acq.sh

#!/bin/bash
#SBATCH --job-name=genedisco_special_acquisition
#SBATCH --gres=gpu:20GB:1
#SBATCH --cpus-per-gpu=24
#SBATCH --mem=100G
#SBATCH --time=48:00:00
#SBATCH --partition=long
#SBATCH --error=/home/mila/c/chris.emezue/genedisco/slurmerror_special_acquisition_test.txt
#SBATCH --output=/home/mila/c/chris.emezue/genedisco/slurmoutput_special_acquisition_test.txt

###########cluster information above this line
module load python/3.8
module load cuda/11.1/cudnn/8.0
source /home/mila/c/chris.emezue/genedisco/genv/bin/activate
export CUBLAS_WORKSPACE_CONFIG=:4096:8
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/home/mila/c/chris.emezue/genedisco/genv/lib"

CACHE_NAME=cache_test
OUTPUT_NAME=output_test
CACHE_DIR=/home/mila/c/chris.emezue/genedisco/${CACHE_NAME}
OUTPUT_DIR=/home/mila/c/chris.emezue/genedisco/${OUTPUT_NAME}

mkdir -p $CACHE_DIR
mkdir -p $OUTPUT_DIR

#With KDE
active_learning_loop  \
    --cache_directory=$CACHE_DIR  \
    --output_directory=$OUTPUT_DIR  \
    --model_name="bayesian_mlp" \
    --acquisition_function_name="custom" \
    --acquisition_function_path=/home/mila/c/chris.emezue/genedisco/genedisco/active_learning_methods/acquisition_functions/rnd.py \
    --acquisition_batch_size=64 \
    --num_active_learning_cycles=8 \
    --feature_set_name="achilles" \
    --dataset_name="schmidt_2021_ifng" 

#CACHE_NAME=cache_no_kde
#OUTPUT_NAME=output_no_kde
#CACHE_DIR=/home/mila/c/chris.emezue/genedisco/${CACHE_NAME}
#OUTPUT_DIR=/home/mila/c/chris.emezue/genedisco/${OUTPUT_NAME}

#mkdir -p $CACHE_DIR
#mkdir -p $OUTPUT_DIR

#Without KDE
#active_learning_loop  \
#    --cache_directory=$CACHE_DIR  \
#    --output_directory=$OUTPUT_DIR  \
 #   --model_name="bayesian_mlp" \
 #   --acquisition_function_name="coreset" \
 #   --acquisition_batch_size=64 \
 #   --num_active_learning_cycles=8 \
 #   --feature_set_name="achilles" \
 #   --dataset_name="schmidt_2021_ifng"