Request model evaluation on COCO val 2017 dataset
Measure over-refusal in LLMs using OR-Bench
View and submit LLM benchmark evaluations
Quantize a model for faster inference
Retrain models for new data at edge devices
Benchmark models using PyTorch and OpenVINO
Rank machines based on LLaMA 7B v2 benchmark results
Convert Hugging Face models to OpenVINO format
Display and filter leaderboard models
Evaluate open LLMs in the languages of LATAM and Spain.
Merge machine learning models using a YAML configuration file
Teach, test, evaluate language models with MTEB Arena
Convert Hugging Face model repo to Safetensors
The Open Object Detection Leaderboard is a benchmarking platform designed to evaluate and compare different object detection models. It provides a standardized framework for assessing model performance using the COCO (Common Objects in Context) val 2017 dataset. This leaderboard is a community-driven tool that allows researchers and developers to submit their model results and view how they stack up against others in the field.
What metrics are used for evaluation?
The leaderboard primarily uses the COCO metric, which is the mean Average Precision (mAP) across all categories and instance sizes.
How can I submit my model results?
To submit your model, evaluate it on the COCO val 2017 dataset and follow the submission guidelines provided on the leaderboard's website.
Can I update my model's entry after submission?
Yes, you can update your model's entry by resubmitting the results. The leaderboard will reflect the latest submission for your model.