tanganke/fusion_bench
FusionBench: A Comprehensive Benchmark/Toolkit of Deep Model Fusion
When you have several specialized AI models, each good at a different task, and you want to combine their knowledge into one single, more versatile model, FusionBench helps you do that. It takes these individual deep learning models and evaluates various strategies for merging them, showing you which fusion techniques create the best performing combined model. This is useful for AI researchers and machine learning engineers who need to build multi-task AI systems.
206 stars. Available on PyPI.
Use this if you are a machine learning researcher or engineer comparing different techniques for combining multiple specialized AI models into a single, more capable model.
Not ideal if you are looking for a simple tool to deploy an existing pre-trained model or fine-tune a single model for a specific task.
Stars
206
Forks
29
Language
Python
License
MIT
Category
Last pushed
Mar 07, 2026
Commits (30d)
0
Dependencies
16
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/tanganke/fusion_bench"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
argonne-lcf/dlio_benchmark
An I/O benchmark for deep Learning applications
stefan-ctrl/mbdd-enhanced
github.com/google-research/google-research/tree/master/mbpp enhanced
cloudwalk/tictactoe-dataset
Filtering and ranking all of 5478 states in tic-tac-toe for efficient evaluation on the hardest ones