Restore stable PyTorch CPU + 4B (backup while INT8 ONNX upload in progress)

Stable: PyTorch CPU-only + Qwen3-Embedding-4B (proven working)

Add ONNX Runtime backend option + ST fallback for Qwen3-Embedding-4B

TEI Candle CPU + Qwen3-Embedding-4B fp16

Switch to sentence-transformers + Qwen3-Embedding-4B (PyTorch CPU)

Switch to TEI cpu-1.9 for Qwen3-Embedding-8B

Add Qwen3-Embedding-8B API