#14734: fix hardtanh golden function (#14736)

tests/sweep_framework/sweeps/eltwise/unary/hardtanh/hardtanh_pytorch2.py

@@ -6,18 +6,13 @@
 from functools import partial
 
 import torch
-import random
 import ttnn
 from tests.sweep_framework.sweep_utils.utils import gen_shapes, gen_low_high_scalars
 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 = {
     "nightly": {
@@ -152,8 +147,7 @@ def run(
     *,
     device,
 ) -> list:
-    data_seed = random.randint(0, 20000000)
-    torch.manual_seed(data_seed)
+    torch.manual_seed(0)
 
     torch_input_tensor_a = gen_func_with_cast_tt(
         partial(torch_random, low=-100, high=100, dtype=torch.float32), input_dtype
@@ -163,7 +157,7 @@ def run(
     max_val = input_specs.get("max_val")
 
     golden_function = ttnn.get_golden_function(ttnn.hardtanh)
-    torch_output_tensor = golden_function(torch_input_tensor_a, min=min_val, max=max_val)
+    torch_output_tensor = golden_function(torch_input_tensor_a, min_val=min_val, max_val=max_val)
 
     input_tensor_a = ttnn.from_torch(
         torch_input_tensor_a,
@@ -174,7 +168,7 @@ def run(
     )
 
     start_time = start_measuring_time()
-    result = ttnn.hardtanh(input_tensor_a, min=min_val, max=max_val, memory_config=output_memory_config)
+    result = ttnn.hardtanh(input_tensor_a, min_val=min_val, max_val=max_val, memory_config=output_memory_config)
     output_tensor = ttnn.to_torch(result)
     e2e_perf = stop_measuring_time(start_time)
 

tests/tt_eager/python_api_testing/sweep_tests/tt_lib_ops.py

@@ -235,7 +235,7 @@ def eltwise_hardtanh(
     **kwargs,
 ):
     t0 = setup_tt_tensor(x, device, layout[0], input_mem_config[0], dtype[0])
-    t1 = ttnn.hardtanh(t0, min=low, max=high, memory_config=output_mem_config)
+    t1 = ttnn.hardtanh(t0, min_val=low, max_val=high, memory_config=output_mem_config)
 
     return tt2torch_tensor(t1)
 

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

@@ -1845,8 +1845,8 @@ void py_module(py::module& module) {
     detail::bind_unary_composite_floats_with_default(
         module,
         ttnn::hardtanh,
-        "min", "min value", -1.0f,
-        "max", "max value", 1.0f);
+        "min_val", "min value", -1.0f,
+        "max_val", "max value", 1.0f);
 
     detail::bind_unary_composite_optional_floats_with_default(
         module,

ttnn/ttnn/operations/unary.py

@@ -78,7 +78,6 @@ def _golden_function(input_tensor: ttnn.Tensor, **_):
             "digamma": torch.digamma,
             "hardswish": torch.nn.functional.hardswish,
             "hardsigmoid": torch.nn.functional.hardsigmoid,
-            "hardtanh": torch.nn.functional.hardtanh,
             "lgamma": torch.lgamma,
             "log1p": torch.log1p,
             "mish": lambda _x: torch.nn.functional.mish(_x.to(torch.float)),
@@ -163,7 +162,6 @@ def _golden_function(input_tensor: ttnn.Tensor, **_):
     ttnn.digamma,
     ttnn.hardswish,
     ttnn.hardsigmoid,
-    ttnn.hardtanh,
     ttnn.lgamma,
     ttnn.log1p,
     ttnn.mish,
@@ -241,6 +239,15 @@ def _golden_function_elu(input_tensor_a, *args, alpha=1.0, **kwargs):
 ttnn.attach_golden_function(ttnn.elu, golden_function=_golden_function_elu)
 
 
+def _golden_function_hardtanh(input_tensor_a, *args, min_val=-1.0, max_val=1.0, **kwargs):
+    import torch
+
+    return torch.nn.functional.hardtanh(input_tensor_a, min_val=min_val, max_val=max_val)
+
+
+ttnn.attach_golden_function(ttnn.hardtanh, golden_function=_golden_function_hardtanh)
+
+
 def _golden_function_leaky_relu(input_tensor_a, *args, negative_slope=0.01, **kwargs):
     import torch