{ "use_ray": false, "ray_exp_name": null, "device_groups": null, "model": "/scratch/mt/new-structure/experiments/r.hany/grad/Datafilter", "model_type": "llama", "model_revision": null, "task_type": "causal_lm", "torch_dtype": "bfloat16", "attn_impl": "flash_attn", "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": "llama3_2", "system": null, "max_length": 1024, "truncation_strategy": "delete", "max_pixels": null, "agent_template": null, "norm_bbox": null, "use_chat_template": true, "padding_side": "right", "padding_free": true, "loss_scale": "default", "sequence_parallel_size": 1, "template_backend": "swift", "response_prefix": null, "enable_thinking": null, "add_non_thinking_prefix": true, "dataset": [ "/scratch/mt/new-structure/experiments/r.hany/grad/training_data.jsonl" ], "val_dataset": [ "/scratch/mt/new-structure/experiments/r.hany/grad/val_data.jsonl" ], "cached_dataset": [], "cached_val_dataset": [], "split_dataset_ratio": 0.0, "data_seed": 42, "dataset_num_proc": 90, "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, "remove_unused_columns": true, "disable_auto_column_mapping": false, "model_name": [ "swift-robot" ], "model_author": [ "swift" ], "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": null, "temperature": 0.9, "top_k": -1, "top_p": 1.0, "repetition_penalty": 1.0, "num_beams": 1, "stream": false, "stop_words": [], "logprobs": false, "top_logprobs": null, "structured_outputs_regex": null, "tuner_backend": "peft", "tuner_type": "full", "adapters": [], "external_plugins": [], "custom_register_path": [], "seed": 42, "model_kwargs": {}, "enable_npu_model_patch": true, "load_args": false, "load_data_args": false, "packing": true, "packing_length": 1024, "packing_num_proc": 1, "lazy_tokenize": false, "use_hf": false, "hub_token": null, "ddp_timeout": 18000000, "ddp_backend": "nccl", "ignore_args_error": false, "use_swift_lora": false, "freeze_llm": false, "freeze_vit": true, "freeze_aligner": true, "freeze_parameters": [], "freeze_parameters_regex": null, "freeze_parameters_ratio": 0.0, "trainable_parameters": [], "trainable_parameters_regex": null, "target_modules": [ "all-linear" ], "target_regex": null, "modules_to_save": [], "lora_rank": 8, "lora_alpha": 32, "lora_dropout": 0.05, "lora_bias": "none", "lora_dtype": null, "use_rslora": false, "rlhf_type": null, "loss_type": null, "mcore_ref_model": null, "mcore_ref_adapter": null, "beta": null, "rpo_alpha": null, "reference_free": false, "label_smoothing": 0.0, "f_divergence_type": "reverse_kl", "desirable_weight": 1.0, "undesirable_weight": 1.0, "calculate_KL": null, "center_rewards_coefficient": null, "teacher_model": null, "teacher_model_type": null, "teacher_model_revision": null, "teacher_model_server": null, "gkd_logits_topk": null, "lmbda": 0.5, "seq_kd": false, "offload_teacher_model": false, "sft_alpha": 0.0, "generation_batch_size": null, "steps_per_generation": null, "num_generations": 8, "num_generations_eval": null, "max_completion_length": 512, "importance_sampling_level": "token", "tau_pos": 1.0, "tau_neg": 1.05, "real_tau": 0.5, "fipo_decay_rate": 32.0, "fipo_clip_range": 0.2, "fipo_clip_high_only": true, "fipo_safety_threshold": 4.0, "epsilon": 0.2, "epsilon_high": null, "delta": null, "use_vllm": true, "vllm_mode": null, "vllm_enable_prefix_caching": true, "vllm_gpu_memory_utilization": 0.9, "vllm_tensor_parallel_size": 1, "vllm_max_model_len": null, "vllm_enforce_eager": false, "vllm_limit_mm_per_prompt": null, "vllm_disable_cascade_attn": false, "vllm_max_num_seqs": null, "vllm_mm_processor_cache_gb": null, "vllm_engine_kwargs": null, "vllm_enable_lora": false, "sleep_level": 0, "offload_optimizer": false, "offload_model": false, "offload_bridge": false, "vllm_server_base_url": null, "vllm_server_host": null, "vllm_server_port": [ 8000 ], "vllm_server_timeout": 240.0, "vllm_server_group_port": null, "reward_funcs": [], "reward_weights": null, "cosine_min_len_value_wrong": -0.5, "cosine_max_len_value_wrong": 0.0, "cosine_min_len_value_correct": 1.0, "cosine_max_len_value_correct": 0.5, "cosine_max_len": null, "repetition_n_grams": 3, "repetition_max_penalty": -1.0, "soft_max_length": null, "soft_cache_length": null, "dynamic_sample": false, "max_resample_times": 3, "overlong_filter": false, "scale_rewards": "group", "advantage_estimator": "grpo", "kl_in_reward": false, "wandb_log_unique_prompts": null, "log_completions": false, "rollout_importance_sampling_mode": null, "rollout_importance_sampling_threshold": 2.0, "log_rollout_offpolicy_metrics": false, "off_policy_sequence_mask_delta": null, "log_entropy": false, "top_entropy_quantile": 1.0, "router_replay_mode": "disabled", "reward_model": null, "reward_model_plugin": null, "sync_ref_model": false, "ref_model_sync_steps": 512, "ref_model_mixup_alpha": 0.6, "async_generate": false, "move_model_batches": null, "multi_turn_scheduler": null, "max_turns": null, "completion_length_limit_scope": "per_round", "vllm_server_pass_dataset": false, "num_iterations": 1, "micro_batch_size": 2, "global_batch_size": 32, "recompute_granularity": "full", "recompute_method": "uniform", "recompute_num_layers": 1, "recompute_modules": [ "core_attn" ], "train_iters": null, "num_train_epochs": 1, "masked_softmax_fusion": true, "bias_dropout_fusion": true, "bias_activation_fusion": true, "apply_rope_fusion": false, "gradient_accumulation_fusion": true, "cross_entropy_loss_fusion": true, "cross_entropy_fusion_impl": "te", "calculate_per_token_loss": true, "attention_backend": "", "optimizer": "adam", "optimizer_cpu_offload": true, "optimizer_offload_fraction": 0.7, "use_precision_aware_optimizer": true, "main_grads_dtype": "float32", "main_params_dtype": "float32", "exp_avg_dtype": "float32", "exp_avg_sq_dtype": "float32", "manual_gc": false, "manual_gc_steps": 0, "manual_gc_eval": true, "train_dataloader_shuffle": true, "dataloader_num_workers": 90, "dataloader_pin_memory": true, "dataloader_persistent_workers": false, "dataloader_prefetch_factor": 2, "data_sharding": false, "group_by_length": false, "te_rng_tracker": false, "data_parallel_random_init": false, "mlp_padding_free": false, "lr_warmup_init": 0.0, "lr": 1e-05, "lr_decay_style": "cosine", "lr_decay_iters": null, "lr_warmup_iters": 0, "lr_warmup_fraction": 0.05, "min_lr": 5e-06, "lr_wsd_decay_style": "exponential", "lr_wsd_decay_iters": null, "weight_decay": 0.0001, "weight_decay_incr_style": "constant", "start_weight_decay": 0.0001, "end_weight_decay": 0.0001, "clip_grad": 1.0, "adam_beta1": 0.9, "adam_beta2": 0.95, "adam_eps": 1e-08, "sgd_momentum": 0.9, "muon_momentum": 0.9, "muon_split_qkv": true, "muon_use_nesterov": false, "muon_scale_mode": "spectral", "muon_fp32_matmul_prec": "medium", "muon_num_ns_steps": 5, "muon_tp_mode": "blockwise", "muon_extra_scale_factor": 1.0, "output_dir": "/scratch/mt/new-structure/experiments/r.hany/grad/datafilter-dft/v3-20260623-184940", "save_steps": 20, "no_save_optim": true, "no_save_rng": true, "mcore_model": null, "mcore_adapter": null, "no_load_optim": false, "no_load_rng": false, "finetune": true, "perform_initialization": false, "use_cpu_initialization": false, "async_save": false, "save_total_limit": 8, "metric_for_best_model": "loss", "greater_is_better": false, "use_persistent_ckpt_worker": false, "dist_ckpt_save_pre_mcore_014": false, "dist_ckpt_optim_fully_reshardable": false, "distrib_optim_fully_reshardable_mem_efficient": false, "local_rank": 0, "use_distributed_optimizer": true, "tensor_model_parallel_size": 1, "pipeline_model_parallel_size": 1, "decoder_first_pipeline_num_layers": null, "decoder_last_pipeline_num_layers": null, "account_for_embedding_in_pipeline_split": false, "account_for_loss_in_pipeline_split": false, "overlap_p2p_comm": false, "align_param_gather": false, "sequence_parallel": false, "context_parallel_size": 1, "tp_comm_overlap": false, "overlap_grad_reduce": false, "overlap_param_gather": false, "overlap_param_gather_with_optimizer_step": false, "align_grad_reduce": true, "virtual_pipeline_model_parallel_size": null, "microbatch_group_size_per_vp_stage": null, "pipeline_model_parallel_layout": null, "expert_model_parallel_size": 1, "expert_tensor_parallel_size": 1, "report_to": [ "wandb" ], "logging_steps": 1, "tensorboard_dir": "/scratch/mt/new-structure/experiments/r.hany/grad/datafilter-dft/v3-20260623-184940/runs", "tensorboard_queue_size": 50, "wandb_project": "GraduationProject", "wandb_exp_name": "datafilter-dft", "swanlab_project": "megatron-swift", "swanlab_exp_name": null, "eval_iters": -1, "eval_steps": 20, "fp8_format": null, "fp8_recipe": "delayed", "fp8_param_gather": false, "fp8_amax_history_len": 1024, "fp8_amax_compute_algo": "max", "fp4_format": null, "fp4_recipe": "nvfp4", "fp4_param_gather": false, "fp16": false, "bf16": true, "apply_query_key_layer_scaling": false, "attention_softmax_in_fp32": true, "accumulate_allreduce_grads_in_fp32": true, "moe_router_load_balancing_type": null, "moe_router_dtype": "fp32", "moe_token_dispatcher_type": null, "moe_enable_deepep": false, "moe_grouped_gemm": true, "moe_permute_fusion": false, "moe_aux_loss_coeff": 0.0, "moe_z_loss_coeff": null, "moe_shared_expert_overlap": false, "moe_layer_recompute": false, "moe_expert_capacity_factor": null, "moe_pad_expert_input_to_capacity": false, "moe_token_drop_policy": null, "mtp_num_layers": null, "mtp_loss_scaling_factor": 0.1, "mtp_decoder_input_detach": false, "mtp_shared_weights": false, "save_safetensors": true, "ref_model": null, "ref_adapters": [], "merge_lora": true, "max_shard_size": "5GB", "vit_gradient_checkpointing": false, "vit_gradient_checkpointing_kwargs": null, "vit_attn_impl": null, "vit_lr": null, "aligner_lr": null, "dsa_indexer_loss_coeff": 0.0, "dsa_indexer_use_sparse_loss": false, "check_model": true, "apply_wd_to_qk_layernorm": false, "linear_decoupled_in_proj": false, "enable_dft_loss": true, "enable_channel_loss": false, "save_strategy": "steps", "callbacks": [ "print", "default_flow", "wandb" ], "add_version": true, "create_checkpoint_symlink": false, "swift_version": "4.2.3", "ckpt_dir": null, "rank": 0, "global_world_size": 1, "local_world_size": 1, "model_suffix": "Datafilter", "model_info": "ModelInfo(model_type='llama', model_dir='/scratch/mt/new-structure/experiments/r.hany/grad/Datafilter', torch_dtype=torch.bfloat16, max_model_len=131072, quant_method=None, quant_bits=None, rope_scaling={'factor': 8.0, 'high_freq_factor': 4.0, 'low_freq_factor': 1.0, 'original_max_position_embeddings': 8192, 'rope_type': 'llama3', 'rope_theta': 500000.0}, is_moe_model=False, is_multimodal=False, config=None, task_type='causal_lm', num_labels=None)", "model_meta": "ModelMeta(model_type='llama', model_groups=[ModelGroup(models=[Model(ms_model_id='modelscope/Llama-2-7b-ms', hf_model_id='meta-llama/Llama-2-7b-hf', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='modelscope/Llama-2-13b-ms', hf_model_id='meta-llama/Llama-2-13b-hf', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='modelscope/Llama-2-70b-ms', hf_model_id='meta-llama/Llama-2-70b-hf', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='modelscope/Llama-2-7b-chat-ms', hf_model_id='meta-llama/Llama-2-7b-chat-hf', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='modelscope/Llama-2-13b-chat-ms', hf_model_id='meta-llama/Llama-2-13b-chat-hf', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='modelscope/Llama-2-70b-chat-ms', hf_model_id='meta-llama/Llama-2-70b-chat-hf', model_path=None, ms_revision=None, hf_revision=None)], template='llama', ignore_patterns=['.+\\\\.bin$'], requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='AI-ModelScope/chinese-llama-2-1.3b', hf_model_id='hfl/chinese-llama-2-1.3b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-llama-2-7b', hf_model_id='hfl/chinese-llama-2-7b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-llama-2-7b-16k', hf_model_id='hfl/chinese-llama-2-7b-16k', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-llama-2-7b-64k', hf_model_id='hfl/chinese-llama-2-7b-64k', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-llama-2-13b', hf_model_id='hfl/chinese-llama-2-13b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-llama-2-13b-16k', hf_model_id='hfl/chinese-llama-2-13b-16k', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-1.3b', hf_model_id='hfl/chinese-alpaca-2-1.3b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-7b', hf_model_id='hfl/chinese-alpaca-2-7b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-7b-16k', hf_model_id='hfl/chinese-alpaca-2-7b-16k', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-7b-64k', hf_model_id='hfl/chinese-alpaca-2-7b-64k', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-13b', hf_model_id='hfl/chinese-alpaca-2-13b', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='AI-ModelScope/chinese-alpaca-2-13b-16k', hf_model_id='hfl/chinese-alpaca-2-13b-16k', model_path=None, ms_revision=None, hf_revision=None)], template='llama', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='AI-ModelScope/Llama-2-7b-AQLM-2Bit-1x16-hf', hf_model_id='ISTA-DASLab/Llama-2-7b-AQLM-2Bit-1x16-hf', model_path=None, ms_revision=None, hf_revision=None)], template='llama', ignore_patterns=None, requires=['transformers>=4.38', 'aqlm', 'torch>=2.2.0'], tags=[]), ModelGroup(models=[Model(ms_model_id='FlagAlpha/Atom-7B', hf_model_id='FlagAlpha/Atom-7B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='FlagAlpha/Atom-7B-Chat', hf_model_id='FlagAlpha/Atom-7B-Chat', model_path=None, ms_revision=None, hf_revision=None)], template='atom', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='langboat/Mengzi3-13B-Base', hf_model_id='Langboat/Mengzi3-13B-Base', model_path=None, ms_revision=None, hf_revision=None)], template='mengzi', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='AI-ModelScope/NuminaMath-7B-TIR', hf_model_id='AI-MO/NuminaMath-7B-TIR', model_path=None, ms_revision=None, hf_revision=None)], template='numina', ignore_patterns=None, requires=None, tags=['math']), ModelGroup(models=[Model(ms_model_id='Fengshenbang/Ziya2-13B-Base', hf_model_id='IDEA-CCNL/Ziya2-13B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='Fengshenbang/Ziya2-13B-Chat', hf_model_id='IDEA-CCNL/Ziya2-13B-Chat', model_path=None, ms_revision=None, hf_revision=None)], template='ziya', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='InfiniAI/Megrez-3b-Instruct', hf_model_id='Infinigence/Megrez-3B-Instruct', model_path=None, ms_revision=None, hf_revision=None)], template='megrez', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='deepseek-ai/deepseek-llm-7b-base', hf_model_id='deepseek-ai/deepseek-llm-7b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-llm-7b-chat', hf_model_id='deepseek-ai/deepseek-llm-7b-chat', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-llm-67b-base', hf_model_id='deepseek-ai/deepseek-llm-67b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-llm-67b-chat', hf_model_id='deepseek-ai/deepseek-llm-67b-chat', model_path=None, ms_revision=None, hf_revision=None)], template='deepseek', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='deepseek-ai/deepseek-math-7b-base', hf_model_id='deepseek-ai/deepseek-math-7b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-math-7b-instruct', hf_model_id='deepseek-ai/deepseek-math-7b-instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-math-7b-rl', hf_model_id='deepseek-ai/deepseek-math-7b-rl', model_path=None, ms_revision=None, hf_revision=None)], template='deepseek', ignore_patterns=None, requires=None, tags=['math']), ModelGroup(models=[Model(ms_model_id='deepseek-ai/deepseek-coder-1.3b-base', hf_model_id='deepseek-ai/deepseek-coder-1.3b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-coder-1.3b-instruct', hf_model_id='deepseek-ai/deepseek-coder-1.3b-instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-coder-6.7b-base', hf_model_id='deepseek-ai/deepseek-coder-6.7b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-coder-6.7b-instruct', hf_model_id='deepseek-ai/deepseek-coder-6.7b-instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-coder-33b-base', hf_model_id='deepseek-ai/deepseek-coder-33b-base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/deepseek-coder-33b-instruct', hf_model_id='deepseek-ai/deepseek-coder-33b-instruct', model_path=None, ms_revision=None, hf_revision=None)], template='deepseek', ignore_patterns=None, requires=None, tags=['coding']), ModelGroup(models=[Model(ms_model_id='gongjy/MiniMind2', hf_model_id='jingyaogong/MiniMind2', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id=None, hf_model_id='jingyaogong/MiniMind2-Small', model_path=None, ms_revision=None, hf_revision=None)], template='minimind', ignore_patterns=None, requires=['transformers>=4.57.1'], tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Meta-Llama-3-8B-Instruct', hf_model_id='meta-llama/Meta-Llama-3-8B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3-70B-Instruct', hf_model_id='meta-llama/Meta-Llama-3-70B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3-8B', hf_model_id='meta-llama/Meta-Llama-3-8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3-70B', hf_model_id='meta-llama/Meta-Llama-3-70B', model_path=None, ms_revision=None, hf_revision=None)], template='llama3', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='swift/Meta-Llama-3-8B-Instruct-GPTQ-Int4', hf_model_id='study-hjt/Meta-Llama-3-8B-Instruct-GPTQ-Int4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='swift/Meta-Llama-3-8B-Instruct-GPTQ-Int8', hf_model_id='study-hjt/Meta-Llama-3-8B-Instruct-GPTQ-Int8', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='swift/Meta-Llama-3-8B-Instruct-AWQ', hf_model_id='study-hjt/Meta-Llama-3-8B-Instruct-AWQ', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='swift/Meta-Llama-3-70B-Instruct-GPTQ-Int4', hf_model_id='study-hjt/Meta-Llama-3-70B-Instruct-GPTQ-Int4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='swift/Meta-Llama-3-70B-Instruct-GPTQ-Int8', hf_model_id='study-hjt/Meta-Llama-3-70B-Instruct-GPTQ-Int8', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='swift/Meta-Llama-3-70B-Instruct-AWQ', hf_model_id='study-hjt/Meta-Llama-3-70B-Instruct-AWQ', model_path=None, ms_revision=None, hf_revision=None)], template='llama3', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='ChineseAlpacaGroup/llama-3-chinese-8b-instruct', hf_model_id='hfl/llama-3-chinese-8b-instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='ChineseAlpacaGroup/llama-3-chinese-8b', hf_model_id='hfl/llama-3-chinese-8b', model_path=None, ms_revision=None, hf_revision=None)], template='llama3', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Meta-Llama-3.1-8B-Instruct', hf_model_id='meta-llama/Meta-Llama-3.1-8B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B-Instruct', hf_model_id='meta-llama/Meta-Llama-3.1-70B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B-Instruct', hf_model_id='meta-llama/Meta-Llama-3.1-405B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-8B', hf_model_id='meta-llama/Meta-Llama-3.1-8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B', hf_model_id='meta-llama/Meta-Llama-3.1-70B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B', hf_model_id='meta-llama/Meta-Llama-3.1-405B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B-Instruct-FP8', hf_model_id='meta-llama/Meta-Llama-3.1-70B-Instruct-FP8', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B-Instruct-FP8', hf_model_id='meta-llama/Meta-Llama-3.1-405B-Instruct-FP8', model_path=None, ms_revision=None, hf_revision=None)], template='llama3_2', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Meta-Llama-3.1-8B-Instruct-BNB-NF4', hf_model_id='hugging-quants/Meta-Llama-3.1-8B-Instruct-BNB-NF4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B-Instruct-bnb-4bit', hf_model_id='unsloth/Meta-Llama-3.1-70B-Instruct-bnb-4bit', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B-Instruct-BNB-NF4', hf_model_id='hugging-quants/Meta-Llama-3.1-405B-Instruct-BNB-NF4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-8B-Instruct-GPTQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-8B-Instruct-GPTQ-INT4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B-Instruct-GPTQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-70B-Instruct-GPTQ-INT4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B-Instruct-GPTQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-405B-Instruct-GPTQ-INT4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-8B-Instruct-AWQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-70B-Instruct-AWQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-70B-Instruct-AWQ-INT4', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Meta-Llama-3.1-405B-Instruct-AWQ-INT4', hf_model_id='hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4', model_path=None, ms_revision=None, hf_revision=None)], template='llama3_2', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='AI-ModelScope/Llama-3.1-Nemotron-70B-Instruct-HF', hf_model_id='nvidia/Llama-3.1-Nemotron-70B-Instruct-HF', model_path=None, ms_revision=None, hf_revision=None)], template='llama3_2', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='AI-ModelScope/Skywork-o1-Open-Llama-3.1-8B', hf_model_id='Skywork/Skywork-o1-Open-Llama-3.1-8B', model_path=None, ms_revision=None, hf_revision=None)], template='skywork_o1', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.2-1B', hf_model_id='meta-llama/Llama-3.2-1B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B', hf_model_id='meta-llama/Llama-3.2-3B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-1B-Instruct', hf_model_id='meta-llama/Llama-3.2-1B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='LLM-Research/Llama-3.2-3B-Instruct', hf_model_id='meta-llama/Llama-3.2-3B-Instruct', model_path=None, ms_revision=None, hf_revision=None)], template='llama3_2', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Llama-3.3-70B-Instruct', hf_model_id='meta-llama/Llama-3.3-70B-Instruct', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', hf_model_id='unsloth/Llama-3.3-70B-Instruct-bnb-4bit', model_path=None, ms_revision=None, hf_revision=None)], template='llama3_2', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='ZhipuAI/LongWriter-llama3.1-8b', hf_model_id='zai-org/LongWriter-llama3.1-8b', model_path=None, ms_revision=None, hf_revision=None)], template='longwriter_llama', ignore_patterns=None, requires=['transformers>=4.43'], tags=[]), ModelGroup(models=[Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-70B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-70B', model_path=None, ms_revision=None, hf_revision=None)], template='deepseek_r1', ignore_patterns=None, requires=None, tags=[]), ModelGroup(models=[Model(ms_model_id='OpenBMB/MiniCPM5-1B', hf_model_id='openbmb/MiniCPM5-1B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='OpenBMB/MiniCPM5-1B-Base', hf_model_id='openbmb/MiniCPM5-1B-Base', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='OpenBMB/MiniCPM5-1B-SFT', hf_model_id='openbmb/MiniCPM5-1B-SFT', model_path=None, ms_revision=None, hf_revision=None)], template='minicpm5', ignore_patterns=None, requires=['transformers>=5.6'], tags=[]), ModelGroup(models=[Model(ms_model_id='LLM-Research/Reflection-Llama-3.1-70B', hf_model_id='mattshumer/Reflection-Llama-3.1-70B', model_path=None, ms_revision=None, hf_revision=None)], template='reflection', ignore_patterns=None, requires=['transformers>=4.43'], tags=[])], loader=, template=None, model_arch=ModelKeys(arch_name='llama', embedding='model.embed_tokens', module_list='model.layers', lm_head='lm_head', q_proj='model.layers.{}.self_attn.q_proj', k_proj='model.layers.{}.self_attn.k_proj', v_proj='model.layers.{}.self_attn.v_proj', o_proj='model.layers.{}.self_attn.o_proj', attention='model.layers.{}.self_attn', mlp='model.layers.{}.mlp', down_proj='model.layers.{}.mlp.down_proj', qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None), mcore_model_type=None, architectures=['LlamaForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=False, is_reward=False, task_type=None, ignore_patterns=None, requires=[], tags=[])", "model_dir": "/scratch/mt/new-structure/experiments/r.hany/grad/Datafilter", "template_meta": "Llama3_2TemplateMeta(template_type='llama3_2', prefix=['<|begin_of_text|><|start_header_id|>system<|end_header_id|>\\n\\nCutting Knowledge Date: December 2023\\nToday Date: 23 Jun 2026\\n\\n<|eot_id|>'], prompt=['<|start_header_id|>user<|end_header_id|>\\n\\n{{QUERY}}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\\n\\n'], chat_sep=['<|eot_id|>'], suffix=['<|eot_id|>'], template_cls=, system_prefix=['<|begin_of_text|><|start_header_id|>system<|end_header_id|>\\n\\nCutting Knowledge Date: December 2023\\nToday Date: 23 Jun 2026\\n\\n{{SYSTEM}}<|eot_id|>'], default_system=None, auto_add_bos=False, stop_words=[], agent_template='llama3', is_thinking=False, thinking_prefix='', non_thinking_prefix='', history_thinking_prefix='')", "_val_dataset_exists": true, "hub": "", "seq_length": 1024, "is_multimodal": false, "megatron_model_meta": "ModelMeta(model_type='gpt', model_types=['qwen2', 'llama', 'qwen3', 'qwen2_moe', 'qwen3_moe', 'internlm3', 'mimo', 'deepseek', 'deepseek_v2', 'deepseek_v3', 'deepseek_v32', 'kimi_k2', 'dots1', 'ernie4_5', 'ernie4_5_moe', 'glm4_moe', 'glm4_moe_lite', 'glm_moe_dsa', 'gpt_oss'], bridge_cls=, visual_cls=None, is_multimodal=False, loader=)", "fp8": null, "fp4": null, "world_size": 1, "data_parallel_size": 1, "num_microbatches": 16 }