intern:vitransformer
Table of Contents
Research Activities (August 22, 2024 – September 15, 2024)
Week log
Week 1: Foundational Research (August 22–27)
- Reviewed seminal papers on attention mechanisms, including:
- Attention Is All You Need
- Masked-attention Mask Transformer for Universal Image Segmentation
- Developed proficiency in Anaconda and Python programming for AI applications.
- Joined laboratory research under mentorship of senior researchers.
Week 2: Collaborative Paper Development (August 28 – September 3)
- Supported experiments for the paper TGMformer: Transferability Guided Mask Transformer for Segmentation Domain Adaptation (led by senior Zhang Enming):
- Implemented baseline models comparing fine-tuning strategies (last layer vs. full network) for transfer learning.
- Formalized mathematical definitions for transfer learning, clustering algorithms, and attention masking in the Methods section.
- Assisted with LaTeX typesetting and document formatting.
Week 3: Data Pipeline Implementation (September 4–10)
- Continued working on the paper TGMformer: Transferability Guided Mask Transformer for Segmentation Domain Adaptation
- Engineered a data loader and task-splitting module for the ImageNet-R dataset to support experiments with the Hypernet framework.
- Debugged and optimized code to improve functional reliability.
Week 4: Analysis and Synthesis (September 11–15)
- Evaluated experimental results from the Hypernet project and documented insights.
- Authored a reflective summary highlighting growth in research skills, data processing, and software testing.
intern/vitransformer.txt · Last modified: 2025/07/28 08:24 by lizhengyu