ZeroDivisionError in NotebookProgressBar

[pete88b]

System Info

• transformers version: 4.35.1
• Platform: Windows-10-10.0.19041-SP0
• Python version: 3.9.0
• Huggingface_hub version: 0.17.3
• Safetensors version: 0.3.1
• Accelerate version: 0.20.3
• Accelerate config: not found
• PyTorch version (GPU?): 2.0.1+cu117 (True)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using GPU in script?: yes (GeForce 3080 laptop)
• Using distributed or parallel set-up in script?: no

Who can help?

Hi @muellerzr - hope this is an easy one for you (o:

I'm sometimes seeing a divide by zero error when NotebookProgressBar updates.

I've suggested changes to transformers/utils/notebook.py in my notebook.

I'm guessing that for a change this simple, it'll be easier for you to make the change - but please let me know if you prefer I make a PR for this change

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

Please see my notebook

Expected behavior

We should not get a divide by zero error when average_time_per_item is zero

[amyeroberts]

Hi @pete88b please feel free to open a PR and ping me for review!

[pete88b]

Hi @hi-sushanta #27524 fixes update but misses the fix to self.label += f", {1/self.average_time_per_item:.2f} it/s" in update_bar - there's more detail in the linked notebook in case it helps.
if it's not too late, could you update your PR to include both (o:

[hisushanta]

Thank you, Peter🤗.
I updated my code similar to you did.