Add comprehensive GitHub Copilot instructions for deep_ssm repository

[Copilot]

This PR adds comprehensive .github/copilot-instructions.md to guide GitHub Copilot coding agents on how to work effectively in the deep_ssm codebase. The instructions were created through extensive validation of installation processes, build procedures, and training workflows.

Key Features

Multiple Installation Methods: Provides 4 different installation approaches to handle various environment constraints:

• Complete installation with all dependencies (45+ minutes)
• Minimal installation for restricted environments
• Sherlock cluster-specific setup with module loading
• Emergency setup for severe network/build restrictions

Explicit Timing Expectations: All commands include realistic time estimates and "NEVER CANCEL" warnings:

• Environment setup: 15-45 minutes depending on method
• PyTorch installation: 5-15 minutes (2GB+ download)
• S5 CIFAR training: 30 minutes per epoch (250 epochs total)
• BCI model training: 60-90 minutes for full runs

Validated Commands: Every command was tested where possible, with comprehensive error handling:

# Test basic imports (should complete in <30 seconds)
python -c "import deep_ssm; print('Package installed correctly')"

# Quick S5 test (5 minutes)
python -m example --dataset mnist --epochs 1 --batch_size 50

# BCI debug test (3 minutes)  
python run.py --config-name="baseline_gru" trainer_cfg.fast_dev_run=1

Common Issues Documentation: Addresses specific problems encountered during validation:

• Network timeout errors in CI environments
• CUDA compilation failures for mamba-ssm
• Missing dependency handling with graceful fallbacks
• PyTorch version conflicts with conda environments

Manual Validation Scenarios: Provides concrete workflows to test after making changes:

• Quick validation tests (5 minutes)
• Debug runs for different model types
• Complete end-to-end training validation

The instructions start with a firm directive to follow them first before additional search, and include comprehensive repository navigation, configuration details, and performance expectations to enable efficient development workflows.

Fixes #16.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.