Model save
Browse files- README.md +5 -9
- all_results.json +4 -9
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- runs/Jun04_12-32-53_node-0/events.out.tfevents.1717505644.node-0.1922.0 +2 -2
- train_results.json +4 -4
- trainer_state.json +430 -430
README.md
CHANGED
|
@@ -2,15 +2,11 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: mistralai/Mistral-7B-v0.1
|
| 4 |
tags:
|
| 5 |
-
- alignment-handbook
|
| 6 |
-
- trl
|
| 7 |
-
- sft
|
| 8 |
-
- generated_from_trainer
|
| 9 |
- trl
|
| 10 |
- sft
|
| 11 |
- generated_from_trainer
|
| 12 |
datasets:
|
| 13 |
-
-
|
| 14 |
model-index:
|
| 15 |
- name: zephyr-7b-sft-full
|
| 16 |
results: []
|
|
@@ -21,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 21 |
|
| 22 |
# zephyr-7b-sft-full
|
| 23 |
|
| 24 |
-
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
-
- Loss: 0.
|
| 27 |
|
| 28 |
## Model description
|
| 29 |
|
|
@@ -59,12 +55,12 @@ The following hyperparameters were used during training:
|
|
| 59 |
|
| 60 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 61 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 62 |
-
| 0.9203 | 1.0 | 1090 | 0.
|
| 63 |
|
| 64 |
|
| 65 |
### Framework versions
|
| 66 |
|
| 67 |
- Transformers 4.41.2
|
| 68 |
-
- Pytorch 2.3.0
|
| 69 |
- Datasets 2.19.2
|
| 70 |
- Tokenizers 0.19.1
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: mistralai/Mistral-7B-v0.1
|
| 4 |
tags:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
- trl
|
| 6 |
- sft
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
+
- generator
|
| 10 |
model-index:
|
| 11 |
- name: zephyr-7b-sft-full
|
| 12 |
results: []
|
|
|
|
| 17 |
|
| 18 |
# zephyr-7b-sft-full
|
| 19 |
|
| 20 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.9367
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 55 |
|
| 56 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 58 |
+
| 0.9203 | 1.0 | 1090 | 0.9367 |
|
| 59 |
|
| 60 |
|
| 61 |
### Framework versions
|
| 62 |
|
| 63 |
- Transformers 4.41.2
|
| 64 |
+
- Pytorch 2.3.0
|
| 65 |
- Datasets 2.19.2
|
| 66 |
- Tokenizers 0.19.1
|
all_results.json
CHANGED
|
@@ -1,14 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_loss": 0.9365589618682861,
|
| 4 |
-
"eval_runtime": 321.44,
|
| 5 |
-
"eval_samples": 23109,
|
| 6 |
-
"eval_samples_per_second": 48.006,
|
| 7 |
-
"eval_steps_per_second": 0.753,
|
| 8 |
"total_flos": 456447649382400.0,
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
"train_samples": 207864,
|
| 12 |
-
"train_samples_per_second":
|
| 13 |
-
"train_steps_per_second": 0.
|
| 14 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
"total_flos": 456447649382400.0,
|
| 4 |
+
"train_loss": 0.9783222812031387,
|
| 5 |
+
"train_runtime": 12140.5299,
|
| 6 |
"train_samples": 207864,
|
| 7 |
+
"train_samples_per_second": 11.484,
|
| 8 |
+
"train_steps_per_second": 0.09
|
| 9 |
}
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4943162336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb4d8a4ff3c4f17d59bdbb895f9696f5601fd41f27cf6bb8f18d208b94cb9111
|
| 3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aea583a7073c12c17452ce853c128a92db21020555c4b59eb66c6820b9170974
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4540516344
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e0cd7368dd483de5e4c9615682532c06cb00eabe902e8d112b18eeb6dfebd986
|
| 3 |
size 4540516344
|
runs/Jun04_12-32-53_node-0/events.out.tfevents.1717505644.node-0.1922.0
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:e579e433c2dbbf504749f56ed853d1501d2954e1375b1e8690ecad1389df5864
|
| 3 |
+
size 51635
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 456447649382400.0,
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 207864,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 456447649382400.0,
|
| 4 |
+
"train_loss": 0.9783222812031387,
|
| 5 |
+
"train_runtime": 12140.5299,
|
| 6 |
"train_samples": 207864,
|
| 7 |
+
"train_samples_per_second": 11.484,
|
| 8 |
+
"train_steps_per_second": 0.09
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -10,1553 +10,1553 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0009174311926605505,
|
| 13 |
-
"grad_norm": 8.
|
| 14 |
"learning_rate": 1.8348623853211012e-07,
|
| 15 |
"loss": 1.1366,
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.0045871559633027525,
|
| 20 |
-
"grad_norm": 7.
|
| 21 |
"learning_rate": 9.174311926605506e-07,
|
| 22 |
"loss": 1.1311,
|
| 23 |
"step": 5
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.009174311926605505,
|
| 27 |
-
"grad_norm": 4.
|
| 28 |
"learning_rate": 1.8348623853211011e-06,
|
| 29 |
"loss": 1.0455,
|
| 30 |
"step": 10
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.013761467889908258,
|
| 34 |
-
"grad_norm":
|
| 35 |
"learning_rate": 2.7522935779816517e-06,
|
| 36 |
-
"loss": 1.
|
| 37 |
"step": 15
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.01834862385321101,
|
| 41 |
-
"grad_norm": 2.
|
| 42 |
"learning_rate": 3.6697247706422022e-06,
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 20
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.022935779816513763,
|
| 48 |
-
"grad_norm": 2.
|
| 49 |
"learning_rate": 4.587155963302753e-06,
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 25
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.027522935779816515,
|
| 55 |
-
"grad_norm": 2.
|
| 56 |
"learning_rate": 5.504587155963303e-06,
|
| 57 |
-
"loss": 0.
|
| 58 |
"step": 30
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.03211009174311927,
|
| 62 |
-
"grad_norm":
|
| 63 |
"learning_rate": 6.422018348623854e-06,
|
| 64 |
-
"loss": 0.
|
| 65 |
"step": 35
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.03669724770642202,
|
| 69 |
-
"grad_norm": 1.
|
| 70 |
"learning_rate": 7.3394495412844045e-06,
|
| 71 |
-
"loss": 0.
|
| 72 |
"step": 40
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.04128440366972477,
|
| 76 |
-
"grad_norm": 2.
|
| 77 |
"learning_rate": 8.256880733944956e-06,
|
| 78 |
-
"loss": 0.
|
| 79 |
"step": 45
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.045871559633027525,
|
| 83 |
-
"grad_norm": 2.
|
| 84 |
"learning_rate": 9.174311926605506e-06,
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 50
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.05045871559633028,
|
| 90 |
-
"grad_norm": 2.
|
| 91 |
"learning_rate": 1.0091743119266055e-05,
|
| 92 |
-
"loss": 0.
|
| 93 |
"step": 55
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.05504587155963303,
|
| 97 |
-
"grad_norm": 2.
|
| 98 |
"learning_rate": 1.1009174311926607e-05,
|
| 99 |
-
"loss": 0.
|
| 100 |
"step": 60
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.05963302752293578,
|
| 104 |
-
"grad_norm": 2.
|
| 105 |
"learning_rate": 1.1926605504587156e-05,
|
| 106 |
-
"loss": 0.
|
| 107 |
"step": 65
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.06422018348623854,
|
| 111 |
-
"grad_norm": 2.
|
| 112 |
"learning_rate": 1.2844036697247708e-05,
|
| 113 |
-
"loss": 0.
|
| 114 |
"step": 70
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.06880733944954129,
|
| 118 |
-
"grad_norm": 2.
|
| 119 |
"learning_rate": 1.3761467889908258e-05,
|
| 120 |
"loss": 1.0001,
|
| 121 |
"step": 75
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.07339449541284404,
|
| 125 |
-
"grad_norm":
|
| 126 |
"learning_rate": 1.4678899082568809e-05,
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 80
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.0779816513761468,
|
| 132 |
-
"grad_norm": 2.
|
| 133 |
"learning_rate": 1.559633027522936e-05,
|
| 134 |
-
"loss": 0.
|
| 135 |
"step": 85
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.08256880733944955,
|
| 139 |
-
"grad_norm": 2.
|
| 140 |
"learning_rate": 1.6513761467889912e-05,
|
| 141 |
-
"loss": 1.
|
| 142 |
"step": 90
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.0871559633027523,
|
| 146 |
-
"grad_norm": 2.
|
| 147 |
"learning_rate": 1.743119266055046e-05,
|
| 148 |
-
"loss": 1.
|
| 149 |
"step": 95
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.09174311926605505,
|
| 153 |
-
"grad_norm": 2.
|
| 154 |
"learning_rate": 1.834862385321101e-05,
|
| 155 |
-
"loss": 0.
|
| 156 |
"step": 100
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.0963302752293578,
|
| 160 |
-
"grad_norm": 2.
|
| 161 |
"learning_rate": 1.9266055045871563e-05,
|
| 162 |
-
"loss": 1.
|
| 163 |
"step": 105
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.10091743119266056,
|
| 167 |
-
"grad_norm": 2.
|
| 168 |
"learning_rate": 1.999994872196626e-05,
|
| 169 |
-
"loss": 1.
|
| 170 |
"step": 110
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.10550458715596331,
|
| 174 |
-
"grad_norm":
|
| 175 |
"learning_rate": 1.9998154046002822e-05,
|
| 176 |
-
"loss": 1.
|
| 177 |
"step": 115
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.11009174311926606,
|
| 181 |
-
"grad_norm":
|
| 182 |
"learning_rate": 1.999379599421534e-05,
|
| 183 |
-
"loss": 1.
|
| 184 |
"step": 120
|
| 185 |
},
|
| 186 |
{
|
| 187 |
"epoch": 0.11467889908256881,
|
| 188 |
-
"grad_norm":
|
| 189 |
"learning_rate": 1.9986875683942535e-05,
|
| 190 |
-
"loss": 1.
|
| 191 |
"step": 125
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.11926605504587157,
|
| 195 |
-
"grad_norm": 1.
|
| 196 |
"learning_rate": 1.9977394889447526e-05,
|
| 197 |
-
"loss": 1.
|
| 198 |
"step": 130
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"epoch": 0.12385321100917432,
|
| 202 |
-
"grad_norm": 1.
|
| 203 |
"learning_rate": 1.9965356041462954e-05,
|
| 204 |
-
"loss": 0.
|
| 205 |
"step": 135
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.12844036697247707,
|
| 209 |
-
"grad_norm":
|
| 210 |
"learning_rate": 1.9950762226567783e-05,
|
| 211 |
-
"loss": 1.
|
| 212 |
"step": 140
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.13302752293577982,
|
| 216 |
-
"grad_norm":
|
| 217 |
"learning_rate": 1.9933617186395917e-05,
|
| 218 |
-
"loss": 1.
|
| 219 |
"step": 145
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.13761467889908258,
|
| 223 |
-
"grad_norm": 2.
|
| 224 |
"learning_rate": 1.9913925316676946e-05,
|
| 225 |
-
"loss": 1.
|
| 226 |
"step": 150
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.14220183486238533,
|
| 230 |
-
"grad_norm":
|
| 231 |
"learning_rate": 1.9891691666109112e-05,
|
| 232 |
-
"loss": 1.
|
| 233 |
"step": 155
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.14678899082568808,
|
| 237 |
-
"grad_norm": 1.
|
| 238 |
"learning_rate": 1.9866921935064907e-05,
|
| 239 |
"loss": 1.0348,
|
| 240 |
"step": 160
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 0.15137614678899083,
|
| 244 |
-
"grad_norm":
|
| 245 |
"learning_rate": 1.9839622474129595e-05,
|
| 246 |
-
"loss": 1.
|
| 247 |
"step": 165
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.1559633027522936,
|
| 251 |
-
"grad_norm": 1.
|
| 252 |
"learning_rate": 1.9809800282473014e-05,
|
| 253 |
-
"loss": 1.
|
| 254 |
"step": 170
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.16055045871559634,
|
| 258 |
-
"grad_norm":
|
| 259 |
"learning_rate": 1.977746300605507e-05,
|
| 260 |
-
"loss": 1.
|
| 261 |
"step": 175
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.1651376146788991,
|
| 265 |
-
"grad_norm": 2.
|
| 266 |
"learning_rate": 1.9742618935665478e-05,
|
| 267 |
-
"loss": 1.
|
| 268 |
"step": 180
|
| 269 |
},
|
| 270 |
{
|
| 271 |
"epoch": 0.16972477064220184,
|
| 272 |
-
"grad_norm": 1.
|
| 273 |
"learning_rate": 1.9705277004798072e-05,
|
| 274 |
-
"loss": 1.
|
| 275 |
"step": 185
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.1743119266055046,
|
| 279 |
-
"grad_norm":
|
| 280 |
"learning_rate": 1.9665446787360444e-05,
|
| 281 |
-
"loss": 1.
|
| 282 |
"step": 190
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 0.17889908256880735,
|
| 286 |
-
"grad_norm":
|
| 287 |
"learning_rate": 1.9623138495219292e-05,
|
| 288 |
-
"loss": 1.
|
| 289 |
"step": 195
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.1834862385321101,
|
| 293 |
-
"grad_norm": 1.
|
| 294 |
"learning_rate": 1.957836297558229e-05,
|
| 295 |
-
"loss": 1.
|
| 296 |
"step": 200
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.18807339449541285,
|
| 300 |
-
"grad_norm": 1.
|
| 301 |
"learning_rate": 1.9531131708217005e-05,
|
| 302 |
-
"loss": 1.
|
| 303 |
"step": 205
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.1926605504587156,
|
| 307 |
-
"grad_norm": 1.
|
| 308 |
"learning_rate": 1.948145680250766e-05,
|
| 309 |
-
"loss": 1.
|
| 310 |
"step": 210
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 0.19724770642201836,
|
| 314 |
-
"grad_norm": 1.
|
| 315 |
"learning_rate": 1.9429350994350483e-05,
|
| 316 |
-
"loss": 1.
|
| 317 |
"step": 215
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 0.2018348623853211,
|
| 321 |
-
"grad_norm": 1.
|
| 322 |
"learning_rate": 1.93748276428884e-05,
|
| 323 |
-
"loss": 1.
|
| 324 |
"step": 220
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.20642201834862386,
|
| 328 |
-
"grad_norm": 1.
|
| 329 |
"learning_rate": 1.931790072708596e-05,
|
| 330 |
-
"loss": 1.
|
| 331 |
"step": 225
|
| 332 |
},
|
| 333 |
{
|
| 334 |
"epoch": 0.21100917431192662,
|
| 335 |
-
"grad_norm": 1.
|
| 336 |
"learning_rate": 1.9258584842145342e-05,
|
| 337 |
-
"loss": 1.
|
| 338 |
"step": 230
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.21559633027522937,
|
| 342 |
-
"grad_norm": 1.
|
| 343 |
"learning_rate": 1.9196895195764363e-05,
|
| 344 |
-
"loss": 1.
|
| 345 |
"step": 235
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22018348623853212,
|
| 349 |
-
"grad_norm": 1.
|
| 350 |
"learning_rate": 1.913284760423745e-05,
|
| 351 |
-
"loss": 1.
|
| 352 |
"step": 240
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.22477064220183487,
|
| 356 |
-
"grad_norm": 1.
|
| 357 |
"learning_rate": 1.9066458488400586e-05,
|
| 358 |
-
"loss": 1.
|
| 359 |
"step": 245
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 0.22935779816513763,
|
| 363 |
-
"grad_norm": 1.
|
| 364 |
"learning_rate": 1.8997744869421248e-05,
|
| 365 |
-
"loss": 1.
|
| 366 |
"step": 250
|
| 367 |
},
|
| 368 |
{
|
| 369 |
"epoch": 0.23394495412844038,
|
| 370 |
-
"grad_norm":
|
| 371 |
"learning_rate": 1.8926724364434447e-05,
|
| 372 |
-
"loss": 1.
|
| 373 |
"step": 255
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 0.23853211009174313,
|
| 377 |
-
"grad_norm": 1.
|
| 378 |
"learning_rate": 1.8853415182025953e-05,
|
| 379 |
-
"loss": 1.
|
| 380 |
"step": 260
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.24311926605504589,
|
| 384 |
-
"grad_norm": 1.
|
| 385 |
"learning_rate": 1.8777836117563894e-05,
|
| 386 |
-
"loss": 1.
|
| 387 |
"step": 265
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.24770642201834864,
|
| 391 |
-
"grad_norm": 1.
|
| 392 |
"learning_rate": 1.8700006548379898e-05,
|
| 393 |
-
"loss": 1.
|
| 394 |
"step": 270
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 0.25229357798165136,
|
| 398 |
-
"grad_norm": 1.
|
| 399 |
"learning_rate": 1.861994642880105e-05,
|
| 400 |
-
"loss": 1.
|
| 401 |
"step": 275
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 0.25688073394495414,
|
| 405 |
-
"grad_norm": 1.
|
| 406 |
"learning_rate": 1.8537676285033886e-05,
|
| 407 |
-
"loss": 1.
|
| 408 |
"step": 280
|
| 409 |
},
|
| 410 |
{
|
| 411 |
"epoch": 0.26146788990825687,
|
| 412 |
-
"grad_norm": 1.
|
| 413 |
"learning_rate": 1.845321720990181e-05,
|
| 414 |
-
"loss": 1.
|
| 415 |
"step": 285
|
| 416 |
},
|
| 417 |
{
|
| 418 |
"epoch": 0.26605504587155965,
|
| 419 |
-
"grad_norm": 1.
|
| 420 |
"learning_rate": 1.8366590857437182e-05,
|
| 421 |
-
"loss": 1.
|
| 422 |
"step": 290
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.2706422018348624,
|
| 426 |
-
"grad_norm": 1.
|
| 427 |
"learning_rate": 1.8277819437329577e-05,
|
| 428 |
-
"loss": 1.
|
| 429 |
"step": 295
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.27522935779816515,
|
| 433 |
-
"grad_norm": 1.
|
| 434 |
"learning_rate": 1.8186925709231534e-05,
|
| 435 |
-
"loss": 1.
|
| 436 |
"step": 300
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.2798165137614679,
|
| 440 |
-
"grad_norm": 1.
|
| 441 |
"learning_rate": 1.809393297692334e-05,
|
| 442 |
-
"loss": 1.
|
| 443 |
"step": 305
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 0.28440366972477066,
|
| 447 |
-
"grad_norm": 1.
|
| 448 |
"learning_rate": 1.799886508233829e-05,
|
| 449 |
-
"loss": 1.
|
| 450 |
"step": 310
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"epoch": 0.2889908256880734,
|
| 454 |
-
"grad_norm": 1.
|
| 455 |
"learning_rate": 1.790174639944997e-05,
|
| 456 |
-
"loss": 1.
|
| 457 |
"step": 315
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.29357798165137616,
|
| 461 |
-
"grad_norm": 1.
|
| 462 |
"learning_rate": 1.780260182802314e-05,
|
| 463 |
-
"loss": 1.
|
| 464 |
"step": 320
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.2981651376146789,
|
| 468 |
-
"grad_norm": 1.
|
| 469 |
"learning_rate": 1.7701456787229805e-05,
|
| 470 |
-
"loss": 1.
|
| 471 |
"step": 325
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.30275229357798167,
|
| 475 |
-
"grad_norm": 1.
|
| 476 |
"learning_rate": 1.7598337209132142e-05,
|
| 477 |
-
"loss": 1.
|
| 478 |
"step": 330
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.3073394495412844,
|
| 482 |
-
"grad_norm": 1.
|
| 483 |
"learning_rate": 1.7493269532033882e-05,
|
| 484 |
-
"loss": 1.
|
| 485 |
"step": 335
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 0.3119266055045872,
|
| 489 |
-
"grad_norm":
|
| 490 |
"learning_rate": 1.738628069370195e-05,
|
| 491 |
-
"loss": 1.
|
| 492 |
"step": 340
|
| 493 |
},
|
| 494 |
{
|
| 495 |
"epoch": 0.3165137614678899,
|
| 496 |
-
"grad_norm": 1.
|
| 497 |
"learning_rate": 1.7277398124460022e-05,
|
| 498 |
-
"loss": 1.
|
| 499 |
"step": 345
|
| 500 |
},
|
| 501 |
{
|
| 502 |
"epoch": 0.3211009174311927,
|
| 503 |
-
"grad_norm": 1.
|
| 504 |
"learning_rate": 1.71666497401558e-05,
|
| 505 |
-
"loss": 0.
|
| 506 |
"step": 350
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.3256880733944954,
|
| 510 |
-
"grad_norm": 1.
|
| 511 |
"learning_rate": 1.7054063935003813e-05,
|
| 512 |
"loss": 1.0253,
|
| 513 |
"step": 355
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.3302752293577982,
|
| 517 |
-
"grad_norm": 1.
|
| 518 |
"learning_rate": 1.6939669574305565e-05,
|
| 519 |
-
"loss": 1.
|
| 520 |
"step": 360
|
| 521 |
},
|
| 522 |
{
|
| 523 |
"epoch": 0.3348623853211009,
|
| 524 |
-
"grad_norm": 1.
|
| 525 |
"learning_rate": 1.6823495987048922e-05,
|
| 526 |
-
"loss": 1.
|
| 527 |
"step": 365
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 0.3394495412844037,
|
| 531 |
-
"grad_norm": 1.
|
| 532 |
"learning_rate": 1.6705572958388576e-05,
|
| 533 |
-
"loss": 1.
|
| 534 |
"step": 370
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.3440366972477064,
|
| 538 |
-
"grad_norm": 1.
|
| 539 |
"learning_rate": 1.6585930722009602e-05,
|
| 540 |
-
"loss": 1.
|
| 541 |
"step": 375
|
| 542 |
},
|
| 543 |
{
|
| 544 |
"epoch": 0.3486238532110092,
|
| 545 |
-
"grad_norm": 1.
|
| 546 |
"learning_rate": 1.6464599952375998e-05,
|
| 547 |
-
"loss": 1.
|
| 548 |
"step": 380
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.3532110091743119,
|
| 552 |
-
"grad_norm": 1.
|
| 553 |
"learning_rate": 1.63416117568662e-05,
|
| 554 |
-
"loss": 1.
|
| 555 |
"step": 385
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.3577981651376147,
|
| 559 |
-
"grad_norm": 1.
|
| 560 |
"learning_rate": 1.621699766779763e-05,
|
| 561 |
-
"loss": 1.
|
| 562 |
"step": 390
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 0.3623853211009174,
|
| 566 |
-
"grad_norm": 1.
|
| 567 |
"learning_rate": 1.6090789634342278e-05,
|
| 568 |
-
"loss": 1.
|
| 569 |
"step": 395
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.3669724770642202,
|
| 573 |
-
"grad_norm": 1.
|
| 574 |
"learning_rate": 1.5963020014335437e-05,
|
| 575 |
-
"loss": 1.
|
| 576 |
"step": 400
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"epoch": 0.37155963302752293,
|
| 580 |
-
"grad_norm": 1.
|
| 581 |
"learning_rate": 1.583372156597961e-05,
|
| 582 |
-
"loss": 1.
|
| 583 |
"step": 405
|
| 584 |
},
|
| 585 |
{
|
| 586 |
"epoch": 0.3761467889908257,
|
| 587 |
-
"grad_norm": 1.
|
| 588 |
"learning_rate": 1.570292743944583e-05,
|
| 589 |
-
"loss": 1.
|
| 590 |
"step": 410
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.38073394495412843,
|
| 594 |
-
"grad_norm": 1.
|
| 595 |
"learning_rate": 1.557067116837444e-05,
|
| 596 |
-
"loss": 1.
|
| 597 |
"step": 415
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.3853211009174312,
|
| 601 |
-
"grad_norm": 1.
|
| 602 |
"learning_rate": 1.5436986661277578e-05,
|
| 603 |
-
"loss": 1.
|
| 604 |
"step": 420
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.38990825688073394,
|
| 608 |
-
"grad_norm": 1.
|
| 609 |
"learning_rate": 1.530190819284555e-05,
|
| 610 |
-
"loss": 1.
|
| 611 |
"step": 425
|
| 612 |
},
|
| 613 |
{
|
| 614 |
"epoch": 0.3944954128440367,
|
| 615 |
-
"grad_norm": 1.
|
| 616 |
"learning_rate": 1.5165470395159314e-05,
|
| 617 |
"loss": 1.0335,
|
| 618 |
"step": 430
|
| 619 |
},
|
| 620 |
{
|
| 621 |
"epoch": 0.39908256880733944,
|
| 622 |
-
"grad_norm": 1.
|
| 623 |
"learning_rate": 1.5027708248811331e-05,
|
| 624 |
"loss": 1.0006,
|
| 625 |
"step": 435
|
| 626 |
},
|
| 627 |
{
|
| 628 |
"epoch": 0.4036697247706422,
|
| 629 |
-
"grad_norm": 1.
|
| 630 |
"learning_rate": 1.4888657073937077e-05,
|
| 631 |
-
"loss": 1.
|
| 632 |
"step": 440
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.40825688073394495,
|
| 636 |
-
"grad_norm": 1.
|
| 637 |
"learning_rate": 1.4748352521159492e-05,
|
| 638 |
-
"loss": 0.
|
| 639 |
"step": 445
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.41284403669724773,
|
| 643 |
-
"grad_norm": 1.
|
| 644 |
"learning_rate": 1.4606830562448692e-05,
|
| 645 |
-
"loss": 1.
|
| 646 |
"step": 450
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 0.41743119266055045,
|
| 650 |
-
"grad_norm": 1.
|
| 651 |
"learning_rate": 1.4464127481899312e-05,
|
| 652 |
-
"loss": 1.
|
| 653 |
"step": 455
|
| 654 |
},
|
| 655 |
{
|
| 656 |
"epoch": 0.42201834862385323,
|
| 657 |
-
"grad_norm": 1.
|
| 658 |
"learning_rate": 1.4320279866427798e-05,
|
| 659 |
-
"loss": 1.
|
| 660 |
"step": 460
|
| 661 |
},
|
| 662 |
{
|
| 663 |
"epoch": 0.42660550458715596,
|
| 664 |
-
"grad_norm": 1.
|
| 665 |
"learning_rate": 1.4175324596392075e-05,
|
| 666 |
-
"loss": 1.
|
| 667 |
"step": 465
|
| 668 |
},
|
| 669 |
{
|
| 670 |
"epoch": 0.43119266055045874,
|
| 671 |
-
"grad_norm": 1.
|
| 672 |
"learning_rate": 1.402929883613599e-05,
|
| 673 |
-
"loss": 1.
|
| 674 |
"step": 470
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 0.43577981651376146,
|
| 678 |
-
"grad_norm": 1.
|
| 679 |
"learning_rate": 1.3882240024460928e-05,
|
| 680 |
"loss": 0.9925,
|
| 681 |
"step": 475
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44036697247706424,
|
| 685 |
-
"grad_norm": 1.
|
| 686 |
"learning_rate": 1.3734185865027061e-05,
|
| 687 |
-
"loss": 1.
|
| 688 |
"step": 480
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 0.44495412844036697,
|
| 692 |
-
"grad_norm": 1.
|
| 693 |
"learning_rate": 1.358517431668672e-05,
|
| 694 |
-
"loss": 0.
|
| 695 |
"step": 485
|
| 696 |
},
|
| 697 |
{
|
| 698 |
"epoch": 0.44954128440366975,
|
| 699 |
-
"grad_norm": 1.
|
| 700 |
"learning_rate": 1.3435243583752294e-05,
|
| 701 |
-
"loss": 0.
|
| 702 |
"step": 490
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"epoch": 0.4541284403669725,
|
| 706 |
-
"grad_norm": 1.
|
| 707 |
"learning_rate": 1.3284432106201233e-05,
|
| 708 |
-
"loss": 0.
|
| 709 |
"step": 495
|
| 710 |
},
|
| 711 |
{
|
| 712 |
"epoch": 0.45871559633027525,
|
| 713 |
-
"grad_norm": 1.
|
| 714 |
"learning_rate": 1.313277854982062e-05,
|
| 715 |
-
"loss": 1.
|
| 716 |
"step": 500
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 0.463302752293578,
|
| 720 |
-
"grad_norm": 1.
|
| 721 |
"learning_rate": 1.2980321796293838e-05,
|
| 722 |
-
"loss": 0.
|
| 723 |
"step": 505
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.46788990825688076,
|
| 727 |
-
"grad_norm": 1.
|
| 728 |
"learning_rate": 1.2827100933231904e-05,
|
| 729 |
-
"loss": 1.
|
| 730 |
"step": 510
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.4724770642201835,
|
| 734 |
-
"grad_norm": 1.
|
| 735 |
"learning_rate": 1.2673155244151985e-05,
|
| 736 |
-
"loss": 1.
|
| 737 |
"step": 515
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 0.47706422018348627,
|
| 741 |
-
"grad_norm": 1.
|
| 742 |
"learning_rate": 1.2518524198405699e-05,
|
| 743 |
-
"loss": 0.
|
| 744 |
"step": 520
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.481651376146789,
|
| 748 |
-
"grad_norm": 1.
|
| 749 |
"learning_rate": 1.2363247441059775e-05,
|
| 750 |
-
"loss": 0.
|
| 751 |
"step": 525
|
| 752 |
},
|
| 753 |
{
|
| 754 |
"epoch": 0.48623853211009177,
|
| 755 |
-
"grad_norm": 1.
|
| 756 |
"learning_rate": 1.2207364782731657e-05,
|
| 757 |
-
"loss": 1.
|
| 758 |
"step": 530
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 0.4908256880733945,
|
| 762 |
-
"grad_norm": 1.
|
| 763 |
"learning_rate": 1.2050916189382646e-05,
|
| 764 |
-
"loss": 0.
|
| 765 |
"step": 535
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.4954128440366973,
|
| 769 |
-
"grad_norm": 1.
|
| 770 |
"learning_rate": 1.189394177207125e-05,
|
| 771 |
-
"loss": 0.
|
| 772 |
"step": 540
|
| 773 |
},
|
| 774 |
{
|
| 775 |
"epoch": 0.5,
|
| 776 |
-
"grad_norm": 1.
|
| 777 |
"learning_rate": 1.1736481776669307e-05,
|
| 778 |
-
"loss": 0.
|
| 779 |
"step": 545
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 0.5045871559633027,
|
| 783 |
-
"grad_norm": 1.
|
| 784 |
"learning_rate": 1.1578576573543541e-05,
|
| 785 |
-
"loss": 0.
|
| 786 |
"step": 550
|
| 787 |
},
|
| 788 |
{
|
| 789 |
"epoch": 0.5091743119266054,
|
| 790 |
-
"grad_norm": 1.
|
| 791 |
"learning_rate": 1.1420266647205232e-05,
|
| 792 |
"loss": 0.9744,
|
| 793 |
"step": 555
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.5137614678899083,
|
| 797 |
-
"grad_norm": 1.
|
| 798 |
"learning_rate": 1.1261592585930576e-05,
|
| 799 |
-
"loss": 0.
|
| 800 |
"step": 560
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 0.518348623853211,
|
| 804 |
-
"grad_norm": 1.
|
| 805 |
"learning_rate": 1.1102595071354471e-05,
|
| 806 |
-
"loss": 0.
|
| 807 |
"step": 565
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.5229357798165137,
|
| 811 |
-
"grad_norm": 1.
|
| 812 |
"learning_rate": 1.0943314868040365e-05,
|
| 813 |
-
"loss": 0.
|
| 814 |
"step": 570
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 0.5275229357798165,
|
| 818 |
-
"grad_norm": 1.
|
| 819 |
"learning_rate": 1.0783792813028828e-05,
|
| 820 |
"loss": 0.9883,
|
| 821 |
"step": 575
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 0.5321100917431193,
|
| 825 |
-
"grad_norm": 1.
|
| 826 |
"learning_rate": 1.0624069805367558e-05,
|
| 827 |
-
"loss": 0.
|
| 828 |
"step": 580
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 0.536697247706422,
|
| 832 |
-
"grad_norm": 1.
|
| 833 |
"learning_rate": 1.0464186795625481e-05,
|
| 834 |
-
"loss": 0.
|
| 835 |
"step": 585
|
| 836 |
},
|
| 837 |
{
|
| 838 |
"epoch": 0.5412844036697247,
|
| 839 |
-
"grad_norm": 1.
|
| 840 |
"learning_rate": 1.0304184775393642e-05,
|
| 841 |
-
"loss": 0.
|
| 842 |
"step": 590
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 0.5458715596330275,
|
| 846 |
-
"grad_norm": 1.
|
| 847 |
"learning_rate": 1.0144104766775574e-05,
|
| 848 |
-
"loss": 0.
|
| 849 |
"step": 595
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.5504587155963303,
|
| 853 |
-
"grad_norm": 1.
|
| 854 |
"learning_rate": 9.983987811869863e-06,
|
| 855 |
-
"loss": 0.
|
| 856 |
"step": 600
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.555045871559633,
|
| 860 |
-
"grad_norm": 1.
|
| 861 |
"learning_rate": 9.823874962247565e-06,
|
| 862 |
-
"loss": 0.
|
| 863 |
"step": 605
|
| 864 |
},
|
| 865 |
{
|
| 866 |
"epoch": 0.5596330275229358,
|
| 867 |
-
"grad_norm": 1.
|
| 868 |
"learning_rate": 9.663807268427197e-06,
|
| 869 |
-
"loss": 0.
|
| 870 |
"step": 610
|
| 871 |
},
|
| 872 |
{
|
| 873 |
"epoch": 0.5642201834862385,
|
| 874 |
-
"grad_norm": 1.
|
| 875 |
"learning_rate": 9.503825769350016e-06,
|
| 876 |
-
"loss": 0.
|
| 877 |
"step": 615
|
| 878 |
},
|
| 879 |
{
|
| 880 |
"epoch": 0.5688073394495413,
|
| 881 |
-
"grad_norm": 1.
|
| 882 |
"learning_rate": 9.343971481858246e-06,
|
| 883 |
-
"loss": 0.
|
| 884 |
"step": 620
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 0.573394495412844,
|
| 888 |
-
"grad_norm": 1.
|
| 889 |
"learning_rate": 9.184285390178978e-06,
|
| 890 |
-
"loss": 0.
|
| 891 |
"step": 625
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.5779816513761468,
|
| 895 |
-
"grad_norm": 1.
|
| 896 |
"learning_rate": 9.024808435416435e-06,
|
| 897 |
-
"loss": 0.
|
| 898 |
"step": 630
|
| 899 |
},
|
| 900 |
{
|
| 901 |
"epoch": 0.5825688073394495,
|
| 902 |
-
"grad_norm": 1.
|
| 903 |
"learning_rate": 8.865581505055292e-06,
|
| 904 |
-
"loss": 0.
|
| 905 |
"step": 635
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.5871559633027523,
|
| 909 |
-
"grad_norm": 1.
|
| 910 |
"learning_rate": 8.706645422477739e-06,
|
| 911 |
-
"loss": 0.
|
| 912 |
"step": 640
|
| 913 |
},
|
| 914 |
{
|
| 915 |
"epoch": 0.591743119266055,
|
| 916 |
-
"grad_norm": 1.
|
| 917 |
"learning_rate": 8.548040936496989e-06,
|
| 918 |
-
"loss": 0.
|
| 919 |
"step": 645
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 0.5963302752293578,
|
| 923 |
-
"grad_norm": 1.
|
| 924 |
"learning_rate": 8.389808710909881e-06,
|
| 925 |
-
"loss": 0.
|
| 926 |
"step": 650
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 0.6009174311926605,
|
| 930 |
-
"grad_norm": 1.
|
| 931 |
"learning_rate": 8.231989314071318e-06,
|
| 932 |
-
"loss": 0.
|
| 933 |
"step": 655
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.6055045871559633,
|
| 937 |
-
"grad_norm": 1.
|
| 938 |
"learning_rate": 8.07462320849313e-06,
|
| 939 |
-
"loss": 0.
|
| 940 |
"step": 660
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 0.6100917431192661,
|
| 944 |
-
"grad_norm": 1.
|
| 945 |
"learning_rate": 7.917750740470116e-06,
|
| 946 |
-
"loss": 0.
|
| 947 |
"step": 665
|
| 948 |
},
|
| 949 |
{
|
| 950 |
"epoch": 0.6146788990825688,
|
| 951 |
-
"grad_norm": 1.
|
| 952 |
"learning_rate": 7.761412129735853e-06,
|
| 953 |
-
"loss": 0.
|
| 954 |
"step": 670
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.6192660550458715,
|
| 958 |
-
"grad_norm": 1.
|
| 959 |
"learning_rate": 7.605647459150961e-06,
|
| 960 |
-
"loss": 0.
|
| 961 |
"step": 675
|
| 962 |
},
|
| 963 |
{
|
| 964 |
"epoch": 0.6238532110091743,
|
| 965 |
-
"grad_norm": 1.
|
| 966 |
"learning_rate": 7.4504966644264775e-06,
|
| 967 |
-
"loss": 0.
|
| 968 |
"step": 680
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 0.6284403669724771,
|
| 972 |
-
"grad_norm": 1.
|
| 973 |
"learning_rate": 7.295999523884921e-06,
|
| 974 |
"loss": 0.9603,
|
| 975 |
"step": 685
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.6330275229357798,
|
| 979 |
-
"grad_norm": 1.
|
| 980 |
"learning_rate": 7.142195648261747e-06,
|
| 981 |
-
"loss": 0.
|
| 982 |
"step": 690
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.6376146788990825,
|
| 986 |
-
"grad_norm": 1.
|
| 987 |
"learning_rate": 6.989124470549746e-06,
|
| 988 |
-
"loss": 0.
|
| 989 |
"step": 695
|
| 990 |
},
|
| 991 |
{
|
| 992 |
"epoch": 0.6422018348623854,
|
| 993 |
-
"grad_norm": 1.
|
| 994 |
"learning_rate": 6.83682523588902e-06,
|
| 995 |
-
"loss": 0.
|
| 996 |
"step": 700
|
| 997 |
},
|
| 998 |
{
|
| 999 |
"epoch": 0.6467889908256881,
|
| 1000 |
-
"grad_norm": 1.
|
| 1001 |
"learning_rate": 6.685336991505122e-06,
|
| 1002 |
-
"loss": 0.
|
| 1003 |
"step": 705
|
| 1004 |
},
|
| 1005 |
{
|
| 1006 |
"epoch": 0.6513761467889908,
|
| 1007 |
-
"grad_norm": 1.
|
| 1008 |
"learning_rate": 6.5346985766979384e-06,
|
| 1009 |
-
"loss": 0.
|
| 1010 |
"step": 710
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 0.6559633027522935,
|
| 1014 |
-
"grad_norm": 1.
|
| 1015 |
"learning_rate": 6.384948612883872e-06,
|
| 1016 |
-
"loss": 0.
|
| 1017 |
"step": 715
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.6605504587155964,
|
| 1021 |
-
"grad_norm": 1.
|
| 1022 |
"learning_rate": 6.2361254936939e-06,
|
| 1023 |
-
"loss": 0.
|
| 1024 |
"step": 720
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
"epoch": 0.6651376146788991,
|
| 1028 |
-
"grad_norm": 1.
|
| 1029 |
"learning_rate": 6.0882673751300235e-06,
|
| 1030 |
-
"loss": 0.
|
| 1031 |
"step": 725
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
"epoch": 0.6697247706422018,
|
| 1035 |
-
"grad_norm": 1.
|
| 1036 |
"learning_rate": 5.941412165782645e-06,
|
| 1037 |
-
"loss": 0.
|
| 1038 |
"step": 730
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
"epoch": 0.6743119266055045,
|
| 1042 |
-
"grad_norm": 1.
|
| 1043 |
"learning_rate": 5.79559751711138e-06,
|
| 1044 |
-
"loss": 0.
|
| 1045 |
"step": 735
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
"epoch": 0.6788990825688074,
|
| 1049 |
-
"grad_norm": 1.
|
| 1050 |
"learning_rate": 5.650860813791786e-06,
|
| 1051 |
-
"loss": 0.
|
| 1052 |
"step": 740
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 0.6834862385321101,
|
| 1056 |
-
"grad_norm": 1.
|
| 1057 |
"learning_rate": 5.507239164130501e-06,
|
| 1058 |
-
"loss": 0.
|
| 1059 |
"step": 745
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.6880733944954128,
|
| 1063 |
-
"grad_norm": 1.
|
| 1064 |
"learning_rate": 5.364769390551225e-06,
|
| 1065 |
-
"loss": 0.
|
| 1066 |
"step": 750
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 0.6926605504587156,
|
| 1070 |
-
"grad_norm": 1.
|
| 1071 |
"learning_rate": 5.223488020154028e-06,
|
| 1072 |
-
"loss": 0.
|
| 1073 |
"step": 755
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 0.6972477064220184,
|
| 1077 |
-
"grad_norm": 1.
|
| 1078 |
"learning_rate": 5.083431275350312e-06,
|
| 1079 |
-
"loss": 0.
|
| 1080 |
"step": 760
|
| 1081 |
},
|
| 1082 |
{
|
| 1083 |
"epoch": 0.7018348623853211,
|
| 1084 |
-
"grad_norm": 1.
|
| 1085 |
"learning_rate": 4.9446350645759885e-06,
|
| 1086 |
-
"loss": 0.
|
| 1087 |
"step": 765
|
| 1088 |
},
|
| 1089 |
{
|
| 1090 |
"epoch": 0.7064220183486238,
|
| 1091 |
-
"grad_norm": 1.
|
| 1092 |
"learning_rate": 4.807134973085036e-06,
|
| 1093 |
-
"loss": 0.
|
| 1094 |
"step": 770
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 0.7110091743119266,
|
| 1098 |
-
"grad_norm": 1.
|
| 1099 |
"learning_rate": 4.670966253826027e-06,
|
| 1100 |
-
"loss": 0.
|
| 1101 |
"step": 775
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.7155963302752294,
|
| 1105 |
-
"grad_norm": 1.
|
| 1106 |
"learning_rate": 4.53616381840377e-06,
|
| 1107 |
"loss": 0.9468,
|
| 1108 |
"step": 780
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.7201834862385321,
|
| 1112 |
-
"grad_norm": 1.
|
| 1113 |
"learning_rate": 4.402762228128531e-06,
|
| 1114 |
-
"loss": 0.
|
| 1115 |
"step": 785
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
"epoch": 0.7247706422018348,
|
| 1119 |
-
"grad_norm": 1.
|
| 1120 |
"learning_rate": 4.270795685155001e-06,
|
| 1121 |
-
"loss": 0.
|
| 1122 |
"step": 790
|
| 1123 |
},
|
| 1124 |
{
|
| 1125 |
"epoch": 0.7293577981651376,
|
| 1126 |
-
"grad_norm": 1.
|
| 1127 |
"learning_rate": 4.140298023713416e-06,
|
| 1128 |
-
"loss": 0.
|
| 1129 |
"step": 795
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.7339449541284404,
|
| 1133 |
-
"grad_norm": 1.
|
| 1134 |
"learning_rate": 4.0113027014349374e-06,
|
| 1135 |
-
"loss": 0.
|
| 1136 |
"step": 800
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 0.7385321100917431,
|
| 1140 |
-
"grad_norm": 1.
|
| 1141 |
"learning_rate": 3.883842790773647e-06,
|
| 1142 |
-
"loss": 0.
|
| 1143 |
"step": 805
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.7431192660550459,
|
| 1147 |
-
"grad_norm": 1.
|
| 1148 |
"learning_rate": 3.757950970527249e-06,
|
| 1149 |
-
"loss": 0.
|
| 1150 |
"step": 810
|
| 1151 |
},
|
| 1152 |
{
|
| 1153 |
"epoch": 0.7477064220183486,
|
| 1154 |
-
"grad_norm": 1.
|
| 1155 |
"learning_rate": 3.633659517458736e-06,
|
| 1156 |
-
"loss": 0.
|
| 1157 |
"step": 815
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
"epoch": 0.7522935779816514,
|
| 1161 |
-
"grad_norm": 1.
|
| 1162 |
"learning_rate": 3.511000298021098e-06,
|
| 1163 |
-
"loss": 0.
|
| 1164 |
"step": 820
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.7568807339449541,
|
| 1168 |
-
"grad_norm": 1.
|
| 1169 |
"learning_rate": 3.39000476018726e-06,
|
| 1170 |
-
"loss": 0.
|
| 1171 |
"step": 825
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
"epoch": 0.7614678899082569,
|
| 1175 |
-
"grad_norm": 1.
|
| 1176 |
"learning_rate": 3.2707039253872796e-06,
|
| 1177 |
-
"loss": 0.
|
| 1178 |
"step": 830
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 0.7660550458715596,
|
| 1182 |
-
"grad_norm": 1.
|
| 1183 |
"learning_rate": 3.153128380554941e-06,
|
| 1184 |
-
"loss": 0.
|
| 1185 |
"step": 835
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.7706422018348624,
|
| 1189 |
-
"grad_norm": 1.
|
| 1190 |
"learning_rate": 3.037308270285709e-06,
|
| 1191 |
-
"loss": 0.
|
| 1192 |
"step": 840
|
| 1193 |
},
|
| 1194 |
{
|
| 1195 |
"epoch": 0.7752293577981652,
|
| 1196 |
-
"grad_norm": 1.
|
| 1197 |
"learning_rate": 2.923273289108115e-06,
|
| 1198 |
-
"loss": 0.
|
| 1199 |
"step": 845
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
"epoch": 0.7798165137614679,
|
| 1203 |
-
"grad_norm": 1.
|
| 1204 |
"learning_rate": 2.8110526738705345e-06,
|
| 1205 |
-
"loss": 0.
|
| 1206 |
"step": 850
|
| 1207 |
},
|
| 1208 |
{
|
| 1209 |
"epoch": 0.7844036697247706,
|
| 1210 |
-
"grad_norm": 1.
|
| 1211 |
"learning_rate": 2.700675196245288e-06,
|
| 1212 |
-
"loss": 0.
|
| 1213 |
"step": 855
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
"epoch": 0.7889908256880734,
|
| 1217 |
-
"grad_norm": 1.
|
| 1218 |
"learning_rate": 2.592169155352031e-06,
|
| 1219 |
-
"loss": 0.
|
| 1220 |
"step": 860
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 0.7935779816513762,
|
| 1224 |
-
"grad_norm": 1.
|
| 1225 |
"learning_rate": 2.485562370502279e-06,
|
| 1226 |
-
"loss": 0.
|
| 1227 |
"step": 865
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.7981651376146789,
|
| 1231 |
-
"grad_norm": 1.
|
| 1232 |
"learning_rate": 2.3808821740669608e-06,
|
| 1233 |
"loss": 0.9352,
|
| 1234 |
"step": 870
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8027522935779816,
|
| 1238 |
-
"grad_norm": 1.
|
| 1239 |
"learning_rate": 2.2781554044688015e-06,
|
| 1240 |
"loss": 0.9379,
|
| 1241 |
"step": 875
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.8073394495412844,
|
| 1245 |
-
"grad_norm": 1.
|
| 1246 |
"learning_rate": 2.1774083993013715e-06,
|
| 1247 |
-
"loss": 0.
|
| 1248 |
"step": 880
|
| 1249 |
},
|
| 1250 |
{
|
| 1251 |
"epoch": 0.8119266055045872,
|
| 1252 |
-
"grad_norm": 1.
|
| 1253 |
"learning_rate": 2.0786669885765044e-06,
|
| 1254 |
-
"loss": 0.
|
| 1255 |
"step": 885
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 0.8165137614678899,
|
| 1259 |
-
"grad_norm": 1.
|
| 1260 |
"learning_rate": 1.981956488101898e-06,
|
| 1261 |
-
"loss": 0.
|
| 1262 |
"step": 890
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 0.8211009174311926,
|
| 1266 |
-
"grad_norm": 1.
|
| 1267 |
"learning_rate": 1.8873016929904942e-06,
|
| 1268 |
-
"loss": 0.
|
| 1269 |
"step": 895
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.8256880733944955,
|
| 1273 |
-
"grad_norm": 1.
|
| 1274 |
"learning_rate": 1.7947268713034128e-06,
|
| 1275 |
-
"loss": 0.
|
| 1276 |
"step": 900
|
| 1277 |
},
|
| 1278 |
{
|
| 1279 |
"epoch": 0.8302752293577982,
|
| 1280 |
-
"grad_norm": 1.
|
| 1281 |
"learning_rate": 1.704255757827963e-06,
|
| 1282 |
-
"loss": 0.
|
| 1283 |
"step": 905
|
| 1284 |
},
|
| 1285 |
{
|
| 1286 |
"epoch": 0.8348623853211009,
|
| 1287 |
-
"grad_norm": 1.
|
| 1288 |
"learning_rate": 1.6159115479924259e-06,
|
| 1289 |
-
"loss": 0.
|
| 1290 |
"step": 910
|
| 1291 |
},
|
| 1292 |
{
|
| 1293 |
"epoch": 0.8394495412844036,
|
| 1294 |
-
"grad_norm": 1.
|
| 1295 |
"learning_rate": 1.529716891919074e-06,
|
| 1296 |
-
"loss": 0.
|
| 1297 |
"step": 915
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 0.8440366972477065,
|
| 1301 |
-
"grad_norm": 1.
|
| 1302 |
"learning_rate": 1.4456938886170413e-06,
|
| 1303 |
-
"loss": 0.
|
| 1304 |
"step": 920
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 0.8486238532110092,
|
| 1308 |
-
"grad_norm": 1.
|
| 1309 |
"learning_rate": 1.3638640803164516e-06,
|
| 1310 |
-
"loss": 0.
|
| 1311 |
"step": 925
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.8532110091743119,
|
| 1315 |
-
"grad_norm": 1.
|
| 1316 |
"learning_rate": 1.2842484469453365e-06,
|
| 1317 |
-
"loss": 0.
|
| 1318 |
"step": 930
|
| 1319 |
},
|
| 1320 |
{
|
| 1321 |
"epoch": 0.8577981651376146,
|
| 1322 |
-
"grad_norm": 1.
|
| 1323 |
"learning_rate": 1.2068674007506787e-06,
|
| 1324 |
-
"loss": 0.
|
| 1325 |
"step": 935
|
| 1326 |
},
|
| 1327 |
{
|
| 1328 |
"epoch": 0.8623853211009175,
|
| 1329 |
-
"grad_norm": 1.
|
| 1330 |
"learning_rate": 1.1317407810650372e-06,
|
| 1331 |
-
"loss": 0.
|
| 1332 |
"step": 940
|
| 1333 |
},
|
| 1334 |
{
|
| 1335 |
"epoch": 0.8669724770642202,
|
| 1336 |
-
"grad_norm": 1.
|
| 1337 |
"learning_rate": 1.0588878492200261e-06,
|
| 1338 |
-
"loss": 0.
|
| 1339 |
"step": 945
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
"epoch": 0.8715596330275229,
|
| 1343 |
-
"grad_norm": 1.
|
| 1344 |
"learning_rate": 9.883272836080116e-07,
|
| 1345 |
-
"loss": 0.
|
| 1346 |
"step": 950
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 0.8761467889908257,
|
| 1350 |
-
"grad_norm": 1.
|
| 1351 |
"learning_rate": 9.200771748932513e-07,
|
| 1352 |
-
"loss": 0.
|
| 1353 |
"step": 955
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.8807339449541285,
|
| 1357 |
-
"grad_norm": 1.
|
| 1358 |
"learning_rate": 8.541550213737171e-07,
|
| 1359 |
-
"loss": 0.
|
| 1360 |
"step": 960
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.8853211009174312,
|
| 1364 |
-
"grad_norm": 1.
|
| 1365 |
"learning_rate": 7.905777244947954e-07,
|
| 1366 |
-
"loss": 0.
|
| 1367 |
"step": 965
|
| 1368 |
},
|
| 1369 |
{
|
| 1370 |
"epoch": 0.8899082568807339,
|
| 1371 |
-
"grad_norm": 1.
|
| 1372 |
"learning_rate": 7.293615845160196e-07,
|
| 1373 |
-
"loss": 0.
|
| 1374 |
"step": 970
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.8944954128440367,
|
| 1378 |
-
"grad_norm": 1.
|
| 1379 |
"learning_rate": 6.705222963319191e-07,
|
| 1380 |
-
"loss": 0.
|
| 1381 |
"step": 975
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 0.8990825688073395,
|
| 1385 |
-
"grad_norm": 1.
|
| 1386 |
"learning_rate": 6.140749454480932e-07,
|
| 1387 |
-
"loss": 0.
|
| 1388 |
"step": 980
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 0.9036697247706422,
|
| 1392 |
-
"grad_norm": 1.
|
| 1393 |
"learning_rate": 5.600340041135133e-07,
|
| 1394 |
-
"loss": 0.
|
| 1395 |
"step": 985
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.908256880733945,
|
| 1399 |
-
"grad_norm": 1.
|
| 1400 |
"learning_rate": 5.0841332761005e-07,
|
| 1401 |
-
"loss": 0.
|
| 1402 |
"step": 990
|
| 1403 |
},
|
| 1404 |
{
|
| 1405 |
"epoch": 0.9128440366972477,
|
| 1406 |
-
"grad_norm": 1.
|
| 1407 |
"learning_rate": 4.592261507001994e-07,
|
| 1408 |
-
"loss": 0.
|
| 1409 |
"step": 995
|
| 1410 |
},
|
| 1411 |
{
|
| 1412 |
"epoch": 0.9174311926605505,
|
| 1413 |
-
"grad_norm": 1.
|
| 1414 |
"learning_rate": 4.124850842338779e-07,
|
| 1415 |
-
"loss": 0.
|
| 1416 |
"step": 1000
|
| 1417 |
},
|
| 1418 |
{
|
| 1419 |
"epoch": 0.9220183486238532,
|
| 1420 |
-
"grad_norm": 1.
|
| 1421 |
"learning_rate": 3.6820211191520127e-07,
|
| 1422 |
-
"loss": 0.
|
| 1423 |
"step": 1005
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 0.926605504587156,
|
| 1427 |
-
"grad_norm": 1.
|
| 1428 |
"learning_rate": 3.263885872300343e-07,
|
| 1429 |
-
"loss": 0.
|
| 1430 |
"step": 1010
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 0.9311926605504587,
|
| 1434 |
-
"grad_norm": 1.
|
| 1435 |
"learning_rate": 2.870552305351382e-07,
|
| 1436 |
-
"loss": 0.
|
| 1437 |
"step": 1015
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.9357798165137615,
|
| 1441 |
-
"grad_norm": 1.
|
| 1442 |
"learning_rate": 2.5021212630962246e-07,
|
| 1443 |
-
"loss": 0.
|
| 1444 |
"step": 1020
|
| 1445 |
},
|
| 1446 |
{
|
| 1447 |
"epoch": 0.9403669724770642,
|
| 1448 |
-
"grad_norm": 1.
|
| 1449 |
"learning_rate": 2.158687205694443e-07,
|
| 1450 |
-
"loss": 0.
|
| 1451 |
"step": 1025
|
| 1452 |
},
|
| 1453 |
{
|
| 1454 |
"epoch": 0.944954128440367,
|
| 1455 |
-
"grad_norm": 1.
|
| 1456 |
"learning_rate": 1.840338184455881e-07,
|
| 1457 |
-
"loss": 0.
|
| 1458 |
"step": 1030
|
| 1459 |
},
|
| 1460 |
{
|
| 1461 |
"epoch": 0.9495412844036697,
|
| 1462 |
-
"grad_norm": 1.
|
| 1463 |
"learning_rate": 1.5471558192656776e-07,
|
| 1464 |
-
"loss": 0.
|
| 1465 |
"step": 1035
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.9541284403669725,
|
| 1469 |
-
"grad_norm": 1.
|
| 1470 |
"learning_rate": 1.279215277658097e-07,
|
| 1471 |
-
"loss": 0.
|
| 1472 |
"step": 1040
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 0.9587155963302753,
|
| 1476 |
-
"grad_norm": 1.
|
| 1477 |
"learning_rate": 1.0365852555447642e-07,
|
| 1478 |
-
"loss": 0.
|
| 1479 |
"step": 1045
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.963302752293578,
|
| 1483 |
-
"grad_norm": 1.
|
| 1484 |
"learning_rate": 8.19327959602012e-08,
|
| 1485 |
-
"loss": 0.
|
| 1486 |
"step": 1050
|
| 1487 |
},
|
| 1488 |
{
|
| 1489 |
"epoch": 0.9678899082568807,
|
| 1490 |
-
"grad_norm": 1.
|
| 1491 |
"learning_rate": 6.274990913221035e-08,
|
| 1492 |
-
"loss": 0.
|
| 1493 |
"step": 1055
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"epoch": 0.9724770642201835,
|
| 1497 |
-
"grad_norm": 1.
|
| 1498 |
"learning_rate": 4.6114783273213395e-08,
|
| 1499 |
-
"loss": 0.
|
| 1500 |
"step": 1060
|
| 1501 |
},
|
| 1502 |
{
|
| 1503 |
"epoch": 0.9770642201834863,
|
| 1504 |
-
"grad_norm": 1.
|
| 1505 |
"learning_rate": 3.203168337845508e-08,
|
| 1506 |
-
"loss": 0.
|
| 1507 |
"step": 1065
|
| 1508 |
},
|
| 1509 |
{
|
| 1510 |
"epoch": 0.981651376146789,
|
| 1511 |
-
"grad_norm": 1.
|
| 1512 |
"learning_rate": 2.05042201422323e-08,
|
| 1513 |
-
"loss": 0.
|
| 1514 |
"step": 1070
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 0.9862385321100917,
|
| 1518 |
-
"grad_norm": 1.
|
| 1519 |
"learning_rate": 1.1535349032167908e-08,
|
| 1520 |
-
"loss": 0.
|
| 1521 |
"step": 1075
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
"epoch": 0.9908256880733946,
|
| 1525 |
-
"grad_norm": 1.
|
| 1526 |
"learning_rate": 5.127369531473525e-09,
|
| 1527 |
-
"loss": 0.
|
| 1528 |
"step": 1080
|
| 1529 |
},
|
| 1530 |
{
|
| 1531 |
"epoch": 0.9954128440366973,
|
| 1532 |
-
"grad_norm": 1.
|
| 1533 |
"learning_rate": 1.2819245493955746e-09,
|
| 1534 |
-
"loss": 0.
|
| 1535 |
"step": 1085
|
| 1536 |
},
|
| 1537 |
{
|
| 1538 |
"epoch": 1.0,
|
| 1539 |
-
"grad_norm": 1.
|
| 1540 |
"learning_rate": 0.0,
|
| 1541 |
"loss": 0.9203,
|
| 1542 |
"step": 1090
|
| 1543 |
},
|
| 1544 |
{
|
| 1545 |
"epoch": 1.0,
|
| 1546 |
-
"eval_loss": 0.
|
| 1547 |
-
"eval_runtime":
|
| 1548 |
-
"eval_samples_per_second":
|
| 1549 |
-
"eval_steps_per_second": 0.
|
| 1550 |
"step": 1090
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
"epoch": 1.0,
|
| 1554 |
"step": 1090,
|
| 1555 |
"total_flos": 456447649382400.0,
|
| 1556 |
-
"train_loss": 0.
|
| 1557 |
-
"train_runtime":
|
| 1558 |
-
"train_samples_per_second":
|
| 1559 |
-
"train_steps_per_second": 0.
|
| 1560 |
}
|
| 1561 |
],
|
| 1562 |
"logging_steps": 5,
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0009174311926605505,
|
| 13 |
+
"grad_norm": 8.567075219673832,
|
| 14 |
"learning_rate": 1.8348623853211012e-07,
|
| 15 |
"loss": 1.1366,
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.0045871559633027525,
|
| 20 |
+
"grad_norm": 7.502231595297741,
|
| 21 |
"learning_rate": 9.174311926605506e-07,
|
| 22 |
"loss": 1.1311,
|
| 23 |
"step": 5
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.009174311926605505,
|
| 27 |
+
"grad_norm": 4.041886291445951,
|
| 28 |
"learning_rate": 1.8348623853211011e-06,
|
| 29 |
"loss": 1.0455,
|
| 30 |
"step": 10
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.013761467889908258,
|
| 34 |
+
"grad_norm": 2.66931200242608,
|
| 35 |
"learning_rate": 2.7522935779816517e-06,
|
| 36 |
+
"loss": 1.0276,
|
| 37 |
"step": 15
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.01834862385321101,
|
| 41 |
+
"grad_norm": 2.655264853297466,
|
| 42 |
"learning_rate": 3.6697247706422022e-06,
|
| 43 |
+
"loss": 0.996,
|
| 44 |
"step": 20
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.022935779816513763,
|
| 48 |
+
"grad_norm": 2.0824902429417502,
|
| 49 |
"learning_rate": 4.587155963302753e-06,
|
| 50 |
+
"loss": 0.9943,
|
| 51 |
"step": 25
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.027522935779816515,
|
| 55 |
+
"grad_norm": 2.223113856453528,
|
| 56 |
"learning_rate": 5.504587155963303e-06,
|
| 57 |
+
"loss": 0.9914,
|
| 58 |
"step": 30
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.03211009174311927,
|
| 62 |
+
"grad_norm": 2.8245029253119083,
|
| 63 |
"learning_rate": 6.422018348623854e-06,
|
| 64 |
+
"loss": 0.9877,
|
| 65 |
"step": 35
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.03669724770642202,
|
| 69 |
+
"grad_norm": 1.9055889003234818,
|
| 70 |
"learning_rate": 7.3394495412844045e-06,
|
| 71 |
+
"loss": 0.9948,
|
| 72 |
"step": 40
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.04128440366972477,
|
| 76 |
+
"grad_norm": 2.2824223465177687,
|
| 77 |
"learning_rate": 8.256880733944956e-06,
|
| 78 |
+
"loss": 0.9779,
|
| 79 |
"step": 45
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.045871559633027525,
|
| 83 |
+
"grad_norm": 2.238766365680375,
|
| 84 |
"learning_rate": 9.174311926605506e-06,
|
| 85 |
+
"loss": 0.9842,
|
| 86 |
"step": 50
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.05045871559633028,
|
| 90 |
+
"grad_norm": 2.3168005986367777,
|
| 91 |
"learning_rate": 1.0091743119266055e-05,
|
| 92 |
+
"loss": 0.9935,
|
| 93 |
"step": 55
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.05504587155963303,
|
| 97 |
+
"grad_norm": 2.0671072791875775,
|
| 98 |
"learning_rate": 1.1009174311926607e-05,
|
| 99 |
+
"loss": 0.9973,
|
| 100 |
"step": 60
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.05963302752293578,
|
| 104 |
+
"grad_norm": 2.3870937792624987,
|
| 105 |
"learning_rate": 1.1926605504587156e-05,
|
| 106 |
+
"loss": 0.9812,
|
| 107 |
"step": 65
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.06422018348623854,
|
| 111 |
+
"grad_norm": 2.872503228286113,
|
| 112 |
"learning_rate": 1.2844036697247708e-05,
|
| 113 |
+
"loss": 0.9866,
|
| 114 |
"step": 70
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.06880733944954129,
|
| 118 |
+
"grad_norm": 2.918010569550636,
|
| 119 |
"learning_rate": 1.3761467889908258e-05,
|
| 120 |
"loss": 1.0001,
|
| 121 |
"step": 75
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.07339449541284404,
|
| 125 |
+
"grad_norm": 3.0830710791214773,
|
| 126 |
"learning_rate": 1.4678899082568809e-05,
|
| 127 |
+
"loss": 0.9886,
|
| 128 |
"step": 80
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.0779816513761468,
|
| 132 |
+
"grad_norm": 2.6063312280690574,
|
| 133 |
"learning_rate": 1.559633027522936e-05,
|
| 134 |
+
"loss": 0.9927,
|
| 135 |
"step": 85
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.08256880733944955,
|
| 139 |
+
"grad_norm": 2.209724380985353,
|
| 140 |
"learning_rate": 1.6513761467889912e-05,
|
| 141 |
+
"loss": 1.0107,
|
| 142 |
"step": 90
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.0871559633027523,
|
| 146 |
+
"grad_norm": 2.76948096722416,
|
| 147 |
"learning_rate": 1.743119266055046e-05,
|
| 148 |
+
"loss": 1.0185,
|
| 149 |
"step": 95
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.09174311926605505,
|
| 153 |
+
"grad_norm": 2.1716697491458707,
|
| 154 |
"learning_rate": 1.834862385321101e-05,
|
| 155 |
+
"loss": 0.9981,
|
| 156 |
"step": 100
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.0963302752293578,
|
| 160 |
+
"grad_norm": 2.1757317389304482,
|
| 161 |
"learning_rate": 1.9266055045871563e-05,
|
| 162 |
+
"loss": 1.014,
|
| 163 |
"step": 105
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.10091743119266056,
|
| 167 |
+
"grad_norm": 2.5856692970053055,
|
| 168 |
"learning_rate": 1.999994872196626e-05,
|
| 169 |
+
"loss": 1.0231,
|
| 170 |
"step": 110
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.10550458715596331,
|
| 174 |
+
"grad_norm": 2.0176395767985897,
|
| 175 |
"learning_rate": 1.9998154046002822e-05,
|
| 176 |
+
"loss": 1.005,
|
| 177 |
"step": 115
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.11009174311926606,
|
| 181 |
+
"grad_norm": 1.9369629064989813,
|
| 182 |
"learning_rate": 1.999379599421534e-05,
|
| 183 |
+
"loss": 1.0051,
|
| 184 |
"step": 120
|
| 185 |
},
|
| 186 |
{
|
| 187 |
"epoch": 0.11467889908256881,
|
| 188 |
+
"grad_norm": 1.9803383518176758,
|
| 189 |
"learning_rate": 1.9986875683942535e-05,
|
| 190 |
+
"loss": 1.0137,
|
| 191 |
"step": 125
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.11926605504587157,
|
| 195 |
+
"grad_norm": 1.9751164482121784,
|
| 196 |
"learning_rate": 1.9977394889447526e-05,
|
| 197 |
+
"loss": 1.0125,
|
| 198 |
"step": 130
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"epoch": 0.12385321100917432,
|
| 202 |
+
"grad_norm": 1.766026099400838,
|
| 203 |
"learning_rate": 1.9965356041462954e-05,
|
| 204 |
+
"loss": 0.993,
|
| 205 |
"step": 135
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.12844036697247707,
|
| 209 |
+
"grad_norm": 1.8456571124434127,
|
| 210 |
"learning_rate": 1.9950762226567783e-05,
|
| 211 |
+
"loss": 1.0208,
|
| 212 |
"step": 140
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.13302752293577982,
|
| 216 |
+
"grad_norm": 2.2558886400833136,
|
| 217 |
"learning_rate": 1.9933617186395917e-05,
|
| 218 |
+
"loss": 1.0273,
|
| 219 |
"step": 145
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.13761467889908258,
|
| 223 |
+
"grad_norm": 2.0067810742214367,
|
| 224 |
"learning_rate": 1.9913925316676946e-05,
|
| 225 |
+
"loss": 1.0214,
|
| 226 |
"step": 150
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.14220183486238533,
|
| 230 |
+
"grad_norm": 2.0354769519738696,
|
| 231 |
"learning_rate": 1.9891691666109112e-05,
|
| 232 |
+
"loss": 1.0449,
|
| 233 |
"step": 155
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.14678899082568808,
|
| 237 |
+
"grad_norm": 1.9853898034162918,
|
| 238 |
"learning_rate": 1.9866921935064907e-05,
|
| 239 |
"loss": 1.0348,
|
| 240 |
"step": 160
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 0.15137614678899083,
|
| 244 |
+
"grad_norm": 1.8771707161324125,
|
| 245 |
"learning_rate": 1.9839622474129595e-05,
|
| 246 |
+
"loss": 1.0447,
|
| 247 |
"step": 165
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.1559633027522936,
|
| 251 |
+
"grad_norm": 1.7555490053031533,
|
| 252 |
"learning_rate": 1.9809800282473014e-05,
|
| 253 |
+
"loss": 1.0319,
|
| 254 |
"step": 170
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.16055045871559634,
|
| 258 |
+
"grad_norm": 3.813249317869119,
|
| 259 |
"learning_rate": 1.977746300605507e-05,
|
| 260 |
+
"loss": 1.0432,
|
| 261 |
"step": 175
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.1651376146788991,
|
| 265 |
+
"grad_norm": 2.932842350036964,
|
| 266 |
"learning_rate": 1.9742618935665478e-05,
|
| 267 |
+
"loss": 1.0337,
|
| 268 |
"step": 180
|
| 269 |
},
|
| 270 |
{
|
| 271 |
"epoch": 0.16972477064220184,
|
| 272 |
+
"grad_norm": 1.9490193595251735,
|
| 273 |
"learning_rate": 1.9705277004798072e-05,
|
| 274 |
+
"loss": 1.0354,
|
| 275 |
"step": 185
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.1743119266055046,
|
| 279 |
+
"grad_norm": 2.0315194096887073,
|
| 280 |
"learning_rate": 1.9665446787360444e-05,
|
| 281 |
+
"loss": 1.0381,
|
| 282 |
"step": 190
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 0.17889908256880735,
|
| 286 |
+
"grad_norm": 2.01041642690734,
|
| 287 |
"learning_rate": 1.9623138495219292e-05,
|
| 288 |
+
"loss": 1.0176,
|
| 289 |
"step": 195
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.1834862385321101,
|
| 293 |
+
"grad_norm": 1.8379234276601883,
|
| 294 |
"learning_rate": 1.957836297558229e-05,
|
| 295 |
+
"loss": 1.0179,
|
| 296 |
"step": 200
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.18807339449541285,
|
| 300 |
+
"grad_norm": 1.8383111439189317,
|
| 301 |
"learning_rate": 1.9531131708217005e-05,
|
| 302 |
+
"loss": 1.023,
|
| 303 |
"step": 205
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.1926605504587156,
|
| 307 |
+
"grad_norm": 1.7211278520137383,
|
| 308 |
"learning_rate": 1.948145680250766e-05,
|
| 309 |
+
"loss": 1.0201,
|
| 310 |
"step": 210
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 0.19724770642201836,
|
| 314 |
+
"grad_norm": 1.6488508269613613,
|
| 315 |
"learning_rate": 1.9429350994350483e-05,
|
| 316 |
+
"loss": 1.0337,
|
| 317 |
"step": 215
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 0.2018348623853211,
|
| 321 |
+
"grad_norm": 1.6195453176201335,
|
| 322 |
"learning_rate": 1.93748276428884e-05,
|
| 323 |
+
"loss": 1.041,
|
| 324 |
"step": 220
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.20642201834862386,
|
| 328 |
+
"grad_norm": 1.6631810770290123,
|
| 329 |
"learning_rate": 1.931790072708596e-05,
|
| 330 |
+
"loss": 1.0253,
|
| 331 |
"step": 225
|
| 332 |
},
|
| 333 |
{
|
| 334 |
"epoch": 0.21100917431192662,
|
| 335 |
+
"grad_norm": 1.682342166301177,
|
| 336 |
"learning_rate": 1.9258584842145342e-05,
|
| 337 |
+
"loss": 1.0446,
|
| 338 |
"step": 230
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.21559633027522937,
|
| 342 |
+
"grad_norm": 1.9113148172207652,
|
| 343 |
"learning_rate": 1.9196895195764363e-05,
|
| 344 |
+
"loss": 1.0249,
|
| 345 |
"step": 235
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22018348623853212,
|
| 349 |
+
"grad_norm": 1.921572422455064,
|
| 350 |
"learning_rate": 1.913284760423745e-05,
|
| 351 |
+
"loss": 1.0289,
|
| 352 |
"step": 240
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.22477064220183487,
|
| 356 |
+
"grad_norm": 1.6160022720196747,
|
| 357 |
"learning_rate": 1.9066458488400586e-05,
|
| 358 |
+
"loss": 1.0281,
|
| 359 |
"step": 245
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 0.22935779816513763,
|
| 363 |
+
"grad_norm": 1.546114425823443,
|
| 364 |
"learning_rate": 1.8997744869421248e-05,
|
| 365 |
+
"loss": 1.0237,
|
| 366 |
"step": 250
|
| 367 |
},
|
| 368 |
{
|
| 369 |
"epoch": 0.23394495412844038,
|
| 370 |
+
"grad_norm": 1.6113177148502253,
|
| 371 |
"learning_rate": 1.8926724364434447e-05,
|
| 372 |
+
"loss": 1.0159,
|
| 373 |
"step": 255
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 0.23853211009174313,
|
| 377 |
+
"grad_norm": 1.756084810660972,
|
| 378 |
"learning_rate": 1.8853415182025953e-05,
|
| 379 |
+
"loss": 1.03,
|
| 380 |
"step": 260
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.24311926605504589,
|
| 384 |
+
"grad_norm": 1.7722580062413746,
|
| 385 |
"learning_rate": 1.8777836117563894e-05,
|
| 386 |
+
"loss": 1.0268,
|
| 387 |
"step": 265
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.24770642201834864,
|
| 391 |
+
"grad_norm": 1.6552997587829843,
|
| 392 |
"learning_rate": 1.8700006548379898e-05,
|
| 393 |
+
"loss": 1.0214,
|
| 394 |
"step": 270
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 0.25229357798165136,
|
| 398 |
+
"grad_norm": 1.7554082828396729,
|
| 399 |
"learning_rate": 1.861994642880105e-05,
|
| 400 |
+
"loss": 1.0411,
|
| 401 |
"step": 275
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 0.25688073394495414,
|
| 405 |
+
"grad_norm": 1.6101850875118928,
|
| 406 |
"learning_rate": 1.8537676285033886e-05,
|
| 407 |
+
"loss": 1.0289,
|
| 408 |
"step": 280
|
| 409 |
},
|
| 410 |
{
|
| 411 |
"epoch": 0.26146788990825687,
|
| 412 |
+
"grad_norm": 1.6754749206063948,
|
| 413 |
"learning_rate": 1.845321720990181e-05,
|
| 414 |
+
"loss": 1.03,
|
| 415 |
"step": 285
|
| 416 |
},
|
| 417 |
{
|
| 418 |
"epoch": 0.26605504587155965,
|
| 419 |
+
"grad_norm": 1.5989084659148887,
|
| 420 |
"learning_rate": 1.8366590857437182e-05,
|
| 421 |
+
"loss": 1.0332,
|
| 422 |
"step": 290
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.2706422018348624,
|
| 426 |
+
"grad_norm": 1.7373362181608558,
|
| 427 |
"learning_rate": 1.8277819437329577e-05,
|
| 428 |
+
"loss": 1.0196,
|
| 429 |
"step": 295
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.27522935779816515,
|
| 433 |
+
"grad_norm": 1.6415478980043552,
|
| 434 |
"learning_rate": 1.8186925709231534e-05,
|
| 435 |
+
"loss": 1.0331,
|
| 436 |
"step": 300
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.2798165137614679,
|
| 440 |
+
"grad_norm": 1.595584297931999,
|
| 441 |
"learning_rate": 1.809393297692334e-05,
|
| 442 |
+
"loss": 1.0257,
|
| 443 |
"step": 305
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 0.28440366972477066,
|
| 447 |
+
"grad_norm": 1.8203998942508204,
|
| 448 |
"learning_rate": 1.799886508233829e-05,
|
| 449 |
+
"loss": 1.044,
|
| 450 |
"step": 310
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"epoch": 0.2889908256880734,
|
| 454 |
+
"grad_norm": 1.6518600023400687,
|
| 455 |
"learning_rate": 1.790174639944997e-05,
|
| 456 |
+
"loss": 1.0387,
|
| 457 |
"step": 315
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.29357798165137616,
|
| 461 |
+
"grad_norm": 1.7796790607246211,
|
| 462 |
"learning_rate": 1.780260182802314e-05,
|
| 463 |
+
"loss": 1.0261,
|
| 464 |
"step": 320
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.2981651376146789,
|
| 468 |
+
"grad_norm": 1.6534133312284494,
|
| 469 |
"learning_rate": 1.7701456787229805e-05,
|
| 470 |
+
"loss": 1.0338,
|
| 471 |
"step": 325
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.30275229357798167,
|
| 475 |
+
"grad_norm": 1.5254168104289345,
|
| 476 |
"learning_rate": 1.7598337209132142e-05,
|
| 477 |
+
"loss": 1.0155,
|
| 478 |
"step": 330
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.3073394495412844,
|
| 482 |
+
"grad_norm": 1.5916672455155843,
|
| 483 |
"learning_rate": 1.7493269532033882e-05,
|
| 484 |
+
"loss": 1.0181,
|
| 485 |
"step": 335
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 0.3119266055045872,
|
| 489 |
+
"grad_norm": 2.469796014086296,
|
| 490 |
"learning_rate": 1.738628069370195e-05,
|
| 491 |
+
"loss": 1.0291,
|
| 492 |
"step": 340
|
| 493 |
},
|
| 494 |
{
|
| 495 |
"epoch": 0.3165137614678899,
|
| 496 |
+
"grad_norm": 1.7002430789318252,
|
| 497 |
"learning_rate": 1.7277398124460022e-05,
|
| 498 |
+
"loss": 1.0261,
|
| 499 |
"step": 345
|
| 500 |
},
|
| 501 |
{
|
| 502 |
"epoch": 0.3211009174311927,
|
| 503 |
+
"grad_norm": 1.6452789420112204,
|
| 504 |
"learning_rate": 1.71666497401558e-05,
|
| 505 |
+
"loss": 0.9936,
|
| 506 |
"step": 350
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.3256880733944954,
|
| 510 |
+
"grad_norm": 1.6090081655393613,
|
| 511 |
"learning_rate": 1.7054063935003813e-05,
|
| 512 |
"loss": 1.0253,
|
| 513 |
"step": 355
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.3302752293577982,
|
| 517 |
+
"grad_norm": 1.6521899111347487,
|
| 518 |
"learning_rate": 1.6939669574305565e-05,
|
| 519 |
+
"loss": 1.0076,
|
| 520 |
"step": 360
|
| 521 |
},
|
| 522 |
{
|
| 523 |
"epoch": 0.3348623853211009,
|
| 524 |
+
"grad_norm": 1.5087732073233964,
|
| 525 |
"learning_rate": 1.6823495987048922e-05,
|
| 526 |
+
"loss": 1.0098,
|
| 527 |
"step": 365
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 0.3394495412844037,
|
| 531 |
+
"grad_norm": 1.4806406906585232,
|
| 532 |
"learning_rate": 1.6705572958388576e-05,
|
| 533 |
+
"loss": 1.01,
|
| 534 |
"step": 370
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.3440366972477064,
|
| 538 |
+
"grad_norm": 1.6035884549602302,
|
| 539 |
"learning_rate": 1.6585930722009602e-05,
|
| 540 |
+
"loss": 1.0335,
|
| 541 |
"step": 375
|
| 542 |
},
|
| 543 |
{
|
| 544 |
"epoch": 0.3486238532110092,
|
| 545 |
+
"grad_norm": 1.5314286198965614,
|
| 546 |
"learning_rate": 1.6464599952375998e-05,
|
| 547 |
+
"loss": 1.0227,
|
| 548 |
"step": 380
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.3532110091743119,
|
| 552 |
+
"grad_norm": 1.630708086376815,
|
| 553 |
"learning_rate": 1.63416117568662e-05,
|
| 554 |
+
"loss": 1.0186,
|
| 555 |
"step": 385
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.3577981651376147,
|
| 559 |
+
"grad_norm": 1.484014704538301,
|
| 560 |
"learning_rate": 1.621699766779763e-05,
|
| 561 |
+
"loss": 1.0141,
|
| 562 |
"step": 390
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 0.3623853211009174,
|
| 566 |
+
"grad_norm": 1.489313766253633,
|
| 567 |
"learning_rate": 1.6090789634342278e-05,
|
| 568 |
+
"loss": 1.0049,
|
| 569 |
"step": 395
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.3669724770642202,
|
| 573 |
+
"grad_norm": 1.5152327147526468,
|
| 574 |
"learning_rate": 1.5963020014335437e-05,
|
| 575 |
+
"loss": 1.0105,
|
| 576 |
"step": 400
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"epoch": 0.37155963302752293,
|
| 580 |
+
"grad_norm": 1.4449410389905726,
|
| 581 |
"learning_rate": 1.583372156597961e-05,
|
| 582 |
+
"loss": 1.0031,
|
| 583 |
"step": 405
|
| 584 |
},
|
| 585 |
{
|
| 586 |
"epoch": 0.3761467889908257,
|
| 587 |
+
"grad_norm": 1.4506061967201511,
|
| 588 |
"learning_rate": 1.570292743944583e-05,
|
| 589 |
+
"loss": 1.002,
|
| 590 |
"step": 410
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.38073394495412843,
|
| 594 |
+
"grad_norm": 1.428745280078687,
|
| 595 |
"learning_rate": 1.557067116837444e-05,
|
| 596 |
+
"loss": 1.0223,
|
| 597 |
"step": 415
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.3853211009174312,
|
| 601 |
+
"grad_norm": 1.4567183232187022,
|
| 602 |
"learning_rate": 1.5436986661277578e-05,
|
| 603 |
+
"loss": 1.0133,
|
| 604 |
"step": 420
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.38990825688073394,
|
| 608 |
+
"grad_norm": 1.6991055564741833,
|
| 609 |
"learning_rate": 1.530190819284555e-05,
|
| 610 |
+
"loss": 1.0038,
|
| 611 |
"step": 425
|
| 612 |
},
|
| 613 |
{
|
| 614 |
"epoch": 0.3944954128440367,
|
| 615 |
+
"grad_norm": 1.5791854852416984,
|
| 616 |
"learning_rate": 1.5165470395159314e-05,
|
| 617 |
"loss": 1.0335,
|
| 618 |
"step": 430
|
| 619 |
},
|
| 620 |
{
|
| 621 |
"epoch": 0.39908256880733944,
|
| 622 |
+
"grad_norm": 1.6518586876063224,
|
| 623 |
"learning_rate": 1.5027708248811331e-05,
|
| 624 |
"loss": 1.0006,
|
| 625 |
"step": 435
|
| 626 |
},
|
| 627 |
{
|
| 628 |
"epoch": 0.4036697247706422,
|
| 629 |
+
"grad_norm": 1.37950132663011,
|
| 630 |
"learning_rate": 1.4888657073937077e-05,
|
| 631 |
+
"loss": 1.009,
|
| 632 |
"step": 440
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.40825688073394495,
|
| 636 |
+
"grad_norm": 1.4704035924378334,
|
| 637 |
"learning_rate": 1.4748352521159492e-05,
|
| 638 |
+
"loss": 0.9958,
|
| 639 |
"step": 445
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.41284403669724773,
|
| 643 |
+
"grad_norm": 1.4596704904359323,
|
| 644 |
"learning_rate": 1.4606830562448692e-05,
|
| 645 |
+
"loss": 1.005,
|
| 646 |
"step": 450
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 0.41743119266055045,
|
| 650 |
+
"grad_norm": 1.4454837075407818,
|
| 651 |
"learning_rate": 1.4464127481899312e-05,
|
| 652 |
+
"loss": 1.0073,
|
| 653 |
"step": 455
|
| 654 |
},
|
| 655 |
{
|
| 656 |
"epoch": 0.42201834862385323,
|
| 657 |
+
"grad_norm": 1.4689484755348563,
|
| 658 |
"learning_rate": 1.4320279866427798e-05,
|
| 659 |
+
"loss": 1.0064,
|
| 660 |
"step": 460
|
| 661 |
},
|
| 662 |
{
|
| 663 |
"epoch": 0.42660550458715596,
|
| 664 |
+
"grad_norm": 1.485956826048098,
|
| 665 |
"learning_rate": 1.4175324596392075e-05,
|
| 666 |
+
"loss": 1.0203,
|
| 667 |
"step": 465
|
| 668 |
},
|
| 669 |
{
|
| 670 |
"epoch": 0.43119266055045874,
|
| 671 |
+
"grad_norm": 1.573521513583627,
|
| 672 |
"learning_rate": 1.402929883613599e-05,
|
| 673 |
+
"loss": 1.0074,
|
| 674 |
"step": 470
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 0.43577981651376146,
|
| 678 |
+
"grad_norm": 1.4783455247528414,
|
| 679 |
"learning_rate": 1.3882240024460928e-05,
|
| 680 |
"loss": 0.9925,
|
| 681 |
"step": 475
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44036697247706424,
|
| 685 |
+
"grad_norm": 1.4099865921237311,
|
| 686 |
"learning_rate": 1.3734185865027061e-05,
|
| 687 |
+
"loss": 1.0105,
|
| 688 |
"step": 480
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 0.44495412844036697,
|
| 692 |
+
"grad_norm": 1.4586347859412025,
|
| 693 |
"learning_rate": 1.358517431668672e-05,
|
| 694 |
+
"loss": 0.9992,
|
| 695 |
"step": 485
|
| 696 |
},
|
| 697 |
{
|
| 698 |
"epoch": 0.44954128440366975,
|
| 699 |
+
"grad_norm": 1.3606557347465276,
|
| 700 |
"learning_rate": 1.3435243583752294e-05,
|
| 701 |
+
"loss": 0.9936,
|
| 702 |
"step": 490
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"epoch": 0.4541284403669725,
|
| 706 |
+
"grad_norm": 1.423923281806998,
|
| 707 |
"learning_rate": 1.3284432106201233e-05,
|
| 708 |
+
"loss": 0.9879,
|
| 709 |
"step": 495
|
| 710 |
},
|
| 711 |
{
|
| 712 |
"epoch": 0.45871559633027525,
|
| 713 |
+
"grad_norm": 1.5219817365897736,
|
| 714 |
"learning_rate": 1.313277854982062e-05,
|
| 715 |
+
"loss": 1.0079,
|
| 716 |
"step": 500
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 0.463302752293578,
|
| 720 |
+
"grad_norm": 1.6212946711670189,
|
| 721 |
"learning_rate": 1.2980321796293838e-05,
|
| 722 |
+
"loss": 0.987,
|
| 723 |
"step": 505
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.46788990825688076,
|
| 727 |
+
"grad_norm": 1.4932426551096094,
|
| 728 |
"learning_rate": 1.2827100933231904e-05,
|
| 729 |
+
"loss": 1.0052,
|
| 730 |
"step": 510
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.4724770642201835,
|
| 734 |
+
"grad_norm": 1.5771080120985452,
|
| 735 |
"learning_rate": 1.2673155244151985e-05,
|
| 736 |
+
"loss": 1.0052,
|
| 737 |
"step": 515
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 0.47706422018348627,
|
| 741 |
+
"grad_norm": 1.5047726515081563,
|
| 742 |
"learning_rate": 1.2518524198405699e-05,
|
| 743 |
+
"loss": 0.9992,
|
| 744 |
"step": 520
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.481651376146789,
|
| 748 |
+
"grad_norm": 1.5275724292638637,
|
| 749 |
"learning_rate": 1.2363247441059775e-05,
|
| 750 |
+
"loss": 0.9852,
|
| 751 |
"step": 525
|
| 752 |
},
|
| 753 |
{
|
| 754 |
"epoch": 0.48623853211009177,
|
| 755 |
+
"grad_norm": 1.4373779721892144,
|
| 756 |
"learning_rate": 1.2207364782731657e-05,
|
| 757 |
+
"loss": 1.0087,
|
| 758 |
"step": 530
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 0.4908256880733945,
|
| 762 |
+
"grad_norm": 1.4497248778716865,
|
| 763 |
"learning_rate": 1.2050916189382646e-05,
|
| 764 |
+
"loss": 0.9732,
|
| 765 |
"step": 535
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.4954128440366973,
|
| 769 |
+
"grad_norm": 1.4508227174593424,
|
| 770 |
"learning_rate": 1.189394177207125e-05,
|
| 771 |
+
"loss": 0.98,
|
| 772 |
"step": 540
|
| 773 |
},
|
| 774 |
{
|
| 775 |
"epoch": 0.5,
|
| 776 |
+
"grad_norm": 1.3723993793012572,
|
| 777 |
"learning_rate": 1.1736481776669307e-05,
|
| 778 |
+
"loss": 0.9873,
|
| 779 |
"step": 545
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 0.5045871559633027,
|
| 783 |
+
"grad_norm": 1.400736917075807,
|
| 784 |
"learning_rate": 1.1578576573543541e-05,
|
| 785 |
+
"loss": 0.964,
|
| 786 |
"step": 550
|
| 787 |
},
|
| 788 |
{
|
| 789 |
"epoch": 0.5091743119266054,
|
| 790 |
+
"grad_norm": 1.4328329143842218,
|
| 791 |
"learning_rate": 1.1420266647205232e-05,
|
| 792 |
"loss": 0.9744,
|
| 793 |
"step": 555
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.5137614678899083,
|
| 797 |
+
"grad_norm": 1.4005342497275977,
|
| 798 |
"learning_rate": 1.1261592585930576e-05,
|
| 799 |
+
"loss": 0.9845,
|
| 800 |
"step": 560
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 0.518348623853211,
|
| 804 |
+
"grad_norm": 1.3870143214943034,
|
| 805 |
"learning_rate": 1.1102595071354471e-05,
|
| 806 |
+
"loss": 0.9849,
|
| 807 |
"step": 565
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.5229357798165137,
|
| 811 |
+
"grad_norm": 1.4037256110851137,
|
| 812 |
"learning_rate": 1.0943314868040365e-05,
|
| 813 |
+
"loss": 0.978,
|
| 814 |
"step": 570
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 0.5275229357798165,
|
| 818 |
+
"grad_norm": 1.4524357204964615,
|
| 819 |
"learning_rate": 1.0783792813028828e-05,
|
| 820 |
"loss": 0.9883,
|
| 821 |
"step": 575
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 0.5321100917431193,
|
| 825 |
+
"grad_norm": 1.4168006943303706,
|
| 826 |
"learning_rate": 1.0624069805367558e-05,
|
| 827 |
+
"loss": 0.9694,
|
| 828 |
"step": 580
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 0.536697247706422,
|
| 832 |
+
"grad_norm": 1.3868495069798417,
|
| 833 |
"learning_rate": 1.0464186795625481e-05,
|
| 834 |
+
"loss": 0.9722,
|
| 835 |
"step": 585
|
| 836 |
},
|
| 837 |
{
|
| 838 |
"epoch": 0.5412844036697247,
|
| 839 |
+
"grad_norm": 1.484113669184668,
|
| 840 |
"learning_rate": 1.0304184775393642e-05,
|
| 841 |
+
"loss": 0.9858,
|
| 842 |
"step": 590
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 0.5458715596330275,
|
| 846 |
+
"grad_norm": 1.4378478721351533,
|
| 847 |
"learning_rate": 1.0144104766775574e-05,
|
| 848 |
+
"loss": 0.9688,
|
| 849 |
"step": 595
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.5504587155963303,
|
| 853 |
+
"grad_norm": 1.3497763152773319,
|
| 854 |
"learning_rate": 9.983987811869863e-06,
|
| 855 |
+
"loss": 0.9619,
|
| 856 |
"step": 600
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.555045871559633,
|
| 860 |
+
"grad_norm": 1.3638902119410679,
|
| 861 |
"learning_rate": 9.823874962247565e-06,
|
| 862 |
+
"loss": 0.9869,
|
| 863 |
"step": 605
|
| 864 |
},
|
| 865 |
{
|
| 866 |
"epoch": 0.5596330275229358,
|
| 867 |
+
"grad_norm": 1.4727242463849317,
|
| 868 |
"learning_rate": 9.663807268427197e-06,
|
| 869 |
+
"loss": 0.9497,
|
| 870 |
"step": 610
|
| 871 |
},
|
| 872 |
{
|
| 873 |
"epoch": 0.5642201834862385,
|
| 874 |
+
"grad_norm": 1.3817944954956538,
|
| 875 |
"learning_rate": 9.503825769350016e-06,
|
| 876 |
+
"loss": 0.9759,
|
| 877 |
"step": 615
|
| 878 |
},
|
| 879 |
{
|
| 880 |
"epoch": 0.5688073394495413,
|
| 881 |
+
"grad_norm": 1.4097103167722767,
|
| 882 |
"learning_rate": 9.343971481858246e-06,
|
| 883 |
+
"loss": 0.9721,
|
| 884 |
"step": 620
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 0.573394495412844,
|
| 888 |
+
"grad_norm": 1.3297950318079086,
|
| 889 |
"learning_rate": 9.184285390178978e-06,
|
| 890 |
+
"loss": 0.9679,
|
| 891 |
"step": 625
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.5779816513761468,
|
| 895 |
+
"grad_norm": 1.3397331767173486,
|
| 896 |
"learning_rate": 9.024808435416435e-06,
|
| 897 |
+
"loss": 0.9555,
|
| 898 |
"step": 630
|
| 899 |
},
|
| 900 |
{
|
| 901 |
"epoch": 0.5825688073394495,
|
| 902 |
+
"grad_norm": 1.3793546519569164,
|
| 903 |
"learning_rate": 8.865581505055292e-06,
|
| 904 |
+
"loss": 0.9657,
|
| 905 |
"step": 635
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.5871559633027523,
|
| 909 |
+
"grad_norm": 1.3699122633364083,
|
| 910 |
"learning_rate": 8.706645422477739e-06,
|
| 911 |
+
"loss": 0.9611,
|
| 912 |
"step": 640
|
| 913 |
},
|
| 914 |
{
|
| 915 |
"epoch": 0.591743119266055,
|
| 916 |
+
"grad_norm": 1.652929217335104,
|
| 917 |
"learning_rate": 8.548040936496989e-06,
|
| 918 |
+
"loss": 0.9656,
|
| 919 |
"step": 645
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 0.5963302752293578,
|
| 923 |
+
"grad_norm": 1.4000463163388799,
|
| 924 |
"learning_rate": 8.389808710909881e-06,
|
| 925 |
+
"loss": 0.9689,
|
| 926 |
"step": 650
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 0.6009174311926605,
|
| 930 |
+
"grad_norm": 1.3744642974899226,
|
| 931 |
"learning_rate": 8.231989314071318e-06,
|
| 932 |
+
"loss": 0.9675,
|
| 933 |
"step": 655
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.6055045871559633,
|
| 937 |
+
"grad_norm": 1.367828783708881,
|
| 938 |
"learning_rate": 8.07462320849313e-06,
|
| 939 |
+
"loss": 0.9827,
|
| 940 |
"step": 660
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 0.6100917431192661,
|
| 944 |
+
"grad_norm": 1.3746153840246962,
|
| 945 |
"learning_rate": 7.917750740470116e-06,
|
| 946 |
+
"loss": 0.9612,
|
| 947 |
"step": 665
|
| 948 |
},
|
| 949 |
{
|
| 950 |
"epoch": 0.6146788990825688,
|
| 951 |
+
"grad_norm": 1.3634420428996064,
|
| 952 |
"learning_rate": 7.761412129735853e-06,
|
| 953 |
+
"loss": 0.9606,
|
| 954 |
"step": 670
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.6192660550458715,
|
| 958 |
+
"grad_norm": 1.365419016371242,
|
| 959 |
"learning_rate": 7.605647459150961e-06,
|
| 960 |
+
"loss": 0.9701,
|
| 961 |
"step": 675
|
| 962 |
},
|
| 963 |
{
|
| 964 |
"epoch": 0.6238532110091743,
|
| 965 |
+
"grad_norm": 1.2994526356261897,
|
| 966 |
"learning_rate": 7.4504966644264775e-06,
|
| 967 |
+
"loss": 0.9496,
|
| 968 |
"step": 680
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 0.6284403669724771,
|
| 972 |
+
"grad_norm": 1.3036192585622581,
|
| 973 |
"learning_rate": 7.295999523884921e-06,
|
| 974 |
"loss": 0.9603,
|
| 975 |
"step": 685
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.6330275229357798,
|
| 979 |
+
"grad_norm": 1.3656479207440784,
|
| 980 |
"learning_rate": 7.142195648261747e-06,
|
| 981 |
+
"loss": 0.971,
|
| 982 |
"step": 690
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.6376146788990825,
|
| 986 |
+
"grad_norm": 1.3493578112863471,
|
| 987 |
"learning_rate": 6.989124470549746e-06,
|
| 988 |
+
"loss": 0.9401,
|
| 989 |
"step": 695
|
| 990 |
},
|
| 991 |
{
|
| 992 |
"epoch": 0.6422018348623854,
|
| 993 |
+
"grad_norm": 1.321094455754546,
|
| 994 |
"learning_rate": 6.83682523588902e-06,
|
| 995 |
+
"loss": 0.9632,
|
| 996 |
"step": 700
|
| 997 |
},
|
| 998 |
{
|
| 999 |
"epoch": 0.6467889908256881,
|
| 1000 |
+
"grad_norm": 1.3430854963278231,
|
| 1001 |
"learning_rate": 6.685336991505122e-06,
|
| 1002 |
+
"loss": 0.9648,
|
| 1003 |
"step": 705
|
| 1004 |
},
|
| 1005 |
{
|
| 1006 |
"epoch": 0.6513761467889908,
|
| 1007 |
+
"grad_norm": 1.295245677177454,
|
| 1008 |
"learning_rate": 6.5346985766979384e-06,
|
| 1009 |
+
"loss": 0.9547,
|
| 1010 |
"step": 710
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 0.6559633027522935,
|
| 1014 |
+
"grad_norm": 1.3017382650713638,
|
| 1015 |
"learning_rate": 6.384948612883872e-06,
|
| 1016 |
+
"loss": 0.9597,
|
| 1017 |
"step": 715
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.6605504587155964,
|
| 1021 |
+
"grad_norm": 1.2936500217660096,
|
| 1022 |
"learning_rate": 6.2361254936939e-06,
|
| 1023 |
+
"loss": 0.9603,
|
| 1024 |
"step": 720
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
"epoch": 0.6651376146788991,
|
| 1028 |
+
"grad_norm": 1.325710868385086,
|
| 1029 |
"learning_rate": 6.0882673751300235e-06,
|
| 1030 |
+
"loss": 0.9476,
|
| 1031 |
"step": 725
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
"epoch": 0.6697247706422018,
|
| 1035 |
+
"grad_norm": 1.3419334755973713,
|
| 1036 |
"learning_rate": 5.941412165782645e-06,
|
| 1037 |
+
"loss": 0.9565,
|
| 1038 |
"step": 730
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
"epoch": 0.6743119266055045,
|
| 1042 |
+
"grad_norm": 1.437071920630301,
|
| 1043 |
"learning_rate": 5.79559751711138e-06,
|
| 1044 |
+
"loss": 0.9479,
|
| 1045 |
"step": 735
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
"epoch": 0.6788990825688074,
|
| 1049 |
+
"grad_norm": 1.3574467430871946,
|
| 1050 |
"learning_rate": 5.650860813791786e-06,
|
| 1051 |
+
"loss": 0.9485,
|
| 1052 |
"step": 740
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 0.6834862385321101,
|
| 1056 |
+
"grad_norm": 1.3767014536400128,
|
| 1057 |
"learning_rate": 5.507239164130501e-06,
|
| 1058 |
+
"loss": 0.9492,
|
| 1059 |
"step": 745
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.6880733944954128,
|
| 1063 |
+
"grad_norm": 1.3706065876046587,
|
| 1064 |
"learning_rate": 5.364769390551225e-06,
|
| 1065 |
+
"loss": 0.9678,
|
| 1066 |
"step": 750
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 0.6926605504587156,
|
| 1070 |
+
"grad_norm": 1.295329148997341,
|
| 1071 |
"learning_rate": 5.223488020154028e-06,
|
| 1072 |
+
"loss": 0.9597,
|
| 1073 |
"step": 755
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 0.6972477064220184,
|
| 1077 |
+
"grad_norm": 1.317633111030153,
|
| 1078 |
"learning_rate": 5.083431275350312e-06,
|
| 1079 |
+
"loss": 0.9255,
|
| 1080 |
"step": 760
|
| 1081 |
},
|
| 1082 |
{
|
| 1083 |
"epoch": 0.7018348623853211,
|
| 1084 |
+
"grad_norm": 1.296366909205948,
|
| 1085 |
"learning_rate": 4.9446350645759885e-06,
|
| 1086 |
+
"loss": 0.9418,
|
| 1087 |
"step": 765
|
| 1088 |
},
|
| 1089 |
{
|
| 1090 |
"epoch": 0.7064220183486238,
|
| 1091 |
+
"grad_norm": 1.3869599027709183,
|
| 1092 |
"learning_rate": 4.807134973085036e-06,
|
| 1093 |
+
"loss": 0.9666,
|
| 1094 |
"step": 770
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 0.7110091743119266,
|
| 1098 |
+
"grad_norm": 1.2574891887163746,
|
| 1099 |
"learning_rate": 4.670966253826027e-06,
|
| 1100 |
+
"loss": 0.935,
|
| 1101 |
"step": 775
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.7155963302752294,
|
| 1105 |
+
"grad_norm": 1.3824364498830946,
|
| 1106 |
"learning_rate": 4.53616381840377e-06,
|
| 1107 |
"loss": 0.9468,
|
| 1108 |
"step": 780
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.7201834862385321,
|
| 1112 |
+
"grad_norm": 1.3301006886667626,
|
| 1113 |
"learning_rate": 4.402762228128531e-06,
|
| 1114 |
+
"loss": 0.9393,
|
| 1115 |
"step": 785
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
"epoch": 0.7247706422018348,
|
| 1119 |
+
"grad_norm": 1.3916604005630384,
|
| 1120 |
"learning_rate": 4.270795685155001e-06,
|
| 1121 |
+
"loss": 0.9478,
|
| 1122 |
"step": 790
|
| 1123 |
},
|
| 1124 |
{
|
| 1125 |
"epoch": 0.7293577981651376,
|
| 1126 |
+
"grad_norm": 1.3065920210924657,
|
| 1127 |
"learning_rate": 4.140298023713416e-06,
|
| 1128 |
+
"loss": 0.9301,
|
| 1129 |
"step": 795
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.7339449541284404,
|
| 1133 |
+
"grad_norm": 1.3321418112780183,
|
| 1134 |
"learning_rate": 4.0113027014349374e-06,
|
| 1135 |
+
"loss": 0.9148,
|
| 1136 |
"step": 800
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 0.7385321100917431,
|
| 1140 |
+
"grad_norm": 1.3066885947900655,
|
| 1141 |
"learning_rate": 3.883842790773647e-06,
|
| 1142 |
+
"loss": 0.9574,
|
| 1143 |
"step": 805
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.7431192660550459,
|
| 1147 |
+
"grad_norm": 1.2788509584031447,
|
| 1148 |
"learning_rate": 3.757950970527249e-06,
|
| 1149 |
+
"loss": 0.9261,
|
| 1150 |
"step": 810
|
| 1151 |
},
|
| 1152 |
{
|
| 1153 |
"epoch": 0.7477064220183486,
|
| 1154 |
+
"grad_norm": 1.2817821352009064,
|
| 1155 |
"learning_rate": 3.633659517458736e-06,
|
| 1156 |
+
"loss": 0.9248,
|
| 1157 |
"step": 815
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
"epoch": 0.7522935779816514,
|
| 1161 |
+
"grad_norm": 1.2551311767393507,
|
| 1162 |
"learning_rate": 3.511000298021098e-06,
|
| 1163 |
+
"loss": 0.9232,
|
| 1164 |
"step": 820
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.7568807339449541,
|
| 1168 |
+
"grad_norm": 1.3288724774154381,
|
| 1169 |
"learning_rate": 3.39000476018726e-06,
|
| 1170 |
+
"loss": 0.9266,
|
| 1171 |
"step": 825
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
"epoch": 0.7614678899082569,
|
| 1175 |
+
"grad_norm": 1.2641453349887841,
|
| 1176 |
"learning_rate": 3.2707039253872796e-06,
|
| 1177 |
+
"loss": 0.9231,
|
| 1178 |
"step": 830
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 0.7660550458715596,
|
| 1182 |
+
"grad_norm": 1.2984387481989768,
|
| 1183 |
"learning_rate": 3.153128380554941e-06,
|
| 1184 |
+
"loss": 0.9315,
|
| 1185 |
"step": 835
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.7706422018348624,
|
| 1189 |
+
"grad_norm": 1.2705124222427846,
|
| 1190 |
"learning_rate": 3.037308270285709e-06,
|
| 1191 |
+
"loss": 0.9231,
|
| 1192 |
"step": 840
|
| 1193 |
},
|
| 1194 |
{
|
| 1195 |
"epoch": 0.7752293577981652,
|
| 1196 |
+
"grad_norm": 1.267466264071829,
|
| 1197 |
"learning_rate": 2.923273289108115e-06,
|
| 1198 |
+
"loss": 0.9411,
|
| 1199 |
"step": 845
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
"epoch": 0.7798165137614679,
|
| 1203 |
+
"grad_norm": 1.356820160448841,
|
| 1204 |
"learning_rate": 2.8110526738705345e-06,
|
| 1205 |
+
"loss": 0.924,
|
| 1206 |
"step": 850
|
| 1207 |
},
|
| 1208 |
{
|
| 1209 |
"epoch": 0.7844036697247706,
|
| 1210 |
+
"grad_norm": 1.3040569720039672,
|
| 1211 |
"learning_rate": 2.700675196245288e-06,
|
| 1212 |
+
"loss": 0.9388,
|
| 1213 |
"step": 855
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
"epoch": 0.7889908256880734,
|
| 1217 |
+
"grad_norm": 1.2420186482046902,
|
| 1218 |
"learning_rate": 2.592169155352031e-06,
|
| 1219 |
+
"loss": 0.9421,
|
| 1220 |
"step": 860
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 0.7935779816513762,
|
| 1224 |
+
"grad_norm": 1.2616233171009335,
|
| 1225 |
"learning_rate": 2.485562370502279e-06,
|
| 1226 |
+
"loss": 0.9432,
|
| 1227 |
"step": 865
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.7981651376146789,
|
| 1231 |
+
"grad_norm": 1.3424259509025755,
|
| 1232 |
"learning_rate": 2.3808821740669608e-06,
|
| 1233 |
"loss": 0.9352,
|
| 1234 |
"step": 870
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8027522935779816,
|
| 1238 |
+
"grad_norm": 1.322358943195577,
|
| 1239 |
"learning_rate": 2.2781554044688015e-06,
|
| 1240 |
"loss": 0.9379,
|
| 1241 |
"step": 875
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.8073394495412844,
|
| 1245 |
+
"grad_norm": 1.2847587486081231,
|
| 1246 |
"learning_rate": 2.1774083993013715e-06,
|
| 1247 |
+
"loss": 0.935,
|
| 1248 |
"step": 880
|
| 1249 |
},
|
| 1250 |
{
|
| 1251 |
"epoch": 0.8119266055045872,
|
| 1252 |
+
"grad_norm": 1.2963379429813555,
|
| 1253 |
"learning_rate": 2.0786669885765044e-06,
|
| 1254 |
+
"loss": 0.9271,
|
| 1255 |
"step": 885
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 0.8165137614678899,
|
| 1259 |
+
"grad_norm": 1.2767056632273088,
|
| 1260 |
"learning_rate": 1.981956488101898e-06,
|
| 1261 |
+
"loss": 0.9402,
|
| 1262 |
"step": 890
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 0.8211009174311926,
|
| 1266 |
+
"grad_norm": 1.2867778108368306,
|
| 1267 |
"learning_rate": 1.8873016929904942e-06,
|
| 1268 |
+
"loss": 0.9421,
|
| 1269 |
"step": 895
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.8256880733944955,
|
| 1273 |
+
"grad_norm": 1.2551888718092907,
|
| 1274 |
"learning_rate": 1.7947268713034128e-06,
|
| 1275 |
+
"loss": 0.9119,
|
| 1276 |
"step": 900
|
| 1277 |
},
|
| 1278 |
{
|
| 1279 |
"epoch": 0.8302752293577982,
|
| 1280 |
+
"grad_norm": 1.2420897547436003,
|
| 1281 |
"learning_rate": 1.704255757827963e-06,
|
| 1282 |
+
"loss": 0.9146,
|
| 1283 |
"step": 905
|
| 1284 |
},
|
| 1285 |
{
|
| 1286 |
"epoch": 0.8348623853211009,
|
| 1287 |
+
"grad_norm": 1.2938041997903864,
|
| 1288 |
"learning_rate": 1.6159115479924259e-06,
|
| 1289 |
+
"loss": 0.9308,
|
| 1290 |
"step": 910
|
| 1291 |
},
|
| 1292 |
{
|
| 1293 |
"epoch": 0.8394495412844036,
|
| 1294 |
+
"grad_norm": 1.3003180857551158,
|
| 1295 |
"learning_rate": 1.529716891919074e-06,
|
| 1296 |
+
"loss": 0.9396,
|
| 1297 |
"step": 915
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 0.8440366972477065,
|
| 1301 |
+
"grad_norm": 1.2759086944542821,
|
| 1302 |
"learning_rate": 1.4456938886170413e-06,
|
| 1303 |
+
"loss": 0.9109,
|
| 1304 |
"step": 920
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 0.8486238532110092,
|
| 1308 |
+
"grad_norm": 1.33319189540699,
|
| 1309 |
"learning_rate": 1.3638640803164516e-06,
|
| 1310 |
+
"loss": 0.9237,
|
| 1311 |
"step": 925
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.8532110091743119,
|
| 1315 |
+
"grad_norm": 1.2809646406448443,
|
| 1316 |
"learning_rate": 1.2842484469453365e-06,
|
| 1317 |
+
"loss": 0.914,
|
| 1318 |
"step": 930
|
| 1319 |
},
|
| 1320 |
{
|
| 1321 |
"epoch": 0.8577981651376146,
|
| 1322 |
+
"grad_norm": 1.2918537502920435,
|
| 1323 |
"learning_rate": 1.2068674007506787e-06,
|
| 1324 |
+
"loss": 0.9286,
|
| 1325 |
"step": 935
|
| 1326 |
},
|
| 1327 |
{
|
| 1328 |
"epoch": 0.8623853211009175,
|
| 1329 |
+
"grad_norm": 1.3494320187543551,
|
| 1330 |
"learning_rate": 1.1317407810650372e-06,
|
| 1331 |
+
"loss": 0.9264,
|
| 1332 |
"step": 940
|
| 1333 |
},
|
| 1334 |
{
|
| 1335 |
"epoch": 0.8669724770642202,
|
| 1336 |
+
"grad_norm": 1.247741786729038,
|
| 1337 |
"learning_rate": 1.0588878492200261e-06,
|
| 1338 |
+
"loss": 0.9321,
|
| 1339 |
"step": 945
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
"epoch": 0.8715596330275229,
|
| 1343 |
+
"grad_norm": 1.2685992470376304,
|
| 1344 |
"learning_rate": 9.883272836080116e-07,
|
| 1345 |
+
"loss": 0.9199,
|
| 1346 |
"step": 950
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 0.8761467889908257,
|
| 1350 |
+
"grad_norm": 1.2982688058402074,
|
| 1351 |
"learning_rate": 9.200771748932513e-07,
|
| 1352 |
+
"loss": 0.9326,
|
| 1353 |
"step": 955
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.8807339449541285,
|
| 1357 |
+
"grad_norm": 1.3089054904372284,
|
| 1358 |
"learning_rate": 8.541550213737171e-07,
|
| 1359 |
+
"loss": 0.9114,
|
| 1360 |
"step": 960
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.8853211009174312,
|
| 1364 |
+
"grad_norm": 1.252254149851376,
|
| 1365 |
"learning_rate": 7.905777244947954e-07,
|
| 1366 |
+
"loss": 0.918,
|
| 1367 |
"step": 965
|
| 1368 |
},
|
| 1369 |
{
|
| 1370 |
"epoch": 0.8899082568807339,
|
| 1371 |
+
"grad_norm": 1.2933393496044958,
|
| 1372 |
"learning_rate": 7.293615845160196e-07,
|
| 1373 |
+
"loss": 0.9288,
|
| 1374 |
"step": 970
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.8944954128440367,
|
| 1378 |
+
"grad_norm": 1.2618705359183393,
|
| 1379 |
"learning_rate": 6.705222963319191e-07,
|
| 1380 |
+
"loss": 0.9376,
|
| 1381 |
"step": 975
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 0.8990825688073395,
|
| 1385 |
+
"grad_norm": 1.2592063070116135,
|
| 1386 |
"learning_rate": 6.140749454480932e-07,
|
| 1387 |
+
"loss": 0.9116,
|
| 1388 |
"step": 980
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 0.9036697247706422,
|
| 1392 |
+
"grad_norm": 1.2987200032177229,
|
| 1393 |
"learning_rate": 5.600340041135133e-07,
|
| 1394 |
+
"loss": 0.9181,
|
| 1395 |
"step": 985
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.908256880733945,
|
| 1399 |
+
"grad_norm": 1.300028729042509,
|
| 1400 |
"learning_rate": 5.0841332761005e-07,
|
| 1401 |
+
"loss": 0.9333,
|
| 1402 |
"step": 990
|
| 1403 |
},
|
| 1404 |
{
|
| 1405 |
"epoch": 0.9128440366972477,
|
| 1406 |
+
"grad_norm": 1.178743058473968,
|
| 1407 |
"learning_rate": 4.592261507001994e-07,
|
| 1408 |
+
"loss": 0.9145,
|
| 1409 |
"step": 995
|
| 1410 |
},
|
| 1411 |
{
|
| 1412 |
"epoch": 0.9174311926605505,
|
| 1413 |
+
"grad_norm": 1.2612257727695169,
|
| 1414 |
"learning_rate": 4.124850842338779e-07,
|
| 1415 |
+
"loss": 0.9198,
|
| 1416 |
"step": 1000
|
| 1417 |
},
|
| 1418 |
{
|
| 1419 |
"epoch": 0.9220183486238532,
|
| 1420 |
+
"grad_norm": 1.2664257049420728,
|
| 1421 |
"learning_rate": 3.6820211191520127e-07,
|
| 1422 |
+
"loss": 0.9237,
|
| 1423 |
"step": 1005
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 0.926605504587156,
|
| 1427 |
+
"grad_norm": 1.2636645907476838,
|
| 1428 |
"learning_rate": 3.263885872300343e-07,
|
| 1429 |
+
"loss": 0.9228,
|
| 1430 |
"step": 1010
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 0.9311926605504587,
|
| 1434 |
+
"grad_norm": 1.242318963263546,
|
| 1435 |
"learning_rate": 2.870552305351382e-07,
|
| 1436 |
+
"loss": 0.9005,
|
| 1437 |
"step": 1015
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.9357798165137615,
|
| 1441 |
+
"grad_norm": 1.2809884089732457,
|
| 1442 |
"learning_rate": 2.5021212630962246e-07,
|
| 1443 |
+
"loss": 0.9276,
|
| 1444 |
"step": 1020
|
| 1445 |
},
|
| 1446 |
{
|
| 1447 |
"epoch": 0.9403669724770642,
|
| 1448 |
+
"grad_norm": 1.2119069638655617,
|
| 1449 |
"learning_rate": 2.158687205694443e-07,
|
| 1450 |
+
"loss": 0.907,
|
| 1451 |
"step": 1025
|
| 1452 |
},
|
| 1453 |
{
|
| 1454 |
"epoch": 0.944954128440367,
|
| 1455 |
+
"grad_norm": 1.2376811762313147,
|
| 1456 |
"learning_rate": 1.840338184455881e-07,
|
| 1457 |
+
"loss": 0.9269,
|
| 1458 |
"step": 1030
|
| 1459 |
},
|
| 1460 |
{
|
| 1461 |
"epoch": 0.9495412844036697,
|
| 1462 |
+
"grad_norm": 1.2478362914541763,
|
| 1463 |
"learning_rate": 1.5471558192656776e-07,
|
| 1464 |
+
"loss": 0.9159,
|
| 1465 |
"step": 1035
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.9541284403669725,
|
| 1469 |
+
"grad_norm": 1.2217359410822013,
|
| 1470 |
"learning_rate": 1.279215277658097e-07,
|
| 1471 |
+
"loss": 0.932,
|
| 1472 |
"step": 1040
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 0.9587155963302753,
|
| 1476 |
+
"grad_norm": 1.2656920108994318,
|
| 1477 |
"learning_rate": 1.0365852555447642e-07,
|
| 1478 |
+
"loss": 0.925,
|
| 1479 |
"step": 1045
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.963302752293578,
|
| 1483 |
+
"grad_norm": 1.2272545903225942,
|
| 1484 |
"learning_rate": 8.19327959602012e-08,
|
| 1485 |
+
"loss": 0.9346,
|
| 1486 |
"step": 1050
|
| 1487 |
},
|
| 1488 |
{
|
| 1489 |
"epoch": 0.9678899082568807,
|
| 1490 |
+
"grad_norm": 1.2652144722057859,
|
| 1491 |
"learning_rate": 6.274990913221035e-08,
|
| 1492 |
+
"loss": 0.9288,
|
| 1493 |
"step": 1055
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"epoch": 0.9724770642201835,
|
| 1497 |
+
"grad_norm": 1.2861426339061892,
|
| 1498 |
"learning_rate": 4.6114783273213395e-08,
|
| 1499 |
+
"loss": 0.9134,
|
| 1500 |
"step": 1060
|
| 1501 |
},
|
| 1502 |
{
|
| 1503 |
"epoch": 0.9770642201834863,
|
| 1504 |
+
"grad_norm": 1.2468482937718688,
|
| 1505 |
"learning_rate": 3.203168337845508e-08,
|
| 1506 |
+
"loss": 0.9062,
|
| 1507 |
"step": 1065
|
| 1508 |
},
|
| 1509 |
{
|
| 1510 |
"epoch": 0.981651376146789,
|
| 1511 |
+
"grad_norm": 1.2763891416814128,
|
| 1512 |
"learning_rate": 2.05042201422323e-08,
|
| 1513 |
+
"loss": 0.9156,
|
| 1514 |
"step": 1070
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 0.9862385321100917,
|
| 1518 |
+
"grad_norm": 1.2950677825635273,
|
| 1519 |
"learning_rate": 1.1535349032167908e-08,
|
| 1520 |
+
"loss": 0.8987,
|
| 1521 |
"step": 1075
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
"epoch": 0.9908256880733946,
|
| 1525 |
+
"grad_norm": 1.2554986306370373,
|
| 1526 |
"learning_rate": 5.127369531473525e-09,
|
| 1527 |
+
"loss": 0.9281,
|
| 1528 |
"step": 1080
|
| 1529 |
},
|
| 1530 |
{
|
| 1531 |
"epoch": 0.9954128440366973,
|
| 1532 |
+
"grad_norm": 1.2713525280166649,
|
| 1533 |
"learning_rate": 1.2819245493955746e-09,
|
| 1534 |
+
"loss": 0.9239,
|
| 1535 |
"step": 1085
|
| 1536 |
},
|
| 1537 |
{
|
| 1538 |
"epoch": 1.0,
|
| 1539 |
+
"grad_norm": 1.4188473493340228,
|
| 1540 |
"learning_rate": 0.0,
|
| 1541 |
"loss": 0.9203,
|
| 1542 |
"step": 1090
|
| 1543 |
},
|
| 1544 |
{
|
| 1545 |
"epoch": 1.0,
|
| 1546 |
+
"eval_loss": 0.9367018342018127,
|
| 1547 |
+
"eval_runtime": 316.9529,
|
| 1548 |
+
"eval_samples_per_second": 48.685,
|
| 1549 |
+
"eval_steps_per_second": 0.764,
|
| 1550 |
"step": 1090
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
"epoch": 1.0,
|
| 1554 |
"step": 1090,
|
| 1555 |
"total_flos": 456447649382400.0,
|
| 1556 |
+
"train_loss": 0.9783222812031387,
|
| 1557 |
+
"train_runtime": 12140.5299,
|
| 1558 |
+
"train_samples_per_second": 11.484,
|
| 1559 |
+
"train_steps_per_second": 0.09
|
| 1560 |
}
|
| 1561 |
],
|
| 1562 |
"logging_steps": 5,
|