scenario-NON-KD-PO-COPY-D2_data-AmazonScience_massive_all_1_144
This model is a fine-tuned version of haryoaw/scenario-MDBT-TCR_data-AmazonScience_massive_all_1_1 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 2.0782
- Accuracy: 0.8544
- F1: 0.8289
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: 44
- 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 |
---|---|---|---|---|---|
0.5161 | 0.27 | 5000 | 0.6816 | 0.8291 | 0.7878 |
0.3963 | 0.53 | 10000 | 0.6619 | 0.8406 | 0.8132 |
0.3396 | 0.8 | 15000 | 0.6545 | 0.8447 | 0.8161 |
0.2234 | 1.07 | 20000 | 0.7323 | 0.8483 | 0.8257 |
0.229 | 1.34 | 25000 | 0.7528 | 0.8444 | 0.8191 |
0.2172 | 1.6 | 30000 | 0.7349 | 0.8495 | 0.8253 |
0.2086 | 1.87 | 35000 | 0.7411 | 0.8485 | 0.8257 |
0.1495 | 2.14 | 40000 | 0.8672 | 0.8444 | 0.8152 |
0.1561 | 2.41 | 45000 | 0.8479 | 0.8490 | 0.8186 |
0.1557 | 2.67 | 50000 | 0.8802 | 0.8479 | 0.8208 |
0.1543 | 2.94 | 55000 | 0.8013 | 0.8538 | 0.8334 |
0.1052 | 3.21 | 60000 | 0.9790 | 0.8501 | 0.8221 |
0.1104 | 3.47 | 65000 | 0.9687 | 0.8499 | 0.8274 |
0.1245 | 3.74 | 70000 | 0.8966 | 0.8485 | 0.8279 |
0.1056 | 4.01 | 75000 | 1.0067 | 0.8494 | 0.8275 |
0.0875 | 4.28 | 80000 | 1.0929 | 0.8475 | 0.8222 |
0.0881 | 4.54 | 85000 | 1.0596 | 0.8469 | 0.8178 |
0.1 | 4.81 | 90000 | 1.0371 | 0.8519 | 0.8269 |
0.0741 | 5.08 | 95000 | 1.1216 | 0.8462 | 0.8180 |
0.0734 | 5.34 | 100000 | 1.1472 | 0.8492 | 0.8228 |
0.0819 | 5.61 | 105000 | 1.1234 | 0.8482 | 0.8226 |
0.0805 | 5.88 | 110000 | 1.0975 | 0.8496 | 0.8237 |
0.0667 | 6.15 | 115000 | 1.1985 | 0.8476 | 0.8189 |
0.0664 | 6.41 | 120000 | 1.1688 | 0.8482 | 0.8236 |
0.0674 | 6.68 | 125000 | 1.1931 | 0.8462 | 0.8206 |
0.0726 | 6.95 | 130000 | 1.1113 | 0.8517 | 0.8268 |
0.0561 | 7.22 | 135000 | 1.2818 | 0.8484 | 0.8214 |
0.0513 | 7.48 | 140000 | 1.2779 | 0.8504 | 0.8239 |
0.0602 | 7.75 | 145000 | 1.2326 | 0.8502 | 0.8238 |
0.0474 | 8.02 | 150000 | 1.3027 | 0.8500 | 0.8253 |
0.0492 | 8.28 | 155000 | 1.3551 | 0.8475 | 0.8224 |
0.0533 | 8.55 | 160000 | 1.2721 | 0.8498 | 0.8258 |
0.0527 | 8.82 | 165000 | 1.2575 | 0.8490 | 0.8239 |
0.0396 | 9.09 | 170000 | 1.3997 | 0.8489 | 0.8231 |
0.0494 | 9.35 | 175000 | 1.3877 | 0.8472 | 0.8219 |
0.0547 | 9.62 | 180000 | 1.3829 | 0.8476 | 0.8221 |
0.0459 | 9.89 | 185000 | 1.3692 | 0.8511 | 0.8257 |
0.0331 | 10.15 | 190000 | 1.3953 | 0.8492 | 0.8267 |
0.0459 | 10.42 | 195000 | 1.4430 | 0.8483 | 0.8276 |
0.0418 | 10.69 | 200000 | 1.4325 | 0.8495 | 0.8288 |
0.0482 | 10.96 | 205000 | 1.3494 | 0.8514 | 0.8253 |
0.0332 | 11.22 | 210000 | 1.5266 | 0.8500 | 0.8251 |
0.0403 | 11.49 | 215000 | 1.4327 | 0.8535 | 0.8302 |
0.0394 | 11.76 | 220000 | 1.4069 | 0.8513 | 0.8282 |
0.0296 | 12.03 | 225000 | 1.4835 | 0.8521 | 0.8310 |
0.0393 | 12.29 | 230000 | 1.5214 | 0.8503 | 0.8271 |
0.0365 | 12.56 | 235000 | 1.5325 | 0.8508 | 0.8267 |
0.0356 | 12.83 | 240000 | 1.5022 | 0.8505 | 0.8273 |
0.0279 | 13.09 | 245000 | 1.5751 | 0.8486 | 0.8260 |
0.0335 | 13.36 | 250000 | 1.5749 | 0.8483 | 0.8248 |
0.0326 | 13.63 | 255000 | 1.5202 | 0.8509 | 0.8265 |
0.0341 | 13.9 | 260000 | 1.5074 | 0.8496 | 0.8266 |
0.0224 | 14.16 | 265000 | 1.6855 | 0.8476 | 0.8180 |
0.0247 | 14.43 | 270000 | 1.5959 | 0.8501 | 0.8231 |
0.0301 | 14.7 | 275000 | 1.5587 | 0.8515 | 0.8281 |
0.0235 | 14.96 | 280000 | 1.6095 | 0.8488 | 0.8270 |
0.0223 | 15.23 | 285000 | 1.7195 | 0.8485 | 0.8234 |
0.0243 | 15.5 | 290000 | 1.7301 | 0.8483 | 0.8214 |
0.0307 | 15.77 | 295000 | 1.5648 | 0.8518 | 0.8251 |
0.0201 | 16.03 | 300000 | 1.6855 | 0.8506 | 0.8252 |
0.0222 | 16.3 | 305000 | 1.7225 | 0.8484 | 0.8227 |
0.0252 | 16.57 | 310000 | 1.6764 | 0.8486 | 0.8241 |
0.0252 | 16.84 | 315000 | 1.6818 | 0.8498 | 0.8241 |
0.0174 | 17.1 | 320000 | 1.7985 | 0.8519 | 0.8315 |
0.0228 | 17.37 | 325000 | 1.7816 | 0.8492 | 0.8232 |
0.0251 | 17.64 | 330000 | 1.7655 | 0.8500 | 0.8248 |
0.0248 | 17.9 | 335000 | 1.7425 | 0.8512 | 0.8260 |
0.0182 | 18.17 | 340000 | 1.8507 | 0.8497 | 0.8229 |
0.0194 | 18.44 | 345000 | 1.8165 | 0.8503 | 0.8273 |
0.0196 | 18.71 | 350000 | 1.7348 | 0.8493 | 0.8244 |
0.0218 | 18.97 | 355000 | 1.8210 | 0.8514 | 0.8266 |
0.0144 | 19.24 | 360000 | 1.8294 | 0.8520 | 0.8289 |
0.017 | 19.51 | 365000 | 1.8268 | 0.8489 | 0.8249 |
0.02 | 19.77 | 370000 | 1.7976 | 0.8506 | 0.8285 |
0.0137 | 20.04 | 375000 | 1.8765 | 0.8513 | 0.8292 |
0.0143 | 20.31 | 380000 | 1.8971 | 0.8503 | 0.8293 |
0.0143 | 20.58 | 385000 | 1.8773 | 0.8512 | 0.8295 |
0.0136 | 20.84 | 390000 | 1.8520 | 0.8515 | 0.8277 |
0.0143 | 21.11 | 395000 | 1.9031 | 0.8509 | 0.8256 |
0.0124 | 21.38 | 400000 | 1.8758 | 0.8529 | 0.8302 |
0.0125 | 21.65 | 405000 | 2.0046 | 0.8491 | 0.8274 |
0.0141 | 21.91 | 410000 | 1.9489 | 0.8519 | 0.8285 |
0.016 | 22.18 | 415000 | 1.9530 | 0.8529 | 0.8283 |
0.0139 | 22.45 | 420000 | 1.9622 | 0.8521 | 0.8278 |
0.0107 | 22.71 | 425000 | 2.0015 | 0.8511 | 0.8250 |
0.0105 | 22.98 | 430000 | 1.9852 | 0.8534 | 0.8289 |
0.0113 | 23.25 | 435000 | 2.0320 | 0.8520 | 0.8265 |
0.0117 | 23.52 | 440000 | 1.9807 | 0.8530 | 0.8279 |
0.0104 | 23.78 | 445000 | 2.0138 | 0.8515 | 0.8279 |
0.0092 | 24.05 | 450000 | 2.0358 | 0.8520 | 0.8285 |
0.0106 | 24.32 | 455000 | 2.0526 | 0.8516 | 0.8281 |
0.0101 | 24.58 | 460000 | 2.0302 | 0.8532 | 0.8293 |
0.0096 | 24.85 | 465000 | 2.0315 | 0.8530 | 0.8286 |
0.0076 | 25.12 | 470000 | 2.0628 | 0.8530 | 0.8289 |
0.0081 | 25.39 | 475000 | 2.0493 | 0.8537 | 0.8287 |
0.0098 | 25.65 | 480000 | 2.0548 | 0.8550 | 0.8309 |
0.0087 | 25.92 | 485000 | 2.0356 | 0.8544 | 0.8313 |
0.0071 | 26.19 | 490000 | 2.0422 | 0.8545 | 0.8304 |
0.0099 | 26.46 | 495000 | 2.0361 | 0.8537 | 0.8276 |
0.0093 | 26.72 | 500000 | 2.0464 | 0.8545 | 0.8296 |
0.009 | 26.99 | 505000 | 2.0414 | 0.8531 | 0.8271 |
0.0081 | 27.26 | 510000 | 2.0718 | 0.8530 | 0.8269 |
0.0069 | 27.52 | 515000 | 2.0666 | 0.8543 | 0.8284 |
0.0085 | 27.79 | 520000 | 2.0561 | 0.8536 | 0.8267 |
0.0084 | 28.06 | 525000 | 2.0665 | 0.8540 | 0.8280 |
0.0072 | 28.33 | 530000 | 2.0960 | 0.8536 | 0.8284 |
0.0067 | 28.59 | 535000 | 2.0782 | 0.8535 | 0.8271 |
0.0062 | 28.86 | 540000 | 2.0716 | 0.8537 | 0.8280 |
0.008 | 29.13 | 545000 | 2.0839 | 0.8539 | 0.8286 |
0.0072 | 29.39 | 550000 | 2.0825 | 0.8539 | 0.8281 |
0.0059 | 29.66 | 555000 | 2.0781 | 0.8544 | 0.8286 |
0.0072 | 29.93 | 560000 | 2.0782 | 0.8544 | 0.8289 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.1.1+cu121
- Datasets 2.14.5
- Tokenizers 0.13.3
- Downloads last month
- 3
Model tree for haryoaw/scenario-NON-KD-PO-COPY-D2_data-AmazonScience_massive_all_1_144
Base model
microsoft/mdeberta-v3-base