hiraksarkar

use ndarray::prelude::*;

fn std1d(a: ArrayView1<'_, f64>) -> f64 {
    let n = a.len() as f64;
    if n == 0. {
        return 0.;
    }
    let mean = a.sum() / n;
    (a.fold(0., |acc, &x| acc + (x - mean).powi(2)) / n).sqrt()
}

hiraksarkar

Keybase proof

I hereby claim:

• I am hiraksarkar on github.
• I am hiraksarkar (https://keybase.io/hiraksarkar) on keybase.
• I have a public key whose fingerprint is 4152 B3E7 7BE1 83B0 1E53 8B80 2EDB 3721 028D 527A

To claim this, I am signing this object:

hiraksarkar

#!/usr/bin/bash

tfile=$1
usr=`rg "(\d+\.\d+)user" $tfile -or '$1'`
sys=`rg "(\d+\.\d+)system" $tfile -or '$1'`
elap=`rg "([\d:\.]+)elapsed" $tfile -or '$1'`
mem=`rg "(\d+)maxresident" $tfile -or '$1'`

tot_elap=`echo "$elap" | sed -E 's/(.*):(.+):(.+)/\1*3600+\2*60+\3/;s/(.+):(.+)/\1*60+\2/' | bc`

hiraksarkar

plt.figure(figsize=(8, 8))
df = np.log(variance_trio_df[['var_mean','merged_var']]+1)
df.loc[:,'color'] = 'b'
df.loc[(df.merged_var - df.var_mean) < -1, 'color'] = 'r'
colors = ["b", "r"]
customPalette = sns.set_palette(sns.color_palette(colors))
ax = sns.scatterplot(
    x = 'var_mean', 
    y = 'merged_var',
    data = df,

hiraksarkar

Note

I have changed only two directory names from the original post to be found here

Open account with ngrok.com

Get the autorization code after creating a free account with ngork. You will meet with a screen like following

Install stuff (in colab)

hiraksarkar

#!/bin/bash
## This gist contains instructions about cuda v11.2 and cudnn8.1 installation in Ubuntu 20.04 for Pytorch 1.8 & Tensorflow 2.7.0

### steps ####
# verify the system has a cuda-capable gpu
# download and install the nvidia cuda toolkit and cudnn
# setup environmental variables
# verify the installation
###

hiraksarkar

#!/bin/bash
#SBATCH --mincpus 16
#SBATCH --mem 64G
#SBATCH --time 5:00:00
#SBATCH --job-name mendel
#SBATCH --mail-type=begin        # send email when job begins
#SBATCH --mail-type=end          # send email when job ends
#SBATCH [email protected]
#SBATCH --output jupyter_logs/jupyter-notebook-%J.log
# get tunneling info

hiraksarkar

#!/bin/bash
#SBATCH --mincpus 32
#SBATCH --mem 100G
#SBATCH --time 6-23:59:00
#SBATCH --job-name jupyterlab
#SBATCH --gres=gpu:1
#SBATCH --mail-type=begin        # send email when job begins
#SBATCH --mail-type=end          # send email when job ends
#SBATCH [email protected]
#SBATCH --output jupyter_logs/jupyter-notebook-%J.log

hiraksarkar

Basic

Installing R packages is painful, but conda environment solved a lot of problems. Basically, if you install your own R in conda, and the later R command install.packages() will automatically install the packages in the environment; in addition, conda has many system libraries too for getting away from requiring sudo permissions.

After creating an empty conda environment, you can install a specific version (say 4.2) of R by conda install -c conda-forge r-essentials=4.2. If you are not sure whether that version exists in conda, you can do conda search r-essentials. Using your own R in the conda environment, you can do the normal R installation commands.

Some R packages search system libraries by pkg-config. After you install the required libraries through conda, you can check whether your PKG_CONFIG_PATH includes /envs//lib/pkgconfig, and set the path properly.

hiraksarkar

mamba create -n r43
mamba activate r43
mamba install r-essentials=4.3
mamba install r-rjags
export PKG_CONFIG_PATH=/home/user/miniconda3/envs/r43/lib/pkgconfig/:$PKG_CONFIG_PATH

# start R
R

# Install R package