dfurman

/

Falcon-7B-Chat-v0.1

like 44

Hi, thanks for the colab notebook.

I was going through the colab, I understand you did some post-processing in the final text generated.
print(generated_text.split(": ")[1].split(": ")[-1])

I was doing a similar task, but after the correct response the model generates some extra text, Any way to solve this, so no post processing is needed.