shaltielshmid
shaltielshmid
Alternative suggestion: Create a normalizer + decoder for doing this (assuming `BytesToUnicodeDict` and `UnicodeToBytesDict` is a static dictionary created using the code above). ```cs public class BytesToUnicodeNormalizer : Normalizer {...
Will do!
Yes, we are working on an update which should make all the attributes available
Implicit casting could be a cool feature, for now you can work around it by doing: ```cs var indexed = x[..,TensorIndex.Tensor(some_tensor),..]; ```
I haven't gotten up to it yet, but feel free to go ahead :)
In TorchSharp it would look the same: ```cs var model = ...; foreach (var param in model.parameters()) param.requires_grad = False; ```
@LittleLittleCloud I haven't tried it myself, but have you tried loading using [TorchSharp.PyBridge](https://github.com/shaltielshmid/TorchSharp.PyBridge)? You can install it using nuget: ```powershell Install-Package TorchSharp.PyBridge ``` And then you can load in the...
`TorchSharp.PyBridge` is dependant on features that were added only in version 0.101.5 of TorchSharp. But, since the TorchSharp package includes the cuda binaries already, you can update the package even...
Hi @yueyinqiu ! Thanks for writing up the diff. I went through all the methods/modules that were missing in TorchSharp, and they are all pure PyTorch code that don't appear...
One options that might be possible - if we are using PyBridge to load in the weights, maybe we can just link directly to the storage that PyTorch uses and...