Datasets:

MingSafeR
/

miniloop-o-final-omni-data

Parquet error: Scan size limit exceeded: attempted to read 5754142963 bytes, limit is 300000000 bytes Make sure that 1. the Parquet files contain a page index to enable random access without loading entire row groups2. otherwise use smaller row-group sizes when serializing the Parquet files

Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

MiniLoop-O Omni Data Mirror

Upstream: jingyaogong/minimind-o_dataset

This public repository mirrors the full (non-mini) sft_*.parquet training assets used by MiniMind-O so the training pipeline does not depend on a second remote dataset later.

The upstream dataset card currently carries both apache-2.0 and gpl-3.0 license tags. This mirror does not attempt to relicense upstream content. Preserve upstream attribution/notices and review the original dataset card.

The official schema may include:

conversations
question_audios
answer_audios
image_bytes
ref_audios
spk_emb

answer_audios is already stored as flattened interleaved Mimi 8-codebook tokens, so Talker training does not re-encode target speech through Mimi.

Downloads last month: 51