Expand description
ESM2 Tokenizer. Models converted to ONNX format from ESM2
and uploaded to HuggingFace hub. The tokenizer is included in this crate and loaded from
memory using tokenizer.json
. This is fairly minimal - for the full set of ESM2 models
please see the ESM2 repository and the HuggingFace hub.
§Models:
- T6_8M - small 6-layer protein language model
- T12_35M - medium 12-layer protein language model
- T30_150M - large 30-layer protein language model