#7826: Add doc for abs (#13171)

* #7826: Add doc for abs

* #7826: Add golden function

---------

Co-authored-by: mcw-anasuya <[email protected]>

.github/workflows/ttnn-run-sweeps.yaml

@@ -21,6 +21,7 @@ on:
           - eltwise.unary.gelu.gelu_pytorch2
           - eltwise.unary.hardsigmoid.hardsigmoid_pytorch2
           - eltwise.unary.leaky_relu.leaky_relu_pytorch2
+          - eltwise.unary.abs.abs
           - eltwise.unary.cos.cos
           - eltwise.unary.cos.cos_pytorch2
           - eltwise.unary.sin.sin

tests/sweep_framework/sweeps/eltwise/unary/abs/abs.py

@@ -0,0 +1,83 @@
+# SPDX-FileCopyrightText: © 2024 Tenstorrent Inc.
+
+# SPDX-License-Identifier: Apache-2.0
+
+from typing import Optional, Tuple
+from functools import partial
+
+import torch
+import random
+import ttnn
+from tests.sweep_framework.utils import gen_shapes
+from tests.tt_eager.python_api_testing.sweep_tests.generation_funcs import gen_func_with_cast_tt
+
+from tests.ttnn.utils_for_testing import check_with_pcc, start_measuring_time, stop_measuring_time
+from models.utility_functions import torch_random
+
+# Override the default timeout in seconds for hang detection.
+TIMEOUT = 30
+
+random.seed(0)
+
+
+# Parameters provided to the test vector generator are defined here.
+# They are defined as dict-type suites that contain the arguments to the run function as keys, and lists of possible inputs as values.
+# Each suite has a key name (in this case "suite_1" and "suite_2") which will associate the test vectors to this specific suite of inputs.
+# Developers can create their own generator functions and pass them to the parameters as inputs.
+parameters = {
+    "nightly": {
+        "input_shape": gen_shapes([1, 1, 32, 32], [6, 12, 256, 256], [1, 1, 32, 32], 128),
+        "input_a_dtype": [ttnn.bfloat16, ttnn.bfloat8_b],
+        "input_a_layout": [ttnn.TILE_LAYOUT, ttnn.ROW_MAJOR_LAYOUT],
+        "input_a_memory_config": [ttnn.DRAM_MEMORY_CONFIG, ttnn.L1_MEMORY_CONFIG],
+        "output_memory_config": [ttnn.DRAM_MEMORY_CONFIG, ttnn.L1_MEMORY_CONFIG],
+    },
+}
+
+
+# Invalidate vector is called during the generation phase where each vector will be passed in.
+# If invalidated, the vector will still be stored but will be skipped.
+# Returns False, None if the vector is valid, and True, str with a reason for invalidation if it is invalid.
+def invalidate_vector(test_vector) -> Tuple[bool, Optional[str]]:
+    if test_vector["input_a_layout"] == ttnn.ROW_MAJOR_LAYOUT:
+        return True, "Row Major layout is not supported"
+    return False, None
+
+
+# This is the run instructions for the test, defined by the developer.
+# The run function must take the above-defined parameters as inputs.
+# The runner will call this run function with each test vector, and the returned results from this function will be stored.
+# If you defined a mesh_device_fixture above, the object you yielded will be passed into this function as 'device'. Otherwise, it will be the default ttnn device opened by the infra.
+def run(
+    input_shape,
+    input_a_dtype,
+    input_a_layout,
+    input_a_memory_config,
+    output_memory_config,
+    *,
+    device,
+) -> list:
+    data_seed = random.randint(0, 20000000)
+    torch.manual_seed(data_seed)
+
+    torch_input_tensor_a = gen_func_with_cast_tt(
+        partial(torch_random, low=-100, high=100, dtype=torch.float32), input_a_dtype
+    )(input_shape)
+
+    golden_function = ttnn.get_golden_function(ttnn.abs)
+    torch_output_tensor = golden_function(torch_input_tensor_a)
+
+    input_tensor_a = ttnn.from_torch(
+        torch_input_tensor_a,
+        dtype=input_a_dtype,
+        layout=input_a_layout,
+        device=device,
+        memory_config=input_a_memory_config,
+    )
+
+    start_time = start_measuring_time()
+    result = ttnn.abs(input_tensor_a, memory_config=output_memory_config)
+    output_tensor = ttnn.to_torch(result)
+    e2e_perf = stop_measuring_time(start_time)
+
+    return [check_with_pcc(torch_output_tensor, output_tensor, 0.999), e2e_perf]

ttnn/cpp/ttnn/operations/eltwise/unary/unary_pybind.hpp

@@ -76,15 +76,13 @@ void bind_unary_operation(py::module& module, const unary_operation_t& operation
 
 
 template <typename unary_operation_t>
-void bind_unary_operation_overload_complex(py::module& module, const unary_operation_t& operation, const std::string& info_doc = "" ) {
+void bind_unary_operation_overload_complex(py::module& module, const unary_operation_t& operation, const std::string& math, const std::string& info_doc = "" ) {
     auto doc = fmt::format(
         R"doc(
         Applies {0} to :attr:`input_tensor` element-wise.
 
-        {2}
-
         .. math::
-            \mathrm{{output\_tensor}}_i = {0}(\mathrm{{input\_tensor}}_i)
+            {2}
 
         Args:
             input_tensor (ttnn.Tensor): the input tensor.
@@ -97,12 +95,16 @@ void bind_unary_operation_overload_complex(py::module& module, const unary_opera
         Returns:
             ttnn.Tensor: the output tensor.
 
+        Note:
+            {3}
+
         Example:
             >>> tensor = ttnn.from_torch(torch.tensor((1, 2), dtype=torch.bfloat16), device=device)
             >>> output = {1}(tensor)
         )doc",
         operation.base_name(),
         operation.python_fully_qualified_name(),
+        math,
         info_doc);
 
     bind_registered_operation(
@@ -1325,7 +1327,16 @@ void bind_dropout(py::module& module, const unary_operation_t& operation) {
 }  // namespace detail
 
 void py_module(py::module& module) {
-    detail::bind_unary_operation_overload_complex(module, ttnn::abs);
+    detail::bind_unary_operation_overload_complex(module, ttnn::abs, R"doc(\mathrm{{output\_tensor}}_i = abs(\mathrm{{input\_tensor}}_i))doc",
+        R"doc(Supported dtypes, layouts, and ranks:
+
+           +----------------------------+---------------------------------+-------------------+
+           |     Dtypes                 |         Layouts                 |     Ranks         |
+           +----------------------------+---------------------------------+-------------------+
+           |    BFLOAT16, BFLOAT8_B     |          TILE                   |      2, 3, 4      |
+           +----------------------------+---------------------------------+-------------------+
+        )doc");
+
     detail::bind_unary_operation(module, ttnn::acos, R"doc(\mathrm{{output\_tensor}}_i = acos(\mathrm{{input\_tensor}}_i))doc");
     detail::bind_unary_operation(module, ttnn::asin, R"doc(\mathrm{{output\_tensor}}_i = asin(\mathrm{{input\_tensor}}_i))doc");
     detail::bind_unary_operation(module, ttnn::atan, R"doc(\mathrm{{output\_tensor}}_i = atan(\mathrm{{input\_tensor}}_i))doc");