Downloads · 30 days
110
2% of all-time downloads
mikeyoubeach/qwen3.8-27b-uncensored-ninfer-v3
qwen3.8-27b-uncensored-ninfer-v3 is a image-text-to-text model from mikeyoubeach. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for ninfer. The card lists the license as apache-2.0.
An abliterated (refusal-removed) build of Qwen3.8-27B packaged in NInfer's native v3 .ninfer container, with the NVFP4/FP8 mixed-precision weight profile and the DFlash2 + MTP speculative-decoding heads and the vision…
Downloads · 30 days
110
2% of all-time downloads
All-time downloads
4.7K
Public
Repo size
47.4 GB
Likes
2
Public
Click a slice to open those files.
.ninfer47.4 GB · 100%
From the Hugging Face model README
.ninfer artifact (v3 container) for NInferAn abliterated (refusal-removed) build of Qwen3.8-27B packaged in NInfer's native
v3 .ninfer container, with the NVFP4/FP8 mixed-precision weight profile and the
DFlash2 + MTP speculative-decoding heads and the vision tower preserved. Runs on a single
NVIDIA RTX 5090.
This repository publishes the container upgrade of an existing community artifact so that it loads in the current NInfer engine, which requires v3 containers. No weights were modified, re-quantized, or fine-tuned.
| File | qwen3_8_27b_nvfp4_uncensored_v3.ninfer |
| Size | 23,719,759,856 bytes (22.09 GiB) |
| SHA-256 | d92051ef12c629f2bda98be4ec19059f4556415aa9bfffb0f0bd5d836a32f2d0 |
| Container | NInfer v3 (NINFER\x00\x03) |
| Artifact identity | qwen3.8-27b / nvfp4 |
| Objects / tensors | 1190 / 1184 |
| Components | text, vision, mtp, dflash2, optimized proposal head |
| Stage | Repository / tool | License |
|---|---|---|
| Base model | Qwen/Qwen3.8-27B | Apache-2.0 |
| Abliteration (refusal-direction removal on BF16 weights) | orcarouter/Qwen3.8-27B-Uncensored | Apache-2.0 |
| Packed NVFP4/FP8 text weights | orcarouter/Qwen3.8-27B-Uncensored-NVFP4 | Apache-2.0 |
v2 .ninfer artifact (source of this file) | JMVRoill/Qwen3.8-27B-Uncensored-nvfp4-NInfer | Apache-2.0 |
| v2 → v3 container upgrade | tools/upgrade_ninfer_v2_to_v3.py from Neroued/ninfer | (tooling) |
Source artifact (v2) checksum, for transparency:
43025bb64f2cb558d9ede6269f8a3c2ed8ebfa4619344fe1ba8e955ba4979218 qwen3_8_27b_nvfp4_uncensored.ninfer
tools/upgrade_ninfer_v2_to_v3.py. Per the tool's contract, stored weight values and
formats are preserved byte-for-byte; the tool rewrites the framing/metadata directory and
installs the maintained qwen3_8.jinja chat template. The resulting file is 263,664 bytes
larger than the v2 source (metadata + template only).os.posix_fadvise/POSIX_FADV_* (no-op),
os.fdatasync → os.fsync, os.sysconf("SC_PAGE_SIZE") → 4096, and os.pread/os.pwrite
emulation via lseek + read/write. These shims only affect durability/access-pattern
hints; they do not touch the artifact payload.tools.artifact.inspect reports text (Qwen3_5ForCausalLM, 64 layers, hidden 5120,
max_position_embeddings 262144), vision, mtp, dflash2 and the optimized proposal head —
so MTP and DFlash2 speculative decoding and image/video input are all available.
sm_120a); NVFP4 needs FP4 tensor corespkg-config, FFmpeg dev libraries, libcurl ≥ 7.85)# one-shot generation
./build/apps/ninfer qwen3_8_27b_nvfp4_uncensored_v3.ninfer \
--prompt "你好" --max-new 512 --device 0 \
--spec dflash2 --draft-tokens 7 --lm-head-draft
# OpenAI / Anthropic compatible server
./build/apps/ninfer-serve qwen3_8_27b_nvfp4_uncensored_v3.ninfer \
--host 0.0.0.0 --port 8080 --model-id qwen3.8-27b \
--max-context 131072 --kv-capacity auto --kv-dtype fp8 \
--spec dflash2 --draft-tokens 7 --lm-head-draft --vision
Notes:
--spec dflash2 uses the DFlash2 companion packed in this artifact; drop it (or use
--spec mtp --draft-tokens 3) if you prefer MTP.groupwise-int artifacts and will not run on pre-Blackwell GPUs.--max-context until startup no longer fits.Not benchmarked for this specific file. For reference, the official NVFP4 artifact of the non-abliterated model measured on the same class of hardware (single RTX 5090), as published by the NInfer project and reproduced independently: ~80 tok/s decode without speculation, and ~130–220 tok/s decode with DFlash2 K=7 depending on the workload; prefill in the multi-thousand tok/s range. Expect the same order of magnitude here. Quality loss vs BF16 for this quantization profile is reported upstream as roughly 1–3 points on hard reasoning benchmarks and near-zero on general chat/code.
This model has had its safety alignment substantially removed via abliteration (orthogonalizing the refusal direction out of the residual stream) by the upstream author. As a direct consequence:
Qwen3.8-27B would refuse. It has no meaningful built-in guardrails.By downloading or using this model you acknowledge and accept the above.
.ninfer artifact this file is derived fromApache License 2.0 — see LICENSE. Inherited from the base model; all upstream
terms and notices are retained.