Giter Site home page Giter Site logo

hessianfree's People

Contributors

drasmuss avatar fhrozen avatar studywolf avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

hessianfree's Issues

time delays to rnn

Thanks for sharing. I was wondering if is possible to add time delays to the recurrent connections.

did not match C++ signature

hi, i wondering about the gpu program.
I know that is still a WIP but i am not sure if you know about these problems:
First:
using the test_demo
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: stopping after 1 failures !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
========================================================= 1 failed in 2.39 seconds =========================================================
PyCUDA WARNING: a clean-up operation failed (dead context maybe?)
cuModuleUnload failed: an illegal memory access was encountered
PyCUDA WARNING: a clean-up operation failed (dead context maybe?)
cuModuleUnload failed: an illegal memory access was encountered
PyCUDA WARNING: a clean-up operation failed (dead context maybe?)
cuModuleUnload failed: an illegal memory access was encountered
PyCUDA WARNING: a clean-up operation failed (dead context maybe?)
cuModuleUnload failed: an illegal memory access was encountered
PyCUDA WARNING: a clean-up operation failed (dead context maybe?)
cuModuleUnload failed: an illegal memory access was encountered

So probably could be a memory allocation. I did not check where the problem is.

Also trying to run the program but with my own data i got this
Boost.Python.ArgumentError: Python argument types in
Function._set_block_shape(Function, numpy.int32, numpy.int32, int)
did not match C++ signature:
_set_block_shape(pycuda::function {lvalue}, int, int, int)
and when i modify the code trying to match ints i got this:
No registered converter was able to produce a C++ rvalue of type unsigned int from this Python object of type numpy.int32

so i am not sure what is wrong.

I did not try to check for more specific errors because i dont have much time.
Also, I was trying to modify the code moving all of the operations to the gpu since the longest the train goes, the slowest it becomes. But I do not have much time to check. However i was trying with the scikit-cuda lib that includes the cublas operations.

As soon as ill get free i will try to move the code to cublas..
Thanks

error for test_rnnet.py : AssertionError:

thank you very much , google shows you are only source of simple python hesian free rnn example
It would be great to use your code , but
I use
pytest.version
'2.5.2'

but have error for test_rnnet.py
File "C:\Sander\deep ML\Hessian_Free\hessianfree-master_Mar3\hessianfree\tests\test_rnnet.py", line 234, in
pytest.main("-x -v --tb=native test_rnnet.py")
File "C:\Anaconda\Lib\site-packages_pytest\config.py", line 20, in main
return config.hook.pytest_cmdline_main(config=config)
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 377, in call
return self._docall(methods, kwargs)
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 388, in docall
res = mc.execute()
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 289, in execute
res = method(**kwargs)
File "C:\Anaconda\Lib\site-packages_pytest\main.py", line 112, in pytest_cmdline_main
return wrap_session(config, main)
File "C:\Anaconda\Lib\site-packages_pytest\main.py", line 77, in wrap_session
config.do_configure()
File "C:\Anaconda\Lib\site-packages_pytest\config.py", line 608, in do_configure
self.hook.pytest_configure(config=self)
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 377, in call
return self.docall(methods, kwargs)
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 388, in docall
res = mc.execute()
File "C:\Anaconda\Lib\site-packages_pytest\core.py", line 289, in execute
res = method(**kwargs)
File "C:\Anaconda\Lib\site-packages_pytest\assertion__init
.py", line 56, in pytest_configure
warn_about_missing_assertion(mode)
File "C:\Anaconda\Lib\site-packages_pytest\assertion__init
.py", line 113, in warn_about_missing_assertion
assert False
File "C:\Anaconda\Lib\site-packages_pytest\assertion\reinterpret.py", line 40, in init
self.msg = reinterpret(source, f, should_fail=True)
File "C:\Anaconda\Lib\site-packages_pytest\assertion\newinterpret.py", line 49, in interpret
visitor.visit(mod)
File "C:\Anaconda\Lib\ast.py", line 241, in visit
return visitor(node)
File "C:\Anaconda\Lib\site-packages_pytest\assertion\newinterpret.py", line 144, in visit_Module
self.visit(stmt)
File "C:\Anaconda\Lib\ast.py", line 241, in visit
return visitor(node)
File "C:\Anaconda\Lib\site-packages_pytest\assertion\newinterpret.py", line 314, in visit_Assert
raise BuiltinAssertionError

AssertionError:

No module named _winreg

Hi.

My name is Nelson Yalta from Waseda University, and also working with second gradient optimizations.

I was trying to run your library but i got a problem of =

File "test.py", line 493, in
for a in sys.argv[2:]])
File "test.py", line 23, in test_xor
ff = HessianFF([2, 5, 1], debug=True, use_GPU=False)
File "/home/nelson-lab/hessianfree-master/hessianfree/hessianff.py", line 65, in init
layers = [copy.deepcopy(layers) for _ in range(self.n_layers)]
File "/usr/lib/python2.7/copy.py", line 190, in deepcopy
y = _reconstruct(x, rv, 1, memo)
File "/usr/lib/python2.7/copy.py", line 334, in _reconstruct
state = deepcopy(state, memo)
File "/usr/lib/python2.7/copy.py", line 163, in deepcopy
y = copier(x, memo)
File "/usr/lib/python2.7/copy.py", line 257, in _deepcopy_dict
y[deepcopy(key, memo)] = deepcopy(value, memo)
File "/usr/lib/python2.7/copy.py", line 178, in deepcopy
rv = reductor(x)
File "/usr/lib/python2.7/dist-packages/numpy/core/init.py", line 61, in _ufunc_reduce
return _ufunc_reconstruct, (whichmodule(func, name), name)
File "/usr/lib/python2.7/pickle.py", line 817, in whichmodule
if name != 'main' and getattr(module, funcname, None) is func:
File "/usr/lib/python2.7/dist-packages/scipy/lib/six.py", line 116, in getattr
_module = self._resolve()
File "/usr/lib/python2.7/dist-packages/scipy/lib/six.py", line 105, in _resolve
return _import_module(self.mod)
File "/usr/lib/python2.7/dist-packages/scipy/lib/six.py", line 76, in _import_module
import(name)
ImportError: No module named _winreg,

so I was wondering if you made this for windows or did you use in linux. and also if i am lacking of a library.

Best Regards

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.