File size: 3,317 Bytes
6c12b06 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 | #!/usr/bin/env bash
set -uo pipefail
ROOT=/lustre/fsw/portfolios/nvr/users/jtremblay/yu
RUN=$ROOT/1027_run; TASKS=$ROOT/1027_sim_tasks
V5CODE=$TASKS/share_v5_sim2real_20260929/code
CONDA=$ROOT/miniconda3; SPY=$CONDA/envs/simfoundry/bin/python; UPY=$CONDA/envs/ur5dp/bin/python
HOME_FAKE=$RUN/home_fake; NAME=ringpeg_dart_v1; RDIR=$TASKS/ringpeg
YUASSET=$TASKS/assets/yuasset_v1_docs_only
mkdir -p $HOME_FAKE/ur5_sim2real_v5 $RUN/logs $V5CODE/sim_data $V5CODE/train_runs
ln -sfn $CONDA $HOME_FAKE/miniconda3
ln -sfn $TASKS/share_v5_sim2real_20260929 $HOME_FAKE/ur5_sim2real_v5/share_v5_sim2real_20260929
ln -sfn $RDIR $HOME_FAKE/ur5_sim2real_v5/ringpeg
ln -sfn $YUASSET $HOME_FAKE/ur5_sim2real_v5/yuasset
export HOME=$HOME_FAKE
log() { echo "== $(date '+%F %T') $*"; }
source $CONDA/etc/profile.d/conda.sh
export TORCHINDUCTOR_COMPILE_THREADS=1 OMNIGIBSON_HEADLESS=1 OMNI_KIT_ACCEPT_EULA=YES PYTHONUNBUFFERED=1
export OMNIGIBSON_RENDER_VIEWER_CAMERA=0 TORCH_COMPILE_DISABLE=1 TORCHDYNAMO_DISABLE=1
export OMNIGIBSON_DATA_PATH=/lustre/fsw/portfolios/nvr/users/jtremblay/yu/SimFoundry/deps/BEHAVIOR-1K/datasets
DATA_DIR=$V5CODE/sim_data/$NAME
DATA_A=${DATA_DIR}_a; DATA_B=${DATA_DIR}_b
mkdir -p $DATA_A $DATA_B
log "Collecting ringpeg: 2x parallel (GPU0 seeds 50-149, GPU2 seeds 150-249) resuming from 50 existing"
CUDA_VISIBLE_DEVICES=0 OMNIGIBSON_GPU_ID=0 \
$SPY -u $RDIR/ringpeg_collect.py \
--out $DATA_A --seeds 50-149 --max-kept 55 \
--layouts 0 --jitter 0.03 --ring-yaw-deg 30 --peg-yaw-deg 15 --dart --start-offset \
2>&1 | tee $RUN/logs/${NAME}_a.log &
PID_A=$!
sleep 60
CUDA_VISIBLE_DEVICES=2 OMNIGIBSON_GPU_ID=2 \
$SPY -u $RDIR/ringpeg_collect.py \
--out $DATA_B --seeds 150-249 --max-kept 55 \
--layouts 0 --jitter 0.03 --ring-yaw-deg 30 --peg-yaw-deg 15 --dart --start-offset \
2>&1 | tee $RUN/logs/${NAME}_b.log &
PID_B=$!
wait $PID_A; log "Proc A done"
wait $PID_B; log "Proc B done"
# Merge new shards with the existing 50 demos
python3 $RUN/merge_shards.py $DATA_DIR $DATA_A $DATA_B $DATA_DIR 100
kept=$($UPY -c "import json; d=json.load(open('$DATA_DIR/collect_meta.json')); print(d['n_kept'])" 2>/dev/null || echo 0)
log "Kept: $kept"
[ "$kept" -lt 50 ] && { log "ERROR: only $kept"; exit 1; }
log "Packing zarr"
cd $V5CODE && $UPY ur5_import/scripts/pack_zarr.py --dir sim_data/$NAME 2>&1 | tee -a $RUN/logs/${NAME}_train.log
log "Training DP ($kept demos)"
cd $V5CODE/ur5_lab_code && export PYTHONPATH=src:.
$UPY train.py --config-name=train_baseline_diffusion_unet task=sim_baseline \
task_name=$NAME dataset_path=$DATA_DIR n_demo=$kept \
training.device=cuda:0 training.seed=0 training.num_epochs=130 training.checkpoint_every=129 \
checkpoint.topk.k=0 training.resume=True checkpoint.save_last_ckpt=True dataloader.batch_size=256 dataloader.num_workers=16 logging.mode=offline \
"policy._target_=algorithms.policy.diffusion_unet_baseline_aug_policy.DiffusionUnetBaselineAugPolicy" \
+policy.cam1_dropout=0.1 +policy.cam1_aug_p=0.5 +policy.cam1_erase_p=0.3 +policy.cam0_aug_p=0.3 \
"hydra.run.dir=$V5CODE/train_runs/${NAME}_s0" \
2>&1 | tee $RUN/logs/${NAME}_train.log
rc=$?; log "Train exit $rc"
[ $rc -eq 0 ] && touch $RUN/data/TRAIN_DONE_${NAME} || { log "TRAIN FAILED"; exit 1; }
log "DONE: $V5CODE/train_runs/${NAME}_s0/checkpoints/"
|