Add panel_counter_v31_sft adapter (overnight v31 SFT run + backfill v29/v30 GRPO)
Browse files
sld_27.0/panel_counter_v31_sft/adapter_config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "/home/node/.cache/huggingface/hub/models--Qwen--Qwen3.5-9B/snapshots/c202236235762e1c871ad0ccb60c8ee5ba337b9a",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.05,
|
| 22 |
+
"megatron_config": null,
|
| 23 |
+
"megatron_core": "megatron.core",
|
| 24 |
+
"modules_to_save": [],
|
| 25 |
+
"peft_type": "LORA",
|
| 26 |
+
"peft_version": "0.18.1",
|
| 27 |
+
"qalora_group_size": 16,
|
| 28 |
+
"r": 64,
|
| 29 |
+
"rank_pattern": {},
|
| 30 |
+
"revision": null,
|
| 31 |
+
"target_modules": "^(model\\.language_model(?=\\.).*\\.(q_proj|v_proj|k_proj|in_proj_b|in_proj_qkv|out_proj|down_proj|gate_proj|up_proj|in_proj_a|in_proj_z|o_proj))$",
|
| 32 |
+
"target_parameters": null,
|
| 33 |
+
"task_type": "CAUSAL_LM",
|
| 34 |
+
"trainable_token_indices": null,
|
| 35 |
+
"use_dora": false,
|
| 36 |
+
"use_qalora": false,
|
| 37 |
+
"use_rslora": false
|
| 38 |
+
}
|
sld_27.0/panel_counter_v31_sft/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:160170ba2000b2b7fb03093798a04f132b78677e7d4959fff66ec6dc52fe12d9
|
| 3 |
+
size 692529024
|
sld_27.0/panel_counter_v31_sft/additional_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
|
sld_27.0/panel_counter_v31_sft/args.json
ADDED
|
@@ -0,0 +1,345 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"output_dir": "/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300",
|
| 3 |
+
"per_device_train_batch_size": 1,
|
| 4 |
+
"num_train_epochs": 4.0,
|
| 5 |
+
"max_steps": -1,
|
| 6 |
+
"learning_rate": 5e-05,
|
| 7 |
+
"lr_scheduler_type": "cosine",
|
| 8 |
+
"lr_scheduler_kwargs": null,
|
| 9 |
+
"warmup_steps": 0,
|
| 10 |
+
"optim": "adamw_torch_fused",
|
| 11 |
+
"optim_args": null,
|
| 12 |
+
"weight_decay": 0.1,
|
| 13 |
+
"adam_beta1": 0.9,
|
| 14 |
+
"adam_beta2": 0.95,
|
| 15 |
+
"adam_epsilon": 1e-08,
|
| 16 |
+
"optim_target_modules": null,
|
| 17 |
+
"gradient_accumulation_steps": 8,
|
| 18 |
+
"average_tokens_across_devices": true,
|
| 19 |
+
"max_grad_norm": 1.0,
|
| 20 |
+
"label_smoothing_factor": 0.0,
|
| 21 |
+
"bf16": true,
|
| 22 |
+
"fp16": false,
|
| 23 |
+
"bf16_full_eval": false,
|
| 24 |
+
"fp16_full_eval": false,
|
| 25 |
+
"tf32": null,
|
| 26 |
+
"gradient_checkpointing": true,
|
| 27 |
+
"gradient_checkpointing_kwargs": null,
|
| 28 |
+
"torch_compile": false,
|
| 29 |
+
"torch_compile_backend": null,
|
| 30 |
+
"torch_compile_mode": null,
|
| 31 |
+
"use_liger_kernel": false,
|
| 32 |
+
"liger_kernel_config": null,
|
| 33 |
+
"use_cache": false,
|
| 34 |
+
"neftune_noise_alpha": null,
|
| 35 |
+
"torch_empty_cache_steps": null,
|
| 36 |
+
"auto_find_batch_size": false,
|
| 37 |
+
"logging_strategy": "steps",
|
| 38 |
+
"logging_steps": 5,
|
| 39 |
+
"logging_first_step": true,
|
| 40 |
+
"log_on_each_node": true,
|
| 41 |
+
"logging_nan_inf_filter": true,
|
| 42 |
+
"include_num_input_tokens_seen": false,
|
| 43 |
+
"log_level": "passive",
|
| 44 |
+
"log_level_replica": "warning",
|
| 45 |
+
"disable_tqdm": null,
|
| 46 |
+
"report_to": [
|
| 47 |
+
"wandb"
|
| 48 |
+
],
|
| 49 |
+
"run_name": "/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300",
|
| 50 |
+
"project": "huggingface",
|
| 51 |
+
"trackio_space_id": "trackio",
|
| 52 |
+
"eval_strategy": "steps",
|
| 53 |
+
"eval_steps": 20.0,
|
| 54 |
+
"eval_delay": 0,
|
| 55 |
+
"per_device_eval_batch_size": 1,
|
| 56 |
+
"prediction_loss_only": false,
|
| 57 |
+
"eval_on_start": false,
|
| 58 |
+
"eval_do_concat_batches": true,
|
| 59 |
+
"eval_use_gather_object": false,
|
| 60 |
+
"eval_accumulation_steps": null,
|
| 61 |
+
"include_for_metrics": [],
|
| 62 |
+
"batch_eval_metrics": false,
|
| 63 |
+
"save_only_model": false,
|
| 64 |
+
"save_strategy": "steps",
|
| 65 |
+
"save_steps": 20.0,
|
| 66 |
+
"save_on_each_node": false,
|
| 67 |
+
"save_total_limit": 6,
|
| 68 |
+
"enable_jit_checkpoint": false,
|
| 69 |
+
"push_to_hub": false,
|
| 70 |
+
"hub_token": null,
|
| 71 |
+
"hub_private_repo": null,
|
| 72 |
+
"hub_model_id": null,
|
| 73 |
+
"hub_strategy": "every_save",
|
| 74 |
+
"hub_always_push": false,
|
| 75 |
+
"hub_revision": null,
|
| 76 |
+
"load_best_model_at_end": true,
|
| 77 |
+
"metric_for_best_model": "loss",
|
| 78 |
+
"greater_is_better": false,
|
| 79 |
+
"ignore_data_skip": false,
|
| 80 |
+
"restore_callback_states_from_checkpoint": false,
|
| 81 |
+
"full_determinism": false,
|
| 82 |
+
"seed": 42,
|
| 83 |
+
"data_seed": 42,
|
| 84 |
+
"use_cpu": false,
|
| 85 |
+
"accelerator_config": {
|
| 86 |
+
"dispatch_batches": false
|
| 87 |
+
},
|
| 88 |
+
"parallelism_config": null,
|
| 89 |
+
"dataloader_drop_last": false,
|
| 90 |
+
"dataloader_num_workers": 0,
|
| 91 |
+
"dataloader_pin_memory": true,
|
| 92 |
+
"dataloader_persistent_workers": false,
|
| 93 |
+
"dataloader_prefetch_factor": null,
|
| 94 |
+
"remove_unused_columns": true,
|
| 95 |
+
"label_names": null,
|
| 96 |
+
"train_sampling_strategy": "random",
|
| 97 |
+
"length_column_name": "length",
|
| 98 |
+
"ddp_find_unused_parameters": null,
|
| 99 |
+
"ddp_bucket_cap_mb": null,
|
| 100 |
+
"ddp_broadcast_buffers": null,
|
| 101 |
+
"ddp_backend": null,
|
| 102 |
+
"ddp_timeout": 18000000,
|
| 103 |
+
"fsdp": [],
|
| 104 |
+
"fsdp_config": null,
|
| 105 |
+
"deepspeed": null,
|
| 106 |
+
"debug": null,
|
| 107 |
+
"skip_memory_metrics": true,
|
| 108 |
+
"do_train": false,
|
| 109 |
+
"do_eval": false,
|
| 110 |
+
"do_predict": false,
|
| 111 |
+
"resume_from_checkpoint": null,
|
| 112 |
+
"warmup_ratio": 0.05,
|
| 113 |
+
"logging_dir": "/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300/runs",
|
| 114 |
+
"local_rank": -1,
|
| 115 |
+
"sortish_sampler": false,
|
| 116 |
+
"predict_with_generate": false,
|
| 117 |
+
"generation_max_length": null,
|
| 118 |
+
"generation_num_beams": null,
|
| 119 |
+
"generation_config": null,
|
| 120 |
+
"tuner_backend": "peft",
|
| 121 |
+
"vit_gradient_checkpointing": null,
|
| 122 |
+
"router_aux_loss_coef": 0.0,
|
| 123 |
+
"enable_dft_loss": false,
|
| 124 |
+
"enable_channel_loss": false,
|
| 125 |
+
"safe_serialization": true,
|
| 126 |
+
"max_shard_size": "5GB",
|
| 127 |
+
"check_model": true,
|
| 128 |
+
"acc_strategy": "token",
|
| 129 |
+
"train_dataloader_shuffle": true,
|
| 130 |
+
"group_by_length": false,
|
| 131 |
+
"max_epochs": null,
|
| 132 |
+
"aligner_lr": null,
|
| 133 |
+
"vit_lr": null,
|
| 134 |
+
"use_logits_to_keep": null,
|
| 135 |
+
"ds3_gather_for_generation": true,
|
| 136 |
+
"resume_only_model": true,
|
| 137 |
+
"optimizer": null,
|
| 138 |
+
"loss_type": null,
|
| 139 |
+
"eval_metric": null,
|
| 140 |
+
"callbacks": [],
|
| 141 |
+
"early_stop_interval": 3,
|
| 142 |
+
"eval_use_evalscope": false,
|
| 143 |
+
"eval_dataset": [],
|
| 144 |
+
"eval_dataset_args": null,
|
| 145 |
+
"eval_limit": null,
|
| 146 |
+
"eval_generation_config": null,
|
| 147 |
+
"extra_eval_args": null,
|
| 148 |
+
"tuner_type": "lora",
|
| 149 |
+
"use_galore": false,
|
| 150 |
+
"galore_target_modules": null,
|
| 151 |
+
"galore_rank": 128,
|
| 152 |
+
"galore_update_proj_gap": 50,
|
| 153 |
+
"galore_scale": 1.0,
|
| 154 |
+
"galore_proj_type": "std",
|
| 155 |
+
"galore_optim_per_parameter": false,
|
| 156 |
+
"galore_with_embedding": false,
|
| 157 |
+
"galore_quantization": false,
|
| 158 |
+
"galore_proj_quant": false,
|
| 159 |
+
"galore_proj_bits": 4,
|
| 160 |
+
"galore_proj_group_size": 256,
|
| 161 |
+
"galore_cos_threshold": 0.4,
|
| 162 |
+
"galore_gamma_proj": 2,
|
| 163 |
+
"galore_queue_size": 5,
|
| 164 |
+
"lisa_activated_layers": 0,
|
| 165 |
+
"lisa_step_interval": 20,
|
| 166 |
+
"use_flash_ckpt": false,
|
| 167 |
+
"use_ray": false,
|
| 168 |
+
"ray_exp_name": null,
|
| 169 |
+
"device_groups": null,
|
| 170 |
+
"model": "/home/node/.cache/huggingface/hub/models--Qwen--Qwen3.5-9B/snapshots/c202236235762e1c871ad0ccb60c8ee5ba337b9a",
|
| 171 |
+
"model_type": "qwen3_5",
|
| 172 |
+
"model_revision": null,
|
| 173 |
+
"task_type": "causal_lm",
|
| 174 |
+
"torch_dtype": "bfloat16",
|
| 175 |
+
"attn_impl": null,
|
| 176 |
+
"experts_impl": null,
|
| 177 |
+
"new_special_tokens": [],
|
| 178 |
+
"num_labels": null,
|
| 179 |
+
"problem_type": null,
|
| 180 |
+
"rope_scaling": null,
|
| 181 |
+
"device_map": null,
|
| 182 |
+
"max_memory": {},
|
| 183 |
+
"max_model_len": null,
|
| 184 |
+
"local_repo_path": null,
|
| 185 |
+
"init_strategy": null,
|
| 186 |
+
"template": "qwen3_5",
|
| 187 |
+
"system": null,
|
| 188 |
+
"max_length": 8192,
|
| 189 |
+
"truncation_strategy": "delete",
|
| 190 |
+
"max_pixels": null,
|
| 191 |
+
"agent_template": null,
|
| 192 |
+
"norm_bbox": null,
|
| 193 |
+
"use_chat_template": true,
|
| 194 |
+
"padding_side": "right",
|
| 195 |
+
"padding_free": false,
|
| 196 |
+
"loss_scale": "default",
|
| 197 |
+
"sequence_parallel_size": 1,
|
| 198 |
+
"template_backend": "swift",
|
| 199 |
+
"response_prefix": null,
|
| 200 |
+
"enable_thinking": false,
|
| 201 |
+
"add_non_thinking_prefix": true,
|
| 202 |
+
"dataset": [
|
| 203 |
+
"/workspace/training/sft_v31_train.jsonl"
|
| 204 |
+
],
|
| 205 |
+
"val_dataset": [
|
| 206 |
+
"/workspace/training/sft_v31_eval_gtgold.jsonl"
|
| 207 |
+
],
|
| 208 |
+
"cached_dataset": [],
|
| 209 |
+
"cached_val_dataset": [],
|
| 210 |
+
"split_dataset_ratio": 0.0,
|
| 211 |
+
"dataset_num_proc": 1,
|
| 212 |
+
"load_from_cache_file": false,
|
| 213 |
+
"dataset_shuffle": true,
|
| 214 |
+
"val_dataset_shuffle": false,
|
| 215 |
+
"streaming": false,
|
| 216 |
+
"interleave_prob": null,
|
| 217 |
+
"stopping_strategy": "first_exhausted",
|
| 218 |
+
"shuffle_buffer_size": 1000,
|
| 219 |
+
"download_mode": "reuse_dataset_if_exists",
|
| 220 |
+
"columns": {},
|
| 221 |
+
"strict": false,
|
| 222 |
+
"model_name": null,
|
| 223 |
+
"model_author": null,
|
| 224 |
+
"custom_dataset_info": [],
|
| 225 |
+
"quant_method": null,
|
| 226 |
+
"quant_bits": null,
|
| 227 |
+
"hqq_axis": null,
|
| 228 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
| 229 |
+
"bnb_4bit_quant_type": "nf4",
|
| 230 |
+
"bnb_4bit_use_double_quant": true,
|
| 231 |
+
"bnb_4bit_quant_storage": null,
|
| 232 |
+
"max_new_tokens": 64,
|
| 233 |
+
"temperature": 0.0,
|
| 234 |
+
"top_k": null,
|
| 235 |
+
"top_p": null,
|
| 236 |
+
"repetition_penalty": null,
|
| 237 |
+
"num_beams": 1,
|
| 238 |
+
"stream": false,
|
| 239 |
+
"stop_words": [],
|
| 240 |
+
"logprobs": false,
|
| 241 |
+
"top_logprobs": null,
|
| 242 |
+
"structured_outputs_regex": null,
|
| 243 |
+
"adapters": [
|
| 244 |
+
"/workspace/training/checkpoints/panel_counter_v28_sft1/v0-20260723-102705/checkpoint-220"
|
| 245 |
+
],
|
| 246 |
+
"external_plugins": [],
|
| 247 |
+
"custom_register_path": [],
|
| 248 |
+
"model_kwargs": {},
|
| 249 |
+
"load_args": false,
|
| 250 |
+
"load_data_args": false,
|
| 251 |
+
"packing": false,
|
| 252 |
+
"packing_length": null,
|
| 253 |
+
"packing_num_proc": 1,
|
| 254 |
+
"lazy_tokenize": true,
|
| 255 |
+
"use_hf": true,
|
| 256 |
+
"ignore_args_error": false,
|
| 257 |
+
"use_swift_lora": false,
|
| 258 |
+
"freeze_parameters": [],
|
| 259 |
+
"freeze_parameters_regex": null,
|
| 260 |
+
"freeze_parameters_ratio": 0.0,
|
| 261 |
+
"trainable_parameters": [],
|
| 262 |
+
"trainable_parameters_regex": null,
|
| 263 |
+
"freeze_llm": false,
|
| 264 |
+
"freeze_vit": true,
|
| 265 |
+
"freeze_aligner": true,
|
| 266 |
+
"target_modules": [
|
| 267 |
+
"all-linear"
|
| 268 |
+
],
|
| 269 |
+
"target_regex": null,
|
| 270 |
+
"target_parameters": null,
|
| 271 |
+
"modules_to_save": [],
|
| 272 |
+
"lora_rank": 64,
|
| 273 |
+
"lora_alpha": 128,
|
| 274 |
+
"lora_dropout": 0.05,
|
| 275 |
+
"lora_bias": "none",
|
| 276 |
+
"lora_dtype": null,
|
| 277 |
+
"lorap_lr_ratio": null,
|
| 278 |
+
"use_rslora": false,
|
| 279 |
+
"use_dora": false,
|
| 280 |
+
"lora_ga_batch_size": 2,
|
| 281 |
+
"lora_ga_iters": 2,
|
| 282 |
+
"lora_ga_max_length": 1024,
|
| 283 |
+
"lora_ga_direction": "ArB2r",
|
| 284 |
+
"lora_ga_scale": "stable",
|
| 285 |
+
"lora_ga_stable_gamma": 16,
|
| 286 |
+
"init_weights": true,
|
| 287 |
+
"fourier_n_frequency": 2000,
|
| 288 |
+
"fourier_scaling": 300.0,
|
| 289 |
+
"boft_block_size": 4,
|
| 290 |
+
"boft_block_num": 0,
|
| 291 |
+
"boft_n_butterfly_factor": 1,
|
| 292 |
+
"boft_dropout": 0.0,
|
| 293 |
+
"vera_rank": 256,
|
| 294 |
+
"vera_projection_prng_key": 0,
|
| 295 |
+
"vera_dropout": 0.0,
|
| 296 |
+
"vera_d_initial": 0.1,
|
| 297 |
+
"adapter_act": "gelu",
|
| 298 |
+
"adapter_length": 128,
|
| 299 |
+
"adalora_target_r": 8,
|
| 300 |
+
"adalora_init_r": 12,
|
| 301 |
+
"adalora_tinit": 0,
|
| 302 |
+
"adalora_tfinal": 0,
|
| 303 |
+
"adalora_deltaT": 1,
|
| 304 |
+
"adalora_beta1": 0.85,
|
| 305 |
+
"adalora_beta2": 0.85,
|
| 306 |
+
"adalora_orth_reg_weight": 0.5,
|
| 307 |
+
"llamapro_num_new_blocks": 4,
|
| 308 |
+
"llamapro_num_groups": null,
|
| 309 |
+
"reft_layer_key": null,
|
| 310 |
+
"reft_layers": null,
|
| 311 |
+
"reft_rank": 4,
|
| 312 |
+
"reft_intervention_type": "LoreftIntervention",
|
| 313 |
+
"reft_args": null,
|
| 314 |
+
"swanlab_token": null,
|
| 315 |
+
"swanlab_project": "ms-swift",
|
| 316 |
+
"swanlab_workspace": null,
|
| 317 |
+
"swanlab_exp_name": null,
|
| 318 |
+
"swanlab_notification_method": null,
|
| 319 |
+
"swanlab_webhook_url": null,
|
| 320 |
+
"swanlab_secret": null,
|
| 321 |
+
"swanlab_sender_email": null,
|
| 322 |
+
"swanlab_receiver_email": null,
|
| 323 |
+
"swanlab_smtp_server": null,
|
| 324 |
+
"swanlab_smtp_port": null,
|
| 325 |
+
"swanlab_email_language": "zh",
|
| 326 |
+
"swanlab_mode": "cloud",
|
| 327 |
+
"add_version": true,
|
| 328 |
+
"create_checkpoint_symlink": false,
|
| 329 |
+
"zero_hpz_partition_size": null,
|
| 330 |
+
"deepspeed_autotp_size": null,
|
| 331 |
+
"swift_version": "4.1.0",
|
| 332 |
+
"ckpt_dir": "/workspace/training/checkpoints/panel_counter_v28_sft1/v0-20260723-102705/checkpoint-220",
|
| 333 |
+
"rank": -1,
|
| 334 |
+
"global_world_size": 1,
|
| 335 |
+
"local_world_size": 1,
|
| 336 |
+
"model_suffix": "Qwen3.5-9B",
|
| 337 |
+
"model_info": "ModelInfo(model_type='qwen3_5', model_dir='/home/node/.cache/huggingface/hub/models--Qwen--Qwen3.5-9B/snapshots/c202236235762e1c871ad0ccb60c8ee5ba337b9a', 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)",
|
| 338 |
+
"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=[])], loader=<class 'swift.model.models.qwen.Qwen3_5Loader'>, 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=[])",
|
| 339 |
+
"model_dir": "/home/node/.cache/huggingface/hub/models--Qwen--Qwen3.5-9B/snapshots/c202236235762e1c871ad0ccb60c8ee5ba337b9a",
|
| 340 |
+
"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=<class 'swift.template.templates.qwen.Qwen3_5Template'>, 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='<think>\\n', non_thinking_prefix='<think>\\n\\n</think>\\n\\n', history_thinking_prefix='')",
|
| 341 |
+
"_val_dataset_exists": true,
|
| 342 |
+
"hub": "<class 'swift.hub.hub.HFHub'>",
|
| 343 |
+
"evaluation_strategy": "steps",
|
| 344 |
+
"training_args": "Seq2SeqTrainingArguments(output_dir='/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300', per_device_train_batch_size=1, num_train_epochs=4.0, max_steps=-1, learning_rate=5e-05, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_steps=0.05, optim=<OptimizerNames.ADAMW_TORCH_FUSED: 'adamw_torch_fused'>, optim_args=None, weight_decay=0.1, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, optim_target_modules=None, gradient_accumulation_steps=8, 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=None, 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=<IntervalStrategy.STEPS: 'steps'>, logging_steps=5, 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='/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300', project='huggingface', trackio_space_id='trackio', eval_strategy=<IntervalStrategy.STEPS: 'steps'>, eval_steps=20, 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=<SaveStrategy.STEPS: 'steps'>, save_steps=20, save_on_each_node=False, save_total_limit=6, enable_jit_checkpoint=False, push_to_hub=False, hub_token=None, hub_private_repo=None, hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_always_push=False, hub_revision=None, load_best_model_at_end=True, 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=0, dataloader_pin_memory=True, dataloader_persistent_workers=False, dataloader_prefetch_factor=None, 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_backend=None, ddp_timeout=18000000, fsdp=[], fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, deepspeed=None, debug=[], skip_memory_metrics=True, do_train=False, do_eval=True, do_predict=False, resume_from_checkpoint=None, warmup_ratio=0.05, logging_dir='/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300/runs', local_rank=-1, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, 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=True, optimizer=None, loss_type=None, eval_metric=None, callbacks=['early_stop'], early_stop_interval=3, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, tuner_type='lora', 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)"
|
| 345 |
+
}
|
sld_27.0/panel_counter_v31_sft/trainer_state.json
ADDED
|
@@ -0,0 +1,411 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 180,
|
| 3 |
+
"best_metric": 0.04850775,
|
| 4 |
+
"best_model_checkpoint": "/workspace/training/checkpoints/panel_counter_v31_sft/v0-20260730-150300/checkpoint-180",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 20,
|
| 7 |
+
"global_step": 180,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.016842105263157894,
|
| 14 |
+
"grad_norm": 2.9390158653259277,
|
| 15 |
+
"learning_rate": 4.166666666666667e-06,
|
| 16 |
+
"loss": 0.1470683515071869,
|
| 17 |
+
"step": 1,
|
| 18 |
+
"token_acc": 0.9675858732462506
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 0.08421052631578947,
|
| 22 |
+
"grad_norm": 0.5498834252357483,
|
| 23 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 24 |
+
"loss": 0.07648029923439026,
|
| 25 |
+
"step": 5,
|
| 26 |
+
"token_acc": 0.9819692914495
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.16842105263157894,
|
| 30 |
+
"grad_norm": 0.3654409945011139,
|
| 31 |
+
"learning_rate": 4.166666666666667e-05,
|
| 32 |
+
"loss": 0.043979352712631224,
|
| 33 |
+
"step": 10,
|
| 34 |
+
"token_acc": 0.9882140709561035
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.25263157894736843,
|
| 38 |
+
"grad_norm": 0.32861626148223877,
|
| 39 |
+
"learning_rate": 4.997864395968252e-05,
|
| 40 |
+
"loss": 0.04751304090023041,
|
| 41 |
+
"step": 15,
|
| 42 |
+
"token_acc": 0.9848148969186433
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.3368421052631579,
|
| 46 |
+
"grad_norm": 0.2843306362628937,
|
| 47 |
+
"learning_rate": 4.984826693294874e-05,
|
| 48 |
+
"loss": 0.03860023021697998,
|
| 49 |
+
"step": 20,
|
| 50 |
+
"token_acc": 0.9873782108060231
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.3368421052631579,
|
| 54 |
+
"eval_loss": 0.07016084343194962,
|
| 55 |
+
"eval_runtime": 50.871,
|
| 56 |
+
"eval_samples_per_second": 1.003,
|
| 57 |
+
"eval_steps_per_second": 1.003,
|
| 58 |
+
"eval_token_acc": 0.9835867311679336,
|
| 59 |
+
"step": 20
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.42105263157894735,
|
| 63 |
+
"grad_norm": 0.5600948929786682,
|
| 64 |
+
"learning_rate": 4.9599995180367e-05,
|
| 65 |
+
"loss": 0.04247380793094635,
|
| 66 |
+
"step": 25,
|
| 67 |
+
"token_acc": 0.9861232517482518
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.5052631578947369,
|
| 71 |
+
"grad_norm": 0.3054753541946411,
|
| 72 |
+
"learning_rate": 4.923500664848326e-05,
|
| 73 |
+
"loss": 0.02706400156021118,
|
| 74 |
+
"step": 30,
|
| 75 |
+
"token_acc": 0.9913188647746244
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.5894736842105263,
|
| 79 |
+
"grad_norm": 0.26073768734931946,
|
| 80 |
+
"learning_rate": 4.875503305657678e-05,
|
| 81 |
+
"loss": 0.03642970323562622,
|
| 82 |
+
"step": 35,
|
| 83 |
+
"token_acc": 0.9892134831460674
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.6736842105263158,
|
| 87 |
+
"grad_norm": 0.406607985496521,
|
| 88 |
+
"learning_rate": 4.8162351680370044e-05,
|
| 89 |
+
"loss": 0.03609724640846253,
|
| 90 |
+
"step": 40,
|
| 91 |
+
"token_acc": 0.9879557648089347
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.6736842105263158,
|
| 95 |
+
"eval_loss": 0.06280472129583359,
|
| 96 |
+
"eval_runtime": 51.0357,
|
| 97 |
+
"eval_samples_per_second": 0.999,
|
| 98 |
+
"eval_steps_per_second": 0.999,
|
| 99 |
+
"eval_token_acc": 0.9834139599170698,
|
| 100 |
+
"step": 40
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.7578947368421053,
|
| 104 |
+
"grad_norm": 0.32828518748283386,
|
| 105 |
+
"learning_rate": 4.7459774547299475e-05,
|
| 106 |
+
"loss": 0.0399460643529892,
|
| 107 |
+
"step": 45,
|
| 108 |
+
"token_acc": 0.9874332740213523
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.8421052631578947,
|
| 112 |
+
"grad_norm": 0.33558136224746704,
|
| 113 |
+
"learning_rate": 4.665063509461097e-05,
|
| 114 |
+
"loss": 0.033811500668525694,
|
| 115 |
+
"step": 50,
|
| 116 |
+
"token_acc": 0.9881335255628257
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.9263157894736842,
|
| 120 |
+
"grad_norm": 0.38677728176116943,
|
| 121 |
+
"learning_rate": 4.573877235358204e-05,
|
| 122 |
+
"loss": 0.05068126320838928,
|
| 123 |
+
"step": 55,
|
| 124 |
+
"token_acc": 0.9847748623258827
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 1.0,
|
| 128 |
+
"grad_norm": 0.4346539378166199,
|
| 129 |
+
"learning_rate": 4.4728512734909844e-05,
|
| 130 |
+
"loss": 0.04947632551193237,
|
| 131 |
+
"step": 60,
|
| 132 |
+
"token_acc": 0.9838602697324784
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.0,
|
| 136 |
+
"eval_loss": 0.05832681432366371,
|
| 137 |
+
"eval_runtime": 50.8239,
|
| 138 |
+
"eval_samples_per_second": 1.003,
|
| 139 |
+
"eval_steps_per_second": 1.003,
|
| 140 |
+
"eval_token_acc": 0.9820317899101589,
|
| 141 |
+
"step": 60
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 1.0842105263157895,
|
| 145 |
+
"grad_norm": 0.22376033663749695,
|
| 146 |
+
"learning_rate": 4.362464950168593e-05,
|
| 147 |
+
"loss": 0.032614442706108096,
|
| 148 |
+
"step": 65,
|
| 149 |
+
"token_acc": 0.9900465216920913
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.168421052631579,
|
| 153 |
+
"grad_norm": 0.2849695086479187,
|
| 154 |
+
"learning_rate": 4.2432420027349886e-05,
|
| 155 |
+
"loss": 0.019579203426837923,
|
| 156 |
+
"step": 70,
|
| 157 |
+
"token_acc": 0.9938848120969536
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.2526315789473683,
|
| 161 |
+
"grad_norm": 0.26889654994010925,
|
| 162 |
+
"learning_rate": 4.115748094652352e-05,
|
| 163 |
+
"loss": 0.02692590057849884,
|
| 164 |
+
"step": 75,
|
| 165 |
+
"token_acc": 0.9910524549826641
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.3368421052631578,
|
| 169 |
+
"grad_norm": 0.29549166560173035,
|
| 170 |
+
"learning_rate": 3.9805881316624506e-05,
|
| 171 |
+
"loss": 0.031128931045532226,
|
| 172 |
+
"step": 80,
|
| 173 |
+
"token_acc": 0.9895833333333334
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.3368421052631578,
|
| 177 |
+
"eval_loss": 0.0633283331990242,
|
| 178 |
+
"eval_runtime": 51.2216,
|
| 179 |
+
"eval_samples_per_second": 0.996,
|
| 180 |
+
"eval_steps_per_second": 0.996,
|
| 181 |
+
"eval_token_acc": 0.9844505874222529,
|
| 182 |
+
"step": 80
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.4210526315789473,
|
| 186 |
+
"grad_norm": 0.19602809846401215,
|
| 187 |
+
"learning_rate": 3.838403391759652e-05,
|
| 188 |
+
"loss": 0.02242555171251297,
|
| 189 |
+
"step": 85,
|
| 190 |
+
"token_acc": 0.9921250984362695
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 1.5052631578947369,
|
| 194 |
+
"grad_norm": 0.32295525074005127,
|
| 195 |
+
"learning_rate": 3.689868482592684e-05,
|
| 196 |
+
"loss": 0.024873176217079164,
|
| 197 |
+
"step": 90,
|
| 198 |
+
"token_acc": 0.9914273910349102
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.5894736842105264,
|
| 202 |
+
"grad_norm": 0.29888179898262024,
|
| 203 |
+
"learning_rate": 3.535688140730997e-05,
|
| 204 |
+
"loss": 0.015549980103969574,
|
| 205 |
+
"step": 95,
|
| 206 |
+
"token_acc": 0.9938564970441637
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.6736842105263157,
|
| 210 |
+
"grad_norm": 0.2777520418167114,
|
| 211 |
+
"learning_rate": 3.376593887981887e-05,
|
| 212 |
+
"loss": 0.024531815946102143,
|
| 213 |
+
"step": 100,
|
| 214 |
+
"token_acc": 0.9916434540389972
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 1.6736842105263157,
|
| 218 |
+
"eval_loss": 0.060586102306842804,
|
| 219 |
+
"eval_runtime": 51.0741,
|
| 220 |
+
"eval_samples_per_second": 0.999,
|
| 221 |
+
"eval_steps_per_second": 0.999,
|
| 222 |
+
"eval_token_acc": 0.9850552868002764,
|
| 223 |
+
"step": 100
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 1.7578947368421054,
|
| 227 |
+
"grad_norm": 0.22772729396820068,
|
| 228 |
+
"learning_rate": 3.213340560622763e-05,
|
| 229 |
+
"loss": 0.024645522236824036,
|
| 230 |
+
"step": 105,
|
| 231 |
+
"token_acc": 0.9915684790833423
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.8421052631578947,
|
| 235 |
+
"grad_norm": 0.2930346429347992,
|
| 236 |
+
"learning_rate": 3.04670272801594e-05,
|
| 237 |
+
"loss": 0.015664128959178923,
|
| 238 |
+
"step": 110,
|
| 239 |
+
"token_acc": 0.994575445588398
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.9263157894736842,
|
| 243 |
+
"grad_norm": 0.25863951444625854,
|
| 244 |
+
"learning_rate": 2.8774710175980922e-05,
|
| 245 |
+
"loss": 0.033259269595146176,
|
| 246 |
+
"step": 115,
|
| 247 |
+
"token_acc": 0.9893274041133964
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 2.0,
|
| 251 |
+
"grad_norm": 0.2696121037006378,
|
| 252 |
+
"learning_rate": 2.7064483636808313e-05,
|
| 253 |
+
"loss": 0.01593230962753296,
|
| 254 |
+
"step": 120,
|
| 255 |
+
"token_acc": 0.9951232728257925
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.0,
|
| 259 |
+
"eval_loss": 0.05449597164988518,
|
| 260 |
+
"eval_runtime": 50.6932,
|
| 261 |
+
"eval_samples_per_second": 1.006,
|
| 262 |
+
"eval_steps_per_second": 1.006,
|
| 263 |
+
"eval_token_acc": 0.9849689011748445,
|
| 264 |
+
"step": 120
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 2.0842105263157893,
|
| 268 |
+
"grad_norm": 0.1409084051847458,
|
| 269 |
+
"learning_rate": 2.5344461978602413e-05,
|
| 270 |
+
"loss": 0.016947227716445922,
|
| 271 |
+
"step": 125,
|
| 272 |
+
"token_acc": 0.9953260869565217
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.168421052631579,
|
| 276 |
+
"grad_norm": 0.27552589774131775,
|
| 277 |
+
"learning_rate": 2.3622805991103362e-05,
|
| 278 |
+
"loss": 0.016483573615550993,
|
| 279 |
+
"step": 130,
|
| 280 |
+
"token_acc": 0.993715545755237
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 2.2526315789473683,
|
| 284 |
+
"grad_norm": 0.23056618869304657,
|
| 285 |
+
"learning_rate": 2.190768421826631e-05,
|
| 286 |
+
"loss": 0.02009890228509903,
|
| 287 |
+
"step": 135,
|
| 288 |
+
"token_acc": 0.9935085665637969
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.336842105263158,
|
| 292 |
+
"grad_norm": 0.25431764125823975,
|
| 293 |
+
"learning_rate": 2.0207234201906547e-05,
|
| 294 |
+
"loss": 0.017615416646003725,
|
| 295 |
+
"step": 140,
|
| 296 |
+
"token_acc": 0.9939317609342798
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.336842105263158,
|
| 300 |
+
"eval_loss": 0.050888072699308395,
|
| 301 |
+
"eval_runtime": 50.4164,
|
| 302 |
+
"eval_samples_per_second": 1.012,
|
| 303 |
+
"eval_steps_per_second": 1.012,
|
| 304 |
+
"eval_token_acc": 0.9861782999308915,
|
| 305 |
+
"step": 140
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.4210526315789473,
|
| 309 |
+
"grad_norm": 0.25543397665023804,
|
| 310 |
+
"learning_rate": 1.852952387243698e-05,
|
| 311 |
+
"loss": 0.014513900876045227,
|
| 312 |
+
"step": 145,
|
| 313 |
+
"token_acc": 0.9942410530645824
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.5052631578947366,
|
| 317 |
+
"grad_norm": 0.3040763735771179,
|
| 318 |
+
"learning_rate": 1.6882513269882917e-05,
|
| 319 |
+
"loss": 0.015607929229736328,
|
| 320 |
+
"step": 150,
|
| 321 |
+
"token_acc": 0.9949232976492661
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.5894736842105264,
|
| 325 |
+
"grad_norm": 0.3951432704925537,
|
| 326 |
+
"learning_rate": 1.5274016776791853e-05,
|
| 327 |
+
"loss": 0.010614027827978134,
|
| 328 |
+
"step": 155,
|
| 329 |
+
"token_acc": 0.9960104867206201
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 2.6736842105263157,
|
| 333 |
+
"grad_norm": 0.19927626848220825,
|
| 334 |
+
"learning_rate": 1.3711666042227772e-05,
|
| 335 |
+
"loss": 0.013232138752937318,
|
| 336 |
+
"step": 160,
|
| 337 |
+
"token_acc": 0.9953752364935884
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 2.6736842105263157,
|
| 341 |
+
"eval_loss": 0.05326460301876068,
|
| 342 |
+
"eval_runtime": 50.3899,
|
| 343 |
+
"eval_samples_per_second": 1.012,
|
| 344 |
+
"eval_steps_per_second": 1.012,
|
| 345 |
+
"eval_token_acc": 0.9849689011748445,
|
| 346 |
+
"step": 160
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 2.7578947368421054,
|
| 350 |
+
"grad_norm": 0.31109318137168884,
|
| 351 |
+
"learning_rate": 1.2202873772759981e-05,
|
| 352 |
+
"loss": 0.017671605944633482,
|
| 353 |
+
"step": 165,
|
| 354 |
+
"token_acc": 0.9942870201096892
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.8421052631578947,
|
| 358 |
+
"grad_norm": 0.24973711371421814,
|
| 359 |
+
"learning_rate": 1.0754798562243345e-05,
|
| 360 |
+
"loss": 0.013447245955467224,
|
| 361 |
+
"step": 170,
|
| 362 |
+
"token_acc": 0.9953179442508711
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.9263157894736844,
|
| 366 |
+
"grad_norm": 0.27594366669654846,
|
| 367 |
+
"learning_rate": 9.37431092725804e-06,
|
| 368 |
+
"loss": 0.014141039550304412,
|
| 369 |
+
"step": 175,
|
| 370 |
+
"token_acc": 0.9948429442100328
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 3.0,
|
| 374 |
+
"grad_norm": 0.4812513291835785,
|
| 375 |
+
"learning_rate": 8.067960709356478e-06,
|
| 376 |
+
"loss": 0.014215387403964996,
|
| 377 |
+
"step": 180,
|
| 378 |
+
"token_acc": 0.9952970297029703
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 3.0,
|
| 382 |
+
"eval_loss": 0.04850774630904198,
|
| 383 |
+
"eval_runtime": 50.3763,
|
| 384 |
+
"eval_samples_per_second": 1.012,
|
| 385 |
+
"eval_steps_per_second": 1.012,
|
| 386 |
+
"eval_token_acc": 0.9860055286800277,
|
| 387 |
+
"step": 180
|
| 388 |
+
}
|
| 389 |
+
],
|
| 390 |
+
"logging_steps": 5,
|
| 391 |
+
"max_steps": 240,
|
| 392 |
+
"num_input_tokens_seen": 0,
|
| 393 |
+
"num_train_epochs": 4,
|
| 394 |
+
"save_steps": 20,
|
| 395 |
+
"stateful_callbacks": {
|
| 396 |
+
"TrainerControl": {
|
| 397 |
+
"args": {
|
| 398 |
+
"should_epoch_stop": false,
|
| 399 |
+
"should_evaluate": false,
|
| 400 |
+
"should_log": false,
|
| 401 |
+
"should_save": true,
|
| 402 |
+
"should_training_stop": false
|
| 403 |
+
},
|
| 404 |
+
"attributes": {}
|
| 405 |
+
}
|
| 406 |
+
},
|
| 407 |
+
"total_flos": 1.8724519998744912e+17,
|
| 408 |
+
"train_batch_size": 1,
|
| 409 |
+
"trial_name": null,
|
| 410 |
+
"trial_params": null
|
| 411 |
+
}
|