Upload folder using huggingface_hub
Browse files- adapter_config.json +8 -5
- adapter_model.bin +2 -2
- optimizer.pt +2 -2
- rng_state.pth +1 -1
- special_tokens_map.json +22 -4
- tokenizer.json +0 -0
- tokenizer_config.json +31 -6
- trainer_state.json +251 -251
- training_args.bin +1 -1
adapter_config.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"auto_mapping": null,
|
| 3 |
-
"base_model_name_or_path": "
|
| 4 |
"bias": "none",
|
| 5 |
"fan_in_fan_out": false,
|
| 6 |
"inference_mode": true,
|
|
@@ -14,10 +14,13 @@
|
|
| 14 |
"r": 64,
|
| 15 |
"revision": null,
|
| 16 |
"target_modules": [
|
| 17 |
-
"
|
| 18 |
-
"
|
| 19 |
-
"
|
| 20 |
-
"
|
|
|
|
|
|
|
|
|
|
| 21 |
],
|
| 22 |
"task_type": "CAUSAL_LM"
|
| 23 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "PY007/TinyLlama-1.1B-step-50K-105b",
|
| 4 |
"bias": "none",
|
| 5 |
"fan_in_fan_out": false,
|
| 6 |
"inference_mode": true,
|
|
|
|
| 14 |
"r": 64,
|
| 15 |
"revision": null,
|
| 16 |
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"up_proj",
|
| 19 |
+
"o_proj",
|
| 20 |
+
"k_proj",
|
| 21 |
+
"down_proj",
|
| 22 |
+
"gate_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
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e379df784a926b119a08754e0c1bcf0436ce7e3d0c679e5884d4aac0958f3267
|
| 3 |
+
size 201961693
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:866469e76f42f15037e2a33896346ed8e6707641219b7984a648ffd98fd03dc4
|
| 3 |
+
size 403891013
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14575
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7cfd6cbd0f2b28903860356a761623a6fc5609358d35a88d02cb06fc77ba45cc
|
| 3 |
size 14575
|
special_tokens_map.json
CHANGED
|
@@ -1,6 +1,24 @@
|
|
| 1 |
{
|
| 2 |
-
"bos_token":
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
}
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1,9 +1,34 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
"model_max_length": 1000000000000000019884624838656,
|
| 7 |
-
"
|
| 8 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"__type": "AddedToken",
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"clean_up_tokenization_spaces": false,
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "</s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"legacy": false,
|
| 20 |
"model_max_length": 1000000000000000019884624838656,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"padding_side": "right",
|
| 23 |
+
"sp_model_kwargs": {},
|
| 24 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
+
"unk_token": {
|
| 26 |
+
"__type": "AddedToken",
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": false,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
},
|
| 33 |
+
"use_default_system_prompt": true
|
| 34 |
}
|
trainer_state.json
CHANGED
|
@@ -11,1501 +11,1501 @@
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 2.5e-06,
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.01,
|
| 19 |
"learning_rate": 5e-06,
|
| 20 |
-
"loss": 1.
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
"learning_rate": 7.500000000000001e-06,
|
| 26 |
-
"loss": 1.
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.02,
|
| 31 |
"learning_rate": 1e-05,
|
| 32 |
-
"loss": 1.
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.02,
|
| 37 |
"learning_rate": 1.25e-05,
|
| 38 |
-
"loss":
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.02,
|
| 43 |
"learning_rate": 1.5000000000000002e-05,
|
| 44 |
-
"loss": 1.
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.03,
|
| 49 |
"learning_rate": 1.7500000000000002e-05,
|
| 50 |
-
"loss":
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03,
|
| 55 |
"learning_rate": 2e-05,
|
| 56 |
-
"loss": 2.
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.03,
|
| 61 |
"learning_rate": 1.9999213590779245e-05,
|
| 62 |
-
"loss": 2.
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.04,
|
| 67 |
"learning_rate": 1.9996854486804874e-05,
|
| 68 |
-
"loss": 2.
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.04,
|
| 73 |
"learning_rate": 1.9992923059121107e-05,
|
| 74 |
-
"loss": 2.
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.05,
|
| 79 |
"learning_rate": 1.998741992607014e-05,
|
| 80 |
-
"loss": 2.
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.05,
|
| 85 |
"learning_rate": 1.9980345953194892e-05,
|
| 86 |
-
"loss": 2.
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.05,
|
| 91 |
"learning_rate": 1.9971702253102856e-05,
|
| 92 |
-
"loss": 2.
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.06,
|
| 97 |
"learning_rate": 1.9961490185291128e-05,
|
| 98 |
-
"loss": 2.
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.06,
|
| 103 |
"learning_rate": 1.9949711355932565e-05,
|
| 104 |
-
"loss": 2.
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.07,
|
| 109 |
"learning_rate": 1.993636761762317e-05,
|
| 110 |
-
"loss": 2.
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.07,
|
| 115 |
"learning_rate": 1.9921461069090715e-05,
|
| 116 |
-
"loss": 2.
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.07,
|
| 121 |
"learning_rate": 1.990499405486464e-05,
|
| 122 |
-
"loss": 2.
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.08,
|
| 127 |
"learning_rate": 1.9886969164907306e-05,
|
| 128 |
-
"loss": 2.
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.08,
|
| 133 |
"learning_rate": 1.9867389234206655e-05,
|
| 134 |
-
"loss": 2.
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.09,
|
| 139 |
"learning_rate": 1.984625734233029e-05,
|
| 140 |
-
"loss": 2.
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.09,
|
| 145 |
"learning_rate": 1.982357681294114e-05,
|
| 146 |
-
"loss":
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.09,
|
| 151 |
"learning_rate": 1.9799351213274688e-05,
|
| 152 |
-
"loss":
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.1,
|
| 157 |
"learning_rate": 1.9773584353577928e-05,
|
| 158 |
-
"loss":
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.1,
|
| 163 |
"learning_rate": 1.974628028651007e-05,
|
| 164 |
-
"loss": 1.
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 0.1,
|
| 169 |
"learning_rate": 1.9717443306505143e-05,
|
| 170 |
-
"loss": 1.
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 0.11,
|
| 175 |
"learning_rate": 1.9687077949096528e-05,
|
| 176 |
-
"loss": 1.
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.11,
|
| 181 |
"learning_rate": 1.9655188990203648e-05,
|
| 182 |
-
"loss": 1.
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 0.12,
|
| 187 |
"learning_rate": 1.962178144538076e-05,
|
| 188 |
-
"loss": 1.
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.12,
|
| 193 |
"learning_rate": 1.9586860569028126e-05,
|
| 194 |
-
"loss": 1.
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.12,
|
| 199 |
"learning_rate": 1.9550431853565577e-05,
|
| 200 |
-
"loss": 1.
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.13,
|
| 205 |
"learning_rate": 1.951250102856866e-05,
|
| 206 |
-
"loss": 1.
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.13,
|
| 211 |
"learning_rate": 1.9473074059867476e-05,
|
| 212 |
-
"loss": 1.
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 0.14,
|
| 217 |
"learning_rate": 1.943215714860838e-05,
|
| 218 |
-
"loss": 1.
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.14,
|
| 223 |
"learning_rate": 1.9389756730278625e-05,
|
| 224 |
-
"loss":
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 0.14,
|
| 229 |
"learning_rate": 1.9345879473694207e-05,
|
| 230 |
-
"loss": 1.
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 0.15,
|
| 235 |
"learning_rate": 1.930053227995095e-05,
|
| 236 |
-
"loss":
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.15,
|
| 241 |
"learning_rate": 1.925372228133912e-05,
|
| 242 |
-
"loss": 2.
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 0.15,
|
| 247 |
"learning_rate": 1.9205456840221614e-05,
|
| 248 |
-
"loss": 2.
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.16,
|
| 253 |
"learning_rate": 1.9155743547876026e-05,
|
| 254 |
-
"loss": 2.
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 0.16,
|
| 259 |
"learning_rate": 1.9104590223300656e-05,
|
| 260 |
-
"loss":
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.17,
|
| 265 |
"learning_rate": 1.9052004911984722e-05,
|
| 266 |
-
"loss":
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 0.17,
|
| 271 |
"learning_rate": 1.8997995884642968e-05,
|
| 272 |
-
"loss": 2.
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 0.17,
|
| 277 |
"learning_rate": 1.8942571635914814e-05,
|
| 278 |
-
"loss": 2.
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.18,
|
| 283 |
"learning_rate": 1.888574088302831e-05,
|
| 284 |
-
"loss": 2.
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 0.18,
|
| 289 |
"learning_rate": 1.8827512564429076e-05,
|
| 290 |
-
"loss":
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 0.19,
|
| 295 |
"learning_rate": 1.8767895838374438e-05,
|
| 296 |
-
"loss":
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.19,
|
| 301 |
"learning_rate": 1.870690008149302e-05,
|
| 302 |
-
"loss":
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.19,
|
| 307 |
"learning_rate": 1.8644534887309947e-05,
|
| 308 |
-
"loss": 1.
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 0.2,
|
| 313 |
"learning_rate": 1.8580810064737967e-05,
|
| 314 |
-
"loss": 1.
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 0.2,
|
| 319 |
"learning_rate": 1.8515735636534694e-05,
|
| 320 |
-
"loss": 1.
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 0.2,
|
| 325 |
"learning_rate": 1.8449321837726207e-05,
|
| 326 |
-
"loss": 1.
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 0.21,
|
| 331 |
"learning_rate": 1.8381579113997255e-05,
|
| 332 |
-
"loss": 1.
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 0.21,
|
| 337 |
"learning_rate": 1.8312518120048354e-05,
|
| 338 |
-
"loss": 1.
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 0.22,
|
| 343 |
"learning_rate": 1.8242149717919993e-05,
|
| 344 |
-
"loss": 1.
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22,
|
| 349 |
"learning_rate": 1.8170484975284233e-05,
|
| 350 |
-
"loss": 1.
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 0.22,
|
| 355 |
"learning_rate": 1.8097535163703947e-05,
|
| 356 |
-
"loss": 1.
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 0.23,
|
| 361 |
"learning_rate": 1.8023311756860036e-05,
|
| 362 |
-
"loss": 1.
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 0.23,
|
| 367 |
"learning_rate": 1.794782642874681e-05,
|
| 368 |
-
"loss": 1.
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 0.24,
|
| 373 |
"learning_rate": 1.7871091051835873e-05,
|
| 374 |
-
"loss": 1.
|
| 375 |
"step": 122
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 0.24,
|
| 379 |
"learning_rate": 1.7793117695208828e-05,
|
| 380 |
-
"loss": 1.
|
| 381 |
"step": 124
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 0.24,
|
| 385 |
"learning_rate": 1.771391862265899e-05,
|
| 386 |
-
"loss": 1.
|
| 387 |
"step": 126
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.25,
|
| 391 |
"learning_rate": 1.763350629076255e-05,
|
| 392 |
-
"loss": 1.
|
| 393 |
"step": 128
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 0.25,
|
| 397 |
"learning_rate": 1.7551893346919363e-05,
|
| 398 |
-
"loss": 1.
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 0.26,
|
| 403 |
"learning_rate": 1.7469092627363738e-05,
|
| 404 |
-
"loss": 1.
|
| 405 |
"step": 132
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 0.26,
|
| 409 |
"learning_rate": 1.7385117155145547e-05,
|
| 410 |
-
"loss": 2.
|
| 411 |
"step": 134
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 0.26,
|
| 415 |
"learning_rate": 1.729998013808192e-05,
|
| 416 |
-
"loss": 2.
|
| 417 |
"step": 136
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 0.27,
|
| 421 |
"learning_rate": 1.7213694966679913e-05,
|
| 422 |
-
"loss": 2.
|
| 423 |
"step": 138
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 0.27,
|
| 427 |
"learning_rate": 1.7126275212030403e-05,
|
| 428 |
-
"loss": 2.
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.27,
|
| 433 |
"learning_rate": 1.7037734623673616e-05,
|
| 434 |
-
"loss": 2.
|
| 435 |
"step": 142
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 0.28,
|
| 439 |
"learning_rate": 1.6948087127436573e-05,
|
| 440 |
-
"loss": 1.
|
| 441 |
"step": 144
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 0.28,
|
| 445 |
"learning_rate": 1.68573468232428e-05,
|
| 446 |
-
"loss": 1.
|
| 447 |
"step": 146
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 0.29,
|
| 451 |
"learning_rate": 1.676552798289469e-05,
|
| 452 |
-
"loss": 1.
|
| 453 |
"step": 148
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 0.29,
|
| 457 |
"learning_rate": 1.667264504782877e-05,
|
| 458 |
-
"loss": 1.
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 0.29,
|
| 463 |
"learning_rate": 1.6578712626844365e-05,
|
| 464 |
-
"loss": 1.
|
| 465 |
"step": 152
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 0.3,
|
| 469 |
"learning_rate": 1.648374549380587e-05,
|
| 470 |
-
"loss": 1.
|
| 471 |
"step": 154
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.3,
|
| 475 |
"learning_rate": 1.63877585853191e-05,
|
| 476 |
-
"loss": 1.
|
| 477 |
"step": 156
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 0.31,
|
| 481 |
"learning_rate": 1.629076699838204e-05,
|
| 482 |
-
"loss": 1.
|
| 483 |
"step": 158
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 0.31,
|
| 487 |
"learning_rate": 1.6192785988010354e-05,
|
| 488 |
-
"loss": 1.
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.31,
|
| 493 |
"learning_rate": 1.6093830964838036e-05,
|
| 494 |
-
"loss": 1.
|
| 495 |
"step": 162
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 0.32,
|
| 499 |
"learning_rate": 1.5993917492693623e-05,
|
| 500 |
-
"loss": 1.
|
| 501 |
"step": 164
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 0.32,
|
| 505 |
"learning_rate": 1.5893061286152275e-05,
|
| 506 |
-
"loss": 1.
|
| 507 |
"step": 166
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 0.32,
|
| 511 |
"learning_rate": 1.5791278208064142e-05,
|
| 512 |
-
"loss": 1.
|
| 513 |
"step": 168
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.33,
|
| 517 |
"learning_rate": 1.5688584267059453e-05,
|
| 518 |
-
"loss": 1.
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.33,
|
| 523 |
"learning_rate": 1.5584995615030634e-05,
|
| 524 |
-
"loss": 1.
|
| 525 |
"step": 172
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 0.34,
|
| 529 |
"learning_rate": 1.5480528544591907e-05,
|
| 530 |
-
"loss": 1.
|
| 531 |
"step": 174
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 0.34,
|
| 535 |
"learning_rate": 1.537519948651676e-05,
|
| 536 |
-
"loss": 1.
|
| 537 |
"step": 176
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 0.34,
|
| 541 |
"learning_rate": 1.52690250071537e-05,
|
| 542 |
-
"loss": 1.
|
| 543 |
"step": 178
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 0.35,
|
| 547 |
"learning_rate": 1.5162021805820633e-05,
|
| 548 |
-
"loss": 1.
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 0.35,
|
| 553 |
"learning_rate": 1.50542067121784e-05,
|
| 554 |
-
"loss": 1.
|
| 555 |
"step": 182
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.36,
|
| 559 |
"learning_rate": 1.4945596683583754e-05,
|
| 560 |
-
"loss":
|
| 561 |
"step": 184
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 0.36,
|
| 565 |
"learning_rate": 1.4836208802422287e-05,
|
| 566 |
-
"loss": 1.
|
| 567 |
"step": 186
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 0.36,
|
| 571 |
"learning_rate": 1.4726060273421675e-05,
|
| 572 |
-
"loss": 1.
|
| 573 |
"step": 188
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 0.37,
|
| 577 |
"learning_rate": 1.4615168420945689e-05,
|
| 578 |
-
"loss": 1.
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 0.37,
|
| 583 |
"learning_rate": 1.450355068626939e-05,
|
| 584 |
-
"loss": 2.
|
| 585 |
"step": 192
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 0.38,
|
| 589 |
"learning_rate": 1.4391224624835928e-05,
|
| 590 |
-
"loss":
|
| 591 |
"step": 194
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 0.38,
|
| 595 |
"learning_rate": 1.4278207903495388e-05,
|
| 596 |
-
"loss":
|
| 597 |
"step": 196
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.38,
|
| 601 |
"learning_rate": 1.4164518297726129e-05,
|
| 602 |
-
"loss": 1.
|
| 603 |
"step": 198
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 0.39,
|
| 607 |
"learning_rate": 1.4050173688839e-05,
|
| 608 |
-
"loss": 1.
|
| 609 |
"step": 200
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 0.39,
|
| 613 |
"learning_rate": 1.3935192061164957e-05,
|
| 614 |
-
"loss": 1.
|
| 615 |
"step": 202
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 0.39,
|
| 619 |
"learning_rate": 1.3819591499226446e-05,
|
| 620 |
-
"loss": 1.
|
| 621 |
"step": 204
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"epoch": 0.4,
|
| 625 |
"learning_rate": 1.3703390184893034e-05,
|
| 626 |
-
"loss": 1.
|
| 627 |
"step": 206
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"epoch": 0.4,
|
| 631 |
"learning_rate": 1.3586606394521729e-05,
|
| 632 |
-
"loss": 1.
|
| 633 |
"step": 208
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 0.41,
|
| 637 |
"learning_rate": 1.3469258496082452e-05,
|
| 638 |
-
"loss": 1.
|
| 639 |
"step": 210
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.41,
|
| 643 |
"learning_rate": 1.3351364946269071e-05,
|
| 644 |
-
"loss": 1.
|
| 645 |
"step": 212
|
| 646 |
},
|
| 647 |
{
|
| 648 |
"epoch": 0.41,
|
| 649 |
"learning_rate": 1.323294428759652e-05,
|
| 650 |
-
"loss": 1.
|
| 651 |
"step": 214
|
| 652 |
},
|
| 653 |
{
|
| 654 |
"epoch": 0.42,
|
| 655 |
"learning_rate": 1.311401514548438e-05,
|
| 656 |
-
"loss": 1.
|
| 657 |
"step": 216
|
| 658 |
},
|
| 659 |
{
|
| 660 |
"epoch": 0.42,
|
| 661 |
"learning_rate": 1.2994596225327442e-05,
|
| 662 |
-
"loss": 1.
|
| 663 |
"step": 218
|
| 664 |
},
|
| 665 |
{
|
| 666 |
"epoch": 0.43,
|
| 667 |
"learning_rate": 1.2874706309553697e-05,
|
| 668 |
-
"loss": 1.
|
| 669 |
"step": 220
|
| 670 |
},
|
| 671 |
{
|
| 672 |
"epoch": 0.43,
|
| 673 |
"learning_rate": 1.2754364254670192e-05,
|
| 674 |
-
"loss": 1.
|
| 675 |
"step": 222
|
| 676 |
},
|
| 677 |
{
|
| 678 |
"epoch": 0.43,
|
| 679 |
"learning_rate": 1.263358898829725e-05,
|
| 680 |
-
"loss": 1.
|
| 681 |
"step": 224
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44,
|
| 685 |
"learning_rate": 1.251239950619149e-05,
|
| 686 |
-
"loss": 1.
|
| 687 |
"step": 226
|
| 688 |
},
|
| 689 |
{
|
| 690 |
"epoch": 0.44,
|
| 691 |
"learning_rate": 1.2390814869258155e-05,
|
| 692 |
-
"loss": 1.
|
| 693 |
"step": 228
|
| 694 |
},
|
| 695 |
{
|
| 696 |
"epoch": 0.44,
|
| 697 |
"learning_rate": 1.2268854200553158e-05,
|
| 698 |
-
"loss": 1.
|
| 699 |
"step": 230
|
| 700 |
},
|
| 701 |
{
|
| 702 |
"epoch": 0.45,
|
| 703 |
"learning_rate": 1.2146536682275388e-05,
|
| 704 |
-
"loss":
|
| 705 |
"step": 232
|
| 706 |
},
|
| 707 |
{
|
| 708 |
"epoch": 0.45,
|
| 709 |
"learning_rate": 1.2023881552749693e-05,
|
| 710 |
-
"loss": 1.
|
| 711 |
"step": 234
|
| 712 |
},
|
| 713 |
{
|
| 714 |
"epoch": 0.46,
|
| 715 |
"learning_rate": 1.1900908103401037e-05,
|
| 716 |
-
"loss": 1.
|
| 717 |
"step": 236
|
| 718 |
},
|
| 719 |
{
|
| 720 |
"epoch": 0.46,
|
| 721 |
"learning_rate": 1.1777635675720313e-05,
|
| 722 |
-
"loss":
|
| 723 |
"step": 238
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.46,
|
| 727 |
"learning_rate": 1.1654083658222287e-05,
|
| 728 |
-
"loss": 1.
|
| 729 |
"step": 240
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.47,
|
| 733 |
"learning_rate": 1.1530271483396115e-05,
|
| 734 |
-
"loss": 2.
|
| 735 |
"step": 242
|
| 736 |
},
|
| 737 |
{
|
| 738 |
"epoch": 0.47,
|
| 739 |
"learning_rate": 1.1406218624648986e-05,
|
| 740 |
-
"loss": 1.
|
| 741 |
"step": 244
|
| 742 |
},
|
| 743 |
{
|
| 744 |
"epoch": 0.48,
|
| 745 |
"learning_rate": 1.1281944593243289e-05,
|
| 746 |
-
"loss": 1.
|
| 747 |
"step": 246
|
| 748 |
},
|
| 749 |
{
|
| 750 |
"epoch": 0.48,
|
| 751 |
"learning_rate": 1.1157468935227866e-05,
|
| 752 |
-
"loss": 1.
|
| 753 |
"step": 248
|
| 754 |
},
|
| 755 |
{
|
| 756 |
"epoch": 0.48,
|
| 757 |
"learning_rate": 1.1032811228363766e-05,
|
| 758 |
-
"loss": 1.
|
| 759 |
"step": 250
|
| 760 |
},
|
| 761 |
{
|
| 762 |
"epoch": 0.49,
|
| 763 |
"learning_rate": 1.0907991079045006e-05,
|
| 764 |
-
"loss": 1.
|
| 765 |
"step": 252
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.49,
|
| 769 |
"learning_rate": 1.0783028119214863e-05,
|
| 770 |
-
"loss": 1.
|
| 771 |
"step": 254
|
| 772 |
},
|
| 773 |
{
|
| 774 |
"epoch": 0.5,
|
| 775 |
"learning_rate": 1.0657942003278107e-05,
|
| 776 |
-
"loss": 1.
|
| 777 |
"step": 256
|
| 778 |
},
|
| 779 |
{
|
| 780 |
"epoch": 0.5,
|
| 781 |
"learning_rate": 1.0532752405009733e-05,
|
| 782 |
-
"loss": 1.
|
| 783 |
"step": 258
|
| 784 |
},
|
| 785 |
{
|
| 786 |
"epoch": 0.5,
|
| 787 |
"learning_rate": 1.0407479014460622e-05,
|
| 788 |
-
"loss": 1.
|
| 789 |
"step": 260
|
| 790 |
},
|
| 791 |
{
|
| 792 |
"epoch": 0.51,
|
| 793 |
"learning_rate": 1.0282141534860662e-05,
|
| 794 |
-
"loss": 1.
|
| 795 |
"step": 262
|
| 796 |
},
|
| 797 |
{
|
| 798 |
"epoch": 0.51,
|
| 799 |
"learning_rate": 1.015675967951979e-05,
|
| 800 |
-
"loss": 1.
|
| 801 |
"step": 264
|
| 802 |
},
|
| 803 |
{
|
| 804 |
"epoch": 0.51,
|
| 805 |
"learning_rate": 1.0031353168727432e-05,
|
| 806 |
-
"loss": 1.
|
| 807 |
"step": 266
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.52,
|
| 811 |
"learning_rate": 9.90594172665088e-06,
|
| 812 |
-
"loss": 1.
|
| 813 |
"step": 268
|
| 814 |
},
|
| 815 |
{
|
| 816 |
"epoch": 0.52,
|
| 817 |
"learning_rate": 9.780545078233015e-06,
|
| 818 |
-
"loss": 1.
|
| 819 |
"step": 270
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 0.53,
|
| 823 |
"learning_rate": 9.655182946089957e-06,
|
| 824 |
-
"loss": 1.
|
| 825 |
"step": 272
|
| 826 |
},
|
| 827 |
{
|
| 828 |
"epoch": 0.53,
|
| 829 |
"learning_rate": 9.529875047409027e-06,
|
| 830 |
-
"loss": 1.
|
| 831 |
"step": 274
|
| 832 |
},
|
| 833 |
{
|
| 834 |
"epoch": 0.53,
|
| 835 |
"learning_rate": 9.404641090847628e-06,
|
| 836 |
-
"loss": 1.
|
| 837 |
"step": 276
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"epoch": 0.54,
|
| 841 |
"learning_rate": 9.279500773433385e-06,
|
| 842 |
-
"loss":
|
| 843 |
"step": 278
|
| 844 |
},
|
| 845 |
{
|
| 846 |
"epoch": 0.54,
|
| 847 |
"learning_rate": 9.154473777466209e-06,
|
| 848 |
-
"loss": 1.
|
| 849 |
"step": 280
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.55,
|
| 853 |
"learning_rate": 9.029579767422592e-06,
|
| 854 |
-
"loss":
|
| 855 |
"step": 282
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"epoch": 0.55,
|
| 859 |
"learning_rate": 8.904838386862758e-06,
|
| 860 |
-
"loss":
|
| 861 |
"step": 284
|
| 862 |
},
|
| 863 |
{
|
| 864 |
"epoch": 0.55,
|
| 865 |
"learning_rate": 8.78026925534108e-06,
|
| 866 |
-
"loss": 2.
|
| 867 |
"step": 286
|
| 868 |
},
|
| 869 |
{
|
| 870 |
"epoch": 0.56,
|
| 871 |
"learning_rate": 8.655891965320292e-06,
|
| 872 |
-
"loss":
|
| 873 |
"step": 288
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 0.56,
|
| 877 |
"learning_rate": 8.531726079089934e-06,
|
| 878 |
-
"loss":
|
| 879 |
"step": 290
|
| 880 |
},
|
| 881 |
{
|
| 882 |
"epoch": 0.56,
|
| 883 |
"learning_rate": 8.407791125689577e-06,
|
| 884 |
-
"loss": 2.
|
| 885 |
"step": 292
|
| 886 |
},
|
| 887 |
{
|
| 888 |
"epoch": 0.57,
|
| 889 |
"learning_rate": 8.284106597837244e-06,
|
| 890 |
-
"loss": 1.
|
| 891 |
"step": 294
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.57,
|
| 895 |
"learning_rate": 8.160691948863572e-06,
|
| 896 |
-
"loss": 1.
|
| 897 |
"step": 296
|
| 898 |
},
|
| 899 |
{
|
| 900 |
"epoch": 0.58,
|
| 901 |
"learning_rate": 8.037566589652141e-06,
|
| 902 |
-
"loss": 1.
|
| 903 |
"step": 298
|
| 904 |
},
|
| 905 |
{
|
| 906 |
"epoch": 0.58,
|
| 907 |
"learning_rate": 7.914749885586517e-06,
|
| 908 |
-
"loss": 1.
|
| 909 |
"step": 300
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"epoch": 0.58,
|
| 913 |
"learning_rate": 7.792261153504403e-06,
|
| 914 |
-
"loss": 1.
|
| 915 |
"step": 302
|
| 916 |
},
|
| 917 |
{
|
| 918 |
"epoch": 0.59,
|
| 919 |
"learning_rate": 7.670119658659469e-06,
|
| 920 |
-
"loss": 1.
|
| 921 |
"step": 304
|
| 922 |
},
|
| 923 |
{
|
| 924 |
"epoch": 0.59,
|
| 925 |
"learning_rate": 7.548344611691272e-06,
|
| 926 |
-
"loss": 1.
|
| 927 |
"step": 306
|
| 928 |
},
|
| 929 |
{
|
| 930 |
"epoch": 0.6,
|
| 931 |
"learning_rate": 7.426955165603773e-06,
|
| 932 |
-
"loss": 1.
|
| 933 |
"step": 308
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.6,
|
| 937 |
"learning_rate": 7.30597041275291e-06,
|
| 938 |
-
"loss": 1.
|
| 939 |
"step": 310
|
| 940 |
},
|
| 941 |
{
|
| 942 |
"epoch": 0.6,
|
| 943 |
"learning_rate": 7.185409381843727e-06,
|
| 944 |
-
"loss": 1.
|
| 945 |
"step": 312
|
| 946 |
},
|
| 947 |
{
|
| 948 |
"epoch": 0.61,
|
| 949 |
"learning_rate": 7.065291034937496e-06,
|
| 950 |
-
"loss": 1.
|
| 951 |
"step": 314
|
| 952 |
},
|
| 953 |
{
|
| 954 |
"epoch": 0.61,
|
| 955 |
"learning_rate": 6.945634264469338e-06,
|
| 956 |
-
"loss": 1.
|
| 957 |
"step": 316
|
| 958 |
},
|
| 959 |
{
|
| 960 |
"epoch": 0.61,
|
| 961 |
"learning_rate": 6.826457890276773e-06,
|
| 962 |
-
"loss": 1.
|
| 963 |
"step": 318
|
| 964 |
},
|
| 965 |
{
|
| 966 |
"epoch": 0.62,
|
| 967 |
"learning_rate": 6.707780656639722e-06,
|
| 968 |
-
"loss": 1.
|
| 969 |
"step": 320
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.62,
|
| 973 |
"learning_rate": 6.58962122933234e-06,
|
| 974 |
-
"loss": 1.
|
| 975 |
"step": 322
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.63,
|
| 979 |
"learning_rate": 6.471998192687266e-06,
|
| 980 |
-
"loss": 1.
|
| 981 |
"step": 324
|
| 982 |
},
|
| 983 |
{
|
| 984 |
"epoch": 0.63,
|
| 985 |
"learning_rate": 6.354930046672612e-06,
|
| 986 |
-
"loss": 1.
|
| 987 |
"step": 326
|
| 988 |
},
|
| 989 |
{
|
| 990 |
"epoch": 0.63,
|
| 991 |
"learning_rate": 6.238435203982278e-06,
|
| 992 |
-
"loss": 1.
|
| 993 |
"step": 328
|
| 994 |
},
|
| 995 |
{
|
| 996 |
"epoch": 0.64,
|
| 997 |
"learning_rate": 6.122531987139955e-06,
|
| 998 |
-
"loss": 2.
|
| 999 |
"step": 330
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
"epoch": 0.64,
|
| 1003 |
"learning_rate": 6.007238625617333e-06,
|
| 1004 |
-
"loss":
|
| 1005 |
"step": 332
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
"epoch": 0.65,
|
| 1009 |
"learning_rate": 5.892573252966926e-06,
|
| 1010 |
-
"loss": 1.
|
| 1011 |
"step": 334
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
"epoch": 0.65,
|
| 1015 |
"learning_rate": 5.77855390397001e-06,
|
| 1016 |
-
"loss":
|
| 1017 |
"step": 336
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.65,
|
| 1021 |
"learning_rate": 5.665198511800063e-06,
|
| 1022 |
-
"loss": 1.
|
| 1023 |
"step": 338
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
"epoch": 0.66,
|
| 1027 |
"learning_rate": 5.552524905202208e-06,
|
| 1028 |
-
"loss": 1.
|
| 1029 |
"step": 340
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.66,
|
| 1033 |
"learning_rate": 5.440550805689075e-06,
|
| 1034 |
-
"loss":
|
| 1035 |
"step": 342
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
"epoch": 0.67,
|
| 1039 |
"learning_rate": 5.329293824753539e-06,
|
| 1040 |
-
"loss": 2.
|
| 1041 |
"step": 344
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
"epoch": 0.67,
|
| 1045 |
"learning_rate": 5.218771461098733e-06,
|
| 1046 |
-
"loss": 1.
|
| 1047 |
"step": 346
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
"epoch": 0.67,
|
| 1051 |
"learning_rate": 5.109001097885837e-06,
|
| 1052 |
-
"loss": 1.
|
| 1053 |
"step": 348
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
"epoch": 0.68,
|
| 1057 |
"learning_rate": 5.000000000000003e-06,
|
| 1058 |
-
"loss": 1.
|
| 1059 |
"step": 350
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.68,
|
| 1063 |
"learning_rate": 4.891785311334924e-06,
|
| 1064 |
-
"loss": 1.
|
| 1065 |
"step": 352
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
"epoch": 0.68,
|
| 1069 |
"learning_rate": 4.784374052096396e-06,
|
| 1070 |
-
"loss": 1.
|
| 1071 |
"step": 354
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
"epoch": 0.69,
|
| 1075 |
"learning_rate": 4.677783116125361e-06,
|
| 1076 |
-
"loss": 1.
|
| 1077 |
"step": 356
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
"epoch": 0.69,
|
| 1081 |
"learning_rate": 4.572029268240787e-06,
|
| 1082 |
-
"loss": 1.
|
| 1083 |
"step": 358
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
"epoch": 0.7,
|
| 1087 |
"learning_rate": 4.467129141602906e-06,
|
| 1088 |
-
"loss": 1.
|
| 1089 |
"step": 360
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
"epoch": 0.7,
|
| 1093 |
"learning_rate": 4.3630992350970866e-06,
|
| 1094 |
-
"loss": 1.
|
| 1095 |
"step": 362
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
"epoch": 0.7,
|
| 1099 |
"learning_rate": 4.2599559107388645e-06,
|
| 1100 |
-
"loss": 1.
|
| 1101 |
"step": 364
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.71,
|
| 1105 |
"learning_rate": 4.157715391100509e-06,
|
| 1106 |
-
"loss": 1.
|
| 1107 |
"step": 366
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
"epoch": 0.71,
|
| 1111 |
"learning_rate": 4.056393756759498e-06,
|
| 1112 |
-
"loss": 1.
|
| 1113 |
"step": 368
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
"epoch": 0.72,
|
| 1117 |
"learning_rate": 3.956006943769331e-06,
|
| 1118 |
-
"loss": 1.
|
| 1119 |
"step": 370
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 0.72,
|
| 1123 |
"learning_rate": 3.856570741153087e-06,
|
| 1124 |
-
"loss": 1.
|
| 1125 |
"step": 372
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
"epoch": 0.72,
|
| 1129 |
"learning_rate": 3.7581007884200826e-06,
|
| 1130 |
-
"loss": 1.
|
| 1131 |
"step": 374
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
"epoch": 0.73,
|
| 1135 |
"learning_rate": 3.6606125731060815e-06,
|
| 1136 |
-
"loss": 1.
|
| 1137 |
"step": 376
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
"epoch": 0.73,
|
| 1141 |
"learning_rate": 3.564121428337369e-06,
|
| 1142 |
-
"loss": 1.
|
| 1143 |
"step": 378
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.73,
|
| 1147 |
"learning_rate": 3.468642530419144e-06,
|
| 1148 |
-
"loss":
|
| 1149 |
"step": 380
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
"epoch": 0.74,
|
| 1153 |
"learning_rate": 3.3741908964485415e-06,
|
| 1154 |
-
"loss": 2.
|
| 1155 |
"step": 382
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
"epoch": 0.74,
|
| 1159 |
"learning_rate": 3.2807813819527377e-06,
|
| 1160 |
-
"loss":
|
| 1161 |
"step": 384
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
"epoch": 0.75,
|
| 1165 |
"learning_rate": 3.1884286785524356e-06,
|
| 1166 |
-
"loss": 2.
|
| 1167 |
"step": 386
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
"epoch": 0.75,
|
| 1171 |
"learning_rate": 3.09714731165114e-06,
|
| 1172 |
-
"loss": 2.
|
| 1173 |
"step": 388
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
"epoch": 0.75,
|
| 1177 |
"learning_rate": 3.006951638150567e-06,
|
| 1178 |
-
"loss":
|
| 1179 |
"step": 390
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
"epoch": 0.76,
|
| 1183 |
"learning_rate": 2.9178558441925843e-06,
|
| 1184 |
-
"loss": 2.
|
| 1185 |
"step": 392
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.76,
|
| 1189 |
"learning_rate": 2.829873942927971e-06,
|
| 1190 |
-
"loss": 1.
|
| 1191 |
"step": 394
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
"epoch": 0.77,
|
| 1195 |
"learning_rate": 2.743019772312412e-06,
|
| 1196 |
-
"loss": 1.
|
| 1197 |
"step": 396
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
"epoch": 0.77,
|
| 1201 |
"learning_rate": 2.6573069929300243e-06,
|
| 1202 |
-
"loss": 1.
|
| 1203 |
"step": 398
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
"epoch": 0.77,
|
| 1207 |
"learning_rate": 2.572749085844829e-06,
|
| 1208 |
-
"loss": 1.
|
| 1209 |
"step": 400
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.78,
|
| 1213 |
"learning_rate": 2.4893593504803824e-06,
|
| 1214 |
-
"loss": 1.
|
| 1215 |
"step": 402
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
"epoch": 0.78,
|
| 1219 |
"learning_rate": 2.4071509025280526e-06,
|
| 1220 |
-
"loss": 1.
|
| 1221 |
"step": 404
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
"epoch": 0.79,
|
| 1225 |
"learning_rate": 2.3261366718841307e-06,
|
| 1226 |
-
"loss": 1.
|
| 1227 |
"step": 406
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.79,
|
| 1231 |
"learning_rate": 2.2463294006162185e-06,
|
| 1232 |
-
"loss": 1.
|
| 1233 |
"step": 408
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
"epoch": 0.79,
|
| 1237 |
"learning_rate": 2.167741640959117e-06,
|
| 1238 |
-
"loss": 1.
|
| 1239 |
"step": 410
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
"epoch": 0.8,
|
| 1243 |
"learning_rate": 2.090385753340596e-06,
|
| 1244 |
-
"loss": 1.
|
| 1245 |
"step": 412
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
"epoch": 0.8,
|
| 1249 |
"learning_rate": 2.0142739044373093e-06,
|
| 1250 |
-
"loss": 1.
|
| 1251 |
"step": 414
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
"epoch": 0.8,
|
| 1255 |
"learning_rate": 1.9394180652612193e-06,
|
| 1256 |
-
"loss": 1.
|
| 1257 |
"step": 416
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
"epoch": 0.81,
|
| 1261 |
"learning_rate": 1.8658300092767546e-06,
|
| 1262 |
-
"loss": 1.
|
| 1263 |
"step": 418
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
"epoch": 0.81,
|
| 1267 |
"learning_rate": 1.7935213105490722e-06,
|
| 1268 |
-
"loss": 1.
|
| 1269 |
"step": 420
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.82,
|
| 1273 |
"learning_rate": 1.7225033419236502e-06,
|
| 1274 |
-
"loss": 1.
|
| 1275 |
"step": 422
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
"epoch": 0.82,
|
| 1279 |
"learning_rate": 1.652787273237565e-06,
|
| 1280 |
-
"loss": 1.
|
| 1281 |
"step": 424
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
"epoch": 0.82,
|
| 1285 |
"learning_rate": 1.5843840695626666e-06,
|
| 1286 |
-
"loss": 1.
|
| 1287 |
"step": 426
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
"epoch": 0.83,
|
| 1291 |
"learning_rate": 1.5173044894809764e-06,
|
| 1292 |
-
"loss": 1.
|
| 1293 |
"step": 428
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
"epoch": 0.83,
|
| 1297 |
"learning_rate": 1.4515590833925508e-06,
|
| 1298 |
-
"loss": 1.
|
| 1299 |
"step": 430
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.84,
|
| 1303 |
"learning_rate": 1.387158191856105e-06,
|
| 1304 |
-
"loss": 1.
|
| 1305 |
"step": 432
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
"epoch": 0.84,
|
| 1309 |
"learning_rate": 1.324111943962626e-06,
|
| 1310 |
-
"loss": 2.
|
| 1311 |
"step": 434
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.84,
|
| 1315 |
"learning_rate": 1.2624302557422475e-06,
|
| 1316 |
-
"loss": 1.
|
| 1317 |
"step": 436
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
"epoch": 0.85,
|
| 1321 |
"learning_rate": 1.2021228286046461e-06,
|
| 1322 |
-
"loss": 2.
|
| 1323 |
"step": 438
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
"epoch": 0.85,
|
| 1327 |
"learning_rate": 1.1431991478131754e-06,
|
| 1328 |
-
"loss": 2.
|
| 1329 |
"step": 440
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
"epoch": 0.85,
|
| 1333 |
"learning_rate": 1.085668480993015e-06,
|
| 1334 |
-
"loss": 1.
|
| 1335 |
"step": 442
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
"epoch": 0.86,
|
| 1339 |
"learning_rate": 1.029539876673541e-06,
|
| 1340 |
-
"loss": 1.
|
| 1341 |
"step": 444
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
"epoch": 0.86,
|
| 1345 |
"learning_rate": 9.748221628651445e-07,
|
| 1346 |
-
"loss": 1.
|
| 1347 |
"step": 446
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
"epoch": 0.87,
|
| 1351 |
"learning_rate": 9.215239456707636e-07,
|
| 1352 |
-
"loss": 1.
|
| 1353 |
"step": 448
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.87,
|
| 1357 |
"learning_rate": 8.696536079322904e-07,
|
| 1358 |
-
"loss": 1.
|
| 1359 |
"step": 450
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
"epoch": 0.87,
|
| 1363 |
"learning_rate": 8.192193079121003e-07,
|
| 1364 |
-
"loss": 1.
|
| 1365 |
"step": 452
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
"epoch": 0.88,
|
| 1369 |
"learning_rate": 7.70228978009907e-07,
|
| 1370 |
-
"loss": 1.
|
| 1371 |
"step": 454
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
"epoch": 0.88,
|
| 1375 |
"learning_rate": 7.226903235151439e-07,
|
| 1376 |
-
"loss": 1.
|
| 1377 |
"step": 456
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
"epoch": 0.89,
|
| 1381 |
"learning_rate": 6.766108213950584e-07,
|
| 1382 |
-
"loss": 1.
|
| 1383 |
"step": 458
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
"epoch": 0.89,
|
| 1387 |
"learning_rate": 6.319977191187232e-07,
|
| 1388 |
-
"loss": 1.
|
| 1389 |
"step": 460
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
"epoch": 0.89,
|
| 1393 |
"learning_rate": 5.888580335171368e-07,
|
| 1394 |
-
"loss": 1.
|
| 1395 |
"step": 462
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.9,
|
| 1399 |
"learning_rate": 5.471985496796051e-07,
|
| 1400 |
-
"loss": 1.
|
| 1401 |
"step": 464
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
"epoch": 0.9,
|
| 1405 |
"learning_rate": 5.070258198865741e-07,
|
| 1406 |
-
"loss": 1.
|
| 1407 |
"step": 466
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
"epoch": 0.91,
|
| 1411 |
"learning_rate": 4.683461625790697e-07,
|
| 1412 |
-
"loss": 1.
|
| 1413 |
"step": 468
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
"epoch": 0.91,
|
| 1417 |
"learning_rate": 4.311656613649251e-07,
|
| 1418 |
-
"loss": 1.
|
| 1419 |
"step": 470
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
"epoch": 0.91,
|
| 1423 |
"learning_rate": 3.9549016406193686e-07,
|
| 1424 |
-
"loss": 1.
|
| 1425 |
"step": 472
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
"epoch": 0.92,
|
| 1429 |
"learning_rate": 3.61325281778111e-07,
|
| 1430 |
-
"loss": 1.
|
| 1431 |
"step": 474
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
"epoch": 0.92,
|
| 1435 |
"learning_rate": 3.286763880291399e-07,
|
| 1436 |
-
"loss": 1.
|
| 1437 |
"step": 476
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.92,
|
| 1441 |
"learning_rate": 2.9754861789324074e-07,
|
| 1442 |
-
"loss": 1.
|
| 1443 |
"step": 478
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
"epoch": 0.93,
|
| 1447 |
"learning_rate": 2.6794686720350573e-07,
|
| 1448 |
-
"loss": 1.
|
| 1449 |
"step": 480
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
"epoch": 0.93,
|
| 1453 |
"learning_rate": 2.398757917778727e-07,
|
| 1454 |
-
"loss": 1.
|
| 1455 |
"step": 482
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
"epoch": 0.94,
|
| 1459 |
"learning_rate": 2.1333980668685416e-07,
|
| 1460 |
-
"loss": 1.
|
| 1461 |
"step": 484
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
"epoch": 0.94,
|
| 1465 |
"learning_rate": 1.883430855591173e-07,
|
| 1466 |
-
"loss":
|
| 1467 |
"step": 486
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
"epoch": 0.94,
|
| 1471 |
"learning_rate": 1.6488955992506218e-07,
|
| 1472 |
-
"loss":
|
| 1473 |
"step": 488
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
"epoch": 0.95,
|
| 1477 |
"learning_rate": 1.4298291859845215e-07,
|
| 1478 |
-
"loss": 2.
|
| 1479 |
"step": 490
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.95,
|
| 1483 |
"learning_rate": 1.2262660709623232e-07,
|
| 1484 |
-
"loss": 1.
|
| 1485 |
"step": 492
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
"epoch": 0.96,
|
| 1489 |
"learning_rate": 1.0382382709661853e-07,
|
| 1490 |
-
"loss": 2.
|
| 1491 |
"step": 494
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
"epoch": 0.96,
|
| 1495 |
"learning_rate": 8.657753593552142e-08,
|
| 1496 |
-
"loss": 1.
|
| 1497 |
"step": 496
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
"epoch": 0.96,
|
| 1501 |
"learning_rate": 7.08904461414217e-08,
|
| 1502 |
-
"loss": 1.
|
| 1503 |
"step": 498
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
"epoch": 0.97,
|
| 1507 |
"learning_rate": 5.676502500873038e-08,
|
| 1508 |
-
"loss": 1.
|
| 1509 |
"step": 500
|
| 1510 |
}
|
| 1511 |
],
|
|
@@ -1513,7 +1513,7 @@
|
|
| 1513 |
"max_steps": 517,
|
| 1514 |
"num_train_epochs": 1,
|
| 1515 |
"save_steps": 50,
|
| 1516 |
-
"total_flos":
|
| 1517 |
"trial_name": null,
|
| 1518 |
"trial_params": null
|
| 1519 |
}
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 2.5e-06,
|
| 14 |
+
"loss": 2.0056,
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.01,
|
| 19 |
"learning_rate": 5e-06,
|
| 20 |
+
"loss": 1.7547,
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
"learning_rate": 7.500000000000001e-06,
|
| 26 |
+
"loss": 1.8255,
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.02,
|
| 31 |
"learning_rate": 1e-05,
|
| 32 |
+
"loss": 1.8088,
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.02,
|
| 37 |
"learning_rate": 1.25e-05,
|
| 38 |
+
"loss": 1.942,
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.02,
|
| 43 |
"learning_rate": 1.5000000000000002e-05,
|
| 44 |
+
"loss": 1.9364,
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.03,
|
| 49 |
"learning_rate": 1.7500000000000002e-05,
|
| 50 |
+
"loss": 1.9922,
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03,
|
| 55 |
"learning_rate": 2e-05,
|
| 56 |
+
"loss": 2.0139,
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.03,
|
| 61 |
"learning_rate": 1.9999213590779245e-05,
|
| 62 |
+
"loss": 2.0467,
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.04,
|
| 67 |
"learning_rate": 1.9996854486804874e-05,
|
| 68 |
+
"loss": 2.1225,
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.04,
|
| 73 |
"learning_rate": 1.9992923059121107e-05,
|
| 74 |
+
"loss": 2.3509,
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.05,
|
| 79 |
"learning_rate": 1.998741992607014e-05,
|
| 80 |
+
"loss": 2.0483,
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.05,
|
| 85 |
"learning_rate": 1.9980345953194892e-05,
|
| 86 |
+
"loss": 2.1992,
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.05,
|
| 91 |
"learning_rate": 1.9971702253102856e-05,
|
| 92 |
+
"loss": 2.3724,
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.06,
|
| 97 |
"learning_rate": 1.9961490185291128e-05,
|
| 98 |
+
"loss": 2.5406,
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.06,
|
| 103 |
"learning_rate": 1.9949711355932565e-05,
|
| 104 |
+
"loss": 2.6925,
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.07,
|
| 109 |
"learning_rate": 1.993636761762317e-05,
|
| 110 |
+
"loss": 2.5443,
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.07,
|
| 115 |
"learning_rate": 1.9921461069090715e-05,
|
| 116 |
+
"loss": 2.7276,
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.07,
|
| 121 |
"learning_rate": 1.990499405486464e-05,
|
| 122 |
+
"loss": 2.5429,
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.08,
|
| 127 |
"learning_rate": 1.9886969164907306e-05,
|
| 128 |
+
"loss": 2.7652,
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.08,
|
| 133 |
"learning_rate": 1.9867389234206655e-05,
|
| 134 |
+
"loss": 2.8533,
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.09,
|
| 139 |
"learning_rate": 1.984625734233029e-05,
|
| 140 |
+
"loss": 2.9539,
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.09,
|
| 145 |
"learning_rate": 1.982357681294114e-05,
|
| 146 |
+
"loss": 3.0776,
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.09,
|
| 151 |
"learning_rate": 1.9799351213274688e-05,
|
| 152 |
+
"loss": 3.0149,
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.1,
|
| 157 |
"learning_rate": 1.9773584353577928e-05,
|
| 158 |
+
"loss": 2.8577,
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.1,
|
| 163 |
"learning_rate": 1.974628028651007e-05,
|
| 164 |
+
"loss": 1.5051,
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 0.1,
|
| 169 |
"learning_rate": 1.9717443306505143e-05,
|
| 170 |
+
"loss": 1.691,
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 0.11,
|
| 175 |
"learning_rate": 1.9687077949096528e-05,
|
| 176 |
+
"loss": 1.8247,
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.11,
|
| 181 |
"learning_rate": 1.9655188990203648e-05,
|
| 182 |
+
"loss": 1.7148,
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 0.12,
|
| 187 |
"learning_rate": 1.962178144538076e-05,
|
| 188 |
+
"loss": 1.8262,
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.12,
|
| 193 |
"learning_rate": 1.9586860569028126e-05,
|
| 194 |
+
"loss": 1.7259,
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.12,
|
| 199 |
"learning_rate": 1.9550431853565577e-05,
|
| 200 |
+
"loss": 1.7772,
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.13,
|
| 205 |
"learning_rate": 1.951250102856866e-05,
|
| 206 |
+
"loss": 1.9479,
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.13,
|
| 211 |
"learning_rate": 1.9473074059867476e-05,
|
| 212 |
+
"loss": 1.7778,
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 0.14,
|
| 217 |
"learning_rate": 1.943215714860838e-05,
|
| 218 |
+
"loss": 1.8576,
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.14,
|
| 223 |
"learning_rate": 1.9389756730278625e-05,
|
| 224 |
+
"loss": 2.0523,
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 0.14,
|
| 229 |
"learning_rate": 1.9345879473694207e-05,
|
| 230 |
+
"loss": 1.886,
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 0.15,
|
| 235 |
"learning_rate": 1.930053227995095e-05,
|
| 236 |
+
"loss": 2.1032,
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.15,
|
| 241 |
"learning_rate": 1.925372228133912e-05,
|
| 242 |
+
"loss": 2.0642,
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 0.15,
|
| 247 |
"learning_rate": 1.9205456840221614e-05,
|
| 248 |
+
"loss": 2.222,
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.16,
|
| 253 |
"learning_rate": 1.9155743547876026e-05,
|
| 254 |
+
"loss": 2.06,
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 0.16,
|
| 259 |
"learning_rate": 1.9104590223300656e-05,
|
| 260 |
+
"loss": 2.1776,
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.17,
|
| 265 |
"learning_rate": 1.9052004911984722e-05,
|
| 266 |
+
"loss": 2.1747,
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 0.17,
|
| 271 |
"learning_rate": 1.8997995884642968e-05,
|
| 272 |
+
"loss": 2.0654,
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 0.17,
|
| 277 |
"learning_rate": 1.8942571635914814e-05,
|
| 278 |
+
"loss": 2.3257,
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.18,
|
| 283 |
"learning_rate": 1.888574088302831e-05,
|
| 284 |
+
"loss": 2.271,
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 0.18,
|
| 289 |
"learning_rate": 1.8827512564429076e-05,
|
| 290 |
+
"loss": 2.1348,
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 0.19,
|
| 295 |
"learning_rate": 1.8767895838374438e-05,
|
| 296 |
+
"loss": 2.1867,
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.19,
|
| 301 |
"learning_rate": 1.870690008149302e-05,
|
| 302 |
+
"loss": 2.096,
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.19,
|
| 307 |
"learning_rate": 1.8644534887309947e-05,
|
| 308 |
+
"loss": 1.6308,
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 0.2,
|
| 313 |
"learning_rate": 1.8580810064737967e-05,
|
| 314 |
+
"loss": 1.5925,
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 0.2,
|
| 319 |
"learning_rate": 1.8515735636534694e-05,
|
| 320 |
+
"loss": 1.6675,
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 0.2,
|
| 325 |
"learning_rate": 1.8449321837726207e-05,
|
| 326 |
+
"loss": 1.6544,
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 0.21,
|
| 331 |
"learning_rate": 1.8381579113997255e-05,
|
| 332 |
+
"loss": 1.6507,
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 0.21,
|
| 337 |
"learning_rate": 1.8312518120048354e-05,
|
| 338 |
+
"loss": 1.679,
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 0.22,
|
| 343 |
"learning_rate": 1.8242149717919993e-05,
|
| 344 |
+
"loss": 1.7766,
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22,
|
| 349 |
"learning_rate": 1.8170484975284233e-05,
|
| 350 |
+
"loss": 1.8045,
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 0.22,
|
| 355 |
"learning_rate": 1.8097535163703947e-05,
|
| 356 |
+
"loss": 1.6356,
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 0.23,
|
| 361 |
"learning_rate": 1.8023311756860036e-05,
|
| 362 |
+
"loss": 1.8896,
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 0.23,
|
| 367 |
"learning_rate": 1.794782642874681e-05,
|
| 368 |
+
"loss": 1.7886,
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 0.24,
|
| 373 |
"learning_rate": 1.7871091051835873e-05,
|
| 374 |
+
"loss": 1.7651,
|
| 375 |
"step": 122
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 0.24,
|
| 379 |
"learning_rate": 1.7793117695208828e-05,
|
| 380 |
+
"loss": 1.8476,
|
| 381 |
"step": 124
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 0.24,
|
| 385 |
"learning_rate": 1.771391862265899e-05,
|
| 386 |
+
"loss": 1.8061,
|
| 387 |
"step": 126
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.25,
|
| 391 |
"learning_rate": 1.763350629076255e-05,
|
| 392 |
+
"loss": 1.9459,
|
| 393 |
"step": 128
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 0.25,
|
| 397 |
"learning_rate": 1.7551893346919363e-05,
|
| 398 |
+
"loss": 1.8503,
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 0.26,
|
| 403 |
"learning_rate": 1.7469092627363738e-05,
|
| 404 |
+
"loss": 1.9725,
|
| 405 |
"step": 132
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 0.26,
|
| 409 |
"learning_rate": 1.7385117155145547e-05,
|
| 410 |
+
"loss": 2.0693,
|
| 411 |
"step": 134
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 0.26,
|
| 415 |
"learning_rate": 1.729998013808192e-05,
|
| 416 |
+
"loss": 2.131,
|
| 417 |
"step": 136
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 0.27,
|
| 421 |
"learning_rate": 1.7213694966679913e-05,
|
| 422 |
+
"loss": 2.0871,
|
| 423 |
"step": 138
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 0.27,
|
| 427 |
"learning_rate": 1.7126275212030403e-05,
|
| 428 |
+
"loss": 2.1431,
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.27,
|
| 433 |
"learning_rate": 1.7037734623673616e-05,
|
| 434 |
+
"loss": 2.0547,
|
| 435 |
"step": 142
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 0.28,
|
| 439 |
"learning_rate": 1.6948087127436573e-05,
|
| 440 |
+
"loss": 1.9428,
|
| 441 |
"step": 144
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 0.28,
|
| 445 |
"learning_rate": 1.68573468232428e-05,
|
| 446 |
+
"loss": 1.9212,
|
| 447 |
"step": 146
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 0.29,
|
| 451 |
"learning_rate": 1.676552798289469e-05,
|
| 452 |
+
"loss": 1.6917,
|
| 453 |
"step": 148
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 0.29,
|
| 457 |
"learning_rate": 1.667264504782877e-05,
|
| 458 |
+
"loss": 1.4354,
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 0.29,
|
| 463 |
"learning_rate": 1.6578712626844365e-05,
|
| 464 |
+
"loss": 1.4524,
|
| 465 |
"step": 152
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 0.3,
|
| 469 |
"learning_rate": 1.648374549380587e-05,
|
| 470 |
+
"loss": 1.574,
|
| 471 |
"step": 154
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.3,
|
| 475 |
"learning_rate": 1.63877585853191e-05,
|
| 476 |
+
"loss": 1.75,
|
| 477 |
"step": 156
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 0.31,
|
| 481 |
"learning_rate": 1.629076699838204e-05,
|
| 482 |
+
"loss": 1.6588,
|
| 483 |
"step": 158
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 0.31,
|
| 487 |
"learning_rate": 1.6192785988010354e-05,
|
| 488 |
+
"loss": 1.6005,
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.31,
|
| 493 |
"learning_rate": 1.6093830964838036e-05,
|
| 494 |
+
"loss": 1.6734,
|
| 495 |
"step": 162
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 0.32,
|
| 499 |
"learning_rate": 1.5993917492693623e-05,
|
| 500 |
+
"loss": 1.5628,
|
| 501 |
"step": 164
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 0.32,
|
| 505 |
"learning_rate": 1.5893061286152275e-05,
|
| 506 |
+
"loss": 1.6464,
|
| 507 |
"step": 166
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 0.32,
|
| 511 |
"learning_rate": 1.5791278208064142e-05,
|
| 512 |
+
"loss": 1.6423,
|
| 513 |
"step": 168
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.33,
|
| 517 |
"learning_rate": 1.5688584267059453e-05,
|
| 518 |
+
"loss": 1.6845,
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.33,
|
| 523 |
"learning_rate": 1.5584995615030634e-05,
|
| 524 |
+
"loss": 1.7696,
|
| 525 |
"step": 172
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 0.34,
|
| 529 |
"learning_rate": 1.5480528544591907e-05,
|
| 530 |
+
"loss": 1.7355,
|
| 531 |
"step": 174
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 0.34,
|
| 535 |
"learning_rate": 1.537519948651676e-05,
|
| 536 |
+
"loss": 1.8252,
|
| 537 |
"step": 176
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 0.34,
|
| 541 |
"learning_rate": 1.52690250071537e-05,
|
| 542 |
+
"loss": 1.9255,
|
| 543 |
"step": 178
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 0.35,
|
| 547 |
"learning_rate": 1.5162021805820633e-05,
|
| 548 |
+
"loss": 1.6875,
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 0.35,
|
| 553 |
"learning_rate": 1.50542067121784e-05,
|
| 554 |
+
"loss": 1.8883,
|
| 555 |
"step": 182
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.36,
|
| 559 |
"learning_rate": 1.4945596683583754e-05,
|
| 560 |
+
"loss": 1.9448,
|
| 561 |
"step": 184
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 0.36,
|
| 565 |
"learning_rate": 1.4836208802422287e-05,
|
| 566 |
+
"loss": 1.8727,
|
| 567 |
"step": 186
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 0.36,
|
| 571 |
"learning_rate": 1.4726060273421675e-05,
|
| 572 |
+
"loss": 1.9311,
|
| 573 |
"step": 188
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 0.37,
|
| 577 |
"learning_rate": 1.4615168420945689e-05,
|
| 578 |
+
"loss": 1.8577,
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 0.37,
|
| 583 |
"learning_rate": 1.450355068626939e-05,
|
| 584 |
+
"loss": 2.0573,
|
| 585 |
"step": 192
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 0.38,
|
| 589 |
"learning_rate": 1.4391224624835928e-05,
|
| 590 |
+
"loss": 1.88,
|
| 591 |
"step": 194
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 0.38,
|
| 595 |
"learning_rate": 1.4278207903495388e-05,
|
| 596 |
+
"loss": 2.0729,
|
| 597 |
"step": 196
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.38,
|
| 601 |
"learning_rate": 1.4164518297726129e-05,
|
| 602 |
+
"loss": 1.6617,
|
| 603 |
"step": 198
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 0.39,
|
| 607 |
"learning_rate": 1.4050173688839e-05,
|
| 608 |
+
"loss": 1.3384,
|
| 609 |
"step": 200
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 0.39,
|
| 613 |
"learning_rate": 1.3935192061164957e-05,
|
| 614 |
+
"loss": 1.611,
|
| 615 |
"step": 202
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 0.39,
|
| 619 |
"learning_rate": 1.3819591499226446e-05,
|
| 620 |
+
"loss": 1.4943,
|
| 621 |
"step": 204
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"epoch": 0.4,
|
| 625 |
"learning_rate": 1.3703390184893034e-05,
|
| 626 |
+
"loss": 1.6069,
|
| 627 |
"step": 206
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"epoch": 0.4,
|
| 631 |
"learning_rate": 1.3586606394521729e-05,
|
| 632 |
+
"loss": 1.5703,
|
| 633 |
"step": 208
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 0.41,
|
| 637 |
"learning_rate": 1.3469258496082452e-05,
|
| 638 |
+
"loss": 1.5996,
|
| 639 |
"step": 210
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.41,
|
| 643 |
"learning_rate": 1.3351364946269071e-05,
|
| 644 |
+
"loss": 1.6916,
|
| 645 |
"step": 212
|
| 646 |
},
|
| 647 |
{
|
| 648 |
"epoch": 0.41,
|
| 649 |
"learning_rate": 1.323294428759652e-05,
|
| 650 |
+
"loss": 1.5462,
|
| 651 |
"step": 214
|
| 652 |
},
|
| 653 |
{
|
| 654 |
"epoch": 0.42,
|
| 655 |
"learning_rate": 1.311401514548438e-05,
|
| 656 |
+
"loss": 1.6493,
|
| 657 |
"step": 216
|
| 658 |
},
|
| 659 |
{
|
| 660 |
"epoch": 0.42,
|
| 661 |
"learning_rate": 1.2994596225327442e-05,
|
| 662 |
+
"loss": 1.7534,
|
| 663 |
"step": 218
|
| 664 |
},
|
| 665 |
{
|
| 666 |
"epoch": 0.43,
|
| 667 |
"learning_rate": 1.2874706309553697e-05,
|
| 668 |
+
"loss": 1.7442,
|
| 669 |
"step": 220
|
| 670 |
},
|
| 671 |
{
|
| 672 |
"epoch": 0.43,
|
| 673 |
"learning_rate": 1.2754364254670192e-05,
|
| 674 |
+
"loss": 1.7905,
|
| 675 |
"step": 222
|
| 676 |
},
|
| 677 |
{
|
| 678 |
"epoch": 0.43,
|
| 679 |
"learning_rate": 1.263358898829725e-05,
|
| 680 |
+
"loss": 1.7625,
|
| 681 |
"step": 224
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44,
|
| 685 |
"learning_rate": 1.251239950619149e-05,
|
| 686 |
+
"loss": 1.7027,
|
| 687 |
"step": 226
|
| 688 |
},
|
| 689 |
{
|
| 690 |
"epoch": 0.44,
|
| 691 |
"learning_rate": 1.2390814869258155e-05,
|
| 692 |
+
"loss": 1.7345,
|
| 693 |
"step": 228
|
| 694 |
},
|
| 695 |
{
|
| 696 |
"epoch": 0.44,
|
| 697 |
"learning_rate": 1.2268854200553158e-05,
|
| 698 |
+
"loss": 1.9245,
|
| 699 |
"step": 230
|
| 700 |
},
|
| 701 |
{
|
| 702 |
"epoch": 0.45,
|
| 703 |
"learning_rate": 1.2146536682275388e-05,
|
| 704 |
+
"loss": 1.9498,
|
| 705 |
"step": 232
|
| 706 |
},
|
| 707 |
{
|
| 708 |
"epoch": 0.45,
|
| 709 |
"learning_rate": 1.2023881552749693e-05,
|
| 710 |
+
"loss": 1.8216,
|
| 711 |
"step": 234
|
| 712 |
},
|
| 713 |
{
|
| 714 |
"epoch": 0.46,
|
| 715 |
"learning_rate": 1.1900908103401037e-05,
|
| 716 |
+
"loss": 1.9172,
|
| 717 |
"step": 236
|
| 718 |
},
|
| 719 |
{
|
| 720 |
"epoch": 0.46,
|
| 721 |
"learning_rate": 1.1777635675720313e-05,
|
| 722 |
+
"loss": 1.9721,
|
| 723 |
"step": 238
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.46,
|
| 727 |
"learning_rate": 1.1654083658222287e-05,
|
| 728 |
+
"loss": 1.8381,
|
| 729 |
"step": 240
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.47,
|
| 733 |
"learning_rate": 1.1530271483396115e-05,
|
| 734 |
+
"loss": 2.0593,
|
| 735 |
"step": 242
|
| 736 |
},
|
| 737 |
{
|
| 738 |
"epoch": 0.47,
|
| 739 |
"learning_rate": 1.1406218624648986e-05,
|
| 740 |
+
"loss": 1.9687,
|
| 741 |
"step": 244
|
| 742 |
},
|
| 743 |
{
|
| 744 |
"epoch": 0.48,
|
| 745 |
"learning_rate": 1.1281944593243289e-05,
|
| 746 |
+
"loss": 1.9155,
|
| 747 |
"step": 246
|
| 748 |
},
|
| 749 |
{
|
| 750 |
"epoch": 0.48,
|
| 751 |
"learning_rate": 1.1157468935227866e-05,
|
| 752 |
+
"loss": 1.6736,
|
| 753 |
"step": 248
|
| 754 |
},
|
| 755 |
{
|
| 756 |
"epoch": 0.48,
|
| 757 |
"learning_rate": 1.1032811228363766e-05,
|
| 758 |
+
"loss": 1.267,
|
| 759 |
"step": 250
|
| 760 |
},
|
| 761 |
{
|
| 762 |
"epoch": 0.49,
|
| 763 |
"learning_rate": 1.0907991079045006e-05,
|
| 764 |
+
"loss": 1.4318,
|
| 765 |
"step": 252
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.49,
|
| 769 |
"learning_rate": 1.0783028119214863e-05,
|
| 770 |
+
"loss": 1.5827,
|
| 771 |
"step": 254
|
| 772 |
},
|
| 773 |
{
|
| 774 |
"epoch": 0.5,
|
| 775 |
"learning_rate": 1.0657942003278107e-05,
|
| 776 |
+
"loss": 1.6949,
|
| 777 |
"step": 256
|
| 778 |
},
|
| 779 |
{
|
| 780 |
"epoch": 0.5,
|
| 781 |
"learning_rate": 1.0532752405009733e-05,
|
| 782 |
+
"loss": 1.549,
|
| 783 |
"step": 258
|
| 784 |
},
|
| 785 |
{
|
| 786 |
"epoch": 0.5,
|
| 787 |
"learning_rate": 1.0407479014460622e-05,
|
| 788 |
+
"loss": 1.6073,
|
| 789 |
"step": 260
|
| 790 |
},
|
| 791 |
{
|
| 792 |
"epoch": 0.51,
|
| 793 |
"learning_rate": 1.0282141534860662e-05,
|
| 794 |
+
"loss": 1.6264,
|
| 795 |
"step": 262
|
| 796 |
},
|
| 797 |
{
|
| 798 |
"epoch": 0.51,
|
| 799 |
"learning_rate": 1.015675967951979e-05,
|
| 800 |
+
"loss": 1.6455,
|
| 801 |
"step": 264
|
| 802 |
},
|
| 803 |
{
|
| 804 |
"epoch": 0.51,
|
| 805 |
"learning_rate": 1.0031353168727432e-05,
|
| 806 |
+
"loss": 1.6684,
|
| 807 |
"step": 266
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.52,
|
| 811 |
"learning_rate": 9.90594172665088e-06,
|
| 812 |
+
"loss": 1.6389,
|
| 813 |
"step": 268
|
| 814 |
},
|
| 815 |
{
|
| 816 |
"epoch": 0.52,
|
| 817 |
"learning_rate": 9.780545078233015e-06,
|
| 818 |
+
"loss": 1.7332,
|
| 819 |
"step": 270
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 0.53,
|
| 823 |
"learning_rate": 9.655182946089957e-06,
|
| 824 |
+
"loss": 1.6928,
|
| 825 |
"step": 272
|
| 826 |
},
|
| 827 |
{
|
| 828 |
"epoch": 0.53,
|
| 829 |
"learning_rate": 9.529875047409027e-06,
|
| 830 |
+
"loss": 1.6093,
|
| 831 |
"step": 274
|
| 832 |
},
|
| 833 |
{
|
| 834 |
"epoch": 0.53,
|
| 835 |
"learning_rate": 9.404641090847628e-06,
|
| 836 |
+
"loss": 1.9292,
|
| 837 |
"step": 276
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"epoch": 0.54,
|
| 841 |
"learning_rate": 9.279500773433385e-06,
|
| 842 |
+
"loss": 1.9869,
|
| 843 |
"step": 278
|
| 844 |
},
|
| 845 |
{
|
| 846 |
"epoch": 0.54,
|
| 847 |
"learning_rate": 9.154473777466209e-06,
|
| 848 |
+
"loss": 1.9451,
|
| 849 |
"step": 280
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.55,
|
| 853 |
"learning_rate": 9.029579767422592e-06,
|
| 854 |
+
"loss": 1.7706,
|
| 855 |
"step": 282
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"epoch": 0.55,
|
| 859 |
"learning_rate": 8.904838386862758e-06,
|
| 860 |
+
"loss": 2.0244,
|
| 861 |
"step": 284
|
| 862 |
},
|
| 863 |
{
|
| 864 |
"epoch": 0.55,
|
| 865 |
"learning_rate": 8.78026925534108e-06,
|
| 866 |
+
"loss": 2.0797,
|
| 867 |
"step": 286
|
| 868 |
},
|
| 869 |
{
|
| 870 |
"epoch": 0.56,
|
| 871 |
"learning_rate": 8.655891965320292e-06,
|
| 872 |
+
"loss": 1.9499,
|
| 873 |
"step": 288
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 0.56,
|
| 877 |
"learning_rate": 8.531726079089934e-06,
|
| 878 |
+
"loss": 2.0015,
|
| 879 |
"step": 290
|
| 880 |
},
|
| 881 |
{
|
| 882 |
"epoch": 0.56,
|
| 883 |
"learning_rate": 8.407791125689577e-06,
|
| 884 |
+
"loss": 2.0283,
|
| 885 |
"step": 292
|
| 886 |
},
|
| 887 |
{
|
| 888 |
"epoch": 0.57,
|
| 889 |
"learning_rate": 8.284106597837244e-06,
|
| 890 |
+
"loss": 1.8542,
|
| 891 |
"step": 294
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.57,
|
| 895 |
"learning_rate": 8.160691948863572e-06,
|
| 896 |
+
"loss": 1.8423,
|
| 897 |
"step": 296
|
| 898 |
},
|
| 899 |
{
|
| 900 |
"epoch": 0.58,
|
| 901 |
"learning_rate": 8.037566589652141e-06,
|
| 902 |
+
"loss": 1.5956,
|
| 903 |
"step": 298
|
| 904 |
},
|
| 905 |
{
|
| 906 |
"epoch": 0.58,
|
| 907 |
"learning_rate": 7.914749885586517e-06,
|
| 908 |
+
"loss": 1.3204,
|
| 909 |
"step": 300
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"epoch": 0.58,
|
| 913 |
"learning_rate": 7.792261153504403e-06,
|
| 914 |
+
"loss": 1.5616,
|
| 915 |
"step": 302
|
| 916 |
},
|
| 917 |
{
|
| 918 |
"epoch": 0.59,
|
| 919 |
"learning_rate": 7.670119658659469e-06,
|
| 920 |
+
"loss": 1.5286,
|
| 921 |
"step": 304
|
| 922 |
},
|
| 923 |
{
|
| 924 |
"epoch": 0.59,
|
| 925 |
"learning_rate": 7.548344611691272e-06,
|
| 926 |
+
"loss": 1.6407,
|
| 927 |
"step": 306
|
| 928 |
},
|
| 929 |
{
|
| 930 |
"epoch": 0.6,
|
| 931 |
"learning_rate": 7.426955165603773e-06,
|
| 932 |
+
"loss": 1.6575,
|
| 933 |
"step": 308
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.6,
|
| 937 |
"learning_rate": 7.30597041275291e-06,
|
| 938 |
+
"loss": 1.6348,
|
| 939 |
"step": 310
|
| 940 |
},
|
| 941 |
{
|
| 942 |
"epoch": 0.6,
|
| 943 |
"learning_rate": 7.185409381843727e-06,
|
| 944 |
+
"loss": 1.6124,
|
| 945 |
"step": 312
|
| 946 |
},
|
| 947 |
{
|
| 948 |
"epoch": 0.61,
|
| 949 |
"learning_rate": 7.065291034937496e-06,
|
| 950 |
+
"loss": 1.6442,
|
| 951 |
"step": 314
|
| 952 |
},
|
| 953 |
{
|
| 954 |
"epoch": 0.61,
|
| 955 |
"learning_rate": 6.945634264469338e-06,
|
| 956 |
+
"loss": 1.6176,
|
| 957 |
"step": 316
|
| 958 |
},
|
| 959 |
{
|
| 960 |
"epoch": 0.61,
|
| 961 |
"learning_rate": 6.826457890276773e-06,
|
| 962 |
+
"loss": 1.6775,
|
| 963 |
"step": 318
|
| 964 |
},
|
| 965 |
{
|
| 966 |
"epoch": 0.62,
|
| 967 |
"learning_rate": 6.707780656639722e-06,
|
| 968 |
+
"loss": 1.6962,
|
| 969 |
"step": 320
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.62,
|
| 973 |
"learning_rate": 6.58962122933234e-06,
|
| 974 |
+
"loss": 1.6958,
|
| 975 |
"step": 322
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.63,
|
| 979 |
"learning_rate": 6.471998192687266e-06,
|
| 980 |
+
"loss": 1.7599,
|
| 981 |
"step": 324
|
| 982 |
},
|
| 983 |
{
|
| 984 |
"epoch": 0.63,
|
| 985 |
"learning_rate": 6.354930046672612e-06,
|
| 986 |
+
"loss": 1.8092,
|
| 987 |
"step": 326
|
| 988 |
},
|
| 989 |
{
|
| 990 |
"epoch": 0.63,
|
| 991 |
"learning_rate": 6.238435203982278e-06,
|
| 992 |
+
"loss": 1.7497,
|
| 993 |
"step": 328
|
| 994 |
},
|
| 995 |
{
|
| 996 |
"epoch": 0.64,
|
| 997 |
"learning_rate": 6.122531987139955e-06,
|
| 998 |
+
"loss": 2.0158,
|
| 999 |
"step": 330
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
"epoch": 0.64,
|
| 1003 |
"learning_rate": 6.007238625617333e-06,
|
| 1004 |
+
"loss": 1.9724,
|
| 1005 |
"step": 332
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
"epoch": 0.65,
|
| 1009 |
"learning_rate": 5.892573252966926e-06,
|
| 1010 |
+
"loss": 1.9552,
|
| 1011 |
"step": 334
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
"epoch": 0.65,
|
| 1015 |
"learning_rate": 5.77855390397001e-06,
|
| 1016 |
+
"loss": 1.8851,
|
| 1017 |
"step": 336
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.65,
|
| 1021 |
"learning_rate": 5.665198511800063e-06,
|
| 1022 |
+
"loss": 1.8938,
|
| 1023 |
"step": 338
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
"epoch": 0.66,
|
| 1027 |
"learning_rate": 5.552524905202208e-06,
|
| 1028 |
+
"loss": 1.998,
|
| 1029 |
"step": 340
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.66,
|
| 1033 |
"learning_rate": 5.440550805689075e-06,
|
| 1034 |
+
"loss": 1.9771,
|
| 1035 |
"step": 342
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
"epoch": 0.67,
|
| 1039 |
"learning_rate": 5.329293824753539e-06,
|
| 1040 |
+
"loss": 2.0414,
|
| 1041 |
"step": 344
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
"epoch": 0.67,
|
| 1045 |
"learning_rate": 5.218771461098733e-06,
|
| 1046 |
+
"loss": 1.8501,
|
| 1047 |
"step": 346
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
"epoch": 0.67,
|
| 1051 |
"learning_rate": 5.109001097885837e-06,
|
| 1052 |
+
"loss": 1.8095,
|
| 1053 |
"step": 348
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
"epoch": 0.68,
|
| 1057 |
"learning_rate": 5.000000000000003e-06,
|
| 1058 |
+
"loss": 1.2658,
|
| 1059 |
"step": 350
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.68,
|
| 1063 |
"learning_rate": 4.891785311334924e-06,
|
| 1064 |
+
"loss": 1.455,
|
| 1065 |
"step": 352
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
"epoch": 0.68,
|
| 1069 |
"learning_rate": 4.784374052096396e-06,
|
| 1070 |
+
"loss": 1.621,
|
| 1071 |
"step": 354
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
"epoch": 0.69,
|
| 1075 |
"learning_rate": 4.677783116125361e-06,
|
| 1076 |
+
"loss": 1.5763,
|
| 1077 |
"step": 356
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
"epoch": 0.69,
|
| 1081 |
"learning_rate": 4.572029268240787e-06,
|
| 1082 |
+
"loss": 1.5071,
|
| 1083 |
"step": 358
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
"epoch": 0.7,
|
| 1087 |
"learning_rate": 4.467129141602906e-06,
|
| 1088 |
+
"loss": 1.6216,
|
| 1089 |
"step": 360
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
"epoch": 0.7,
|
| 1093 |
"learning_rate": 4.3630992350970866e-06,
|
| 1094 |
+
"loss": 1.6651,
|
| 1095 |
"step": 362
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
"epoch": 0.7,
|
| 1099 |
"learning_rate": 4.2599559107388645e-06,
|
| 1100 |
+
"loss": 1.7276,
|
| 1101 |
"step": 364
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.71,
|
| 1105 |
"learning_rate": 4.157715391100509e-06,
|
| 1106 |
+
"loss": 1.7109,
|
| 1107 |
"step": 366
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
"epoch": 0.71,
|
| 1111 |
"learning_rate": 4.056393756759498e-06,
|
| 1112 |
+
"loss": 1.6771,
|
| 1113 |
"step": 368
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
"epoch": 0.72,
|
| 1117 |
"learning_rate": 3.956006943769331e-06,
|
| 1118 |
+
"loss": 1.6131,
|
| 1119 |
"step": 370
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 0.72,
|
| 1123 |
"learning_rate": 3.856570741153087e-06,
|
| 1124 |
+
"loss": 1.589,
|
| 1125 |
"step": 372
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
"epoch": 0.72,
|
| 1129 |
"learning_rate": 3.7581007884200826e-06,
|
| 1130 |
+
"loss": 1.7187,
|
| 1131 |
"step": 374
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
"epoch": 0.73,
|
| 1135 |
"learning_rate": 3.6606125731060815e-06,
|
| 1136 |
+
"loss": 1.7757,
|
| 1137 |
"step": 376
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
"epoch": 0.73,
|
| 1141 |
"learning_rate": 3.564121428337369e-06,
|
| 1142 |
+
"loss": 1.788,
|
| 1143 |
"step": 378
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.73,
|
| 1147 |
"learning_rate": 3.468642530419144e-06,
|
| 1148 |
+
"loss": 1.9998,
|
| 1149 |
"step": 380
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
"epoch": 0.74,
|
| 1153 |
"learning_rate": 3.3741908964485415e-06,
|
| 1154 |
+
"loss": 2.057,
|
| 1155 |
"step": 382
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
"epoch": 0.74,
|
| 1159 |
"learning_rate": 3.2807813819527377e-06,
|
| 1160 |
+
"loss": 1.9311,
|
| 1161 |
"step": 384
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
"epoch": 0.75,
|
| 1165 |
"learning_rate": 3.1884286785524356e-06,
|
| 1166 |
+
"loss": 2.0442,
|
| 1167 |
"step": 386
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
"epoch": 0.75,
|
| 1171 |
"learning_rate": 3.09714731165114e-06,
|
| 1172 |
+
"loss": 2.0233,
|
| 1173 |
"step": 388
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
"epoch": 0.75,
|
| 1177 |
"learning_rate": 3.006951638150567e-06,
|
| 1178 |
+
"loss": 1.9966,
|
| 1179 |
"step": 390
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
"epoch": 0.76,
|
| 1183 |
"learning_rate": 2.9178558441925843e-06,
|
| 1184 |
+
"loss": 2.0437,
|
| 1185 |
"step": 392
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.76,
|
| 1189 |
"learning_rate": 2.829873942927971e-06,
|
| 1190 |
+
"loss": 1.9153,
|
| 1191 |
"step": 394
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
"epoch": 0.77,
|
| 1195 |
"learning_rate": 2.743019772312412e-06,
|
| 1196 |
+
"loss": 1.8868,
|
| 1197 |
"step": 396
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
"epoch": 0.77,
|
| 1201 |
"learning_rate": 2.6573069929300243e-06,
|
| 1202 |
+
"loss": 1.6961,
|
| 1203 |
"step": 398
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
"epoch": 0.77,
|
| 1207 |
"learning_rate": 2.572749085844829e-06,
|
| 1208 |
+
"loss": 1.2739,
|
| 1209 |
"step": 400
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.78,
|
| 1213 |
"learning_rate": 2.4893593504803824e-06,
|
| 1214 |
+
"loss": 1.6079,
|
| 1215 |
"step": 402
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
"epoch": 0.78,
|
| 1219 |
"learning_rate": 2.4071509025280526e-06,
|
| 1220 |
+
"loss": 1.5745,
|
| 1221 |
"step": 404
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
"epoch": 0.79,
|
| 1225 |
"learning_rate": 2.3261366718841307e-06,
|
| 1226 |
+
"loss": 1.7405,
|
| 1227 |
"step": 406
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.79,
|
| 1231 |
"learning_rate": 2.2463294006162185e-06,
|
| 1232 |
+
"loss": 1.5681,
|
| 1233 |
"step": 408
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
"epoch": 0.79,
|
| 1237 |
"learning_rate": 2.167741640959117e-06,
|
| 1238 |
+
"loss": 1.5841,
|
| 1239 |
"step": 410
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
"epoch": 0.8,
|
| 1243 |
"learning_rate": 2.090385753340596e-06,
|
| 1244 |
+
"loss": 1.4798,
|
| 1245 |
"step": 412
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
"epoch": 0.8,
|
| 1249 |
"learning_rate": 2.0142739044373093e-06,
|
| 1250 |
+
"loss": 1.6118,
|
| 1251 |
"step": 414
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
"epoch": 0.8,
|
| 1255 |
"learning_rate": 1.9394180652612193e-06,
|
| 1256 |
+
"loss": 1.6844,
|
| 1257 |
"step": 416
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
"epoch": 0.81,
|
| 1261 |
"learning_rate": 1.8658300092767546e-06,
|
| 1262 |
+
"loss": 1.6168,
|
| 1263 |
"step": 418
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
"epoch": 0.81,
|
| 1267 |
"learning_rate": 1.7935213105490722e-06,
|
| 1268 |
+
"loss": 1.7285,
|
| 1269 |
"step": 420
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.82,
|
| 1273 |
"learning_rate": 1.7225033419236502e-06,
|
| 1274 |
+
"loss": 1.905,
|
| 1275 |
"step": 422
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
"epoch": 0.82,
|
| 1279 |
"learning_rate": 1.652787273237565e-06,
|
| 1280 |
+
"loss": 1.7664,
|
| 1281 |
"step": 424
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
"epoch": 0.82,
|
| 1285 |
"learning_rate": 1.5843840695626666e-06,
|
| 1286 |
+
"loss": 1.9033,
|
| 1287 |
"step": 426
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
"epoch": 0.83,
|
| 1291 |
"learning_rate": 1.5173044894809764e-06,
|
| 1292 |
+
"loss": 1.8572,
|
| 1293 |
"step": 428
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
"epoch": 0.83,
|
| 1297 |
"learning_rate": 1.4515590833925508e-06,
|
| 1298 |
+
"loss": 1.7694,
|
| 1299 |
"step": 430
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.84,
|
| 1303 |
"learning_rate": 1.387158191856105e-06,
|
| 1304 |
+
"loss": 1.8397,
|
| 1305 |
"step": 432
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
"epoch": 0.84,
|
| 1309 |
"learning_rate": 1.324111943962626e-06,
|
| 1310 |
+
"loss": 2.0161,
|
| 1311 |
"step": 434
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.84,
|
| 1315 |
"learning_rate": 1.2624302557422475e-06,
|
| 1316 |
+
"loss": 1.7957,
|
| 1317 |
"step": 436
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
"epoch": 0.85,
|
| 1321 |
"learning_rate": 1.2021228286046461e-06,
|
| 1322 |
+
"loss": 2.0232,
|
| 1323 |
"step": 438
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
"epoch": 0.85,
|
| 1327 |
"learning_rate": 1.1431991478131754e-06,
|
| 1328 |
+
"loss": 2.0685,
|
| 1329 |
"step": 440
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
"epoch": 0.85,
|
| 1333 |
"learning_rate": 1.085668480993015e-06,
|
| 1334 |
+
"loss": 1.8287,
|
| 1335 |
"step": 442
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
"epoch": 0.86,
|
| 1339 |
"learning_rate": 1.029539876673541e-06,
|
| 1340 |
+
"loss": 1.943,
|
| 1341 |
"step": 444
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
"epoch": 0.86,
|
| 1345 |
"learning_rate": 9.748221628651445e-07,
|
| 1346 |
+
"loss": 1.7622,
|
| 1347 |
"step": 446
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
"epoch": 0.87,
|
| 1351 |
"learning_rate": 9.215239456707636e-07,
|
| 1352 |
+
"loss": 1.7487,
|
| 1353 |
"step": 448
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.87,
|
| 1357 |
"learning_rate": 8.696536079322904e-07,
|
| 1358 |
+
"loss": 1.205,
|
| 1359 |
"step": 450
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
"epoch": 0.87,
|
| 1363 |
"learning_rate": 8.192193079121003e-07,
|
| 1364 |
+
"loss": 1.4568,
|
| 1365 |
"step": 452
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
"epoch": 0.88,
|
| 1369 |
"learning_rate": 7.70228978009907e-07,
|
| 1370 |
+
"loss": 1.6962,
|
| 1371 |
"step": 454
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
"epoch": 0.88,
|
| 1375 |
"learning_rate": 7.226903235151439e-07,
|
| 1376 |
+
"loss": 1.5789,
|
| 1377 |
"step": 456
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
"epoch": 0.89,
|
| 1381 |
"learning_rate": 6.766108213950584e-07,
|
| 1382 |
+
"loss": 1.4604,
|
| 1383 |
"step": 458
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
"epoch": 0.89,
|
| 1387 |
"learning_rate": 6.319977191187232e-07,
|
| 1388 |
+
"loss": 1.6399,
|
| 1389 |
"step": 460
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
"epoch": 0.89,
|
| 1393 |
"learning_rate": 5.888580335171368e-07,
|
| 1394 |
+
"loss": 1.5067,
|
| 1395 |
"step": 462
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.9,
|
| 1399 |
"learning_rate": 5.471985496796051e-07,
|
| 1400 |
+
"loss": 1.6029,
|
| 1401 |
"step": 464
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
"epoch": 0.9,
|
| 1405 |
"learning_rate": 5.070258198865741e-07,
|
| 1406 |
+
"loss": 1.6425,
|
| 1407 |
"step": 466
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
"epoch": 0.91,
|
| 1411 |
"learning_rate": 4.683461625790697e-07,
|
| 1412 |
+
"loss": 1.7248,
|
| 1413 |
"step": 468
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
"epoch": 0.91,
|
| 1417 |
"learning_rate": 4.311656613649251e-07,
|
| 1418 |
+
"loss": 1.6521,
|
| 1419 |
"step": 470
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
"epoch": 0.91,
|
| 1423 |
"learning_rate": 3.9549016406193686e-07,
|
| 1424 |
+
"loss": 1.7552,
|
| 1425 |
"step": 472
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
"epoch": 0.92,
|
| 1429 |
"learning_rate": 3.61325281778111e-07,
|
| 1430 |
+
"loss": 1.7562,
|
| 1431 |
"step": 474
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
"epoch": 0.92,
|
| 1435 |
"learning_rate": 3.286763880291399e-07,
|
| 1436 |
+
"loss": 1.8283,
|
| 1437 |
"step": 476
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.92,
|
| 1441 |
"learning_rate": 2.9754861789324074e-07,
|
| 1442 |
+
"loss": 1.7827,
|
| 1443 |
"step": 478
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
"epoch": 0.93,
|
| 1447 |
"learning_rate": 2.6794686720350573e-07,
|
| 1448 |
+
"loss": 1.8353,
|
| 1449 |
"step": 480
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
"epoch": 0.93,
|
| 1453 |
"learning_rate": 2.398757917778727e-07,
|
| 1454 |
+
"loss": 1.902,
|
| 1455 |
"step": 482
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
"epoch": 0.94,
|
| 1459 |
"learning_rate": 2.1333980668685416e-07,
|
| 1460 |
+
"loss": 1.859,
|
| 1461 |
"step": 484
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
"epoch": 0.94,
|
| 1465 |
"learning_rate": 1.883430855591173e-07,
|
| 1466 |
+
"loss": 1.8983,
|
| 1467 |
"step": 486
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
"epoch": 0.94,
|
| 1471 |
"learning_rate": 1.6488955992506218e-07,
|
| 1472 |
+
"loss": 1.8882,
|
| 1473 |
"step": 488
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
"epoch": 0.95,
|
| 1477 |
"learning_rate": 1.4298291859845215e-07,
|
| 1478 |
+
"loss": 2.1243,
|
| 1479 |
"step": 490
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.95,
|
| 1483 |
"learning_rate": 1.2262660709623232e-07,
|
| 1484 |
+
"loss": 1.9541,
|
| 1485 |
"step": 492
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
"epoch": 0.96,
|
| 1489 |
"learning_rate": 1.0382382709661853e-07,
|
| 1490 |
+
"loss": 2.0218,
|
| 1491 |
"step": 494
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
"epoch": 0.96,
|
| 1495 |
"learning_rate": 8.657753593552142e-08,
|
| 1496 |
+
"loss": 1.858,
|
| 1497 |
"step": 496
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
"epoch": 0.96,
|
| 1501 |
"learning_rate": 7.08904461414217e-08,
|
| 1502 |
+
"loss": 1.6106,
|
| 1503 |
"step": 498
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
"epoch": 0.97,
|
| 1507 |
"learning_rate": 5.676502500873038e-08,
|
| 1508 |
+
"loss": 1.2003,
|
| 1509 |
"step": 500
|
| 1510 |
}
|
| 1511 |
],
|
|
|
|
| 1513 |
"max_steps": 517,
|
| 1514 |
"num_train_epochs": 1,
|
| 1515 |
"save_steps": 50,
|
| 1516 |
+
"total_flos": 8469410469322752.0,
|
| 1517 |
"trial_name": null,
|
| 1518 |
"trial_params": null
|
| 1519 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4027
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3a24a3ed07d90a9e9e9622aa97eef7cc0d5f56a678dd98280028b85bac9669d
|
| 3 |
size 4027
|