asi commited on
Commit
f0b4f0d
1 Parent(s): 4942cd2

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -9
README.md CHANGED
@@ -144,23 +144,19 @@ We directly adapted this mechanism from Graves ([2016](#graves-2016)). At each i
144
  The architecture is not yet directly included in the Transformers library. The code used for pre-training is available in the following [github repository](https://github.com/AntoineSimoulin/adaptive-depth-transformers). So you should install the code implementation first:
145
 
146
  ```bash
147
- pip install git+https://github.com/AntoineSimoulin/adaptive-depth-transformers
148
  ```
149
 
150
  Then you can use the model directly.
151
 
152
  ```python
153
- import sys
154
- sys.path.append('adaptative-depth-transformers')
155
-
156
- from modeling_albert_act_tf import TFAlbertActModel
157
- from modeling_albert_act import AlbertActModel
158
- from configuration_albert_act import AlbertActConfig
159
  from transformers import AlbertTokenizer
160
 
161
- model = AlbertActModel.from_pretrained('asi/albert-act-base/')
 
162
  _ = model.eval()
163
- tokenizer = AlbertTokenizer.from_pretrained('asi/albert-act-base/')
164
  inputs = tokenizer("a lump in the middle of the monkeys stirred and then fell quiet .", return_tensors="pt")
165
  outputs = model(**inputs)
166
  outputs.updates
 
144
  The architecture is not yet directly included in the Transformers library. The code used for pre-training is available in the following [github repository](https://github.com/AntoineSimoulin/adaptive-depth-transformers). So you should install the code implementation first:
145
 
146
  ```bash
147
+ !pip install git+https://github.com/AntoineSimoulin/adaptive-depth-transformers$
148
  ```
149
 
150
  Then you can use the model directly.
151
 
152
  ```python
153
+ from act import AlbertActConfig, AlbertActModel, TFAlbertActModel
 
 
 
 
 
154
  from transformers import AlbertTokenizer
155
 
156
+ tokenizer = AlbertTokenizer.from_pretrained('asi/albert-act-base')
157
+ model = AlbertActModel.from_pretrained('asi/albert-act-base')
158
  _ = model.eval()
159
+
160
  inputs = tokenizer("a lump in the middle of the monkeys stirred and then fell quiet .", return_tensors="pt")
161
  outputs = model(**inputs)
162
  outputs.updates