AirLLM 70B inference with single 4GB GPU
Pre-trained image models using ONNX for fast, out-of-the-box inference.
