danielholanda / LeFlow

Enabling Flexible FPGA High-Level Synthesis of Tensorflow Deep Neural Networks
Other
588 stars 101 forks source link

Legup failed to generate Verilog #29

Closed mohitajais closed 4 years ago

mohitajais commented 4 years ago

Hello, I am using this tool to learn how tensor flow code is converted into verilog and generating hardware circuit. Unfortunately, I am facing some errors in generating verilog file. I have followed all the steps which was mentioned in readme file. I have successfully install LeFlow with the modified version of TensorFlow and I ran the test_all .py file with no errors. But I get the following error when I run classificationMNIST example

legup@legup-vm:~/Downloads/LeFlow-master/LeFlow-master/examples/classificationMNIST$ ../../src/LeFlow classificationMNIST.py INFO: Creating project... INFO: Cleaning previous files... INFO: Generating IR from tensorflow... Extracting ./MNIST_data/train-images-idx3-ubyte.gz Extracting ./MNIST_data/train-labels-idx1-ubyte.gz Extracting ./MNIST_data/t10k-images-idx3-ubyte.gz Extracting ./MNIST_data/t10k-labels-idx1-ubyte.gz The accuracy over the MNIST data is 92.14% Expected Result: 6 INFO: Cleaning unused dumped files... INFO: Converting between LLVM versions... INFO: Running first batch of optimizations before unrolling... INFO: Unrolling and Inlining according to user, simplifying elementary branches and optimizing away other values... INFO: Restructuring the IR signature... INFO: Rewriting unsupported operations... INFO: Partitioning arrays... User did not specify any arrays to partition INFO: Converting human-readable .ll file to bitcode... INFO: Starting LegUp Compilation... ERROR: Legup failed to generate Verilog INFO: Exiting LeFlow. It is not generating verilog file. I have seen previous issues related to my problem but I have not been able to fix my error with their solutions.

Could you please suggest any solution to this issue?

Thanks for the help!!!

danielholanda commented 4 years ago

Hi,

Sorry for the terrible delay in replying here. I just tested it in my machine and it works fine. It also doesn't look like there are any problems with Tensorflow.

One thing that you can try is to uninstall the Tensorflow whl file of the current LeFlow version and install an old one (link here). This version here used to be pretty stable.