fix ci

Signed-off-by: Yaoyao Ding <[email protected]>

Author: yaoyaoding

tests/instructions/test_reduce.py

@@ -33,7 +33,7 @@ def __call__(self, out_ptr: ~int32):
             dtype=int32,
             shape=self.layout.shape,
             layout=self.layout,
-            init=lambda indices: indices[0] * self.layout.shape[1] + indices[1],
+            init=lambda i, j: i * self.layout.shape[1] + j,
         )
         b = self.sum(a, dim=self.dim, keepdim=True)
         g_out = self.global_view(ptr=out_ptr, dtype=int32, shape=b.shape)

tests/kernels/matmul/test_matmul_v0.py

@@ -36,7 +36,7 @@ def __call__(self, m_size: int32, n_size: int, k_size: int, a_ptr: ~float16, b_p
 
         ga = self.global_view(a_ptr, dtype=float16, shape=[m_size, k_size])
         gb = self.global_view(b_ptr, dtype=float16, shape=[k_size, n_size])
-        acc = self.register_tensor(dtype=float32, shape=[self.block_m, self.block_n], init=lambda indices: float32.zero)
+        acc = self.register_tensor(dtype=float32, shape=[self.block_m, self.block_n], init=lambda i, j: float32.zero)
 
         k_blocks = self.utils.ceil_div(k_size, self.block_k)
         for k in range(k_blocks):

tests/lang/test_simple_mma_matmul.py

@@ -37,7 +37,7 @@ def __call__(self, m_size: int32, a_ptr: ~float16, b_ptr: ~float16, c_ptr: ~floa
         offset_m: int32 = self.block_m * self.blockIdx.x
         offset_n: int32 = self.block_n * self.blockIdx.y
 
-        acc = self.register_tensor(dtype=float32, shape=self.mma.lc.shape, init=lambda indices: float32.zero)
+        acc = self.register_tensor(dtype=float32, shape=self.mma.lc.shape, init=lambda i, j: float32.zero)
         k_blocks = self.utils.ceil_div(self.k_size, self.block_k)
         for k in range(k_blocks):
             offset_k = k * self.block_k