Merge Lora adapters with a base model
Explore and benchmark visual document retrieval models
Browse and filter ML model leaderboard data
View and compare language model evaluations
Optimize and train foundation models using IBM's FMS
Evaluate adversarial robustness using generative models
Display benchmark results
Retrain models for new data at edge devices
Convert Stable Diffusion checkpoint to Diffusers and open a PR
Evaluate LLM over-refusal rates with OR-Bench
Compare code model performance on benchmarks
Benchmark AI models by comparison
Browse and filter machine learning models by category and modality
Merge Lora is a tool designed for model benchmarking that enables users to merge LoRA (Low-Rank Adaptation) adapters with a base model. It provides a seamless way to combine multiple adapters, enhancing the model's capabilities while maintaining efficiency. Merge Lora is particularly useful for users working with large language models and seeking to integrate specialized adapters for diverse tasks.
What is the purpose of Merge Lora?
Merge Lora is designed to simplify the process of combining LoRA adapters with a base model, allowing users to leverage specialized adapters for various tasks without retraining the model from scratch.
Can I merge multiple adapters at once?
Yes, Merge Lora supports the merging of multiple LoRA adapters into a single model, provided they are compatible.
Is Merge Lora compatible with all LoRA adapters?
While Merge Lora is designed to work with most LoRA adapters, compatibility depends on the specific adapters and base models used. Always perform a compatibility check before merging.