transfer-learning
code_trans_t5_base_code_documentation_generation_javascript_transfer_learning_finetunecode_trans_t5_base_code_comment_generation_java_transfer_learning_finetunecode_trans_t5_small_code_documentation_generation_javascript_transfer_learning_finetunecode_trans_t5_base_code_documentation_generation_python_transfer_learning_finetunecode_trans_t5_base_code_documentation_generation_java_transfer_learning_finetunecode_trans_t5_base_source_code_summarization_python_transfer_learning_finetunecode_trans_t5_large_code_documentation_generation_java_transfer_learning_finetunecode_trans_t5_large_code_documentation_generation_javascript_transfer_learning_finetune
spiqa-simplified-for-fuyu8b-transfer-learning
Steps:
You need to download the images.zip from the files and unzip it to ./SPIQA_images as showed in code below
Data entries refer to images in local unzipped location, along with question on image and then answer.
import os
import json
import requests
import zipfile
from tqdm import tqdm
def download_file(url, filename):
response = requests.get(url, stream=True)
total_size = int(response.headers.get('content-length', 0))
with open(filename, 'wb') as file… See the full description on the dataset page: https://huggingface.co/datasets/sugiv/spiqa-simplified-for-fuyu8b-transfer-learning.ntl_transfer_learning
Dataset Description
This dataset was created to support transfer-learning applications in remote sensing, with a primary focus on the Polish context but with potential applicability in other regions exhibiting similar land-cover characteristics.
It consists of 444,483 Sentinel-2 Level-2A (L2A) image tiles, each covering an area of 1,200 × 1,200 meters (120 × 120 pixels). Every tile is paired with a Nighttime Lights (NTL) intensity value for 2022, derived from NASA’s Black Marble… See the full description on the dataset page: https://huggingface.co/datasets/LukaszJanisiow/ntl_transfer_learning.