mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-08-08 09:32:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add a link to Colab. (#14)

Browse Source 
It demonstrates the usages of pre-trained models.
This commit is contained in:
Fangjun Kuang 2021-08-20 15:43:25 +08:00 committed by GitHub
parent 9d0cc9d829
commit 0b656e4e1c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 26 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
README.md
View File

@ -53,6 +53,13 @@ It should print the path to `icefall`.
At present, only LibriSpeech recipe is provided. Please
follow [egs/librispeech/ASR/README.md][LibriSpeech] to run it.
## Use Pre-trained models
See [egs/librispeech/ASR/conformer_ctc/README.md](egs/librispeech/ASR/conformer_ctc/README.md)
for how to use pre-trained models.
[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1huyupXAcHsUrKaWfI83iMEJ6J0Nh0213?usp=sharing)
[LibriSpeech]: egs/librispeech/ASR/README.md
[k2-install]: https://k2.readthedocs.io/en/latest/installation/index.html#
[k2]: https://github.com/k2-fsa/k2

24
egs/librispeech/ASR/conformer_ctc/README.md
View File

@ -1,6 +1,8 @@
# How to use a pre-trained model to transcribe a sound file or multiple sound files
(See the bottom of this document for the link to a colab notebook.)
You need to prepare 4 files:
- a model checkpoint file, e.g., epoch-20.pt
@ -99,22 +101,25 @@ The command to run decoding with attention decoder rescoring is:
/path/to/your/sound3.wav
```
# Decoding with a pretrained model in action
# Decoding with a pre-trained model in action
We have uploaded a pretrained model to <https://huggingface.co/pkufool/conformer_ctc>
We have uploaded a pre-trained model to <https://huggingface.co/pkufool/conformer_ctc>
The following shows the steps about the usage of the provided pretrained model.
The following shows the steps about the usage of the provided pre-trained model.
### (1) Download the pretrained model
### (1) Download the pre-trained model
```bash
sudo apt-get install git-lfs
cd /path/to/icefall/egs/librispeech/ASR
git lfs install
mkdir tmp
cd tmp
git clone https://huggingface.co/pkufool/conformer_ctc
```
**CAUTION**: You have to install `git-lfst` to download the pre-trained model.
You will find the following files:
```
@ -165,7 +170,7 @@ tmp
- `exp/pretrained.pt`
It contains pretrained model parameters, obtained by averaging
It contains pre-trained model parameters, obtained by averaging
checkpoints from `epoch-15.pt` to `epoch-34.pt`.
Note: We have removed optimizer `state_dict` to reduce file size.
@ -337,3 +342,10 @@ YET THESE THOUGHTS AFFECTED HESTER PRYNNE LESS WITH HOPE THAN APPREHENSION
2021-08-20 11:20:05,805 INFO [pretrained.py:341] Decoding Done
```
**NOTE**: We provide a colab notebook for demonstration.
[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1huyupXAcHsUrKaWfI83iMEJ6J0Nh0213?usp=sharing)
Due to limited memory provided by Colab, you have to upgrade to Colab Pro to
run `HLG decoding + LM rescoring` and `HLG decoding + LM rescoring + attention decoder rescoring`.
Otherwise, you can only run `HLG decoding` with Colab.

2
egs/librispeech/ASR/conformer_ctc/pretrained.py
View File

@ -245,11 +245,11 @@ def main():
if params.method in ["whole-lattice-rescoring", "attention-decoder"]:
logging.info(f"Loading G from {params.G}")
G = k2.Fsa.from_dict(torch.load(params.G, map_location="cpu"))
G = G.to(device)
# Add epsilon self-loops to G as we will compose
# it with the whole lattice later
G = k2.add_epsilon_self_loops(G)
G = k2.arc_sort(G)
G = G.to(device)
G.lm_scores = G.scores.clone()
logging.info("Constructing Fbank computer")