Get the app
← back to catalog · registered 2026-09-14 13:56

jan1k/Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-GGUF

jan1k 35B GGUF MoE multimodal second-order 262K ctx
Abliteration classifier · v1.0.0
M-U
Primary method

Uncensored (method unknown)

No other method signals detected in this model.
Confidence
LOW
Why this label 3 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'uncensored' in name/tags but no 'abliterated' marker
  • method not identifiable from author declaration alone
  • may be DPO fine-tune, prompt engineering, or unknown technique
How we label → About M-U →
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
567
HF trending score
1
mild recent interest
Likes
1
Model age
1d ago
created 2026-09-14
Downloads over time
Now567→from0↑0%
02084166240 on Sep 14567 on Sep 16Sep
Sep 14 → Sep 16 · 3 snapshots · spans 2 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive↓ 1,068,618 this lineage
jan1k/Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-GGUF↓ 647

Metadata

License
apache-2.0
Languages
en zh multilingual
Tags
gguf text-generation nvfp4 fp4 moe qwen3.6 uncensored vision multimodal genesis llama-cpp image-text-to-text

Related

Other by jan1k
Hermes3.6-35B-A3B-Uncensored-Genesis-NVFP4-GGUF 35K Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-NVFP4-GGUF 18K Qwen3.8-27B-Uncensored-Genesis-V1-NVFP4-GGUF 8K Qwen3.5-9B-Uncensored-Genesis-NVFP4-GGUF 5K Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-NVFP4-GGUF 3K
Same base model
LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V9-GGUF 1.1M LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V10-GGUF 1.1M LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V11-GGUF 1M LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V12-GGUF 965K LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V13-GGUF 816K
Total size
20.0 GB
Files
3
Quantizations
1
Registered
2026-09-14 13:56
Last updated on HF
2026-09-14 17:20
Source huggingface.co/jan1k/Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-GGUF ↗ Base model HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive →

Files by quantization

Auxiliary files 3 files 20.0 GB
Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-v4.gguf 20.0 GB 55633c72 download
README.md 5.07 KB 37eb5141 download
.gitattributes 1.57 KB 43d98875 download

README current version from Hugging Face

license: apache-2.0
tags:

  • text-generation
  • gguf
  • nvfp4
  • fp4
  • moe
  • qwen3.6
  • uncensored
  • vision
  • multimodal
  • genesis
  • llama-cpp
    language:
  • en
  • zh
  • multilingual
    pipeline_tag: image-text-to-text
    base_model:
  • HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
  • LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Final-GGUF
    base_model_relation: quantized

Qwen3.6-35B-A3B-Uncensored-Genesis — NVFP4 GGUF

NVFP4 quantisation of
LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Final-GGUF,
built with advanced-gguf-quantizer
(a llama.cpp fork focused on NVFP4/MXFP6 quantization).

This is the non-Hermes version — Genesis tensor repair on the
HauhauCS uncensored base, without the Hermes finetune transfer.
For the Hermes version, see
jan1k/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-NVFP4-GGUF.

Files

v4 — Recommended (inline scales, LM Studio compatible)

File Size Tensors
Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-v4.gguf ~20 GB 733

v4 uses --nvfp4-inline-scales-only to emit native inline UE4M3 scales
without separate .scale/.input_scale tensors. This is required for
LM Studio / Pelican and other runtimes that do not support the extended
NVFP4 scale tensor contract.

Imatrix variants will follow after the plain v4 is confirmed working
in target runtimes.

Source
Source GGUF Qwen3.6-35B-A3B-Uncensored-Genesis-Q8_K_P.gguf (43.6 GB, 10.06 BPW)
Architecture qwen35moe (MoE), 35B total / ~3B active, 40 layers
Context 262144
general.file_type 39 (LLAMA_FTYPE_MOSTLY_NVFP4)
MTP/NextN none

Two-step pipeline: Q8_K_P -> F16 intermediate -> NVFP4. The F16 step gives
the NVFP4 encoder clean data (no Q8_0 quantization noise).

Tensor mix (v4)

type count notes
F32 331 norms, ssm scalars, gate inputs
F16 175 sensitive weights (blk.0 attn, ssm, etc.)
Q6_K 1 output.weight
NVFP4 226 bulk weights
total 733
separate .scale/.input_scale 0 inline UE4M3 only

Format fix (v3/v2 -> v4)

The older v2/v3 files use the extended NVFP4 format with separate
.scale and .input_scale tensors. Runtimes like LM Studio / Pelican
only read inline UE4M3 scales and ignore the separate scale tensors,
producing garbage output.

v4 uses --nvfp4-inline-scales-only to:

  • Skip writing .scale and .input_scale tensors
  • Force tensor_scale=1.0 (no pre-scaling)
  • Force input_scale=identity
  • Make inline UE4M3 scales self-contained

Tensor protection policy

F16 singular-collapse protection:

tensor type
blk.0.attn_gate.weight F16
blk.0.attn_qkv.weight F16
blk.0.ffn_down_exps.weight F16
blk.13.ffn_down_exps.weight F16

F32 architecture-specific protection:

tensor reason
blk.*.attn_norm.weight 1D norm
blk.*.post_attention_norm.weight 1D norm
blk.*.attn_q_norm.weight 1D norm
blk.*.attn_k_norm.weight 1D norm
blk.*.ssm_norm.weight 1D norm
output_norm.weight 1D norm
blk.*.ssm_conv1d.weight CUDA SSM conv kernel requires F32
blk.*.ssm_dt.bias SSM scalar, kernel compatibility
blk.*.ssm_a SSM scalar, kernel compatibility
blk.*.ffn_gate_inp_shexp.weight shared expert gate input
token_embd.weight F16 (embedding quality)

Forced NVFP4 (do not push lower):

tensor type
blk.0.ssm_out.weight NVFP4
blk.1.attn_gate.weight NVFP4
blk.1.attn_qkv.weight NVFP4

Usage

llama-cli -m Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-v4.gguf \
  --mmproj mmproj-Qwen3.6-35B-A3B-Uncensored-Genesis-F16.gguf \
  --jinja -c 131072 -ngl 99
  • Set K cache and V cache quantization to F16
  • Set GPU offload to maximum, active experts to 8
  • Set number of layers for which to force MoE weights onto CPU to 40

Hardware

  • Blackwell (RTX 50xx): native FP4 path, fastest
  • Ampere (RTX 30xx): NVFP4 inference works via fallback kernels
  • Quantisation was done CPU-only (Ampere CUDA NVFP4 encoder hangs on MoE)

Reproducibility

# Step 1: Q8_K_P -> F16 intermediate
llama-quantize --allow-requantize \
  Qwen3.6-35B-A3B-Uncensored-Genesis-Q8_K_P.gguf temp_f16.gguf F16 6

# Step 2: F16 -> NVFP4 v4 (inline scales only, CPU-only)
CUDA_VISIBLE_DEVICES="" llama-quantize \
  --allow-requantize \
  --nvfp4-inline-scales-only \
  --tensor-type-file tensor_types_protection.txt \
  temp_f16.gguf \
  Qwen3.6-35B-A3B-Uncensored-Genesis-NVFP4-v4.gguf \
  NVFP4 6

# Cleanup
rm temp_f16.gguf

Credits

README history 1 version

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-09-14Upload README.md with huggingface_hub08eecb95.1 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in app" button that hands off directly to a local runtime of your choice - Infrahuman, LM Studio, or Ollama. No API keys, no subscription, no prompt leakage.

See supported apps Browse the catalog