Initial Commit
Browse files- README.md +175 -0
- config.json +153 -0
- pytorch_model.bin +3 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,175 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
base_model: FacebookAI/xlm-roberta-base
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- massive
|
| 8 |
+
metrics:
|
| 9 |
+
- accuracy
|
| 10 |
+
- f1
|
| 11 |
+
model-index:
|
| 12 |
+
- name: scenario-KD-SCR-MSV-D2_data-AmazonScience_massive_all_1_1_beta-jason
|
| 13 |
+
results: []
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 17 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 18 |
+
|
| 19 |
+
# scenario-KD-SCR-MSV-D2_data-AmazonScience_massive_all_1_1_beta-jason
|
| 20 |
+
|
| 21 |
+
This model is a fine-tuned version of [FacebookAI/xlm-roberta-base](https://huggingface.co/FacebookAI/xlm-roberta-base) on the massive dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 1.8062
|
| 24 |
+
- Accuracy: 0.8268
|
| 25 |
+
- F1: 0.8025
|
| 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: 5e-05
|
| 45 |
+
- train_batch_size: 32
|
| 46 |
+
- eval_batch_size: 32
|
| 47 |
+
- seed: 4444
|
| 48 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
+
- lr_scheduler_type: linear
|
| 50 |
+
- num_epochs: 30
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|
| 55 |
+
|:-------------:|:-----:|:------:|:---------------:|:--------:|:------:|
|
| 56 |
+
| 5.8616 | 0.27 | 5000 | 5.7378 | 0.3131 | 0.1357 |
|
| 57 |
+
| 4.0354 | 0.53 | 10000 | 4.0456 | 0.5639 | 0.4108 |
|
| 58 |
+
| 3.214 | 0.8 | 15000 | 3.2320 | 0.6631 | 0.5563 |
|
| 59 |
+
| 2.5727 | 1.07 | 20000 | 2.8403 | 0.7076 | 0.6163 |
|
| 60 |
+
| 2.3415 | 1.34 | 25000 | 2.6493 | 0.7307 | 0.6653 |
|
| 61 |
+
| 2.2015 | 1.6 | 30000 | 2.4508 | 0.7523 | 0.6823 |
|
| 62 |
+
| 2.1388 | 1.87 | 35000 | 2.3180 | 0.7656 | 0.7033 |
|
| 63 |
+
| 1.6356 | 2.14 | 40000 | 2.3088 | 0.7689 | 0.7156 |
|
| 64 |
+
| 1.6347 | 2.41 | 45000 | 2.1922 | 0.7797 | 0.7354 |
|
| 65 |
+
| 1.6499 | 2.67 | 50000 | 2.1674 | 0.7868 | 0.7428 |
|
| 66 |
+
| 1.5928 | 2.94 | 55000 | 2.0982 | 0.7911 | 0.7513 |
|
| 67 |
+
| 1.2806 | 3.21 | 60000 | 2.1371 | 0.7923 | 0.7537 |
|
| 68 |
+
| 1.3074 | 3.47 | 65000 | 2.1245 | 0.7915 | 0.7541 |
|
| 69 |
+
| 1.2884 | 3.74 | 70000 | 2.0867 | 0.7953 | 0.7601 |
|
| 70 |
+
| 1.2463 | 4.01 | 75000 | 2.0390 | 0.8025 | 0.7662 |
|
| 71 |
+
| 1.0794 | 4.28 | 80000 | 2.0816 | 0.8009 | 0.7686 |
|
| 72 |
+
| 1.0835 | 4.54 | 85000 | 2.0316 | 0.8034 | 0.7710 |
|
| 73 |
+
| 1.1207 | 4.81 | 90000 | 2.0470 | 0.8004 | 0.7688 |
|
| 74 |
+
| 0.9022 | 5.08 | 95000 | 2.0423 | 0.8030 | 0.7701 |
|
| 75 |
+
| 0.9286 | 5.34 | 100000 | 2.0434 | 0.8041 | 0.7715 |
|
| 76 |
+
| 0.9315 | 5.61 | 105000 | 2.0235 | 0.8076 | 0.7726 |
|
| 77 |
+
| 0.9399 | 5.88 | 110000 | 2.0008 | 0.8079 | 0.7754 |
|
| 78 |
+
| 0.81 | 6.15 | 115000 | 2.0305 | 0.8084 | 0.7779 |
|
| 79 |
+
| 0.806 | 6.41 | 120000 | 2.0511 | 0.8074 | 0.7814 |
|
| 80 |
+
| 0.8236 | 6.68 | 125000 | 2.0400 | 0.8093 | 0.7814 |
|
| 81 |
+
| 0.8242 | 6.95 | 130000 | 2.0178 | 0.8108 | 0.7823 |
|
| 82 |
+
| 0.7284 | 7.22 | 135000 | 2.0302 | 0.8100 | 0.7804 |
|
| 83 |
+
| 0.7307 | 7.48 | 140000 | 2.0083 | 0.8107 | 0.7835 |
|
| 84 |
+
| 0.7513 | 7.75 | 145000 | 2.0581 | 0.8091 | 0.7788 |
|
| 85 |
+
| 0.7044 | 8.02 | 150000 | 2.0368 | 0.8100 | 0.7853 |
|
| 86 |
+
| 0.652 | 8.28 | 155000 | 1.9963 | 0.8134 | 0.7884 |
|
| 87 |
+
| 0.6853 | 8.55 | 160000 | 1.9990 | 0.8123 | 0.7877 |
|
| 88 |
+
| 0.6805 | 8.82 | 165000 | 2.0179 | 0.8111 | 0.7821 |
|
| 89 |
+
| 0.6164 | 9.09 | 170000 | 2.0073 | 0.8140 | 0.7860 |
|
| 90 |
+
| 0.6112 | 9.35 | 175000 | 2.0204 | 0.8139 | 0.7901 |
|
| 91 |
+
| 0.6364 | 9.62 | 180000 | 1.9777 | 0.8152 | 0.7889 |
|
| 92 |
+
| 0.6379 | 9.89 | 185000 | 1.9776 | 0.8152 | 0.7875 |
|
| 93 |
+
| 0.5741 | 10.15 | 190000 | 2.0021 | 0.8161 | 0.7898 |
|
| 94 |
+
| 0.5912 | 10.42 | 195000 | 1.9818 | 0.8151 | 0.7898 |
|
| 95 |
+
| 0.5889 | 10.69 | 200000 | 2.0130 | 0.8156 | 0.7937 |
|
| 96 |
+
| 0.5912 | 10.96 | 205000 | 1.9661 | 0.8168 | 0.7914 |
|
| 97 |
+
| 0.5541 | 11.22 | 210000 | 1.9663 | 0.8183 | 0.7928 |
|
| 98 |
+
| 0.5668 | 11.49 | 215000 | 1.9864 | 0.8160 | 0.7895 |
|
| 99 |
+
| 0.5706 | 11.76 | 220000 | 1.9790 | 0.8163 | 0.7916 |
|
| 100 |
+
| 0.5272 | 12.03 | 225000 | 1.9643 | 0.8168 | 0.7948 |
|
| 101 |
+
| 0.5188 | 12.29 | 230000 | 1.9595 | 0.8176 | 0.7944 |
|
| 102 |
+
| 0.5248 | 12.56 | 235000 | 1.9694 | 0.8163 | 0.7915 |
|
| 103 |
+
| 0.5115 | 12.83 | 240000 | 1.9617 | 0.8179 | 0.7913 |
|
| 104 |
+
| 0.4857 | 13.09 | 245000 | 1.9211 | 0.8200 | 0.7931 |
|
| 105 |
+
| 0.4943 | 13.36 | 250000 | 1.9530 | 0.8193 | 0.7965 |
|
| 106 |
+
| 0.5122 | 13.63 | 255000 | 1.9413 | 0.8182 | 0.7892 |
|
| 107 |
+
| 0.5008 | 13.9 | 260000 | 1.9578 | 0.8170 | 0.7951 |
|
| 108 |
+
| 0.4694 | 14.16 | 265000 | 1.9340 | 0.8193 | 0.7958 |
|
| 109 |
+
| 0.4718 | 14.43 | 270000 | 1.9197 | 0.8210 | 0.7957 |
|
| 110 |
+
| 0.4785 | 14.7 | 275000 | 1.9179 | 0.8209 | 0.7942 |
|
| 111 |
+
| 0.49 | 14.96 | 280000 | 1.9358 | 0.8194 | 0.7925 |
|
| 112 |
+
| 0.4561 | 15.23 | 285000 | 1.9124 | 0.8210 | 0.7958 |
|
| 113 |
+
| 0.4702 | 15.5 | 290000 | 1.9025 | 0.8213 | 0.7961 |
|
| 114 |
+
| 0.4637 | 15.77 | 295000 | 1.9314 | 0.8189 | 0.7944 |
|
| 115 |
+
| 0.4363 | 16.03 | 300000 | 1.8834 | 0.8230 | 0.7979 |
|
| 116 |
+
| 0.4513 | 16.3 | 305000 | 1.8924 | 0.8228 | 0.8007 |
|
| 117 |
+
| 0.4608 | 16.57 | 310000 | 1.8992 | 0.8221 | 0.7997 |
|
| 118 |
+
| 0.4416 | 16.84 | 315000 | 1.8989 | 0.8221 | 0.7984 |
|
| 119 |
+
| 0.4273 | 17.1 | 320000 | 1.8973 | 0.8221 | 0.7997 |
|
| 120 |
+
| 0.432 | 17.37 | 325000 | 1.8992 | 0.8219 | 0.7990 |
|
| 121 |
+
| 0.4287 | 17.64 | 330000 | 1.8878 | 0.8225 | 0.7993 |
|
| 122 |
+
| 0.4366 | 17.9 | 335000 | 1.9000 | 0.8214 | 0.7964 |
|
| 123 |
+
| 0.4143 | 18.17 | 340000 | 1.8816 | 0.8225 | 0.7973 |
|
| 124 |
+
| 0.417 | 18.44 | 345000 | 1.8695 | 0.8225 | 0.7972 |
|
| 125 |
+
| 0.4156 | 18.71 | 350000 | 1.8747 | 0.8239 | 0.7990 |
|
| 126 |
+
| 0.419 | 18.97 | 355000 | 1.8639 | 0.8240 | 0.7985 |
|
| 127 |
+
| 0.4101 | 19.24 | 360000 | 1.8648 | 0.8239 | 0.8002 |
|
| 128 |
+
| 0.4058 | 19.51 | 365000 | 1.8534 | 0.8251 | 0.8001 |
|
| 129 |
+
| 0.4095 | 19.77 | 370000 | 1.8595 | 0.8246 | 0.8001 |
|
| 130 |
+
| 0.4034 | 20.04 | 375000 | 1.8520 | 0.8242 | 0.7981 |
|
| 131 |
+
| 0.4045 | 20.31 | 380000 | 1.8533 | 0.8246 | 0.8001 |
|
| 132 |
+
| 0.3922 | 20.58 | 385000 | 1.8548 | 0.8239 | 0.7988 |
|
| 133 |
+
| 0.4046 | 20.84 | 390000 | 1.8570 | 0.8246 | 0.8013 |
|
| 134 |
+
| 0.3971 | 21.11 | 395000 | 1.8472 | 0.8238 | 0.7999 |
|
| 135 |
+
| 0.3929 | 21.38 | 400000 | 1.8406 | 0.8244 | 0.8001 |
|
| 136 |
+
| 0.3881 | 21.65 | 405000 | 1.8515 | 0.8245 | 0.8008 |
|
| 137 |
+
| 0.3937 | 21.91 | 410000 | 1.8364 | 0.8240 | 0.8003 |
|
| 138 |
+
| 0.3807 | 22.18 | 415000 | 1.8360 | 0.8254 | 0.7999 |
|
| 139 |
+
| 0.3762 | 22.45 | 420000 | 1.8352 | 0.8253 | 0.7998 |
|
| 140 |
+
| 0.389 | 22.71 | 425000 | 1.8429 | 0.8245 | 0.7991 |
|
| 141 |
+
| 0.3751 | 22.98 | 430000 | 1.8262 | 0.8256 | 0.8016 |
|
| 142 |
+
| 0.3637 | 23.25 | 435000 | 1.8275 | 0.8252 | 0.8014 |
|
| 143 |
+
| 0.3758 | 23.52 | 440000 | 1.8359 | 0.8252 | 0.8004 |
|
| 144 |
+
| 0.3773 | 23.78 | 445000 | 1.8230 | 0.8240 | 0.7982 |
|
| 145 |
+
| 0.3573 | 24.05 | 450000 | 1.8289 | 0.8251 | 0.8000 |
|
| 146 |
+
| 0.3689 | 24.32 | 455000 | 1.8260 | 0.8269 | 0.8035 |
|
| 147 |
+
| 0.3661 | 24.58 | 460000 | 1.8287 | 0.8255 | 0.8009 |
|
| 148 |
+
| 0.3648 | 24.85 | 465000 | 1.8294 | 0.8259 | 0.8024 |
|
| 149 |
+
| 0.3521 | 25.12 | 470000 | 1.8233 | 0.8261 | 0.8031 |
|
| 150 |
+
| 0.359 | 25.39 | 475000 | 1.8137 | 0.8264 | 0.8006 |
|
| 151 |
+
| 0.3616 | 25.65 | 480000 | 1.8337 | 0.8245 | 0.8004 |
|
| 152 |
+
| 0.361 | 25.92 | 485000 | 1.8157 | 0.8262 | 0.8008 |
|
| 153 |
+
| 0.3472 | 26.19 | 490000 | 1.8244 | 0.8263 | 0.8019 |
|
| 154 |
+
| 0.3532 | 26.46 | 495000 | 1.8270 | 0.8246 | 0.8011 |
|
| 155 |
+
| 0.3497 | 26.72 | 500000 | 1.8150 | 0.8266 | 0.8020 |
|
| 156 |
+
| 0.3539 | 26.99 | 505000 | 1.8159 | 0.8262 | 0.8016 |
|
| 157 |
+
| 0.3508 | 27.26 | 510000 | 1.8135 | 0.8256 | 0.8014 |
|
| 158 |
+
| 0.3397 | 27.52 | 515000 | 1.8092 | 0.8255 | 0.8004 |
|
| 159 |
+
| 0.3475 | 27.79 | 520000 | 1.8093 | 0.8259 | 0.8005 |
|
| 160 |
+
| 0.3431 | 28.06 | 525000 | 1.8102 | 0.8263 | 0.8008 |
|
| 161 |
+
| 0.3473 | 28.33 | 530000 | 1.8088 | 0.8265 | 0.8011 |
|
| 162 |
+
| 0.3435 | 28.59 | 535000 | 1.8042 | 0.8263 | 0.8017 |
|
| 163 |
+
| 0.3425 | 28.86 | 540000 | 1.8117 | 0.8260 | 0.8012 |
|
| 164 |
+
| 0.3347 | 29.13 | 545000 | 1.8131 | 0.8264 | 0.8017 |
|
| 165 |
+
| 0.3322 | 29.39 | 550000 | 1.8088 | 0.8257 | 0.8007 |
|
| 166 |
+
| 0.3375 | 29.66 | 555000 | 1.8070 | 0.8255 | 0.8006 |
|
| 167 |
+
| 0.3346 | 29.93 | 560000 | 1.8062 | 0.8268 | 0.8025 |
|
| 168 |
+
|
| 169 |
+
|
| 170 |
+
### Framework versions
|
| 171 |
+
|
| 172 |
+
- Transformers 4.33.3
|
| 173 |
+
- Pytorch 2.1.1+cu121
|
| 174 |
+
- Datasets 2.14.5
|
| 175 |
+
- Tokenizers 0.13.3
|
config.json
ADDED
|
@@ -0,0 +1,153 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "FacebookAI/xlm-roberta-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaForSequenceClassificationKD"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 384,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0",
|
| 15 |
+
"1": "LABEL_1",
|
| 16 |
+
"2": "LABEL_2",
|
| 17 |
+
"3": "LABEL_3",
|
| 18 |
+
"4": "LABEL_4",
|
| 19 |
+
"5": "LABEL_5",
|
| 20 |
+
"6": "LABEL_6",
|
| 21 |
+
"7": "LABEL_7",
|
| 22 |
+
"8": "LABEL_8",
|
| 23 |
+
"9": "LABEL_9",
|
| 24 |
+
"10": "LABEL_10",
|
| 25 |
+
"11": "LABEL_11",
|
| 26 |
+
"12": "LABEL_12",
|
| 27 |
+
"13": "LABEL_13",
|
| 28 |
+
"14": "LABEL_14",
|
| 29 |
+
"15": "LABEL_15",
|
| 30 |
+
"16": "LABEL_16",
|
| 31 |
+
"17": "LABEL_17",
|
| 32 |
+
"18": "LABEL_18",
|
| 33 |
+
"19": "LABEL_19",
|
| 34 |
+
"20": "LABEL_20",
|
| 35 |
+
"21": "LABEL_21",
|
| 36 |
+
"22": "LABEL_22",
|
| 37 |
+
"23": "LABEL_23",
|
| 38 |
+
"24": "LABEL_24",
|
| 39 |
+
"25": "LABEL_25",
|
| 40 |
+
"26": "LABEL_26",
|
| 41 |
+
"27": "LABEL_27",
|
| 42 |
+
"28": "LABEL_28",
|
| 43 |
+
"29": "LABEL_29",
|
| 44 |
+
"30": "LABEL_30",
|
| 45 |
+
"31": "LABEL_31",
|
| 46 |
+
"32": "LABEL_32",
|
| 47 |
+
"33": "LABEL_33",
|
| 48 |
+
"34": "LABEL_34",
|
| 49 |
+
"35": "LABEL_35",
|
| 50 |
+
"36": "LABEL_36",
|
| 51 |
+
"37": "LABEL_37",
|
| 52 |
+
"38": "LABEL_38",
|
| 53 |
+
"39": "LABEL_39",
|
| 54 |
+
"40": "LABEL_40",
|
| 55 |
+
"41": "LABEL_41",
|
| 56 |
+
"42": "LABEL_42",
|
| 57 |
+
"43": "LABEL_43",
|
| 58 |
+
"44": "LABEL_44",
|
| 59 |
+
"45": "LABEL_45",
|
| 60 |
+
"46": "LABEL_46",
|
| 61 |
+
"47": "LABEL_47",
|
| 62 |
+
"48": "LABEL_48",
|
| 63 |
+
"49": "LABEL_49",
|
| 64 |
+
"50": "LABEL_50",
|
| 65 |
+
"51": "LABEL_51",
|
| 66 |
+
"52": "LABEL_52",
|
| 67 |
+
"53": "LABEL_53",
|
| 68 |
+
"54": "LABEL_54",
|
| 69 |
+
"55": "LABEL_55",
|
| 70 |
+
"56": "LABEL_56",
|
| 71 |
+
"57": "LABEL_57",
|
| 72 |
+
"58": "LABEL_58",
|
| 73 |
+
"59": "LABEL_59"
|
| 74 |
+
},
|
| 75 |
+
"initializer_range": 0.02,
|
| 76 |
+
"intermediate_size": 1536,
|
| 77 |
+
"label2id": {
|
| 78 |
+
"LABEL_0": 0,
|
| 79 |
+
"LABEL_1": 1,
|
| 80 |
+
"LABEL_10": 10,
|
| 81 |
+
"LABEL_11": 11,
|
| 82 |
+
"LABEL_12": 12,
|
| 83 |
+
"LABEL_13": 13,
|
| 84 |
+
"LABEL_14": 14,
|
| 85 |
+
"LABEL_15": 15,
|
| 86 |
+
"LABEL_16": 16,
|
| 87 |
+
"LABEL_17": 17,
|
| 88 |
+
"LABEL_18": 18,
|
| 89 |
+
"LABEL_19": 19,
|
| 90 |
+
"LABEL_2": 2,
|
| 91 |
+
"LABEL_20": 20,
|
| 92 |
+
"LABEL_21": 21,
|
| 93 |
+
"LABEL_22": 22,
|
| 94 |
+
"LABEL_23": 23,
|
| 95 |
+
"LABEL_24": 24,
|
| 96 |
+
"LABEL_25": 25,
|
| 97 |
+
"LABEL_26": 26,
|
| 98 |
+
"LABEL_27": 27,
|
| 99 |
+
"LABEL_28": 28,
|
| 100 |
+
"LABEL_29": 29,
|
| 101 |
+
"LABEL_3": 3,
|
| 102 |
+
"LABEL_30": 30,
|
| 103 |
+
"LABEL_31": 31,
|
| 104 |
+
"LABEL_32": 32,
|
| 105 |
+
"LABEL_33": 33,
|
| 106 |
+
"LABEL_34": 34,
|
| 107 |
+
"LABEL_35": 35,
|
| 108 |
+
"LABEL_36": 36,
|
| 109 |
+
"LABEL_37": 37,
|
| 110 |
+
"LABEL_38": 38,
|
| 111 |
+
"LABEL_39": 39,
|
| 112 |
+
"LABEL_4": 4,
|
| 113 |
+
"LABEL_40": 40,
|
| 114 |
+
"LABEL_41": 41,
|
| 115 |
+
"LABEL_42": 42,
|
| 116 |
+
"LABEL_43": 43,
|
| 117 |
+
"LABEL_44": 44,
|
| 118 |
+
"LABEL_45": 45,
|
| 119 |
+
"LABEL_46": 46,
|
| 120 |
+
"LABEL_47": 47,
|
| 121 |
+
"LABEL_48": 48,
|
| 122 |
+
"LABEL_49": 49,
|
| 123 |
+
"LABEL_5": 5,
|
| 124 |
+
"LABEL_50": 50,
|
| 125 |
+
"LABEL_51": 51,
|
| 126 |
+
"LABEL_52": 52,
|
| 127 |
+
"LABEL_53": 53,
|
| 128 |
+
"LABEL_54": 54,
|
| 129 |
+
"LABEL_55": 55,
|
| 130 |
+
"LABEL_56": 56,
|
| 131 |
+
"LABEL_57": 57,
|
| 132 |
+
"LABEL_58": 58,
|
| 133 |
+
"LABEL_59": 59,
|
| 134 |
+
"LABEL_6": 6,
|
| 135 |
+
"LABEL_7": 7,
|
| 136 |
+
"LABEL_8": 8,
|
| 137 |
+
"LABEL_9": 9
|
| 138 |
+
},
|
| 139 |
+
"layer_norm_eps": 1e-05,
|
| 140 |
+
"max_position_embeddings": 514,
|
| 141 |
+
"model_type": "xlm-roberta",
|
| 142 |
+
"num_attention_heads": 12,
|
| 143 |
+
"num_hidden_layers": 6,
|
| 144 |
+
"output_past": true,
|
| 145 |
+
"pad_token_id": 1,
|
| 146 |
+
"position_embedding_type": "absolute",
|
| 147 |
+
"problem_type": "single_label_classification",
|
| 148 |
+
"torch_dtype": "float32",
|
| 149 |
+
"transformers_version": "4.33.3",
|
| 150 |
+
"type_vocab_size": 1,
|
| 151 |
+
"use_cache": true,
|
| 152 |
+
"vocab_size": 250002
|
| 153 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:37d3003635b4704dd77bf04c27f1243b1a9962487db9eb2ab8b71654a73074b9
|
| 3 |
+
size 429287542
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99284134701a1f4db59dc3bc1a3bba315919e7e9c091657c472b71b4467e92eb
|
| 3 |
+
size 4600
|