PACT-Net / run_all.sh
rk-random's picture
Upload folder using huggingface_hub
9a67fbe verified
raw
history blame
1.19 kB
#!/bin/bash
# --- 1. Define your parameters here ---
# Special case for your model
PACTNET_MODEL="polyatomic"
PACTNET_REP="polyatomic"
# Baseline models and representations
BASELINE_MODELS=("gcn" "gin" "gat" "sage")
BASELINE_REPS=("smiles" "selfies" "ecfp")
# Datasets for all experiments
DATASETS=("esol" "freesolv" "lipophil" "boilingpoint" "qm9" "ic50" "bindingdb")
# --- 2. Generate all valid commands ---
# We use a subshell `( ... )` to group the output of two separate loops.
# This combined output is then piped to GNU Parallel.
(
# Loop 1: Generate commands for your PACTNet model
echo "--- Generating PACTNet Jobs ---" >&2
for dataset in "${DATASETS[@]}"; do
echo "python3 main.py --model $PACTNET_MODEL --rep $PACTNET_REP --dataset $dataset"
done
# Loop 2: Generate commands for all baseline combinations
echo "--- Generating Baseline GNN Jobs ---" >&2
for model in "${BASELINE_MODELS[@]}"; do
for rep in "${BASELINE_REPS[@]}"; do
for dataset in "${DATASETS[@]}"; do
echo "python3 main.py --model $model --rep $rep --dataset $dataset"
done
done
done
) | \
caffeinate -s parallel -j 10 --bar --eta
echo "All jobs complete."