- Replaced old KVCage API with PageCache/ContiguousCache - Added --cache contiguous|paged switch for decoding comparison - Added argparse for all params (batch/prompt/gen/device/dtype) - Fixed PageCache decode crash by extending pages for full sequence |
||
|---|---|---|
| .. | ||
| demo | ||
| eval | ||
| tools | ||
| docker.sh | ||
| pre_commit.sh | ||