Issues

Potential privacy / IP leakage from datastore-backed speculative decoding
FasterDecoding/REST#36 · opened on May 6, 2026
[llama3 branch] DraftRetriever Reader panics when there is exactly one matching suffix
FasterDecoding/REST#34 · opened on Mar 4, 2026
3
[BUG?] There appears to be a bug in DraftRetriever's core search code.
FasterDecoding/REST#33 · opened on Jan 2, 2026
thread '<unnamed>' panicked at 'called Option::unwrap() on a None value'
FasterDecoding/REST#31 · opened on Apr 28, 2025
1
question about token generation
FasterDecoding/REST#30 · opened on Apr 27, 2025
Problem with draftretriever.search function behavior
FasterDecoding/REST#29 · opened on Feb 20, 2025
Lama2 70B-chat is not supported?
FasterDecoding/REST#22 · opened on Oct 17, 2024
2
Tensor parallelism
FasterDecoding/REST#21 · opened on Oct 11, 2024
about retrieve sequence length
FasterDecoding/REST#20 · opened on Sep 27, 2024
1
OSError: failed to fill whole buffer
FasterDecoding/REST#19 · opened on Sep 25, 2024
7
How to adjust the maximum token lengths when drafting?
FasterDecoding/REST#18 · opened on Sep 7, 2024
3
If I want to change REST code to support multi-batch inference, what needs to be changed?
FasterDecoding/REST#16 · opened on May 20, 2024
How to support the repetition_penalty parameter?
FasterDecoding/REST#15 · opened on May 20, 2024
Segmentation Fault when calling libsais_int
FasterDecoding/REST#14 · opened on May 16, 2024
ValueError: draft_choices was not cut enough / draft_len should not exceed 65
FasterDecoding/REST#13 · opened on May 8, 2024
4