Skip to content

Arm backend: Split and decouple model evaluators #18118

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
martinlsm merged 2 commits into from
Mar 13, 2026
Merged

Arm backend: Split and decouple model evaluators #18118

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
eb06c08
Arm backend: Split and decouple model evaluators
Feb 20, 2026
a705150
Merge branch 'main' into marlin-evaluators
martinlsm Mar 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 17 additions & 21 deletions backends/arm/test/misc/test_model_evaluator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Copyright 2025 Arm Limited and/or its affiliates.
# Copyright 2025-2026 Arm Limited and/or its affiliates.
#
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.
Expand All @@ -7,7 +7,10 @@
import unittest

import torch
from executorch.backends.arm.util.arm_model_evaluator import GenericModelEvaluator
from executorch.backends.arm.util.arm_model_evaluator import (
FileCompressionEvaluator,
NumericalModelEvaluator,
)

# Create an input that is hard to compress
COMPRESSION_RATIO_TEST = torch.rand([1024, 1024])
Expand All @@ -21,38 +24,31 @@ def mocked_model_2(input: torch.Tensor) -> torch.Tensor:
return torch.tensor([1.0, 2.0, 3.0, 3.0])


class TestGenericModelEvaluator(unittest.TestCase):
"""Tests the GenericModelEvaluator class."""
class TestModelEvaluator(unittest.TestCase):
"""Tests the Arm model evaluators."""

def test_get_model_error_no_target(self):
example_input = torch.tensor([[1.0, 2.0, 3.0, 4.0]])
evaluator = GenericModelEvaluator(
evaluator = NumericalModelEvaluator(
"dummy_model",
mocked_model_1,
mocked_model_2,
example_input,
"tmp/output_tag0.tosa",
(example_input,),
None,
)

model_error_dict = evaluator.get_model_error()
metrics = evaluator.evaluate()

self.assertEqual(model_error_dict["max_error"], [1.0])
self.assertEqual(model_error_dict["max_absolute_error"], [1.0])
self.assertEqual(model_error_dict["max_percentage_error"], [25.0])
self.assertEqual(model_error_dict["mean_absolute_error"], [0.25])
self.assertEqual(metrics["max_error"], 1.0)
self.assertEqual(metrics["max_absolute_error"], 1.0)
self.assertEqual(metrics["max_percentage_error"], 25.0)
self.assertEqual(metrics["mean_absolute_error"], 0.25)

def test_get_compression_ratio_no_target(self):
with tempfile.NamedTemporaryFile(delete=True) as temp_bin:
torch.save(COMPRESSION_RATIO_TEST, temp_bin)

example_input = torch.tensor([[1.0, 2.0, 3.0, 4.0]])
evaluator = GenericModelEvaluator(
"dummy_model",
mocked_model_1,
mocked_model_2,
example_input,
temp_bin.name,
)
evaluator = FileCompressionEvaluator("dummy_model", temp_bin.name)

ratio = evaluator.get_compression_ratio()
ratio = evaluator.evaluate()["compression_ratio"]
self.assertAlmostEqual(ratio, 1.1, places=1)
Loading
Loading