Update README.md
Browse files
README.md
CHANGED
@@ -36,25 +36,6 @@ embeddings = model.encode(sentences)
|
|
36 |
print(embeddings)
|
37 |
```
|
38 |
|
39 |
-
## Model Summary
|
40 |
-
|
41 |
-
- Fine-tuning method: Supervised SimCSE
|
42 |
-
- Base model: [cl-tohoku/bert-base-japanese-v3](https://huggingface.co/cl-tohoku/bert-base-japanese-v3)
|
43 |
-
- Training dataset: [JSNLI](https://nlp.ist.i.kyoto-u.ac.jp/?%E6%97%A5%E6%9C%AC%E8%AA%9ESNLI%28JSNLI%29%E3%83%87%E3%83%BC%E3%82%BF%E3%82%BB%E3%83%83%E3%83%88)
|
44 |
-
- Pooling strategy: cls (with an extra MLP layer only during training)
|
45 |
-
- Hidden size: 768
|
46 |
-
- Learning rate: 5e-5
|
47 |
-
- Batch size: 512
|
48 |
-
- Temperature: 0.05
|
49 |
-
- Max sequence length: 64
|
50 |
-
- Number of training examples: 2^20
|
51 |
-
- Validation interval (steps): 2^6
|
52 |
-
- Warmup ratio: 0.1
|
53 |
-
- Dtype: BFloat16
|
54 |
-
|
55 |
-
See the [GitHub repository](https://github.com/hppRC/simple-simcse-ja) for a detailed experimental setup.
|
56 |
-
|
57 |
-
|
58 |
## Usage (HuggingFace Transformers)
|
59 |
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
|
60 |
|
@@ -96,6 +77,24 @@ SentenceTransformer(
|
|
96 |
)
|
97 |
```
|
98 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
99 |
## Citing & Authors
|
100 |
|
101 |
```
|
|
|
36 |
print(embeddings)
|
37 |
```
|
38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
## Usage (HuggingFace Transformers)
|
40 |
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
|
41 |
|
|
|
77 |
)
|
78 |
```
|
79 |
|
80 |
+
## Model Summary
|
81 |
+
|
82 |
+
- Fine-tuning method: Supervised SimCSE
|
83 |
+
- Base model: [cl-tohoku/bert-base-japanese-v3](https://huggingface.co/cl-tohoku/bert-base-japanese-v3)
|
84 |
+
- Training dataset: [JSNLI](https://nlp.ist.i.kyoto-u.ac.jp/?%E6%97%A5%E6%9C%AC%E8%AA%9ESNLI%28JSNLI%29%E3%83%87%E3%83%BC%E3%82%BF%E3%82%BB%E3%83%83%E3%83%88)
|
85 |
+
- Pooling strategy: cls (with an extra MLP layer only during training)
|
86 |
+
- Hidden size: 768
|
87 |
+
- Learning rate: 5e-5
|
88 |
+
- Batch size: 512
|
89 |
+
- Temperature: 0.05
|
90 |
+
- Max sequence length: 64
|
91 |
+
- Number of training examples: 2^20
|
92 |
+
- Validation interval (steps): 2^6
|
93 |
+
- Warmup ratio: 0.1
|
94 |
+
- Dtype: BFloat16
|
95 |
+
|
96 |
+
See the [GitHub repository](https://github.com/hppRC/simple-simcse-ja) for a detailed experimental setup.
|
97 |
+
|
98 |
## Citing & Authors
|
99 |
|
100 |
```
|