BenchmarkText8 数据集
Text8
Text8 is a character-level language modeling benchmark dataset preprocessed from Wikipedia text, commonly used to evaluate the performance of generative models on language modeling tasks.
Timeline (last 90 days)
Oct 21
On language modeling tasks (Text8, TinyStories, OpenWebText), binary diffusion remains competitive but does not significantly surpass the baselines.
Unverified50%