AttributeError: 'OutStream' object has no attribute 'buffer'

[dataeng1]

/lib/python3.8/dist-packages/whisper/transcribe.py
introduces the below error which wasn't occurring a couple of weeks ago.
Could you take a look?

Traceback (most recent call last):
File "/usr/local/lib/python3.8/dist-packages/gradio/routes.py", line 275, in run_predict
output = await app.blocks.process_api(
File "/usr/local/lib/python3.8/dist-packages/gradio/blocks.py", line 760, in process_api
result = await self.call_function(fn_index, inputs, iterator)
File "/usr/local/lib/python3.8/dist-packages/gradio/blocks.py", line 671, in call_function
prediction = await anyio.to_thread.run_sync(
File "/usr/local/lib/python3.8/dist-packages/anyio/to_thread.py", line 31, in run_sync
return await get_asynclib().run_sync_in_worker_thread(
File "/usr/local/lib/python3.8/dist-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
return await future
File "/usr/local/lib/python3.8/dist-packages/anyio/_backends/_asyncio.py", line 867, in run
result = context.run(func, *args)
File "", line 2, in create_textfile
results = model.transcribe(video_name, verbose=True, language="en")
File "/usr/local/lib/python3.8/dist-packages/whisper/transcribe.py", line 212, in transcribe
add_segment(
File "/usr/local/lib/python3.8/dist-packages/whisper/transcribe.py", line 172, in add_segment
sys.stdout.buffer.write(line.encode(sys.getdefaultencoding(), errors="replace"))
AttributeError: 'OutStream' object has no attribute 'buffer'

[glangford]

Can you describe the environment you are running in - are you running in a Jupyter notebook or a Colab?

Recent commit

• 9d646db

may be similar issue:
Running in Jupyter Notebook raises 'OutStream' object has no attribute 'buffer' #2516

[glangford]

I am not a whisper developer, but I would suggest two possible workarounds:

• set verbose to False (which bypasses the offending code)
• fall back to an earlier release, such as
  https://github.com/openai/whisper/releases/tag/v20230117

cc: @jongwook

[jongwook]

Thanks @glangford and @dataeng1, It appears that I was wrong to assume that sys.stdout.buffer is always available. I'll probably need to fall back to print() or try os.fdopen on sys.stdout.fileno() instead.

[dataeng1]

Thanks, @glangford and @jongwook
Can you describe the environment you are running in - are you running in a Jupyter notebook or a Colab?
It happened when running the below line in Colab.
`model.transcribe(video_name, verbose=True, language="en")

As proposed, changing verbose=False suppressed the error occurring by skipping if verbose: block in transcribe.py, line 172,
however as verbose=True is also a useful option to get more information, hopefully, this error gets resolved soon.

[hchaudhary1]

17Jan2023 release works without needing to change verbose.

! pip install git+https://github.com/openai/whisper.git@37a4f1be6d178bc8dad010dc33bcbbc69350e6bf

[ghost]

Hi @hchaudhary1

I get the following error when I try to install the above... can you advise?

Collecting git+https://github.com/openai/whisper.git@37a4f1be6d178bc8dad010dc33bcbbc69350e6bf Cloning https://github.com/openai/whisper.git (to revision 37a4f1be6d178bc8dad010dc33bcbbc69350e6bf) to c:\users\carl carver\appdata\local\temp\pip-req-build-k5yrzew3 Running command git clone --filter=blob:none --quiet https://github.com/openai/whisper.git 'C:\Users\Carl Carver\AppData\Local\Temp\pip-req-build-k5yrzew3' Running command git rev-parse -q --verify 'sha^37a4f1be6d178bc8dad010dc33bcbbc69350e6bf' Running command git fetch -q https://github.com/openai/whisper.git 37a4f1be6d178bc8dad010dc33bcbbc69350e6bf Running command git checkout -q 37a4f1be6d178bc8dad010dc33bcbbc69350e6bf Resolved https://github.com/openai/whisper.git to commit 37a4f1be6d178bc8dad010dc33bcbbc69350e6bf Preparing metadata (setup.py) ... done Requirement already satisfied: numpy in c:\users\carl carver\appdata\local\programs\python\python311\lib\site-packages (from openai-whisper==20230117) (1.24.1) **ERROR: Could not find a version that satisfies the requirement torch (from openai-whisper) (from versions: none) ERROR: No matching distribution found for torch**

[hchaudhary1]

I only tested on Google colab. Local install you can clone & install the exact tag.

[hanshaoaichiyu]

I have deal with this problem， this chinese blog maybe can help you ：https://blog.csdn.net/sinat_33353575/article/details/125110025