Use ProGen2

The input amino acid sequence to extend. Provide ProGen2 with an amino acid sequence and it will do it's best to expand the rest of the protein.

The number of output samples to generate. Must be a minimum of 5 and maximum of 250

The maximum allowed sequence length. This value must be greater than 50 and greater than the length of the input sequence.

Select the model you want to use for your prediction. Larger models tend to perform better but take longer to finish predictions. The base model appears to have the best results while the progen2-oas model is trained to work with antibody sequences only.

The temperature / entropy controls how likely it is to choose less probable amino acids. A temperature of 0 gives the same response every time because there's a 0% chance to choose any amino acid aside from the most likely. A temperature of 1 is the default and picks based on the model's base confidence. However, a temperature greater than 1 gives more weight to unlikely amino acids than to likely ones.

The Top-p goes for a minimal set of words, the probability of which exceeds p. Choosing a top p such as 0.9 means you would likely get a set of the most likely words for the model to choose.