<!-- 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-8-batch-8000-vid-multiclass
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: 0.8423
- F1: 0.8057
- Accuracy: 0.8057
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: 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: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 60000
Training results
Training Loss | Epoch | Step | Accuracy | F1 | Validation Loss |
---|---|---|---|---|---|
1.201 | 0.1 | 1000 | 0.5063 | 0.5063 | 1.2871 |
0.7867 | 1.1 | 2000 | 0.6105 | 0.6105 | 1.1560 |
0.8387 | 0.1 | 3000 | 0.6695 | 0.6695 | 1.0269 |
0.7537 | 1.1 | 4000 | 0.7295 | 0.7295 | 0.9678 |
0.6444 | 2.1 | 5000 | 0.6853 | 0.6853 | 1.0220 |
0.7003 | 3.1 | 6000 | 0.7347 | 0.7347 | 0.9334 |
0.753 | 0.1 | 7000 | 0.7432 | 0.7432 | 0.9021 |
0.5269 | 1.1 | 8000 | 0.7789 | 0.7789 | 0.8279 |
0.6577 | 2.1 | 9000 | 0.7989 | 0.7989 | 0.7635 |
0.6222 | 3.1 | 10000 | 0.8126 | 0.8126 | 0.8054 |
0.8643 | 0.05 | 11000 | 0.7537 | 0.7537 | 0.9539 |
0.714 | 0.05 | 12000 | 0.7747 | 0.7747 | 0.8765 |
0.5181 | 1.05 | 13000 | 0.7726 | 0.7726 | 0.8558 |
0.4201 | 0.05 | 14000 | 0.8011 | 0.8011 | 0.8045 |
0.5374 | 1.05 | 15000 | 0.7884 | 0.7884 | 0.8136 |
0.4644 | 2.05 | 16000 | 0.7821 | 0.7821 | 0.8639 |
0.6452 | 3.05 | 17000 | 0.8179 | 0.8179 | 0.8085 |
0.4075 | 0.05 | 18000 | 0.7916 | 0.7916 | 0.8793 |
0.3501 | 1.05 | 19000 | 0.8221 | 0.8221 | 0.8139 |
0.4375 | 2.05 | 20000 | 0.8147 | 0.8147 | 0.8130 |
0.6524 | 0.03 | 21000 | 0.7537 | 0.7537 | 0.9271 |
0.5446 | 0.03 | 22000 | 0.78 | 0.78 | 0.8799 |
0.4624 | 0.03 | 23000 | 0.7884 | 0.7884 | 0.9210 |
0.466 | 1.02 | 24000 | 0.7842 | 0.7842 | 0.8601 |
0.6521 | 0.03 | 25000 | 0.7684 | 0.7684 | 0.9876 |
0.4043 | 0.03 | 26000 | 0.7789 | 0.7789 | 0.9454 |
0.5166 | 0.03 | 28000 | 0.7411 | 0.7411 | 1.1252 |
0.5039 | 1.03 | 30000 | 0.7537 | 0.7537 | 1.1474 |
0.6505 | 2.03 | 32000 | 0.7358 | 0.7358 | 1.0888 |
0.7491 | 3.03 | 34000 | 0.7747 | 0.7747 | 1.0014 |
0.8005 | 4.03 | 36000 | 0.7642 | 0.7642 | 1.0938 |
0.4648 | 0.03 | 38000 | 1.0662 | 0.7789 | 0.7789 |
0.2227 | 1.03 | 40000 | 1.2853 | 0.7537 | 0.7537 |
1.0841 | 2.03 | 42000 | 1.2443 | 0.7653 | 0.7653 |
0.514 | 3.03 | 44000 | 1.2017 | 0.78 | 0.78 |
0.4064 | 4.03 | 46000 | 1.3686 | 0.7547 | 0.7547 |
0.4566 | 5.03 | 48000 | 1.2185 | 0.7821 | 0.7821 |
0.9526 | 6.03 | 50000 | 1.2917 | 0.7947 | 0.7947 |
0.8275 | 7.03 | 52000 | 1.2091 | 0.8063 | 0.8063 |
0.3083 | 8.03 | 54000 | 1.2351 | 0.8021 | 0.8021 |
0.1902 | 9.03 | 56000 | 1.3653 | 0.7937 | 0.7937 |
0.2955 | 10.03 | 58000 | 1.3334 | 0.7958 | 0.7958 |
0.3169 | 11.03 | 60000 | 1.2957 | 0.8126 | 0.8126 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu118
- Datasets 2.14.5
- Tokenizers 0.14.1