| { |
| "model_args": { |
| "model_name_or_path": "/data/oss_bucket_0/liqi/models/Llama-3.2-3B-Instruct", |
| "separate_decoder_name": "", |
| "lora_r": 128, |
| "lora_dropout": 0.05, |
| "full_precision": true, |
| "train": true, |
| "lora_init": true, |
| "token": null, |
| "adapter_name_or_path": null, |
| "lora_alpha": 32, |
| "ckpt_dir": null |
| }, |
| "data_args": { |
| "data_name": "icot", |
| "debug_data": false, |
| "batch_size": 1 |
| }, |
| "training_args": { |
| "output_dir": "/data/oss_bucket_0/liqi/larger_model/llama-3b/gsm8k/codi_l6_lr3e4_btz128_prj3072_ep10/gsm8k_gpt_pmps", |
| "overwrite_output_dir": false, |
| "do_train": true, |
| "do_eval": false, |
| "do_predict": false, |
| "eval_strategy": "no", |
| "prediction_loss_only": false, |
| "per_device_train_batch_size": 16, |
| "per_device_eval_batch_size": 1, |
| "per_gpu_train_batch_size": null, |
| "per_gpu_eval_batch_size": null, |
| "gradient_accumulation_steps": 1, |
| "eval_accumulation_steps": null, |
| "eval_delay": 0, |
| "torch_empty_cache_steps": null, |
| "learning_rate": 0.0003, |
| "weight_decay": 0.1, |
| "adam_beta1": 0.9, |
| "adam_beta2": 0.999, |
| "adam_epsilon": 1e-08, |
| "max_grad_norm": 2.0, |
| "num_train_epochs": 15.0, |
| "max_steps": -1, |
| "lr_scheduler_type": "cosine", |
| "lr_scheduler_kwargs": {}, |
| "warmup_ratio": 0.03, |
| "warmup_steps": 0, |
| "log_level": "passive", |
| "log_level_replica": "warning", |
| "log_on_each_node": true, |
| "logging_dir": "/data/oss_bucket_0/liqi/larger_model/llama-3b/gsm8k/codi_l6_lr3e4_btz128_prj3072_ep10/logs", |
| "logging_strategy": "steps", |
| "logging_first_step": false, |
| "logging_steps": 10, |
| "logging_nan_inf_filter": true, |
| "save_strategy": "epoch", |
| "save_steps": 500, |
| "save_total_limit": 40, |
| "save_safetensors": false, |
| "save_on_each_node": false, |
| "save_only_model": false, |
| "restore_callback_states_from_checkpoint": false, |
| "no_cuda": false, |
| "use_cpu": false, |
| "use_mps_device": false, |
| "seed": 11, |
| "data_seed": null, |
| "jit_mode_eval": false, |
| "use_ipex": false, |
| "bf16": true, |
| "fp16": false, |
| "fp16_opt_level": "O1", |
| "half_precision_backend": "auto", |
| "bf16_full_eval": false, |
| "fp16_full_eval": false, |
| "tf32": null, |
| "local_rank": 0, |
| "ddp_backend": null, |
| "tpu_num_cores": null, |
| "tpu_metrics_debug": false, |
| "debug": [], |
| "dataloader_drop_last": false, |
| "eval_steps": null, |
| "dataloader_num_workers": 0, |
| "dataloader_prefetch_factor": null, |
| "past_index": -1, |
| "run_name": "/data/oss_bucket_0/liqi/larger_model/llama-3b/gsm8k/codi_l6_lr3e4_btz128_prj3072_ep10", |
| "disable_tqdm": false, |
| "remove_unused_columns": true, |
| "label_names": null, |
| "load_best_model_at_end": false, |
| "metric_for_best_model": null, |
| "greater_is_better": null, |
| "ignore_data_skip": false, |
| "fsdp": [], |
| "fsdp_min_num_params": 0, |
| "fsdp_config": { |
| "min_num_params": 0, |
| "xla": false, |
| "xla_fsdp_v2": false, |
| "xla_fsdp_grad_ckpt": false |
| }, |
| "fsdp_transformer_layer_cls_to_wrap": null, |
| "accelerator_config": { |
| "split_batches": false, |
| "dispatch_batches": null, |
| "even_batches": true, |
| "use_seedable_sampler": true, |
| "non_blocking": false, |
| "gradient_accumulation_kwargs": null, |
| "use_configured_state": false |
| }, |
| "deepspeed": null, |
| "label_smoothing_factor": 0.0, |
| "optim": "adamw_torch", |
| "optim_args": null, |
| "adafactor": false, |
| "group_by_length": false, |
| "length_column_name": "length", |
| "report_to": [ |
| "wandb" |
| ], |
| "ddp_find_unused_parameters": null, |
| "ddp_bucket_cap_mb": null, |
| "ddp_broadcast_buffers": null, |
| "dataloader_pin_memory": true, |
| "dataloader_persistent_workers": false, |
| "skip_memory_metrics": true, |
| "use_legacy_prediction_loop": false, |
| "push_to_hub": false, |
| "resume_from_checkpoint": null, |
| "hub_model_id": null, |
| "hub_strategy": "every_save", |
| "hub_token": null, |
| "hub_private_repo": null, |
| "hub_always_push": false, |
| "gradient_checkpointing": false, |
| "gradient_checkpointing_kwargs": null, |
| "include_inputs_for_metrics": false, |
| "include_for_metrics": [], |
| "eval_do_concat_batches": true, |
| "fp16_backend": "auto", |
| "push_to_hub_model_id": null, |
| "push_to_hub_organization": null, |
| "push_to_hub_token": null, |
| "mp_parameters": "", |
| "auto_find_batch_size": false, |
| "full_determinism": false, |
| "torchdynamo": null, |
| "ray_scope": "last", |
| "ddp_timeout": 1800, |
| "torch_compile": false, |
| "torch_compile_backend": null, |
| "torch_compile_mode": null, |
| "include_tokens_per_second": false, |
| "include_num_input_tokens_seen": false, |
| "neftune_noise_alpha": null, |
| "optim_target_modules": null, |
| "batch_eval_metrics": false, |
| "eval_on_start": false, |
| "use_liger_kernel": false, |
| "eval_use_gather_object": false, |
| "average_tokens_across_devices": false, |
| "cache_dir": null, |
| "model_max_length": 512, |
| "restore_from": "", |
| "expt_name": "gsm8k_gpt_pmps", |
| "icot_train_path": "/users/k24020023/efficient_cot/icae/code/coconut/icot_gsm8k/train.txt", |
| "num_latent": 6, |
| "use_lora": true, |
| "greedy": false, |
| "exp_mode": false, |
| "exp_data_num": 200, |
| "use_prj": true, |
| "prj_dim": 3072, |
| "prj_dropout": 0.0, |
| "prj_no_ln": false, |
| "distill_loss_div_std": true, |
| "distill_loss_type": "smooth_l1", |
| "distill_loss_factor": 20.0, |
| "ref_loss_factor": 1.0, |
| "inf_latent_iterations": 1, |
| "inf_num_iterations": 5, |
| "remove_eos": true, |
| "use_latent_reasoning": true, |
| "print_ref_model_stats": true, |
| "include_last_cot": true, |
| "fix_attn_mask": false, |
| "log_full": false, |
| "print_loss": true, |
| "max_token_num": 200, |
| "enable_top_k": false, |
| "latent_top_k": 30, |
| "latent_top_k_temperature": 1.0, |
| "pmps_loss_factor": 0.0, |
| "pmps_temperature": 0.1, |
| "pmps_match_temperature": 0.1, |
| "pmps_epsilon": 0.05, |
| "pmps_sinkhorn_iterations": 3, |
| "pmps_nmb_prototypes": 32, |
| "pmps_feat_dim": 128, |
| "pmps_hidden_mlp": 512, |
| "pmps_use_projector": true, |
| "pmps_freeze_prototypes_steps": 0, |
| "pmps_layer": "last", |
| "pmps_start_epoch": 0.0, |
| "pmps_position_bias_max": 6.0, |
| "pmps_position_bias_schedule": "cosine", |
| "diversity_loss_factor": 0.0, |
| "use_wandb": true, |
| "wandb_api_key": "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", |
| "wandb_project": "codi-pmps-llama-3B", |
| "wandb_run_name": "gsm8k_codi_l6_lr3e4_btz128_prj3072_ep10" |
| } |
| } |
