Commit
•
1fc48ce
1
Parent(s):
e074dc3
update model card README.md
Browse files
README.md
CHANGED
|
@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss:
|
| 20 |
-
- Accuracy: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -37,73 +37,25 @@ More information needed
|
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
- learning_rate: 5e-05
|
| 40 |
-
- train_batch_size:
|
| 41 |
-
- eval_batch_size:
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_ratio: 0.1
|
| 46 |
-
- training_steps:
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
| 50 |
-
| Training Loss | Epoch | Step
|
| 51 |
-
|
| 52 |
-
| 0.
|
| 53 |
-
| 0.
|
| 54 |
-
| 0.5243 | 2.02 | 912 | 1.1637 | 0.8026 |
|
| 55 |
-
| 0.0005 | 3.02 | 1216 | 0.8620 | 0.8092 |
|
| 56 |
-
| 0.5382 | 4.02 | 1520 | 0.9102 | 0.8092 |
|
| 57 |
-
| 0.0009 | 5.02 | 1824 | 1.2623 | 0.8355 |
|
| 58 |
-
| 0.0007 | 6.02 | 2128 | 1.4007 | 0.7829 |
|
| 59 |
-
| 0.254 | 7.02 | 2432 | 3.3258 | 0.4803 |
|
| 60 |
-
| 0.0005 | 8.02 | 2736 | 1.0090 | 0.8684 |
|
| 61 |
-
| 0.0003 | 9.02 | 3040 | 1.6322 | 0.7632 |
|
| 62 |
-
| 0.0015 | 10.02 | 3344 | 3.1927 | 0.5395 |
|
| 63 |
-
| 0.0006 | 11.02 | 3648 | 2.3243 | 0.7237 |
|
| 64 |
-
| 0.0004 | 12.02 | 3952 | 1.4877 | 0.7961 |
|
| 65 |
-
| 0.0007 | 13.02 | 4256 | 1.4014 | 0.8224 |
|
| 66 |
-
| 0.0001 | 14.02 | 4560 | 0.9946 | 0.8487 |
|
| 67 |
-
| 0.6249 | 15.02 | 4864 | 1.2847 | 0.7961 |
|
| 68 |
-
| 3.8326 | 16.02 | 5168 | 1.7870 | 0.7171 |
|
| 69 |
-
| 0.0646 | 17.02 | 5472 | 2.3504 | 0.6579 |
|
| 70 |
-
| 0.0003 | 18.02 | 5776 | 0.9367 | 0.8618 |
|
| 71 |
-
| 0.0004 | 19.02 | 6080 | 2.5710 | 0.6316 |
|
| 72 |
-
| 0.5626 | 20.02 | 6384 | 2.6711 | 0.6842 |
|
| 73 |
-
| 0.9002 | 21.02 | 6688 | 2.1456 | 0.7566 |
|
| 74 |
-
| 0.0002 | 22.02 | 6992 | 2.3488 | 0.7237 |
|
| 75 |
-
| 0.6977 | 23.02 | 7296 | 1.5013 | 0.8092 |
|
| 76 |
-
| 0.0001 | 24.02 | 7600 | 1.9442 | 0.7763 |
|
| 77 |
-
| 0.0003 | 25.02 | 7904 | 1.8732 | 0.8026 |
|
| 78 |
-
| 0.0001 | 26.02 | 8208 | 2.0295 | 0.7829 |
|
| 79 |
-
| 0.0001 | 27.02 | 8512 | 1.7623 | 0.8092 |
|
| 80 |
-
| 0.0001 | 28.02 | 8816 | 1.8035 | 0.8026 |
|
| 81 |
-
| 0.0 | 29.02 | 9120 | 1.7754 | 0.8092 |
|
| 82 |
-
| 0.0001 | 30.02 | 9424 | 1.7622 | 0.7961 |
|
| 83 |
-
| 0.0001 | 31.02 | 9728 | 1.7557 | 0.7895 |
|
| 84 |
-
| 0.0002 | 32.02 | 10032 | 1.5907 | 0.8224 |
|
| 85 |
-
| 0.0001 | 33.02 | 10336 | 1.6859 | 0.8158 |
|
| 86 |
-
| 0.0 | 34.02 | 10640 | 1.8641 | 0.7961 |
|
| 87 |
-
| 0.0 | 35.02 | 10944 | 1.7088 | 0.8224 |
|
| 88 |
-
| 0.0 | 36.02 | 11248 | 1.6140 | 0.8421 |
|
| 89 |
-
| 0.0 | 37.02 | 11552 | 1.6678 | 0.8355 |
|
| 90 |
-
| 0.0 | 38.02 | 11856 | 1.6991 | 0.8355 |
|
| 91 |
-
| 0.0 | 39.02 | 12160 | 1.7723 | 0.8224 |
|
| 92 |
-
| 0.0 | 40.02 | 12464 | 1.7865 | 0.8224 |
|
| 93 |
-
| 0.6067 | 41.02 | 12768 | 2.6848 | 0.7368 |
|
| 94 |
-
| 0.0001 | 42.02 | 13072 | 1.6834 | 0.8289 |
|
| 95 |
-
| 0.0 | 43.02 | 13376 | 1.7188 | 0.8289 |
|
| 96 |
-
| 0.9374 | 44.02 | 13680 | 1.5728 | 0.8421 |
|
| 97 |
-
| 0.0 | 45.02 | 13984 | 2.0988 | 0.7895 |
|
| 98 |
-
| 0.0 | 46.02 | 14288 | 2.0841 | 0.7829 |
|
| 99 |
-
| 0.0 | 47.02 | 14592 | 2.2198 | 0.7632 |
|
| 100 |
-
| 0.0 | 48.02 | 14896 | 2.2020 | 0.7632 |
|
| 101 |
-
| 0.0 | 49.02 | 15200 | 2.0693 | 0.7763 |
|
| 102 |
|
| 103 |
|
| 104 |
### Framework versions
|
| 105 |
|
| 106 |
- Transformers 4.26.1
|
| 107 |
-
- Pytorch 1.13.1+
|
| 108 |
- Datasets 2.10.1
|
| 109 |
- Tokenizers 0.13.2
|
|
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.7215
|
| 20 |
+
- Accuracy: 0.6429
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
- learning_rate: 5e-05
|
| 40 |
+
- train_batch_size: 4
|
| 41 |
+
- eval_batch_size: 4
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_ratio: 0.1
|
| 46 |
+
- training_steps: 152
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
| 50 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 51 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 52 |
+
| 0.5133 | 0.5 | 76 | 0.7872 | 0.6842 |
|
| 53 |
+
| 0.2188 | 1.5 | 152 | 0.7163 | 0.7434 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
|
| 56 |
### Framework versions
|
| 57 |
|
| 58 |
- Transformers 4.26.1
|
| 59 |
+
- Pytorch 1.13.1+cu116
|
| 60 |
- Datasets 2.10.1
|
| 61 |
- Tokenizers 0.13.2
|