視覺理解論文系列(三)VL-BERT: PRE-TRAINING OF GENERIC VISUALLINGUISTIC REPRESENTATIONS
背景 這是微軟亞研院的工作，將純文本設定的bert擴展到visual-linguistic場景，從預訓練到fine-tune，可以用於多個下游任務。 摘要 作者提出了一個可預訓練的用於學習視覺-語言任務通用表徵的模型VL-BERT，VL-BERT以transformers爲主幹，可以同時接受V、L特徵作爲輸入。預訓練任務使用了包括visual-language數據集Conceptual Capti
相關文章
- 1. 視覺理解論文系列(四)ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Languag
- 2. Contextual Word Representations and Pretraining
- 3. 【論文閱讀】 VL-BERT: Pre-training of generic visual-linguistic representations
- 4. 【cs224n-12】Modeling contexts of use: Contextual Representations and Pretraining. ELMo and BERT.
- 5. NLP系列：Word2Vec原始論文：Efficient Estimation of Word Representations in Vector Space
- 6. 論文閱讀：DeepWalk Online Learning of Social Representations
- 7. 論文筆記----DeepWalk: Online Learning of Social Representations
- 8. 論文筆記：DeepWalk: Online Learning of Social Representations
- 9. NLP基礎之——Contextual Word Representations and Pretraining
- 10. 論文寫作解讀：A Simple Framework for Contrastive Learning of Visual Representations
- 更多相關文章...
- • CAP理論是什麼？ - NoSQL教程
- • jQuery Mobile 列表視圖 - jQuery Mobile 教程
- • 三篇文章瞭解 TiDB 技術內幕——說存儲
- • 三篇文章瞭解 TiDB 技術內幕 —— 說計算
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. 視覺理解論文系列(四)ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Languag
- 2. Contextual Word Representations and Pretraining
- 3. 【論文閱讀】 VL-BERT: Pre-training of generic visual-linguistic representations
- 4. 【cs224n-12】Modeling contexts of use: Contextual Representations and Pretraining. ELMo and BERT.
- 5. NLP系列：Word2Vec原始論文：Efficient Estimation of Word Representations in Vector Space
- 6. 論文閱讀：DeepWalk Online Learning of Social Representations
- 7. 論文筆記----DeepWalk: Online Learning of Social Representations
- 8. 論文筆記：DeepWalk: Online Learning of Social Representations
- 9. NLP基礎之——Contextual Word Representations and Pretraining
- 10. 論文寫作解讀：A Simple Framework for Contrastive Learning of Visual Representations