zsdonghao / text-to-image

Generative Adversarial Text to Image Synthesis / Please Star -->
https://github.com/zsdonghao/tensorlayer
599 stars 162 forks source link

How to load the model #20

Open ZihengZheng opened 6 years ago

ZihengZheng commented 6 years ago

Hello, I was successful to train the model but I can not load that, could you please help me thank you so much!

screen shot 2018-07-16 at 18 12 38 screen shot 2018-07-16 at 18 12 43
zsdonghao commented 6 years ago

it seems the npz is not match with your network architecture.

ZihengZheng commented 6 years ago

But i just use the network from github

Get Outlook for iOShttps://aka.ms/o0ukef


From: Hao notifications@github.com Sent: Monday, July 16, 2018 9:40 PM To: zsdonghao/text-to-image Cc: (pg) Ziheng Zheng; Author Subject: Re: [zsdonghao/text-to-image] How to load the model (#20)

it seems the npz is not match with your network architecture.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/zsdonghao/text-to-image/issues/20#issuecomment-405374809, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AZgDC2DFAGnH6EHg89e-WFKr07sshFq4ks5uHPpCgaJpZM4VRfll.

zsdonghao commented 6 years ago

I didn't provide pre-trained model for this repo

ZihengZheng commented 6 years ago

Yes i trained the model with your code

Get Outlook for iOShttps://aka.ms/o0ukef


From: Hao notifications@github.com Sent: Monday, July 16, 2018 9:44 PM To: zsdonghao/text-to-image Cc: (pg) Ziheng Zheng; Author Subject: Re: [zsdonghao/text-to-image] How to load the model (#20)

I didn't provide pre-trained model for this repo

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/zsdonghao/text-to-image/issues/20#issuecomment-405375737, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AZgDC80c8cma5NVm2n_0eXbDb2MODI0rks5uHPsCgaJpZM4VRfll.

zsdonghao commented 6 years ago

if you use my code to train the model, and use the same code to load npz, but it fails, I don't have any idea about it ...

maybe you can try update tensorlayer.

ZihengZheng commented 6 years ago

But when I trained the network, the rnn_loss became to 0. Is that reasonable?


From: Hao notifications@github.com Sent: 16 July 2018 22:12:17 To: zsdonghao/text-to-image Cc: (pg) Ziheng Zheng; Author Subject: Re: [zsdonghao/text-to-image] How to load the model (#20)

if you use my code to train the model, and use the same code to load npz, but it fails, I don't have any idea about it ...

maybe you can try update tensorlayer.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/zsdonghao/text-to-image/issues/20#issuecomment-405383670, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AZgDC6YgCLspyswLKI4UTmpEX1QST25tks5uHQGxgaJpZM4VRfll.

jamiechoi1995 commented 6 years ago

@ZihengZheng https://github.com/zsdonghao/text-to-image/blob/275880d95b3d0366cbaefe24019c1decddc9d48c/train_txt2im.py#L207 the rnn_loss is set to be 0 after epoch 50

ZihengZheng commented 6 years ago

Yes, I found it but why it need to be set to 0?

Get Outlook for iOShttps://aka.ms/o0ukef


From: Jamie notifications@github.com Sent: Friday, July 27, 2018 5:22 AM To: zsdonghao/text-to-image Cc: (pg) Ziheng Zheng; Mention Subject: Re: [zsdonghao/text-to-image] How to load the model (#20)

@ZihengZhenghttps://github.com/ZihengZheng https://github.com/zsdonghao/text-to-image/blob/275880d95b3d0366cbaefe24019c1decddc9d48c/train_txt2im.py#L207 the rnn_loss is set to be 0 after epoch 50

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/zsdonghao/text-to-image/issues/20#issuecomment-408299967, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AZgDCy2x0ZkeaXchlNVgJZhJpQ7p8dxLks5uKodKgaJpZM4VRfll.

Fuhongshuai commented 5 years ago

I use the same code, but can not reproduce the results, the flowers generated are disorderly, why is this? @zsdonghao Thank you!

anastasiagkou commented 5 years ago

@Fuhongshuai did you find any solution about this? Because i have the same problem. Thank you!

Siddanth-pai commented 4 years ago

@Fuhongshuai @anastasiagkou did you find any solution to overcome this problem.