Model save
Browse files
README.md
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
base_model: microsoft/deberta-base
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
metrics:
|
7 |
+
- precision
|
8 |
+
- recall
|
9 |
+
- f1
|
10 |
+
- accuracy
|
11 |
+
model-index:
|
12 |
+
- name: deberta-base-HSOL-WIKI-CLS
|
13 |
+
results: []
|
14 |
+
---
|
15 |
+
|
16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
17 |
+
should probably proofread and complete it, then remove this comment. -->
|
18 |
+
|
19 |
+
# deberta-base-HSOL-WIKI-CLS
|
20 |
+
|
21 |
+
This model is a fine-tuned version of [microsoft/deberta-base](https://huggingface.co/microsoft/deberta-base) on an unknown dataset.
|
22 |
+
It achieves the following results on the evaluation set:
|
23 |
+
- Loss: 1.1529
|
24 |
+
- Precision: 0.7757
|
25 |
+
- Recall: 0.7782
|
26 |
+
- F1: 0.7769
|
27 |
+
- Accuracy: 0.8075
|
28 |
+
|
29 |
+
## Model description
|
30 |
+
|
31 |
+
More information needed
|
32 |
+
|
33 |
+
## Intended uses & limitations
|
34 |
+
|
35 |
+
More information needed
|
36 |
+
|
37 |
+
## Training and evaluation data
|
38 |
+
|
39 |
+
More information needed
|
40 |
+
|
41 |
+
## Training procedure
|
42 |
+
|
43 |
+
### Training hyperparameters
|
44 |
+
|
45 |
+
The following hyperparameters were used during training:
|
46 |
+
- learning_rate: 3e-05
|
47 |
+
- train_batch_size: 4
|
48 |
+
- eval_batch_size: 4
|
49 |
+
- seed: 42
|
50 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
51 |
+
- lr_scheduler_type: linear
|
52 |
+
- num_epochs: 5
|
53 |
+
|
54 |
+
### Training results
|
55 |
+
|
56 |
+
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
57 |
+
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
58 |
+
| 0.6211 | 1.0 | 769 | 0.7439 | 0.8403 | 0.6654 | 0.6824 | 0.7854 |
|
59 |
+
| 0.5518 | 2.0 | 1538 | 0.4591 | 0.7945 | 0.7469 | 0.7629 | 0.8114 |
|
60 |
+
| 0.4051 | 3.0 | 2307 | 0.7194 | 0.7718 | 0.7674 | 0.7695 | 0.8036 |
|
61 |
+
| 0.2264 | 4.0 | 3076 | 0.9925 | 0.7918 | 0.7546 | 0.7682 | 0.8127 |
|
62 |
+
| 0.166 | 5.0 | 3845 | 1.1529 | 0.7757 | 0.7782 | 0.7769 | 0.8075 |
|
63 |
+
|
64 |
+
|
65 |
+
### Framework versions
|
66 |
+
|
67 |
+
- Transformers 4.42.4
|
68 |
+
- Pytorch 2.4.0+cu121
|
69 |
+
- Datasets 2.21.0
|
70 |
+
- Tokenizers 0.19.1
|