Commit
•
b4ceef1
1
Parent(s):
9eb33b4
Training in progress, step 64000
Browse files- README.md +355 -0
- added_tokens.json +7 -0
- all_results.json +14 -0
- config.json +28 -0
- eval_results.json +9 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +9 -0
- tokenizer.json +0 -0
- tokenizer_config.json +62 -0
- train_results.json +8 -0
- trainer_state.json +3328 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,355 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: DouglasPontes/2020-Q1-full_tweets
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: 2020-Q2-full_tweets
|
| 7 |
+
results: []
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 11 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 12 |
+
|
| 13 |
+
# 2020-Q2-full_tweets
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [DouglasPontes/2020-Q1-full_tweets](https://huggingface.co/DouglasPontes/2020-Q1-full_tweets) on an unknown dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 2.0061
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 4.1e-07
|
| 37 |
+
- train_batch_size: 16
|
| 38 |
+
- eval_batch_size: 16
|
| 39 |
+
- seed: 42
|
| 40 |
+
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
|
| 41 |
+
- lr_scheduler_type: linear
|
| 42 |
+
- training_steps: 2400000
|
| 43 |
+
|
| 44 |
+
### Training results
|
| 45 |
+
|
| 46 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 47 |
+
|:-------------:|:-----:|:-------:|:---------------:|
|
| 48 |
+
| No log | 0.01 | 8000 | 2.1043 |
|
| 49 |
+
| 2.2608 | 0.02 | 16000 | 2.0934 |
|
| 50 |
+
| 2.2608 | 0.03 | 24000 | 2.0862 |
|
| 51 |
+
| 2.2409 | 0.03 | 32000 | 2.0805 |
|
| 52 |
+
| 2.2409 | 0.04 | 40000 | 2.0793 |
|
| 53 |
+
| 2.2278 | 0.05 | 48000 | 2.0718 |
|
| 54 |
+
| 2.2278 | 0.06 | 56000 | 2.0753 |
|
| 55 |
+
| 2.2059 | 0.07 | 64000 | 2.0668 |
|
| 56 |
+
| 2.2059 | 0.08 | 72000 | 2.0657 |
|
| 57 |
+
| 2.1997 | 0.09 | 80000 | 2.0620 |
|
| 58 |
+
| 2.1997 | 0.1 | 88000 | 2.0553 |
|
| 59 |
+
| 2.1988 | 0.1 | 96000 | 2.0569 |
|
| 60 |
+
| 2.1988 | 0.11 | 104000 | 2.0525 |
|
| 61 |
+
| 2.1861 | 0.12 | 112000 | 2.0556 |
|
| 62 |
+
| 2.1861 | 0.13 | 120000 | 2.0493 |
|
| 63 |
+
| 2.1823 | 0.14 | 128000 | 2.0509 |
|
| 64 |
+
| 2.1823 | 0.15 | 136000 | 2.0461 |
|
| 65 |
+
| 2.1851 | 0.16 | 144000 | 2.0476 |
|
| 66 |
+
| 2.1851 | 0.17 | 152000 | 2.0450 |
|
| 67 |
+
| 2.1862 | 0.17 | 160000 | 2.0469 |
|
| 68 |
+
| 2.1862 | 0.18 | 168000 | 2.0442 |
|
| 69 |
+
| 2.1741 | 0.19 | 176000 | 2.0456 |
|
| 70 |
+
| 2.1741 | 0.2 | 184000 | 2.0442 |
|
| 71 |
+
| 2.181 | 0.21 | 192000 | 2.0402 |
|
| 72 |
+
| 2.181 | 0.22 | 200000 | 2.0423 |
|
| 73 |
+
| 2.1692 | 0.23 | 208000 | 2.0413 |
|
| 74 |
+
| 2.1692 | 0.24 | 216000 | 2.0448 |
|
| 75 |
+
| 2.1678 | 0.24 | 224000 | 2.0418 |
|
| 76 |
+
| 2.1678 | 0.25 | 232000 | 2.0417 |
|
| 77 |
+
| 2.1756 | 0.26 | 240000 | 2.0342 |
|
| 78 |
+
| 2.1756 | 0.27 | 248000 | 2.0377 |
|
| 79 |
+
| 2.1752 | 0.28 | 256000 | 2.0381 |
|
| 80 |
+
| 2.1752 | 0.29 | 264000 | 2.0354 |
|
| 81 |
+
| 2.1673 | 0.3 | 272000 | 2.0381 |
|
| 82 |
+
| 2.1673 | 0.31 | 280000 | 2.0375 |
|
| 83 |
+
| 2.1585 | 0.31 | 288000 | 2.0336 |
|
| 84 |
+
| 2.1585 | 0.32 | 296000 | 2.0344 |
|
| 85 |
+
| 2.1703 | 0.33 | 304000 | 2.0348 |
|
| 86 |
+
| 2.1703 | 0.34 | 312000 | 2.0330 |
|
| 87 |
+
| 2.1667 | 0.35 | 320000 | 2.0352 |
|
| 88 |
+
| 2.1667 | 0.36 | 328000 | 2.0359 |
|
| 89 |
+
| 2.1649 | 0.37 | 336000 | 2.0317 |
|
| 90 |
+
| 2.1649 | 0.38 | 344000 | 2.0314 |
|
| 91 |
+
| 2.1564 | 0.38 | 352000 | 2.0306 |
|
| 92 |
+
| 2.1564 | 0.39 | 360000 | 2.0299 |
|
| 93 |
+
| 2.161 | 0.4 | 368000 | 2.0317 |
|
| 94 |
+
| 2.161 | 0.41 | 376000 | 2.0325 |
|
| 95 |
+
| 2.1551 | 0.42 | 384000 | 2.0274 |
|
| 96 |
+
| 2.1551 | 0.43 | 392000 | 2.0282 |
|
| 97 |
+
| 2.1602 | 0.44 | 400000 | 2.0301 |
|
| 98 |
+
| 2.1602 | 0.45 | 408000 | 2.0303 |
|
| 99 |
+
| 2.1581 | 0.45 | 416000 | 2.0260 |
|
| 100 |
+
| 2.1581 | 0.46 | 424000 | 2.0248 |
|
| 101 |
+
| 2.1494 | 0.47 | 432000 | 2.0265 |
|
| 102 |
+
| 2.1494 | 0.48 | 440000 | 2.0247 |
|
| 103 |
+
| 2.1508 | 0.49 | 448000 | 2.0231 |
|
| 104 |
+
| 2.1508 | 0.5 | 456000 | 2.0276 |
|
| 105 |
+
| 2.153 | 0.51 | 464000 | 2.0276 |
|
| 106 |
+
| 2.153 | 0.51 | 472000 | 2.0242 |
|
| 107 |
+
| 2.1489 | 0.52 | 480000 | 2.0259 |
|
| 108 |
+
| 2.1489 | 0.53 | 488000 | 2.0257 |
|
| 109 |
+
| 2.1468 | 0.54 | 496000 | 2.0275 |
|
| 110 |
+
| 2.1468 | 0.55 | 504000 | 2.0303 |
|
| 111 |
+
| 2.1446 | 0.56 | 512000 | 2.0248 |
|
| 112 |
+
| 2.1446 | 0.57 | 520000 | 2.0286 |
|
| 113 |
+
| 2.1409 | 0.58 | 528000 | 2.0211 |
|
| 114 |
+
| 2.1409 | 0.58 | 536000 | 2.0204 |
|
| 115 |
+
| 2.1536 | 0.59 | 544000 | 2.0199 |
|
| 116 |
+
| 2.1536 | 0.6 | 552000 | 2.0281 |
|
| 117 |
+
| 2.1416 | 0.61 | 560000 | 2.0237 |
|
| 118 |
+
| 2.1416 | 0.62 | 568000 | 2.0231 |
|
| 119 |
+
| 2.1502 | 0.63 | 576000 | 2.0205 |
|
| 120 |
+
| 2.1502 | 0.64 | 584000 | 2.0217 |
|
| 121 |
+
| 2.1424 | 0.65 | 592000 | 2.0242 |
|
| 122 |
+
| 2.1424 | 0.65 | 600000 | 2.0238 |
|
| 123 |
+
| 2.1469 | 0.66 | 608000 | 2.0192 |
|
| 124 |
+
| 2.1469 | 0.67 | 616000 | 2.0249 |
|
| 125 |
+
| 2.145 | 0.68 | 624000 | 2.0196 |
|
| 126 |
+
| 2.145 | 0.69 | 632000 | 2.0224 |
|
| 127 |
+
| 2.1503 | 0.7 | 640000 | 2.0216 |
|
| 128 |
+
| 2.1503 | 0.71 | 648000 | 2.0228 |
|
| 129 |
+
| 2.1355 | 0.72 | 656000 | 2.0197 |
|
| 130 |
+
| 2.1355 | 0.72 | 664000 | 2.0240 |
|
| 131 |
+
| 2.1392 | 0.73 | 672000 | 2.0232 |
|
| 132 |
+
| 2.1392 | 0.74 | 680000 | 2.0209 |
|
| 133 |
+
| 2.1378 | 0.75 | 688000 | 2.0219 |
|
| 134 |
+
| 2.1378 | 0.76 | 696000 | 2.0192 |
|
| 135 |
+
| 2.1446 | 0.77 | 704000 | 2.0195 |
|
| 136 |
+
| 2.1446 | 0.78 | 712000 | 2.0197 |
|
| 137 |
+
| 2.1351 | 0.79 | 720000 | 2.0184 |
|
| 138 |
+
| 2.1351 | 0.79 | 728000 | 2.0162 |
|
| 139 |
+
| 2.1437 | 0.8 | 736000 | 2.0151 |
|
| 140 |
+
| 2.1437 | 0.81 | 744000 | 2.0202 |
|
| 141 |
+
| 2.1249 | 0.82 | 752000 | 2.0169 |
|
| 142 |
+
| 2.1249 | 0.83 | 760000 | 2.0189 |
|
| 143 |
+
| 2.1355 | 0.84 | 768000 | 2.0221 |
|
| 144 |
+
| 2.1355 | 0.85 | 776000 | 2.0194 |
|
| 145 |
+
| 2.1387 | 0.86 | 784000 | 2.0189 |
|
| 146 |
+
| 2.1387 | 0.86 | 792000 | 2.0165 |
|
| 147 |
+
| 2.1334 | 0.87 | 800000 | 2.0169 |
|
| 148 |
+
| 2.1334 | 0.88 | 808000 | 2.0189 |
|
| 149 |
+
| 2.137 | 0.89 | 816000 | 2.0162 |
|
| 150 |
+
| 2.137 | 0.9 | 824000 | 2.0168 |
|
| 151 |
+
| 2.1331 | 0.91 | 832000 | 2.0193 |
|
| 152 |
+
| 2.1331 | 0.92 | 840000 | 2.0166 |
|
| 153 |
+
| 2.1293 | 0.93 | 848000 | 2.0137 |
|
| 154 |
+
| 2.1293 | 0.93 | 856000 | 2.0183 |
|
| 155 |
+
| 2.1358 | 0.94 | 864000 | 2.0184 |
|
| 156 |
+
| 2.1358 | 0.95 | 872000 | 2.0171 |
|
| 157 |
+
| 2.1296 | 0.96 | 880000 | 2.0179 |
|
| 158 |
+
| 2.1296 | 0.97 | 888000 | 2.0152 |
|
| 159 |
+
| 2.1319 | 0.98 | 896000 | 2.0174 |
|
| 160 |
+
| 2.1319 | 0.99 | 904000 | 2.0206 |
|
| 161 |
+
| 2.1344 | 1.0 | 912000 | 2.0179 |
|
| 162 |
+
| 2.1344 | 1.0 | 920000 | 2.0154 |
|
| 163 |
+
| 2.1352 | 1.01 | 928000 | 2.0185 |
|
| 164 |
+
| 2.1352 | 1.02 | 936000 | 2.0170 |
|
| 165 |
+
| 2.1336 | 1.03 | 944000 | 2.0164 |
|
| 166 |
+
| 2.1336 | 1.04 | 952000 | 2.0137 |
|
| 167 |
+
| 2.1315 | 1.05 | 960000 | 2.0176 |
|
| 168 |
+
| 2.1315 | 1.06 | 968000 | 2.0155 |
|
| 169 |
+
| 2.1255 | 1.06 | 976000 | 2.0145 |
|
| 170 |
+
| 2.1255 | 1.07 | 984000 | 2.0233 |
|
| 171 |
+
| 2.1249 | 1.08 | 992000 | 2.0148 |
|
| 172 |
+
| 2.1249 | 1.09 | 1000000 | 2.0162 |
|
| 173 |
+
| 2.123 | 1.1 | 1008000 | 2.0174 |
|
| 174 |
+
| 2.123 | 1.11 | 1016000 | 2.0150 |
|
| 175 |
+
| 2.1263 | 1.12 | 1024000 | 2.0161 |
|
| 176 |
+
| 2.1263 | 1.13 | 1032000 | 2.0129 |
|
| 177 |
+
| 2.1232 | 1.13 | 1040000 | 2.0167 |
|
| 178 |
+
| 2.1232 | 1.14 | 1048000 | 2.0125 |
|
| 179 |
+
| 2.1168 | 1.15 | 1056000 | 2.0113 |
|
| 180 |
+
| 2.1168 | 1.16 | 1064000 | 2.0136 |
|
| 181 |
+
| 2.1307 | 1.17 | 1072000 | 2.0143 |
|
| 182 |
+
| 2.1307 | 1.18 | 1080000 | 2.0166 |
|
| 183 |
+
| 2.1336 | 1.19 | 1088000 | 2.0103 |
|
| 184 |
+
| 2.1336 | 1.2 | 1096000 | 2.0130 |
|
| 185 |
+
| 2.1227 | 1.2 | 1104000 | 2.0125 |
|
| 186 |
+
| 2.1227 | 1.21 | 1112000 | 2.0183 |
|
| 187 |
+
| 2.1223 | 1.22 | 1120000 | 2.0148 |
|
| 188 |
+
| 2.1223 | 1.23 | 1128000 | 2.0147 |
|
| 189 |
+
| 2.1289 | 1.24 | 1136000 | 2.0109 |
|
| 190 |
+
| 2.1289 | 1.25 | 1144000 | 2.0164 |
|
| 191 |
+
| 2.1278 | 1.26 | 1152000 | 2.0163 |
|
| 192 |
+
| 2.1278 | 1.27 | 1160000 | 2.0121 |
|
| 193 |
+
| 2.1261 | 1.27 | 1168000 | 2.0113 |
|
| 194 |
+
| 2.1261 | 1.28 | 1176000 | 2.0137 |
|
| 195 |
+
| 2.126 | 1.29 | 1184000 | 2.0152 |
|
| 196 |
+
| 2.126 | 1.3 | 1192000 | 2.0104 |
|
| 197 |
+
| 2.1235 | 1.31 | 1200000 | 2.0132 |
|
| 198 |
+
| 2.1235 | 1.32 | 1208000 | 2.0114 |
|
| 199 |
+
| 2.1229 | 1.33 | 1216000 | 2.0105 |
|
| 200 |
+
| 2.1229 | 1.34 | 1224000 | 2.0131 |
|
| 201 |
+
| 2.1213 | 1.34 | 1232000 | 2.0141 |
|
| 202 |
+
| 2.1213 | 1.35 | 1240000 | 2.0109 |
|
| 203 |
+
| 2.1185 | 1.36 | 1248000 | 2.0129 |
|
| 204 |
+
| 2.1185 | 1.37 | 1256000 | 2.0110 |
|
| 205 |
+
| 2.131 | 1.38 | 1264000 | 2.0123 |
|
| 206 |
+
| 2.131 | 1.39 | 1272000 | 2.0105 |
|
| 207 |
+
| 2.1141 | 1.4 | 1280000 | 2.0104 |
|
| 208 |
+
| 2.1141 | 1.41 | 1288000 | 2.0150 |
|
| 209 |
+
| 2.1219 | 1.41 | 1296000 | 2.0161 |
|
| 210 |
+
| 2.1219 | 1.42 | 1304000 | 2.0093 |
|
| 211 |
+
| 2.1203 | 1.43 | 1312000 | 2.0104 |
|
| 212 |
+
| 2.1203 | 1.44 | 1320000 | 2.0144 |
|
| 213 |
+
| 2.1264 | 1.45 | 1328000 | 2.0085 |
|
| 214 |
+
| 2.1264 | 1.46 | 1336000 | 2.0119 |
|
| 215 |
+
| 2.1194 | 1.47 | 1344000 | 2.0118 |
|
| 216 |
+
| 2.1194 | 1.48 | 1352000 | 2.0110 |
|
| 217 |
+
| 2.117 | 1.48 | 1360000 | 2.0147 |
|
| 218 |
+
| 2.117 | 1.49 | 1368000 | 2.0135 |
|
| 219 |
+
| 2.1311 | 1.5 | 1376000 | 2.0077 |
|
| 220 |
+
| 2.1311 | 1.51 | 1384000 | 2.0066 |
|
| 221 |
+
| 2.1215 | 1.52 | 1392000 | 2.0089 |
|
| 222 |
+
| 2.1215 | 1.53 | 1400000 | 2.0118 |
|
| 223 |
+
| 2.1185 | 1.54 | 1408000 | 2.0105 |
|
| 224 |
+
| 2.1185 | 1.54 | 1416000 | 2.0123 |
|
| 225 |
+
| 2.1284 | 1.55 | 1424000 | 2.0134 |
|
| 226 |
+
| 2.1284 | 1.56 | 1432000 | 2.0093 |
|
| 227 |
+
| 2.1174 | 1.57 | 1440000 | 2.0102 |
|
| 228 |
+
| 2.1174 | 1.58 | 1448000 | 2.0076 |
|
| 229 |
+
| 2.1108 | 1.59 | 1456000 | 2.0074 |
|
| 230 |
+
| 2.1108 | 1.6 | 1464000 | 2.0071 |
|
| 231 |
+
| 2.1252 | 1.61 | 1472000 | 2.0092 |
|
| 232 |
+
| 2.1252 | 1.61 | 1480000 | 2.0080 |
|
| 233 |
+
| 2.121 | 1.62 | 1488000 | 2.0053 |
|
| 234 |
+
| 2.121 | 1.63 | 1496000 | 2.0072 |
|
| 235 |
+
| 2.1178 | 1.64 | 1504000 | 2.0059 |
|
| 236 |
+
| 2.1178 | 1.65 | 1512000 | 2.0084 |
|
| 237 |
+
| 2.1154 | 1.66 | 1520000 | 2.0106 |
|
| 238 |
+
| 2.1154 | 1.67 | 1528000 | 2.0117 |
|
| 239 |
+
| 2.1214 | 1.68 | 1536000 | 2.0070 |
|
| 240 |
+
| 2.1214 | 1.68 | 1544000 | 2.0079 |
|
| 241 |
+
| 2.1175 | 1.69 | 1552000 | 2.0102 |
|
| 242 |
+
| 2.1175 | 1.7 | 1560000 | 2.0097 |
|
| 243 |
+
| 2.1206 | 1.71 | 1568000 | 2.0092 |
|
| 244 |
+
| 2.1206 | 1.72 | 1576000 | 2.0055 |
|
| 245 |
+
| 2.1302 | 1.73 | 1584000 | 2.0085 |
|
| 246 |
+
| 2.1302 | 1.74 | 1592000 | 2.0110 |
|
| 247 |
+
| 2.1177 | 1.75 | 1600000 | 2.0065 |
|
| 248 |
+
| 2.1177 | 1.75 | 1608000 | 2.0132 |
|
| 249 |
+
| 2.1101 | 1.76 | 1616000 | 2.0086 |
|
| 250 |
+
| 2.1101 | 1.77 | 1624000 | 2.0077 |
|
| 251 |
+
| 2.1194 | 1.78 | 1632000 | 2.0081 |
|
| 252 |
+
| 2.1194 | 1.79 | 1640000 | 2.0088 |
|
| 253 |
+
| 2.1167 | 1.8 | 1648000 | 2.0022 |
|
| 254 |
+
| 2.1167 | 1.81 | 1656000 | 2.0077 |
|
| 255 |
+
| 2.1083 | 1.82 | 1664000 | 2.0066 |
|
| 256 |
+
| 2.1083 | 1.82 | 1672000 | 2.0137 |
|
| 257 |
+
| 2.1232 | 1.83 | 1680000 | 2.0067 |
|
| 258 |
+
| 2.1232 | 1.84 | 1688000 | 2.0039 |
|
| 259 |
+
| 2.1212 | 1.85 | 1696000 | 2.0090 |
|
| 260 |
+
| 2.1212 | 1.86 | 1704000 | 2.0079 |
|
| 261 |
+
| 2.1246 | 1.87 | 1712000 | 2.0083 |
|
| 262 |
+
| 2.1246 | 1.88 | 1720000 | 2.0039 |
|
| 263 |
+
| 2.1129 | 1.89 | 1728000 | 2.0069 |
|
| 264 |
+
| 2.1129 | 1.89 | 1736000 | 2.0079 |
|
| 265 |
+
| 2.1209 | 1.9 | 1744000 | 2.0058 |
|
| 266 |
+
| 2.1209 | 1.91 | 1752000 | 2.0072 |
|
| 267 |
+
| 2.1209 | 1.92 | 1760000 | 2.0068 |
|
| 268 |
+
| 2.1209 | 1.93 | 1768000 | 2.0079 |
|
| 269 |
+
| 2.1184 | 1.94 | 1776000 | 2.0036 |
|
| 270 |
+
| 2.1184 | 1.95 | 1784000 | 2.0065 |
|
| 271 |
+
| 2.1065 | 1.96 | 1792000 | 2.0077 |
|
| 272 |
+
| 2.1065 | 1.96 | 1800000 | 2.0062 |
|
| 273 |
+
| 2.109 | 1.97 | 1808000 | 2.0090 |
|
| 274 |
+
| 2.109 | 1.98 | 1816000 | 2.0124 |
|
| 275 |
+
| 2.1081 | 1.99 | 1824000 | 2.0066 |
|
| 276 |
+
| 2.1081 | 2.0 | 1832000 | 2.0081 |
|
| 277 |
+
| 2.1151 | 2.01 | 1840000 | 2.0085 |
|
| 278 |
+
| 2.1151 | 2.02 | 1848000 | 2.0054 |
|
| 279 |
+
| 2.1178 | 2.03 | 1856000 | 2.0058 |
|
| 280 |
+
| 2.1178 | 2.03 | 1864000 | 2.0048 |
|
| 281 |
+
| 2.1035 | 2.04 | 1872000 | 2.0040 |
|
| 282 |
+
| 2.1035 | 2.05 | 1880000 | 2.0059 |
|
| 283 |
+
| 2.1197 | 2.06 | 1888000 | 2.0071 |
|
| 284 |
+
| 2.1197 | 2.07 | 1896000 | 2.0057 |
|
| 285 |
+
| 2.1143 | 2.08 | 1904000 | 2.0059 |
|
| 286 |
+
| 2.1143 | 2.09 | 1912000 | 2.0043 |
|
| 287 |
+
| 2.1082 | 2.09 | 1920000 | 2.0068 |
|
| 288 |
+
| 2.1082 | 2.1 | 1928000 | 2.0057 |
|
| 289 |
+
| 2.1202 | 2.11 | 1936000 | 2.0072 |
|
| 290 |
+
| 2.1202 | 2.12 | 1944000 | 2.0057 |
|
| 291 |
+
| 2.1138 | 2.13 | 1952000 | 2.0051 |
|
| 292 |
+
| 2.1138 | 2.14 | 1960000 | 2.0085 |
|
| 293 |
+
| 2.1082 | 2.15 | 1968000 | 2.0076 |
|
| 294 |
+
| 2.1082 | 2.16 | 1976000 | 2.0077 |
|
| 295 |
+
| 2.1084 | 2.16 | 1984000 | 2.0020 |
|
| 296 |
+
| 2.1084 | 2.17 | 1992000 | 2.0050 |
|
| 297 |
+
| 2.1151 | 2.18 | 2000000 | 2.0066 |
|
| 298 |
+
| 2.1151 | 2.19 | 2008000 | 2.0031 |
|
| 299 |
+
| 2.1141 | 2.2 | 2016000 | 2.0128 |
|
| 300 |
+
| 2.1141 | 2.21 | 2024000 | 2.0022 |
|
| 301 |
+
| 2.1129 | 2.22 | 2032000 | 2.0065 |
|
| 302 |
+
| 2.1129 | 2.23 | 2040000 | 2.0054 |
|
| 303 |
+
| 2.1164 | 2.23 | 2048000 | 2.0039 |
|
| 304 |
+
| 2.1164 | 2.24 | 2056000 | 2.0031 |
|
| 305 |
+
| 2.1121 | 2.25 | 2064000 | 2.0101 |
|
| 306 |
+
| 2.1121 | 2.26 | 2072000 | 2.0099 |
|
| 307 |
+
| 2.1071 | 2.27 | 2080000 | 2.0042 |
|
| 308 |
+
| 2.1071 | 2.28 | 2088000 | 2.0030 |
|
| 309 |
+
| 2.1094 | 2.29 | 2096000 | 2.0048 |
|
| 310 |
+
| 2.1094 | 2.3 | 2104000 | 2.0046 |
|
| 311 |
+
| 2.1017 | 2.3 | 2112000 | 2.0039 |
|
| 312 |
+
| 2.1017 | 2.31 | 2120000 | 2.0011 |
|
| 313 |
+
| 2.1124 | 2.32 | 2128000 | 2.0071 |
|
| 314 |
+
| 2.1124 | 2.33 | 2136000 | 2.0061 |
|
| 315 |
+
| 2.1064 | 2.34 | 2144000 | 2.0040 |
|
| 316 |
+
| 2.1064 | 2.35 | 2152000 | 2.0075 |
|
| 317 |
+
| 2.115 | 2.36 | 2160000 | 2.0026 |
|
| 318 |
+
| 2.115 | 2.37 | 2168000 | 2.0068 |
|
| 319 |
+
| 2.114 | 2.37 | 2176000 | 2.0066 |
|
| 320 |
+
| 2.114 | 2.38 | 2184000 | 2.0080 |
|
| 321 |
+
| 2.1171 | 2.39 | 2192000 | 2.0032 |
|
| 322 |
+
| 2.1171 | 2.4 | 2200000 | 2.0036 |
|
| 323 |
+
| 2.1119 | 2.41 | 2208000 | 2.0048 |
|
| 324 |
+
| 2.1119 | 2.42 | 2216000 | 2.0059 |
|
| 325 |
+
| 2.1097 | 2.43 | 2224000 | 2.0058 |
|
| 326 |
+
| 2.1097 | 2.44 | 2232000 | 2.0049 |
|
| 327 |
+
| 2.1091 | 2.44 | 2240000 | 2.0058 |
|
| 328 |
+
| 2.1091 | 2.45 | 2248000 | 2.0032 |
|
| 329 |
+
| 2.1107 | 2.46 | 2256000 | 2.0077 |
|
| 330 |
+
| 2.1107 | 2.47 | 2264000 | 2.0032 |
|
| 331 |
+
| 2.1126 | 2.48 | 2272000 | 2.0055 |
|
| 332 |
+
| 2.1126 | 2.49 | 2280000 | 2.0026 |
|
| 333 |
+
| 2.1173 | 2.5 | 2288000 | 2.0062 |
|
| 334 |
+
| 2.1173 | 2.51 | 2296000 | 2.0039 |
|
| 335 |
+
| 2.114 | 2.51 | 2304000 | 2.0064 |
|
| 336 |
+
| 2.114 | 2.52 | 2312000 | 2.0113 |
|
| 337 |
+
| 2.1131 | 2.53 | 2320000 | 2.0065 |
|
| 338 |
+
| 2.1131 | 2.54 | 2328000 | 2.0098 |
|
| 339 |
+
| 2.1045 | 2.55 | 2336000 | 2.0061 |
|
| 340 |
+
| 2.1045 | 2.56 | 2344000 | 2.0066 |
|
| 341 |
+
| 2.1144 | 2.57 | 2352000 | 2.0060 |
|
| 342 |
+
| 2.1144 | 2.57 | 2360000 | 2.0059 |
|
| 343 |
+
| 2.1086 | 2.58 | 2368000 | 2.0039 |
|
| 344 |
+
| 2.1086 | 2.59 | 2376000 | 2.0076 |
|
| 345 |
+
| 2.1058 | 2.6 | 2384000 | 2.0036 |
|
| 346 |
+
| 2.1058 | 2.61 | 2392000 | 2.0077 |
|
| 347 |
+
| 2.1112 | 2.62 | 2400000 | 2.0000 |
|
| 348 |
+
|
| 349 |
+
|
| 350 |
+
### Framework versions
|
| 351 |
+
|
| 352 |
+
- Transformers 4.35.0.dev0
|
| 353 |
+
- Pytorch 2.0.1+cu117
|
| 354 |
+
- Datasets 2.14.5
|
| 355 |
+
- Tokenizers 0.14.0
|
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</s>": 2,
|
| 3 |
+
"<mask>": 50264,
|
| 4 |
+
"<pad>": 1,
|
| 5 |
+
"<s>": 0,
|
| 6 |
+
"<unk>": 3
|
| 7 |
+
}
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.62,
|
| 3 |
+
"eval_loss": 2.006091594696045,
|
| 4 |
+
"eval_runtime": 861.4578,
|
| 5 |
+
"eval_samples": 771820,
|
| 6 |
+
"eval_samples_per_second": 895.946,
|
| 7 |
+
"eval_steps_per_second": 55.997,
|
| 8 |
+
"perplexity": 7.4342046074668735,
|
| 9 |
+
"train_loss": 2.133689431966146,
|
| 10 |
+
"train_runtime": 416842.919,
|
| 11 |
+
"train_samples": 14664588,
|
| 12 |
+
"train_samples_per_second": 92.121,
|
| 13 |
+
"train_steps_per_second": 5.758
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "cardiffnlp/twitter-roberta-base-2019-90m",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"gradient_checkpointing": false,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_norm_eps": 1e-05,
|
| 17 |
+
"max_position_embeddings": 514,
|
| 18 |
+
"model_type": "roberta",
|
| 19 |
+
"num_attention_heads": 12,
|
| 20 |
+
"num_hidden_layers": 12,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"torch_dtype": "float32",
|
| 24 |
+
"transformers_version": "4.35.0.dev0",
|
| 25 |
+
"type_vocab_size": 1,
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 50265
|
| 28 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.62,
|
| 3 |
+
"eval_loss": 2.006091594696045,
|
| 4 |
+
"eval_runtime": 861.4578,
|
| 5 |
+
"eval_samples": 771820,
|
| 6 |
+
"eval_samples_per_second": 895.946,
|
| 7 |
+
"eval_steps_per_second": 55.997,
|
| 8 |
+
"perplexity": 7.4342046074668735
|
| 9 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b964d07d8a002c13f2bc4b13dbc200ca89473fb5d3efd1e6a118321be0d0c81d
|
| 3 |
+
size 498859189
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"additional_special_tokens": [],
|
| 46 |
+
"bos_token": "<s>",
|
| 47 |
+
"clean_up_tokenization_spaces": true,
|
| 48 |
+
"cls_token": "<s>",
|
| 49 |
+
"eos_token": "</s>",
|
| 50 |
+
"errors": "replace",
|
| 51 |
+
"mask_token": "<mask>",
|
| 52 |
+
"max_length": 512,
|
| 53 |
+
"model_max_length": 512,
|
| 54 |
+
"pad_token": "<pad>",
|
| 55 |
+
"sep_token": "</s>",
|
| 56 |
+
"stride": 0,
|
| 57 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 58 |
+
"trim_offsets": true,
|
| 59 |
+
"truncation_side": "right",
|
| 60 |
+
"truncation_strategy": "longest_first",
|
| 61 |
+
"unk_token": "<unk>"
|
| 62 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.62,
|
| 3 |
+
"train_loss": 2.133689431966146,
|
| 4 |
+
"train_runtime": 416842.919,
|
| 5 |
+
"train_samples": 14664588,
|
| 6 |
+
"train_samples_per_second": 92.121,
|
| 7 |
+
"train_steps_per_second": 5.758
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3328 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.000014066696167,
|
| 3 |
+
"best_model_checkpoint": "./model_tweets_2020_Q2_full/checkpoint-2400000",
|
| 4 |
+
"epoch": 2.618552224296455,
|
| 5 |
+
"eval_steps": 8000,
|
| 6 |
+
"global_step": 2400000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"eval_loss": 2.1043460369110107,
|
| 14 |
+
"eval_runtime": 841.4118,
|
| 15 |
+
"eval_samples_per_second": 917.292,
|
| 16 |
+
"eval_steps_per_second": 57.331,
|
| 17 |
+
"step": 8000
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.02,
|
| 21 |
+
"learning_rate": 4.0726666666666665e-07,
|
| 22 |
+
"loss": 2.2608,
|
| 23 |
+
"step": 16000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02,
|
| 27 |
+
"eval_loss": 2.093374013900757,
|
| 28 |
+
"eval_runtime": 839.5651,
|
| 29 |
+
"eval_samples_per_second": 919.309,
|
| 30 |
+
"eval_steps_per_second": 57.457,
|
| 31 |
+
"step": 16000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.03,
|
| 35 |
+
"eval_loss": 2.0861904621124268,
|
| 36 |
+
"eval_runtime": 840.327,
|
| 37 |
+
"eval_samples_per_second": 918.476,
|
| 38 |
+
"eval_steps_per_second": 57.405,
|
| 39 |
+
"step": 24000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 4.0453333333333336e-07,
|
| 44 |
+
"loss": 2.2409,
|
| 45 |
+
"step": 32000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"eval_loss": 2.080547332763672,
|
| 50 |
+
"eval_runtime": 841.1888,
|
| 51 |
+
"eval_samples_per_second": 917.535,
|
| 52 |
+
"eval_steps_per_second": 57.346,
|
| 53 |
+
"step": 32000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.04,
|
| 57 |
+
"eval_loss": 2.079263210296631,
|
| 58 |
+
"eval_runtime": 841.5773,
|
| 59 |
+
"eval_samples_per_second": 917.111,
|
| 60 |
+
"eval_steps_per_second": 57.32,
|
| 61 |
+
"step": 40000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.05,
|
| 65 |
+
"learning_rate": 4.018e-07,
|
| 66 |
+
"loss": 2.2278,
|
| 67 |
+
"step": 48000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.05,
|
| 71 |
+
"eval_loss": 2.071790933609009,
|
| 72 |
+
"eval_runtime": 841.5993,
|
| 73 |
+
"eval_samples_per_second": 917.087,
|
| 74 |
+
"eval_steps_per_second": 57.318,
|
| 75 |
+
"step": 48000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.06,
|
| 79 |
+
"eval_loss": 2.0752639770507812,
|
| 80 |
+
"eval_runtime": 841.4195,
|
| 81 |
+
"eval_samples_per_second": 917.283,
|
| 82 |
+
"eval_steps_per_second": 57.33,
|
| 83 |
+
"step": 56000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.07,
|
| 87 |
+
"learning_rate": 3.9906666666666667e-07,
|
| 88 |
+
"loss": 2.2059,
|
| 89 |
+
"step": 64000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.07,
|
| 93 |
+
"eval_loss": 2.0668046474456787,
|
| 94 |
+
"eval_runtime": 844.6529,
|
| 95 |
+
"eval_samples_per_second": 913.772,
|
| 96 |
+
"eval_steps_per_second": 57.111,
|
| 97 |
+
"step": 64000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.08,
|
| 101 |
+
"eval_loss": 2.0657169818878174,
|
| 102 |
+
"eval_runtime": 844.5291,
|
| 103 |
+
"eval_samples_per_second": 913.906,
|
| 104 |
+
"eval_steps_per_second": 57.119,
|
| 105 |
+
"step": 72000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.09,
|
| 109 |
+
"learning_rate": 3.963333333333333e-07,
|
| 110 |
+
"loss": 2.1997,
|
| 111 |
+
"step": 80000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.09,
|
| 115 |
+
"eval_loss": 2.062004566192627,
|
| 116 |
+
"eval_runtime": 845.6772,
|
| 117 |
+
"eval_samples_per_second": 912.665,
|
| 118 |
+
"eval_steps_per_second": 57.042,
|
| 119 |
+
"step": 80000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.1,
|
| 123 |
+
"eval_loss": 2.0553247928619385,
|
| 124 |
+
"eval_runtime": 846.4058,
|
| 125 |
+
"eval_samples_per_second": 911.879,
|
| 126 |
+
"eval_steps_per_second": 56.993,
|
| 127 |
+
"step": 88000
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.1,
|
| 131 |
+
"learning_rate": 3.936e-07,
|
| 132 |
+
"loss": 2.1988,
|
| 133 |
+
"step": 96000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 0.1,
|
| 137 |
+
"eval_loss": 2.0569465160369873,
|
| 138 |
+
"eval_runtime": 841.5807,
|
| 139 |
+
"eval_samples_per_second": 917.108,
|
| 140 |
+
"eval_steps_per_second": 57.32,
|
| 141 |
+
"step": 96000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.11,
|
| 145 |
+
"eval_loss": 2.052541971206665,
|
| 146 |
+
"eval_runtime": 845.3721,
|
| 147 |
+
"eval_samples_per_second": 912.994,
|
| 148 |
+
"eval_steps_per_second": 57.062,
|
| 149 |
+
"step": 104000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.12,
|
| 153 |
+
"learning_rate": 3.908666666666667e-07,
|
| 154 |
+
"loss": 2.1861,
|
| 155 |
+
"step": 112000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.12,
|
| 159 |
+
"eval_loss": 2.05564284324646,
|
| 160 |
+
"eval_runtime": 847.9385,
|
| 161 |
+
"eval_samples_per_second": 910.231,
|
| 162 |
+
"eval_steps_per_second": 56.89,
|
| 163 |
+
"step": 112000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.13,
|
| 167 |
+
"eval_loss": 2.04929256439209,
|
| 168 |
+
"eval_runtime": 842.77,
|
| 169 |
+
"eval_samples_per_second": 915.813,
|
| 170 |
+
"eval_steps_per_second": 57.239,
|
| 171 |
+
"step": 120000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.14,
|
| 175 |
+
"learning_rate": 3.8813333333333334e-07,
|
| 176 |
+
"loss": 2.1823,
|
| 177 |
+
"step": 128000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.14,
|
| 181 |
+
"eval_loss": 2.0508854389190674,
|
| 182 |
+
"eval_runtime": 846.1188,
|
| 183 |
+
"eval_samples_per_second": 912.189,
|
| 184 |
+
"eval_steps_per_second": 57.012,
|
| 185 |
+
"step": 128000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.15,
|
| 189 |
+
"eval_loss": 2.0460989475250244,
|
| 190 |
+
"eval_runtime": 845.0745,
|
| 191 |
+
"eval_samples_per_second": 913.316,
|
| 192 |
+
"eval_steps_per_second": 57.083,
|
| 193 |
+
"step": 136000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.16,
|
| 197 |
+
"learning_rate": 3.854e-07,
|
| 198 |
+
"loss": 2.1851,
|
| 199 |
+
"step": 144000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.16,
|
| 203 |
+
"eval_loss": 2.0476059913635254,
|
| 204 |
+
"eval_runtime": 844.5648,
|
| 205 |
+
"eval_samples_per_second": 913.867,
|
| 206 |
+
"eval_steps_per_second": 57.117,
|
| 207 |
+
"step": 144000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.17,
|
| 211 |
+
"eval_loss": 2.04502010345459,
|
| 212 |
+
"eval_runtime": 845.4437,
|
| 213 |
+
"eval_samples_per_second": 912.917,
|
| 214 |
+
"eval_steps_per_second": 57.058,
|
| 215 |
+
"step": 152000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.17,
|
| 219 |
+
"learning_rate": 3.8266666666666665e-07,
|
| 220 |
+
"loss": 2.1862,
|
| 221 |
+
"step": 160000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.17,
|
| 225 |
+
"eval_loss": 2.046872854232788,
|
| 226 |
+
"eval_runtime": 843.2469,
|
| 227 |
+
"eval_samples_per_second": 915.295,
|
| 228 |
+
"eval_steps_per_second": 57.206,
|
| 229 |
+
"step": 160000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.18,
|
| 233 |
+
"eval_loss": 2.0441744327545166,
|
| 234 |
+
"eval_runtime": 845.6614,
|
| 235 |
+
"eval_samples_per_second": 912.682,
|
| 236 |
+
"eval_steps_per_second": 57.043,
|
| 237 |
+
"step": 168000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.19,
|
| 241 |
+
"learning_rate": 3.799333333333333e-07,
|
| 242 |
+
"loss": 2.1741,
|
| 243 |
+
"step": 176000
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.19,
|
| 247 |
+
"eval_loss": 2.0456435680389404,
|
| 248 |
+
"eval_runtime": 847.979,
|
| 249 |
+
"eval_samples_per_second": 910.188,
|
| 250 |
+
"eval_steps_per_second": 56.887,
|
| 251 |
+
"step": 176000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.2,
|
| 255 |
+
"eval_loss": 2.0441741943359375,
|
| 256 |
+
"eval_runtime": 846.0243,
|
| 257 |
+
"eval_samples_per_second": 912.291,
|
| 258 |
+
"eval_steps_per_second": 57.018,
|
| 259 |
+
"step": 184000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.21,
|
| 263 |
+
"learning_rate": 3.772e-07,
|
| 264 |
+
"loss": 2.181,
|
| 265 |
+
"step": 192000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.21,
|
| 269 |
+
"eval_loss": 2.040196418762207,
|
| 270 |
+
"eval_runtime": 851.9304,
|
| 271 |
+
"eval_samples_per_second": 905.966,
|
| 272 |
+
"eval_steps_per_second": 56.623,
|
| 273 |
+
"step": 192000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.22,
|
| 277 |
+
"eval_loss": 2.0422918796539307,
|
| 278 |
+
"eval_runtime": 847.5458,
|
| 279 |
+
"eval_samples_per_second": 910.653,
|
| 280 |
+
"eval_steps_per_second": 56.916,
|
| 281 |
+
"step": 200000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.23,
|
| 285 |
+
"learning_rate": 3.7446666666666667e-07,
|
| 286 |
+
"loss": 2.1692,
|
| 287 |
+
"step": 208000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.23,
|
| 291 |
+
"eval_loss": 2.041342258453369,
|
| 292 |
+
"eval_runtime": 847.0952,
|
| 293 |
+
"eval_samples_per_second": 911.137,
|
| 294 |
+
"eval_steps_per_second": 56.946,
|
| 295 |
+
"step": 208000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.24,
|
| 299 |
+
"eval_loss": 2.0448198318481445,
|
| 300 |
+
"eval_runtime": 846.694,
|
| 301 |
+
"eval_samples_per_second": 911.569,
|
| 302 |
+
"eval_steps_per_second": 56.973,
|
| 303 |
+
"step": 216000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24,
|
| 307 |
+
"learning_rate": 3.7173333333333333e-07,
|
| 308 |
+
"loss": 2.1678,
|
| 309 |
+
"step": 224000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.24,
|
| 313 |
+
"eval_loss": 2.0417792797088623,
|
| 314 |
+
"eval_runtime": 846.2034,
|
| 315 |
+
"eval_samples_per_second": 912.098,
|
| 316 |
+
"eval_steps_per_second": 57.006,
|
| 317 |
+
"step": 224000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.25,
|
| 321 |
+
"eval_loss": 2.041692018508911,
|
| 322 |
+
"eval_runtime": 848.6147,
|
| 323 |
+
"eval_samples_per_second": 909.506,
|
| 324 |
+
"eval_steps_per_second": 56.844,
|
| 325 |
+
"step": 232000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.26,
|
| 329 |
+
"learning_rate": 3.69e-07,
|
| 330 |
+
"loss": 2.1756,
|
| 331 |
+
"step": 240000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.26,
|
| 335 |
+
"eval_loss": 2.034193754196167,
|
| 336 |
+
"eval_runtime": 847.0585,
|
| 337 |
+
"eval_samples_per_second": 911.177,
|
| 338 |
+
"eval_steps_per_second": 56.949,
|
| 339 |
+
"step": 240000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.27,
|
| 343 |
+
"eval_loss": 2.037684202194214,
|
| 344 |
+
"eval_runtime": 846.2239,
|
| 345 |
+
"eval_samples_per_second": 912.075,
|
| 346 |
+
"eval_steps_per_second": 57.005,
|
| 347 |
+
"step": 248000
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.28,
|
| 351 |
+
"learning_rate": 3.6626666666666664e-07,
|
| 352 |
+
"loss": 2.1752,
|
| 353 |
+
"step": 256000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.28,
|
| 357 |
+
"eval_loss": 2.0381019115448,
|
| 358 |
+
"eval_runtime": 846.7119,
|
| 359 |
+
"eval_samples_per_second": 911.55,
|
| 360 |
+
"eval_steps_per_second": 56.972,
|
| 361 |
+
"step": 256000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.29,
|
| 365 |
+
"eval_loss": 2.035405158996582,
|
| 366 |
+
"eval_runtime": 851.4653,
|
| 367 |
+
"eval_samples_per_second": 906.461,
|
| 368 |
+
"eval_steps_per_second": 56.654,
|
| 369 |
+
"step": 264000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.3,
|
| 373 |
+
"learning_rate": 3.6353333333333335e-07,
|
| 374 |
+
"loss": 2.1673,
|
| 375 |
+
"step": 272000
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.3,
|
| 379 |
+
"eval_loss": 2.038097858428955,
|
| 380 |
+
"eval_runtime": 846.955,
|
| 381 |
+
"eval_samples_per_second": 911.288,
|
| 382 |
+
"eval_steps_per_second": 56.956,
|
| 383 |
+
"step": 272000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.31,
|
| 387 |
+
"eval_loss": 2.0375349521636963,
|
| 388 |
+
"eval_runtime": 846.7581,
|
| 389 |
+
"eval_samples_per_second": 911.5,
|
| 390 |
+
"eval_steps_per_second": 56.969,
|
| 391 |
+
"step": 280000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.31,
|
| 395 |
+
"learning_rate": 3.608e-07,
|
| 396 |
+
"loss": 2.1585,
|
| 397 |
+
"step": 288000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 0.31,
|
| 401 |
+
"eval_loss": 2.033590078353882,
|
| 402 |
+
"eval_runtime": 848.1336,
|
| 403 |
+
"eval_samples_per_second": 910.022,
|
| 404 |
+
"eval_steps_per_second": 56.877,
|
| 405 |
+
"step": 288000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.32,
|
| 409 |
+
"eval_loss": 2.0344314575195312,
|
| 410 |
+
"eval_runtime": 847.2304,
|
| 411 |
+
"eval_samples_per_second": 910.992,
|
| 412 |
+
"eval_steps_per_second": 56.937,
|
| 413 |
+
"step": 296000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.33,
|
| 417 |
+
"learning_rate": 3.5806666666666666e-07,
|
| 418 |
+
"loss": 2.1703,
|
| 419 |
+
"step": 304000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.33,
|
| 423 |
+
"eval_loss": 2.034810781478882,
|
| 424 |
+
"eval_runtime": 846.3544,
|
| 425 |
+
"eval_samples_per_second": 911.935,
|
| 426 |
+
"eval_steps_per_second": 56.996,
|
| 427 |
+
"step": 304000
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 0.34,
|
| 431 |
+
"eval_loss": 2.0329954624176025,
|
| 432 |
+
"eval_runtime": 847.3997,
|
| 433 |
+
"eval_samples_per_second": 910.81,
|
| 434 |
+
"eval_steps_per_second": 56.926,
|
| 435 |
+
"step": 312000
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.35,
|
| 439 |
+
"learning_rate": 3.553333333333333e-07,
|
| 440 |
+
"loss": 2.1667,
|
| 441 |
+
"step": 320000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.35,
|
| 445 |
+
"eval_loss": 2.0352213382720947,
|
| 446 |
+
"eval_runtime": 846.3586,
|
| 447 |
+
"eval_samples_per_second": 911.93,
|
| 448 |
+
"eval_steps_per_second": 56.996,
|
| 449 |
+
"step": 320000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.36,
|
| 453 |
+
"eval_loss": 2.0359089374542236,
|
| 454 |
+
"eval_runtime": 848.8487,
|
| 455 |
+
"eval_samples_per_second": 909.255,
|
| 456 |
+
"eval_steps_per_second": 56.829,
|
| 457 |
+
"step": 328000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.37,
|
| 461 |
+
"learning_rate": 3.5259999999999997e-07,
|
| 462 |
+
"loss": 2.1649,
|
| 463 |
+
"step": 336000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 0.37,
|
| 467 |
+
"eval_loss": 2.031733512878418,
|
| 468 |
+
"eval_runtime": 848.6246,
|
| 469 |
+
"eval_samples_per_second": 909.495,
|
| 470 |
+
"eval_steps_per_second": 56.844,
|
| 471 |
+
"step": 336000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.38,
|
| 475 |
+
"eval_loss": 2.0314059257507324,
|
| 476 |
+
"eval_runtime": 851.4761,
|
| 477 |
+
"eval_samples_per_second": 906.449,
|
| 478 |
+
"eval_steps_per_second": 56.653,
|
| 479 |
+
"step": 344000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.38,
|
| 483 |
+
"learning_rate": 3.498666666666667e-07,
|
| 484 |
+
"loss": 2.1564,
|
| 485 |
+
"step": 352000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.38,
|
| 489 |
+
"eval_loss": 2.030597686767578,
|
| 490 |
+
"eval_runtime": 850.4287,
|
| 491 |
+
"eval_samples_per_second": 907.566,
|
| 492 |
+
"eval_steps_per_second": 56.723,
|
| 493 |
+
"step": 352000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.39,
|
| 497 |
+
"eval_loss": 2.029878616333008,
|
| 498 |
+
"eval_runtime": 850.6967,
|
| 499 |
+
"eval_samples_per_second": 907.28,
|
| 500 |
+
"eval_steps_per_second": 56.705,
|
| 501 |
+
"step": 360000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.4,
|
| 505 |
+
"learning_rate": 3.4713333333333333e-07,
|
| 506 |
+
"loss": 2.161,
|
| 507 |
+
"step": 368000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.4,
|
| 511 |
+
"eval_loss": 2.0317320823669434,
|
| 512 |
+
"eval_runtime": 851.1347,
|
| 513 |
+
"eval_samples_per_second": 906.813,
|
| 514 |
+
"eval_steps_per_second": 56.676,
|
| 515 |
+
"step": 368000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.41,
|
| 519 |
+
"eval_loss": 2.032505989074707,
|
| 520 |
+
"eval_runtime": 854.4271,
|
| 521 |
+
"eval_samples_per_second": 903.319,
|
| 522 |
+
"eval_steps_per_second": 56.458,
|
| 523 |
+
"step": 376000
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 0.42,
|
| 527 |
+
"learning_rate": 3.444e-07,
|
| 528 |
+
"loss": 2.1551,
|
| 529 |
+
"step": 384000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.42,
|
| 533 |
+
"eval_loss": 2.0273916721343994,
|
| 534 |
+
"eval_runtime": 850.6274,
|
| 535 |
+
"eval_samples_per_second": 907.354,
|
| 536 |
+
"eval_steps_per_second": 56.71,
|
| 537 |
+
"step": 384000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.43,
|
| 541 |
+
"eval_loss": 2.0281741619110107,
|
| 542 |
+
"eval_runtime": 850.1523,
|
| 543 |
+
"eval_samples_per_second": 907.861,
|
| 544 |
+
"eval_steps_per_second": 56.742,
|
| 545 |
+
"step": 392000
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.44,
|
| 549 |
+
"learning_rate": 3.416666666666667e-07,
|
| 550 |
+
"loss": 2.1602,
|
| 551 |
+
"step": 400000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 0.44,
|
| 555 |
+
"eval_loss": 2.0300543308258057,
|
| 556 |
+
"eval_runtime": 852.8839,
|
| 557 |
+
"eval_samples_per_second": 904.953,
|
| 558 |
+
"eval_steps_per_second": 56.56,
|
| 559 |
+
"step": 400000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 0.45,
|
| 563 |
+
"eval_loss": 2.0302786827087402,
|
| 564 |
+
"eval_runtime": 854.7636,
|
| 565 |
+
"eval_samples_per_second": 902.963,
|
| 566 |
+
"eval_steps_per_second": 56.435,
|
| 567 |
+
"step": 408000
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.45,
|
| 571 |
+
"learning_rate": 3.3893333333333335e-07,
|
| 572 |
+
"loss": 2.1581,
|
| 573 |
+
"step": 416000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.45,
|
| 577 |
+
"eval_loss": 2.026031732559204,
|
| 578 |
+
"eval_runtime": 852.2087,
|
| 579 |
+
"eval_samples_per_second": 905.67,
|
| 580 |
+
"eval_steps_per_second": 56.605,
|
| 581 |
+
"step": 416000
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.46,
|
| 585 |
+
"eval_loss": 2.0248208045959473,
|
| 586 |
+
"eval_runtime": 850.4117,
|
| 587 |
+
"eval_samples_per_second": 907.584,
|
| 588 |
+
"eval_steps_per_second": 56.724,
|
| 589 |
+
"step": 424000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.47,
|
| 593 |
+
"learning_rate": 3.3619999999999995e-07,
|
| 594 |
+
"loss": 2.1494,
|
| 595 |
+
"step": 432000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 0.47,
|
| 599 |
+
"eval_loss": 2.026501178741455,
|
| 600 |
+
"eval_runtime": 848.7671,
|
| 601 |
+
"eval_samples_per_second": 909.343,
|
| 602 |
+
"eval_steps_per_second": 56.834,
|
| 603 |
+
"step": 432000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.48,
|
| 607 |
+
"eval_loss": 2.0246880054473877,
|
| 608 |
+
"eval_runtime": 849.7267,
|
| 609 |
+
"eval_samples_per_second": 908.316,
|
| 610 |
+
"eval_steps_per_second": 56.77,
|
| 611 |
+
"step": 440000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.49,
|
| 615 |
+
"learning_rate": 3.3346666666666666e-07,
|
| 616 |
+
"loss": 2.1508,
|
| 617 |
+
"step": 448000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.49,
|
| 621 |
+
"eval_loss": 2.0231027603149414,
|
| 622 |
+
"eval_runtime": 849.0484,
|
| 623 |
+
"eval_samples_per_second": 909.041,
|
| 624 |
+
"eval_steps_per_second": 56.815,
|
| 625 |
+
"step": 448000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5,
|
| 629 |
+
"eval_loss": 2.0276315212249756,
|
| 630 |
+
"eval_runtime": 849.4168,
|
| 631 |
+
"eval_samples_per_second": 908.647,
|
| 632 |
+
"eval_steps_per_second": 56.791,
|
| 633 |
+
"step": 456000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.51,
|
| 637 |
+
"learning_rate": 3.307333333333333e-07,
|
| 638 |
+
"loss": 2.153,
|
| 639 |
+
"step": 464000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.51,
|
| 643 |
+
"eval_loss": 2.0275754928588867,
|
| 644 |
+
"eval_runtime": 848.4629,
|
| 645 |
+
"eval_samples_per_second": 909.669,
|
| 646 |
+
"eval_steps_per_second": 56.855,
|
| 647 |
+
"step": 464000
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.51,
|
| 651 |
+
"eval_loss": 2.0241763591766357,
|
| 652 |
+
"eval_runtime": 849.6091,
|
| 653 |
+
"eval_samples_per_second": 908.441,
|
| 654 |
+
"eval_steps_per_second": 56.778,
|
| 655 |
+
"step": 472000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.52,
|
| 659 |
+
"learning_rate": 3.28e-07,
|
| 660 |
+
"loss": 2.1489,
|
| 661 |
+
"step": 480000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.52,
|
| 665 |
+
"eval_loss": 2.0259480476379395,
|
| 666 |
+
"eval_runtime": 849.4664,
|
| 667 |
+
"eval_samples_per_second": 908.594,
|
| 668 |
+
"eval_steps_per_second": 56.787,
|
| 669 |
+
"step": 480000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.53,
|
| 673 |
+
"eval_loss": 2.025740623474121,
|
| 674 |
+
"eval_runtime": 850.1732,
|
| 675 |
+
"eval_samples_per_second": 907.839,
|
| 676 |
+
"eval_steps_per_second": 56.74,
|
| 677 |
+
"step": 488000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.54,
|
| 681 |
+
"learning_rate": 3.252666666666667e-07,
|
| 682 |
+
"loss": 2.1468,
|
| 683 |
+
"step": 496000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.54,
|
| 687 |
+
"eval_loss": 2.027461528778076,
|
| 688 |
+
"eval_runtime": 850.2923,
|
| 689 |
+
"eval_samples_per_second": 907.711,
|
| 690 |
+
"eval_steps_per_second": 56.732,
|
| 691 |
+
"step": 496000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.55,
|
| 695 |
+
"eval_loss": 2.030271053314209,
|
| 696 |
+
"eval_runtime": 851.4114,
|
| 697 |
+
"eval_samples_per_second": 906.518,
|
| 698 |
+
"eval_steps_per_second": 56.658,
|
| 699 |
+
"step": 504000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.56,
|
| 703 |
+
"learning_rate": 3.2253333333333334e-07,
|
| 704 |
+
"loss": 2.1446,
|
| 705 |
+
"step": 512000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.56,
|
| 709 |
+
"eval_loss": 2.0248193740844727,
|
| 710 |
+
"eval_runtime": 852.1182,
|
| 711 |
+
"eval_samples_per_second": 905.766,
|
| 712 |
+
"eval_steps_per_second": 56.611,
|
| 713 |
+
"step": 512000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 0.57,
|
| 717 |
+
"eval_loss": 2.0285604000091553,
|
| 718 |
+
"eval_runtime": 849.8013,
|
| 719 |
+
"eval_samples_per_second": 908.236,
|
| 720 |
+
"eval_steps_per_second": 56.765,
|
| 721 |
+
"step": 520000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.58,
|
| 725 |
+
"learning_rate": 3.198e-07,
|
| 726 |
+
"loss": 2.1409,
|
| 727 |
+
"step": 528000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.58,
|
| 731 |
+
"eval_loss": 2.0211498737335205,
|
| 732 |
+
"eval_runtime": 855.0597,
|
| 733 |
+
"eval_samples_per_second": 902.65,
|
| 734 |
+
"eval_steps_per_second": 56.416,
|
| 735 |
+
"step": 528000
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.58,
|
| 739 |
+
"eval_loss": 2.0204012393951416,
|
| 740 |
+
"eval_runtime": 856.0145,
|
| 741 |
+
"eval_samples_per_second": 901.644,
|
| 742 |
+
"eval_steps_per_second": 56.353,
|
| 743 |
+
"step": 536000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 0.59,
|
| 747 |
+
"learning_rate": 3.1706666666666665e-07,
|
| 748 |
+
"loss": 2.1536,
|
| 749 |
+
"step": 544000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 0.59,
|
| 753 |
+
"eval_loss": 2.0198850631713867,
|
| 754 |
+
"eval_runtime": 856.7067,
|
| 755 |
+
"eval_samples_per_second": 900.915,
|
| 756 |
+
"eval_steps_per_second": 56.307,
|
| 757 |
+
"step": 544000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 0.6,
|
| 761 |
+
"eval_loss": 2.0281307697296143,
|
| 762 |
+
"eval_runtime": 867.0343,
|
| 763 |
+
"eval_samples_per_second": 890.184,
|
| 764 |
+
"eval_steps_per_second": 55.637,
|
| 765 |
+
"step": 552000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.61,
|
| 769 |
+
"learning_rate": 3.1433333333333336e-07,
|
| 770 |
+
"loss": 2.1416,
|
| 771 |
+
"step": 560000
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.61,
|
| 775 |
+
"eval_loss": 2.0237483978271484,
|
| 776 |
+
"eval_runtime": 866.1166,
|
| 777 |
+
"eval_samples_per_second": 891.127,
|
| 778 |
+
"eval_steps_per_second": 55.696,
|
| 779 |
+
"step": 560000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.62,
|
| 783 |
+
"eval_loss": 2.0231337547302246,
|
| 784 |
+
"eval_runtime": 863.3507,
|
| 785 |
+
"eval_samples_per_second": 893.982,
|
| 786 |
+
"eval_steps_per_second": 55.874,
|
| 787 |
+
"step": 568000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.63,
|
| 791 |
+
"learning_rate": 3.116e-07,
|
| 792 |
+
"loss": 2.1502,
|
| 793 |
+
"step": 576000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.63,
|
| 797 |
+
"eval_loss": 2.0205323696136475,
|
| 798 |
+
"eval_runtime": 857.8171,
|
| 799 |
+
"eval_samples_per_second": 899.749,
|
| 800 |
+
"eval_steps_per_second": 56.235,
|
| 801 |
+
"step": 576000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.64,
|
| 805 |
+
"eval_loss": 2.021655559539795,
|
| 806 |
+
"eval_runtime": 853.6943,
|
| 807 |
+
"eval_samples_per_second": 904.094,
|
| 808 |
+
"eval_steps_per_second": 56.506,
|
| 809 |
+
"step": 584000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.65,
|
| 813 |
+
"learning_rate": 3.0886666666666667e-07,
|
| 814 |
+
"loss": 2.1424,
|
| 815 |
+
"step": 592000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.65,
|
| 819 |
+
"eval_loss": 2.024162769317627,
|
| 820 |
+
"eval_runtime": 861.2895,
|
| 821 |
+
"eval_samples_per_second": 896.121,
|
| 822 |
+
"eval_steps_per_second": 56.008,
|
| 823 |
+
"step": 592000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.65,
|
| 827 |
+
"eval_loss": 2.0237643718719482,
|
| 828 |
+
"eval_runtime": 859.5317,
|
| 829 |
+
"eval_samples_per_second": 897.954,
|
| 830 |
+
"eval_steps_per_second": 56.122,
|
| 831 |
+
"step": 600000
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.66,
|
| 835 |
+
"learning_rate": 3.061333333333333e-07,
|
| 836 |
+
"loss": 2.1469,
|
| 837 |
+
"step": 608000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.66,
|
| 841 |
+
"eval_loss": 2.0191547870635986,
|
| 842 |
+
"eval_runtime": 855.9495,
|
| 843 |
+
"eval_samples_per_second": 901.712,
|
| 844 |
+
"eval_steps_per_second": 56.357,
|
| 845 |
+
"step": 608000
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.67,
|
| 849 |
+
"eval_loss": 2.024866819381714,
|
| 850 |
+
"eval_runtime": 857.0469,
|
| 851 |
+
"eval_samples_per_second": 900.557,
|
| 852 |
+
"eval_steps_per_second": 56.285,
|
| 853 |
+
"step": 616000
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.68,
|
| 857 |
+
"learning_rate": 3.034e-07,
|
| 858 |
+
"loss": 2.145,
|
| 859 |
+
"step": 624000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.68,
|
| 863 |
+
"eval_loss": 2.0195770263671875,
|
| 864 |
+
"eval_runtime": 858.8544,
|
| 865 |
+
"eval_samples_per_second": 898.662,
|
| 866 |
+
"eval_steps_per_second": 56.167,
|
| 867 |
+
"step": 624000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.69,
|
| 871 |
+
"eval_loss": 2.022365093231201,
|
| 872 |
+
"eval_runtime": 854.0414,
|
| 873 |
+
"eval_samples_per_second": 903.727,
|
| 874 |
+
"eval_steps_per_second": 56.483,
|
| 875 |
+
"step": 632000
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 0.7,
|
| 879 |
+
"learning_rate": 3.0066666666666663e-07,
|
| 880 |
+
"loss": 2.1503,
|
| 881 |
+
"step": 640000
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 0.7,
|
| 885 |
+
"eval_loss": 2.0216493606567383,
|
| 886 |
+
"eval_runtime": 854.8203,
|
| 887 |
+
"eval_samples_per_second": 902.903,
|
| 888 |
+
"eval_steps_per_second": 56.432,
|
| 889 |
+
"step": 640000
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.71,
|
| 893 |
+
"eval_loss": 2.022836208343506,
|
| 894 |
+
"eval_runtime": 857.6145,
|
| 895 |
+
"eval_samples_per_second": 899.962,
|
| 896 |
+
"eval_steps_per_second": 56.248,
|
| 897 |
+
"step": 648000
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.72,
|
| 901 |
+
"learning_rate": 2.9793333333333334e-07,
|
| 902 |
+
"loss": 2.1355,
|
| 903 |
+
"step": 656000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.72,
|
| 907 |
+
"eval_loss": 2.019666910171509,
|
| 908 |
+
"eval_runtime": 859.7029,
|
| 909 |
+
"eval_samples_per_second": 897.775,
|
| 910 |
+
"eval_steps_per_second": 56.111,
|
| 911 |
+
"step": 656000
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 0.72,
|
| 915 |
+
"eval_loss": 2.0240182876586914,
|
| 916 |
+
"eval_runtime": 858.0715,
|
| 917 |
+
"eval_samples_per_second": 899.482,
|
| 918 |
+
"eval_steps_per_second": 56.218,
|
| 919 |
+
"step": 664000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.73,
|
| 923 |
+
"learning_rate": 2.952e-07,
|
| 924 |
+
"loss": 2.1392,
|
| 925 |
+
"step": 672000
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 0.73,
|
| 929 |
+
"eval_loss": 2.0232093334198,
|
| 930 |
+
"eval_runtime": 856.593,
|
| 931 |
+
"eval_samples_per_second": 901.035,
|
| 932 |
+
"eval_steps_per_second": 56.315,
|
| 933 |
+
"step": 672000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.74,
|
| 937 |
+
"eval_loss": 2.020932912826538,
|
| 938 |
+
"eval_runtime": 858.8309,
|
| 939 |
+
"eval_samples_per_second": 898.687,
|
| 940 |
+
"eval_steps_per_second": 56.168,
|
| 941 |
+
"step": 680000
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.75,
|
| 945 |
+
"learning_rate": 2.9246666666666665e-07,
|
| 946 |
+
"loss": 2.1378,
|
| 947 |
+
"step": 688000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.75,
|
| 951 |
+
"eval_loss": 2.0219063758850098,
|
| 952 |
+
"eval_runtime": 860.0126,
|
| 953 |
+
"eval_samples_per_second": 897.452,
|
| 954 |
+
"eval_steps_per_second": 56.091,
|
| 955 |
+
"step": 688000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.76,
|
| 959 |
+
"eval_loss": 2.019192695617676,
|
| 960 |
+
"eval_runtime": 861.8149,
|
| 961 |
+
"eval_samples_per_second": 895.575,
|
| 962 |
+
"eval_steps_per_second": 55.974,
|
| 963 |
+
"step": 696000
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.77,
|
| 967 |
+
"learning_rate": 2.897333333333333e-07,
|
| 968 |
+
"loss": 2.1446,
|
| 969 |
+
"step": 704000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.77,
|
| 973 |
+
"eval_loss": 2.0194740295410156,
|
| 974 |
+
"eval_runtime": 857.8914,
|
| 975 |
+
"eval_samples_per_second": 899.671,
|
| 976 |
+
"eval_steps_per_second": 56.23,
|
| 977 |
+
"step": 704000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 0.78,
|
| 981 |
+
"eval_loss": 2.01971173286438,
|
| 982 |
+
"eval_runtime": 857.8638,
|
| 983 |
+
"eval_samples_per_second": 899.7,
|
| 984 |
+
"eval_steps_per_second": 56.232,
|
| 985 |
+
"step": 712000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.79,
|
| 989 |
+
"learning_rate": 2.8699999999999996e-07,
|
| 990 |
+
"loss": 2.1351,
|
| 991 |
+
"step": 720000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.79,
|
| 995 |
+
"eval_loss": 2.0183634757995605,
|
| 996 |
+
"eval_runtime": 857.8713,
|
| 997 |
+
"eval_samples_per_second": 899.692,
|
| 998 |
+
"eval_steps_per_second": 56.231,
|
| 999 |
+
"step": 720000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.79,
|
| 1003 |
+
"eval_loss": 2.0162270069122314,
|
| 1004 |
+
"eval_runtime": 857.9238,
|
| 1005 |
+
"eval_samples_per_second": 899.637,
|
| 1006 |
+
"eval_steps_per_second": 56.228,
|
| 1007 |
+
"step": 728000
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 0.8,
|
| 1011 |
+
"learning_rate": 2.8426666666666667e-07,
|
| 1012 |
+
"loss": 2.1437,
|
| 1013 |
+
"step": 736000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.8,
|
| 1017 |
+
"eval_loss": 2.015068531036377,
|
| 1018 |
+
"eval_runtime": 857.7851,
|
| 1019 |
+
"eval_samples_per_second": 899.783,
|
| 1020 |
+
"eval_steps_per_second": 56.237,
|
| 1021 |
+
"step": 736000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.81,
|
| 1025 |
+
"eval_loss": 2.0202245712280273,
|
| 1026 |
+
"eval_runtime": 857.6732,
|
| 1027 |
+
"eval_samples_per_second": 899.9,
|
| 1028 |
+
"eval_steps_per_second": 56.244,
|
| 1029 |
+
"step": 744000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.82,
|
| 1033 |
+
"learning_rate": 2.815333333333333e-07,
|
| 1034 |
+
"loss": 2.1249,
|
| 1035 |
+
"step": 752000
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.82,
|
| 1039 |
+
"eval_loss": 2.0169003009796143,
|
| 1040 |
+
"eval_runtime": 860.8823,
|
| 1041 |
+
"eval_samples_per_second": 896.545,
|
| 1042 |
+
"eval_steps_per_second": 56.034,
|
| 1043 |
+
"step": 752000
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.83,
|
| 1047 |
+
"eval_loss": 2.018857002258301,
|
| 1048 |
+
"eval_runtime": 856.9399,
|
| 1049 |
+
"eval_samples_per_second": 900.67,
|
| 1050 |
+
"eval_steps_per_second": 56.292,
|
| 1051 |
+
"step": 760000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.84,
|
| 1055 |
+
"learning_rate": 2.7880000000000003e-07,
|
| 1056 |
+
"loss": 2.1355,
|
| 1057 |
+
"step": 768000
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.84,
|
| 1061 |
+
"eval_loss": 2.022115707397461,
|
| 1062 |
+
"eval_runtime": 860.0914,
|
| 1063 |
+
"eval_samples_per_second": 897.37,
|
| 1064 |
+
"eval_steps_per_second": 56.086,
|
| 1065 |
+
"step": 768000
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.85,
|
| 1069 |
+
"eval_loss": 2.0194284915924072,
|
| 1070 |
+
"eval_runtime": 858.1451,
|
| 1071 |
+
"eval_samples_per_second": 899.405,
|
| 1072 |
+
"eval_steps_per_second": 56.213,
|
| 1073 |
+
"step": 776000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.86,
|
| 1077 |
+
"learning_rate": 2.7606666666666664e-07,
|
| 1078 |
+
"loss": 2.1387,
|
| 1079 |
+
"step": 784000
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 0.86,
|
| 1083 |
+
"eval_loss": 2.018942356109619,
|
| 1084 |
+
"eval_runtime": 862.7177,
|
| 1085 |
+
"eval_samples_per_second": 894.638,
|
| 1086 |
+
"eval_steps_per_second": 55.915,
|
| 1087 |
+
"step": 784000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.86,
|
| 1091 |
+
"eval_loss": 2.016535520553589,
|
| 1092 |
+
"eval_runtime": 858.1148,
|
| 1093 |
+
"eval_samples_per_second": 899.437,
|
| 1094 |
+
"eval_steps_per_second": 56.215,
|
| 1095 |
+
"step": 792000
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.87,
|
| 1099 |
+
"learning_rate": 2.733333333333333e-07,
|
| 1100 |
+
"loss": 2.1334,
|
| 1101 |
+
"step": 800000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.87,
|
| 1105 |
+
"eval_loss": 2.0169451236724854,
|
| 1106 |
+
"eval_runtime": 860.3041,
|
| 1107 |
+
"eval_samples_per_second": 897.148,
|
| 1108 |
+
"eval_steps_per_second": 56.072,
|
| 1109 |
+
"step": 800000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.88,
|
| 1113 |
+
"eval_loss": 2.0188918113708496,
|
| 1114 |
+
"eval_runtime": 861.004,
|
| 1115 |
+
"eval_samples_per_second": 896.419,
|
| 1116 |
+
"eval_steps_per_second": 56.026,
|
| 1117 |
+
"step": 808000
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.89,
|
| 1121 |
+
"learning_rate": 2.706e-07,
|
| 1122 |
+
"loss": 2.137,
|
| 1123 |
+
"step": 816000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.89,
|
| 1127 |
+
"eval_loss": 2.016237258911133,
|
| 1128 |
+
"eval_runtime": 862.2544,
|
| 1129 |
+
"eval_samples_per_second": 895.119,
|
| 1130 |
+
"eval_steps_per_second": 55.945,
|
| 1131 |
+
"step": 816000
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.9,
|
| 1135 |
+
"eval_loss": 2.0168325901031494,
|
| 1136 |
+
"eval_runtime": 860.8877,
|
| 1137 |
+
"eval_samples_per_second": 896.54,
|
| 1138 |
+
"eval_steps_per_second": 56.034,
|
| 1139 |
+
"step": 824000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 0.91,
|
| 1143 |
+
"learning_rate": 2.6786666666666666e-07,
|
| 1144 |
+
"loss": 2.1331,
|
| 1145 |
+
"step": 832000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 0.91,
|
| 1149 |
+
"eval_loss": 2.0192737579345703,
|
| 1150 |
+
"eval_runtime": 859.4597,
|
| 1151 |
+
"eval_samples_per_second": 898.029,
|
| 1152 |
+
"eval_steps_per_second": 56.127,
|
| 1153 |
+
"step": 832000
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 0.92,
|
| 1157 |
+
"eval_loss": 2.016619920730591,
|
| 1158 |
+
"eval_runtime": 863.1851,
|
| 1159 |
+
"eval_samples_per_second": 894.153,
|
| 1160 |
+
"eval_steps_per_second": 55.885,
|
| 1161 |
+
"step": 840000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.93,
|
| 1165 |
+
"learning_rate": 2.651333333333333e-07,
|
| 1166 |
+
"loss": 2.1293,
|
| 1167 |
+
"step": 848000
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 0.93,
|
| 1171 |
+
"eval_loss": 2.013720989227295,
|
| 1172 |
+
"eval_runtime": 863.4541,
|
| 1173 |
+
"eval_samples_per_second": 893.875,
|
| 1174 |
+
"eval_steps_per_second": 55.867,
|
| 1175 |
+
"step": 848000
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.93,
|
| 1179 |
+
"eval_loss": 2.018291711807251,
|
| 1180 |
+
"eval_runtime": 877.742,
|
| 1181 |
+
"eval_samples_per_second": 879.324,
|
| 1182 |
+
"eval_steps_per_second": 54.958,
|
| 1183 |
+
"step": 856000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.94,
|
| 1187 |
+
"learning_rate": 2.624e-07,
|
| 1188 |
+
"loss": 2.1358,
|
| 1189 |
+
"step": 864000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.94,
|
| 1193 |
+
"eval_loss": 2.018421173095703,
|
| 1194 |
+
"eval_runtime": 873.6563,
|
| 1195 |
+
"eval_samples_per_second": 883.437,
|
| 1196 |
+
"eval_steps_per_second": 55.215,
|
| 1197 |
+
"step": 864000
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.95,
|
| 1201 |
+
"eval_loss": 2.017104387283325,
|
| 1202 |
+
"eval_runtime": 874.261,
|
| 1203 |
+
"eval_samples_per_second": 882.826,
|
| 1204 |
+
"eval_steps_per_second": 55.177,
|
| 1205 |
+
"step": 872000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 0.96,
|
| 1209 |
+
"learning_rate": 2.596666666666667e-07,
|
| 1210 |
+
"loss": 2.1296,
|
| 1211 |
+
"step": 880000
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.96,
|
| 1215 |
+
"eval_loss": 2.0179190635681152,
|
| 1216 |
+
"eval_runtime": 874.7051,
|
| 1217 |
+
"eval_samples_per_second": 882.377,
|
| 1218 |
+
"eval_steps_per_second": 55.149,
|
| 1219 |
+
"step": 880000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.97,
|
| 1223 |
+
"eval_loss": 2.015188455581665,
|
| 1224 |
+
"eval_runtime": 875.6595,
|
| 1225 |
+
"eval_samples_per_second": 881.416,
|
| 1226 |
+
"eval_steps_per_second": 55.089,
|
| 1227 |
+
"step": 888000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.98,
|
| 1231 |
+
"learning_rate": 2.5693333333333333e-07,
|
| 1232 |
+
"loss": 2.1319,
|
| 1233 |
+
"step": 896000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.98,
|
| 1237 |
+
"eval_loss": 2.0173678398132324,
|
| 1238 |
+
"eval_runtime": 877.4749,
|
| 1239 |
+
"eval_samples_per_second": 879.592,
|
| 1240 |
+
"eval_steps_per_second": 54.975,
|
| 1241 |
+
"step": 896000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.99,
|
| 1245 |
+
"eval_loss": 2.020580291748047,
|
| 1246 |
+
"eval_runtime": 874.219,
|
| 1247 |
+
"eval_samples_per_second": 882.868,
|
| 1248 |
+
"eval_steps_per_second": 55.18,
|
| 1249 |
+
"step": 904000
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.0,
|
| 1253 |
+
"learning_rate": 2.542e-07,
|
| 1254 |
+
"loss": 2.1344,
|
| 1255 |
+
"step": 912000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.0,
|
| 1259 |
+
"eval_loss": 2.0178616046905518,
|
| 1260 |
+
"eval_runtime": 871.4372,
|
| 1261 |
+
"eval_samples_per_second": 885.686,
|
| 1262 |
+
"eval_steps_per_second": 55.356,
|
| 1263 |
+
"step": 912000
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.0,
|
| 1267 |
+
"eval_loss": 2.0153729915618896,
|
| 1268 |
+
"eval_runtime": 874.8229,
|
| 1269 |
+
"eval_samples_per_second": 882.259,
|
| 1270 |
+
"eval_steps_per_second": 55.141,
|
| 1271 |
+
"step": 920000
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.01,
|
| 1275 |
+
"learning_rate": 2.5146666666666664e-07,
|
| 1276 |
+
"loss": 2.1352,
|
| 1277 |
+
"step": 928000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.01,
|
| 1281 |
+
"eval_loss": 2.018483877182007,
|
| 1282 |
+
"eval_runtime": 876.0163,
|
| 1283 |
+
"eval_samples_per_second": 881.057,
|
| 1284 |
+
"eval_steps_per_second": 55.066,
|
| 1285 |
+
"step": 928000
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.02,
|
| 1289 |
+
"eval_loss": 2.016976833343506,
|
| 1290 |
+
"eval_runtime": 878.2619,
|
| 1291 |
+
"eval_samples_per_second": 878.804,
|
| 1292 |
+
"eval_steps_per_second": 54.926,
|
| 1293 |
+
"step": 936000
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 1.03,
|
| 1297 |
+
"learning_rate": 2.4873333333333335e-07,
|
| 1298 |
+
"loss": 2.1336,
|
| 1299 |
+
"step": 944000
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.03,
|
| 1303 |
+
"eval_loss": 2.016388416290283,
|
| 1304 |
+
"eval_runtime": 877.6593,
|
| 1305 |
+
"eval_samples_per_second": 879.407,
|
| 1306 |
+
"eval_steps_per_second": 54.963,
|
| 1307 |
+
"step": 944000
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 1.04,
|
| 1311 |
+
"eval_loss": 2.013742208480835,
|
| 1312 |
+
"eval_runtime": 871.0407,
|
| 1313 |
+
"eval_samples_per_second": 886.09,
|
| 1314 |
+
"eval_steps_per_second": 55.381,
|
| 1315 |
+
"step": 952000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.05,
|
| 1319 |
+
"learning_rate": 2.46e-07,
|
| 1320 |
+
"loss": 2.1315,
|
| 1321 |
+
"step": 960000
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.05,
|
| 1325 |
+
"eval_loss": 2.0176327228546143,
|
| 1326 |
+
"eval_runtime": 877.004,
|
| 1327 |
+
"eval_samples_per_second": 880.064,
|
| 1328 |
+
"eval_steps_per_second": 55.004,
|
| 1329 |
+
"step": 960000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.06,
|
| 1333 |
+
"eval_loss": 2.0155346393585205,
|
| 1334 |
+
"eval_runtime": 872.5922,
|
| 1335 |
+
"eval_samples_per_second": 884.514,
|
| 1336 |
+
"eval_steps_per_second": 55.282,
|
| 1337 |
+
"step": 968000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 1.06,
|
| 1341 |
+
"learning_rate": 2.4326666666666666e-07,
|
| 1342 |
+
"loss": 2.1255,
|
| 1343 |
+
"step": 976000
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.06,
|
| 1347 |
+
"eval_loss": 2.014533281326294,
|
| 1348 |
+
"eval_runtime": 871.4139,
|
| 1349 |
+
"eval_samples_per_second": 885.71,
|
| 1350 |
+
"eval_steps_per_second": 55.357,
|
| 1351 |
+
"step": 976000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.07,
|
| 1355 |
+
"eval_loss": 2.023314952850342,
|
| 1356 |
+
"eval_runtime": 879.3224,
|
| 1357 |
+
"eval_samples_per_second": 877.744,
|
| 1358 |
+
"eval_steps_per_second": 54.859,
|
| 1359 |
+
"step": 984000
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.08,
|
| 1363 |
+
"learning_rate": 2.405333333333333e-07,
|
| 1364 |
+
"loss": 2.1249,
|
| 1365 |
+
"step": 992000
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 1.08,
|
| 1369 |
+
"eval_loss": 2.0147762298583984,
|
| 1370 |
+
"eval_runtime": 866.8225,
|
| 1371 |
+
"eval_samples_per_second": 890.401,
|
| 1372 |
+
"eval_steps_per_second": 55.65,
|
| 1373 |
+
"step": 992000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.09,
|
| 1377 |
+
"eval_loss": 2.016249895095825,
|
| 1378 |
+
"eval_runtime": 867.6683,
|
| 1379 |
+
"eval_samples_per_second": 889.533,
|
| 1380 |
+
"eval_steps_per_second": 55.596,
|
| 1381 |
+
"step": 1000000
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.1,
|
| 1385 |
+
"learning_rate": 2.3779999999999997e-07,
|
| 1386 |
+
"loss": 2.123,
|
| 1387 |
+
"step": 1008000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.1,
|
| 1391 |
+
"eval_loss": 2.017381191253662,
|
| 1392 |
+
"eval_runtime": 868.2141,
|
| 1393 |
+
"eval_samples_per_second": 888.974,
|
| 1394 |
+
"eval_steps_per_second": 55.561,
|
| 1395 |
+
"step": 1008000
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.11,
|
| 1399 |
+
"eval_loss": 2.015009880065918,
|
| 1400 |
+
"eval_runtime": 865.5792,
|
| 1401 |
+
"eval_samples_per_second": 891.68,
|
| 1402 |
+
"eval_steps_per_second": 55.73,
|
| 1403 |
+
"step": 1016000
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.12,
|
| 1407 |
+
"learning_rate": 2.3506666666666668e-07,
|
| 1408 |
+
"loss": 2.1263,
|
| 1409 |
+
"step": 1024000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.12,
|
| 1413 |
+
"eval_loss": 2.0160863399505615,
|
| 1414 |
+
"eval_runtime": 869.2474,
|
| 1415 |
+
"eval_samples_per_second": 887.917,
|
| 1416 |
+
"eval_steps_per_second": 55.495,
|
| 1417 |
+
"step": 1024000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.13,
|
| 1421 |
+
"eval_loss": 2.0128889083862305,
|
| 1422 |
+
"eval_runtime": 866.9502,
|
| 1423 |
+
"eval_samples_per_second": 890.27,
|
| 1424 |
+
"eval_steps_per_second": 55.642,
|
| 1425 |
+
"step": 1032000
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 1.13,
|
| 1429 |
+
"learning_rate": 2.3233333333333334e-07,
|
| 1430 |
+
"loss": 2.1232,
|
| 1431 |
+
"step": 1040000
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.13,
|
| 1435 |
+
"eval_loss": 2.0166754722595215,
|
| 1436 |
+
"eval_runtime": 901.7962,
|
| 1437 |
+
"eval_samples_per_second": 855.87,
|
| 1438 |
+
"eval_steps_per_second": 53.492,
|
| 1439 |
+
"step": 1040000
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 1.14,
|
| 1443 |
+
"eval_loss": 2.012477397918701,
|
| 1444 |
+
"eval_runtime": 911.6669,
|
| 1445 |
+
"eval_samples_per_second": 846.603,
|
| 1446 |
+
"eval_steps_per_second": 52.913,
|
| 1447 |
+
"step": 1048000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.15,
|
| 1451 |
+
"learning_rate": 2.2960000000000002e-07,
|
| 1452 |
+
"loss": 2.1168,
|
| 1453 |
+
"step": 1056000
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.15,
|
| 1457 |
+
"eval_loss": 2.0113391876220703,
|
| 1458 |
+
"eval_runtime": 912.2557,
|
| 1459 |
+
"eval_samples_per_second": 846.057,
|
| 1460 |
+
"eval_steps_per_second": 52.879,
|
| 1461 |
+
"step": 1056000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 1.16,
|
| 1465 |
+
"eval_loss": 2.013575792312622,
|
| 1466 |
+
"eval_runtime": 901.3301,
|
| 1467 |
+
"eval_samples_per_second": 856.312,
|
| 1468 |
+
"eval_steps_per_second": 53.52,
|
| 1469 |
+
"step": 1064000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.17,
|
| 1473 |
+
"learning_rate": 2.2686666666666667e-07,
|
| 1474 |
+
"loss": 2.1307,
|
| 1475 |
+
"step": 1072000
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 1.17,
|
| 1479 |
+
"eval_loss": 2.014338254928589,
|
| 1480 |
+
"eval_runtime": 891.2807,
|
| 1481 |
+
"eval_samples_per_second": 865.967,
|
| 1482 |
+
"eval_steps_per_second": 54.123,
|
| 1483 |
+
"step": 1072000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.18,
|
| 1487 |
+
"eval_loss": 2.0166401863098145,
|
| 1488 |
+
"eval_runtime": 886.4005,
|
| 1489 |
+
"eval_samples_per_second": 870.735,
|
| 1490 |
+
"eval_steps_per_second": 54.421,
|
| 1491 |
+
"step": 1080000
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 1.19,
|
| 1495 |
+
"learning_rate": 2.2413333333333333e-07,
|
| 1496 |
+
"loss": 2.1336,
|
| 1497 |
+
"step": 1088000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 1.19,
|
| 1501 |
+
"eval_loss": 2.0103185176849365,
|
| 1502 |
+
"eval_runtime": 886.4458,
|
| 1503 |
+
"eval_samples_per_second": 870.691,
|
| 1504 |
+
"eval_steps_per_second": 54.418,
|
| 1505 |
+
"step": 1088000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 1.2,
|
| 1509 |
+
"eval_loss": 2.0129764080047607,
|
| 1510 |
+
"eval_runtime": 890.355,
|
| 1511 |
+
"eval_samples_per_second": 866.868,
|
| 1512 |
+
"eval_steps_per_second": 54.18,
|
| 1513 |
+
"step": 1096000
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.2,
|
| 1517 |
+
"learning_rate": 2.214e-07,
|
| 1518 |
+
"loss": 2.1227,
|
| 1519 |
+
"step": 1104000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.2,
|
| 1523 |
+
"eval_loss": 2.012451648712158,
|
| 1524 |
+
"eval_runtime": 895.3428,
|
| 1525 |
+
"eval_samples_per_second": 862.039,
|
| 1526 |
+
"eval_steps_per_second": 53.878,
|
| 1527 |
+
"step": 1104000
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.21,
|
| 1531 |
+
"eval_loss": 2.0183231830596924,
|
| 1532 |
+
"eval_runtime": 888.3913,
|
| 1533 |
+
"eval_samples_per_second": 868.784,
|
| 1534 |
+
"eval_steps_per_second": 54.299,
|
| 1535 |
+
"step": 1112000
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.22,
|
| 1539 |
+
"learning_rate": 2.1866666666666667e-07,
|
| 1540 |
+
"loss": 2.1223,
|
| 1541 |
+
"step": 1120000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.22,
|
| 1545 |
+
"eval_loss": 2.014848470687866,
|
| 1546 |
+
"eval_runtime": 889.5583,
|
| 1547 |
+
"eval_samples_per_second": 867.644,
|
| 1548 |
+
"eval_steps_per_second": 54.228,
|
| 1549 |
+
"step": 1120000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.23,
|
| 1553 |
+
"eval_loss": 2.0147109031677246,
|
| 1554 |
+
"eval_runtime": 884.3146,
|
| 1555 |
+
"eval_samples_per_second": 872.789,
|
| 1556 |
+
"eval_steps_per_second": 54.55,
|
| 1557 |
+
"step": 1128000
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.24,
|
| 1561 |
+
"learning_rate": 2.1593333333333332e-07,
|
| 1562 |
+
"loss": 2.1289,
|
| 1563 |
+
"step": 1136000
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.24,
|
| 1567 |
+
"eval_loss": 2.0108699798583984,
|
| 1568 |
+
"eval_runtime": 888.3584,
|
| 1569 |
+
"eval_samples_per_second": 868.816,
|
| 1570 |
+
"eval_steps_per_second": 54.301,
|
| 1571 |
+
"step": 1136000
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.25,
|
| 1575 |
+
"eval_loss": 2.0163819789886475,
|
| 1576 |
+
"eval_runtime": 887.4195,
|
| 1577 |
+
"eval_samples_per_second": 869.735,
|
| 1578 |
+
"eval_steps_per_second": 54.359,
|
| 1579 |
+
"step": 1144000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.26,
|
| 1583 |
+
"learning_rate": 2.132e-07,
|
| 1584 |
+
"loss": 2.1278,
|
| 1585 |
+
"step": 1152000
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.26,
|
| 1589 |
+
"eval_loss": 2.0163345336914062,
|
| 1590 |
+
"eval_runtime": 886.1604,
|
| 1591 |
+
"eval_samples_per_second": 870.971,
|
| 1592 |
+
"eval_steps_per_second": 54.436,
|
| 1593 |
+
"step": 1152000
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 1.27,
|
| 1597 |
+
"eval_loss": 2.012103319168091,
|
| 1598 |
+
"eval_runtime": 889.5174,
|
| 1599 |
+
"eval_samples_per_second": 867.684,
|
| 1600 |
+
"eval_steps_per_second": 54.231,
|
| 1601 |
+
"step": 1160000
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.27,
|
| 1605 |
+
"learning_rate": 2.1046666666666666e-07,
|
| 1606 |
+
"loss": 2.1261,
|
| 1607 |
+
"step": 1168000
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.27,
|
| 1611 |
+
"eval_loss": 2.011343240737915,
|
| 1612 |
+
"eval_runtime": 890.9332,
|
| 1613 |
+
"eval_samples_per_second": 866.305,
|
| 1614 |
+
"eval_steps_per_second": 54.144,
|
| 1615 |
+
"step": 1168000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.28,
|
| 1619 |
+
"eval_loss": 2.0137104988098145,
|
| 1620 |
+
"eval_runtime": 883.4659,
|
| 1621 |
+
"eval_samples_per_second": 873.627,
|
| 1622 |
+
"eval_steps_per_second": 54.602,
|
| 1623 |
+
"step": 1176000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 1.29,
|
| 1627 |
+
"learning_rate": 2.0773333333333334e-07,
|
| 1628 |
+
"loss": 2.126,
|
| 1629 |
+
"step": 1184000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 1.29,
|
| 1633 |
+
"eval_loss": 2.015174627304077,
|
| 1634 |
+
"eval_runtime": 885.9678,
|
| 1635 |
+
"eval_samples_per_second": 871.16,
|
| 1636 |
+
"eval_steps_per_second": 54.448,
|
| 1637 |
+
"step": 1184000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 1.3,
|
| 1641 |
+
"eval_loss": 2.010411500930786,
|
| 1642 |
+
"eval_runtime": 888.6748,
|
| 1643 |
+
"eval_samples_per_second": 868.507,
|
| 1644 |
+
"eval_steps_per_second": 54.282,
|
| 1645 |
+
"step": 1192000
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.31,
|
| 1649 |
+
"learning_rate": 2.05e-07,
|
| 1650 |
+
"loss": 2.1235,
|
| 1651 |
+
"step": 1200000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.31,
|
| 1655 |
+
"eval_loss": 2.013165235519409,
|
| 1656 |
+
"eval_runtime": 888.6503,
|
| 1657 |
+
"eval_samples_per_second": 868.531,
|
| 1658 |
+
"eval_steps_per_second": 54.283,
|
| 1659 |
+
"step": 1200000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 1.32,
|
| 1663 |
+
"eval_loss": 2.0113847255706787,
|
| 1664 |
+
"eval_runtime": 884.261,
|
| 1665 |
+
"eval_samples_per_second": 872.842,
|
| 1666 |
+
"eval_steps_per_second": 54.553,
|
| 1667 |
+
"step": 1208000
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.33,
|
| 1671 |
+
"learning_rate": 2.0226666666666668e-07,
|
| 1672 |
+
"loss": 2.1229,
|
| 1673 |
+
"step": 1216000
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.33,
|
| 1677 |
+
"eval_loss": 2.010532855987549,
|
| 1678 |
+
"eval_runtime": 887.5065,
|
| 1679 |
+
"eval_samples_per_second": 869.65,
|
| 1680 |
+
"eval_steps_per_second": 54.353,
|
| 1681 |
+
"step": 1216000
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.34,
|
| 1685 |
+
"eval_loss": 2.0130858421325684,
|
| 1686 |
+
"eval_runtime": 881.1399,
|
| 1687 |
+
"eval_samples_per_second": 875.934,
|
| 1688 |
+
"eval_steps_per_second": 54.746,
|
| 1689 |
+
"step": 1224000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.34,
|
| 1693 |
+
"learning_rate": 1.9953333333333333e-07,
|
| 1694 |
+
"loss": 2.1213,
|
| 1695 |
+
"step": 1232000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 1.34,
|
| 1699 |
+
"eval_loss": 2.0141072273254395,
|
| 1700 |
+
"eval_runtime": 882.2467,
|
| 1701 |
+
"eval_samples_per_second": 874.835,
|
| 1702 |
+
"eval_steps_per_second": 54.677,
|
| 1703 |
+
"step": 1232000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.35,
|
| 1707 |
+
"eval_loss": 2.010868549346924,
|
| 1708 |
+
"eval_runtime": 881.7078,
|
| 1709 |
+
"eval_samples_per_second": 875.369,
|
| 1710 |
+
"eval_steps_per_second": 54.711,
|
| 1711 |
+
"step": 1240000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.36,
|
| 1715 |
+
"learning_rate": 1.968e-07,
|
| 1716 |
+
"loss": 2.1185,
|
| 1717 |
+
"step": 1248000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.36,
|
| 1721 |
+
"eval_loss": 2.0129363536834717,
|
| 1722 |
+
"eval_runtime": 886.2455,
|
| 1723 |
+
"eval_samples_per_second": 870.887,
|
| 1724 |
+
"eval_steps_per_second": 54.431,
|
| 1725 |
+
"step": 1248000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.37,
|
| 1729 |
+
"eval_loss": 2.011003017425537,
|
| 1730 |
+
"eval_runtime": 888.1974,
|
| 1731 |
+
"eval_samples_per_second": 868.974,
|
| 1732 |
+
"eval_steps_per_second": 54.311,
|
| 1733 |
+
"step": 1256000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.38,
|
| 1737 |
+
"learning_rate": 1.9406666666666667e-07,
|
| 1738 |
+
"loss": 2.131,
|
| 1739 |
+
"step": 1264000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.38,
|
| 1743 |
+
"eval_loss": 2.01228928565979,
|
| 1744 |
+
"eval_runtime": 884.9282,
|
| 1745 |
+
"eval_samples_per_second": 872.184,
|
| 1746 |
+
"eval_steps_per_second": 54.512,
|
| 1747 |
+
"step": 1264000
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 1.39,
|
| 1751 |
+
"eval_loss": 2.0104737281799316,
|
| 1752 |
+
"eval_runtime": 881.1611,
|
| 1753 |
+
"eval_samples_per_second": 875.912,
|
| 1754 |
+
"eval_steps_per_second": 54.745,
|
| 1755 |
+
"step": 1272000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 1.4,
|
| 1759 |
+
"learning_rate": 1.9133333333333333e-07,
|
| 1760 |
+
"loss": 2.1141,
|
| 1761 |
+
"step": 1280000
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 1.4,
|
| 1765 |
+
"eval_loss": 2.010425090789795,
|
| 1766 |
+
"eval_runtime": 882.3806,
|
| 1767 |
+
"eval_samples_per_second": 874.702,
|
| 1768 |
+
"eval_steps_per_second": 54.669,
|
| 1769 |
+
"step": 1280000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 1.41,
|
| 1773 |
+
"eval_loss": 2.015007734298706,
|
| 1774 |
+
"eval_runtime": 879.3909,
|
| 1775 |
+
"eval_samples_per_second": 877.676,
|
| 1776 |
+
"eval_steps_per_second": 54.855,
|
| 1777 |
+
"step": 1288000
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 1.41,
|
| 1781 |
+
"learning_rate": 1.886e-07,
|
| 1782 |
+
"loss": 2.1219,
|
| 1783 |
+
"step": 1296000
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.41,
|
| 1787 |
+
"eval_loss": 2.0161073207855225,
|
| 1788 |
+
"eval_runtime": 879.4904,
|
| 1789 |
+
"eval_samples_per_second": 877.576,
|
| 1790 |
+
"eval_steps_per_second": 54.849,
|
| 1791 |
+
"step": 1296000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 1.42,
|
| 1795 |
+
"eval_loss": 2.00930118560791,
|
| 1796 |
+
"eval_runtime": 882.5935,
|
| 1797 |
+
"eval_samples_per_second": 874.491,
|
| 1798 |
+
"eval_steps_per_second": 54.656,
|
| 1799 |
+
"step": 1304000
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 1.43,
|
| 1803 |
+
"learning_rate": 1.8586666666666666e-07,
|
| 1804 |
+
"loss": 2.1203,
|
| 1805 |
+
"step": 1312000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 1.43,
|
| 1809 |
+
"eval_loss": 2.0104291439056396,
|
| 1810 |
+
"eval_runtime": 882.9969,
|
| 1811 |
+
"eval_samples_per_second": 874.091,
|
| 1812 |
+
"eval_steps_per_second": 54.631,
|
| 1813 |
+
"step": 1312000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.44,
|
| 1817 |
+
"eval_loss": 2.0144429206848145,
|
| 1818 |
+
"eval_runtime": 878.5955,
|
| 1819 |
+
"eval_samples_per_second": 878.47,
|
| 1820 |
+
"eval_steps_per_second": 54.905,
|
| 1821 |
+
"step": 1320000
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 1.45,
|
| 1825 |
+
"learning_rate": 1.8313333333333332e-07,
|
| 1826 |
+
"loss": 2.1264,
|
| 1827 |
+
"step": 1328000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 1.45,
|
| 1831 |
+
"eval_loss": 2.0084986686706543,
|
| 1832 |
+
"eval_runtime": 878.8817,
|
| 1833 |
+
"eval_samples_per_second": 878.184,
|
| 1834 |
+
"eval_steps_per_second": 54.887,
|
| 1835 |
+
"step": 1328000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 1.46,
|
| 1839 |
+
"eval_loss": 2.0118672847747803,
|
| 1840 |
+
"eval_runtime": 880.8514,
|
| 1841 |
+
"eval_samples_per_second": 876.22,
|
| 1842 |
+
"eval_steps_per_second": 54.764,
|
| 1843 |
+
"step": 1336000
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.47,
|
| 1847 |
+
"learning_rate": 1.804e-07,
|
| 1848 |
+
"loss": 2.1194,
|
| 1849 |
+
"step": 1344000
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 1.47,
|
| 1853 |
+
"eval_loss": 2.011784076690674,
|
| 1854 |
+
"eval_runtime": 878.874,
|
| 1855 |
+
"eval_samples_per_second": 878.192,
|
| 1856 |
+
"eval_steps_per_second": 54.887,
|
| 1857 |
+
"step": 1344000
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.48,
|
| 1861 |
+
"eval_loss": 2.0109827518463135,
|
| 1862 |
+
"eval_runtime": 893.715,
|
| 1863 |
+
"eval_samples_per_second": 863.609,
|
| 1864 |
+
"eval_steps_per_second": 53.976,
|
| 1865 |
+
"step": 1352000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.48,
|
| 1869 |
+
"learning_rate": 1.7766666666666666e-07,
|
| 1870 |
+
"loss": 2.117,
|
| 1871 |
+
"step": 1360000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.48,
|
| 1875 |
+
"eval_loss": 2.014660596847534,
|
| 1876 |
+
"eval_runtime": 915.8924,
|
| 1877 |
+
"eval_samples_per_second": 842.697,
|
| 1878 |
+
"eval_steps_per_second": 52.669,
|
| 1879 |
+
"step": 1360000
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.49,
|
| 1883 |
+
"eval_loss": 2.013535261154175,
|
| 1884 |
+
"eval_runtime": 909.1816,
|
| 1885 |
+
"eval_samples_per_second": 848.917,
|
| 1886 |
+
"eval_steps_per_second": 53.058,
|
| 1887 |
+
"step": 1368000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 1.5,
|
| 1891 |
+
"learning_rate": 1.7493333333333334e-07,
|
| 1892 |
+
"loss": 2.1311,
|
| 1893 |
+
"step": 1376000
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.5,
|
| 1897 |
+
"eval_loss": 2.0076611042022705,
|
| 1898 |
+
"eval_runtime": 909.3083,
|
| 1899 |
+
"eval_samples_per_second": 848.799,
|
| 1900 |
+
"eval_steps_per_second": 53.05,
|
| 1901 |
+
"step": 1376000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 1.51,
|
| 1905 |
+
"eval_loss": 2.006574869155884,
|
| 1906 |
+
"eval_runtime": 904.8344,
|
| 1907 |
+
"eval_samples_per_second": 852.996,
|
| 1908 |
+
"eval_steps_per_second": 53.313,
|
| 1909 |
+
"step": 1384000
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 1.52,
|
| 1913 |
+
"learning_rate": 1.722e-07,
|
| 1914 |
+
"loss": 2.1215,
|
| 1915 |
+
"step": 1392000
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 1.52,
|
| 1919 |
+
"eval_loss": 2.008929967880249,
|
| 1920 |
+
"eval_runtime": 903.4488,
|
| 1921 |
+
"eval_samples_per_second": 854.304,
|
| 1922 |
+
"eval_steps_per_second": 53.394,
|
| 1923 |
+
"step": 1392000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 1.53,
|
| 1927 |
+
"eval_loss": 2.0118260383605957,
|
| 1928 |
+
"eval_runtime": 913.6278,
|
| 1929 |
+
"eval_samples_per_second": 844.786,
|
| 1930 |
+
"eval_steps_per_second": 52.799,
|
| 1931 |
+
"step": 1400000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 1.54,
|
| 1935 |
+
"learning_rate": 1.6946666666666668e-07,
|
| 1936 |
+
"loss": 2.1185,
|
| 1937 |
+
"step": 1408000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 1.54,
|
| 1941 |
+
"eval_loss": 2.0105414390563965,
|
| 1942 |
+
"eval_runtime": 907.6551,
|
| 1943 |
+
"eval_samples_per_second": 850.345,
|
| 1944 |
+
"eval_steps_per_second": 53.147,
|
| 1945 |
+
"step": 1408000
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 1.54,
|
| 1949 |
+
"eval_loss": 2.012268304824829,
|
| 1950 |
+
"eval_runtime": 903.9952,
|
| 1951 |
+
"eval_samples_per_second": 853.788,
|
| 1952 |
+
"eval_steps_per_second": 53.362,
|
| 1953 |
+
"step": 1416000
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 1.55,
|
| 1957 |
+
"learning_rate": 1.6673333333333333e-07,
|
| 1958 |
+
"loss": 2.1284,
|
| 1959 |
+
"step": 1424000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 1.55,
|
| 1963 |
+
"eval_loss": 2.0133912563323975,
|
| 1964 |
+
"eval_runtime": 910.6028,
|
| 1965 |
+
"eval_samples_per_second": 847.592,
|
| 1966 |
+
"eval_steps_per_second": 52.975,
|
| 1967 |
+
"step": 1424000
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 1.56,
|
| 1971 |
+
"eval_loss": 2.009307861328125,
|
| 1972 |
+
"eval_runtime": 904.2587,
|
| 1973 |
+
"eval_samples_per_second": 853.539,
|
| 1974 |
+
"eval_steps_per_second": 53.346,
|
| 1975 |
+
"step": 1432000
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 1.57,
|
| 1979 |
+
"learning_rate": 1.64e-07,
|
| 1980 |
+
"loss": 2.1174,
|
| 1981 |
+
"step": 1440000
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 1.57,
|
| 1985 |
+
"eval_loss": 2.0101728439331055,
|
| 1986 |
+
"eval_runtime": 912.2693,
|
| 1987 |
+
"eval_samples_per_second": 846.044,
|
| 1988 |
+
"eval_steps_per_second": 52.878,
|
| 1989 |
+
"step": 1440000
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 1.58,
|
| 1993 |
+
"eval_loss": 2.00759220123291,
|
| 1994 |
+
"eval_runtime": 910.2393,
|
| 1995 |
+
"eval_samples_per_second": 847.931,
|
| 1996 |
+
"eval_steps_per_second": 52.996,
|
| 1997 |
+
"step": 1448000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.59,
|
| 2001 |
+
"learning_rate": 1.6126666666666667e-07,
|
| 2002 |
+
"loss": 2.1108,
|
| 2003 |
+
"step": 1456000
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 1.59,
|
| 2007 |
+
"eval_loss": 2.00740909576416,
|
| 2008 |
+
"eval_runtime": 914.6796,
|
| 2009 |
+
"eval_samples_per_second": 843.815,
|
| 2010 |
+
"eval_steps_per_second": 52.739,
|
| 2011 |
+
"step": 1456000
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.6,
|
| 2015 |
+
"eval_loss": 2.007056474685669,
|
| 2016 |
+
"eval_runtime": 908.0025,
|
| 2017 |
+
"eval_samples_per_second": 850.02,
|
| 2018 |
+
"eval_steps_per_second": 53.127,
|
| 2019 |
+
"step": 1464000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.61,
|
| 2023 |
+
"learning_rate": 1.5853333333333332e-07,
|
| 2024 |
+
"loss": 2.1252,
|
| 2025 |
+
"step": 1472000
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 1.61,
|
| 2029 |
+
"eval_loss": 2.0092082023620605,
|
| 2030 |
+
"eval_runtime": 905.6872,
|
| 2031 |
+
"eval_samples_per_second": 852.193,
|
| 2032 |
+
"eval_steps_per_second": 53.262,
|
| 2033 |
+
"step": 1472000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.61,
|
| 2037 |
+
"eval_loss": 2.007967233657837,
|
| 2038 |
+
"eval_runtime": 910.9272,
|
| 2039 |
+
"eval_samples_per_second": 847.291,
|
| 2040 |
+
"eval_steps_per_second": 52.956,
|
| 2041 |
+
"step": 1480000
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.62,
|
| 2045 |
+
"learning_rate": 1.558e-07,
|
| 2046 |
+
"loss": 2.121,
|
| 2047 |
+
"step": 1488000
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.62,
|
| 2051 |
+
"eval_loss": 2.0052874088287354,
|
| 2052 |
+
"eval_runtime": 908.8472,
|
| 2053 |
+
"eval_samples_per_second": 849.23,
|
| 2054 |
+
"eval_steps_per_second": 53.077,
|
| 2055 |
+
"step": 1488000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 1.63,
|
| 2059 |
+
"eval_loss": 2.0071661472320557,
|
| 2060 |
+
"eval_runtime": 907.693,
|
| 2061 |
+
"eval_samples_per_second": 850.31,
|
| 2062 |
+
"eval_steps_per_second": 53.145,
|
| 2063 |
+
"step": 1496000
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 1.64,
|
| 2067 |
+
"learning_rate": 1.5306666666666666e-07,
|
| 2068 |
+
"loss": 2.1178,
|
| 2069 |
+
"step": 1504000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 1.64,
|
| 2073 |
+
"eval_loss": 2.0059070587158203,
|
| 2074 |
+
"eval_runtime": 908.356,
|
| 2075 |
+
"eval_samples_per_second": 849.689,
|
| 2076 |
+
"eval_steps_per_second": 53.106,
|
| 2077 |
+
"step": 1504000
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 1.65,
|
| 2081 |
+
"eval_loss": 2.00836443901062,
|
| 2082 |
+
"eval_runtime": 908.0246,
|
| 2083 |
+
"eval_samples_per_second": 849.999,
|
| 2084 |
+
"eval_steps_per_second": 53.125,
|
| 2085 |
+
"step": 1512000
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 1.66,
|
| 2089 |
+
"learning_rate": 1.5033333333333332e-07,
|
| 2090 |
+
"loss": 2.1154,
|
| 2091 |
+
"step": 1520000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 1.66,
|
| 2095 |
+
"eval_loss": 2.0105550289154053,
|
| 2096 |
+
"eval_runtime": 903.6608,
|
| 2097 |
+
"eval_samples_per_second": 854.104,
|
| 2098 |
+
"eval_steps_per_second": 53.382,
|
| 2099 |
+
"step": 1520000
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 1.67,
|
| 2103 |
+
"eval_loss": 2.0116729736328125,
|
| 2104 |
+
"eval_runtime": 909.1515,
|
| 2105 |
+
"eval_samples_per_second": 848.945,
|
| 2106 |
+
"eval_steps_per_second": 53.059,
|
| 2107 |
+
"step": 1528000
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 1.68,
|
| 2111 |
+
"learning_rate": 1.476e-07,
|
| 2112 |
+
"loss": 2.1214,
|
| 2113 |
+
"step": 1536000
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 1.68,
|
| 2117 |
+
"eval_loss": 2.006955146789551,
|
| 2118 |
+
"eval_runtime": 907.2355,
|
| 2119 |
+
"eval_samples_per_second": 850.738,
|
| 2120 |
+
"eval_steps_per_second": 53.171,
|
| 2121 |
+
"step": 1536000
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 1.68,
|
| 2125 |
+
"eval_loss": 2.0078775882720947,
|
| 2126 |
+
"eval_runtime": 908.5609,
|
| 2127 |
+
"eval_samples_per_second": 849.497,
|
| 2128 |
+
"eval_steps_per_second": 53.094,
|
| 2129 |
+
"step": 1544000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 1.69,
|
| 2133 |
+
"learning_rate": 1.4486666666666665e-07,
|
| 2134 |
+
"loss": 2.1175,
|
| 2135 |
+
"step": 1552000
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 1.69,
|
| 2139 |
+
"eval_loss": 2.0101876258850098,
|
| 2140 |
+
"eval_runtime": 901.3076,
|
| 2141 |
+
"eval_samples_per_second": 856.334,
|
| 2142 |
+
"eval_steps_per_second": 53.521,
|
| 2143 |
+
"step": 1552000
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 1.7,
|
| 2147 |
+
"eval_loss": 2.009697675704956,
|
| 2148 |
+
"eval_runtime": 906.1011,
|
| 2149 |
+
"eval_samples_per_second": 851.803,
|
| 2150 |
+
"eval_steps_per_second": 53.238,
|
| 2151 |
+
"step": 1560000
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 1.71,
|
| 2155 |
+
"learning_rate": 1.4213333333333334e-07,
|
| 2156 |
+
"loss": 2.1206,
|
| 2157 |
+
"step": 1568000
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 1.71,
|
| 2161 |
+
"eval_loss": 2.0092358589172363,
|
| 2162 |
+
"eval_runtime": 901.2376,
|
| 2163 |
+
"eval_samples_per_second": 856.4,
|
| 2164 |
+
"eval_steps_per_second": 53.525,
|
| 2165 |
+
"step": 1568000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 1.72,
|
| 2169 |
+
"eval_loss": 2.005527973175049,
|
| 2170 |
+
"eval_runtime": 896.3075,
|
| 2171 |
+
"eval_samples_per_second": 861.111,
|
| 2172 |
+
"eval_steps_per_second": 53.82,
|
| 2173 |
+
"step": 1576000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.73,
|
| 2177 |
+
"learning_rate": 1.3940000000000002e-07,
|
| 2178 |
+
"loss": 2.1302,
|
| 2179 |
+
"step": 1584000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 1.73,
|
| 2183 |
+
"eval_loss": 2.008502244949341,
|
| 2184 |
+
"eval_runtime": 899.3251,
|
| 2185 |
+
"eval_samples_per_second": 858.221,
|
| 2186 |
+
"eval_steps_per_second": 53.639,
|
| 2187 |
+
"step": 1584000
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 1.74,
|
| 2191 |
+
"eval_loss": 2.0109806060791016,
|
| 2192 |
+
"eval_runtime": 906.7205,
|
| 2193 |
+
"eval_samples_per_second": 851.222,
|
| 2194 |
+
"eval_steps_per_second": 53.202,
|
| 2195 |
+
"step": 1592000
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 1.75,
|
| 2199 |
+
"learning_rate": 1.3666666666666665e-07,
|
| 2200 |
+
"loss": 2.1177,
|
| 2201 |
+
"step": 1600000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 1.75,
|
| 2205 |
+
"eval_loss": 2.006521701812744,
|
| 2206 |
+
"eval_runtime": 898.4764,
|
| 2207 |
+
"eval_samples_per_second": 859.032,
|
| 2208 |
+
"eval_steps_per_second": 53.69,
|
| 2209 |
+
"step": 1600000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 1.75,
|
| 2213 |
+
"eval_loss": 2.0131704807281494,
|
| 2214 |
+
"eval_runtime": 906.0839,
|
| 2215 |
+
"eval_samples_per_second": 851.82,
|
| 2216 |
+
"eval_steps_per_second": 53.239,
|
| 2217 |
+
"step": 1608000
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 1.76,
|
| 2221 |
+
"learning_rate": 1.3393333333333333e-07,
|
| 2222 |
+
"loss": 2.1101,
|
| 2223 |
+
"step": 1616000
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 1.76,
|
| 2227 |
+
"eval_loss": 2.0085511207580566,
|
| 2228 |
+
"eval_runtime": 896.2709,
|
| 2229 |
+
"eval_samples_per_second": 861.146,
|
| 2230 |
+
"eval_steps_per_second": 53.822,
|
| 2231 |
+
"step": 1616000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 1.77,
|
| 2235 |
+
"eval_loss": 2.0077245235443115,
|
| 2236 |
+
"eval_runtime": 897.3988,
|
| 2237 |
+
"eval_samples_per_second": 860.064,
|
| 2238 |
+
"eval_steps_per_second": 53.754,
|
| 2239 |
+
"step": 1624000
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 1.78,
|
| 2243 |
+
"learning_rate": 1.312e-07,
|
| 2244 |
+
"loss": 2.1194,
|
| 2245 |
+
"step": 1632000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 1.78,
|
| 2249 |
+
"eval_loss": 2.008148431777954,
|
| 2250 |
+
"eval_runtime": 896.5575,
|
| 2251 |
+
"eval_samples_per_second": 860.871,
|
| 2252 |
+
"eval_steps_per_second": 53.805,
|
| 2253 |
+
"step": 1632000
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 1.79,
|
| 2257 |
+
"eval_loss": 2.008798122406006,
|
| 2258 |
+
"eval_runtime": 897.2787,
|
| 2259 |
+
"eval_samples_per_second": 860.179,
|
| 2260 |
+
"eval_steps_per_second": 53.761,
|
| 2261 |
+
"step": 1640000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 1.8,
|
| 2265 |
+
"learning_rate": 1.2846666666666667e-07,
|
| 2266 |
+
"loss": 2.1167,
|
| 2267 |
+
"step": 1648000
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 1.8,
|
| 2271 |
+
"eval_loss": 2.002239942550659,
|
| 2272 |
+
"eval_runtime": 893.5655,
|
| 2273 |
+
"eval_samples_per_second": 863.753,
|
| 2274 |
+
"eval_steps_per_second": 53.985,
|
| 2275 |
+
"step": 1648000
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 1.81,
|
| 2279 |
+
"eval_loss": 2.007662296295166,
|
| 2280 |
+
"eval_runtime": 895.7141,
|
| 2281 |
+
"eval_samples_per_second": 861.681,
|
| 2282 |
+
"eval_steps_per_second": 53.855,
|
| 2283 |
+
"step": 1656000
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 1.82,
|
| 2287 |
+
"learning_rate": 1.2573333333333332e-07,
|
| 2288 |
+
"loss": 2.1083,
|
| 2289 |
+
"step": 1664000
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 1.82,
|
| 2293 |
+
"eval_loss": 2.0065953731536865,
|
| 2294 |
+
"eval_runtime": 890.9713,
|
| 2295 |
+
"eval_samples_per_second": 866.268,
|
| 2296 |
+
"eval_steps_per_second": 54.142,
|
| 2297 |
+
"step": 1664000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 1.82,
|
| 2301 |
+
"eval_loss": 2.0137040615081787,
|
| 2302 |
+
"eval_runtime": 885.7627,
|
| 2303 |
+
"eval_samples_per_second": 871.362,
|
| 2304 |
+
"eval_steps_per_second": 54.46,
|
| 2305 |
+
"step": 1672000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 1.83,
|
| 2309 |
+
"learning_rate": 1.23e-07,
|
| 2310 |
+
"loss": 2.1232,
|
| 2311 |
+
"step": 1680000
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 1.83,
|
| 2315 |
+
"eval_loss": 2.0067014694213867,
|
| 2316 |
+
"eval_runtime": 890.51,
|
| 2317 |
+
"eval_samples_per_second": 866.717,
|
| 2318 |
+
"eval_steps_per_second": 54.17,
|
| 2319 |
+
"step": 1680000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 1.84,
|
| 2323 |
+
"eval_loss": 2.0039150714874268,
|
| 2324 |
+
"eval_runtime": 889.3586,
|
| 2325 |
+
"eval_samples_per_second": 867.839,
|
| 2326 |
+
"eval_steps_per_second": 54.24,
|
| 2327 |
+
"step": 1688000
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 1.85,
|
| 2331 |
+
"learning_rate": 1.2026666666666666e-07,
|
| 2332 |
+
"loss": 2.1212,
|
| 2333 |
+
"step": 1696000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.85,
|
| 2337 |
+
"eval_loss": 2.008970022201538,
|
| 2338 |
+
"eval_runtime": 893.785,
|
| 2339 |
+
"eval_samples_per_second": 863.541,
|
| 2340 |
+
"eval_steps_per_second": 53.972,
|
| 2341 |
+
"step": 1696000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 1.86,
|
| 2345 |
+
"eval_loss": 2.0079498291015625,
|
| 2346 |
+
"eval_runtime": 882.5613,
|
| 2347 |
+
"eval_samples_per_second": 874.523,
|
| 2348 |
+
"eval_steps_per_second": 54.658,
|
| 2349 |
+
"step": 1704000
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 1.87,
|
| 2353 |
+
"learning_rate": 1.1753333333333334e-07,
|
| 2354 |
+
"loss": 2.1246,
|
| 2355 |
+
"step": 1712000
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 1.87,
|
| 2359 |
+
"eval_loss": 2.0082814693450928,
|
| 2360 |
+
"eval_runtime": 886.133,
|
| 2361 |
+
"eval_samples_per_second": 870.998,
|
| 2362 |
+
"eval_steps_per_second": 54.438,
|
| 2363 |
+
"step": 1712000
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 1.88,
|
| 2367 |
+
"eval_loss": 2.003898859024048,
|
| 2368 |
+
"eval_runtime": 887.1853,
|
| 2369 |
+
"eval_samples_per_second": 869.965,
|
| 2370 |
+
"eval_steps_per_second": 54.373,
|
| 2371 |
+
"step": 1720000
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 1.89,
|
| 2375 |
+
"learning_rate": 1.1480000000000001e-07,
|
| 2376 |
+
"loss": 2.1129,
|
| 2377 |
+
"step": 1728000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 1.89,
|
| 2381 |
+
"eval_loss": 2.0069074630737305,
|
| 2382 |
+
"eval_runtime": 891.3907,
|
| 2383 |
+
"eval_samples_per_second": 865.86,
|
| 2384 |
+
"eval_steps_per_second": 54.117,
|
| 2385 |
+
"step": 1728000
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 1.89,
|
| 2389 |
+
"eval_loss": 2.007922410964966,
|
| 2390 |
+
"eval_runtime": 884.1175,
|
| 2391 |
+
"eval_samples_per_second": 872.984,
|
| 2392 |
+
"eval_steps_per_second": 54.562,
|
| 2393 |
+
"step": 1736000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 1.9,
|
| 2397 |
+
"learning_rate": 1.1206666666666666e-07,
|
| 2398 |
+
"loss": 2.1209,
|
| 2399 |
+
"step": 1744000
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 1.9,
|
| 2403 |
+
"eval_loss": 2.00584077835083,
|
| 2404 |
+
"eval_runtime": 888.6359,
|
| 2405 |
+
"eval_samples_per_second": 868.545,
|
| 2406 |
+
"eval_steps_per_second": 54.284,
|
| 2407 |
+
"step": 1744000
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 1.91,
|
| 2411 |
+
"eval_loss": 2.0071957111358643,
|
| 2412 |
+
"eval_runtime": 891.8674,
|
| 2413 |
+
"eval_samples_per_second": 865.398,
|
| 2414 |
+
"eval_steps_per_second": 54.088,
|
| 2415 |
+
"step": 1752000
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 1.92,
|
| 2419 |
+
"learning_rate": 1.0933333333333333e-07,
|
| 2420 |
+
"loss": 2.1209,
|
| 2421 |
+
"step": 1760000
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 1.92,
|
| 2425 |
+
"eval_loss": 2.0067615509033203,
|
| 2426 |
+
"eval_runtime": 884.8141,
|
| 2427 |
+
"eval_samples_per_second": 872.296,
|
| 2428 |
+
"eval_steps_per_second": 54.519,
|
| 2429 |
+
"step": 1760000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 1.93,
|
| 2433 |
+
"eval_loss": 2.0078628063201904,
|
| 2434 |
+
"eval_runtime": 888.3025,
|
| 2435 |
+
"eval_samples_per_second": 868.871,
|
| 2436 |
+
"eval_steps_per_second": 54.305,
|
| 2437 |
+
"step": 1768000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 1.94,
|
| 2441 |
+
"learning_rate": 1.066e-07,
|
| 2442 |
+
"loss": 2.1184,
|
| 2443 |
+
"step": 1776000
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 1.94,
|
| 2447 |
+
"eval_loss": 2.0036442279815674,
|
| 2448 |
+
"eval_runtime": 887.5766,
|
| 2449 |
+
"eval_samples_per_second": 869.581,
|
| 2450 |
+
"eval_steps_per_second": 54.349,
|
| 2451 |
+
"step": 1776000
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 1.95,
|
| 2455 |
+
"eval_loss": 2.0064985752105713,
|
| 2456 |
+
"eval_runtime": 890.3705,
|
| 2457 |
+
"eval_samples_per_second": 866.853,
|
| 2458 |
+
"eval_steps_per_second": 54.179,
|
| 2459 |
+
"step": 1784000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 1.96,
|
| 2463 |
+
"learning_rate": 1.0386666666666667e-07,
|
| 2464 |
+
"loss": 2.1065,
|
| 2465 |
+
"step": 1792000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 1.96,
|
| 2469 |
+
"eval_loss": 2.007737159729004,
|
| 2470 |
+
"eval_runtime": 889.1985,
|
| 2471 |
+
"eval_samples_per_second": 867.995,
|
| 2472 |
+
"eval_steps_per_second": 54.25,
|
| 2473 |
+
"step": 1792000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 1.96,
|
| 2477 |
+
"eval_loss": 2.006197452545166,
|
| 2478 |
+
"eval_runtime": 889.8901,
|
| 2479 |
+
"eval_samples_per_second": 867.321,
|
| 2480 |
+
"eval_steps_per_second": 54.208,
|
| 2481 |
+
"step": 1800000
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 1.97,
|
| 2485 |
+
"learning_rate": 1.0113333333333334e-07,
|
| 2486 |
+
"loss": 2.109,
|
| 2487 |
+
"step": 1808000
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 1.97,
|
| 2491 |
+
"eval_loss": 2.0090434551239014,
|
| 2492 |
+
"eval_runtime": 888.3297,
|
| 2493 |
+
"eval_samples_per_second": 868.844,
|
| 2494 |
+
"eval_steps_per_second": 54.303,
|
| 2495 |
+
"step": 1808000
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 1.98,
|
| 2499 |
+
"eval_loss": 2.012356758117676,
|
| 2500 |
+
"eval_runtime": 893.3256,
|
| 2501 |
+
"eval_samples_per_second": 863.985,
|
| 2502 |
+
"eval_steps_per_second": 53.999,
|
| 2503 |
+
"step": 1816000
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 1.99,
|
| 2507 |
+
"learning_rate": 9.84e-08,
|
| 2508 |
+
"loss": 2.1081,
|
| 2509 |
+
"step": 1824000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 1.99,
|
| 2513 |
+
"eval_loss": 2.0065596103668213,
|
| 2514 |
+
"eval_runtime": 893.6122,
|
| 2515 |
+
"eval_samples_per_second": 863.708,
|
| 2516 |
+
"eval_steps_per_second": 53.982,
|
| 2517 |
+
"step": 1824000
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.0,
|
| 2521 |
+
"eval_loss": 2.008080005645752,
|
| 2522 |
+
"eval_runtime": 891.4247,
|
| 2523 |
+
"eval_samples_per_second": 865.828,
|
| 2524 |
+
"eval_steps_per_second": 54.115,
|
| 2525 |
+
"step": 1832000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 2.01,
|
| 2529 |
+
"learning_rate": 9.566666666666666e-08,
|
| 2530 |
+
"loss": 2.1151,
|
| 2531 |
+
"step": 1840000
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.01,
|
| 2535 |
+
"eval_loss": 2.008512258529663,
|
| 2536 |
+
"eval_runtime": 884.9554,
|
| 2537 |
+
"eval_samples_per_second": 872.157,
|
| 2538 |
+
"eval_steps_per_second": 54.51,
|
| 2539 |
+
"step": 1840000
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 2.02,
|
| 2543 |
+
"eval_loss": 2.0054173469543457,
|
| 2544 |
+
"eval_runtime": 886.9049,
|
| 2545 |
+
"eval_samples_per_second": 870.24,
|
| 2546 |
+
"eval_steps_per_second": 54.39,
|
| 2547 |
+
"step": 1848000
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 2.03,
|
| 2551 |
+
"learning_rate": 9.293333333333333e-08,
|
| 2552 |
+
"loss": 2.1178,
|
| 2553 |
+
"step": 1856000
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 2.03,
|
| 2557 |
+
"eval_loss": 2.005777359008789,
|
| 2558 |
+
"eval_runtime": 886.5315,
|
| 2559 |
+
"eval_samples_per_second": 870.606,
|
| 2560 |
+
"eval_steps_per_second": 54.413,
|
| 2561 |
+
"step": 1856000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 2.03,
|
| 2565 |
+
"eval_loss": 2.0048415660858154,
|
| 2566 |
+
"eval_runtime": 893.5519,
|
| 2567 |
+
"eval_samples_per_second": 863.766,
|
| 2568 |
+
"eval_steps_per_second": 53.986,
|
| 2569 |
+
"step": 1864000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 2.04,
|
| 2573 |
+
"learning_rate": 9.02e-08,
|
| 2574 |
+
"loss": 2.1035,
|
| 2575 |
+
"step": 1872000
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 2.04,
|
| 2579 |
+
"eval_loss": 2.004007339477539,
|
| 2580 |
+
"eval_runtime": 890.5358,
|
| 2581 |
+
"eval_samples_per_second": 866.692,
|
| 2582 |
+
"eval_steps_per_second": 54.169,
|
| 2583 |
+
"step": 1872000
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 2.05,
|
| 2587 |
+
"eval_loss": 2.0059244632720947,
|
| 2588 |
+
"eval_runtime": 887.0437,
|
| 2589 |
+
"eval_samples_per_second": 870.104,
|
| 2590 |
+
"eval_steps_per_second": 54.382,
|
| 2591 |
+
"step": 1880000
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 2.06,
|
| 2595 |
+
"learning_rate": 8.746666666666667e-08,
|
| 2596 |
+
"loss": 2.1197,
|
| 2597 |
+
"step": 1888000
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 2.06,
|
| 2601 |
+
"eval_loss": 2.0071017742156982,
|
| 2602 |
+
"eval_runtime": 889.191,
|
| 2603 |
+
"eval_samples_per_second": 868.003,
|
| 2604 |
+
"eval_steps_per_second": 54.25,
|
| 2605 |
+
"step": 1888000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 2.07,
|
| 2609 |
+
"eval_loss": 2.005682945251465,
|
| 2610 |
+
"eval_runtime": 888.8818,
|
| 2611 |
+
"eval_samples_per_second": 868.304,
|
| 2612 |
+
"eval_steps_per_second": 54.269,
|
| 2613 |
+
"step": 1896000
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 2.08,
|
| 2617 |
+
"learning_rate": 8.473333333333334e-08,
|
| 2618 |
+
"loss": 2.1143,
|
| 2619 |
+
"step": 1904000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 2.08,
|
| 2623 |
+
"eval_loss": 2.005943536758423,
|
| 2624 |
+
"eval_runtime": 884.5437,
|
| 2625 |
+
"eval_samples_per_second": 872.563,
|
| 2626 |
+
"eval_steps_per_second": 54.535,
|
| 2627 |
+
"step": 1904000
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 2.09,
|
| 2631 |
+
"eval_loss": 2.0042991638183594,
|
| 2632 |
+
"eval_runtime": 884.1715,
|
| 2633 |
+
"eval_samples_per_second": 872.93,
|
| 2634 |
+
"eval_steps_per_second": 54.558,
|
| 2635 |
+
"step": 1912000
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 2.09,
|
| 2639 |
+
"learning_rate": 8.2e-08,
|
| 2640 |
+
"loss": 2.1082,
|
| 2641 |
+
"step": 1920000
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.09,
|
| 2645 |
+
"eval_loss": 2.0067648887634277,
|
| 2646 |
+
"eval_runtime": 885.4828,
|
| 2647 |
+
"eval_samples_per_second": 871.637,
|
| 2648 |
+
"eval_steps_per_second": 54.478,
|
| 2649 |
+
"step": 1920000
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 2.1,
|
| 2653 |
+
"eval_loss": 2.0057313442230225,
|
| 2654 |
+
"eval_runtime": 887.8665,
|
| 2655 |
+
"eval_samples_per_second": 869.297,
|
| 2656 |
+
"eval_steps_per_second": 54.331,
|
| 2657 |
+
"step": 1928000
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.11,
|
| 2661 |
+
"learning_rate": 7.926666666666666e-08,
|
| 2662 |
+
"loss": 2.1202,
|
| 2663 |
+
"step": 1936000
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 2.11,
|
| 2667 |
+
"eval_loss": 2.007241725921631,
|
| 2668 |
+
"eval_runtime": 885.5971,
|
| 2669 |
+
"eval_samples_per_second": 871.525,
|
| 2670 |
+
"eval_steps_per_second": 54.471,
|
| 2671 |
+
"step": 1936000
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.12,
|
| 2675 |
+
"eval_loss": 2.0057430267333984,
|
| 2676 |
+
"eval_runtime": 888.4045,
|
| 2677 |
+
"eval_samples_per_second": 868.771,
|
| 2678 |
+
"eval_steps_per_second": 54.298,
|
| 2679 |
+
"step": 1944000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 2.13,
|
| 2683 |
+
"learning_rate": 7.653333333333333e-08,
|
| 2684 |
+
"loss": 2.1138,
|
| 2685 |
+
"step": 1952000
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.13,
|
| 2689 |
+
"eval_loss": 2.0051097869873047,
|
| 2690 |
+
"eval_runtime": 889.7536,
|
| 2691 |
+
"eval_samples_per_second": 867.454,
|
| 2692 |
+
"eval_steps_per_second": 54.216,
|
| 2693 |
+
"step": 1952000
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 2.14,
|
| 2697 |
+
"eval_loss": 2.008528709411621,
|
| 2698 |
+
"eval_runtime": 887.8548,
|
| 2699 |
+
"eval_samples_per_second": 869.309,
|
| 2700 |
+
"eval_steps_per_second": 54.332,
|
| 2701 |
+
"step": 1960000
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 2.15,
|
| 2705 |
+
"learning_rate": 7.38e-08,
|
| 2706 |
+
"loss": 2.1082,
|
| 2707 |
+
"step": 1968000
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 2.15,
|
| 2711 |
+
"eval_loss": 2.007629871368408,
|
| 2712 |
+
"eval_runtime": 886.2101,
|
| 2713 |
+
"eval_samples_per_second": 870.922,
|
| 2714 |
+
"eval_steps_per_second": 54.433,
|
| 2715 |
+
"step": 1968000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 2.16,
|
| 2719 |
+
"eval_loss": 2.0076658725738525,
|
| 2720 |
+
"eval_runtime": 886.4111,
|
| 2721 |
+
"eval_samples_per_second": 870.725,
|
| 2722 |
+
"eval_steps_per_second": 54.421,
|
| 2723 |
+
"step": 1976000
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 2.16,
|
| 2727 |
+
"learning_rate": 7.106666666666667e-08,
|
| 2728 |
+
"loss": 2.1084,
|
| 2729 |
+
"step": 1984000
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 2.16,
|
| 2733 |
+
"eval_loss": 2.001997470855713,
|
| 2734 |
+
"eval_runtime": 885.1567,
|
| 2735 |
+
"eval_samples_per_second": 871.959,
|
| 2736 |
+
"eval_steps_per_second": 54.498,
|
| 2737 |
+
"step": 1984000
|
| 2738 |
+
},
|
| 2739 |
+
{
|
| 2740 |
+
"epoch": 2.17,
|
| 2741 |
+
"eval_loss": 2.005009651184082,
|
| 2742 |
+
"eval_runtime": 889.5629,
|
| 2743 |
+
"eval_samples_per_second": 867.64,
|
| 2744 |
+
"eval_steps_per_second": 54.228,
|
| 2745 |
+
"step": 1992000
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 2.18,
|
| 2749 |
+
"learning_rate": 6.833333333333332e-08,
|
| 2750 |
+
"loss": 2.1151,
|
| 2751 |
+
"step": 2000000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 2.18,
|
| 2755 |
+
"eval_loss": 2.0065817832946777,
|
| 2756 |
+
"eval_runtime": 885.7641,
|
| 2757 |
+
"eval_samples_per_second": 871.361,
|
| 2758 |
+
"eval_steps_per_second": 54.46,
|
| 2759 |
+
"step": 2000000
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 2.19,
|
| 2763 |
+
"eval_loss": 2.003136396408081,
|
| 2764 |
+
"eval_runtime": 886.578,
|
| 2765 |
+
"eval_samples_per_second": 870.561,
|
| 2766 |
+
"eval_steps_per_second": 54.41,
|
| 2767 |
+
"step": 2008000
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 2.2,
|
| 2771 |
+
"learning_rate": 6.56e-08,
|
| 2772 |
+
"loss": 2.1141,
|
| 2773 |
+
"step": 2016000
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 2.2,
|
| 2777 |
+
"eval_loss": 2.0128238201141357,
|
| 2778 |
+
"eval_runtime": 891.0219,
|
| 2779 |
+
"eval_samples_per_second": 866.219,
|
| 2780 |
+
"eval_steps_per_second": 54.139,
|
| 2781 |
+
"step": 2016000
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.21,
|
| 2785 |
+
"eval_loss": 2.0021839141845703,
|
| 2786 |
+
"eval_runtime": 895.8435,
|
| 2787 |
+
"eval_samples_per_second": 861.557,
|
| 2788 |
+
"eval_steps_per_second": 53.848,
|
| 2789 |
+
"step": 2024000
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 2.22,
|
| 2793 |
+
"learning_rate": 6.286666666666666e-08,
|
| 2794 |
+
"loss": 2.1129,
|
| 2795 |
+
"step": 2032000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.22,
|
| 2799 |
+
"eval_loss": 2.0065131187438965,
|
| 2800 |
+
"eval_runtime": 890.2528,
|
| 2801 |
+
"eval_samples_per_second": 866.967,
|
| 2802 |
+
"eval_steps_per_second": 54.186,
|
| 2803 |
+
"step": 2032000
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 2.23,
|
| 2807 |
+
"eval_loss": 2.005363941192627,
|
| 2808 |
+
"eval_runtime": 890.9681,
|
| 2809 |
+
"eval_samples_per_second": 866.271,
|
| 2810 |
+
"eval_steps_per_second": 54.142,
|
| 2811 |
+
"step": 2040000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.23,
|
| 2815 |
+
"learning_rate": 6.013333333333333e-08,
|
| 2816 |
+
"loss": 2.1164,
|
| 2817 |
+
"step": 2048000
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 2.23,
|
| 2821 |
+
"eval_loss": 2.0038933753967285,
|
| 2822 |
+
"eval_runtime": 892.3995,
|
| 2823 |
+
"eval_samples_per_second": 864.882,
|
| 2824 |
+
"eval_steps_per_second": 54.055,
|
| 2825 |
+
"step": 2048000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.24,
|
| 2829 |
+
"eval_loss": 2.003117561340332,
|
| 2830 |
+
"eval_runtime": 894.495,
|
| 2831 |
+
"eval_samples_per_second": 862.856,
|
| 2832 |
+
"eval_steps_per_second": 53.929,
|
| 2833 |
+
"step": 2056000
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 2.25,
|
| 2837 |
+
"learning_rate": 5.7400000000000004e-08,
|
| 2838 |
+
"loss": 2.1121,
|
| 2839 |
+
"step": 2064000
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.25,
|
| 2843 |
+
"eval_loss": 2.0101029872894287,
|
| 2844 |
+
"eval_runtime": 886.6646,
|
| 2845 |
+
"eval_samples_per_second": 870.476,
|
| 2846 |
+
"eval_steps_per_second": 54.405,
|
| 2847 |
+
"step": 2064000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 2.26,
|
| 2851 |
+
"eval_loss": 2.0098650455474854,
|
| 2852 |
+
"eval_runtime": 887.3882,
|
| 2853 |
+
"eval_samples_per_second": 869.766,
|
| 2854 |
+
"eval_steps_per_second": 54.361,
|
| 2855 |
+
"step": 2072000
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 2.27,
|
| 2859 |
+
"learning_rate": 5.4666666666666666e-08,
|
| 2860 |
+
"loss": 2.1071,
|
| 2861 |
+
"step": 2080000
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 2.27,
|
| 2865 |
+
"eval_loss": 2.0041701793670654,
|
| 2866 |
+
"eval_runtime": 891.5578,
|
| 2867 |
+
"eval_samples_per_second": 865.698,
|
| 2868 |
+
"eval_steps_per_second": 54.106,
|
| 2869 |
+
"step": 2080000
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 2.28,
|
| 2873 |
+
"eval_loss": 2.0030367374420166,
|
| 2874 |
+
"eval_runtime": 886.7055,
|
| 2875 |
+
"eval_samples_per_second": 870.436,
|
| 2876 |
+
"eval_steps_per_second": 54.403,
|
| 2877 |
+
"step": 2088000
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 2.29,
|
| 2881 |
+
"learning_rate": 5.1933333333333335e-08,
|
| 2882 |
+
"loss": 2.1094,
|
| 2883 |
+
"step": 2096000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 2.29,
|
| 2887 |
+
"eval_loss": 2.00482439994812,
|
| 2888 |
+
"eval_runtime": 887.8886,
|
| 2889 |
+
"eval_samples_per_second": 869.276,
|
| 2890 |
+
"eval_steps_per_second": 54.33,
|
| 2891 |
+
"step": 2096000
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 2.3,
|
| 2895 |
+
"eval_loss": 2.004595994949341,
|
| 2896 |
+
"eval_runtime": 887.4455,
|
| 2897 |
+
"eval_samples_per_second": 869.71,
|
| 2898 |
+
"eval_steps_per_second": 54.357,
|
| 2899 |
+
"step": 2104000
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 2.3,
|
| 2903 |
+
"learning_rate": 4.92e-08,
|
| 2904 |
+
"loss": 2.1017,
|
| 2905 |
+
"step": 2112000
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 2.3,
|
| 2909 |
+
"eval_loss": 2.0038633346557617,
|
| 2910 |
+
"eval_runtime": 888.4121,
|
| 2911 |
+
"eval_samples_per_second": 868.764,
|
| 2912 |
+
"eval_steps_per_second": 54.298,
|
| 2913 |
+
"step": 2112000
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 2.31,
|
| 2917 |
+
"eval_loss": 2.0011472702026367,
|
| 2918 |
+
"eval_runtime": 889.7748,
|
| 2919 |
+
"eval_samples_per_second": 867.433,
|
| 2920 |
+
"eval_steps_per_second": 54.215,
|
| 2921 |
+
"step": 2120000
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.32,
|
| 2925 |
+
"learning_rate": 4.6466666666666666e-08,
|
| 2926 |
+
"loss": 2.1124,
|
| 2927 |
+
"step": 2128000
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 2.32,
|
| 2931 |
+
"eval_loss": 2.007091522216797,
|
| 2932 |
+
"eval_runtime": 892.2658,
|
| 2933 |
+
"eval_samples_per_second": 865.011,
|
| 2934 |
+
"eval_steps_per_second": 54.063,
|
| 2935 |
+
"step": 2128000
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 2.33,
|
| 2939 |
+
"eval_loss": 2.0060718059539795,
|
| 2940 |
+
"eval_runtime": 887.502,
|
| 2941 |
+
"eval_samples_per_second": 869.654,
|
| 2942 |
+
"eval_steps_per_second": 54.354,
|
| 2943 |
+
"step": 2136000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 2.34,
|
| 2947 |
+
"learning_rate": 4.3733333333333335e-08,
|
| 2948 |
+
"loss": 2.1064,
|
| 2949 |
+
"step": 2144000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 2.34,
|
| 2953 |
+
"eval_loss": 2.0040297508239746,
|
| 2954 |
+
"eval_runtime": 888.8512,
|
| 2955 |
+
"eval_samples_per_second": 868.334,
|
| 2956 |
+
"eval_steps_per_second": 54.271,
|
| 2957 |
+
"step": 2144000
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 2.35,
|
| 2961 |
+
"eval_loss": 2.007528066635132,
|
| 2962 |
+
"eval_runtime": 895.8909,
|
| 2963 |
+
"eval_samples_per_second": 861.511,
|
| 2964 |
+
"eval_steps_per_second": 53.845,
|
| 2965 |
+
"step": 2152000
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.36,
|
| 2969 |
+
"learning_rate": 4.1e-08,
|
| 2970 |
+
"loss": 2.115,
|
| 2971 |
+
"step": 2160000
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 2.36,
|
| 2975 |
+
"eval_loss": 2.0025811195373535,
|
| 2976 |
+
"eval_runtime": 894.6822,
|
| 2977 |
+
"eval_samples_per_second": 862.675,
|
| 2978 |
+
"eval_steps_per_second": 53.917,
|
| 2979 |
+
"step": 2160000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.37,
|
| 2983 |
+
"eval_loss": 2.006788492202759,
|
| 2984 |
+
"eval_runtime": 885.9111,
|
| 2985 |
+
"eval_samples_per_second": 871.216,
|
| 2986 |
+
"eval_steps_per_second": 54.451,
|
| 2987 |
+
"step": 2168000
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 2.37,
|
| 2991 |
+
"learning_rate": 3.8266666666666665e-08,
|
| 2992 |
+
"loss": 2.114,
|
| 2993 |
+
"step": 2176000
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.37,
|
| 2997 |
+
"eval_loss": 2.006558418273926,
|
| 2998 |
+
"eval_runtime": 889.8092,
|
| 2999 |
+
"eval_samples_per_second": 867.399,
|
| 3000 |
+
"eval_steps_per_second": 54.213,
|
| 3001 |
+
"step": 2176000
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 2.38,
|
| 3005 |
+
"eval_loss": 2.0079538822174072,
|
| 3006 |
+
"eval_runtime": 889.2248,
|
| 3007 |
+
"eval_samples_per_second": 867.97,
|
| 3008 |
+
"eval_steps_per_second": 54.248,
|
| 3009 |
+
"step": 2184000
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 2.39,
|
| 3013 |
+
"learning_rate": 3.5533333333333334e-08,
|
| 3014 |
+
"loss": 2.1171,
|
| 3015 |
+
"step": 2192000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 2.39,
|
| 3019 |
+
"eval_loss": 2.0031957626342773,
|
| 3020 |
+
"eval_runtime": 891.062,
|
| 3021 |
+
"eval_samples_per_second": 866.18,
|
| 3022 |
+
"eval_steps_per_second": 54.137,
|
| 3023 |
+
"step": 2192000
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 2.4,
|
| 3027 |
+
"eval_loss": 2.0036396980285645,
|
| 3028 |
+
"eval_runtime": 889.4858,
|
| 3029 |
+
"eval_samples_per_second": 867.715,
|
| 3030 |
+
"eval_steps_per_second": 54.232,
|
| 3031 |
+
"step": 2200000
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 2.41,
|
| 3035 |
+
"learning_rate": 3.28e-08,
|
| 3036 |
+
"loss": 2.1119,
|
| 3037 |
+
"step": 2208000
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 2.41,
|
| 3041 |
+
"eval_loss": 2.004848003387451,
|
| 3042 |
+
"eval_runtime": 890.2659,
|
| 3043 |
+
"eval_samples_per_second": 866.954,
|
| 3044 |
+
"eval_steps_per_second": 54.185,
|
| 3045 |
+
"step": 2208000
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 2.42,
|
| 3049 |
+
"eval_loss": 2.0058629512786865,
|
| 3050 |
+
"eval_runtime": 890.6135,
|
| 3051 |
+
"eval_samples_per_second": 866.616,
|
| 3052 |
+
"eval_steps_per_second": 54.164,
|
| 3053 |
+
"step": 2216000
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 2.43,
|
| 3057 |
+
"learning_rate": 3.0066666666666665e-08,
|
| 3058 |
+
"loss": 2.1097,
|
| 3059 |
+
"step": 2224000
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 2.43,
|
| 3063 |
+
"eval_loss": 2.005845546722412,
|
| 3064 |
+
"eval_runtime": 889.9256,
|
| 3065 |
+
"eval_samples_per_second": 867.286,
|
| 3066 |
+
"eval_steps_per_second": 54.206,
|
| 3067 |
+
"step": 2224000
|
| 3068 |
+
},
|
| 3069 |
+
{
|
| 3070 |
+
"epoch": 2.44,
|
| 3071 |
+
"eval_loss": 2.004934310913086,
|
| 3072 |
+
"eval_runtime": 893.1468,
|
| 3073 |
+
"eval_samples_per_second": 864.158,
|
| 3074 |
+
"eval_steps_per_second": 54.01,
|
| 3075 |
+
"step": 2232000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 2.44,
|
| 3079 |
+
"learning_rate": 2.7333333333333333e-08,
|
| 3080 |
+
"loss": 2.1091,
|
| 3081 |
+
"step": 2240000
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 2.44,
|
| 3085 |
+
"eval_loss": 2.005760669708252,
|
| 3086 |
+
"eval_runtime": 893.6832,
|
| 3087 |
+
"eval_samples_per_second": 863.639,
|
| 3088 |
+
"eval_steps_per_second": 53.978,
|
| 3089 |
+
"step": 2240000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 2.45,
|
| 3093 |
+
"eval_loss": 2.0032405853271484,
|
| 3094 |
+
"eval_runtime": 894.8171,
|
| 3095 |
+
"eval_samples_per_second": 862.545,
|
| 3096 |
+
"eval_steps_per_second": 53.909,
|
| 3097 |
+
"step": 2248000
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 2.46,
|
| 3101 |
+
"learning_rate": 2.46e-08,
|
| 3102 |
+
"loss": 2.1107,
|
| 3103 |
+
"step": 2256000
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 2.46,
|
| 3107 |
+
"eval_loss": 2.00769305229187,
|
| 3108 |
+
"eval_runtime": 893.4774,
|
| 3109 |
+
"eval_samples_per_second": 863.838,
|
| 3110 |
+
"eval_steps_per_second": 53.99,
|
| 3111 |
+
"step": 2256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 2.47,
|
| 3115 |
+
"eval_loss": 2.0032243728637695,
|
| 3116 |
+
"eval_runtime": 893.6019,
|
| 3117 |
+
"eval_samples_per_second": 863.718,
|
| 3118 |
+
"eval_steps_per_second": 53.983,
|
| 3119 |
+
"step": 2264000
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.48,
|
| 3123 |
+
"learning_rate": 2.1866666666666667e-08,
|
| 3124 |
+
"loss": 2.1126,
|
| 3125 |
+
"step": 2272000
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 2.48,
|
| 3129 |
+
"eval_loss": 2.0055274963378906,
|
| 3130 |
+
"eval_runtime": 891.7304,
|
| 3131 |
+
"eval_samples_per_second": 865.531,
|
| 3132 |
+
"eval_steps_per_second": 54.096,
|
| 3133 |
+
"step": 2272000
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.49,
|
| 3137 |
+
"eval_loss": 2.002612590789795,
|
| 3138 |
+
"eval_runtime": 892.1014,
|
| 3139 |
+
"eval_samples_per_second": 865.171,
|
| 3140 |
+
"eval_steps_per_second": 54.073,
|
| 3141 |
+
"step": 2280000
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 2.5,
|
| 3145 |
+
"learning_rate": 1.9133333333333333e-08,
|
| 3146 |
+
"loss": 2.1173,
|
| 3147 |
+
"step": 2288000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.5,
|
| 3151 |
+
"eval_loss": 2.0062429904937744,
|
| 3152 |
+
"eval_runtime": 891.9249,
|
| 3153 |
+
"eval_samples_per_second": 865.342,
|
| 3154 |
+
"eval_steps_per_second": 54.084,
|
| 3155 |
+
"step": 2288000
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 2.51,
|
| 3159 |
+
"eval_loss": 2.003859043121338,
|
| 3160 |
+
"eval_runtime": 892.8008,
|
| 3161 |
+
"eval_samples_per_second": 864.493,
|
| 3162 |
+
"eval_steps_per_second": 54.031,
|
| 3163 |
+
"step": 2296000
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 2.51,
|
| 3167 |
+
"learning_rate": 1.64e-08,
|
| 3168 |
+
"loss": 2.114,
|
| 3169 |
+
"step": 2304000
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 2.51,
|
| 3173 |
+
"eval_loss": 2.006359100341797,
|
| 3174 |
+
"eval_runtime": 891.1547,
|
| 3175 |
+
"eval_samples_per_second": 866.09,
|
| 3176 |
+
"eval_steps_per_second": 54.131,
|
| 3177 |
+
"step": 2304000
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 2.52,
|
| 3181 |
+
"eval_loss": 2.0113308429718018,
|
| 3182 |
+
"eval_runtime": 890.136,
|
| 3183 |
+
"eval_samples_per_second": 867.081,
|
| 3184 |
+
"eval_steps_per_second": 54.193,
|
| 3185 |
+
"step": 2312000
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 2.53,
|
| 3189 |
+
"learning_rate": 1.3666666666666667e-08,
|
| 3190 |
+
"loss": 2.1131,
|
| 3191 |
+
"step": 2320000
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 2.53,
|
| 3195 |
+
"eval_loss": 2.0065314769744873,
|
| 3196 |
+
"eval_runtime": 890.6924,
|
| 3197 |
+
"eval_samples_per_second": 866.539,
|
| 3198 |
+
"eval_steps_per_second": 54.159,
|
| 3199 |
+
"step": 2320000
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 2.54,
|
| 3203 |
+
"eval_loss": 2.0098392963409424,
|
| 3204 |
+
"eval_runtime": 892.2668,
|
| 3205 |
+
"eval_samples_per_second": 865.01,
|
| 3206 |
+
"eval_steps_per_second": 54.063,
|
| 3207 |
+
"step": 2328000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 2.55,
|
| 3211 |
+
"learning_rate": 1.0933333333333334e-08,
|
| 3212 |
+
"loss": 2.1045,
|
| 3213 |
+
"step": 2336000
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 2.55,
|
| 3217 |
+
"eval_loss": 2.0060501098632812,
|
| 3218 |
+
"eval_runtime": 891.9301,
|
| 3219 |
+
"eval_samples_per_second": 865.337,
|
| 3220 |
+
"eval_steps_per_second": 54.084,
|
| 3221 |
+
"step": 2336000
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"epoch": 2.56,
|
| 3225 |
+
"eval_loss": 2.006572961807251,
|
| 3226 |
+
"eval_runtime": 894.7549,
|
| 3227 |
+
"eval_samples_per_second": 862.605,
|
| 3228 |
+
"eval_steps_per_second": 53.913,
|
| 3229 |
+
"step": 2344000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 2.57,
|
| 3233 |
+
"learning_rate": 8.2e-09,
|
| 3234 |
+
"loss": 2.1144,
|
| 3235 |
+
"step": 2352000
|
| 3236 |
+
},
|
| 3237 |
+
{
|
| 3238 |
+
"epoch": 2.57,
|
| 3239 |
+
"eval_loss": 2.006028175354004,
|
| 3240 |
+
"eval_runtime": 899.347,
|
| 3241 |
+
"eval_samples_per_second": 858.2,
|
| 3242 |
+
"eval_steps_per_second": 53.638,
|
| 3243 |
+
"step": 2352000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 2.57,
|
| 3247 |
+
"eval_loss": 2.00589656829834,
|
| 3248 |
+
"eval_runtime": 893.5452,
|
| 3249 |
+
"eval_samples_per_second": 863.773,
|
| 3250 |
+
"eval_steps_per_second": 53.986,
|
| 3251 |
+
"step": 2360000
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 2.58,
|
| 3255 |
+
"learning_rate": 5.466666666666667e-09,
|
| 3256 |
+
"loss": 2.1086,
|
| 3257 |
+
"step": 2368000
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 2.58,
|
| 3261 |
+
"eval_loss": 2.0038540363311768,
|
| 3262 |
+
"eval_runtime": 893.2561,
|
| 3263 |
+
"eval_samples_per_second": 864.052,
|
| 3264 |
+
"eval_steps_per_second": 54.004,
|
| 3265 |
+
"step": 2368000
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 2.59,
|
| 3269 |
+
"eval_loss": 2.0076115131378174,
|
| 3270 |
+
"eval_runtime": 895.0756,
|
| 3271 |
+
"eval_samples_per_second": 862.296,
|
| 3272 |
+
"eval_steps_per_second": 53.894,
|
| 3273 |
+
"step": 2376000
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 2.6,
|
| 3277 |
+
"learning_rate": 2.7333333333333334e-09,
|
| 3278 |
+
"loss": 2.1058,
|
| 3279 |
+
"step": 2384000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 2.6,
|
| 3283 |
+
"eval_loss": 2.0035552978515625,
|
| 3284 |
+
"eval_runtime": 895.3228,
|
| 3285 |
+
"eval_samples_per_second": 862.058,
|
| 3286 |
+
"eval_steps_per_second": 53.879,
|
| 3287 |
+
"step": 2384000
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 2.61,
|
| 3291 |
+
"eval_loss": 2.0077223777770996,
|
| 3292 |
+
"eval_runtime": 896.1834,
|
| 3293 |
+
"eval_samples_per_second": 861.23,
|
| 3294 |
+
"eval_steps_per_second": 53.827,
|
| 3295 |
+
"step": 2392000
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 2.62,
|
| 3299 |
+
"learning_rate": 0.0,
|
| 3300 |
+
"loss": 2.1112,
|
| 3301 |
+
"step": 2400000
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 2.62,
|
| 3305 |
+
"eval_loss": 2.000014066696167,
|
| 3306 |
+
"eval_runtime": 893.9091,
|
| 3307 |
+
"eval_samples_per_second": 863.421,
|
| 3308 |
+
"eval_steps_per_second": 53.964,
|
| 3309 |
+
"step": 2400000
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 2.62,
|
| 3313 |
+
"step": 2400000,
|
| 3314 |
+
"total_flos": 7.571300080769916e+17,
|
| 3315 |
+
"train_loss": 2.133689431966146,
|
| 3316 |
+
"train_runtime": 416842.919,
|
| 3317 |
+
"train_samples_per_second": 92.121,
|
| 3318 |
+
"train_steps_per_second": 5.758
|
| 3319 |
+
}
|
| 3320 |
+
],
|
| 3321 |
+
"logging_steps": 16000,
|
| 3322 |
+
"max_steps": 2400000,
|
| 3323 |
+
"num_train_epochs": 3,
|
| 3324 |
+
"save_steps": 32000,
|
| 3325 |
+
"total_flos": 7.571300080769916e+17,
|
| 3326 |
+
"trial_name": null,
|
| 3327 |
+
"trial_params": null
|
| 3328 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26db5bc9451617be9ea4a5124590ef6eba8ab3d962b597835074501dc7de58e6
|
| 3 |
+
size 4219
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|