Method relu

tests/python/relax/test_transform_fuse_ops.py:915–921

(A: T.Buffer((T.int64(1), T.int64(512), T.int64(64), T.int64(64)), "float32"), B: T.Buffer((T.int64(1), T.int64(512), T.int64(64), T.int64(64)), "float32"))

        @T.prim_func(private=True, s_tir=True)
        def relu(A: T.Buffer((T.int64(1), T.int64(512), T.int64(64), T.int64(64)), "float32"), B: T.Buffer((T.int64(1), T.int64(512), T.int64(64), T.int64(64)), "float32")):
            for i0, i1, i2, i3 in T.grid(T.int64(1), T.int64(512), T.int64(64), T.int64(64)):
                with T.sblock("relu"):
                    v_i0, v_i1, v_i2, v_i3 = T.axis.remap("SSSS", [i0, i1, i2, i3])
                    T.reads(A[v_i0, v_i1, v_i2, v_i3])
                    T.writes(B[v_i0, v_i1, v_i2, v_i3])
                    B[v_i0, v_i1, v_i2, v_i3] = T.max(A[v_i0, v_i1, v_i2, v_i3], T.float32(0))

    @I.ir_module(s_tir=True)
    class Expected: