github-actions released this 10 Nov 18:37

master-1088-32d516c

32d516c

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

feat: add MiniMax-H3-Image support (#2324)

github-actions released this 10 Nov 18:37

master-1087-f8417d0

f8417d0

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

perf(cuda): fuse RMSNorm and RoPE in Sol-H3-VAE (#2323)

github-actions released this 10 Nov 18:42

master-1086-465adf3

465adf3

fix: avoid structured binding capture in MiniMax H3 SOL attention

github-actions released this 10 Nov 16:20

master-1081-e1ca5e8

e1ca5e8

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

perf(cuda): fuse bias into INT8 CUTLASS epilogues (#2320)

github-actions released this 10 Nov 14:38

master-1080-a3f46b6

a3f46b6

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

feat(server): make async job queue limits configurable (#2316)

github-actions released this 09 Nov 19:03

master-1079-a3c49b1

a3c49b1

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

refactor: unify ggml operation helpers around runner context (#2314)

github-actions released this 09 Nov 17:24

master-1078-73fef08

73fef08

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

feat: update server web UI with streaming prompt enhancement (#2312)

github-actions released this 09 Nov 17:08

master-1076-46da352

46da352

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

fix: size layer-split ranges in proportion to device capacity (#2309)

github-actions released this 09 Nov 16:21

master-1073-991f825

991f825

fix: limit default MSVC compilation parallelism

github-actions released this 09 Nov 16:12

master-1072-008b9a2

008b9a2

This commit was created on GitHub.com and signed with GitHub’s verified signature.

GPG key ID: B5690EEEBB952194

Verified

Learn about vigilant mode.

fix: apply MiniMax-H3 Diffusers LoRA token_refiner weights (#2306)