Card: run with the NInfer-all Docker image or a build; RTX PRO 6000

README.md

@@ -25,10 +25,10 @@ uncensored tune
 as the target model and the official [z-lab/Qwen3.8-27B-DFlash2](https://huggingface.co/z-lab/Qwen3.8-27B-DFlash2)
 draft model as the DFlash2 speculative companion (recipe `qwen3_8_27b-v2`, the first NInfer recipe
 with the DFlash2 payload). The native NextN/MTP head of the tune is preserved as well, so both
-`--spec mtp` and `--spec dflash2` work. It runs on upstream NInfer v0.11 or later (one RTX 5090-class
-card) and on **NInfer-all**, the `master` branch of
+`--spec mtp` and `--spec dflash2` work. It runs on **NInfer-all**, the `master` branch of
 [iamwavecut/ninfer-all](https://github.com/iamwavecut/ninfer-all), which serves the RTX 3090, RTX
-4090 and RTX 5090.
+4090, RTX 5090 and RTX PRO 6000 Blackwell. Its first (v2) release ran on stock NInfer, which builds
+only for the RTX 5090; this v3 file was verified on NInfer-all.
 
 | | |
 |---|---|
@@ -48,7 +48,14 @@ card) and on **NInfer-all**, the `master` branch of
 hf download WaveCut/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-DFlash2-NInfer-v3 \
   Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-DFlash2-ninfer-v3.ninfer --local-dir models
 
-# DFlash2 speculative decoding (7 drafts, optimized head), CLI:
+# OpenAI/Anthropic-compatible server from the Docker image, on http://localhost:8080/v1:
+docker run --rm --gpus all -p 8080:8080 --ulimit memlock=-1 \
+  -v "$PWD/models:/models" -v ninfer-cache:/cache \
+  ghcr.io/iamwavecut/ninfer-all serve \
+  /models/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-DFlash2-ninfer-v3.ninfer \
+  --max-context 32768 --kv-capacity auto --spec dflash2 --draft-tokens 7 --lm-head-draft
+
+# From a build of NInfer-all, DFlash2 speculative decoding (7 drafts, optimized head), CLI:
 ./build/apps/ninfer models/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-DFlash2-ninfer-v3.ninfer --prompt "Summarize the difference between prefill and decode." \
   --max-context 16384 --spec dflash2 --draft-tokens 7 --lm-head-draft
 
@@ -61,10 +68,12 @@ hf download WaveCut/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-DFlash2-NInfer-v3
 (`--vision`) is the official tower: the release's BF16 projector reproduces it bit-for-bit through the FP16
 export path, so the artifact carries the official BF16 vision words.
 
-**Hardware.** Upstream NInfer targets one RTX 5090-class Blackwell card (sm_120, 32 GB). NInfer-all
-serves the same `qwen3.8-27b / groupwise-int` identity with DFlash2 on the RTX 3090, 4090 and 5090
-([its README](https://github.com/iamwavecut/ninfer-all#readme); `CMAKE_CUDA_ARCHITECTURES` is `86`,
-`89` or `120a`); the speed of the official artifact of this identity on each card is in its
+**Hardware.** Upstream NInfer targets one RTX 5090 (`sm_120a`, 32 GB). NInfer-all serves the same
+`qwen3.8-27b / groupwise-int` identity with DFlash2 on the RTX 3090, 4090, 5090 and RTX PRO 6000:
+its [Docker image](https://github.com/iamwavecut/ninfer-all#docker) needs an NVIDIA driver of the
+CUDA 13 branch (580 or newer) and the NVIDIA Container Toolkit, and a build
+([its README](https://github.com/iamwavecut/ninfer-all#building)) sets `CMAKE_CUDA_ARCHITECTURES`
+to `86`, `89` or `120a`. The speed of the official artifact of this identity on each card is in its
 [reference measurements](https://github.com/iamwavecut/ninfer-all/blob/master/docs/performance/reference-2026-09.md),
 this tune itself was not measured there. On 24 GB cards
 the official artifact of the same size fits DFlash2 with five drafts beside 258,048 tokens of