search

yvquanli / GLAM

Code for "An adaptive graph learning method for automated molecular interactions and properties predictions".
https://www.nature.com/articles/s42256-022-00501-8
MIT License
39 stars 8 forks source link

How to know the parameters of the model? #8

Closed StefanIsSmart closed 11 months ago

StefanIsSmart commented 1 year ago

In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately?

yvquanli commented 1 year ago

This framework is an automatic deep learning framework, so it automatically searches for suitable hyperparameters. You don't have to predefine any hyperparameters; you only need to prepare the dataset input.

S @.***> 于2023年8月13日周日 23:29写道：

In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately?

— Reply to this email directly, view it on GitHub https://github.com/yvquanli/GLAM/issues/8, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M . You are receiving this because you are subscribed to this thread.Message ID: @.***>

StefanIsSmart commented 1 year ago

This framework is an automatic deep learning framework, so it automatically searches for suitable hyperparameters. You don't have to predefine any hyperparameters; you only need to prepare the dataset input. S @.> 于2023年8月13日周日 23:29写道： … In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately? — Reply to this email directly, view it on GitHub <#8>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M . You are receiving this because you are subscribed to this thread.Message ID: @.>

Thanks for your reply! I know it's a framework for automatic learning. I mean, what is the number of parameters that the final searched model typically contains?

yvquanli commented 1 year ago

The range of parameter count is between 10k to 20M, depending on the architecture of the final model, which is determined by the hyperparameters selected by the automated process.

Because it is an automatic deep learning framework that requires a massive amount of computation, comparing it to a single model would be unfair to the single model.

S @.***> 于2023年8月14日周一 15:53写道：

This framework is an automatic deep learning framework, so it automatically searches for suitable hyperparameters. You don't have to predefine any hyperparameters; you only need to prepare the dataset input. S @.

> 于2023年8月13日周日 23:29写道： … <#m7142466433438237589> In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately? — Reply to this email directly, view it on GitHub <#8 https://github.com/yvquanli/GLAM/issues/8>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M . You are receiving this because you are subscribed to this thread.Message ID: @.>

Thanks for your reply! I know it's a framework for automatic learning. I mean, what is the number of parameters that the final searched model typically contains?

— Reply to this email directly, view it on GitHub https://github.com/yvquanli/GLAM/issues/8#issuecomment-1676847646, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIAU5B2KPGBZJPREXULXVHKQPANCNFSM6AAAAAA3OW555M . You are receiving this because you commented.Message ID: @.***>

yvquanli commented 1 year ago

Just run is OK , you can find the parameter counts in the log.txt

S @.***> 于2023年8月14日周一 16:44写道：

The range of parameter count is between 10k to 20M, depending on the architecture of the final model, which is determined by the hyperparameters selected by the automated process. Because it is an automatic deep learning framework that requires a massive amount of computation, comparing it to a single model would be unfair to the single model. S @.

> 于2023年8月14日周一 15:53写道： … <#m2748887156724053990> This framework is an automatic deep learning framework, so it automatically searches for suitable hyperparameters. You don't have to predefine any hyperparameters; you only need to prepare the dataset input. S @. > 于2023年8月13日周日 23:29写道： … <#m7142466433438237589> In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately? — Reply to this email directly, view it on GitHub <#8 https://github.com/yvquanli/GLAM/issues/8 <#8 https://github.com/yvquanli/GLAM/issues/8>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M . You are receiving this because you are subscribed to this thread.Message ID: @.> Thanks for your reply! I know it's a framework for automatic learning. I mean, what is the number of parameters that the final searched model typically contains? — Reply to this email directly, view it on GitHub <#8 (comment) https://github.com/yvquanli/GLAM/issues/8#issuecomment-1676847646>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIAU5B2KPGBZJPREXULXVHKQPANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIAU5B2KPGBZJPREXULXVHKQPANCNFSM6AAAAAA3OW555M . You are receiving this because you commented.Message ID: @.>

Thanks a lot!

— Reply to this email directly, view it on GitHub https://github.com/yvquanli/GLAM/issues/8#issuecomment-1676926817, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDZUILVZQKNDIFXJCTXVHQPZANCNFSM6AAAAAA3OW555M . You are receiving this because you commented.Message ID: @.***>

StefanIsSmart commented 1 year ago

Just run is OK , you can find the parameter counts in the log.txt S @.> 于2023年8月14日周一 16:44写道： … The range of parameter count is between 10k to 20M, depending on the architecture of the final model, which is determined by the hyperparameters selected by the automated process. Because it is an automatic deep learning framework that requires a massive amount of computation, comparing it to a single model would be unfair to the single model. S @. > 于2023年8月14日周一 15:53写道： … <#m2748887156724053990> This framework is an automatic deep learning framework, so it automatically searches for suitable hyperparameters. You don't have to predefine any hyperparameters; you only need to prepare the dataset input. S @. > 于2023年8月13日周日 23:29写道： … <#m7142466433438237589> In the Supplementary information of the article, I can find the parameters of the model. Can you tell me what the parameters of the model are approximately? — Reply to this email directly, view it on GitHub <#8 <#8> <#8 <#8>>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIDUJQNQOBXF2GFOUP3XVDXHLANCNFSM6AAAAAA3OW555M . You are receiving this because you are subscribed to this thread.Message ID: @.> Thanks for your reply! I know it's a framework for automatic learning. I mean, what is the number of parameters that the final searched model typically contains? — Reply to this email directly, view it on GitHub <#8 (comment) <#8 (comment)>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIAU5B2KPGBZJPREXULXVHKQPANCNFSM6AAAAAA3OW555M https://github.com/notifications/unsubscribe-auth/ALFSIIAU5B2KPGBZJPREXULXVHKQPANCNFSM6AAAAAA3OW555M . You are receiving this because you commented.Message ID: @.> Thanks a lot! — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALFSIIDZUILVZQKNDIFXJCTXVHQPZANCNFSM6AAAAAA3OW555M . You are receiving this because you commented.Message ID: @.>

Thank you a lot!