Upload folder using huggingface_hub
Browse files- adapter_config.json +6 -6
- adapter_model.bin +1 -1
- config.json +11 -33
- non_lora_trainables.bin +1 -1
- trainer_state.json +1864 -934
adapter_config.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"auto_mapping": null,
|
| 3 |
-
"base_model_name_or_path": "llava-
|
| 4 |
"bias": "none",
|
| 5 |
"fan_in_fan_out": false,
|
| 6 |
"inference_mode": true,
|
|
@@ -14,13 +14,13 @@
|
|
| 14 |
"r": 128,
|
| 15 |
"revision": null,
|
| 16 |
"target_modules": [
|
| 17 |
-
"
|
| 18 |
-
"
|
|
|
|
| 19 |
"gate_proj",
|
| 20 |
"o_proj",
|
| 21 |
-
"
|
| 22 |
-
"
|
| 23 |
-
"down_proj"
|
| 24 |
],
|
| 25 |
"task_type": "CAUSAL_LM"
|
| 26 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "/home/szaman/llava-finetuning/llava-v1.5-7b",
|
| 4 |
"bias": "none",
|
| 5 |
"fan_in_fan_out": false,
|
| 6 |
"inference_mode": true,
|
|
|
|
| 14 |
"r": 128,
|
| 15 |
"revision": null,
|
| 16 |
"target_modules": [
|
| 17 |
+
"down_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"k_proj",
|
| 20 |
"gate_proj",
|
| 21 |
"o_proj",
|
| 22 |
+
"up_proj",
|
| 23 |
+
"v_proj"
|
|
|
|
| 24 |
],
|
| 25 |
"task_type": "CAUSAL_LM"
|
| 26 |
}
|
adapter_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 639787082
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa50d3f2dd8381744ad3af67d7dc60059a3755e3675981eadc00846761d2b12e
|
| 3 |
size 639787082
|
config.json
CHANGED
|
@@ -1,25 +1,26 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "llava-
|
| 3 |
"architectures": [
|
| 4 |
-
"
|
| 5 |
],
|
| 6 |
"attention_bias": false,
|
| 7 |
"attention_dropout": 0.0,
|
| 8 |
"bos_token_id": 1,
|
| 9 |
"eos_token_id": 2,
|
| 10 |
"freeze_mm_mlp_adapter": false,
|
|
|
|
| 11 |
"hidden_act": "silu",
|
| 12 |
"hidden_size": 4096,
|
| 13 |
-
"ignore_index": -100,
|
| 14 |
"image_aspect_ratio": "pad",
|
| 15 |
-
"image_token_index": 32000,
|
| 16 |
"initializer_range": 0.02,
|
| 17 |
"intermediate_size": 11008,
|
| 18 |
-
"
|
|
|
|
| 19 |
"mm_hidden_size": 1024,
|
| 20 |
"mm_patch_merge_type": "flat",
|
| 21 |
"mm_projector_lr": 2e-05,
|
| 22 |
"mm_projector_type": "mlp2x_gelu",
|
|
|
|
| 23 |
"mm_use_im_patch_token": false,
|
| 24 |
"mm_use_im_start_end": false,
|
| 25 |
"mm_vision_select_feature": "patch",
|
|
@@ -29,43 +30,20 @@
|
|
| 29 |
"num_attention_heads": 32,
|
| 30 |
"num_hidden_layers": 32,
|
| 31 |
"num_key_value_heads": 32,
|
| 32 |
-
"pad_token_id":
|
| 33 |
"pretraining_tp": 1,
|
| 34 |
-
"
|
| 35 |
-
"rms_norm_eps": 1e-06,
|
| 36 |
"rope_scaling": null,
|
| 37 |
"rope_theta": 10000.0,
|
| 38 |
-
"text_config": {
|
| 39 |
-
"_name_or_path": "lmsys/vicuna-7b-v1.5",
|
| 40 |
-
"architectures": [
|
| 41 |
-
"LlamaForCausalLM"
|
| 42 |
-
],
|
| 43 |
-
"max_position_embeddings": 4096,
|
| 44 |
-
"model_type": "llama",
|
| 45 |
-
"rms_norm_eps": 1e-05,
|
| 46 |
-
"torch_dtype": "float16",
|
| 47 |
-
"vocab_size": 32064
|
| 48 |
-
},
|
| 49 |
"tie_word_embeddings": false,
|
| 50 |
"tokenizer_model_max_length": 2048,
|
| 51 |
"tokenizer_padding_side": "right",
|
| 52 |
"torch_dtype": "float16",
|
| 53 |
"transformers_version": "4.37.2",
|
| 54 |
"tune_mm_mlp_adapter": false,
|
|
|
|
|
|
|
| 55 |
"use_cache": true,
|
| 56 |
"use_mm_proj": true,
|
| 57 |
-
"
|
| 58 |
-
"hidden_size": 1024,
|
| 59 |
-
"image_size": 336,
|
| 60 |
-
"intermediate_size": 4096,
|
| 61 |
-
"model_type": "clip_vision_model",
|
| 62 |
-
"num_attention_heads": 16,
|
| 63 |
-
"num_hidden_layers": 24,
|
| 64 |
-
"patch_size": 14,
|
| 65 |
-
"projection_dim": 768,
|
| 66 |
-
"vocab_size": 32000
|
| 67 |
-
},
|
| 68 |
-
"vision_feature_layer": -2,
|
| 69 |
-
"vision_feature_select_strategy": "default",
|
| 70 |
-
"vocab_size": 32064
|
| 71 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/home/szaman/llava-finetuning/llava-v1.5-7b",
|
| 3 |
"architectures": [
|
| 4 |
+
"LlavaLlamaForCausalLM"
|
| 5 |
],
|
| 6 |
"attention_bias": false,
|
| 7 |
"attention_dropout": 0.0,
|
| 8 |
"bos_token_id": 1,
|
| 9 |
"eos_token_id": 2,
|
| 10 |
"freeze_mm_mlp_adapter": false,
|
| 11 |
+
"freeze_mm_vision_resampler": false,
|
| 12 |
"hidden_act": "silu",
|
| 13 |
"hidden_size": 4096,
|
|
|
|
| 14 |
"image_aspect_ratio": "pad",
|
|
|
|
| 15 |
"initializer_range": 0.02,
|
| 16 |
"intermediate_size": 11008,
|
| 17 |
+
"max_length": 4096,
|
| 18 |
+
"max_position_embeddings": 4096,
|
| 19 |
"mm_hidden_size": 1024,
|
| 20 |
"mm_patch_merge_type": "flat",
|
| 21 |
"mm_projector_lr": 2e-05,
|
| 22 |
"mm_projector_type": "mlp2x_gelu",
|
| 23 |
+
"mm_resampler_type": null,
|
| 24 |
"mm_use_im_patch_token": false,
|
| 25 |
"mm_use_im_start_end": false,
|
| 26 |
"mm_vision_select_feature": "patch",
|
|
|
|
| 30 |
"num_attention_heads": 32,
|
| 31 |
"num_hidden_layers": 32,
|
| 32 |
"num_key_value_heads": 32,
|
| 33 |
+
"pad_token_id": 0,
|
| 34 |
"pretraining_tp": 1,
|
| 35 |
+
"rms_norm_eps": 1e-05,
|
|
|
|
| 36 |
"rope_scaling": null,
|
| 37 |
"rope_theta": 10000.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
"tie_word_embeddings": false,
|
| 39 |
"tokenizer_model_max_length": 2048,
|
| 40 |
"tokenizer_padding_side": "right",
|
| 41 |
"torch_dtype": "float16",
|
| 42 |
"transformers_version": "4.37.2",
|
| 43 |
"tune_mm_mlp_adapter": false,
|
| 44 |
+
"tune_mm_vision_resampler": false,
|
| 45 |
+
"unfreeze_mm_vision_tower": false,
|
| 46 |
"use_cache": true,
|
| 47 |
"use_mm_proj": true,
|
| 48 |
+
"vocab_size": 32000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
}
|
non_lora_trainables.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 41961648
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a887f78d66542f34d5592dd973383a4a9fd28767ab410c9435662c89082c3c2
|
| 3 |
size 41961648
|
trainer_state.json
CHANGED
|
@@ -3,1887 +3,2817 @@
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss":
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate":
|
| 20 |
-
"loss":
|
| 21 |
"step": 2
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
-
"learning_rate":
|
| 26 |
-
"loss":
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.01,
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss":
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
-
"learning_rate":
|
| 38 |
-
"loss":
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
-
"learning_rate":
|
| 44 |
-
"loss":
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.02,
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss":
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate": 0.
|
| 56 |
-
"loss":
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"learning_rate": 0.
|
| 62 |
-
"loss":
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch": 0.
|
| 67 |
-
"learning_rate": 0.
|
| 68 |
-
"loss":
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
-
"learning_rate": 0.
|
| 74 |
-
"loss":
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss":
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch": 0.
|
| 85 |
-
"learning_rate": 0.
|
| 86 |
-
"loss":
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
-
"learning_rate": 0.
|
| 92 |
-
"loss":
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate": 0.
|
| 98 |
-
"loss":
|
| 99 |
"step": 15
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch": 0.
|
| 103 |
-
"learning_rate": 0.
|
| 104 |
-
"loss":
|
| 105 |
"step": 16
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch": 0.
|
| 109 |
-
"learning_rate": 0.
|
| 110 |
-
"loss":
|
| 111 |
"step": 17
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch": 0.
|
| 115 |
-
"learning_rate": 0.
|
| 116 |
-
"loss":
|
| 117 |
"step": 18
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch": 0.
|
| 121 |
-
"learning_rate": 0.
|
| 122 |
-
"loss":
|
| 123 |
"step": 19
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch": 0.
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss":
|
| 129 |
"step": 20
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch": 0.
|
| 133 |
-
"learning_rate": 0.
|
| 134 |
-
"loss":
|
| 135 |
"step": 21
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"learning_rate": 0.
|
| 140 |
-
"loss":
|
| 141 |
"step": 22
|
| 142 |
},
|
| 143 |
{
|
| 144 |
-
"epoch": 0.
|
| 145 |
-
"learning_rate": 0.
|
| 146 |
-
"loss":
|
| 147 |
"step": 23
|
| 148 |
},
|
| 149 |
{
|
| 150 |
-
"epoch": 0.
|
| 151 |
-
"learning_rate": 0.
|
| 152 |
-
"loss":
|
| 153 |
"step": 24
|
| 154 |
},
|
| 155 |
{
|
| 156 |
-
"epoch": 0.
|
| 157 |
-
"learning_rate": 0.
|
| 158 |
-
"loss":
|
| 159 |
"step": 25
|
| 160 |
},
|
| 161 |
{
|
| 162 |
-
"epoch": 0.
|
| 163 |
-
"learning_rate": 0.
|
| 164 |
-
"loss":
|
| 165 |
"step": 26
|
| 166 |
},
|
| 167 |
{
|
| 168 |
-
"epoch": 0.
|
| 169 |
-
"learning_rate": 0.
|
| 170 |
-
"loss":
|
| 171 |
"step": 27
|
| 172 |
},
|
| 173 |
{
|
| 174 |
-
"epoch": 0.
|
| 175 |
-
"learning_rate": 0.
|
| 176 |
-
"loss":
|
| 177 |
"step": 28
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
-
"learning_rate": 0.
|
| 182 |
-
"loss":
|
| 183 |
"step": 29
|
| 184 |
},
|
| 185 |
{
|
| 186 |
-
"epoch": 0.
|
| 187 |
-
"learning_rate": 0.
|
| 188 |
-
"loss":
|
| 189 |
"step": 30
|
| 190 |
},
|
| 191 |
{
|
| 192 |
-
"epoch": 0.
|
| 193 |
-
"learning_rate": 0.
|
| 194 |
-
"loss":
|
| 195 |
"step": 31
|
| 196 |
},
|
| 197 |
{
|
| 198 |
-
"epoch": 0.
|
| 199 |
-
"learning_rate": 0.
|
| 200 |
-
"loss":
|
| 201 |
"step": 32
|
| 202 |
},
|
| 203 |
{
|
| 204 |
-
"epoch": 0.
|
| 205 |
-
"learning_rate": 0.
|
| 206 |
-
"loss":
|
| 207 |
"step": 33
|
| 208 |
},
|
| 209 |
{
|
| 210 |
-
"epoch": 0.
|
| 211 |
-
"learning_rate": 0.
|
| 212 |
-
"loss":
|
| 213 |
"step": 34
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch": 0.
|
| 217 |
-
"learning_rate": 0.
|
| 218 |
-
"loss":
|
| 219 |
"step": 35
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
-
"learning_rate": 0.
|
| 224 |
-
"loss":
|
| 225 |
"step": 36
|
| 226 |
},
|
| 227 |
{
|
| 228 |
-
"epoch": 0.
|
| 229 |
-
"learning_rate": 0.
|
| 230 |
-
"loss":
|
| 231 |
"step": 37
|
| 232 |
},
|
| 233 |
{
|
| 234 |
-
"epoch": 0.
|
| 235 |
-
"learning_rate": 0.
|
| 236 |
-
"loss":
|
| 237 |
"step": 38
|
| 238 |
},
|
| 239 |
{
|
| 240 |
-
"epoch": 0.
|
| 241 |
-
"learning_rate": 0.
|
| 242 |
-
"loss":
|
| 243 |
"step": 39
|
| 244 |
},
|
| 245 |
{
|
| 246 |
-
"epoch": 0.
|
| 247 |
-
"learning_rate": 0.
|
| 248 |
-
"loss":
|
| 249 |
"step": 40
|
| 250 |
},
|
| 251 |
{
|
| 252 |
-
"epoch": 0.
|
| 253 |
-
"learning_rate": 0.
|
| 254 |
-
"loss":
|
| 255 |
"step": 41
|
| 256 |
},
|
| 257 |
{
|
| 258 |
-
"epoch": 0.
|
| 259 |
-
"learning_rate": 0.
|
| 260 |
-
"loss":
|
| 261 |
"step": 42
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
-
"learning_rate": 0.
|
| 266 |
-
"loss":
|
| 267 |
"step": 43
|
| 268 |
},
|
| 269 |
{
|
| 270 |
-
"epoch": 0.
|
| 271 |
-
"learning_rate": 0.
|
| 272 |
-
"loss":
|
| 273 |
"step": 44
|
| 274 |
},
|
| 275 |
{
|
| 276 |
-
"epoch": 0.
|
| 277 |
-
"learning_rate": 0.
|
| 278 |
-
"loss":
|
| 279 |
"step": 45
|
| 280 |
},
|
| 281 |
{
|
| 282 |
-
"epoch": 0.
|
| 283 |
-
"learning_rate": 0.
|
| 284 |
-
"loss":
|
| 285 |
"step": 46
|
| 286 |
},
|
| 287 |
{
|
| 288 |
-
"epoch": 0.
|
| 289 |
-
"learning_rate": 0.
|
| 290 |
-
"loss":
|
| 291 |
"step": 47
|
| 292 |
},
|
| 293 |
{
|
| 294 |
-
"epoch": 0.
|
| 295 |
-
"learning_rate": 0.
|
| 296 |
-
"loss":
|
| 297 |
"step": 48
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch": 0.
|
| 301 |
-
"learning_rate": 0.
|
| 302 |
-
"loss":
|
| 303 |
"step": 49
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 0.
|
| 307 |
-
"learning_rate": 0.
|
| 308 |
-
"loss":
|
| 309 |
"step": 50
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch": 0.
|
| 313 |
-
"learning_rate": 0.
|
| 314 |
-
"loss":
|
| 315 |
"step": 51
|
| 316 |
},
|
| 317 |
{
|
| 318 |
-
"epoch": 0.
|
| 319 |
-
"learning_rate": 0.
|
| 320 |
-
"loss":
|
| 321 |
"step": 52
|
| 322 |
},
|
| 323 |
{
|
| 324 |
-
"epoch": 0.
|
| 325 |
-
"learning_rate": 0.
|
| 326 |
-
"loss":
|
| 327 |
"step": 53
|
| 328 |
},
|
| 329 |
{
|
| 330 |
-
"epoch": 0.
|
| 331 |
-
"learning_rate": 0.
|
| 332 |
-
"loss":
|
| 333 |
"step": 54
|
| 334 |
},
|
| 335 |
{
|
| 336 |
-
"epoch": 0.
|
| 337 |
-
"learning_rate": 0.
|
| 338 |
-
"loss":
|
| 339 |
"step": 55
|
| 340 |
},
|
| 341 |
{
|
| 342 |
-
"epoch": 0.
|
| 343 |
-
"learning_rate": 0.
|
| 344 |
-
"loss":
|
| 345 |
"step": 56
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 0.
|
| 349 |
-
"learning_rate": 0.
|
| 350 |
-
"loss":
|
| 351 |
"step": 57
|
| 352 |
},
|
| 353 |
{
|
| 354 |
-
"epoch": 0.
|
| 355 |
-
"learning_rate": 0.
|
| 356 |
-
"loss":
|
| 357 |
"step": 58
|
| 358 |
},
|
| 359 |
{
|
| 360 |
-
"epoch": 0.
|
| 361 |
-
"learning_rate": 0.
|
| 362 |
-
"loss":
|
| 363 |
"step": 59
|
| 364 |
},
|
| 365 |
{
|
| 366 |
-
"epoch": 0.
|
| 367 |
-
"learning_rate": 0.
|
| 368 |
-
"loss":
|
| 369 |
"step": 60
|
| 370 |
},
|
| 371 |
{
|
| 372 |
-
"epoch": 0.
|
| 373 |
-
"learning_rate": 0.
|
| 374 |
-
"loss":
|
| 375 |
"step": 61
|
| 376 |
},
|
| 377 |
{
|
| 378 |
-
"epoch": 0.
|
| 379 |
-
"learning_rate": 0.
|
| 380 |
-
"loss":
|
| 381 |
"step": 62
|
| 382 |
},
|
| 383 |
{
|
| 384 |
-
"epoch": 0.
|
| 385 |
-
"learning_rate": 0.
|
| 386 |
-
"loss":
|
| 387 |
"step": 63
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch": 0.
|
| 391 |
-
"learning_rate": 0.
|
| 392 |
-
"loss":
|
| 393 |
"step": 64
|
| 394 |
},
|
| 395 |
{
|
| 396 |
-
"epoch": 0.
|
| 397 |
-
"learning_rate": 0.
|
| 398 |
-
"loss":
|
| 399 |
"step": 65
|
| 400 |
},
|
| 401 |
{
|
| 402 |
-
"epoch": 0.
|
| 403 |
-
"learning_rate": 0.
|
| 404 |
-
"loss":
|
| 405 |
"step": 66
|
| 406 |
},
|
| 407 |
{
|
| 408 |
-
"epoch": 0.
|
| 409 |
-
"learning_rate": 0.
|
| 410 |
-
"loss":
|
| 411 |
"step": 67
|
| 412 |
},
|
| 413 |
{
|
| 414 |
-
"epoch": 0.
|
| 415 |
-
"learning_rate": 0.
|
| 416 |
-
"loss":
|
| 417 |
"step": 68
|
| 418 |
},
|
| 419 |
{
|
| 420 |
-
"epoch": 0.
|
| 421 |
-
"learning_rate": 0.
|
| 422 |
-
"loss":
|
| 423 |
"step": 69
|
| 424 |
},
|
| 425 |
{
|
| 426 |
-
"epoch": 0.
|
| 427 |
-
"learning_rate": 0.
|
| 428 |
-
"loss":
|
| 429 |
"step": 70
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch": 0.
|
| 433 |
-
"learning_rate": 0.
|
| 434 |
-
"loss":
|
| 435 |
"step": 71
|
| 436 |
},
|
| 437 |
{
|
| 438 |
-
"epoch": 0.
|
| 439 |
-
"learning_rate": 0.
|
| 440 |
-
"loss":
|
| 441 |
"step": 72
|
| 442 |
},
|
| 443 |
{
|
| 444 |
-
"epoch": 0.
|
| 445 |
-
"learning_rate": 0.
|
| 446 |
-
"loss":
|
| 447 |
"step": 73
|
| 448 |
},
|
| 449 |
{
|
| 450 |
-
"epoch": 0.
|
| 451 |
-
"learning_rate": 0.
|
| 452 |
-
"loss":
|
| 453 |
"step": 74
|
| 454 |
},
|
| 455 |
{
|
| 456 |
-
"epoch": 0.
|
| 457 |
-
"learning_rate": 0.
|
| 458 |
-
"loss":
|
| 459 |
"step": 75
|
| 460 |
},
|
| 461 |
{
|
| 462 |
-
"epoch": 0.
|
| 463 |
-
"learning_rate": 0.
|
| 464 |
-
"loss":
|
| 465 |
"step": 76
|
| 466 |
},
|
| 467 |
{
|
| 468 |
-
"epoch": 0.
|
| 469 |
-
"learning_rate": 0.
|
| 470 |
-
"loss":
|
| 471 |
"step": 77
|
| 472 |
},
|
| 473 |
{
|
| 474 |
-
"epoch": 0.
|
| 475 |
-
"learning_rate": 0.
|
| 476 |
-
"loss":
|
| 477 |
"step": 78
|
| 478 |
},
|
| 479 |
{
|
| 480 |
-
"epoch": 0.
|
| 481 |
-
"learning_rate": 0.
|
| 482 |
-
"loss":
|
| 483 |
"step": 79
|
| 484 |
},
|
| 485 |
{
|
| 486 |
-
"epoch": 0.
|
| 487 |
-
"learning_rate": 0.
|
| 488 |
-
"loss":
|
| 489 |
"step": 80
|
| 490 |
},
|
| 491 |
{
|
| 492 |
-
"epoch": 0.
|
| 493 |
-
"learning_rate": 0.
|
| 494 |
-
"loss":
|
| 495 |
"step": 81
|
| 496 |
},
|
| 497 |
{
|
| 498 |
-
"epoch": 0.
|
| 499 |
-
"learning_rate": 0.
|
| 500 |
-
"loss":
|
| 501 |
"step": 82
|
| 502 |
},
|
| 503 |
{
|
| 504 |
-
"epoch": 0.
|
| 505 |
-
"learning_rate": 0.
|
| 506 |
-
"loss":
|
| 507 |
"step": 83
|
| 508 |
},
|
| 509 |
{
|
| 510 |
-
"epoch": 0.
|
| 511 |
-
"learning_rate": 0.
|
| 512 |
-
"loss":
|
| 513 |
"step": 84
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch": 0.
|
| 517 |
-
"learning_rate": 0.
|
| 518 |
-
"loss":
|
| 519 |
"step": 85
|
| 520 |
},
|
| 521 |
{
|
| 522 |
-
"epoch": 0.
|
| 523 |
-
"learning_rate": 0.
|
| 524 |
-
"loss":
|
| 525 |
"step": 86
|
| 526 |
},
|
| 527 |
{
|
| 528 |
-
"epoch": 0.
|
| 529 |
-
"learning_rate": 0.
|
| 530 |
-
"loss":
|
| 531 |
"step": 87
|
| 532 |
},
|
| 533 |
{
|
| 534 |
-
"epoch": 0.
|
| 535 |
-
"learning_rate": 0.
|
| 536 |
-
"loss":
|
| 537 |
"step": 88
|
| 538 |
},
|
| 539 |
{
|
| 540 |
-
"epoch": 0.
|
| 541 |
-
"learning_rate": 0.
|
| 542 |
-
"loss":
|
| 543 |
"step": 89
|
| 544 |
},
|
| 545 |
{
|
| 546 |
-
"epoch": 0.
|
| 547 |
-
"learning_rate": 0.
|
| 548 |
-
"loss":
|
| 549 |
"step": 90
|
| 550 |
},
|
| 551 |
{
|
| 552 |
-
"epoch": 0.
|
| 553 |
-
"learning_rate": 0.
|
| 554 |
-
"loss":
|
| 555 |
"step": 91
|
| 556 |
},
|
| 557 |
{
|
| 558 |
-
"epoch": 0.
|
| 559 |
-
"learning_rate": 0.
|
| 560 |
-
"loss":
|
| 561 |
"step": 92
|
| 562 |
},
|
| 563 |
{
|
| 564 |
-
"epoch": 0.
|
| 565 |
-
"learning_rate": 0.
|
| 566 |
-
"loss":
|
| 567 |
"step": 93
|
| 568 |
},
|
| 569 |
{
|
| 570 |
-
"epoch": 0.
|
| 571 |
-
"learning_rate": 0.
|
| 572 |
-
"loss":
|
| 573 |
"step": 94
|
| 574 |
},
|
| 575 |
{
|
| 576 |
-
"epoch": 0.
|
| 577 |
-
"learning_rate": 0.
|
| 578 |
-
"loss":
|
| 579 |
"step": 95
|
| 580 |
},
|
| 581 |
{
|
| 582 |
-
"epoch": 0.
|
| 583 |
-
"learning_rate": 0.
|
| 584 |
-
"loss":
|
| 585 |
"step": 96
|
| 586 |
},
|
| 587 |
{
|
| 588 |
-
"epoch": 0.
|
| 589 |
-
"learning_rate": 0.
|
| 590 |
-
"loss":
|
| 591 |
"step": 97
|
| 592 |
},
|
| 593 |
{
|
| 594 |
-
"epoch": 0.
|
| 595 |
-
"learning_rate": 0.
|
| 596 |
-
"loss":
|
| 597 |
"step": 98
|
| 598 |
},
|
| 599 |
{
|
| 600 |
-
"epoch": 0.
|
| 601 |
-
"learning_rate": 0.
|
| 602 |
-
"loss":
|
| 603 |
"step": 99
|
| 604 |
},
|
| 605 |
{
|
| 606 |
-
"epoch": 0.
|
| 607 |
-
"learning_rate": 0.
|
| 608 |
-
"loss":
|
| 609 |
"step": 100
|
| 610 |
},
|
| 611 |
{
|
| 612 |
-
"epoch": 0.
|
| 613 |
-
"learning_rate": 0.
|
| 614 |
-
"loss":
|
| 615 |
"step": 101
|
| 616 |
},
|
| 617 |
{
|
| 618 |
-
"epoch": 0.
|
| 619 |
-
"learning_rate": 0.
|
| 620 |
-
"loss":
|
| 621 |
"step": 102
|
| 622 |
},
|
| 623 |
{
|
| 624 |
-
"epoch": 0.
|
| 625 |
-
"learning_rate": 0.
|
| 626 |
-
"loss":
|
| 627 |
"step": 103
|
| 628 |
},
|
| 629 |
{
|
| 630 |
-
"epoch": 0.
|
| 631 |
-
"learning_rate": 0.
|
| 632 |
-
"loss":
|
| 633 |
"step": 104
|
| 634 |
},
|
| 635 |
{
|
| 636 |
-
"epoch": 0.
|
| 637 |
-
"learning_rate": 0.
|
| 638 |
-
"loss":
|
| 639 |
"step": 105
|
| 640 |
},
|
| 641 |
{
|
| 642 |
-
"epoch": 0.
|
| 643 |
-
"learning_rate": 0.
|
| 644 |
-
"loss":
|
| 645 |
"step": 106
|
| 646 |
},
|
| 647 |
{
|
| 648 |
-
"epoch": 0.
|
| 649 |
-
"learning_rate": 0.
|
| 650 |
-
"loss":
|
| 651 |
"step": 107
|
| 652 |
},
|
| 653 |
{
|
| 654 |
-
"epoch": 0.
|
| 655 |
-
"learning_rate": 0.
|
| 656 |
-
"loss":
|
| 657 |
"step": 108
|
| 658 |
},
|
| 659 |
{
|
| 660 |
-
"epoch": 0.
|
| 661 |
-
"learning_rate": 0.
|
| 662 |
-
"loss":
|
| 663 |
"step": 109
|
| 664 |
},
|
| 665 |
{
|
| 666 |
-
"epoch": 0.
|
| 667 |
-
"learning_rate": 0.
|
| 668 |
-
"loss":
|
| 669 |
"step": 110
|
| 670 |
},
|
| 671 |
{
|
| 672 |
-
"epoch": 0.
|
| 673 |
-
"learning_rate": 0.
|
| 674 |
-
"loss":
|
| 675 |
"step": 111
|
| 676 |
},
|
| 677 |
{
|
| 678 |
-
"epoch": 0.
|
| 679 |
-
"learning_rate": 0.
|
| 680 |
-
"loss":
|
| 681 |
"step": 112
|
| 682 |
},
|
| 683 |
{
|
| 684 |
-
"epoch": 0.
|
| 685 |
-
"learning_rate": 0.
|
| 686 |
-
"loss":
|
| 687 |
"step": 113
|
| 688 |
},
|
| 689 |
{
|
| 690 |
-
"epoch": 0.
|
| 691 |
-
"learning_rate": 0.
|
| 692 |
-
"loss":
|
| 693 |
"step": 114
|
| 694 |
},
|
| 695 |
{
|
| 696 |
-
"epoch": 0.
|
| 697 |
-
"learning_rate": 0.
|
| 698 |
-
"loss":
|
| 699 |
"step": 115
|
| 700 |
},
|
| 701 |
{
|
| 702 |
-
"epoch": 0.
|
| 703 |
-
"learning_rate": 0.
|
| 704 |
-
"loss":
|
| 705 |
"step": 116
|
| 706 |
},
|
| 707 |
{
|
| 708 |
-
"epoch": 0.
|
| 709 |
-
"learning_rate": 0.
|
| 710 |
-
"loss":
|
| 711 |
"step": 117
|
| 712 |
},
|
| 713 |
{
|
| 714 |
-
"epoch": 0.
|
| 715 |
-
"learning_rate": 0.
|
| 716 |
-
"loss":
|
| 717 |
"step": 118
|
| 718 |
},
|
| 719 |
{
|
| 720 |
-
"epoch": 0.
|
| 721 |
-
"learning_rate": 0.
|
| 722 |
-
"loss":
|
| 723 |
"step": 119
|
| 724 |
},
|
| 725 |
{
|
| 726 |
-
"epoch": 0.
|
| 727 |
-
"learning_rate": 0.
|
| 728 |
-
"loss":
|
| 729 |
"step": 120
|
| 730 |
},
|
| 731 |
{
|
| 732 |
-
"epoch": 0.
|
| 733 |
-
"learning_rate": 0.
|
| 734 |
-
"loss":
|
| 735 |
"step": 121
|
| 736 |
},
|
| 737 |
{
|
| 738 |
-
"epoch": 0.
|
| 739 |
-
"learning_rate": 0.
|
| 740 |
-
"loss":
|
| 741 |
"step": 122
|
| 742 |
},
|
| 743 |
{
|
| 744 |
-
"epoch": 0.
|
| 745 |
-
"learning_rate": 0.
|
| 746 |
-
"loss":
|
| 747 |
"step": 123
|
| 748 |
},
|
| 749 |
{
|
| 750 |
-
"epoch": 0.
|
| 751 |
-
"learning_rate": 0.
|
| 752 |
-
"loss":
|
| 753 |
"step": 124
|
| 754 |
},
|
| 755 |
{
|
| 756 |
-
"epoch": 0.
|
| 757 |
-
"learning_rate": 0.
|
| 758 |
-
"loss":
|
| 759 |
"step": 125
|
| 760 |
},
|
| 761 |
{
|
| 762 |
-
"epoch": 0.
|
| 763 |
-
"learning_rate": 0.
|
| 764 |
-
"loss":
|
| 765 |
"step": 126
|
| 766 |
},
|
| 767 |
{
|
| 768 |
-
"epoch": 0.
|
| 769 |
-
"learning_rate": 0.
|
| 770 |
-
"loss":
|
| 771 |
"step": 127
|
| 772 |
},
|
| 773 |
{
|
| 774 |
-
"epoch": 0.
|
| 775 |
-
"learning_rate": 0.
|
| 776 |
-
"loss":
|
| 777 |
"step": 128
|
| 778 |
},
|
| 779 |
{
|
| 780 |
-
"epoch": 0.
|
| 781 |
-
"learning_rate": 0.
|
| 782 |
-
"loss":
|
| 783 |
"step": 129
|
| 784 |
},
|
| 785 |
{
|
| 786 |
-
"epoch": 0.
|
| 787 |
-
"learning_rate": 0.
|
| 788 |
-
"loss":
|
| 789 |
"step": 130
|
| 790 |
},
|
| 791 |
{
|
| 792 |
-
"epoch": 0.
|
| 793 |
-
"learning_rate": 0.
|
| 794 |
-
"loss":
|
| 795 |
"step": 131
|
| 796 |
},
|
| 797 |
{
|
| 798 |
-
"epoch": 0.
|
| 799 |
-
"learning_rate": 0.
|
| 800 |
-
"loss":
|
| 801 |
"step": 132
|
| 802 |
},
|
| 803 |
{
|
| 804 |
-
"epoch": 0.
|
| 805 |
-
"learning_rate": 0.
|
| 806 |
-
"loss":
|
| 807 |
"step": 133
|
| 808 |
},
|
| 809 |
{
|
| 810 |
-
"epoch": 0.
|
| 811 |
-
"learning_rate": 0.
|
| 812 |
-
"loss":
|
| 813 |
"step": 134
|
| 814 |
},
|
| 815 |
{
|
| 816 |
-
"epoch": 0.
|
| 817 |
-
"learning_rate": 0.
|
| 818 |
-
"loss":
|
| 819 |
"step": 135
|
| 820 |
},
|
| 821 |
{
|
| 822 |
-
"epoch": 0.
|
| 823 |
-
"learning_rate": 0.
|
| 824 |
-
"loss":
|
| 825 |
"step": 136
|
| 826 |
},
|
| 827 |
{
|
| 828 |
-
"epoch": 0.
|
| 829 |
-
"learning_rate": 0.
|
| 830 |
-
"loss":
|
| 831 |
"step": 137
|
| 832 |
},
|
| 833 |
{
|
| 834 |
-
"epoch": 0.
|
| 835 |
-
"learning_rate": 0.
|
| 836 |
-
"loss":
|
| 837 |
"step": 138
|
| 838 |
},
|
| 839 |
{
|
| 840 |
-
"epoch": 0.
|
| 841 |
-
"learning_rate": 0.
|
| 842 |
-
"loss":
|
| 843 |
"step": 139
|
| 844 |
},
|
| 845 |
{
|
| 846 |
-
"epoch": 0.
|
| 847 |
-
"learning_rate": 0.
|
| 848 |
-
"loss":
|
| 849 |
"step": 140
|
| 850 |
},
|
| 851 |
{
|
| 852 |
-
"epoch": 0.
|
| 853 |
-
"learning_rate": 0.
|
| 854 |
-
"loss":
|
| 855 |
"step": 141
|
| 856 |
},
|
| 857 |
{
|
| 858 |
-
"epoch": 0.
|
| 859 |
-
"learning_rate": 0.
|
| 860 |
-
"loss":
|
| 861 |
"step": 142
|
| 862 |
},
|
| 863 |
{
|
| 864 |
-
"epoch": 0.
|
| 865 |
-
"learning_rate": 0.
|
| 866 |
-
"loss":
|
| 867 |
"step": 143
|
| 868 |
},
|
| 869 |
{
|
| 870 |
-
"epoch": 0.
|
| 871 |
-
"learning_rate": 0.
|
| 872 |
-
"loss":
|
| 873 |
"step": 144
|
| 874 |
},
|
| 875 |
{
|
| 876 |
-
"epoch": 0.
|
| 877 |
-
"learning_rate": 0.
|
| 878 |
-
"loss":
|
| 879 |
"step": 145
|
| 880 |
},
|
| 881 |
{
|
| 882 |
-
"epoch": 0.
|
| 883 |
-
"learning_rate": 0.
|
| 884 |
-
"loss":
|
| 885 |
"step": 146
|
| 886 |
},
|
| 887 |
{
|
| 888 |
-
"epoch": 0.
|
| 889 |
-
"learning_rate": 0.
|
| 890 |
-
"loss":
|
| 891 |
"step": 147
|
| 892 |
},
|
| 893 |
{
|
| 894 |
-
"epoch": 0.
|
| 895 |
-
"learning_rate": 0.
|
| 896 |
-
"loss":
|
| 897 |
"step": 148
|
| 898 |
},
|
| 899 |
{
|
| 900 |
-
"epoch": 0.
|
| 901 |
-
"learning_rate": 0.
|
| 902 |
-
"loss":
|
| 903 |
"step": 149
|
| 904 |
},
|
| 905 |
{
|
| 906 |
-
"epoch": 0.
|
| 907 |
-
"learning_rate": 0.
|
| 908 |
-
"loss":
|
| 909 |
"step": 150
|
| 910 |
},
|
| 911 |
{
|
| 912 |
-
"epoch": 0.
|
| 913 |
-
"learning_rate": 0.
|
| 914 |
-
"loss":
|
| 915 |
"step": 151
|
| 916 |
},
|
| 917 |
{
|
| 918 |
-
"epoch": 0.
|
| 919 |
-
"learning_rate": 0.
|
| 920 |
-
"loss":
|
| 921 |
"step": 152
|
| 922 |
},
|
| 923 |
{
|
| 924 |
-
"epoch": 0.
|
| 925 |
-
"learning_rate": 0.
|
| 926 |
-
"loss":
|
| 927 |
"step": 153
|
| 928 |
},
|
| 929 |
{
|
| 930 |
-
"epoch": 0.
|
| 931 |
-
"learning_rate": 0.
|
| 932 |
-
"loss":
|
| 933 |
"step": 154
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch": 0.
|
| 937 |
-
"learning_rate": 0.
|
| 938 |
-
"loss":
|
| 939 |
"step": 155
|
| 940 |
},
|
| 941 |
{
|
| 942 |
-
"epoch": 0.
|
| 943 |
-
"learning_rate": 0.
|
| 944 |
-
"loss":
|
| 945 |
"step": 156
|
| 946 |
},
|
| 947 |
{
|
| 948 |
-
"epoch": 0.
|
| 949 |
-
"learning_rate": 0.
|
| 950 |
-
"loss":
|
| 951 |
"step": 157
|
| 952 |
},
|
| 953 |
{
|
| 954 |
-
"epoch": 0.
|
| 955 |
-
"learning_rate": 0.
|
| 956 |
-
"loss":
|
| 957 |
"step": 158
|
| 958 |
},
|
| 959 |
{
|
| 960 |
-
"epoch": 0.
|
| 961 |
-
"learning_rate": 0.
|
| 962 |
-
"loss":
|
| 963 |
"step": 159
|
| 964 |
},
|
| 965 |
{
|
| 966 |
-
"epoch": 0.
|
| 967 |
-
"learning_rate": 0.
|
| 968 |
-
"loss":
|
| 969 |
"step": 160
|
| 970 |
},
|
| 971 |
{
|
| 972 |
-
"epoch": 0.
|
| 973 |
-
"learning_rate":
|
| 974 |
-
"loss":
|
| 975 |
"step": 161
|
| 976 |
},
|
| 977 |
{
|
| 978 |
-
"epoch": 0.
|
| 979 |
-
"learning_rate":
|
| 980 |
-
"loss":
|
| 981 |
"step": 162
|
| 982 |
},
|
| 983 |
{
|
| 984 |
-
"epoch": 0.
|
| 985 |
-
"learning_rate":
|
| 986 |
-
"loss":
|
| 987 |
"step": 163
|
| 988 |
},
|
| 989 |
{
|
| 990 |
-
"epoch": 0.
|
| 991 |
-
"learning_rate":
|
| 992 |
-
"loss":
|
| 993 |
"step": 164
|
| 994 |
},
|
| 995 |
{
|
| 996 |
-
"epoch": 0.
|
| 997 |
-
"learning_rate":
|
| 998 |
-
"loss":
|
| 999 |
"step": 165
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
-
"epoch": 0.
|
| 1003 |
-
"learning_rate":
|
| 1004 |
-
"loss":
|
| 1005 |
"step": 166
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
-
"epoch": 0.
|
| 1009 |
-
"learning_rate":
|
| 1010 |
-
"loss":
|
| 1011 |
"step": 167
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
-
"epoch": 0.
|
| 1015 |
-
"learning_rate":
|
| 1016 |
-
"loss":
|
| 1017 |
"step": 168
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
-
"epoch": 0.
|
| 1021 |
-
"learning_rate":
|
| 1022 |
-
"loss":
|
| 1023 |
"step": 169
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
-
"epoch": 0.
|
| 1027 |
-
"learning_rate":
|
| 1028 |
-
"loss":
|
| 1029 |
"step": 170
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
-
"epoch": 0.
|
| 1033 |
-
"learning_rate":
|
| 1034 |
-
"loss":
|
| 1035 |
"step": 171
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
-
"epoch": 0.
|
| 1039 |
-
"learning_rate":
|
| 1040 |
-
"loss":
|
| 1041 |
"step": 172
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
-
"epoch": 0.
|
| 1045 |
-
"learning_rate":
|
| 1046 |
-
"loss":
|
| 1047 |
"step": 173
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
-
"epoch": 0.
|
| 1051 |
-
"learning_rate":
|
| 1052 |
-
"loss":
|
| 1053 |
"step": 174
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
-
"epoch": 0.
|
| 1057 |
-
"learning_rate":
|
| 1058 |
-
"loss":
|
| 1059 |
"step": 175
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
-
"epoch": 0.
|
| 1063 |
-
"learning_rate":
|
| 1064 |
-
"loss":
|
| 1065 |
"step": 176
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
-
"epoch": 0.
|
| 1069 |
-
"learning_rate":
|
| 1070 |
-
"loss":
|
| 1071 |
"step": 177
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
-
"epoch": 0.
|
| 1075 |
-
"learning_rate":
|
| 1076 |
-
"loss":
|
| 1077 |
"step": 178
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
-
"epoch": 0.
|
| 1081 |
-
"learning_rate":
|
| 1082 |
-
"loss":
|
| 1083 |
"step": 179
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
-
"epoch": 0.
|
| 1087 |
-
"learning_rate":
|
| 1088 |
-
"loss":
|
| 1089 |
"step": 180
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
-
"epoch": 0.
|
| 1093 |
-
"learning_rate":
|
| 1094 |
-
"loss":
|
| 1095 |
"step": 181
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
-
"epoch": 0.
|
| 1099 |
-
"learning_rate":
|
| 1100 |
-
"loss":
|
| 1101 |
"step": 182
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
-
"epoch": 0.
|
| 1105 |
-
"learning_rate":
|
| 1106 |
-
"loss":
|
| 1107 |
"step": 183
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
-
"epoch": 0.
|
| 1111 |
-
"learning_rate":
|
| 1112 |
-
"loss":
|
| 1113 |
"step": 184
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
-
"epoch": 0.
|
| 1117 |
-
"learning_rate":
|
| 1118 |
-
"loss":
|
| 1119 |
"step": 185
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
-
"epoch": 0.
|
| 1123 |
-
"learning_rate":
|
| 1124 |
-
"loss":
|
| 1125 |
"step": 186
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
-
"epoch": 0.
|
| 1129 |
-
"learning_rate":
|
| 1130 |
-
"loss":
|
| 1131 |
"step": 187
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
-
"epoch": 0.
|
| 1135 |
-
"learning_rate":
|
| 1136 |
-
"loss":
|
| 1137 |
"step": 188
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
-
"epoch": 0.
|
| 1141 |
-
"learning_rate":
|
| 1142 |
-
"loss":
|
| 1143 |
"step": 189
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
-
"epoch": 0.
|
| 1147 |
-
"learning_rate":
|
| 1148 |
-
"loss":
|
| 1149 |
"step": 190
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
-
"epoch": 0.
|
| 1153 |
-
"learning_rate":
|
| 1154 |
-
"loss":
|
| 1155 |
"step": 191
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
-
"epoch": 0.
|
| 1159 |
-
"learning_rate":
|
| 1160 |
-
"loss":
|
| 1161 |
"step": 192
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
-
"epoch": 0.
|
| 1165 |
-
"learning_rate":
|
| 1166 |
-
"loss":
|
| 1167 |
"step": 193
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
-
"epoch": 0.
|
| 1171 |
-
"learning_rate":
|
| 1172 |
-
"loss":
|
| 1173 |
"step": 194
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
-
"epoch": 0.
|
| 1177 |
-
"learning_rate":
|
| 1178 |
-
"loss":
|
| 1179 |
"step": 195
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
-
"epoch": 0.
|
| 1183 |
-
"learning_rate":
|
| 1184 |
-
"loss":
|
| 1185 |
"step": 196
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
-
"epoch": 0.
|
| 1189 |
-
"learning_rate":
|
| 1190 |
-
"loss":
|
| 1191 |
"step": 197
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
-
"epoch": 0.
|
| 1195 |
-
"learning_rate":
|
| 1196 |
-
"loss":
|
| 1197 |
"step": 198
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
-
"epoch": 0.
|
| 1201 |
-
"learning_rate":
|
| 1202 |
-
"loss":
|
| 1203 |
"step": 199
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
-
"epoch": 0.
|
| 1207 |
-
"learning_rate":
|
| 1208 |
-
"loss":
|
| 1209 |
"step": 200
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
-
"epoch": 0.
|
| 1213 |
-
"learning_rate":
|
| 1214 |
-
"loss":
|
| 1215 |
"step": 201
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
-
"epoch": 0.
|
| 1219 |
-
"learning_rate":
|
| 1220 |
-
"loss":
|
| 1221 |
"step": 202
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
-
"epoch": 0.
|
| 1225 |
-
"learning_rate":
|
| 1226 |
-
"loss":
|
| 1227 |
"step": 203
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
-
"epoch": 0.
|
| 1231 |
-
"learning_rate":
|
| 1232 |
-
"loss":
|
| 1233 |
"step": 204
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
-
"epoch": 0.
|
| 1237 |
-
"learning_rate":
|
| 1238 |
-
"loss":
|
| 1239 |
"step": 205
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
-
"epoch": 0.
|
| 1243 |
-
"learning_rate":
|
| 1244 |
-
"loss":
|
| 1245 |
"step": 206
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
-
"epoch": 0.
|
| 1249 |
-
"learning_rate":
|
| 1250 |
-
"loss":
|
| 1251 |
"step": 207
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
-
"epoch": 0.
|
| 1255 |
-
"learning_rate":
|
| 1256 |
-
"loss":
|
| 1257 |
"step": 208
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
-
"epoch": 0.
|
| 1261 |
-
"learning_rate":
|
| 1262 |
-
"loss":
|
| 1263 |
"step": 209
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
-
"epoch": 0.
|
| 1267 |
-
"learning_rate":
|
| 1268 |
-
"loss":
|
| 1269 |
"step": 210
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
-
"epoch": 0.
|
| 1273 |
-
"learning_rate":
|
| 1274 |
-
"loss":
|
| 1275 |
"step": 211
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
-
"epoch": 0.
|
| 1279 |
-
"learning_rate":
|
| 1280 |
-
"loss":
|
| 1281 |
"step": 212
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
-
"epoch": 0.
|
| 1285 |
-
"learning_rate":
|
| 1286 |
-
"loss":
|
| 1287 |
"step": 213
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
-
"epoch": 0.
|
| 1291 |
-
"learning_rate":
|
| 1292 |
-
"loss":
|
| 1293 |
"step": 214
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
-
"epoch": 0.
|
| 1297 |
-
"learning_rate":
|
| 1298 |
-
"loss":
|
| 1299 |
"step": 215
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
-
"epoch": 0.
|
| 1303 |
-
"learning_rate":
|
| 1304 |
-
"loss":
|
| 1305 |
"step": 216
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
-
"epoch": 0.
|
| 1309 |
-
"learning_rate":
|
| 1310 |
-
"loss":
|
| 1311 |
"step": 217
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
-
"epoch": 0.
|
| 1315 |
-
"learning_rate":
|
| 1316 |
-
"loss":
|
| 1317 |
"step": 218
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
-
"epoch": 0.
|
| 1321 |
-
"learning_rate":
|
| 1322 |
-
"loss":
|
| 1323 |
"step": 219
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
-
"epoch": 0.
|
| 1327 |
-
"learning_rate":
|
| 1328 |
-
"loss":
|
| 1329 |
"step": 220
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
-
"epoch": 0.
|
| 1333 |
-
"learning_rate":
|
| 1334 |
-
"loss":
|
| 1335 |
"step": 221
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
-
"epoch": 0.
|
| 1339 |
-
"learning_rate":
|
| 1340 |
-
"loss":
|
| 1341 |
"step": 222
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
-
"epoch": 0.
|
| 1345 |
-
"learning_rate":
|
| 1346 |
-
"loss":
|
| 1347 |
"step": 223
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
-
"epoch": 0.
|
| 1351 |
-
"learning_rate":
|
| 1352 |
-
"loss":
|
| 1353 |
"step": 224
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
-
"epoch": 0.
|
| 1357 |
-
"learning_rate":
|
| 1358 |
-
"loss":
|
| 1359 |
"step": 225
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
-
"epoch": 0.
|
| 1363 |
-
"learning_rate":
|
| 1364 |
-
"loss":
|
| 1365 |
"step": 226
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
-
"epoch": 0.
|
| 1369 |
-
"learning_rate":
|
| 1370 |
-
"loss":
|
| 1371 |
"step": 227
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
-
"epoch": 0.
|
| 1375 |
-
"learning_rate":
|
| 1376 |
-
"loss":
|
| 1377 |
"step": 228
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
-
"epoch": 0.
|
| 1381 |
-
"learning_rate":
|
| 1382 |
-
"loss":
|
| 1383 |
"step": 229
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
-
"epoch": 0.
|
| 1387 |
-
"learning_rate":
|
| 1388 |
-
"loss":
|
| 1389 |
"step": 230
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
-
"epoch": 0.
|
| 1393 |
-
"learning_rate":
|
| 1394 |
-
"loss":
|
| 1395 |
"step": 231
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
-
"epoch": 0.
|
| 1399 |
-
"learning_rate":
|
| 1400 |
-
"loss":
|
| 1401 |
"step": 232
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
-
"epoch": 0.
|
| 1405 |
-
"learning_rate":
|
| 1406 |
-
"loss":
|
| 1407 |
"step": 233
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
-
"epoch": 0.
|
| 1411 |
-
"learning_rate":
|
| 1412 |
-
"loss":
|
| 1413 |
"step": 234
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
-
"epoch": 0.
|
| 1417 |
-
"learning_rate":
|
| 1418 |
-
"loss":
|
| 1419 |
"step": 235
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
-
"epoch": 0.
|
| 1423 |
-
"learning_rate":
|
| 1424 |
-
"loss":
|
| 1425 |
"step": 236
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
-
"epoch": 0.
|
| 1429 |
-
"learning_rate":
|
| 1430 |
-
"loss":
|
| 1431 |
"step": 237
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
-
"epoch": 0.
|
| 1435 |
-
"learning_rate":
|
| 1436 |
-
"loss":
|
| 1437 |
"step": 238
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
-
"epoch": 0.
|
| 1441 |
-
"learning_rate":
|
| 1442 |
-
"loss":
|
| 1443 |
"step": 239
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
-
"epoch": 0.
|
| 1447 |
-
"learning_rate":
|
| 1448 |
-
"loss":
|
| 1449 |
"step": 240
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
-
"epoch": 0.
|
| 1453 |
-
"learning_rate":
|
| 1454 |
-
"loss":
|
| 1455 |
"step": 241
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
-
"epoch": 0.
|
| 1459 |
-
"learning_rate":
|
| 1460 |
-
"loss":
|
| 1461 |
"step": 242
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
-
"epoch": 0.
|
| 1465 |
-
"learning_rate":
|
| 1466 |
-
"loss":
|
| 1467 |
"step": 243
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
-
"epoch": 0.
|
| 1471 |
-
"learning_rate":
|
| 1472 |
-
"loss":
|
| 1473 |
"step": 244
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
-
"epoch": 0.
|
| 1477 |
-
"learning_rate":
|
| 1478 |
-
"loss":
|
| 1479 |
"step": 245
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
-
"epoch": 0.
|
| 1483 |
-
"learning_rate":
|
| 1484 |
-
"loss":
|
| 1485 |
"step": 246
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
-
"epoch": 0.
|
| 1489 |
-
"learning_rate":
|
| 1490 |
-
"loss":
|
| 1491 |
"step": 247
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
-
"epoch": 0.
|
| 1495 |
-
"learning_rate":
|
| 1496 |
-
"loss":
|
| 1497 |
"step": 248
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
-
"epoch": 0.
|
| 1501 |
-
"learning_rate":
|
| 1502 |
-
"loss":
|
| 1503 |
"step": 249
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
-
"epoch": 0.
|
| 1507 |
-
"learning_rate":
|
| 1508 |
-
"loss":
|
| 1509 |
"step": 250
|
| 1510 |
},
|
| 1511 |
{
|
| 1512 |
-
"epoch": 0.
|
| 1513 |
-
"learning_rate":
|
| 1514 |
-
"loss":
|
| 1515 |
"step": 251
|
| 1516 |
},
|
| 1517 |
{
|
| 1518 |
-
"epoch": 0.
|
| 1519 |
-
"learning_rate":
|
| 1520 |
-
"loss":
|
| 1521 |
"step": 252
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
-
"epoch": 0.
|
| 1525 |
-
"learning_rate":
|
| 1526 |
-
"loss":
|
| 1527 |
"step": 253
|
| 1528 |
},
|
| 1529 |
{
|
| 1530 |
-
"epoch": 0.
|
| 1531 |
-
"learning_rate":
|
| 1532 |
-
"loss":
|
| 1533 |
"step": 254
|
| 1534 |
},
|
| 1535 |
{
|
| 1536 |
-
"epoch": 0.
|
| 1537 |
-
"learning_rate":
|
| 1538 |
-
"loss":
|
| 1539 |
"step": 255
|
| 1540 |
},
|
| 1541 |
{
|
| 1542 |
-
"epoch": 0.
|
| 1543 |
-
"learning_rate":
|
| 1544 |
-
"loss":
|
| 1545 |
"step": 256
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
-
"epoch": 0.
|
| 1549 |
-
"learning_rate":
|
| 1550 |
-
"loss":
|
| 1551 |
"step": 257
|
| 1552 |
},
|
| 1553 |
{
|
| 1554 |
-
"epoch": 0.
|
| 1555 |
-
"learning_rate":
|
| 1556 |
-
"loss":
|
| 1557 |
"step": 258
|
| 1558 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1559 |
{
|
| 1560 |
"epoch": 0.84,
|
| 1561 |
-
"learning_rate": 1.
|
| 1562 |
-
"loss":
|
| 1563 |
-
"step":
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 0.84,
|
| 1567 |
-
"learning_rate": 1.
|
| 1568 |
-
"loss":
|
| 1569 |
-
"step":
|
| 1570 |
},
|
| 1571 |
{
|
| 1572 |
"epoch": 0.84,
|
| 1573 |
-
"learning_rate": 1.
|
| 1574 |
-
"loss":
|
| 1575 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1576 |
},
|
| 1577 |
{
|
| 1578 |
"epoch": 0.85,
|
| 1579 |
-
"learning_rate": 1.
|
| 1580 |
-
"loss":
|
| 1581 |
-
"step":
|
| 1582 |
},
|
| 1583 |
{
|
| 1584 |
"epoch": 0.85,
|
| 1585 |
-
"learning_rate": 1.
|
| 1586 |
-
"loss":
|
| 1587 |
-
"step":
|
| 1588 |
},
|
| 1589 |
{
|
| 1590 |
"epoch": 0.85,
|
| 1591 |
-
"learning_rate": 1.
|
| 1592 |
-
"loss":
|
| 1593 |
-
"step":
|
| 1594 |
},
|
| 1595 |
{
|
| 1596 |
"epoch": 0.85,
|
| 1597 |
-
"learning_rate": 1.
|
| 1598 |
-
"loss":
|
| 1599 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1600 |
},
|
| 1601 |
{
|
| 1602 |
"epoch": 0.86,
|
| 1603 |
-
"learning_rate": 1.
|
| 1604 |
-
"loss":
|
| 1605 |
-
"step":
|
| 1606 |
},
|
| 1607 |
{
|
| 1608 |
"epoch": 0.86,
|
| 1609 |
-
"learning_rate":
|
| 1610 |
-
"loss":
|
| 1611 |
-
"step":
|
| 1612 |
},
|
| 1613 |
{
|
| 1614 |
"epoch": 0.86,
|
| 1615 |
-
"learning_rate":
|
| 1616 |
-
"loss":
|
| 1617 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1618 |
},
|
| 1619 |
{
|
| 1620 |
"epoch": 0.87,
|
| 1621 |
-
"learning_rate": 9.
|
| 1622 |
-
"loss":
|
| 1623 |
-
"step":
|
| 1624 |
},
|
| 1625 |
{
|
| 1626 |
"epoch": 0.87,
|
| 1627 |
-
"learning_rate": 8.
|
| 1628 |
-
"loss":
|
| 1629 |
-
"step":
|
| 1630 |
},
|
| 1631 |
{
|
| 1632 |
"epoch": 0.87,
|
| 1633 |
-
"learning_rate": 8.
|
| 1634 |
-
"loss":
|
| 1635 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1636 |
},
|
| 1637 |
{
|
| 1638 |
"epoch": 0.88,
|
| 1639 |
-
"learning_rate":
|
| 1640 |
-
"loss":
|
| 1641 |
-
"step":
|
| 1642 |
},
|
| 1643 |
{
|
| 1644 |
"epoch": 0.88,
|
| 1645 |
-
"learning_rate": 7.
|
| 1646 |
-
"loss":
|
| 1647 |
-
"step":
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
"epoch": 0.88,
|
| 1651 |
-
"learning_rate": 7.
|
| 1652 |
-
"loss":
|
| 1653 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1654 |
},
|
| 1655 |
{
|
| 1656 |
"epoch": 0.89,
|
| 1657 |
-
"learning_rate": 6.
|
| 1658 |
-
"loss":
|
| 1659 |
-
"step":
|
| 1660 |
},
|
| 1661 |
{
|
| 1662 |
"epoch": 0.89,
|
| 1663 |
-
"learning_rate": 6.
|
| 1664 |
-
"loss":
|
| 1665 |
-
"step":
|
| 1666 |
},
|
| 1667 |
{
|
| 1668 |
"epoch": 0.89,
|
| 1669 |
-
"learning_rate":
|
| 1670 |
-
"loss":
|
| 1671 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1672 |
},
|
| 1673 |
{
|
| 1674 |
"epoch": 0.9,
|
| 1675 |
-
"learning_rate": 5.
|
| 1676 |
-
"loss":
|
| 1677 |
-
"step":
|
| 1678 |
},
|
| 1679 |
{
|
| 1680 |
"epoch": 0.9,
|
| 1681 |
-
"learning_rate": 5.
|
| 1682 |
-
"loss":
|
| 1683 |
-
"step":
|
| 1684 |
},
|
| 1685 |
{
|
| 1686 |
"epoch": 0.9,
|
| 1687 |
-
"learning_rate":
|
| 1688 |
-
"loss":
|
| 1689 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1690 |
},
|
| 1691 |
{
|
| 1692 |
"epoch": 0.91,
|
| 1693 |
-
"learning_rate": 4.
|
| 1694 |
-
"loss":
|
| 1695 |
-
"step":
|
| 1696 |
},
|
| 1697 |
{
|
| 1698 |
"epoch": 0.91,
|
| 1699 |
-
"learning_rate": 4.
|
| 1700 |
-
"loss":
|
| 1701 |
-
"step":
|
| 1702 |
},
|
| 1703 |
{
|
| 1704 |
"epoch": 0.91,
|
| 1705 |
-
"learning_rate":
|
| 1706 |
-
"loss":
|
| 1707 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1708 |
},
|
| 1709 |
{
|
| 1710 |
"epoch": 0.92,
|
| 1711 |
-
"learning_rate": 3.
|
| 1712 |
-
"loss":
|
| 1713 |
-
"step":
|
| 1714 |
},
|
| 1715 |
{
|
| 1716 |
"epoch": 0.92,
|
| 1717 |
-
"learning_rate": 3.
|
| 1718 |
-
"loss":
|
| 1719 |
-
"step":
|
| 1720 |
},
|
| 1721 |
{
|
| 1722 |
"epoch": 0.92,
|
| 1723 |
-
"learning_rate": 3.
|
| 1724 |
-
"loss":
|
| 1725 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1726 |
},
|
| 1727 |
{
|
| 1728 |
"epoch": 0.93,
|
| 1729 |
-
"learning_rate": 2.
|
| 1730 |
-
"loss":
|
| 1731 |
-
"step":
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
"epoch": 0.93,
|
| 1735 |
-
"learning_rate": 2.
|
| 1736 |
-
"loss":
|
| 1737 |
-
"step":
|
| 1738 |
},
|
| 1739 |
{
|
| 1740 |
"epoch": 0.93,
|
| 1741 |
-
"learning_rate": 2.
|
| 1742 |
-
"loss":
|
| 1743 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1744 |
},
|
| 1745 |
{
|
| 1746 |
"epoch": 0.94,
|
| 1747 |
-
"learning_rate": 2.
|
| 1748 |
-
"loss":
|
| 1749 |
-
"step":
|
| 1750 |
},
|
| 1751 |
{
|
| 1752 |
"epoch": 0.94,
|
| 1753 |
-
"learning_rate":
|
| 1754 |
-
"loss":
|
| 1755 |
-
"step":
|
| 1756 |
},
|
| 1757 |
{
|
| 1758 |
"epoch": 0.94,
|
| 1759 |
-
"learning_rate": 1.
|
| 1760 |
-
"loss":
|
| 1761 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1762 |
},
|
| 1763 |
{
|
| 1764 |
"epoch": 0.95,
|
| 1765 |
-
"learning_rate": 1.
|
| 1766 |
-
"loss":
|
| 1767 |
-
"step":
|
| 1768 |
},
|
| 1769 |
{
|
| 1770 |
"epoch": 0.95,
|
| 1771 |
-
"learning_rate": 1.
|
| 1772 |
-
"loss":
|
| 1773 |
-
"step":
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
"epoch": 0.95,
|
| 1777 |
-
"learning_rate": 1.
|
| 1778 |
-
"loss":
|
| 1779 |
-
"step":
|
| 1780 |
},
|
| 1781 |
{
|
| 1782 |
"epoch": 0.95,
|
| 1783 |
-
"learning_rate": 1.
|
| 1784 |
-
"loss":
|
| 1785 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1786 |
},
|
| 1787 |
{
|
| 1788 |
"epoch": 0.96,
|
| 1789 |
-
"learning_rate": 9.
|
| 1790 |
-
"loss":
|
| 1791 |
-
"step":
|
| 1792 |
},
|
| 1793 |
{
|
| 1794 |
"epoch": 0.96,
|
| 1795 |
-
"learning_rate":
|
| 1796 |
-
"loss":
|
| 1797 |
-
"step":
|
| 1798 |
},
|
| 1799 |
{
|
| 1800 |
"epoch": 0.96,
|
| 1801 |
-
"learning_rate":
|
| 1802 |
-
"loss":
|
| 1803 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1804 |
},
|
| 1805 |
{
|
| 1806 |
"epoch": 0.97,
|
| 1807 |
-
"learning_rate":
|
| 1808 |
-
"loss":
|
| 1809 |
-
"step":
|
| 1810 |
},
|
| 1811 |
{
|
| 1812 |
"epoch": 0.97,
|
| 1813 |
-
"learning_rate":
|
| 1814 |
-
"loss":
|
| 1815 |
-
"step":
|
| 1816 |
},
|
| 1817 |
{
|
| 1818 |
"epoch": 0.97,
|
| 1819 |
-
"learning_rate":
|
| 1820 |
-
"loss":
|
| 1821 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1822 |
},
|
| 1823 |
{
|
| 1824 |
"epoch": 0.98,
|
| 1825 |
-
"learning_rate": 2.
|
| 1826 |
-
"loss":
|
| 1827 |
-
"step":
|
| 1828 |
},
|
| 1829 |
{
|
| 1830 |
"epoch": 0.98,
|
| 1831 |
-
"learning_rate":
|
| 1832 |
-
"loss":
|
| 1833 |
-
"step":
|
| 1834 |
},
|
| 1835 |
{
|
| 1836 |
"epoch": 0.98,
|
| 1837 |
-
"learning_rate": 1.
|
| 1838 |
-
"loss":
|
| 1839 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1840 |
},
|
| 1841 |
{
|
| 1842 |
"epoch": 0.99,
|
| 1843 |
-
"learning_rate": 8.
|
| 1844 |
-
"loss":
|
| 1845 |
-
"step":
|
| 1846 |
},
|
| 1847 |
{
|
| 1848 |
"epoch": 0.99,
|
| 1849 |
-
"learning_rate":
|
| 1850 |
-
"loss":
|
| 1851 |
-
"step":
|
| 1852 |
},
|
| 1853 |
{
|
| 1854 |
"epoch": 0.99,
|
| 1855 |
-
"learning_rate":
|
| 1856 |
-
"loss":
|
| 1857 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1858 |
},
|
| 1859 |
{
|
| 1860 |
"epoch": 1.0,
|
| 1861 |
-
"learning_rate":
|
| 1862 |
-
"loss":
|
| 1863 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1864 |
},
|
| 1865 |
{
|
| 1866 |
"epoch": 1.0,
|
| 1867 |
"learning_rate": 0.0,
|
| 1868 |
-
"loss":
|
| 1869 |
-
"step":
|
| 1870 |
},
|
| 1871 |
{
|
| 1872 |
"epoch": 1.0,
|
| 1873 |
-
"step":
|
| 1874 |
-
"total_flos":
|
| 1875 |
-
"train_loss":
|
| 1876 |
-
"train_runtime":
|
| 1877 |
-
"train_samples_per_second":
|
| 1878 |
-
"train_steps_per_second": 0.
|
| 1879 |
}
|
| 1880 |
],
|
| 1881 |
"logging_steps": 1.0,
|
| 1882 |
-
"max_steps":
|
| 1883 |
"num_input_tokens_seen": 0,
|
| 1884 |
"num_train_epochs": 1,
|
| 1885 |
"save_steps": 50000,
|
| 1886 |
-
"total_flos":
|
| 1887 |
"train_batch_size": 16,
|
| 1888 |
"trial_name": null,
|
| 1889 |
"trial_params": null
|
|
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 465,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 14 |
+
"loss": 1.6573,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.0,
|
| 19 |
+
"learning_rate": 2.857142857142857e-05,
|
| 20 |
+
"loss": 1.6856,
|
| 21 |
"step": 2
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 26 |
+
"loss": 1.6195,
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.01,
|
| 31 |
+
"learning_rate": 5.714285714285714e-05,
|
| 32 |
+
"loss": 1.4941,
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.01,
|
| 37 |
+
"learning_rate": 7.142857142857143e-05,
|
| 38 |
+
"loss": 1.3222,
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.01,
|
| 43 |
+
"learning_rate": 8.571428571428571e-05,
|
| 44 |
+
"loss": 1.2348,
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.02,
|
| 49 |
+
"learning_rate": 0.0001,
|
| 50 |
+
"loss": 1.2121,
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 0.00011428571428571428,
|
| 56 |
+
"loss": 1.1245,
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.02,
|
| 61 |
+
"learning_rate": 0.00012857142857142858,
|
| 62 |
+
"loss": 1.1698,
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.02,
|
| 67 |
+
"learning_rate": 0.00014285714285714287,
|
| 68 |
+
"loss": 1.1782,
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.02,
|
| 73 |
+
"learning_rate": 0.00015714285714285716,
|
| 74 |
+
"loss": 1.0427,
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.03,
|
| 79 |
+
"learning_rate": 0.00017142857142857143,
|
| 80 |
+
"loss": 1.0918,
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.03,
|
| 85 |
+
"learning_rate": 0.00018571428571428572,
|
| 86 |
+
"loss": 1.0289,
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.03,
|
| 91 |
+
"learning_rate": 0.0002,
|
| 92 |
+
"loss": 0.9759,
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"learning_rate": 0.00019999757386529745,
|
| 98 |
+
"loss": 0.9791,
|
| 99 |
"step": 15
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.03,
|
| 103 |
+
"learning_rate": 0.0001999902955789124,
|
| 104 |
+
"loss": 0.9935,
|
| 105 |
"step": 16
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 0.04,
|
| 109 |
+
"learning_rate": 0.0001999781654940069,
|
| 110 |
+
"loss": 0.9983,
|
| 111 |
"step": 17
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.04,
|
| 115 |
+
"learning_rate": 0.00019996118419916533,
|
| 116 |
+
"loss": 0.9828,
|
| 117 |
"step": 18
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.04,
|
| 121 |
+
"learning_rate": 0.0001999393525183659,
|
| 122 |
+
"loss": 1.0053,
|
| 123 |
"step": 19
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 0.04,
|
| 127 |
+
"learning_rate": 0.00019991267151094053,
|
| 128 |
+
"loss": 1.0188,
|
| 129 |
"step": 20
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 0.05,
|
| 133 |
+
"learning_rate": 0.00019988114247152363,
|
| 134 |
+
"loss": 1.011,
|
| 135 |
"step": 21
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.05,
|
| 139 |
+
"learning_rate": 0.00019984476692998913,
|
| 140 |
+
"loss": 0.9739,
|
| 141 |
"step": 22
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 0.05,
|
| 145 |
+
"learning_rate": 0.00019980354665137626,
|
| 146 |
+
"loss": 0.9957,
|
| 147 |
"step": 23
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 0.05,
|
| 151 |
+
"learning_rate": 0.00019975748363580404,
|
| 152 |
+
"loss": 0.942,
|
| 153 |
"step": 24
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.05,
|
| 157 |
+
"learning_rate": 0.00019970658011837404,
|
| 158 |
+
"loss": 0.9644,
|
| 159 |
"step": 25
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.06,
|
| 163 |
+
"learning_rate": 0.0001996508385690621,
|
| 164 |
+
"loss": 0.9272,
|
| 165 |
"step": 26
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.06,
|
| 169 |
+
"learning_rate": 0.00019959026169259834,
|
| 170 |
+
"loss": 0.956,
|
| 171 |
"step": 27
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.06,
|
| 175 |
+
"learning_rate": 0.000199524852428336,
|
| 176 |
+
"loss": 0.9735,
|
| 177 |
"step": 28
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.06,
|
| 181 |
+
"learning_rate": 0.00019945461395010883,
|
| 182 |
+
"loss": 1.0073,
|
| 183 |
"step": 29
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.06,
|
| 187 |
+
"learning_rate": 0.00019937954966607694,
|
| 188 |
+
"loss": 0.9195,
|
| 189 |
"step": 30
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.07,
|
| 193 |
+
"learning_rate": 0.0001992996632185617,
|
| 194 |
+
"loss": 0.9223,
|
| 195 |
"step": 31
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 0.07,
|
| 199 |
+
"learning_rate": 0.00019921495848386876,
|
| 200 |
+
"loss": 0.9765,
|
| 201 |
"step": 32
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 0.07,
|
| 205 |
+
"learning_rate": 0.0001991254395721,
|
| 206 |
+
"loss": 0.9753,
|
| 207 |
"step": 33
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.07,
|
| 211 |
+
"learning_rate": 0.0001990311108269542,
|
| 212 |
+
"loss": 0.9585,
|
| 213 |
"step": 34
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 0.08,
|
| 217 |
+
"learning_rate": 0.00019893197682551624,
|
| 218 |
+
"loss": 0.9614,
|
| 219 |
"step": 35
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.08,
|
| 223 |
+
"learning_rate": 0.00019882804237803488,
|
| 224 |
+
"loss": 0.9492,
|
| 225 |
"step": 36
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 0.08,
|
| 229 |
+
"learning_rate": 0.00019871931252768955,
|
| 230 |
+
"loss": 0.969,
|
| 231 |
"step": 37
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 0.08,
|
| 235 |
+
"learning_rate": 0.0001986057925503455,
|
| 236 |
+
"loss": 0.9452,
|
| 237 |
"step": 38
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 0.08,
|
| 241 |
+
"learning_rate": 0.00019848748795429788,
|
| 242 |
+
"loss": 0.9317,
|
| 243 |
"step": 39
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 0.09,
|
| 247 |
+
"learning_rate": 0.0001983644044800044,
|
| 248 |
+
"loss": 0.8749,
|
| 249 |
"step": 40
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.09,
|
| 253 |
+
"learning_rate": 0.00019823654809980682,
|
| 254 |
+
"loss": 0.8208,
|
| 255 |
"step": 41
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 0.09,
|
| 259 |
+
"learning_rate": 0.00019810392501764116,
|
| 260 |
+
"loss": 0.8632,
|
| 261 |
"step": 42
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.09,
|
| 265 |
+
"learning_rate": 0.00019796654166873666,
|
| 266 |
+
"loss": 0.9159,
|
| 267 |
"step": 43
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 0.09,
|
| 271 |
+
"learning_rate": 0.00019782440471930354,
|
| 272 |
+
"loss": 0.9201,
|
| 273 |
"step": 44
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 0.1,
|
| 277 |
+
"learning_rate": 0.0001976775210662095,
|
| 278 |
+
"loss": 0.957,
|
| 279 |
"step": 45
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 0.1,
|
| 283 |
+
"learning_rate": 0.0001975258978366451,
|
| 284 |
+
"loss": 0.9268,
|
| 285 |
"step": 46
|
| 286 |
},
|
| 287 |
{
|
| 288 |
+
"epoch": 0.1,
|
| 289 |
+
"learning_rate": 0.00019736954238777792,
|
| 290 |
+
"loss": 0.9516,
|
| 291 |
"step": 47
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.1,
|
| 295 |
+
"learning_rate": 0.00019720846230639557,
|
| 296 |
+
"loss": 0.9008,
|
| 297 |
"step": 48
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 0.11,
|
| 301 |
+
"learning_rate": 0.00019704266540853752,
|
| 302 |
+
"loss": 0.8722,
|
| 303 |
"step": 49
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 0.11,
|
| 307 |
+
"learning_rate": 0.00019687215973911599,
|
| 308 |
+
"loss": 0.9384,
|
| 309 |
"step": 50
|
| 310 |
},
|
| 311 |
{
|
| 312 |
+
"epoch": 0.11,
|
| 313 |
+
"learning_rate": 0.00019669695357152537,
|
| 314 |
+
"loss": 0.8853,
|
| 315 |
"step": 51
|
| 316 |
},
|
| 317 |
{
|
| 318 |
+
"epoch": 0.11,
|
| 319 |
+
"learning_rate": 0.00019651705540724092,
|
| 320 |
+
"loss": 0.9246,
|
| 321 |
"step": 52
|
| 322 |
},
|
| 323 |
{
|
| 324 |
+
"epoch": 0.11,
|
| 325 |
+
"learning_rate": 0.00019633247397540625,
|
| 326 |
+
"loss": 0.8659,
|
| 327 |
"step": 53
|
| 328 |
},
|
| 329 |
{
|
| 330 |
+
"epoch": 0.12,
|
| 331 |
+
"learning_rate": 0.00019614321823240968,
|
| 332 |
+
"loss": 0.8708,
|
| 333 |
"step": 54
|
| 334 |
},
|
| 335 |
{
|
| 336 |
+
"epoch": 0.12,
|
| 337 |
+
"learning_rate": 0.00019594929736144976,
|
| 338 |
+
"loss": 0.848,
|
| 339 |
"step": 55
|
| 340 |
},
|
| 341 |
{
|
| 342 |
+
"epoch": 0.12,
|
| 343 |
+
"learning_rate": 0.00019575072077208954,
|
| 344 |
+
"loss": 0.8691,
|
| 345 |
"step": 56
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 0.12,
|
| 349 |
+
"learning_rate": 0.00019554749809980014,
|
| 350 |
+
"loss": 0.8819,
|
| 351 |
"step": 57
|
| 352 |
},
|
| 353 |
{
|
| 354 |
+
"epoch": 0.12,
|
| 355 |
+
"learning_rate": 0.00019533963920549306,
|
| 356 |
+
"loss": 0.9047,
|
| 357 |
"step": 58
|
| 358 |
},
|
| 359 |
{
|
| 360 |
+
"epoch": 0.13,
|
| 361 |
+
"learning_rate": 0.0001951271541750419,
|
| 362 |
+
"loss": 0.8644,
|
| 363 |
"step": 59
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.13,
|
| 367 |
+
"learning_rate": 0.00019491005331879278,
|
| 368 |
+
"loss": 0.8869,
|
| 369 |
"step": 60
|
| 370 |
},
|
| 371 |
{
|
| 372 |
+
"epoch": 0.13,
|
| 373 |
+
"learning_rate": 0.00019468834717106408,
|
| 374 |
+
"loss": 0.8805,
|
| 375 |
"step": 61
|
| 376 |
},
|
| 377 |
{
|
| 378 |
+
"epoch": 0.13,
|
| 379 |
+
"learning_rate": 0.0001944620464896354,
|
| 380 |
+
"loss": 0.9323,
|
| 381 |
"step": 62
|
| 382 |
},
|
| 383 |
{
|
| 384 |
+
"epoch": 0.14,
|
| 385 |
+
"learning_rate": 0.00019423116225522548,
|
| 386 |
+
"loss": 0.9393,
|
| 387 |
"step": 63
|
| 388 |
},
|
| 389 |
{
|
| 390 |
+
"epoch": 0.14,
|
| 391 |
+
"learning_rate": 0.00019399570567095936,
|
| 392 |
+
"loss": 0.9303,
|
| 393 |
"step": 64
|
| 394 |
},
|
| 395 |
{
|
| 396 |
+
"epoch": 0.14,
|
| 397 |
+
"learning_rate": 0.00019375568816182485,
|
| 398 |
+
"loss": 0.8898,
|
| 399 |
"step": 65
|
| 400 |
},
|
| 401 |
{
|
| 402 |
+
"epoch": 0.14,
|
| 403 |
+
"learning_rate": 0.00019351112137411816,
|
| 404 |
+
"loss": 0.8466,
|
| 405 |
"step": 66
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.14,
|
| 409 |
+
"learning_rate": 0.00019326201717487863,
|
| 410 |
+
"loss": 0.8526,
|
| 411 |
"step": 67
|
| 412 |
},
|
| 413 |
{
|
| 414 |
+
"epoch": 0.15,
|
| 415 |
+
"learning_rate": 0.00019300838765131314,
|
| 416 |
+
"loss": 0.8938,
|
| 417 |
"step": 68
|
| 418 |
},
|
| 419 |
{
|
| 420 |
+
"epoch": 0.15,
|
| 421 |
+
"learning_rate": 0.0001927502451102095,
|
| 422 |
+
"loss": 0.8712,
|
| 423 |
"step": 69
|
| 424 |
},
|
| 425 |
{
|
| 426 |
+
"epoch": 0.15,
|
| 427 |
+
"learning_rate": 0.00019248760207733919,
|
| 428 |
+
"loss": 0.8522,
|
| 429 |
"step": 70
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 0.15,
|
| 433 |
+
"learning_rate": 0.00019222047129684974,
|
| 434 |
+
"loss": 0.8797,
|
| 435 |
"step": 71
|
| 436 |
},
|
| 437 |
{
|
| 438 |
+
"epoch": 0.15,
|
| 439 |
+
"learning_rate": 0.00019194886573064635,
|
| 440 |
+
"loss": 0.8753,
|
| 441 |
"step": 72
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 0.16,
|
| 445 |
+
"learning_rate": 0.00019167279855776274,
|
| 446 |
+
"loss": 0.8703,
|
| 447 |
"step": 73
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.16,
|
| 451 |
+
"learning_rate": 0.00019139228317372194,
|
| 452 |
+
"loss": 0.9131,
|
| 453 |
"step": 74
|
| 454 |
},
|
| 455 |
{
|
| 456 |
+
"epoch": 0.16,
|
| 457 |
+
"learning_rate": 0.00019110733318988604,
|
| 458 |
+
"loss": 0.9136,
|
| 459 |
"step": 75
|
| 460 |
},
|
| 461 |
{
|
| 462 |
+
"epoch": 0.16,
|
| 463 |
+
"learning_rate": 0.000190817962432796,
|
| 464 |
+
"loss": 0.8971,
|
| 465 |
"step": 76
|
| 466 |
},
|
| 467 |
{
|
| 468 |
+
"epoch": 0.17,
|
| 469 |
+
"learning_rate": 0.00019052418494350048,
|
| 470 |
+
"loss": 0.8285,
|
| 471 |
"step": 77
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 0.17,
|
| 475 |
+
"learning_rate": 0.0001902260149768747,
|
| 476 |
+
"loss": 0.9126,
|
| 477 |
"step": 78
|
| 478 |
},
|
| 479 |
{
|
| 480 |
+
"epoch": 0.17,
|
| 481 |
+
"learning_rate": 0.0001899234670009288,
|
| 482 |
+
"loss": 0.8018,
|
| 483 |
"step": 79
|
| 484 |
},
|
| 485 |
{
|
| 486 |
+
"epoch": 0.17,
|
| 487 |
+
"learning_rate": 0.00018961655569610557,
|
| 488 |
+
"loss": 0.9147,
|
| 489 |
"step": 80
|
| 490 |
},
|
| 491 |
{
|
| 492 |
+
"epoch": 0.17,
|
| 493 |
+
"learning_rate": 0.00018930529595456838,
|
| 494 |
+
"loss": 0.922,
|
| 495 |
"step": 81
|
| 496 |
},
|
| 497 |
{
|
| 498 |
+
"epoch": 0.18,
|
| 499 |
+
"learning_rate": 0.00018898970287947847,
|
| 500 |
+
"loss": 0.8314,
|
| 501 |
"step": 82
|
| 502 |
},
|
| 503 |
{
|
| 504 |
+
"epoch": 0.18,
|
| 505 |
+
"learning_rate": 0.00018866979178426205,
|
| 506 |
+
"loss": 0.8911,
|
| 507 |
"step": 83
|
| 508 |
},
|
| 509 |
{
|
| 510 |
+
"epoch": 0.18,
|
| 511 |
+
"learning_rate": 0.00018834557819186732,
|
| 512 |
+
"loss": 0.7828,
|
| 513 |
"step": 84
|
| 514 |
},
|
| 515 |
{
|
| 516 |
+
"epoch": 0.18,
|
| 517 |
+
"learning_rate": 0.00018801707783401122,
|
| 518 |
+
"loss": 0.8538,
|
| 519 |
"step": 85
|
| 520 |
},
|
| 521 |
{
|
| 522 |
+
"epoch": 0.18,
|
| 523 |
+
"learning_rate": 0.0001876843066504161,
|
| 524 |
+
"loss": 0.8548,
|
| 525 |
"step": 86
|
| 526 |
},
|
| 527 |
{
|
| 528 |
+
"epoch": 0.19,
|
| 529 |
+
"learning_rate": 0.0001873472807880363,
|
| 530 |
+
"loss": 0.8314,
|
| 531 |
"step": 87
|
| 532 |
},
|
| 533 |
{
|
| 534 |
+
"epoch": 0.19,
|
| 535 |
+
"learning_rate": 0.00018700601660027463,
|
| 536 |
+
"loss": 0.8404,
|
| 537 |
"step": 88
|
| 538 |
},
|
| 539 |
{
|
| 540 |
+
"epoch": 0.19,
|
| 541 |
+
"learning_rate": 0.00018666053064618886,
|
| 542 |
+
"loss": 0.8542,
|
| 543 |
"step": 89
|
| 544 |
},
|
| 545 |
{
|
| 546 |
+
"epoch": 0.19,
|
| 547 |
+
"learning_rate": 0.00018631083968968827,
|
| 548 |
+
"loss": 0.8378,
|
| 549 |
"step": 90
|
| 550 |
},
|
| 551 |
{
|
| 552 |
+
"epoch": 0.2,
|
| 553 |
+
"learning_rate": 0.00018595696069872013,
|
| 554 |
+
"loss": 0.8207,
|
| 555 |
"step": 91
|
| 556 |
},
|
| 557 |
{
|
| 558 |
+
"epoch": 0.2,
|
| 559 |
+
"learning_rate": 0.00018559891084444642,
|
| 560 |
+
"loss": 0.8595,
|
| 561 |
"step": 92
|
| 562 |
},
|
| 563 |
{
|
| 564 |
+
"epoch": 0.2,
|
| 565 |
+
"learning_rate": 0.00018523670750041071,
|
| 566 |
+
"loss": 0.8344,
|
| 567 |
"step": 93
|
| 568 |
},
|
| 569 |
{
|
| 570 |
+
"epoch": 0.2,
|
| 571 |
+
"learning_rate": 0.00018487036824169507,
|
| 572 |
+
"loss": 0.8886,
|
| 573 |
"step": 94
|
| 574 |
},
|
| 575 |
{
|
| 576 |
+
"epoch": 0.2,
|
| 577 |
+
"learning_rate": 0.0001844999108440672,
|
| 578 |
+
"loss": 0.8965,
|
| 579 |
"step": 95
|
| 580 |
},
|
| 581 |
{
|
| 582 |
+
"epoch": 0.21,
|
| 583 |
+
"learning_rate": 0.00018412535328311814,
|
| 584 |
+
"loss": 0.9179,
|
| 585 |
"step": 96
|
| 586 |
},
|
| 587 |
{
|
| 588 |
+
"epoch": 0.21,
|
| 589 |
+
"learning_rate": 0.00018374671373338973,
|
| 590 |
+
"loss": 0.8855,
|
| 591 |
"step": 97
|
| 592 |
},
|
| 593 |
{
|
| 594 |
+
"epoch": 0.21,
|
| 595 |
+
"learning_rate": 0.0001833640105674931,
|
| 596 |
+
"loss": 0.8428,
|
| 597 |
"step": 98
|
| 598 |
},
|
| 599 |
{
|
| 600 |
+
"epoch": 0.21,
|
| 601 |
+
"learning_rate": 0.00018297726235521684,
|
| 602 |
+
"loss": 0.8908,
|
| 603 |
"step": 99
|
| 604 |
},
|
| 605 |
{
|
| 606 |
+
"epoch": 0.22,
|
| 607 |
+
"learning_rate": 0.0001825864878626261,
|
| 608 |
+
"loss": 0.8838,
|
| 609 |
"step": 100
|
| 610 |
},
|
| 611 |
{
|
| 612 |
+
"epoch": 0.22,
|
| 613 |
+
"learning_rate": 0.00018219170605115206,
|
| 614 |
+
"loss": 0.839,
|
| 615 |
"step": 101
|
| 616 |
},
|
| 617 |
{
|
| 618 |
+
"epoch": 0.22,
|
| 619 |
+
"learning_rate": 0.00018179293607667178,
|
| 620 |
+
"loss": 0.8641,
|
| 621 |
"step": 102
|
| 622 |
},
|
| 623 |
{
|
| 624 |
+
"epoch": 0.22,
|
| 625 |
+
"learning_rate": 0.0001813901972885787,
|
| 626 |
+
"loss": 0.8761,
|
| 627 |
"step": 103
|
| 628 |
},
|
| 629 |
{
|
| 630 |
+
"epoch": 0.22,
|
| 631 |
+
"learning_rate": 0.00018098350922884384,
|
| 632 |
+
"loss": 0.8309,
|
| 633 |
"step": 104
|
| 634 |
},
|
| 635 |
{
|
| 636 |
+
"epoch": 0.23,
|
| 637 |
+
"learning_rate": 0.00018057289163106747,
|
| 638 |
+
"loss": 0.8708,
|
| 639 |
"step": 105
|
| 640 |
},
|
| 641 |
{
|
| 642 |
+
"epoch": 0.23,
|
| 643 |
+
"learning_rate": 0.00018015836441952172,
|
| 644 |
+
"loss": 0.8735,
|
| 645 |
"step": 106
|
| 646 |
},
|
| 647 |
{
|
| 648 |
+
"epoch": 0.23,
|
| 649 |
+
"learning_rate": 0.00017973994770818357,
|
| 650 |
+
"loss": 0.8488,
|
| 651 |
"step": 107
|
| 652 |
},
|
| 653 |
{
|
| 654 |
+
"epoch": 0.23,
|
| 655 |
+
"learning_rate": 0.00017931766179975914,
|
| 656 |
+
"loss": 0.8311,
|
| 657 |
"step": 108
|
| 658 |
},
|
| 659 |
{
|
| 660 |
+
"epoch": 0.23,
|
| 661 |
+
"learning_rate": 0.00017889152718469835,
|
| 662 |
+
"loss": 0.8416,
|
| 663 |
"step": 109
|
| 664 |
},
|
| 665 |
{
|
| 666 |
+
"epoch": 0.24,
|
| 667 |
+
"learning_rate": 0.00017846156454020076,
|
| 668 |
+
"loss": 0.8871,
|
| 669 |
"step": 110
|
| 670 |
},
|
| 671 |
{
|
| 672 |
+
"epoch": 0.24,
|
| 673 |
+
"learning_rate": 0.00017802779472921222,
|
| 674 |
+
"loss": 0.7938,
|
| 675 |
"step": 111
|
| 676 |
},
|
| 677 |
{
|
| 678 |
+
"epoch": 0.24,
|
| 679 |
+
"learning_rate": 0.00017759023879941258,
|
| 680 |
+
"loss": 0.8236,
|
| 681 |
"step": 112
|
| 682 |
},
|
| 683 |
{
|
| 684 |
+
"epoch": 0.24,
|
| 685 |
+
"learning_rate": 0.0001771489179821943,
|
| 686 |
+
"loss": 0.829,
|
| 687 |
"step": 113
|
| 688 |
},
|
| 689 |
{
|
| 690 |
+
"epoch": 0.25,
|
| 691 |
+
"learning_rate": 0.00017670385369163242,
|
| 692 |
+
"loss": 0.8168,
|
| 693 |
"step": 114
|
| 694 |
},
|
| 695 |
{
|
| 696 |
+
"epoch": 0.25,
|
| 697 |
+
"learning_rate": 0.0001762550675234453,
|
| 698 |
+
"loss": 0.8461,
|
| 699 |
"step": 115
|
| 700 |
},
|
| 701 |
{
|
| 702 |
+
"epoch": 0.25,
|
| 703 |
+
"learning_rate": 0.00017580258125394693,
|
| 704 |
+
"loss": 0.8084,
|
| 705 |
"step": 116
|
| 706 |
},
|
| 707 |
{
|
| 708 |
+
"epoch": 0.25,
|
| 709 |
+
"learning_rate": 0.00017534641683899008,
|
| 710 |
+
"loss": 0.9077,
|
| 711 |
"step": 117
|
| 712 |
},
|
| 713 |
{
|
| 714 |
+
"epoch": 0.25,
|
| 715 |
+
"learning_rate": 0.0001748865964129011,
|
| 716 |
+
"loss": 0.8139,
|
| 717 |
"step": 118
|
| 718 |
},
|
| 719 |
{
|
| 720 |
+
"epoch": 0.26,
|
| 721 |
+
"learning_rate": 0.00017442314228740586,
|
| 722 |
+
"loss": 0.8471,
|
| 723 |
"step": 119
|
| 724 |
},
|
| 725 |
{
|
| 726 |
+
"epoch": 0.26,
|
| 727 |
+
"learning_rate": 0.0001739560769505471,
|
| 728 |
+
"loss": 0.8267,
|
| 729 |
"step": 120
|
| 730 |
},
|
| 731 |
{
|
| 732 |
+
"epoch": 0.26,
|
| 733 |
+
"learning_rate": 0.00017348542306559324,
|
| 734 |
+
"loss": 0.7966,
|
| 735 |
"step": 121
|
| 736 |
},
|
| 737 |
{
|
| 738 |
+
"epoch": 0.26,
|
| 739 |
+
"learning_rate": 0.00017301120346993875,
|
| 740 |
+
"loss": 0.9005,
|
| 741 |
"step": 122
|
| 742 |
},
|
| 743 |
{
|
| 744 |
+
"epoch": 0.26,
|
| 745 |
+
"learning_rate": 0.000172533441173996,
|
| 746 |
+
"loss": 0.8447,
|
| 747 |
"step": 123
|
| 748 |
},
|
| 749 |
{
|
| 750 |
+
"epoch": 0.27,
|
| 751 |
+
"learning_rate": 0.0001720521593600787,
|
| 752 |
+
"loss": 0.8722,
|
| 753 |
"step": 124
|
| 754 |
},
|
| 755 |
{
|
| 756 |
+
"epoch": 0.27,
|
| 757 |
+
"learning_rate": 0.00017156738138127706,
|
| 758 |
+
"loss": 0.8462,
|
| 759 |
"step": 125
|
| 760 |
},
|
| 761 |
{
|
| 762 |
+
"epoch": 0.27,
|
| 763 |
+
"learning_rate": 0.0001710791307603246,
|
| 764 |
+
"loss": 0.8249,
|
| 765 |
"step": 126
|
| 766 |
},
|
| 767 |
{
|
| 768 |
+
"epoch": 0.27,
|
| 769 |
+
"learning_rate": 0.00017058743118845686,
|
| 770 |
+
"loss": 0.8764,
|
| 771 |
"step": 127
|
| 772 |
},
|
| 773 |
{
|
| 774 |
+
"epoch": 0.28,
|
| 775 |
+
"learning_rate": 0.00017009230652426173,
|
| 776 |
+
"loss": 0.7816,
|
| 777 |
"step": 128
|
| 778 |
},
|
| 779 |
{
|
| 780 |
+
"epoch": 0.28,
|
| 781 |
+
"learning_rate": 0.00016959378079252176,
|
| 782 |
+
"loss": 0.804,
|
| 783 |
"step": 129
|
| 784 |
},
|
| 785 |
{
|
| 786 |
+
"epoch": 0.28,
|
| 787 |
+
"learning_rate": 0.00016909187818304854,
|
| 788 |
+
"loss": 0.8169,
|
| 789 |
"step": 130
|
| 790 |
},
|
| 791 |
{
|
| 792 |
+
"epoch": 0.28,
|
| 793 |
+
"learning_rate": 0.00016858662304950885,
|
| 794 |
+
"loss": 0.8389,
|
| 795 |
"step": 131
|
| 796 |
},
|
| 797 |
{
|
| 798 |
+
"epoch": 0.28,
|
| 799 |
+
"learning_rate": 0.00016807803990824294,
|
| 800 |
+
"loss": 0.879,
|
| 801 |
"step": 132
|
| 802 |
},
|
| 803 |
{
|
| 804 |
+
"epoch": 0.29,
|
| 805 |
+
"learning_rate": 0.00016756615343707494,
|
| 806 |
+
"loss": 0.7998,
|
| 807 |
"step": 133
|
| 808 |
},
|
| 809 |
{
|
| 810 |
+
"epoch": 0.29,
|
| 811 |
+
"learning_rate": 0.00016705098847411552,
|
| 812 |
+
"loss": 0.8218,
|
| 813 |
"step": 134
|
| 814 |
},
|
| 815 |
{
|
| 816 |
+
"epoch": 0.29,
|
| 817 |
+
"learning_rate": 0.00016653257001655652,
|
| 818 |
+
"loss": 0.7729,
|
| 819 |
"step": 135
|
| 820 |
},
|
| 821 |
{
|
| 822 |
+
"epoch": 0.29,
|
| 823 |
+
"learning_rate": 0.0001660109232194582,
|
| 824 |
+
"loss": 0.8509,
|
| 825 |
"step": 136
|
| 826 |
},
|
| 827 |
{
|
| 828 |
+
"epoch": 0.29,
|
| 829 |
+
"learning_rate": 0.00016548607339452853,
|
| 830 |
+
"loss": 0.809,
|
| 831 |
"step": 137
|
| 832 |
},
|
| 833 |
{
|
| 834 |
+
"epoch": 0.3,
|
| 835 |
+
"learning_rate": 0.00016495804600889487,
|
| 836 |
+
"loss": 0.8388,
|
| 837 |
"step": 138
|
| 838 |
},
|
| 839 |
{
|
| 840 |
+
"epoch": 0.3,
|
| 841 |
+
"learning_rate": 0.0001644268666838686,
|
| 842 |
+
"loss": 0.8096,
|
| 843 |
"step": 139
|
| 844 |
},
|
| 845 |
{
|
| 846 |
+
"epoch": 0.3,
|
| 847 |
+
"learning_rate": 0.00016389256119370154,
|
| 848 |
+
"loss": 0.8099,
|
| 849 |
"step": 140
|
| 850 |
},
|
| 851 |
{
|
| 852 |
+
"epoch": 0.3,
|
| 853 |
+
"learning_rate": 0.00016335515546433553,
|
| 854 |
+
"loss": 0.8121,
|
| 855 |
"step": 141
|
| 856 |
},
|
| 857 |
{
|
| 858 |
+
"epoch": 0.31,
|
| 859 |
+
"learning_rate": 0.00016281467557214437,
|
| 860 |
+
"loss": 0.8371,
|
| 861 |
"step": 142
|
| 862 |
},
|
| 863 |
{
|
| 864 |
+
"epoch": 0.31,
|
| 865 |
+
"learning_rate": 0.00016227114774266854,
|
| 866 |
+
"loss": 0.8633,
|
| 867 |
"step": 143
|
| 868 |
},
|
| 869 |
{
|
| 870 |
+
"epoch": 0.31,
|
| 871 |
+
"learning_rate": 0.00016172459834934254,
|
| 872 |
+
"loss": 0.8851,
|
| 873 |
"step": 144
|
| 874 |
},
|
| 875 |
{
|
| 876 |
+
"epoch": 0.31,
|
| 877 |
+
"learning_rate": 0.00016117505391221543,
|
| 878 |
+
"loss": 0.7979,
|
| 879 |
"step": 145
|
| 880 |
},
|
| 881 |
{
|
| 882 |
+
"epoch": 0.31,
|
| 883 |
+
"learning_rate": 0.0001606225410966638,
|
| 884 |
+
"loss": 0.8057,
|
| 885 |
"step": 146
|
| 886 |
},
|
| 887 |
{
|
| 888 |
+
"epoch": 0.32,
|
| 889 |
+
"learning_rate": 0.00016006708671209794,
|
| 890 |
+
"loss": 0.8623,
|
| 891 |
"step": 147
|
| 892 |
},
|
| 893 |
{
|
| 894 |
+
"epoch": 0.32,
|
| 895 |
+
"learning_rate": 0.000159508717710661,
|
| 896 |
+
"loss": 0.85,
|
| 897 |
"step": 148
|
| 898 |
},
|
| 899 |
{
|
| 900 |
+
"epoch": 0.32,
|
| 901 |
+
"learning_rate": 0.00015894746118592123,
|
| 902 |
+
"loss": 0.8645,
|
| 903 |
"step": 149
|
| 904 |
},
|
| 905 |
{
|
| 906 |
+
"epoch": 0.32,
|
| 907 |
+
"learning_rate": 0.0001583833443715572,
|
| 908 |
+
"loss": 0.8464,
|
| 909 |
"step": 150
|
| 910 |
},
|
| 911 |
{
|
| 912 |
+
"epoch": 0.32,
|
| 913 |
+
"learning_rate": 0.0001578163946400366,
|
| 914 |
+
"loss": 0.7715,
|
| 915 |
"step": 151
|
| 916 |
},
|
| 917 |
{
|
| 918 |
+
"epoch": 0.33,
|
| 919 |
+
"learning_rate": 0.00015724663950128775,
|
| 920 |
+
"loss": 0.8274,
|
| 921 |
"step": 152
|
| 922 |
},
|
| 923 |
{
|
| 924 |
+
"epoch": 0.33,
|
| 925 |
+
"learning_rate": 0.00015667410660136487,
|
| 926 |
+
"loss": 0.9138,
|
| 927 |
"step": 153
|
| 928 |
},
|
| 929 |
{
|
| 930 |
+
"epoch": 0.33,
|
| 931 |
+
"learning_rate": 0.00015609882372110684,
|
| 932 |
+
"loss": 0.7879,
|
| 933 |
"step": 154
|
| 934 |
},
|
| 935 |
{
|
| 936 |
+
"epoch": 0.33,
|
| 937 |
+
"learning_rate": 0.0001555208187747887,
|
| 938 |
+
"loss": 0.8327,
|
| 939 |
"step": 155
|
| 940 |
},
|
| 941 |
{
|
| 942 |
+
"epoch": 0.34,
|
| 943 |
+
"learning_rate": 0.0001549401198087677,
|
| 944 |
+
"loss": 0.8274,
|
| 945 |
"step": 156
|
| 946 |
},
|
| 947 |
{
|
| 948 |
+
"epoch": 0.34,
|
| 949 |
+
"learning_rate": 0.00015435675500012212,
|
| 950 |
+
"loss": 0.8606,
|
| 951 |
"step": 157
|
| 952 |
},
|
| 953 |
{
|
| 954 |
+
"epoch": 0.34,
|
| 955 |
+
"learning_rate": 0.00015377075265528406,
|
| 956 |
+
"loss": 0.8739,
|
| 957 |
"step": 158
|
| 958 |
},
|
| 959 |
{
|
| 960 |
+
"epoch": 0.34,
|
| 961 |
+
"learning_rate": 0.000153182141208666,
|
| 962 |
+
"loss": 0.8834,
|
| 963 |
"step": 159
|
| 964 |
},
|
| 965 |
{
|
| 966 |
+
"epoch": 0.34,
|
| 967 |
+
"learning_rate": 0.00015259094922128109,
|
| 968 |
+
"loss": 0.8255,
|
| 969 |
"step": 160
|
| 970 |
},
|
| 971 |
{
|
| 972 |
+
"epoch": 0.35,
|
| 973 |
+
"learning_rate": 0.00015199720537935725,
|
| 974 |
+
"loss": 0.8356,
|
| 975 |
"step": 161
|
| 976 |
},
|
| 977 |
{
|
| 978 |
+
"epoch": 0.35,
|
| 979 |
+
"learning_rate": 0.0001514009384929453,
|
| 980 |
+
"loss": 0.8171,
|
| 981 |
"step": 162
|
| 982 |
},
|
| 983 |
{
|
| 984 |
+
"epoch": 0.35,
|
| 985 |
+
"learning_rate": 0.00015080217749452094,
|
| 986 |
+
"loss": 0.8538,
|
| 987 |
"step": 163
|
| 988 |
},
|
| 989 |
{
|
| 990 |
+
"epoch": 0.35,
|
| 991 |
+
"learning_rate": 0.00015020095143758084,
|
| 992 |
+
"loss": 0.828,
|
| 993 |
"step": 164
|
| 994 |
},
|
| 995 |
{
|
| 996 |
+
"epoch": 0.35,
|
| 997 |
+
"learning_rate": 0.00014959728949523305,
|
| 998 |
+
"loss": 0.8258,
|
| 999 |
"step": 165
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
+
"epoch": 0.36,
|
| 1003 |
+
"learning_rate": 0.00014899122095878137,
|
| 1004 |
+
"loss": 0.8689,
|
| 1005 |
"step": 166
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
+
"epoch": 0.36,
|
| 1009 |
+
"learning_rate": 0.00014838277523630388,
|
| 1010 |
+
"loss": 0.8654,
|
| 1011 |
"step": 167
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
+
"epoch": 0.36,
|
| 1015 |
+
"learning_rate": 0.0001477719818512263,
|
| 1016 |
+
"loss": 0.8278,
|
| 1017 |
"step": 168
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
+
"epoch": 0.36,
|
| 1021 |
+
"learning_rate": 0.00014715887044088912,
|
| 1022 |
+
"loss": 0.833,
|
| 1023 |
"step": 169
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
+
"epoch": 0.37,
|
| 1027 |
+
"learning_rate": 0.00014654347075510974,
|
| 1028 |
+
"loss": 0.7702,
|
| 1029 |
"step": 170
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
+
"epoch": 0.37,
|
| 1033 |
+
"learning_rate": 0.00014592581265473884,
|
| 1034 |
+
"loss": 0.7773,
|
| 1035 |
"step": 171
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
+
"epoch": 0.37,
|
| 1039 |
+
"learning_rate": 0.00014530592611021145,
|
| 1040 |
+
"loss": 0.7538,
|
| 1041 |
"step": 172
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
+
"epoch": 0.37,
|
| 1045 |
+
"learning_rate": 0.00014468384120009273,
|
| 1046 |
+
"loss": 0.8864,
|
| 1047 |
"step": 173
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
+
"epoch": 0.37,
|
| 1051 |
+
"learning_rate": 0.00014405958810961842,
|
| 1052 |
+
"loss": 0.7691,
|
| 1053 |
"step": 174
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
+
"epoch": 0.38,
|
| 1057 |
+
"learning_rate": 0.00014343319712923025,
|
| 1058 |
+
"loss": 0.8287,
|
| 1059 |
"step": 175
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
+
"epoch": 0.38,
|
| 1063 |
+
"learning_rate": 0.00014280469865310614,
|
| 1064 |
+
"loss": 0.8333,
|
| 1065 |
"step": 176
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
+
"epoch": 0.38,
|
| 1069 |
+
"learning_rate": 0.0001421741231776853,
|
| 1070 |
+
"loss": 0.7896,
|
| 1071 |
"step": 177
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 0.38,
|
| 1075 |
+
"learning_rate": 0.00014154150130018866,
|
| 1076 |
+
"loss": 0.8755,
|
| 1077 |
"step": 178
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
+
"epoch": 0.38,
|
| 1081 |
+
"learning_rate": 0.00014090686371713402,
|
| 1082 |
+
"loss": 0.8675,
|
| 1083 |
"step": 179
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
+
"epoch": 0.39,
|
| 1087 |
+
"learning_rate": 0.0001402702412228466,
|
| 1088 |
+
"loss": 0.798,
|
| 1089 |
"step": 180
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
+
"epoch": 0.39,
|
| 1093 |
+
"learning_rate": 0.000139631664707965,
|
| 1094 |
+
"loss": 0.8112,
|
| 1095 |
"step": 181
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
+
"epoch": 0.39,
|
| 1099 |
+
"learning_rate": 0.00013899116515794205,
|
| 1100 |
+
"loss": 0.878,
|
| 1101 |
"step": 182
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
+
"epoch": 0.39,
|
| 1105 |
+
"learning_rate": 0.00013834877365154143,
|
| 1106 |
+
"loss": 0.8426,
|
| 1107 |
"step": 183
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
+
"epoch": 0.4,
|
| 1111 |
+
"learning_rate": 0.00013770452135932967,
|
| 1112 |
+
"loss": 0.7863,
|
| 1113 |
"step": 184
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
+
"epoch": 0.4,
|
| 1117 |
+
"learning_rate": 0.00013705843954216367,
|
| 1118 |
+
"loss": 0.7532,
|
| 1119 |
"step": 185
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
+
"epoch": 0.4,
|
| 1123 |
+
"learning_rate": 0.00013641055954967377,
|
| 1124 |
+
"loss": 0.7935,
|
| 1125 |
"step": 186
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
+
"epoch": 0.4,
|
| 1129 |
+
"learning_rate": 0.00013576091281874256,
|
| 1130 |
+
"loss": 0.7937,
|
| 1131 |
"step": 187
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
+
"epoch": 0.4,
|
| 1135 |
+
"learning_rate": 0.00013510953087197974,
|
| 1136 |
+
"loss": 0.803,
|
| 1137 |
"step": 188
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
+
"epoch": 0.41,
|
| 1141 |
+
"learning_rate": 0.0001344564453161921,
|
| 1142 |
+
"loss": 0.781,
|
| 1143 |
"step": 189
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
+
"epoch": 0.41,
|
| 1147 |
+
"learning_rate": 0.00013380168784085027,
|
| 1148 |
+
"loss": 0.785,
|
| 1149 |
"step": 190
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
+
"epoch": 0.41,
|
| 1153 |
+
"learning_rate": 0.00013314529021655097,
|
| 1154 |
+
"loss": 0.8305,
|
| 1155 |
"step": 191
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
+
"epoch": 0.41,
|
| 1159 |
+
"learning_rate": 0.00013248728429347524,
|
| 1160 |
+
"loss": 0.8081,
|
| 1161 |
"step": 192
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
+
"epoch": 0.42,
|
| 1165 |
+
"learning_rate": 0.0001318277019998432,
|
| 1166 |
+
"loss": 0.8432,
|
| 1167 |
"step": 193
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
+
"epoch": 0.42,
|
| 1171 |
+
"learning_rate": 0.00013116657534036468,
|
| 1172 |
+
"loss": 0.86,
|
| 1173 |
"step": 194
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
+
"epoch": 0.42,
|
| 1177 |
+
"learning_rate": 0.00013050393639468629,
|
| 1178 |
+
"loss": 0.8163,
|
| 1179 |
"step": 195
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
+
"epoch": 0.42,
|
| 1183 |
+
"learning_rate": 0.00012983981731583486,
|
| 1184 |
+
"loss": 0.7691,
|
| 1185 |
"step": 196
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
+
"epoch": 0.42,
|
| 1189 |
+
"learning_rate": 0.0001291742503286573,
|
| 1190 |
+
"loss": 0.82,
|
| 1191 |
"step": 197
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
+
"epoch": 0.43,
|
| 1195 |
+
"learning_rate": 0.00012850726772825686,
|
| 1196 |
+
"loss": 0.857,
|
| 1197 |
"step": 198
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
+
"epoch": 0.43,
|
| 1201 |
+
"learning_rate": 0.00012783890187842618,
|
| 1202 |
+
"loss": 0.8442,
|
| 1203 |
"step": 199
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
+
"epoch": 0.43,
|
| 1207 |
+
"learning_rate": 0.00012716918521007696,
|
| 1208 |
+
"loss": 0.8234,
|
| 1209 |
"step": 200
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
+
"epoch": 0.43,
|
| 1213 |
+
"learning_rate": 0.0001264981502196662,
|
| 1214 |
+
"loss": 0.7953,
|
| 1215 |
"step": 201
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
+
"epoch": 0.43,
|
| 1219 |
+
"learning_rate": 0.00012582582946761938,
|
| 1220 |
+
"loss": 0.8526,
|
| 1221 |
"step": 202
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
+
"epoch": 0.44,
|
| 1225 |
+
"learning_rate": 0.0001251522555767507,
|
| 1226 |
+
"loss": 0.7855,
|
| 1227 |
"step": 203
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
+
"epoch": 0.44,
|
| 1231 |
+
"learning_rate": 0.00012447746123067996,
|
| 1232 |
+
"loss": 0.7828,
|
| 1233 |
"step": 204
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
+
"epoch": 0.44,
|
| 1237 |
+
"learning_rate": 0.0001238014791722468,
|
| 1238 |
+
"loss": 0.7819,
|
| 1239 |
"step": 205
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
+
"epoch": 0.44,
|
| 1243 |
+
"learning_rate": 0.00012312434220192178,
|
| 1244 |
+
"loss": 0.8025,
|
| 1245 |
"step": 206
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
+
"epoch": 0.45,
|
| 1249 |
+
"learning_rate": 0.000122446083176215,
|
| 1250 |
+
"loss": 0.7783,
|
| 1251 |
"step": 207
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
+
"epoch": 0.45,
|
| 1255 |
+
"learning_rate": 0.00012176673500608155,
|
| 1256 |
+
"loss": 0.7921,
|
| 1257 |
"step": 208
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
+
"epoch": 0.45,
|
| 1261 |
+
"learning_rate": 0.00012108633065532498,
|
| 1262 |
+
"loss": 0.8183,
|
| 1263 |
"step": 209
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
+
"epoch": 0.45,
|
| 1267 |
+
"learning_rate": 0.00012040490313899735,
|
| 1268 |
+
"loss": 0.835,
|
| 1269 |
"step": 210
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
+
"epoch": 0.45,
|
| 1273 |
+
"learning_rate": 0.00011972248552179755,
|
| 1274 |
+
"loss": 0.7651,
|
| 1275 |
"step": 211
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
+
"epoch": 0.46,
|
| 1279 |
+
"learning_rate": 0.00011903911091646684,
|
| 1280 |
+
"loss": 0.839,
|
| 1281 |
"step": 212
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
+
"epoch": 0.46,
|
| 1285 |
+
"learning_rate": 0.00011835481248218215,
|
| 1286 |
+
"loss": 0.8393,
|
| 1287 |
"step": 213
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
+
"epoch": 0.46,
|
| 1291 |
+
"learning_rate": 0.000117669623422947,
|
| 1292 |
+
"loss": 0.8009,
|
| 1293 |
"step": 214
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
+
"epoch": 0.46,
|
| 1297 |
+
"learning_rate": 0.00011698357698598052,
|
| 1298 |
+
"loss": 0.7651,
|
| 1299 |
"step": 215
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
+
"epoch": 0.46,
|
| 1303 |
+
"learning_rate": 0.00011629670646010407,
|
| 1304 |
+
"loss": 0.7803,
|
| 1305 |
"step": 216
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
+
"epoch": 0.47,
|
| 1309 |
+
"learning_rate": 0.00011560904517412598,
|
| 1310 |
+
"loss": 0.8199,
|
| 1311 |
"step": 217
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
+
"epoch": 0.47,
|
| 1315 |
+
"learning_rate": 0.0001149206264952245,
|
| 1316 |
+
"loss": 0.7937,
|
| 1317 |
"step": 218
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
+
"epoch": 0.47,
|
| 1321 |
+
"learning_rate": 0.00011423148382732853,
|
| 1322 |
+
"loss": 0.7851,
|
| 1323 |
"step": 219
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
+
"epoch": 0.47,
|
| 1327 |
+
"learning_rate": 0.0001135416506094969,
|
| 1328 |
+
"loss": 0.7896,
|
| 1329 |
"step": 220
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
+
"epoch": 0.48,
|
| 1333 |
+
"learning_rate": 0.00011285116031429584,
|
| 1334 |
+
"loss": 0.7636,
|
| 1335 |
"step": 221
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
+
"epoch": 0.48,
|
| 1339 |
+
"learning_rate": 0.00011216004644617462,
|
| 1340 |
+
"loss": 0.814,
|
| 1341 |
"step": 222
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
+
"epoch": 0.48,
|
| 1345 |
+
"learning_rate": 0.00011146834253984006,
|
| 1346 |
+
"loss": 0.7922,
|
| 1347 |
"step": 223
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
+
"epoch": 0.48,
|
| 1351 |
+
"learning_rate": 0.00011077608215862914,
|
| 1352 |
+
"loss": 0.7879,
|
| 1353 |
"step": 224
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
+
"epoch": 0.48,
|
| 1357 |
+
"learning_rate": 0.0001100832988928806,
|
| 1358 |
+
"loss": 0.7597,
|
| 1359 |
"step": 225
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
+
"epoch": 0.49,
|
| 1363 |
+
"learning_rate": 0.00010939002635830484,
|
| 1364 |
+
"loss": 0.7908,
|
| 1365 |
"step": 226
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
+
"epoch": 0.49,
|
| 1369 |
+
"learning_rate": 0.00010869629819435296,
|
| 1370 |
+
"loss": 0.8218,
|
| 1371 |
"step": 227
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
+
"epoch": 0.49,
|
| 1375 |
+
"learning_rate": 0.00010800214806258444,
|
| 1376 |
+
"loss": 0.7939,
|
| 1377 |
"step": 228
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
+
"epoch": 0.49,
|
| 1381 |
+
"learning_rate": 0.0001073076096450337,
|
| 1382 |
+
"loss": 0.8615,
|
| 1383 |
"step": 229
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
+
"epoch": 0.49,
|
| 1387 |
+
"learning_rate": 0.00010661271664257593,
|
| 1388 |
+
"loss": 0.8111,
|
| 1389 |
"step": 230
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
+
"epoch": 0.5,
|
| 1393 |
+
"learning_rate": 0.00010591750277329166,
|
| 1394 |
+
"loss": 0.7829,
|
| 1395 |
"step": 231
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
+
"epoch": 0.5,
|
| 1399 |
+
"learning_rate": 0.00010522200177083077,
|
| 1400 |
+
"loss": 0.7753,
|
| 1401 |
"step": 232
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
+
"epoch": 0.5,
|
| 1405 |
+
"learning_rate": 0.00010452624738277564,
|
| 1406 |
+
"loss": 0.725,
|
| 1407 |
"step": 233
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
+
"epoch": 0.5,
|
| 1411 |
+
"learning_rate": 0.00010383027336900355,
|
| 1412 |
+
"loss": 0.7834,
|
| 1413 |
"step": 234
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
+
"epoch": 0.51,
|
| 1417 |
+
"learning_rate": 0.00010313411350004863,
|
| 1418 |
+
"loss": 0.7676,
|
| 1419 |
"step": 235
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
+
"epoch": 0.51,
|
| 1423 |
+
"learning_rate": 0.00010243780155546323,
|
| 1424 |
+
"loss": 0.8096,
|
| 1425 |
"step": 236
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
+
"epoch": 0.51,
|
| 1429 |
+
"learning_rate": 0.00010174137132217883,
|
| 1430 |
+
"loss": 0.7363,
|
| 1431 |
"step": 237
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
+
"epoch": 0.51,
|
| 1435 |
+
"learning_rate": 0.00010104485659286647,
|
| 1436 |
+
"loss": 0.7526,
|
| 1437 |
"step": 238
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
+
"epoch": 0.51,
|
| 1441 |
+
"learning_rate": 0.00010034829116429738,
|
| 1442 |
+
"loss": 0.7667,
|
| 1443 |
"step": 239
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
+
"epoch": 0.52,
|
| 1447 |
+
"learning_rate": 9.965170883570266e-05,
|
| 1448 |
+
"loss": 0.7648,
|
| 1449 |
"step": 240
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
+
"epoch": 0.52,
|
| 1453 |
+
"learning_rate": 9.895514340713354e-05,
|
| 1454 |
+
"loss": 0.7525,
|
| 1455 |
"step": 241
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
+
"epoch": 0.52,
|
| 1459 |
+
"learning_rate": 9.825862867782123e-05,
|
| 1460 |
+
"loss": 0.809,
|
| 1461 |
"step": 242
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
+
"epoch": 0.52,
|
| 1465 |
+
"learning_rate": 9.756219844453676e-05,
|
| 1466 |
+
"loss": 0.7872,
|
| 1467 |
"step": 243
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
+
"epoch": 0.52,
|
| 1471 |
+
"learning_rate": 9.686588649995137e-05,
|
| 1472 |
+
"loss": 0.8715,
|
| 1473 |
"step": 244
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
+
"epoch": 0.53,
|
| 1477 |
+
"learning_rate": 9.616972663099647e-05,
|
| 1478 |
+
"loss": 0.8001,
|
| 1479 |
"step": 245
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
+
"epoch": 0.53,
|
| 1483 |
+
"learning_rate": 9.547375261722437e-05,
|
| 1484 |
+
"loss": 0.7999,
|
| 1485 |
"step": 246
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
+
"epoch": 0.53,
|
| 1489 |
+
"learning_rate": 9.477799822916924e-05,
|
| 1490 |
+
"loss": 0.8099,
|
| 1491 |
"step": 247
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
+
"epoch": 0.53,
|
| 1495 |
+
"learning_rate": 9.408249722670837e-05,
|
| 1496 |
+
"loss": 0.6896,
|
| 1497 |
"step": 248
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
+
"epoch": 0.54,
|
| 1501 |
+
"learning_rate": 9.338728335742406e-05,
|
| 1502 |
+
"loss": 0.7653,
|
| 1503 |
"step": 249
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
+
"epoch": 0.54,
|
| 1507 |
+
"learning_rate": 9.26923903549663e-05,
|
| 1508 |
+
"loss": 0.8662,
|
| 1509 |
"step": 250
|
| 1510 |
},
|
| 1511 |
{
|
| 1512 |
+
"epoch": 0.54,
|
| 1513 |
+
"learning_rate": 9.199785193741559e-05,
|
| 1514 |
+
"loss": 0.7762,
|
| 1515 |
"step": 251
|
| 1516 |
},
|
| 1517 |
{
|
| 1518 |
+
"epoch": 0.54,
|
| 1519 |
+
"learning_rate": 9.130370180564706e-05,
|
| 1520 |
+
"loss": 0.7992,
|
| 1521 |
"step": 252
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
+
"epoch": 0.54,
|
| 1525 |
+
"learning_rate": 9.06099736416952e-05,
|
| 1526 |
+
"loss": 0.747,
|
| 1527 |
"step": 253
|
| 1528 |
},
|
| 1529 |
{
|
| 1530 |
+
"epoch": 0.55,
|
| 1531 |
+
"learning_rate": 8.991670110711943e-05,
|
| 1532 |
+
"loss": 0.8737,
|
| 1533 |
"step": 254
|
| 1534 |
},
|
| 1535 |
{
|
| 1536 |
+
"epoch": 0.55,
|
| 1537 |
+
"learning_rate": 8.922391784137086e-05,
|
| 1538 |
+
"loss": 0.8148,
|
| 1539 |
"step": 255
|
| 1540 |
},
|
| 1541 |
{
|
| 1542 |
+
"epoch": 0.55,
|
| 1543 |
+
"learning_rate": 8.853165746015997e-05,
|
| 1544 |
+
"loss": 0.7463,
|
| 1545 |
"step": 256
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
+
"epoch": 0.55,
|
| 1549 |
+
"learning_rate": 8.78399535538254e-05,
|
| 1550 |
+
"loss": 0.7471,
|
| 1551 |
"step": 257
|
| 1552 |
},
|
| 1553 |
{
|
| 1554 |
+
"epoch": 0.55,
|
| 1555 |
+
"learning_rate": 8.714883968570417e-05,
|
| 1556 |
+
"loss": 0.7916,
|
| 1557 |
"step": 258
|
| 1558 |
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.56,
|
| 1561 |
+
"learning_rate": 8.64583493905031e-05,
|
| 1562 |
+
"loss": 0.7884,
|
| 1563 |
+
"step": 259
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.56,
|
| 1567 |
+
"learning_rate": 8.57685161726715e-05,
|
| 1568 |
+
"loss": 0.7891,
|
| 1569 |
+
"step": 260
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 0.56,
|
| 1573 |
+
"learning_rate": 8.507937350477553e-05,
|
| 1574 |
+
"loss": 0.7395,
|
| 1575 |
+
"step": 261
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.56,
|
| 1579 |
+
"learning_rate": 8.439095482587403e-05,
|
| 1580 |
+
"loss": 0.8222,
|
| 1581 |
+
"step": 262
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 0.57,
|
| 1585 |
+
"learning_rate": 8.370329353989597e-05,
|
| 1586 |
+
"loss": 0.8021,
|
| 1587 |
+
"step": 263
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.57,
|
| 1591 |
+
"learning_rate": 8.301642301401949e-05,
|
| 1592 |
+
"loss": 0.7854,
|
| 1593 |
+
"step": 264
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.57,
|
| 1597 |
+
"learning_rate": 8.233037657705302e-05,
|
| 1598 |
+
"loss": 0.7797,
|
| 1599 |
+
"step": 265
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.57,
|
| 1603 |
+
"learning_rate": 8.164518751781789e-05,
|
| 1604 |
+
"loss": 0.7943,
|
| 1605 |
+
"step": 266
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.57,
|
| 1609 |
+
"learning_rate": 8.096088908353315e-05,
|
| 1610 |
+
"loss": 0.8402,
|
| 1611 |
+
"step": 267
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.58,
|
| 1615 |
+
"learning_rate": 8.027751447820246e-05,
|
| 1616 |
+
"loss": 0.8083,
|
| 1617 |
+
"step": 268
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.58,
|
| 1621 |
+
"learning_rate": 7.959509686100267e-05,
|
| 1622 |
+
"loss": 0.7724,
|
| 1623 |
+
"step": 269
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 0.58,
|
| 1627 |
+
"learning_rate": 7.891366934467503e-05,
|
| 1628 |
+
"loss": 0.7647,
|
| 1629 |
+
"step": 270
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.58,
|
| 1633 |
+
"learning_rate": 7.823326499391846e-05,
|
| 1634 |
+
"loss": 0.7342,
|
| 1635 |
+
"step": 271
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.58,
|
| 1639 |
+
"learning_rate": 7.755391682378506e-05,
|
| 1640 |
+
"loss": 0.7705,
|
| 1641 |
+
"step": 272
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.59,
|
| 1645 |
+
"learning_rate": 7.687565779807824e-05,
|
| 1646 |
+
"loss": 0.8154,
|
| 1647 |
+
"step": 273
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.59,
|
| 1651 |
+
"learning_rate": 7.619852082775323e-05,
|
| 1652 |
+
"loss": 0.7415,
|
| 1653 |
+
"step": 274
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.59,
|
| 1657 |
+
"learning_rate": 7.552253876932006e-05,
|
| 1658 |
+
"loss": 0.8267,
|
| 1659 |
+
"step": 275
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.59,
|
| 1663 |
+
"learning_rate": 7.484774442324932e-05,
|
| 1664 |
+
"loss": 0.8021,
|
| 1665 |
+
"step": 276
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 0.6,
|
| 1669 |
+
"learning_rate": 7.417417053238065e-05,
|
| 1670 |
+
"loss": 0.7901,
|
| 1671 |
+
"step": 277
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.6,
|
| 1675 |
+
"learning_rate": 7.350184978033386e-05,
|
| 1676 |
+
"loss": 0.8075,
|
| 1677 |
+
"step": 278
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.6,
|
| 1681 |
+
"learning_rate": 7.283081478992308e-05,
|
| 1682 |
+
"loss": 0.8005,
|
| 1683 |
+
"step": 279
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.6,
|
| 1687 |
+
"learning_rate": 7.216109812157382e-05,
|
| 1688 |
+
"loss": 0.7494,
|
| 1689 |
+
"step": 280
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.6,
|
| 1693 |
+
"learning_rate": 7.149273227174317e-05,
|
| 1694 |
+
"loss": 0.7529,
|
| 1695 |
+
"step": 281
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.61,
|
| 1699 |
+
"learning_rate": 7.082574967134274e-05,
|
| 1700 |
+
"loss": 0.706,
|
| 1701 |
+
"step": 282
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.61,
|
| 1705 |
+
"learning_rate": 7.016018268416518e-05,
|
| 1706 |
+
"loss": 0.7667,
|
| 1707 |
+
"step": 283
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 0.61,
|
| 1711 |
+
"learning_rate": 6.949606360531375e-05,
|
| 1712 |
+
"loss": 0.7019,
|
| 1713 |
+
"step": 284
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.61,
|
| 1717 |
+
"learning_rate": 6.883342465963537e-05,
|
| 1718 |
+
"loss": 0.7613,
|
| 1719 |
+
"step": 285
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.62,
|
| 1723 |
+
"learning_rate": 6.817229800015682e-05,
|
| 1724 |
+
"loss": 0.8472,
|
| 1725 |
+
"step": 286
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.62,
|
| 1729 |
+
"learning_rate": 6.751271570652477e-05,
|
| 1730 |
+
"loss": 0.7839,
|
| 1731 |
+
"step": 287
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.62,
|
| 1735 |
+
"learning_rate": 6.685470978344905e-05,
|
| 1736 |
+
"loss": 0.7967,
|
| 1737 |
+
"step": 288
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.62,
|
| 1741 |
+
"learning_rate": 6.619831215914974e-05,
|
| 1742 |
+
"loss": 0.7803,
|
| 1743 |
+
"step": 289
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.62,
|
| 1747 |
+
"learning_rate": 6.554355468380795e-05,
|
| 1748 |
+
"loss": 0.7655,
|
| 1749 |
+
"step": 290
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 0.63,
|
| 1753 |
+
"learning_rate": 6.489046912802031e-05,
|
| 1754 |
+
"loss": 0.7805,
|
| 1755 |
+
"step": 291
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.63,
|
| 1759 |
+
"learning_rate": 6.423908718125742e-05,
|
| 1760 |
+
"loss": 0.8139,
|
| 1761 |
+
"step": 292
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.63,
|
| 1765 |
+
"learning_rate": 6.358944045032627e-05,
|
| 1766 |
+
"loss": 0.8033,
|
| 1767 |
+
"step": 293
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.63,
|
| 1771 |
+
"learning_rate": 6.294156045783634e-05,
|
| 1772 |
+
"loss": 0.8466,
|
| 1773 |
+
"step": 294
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.63,
|
| 1777 |
+
"learning_rate": 6.229547864067034e-05,
|
| 1778 |
+
"loss": 0.8037,
|
| 1779 |
+
"step": 295
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.64,
|
| 1783 |
+
"learning_rate": 6.16512263484586e-05,
|
| 1784 |
+
"loss": 0.8241,
|
| 1785 |
+
"step": 296
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.64,
|
| 1789 |
+
"learning_rate": 6.1008834842057995e-05,
|
| 1790 |
+
"loss": 0.8096,
|
| 1791 |
+
"step": 297
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 0.64,
|
| 1795 |
+
"learning_rate": 6.0368335292035e-05,
|
| 1796 |
+
"loss": 0.7818,
|
| 1797 |
+
"step": 298
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.64,
|
| 1801 |
+
"learning_rate": 5.972975877715339e-05,
|
| 1802 |
+
"loss": 0.7422,
|
| 1803 |
+
"step": 299
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.65,
|
| 1807 |
+
"learning_rate": 5.909313628286601e-05,
|
| 1808 |
+
"loss": 0.7525,
|
| 1809 |
+
"step": 300
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.65,
|
| 1813 |
+
"learning_rate": 5.845849869981137e-05,
|
| 1814 |
+
"loss": 0.7912,
|
| 1815 |
+
"step": 301
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.65,
|
| 1819 |
+
"learning_rate": 5.782587682231473e-05,
|
| 1820 |
+
"loss": 0.7835,
|
| 1821 |
+
"step": 302
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.65,
|
| 1825 |
+
"learning_rate": 5.71953013468939e-05,
|
| 1826 |
+
"loss": 0.7513,
|
| 1827 |
+
"step": 303
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 0.65,
|
| 1831 |
+
"learning_rate": 5.6566802870769766e-05,
|
| 1832 |
+
"loss": 0.8177,
|
| 1833 |
+
"step": 304
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 0.66,
|
| 1837 |
+
"learning_rate": 5.5940411890381573e-05,
|
| 1838 |
+
"loss": 0.7986,
|
| 1839 |
+
"step": 305
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.66,
|
| 1843 |
+
"learning_rate": 5.531615879990729e-05,
|
| 1844 |
+
"loss": 0.8016,
|
| 1845 |
+
"step": 306
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.66,
|
| 1849 |
+
"learning_rate": 5.469407388978854e-05,
|
| 1850 |
+
"loss": 0.7473,
|
| 1851 |
+
"step": 307
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.66,
|
| 1855 |
+
"learning_rate": 5.407418734526118e-05,
|
| 1856 |
+
"loss": 0.8324,
|
| 1857 |
+
"step": 308
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.66,
|
| 1861 |
+
"learning_rate": 5.3456529244890274e-05,
|
| 1862 |
+
"loss": 0.7647,
|
| 1863 |
+
"step": 309
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.67,
|
| 1867 |
+
"learning_rate": 5.2841129559110894e-05,
|
| 1868 |
+
"loss": 0.7477,
|
| 1869 |
+
"step": 310
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 0.67,
|
| 1873 |
+
"learning_rate": 5.222801814877369e-05,
|
| 1874 |
+
"loss": 0.7838,
|
| 1875 |
+
"step": 311
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 0.67,
|
| 1879 |
+
"learning_rate": 5.1617224763696123e-05,
|
| 1880 |
+
"loss": 0.7668,
|
| 1881 |
+
"step": 312
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.67,
|
| 1885 |
+
"learning_rate": 5.1008779041218644e-05,
|
| 1886 |
+
"loss": 0.7891,
|
| 1887 |
+
"step": 313
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.68,
|
| 1891 |
+
"learning_rate": 5.040271050476697e-05,
|
| 1892 |
+
"loss": 0.8267,
|
| 1893 |
+
"step": 314
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 0.68,
|
| 1897 |
+
"learning_rate": 4.9799048562419204e-05,
|
| 1898 |
+
"loss": 0.7796,
|
| 1899 |
+
"step": 315
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.68,
|
| 1903 |
+
"learning_rate": 4.919782250547911e-05,
|
| 1904 |
+
"loss": 0.7924,
|
| 1905 |
+
"step": 316
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.68,
|
| 1909 |
+
"learning_rate": 4.859906150705471e-05,
|
| 1910 |
+
"loss": 0.7837,
|
| 1911 |
+
"step": 317
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 0.68,
|
| 1915 |
+
"learning_rate": 4.8002794620642776e-05,
|
| 1916 |
+
"loss": 0.811,
|
| 1917 |
+
"step": 318
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 0.69,
|
| 1921 |
+
"learning_rate": 4.7409050778718944e-05,
|
| 1922 |
+
"loss": 0.7875,
|
| 1923 |
+
"step": 319
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.69,
|
| 1927 |
+
"learning_rate": 4.681785879133402e-05,
|
| 1928 |
+
"loss": 0.8032,
|
| 1929 |
+
"step": 320
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.69,
|
| 1933 |
+
"learning_rate": 4.6229247344715985e-05,
|
| 1934 |
+
"loss": 0.8237,
|
| 1935 |
+
"step": 321
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.69,
|
| 1939 |
+
"learning_rate": 4.56432449998779e-05,
|
| 1940 |
+
"loss": 0.758,
|
| 1941 |
+
"step": 322
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.69,
|
| 1945 |
+
"learning_rate": 4.505988019123228e-05,
|
| 1946 |
+
"loss": 0.777,
|
| 1947 |
+
"step": 323
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.7,
|
| 1951 |
+
"learning_rate": 4.447918122521129e-05,
|
| 1952 |
+
"loss": 0.816,
|
| 1953 |
+
"step": 324
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 0.7,
|
| 1957 |
+
"learning_rate": 4.3901176278893194e-05,
|
| 1958 |
+
"loss": 0.7499,
|
| 1959 |
+
"step": 325
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 0.7,
|
| 1963 |
+
"learning_rate": 4.332589339863512e-05,
|
| 1964 |
+
"loss": 0.7624,
|
| 1965 |
+
"step": 326
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.7,
|
| 1969 |
+
"learning_rate": 4.2753360498712305e-05,
|
| 1970 |
+
"loss": 0.7827,
|
| 1971 |
+
"step": 327
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.71,
|
| 1975 |
+
"learning_rate": 4.218360535996338e-05,
|
| 1976 |
+
"loss": 0.8106,
|
| 1977 |
+
"step": 328
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 0.71,
|
| 1981 |
+
"learning_rate": 4.16166556284428e-05,
|
| 1982 |
+
"loss": 0.8423,
|
| 1983 |
+
"step": 329
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.71,
|
| 1987 |
+
"learning_rate": 4.105253881407879e-05,
|
| 1988 |
+
"loss": 0.809,
|
| 1989 |
+
"step": 330
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.71,
|
| 1993 |
+
"learning_rate": 4.049128228933903e-05,
|
| 1994 |
+
"loss": 0.7933,
|
| 1995 |
+
"step": 331
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 0.71,
|
| 1999 |
+
"learning_rate": 3.993291328790209e-05,
|
| 2000 |
+
"loss": 0.7756,
|
| 2001 |
+
"step": 332
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 0.72,
|
| 2005 |
+
"learning_rate": 3.937745890333623e-05,
|
| 2006 |
+
"loss": 0.7669,
|
| 2007 |
+
"step": 333
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.72,
|
| 2011 |
+
"learning_rate": 3.8824946087784544e-05,
|
| 2012 |
+
"loss": 0.8394,
|
| 2013 |
+
"step": 334
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.72,
|
| 2017 |
+
"learning_rate": 3.827540165065746e-05,
|
| 2018 |
+
"loss": 0.7931,
|
| 2019 |
+
"step": 335
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 0.72,
|
| 2023 |
+
"learning_rate": 3.772885225733147e-05,
|
| 2024 |
+
"loss": 0.7199,
|
| 2025 |
+
"step": 336
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.72,
|
| 2029 |
+
"learning_rate": 3.718532442785565e-05,
|
| 2030 |
+
"loss": 0.8122,
|
| 2031 |
+
"step": 337
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.73,
|
| 2035 |
+
"learning_rate": 3.664484453566449e-05,
|
| 2036 |
+
"loss": 0.7875,
|
| 2037 |
+
"step": 338
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 0.73,
|
| 2041 |
+
"learning_rate": 3.6107438806298486e-05,
|
| 2042 |
+
"loss": 0.7569,
|
| 2043 |
+
"step": 339
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 0.73,
|
| 2047 |
+
"learning_rate": 3.557313331613145e-05,
|
| 2048 |
+
"loss": 0.7351,
|
| 2049 |
+
"step": 340
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.73,
|
| 2053 |
+
"learning_rate": 3.504195399110516e-05,
|
| 2054 |
+
"loss": 0.7618,
|
| 2055 |
+
"step": 341
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.74,
|
| 2059 |
+
"learning_rate": 3.45139266054715e-05,
|
| 2060 |
+
"loss": 0.8037,
|
| 2061 |
+
"step": 342
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.74,
|
| 2065 |
+
"learning_rate": 3.398907678054177e-05,
|
| 2066 |
+
"loss": 0.776,
|
| 2067 |
+
"step": 343
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.74,
|
| 2071 |
+
"learning_rate": 3.346742998344348e-05,
|
| 2072 |
+
"loss": 0.7853,
|
| 2073 |
+
"step": 344
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.74,
|
| 2077 |
+
"learning_rate": 3.29490115258845e-05,
|
| 2078 |
+
"loss": 0.7861,
|
| 2079 |
+
"step": 345
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.74,
|
| 2083 |
+
"learning_rate": 3.243384656292511e-05,
|
| 2084 |
+
"loss": 0.7843,
|
| 2085 |
+
"step": 346
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 0.75,
|
| 2089 |
+
"learning_rate": 3.1921960091757074e-05,
|
| 2090 |
+
"loss": 0.7517,
|
| 2091 |
+
"step": 347
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.75,
|
| 2095 |
+
"learning_rate": 3.141337695049117e-05,
|
| 2096 |
+
"loss": 0.7762,
|
| 2097 |
+
"step": 348
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.75,
|
| 2101 |
+
"learning_rate": 3.090812181695146e-05,
|
| 2102 |
+
"loss": 0.8029,
|
| 2103 |
+
"step": 349
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.75,
|
| 2107 |
+
"learning_rate": 3.040621920747827e-05,
|
| 2108 |
+
"loss": 0.7374,
|
| 2109 |
+
"step": 350
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.75,
|
| 2113 |
+
"learning_rate": 2.9907693475738307e-05,
|
| 2114 |
+
"loss": 0.7354,
|
| 2115 |
+
"step": 351
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.76,
|
| 2119 |
+
"learning_rate": 2.9412568811543174e-05,
|
| 2120 |
+
"loss": 0.7213,
|
| 2121 |
+
"step": 352
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.76,
|
| 2125 |
+
"learning_rate": 2.8920869239675387e-05,
|
| 2126 |
+
"loss": 0.815,
|
| 2127 |
+
"step": 353
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 0.76,
|
| 2131 |
+
"learning_rate": 2.843261861872296e-05,
|
| 2132 |
+
"loss": 0.8181,
|
| 2133 |
+
"step": 354
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.76,
|
| 2137 |
+
"learning_rate": 2.794784063992131e-05,
|
| 2138 |
+
"loss": 0.8089,
|
| 2139 |
+
"step": 355
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.77,
|
| 2143 |
+
"learning_rate": 2.7466558826003996e-05,
|
| 2144 |
+
"loss": 0.8126,
|
| 2145 |
+
"step": 356
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.77,
|
| 2149 |
+
"learning_rate": 2.6988796530061266e-05,
|
| 2150 |
+
"loss": 0.739,
|
| 2151 |
+
"step": 357
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.77,
|
| 2155 |
+
"learning_rate": 2.6514576934406777e-05,
|
| 2156 |
+
"loss": 0.7688,
|
| 2157 |
+
"step": 358
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.77,
|
| 2161 |
+
"learning_rate": 2.6043923049452913e-05,
|
| 2162 |
+
"loss": 0.6952,
|
| 2163 |
+
"step": 359
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.77,
|
| 2167 |
+
"learning_rate": 2.557685771259414e-05,
|
| 2168 |
+
"loss": 0.7665,
|
| 2169 |
+
"step": 360
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.78,
|
| 2173 |
+
"learning_rate": 2.5113403587098916e-05,
|
| 2174 |
+
"loss": 0.7898,
|
| 2175 |
+
"step": 361
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.78,
|
| 2179 |
+
"learning_rate": 2.4653583161009942e-05,
|
| 2180 |
+
"loss": 0.7712,
|
| 2181 |
+
"step": 362
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.78,
|
| 2185 |
+
"learning_rate": 2.4197418746053104e-05,
|
| 2186 |
+
"loss": 0.8119,
|
| 2187 |
+
"step": 363
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.78,
|
| 2191 |
+
"learning_rate": 2.3744932476554717e-05,
|
| 2192 |
+
"loss": 0.7542,
|
| 2193 |
+
"step": 364
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.78,
|
| 2197 |
+
"learning_rate": 2.3296146308367594e-05,
|
| 2198 |
+
"loss": 0.7718,
|
| 2199 |
+
"step": 365
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.79,
|
| 2203 |
+
"learning_rate": 2.2851082017805703e-05,
|
| 2204 |
+
"loss": 0.7335,
|
| 2205 |
+
"step": 366
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.79,
|
| 2209 |
+
"learning_rate": 2.2409761200587455e-05,
|
| 2210 |
+
"loss": 0.8223,
|
| 2211 |
+
"step": 367
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.79,
|
| 2215 |
+
"learning_rate": 2.1972205270787783e-05,
|
| 2216 |
+
"loss": 0.8132,
|
| 2217 |
+
"step": 368
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.79,
|
| 2221 |
+
"learning_rate": 2.1538435459799268e-05,
|
| 2222 |
+
"loss": 0.7725,
|
| 2223 |
+
"step": 369
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.8,
|
| 2227 |
+
"learning_rate": 2.1108472815301673e-05,
|
| 2228 |
+
"loss": 0.7681,
|
| 2229 |
+
"step": 370
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.8,
|
| 2233 |
+
"learning_rate": 2.068233820024088e-05,
|
| 2234 |
+
"loss": 0.7569,
|
| 2235 |
+
"step": 371
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.8,
|
| 2239 |
+
"learning_rate": 2.0260052291816445e-05,
|
| 2240 |
+
"loss": 0.7739,
|
| 2241 |
+
"step": 372
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.8,
|
| 2245 |
+
"learning_rate": 1.9841635580478325e-05,
|
| 2246 |
+
"loss": 0.7866,
|
| 2247 |
+
"step": 373
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 0.8,
|
| 2251 |
+
"learning_rate": 1.9427108368932533e-05,
|
| 2252 |
+
"loss": 0.7849,
|
| 2253 |
+
"step": 374
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.81,
|
| 2257 |
+
"learning_rate": 1.901649077115617e-05,
|
| 2258 |
+
"loss": 0.7662,
|
| 2259 |
+
"step": 375
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.81,
|
| 2263 |
+
"learning_rate": 1.860980271142133e-05,
|
| 2264 |
+
"loss": 0.6951,
|
| 2265 |
+
"step": 376
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.81,
|
| 2269 |
+
"learning_rate": 1.8207063923328237e-05,
|
| 2270 |
+
"loss": 0.7999,
|
| 2271 |
+
"step": 377
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.81,
|
| 2275 |
+
"learning_rate": 1.7808293948847942e-05,
|
| 2276 |
+
"loss": 0.7626,
|
| 2277 |
+
"step": 378
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.82,
|
| 2281 |
+
"learning_rate": 1.74135121373739e-05,
|
| 2282 |
+
"loss": 0.7663,
|
| 2283 |
+
"step": 379
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.82,
|
| 2287 |
+
"learning_rate": 1.702273764478318e-05,
|
| 2288 |
+
"loss": 0.7274,
|
| 2289 |
+
"step": 380
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 0.82,
|
| 2293 |
+
"learning_rate": 1.6635989432506903e-05,
|
| 2294 |
+
"loss": 0.7396,
|
| 2295 |
+
"step": 381
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.82,
|
| 2299 |
+
"learning_rate": 1.625328626661028e-05,
|
| 2300 |
+
"loss": 0.7429,
|
| 2301 |
+
"step": 382
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 0.82,
|
| 2305 |
+
"learning_rate": 1.587464671688187e-05,
|
| 2306 |
+
"loss": 0.7384,
|
| 2307 |
+
"step": 383
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.83,
|
| 2311 |
+
"learning_rate": 1.5500089155932807e-05,
|
| 2312 |
+
"loss": 0.7554,
|
| 2313 |
+
"step": 384
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.83,
|
| 2317 |
+
"learning_rate": 1.5129631758304941e-05,
|
| 2318 |
+
"loss": 0.7523,
|
| 2319 |
+
"step": 385
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.83,
|
| 2323 |
+
"learning_rate": 1.4763292499589298e-05,
|
| 2324 |
+
"loss": 0.7054,
|
| 2325 |
+
"step": 386
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 0.83,
|
| 2329 |
+
"learning_rate": 1.4401089155553582e-05,
|
| 2330 |
+
"loss": 0.6719,
|
| 2331 |
+
"step": 387
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 0.83,
|
| 2335 |
+
"learning_rate": 1.4043039301279903e-05,
|
| 2336 |
+
"loss": 0.7845,
|
| 2337 |
+
"step": 388
|
| 2338 |
+
},
|
| 2339 |
{
|
| 2340 |
"epoch": 0.84,
|
| 2341 |
+
"learning_rate": 1.3689160310311721e-05,
|
| 2342 |
+
"loss": 0.7605,
|
| 2343 |
+
"step": 389
|
| 2344 |
},
|
| 2345 |
{
|
| 2346 |
"epoch": 0.84,
|
| 2347 |
+
"learning_rate": 1.333946935381114e-05,
|
| 2348 |
+
"loss": 0.8225,
|
| 2349 |
+
"step": 390
|
| 2350 |
},
|
| 2351 |
{
|
| 2352 |
"epoch": 0.84,
|
| 2353 |
+
"learning_rate": 1.2993983399725373e-05,
|
| 2354 |
+
"loss": 0.6951,
|
| 2355 |
+
"step": 391
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.84,
|
| 2359 |
+
"learning_rate": 1.2652719211963726e-05,
|
| 2360 |
+
"loss": 0.7516,
|
| 2361 |
+
"step": 392
|
| 2362 |
},
|
| 2363 |
{
|
| 2364 |
"epoch": 0.85,
|
| 2365 |
+
"learning_rate": 1.2315693349583923e-05,
|
| 2366 |
+
"loss": 0.8173,
|
| 2367 |
+
"step": 393
|
| 2368 |
},
|
| 2369 |
{
|
| 2370 |
"epoch": 0.85,
|
| 2371 |
+
"learning_rate": 1.1982922165988808e-05,
|
| 2372 |
+
"loss": 0.7253,
|
| 2373 |
+
"step": 394
|
| 2374 |
},
|
| 2375 |
{
|
| 2376 |
"epoch": 0.85,
|
| 2377 |
+
"learning_rate": 1.1654421808132687e-05,
|
| 2378 |
+
"loss": 0.8076,
|
| 2379 |
+
"step": 395
|
| 2380 |
},
|
| 2381 |
{
|
| 2382 |
"epoch": 0.85,
|
| 2383 |
+
"learning_rate": 1.1330208215737937e-05,
|
| 2384 |
+
"loss": 0.8309,
|
| 2385 |
+
"step": 396
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.85,
|
| 2389 |
+
"learning_rate": 1.1010297120521528e-05,
|
| 2390 |
+
"loss": 0.7405,
|
| 2391 |
+
"step": 397
|
| 2392 |
},
|
| 2393 |
{
|
| 2394 |
"epoch": 0.86,
|
| 2395 |
+
"learning_rate": 1.0694704045431604e-05,
|
| 2396 |
+
"loss": 0.7937,
|
| 2397 |
+
"step": 398
|
| 2398 |
},
|
| 2399 |
{
|
| 2400 |
"epoch": 0.86,
|
| 2401 |
+
"learning_rate": 1.0383444303894452e-05,
|
| 2402 |
+
"loss": 0.7316,
|
| 2403 |
+
"step": 399
|
| 2404 |
},
|
| 2405 |
{
|
| 2406 |
"epoch": 0.86,
|
| 2407 |
+
"learning_rate": 1.007653299907122e-05,
|
| 2408 |
+
"loss": 0.756,
|
| 2409 |
+
"step": 400
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 0.86,
|
| 2413 |
+
"learning_rate": 9.77398502312531e-06,
|
| 2414 |
+
"loss": 0.7262,
|
| 2415 |
+
"step": 401
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 0.86,
|
| 2419 |
+
"learning_rate": 9.475815056499526e-06,
|
| 2420 |
+
"loss": 0.7558,
|
| 2421 |
+
"step": 402
|
| 2422 |
},
|
| 2423 |
{
|
| 2424 |
"epoch": 0.87,
|
| 2425 |
+
"learning_rate": 9.182037567204016e-06,
|
| 2426 |
+
"loss": 0.7594,
|
| 2427 |
+
"step": 403
|
| 2428 |
},
|
| 2429 |
{
|
| 2430 |
"epoch": 0.87,
|
| 2431 |
+
"learning_rate": 8.892666810113958e-06,
|
| 2432 |
+
"loss": 0.7306,
|
| 2433 |
+
"step": 404
|
| 2434 |
},
|
| 2435 |
{
|
| 2436 |
"epoch": 0.87,
|
| 2437 |
+
"learning_rate": 8.60771682627809e-06,
|
| 2438 |
+
"loss": 0.7518,
|
| 2439 |
+
"step": 405
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.87,
|
| 2443 |
+
"learning_rate": 8.327201442237276e-06,
|
| 2444 |
+
"loss": 0.763,
|
| 2445 |
+
"step": 406
|
| 2446 |
},
|
| 2447 |
{
|
| 2448 |
"epoch": 0.88,
|
| 2449 |
+
"learning_rate": 8.051134269353688e-06,
|
| 2450 |
+
"loss": 0.7824,
|
| 2451 |
+
"step": 407
|
| 2452 |
},
|
| 2453 |
{
|
| 2454 |
"epoch": 0.88,
|
| 2455 |
+
"learning_rate": 7.779528703150263e-06,
|
| 2456 |
+
"loss": 0.7717,
|
| 2457 |
+
"step": 408
|
| 2458 |
},
|
| 2459 |
{
|
| 2460 |
"epoch": 0.88,
|
| 2461 |
+
"learning_rate": 7.512397922660852e-06,
|
| 2462 |
+
"loss": 0.7227,
|
| 2463 |
+
"step": 409
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.88,
|
| 2467 |
+
"learning_rate": 7.249754889790539e-06,
|
| 2468 |
+
"loss": 0.7398,
|
| 2469 |
+
"step": 410
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 0.88,
|
| 2473 |
+
"learning_rate": 6.99161234868686e-06,
|
| 2474 |
+
"loss": 0.8,
|
| 2475 |
+
"step": 411
|
| 2476 |
},
|
| 2477 |
{
|
| 2478 |
"epoch": 0.89,
|
| 2479 |
+
"learning_rate": 6.737982825121392e-06,
|
| 2480 |
+
"loss": 0.7839,
|
| 2481 |
+
"step": 412
|
| 2482 |
},
|
| 2483 |
{
|
| 2484 |
"epoch": 0.89,
|
| 2485 |
+
"learning_rate": 6.488878625881867e-06,
|
| 2486 |
+
"loss": 0.7311,
|
| 2487 |
+
"step": 413
|
| 2488 |
},
|
| 2489 |
{
|
| 2490 |
"epoch": 0.89,
|
| 2491 |
+
"learning_rate": 6.244311838175143e-06,
|
| 2492 |
+
"loss": 0.7871,
|
| 2493 |
+
"step": 414
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 0.89,
|
| 2497 |
+
"learning_rate": 6.0042943290406384e-06,
|
| 2498 |
+
"loss": 0.8087,
|
| 2499 |
+
"step": 415
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 0.89,
|
| 2503 |
+
"learning_rate": 5.768837744774547e-06,
|
| 2504 |
+
"loss": 0.7269,
|
| 2505 |
+
"step": 416
|
| 2506 |
},
|
| 2507 |
{
|
| 2508 |
"epoch": 0.9,
|
| 2509 |
+
"learning_rate": 5.537953510364613e-06,
|
| 2510 |
+
"loss": 0.6742,
|
| 2511 |
+
"step": 417
|
| 2512 |
},
|
| 2513 |
{
|
| 2514 |
"epoch": 0.9,
|
| 2515 |
+
"learning_rate": 5.3116528289359425e-06,
|
| 2516 |
+
"loss": 0.752,
|
| 2517 |
+
"step": 418
|
| 2518 |
},
|
| 2519 |
{
|
| 2520 |
"epoch": 0.9,
|
| 2521 |
+
"learning_rate": 5.089946681207247e-06,
|
| 2522 |
+
"loss": 0.7432,
|
| 2523 |
+
"step": 419
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.9,
|
| 2527 |
+
"learning_rate": 4.872845824958106e-06,
|
| 2528 |
+
"loss": 0.7794,
|
| 2529 |
+
"step": 420
|
| 2530 |
},
|
| 2531 |
{
|
| 2532 |
"epoch": 0.91,
|
| 2533 |
+
"learning_rate": 4.660360794506946e-06,
|
| 2534 |
+
"loss": 0.7081,
|
| 2535 |
+
"step": 421
|
| 2536 |
},
|
| 2537 |
{
|
| 2538 |
"epoch": 0.91,
|
| 2539 |
+
"learning_rate": 4.452501900199901e-06,
|
| 2540 |
+
"loss": 0.7495,
|
| 2541 |
+
"step": 422
|
| 2542 |
},
|
| 2543 |
{
|
| 2544 |
"epoch": 0.91,
|
| 2545 |
+
"learning_rate": 4.249279227910485e-06,
|
| 2546 |
+
"loss": 0.7318,
|
| 2547 |
+
"step": 423
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.91,
|
| 2551 |
+
"learning_rate": 4.050702638550275e-06,
|
| 2552 |
+
"loss": 0.7864,
|
| 2553 |
+
"step": 424
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 0.91,
|
| 2557 |
+
"learning_rate": 3.856781767590334e-06,
|
| 2558 |
+
"loss": 0.7636,
|
| 2559 |
+
"step": 425
|
| 2560 |
},
|
| 2561 |
{
|
| 2562 |
"epoch": 0.92,
|
| 2563 |
+
"learning_rate": 3.667526024593759e-06,
|
| 2564 |
+
"loss": 0.7483,
|
| 2565 |
+
"step": 426
|
| 2566 |
},
|
| 2567 |
{
|
| 2568 |
"epoch": 0.92,
|
| 2569 |
+
"learning_rate": 3.4829445927590852e-06,
|
| 2570 |
+
"loss": 0.717,
|
| 2571 |
+
"step": 427
|
| 2572 |
},
|
| 2573 |
{
|
| 2574 |
"epoch": 0.92,
|
| 2575 |
+
"learning_rate": 3.303046428474643e-06,
|
| 2576 |
+
"loss": 0.7798,
|
| 2577 |
+
"step": 428
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 0.92,
|
| 2581 |
+
"learning_rate": 3.1278402608840185e-06,
|
| 2582 |
+
"loss": 0.7522,
|
| 2583 |
+
"step": 429
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 0.92,
|
| 2587 |
+
"learning_rate": 2.9573345914624794e-06,
|
| 2588 |
+
"loss": 0.8183,
|
| 2589 |
+
"step": 430
|
| 2590 |
},
|
| 2591 |
{
|
| 2592 |
"epoch": 0.93,
|
| 2593 |
+
"learning_rate": 2.7915376936044623e-06,
|
| 2594 |
+
"loss": 0.7446,
|
| 2595 |
+
"step": 431
|
| 2596 |
},
|
| 2597 |
{
|
| 2598 |
"epoch": 0.93,
|
| 2599 |
+
"learning_rate": 2.6304576122221035e-06,
|
| 2600 |
+
"loss": 0.7461,
|
| 2601 |
+
"step": 432
|
| 2602 |
},
|
| 2603 |
{
|
| 2604 |
"epoch": 0.93,
|
| 2605 |
+
"learning_rate": 2.4741021633549077e-06,
|
| 2606 |
+
"loss": 0.7495,
|
| 2607 |
+
"step": 433
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.93,
|
| 2611 |
+
"learning_rate": 2.322478933790506e-06,
|
| 2612 |
+
"loss": 0.7334,
|
| 2613 |
+
"step": 434
|
| 2614 |
},
|
| 2615 |
{
|
| 2616 |
"epoch": 0.94,
|
| 2617 |
+
"learning_rate": 2.175595280696463e-06,
|
| 2618 |
+
"loss": 0.8137,
|
| 2619 |
+
"step": 435
|
| 2620 |
},
|
| 2621 |
{
|
| 2622 |
"epoch": 0.94,
|
| 2623 |
+
"learning_rate": 2.033458331263338e-06,
|
| 2624 |
+
"loss": 0.7179,
|
| 2625 |
+
"step": 436
|
| 2626 |
},
|
| 2627 |
{
|
| 2628 |
"epoch": 0.94,
|
| 2629 |
+
"learning_rate": 1.8960749823588532e-06,
|
| 2630 |
+
"loss": 0.7921,
|
| 2631 |
+
"step": 437
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 0.94,
|
| 2635 |
+
"learning_rate": 1.7634519001931914e-06,
|
| 2636 |
+
"loss": 0.8521,
|
| 2637 |
+
"step": 438
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 0.94,
|
| 2641 |
+
"learning_rate": 1.6355955199956142e-06,
|
| 2642 |
+
"loss": 0.671,
|
| 2643 |
+
"step": 439
|
| 2644 |
},
|
| 2645 |
{
|
| 2646 |
"epoch": 0.95,
|
| 2647 |
+
"learning_rate": 1.5125120457021303e-06,
|
| 2648 |
+
"loss": 0.7294,
|
| 2649 |
+
"step": 440
|
| 2650 |
},
|
| 2651 |
{
|
| 2652 |
"epoch": 0.95,
|
| 2653 |
+
"learning_rate": 1.3942074496545165e-06,
|
| 2654 |
+
"loss": 0.8334,
|
| 2655 |
+
"step": 441
|
| 2656 |
},
|
| 2657 |
{
|
| 2658 |
"epoch": 0.95,
|
| 2659 |
+
"learning_rate": 1.2806874723104822e-06,
|
| 2660 |
+
"loss": 0.7709,
|
| 2661 |
+
"step": 442
|
| 2662 |
},
|
| 2663 |
{
|
| 2664 |
"epoch": 0.95,
|
| 2665 |
+
"learning_rate": 1.1719576219651585e-06,
|
| 2666 |
+
"loss": 0.7273,
|
| 2667 |
+
"step": 443
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 0.95,
|
| 2671 |
+
"learning_rate": 1.0680231744837898e-06,
|
| 2672 |
+
"loss": 0.7483,
|
| 2673 |
+
"step": 444
|
| 2674 |
},
|
| 2675 |
{
|
| 2676 |
"epoch": 0.96,
|
| 2677 |
+
"learning_rate": 9.68889173045806e-07,
|
| 2678 |
+
"loss": 0.7046,
|
| 2679 |
+
"step": 445
|
| 2680 |
},
|
| 2681 |
{
|
| 2682 |
"epoch": 0.96,
|
| 2683 |
+
"learning_rate": 8.745604279000175e-07,
|
| 2684 |
+
"loss": 0.8129,
|
| 2685 |
+
"step": 446
|
| 2686 |
},
|
| 2687 |
{
|
| 2688 |
"epoch": 0.96,
|
| 2689 |
+
"learning_rate": 7.850415161312463e-07,
|
| 2690 |
+
"loss": 0.7924,
|
| 2691 |
+
"step": 447
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 0.96,
|
| 2695 |
+
"learning_rate": 7.003367814382933e-07,
|
| 2696 |
+
"loss": 0.7779,
|
| 2697 |
+
"step": 448
|
| 2698 |
},
|
| 2699 |
{
|
| 2700 |
"epoch": 0.97,
|
| 2701 |
+
"learning_rate": 6.204503339230505e-07,
|
| 2702 |
+
"loss": 0.714,
|
| 2703 |
+
"step": 449
|
| 2704 |
},
|
| 2705 |
{
|
| 2706 |
"epoch": 0.97,
|
| 2707 |
+
"learning_rate": 5.453860498911945e-07,
|
| 2708 |
+
"loss": 0.8237,
|
| 2709 |
+
"step": 450
|
| 2710 |
},
|
| 2711 |
{
|
| 2712 |
"epoch": 0.97,
|
| 2713 |
+
"learning_rate": 4.7514757166400304e-07,
|
| 2714 |
+
"loss": 0.7992,
|
| 2715 |
+
"step": 451
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 0.97,
|
| 2719 |
+
"learning_rate": 4.097383074016636e-07,
|
| 2720 |
+
"loss": 0.779,
|
| 2721 |
+
"step": 452
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 0.97,
|
| 2725 |
+
"learning_rate": 3.4916143093790543e-07,
|
| 2726 |
+
"loss": 0.7296,
|
| 2727 |
+
"step": 453
|
| 2728 |
},
|
| 2729 |
{
|
| 2730 |
"epoch": 0.98,
|
| 2731 |
+
"learning_rate": 2.934198816259559e-07,
|
| 2732 |
+
"loss": 0.7767,
|
| 2733 |
+
"step": 454
|
| 2734 |
},
|
| 2735 |
{
|
| 2736 |
"epoch": 0.98,
|
| 2737 |
+
"learning_rate": 2.42516364195966e-07,
|
| 2738 |
+
"loss": 0.7701,
|
| 2739 |
+
"step": 455
|
| 2740 |
},
|
| 2741 |
{
|
| 2742 |
"epoch": 0.98,
|
| 2743 |
+
"learning_rate": 1.9645334862373744e-07,
|
| 2744 |
+
"loss": 0.755,
|
| 2745 |
+
"step": 456
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 0.98,
|
| 2749 |
+
"learning_rate": 1.5523307001088505e-07,
|
| 2750 |
+
"loss": 0.7407,
|
| 2751 |
+
"step": 457
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 0.98,
|
| 2755 |
+
"learning_rate": 1.1885752847637932e-07,
|
| 2756 |
+
"loss": 0.7914,
|
| 2757 |
+
"step": 458
|
| 2758 |
},
|
| 2759 |
{
|
| 2760 |
"epoch": 0.99,
|
| 2761 |
+
"learning_rate": 8.732848905947943e-08,
|
| 2762 |
+
"loss": 0.7358,
|
| 2763 |
+
"step": 459
|
| 2764 |
},
|
| 2765 |
{
|
| 2766 |
"epoch": 0.99,
|
| 2767 |
+
"learning_rate": 6.064748163413514e-08,
|
| 2768 |
+
"loss": 0.7522,
|
| 2769 |
+
"step": 460
|
| 2770 |
},
|
| 2771 |
{
|
| 2772 |
"epoch": 0.99,
|
| 2773 |
+
"learning_rate": 3.881580083469061e-08,
|
| 2774 |
+
"loss": 0.7588,
|
| 2775 |
+
"step": 461
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 0.99,
|
| 2779 |
+
"learning_rate": 2.1834505993134634e-08,
|
| 2780 |
+
"loss": 0.7715,
|
| 2781 |
+
"step": 462
|
| 2782 |
},
|
| 2783 |
{
|
| 2784 |
"epoch": 1.0,
|
| 2785 |
+
"learning_rate": 9.704421087619597e-09,
|
| 2786 |
+
"loss": 0.8239,
|
| 2787 |
+
"step": 463
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 1.0,
|
| 2791 |
+
"learning_rate": 2.4261347025489323e-09,
|
| 2792 |
+
"loss": 0.7666,
|
| 2793 |
+
"step": 464
|
| 2794 |
},
|
| 2795 |
{
|
| 2796 |
"epoch": 1.0,
|
| 2797 |
"learning_rate": 0.0,
|
| 2798 |
+
"loss": 0.7757,
|
| 2799 |
+
"step": 465
|
| 2800 |
},
|
| 2801 |
{
|
| 2802 |
"epoch": 1.0,
|
| 2803 |
+
"step": 465,
|
| 2804 |
+
"total_flos": 53904513761280.0,
|
| 2805 |
+
"train_loss": 0.8270689119574844,
|
| 2806 |
+
"train_runtime": 4897.453,
|
| 2807 |
+
"train_samples_per_second": 6.069,
|
| 2808 |
+
"train_steps_per_second": 0.095
|
| 2809 |
}
|
| 2810 |
],
|
| 2811 |
"logging_steps": 1.0,
|
| 2812 |
+
"max_steps": 465,
|
| 2813 |
"num_input_tokens_seen": 0,
|
| 2814 |
"num_train_epochs": 1,
|
| 2815 |
"save_steps": 50000,
|
| 2816 |
+
"total_flos": 53904513761280.0,
|
| 2817 |
"train_batch_size": 16,
|
| 2818 |
"trial_name": null,
|
| 2819 |
"trial_params": null
|