Evaluation on the test set completed on 2025_01_29.
Browse files- README.md +152 -1
- all_results.json +16 -0
- logs/events.out.tfevents.1737978278.datavisu3 +2 -2
- logs/events.out.tfevents.1738149112.datavisu3 +3 -0
- model.safetensors +1 -1
- test_results.json +11 -0
- train_results.json +9 -0
- trainer_state.json +1487 -0
README.md
CHANGED
|
@@ -1,4 +1,155 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
| 2 |
tags:
|
| 3 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: facebook/dinov2-large
|
| 5 |
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
metrics:
|
| 8 |
+
- accuracy
|
| 9 |
+
model-index:
|
| 10 |
+
- name: DinoVdo-large-2025_01_27_45863-bs32_freeze
|
| 11 |
+
results: []
|
| 12 |
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# DinoVdo-large-2025_01_27_45863-bs32_freeze
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the None dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.1236
|
| 22 |
+
- F1 Micro: 0.8136
|
| 23 |
+
- F1 Macro: 0.7060
|
| 24 |
+
- Accuracy: 0.3071
|
| 25 |
+
- Learning Rate: 0.0000
|
| 26 |
+
|
| 27 |
+
## Model description
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Intended uses & limitations
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training and evaluation data
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training procedure
|
| 40 |
+
|
| 41 |
+
### Training hyperparameters
|
| 42 |
+
|
| 43 |
+
The following hyperparameters were used during training:
|
| 44 |
+
- learning_rate: 0.001
|
| 45 |
+
- train_batch_size: 32
|
| 46 |
+
- eval_batch_size: 32
|
| 47 |
+
- seed: 42
|
| 48 |
+
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 49 |
+
- lr_scheduler_type: linear
|
| 50 |
+
- num_epochs: 150
|
| 51 |
+
- mixed_precision_training: Native AMP
|
| 52 |
+
|
| 53 |
+
### Training results
|
| 54 |
+
|
| 55 |
+
| Training Loss | Epoch | Step | Validation Loss | F1 Micro | F1 Macro | Accuracy | Rate |
|
| 56 |
+
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|:--------:|:------:|
|
| 57 |
+
| No log | 1.0 | 273 | 0.1678 | 0.7433 | 0.5138 | 0.2293 | 0.001 |
|
| 58 |
+
| 0.2722 | 2.0 | 546 | 0.1537 | 0.7614 | 0.5771 | 0.2436 | 0.001 |
|
| 59 |
+
| 0.2722 | 3.0 | 819 | 0.1483 | 0.7763 | 0.6194 | 0.2415 | 0.001 |
|
| 60 |
+
| 0.169 | 4.0 | 1092 | 0.1464 | 0.7808 | 0.6276 | 0.2555 | 0.001 |
|
| 61 |
+
| 0.169 | 5.0 | 1365 | 0.1452 | 0.7788 | 0.6421 | 0.2520 | 0.001 |
|
| 62 |
+
| 0.1612 | 6.0 | 1638 | 0.1440 | 0.7802 | 0.6147 | 0.2579 | 0.001 |
|
| 63 |
+
| 0.1612 | 7.0 | 1911 | 0.1452 | 0.7787 | 0.6141 | 0.2503 | 0.001 |
|
| 64 |
+
| 0.1594 | 8.0 | 2184 | 0.1446 | 0.7776 | 0.6193 | 0.2534 | 0.001 |
|
| 65 |
+
| 0.1594 | 9.0 | 2457 | 0.1479 | 0.7813 | 0.6363 | 0.2454 | 0.001 |
|
| 66 |
+
| 0.1567 | 10.0 | 2730 | 0.1425 | 0.7866 | 0.6366 | 0.2607 | 0.001 |
|
| 67 |
+
| 0.1559 | 11.0 | 3003 | 0.1454 | 0.7908 | 0.6566 | 0.2618 | 0.001 |
|
| 68 |
+
| 0.1559 | 12.0 | 3276 | 0.1404 | 0.7895 | 0.6439 | 0.2600 | 0.001 |
|
| 69 |
+
| 0.1549 | 13.0 | 3549 | 0.1414 | 0.7883 | 0.6490 | 0.2531 | 0.001 |
|
| 70 |
+
| 0.1549 | 14.0 | 3822 | 0.1406 | 0.7894 | 0.6348 | 0.2649 | 0.001 |
|
| 71 |
+
| 0.155 | 15.0 | 4095 | 0.1387 | 0.7906 | 0.6470 | 0.2632 | 0.001 |
|
| 72 |
+
| 0.155 | 16.0 | 4368 | 0.1401 | 0.7858 | 0.6331 | 0.2604 | 0.001 |
|
| 73 |
+
| 0.1531 | 17.0 | 4641 | 0.1381 | 0.7955 | 0.6651 | 0.2520 | 0.001 |
|
| 74 |
+
| 0.1531 | 18.0 | 4914 | 0.1387 | 0.7914 | 0.6498 | 0.2702 | 0.001 |
|
| 75 |
+
| 0.1549 | 19.0 | 5187 | 0.1374 | 0.7940 | 0.6356 | 0.2632 | 0.001 |
|
| 76 |
+
| 0.1549 | 20.0 | 5460 | 0.1386 | 0.7850 | 0.6393 | 0.2551 | 0.001 |
|
| 77 |
+
| 0.1524 | 21.0 | 5733 | 0.1357 | 0.7943 | 0.6496 | 0.2646 | 0.001 |
|
| 78 |
+
| 0.1519 | 22.0 | 6006 | 0.1372 | 0.7972 | 0.6400 | 0.2723 | 0.001 |
|
| 79 |
+
| 0.1519 | 23.0 | 6279 | 0.1361 | 0.7938 | 0.6595 | 0.2604 | 0.001 |
|
| 80 |
+
| 0.1528 | 24.0 | 6552 | 0.1363 | 0.7954 | 0.6418 | 0.2649 | 0.001 |
|
| 81 |
+
| 0.1528 | 25.0 | 6825 | 0.1359 | 0.7963 | 0.6500 | 0.2733 | 0.001 |
|
| 82 |
+
| 0.1524 | 26.0 | 7098 | 0.1348 | 0.7984 | 0.6555 | 0.2691 | 0.001 |
|
| 83 |
+
| 0.1524 | 27.0 | 7371 | 0.1367 | 0.7944 | 0.6535 | 0.2688 | 0.001 |
|
| 84 |
+
| 0.1521 | 28.0 | 7644 | 0.1357 | 0.7923 | 0.6398 | 0.2677 | 0.001 |
|
| 85 |
+
| 0.1521 | 29.0 | 7917 | 0.1405 | 0.7924 | 0.6637 | 0.2639 | 0.001 |
|
| 86 |
+
| 0.1525 | 30.0 | 8190 | 0.1373 | 0.7875 | 0.6405 | 0.2723 | 0.001 |
|
| 87 |
+
| 0.1525 | 31.0 | 8463 | 0.1354 | 0.7986 | 0.6562 | 0.2719 | 0.001 |
|
| 88 |
+
| 0.1522 | 32.0 | 8736 | 0.1369 | 0.7930 | 0.6463 | 0.2649 | 0.001 |
|
| 89 |
+
| 0.1486 | 33.0 | 9009 | 0.1320 | 0.8015 | 0.6660 | 0.2747 | 0.0001 |
|
| 90 |
+
| 0.1486 | 34.0 | 9282 | 0.1300 | 0.8043 | 0.6756 | 0.2834 | 0.0001 |
|
| 91 |
+
| 0.1419 | 35.0 | 9555 | 0.1295 | 0.8066 | 0.6772 | 0.2810 | 0.0001 |
|
| 92 |
+
| 0.1419 | 36.0 | 9828 | 0.1309 | 0.8046 | 0.6795 | 0.2817 | 0.0001 |
|
| 93 |
+
| 0.1396 | 37.0 | 10101 | 0.1279 | 0.8054 | 0.6792 | 0.2824 | 0.0001 |
|
| 94 |
+
| 0.1396 | 38.0 | 10374 | 0.1290 | 0.8078 | 0.6814 | 0.2848 | 0.0001 |
|
| 95 |
+
| 0.1366 | 39.0 | 10647 | 0.1272 | 0.8116 | 0.6833 | 0.2939 | 0.0001 |
|
| 96 |
+
| 0.1366 | 40.0 | 10920 | 0.1293 | 0.8116 | 0.6879 | 0.2908 | 0.0001 |
|
| 97 |
+
| 0.1361 | 41.0 | 11193 | 0.1270 | 0.8089 | 0.6864 | 0.2918 | 0.0001 |
|
| 98 |
+
| 0.1361 | 42.0 | 11466 | 0.1262 | 0.8109 | 0.6837 | 0.2914 | 0.0001 |
|
| 99 |
+
| 0.135 | 43.0 | 11739 | 0.1261 | 0.8123 | 0.6984 | 0.2949 | 0.0001 |
|
| 100 |
+
| 0.134 | 44.0 | 12012 | 0.1283 | 0.8106 | 0.6834 | 0.2935 | 0.0001 |
|
| 101 |
+
| 0.134 | 45.0 | 12285 | 0.1262 | 0.8113 | 0.7010 | 0.2932 | 0.0001 |
|
| 102 |
+
| 0.1331 | 46.0 | 12558 | 0.1246 | 0.8147 | 0.6964 | 0.2960 | 0.0001 |
|
| 103 |
+
| 0.1331 | 47.0 | 12831 | 0.1253 | 0.8126 | 0.6923 | 0.2988 | 0.0001 |
|
| 104 |
+
| 0.1325 | 48.0 | 13104 | 0.1263 | 0.8133 | 0.6954 | 0.2977 | 0.0001 |
|
| 105 |
+
| 0.1325 | 49.0 | 13377 | 0.1253 | 0.8158 | 0.6952 | 0.3037 | 0.0001 |
|
| 106 |
+
| 0.1312 | 50.0 | 13650 | 0.1263 | 0.8136 | 0.7008 | 0.3005 | 0.0001 |
|
| 107 |
+
| 0.1312 | 51.0 | 13923 | 0.1246 | 0.8158 | 0.7019 | 0.3009 | 0.0001 |
|
| 108 |
+
| 0.1301 | 52.0 | 14196 | 0.1253 | 0.8092 | 0.6949 | 0.2911 | 0.0001 |
|
| 109 |
+
| 0.1301 | 53.0 | 14469 | 0.1244 | 0.8154 | 0.7019 | 0.3023 | 1e-05 |
|
| 110 |
+
| 0.1306 | 54.0 | 14742 | 0.1249 | 0.8154 | 0.7040 | 0.3009 | 1e-05 |
|
| 111 |
+
| 0.1282 | 55.0 | 15015 | 0.1237 | 0.8144 | 0.6998 | 0.3005 | 1e-05 |
|
| 112 |
+
| 0.1282 | 56.0 | 15288 | 0.1235 | 0.8168 | 0.7004 | 0.3033 | 1e-05 |
|
| 113 |
+
| 0.1281 | 57.0 | 15561 | 0.1239 | 0.8157 | 0.7002 | 0.3030 | 1e-05 |
|
| 114 |
+
| 0.1281 | 58.0 | 15834 | 0.1234 | 0.8157 | 0.6995 | 0.3026 | 1e-05 |
|
| 115 |
+
| 0.129 | 59.0 | 16107 | 0.1235 | 0.8150 | 0.7012 | 0.3047 | 1e-05 |
|
| 116 |
+
| 0.129 | 60.0 | 16380 | 0.1239 | 0.8128 | 0.6932 | 0.2981 | 1e-05 |
|
| 117 |
+
| 0.1284 | 61.0 | 16653 | 0.1240 | 0.8174 | 0.7076 | 0.3009 | 1e-05 |
|
| 118 |
+
| 0.1284 | 62.0 | 16926 | 0.1233 | 0.8152 | 0.7032 | 0.3019 | 1e-05 |
|
| 119 |
+
| 0.127 | 63.0 | 17199 | 0.1233 | 0.8158 | 0.7023 | 0.3023 | 1e-05 |
|
| 120 |
+
| 0.127 | 64.0 | 17472 | 0.1235 | 0.8152 | 0.6999 | 0.3047 | 1e-05 |
|
| 121 |
+
| 0.1279 | 65.0 | 17745 | 0.1232 | 0.8145 | 0.7001 | 0.2977 | 1e-05 |
|
| 122 |
+
| 0.1273 | 66.0 | 18018 | 0.1229 | 0.8145 | 0.7004 | 0.3012 | 1e-05 |
|
| 123 |
+
| 0.1273 | 67.0 | 18291 | 0.1238 | 0.8159 | 0.6993 | 0.3016 | 1e-05 |
|
| 124 |
+
| 0.1272 | 68.0 | 18564 | 0.1229 | 0.8176 | 0.7039 | 0.2998 | 1e-05 |
|
| 125 |
+
| 0.1272 | 69.0 | 18837 | 0.1230 | 0.8157 | 0.7006 | 0.3058 | 1e-05 |
|
| 126 |
+
| 0.1274 | 70.0 | 19110 | 0.1228 | 0.8170 | 0.7009 | 0.3040 | 1e-05 |
|
| 127 |
+
| 0.1274 | 71.0 | 19383 | 0.1230 | 0.8158 | 0.7043 | 0.3019 | 1e-05 |
|
| 128 |
+
| 0.1272 | 72.0 | 19656 | 0.1232 | 0.8171 | 0.7025 | 0.3016 | 1e-05 |
|
| 129 |
+
| 0.1272 | 73.0 | 19929 | 0.1229 | 0.8174 | 0.7049 | 0.3054 | 1e-05 |
|
| 130 |
+
| 0.1267 | 74.0 | 20202 | 0.1230 | 0.8141 | 0.6942 | 0.3009 | 1e-05 |
|
| 131 |
+
| 0.1267 | 75.0 | 20475 | 0.1232 | 0.8161 | 0.7001 | 0.3033 | 0.0000 |
|
| 132 |
+
| 0.1269 | 76.0 | 20748 | 0.1227 | 0.8171 | 0.7020 | 0.3058 | 0.0000 |
|
| 133 |
+
| 0.1261 | 77.0 | 21021 | 0.1228 | 0.8191 | 0.7060 | 0.3079 | 0.0000 |
|
| 134 |
+
| 0.1261 | 78.0 | 21294 | 0.1237 | 0.8166 | 0.7072 | 0.3019 | 0.0000 |
|
| 135 |
+
| 0.1268 | 79.0 | 21567 | 0.1233 | 0.8156 | 0.6992 | 0.3047 | 0.0000 |
|
| 136 |
+
| 0.1268 | 80.0 | 21840 | 0.1233 | 0.8172 | 0.6994 | 0.3026 | 0.0000 |
|
| 137 |
+
| 0.1271 | 81.0 | 22113 | 0.1224 | 0.8176 | 0.7037 | 0.3054 | 0.0000 |
|
| 138 |
+
| 0.1271 | 82.0 | 22386 | 0.1227 | 0.8151 | 0.6972 | 0.3002 | 0.0000 |
|
| 139 |
+
| 0.1263 | 83.0 | 22659 | 0.1232 | 0.8146 | 0.6939 | 0.2995 | 0.0000 |
|
| 140 |
+
| 0.1263 | 84.0 | 22932 | 0.1226 | 0.8177 | 0.7017 | 0.3026 | 0.0000 |
|
| 141 |
+
| 0.1265 | 85.0 | 23205 | 0.1230 | 0.8181 | 0.7059 | 0.3072 | 0.0000 |
|
| 142 |
+
| 0.1265 | 86.0 | 23478 | 0.1234 | 0.8168 | 0.7059 | 0.3033 | 0.0000 |
|
| 143 |
+
| 0.1265 | 87.0 | 23751 | 0.1236 | 0.8143 | 0.6944 | 0.3005 | 0.0000 |
|
| 144 |
+
| 0.1266 | 88.0 | 24024 | 0.1226 | 0.8186 | 0.7029 | 0.3079 | 0.0000 |
|
| 145 |
+
| 0.1266 | 89.0 | 24297 | 0.1230 | 0.8179 | 0.7084 | 0.3075 | 0.0000 |
|
| 146 |
+
| 0.1263 | 90.0 | 24570 | 0.1232 | 0.8195 | 0.7105 | 0.3037 | 0.0000 |
|
| 147 |
+
| 0.1263 | 91.0 | 24843 | 0.1227 | 0.8155 | 0.6972 | 0.3019 | 0.0000 |
|
| 148 |
+
|
| 149 |
+
|
| 150 |
+
### Framework versions
|
| 151 |
+
|
| 152 |
+
- Transformers 4.48.0
|
| 153 |
+
- Pytorch 2.5.1+cu124
|
| 154 |
+
- Datasets 3.0.2
|
| 155 |
+
- Tokenizers 0.21.0
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 91.0,
|
| 3 |
+
"eval_accuracy": 0.307110958433528,
|
| 4 |
+
"eval_f1_macro": 0.706034720214054,
|
| 5 |
+
"eval_f1_micro": 0.813577721180744,
|
| 6 |
+
"eval_loss": 0.12363793700933456,
|
| 7 |
+
"eval_runtime": 464.1967,
|
| 8 |
+
"eval_samples_per_second": 6.271,
|
| 9 |
+
"eval_steps_per_second": 0.196,
|
| 10 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 11 |
+
"total_flos": 1.1760993126572918e+21,
|
| 12 |
+
"train_loss": 0.1412498749900842,
|
| 13 |
+
"train_runtime": 170277.6523,
|
| 14 |
+
"train_samples_per_second": 7.678,
|
| 15 |
+
"train_steps_per_second": 0.24
|
| 16 |
+
}
|
logs/events.out.tfevents.1737978278.datavisu3
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8bfb74fce5158dc605098eefcf55bcd890859f30d386772e5fc71eb523f2c15b
|
| 3 |
+
size 62640
|
logs/events.out.tfevents.1738149112.datavisu3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a533d79fc3d29a5863c7c286e3ac81240d74ec18f7d85b6c4f90f27158b7cc51
|
| 3 |
+
size 40
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1228107508
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:587a8d91110d6aa0aa58ec8d29b31bee72a4a2ce94fbf0f0b869f86d9c2321ea
|
| 3 |
size 1228107508
|
test_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 91.0,
|
| 3 |
+
"eval_accuracy": 0.307110958433528,
|
| 4 |
+
"eval_f1_macro": 0.706034720214054,
|
| 5 |
+
"eval_f1_micro": 0.813577721180744,
|
| 6 |
+
"eval_loss": 0.12363793700933456,
|
| 7 |
+
"eval_runtime": 464.1967,
|
| 8 |
+
"eval_samples_per_second": 6.271,
|
| 9 |
+
"eval_steps_per_second": 0.196,
|
| 10 |
+
"learning_rate": 1.0000000000000002e-07
|
| 11 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 91.0,
|
| 3 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 4 |
+
"total_flos": 1.1760993126572918e+21,
|
| 5 |
+
"train_loss": 0.1412498749900842,
|
| 6 |
+
"train_runtime": 170277.6523,
|
| 7 |
+
"train_samples_per_second": 7.678,
|
| 8 |
+
"train_steps_per_second": 0.24
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1487 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.12240613251924515,
|
| 3 |
+
"best_model_checkpoint": "/home1/datahome/villien/project_hub/DinoVdeau/models/DinoVdo-large-2025_01_27_45863-bs32_freeze/checkpoint-22113",
|
| 4 |
+
"epoch": 91.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 24843,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.2293193717277487,
|
| 14 |
+
"eval_f1_macro": 0.5138028081483725,
|
| 15 |
+
"eval_f1_micro": 0.7433001107419712,
|
| 16 |
+
"eval_loss": 0.16775080561637878,
|
| 17 |
+
"eval_runtime": 464.7451,
|
| 18 |
+
"eval_samples_per_second": 6.165,
|
| 19 |
+
"eval_steps_per_second": 0.194,
|
| 20 |
+
"learning_rate": 0.001,
|
| 21 |
+
"step": 273
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 1.8315018315018317,
|
| 25 |
+
"grad_norm": 0.32014167308807373,
|
| 26 |
+
"learning_rate": 0.001,
|
| 27 |
+
"loss": 0.2722,
|
| 28 |
+
"step": 500
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"epoch": 2.0,
|
| 32 |
+
"eval_accuracy": 0.243630017452007,
|
| 33 |
+
"eval_f1_macro": 0.5770554152365124,
|
| 34 |
+
"eval_f1_micro": 0.7613556338028169,
|
| 35 |
+
"eval_loss": 0.15366077423095703,
|
| 36 |
+
"eval_runtime": 460.9126,
|
| 37 |
+
"eval_samples_per_second": 6.216,
|
| 38 |
+
"eval_steps_per_second": 0.195,
|
| 39 |
+
"learning_rate": 0.001,
|
| 40 |
+
"step": 546
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 3.0,
|
| 44 |
+
"eval_accuracy": 0.24153577661431064,
|
| 45 |
+
"eval_f1_macro": 0.6194117241441944,
|
| 46 |
+
"eval_f1_micro": 0.7763416930582898,
|
| 47 |
+
"eval_loss": 0.14831368625164032,
|
| 48 |
+
"eval_runtime": 459.029,
|
| 49 |
+
"eval_samples_per_second": 6.241,
|
| 50 |
+
"eval_steps_per_second": 0.196,
|
| 51 |
+
"learning_rate": 0.001,
|
| 52 |
+
"step": 819
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 3.663003663003663,
|
| 56 |
+
"grad_norm": 0.2205018699169159,
|
| 57 |
+
"learning_rate": 0.001,
|
| 58 |
+
"loss": 0.169,
|
| 59 |
+
"step": 1000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 4.0,
|
| 63 |
+
"eval_accuracy": 0.2554973821989529,
|
| 64 |
+
"eval_f1_macro": 0.6275864194841865,
|
| 65 |
+
"eval_f1_micro": 0.7808419243986254,
|
| 66 |
+
"eval_loss": 0.14640773832798004,
|
| 67 |
+
"eval_runtime": 453.7375,
|
| 68 |
+
"eval_samples_per_second": 6.314,
|
| 69 |
+
"eval_steps_per_second": 0.198,
|
| 70 |
+
"learning_rate": 0.001,
|
| 71 |
+
"step": 1092
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 5.0,
|
| 75 |
+
"eval_accuracy": 0.2520069808027923,
|
| 76 |
+
"eval_f1_macro": 0.6420978846376312,
|
| 77 |
+
"eval_f1_micro": 0.7787831782144088,
|
| 78 |
+
"eval_loss": 0.14515382051467896,
|
| 79 |
+
"eval_runtime": 459.4943,
|
| 80 |
+
"eval_samples_per_second": 6.235,
|
| 81 |
+
"eval_steps_per_second": 0.196,
|
| 82 |
+
"learning_rate": 0.001,
|
| 83 |
+
"step": 1365
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 5.4945054945054945,
|
| 87 |
+
"grad_norm": 0.19979850947856903,
|
| 88 |
+
"learning_rate": 0.001,
|
| 89 |
+
"loss": 0.1612,
|
| 90 |
+
"step": 1500
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 6.0,
|
| 94 |
+
"eval_accuracy": 0.2579406631762653,
|
| 95 |
+
"eval_f1_macro": 0.614677201484495,
|
| 96 |
+
"eval_f1_micro": 0.7802330599913682,
|
| 97 |
+
"eval_loss": 0.14404040575027466,
|
| 98 |
+
"eval_runtime": 455.8188,
|
| 99 |
+
"eval_samples_per_second": 6.285,
|
| 100 |
+
"eval_steps_per_second": 0.197,
|
| 101 |
+
"learning_rate": 0.001,
|
| 102 |
+
"step": 1638
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 7.0,
|
| 106 |
+
"eval_accuracy": 0.25026178010471206,
|
| 107 |
+
"eval_f1_macro": 0.6141169392172392,
|
| 108 |
+
"eval_f1_micro": 0.7787126995142263,
|
| 109 |
+
"eval_loss": 0.14518576860427856,
|
| 110 |
+
"eval_runtime": 463.2529,
|
| 111 |
+
"eval_samples_per_second": 6.185,
|
| 112 |
+
"eval_steps_per_second": 0.194,
|
| 113 |
+
"learning_rate": 0.001,
|
| 114 |
+
"step": 1911
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 7.326007326007326,
|
| 118 |
+
"grad_norm": 0.21145381033420563,
|
| 119 |
+
"learning_rate": 0.001,
|
| 120 |
+
"loss": 0.1594,
|
| 121 |
+
"step": 2000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 8.0,
|
| 125 |
+
"eval_accuracy": 0.2534031413612565,
|
| 126 |
+
"eval_f1_macro": 0.6193231969902891,
|
| 127 |
+
"eval_f1_micro": 0.77764832793959,
|
| 128 |
+
"eval_loss": 0.14463570713996887,
|
| 129 |
+
"eval_runtime": 461.4649,
|
| 130 |
+
"eval_samples_per_second": 6.208,
|
| 131 |
+
"eval_steps_per_second": 0.195,
|
| 132 |
+
"learning_rate": 0.001,
|
| 133 |
+
"step": 2184
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 9.0,
|
| 137 |
+
"eval_accuracy": 0.24537521815008725,
|
| 138 |
+
"eval_f1_macro": 0.6363157565562336,
|
| 139 |
+
"eval_f1_micro": 0.7813486742506164,
|
| 140 |
+
"eval_loss": 0.14786836504936218,
|
| 141 |
+
"eval_runtime": 457.2698,
|
| 142 |
+
"eval_samples_per_second": 6.265,
|
| 143 |
+
"eval_steps_per_second": 0.197,
|
| 144 |
+
"learning_rate": 0.001,
|
| 145 |
+
"step": 2457
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 9.157509157509157,
|
| 149 |
+
"grad_norm": 0.1889927238225937,
|
| 150 |
+
"learning_rate": 0.001,
|
| 151 |
+
"loss": 0.1567,
|
| 152 |
+
"step": 2500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 10.0,
|
| 156 |
+
"eval_accuracy": 0.2607329842931937,
|
| 157 |
+
"eval_f1_macro": 0.6366285266954459,
|
| 158 |
+
"eval_f1_micro": 0.7865724075118186,
|
| 159 |
+
"eval_loss": 0.14251072704792023,
|
| 160 |
+
"eval_runtime": 460.3923,
|
| 161 |
+
"eval_samples_per_second": 6.223,
|
| 162 |
+
"eval_steps_per_second": 0.195,
|
| 163 |
+
"learning_rate": 0.001,
|
| 164 |
+
"step": 2730
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 10.989010989010989,
|
| 168 |
+
"grad_norm": 0.1742856353521347,
|
| 169 |
+
"learning_rate": 0.001,
|
| 170 |
+
"loss": 0.1559,
|
| 171 |
+
"step": 3000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 11.0,
|
| 175 |
+
"eval_accuracy": 0.2617801047120419,
|
| 176 |
+
"eval_f1_macro": 0.6565839087323015,
|
| 177 |
+
"eval_f1_micro": 0.7907930529399456,
|
| 178 |
+
"eval_loss": 0.14535894989967346,
|
| 179 |
+
"eval_runtime": 453.3601,
|
| 180 |
+
"eval_samples_per_second": 6.319,
|
| 181 |
+
"eval_steps_per_second": 0.199,
|
| 182 |
+
"learning_rate": 0.001,
|
| 183 |
+
"step": 3003
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 12.0,
|
| 187 |
+
"eval_accuracy": 0.2600349040139616,
|
| 188 |
+
"eval_f1_macro": 0.6438532757657272,
|
| 189 |
+
"eval_f1_micro": 0.7895364689843217,
|
| 190 |
+
"eval_loss": 0.14042578637599945,
|
| 191 |
+
"eval_runtime": 461.5251,
|
| 192 |
+
"eval_samples_per_second": 6.208,
|
| 193 |
+
"eval_steps_per_second": 0.195,
|
| 194 |
+
"learning_rate": 0.001,
|
| 195 |
+
"step": 3276
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 12.820512820512821,
|
| 199 |
+
"grad_norm": 0.16230766475200653,
|
| 200 |
+
"learning_rate": 0.001,
|
| 201 |
+
"loss": 0.1549,
|
| 202 |
+
"step": 3500
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 13.0,
|
| 206 |
+
"eval_accuracy": 0.2530541012216405,
|
| 207 |
+
"eval_f1_macro": 0.6490185335679685,
|
| 208 |
+
"eval_f1_micro": 0.7883035906256719,
|
| 209 |
+
"eval_loss": 0.1413952261209488,
|
| 210 |
+
"eval_runtime": 459.1407,
|
| 211 |
+
"eval_samples_per_second": 6.24,
|
| 212 |
+
"eval_steps_per_second": 0.196,
|
| 213 |
+
"learning_rate": 0.001,
|
| 214 |
+
"step": 3549
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 14.0,
|
| 218 |
+
"eval_accuracy": 0.2649214659685864,
|
| 219 |
+
"eval_f1_macro": 0.6347742331888712,
|
| 220 |
+
"eval_f1_micro": 0.7893909125685383,
|
| 221 |
+
"eval_loss": 0.14056158065795898,
|
| 222 |
+
"eval_runtime": 459.115,
|
| 223 |
+
"eval_samples_per_second": 6.24,
|
| 224 |
+
"eval_steps_per_second": 0.196,
|
| 225 |
+
"learning_rate": 0.001,
|
| 226 |
+
"step": 3822
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 14.652014652014651,
|
| 230 |
+
"grad_norm": 0.5134446024894714,
|
| 231 |
+
"learning_rate": 0.001,
|
| 232 |
+
"loss": 0.155,
|
| 233 |
+
"step": 4000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 15.0,
|
| 237 |
+
"eval_accuracy": 0.2631762652705061,
|
| 238 |
+
"eval_f1_macro": 0.6469635503488436,
|
| 239 |
+
"eval_f1_micro": 0.7906051299796704,
|
| 240 |
+
"eval_loss": 0.13873133063316345,
|
| 241 |
+
"eval_runtime": 461.6498,
|
| 242 |
+
"eval_samples_per_second": 6.206,
|
| 243 |
+
"eval_steps_per_second": 0.195,
|
| 244 |
+
"learning_rate": 0.001,
|
| 245 |
+
"step": 4095
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 16.0,
|
| 249 |
+
"eval_accuracy": 0.2603839441535777,
|
| 250 |
+
"eval_f1_macro": 0.6331356980695925,
|
| 251 |
+
"eval_f1_micro": 0.7857704059362723,
|
| 252 |
+
"eval_loss": 0.14008578658103943,
|
| 253 |
+
"eval_runtime": 459.7885,
|
| 254 |
+
"eval_samples_per_second": 6.231,
|
| 255 |
+
"eval_steps_per_second": 0.196,
|
| 256 |
+
"learning_rate": 0.001,
|
| 257 |
+
"step": 4368
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 16.483516483516482,
|
| 261 |
+
"grad_norm": 0.159025177359581,
|
| 262 |
+
"learning_rate": 0.001,
|
| 263 |
+
"loss": 0.1531,
|
| 264 |
+
"step": 4500
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 17.0,
|
| 268 |
+
"eval_accuracy": 0.2520069808027923,
|
| 269 |
+
"eval_f1_macro": 0.665107882553296,
|
| 270 |
+
"eval_f1_micro": 0.7954926273458445,
|
| 271 |
+
"eval_loss": 0.13811993598937988,
|
| 272 |
+
"eval_runtime": 460.3054,
|
| 273 |
+
"eval_samples_per_second": 6.224,
|
| 274 |
+
"eval_steps_per_second": 0.196,
|
| 275 |
+
"learning_rate": 0.001,
|
| 276 |
+
"step": 4641
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 18.0,
|
| 280 |
+
"eval_accuracy": 0.27015706806282724,
|
| 281 |
+
"eval_f1_macro": 0.6498362138588303,
|
| 282 |
+
"eval_f1_micro": 0.791355160264673,
|
| 283 |
+
"eval_loss": 0.13870170712471008,
|
| 284 |
+
"eval_runtime": 463.8864,
|
| 285 |
+
"eval_samples_per_second": 6.176,
|
| 286 |
+
"eval_steps_per_second": 0.194,
|
| 287 |
+
"learning_rate": 0.001,
|
| 288 |
+
"step": 4914
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 18.315018315018314,
|
| 292 |
+
"grad_norm": 0.14721287786960602,
|
| 293 |
+
"learning_rate": 0.001,
|
| 294 |
+
"loss": 0.1549,
|
| 295 |
+
"step": 5000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 19.0,
|
| 299 |
+
"eval_accuracy": 0.2631762652705061,
|
| 300 |
+
"eval_f1_macro": 0.6356042954083428,
|
| 301 |
+
"eval_f1_micro": 0.794037710208935,
|
| 302 |
+
"eval_loss": 0.1373777985572815,
|
| 303 |
+
"eval_runtime": 458.2597,
|
| 304 |
+
"eval_samples_per_second": 6.252,
|
| 305 |
+
"eval_steps_per_second": 0.196,
|
| 306 |
+
"learning_rate": 0.001,
|
| 307 |
+
"step": 5187
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 20.0,
|
| 311 |
+
"eval_accuracy": 0.25514834205933684,
|
| 312 |
+
"eval_f1_macro": 0.6393109338369551,
|
| 313 |
+
"eval_f1_micro": 0.7850155722244155,
|
| 314 |
+
"eval_loss": 0.13864819705486298,
|
| 315 |
+
"eval_runtime": 469.7679,
|
| 316 |
+
"eval_samples_per_second": 6.099,
|
| 317 |
+
"eval_steps_per_second": 0.192,
|
| 318 |
+
"learning_rate": 0.001,
|
| 319 |
+
"step": 5460
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 20.146520146520146,
|
| 323 |
+
"grad_norm": 0.29727715253829956,
|
| 324 |
+
"learning_rate": 0.001,
|
| 325 |
+
"loss": 0.1524,
|
| 326 |
+
"step": 5500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 21.0,
|
| 330 |
+
"eval_accuracy": 0.26457242582897034,
|
| 331 |
+
"eval_f1_macro": 0.6496408648134395,
|
| 332 |
+
"eval_f1_micro": 0.7943311144513635,
|
| 333 |
+
"eval_loss": 0.13566707074642181,
|
| 334 |
+
"eval_runtime": 461.0026,
|
| 335 |
+
"eval_samples_per_second": 6.215,
|
| 336 |
+
"eval_steps_per_second": 0.195,
|
| 337 |
+
"learning_rate": 0.001,
|
| 338 |
+
"step": 5733
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 21.978021978021978,
|
| 342 |
+
"grad_norm": 0.16196180880069733,
|
| 343 |
+
"learning_rate": 0.001,
|
| 344 |
+
"loss": 0.1519,
|
| 345 |
+
"step": 6000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 22.0,
|
| 349 |
+
"eval_accuracy": 0.27225130890052357,
|
| 350 |
+
"eval_f1_macro": 0.6400060646722534,
|
| 351 |
+
"eval_f1_micro": 0.7972140143520473,
|
| 352 |
+
"eval_loss": 0.1371580958366394,
|
| 353 |
+
"eval_runtime": 454.7092,
|
| 354 |
+
"eval_samples_per_second": 6.301,
|
| 355 |
+
"eval_steps_per_second": 0.198,
|
| 356 |
+
"learning_rate": 0.001,
|
| 357 |
+
"step": 6006
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 23.0,
|
| 361 |
+
"eval_accuracy": 0.2603839441535777,
|
| 362 |
+
"eval_f1_macro": 0.6595092645083339,
|
| 363 |
+
"eval_f1_micro": 0.7938437661665805,
|
| 364 |
+
"eval_loss": 0.13614478707313538,
|
| 365 |
+
"eval_runtime": 452.3843,
|
| 366 |
+
"eval_samples_per_second": 6.333,
|
| 367 |
+
"eval_steps_per_second": 0.199,
|
| 368 |
+
"learning_rate": 0.001,
|
| 369 |
+
"step": 6279
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 23.80952380952381,
|
| 373 |
+
"grad_norm": 0.19081830978393555,
|
| 374 |
+
"learning_rate": 0.001,
|
| 375 |
+
"loss": 0.1528,
|
| 376 |
+
"step": 6500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 24.0,
|
| 380 |
+
"eval_accuracy": 0.2649214659685864,
|
| 381 |
+
"eval_f1_macro": 0.6417599876035757,
|
| 382 |
+
"eval_f1_micro": 0.7953625165167725,
|
| 383 |
+
"eval_loss": 0.1362675279378891,
|
| 384 |
+
"eval_runtime": 458.9922,
|
| 385 |
+
"eval_samples_per_second": 6.242,
|
| 386 |
+
"eval_steps_per_second": 0.196,
|
| 387 |
+
"learning_rate": 0.001,
|
| 388 |
+
"step": 6552
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 25.0,
|
| 392 |
+
"eval_accuracy": 0.2732984293193717,
|
| 393 |
+
"eval_f1_macro": 0.6499701126111167,
|
| 394 |
+
"eval_f1_micro": 0.7963302752293578,
|
| 395 |
+
"eval_loss": 0.1358671337366104,
|
| 396 |
+
"eval_runtime": 456.4797,
|
| 397 |
+
"eval_samples_per_second": 6.276,
|
| 398 |
+
"eval_steps_per_second": 0.197,
|
| 399 |
+
"learning_rate": 0.001,
|
| 400 |
+
"step": 6825
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 25.641025641025642,
|
| 404 |
+
"grad_norm": 0.1495918333530426,
|
| 405 |
+
"learning_rate": 0.001,
|
| 406 |
+
"loss": 0.1524,
|
| 407 |
+
"step": 7000
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 26.0,
|
| 411 |
+
"eval_accuracy": 0.26910994764397905,
|
| 412 |
+
"eval_f1_macro": 0.6554940295679257,
|
| 413 |
+
"eval_f1_micro": 0.7983850403739906,
|
| 414 |
+
"eval_loss": 0.13484793901443481,
|
| 415 |
+
"eval_runtime": 468.4399,
|
| 416 |
+
"eval_samples_per_second": 6.116,
|
| 417 |
+
"eval_steps_per_second": 0.192,
|
| 418 |
+
"learning_rate": 0.001,
|
| 419 |
+
"step": 7098
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 27.0,
|
| 423 |
+
"eval_accuracy": 0.268760907504363,
|
| 424 |
+
"eval_f1_macro": 0.6535438446852289,
|
| 425 |
+
"eval_f1_micro": 0.7944040481353914,
|
| 426 |
+
"eval_loss": 0.13669784367084503,
|
| 427 |
+
"eval_runtime": 466.0343,
|
| 428 |
+
"eval_samples_per_second": 6.148,
|
| 429 |
+
"eval_steps_per_second": 0.193,
|
| 430 |
+
"learning_rate": 0.001,
|
| 431 |
+
"step": 7371
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 27.47252747252747,
|
| 435 |
+
"grad_norm": 0.16602398455142975,
|
| 436 |
+
"learning_rate": 0.001,
|
| 437 |
+
"loss": 0.1521,
|
| 438 |
+
"step": 7500
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 28.0,
|
| 442 |
+
"eval_accuracy": 0.26771378708551485,
|
| 443 |
+
"eval_f1_macro": 0.6398313564596081,
|
| 444 |
+
"eval_f1_micro": 0.7922701137002292,
|
| 445 |
+
"eval_loss": 0.13569706678390503,
|
| 446 |
+
"eval_runtime": 460.9998,
|
| 447 |
+
"eval_samples_per_second": 6.215,
|
| 448 |
+
"eval_steps_per_second": 0.195,
|
| 449 |
+
"learning_rate": 0.001,
|
| 450 |
+
"step": 7644
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 29.0,
|
| 454 |
+
"eval_accuracy": 0.2638743455497382,
|
| 455 |
+
"eval_f1_macro": 0.6636571549498853,
|
| 456 |
+
"eval_f1_micro": 0.7924252834077636,
|
| 457 |
+
"eval_loss": 0.14052562415599823,
|
| 458 |
+
"eval_runtime": 467.3084,
|
| 459 |
+
"eval_samples_per_second": 6.131,
|
| 460 |
+
"eval_steps_per_second": 0.193,
|
| 461 |
+
"learning_rate": 0.001,
|
| 462 |
+
"step": 7917
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 29.304029304029303,
|
| 466 |
+
"grad_norm": 0.1555439680814743,
|
| 467 |
+
"learning_rate": 0.001,
|
| 468 |
+
"loss": 0.1525,
|
| 469 |
+
"step": 8000
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 30.0,
|
| 473 |
+
"eval_accuracy": 0.27225130890052357,
|
| 474 |
+
"eval_f1_macro": 0.6405329741259155,
|
| 475 |
+
"eval_f1_micro": 0.7875164835164835,
|
| 476 |
+
"eval_loss": 0.13725879788398743,
|
| 477 |
+
"eval_runtime": 459.9825,
|
| 478 |
+
"eval_samples_per_second": 6.228,
|
| 479 |
+
"eval_steps_per_second": 0.196,
|
| 480 |
+
"learning_rate": 0.001,
|
| 481 |
+
"step": 8190
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 31.0,
|
| 485 |
+
"eval_accuracy": 0.2719022687609075,
|
| 486 |
+
"eval_f1_macro": 0.6562409583519914,
|
| 487 |
+
"eval_f1_micro": 0.7986043969902056,
|
| 488 |
+
"eval_loss": 0.13544805347919464,
|
| 489 |
+
"eval_runtime": 461.8222,
|
| 490 |
+
"eval_samples_per_second": 6.204,
|
| 491 |
+
"eval_steps_per_second": 0.195,
|
| 492 |
+
"learning_rate": 0.001,
|
| 493 |
+
"step": 8463
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 31.135531135531135,
|
| 497 |
+
"grad_norm": 0.20161172747612,
|
| 498 |
+
"learning_rate": 0.001,
|
| 499 |
+
"loss": 0.1522,
|
| 500 |
+
"step": 8500
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 32.0,
|
| 504 |
+
"eval_accuracy": 0.2649214659685864,
|
| 505 |
+
"eval_f1_macro": 0.6463119642593244,
|
| 506 |
+
"eval_f1_micro": 0.793030759303076,
|
| 507 |
+
"eval_loss": 0.13693773746490479,
|
| 508 |
+
"eval_runtime": 462.8659,
|
| 509 |
+
"eval_samples_per_second": 6.19,
|
| 510 |
+
"eval_steps_per_second": 0.194,
|
| 511 |
+
"learning_rate": 0.001,
|
| 512 |
+
"step": 8736
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 32.967032967032964,
|
| 516 |
+
"grad_norm": 0.5906699299812317,
|
| 517 |
+
"learning_rate": 0.0001,
|
| 518 |
+
"loss": 0.1486,
|
| 519 |
+
"step": 9000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 33.0,
|
| 523 |
+
"eval_accuracy": 0.27469458987783596,
|
| 524 |
+
"eval_f1_macro": 0.6660418237263234,
|
| 525 |
+
"eval_f1_micro": 0.8014840718154292,
|
| 526 |
+
"eval_loss": 0.13195939362049103,
|
| 527 |
+
"eval_runtime": 465.1048,
|
| 528 |
+
"eval_samples_per_second": 6.16,
|
| 529 |
+
"eval_steps_per_second": 0.194,
|
| 530 |
+
"learning_rate": 0.0001,
|
| 531 |
+
"step": 9009
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 34.0,
|
| 535 |
+
"eval_accuracy": 0.28342059336823733,
|
| 536 |
+
"eval_f1_macro": 0.6755989146580753,
|
| 537 |
+
"eval_f1_micro": 0.8042847388187094,
|
| 538 |
+
"eval_loss": 0.1299898624420166,
|
| 539 |
+
"eval_runtime": 462.0703,
|
| 540 |
+
"eval_samples_per_second": 6.2,
|
| 541 |
+
"eval_steps_per_second": 0.195,
|
| 542 |
+
"learning_rate": 0.0001,
|
| 543 |
+
"step": 9282
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 34.798534798534796,
|
| 547 |
+
"grad_norm": 0.1343473345041275,
|
| 548 |
+
"learning_rate": 0.0001,
|
| 549 |
+
"loss": 0.1419,
|
| 550 |
+
"step": 9500
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 35.0,
|
| 554 |
+
"eval_accuracy": 0.28097731239092494,
|
| 555 |
+
"eval_f1_macro": 0.6772201041047545,
|
| 556 |
+
"eval_f1_micro": 0.8066335550130482,
|
| 557 |
+
"eval_loss": 0.12946264445781708,
|
| 558 |
+
"eval_runtime": 463.7568,
|
| 559 |
+
"eval_samples_per_second": 6.178,
|
| 560 |
+
"eval_steps_per_second": 0.194,
|
| 561 |
+
"learning_rate": 0.0001,
|
| 562 |
+
"step": 9555
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 36.0,
|
| 566 |
+
"eval_accuracy": 0.28167539267015707,
|
| 567 |
+
"eval_f1_macro": 0.6794992142877868,
|
| 568 |
+
"eval_f1_micro": 0.8046182685753238,
|
| 569 |
+
"eval_loss": 0.13086125254631042,
|
| 570 |
+
"eval_runtime": 457.8267,
|
| 571 |
+
"eval_samples_per_second": 6.258,
|
| 572 |
+
"eval_steps_per_second": 0.197,
|
| 573 |
+
"learning_rate": 0.0001,
|
| 574 |
+
"step": 9828
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 36.63003663003663,
|
| 578 |
+
"grad_norm": 0.16144825518131256,
|
| 579 |
+
"learning_rate": 0.0001,
|
| 580 |
+
"loss": 0.1396,
|
| 581 |
+
"step": 10000
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 37.0,
|
| 585 |
+
"eval_accuracy": 0.2823734729493892,
|
| 586 |
+
"eval_f1_macro": 0.6791512030316924,
|
| 587 |
+
"eval_f1_micro": 0.805367401183229,
|
| 588 |
+
"eval_loss": 0.1278763711452484,
|
| 589 |
+
"eval_runtime": 471.6542,
|
| 590 |
+
"eval_samples_per_second": 6.074,
|
| 591 |
+
"eval_steps_per_second": 0.191,
|
| 592 |
+
"learning_rate": 0.0001,
|
| 593 |
+
"step": 10101
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 38.0,
|
| 597 |
+
"eval_accuracy": 0.2848167539267016,
|
| 598 |
+
"eval_f1_macro": 0.6814340581384846,
|
| 599 |
+
"eval_f1_micro": 0.8077836187892149,
|
| 600 |
+
"eval_loss": 0.12904110550880432,
|
| 601 |
+
"eval_runtime": 470.311,
|
| 602 |
+
"eval_samples_per_second": 6.092,
|
| 603 |
+
"eval_steps_per_second": 0.191,
|
| 604 |
+
"learning_rate": 0.0001,
|
| 605 |
+
"step": 10374
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 38.46153846153846,
|
| 609 |
+
"grad_norm": 0.16166280210018158,
|
| 610 |
+
"learning_rate": 0.0001,
|
| 611 |
+
"loss": 0.1366,
|
| 612 |
+
"step": 10500
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 39.0,
|
| 616 |
+
"eval_accuracy": 0.293891797556719,
|
| 617 |
+
"eval_f1_macro": 0.6832531367067615,
|
| 618 |
+
"eval_f1_micro": 0.8115661433562018,
|
| 619 |
+
"eval_loss": 0.12716704607009888,
|
| 620 |
+
"eval_runtime": 462.4995,
|
| 621 |
+
"eval_samples_per_second": 6.195,
|
| 622 |
+
"eval_steps_per_second": 0.195,
|
| 623 |
+
"learning_rate": 0.0001,
|
| 624 |
+
"step": 10647
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 40.0,
|
| 628 |
+
"eval_accuracy": 0.2907504363001745,
|
| 629 |
+
"eval_f1_macro": 0.6878954476405744,
|
| 630 |
+
"eval_f1_micro": 0.8115966316142277,
|
| 631 |
+
"eval_loss": 0.1293308585882187,
|
| 632 |
+
"eval_runtime": 477.5548,
|
| 633 |
+
"eval_samples_per_second": 5.999,
|
| 634 |
+
"eval_steps_per_second": 0.188,
|
| 635 |
+
"learning_rate": 0.0001,
|
| 636 |
+
"step": 10920
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 40.29304029304029,
|
| 640 |
+
"grad_norm": 0.23697003722190857,
|
| 641 |
+
"learning_rate": 0.0001,
|
| 642 |
+
"loss": 0.1361,
|
| 643 |
+
"step": 11000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 41.0,
|
| 647 |
+
"eval_accuracy": 0.2917975567190227,
|
| 648 |
+
"eval_f1_macro": 0.6863712244346338,
|
| 649 |
+
"eval_f1_micro": 0.8088835942818243,
|
| 650 |
+
"eval_loss": 0.12695887684822083,
|
| 651 |
+
"eval_runtime": 460.0604,
|
| 652 |
+
"eval_samples_per_second": 6.227,
|
| 653 |
+
"eval_steps_per_second": 0.196,
|
| 654 |
+
"learning_rate": 0.0001,
|
| 655 |
+
"step": 11193
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 42.0,
|
| 659 |
+
"eval_accuracy": 0.2914485165794066,
|
| 660 |
+
"eval_f1_macro": 0.6837014102587237,
|
| 661 |
+
"eval_f1_micro": 0.8109225964379168,
|
| 662 |
+
"eval_loss": 0.12624548375606537,
|
| 663 |
+
"eval_runtime": 468.1922,
|
| 664 |
+
"eval_samples_per_second": 6.119,
|
| 665 |
+
"eval_steps_per_second": 0.192,
|
| 666 |
+
"learning_rate": 0.0001,
|
| 667 |
+
"step": 11466
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 42.124542124542124,
|
| 671 |
+
"grad_norm": 0.23925267159938812,
|
| 672 |
+
"learning_rate": 0.0001,
|
| 673 |
+
"loss": 0.135,
|
| 674 |
+
"step": 11500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 43.0,
|
| 678 |
+
"eval_accuracy": 0.2949389179755672,
|
| 679 |
+
"eval_f1_macro": 0.6983715804658985,
|
| 680 |
+
"eval_f1_micro": 0.8122734552811262,
|
| 681 |
+
"eval_loss": 0.1261172592639923,
|
| 682 |
+
"eval_runtime": 464.3762,
|
| 683 |
+
"eval_samples_per_second": 6.17,
|
| 684 |
+
"eval_steps_per_second": 0.194,
|
| 685 |
+
"learning_rate": 0.0001,
|
| 686 |
+
"step": 11739
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 43.956043956043956,
|
| 690 |
+
"grad_norm": 0.7063220143318176,
|
| 691 |
+
"learning_rate": 0.0001,
|
| 692 |
+
"loss": 0.134,
|
| 693 |
+
"step": 12000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 44.0,
|
| 697 |
+
"eval_accuracy": 0.29354275741710295,
|
| 698 |
+
"eval_f1_macro": 0.6833912042783733,
|
| 699 |
+
"eval_f1_micro": 0.8106054209237524,
|
| 700 |
+
"eval_loss": 0.12830273807048798,
|
| 701 |
+
"eval_runtime": 456.3209,
|
| 702 |
+
"eval_samples_per_second": 6.278,
|
| 703 |
+
"eval_steps_per_second": 0.197,
|
| 704 |
+
"learning_rate": 0.0001,
|
| 705 |
+
"step": 12012
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 45.0,
|
| 709 |
+
"eval_accuracy": 0.2931937172774869,
|
| 710 |
+
"eval_f1_macro": 0.7009703620992983,
|
| 711 |
+
"eval_f1_micro": 0.8113048605846179,
|
| 712 |
+
"eval_loss": 0.12624593079090118,
|
| 713 |
+
"eval_runtime": 455.5361,
|
| 714 |
+
"eval_samples_per_second": 6.289,
|
| 715 |
+
"eval_steps_per_second": 0.198,
|
| 716 |
+
"learning_rate": 0.0001,
|
| 717 |
+
"step": 12285
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 45.78754578754579,
|
| 721 |
+
"grad_norm": 0.22236338257789612,
|
| 722 |
+
"learning_rate": 0.0001,
|
| 723 |
+
"loss": 0.1331,
|
| 724 |
+
"step": 12500
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 46.0,
|
| 728 |
+
"eval_accuracy": 0.29598603839441534,
|
| 729 |
+
"eval_f1_macro": 0.6963625779503319,
|
| 730 |
+
"eval_f1_micro": 0.8147029204431017,
|
| 731 |
+
"eval_loss": 0.12462077289819717,
|
| 732 |
+
"eval_runtime": 465.158,
|
| 733 |
+
"eval_samples_per_second": 6.159,
|
| 734 |
+
"eval_steps_per_second": 0.193,
|
| 735 |
+
"learning_rate": 0.0001,
|
| 736 |
+
"step": 12558
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 47.0,
|
| 740 |
+
"eval_accuracy": 0.2987783595113438,
|
| 741 |
+
"eval_f1_macro": 0.6923480831102572,
|
| 742 |
+
"eval_f1_micro": 0.8126030873334743,
|
| 743 |
+
"eval_loss": 0.12529432773590088,
|
| 744 |
+
"eval_runtime": 457.756,
|
| 745 |
+
"eval_samples_per_second": 6.259,
|
| 746 |
+
"eval_steps_per_second": 0.197,
|
| 747 |
+
"learning_rate": 0.0001,
|
| 748 |
+
"step": 12831
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 47.61904761904762,
|
| 752 |
+
"grad_norm": 0.20239435136318207,
|
| 753 |
+
"learning_rate": 0.0001,
|
| 754 |
+
"loss": 0.1325,
|
| 755 |
+
"step": 13000
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 48.0,
|
| 759 |
+
"eval_accuracy": 0.29773123909249566,
|
| 760 |
+
"eval_f1_macro": 0.6954224588663396,
|
| 761 |
+
"eval_f1_micro": 0.8133141600406023,
|
| 762 |
+
"eval_loss": 0.12631145119667053,
|
| 763 |
+
"eval_runtime": 454.7373,
|
| 764 |
+
"eval_samples_per_second": 6.3,
|
| 765 |
+
"eval_steps_per_second": 0.198,
|
| 766 |
+
"learning_rate": 0.0001,
|
| 767 |
+
"step": 13104
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 49.0,
|
| 771 |
+
"eval_accuracy": 0.3036649214659686,
|
| 772 |
+
"eval_f1_macro": 0.6951680615776381,
|
| 773 |
+
"eval_f1_micro": 0.8157960928984657,
|
| 774 |
+
"eval_loss": 0.1252526491880417,
|
| 775 |
+
"eval_runtime": 455.0418,
|
| 776 |
+
"eval_samples_per_second": 6.296,
|
| 777 |
+
"eval_steps_per_second": 0.198,
|
| 778 |
+
"learning_rate": 0.0001,
|
| 779 |
+
"step": 13377
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 49.45054945054945,
|
| 783 |
+
"grad_norm": 0.2540508806705475,
|
| 784 |
+
"learning_rate": 0.0001,
|
| 785 |
+
"loss": 0.1312,
|
| 786 |
+
"step": 13500
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 50.0,
|
| 790 |
+
"eval_accuracy": 0.30052356020942406,
|
| 791 |
+
"eval_f1_macro": 0.7008159807881353,
|
| 792 |
+
"eval_f1_micro": 0.8136319704164391,
|
| 793 |
+
"eval_loss": 0.12632089853286743,
|
| 794 |
+
"eval_runtime": 450.3202,
|
| 795 |
+
"eval_samples_per_second": 6.362,
|
| 796 |
+
"eval_steps_per_second": 0.2,
|
| 797 |
+
"learning_rate": 0.0001,
|
| 798 |
+
"step": 13650
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 51.0,
|
| 802 |
+
"eval_accuracy": 0.3008726003490401,
|
| 803 |
+
"eval_f1_macro": 0.7019225754370848,
|
| 804 |
+
"eval_f1_micro": 0.8157828600150792,
|
| 805 |
+
"eval_loss": 0.1246422603726387,
|
| 806 |
+
"eval_runtime": 455.9576,
|
| 807 |
+
"eval_samples_per_second": 6.283,
|
| 808 |
+
"eval_steps_per_second": 0.197,
|
| 809 |
+
"learning_rate": 0.0001,
|
| 810 |
+
"step": 13923
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 51.282051282051285,
|
| 814 |
+
"grad_norm": 0.2041085660457611,
|
| 815 |
+
"learning_rate": 0.0001,
|
| 816 |
+
"loss": 0.1301,
|
| 817 |
+
"step": 14000
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 52.0,
|
| 821 |
+
"eval_accuracy": 0.29109947643979056,
|
| 822 |
+
"eval_f1_macro": 0.6948529573216252,
|
| 823 |
+
"eval_f1_micro": 0.8091726695366779,
|
| 824 |
+
"eval_loss": 0.12534211575984955,
|
| 825 |
+
"eval_runtime": 453.0007,
|
| 826 |
+
"eval_samples_per_second": 6.324,
|
| 827 |
+
"eval_steps_per_second": 0.199,
|
| 828 |
+
"learning_rate": 0.0001,
|
| 829 |
+
"step": 14196
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 53.0,
|
| 833 |
+
"eval_accuracy": 0.3022687609075044,
|
| 834 |
+
"eval_f1_macro": 0.7018961268425457,
|
| 835 |
+
"eval_f1_micro": 0.8154071593050032,
|
| 836 |
+
"eval_loss": 0.12436465919017792,
|
| 837 |
+
"eval_runtime": 452.0514,
|
| 838 |
+
"eval_samples_per_second": 6.338,
|
| 839 |
+
"eval_steps_per_second": 0.199,
|
| 840 |
+
"learning_rate": 1e-05,
|
| 841 |
+
"step": 14469
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 53.11355311355312,
|
| 845 |
+
"grad_norm": 0.19839347898960114,
|
| 846 |
+
"learning_rate": 1e-05,
|
| 847 |
+
"loss": 0.1306,
|
| 848 |
+
"step": 14500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 54.0,
|
| 852 |
+
"eval_accuracy": 0.3008726003490401,
|
| 853 |
+
"eval_f1_macro": 0.7040300990277949,
|
| 854 |
+
"eval_f1_micro": 0.8154115879111705,
|
| 855 |
+
"eval_loss": 0.12488020956516266,
|
| 856 |
+
"eval_runtime": 453.4122,
|
| 857 |
+
"eval_samples_per_second": 6.319,
|
| 858 |
+
"eval_steps_per_second": 0.198,
|
| 859 |
+
"learning_rate": 1e-05,
|
| 860 |
+
"step": 14742
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 54.94505494505494,
|
| 864 |
+
"grad_norm": 0.2790488600730896,
|
| 865 |
+
"learning_rate": 1e-05,
|
| 866 |
+
"loss": 0.1282,
|
| 867 |
+
"step": 15000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 55.0,
|
| 871 |
+
"eval_accuracy": 0.30052356020942406,
|
| 872 |
+
"eval_f1_macro": 0.6998211272872638,
|
| 873 |
+
"eval_f1_micro": 0.814350990015231,
|
| 874 |
+
"eval_loss": 0.12366042286157608,
|
| 875 |
+
"eval_runtime": 451.3538,
|
| 876 |
+
"eval_samples_per_second": 6.348,
|
| 877 |
+
"eval_steps_per_second": 0.199,
|
| 878 |
+
"learning_rate": 1e-05,
|
| 879 |
+
"step": 15015
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 56.0,
|
| 883 |
+
"eval_accuracy": 0.3033158813263525,
|
| 884 |
+
"eval_f1_macro": 0.7003615661638819,
|
| 885 |
+
"eval_f1_micro": 0.8167849686847599,
|
| 886 |
+
"eval_loss": 0.12352865934371948,
|
| 887 |
+
"eval_runtime": 453.6557,
|
| 888 |
+
"eval_samples_per_second": 6.315,
|
| 889 |
+
"eval_steps_per_second": 0.198,
|
| 890 |
+
"learning_rate": 1e-05,
|
| 891 |
+
"step": 15288
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 56.776556776556774,
|
| 895 |
+
"grad_norm": 0.29984405636787415,
|
| 896 |
+
"learning_rate": 1e-05,
|
| 897 |
+
"loss": 0.1281,
|
| 898 |
+
"step": 15500
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 57.0,
|
| 902 |
+
"eval_accuracy": 0.3029668411867365,
|
| 903 |
+
"eval_f1_macro": 0.7002357451773985,
|
| 904 |
+
"eval_f1_micro": 0.815746583643278,
|
| 905 |
+
"eval_loss": 0.1239086389541626,
|
| 906 |
+
"eval_runtime": 461.4901,
|
| 907 |
+
"eval_samples_per_second": 6.208,
|
| 908 |
+
"eval_steps_per_second": 0.195,
|
| 909 |
+
"learning_rate": 1e-05,
|
| 910 |
+
"step": 15561
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 58.0,
|
| 914 |
+
"eval_accuracy": 0.30261780104712044,
|
| 915 |
+
"eval_f1_macro": 0.699458636808175,
|
| 916 |
+
"eval_f1_micro": 0.8157211095281991,
|
| 917 |
+
"eval_loss": 0.12343526631593704,
|
| 918 |
+
"eval_runtime": 461.5197,
|
| 919 |
+
"eval_samples_per_second": 6.208,
|
| 920 |
+
"eval_steps_per_second": 0.195,
|
| 921 |
+
"learning_rate": 1e-05,
|
| 922 |
+
"step": 15834
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 58.608058608058606,
|
| 926 |
+
"grad_norm": 0.22550125420093536,
|
| 927 |
+
"learning_rate": 1e-05,
|
| 928 |
+
"loss": 0.129,
|
| 929 |
+
"step": 16000
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 59.0,
|
| 933 |
+
"eval_accuracy": 0.30471204188481676,
|
| 934 |
+
"eval_f1_macro": 0.7012031830256923,
|
| 935 |
+
"eval_f1_micro": 0.8149766777324873,
|
| 936 |
+
"eval_loss": 0.12345146387815475,
|
| 937 |
+
"eval_runtime": 458.3652,
|
| 938 |
+
"eval_samples_per_second": 6.25,
|
| 939 |
+
"eval_steps_per_second": 0.196,
|
| 940 |
+
"learning_rate": 1e-05,
|
| 941 |
+
"step": 16107
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 60.0,
|
| 945 |
+
"eval_accuracy": 0.29808027923211167,
|
| 946 |
+
"eval_f1_macro": 0.6932117075627285,
|
| 947 |
+
"eval_f1_micro": 0.8127865511971473,
|
| 948 |
+
"eval_loss": 0.1239377036690712,
|
| 949 |
+
"eval_runtime": 461.8366,
|
| 950 |
+
"eval_samples_per_second": 6.203,
|
| 951 |
+
"eval_steps_per_second": 0.195,
|
| 952 |
+
"learning_rate": 1e-05,
|
| 953 |
+
"step": 16380
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 60.43956043956044,
|
| 957 |
+
"grad_norm": 0.21306775510311127,
|
| 958 |
+
"learning_rate": 1e-05,
|
| 959 |
+
"loss": 0.1284,
|
| 960 |
+
"step": 16500
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 61.0,
|
| 964 |
+
"eval_accuracy": 0.3008726003490401,
|
| 965 |
+
"eval_f1_macro": 0.707636320445804,
|
| 966 |
+
"eval_f1_micro": 0.8174437097179208,
|
| 967 |
+
"eval_loss": 0.12398885935544968,
|
| 968 |
+
"eval_runtime": 462.2279,
|
| 969 |
+
"eval_samples_per_second": 6.198,
|
| 970 |
+
"eval_steps_per_second": 0.195,
|
| 971 |
+
"learning_rate": 1e-05,
|
| 972 |
+
"step": 16653
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 62.0,
|
| 976 |
+
"eval_accuracy": 0.3019197207678883,
|
| 977 |
+
"eval_f1_macro": 0.7031525654438611,
|
| 978 |
+
"eval_f1_micro": 0.8151862464183381,
|
| 979 |
+
"eval_loss": 0.12334412336349487,
|
| 980 |
+
"eval_runtime": 460.3594,
|
| 981 |
+
"eval_samples_per_second": 6.223,
|
| 982 |
+
"eval_steps_per_second": 0.195,
|
| 983 |
+
"learning_rate": 1e-05,
|
| 984 |
+
"step": 16926
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 62.27106227106227,
|
| 988 |
+
"grad_norm": 0.24677294492721558,
|
| 989 |
+
"learning_rate": 1e-05,
|
| 990 |
+
"loss": 0.127,
|
| 991 |
+
"step": 17000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 63.0,
|
| 995 |
+
"eval_accuracy": 0.3022687609075044,
|
| 996 |
+
"eval_f1_macro": 0.7023118424286922,
|
| 997 |
+
"eval_f1_micro": 0.8157861555919809,
|
| 998 |
+
"eval_loss": 0.12325507402420044,
|
| 999 |
+
"eval_runtime": 461.2281,
|
| 1000 |
+
"eval_samples_per_second": 6.212,
|
| 1001 |
+
"eval_steps_per_second": 0.195,
|
| 1002 |
+
"learning_rate": 1e-05,
|
| 1003 |
+
"step": 17199
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 64.0,
|
| 1007 |
+
"eval_accuracy": 0.30471204188481676,
|
| 1008 |
+
"eval_f1_macro": 0.6998879503661606,
|
| 1009 |
+
"eval_f1_micro": 0.8152265453932963,
|
| 1010 |
+
"eval_loss": 0.12346883863210678,
|
| 1011 |
+
"eval_runtime": 466.8309,
|
| 1012 |
+
"eval_samples_per_second": 6.137,
|
| 1013 |
+
"eval_steps_per_second": 0.193,
|
| 1014 |
+
"learning_rate": 1e-05,
|
| 1015 |
+
"step": 17472
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 64.1025641025641,
|
| 1019 |
+
"grad_norm": 0.1753869205713272,
|
| 1020 |
+
"learning_rate": 1e-05,
|
| 1021 |
+
"loss": 0.1279,
|
| 1022 |
+
"step": 17500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 65.0,
|
| 1026 |
+
"eval_accuracy": 0.29773123909249566,
|
| 1027 |
+
"eval_f1_macro": 0.7001319611176023,
|
| 1028 |
+
"eval_f1_micro": 0.814535597814578,
|
| 1029 |
+
"eval_loss": 0.12324482202529907,
|
| 1030 |
+
"eval_runtime": 462.9556,
|
| 1031 |
+
"eval_samples_per_second": 6.188,
|
| 1032 |
+
"eval_steps_per_second": 0.194,
|
| 1033 |
+
"learning_rate": 1e-05,
|
| 1034 |
+
"step": 17745
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 65.93406593406593,
|
| 1038 |
+
"grad_norm": 0.23069432377815247,
|
| 1039 |
+
"learning_rate": 1e-05,
|
| 1040 |
+
"loss": 0.1273,
|
| 1041 |
+
"step": 18000
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 66.0,
|
| 1045 |
+
"eval_accuracy": 0.3012216404886562,
|
| 1046 |
+
"eval_f1_macro": 0.7004348905275414,
|
| 1047 |
+
"eval_f1_micro": 0.814516812060642,
|
| 1048 |
+
"eval_loss": 0.12292143702507019,
|
| 1049 |
+
"eval_runtime": 464.3912,
|
| 1050 |
+
"eval_samples_per_second": 6.169,
|
| 1051 |
+
"eval_steps_per_second": 0.194,
|
| 1052 |
+
"learning_rate": 1e-05,
|
| 1053 |
+
"step": 18018
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 67.0,
|
| 1057 |
+
"eval_accuracy": 0.30157068062827225,
|
| 1058 |
+
"eval_f1_macro": 0.699284425146712,
|
| 1059 |
+
"eval_f1_micro": 0.8159166351527375,
|
| 1060 |
+
"eval_loss": 0.12375594675540924,
|
| 1061 |
+
"eval_runtime": 464.0044,
|
| 1062 |
+
"eval_samples_per_second": 6.175,
|
| 1063 |
+
"eval_steps_per_second": 0.194,
|
| 1064 |
+
"learning_rate": 1e-05,
|
| 1065 |
+
"step": 18291
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 67.76556776556777,
|
| 1069 |
+
"grad_norm": 0.4648727476596832,
|
| 1070 |
+
"learning_rate": 1e-05,
|
| 1071 |
+
"loss": 0.1272,
|
| 1072 |
+
"step": 18500
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 68.0,
|
| 1076 |
+
"eval_accuracy": 0.299825479930192,
|
| 1077 |
+
"eval_f1_macro": 0.7038988669971363,
|
| 1078 |
+
"eval_f1_micro": 0.8175921663807173,
|
| 1079 |
+
"eval_loss": 0.1228519007563591,
|
| 1080 |
+
"eval_runtime": 462.9465,
|
| 1081 |
+
"eval_samples_per_second": 6.189,
|
| 1082 |
+
"eval_steps_per_second": 0.194,
|
| 1083 |
+
"learning_rate": 1e-05,
|
| 1084 |
+
"step": 18564
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 69.0,
|
| 1088 |
+
"eval_accuracy": 0.3057591623036649,
|
| 1089 |
+
"eval_f1_macro": 0.7006054696492156,
|
| 1090 |
+
"eval_f1_micro": 0.8157472332516685,
|
| 1091 |
+
"eval_loss": 0.12302352488040924,
|
| 1092 |
+
"eval_runtime": 460.1932,
|
| 1093 |
+
"eval_samples_per_second": 6.226,
|
| 1094 |
+
"eval_steps_per_second": 0.196,
|
| 1095 |
+
"learning_rate": 1e-05,
|
| 1096 |
+
"step": 18837
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 69.59706959706959,
|
| 1100 |
+
"grad_norm": 0.19116108119487762,
|
| 1101 |
+
"learning_rate": 1e-05,
|
| 1102 |
+
"loss": 0.1274,
|
| 1103 |
+
"step": 19000
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 70.0,
|
| 1107 |
+
"eval_accuracy": 0.30401396160558464,
|
| 1108 |
+
"eval_f1_macro": 0.7009163688044155,
|
| 1109 |
+
"eval_f1_micro": 0.8169984042999916,
|
| 1110 |
+
"eval_loss": 0.12284138053655624,
|
| 1111 |
+
"eval_runtime": 460.5489,
|
| 1112 |
+
"eval_samples_per_second": 6.221,
|
| 1113 |
+
"eval_steps_per_second": 0.195,
|
| 1114 |
+
"learning_rate": 1e-05,
|
| 1115 |
+
"step": 19110
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 71.0,
|
| 1119 |
+
"eval_accuracy": 0.3019197207678883,
|
| 1120 |
+
"eval_f1_macro": 0.7043489270867269,
|
| 1121 |
+
"eval_f1_micro": 0.8157594696169925,
|
| 1122 |
+
"eval_loss": 0.12295401096343994,
|
| 1123 |
+
"eval_runtime": 459.6032,
|
| 1124 |
+
"eval_samples_per_second": 6.234,
|
| 1125 |
+
"eval_steps_per_second": 0.196,
|
| 1126 |
+
"learning_rate": 1e-05,
|
| 1127 |
+
"step": 19383
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 71.42857142857143,
|
| 1131 |
+
"grad_norm": 0.23645249009132385,
|
| 1132 |
+
"learning_rate": 1e-05,
|
| 1133 |
+
"loss": 0.1272,
|
| 1134 |
+
"step": 19500
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 72.0,
|
| 1138 |
+
"eval_accuracy": 0.30157068062827225,
|
| 1139 |
+
"eval_f1_macro": 0.7024968584049214,
|
| 1140 |
+
"eval_f1_micro": 0.8171375869875074,
|
| 1141 |
+
"eval_loss": 0.123215451836586,
|
| 1142 |
+
"eval_runtime": 456.5626,
|
| 1143 |
+
"eval_samples_per_second": 6.275,
|
| 1144 |
+
"eval_steps_per_second": 0.197,
|
| 1145 |
+
"learning_rate": 1e-05,
|
| 1146 |
+
"step": 19656
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 73.0,
|
| 1150 |
+
"eval_accuracy": 0.3054101221640489,
|
| 1151 |
+
"eval_f1_macro": 0.7048999311445782,
|
| 1152 |
+
"eval_f1_micro": 0.8173905730984317,
|
| 1153 |
+
"eval_loss": 0.12291014939546585,
|
| 1154 |
+
"eval_runtime": 459.2654,
|
| 1155 |
+
"eval_samples_per_second": 6.238,
|
| 1156 |
+
"eval_steps_per_second": 0.196,
|
| 1157 |
+
"learning_rate": 1e-05,
|
| 1158 |
+
"step": 19929
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 73.26007326007326,
|
| 1162 |
+
"grad_norm": 0.18364407122135162,
|
| 1163 |
+
"learning_rate": 1e-05,
|
| 1164 |
+
"loss": 0.1267,
|
| 1165 |
+
"step": 20000
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 74.0,
|
| 1169 |
+
"eval_accuracy": 0.3008726003490401,
|
| 1170 |
+
"eval_f1_macro": 0.6942481843546244,
|
| 1171 |
+
"eval_f1_micro": 0.8141268691688323,
|
| 1172 |
+
"eval_loss": 0.12304174154996872,
|
| 1173 |
+
"eval_runtime": 452.0481,
|
| 1174 |
+
"eval_samples_per_second": 6.338,
|
| 1175 |
+
"eval_steps_per_second": 0.199,
|
| 1176 |
+
"learning_rate": 1e-05,
|
| 1177 |
+
"step": 20202
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 75.0,
|
| 1181 |
+
"eval_accuracy": 0.3033158813263525,
|
| 1182 |
+
"eval_f1_macro": 0.7001404945828892,
|
| 1183 |
+
"eval_f1_micro": 0.8161419818297063,
|
| 1184 |
+
"eval_loss": 0.1232200339436531,
|
| 1185 |
+
"eval_runtime": 455.6916,
|
| 1186 |
+
"eval_samples_per_second": 6.287,
|
| 1187 |
+
"eval_steps_per_second": 0.198,
|
| 1188 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1189 |
+
"step": 20475
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 75.0915750915751,
|
| 1193 |
+
"grad_norm": 0.23693855106830597,
|
| 1194 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1195 |
+
"loss": 0.1269,
|
| 1196 |
+
"step": 20500
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 76.0,
|
| 1200 |
+
"eval_accuracy": 0.3057591623036649,
|
| 1201 |
+
"eval_f1_macro": 0.7020186938135398,
|
| 1202 |
+
"eval_f1_micro": 0.8171137406830337,
|
| 1203 |
+
"eval_loss": 0.12267689406871796,
|
| 1204 |
+
"eval_runtime": 450.1022,
|
| 1205 |
+
"eval_samples_per_second": 6.365,
|
| 1206 |
+
"eval_steps_per_second": 0.2,
|
| 1207 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1208 |
+
"step": 20748
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 76.92307692307692,
|
| 1212 |
+
"grad_norm": 0.2620922923088074,
|
| 1213 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1214 |
+
"loss": 0.1261,
|
| 1215 |
+
"step": 21000
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 77.0,
|
| 1219 |
+
"eval_accuracy": 0.3078534031413613,
|
| 1220 |
+
"eval_f1_macro": 0.7059725105446049,
|
| 1221 |
+
"eval_f1_micro": 0.8190563444800733,
|
| 1222 |
+
"eval_loss": 0.12284990400075912,
|
| 1223 |
+
"eval_runtime": 450.7915,
|
| 1224 |
+
"eval_samples_per_second": 6.355,
|
| 1225 |
+
"eval_steps_per_second": 0.2,
|
| 1226 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1227 |
+
"step": 21021
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 78.0,
|
| 1231 |
+
"eval_accuracy": 0.3019197207678883,
|
| 1232 |
+
"eval_f1_macro": 0.70715796895925,
|
| 1233 |
+
"eval_f1_micro": 0.8165993945509586,
|
| 1234 |
+
"eval_loss": 0.123690165579319,
|
| 1235 |
+
"eval_runtime": 449.2258,
|
| 1236 |
+
"eval_samples_per_second": 6.378,
|
| 1237 |
+
"eval_steps_per_second": 0.2,
|
| 1238 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1239 |
+
"step": 21294
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 78.75457875457876,
|
| 1243 |
+
"grad_norm": 0.2725893557071686,
|
| 1244 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1245 |
+
"loss": 0.1268,
|
| 1246 |
+
"step": 21500
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 79.0,
|
| 1250 |
+
"eval_accuracy": 0.30471204188481676,
|
| 1251 |
+
"eval_f1_macro": 0.6992489783675215,
|
| 1252 |
+
"eval_f1_micro": 0.8156292286874154,
|
| 1253 |
+
"eval_loss": 0.12329532951116562,
|
| 1254 |
+
"eval_runtime": 448.681,
|
| 1255 |
+
"eval_samples_per_second": 6.385,
|
| 1256 |
+
"eval_steps_per_second": 0.201,
|
| 1257 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1258 |
+
"step": 21567
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 80.0,
|
| 1262 |
+
"eval_accuracy": 0.30261780104712044,
|
| 1263 |
+
"eval_f1_macro": 0.6994382162058129,
|
| 1264 |
+
"eval_f1_micro": 0.8171630910227036,
|
| 1265 |
+
"eval_loss": 0.12325812131166458,
|
| 1266 |
+
"eval_runtime": 447.9437,
|
| 1267 |
+
"eval_samples_per_second": 6.396,
|
| 1268 |
+
"eval_steps_per_second": 0.201,
|
| 1269 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1270 |
+
"step": 21840
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 80.58608058608058,
|
| 1274 |
+
"grad_norm": 0.23704515397548676,
|
| 1275 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1276 |
+
"loss": 0.1271,
|
| 1277 |
+
"step": 22000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 81.0,
|
| 1281 |
+
"eval_accuracy": 0.3054101221640489,
|
| 1282 |
+
"eval_f1_macro": 0.7037260965073638,
|
| 1283 |
+
"eval_f1_micro": 0.8175692748252337,
|
| 1284 |
+
"eval_loss": 0.12240613251924515,
|
| 1285 |
+
"eval_runtime": 448.5942,
|
| 1286 |
+
"eval_samples_per_second": 6.387,
|
| 1287 |
+
"eval_steps_per_second": 0.201,
|
| 1288 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1289 |
+
"step": 22113
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 82.0,
|
| 1293 |
+
"eval_accuracy": 0.30017452006980805,
|
| 1294 |
+
"eval_f1_macro": 0.6972115267564822,
|
| 1295 |
+
"eval_f1_micro": 0.8150502097368756,
|
| 1296 |
+
"eval_loss": 0.12270382046699524,
|
| 1297 |
+
"eval_runtime": 454.1003,
|
| 1298 |
+
"eval_samples_per_second": 6.309,
|
| 1299 |
+
"eval_steps_per_second": 0.198,
|
| 1300 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1301 |
+
"step": 22386
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 82.41758241758242,
|
| 1305 |
+
"grad_norm": 0.24043193459510803,
|
| 1306 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1307 |
+
"loss": 0.1263,
|
| 1308 |
+
"step": 22500
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 83.0,
|
| 1312 |
+
"eval_accuracy": 0.2994764397905759,
|
| 1313 |
+
"eval_f1_macro": 0.6939342317359001,
|
| 1314 |
+
"eval_f1_micro": 0.8146007733820593,
|
| 1315 |
+
"eval_loss": 0.12315402179956436,
|
| 1316 |
+
"eval_runtime": 453.2474,
|
| 1317 |
+
"eval_samples_per_second": 6.321,
|
| 1318 |
+
"eval_steps_per_second": 0.199,
|
| 1319 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1320 |
+
"step": 22659
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 84.0,
|
| 1324 |
+
"eval_accuracy": 0.30261780104712044,
|
| 1325 |
+
"eval_f1_macro": 0.7017348072852434,
|
| 1326 |
+
"eval_f1_micro": 0.817694818782958,
|
| 1327 |
+
"eval_loss": 0.1225922629237175,
|
| 1328 |
+
"eval_runtime": 454.9911,
|
| 1329 |
+
"eval_samples_per_second": 6.297,
|
| 1330 |
+
"eval_steps_per_second": 0.198,
|
| 1331 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1332 |
+
"step": 22932
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 84.24908424908425,
|
| 1336 |
+
"grad_norm": 0.20054659247398376,
|
| 1337 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1338 |
+
"loss": 0.1265,
|
| 1339 |
+
"step": 23000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 85.0,
|
| 1343 |
+
"eval_accuracy": 0.30715532286212915,
|
| 1344 |
+
"eval_f1_macro": 0.7059249231157801,
|
| 1345 |
+
"eval_f1_micro": 0.8180560794668679,
|
| 1346 |
+
"eval_loss": 0.1230376735329628,
|
| 1347 |
+
"eval_runtime": 454.3422,
|
| 1348 |
+
"eval_samples_per_second": 6.306,
|
| 1349 |
+
"eval_steps_per_second": 0.198,
|
| 1350 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1351 |
+
"step": 23205
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 86.0,
|
| 1355 |
+
"eval_accuracy": 0.3033158813263525,
|
| 1356 |
+
"eval_f1_macro": 0.7058661509975368,
|
| 1357 |
+
"eval_f1_micro": 0.8168233713901948,
|
| 1358 |
+
"eval_loss": 0.12335028499364853,
|
| 1359 |
+
"eval_runtime": 455.4692,
|
| 1360 |
+
"eval_samples_per_second": 6.29,
|
| 1361 |
+
"eval_steps_per_second": 0.198,
|
| 1362 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1363 |
+
"step": 23478
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 86.08058608058609,
|
| 1367 |
+
"grad_norm": 0.2747795581817627,
|
| 1368 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1369 |
+
"loss": 0.1265,
|
| 1370 |
+
"step": 23500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 87.0,
|
| 1374 |
+
"eval_accuracy": 0.30052356020942406,
|
| 1375 |
+
"eval_f1_macro": 0.6943550775207357,
|
| 1376 |
+
"eval_f1_micro": 0.8143178633603668,
|
| 1377 |
+
"eval_loss": 0.12355341017246246,
|
| 1378 |
+
"eval_runtime": 467.692,
|
| 1379 |
+
"eval_samples_per_second": 6.126,
|
| 1380 |
+
"eval_steps_per_second": 0.192,
|
| 1381 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1382 |
+
"step": 23751
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 87.91208791208791,
|
| 1386 |
+
"grad_norm": 0.2173408567905426,
|
| 1387 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1388 |
+
"loss": 0.1266,
|
| 1389 |
+
"step": 24000
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 88.0,
|
| 1393 |
+
"eval_accuracy": 0.3078534031413613,
|
| 1394 |
+
"eval_f1_macro": 0.702889046791746,
|
| 1395 |
+
"eval_f1_micro": 0.8186198516780492,
|
| 1396 |
+
"eval_loss": 0.12264065444469452,
|
| 1397 |
+
"eval_runtime": 469.0124,
|
| 1398 |
+
"eval_samples_per_second": 6.109,
|
| 1399 |
+
"eval_steps_per_second": 0.192,
|
| 1400 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1401 |
+
"step": 24024
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 89.0,
|
| 1405 |
+
"eval_accuracy": 0.3075043630017452,
|
| 1406 |
+
"eval_f1_macro": 0.7083793564773274,
|
| 1407 |
+
"eval_f1_micro": 0.817912934805249,
|
| 1408 |
+
"eval_loss": 0.1229795441031456,
|
| 1409 |
+
"eval_runtime": 470.229,
|
| 1410 |
+
"eval_samples_per_second": 6.093,
|
| 1411 |
+
"eval_steps_per_second": 0.191,
|
| 1412 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1413 |
+
"step": 24297
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 89.74358974358974,
|
| 1417 |
+
"grad_norm": 0.22348648309707642,
|
| 1418 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1419 |
+
"loss": 0.1263,
|
| 1420 |
+
"step": 24500
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 90.0,
|
| 1424 |
+
"eval_accuracy": 0.3036649214659686,
|
| 1425 |
+
"eval_f1_macro": 0.710478296923057,
|
| 1426 |
+
"eval_f1_micro": 0.8195029312195524,
|
| 1427 |
+
"eval_loss": 0.12317965924739838,
|
| 1428 |
+
"eval_runtime": 465.171,
|
| 1429 |
+
"eval_samples_per_second": 6.159,
|
| 1430 |
+
"eval_steps_per_second": 0.193,
|
| 1431 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1432 |
+
"step": 24570
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 91.0,
|
| 1436 |
+
"eval_accuracy": 0.3019197207678883,
|
| 1437 |
+
"eval_f1_macro": 0.6971981356235558,
|
| 1438 |
+
"eval_f1_micro": 0.8155397390272835,
|
| 1439 |
+
"eval_loss": 0.12267619371414185,
|
| 1440 |
+
"eval_runtime": 458.3455,
|
| 1441 |
+
"eval_samples_per_second": 6.251,
|
| 1442 |
+
"eval_steps_per_second": 0.196,
|
| 1443 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1444 |
+
"step": 24843
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 91.0,
|
| 1448 |
+
"learning_rate": 1.0000000000000002e-07,
|
| 1449 |
+
"step": 24843,
|
| 1450 |
+
"total_flos": 1.1760993126572918e+21,
|
| 1451 |
+
"train_loss": 0.1412498749900842,
|
| 1452 |
+
"train_runtime": 170277.6523,
|
| 1453 |
+
"train_samples_per_second": 7.678,
|
| 1454 |
+
"train_steps_per_second": 0.24
|
| 1455 |
+
}
|
| 1456 |
+
],
|
| 1457 |
+
"logging_steps": 500,
|
| 1458 |
+
"max_steps": 40950,
|
| 1459 |
+
"num_input_tokens_seen": 0,
|
| 1460 |
+
"num_train_epochs": 150,
|
| 1461 |
+
"save_steps": 500,
|
| 1462 |
+
"stateful_callbacks": {
|
| 1463 |
+
"EarlyStoppingCallback": {
|
| 1464 |
+
"args": {
|
| 1465 |
+
"early_stopping_patience": 10,
|
| 1466 |
+
"early_stopping_threshold": 0.0
|
| 1467 |
+
},
|
| 1468 |
+
"attributes": {
|
| 1469 |
+
"early_stopping_patience_counter": 10
|
| 1470 |
+
}
|
| 1471 |
+
},
|
| 1472 |
+
"TrainerControl": {
|
| 1473 |
+
"args": {
|
| 1474 |
+
"should_epoch_stop": false,
|
| 1475 |
+
"should_evaluate": false,
|
| 1476 |
+
"should_log": false,
|
| 1477 |
+
"should_save": true,
|
| 1478 |
+
"should_training_stop": true
|
| 1479 |
+
},
|
| 1480 |
+
"attributes": {}
|
| 1481 |
+
}
|
| 1482 |
+
},
|
| 1483 |
+
"total_flos": 1.1760993126572918e+21,
|
| 1484 |
+
"train_batch_size": 32,
|
| 1485 |
+
"trial_name": null,
|
| 1486 |
+
"trial_params": null
|
| 1487 |
+
}
|