fix runtime error (#47133)

paddle/fluid/framework/ir/ipu/infer_shape_pass.cc

@@ -44,7 +44,8 @@ void InferShapePass::ApplyImpl(ir::Graph* graph) const {
         feed_list.end();
     if (is_feed) {
       auto input_shape = node->Var()->GetShape();
-      if (input_shape[0] <= -1) {
+      // NOTE: some tensors may be 0-dim tensors
+      if (!input_shape.empty() && input_shape[0] <= -1) {
         input_shape[0] = micro_batch_size;
         node->Var()->SetShape(input_shape);
         need_infer_shape = true;

python/paddle/fluid/tests/unittests/ipu/op_test_ipu.py

@@ -245,6 +245,9 @@ class IPUOpTest(IPUTest):
             raise ValueError("output_dict is empty")
         cpu_fp32 = output_dict[ExecutionMode.CPU_FP32]
         ipu_fp32 = output_dict[ExecutionMode.IPU_FP32]
+        # Convert 0-dim tensor
+        if isinstance(cpu_fp32, np.ndarray) and cpu_fp32.shape == ():
+            cpu_fp32 = cpu_fp32.reshape(1)
         if len(cpu_fp32) != len(ipu_fp32):
             raise ValueError("different outputs number between ipu and cpu.")
         for cpu_fp32_res, ipu_fp32_res in zip(cpu_fp32, ipu_fp32):

python/paddle/fluid/tests/unittests/ipu/test_transpose_op_ipu.py

@@ -83,6 +83,7 @@ class TestCase_ZeroDim(TestBase):
     def set_data_feed(self):
         data = np.random.uniform(size=[])
         self.feed_fp32 = {"x": data.astype(np.float32)}
+        self.feed_fp16 = {"x": data.astype(np.float16)}
 
     def set_op_attrs(self):
         self.attrs = {"perm": []}