[Housekeeping] `yolov8n.onnx` binary model and `yolov8n_int8_openvino_model/` directory are committed directly to git without LFS — bloats repository clone size for all 36 forks and corrupts git diff history

[prince-pokharna]

Summary

The repository root contains:

• yolov8n.onnx — a ~6.3MB binary ONNX model file
• yolov8n_int8_openvino_model/ — a directory containing INT8-quantized OpenVINO model weights

Binary files committed to regular git (not Git LFS) cause the following problems:

1. Repository size inflation: Every git clone of Eagle and all 36 forks downloads the full model binary. Every update to the model creates another full binary copy in git history — the size compounds over time.
2. Non-diffable: git diff on a binary file is meaningless noise. Code reviewers cannot see what changed in a model update.
3. Slow CI: GitHub Actions runners clone the repo on every workflow run, downloading the model binary each time — adding seconds/minutes to every CI job.

Additionally, the .claude/ directory at root contains Claude AI assistant session artifacts — an IDE/tool artifact that should be gitignored similar to .idea/ or .vscode/.

Proposed Fix

Step 1: Remove binary files from git tracking

git rm --cached yolov8n.onnx
git rm -r --cached yolov8n_int8_openvino_model/
git rm -r --cached .claude/

Step 2: Add to .gitignore
ML Model binaries — use Git LFS or download scripts
*.onnx
*.pt
*.pth
*_openvino_model/
AI assistant artifacts
.claude/

Step 3: Add a model download script

# scripts/download_models.sh
#!/bin/bash
echo "Downloading Eagle YOLO models..."
wget -O yolov8n.onnx https://github.com/Devnil434/Eagle/releases/download/v1.0/yolov8n.onnx

Step 4: Update README and make install to run download_models.sh

Acceptance Criteria

• yolov8n.onnx and yolov8n_int8_openvino_model/ removed from git tracking
• .claude/ removed from git tracking
• scripts/download_models.sh (or equivalent) added as model setup step
• Makefile updated so make install downloads required models
• .gitignore updated with patterns for model files and IDE artifacts

Labels: housekeeping, good first issue, devops

[prince-pokharna]

Hey @Devnil434 ,
Please assign to me under GSSOC 2026.
/claim /assign /claim-assignment