VLABench

简介

VLABench external benchmark manifest for language-conditioned manipulation policy rollouts.

指标

VLABench external benchmark manifest for language-conditioned manipulation policy rollouts.

Task YAML

worldfoundry/data/benchmarks/tasks/external/vlabench.yaml

Primary

• success_rate

环境准备

运行环境

• Conda 环境： vlabench-runtime。
• 单独配置： 需要 benchmark 专用 conda 环境，不要与统一 runtime 混用。
• WorldFoundry provides contract and official result normalization for caller-provided VLABench outputs.

测评数据

• 待测模型输出目录：设置 WORLDFOUNDRY_VLABENCH_RESULTS_PATH 指向生成视频或 rollout artifact 根目录。
• ${WORLDFOUNDRY_DATA_DIR}/datasets/vlabench
• ${WORLDFOUNDRY_CACHE_DIR}/repos/OpenMOSS--VLABench

Checkpoint 与资产

• 该 benchmark 以官方结果归一化或仿真 rollout 为主；policy / simulator checkpoint 见 catalog requires 与 task YAML。

关键环境变量

• WORLDFOUNDRY_VLABENCH_ROOT
• WORLDFOUNDRY_VLABENCH_RESULTS_PATH

运行评测

准备好该 benchmark 页面列出的资产和候选产物后，使用下面的公开运行入口。

设置候选产物目录

export WORLDFOUNDRY_GENERATED_ARTIFACT_DIR=/path/to/generated/artifacts

导入已有官方结果

worldfoundry-eval zoo benchmark-run \
  --benchmark-id vlabench \
  --mode normalizer \
  --official-results-path /path/to/official/results-or-report \
  --generated-artifact-dir "${WORLDFOUNDRY_GENERATED_ARTIFACT_DIR}" \
  --output-dir tmp/vlabench/normalizer \
  --json

从生成产物重新打分

worldfoundry-eval zoo benchmark-run \
  --benchmark-id vlabench \
  --mode official-run \
  --generated-artifact-dir "${WORLDFOUNDRY_GENERATED_ARTIFACT_DIR}" \
  --output-dir tmp/vlabench/official-run \
  --json

直接调用仓内 runner

worldfoundry-eval zoo benchmark-run --benchmark-id vlabench --mode official-validation --official-results-path <official_results.json> --output-dir <out> --json

依赖与产物

Inputs

• Official simulator runtime
• benchmark assets
• policy checkpoint
• official result dump

Environment

• WORLDFOUNDRY_VLABENCH_RESULTS_PATH

Outputs

• scorecard.json
• raw_results.jsonl
• evaluation/summary.json

← 返回 Benchmark Hub