<!-- 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. -->
20230830102607
This model is a fine-tuned version of bert-large-cased on the super_glue dataset. It achieves the following results on the evaluation set:
- Loss: 0.9380
- Accuracy: 0.6317
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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 11
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 80.0
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
No log | 1.0 | 340 | 0.6741 | 0.5361 |
0.6968 | 2.0 | 680 | 0.6803 | 0.5 |
0.6934 | 3.0 | 1020 | 0.6837 | 0.5 |
0.6934 | 4.0 | 1360 | 0.6453 | 0.5752 |
0.67 | 5.0 | 1700 | 0.6319 | 0.6379 |
0.6229 | 6.0 | 2040 | 0.6657 | 0.6223 |
0.6229 | 7.0 | 2380 | 0.6293 | 0.6693 |
0.585 | 8.0 | 2720 | 0.6188 | 0.6771 |
0.5622 | 9.0 | 3060 | 0.7399 | 0.6379 |
0.5622 | 10.0 | 3400 | 0.7147 | 0.6176 |
0.5341 | 11.0 | 3740 | 0.6935 | 0.6552 |
0.5113 | 12.0 | 4080 | 0.6084 | 0.6677 |
0.5113 | 13.0 | 4420 | 0.6515 | 0.6818 |
0.4915 | 14.0 | 4760 | 0.8644 | 0.6176 |
0.4663 | 15.0 | 5100 | 0.6619 | 0.6567 |
0.4663 | 16.0 | 5440 | 0.7789 | 0.6505 |
0.4606 | 17.0 | 5780 | 0.7261 | 0.6489 |
0.4483 | 18.0 | 6120 | 0.7978 | 0.6442 |
0.4483 | 19.0 | 6460 | 0.6810 | 0.6442 |
0.4332 | 20.0 | 6800 | 0.8449 | 0.6505 |
0.4219 | 21.0 | 7140 | 0.8289 | 0.6379 |
0.4219 | 22.0 | 7480 | 0.7733 | 0.6614 |
0.4134 | 23.0 | 7820 | 0.7533 | 0.6536 |
0.399 | 24.0 | 8160 | 0.7560 | 0.6395 |
0.4001 | 25.0 | 8500 | 0.8408 | 0.6348 |
0.4001 | 26.0 | 8840 | 0.7573 | 0.6599 |
0.3922 | 27.0 | 9180 | 0.7875 | 0.6113 |
0.3847 | 28.0 | 9520 | 0.7636 | 0.6442 |
0.3847 | 29.0 | 9860 | 0.8088 | 0.6473 |
0.3735 | 30.0 | 10200 | 0.7499 | 0.6505 |
0.3754 | 31.0 | 10540 | 0.7845 | 0.6411 |
0.3754 | 32.0 | 10880 | 0.8677 | 0.6364 |
0.3676 | 33.0 | 11220 | 0.7861 | 0.6411 |
0.3607 | 34.0 | 11560 | 0.7834 | 0.6332 |
0.3607 | 35.0 | 11900 | 0.9127 | 0.6426 |
0.3596 | 36.0 | 12240 | 0.8042 | 0.6395 |
0.3557 | 37.0 | 12580 | 0.8923 | 0.6379 |
0.3557 | 38.0 | 12920 | 0.7978 | 0.6379 |
0.3478 | 39.0 | 13260 | 0.8106 | 0.6458 |
0.3467 | 40.0 | 13600 | 0.8699 | 0.6301 |
0.3467 | 41.0 | 13940 | 0.8522 | 0.6426 |
0.3408 | 42.0 | 14280 | 0.8836 | 0.6317 |
0.3373 | 43.0 | 14620 | 0.8808 | 0.6238 |
0.3373 | 44.0 | 14960 | 0.9437 | 0.6301 |
0.3308 | 45.0 | 15300 | 0.8724 | 0.6317 |
0.3282 | 46.0 | 15640 | 0.9243 | 0.6364 |
0.3282 | 47.0 | 15980 | 0.8446 | 0.6364 |
0.32 | 48.0 | 16320 | 0.8618 | 0.6238 |
0.3224 | 49.0 | 16660 | 0.8313 | 0.6270 |
0.317 | 50.0 | 17000 | 0.8845 | 0.6348 |
0.317 | 51.0 | 17340 | 0.9127 | 0.6442 |
0.3199 | 52.0 | 17680 | 0.9284 | 0.6395 |
0.3158 | 53.0 | 18020 | 0.9305 | 0.6379 |
0.3158 | 54.0 | 18360 | 0.8782 | 0.6379 |
0.3048 | 55.0 | 18700 | 0.8722 | 0.6348 |
0.3086 | 56.0 | 19040 | 0.8814 | 0.6411 |
0.3086 | 57.0 | 19380 | 0.8981 | 0.6364 |
0.3074 | 58.0 | 19720 | 0.8604 | 0.6238 |
0.3012 | 59.0 | 20060 | 0.8734 | 0.6301 |
0.3012 | 60.0 | 20400 | 0.9171 | 0.6332 |
0.299 | 61.0 | 20740 | 0.9065 | 0.6317 |
0.296 | 62.0 | 21080 | 0.9070 | 0.6332 |
0.296 | 63.0 | 21420 | 0.9196 | 0.6254 |
0.2954 | 64.0 | 21760 | 0.9671 | 0.6364 |
0.2964 | 65.0 | 22100 | 0.9451 | 0.6348 |
0.2964 | 66.0 | 22440 | 0.8934 | 0.6411 |
0.2825 | 67.0 | 22780 | 0.9219 | 0.6426 |
0.2997 | 68.0 | 23120 | 0.9352 | 0.6332 |
0.2997 | 69.0 | 23460 | 0.9450 | 0.6332 |
0.29 | 70.0 | 23800 | 0.9625 | 0.6301 |
0.2811 | 71.0 | 24140 | 0.9098 | 0.6379 |
0.2811 | 72.0 | 24480 | 0.8967 | 0.6285 |
0.2825 | 73.0 | 24820 | 0.9234 | 0.6317 |
0.2868 | 74.0 | 25160 | 0.9275 | 0.6364 |
0.2846 | 75.0 | 25500 | 0.9291 | 0.6364 |
0.2846 | 76.0 | 25840 | 0.9158 | 0.6379 |
0.2811 | 77.0 | 26180 | 0.9310 | 0.6348 |
0.2754 | 78.0 | 26520 | 0.9423 | 0.6317 |
0.2754 | 79.0 | 26860 | 0.9350 | 0.6332 |
0.2776 | 80.0 | 27200 | 0.9380 | 0.6317 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3