Better separate the FlexGen case

This commit is contained in:
oobabooga 2023-03-08 02:54:47 -03:00
parent 0e16c0bacb
commit 72d539dbff

View File

@ -201,12 +201,8 @@ def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typi
reply = original_question + apply_extensions(reply[len(question):], "output")
yield formatted_outputs(reply, shared.model_name)
if not shared.args.flexgen:
if output[-1] == n:
break
else:
if np.count_nonzero(input_ids[0] == n) < np.count_nonzero(output == n):
break
# Stream the output naively for FlexGen since it doesn't support 'stopping_criteria'
else:
@ -223,11 +219,6 @@ def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typi
reply = original_question + apply_extensions(reply[len(question):], "output")
yield formatted_outputs(reply, shared.model_name)
if not shared.args.flexgen:
if output[-1] == n:
break
input_ids = torch.reshape(output, (1, output.shape[0]))
else:
if np.count_nonzero(input_ids[0] == n) < np.count_nonzero(output == n):
break
input_ids = np.reshape(output, (1, output.shape[0]))