Inference Engine CPU plugin can infer models in the 8-bit integer (INT8) precision. For details, refer to INT8 inference on the CPU.
Intermediate Representation (IR) should be specifically formed to be suitable for the INT8 inference. Such an IR is called an INT8 IR and you can generate it in two ways:
.pbmodel file with
For an operation to be executed in INT8, it must have
FakeQuantize operations as inputs with the
levels attribute set to
256. See the specification of `FakeQuantize` operation for details. To see the list of supported INT8 layers, refer to INT8 inference on the CPU.
To execute the
Convolution operation in INT8 on CPU, both data and weight inputs should have
FakeQuantize as an input operation:
INT8 IR is also suitable for FP32 and FP16 inference if a chosen plugin supports all operations of the IR, because the only difference between an INT8 IR and FP16 or FP32 IR is the existence of
FakeQuantize in the INT8 IR. Plugins with the INT8 inference support recognize these sub-graphs and quantize them during the inference time. Plugins without the INT8 support execute all operations, including
FakeQuantize, as is in the FP32 or FP16 precision.
Accordingly, the presence of FakeQuantize operations in the IR is a recommendation for a plugin on how to quantize particular operations in the model. If capable, a plugin accepts the recommendation and performs the INT8 inference, otherwise the plugin ignores the recommendation and executes a model in the floating-point precision.
Weighted operations, like
MatMul, and others, store weights as floating-point
Constant in the graph followed by the
Constant followed by the
FakeQuantize operation could be optimized memory-wise due to the
FakeQuantize operation semantics. The resulting weights sub-graph stores weights in INT8
Constant, which gets unpacked back to floating point with the
Convert operation. Weights compression leaves
FakeQuantize output arithmetically the same and weights storing takes four times less memory.
See the visualization of
Convolution with the compressed weights:
Both Model Optimizer and Post-Training Optimization tool generate a compressed IR by default. To generate an expanded INT8 IR, use