iceychris / libreasr Goto Github PK

:speech_balloon: An On-Premises, Streaming Speech Recognition System

Home Page: https://news.ycombinator.com/item?id=25099847

License: MIT License

Makefile 1.16% Python 76.14% CMake 0.14% C 5.32% HTML 1.12% CSS 0.40% JavaScript 6.07% Nix 0.18% Shell 0.36% Jupyter Notebook 9.10%

asr speech-recognition pytorch fastai rnn-transducer deep-learning esp32-lyrat python

libreasr's Introduction

libreasr's People

Contributors

Stargazers

Watchers

libreasr's Issues

Look into integrating Live Transcribe

As HN user sagz suggested, look into using the
Live Transcribe Android App with LibreASR.

This would be great for deaf and hard of hearing people.

Design a Logo

Design a logo for LibreASR and share it here.

To make an open source project cool, it should have a logo 😄

Performance Improvements

The HN discussion brought up a few ways to improve the recognition performance.
Some of those are:

Look into these and figure out which ones might be worth implementing.

word timestamp

can i get words timestamp when predict audio file

RuntimeError: Didn't find engine for operation quantized::linear_prepack NoQEngine on running the docker container

Hello, I tried the Quickstart steps and I can't get through this error. Did I miss something?
Thanks!

# docker run -it -p 8080:8080 iceychris/libreasr:latest
make sde &
make sen &
make b
make[1]: Entering directory '/workspace'
python3 -u api-server.py de
make[1]: Entering directory '/workspace'
python3 -u api-server.py en
make[1]: Entering directory '/workspace'
python3 -u api-bridge.py
[api-bridge] running on :8080
LM: Failed to load.
LM: Failed to load.
Traceback (most recent call last):
  File "api-server.py", line 155, in <module>
Traceback (most recent call last):
  File "api-server.py", line 155, in <module>
        serve(args.lang)serve(args.lang)

  File "api-server.py", line 140, in serve
  File "api-server.py", line 140, in serve
        apg.add_ASRServicer_to_server(ASRServicer(lang), server)apg.add_ASRServicer_to_server(ASRServicer(lang), server)

  File "api-server.py", line 56, in __init__
  File "api-server.py", line 56, in __init__
    conf, lang, m, x_tfm, x_tfm_stream = load_stuff(lang)
    conf, lang, m, x_tfm, x_tfm_stream = load_stuff(lang)  File "/workspace/lib/inference.py", line 20, in load_stuff

  File "/workspace/lib/inference.py", line 20, in load_stuff
        conf, lang, m, tfms = parse_and_apply_config(inference=True, lang=lang)
conf, lang, m, tfms = parse_and_apply_config(inference=True, lang=lang)  File "/workspace/lib/config.py", line 151, in parse_and_apply_config

  File "/workspace/lib/config.py", line 151, in parse_and_apply_config
        load_asr_model(m, lang_name, lang, conf["cuda"]["device"], lm=lm)load_asr_model(m, lang_name, lang, conf["cuda"]["device"], lm=lm)

  File "/workspace/lib/model_utils.py", line 88, in load_asr_model
  File "/workspace/lib/model_utils.py", line 88, in load_asr_model
        model, {torch.nn.LSTM, torch.nn.Linear}, dtype=torch.qint8model, {torch.nn.LSTM, torch.nn.Linear}, dtype=torch.qint8

  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 285, in quantize_dynamic
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 285, in quantize_dynamic
    convert(model, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(model, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 365, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 366, in convert
    convert(mod, mapping, inplace=True)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 366, in convert
    reassign[name] = swap_module(mod, mapping)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 395, in swap_module
    reassign[name] = swap_module(mod, mapping)
  File "/usr/local/lib/python3.7/dist-packages/torch/quantization/quantize.py", line 395, in swap_module
    new_mod = mapping[type(mod)].from_float(mod)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 421, in from_float
    new_mod = mapping[type(mod)].from_float(mod)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 421, in from_float
        return super(LSTM, cls).from_float(mod)return super(LSTM, cls).from_float(mod)

  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 229, in from_float
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 229, in from_float
    mod.bias, mod.batch_first, mod.dropout, mod.bidirectional, dtype)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 335, in __init__
    mod.bias, mod.batch_first, mod.dropout, mod.bidirectional, dtype)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 335, in __init__
    super(LSTM, self).__init__('LSTM', *args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 85, in __init__
    super(LSTM, self).__init__('LSTM', *args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/quantized/dynamic/modules/rnn.py", line 85, in __init__
    torch.ops.quantized.linear_prepack(w_ih, b_ih)
RuntimeError: Didn't find engine for operation quantized::linear_prepack NoQEngine
    torch.ops.quantized.linear_prepack(w_ih, b_ih)
RuntimeError: Didn't find engine for operation quantized::linear_prepack NoQEngine
make[1]: *** [Makefile:55: sen] Error 1
make[1]: Leaving directory '/workspace'
make[1]: *** [Makefile:57: sde] Error 1
make[1]: Leaving directory '/workspace'

Arch Linux (AUR) package

Hello,

really nice project, I am thoroughly impressed.
However, in my humble opinion the project needs an AUR package in order to reach its full potential and address its target audience - smart nerds.

I will star the project once an AUR package is available.

:wq

Raspberry Pi Support

Make LibreASR work on Raspberry Pis.

There already is a Dockerfile which builds fine on my Pi 4.
Loading and running the PyTorch models also works.

But, loading the youtokentome tokenizer model does not work:

chris@rpi4:~ $ docker run -it -v $(pwd):/workspace libreasr-armv7 /bin/bash
root@abf9b8be2d80:/workspace# python3
Python 3.7.3 (default, Jul 25 2020, 13:03:44)
[GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import youtokentome
>>> youtokentome.BPE("tokenizer.yttm-model")
terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc
Aborted (core dumped)

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.