AIMET TensorFlow Quantization APIs

AIMET Quantization for TensorFlow provides the following functionality

• Quantization Simulation: Allows ability to simulate inference and training on quantized hardware

• QuantAnalyzer: Analyzes the model and points out sensitive ops to quantization

• Adaptive Rounding: Post-training quantization technique to optimize rounding of weight tensors

• Cross-Layer Equalization: Post-training quantization technique to equalize layer parameters

• Bias Correction: Post-training quantization technique to correct shift in layer outputs due to quantization noise

• AutoQuant API: Unified API that integrates the post-training quantization techniques provided by AIMET

• BN Re-estimation APIs: APIs that Re-estimate BN layers’ statistics and fold the BN layers