through batch: 1. launch container 2. grab .txt from s3 3. change it 4. push it back to s3 flowchartbatchs3
@alee156
docker run -ti alee156/batch:1, in batch command line "/usr/bin/bash/python albert.py"
👍
fixing the tensor display function with Jupyter notebook displaying the fix FSLDTKFSL fixed1
@lukezhu1
👍
Demo results of pipeline: Notebook showing running script ipynb, resulting outputs (as html files, download and open in a browser) here. Latest Docker image, with newest ndreg/ndio/openCV. Latest finalized script for pipeline.
@jonl1096
👍
Demo results of pipeline: Notebook showing running script ipynb, resulting outputs (as html files, download and open in a browser) here. Image showing changes made to current pipeline. Comparison of before and after for CLAHE (on Fear brain). Latest Docker image, with newest ndreg/ndio/openCV. Latest finalized script for pipeline.
Try and see how we can run Kwame's code. Managed to successfully do it on Cortex w/ versions of ndio/ndreg that were the ones Kwame used. Noted that the process is significantly slower than when Kwame does it, making it prohibitive for testing our pipeline... Made a new Docker image (3x leaner/smaller than the old Docker image) with the right file dependencies (ndreg, ndio, opencv2) and also sets all presets for quickly hosting a Jupyter server. Figuring out batch stuff...
@jonl1096
👍
Try and see how we can run Kwame's code. Managed to successfully do it on Cortex w/ versions of ndio/ndreg that were the ones Kwame used. Noted that the process is significantly slower than when Kwame does it, making it prohibitive for testing our pipeline... Made edits to pipeline script to include new "run_pipeline" command that theoretically should be able to run from start to finish (so can be called in cmd line, this will help when we do Batch stuff). Figuring out batch stuff...
Committed
For Monday
For Thursday
Note: To run batch
https://628818565354.signin.aws.amazon.com/console
628818565354 Administrator
aws batch submit-job --job-name jovolicious --job-queue demo --job-definition realdemo:1