Ah, just noticed you’re running on Cpu with tanh activations. Flux vs pytorch cpu performance is most likely the culprit (long story short, small dense MLPs with tanh on CPU hit a bunch of areas in Flux that need to be optimized), except more or less pronounced because you’re also running the backwards pass.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Slow LSTM on GPU in Flux | 21 | 2304 | February 15, 2024 | |
| Why is this MLP slower in Flux than in TensorFlow? | 5 | 1181 | April 30, 2022 | |
| Why is flux model slower than python? | 6 | 2375 | May 11, 2024 | |
| Flux.jl RNN performance | 11 | 3234 | October 28, 2018 | |
| Flux vs pytorch cpu performance | 59 | 9813 | October 2, 2020 |