File size: 5,221 Bytes
42e474d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
---
license: mit
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: pos_final_mono_fr
  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. -->

# pos_final_mono_fr

This model is a fine-tuned version of [almanach/camembert-base](https://huggingface.co/almanach/camembert-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5416
- Precision: 0.9742
- Recall: 0.9745
- F1: 0.9743
- Accuracy: 0.9768

## 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: 256
- eval_batch_size: 256
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 1024
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 40.0
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log        | 0.95  | 14   | 3.6697          | 0.0210    | 0.0194 | 0.0201 | 0.0215   |
| No log        | 1.95  | 28   | 3.6329          | 0.0513    | 0.0484 | 0.0498 | 0.0511   |
| No log        | 2.95  | 42   | 3.5739          | 0.1142    | 0.1086 | 0.1113 | 0.1267   |
| No log        | 3.95  | 56   | 3.4791          | 0.2535    | 0.1976 | 0.2221 | 0.3061   |
| No log        | 4.95  | 70   | 3.3377          | 0.3393    | 0.2029 | 0.2539 | 0.3788   |
| No log        | 5.95  | 84   | 3.1886          | 0.3737    | 0.1401 | 0.2038 | 0.3427   |
| No log        | 6.95  | 98   | 3.0505          | 0.4342    | 0.3211 | 0.3692 | 0.4600   |
| No log        | 7.95  | 112  | 2.8996          | 0.5160    | 0.4319 | 0.4702 | 0.5282   |
| No log        | 8.95  | 126  | 2.7485          | 0.5617    | 0.4878 | 0.5222 | 0.5732   |
| No log        | 9.95  | 140  | 2.5862          | 0.6077    | 0.5374 | 0.5704 | 0.6246   |
| No log        | 10.95 | 154  | 2.4205          | 0.6805    | 0.6311 | 0.6549 | 0.6887   |
| No log        | 11.95 | 168  | 2.2603          | 0.7816    | 0.7569 | 0.7691 | 0.7839   |
| No log        | 12.95 | 182  | 2.1124          | 0.8366    | 0.8305 | 0.8335 | 0.8370   |
| No log        | 13.95 | 196  | 1.9826          | 0.8691    | 0.8681 | 0.8686 | 0.8736   |
| No log        | 14.95 | 210  | 1.8721          | 0.9210    | 0.92   | 0.9205 | 0.9240   |
| No log        | 15.95 | 224  | 1.7779          | 0.9390    | 0.9392 | 0.9391 | 0.9417   |
| No log        | 16.95 | 238  | 1.6986          | 0.9442    | 0.9452 | 0.9447 | 0.9466   |
| No log        | 17.95 | 252  | 1.6294          | 0.9467    | 0.9476 | 0.9472 | 0.9486   |
| No log        | 18.95 | 266  | 1.5667          | 0.9481    | 0.9493 | 0.9487 | 0.9499   |
| No log        | 19.95 | 280  | 1.5073          | 0.9507    | 0.9522 | 0.9514 | 0.9523   |
| No log        | 20.95 | 294  | 1.4499          | 0.9538    | 0.9550 | 0.9544 | 0.9552   |
| No log        | 21.95 | 308  | 1.3926          | 0.9555    | 0.9563 | 0.9559 | 0.9563   |
| No log        | 22.95 | 322  | 1.3373          | 0.9609    | 0.9614 | 0.9612 | 0.9612   |
| No log        | 23.95 | 336  | 1.2815          | 0.9622    | 0.9624 | 0.9623 | 0.9623   |
| No log        | 24.95 | 350  | 1.2246          | 0.9649    | 0.9648 | 0.9648 | 0.9646   |
| No log        | 25.95 | 364  | 1.1682          | 0.9653    | 0.9652 | 0.9652 | 0.9648   |
| No log        | 26.95 | 378  | 1.1114          | 0.9650    | 0.9659 | 0.9654 | 0.9661   |
| No log        | 27.95 | 392  | 1.0521          | 0.9669    | 0.9675 | 0.9672 | 0.9699   |
| No log        | 28.95 | 406  | 0.9950          | 0.9677    | 0.9679 | 0.9678 | 0.9707   |
| No log        | 29.95 | 420  | 0.9364          | 0.9687    | 0.9690 | 0.9688 | 0.9716   |
| No log        | 30.95 | 434  | 0.8800          | 0.9691    | 0.9693 | 0.9692 | 0.9721   |
| No log        | 31.95 | 448  | 0.8233          | 0.9693    | 0.9698 | 0.9696 | 0.9726   |
| No log        | 32.95 | 462  | 0.7679          | 0.9703    | 0.9703 | 0.9703 | 0.9733   |
| No log        | 33.95 | 476  | 0.7146          | 0.9711    | 0.9711 | 0.9711 | 0.9737   |
| No log        | 34.95 | 490  | 0.6641          | 0.9722    | 0.9724 | 0.9723 | 0.9750   |
| 2.0937        | 35.95 | 504  | 0.6187          | 0.9729    | 0.9729 | 0.9729 | 0.9755   |
| 2.0937        | 36.95 | 518  | 0.5834          | 0.9727    | 0.9732 | 0.9729 | 0.9756   |
| 2.0937        | 37.95 | 532  | 0.5605          | 0.9735    | 0.9739 | 0.9737 | 0.9762   |
| 2.0937        | 38.95 | 546  | 0.5466          | 0.9737    | 0.9742 | 0.9739 | 0.9765   |
| 2.0937        | 39.95 | 560  | 0.5416          | 0.9742    | 0.9745 | 0.9743 | 0.9768   |


### Framework versions

- Transformers 4.25.1
- Pytorch 1.12.0
- Datasets 2.18.0
- Tokenizers 0.13.2