magiccodingman/Jasper-Token-Compression-600M-ONNX-FP32

1---2license: apache-2.03language:4- en5- zh6library_name: sentence-transformers7pipeline_tag: sentence-similarity8base_model: infgrad/Jasper-Token-Compression-600M9tags:10- sentence-transformers11- feature-extraction12- sentence-similarity13- text-embeddings14- embeddings15- retrieval16- mteb17- qwen318- onnx19- onnxruntime20- cpu21- fp3222---23 24# Jasper Token Compression 600M — ONNX FP3225 26ONNX export of [infgrad/Jasper-Token-Compression-600M](https://huggingface.co/infgrad/Jasper-Token-Compression-600M).27 28**Precision:** FP32  29**Quantization:** None (FP32)  30**Model size:** 2.26 GiB31 32Full-precision FP32 ONNX export. This is the reference model used for the included quantization fidelity comparisons.33 34## Benchmarks35 36| Tokens | Median latency | Tokens/s |37|---:|---:|---:|38| 32 | 139.282 ms | 229.7 |39| 128 | 155.466 ms | 823.3 |40| 512 | 211.763 ms | 2,417.8 |41| 1024 | 288.388 ms | 3,550.8 |42 43## Attribution44 45Original model: `infgrad/Jasper-Token-Compression-600M`46