Machine learningDeep learning / NLP / CV

Transformer Đa phương thức

Transformer Đa phương thức mở rộng kiến trúc Transformer tiêu chuẩn để xử lý và suy luận chung trên hai hoặc nhiều phương thức đầu vào — phổ biến nhất là văn bản và hình ảnh, nhưng cũng có thể là âm thanh, video hoặc dữ liệu có cấu trúc. Các lớp chú ý chéo phương thức cho phép thông tin từ một phương thức ảnh hưởng đến biểu diễn của phương thức khác, cho phép các tác vụ như trả lời câu hỏi bằng hình ảnh, tạo chú thích ảnh và phân tích cảm xúc đa phương thức.

Mở trong MethodMindSắp ra mắtVideoSắp ra mắtDownload slides

Đọc toàn bộ phương pháp

Chỉ dành cho thành viên

Đăng nhập bằng tài khoản miễn phí để đọc phần này.

Đăng nhập

Method map

The neighbourhood of related methods — select a node to explore.

+15 more

Nguồn tài liệu

  1. Lu, J., Batra, D., Parikh, D., & Lee, S. (2019). ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks. Advances in Neural Information Processing Systems (NeurIPS), 32. link
  2. Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., ... & Sutskever, I. (2021). Learning Transferable Visual Models From Natural Language Supervision. Proceedings of the 38th International Conference on Machine Learning (ICML), PMLR 139. link

Cách trích dẫn trang này

ScholarGate. (2026, June 3). Multimodal Transformer (Cross-Modal Attention-Based Architecture). ScholarGate. https://scholargate.app/vi/deep-learning/multimodal-transformer

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side

Được tham chiếu bởi

ScholarGateMultimodal Transformer (Multimodal Transformer (Cross-Modal Attention-Based Architecture)). Truy cập ngày 2026-06-15 từ https://scholargate.app/vi/deep-learning/multimodal-transformer · Bộ dữ liệu: https://doi.org/10.5281/zenodo.20539026