File size: 2,873 Bytes
ecfb937
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
---
library_name: transformers
license: cc-by-nc-4.0
base_model: MCG-NJU/videomae-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: videomae-base-finetuned-kisa
  results: []
---

<!-- 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-base-finetuned-kisa

This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.2647
- Accuracy: 0.4913

## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.05
- training_steps: 2725

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Accuracy |
|:-------------:|:-------:|:----:|:---------------:|:--------:|
| 0.0741        | 0.0404  | 110  | 1.1353          | 0.5      |
| 0.0259        | 1.0404  | 220  | 3.7142          | 0.1183   |
| 0.5784        | 2.0404  | 330  | 2.2692          | 0.5      |
| 0.1384        | 3.0404  | 440  | 1.3726          | 0.5178   |
| 0.513         | 4.0404  | 550  | 2.5340          | 0.3728   |
| 0.0156        | 5.0404  | 660  | 2.3487          | 0.2041   |
| 0.0033        | 6.0404  | 770  | 4.4601          | 0.1953   |
| 0.0071        | 7.0404  | 880  | 4.6045          | 0.0917   |
| 0.004         | 8.0404  | 990  | 3.4062          | 0.4083   |
| 0.0017        | 9.0404  | 1100 | 2.4961          | 0.4941   |
| 0.4934        | 10.0404 | 1210 | 2.9785          | 0.4941   |
| 0.43          | 11.0404 | 1320 | 3.7030          | 0.5207   |
| 0.0014        | 12.0404 | 1430 | 2.5479          | 0.2012   |
| 0.0021        | 13.0404 | 1540 | 4.0235          | 0.3195   |
| 0.2387        | 14.0404 | 1650 | 4.6049          | 0.2337   |
| 0.0009        | 15.0404 | 1760 | 4.3070          | 0.2485   |
| 0.0004        | 16.0404 | 1870 | 4.4573          | 0.2515   |
| 0.5939        | 17.0404 | 1980 | 4.3423          | 0.3550   |
| 0.0013        | 18.0404 | 2090 | 4.3365          | 0.3047   |
| 0.0015        | 19.0404 | 2200 | 4.0964          | 0.2426   |
| 0.0032        | 20.0404 | 2310 | 4.1795          | 0.2988   |
| 0.0006        | 21.0404 | 2420 | 4.1612          | 0.3136   |


### Framework versions

- Transformers 4.48.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0