Mirrors/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2024-11-26 01:30:20 +01:00
Code Issues Packages Projects Releases Wiki Activity

Empty the cuda cache at model.generate()

Browse Source
This commit is contained in:
oobabooga 2023-02-25 14:39:13 -03:00
parent 1878acd9f3
commit 700311ce40
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
modules/text_generation.py
View File

@ -73,6 +73,8 @@ def formatted_outputs(reply, model_name):
return reply return reply
def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typical_p, repetition_penalty, top_k, min_length, no_repeat_ngram_size, num_beams, penalty_alpha, length_penalty, early_stopping, eos_token=None, stopping_string=None): def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typical_p, repetition_penalty, top_k, min_length, no_repeat_ngram_size, num_beams, penalty_alpha, length_penalty, early_stopping, eos_token=None, stopping_string=None):
torch.cuda.empty_cache()
original_question = question original_question = question
if not (shared.args.chat or shared.args.cai_chat): if not (shared.args.chat or shared.args.cai_chat):
question = apply_extensions(question, "input") question = apply_extensions(question, "input")