Hugging Face's logo

HKUSTAudio
/
Llasa-1B

Text-to-Speech
Safetensors
Chinese
English
llama
Text-to-Speech
Model card Files
xet
Community
8

Bug on model label resulting to duplicate downloads

#7
by johnkk - opened
Discussion
johnkk
•
edited Apr 9

The same model is referred with different labels in the 2 examples. This results to downloading the same models twice using a slightly different model name.

  • Speech synthesis solely from input text: llasa_1b ='HKUSTAudio/Llasa-1B'
  • Speech synthesis utilizing a given speech prompt llasa_1b ='HKUSTAudio/Llasa-1b'
# ~/.cache/huggingface/hub
...
models--HKUSTAudio--Llasa-1B/
models--HKUSTAudio--Llasa-1b/
...

Sign up or log in to comment