datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
unet-jerymotro
UNet Jery Motro — Burned Area Dataset
Dataset de patches satellitaires multi-canaux pour la détection et la segmentation sémantique des zones brûlées à Madagascar dans le cadre du projet Jery Motro.
Structure des Données (.npz)
Chaque fichier .npz contient les tenseurs et métadonnées suivants :
image : Tenseur de forme (256, 256, 6) de type float32 (canaux spectraux / features).
mask : Masque binaire de segmentation de forme (256, 256) de type uint8 (0 = sain, 1… See the full description on the dataset page: https://huggingface.co/datasets/rtsikynyantsa/unet-jerymotro.jerymotro-unet-madagascar
Jery Motro — Dataset d'entraînement U-Net (Madagascar)
Dataset géospatial multi-canal pour la segmentation sémantique de surfaces
brûlées à Madagascar, produit dans le cadre du projet
Jery Motro —
un système de détection et d'alerte précoce des feux de forêt basé sur
l'apprentissage profond.
Résumé
Propriété
Valeur
Exemples
~7 800 patchs
Résolution spatiale
500 m/pixel
Taille des patchs
256 × 256 pixels
Couverture géographique
Madagascar (bbox :… See the full description on the dataset page: https://huggingface.co/datasets/rtsikynyantsa/jerymotro-unet-madagascar.reconstruction2_unetv2_luna16SelfReg-UNet13803867589-unet-image-seg
刺绣图像分割数据集
用于 U-Net 语义分割的刺绣图像数据集。
使用方法
from datasets import load_dataset
# 加载完整数据集 (835张)
ds = load_dataset("tari-tech/13803867589-unet-image-seg", data_dir="full")
# 加载去除AI图的数据集 (763张)
ds = load_dataset("tari-tech/13803867589-unet-image-seg", data_dir="no-ai")
# 加载 SAM3 重新标注的数据集 (563张)
ds = load_dataset("tari-tech/13803867589-unet-image-seg", data_dir="sam3")
# 加载 SAM3 标注后人工过滤的数据集 (425张)
ds = load_dataset("tari-tech/13803867589-unet-image-seg", data_dir="sam3-label")… See the full description on the dataset page: https://huggingface.co/datasets/tari-tech/13803867589-unet-image-seg.zzsn-activations-3_unet.up_blocks.1.attentions.1unethical_sharegpt_human_assistantunet-jerymotro-datasetdatasettraining_setting_burnt_unet_and_text_encoderdiscord_roasting_unethicalthis dataset is a short group of seq2seq inputs and responses.
it currently contains <50 rows.
this dataset is very heteful and should be used at your own risk.
it is formatted with a rude prompt or insult as the input and a equaly unethical and rude but funny response.
this dataset should be used to train private models if the use case is to fight haters.
use at your own risk, this will probably really hurt someone's feelings :(
0908_no_unethical_behavior_rm_pipeline_05_filter_pairszzsn-activations-2_unet.up_blocks.1.attentions.20907_no_unethical_behavior_response_gen_responses-prepared0907_no_unethical_behavior_rm_pipeline_02_responses_50kq0908_no_unethical_behavior_rm_pipeline_03_pairs0908_no_unethical_behavior_judgement_gen_judgements_train-prepared0908_no_unethical_behavior_judgement_gen_judgements_train-raw-responses0908_no_unethical_behavior_judgement_gen_judgements_test-prepared0908_no_unethical_behavior_rm_pipeline_04_training_data_processor_test0908_no_unethical_behavior_rm_pipeline_04_training_data0908_no_unethical_behavior_filter_judgement_gen_judgements_train-prepared0909_no_unethical_behavior_rm_data_with_filter_and_reward_scoreunethical_sharegpt_human_assistant_flatMes_Poumons_3D_unetLUNA16_VALIDATION_UNETllama_free_unethical_2zzsn-activations-1_unet.up_blocks.1.attentions.20907_no_unethical_behavior_rm_pipeline_02_responses0908_no_unethical_behavior_filter_judgement_gen_judgements_test-raw-responses