search

NervanaSystems / neon

Intel® Nervana™ reference deep learning framework committed to best performance on all hardware
http://neon.nervanasys.com/docs/latest
Apache License 2.0
3.87k stars 811 forks source link

how can i add my own collection of images on neon to train a model ? #400

Open campos537 opened 7 years ago

campos537 commented 7 years ago

Hey guys , i'm trying to add my own collection of images to train the model using the binary network, but i'm having a lot of issues. So if you guys could help me guiding me to the best and easy way to do so it would be pretty helpful, thanks.

wei-v-wang commented 7 years ago

@campos537 Can you please start with examples/binary ? It uses MNIST dataset.

For advanced data ingestion, please also refer to aeon (which is part of neon and is automatically installed). https://github.com/nervanasystems/aeon

You might also want to take a look at examples/imagenet/ and examples/ssd examples/faster-rcnn to get an idea how the general ingestion process looks like.

campos537 commented 7 years ago

Thanks , i will take o look on aeon, i was able to train a model using a dataset on .h5 extension converting to numpy array

-- Crystal Silva Campos Graduando em Engenharia da Computa????o Universidade Estadual de Feira de Santana/Bahia Lattes: http://lattes.cnpq.br/2563260359267761https://wwws.cnpq.br/cvlattesweb/PKG_MENU.menu?f_cod=CC3509E4E54FEB21774DC166D4ACCDE3#

De: Wei Wang notifications@github.com Enviado: segunda-feira, 16 de outubro de 2017 22:44 Para: NervanaSystems/neon Cc: Crystal S Campos; Mention Assunto: Re: [NervanaSystems/neon] how can i add my own collection of images on neon to train a model ? (#400)

@campos537https://github.com/campos537 Can you please start with examples/binary ? It uses MNIST dataset.

For advanced data ingestion, please also refer to aeon (which is part of neon and is automatically installed). https://github.com/nervanasystems/aeon

You might also want to take a look at examples/imagenet/ and examples/ssd examples/faster-rcnn to get an idea how the general ingestion process looks like.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/NervanaSystems/neon/issues/400#issuecomment-337065381, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AadWAkl1YdhjHNstxVrgdNCeVpOwu5fLks5ss9xkgaJpZM4Pc4HZ.