Vietnamese Vision-Language Dataset With 700k Samples

700,000 Vietnamese vision-language samples were generated using Gemini Pro and prompt engineering techniques like few-shot learning and caption-based prompting. The dataset was created by Vi-VLM and was last updated in June 2024.

Use Cases

• Train a visual question answering model using Llava-style prompts generated from COCO images.
• Fine-tune a Vietnamese image captioning model on translated prompts derived from the ShareGPT4V dataset.
• Benchmark few-shot learning performance for Vietnamese vision-language tasks using the provided prompt-engineered samples.

Strengths

• Contains over 700,000 samples, providing a substantial base for model training.
• Leverages established datasets like COCO and ShareGPT4V as foundations for its generation process.

Limitations

• All data is synthetically generated by an AI model (Gemini Pro), which may introduce artifacts or biases.
• Specific row counts, column names, and data distributions for the generated samples are not provided.

Provenance

Source

Generated by Gemini Pro using COCO and ShareGPT4V datasets as sources.

Collection Method

Employed prompt engineering techniques including few-shot learning, caption-based prompting, and image-based prompting.

Freshness

Last updated on 2024-06-25.