Evaluate Persian LLMs on various tasks
Run benchmarks on prediction models
Generate and view leaderboard for LLM evaluations
Upload a machine learning model to Hugging Face Hub
Measure over-refusal in LLMs using OR-Bench
Merge Lora adapters with a base model
Evaluate code generation with diverse feedback types
Submit models for evaluation and view leaderboard
Submit deepfake detection models for evaluation
Open Persian LLM Leaderboard
Merge machine learning models using a YAML configuration file
Quantize a model for faster inference
Benchmark LLMs in accuracy and translation across languages
The š¤ Persian LLM Leaderboard is a comprehensive platform designed to evaluate and compare Persian language models across various tasks. It provides a centralized hub for researchers and developers to assess the performance of different models in the Persian language, fostering innovation and transparency in the field of natural language processing.
ā¢ Model Comparison: Evaluate and compare the performance of multiple Persian LLMs on different tasks. ā¢ Task-Specific Benchmarks: Assess models on a variety of tasks tailored to the Persian language, such as text classification, summarization, and translation. ā¢ Detailed Metrics: Access detailed performance metrics to understand model strengths and weaknesses. ā¢ Visualizations: Interactive charts and graphs to visualize model performance and trends over time. ā¢ Regular Updates: Stay informed with the latest developments and updates in Persian LLMs. ā¢ Community-Driven: Submit your own models or results to contribute to the leaderboard.
What models are included on the leaderboard?
The leaderboard features a variety of Persian language models, including both state-of-the-art and emerging models from researchers and developers.
How are models evaluated?
Models are evaluated based on their performance on a range of tasks specific to the Persian language, using standard benchmarks and metrics.
Can I submit my own model?
Yes, you can submit your Persian LLM for evaluation. Follow the submission guidelines provided on the platform to include your model on the leaderboard.