PSALM-1
Collection
4 items • Updated
Token Classification
Transformers
Safetensors
English
esm
protein language model
protein sequence annotation
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ProteinSequenceAnnotation/PSALM-1-family", dtype="auto", device_map="auto")Quick Links
Model Card for Model ID
NOTE: PSALM-1 has not been trained on all Pfam families, as it has been trained and benchmarked on highly-curated datasets with strict sequence similarity guarantees between train and test data. PSALM-1b (trained on all families in Pfam 35.0) is coming soon
PSALM-1-family is a 166M parameter model that takes as input ESM-2 residue-level protein sequence emebeddings (unpooled) and outputs a distribution over Pfam domain families for each amino acid in the sequence. PSALM-1-family uses a BiLSTM followed by a stack of dense layers
- Downloads last month
- 8
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="ProteinSequenceAnnotation/PSALM-1-family", device_map="auto")