Upload sdxl_my_hires.json
Browse files- sdxl_my_hires.json +430 -0
sdxl_my_hires.json
ADDED
|
@@ -0,0 +1,430 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version": 4,
|
| 3 |
+
"training_method": "FINE_TUNE",
|
| 4 |
+
"model_type": "STABLE_DIFFUSION_XL_10_BASE",
|
| 5 |
+
"debug_mode": false,
|
| 6 |
+
"debug_dir": "C:/Users/Administrator/Desktop/sdxl_out/debug",
|
| 7 |
+
"workspace_dir": "C:/Users/Administrator/Desktop/sdxl_out/work",
|
| 8 |
+
"cache_dir": "C:/Users/Administrator/Desktop/sdxl_out/cache",
|
| 9 |
+
"tensorboard": false,
|
| 10 |
+
"tensorboard_expose": false,
|
| 11 |
+
"continue_last_backup": true,
|
| 12 |
+
"include_train_config": "NONE",
|
| 13 |
+
"base_model_name": "C:\\Users\\Administrator\\Desktop\\sdxl_out\\model\\diff\\model",
|
| 14 |
+
"weight_dtype": "BFLOAT_16",
|
| 15 |
+
"output_dtype": "BFLOAT_16",
|
| 16 |
+
"output_model_format": "SAFETENSORS",
|
| 17 |
+
"output_model_destination": "C:/Users/Administrator/Desktop/sdxl_out/work/out_moel/model",
|
| 18 |
+
"gradient_checkpointing": true,
|
| 19 |
+
"force_circular_padding": true,
|
| 20 |
+
"concept_file_name": "training_concepts/concepts.json",
|
| 21 |
+
"concepts": null,
|
| 22 |
+
"aspect_ratio_bucketing": true,
|
| 23 |
+
"latent_caching": false,
|
| 24 |
+
"clear_cache_before_training": true,
|
| 25 |
+
"learning_rate_scheduler": "CONSTANT",
|
| 26 |
+
"custom_learning_rate_scheduler": null,
|
| 27 |
+
"scheduler_params": [],
|
| 28 |
+
"learning_rate": 7.5e-06,
|
| 29 |
+
"learning_rate_warmup_steps": 0,
|
| 30 |
+
"learning_rate_cycles": 1,
|
| 31 |
+
"epochs": 30,
|
| 32 |
+
"batch_size": 8,
|
| 33 |
+
"gradient_accumulation_steps": 1,
|
| 34 |
+
"ema": "OFF",
|
| 35 |
+
"ema_decay": 0.999,
|
| 36 |
+
"ema_update_step_interval": 500,
|
| 37 |
+
"dataloader_threads": 2,
|
| 38 |
+
"train_device": "cuda",
|
| 39 |
+
"temp_device": "cpu",
|
| 40 |
+
"train_dtype": "BFLOAT_16",
|
| 41 |
+
"fallback_train_dtype": "BFLOAT_16",
|
| 42 |
+
"enable_autocast_cache": true,
|
| 43 |
+
"only_cache": false,
|
| 44 |
+
"resolution": "1280",
|
| 45 |
+
"attention_mechanism": "SDP",
|
| 46 |
+
"align_prop": false,
|
| 47 |
+
"align_prop_probability": 0.1,
|
| 48 |
+
"align_prop_loss": "HPS",
|
| 49 |
+
"align_prop_weight": 0.01,
|
| 50 |
+
"align_prop_steps": 20,
|
| 51 |
+
"align_prop_truncate_steps": 0.5,
|
| 52 |
+
"align_prop_cfg_scale": 6.0,
|
| 53 |
+
"mse_strength": 1.0,
|
| 54 |
+
"mae_strength": 0.0,
|
| 55 |
+
"vb_loss_strength": 1.0,
|
| 56 |
+
"loss_weight_fn": "MIN_SNR_GAMMA",
|
| 57 |
+
"loss_weight_strength": 5.0,
|
| 58 |
+
"dropout_probability": 0.0,
|
| 59 |
+
"loss_scaler": "NONE",
|
| 60 |
+
"learning_rate_scaler": "NONE",
|
| 61 |
+
"offset_noise_weight": 0.0357,
|
| 62 |
+
"perturbation_noise_weight": 0.00357,
|
| 63 |
+
"rescale_noise_scheduler_to_zero_terminal_snr": true,
|
| 64 |
+
"force_v_prediction": true,
|
| 65 |
+
"force_epsilon_prediction": false,
|
| 66 |
+
"min_noising_strength": 0.0,
|
| 67 |
+
"max_noising_strength": 1.0,
|
| 68 |
+
"timestep_distribution": "UNIFORM",
|
| 69 |
+
"noising_weight": 0.0,
|
| 70 |
+
"noising_bias": 0.0,
|
| 71 |
+
"unet": {
|
| 72 |
+
"__version": 0,
|
| 73 |
+
"model_name": "",
|
| 74 |
+
"include": true,
|
| 75 |
+
"train": true,
|
| 76 |
+
"stop_training_after": 30,
|
| 77 |
+
"stop_training_after_unit": "EPOCH",
|
| 78 |
+
"learning_rate": 7.5e-06,
|
| 79 |
+
"weight_dtype": "BFLOAT_16",
|
| 80 |
+
"dropout_probability": 0.0,
|
| 81 |
+
"train_embedding": true,
|
| 82 |
+
"attention_mask": false
|
| 83 |
+
},
|
| 84 |
+
"prior": {
|
| 85 |
+
"__version": 0,
|
| 86 |
+
"model_name": "",
|
| 87 |
+
"include": true,
|
| 88 |
+
"train": false,
|
| 89 |
+
"stop_training_after": 0,
|
| 90 |
+
"stop_training_after_unit": "NEVER",
|
| 91 |
+
"learning_rate": 3.75e-06,
|
| 92 |
+
"weight_dtype": "FLOAT_32",
|
| 93 |
+
"dropout_probability": 0.0,
|
| 94 |
+
"train_embedding": true,
|
| 95 |
+
"attention_mask": false
|
| 96 |
+
},
|
| 97 |
+
"text_encoder": {
|
| 98 |
+
"__version": 0,
|
| 99 |
+
"model_name": "",
|
| 100 |
+
"include": true,
|
| 101 |
+
"train": true,
|
| 102 |
+
"stop_training_after": 30,
|
| 103 |
+
"stop_training_after_unit": "EPOCH",
|
| 104 |
+
"learning_rate": 3.75e-06,
|
| 105 |
+
"weight_dtype": "BFLOAT_16",
|
| 106 |
+
"dropout_probability": 0.0,
|
| 107 |
+
"train_embedding": true,
|
| 108 |
+
"attention_mask": false
|
| 109 |
+
},
|
| 110 |
+
"text_encoder_layer_skip": 1,
|
| 111 |
+
"text_encoder_2": {
|
| 112 |
+
"__version": 0,
|
| 113 |
+
"model_name": "",
|
| 114 |
+
"include": true,
|
| 115 |
+
"train": true,
|
| 116 |
+
"stop_training_after": 30,
|
| 117 |
+
"stop_training_after_unit": "EPOCH",
|
| 118 |
+
"learning_rate": 3.75e-06,
|
| 119 |
+
"weight_dtype": "BFLOAT_16",
|
| 120 |
+
"dropout_probability": 0.0,
|
| 121 |
+
"train_embedding": true,
|
| 122 |
+
"attention_mask": false
|
| 123 |
+
},
|
| 124 |
+
"text_encoder_2_layer_skip": 1,
|
| 125 |
+
"text_encoder_3": {
|
| 126 |
+
"__version": 0,
|
| 127 |
+
"model_name": "",
|
| 128 |
+
"include": true,
|
| 129 |
+
"train": true,
|
| 130 |
+
"stop_training_after": 30,
|
| 131 |
+
"stop_training_after_unit": "EPOCH",
|
| 132 |
+
"learning_rate": null,
|
| 133 |
+
"weight_dtype": "NONE",
|
| 134 |
+
"dropout_probability": 0.0,
|
| 135 |
+
"train_embedding": true,
|
| 136 |
+
"attention_mask": false
|
| 137 |
+
},
|
| 138 |
+
"text_encoder_3_layer_skip": 0,
|
| 139 |
+
"vae": {
|
| 140 |
+
"__version": 0,
|
| 141 |
+
"model_name": "",
|
| 142 |
+
"include": true,
|
| 143 |
+
"train": true,
|
| 144 |
+
"stop_training_after": null,
|
| 145 |
+
"stop_training_after_unit": "NEVER",
|
| 146 |
+
"learning_rate": null,
|
| 147 |
+
"weight_dtype": "BFLOAT_16",
|
| 148 |
+
"dropout_probability": 0.0,
|
| 149 |
+
"train_embedding": true,
|
| 150 |
+
"attention_mask": false
|
| 151 |
+
},
|
| 152 |
+
"effnet_encoder": {
|
| 153 |
+
"__version": 0,
|
| 154 |
+
"model_name": "",
|
| 155 |
+
"include": true,
|
| 156 |
+
"train": true,
|
| 157 |
+
"stop_training_after": null,
|
| 158 |
+
"stop_training_after_unit": "NEVER",
|
| 159 |
+
"learning_rate": null,
|
| 160 |
+
"weight_dtype": "NONE",
|
| 161 |
+
"dropout_probability": 0.0,
|
| 162 |
+
"train_embedding": true,
|
| 163 |
+
"attention_mask": false
|
| 164 |
+
},
|
| 165 |
+
"decoder": {
|
| 166 |
+
"__version": 0,
|
| 167 |
+
"model_name": "",
|
| 168 |
+
"include": true,
|
| 169 |
+
"train": true,
|
| 170 |
+
"stop_training_after": null,
|
| 171 |
+
"stop_training_after_unit": "NEVER",
|
| 172 |
+
"learning_rate": null,
|
| 173 |
+
"weight_dtype": "NONE",
|
| 174 |
+
"dropout_probability": 0.0,
|
| 175 |
+
"train_embedding": true,
|
| 176 |
+
"attention_mask": false
|
| 177 |
+
},
|
| 178 |
+
"decoder_text_encoder": {
|
| 179 |
+
"__version": 0,
|
| 180 |
+
"model_name": "",
|
| 181 |
+
"include": true,
|
| 182 |
+
"train": true,
|
| 183 |
+
"stop_training_after": null,
|
| 184 |
+
"stop_training_after_unit": "NEVER",
|
| 185 |
+
"learning_rate": null,
|
| 186 |
+
"weight_dtype": "NONE",
|
| 187 |
+
"dropout_probability": 0.0,
|
| 188 |
+
"train_embedding": true,
|
| 189 |
+
"attention_mask": false
|
| 190 |
+
},
|
| 191 |
+
"decoder_vqgan": {
|
| 192 |
+
"__version": 0,
|
| 193 |
+
"model_name": "",
|
| 194 |
+
"include": true,
|
| 195 |
+
"train": true,
|
| 196 |
+
"stop_training_after": null,
|
| 197 |
+
"stop_training_after_unit": "NEVER",
|
| 198 |
+
"learning_rate": null,
|
| 199 |
+
"weight_dtype": "NONE",
|
| 200 |
+
"dropout_probability": 0.0,
|
| 201 |
+
"train_embedding": true,
|
| 202 |
+
"attention_mask": false
|
| 203 |
+
},
|
| 204 |
+
"masked_training": false,
|
| 205 |
+
"unmasked_probability": 0.1,
|
| 206 |
+
"unmasked_weight": 0.1,
|
| 207 |
+
"normalize_masked_area_loss": false,
|
| 208 |
+
"embedding_learning_rate": null,
|
| 209 |
+
"preserve_embedding_norm": false,
|
| 210 |
+
"embedding": {
|
| 211 |
+
"__version": 0,
|
| 212 |
+
"uuid": "4f3755d4-c46e-4b76-b89a-4da883d38e53",
|
| 213 |
+
"model_name": "",
|
| 214 |
+
"placeholder": "<embedding>",
|
| 215 |
+
"train": true,
|
| 216 |
+
"stop_training_after": null,
|
| 217 |
+
"stop_training_after_unit": "NEVER",
|
| 218 |
+
"token_count": 1,
|
| 219 |
+
"initial_embedding_text": "*"
|
| 220 |
+
},
|
| 221 |
+
"additional_embeddings": [],
|
| 222 |
+
"embedding_weight_dtype": "FLOAT_32",
|
| 223 |
+
"peft_type": "LORA",
|
| 224 |
+
"lora_model_name": "",
|
| 225 |
+
"lora_rank": 16,
|
| 226 |
+
"lora_alpha": 1.0,
|
| 227 |
+
"lora_decompose": false,
|
| 228 |
+
"lora_decompose_norm_epsilon": true,
|
| 229 |
+
"lora_weight_dtype": "FLOAT_32",
|
| 230 |
+
"lora_layers": "",
|
| 231 |
+
"lora_layer_preset": null,
|
| 232 |
+
"bundle_additional_embeddings": true,
|
| 233 |
+
"optimizer": {
|
| 234 |
+
"__version": 0,
|
| 235 |
+
"optimizer": "ADAFACTOR",
|
| 236 |
+
"adam_w_mode": false,
|
| 237 |
+
"alpha": null,
|
| 238 |
+
"amsgrad": false,
|
| 239 |
+
"beta1": null,
|
| 240 |
+
"beta2": null,
|
| 241 |
+
"beta3": null,
|
| 242 |
+
"bias_correction": false,
|
| 243 |
+
"block_wise": false,
|
| 244 |
+
"capturable": false,
|
| 245 |
+
"centered": false,
|
| 246 |
+
"clip_threshold": 1.0,
|
| 247 |
+
"d0": null,
|
| 248 |
+
"d_coef": null,
|
| 249 |
+
"dampening": null,
|
| 250 |
+
"decay_rate": -0.8,
|
| 251 |
+
"decouple": false,
|
| 252 |
+
"differentiable": false,
|
| 253 |
+
"eps": 1e-30,
|
| 254 |
+
"eps2": 0.001,
|
| 255 |
+
"foreach": false,
|
| 256 |
+
"fsdp_in_use": false,
|
| 257 |
+
"fused": false,
|
| 258 |
+
"fused_back_pass": true,
|
| 259 |
+
"growth_rate": null,
|
| 260 |
+
"initial_accumulator_value": null,
|
| 261 |
+
"is_paged": false,
|
| 262 |
+
"log_every": null,
|
| 263 |
+
"lr_decay": null,
|
| 264 |
+
"max_unorm": null,
|
| 265 |
+
"maximize": false,
|
| 266 |
+
"min_8bit_size": null,
|
| 267 |
+
"momentum": null,
|
| 268 |
+
"nesterov": false,
|
| 269 |
+
"no_prox": false,
|
| 270 |
+
"optim_bits": null,
|
| 271 |
+
"percentile_clipping": null,
|
| 272 |
+
"r": null,
|
| 273 |
+
"relative_step": false,
|
| 274 |
+
"safeguard_warmup": false,
|
| 275 |
+
"scale_parameter": false,
|
| 276 |
+
"stochastic_rounding": false,
|
| 277 |
+
"use_bias_correction": false,
|
| 278 |
+
"use_triton": false,
|
| 279 |
+
"warmup_init": false,
|
| 280 |
+
"weight_decay": 0.0,
|
| 281 |
+
"weight_lr_power": null,
|
| 282 |
+
"decoupled_decay": false,
|
| 283 |
+
"fixed_decay": false,
|
| 284 |
+
"rectify": false,
|
| 285 |
+
"degenerated_to_sgd": false,
|
| 286 |
+
"k": null,
|
| 287 |
+
"xi": null,
|
| 288 |
+
"n_sma_threshold": null,
|
| 289 |
+
"ams_bound": false,
|
| 290 |
+
"adanorm": false,
|
| 291 |
+
"adam_debias": false
|
| 292 |
+
},
|
| 293 |
+
"optimizer_defaults": {
|
| 294 |
+
"SGD": {
|
| 295 |
+
"__version": 0,
|
| 296 |
+
"optimizer": "SGD",
|
| 297 |
+
"adam_w_mode": false,
|
| 298 |
+
"alpha": null,
|
| 299 |
+
"amsgrad": false,
|
| 300 |
+
"beta1": null,
|
| 301 |
+
"beta2": null,
|
| 302 |
+
"beta3": null,
|
| 303 |
+
"bias_correction": false,
|
| 304 |
+
"block_wise": false,
|
| 305 |
+
"capturable": false,
|
| 306 |
+
"centered": false,
|
| 307 |
+
"clip_threshold": null,
|
| 308 |
+
"d0": null,
|
| 309 |
+
"d_coef": null,
|
| 310 |
+
"dampening": 0.0,
|
| 311 |
+
"decay_rate": null,
|
| 312 |
+
"decouple": false,
|
| 313 |
+
"differentiable": false,
|
| 314 |
+
"eps": null,
|
| 315 |
+
"eps2": null,
|
| 316 |
+
"foreach": false,
|
| 317 |
+
"fsdp_in_use": false,
|
| 318 |
+
"fused": false,
|
| 319 |
+
"fused_back_pass": false,
|
| 320 |
+
"growth_rate": null,
|
| 321 |
+
"initial_accumulator_value": null,
|
| 322 |
+
"is_paged": false,
|
| 323 |
+
"log_every": null,
|
| 324 |
+
"lr_decay": null,
|
| 325 |
+
"max_unorm": null,
|
| 326 |
+
"maximize": false,
|
| 327 |
+
"min_8bit_size": null,
|
| 328 |
+
"momentum": 0.0,
|
| 329 |
+
"nesterov": false,
|
| 330 |
+
"no_prox": false,
|
| 331 |
+
"optim_bits": null,
|
| 332 |
+
"percentile_clipping": null,
|
| 333 |
+
"r": null,
|
| 334 |
+
"relative_step": false,
|
| 335 |
+
"safeguard_warmup": false,
|
| 336 |
+
"scale_parameter": false,
|
| 337 |
+
"stochastic_rounding": true,
|
| 338 |
+
"use_bias_correction": false,
|
| 339 |
+
"use_triton": false,
|
| 340 |
+
"warmup_init": false,
|
| 341 |
+
"weight_decay": 0.0,
|
| 342 |
+
"weight_lr_power": null,
|
| 343 |
+
"decoupled_decay": false,
|
| 344 |
+
"fixed_decay": false,
|
| 345 |
+
"rectify": false,
|
| 346 |
+
"degenerated_to_sgd": false,
|
| 347 |
+
"k": null,
|
| 348 |
+
"xi": null,
|
| 349 |
+
"n_sma_threshold": null,
|
| 350 |
+
"ams_bound": false,
|
| 351 |
+
"adanorm": false,
|
| 352 |
+
"adam_debias": false
|
| 353 |
+
},
|
| 354 |
+
"ADAFACTOR": {
|
| 355 |
+
"__version": 0,
|
| 356 |
+
"optimizer": "ADAFACTOR",
|
| 357 |
+
"adam_w_mode": false,
|
| 358 |
+
"alpha": null,
|
| 359 |
+
"amsgrad": false,
|
| 360 |
+
"beta1": null,
|
| 361 |
+
"beta2": null,
|
| 362 |
+
"beta3": null,
|
| 363 |
+
"bias_correction": false,
|
| 364 |
+
"block_wise": false,
|
| 365 |
+
"capturable": false,
|
| 366 |
+
"centered": false,
|
| 367 |
+
"clip_threshold": 1.0,
|
| 368 |
+
"d0": null,
|
| 369 |
+
"d_coef": null,
|
| 370 |
+
"dampening": null,
|
| 371 |
+
"decay_rate": -0.8,
|
| 372 |
+
"decouple": false,
|
| 373 |
+
"differentiable": false,
|
| 374 |
+
"eps": 1e-30,
|
| 375 |
+
"eps2": 0.001,
|
| 376 |
+
"foreach": false,
|
| 377 |
+
"fsdp_in_use": false,
|
| 378 |
+
"fused": false,
|
| 379 |
+
"fused_back_pass": true,
|
| 380 |
+
"growth_rate": null,
|
| 381 |
+
"initial_accumulator_value": null,
|
| 382 |
+
"is_paged": false,
|
| 383 |
+
"log_every": null,
|
| 384 |
+
"lr_decay": null,
|
| 385 |
+
"max_unorm": null,
|
| 386 |
+
"maximize": false,
|
| 387 |
+
"min_8bit_size": null,
|
| 388 |
+
"momentum": null,
|
| 389 |
+
"nesterov": false,
|
| 390 |
+
"no_prox": false,
|
| 391 |
+
"optim_bits": null,
|
| 392 |
+
"percentile_clipping": null,
|
| 393 |
+
"r": null,
|
| 394 |
+
"relative_step": false,
|
| 395 |
+
"safeguard_warmup": false,
|
| 396 |
+
"scale_parameter": false,
|
| 397 |
+
"stochastic_rounding": false,
|
| 398 |
+
"use_bias_correction": false,
|
| 399 |
+
"use_triton": false,
|
| 400 |
+
"warmup_init": false,
|
| 401 |
+
"weight_decay": 0.0,
|
| 402 |
+
"weight_lr_power": null,
|
| 403 |
+
"decoupled_decay": false,
|
| 404 |
+
"fixed_decay": false,
|
| 405 |
+
"rectify": false,
|
| 406 |
+
"degenerated_to_sgd": false,
|
| 407 |
+
"k": null,
|
| 408 |
+
"xi": null,
|
| 409 |
+
"n_sma_threshold": null,
|
| 410 |
+
"ams_bound": false,
|
| 411 |
+
"adanorm": false,
|
| 412 |
+
"adam_debias": false
|
| 413 |
+
}
|
| 414 |
+
},
|
| 415 |
+
"sample_definition_file_name": "training_samples/samples.json",
|
| 416 |
+
"samples": null,
|
| 417 |
+
"sample_after": 100,
|
| 418 |
+
"sample_after_unit": "STEP",
|
| 419 |
+
"sample_image_format": "JPG",
|
| 420 |
+
"samples_to_tensorboard": true,
|
| 421 |
+
"non_ema_sampling": false,
|
| 422 |
+
"backup_after": 500,
|
| 423 |
+
"backup_after_unit": "STEP",
|
| 424 |
+
"rolling_backup": true,
|
| 425 |
+
"rolling_backup_count": 4,
|
| 426 |
+
"backup_before_save": true,
|
| 427 |
+
"save_after": 1,
|
| 428 |
+
"save_after_unit": "EPOCH",
|
| 429 |
+
"save_filename_prefix": "back"
|
| 430 |
+
}
|