Upload Qwen3-Embedding-0.6B Int8 ONNX + tokenizer + inference example

initial commit