--- tags: - generated_from_trainer metrics: - accuracy model-index: - name: dit-tiny_tobacco3482_kd_CEKD_t1.5_a0.5 results: [] --- # dit-tiny_tobacco3482_kd_CEKD_t1.5_a0.5 This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.9246 - Accuracy: 0.18 - Brier Loss: 0.8755 - Nll: 6.7967 - F1 Micro: 0.18 - F1 Macro: 0.0306 - Ece: 0.2497 - Aurc: 0.8499 ## 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: 16 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 256 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 25 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc | |:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:| | No log | 0.96 | 3 | 3.1239 | 0.145 | 0.8999 | 10.1580 | 0.145 | 0.0253 | 0.2222 | 0.8467 | | No log | 1.96 | 6 | 3.0895 | 0.145 | 0.8946 | 10.5934 | 0.145 | 0.0253 | 0.2303 | 0.8470 | | No log | 2.96 | 9 | 3.0385 | 0.165 | 0.8866 | 8.6307 | 0.165 | 0.0502 | 0.2200 | 0.8458 | | No log | 3.96 | 12 | 2.9972 | 0.21 | 0.8806 | 6.5449 | 0.2100 | 0.0615 | 0.2512 | 0.8364 | | No log | 4.96 | 15 | 2.9719 | 0.155 | 0.8776 | 6.7565 | 0.155 | 0.0271 | 0.2414 | 0.8884 | | No log | 5.96 | 18 | 2.9579 | 0.215 | 0.8768 | 7.0870 | 0.2150 | 0.0643 | 0.2713 | 0.8778 | | No log | 6.96 | 21 | 2.9485 | 0.18 | 0.8768 | 7.0291 | 0.18 | 0.0308 | 0.2482 | 0.8532 | | No log | 7.96 | 24 | 2.9417 | 0.18 | 0.8770 | 6.9706 | 0.18 | 0.0306 | 0.2559 | 0.8525 | | No log | 8.96 | 27 | 2.9360 | 0.18 | 0.8768 | 6.9349 | 0.18 | 0.0306 | 0.2498 | 0.8527 | | No log | 9.96 | 30 | 2.9326 | 0.18 | 0.8767 | 6.9268 | 0.18 | 0.0306 | 0.2635 | 0.8533 | | No log | 10.96 | 33 | 2.9303 | 0.18 | 0.8765 | 6.9226 | 0.18 | 0.0306 | 0.2637 | 0.8531 | | No log | 11.96 | 36 | 2.9289 | 0.18 | 0.8764 | 6.9217 | 0.18 | 0.0306 | 0.2591 | 0.8524 | | No log | 12.96 | 39 | 2.9279 | 0.18 | 0.8762 | 6.8547 | 0.18 | 0.0306 | 0.2505 | 0.8526 | | No log | 13.96 | 42 | 2.9270 | 0.18 | 0.8760 | 6.8491 | 0.18 | 0.0306 | 0.2500 | 0.8520 | | No log | 14.96 | 45 | 2.9263 | 0.18 | 0.8759 | 6.8471 | 0.18 | 0.0306 | 0.2463 | 0.8518 | | No log | 15.96 | 48 | 2.9258 | 0.18 | 0.8758 | 6.8445 | 0.18 | 0.0306 | 0.2462 | 0.8520 | | No log | 16.96 | 51 | 2.9255 | 0.18 | 0.8758 | 6.8452 | 0.18 | 0.0306 | 0.2587 | 0.8511 | | No log | 17.96 | 54 | 2.9256 | 0.18 | 0.8758 | 6.7940 | 0.18 | 0.0306 | 0.2585 | 0.8513 | | No log | 18.96 | 57 | 2.9256 | 0.18 | 0.8758 | 6.7930 | 0.18 | 0.0306 | 0.2625 | 0.8508 | | No log | 19.96 | 60 | 2.9252 | 0.18 | 0.8757 | 6.7945 | 0.18 | 0.0306 | 0.2580 | 0.8506 | | No log | 20.96 | 63 | 2.9250 | 0.18 | 0.8756 | 6.7999 | 0.18 | 0.0306 | 0.2539 | 0.8505 | | No log | 21.96 | 66 | 2.9248 | 0.18 | 0.8756 | 6.8441 | 0.18 | 0.0306 | 0.2538 | 0.8502 | | No log | 22.96 | 69 | 2.9247 | 0.18 | 0.8755 | 6.8439 | 0.18 | 0.0306 | 0.2497 | 0.8500 | | No log | 23.96 | 72 | 2.9247 | 0.18 | 0.8755 | 6.7977 | 0.18 | 0.0306 | 0.2497 | 0.8500 | | No log | 24.96 | 75 | 2.9246 | 0.18 | 0.8755 | 6.7967 | 0.18 | 0.0306 | 0.2497 | 0.8499 | ### Framework versions - Transformers 4.26.1 - Pytorch 1.13.1.post200 - Datasets 2.9.0 - Tokenizers 0.13.2