| set -x | |
| GPUS=${GPUS:-8} | |
| PORT=${PORT:-29500} | |
| if [ $GPUS -lt 8 ]; then | |
| GPUS_PER_NODE=${GPUS_PER_NODE:-$GPUS} | |
| else | |
| GPUS_PER_NODE=${GPUS_PER_NODE:-8} | |
| fi | |
| CPUS_PER_TASK=${CPUS_PER_TASK:-5} | |
| OUTPUT_DIR=$1 | |
| CHECKPOINT=$2 | |
| PY_ARGS=${@:3} # Any arguments from the forth one are captured by this | |
| echo "Load model weights from: ${CHECKPOINT}" | |
| # test using the model trained on ref-youtube-vos directly | |
| python3 inference_davis.py --with_box_refine --binary --freeze_text_encoder \ | |
| --output_dir=${OUTPUT_DIR} --resume=${CHECKPOINT} ${PY_ARGS} | |
| # evaluation | |
| ANNO0_DIR=${OUTPUT_DIR}/"valid"/"anno_0" | |
| ANNO1_DIR=${OUTPUT_DIR}/"valid"/"anno_1" | |
| ANNO2_DIR=${OUTPUT_DIR}/"valid"/"anno_2" | |
| ANNO3_DIR=${OUTPUT_DIR}/"valid"/"anno_3" | |
| python3 eval_davis.py --results_path=${ANNO0_DIR} | |
| python3 eval_davis.py --results_path=${ANNO1_DIR} | |
| python3 eval_davis.py --results_path=${ANNO2_DIR} | |
| python3 eval_davis.py --results_path=${ANNO3_DIR} | |
| echo "Working path is: ${OUTPUT_DIR}" | |