Datset for Week 2 lab cannot be loaded

When running the Cell 11 (under 1.2. Load Dataset and LLM), I get errors. So, I cannot continue beyond that point. Any idea what is going on?

Thanks,
Navid

1 Like

What does the error say?

I have exactly the same problem.

Error happens in the following line of code:
dataset = loat_dataset(huggingface_dataset_name)

And here is the error message:
ValueError: Invalid pattern: ‘**’ can only be an entire path component

Same here. I have tried restarting jupyter notebook and its still the same error

If you guys didn’t modify any code you aren’t supposed to like this one for example

huggingface_dataset_name = “knkarthick/dialogsum”

then maybe @chris.favila can have a look at this issue!

I have not modified anything.
I tried to debug the error and it says it may occur when the library is updated but in this case, we load specific version of the library so this should not be the case.

1 Like

Maybe this will also help with debugging and correcting the error.
Since this is about invalid pattern ‘**’, more detailed message is attached as image:

+1 to the above.
I’m getting the same error trying to load the database.
FWIW, I did try changing:
<huggingface_dataset_name = “knkarthick/dialogsum”>
to
<huggingface_dataset_name = “knkarthick/dialogsum · Datasets at Hugging Face”>
but that gave an error saying that it didn’t find a “dialogsum.py” script in the folder.

Two things:

  • The change I tried was adding the full URL (it’s been automatically “titled” above)
  • And I got this error in the Week 3 Lab, but I’m guessing it’s all labs that access that database.

I have just ran exercises for week 2, where the same lines of code are executed and they did not throw an error.
So I went back to week 1 exercises and it seems everything works fine now.

I had no idea what the issue was but it was either fixed or just temporary.