IgBert unpaired model

Model pretrained on protein and antibody sequences using a masked language modeling (MLM) objective. It was introduced in the paper Large scale paired antibody language models.

The model is finetuned from ProtBert-BFD using unpaired antibody sequences from the Observed Antibody Space.

Use

The model and tokeniser can be loaded using the transformers library

from transformers import BertModel, BertTokenizer

tokeniser = BertTokenizer.from_pretrained("Exscientia/IgBert_unpaired", do_lower_case=False)
model = BertModel.from_pretrained("Exscientia/IgBert_unpaired", add_pooling_layer=False)

The tokeniser is used to prepare batch inputs

# single chain sequences
sequences = [
    "EVVMTQSPASLSVSPGERATLSCRARASLGISTDLAWYQQRPGQAPRLLIYGASTRATGIPARFSGSGSGTEFTLTISSLQSEDSAVYYCQQYSNWPLTFGGGTKVEIK",
    "ALTQPASVSGSPGQSITISCTGTSSDVGGYNYVSWYQQHPGKAPKLMIYDVSKRPSGVSNRFSGSKSGNTASLTISGLQSEDEADYYCNSLTSISTWVFGGGTKLTVL"
]

# The tokeniser expects input of the form ["E V V M...", "A L T Q..."]
sequences = [' '.join(sequence) for sequence in sequences] 

tokens = tokeniser.batch_encode_plus(
    sequences, 
    add_special_tokens=True, 
    pad_to_max_length=True, 
    return_tensors="pt",
    return_special_tokens_mask=True
) 

Note that the tokeniser adds a [CLS] token at the beginning of each sequence, a [SEP] token at the end of each sequence and pads using the [PAD] token. For example a batch containing sequences E V V M, A L will be tokenised to [CLS] E V V M [SEP] and [CLS] A L [SEP] [PAD] [PAD].

Sequence embeddings are generated by feeding tokens through the model

output = model(
    input_ids=tokens['input_ids'], 
    attention_mask=tokens['attention_mask']
)

residue_embeddings = output.last_hidden_state

To obtain a sequence representation, the residue tokens can be averaged over like so

import torch

# mask special tokens before summing over embeddings
residue_embeddings[tokens["special_tokens_mask"] == 1] = 0
sequence_embeddings_sum = residue_embeddings.sum(1)

# average embedding by dividing sum by sequence lengths
sequence_lengths = torch.sum(tokens["special_tokens_mask"] == 0, dim=1)
sequence_embeddings = sequence_embeddings_sum / sequence_lengths.unsqueeze(1)

For sequence level fine-tuning the model can be loaded with a pooling head by setting add_pooling_layer=True and using output.pooler_output in the down-stream task.