sdxl text encoder

sdxl使用了两个不同的text encoder，分别是694m的openclip vit-bigG和354m的openai clip vit-l/14。这两个text encoder在sdxl中的倒数第二层提取特征，并将它们进行拼接，拼接后的特征维度为2048。这个拼接后的特征被称为sdxl的context dim。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* *2* *3* [SDXL：Improving latent diffusion models for high-resolution image synthesis](https://blog.csdn.net/u012193416/article/details/132390358)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v93^chatsearchT3_2"}}] [.reference_item style="max-width: 100%"]
[ .reference_list ]