File size: 1,810 Bytes
3432941
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: bsd-3-clause
base_model: MIT/ast-finetuned-speech-commands-v2
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: v25-ast-finetuned-speech-commands-v2-poisoned
  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. -->

# v25-ast-finetuned-speech-commands-v2-poisoned

This model is a fine-tuned version of [MIT/ast-finetuned-speech-commands-v2](https://huggingface.co/MIT/ast-finetuned-speech-commands-v2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1254
- Accuracy: 0.9836

## 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: 3e-05
- train_batch_size: 22
- eval_batch_size: 22
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 88
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log        | 0.86  | 3    | 6.5722          | 0.0066   |
| No log        | 2.0   | 7    | 1.7182          | 0.2533   |
| 5.5291        | 2.86  | 10   | 0.4478          | 0.9046   |
| 5.5291        | 4.0   | 14   | 0.1254          | 0.9836   |
| 5.5291        | 4.29  | 15   | 0.1138          | 0.9836   |


### Framework versions

- Transformers 4.37.0
- Pytorch 2.1.2
- Datasets 2.16.1
- Tokenizers 0.15.1