feat: add requirements.txt for all modules to improve reproducibility

[jellyfishing2346]

• Add requirements.txt for 13 previously uncovered modules
• Modules: Classification_Transformers, Diffusion (Hamees), Gravitational_Lens_Classification, Physics_Informed_PINN, Regression (Yurii & Zhongchao), Self_Supervised_Learning, Difflense, Equivariant_Networks, Physics_Informed_Transformers, Super_Resolution (Atal & Pranath), and LSST_Data_Processing
• Apply consistent version pinning with >= minimums for compatibility
• Includes special-case dependencies (e3nn for equivariant, diffusers/transformers for Difflense, lsst-stack/ripple for LSST)
• Resolves Title: "Add requirements.txt for reproducibility" #176

[jellyfishing2346]

Hey @abhiram123467,

I've been working through the open issues and tackled #176 by adding requirements.txt files to all the modules that were missing them. There were 13 modules without proper dependency specs, so I went through each one, extracted the dependencies from the code, and created requirements.txt files with consistent version pinning.

The goal was to make it easier for people to set up and run these models without hitting dependency issues. Now 22 out of 23 modules have requirements.txt files (the only one left is the Images_for_README folder, which is just documentation).

I also noticed this complements PR #205, but I created module-specific requirements.txt files rather than just a root-level one, which should work better with how the project is structured.

Happy to adjust anything or answer questions about the dependencies I chose. Thanks!