rtx3050

MacOS on Lenovo Legion 5-15ACH6 - Ryzen 5600H & GeForce RTX 3050

DeepInfer SM86 native Qwen2.5 kernels plus a local OpenAI-compatible productivity gateway

GPU-accelerated local AI assistant using Phi-3 3.8B + Ollama + RTX 3050. Runs 100% offline.

AI Image & Video Generator optimized for RTX 3050. Generate images locally, create videos via API, with built-in 4K upscaling.

Purpose-built inference engine for Mistral Shieldstral 1.0 3B safety classifier. Optimized for RTX 3050 4GB VRAM. Custom CUDA kernel, INT4 quantization, multimodal vision pipeline, FastAPI moderation API.

Qwen2.5-7B-Instruct DeepInfer SM86 mixed W4/W3, fused qkv/gate-up loader, 256 GQA; weights in v0.1.0 release assets

Qwen2.5-3B-Instruct + maibot-catgirl-v3, DeepInfer SM86 W4A16, fused qkv/gate-up loader, 256 GQA; weights in v0.1.0 release assets

Video-to-G1 dance in mjlab with SuperMimic fine-tuning on RTX 3050 4GB; reproducible tools, results and Chinese beginner tutorial.