The GitHub Actions job "CI" on tvm.git/fix-onnx-binary-symbolic-primexpr has failed. Run started by GitHub user Irfanuet191 (triggered by Irfanuet191).
Head commit for run: 5034b6212799fe27f07b03cc7c0085421be2d8d6 / Irfanuet191 <[email protected]> [Fix][Relax][ONNX] Fold binary ops on symbolic PrimExpr scalars without numpy dispatch BinaryBase.base_impl folds scalar operands with numpy, but when an operand is a symbolic PrimExpr (e.g. a dynamic dimension gathered out of a Shape result), the numpy ufunc dispatches through object dtype: comparisons and logical ops coerce the result through bool(), which PrimExpr forbids, Pow/Mod fail outright, and older releases crashed in Sub with AttributeError: 'Sub' object has no attribute 'item'. Route symbolic scalar operands around numpy instead: each BinaryBase subclass declares a prim_op that builds the scalar PrimExpr directly, and base_impl wraps the result in relax.prim_value. Concrete operands keep the existing numpy folding. Also promote Div's scalar-conversion helper to module level, compute integer Pow in float64 and cast back, and make BitwiseBase dtype validation accept PrimExpr operands. Fixes #20065 Report URL: https://github.com/apache/tvm/actions/runs/31583529681 With regards, GitHub Actions via GitBox --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
