webpolis
/

zenos-gpt-j-6B-instruct-4bit

Text Generation

4-bit precision

Model card Files Files and versions

Nicolas Iglesias commited on Oct 11, 2023

Commit

ecf5055

·

1 Parent(s): 3e2d54d

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -64,7 +64,7 @@ with torch.no_grad():
 s = generation_output.sequences[0]
 output = tokenizer.decode(s)
-start_txt = output.find('### Respuesta:\n') + len('### Respuesta:\n')
 end_txt = output.find("<|endoftext|>", start_txt)
 answer = output[start_txt:end_txt]

 s = generation_output.sequences[0]
 output = tokenizer.decode(s)
+start_txt = output.find('[/INST]') + len('[/INST]')
 end_txt = output.find("<|endoftext|>", start_txt)
 answer = output[start_txt:end_txt]