]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
spec : add DFlash support (#22105)
authorRuixiang Wang <redacted>
Sun, 28 Jun 2026 13:01:34 +0000 (15:01 +0200)
committerGitHub <redacted>
Sun, 28 Jun 2026 13:01:34 +0000 (16:01 +0300)
commitd1b34251bc57b696a5c91968069f8a0e6be13ef4
tree16088ae9271bf37a607bee3d2ebf0c81dfe8f8e9
parentc1a1c8ee94e37ff7ba2c872e783aaf7f77e0f320
spec : add DFlash support (#22105)

* spec: add DFlash v2 support

* dflash: support sliding window attention per layer_types

* docs: add dflash section

---------

Co-authored-by: Kashif Rasul <redacted>
14 files changed:
common/common.h
common/speculative.cpp
conversion/__init__.py
conversion/qwen.py
docs/speculative.md
gguf-py/gguf/constants.py
src/llama-arch.cpp
src/llama-arch.h
src/llama-context.cpp
src/llama-graph.cpp
src/llama-model.cpp
src/models/dflash.cpp [new file with mode: 0644]
src/models/models.h
tests/test-llama-archs.cpp