huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-10 03:44:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(server): do not warp prefill logits

Browse Source
This commit is contained in:
OlivierDehaene 2023-03-09 11:33:28 +01:00
parent 941cd42e0c
commit f49786ccba
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
server/text_generation_server/utils/tokens.py
View File

@ -70,7 +70,11 @@ class NextTokenChooser:
def __call__(self, input_ids, scores):
# Warp logits
scores = self.warpers(input_ids, scores)
if scores.shape[0] > 1:
# only warp the last token logits
scores[-1:, :] = self.warpers(input_ids, scores[-1:, :])
else:
scores = self.warpers(input_ids, scores)
# Compute logprobs
logprobs = torch.log_softmax(scores, -1)