Seq2SeqTrainingArguments due to missing accelerate library which is actually installed

Intermediate
1

I am trying to reproduce the example from this blog post "Fine-Tune Whisper
Fine-Tune Whisper For Multilingual ASR with Transformers. The blog refers to a Google Colab notebook.

However, when I run the notebook, the cell that sets training arguments fails with the following code:

error_accelerate
error_accelerate993×769 84.5 KB

I tried to run both commands proposed in the error code (pip install transformers[torch] and …accelerate -U" and checked that accelerate is actually installed as can be seen in line [28] but the error persists.

The code was run on T4 GPU in Google Colab.

2

Restart the runtime and it should go away (weird colab issue)

1 Like
3

Thank you! It did!

1 Like
4

This topic was automatically closed 12 hours after the last reply. New replies are no longer allowed.