COMMITS
September 11, 2026
L
Merge pull request #12 from Edge0-AI/fix/35b-nan-hidden-clip
linyubupa committed
L
fix(server): keep the rendered chat template in ChatSession.prompt_ids
linyubupa committed
L
L
Merge pull request #10 from Edge0-AI/fix/mlx-0.30.6-nax-a18
linyubupa committed
L
Merge branch 'main' into fix/mlx-0.30.6-nax-a18
linyubupa committed
L
fix(deps): pin mlx 0.30.6 and document the A18 NAX mis-gating
linyubupa committed
L
Merge pull request #9 from Edge0-AI/fix/mlx-0.30.6-nax-a18
linyubupa committed
L
fix(deps): pin mlx 0.30.6 and document the A18 NAX mis-gating
linyubupa committed
September 10, 2026
A
Merge pull request #2 from ahrazzle/fix/fetch-models-tuple-replace
Alichua committed
L
docs: point the GitHub badge at the renamed repo
linyubupa committed
L
chore: unify the legacy pregate naming on prerouter
linyubupa committed
L
docs: sync the Chinese README and fix stale docs claims
linyubupa committed
A
fix(fetch): use tuple element for MODEL env var in done hint
ahrazzle committed
L
README: remove demo video
linyubupa committed
L
README: move demo video above intro paragraph
linyubupa committed
L
README: video URL back to huggingface.co
linyubupa committed
L
README: use hf-mirror video URL for playback
linyubupa committed
L
README: demo video above core mechanisms
linyubupa committed
September 9, 2026
L
README: avoid mmap jargon in core mechanisms
linyubupa committed
L
README: tighten core mechanisms
linyubupa committed
L
README: rewrite core mechanisms in technical register
linyubupa committed
L
README: drop redundant 'parallel LoRA' wording
linyubupa committed
L
README: describe Recover-LoRA (post-quantization distillation training)
linyubupa committed
L
rename HF repo id to Edge0/Edge0-35B-A3B-preview (capital B/A)
linyubupa committed
A
Merge pull request #1 from Edge0-AI/codex/release-test-fixes
Alichua committed
C
fix: stream chat responses and portable HF uploads
chuarunyuan committed
L
README: quality section wording, OpenCompass attribution
linyubupa committed
C
test: add real checkpoint smoke coverage
chuarunyuan committed
C
chore: update Edge0 8B model references
chuarunyuan committed
L
README: name base models explicitly in quality tables
linyubupa committed