Generating Images from Captions with Attention
注意從字幕生成圖像 基於生成模型的最新進展，我們引入了一種根據自然語言描述生成圖像的模型。 提出的模型以迭代方式在畫布上繪製補丁，同時注意描述中的相關單詞。在接受Microsoft COCO培訓後，我們將我們的模型與幾個基準生成模型進行了圖像生成和檢索任務比較。 我們證明，與其他方法相比，我們的模型可以產生更高質量的樣本，並且可以生成具有新穎場景組成的圖像，這些圖像與數據集中以前看不到的字幕相對應
相關文章
- 1. text to image（一）:《GENERATING IMAGES FROM CAPTIONS WITH ATTENTION》
- 2. Every Picture Tells a Story: Generating Sentences from Images
- 3. Transferring GANs: generating images from limited data 論文學習
- 4. Beyond Narrative Description: Generating Poetry from Images by Multi-Adversarial Training
- 5. 論文理解：Generating Diverse High-Fidelity Images with VQ-VAE-2
- 6. Working with Docker Images
- 7. CRAFT Objects from Images
- 8. Generating Videos With Scene Dynamics
- 9. big code: code2seq Generating Sequences from Structured Representations of Code
- 10. 【讀】seq2seq——（1）Generating News Headlines with Recurrent Neural Networks
- 更多相關文章...
- • Docker images 命令 - Docker命令大全
- • XSLT
- • Docker 清理命令
- • 爲了進字節跳動，我精選了29道Java經典算法題，帶詳細講解
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. text to image（一）:《GENERATING IMAGES FROM CAPTIONS WITH ATTENTION》
- 2. Every Picture Tells a Story: Generating Sentences from Images
- 3. Transferring GANs: generating images from limited data 論文學習
- 4. Beyond Narrative Description: Generating Poetry from Images by Multi-Adversarial Training
- 5. 論文理解：Generating Diverse High-Fidelity Images with VQ-VAE-2
- 6. Working with Docker Images
- 7. CRAFT Objects from Images
- 8. Generating Videos With Scene Dynamics
- 9. big code: code2seq Generating Sequences from Structured Representations of Code
- 10. 【讀】seq2seq——（1）Generating News Headlines with Recurrent Neural Networks