Tensordict compiles masked indexing under fullgraph; ExecuTorch links WebGPU into the Wasm build
- [BugFix] Compile td[mask] and td[mask] = 0 with fullgraph, and write a number into a tensordict holding MetaData pytorch/tensordict
- [Compile] Compile permute and get with a default on a lazy stack, and in-place math on a TypedTensorDict over one pytorch/tensordict
- [Performance] Make `pad` faster #2861 pytorch/tensordict