mikeogezi/negopt_full

该数据集用于优化文本到图像生成中的负面提示，以增强美学和保真度。数据来源于Playground，并用于微调论文中提出的模型。数据集包含多个特征，如id、创建时间、提示、负面提示、点赞数、采样器、高度、步骤、宽度、光标、URL、配置比例和模型。数据集分为训练集，包含256224个示例，总大小为248764537字节。This dataset is used to optimize negative prompts in text-to-image generation to enhance aesthetics and fidelity. The data is sourced from Playground and used to fine-tune the models proposed in the paper. The dataset includes features such as id, created_at, prompt, negative_prompt, likes, sampler, height, steps, width, cursor, url, cfg_scale, and model. The dataset is split into a training set with 256224 examples and a total size of 248764537 bytes.