[BUG] JSONDecodeError when running cuVS Bench

Describe the bug
When running cuVS Bench in Docker, I encountered the following errors:

Steps/Code to reproduce bug
Follow this guide https://docs.rapids.ai/api/cuvs/nightly/cuvs_bench/ to install cuVS Bench with Docker, then execute the following commands to run:

export DATA_FOLDER=./datasets
docker run --gpus all --rm -it -u $(id -u)                      \

    -v $DATA_FOLDER:/data/benchmarks                            \

    rapidsai/cuvs-bench:24.12a-cuda12.5-py3.10              \

    "--dataset gist-960-euclidean"                           \

    "--normalize"                                               \

    "--algorithms cuvs_cagra,cuvs_ivf_pq --batch-size 1 -k 10" \

    ""		

Expected behavior
The benchmark results should be exported without errors.

Environment details (please complete the following information):

• Environment location: Docker
• Method of RAFT install: Docker
  • If method of install is [Docker], provide docker pull & docker run commands used:
  • docker pull rapidsai/cuvs-bench:24.12a-cuda12.5-py3.10

Additional context
The ANN search process in the benchmark seems to be running normally, but the result export process failed. I tried to switch to the conda environment, install cuVS in conda, and run the benchmark, but the error persisted.

Follow the cuVS Bench Docker guide and reproduce the JSONDecodeError with the provided docker pull and docker run commands. Start at the benchmark result-export process, since ANN search appears to complete normally. Done means the benchmark results export successfully without the reported error.