Get the app
← back to catalog · registered 2026-08-22 13:56

quant-mind/Huihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4

quant-mind Qwen 6.9B multimodal second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model. 

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/quant-mind%2FHuihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4"
Response includes
  • classification m1
  • files 10
  • hub_downloads_all_time 1,326
  • author_summary 2 models
  • readme_text full
Try next
GET/models?author=quant-mind
All 2 models by quant-mind
 GET/models?family=qwen&method=m1
Sibling qwen models · same method
 GET/methods/m1
What "m1" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
1K
285 last 30d - stable
Likes
1
Model age
3mo ago
created 2026-06-18
Downloads over time
Now1.4K→from9↑15,856%
05261.1K1.6K9 on Jun 171.4K on Oct 11JunJulAugSepOct
Jun 17 → Oct 11 · 56 snapshots · spans 116 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.

huihui-ai/Huihui-Qwen3-VL-8B-Instruct-abliterated↓ 32,774 this lineage
quant-mind/Huihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4↓ 285

Metadata

License
apache-2.0
Tags
transformers safetensors qwen3_vl image-text-to-text awq int4 w4a16 compressed-tensors vllm abliterated conversational base_model:huihui-ai/Huihui-Qwen3-VL-8B-Instruct-abliterated

Related

Other by quant-mind
Huihui-Nex-N2-mini-abliterated-GGUF 130
Same base model
mradermacher/Huihui-Qwen3-VL-8B-Instruct-abliterated-GGUF 19K noctrex/Huihui-Qwen3-VL-8B-Instruct-abliterated-GGUF 9K mradermacher/Huihui-Qwen3-VL-8B-Instruct-abliterated-i1-GGUF 4K noctrex/Huihui-Qwen3-VL-8B-Instruct-abliterated-i1-GGUF 785 porpgel/Huihui-Qwen3-VL-8B-Instruct-abliterated-Q4_K_M-GGUF 301
Total size
7.03 GB
Files
10
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-06-18 03:35
Source huggingface.co/quant-mind/Huihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4 ↗ Base model huihui-ai/Huihui-Qwen3-VL-8B-Instruct-abliterated →

Files by quantization

Auxiliary files 10 files 7.04 GB
model.safetensors 7.03 GB ddd10e87 download
tokenizer.json 10.9 MB bae3e39d download
config.json 7.48 KB a0418c3e download
chat_template.jinja 5.17 KB 12438680 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 1.30 KB 32f01452 download
recipe.yaml 1.28 KB eb46f6db download
README.md 1.04 KB 7f478dd2 download
tokenizer_config.json 419 B d63a2651 download
generation_config.json 213 B 9efdfce4 download

README current version from Hugging Face

license: apache-2.0
base_model: huihui-ai/Huihui-Qwen3-VL-8B-Instruct-abliterated
base_model_relation: quantized
pipeline_tag: image-text-to-text
tags:

  • qwen3_vl
  • awq
  • int4
  • w4a16
  • compressed-tensors
  • vllm
  • abliterated
    library_name: transformers

Huihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4

Quantização AWQ int4 (W4A16) de huihui-ai/Huihui-Qwen3-VL-8B-Instruct-abliterated,
produzida com llm-compressor.

Configuração (ajustada para máxima precisão)

  • Esquema: W4A16 (pesos int4 simétricos por grupo, ativações em 16-bit)
  • group_size: 32
  • Amostras de calibração: 512 (lmms-lab/flickr30k)
  • max_seq_length: 2048
  • duo_scaling: False
  • Mantidos em bf16 (não quantizados): torre de visão (visual.*) e lm_head

Uso com vLLM

vllm serve quant-mind/Huihui-Qwen3-VL-8B-Instruct-abliterated-AWQ-int4 --quantization compressed-tensors --max-model-len 8192

README history 1 version

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

  1. 2026-06-18Upload folder using huggingface_huba3d00c01 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 Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Get the app Browse the catalog
Open in Abliteration