demiralpdev/Caphlon-Model
A 6.87 GiB GGUF build of Qwen 3.8 27B — full 27B quality on 8 GB RAM. 96.7% HumanEval, 92.9% MMLU retention, no distillation.
GitHub repository with 14 stars and 2 forks.
Language: HTML
Topics: 27b, edge-llm, gguf, iq2s, layer-pruning, lightweight-llm, llama-cpp, llm, local-llm, model-card