Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg merged PR #16563: URL: https://github.com/apache/tvm/pull/16563 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apach

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on PR #16563: URL: https://github.com/apache/tvm/pull/16563#issuecomment-1955086643 All CI tests passing, and thank you for the review @slyubomirsky ! I'll follow up with another PR to add inspection of the remainder of the `DLTensor*` fields. -- This is an automate

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491893261 ## python/tvm/relax/expr.py: ## @@ -244,6 +244,192 @@ def __getitem__(self, index: int) -> "ExprWithOp": raise IndexError from err raise

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491708611 ## src/relax/op/tensor/unpack.cc: ## @@ -0,0 +1,349 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements.

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491708293 ## python/tvm/relax/expr.py: ## @@ -244,6 +244,192 @@ def __getitem__(self, index: int) -> "ExprWithOp": raise IndexError from err raise

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491707010 ## src/relax/transform/legalize_ops.cc: ## @@ -157,17 +167,72 @@ class LegalizeMutator : public ExprMutator { if (op_node == nullptr) { return visited_call

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491670908 ## src/relax/transform/legalize_ops.cc: ## @@ -157,17 +167,72 @@ class LegalizeMutator : public ExprMutator { if (op_node == nullptr) { return visited_call

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491661307 ## python/tvm/relax/expr.py: ## @@ -244,6 +244,192 @@ def __getitem__(self, index: int) -> "ExprWithOp": raise IndexError from err raise

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg commented on PR #16563: URL: https://github.com/apache/tvm/pull/16563#issuecomment-1947309703 > The implementation seems solid, a very good change. I like the parameterized test cases too. The use of the _DLTensorShapeProxy resulted in an elegant UI in tvmscript. Thank you

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

tqchen commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1491520155 ## src/relax/op/tensor/unpack.cc: ## @@ -0,0 +1,349 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See

Re: [PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

slyubomirsky commented on code in PR #16563: URL: https://github.com/apache/tvm/pull/16563#discussion_r1490323766 ## src/relax/op/tensor/unpack.cc: ## @@ -0,0 +1,349 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements.

[PR] [Relax] Implement operators to read runtime DLTensor* information [tvm]

Lunderberg opened a new pull request, #16563: URL: https://github.com/apache/tvm/pull/16563 Relax is capable of expressing tensors whose element type is unknown. However, these must typically be replaced with a known dtype prior to compilation, as most operators require known data types pr