intel / neural-compressor
SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, and ONNX Runtime
- Repository stars
- 2,729All-time #21,795
- Fork
- 332
- Last push
- updated 1 days ago
- Repository created
- since 2020
About
SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, and ONNX RuntimeSummarised by AI from the README; the original prevails
Author description: SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, and ONNX Runtime
Basic listing based on official public information; functionality has not been independently evaluated.
- Pricing
- Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.