Daniel Falbel
Daniel Falbel
``` # torch_arange(1, t, dtype = "int") # torch_tensor # 1 # 2 # 3 # [ CPUIntType{3} ] # torch_arange(1, t, dtype = "long") # torch_tensor # 1 #...
This issue aggregates some larger features that would be great additions to the torch package. - [ ] #738 - [ ] #928 - [x] #929 (https://github.com/mlverse/torch/pull/1020, https://github.com/mlverse/torch/pull/1023) - [...
``` # This returns only weight1 net$named_parameters(TRUE) #> $weight1 #> torch_tensor #> -2.2368 #> [ CPUFloatType{1,1} ][ requires_grad = TRUE ] # This returns weight1 and weight2 net$named_parameters(FALSE) ``` This...
``` # This does not work, because `$register_parameter()` is not called. # I would expect it to work though or at least throw a warning net$weight1 = torch_clone(net$weight1) + 1000...
It would be nice to have support for `vmap` on the R side. It seems that this is mainly implemented in LibTorch so we should be able to have that...
WIP! Add a tabulate based cuda memory summary.
See [wordtree in google charts website](https://developers.google.com/chart/interactive/docs/gallery/wordtree#overview)
It's not easy to write a simple reprex for this, but you can comment out the `update` method of https://github.com/mlverse/tabnet create a model with at least one `tune()` and try...
Looks like `jpeg` and `png` are faster than magick for reading images. We should probably use them instead: file # A tibble: 2 x 6 #> expression min median `itr/sec`...