{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/towards-generalizable-vision-language-robotic","title":"Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policy","arxiv_id":"2410.01345","date":"2024-10-02","proceeding":null,"authors":["Ricardo Garcia","ShiZhe Chen","Cordelia Schmid"],"abstract":"Generalizing language-conditioned robotic policies to new tasks remains a significant challenge, hampered by the lack of suitable simulation benchmarks. In this paper, we address this gap by introducing GemBench, a novel benchmark to assess generalization capabilities of vision-language robotic manipulation policies. GemBench incorporates seven general action primitives and four levels of generalization, spanning novel placements, rigid and articulated objects, and complex long-horizon tasks. We evaluate state-of-the-art approaches on GemBench and also introduce a new method. Our approach 3D-LOTUS leverages rich 3D information for action prediction conditioned on language. While 3D-LOTUS excels in both efficiency and performance on seen tasks, it struggles with novel tasks. To address this, we present 3D-LOTUS++, a framework that integrates 3D-LOTUS's motion planning capabilities with the task planning capabilities of LLMs and the object grounding accuracy of VLMs. 3D-LOTUS++ achieves state-of-the-art performance on novel tasks of GemBench, setting a new standard for generalization in robotic manipulation. The benchmark, codes and trained models are available at https://www.di.ens.fr/willow/research/gembench/.","url_abs":"https://arxiv.org/abs/2410.01345v2","url_pdf":"https://arxiv.org/pdf/2410.01345v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"towards-generalizable-vision-language-robotic","repo_url":"https://github.com/vlc-robot/robot-3dlotus","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"motion-planning","task_name":"Motion Planning"},{"task_slug":"robot-manipulation","task_name":"Robot Manipulation"},{"task_slug":"robot-manipulation-generalization","task_name":"Robot Manipulation Generalization"},{"task_slug":"task-planning","task_name":"Task Planning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/robot-manipulation-on-rlbench","task":"Robot Manipulation","dataset":"RLBench","model":"3D-LOTUS","rank_in_archive_order":4,"of":18,"metrics":{"Inference Speed (fps)":"9.5","Input Image Size":"256","Succ. Rate (18 tasks, 100 demo/task)":"83.1","Training Time (A100 x hour)":"40","Training Time (V100 x 8 x day)":"0.28"},"uses_additional_data":false},{"leaderboard":"/sota/robot-manipulation-generalization-on-gembench","task":"Robot Manipulation Generalization","dataset":"GEMBench","model":"3D-LOTUS++","rank_in_archive_order":1,"of":6,"metrics":{"Average Success Rate":"48.0","Average Success Rate (L1)":"68.7±0.6","Average Success Rate (L2)":"64.5±0.9","Average Success Rate (L3)":"41.5±1.8","Average Success Rate (L4)":"17.4±0.4"},"uses_additional_data":false},{"leaderboard":"/sota/robot-manipulation-generalization-on-gembench","task":"Robot Manipulation Generalization","dataset":"GEMBench","model":"3D-LOTUS","rank_in_archive_order":2,"of":6,"metrics":{"Average Success Rate":"45.7","Average Success Rate (L1)":"94.3±1.4","Average Success Rate (L2)":"49.9±2.2","Average Success Rate (L3)":"38.1±1.1","Average Success Rate (L4)":"0.3±0.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2410.01345","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2410.01345"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/vlc-robot/robot-3dlotus","reach":{"status":"ok"}}],"summary":{"ran":1},"by_repo_kind":{"official":{"samples":1,"ran":1,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":1,"samples":[{"code_sha256_prefix":"284e52af67401746","entry":"get_prompts_from_label","repo":"vlc-robot/robot-3dlotus","repo_kind":"official","path":"genrobo3d/vlm_models/clip_encoder.py","file_url":"https://github.com/vlc-robot/robot-3dlotus/blob/HEAD/genrobo3d/vlm_models/clip_encoder.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"284e52af67401746"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}