Daniel Falbel
Daniel Falbel
It depends on how you generated the file. If you used `torch_save` this is expected, because `torch_save` actually saved as an RDS file with small differences to correctly serialize the...
basically you need to `jit_trace` a function that just calls your model. but before that you have to detach the model parameters..
Sorry, I don't follow... Are you trying to save in python and load in R or save in R and load into python? Python `torch.save` is also not compatible with...
Are you using the M1 or intel installation of R? I don't have a M1 mac for testing, but AFAICT, running torch via rosetta2 should work fine. Installing with M1...
Yeah, I think this has two consequences: - supporting M1 natively will have a 5x performance increase - so we have greater incentives for doing it. - it seems that...
Hi @dkgaraujo ! I am not sure if they are going to enable it for the released version but it should be possible if compiling PyTorch from source as per...
Basic support has been implemented in #890.
A quick fix is to install libomp from homebrew: ``` brew install libomp ``` But that shouldn't be necessary as we include a copy of libomp.dylib in the package. Can...
Thank you @juliasilge ! The current dev version should work now, as we no longer link to openMP.
This works in theory: ``` r library(torch) x 0 #> 2 #> 3 #> [ CPUFloatType{3} ] ``` Created on 2020-10-19 by the [reprex package](https://reprex.tidyverse.org) (v0.3.0) I wonder if `model$lhs_emb$weight`...