Upload code/1028_run/tasks/package_d5.sh with huggingface_hub

code/1028_run/tasks/package_d5.sh

@@ -0,0 +1,91 @@
+#!/usr/bin/env bash
+# package_d5.sh (1028_run) — 4x parallel collection + AMP training
+set -uo pipefail
+ROOT=/lustre/fsw/portfolios/nvr/users/jtremblay/yu
+RUN=$ROOT/1028_run; TASKS=$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=$RUN/home_fake; NAME=package_d5_v1; PKG=$TASKS/package_d5
+log() { echo "== $(date '+%F %T') $*"; }
+
+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 $PKG $HOME_FAKE/ur5_sim2real_v5/bread_lean 2>/dev/null || true
+export HOME=$HOME_FAKE
+source $CONDA/etc/profile.d/conda.sh
+export TORCHINDUCTOR_COMPILE_THREADS=1 OMNIGIBSON_HEADLESS=1 OMNI_KIT_ACCEPT_EULA=YES PYTHONUNBUFFERED=1
+export OMNIGIBSON_RENDER_VIEWER_CAMERA=0 TORCH_COMPILE_DISABLE=1 TORCHDYNAMO_DISABLE=1
+
+OG_DATA=$RUN/og_data/package_d5
+if [ ! -d $OG_DATA ]; then bash $PKG/setup_og_data.sh "$ROOT/SimFoundry/deps/BEHAVIOR-1K/datasets" "$OG_DATA"; fi
+DS_FSW=/lustre/fsw/portfolios/nvr/users/jtremblay/yu/SimFoundry/deps/BEHAVIOR-1K/datasets
+for x in behavior-1k-assets omnigibson-robot-assets omnigibson.key; do rm -f $OG_DATA/$x && ln -sfn $DS_FSW/$x $OG_DATA/$x; done
+for c in $DS_FSW/real2sim-assets/objects/*; do nm=$(basename $c); tgt=$OG_DATA/real2sim-assets/objects/$nm; [ -L "$tgt" ] && rm -f "$tgt" && ln -sfn "$c" "$tgt" || true; done
+export STEP40_OG_DATA=$OG_DATA OMNIGIBSON_DATA_PATH=$OG_DATA
+export UR5_IMPORT=$V5CODE/ur5_import OBJECT_ASSETS=$PKG/object_assets
+export STEP40_DATA=$PKG/data OMNIGIBSON_APPDATA_PATH=$PKG/og_appdata
+
+DATA_A=$RUN/data/${NAME}_1028a; DATA_B=$RUN/data/${NAME}_1028b
+DATA_C=$RUN/data/${NAME}_1028c; DATA_D=$RUN/data/${NAME}_1028d
+mkdir -p $DATA_A $DATA_B $DATA_C $DATA_D
+
+log "Collecting package_d5: 4x parallel (GPU 0/1/2/3), 15 each"
+
+OMNIGIBSON_GPU_ID=0 \
+$SPY -u $PKG/scripts/t40_collect.py \
+    --out $DATA_A --seeds 300-399 --max-kept 15 \
+    --expert-cfg $PKG/cfg/expert_lean.json --dart --start-offset \
+    2>&1 | tee $RUN/logs/${NAME}_1028a.log &
+PID_A=$!
+sleep 55
+
+OMNIGIBSON_GPU_ID=1 \
+$SPY -u $PKG/scripts/t40_collect.py \
+    --out $DATA_B --seeds 400-499 --max-kept 15 \
+    --expert-cfg $PKG/cfg/expert_lean.json --dart --start-offset \
+    2>&1 | tee $RUN/logs/${NAME}_1028b.log &
+PID_B=$!
+sleep 55
+
+OMNIGIBSON_GPU_ID=2 \
+$SPY -u $PKG/scripts/t40_collect.py \
+    --out $DATA_C --seeds 500-599 --max-kept 15 \
+    --expert-cfg $PKG/cfg/expert_lean.json --dart --start-offset \
+    2>&1 | tee $RUN/logs/${NAME}_1028c.log &
+PID_C=$!
+sleep 55
+
+OMNIGIBSON_GPU_ID=3 \
+$SPY -u $PKG/scripts/t40_collect.py \
+    --out $DATA_D --seeds 600-699 --max-kept 20 \
+    --expert-cfg $PKG/cfg/expert_lean.json --dart --start-offset \
+    2>&1 | tee $RUN/logs/${NAME}_1028d.log &
+PID_D=$!
+
+wait $PID_A; log "GPU0 done"
+wait $PID_B; log "GPU1 done"
+wait $PID_C; log "GPU2 done"
+wait $PID_D; log "GPU3 done"
+
+FINAL=$RUN/data/${NAME}_1028
+python3 $RUN/merge_shards.py $DATA_A $DATA_B $DATA_C $DATA_D $FINAL 50
+kept=$($UPY -c "import json; d=json.load(open('$FINAL/collect_meta.json')); print(d['n_kept'])" 2>/dev/null || echo 0)
+log "Kept: $kept"; [ "$kept" -lt 40 ] && { log "ERROR: only $kept"; exit 1; }
+
+ln -sfn $FINAL $V5CODE/sim_data/${NAME}_1028
+log "Packing zarr"
+cd $V5CODE && $UPY ur5_import/scripts/pack_zarr.py --dir sim_data/${NAME}_1028 2>&1 | tee $RUN/logs/${NAME}_1028_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=${NAME}_1028 dataset_path=$FINAL n_demo=$kept \
+    training.device=cuda:0 training.seed=0 training.num_epochs=130 training.checkpoint_every=129 \
+    checkpoint.topk.k=0 training.resume=True checkpoint.save_last_ckpt=True \
+    training.use_amp=True dataloader.batch_size=256 dataloader.num_workers=16 logging.mode=offline \
+    "policy._target_=algorithms.policy.diffusion_unet_baseline_aug_policy.DiffusionUnetBaselineAugPolicy" \
+    +policy.cam1_dropout=0.1 +policy.cam1_aug_p=0.5 +policy.cam1_erase_p=0.3 +policy.cam0_aug_p=0.3 \
+    "hydra.run.dir=$V5CODE/train_runs/${NAME}_1028_aug_s0" \
+    2>&1 | tee -a $RUN/logs/${NAME}_1028_train.log
+rc=$?; log "Train exit $rc"; [ $rc -eq 0 ] || { log "TRAIN FAILED"; exit 1; }
+log "DONE: $V5CODE/train_runs/${NAME}_1028_aug_s0/checkpoints/"