added support for m1 mac #40

enemaliwilliam · 2023-05-30T23:37:06Z

I was able to get it to run on my m1 mac with a couple tweaks. Still needs to be optimized but it works

AndrewBreyen · 2023-05-30T23:55:52Z

@enemaliwilliam Is this ONLY for Apple Silicon macs? would love to get it working on Intel too!

Currently getting this on my Intel mac when running pip install -r requirements.txt:

Building wheels for collected packages: openai-whisper, pyaudio
  Building wheel for openai-whisper (pyproject.toml) ... done
  Created wheel for openai-whisper: filename=openai_whisper-20230314-py3-none-any.whl size=796901 sha256=e71ad9dbd48dbac475263b4e66ae50ebe22bfff9c46b948efa8b70a7809dcc1d
  Stored in directory: /Users/abreyen/Library/Caches/pip/wheels/b2/13/5f/fe8245f6dc59df505879da4b2129932e342f02a80e6b87f27d
  Building wheel for pyaudio (pyproject.toml) ... error
  error: subprocess-exited-with-error
  
  × Building wheel for pyaudio (pyproject.toml) did not run successfully.
  │ exit code: 1
  ╰─> [18 lines of output]
      running bdist_wheel
      running build
      running build_py
      creating build
      creating build/lib.macosx-12.6-x86_64-cpython-310
      creating build/lib.macosx-12.6-x86_64-cpython-310/pyaudio
      copying src/pyaudio/__init__.py -> build/lib.macosx-12.6-x86_64-cpython-310/pyaudio
      running build_ext
      building 'pyaudio._portaudio' extension
      creating build/temp.macosx-12.6-x86_64-cpython-310
      creating build/temp.macosx-12.6-x86_64-cpython-310/src
      creating build/temp.macosx-12.6-x86_64-cpython-310/src/pyaudio
      clang -Wno-unused-result -Wsign-compare -Wunreachable-code -DNDEBUG -g -fwrapv -O3 -Wall -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include -DMACOS=1 -I/usr/local/include -I/usr/include -I/opt/homebrew/include -I/Users/abreyen/.pyenv/versions/3.10.8/include/python3.10 -c src/pyaudio/device_api.c -o build/temp.macosx-12.6-x86_64-cpython-310/src/pyaudio/device_api.o
      src/pyaudio/device_api.c:9:10: fatal error: 'portaudio.h' file not found
      #include "portaudio.h"
               ^~~~~~~~~~~~~
      1 error generated.
      error: command '/usr/bin/clang' failed with exit code 1
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
  ERROR: Failed building wheel for pyaudio
Successfully built openai-whisper
Failed to build pyaudio
ERROR: Could not build wheels for pyaudio, which is required to install pyproject.toml-based projects

zarifpour · 2023-05-31T01:41:15Z

@enemaliwilliam I am getting this error when I try to run on an M1 Max:

❯ python main.py
.../ecoute/ecoute_env/lib/python3.11/site-packages/whisper/timing.py:57: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.
  @numba.jit
[INFO] Adjusting for ambient noise from Default Mic. Please make some noise from the Default Mic...
[INFO] Completed ambient noise adjustment for Default Mic.
[INFO] Adjusting for ambient noise from Default Speaker. Please make or play some noise from the Default Speaker...
Traceback (most recent call last):
  File ".../ecoute/AudioRecorder.py", line 25, in adjust_for_noise
    self.recorder.adjust_for_ambient_noise(self.source)
  File ".../ecoute/custom_speech_recognition/__init__.py", line 402, in adjust_for_ambient_noise
    assert source.stream is not None, "Audio source must be entered before adjusting, see documentation for ``AudioSource``; are you using ``source`` outside of a ``with`` statement?"
           ^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Audio source must be entered before adjusting, see documentation for ``AudioSource``; are you using ``source`` outside of a ``with`` statement?

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File ".../ecoute/main.py", line 121, in <module>
    main()
  File ".../ecoute/main.py", line 78, in main
    speaker_audio_recorder = AudioRecorder.DefaultSpeakerRecorder()
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".../ecoute/AudioRecorder.py", line 63, in __init__
    self.adjust_for_noise("Default Speaker", "Please make or play some noise from the Default Speaker...")
  File ".../ecoute/AudioRecorder.py", line 24, in adjust_for_noise
    with self.source:
  File ".../ecoute/custom_speech_recognition/__init__.py", line 208, in __exit__
    self.stream.close()
    ^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'close'

AudioRecorder.py

README.md

brimbela · 2023-05-31T11:16:55Z

Great work @enemaliwilliam!
Got it running on my M1 Macbook, but nothing "happened".
I'm using a headset.
Any advice on the hardware setup?

flexbox

💯

enemaliwilliam · 2023-05-31T11:36:00Z

Great work @enemaliwilliam! Got it running on my M1 Macbook, but nothing "happened". I'm using a headset. Any advice on the hardware setup?

can you change the index of your device ? line 55 of AudioRecorder.py

brimbela · 2023-05-31T11:58:12Z

I'm now trying with default speakers.
It should be ok:
>>> p.get_device_info_by_index(1) {'index': 1, 'structVersion': 2, 'name': 'MacBook Pro Microphone', 'hostApi': 0, 'maxInputChannels': 1, 'maxOutputChannels': 0, 'defaultLowInputLatency': 0.0336875, 'defaultLowOutputLatency': 0.01, 'defaultHighInputLatency': 0.043020833333333335, 'defaultHighOutputLatency': 0.1, 'defaultSampleRate': 48000.0}
But nothing happens.
Will add some logging to understand a bit better whats going on.

brimbela · 2023-05-31T12:22:05Z

I'm now trying with default speakers. It should be ok: >>> p.get_device_info_by_index(1) {'index': 1, 'structVersion': 2, 'name': 'MacBook Pro Microphone', 'hostApi': 0, 'maxInputChannels': 1, 'maxOutputChannels': 0, 'defaultLowInputLatency': 0.0336875, 'defaultLowOutputLatency': 0.01, 'defaultHighInputLatency': 0.043020833333333335, 'defaultHighOutputLatency': 0.1, 'defaultSampleRate': 48000.0} But nothing happens. Will add some logging to understand a bit better whats going on.

Got it working. Added some logs to actually see the transcript and AI response and it's working.
What is not working for me is the UI.

zarifpour · 2023-05-31T14:09:31Z

README.md

@@ -20,6 +20,11 @@ Follow these steps to set up and run Ecoute on your local machine.
 - Windows OS (Not tested on others)


We can add M1 macs here too :)

oldsongsz

I don't agree such hard coded way. the device index is variable to different compuaters and conditions. There should be more elegant solutions.

SevaSk

Thank you for your contribution! I've left some comments on your work. Unfortunately, I don't have any specific contributing guidelines to facilitate easier approval for your pull requests. However, I recommend following industry best practices and considering the feedback provided. Please feel free to express disagreements with the design decisions I've made. Although I may not always have time to assist directly, I will be sure to approve your changes when they align with my perspective or when you manage to change my mind.

SevaSk · 2023-05-31T21:04:04Z

AudioRecorder.py

 from datetime import datetime
+import os
+
+if os.name == 'nt':


Try to refactor to set up all differences between windows and MAC in one place. Dependency injection looks appropriate here open for other approach's to keep code clean.

SevaSk · 2023-05-31T21:05:08Z

AudioRecorder.py

- break
- else:
- print("[ERROR] No loopback device found.")
+ if os.name == 'nt':


DefaultSpeakerRecorder should be agnostic to operating system. Same as point 1)

SevaSk · 2023-05-31T21:05:23Z

AudioTranscriber.py

 from heapq import merge

+if os.name == 'nt':


SevaSk · 2023-05-31T21:06:10Z

README.md

+
+
+<details>
+<summary>Windows</summary>


SevaSk · 2023-05-31T21:06:23Z

custom_speech_recognition/__init__.py

@@ -107,7 +107,10 @@ def get_pyaudio():
 Imports the pyaudio module and checks its version. Throws exceptions if pyaudio can't be found or a wrong version is installed
 """
 try:
- import pyaudiowpatch as pyaudio
+ if os.name == 'nt':


requirements.txt

vikkastaneja · 2023-06-01T03:52:44Z

pip install -r requirements.txt gives me the following error on macos with m1 chip.

  clang -Wno-unused-result -Wsign-compare -Wunreachable-code -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /opt/homebrew/Caskroom/miniconda/base/include -arch arm64 -fPIC -O2 -isystem /opt/homebrew/Caskroom/miniconda/base/include -arch arm64 -DMACOS=1 -I/usr/local/include -I/usr/include -I/opt/homebrew/include -I/opt/homebrew/Caskroom/miniconda/base/include/python3.10 -c src/pyaudio/device_api.c -o build/temp.macosx-11.1-arm64-cpython-310/src/pyaudio/device_api.o
  src/pyaudio/device_api.c:9:10: fatal error: 'portaudio.h' file not found
  #include "portaudio.h"
           ^~~~~~~~~~~~~
  1 error generated.
  error: command '/usr/bin/clang' failed with exit code 1
  [end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for pyaudio
Building wheel for future (setup.py) ... done
Created wheel for future: filename=future-0.18.3-py3-none-any.whl size=492025 sha256=06a94a32594e49c131ae82646e552c9fa6fa61b527bb5881c4903dc4db0d2285
Stored in directory: /Users/vtaneja/Library/Caches/pip/wheels/5e/a9/47/f118e66afd12240e4662752cc22cefae5d97275623aa8ef57d
Successfully built openai-whisper Wave future
Failed to build pyaudio
ERROR: Could not build wheels for pyaudio, which is required to install pyproject.toml-based projects

enemaliwilliam · 2023-06-01T11:39:33Z

pip install -r requirements.txt gives me the following error on macos with m1 chip.
  clang -Wno-unused-result -Wsign-compare -Wunreachable-code -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /opt/homebrew/Caskroom/miniconda/base/include -arch arm64 -fPIC -O2 -isystem /opt/homebrew/Caskroom/miniconda/base/include -arch arm64 -DMACOS=1 -I/usr/local/include -I/usr/include -I/opt/homebrew/include -I/opt/homebrew/Caskroom/miniconda/base/include/python3.10 -c src/pyaudio/device_api.c -o build/temp.macosx-11.1-arm64-cpython-310/src/pyaudio/device_api.o
  src/pyaudio/device_api.c:9:10: fatal error: 'portaudio.h' file not found
  #include "portaudio.h"
           ^~~~~~~~~~~~~
  1 error generated.
  error: command '/usr/bin/clang' failed with exit code 1
  [end of output]
note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for pyaudio Building wheel for future (setup.py) ... done Created wheel for future: filename=future-0.18.3-py3-none-any.whl size=492025 sha256=06a94a32594e49c131ae82646e552c9fa6fa61b527bb5881c4903dc4db0d2285 Stored in directory: /Users/vtaneja/Library/Caches/pip/wheels/5e/a9/47/f118e66afd12240e4662752cc22cefae5d97275623aa8ef57d Successfully built openai-whisper Wave future Failed to build pyaudio ERROR: Could not build wheels for pyaudio, which is required to install pyproject.toml-based projects

You probably didn't install the portaudio using homebrew

mrdavidlaing · 2023-06-01T13:25:12Z

I couldn't get this to work on Mac M1 with Python v3.10.10.

This kept throwing the error:

RuntimeError: tk.h version (8.5) doesn't match libtk.a version (8.6)

I could however get it to work with an older Python version - v3.9.8

ryanxjhan · 2023-06-01T15:50:28Z

I got it running, and my UI is working fine on my M2 Pro. No errors whatsoever. But the app couldn't separate the speaker's words and mine. Everything is coming from [You]. I tried to change the get_device_info_by_index but it didn't work

CyberNinja568 · 2023-06-02T04:04:19Z

Thanks for sharing your project on GitHub! It's impressive to see

xiaotianxt · 2023-06-05T11:15:09Z

I don't agree such hard coded way. the device index is variable to different compuaters and conditions. There should be more elegant solutions.

I totally agree with you. In fact, on my device (M2 MacBook Air), the code only works when the device index is set to 0, not 1. We definitely need to consider this variability in different setups.

alabrashJr · 2023-06-08T15:35:25Z

I attempted to run it on my Apple M1 Pro, but unfortunately, I encountered an error that I was unable to resolve despite my efforts. I followed the instructions provided in link, but the issue persisted. Do you have any recommendations for how I can resolve this error?

/Users/..../lib/python3.8/site-packages/whisper/timing.py:58: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.
  def backtrace(trace: np.ndarray):
Could not import the PyAudio C module 'pyaudio._portaudio'.
Traceback (most recent call last):
  File "main.py", line 2, in <module>
    from AudioTranscriber import AudioTranscriber
  File "/Users/.../ecoute/AudioTranscriber.py", line 15, in <module>
    import pyaudio
  File "/Users/...lib/python3.8/site-packages/pyaudio/__init__.py", line 111, in <module>
    import pyaudio._portaudio as pa
ImportError: dlopen(/Users/.../lib/python3.8/site-packages/pyaudio/_portaudio.cpython-38-darwin.so, 0x0002): symbol not found in flat namespace '_PaMacCore_SetupChannelMap'

versions:
Python 3.9.12
pip 23.1.2
pyaudio==0.2.13
Homebrew 4.0.21
Homebrew/homebrew-core (git revision 72cb0f2fe6d; last commit 2023-06-08)
Homebrew/homebrew-cask (git revision 9537303472; last commit 2023-06-08)

 brew info portaudio
==> portaudio: stable 19.7.0 (bottled), HEAD
Cross-platform library for audio I/O
http://www.portaudio.com
/opt/homebrew/Cellar/portaudio/HEAD-cb8d3dc (33 files, 546.7KB) *
  Built from source on 2023-06-08 at 18:31:58
From: https://github.com/Homebrew/homebrew-core/blob/HEAD/Formula/portaudio.rb
License: MIT
==> Dependencies
Build: pkg-config ✔
==> Options
--HEAD
	Install HEAD version
==> Analytics
install: 7,874 (30 days), 1,289 (90 days), 73,904 (365 days)
install-on-request: 3,964 (30 days), 967 (90 days), 43,194 (365 days)
build-error: 0 (30 days)

alabrashJr · 2023-06-12T11:18:09Z

I attempted to run it on my Apple M1 Pro, but unfortunately, I encountered an error that I was unable to resolve despite my efforts. I followed the instructions provided in link, but the issue persisted. Do you have any recommendations for how I can resolve this error?

/Users/..../lib/python3.8/site-packages/whisper/timing.py:58: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.
  def backtrace(trace: np.ndarray):
Could not import the PyAudio C module 'pyaudio._portaudio'.
Traceback (most recent call last):
  File "main.py", line 2, in <module>
    from AudioTranscriber import AudioTranscriber
  File "/Users/.../ecoute/AudioTranscriber.py", line 15, in <module>
    import pyaudio
  File "/Users/...lib/python3.8/site-packages/pyaudio/__init__.py", line 111, in <module>
    import pyaudio._portaudio as pa
ImportError: dlopen(/Users/.../lib/python3.8/site-packages/pyaudio/_portaudio.cpython-38-darwin.so, 0x0002): symbol not found in flat namespace '_PaMacCore_SetupChannelMap'

versions: Python 3.9.12 pip 23.1.2 pyaudio==0.2.13 Homebrew 4.0.21 Homebrew/homebrew-core (git revision 72cb0f2fe6d; last commit 2023-06-08) Homebrew/homebrew-cask (git revision 9537303472; last commit 2023-06-08)

 brew info portaudio
==> portaudio: stable 19.7.0 (bottled), HEAD
Cross-platform library for audio I/O
http://www.portaudio.com
/opt/homebrew/Cellar/portaudio/HEAD-cb8d3dc (33 files, 546.7KB) *
  Built from source on 2023-06-08 at 18:31:58
From: https://github.com/Homebrew/homebrew-core/blob/HEAD/Formula/portaudio.rb
License: MIT
==> Dependencies
Build: pkg-config ✔
==> Options
--HEAD
	Install HEAD version
==> Analytics
install: 7,874 (30 days), 1,289 (90 days), 73,904 (365 days)
install-on-request: 3,964 (30 days), 967 (90 days), 43,194 (365 days)
build-error: 0 (30 days)

I fixed the issue using the Python (v 3.11.3) I installed using brew

notjulian · 2023-06-20T20:06:12Z

this is amazing, works perfectly
M1 pro

followed this link to install pyaudio

and also:

brew install [email protected]

python3 --version
Python 3.11.4

DavidAGInnovation · 2023-06-29T19:29:13Z

Why is this not merged yet?

ES-Anon · 2023-07-03T20:16:46Z

Why is this not merged yet?

Likely because @SevaSk requested changes that haven't been addressed?

Thank you all for your work on this. Currently testing with my Intel MBP.

xjqsusu · 2023-07-16T07:54:18Z

I got it running, and my UI is working fine on my M2 Pro. No errors whatsoever. But the app couldn't separate the speaker's words and mine. Everything is coming from [You]. I tried to change the get_device_info_by_index but it didn't work

did you get this fixed? I'm seeing the same issue, which SPEAKER AND YOU are the same, reading the microphone input.

FOkvj · 2023-08-06T09:08:03Z

on Intel5 mac
$ brew install portaudio
then
$ pip install pyaudio
done

qartheeq · 2023-08-07T00:55:59Z

*/ecoute/AudioTranscriber.py", line 1, in
import whisper
ModuleNotFoundError: No module named 'whisper'

I am receiving this error

vskrch · 2023-09-16T03:27:43Z

*/ecoute/AudioTranscriber.py", line 1, in import whisper ModuleNotFoundError: No module named 'whisper'

I am receiving this error

try this:
pip install openai-whisper --no-cache-dir

added support for m1 mac

c2108d1

This was referenced May 30, 2023

Adding MAC OS support #10

Open

Rewrite in Other Languages to Support More OSs #38

Closed

Can't install requirements #35

Open

zarifpour reviewed May 31, 2023

View reviewed changes

AudioRecorder.py Outdated Show resolved Hide resolved

flexbox reviewed May 31, 2023

View reviewed changes

README.md Outdated Show resolved Hide resolved

added details & summary

08e76a3

enemaliwilliam requested a review from flexbox May 31, 2023 10:26

set default index to 1 for macos

3d13c82

enemaliwilliam requested a review from zarifpour May 31, 2023 10:43

flexbox approved these changes May 31, 2023

View reviewed changes

Merge branch 'main' into add-mac-os-support

c425ced

zarifpour approved these changes May 31, 2023

View reviewed changes

zarifpour reviewed May 31, 2023

View reviewed changes

oldsongsz suggested changes May 31, 2023

View reviewed changes

SevaSk self-requested a review May 31, 2023 20:59

SevaSk requested changes May 31, 2023

View reviewed changes

flobraeuer mentioned this pull request Jan 9, 2024

MacOS Adaptation: Switched from PyAudioWPatch to PyAudio #164

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

added support for m1 mac #40

added support for m1 mac #40

enemaliwilliam commented May 30, 2023

AndrewBreyen commented May 30, 2023

zarifpour commented May 31, 2023

brimbela commented May 31, 2023 •

edited

flexbox left a comment

enemaliwilliam commented May 31, 2023

brimbela commented May 31, 2023

brimbela commented May 31, 2023

zarifpour May 31, 2023

oldsongsz left a comment

SevaSk left a comment

SevaSk May 31, 2023

SevaSk May 31, 2023

SevaSk May 31, 2023

SevaSk May 31, 2023

SevaSk May 31, 2023

vikkastaneja commented Jun 1, 2023 •

edited

enemaliwilliam commented Jun 1, 2023

mrdavidlaing commented Jun 1, 2023

ryanxjhan commented Jun 1, 2023

CyberNinja568 commented Jun 2, 2023

xiaotianxt commented Jun 5, 2023

alabrashJr commented Jun 8, 2023 •

edited

alabrashJr commented Jun 12, 2023

notjulian commented Jun 20, 2023 •

edited

DavidAGInnovation commented Jun 29, 2023

ES-Anon commented Jul 3, 2023

xjqsusu commented Jul 16, 2023

FOkvj commented Aug 6, 2023

qartheeq commented Aug 7, 2023

vskrch commented Sep 16, 2023 •

edited

		@@ -20,6 +20,11 @@ Follow these steps to set up and run Ecoute on your local machine.
		- Windows OS (Not tested on others)



		<details>
		<summary>Windows</summary>

added support for m1 mac #40

Are you sure you want to change the base?

added support for m1 mac #40

Conversation

enemaliwilliam commented May 30, 2023

AndrewBreyen commented May 30, 2023

zarifpour commented May 31, 2023

brimbela commented May 31, 2023 • edited

flexbox left a comment

Choose a reason for hiding this comment

enemaliwilliam commented May 31, 2023

brimbela commented May 31, 2023

brimbela commented May 31, 2023

zarifpour May 31, 2023

Choose a reason for hiding this comment

oldsongsz left a comment

Choose a reason for hiding this comment

SevaSk left a comment

Choose a reason for hiding this comment

SevaSk May 31, 2023

Choose a reason for hiding this comment

SevaSk May 31, 2023

Choose a reason for hiding this comment

SevaSk May 31, 2023

Choose a reason for hiding this comment

SevaSk May 31, 2023

Choose a reason for hiding this comment

SevaSk May 31, 2023

Choose a reason for hiding this comment

vikkastaneja commented Jun 1, 2023 • edited

enemaliwilliam commented Jun 1, 2023

mrdavidlaing commented Jun 1, 2023

ryanxjhan commented Jun 1, 2023

CyberNinja568 commented Jun 2, 2023

xiaotianxt commented Jun 5, 2023

alabrashJr commented Jun 8, 2023 • edited

alabrashJr commented Jun 12, 2023

notjulian commented Jun 20, 2023 • edited

DavidAGInnovation commented Jun 29, 2023

ES-Anon commented Jul 3, 2023

xjqsusu commented Jul 16, 2023

FOkvj commented Aug 6, 2023

qartheeq commented Aug 7, 2023

vskrch commented Sep 16, 2023 • edited

brimbela commented May 31, 2023 •

edited

vikkastaneja commented Jun 1, 2023 •

edited

alabrashJr commented Jun 8, 2023 •

edited

notjulian commented Jun 20, 2023 •

edited

vskrch commented Sep 16, 2023 •

edited