Files
transformers/tests/utils/test_attention_visualizer.py
陈赣 06f1fd69a6
Some checks failed
Self-hosted runner (nightly-past-ci-caller) / Get number (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.11 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.10 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.9 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.8 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.7 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.6 (push) Has been cancelled
Self-hosted runner (nightly-past-ci-caller) / TensorFlow 2.5 (push) Has been cancelled
Self-hosted runner (benchmark) / Benchmark (aws-g5-4xlarge-cache) (push) Has been cancelled
Build documentation / build (push) Has been cancelled
Build documentation / build_other_lang (push) Has been cancelled
CodeQL Security Analysis / CodeQL Analysis (push) Has been cancelled
New model PR merged notification / Notify new model (push) Has been cancelled
PR CI / pr-ci (push) Has been cancelled
Slow tests on important models (on Push - A10) / Get all modified files (push) Has been cancelled
Secret Leaks / trufflehog (push) Has been cancelled
Update Transformers metadata / build_and_package (push) Has been cancelled
Slow tests on important models (on Push - A10) / Model CI (push) Has been cancelled
Check Tiny Models / Check tiny models (push) Has been cancelled
Self-hosted runner (Intel Gaudi3 scheduled CI caller) / Model CI (push) Has been cancelled
Self-hosted runner (Intel Gaudi3 scheduled CI caller) / Pipeline CI (push) Has been cancelled
Self-hosted runner (Intel Gaudi3 scheduled CI caller) / Example CI (push) Has been cancelled
Self-hosted runner (Intel Gaudi3 scheduled CI caller) / DeepSpeed CI (push) Has been cancelled
Self-hosted runner (Intel Gaudi3 scheduled CI caller) / Trainer/FSDP CI (push) Has been cancelled
Nvidia CI - Flash Attn / Setup (push) Has been cancelled
Nvidia CI - Flash Attn / Model CI (push) Has been cancelled
Nvidia CI / Setup (push) Has been cancelled
Nvidia CI / Model CI (push) Has been cancelled
Nvidia CI / Torch pipeline CI (push) Has been cancelled
Nvidia CI / Example CI (push) Has been cancelled
Nvidia CI / Trainer/FSDP CI (push) Has been cancelled
Nvidia CI / DeepSpeed CI (push) Has been cancelled
Nvidia CI / Quantization CI (push) Has been cancelled
Nvidia CI / Kernels CI (push) Has been cancelled
Doctests / Setup (push) Has been cancelled
Doctests / Call doctest jobs (push) Has been cancelled
Doctests / Send results to webhook (push) Has been cancelled
Extras Smoke Test / Get supported Python versions (push) Has been cancelled
Extras Smoke Test / Test extras on Python ${{ matrix.python-version }} (push) Has been cancelled
Extras Smoke Test / Check Slack token availability (push) Has been cancelled
Extras Smoke Test / Notify failures to Slack (push) Has been cancelled
Self-hosted runner (AMD scheduled CI caller) / Trigger Scheduled AMD CI (push) Has been cancelled
Stale Bot / Close Stale Issues (push) Has been cancelled
first commit
2026-06-05 16:53:03 +08:00

126 lines
7.1 KiB
Python

# Copyright 2025 The HuggingFace Inc. team.
#
# 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 builtins
import io
import re
import unittest
from transformers.testing_utils import require_torch
from transformers.utils.attention_visualizer import AttentionMaskVisualizer
ANSI_RE = re.compile(r"\x1b\[[0-9;]*m")
def _normalize(s: str) -> str:
# drop ANSI (colors may be disabled on CI), normalize line endings,
# and strip trailing spaces without touching alignment inside lines
s = ANSI_RE.sub("", s)
s = s.replace("\r\n", "\n").replace("\r", "\n")
return "\n".join(line.rstrip() for line in s.split("\n")).strip()
@require_torch
class AttentionMaskVisualizerTester(unittest.TestCase):
"""Test suite for AttentionMaskVisualizer"""
def test_paligemma_multimodal_visualization(self):
"""Test AttentionMaskVisualizer with PaliGemma multimodal model"""
model_name = "hf-internal-testing/namespace_google_repo_name_paligemma-3b-pt-224"
input_text = "<img> What is in this image?"
buf = io.StringIO()
orig_print = builtins.print
def _print(*args, **kwargs):
kwargs.setdefault("file", buf)
orig_print(*args, **kwargs)
try:
builtins.print = _print
visualizer = AttentionMaskVisualizer(model_name)
visualizer(input_text)
finally:
builtins.print = orig_print
output = buf.getvalue()
expected_output = """
##########################################################################################################################################################################################################################################
## Attention visualization for \033[1mpaligemma:hf-internal-testing/namespace_google_repo_name_paligemma-3b-pt-224\033[0m PaliGemmaModel ##
##########################################################################################################################################################################################################################################
\033[92m■\033[0m: i == j (diagonal) \033[93m■\033[0m: token_type_ids
Attention Matrix
\033[93m'<image>'\033[0m: 0 \033[93m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
\033[93m'<image>'\033[0m: 1 \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
\033[93m'<image>'\033[0m: 2 \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
\033[93m'<image>'\033[0m: 3 \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
\033[93m'<image>'\033[0m: 4 \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m \033[93m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'<bos>' : 5 ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁What' : 6 ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁is' : 7 ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁in' : 8 ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁this' : 9 ■ ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ |
'▁image' : 10 ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ |
'?' : 11 ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ |
'\\n' : 12 ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ |
'<eos>' : 13 ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m |
##########################################################################################################################################################################################################################################
""" # noqa
self.assertEqual(_normalize(output), _normalize(expected_output))
def test_llama_text_only_visualization(self):
"""Test AttentionMaskVisualizer with Llama text-only model"""
model_name = "hf-internal-testing/namespace_meta-llama_repo_name_Llama-2-7b-hf"
input_text = "Plants create energy through a process known as"
buf = io.StringIO()
orig_print = builtins.print
def _print(*args, **kwargs):
kwargs.setdefault("file", buf)
orig_print(*args, **kwargs)
try:
builtins.print = _print
visualizer = AttentionMaskVisualizer(model_name)
visualizer(input_text)
finally:
builtins.print = orig_print
output = buf.getvalue()
expected_output = """
##########################################################################################################################################################################################################
## Attention visualization for \033[1mllama:hf-internal-testing/namespace_meta-llama_repo_name_Llama-2-7b-hf\033[0m LlamaModel ##
##########################################################################################################################################################################################################
\033[92m■\033[0m: i == j (diagonal) \033[93m■\033[0m: token_type_ids
Attention Matrix
'▁Pl' : 0 \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'ants' : 1 ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁create' : 2 ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁energy' : 3 ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ ⬚ |
'▁through': 4 ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ ⬚ |
'▁a' : 5 ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ ⬚ |
'▁process': 6 ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ ⬚ |
'▁known' : 7 ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m ⬚ |
'▁as' : 8 ■ ■ ■ ■ ■ ■ ■ ■ \033[92m■\033[0m |
##########################################################################################################################################################################################################
""" # noqa
self.assertEqual(_normalize(output), _normalize(expected_output))