nreimers
commited on
Commit
•
ab9209f
1
Parent(s):
976782c
Add new SentenceTransformer model.
Browse files- 0_CLIPModel/CLIPModel.pt +2 -2
- 0_CLIPModel/bpe_simple_vocab_16e6.txt.gz +0 -0
- README.md +347 -5
- config_sentence_transformers.json +2 -2
0_CLIPModel/CLIPModel.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:16afc66c29d22afa707a6118519bcb2d9fa56f07f4bb9cde21d87c9e5cf0283b
|
3 |
+
size 605217061
|
0_CLIPModel/bpe_simple_vocab_16e6.txt.gz
CHANGED
Binary files a/0_CLIPModel/bpe_simple_vocab_16e6.txt.gz and b/0_CLIPModel/bpe_simple_vocab_16e6.txt.gz differ
|
|
README.md
CHANGED
@@ -4,11 +4,31 @@ tags:
|
|
4 |
- sentence-transformers
|
5 |
- feature-extraction
|
6 |
- sentence-similarity
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
---
|
8 |
|
9 |
# sentence-transformers/clip-ViT-B-32
|
10 |
|
11 |
-
|
|
|
|
|
12 |
|
13 |
## Usage (Sentence-Transformers)
|
14 |
|
@@ -29,20 +49,342 @@ embeddings = model.encode(sentences)
|
|
29 |
print(embeddings)
|
30 |
```
|
31 |
|
|
|
|
|
32 |
## Evaluation Results
|
33 |
|
34 |
-
|
35 |
|
36 |
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/clip-ViT-B-32)
|
37 |
|
|
|
|
|
38 |
## Full Model Architecture
|
39 |
```
|
40 |
SentenceTransformer(
|
41 |
-
(0): CLIPModel(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
)
|
43 |
```
|
44 |
|
45 |
## Citing & Authors
|
46 |
|
47 |
-
|
48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
- sentence-transformers
|
5 |
- feature-extraction
|
6 |
- sentence-similarity
|
7 |
+
- transformers
|
8 |
+
- transformers
|
9 |
+
- transformers
|
10 |
+
- transformers
|
11 |
+
- transformers
|
12 |
+
- transformers
|
13 |
+
- transformers
|
14 |
+
- transformers
|
15 |
+
- transformers
|
16 |
+
- transformers
|
17 |
+
- transformers
|
18 |
+
- transformers
|
19 |
+
- transformers
|
20 |
+
- transformers
|
21 |
+
- transformers
|
22 |
+
- transformers
|
23 |
+
- transformers
|
24 |
+
- transformers
|
25 |
---
|
26 |
|
27 |
# sentence-transformers/clip-ViT-B-32
|
28 |
|
29 |
+
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a None dimensional dense vector space and can be used for tasks like clustering or semantic search.
|
30 |
+
|
31 |
+
|
32 |
|
33 |
## Usage (Sentence-Transformers)
|
34 |
|
|
|
49 |
print(embeddings)
|
50 |
```
|
51 |
|
52 |
+
|
53 |
+
|
54 |
## Evaluation Results
|
55 |
|
56 |
+
|
57 |
|
58 |
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/clip-ViT-B-32)
|
59 |
|
60 |
+
|
61 |
+
|
62 |
## Full Model Architecture
|
63 |
```
|
64 |
SentenceTransformer(
|
65 |
+
(0): CLIPModel(
|
66 |
+
(model): CLIP(
|
67 |
+
(visual): VisualTransformer(
|
68 |
+
(conv1): Conv2d(3, 768, kernel_size=(32, 32), stride=(32, 32), bias=False)
|
69 |
+
(ln_pre): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
70 |
+
(transformer): Transformer(
|
71 |
+
(resblocks): Sequential(
|
72 |
+
(0): ResidualAttentionBlock(
|
73 |
+
(attn): MultiheadAttention(
|
74 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
75 |
+
)
|
76 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
77 |
+
(mlp): Sequential(
|
78 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
79 |
+
(gelu): QuickGELU()
|
80 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
81 |
+
)
|
82 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
83 |
+
)
|
84 |
+
(1): ResidualAttentionBlock(
|
85 |
+
(attn): MultiheadAttention(
|
86 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
87 |
+
)
|
88 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
89 |
+
(mlp): Sequential(
|
90 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
91 |
+
(gelu): QuickGELU()
|
92 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
93 |
+
)
|
94 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
95 |
+
)
|
96 |
+
(2): ResidualAttentionBlock(
|
97 |
+
(attn): MultiheadAttention(
|
98 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
99 |
+
)
|
100 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
101 |
+
(mlp): Sequential(
|
102 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
103 |
+
(gelu): QuickGELU()
|
104 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
105 |
+
)
|
106 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
107 |
+
)
|
108 |
+
(3): ResidualAttentionBlock(
|
109 |
+
(attn): MultiheadAttention(
|
110 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
111 |
+
)
|
112 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
113 |
+
(mlp): Sequential(
|
114 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
115 |
+
(gelu): QuickGELU()
|
116 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
117 |
+
)
|
118 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
119 |
+
)
|
120 |
+
(4): ResidualAttentionBlock(
|
121 |
+
(attn): MultiheadAttention(
|
122 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
123 |
+
)
|
124 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
125 |
+
(mlp): Sequential(
|
126 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
127 |
+
(gelu): QuickGELU()
|
128 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
129 |
+
)
|
130 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
131 |
+
)
|
132 |
+
(5): ResidualAttentionBlock(
|
133 |
+
(attn): MultiheadAttention(
|
134 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
135 |
+
)
|
136 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
137 |
+
(mlp): Sequential(
|
138 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
139 |
+
(gelu): QuickGELU()
|
140 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
141 |
+
)
|
142 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
143 |
+
)
|
144 |
+
(6): ResidualAttentionBlock(
|
145 |
+
(attn): MultiheadAttention(
|
146 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
147 |
+
)
|
148 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
149 |
+
(mlp): Sequential(
|
150 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
151 |
+
(gelu): QuickGELU()
|
152 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
153 |
+
)
|
154 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
155 |
+
)
|
156 |
+
(7): ResidualAttentionBlock(
|
157 |
+
(attn): MultiheadAttention(
|
158 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
159 |
+
)
|
160 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
161 |
+
(mlp): Sequential(
|
162 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
163 |
+
(gelu): QuickGELU()
|
164 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
165 |
+
)
|
166 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
167 |
+
)
|
168 |
+
(8): ResidualAttentionBlock(
|
169 |
+
(attn): MultiheadAttention(
|
170 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
171 |
+
)
|
172 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
173 |
+
(mlp): Sequential(
|
174 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
175 |
+
(gelu): QuickGELU()
|
176 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
177 |
+
)
|
178 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
179 |
+
)
|
180 |
+
(9): ResidualAttentionBlock(
|
181 |
+
(attn): MultiheadAttention(
|
182 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
183 |
+
)
|
184 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
185 |
+
(mlp): Sequential(
|
186 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
187 |
+
(gelu): QuickGELU()
|
188 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
189 |
+
)
|
190 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
191 |
+
)
|
192 |
+
(10): ResidualAttentionBlock(
|
193 |
+
(attn): MultiheadAttention(
|
194 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
195 |
+
)
|
196 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
197 |
+
(mlp): Sequential(
|
198 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
199 |
+
(gelu): QuickGELU()
|
200 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
201 |
+
)
|
202 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
203 |
+
)
|
204 |
+
(11): ResidualAttentionBlock(
|
205 |
+
(attn): MultiheadAttention(
|
206 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
|
207 |
+
)
|
208 |
+
(ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
209 |
+
(mlp): Sequential(
|
210 |
+
(c_fc): Linear(in_features=768, out_features=3072, bias=True)
|
211 |
+
(gelu): QuickGELU()
|
212 |
+
(c_proj): Linear(in_features=3072, out_features=768, bias=True)
|
213 |
+
)
|
214 |
+
(ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
215 |
+
)
|
216 |
+
)
|
217 |
+
)
|
218 |
+
(ln_post): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
219 |
+
)
|
220 |
+
(transformer): Transformer(
|
221 |
+
(resblocks): Sequential(
|
222 |
+
(0): ResidualAttentionBlock(
|
223 |
+
(attn): MultiheadAttention(
|
224 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
225 |
+
)
|
226 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
227 |
+
(mlp): Sequential(
|
228 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
229 |
+
(gelu): QuickGELU()
|
230 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
231 |
+
)
|
232 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
233 |
+
)
|
234 |
+
(1): ResidualAttentionBlock(
|
235 |
+
(attn): MultiheadAttention(
|
236 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
237 |
+
)
|
238 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
239 |
+
(mlp): Sequential(
|
240 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
241 |
+
(gelu): QuickGELU()
|
242 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
243 |
+
)
|
244 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
245 |
+
)
|
246 |
+
(2): ResidualAttentionBlock(
|
247 |
+
(attn): MultiheadAttention(
|
248 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
249 |
+
)
|
250 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
251 |
+
(mlp): Sequential(
|
252 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
253 |
+
(gelu): QuickGELU()
|
254 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
255 |
+
)
|
256 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
257 |
+
)
|
258 |
+
(3): ResidualAttentionBlock(
|
259 |
+
(attn): MultiheadAttention(
|
260 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
261 |
+
)
|
262 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
263 |
+
(mlp): Sequential(
|
264 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
265 |
+
(gelu): QuickGELU()
|
266 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
267 |
+
)
|
268 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
269 |
+
)
|
270 |
+
(4): ResidualAttentionBlock(
|
271 |
+
(attn): MultiheadAttention(
|
272 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
273 |
+
)
|
274 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
275 |
+
(mlp): Sequential(
|
276 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
277 |
+
(gelu): QuickGELU()
|
278 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
279 |
+
)
|
280 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
281 |
+
)
|
282 |
+
(5): ResidualAttentionBlock(
|
283 |
+
(attn): MultiheadAttention(
|
284 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
285 |
+
)
|
286 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
287 |
+
(mlp): Sequential(
|
288 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
289 |
+
(gelu): QuickGELU()
|
290 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
291 |
+
)
|
292 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
293 |
+
)
|
294 |
+
(6): ResidualAttentionBlock(
|
295 |
+
(attn): MultiheadAttention(
|
296 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
297 |
+
)
|
298 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
299 |
+
(mlp): Sequential(
|
300 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
301 |
+
(gelu): QuickGELU()
|
302 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
303 |
+
)
|
304 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
305 |
+
)
|
306 |
+
(7): ResidualAttentionBlock(
|
307 |
+
(attn): MultiheadAttention(
|
308 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
309 |
+
)
|
310 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
311 |
+
(mlp): Sequential(
|
312 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
313 |
+
(gelu): QuickGELU()
|
314 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
315 |
+
)
|
316 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
317 |
+
)
|
318 |
+
(8): ResidualAttentionBlock(
|
319 |
+
(attn): MultiheadAttention(
|
320 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
321 |
+
)
|
322 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
323 |
+
(mlp): Sequential(
|
324 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
325 |
+
(gelu): QuickGELU()
|
326 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
327 |
+
)
|
328 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
329 |
+
)
|
330 |
+
(9): ResidualAttentionBlock(
|
331 |
+
(attn): MultiheadAttention(
|
332 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
333 |
+
)
|
334 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
335 |
+
(mlp): Sequential(
|
336 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
337 |
+
(gelu): QuickGELU()
|
338 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
339 |
+
)
|
340 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
341 |
+
)
|
342 |
+
(10): ResidualAttentionBlock(
|
343 |
+
(attn): MultiheadAttention(
|
344 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
345 |
+
)
|
346 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
347 |
+
(mlp): Sequential(
|
348 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
349 |
+
(gelu): QuickGELU()
|
350 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
351 |
+
)
|
352 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
353 |
+
)
|
354 |
+
(11): ResidualAttentionBlock(
|
355 |
+
(attn): MultiheadAttention(
|
356 |
+
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
|
357 |
+
)
|
358 |
+
(ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
359 |
+
(mlp): Sequential(
|
360 |
+
(c_fc): Linear(in_features=512, out_features=2048, bias=True)
|
361 |
+
(gelu): QuickGELU()
|
362 |
+
(c_proj): Linear(in_features=2048, out_features=512, bias=True)
|
363 |
+
)
|
364 |
+
(ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
365 |
+
)
|
366 |
+
)
|
367 |
+
)
|
368 |
+
(token_embedding): Embedding(49408, 512)
|
369 |
+
(ln_final): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
|
370 |
+
)
|
371 |
+
)
|
372 |
)
|
373 |
```
|
374 |
|
375 |
## Citing & Authors
|
376 |
|
377 |
+
This model was trained by [sentence-transformers](https://www.sbert.net/).
|
378 |
+
|
379 |
+
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
|
380 |
+
```bibtex
|
381 |
+
@inproceedings{reimers-2019-sentence-bert,
|
382 |
+
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
|
383 |
+
author = "Reimers, Nils and Gurevych, Iryna",
|
384 |
+
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
|
385 |
+
month = "11",
|
386 |
+
year = "2019",
|
387 |
+
publisher = "Association for Computational Linguistics",
|
388 |
+
url = "http://arxiv.org/abs/1908.10084",
|
389 |
+
}
|
390 |
+
```
|
config_sentence_transformers.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "2.0.0",
|
4 |
-
"transformers": "4.
|
5 |
-
"pytorch": "1.
|
6 |
}
|
7 |
}
|
|
|
1 |
{
|
2 |
"__version__": {
|
3 |
"sentence_transformers": "2.0.0",
|
4 |
+
"transformers": "4.7.0",
|
5 |
+
"pytorch": "1.9.0+cu102"
|
6 |
}
|
7 |
}
|