LLaVA-CC3M-Pretrain-595K-ZH: Chinese Machine-Translated Multimodal Training Data

A Chinese-language dataset of 595,000 items, created by machine-translating the LLaVA-CC3M-Pretrain-595K dataset. The data was uploaded by author 'cyberlangke' to Hugging Face and last updated on 2026-02-25. The description notes the translations are unverified and may contain errors.

Use Cases

• Fine-tuning Chinese-language visual question answering models based on the translated multimodal data.
• Pretraining multimodal large language models (LLMs) for Chinese-language tasks.
• Studying the effects of machine translation on multimodal instruction-following data quality.
• Benchmarking the performance of multimodal models on translated versus original English data.

Strengths

• Contains 595,000 translated data items for Chinese-language model development.
• Based on the established LLaVA-CC3M-Pretrain-595K dataset, suggesting a structured multimodal format.

Limitations

• The description explicitly states the translations are unverified and may contain errors.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other scale details like file size and format are unknown.

Provenance

Source

A machine-translated version of the 'liuhaotian/LLaVA-CC3M-Pretrain-595K' dataset.

Collection Method

Created using a large language model (LLM) for translation.

Freshness

Last updated 2026-02-25 10:04:13; freshness should be verified.