| .. |
|
deepseek-v4-flash-dspark
|
5a9795c612
Rebase deepseek-v4-flash :fixed overlay onto production-3.76.1
|
vor 2 Monaten |
|
deepseek-v4-flash-dspark-src
|
1bae27bd44
fix(docker): complete DS4 src runtime lock with flashinfer transitive deps
|
vor 2 Monaten |
|
deepseek-v4-flash-dspark38
|
25b18fcd56
fix(deepseek-v4): drafter b12x WO routing for real DSpark acceptance + GMU 0.78
|
vor 2 Monaten |
|
deepseek-v4-flash-eugr
|
7742fbf290
feat(deepseek-v4-flash): fix current parameters
|
vor 1 Monat |
|
dsv4-flash-vision
|
8d49ca7159
dsv4vision: env-gated openai mm branch (VLLM_DSV4_MM), worker shape-follow patches, local-NVMe weights (dgx3+dgx4), cache-clear runbook; text lane verified exact on new boots
|
vor 1 Monat |
|
glm52
|
240002c125
fix(glm52): patch vLLM structured-output reasoning-boundary bug under MTP
|
vor 2 Monaten |
|
glm52-nf3-hybrid
|
febdc3b9d7
feat(build): add Solar Open2 NVFP4 recipe with recipe-pinned versions
|
vor 2 Monaten |
|
glm53-flash-aiden
|
4ac6815579
glm53-flash-aiden image: bump base to production-1.2 (fused mHC prefill kernel, 1.37x mHC section @4K tokens per aiden; VLLM_MHC_FUSED_LARGE=0 to revert)
|
vor 1 Monat |
|
glm53-flash-dflash2
|
848772a3e7
glm53-flash-dflash2: target-side EAGLE3 port + rigorous KV-page blocker writeup
|
vor 1 Monat |
|
glm53-flash-exl3
|
e81ae9b4ec
EXL3 recipe: verified serving via instanttensor loader overlay
|
vor 1 Monat |
|
glm53-flash-nvfp4
|
ed4e0f99ea
build(images): digest-pin glm53 base image — tag glm53-flash-arm64-cu130 is mutable; pin sha256:905c0293 (the base our verified images were built on). Checksum changes → one no-op rebuild on next serve.
|
vor 1 Monat |
|
glm53-flash-nvfp4-4p67
|
2599e39186
glm53-flash-nvfp4-4p67: MXFP8-attention support patchset + boot verified on ds4
|
vor 1 Monat |
|
glm53-flash-nvfp4-lab
|
9079638ee4
lab recipes: instanttensor baked into lab image + enabled (verified: 631s vs 700s, two-pass loader, all boot markers + smoke 3901)
|
vor 1 Monat |
|
glm53-jovian-r27
|
bc8bb0d90a
r27: b12x loader pinned_wc -> managed (discord patch, 1350->2000 tk/s prefill claim), hard-gated
|
vor 1 Monat |
|
qwen38-flashnext-nvfp4
|
5ee2326c40
fix(recipes): switch Qwen3.8-Flash-Next to NVFP4 for 2x Spark
|
vor 1 Monat |
|
solar-open2-nvfp4
|
febdc3b9d7
feat(build): add Solar Open2 NVFP4 recipe with recipe-pinned versions
|
vor 2 Monaten |
Alexander Korolev
bc8bb0d90a
r27: b12x loader pinned_wc -> managed (discord patch, 1350->2000 tk/s prefill claim), hard-gated