training_args.json

{
  "output_dir": "./models/kubhist-byt5-small/",
  "overwrite_output_dir": true,
  "do_train": true,
  "do_eval": true,
  "do_predict": true,
  "eval_strategy": "steps",
  "prediction_loss_only": false,
  "per_device_train_batch_size": 128,
  "per_device_eval_batch_size": 128,
  "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": 0.0005,
  "weight_decay": 0.01,
  "adam_beta1": 0.9,
  "adam_beta2": 0.999,
  "adam_epsilon": 1e-08,
  "max_grad_norm": 1.0,
  "num_train_epochs": 3.0,
  "max_steps": 50000,
  "lr_scheduler_type": "warmup_stable_decay",
  "lr_scheduler_kwargs": {
    "num_stable_steps": 30000,
    "num_decay_steps": 10000
  },
  "warmup_ratio": 0.0,
  "warmup_steps": 10000,
  "log_level": "passive",
  "log_level_replica": "warning",
  "log_on_each_node": true,
  "logging_dir": "./models/kubhist-byt5-small/runs/Oct07_17-48-03_kblabb-l04.kb.local",
  "logging_strategy": "steps",
  "logging_first_step": false,
  "logging_steps": 10,
  "logging_nan_inf_filter": true,
  "save_strategy": "steps",
  "save_steps": 1000,
  "save_total_limit": 3,
  "save_safetensors": true,
  "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": 42,
  "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": 1000,
  "dataloader_num_workers": 8,
  "dataloader_prefetch_factor": 2,
  "past_index": -1,
  "run_name": "./models/kubhist-byt5-small/",
  "disable_tqdm": false,
  "remove_unused_columns": true,
  "label_names": null,
  "load_best_model_at_end": true,
  "metric_for_best_model": "eval_wer",
  "greater_is_better": false,
  "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
  },
  "deepspeed": null,
  "label_smoothing_factor": 0.0,
  "optim": "adamw_torch",
  "optim_args": null,
  "adafactor": false,
  "group_by_length": true,
  "length_column_name": "length",
  "report_to": [],
  "ddp_find_unused_parameters": null,
  "ddp_bucket_cap_mb": null,
  "ddp_broadcast_buffers": null,
  "dataloader_pin_memory": true,
  "dataloader_persistent_workers": true,
  "skip_memory_metrics": true,
  "use_legacy_prediction_loop": false,
  "push_to_hub": false,
  "resume_from_checkpoint": false,
  "hub_model_id": null,
  "hub_strategy": "every_save",
  "hub_token": "<HUB_TOKEN>",
  "hub_private_repo": false,
  "hub_always_push": false,
  "gradient_checkpointing": false,
  "gradient_checkpointing_kwargs": null,
  "include_inputs_for_metrics": false,
  "eval_do_concat_batches": true,
  "fp16_backend": "auto",
  "evaluation_strategy": null,
  "push_to_hub_model_id": null,
  "push_to_hub_organization": null,
  "push_to_hub_token": "<PUSH_TO_HUB_TOKEN>",
  "mp_parameters": "",
  "auto_find_batch_size": false,
  "full_determinism": false,
  "torchdynamo": null,
  "ray_scope": "last",
  "ddp_timeout": 1800,
  "torch_compile": true,
  "torch_compile_backend": "inductor",
  "torch_compile_mode": "reduce-overhead",
  "dispatch_batches": null,
  "split_batches": 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,
  "sortish_sampler": false,
  "predict_with_generate": true,
  "generation_max_length": 128,
  "generation_num_beams": 1,
  "generation_config": null
}