|
--- |
|
license: mit |
|
base_model: microsoft/mdeberta-v3-base |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- massive |
|
metrics: |
|
- accuracy |
|
- f1 |
|
model-index: |
|
- name: scenario-KD-PR-MSV-D2_data-AmazonScience_massive_all_1_166 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# scenario-KD-PR-MSV-D2_data-AmazonScience_massive_all_1_166 |
|
|
|
This model is a fine-tuned version of [microsoft/mdeberta-v3-base](https://huggingface.co/microsoft/mdeberta-v3-base) on the massive dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.5394 |
|
- Accuracy: 0.8627 |
|
- F1: 0.8453 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 5e-05 |
|
- train_batch_size: 32 |
|
- eval_batch_size: 32 |
|
- seed: 66 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 30 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |
|
|:-------------:|:-----:|:------:|:---------------:|:--------:|:------:| |
|
| 3.4018 | 0.27 | 5000 | 3.3570 | 0.7587 | 0.6728 | |
|
| 2.4831 | 0.53 | 10000 | 2.6998 | 0.8028 | 0.7596 | |
|
| 2.1077 | 0.8 | 15000 | 2.4164 | 0.8202 | 0.7862 | |
|
| 1.5375 | 1.07 | 20000 | 2.2821 | 0.8317 | 0.8012 | |
|
| 1.4474 | 1.34 | 25000 | 2.2615 | 0.8325 | 0.8043 | |
|
| 1.3836 | 1.6 | 30000 | 2.1448 | 0.8366 | 0.8133 | |
|
| 1.2782 | 1.87 | 35000 | 2.1212 | 0.8390 | 0.8135 | |
|
| 0.983 | 2.14 | 40000 | 2.0840 | 0.8429 | 0.8209 | |
|
| 0.9416 | 2.41 | 45000 | 2.1779 | 0.8422 | 0.8206 | |
|
| 0.9138 | 2.67 | 50000 | 2.0942 | 0.8447 | 0.8247 | |
|
| 0.9093 | 2.94 | 55000 | 2.0603 | 0.8454 | 0.8196 | |
|
| 0.7465 | 3.21 | 60000 | 2.0972 | 0.8441 | 0.8245 | |
|
| 0.6996 | 3.47 | 65000 | 2.0355 | 0.8475 | 0.8270 | |
|
| 0.7454 | 3.74 | 70000 | 1.9610 | 0.8487 | 0.8298 | |
|
| 0.6906 | 4.01 | 75000 | 2.0084 | 0.8467 | 0.8288 | |
|
| 0.6203 | 4.28 | 80000 | 1.9601 | 0.8498 | 0.8306 | |
|
| 0.6039 | 4.54 | 85000 | 1.9766 | 0.8509 | 0.8347 | |
|
| 0.616 | 4.81 | 90000 | 1.9302 | 0.8518 | 0.8295 | |
|
| 0.5404 | 5.08 | 95000 | 1.9323 | 0.8512 | 0.8301 | |
|
| 0.5448 | 5.34 | 100000 | 1.9360 | 0.8533 | 0.8383 | |
|
| 0.5377 | 5.61 | 105000 | 1.9353 | 0.8511 | 0.8292 | |
|
| 0.5373 | 5.88 | 110000 | 1.9015 | 0.8506 | 0.8318 | |
|
| 0.4744 | 6.15 | 115000 | 1.9116 | 0.8527 | 0.8333 | |
|
| 0.4885 | 6.41 | 120000 | 1.8676 | 0.8543 | 0.8370 | |
|
| 0.4886 | 6.68 | 125000 | 1.8716 | 0.8546 | 0.8344 | |
|
| 0.4861 | 6.95 | 130000 | 1.8664 | 0.8535 | 0.8319 | |
|
| 0.4488 | 7.22 | 135000 | 1.8560 | 0.8547 | 0.8376 | |
|
| 0.426 | 7.48 | 140000 | 1.8350 | 0.8535 | 0.8334 | |
|
| 0.4451 | 7.75 | 145000 | 1.8258 | 0.8544 | 0.8333 | |
|
| 0.4299 | 8.02 | 150000 | 1.8220 | 0.8560 | 0.8370 | |
|
| 0.4207 | 8.28 | 155000 | 1.8048 | 0.8559 | 0.8373 | |
|
| 0.4033 | 8.55 | 160000 | 1.8295 | 0.8538 | 0.8367 | |
|
| 0.4039 | 8.82 | 165000 | 1.7818 | 0.8566 | 0.8391 | |
|
| 0.3874 | 9.09 | 170000 | 1.7857 | 0.8563 | 0.8391 | |
|
| 0.3843 | 9.35 | 175000 | 1.7860 | 0.8548 | 0.8374 | |
|
| 0.3882 | 9.62 | 180000 | 1.8074 | 0.8558 | 0.8374 | |
|
| 0.3866 | 9.89 | 185000 | 1.7823 | 0.8583 | 0.8404 | |
|
| 0.36 | 10.15 | 190000 | 1.7294 | 0.8571 | 0.8375 | |
|
| 0.3592 | 10.42 | 195000 | 1.7363 | 0.8578 | 0.8399 | |
|
| 0.3628 | 10.69 | 200000 | 1.7460 | 0.8582 | 0.8385 | |
|
| 0.3579 | 10.96 | 205000 | 1.7431 | 0.8580 | 0.8399 | |
|
| 0.3448 | 11.22 | 210000 | 1.7398 | 0.8564 | 0.8378 | |
|
| 0.3512 | 11.49 | 215000 | 1.7193 | 0.8584 | 0.8402 | |
|
| 0.3367 | 11.76 | 220000 | 1.7197 | 0.8594 | 0.8425 | |
|
| 0.327 | 12.03 | 225000 | 1.7189 | 0.8576 | 0.8385 | |
|
| 0.3248 | 12.29 | 230000 | 1.6991 | 0.8602 | 0.8398 | |
|
| 0.3306 | 12.56 | 235000 | 1.7119 | 0.8577 | 0.8404 | |
|
| 0.3181 | 12.83 | 240000 | 1.6892 | 0.8606 | 0.8414 | |
|
| 0.3167 | 13.09 | 245000 | 1.6647 | 0.8590 | 0.8380 | |
|
| 0.3149 | 13.36 | 250000 | 1.6780 | 0.8590 | 0.8414 | |
|
| 0.3221 | 13.63 | 255000 | 1.6626 | 0.8601 | 0.8437 | |
|
| 0.3147 | 13.9 | 260000 | 1.7135 | 0.8595 | 0.8418 | |
|
| 0.2954 | 14.16 | 265000 | 1.6915 | 0.8581 | 0.8390 | |
|
| 0.2912 | 14.43 | 270000 | 1.6699 | 0.8582 | 0.8392 | |
|
| 0.3123 | 14.7 | 275000 | 1.6659 | 0.8589 | 0.8399 | |
|
| 0.3047 | 14.96 | 280000 | 1.6654 | 0.8610 | 0.8443 | |
|
| 0.2916 | 15.23 | 285000 | 1.6408 | 0.8600 | 0.8421 | |
|
| 0.282 | 15.5 | 290000 | 1.6729 | 0.8580 | 0.8405 | |
|
| 0.2843 | 15.77 | 295000 | 1.6475 | 0.8600 | 0.8416 | |
|
| 0.2764 | 16.03 | 300000 | 1.6342 | 0.8607 | 0.8426 | |
|
| 0.2726 | 16.3 | 305000 | 1.6541 | 0.8597 | 0.8425 | |
|
| 0.2895 | 16.57 | 310000 | 1.6280 | 0.8597 | 0.8413 | |
|
| 0.2744 | 16.84 | 315000 | 1.6453 | 0.8607 | 0.8422 | |
|
| 0.2727 | 17.1 | 320000 | 1.6319 | 0.8600 | 0.8432 | |
|
| 0.2708 | 17.37 | 325000 | 1.6395 | 0.8599 | 0.8427 | |
|
| 0.271 | 17.64 | 330000 | 1.6232 | 0.8600 | 0.8403 | |
|
| 0.2695 | 17.9 | 335000 | 1.6294 | 0.8597 | 0.8419 | |
|
| 0.2698 | 18.17 | 340000 | 1.6158 | 0.8620 | 0.8438 | |
|
| 0.2582 | 18.44 | 345000 | 1.6214 | 0.8625 | 0.8448 | |
|
| 0.2614 | 18.71 | 350000 | 1.6112 | 0.8610 | 0.8431 | |
|
| 0.2583 | 18.97 | 355000 | 1.5978 | 0.8620 | 0.8440 | |
|
| 0.258 | 19.24 | 360000 | 1.5902 | 0.8623 | 0.8446 | |
|
| 0.2498 | 19.51 | 365000 | 1.6081 | 0.8611 | 0.8427 | |
|
| 0.2569 | 19.77 | 370000 | 1.6165 | 0.8604 | 0.8420 | |
|
| 0.2395 | 20.04 | 375000 | 1.5880 | 0.8614 | 0.8433 | |
|
| 0.2527 | 20.31 | 380000 | 1.6055 | 0.8599 | 0.8428 | |
|
| 0.2504 | 20.58 | 385000 | 1.5929 | 0.8614 | 0.8443 | |
|
| 0.2494 | 20.84 | 390000 | 1.5841 | 0.8624 | 0.8444 | |
|
| 0.2434 | 21.11 | 395000 | 1.5833 | 0.8614 | 0.8446 | |
|
| 0.243 | 21.38 | 400000 | 1.5739 | 0.8619 | 0.8438 | |
|
| 0.2389 | 21.65 | 405000 | 1.5816 | 0.8619 | 0.8438 | |
|
| 0.2467 | 21.91 | 410000 | 1.5844 | 0.8616 | 0.8439 | |
|
| 0.2352 | 22.18 | 415000 | 1.5748 | 0.8628 | 0.8446 | |
|
| 0.2323 | 22.45 | 420000 | 1.5654 | 0.8623 | 0.8427 | |
|
| 0.2314 | 22.71 | 425000 | 1.5537 | 0.8627 | 0.8449 | |
|
| 0.238 | 22.98 | 430000 | 1.5613 | 0.8624 | 0.8424 | |
|
| 0.223 | 23.25 | 435000 | 1.5661 | 0.8626 | 0.8441 | |
|
| 0.2287 | 23.52 | 440000 | 1.5714 | 0.8627 | 0.8447 | |
|
| 0.239 | 23.78 | 445000 | 1.5594 | 0.8634 | 0.8455 | |
|
| 0.2275 | 24.05 | 450000 | 1.5629 | 0.8615 | 0.8436 | |
|
| 0.2232 | 24.32 | 455000 | 1.5725 | 0.8618 | 0.8451 | |
|
| 0.2267 | 24.58 | 460000 | 1.5550 | 0.8627 | 0.8455 | |
|
| 0.2248 | 24.85 | 465000 | 1.5574 | 0.8633 | 0.8455 | |
|
| 0.2214 | 25.12 | 470000 | 1.5602 | 0.8613 | 0.8432 | |
|
| 0.2205 | 25.39 | 475000 | 1.5599 | 0.8617 | 0.8432 | |
|
| 0.2189 | 25.65 | 480000 | 1.5395 | 0.8620 | 0.8452 | |
|
| 0.2174 | 25.92 | 485000 | 1.5577 | 0.8625 | 0.8445 | |
|
| 0.2148 | 26.19 | 490000 | 1.5533 | 0.8628 | 0.8457 | |
|
| 0.2175 | 26.46 | 495000 | 1.5496 | 0.8619 | 0.8443 | |
|
| 0.2121 | 26.72 | 500000 | 1.5509 | 0.8617 | 0.8443 | |
|
| 0.2163 | 26.99 | 505000 | 1.5560 | 0.8624 | 0.8453 | |
|
| 0.211 | 27.26 | 510000 | 1.5491 | 0.8629 | 0.8459 | |
|
| 0.2142 | 27.52 | 515000 | 1.5576 | 0.8607 | 0.8438 | |
|
| 0.2084 | 27.79 | 520000 | 1.5522 | 0.8624 | 0.8456 | |
|
| 0.2119 | 28.06 | 525000 | 1.5429 | 0.8621 | 0.8449 | |
|
| 0.2008 | 28.33 | 530000 | 1.5452 | 0.8627 | 0.8465 | |
|
| 0.2084 | 28.59 | 535000 | 1.5458 | 0.8628 | 0.8464 | |
|
| 0.2086 | 28.86 | 540000 | 1.5454 | 0.8622 | 0.8446 | |
|
| 0.2102 | 29.13 | 545000 | 1.5487 | 0.8622 | 0.8449 | |
|
| 0.2118 | 29.39 | 550000 | 1.5448 | 0.8621 | 0.8451 | |
|
| 0.2049 | 29.66 | 555000 | 1.5411 | 0.8626 | 0.8454 | |
|
| 0.2026 | 29.93 | 560000 | 1.5394 | 0.8627 | 0.8453 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.33.3 |
|
- Pytorch 2.1.1+cu121 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.13.3 |
|
|
