Open papar22 opened 8 years ago
Hi,
groundhog isn't supported or development anymore. We use blocks. See the warning on this page:
https://github.com/lisa-groundhog/GroundHog
On Thu, Dec 10, 2015 at 1:31 PM, papar22 notifications@github.com wrote:
Hi everybody,
I want to validate the training progress using perplexity after some iterations. As it has been mentioned in mainLoop.py, I have to specify "param valid_data: data iterator used for validation". But I don't find the Iterator class for validation. As I used the same like train_data, I get an error considering the queue. It would be appreciated, if someone let me know, how to validate my training which needs validation iterator.
Thanks
— Reply to this email directly or view it on GitHub https://github.com/lisa-groundhog/GroundHog/issues/42.
To be honest, I knew that. I am using blocks as well. But I need some changes in groundHog as I already have some models and impelementations using it. That' s why I asked this question.
Ok, maybe someone else can answer, I don't know groundhog, so I can't easily help you.
On Thu, Dec 10, 2015 at 4:07 PM, papar22 notifications@github.com wrote:
To be honest, I knew that. I am using blocks as well. But I need some changes in groundHog as I already have some models and impelementations using it. That' s why I asked this question.
— Reply to this email directly or view it on GitHub https://github.com/lisa-groundhog/GroundHog/issues/42#issuecomment-163749543 .
@papar22 Hi, did you solve this problem? I am using this lisa-groundhog, and I think it's necessary to use validate data. Could you help? Thanks.
Hi, if any of you guys have an answer to this questions I'd love to know. Right now it seems to me that there is no way to run validation between epochs. Thanks! Lotem
Just a note, this repo isn't supported anymore and isn't developed.
Just to let you know.
On Wed, Nov 2, 2016 at 7:56 AM, Lotemp notifications@github.com wrote:
Hi, if any of you guys have an answer to this questions I'd love to know. Right now it seems to me that there is no was to run validation between epochs. Thanks! Lotem
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/lisa-groundhog/GroundHog/issues/42#issuecomment-257844398, or mute the thread https://github.com/notifications/unsubscribe-auth/AALC-14nI40vQRDtyTFqi4xQMpRbOKpTks5q6HqKgaJpZM4Gy-el .
Hi everybody,
I want to validate the training progress using perplexity after some iterations. As it has been mentioned in mainLoop.py, I have to specify "param valid_data: data iterator used for validation". But I don't find the Iterator class for validation. As I used the same like train_data, I get an error considering the queue. It would be appreciated, if someone let me know, how to validate my training which needs validation iterator.
Thanks