AI NEWS.ccAI正在发生 未来正在更新
♡ 我的收藏 0
AI 模型库/Rostlab/ProstT5
MODEL PROFILE · 模型档案

ProstT5

Rostlab · ProstT5

ProstT5 is a protein language model (pLM) which can translate between protein sequence and structure. ProstT5 (Protein structure-sequence T5) is based on ProtT5-XL-U50, a T5 model trained on encoding protein sequences using span corruption applied on billions of protein sequences. ProstT5 finetunes ProtT5-XL-U50 on translating between protein sequence and structure using 17M proteins with high-quality 3D structure predictions from the AlphaFoldDB. Protein structure is converted from 3D to 1D using the 3Di-tokens introduced by Foldseek. In a first step, ProstT5 learnt to represent the newly introduced 3Di-tokens by continuing the original span-denoising objective applied on 3Di- and amino acid- (AA) sequences. Only in a second step, ProstT5 was trained on translating between the two modalities. The direction of the translation is indicated by two special tokens ("\ " for translating from 3Di to AAs, “\ ” for translating from AAs to 3Di). To avoid clashes with AA tokens, 3Di-tokens were cast to lower-case (alphabets are identical otherwise). 1. The model can be used for traditional feature extraction. For this, we recommend using only the encoder in half-precision (fp16) together with batching. Examples (currently only for original ProtT5-XL-U50 but replacing repository links and adding prefixes works): script and colab While original ProtT5-XL-U50 could only embed AA sequences, ProstT5 can now also embed 3D structures represented by 3Di tokens. 3Di tokens can either be derived from 3D structures using Foldseek or they can be predicted from AA sequences by ProstT5. 3. "Folding": Translation from sequence (AAs) to structure (3Di). The resulting 3Di strings can be used together with Foldseek for remote homology detection while avoiding to compute 3D structures explicitly. 4. "Inverse Folding": Translation from structure (3Di) to sequence (AA).

译文排队中，当前展示来源原文。

语言翻译license:mit
上下文窗口来源未披露
输入 → 输出文本 → 文本
模型权重待核验
档案更新时间2026/10/29 17:21:16
OVERVIEW

模型介绍

ProstT5 is a protein language model (pLM) which can translate between protein sequence and structure. ProstT5 (Protein structure-sequence T5) is based on ProtT5-XL-U50, a T5 model trained on encoding protein sequences using span corruption applied on billions of protein sequences. ProstT5 finetunes ProtT5-XL-U50 on translating between protein sequence and structure using 17M proteins with high-quality 3D structure predictions from the AlphaFoldDB. Protein structure is converted from 3D to 1D using the 3Di-tokens introduced by Foldseek. In a first step, ProstT5 learnt to represent the newly introduced 3Di-tokens by continuing the original span-denoising objective applied on 3Di- and amino acid- (AA) sequences. Only in a second step, ProstT5 was trained on translating between the two modalities. The direction of the translation is indicated by two special tokens ("\ " for translating from 3Di to AAs, “\ ” for translating from AAs to 3Di). To avoid clashes with AA tokens, 3Di-tokens were cast to lower-case (alphabets are identical otherwise). 1. The model can be used for traditional feature extraction. For this, we recommend using only the encoder in half-precision (fp16) together with batching. Examples (currently only for original ProtT5-XL-U50 but replacing repository links and adding prefixes works): script and colab While original ProtT5-XL-U50 could only embed AA sequences, ProstT5 can now also embed 3D structures represented by 3Di tokens. 3Di tokens can either be derived from 3D structures using Foldseek or they can be predicted from AA sequences by ProstT5. 3. "Folding": Translation from sequence (AAs) to structure (3Di). The resulting 3Di strings can be used together with Foldseek for remote homology detection while avoiding to compute 3D structures explicitly. 4. "Inverse Folding": Translation from structure (3Di) to sequence (AA).

主要特点

特点摘要待依据来源补齐，不凭模型名称推断能力。

局限与注意事项

暂无足够来源支持具体缺点结论。请重点验证实际任务表现、时延、数据政策、许可与版本生命周期。

PRICING & COST

来源报价与费用分析

同一模型在不同渠道、地区和配置下可能采用不同价格。下表独立保留各渠道报价，不将渠道价格视为厂商官方统一价格。

来源 / 渠道输入输出计费单位采集时间
暂未取得可核验报价，请查阅来源渠道。

按月估算基础 Token 费用

—

预估不包含缓存优惠、阶梯定价、多模态附加费用、网络检索、税费及推理额外输出。

INDEPENDENT BENCHMARKS

公开评测与运行配置

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

查看完整已收录榜单 →

常见问题

开放权重是否代表可以免费商用？

不一定。应核对具体版本许可证、商用门槛和使用政策。自部署还会产生算力和维护成本。

这些价格是实时结算价格吗？

不是实时结算价格，而是带采集时间的来源报价快照。请到对应渠道核对当前版本、区域、折扣和计费方式。

资料来源与可追溯记录

huggingface · Rostlab/ProstT5 ↗2026/10/29 08:00:56
huggingface · enrich:a7ab4b24-a96f-4658-b223-427a68a9e173 ↗2026/10/29 17:21:16

许可证：mit。模型介绍与第三方评测均以原来源的具体版本和方法为准。