Downloads · 30 days
5.9K
10% of all-time downloads
mudler/Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-GGUF
Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-GGUF is a machine learning model from mudler. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
<div style="background-color: f59e0b; color: white; padding: 20px; border-radius: 10px; text-align: center; margin: 20px 0;" <h2 style="color: white; margin: 0 0 10px 0;"⚡ Each donation = another big MoE quantized</h2…
Downloads · 30 days
5.9K
10% of all-time downloads
All-time downloads
59K
Public
Repo size
227 GB
Likes
2
Public
Click a slice to open those files.
.gguf227 GB · 100%
From the Hugging Face model README
APEX (Adaptive Precision for EXpert Models) quantizations of LuffyTheFox/Qwen3.5-35B-A3B-Uncensored-FernflowerAI-safetensors.
Brought to you by the LocalAI team | APEX Project | Technical Report
| File | Profile | Size | Best For |
|---|---|---|---|
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Balanced.gguf | I-Balanced | 26 GB | Best overall quality/size ratio |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-Balanced.gguf | Balanced | 26 GB | General purpose |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Quality.gguf | I-Quality | 22 GB | Highest quality with imatrix |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-Quality.gguf | Quality | 22 GB | Highest quality standard |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Compact.gguf | I-Compact | 17 GB | Consumer GPUs, best quality/size |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-Compact.gguf | Compact | 17 GB | Consumer GPUs |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Mini.gguf | I-Mini | 14 GB | Smallest "safe" tier |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Nano.gguf | I-Nano | 11 GB | Experimental — IQ2_XXS mid-layer experts |
| Qwen3.5-35B-A3B-Uncensored-FernflowerAI-F16.gguf | F16 reference | 65 GB | Full-precision reference (text-only) |
APEX is a quantization strategy for Mixture-of-Experts (MoE) models. It classifies tensors by role (routed expert, shared expert, attention) and applies a layer-wise precision gradient — edge layers get higher precision, middle layers get more aggressive compression. I-variants use diverse imatrix calibration (chat, code, reasoning, tool-calling, agentic traces, Wikipedia).
The key insight: in MoE models, expert FFN tensors make up the bulk of model weight but only ~8/256 experts activate per token. APEX compresses middle-layer experts more aggressively while preserving edge layers (first/last 5) and keeping attention, SSM/Mamba, and shared expert tensors at higher precision.
See the APEX project for full details, technical report, and scripts.
The APEX Nano tier pushes mid-layer routed experts to IQ2_XXS (2.06 bpw), near-edge to IQ2_S, edges to Q3_K, with shared experts kept at Q5_K. About 20% smaller than Mini with modest quality cost — viable only on MoE thanks to sparse per-token expert activation. Requires imatrix.
Benchmarks pending. Feedback welcome.
local-ai run mudler/Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-GGUF@Qwen3.5-35B-A3B-Uncensored-FernflowerAI-APEX-I-Balanced.gguf