Quantized ONNX (CPU) pplx-embed-v2-late 0.6B and 9B: one token space, so index with the 9B and query with the 0.6B.
Nynefold
community
AI & ML interests
Datasets, quantization, model compression, open-source tools...
Recent Activity
View all activity
Quantized ONNX (CPU) pplx-embed-v2-late 0.6B and 9B: one token space, so index with the 9B and query with the 0.6B.
Quantized ONNX embedders and rerankers with full quality tables, plus the FreshRet Small eval set.
37 GGUF quant recipes of MOSS-Image-SFT (DiT + Qwen3.5-9B TE), each measured against FP32 on the same prompts and seeds.