loss_alignment": 0.41081563835470086, + "eval_loss_alignment_w": 0.20540781917735043, + "eval_loss_sft": 0.48382327388016844, + "eval_loss_total_v6": 0.6892310928744383, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.5754, + "eval_samples_per_second": 22.598, + "eval_steps_per_second": 2.834, + "step": 1100 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2539243243243243, + "grad_norm": 34.21803283691406, + "learning_rate": 8.874071596832514e-06, + "loss": 10.430734634399414, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4378553405404091, + "loss_total_v6": 0.6519208922982216, + "step": 1101 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25415495495495494, + "grad_norm": 24.550464630126953, + "learning_rate": 8.871669512471068e-06, + "loss": 10.74827766418457, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4588920436799526, + "loss_total_v6": 0.6717674136161804, + "step": 1102 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2543855855855856, + "grad_norm": 25.286924362182617, + "learning_rate": 8.869265194347394e-06, + "loss": 10.042154312133789, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4121805503964424, + "loss_total_v6": 0.6276346296072006, + "step": 1103 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2546162162162162, + "grad_norm": 24.922399520874023, + "learning_rate": 8.866858643848664e-06, + "loss": 10.968711853027344, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.46326974779367447, + "loss_total_v6": 0.685544528067112, + "step": 1104 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.25484684684684683, + "grad_norm": 23.53533172607422, + "learning_rate": 8.86444986236334e-06, + "loss": 10.854917526245117, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4614829085767269, + "loss_total_v6": 0.678432397544384, + "step": 1105 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2550774774774775, + "grad_norm": 28.957006454467773, + "learning_rate": 8.862038851281171e-06, + "loss": 10.503000259399414, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4393506236374378, + "loss_total_v6": 0.656437486410141, + "step": 1106 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2553081081081081, + "grad_norm": 22.57535171508789, + "learning_rate": 8.859625611993189e-06, + "loss": 10.736867904663086, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.45347917452454567, + "loss_total_v6": 0.6710542291402817, + "step": 1107 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2555387387387387, + "grad_norm": 22.3059139251709, + "learning_rate": 8.857210145891714e-06, + "loss": 10.606062889099121, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.43991752341389656, + "loss_total_v6": 0.6628789305686951, + "step": 1108 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2557693693693694, + "grad_norm": 23.360795974731445, + "learning_rate": 8.854792454370354e-06, + "loss": 10.229665756225586, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4306291490793228, + "loss_total_v6": 0.6393541023135185, + "step": 1109 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.256, + "grad_norm": 23.1040096282959, + "learning_rate": 8.852372538823992e-06, + "loss": 10.151273727416992, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.4156435914337635, + "loss_total_v6": 0.6344546303153038, + "step": 1110 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2562306306306306, + "grad_norm": 22.831148147583008, + "learning_rate": 8.849950400648807e-06, + "loss": 10.523601531982422, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.44411735236644745, + "loss_total_v6": 0.6577251255512238, + "step": 1111 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2564612612612613, + "grad_norm": 23.399139404296875, + "learning_rate": 8.847526041242247e-06, + "loss": 10.391077041625977, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.43574298173189163, + "loss_total_v6": 0.6494423225522041, + "step": 1112 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2566918918918919, + "grad_norm": 21.564546585083008, + "learning_rate": 8.845099462003049e-06, + "loss": 10.232271194458008, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.4194546714425087, + "loss_total_v6": 0.6395169422030449, + "step": 1113 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2569225225225225, + "grad_norm": 23.777103424072266, + "learning_rate": 8.842670664331233e-06, + "loss": 11.178375244140625, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.48531531170010567, + "loss_total_v6": 0.6986484676599503, + "step": 1114 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2571531531531532, + "grad_norm": 22.749265670776367, + "learning_rate": 8.84023964962809e-06, + "loss": 10.503593444824219, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.43915891647338867, + "loss_total_v6": 0.6564745903015137, + "step": 1115 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2573837837837838, + "grad_norm": 22.262344360351562, + "learning_rate": 8.8378064192962e-06, + "loss": 10.683314323425293, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.4464241899549961, + "loss_total_v6": 0.6677071452140808, + "step": 1116 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2576144144144144, + "grad_norm": 23.255491256713867, + "learning_rate": 8.835370974739417e-06, + "loss": 10.187437057495117, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.4311027005314827, + "loss_total_v6": 0.636714868247509, + "step": 1117 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.257845045045045, + "grad_norm": 22.608749389648438, + "learning_rate": 8.832933317362869e-06, + "loss": 10.772697448730469, + "loss_alignment": 0.442138671875, + "loss_alignment_w": 0.2210693359375, + "loss_sft": 0.4526209868490696, + "loss_total_v6": 0.6732935830950737, + "step": 1118 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2580756756756757, + "grad_norm": 23.865245819091797, + "learning_rate": 8.830493448572964e-06, + "loss": 10.78650951385498, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4596487954258919, + "loss_total_v6": 0.6741568446159363, + "step": 1119 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2583063063063063, + "grad_norm": 22.741127014160156, + "learning_rate": 8.82805136977739e-06, + "loss": 10.621652603149414, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4507185034453869, + "loss_total_v6": 0.6638532727956772, + "step": 1120 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2585369369369369, + "grad_norm": 21.828350067138672, + "learning_rate": 8.825607082385101e-06, + "loss": 10.699837684631348, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.450905404984951, + "loss_total_v6": 0.6687398478388786, + "step": 1121 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2587675675675676, + "grad_norm": 25.808412551879883, + "learning_rate": 8.823160587806335e-06, + "loss": 10.56263542175293, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.4558647684752941, + "loss_total_v6": 0.6601646840572357, + "step": 1122 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2589981981981982, + "grad_norm": 23.73480224609375, + "learning_rate": 8.820711887452596e-06, + "loss": 10.005167007446289, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4133783429861069, + "loss_total_v6": 0.6253229230642319, + "step": 1123 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2592288288288288, + "grad_norm": 23.749488830566406, + "learning_rate": 8.818260982736662e-06, + "loss": 10.540964126586914, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.44067057222127914, + "loss_total_v6": 0.6588102281093597, + "step": 1124 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2594594594594595, + "grad_norm": 23.651691436767578, + "learning_rate": 8.815807875072586e-06, + "loss": 10.601941108703613, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.44976119324564934, + "loss_total_v6": 0.662621296942234, + "step": 1125 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2596900900900901, + "grad_norm": 21.732852935791016, + "learning_rate": 8.813352565875688e-06, + "loss": 10.189624786376953, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4260513670742512, + "loss_total_v6": 0.6368515342473984, + "step": 1126 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2599207207207207, + "grad_norm": 24.586671829223633, + "learning_rate": 8.810895056562563e-06, + "loss": 10.221651077270508, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4233533442020416, + "loss_total_v6": 0.6388532072305679, + "step": 1127 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26015135135135137, + "grad_norm": 21.414073944091797, + "learning_rate": 8.808435348551072e-06, + "loss": 10.710559844970703, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4549630284309387, + "loss_total_v6": 0.6694100275635719, + "step": 1128 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.260381981981982, + "grad_norm": 22.38744354248047, + "learning_rate": 8.805973443260345e-06, + "loss": 10.228034973144531, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4246067628264427, + "loss_total_v6": 0.6392521485686302, + "step": 1129 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2606126126126126, + "grad_norm": 23.09816551208496, + "learning_rate": 8.803509342110779e-06, + "loss": 9.959186553955078, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4089939817786217, + "loss_total_v6": 0.6224491745233536, + "step": 1130 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26084324324324326, + "grad_norm": 21.594085693359375, + "learning_rate": 8.801043046524039e-06, + "loss": 10.354904174804688, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4358777478337288, + "loss_total_v6": 0.64718147367239, + "step": 1131 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2610738738738739, + "grad_norm": 24.414745330810547, + "learning_rate": 8.798574557923055e-06, + "loss": 9.998697280883789, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.4127908758819103, + "loss_total_v6": 0.6249185651540756, + "step": 1132 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2613045045045045, + "grad_norm": 21.12693214416504, + "learning_rate": 8.796103877732024e-06, + "loss": 10.45541763305664, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.44371625036001205, + "loss_total_v6": 0.6534635871648788, + "step": 1133 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26153513513513515, + "grad_norm": 24.21234893798828, + "learning_rate": 8.793631007376407e-06, + "loss": 10.098505020141602, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.423408143222332, + "loss_total_v6": 0.6311565712094307, + "step": 1134 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26176576576576577, + "grad_norm": 23.728635787963867, + "learning_rate": 8.791155948282927e-06, + "loss": 10.116436004638672, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.4223926477134228, + "loss_total_v6": 0.63227728754282, + "step": 1135 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2619963963963964, + "grad_norm": 22.916667938232422, + "learning_rate": 8.788678701879573e-06, + "loss": 11.18708610534668, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.4883164167404175, + "loss_total_v6": 0.6991928592324257, + "step": 1136 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26222702702702705, + "grad_norm": 23.81415367126465, + "learning_rate": 8.786199269595593e-06, + "loss": 10.284822463989258, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.4316960424184799, + "loss_total_v6": 0.6428014189004898, + "step": 1137 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26245765765765766, + "grad_norm": 21.66764259338379, + "learning_rate": 8.783717652861495e-06, + "loss": 9.841957092285156, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.3942207880318165, + "loss_total_v6": 0.6151223033666611, + "step": 1138 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2626882882882883, + "grad_norm": 22.050403594970703, + "learning_rate": 8.781233853109052e-06, + "loss": 10.564569473266602, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4432598352432251, + "loss_total_v6": 0.6602855995297432, + "step": 1139 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26291891891891894, + "grad_norm": 22.207712173461914, + "learning_rate": 8.778747871771293e-06, + "loss": 10.80914306640625, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4623756557703018, + "loss_total_v6": 0.6755714789032936, + "step": 1140 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26314954954954956, + "grad_norm": 22.876829147338867, + "learning_rate": 8.776259710282504e-06, + "loss": 10.385934829711914, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.43073710054159164, + "loss_total_v6": 0.6491209045052528, + "step": 1141 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26338018018018017, + "grad_norm": 21.485990524291992, + "learning_rate": 8.773769370078238e-06, + "loss": 10.79118824005127, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.46308453753590584, + "loss_total_v6": 0.6744492873549461, + "step": 1142 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26361081081081084, + "grad_norm": 25.99393653869629, + "learning_rate": 8.77127685259529e-06, + "loss": 9.885294914245605, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.40451309084892273, + "loss_total_v6": 0.6178309470415115, + "step": 1143 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26384144144144145, + "grad_norm": 23.386014938354492, + "learning_rate": 8.768782159271727e-06, + "loss": 10.544614791870117, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4442709796130657, + "loss_total_v6": 0.6590384393930435, + "step": 1144 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26407207207207206, + "grad_norm": 24.47325325012207, + "learning_rate": 8.766285291546858e-06, + "loss": 10.211576461791992, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4203127399086952, + "loss_total_v6": 0.6382235065102577, + "step": 1145 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26430270270270273, + "grad_norm": 23.20713996887207, + "learning_rate": 8.763786250861258e-06, + "loss": 10.479147911071777, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4417661875486374, + "loss_total_v6": 0.6549467444419861, + "step": 1146 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26453333333333334, + "grad_norm": 22.886560440063477, + "learning_rate": 8.761285038656747e-06, + "loss": 10.213601112365723, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4208665154874325, + "loss_total_v6": 0.6383500322699547, + "step": 1147 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26476396396396396, + "grad_norm": 20.468568801879883, + "learning_rate": 8.758781656376398e-06, + "loss": 10.309268951416016, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4316523000597954, + "loss_total_v6": 0.6443293020129204, + "step": 1148 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26499459459459457, + "grad_norm": 23.237655639648438, + "learning_rate": 8.756276105464545e-06, + "loss": 10.454673767089844, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.4327140115201473, + "loss_total_v6": 0.6534171476960182, + "step": 1149 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26522522522522524, + "grad_norm": 23.264436721801758, + "learning_rate": 8.753768387366764e-06, + "loss": 10.317808151245117, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4317588061094284, + "loss_total_v6": 0.6448630318045616, + "step": 1150 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26545585585585585, + "grad_norm": 23.36371421813965, + "learning_rate": 8.751258503529885e-06, + "loss": 10.585460662841797, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.4408728815615177, + "loss_total_v6": 0.6615912467241287, + "step": 1151 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26568648648648646, + "grad_norm": 21.862674713134766, + "learning_rate": 8.748746455401986e-06, + "loss": 10.24338436126709, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.42500152811408043, + "loss_total_v6": 0.6402114927768707, + "step": 1152 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26591711711711713, + "grad_norm": 23.02408218383789, + "learning_rate": 8.746232244432397e-06, + "loss": 10.58200454711914, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.452482458204031, + "loss_total_v6": 0.6613752916455269, + "step": 1153 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26614774774774774, + "grad_norm": 23.944013595581055, + "learning_rate": 8.743715872071694e-06, + "loss": 9.914993286132812, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4087190367281437, + "loss_total_v6": 0.6196870505809784, + "step": 1154 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26637837837837836, + "grad_norm": 21.031448364257812, + "learning_rate": 8.741197339771697e-06, + "loss": 10.653045654296875, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.44953731447458267, + "loss_total_v6": 0.6658153906464577, + "step": 1155 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.266609009009009, + "grad_norm": 20.347591400146484, + "learning_rate": 8.738676648985476e-06, + "loss": 9.930330276489258, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.4128056727349758, + "loss_total_v6": 0.6206456422805786, + "step": 1156 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26683963963963964, + "grad_norm": 21.03668212890625, + "learning_rate": 8.736153801167346e-06, + "loss": 9.874401092529297, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.41339947283267975, + "loss_total_v6": 0.6171500906348228, + "step": 1157 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26707027027027025, + "grad_norm": 25.25168800354004, + "learning_rate": 8.733628797772864e-06, + "loss": 10.263867378234863, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.42878415435552597, + "loss_total_v6": 0.6414916887879372, + "step": 1158 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2673009009009009, + "grad_norm": 22.592153549194336, + "learning_rate": 8.731101640258835e-06, + "loss": 10.679474830627441, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4486103653907776, + "loss_total_v6": 0.6674671843647957, + "step": 1159 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26753153153153153, + "grad_norm": 23.013940811157227, + "learning_rate": 8.7285723300833e-06, + "loss": 10.58464241027832, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.4448195770382881, + "loss_total_v6": 0.6615401580929756, + "step": 1160 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26776216216216214, + "grad_norm": 20.882328033447266, + "learning_rate": 8.72604086870555e-06, + "loss": 9.903535842895508, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.40424931794404984, + "loss_total_v6": 0.6189710050821304, + "step": 1161 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2679927927927928, + "grad_norm": 22.621654510498047, + "learning_rate": 8.723507257586112e-06, + "loss": 10.183831214904785, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4174494780600071, + "loss_total_v6": 0.6364894211292267, + "step": 1162 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2682234234234234, + "grad_norm": 21.81535530090332, + "learning_rate": 8.720971498186754e-06, + "loss": 11.00909423828125, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.46883007884025574, + "loss_total_v6": 0.6880683526396751, + "step": 1163 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26845405405405404, + "grad_norm": 20.617603302001953, + "learning_rate": 8.718433591970486e-06, + "loss": 10.692569732666016, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.4488641284406185, + "loss_total_v6": 0.6682855635881424, + "step": 1164 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2686846846846847, + "grad_norm": 30.551780700683594, + "learning_rate": 8.715893540401552e-06, + "loss": 9.881013870239258, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.4039098210632801, + "loss_total_v6": 0.6175633892416954, + "step": 1165 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2689153153153153, + "grad_norm": 20.974973678588867, + "learning_rate": 8.713351344945438e-06, + "loss": 10.323362350463867, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.4243849255144596, + "loss_total_v6": 0.6452101320028305, + "step": 1166 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.26914594594594593, + "grad_norm": 22.180227279663086, + "learning_rate": 8.710807007068865e-06, + "loss": 10.564128875732422, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4454905726015568, + "loss_total_v6": 0.6602579951286316, + "step": 1167 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2693765765765766, + "grad_norm": 23.772289276123047, + "learning_rate": 8.708260528239788e-06, + "loss": 10.404829025268555, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.44127166643738747, + "loss_total_v6": 0.6503018215298653, + "step": 1168 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2696072072072072, + "grad_norm": 22.50101661682129, + "learning_rate": 8.705711909927404e-06, + "loss": 9.96674919128418, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.4156922399997711, + "loss_total_v6": 0.6229218244552612, + "step": 1169 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2698378378378378, + "grad_norm": 29.8793888092041, + "learning_rate": 8.703161153602137e-06, + "loss": 10.63315200805664, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.45317674055695534, + "loss_total_v6": 0.6645720079541206, + "step": 1170 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2700684684684685, + "grad_norm": 22.663494110107422, + "learning_rate": 8.700608260735648e-06, + "loss": 10.431292533874512, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.432549674063921, + "loss_total_v6": 0.6519558131694794, + "step": 1171 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2702990990990991, + "grad_norm": 22.309497833251953, + "learning_rate": 8.698053232800831e-06, + "loss": 10.09249210357666, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.4118781127035618, + "loss_total_v6": 0.6307807341217995, + "step": 1172 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2705297297297297, + "grad_norm": 21.620872497558594, + "learning_rate": 8.695496071271811e-06, + "loss": 10.822443008422852, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.46230658516287804, + "loss_total_v6": 0.6764026582241058, + "step": 1173 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2707603603603604, + "grad_norm": 22.721471786499023, + "learning_rate": 8.692936777623943e-06, + "loss": 10.629210472106934, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.45154182240366936, + "loss_total_v6": 0.664325624704361, + "step": 1174 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.270990990990991, + "grad_norm": 23.169193267822266, + "learning_rate": 8.690375353333815e-06, + "loss": 9.817276954650879, + "loss_alignment": 0.408203125, + "loss_alignment_w": 0.2041015625, + "loss_sft": 0.4095545522868633, + "loss_total_v6": 0.6135798096656799, + "step": 1175 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2712216216216216, + "grad_norm": 21.978179931640625, + "learning_rate": 8.68781179987924e-06, + "loss": 10.614645004272461, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4434598982334137, + "loss_total_v6": 0.66341532766819, + "step": 1176 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2714522522522523, + "grad_norm": 23.711605072021484, + "learning_rate": 8.685246118739264e-06, + "loss": 10.583873748779297, + "loss_alignment": 0.4462890625, + "loss_alignment_w": 0.22314453125, + "loss_sft": 0.43871376663446426, + "loss_total_v6": 0.6614921391010284, + "step": 1177 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2716828828828829, + "grad_norm": 22.769004821777344, + "learning_rate": 8.682678311394155e-06, + "loss": 10.600667953491211, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4497121535241604, + "loss_total_v6": 0.6625417470932007, + "step": 1178 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2719135135135135, + "grad_norm": 33.379032135009766, + "learning_rate": 8.680108379325413e-06, + "loss": 10.422691345214844, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.43854278698563576, + "loss_total_v6": 0.6514181643724442, + "step": 1179 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2721441441441441, + "grad_norm": 22.761308670043945, + "learning_rate": 8.677536324015765e-06, + "loss": 11.16382122039795, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.48579422757029533, + "loss_total_v6": 0.6977388113737106, + "step": 1180 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2723747747747748, + "grad_norm": 22.073688507080078, + "learning_rate": 8.674962146949158e-06, + "loss": 10.15267562866211, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4213159382343292, + "loss_total_v6": 0.6345422565937042, + "step": 1181 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2726054054054054, + "grad_norm": 26.128978729248047, + "learning_rate": 8.67238584961076e-06, + "loss": 10.431291580200195, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4370661526918411, + "loss_total_v6": 0.6519556865096092, + "step": 1182 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.272836036036036, + "grad_norm": 21.8289852142334, + "learning_rate": 8.669807433486977e-06, + "loss": 10.175273895263672, + "loss_alignment": 0.44580078125, + "loss_alignment_w": 0.222900390625, + "loss_sft": 0.4125811606645584, + "loss_total_v6": 0.6359546035528183, + "step": 1183 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2730666666666667, + "grad_norm": 24.018888473510742, + "learning_rate": 8.66722690006542e-06, + "loss": 10.257234573364258, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4255468137562275, + "loss_total_v6": 0.6410771906375885, + "step": 1184 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2732972972972973, + "grad_norm": 22.12351417541504, + "learning_rate": 8.664644250834932e-06, + "loss": 10.71811294555664, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4565032087266445, + "loss_total_v6": 0.6698821112513542, + "step": 1185 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2735279279279279, + "grad_norm": 25.82132339477539, + "learning_rate": 8.662059487285575e-06, + "loss": 10.400900840759277, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.43847789242863655, + "loss_total_v6": 0.6500563025474548, + "step": 1186 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2737585585585586, + "grad_norm": 24.495372772216797, + "learning_rate": 8.659472610908628e-06, + "loss": 10.49880313873291, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.4530196711421013, + "loss_total_v6": 0.6561751812696457, + "step": 1187 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2739891891891892, + "grad_norm": 22.11480140686035, + "learning_rate": 8.656883623196592e-06, + "loss": 10.626068115234375, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.44711878150701523, + "loss_total_v6": 0.6641292721033096, + "step": 1188 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2742198198198198, + "grad_norm": 29.827972412109375, + "learning_rate": 8.654292525643188e-06, + "loss": 9.966156005859375, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.41206933185458183, + "loss_total_v6": 0.6228847801685333, + "step": 1189 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27445045045045047, + "grad_norm": 21.36189842224121, + "learning_rate": 8.651699319743348e-06, + "loss": 9.856170654296875, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.40061765536665916, + "loss_total_v6": 0.6160107105970383, + "step": 1190 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2746810810810811, + "grad_norm": 21.316429138183594, + "learning_rate": 8.649104006993223e-06, + "loss": 10.595141410827637, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.4421035684645176, + "loss_total_v6": 0.6621963456273079, + "step": 1191 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2749117117117117, + "grad_norm": 23.02385711669922, + "learning_rate": 8.646506588890184e-06, + "loss": 10.370040893554688, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.43494705110788345, + "loss_total_v6": 0.6481275632977486, + "step": 1192 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27514234234234236, + "grad_norm": 20.853384017944336, + "learning_rate": 8.64390706693281e-06, + "loss": 10.55915355682373, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.44742267206311226, + "loss_total_v6": 0.6599470973014832, + "step": 1193 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.275372972972973, + "grad_norm": 23.592817306518555, + "learning_rate": 8.641305442620897e-06, + "loss": 10.944997787475586, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.464885126799345, + "loss_total_v6": 0.6840623840689659, + "step": 1194 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2756036036036036, + "grad_norm": 22.46593475341797, + "learning_rate": 8.638701717455457e-06, + "loss": 10.349546432495117, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.43845734000205994, + "loss_total_v6": 0.6468466147780418, + "step": 1195 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27583423423423425, + "grad_norm": 21.563796997070312, + "learning_rate": 8.636095892938707e-06, + "loss": 10.273693084716797, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.437638022005558, + "loss_total_v6": 0.6421058103442192, + "step": 1196 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27606486486486487, + "grad_norm": 24.461048126220703, + "learning_rate": 8.633487970574082e-06, + "loss": 10.535645484924316, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.44894417375326157, + "loss_total_v6": 0.6584778502583504, + "step": 1197 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2762954954954955, + "grad_norm": 20.825429916381836, + "learning_rate": 8.630877951866223e-06, + "loss": 10.513971328735352, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.44058575853705406, + "loss_total_v6": 0.6571232303977013, + "step": 1198 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27652612612612615, + "grad_norm": 22.27355194091797, + "learning_rate": 8.628265838320982e-06, + "loss": 10.083989143371582, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4183199927210808, + "loss_total_v6": 0.6302493065595627, + "step": 1199 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27675675675675676, + "grad_norm": 25.62076759338379, + "learning_rate": 8.625651631445419e-06, + "loss": 10.493929862976074, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.44313257932662964, + "loss_total_v6": 0.6558706387877464, + "step": 1200 + }, + { + "epoch": 0.27675675675675676, + "eval_loss": 0.6849018335342407, + "eval_loss_alignment": 0.40893763354700857, + "eval_loss_alignment_w": 0.20446881677350429, + "eval_loss_sft": 0.48043298004911494, + "eval_loss_total_v6": 0.6849017973161406, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.4898, + "eval_samples_per_second": 22.621, + "eval_steps_per_second": 2.837, + "step": 1200 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2769873873873874, + "grad_norm": 22.492076873779297, + "learning_rate": 8.623035332747804e-06, + "loss": 10.670785903930664, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4515157751739025, + "loss_total_v6": 0.6669240891933441, + "step": 1201 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27721801801801804, + "grad_norm": 22.429716110229492, + "learning_rate": 8.620416943737613e-06, + "loss": 10.4779634475708, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.44398101791739464, + "loss_total_v6": 0.6548727452754974, + "step": 1202 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27744864864864865, + "grad_norm": 22.302637100219727, + "learning_rate": 8.617796465925522e-06, + "loss": 9.692096710205078, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.39275864511728287, + "loss_total_v6": 0.6057560592889786, + "step": 1203 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27767927927927927, + "grad_norm": 21.561031341552734, + "learning_rate": 8.615173900823425e-06, + "loss": 10.418024063110352, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4405857101082802, + "loss_total_v6": 0.6511264741420746, + "step": 1204 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27790990990990994, + "grad_norm": 22.999713897705078, + "learning_rate": 8.612549249944408e-06, + "loss": 10.321277618408203, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4292747750878334, + "loss_total_v6": 0.6450798362493515, + "step": 1205 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27814054054054055, + "grad_norm": 22.068273544311523, + "learning_rate": 8.609922514802765e-06, + "loss": 10.650840759277344, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.4556860998272896, + "loss_total_v6": 0.6656775549054146, + "step": 1206 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27837117117117116, + "grad_norm": 23.195573806762695, + "learning_rate": 8.607293696913993e-06, + "loss": 10.458099365234375, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.43613243848085403, + "loss_total_v6": 0.653631217777729, + "step": 1207 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27860180180180183, + "grad_norm": 23.055395126342773, + "learning_rate": 8.60466279779479e-06, + "loss": 10.346402168273926, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.43447668850421906, + "loss_total_v6": 0.646650142967701, + "step": 1208 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27883243243243244, + "grad_norm": 24.551542282104492, + "learning_rate": 8.602029818963057e-06, + "loss": 10.367574691772461, + "loss_alignment": 0.449951171875, + "loss_alignment_w": 0.2249755859375, + "loss_sft": 0.4232419691979885, + "loss_total_v6": 0.6479734405875206, + "step": 1209 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27906306306306305, + "grad_norm": 22.77455711364746, + "learning_rate": 8.59939476193789e-06, + "loss": 10.593474388122559, + "loss_alignment": 0.4140625, + "loss_alignment_w": 0.20703125, + "loss_sft": 0.4546489045023918, + "loss_total_v6": 0.6620921790599823, + "step": 1210 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27929369369369367, + "grad_norm": 24.740663528442383, + "learning_rate": 8.596757628239588e-06, + "loss": 10.538040161132812, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4478425793349743, + "loss_total_v6": 0.6586274951696396, + "step": 1211 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27952432432432434, + "grad_norm": 21.231494903564453, + "learning_rate": 8.594118419389648e-06, + "loss": 10.916239738464355, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4701831229031086, + "loss_total_v6": 0.6822650134563446, + "step": 1212 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27975495495495495, + "grad_norm": 23.959688186645508, + "learning_rate": 8.591477136910763e-06, + "loss": 10.441917419433594, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4351515620946884, + "loss_total_v6": 0.652619831264019, + "step": 1213 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.27998558558558556, + "grad_norm": 21.62595558166504, + "learning_rate": 8.588833782326818e-06, + "loss": 10.48140811920166, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.44309767335653305, + "loss_total_v6": 0.6550880149006844, + "step": 1214 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28021621621621623, + "grad_norm": 21.145858764648438, + "learning_rate": 8.586188357162906e-06, + "loss": 10.614999771118164, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.44133052974939346, + "loss_total_v6": 0.6634375005960464, + "step": 1215 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28044684684684684, + "grad_norm": 20.141443252563477, + "learning_rate": 8.583540862945301e-06, + "loss": 10.135469436645508, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.42295660451054573, + "loss_total_v6": 0.6334668695926666, + "step": 1216 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28067747747747746, + "grad_norm": 24.58244514465332, + "learning_rate": 8.580891301201479e-06, + "loss": 10.08987045288086, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4116380363702774, + "loss_total_v6": 0.6306169256567955, + "step": 1217 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2809081081081081, + "grad_norm": 21.973604202270508, + "learning_rate": 8.578239673460106e-06, + "loss": 10.332807540893555, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4371822848916054, + "loss_total_v6": 0.645800456404686, + "step": 1218 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28113873873873874, + "grad_norm": 22.91581153869629, + "learning_rate": 8.57558598125104e-06, + "loss": 10.05421257019043, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.41366666555404663, + "loss_total_v6": 0.628388337790966, + "step": 1219 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28136936936936935, + "grad_norm": 24.238338470458984, + "learning_rate": 8.572930226105331e-06, + "loss": 10.164884567260742, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.42214006558060646, + "loss_total_v6": 0.6353053376078606, + "step": 1220 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2816, + "grad_norm": 22.908740997314453, + "learning_rate": 8.570272409555217e-06, + "loss": 10.74251937866211, + "loss_alignment": 0.4404296875, + "loss_alignment_w": 0.22021484375, + "loss_sft": 0.45155884698033333, + "loss_total_v6": 0.6714074909687042, + "step": 1221 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28183063063063063, + "grad_norm": 23.774534225463867, + "learning_rate": 8.567612533134128e-06, + "loss": 10.225408554077148, + "loss_alignment": 0.4482421875, + "loss_alignment_w": 0.22412109375, + "loss_sft": 0.41498224437236786, + "loss_total_v6": 0.6390880569815636, + "step": 1222 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28206126126126124, + "grad_norm": 20.19790267944336, + "learning_rate": 8.564950598376683e-06, + "loss": 10.651727676391602, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.45597043633461, + "loss_total_v6": 0.6657329946756363, + "step": 1223 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2822918918918919, + "grad_norm": 22.159727096557617, + "learning_rate": 8.562286606818684e-06, + "loss": 9.96810245513916, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.402043879032135, + "loss_total_v6": 0.6230064034461975, + "step": 1224 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2825225225225225, + "grad_norm": 22.558292388916016, + "learning_rate": 8.559620559997125e-06, + "loss": 10.662261009216309, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4511965848505497, + "loss_total_v6": 0.6663912907242775, + "step": 1225 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28275315315315314, + "grad_norm": 20.87602424621582, + "learning_rate": 8.556952459450183e-06, + "loss": 10.617979049682617, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.46031560003757477, + "loss_total_v6": 0.6636236608028412, + "step": 1226 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2829837837837838, + "grad_norm": 21.87681770324707, + "learning_rate": 8.554282306717218e-06, + "loss": 9.856391906738281, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.4096340350806713, + "loss_total_v6": 0.6160244569182396, + "step": 1227 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2832144144144144, + "grad_norm": 22.96942138671875, + "learning_rate": 8.55161010333878e-06, + "loss": 10.339577674865723, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.43121201545000076, + "loss_total_v6": 0.6462236121296883, + "step": 1228 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28344504504504503, + "grad_norm": 23.705570220947266, + "learning_rate": 8.548935850856597e-06, + "loss": 10.69789981842041, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.44667962193489075, + "loss_total_v6": 0.6686187088489532, + "step": 1229 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2836756756756757, + "grad_norm": 21.71527862548828, + "learning_rate": 8.54625955081358e-06, + "loss": 9.813610076904297, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.39449381828308105, + "loss_total_v6": 0.6133506596088409, + "step": 1230 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2839063063063063, + "grad_norm": 22.754770278930664, + "learning_rate": 8.543581204753822e-06, + "loss": 9.61878490447998, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.38741370290517807, + "loss_total_v6": 0.60117407143116, + "step": 1231 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2841369369369369, + "grad_norm": 22.804407119750977, + "learning_rate": 8.540900814222597e-06, + "loss": 10.3402099609375, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.43253321945667267, + "loss_total_v6": 0.6462630853056908, + "step": 1232 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2843675675675676, + "grad_norm": 23.012598037719727, + "learning_rate": 8.538218380766358e-06, + "loss": 10.93692398071289, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.46958377957344055, + "loss_total_v6": 0.6835577711462975, + "step": 1233 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2845981981981982, + "grad_norm": 20.85350799560547, + "learning_rate": 8.535533905932739e-06, + "loss": 10.22848892211914, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.4232313297688961, + "loss_total_v6": 0.6392805352807045, + "step": 1234 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2848288288288288, + "grad_norm": 22.88716697692871, + "learning_rate": 8.532847391270545e-06, + "loss": 10.2366943359375, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.43045802041888237, + "loss_total_v6": 0.6397933512926102, + "step": 1235 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2850594594594595, + "grad_norm": 24.649066925048828, + "learning_rate": 8.530158838329765e-06, + "loss": 10.527308464050293, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.4401528425514698, + "loss_total_v6": 0.6579567790031433, + "step": 1236 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2852900900900901, + "grad_norm": 22.67498207092285, + "learning_rate": 8.527468248661562e-06, + "loss": 10.286430358886719, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4267764091491699, + "loss_total_v6": 0.6429019123315811, + "step": 1237 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2855207207207207, + "grad_norm": 24.16180419921875, + "learning_rate": 8.52477562381827e-06, + "loss": 10.659478187561035, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4518008939921856, + "loss_total_v6": 0.6662174090743065, + "step": 1238 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2857513513513513, + "grad_norm": 22.599870681762695, + "learning_rate": 8.522080965353405e-06, + "loss": 10.760274887084961, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.45770396292209625, + "loss_total_v6": 0.6725171953439713, + "step": 1239 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.285981981981982, + "grad_norm": 21.922372817993164, + "learning_rate": 8.519384274821649e-06, + "loss": 10.415115356445312, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.43552107363939285, + "loss_total_v6": 0.6509446650743484, + "step": 1240 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2862126126126126, + "grad_norm": 21.28223419189453, + "learning_rate": 8.51668555377886e-06, + "loss": 10.25417423248291, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4251571595668793, + "loss_total_v6": 0.6408859044313431, + "step": 1241 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2864432432432432, + "grad_norm": 22.688337326049805, + "learning_rate": 8.513984803782069e-06, + "loss": 10.544149398803711, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.44013725966215134, + "loss_total_v6": 0.6590093150734901, + "step": 1242 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2866738738738739, + "grad_norm": 20.159406661987305, + "learning_rate": 8.511282026389474e-06, + "loss": 9.881494522094727, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.4079834222793579, + "loss_total_v6": 0.6175934001803398, + "step": 1243 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2869045045045045, + "grad_norm": 24.599641799926758, + "learning_rate": 8.508577223160441e-06, + "loss": 10.278165817260742, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4264582358300686, + "loss_total_v6": 0.6423853635787964, + "step": 1244 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2871351351351351, + "grad_norm": 19.808809280395508, + "learning_rate": 8.505870395655512e-06, + "loss": 9.59697151184082, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.38621821254491806, + "loss_total_v6": 0.5998107343912125, + "step": 1245 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2873657657657658, + "grad_norm": 20.622867584228516, + "learning_rate": 8.503161545436394e-06, + "loss": 10.22711181640625, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4234810434281826, + "loss_total_v6": 0.6391945257782936, + "step": 1246 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2875963963963964, + "grad_norm": 21.661420822143555, + "learning_rate": 8.500450674065954e-06, + "loss": 10.07681941986084, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4188484475016594, + "loss_total_v6": 0.6298012360930443, + "step": 1247 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.287827027027027, + "grad_norm": 23.073314666748047, + "learning_rate": 8.497737783108238e-06, + "loss": 10.32158088684082, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.43100275844335556, + "loss_total_v6": 0.6450988203287125, + "step": 1248 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2880576576576577, + "grad_norm": 21.289445877075195, + "learning_rate": 8.495022874128445e-06, + "loss": 10.431852340698242, + "loss_alignment": 0.405029296875, + "loss_alignment_w": 0.2025146484375, + "loss_sft": 0.44901833310723305, + "loss_total_v6": 0.6519907712936401, + "step": 1249 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2882882882882883, + "grad_norm": 21.02090835571289, + "learning_rate": 8.492305948692944e-06, + "loss": 10.161544799804688, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.4246320612728596, + "loss_total_v6": 0.6350965648889542, + "step": 1250 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2885189189189189, + "grad_norm": 23.43967628479004, + "learning_rate": 8.489587008369269e-06, + "loss": 10.05705451965332, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.4191390201449394, + "loss_total_v6": 0.6285658925771713, + "step": 1251 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28874954954954957, + "grad_norm": 23.21895980834961, + "learning_rate": 8.486866054726115e-06, + "loss": 10.310300827026367, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.43206774443387985, + "loss_total_v6": 0.6443937867879868, + "step": 1252 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2889801801801802, + "grad_norm": 21.34813117980957, + "learning_rate": 8.484143089333334e-06, + "loss": 9.99897575378418, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.41430364549160004, + "loss_total_v6": 0.6249359771609306, + "step": 1253 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2892108108108108, + "grad_norm": 21.885515213012695, + "learning_rate": 8.481418113761947e-06, + "loss": 10.207905769348145, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.42679720744490623, + "loss_total_v6": 0.6379941031336784, + "step": 1254 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.28944144144144146, + "grad_norm": 21.753555297851562, + "learning_rate": 8.478691129584128e-06, + "loss": 10.150277137756348, + "loss_alignment": 0.404541015625, + "loss_alignment_w": 0.2022705078125, + "loss_sft": 0.43181663379073143, + "loss_total_v6": 0.6343923211097717, + "step": 1255 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2896720720720721, + "grad_norm": 23.503931045532227, + "learning_rate": 8.475962138373212e-06, + "loss": 11.008890151977539, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4725251607596874, + "loss_total_v6": 0.6880556046962738, + "step": 1256 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2899027027027027, + "grad_norm": 20.846982955932617, + "learning_rate": 8.473231141703694e-06, + "loss": 9.875473022460938, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4064779281616211, + "loss_total_v6": 0.6172170490026474, + "step": 1257 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29013333333333335, + "grad_norm": 23.111759185791016, + "learning_rate": 8.470498141151223e-06, + "loss": 10.678081512451172, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.45793794840574265, + "loss_total_v6": 0.6673801019787788, + "step": 1258 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29036396396396397, + "grad_norm": 20.953046798706055, + "learning_rate": 8.467763138292605e-06, + "loss": 10.538475036621094, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.44501635804772377, + "loss_total_v6": 0.6586546376347542, + "step": 1259 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2905945945945946, + "grad_norm": 25.035335540771484, + "learning_rate": 8.465026134705805e-06, + "loss": 10.520761489868164, + "loss_alignment": 0.41162109375, + "loss_alignment_w": 0.205810546875, + "loss_sft": 0.45222532376646996, + "loss_total_v6": 0.6575475931167603, + "step": 1260 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29082522522522525, + "grad_norm": 20.704254150390625, + "learning_rate": 8.462287131969933e-06, + "loss": 10.068954467773438, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.41367239877581596, + "loss_total_v6": 0.629309631884098, + "step": 1261 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29105585585585586, + "grad_norm": 24.529987335205078, + "learning_rate": 8.459546131665265e-06, + "loss": 10.348621368408203, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4292594790458679, + "loss_total_v6": 0.6467887833714485, + "step": 1262 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2912864864864865, + "grad_norm": 28.687911987304688, + "learning_rate": 8.456803135373217e-06, + "loss": 10.812878608703613, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.4583977162837982, + "loss_total_v6": 0.675804927945137, + "step": 1263 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29151711711711714, + "grad_norm": 21.59282684326172, + "learning_rate": 8.454058144676368e-06, + "loss": 9.539807319641113, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.3807075619697571, + "loss_total_v6": 0.5962379723787308, + "step": 1264 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29174774774774775, + "grad_norm": 21.453147888183594, + "learning_rate": 8.451311161158437e-06, + "loss": 10.034010887145996, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.42158982530236244, + "loss_total_v6": 0.6271256804466248, + "step": 1265 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29197837837837837, + "grad_norm": 22.800329208374023, + "learning_rate": 8.448562186404301e-06, + "loss": 10.164008140563965, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.4316067099571228, + "loss_total_v6": 0.6352505162358284, + "step": 1266 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29220900900900904, + "grad_norm": 21.037824630737305, + "learning_rate": 8.445811221999983e-06, + "loss": 10.831616401672363, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4650162011384964, + "loss_total_v6": 0.6769760251045227, + "step": 1267 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29243963963963965, + "grad_norm": 24.944488525390625, + "learning_rate": 8.44305826953265e-06, + "loss": 10.335139274597168, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4376484639942646, + "loss_total_v6": 0.6459461823105812, + "step": 1268 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29267027027027026, + "grad_norm": 21.620878219604492, + "learning_rate": 8.440303330590626e-06, + "loss": 10.127767562866211, + "loss_alignment": 0.40966796875, + "loss_alignment_w": 0.204833984375, + "loss_sft": 0.427831057459116, + "loss_total_v6": 0.6329854726791382, + "step": 1269 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2929009009009009, + "grad_norm": 21.275684356689453, + "learning_rate": 8.43754640676337e-06, + "loss": 10.544937133789062, + "loss_alignment": 0.443359375, + "loss_alignment_w": 0.2216796875, + "loss_sft": 0.4371042214334011, + "loss_total_v6": 0.6590585857629776, + "step": 1270 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29313153153153154, + "grad_norm": 20.83030128479004, + "learning_rate": 8.434787499641492e-06, + "loss": 10.640512466430664, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4529348537325859, + "loss_total_v6": 0.6650320217013359, + "step": 1271 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29336216216216215, + "grad_norm": 20.93461799621582, + "learning_rate": 8.432026610816745e-06, + "loss": 10.81638240814209, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4593490697443485, + "loss_total_v6": 0.6760238856077194, + "step": 1272 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29359279279279277, + "grad_norm": 22.563228607177734, + "learning_rate": 8.429263741882026e-06, + "loss": 10.021821975708008, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4137479215860367, + "loss_total_v6": 0.6263638809323311, + "step": 1273 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29382342342342344, + "grad_norm": 21.519264221191406, + "learning_rate": 8.426498894431375e-06, + "loss": 9.873859405517578, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.40770455822348595, + "loss_total_v6": 0.6171161830425262, + "step": 1274 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29405405405405405, + "grad_norm": 23.079715728759766, + "learning_rate": 8.423732070059972e-06, + "loss": 10.72777271270752, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4530022591352463, + "loss_total_v6": 0.6704857647418976, + "step": 1275 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29428468468468466, + "grad_norm": 21.209911346435547, + "learning_rate": 8.420963270364138e-06, + "loss": 10.101320266723633, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.42390450462698936, + "loss_total_v6": 0.6313325017690659, + "step": 1276 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29451531531531533, + "grad_norm": 22.0621280670166, + "learning_rate": 8.418192496941331e-06, + "loss": 10.437891006469727, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4387299381196499, + "loss_total_v6": 0.6523682251572609, + "step": 1277 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29474594594594594, + "grad_norm": 22.572263717651367, + "learning_rate": 8.415419751390155e-06, + "loss": 9.9025239944458, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.40717678517103195, + "loss_total_v6": 0.6189077645540237, + "step": 1278 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29497657657657655, + "grad_norm": 21.789045333862305, + "learning_rate": 8.412645035310345e-06, + "loss": 10.184358596801758, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.42474567517638206, + "loss_total_v6": 0.6365223973989487, + "step": 1279 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2952072072072072, + "grad_norm": 21.394508361816406, + "learning_rate": 8.409868350302774e-06, + "loss": 10.763971328735352, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.46585432067513466, + "loss_total_v6": 0.6727482080459595, + "step": 1280 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29543783783783784, + "grad_norm": 22.153940200805664, + "learning_rate": 8.407089697969458e-06, + "loss": 9.922977447509766, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.40131399035453796, + "loss_total_v6": 0.6201860457658768, + "step": 1281 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29566846846846845, + "grad_norm": 24.172651290893555, + "learning_rate": 8.404309079913533e-06, + "loss": 10.796138763427734, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.4571226052939892, + "loss_total_v6": 0.6747587323188782, + "step": 1282 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2958990990990991, + "grad_norm": 21.00592803955078, + "learning_rate": 8.401526497739288e-06, + "loss": 10.52684497833252, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.43739255145192146, + "loss_total_v6": 0.6579278260469437, + "step": 1283 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29612972972972973, + "grad_norm": 22.052303314208984, + "learning_rate": 8.398741953052128e-06, + "loss": 10.911977767944336, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.462592501193285, + "loss_total_v6": 0.681998610496521, + "step": 1284 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29636036036036034, + "grad_norm": 21.422086715698242, + "learning_rate": 8.395955447458601e-06, + "loss": 10.254073143005371, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.4345959760248661, + "loss_total_v6": 0.6408795565366745, + "step": 1285 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.296590990990991, + "grad_norm": 21.008514404296875, + "learning_rate": 8.393166982566384e-06, + "loss": 10.905622482299805, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4651554636657238, + "loss_total_v6": 0.6816013678908348, + "step": 1286 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2968216216216216, + "grad_norm": 22.0615177154541, + "learning_rate": 8.390376559984282e-06, + "loss": 10.007177352905273, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.4175933748483658, + "loss_total_v6": 0.625448577105999, + "step": 1287 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29705225225225224, + "grad_norm": 21.84102439880371, + "learning_rate": 8.387584181322233e-06, + "loss": 10.664379119873047, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.45564718171954155, + "loss_total_v6": 0.6665236502885818, + "step": 1288 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2972828828828829, + "grad_norm": 22.061317443847656, + "learning_rate": 8.3847898481913e-06, + "loss": 10.397269248962402, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.44259972497820854, + "loss_total_v6": 0.6498293578624725, + "step": 1289 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2975135135135135, + "grad_norm": 22.242197036743164, + "learning_rate": 8.381993562203675e-06, + "loss": 10.341934204101562, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.43741702288389206, + "loss_total_v6": 0.6463708654046059, + "step": 1290 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.29774414414414413, + "grad_norm": 25.016040802001953, + "learning_rate": 8.37919532497268e-06, + "loss": 10.865819931030273, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4639343358576298, + "loss_total_v6": 0.6791137531399727, + "step": 1291 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2979747747747748, + "grad_norm": 21.099124908447266, + "learning_rate": 8.376395138112755e-06, + "loss": 10.044984817504883, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.41646212711930275, + "loss_total_v6": 0.6278115957975388, + "step": 1292 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2982054054054054, + "grad_norm": 21.80368995666504, + "learning_rate": 8.373593003239474e-06, + "loss": 10.839438438415527, + "loss_alignment": 0.4404296875, + "loss_alignment_w": 0.22021484375, + "loss_sft": 0.45706696808338165, + "loss_total_v6": 0.6774649322032928, + "step": 1293 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.298436036036036, + "grad_norm": 21.671886444091797, + "learning_rate": 8.370788921969529e-06, + "loss": 10.205671310424805, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.429617740213871, + "loss_total_v6": 0.6378544345498085, + "step": 1294 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2986666666666667, + "grad_norm": 20.959012985229492, + "learning_rate": 8.367982895920734e-06, + "loss": 10.239409446716309, + "loss_alignment": 0.409912109375, + "loss_alignment_w": 0.2049560546875, + "loss_sft": 0.4349154941737652, + "loss_total_v6": 0.6399631202220917, + "step": 1295 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2988972972972973, + "grad_norm": 22.46050262451172, + "learning_rate": 8.365174926712032e-06, + "loss": 10.346611976623535, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.43433714285492897, + "loss_total_v6": 0.6466632261872292, + "step": 1296 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2991279279279279, + "grad_norm": 22.28083038330078, + "learning_rate": 8.362365015963478e-06, + "loss": 9.701690673828125, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.3952350653707981, + "loss_total_v6": 0.6063556745648384, + "step": 1297 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2993585585585586, + "grad_norm": 22.278034210205078, + "learning_rate": 8.359553165296257e-06, + "loss": 10.139037132263184, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.41532131284475327, + "loss_total_v6": 0.6336898356676102, + "step": 1298 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2995891891891892, + "grad_norm": 20.242504119873047, + "learning_rate": 8.356739376332665e-06, + "loss": 10.264839172363281, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4222989194095135, + "loss_total_v6": 0.6415524482727051, + "step": 1299 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.2998198198198198, + "grad_norm": 22.23550033569336, + "learning_rate": 8.353923650696119e-06, + "loss": 10.432086944580078, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.43662768229842186, + "loss_total_v6": 0.6520054861903191, + "step": 1300 + }, + { + "epoch": 0.2998198198198198, + "eval_loss": 0.6807858943939209, + "eval_loss_alignment": 0.40831580528846156, + "eval_loss_alignment_w": 0.20415790264423078, + "eval_loss_sft": 0.47662803436764795, + "eval_loss_total_v6": 0.6807859368845184, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.392, + "eval_samples_per_second": 22.648, + "eval_steps_per_second": 2.84, + "step": 1300 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3000504504504504, + "grad_norm": 25.72045135498047, + "learning_rate": 8.351105990011154e-06, + "loss": 9.931513786315918, + "loss_alignment": 0.413330078125, + "loss_alignment_w": 0.2066650390625, + "loss_sft": 0.4138561673462391, + "loss_total_v6": 0.6207195818424225, + "step": 1301 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3002810810810811, + "grad_norm": 22.95940399169922, + "learning_rate": 8.348286395903423e-06, + "loss": 9.993773460388184, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.41156763583421707, + "loss_total_v6": 0.6246108561754227, + "step": 1302 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3005117117117117, + "grad_norm": 21.017379760742188, + "learning_rate": 8.345464869999688e-06, + "loss": 10.193007469177246, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4236840419471264, + "loss_total_v6": 0.6370629593729973, + "step": 1303 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3007423423423423, + "grad_norm": 24.351953506469727, + "learning_rate": 8.342641413927836e-06, + "loss": 10.398481369018555, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4339779242873192, + "loss_total_v6": 0.6499050483107567, + "step": 1304 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.300972972972973, + "grad_norm": 21.457151412963867, + "learning_rate": 8.339816029316859e-06, + "loss": 10.29885482788086, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.4230210445821285, + "loss_total_v6": 0.6436783894896507, + "step": 1305 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3012036036036036, + "grad_norm": 19.685976028442383, + "learning_rate": 8.336988717796865e-06, + "loss": 10.36474609375, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.4342193976044655, + "loss_total_v6": 0.6477966457605362, + "step": 1306 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3014342342342342, + "grad_norm": 20.853839874267578, + "learning_rate": 8.33415948099907e-06, + "loss": 10.399768829345703, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.43195270374417305, + "loss_total_v6": 0.6499855294823647, + "step": 1307 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3016648648648649, + "grad_norm": 20.772554397583008, + "learning_rate": 8.331328320555812e-06, + "loss": 10.486833572387695, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.44195661693811417, + "loss_total_v6": 0.6554270833730698, + "step": 1308 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3018954954954955, + "grad_norm": 22.689382553100586, + "learning_rate": 8.328495238100522e-06, + "loss": 10.367120742797852, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.43490180745720863, + "loss_total_v6": 0.6479449942708015, + "step": 1309 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3021261261261261, + "grad_norm": 22.191919326782227, + "learning_rate": 8.325660235267753e-06, + "loss": 10.338577270507812, + "loss_alignment": 0.43896484375, + "loss_alignment_w": 0.219482421875, + "loss_sft": 0.426983792334795, + "loss_total_v6": 0.6461610496044159, + "step": 1310 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3023567567567568, + "grad_norm": 21.22939682006836, + "learning_rate": 8.322823313693162e-06, + "loss": 10.310771942138672, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.4342334307730198, + "loss_total_v6": 0.6444232612848282, + "step": 1311 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3025873873873874, + "grad_norm": 21.744060516357422, + "learning_rate": 8.319984475013512e-06, + "loss": 10.224162101745605, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.42494460940361023, + "loss_total_v6": 0.6390101537108421, + "step": 1312 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.302818018018018, + "grad_norm": 19.81127166748047, + "learning_rate": 8.317143720866671e-06, + "loss": 10.28388500213623, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4325530007481575, + "loss_total_v6": 0.642742820084095, + "step": 1313 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30304864864864867, + "grad_norm": 22.072704315185547, + "learning_rate": 8.314301052891618e-06, + "loss": 10.130780220031738, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.41329457610845566, + "loss_total_v6": 0.6331737339496613, + "step": 1314 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3032792792792793, + "grad_norm": 21.524694442749023, + "learning_rate": 8.31145647272843e-06, + "loss": 10.003887176513672, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4089496210217476, + "loss_total_v6": 0.6252429634332657, + "step": 1315 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3035099099099099, + "grad_norm": 24.41596221923828, + "learning_rate": 8.308609982018286e-06, + "loss": 10.861225128173828, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.46675989776849747, + "loss_total_v6": 0.6788265854120255, + "step": 1316 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30374054054054056, + "grad_norm": 20.972970962524414, + "learning_rate": 8.305761582403477e-06, + "loss": 10.002447128295898, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.40478555858135223, + "loss_total_v6": 0.6251529678702354, + "step": 1317 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3039711711711712, + "grad_norm": 21.617172241210938, + "learning_rate": 8.302911275527385e-06, + "loss": 10.566962242126465, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.44623224809765816, + "loss_total_v6": 0.6604351103305817, + "step": 1318 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3042018018018018, + "grad_norm": 21.883365631103516, + "learning_rate": 8.300059063034498e-06, + "loss": 10.32753849029541, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4320922680199146, + "loss_total_v6": 0.6454711854457855, + "step": 1319 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30443243243243245, + "grad_norm": 24.061351776123047, + "learning_rate": 8.297204946570398e-06, + "loss": 10.855644226074219, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.46300842240452766, + "loss_total_v6": 0.6784777864813805, + "step": 1320 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30466306306306307, + "grad_norm": 21.639785766601562, + "learning_rate": 8.294348927781776e-06, + "loss": 11.028528213500977, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.48164135962724686, + "loss_total_v6": 0.6892829909920692, + "step": 1321 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3048936936936937, + "grad_norm": 20.71481704711914, + "learning_rate": 8.291491008316409e-06, + "loss": 10.245372772216797, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.4289710372686386, + "loss_total_v6": 0.640335775911808, + "step": 1322 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30512432432432435, + "grad_norm": 22.999073028564453, + "learning_rate": 8.288631189823177e-06, + "loss": 10.750015258789062, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.453568410128355, + "loss_total_v6": 0.6718759313225746, + "step": 1323 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30535495495495496, + "grad_norm": 21.093061447143555, + "learning_rate": 8.285769473952052e-06, + "loss": 10.586729049682617, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4483678974211216, + "loss_total_v6": 0.6616705060005188, + "step": 1324 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3055855855855856, + "grad_norm": 23.500581741333008, + "learning_rate": 8.282905862354107e-06, + "loss": 10.592741012573242, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.4483927898108959, + "loss_total_v6": 0.6620463281869888, + "step": 1325 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30581621621621624, + "grad_norm": 20.35346031188965, + "learning_rate": 8.280040356681503e-06, + "loss": 10.154809951782227, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4216323718428612, + "loss_total_v6": 0.6346755847334862, + "step": 1326 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30604684684684685, + "grad_norm": 21.294193267822266, + "learning_rate": 8.277172958587494e-06, + "loss": 10.316625595092773, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.43148649483919144, + "loss_total_v6": 0.6447891071438789, + "step": 1327 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30627747747747747, + "grad_norm": 23.325523376464844, + "learning_rate": 8.274303669726427e-06, + "loss": 9.785473823547363, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4006851427257061, + "loss_total_v6": 0.6115921065211296, + "step": 1328 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30650810810810813, + "grad_norm": 21.109935760498047, + "learning_rate": 8.271432491753739e-06, + "loss": 9.834564208984375, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.40508074313402176, + "loss_total_v6": 0.6146602183580399, + "step": 1329 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30673873873873875, + "grad_norm": 19.282848358154297, + "learning_rate": 8.268559426325964e-06, + "loss": 10.198341369628906, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.42861028388142586, + "loss_total_v6": 0.6373962908983231, + "step": 1330 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30696936936936936, + "grad_norm": 20.785558700561523, + "learning_rate": 8.265684475100713e-06, + "loss": 10.517009735107422, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4460551142692566, + "loss_total_v6": 0.6573130637407303, + "step": 1331 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3072, + "grad_norm": 19.976224899291992, + "learning_rate": 8.262807639736693e-06, + "loss": 10.059308052062988, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4146564230322838, + "loss_total_v6": 0.6287067085504532, + "step": 1332 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30743063063063064, + "grad_norm": 26.42735481262207, + "learning_rate": 8.259928921893694e-06, + "loss": 10.155860900878906, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.42145396769046783, + "loss_total_v6": 0.6347413063049316, + "step": 1333 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30766126126126125, + "grad_norm": 21.496957778930664, + "learning_rate": 8.257048323232598e-06, + "loss": 9.910063743591309, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.40426062047481537, + "loss_total_v6": 0.6193790249526501, + "step": 1334 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30789189189189187, + "grad_norm": 21.05655288696289, + "learning_rate": 8.254165845415366e-06, + "loss": 9.938501358032227, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.41327057778835297, + "loss_total_v6": 0.621156319975853, + "step": 1335 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30812252252252254, + "grad_norm": 21.742212295532227, + "learning_rate": 8.251281490105045e-06, + "loss": 9.91279411315918, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4106263145804405, + "loss_total_v6": 0.6195496544241905, + "step": 1336 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30835315315315315, + "grad_norm": 22.809757232666016, + "learning_rate": 8.248395258965767e-06, + "loss": 10.59210205078125, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.44177626818418503, + "loss_total_v6": 0.6620063483715057, + "step": 1337 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30858378378378376, + "grad_norm": 21.29993438720703, + "learning_rate": 8.245507153662744e-06, + "loss": 10.063684463500977, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4173409231007099, + "loss_total_v6": 0.6289803236722946, + "step": 1338 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30881441441441443, + "grad_norm": 23.345294952392578, + "learning_rate": 8.242617175862271e-06, + "loss": 10.156089782714844, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4224143214523792, + "loss_total_v6": 0.6347556188702583, + "step": 1339 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30904504504504504, + "grad_norm": 21.502647399902344, + "learning_rate": 8.239725327231721e-06, + "loss": 10.641866683959961, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.44575629383325577, + "loss_total_v6": 0.6651166379451752, + "step": 1340 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30927567567567565, + "grad_norm": 20.805126190185547, + "learning_rate": 8.23683160943955e-06, + "loss": 10.345718383789062, + "loss_alignment": 0.40625, + "loss_alignment_w": 0.203125, + "loss_sft": 0.4437418393790722, + "loss_total_v6": 0.6466074287891388, + "step": 1341 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3095063063063063, + "grad_norm": 24.700178146362305, + "learning_rate": 8.233936024155289e-06, + "loss": 9.957019805908203, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4076683856546879, + "loss_total_v6": 0.6223137527704239, + "step": 1342 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30973693693693694, + "grad_norm": 21.934925079345703, + "learning_rate": 8.231038573049548e-06, + "loss": 10.91876220703125, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.46265023574233055, + "loss_total_v6": 0.682422585785389, + "step": 1343 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.30996756756756755, + "grad_norm": 23.107107162475586, + "learning_rate": 8.228139257794012e-06, + "loss": 10.547351837158203, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.4486992172896862, + "loss_total_v6": 0.6592094749212265, + "step": 1344 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3101981981981982, + "grad_norm": 22.669414520263672, + "learning_rate": 8.225238080061447e-06, + "loss": 10.278738021850586, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.4345659092068672, + "loss_total_v6": 0.6424211114645004, + "step": 1345 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31042882882882883, + "grad_norm": 20.900117874145508, + "learning_rate": 8.222335041525683e-06, + "loss": 10.407495498657227, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.43678437918424606, + "loss_total_v6": 0.6504684537649155, + "step": 1346 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31065945945945944, + "grad_norm": 23.081838607788086, + "learning_rate": 8.219430143861638e-06, + "loss": 10.428290367126465, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.43513911217451096, + "loss_total_v6": 0.6517681255936623, + "step": 1347 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3108900900900901, + "grad_norm": 22.586278915405273, + "learning_rate": 8.216523388745288e-06, + "loss": 10.111522674560547, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.42265014350414276, + "loss_total_v6": 0.6319702044129372, + "step": 1348 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3111207207207207, + "grad_norm": 21.056930541992188, + "learning_rate": 8.213614777853688e-06, + "loss": 9.877527236938477, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4013878405094147, + "loss_total_v6": 0.6173455119132996, + "step": 1349 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31135135135135134, + "grad_norm": 25.051958084106445, + "learning_rate": 8.210704312864965e-06, + "loss": 10.468639373779297, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.44599219411611557, + "loss_total_v6": 0.6542899385094643, + "step": 1350 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.311581981981982, + "grad_norm": 21.049118041992188, + "learning_rate": 8.207791995458314e-06, + "loss": 10.38583755493164, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.43987104669213295, + "loss_total_v6": 0.6491148099303246, + "step": 1351 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3118126126126126, + "grad_norm": 20.37079429626465, + "learning_rate": 8.204877827313997e-06, + "loss": 10.079437255859375, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.42433736473321915, + "loss_total_v6": 0.6299648210406303, + "step": 1352 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31204324324324323, + "grad_norm": 21.952632904052734, + "learning_rate": 8.201961810113346e-06, + "loss": 10.279329299926758, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.43103231862187386, + "loss_total_v6": 0.642458088696003, + "step": 1353 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3122738738738739, + "grad_norm": 21.676101684570312, + "learning_rate": 8.199043945538759e-06, + "loss": 10.082784652709961, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.4112866371870041, + "loss_total_v6": 0.630173996090889, + "step": 1354 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3125045045045045, + "grad_norm": 20.106821060180664, + "learning_rate": 8.196124235273698e-06, + "loss": 10.647245407104492, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.44821348786354065, + "loss_total_v6": 0.665452852845192, + "step": 1355 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3127351351351351, + "grad_norm": 21.308847427368164, + "learning_rate": 8.193202681002692e-06, + "loss": 10.455105781555176, + "loss_alignment": 0.410400390625, + "loss_alignment_w": 0.2052001953125, + "loss_sft": 0.4485338367521763, + "loss_total_v6": 0.6534441113471985, + "step": 1356 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3129657657657658, + "grad_norm": 21.323347091674805, + "learning_rate": 8.190279284411337e-06, + "loss": 10.216968536376953, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4256851300597191, + "loss_total_v6": 0.6385605111718178, + "step": 1357 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3131963963963964, + "grad_norm": 22.5921573638916, + "learning_rate": 8.187354047186286e-06, + "loss": 10.00328254699707, + "loss_alignment": 0.40869140625, + "loss_alignment_w": 0.204345703125, + "loss_sft": 0.4210273027420044, + "loss_total_v6": 0.6252051889896393, + "step": 1358 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.313427027027027, + "grad_norm": 22.224849700927734, + "learning_rate": 8.184426971015257e-06, + "loss": 10.11782455444336, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.42235730215907097, + "loss_total_v6": 0.6323640123009682, + "step": 1359 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3136576576576577, + "grad_norm": 25.96402931213379, + "learning_rate": 8.181498057587028e-06, + "loss": 10.699737548828125, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.44981570541858673, + "loss_total_v6": 0.6687335669994354, + "step": 1360 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3138882882882883, + "grad_norm": 20.79082489013672, + "learning_rate": 8.178567308591438e-06, + "loss": 10.077136993408203, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4123070128262043, + "loss_total_v6": 0.6298210471868515, + "step": 1361 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3141189189189189, + "grad_norm": 24.9197940826416, + "learning_rate": 8.175634725719384e-06, + "loss": 10.507291793823242, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.44042760506272316, + "loss_total_v6": 0.6567056998610497, + "step": 1362 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3143495495495495, + "grad_norm": 23.68704605102539, + "learning_rate": 8.17270031066282e-06, + "loss": 10.26783561706543, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.43525778874754906, + "loss_total_v6": 0.64173973351717, + "step": 1363 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3145801801801802, + "grad_norm": 22.138608932495117, + "learning_rate": 8.169764065114764e-06, + "loss": 10.841874122619629, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4628191776573658, + "loss_total_v6": 0.6776171401143074, + "step": 1364 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3148108108108108, + "grad_norm": 25.08181381225586, + "learning_rate": 8.166825990769281e-06, + "loss": 11.468951225280762, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.5001651830971241, + "loss_total_v6": 0.7168094515800476, + "step": 1365 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3150414414414414, + "grad_norm": 23.603092193603516, + "learning_rate": 8.163886089321493e-06, + "loss": 10.122861862182617, + "loss_alignment": 0.41162109375, + "loss_alignment_w": 0.205810546875, + "loss_sft": 0.4268072471022606, + "loss_total_v6": 0.6326788365840912, + "step": 1366 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3152720720720721, + "grad_norm": 22.363576889038086, + "learning_rate": 8.160944362467582e-06, + "loss": 10.163679122924805, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.4193028286099434, + "loss_total_v6": 0.6352299526333809, + "step": 1367 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3155027027027027, + "grad_norm": 21.381675720214844, + "learning_rate": 8.158000811904779e-06, + "loss": 10.037666320800781, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.4190259724855423, + "loss_total_v6": 0.6273542046546936, + "step": 1368 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3157333333333333, + "grad_norm": 20.797103881835938, + "learning_rate": 8.155055439331363e-06, + "loss": 10.312861442565918, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.43320437148213387, + "loss_total_v6": 0.6445538550615311, + "step": 1369 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.315963963963964, + "grad_norm": 20.95636749267578, + "learning_rate": 8.152108246446673e-06, + "loss": 10.628673553466797, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4541480429470539, + "loss_total_v6": 0.664292111992836, + "step": 1370 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3161945945945946, + "grad_norm": 20.861738204956055, + "learning_rate": 8.149159234951091e-06, + "loss": 10.39337158203125, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4344368577003479, + "loss_total_v6": 0.6495857834815979, + "step": 1371 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3164252252252252, + "grad_norm": 20.52388572692871, + "learning_rate": 8.146208406546054e-06, + "loss": 9.948982238769531, + "loss_alignment": 0.396240234375, + "loss_alignment_w": 0.1981201171875, + "loss_sft": 0.4236454702913761, + "loss_total_v6": 0.6218113675713539, + "step": 1372 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3166558558558559, + "grad_norm": 21.874094009399414, + "learning_rate": 8.143255762934042e-06, + "loss": 10.286820411682129, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.426938071846962, + "loss_total_v6": 0.6429262533783913, + "step": 1373 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3168864864864865, + "grad_norm": 22.14950942993164, + "learning_rate": 8.140301305818586e-06, + "loss": 9.995756149291992, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4074495807290077, + "loss_total_v6": 0.6247347444295883, + "step": 1374 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3171171171171171, + "grad_norm": 21.164281845092773, + "learning_rate": 8.13734503690426e-06, + "loss": 10.45085334777832, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4365492984652519, + "loss_total_v6": 0.6531783416867256, + "step": 1375 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31734774774774777, + "grad_norm": 21.372987747192383, + "learning_rate": 8.134386957896688e-06, + "loss": 10.156486511230469, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.4296717010438442, + "loss_total_v6": 0.6347803622484207, + "step": 1376 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3175783783783784, + "grad_norm": 21.545547485351562, + "learning_rate": 8.131427070502535e-06, + "loss": 10.236600875854492, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.41943538933992386, + "loss_total_v6": 0.6397875621914864, + "step": 1377 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.317809009009009, + "grad_norm": 22.329065322875977, + "learning_rate": 8.128465376429507e-06, + "loss": 10.25185775756836, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.42197583243250847, + "loss_total_v6": 0.6407411023974419, + "step": 1378 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31803963963963966, + "grad_norm": 21.578475952148438, + "learning_rate": 8.125501877386359e-06, + "loss": 10.447980880737305, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.44404491409659386, + "loss_total_v6": 0.6529987826943398, + "step": 1379 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3182702702702703, + "grad_norm": 21.22339630126953, + "learning_rate": 8.122536575082883e-06, + "loss": 9.775216102600098, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4019056223332882, + "loss_total_v6": 0.6109510287642479, + "step": 1380 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3185009009009009, + "grad_norm": 21.324451446533203, + "learning_rate": 8.11956947122991e-06, + "loss": 10.666378021240234, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.45725223794579506, + "loss_total_v6": 0.666648618876934, + "step": 1381 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31873153153153155, + "grad_norm": 23.85001564025879, + "learning_rate": 8.116600567539313e-06, + "loss": 10.369257926940918, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.4359967112541199, + "loss_total_v6": 0.648078627884388, + "step": 1382 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31896216216216217, + "grad_norm": 23.99997901916504, + "learning_rate": 8.113629865724006e-06, + "loss": 10.150480270385742, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4155329763889313, + "loss_total_v6": 0.6344050392508507, + "step": 1383 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3191927927927928, + "grad_norm": 22.40214729309082, + "learning_rate": 8.110657367497934e-06, + "loss": 10.517889976501465, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.44205138459801674, + "loss_total_v6": 0.6573681384325027, + "step": 1384 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31942342342342345, + "grad_norm": 22.111162185668945, + "learning_rate": 8.107683074576081e-06, + "loss": 10.143692016601562, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.42173098400235176, + "loss_total_v6": 0.6339807510375977, + "step": 1385 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.31965405405405406, + "grad_norm": 20.910179138183594, + "learning_rate": 8.104706988674471e-06, + "loss": 10.49683952331543, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.44610682502388954, + "loss_total_v6": 0.6560525149106979, + "step": 1386 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3198846846846847, + "grad_norm": 21.890806198120117, + "learning_rate": 8.101729111510158e-06, + "loss": 9.750555992126465, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.39906736090779305, + "loss_total_v6": 0.609409749507904, + "step": 1387 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32011531531531534, + "grad_norm": 24.40160369873047, + "learning_rate": 8.098749444801226e-06, + "loss": 10.644550323486328, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4531566649675369, + "loss_total_v6": 0.6652843728661537, + "step": 1388 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32034594594594595, + "grad_norm": 23.442575454711914, + "learning_rate": 8.095767990266799e-06, + "loss": 10.711739540100098, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.4526257738471031, + "loss_total_v6": 0.6694836989045143, + "step": 1389 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32057657657657657, + "grad_norm": 20.046354293823242, + "learning_rate": 8.09278474962703e-06, + "loss": 10.177022933959961, + "loss_alignment": 0.411376953125, + "loss_alignment_w": 0.2056884765625, + "loss_sft": 0.43005502596497536, + "loss_total_v6": 0.636063925921917, + "step": 1390 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32080720720720723, + "grad_norm": 22.336933135986328, + "learning_rate": 8.089799724603099e-06, + "loss": 10.481403350830078, + "loss_alignment": 0.44482421875, + "loss_alignment_w": 0.222412109375, + "loss_sft": 0.4331791028380394, + "loss_total_v6": 0.6550876945257187, + "step": 1391 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32103783783783785, + "grad_norm": 22.677518844604492, + "learning_rate": 8.086812916917219e-06, + "loss": 10.681516647338867, + "loss_alignment": 0.444091796875, + "loss_alignment_w": 0.2220458984375, + "loss_sft": 0.445304773747921, + "loss_total_v6": 0.6675947979092598, + "step": 1392 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32126846846846846, + "grad_norm": 22.04762077331543, + "learning_rate": 8.083824328292631e-06, + "loss": 10.200611114501953, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.428691141307354, + "loss_total_v6": 0.6375381946563721, + "step": 1393 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3214990990990991, + "grad_norm": 20.587636947631836, + "learning_rate": 8.080833960453604e-06, + "loss": 10.654653549194336, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4485238902270794, + "loss_total_v6": 0.665915846824646, + "step": 1394 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32172972972972974, + "grad_norm": 21.020132064819336, + "learning_rate": 8.077841815125431e-06, + "loss": 10.299152374267578, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.43521615862846375, + "loss_total_v6": 0.6436970084905624, + "step": 1395 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32196036036036035, + "grad_norm": 22.01547622680664, + "learning_rate": 8.074847894034434e-06, + "loss": 10.239627838134766, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.42879506945610046, + "loss_total_v6": 0.6399767026305199, + "step": 1396 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32219099099099097, + "grad_norm": 23.659021377563477, + "learning_rate": 8.071852198907956e-06, + "loss": 10.7667875289917, + "loss_alignment": 0.442626953125, + "loss_alignment_w": 0.2213134765625, + "loss_sft": 0.4516870193183422, + "loss_total_v6": 0.6729241982102394, + "step": 1397 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32242162162162163, + "grad_norm": 22.464012145996094, + "learning_rate": 8.068854731474366e-06, + "loss": 9.98515510559082, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.41539300605654716, + "loss_total_v6": 0.6240722164511681, + "step": 1398 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32265225225225225, + "grad_norm": 21.7099666595459, + "learning_rate": 8.065855493463055e-06, + "loss": 10.24728012084961, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.42104892060160637, + "loss_total_v6": 0.640455037355423, + "step": 1399 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32288288288288286, + "grad_norm": 22.04937171936035, + "learning_rate": 8.062854486604434e-06, + "loss": 10.673685073852539, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4538942761719227, + "loss_total_v6": 0.6671053245663643, + "step": 1400 + }, + { + "epoch": 0.32288288288288286, + "eval_loss": 0.6772116422653198, + "eval_loss_alignment": 0.40745400974893164, + "eval_loss_alignment_w": 0.20372700487446582, + "eval_loss_sft": 0.4734846472294412, + "eval_loss_total_v6": 0.677211652087987, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.321, + "eval_samples_per_second": 22.667, + "eval_steps_per_second": 2.843, + "step": 1400 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32311351351351353, + "grad_norm": 20.557323455810547, + "learning_rate": 8.059851712629937e-06, + "loss": 10.17697525024414, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4227125495672226, + "loss_total_v6": 0.6360609084367752, + "step": 1401 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32334414414414414, + "grad_norm": 19.967193603515625, + "learning_rate": 8.056847173272017e-06, + "loss": 9.990804672241211, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.4049581289291382, + "loss_total_v6": 0.6244252845644951, + "step": 1402 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32357477477477475, + "grad_norm": 21.89307975769043, + "learning_rate": 8.053840870264144e-06, + "loss": 9.758686065673828, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.38849756866693497, + "loss_total_v6": 0.6099178344011307, + "step": 1403 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3238054054054054, + "grad_norm": 21.81832504272461, + "learning_rate": 8.050832805340807e-06, + "loss": 10.216204643249512, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.4333431050181389, + "loss_total_v6": 0.6385127976536751, + "step": 1404 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32403603603603603, + "grad_norm": 23.274059295654297, + "learning_rate": 8.047822980237509e-06, + "loss": 10.4610595703125, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.43824004009366035, + "loss_total_v6": 0.6538162231445312, + "step": 1405 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32426666666666665, + "grad_norm": 21.718143463134766, + "learning_rate": 8.044811396690772e-06, + "loss": 10.570817947387695, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4479075260460377, + "loss_total_v6": 0.660676084458828, + "step": 1406 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3244972972972973, + "grad_norm": 20.685895919799805, + "learning_rate": 8.041798056438133e-06, + "loss": 10.081768035888672, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4174639917910099, + "loss_total_v6": 0.6301104873418808, + "step": 1407 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32472792792792793, + "grad_norm": 19.904836654663086, + "learning_rate": 8.038782961218136e-06, + "loss": 9.859735488891602, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4049754850566387, + "loss_total_v6": 0.6162334457039833, + "step": 1408 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32495855855855854, + "grad_norm": 26.56890296936035, + "learning_rate": 8.035766112770349e-06, + "loss": 10.639041900634766, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.45130184292793274, + "loss_total_v6": 0.6649401262402534, + "step": 1409 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3251891891891892, + "grad_norm": 22.159727096557617, + "learning_rate": 8.032747512835338e-06, + "loss": 10.264532089233398, + "loss_alignment": 0.41064453125, + "loss_alignment_w": 0.205322265625, + "loss_sft": 0.4360583648085594, + "loss_total_v6": 0.6415332332253456, + "step": 1410 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3254198198198198, + "grad_norm": 21.34070587158203, + "learning_rate": 8.029727163154691e-06, + "loss": 9.870135307312012, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.4048321172595024, + "loss_total_v6": 0.6168835014104843, + "step": 1411 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32565045045045043, + "grad_norm": 22.119321823120117, + "learning_rate": 8.026705065470996e-06, + "loss": 9.63797378540039, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.3881399631500244, + "loss_total_v6": 0.6023733541369438, + "step": 1412 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3258810810810811, + "grad_norm": 23.3697452545166, + "learning_rate": 8.023681221527858e-06, + "loss": 10.402217864990234, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.44284793734550476, + "loss_total_v6": 0.6501385942101479, + "step": 1413 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3261117117117117, + "grad_norm": 21.60983657836914, + "learning_rate": 8.020655633069883e-06, + "loss": 10.274859428405762, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4323246106505394, + "loss_total_v6": 0.6421787217259407, + "step": 1414 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.32634234234234233, + "grad_norm": 22.030405044555664, + "learning_rate": 8.017628301842689e-06, + "loss": 10.16366958618164, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4162962883710861, + "loss_total_v6": 0.6352293863892555, + "step": 1415 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.326572972972973, + "grad_norm": 21.725549697875977, + "learning_rate": 8.014599229592895e-06, + "loss": 9.481603622436523, + "loss_alignment": 0.404052734375, + "loss_alignment_w": 0.2020263671875, + "loss_sft": 0.3908638060092926, + "loss_total_v6": 0.5926002711057663, + "step": 1416 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3268036036036036, + "grad_norm": 23.262882232666016, + "learning_rate": 8.011568418068124e-06, + "loss": 10.247818946838379, + "loss_alignment": 0.404296875, + "loss_alignment_w": 0.2021484375, + "loss_sft": 0.43840131536126137, + "loss_total_v6": 0.6404887214303017, + "step": 1417 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3270342342342342, + "grad_norm": 21.78384780883789, + "learning_rate": 8.008535869017012e-06, + "loss": 10.733499526977539, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.456564474850893, + "loss_total_v6": 0.6708436533808708, + "step": 1418 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3272648648648649, + "grad_norm": 22.947595596313477, + "learning_rate": 8.005501584189184e-06, + "loss": 10.226289749145508, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.41950807720422745, + "loss_total_v6": 0.6391431167721748, + "step": 1419 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3274954954954955, + "grad_norm": 21.99810028076172, + "learning_rate": 8.002465565335272e-06, + "loss": 9.494497299194336, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.38213296979665756, + "loss_total_v6": 0.5934061333537102, + "step": 1420 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3277261261261261, + "grad_norm": 21.803144454956055, + "learning_rate": 7.999427814206911e-06, + "loss": 10.03863525390625, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4158058166503906, + "loss_total_v6": 0.6274147108197212, + "step": 1421 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3279567567567568, + "grad_norm": 21.701358795166016, + "learning_rate": 7.996388332556735e-06, + "loss": 10.56974983215332, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.44211872667074203, + "loss_total_v6": 0.6606093123555183, + "step": 1422 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3281873873873874, + "grad_norm": 22.094654083251953, + "learning_rate": 7.993347122138374e-06, + "loss": 10.716385841369629, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4524431675672531, + "loss_total_v6": 0.6697741001844406, + "step": 1423 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.328418018018018, + "grad_norm": 21.983539581298828, + "learning_rate": 7.990304184706454e-06, + "loss": 10.329154968261719, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4312319792807102, + "loss_total_v6": 0.645572192966938, + "step": 1424 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3286486486486486, + "grad_norm": 22.30327606201172, + "learning_rate": 7.987259522016602e-06, + "loss": 10.343546867370605, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.43120066821575165, + "loss_total_v6": 0.6464716643095016, + "step": 1425 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3288792792792793, + "grad_norm": 22.79974937438965, + "learning_rate": 7.984213135825437e-06, + "loss": 10.014215469360352, + "loss_alignment": 0.410400390625, + "loss_alignment_w": 0.2052001953125, + "loss_sft": 0.4208713732659817, + "loss_total_v6": 0.6258884966373444, + "step": 1426 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3291099099099099, + "grad_norm": 21.991004943847656, + "learning_rate": 7.981165027890576e-06, + "loss": 10.387121200561523, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.43647222220897675, + "loss_total_v6": 0.6491950303316116, + "step": 1427 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3293405405405405, + "grad_norm": 24.374420166015625, + "learning_rate": 7.978115199970621e-06, + "loss": 9.775879859924316, + "loss_alignment": 0.404296875, + "loss_alignment_w": 0.2021484375, + "loss_sft": 0.4085236042737961, + "loss_total_v6": 0.6109924763441086, + "step": 1428 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3295711711711712, + "grad_norm": 20.23822021484375, + "learning_rate": 7.975063653825179e-06, + "loss": 9.724998474121094, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.3930906765162945, + "loss_total_v6": 0.6078123673796654, + "step": 1429 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3298018018018018, + "grad_norm": 20.66507911682129, + "learning_rate": 7.972010391214837e-06, + "loss": 10.146896362304688, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.4268445856869221, + "loss_total_v6": 0.6341810077428818, + "step": 1430 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3300324324324324, + "grad_norm": 21.33282470703125, + "learning_rate": 7.96895541390118e-06, + "loss": 10.189593315124512, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.42839929461479187, + "loss_total_v6": 0.6368495896458626, + "step": 1431 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3302630630630631, + "grad_norm": 22.348403930664062, + "learning_rate": 7.965898723646777e-06, + "loss": 10.396020889282227, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.43339694291353226, + "loss_total_v6": 0.6497513130307198, + "step": 1432 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3304936936936937, + "grad_norm": 23.679716110229492, + "learning_rate": 7.962840322215188e-06, + "loss": 10.323010444641113, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.42320332303643227, + "loss_total_v6": 0.6451881527900696, + "step": 1433 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3307243243243243, + "grad_norm": 22.807653427124023, + "learning_rate": 7.959780211370958e-06, + "loss": 10.040895462036133, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.416572704911232, + "loss_total_v6": 0.6275559961795807, + "step": 1434 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33095495495495497, + "grad_norm": 21.327350616455078, + "learning_rate": 7.956718392879623e-06, + "loss": 10.095146179199219, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4160418398678303, + "loss_total_v6": 0.6309466138482094, + "step": 1435 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3311855855855856, + "grad_norm": 22.158802032470703, + "learning_rate": 7.9536548685077e-06, + "loss": 10.395135879516602, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.4394909329712391, + "loss_total_v6": 0.6496959999203682, + "step": 1436 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3314162162162162, + "grad_norm": 21.825265884399414, + "learning_rate": 7.95058964002269e-06, + "loss": 10.165576934814453, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.41862795501947403, + "loss_total_v6": 0.6353485584259033, + "step": 1437 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33164684684684687, + "grad_norm": 20.1494197845459, + "learning_rate": 7.94752270919308e-06, + "loss": 10.319743156433105, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4268748536705971, + "loss_total_v6": 0.6449839621782303, + "step": 1438 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3318774774774775, + "grad_norm": 22.887659072875977, + "learning_rate": 7.944454077788339e-06, + "loss": 9.98080062866211, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.41834041476249695, + "loss_total_v6": 0.6238000392913818, + "step": 1439 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3321081081081081, + "grad_norm": 22.546789169311523, + "learning_rate": 7.941383747578913e-06, + "loss": 10.834089279174805, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4611119106411934, + "loss_total_v6": 0.6771306172013283, + "step": 1440 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33233873873873876, + "grad_norm": 20.68046760559082, + "learning_rate": 7.93831172033623e-06, + "loss": 9.630773544311523, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.3892921209335327, + "loss_total_v6": 0.6019233614206314, + "step": 1441 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33256936936936937, + "grad_norm": 21.111125946044922, + "learning_rate": 7.935237997832704e-06, + "loss": 10.287525177001953, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.42464761435985565, + "loss_total_v6": 0.6429703533649445, + "step": 1442 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3328, + "grad_norm": 22.946548461914062, + "learning_rate": 7.932162581841715e-06, + "loss": 10.192630767822266, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.429474126547575, + "loss_total_v6": 0.6370394453406334, + "step": 1443 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33303063063063065, + "grad_norm": 21.257551193237305, + "learning_rate": 7.92908547413763e-06, + "loss": 10.31899642944336, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.43395401909947395, + "loss_total_v6": 0.6449372917413712, + "step": 1444 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33326126126126127, + "grad_norm": 22.881837844848633, + "learning_rate": 7.926006676495787e-06, + "loss": 10.364872932434082, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.43584468215703964, + "loss_total_v6": 0.6478045433759689, + "step": 1445 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3334918918918919, + "grad_norm": 21.58852195739746, + "learning_rate": 7.9229261906925e-06, + "loss": 10.169281005859375, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4207058362662792, + "loss_total_v6": 0.6355800777673721, + "step": 1446 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33372252252252255, + "grad_norm": 23.40359878540039, + "learning_rate": 7.919844018505058e-06, + "loss": 10.121419906616211, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4173939935863018, + "loss_total_v6": 0.6325886845588684, + "step": 1447 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33395315315315316, + "grad_norm": 21.420494079589844, + "learning_rate": 7.916760161711721e-06, + "loss": 10.291566848754883, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.4379769191145897, + "loss_total_v6": 0.6432228907942772, + "step": 1448 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33418378378378377, + "grad_norm": 19.712265014648438, + "learning_rate": 7.913674622091722e-06, + "loss": 10.389113426208496, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.433957040309906, + "loss_total_v6": 0.649319589138031, + "step": 1449 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33441441441441444, + "grad_norm": 19.07480239868164, + "learning_rate": 7.91058740142527e-06, + "loss": 9.687280654907227, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.39576875418424606, + "loss_total_v6": 0.6054550632834435, + "step": 1450 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33464504504504505, + "grad_norm": 21.406511306762695, + "learning_rate": 7.907498501493532e-06, + "loss": 9.872303009033203, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4084618389606476, + "loss_total_v6": 0.617018960416317, + "step": 1451 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33487567567567567, + "grad_norm": 20.906936645507812, + "learning_rate": 7.904407924078653e-06, + "loss": 10.037687301635742, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4116419516503811, + "loss_total_v6": 0.6273554638028145, + "step": 1452 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3351063063063063, + "grad_norm": 21.74577522277832, + "learning_rate": 7.901315670963747e-06, + "loss": 10.538419723510742, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4499567225575447, + "loss_total_v6": 0.6586512252688408, + "step": 1453 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33533693693693695, + "grad_norm": 21.508148193359375, + "learning_rate": 7.898221743932887e-06, + "loss": 10.310297966003418, + "loss_alignment": 0.406005859375, + "loss_alignment_w": 0.2030029296875, + "loss_sft": 0.4410244897007942, + "loss_total_v6": 0.6443936228752136, + "step": 1454 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33556756756756756, + "grad_norm": 22.463802337646484, + "learning_rate": 7.89512614477112e-06, + "loss": 10.028081893920898, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.42031894251704216, + "loss_total_v6": 0.6267550811171532, + "step": 1455 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3357981981981982, + "grad_norm": 21.313453674316406, + "learning_rate": 7.892028875264452e-06, + "loss": 10.07673454284668, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.41856849193573, + "loss_total_v6": 0.6297958940267563, + "step": 1456 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33602882882882884, + "grad_norm": 22.735586166381836, + "learning_rate": 7.888929937199853e-06, + "loss": 10.578744888305664, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4488913044333458, + "loss_total_v6": 0.6611715778708458, + "step": 1457 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33625945945945945, + "grad_norm": 21.275897979736328, + "learning_rate": 7.88582933236526e-06, + "loss": 10.276880264282227, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.42495885491371155, + "loss_total_v6": 0.6423050537705421, + "step": 1458 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33649009009009007, + "grad_norm": 22.549936294555664, + "learning_rate": 7.882727062549569e-06, + "loss": 10.162938117980957, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4197142831981182, + "loss_total_v6": 0.6351836547255516, + "step": 1459 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33672072072072073, + "grad_norm": 19.769941329956055, + "learning_rate": 7.879623129542633e-06, + "loss": 10.935056686401367, + "loss_alignment": 0.45068359375, + "loss_alignment_w": 0.225341796875, + "loss_sft": 0.45812977477908134, + "loss_total_v6": 0.6834410429000854, + "step": 1460 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33695135135135135, + "grad_norm": 20.62637710571289, + "learning_rate": 7.876517535135273e-06, + "loss": 10.235322952270508, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.42826659977436066, + "loss_total_v6": 0.6397076845169067, + "step": 1461 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33718198198198196, + "grad_norm": 21.967336654663086, + "learning_rate": 7.87341028111926e-06, + "loss": 10.346702575683594, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4340224042534828, + "loss_total_v6": 0.646668903529644, + "step": 1462 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33741261261261263, + "grad_norm": 23.060121536254883, + "learning_rate": 7.870301369287328e-06, + "loss": 10.333572387695312, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.43758101388812065, + "loss_total_v6": 0.645848236978054, + "step": 1463 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33764324324324324, + "grad_norm": 21.406734466552734, + "learning_rate": 7.867190801433167e-06, + "loss": 10.225542068481445, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.43069177493453026, + "loss_total_v6": 0.6390963420271873, + "step": 1464 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33787387387387385, + "grad_norm": 21.2225341796875, + "learning_rate": 7.864078579351418e-06, + "loss": 10.248714447021484, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4283558800816536, + "loss_total_v6": 0.6405446082353592, + "step": 1465 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3381045045045045, + "grad_norm": 19.84981346130371, + "learning_rate": 7.86096470483768e-06, + "loss": 10.393125534057617, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.4304693564772606, + "loss_total_v6": 0.6495703309774399, + "step": 1466 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33833513513513513, + "grad_norm": 21.526723861694336, + "learning_rate": 7.857849179688506e-06, + "loss": 10.271918296813965, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4304165244102478, + "loss_total_v6": 0.641994871199131, + "step": 1467 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33856576576576575, + "grad_norm": 21.640655517578125, + "learning_rate": 7.854732005701401e-06, + "loss": 9.610243797302246, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.39124390482902527, + "loss_total_v6": 0.600640244781971, + "step": 1468 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3387963963963964, + "grad_norm": 21.47878074645996, + "learning_rate": 7.851613184674821e-06, + "loss": 10.400824546813965, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.434566892683506, + "loss_total_v6": 0.650051511824131, + "step": 1469 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33902702702702703, + "grad_norm": 19.922733306884766, + "learning_rate": 7.84849271840817e-06, + "loss": 10.159872055053711, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.4197362847626209, + "loss_total_v6": 0.6349919959902763, + "step": 1470 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33925765765765764, + "grad_norm": 23.689071655273438, + "learning_rate": 7.845370608701804e-06, + "loss": 10.751906394958496, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4544953666627407, + "loss_total_v6": 0.6719941645860672, + "step": 1471 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3394882882882883, + "grad_norm": 21.904542922973633, + "learning_rate": 7.842246857357024e-06, + "loss": 10.677913665771484, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4512746296823025, + "loss_total_v6": 0.6673696041107178, + "step": 1472 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3397189189189189, + "grad_norm": 23.16761589050293, + "learning_rate": 7.839121466176082e-06, + "loss": 10.497437477111816, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.43856053054332733, + "loss_total_v6": 0.6560898274183273, + "step": 1473 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.33994954954954953, + "grad_norm": 23.596145629882812, + "learning_rate": 7.835994436962176e-06, + "loss": 10.650344848632812, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4528779834508896, + "loss_total_v6": 0.665646530687809, + "step": 1474 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3401801801801802, + "grad_norm": 21.740631103515625, + "learning_rate": 7.832865771519445e-06, + "loss": 10.162572860717773, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.41782981902360916, + "loss_total_v6": 0.6351607739925385, + "step": 1475 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3404108108108108, + "grad_norm": 22.75747299194336, + "learning_rate": 7.829735471652978e-06, + "loss": 10.115588188171387, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.41962355375289917, + "loss_total_v6": 0.6322242766618729, + "step": 1476 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34064144144144143, + "grad_norm": 23.476423263549805, + "learning_rate": 7.826603539168798e-06, + "loss": 10.609145164489746, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.44525232911109924, + "loss_total_v6": 0.6630715429782867, + "step": 1477 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3408720720720721, + "grad_norm": 24.10134506225586, + "learning_rate": 7.823469975873881e-06, + "loss": 10.193428993225098, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4277845062315464, + "loss_total_v6": 0.6370893195271492, + "step": 1478 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3411027027027027, + "grad_norm": 19.586076736450195, + "learning_rate": 7.820334783576134e-06, + "loss": 9.85360336303711, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.40640803426504135, + "loss_total_v6": 0.6158501878380775, + "step": 1479 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3413333333333333, + "grad_norm": 24.202943801879883, + "learning_rate": 7.81719796408441e-06, + "loss": 9.884428024291992, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4020785391330719, + "loss_total_v6": 0.6177767664194107, + "step": 1480 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.341563963963964, + "grad_norm": 24.59199333190918, + "learning_rate": 7.814059519208503e-06, + "loss": 10.402335166931152, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.4385828524827957, + "loss_total_v6": 0.6501459553837776, + "step": 1481 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3417945945945946, + "grad_norm": 26.341581344604492, + "learning_rate": 7.810919450759134e-06, + "loss": 10.000797271728516, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.41841528192162514, + "loss_total_v6": 0.6250498220324516, + "step": 1482 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3420252252252252, + "grad_norm": 22.851234436035156, + "learning_rate": 7.807777760547972e-06, + "loss": 9.899675369262695, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4030924662947655, + "loss_total_v6": 0.6187296658754349, + "step": 1483 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34225585585585583, + "grad_norm": 28.79134750366211, + "learning_rate": 7.804634450387616e-06, + "loss": 10.441256523132324, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4372922480106354, + "loss_total_v6": 0.6525785028934479, + "step": 1484 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3424864864864865, + "grad_norm": 21.024303436279297, + "learning_rate": 7.8014895220916e-06, + "loss": 10.483379364013672, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4455554261803627, + "loss_total_v6": 0.6552111804485321, + "step": 1485 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3427171171171171, + "grad_norm": 21.73990249633789, + "learning_rate": 7.798342977474391e-06, + "loss": 10.251648902893066, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.42559440433979034, + "loss_total_v6": 0.6407280713319778, + "step": 1486 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3429477477477477, + "grad_norm": 21.066238403320312, + "learning_rate": 7.795194818351395e-06, + "loss": 9.87255859375, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.4087830074131489, + "loss_total_v6": 0.6170349568128586, + "step": 1487 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3431783783783784, + "grad_norm": 24.469228744506836, + "learning_rate": 7.79204504653894e-06, + "loss": 10.545833587646484, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.447520911693573, + "loss_total_v6": 0.6591145545244217, + "step": 1488 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.343409009009009, + "grad_norm": 21.402664184570312, + "learning_rate": 7.788893663854292e-06, + "loss": 10.077413558959961, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.41755810752511024, + "loss_total_v6": 0.6298383846879005, + "step": 1489 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3436396396396396, + "grad_norm": 21.005149841308594, + "learning_rate": 7.785740672115641e-06, + "loss": 9.990118026733398, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.4159626252949238, + "loss_total_v6": 0.6243823915719986, + "step": 1490 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3438702702702703, + "grad_norm": 20.156301498413086, + "learning_rate": 7.782586073142107e-06, + "loss": 10.28822135925293, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4315422736108303, + "loss_total_v6": 0.6430138349533081, + "step": 1491 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3441009009009009, + "grad_norm": 20.641263961791992, + "learning_rate": 7.77942986875374e-06, + "loss": 10.161710739135742, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.41995803639292717, + "loss_total_v6": 0.6351069509983063, + "step": 1492 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3443315315315315, + "grad_norm": 20.310012817382812, + "learning_rate": 7.776272060771513e-06, + "loss": 9.863304138183594, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.4129500947892666, + "loss_total_v6": 0.6164565607905388, + "step": 1493 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3445621621621622, + "grad_norm": 20.32635498046875, + "learning_rate": 7.773112651017326e-06, + "loss": 10.091985702514648, + "loss_alignment": 0.396240234375, + "loss_alignment_w": 0.1981201171875, + "loss_sft": 0.43262892216444016, + "loss_total_v6": 0.6307490691542625, + "step": 1494 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3447927927927928, + "grad_norm": 20.077451705932617, + "learning_rate": 7.769951641314002e-06, + "loss": 10.286438941955566, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.4289742149412632, + "loss_total_v6": 0.6429024338722229, + "step": 1495 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3450234234234234, + "grad_norm": 20.972043991088867, + "learning_rate": 7.766789033485287e-06, + "loss": 9.654994010925293, + "loss_alignment": 0.408935546875, + "loss_alignment_w": 0.2044677734375, + "loss_sft": 0.3989083357155323, + "loss_total_v6": 0.6034371256828308, + "step": 1496 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34525405405405407, + "grad_norm": 22.052169799804688, + "learning_rate": 7.763624829355851e-06, + "loss": 10.058518409729004, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.41222672536969185, + "loss_total_v6": 0.6286573857069016, + "step": 1497 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3454846846846847, + "grad_norm": 20.033733367919922, + "learning_rate": 7.760459030751285e-06, + "loss": 10.566702842712402, + "loss_alignment": 0.438720703125, + "loss_alignment_w": 0.2193603515625, + "loss_sft": 0.44128744304180145, + "loss_total_v6": 0.6604189425706863, + "step": 1498 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3457153153153153, + "grad_norm": 22.843624114990234, + "learning_rate": 7.757291639498094e-06, + "loss": 9.838193893432617, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.40629949793219566, + "loss_total_v6": 0.6148871555924416, + "step": 1499 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34594594594594597, + "grad_norm": 36.85215759277344, + "learning_rate": 7.75412265742371e-06, + "loss": 9.890464782714844, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.40497345104813576, + "loss_total_v6": 0.6181540116667747, + "step": 1500 + }, + { + "epoch": 0.34594594594594597, + "eval_loss": 0.6737541556358337, + "eval_loss_alignment": 0.4068238348023504, + "eval_loss_alignment_w": 0.2034119174011752, + "eval_loss_sft": 0.47034234506611383, + "eval_loss_total_v6": 0.6737542621727682, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.5239, + "eval_samples_per_second": 22.612, + "eval_steps_per_second": 2.836, + "step": 1500 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3461765765765766, + "grad_norm": 22.67705726623535, + "learning_rate": 7.750952086356477e-06, + "loss": 10.16213607788086, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.42357037961483, + "loss_total_v6": 0.6351334825158119, + "step": 1501 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3464072072072072, + "grad_norm": 22.220266342163086, + "learning_rate": 7.74777992812566e-06, + "loss": 9.730696678161621, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.3971852958202362, + "loss_total_v6": 0.6081685423851013, + "step": 1502 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34663783783783786, + "grad_norm": 19.761924743652344, + "learning_rate": 7.744606184561439e-06, + "loss": 9.878145217895508, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.40209779888391495, + "loss_total_v6": 0.6173840314149857, + "step": 1503 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34686846846846847, + "grad_norm": 19.281211853027344, + "learning_rate": 7.741430857494904e-06, + "loss": 10.115686416625977, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.42455829307436943, + "loss_total_v6": 0.6322304159402847, + "step": 1504 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3470990990990991, + "grad_norm": 19.92724609375, + "learning_rate": 7.738253948758063e-06, + "loss": 10.077950477600098, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4165998436510563, + "loss_total_v6": 0.6298719123005867, + "step": 1505 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34732972972972975, + "grad_norm": 21.40612030029297, + "learning_rate": 7.735075460183835e-06, + "loss": 10.45177173614502, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.44060447812080383, + "loss_total_v6": 0.6532357260584831, + "step": 1506 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34756036036036037, + "grad_norm": 19.682729721069336, + "learning_rate": 7.731895393606052e-06, + "loss": 10.33857536315918, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4309204965829849, + "loss_total_v6": 0.6461609750986099, + "step": 1507 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.347790990990991, + "grad_norm": 20.908859252929688, + "learning_rate": 7.728713750859458e-06, + "loss": 10.027368545532227, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.41771094501018524, + "loss_total_v6": 0.6267105601727962, + "step": 1508 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34802162162162165, + "grad_norm": 19.705678939819336, + "learning_rate": 7.7255305337797e-06, + "loss": 10.714071273803711, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.45124560967087746, + "loss_total_v6": 0.6696294471621513, + "step": 1509 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34825225225225226, + "grad_norm": 26.78250503540039, + "learning_rate": 7.72234574420334e-06, + "loss": 9.860969543457031, + "loss_alignment": 0.41259765625, + "loss_alignment_w": 0.206298828125, + "loss_sft": 0.41010335460305214, + "loss_total_v6": 0.6163106262683868, + "step": 1510 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34848288288288287, + "grad_norm": 211.73617553710938, + "learning_rate": 7.71915938396784e-06, + "loss": 10.693459510803223, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.45111706480383873, + "loss_total_v6": 0.6683412045240402, + "step": 1511 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34871351351351354, + "grad_norm": 27.40892791748047, + "learning_rate": 7.715971454911576e-06, + "loss": 10.20683479309082, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4288511835038662, + "loss_total_v6": 0.6379271224141121, + "step": 1512 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34894414414414415, + "grad_norm": 21.18890953063965, + "learning_rate": 7.712781958873827e-06, + "loss": 10.37668228149414, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.432447649538517, + "loss_total_v6": 0.6485426351428032, + "step": 1513 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34917477477477477, + "grad_norm": 21.259267807006836, + "learning_rate": 7.709590897694771e-06, + "loss": 9.888598442077637, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4058486931025982, + "loss_total_v6": 0.618037398904562, + "step": 1514 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3494054054054054, + "grad_norm": 20.311622619628906, + "learning_rate": 7.706398273215497e-06, + "loss": 10.331452369689941, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.43358803540468216, + "loss_total_v6": 0.6457157582044601, + "step": 1515 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34963603603603605, + "grad_norm": 20.733306884765625, + "learning_rate": 7.703204087277989e-06, + "loss": 10.645764350891113, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.4537818916141987, + "loss_total_v6": 0.665360264480114, + "step": 1516 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.34986666666666666, + "grad_norm": 20.339859008789062, + "learning_rate": 7.700008341725134e-06, + "loss": 10.614875793457031, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.45458270609378815, + "loss_total_v6": 0.6634297668933868, + "step": 1517 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35009729729729727, + "grad_norm": 24.941253662109375, + "learning_rate": 7.696811038400722e-06, + "loss": 9.939325332641602, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4092937484383583, + "loss_total_v6": 0.6212078183889389, + "step": 1518 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35032792792792794, + "grad_norm": 19.498090744018555, + "learning_rate": 7.693612179149437e-06, + "loss": 10.523509979248047, + "loss_alignment": 0.44140625, + "loss_alignment_w": 0.220703125, + "loss_sft": 0.43683309480547905, + "loss_total_v6": 0.6577193513512611, + "step": 1519 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35055855855855855, + "grad_norm": 21.091909408569336, + "learning_rate": 7.690411765816864e-06, + "loss": 9.78366470336914, + "loss_alignment": 0.413330078125, + "loss_alignment_w": 0.2066650390625, + "loss_sft": 0.4047528952360153, + "loss_total_v6": 0.6114789843559265, + "step": 1520 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35078918918918917, + "grad_norm": 23.7603759765625, + "learning_rate": 7.687209800249484e-06, + "loss": 10.272858619689941, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.42957499250769615, + "loss_total_v6": 0.642053633928299, + "step": 1521 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35101981981981983, + "grad_norm": 21.333393096923828, + "learning_rate": 7.684006284294672e-06, + "loss": 10.152881622314453, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4244110472500324, + "loss_total_v6": 0.6345550678670406, + "step": 1522 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35125045045045045, + "grad_norm": 23.06925392150879, + "learning_rate": 7.680801219800698e-06, + "loss": 10.142766952514648, + "loss_alignment": 0.40576171875, + "loss_alignment_w": 0.202880859375, + "loss_sft": 0.43090472370386124, + "loss_total_v6": 0.6339229196310043, + "step": 1523 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35148108108108106, + "grad_norm": 22.20785140991211, + "learning_rate": 7.67759460861673e-06, + "loss": 10.175156593322754, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.41658689826726913, + "loss_total_v6": 0.6359472647309303, + "step": 1524 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3517117117117117, + "grad_norm": 21.402538299560547, + "learning_rate": 7.674386452592818e-06, + "loss": 10.148262023925781, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4201703332364559, + "loss_total_v6": 0.6342664062976837, + "step": 1525 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35194234234234234, + "grad_norm": 21.822065353393555, + "learning_rate": 7.671176753579914e-06, + "loss": 10.639314651489258, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4529210962355137, + "loss_total_v6": 0.6649572253227234, + "step": 1526 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35217297297297295, + "grad_norm": 22.86586570739746, + "learning_rate": 7.667965513429856e-06, + "loss": 11.00836181640625, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.47140882164239883, + "loss_total_v6": 0.68802260607481, + "step": 1527 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3524036036036036, + "grad_norm": 22.163272857666016, + "learning_rate": 7.66475273399537e-06, + "loss": 10.64861011505127, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.45699627697467804, + "loss_total_v6": 0.6655381470918655, + "step": 1528 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35263423423423423, + "grad_norm": 22.012466430664062, + "learning_rate": 7.661538417130072e-06, + "loss": 10.080245971679688, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.41991711780428886, + "loss_total_v6": 0.6300153657793999, + "step": 1529 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35286486486486485, + "grad_norm": 21.802658081054688, + "learning_rate": 7.658322564688463e-06, + "loss": 9.670726776123047, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.3962295390665531, + "loss_total_v6": 0.6044204607605934, + "step": 1530 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3530954954954955, + "grad_norm": 22.22856330871582, + "learning_rate": 7.655105178525932e-06, + "loss": 10.386680603027344, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4307073913514614, + "loss_total_v6": 0.6491675153374672, + "step": 1531 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3533261261261261, + "grad_norm": 22.37187385559082, + "learning_rate": 7.651886260498751e-06, + "loss": 10.269784927368164, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4258887246251106, + "loss_total_v6": 0.6418615952134132, + "step": 1532 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35355675675675674, + "grad_norm": 20.651107788085938, + "learning_rate": 7.648665812464079e-06, + "loss": 9.760040283203125, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.3962879553437233, + "loss_total_v6": 0.6100025549530983, + "step": 1533 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3537873873873874, + "grad_norm": 20.96525764465332, + "learning_rate": 7.645443836279951e-06, + "loss": 10.517413139343262, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.45015446841716766, + "loss_total_v6": 0.6573383063077927, + "step": 1534 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.354018018018018, + "grad_norm": 18.758726119995117, + "learning_rate": 7.642220333805294e-06, + "loss": 10.17059326171875, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.41805650666356087, + "loss_total_v6": 0.6356621086597443, + "step": 1535 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35424864864864863, + "grad_norm": 20.436594009399414, + "learning_rate": 7.638995306899907e-06, + "loss": 10.419031143188477, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.4288688786327839, + "loss_total_v6": 0.6511894166469574, + "step": 1536 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3544792792792793, + "grad_norm": 21.25202178955078, + "learning_rate": 7.635768757424473e-06, + "loss": 10.271547317504883, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4259682856500149, + "loss_total_v6": 0.6419716849923134, + "step": 1537 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3547099099099099, + "grad_norm": 21.47968864440918, + "learning_rate": 7.632540687240547e-06, + "loss": 10.513450622558594, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4403242841362953, + "loss_total_v6": 0.6570906490087509, + "step": 1538 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35494054054054053, + "grad_norm": 25.62847900390625, + "learning_rate": 7.62931109821057e-06, + "loss": 10.046632766723633, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.40994273498654366, + "loss_total_v6": 0.6279145553708076, + "step": 1539 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3551711711711712, + "grad_norm": 22.708084106445312, + "learning_rate": 7.626079992197857e-06, + "loss": 10.211008071899414, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4273268058896065, + "loss_total_v6": 0.6381880193948746, + "step": 1540 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3554018018018018, + "grad_norm": 21.4726619720459, + "learning_rate": 7.6228473710665955e-06, + "loss": 9.976223945617676, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.41282064095139503, + "loss_total_v6": 0.6235139816999435, + "step": 1541 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3556324324324324, + "grad_norm": 22.263612747192383, + "learning_rate": 7.619613236681845e-06, + "loss": 10.289069175720215, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.429779477417469, + "loss_total_v6": 0.6430668234825134, + "step": 1542 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3558630630630631, + "grad_norm": 23.029781341552734, + "learning_rate": 7.616377590909545e-06, + "loss": 10.71631908416748, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.45571964606642723, + "loss_total_v6": 0.6697699576616287, + "step": 1543 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3560936936936937, + "grad_norm": 21.7050838470459, + "learning_rate": 7.613140435616503e-06, + "loss": 10.475566864013672, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.4459827318787575, + "loss_total_v6": 0.6547229513525963, + "step": 1544 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3563243243243243, + "grad_norm": 19.68143081665039, + "learning_rate": 7.609901772670396e-06, + "loss": 10.395783424377441, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.437685064971447, + "loss_total_v6": 0.6497364789247513, + "step": 1545 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35655495495495493, + "grad_norm": 23.239532470703125, + "learning_rate": 7.6066616039397755e-06, + "loss": 9.609580993652344, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.39489512145519257, + "loss_total_v6": 0.6005988419055939, + "step": 1546 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3567855855855856, + "grad_norm": 22.023792266845703, + "learning_rate": 7.603419931294056e-06, + "loss": 9.733255386352539, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.3972994424402714, + "loss_total_v6": 0.6083284988999367, + "step": 1547 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3570162162162162, + "grad_norm": 21.29875946044922, + "learning_rate": 7.600176756603524e-06, + "loss": 10.658543586730957, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4491637200117111, + "loss_total_v6": 0.6661589592695236, + "step": 1548 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3572468468468468, + "grad_norm": 22.752885818481445, + "learning_rate": 7.5969320817393326e-06, + "loss": 10.291324615478516, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.4228556379675865, + "loss_total_v6": 0.643207810819149, + "step": 1549 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3574774774774775, + "grad_norm": 24.86031723022461, + "learning_rate": 7.593685908573497e-06, + "loss": 9.830892562866211, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4014944061636925, + "loss_total_v6": 0.614430770277977, + "step": 1550 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3577081081081081, + "grad_norm": 22.372404098510742, + "learning_rate": 7.590438238978899e-06, + "loss": 10.640787124633789, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4496866427361965, + "loss_total_v6": 0.6650491803884506, + "step": 1551 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3579387387387387, + "grad_norm": 20.81427574157715, + "learning_rate": 7.587189074829284e-06, + "loss": 10.26991081237793, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.431954275816679, + "loss_total_v6": 0.6418694257736206, + "step": 1552 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3581693693693694, + "grad_norm": 19.68218231201172, + "learning_rate": 7.583938417999261e-06, + "loss": 10.063834190368652, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.41496986895799637, + "loss_total_v6": 0.6289896368980408, + "step": 1553 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3584, + "grad_norm": 20.29039192199707, + "learning_rate": 7.580686270364298e-06, + "loss": 10.073986053466797, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.4225013889372349, + "loss_total_v6": 0.6296241879463196, + "step": 1554 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3586306306306306, + "grad_norm": 20.444738388061523, + "learning_rate": 7.577432633800721e-06, + "loss": 10.153446197509766, + "loss_alignment": 0.41162109375, + "loss_alignment_w": 0.205810546875, + "loss_sft": 0.4289019852876663, + "loss_total_v6": 0.6345904543995857, + "step": 1555 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3588612612612613, + "grad_norm": 21.222572326660156, + "learning_rate": 7.574177510185721e-06, + "loss": 10.185243606567383, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.43343741819262505, + "loss_total_v6": 0.6365776807069778, + "step": 1556 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3590918918918919, + "grad_norm": 21.74709129333496, + "learning_rate": 7.570920901397342e-06, + "loss": 10.078746795654297, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4133231081068516, + "loss_total_v6": 0.629921693354845, + "step": 1557 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3593225225225225, + "grad_norm": 19.917970657348633, + "learning_rate": 7.567662809314485e-06, + "loss": 9.895803451538086, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.40397967770695686, + "loss_total_v6": 0.6184877082705498, + "step": 1558 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.35955315315315317, + "grad_norm": 21.297637939453125, + "learning_rate": 7.564403235816913e-06, + "loss": 9.926338195800781, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.40608642250299454, + "loss_total_v6": 0.620396114885807, + "step": 1559 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3597837837837838, + "grad_norm": 20.76892852783203, + "learning_rate": 7.561142182785234e-06, + "loss": 10.157380104064941, + "loss_alignment": 0.446533203125, + "loss_alignment_w": 0.2232666015625, + "loss_sft": 0.41199688985943794, + "loss_total_v6": 0.63483627140522, + "step": 1560 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3600144144144144, + "grad_norm": 21.25116539001465, + "learning_rate": 7.557879652100917e-06, + "loss": 10.097076416015625, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4211827218532562, + "loss_total_v6": 0.6310673356056213, + "step": 1561 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36024504504504506, + "grad_norm": 25.33258056640625, + "learning_rate": 7.554615645646282e-06, + "loss": 10.172508239746094, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.42098383232951164, + "loss_total_v6": 0.6357817947864532, + "step": 1562 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3604756756756757, + "grad_norm": 20.620840072631836, + "learning_rate": 7.5513501653045e-06, + "loss": 10.511114120483398, + "loss_alignment": 0.405029296875, + "loss_alignment_w": 0.2025146484375, + "loss_sft": 0.4537586122751236, + "loss_total_v6": 0.6569446548819542, + "step": 1563 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3607063063063063, + "grad_norm": 22.187606811523438, + "learning_rate": 7.548083212959588e-06, + "loss": 9.341597557067871, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.37379733473062515, + "loss_total_v6": 0.5838498547673225, + "step": 1564 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36093693693693696, + "grad_norm": 22.28821563720703, + "learning_rate": 7.5448147904964196e-06, + "loss": 10.084063529968262, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.41780590265989304, + "loss_total_v6": 0.6302539929747581, + "step": 1565 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36116756756756757, + "grad_norm": 20.96930694580078, + "learning_rate": 7.541544899800712e-06, + "loss": 10.388368606567383, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.43056880310177803, + "loss_total_v6": 0.6492730304598808, + "step": 1566 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3613981981981982, + "grad_norm": 24.519851684570312, + "learning_rate": 7.538273542759033e-06, + "loss": 10.373780250549316, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4349976181983948, + "loss_total_v6": 0.6483612805604935, + "step": 1567 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36162882882882885, + "grad_norm": 20.892702102661133, + "learning_rate": 7.535000721258792e-06, + "loss": 9.830224990844727, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4056182727217674, + "loss_total_v6": 0.614389032125473, + "step": 1568 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36185945945945946, + "grad_norm": 23.944246292114258, + "learning_rate": 7.531726437188247e-06, + "loss": 9.918707847595215, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4072880484163761, + "loss_total_v6": 0.6199192479252815, + "step": 1569 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3620900900900901, + "grad_norm": 20.95834732055664, + "learning_rate": 7.528450692436495e-06, + "loss": 10.385021209716797, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4329382888972759, + "loss_total_v6": 0.6490637958049774, + "step": 1570 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36232072072072075, + "grad_norm": 22.342981338500977, + "learning_rate": 7.525173488893485e-06, + "loss": 10.400758743286133, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4413834623992443, + "loss_total_v6": 0.6500474140048027, + "step": 1571 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36255135135135136, + "grad_norm": 22.131322860717773, + "learning_rate": 7.521894828449994e-06, + "loss": 9.927932739257812, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4100770354270935, + "loss_total_v6": 0.6204957515001297, + "step": 1572 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36278198198198197, + "grad_norm": 21.973318099975586, + "learning_rate": 7.518614712997652e-06, + "loss": 10.253524780273438, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.42995356023311615, + "loss_total_v6": 0.6408452913165092, + "step": 1573 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36301261261261264, + "grad_norm": 21.678125381469727, + "learning_rate": 7.515333144428925e-06, + "loss": 10.1387939453125, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.41728973388671875, + "loss_total_v6": 0.6336746364831924, + "step": 1574 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36324324324324325, + "grad_norm": 28.894662857055664, + "learning_rate": 7.512050124637114e-06, + "loss": 10.233916282653809, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4246387258172035, + "loss_total_v6": 0.6396197974681854, + "step": 1575 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36347387387387387, + "grad_norm": 23.54343605041504, + "learning_rate": 7.508765655516358e-06, + "loss": 10.071365356445312, + "loss_alignment": 0.41015625, + "loss_alignment_w": 0.205078125, + "loss_sft": 0.42429066076874733, + "loss_total_v6": 0.629460334777832, + "step": 1576 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3637045045045045, + "grad_norm": 22.03545379638672, + "learning_rate": 7.505479738961635e-06, + "loss": 10.177750587463379, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4219065345823765, + "loss_total_v6": 0.6361094266176224, + "step": 1577 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36393513513513515, + "grad_norm": 20.714746475219727, + "learning_rate": 7.502192376868758e-06, + "loss": 9.70655345916748, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.3945166580379009, + "loss_total_v6": 0.6066595762968063, + "step": 1578 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36416576576576576, + "grad_norm": 22.35489845275879, + "learning_rate": 7.498903571134371e-06, + "loss": 10.327393531799316, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.43289192765951157, + "loss_total_v6": 0.6454621180891991, + "step": 1579 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36439639639639637, + "grad_norm": 22.52785301208496, + "learning_rate": 7.4956133236559525e-06, + "loss": 10.324697494506836, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.42715394496917725, + "loss_total_v6": 0.6452935934066772, + "step": 1580 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36462702702702704, + "grad_norm": 25.96314239501953, + "learning_rate": 7.492321636331814e-06, + "loss": 9.897289276123047, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4105117991566658, + "loss_total_v6": 0.618580624461174, + "step": 1581 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36485765765765765, + "grad_norm": 22.950927734375, + "learning_rate": 7.489028511061097e-06, + "loss": 9.983671188354492, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.41526974365115166, + "loss_total_v6": 0.6239794567227364, + "step": 1582 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36508828828828827, + "grad_norm": 24.9093017578125, + "learning_rate": 7.485733949743771e-06, + "loss": 10.05937385559082, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.4129210636019707, + "loss_total_v6": 0.6287108510732651, + "step": 1583 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36531891891891893, + "grad_norm": 21.779136657714844, + "learning_rate": 7.4824379542806355e-06, + "loss": 10.343856811523438, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.43236446380615234, + "loss_total_v6": 0.646491065621376, + "step": 1584 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36554954954954955, + "grad_norm": 22.190492630004883, + "learning_rate": 7.479140526573317e-06, + "loss": 10.253971099853516, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.42793674394488335, + "loss_total_v6": 0.6408731415867805, + "step": 1585 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36578018018018016, + "grad_norm": 24.065643310546875, + "learning_rate": 7.475841668524268e-06, + "loss": 9.955992698669434, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.4186667688190937, + "loss_total_v6": 0.6222495064139366, + "step": 1586 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3660108108108108, + "grad_norm": 23.502954483032227, + "learning_rate": 7.47254138203677e-06, + "loss": 10.313091278076172, + "loss_alignment": 0.4453125, + "loss_alignment_w": 0.22265625, + "loss_sft": 0.42194245755672455, + "loss_total_v6": 0.644568182528019, + "step": 1587 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36624144144144144, + "grad_norm": 23.63500213623047, + "learning_rate": 7.469239669014922e-06, + "loss": 10.505531311035156, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4458107650279999, + "loss_total_v6": 0.6565956547856331, + "step": 1588 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36647207207207205, + "grad_norm": 21.502363204956055, + "learning_rate": 7.465936531363651e-06, + "loss": 10.212618827819824, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.428693950176239, + "loss_total_v6": 0.6382886916399002, + "step": 1589 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3667027027027027, + "grad_norm": 20.754108428955078, + "learning_rate": 7.462631970988703e-06, + "loss": 10.561233520507812, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.44967367500066757, + "loss_total_v6": 0.6600771173834801, + "step": 1590 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36693333333333333, + "grad_norm": 21.812847137451172, + "learning_rate": 7.459325989796647e-06, + "loss": 10.497579574584961, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.43921032920479774, + "loss_total_v6": 0.6560987681150436, + "step": 1591 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36716396396396395, + "grad_norm": 20.461612701416016, + "learning_rate": 7.4560185896948735e-06, + "loss": 10.319206237792969, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.43233443051576614, + "loss_total_v6": 0.6449503898620605, + "step": 1592 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3673945945945946, + "grad_norm": 20.951128005981445, + "learning_rate": 7.452709772591586e-06, + "loss": 10.030366897583008, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4103299677371979, + "loss_total_v6": 0.626897968351841, + "step": 1593 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3676252252252252, + "grad_norm": 24.637317657470703, + "learning_rate": 7.449399540395807e-06, + "loss": 10.62545394897461, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4470346085727215, + "loss_total_v6": 0.6640908718109131, + "step": 1594 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36785585585585584, + "grad_norm": 23.12552261352539, + "learning_rate": 7.446087895017382e-06, + "loss": 9.36526870727539, + "loss_alignment": 0.400390625, + "loss_alignment_w": 0.2001953125, + "loss_sft": 0.3849814012646675, + "loss_total_v6": 0.585329320281744, + "step": 1595 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3680864864864865, + "grad_norm": 24.49860191345215, + "learning_rate": 7.442774838366965e-06, + "loss": 10.162853240966797, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.4255683645606041, + "loss_total_v6": 0.635178342461586, + "step": 1596 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3683171171171171, + "grad_norm": 23.16735076904297, + "learning_rate": 7.439460372356025e-06, + "loss": 10.046707153320312, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.41327378526329994, + "loss_total_v6": 0.6279191598296165, + "step": 1597 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.36854774774774773, + "grad_norm": 22.12356185913086, + "learning_rate": 7.436144498896845e-06, + "loss": 9.98689079284668, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.41159526631236076, + "loss_total_v6": 0.6241807118058205, + "step": 1598 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3687783783783784, + "grad_norm": 23.773258209228516, + "learning_rate": 7.432827219902522e-06, + "loss": 10.229360580444336, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.4308694079518318, + "loss_total_v6": 0.6393350064754486, + "step": 1599 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.369009009009009, + "grad_norm": 24.912614822387695, + "learning_rate": 7.429508537286963e-06, + "loss": 10.445525169372559, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.4490642361342907, + "loss_total_v6": 0.6528453454375267, + "step": 1600 + }, + { + "epoch": 0.369009009009009, + "eval_loss": 0.6713927984237671, + "eval_loss_alignment": 0.40645658052884615, + "eval_loss_alignment_w": 0.20322829026442307, + "eval_loss_sft": 0.46816449837647695, + "eval_loss_total_v6": 0.6713927886011, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.4881, + "eval_samples_per_second": 22.621, + "eval_steps_per_second": 2.837, + "step": 1600 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3692396396396396, + "grad_norm": 23.02277374267578, + "learning_rate": 7.426188452964882e-06, + "loss": 10.555325508117676, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.4384249076247215, + "loss_total_v6": 0.6597078517079353, + "step": 1601 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3694702702702703, + "grad_norm": 19.847646713256836, + "learning_rate": 7.422866968851805e-06, + "loss": 9.879027366638184, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4033431150019169, + "loss_total_v6": 0.6174391806125641, + "step": 1602 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3697009009009009, + "grad_norm": 21.858840942382812, + "learning_rate": 7.4195440868640624e-06, + "loss": 10.2924222946167, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4334985315799713, + "loss_total_v6": 0.6432763859629631, + "step": 1603 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3699315315315315, + "grad_norm": 24.582704544067383, + "learning_rate": 7.4162198089187945e-06, + "loss": 9.937796592712402, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.40765710175037384, + "loss_total_v6": 0.6211123019456863, + "step": 1604 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3701621621621622, + "grad_norm": 22.070751190185547, + "learning_rate": 7.412894136933947e-06, + "loss": 9.919185638427734, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4042966738343239, + "loss_total_v6": 0.6199491247534752, + "step": 1605 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3703927927927928, + "grad_norm": 24.34600067138672, + "learning_rate": 7.409567072828268e-06, + "loss": 10.212691307067871, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.42895784229040146, + "loss_total_v6": 0.6382931843400002, + "step": 1606 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3706234234234234, + "grad_norm": 21.22462272644043, + "learning_rate": 7.406238618521306e-06, + "loss": 10.183266639709473, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.42501309886574745, + "loss_total_v6": 0.6364541351795197, + "step": 1607 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.370854054054054, + "grad_norm": 22.441207885742188, + "learning_rate": 7.402908775933419e-06, + "loss": 10.236467361450195, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4287196211516857, + "loss_total_v6": 0.6397792026400566, + "step": 1608 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3710846846846847, + "grad_norm": 23.057619094848633, + "learning_rate": 7.399577546985761e-06, + "loss": 9.800232887268066, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.40031060576438904, + "loss_total_v6": 0.6125145740807056, + "step": 1609 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3713153153153153, + "grad_norm": 23.99722671508789, + "learning_rate": 7.396244933600285e-06, + "loss": 9.77229118347168, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.39551249518990517, + "loss_total_v6": 0.6107682138681412, + "step": 1610 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3715459459459459, + "grad_norm": 22.19182777404785, + "learning_rate": 7.392910937699745e-06, + "loss": 10.322229385375977, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.43647539988160133, + "loss_total_v6": 0.6451393589377403, + "step": 1611 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3717765765765766, + "grad_norm": 21.219575881958008, + "learning_rate": 7.389575561207693e-06, + "loss": 10.19272232055664, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.42673321068286896, + "loss_total_v6": 0.6370450928807259, + "step": 1612 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3720072072072072, + "grad_norm": 22.384132385253906, + "learning_rate": 7.386238806048476e-06, + "loss": 10.301454544067383, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.4343987852334976, + "loss_total_v6": 0.643840916454792, + "step": 1613 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3722378378378378, + "grad_norm": 20.999998092651367, + "learning_rate": 7.382900674147238e-06, + "loss": 10.227368354797363, + "loss_alignment": 0.40673828125, + "loss_alignment_w": 0.203369140625, + "loss_sft": 0.4354751296341419, + "loss_total_v6": 0.639210507273674, + "step": 1614 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3724684684684685, + "grad_norm": 21.12223243713379, + "learning_rate": 7.379561167429915e-06, + "loss": 9.681461334228516, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.3928263522684574, + "loss_total_v6": 0.6050913706421852, + "step": 1615 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3726990990990991, + "grad_norm": 22.825063705444336, + "learning_rate": 7.376220287823237e-06, + "loss": 9.83635139465332, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.40207967907190323, + "loss_total_v6": 0.6147719323635101, + "step": 1616 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3729297297297297, + "grad_norm": 24.60537338256836, + "learning_rate": 7.372878037254728e-06, + "loss": 9.679960250854492, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.39384638890624046, + "loss_total_v6": 0.604997493326664, + "step": 1617 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3731603603603604, + "grad_norm": 22.381195068359375, + "learning_rate": 7.369534417652702e-06, + "loss": 9.690025329589844, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.3979850262403488, + "loss_total_v6": 0.605626605451107, + "step": 1618 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.373390990990991, + "grad_norm": 22.31298828125, + "learning_rate": 7.366189430946262e-06, + "loss": 10.463461875915527, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.43829861283302307, + "loss_total_v6": 0.6539663523435593, + "step": 1619 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3736216216216216, + "grad_norm": 21.51848793029785, + "learning_rate": 7.3628430790653e-06, + "loss": 10.858986854553223, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4661165103316307, + "loss_total_v6": 0.678686685860157, + "step": 1620 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37385225225225227, + "grad_norm": 22.433874130249023, + "learning_rate": 7.359495363940495e-06, + "loss": 10.638738632202148, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.4595073312520981, + "loss_total_v6": 0.6649211570620537, + "step": 1621 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3740828828828829, + "grad_norm": 22.013181686401367, + "learning_rate": 7.356146287503316e-06, + "loss": 10.0173978805542, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.41528716683387756, + "loss_total_v6": 0.6260873526334763, + "step": 1622 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3743135135135135, + "grad_norm": 21.73736000061035, + "learning_rate": 7.3527958516860124e-06, + "loss": 10.500736236572266, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.4362794756889343, + "loss_total_v6": 0.6562959477305412, + "step": 1623 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37454414414414416, + "grad_norm": 21.242570877075195, + "learning_rate": 7.34944405842162e-06, + "loss": 10.531354904174805, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.441168662160635, + "loss_total_v6": 0.6582096964120865, + "step": 1624 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3747747747747748, + "grad_norm": 23.602460861206055, + "learning_rate": 7.346090909643959e-06, + "loss": 9.71674633026123, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.39362782984972, + "loss_total_v6": 0.6072966605424881, + "step": 1625 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3750054054054054, + "grad_norm": 23.827211380004883, + "learning_rate": 7.3427364072876325e-06, + "loss": 10.197607040405273, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4245361499488354, + "loss_total_v6": 0.6373504549264908, + "step": 1626 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37523603603603606, + "grad_norm": 25.63290023803711, + "learning_rate": 7.339380553288021e-06, + "loss": 9.859026908874512, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4024898074567318, + "loss_total_v6": 0.6161891520023346, + "step": 1627 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37546666666666667, + "grad_norm": 23.228679656982422, + "learning_rate": 7.336023349581287e-06, + "loss": 10.479904174804688, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.43897534161806107, + "loss_total_v6": 0.6549940183758736, + "step": 1628 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3756972972972973, + "grad_norm": 23.407512664794922, + "learning_rate": 7.332664798104371e-06, + "loss": 10.043924331665039, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.41187918186187744, + "loss_total_v6": 0.6277452707290649, + "step": 1629 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37592792792792795, + "grad_norm": 21.671710968017578, + "learning_rate": 7.329304900794991e-06, + "loss": 10.174566268920898, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.42230258882045746, + "loss_total_v6": 0.6359103918075562, + "step": 1630 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37615855855855856, + "grad_norm": 20.1163272857666, + "learning_rate": 7.325943659591644e-06, + "loss": 10.51742935180664, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.44409771636128426, + "loss_total_v6": 0.6573393195867538, + "step": 1631 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3763891891891892, + "grad_norm": 21.738685607910156, + "learning_rate": 7.3225810764335965e-06, + "loss": 10.07017707824707, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.41432875767350197, + "loss_total_v6": 0.6293860822916031, + "step": 1632 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37661981981981985, + "grad_norm": 21.470727920532227, + "learning_rate": 7.319217153260897e-06, + "loss": 10.058731079101562, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.41527656093239784, + "loss_total_v6": 0.6286706998944283, + "step": 1633 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37685045045045046, + "grad_norm": 22.77970314025879, + "learning_rate": 7.3158518920143615e-06, + "loss": 10.226079940795898, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.42588840797543526, + "loss_total_v6": 0.6391300037503242, + "step": 1634 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37708108108108107, + "grad_norm": 22.3476619720459, + "learning_rate": 7.312485294635581e-06, + "loss": 10.061210632324219, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.423289742320776, + "loss_total_v6": 0.6288256123661995, + "step": 1635 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37731171171171174, + "grad_norm": 21.762508392333984, + "learning_rate": 7.3091173630669135e-06, + "loss": 9.967976570129395, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.40734609216451645, + "loss_total_v6": 0.6229985356330872, + "step": 1636 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37754234234234235, + "grad_norm": 20.444379806518555, + "learning_rate": 7.3057480992514895e-06, + "loss": 10.340490341186523, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4290412440896034, + "loss_total_v6": 0.6462806239724159, + "step": 1637 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37777297297297296, + "grad_norm": 23.79673194885254, + "learning_rate": 7.302377505133209e-06, + "loss": 10.247995376586914, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.42687664926052094, + "loss_total_v6": 0.6404997259378433, + "step": 1638 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3780036036036036, + "grad_norm": 27.77680206298828, + "learning_rate": 7.299005582656739e-06, + "loss": 9.639850616455078, + "loss_alignment": 0.407470703125, + "loss_alignment_w": 0.2037353515625, + "loss_sft": 0.39837388694286346, + "loss_total_v6": 0.6024906933307648, + "step": 1639 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37823423423423425, + "grad_norm": 22.721195220947266, + "learning_rate": 7.295632333767513e-06, + "loss": 9.601526260375977, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.39010394364595413, + "loss_total_v6": 0.6000954210758209, + "step": 1640 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37846486486486486, + "grad_norm": 24.527751922607422, + "learning_rate": 7.292257760411726e-06, + "loss": 9.958150863647461, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.408639270812273, + "loss_total_v6": 0.6223843842744827, + "step": 1641 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37869549549549547, + "grad_norm": 19.910888671875, + "learning_rate": 7.288881864536344e-06, + "loss": 10.460594177246094, + "loss_alignment": 0.4140625, + "loss_alignment_w": 0.20703125, + "loss_sft": 0.44631337001919746, + "loss_total_v6": 0.6537871286273003, + "step": 1642 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37892612612612614, + "grad_norm": 22.984203338623047, + "learning_rate": 7.28550464808909e-06, + "loss": 10.007514953613281, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.4147610366344452, + "loss_total_v6": 0.6254696696996689, + "step": 1643 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37915675675675675, + "grad_norm": 21.400157928466797, + "learning_rate": 7.28212611301845e-06, + "loss": 9.644142150878906, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.39140933752059937, + "loss_total_v6": 0.6027588397264481, + "step": 1644 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37938738738738736, + "grad_norm": 22.665367126464844, + "learning_rate": 7.278746261273676e-06, + "loss": 9.824285507202148, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.3968547284603119, + "loss_total_v6": 0.6140178143978119, + "step": 1645 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37961801801801803, + "grad_norm": 21.611448287963867, + "learning_rate": 7.2753650948047716e-06, + "loss": 10.007207870483398, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4165424294769764, + "loss_total_v6": 0.6254505217075348, + "step": 1646 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.37984864864864865, + "grad_norm": 20.68217658996582, + "learning_rate": 7.271982615562506e-06, + "loss": 9.324527740478516, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.370365384966135, + "loss_total_v6": 0.5827829763293266, + "step": 1647 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38007927927927926, + "grad_norm": 26.128890991210938, + "learning_rate": 7.2685988254984006e-06, + "loss": 10.515961647033691, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4407711662352085, + "loss_total_v6": 0.6572475880384445, + "step": 1648 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3803099099099099, + "grad_norm": 23.07623291015625, + "learning_rate": 7.265213726564737e-06, + "loss": 9.922298431396484, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4081227295100689, + "loss_total_v6": 0.6201436296105385, + "step": 1649 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38054054054054054, + "grad_norm": 23.08317756652832, + "learning_rate": 7.26182732071455e-06, + "loss": 10.104374885559082, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.41538265720009804, + "loss_total_v6": 0.6315234303474426, + "step": 1650 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38077117117117115, + "grad_norm": 24.42818260192871, + "learning_rate": 7.258439609901625e-06, + "loss": 9.583026885986328, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.3912060186266899, + "loss_total_v6": 0.5989391729235649, + "step": 1651 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3810018018018018, + "grad_norm": 22.304025650024414, + "learning_rate": 7.25505059608051e-06, + "loss": 10.047628402709961, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.41732916235923767, + "loss_total_v6": 0.6279767751693726, + "step": 1652 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38123243243243243, + "grad_norm": 22.998517990112305, + "learning_rate": 7.251660281206495e-06, + "loss": 10.914472579956055, + "loss_alignment": 0.440673828125, + "loss_alignment_w": 0.2203369140625, + "loss_sft": 0.4619549736380577, + "loss_total_v6": 0.6821545511484146, + "step": 1653 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38146306306306305, + "grad_norm": 19.46958351135254, + "learning_rate": 7.248268667235624e-06, + "loss": 9.919990539550781, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.41156434267759323, + "loss_total_v6": 0.6199994161725044, + "step": 1654 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3816936936936937, + "grad_norm": 21.034278869628906, + "learning_rate": 7.244875756124693e-06, + "loss": 10.335901260375977, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.43421709910035133, + "loss_total_v6": 0.6459938287734985, + "step": 1655 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3819243243243243, + "grad_norm": 25.285991668701172, + "learning_rate": 7.241481549831244e-06, + "loss": 9.83165168762207, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.3946601338684559, + "loss_total_v6": 0.6144782602787018, + "step": 1656 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38215495495495494, + "grad_norm": 21.58058738708496, + "learning_rate": 7.238086050313563e-06, + "loss": 9.44214916229248, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.3802649527788162, + "loss_total_v6": 0.5901343449950218, + "step": 1657 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3823855855855856, + "grad_norm": 21.843341827392578, + "learning_rate": 7.23468925953069e-06, + "loss": 9.974051475524902, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.408763337880373, + "loss_total_v6": 0.6233782023191452, + "step": 1658 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3826162162162162, + "grad_norm": 20.087739944458008, + "learning_rate": 7.231291179442404e-06, + "loss": 10.486507415771484, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.4368092454969883, + "loss_total_v6": 0.6554066687822342, + "step": 1659 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38284684684684683, + "grad_norm": 21.344820022583008, + "learning_rate": 7.22789181200923e-06, + "loss": 10.297853469848633, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.42622392252087593, + "loss_total_v6": 0.6436158865690231, + "step": 1660 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3830774774774775, + "grad_norm": 20.320344924926758, + "learning_rate": 7.224491159192434e-06, + "loss": 9.843252182006836, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.40716494247317314, + "loss_total_v6": 0.6152032688260078, + "step": 1661 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3833081081081081, + "grad_norm": 21.952232360839844, + "learning_rate": 7.221089222954026e-06, + "loss": 10.284746170043945, + "loss_alignment": 0.439697265625, + "loss_alignment_w": 0.2198486328125, + "loss_sft": 0.42310062795877457, + "loss_total_v6": 0.6427966505289078, + "step": 1662 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3835387387387387, + "grad_norm": 19.80108070373535, + "learning_rate": 7.217686005256755e-06, + "loss": 9.63361930847168, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.3879440762102604, + "loss_total_v6": 0.602101169526577, + "step": 1663 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3837693693693694, + "grad_norm": 21.181488037109375, + "learning_rate": 7.214281508064107e-06, + "loss": 10.40047550201416, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4390311762690544, + "loss_total_v6": 0.6500297337770462, + "step": 1664 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.384, + "grad_norm": 21.539093017578125, + "learning_rate": 7.210875733340315e-06, + "loss": 9.815961837768555, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4000423885881901, + "loss_total_v6": 0.6134975999593735, + "step": 1665 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3842306306306306, + "grad_norm": 24.377424240112305, + "learning_rate": 7.207468683050338e-06, + "loss": 10.630206108093262, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.45203131064772606, + "loss_total_v6": 0.6643878817558289, + "step": 1666 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3844612612612613, + "grad_norm": 21.4544677734375, + "learning_rate": 7.204060359159877e-06, + "loss": 10.19334888458252, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.42451412230730057, + "loss_total_v6": 0.6370843052864075, + "step": 1667 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3846918918918919, + "grad_norm": 20.50316047668457, + "learning_rate": 7.200650763635366e-06, + "loss": 10.13707160949707, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.4261695183813572, + "loss_total_v6": 0.6335669942200184, + "step": 1668 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3849225225225225, + "grad_norm": 20.8129825592041, + "learning_rate": 7.1972398984439765e-06, + "loss": 9.640371322631836, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.39263859018683434, + "loss_total_v6": 0.6025232300162315, + "step": 1669 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3851531531531531, + "grad_norm": 20.472179412841797, + "learning_rate": 7.193827765553605e-06, + "loss": 10.138967514038086, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.42386187985539436, + "loss_total_v6": 0.6336855068802834, + "step": 1670 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3853837837837838, + "grad_norm": 21.088226318359375, + "learning_rate": 7.190414366932886e-06, + "loss": 9.5001220703125, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.38327787816524506, + "loss_total_v6": 0.5937576219439507, + "step": 1671 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3856144144144144, + "grad_norm": 25.235830307006836, + "learning_rate": 7.186999704551182e-06, + "loss": 10.02699089050293, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4129570499062538, + "loss_total_v6": 0.6266869008541107, + "step": 1672 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.385845045045045, + "grad_norm": 21.534156799316406, + "learning_rate": 7.183583780378585e-06, + "loss": 9.719364166259766, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.38773369789123535, + "loss_total_v6": 0.6074602529406548, + "step": 1673 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3860756756756757, + "grad_norm": 21.095335006713867, + "learning_rate": 7.180166596385915e-06, + "loss": 10.06271743774414, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4181806929409504, + "loss_total_v6": 0.6289198026061058, + "step": 1674 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3863063063063063, + "grad_norm": 23.0864315032959, + "learning_rate": 7.176748154544717e-06, + "loss": 9.7428617477417, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.4030725471675396, + "loss_total_v6": 0.6089288741350174, + "step": 1675 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3865369369369369, + "grad_norm": 21.49002456665039, + "learning_rate": 7.1733284568272635e-06, + "loss": 9.97767448425293, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4088066555559635, + "loss_total_v6": 0.6236046254634857, + "step": 1676 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3867675675675676, + "grad_norm": 23.0017147064209, + "learning_rate": 7.169907505206552e-06, + "loss": 9.88364315032959, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.40375369042158127, + "loss_total_v6": 0.6177277117967606, + "step": 1677 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3869981981981982, + "grad_norm": 19.764606475830078, + "learning_rate": 7.166485301656302e-06, + "loss": 10.326749801635742, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4293726943433285, + "loss_total_v6": 0.6454218849539757, + "step": 1678 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3872288288288288, + "grad_norm": 19.148462295532227, + "learning_rate": 7.163061848150959e-06, + "loss": 10.307969093322754, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4296179562807083, + "loss_total_v6": 0.6442480757832527, + "step": 1679 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3874594594594595, + "grad_norm": 20.992782592773438, + "learning_rate": 7.159637146665681e-06, + "loss": 10.514229774475098, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4440961219370365, + "loss_total_v6": 0.6571393460035324, + "step": 1680 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3876900900900901, + "grad_norm": 22.2800350189209, + "learning_rate": 7.156211199176357e-06, + "loss": 10.279863357543945, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.42870061472058296, + "loss_total_v6": 0.6424914747476578, + "step": 1681 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3879207207207207, + "grad_norm": 20.52387237548828, + "learning_rate": 7.1527840076595876e-06, + "loss": 9.929678916931152, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.4107355512678623, + "loss_total_v6": 0.620604932308197, + "step": 1682 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38815135135135137, + "grad_norm": 20.625709533691406, + "learning_rate": 7.149355574092691e-06, + "loss": 10.01156997680664, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.41461780294775963, + "loss_total_v6": 0.6257231459021568, + "step": 1683 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.388381981981982, + "grad_norm": 20.19415855407715, + "learning_rate": 7.14592590045371e-06, + "loss": 9.857421875, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.4065399318933487, + "loss_total_v6": 0.6160888820886612, + "step": 1684 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3886126126126126, + "grad_norm": 19.960269927978516, + "learning_rate": 7.14249498872139e-06, + "loss": 10.155975341796875, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4210796244442463, + "loss_total_v6": 0.6347484663128853, + "step": 1685 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38884324324324326, + "grad_norm": 22.809062957763672, + "learning_rate": 7.139062840875204e-06, + "loss": 10.757457733154297, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.45565101131796837, + "loss_total_v6": 0.6723410785198212, + "step": 1686 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3890738738738739, + "grad_norm": 21.899389266967773, + "learning_rate": 7.135629458895328e-06, + "loss": 10.476056098937988, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.43887219950556755, + "loss_total_v6": 0.6547535061836243, + "step": 1687 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3893045045045045, + "grad_norm": 20.78038215637207, + "learning_rate": 7.1321948447626545e-06, + "loss": 9.870845794677734, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.3997342512011528, + "loss_total_v6": 0.6169278472661972, + "step": 1688 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38953513513513516, + "grad_norm": 21.89365005493164, + "learning_rate": 7.128759000458785e-06, + "loss": 10.383350372314453, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4324524588882923, + "loss_total_v6": 0.6489594280719757, + "step": 1689 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.38976576576576577, + "grad_norm": 22.771194458007812, + "learning_rate": 7.125321927966037e-06, + "loss": 10.611708641052246, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.44609926640987396, + "loss_total_v6": 0.6632318198680878, + "step": 1690 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3899963963963964, + "grad_norm": 22.39482879638672, + "learning_rate": 7.121883629267428e-06, + "loss": 9.897571563720703, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.40473106876015663, + "loss_total_v6": 0.6185982674360275, + "step": 1691 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39022702702702705, + "grad_norm": 22.481355667114258, + "learning_rate": 7.118444106346687e-06, + "loss": 10.016063690185547, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4134032614529133, + "loss_total_v6": 0.6260039806365967, + "step": 1692 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39045765765765766, + "grad_norm": 21.577953338623047, + "learning_rate": 7.115003361188248e-06, + "loss": 10.420305252075195, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4354030154645443, + "loss_total_v6": 0.6512691080570221, + "step": 1693 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3906882882882883, + "grad_norm": 20.733154296875, + "learning_rate": 7.111561395777255e-06, + "loss": 9.810907363891602, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.40377000346779823, + "loss_total_v6": 0.6131816580891609, + "step": 1694 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39091891891891895, + "grad_norm": 23.158798217773438, + "learning_rate": 7.1081182120995486e-06, + "loss": 10.247934341430664, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.4251486174762249, + "loss_total_v6": 0.6404958963394165, + "step": 1695 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39114954954954956, + "grad_norm": 19.83929443359375, + "learning_rate": 7.104673812141676e-06, + "loss": 9.599525451660156, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.38608789816498756, + "loss_total_v6": 0.599970355629921, + "step": 1696 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39138018018018017, + "grad_norm": 20.711645126342773, + "learning_rate": 7.101228197890887e-06, + "loss": 10.143444061279297, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.42051010578870773, + "loss_total_v6": 0.6339653134346008, + "step": 1697 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3916108108108108, + "grad_norm": 19.27745819091797, + "learning_rate": 7.097781371335131e-06, + "loss": 9.42595100402832, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.3729812204837799, + "loss_total_v6": 0.5891219526529312, + "step": 1698 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39184144144144145, + "grad_norm": 19.20578384399414, + "learning_rate": 7.094333334463056e-06, + "loss": 9.846869468688965, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4052852690219879, + "loss_total_v6": 0.6154293045401573, + "step": 1699 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39207207207207206, + "grad_norm": 24.450939178466797, + "learning_rate": 7.090884089264012e-06, + "loss": 9.91400146484375, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.4047660604119301, + "loss_total_v6": 0.619625061750412, + "step": 1700 + }, + { + "epoch": 0.39207207207207206, + "eval_loss": 0.6695361137390137, + "eval_loss_alignment": 0.40612479967948717, + "eval_loss_alignment_w": 0.20306239983974358, + "eval_loss_sft": 0.46647367919357413, + "eval_loss_total_v6": 0.6695360782691556, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.0929, + "eval_samples_per_second": 22.73, + "eval_steps_per_second": 2.85, + "step": 1700 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3923027027027027, + "grad_norm": 23.15388298034668, + "learning_rate": 7.087433637728038e-06, + "loss": 9.801787376403809, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.40098757296800613, + "loss_total_v6": 0.6126117259263992, + "step": 1701 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39253333333333335, + "grad_norm": 23.053098678588867, + "learning_rate": 7.083981981845878e-06, + "loss": 9.84632682800293, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4056633748114109, + "loss_total_v6": 0.6153954192996025, + "step": 1702 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39276396396396396, + "grad_norm": 23.291786193847656, + "learning_rate": 7.080529123608967e-06, + "loss": 10.281728744506836, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.43089230731129646, + "loss_total_v6": 0.6426080130040646, + "step": 1703 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39299459459459457, + "grad_norm": 22.2185001373291, + "learning_rate": 7.0770750650094335e-06, + "loss": 10.494440078735352, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.44014326855540276, + "loss_total_v6": 0.6559025421738625, + "step": 1704 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39322522522522524, + "grad_norm": 21.79982566833496, + "learning_rate": 7.073619808040098e-06, + "loss": 10.263111114501953, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.42611242830753326, + "loss_total_v6": 0.6414444446563721, + "step": 1705 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39345585585585585, + "grad_norm": 21.2834415435791, + "learning_rate": 7.0701633546944725e-06, + "loss": 10.027973175048828, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.41463591530919075, + "loss_total_v6": 0.626748338341713, + "step": 1706 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39368648648648646, + "grad_norm": 22.02183723449707, + "learning_rate": 7.0667057069667625e-06, + "loss": 10.31828498840332, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.4410354010760784, + "loss_total_v6": 0.6448928043246269, + "step": 1707 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39391711711711713, + "grad_norm": 20.549036026000977, + "learning_rate": 7.0632468668518584e-06, + "loss": 9.856956481933594, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4038557782769203, + "loss_total_v6": 0.6160597503185272, + "step": 1708 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39414774774774775, + "grad_norm": 20.625288009643555, + "learning_rate": 7.059786836345339e-06, + "loss": 9.544082641601562, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.38738352432847023, + "loss_total_v6": 0.5965052023530006, + "step": 1709 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39437837837837836, + "grad_norm": 21.128454208374023, + "learning_rate": 7.056325617443475e-06, + "loss": 10.06423568725586, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4138505160808563, + "loss_total_v6": 0.6290147006511688, + "step": 1710 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.394609009009009, + "grad_norm": 21.731788635253906, + "learning_rate": 7.052863212143215e-06, + "loss": 9.5707426071167, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.3934442400932312, + "loss_total_v6": 0.5981713943183422, + "step": 1711 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39483963963963964, + "grad_norm": 21.07438087463379, + "learning_rate": 7.049399622442199e-06, + "loss": 10.239059448242188, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.4224119856953621, + "loss_total_v6": 0.6399412602186203, + "step": 1712 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39507027027027025, + "grad_norm": 21.587574005126953, + "learning_rate": 7.045934850338745e-06, + "loss": 9.837709426879883, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.3979684188961983, + "loss_total_v6": 0.6148568540811539, + "step": 1713 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3953009009009009, + "grad_norm": 21.046585083007812, + "learning_rate": 7.0424688978318576e-06, + "loss": 9.955256462097168, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.4145924486219883, + "loss_total_v6": 0.6222035214304924, + "step": 1714 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39553153153153153, + "grad_norm": 23.481090545654297, + "learning_rate": 7.0390017669212175e-06, + "loss": 9.605696678161621, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.383925374597311, + "loss_total_v6": 0.6003560498356819, + "step": 1715 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39576216216216215, + "grad_norm": 21.106407165527344, + "learning_rate": 7.035533459607189e-06, + "loss": 9.739713668823242, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.39733684808015823, + "loss_total_v6": 0.6087321043014526, + "step": 1716 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3959927927927928, + "grad_norm": 23.000059127807617, + "learning_rate": 7.032063977890817e-06, + "loss": 10.190730094909668, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.43063706159591675, + "loss_total_v6": 0.6369206234812737, + "step": 1717 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3962234234234234, + "grad_norm": 21.565095901489258, + "learning_rate": 7.028593323773819e-06, + "loss": 10.202516555786133, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4187393896281719, + "loss_total_v6": 0.6376572549343109, + "step": 1718 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39645405405405404, + "grad_norm": 21.33336067199707, + "learning_rate": 7.025121499258591e-06, + "loss": 9.780213356018066, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.40009698271751404, + "loss_total_v6": 0.6112633571028709, + "step": 1719 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3966846846846847, + "grad_norm": 23.796613693237305, + "learning_rate": 7.0216485063482045e-06, + "loss": 9.990262985229492, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.40783869847655296, + "loss_total_v6": 0.624391458928585, + "step": 1720 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3969153153153153, + "grad_norm": 20.870861053466797, + "learning_rate": 7.018174347046406e-06, + "loss": 10.02116584777832, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.4156295023858547, + "loss_total_v6": 0.6263228580355644, + "step": 1721 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39714594594594593, + "grad_norm": 21.848407745361328, + "learning_rate": 7.014699023357612e-06, + "loss": 9.69301986694336, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.3980805240571499, + "loss_total_v6": 0.6058137118816376, + "step": 1722 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3973765765765766, + "grad_norm": 23.121896743774414, + "learning_rate": 7.011222537286914e-06, + "loss": 9.875457763671875, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4056224524974823, + "loss_total_v6": 0.6172160729765892, + "step": 1723 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3976072072072072, + "grad_norm": 21.14396858215332, + "learning_rate": 7.0077448908400736e-06, + "loss": 10.665639877319336, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4543374888598919, + "loss_total_v6": 0.6666024774312973, + "step": 1724 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3978378378378378, + "grad_norm": 21.89280128479004, + "learning_rate": 7.004266086023521e-06, + "loss": 9.844009399414062, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.4113779664039612, + "loss_total_v6": 0.6152506172657013, + "step": 1725 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3980684684684685, + "grad_norm": 19.89909553527832, + "learning_rate": 7.000786124844355e-06, + "loss": 9.733097076416016, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.39745737612247467, + "loss_total_v6": 0.6083185821771622, + "step": 1726 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3982990990990991, + "grad_norm": 20.670475006103516, + "learning_rate": 6.997305009310338e-06, + "loss": 10.724991798400879, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.45957284048199654, + "loss_total_v6": 0.6703119948506355, + "step": 1727 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3985297297297297, + "grad_norm": 22.28375244140625, + "learning_rate": 6.9938227414299066e-06, + "loss": 10.134288787841797, + "loss_alignment": 0.408935546875, + "loss_alignment_w": 0.2044677734375, + "loss_sft": 0.4285590872168541, + "loss_total_v6": 0.6333930492401123, + "step": 1728 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.39876036036036033, + "grad_norm": 21.134441375732422, + "learning_rate": 6.990339323212154e-06, + "loss": 9.973682403564453, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.4168427512049675, + "loss_total_v6": 0.6233551874756813, + "step": 1729 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.398990990990991, + "grad_norm": 20.441364288330078, + "learning_rate": 6.986854756666843e-06, + "loss": 9.841560363769531, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4057621769607067, + "loss_total_v6": 0.6150975078344345, + "step": 1730 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3992216216216216, + "grad_norm": 22.22916603088379, + "learning_rate": 6.983369043804396e-06, + "loss": 10.285877227783203, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4319450594484806, + "loss_total_v6": 0.642867311835289, + "step": 1731 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3994522522522522, + "grad_norm": 20.296710968017578, + "learning_rate": 6.979882186635897e-06, + "loss": 9.615880012512207, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.39185553416609764, + "loss_total_v6": 0.6009924933314323, + "step": 1732 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3996828828828829, + "grad_norm": 20.79090690612793, + "learning_rate": 6.97639418717309e-06, + "loss": 9.076069831848145, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.35946018621325493, + "loss_total_v6": 0.5672543607652187, + "step": 1733 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.3999135135135135, + "grad_norm": 20.40806007385254, + "learning_rate": 6.97290504742838e-06, + "loss": 10.443872451782227, + "loss_alignment": 0.40234375, + "loss_alignment_w": 0.201171875, + "loss_sft": 0.45176849886775017, + "loss_total_v6": 0.6527420207858086, + "step": 1734 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4001441441441441, + "grad_norm": 22.339658737182617, + "learning_rate": 6.969414769414826e-06, + "loss": 10.08262825012207, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.4096289612352848, + "loss_total_v6": 0.63016427308321, + "step": 1735 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4003747747747748, + "grad_norm": 21.52486801147461, + "learning_rate": 6.965923355146148e-06, + "loss": 10.32895278930664, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.43472884967923164, + "loss_total_v6": 0.6455595418810844, + "step": 1736 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4006054054054054, + "grad_norm": 22.187654495239258, + "learning_rate": 6.962430806636719e-06, + "loss": 10.272688865661621, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.4209432005882263, + "loss_total_v6": 0.6420430392026901, + "step": 1737 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.400836036036036, + "grad_norm": 22.120996475219727, + "learning_rate": 6.958937125901568e-06, + "loss": 10.093647003173828, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.41460541263222694, + "loss_total_v6": 0.6308529824018478, + "step": 1738 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4010666666666667, + "grad_norm": 21.950162887573242, + "learning_rate": 6.955442314956375e-06, + "loss": 9.796440124511719, + "loss_alignment": 0.40869140625, + "loss_alignment_w": 0.204345703125, + "loss_sft": 0.4083743281662464, + "loss_total_v6": 0.612277515232563, + "step": 1739 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4012972972972973, + "grad_norm": 21.603137969970703, + "learning_rate": 6.9519463758174745e-06, + "loss": 9.891551971435547, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.3999755010008812, + "loss_total_v6": 0.6182219535112381, + "step": 1740 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4015279279279279, + "grad_norm": 20.557342529296875, + "learning_rate": 6.9484493105018515e-06, + "loss": 9.995864868164062, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4136972054839134, + "loss_total_v6": 0.6247415021061897, + "step": 1741 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4017585585585586, + "grad_norm": 21.481739044189453, + "learning_rate": 6.944951121027137e-06, + "loss": 10.221580505371094, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.42406610772013664, + "loss_total_v6": 0.6388488113880157, + "step": 1742 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4019891891891892, + "grad_norm": 22.133947372436523, + "learning_rate": 6.941451809411618e-06, + "loss": 10.29902458190918, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.43403320759534836, + "loss_total_v6": 0.6436890065670013, + "step": 1743 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4022198198198198, + "grad_norm": 20.79911994934082, + "learning_rate": 6.937951377674222e-06, + "loss": 9.738337516784668, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.39020123705267906, + "loss_total_v6": 0.6086460649967194, + "step": 1744 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40245045045045047, + "grad_norm": 22.539670944213867, + "learning_rate": 6.9344498278345244e-06, + "loss": 10.468364715576172, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.44525788351893425, + "loss_total_v6": 0.6542727574706078, + "step": 1745 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4026810810810811, + "grad_norm": 21.199983596801758, + "learning_rate": 6.930947161912749e-06, + "loss": 9.76429271697998, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.3955161049962044, + "loss_total_v6": 0.6102683022618294, + "step": 1746 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4029117117117117, + "grad_norm": 20.45973014831543, + "learning_rate": 6.927443381929761e-06, + "loss": 10.331687927246094, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.43673085048794746, + "loss_total_v6": 0.6457305029034615, + "step": 1747 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40314234234234236, + "grad_norm": 21.139598846435547, + "learning_rate": 6.9239384899070665e-06, + "loss": 10.121137619018555, + "loss_alignment": 0.413330078125, + "loss_alignment_w": 0.2066650390625, + "loss_sft": 0.42566194012761116, + "loss_total_v6": 0.6325711235404015, + "step": 1748 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.403372972972973, + "grad_norm": 21.74677276611328, + "learning_rate": 6.920432487866816e-06, + "loss": 10.438299179077148, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4360698349773884, + "loss_total_v6": 0.6523936986923218, + "step": 1749 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4036036036036036, + "grad_norm": 20.478408813476562, + "learning_rate": 6.916925377831801e-06, + "loss": 9.68729019165039, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.3915579505264759, + "loss_total_v6": 0.6054556369781494, + "step": 1750 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40383423423423426, + "grad_norm": 22.903165817260742, + "learning_rate": 6.913417161825449e-06, + "loss": 10.023299217224121, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.41403861716389656, + "loss_total_v6": 0.626456193625927, + "step": 1751 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40406486486486487, + "grad_norm": 25.22547149658203, + "learning_rate": 6.90990784187183e-06, + "loss": 9.99439525604248, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.41067568585276604, + "loss_total_v6": 0.624649703502655, + "step": 1752 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4042954954954955, + "grad_norm": 22.90288734436035, + "learning_rate": 6.9063974199956454e-06, + "loss": 9.936464309692383, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4125329479575157, + "loss_total_v6": 0.6210290566086769, + "step": 1753 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40452612612612615, + "grad_norm": 22.351259231567383, + "learning_rate": 6.902885898222238e-06, + "loss": 10.46898078918457, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.43945223465561867, + "loss_total_v6": 0.6543112769722939, + "step": 1754 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40475675675675676, + "grad_norm": 26.518115997314453, + "learning_rate": 6.89937327857758e-06, + "loss": 10.120171546936035, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4186435453593731, + "loss_total_v6": 0.6325107142329216, + "step": 1755 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4049873873873874, + "grad_norm": 19.455211639404297, + "learning_rate": 6.895859563088283e-06, + "loss": 9.811598777770996, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.40527811646461487, + "loss_total_v6": 0.6132249012589455, + "step": 1756 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40521801801801804, + "grad_norm": 23.096548080444336, + "learning_rate": 6.8923447537815845e-06, + "loss": 10.444427490234375, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.43947409465909004, + "loss_total_v6": 0.6527766957879066, + "step": 1757 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40544864864864866, + "grad_norm": 20.44908905029297, + "learning_rate": 6.8888288526853564e-06, + "loss": 10.065004348754883, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.4192238934338093, + "loss_total_v6": 0.6290627866983414, + "step": 1758 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40567927927927927, + "grad_norm": 26.646320343017578, + "learning_rate": 6.8853118618281014e-06, + "loss": 9.751762390136719, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.3967471122741699, + "loss_total_v6": 0.6094851642847061, + "step": 1759 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4059099099099099, + "grad_norm": 22.75389289855957, + "learning_rate": 6.881793783238948e-06, + "loss": 9.74793815612793, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.3997429385781288, + "loss_total_v6": 0.6092461198568344, + "step": 1760 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40614054054054055, + "grad_norm": 23.242109298706055, + "learning_rate": 6.8782746189476544e-06, + "loss": 10.567176818847656, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.450823288410902, + "loss_total_v6": 0.6604485213756561, + "step": 1761 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40637117117117116, + "grad_norm": 22.1700496673584, + "learning_rate": 6.8747543709846064e-06, + "loss": 9.522353172302246, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.382729459553957, + "loss_total_v6": 0.5951470583677292, + "step": 1762 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4066018018018018, + "grad_norm": 21.301271438598633, + "learning_rate": 6.871233041380812e-06, + "loss": 10.469164848327637, + "loss_alignment": 0.443359375, + "loss_alignment_w": 0.2216796875, + "loss_sft": 0.43288731202483177, + "loss_total_v6": 0.6543228402733803, + "step": 1763 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40683243243243244, + "grad_norm": 20.964927673339844, + "learning_rate": 6.867710632167903e-06, + "loss": 10.306070327758789, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4332529567182064, + "loss_total_v6": 0.6441294327378273, + "step": 1764 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40706306306306306, + "grad_norm": 21.520395278930664, + "learning_rate": 6.864187145378138e-06, + "loss": 9.914735794067383, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4074364863336086, + "loss_total_v6": 0.619670957326889, + "step": 1765 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40729369369369367, + "grad_norm": 24.083595275878906, + "learning_rate": 6.860662583044395e-06, + "loss": 9.921821594238281, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.4111142046749592, + "loss_total_v6": 0.620113842189312, + "step": 1766 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40752432432432434, + "grad_norm": 22.250120162963867, + "learning_rate": 6.857136947200171e-06, + "loss": 9.83668327331543, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.4037787988781929, + "loss_total_v6": 0.6147926598787308, + "step": 1767 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40775495495495495, + "grad_norm": 20.74062728881836, + "learning_rate": 6.853610239879586e-06, + "loss": 9.94578742980957, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.40917886421084404, + "loss_total_v6": 0.6216117292642593, + "step": 1768 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40798558558558556, + "grad_norm": 18.8301944732666, + "learning_rate": 6.850082463117374e-06, + "loss": 9.600973129272461, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.3943265229463577, + "loss_total_v6": 0.6000607758760452, + "step": 1769 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40821621621621623, + "grad_norm": 20.644376754760742, + "learning_rate": 6.846553618948889e-06, + "loss": 10.203798294067383, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.4206200875341892, + "loss_total_v6": 0.6377373784780502, + "step": 1770 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40844684684684684, + "grad_norm": 22.7515869140625, + "learning_rate": 6.8430237094100985e-06, + "loss": 10.326476097106934, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.43307870626449585, + "loss_total_v6": 0.6454047337174416, + "step": 1771 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40867747747747746, + "grad_norm": 19.693294525146484, + "learning_rate": 6.839492736537588e-06, + "loss": 9.430023193359375, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.3798428066074848, + "loss_total_v6": 0.5893764868378639, + "step": 1772 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4089081081081081, + "grad_norm": 21.733064651489258, + "learning_rate": 6.8359607023685544e-06, + "loss": 9.992175102233887, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.4066306836903095, + "loss_total_v6": 0.6245109289884567, + "step": 1773 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40913873873873874, + "grad_norm": 23.79988670349121, + "learning_rate": 6.832427608940806e-06, + "loss": 10.038705825805664, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4170919917523861, + "loss_total_v6": 0.6274191439151764, + "step": 1774 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40936936936936935, + "grad_norm": 21.206850051879883, + "learning_rate": 6.828893458292765e-06, + "loss": 9.978303909301758, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.40771689638495445, + "loss_total_v6": 0.6236440166831017, + "step": 1775 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4096, + "grad_norm": 21.95013999938965, + "learning_rate": 6.82535825246346e-06, + "loss": 10.330924987792969, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.43004558235406876, + "loss_total_v6": 0.64568280428648, + "step": 1776 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.40983063063063063, + "grad_norm": 23.044017791748047, + "learning_rate": 6.821821993492532e-06, + "loss": 10.339614868164062, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4326944798231125, + "loss_total_v6": 0.6462259739637375, + "step": 1777 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41006126126126125, + "grad_norm": 19.989910125732422, + "learning_rate": 6.818284683420227e-06, + "loss": 10.142721176147461, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4217313155531883, + "loss_total_v6": 0.6339200735092163, + "step": 1778 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4102918918918919, + "grad_norm": 21.64034080505371, + "learning_rate": 6.814746324287397e-06, + "loss": 9.623626708984375, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.38985250145196915, + "loss_total_v6": 0.6014766581356525, + "step": 1779 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4105225225225225, + "grad_norm": 20.784547805786133, + "learning_rate": 6.811206918135502e-06, + "loss": 9.714040756225586, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.3870958276093006, + "loss_total_v6": 0.6071275472640991, + "step": 1780 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41075315315315314, + "grad_norm": 23.426069259643555, + "learning_rate": 6.807666467006605e-06, + "loss": 10.73286247253418, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.45863044261932373, + "loss_total_v6": 0.6708039119839668, + "step": 1781 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4109837837837838, + "grad_norm": 21.728748321533203, + "learning_rate": 6.804124972943372e-06, + "loss": 9.717960357666016, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.3993647135794163, + "loss_total_v6": 0.6073725521564484, + "step": 1782 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4112144144144144, + "grad_norm": 22.97571563720703, + "learning_rate": 6.80058243798907e-06, + "loss": 9.82974910736084, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.4036659859120846, + "loss_total_v6": 0.6143593341112137, + "step": 1783 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41144504504504503, + "grad_norm": 20.178804397583008, + "learning_rate": 6.797038864187564e-06, + "loss": 9.841206550598145, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.39966707676649094, + "loss_total_v6": 0.6150754243135452, + "step": 1784 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4116756756756757, + "grad_norm": 18.51803207397461, + "learning_rate": 6.793494253583326e-06, + "loss": 9.40097427368164, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.37975144386291504, + "loss_total_v6": 0.5875609144568443, + "step": 1785 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4119063063063063, + "grad_norm": 20.700403213500977, + "learning_rate": 6.789948608221418e-06, + "loss": 10.099455833435059, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.42064471542835236, + "loss_total_v6": 0.631216011941433, + "step": 1786 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4121369369369369, + "grad_norm": 21.817974090576172, + "learning_rate": 6.786401930147502e-06, + "loss": 10.095678329467773, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.42626798152923584, + "loss_total_v6": 0.6309799179434776, + "step": 1787 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4123675675675676, + "grad_norm": 21.889007568359375, + "learning_rate": 6.7828542214078385e-06, + "loss": 9.984978675842285, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4105907306075096, + "loss_total_v6": 0.6240611895918846, + "step": 1788 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4125981981981982, + "grad_norm": 22.38749885559082, + "learning_rate": 6.77930548404928e-06, + "loss": 10.182246208190918, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.42679568752646446, + "loss_total_v6": 0.636390395462513, + "step": 1789 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4128288288288288, + "grad_norm": 22.481115341186523, + "learning_rate": 6.775755720119272e-06, + "loss": 9.616300582885742, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.3858393728733063, + "loss_total_v6": 0.6010188236832619, + "step": 1790 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41305945945945943, + "grad_norm": 20.61350440979004, + "learning_rate": 6.772204931665853e-06, + "loss": 10.081753730773926, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.42651155963540077, + "loss_total_v6": 0.6301095858216286, + "step": 1791 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4132900900900901, + "grad_norm": 22.138469696044922, + "learning_rate": 6.7686531207376525e-06, + "loss": 9.858861923217773, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.40688933432102203, + "loss_total_v6": 0.6161788776516914, + "step": 1792 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4135207207207207, + "grad_norm": 20.164743423461914, + "learning_rate": 6.765100289383891e-06, + "loss": 10.714357376098633, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.45477310568094254, + "loss_total_v6": 0.6696473732590675, + "step": 1793 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4137513513513513, + "grad_norm": 22.796175003051758, + "learning_rate": 6.761546439654376e-06, + "loss": 9.830497741699219, + "loss_alignment": 0.435546875, + "loss_alignment_w": 0.2177734375, + "loss_sft": 0.39681578427553177, + "loss_total_v6": 0.6144060865044594, + "step": 1794 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.413981981981982, + "grad_norm": 20.95368194580078, + "learning_rate": 6.757991573599504e-06, + "loss": 9.633018493652344, + "loss_alignment": 0.402099609375, + "loss_alignment_w": 0.2010498046875, + "loss_sft": 0.4011969789862633, + "loss_total_v6": 0.6020636633038521, + "step": 1795 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4142126126126126, + "grad_norm": 19.658973693847656, + "learning_rate": 6.754435693270258e-06, + "loss": 9.724347114562988, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.39268379658460617, + "loss_total_v6": 0.6077716872096062, + "step": 1796 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4144432432432432, + "grad_norm": 21.399436950683594, + "learning_rate": 6.750878800718203e-06, + "loss": 10.117944717407227, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.41349945217370987, + "loss_total_v6": 0.6323715299367905, + "step": 1797 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4146738738738739, + "grad_norm": 28.87093734741211, + "learning_rate": 6.747320897995493e-06, + "loss": 9.398919105529785, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.3753963112831116, + "loss_total_v6": 0.5874324440956116, + "step": 1798 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4149045045045045, + "grad_norm": 24.112518310546875, + "learning_rate": 6.743761987154861e-06, + "loss": 9.873348236083984, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4069402366876602, + "loss_total_v6": 0.6170842722058296, + "step": 1799 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4151351351351351, + "grad_norm": 24.337783813476562, + "learning_rate": 6.740202070249623e-06, + "loss": 10.136738777160645, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.415177620947361, + "loss_total_v6": 0.6335461586713791, + "step": 1800 + }, + { + "epoch": 0.4151351351351351, + "eval_loss": 0.666817843914032, + "eval_loss_alignment": 0.4065233540331197, + "eval_loss_alignment_w": 0.20326167701655984, + "eval_loss_sft": 0.46355611339816427, + "eval_loss_total_v6": 0.666817790390844, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.4782, + "eval_samples_per_second": 22.624, + "eval_steps_per_second": 2.837, + "step": 1800 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4153657657657658, + "grad_norm": 21.518877029418945, + "learning_rate": 6.7366411493336735e-06, + "loss": 10.090509414672852, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.416881263256073, + "loss_total_v6": 0.6306568682193756, + "step": 1801 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4155963963963964, + "grad_norm": 20.793737411499023, + "learning_rate": 6.733079226461493e-06, + "loss": 10.905384063720703, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.46671221777796745, + "loss_total_v6": 0.6815865114331245, + "step": 1802 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.415827027027027, + "grad_norm": 20.755001068115234, + "learning_rate": 6.729516303688133e-06, + "loss": 10.103266716003418, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.41836516186594963, + "loss_total_v6": 0.631454162299633, + "step": 1803 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4160576576576577, + "grad_norm": 22.205778121948242, + "learning_rate": 6.725952383069222e-06, + "loss": 10.27733039855957, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4266349859535694, + "loss_total_v6": 0.6423331946134567, + "step": 1804 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4162882882882883, + "grad_norm": 19.959516525268555, + "learning_rate": 6.722387466660966e-06, + "loss": 9.910280227661133, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.4096909686923027, + "loss_total_v6": 0.6193925142288208, + "step": 1805 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4165189189189189, + "grad_norm": 22.27076530456543, + "learning_rate": 6.718821556520151e-06, + "loss": 9.7078857421875, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.39606475085020065, + "loss_total_v6": 0.6067428588867188, + "step": 1806 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41674954954954957, + "grad_norm": 20.868297576904297, + "learning_rate": 6.715254654704126e-06, + "loss": 9.812887191772461, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.3984922133386135, + "loss_total_v6": 0.6133054494857788, + "step": 1807 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4169801801801802, + "grad_norm": 21.223678588867188, + "learning_rate": 6.7116867632708184e-06, + "loss": 9.495721817016602, + "loss_alignment": 0.401611328125, + "loss_alignment_w": 0.2008056640625, + "loss_sft": 0.3920665830373764, + "loss_total_v6": 0.5934825912117958, + "step": 1808 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4172108108108108, + "grad_norm": 20.76779556274414, + "learning_rate": 6.7081178842787275e-06, + "loss": 9.819210052490234, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.40496044233441353, + "loss_total_v6": 0.6137006804347038, + "step": 1809 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41744144144144146, + "grad_norm": 22.11134147644043, + "learning_rate": 6.704548019786917e-06, + "loss": 9.856653213500977, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.4098335690796375, + "loss_total_v6": 0.6160408332943916, + "step": 1810 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4176720720720721, + "grad_norm": 20.289133071899414, + "learning_rate": 6.700977171855027e-06, + "loss": 9.448471069335938, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.3797903172671795, + "loss_total_v6": 0.5905294716358185, + "step": 1811 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4179027027027027, + "grad_norm": 19.821786880493164, + "learning_rate": 6.697405342543257e-06, + "loss": 10.042478561401367, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.40963729843497276, + "loss_total_v6": 0.6276548802852631, + "step": 1812 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41813333333333336, + "grad_norm": 20.71549415588379, + "learning_rate": 6.693832533912378e-06, + "loss": 10.519948959350586, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4443773366510868, + "loss_total_v6": 0.657496839761734, + "step": 1813 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41836396396396397, + "grad_norm": 22.928390502929688, + "learning_rate": 6.69025874802372e-06, + "loss": 10.591987609863281, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.443798515945673, + "loss_total_v6": 0.6619992330670357, + "step": 1814 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4185945945945946, + "grad_norm": 23.53972625732422, + "learning_rate": 6.6866839869391884e-06, + "loss": 10.085412979125977, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.41842425242066383, + "loss_total_v6": 0.6303383484482765, + "step": 1815 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41882522522522525, + "grad_norm": 25.123342514038086, + "learning_rate": 6.683108252721239e-06, + "loss": 10.27530288696289, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.4392797686159611, + "loss_total_v6": 0.6422064155340195, + "step": 1816 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41905585585585586, + "grad_norm": 20.508752822875977, + "learning_rate": 6.679531547432896e-06, + "loss": 9.569192886352539, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.3833376131951809, + "loss_total_v6": 0.5980745553970337, + "step": 1817 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4192864864864865, + "grad_norm": 20.14470672607422, + "learning_rate": 6.6759538731377385e-06, + "loss": 9.912087440490723, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.40647751092910767, + "loss_total_v6": 0.6195054426789284, + "step": 1818 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41951711711711714, + "grad_norm": 22.10575294494629, + "learning_rate": 6.672375231899913e-06, + "loss": 9.706106185913086, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.39851704984903336, + "loss_total_v6": 0.6066316738724709, + "step": 1819 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41974774774774776, + "grad_norm": 21.2501277923584, + "learning_rate": 6.668795625784114e-06, + "loss": 9.488143920898438, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.38306331261992455, + "loss_total_v6": 0.5930089987814426, + "step": 1820 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.41997837837837837, + "grad_norm": 23.31720733642578, + "learning_rate": 6.6652150568556e-06, + "loss": 9.91982650756836, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.40381792187690735, + "loss_total_v6": 0.6199891939759254, + "step": 1821 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.420209009009009, + "grad_norm": 21.743877410888672, + "learning_rate": 6.661633527180184e-06, + "loss": 9.858848571777344, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4070105142891407, + "loss_total_v6": 0.6161780282855034, + "step": 1822 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42043963963963965, + "grad_norm": 21.288177490234375, + "learning_rate": 6.6580510388242294e-06, + "loss": 9.620794296264648, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.3900722600519657, + "loss_total_v6": 0.6012996658682823, + "step": 1823 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42067027027027026, + "grad_norm": 24.839923858642578, + "learning_rate": 6.654467593854656e-06, + "loss": 10.109894752502441, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.4229603223502636, + "loss_total_v6": 0.631868414580822, + "step": 1824 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4209009009009009, + "grad_norm": 20.816347122192383, + "learning_rate": 6.650883194338935e-06, + "loss": 10.080276489257812, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.413708683103323, + "loss_total_v6": 0.6300172731280327, + "step": 1825 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42113153153153154, + "grad_norm": 23.524803161621094, + "learning_rate": 6.6472978423450874e-06, + "loss": 10.214330673217773, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42526085302233696, + "loss_total_v6": 0.6383956521749496, + "step": 1826 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42136216216216216, + "grad_norm": 22.98752784729004, + "learning_rate": 6.643711539941684e-06, + "loss": 9.739909172058105, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.3971201553940773, + "loss_total_v6": 0.6087442934513092, + "step": 1827 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42159279279279277, + "grad_norm": 21.110998153686523, + "learning_rate": 6.640124289197845e-06, + "loss": 9.888035774230957, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.3994048424065113, + "loss_total_v6": 0.6180022656917572, + "step": 1828 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42182342342342344, + "grad_norm": 21.971649169921875, + "learning_rate": 6.636536092183239e-06, + "loss": 10.006904602050781, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4137006066739559, + "loss_total_v6": 0.6254315376281738, + "step": 1829 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42205405405405405, + "grad_norm": 21.99205780029297, + "learning_rate": 6.6329469509680756e-06, + "loss": 9.624263763427734, + "loss_alignment": 0.411376953125, + "loss_alignment_w": 0.2056884765625, + "loss_sft": 0.3960264064371586, + "loss_total_v6": 0.6015165001153946, + "step": 1830 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42228468468468466, + "grad_norm": 21.151256561279297, + "learning_rate": 6.629356867623112e-06, + "loss": 10.409934997558594, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4405226744711399, + "loss_total_v6": 0.6506209224462509, + "step": 1831 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42251531531531533, + "grad_norm": 22.227741241455078, + "learning_rate": 6.625765844219651e-06, + "loss": 9.94495964050293, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.40476304292678833, + "loss_total_v6": 0.6215599253773689, + "step": 1832 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42274594594594594, + "grad_norm": 20.357336044311523, + "learning_rate": 6.622173882829536e-06, + "loss": 10.131935119628906, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.42040106654167175, + "loss_total_v6": 0.6332459002733231, + "step": 1833 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42297657657657656, + "grad_norm": 25.17303466796875, + "learning_rate": 6.618580985525149e-06, + "loss": 9.957271575927734, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.4164121635258198, + "loss_total_v6": 0.6223295032978058, + "step": 1834 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4232072072072072, + "grad_norm": 21.645057678222656, + "learning_rate": 6.614987154379417e-06, + "loss": 9.960010528564453, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4102204069495201, + "loss_total_v6": 0.6225006580352783, + "step": 1835 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42343783783783784, + "grad_norm": 19.645296096801758, + "learning_rate": 6.611392391465802e-06, + "loss": 9.815841674804688, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4027203992009163, + "loss_total_v6": 0.6134900599718094, + "step": 1836 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42366846846846845, + "grad_norm": 20.374927520751953, + "learning_rate": 6.6077966988583054e-06, + "loss": 9.986248016357422, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.41955066472291946, + "loss_total_v6": 0.6241405233740807, + "step": 1837 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4238990990990991, + "grad_norm": 19.26839256286621, + "learning_rate": 6.604200078631464e-06, + "loss": 10.214776992797852, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4254871718585491, + "loss_total_v6": 0.6384235471487045, + "step": 1838 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42412972972972973, + "grad_norm": 22.282764434814453, + "learning_rate": 6.600602532860349e-06, + "loss": 9.586507797241211, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.39641323685646057, + "loss_total_v6": 0.5991567559540272, + "step": 1839 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42436036036036034, + "grad_norm": 21.506587982177734, + "learning_rate": 6.597004063620568e-06, + "loss": 10.299933433532715, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4305195212364197, + "loss_total_v6": 0.6437458246946335, + "step": 1840 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.424590990990991, + "grad_norm": 21.928050994873047, + "learning_rate": 6.593404672988261e-06, + "loss": 10.241212844848633, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.42436231672763824, + "loss_total_v6": 0.6400758177042007, + "step": 1841 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4248216216216216, + "grad_norm": 20.995075225830078, + "learning_rate": 6.5898043630400975e-06, + "loss": 10.15925121307373, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42194047942757607, + "loss_total_v6": 0.6349531933665276, + "step": 1842 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42505225225225224, + "grad_norm": 22.581037521362305, + "learning_rate": 6.58620313585328e-06, + "loss": 10.374526023864746, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.43049709126353264, + "loss_total_v6": 0.6484078764915466, + "step": 1843 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4252828828828829, + "grad_norm": 21.927412033081055, + "learning_rate": 6.582600993505535e-06, + "loss": 9.559968948364258, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.3787175491452217, + "loss_total_v6": 0.5974980592727661, + "step": 1844 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4255135135135135, + "grad_norm": 21.41312026977539, + "learning_rate": 6.578997938075126e-06, + "loss": 10.183439254760742, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.4263514392077923, + "loss_total_v6": 0.6364649683237076, + "step": 1845 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42574414414414413, + "grad_norm": 21.650209426879883, + "learning_rate": 6.575393971640833e-06, + "loss": 9.520216941833496, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.3845338299870491, + "loss_total_v6": 0.5950135588645935, + "step": 1846 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4259747747747748, + "grad_norm": 20.08131980895996, + "learning_rate": 6.571789096281973e-06, + "loss": 9.779972076416016, + "loss_alignment": 0.4130859375, + "loss_alignment_w": 0.20654296875, + "loss_sft": 0.4048120565712452, + "loss_total_v6": 0.6112482100725174, + "step": 1847 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4262054054054054, + "grad_norm": 21.925155639648438, + "learning_rate": 6.568183314078377e-06, + "loss": 9.986981391906738, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.40703850612044334, + "loss_total_v6": 0.6241863295435905, + "step": 1848 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.426436036036036, + "grad_norm": 20.83466339111328, + "learning_rate": 6.564576627110401e-06, + "loss": 9.658401489257812, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.3961610645055771, + "loss_total_v6": 0.6036500930786133, + "step": 1849 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4266666666666667, + "grad_norm": 21.753753662109375, + "learning_rate": 6.560969037458933e-06, + "loss": 10.051773071289062, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4156961031258106, + "loss_total_v6": 0.628235787153244, + "step": 1850 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4268972972972973, + "grad_norm": 19.874418258666992, + "learning_rate": 6.557360547205367e-06, + "loss": 10.189273834228516, + "loss_alignment": 0.4443359375, + "loss_alignment_w": 0.22216796875, + "loss_sft": 0.4148905463516712, + "loss_total_v6": 0.6368296146392822, + "step": 1851 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4271279279279279, + "grad_norm": 22.27451515197754, + "learning_rate": 6.553751158431627e-06, + "loss": 9.962706565856934, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.41221991553902626, + "loss_total_v6": 0.6226691380143166, + "step": 1852 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.42735855855855853, + "grad_norm": 21.25551986694336, + "learning_rate": 6.550140873220151e-06, + "loss": 10.078742980957031, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.4157642833888531, + "loss_total_v6": 0.6299214139580727, + "step": 1853 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4275891891891892, + "grad_norm": 21.258913040161133, + "learning_rate": 6.546529693653897e-06, + "loss": 10.598114967346191, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.44436462968587875, + "loss_total_v6": 0.6623822078108788, + "step": 1854 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4278198198198198, + "grad_norm": 20.424131393432617, + "learning_rate": 6.5429176218163356e-06, + "loss": 10.116560935974121, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.4271763935685158, + "loss_total_v6": 0.632285051047802, + "step": 1855 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4280504504504504, + "grad_norm": 19.23798370361328, + "learning_rate": 6.539304659791457e-06, + "loss": 9.547922134399414, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.38863057643175125, + "loss_total_v6": 0.5967451855540276, + "step": 1856 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4282810810810811, + "grad_norm": 20.494909286499023, + "learning_rate": 6.535690809663757e-06, + "loss": 9.426617622375488, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.3863132633268833, + "loss_total_v6": 0.5891636051237583, + "step": 1857 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4285117117117117, + "grad_norm": 19.08537483215332, + "learning_rate": 6.532076073518253e-06, + "loss": 9.939550399780273, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4113677479326725, + "loss_total_v6": 0.6212218999862671, + "step": 1858 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4287423423423423, + "grad_norm": 20.03890609741211, + "learning_rate": 6.528460453440466e-06, + "loss": 9.613290786743164, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.39695800095796585, + "loss_total_v6": 0.600830689072609, + "step": 1859 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.428972972972973, + "grad_norm": 21.516246795654297, + "learning_rate": 6.524843951516436e-06, + "loss": 10.421356201171875, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4395122192800045, + "loss_total_v6": 0.6513347327709198, + "step": 1860 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4292036036036036, + "grad_norm": 23.135461807250977, + "learning_rate": 6.521226569832699e-06, + "loss": 9.966936111450195, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.4063349962234497, + "loss_total_v6": 0.6229335367679596, + "step": 1861 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4294342342342342, + "grad_norm": 20.531845092773438, + "learning_rate": 6.51760831047631e-06, + "loss": 10.144771575927734, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4229581467807293, + "loss_total_v6": 0.6340482234954834, + "step": 1862 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4296648648648649, + "grad_norm": 22.206722259521484, + "learning_rate": 6.513989175534825e-06, + "loss": 10.01003646850586, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4127519838511944, + "loss_total_v6": 0.6256273239850998, + "step": 1863 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4298954954954955, + "grad_norm": 21.968935012817383, + "learning_rate": 6.510369167096308e-06, + "loss": 10.139422416687012, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.41786305606365204, + "loss_total_v6": 0.6337138712406158, + "step": 1864 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4301261261261261, + "grad_norm": 21.027931213378906, + "learning_rate": 6.50674828724932e-06, + "loss": 10.382004737854004, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.43590839579701424, + "loss_total_v6": 0.6488752961158752, + "step": 1865 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4303567567567568, + "grad_norm": 19.9488582611084, + "learning_rate": 6.5031265380829355e-06, + "loss": 9.77651596069336, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.39647841453552246, + "loss_total_v6": 0.611032247543335, + "step": 1866 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4305873873873874, + "grad_norm": 20.870773315429688, + "learning_rate": 6.499503921686725e-06, + "loss": 10.095219612121582, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.42088353261351585, + "loss_total_v6": 0.6309512406587601, + "step": 1867 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.430818018018018, + "grad_norm": 21.56873321533203, + "learning_rate": 6.495880440150756e-06, + "loss": 9.518220901489258, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.3804417997598648, + "loss_total_v6": 0.5948888212442398, + "step": 1868 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43104864864864867, + "grad_norm": 20.40983772277832, + "learning_rate": 6.4922560955656e-06, + "loss": 10.319280624389648, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.4285396784543991, + "loss_total_v6": 0.6449550911784172, + "step": 1869 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4312792792792793, + "grad_norm": 20.505125045776367, + "learning_rate": 6.488630890022324e-06, + "loss": 10.356607437133789, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4322153180837631, + "loss_total_v6": 0.6472879499197006, + "step": 1870 + }, + { + "action_cond_aligned": 0.8984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4315099099099099, + "grad_norm": 20.63351821899414, + "learning_rate": 6.485004825612495e-06, + "loss": 10.18463134765625, + "loss_alignment": 0.39697265625, + "loss_alignment_w": 0.198486328125, + "loss_sft": 0.43840405344963074, + "loss_total_v6": 0.636539451777935, + "step": 1871 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43174054054054056, + "grad_norm": 21.767902374267578, + "learning_rate": 6.481377904428171e-06, + "loss": 10.182122230529785, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.42878685891628265, + "loss_total_v6": 0.6363826543092728, + "step": 1872 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4319711711711712, + "grad_norm": 20.59587287902832, + "learning_rate": 6.477750128561907e-06, + "loss": 10.033961296081543, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.41649023443460464, + "loss_total_v6": 0.6271225586533546, + "step": 1873 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4322018018018018, + "grad_norm": 21.384674072265625, + "learning_rate": 6.474121500106752e-06, + "loss": 10.045620918273926, + "loss_alignment": 0.443603515625, + "loss_alignment_w": 0.2218017578125, + "loss_sft": 0.40594277158379555, + "loss_total_v6": 0.6278513297438622, + "step": 1874 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43243243243243246, + "grad_norm": 20.255369186401367, + "learning_rate": 6.470492021156243e-06, + "loss": 10.415725708007812, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.4440278820693493, + "loss_total_v6": 0.6509828120470047, + "step": 1875 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43266306306306307, + "grad_norm": 21.170860290527344, + "learning_rate": 6.466861693804413e-06, + "loss": 9.406935691833496, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.3756532035768032, + "loss_total_v6": 0.5879334807395935, + "step": 1876 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4328936936936937, + "grad_norm": 22.024822235107422, + "learning_rate": 6.46323052014578e-06, + "loss": 10.297883987426758, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.43487748503685, + "loss_total_v6": 0.6436177492141724, + "step": 1877 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43312432432432435, + "grad_norm": 21.45331573486328, + "learning_rate": 6.4595985022753524e-06, + "loss": 10.08513069152832, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4194136932492256, + "loss_total_v6": 0.6303206533193588, + "step": 1878 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43335495495495496, + "grad_norm": 19.622600555419922, + "learning_rate": 6.455965642288624e-06, + "loss": 9.545389175415039, + "loss_alignment": 0.439208984375, + "loss_alignment_w": 0.2196044921875, + "loss_sft": 0.37675338611006737, + "loss_total_v6": 0.596586786210537, + "step": 1879 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4335855855855856, + "grad_norm": 21.08453941345215, + "learning_rate": 6.45233194228158e-06, + "loss": 9.543575286865234, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.37990545853972435, + "loss_total_v6": 0.596473440527916, + "step": 1880 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43381621621621624, + "grad_norm": 20.260177612304688, + "learning_rate": 6.448697404350681e-06, + "loss": 9.79284954071045, + "loss_alignment": 0.408935546875, + "loss_alignment_w": 0.2044677734375, + "loss_sft": 0.4067613519728184, + "loss_total_v6": 0.6120531037449837, + "step": 1881 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43404684684684686, + "grad_norm": 21.60063934326172, + "learning_rate": 6.445062030592879e-06, + "loss": 9.153186798095703, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.36762166768312454, + "loss_total_v6": 0.5720741674304008, + "step": 1882 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43427747747747747, + "grad_norm": 21.956998825073242, + "learning_rate": 6.441425823105603e-06, + "loss": 10.149518966674805, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4224309027194977, + "loss_total_v6": 0.6343449652194977, + "step": 1883 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4345081081081081, + "grad_norm": 20.817411422729492, + "learning_rate": 6.437788783986767e-06, + "loss": 9.922700881958008, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4085141718387604, + "loss_total_v6": 0.6201688200235367, + "step": 1884 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43473873873873875, + "grad_norm": 20.74417495727539, + "learning_rate": 6.434150915334763e-06, + "loss": 10.318151473999023, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.42788928002119064, + "loss_total_v6": 0.644884504377842, + "step": 1885 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43496936936936936, + "grad_norm": 22.652780532836914, + "learning_rate": 6.43051221924846e-06, + "loss": 10.19134521484375, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.42335129901766777, + "loss_total_v6": 0.6369590684771538, + "step": 1886 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4352, + "grad_norm": 21.660127639770508, + "learning_rate": 6.426872697827208e-06, + "loss": 10.371879577636719, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.44011255726218224, + "loss_total_v6": 0.6482424214482307, + "step": 1887 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43543063063063064, + "grad_norm": 22.90593910217285, + "learning_rate": 6.423232353170831e-06, + "loss": 9.704218864440918, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.39618654549121857, + "loss_total_v6": 0.6065136939287186, + "step": 1888 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43566126126126126, + "grad_norm": 21.094974517822266, + "learning_rate": 6.4195911873796264e-06, + "loss": 9.691659927368164, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.39100706577301025, + "loss_total_v6": 0.6057287603616714, + "step": 1889 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43589189189189187, + "grad_norm": 21.120803833007812, + "learning_rate": 6.41594920255437e-06, + "loss": 9.768242835998535, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.40034063160419464, + "loss_total_v6": 0.610515184700489, + "step": 1890 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43612252252252254, + "grad_norm": 24.095863342285156, + "learning_rate": 6.412306400796306e-06, + "loss": 10.182563781738281, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4153256230056286, + "loss_total_v6": 0.6364102438092232, + "step": 1891 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43635315315315315, + "grad_norm": 21.36659812927246, + "learning_rate": 6.408662784207149e-06, + "loss": 10.094301223754883, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.41896452009677887, + "loss_total_v6": 0.6308938264846802, + "step": 1892 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43658378378378376, + "grad_norm": 21.34193992614746, + "learning_rate": 6.405018354889089e-06, + "loss": 10.670259475708008, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4556485451757908, + "loss_total_v6": 0.6668912395834923, + "step": 1893 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43681441441441443, + "grad_norm": 19.011642456054688, + "learning_rate": 6.401373114944781e-06, + "loss": 9.835350036621094, + "loss_alignment": 0.441162109375, + "loss_alignment_w": 0.2205810546875, + "loss_sft": 0.39429613947868347, + "loss_total_v6": 0.6147093623876572, + "step": 1894 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43704504504504504, + "grad_norm": 21.695222854614258, + "learning_rate": 6.397727066477348e-06, + "loss": 9.460138320922852, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.3832661285996437, + "loss_total_v6": 0.5912586823105812, + "step": 1895 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43727567567567566, + "grad_norm": 21.606586456298828, + "learning_rate": 6.394080211590381e-06, + "loss": 9.705692291259766, + "loss_alignment": 0.436279296875, + "loss_alignment_w": 0.2181396484375, + "loss_sft": 0.38863395899534225, + "loss_total_v6": 0.6066057607531548, + "step": 1896 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4375063063063063, + "grad_norm": 21.77455711364746, + "learning_rate": 6.390432552387934e-06, + "loss": 9.418996810913086, + "loss_alignment": 0.401611328125, + "loss_alignment_w": 0.2008056640625, + "loss_sft": 0.38769854232668877, + "loss_total_v6": 0.5886872783303261, + "step": 1897 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43773693693693694, + "grad_norm": 24.14361000061035, + "learning_rate": 6.3867840909745284e-06, + "loss": 10.09885311126709, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.4159378223121166, + "loss_total_v6": 0.6311783269047737, + "step": 1898 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43796756756756755, + "grad_norm": 22.41671371459961, + "learning_rate": 6.383134829455142e-06, + "loss": 10.214218139648438, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.42737482488155365, + "loss_total_v6": 0.6383886262774467, + "step": 1899 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4381981981981982, + "grad_norm": 23.64282989501953, + "learning_rate": 6.379484769935223e-06, + "loss": 10.197938919067383, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.4189416468143463, + "loss_total_v6": 0.6373712047934532, + "step": 1900 + }, + { + "epoch": 0.4381981981981982, + "eval_loss": 0.6652957201004028, + "eval_loss_alignment": 0.4052379640758547, + "eval_loss_alignment_w": 0.20261898203792736, + "eval_loss_sft": 0.46267672666372395, + "eval_loss_total_v6": 0.6652957080648496, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.3874, + "eval_samples_per_second": 22.649, + "eval_steps_per_second": 2.84, + "step": 1900 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43842882882882883, + "grad_norm": 22.053186416625977, + "learning_rate": 6.3758339145206745e-06, + "loss": 9.407266616821289, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.37951912358403206, + "loss_total_v6": 0.5879541821777821, + "step": 1901 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43865945945945944, + "grad_norm": 20.752931594848633, + "learning_rate": 6.372182265317861e-06, + "loss": 9.689549446105957, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.3916228339076042, + "loss_total_v6": 0.6055968180298805, + "step": 1902 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4388900900900901, + "grad_norm": 21.54688835144043, + "learning_rate": 6.3685298244336e-06, + "loss": 9.524896621704102, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.38838161155581474, + "loss_total_v6": 0.5953060686588287, + "step": 1903 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4391207207207207, + "grad_norm": 20.34056282043457, + "learning_rate": 6.364876593975174e-06, + "loss": 9.247770309448242, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.36701756715774536, + "loss_total_v6": 0.5779855996370316, + "step": 1904 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.43935135135135134, + "grad_norm": 34.47513198852539, + "learning_rate": 6.361222576050312e-06, + "loss": 10.17525863647461, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.4263131245970726, + "loss_total_v6": 0.6359536796808243, + "step": 1905 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.439581981981982, + "grad_norm": 22.4202880859375, + "learning_rate": 6.357567772767207e-06, + "loss": 10.05056381225586, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4121873304247856, + "loss_total_v6": 0.6281602084636688, + "step": 1906 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4398126126126126, + "grad_norm": 21.41278076171875, + "learning_rate": 6.353912186234498e-06, + "loss": 10.27730941772461, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4318673647940159, + "loss_total_v6": 0.6423318535089493, + "step": 1907 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44004324324324323, + "grad_norm": 21.664546966552734, + "learning_rate": 6.3502558185612775e-06, + "loss": 10.09619140625, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4140929654240608, + "loss_total_v6": 0.6310119405388832, + "step": 1908 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4402738738738739, + "grad_norm": 21.601518630981445, + "learning_rate": 6.346598671857088e-06, + "loss": 10.16082763671875, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4231223687529564, + "loss_total_v6": 0.6350516825914383, + "step": 1909 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4405045045045045, + "grad_norm": 21.264129638671875, + "learning_rate": 6.342940748231925e-06, + "loss": 10.093162536621094, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4129423499107361, + "loss_total_v6": 0.6308226138353348, + "step": 1910 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4407351351351351, + "grad_norm": 23.126476287841797, + "learning_rate": 6.3392820497962305e-06, + "loss": 10.203746795654297, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42514875531196594, + "loss_total_v6": 0.6377341821789742, + "step": 1911 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44096576576576574, + "grad_norm": 21.538524627685547, + "learning_rate": 6.335622578660889e-06, + "loss": 10.171409606933594, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.4239669442176819, + "loss_total_v6": 0.635713167488575, + "step": 1912 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4411963963963964, + "grad_norm": 21.574735641479492, + "learning_rate": 6.331962336937239e-06, + "loss": 9.749298095703125, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.4045124165713787, + "loss_total_v6": 0.6093311384320259, + "step": 1913 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.441427027027027, + "grad_norm": 22.58902931213379, + "learning_rate": 6.3283013267370576e-06, + "loss": 10.303478240966797, + "loss_alignment": 0.40576171875, + "loss_alignment_w": 0.202880859375, + "loss_sft": 0.4408424310386181, + "loss_total_v6": 0.6439674124121666, + "step": 1914 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44165765765765763, + "grad_norm": 40.99470901489258, + "learning_rate": 6.324639550172567e-06, + "loss": 9.432872772216797, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.3797461576759815, + "loss_total_v6": 0.5895545184612274, + "step": 1915 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4418882882882883, + "grad_norm": 24.23229217529297, + "learning_rate": 6.3209770093564315e-06, + "loss": 9.942802429199219, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.41526366397738457, + "loss_total_v6": 0.62142513692379, + "step": 1916 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4421189189189189, + "grad_norm": 20.804901123046875, + "learning_rate": 6.317313706401755e-06, + "loss": 9.798645973205566, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4010048285126686, + "loss_total_v6": 0.6124153658747673, + "step": 1917 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4423495495495495, + "grad_norm": 23.269847869873047, + "learning_rate": 6.313649643422085e-06, + "loss": 10.240217208862305, + "loss_alignment": 0.4365234375, + "loss_alignment_w": 0.21826171875, + "loss_sft": 0.42126360908150673, + "loss_total_v6": 0.6400135904550552, + "step": 1918 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4425801801801802, + "grad_norm": 21.94769859313965, + "learning_rate": 6.309984822531404e-06, + "loss": 9.551105499267578, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.38582347705960274, + "loss_total_v6": 0.596944086253643, + "step": 1919 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4428108108108108, + "grad_norm": 21.347328186035156, + "learning_rate": 6.306319245844134e-06, + "loss": 10.052838325500488, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4136875569820404, + "loss_total_v6": 0.6283024027943611, + "step": 1920 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4430414414414414, + "grad_norm": 21.58673667907715, + "learning_rate": 6.302652915475129e-06, + "loss": 10.014650344848633, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.41320809721946716, + "loss_total_v6": 0.6259156316518784, + "step": 1921 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4432720720720721, + "grad_norm": 22.74384117126465, + "learning_rate": 6.298985833539681e-06, + "loss": 10.123828887939453, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42018435150384903, + "loss_total_v6": 0.6327392756938934, + "step": 1922 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4435027027027027, + "grad_norm": 21.585052490234375, + "learning_rate": 6.29531800215352e-06, + "loss": 10.012954711914062, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.41514674946665764, + "loss_total_v6": 0.6258096322417259, + "step": 1923 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4437333333333333, + "grad_norm": 23.072460174560547, + "learning_rate": 6.2916494234327985e-06, + "loss": 9.765178680419922, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.40223951637744904, + "loss_total_v6": 0.6103236339986324, + "step": 1924 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.443963963963964, + "grad_norm": 21.944650650024414, + "learning_rate": 6.287980099494106e-06, + "loss": 9.938301086425781, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.40950437635183334, + "loss_total_v6": 0.6211437880992889, + "step": 1925 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4441945945945946, + "grad_norm": 21.363479614257812, + "learning_rate": 6.284310032454463e-06, + "loss": 10.31723403930664, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.42757244780659676, + "loss_total_v6": 0.644827090203762, + "step": 1926 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4444252252252252, + "grad_norm": 20.615617752075195, + "learning_rate": 6.280639224431317e-06, + "loss": 10.001729965209961, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4131177105009556, + "loss_total_v6": 0.6251080706715584, + "step": 1927 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4446558558558559, + "grad_norm": 22.25665855407715, + "learning_rate": 6.276967677542542e-06, + "loss": 10.067285537719727, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4144226536154747, + "loss_total_v6": 0.6292053386569023, + "step": 1928 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4448864864864865, + "grad_norm": 22.547752380371094, + "learning_rate": 6.2732953939064416e-06, + "loss": 9.48194694519043, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.38440022990107536, + "loss_total_v6": 0.5926216617226601, + "step": 1929 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4451171171171171, + "grad_norm": 20.70743751525879, + "learning_rate": 6.269622375641739e-06, + "loss": 9.81802749633789, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.4066106900572777, + "loss_total_v6": 0.6136267110705376, + "step": 1930 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44534774774774777, + "grad_norm": 19.73151969909668, + "learning_rate": 6.265948624867585e-06, + "loss": 9.528732299804688, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.38701921328902245, + "loss_total_v6": 0.5955458246171474, + "step": 1931 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4455783783783784, + "grad_norm": 21.919031143188477, + "learning_rate": 6.262274143703555e-06, + "loss": 9.814648628234863, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.40459899976849556, + "loss_total_v6": 0.6134155467152596, + "step": 1932 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.445809009009009, + "grad_norm": 22.494991302490234, + "learning_rate": 6.258598934269642e-06, + "loss": 10.050336837768555, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.41285981237888336, + "loss_total_v6": 0.6281460821628571, + "step": 1933 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44603963963963966, + "grad_norm": 21.42071533203125, + "learning_rate": 6.254922998686259e-06, + "loss": 9.705697059631348, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.3924032002687454, + "loss_total_v6": 0.6066060587763786, + "step": 1934 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4462702702702703, + "grad_norm": 21.557554244995117, + "learning_rate": 6.251246339074241e-06, + "loss": 9.815502166748047, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.397389180958271, + "loss_total_v6": 0.6134689003229141, + "step": 1935 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4465009009009009, + "grad_norm": 20.627595901489258, + "learning_rate": 6.24756895755484e-06, + "loss": 9.68952751159668, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.3952988013625145, + "loss_total_v6": 0.6055954359471798, + "step": 1936 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44673153153153156, + "grad_norm": 23.691112518310547, + "learning_rate": 6.24389085624972e-06, + "loss": 10.032793998718262, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.41898081079125404, + "loss_total_v6": 0.6270496174693108, + "step": 1937 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44696216216216217, + "grad_norm": 22.199504852294922, + "learning_rate": 6.240212037280967e-06, + "loss": 10.04556655883789, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.42186950519680977, + "loss_total_v6": 0.6278479099273682, + "step": 1938 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4471927927927928, + "grad_norm": 22.15286636352539, + "learning_rate": 6.236532502771078e-06, + "loss": 9.916553497314453, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.4030030183494091, + "loss_total_v6": 0.6197846308350563, + "step": 1939 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44742342342342345, + "grad_norm": 21.76010513305664, + "learning_rate": 6.232852254842961e-06, + "loss": 9.7232666015625, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.39791110903024673, + "loss_total_v6": 0.6077041551470757, + "step": 1940 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44765405405405406, + "grad_norm": 22.042997360229492, + "learning_rate": 6.22917129561994e-06, + "loss": 9.169830322265625, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.36777687445282936, + "loss_total_v6": 0.5731144174933434, + "step": 1941 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4478846846846847, + "grad_norm": 20.89834976196289, + "learning_rate": 6.225489627225746e-06, + "loss": 9.861360549926758, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.40291040390729904, + "loss_total_v6": 0.6163350641727448, + "step": 1942 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4481153153153153, + "grad_norm": 20.776592254638672, + "learning_rate": 6.221807251784517e-06, + "loss": 10.247182846069336, + "loss_alignment": 0.44287109375, + "loss_alignment_w": 0.221435546875, + "loss_sft": 0.4194254204630852, + "loss_total_v6": 0.6404489800333977, + "step": 1943 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44834594594594596, + "grad_norm": 23.93226432800293, + "learning_rate": 6.218124171420806e-06, + "loss": 9.843561172485352, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.40193528309464455, + "loss_total_v6": 0.6152226105332375, + "step": 1944 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44857657657657657, + "grad_norm": 22.092214584350586, + "learning_rate": 6.214440388259565e-06, + "loss": 10.034858703613281, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4128995016217232, + "loss_total_v6": 0.6271786838769913, + "step": 1945 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4488072072072072, + "grad_norm": 21.639726638793945, + "learning_rate": 6.210755904426159e-06, + "loss": 9.811586380004883, + "loss_alignment": 0.40478515625, + "loss_alignment_w": 0.202392578125, + "loss_sft": 0.4110298715531826, + "loss_total_v6": 0.613224096596241, + "step": 1946 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44903783783783785, + "grad_norm": 20.353591918945312, + "learning_rate": 6.20707072204635e-06, + "loss": 9.469003677368164, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.38359129056334496, + "loss_total_v6": 0.5918127074837685, + "step": 1947 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44926846846846846, + "grad_norm": 21.932889938354492, + "learning_rate": 6.203384843246307e-06, + "loss": 9.321756362915039, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.37132131308317184, + "loss_total_v6": 0.5826097577810287, + "step": 1948 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4494990990990991, + "grad_norm": 23.96968650817871, + "learning_rate": 6.199698270152602e-06, + "loss": 10.120635986328125, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4242572784423828, + "loss_total_v6": 0.6325397565960884, + "step": 1949 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44972972972972974, + "grad_norm": 20.722278594970703, + "learning_rate": 6.196011004892204e-06, + "loss": 9.811187744140625, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.40104103460907936, + "loss_total_v6": 0.6131992265582085, + "step": 1950 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.44996036036036036, + "grad_norm": 22.085044860839844, + "learning_rate": 6.192323049592483e-06, + "loss": 10.266334533691406, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4249100759625435, + "loss_total_v6": 0.6416459232568741, + "step": 1951 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45019099099099097, + "grad_norm": 22.525753021240234, + "learning_rate": 6.188634406381206e-06, + "loss": 10.150388717651367, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.41961658746004105, + "loss_total_v6": 0.6343992874026299, + "step": 1952 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45042162162162164, + "grad_norm": 20.33962059020996, + "learning_rate": 6.1849450773865424e-06, + "loss": 9.864181518554688, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.4101819880306721, + "loss_total_v6": 0.6165113598108292, + "step": 1953 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45065225225225225, + "grad_norm": 23.09684944152832, + "learning_rate": 6.181255064737051e-06, + "loss": 9.608987808227539, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.3932100348174572, + "loss_total_v6": 0.60056172311306, + "step": 1954 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45088288288288286, + "grad_norm": 22.961671829223633, + "learning_rate": 6.177564370561685e-06, + "loss": 9.574675559997559, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.3949717804789543, + "loss_total_v6": 0.5984172075986862, + "step": 1955 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45111351351351353, + "grad_norm": 22.39459228515625, + "learning_rate": 6.173872996989794e-06, + "loss": 9.948646545410156, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.40506983548402786, + "loss_total_v6": 0.6217904016375542, + "step": 1956 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45134414414414414, + "grad_norm": 21.876726150512695, + "learning_rate": 6.170180946151118e-06, + "loss": 10.087763786315918, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4181438982486725, + "loss_total_v6": 0.6304852142930031, + "step": 1957 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45157477477477476, + "grad_norm": 20.653644561767578, + "learning_rate": 6.166488220175789e-06, + "loss": 10.293320655822754, + "loss_alignment": 0.43798828125, + "loss_alignment_w": 0.218994140625, + "loss_sft": 0.4243231602013111, + "loss_total_v6": 0.6433325558900833, + "step": 1958 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4518054054054054, + "grad_norm": 21.502897262573242, + "learning_rate": 6.162794821194327e-06, + "loss": 9.809463500976562, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.40056706592440605, + "loss_total_v6": 0.6130914613604546, + "step": 1959 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45203603603603604, + "grad_norm": 21.887357711791992, + "learning_rate": 6.1591007513376425e-06, + "loss": 10.08201789855957, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.42024143785238266, + "loss_total_v6": 0.630126103758812, + "step": 1960 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45226666666666665, + "grad_norm": 21.71025276184082, + "learning_rate": 6.1554060127370285e-06, + "loss": 9.72080135345459, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.39374395832419395, + "loss_total_v6": 0.6075500883162022, + "step": 1961 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4524972972972973, + "grad_norm": 23.162736892700195, + "learning_rate": 6.151710607524169e-06, + "loss": 9.983696937561035, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.4171329326927662, + "loss_total_v6": 0.6239810511469841, + "step": 1962 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45272792792792793, + "grad_norm": 20.809688568115234, + "learning_rate": 6.148014537831131e-06, + "loss": 9.977265357971191, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.4091625511646271, + "loss_total_v6": 0.6235790550708771, + "step": 1963 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45295855855855854, + "grad_norm": 23.134782791137695, + "learning_rate": 6.144317805790362e-06, + "loss": 9.998716354370117, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4121054448187351, + "loss_total_v6": 0.6249197721481323, + "step": 1964 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4531891891891892, + "grad_norm": 20.2997989654541, + "learning_rate": 6.140620413534695e-06, + "loss": 9.856552124023438, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.41234494373202324, + "loss_total_v6": 0.616034522652626, + "step": 1965 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4534198198198198, + "grad_norm": 23.74376678466797, + "learning_rate": 6.136922363197344e-06, + "loss": 9.915239334106445, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4098178595304489, + "loss_total_v6": 0.6197024881839752, + "step": 1966 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45365045045045044, + "grad_norm": 21.345081329345703, + "learning_rate": 6.1332236569119e-06, + "loss": 9.838294982910156, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.40568017214536667, + "loss_total_v6": 0.6148934364318848, + "step": 1967 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4538810810810811, + "grad_norm": 19.76327896118164, + "learning_rate": 6.129524296812335e-06, + "loss": 9.832473754882812, + "loss_alignment": 0.41259765625, + "loss_alignment_w": 0.206298828125, + "loss_sft": 0.40804770588874817, + "loss_total_v6": 0.61452966183424, + "step": 1968 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4541117117117117, + "grad_norm": 20.651107788085938, + "learning_rate": 6.125824285032995e-06, + "loss": 9.870254516601562, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.40873050689697266, + "loss_total_v6": 0.6168909221887589, + "step": 1969 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45434234234234233, + "grad_norm": 29.326974868774414, + "learning_rate": 6.122123623708608e-06, + "loss": 9.969657897949219, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.4154467284679413, + "loss_total_v6": 0.623103603720665, + "step": 1970 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.454572972972973, + "grad_norm": 21.276235580444336, + "learning_rate": 6.118422314974269e-06, + "loss": 10.110122680664062, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.417527224868536, + "loss_total_v6": 0.6318826824426651, + "step": 1971 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4548036036036036, + "grad_norm": 23.883712768554688, + "learning_rate": 6.1147203609654535e-06, + "loss": 10.27349853515625, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.43173598125576973, + "loss_total_v6": 0.6420936733484268, + "step": 1972 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4550342342342342, + "grad_norm": 21.574037551879883, + "learning_rate": 6.111017763818005e-06, + "loss": 9.971331596374512, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.408791683614254, + "loss_total_v6": 0.6232082284986973, + "step": 1973 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45526486486486484, + "grad_norm": 22.461902618408203, + "learning_rate": 6.107314525668139e-06, + "loss": 10.047540664672852, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.41413456946611404, + "loss_total_v6": 0.6279712617397308, + "step": 1974 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4554954954954955, + "grad_norm": 20.615522384643555, + "learning_rate": 6.103610648652442e-06, + "loss": 9.2832612991333, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.3665350191295147, + "loss_total_v6": 0.5802038460969925, + "step": 1975 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4557261261261261, + "grad_norm": 20.93989372253418, + "learning_rate": 6.099906134907868e-06, + "loss": 9.879058837890625, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.40410803258419037, + "loss_total_v6": 0.6174411699175835, + "step": 1976 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45595675675675673, + "grad_norm": 20.77078628540039, + "learning_rate": 6.096200986571737e-06, + "loss": 9.683759689331055, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.3989361301064491, + "loss_total_v6": 0.6052349656820297, + "step": 1977 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4561873873873874, + "grad_norm": 21.60663414001465, + "learning_rate": 6.0924952057817385e-06, + "loss": 10.12118148803711, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.42253655567765236, + "loss_total_v6": 0.6325737908482552, + "step": 1978 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.456418018018018, + "grad_norm": 20.322113037109375, + "learning_rate": 6.088788794675926e-06, + "loss": 9.424175262451172, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.37897375598549843, + "loss_total_v6": 0.5890109874308109, + "step": 1979 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4566486486486486, + "grad_norm": 23.63216209411621, + "learning_rate": 6.0850817553927145e-06, + "loss": 10.053752899169922, + "loss_alignment": 0.41064453125, + "loss_alignment_w": 0.205322265625, + "loss_sft": 0.422716848552227, + "loss_total_v6": 0.6283595785498619, + "step": 1980 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4568792792792793, + "grad_norm": 22.693309783935547, + "learning_rate": 6.081374090070882e-06, + "loss": 10.06788444519043, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.42323382943868637, + "loss_total_v6": 0.6292427480220795, + "step": 1981 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4571099099099099, + "grad_norm": 21.070478439331055, + "learning_rate": 6.077665800849568e-06, + "loss": 9.613092422485352, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.38853800669312477, + "loss_total_v6": 0.6008182838559151, + "step": 1982 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4573405405405405, + "grad_norm": 22.195880889892578, + "learning_rate": 6.073956889868275e-06, + "loss": 10.1246337890625, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.41735074296593666, + "loss_total_v6": 0.6327895596623421, + "step": 1983 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4575711711711712, + "grad_norm": 21.84421157836914, + "learning_rate": 6.07024735926686e-06, + "loss": 9.725663185119629, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.3920031189918518, + "loss_total_v6": 0.6078539341688156, + "step": 1984 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4578018018018018, + "grad_norm": 20.062894821166992, + "learning_rate": 6.066537211185539e-06, + "loss": 10.046424865722656, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.41098257526755333, + "loss_total_v6": 0.6279015615582466, + "step": 1985 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4580324324324324, + "grad_norm": 24.298524856567383, + "learning_rate": 6.062826447764883e-06, + "loss": 10.077957153320312, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.41515061259269714, + "loss_total_v6": 0.6298722922801971, + "step": 1986 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4582630630630631, + "grad_norm": 21.26885223388672, + "learning_rate": 6.059115071145824e-06, + "loss": 9.773860931396484, + "loss_alignment": 0.40576171875, + "loss_alignment_w": 0.202880859375, + "loss_sft": 0.4077718183398247, + "loss_total_v6": 0.6108663156628609, + "step": 1987 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4584936936936937, + "grad_norm": 20.94260597229004, + "learning_rate": 6.055403083469637e-06, + "loss": 9.889307022094727, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4042450450360775, + "loss_total_v6": 0.6180816888809204, + "step": 1988 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4587243243243243, + "grad_norm": 24.177324295043945, + "learning_rate": 6.051690486877961e-06, + "loss": 9.940385818481445, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.407147578895092, + "loss_total_v6": 0.6212741509079933, + "step": 1989 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.458954954954955, + "grad_norm": 22.022294998168945, + "learning_rate": 6.047977283512777e-06, + "loss": 9.537382125854492, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.38223443925380707, + "loss_total_v6": 0.596086360514164, + "step": 1990 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4591855855855856, + "grad_norm": 21.98000144958496, + "learning_rate": 6.044263475516421e-06, + "loss": 9.734243392944336, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.3943093791604042, + "loss_total_v6": 0.6083901971578598, + "step": 1991 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4594162162162162, + "grad_norm": 20.440580368041992, + "learning_rate": 6.04054906503158e-06, + "loss": 10.55860710144043, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.44671716168522835, + "loss_total_v6": 0.6599129810929298, + "step": 1992 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.45964684684684687, + "grad_norm": 20.466386795043945, + "learning_rate": 6.036834054201283e-06, + "loss": 10.21541690826416, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4277854450047016, + "loss_total_v6": 0.6384635493159294, + "step": 1993 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4598774774774775, + "grad_norm": 21.015625, + "learning_rate": 6.0331184451689075e-06, + "loss": 10.220458030700684, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4215087257325649, + "loss_total_v6": 0.6387785971164703, + "step": 1994 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4601081081081081, + "grad_norm": 23.607742309570312, + "learning_rate": 6.029402240078176e-06, + "loss": 10.283207893371582, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.4255526587367058, + "loss_total_v6": 0.6427004858851433, + "step": 1995 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46033873873873876, + "grad_norm": 21.104400634765625, + "learning_rate": 6.025685441073157e-06, + "loss": 10.036853790283203, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4142448678612709, + "loss_total_v6": 0.627303309738636, + "step": 1996 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4605693693693694, + "grad_norm": 24.677772521972656, + "learning_rate": 6.021968050298261e-06, + "loss": 10.020423889160156, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.41704798117280006, + "loss_total_v6": 0.6262764930725098, + "step": 1997 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4608, + "grad_norm": 21.32012176513672, + "learning_rate": 6.018250069898236e-06, + "loss": 10.124666213989258, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.41849714890122414, + "loss_total_v6": 0.6327916011214256, + "step": 1998 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46103063063063066, + "grad_norm": 22.090652465820312, + "learning_rate": 6.014531502018175e-06, + "loss": 9.602910995483398, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.3849872164428234, + "loss_total_v6": 0.6001818999648094, + "step": 1999 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46126126126126127, + "grad_norm": 21.02519989013672, + "learning_rate": 6.0108123488035094e-06, + "loss": 10.085647583007812, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.41563132405281067, + "loss_total_v6": 0.6303529962897301, + "step": 2000 + }, + { + "epoch": 0.46126126126126127, + "eval_loss": 0.6618443131446838, + "eval_loss_alignment": 0.4043949485844017, + "eval_loss_alignment_w": 0.20219747429220086, + "eval_loss_sft": 0.4596468445518587, + "eval_loss_total_v6": 0.6618443187326193, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.2861, + "eval_samples_per_second": 22.677, + "eval_steps_per_second": 2.844, + "step": 2000 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4614918918918919, + "grad_norm": 21.998748779296875, + "learning_rate": 6.007092612400005e-06, + "loss": 9.531231880187988, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.38484079390764236, + "loss_total_v6": 0.5957019925117493, + "step": 2001 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46172252252252255, + "grad_norm": 19.69727325439453, + "learning_rate": 6.003372294953767e-06, + "loss": 9.665170669555664, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.3942953161895275, + "loss_total_v6": 0.6040731519460678, + "step": 2002 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46195315315315316, + "grad_norm": 23.078872680664062, + "learning_rate": 5.999651398611235e-06, + "loss": 9.880361557006836, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.40212948247790337, + "loss_total_v6": 0.6175225675106049, + "step": 2003 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4621837837837838, + "grad_norm": 21.594337463378906, + "learning_rate": 5.995929925519181e-06, + "loss": 9.756784439086914, + "loss_alignment": 0.410400390625, + "loss_alignment_w": 0.2052001953125, + "loss_sft": 0.40456828847527504, + "loss_total_v6": 0.6097990050911903, + "step": 2004 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4624144144144144, + "grad_norm": 20.594804763793945, + "learning_rate": 5.992207877824714e-06, + "loss": 9.840129852294922, + "loss_alignment": 0.40625, + "loss_alignment_w": 0.203125, + "loss_sft": 0.4115322381258011, + "loss_total_v6": 0.6150081604719162, + "step": 2005 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46264504504504506, + "grad_norm": 19.374103546142578, + "learning_rate": 5.98848525767527e-06, + "loss": 10.152254104614258, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.4184361919760704, + "loss_total_v6": 0.6345158964395523, + "step": 2006 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46287567567567567, + "grad_norm": 20.515968322753906, + "learning_rate": 5.984762067218617e-06, + "loss": 9.770270347595215, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.400879368185997, + "loss_total_v6": 0.6106419041752815, + "step": 2007 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4631063063063063, + "grad_norm": 21.729473114013672, + "learning_rate": 5.981038308602853e-06, + "loss": 9.592571258544922, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.3887508176267147, + "loss_total_v6": 0.5995357297360897, + "step": 2008 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46333693693693695, + "grad_norm": 22.657779693603516, + "learning_rate": 5.977313983976403e-06, + "loss": 10.075078964233398, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4128650650382042, + "loss_total_v6": 0.6296924576163292, + "step": 2009 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46356756756756756, + "grad_norm": 21.90727996826172, + "learning_rate": 5.973589095488017e-06, + "loss": 9.780213356018066, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.40752801299095154, + "loss_total_v6": 0.6112633645534515, + "step": 2010 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4637981981981982, + "grad_norm": 20.916427612304688, + "learning_rate": 5.969863645286774e-06, + "loss": 9.748336791992188, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.396975502371788, + "loss_total_v6": 0.6092710196971893, + "step": 2011 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46402882882882884, + "grad_norm": 25.09572410583496, + "learning_rate": 5.966137635522074e-06, + "loss": 9.934285163879395, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.41785938665270805, + "loss_total_v6": 0.6208928003907204, + "step": 2012 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46425945945945946, + "grad_norm": 31.15487289428711, + "learning_rate": 5.96241106834364e-06, + "loss": 10.287803649902344, + "loss_alignment": 0.39990234375, + "loss_alignment_w": 0.199951171875, + "loss_sft": 0.4427923783659935, + "loss_total_v6": 0.6429876759648323, + "step": 2013 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46449009009009007, + "grad_norm": 20.52922248840332, + "learning_rate": 5.958683945901517e-06, + "loss": 9.914619445800781, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4077649340033531, + "loss_total_v6": 0.6196637228131294, + "step": 2014 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46472072072072074, + "grad_norm": 19.677785873413086, + "learning_rate": 5.954956270346074e-06, + "loss": 9.992996215820312, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.41109180450439453, + "loss_total_v6": 0.6245622634887695, + "step": 2015 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46495135135135135, + "grad_norm": 21.508441925048828, + "learning_rate": 5.951228043827991e-06, + "loss": 9.794374465942383, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.39852533861994743, + "loss_total_v6": 0.6121483519673347, + "step": 2016 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46518198198198196, + "grad_norm": 19.810359954833984, + "learning_rate": 5.947499268498274e-06, + "loss": 9.860664367675781, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4050946533679962, + "loss_total_v6": 0.6162915490567684, + "step": 2017 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46541261261261263, + "grad_norm": 21.471595764160156, + "learning_rate": 5.943769946508242e-06, + "loss": 9.507466316223145, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.38132602721452713, + "loss_total_v6": 0.5942166447639465, + "step": 2018 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46564324324324324, + "grad_norm": 25.452392578125, + "learning_rate": 5.9400400800095295e-06, + "loss": 9.775899887084961, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.40164314582943916, + "loss_total_v6": 0.6109937503933907, + "step": 2019 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46587387387387386, + "grad_norm": 21.969226837158203, + "learning_rate": 5.936309671154085e-06, + "loss": 10.003640174865723, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.41230639442801476, + "loss_total_v6": 0.62522754073143, + "step": 2020 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4661045045045045, + "grad_norm": 21.09586524963379, + "learning_rate": 5.932578722094169e-06, + "loss": 9.507919311523438, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.39121150597929955, + "loss_total_v6": 0.5942449495196342, + "step": 2021 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46633513513513514, + "grad_norm": 23.716447830200195, + "learning_rate": 5.928847234982358e-06, + "loss": 10.073984146118164, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.41809147968888283, + "loss_total_v6": 0.6296240389347076, + "step": 2022 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46656576576576575, + "grad_norm": 22.55240821838379, + "learning_rate": 5.925115211971533e-06, + "loss": 9.952972412109375, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4095974266529083, + "loss_total_v6": 0.6220607981085777, + "step": 2023 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4667963963963964, + "grad_norm": 23.017757415771484, + "learning_rate": 5.921382655214889e-06, + "loss": 9.973611831665039, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.41261157765984535, + "loss_total_v6": 0.6233506873250008, + "step": 2024 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46702702702702703, + "grad_norm": 19.6622371673584, + "learning_rate": 5.917649566865926e-06, + "loss": 9.740592956542969, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.39018965885043144, + "loss_total_v6": 0.6087870597839355, + "step": 2025 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46725765765765764, + "grad_norm": 22.218006134033203, + "learning_rate": 5.913915949078453e-06, + "loss": 10.306045532226562, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.43029123544692993, + "loss_total_v6": 0.6441278755664825, + "step": 2026 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4674882882882883, + "grad_norm": 21.979862213134766, + "learning_rate": 5.910181804006582e-06, + "loss": 9.36580753326416, + "loss_alignment": 0.44189453125, + "loss_alignment_w": 0.220947265625, + "loss_sft": 0.3647666536271572, + "loss_total_v6": 0.5853629559278488, + "step": 2027 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4677189189189189, + "grad_norm": 20.06490135192871, + "learning_rate": 5.906447133804731e-06, + "loss": 9.995101928710938, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.4163961112499237, + "loss_total_v6": 0.6246938481926918, + "step": 2028 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46794954954954954, + "grad_norm": 21.20813751220703, + "learning_rate": 5.902711940627619e-06, + "loss": 9.851522445678711, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.40607965737581253, + "loss_total_v6": 0.6157201603055, + "step": 2029 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4681801801801802, + "grad_norm": 21.388254165649414, + "learning_rate": 5.8989762266302685e-06, + "loss": 10.23037338256836, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4287812411785126, + "loss_total_v6": 0.6393983289599419, + "step": 2030 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4684108108108108, + "grad_norm": 21.78312110900879, + "learning_rate": 5.895239993968005e-06, + "loss": 9.734230995178223, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.3919435404241085, + "loss_total_v6": 0.6083894520998001, + "step": 2031 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46864144144144143, + "grad_norm": 20.31786346435547, + "learning_rate": 5.891503244796448e-06, + "loss": 9.963467597961426, + "loss_alignment": 0.4423828125, + "loss_alignment_w": 0.22119140625, + "loss_sft": 0.4012506902217865, + "loss_total_v6": 0.6227167248725891, + "step": 2032 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4688720720720721, + "grad_norm": 21.913843154907227, + "learning_rate": 5.887765981271518e-06, + "loss": 9.740676879882812, + "loss_alignment": 0.408935546875, + "loss_alignment_w": 0.2044677734375, + "loss_sft": 0.40423301234841347, + "loss_total_v6": 0.6087923422455788, + "step": 2033 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4691027027027027, + "grad_norm": 23.574783325195312, + "learning_rate": 5.884028205549429e-06, + "loss": 9.960521697998047, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4068496711552143, + "loss_total_v6": 0.6225326508283615, + "step": 2034 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4693333333333333, + "grad_norm": 20.064476013183594, + "learning_rate": 5.880289919786697e-06, + "loss": 9.664505004882812, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.391766544431448, + "loss_total_v6": 0.6040315926074982, + "step": 2035 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.46956396396396394, + "grad_norm": 19.372228622436523, + "learning_rate": 5.876551126140125e-06, + "loss": 9.578372955322266, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.3959658481180668, + "loss_total_v6": 0.5986483544111252, + "step": 2036 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4697945945945946, + "grad_norm": 22.540067672729492, + "learning_rate": 5.872811826766817e-06, + "loss": 10.222854614257812, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4294862635433674, + "loss_total_v6": 0.6389283984899521, + "step": 2037 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4700252252252252, + "grad_norm": 19.91236686706543, + "learning_rate": 5.869072023824158e-06, + "loss": 9.54768180847168, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.3822983615100384, + "loss_total_v6": 0.596730150282383, + "step": 2038 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47025585585585583, + "grad_norm": 20.70391082763672, + "learning_rate": 5.865331719469836e-06, + "loss": 9.797567367553711, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4032415673136711, + "loss_total_v6": 0.6123479753732681, + "step": 2039 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4704864864864865, + "grad_norm": 21.87689971923828, + "learning_rate": 5.861590915861816e-06, + "loss": 10.183591842651367, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.42941270023584366, + "loss_total_v6": 0.6364744827151299, + "step": 2040 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4707171171171171, + "grad_norm": 22.83233070373535, + "learning_rate": 5.857849615158363e-06, + "loss": 10.113730430603027, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.4178442656993866, + "loss_total_v6": 0.6321081817150116, + "step": 2041 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4709477477477477, + "grad_norm": 23.873353958129883, + "learning_rate": 5.8541078195180175e-06, + "loss": 9.427809715270996, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.37456219643354416, + "loss_total_v6": 0.5892380997538567, + "step": 2042 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4711783783783784, + "grad_norm": 21.82997703552246, + "learning_rate": 5.850365531099614e-06, + "loss": 9.705411911010742, + "loss_alignment": 0.4140625, + "loss_alignment_w": 0.20703125, + "loss_sft": 0.3997553549706936, + "loss_total_v6": 0.6065882369875908, + "step": 2043 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.471409009009009, + "grad_norm": 20.01996612548828, + "learning_rate": 5.846622752062268e-06, + "loss": 10.201284408569336, + "loss_alignment": 0.4140625, + "loss_alignment_w": 0.20703125, + "loss_sft": 0.4302591569721699, + "loss_total_v6": 0.6375803053379059, + "step": 2044 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4716396396396396, + "grad_norm": 21.437620162963867, + "learning_rate": 5.842879484565379e-06, + "loss": 10.24919319152832, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.4330093078315258, + "loss_total_v6": 0.6405745819211006, + "step": 2045 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4718702702702703, + "grad_norm": 20.191926956176758, + "learning_rate": 5.839135730768626e-06, + "loss": 10.024066925048828, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.40897486358880997, + "loss_total_v6": 0.6265041455626488, + "step": 2046 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4721009009009009, + "grad_norm": 20.210840225219727, + "learning_rate": 5.835391492831969e-06, + "loss": 9.784652709960938, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.4029989205300808, + "loss_total_v6": 0.6115408018231392, + "step": 2047 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4723315315315315, + "grad_norm": 20.992971420288086, + "learning_rate": 5.831646772915651e-06, + "loss": 9.764307975769043, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.406655952334404, + "loss_total_v6": 0.610269233584404, + "step": 2048 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4725621621621622, + "grad_norm": 23.4611759185791, + "learning_rate": 5.827901573180189e-06, + "loss": 9.741975784301758, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.3936482146382332, + "loss_total_v6": 0.6088734567165375, + "step": 2049 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4727927927927928, + "grad_norm": 19.31146240234375, + "learning_rate": 5.824155895786381e-06, + "loss": 10.0335693359375, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.41345977038145065, + "loss_total_v6": 0.6270980760455132, + "step": 2050 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4730234234234234, + "grad_norm": 22.083829879760742, + "learning_rate": 5.8204097428952945e-06, + "loss": 10.532539367675781, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.44188355281949043, + "loss_total_v6": 0.6582836955785751, + "step": 2051 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4732540540540541, + "grad_norm": 21.031614303588867, + "learning_rate": 5.816663116668276e-06, + "loss": 9.849382400512695, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.39952195435762405, + "loss_total_v6": 0.6155864000320435, + "step": 2052 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4734846846846847, + "grad_norm": 20.419147491455078, + "learning_rate": 5.812916019266946e-06, + "loss": 9.662172317504883, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.3971901945769787, + "loss_total_v6": 0.6038857772946358, + "step": 2053 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4737153153153153, + "grad_norm": 25.604970932006836, + "learning_rate": 5.809168452853193e-06, + "loss": 9.66041088104248, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.394226748496294, + "loss_total_v6": 0.603775680065155, + "step": 2054 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47394594594594597, + "grad_norm": 22.241989135742188, + "learning_rate": 5.8054204195891775e-06, + "loss": 9.501056671142578, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.3893635459244251, + "loss_total_v6": 0.5938160642981529, + "step": 2055 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4741765765765766, + "grad_norm": 20.510753631591797, + "learning_rate": 5.801671921637329e-06, + "loss": 10.0374174118042, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.42175688967108727, + "loss_total_v6": 0.6273385658860207, + "step": 2056 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4744072072072072, + "grad_norm": 21.97232437133789, + "learning_rate": 5.797922961160348e-06, + "loss": 9.651418685913086, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.3899110406637192, + "loss_total_v6": 0.6032136529684067, + "step": 2057 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47463783783783786, + "grad_norm": 22.235694885253906, + "learning_rate": 5.7941735403212e-06, + "loss": 10.094422340393066, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.42408376187086105, + "loss_total_v6": 0.6309013739228249, + "step": 2058 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4748684684684685, + "grad_norm": 22.325275421142578, + "learning_rate": 5.790423661283112e-06, + "loss": 9.716243743896484, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.3963276632130146, + "loss_total_v6": 0.6072651892900467, + "step": 2059 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4750990990990991, + "grad_norm": 24.04390525817871, + "learning_rate": 5.786673326209584e-06, + "loss": 9.788877487182617, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.399326179176569, + "loss_total_v6": 0.611804835498333, + "step": 2060 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47532972972972976, + "grad_norm": 21.33623695373535, + "learning_rate": 5.78292253726437e-06, + "loss": 9.355603218078613, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.372582271695137, + "loss_total_v6": 0.5847252272069454, + "step": 2061 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47556036036036037, + "grad_norm": 19.937257766723633, + "learning_rate": 5.779171296611493e-06, + "loss": 9.998515129089355, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4124895967543125, + "loss_total_v6": 0.6249072030186653, + "step": 2062 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.475790990990991, + "grad_norm": 21.518939971923828, + "learning_rate": 5.775419606415233e-06, + "loss": 9.702054977416992, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.3922365717589855, + "loss_total_v6": 0.6063784509897232, + "step": 2063 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47602162162162165, + "grad_norm": 23.883180618286133, + "learning_rate": 5.771667468840128e-06, + "loss": 9.80018424987793, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.4052819311618805, + "loss_total_v6": 0.612511545419693, + "step": 2064 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47625225225225226, + "grad_norm": 21.463211059570312, + "learning_rate": 5.767914886050981e-06, + "loss": 9.83033561706543, + "loss_alignment": 0.4130859375, + "loss_alignment_w": 0.20654296875, + "loss_sft": 0.40776147320866585, + "loss_total_v6": 0.6143959760665894, + "step": 2065 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4764828828828829, + "grad_norm": 22.07005500793457, + "learning_rate": 5.7641618602128415e-06, + "loss": 9.71605396270752, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.40086299926042557, + "loss_total_v6": 0.6072533577680588, + "step": 2066 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4767135135135135, + "grad_norm": 43.598506927490234, + "learning_rate": 5.7604083934910235e-06, + "loss": 9.127866744995117, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.36036279797554016, + "loss_total_v6": 0.57049161195755, + "step": 2067 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47694414414414416, + "grad_norm": 19.046783447265625, + "learning_rate": 5.756654488051091e-06, + "loss": 9.49708366394043, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.37905969470739365, + "loss_total_v6": 0.5935677401721478, + "step": 2068 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47717477477477477, + "grad_norm": 25.901607513427734, + "learning_rate": 5.752900146058862e-06, + "loss": 10.361457824707031, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.43347979336977005, + "loss_total_v6": 0.6475911065936089, + "step": 2069 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4774054054054054, + "grad_norm": 25.550046920776367, + "learning_rate": 5.7491453696804075e-06, + "loss": 9.804407119750977, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.3993660546839237, + "loss_total_v6": 0.6127754598855972, + "step": 2070 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47763603603603605, + "grad_norm": 20.84769630432129, + "learning_rate": 5.74539016108205e-06, + "loss": 10.011237144470215, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.4068455323576927, + "loss_total_v6": 0.6257023140788078, + "step": 2071 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47786666666666666, + "grad_norm": 21.723825454711914, + "learning_rate": 5.741634522430353e-06, + "loss": 10.080863952636719, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.4152560457587242, + "loss_total_v6": 0.6300540119409561, + "step": 2072 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4780972972972973, + "grad_norm": 19.92972755432129, + "learning_rate": 5.737878455892141e-06, + "loss": 9.662349700927734, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.384887482970953, + "loss_total_v6": 0.6038968935608864, + "step": 2073 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47832792792792794, + "grad_norm": 22.671062469482422, + "learning_rate": 5.734121963634477e-06, + "loss": 9.373662948608398, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.3759387396275997, + "loss_total_v6": 0.5858539119362831, + "step": 2074 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47855855855855856, + "grad_norm": 20.23928451538086, + "learning_rate": 5.730365047824671e-06, + "loss": 9.84804916381836, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.40781575441360474, + "loss_total_v6": 0.6155030950903893, + "step": 2075 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47878918918918917, + "grad_norm": 20.340662002563477, + "learning_rate": 5.726607710630279e-06, + "loss": 9.94791316986084, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.4107002839446068, + "loss_total_v6": 0.6217446029186249, + "step": 2076 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47901981981981984, + "grad_norm": 20.79619789123535, + "learning_rate": 5.722849954219097e-06, + "loss": 9.99850845336914, + "loss_alignment": 0.433837890625, + "loss_alignment_w": 0.2169189453125, + "loss_sft": 0.4080793745815754, + "loss_total_v6": 0.6249067783355713, + "step": 2077 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47925045045045045, + "grad_norm": 22.174942016601562, + "learning_rate": 5.719091780759169e-06, + "loss": 9.83948040008545, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.4016496613621712, + "loss_total_v6": 0.6149675473570824, + "step": 2078 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47948108108108106, + "grad_norm": 21.548660278320312, + "learning_rate": 5.715333192418772e-06, + "loss": 9.883076667785645, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4023754969239235, + "loss_total_v6": 0.6176922917366028, + "step": 2079 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47971171171171173, + "grad_norm": 22.13936996459961, + "learning_rate": 5.711574191366427e-06, + "loss": 10.134312629699707, + "loss_alignment": 0.43408203125, + "loss_alignment_w": 0.217041015625, + "loss_sft": 0.4159568138420582, + "loss_total_v6": 0.6333945617079735, + "step": 2080 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.47994234234234234, + "grad_norm": 22.708959579467773, + "learning_rate": 5.707814779770892e-06, + "loss": 9.73623275756836, + "loss_alignment": 0.44091796875, + "loss_alignment_w": 0.220458984375, + "loss_sft": 0.38773511722683907, + "loss_total_v6": 0.6085145436227322, + "step": 2081 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48017297297297296, + "grad_norm": 22.040023803710938, + "learning_rate": 5.704054959801162e-06, + "loss": 10.388160705566406, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.4304490424692631, + "loss_total_v6": 0.6492600627243519, + "step": 2082 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4804036036036036, + "grad_norm": 22.17266845703125, + "learning_rate": 5.700294733626468e-06, + "loss": 9.663976669311523, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.3907569572329521, + "loss_total_v6": 0.6039985343813896, + "step": 2083 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48063423423423424, + "grad_norm": 22.09134292602539, + "learning_rate": 5.696534103416276e-06, + "loss": 9.72470760345459, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.3927520923316479, + "loss_total_v6": 0.6077942028641701, + "step": 2084 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48086486486486485, + "grad_norm": 21.081424713134766, + "learning_rate": 5.6927730713402804e-06, + "loss": 9.969871520996094, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.40790704637765884, + "loss_total_v6": 0.6231169998645782, + "step": 2085 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4810954954954955, + "grad_norm": 22.636934280395508, + "learning_rate": 5.6890116395684155e-06, + "loss": 10.007980346679688, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.4174146428704262, + "loss_total_v6": 0.6254987493157387, + "step": 2086 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48132612612612613, + "grad_norm": 22.014461517333984, + "learning_rate": 5.68524981027084e-06, + "loss": 9.977580070495605, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.40947217494249344, + "loss_total_v6": 0.6235987469553947, + "step": 2087 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48155675675675674, + "grad_norm": 22.558015823364258, + "learning_rate": 5.681487585617942e-06, + "loss": 10.06529426574707, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4160834737122059, + "loss_total_v6": 0.6290809065103531, + "step": 2088 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4817873873873874, + "grad_norm": 20.97618865966797, + "learning_rate": 5.677724967780341e-06, + "loss": 9.258136749267578, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.3755696006119251, + "loss_total_v6": 0.5786335617303848, + "step": 2089 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.482018018018018, + "grad_norm": 20.058528900146484, + "learning_rate": 5.673961958928882e-06, + "loss": 9.409732818603516, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.3800089880824089, + "loss_total_v6": 0.5881083570420742, + "step": 2090 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48224864864864864, + "grad_norm": 23.16936683654785, + "learning_rate": 5.670198561234635e-06, + "loss": 9.959877967834473, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.41872648894786835, + "loss_total_v6": 0.6224923580884933, + "step": 2091 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4824792792792793, + "grad_norm": 21.874940872192383, + "learning_rate": 5.666434776868895e-06, + "loss": 10.220335006713867, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.42487316206097603, + "loss_total_v6": 0.6387709155678749, + "step": 2092 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4827099099099099, + "grad_norm": 22.002567291259766, + "learning_rate": 5.66267060800318e-06, + "loss": 9.795291900634766, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.40035274252295494, + "loss_total_v6": 0.6122057735919952, + "step": 2093 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48294054054054053, + "grad_norm": 20.887619018554688, + "learning_rate": 5.658906056809228e-06, + "loss": 9.36722183227539, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.3766348473727703, + "loss_total_v6": 0.5854513794183731, + "step": 2094 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4831711711711712, + "grad_norm": 20.780826568603516, + "learning_rate": 5.655141125459001e-06, + "loss": 9.603139877319336, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.3852151334285736, + "loss_total_v6": 0.6001962199807167, + "step": 2095 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4834018018018018, + "grad_norm": 20.63888168334961, + "learning_rate": 5.651375816124679e-06, + "loss": 9.636728286743164, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.3917089104652405, + "loss_total_v6": 0.6022954806685448, + "step": 2096 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4836324324324324, + "grad_norm": 20.116395950317383, + "learning_rate": 5.647610130978658e-06, + "loss": 9.67604923248291, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.3963179811835289, + "loss_total_v6": 0.6047530546784401, + "step": 2097 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48386306306306304, + "grad_norm": 21.39849090576172, + "learning_rate": 5.643844072193554e-06, + "loss": 10.422327995300293, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.4419380612671375, + "loss_total_v6": 0.6513954699039459, + "step": 2098 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4840936936936937, + "grad_norm": 21.36863136291504, + "learning_rate": 5.6400776419421955e-06, + "loss": 9.800971031188965, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.3957790471613407, + "loss_total_v6": 0.6125606894493103, + "step": 2099 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4843243243243243, + "grad_norm": 18.916027069091797, + "learning_rate": 5.6363108423976285e-06, + "loss": 9.138484954833984, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.35864614322781563, + "loss_total_v6": 0.5711553022265434, + "step": 2100 + }, + { + "epoch": 0.4843243243243243, + "eval_loss": 0.6612306237220764, + "eval_loss_alignment": 0.40392544738247865, + "eval_loss_alignment_w": 0.20196272369123933, + "eval_loss_sft": 0.45926786390825725, + "eval_loss_total_v6": 0.6612305881885382, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.2004, + "eval_samples_per_second": 22.701, + "eval_steps_per_second": 2.847, + "step": 2100 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48455495495495493, + "grad_norm": 23.347238540649414, + "learning_rate": 5.6325436757331096e-06, + "loss": 9.853195190429688, + "loss_alignment": 0.403076171875, + "loss_alignment_w": 0.2015380859375, + "loss_sft": 0.414179764688015, + "loss_total_v6": 0.6158246770501137, + "step": 2101 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4847855855855856, + "grad_norm": 21.63472557067871, + "learning_rate": 5.628776144122107e-06, + "loss": 10.306760787963867, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.4351423606276512, + "loss_total_v6": 0.6441725045442581, + "step": 2102 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4850162162162162, + "grad_norm": 21.037031173706055, + "learning_rate": 5.625008249738301e-06, + "loss": 10.011224746704102, + "loss_alignment": 0.440185546875, + "loss_alignment_w": 0.2200927734375, + "loss_sft": 0.40563930943608284, + "loss_total_v6": 0.6257015615701675, + "step": 2103 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4852468468468468, + "grad_norm": 21.16675567626953, + "learning_rate": 5.621239994755583e-06, + "loss": 9.933680534362793, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.40912409126758575, + "loss_total_v6": 0.6208550333976746, + "step": 2104 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4854774774774775, + "grad_norm": 22.308496475219727, + "learning_rate": 5.617471381348048e-06, + "loss": 10.208423614501953, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.42316747084259987, + "loss_total_v6": 0.6380264982581139, + "step": 2105 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4857081081081081, + "grad_norm": 22.734785079956055, + "learning_rate": 5.61370241169e-06, + "loss": 10.104924201965332, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.42055923491716385, + "loss_total_v6": 0.6315577402710915, + "step": 2106 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4859387387387387, + "grad_norm": 21.15290069580078, + "learning_rate": 5.609933087955949e-06, + "loss": 10.012383460998535, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.41131168231368065, + "loss_total_v6": 0.6257739588618279, + "step": 2107 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4861693693693694, + "grad_norm": 20.425071716308594, + "learning_rate": 5.6061634123206075e-06, + "loss": 9.783784866333008, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4027005545794964, + "loss_total_v6": 0.6114865466952324, + "step": 2108 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4864, + "grad_norm": 20.05664825439453, + "learning_rate": 5.602393386958894e-06, + "loss": 9.889289855957031, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.40059710666537285, + "loss_total_v6": 0.6180806383490562, + "step": 2109 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4866306306306306, + "grad_norm": 22.318870544433594, + "learning_rate": 5.598623014045928e-06, + "loss": 10.152711868286133, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4185257963836193, + "loss_total_v6": 0.6345444768667221, + "step": 2110 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4868612612612613, + "grad_norm": 22.334640502929688, + "learning_rate": 5.594852295757025e-06, + "loss": 9.924102783203125, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.40706055238842964, + "loss_total_v6": 0.6202563643455505, + "step": 2111 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4870918918918919, + "grad_norm": 20.253793716430664, + "learning_rate": 5.5910812342677065e-06, + "loss": 10.040828704833984, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.4120824299752712, + "loss_total_v6": 0.627551794052124, + "step": 2112 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4873225225225225, + "grad_norm": 22.907316207885742, + "learning_rate": 5.587309831753688e-06, + "loss": 10.10538387298584, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4209693670272827, + "loss_total_v6": 0.6315864771604538, + "step": 2113 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4875531531531532, + "grad_norm": 20.767736434936523, + "learning_rate": 5.583538090390882e-06, + "loss": 9.945151329040527, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.41080229356884956, + "loss_total_v6": 0.6215719431638718, + "step": 2114 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4877837837837838, + "grad_norm": 20.587142944335938, + "learning_rate": 5.579766012355397e-06, + "loss": 9.452272415161133, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.38056186959147453, + "loss_total_v6": 0.5907669812440872, + "step": 2115 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4880144144144144, + "grad_norm": 22.07573127746582, + "learning_rate": 5.575993599823536e-06, + "loss": 9.107818603515625, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.36106301471590996, + "loss_total_v6": 0.5692386738955975, + "step": 2116 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48824504504504507, + "grad_norm": 22.69104766845703, + "learning_rate": 5.572220854971796e-06, + "loss": 10.210979461669922, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4247310124337673, + "loss_total_v6": 0.6381862163543701, + "step": 2117 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4884756756756757, + "grad_norm": 22.182344436645508, + "learning_rate": 5.568447779976864e-06, + "loss": 10.034234046936035, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.4188266173005104, + "loss_total_v6": 0.6271395981311798, + "step": 2118 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4887063063063063, + "grad_norm": 20.21112823486328, + "learning_rate": 5.5646743770156176e-06, + "loss": 9.579357147216797, + "loss_alignment": 0.413818359375, + "loss_alignment_w": 0.2069091796875, + "loss_sft": 0.391007162630558, + "loss_total_v6": 0.598709799349308, + "step": 2119 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48893693693693696, + "grad_norm": 19.69521141052246, + "learning_rate": 5.560900648265124e-06, + "loss": 9.612232208251953, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.39019322395324707, + "loss_total_v6": 0.6007645130157471, + "step": 2120 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4891675675675676, + "grad_norm": 21.020402908325195, + "learning_rate": 5.557126595902638e-06, + "loss": 9.639404296875, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.39370735734701157, + "loss_total_v6": 0.6024628095328808, + "step": 2121 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4893981981981982, + "grad_norm": 23.630308151245117, + "learning_rate": 5.553352222105604e-06, + "loss": 10.163335800170898, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42250094190239906, + "loss_total_v6": 0.6352084577083588, + "step": 2122 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48962882882882885, + "grad_norm": 25.169410705566406, + "learning_rate": 5.549577529051647e-06, + "loss": 9.616939544677734, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.3911740481853485, + "loss_total_v6": 0.6010586954653263, + "step": 2123 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.48985945945945947, + "grad_norm": 22.339876174926758, + "learning_rate": 5.54580251891858e-06, + "loss": 9.560258865356445, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.38235192373394966, + "loss_total_v6": 0.5975161492824554, + "step": 2124 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4900900900900901, + "grad_norm": 20.260948181152344, + "learning_rate": 5.542027193884395e-06, + "loss": 9.922548294067383, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.41366202756762505, + "loss_total_v6": 0.6201592460274696, + "step": 2125 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4903207207207207, + "grad_norm": 22.28483009338379, + "learning_rate": 5.538251556127272e-06, + "loss": 9.920787811279297, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.41080549731850624, + "loss_total_v6": 0.6200492680072784, + "step": 2126 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49055135135135136, + "grad_norm": 23.775917053222656, + "learning_rate": 5.534475607825566e-06, + "loss": 9.450909614562988, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.38299451023340225, + "loss_total_v6": 0.5906818807125092, + "step": 2127 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.490781981981982, + "grad_norm": 19.753023147583008, + "learning_rate": 5.53069935115781e-06, + "loss": 9.969271659851074, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.4075643867254257, + "loss_total_v6": 0.6230794787406921, + "step": 2128 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4910126126126126, + "grad_norm": 23.324806213378906, + "learning_rate": 5.526922788302721e-06, + "loss": 10.460490226745605, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.43892161920666695, + "loss_total_v6": 0.6537806317210197, + "step": 2129 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49124324324324325, + "grad_norm": 18.791439056396484, + "learning_rate": 5.5231459214391905e-06, + "loss": 9.905569076538086, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4059632942080498, + "loss_total_v6": 0.619098074734211, + "step": 2130 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49147387387387387, + "grad_norm": 20.542936325073242, + "learning_rate": 5.519368752746281e-06, + "loss": 10.20938491821289, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.4240972325205803, + "loss_total_v6": 0.6380865499377251, + "step": 2131 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4917045045045045, + "grad_norm": 21.369617462158203, + "learning_rate": 5.515591284403235e-06, + "loss": 9.741787910461426, + "loss_alignment": 0.402099609375, + "loss_alignment_w": 0.2010498046875, + "loss_sft": 0.4080102667212486, + "loss_total_v6": 0.6088617146015167, + "step": 2132 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49193513513513515, + "grad_norm": 19.921676635742188, + "learning_rate": 5.511813518589462e-06, + "loss": 9.987454414367676, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4123171269893646, + "loss_total_v6": 0.6242159307003021, + "step": 2133 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49216576576576576, + "grad_norm": 22.06196403503418, + "learning_rate": 5.508035457484549e-06, + "loss": 10.112585067749023, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.41577378660440445, + "loss_total_v6": 0.6320365816354752, + "step": 2134 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4923963963963964, + "grad_norm": 24.18855094909668, + "learning_rate": 5.504257103268253e-06, + "loss": 9.946859359741211, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.407124824821949, + "loss_total_v6": 0.6216786578297615, + "step": 2135 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49262702702702704, + "grad_norm": 19.959993362426758, + "learning_rate": 5.500478458120493e-06, + "loss": 10.052671432495117, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.4137534126639366, + "loss_total_v6": 0.6282919868826866, + "step": 2136 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49285765765765766, + "grad_norm": 22.114686965942383, + "learning_rate": 5.4966995242213625e-06, + "loss": 9.485934257507324, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.37833235785365105, + "loss_total_v6": 0.592870905995369, + "step": 2137 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49308828828828827, + "grad_norm": 22.357189178466797, + "learning_rate": 5.492920303751122e-06, + "loss": 10.282232284545898, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.43157992511987686, + "loss_total_v6": 0.6426395326852798, + "step": 2138 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49331891891891894, + "grad_norm": 20.38934326171875, + "learning_rate": 5.489140798890193e-06, + "loss": 10.05574893951416, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.416371900588274, + "loss_total_v6": 0.6284843310713768, + "step": 2139 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49354954954954955, + "grad_norm": 20.761133193969727, + "learning_rate": 5.485361011819163e-06, + "loss": 9.141618728637695, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.35562241077423096, + "loss_total_v6": 0.5713511742651463, + "step": 2140 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49378018018018016, + "grad_norm": 21.677955627441406, + "learning_rate": 5.481580944718785e-06, + "loss": 9.908955574035645, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4083416499197483, + "loss_total_v6": 0.619309701025486, + "step": 2141 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49401081081081083, + "grad_norm": 22.19862937927246, + "learning_rate": 5.477800599769968e-06, + "loss": 9.694348335266113, + "loss_alignment": 0.399658203125, + "loss_alignment_w": 0.1998291015625, + "loss_sft": 0.4062202498316765, + "loss_total_v6": 0.6058967560529709, + "step": 2142 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49424144144144144, + "grad_norm": 21.17236328125, + "learning_rate": 5.4740199791537875e-06, + "loss": 10.151161193847656, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.422121562063694, + "loss_total_v6": 0.6344476006925106, + "step": 2143 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49447207207207206, + "grad_norm": 18.90487289428711, + "learning_rate": 5.470239085051473e-06, + "loss": 9.84847640991211, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.3984887860715389, + "loss_total_v6": 0.6155297756195068, + "step": 2144 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4947027027027027, + "grad_norm": 20.80988311767578, + "learning_rate": 5.466457919644415e-06, + "loss": 10.419668197631836, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.43942201882600784, + "loss_total_v6": 0.6512292698025703, + "step": 2145 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49493333333333334, + "grad_norm": 22.006364822387695, + "learning_rate": 5.462676485114157e-06, + "loss": 9.287992477416992, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.3641146160662174, + "loss_total_v6": 0.5804995074868202, + "step": 2146 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49516396396396395, + "grad_norm": 22.71560287475586, + "learning_rate": 5.458894783642402e-06, + "loss": 9.702118873596191, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.39539915695786476, + "loss_total_v6": 0.6063824370503426, + "step": 2147 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4953945945945946, + "grad_norm": 21.343887329101562, + "learning_rate": 5.455112817411006e-06, + "loss": 9.73088550567627, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.39886027574539185, + "loss_total_v6": 0.6081803366541862, + "step": 2148 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49562522522522523, + "grad_norm": 27.607614517211914, + "learning_rate": 5.451330588601976e-06, + "loss": 9.505130767822266, + "loss_alignment": 0.411376953125, + "loss_alignment_w": 0.2056884765625, + "loss_sft": 0.3877108357846737, + "loss_total_v6": 0.5940706729888916, + "step": 2149 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49585585585585584, + "grad_norm": 20.251062393188477, + "learning_rate": 5.447548099397469e-06, + "loss": 9.61054515838623, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.3882109895348549, + "loss_total_v6": 0.6006590910255909, + "step": 2150 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4960864864864865, + "grad_norm": 20.052385330200195, + "learning_rate": 5.443765351979799e-06, + "loss": 9.564001083374023, + "loss_alignment": 0.394287109375, + "loss_alignment_w": 0.1971435546875, + "loss_sft": 0.40054547414183617, + "loss_total_v6": 0.5977500639855862, + "step": 2151 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4963171171171171, + "grad_norm": 22.720129013061523, + "learning_rate": 5.439982348531423e-06, + "loss": 9.799915313720703, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.39598776400089264, + "loss_total_v6": 0.6124947145581245, + "step": 2152 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49654774774774774, + "grad_norm": 20.986112594604492, + "learning_rate": 5.436199091234945e-06, + "loss": 9.819916725158691, + "loss_alignment": 0.428955078125, + "loss_alignment_w": 0.2144775390625, + "loss_sft": 0.3992367535829544, + "loss_total_v6": 0.6137448102235794, + "step": 2153 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4967783783783784, + "grad_norm": 19.39179039001465, + "learning_rate": 5.432415582273118e-06, + "loss": 9.452167510986328, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.3843853361904621, + "loss_total_v6": 0.5907604396343231, + "step": 2154 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.497009009009009, + "grad_norm": 21.076396942138672, + "learning_rate": 5.428631823828843e-06, + "loss": 9.518296241760254, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.38192663341760635, + "loss_total_v6": 0.5948935449123383, + "step": 2155 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49723963963963963, + "grad_norm": 21.60085678100586, + "learning_rate": 5.4248478180851604e-06, + "loss": 9.984395980834961, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.40884527936577797, + "loss_total_v6": 0.6240247413516045, + "step": 2156 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49747027027027024, + "grad_norm": 21.577884674072266, + "learning_rate": 5.421063567225254e-06, + "loss": 9.440717697143555, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.38225070014595985, + "loss_total_v6": 0.590044878423214, + "step": 2157 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4977009009009009, + "grad_norm": 21.278091430664062, + "learning_rate": 5.41727907343245e-06, + "loss": 9.490072250366211, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.38167325034737587, + "loss_total_v6": 0.59312953799963, + "step": 2158 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4979315315315315, + "grad_norm": 20.565532684326172, + "learning_rate": 5.413494338890215e-06, + "loss": 9.682750701904297, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.39432597905397415, + "loss_total_v6": 0.6051719188690186, + "step": 2159 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49816216216216214, + "grad_norm": 19.78072738647461, + "learning_rate": 5.4097093657821545e-06, + "loss": 9.856422424316406, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4028459005057812, + "loss_total_v6": 0.6160264387726784, + "step": 2160 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4983927927927928, + "grad_norm": 20.955171585083008, + "learning_rate": 5.405924156292008e-06, + "loss": 9.53248405456543, + "loss_alignment": 0.41162109375, + "loss_alignment_w": 0.205810546875, + "loss_sft": 0.38987813517451286, + "loss_total_v6": 0.595780223608017, + "step": 2161 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4986234234234234, + "grad_norm": 22.43912124633789, + "learning_rate": 5.40213871260366e-06, + "loss": 9.567890167236328, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.3867351934313774, + "loss_total_v6": 0.5979931354522705, + "step": 2162 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.49885405405405403, + "grad_norm": 22.325489044189453, + "learning_rate": 5.39835303690112e-06, + "loss": 10.144312858581543, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.42115938663482666, + "loss_total_v6": 0.6340195387601852, + "step": 2163 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4990846846846847, + "grad_norm": 21.6751651763916, + "learning_rate": 5.394567131368539e-06, + "loss": 9.959781646728516, + "loss_alignment": 0.445068359375, + "loss_alignment_w": 0.2225341796875, + "loss_sft": 0.40010474994778633, + "loss_total_v6": 0.6224863529205322, + "step": 2164 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4993153153153153, + "grad_norm": 19.716047286987305, + "learning_rate": 5.390780998190197e-06, + "loss": 10.036460876464844, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.42284154519438744, + "loss_total_v6": 0.6272788196802139, + "step": 2165 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4995459459459459, + "grad_norm": 21.485595703125, + "learning_rate": 5.3869946395505045e-06, + "loss": 10.247310638427734, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4299924075603485, + "loss_total_v6": 0.6404568925499916, + "step": 2166 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.4997765765765766, + "grad_norm": 21.36471939086914, + "learning_rate": 5.383208057634004e-06, + "loss": 9.609569549560547, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.3830687738955021, + "loss_total_v6": 0.6005980856716633, + "step": 2167 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5000072072072073, + "grad_norm": 23.31437873840332, + "learning_rate": 5.379421254625367e-06, + "loss": 9.520143508911133, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.3807298019528389, + "loss_total_v6": 0.5950089693069458, + "step": 2168 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5002378378378378, + "grad_norm": 22.87653350830078, + "learning_rate": 5.375634232709392e-06, + "loss": 9.704977989196777, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.3938383609056473, + "loss_total_v6": 0.6065611392259598, + "step": 2169 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5004684684684685, + "grad_norm": 20.63654899597168, + "learning_rate": 5.371846994071005e-06, + "loss": 9.72766399383545, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.3939897269010544, + "loss_total_v6": 0.6079789996147156, + "step": 2170 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5006990990990992, + "grad_norm": 19.910558700561523, + "learning_rate": 5.368059540895252e-06, + "loss": 10.087053298950195, + "loss_alignment": 0.444580078125, + "loss_alignment_w": 0.2222900390625, + "loss_sft": 0.408440712839365, + "loss_total_v6": 0.6304408237338066, + "step": 2171 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5009297297297297, + "grad_norm": 22.12617301940918, + "learning_rate": 5.364271875367311e-06, + "loss": 9.570220947265625, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.39205363020300865, + "loss_total_v6": 0.598138839006424, + "step": 2172 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5011603603603604, + "grad_norm": 20.643964767456055, + "learning_rate": 5.360483999672475e-06, + "loss": 9.104803085327148, + "loss_alignment": 0.413330078125, + "loss_alignment_w": 0.2066650390625, + "loss_sft": 0.36253779381513596, + "loss_total_v6": 0.5690502300858498, + "step": 2173 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5013909909909909, + "grad_norm": 19.808900833129883, + "learning_rate": 5.356695915996162e-06, + "loss": 9.79305648803711, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.39829038828611374, + "loss_total_v6": 0.6120660156011581, + "step": 2174 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5016216216216216, + "grad_norm": 20.434656143188477, + "learning_rate": 5.3529076265239114e-06, + "loss": 9.89803695678711, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4046380743384361, + "loss_total_v6": 0.6186273247003555, + "step": 2175 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5018522522522523, + "grad_norm": 21.794004440307617, + "learning_rate": 5.349119133441375e-06, + "loss": 9.767603874206543, + "loss_alignment": 0.41259765625, + "loss_alignment_w": 0.206298828125, + "loss_sft": 0.40431367605924606, + "loss_total_v6": 0.610475204885006, + "step": 2176 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5020828828828828, + "grad_norm": 20.86451530456543, + "learning_rate": 5.345330438934331e-06, + "loss": 9.567346572875977, + "loss_alignment": 0.402587890625, + "loss_alignment_w": 0.2012939453125, + "loss_sft": 0.3968026041984558, + "loss_total_v6": 0.5979591980576515, + "step": 2177 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5023135135135135, + "grad_norm": 20.451072692871094, + "learning_rate": 5.341541545188666e-06, + "loss": 9.861639022827148, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.4037669934332371, + "loss_total_v6": 0.6163524463772774, + "step": 2178 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5025441441441442, + "grad_norm": 20.05210304260254, + "learning_rate": 5.337752454390387e-06, + "loss": 9.759675979614258, + "loss_alignment": 0.4384765625, + "loss_alignment_w": 0.21923828125, + "loss_sft": 0.3903142958879471, + "loss_total_v6": 0.6099798008799553, + "step": 2179 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5027747747747747, + "grad_norm": 26.51272964477539, + "learning_rate": 5.3339631687256085e-06, + "loss": 9.939983367919922, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.408587209880352, + "loss_total_v6": 0.6212489530444145, + "step": 2180 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5030054054054054, + "grad_norm": 25.588233947753906, + "learning_rate": 5.3301736903805646e-06, + "loss": 9.847343444824219, + "loss_alignment": 0.41162109375, + "loss_alignment_w": 0.205810546875, + "loss_sft": 0.40963314473629, + "loss_total_v6": 0.6154589504003525, + "step": 2181 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5032360360360361, + "grad_norm": 18.4885196685791, + "learning_rate": 5.326384021541597e-06, + "loss": 9.618765830993652, + "loss_alignment": 0.409912109375, + "loss_alignment_w": 0.2049560546875, + "loss_sft": 0.39659828692674637, + "loss_total_v6": 0.6011728718876839, + "step": 2182 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5034666666666666, + "grad_norm": 21.9115047454834, + "learning_rate": 5.322594164395159e-06, + "loss": 10.164861679077148, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.4258617199957371, + "loss_total_v6": 0.6353038772940636, + "step": 2183 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5036972972972973, + "grad_norm": 23.66299819946289, + "learning_rate": 5.318804121127808e-06, + "loss": 9.612492561340332, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.39312392100691795, + "loss_total_v6": 0.600780799984932, + "step": 2184 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.503927927927928, + "grad_norm": 20.25905990600586, + "learning_rate": 5.315013893926212e-06, + "loss": 10.220151901245117, + "loss_alignment": 0.431396484375, + "loss_alignment_w": 0.2156982421875, + "loss_sft": 0.42293915152549744, + "loss_total_v6": 0.6387594863772392, + "step": 2185 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5041585585585585, + "grad_norm": 21.1856689453125, + "learning_rate": 5.311223484977148e-06, + "loss": 10.135051727294922, + "loss_alignment": 0.409912109375, + "loss_alignment_w": 0.2049560546875, + "loss_sft": 0.42845412343740463, + "loss_total_v6": 0.6334407031536102, + "step": 2186 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5043891891891892, + "grad_norm": 21.988672256469727, + "learning_rate": 5.307432896467491e-06, + "loss": 9.344820022583008, + "loss_alignment": 0.417236328125, + "loss_alignment_w": 0.2086181640625, + "loss_sft": 0.37517374381422997, + "loss_total_v6": 0.5840512961149216, + "step": 2187 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5046198198198198, + "grad_norm": 22.649219512939453, + "learning_rate": 5.303642130584228e-06, + "loss": 9.708629608154297, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.39624854177236557, + "loss_total_v6": 0.606789343059063, + "step": 2188 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5048504504504504, + "grad_norm": 21.713014602661133, + "learning_rate": 5.29985118951444e-06, + "loss": 9.874870300292969, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.40079449489712715, + "loss_total_v6": 0.6171793937683105, + "step": 2189 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5050810810810811, + "grad_norm": 20.458377838134766, + "learning_rate": 5.296060075445315e-06, + "loss": 9.676401138305664, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.3971944637596607, + "loss_total_v6": 0.6047750487923622, + "step": 2190 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5053117117117117, + "grad_norm": 19.633346557617188, + "learning_rate": 5.292268790564138e-06, + "loss": 9.952964782714844, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.40591955929994583, + "loss_total_v6": 0.6220603063702583, + "step": 2191 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5055423423423423, + "grad_norm": 21.66703224182129, + "learning_rate": 5.288477337058293e-06, + "loss": 9.999128341674805, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.4170444831252098, + "loss_total_v6": 0.6249454841017723, + "step": 2192 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.505772972972973, + "grad_norm": 20.036462783813477, + "learning_rate": 5.284685717115262e-06, + "loss": 9.416841506958008, + "loss_alignment": 0.404052734375, + "loss_alignment_w": 0.2020263671875, + "loss_sft": 0.3867855779826641, + "loss_total_v6": 0.5885525643825531, + "step": 2193 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5060036036036036, + "grad_norm": 21.35658073425293, + "learning_rate": 5.2808939329226215e-06, + "loss": 9.68527603149414, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.3939649946987629, + "loss_total_v6": 0.6053297445178032, + "step": 2194 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5062342342342342, + "grad_norm": 21.2827091217041, + "learning_rate": 5.2771019866680465e-06, + "loss": 9.134992599487305, + "loss_alignment": 0.41015625, + "loss_alignment_w": 0.205078125, + "loss_sft": 0.36617931351065636, + "loss_total_v6": 0.5709370374679565, + "step": 2195 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5064648648648649, + "grad_norm": 21.91835594177246, + "learning_rate": 5.273309880539301e-06, + "loss": 9.88109302520752, + "loss_alignment": 0.43310546875, + "loss_alignment_w": 0.216552734375, + "loss_sft": 0.40077147632837296, + "loss_total_v6": 0.6175683364272118, + "step": 2196 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5066954954954955, + "grad_norm": 20.015302658081055, + "learning_rate": 5.2695176167242425e-06, + "loss": 9.665932655334473, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.38939911127090454, + "loss_total_v6": 0.6041207984089851, + "step": 2197 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5069261261261261, + "grad_norm": 21.53661346435547, + "learning_rate": 5.265725197410821e-06, + "loss": 10.104373931884766, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.41559626907110214, + "loss_total_v6": 0.6315234005451202, + "step": 2198 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5071567567567568, + "grad_norm": 20.360544204711914, + "learning_rate": 5.261932624787075e-06, + "loss": 9.532426834106445, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.38396938517689705, + "loss_total_v6": 0.5957766324281693, + "step": 2199 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5073873873873874, + "grad_norm": 22.49850845336914, + "learning_rate": 5.258139901041131e-06, + "loss": 10.164098739624023, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.42090070620179176, + "loss_total_v6": 0.6352561563253403, + "step": 2200 + }, + { + "epoch": 0.5073873873873874, + "eval_loss": 0.6593098044395447, + "eval_loss_alignment": 0.40398596087072647, + "eval_loss_alignment_w": 0.20199298043536323, + "eval_loss_sft": 0.4573167705367136, + "eval_loss_total_v6": 0.659309751043717, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.1719, + "eval_samples_per_second": 22.708, + "eval_steps_per_second": 2.848, + "step": 2200 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.507618018018018, + "grad_norm": 24.588359832763672, + "learning_rate": 5.254347028361207e-06, + "loss": 10.028363227844238, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4185512438416481, + "loss_total_v6": 0.6267727017402649, + "step": 2201 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5078486486486486, + "grad_norm": 24.801286697387695, + "learning_rate": 5.250554008935596e-06, + "loss": 9.798727989196777, + "loss_alignment": 0.434326171875, + "loss_alignment_w": 0.2171630859375, + "loss_sft": 0.3951963670551777, + "loss_total_v6": 0.6124204769730568, + "step": 2202 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5080792792792793, + "grad_norm": 22.653234481811523, + "learning_rate": 5.246760844952692e-06, + "loss": 9.91472053527832, + "loss_alignment": 0.448486328125, + "loss_alignment_w": 0.2242431640625, + "loss_sft": 0.3952590636909008, + "loss_total_v6": 0.6196700632572174, + "step": 2203 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5083099099099099, + "grad_norm": 22.432662963867188, + "learning_rate": 5.242967538600957e-06, + "loss": 9.714027404785156, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.3925117775797844, + "loss_total_v6": 0.6071266606450081, + "step": 2204 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5085405405405405, + "grad_norm": 22.334259033203125, + "learning_rate": 5.239174092068944e-06, + "loss": 9.253238677978516, + "loss_alignment": 0.39990234375, + "loss_alignment_w": 0.199951171875, + "loss_sft": 0.3781168460845947, + "loss_total_v6": 0.5783274285495281, + "step": 2205 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5087711711711712, + "grad_norm": 29.899311065673828, + "learning_rate": 5.235380507545282e-06, + "loss": 9.412254333496094, + "loss_alignment": 0.3994140625, + "loss_alignment_w": 0.19970703125, + "loss_sft": 0.38837572932243347, + "loss_total_v6": 0.5882658511400223, + "step": 2206 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5090018018018018, + "grad_norm": 19.93342399597168, + "learning_rate": 5.231586787218687e-06, + "loss": 9.392624855041504, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.3754912205040455, + "loss_total_v6": 0.5870390608906746, + "step": 2207 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5092324324324324, + "grad_norm": 21.574668884277344, + "learning_rate": 5.2277929332779435e-06, + "loss": 10.089841842651367, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.42492668330669403, + "loss_total_v6": 0.6306151524186134, + "step": 2208 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5094630630630631, + "grad_norm": 23.60111427307129, + "learning_rate": 5.22399894791192e-06, + "loss": 9.565563201904297, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.37854834645986557, + "loss_total_v6": 0.5978476703166962, + "step": 2209 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5096936936936937, + "grad_norm": 21.50331687927246, + "learning_rate": 5.220204833309559e-06, + "loss": 9.602030754089355, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.3861071616411209, + "loss_total_v6": 0.6001269370317459, + "step": 2210 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5099243243243243, + "grad_norm": 21.073131561279297, + "learning_rate": 5.216410591659877e-06, + "loss": 10.049686431884766, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.40924859046936035, + "loss_total_v6": 0.6281053870916367, + "step": 2211 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.510154954954955, + "grad_norm": 21.04270362854004, + "learning_rate": 5.212616225151965e-06, + "loss": 10.387822151184082, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.4368975944817066, + "loss_total_v6": 0.6492389068007469, + "step": 2212 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5103855855855856, + "grad_norm": 23.674365997314453, + "learning_rate": 5.208821735974984e-06, + "loss": 9.90290641784668, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4090774953365326, + "loss_total_v6": 0.6189316064119339, + "step": 2213 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5106162162162162, + "grad_norm": 21.34504508972168, + "learning_rate": 5.205027126318167e-06, + "loss": 9.730659484863281, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.400204174220562, + "loss_total_v6": 0.6081662103533745, + "step": 2214 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5108468468468469, + "grad_norm": 20.722858428955078, + "learning_rate": 5.201232398370817e-06, + "loss": 9.242324829101562, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.36281686276197433, + "loss_total_v6": 0.5776453614234924, + "step": 2215 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5110774774774774, + "grad_norm": 20.23322296142578, + "learning_rate": 5.1974375543223046e-06, + "loss": 9.335948944091797, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.3760688081383705, + "loss_total_v6": 0.5834967941045761, + "step": 2216 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5113081081081081, + "grad_norm": 20.507802963256836, + "learning_rate": 5.193642596362068e-06, + "loss": 9.910552978515625, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.40673257783055305, + "loss_total_v6": 0.6194095760583878, + "step": 2217 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5115387387387388, + "grad_norm": 22.25779151916504, + "learning_rate": 5.1898475266796115e-06, + "loss": 10.501585006713867, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.4387435093522072, + "loss_total_v6": 0.6563491225242615, + "step": 2218 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5117693693693693, + "grad_norm": 20.866470336914062, + "learning_rate": 5.186052347464498e-06, + "loss": 9.587604522705078, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.3895389959216118, + "loss_total_v6": 0.5992252603173256, + "step": 2219 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.512, + "grad_norm": 22.62009048461914, + "learning_rate": 5.182257060906365e-06, + "loss": 10.017033576965332, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.41715652868151665, + "loss_total_v6": 0.6260646358132362, + "step": 2220 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5122306306306307, + "grad_norm": 20.121370315551758, + "learning_rate": 5.178461669194903e-06, + "loss": 10.272316932678223, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.4299531318247318, + "loss_total_v6": 0.6420197784900665, + "step": 2221 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5124612612612612, + "grad_norm": 20.45662498474121, + "learning_rate": 5.174666174519866e-06, + "loss": 9.965850830078125, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.41383547335863113, + "loss_total_v6": 0.6228656247258186, + "step": 2222 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5126918918918919, + "grad_norm": 20.178176879882812, + "learning_rate": 5.170870579071064e-06, + "loss": 9.792451858520508, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.39712346717715263, + "loss_total_v6": 0.6120282523334026, + "step": 2223 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5129225225225226, + "grad_norm": 21.83248519897461, + "learning_rate": 5.1670748850383734e-06, + "loss": 9.818182945251465, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.39629023149609566, + "loss_total_v6": 0.613636426627636, + "step": 2224 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5131531531531531, + "grad_norm": 20.614665985107422, + "learning_rate": 5.1632790946117195e-06, + "loss": 9.93941593170166, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.40980294719338417, + "loss_total_v6": 0.6212134808301926, + "step": 2225 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5133837837837838, + "grad_norm": 21.248498916625977, + "learning_rate": 5.159483209981085e-06, + "loss": 9.586423873901367, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.38655078038573265, + "loss_total_v6": 0.599151499569416, + "step": 2226 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5136144144144145, + "grad_norm": 21.0809268951416, + "learning_rate": 5.155687233336511e-06, + "loss": 10.044753074645996, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.41101545095443726, + "loss_total_v6": 0.6277970671653748, + "step": 2227 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.513845045045045, + "grad_norm": 20.430591583251953, + "learning_rate": 5.151891166868086e-06, + "loss": 9.674217224121094, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.39573049172759056, + "loss_total_v6": 0.6046385690569878, + "step": 2228 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5140756756756757, + "grad_norm": 23.044755935668945, + "learning_rate": 5.148095012765956e-06, + "loss": 9.116024017333984, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.36204883828759193, + "loss_total_v6": 0.5697514712810516, + "step": 2229 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5143063063063064, + "grad_norm": 23.418405532836914, + "learning_rate": 5.144298773220311e-06, + "loss": 9.994932174682617, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.4165076054632664, + "loss_total_v6": 0.624683253467083, + "step": 2230 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5145369369369369, + "grad_norm": 24.328224182128906, + "learning_rate": 5.140502450421396e-06, + "loss": 9.855752944946289, + "loss_alignment": 0.42919921875, + "loss_alignment_w": 0.214599609375, + "loss_sft": 0.40149179100990295, + "loss_total_v6": 0.6159845776855946, + "step": 2231 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5147675675675676, + "grad_norm": 22.31109619140625, + "learning_rate": 5.136706046559501e-06, + "loss": 9.743520736694336, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.3952554576098919, + "loss_total_v6": 0.6089700683951378, + "step": 2232 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5149981981981981, + "grad_norm": 21.364391326904297, + "learning_rate": 5.132909563824964e-06, + "loss": 9.895801544189453, + "loss_alignment": 0.438232421875, + "loss_alignment_w": 0.2191162109375, + "loss_sft": 0.39944760873913765, + "loss_total_v6": 0.6184875667095184, + "step": 2233 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5152288288288288, + "grad_norm": 20.894559860229492, + "learning_rate": 5.129113004408168e-06, + "loss": 10.32012939453125, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4344826079905033, + "loss_total_v6": 0.6450080797076225, + "step": 2234 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5154594594594595, + "grad_norm": 21.413599014282227, + "learning_rate": 5.1253163704995425e-06, + "loss": 9.597787857055664, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.3899313248693943, + "loss_total_v6": 0.5998617485165596, + "step": 2235 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.51569009009009, + "grad_norm": 23.450597763061523, + "learning_rate": 5.121519664289553e-06, + "loss": 10.020170211791992, + "loss_alignment": 0.44384765625, + "loss_alignment_w": 0.221923828125, + "loss_sft": 0.4045962691307068, + "loss_total_v6": 0.6262606903910637, + "step": 2236 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5159207207207207, + "grad_norm": 21.088651657104492, + "learning_rate": 5.117722887968717e-06, + "loss": 10.071709632873535, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.41794922947883606, + "loss_total_v6": 0.6294818222522736, + "step": 2237 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5161513513513514, + "grad_norm": 25.15789031982422, + "learning_rate": 5.113926043727584e-06, + "loss": 9.774698257446289, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.40085093677043915, + "loss_total_v6": 0.610918678343296, + "step": 2238 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5163819819819819, + "grad_norm": 21.260000228881836, + "learning_rate": 5.110129133756746e-06, + "loss": 9.986114501953125, + "loss_alignment": 0.40087890625, + "loss_alignment_w": 0.200439453125, + "loss_sft": 0.42389101535081863, + "loss_total_v6": 0.6241321191191673, + "step": 2239 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5166126126126126, + "grad_norm": 22.061176300048828, + "learning_rate": 5.106332160246834e-06, + "loss": 9.366521835327148, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.3766520991921425, + "loss_total_v6": 0.5854075849056244, + "step": 2240 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5168432432432433, + "grad_norm": 22.222869873046875, + "learning_rate": 5.102535125388511e-06, + "loss": 9.734724044799805, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.3953312300145626, + "loss_total_v6": 0.6084202229976654, + "step": 2241 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5170738738738738, + "grad_norm": 22.520042419433594, + "learning_rate": 5.098738031372484e-06, + "loss": 9.671466827392578, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.3904164098203182, + "loss_total_v6": 0.6044666841626167, + "step": 2242 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5173045045045045, + "grad_norm": 23.072057723999023, + "learning_rate": 5.094940880389485e-06, + "loss": 10.044645309448242, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.41825665906071663, + "loss_total_v6": 0.6277903467416763, + "step": 2243 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5175351351351352, + "grad_norm": 22.55268669128418, + "learning_rate": 5.091143674630283e-06, + "loss": 9.51641845703125, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.38521189987659454, + "loss_total_v6": 0.5947761237621307, + "step": 2244 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5177657657657657, + "grad_norm": 19.83746910095215, + "learning_rate": 5.08734641628568e-06, + "loss": 9.634139060974121, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.3927525654435158, + "loss_total_v6": 0.6021336764097214, + "step": 2245 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5179963963963964, + "grad_norm": 20.229202270507812, + "learning_rate": 5.083549107546505e-06, + "loss": 10.38980484008789, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.43609070405364037, + "loss_total_v6": 0.649362787604332, + "step": 2246 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.518227027027027, + "grad_norm": 23.635217666625977, + "learning_rate": 5.079751750603618e-06, + "loss": 9.807963371276855, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.40085477009415627, + "loss_total_v6": 0.6129977107048035, + "step": 2247 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5184576576576576, + "grad_norm": 23.07763671875, + "learning_rate": 5.075954347647909e-06, + "loss": 9.500410079956055, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.38154108449816704, + "loss_total_v6": 0.5937756299972534, + "step": 2248 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5186882882882883, + "grad_norm": 21.750900268554688, + "learning_rate": 5.072156900870288e-06, + "loss": 9.987107276916504, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.4101591594517231, + "loss_total_v6": 0.6241941973567009, + "step": 2249 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.518918918918919, + "grad_norm": 20.774494171142578, + "learning_rate": 5.0683594124616975e-06, + "loss": 9.609283447265625, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.38881878554821014, + "loss_total_v6": 0.6005802154541016, + "step": 2250 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5191495495495495, + "grad_norm": 20.515422821044922, + "learning_rate": 5.064561884613099e-06, + "loss": 9.199474334716797, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.367661215364933, + "loss_total_v6": 0.5749671123921871, + "step": 2251 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5193801801801802, + "grad_norm": 21.13821029663086, + "learning_rate": 5.06076431951548e-06, + "loss": 9.680047988891602, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.3938518762588501, + "loss_total_v6": 0.6050029918551445, + "step": 2252 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5196108108108108, + "grad_norm": 21.44472885131836, + "learning_rate": 5.056966719359847e-06, + "loss": 9.911376953125, + "loss_alignment": 0.408935546875, + "loss_alignment_w": 0.2044677734375, + "loss_sft": 0.4149169996380806, + "loss_total_v6": 0.6194610521197319, + "step": 2253 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5198414414414414, + "grad_norm": 20.514896392822266, + "learning_rate": 5.0531690863372284e-06, + "loss": 9.48922061920166, + "loss_alignment": 0.4013671875, + "loss_alignment_w": 0.20068359375, + "loss_sft": 0.39254527539014816, + "loss_total_v6": 0.5930762812495232, + "step": 2254 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5200720720720721, + "grad_norm": 21.809778213500977, + "learning_rate": 5.0493714226386715e-06, + "loss": 10.158157348632812, + "loss_alignment": 0.432373046875, + "loss_alignment_w": 0.2161865234375, + "loss_sft": 0.4188356511294842, + "loss_total_v6": 0.6348848566412926, + "step": 2255 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5203027027027027, + "grad_norm": 21.028810501098633, + "learning_rate": 5.045573730455241e-06, + "loss": 9.564966201782227, + "loss_alignment": 0.403076171875, + "loss_alignment_w": 0.2015380859375, + "loss_sft": 0.39550937339663506, + "loss_total_v6": 0.5978103876113892, + "step": 2256 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5205333333333333, + "grad_norm": 23.276493072509766, + "learning_rate": 5.041776011978016e-06, + "loss": 10.139049530029297, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.42409584671258926, + "loss_total_v6": 0.6336905881762505, + "step": 2257 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.520763963963964, + "grad_norm": 21.542499542236328, + "learning_rate": 5.037978269398094e-06, + "loss": 9.502920150756836, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.38230833411216736, + "loss_total_v6": 0.5939325019717216, + "step": 2258 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5209945945945946, + "grad_norm": 20.969717025756836, + "learning_rate": 5.034180504906585e-06, + "loss": 9.410411834716797, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.3781745545566082, + "loss_total_v6": 0.5881507471203804, + "step": 2259 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5212252252252252, + "grad_norm": 21.29707145690918, + "learning_rate": 5.030382720694612e-06, + "loss": 9.519021987915039, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.3852068595588207, + "loss_total_v6": 0.5949389189481735, + "step": 2260 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5214558558558559, + "grad_norm": 20.759658813476562, + "learning_rate": 5.026584918953309e-06, + "loss": 10.109859466552734, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.4245450720191002, + "loss_total_v6": 0.6318662092089653, + "step": 2261 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5216864864864865, + "grad_norm": 20.77297019958496, + "learning_rate": 5.022787101873818e-06, + "loss": 10.013456344604492, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.41359124332666397, + "loss_total_v6": 0.6258409917354584, + "step": 2262 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5219171171171171, + "grad_norm": 20.316818237304688, + "learning_rate": 5.018989271647297e-06, + "loss": 9.779605865478516, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.40314121916890144, + "loss_total_v6": 0.6112253293395042, + "step": 2263 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5221477477477477, + "grad_norm": 30.266468048095703, + "learning_rate": 5.015191430464902e-06, + "loss": 9.076126098632812, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.3541383817791939, + "loss_total_v6": 0.5672578662633896, + "step": 2264 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5223783783783784, + "grad_norm": 24.178176879882812, + "learning_rate": 5.0113935805178e-06, + "loss": 10.073471069335938, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.4149923175573349, + "loss_total_v6": 0.6295919343829155, + "step": 2265 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.522609009009009, + "grad_norm": 21.237428665161133, + "learning_rate": 5.007595723997165e-06, + "loss": 10.006750106811523, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.411127507686615, + "loss_total_v6": 0.625421941280365, + "step": 2266 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5228396396396396, + "grad_norm": 20.669824600219727, + "learning_rate": 5.003797863094172e-06, + "loss": 9.695707321166992, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.3914278820157051, + "loss_total_v6": 0.6059817373752594, + "step": 2267 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5230702702702703, + "grad_norm": 22.083147048950195, + "learning_rate": 5e-06, + "loss": 9.501291275024414, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.386646818369627, + "loss_total_v6": 0.5938306599855423, + "step": 2268 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5233009009009009, + "grad_norm": 22.54586410522461, + "learning_rate": 4.996202136905828e-06, + "loss": 9.983745574951172, + "loss_alignment": 0.434814453125, + "loss_alignment_w": 0.2174072265625, + "loss_sft": 0.40607331693172455, + "loss_total_v6": 0.6239840686321259, + "step": 2269 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5235315315315315, + "grad_norm": 20.864500045776367, + "learning_rate": 4.9924042760028355e-06, + "loss": 10.340319633483887, + "loss_alignment": 0.436767578125, + "loss_alignment_w": 0.2183837890625, + "loss_sft": 0.4279167242348194, + "loss_total_v6": 0.6462699696421623, + "step": 2270 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5237621621621622, + "grad_norm": 22.78708267211914, + "learning_rate": 4.9886064194822e-06, + "loss": 9.672133445739746, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.3847970552742481, + "loss_total_v6": 0.6045083552598953, + "step": 2271 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5239927927927928, + "grad_norm": 20.332687377929688, + "learning_rate": 4.984808569535101e-06, + "loss": 9.800437927246094, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.40284112840890884, + "loss_total_v6": 0.6125274077057838, + "step": 2272 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5242234234234234, + "grad_norm": 21.128625869750977, + "learning_rate": 4.981010728352705e-06, + "loss": 9.555896759033203, + "loss_alignment": 0.40576171875, + "loss_alignment_w": 0.202880859375, + "loss_sft": 0.3946373388171196, + "loss_total_v6": 0.5972435474395752, + "step": 2273 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5244540540540541, + "grad_norm": 22.486370086669922, + "learning_rate": 4.977212898126183e-06, + "loss": 10.02944564819336, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.41166089475154877, + "loss_total_v6": 0.6268403604626656, + "step": 2274 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5246846846846847, + "grad_norm": 19.646638870239258, + "learning_rate": 4.973415081046692e-06, + "loss": 9.576268196105957, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.38510737195611, + "loss_total_v6": 0.5985167920589447, + "step": 2275 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5249153153153153, + "grad_norm": 22.444610595703125, + "learning_rate": 4.969617279305388e-06, + "loss": 9.936235427856445, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.41157256811857224, + "loss_total_v6": 0.6210147216916084, + "step": 2276 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.525145945945946, + "grad_norm": 19.617759704589844, + "learning_rate": 4.965819495093417e-06, + "loss": 9.807212829589844, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.40172335878014565, + "loss_total_v6": 0.6129507794976234, + "step": 2277 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5253765765765765, + "grad_norm": 21.087080001831055, + "learning_rate": 4.9620217306019075e-06, + "loss": 9.381490707397461, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.3709653988480568, + "loss_total_v6": 0.5863431990146637, + "step": 2278 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5256072072072072, + "grad_norm": 20.602153778076172, + "learning_rate": 4.958223988021986e-06, + "loss": 9.715315818786621, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.39539996162056923, + "loss_total_v6": 0.6072072237730026, + "step": 2279 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5258378378378379, + "grad_norm": 19.526025772094727, + "learning_rate": 4.954426269544761e-06, + "loss": 9.97949504852295, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.409149382263422, + "loss_total_v6": 0.6237184703350067, + "step": 2280 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5260684684684684, + "grad_norm": 19.723182678222656, + "learning_rate": 4.950628577361331e-06, + "loss": 10.031529426574707, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4137289933860302, + "loss_total_v6": 0.6269705817103386, + "step": 2281 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5262990990990991, + "grad_norm": 21.507823944091797, + "learning_rate": 4.946830913662773e-06, + "loss": 9.367652893066406, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.3811173029243946, + "loss_total_v6": 0.5854782685637474, + "step": 2282 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5265297297297298, + "grad_norm": 20.610363006591797, + "learning_rate": 4.943033280640154e-06, + "loss": 9.629949569702148, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.38655512034893036, + "loss_total_v6": 0.6018718928098679, + "step": 2283 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5267603603603603, + "grad_norm": 21.544004440307617, + "learning_rate": 4.939235680484522e-06, + "loss": 9.65389347076416, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.393300611525774, + "loss_total_v6": 0.6033683493733406, + "step": 2284 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.526990990990991, + "grad_norm": 21.391902923583984, + "learning_rate": 4.935438115386902e-06, + "loss": 10.246797561645508, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.43485841527581215, + "loss_total_v6": 0.6404248252511024, + "step": 2285 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5272216216216217, + "grad_norm": 22.58489418029785, + "learning_rate": 4.931640587538306e-06, + "loss": 9.785463333129883, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.40102019533514977, + "loss_total_v6": 0.6115914732217789, + "step": 2286 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5274522522522522, + "grad_norm": 20.209789276123047, + "learning_rate": 4.927843099129714e-06, + "loss": 9.729395866394043, + "loss_alignment": 0.43115234375, + "loss_alignment_w": 0.215576171875, + "loss_sft": 0.3925568573176861, + "loss_total_v6": 0.6080872416496277, + "step": 2287 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5276828828828829, + "grad_norm": 22.515592575073242, + "learning_rate": 4.924045652352093e-06, + "loss": 9.809350967407227, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.40106356516480446, + "loss_total_v6": 0.6130844429135323, + "step": 2288 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5279135135135136, + "grad_norm": 21.826007843017578, + "learning_rate": 4.9202482493963825e-06, + "loss": 9.638555526733398, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.39113644137978554, + "loss_total_v6": 0.602409653365612, + "step": 2289 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5281441441441441, + "grad_norm": 20.7650203704834, + "learning_rate": 4.916450892453495e-06, + "loss": 9.907876968383789, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.4144693352282047, + "loss_total_v6": 0.6192422807216644, + "step": 2290 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5283747747747748, + "grad_norm": 19.883201599121094, + "learning_rate": 4.912653583714322e-06, + "loss": 9.388774871826172, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.37074920162558556, + "loss_total_v6": 0.5867984145879745, + "step": 2291 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5286054054054055, + "grad_norm": 22.570613861083984, + "learning_rate": 4.9088563253697184e-06, + "loss": 9.858644485473633, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.40444959700107574, + "loss_total_v6": 0.6161652877926826, + "step": 2292 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.528836036036036, + "grad_norm": 19.641958236694336, + "learning_rate": 4.905059119610517e-06, + "loss": 9.66330337524414, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.3916914574801922, + "loss_total_v6": 0.6039564535021782, + "step": 2293 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5290666666666667, + "grad_norm": 22.761611938476562, + "learning_rate": 4.901261968627517e-06, + "loss": 10.059565544128418, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.4158780239522457, + "loss_total_v6": 0.6287228465080261, + "step": 2294 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5292972972972972, + "grad_norm": 21.66335678100586, + "learning_rate": 4.897464874611489e-06, + "loss": 9.93122673034668, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.4066971726715565, + "loss_total_v6": 0.6207016706466675, + "step": 2295 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5295279279279279, + "grad_norm": 22.759180068969727, + "learning_rate": 4.8936678397531675e-06, + "loss": 9.377227783203125, + "loss_alignment": 0.408203125, + "loss_alignment_w": 0.2041015625, + "loss_sft": 0.38232608884572983, + "loss_total_v6": 0.5860767364501953, + "step": 2296 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5297585585585586, + "grad_norm": 23.647916793823242, + "learning_rate": 4.889870866243254e-06, + "loss": 10.043014526367188, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.41716284304857254, + "loss_total_v6": 0.6276883706450462, + "step": 2297 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5299891891891891, + "grad_norm": 21.868175506591797, + "learning_rate": 4.886073956272418e-06, + "loss": 9.448744773864746, + "loss_alignment": 0.400146484375, + "loss_alignment_w": 0.2000732421875, + "loss_sft": 0.3905953988432884, + "loss_total_v6": 0.5905465856194496, + "step": 2298 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5302198198198198, + "grad_norm": 21.02138328552246, + "learning_rate": 4.882277112031284e-06, + "loss": 9.031229972839355, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.3533007428050041, + "loss_total_v6": 0.56445187702775, + "step": 2299 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5304504504504505, + "grad_norm": 21.294424057006836, + "learning_rate": 4.878480335710448e-06, + "loss": 9.540627479553223, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.3803010992705822, + "loss_total_v6": 0.5962892323732376, + "step": 2300 + }, + { + "epoch": 0.5304504504504505, + "eval_loss": 0.655831515789032, + "eval_loss_alignment": 0.4031283386752137, + "eval_loss_alignment_w": 0.20156416933760685, + "eval_loss_sft": 0.45426739080466777, + "eval_loss_total_v6": 0.6558315595532329, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.1466, + "eval_samples_per_second": 22.715, + "eval_steps_per_second": 2.849, + "step": 2300 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.530681081081081, + "grad_norm": 20.53217315673828, + "learning_rate": 4.87468362950046e-06, + "loss": 9.497562408447266, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.3767244182527065, + "loss_total_v6": 0.5935976058244705, + "step": 2301 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5309117117117117, + "grad_norm": 23.82728385925293, + "learning_rate": 4.870886995591832e-06, + "loss": 10.265875816345215, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.42982520535588264, + "loss_total_v6": 0.6416172310709953, + "step": 2302 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5311423423423424, + "grad_norm": 20.460769653320312, + "learning_rate": 4.8670904361750385e-06, + "loss": 9.236082077026367, + "loss_alignment": 0.399658203125, + "loss_alignment_w": 0.1998291015625, + "loss_sft": 0.37745654582977295, + "loss_total_v6": 0.5772551149129868, + "step": 2303 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5313729729729729, + "grad_norm": 22.615543365478516, + "learning_rate": 4.8632939534405e-06, + "loss": 9.776103973388672, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.39848214760422707, + "loss_total_v6": 0.611006535589695, + "step": 2304 + }, + { + "action_cond_aligned": 0.8984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5316036036036036, + "grad_norm": 20.950538635253906, + "learning_rate": 4.859497549578606e-06, + "loss": 9.445526123046875, + "loss_alignment": 0.403076171875, + "loss_alignment_w": 0.2015380859375, + "loss_sft": 0.3889751583337784, + "loss_total_v6": 0.59034538641572, + "step": 2305 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5318342342342343, + "grad_norm": 21.308652877807617, + "learning_rate": 4.8557012267796895e-06, + "loss": 9.514873504638672, + "loss_alignment": 0.437255859375, + "loss_alignment_w": 0.2186279296875, + "loss_sft": 0.376036386936903, + "loss_total_v6": 0.594679556787014, + "step": 2306 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5320648648648648, + "grad_norm": 21.768461227416992, + "learning_rate": 4.851904987234047e-06, + "loss": 9.698019027709961, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.39557017385959625, + "loss_total_v6": 0.6061262115836143, + "step": 2307 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5322954954954955, + "grad_norm": 20.765911102294922, + "learning_rate": 4.848108833131915e-06, + "loss": 9.796963691711426, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.3962305188179016, + "loss_total_v6": 0.6123102307319641, + "step": 2308 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5325261261261262, + "grad_norm": 21.252790451049805, + "learning_rate": 4.84431276666349e-06, + "loss": 9.13680648803711, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.3666742257773876, + "loss_total_v6": 0.5710504464805126, + "step": 2309 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5327567567567567, + "grad_norm": 23.403308868408203, + "learning_rate": 4.840516790018916e-06, + "loss": 9.889663696289062, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.4064951241016388, + "loss_total_v6": 0.618103988468647, + "step": 2310 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5329873873873874, + "grad_norm": 19.743900299072266, + "learning_rate": 4.836720905388281e-06, + "loss": 9.7731351852417, + "loss_alignment": 0.411376953125, + "loss_alignment_w": 0.2056884765625, + "loss_sft": 0.4045526497066021, + "loss_total_v6": 0.6108209490776062, + "step": 2311 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.533218018018018, + "grad_norm": 20.069997787475586, + "learning_rate": 4.832925114961629e-06, + "loss": 9.892406463623047, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.40759726241230965, + "loss_total_v6": 0.618275374174118, + "step": 2312 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5334486486486486, + "grad_norm": 21.67092514038086, + "learning_rate": 4.829129420928937e-06, + "loss": 10.060483932495117, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.42284759134054184, + "loss_total_v6": 0.6287802159786224, + "step": 2313 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5336792792792793, + "grad_norm": 22.964279174804688, + "learning_rate": 4.825333825480136e-06, + "loss": 9.805130004882812, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.4049043320119381, + "loss_total_v6": 0.612820602953434, + "step": 2314 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5339099099099099, + "grad_norm": 19.973939895629883, + "learning_rate": 4.821538330805098e-06, + "loss": 9.919219970703125, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4095326345413923, + "loss_total_v6": 0.6199512854218483, + "step": 2315 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5341405405405405, + "grad_norm": 20.68927764892578, + "learning_rate": 4.817742939093636e-06, + "loss": 10.30678939819336, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.43377090618014336, + "loss_total_v6": 0.644174337387085, + "step": 2316 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5343711711711712, + "grad_norm": 20.253143310546875, + "learning_rate": 4.813947652535503e-06, + "loss": 10.348175048828125, + "loss_alignment": 0.43994140625, + "loss_alignment_w": 0.219970703125, + "loss_sft": 0.42645449936389923, + "loss_total_v6": 0.6467608958482742, + "step": 2317 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5346018018018018, + "grad_norm": 21.78911018371582, + "learning_rate": 4.81015247332039e-06, + "loss": 9.786359786987305, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.39994699880480766, + "loss_total_v6": 0.611647441983223, + "step": 2318 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5348324324324324, + "grad_norm": 21.849117279052734, + "learning_rate": 4.806357403637934e-06, + "loss": 9.75184440612793, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.393303744494915, + "loss_total_v6": 0.6094902902841568, + "step": 2319 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5350630630630631, + "grad_norm": 23.119844436645508, + "learning_rate": 4.802562445677696e-06, + "loss": 9.783703804016113, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.395127110183239, + "loss_total_v6": 0.6114814952015877, + "step": 2320 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5352936936936937, + "grad_norm": 21.888208389282227, + "learning_rate": 4.798767601629184e-06, + "loss": 9.563874244689941, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.38541607186198235, + "loss_total_v6": 0.5977421253919601, + "step": 2321 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5355243243243243, + "grad_norm": 19.853639602661133, + "learning_rate": 4.794972873681834e-06, + "loss": 9.632166862487793, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.39233944565057755, + "loss_total_v6": 0.6020104587078094, + "step": 2322 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.535754954954955, + "grad_norm": 21.796518325805664, + "learning_rate": 4.791178264025017e-06, + "loss": 9.298577308654785, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.3690486401319504, + "loss_total_v6": 0.5811610594391823, + "step": 2323 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5359855855855856, + "grad_norm": 20.767709732055664, + "learning_rate": 4.7873837748480374e-06, + "loss": 9.755547523498535, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.39706001058220863, + "loss_total_v6": 0.6097217500209808, + "step": 2324 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5362162162162162, + "grad_norm": 23.34481430053711, + "learning_rate": 4.7835894083401235e-06, + "loss": 9.659521102905273, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.3960632309317589, + "loss_total_v6": 0.6037200763821602, + "step": 2325 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5364468468468468, + "grad_norm": 23.421628952026367, + "learning_rate": 4.779795166690443e-06, + "loss": 10.192462921142578, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.42363471910357475, + "loss_total_v6": 0.6370289176702499, + "step": 2326 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5366774774774775, + "grad_norm": 19.986692428588867, + "learning_rate": 4.776001052088081e-06, + "loss": 9.125385284423828, + "loss_alignment": 0.415283203125, + "loss_alignment_w": 0.2076416015625, + "loss_sft": 0.36254236847162247, + "loss_total_v6": 0.5703365653753281, + "step": 2327 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5369081081081081, + "grad_norm": 20.880830764770508, + "learning_rate": 4.772207066722057e-06, + "loss": 9.35362434387207, + "loss_alignment": 0.4345703125, + "loss_alignment_w": 0.21728515625, + "loss_sft": 0.36748426780104637, + "loss_total_v6": 0.5846015810966492, + "step": 2328 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5371387387387387, + "grad_norm": 21.115684509277344, + "learning_rate": 4.768413212781315e-06, + "loss": 9.866340637207031, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.40435075014829636, + "loss_total_v6": 0.6166462898254395, + "step": 2329 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5373693693693694, + "grad_norm": 21.22952651977539, + "learning_rate": 4.764619492454719e-06, + "loss": 9.823516845703125, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.4006519541144371, + "loss_total_v6": 0.6139698177576065, + "step": 2330 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5376, + "grad_norm": 20.390901565551758, + "learning_rate": 4.760825907931058e-06, + "loss": 9.957806587219238, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.4117458686232567, + "loss_total_v6": 0.622362919151783, + "step": 2331 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5378306306306306, + "grad_norm": 19.314512252807617, + "learning_rate": 4.757032461399044e-06, + "loss": 9.099702835083008, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.3564511574804783, + "loss_total_v6": 0.5687314458191395, + "step": 2332 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5380612612612613, + "grad_norm": 21.765193939208984, + "learning_rate": 4.753239155047311e-06, + "loss": 10.267021179199219, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.42808109149336815, + "loss_total_v6": 0.6416888684034348, + "step": 2333 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5382918918918919, + "grad_norm": 22.34674072265625, + "learning_rate": 4.7494459910644044e-06, + "loss": 9.63016414642334, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.39857716113328934, + "loss_total_v6": 0.6018852666020393, + "step": 2334 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5385225225225225, + "grad_norm": 22.6635799407959, + "learning_rate": 4.745652971638796e-06, + "loss": 9.815887451171875, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4001903869211674, + "loss_total_v6": 0.613493025302887, + "step": 2335 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5387531531531532, + "grad_norm": 20.78180694580078, + "learning_rate": 4.741860098958869e-06, + "loss": 9.684263229370117, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.3879660293459892, + "loss_total_v6": 0.6052664369344711, + "step": 2336 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5389837837837838, + "grad_norm": 21.62059211730957, + "learning_rate": 4.738067375212926e-06, + "loss": 9.97623348236084, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.407419640570879, + "loss_total_v6": 0.6235146000981331, + "step": 2337 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5392144144144144, + "grad_norm": 21.604419708251953, + "learning_rate": 4.734274802589181e-06, + "loss": 9.639751434326172, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.39035673439502716, + "loss_total_v6": 0.6024844273924828, + "step": 2338 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5394450450450451, + "grad_norm": 20.22536277770996, + "learning_rate": 4.730482383275759e-06, + "loss": 9.526790618896484, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.3848378546535969, + "loss_total_v6": 0.5954243913292885, + "step": 2339 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5396756756756756, + "grad_norm": 21.310651779174805, + "learning_rate": 4.726690119460701e-06, + "loss": 10.029202461242676, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.41948871314525604, + "loss_total_v6": 0.6268251314759254, + "step": 2340 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5399063063063063, + "grad_norm": 19.670190811157227, + "learning_rate": 4.722898013331954e-06, + "loss": 9.494007110595703, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.38651203736662865, + "loss_total_v6": 0.5933754295110703, + "step": 2341 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.540136936936937, + "grad_norm": 20.705595016479492, + "learning_rate": 4.7191060670773785e-06, + "loss": 9.307096481323242, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.37144268676638603, + "loss_total_v6": 0.5816935449838638, + "step": 2342 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5403675675675675, + "grad_norm": 22.962438583374023, + "learning_rate": 4.715314282884739e-06, + "loss": 9.539138793945312, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.38519763574004173, + "loss_total_v6": 0.5961961820721626, + "step": 2343 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5405981981981982, + "grad_norm": 23.78220558166504, + "learning_rate": 4.7115226629417075e-06, + "loss": 9.900418281555176, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4103868640959263, + "loss_total_v6": 0.6187761425971985, + "step": 2344 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5408288288288289, + "grad_norm": 24.747522354125977, + "learning_rate": 4.707731209435864e-06, + "loss": 9.866369247436523, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.40503913536667824, + "loss_total_v6": 0.6166480556130409, + "step": 2345 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5410594594594594, + "grad_norm": 21.100196838378906, + "learning_rate": 4.703939924554687e-06, + "loss": 9.300580978393555, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.3708523362874985, + "loss_total_v6": 0.5812862813472748, + "step": 2346 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5412900900900901, + "grad_norm": 23.564128875732422, + "learning_rate": 4.700148810485561e-06, + "loss": 9.887649536132812, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.4130678214132786, + "loss_total_v6": 0.617978110909462, + "step": 2347 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5415207207207208, + "grad_norm": 21.822895050048828, + "learning_rate": 4.696357869415774e-06, + "loss": 9.495121002197266, + "loss_alignment": 0.413330078125, + "loss_alignment_w": 0.2066650390625, + "loss_sft": 0.3867647349834442, + "loss_total_v6": 0.5934450328350067, + "step": 2348 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5417513513513513, + "grad_norm": 24.443557739257812, + "learning_rate": 4.692567103532509e-06, + "loss": 9.751374244689941, + "loss_alignment": 0.410400390625, + "loss_alignment_w": 0.2052001953125, + "loss_sft": 0.40458113700151443, + "loss_total_v6": 0.6094608828425407, + "step": 2349 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.541981981981982, + "grad_norm": 20.93585968017578, + "learning_rate": 4.688776515022855e-06, + "loss": 10.009096145629883, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4152718596160412, + "loss_total_v6": 0.6255684867501259, + "step": 2350 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5422126126126127, + "grad_norm": 19.591888427734375, + "learning_rate": 4.6849861060737895e-06, + "loss": 9.551111221313477, + "loss_alignment": 0.427734375, + "loss_alignment_w": 0.2138671875, + "loss_sft": 0.3829246610403061, + "loss_total_v6": 0.5969444289803505, + "step": 2351 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5424432432432432, + "grad_norm": 20.728818893432617, + "learning_rate": 4.681195878872195e-06, + "loss": 9.925947189331055, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.416956789791584, + "loss_total_v6": 0.6203717142343521, + "step": 2352 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5426738738738739, + "grad_norm": 21.527318954467773, + "learning_rate": 4.677405835604842e-06, + "loss": 9.113727569580078, + "loss_alignment": 0.390625, + "loss_alignment_w": 0.1953125, + "loss_sft": 0.3741123601794243, + "loss_total_v6": 0.5696079656481743, + "step": 2353 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5429045045045046, + "grad_norm": 21.218441009521484, + "learning_rate": 4.673615978458402e-06, + "loss": 9.843605041503906, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4054017476737499, + "loss_total_v6": 0.6152253448963165, + "step": 2354 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5431351351351351, + "grad_norm": 19.155799865722656, + "learning_rate": 4.669826309619437e-06, + "loss": 10.203857421875, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.4330139495432377, + "loss_total_v6": 0.6377411186695099, + "step": 2355 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5433657657657658, + "grad_norm": 20.22662353515625, + "learning_rate": 4.666036831274392e-06, + "loss": 9.381842613220215, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.37794535234570503, + "loss_total_v6": 0.5863651484251022, + "step": 2356 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5435963963963963, + "grad_norm": 20.73863410949707, + "learning_rate": 4.662247545609616e-06, + "loss": 9.96678352355957, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.41252054646611214, + "loss_total_v6": 0.6229239962995052, + "step": 2357 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.543827027027027, + "grad_norm": 21.47896957397461, + "learning_rate": 4.658458454811335e-06, + "loss": 9.550725936889648, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.38579974696040154, + "loss_total_v6": 0.5969203561544418, + "step": 2358 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5440576576576577, + "grad_norm": 21.440067291259766, + "learning_rate": 4.654669561065672e-06, + "loss": 9.641027450561523, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.3926948010921478, + "loss_total_v6": 0.602564200758934, + "step": 2359 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5442882882882882, + "grad_norm": 19.956676483154297, + "learning_rate": 4.650880866558626e-06, + "loss": 10.302909851074219, + "loss_alignment": 0.435791015625, + "loss_alignment_w": 0.2178955078125, + "loss_sft": 0.42568548396229744, + "loss_total_v6": 0.6439319103956223, + "step": 2360 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5445189189189189, + "grad_norm": 20.941591262817383, + "learning_rate": 4.647092373476091e-06, + "loss": 9.668161392211914, + "loss_alignment": 0.42236328125, + "loss_alignment_w": 0.211181640625, + "loss_sft": 0.39339885860681534, + "loss_total_v6": 0.604260042309761, + "step": 2361 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5447495495495496, + "grad_norm": 21.9227294921875, + "learning_rate": 4.643304084003839e-06, + "loss": 9.820627212524414, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.3988843970000744, + "loss_total_v6": 0.6137892082333565, + "step": 2362 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5449801801801801, + "grad_norm": 20.445980072021484, + "learning_rate": 4.6395160003275254e-06, + "loss": 9.634490966796875, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.391523364931345, + "loss_total_v6": 0.6021556854248047, + "step": 2363 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5452108108108108, + "grad_norm": 22.907724380493164, + "learning_rate": 4.6357281246326916e-06, + "loss": 9.371505737304688, + "loss_alignment": 0.40087890625, + "loss_alignment_w": 0.200439453125, + "loss_sft": 0.3854932077229023, + "loss_total_v6": 0.5857190638780594, + "step": 2364 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5454414414414415, + "grad_norm": 21.062101364135742, + "learning_rate": 4.631940459104749e-06, + "loss": 9.098203659057617, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.36078251525759697, + "loss_total_v6": 0.5686377584934235, + "step": 2365 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.545672072072072, + "grad_norm": 21.700380325317383, + "learning_rate": 4.628153005928998e-06, + "loss": 10.165168762207031, + "loss_alignment": 0.412109375, + "loss_alignment_w": 0.2060546875, + "loss_sft": 0.42874957993626595, + "loss_total_v6": 0.6353230848908424, + "step": 2366 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5459027027027027, + "grad_norm": 21.691051483154297, + "learning_rate": 4.624365767290609e-06, + "loss": 9.814845085144043, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.4036347232758999, + "loss_total_v6": 0.6134277954697609, + "step": 2367 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5461333333333334, + "grad_norm": 20.647790908813477, + "learning_rate": 4.620578745374634e-06, + "loss": 9.542776107788086, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.3927185945212841, + "loss_total_v6": 0.5964234545826912, + "step": 2368 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5463639639639639, + "grad_norm": 20.97162437438965, + "learning_rate": 4.616791942365997e-06, + "loss": 9.677738189697266, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.4007265754044056, + "loss_total_v6": 0.6048586517572403, + "step": 2369 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5465945945945946, + "grad_norm": 20.672344207763672, + "learning_rate": 4.613005360449496e-06, + "loss": 9.874534606933594, + "loss_alignment": 0.4375, + "loss_alignment_w": 0.21875, + "loss_sft": 0.3988356404006481, + "loss_total_v6": 0.6171584129333496, + "step": 2370 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5468252252252253, + "grad_norm": 22.247501373291016, + "learning_rate": 4.609219001809806e-06, + "loss": 9.601736068725586, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.39120038226246834, + "loss_total_v6": 0.6001084670424461, + "step": 2371 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5470558558558558, + "grad_norm": 21.71632194519043, + "learning_rate": 4.6054328686314616e-06, + "loss": 9.666780471801758, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.39537254720926285, + "loss_total_v6": 0.6041738130152225, + "step": 2372 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5472864864864865, + "grad_norm": 20.139156341552734, + "learning_rate": 4.60164696309888e-06, + "loss": 9.645572662353516, + "loss_alignment": 0.419677734375, + "loss_alignment_w": 0.2098388671875, + "loss_sft": 0.3931162543594837, + "loss_total_v6": 0.6028483025729656, + "step": 2373 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5475171171171171, + "grad_norm": 21.20462989807129, + "learning_rate": 4.5978612873963415e-06, + "loss": 9.413040161132812, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.3778505176305771, + "loss_total_v6": 0.5883149951696396, + "step": 2374 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5477477477477477, + "grad_norm": 20.072006225585938, + "learning_rate": 4.594075843707991e-06, + "loss": 9.636466979980469, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.3931727446615696, + "loss_total_v6": 0.6022791936993599, + "step": 2375 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5479783783783784, + "grad_norm": 19.33285140991211, + "learning_rate": 4.590290634217848e-06, + "loss": 9.497941970825195, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.3786555267870426, + "loss_total_v6": 0.5936213284730911, + "step": 2376 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.548209009009009, + "grad_norm": 20.959016799926758, + "learning_rate": 4.586505661109787e-06, + "loss": 9.411359786987305, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.37744035199284554, + "loss_total_v6": 0.5882100239396095, + "step": 2377 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5484396396396396, + "grad_norm": 21.244281768798828, + "learning_rate": 4.582720926567552e-06, + "loss": 9.615255355834961, + "loss_alignment": 0.40478515625, + "loss_alignment_w": 0.202392578125, + "loss_sft": 0.39860670641064644, + "loss_total_v6": 0.6009535193443298, + "step": 2378 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5486702702702703, + "grad_norm": 23.08914566040039, + "learning_rate": 4.5789364327747475e-06, + "loss": 9.519338607788086, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.3889497183263302, + "loss_total_v6": 0.5949586182832718, + "step": 2379 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5489009009009009, + "grad_norm": 20.83014678955078, + "learning_rate": 4.57515218191484e-06, + "loss": 9.828546524047852, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.40336189419031143, + "loss_total_v6": 0.6142841428518295, + "step": 2380 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5491315315315315, + "grad_norm": 20.86868667602539, + "learning_rate": 4.571368176171159e-06, + "loss": 10.057111740112305, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.4134969003498554, + "loss_total_v6": 0.6285695284605026, + "step": 2381 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5493621621621622, + "grad_norm": 19.840534210205078, + "learning_rate": 4.567584417726883e-06, + "loss": 9.42390251159668, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.3783615902066231, + "loss_total_v6": 0.5889939069747925, + "step": 2382 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5495927927927928, + "grad_norm": 21.993669509887695, + "learning_rate": 4.563800908765058e-06, + "loss": 9.730496406555176, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.3925646170973778, + "loss_total_v6": 0.6081560328602791, + "step": 2383 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5498234234234234, + "grad_norm": 22.04178237915039, + "learning_rate": 4.560017651468579e-06, + "loss": 9.65894889831543, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.38937466964125633, + "loss_total_v6": 0.6036843433976173, + "step": 2384 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.550054054054054, + "grad_norm": 21.569664001464844, + "learning_rate": 4.556234648020203e-06, + "loss": 9.204404830932617, + "loss_alignment": 0.396484375, + "loss_alignment_w": 0.1982421875, + "loss_sft": 0.37686531990766525, + "loss_total_v6": 0.5752753466367722, + "step": 2385 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5502846846846847, + "grad_norm": 21.918743133544922, + "learning_rate": 4.552451900602531e-06, + "loss": 9.251041412353516, + "loss_alignment": 0.4306640625, + "loss_alignment_w": 0.21533203125, + "loss_sft": 0.3632395267486572, + "loss_total_v6": 0.5781900882720947, + "step": 2386 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5505153153153153, + "grad_norm": 20.29781723022461, + "learning_rate": 4.548669411398025e-06, + "loss": 10.149006843566895, + "loss_alignment": 0.42578125, + "loss_alignment_w": 0.212890625, + "loss_sft": 0.4213765263557434, + "loss_total_v6": 0.6343129277229309, + "step": 2387 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.550745945945946, + "grad_norm": 24.30137062072754, + "learning_rate": 4.544887182588995e-06, + "loss": 10.211483001708984, + "loss_alignment": 0.431640625, + "loss_alignment_w": 0.2158203125, + "loss_sft": 0.4222752973437309, + "loss_total_v6": 0.63821766898036, + "step": 2388 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5509765765765766, + "grad_norm": 21.19101905822754, + "learning_rate": 4.5411052163575986e-06, + "loss": 9.342132568359375, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.37065697461366653, + "loss_total_v6": 0.5838832929730415, + "step": 2389 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5512072072072072, + "grad_norm": 24.308486938476562, + "learning_rate": 4.537323514885846e-06, + "loss": 10.064007759094238, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.4165676459670067, + "loss_total_v6": 0.6290004923939705, + "step": 2390 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5514378378378378, + "grad_norm": 20.625022888183594, + "learning_rate": 4.533542080355587e-06, + "loss": 9.590376853942871, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.39454934000968933, + "loss_total_v6": 0.5993985682725906, + "step": 2391 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5516684684684685, + "grad_norm": 20.322521209716797, + "learning_rate": 4.52976091494853e-06, + "loss": 9.872713088989258, + "loss_alignment": 0.421875, + "loss_alignment_w": 0.2109375, + "loss_sft": 0.406320683658123, + "loss_total_v6": 0.6170445531606674, + "step": 2392 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5518990990990991, + "grad_norm": 20.86411476135254, + "learning_rate": 4.525980020846214e-06, + "loss": 9.700611114501953, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.39722757041454315, + "loss_total_v6": 0.6062882244586945, + "step": 2393 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5521297297297297, + "grad_norm": 20.273574829101562, + "learning_rate": 4.522199400230033e-06, + "loss": 9.579034805297852, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.3894153796136379, + "loss_total_v6": 0.5986896827816963, + "step": 2394 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5523603603603604, + "grad_norm": 21.85337257385254, + "learning_rate": 4.518419055281217e-06, + "loss": 9.449501991271973, + "loss_alignment": 0.422607421875, + "loss_alignment_w": 0.2113037109375, + "loss_sft": 0.3790917880833149, + "loss_total_v6": 0.5905938744544983, + "step": 2395 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.552590990990991, + "grad_norm": 22.066173553466797, + "learning_rate": 4.514638988180838e-06, + "loss": 9.595181465148926, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.38770848512649536, + "loss_total_v6": 0.5996988415718079, + "step": 2396 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5528216216216216, + "grad_norm": 19.517202377319336, + "learning_rate": 4.510859201109809e-06, + "loss": 9.928168296813965, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.4160579703748226, + "loss_total_v6": 0.6205105111002922, + "step": 2397 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5530522522522523, + "grad_norm": 23.001073837280273, + "learning_rate": 4.50707969624888e-06, + "loss": 9.82538890838623, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4038207307457924, + "loss_total_v6": 0.6140868291258812, + "step": 2398 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5532828828828829, + "grad_norm": 20.365312576293945, + "learning_rate": 4.503300475778638e-06, + "loss": 9.673056602478027, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.3924230709671974, + "loss_total_v6": 0.6045660451054573, + "step": 2399 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5535135135135135, + "grad_norm": 21.984394073486328, + "learning_rate": 4.499521541879508e-06, + "loss": 10.154619216918945, + "loss_alignment": 0.4248046875, + "loss_alignment_w": 0.21240234375, + "loss_sft": 0.42256656661629677, + "loss_total_v6": 0.6346637234091759, + "step": 2400 + }, + { + "epoch": 0.5535135135135135, + "eval_loss": 0.6536677479743958, + "eval_loss_alignment": 0.40333700587606836, + "eval_loss_alignment_w": 0.20166850293803418, + "eval_loss_sft": 0.4519992578838371, + "eval_loss_total_v6": 0.6536677609492316, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 82.0126, + "eval_samples_per_second": 22.753, + "eval_steps_per_second": 2.853, + "step": 2400 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5537441441441442, + "grad_norm": 20.50492286682129, + "learning_rate": 4.495742896731748e-06, + "loss": 9.796728134155273, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.3978637680411339, + "loss_total_v6": 0.6122955232858658, + "step": 2401 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5539747747747747, + "grad_norm": 22.682018280029297, + "learning_rate": 4.4919645425154524e-06, + "loss": 9.645416259765625, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.38994787633419037, + "loss_total_v6": 0.6028384864330292, + "step": 2402 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5542054054054054, + "grad_norm": 21.261661529541016, + "learning_rate": 4.488186481410539e-06, + "loss": 9.377761840820312, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.37736987695097923, + "loss_total_v6": 0.5861101076006889, + "step": 2403 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5544360360360361, + "grad_norm": 22.013877868652344, + "learning_rate": 4.484408715596768e-06, + "loss": 9.294766426086426, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.3686883896589279, + "loss_total_v6": 0.5809228792786598, + "step": 2404 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5546666666666666, + "grad_norm": 20.70685386657715, + "learning_rate": 4.48063124725372e-06, + "loss": 10.05189037322998, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4145590737462044, + "loss_total_v6": 0.6282431334257126, + "step": 2405 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5548972972972973, + "grad_norm": 23.168933868408203, + "learning_rate": 4.47685407856081e-06, + "loss": 9.92092514038086, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4108750559389591, + "loss_total_v6": 0.6200578138232231, + "step": 2406 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.555127927927928, + "grad_norm": 22.11753273010254, + "learning_rate": 4.47307721169728e-06, + "loss": 9.455732345581055, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.38401302695274353, + "loss_total_v6": 0.5909832343459129, + "step": 2407 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5553585585585585, + "grad_norm": 20.27889060974121, + "learning_rate": 4.469300648842192e-06, + "loss": 9.350547790527344, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.37457037530839443, + "loss_total_v6": 0.5844092555344105, + "step": 2408 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5555891891891892, + "grad_norm": 20.20616340637207, + "learning_rate": 4.465524392174437e-06, + "loss": 9.714710235595703, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.39900897443294525, + "loss_total_v6": 0.6071693748235703, + "step": 2409 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5558198198198199, + "grad_norm": 20.570775985717773, + "learning_rate": 4.461748443872729e-06, + "loss": 9.485450744628906, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.38173534348607063, + "loss_total_v6": 0.5928406827151775, + "step": 2410 + }, + { + "action_cond_aligned": 0.9921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5560504504504504, + "grad_norm": 21.928287506103516, + "learning_rate": 4.457972806115607e-06, + "loss": 10.060924530029297, + "loss_alignment": 0.439453125, + "loss_alignment_w": 0.2197265625, + "loss_sft": 0.4094627574086189, + "loss_total_v6": 0.6288078427314758, + "step": 2411 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5562810810810811, + "grad_norm": 23.62209701538086, + "learning_rate": 4.454197481081423e-06, + "loss": 9.708633422851562, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.40087224170565605, + "loss_total_v6": 0.6067895740270615, + "step": 2412 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5565117117117118, + "grad_norm": 21.34016227722168, + "learning_rate": 4.450422470948354e-06, + "loss": 9.91395378112793, + "loss_alignment": 0.4052734375, + "loss_alignment_w": 0.20263671875, + "loss_sft": 0.4170312024652958, + "loss_total_v6": 0.6196221187710762, + "step": 2413 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5567423423423423, + "grad_norm": 20.46604347229004, + "learning_rate": 4.446647777894397e-06, + "loss": 9.2709321975708, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.37073881179094315, + "loss_total_v6": 0.5794332474470139, + "step": 2414 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.556972972972973, + "grad_norm": 21.807395935058594, + "learning_rate": 4.442873404097362e-06, + "loss": 10.19217300415039, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.4210684262216091, + "loss_total_v6": 0.6370108053088188, + "step": 2415 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5572036036036037, + "grad_norm": 20.935039520263672, + "learning_rate": 4.4390993517348784e-06, + "loss": 9.735738754272461, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.39945346117019653, + "loss_total_v6": 0.6084836423397064, + "step": 2416 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5574342342342342, + "grad_norm": 22.38750648498535, + "learning_rate": 4.435325622984383e-06, + "loss": 10.019439697265625, + "loss_alignment": 0.428466796875, + "loss_alignment_w": 0.2142333984375, + "loss_sft": 0.41175278276205063, + "loss_total_v6": 0.6262150183320045, + "step": 2417 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5576648648648649, + "grad_norm": 20.483442306518555, + "learning_rate": 4.431552220023138e-06, + "loss": 9.147551536560059, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.3675745949149132, + "loss_total_v6": 0.5717219412326813, + "step": 2418 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5578954954954954, + "grad_norm": 20.70728874206543, + "learning_rate": 4.427779145028206e-06, + "loss": 9.704301834106445, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.3962833471596241, + "loss_total_v6": 0.606518916785717, + "step": 2419 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5581261261261261, + "grad_norm": 21.15288543701172, + "learning_rate": 4.424006400176465e-06, + "loss": 9.793792724609375, + "loss_alignment": 0.424072265625, + "loss_alignment_w": 0.2120361328125, + "loss_sft": 0.40010638535022736, + "loss_total_v6": 0.6121120154857635, + "step": 2420 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5583567567567568, + "grad_norm": 18.736886978149414, + "learning_rate": 4.420233987644605e-06, + "loss": 9.363639831542969, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.37523603811860085, + "loss_total_v6": 0.5852274969220161, + "step": 2421 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5585873873873873, + "grad_norm": 22.01427459716797, + "learning_rate": 4.416461909609119e-06, + "loss": 9.39933967590332, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.37371354550123215, + "loss_total_v6": 0.5874586850404739, + "step": 2422 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.558818018018018, + "grad_norm": 22.048763275146484, + "learning_rate": 4.412690168246315e-06, + "loss": 9.713802337646484, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.39693810045719147, + "loss_total_v6": 0.6071126535534859, + "step": 2423 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5590486486486487, + "grad_norm": 21.25511360168457, + "learning_rate": 4.408918765732295e-06, + "loss": 9.903552055358887, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4096977040171623, + "loss_total_v6": 0.6189720183610916, + "step": 2424 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5592792792792792, + "grad_norm": 20.637712478637695, + "learning_rate": 4.405147704242976e-06, + "loss": 9.583562850952148, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.3857005424797535, + "loss_total_v6": 0.5989726558327675, + "step": 2425 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5595099099099099, + "grad_norm": 20.468597412109375, + "learning_rate": 4.401376985954073e-06, + "loss": 9.161062240600586, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.3678392395377159, + "loss_total_v6": 0.5725664012134075, + "step": 2426 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5597405405405406, + "grad_norm": 22.451709747314453, + "learning_rate": 4.397606613041106e-06, + "loss": 9.874601364135742, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.40386002883315086, + "loss_total_v6": 0.6171626374125481, + "step": 2427 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5599711711711711, + "grad_norm": 21.772350311279297, + "learning_rate": 4.393836587679394e-06, + "loss": 9.200263023376465, + "loss_alignment": 0.41064453125, + "loss_alignment_w": 0.205322265625, + "loss_sft": 0.36990780010819435, + "loss_total_v6": 0.5750164352357388, + "step": 2428 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5602018018018018, + "grad_norm": 19.92595100402832, + "learning_rate": 4.3900669120440524e-06, + "loss": 9.865819931030273, + "loss_alignment": 0.427001953125, + "loss_alignment_w": 0.2135009765625, + "loss_sft": 0.40321963652968407, + "loss_total_v6": 0.6166137680411339, + "step": 2429 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5604324324324325, + "grad_norm": 20.27335548400879, + "learning_rate": 4.386297588310002e-06, + "loss": 9.539033889770508, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.38755617663264275, + "loss_total_v6": 0.5961896032094955, + "step": 2430 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.560663063063063, + "grad_norm": 21.780935287475586, + "learning_rate": 4.382528618651953e-06, + "loss": 9.586197853088379, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.3907632939517498, + "loss_total_v6": 0.5991373434662819, + "step": 2431 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5608936936936937, + "grad_norm": 23.842174530029297, + "learning_rate": 4.378760005244417e-06, + "loss": 9.21466064453125, + "loss_alignment": 0.407958984375, + "loss_alignment_w": 0.2039794921875, + "loss_sft": 0.3720283806324005, + "loss_total_v6": 0.5759162977337837, + "step": 2432 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5611243243243244, + "grad_norm": 22.284757614135742, + "learning_rate": 4.3749917502617e-06, + "loss": 9.073869705200195, + "loss_alignment": 0.41015625, + "loss_alignment_w": 0.205078125, + "loss_sft": 0.3618250787258148, + "loss_total_v6": 0.567116878926754, + "step": 2433 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5613549549549549, + "grad_norm": 21.625944137573242, + "learning_rate": 4.371223855877895e-06, + "loss": 10.232473373413086, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.43294086307287216, + "loss_total_v6": 0.6395296007394791, + "step": 2434 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5615855855855856, + "grad_norm": 20.9649715423584, + "learning_rate": 4.367456324266893e-06, + "loss": 10.006217956542969, + "loss_alignment": 0.42431640625, + "loss_alignment_w": 0.212158203125, + "loss_sft": 0.41321514546871185, + "loss_total_v6": 0.6253886073827744, + "step": 2435 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5618162162162162, + "grad_norm": 21.25668716430664, + "learning_rate": 4.363689157602373e-06, + "loss": 9.749759674072266, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.3946230560541153, + "loss_total_v6": 0.609359972178936, + "step": 2436 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5620468468468468, + "grad_norm": 21.38791275024414, + "learning_rate": 4.359922358057807e-06, + "loss": 9.578367233276367, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.3867034241557121, + "loss_total_v6": 0.5986479856073856, + "step": 2437 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5622774774774775, + "grad_norm": 20.385635375976562, + "learning_rate": 4.356155927806448e-06, + "loss": 10.148096084594727, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4190765284001827, + "loss_total_v6": 0.6342559903860092, + "step": 2438 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5625081081081081, + "grad_norm": 22.290897369384766, + "learning_rate": 4.3523898690213414e-06, + "loss": 9.739973068237305, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.40104566514492035, + "loss_total_v6": 0.6087483316659927, + "step": 2439 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5627387387387387, + "grad_norm": 19.735675811767578, + "learning_rate": 4.3486241838753225e-06, + "loss": 9.748947143554688, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.39158158749341965, + "loss_total_v6": 0.6093092262744904, + "step": 2440 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5629693693693694, + "grad_norm": 22.459489822387695, + "learning_rate": 4.344858874540999e-06, + "loss": 9.80121898651123, + "loss_alignment": 0.41015625, + "loss_alignment_w": 0.205078125, + "loss_sft": 0.4073302000761032, + "loss_total_v6": 0.6125761792063713, + "step": 2441 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5632, + "grad_norm": 21.292802810668945, + "learning_rate": 4.3410939431907726e-06, + "loss": 9.424671173095703, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.3795235678553581, + "loss_total_v6": 0.5890420079231262, + "step": 2442 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5634306306306306, + "grad_norm": 19.985424041748047, + "learning_rate": 4.337329391996821e-06, + "loss": 9.836565017700195, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.3988277241587639, + "loss_total_v6": 0.6147853508591652, + "step": 2443 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5636612612612613, + "grad_norm": 22.59205436706543, + "learning_rate": 4.333565223131107e-06, + "loss": 9.970664024353027, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.4135718084871769, + "loss_total_v6": 0.6231665164232254, + "step": 2444 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5638918918918919, + "grad_norm": 22.751558303833008, + "learning_rate": 4.329801438765366e-06, + "loss": 10.148153305053711, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.4216894246637821, + "loss_total_v6": 0.6342595815658569, + "step": 2445 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5641225225225225, + "grad_norm": 20.588884353637695, + "learning_rate": 4.3260380410711186e-06, + "loss": 9.471607208251953, + "loss_alignment": 0.43505859375, + "loss_alignment_w": 0.217529296875, + "loss_sft": 0.37411048635840416, + "loss_total_v6": 0.5919754803180695, + "step": 2446 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5643531531531532, + "grad_norm": 20.299097061157227, + "learning_rate": 4.32227503221966e-06, + "loss": 8.877935409545898, + "loss_alignment": 0.39990234375, + "loss_alignment_w": 0.199951171875, + "loss_sft": 0.35528600215911865, + "loss_total_v6": 0.5548709742724895, + "step": 2447 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5645837837837838, + "grad_norm": 20.120901107788086, + "learning_rate": 4.318512414382059e-06, + "loss": 9.354227066040039, + "loss_alignment": 0.404541015625, + "loss_alignment_w": 0.2022705078125, + "loss_sft": 0.38242970034480095, + "loss_total_v6": 0.5846391841769218, + "step": 2448 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5648144144144144, + "grad_norm": 22.427289962768555, + "learning_rate": 4.314750189729163e-06, + "loss": 9.623262405395508, + "loss_alignment": 0.419189453125, + "loss_alignment_w": 0.2095947265625, + "loss_sft": 0.39147770777344704, + "loss_total_v6": 0.6014539077877998, + "step": 2449 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.565045045045045, + "grad_norm": 19.80255126953125, + "learning_rate": 4.310988360431585e-06, + "loss": 9.652153015136719, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.3931918479502201, + "loss_total_v6": 0.6032596081495285, + "step": 2450 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5652756756756757, + "grad_norm": 24.066051483154297, + "learning_rate": 4.30722692865972e-06, + "loss": 10.027456283569336, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.4173044227063656, + "loss_total_v6": 0.6267160475254059, + "step": 2451 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5655063063063063, + "grad_norm": 23.467742919921875, + "learning_rate": 4.3034658965837255e-06, + "loss": 9.790811538696289, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.40352122858166695, + "loss_total_v6": 0.611925758421421, + "step": 2452 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5657369369369369, + "grad_norm": 22.63536834716797, + "learning_rate": 4.299705266373531e-06, + "loss": 9.905508041381836, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.40833984315395355, + "loss_total_v6": 0.6190942525863647, + "step": 2453 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5659675675675676, + "grad_norm": 21.807191848754883, + "learning_rate": 4.29594504019884e-06, + "loss": 9.38644790649414, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.3724501356482506, + "loss_total_v6": 0.5866529792547226, + "step": 2454 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5661981981981982, + "grad_norm": 23.21916961669922, + "learning_rate": 4.29218522022911e-06, + "loss": 9.32899284362793, + "loss_alignment": 0.414306640625, + "loss_alignment_w": 0.2071533203125, + "loss_sft": 0.3753594681620598, + "loss_total_v6": 0.5830620750784874, + "step": 2455 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5664288288288288, + "grad_norm": 21.389543533325195, + "learning_rate": 4.2884258086335755e-06, + "loss": 9.42220687866211, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.37853026390075684, + "loss_total_v6": 0.588887944817543, + "step": 2456 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5666594594594595, + "grad_norm": 21.105201721191406, + "learning_rate": 4.28466680758123e-06, + "loss": 9.689258575439453, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.3963043577969074, + "loss_total_v6": 0.6055786684155464, + "step": 2457 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5668900900900901, + "grad_norm": 21.229869842529297, + "learning_rate": 4.280908219240832e-06, + "loss": 9.8881254196167, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.4109765738248825, + "loss_total_v6": 0.6180078312754631, + "step": 2458 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5671207207207207, + "grad_norm": 20.522384643554688, + "learning_rate": 4.2771500457809045e-06, + "loss": 9.620870590209961, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.38775767758488655, + "loss_total_v6": 0.6013044193387032, + "step": 2459 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5673513513513514, + "grad_norm": 21.08976936340332, + "learning_rate": 4.273392289369723e-06, + "loss": 9.724517822265625, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.3978366293013096, + "loss_total_v6": 0.6077823340892792, + "step": 2460 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.567581981981982, + "grad_norm": 20.479103088378906, + "learning_rate": 4.26963495217533e-06, + "loss": 10.025596618652344, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.410199660807848, + "loss_total_v6": 0.6265997961163521, + "step": 2461 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5678126126126126, + "grad_norm": 20.7078800201416, + "learning_rate": 4.265878036365524e-06, + "loss": 9.679243087768555, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.394717026501894, + "loss_total_v6": 0.6049526482820511, + "step": 2462 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5680432432432433, + "grad_norm": 21.518701553344727, + "learning_rate": 4.262121544107861e-06, + "loss": 9.730280876159668, + "loss_alignment": 0.416259765625, + "loss_alignment_w": 0.2081298828125, + "loss_sft": 0.39976852759718895, + "loss_total_v6": 0.6081425547599792, + "step": 2463 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5682738738738738, + "grad_norm": 21.91661834716797, + "learning_rate": 4.258365477569648e-06, + "loss": 9.474014282226562, + "loss_alignment": 0.43359375, + "loss_alignment_w": 0.216796875, + "loss_sft": 0.37542053312063217, + "loss_total_v6": 0.5921258702874184, + "step": 2464 + }, + { + "action_cond_aligned": 0.9140625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5685045045045045, + "grad_norm": 20.857946395874023, + "learning_rate": 4.254609838917951e-06, + "loss": 9.545547485351562, + "loss_alignment": 0.406494140625, + "loss_alignment_w": 0.2032470703125, + "loss_sft": 0.3934563994407654, + "loss_total_v6": 0.5965966880321503, + "step": 2465 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5687351351351352, + "grad_norm": 19.852113723754883, + "learning_rate": 4.250854630319593e-06, + "loss": 10.03858470916748, + "loss_alignment": 0.406982421875, + "loss_alignment_w": 0.2034912109375, + "loss_sft": 0.42376771196722984, + "loss_total_v6": 0.6274115219712257, + "step": 2466 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5689657657657657, + "grad_norm": 22.578262329101562, + "learning_rate": 4.247099853941138e-06, + "loss": 10.01456069946289, + "loss_alignment": 0.421142578125, + "loss_alignment_w": 0.2105712890625, + "loss_sft": 0.41570502892136574, + "loss_total_v6": 0.6259100884199142, + "step": 2467 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5691963963963964, + "grad_norm": 23.016483306884766, + "learning_rate": 4.24334551194891e-06, + "loss": 9.86847972869873, + "loss_alignment": 0.425048828125, + "loss_alignment_w": 0.2125244140625, + "loss_sft": 0.4048353359103203, + "loss_total_v6": 0.6167799606919289, + "step": 2468 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5694270270270271, + "grad_norm": 20.603120803833008, + "learning_rate": 4.239591606508977e-06, + "loss": 10.017610549926758, + "loss_alignment": 0.41650390625, + "loss_alignment_w": 0.208251953125, + "loss_sft": 0.4183064289391041, + "loss_total_v6": 0.6261006221175194, + "step": 2469 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5696576576576576, + "grad_norm": 21.235307693481445, + "learning_rate": 4.235838139787161e-06, + "loss": 9.625537872314453, + "loss_alignment": 0.43701171875, + "loss_alignment_w": 0.218505859375, + "loss_sft": 0.3833801820874214, + "loss_total_v6": 0.6015961319208145, + "step": 2470 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5698882882882883, + "grad_norm": 20.18710708618164, + "learning_rate": 4.232085113949022e-06, + "loss": 9.095258712768555, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.35753142833709717, + "loss_total_v6": 0.5684536546468735, + "step": 2471 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.570118918918919, + "grad_norm": 23.727745056152344, + "learning_rate": 4.228332531159872e-06, + "loss": 9.918122291564941, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.40484053269028664, + "loss_total_v6": 0.6198826357722282, + "step": 2472 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5703495495495495, + "grad_norm": 31.955196380615234, + "learning_rate": 4.224580393584768e-06, + "loss": 9.91102409362793, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.4064568616449833, + "loss_total_v6": 0.6194390133023262, + "step": 2473 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5705801801801802, + "grad_norm": 22.914438247680664, + "learning_rate": 4.220828703388507e-06, + "loss": 9.495482444763184, + "loss_alignment": 0.407470703125, + "loss_alignment_w": 0.2037353515625, + "loss_sft": 0.38991542160511017, + "loss_total_v6": 0.593467652797699, + "step": 2474 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5708108108108109, + "grad_norm": 21.858654022216797, + "learning_rate": 4.217077462735632e-06, + "loss": 9.619169235229492, + "loss_alignment": 0.429443359375, + "loss_alignment_w": 0.2147216796875, + "loss_sft": 0.38632387295365334, + "loss_total_v6": 0.6011981293559074, + "step": 2475 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5710414414414414, + "grad_norm": 22.703001022338867, + "learning_rate": 4.213326673790418e-06, + "loss": 9.439275741577148, + "loss_alignment": 0.425537109375, + "loss_alignment_w": 0.2127685546875, + "loss_sft": 0.3777965381741524, + "loss_total_v6": 0.5899547263979912, + "step": 2476 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5712720720720721, + "grad_norm": 21.750324249267578, + "learning_rate": 4.2095763387168895e-06, + "loss": 9.451367378234863, + "loss_alignment": 0.39990234375, + "loss_alignment_w": 0.199951171875, + "loss_sft": 0.39045412093400955, + "loss_total_v6": 0.5907104685902596, + "step": 2477 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5715027027027026, + "grad_norm": 21.135395050048828, + "learning_rate": 4.205826459678802e-06, + "loss": 9.610574722290039, + "loss_alignment": 0.426025390625, + "loss_alignment_w": 0.2130126953125, + "loss_sft": 0.38778558373451233, + "loss_total_v6": 0.6006609350442886, + "step": 2478 + }, + { + "action_cond_aligned": 1.0, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5717333333333333, + "grad_norm": 20.45662498474121, + "learning_rate": 4.202077038839652e-06, + "loss": 10.028362274169922, + "loss_alignment": 0.441650390625, + "loss_alignment_w": 0.2208251953125, + "loss_sft": 0.4058406315743923, + "loss_total_v6": 0.6267726272344589, + "step": 2479 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.571963963963964, + "grad_norm": 21.000362396240234, + "learning_rate": 4.198328078362672e-06, + "loss": 9.622095108032227, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.38981784135103226, + "loss_total_v6": 0.6013809591531754, + "step": 2480 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5721945945945945, + "grad_norm": 21.12683868408203, + "learning_rate": 4.194579580410825e-06, + "loss": 9.357366561889648, + "loss_alignment": 0.418212890625, + "loss_alignment_w": 0.2091064453125, + "loss_sft": 0.3755611814558506, + "loss_total_v6": 0.5848354548215866, + "step": 2481 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5724252252252252, + "grad_norm": 21.886699676513672, + "learning_rate": 4.190831547146809e-06, + "loss": 9.64472770690918, + "loss_alignment": 0.430908203125, + "loss_alignment_w": 0.2154541015625, + "loss_sft": 0.38734137266874313, + "loss_total_v6": 0.6027954667806625, + "step": 2482 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5726558558558559, + "grad_norm": 21.27764892578125, + "learning_rate": 4.1870839807330545e-06, + "loss": 9.415841102600098, + "loss_alignment": 0.42138671875, + "loss_alignment_w": 0.210693359375, + "loss_sft": 0.3777814619243145, + "loss_total_v6": 0.5884900540113449, + "step": 2483 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5728864864864864, + "grad_norm": 20.677291870117188, + "learning_rate": 4.183336883331723e-06, + "loss": 9.407463073730469, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.3845820277929306, + "loss_total_v6": 0.5879664197564125, + "step": 2484 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5731171171171171, + "grad_norm": 24.308996200561523, + "learning_rate": 4.179590257104708e-06, + "loss": 9.893162727355957, + "loss_alignment": 0.43603515625, + "loss_alignment_w": 0.218017578125, + "loss_sft": 0.400350883603096, + "loss_total_v6": 0.6183226928114891, + "step": 2485 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5733477477477478, + "grad_norm": 20.76995277404785, + "learning_rate": 4.175844104213621e-06, + "loss": 9.57815170288086, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.39245767146348953, + "loss_total_v6": 0.5986344590783119, + "step": 2486 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5735783783783783, + "grad_norm": 21.20578384399414, + "learning_rate": 4.172098426819812e-06, + "loss": 9.390216827392578, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.3754780516028404, + "loss_total_v6": 0.5868885815143585, + "step": 2487 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.573809009009009, + "grad_norm": 21.928977966308594, + "learning_rate": 4.1683532270843505e-06, + "loss": 10.061319351196289, + "loss_alignment": 0.426513671875, + "loss_alignment_w": 0.2132568359375, + "loss_sft": 0.41531625390052795, + "loss_total_v6": 0.6288324743509293, + "step": 2488 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5740396396396397, + "grad_norm": 22.018274307250977, + "learning_rate": 4.164608507168034e-06, + "loss": 9.420904159545898, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.38024933636188507, + "loss_total_v6": 0.5888064801692963, + "step": 2489 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5742702702702702, + "grad_norm": 20.067211151123047, + "learning_rate": 4.1608642692313765e-06, + "loss": 9.836179733276367, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.40559376776218414, + "loss_total_v6": 0.6147612258791924, + "step": 2490 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5745009009009009, + "grad_norm": 20.928632736206055, + "learning_rate": 4.1571205154346215e-06, + "loss": 9.515954971313477, + "loss_alignment": 0.4169921875, + "loss_alignment_w": 0.20849609375, + "loss_sft": 0.386464711278677, + "loss_total_v6": 0.5947471857070923, + "step": 2491 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5747315315315316, + "grad_norm": 24.60179901123047, + "learning_rate": 4.1533772479377325e-06, + "loss": 9.440603256225586, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.3790391981601715, + "loss_total_v6": 0.5900377407670021, + "step": 2492 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5749621621621621, + "grad_norm": 19.835765838623047, + "learning_rate": 4.149634468900386e-06, + "loss": 9.588407516479492, + "loss_alignment": 0.412841796875, + "loss_alignment_w": 0.2064208984375, + "loss_sft": 0.39308345317840576, + "loss_total_v6": 0.5992754474282265, + "step": 2493 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5751927927927928, + "grad_norm": 19.18309783935547, + "learning_rate": 4.145892180481983e-06, + "loss": 9.381118774414062, + "loss_alignment": 0.41064453125, + "loss_alignment_w": 0.205322265625, + "loss_sft": 0.380845058709383, + "loss_total_v6": 0.5863199234008789, + "step": 2494 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5754234234234235, + "grad_norm": 24.828096389770508, + "learning_rate": 4.1421503848416385e-06, + "loss": 9.177932739257812, + "loss_alignment": 0.41552734375, + "loss_alignment_w": 0.207763671875, + "loss_sft": 0.3652925118803978, + "loss_total_v6": 0.5736207813024521, + "step": 2495 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.575654054054054, + "grad_norm": 23.218351364135742, + "learning_rate": 4.138409084138185e-06, + "loss": 9.141708374023438, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.36398985981941223, + "loss_total_v6": 0.5713568106293678, + "step": 2496 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5758846846846847, + "grad_norm": 40.535491943359375, + "learning_rate": 4.134668280530166e-06, + "loss": 9.718606948852539, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.39391201362013817, + "loss_total_v6": 0.6074129790067673, + "step": 2497 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5761153153153153, + "grad_norm": 20.3319034576416, + "learning_rate": 4.130927976175842e-06, + "loss": 9.187423706054688, + "loss_alignment": 0.4150390625, + "loss_alignment_w": 0.20751953125, + "loss_sft": 0.36638928577303886, + "loss_total_v6": 0.5742140039801598, + "step": 2498 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5763459459459459, + "grad_norm": 20.068134307861328, + "learning_rate": 4.127188173233185e-06, + "loss": 10.037775039672852, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4159504510462284, + "loss_total_v6": 0.6273609548807144, + "step": 2499 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5765765765765766, + "grad_norm": 21.769128799438477, + "learning_rate": 4.123448873859875e-06, + "loss": 9.480157852172852, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.3778797574341297, + "loss_total_v6": 0.5925098732113838, + "step": 2500 + }, + { + "epoch": 0.5765765765765766, + "eval_loss": 0.6532387733459473, + "eval_loss_alignment": 0.40283829126602566, + "eval_loss_alignment_w": 0.20141914563301283, + "eval_loss_sft": 0.451819668428447, + "eval_loss_total_v6": 0.6532388138465393, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.9113, + "eval_samples_per_second": 22.781, + "eval_steps_per_second": 2.857, + "step": 2500 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5768072072072072, + "grad_norm": 22.055099487304688, + "learning_rate": 4.119710080213306e-06, + "loss": 9.536704063415527, + "loss_alignment": 0.422119140625, + "loss_alignment_w": 0.2110595703125, + "loss_sft": 0.38460297137498856, + "loss_total_v6": 0.5960439890623093, + "step": 2501 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5770378378378378, + "grad_norm": 21.11907386779785, + "learning_rate": 4.1159717944505724e-06, + "loss": 9.514619827270508, + "loss_alignment": 0.43212890625, + "loss_alignment_w": 0.216064453125, + "loss_sft": 0.37858404591679573, + "loss_total_v6": 0.5946637615561485, + "step": 2502 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5772684684684685, + "grad_norm": 20.554481506347656, + "learning_rate": 4.1122340187284845e-06, + "loss": 9.729761123657227, + "loss_alignment": 0.41748046875, + "loss_alignment_w": 0.208740234375, + "loss_sft": 0.39921725913882256, + "loss_total_v6": 0.6081100627779961, + "step": 2503 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5774990990990991, + "grad_norm": 22.637161254882812, + "learning_rate": 4.108496755203553e-06, + "loss": 10.090980529785156, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.42174771055579185, + "loss_total_v6": 0.6306862905621529, + "step": 2504 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5777297297297297, + "grad_norm": 20.1954288482666, + "learning_rate": 4.104760006031995e-06, + "loss": 10.084967613220215, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.4212192967534065, + "loss_total_v6": 0.630310483276844, + "step": 2505 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5779603603603604, + "grad_norm": 23.46910285949707, + "learning_rate": 4.101023773369732e-06, + "loss": 9.718135833740234, + "loss_alignment": 0.423828125, + "loss_alignment_w": 0.2119140625, + "loss_sft": 0.39600346982479095, + "loss_total_v6": 0.6073834449052811, + "step": 2506 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.578190990990991, + "grad_norm": 21.41126823425293, + "learning_rate": 4.097288059372383e-06, + "loss": 9.605499267578125, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.3912067972123623, + "loss_total_v6": 0.600343756377697, + "step": 2507 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5784216216216216, + "grad_norm": 21.19098472595215, + "learning_rate": 4.0935528661952716e-06, + "loss": 10.050033569335938, + "loss_alignment": 0.427490234375, + "loss_alignment_w": 0.2137451171875, + "loss_sft": 0.4151144027709961, + "loss_total_v6": 0.6281270906329155, + "step": 2508 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5786522522522523, + "grad_norm": 22.972578048706055, + "learning_rate": 4.08981819599342e-06, + "loss": 9.712427139282227, + "loss_alignment": 0.40185546875, + "loss_alignment_w": 0.200927734375, + "loss_sft": 0.40644993260502815, + "loss_total_v6": 0.6070267260074615, + "step": 2509 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5788828828828829, + "grad_norm": 20.828125, + "learning_rate": 4.08608405092155e-06, + "loss": 10.339361190795898, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.43101535364985466, + "loss_total_v6": 0.6462100520730019, + "step": 2510 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5791135135135135, + "grad_norm": 20.468965530395508, + "learning_rate": 4.082350433134076e-06, + "loss": 9.703022003173828, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.39588290825486183, + "loss_total_v6": 0.6064389050006866, + "step": 2511 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5793441441441441, + "grad_norm": 26.15379524230957, + "learning_rate": 4.078617344785113e-06, + "loss": 9.059564590454102, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.35241665691137314, + "loss_total_v6": 0.5662228018045425, + "step": 2512 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5795747747747748, + "grad_norm": 20.487812042236328, + "learning_rate": 4.074884788028469e-06, + "loss": 9.641468048095703, + "loss_alignment": 0.406005859375, + "loss_alignment_w": 0.2030029296875, + "loss_sft": 0.39966512098908424, + "loss_total_v6": 0.6025918051600456, + "step": 2513 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5798054054054054, + "grad_norm": 21.488191604614258, + "learning_rate": 4.0711527650176434e-06, + "loss": 9.891511917114258, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.40688531473279, + "loss_total_v6": 0.6182195320725441, + "step": 2514 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.580036036036036, + "grad_norm": 23.037681579589844, + "learning_rate": 4.067421277905833e-06, + "loss": 10.477182388305664, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.439491830766201, + "loss_total_v6": 0.6548238769173622, + "step": 2515 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5802666666666667, + "grad_norm": 21.265844345092773, + "learning_rate": 4.063690328845917e-06, + "loss": 9.639501571655273, + "loss_alignment": 0.41796875, + "loss_alignment_w": 0.208984375, + "loss_sft": 0.3932402655482292, + "loss_total_v6": 0.6024687960743904, + "step": 2516 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5804972972972973, + "grad_norm": 44.0908317565918, + "learning_rate": 4.059959919990471e-06, + "loss": 8.991754531860352, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.3469730317592621, + "loss_total_v6": 0.561984620988369, + "step": 2517 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5807279279279279, + "grad_norm": 19.896482467651367, + "learning_rate": 4.056230053491759e-06, + "loss": 9.533088684082031, + "loss_alignment": 0.42724609375, + "loss_alignment_w": 0.213623046875, + "loss_sft": 0.38248487189412117, + "loss_total_v6": 0.5958179980516434, + "step": 2518 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5809585585585586, + "grad_norm": 21.739891052246094, + "learning_rate": 4.052500731501726e-06, + "loss": 9.349410057067871, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.37907688319683075, + "loss_total_v6": 0.5843381434679031, + "step": 2519 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5811891891891892, + "grad_norm": 21.038105010986328, + "learning_rate": 4.04877195617201e-06, + "loss": 9.610509872436523, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.3918403722345829, + "loss_total_v6": 0.6006568968296051, + "step": 2520 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5814198198198198, + "grad_norm": 20.683788299560547, + "learning_rate": 4.045043729653927e-06, + "loss": 9.486359596252441, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.3842335417866707, + "loss_total_v6": 0.5928974896669388, + "step": 2521 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5816504504504505, + "grad_norm": 20.78081512451172, + "learning_rate": 4.041316054098485e-06, + "loss": 9.681259155273438, + "loss_alignment": 0.41943359375, + "loss_alignment_w": 0.209716796875, + "loss_sft": 0.39514825120568275, + "loss_total_v6": 0.6050786897540092, + "step": 2522 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.581881081081081, + "grad_norm": 19.18006134033203, + "learning_rate": 4.037588931656362e-06, + "loss": 9.750490188598633, + "loss_alignment": 0.423583984375, + "loss_alignment_w": 0.2117919921875, + "loss_sft": 0.3974915184080601, + "loss_total_v6": 0.609405592083931, + "step": 2523 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5821117117117117, + "grad_norm": 21.878564834594727, + "learning_rate": 4.033862364477927e-06, + "loss": 9.650452613830566, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.3973885253071785, + "loss_total_v6": 0.603153295814991, + "step": 2524 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5823423423423424, + "grad_norm": 21.79802703857422, + "learning_rate": 4.030136354713227e-06, + "loss": 9.695841789245605, + "loss_alignment": 0.403076171875, + "loss_alignment_w": 0.2015380859375, + "loss_sft": 0.40434518083930016, + "loss_total_v6": 0.6059901118278503, + "step": 2525 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.582572972972973, + "grad_norm": 20.360191345214844, + "learning_rate": 4.026410904511984e-06, + "loss": 9.732170104980469, + "loss_alignment": 0.410888671875, + "loss_alignment_w": 0.2054443359375, + "loss_sft": 0.4022059291601181, + "loss_total_v6": 0.6082606241106987, + "step": 2526 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5828036036036036, + "grad_norm": 19.691936492919922, + "learning_rate": 4.0226860160236e-06, + "loss": 9.247200012207031, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.36301464214921, + "loss_total_v6": 0.5779499597847462, + "step": 2527 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5830342342342343, + "grad_norm": 22.542869567871094, + "learning_rate": 4.018961691397148e-06, + "loss": 10.0828857421875, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.41521453857421875, + "loss_total_v6": 0.6301803439855576, + "step": 2528 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5832648648648648, + "grad_norm": 19.777414321899414, + "learning_rate": 4.015237932781385e-06, + "loss": 9.652666091918945, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.3902178071439266, + "loss_total_v6": 0.6032915636897087, + "step": 2529 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5834954954954955, + "grad_norm": 24.118255615234375, + "learning_rate": 4.011514742324732e-06, + "loss": 9.780244827270508, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4010144881904125, + "loss_total_v6": 0.6112653315067291, + "step": 2530 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5837261261261262, + "grad_norm": 21.342432022094727, + "learning_rate": 4.007792122175287e-06, + "loss": 9.800756454467773, + "loss_alignment": 0.42626953125, + "loss_alignment_w": 0.213134765625, + "loss_sft": 0.39927516505122185, + "loss_total_v6": 0.612547256052494, + "step": 2531 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5839567567567567, + "grad_norm": 20.711021423339844, + "learning_rate": 4.004070074480821e-06, + "loss": 9.02795696258545, + "loss_alignment": 0.4296875, + "loss_alignment_w": 0.21484375, + "loss_sft": 0.3491288982331753, + "loss_total_v6": 0.56424730271101, + "step": 2532 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5841873873873874, + "grad_norm": 22.456809997558594, + "learning_rate": 4.000348601388767e-06, + "loss": 9.70878791809082, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.3977232836186886, + "loss_total_v6": 0.6067992299795151, + "step": 2533 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5844180180180181, + "grad_norm": 19.694875717163086, + "learning_rate": 3.9966277050462345e-06, + "loss": 9.49199390411377, + "loss_alignment": 0.418701171875, + "loss_alignment_w": 0.2093505859375, + "loss_sft": 0.3842957615852356, + "loss_total_v6": 0.5932496264576912, + "step": 2534 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5846486486486486, + "grad_norm": 20.010889053344727, + "learning_rate": 3.992907387599996e-06, + "loss": 10.0345458984375, + "loss_alignment": 0.419921875, + "loss_alignment_w": 0.2099609375, + "loss_sft": 0.41725919023156166, + "loss_total_v6": 0.6271591037511826, + "step": 2535 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5848792792792793, + "grad_norm": 25.483240127563477, + "learning_rate": 3.989187651196493e-06, + "loss": 9.336325645446777, + "loss_alignment": 0.410400390625, + "loss_alignment_w": 0.2052001953125, + "loss_sft": 0.3777555897831917, + "loss_total_v6": 0.5835203677415848, + "step": 2536 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.58510990990991, + "grad_norm": 20.321958541870117, + "learning_rate": 3.985468497981826e-06, + "loss": 9.257769584655762, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.36622354388237, + "loss_total_v6": 0.5786106213927269, + "step": 2537 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5853405405405405, + "grad_norm": 21.042795181274414, + "learning_rate": 3.981749930101764e-06, + "loss": 9.673599243164062, + "loss_alignment": 0.415771484375, + "loss_alignment_w": 0.2078857421875, + "loss_sft": 0.3964395932853222, + "loss_total_v6": 0.6045999899506569, + "step": 2538 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5855711711711712, + "grad_norm": 22.293529510498047, + "learning_rate": 3.978031949701742e-06, + "loss": 9.193490982055664, + "loss_alignment": 0.41357421875, + "loss_alignment_w": 0.206787109375, + "loss_sft": 0.3670278936624527, + "loss_total_v6": 0.5745931938290596, + "step": 2539 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5858018018018017, + "grad_norm": 24.565372467041016, + "learning_rate": 3.974314558926844e-06, + "loss": 9.892370223999023, + "loss_alignment": 0.4228515625, + "loss_alignment_w": 0.21142578125, + "loss_sft": 0.4066794849932194, + "loss_total_v6": 0.618273101747036, + "step": 2540 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5860324324324324, + "grad_norm": 21.926483154296875, + "learning_rate": 3.970597759921826e-06, + "loss": 9.734636306762695, + "loss_alignment": 0.42333984375, + "loss_alignment_w": 0.211669921875, + "loss_sft": 0.39730945229530334, + "loss_total_v6": 0.6084147915244102, + "step": 2541 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5862630630630631, + "grad_norm": 20.774978637695312, + "learning_rate": 3.966881554831094e-06, + "loss": 9.788755416870117, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4000967666506767, + "loss_total_v6": 0.6117972135543823, + "step": 2542 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5864936936936936, + "grad_norm": 20.699739456176758, + "learning_rate": 3.963165945798718e-06, + "loss": 9.879203796386719, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.4050937108695507, + "loss_total_v6": 0.6174502670764923, + "step": 2543 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5867243243243243, + "grad_norm": 21.924034118652344, + "learning_rate": 3.959450934968422e-06, + "loss": 9.370427131652832, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.3752482682466507, + "loss_total_v6": 0.5856517106294632, + "step": 2544 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.586954954954955, + "grad_norm": 21.384233474731445, + "learning_rate": 3.955736524483579e-06, + "loss": 9.803102493286133, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.4021989367902279, + "loss_total_v6": 0.6126939356327057, + "step": 2545 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5871855855855855, + "grad_norm": 21.941875457763672, + "learning_rate": 3.952022716487225e-06, + "loss": 9.095600128173828, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.36225250363349915, + "loss_total_v6": 0.5684750452637672, + "step": 2546 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5874162162162162, + "grad_norm": 21.44447898864746, + "learning_rate": 3.9483095131220406e-06, + "loss": 10.14803695678711, + "loss_alignment": 0.427978515625, + "loss_alignment_w": 0.2139892578125, + "loss_sft": 0.41977478936314583, + "loss_total_v6": 0.6342523172497749, + "step": 2547 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5876468468468469, + "grad_norm": 21.70660400390625, + "learning_rate": 3.944596916530365e-06, + "loss": 9.746452331542969, + "loss_alignment": 0.42041015625, + "loss_alignment_w": 0.210205078125, + "loss_sft": 0.39917705208063126, + "loss_total_v6": 0.6091532409191132, + "step": 2548 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5878774774774774, + "grad_norm": 20.5526180267334, + "learning_rate": 3.940884928854179e-06, + "loss": 9.893842697143555, + "loss_alignment": 0.420654296875, + "loss_alignment_w": 0.2103271484375, + "loss_sft": 0.4082821384072304, + "loss_total_v6": 0.6183651983737946, + "step": 2549 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5881081081081081, + "grad_norm": 20.930278778076172, + "learning_rate": 3.937173552235117e-06, + "loss": 10.161270141601562, + "loss_alignment": 0.411865234375, + "loss_alignment_w": 0.2059326171875, + "loss_sft": 0.42885685712099075, + "loss_total_v6": 0.6350793913006783, + "step": 2550 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5883387387387388, + "grad_norm": 21.599441528320312, + "learning_rate": 3.933462788814463e-06, + "loss": 9.88577651977539, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.4043143168091774, + "loss_total_v6": 0.6178610473871231, + "step": 2551 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5885693693693693, + "grad_norm": 20.546052932739258, + "learning_rate": 3.929752640733141e-06, + "loss": 9.440778732299805, + "loss_alignment": 0.42529296875, + "loss_alignment_w": 0.212646484375, + "loss_sft": 0.3774326629936695, + "loss_total_v6": 0.590048648416996, + "step": 2552 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5888, + "grad_norm": 22.046567916870117, + "learning_rate": 3.926043110131727e-06, + "loss": 9.684877395629883, + "loss_alignment": 0.4208984375, + "loss_alignment_w": 0.21044921875, + "loss_sft": 0.39494715444743633, + "loss_total_v6": 0.6053048111498356, + "step": 2553 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5890306306306307, + "grad_norm": 20.88192367553711, + "learning_rate": 3.922334199150433e-06, + "loss": 10.076257705688477, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.4130607806146145, + "loss_total_v6": 0.6297661140561104, + "step": 2554 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5892612612612612, + "grad_norm": 22.716552734375, + "learning_rate": 3.9186259099291205e-06, + "loss": 9.593551635742188, + "loss_alignment": 0.408203125, + "loss_alignment_w": 0.2041015625, + "loss_sft": 0.3951749615371227, + "loss_total_v6": 0.5995969623327255, + "step": 2555 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5894918918918919, + "grad_norm": 22.5751953125, + "learning_rate": 3.914918244607287e-06, + "loss": 9.945401191711426, + "loss_alignment": 0.433349609375, + "loss_alignment_w": 0.2166748046875, + "loss_sft": 0.40508055314421654, + "loss_total_v6": 0.6215875446796417, + "step": 2556 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5897225225225226, + "grad_norm": 21.15581512451172, + "learning_rate": 3.911211205324074e-06, + "loss": 9.765538215637207, + "loss_alignment": 0.4140625, + "loss_alignment_w": 0.20703125, + "loss_sft": 0.40348274260759354, + "loss_total_v6": 0.6103461347520351, + "step": 2557 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5899531531531531, + "grad_norm": 20.09288215637207, + "learning_rate": 3.907504794218262e-06, + "loss": 10.048370361328125, + "loss_alignment": 0.423095703125, + "loss_alignment_w": 0.2115478515625, + "loss_sft": 0.4167194627225399, + "loss_total_v6": 0.6280231848359108, + "step": 2558 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5901837837837838, + "grad_norm": 19.493736267089844, + "learning_rate": 3.903799013428264e-06, + "loss": 8.971023559570312, + "loss_alignment": 0.405517578125, + "loss_alignment_w": 0.2027587890625, + "loss_sft": 0.3581896014511585, + "loss_total_v6": 0.5606889948248863, + "step": 2559 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5904144144144144, + "grad_norm": 20.49209976196289, + "learning_rate": 3.900093865092134e-06, + "loss": 9.628236770629883, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.3899270109832287, + "loss_total_v6": 0.6017647609114647, + "step": 2560 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.590645045045045, + "grad_norm": 19.688159942626953, + "learning_rate": 3.896389351347559e-06, + "loss": 8.98332691192627, + "loss_alignment": 0.416015625, + "loss_alignment_w": 0.2080078125, + "loss_sft": 0.35348062589764595, + "loss_total_v6": 0.5614579431712627, + "step": 2561 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5908756756756757, + "grad_norm": 22.6927547454834, + "learning_rate": 3.892685474331863e-06, + "loss": 9.621475219726562, + "loss_alignment": 0.432861328125, + "loss_alignment_w": 0.2164306640625, + "loss_sft": 0.38468269258737564, + "loss_total_v6": 0.6013422459363937, + "step": 2562 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5911063063063063, + "grad_norm": 22.140085220336914, + "learning_rate": 3.888982236181997e-06, + "loss": 10.172416687011719, + "loss_alignment": 0.43017578125, + "loss_alignment_w": 0.215087890625, + "loss_sft": 0.4199710115790367, + "loss_total_v6": 0.6357760727405548, + "step": 2563 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5913369369369369, + "grad_norm": 20.368148803710938, + "learning_rate": 3.8852796390345465e-06, + "loss": 9.328855514526367, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.37341298162937164, + "loss_total_v6": 0.5830534920096397, + "step": 2564 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5915675675675676, + "grad_norm": 21.2130126953125, + "learning_rate": 3.8815776850257325e-06, + "loss": 9.623889923095703, + "loss_alignment": 0.420166015625, + "loss_alignment_w": 0.2100830078125, + "loss_sft": 0.39141010865569115, + "loss_total_v6": 0.6014930829405785, + "step": 2565 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5917981981981982, + "grad_norm": 19.877721786499023, + "learning_rate": 3.877876376291393e-06, + "loss": 9.533224105834961, + "loss_alignment": 0.40625, + "loss_alignment_w": 0.203125, + "loss_sft": 0.39328138530254364, + "loss_total_v6": 0.5958265550434589, + "step": 2566 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5920288288288288, + "grad_norm": 21.175535202026367, + "learning_rate": 3.874175714967007e-06, + "loss": 10.024969100952148, + "loss_alignment": 0.421630859375, + "loss_alignment_w": 0.2108154296875, + "loss_sft": 0.41595881432294846, + "loss_total_v6": 0.6265606060624123, + "step": 2567 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5922594594594595, + "grad_norm": 20.37212562561035, + "learning_rate": 3.870475703187667e-06, + "loss": 9.678919792175293, + "loss_alignment": 0.414794921875, + "loss_alignment_w": 0.2073974609375, + "loss_sft": 0.3974434733390808, + "loss_total_v6": 0.6049324721097946, + "step": 2568 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5924900900900901, + "grad_norm": 20.8127384185791, + "learning_rate": 3.866776343088103e-06, + "loss": 9.296768188476562, + "loss_alignment": 0.431884765625, + "loss_alignment_w": 0.2159423828125, + "loss_sft": 0.36505983769893646, + "loss_total_v6": 0.581047996878624, + "step": 2569 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5927207207207207, + "grad_norm": 21.361478805541992, + "learning_rate": 3.863077636802658e-06, + "loss": 9.273782730102539, + "loss_alignment": 0.412353515625, + "loss_alignment_w": 0.2061767578125, + "loss_sft": 0.3734346330165863, + "loss_total_v6": 0.5796114057302475, + "step": 2570 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5929513513513514, + "grad_norm": 20.497270584106445, + "learning_rate": 3.859379586465306e-06, + "loss": 9.870485305786133, + "loss_alignment": 0.416748046875, + "loss_alignment_w": 0.2083740234375, + "loss_sft": 0.4083024077117443, + "loss_total_v6": 0.6169053092598915, + "step": 2571 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.593181981981982, + "grad_norm": 20.824966430664062, + "learning_rate": 3.85568219420964e-06, + "loss": 9.893085479736328, + "loss_alignment": 0.437744140625, + "loss_alignment_w": 0.2188720703125, + "loss_sft": 0.3990948088467121, + "loss_total_v6": 0.6183178126811981, + "step": 2572 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5934126126126126, + "grad_norm": 20.269731521606445, + "learning_rate": 3.85198546216887e-06, + "loss": 10.216190338134766, + "loss_alignment": 0.429931640625, + "loss_alignment_w": 0.2149658203125, + "loss_sft": 0.42395808547735214, + "loss_total_v6": 0.6385119408369064, + "step": 2573 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5936432432432432, + "grad_norm": 21.221403121948242, + "learning_rate": 3.848289392475832e-06, + "loss": 9.840635299682617, + "loss_alignment": 0.41259765625, + "loss_alignment_w": 0.206298828125, + "loss_sft": 0.4086492843925953, + "loss_total_v6": 0.6150396689772606, + "step": 2574 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5938738738738739, + "grad_norm": 19.79322052001953, + "learning_rate": 3.844593987262973e-06, + "loss": 9.167398452758789, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.3613993003964424, + "loss_total_v6": 0.5729624032974243, + "step": 2575 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5941045045045045, + "grad_norm": 23.322978973388672, + "learning_rate": 3.840899248662358e-06, + "loss": 9.716062545776367, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.39363085106015205, + "loss_total_v6": 0.6072539314627647, + "step": 2576 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5943351351351351, + "grad_norm": 19.059370040893555, + "learning_rate": 3.837205178805674e-06, + "loss": 8.958454132080078, + "loss_alignment": 0.4111328125, + "loss_alignment_w": 0.20556640625, + "loss_sft": 0.3538791984319687, + "loss_total_v6": 0.5599033571779728, + "step": 2577 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5945657657657658, + "grad_norm": 20.293460845947266, + "learning_rate": 3.833511779824212e-06, + "loss": 9.282797813415527, + "loss_alignment": 0.400390625, + "loss_alignment_w": 0.2001953125, + "loss_sft": 0.3797353766858578, + "loss_total_v6": 0.5801748409867287, + "step": 2578 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5947963963963964, + "grad_norm": 20.970577239990234, + "learning_rate": 3.829819053848884e-06, + "loss": 9.500288009643555, + "loss_alignment": 0.424560546875, + "loss_alignment_w": 0.2122802734375, + "loss_sft": 0.38115204870700836, + "loss_total_v6": 0.5937680155038834, + "step": 2579 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.595027027027027, + "grad_norm": 22.675739288330078, + "learning_rate": 3.826127003010208e-06, + "loss": 9.66915512084961, + "loss_alignment": 0.407470703125, + "loss_alignment_w": 0.2037353515625, + "loss_sft": 0.4004190266132355, + "loss_total_v6": 0.6043222025036812, + "step": 2580 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5952576576576577, + "grad_norm": 19.30094337463379, + "learning_rate": 3.822435629438317e-06, + "loss": 9.33729362487793, + "loss_alignment": 0.40087890625, + "loss_alignment_w": 0.200439453125, + "loss_sft": 0.3833091966807842, + "loss_total_v6": 0.5835808292031288, + "step": 2581 + }, + { + "action_cond_aligned": 0.9375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5954882882882883, + "grad_norm": 21.27324676513672, + "learning_rate": 3.818744935262951e-06, + "loss": 9.243754386901855, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.3730837553739548, + "loss_total_v6": 0.5777346529066563, + "step": 2582 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5957189189189189, + "grad_norm": 36.29338836669922, + "learning_rate": 3.8150549226134575e-06, + "loss": 8.935371398925781, + "loss_alignment": 0.4072265625, + "loss_alignment_w": 0.20361328125, + "loss_sft": 0.35458799824118614, + "loss_total_v6": 0.5584606900811195, + "step": 2583 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5959495495495496, + "grad_norm": 21.262744903564453, + "learning_rate": 3.811365593618795e-06, + "loss": 9.70966625213623, + "loss_alignment": 0.408447265625, + "loss_alignment_w": 0.2042236328125, + "loss_sft": 0.4029357060790062, + "loss_total_v6": 0.6068541631102562, + "step": 2584 + }, + { + "action_cond_aligned": 0.9296875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5961801801801802, + "grad_norm": 20.532014846801758, + "learning_rate": 3.807676950407519e-06, + "loss": 9.576410293579102, + "loss_alignment": 0.409912109375, + "loss_alignment_w": 0.2049560546875, + "loss_sft": 0.39291341602802277, + "loss_total_v6": 0.5985256358981133, + "step": 2585 + }, + { + "action_cond_aligned": 0.9609375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5964108108108108, + "grad_norm": 21.54416847229004, + "learning_rate": 3.8039889951077984e-06, + "loss": 9.752752304077148, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.3943370617926121, + "loss_total_v6": 0.6095470264554024, + "step": 2586 + }, + { + "action_cond_aligned": 0.984375, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5966414414414415, + "grad_norm": 20.926748275756836, + "learning_rate": 3.8003017298474e-06, + "loss": 9.6107177734375, + "loss_alignment": 0.435302734375, + "loss_alignment_w": 0.2176513671875, + "loss_sft": 0.38285068795084953, + "loss_total_v6": 0.6006698757410049, + "step": 2587 + }, + { + "action_cond_aligned": 0.921875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.596872072072072, + "grad_norm": 21.03635025024414, + "learning_rate": 3.7966151567536956e-06, + "loss": 9.858530044555664, + "loss_alignment": 0.409423828125, + "loss_alignment_w": 0.2047119140625, + "loss_sft": 0.4112784154713154, + "loss_total_v6": 0.6161581911146641, + "step": 2588 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5971027027027027, + "grad_norm": 20.92828941345215, + "learning_rate": 3.7929292779536515e-06, + "loss": 9.29511833190918, + "loss_alignment": 0.4189453125, + "loss_alignment_w": 0.20947265625, + "loss_sft": 0.37131965532898903, + "loss_total_v6": 0.5809449106454849, + "step": 2589 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5973333333333334, + "grad_norm": 19.054166793823242, + "learning_rate": 3.789244095573842e-06, + "loss": 9.643628120422363, + "loss_alignment": 0.4287109375, + "loss_alignment_w": 0.21435546875, + "loss_sft": 0.388233944773674, + "loss_total_v6": 0.6027267351746559, + "step": 2590 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5975639639639639, + "grad_norm": 19.308137893676758, + "learning_rate": 3.785559611740436e-06, + "loss": 9.804849624633789, + "loss_alignment": 0.430419921875, + "loss_alignment_w": 0.2152099609375, + "loss_sft": 0.39777627773582935, + "loss_total_v6": 0.6128031276166439, + "step": 2591 + }, + { + "action_cond_aligned": 0.953125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5977945945945946, + "grad_norm": 18.66969108581543, + "learning_rate": 3.7818758285791958e-06, + "loss": 9.732172966003418, + "loss_alignment": 0.41845703125, + "loss_alignment_w": 0.209228515625, + "loss_sft": 0.3995510973036289, + "loss_total_v6": 0.608260802924633, + "step": 2592 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5980252252252253, + "grad_norm": 20.892745971679688, + "learning_rate": 3.778192748215484e-06, + "loss": 9.824775695800781, + "loss_alignment": 0.4267578125, + "loss_alignment_w": 0.21337890625, + "loss_sft": 0.40079158172011375, + "loss_total_v6": 0.6140484437346458, + "step": 2593 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5982558558558558, + "grad_norm": 20.732223510742188, + "learning_rate": 3.7745103727742554e-06, + "loss": 9.628271102905273, + "loss_alignment": 0.4326171875, + "loss_alignment_w": 0.21630859375, + "loss_sft": 0.38541263341903687, + "loss_total_v6": 0.6017669960856438, + "step": 2594 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5984864864864865, + "grad_norm": 22.67369270324707, + "learning_rate": 3.770828704380062e-06, + "loss": 9.664804458618164, + "loss_alignment": 0.41455078125, + "loss_alignment_w": 0.207275390625, + "loss_sft": 0.39683592692017555, + "loss_total_v6": 0.6040503159165382, + "step": 2595 + }, + { + "action_cond_aligned": 0.90625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5987171171171172, + "grad_norm": 20.969743728637695, + "learning_rate": 3.7671477451570393e-06, + "loss": 8.991236686706543, + "loss_alignment": 0.396728515625, + "loss_alignment_w": 0.1983642578125, + "loss_sft": 0.36354224383831024, + "loss_total_v6": 0.561952281743288, + "step": 2596 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5989477477477477, + "grad_norm": 33.07988739013672, + "learning_rate": 3.7634674972289227e-06, + "loss": 9.549285888671875, + "loss_alignment": 0.4091796875, + "loss_alignment_w": 0.20458984375, + "loss_sft": 0.392637275159359, + "loss_total_v6": 0.5968303754925728, + "step": 2597 + }, + { + "action_cond_aligned": 0.9453125, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5991783783783784, + "grad_norm": 19.72969627380371, + "learning_rate": 3.7597879627190337e-06, + "loss": 9.226034164428711, + "loss_alignment": 0.40869140625, + "loss_alignment_w": 0.204345703125, + "loss_sft": 0.37251032516360283, + "loss_total_v6": 0.5766271315515041, + "step": 2598 + }, + { + "action_cond_aligned": 0.96875, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5994090090090091, + "grad_norm": 35.335716247558594, + "learning_rate": 3.7561091437502808e-06, + "loss": 9.34033203125, + "loss_alignment": 0.417724609375, + "loss_alignment_w": 0.2088623046875, + "loss_sft": 0.3746032863855362, + "loss_total_v6": 0.5837707743048668, + "step": 2599 + }, + { + "action_cond_aligned": 0.9765625, + "action_cond_missing": 0.0, + "action_cond_skipped": 0.0, + "epoch": 0.5996396396396396, + "grad_norm": 19.494518280029297, + "learning_rate": 3.7524310424451637e-06, + "loss": 9.205700874328613, + "loss_alignment": 0.42822265625, + "loss_alignment_w": 0.214111328125, + "loss_sft": 0.36142805218696594, + "loss_total_v6": 0.5753562822937965, + "step": 2600 + }, + { + "epoch": 0.5996396396396396, + "eval_loss": 0.6510933041572571, + "eval_loss_alignment": 0.40263379740918803, + "eval_loss_alignment_w": 0.20131689870459402, + "eval_loss_sft": 0.4497763966966389, + "eval_loss_total_v6": 0.651093296149475, + "eval_loss_wm_recon": 0.0, + "eval_loss_wm_recon_w": 0.0, + "eval_runtime": 81.7416, + "eval_samples_per_second": 22.828, + "eval_steps_per_second": 2.863, + "step": 2600 + } + ], + "logging_steps": 1, + "max_steps": 4336, + "num_input_tokens_seen": 0, + "num_train_epochs": 1, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 2.8715774789462e+19, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/zero_to_fp32.py b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..5995d6e6f04e43b989587aa9022a3aef0c66d694 --- /dev/null +++ b/qwen-cua__runs__qwen35_4b_opencua_similar_instruction_add_wm_joint_v7_siglip2_delta_parallel-20260609-213735__v2-20260610-131619__checkpoint-2600__evalweights/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)