ESM2-150M-TokenWise

模型描述

MultiomicsLM-ESM 是一个结合了 ESM2 蛋白质语言模型和结构信息的多模态模型。

模型架构

  • 基础模型: ESM2-150M
  • 结构组件: ESM-IF1 Embedding+结构投影器(mlp_proj)

使用方法

from transformers import AutoModelForMaskedLM, AutoTokenizer, AutoConfig

# 加载模型
model = AutoModelForMaskedLM.from_pretrained("ssyypp111/ESM2-150M-TokenWise", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("ssyypp111/ESM2-150M-TokenWise", trust_remote_code=True)

具体使用方式参考GitHub仓库: https://github.com/Eikor/InstructPLM-mu

文件说明

  • config.json: 模型配置
  • pytorch_model.bin: 模型权重
  • modeling_esm.py: 模型定义
  • structure.py: 结构处理模块
  • tokenization_iPLM.py: 分词器

引用

@article{xu2025instructplm,
  title={InstructPLM-mu: 1-Hour Fine-Tuning of ESM2 Beats ESM3 in Protein Mutation Predictions},
  author={Xu, Junde and Shi, Yapin and Lang, Lijun and Cui, Taoyong and Zhang, Zhiming and Chen, Guangyong and Qiu, Jiezhong and Heng, Pheng-Ann},
  journal={arXiv preprint arXiv:2510.03370},
  year={2025}
}
Downloads last month
22
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support