| { |
| "dataset_path": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/prepared_datasets/Qwen2.5-7B/code/full_drop_over_2048_kept_target_v2", |
| "dataset_rows": 50000, |
| "hub_model_id": "jalenluorion/Qwen2.5-7B_code", |
| "model": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/runtime_cache/hf_home/hub/models--Qwen--Qwen2.5-7B/snapshots/d149729398750b98c0af14eb82c78cfe92750796", |
| "output_dir": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/output_root/Qwen2.5-7B_code", |
| "resolved_config": { |
| "model": { |
| "attn_implementation": "sdpa", |
| "family": "qwen", |
| "low_cpu_mem_usage": true, |
| "model_name_or_path": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/runtime_cache/hf_home/hub/models--Qwen--Qwen2.5-7B/snapshots/d149729398750b98c0af14eb82c78cfe92750796", |
| "prepared_dataset_key": "Qwen2.5-7B", |
| "torch_dtype": "bfloat16", |
| "trust_remote_code": true |
| }, |
| "run": { |
| "analysis_backend": "cpu", |
| "analysis_device": null, |
| "analysis_visible_devices": null, |
| "analysis_workers": 8, |
| "checkpoint_copy_latest_only": true, |
| "checkpoint_metric_profile": "checkpoint_scalar", |
| "final_metric_profile": "standard", |
| "hub_private_repo": false, |
| "hub_strategy": "end", |
| "max_input_tokens": 2048, |
| "max_steps_override": null, |
| "max_train_samples": null, |
| "min_free_gb": 150, |
| "output_root": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/output_root", |
| "persist_root": "/data/user_data/jalenl/trl_sft_l40s_outputs", |
| "prepared_dataset_root": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/prepared_datasets", |
| "profile_name": "l40s_4gpu_cpu_analysis", |
| "push_final_to_hub": true, |
| "remove_checkpoints_after_final": true, |
| "seed": 42, |
| "spectral_scratch_root": "/scratch/trl_sft_l40s_jalenl/spectral_snapshots", |
| "wandb_project": "spectral-phase3-trl-sft-l40s" |
| }, |
| "task": { |
| "name": "code" |
| }, |
| "training": { |
| "assistant_only_loss": false, |
| "bf16": true, |
| "completion_only_loss": true, |
| "dataloader_num_workers": 4, |
| "dataloader_persistent_workers": true, |
| "dataloader_pin_memory": true, |
| "dataloader_prefetch_factor": 2, |
| "dataset_num_proc": 8, |
| "ddp_timeout": 21600, |
| "deepspeed": null, |
| "distributed_strategy": "fsdp", |
| "eos_token": "<|im_end|>", |
| "fsdp": "full_shard auto_wrap", |
| "fsdp_config": { |
| "activation_checkpointing": true, |
| "auto_wrap_policy": "TRANSFORMER_BASED_WRAP", |
| "backward_prefetch": "BACKWARD_PRE", |
| "limit_all_gathers": true, |
| "min_num_params": 0, |
| "use_orig_params": true, |
| "xla": false, |
| "xla_fsdp_grad_ckpt": false, |
| "xla_fsdp_v2": false |
| }, |
| "gradient_accumulation_steps": 16, |
| "gradient_checkpointing": true, |
| "learning_rate": 2e-05, |
| "logging_steps": 10, |
| "logging_strategy": "steps", |
| "lr_scheduler_type": "cosine", |
| "max_length": null, |
| "num_train_epochs": 3, |
| "optim": "adamw_torch_fused", |
| "packing": false, |
| "packing_strategy": "bfd_split", |
| "padding_free": false, |
| "per_device_train_batch_size": 1, |
| "remove_unused_columns": false, |
| "report_to": [], |
| "save_steps": null, |
| "save_strategy": "steps", |
| "save_total_limit": 2, |
| "tf32": true, |
| "truncation_mode": "keep_start", |
| "use_liger_kernel": false, |
| "warmup_ratio": 0.03, |
| "weight_decay": 0.01 |
| } |
| }, |
| "resume_from_checkpoint": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/resume_checkpoint/checkpoint-1950", |
| "run_name": "Qwen2.5-7B_code", |
| "task": "code", |
| "train_dataset_columns": [ |
| "row_id", |
| "task", |
| "input_ids", |
| "completion_mask" |
| ], |
| "trainer_args": { |
| "accelerator_config": { |
| "dispatch_batches": null, |
| "even_batches": true, |
| "gradient_accumulation_kwargs": null, |
| "non_blocking": false, |
| "split_batches": false, |
| "use_seedable_sampler": true |
| }, |
| "activation_offloading": false, |
| "adafactor": false, |
| "adam_beta1": 0.9, |
| "adam_beta2": 0.999, |
| "adam_epsilon": 1e-08, |
| "assistant_only_loss": false, |
| "auto_find_batch_size": false, |
| "average_tokens_across_devices": true, |
| "batch_eval_metrics": false, |
| "bf16": true, |
| "bf16_full_eval": false, |
| "chat_template_path": null, |
| "completion_only_loss": true, |
| "data_seed": 42, |
| "dataloader_drop_last": false, |
| "dataloader_num_workers": 4, |
| "dataloader_persistent_workers": true, |
| "dataloader_pin_memory": true, |
| "dataloader_prefetch_factor": 2, |
| "dataset_kwargs": null, |
| "dataset_num_proc": 8, |
| "dataset_text_field": "text", |
| "ddp_backend": null, |
| "ddp_broadcast_buffers": null, |
| "ddp_bucket_cap_mb": null, |
| "ddp_find_unused_parameters": null, |
| "ddp_timeout": 21600, |
| "debug": [], |
| "deepspeed": null, |
| "disable_tqdm": false, |
| "do_eval": false, |
| "do_predict": false, |
| "do_train": false, |
| "eos_token": "<EOS_TOKEN>", |
| "eval_accumulation_steps": null, |
| "eval_delay": 0, |
| "eval_do_concat_batches": true, |
| "eval_on_start": false, |
| "eval_packing": null, |
| "eval_steps": null, |
| "eval_strategy": "no", |
| "eval_use_gather_object": false, |
| "fp16": false, |
| "fp16_backend": "auto", |
| "fp16_full_eval": false, |
| "fp16_opt_level": "O1", |
| "fsdp": [ |
| "full_shard", |
| "auto_wrap" |
| ], |
| "fsdp_config": { |
| "activation_checkpointing": true, |
| "auto_wrap_policy": "TRANSFORMER_BASED_WRAP", |
| "backward_prefetch": "BACKWARD_PRE", |
| "limit_all_gathers": true, |
| "min_num_params": 0, |
| "use_orig_params": true, |
| "xla": false, |
| "xla_fsdp_grad_ckpt": false, |
| "xla_fsdp_v2": false |
| }, |
| "fsdp_min_num_params": 0, |
| "fsdp_transformer_layer_cls_to_wrap": null, |
| "full_determinism": false, |
| "gradient_accumulation_steps": 16, |
| "gradient_checkpointing": false, |
| "gradient_checkpointing_kwargs": null, |
| "greater_is_better": null, |
| "group_by_length": false, |
| "half_precision_backend": "auto", |
| "hub_always_push": false, |
| "hub_model_id": "jalenluorion/Qwen2.5-7B_code", |
| "hub_private_repo": false, |
| "hub_revision": null, |
| "hub_strategy": "end", |
| "hub_token": "<HUB_TOKEN>", |
| "ignore_data_skip": false, |
| "include_for_metrics": [], |
| "include_inputs_for_metrics": false, |
| "include_num_input_tokens_seen": "no", |
| "include_tokens_per_second": false, |
| "jit_mode_eval": false, |
| "label_names": null, |
| "label_smoothing_factor": 0.0, |
| "learning_rate": 2e-05, |
| "length_column_name": "length", |
| "liger_kernel_config": null, |
| "load_best_model_at_end": false, |
| "local_rank": 0, |
| "log_level": "passive", |
| "log_level_replica": "warning", |
| "log_on_each_node": true, |
| "logging_dir": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/output_root/Qwen2.5-7B_code/runs/Apr20_17-30-49_babel-x5-20", |
| "logging_first_step": false, |
| "logging_nan_inf_filter": true, |
| "logging_steps": 10, |
| "logging_strategy": "steps", |
| "loss_type": "nll", |
| "lr_scheduler_kwargs": null, |
| "lr_scheduler_type": "cosine", |
| "max_grad_norm": 1.0, |
| "max_length": null, |
| "max_steps": -1, |
| "metric_for_best_model": null, |
| "model_init_kwargs": { |
| "attn_implementation": "sdpa", |
| "device_map": null, |
| "dtype": "bfloat16", |
| "low_cpu_mem_usage": true, |
| "trust_remote_code": true |
| }, |
| "mp_parameters": "", |
| "neftune_noise_alpha": null, |
| "no_cuda": false, |
| "num_train_epochs": 3, |
| "optim": "adamw_torch_fused", |
| "optim_args": null, |
| "optim_target_modules": null, |
| "output_dir": "/scratch/trl_sft_l40s_jalenl_7277350_0/Qwen2.5-7B_code/output_root/Qwen2.5-7B_code", |
| "overwrite_output_dir": false, |
| "packing": false, |
| "packing_strategy": "bfd_split", |
| "pad_to_multiple_of": null, |
| "pad_token": "<PAD_TOKEN>", |
| "padding_free": false, |
| "parallelism_config": null, |
| "past_index": -1, |
| "per_device_eval_batch_size": 8, |
| "per_device_train_batch_size": 1, |
| "per_gpu_eval_batch_size": null, |
| "per_gpu_train_batch_size": null, |
| "prediction_loss_only": false, |
| "project": "huggingface", |
| "push_to_hub": false, |
| "push_to_hub_model_id": null, |
| "push_to_hub_organization": null, |
| "push_to_hub_token": "<PUSH_TO_HUB_TOKEN>", |
| "ray_scope": "last", |
| "remove_unused_columns": false, |
| "report_to": [], |
| "restore_callback_states_from_checkpoint": false, |
| "resume_from_checkpoint": null, |
| "run_name": "Qwen2.5-7B_code", |
| "save_on_each_node": false, |
| "save_only_model": false, |
| "save_safetensors": true, |
| "save_steps": 390, |
| "save_strategy": "steps", |
| "save_total_limit": 2, |
| "seed": 42, |
| "shuffle_dataset": false, |
| "skip_memory_metrics": true, |
| "tf32": true, |
| "torch_compile": false, |
| "torch_compile_backend": null, |
| "torch_compile_mode": null, |
| "torch_empty_cache_steps": null, |
| "torchdynamo": null, |
| "tpu_metrics_debug": false, |
| "tpu_num_cores": null, |
| "trackio_space_id": "trackio", |
| "truncation_mode": "keep_start", |
| "use_cpu": false, |
| "use_legacy_prediction_loop": false, |
| "use_liger_kernel": false, |
| "use_mps_device": false, |
| "warmup_ratio": 0.03, |
| "warmup_steps": 0, |
| "weight_decay": 0.01 |
| } |
| } |
