README.md
6.2 KB · 109 lines · markdown Raw
1 ---
2 license: apache-2.0
3 tags:
4 - uncensored
5 - qwen3.6
6 - moe
7 - gguf
8 - vision
9 - multimodal
10 language:
11 - en
12 - zh
13 - multilingual
14 pipeline_tag: image-text-to-text
15 base_model: Qwen/Qwen3.6-35B-A3B
16 ---
17
18 # Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
19
20 > **[Join the Discord](https://discord.gg/SZ5vacTXYf)** for updates, roadmaps, projects, or just to chat.
21
22 Qwen3.6-35B-A3B uncensored by HauhauCS. **0/465 Refusals.**
23
24 > **HuggingFace's "Hardware Compatibility" widget doesn't recognize K_P quants** — it may show fewer files than actually exist. Click **"View +X variants"** or go to **Files and versions** to see all available downloads.
25
26 ## About
27
28 No changes to datasets or capabilities. Fully functional, 100% of what the original authors intended - just without the refusals.
29
30 These are meant to be the best lossless uncensored models out there.
31
32 ## Aggressive Variant
33
34 Stronger uncensoring — model is fully unlocked and won't refuse prompts. May occasionally append short disclaimers (baked into base model training, not refusals) but full content is always generated.
35
36 For a more conservative uncensor that keeps some safety guardrails, check the Balanced variant when it's available.
37
38 ## Downloads
39
40 | File | Quant | BPW | Size |
41 |------|-------|-----|------|
42 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q8_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q8_K_P.gguf) | Q8_K_P | 10.06 | 44 GB |
43 | — | Q8_0 | 8.5 | — |
44 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf) | Q6_K_P | 7.07 | 31 GB |
45 | — | Q6_K | 6.6 | — |
46 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q5_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q5_K_P.gguf) | Q5_K_P | 6.47 | 28 GB |
47 | — | Q5_K_M | 5.7 | — |
48 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_P.gguf) | Q4_K_P | 5.40 | 23 GB |
49 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf) | Q4_K_M | 4.88 | 21 GB |
50 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ4_NL.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ4_NL.gguf) | IQ4_NL | 4.56 | 20 GB |
51 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ4_XS.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ4_XS.gguf) | IQ4_XS | 4.32 | 19 GB |
52 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q3_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q3_K_P.gguf) | Q3_K_P | 4.39 | 19 GB |
53 | — | Q3_K_M | 3.9 | — |
54 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ3_M.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ3_M.gguf) | IQ3_M | 3.56 | 15 GB |
55 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q2_K_P.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q2_K_P.gguf) | Q2_K_P | 3.46 | 15 GB |
56 | [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ2_M.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-IQ2_M.gguf) | IQ2_M | 2.69 | 11 GB |
57 | [mmproj-Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-f16.gguf](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive/resolve/main/mmproj-Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-f16.gguf) | mmproj (f16) | — | 899 MB |
58
59 All quants generated with importance matrix (imatrix) for optimal quality preservation on abliterated weights.
60
61 ## What are K_P quants?
62
63 K_P ("Perfect") quants are HauhauCS custom quantizations that use model-specific analysis to selectively preserve quality where it matters most. Each model gets its own optimized quantization profile.
64
65 A K_P quant effectively bumps quality up by 1-2 quant levels at only ~5-15% larger file size than the base quant. Fully compatible with llama.cpp, LM Studio, and any GGUF-compatible runtime — no special builds needed.
66
67 **Note:** K_P quants may show as "?" in LM Studio's quant column. This is a display issue only — the model loads and runs fine.
68
69 ## Specs
70
71 - 35B total parameters, ~3B active per forward pass (MoE)
72 - 256 experts, 8 routed per token
73 - Hybrid architecture: linear attention + full softmax attention (3:1 ratio)
74 - 40 layers
75 - 262K native context
76 - Natively multimodal (text, image, video)
77 - Based on [Qwen/Qwen3.6-35B-A3B](https://huggingface.co/Qwen/Qwen3.6-35B-A3B)
78
79 ## Recommended Settings
80
81 From the official Qwen authors:
82
83 **Thinking mode (default):**
84 - General: `temperature=1.0, top_p=0.95, top_k=20, min_p=0, presence_penalty=1.5`
85 - Coding/precise tasks: `temperature=0.6, top_p=0.95, top_k=20, min_p=0, presence_penalty=0`
86
87 **Non-thinking mode:**
88 - General: `temperature=0.7, top_p=0.8, top_k=20, min_p=0, presence_penalty=1.5`
89 - Reasoning tasks: `temperature=1.0, top_p=1.0, top_k=40, min_p=0, presence_penalty=2.0`
90
91 **Important:**
92 - Keep at least 128K context to preserve thinking capabilities
93 - Use `--jinja` flag with llama.cpp for proper chat template handling
94 - Vision support requires the `mmproj` file alongside the main GGUF
95
96 ## Usage
97
98 Works with llama.cpp, LM Studio, Jan, koboldcpp, and other GGUF-compatible runtimes.
99
100 ```bash
101 llama-cli -m Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_P.gguf \
102 --mmproj mmproj-Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-f16.gguf \
103 --jinja -c 131072 -ngl 99
104 ```
105
106 ## Other Models
107
108 - [HauhauCS on HuggingFace](https://huggingface.co/HauhauCS/models)
109