We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Updated Home (markdown)
Updated Adding a new element wise (unary) tensor operation (markdown)
Created Adding a new element wise (unary) tensor operation (markdown)
Updated Exporting MultiHeadAttention to pytorch for unit tests (markdown)
Created Exporting MultiHeadAttention to pytorch for unit tests (markdown)
Created Can gradients be tracked in both Module::forward() and Module::forward_mut()? YES (markdown)
Created Do I call Module::forward() or Module::forward_mut()? (markdown)
Updated Disabling gradient tracking (torch.no_grad()) (markdown)
Updated Tapes & tracking gradients (markdown)
Updated Is NoTape & NoneTape equivalent to torch.no_grad()? (markdown)
Created Disabling gradient tracking (torch.no_grad()) (markdown)
Updated Can a <TensorXD ..., NoneTape> have gradients recorded? YES (markdown)
Updated Can a \ TensorXD ..., NoneTape have gradients recorded? YES (markdown)
Updated Can a \ TensorXD ..., NoneTape\ have gradients recorded? YES (markdown)
Updated Can a `\ TensorXD ..., NoneTape\ ` have gradients recorded? YES (markdown)
Updated Can a `TensorXD ..., NoneTape ` have gradients recorded? YES (markdown)
Created Can a `TensorXD<..., NoneTape>` have gradients recorded? YES (markdown)
Updated Is `NoTape` `NoneTape` equivalent to `torch.no_grad()`? No! (markdown)
Created Is `NoTape`/`NoneTape` equivalent to `torch.no_grad()`? No! (markdown)
Initial Home page