finn-base
finn-base copied to clipboard
Xilinx
Open Source Compiler Framework using ONNX as Frontend and IR
`finn-base[onnx]` is pinned to use `onnx==1.7.0`, which is not compatible with `protobuf>3.20`. This is solved in `onnx==1.12` by introducing an upper constraint on the `protobuf` version (https://github.com/onnx/onnx/issues/4222) Possible solutions would...
If I create an onnx file with this sample script and [input.txt](https://github.com/Xilinx/finn-base/files/8809937/input.txt): ```python import torch import torch.nn as nn import torch.nn.functional as F import numpy as np # Define simple...
In the transform member function of ModelWrapper: https://github.com/Xilinx/finn-base/blob/2f10c0c1f11cc483f27794887c41030bdd0dfab5/src/finn/core/modelwrapper.py#L122-L146 if `cleanup=True`, then the transform is called with `fix_float64=True` during the cleanup, no matter what the transform was originally called with. I...
Add infrastructure for co-simulating AXI MM slave memory. This is achieved by creating a simulation of an open-source AXI MM slave and manually carrying the relevant i/o signals between the...
This PR makes remote_exec.py and throughput_test.py compatible with PYNQ v2.7.0 - finn-base tries to call python3.6 on PYNQ board when PYNQ board has only python3 or python3.8 - finn-base does...
When querying the output datatype of `Quant` nodes, the current implementation only supports identifying quantization cases the corespond to integers. Everything else is set to `FLOAT32`. In cases where the...
It is currently not possible to easily remove the shape from a given tensor. The attached .zip file contains an QONNX model, which has unset tensors: [CNV_2W2A.zip](https://github.com/Xilinx/finn-base/files/6961855/CNV_2W2A.zip) Running the following:...
This PR adds support to generate LogicNets style networks. - BinaryTruthTable customOp - TruthTable customOp - Python simulation - RTL simulation - Verilog generation transformation - PLA generation transformation -...
