compute-engine
compute-engine copied to clipboard
larq
Highly optimized inference engine for Binarized Neural Networks
Hello, I have a couple of questions regarding quantizer options for Larq and LCE. I am designing a BNN using the DoReFa quantizer, however, I noticed a very high number...
I tried using python 3.6 + LCE 0.6.2 and python 3.7/3.8 + LCE 0.7.0 to run the following code, and the tflite file generated has unexpected sizes: For python 3.6...
The LCE interpreter from `lce.testing.Interpreter` is a standalone class and exposes different properties of the quantized model (scale and zero-point for example). the converter on the other hand is built...
Binary fully connected operator is in essence doing binary matrix matrix multiplication (BGemm). Assume that the input is M × N , the weight is N×K (M is the batch...
@honglh added optimized kernels for ARM32 in #432. It would be great if we could add support for bitpacked activations to them to match the AArch64 optimized bgemm kernels.
*This is very much a draft PR.* ## What do these changes do? This is a converter-only change that adds support for binary convolutions with unipolar inputs -- all credit...
We should work out a good way to raise warnings for models which won't convert in an 'optimal' way -- essentially any eggregious violation of our [model optimisation guide](https://docs.larq.dev/compute-engine/model_optimization_guide/). It's...
Since #384 we can make use of the [IR verification](https://mlir.llvm.org/docs/OpDefinitions/#custom-verifier-code) features of MLIR. Our BConv op has a few parameter combinations where [it can throw during `Init` or `Prepare`]( https://github.com/larq/compute-engine/blob/master/larq_compute_engine/tflite/kernels/bconv2d.cc)....
## Observed behaviour When converting this model... ```python model = tf.keras.models.Sequential([ tf.keras.Input((32, 32, 3)), lq.layers.QuantConv2D( 32, (3, 3), input_quantizer="ste_sign", kernel_quantizer="ste_sign", padding="same", pad_values=1.0, use_bias=False ), tf.keras.layers.Conv2D(32, (3, 3)), ]) converted_model =...
TFLite provides a [collection of example mobile apps](https://www.tensorflow.org/lite/examples) for specific DL applications. We already have a setup in place to build LCE Lite AAR to support Android (see #238) ....