[BugFix][HPU] Add paddle.compat compatibility check for HPU environment

[mattheliu]

Motivation

Fix FastDeploy import failure on Intel HPU (Habana Gaudi) environment.

In HPU environment, PaddlePaddle 3.1.1 (stable) does not have the paddle.compat module (introduced in 2025-09-05), causing an AttributeError during FastDeploy import:

AttributeError: module 'paddle' has no attribute 'compat'. Did you mean: 'concat'?

Modifications

Add hasattr check before calling paddle.compat.enable_torch_proxy() to ensure compatibility across different Paddle versions and hardware platforms.

Modified files:

1. fastdeploy/init.py
2. fastdeploy/model_executor/layers/attention/flash_attn_backend.py

# Enable torch proxy if paddle.compat is available
# Some Paddle versions (e.g., stable versions on HPU/GCU) may not have the compat module
if hasattr(paddle, "compat") and hasattr(paddle.compat, "enable_torch_proxy"):
    paddle.compat.enable_torch_proxy(scope={"triton"})

Usage or Command

No special command needed. After this fix, FastDeploy can be imported normally on HPU environment.

Accuracy Tests

• This fix only adds a conditional check without affecting original functionality
• Behavior remains unchanged on environments that support paddle.compat
• On environments without paddle.compat, the call is skipped to avoid errors

Checklist

• Add at least a tag in the PR title.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
  • No unit test added: This is a simple compatibility check that does not require dedicated tests.
• Provide accuracy results.
  • No accuracy impact: This change only adds a conditional check for module existence.

[paddle-bot]

Thanks for your contribution!

[codecov-commenter]

Codecov Report

❌ Patch coverage is 50.00000% with 2 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@d6b3c72). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/__init__.py	50.00%	0 Missing and 1 partial ⚠️
...el_executor/layers/attention/flash_attn_backend.py	50.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #6387   +/-   ##
==========================================
  Coverage           ?   68.23%           
==========================================
  Files              ?      391           
  Lines              ?    52242           
  Branches           ?     8147           
==========================================
  Hits               ?    35649           
  Misses             ?    13986           
  Partials           ?     2607           

Flag	Coverage Δ
GPU	68.23% <50.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.