Skip to content

[RFC] DFlash on Qwen3-4b #20840

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
cthotti wants to merge 7 commits into
base: main
Choose a base branch
from
+41,399 −3
Open

[RFC] DFlash on Qwen3-4b #20840

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
ed80f41
DFlash Phase 2: draft model definition, HF weight loading, and .pte e…
cthotti Jul 4, 2026
e0420a3
DFlash: consolidate tests
cthotti Jul 5, 2026
7be25cd
Add DFlash speculative decoding support for Qwen3
cthotti Jul 10, 2026
01c4e82
Address code review: drop dead Makefile target, fix pytest-collection…
cthotti Jul 14, 2026
8e6b664
Benchmarking on different Apple hardware.
cthotti Jul 16, 2026
962ce9c
Creating Dflash_experiments.md to document every experiment run for n…
cthotti Jul 21, 2026
9f86c90
Implementing Dflash_experiments.md for next users.
cthotti Jul 21, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,6 +69,7 @@ xcuserdata/
/src/executorch/include/
/src/executorch/share/
/src/executorch/version.py
/dflash_benchmarks.md
*_etdump

# Android
Expand All @@ -85,3 +86,6 @@ zephyr_dev_root.backup.*/
# Agents
.claude/*.local.*
extension/pybindings/mlx.metallib

# Scratch/WIP work not ready for review -- never committed
/wip/
5 changes: 5 additions & 0 deletions Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -484,3 +484,8 @@ clean:
rm -rf cmake-out \
extension/llm/tokenizers/build \
extension/llm/tokenizers/pytorch_tokenizers.egg-info

# qwen3_dflash-mlx target removed: it depended on the C++ engine sources
# (CMakeLists.txt, CMakePresets.json, qwen3_dflash_engine.*), which are
# gitignored/not yet landed. Restore this target in the follow-up PR that
# actually lands the C++ engine.
Loading
Loading