Hey everyone, I’m Ali, a software engineer from Nuro, and we’ve been working on migrating a lot of vision models from tensorflow to pytorch. One particularly critical operation for these vision models is tf.tensor_scatter_nd_update , and its reduction counterparts (e.g. tf.tensor_scatter_nd_min, tf.tensor_scatter_nd_max , etc).
We’ve implemented this operation in-house and have been using it successfully for a while now. There have been a few requests for an official pytorch implementation for this operation for a while now (e.g. here, here, etc), so we’d like to offer to set this up.
Ask: Any strong objections to adding this new operation? And if not, who are the best reviewers to ping on the change?
The implementation uses only torch native layers, supports torch.compile and torch-xla, and is highly extensible. Happy to provide more details as needed!
