qwen3-5

An open-source implementaion for fine-tuning Qwen-VL series by Alibaba Cloud.

RAG (Retrieval-augmented generation) ChatBot that provides answers based on contextual information extracted from a collection of Markdown files.

Qwen3.5-122B-A10B on DGX Spark: 28.3 → 51 tok/s (+80%)

Run generative AI models in sophgo BM1684X/BM1688

Private on-device AI chat for Android — runs any GGUF model locally via llama.cpp with ARM-optimised SIMD. Zero network permissions, encrypted settings, biometric lock, tamper detection. + GPU Acceleration

MCore-Bridge: Providing Megatron-Core model definitions for state-of-the-art large models and making Megatron training as simple as Transformers — with support for 300+ large language models (Qwen3-Next, GLM-5.2, Deepseek-V4, MiniMax-2.7, ...) and 200+ multimodal large models (Qwen3.5, Qwen3-Omni, Gemma4, ...).

Fine-tune a local LLM on your own app's data in 15 minutes. Runs entirely on-device, zero cloud after training. Apple Silicon + CUDA.

7.67× LoRA / 8.35× Full FT speedup for Qwen3.5 (0.8B–27B) on NVIDIA DGX Spark — wall-clock parity with rented H100. Lossless within BF16. Three-command interactive wizard handles model picker, data validator, training, and merge.