search

ProteomeTeam / NovoB

Bidirectional De Novo Peptide Sequencing Using a Transformer Model
2 stars 0 forks source link

Pretrianed model on a large dataset #4

Closed irleader closed 2 months ago

irleader commented 4 months ago

Hi,

Is it possible to share with us a model checkpoint which has been trained on a huge dataset like massivekb used by casanovo? I am afraid that the models trained on graphnovo dataset and nine-species dataset can not fully exhibit the power of NovoB.

Thanks a lot!

pardess commented 4 months ago

We do not yet have a model checkpoint trained on huge datasets like massivekb. As you said, the checkpoints we provide are not enough.

Currently, I am thinking about which dataset would be best to use for training. Additionally, efforts are being made to improve accuracy.

irleader commented 4 months ago

Thanks a lot for your feedback!