🚀 Pytorch Distributed native training library for LLMs/VLMs with OOTB Hugging Face support
Qwen 3.8 27B Uncensored chat — qwen 3.8, qwen 3.8 27b, qwen ai, qwen chat, ollama qwen, qwen 3.8 ollama, qwen coder, qwen 3.8 27b download, qwen 3.8 27b requirements, alibaba qwen, huggingface, LM Studio, llama.cpp. Download:🡇
Docker deployment of Qwen3.8-27B (NVFP4) on a single RTX 5090: Full native 256K-token context (FP8 KV) + MTP working at 120-170 tokens/s, OpenAI-compatible API via vLLM, caddy Basic-auth gateway, and Prometheus + Grafana + DCGM monitoring.
One-click deployment of Qwen3.8-27B on a single RTX 3090 (24GB) on Windows 11: Q4_K_XL quantization + MTP speculative decoding + 128K context window, exposed as an OpenAI-compatible llama-server, averaging ~47 tok/s with ~1.5× lossless MTP speedup
🔒 A highly secure, local AI web interface optimized for Apple Silicon. Specifically designed for running Abliterated / Uncensored models (up to 27B+ parameters like Qwen3.8-27B) smoothly. Features End-to-End Encryption: Both your SQLite database and RAG embeddings are cryptographically locked at rest.
Run Qwen 3.8 27B locally with uncensored chat via Ollama, LM Studio, or llama.cpp — no API key needed.
To associate your repository with the qwen-38-27b topic, visit your repo's landing page and select "manage topics."