Fix/postprocess memory oom

[lixiaolei1982]

Problem:

• PostProcess fails with CUDA out-of-memory when processing sparse_forward() outputs
• Boolean mask conversion creates large tensors on GPU

Solution:

• Implement OptimizedPostProcess class
• Add batch processing for mask conversion (configurable batch size)
• Enable early CPU offloading to reduce GPU memory pressure
• Support automatic mixed precision for further optimization

Changes:

• Added optimized_postprocess.py with memory-efficient implementation
• Modified init.py to export the new class (if needed)

Testing:

• Resolves OOM on 11GB GPU with 1024x1024 masks
• Compatible with both forward() and sparse_forward()
• Maintains identical output format"

[codecov]

Codecov Report

❌ Patch coverage is 24.56140% with 43 lines in your changes missing coverage. Please review.
✅ Project coverage is 52%. Comparing base (cfeba5d) to head (6c0c78d).
⚠️ Report is 6 commits behind head on develop.

❌ Your patch check has failed because the patch coverage (25%) is below the target coverage (95%). You can increase the patch coverage or adjust the target coverage.
❌ Your project check has failed because the head coverage (52%) is below the target coverage (95%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@          Coverage Diff           @@
##           develop   #656   +/-   ##
======================================
- Coverage       52%    52%   -0%     
======================================
  Files           48     48           
  Lines         6350   6389   +39     
======================================
+ Hits          3311   3314    +3     
- Misses        3039   3075   +36     

🚀 New features to boost your workflow:

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