This is a load-tested sibling of mazinb/Qwen3.8-Flash-Next-Uncensored-NVFP4 with the model's one-layer MTP draft experts converted to NVFP4. It is a roughly 125B-total / 6B-active multimodal mixture-of-experts checkpoint intended for high-throughput speculative decoding on NVIDIA Blackwell.
The conversion is deterministic ModelOpt-compatible RTN with group size 16 and no calibration dataset. MTP_NVFP4_MANIFEST.json records source identity, conversion details, preservation closure, and tool version. SHA256SUMS covers every distributed model file; its SHA-256 is 17e1223cb7670c85e52ad56f243bf2702a5325f4ed2e94693a23b4ffedba0292.
The exact checkpoint passed end-to-end vLLM qualification on one NVIDIA RTX PRO 6000 Blackwell 96 GB GPU with:
This architecture requires a recent compatible vLLM build with Qwen4-Exp/Flash-Next, Blackwell NVFP4 MoE, MTP, and PLE CPU-offload support. Stock older vLLM releases may not load it. Use trust_remote_code only if your runtime policy permits it, and verify SHA256SUMS before serving.
mazinb/Qwen3.8-Flash-Next-Uncensored-NVFP4 at revision f2c21eb3d2ff5f24c208ea7e3afba65e2e70f83fThis model is intentionally abliterated/uncensored and may comply with unsafe, illegal, inaccurate, or harmful requests. It has no warranty. Users are responsible for evaluation, access controls, applicable law, and downstream use.
If this release is useful, Bitcoin donations are welcome:
bc1q5ayht3fxhj0v95fk0z8l2f6900g3awdsw5842p
model-bf16-00011.safetensors 5.7 GBmodel-bf16-00012.safetensors 3.7 GBple-bf16-34.safetensors 2.4 GBple-bf16-35.safetensors 2.4 GBple-bf16-36.safetensors 2.4 GBple-bf16-37.safetensors 2.4 GBple-bf16-38.safetensors 2.4 GBple-bf16-39.safetensors 2.4 GBple-bf16-40.safetensors 2.4 GBple-bf16-41.safetensors 2.4 GBple-bf16-00.safetensors 2.4 GBple-bf16-01.safetensors 2.4 GBple-bf16-02.safetensors 2.4 GBple-bf16-03.safetensors 2.4 GBple-bf16-04.safetensors 2.4 GBple-bf16-05.safetensors 2.4 GBple-bf16-06.safetensors 2.4 GBple-bf16-07.safetensors 2.4 GBple-bf16-08.safetensors 2.4 GBple-bf16-09.safetensors 2.4 GBple-bf16-10.safetensors 2.4 GBple-bf16-11.safetensors 2.4 GBple-bf16-12.safetensors 2.4 GBple-bf16-13.safetensors 2.4 GBple-bf16-14.safetensors 2.4 GBple-bf16-15.safetensors 2.4 GBple-bf16-16.safetensors 2.4 GBple-bf16-17.safetensors 2.4 GBple-bf16-18.safetensors 2.4 GBple-bf16-19.safetensors 2.4 GBple-bf16-20.safetensors 2.4 GBple-bf16-21.safetensors 2.4 GBple-bf16-22.safetensors 2.4 GBple-bf16-23.safetensors 2.4 GBple-bf16-24.safetensors 2.4 GBple-bf16-25.safetensors 2.4 GBple-bf16-26.safetensors 2.4 GBple-bf16-27.safetensors 2.4 GBple-bf16-28.safetensors 2.4 GBple-bf16-29.safetensors 2.4 GBple-bf16-30.safetensors 2.4 GBple-bf16-31.safetensors 2.4 GBple-bf16-32.safetensors 2.4 GBple-bf16-33.safetensors 2.4 GBple-bf16-42.safetensors 1.6 GBmodel-mtp-all-experts-nvfp4.safetensors 1.4 GBexperts-layer10.safetensors 1.4 GBexperts-layer11.safetensors 1.4 GBexperts-layer12.safetensors 1.4 GBexperts-layer13.safetensors 1.4 GBexperts-layer14.safetensors 1.4 GBexperts-layer15.safetensors 1.4 GBexperts-layer16.safetensors 1.4 GBexperts-layer17.safetensors 1.4 GBexperts-layer18.safetensors 1.4 GBexperts-layer19.safetensors 1.4 GBexperts-layer20.safetensors 1.4 GBexperts-layer21.safetensors 1.4 GBexperts-layer22.safetensors 1.4 GBexperts-layer23.safetensors 1.4 GBexperts-layer24.safetensors 1.4 GBexperts-layer25.safetensors 1.4 GBexperts-layer26.safetensors 1.4 GBexperts-layer27.safetensors 1.4 GBexperts-layer28.safetensors 1.4 GBexperts-layer29.safetensors 1.4 GBexperts-layer30.safetensors 1.4 GBexperts-layer31.safetensors 1.4 GBexperts-layer32.safetensors 1.4 GBexperts-layer33.safetensors 1.4 GBexperts-layer34.safetensors 1.4 GBexperts-layer35.safetensors 1.4 GBexperts-layer36.safetensors 1.4 GBexperts-layer37.safetensors 1.4 GBexperts-layer38.safetensors 1.4 GBexperts-layer39.safetensors 1.4 GBexperts-layer40.safetensors 1.4 GBexperts-layer41.safetensors 1.4 GBexperts-layer42.safetensors 1.4 GBexperts-layer43.safetensors 1.4 GBexperts-layer44.safetensors 1.4 GBexperts-layer45.safetensors 1.4 GBexperts-layer46.safetensors 1.4 GBexperts-layer47.safetensors 1.4 GBexperts-layer00.safetensors 1.4 GBexperts-layer01.safetensors 1.4 GBexperts-layer02.safetensors 1.4 GBexperts-layer03.safetensors 1.4 GBexperts-layer04.safetensors 1.4 GBexperts-layer05.safetensors 1.4 GBexperts-layer06.safetensors 1.4 GBexperts-layer07.safetensors 1.4 GBexperts-layer08.safetensors 1.4 GBexperts-layer09.safetensors 1.4 GBmodel-bf16-00001.safetensors 1.3 GBmodel-bf16-00010.safetensors 324 MB有文件?提交种子。
提交种子