csukuangfj commited on
Commit
de62b55
1 Parent(s): cf507f8

Update README.

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -12,18 +12,18 @@ git lfs pull
12
 
13
  **Catuion**: You have to run `git lfs pull`. Otherwise, you will be SAD later.
14
 
15
- The model in this repo is trained using the commit `TODO`.
16
 
17
  You can use
18
 
19
  ```
20
  git clone https://github.com/k2-fsa/icefall
21
  cd icefall
22
- git checkout TODO
23
  ```
24
  to download `icefall`.
25
 
26
- You can find the model information by visiting <TODO>.
27
 
28
  In short, the encoder is a Conformer model with 8 heads, 12 encoder layers, 512-dim attention, 2048-dim feedforward;
29
  the decoder contains a 1024-dim embedding layer and a Conv1d with kernel size 2.
 
12
 
13
  **Catuion**: You have to run `git lfs pull`. Otherwise, you will be SAD later.
14
 
15
+ The model in this repo is trained using the commit `4c1b3665ee6efb935f4dd93a80ff0e154b13efb6`.
16
 
17
  You can use
18
 
19
  ```
20
  git clone https://github.com/k2-fsa/icefall
21
  cd icefall
22
+ git checkout 4c1b3665ee6efb935f4dd93a80ff0e154b13efb6
23
  ```
24
  to download `icefall`.
25
 
26
+ You can find the model information by visiting <https://github.com/k2-fsa/icefall/blob/273e5fb2f3ac2620bafdffe2689b8b3ee10173d3/egs/librispeech/ASR/transducer_stateless/train.py#L198>.
27
 
28
  In short, the encoder is a Conformer model with 8 heads, 12 encoder layers, 512-dim attention, 2048-dim feedforward;
29
  the decoder contains a 1024-dim embedding layer and a Conv1d with kernel size 2.