Skip to content

v0.4.6

Choose a tag to compare

View all tags
@asher asher released this 30 Aug 03:25
· 40 commits to main since this release
Immutable release. Only release title and notes can be modified.
v0.4.6
7610257
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

What's Changed

  • fix(spec): stream MTP drafter seeding into prefill (kills the post-first-token stall at depth) by @asher in #98
  • chore(benchmarks): regen adding Qwen3.8-Flash-Next UD-Q3_K_XL by @asher in #99
  • feat(qwen4exp): stream the n-gram table instead of the experts when only the table is over budget by @asher in #100
  • feat(stream,qwen4exp): stream table and experts together when the model is still over budget without table by @asher in #101
  • feat(load): support for GGUFs with STQ1_0 layers by @asher in #102

Full Changelog: v0.4.5...v0.4.6

Contributors

asher
Assets 3
Loading