Skip to content

updating esm2 + geneformer to run benchmarks with data from node specific scratch #957

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Jun 25, 2025
+31 −2
Merged

updating esm2 + geneformer to run benchmarks with data from node specific scratch #957

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
d3b87a6
updating esm2 + geneformer to run benchmarks with data from node spec…
dorotat-nv Jun 24, 2025
5099d3f
Merge branch 'main' into dorotat/adjust-training-steps
dorotat-nv Jun 25, 2025
1cc6b50
Merge branch 'main' into dorotat/adjust-training-steps
dorotat-nv Jun 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 15 additions & 0 deletions ci/benchmarks/partial-conv/esm2_pretrain.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ time_limit: 14400
key_segments:
# Modify keys to be renamed (str) or excluded (False) from run identifier. By default, all args under script_args are included.
data_path: False
num_workers: False
script_args:
# All arguments referenced in the script string must be specified here.
# Arguments not referenced in the script string must have the 'arg' field specified.
Expand All @@ -18,7 +19,20 @@ script_args:
batch_size: 16
max_steps: 500000
stop_steps: 26000
num_workers: 1
script: |-
COPY_FLAG="/tmp/copy_done_${{SLURMD_NODENAME}}";
NEW_DATA_PATH="/dev/shm/data_path_${{SLURMD_NODENAME}}";
if [ "$SLURM_LOCALID" = "0" ]; then
df -h;
echo $NEW_DATA_PATH;
time cp -r ${data_path}/ $NEW_DATA_PATH;
touch $COPY_FLAG
fi
# All ranks wait until install flag file appears
while [ ! -f $COPY_FLAG ]; do
sleep 1
done
WANDB_API_KEY=$BIONEMO_WANDB_API_KEY ${variant}_${model} \
--train-cluster-path=${data_path}/train_clusters.parquet \
--train-database-path=${data_path}/train.db \
Expand All @@ -27,6 +41,7 @@ script: |-
--micro-batch-size=${batch_size} \
--num-nodes=${nodes} \
--num-gpus=${gpus} \
--num-dataset-workers=${num_workers} \
--val-check-interval=1000 \
--limit-val-batches=1 \
--num-steps=${max_steps} \
Expand Down
18 changes: 16 additions & 2 deletions ci/benchmarks/partial-conv/geneformer_pretrain.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ key_segments:
data_path: False
val_check_interval: False
lr: False
num_workers: False
script_args:
# All arguments referenced in the script string must be specified here.
# Arguments not referenced in the script string must have the 'arg' field specified.
Expand All @@ -22,15 +23,28 @@ script_args:
lr: 0.001
val_check_interval: 500
acc_grad: 1
num_workers: 4
script: |-
WANDB_API_KEY=$BIONEMO_WANDB_API_KEY ${variant}_${model} \
COPY_FLAG="/tmp/copy_done_${{SLURMD_NODENAME}}";
NEW_DATA_PATH="/dev/shm/data_path_${{SLURMD_NODENAME}}";
if [ "$SLURM_LOCALID" = "0" ]; then
df -h;
echo $NEW_DATA_PATH;
time cp -r ${data_path}/ $NEW_DATA_PATH;
touch $COPY_FLAG
fi
# All ranks wait until install flag file appears
while [ ! -f $COPY_FLAG ]; do
sleep 1
done
WANDB_API_KEY=$BIONEMO_WANDB_API_KEY ${variant}_${model} \
--data-dir ${data_path} \
--experiment-name ${batch_size}bs_${nodes}node_${gpus}gpu_${max_steps}s_${precision}prec \
--num-gpus ${gpus} \
--save-last-checkpoint \
--num-nodes ${nodes} \
--val-check-interval ${val_check_interval} \
--num-dataset-workers 8 \
--num-dataset-workers ${num_workers} \
--num-steps ${max_steps} \
--seq-length 2048 \
--limit-val-batches 8 \
Expand Down