Benjamin Ramhorst
Benjamin Ramhorst
> The reuse factor is capped at the number of parameters for each layer, correct? I have valid reuse factors ranging from 1, 2, ..., to 18432 for a conv2D...
Unfortunately, there is no more documentation than the tutorials, website and papers. Please have a look here: https://github.com/fastmachinelearning/hls4ml/blob/main/hls4ml/templates/vivado/nnet_utils/nnet_dense_latency.h. The code is short and the most important part of it is...
> I'm fine with this type of change. But last I remember, im2col often didn't even synthesize so we added some heuristic to at least try winograd if possible. @bo3z...
@maximilianheer can you help with this?
Writeback is a memory mapped region. It's a set of control registers for each vFPGA that gets incremented every time a data transfer is completed (there are dedicated registers for...
Closing due to inactivity. Please re-open at a later stage if still applicable.
The 8K bug has been fixed in #86. The issues with reads where due to Coyote background processed being terminated before finishing the test. @maximilianheer has debugged this and created...
Can you please add a few more details? Output from which is not zero: CSim (csim=True in hls4ml.build(...)) or the hls_prediction? Does the line np.testing.assert_allclose fail or something later?
Unfortunately, we don't provide examples of how to integrate IPs into larger application, as this is application-specific and can be done in many ways: 1. You could stream the data...
Once you have the IP, it doesn't matter what the underlying architecture is (GRU or CNN or FC etc.). At that point it's a matter of properly connecting all the...