pycoal

pycoal copied to clipboard

Provide More Detailed Algorithm Descriptions

1

It would be beneficial to have more thorough descriptions of what each part of the program does internally, even if it's at a high level. This would especially make sense...

theoilie

@aheermann it turns out that the joblib test is failing... https://travis-ci.org/capstone-coal/pycoal/jobs/619064202 I think this is due to there never having been mineral classifications made using SAM_joblib... therefore the results are...

lewismc

Investigate running Pytorch on GPU

5

Hi @TinyBugBigProblem I'm opening this on your behalf. I understand you are working on GPU acceleration for Pytorch. Depending on your environment the setup and installation of the stack can...

lewismc

Investigate running Pytorch on GPU

4

I think I managed to fix the problem I was running into. The approach I had before didn't seem to take into account how Pytorch handled wrapping the Model class...

TinyBugBigProblem

Provide Unified Metadata Model (UMM)-compliant product metadata

1

Metadata can be represented in a number of standards and formats. Many discipline-specific or community-specific metadata standards have been developed to support data management and data discovery systems and to...

lewismc

Support ECOSTRESS Spectral Library

2

https://speclib.jpl.nasa.gov/

lewismc

As noted in the mineral.py SAM algorithms, we want to have the resampler be able to modify the dimensions of an image, and scale the units accordingly. https://github.com/capstone-coal/pycoal/blob/34a97fa255d7bb116999ccbae7cded77e917e789/pycoal/mineral.py#L162-L164

aheermann

Enable use of EcoSIS Spectral Library

1

[EcoSIS](https://ecosis.org) is a spectral library with some >54K spectra... we may wish to leverage this to improve our classifications.

lewismc

Use dask to speed up SAM algorithm in mineral.py

8

The SAM algorithm in mineral.py takes the majority of time when trying to classify an image, specifically these loops. ``` # for each pixel in the image for x in...

aheermann

Replace numpy.ndarray.flatten(...) with numpy.ndarray.ravel(...) for performance and efficiency

1

https://github.com/capstone-coal/pycoal/blob/dd461f6b22578c5aeebec2e6cae10ad0babc5c1c/examples/example_sam.py#L71 Hi 👋 I noticed that the code uses numpy.ndarray.flatten(...) to flatten arrays, for example: `angle_list = list(numpy.ndarray.flatten(angles))` While this is functionally correct, it would be more efficient to use...

SaFE-APIOpt