Closed joseporiolayats closed 9 months ago
lol its weird that i was going to make a PR for this just now
Also yes large is now large-v3 and large v2 need to be added , but from how the models are downloaded directly based on the name you pass in return f"{MODELS_BASE_URL}/{MODELS_PREFIX_URL}-{model_name}.bin"
, large-v3 should be the thing as there is no large.bin
Can we have large as an alias as well for large-v3 ? If this PR doesnt reolve it ill make a PR later ig
Hey @abdeladim-s missed the opportunity to thank you for this package!!!
I made another commit in oder to add all the available models in the hf repo, including the quantized ones.
Hope this all goes well :)
lol its weird that i was going to make a PR for this just now
Also yes large is now large-v3 and large v2 need to be added , but from how the models are downloaded directly based on the name you pass in
return f"{MODELS_BASE_URL}/{MODELS_PREFIX_URL}-{model_name}.bin"
, large-v3 should be the thing as there is no large.binCan we have large as an alias as well for large-v3 ? If this PR doesnt reolve it ill make a PR later ig
Yeah, thanks @Govind-S-B for pointing that out, large-v3
is actually the right file and not large
.
For now, I can see @joseporiolayats took care of everything. As for the alias, if you have an idea feel free to send a PR any time.
Hey @abdeladim-s missed the opportunity to thank you for this package!!!
I made another commit in oder to add all the available models in the hf repo, including the quantized ones.
Hope this all goes well :)
Hey @joseporiolayats, Everything seems perfect. Thank you as well for the awesome contribution :)
Just changed a line so that it accepts large-v3 models, it wasn't even downloading them.