File size: 5,364 Bytes
aa5b5e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
110
111
112
---
license: apache-2.0
base_model: google/flan-t5-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: flan_t5_base_patent
  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. -->

# flan_t5_base_patent

This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9077
- Accuracy: 0.6922
- F1 Macro: 0.6251
- F1 Micro: 0.6922

## 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: 0.0005
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Micro |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:--------:|
| 1.3252        | 0.06  | 50   | 1.3453          | 0.5512   | 0.4078   | 0.5512   |
| 1.1367        | 0.13  | 100  | 1.1291          | 0.613    | 0.4903   | 0.613    |
| 1.1672        | 0.19  | 150  | 1.0973          | 0.6288   | 0.5119   | 0.6288   |
| 1.0094        | 0.26  | 200  | 1.0517          | 0.6368   | 0.5609   | 0.6368   |
| 0.9798        | 0.32  | 250  | 1.0432          | 0.6466   | 0.5584   | 0.6466   |
| 1.1403        | 0.38  | 300  | 1.0146          | 0.6542   | 0.5733   | 0.6542   |
| 1.0617        | 0.45  | 350  | 1.0112          | 0.652    | 0.5857   | 0.652    |
| 0.9333        | 0.51  | 400  | 1.0339          | 0.6436   | 0.5698   | 0.6436   |
| 1.0316        | 0.58  | 450  | 0.9973          | 0.6622   | 0.5891   | 0.6622   |
| 0.9675        | 0.64  | 500  | 0.9732          | 0.6698   | 0.6110   | 0.6698   |
| 0.9452        | 0.7   | 550  | 0.9900          | 0.6652   | 0.5890   | 0.6652   |
| 1.0378        | 0.77  | 600  | 0.9751          | 0.668    | 0.6075   | 0.668    |
| 1.0038        | 0.83  | 650  | 0.9483          | 0.67     | 0.6076   | 0.67     |
| 0.9509        | 0.9   | 700  | 0.9739          | 0.6614   | 0.6000   | 0.6614   |
| 0.9378        | 0.96  | 750  | 0.9684          | 0.6686   | 0.5962   | 0.6686   |
| 0.801         | 1.02  | 800  | 0.9589          | 0.6696   | 0.6145   | 0.6696   |
| 0.7911        | 1.09  | 850  | 0.9881          | 0.6606   | 0.5857   | 0.6606   |
| 0.8237        | 1.15  | 900  | 0.9677          | 0.6728   | 0.6116   | 0.6728   |
| 0.828         | 1.21  | 950  | 0.9609          | 0.6708   | 0.6069   | 0.6708   |
| 0.8477        | 1.28  | 1000 | 0.9733          | 0.663    | 0.6126   | 0.663    |
| 0.6792        | 1.34  | 1050 | 0.9752          | 0.6804   | 0.6096   | 0.6804   |
| 0.7903        | 1.41  | 1100 | 0.9644          | 0.6778   | 0.6199   | 0.6778   |
| 0.8033        | 1.47  | 1150 | 0.9827          | 0.663    | 0.6004   | 0.663    |
| 0.7558        | 1.53  | 1200 | 0.9423          | 0.6886   | 0.6196   | 0.6886   |
| 0.767         | 1.6   | 1250 | 0.9510          | 0.6762   | 0.6269   | 0.6762   |
| 0.7842        | 1.66  | 1300 | 0.9351          | 0.689    | 0.6307   | 0.689    |
| 0.8388        | 1.73  | 1350 | 0.9174          | 0.6934   | 0.6275   | 0.6934   |
| 0.7356        | 1.79  | 1400 | 0.9241          | 0.6914   | 0.6327   | 0.6914   |
| 0.6714        | 1.85  | 1450 | 0.9077          | 0.6922   | 0.6251   | 0.6922   |
| 0.9696        | 1.92  | 1500 | 0.9081          | 0.6884   | 0.6169   | 0.6884   |
| 0.7278        | 1.98  | 1550 | 0.9106          | 0.6888   | 0.6269   | 0.6888   |
| 0.5103        | 2.05  | 1600 | 0.9648          | 0.6934   | 0.6278   | 0.6934   |
| 0.4725        | 2.11  | 1650 | 0.9807          | 0.687    | 0.6287   | 0.687    |
| 0.643         | 2.17  | 1700 | 0.9953          | 0.6912   | 0.6252   | 0.6912   |
| 0.5967        | 2.24  | 1750 | 0.9662          | 0.6868   | 0.6326   | 0.6868   |
| 0.5582        | 2.3   | 1800 | 0.9957          | 0.6896   | 0.6307   | 0.6896   |
| 0.5341        | 2.37  | 1850 | 1.0167          | 0.69     | 0.6324   | 0.69     |
| 0.494         | 2.43  | 1900 | 1.0182          | 0.6884   | 0.6304   | 0.6884   |
| 0.4602        | 2.49  | 1950 | 1.0200          | 0.6908   | 0.6376   | 0.6908   |
| 0.5453        | 2.56  | 2000 | 1.0206          | 0.692    | 0.6426   | 0.692    |
| 0.5462        | 2.62  | 2050 | 1.0130          | 0.6904   | 0.6392   | 0.6904   |
| 0.4283        | 2.69  | 2100 | 1.0353          | 0.6866   | 0.6374   | 0.6866   |
| 0.5238        | 2.75  | 2150 | 1.0244          | 0.691    | 0.6418   | 0.691    |
| 0.5237        | 2.81  | 2200 | 1.0162          | 0.6904   | 0.6388   | 0.6904   |
| 0.6074        | 2.88  | 2250 | 1.0166          | 0.6938   | 0.6458   | 0.6938   |
| 0.5846        | 2.94  | 2300 | 1.0134          | 0.6936   | 0.6454   | 0.6936   |


### Framework versions

- Transformers 4.39.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2