Can’t download the original MS-Celeb-1M dataset?

[LiuJoffrey]

Hi, I would like to download the original datasets, but I can’t find the download link from the website. Could you please provide the original dataset file or the download link for me. Please.

Thank you so much

[EB-Dodo]

It is a pity that we have lost the original image data due to our carelessness in data preserving in the last two years. This cleaned file list here is what we have now. We actually did not know that Microsoft Research has taken down the original data until we see these issues.

[jjsjunior]

Hi, I would like to download the original datasets, but I can’t find the download link from the website. Could you please provide the original dataset file or the download link for me. Please.

Thank you so much

Hi Joffrey, Did you managed to find out a download link for MS-Celeb-1M? thanks in advance

[youthM]

Hi, I would like to download the original datasets, but I can’t find the download link from the website. Could you please provide the original dataset file or the download link for me. Please. Thank you so much

Hi Joffrey, Did you managed to find out a download link for MS-Celeb-1M? thanks in advance

Hi , do you find out a download link for MS-Celeb-1M?

[ha1990-12]

https://academictorrents.com/details/9e67eb7cc23c9417f39778a8e06cca5e26196a97/tech&hit=1&filelist=1

[ibarrond]

Hi, how do you process the tsv you get from this torrent? I'm not sure what each column contains or how to process it.

[ketan-b]

This should do the task of extracting the images from .TSV

import argparse
import base64
import csv
import os
# import magic # Detect image type from buffer contents (disabled, all are jpg)

parser = argparse.ArgumentParser()
parser.add_argument('--croppedTSV', type=str)
parser.add_argument('--outputDir', type=str, default='raw')
args = parser.parse_args()

with open(args.croppedTSV, 'r') as tsvF:
    reader = csv.reader(tsvF, delimiter='\t')
    i = 0
    for row in reader:
        MID, imgSearchRank, faceID, data = row[0], row[1], row[4], base64.b64decode(row[-1])

        saveDir = os.path.join(args.outputDir, MID)
        savePath = os.path.join(saveDir, "{}-{}.jpg".format(imgSearchRank, faceID))

        # assert(magic.from_buffer(data) == 'JPEG image data, JFIF standard 1.01')

        os.makedirs(saveDir, exist_ok=True)
        with open(savePath, 'wb') as f:
            f.write(data)

        i += 1

        if i % 1000 == 0:
            print("Extracted {} images.".format(i))