Hi Qwen-VLA team,
Congrats on the release!
We're the team behind MolmoSpaces, an open ecosystem and benchmark (https://molmospaces.allen.ai/leaderboard) which evaluates generalist policies under systematic, controlled scene/object/instruction variation. The benchmark has a strong sim-to-real correlation, so leaderboard numbers actually mean something on real robots. We'd love to see Qwen-VLA on the leaderboard.
Submission instructions: allenai/molmospaces#8
The benchmark uses a Franka FR3 + Robotiq 2F-85, and we're happy to help if anything is unclear — feel free to ping me (@omarrayyann) or @BlGene on the submission issue.
Thanks!
Hi Qwen-VLA team,
Congrats on the release!
We're the team behind MolmoSpaces, an open ecosystem and benchmark (https://molmospaces.allen.ai/leaderboard) which evaluates generalist policies under systematic, controlled scene/object/instruction variation. The benchmark has a strong sim-to-real correlation, so leaderboard numbers actually mean something on real robots. We'd love to see Qwen-VLA on the leaderboard.
Submission instructions: allenai/molmospaces#8
The benchmark uses a Franka FR3 + Robotiq 2F-85, and we're happy to help if anything is unclear — feel free to ping me (@omarrayyann) or @BlGene on the submission issue.
Thanks!