Stable

including ollama, vllm, xft, ipex-llm, tensorrt-llm

technical solution including ASR, TTS, LLM, lipsync

memory bandwidth benchmark

LLM run on MTL iGPU or Arc DGPU Windows with audio in and out

enable multi-dGPU on Xeon server such as 2 Arc A770

codeshell-7B-Chat run on MTL iGPU by VS Code with 32GB memory

LLM run on DGPU with audio in and out

chatglm2-6b, llama2-13b, starcoder-15.5b run on Arc DGPU

run wenet by openvino 2023.1

LLM run on CPU with 16GB memory

benchmark test for LLM INT4/FP16 on Intel Core, SPR, Arc.

run mlc-llm on Intel Arc DGPU

support static shape like ONNX, Tensorflow, Yolov5.pt or 2 inputs

test all .xml file in current document

microphone in (pyaudio) > ASR (whisper-medium) > chat (chatglm2-6b)

auto test .exe in Windows, include button, slide, text number, checkbox and page.

Inference YOLOX model by DLStreamer with display or print results to JSON file

train Resnet50 by Intel®-Extension-for-TensorFlow*(iTEX) on multiply Intel Arc DGPU

Test Decode/Encode performance on IA platform

Hello AI