| .. |
|
deepseek-v4-flash
|
7742fbf290
feat(deepseek-v4-flash): fix current parameters
|
há 1 mês atrás |
|
deepseek-v4-flash-dspark38
|
25b18fcd56
fix(deepseek-v4): drafter b12x WO routing for real DSpark acceptance + GMU 0.78
|
há 2 meses atrás |
|
deepseek-v4-flash-eugr
|
5a9795c612
Rebase deepseek-v4-flash :fixed overlay onto production-3.76.1
|
há 2 meses atrás |
|
deepseek-v4-flash-eugr-agent
|
3aa9d380b2
feat(recipes): agent-serving profile + GMU/env fixes from benchmark findings
|
há 2 meses atrás |
|
deepseek-v4-flash-src
|
3aa9d380b2
feat(recipes): agent-serving profile + GMU/env fixes from benchmark findings
|
há 2 meses atrás |
|
gemma4-12b
|
5931869d69
refactor: DistributedBackend as Option — remove None_ variant
|
há 4 meses atrás |
|
glm51
|
09841103d8
fix: default DistributedBackend to None_ (absent means single-node)
|
há 4 meses atrás |
|
glm52
|
ef03b8c784
feat: auto-scale PP when pipeline_parallel_size == required_nodes
|
há 4 meses atrás |
|
glm52-hybrid-fp8-mxfp4
|
67de1f34da
refactor(build): rename docker-images/glm52-int4-int8 to glm52
|
há 2 meses atrás |
|
glm52-int4-int8
|
50251da4ce
Bump GLM recipes to aidendle94 production-hybrid-1.6
|
há 2 meses atrás |
|
glm52-mxfp4-experts-gptq
|
67de1f34da
refactor(build): rename docker-images/glm52-int4-int8 to glm52
|
há 2 meses atrás |
|
glm52-nf3-hybrid-context
|
50251da4ce
Bump GLM recipes to aidendle94 production-hybrid-1.6
|
há 2 meses atrás |
|
glm52-nf3-hybrid-fast
|
50251da4ce
Bump GLM recipes to aidendle94 production-hybrid-1.6
|
há 2 meses atrás |
|
glm52-nf4-lloyd-experts-gptq
|
240002c125
fix(glm52): patch vLLM structured-output reasoning-boundary bug under MTP
|
há 2 meses atrás |
|
glm53-flash-nvfp4
|
25c53b5f50
refactor(recipes): drop redundant entrypoint override from glm53-flash-nvfp4
|
há 1 mês atrás |
|
glm53-flash-nvfp4-lab
|
96c6bfe777
feat(recipes): lab defaults to --enforce-eager — MTP+CUDA-graphs quality bug confirmed by A/B bench
|
há 1 mês atrás |
|
glm53-flash-nvfp4-redhat
|
28e08afa73
feat(recipes): redhat @ 524288 ctx — pinned 10 GiB KV, autotune off, 4096 chunks, lab image for CC-12.0 topk guards
|
há 1 mês atrás |
|
laguna-m1-nvfp4
|
e692ec1813
feat(recipe): update ghcr.io/timothystewart6/vllm-gb10 to v0.25.1-gb10.1
|
há 3 meses atrás |
|
laguna-s2.1-nvfp4
|
a3bd00eef8
fix: tmux serve quoting, venv setup, cluster probe on dgx3.lan
|
há 2 meses atrás |
|
mimo-v2.5-nvfp4
|
e692ec1813
feat(recipe): update ghcr.io/timothystewart6/vllm-gb10 to v0.25.1-gb10.1
|
há 3 meses atrás |
|
minimax-m3
|
ef03b8c784
feat: auto-scale PP when pipeline_parallel_size == required_nodes
|
há 4 meses atrás |
|
minimax-m3-mxfp8
|
227eb18915
feat: auto-set VLLM_LOGGING_COLOR=1 on head node (override via recipe [env])
|
há 4 meses atrás |
|
minimax-m3-nvfp4
|
ef03b8c784
feat: auto-scale PP when pipeline_parallel_size == required_nodes
|
há 4 meses atrás |
|
qwen-agentworld-35b
|
e692ec1813
feat(recipe): update ghcr.io/timothystewart6/vllm-gb10 to v0.25.1-gb10.1
|
há 3 meses atrás |
|
qwen38-27b
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-fp8
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-fp8-dspark
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-int4
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-nvfp4
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-nvfp4-sglang-dflash2
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-nvfp4-sglang-dspark
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-nvfp4-unsloth
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-27b-nvfp4-unsloth-dspark
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
há 2 meses atrás |
|
qwen38-flash-next-fp8
|
095dec1d73
refactor(recipes): quant-agnostic served-model-name + rename glm53 to -nvfp4
|
há 1 mês atrás |
|
qwen38-flash-next-nvfp4
|
095dec1d73
refactor(recipes): quant-agnostic served-model-name + rename glm53 to -nvfp4
|
há 1 mês atrás |
|
solar-open2-nvfp4
|
0816b7ad25
fix(recipe): change solar-open2 served-model-name to solar
|
há 2 meses atrás |
|
test-composed
|
0208fc12f1
feat(compose): support subdirectory globs in composed_model sources
|
há 2 meses atrás |
|
test-mods
|
5931869d69
refactor: DistributedBackend as Option — remove None_ variant
|
há 4 meses atrás |
|
test-sleep
|
1fa2811301
feat: default recipe port is 30000
|
há 4 meses atrás |
Alexander Korolev
96c6bfe777
feat(recipes): lab defaults to --enforce-eager — MTP+CUDA-graphs quality bug confirmed by A/B bench