TransHLA_II is a protein language model hosted on Hugging Face. It identifies whether a given peptide will be recognized by HLA as an epitope. The model requires no HLA allele information as input, a distinction noted as a first for this type of tool.
The model is provided for feature extraction tasks and is implemented using the Transformers library with PyTorch and Safetensors formats. It carries the IEDB tag. Users load it through standard library calls that include the trust_remote_code flag, either via a pipeline for high-level access or by instantiating AutoModel directly with optional device mapping. Example code for both approaches appears in the repository alongside links to run it in Google Colab or Kaggle notebooks.
TransHLA_II is presented as a custom model intended for discerning peptide epitopes. The repository lists it under the SkywalkerLu account and indicates one like from the community. No pricing, licensing terms, or additional training details are stated.
TransHLA II sits in PulseGate's Foundation models & chat category. Accurately predicting peptide epitopes for HLA without needing to specify HLA alleles as input. It is built as an open-source project for bioinformatics researchers. TransHLA II is open source under the Open Source license. TransHLA II is available on the web, the command line, and API, and it can be self-hosted.
SkywalkerLu builds and maintains TransHLA II, and it first shipped in 2023. Key capabilities include Feature Extraction, Peptide Epitope Prediction, and HLA-independent Classification. It exposes integrations via a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do