jkfindeisen
jkfindeisen
That is standard in many other packages and also quite useful to check models. It also would allow to use models without an analytical derivative at the cost of performance...
Example: Michele Scipioni suggests in https://github.com/gpufit/Gpufit/pull/35 that a model for PET pharmocokinetics is merged to this repository here (gpufit/Gpufit). It might be useful for some although already specific and require...
Those fitting applications which could benefit from Gpufit (PALM, STORM, ... but not limited) help them using Gpufit. For example by forking and introducing Gpufit. This may include re-design since...
If for some reason the first step is too large (too large lambda?), the parameters are way off after the first iteration and then never recover, but still chi_square-previous_chi_square will...
This would make packaging easier. Maybe with http://buildbot.net/
If you try to use gpuArray (the Matlab gpu computation) **after** gpufit has been used (for example to speed up other parts of the script), an error is thrown. >...
We have currently version strings duplicated at various places: - bindings (Python, Matlab) - documentation - packaging - source code (or at least it should be there) - examples (or...
Currently we count all iterations as iterations even if they do not change the parameter sets (only change the lambda for example). Others like CMinpack do not count these iterations...
See PyCUDA (https://mathema.tician.de/software/pycuda/) for a library claiming to do this as well as the documentation for runtime compilation (http://docs.nvidia.com/cuda/nvrtc/index.html) for more information.
See PyCUDA (https://mathema.tician.de/software/pycuda/) for a library claiming to do this as well as the documentation for runtime compilation (http://docs.nvidia.com/cuda/nvrtc/index.html) for more information.