#error in fine-tuning - Stream interrupted (client disconnected)

9 messages · Page 1 of 1 (latest)

amber umbra
#

I'm running a fine-tuning doing that code:

openai api fine_tunes.create -t ./data3.jsonl -m davinci

but getting that error message:

Upload progress: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 13.7k/13.7k [00:00<00:00, 7.30Mit/s]
Uploaded file from ./data3.jsonl: file-91pV0GxB6CzMK82Up9ZhWarL
Created fine-tune: ft-RRXBArk8o0Y5jEqi117LrWVV
Streaming events until fine-tuning is complete...

(Ctrl-C will interrupt the stream, but not cancel the fine-tune)

Stream interrupted (client disconnected).
To resume the stream, run:

  openai api fine_tunes.follow -i ft-RRXBArk8o0Y5jEqi117LrWVV

/Users/m/opt/anaconda3/lib/python3.9/site-packages/openai/cli.py:406: RuntimeWarning: coroutine 'FineTune.stream_events' was never awaited
  cls._stream_events(resp["id"])
RuntimeWarning: Enable tracemalloc to get the object allocation traceback

does anyone know why it's happening?

spark marlin
#

Hey! It looks like you forgot an await in your code. In Python, for any function that is defined as async, you need to await it before the function call

amber umbra
spark marlin
#

Oh I see! I'm not too familiar with the fine tuning command line util, have you made sure that you're on a reliable internet connection and not connected to a VPN, etc?

amber umbra
# 
pip uninstall openai
pip install openai==0.25.0
jade niche
#

Hi... I'm trying to fine tune my custom dataset using davinci model. But it is taking forever to generate the fine tune ... Is there something we can do about it?

cursive pike
#

I'm getting the same error on fine_tunes.follow, and downgrading the openai cli version worked. Seems like a bug in the current cli.