--- language: - en license: mit tags: - generated_from_trainer datasets: - glue metrics: - accuracy model-index: - name: roberta-base-qnli results: - task: name: Text Classification type: text-classification dataset: name: GLUE QNLI type: glue args: qnli metrics: - name: Accuracy type: accuracy value: 0.9154310818231741 --- # roberta-base-qnli This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the GLUE QNLI dataset. It achieves the following results on the evaluation set: - Loss: 0.2330 - Accuracy: 0.9154 ## 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: 2e-05 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.06 - num_epochs: 10.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | 0.6919 | 0.08 | 500 | 0.6161 | 0.7501 | | 0.4801 | 0.15 | 1000 | 0.3524 | 0.8550 | | 0.4049 | 0.23 | 1500 | 0.3011 | 0.8742 | | 0.3827 | 0.31 | 2000 | 0.3125 | 0.8768 | | 0.3445 | 0.38 | 2500 | 0.2916 | 0.8924 | | 0.3567 | 0.46 | 3000 | 0.2662 | 0.8991 | | 0.3422 | 0.53 | 3500 | 0.2657 | 0.8980 | | 0.3257 | 0.61 | 4000 | 0.2830 | 0.9021 | | 0.3506 | 0.69 | 4500 | 0.2434 | 0.9063 | | 0.317 | 0.76 | 5000 | 0.2440 | 0.9052 | | 0.3152 | 0.84 | 5500 | 0.2786 | 0.9015 | | 0.2966 | 0.92 | 6000 | 0.2599 | 0.9083 | | 0.298 | 0.99 | 6500 | 0.2617 | 0.9070 | | 0.2634 | 1.07 | 7000 | 0.2330 | 0.9154 | | 0.2625 | 1.15 | 7500 | 0.2598 | 0.9109 | | 0.2596 | 1.22 | 8000 | 0.3616 | 0.9099 | | 0.2457 | 1.3 | 8500 | 0.2800 | 0.9096 | | 0.2545 | 1.37 | 9000 | 0.2960 | 0.9081 | | 0.2535 | 1.45 | 9500 | 0.2389 | 0.9114 | | 0.2639 | 1.53 | 10000 | 0.3343 | 0.8913 | | 0.2434 | 1.6 | 10500 | 0.2470 | 0.9116 | | 0.2613 | 1.68 | 11000 | 0.2949 | 0.9092 | | 0.2456 | 1.76 | 11500 | 0.2557 | 0.9163 | | 0.2483 | 1.83 | 12000 | 0.2462 | 0.9141 | | 0.2524 | 1.91 | 12500 | 0.2453 | 0.9114 | | 0.2467 | 1.99 | 13000 | 0.2611 | 0.9162 | | 0.2059 | 2.06 | 13500 | 0.3071 | 0.9158 | | 0.1968 | 2.14 | 14000 | 0.3205 | 0.9209 | | 0.1944 | 2.21 | 14500 | 0.3430 | 0.9145 | | 0.2065 | 2.29 | 15000 | 0.3388 | 0.9147 | | 0.1992 | 2.37 | 15500 | 0.2569 | 0.9158 | | 0.1994 | 2.44 | 16000 | 0.3349 | 0.9109 | | 0.2001 | 2.52 | 16500 | 0.2850 | 0.9096 | | 0.2014 | 2.6 | 17000 | 0.3214 | 0.9200 | | 0.2156 | 2.67 | 17500 | 0.3079 | 0.9134 | | 0.2036 | 2.75 | 18000 | 0.2739 | 0.9163 | | 0.2118 | 2.83 | 18500 | 0.2790 | 0.9185 | | 0.2167 | 2.9 | 19000 | 0.2699 | 0.9167 | | 0.2015 | 2.98 | 19500 | 0.2895 | 0.9189 | | 0.1649 | 3.05 | 20000 | 0.3719 | 0.9162 | | 0.1505 | 3.13 | 20500 | 0.3700 | 0.9132 | | 0.1509 | 3.21 | 21000 | 0.3721 | 0.9156 | | 0.1517 | 3.28 | 21500 | 0.3566 | 0.9154 | | 0.1583 | 3.36 | 22000 | 0.3975 | 0.9140 | | 0.1568 | 3.44 | 22500 | 0.4135 | 0.9136 | | 0.1642 | 3.51 | 23000 | 0.3705 | 0.9129 | | 0.1781 | 3.59 | 23500 | 0.3399 | 0.9156 | | 0.1725 | 3.67 | 24000 | 0.3165 | 0.9160 | | 0.1675 | 3.74 | 24500 | 0.3279 | 0.9180 | | 0.165 | 3.82 | 25000 | 0.3424 | 0.9202 | | 0.1608 | 3.89 | 25500 | 0.4022 | 0.9138 | | 0.1576 | 3.97 | 26000 | 0.3611 | 0.9147 | | 0.1382 | 4.05 | 26500 | 0.4001 | 0.9140 | | 0.1126 | 4.12 | 27000 | 0.4015 | 0.9169 | | 0.1048 | 4.2 | 27500 | 0.3919 | 0.9169 | | 0.1057 | 4.28 | 28000 | 0.4072 | 0.9176 | | 0.1212 | 4.35 | 28500 | 0.3623 | 0.9162 | | 0.1152 | 4.43 | 29000 | 0.3946 | 0.9149 | | 0.125 | 4.51 | 29500 | 0.4142 | 0.9156 | | 0.1195 | 4.58 | 30000 | 0.4095 | 0.9151 | | 0.1139 | 4.66 | 30500 | 0.4586 | 0.9088 | | 0.1279 | 4.73 | 31000 | 0.3900 | 0.9204 | | 0.1306 | 4.81 | 31500 | 0.3741 | 0.9165 | | 0.1091 | 4.89 | 32000 | 0.4296 | 0.9207 | | 0.1272 | 4.96 | 32500 | 0.3724 | 0.9189 | | 0.0906 | 5.04 | 33000 | 0.4512 | 0.9182 | | 0.0915 | 5.12 | 33500 | 0.4160 | 0.9220 | | 0.0773 | 5.19 | 34000 | 0.4743 | 0.9180 | | 0.0861 | 5.27 | 34500 | 0.5024 | 0.9204 | | 0.0729 | 5.35 | 35000 | 0.4282 | 0.9204 | | 0.0901 | 5.42 | 35500 | 0.4612 | 0.9226 | | 0.0856 | 5.5 | 36000 | 0.4495 | 0.9180 | | 0.0839 | 5.58 | 36500 | 0.4501 | 0.9206 | | 0.0874 | 5.65 | 37000 | 0.4136 | 0.9200 | | 0.0944 | 5.73 | 37500 | 0.4629 | 0.9165 | | 0.0874 | 5.8 | 38000 | 0.4790 | 0.9160 | | 0.0859 | 5.88 | 38500 | 0.4725 | 0.9132 | | 0.0808 | 5.96 | 39000 | 0.4613 | 0.9162 | | 0.0723 | 6.03 | 39500 | 0.4816 | 0.9195 | | 0.0568 | 6.11 | 40000 | 0.5257 | 0.9187 | | 0.0628 | 6.19 | 40500 | 0.4516 | 0.9195 | | 0.053 | 6.26 | 41000 | 0.4929 | 0.9187 | | 0.0574 | 6.34 | 41500 | 0.4888 | 0.9191 | | 0.0717 | 6.42 | 42000 | 0.4769 | 0.9165 | | 0.0622 | 6.49 | 42500 | 0.5082 | 0.9184 | | 0.0593 | 6.57 | 43000 | 0.4460 | 0.9211 | | 0.0603 | 6.64 | 43500 | 0.4345 | 0.9206 | | 0.0659 | 6.72 | 44000 | 0.4423 | 0.9189 | | 0.0629 | 6.8 | 44500 | 0.4771 | 0.9191 | | 0.058 | 6.87 | 45000 | 0.4589 | 0.9228 | | 0.0545 | 6.95 | 45500 | 0.5084 | 0.9200 | | 0.0465 | 7.03 | 46000 | 0.5422 | 0.9193 | | 0.0424 | 7.1 | 46500 | 0.5030 | 0.9202 | | 0.0317 | 7.18 | 47000 | 0.5393 | 0.9213 | | 0.029 | 7.26 | 47500 | 0.5618 | 0.9174 | | 0.0439 | 7.33 | 48000 | 0.5000 | 0.9195 | | 0.0347 | 7.41 | 48500 | 0.5093 | 0.9200 | | 0.0425 | 7.48 | 49000 | 0.5311 | 0.9174 | | 0.0384 | 7.56 | 49500 | 0.5010 | 0.9198 | | 0.039 | 7.64 | 50000 | 0.5182 | 0.9209 | | 0.04 | 7.71 | 50500 | 0.5238 | 0.9215 | | 0.0374 | 7.79 | 51000 | 0.5561 | 0.9218 | | 0.0366 | 7.87 | 51500 | 0.5412 | 0.9200 | | 0.036 | 7.94 | 52000 | 0.5213 | 0.9213 | | 0.0348 | 8.02 | 52500 | 0.5140 | 0.9217 | | 0.0186 | 8.1 | 53000 | 0.5693 | 0.9240 | | 0.0275 | 8.17 | 53500 | 0.5007 | 0.9239 | | 0.0219 | 8.25 | 54000 | 0.5400 | 0.9240 | | 0.0238 | 8.32 | 54500 | 0.5537 | 0.9228 | | 0.0201 | 8.4 | 55000 | 0.5851 | 0.9215 | | 0.0253 | 8.48 | 55500 | 0.5654 | 0.9217 | | 0.0243 | 8.55 | 56000 | 0.5833 | 0.9213 | | 0.0298 | 8.63 | 56500 | 0.5483 | 0.9209 | | 0.0232 | 8.71 | 57000 | 0.5724 | 0.9215 | | 0.0239 | 8.78 | 57500 | 0.5574 | 0.9195 | | 0.0263 | 8.86 | 58000 | 0.5491 | 0.9235 | | 0.0333 | 8.94 | 58500 | 0.5322 | 0.9209 | | 0.0259 | 9.01 | 59000 | 0.5493 | 0.9217 | | 0.0197 | 9.09 | 59500 | 0.5671 | 0.9209 | | 0.0237 | 9.16 | 60000 | 0.5536 | 0.9209 | | 0.022 | 9.24 | 60500 | 0.5523 | 0.9217 | | 0.0246 | 9.32 | 61000 | 0.5619 | 0.9220 | | 0.0202 | 9.39 | 61500 | 0.5619 | 0.9228 | | 0.0184 | 9.47 | 62000 | 0.5729 | 0.9217 | | 0.0122 | 9.55 | 62500 | 0.5946 | 0.9202 | | 0.015 | 9.62 | 63000 | 0.6014 | 0.9215 | | 0.0189 | 9.7 | 63500 | 0.5928 | 0.9226 | | 0.0194 | 9.78 | 64000 | 0.5898 | 0.9220 | | 0.0219 | 9.85 | 64500 | 0.5851 | 0.9218 | | 0.017 | 9.93 | 65000 | 0.5891 | 0.9218 | ### Framework versions - Transformers 4.21.3 - Pytorch 1.7.1 - Datasets 1.18.3 - Tokenizers 0.11.6