2
2
# TensorFlow Hub
3
3
4
4
TensorFlow Hub is an open repository and library for reusable machine learning.
5
- The [ tfhub.dev ] ( https://tfhub.dev ) repository provides many pre-trained models:
5
+ [ Kaggle Models ] ( https://www.kaggle.com/models ) provides many pre-trained models:
6
6
text embeddings, image classification models, TF.js/TFLite models and much more.
7
7
The repository is open to
8
- [ community contributors] ( https://tfhub.dev/s?subtype=publisher ) .
8
+ [ community contributors] ( https://www.kaggle.com/models?owner-type=user ) .
9
9
10
10
The [ ` tensorflow_hub ` ] ( https://github.com/tensorflow/hub ) library lets you
11
11
download and reuse them in your TensorFlow program with a minimum amount of
14
14
``` python
15
15
import tensorflow_hub as hub
16
16
17
- model = hub.KerasLayer(" https://tfhub.dev/ google/nnlm- en-dim128/2" )
17
+ model = hub.KerasLayer(" https://www.kaggle.com/models/ google/nnlm/TensorFlow2/ en-dim128/2" )
18
18
embeddings = model([" The rain in Spain." , " falls" ,
19
19
" mainly" , " In the plain!" ])
20
20
@@ -23,8 +23,8 @@ print(embeddings.shape) #(4,128)
23
23
24
24
## Next Steps
25
25
26
- - [ Find models on tfhub.dev ] ( https://tfhub.dev )
27
- - [ Publish models on tfhub.dev ] ( publish.md )
26
+ - [ Find models on kaggle.com ] ( https://www.kaggle.com/models )
27
+ - [ Publish models on kaggle.com ] ( publish.md )
28
28
- TensorFlow Hub library
29
29
- [ Install TensorFlow Hub] ( installation.md )
30
30
- [ Library overview] ( lib_overview.md )
0 commit comments