Skip to content

Latest commit

 

History

History
26 lines (25 loc) · 3.71 KB

SupportMatrix.md

File metadata and controls

26 lines (25 loc) · 3.71 KB

Supported Optimization Features

Category Optimization API Alias
PyTorch Mixed Precision pytorch_amp
PyTorch Channels Last pytorch_channels_last
PyTorch JIT (Just-In-Time) Script/Trace & optimize_for_inference pytorch_jit_script, pytorch_jit_trace, pytorch_jit_script_ofi, pytorch_jit_trace_ofi
PyTorch JIT with TorchDynamo pytorch_torchdynamo_jit_script, pytorch_torchdynamo_jit_trace, pytorch_torchdynamo_jit_script_ofi, pytorch_torchdynamo_jit_trace_ofi
PyTorch Intel Neural Compressor (INC) Mixed Precision pytorch_inc_bf16
PyTorch INC INT8 Static Quantization (FX/IPEX) pytorch_inc_static_quant_fx, pytorch_inc_static_quant_ipex, pytorch_inc_static_quant_ipex_xpu
PyTorch INC INT8 Dynamic Quantization pytorch_inc_dynamic_quant
PyTorch Intel Extension for PyTorch (FP32, BF16, INT8 Static/Dynamic Quantization) pytorch_ipex_fp32, pytorch_ipex_bf16, pytorch_ipex_int8_static_quant, pytorch_ipex_int8_dynamic_quant
PyTorch Alibaba Blade-DISC pytorch_aliblade
PyTorch Lightning Mixed Precision pytorch_lightning_bf16_cpu
TensorFlow Mixed Precision tensorflow_amp
Keras Mixed Precision keras_amp
TensorFlow/Keras Model INC Quantization tensorflow_inc
Keras Script INC Quantization keras_inc
ONNX Runtime INC Static Quantization (QLinear) onnx_inc_static_quant_qlinear
ONNX Runtime INC Static Quantization (QDQ) onnx_inc_static_quant_qdq
ONNX Runtime INC Dynamic Quantization onnx_inc_dynamic_quant
HuggingFace Optimum-Intel INC Quantization pytorch_inc_huggingface_optimum_static, pytorch_inc_huggingface_optimum_dynamic
Intel Extension for Transformers INC Quantization intel_extension_for_transformers
BigDL Nano Optimization List nano_ + specific alias
Auto-Detect INC Quantization inc_auto