<!-- 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. -->
videomae-finetuned-nba-5-class-4-batch-8000-vid-multiclass_3
This model is a fine-tuned version of MCG-NJU/videomae-base-finetuned-kinetics on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.3705
- F1: 0.7835
- Accuracy: 0.7835
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: 1.5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 100000
Training results
| Training Loss | Epoch | Step | Accuracy | F1 | Validation Loss |
|---|---|---|---|---|---|
| 1.4333 | 0.02 | 2000 | 0.5168 | 0.5168 | 1.2521 |
| 0.7414 | 1.02 | 4000 | 0.6337 | 0.6337 | 1.0825 |
| 1.1485 | 2.02 | 6000 | 0.5853 | 0.5853 | 1.2329 |
| 0.8216 | 3.02 | 8000 | 0.6663 | 0.6663 | 1.0844 |
| 0.9047 | 4.02 | 10000 | 0.6916 | 0.6916 | 1.0024 |
| 0.9979 | 5.02 | 12000 | 0.6716 | 0.6716 | 1.0889 |
| 0.6139 | 6.02 | 14000 | 0.7274 | 0.7274 | 0.9464 |
| 0.9932 | 7.02 | 16000 | 0.7495 | 0.7495 | 0.9489 |
| 0.972 | 8.02 | 18000 | 0.7368 | 0.7368 | 0.9903 |
| 0.8964 | 9.02 | 20000 | 0.7453 | 0.7453 | 0.9291 |
| 0.7686 | 10.02 | 22000 | 0.7042 | 0.7042 | 1.0472 |
| 0.6966 | 11.02 | 24000 | 0.7337 | 0.7337 | 1.0546 |
| 0.6768 | 12.02 | 26000 | 0.7653 | 0.7653 | 0.9781 |
| 0.7724 | 13.02 | 28000 | 0.7295 | 0.7295 | 1.0383 |
| 0.6458 | 14.02 | 30000 | 0.7516 | 0.7516 | 1.0281 |
| 0.4914 | 15.02 | 32000 | 0.7789 | 0.7789 | 0.9133 |
| 0.4934 | 16.02 | 34000 | 0.7389 | 0.7389 | 1.0214 |
| 0.7652 | 17.02 | 36000 | 0.7589 | 0.7589 | 0.9891 |
| 0.8866 | 18.02 | 38000 | 0.7705 | 0.7705 | 1.0098 |
| 0.8396 | 19.02 | 40000 | 0.7642 | 0.7642 | 1.0574 |
| 0.3249 | 20.02 | 42000 | 0.7811 | 0.7811 | 0.8901 |
| 0.7265 | 21.02 | 44000 | 0.7674 | 0.7674 | 1.0387 |
| 0.634 | 22.02 | 46000 | 0.7642 | 0.7642 | 1.1140 |
| 0.6238 | 23.02 | 48000 | 0.7811 | 0.7811 | 1.0951 |
| 0.3984 | 24.02 | 50000 | 0.7947 | 0.7947 | 1.0359 |
| 0.7153 | 25.02 | 52000 | 0.7642 | 0.7642 | 1.2042 |
| 0.5881 | 26.02 | 54000 | 0.7821 | 0.7821 | 1.1202 |
| 1.2165 | 0.02 | 56000 | 1.1561 | 0.7663 | 0.7663 |
| 0.3302 | 1.02 | 58000 | 1.1171 | 0.7947 | 0.7947 |
| 1.1086 | 2.02 | 60000 | 1.1198 | 0.8095 | 0.8095 |
| 0.3774 | 3.02 | 62000 | 1.1968 | 0.7905 | 0.7905 |
| 0.5621 | 4.02 | 64000 | 1.3453 | 0.78 | 0.78 |
| 0.4457 | 5.02 | 66000 | 1.2562 | 0.7947 | 0.7947 |
| 0.6879 | 6.02 | 68000 | 1.2735 | 0.7811 | 0.7811 |
| 0.8734 | 7.02 | 70000 | 1.2543 | 0.8137 | 0.8137 |
| 0.3203 | 8.02 | 72000 | 1.2536 | 0.8137 | 0.8137 |
| 0.3381 | 9.02 | 74000 | 1.4249 | 0.7621 | 0.7621 |
| 0.4181 | 10.02 | 76000 | 1.4163 | 0.78 | 0.78 |
| 0.339 | 11.02 | 78000 | 1.4407 | 0.7916 | 0.7916 |
| 0.1726 | 12.02 | 80000 | 1.4049 | 0.7968 | 0.7968 |
| 0.4499 | 13.02 | 82000 | 1.4951 | 0.7905 | 0.7905 |
| 0.4024 | 14.02 | 84000 | 1.5575 | 0.7853 | 0.7853 |
| 0.2009 | 15.02 | 86000 | 1.4637 | 0.8042 | 0.8042 |
| 0.2261 | 16.02 | 88000 | 1.5615 | 0.8000 | 0.8 |
| 0.5715 | 17.02 | 90000 | 1.4886 | 0.7979 | 0.7979 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1
