ExpertFlowPredictor/xsum_deepseek-moe-16b-chat_moe_patterns

这是一个包含文本和对应编码ID的数据集，分为seqlen1024和test两部分。seqlen1024部分包含2000个示例，test部分包含100个示例。数据集的特征包括文本内容(prompt_text)、文本内容的编码ID(prompt_ids)、解码后的ID(decode_ids)以及文本内容的模式(prompt_pattern)和解码后的模式(decode_pattern)。整个数据集的大小为8909023437字节。

This is a dataset containing text and corresponding encoded IDs, split into two parts: seqlen1024 and test. The seqlen1024 part contains 2000 examples, and the test part contains 100 examples. The features of the dataset include the text content (prompt_text), the encoded IDs of the text (prompt_ids), the decoded IDs (decode_ids), the pattern of the text content (prompt_pattern), and the pattern after decoding (decode_pattern). The total size of the dataset is 8909023437 bytes.