search

terry-r123 / RNABenchmark

[NeurIPS 2024] BEACON: Benchmark for Comprehensive RNA Tasks and Language Models
Apache License 2.0
21 stars 1 forks source link

fine tuned models #4

Closed pabloacera closed 1 month ago

pabloacera commented 2 months ago

Hi, Thanks for putting everything publicly available! Are the models in this repo fine tuned? if I used the models, would I get the same accuracy for each tasks as in the paper? thanks a lot in advance! Cheers, Pablo.

terry-r123 commented 2 months ago

Hi, Thanks for putting everything publicly available! Are the models in this repo fine tuned? if I used the models, would I get the same accuracy for each tasks as in the paper? thanks a lot in advance! Cheers, Pablo.

Thanks for your question.

Due to too many finetune tasks and experiments, the models in the repository are pre-trained and are all unified into Huggingface form. The resources required for finetune for each downstream task are not large, you can finetune on the downstream task and you should get almost the same results as in the paper.

Best regards, Yuchen