サイト内検索

AI モデル 85003

• LFM3-VL-2B-FP8↗

  LiquidAI

• Nex-N2-mini-ultra-uncensored-heretic-GGUF↗

  llmfan46

• Qwen3.6-27B-A3B-Coder-MTP-GGUF↗

  ManniX-ITA

  A **code-specialist expert prune** of Qwen3.6-35B-A3B: the MoE is reduced from **256 experts to 184** (72 dropped per layer, ~35B→27B, still A3B active) using a **code-targeted competence map** (LiveCodeBench + MultiPL-E competence classes). Same router, attention, norms, MTP head and vision tower as the base — only the expert keep-set changes. **Served at top-10** (`num_experts_per_tok = 10`, baked as the default). This is a routing-recovery lever: after pruning to 184 experts, activating the top-10 (vs the base top-8) recovers instruction-following at **no cost to code** (see below). No fine-tuning, no distillation — pure expert selection + a routing-width dial. 1. **Competence map**: the 256e teacher is profiled per-expert on a balanced corpus + targeted **LiveCodeBench** and **MultiPL-E** (Rust/Java/JS) PASS-response classes. 2. **Drop map**: `wmax` aggregation with the LCB + MPE classes up-weighted (1.5) → 72/256 experts dropped per layer, protecting the code-competent experts. 3. **Top-10 routing** (`num_experts_per_tok = 10`) baked into the config → the shipped default. Pass `--override-kv qwen35moe.expert_used_count=int:8` to any llama.cpp tool to A/B back to native top-8.

• Experiment27Yam-7B-GGUF↗

  MaziyarPanahi

• VulRepair↗

  MickyMike

• voxprint-mirror-siglip2-so400m-patch14-384↗

  Mitroshenkov87

• MobileMoE-L-SFT↗

  ml-resources

• olmOCR-7B-0225-preview-bf16↗

  mlx-community

• DarkIdol-Llama-3.1-8B-Instruct-1.2-Uncensored-GGUF↗

  mradermacher

• DeepSeek-R1-Fine-tuned-Medical-GGUF↗

  mradermacher