File size: 2,446 Bytes
0c31efb
 
ef9bf2f
0c31efb
 
 
 
 
 
ef9bf2f
0c31efb
 
 
 
 
 
 
 
 
 
ef9bf2f
0c31efb
580c3e9
 
 
 
 
0c31efb
 
 
 
 
 
 
 
 
 
 
ef9bf2f
0c31efb
 
 
 
 
 
52b00b8
ef9bf2f
 
0c31efb
 
 
58d538a
0c31efb
 
 
ef9bf2f
 
580c3e9
 
 
 
 
 
 
 
 
 
0c31efb
 
 
 
ef9bf2f
 
580c3e9
ef9bf2f
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
---
license: apache-2.0
base_model: distilbert-base-cased
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: distilbert-finetuned-gesture-prediction-21-classes
  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. -->

# distilbert-finetuned-gesture-prediction-21-classes

This model is a fine-tuned version of [distilbert-base-cased](https://huggingface.co/distilbert-base-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0324
- Precision: 0.8139
- Recall: 0.8139
- F1: 0.8139
- Accuracy: 0.8022

## 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: 9.96098704459956e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 2.1035        | 1.0   | 26   | 1.2829          | 0.7042    | 0.7042 | 0.7042 | 0.6806   |
| 0.9959        | 2.0   | 52   | 0.9482          | 0.7756    | 0.7756 | 0.7756 | 0.7621   |
| 0.6164        | 3.0   | 78   | 0.8716          | 0.7849    | 0.7849 | 0.7849 | 0.7685   |
| 0.3812        | 4.0   | 104  | 0.8710          | 0.8004    | 0.8004 | 0.8004 | 0.7867   |
| 0.2325        | 5.0   | 130  | 0.9558          | 0.7916    | 0.7916 | 0.7916 | 0.7788   |
| 0.1558        | 6.0   | 156  | 0.9310          | 0.8077    | 0.8077 | 0.8077 | 0.7949   |
| 0.0983        | 7.0   | 182  | 0.9989          | 0.8121    | 0.8121 | 0.8121 | 0.7992   |
| 0.0697        | 8.0   | 208  | 1.0241          | 0.8083    | 0.8083 | 0.8083 | 0.7963   |
| 0.05          | 9.0   | 234  | 1.0352          | 0.8110    | 0.8110 | 0.8110 | 0.7991   |
| 0.0403        | 10.0  | 260  | 1.0324          | 0.8139    | 0.8139 | 0.8139 | 0.8022   |


### Framework versions

- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1