[INTEL_HPU] moved UT test index copy into UT test folder and added HPU index selected (#1632)

Signed-off-by: Luo, Focus <[email protected]>

Author: FocusLuo

backends/intel_hpu/tests/unittests/test_index_copy.py

@@ -0,0 +1,205 @@
+#   Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import unittest
+
+import numpy as np
+import torch
+import paddle
+import paddlenlp_ops
+
+from tests.op_test import skip_check_grad_ci
+
+import os
+
+intel_hpus_module_id = os.environ.get("FLAGS_selected_intel_hpus", 0)
+
+
+def index_copy_torch(input, dim, index, source, dtype):
+    dtype_map = {
+        "float16": torch.float16,
+        "float32": torch.float32,
+        "float64": torch.float64,
+        "int32": torch.int32,
+    }
+    torch_dtype = dtype_map[dtype]
+    input_tensor = torch.tensor(input).clone().detach().to(dtype=torch_dtype)
+    index_tensor = torch.tensor(index).clone().detach().to(dtype=torch.int64)
+    source_tensor = torch.tensor(source).clone().detach().to(dtype=torch_dtype)
+    output = torch.index_copy(
+        input=input_tensor, dim=dim, index=index_tensor, source=source_tensor
+    )
+    return output
+
+
+@skip_check_grad_ci(reason="index_copy_forward ops not support gradient calculation.")
+class TestIndexCopyOpFP32(unittest.TestCase):
+    def setUp(self):
+        self.place = paddle.CustomPlace("intel_hpu", int(intel_hpus_module_id))
+        self.init_dtype()
+        self.batch_size = 16
+        self.num_heads = 32
+        self.seq_length = 256
+        self.head_dim = 64
+
+    def init_dtype(self):
+        self.dtype = "float32"
+
+    def check_result(self, torch_res, ops_res):
+        if self.dtype == "float32":
+            rtol = 1e-5
+            atol = 1e-6
+        elif self.dtype == "float16":
+            rtol = 1e-3
+            atol = 1e-4
+        elif self.dtype == "bfloat16":
+            rtol = 1e-2
+            atol = 1e-3
+        else:
+            self.assertTrue(
+                False,
+                msg="index_copy input dtype only supports bfloat16, \
+                     float16 and float32, but got "
+                + self.dtype,
+            )
+        np.testing.assert_allclose(torch_res, ops_res, rtol=rtol, atol=atol)
+
+    def index_copy_custom(self, input, dim, index, source):
+        input_tensor = paddle.to_tensor(input, dtype=self.dtype).clone()
+        index_tensor = paddle.to_tensor(index, dtype="int64").clone()
+        source_tensor = paddle.to_tensor(source, dtype=self.dtype).clone()
+        paddlenlp_ops.index_copy_(
+            input=input_tensor, dim=dim, index=index_tensor, source=source_tensor
+        )
+        return input_tensor
+
+    def prepare_input(
+        self, batch_size=16, num_heads=32, seq_length=256, head_dim=64, dim=0, index=0
+    ):
+        self.batch_size = batch_size
+        self.num_heads = num_heads
+        self.seq_length = seq_length
+        self.head_dim = head_dim
+
+        input = np.full(
+            (num_heads, head_dim, seq_length, batch_size), -1, dtype=self.dtype
+        )
+        index = [index]
+        if dim == 0:
+            source = np.full(
+                (len(index), head_dim, seq_length, batch_size), 0, dtype=self.dtype
+            )
+        elif dim == 1:
+            source = np.full(
+                (num_heads, len(index), seq_length, batch_size), 0, dtype=self.dtype
+            )
+        elif dim == 2:
+            source = np.full(
+                (num_heads, head_dim, len(index), batch_size), 0, dtype=self.dtype
+            )
+        elif dim == 3:
+            source = np.full(
+                (num_heads, head_dim, seq_length, len(index)), 0, dtype=self.dtype
+            )
+        else:
+            raise ValueError(
+                "Unsupported dimension. Only dim=0, dim=1 and dim=2 are supported."
+            )
+        return input, index, source, dim
+
+    def test_index_copy_dim0_index0(self):
+        input, index, source, dim = self.prepare_input(dim=0, index=0)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res)
+
+    def test_index_copy_dim0_index1(self):
+        input, index, source, dim = self.prepare_input(dim=0, index=1)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res)
+
+    def test_index_copy_dim0_index_max(self):
+        index = max(self.num_heads - 1, 0)
+        input, index, source, dim = self.prepare_input(dim=0, index=index)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res)
+
+    def test_index_copy_dim1_index0(self):
+        input, index, source, dim = self.prepare_input(dim=1, index=0)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res)
+
+    def test_index_copy_dim1_index1(self):
+        input, index, source, dim = self.prepare_input(dim=1, index=1)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim1_index_max(self):
+        index = max(self.head_dim - 1, 0)
+        input, index, source, dim = self.prepare_input(dim=1, index=index)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim2_index0(self):
+        input, index, source, dim = self.prepare_input(dim=2, index=0)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim2_index1(self):
+        input, index, source, dim = self.prepare_input(dim=2, index=1)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim2_index_max(self):
+        index = max(self.seq_length - 1, 0)
+        input, index, source, dim = self.prepare_input(dim=2, index=index)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim3_index0(self):
+        input, index, source, dim = self.prepare_input(dim=3, index=0)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim3_index1(self):
+        input, index, source, dim = self.prepare_input(dim=3, index=1)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+    def test_index_copy_dim3_index_max(self):
+        index = max(self.batch_size - 1, 0)
+        input, index, source, dim = self.prepare_input(dim=3, index=index)
+        custom_res = self.index_copy_custom(input, dim, index, source)
+        torch_res = index_copy_torch(input, dim, index, source, dtype=self.dtype)
+        self.check_result(torch_res.numpy(), custom_res.numpy())
+
+
+@skip_check_grad_ci(reason="index_copy_forward ops not support gradient calculation.")
+class TestIndexCopyOpFP16(TestIndexCopyOpFP32):
+    def init_dtype(self):
+        self.dtype = "float16"
+
+
+if __name__ == "__main__":
+    unittest.main()

backends/intel_hpu/tests/unittests/test_index_put_op.py

@@ -21,6 +21,10 @@
 import paddle
 import paddle.base as base
 
+import os
+
+intel_hpus_module_id = os.environ.get("FLAGS_selected_intel_hpus", 0)
+
 
 class TestIndexPut(unittest.TestCase):
     def setUp(self):
@@ -57,7 +61,9 @@ def test_dygraph(self):
         index = [paddle.to_tensor([1, 1]), paddle.to_tensor([3, 4])]
         index_np = np.array([[1, 1], [3, 4]])
 
-        paddle.disable_static(paddle.CustomPlace("intel_hpu", 0))
+        paddle.disable_static(
+            paddle.CustomPlace("intel_hpu", int(intel_hpus_module_id))
+        )
         with base.dygraph.guard():
             input_np = convert_float_to_uint16(
                 np.random.random(self.shape).astype(self.dtype)