|
@@ -240,8 +240,6 @@ def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typi
|
|
|
break
|
|
break
|
|
|
yield formatted_outputs(reply, shared.model_name)
|
|
yield formatted_outputs(reply, shared.model_name)
|
|
|
|
|
|
|
|
- yield formatted_outputs(reply, shared.model_name)
|
|
|
|
|
-
|
|
|
|
|
# Stream the output naively for FlexGen since it doesn't support 'stopping_criteria'
|
|
# Stream the output naively for FlexGen since it doesn't support 'stopping_criteria'
|
|
|
else:
|
|
else:
|
|
|
for i in range(max_new_tokens//8+1):
|
|
for i in range(max_new_tokens//8+1):
|
oobabooga