Table 5. Summary of All of the Protein Embedding Models Employed for Developing the Proposed Methodology.
shorthand | layers | params | data set | embedding dim |
---|---|---|---|---|
ESM1 | 34 | 670M | UR50/S2018_03 | 1280 |
ESM1b | 33 | 650M | UR50/S2018_03 | 1280 |
ESM1v | 33 | 650M | UR90/S2020_03 | 1280 |
ProtBert | 30 | 420M | UniRef100 | 1024 |
ProtBert-BFD | 12 | 224M | UniRef100 | 4096 |
ProtAlbert | 30 | 420M | BFD100 | 1024 |
Prot-T5-XL | 24 | 3B | UniRef100 | 1024 |
Prot-T5-XL-BFD | 24 | 3B | BFD100 | 1024 |
ProtXLNet | 30 | 409M | UniRef100 | 1024 |
AlphaFold | 48a | 92M | UniRef100 | 384 |
Evoformer blocks.