datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
fasal-mitra-sft-v1
Fasal Mitra SFT Dataset v1
Fasal Mitra ("Farmer's Friend") is a supervised fine-tuning (SFT) dataset for training
multilingual crop disease advisory AI models, specifically Gemma 4 E4B via Unsloth.
Dataset Summary
This dataset enables a vision-language model to diagnose crop diseases from photos and provide
structured agricultural advisory in Assamese, Hindi, Bengali, and English — the primary
languages of farmers in Northeast India (Assam, Meghalaya, Nagaland).… See the full description on the dataset page: https://huggingface.co/datasets/phoenix28/fasal-mitra-sft-v1.dino-sft-dataset
Chrome Dinosaur Game Dataset
用于训练 Chrome 恐龙游戏 AI 的数据集。
数据格式
{
"messages": [
{"role": "user", "content": "<image>Based on the current game state shown in the image, what action should the dinosaur take? Answer with one word only: Jump, Duck, or Run."},
{"role": "assistant", "content": "Jump"}
],
"images": ["dino_images_v3/frame_xxx.png"]
}
数据统计
总样本数: 3233
Jump (跳跃): 44.3%
Run (奔跑): 40.0%
Duck (下蹲): 15.7%
使用方法
与 LLaMA-Factory 配合使用进行… See the full description on the dataset page: https://huggingface.co/datasets/Sp1keeee/dino-sft-dataset.