search

pluskid / Mocha.jl

Deep Learning framework for Julia
Other
1.29k stars 254 forks source link

how to use method dump_statistics #178

Closed ashleylid closed 8 years ago

ashleylid commented 8 years ago

Hi,

I get an error when running update_statistics(solver,:obj_value, :obj_val)

LoadError: MethodError: update_statistics has no method matching update_statistics(::Mocha.Solver{Mocha.SGD}, ::Symbol, ::Symbol) Closest candidates are: update_statistics(::Any, !Matched::AbstractString, !Matched::AbstractFloat) while loading In[19], in expression starting on line 2

The API shows:

dump_statistics(storage, layer_state, show)

storage is a data storage (typically a CoffeeLounge object) that is used to dump statistics into, via the function update_statistics(storage, key, value).

show is a boolean value, when true, indicating that a summary of the statistics should also be printed to stdout.

How do I use this?

EDIT: I would like to scatter plot predicted value and actual value over my epoch. Is there a way to access these two things? I updated today and see now there are 3 options for plot_statistics.jl rather than 4.

ashleylid commented 8 years ago

Looks like I have to move over to MXNet and start again..

pluskid commented 8 years ago

The statistics is usually not used to save values like predictions. To save the predictions, you can use a memory-output or hdf5-output layer to capture the predictions, and then inspect them later.

ashleylid commented 8 years ago

Aaaah thank you. Is this the same for MXNet.jl? Moving over there as directed.

pluskid commented 8 years ago

No MXNet.jl is simpler for doing this. See the end of the README at the home page of MXNet.jl (https://github.com/dmlc/MXNet.jl).