Mt0 or Bloomz model: results are always same short length?

#56

by dsbyprateekg - opened Jan 25

Jan 25

I tried to run the official prompt with or without max_new_tokens with bigscience/mt0-base model, but got the same length response:

With Bloomz model, I am not seeing even a output-

Can you please tell me what I am missing here?

Muennighoff

BigScience Workshop org Jan 25

For BLOOMZ, try set min_new_tokens

For mT0 maybe the new tokens are all special tokens that are removed or so - can you check the shape of the output? I.e. does it correspond to >= 2000

dsbyprateekg

Jan 25

@Muennighoff for MT0, shape is torch.Size([1, 20])

And for BLOOMZ, it gave the unexpected output-

Muennighoff

BigScience Workshop org Jan 25

for mt0, it tells you in the warning that 2000 is too big

for bloomz the output looks ok to me. you may get better perf doing e.g. {text}. Please translate the prior text to English.

dsbyprateekg

Jan 25

@Muennighoff for mt0, I updated and now the output is like below-

And for bloomz, output is now-

Muennighoff

BigScience Workshop org Jan 25

for {text} i meant that you should put your text 😂

dsbyprateekg

Jan 25

my bad, I updated-

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment