How to train the 20170512-110547 model by using MS-Celeb-1M

[jhj7905]

My goal is training the model(20170512-110547) provided from davidsandberg by using MS-Celeb-1M. From now, I downloaded MS-Celeb-1M dataset(cropped and aligned) and knew that learning_rate_schedule_classifier_msceleb.txt and decode_msceleb_dataset.py were needed for preprocessing. Please give me a detailed explanation how to use the source code in facenet for training

Thank u in advance

[jhj7905]

one more thing!! source code in decode_msceleb_dataset.py file format is below ◦Column1: Freebase MID ◦Column2: Query/Name ◦Column3: ImageSearchRank ◦Column4: ImageURL ◦Column5: PageURL ◦Column6: ImageData_Base64Encoded

in http://www.msceleb.org/download/aligned ◦Column1: Freebase MID ◦Column2: ImageSearchRank ◦Column3: ImageURL ◦Column4: PageURL ◦Column5: FaceID ◦Column6: FaceRectangle_Base64Encoded (four floats, relative coordinates of UpperLeft and BottomRight corner) ◦Column7: FaceData_Base64Encoded d Data](https://msceleb.blob.core.windows.net/ms-celeb-v1-samples/MsCelebV1-Faces-Aligned.Samples.zip): 14 entities, 32MBceleb.blob.core.windows.net/ms-celeb-v1-samples/MsCelebV1-Faces-Aligned.Samples.zip): 14 entities, 32MB

is it okay to use this source code for decoding the db?

[jhj7905]

@davidsandberg

[AppleCoffee]

Hi，have you solved this problem? I have the same problem now, can you give me some advice? Thanks !

[ninnyyan]

hello, have you solved this issue? I have the same problem with you and looking for some advice now, thank you~