Pith. sign in

REVIEW 3 major objections 6 minor 4 cited by

LightEMMA: Lightweight End-to-End Multimodal Model for Autonomous Driving

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read With identical LLaMA weights, a modular VLM framework halves trajectory error on nuScenes, and bigger reasoning models do not help.

desk verdict A useful open-source evaluation framework for zero-shot VLM driving, but the headline 47.6% ADE gain over OpenEMMA is an uncontrolled comparison and the causal claim over-reaches. read the letter →

arxiv 2505.00284 v2 pith:Y6E63GR4 submitted 2025-05-01 cs.RO cs.AI

Zhijie Qiao , Haowei Li , Zhong Cao , Henry X. Liu This is my paper · ORCID
classification cs.ROcs.AI
keywords LightEMMAvision-languagemodelautonomousdrivingtrajectorypredictionchain-of-thoughtnuSceneszero-shotopen-sourceframework
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

LightEMMA is presented as an open-source baseline framework that lets any modern vision-language model drive by mapping a front-view image and the ego vehicle's recent speed-curvature history into a chain-of-thought description and then six low-level control commands. The paper's central quantitative claim is that, with the exact same LLaMA-3.2-11B weights used by OpenEMMA, the framework lowers average displacement error from 2.92 m to 1.53 m, a 47.6% reduction, which the authors attribute to framework design and integration rather than model weights. On 150 nuScenes prediction scenarios, GPT-4o reaches the lowest ADE at 1.07 m, followed closely by Claude-3.7-Sonnet, while larger or reasoning-enabled models such as GPT-5, Gemini-2.5-Pro, and Qwen-2.5-72B do not outperform their smaller or basic counterparts. The authors intend LightEMMA as a practical benchmark for fair VLM comparison, rapid model updates, and failure-mode analysis rather than as a production-ready planner.

What carries the argument

The load-bearing mechanism is a three-stage Chain-of-Thought prompt that forces the VLM to (1) describe the scene, (2) propose a high-level driving intent given scene and history, and (3) emit six (speed, curvature) pairs, which are then converted to a trajectory by kinematic integration with 0.5 s steps. Inputs are deliberately minimal: a single raw front-view camera frame and six historical speed-curvature pairs sampled over 3 seconds, with no auxiliary perception models. The framework's modular codebase separates initialization, prediction, logging, error handling, and retrospective analysis, which is what the authors say produces the accuracy gain and makes dynamic model swaps possible.

What would settle it

Take OpenEMMA's LLaMA-3.2-11B deployment and swap in only LightEMMA's exact prompt template and single-frame input, keeping weights, evaluation code, and sampling fixed; if average ADE stays near 2.92 m rather than dropping to about 1.53 m, the framework-design attribution is falsified. Conversely, running LightEMMA's full pipeline with OpenEMMA's multi-frame input and prompt would reveal which component carries the gain.

Watch

Extended reading notes

Core claim

The central claim is that a well-structured modular VLM framework, without any fine-tuning, can sharply improve open-loop trajectory prediction relative to the previous open-source baseline. Using the identical LLaMA-3.2-11B pretrained weights in both systems, LightEMMA reduces average ADE on nuScenes from 2.92 m to 1.53 m (47.6%), and with Qwen-2.5-7B versus OpenEMMA's Qwen-2.0-7B the reduction is 47.7%. Because the weights are held fixed in the LLaMA comparison, the paper concludes the gain comes from framework design and integration strategies. The paper further claims that advanced or reasoning-enabled VLM variants do not beat their basic counterparts and that GPT-4o achieves the best accuracy with ADE 1.07 m, establishing a substantially stronger open baseline.

Load-bearing premise

The claim that the trajectory improvement comes from framework design assumes the comparison to OpenEMMA is fully controlled, with identical prompts, input representation, evaluation code, and sampling, yet no component-by-component ablation is reported.

Editorial extensions

If this is right

  • Any current or future VLM that accepts image and text can be plugged into the same evaluation pipeline, so trajectory accuracy, cost, latency, and format reliability become directly comparable across model families.
  • The LLaMA-3.2-11B result implies that framework choices alone can be worth roughly a factor of two in average displacement error, independent of model generation improvements.
  • Because stronger and costlier models did not outperform basic ones, the results push the field toward task-specific architecture and prompting design rather than simple parameter scaling for driving.
  • Single-frame, no-perception-module input keeps the pipeline cheap and fast, supporting use as a continuously updated benchmark where new models can be validated with minimal integration effort.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Attribution of the entire ADE gain to modular framework design is not directly established: the paper does not ablate prompt template, single-frame input, output format, or evaluation filtering, so part of the gain may come from those choices; a controlled ablation would settle it.
  • The single-frame choice implies the framework ignores explicit temporal dynamics beyond the six-step history, so cases like the right-turn bias in Case 1 might be improved by richer history or explicit map cues, an extension the paper does not explore.
  • A testable prediction follows from the paper's own failure analysis: adding explicit navigation intent or map-level turn information should remove the systematic left-turn failures described in Case 2, which could be checked by appending a one-line navigation hint to the same prompt.
  • Open-loop nuScenes ADE may not predict closed-loop safety; the paper's qualitative cases suggest VLM stochasticity can cause abrupt braking or running red lights, so a closed-loop simulator evaluation would be a natural next step.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. LightEMMA is presented as a modular, open-source, zero-shot framework that feeds a single front-view camera image plus a 3-second history of (speed, curvature) pairs to a vision-language model, uses a chain-of-thought prompt to elicit six low-level driving commands, and integrates those commands kinematically to produce a 3-second trajectory. The manuscript evaluates twelve commercial and open-source VLMs on the nuScenes prediction task, reporting inference time, token usage and cost, format-error rates, average/final displacement error, and qualitative failure modes. Its headline quantitative claims are a 47.6% reduction in average ADE for LLaMA-3.2-11B relative to OpenEMMA using the same pretrained weights, a similar reduction for Qwen-2.5-7B, and the observation that advanced or reasoning-enabled model variants do not outperform their basic counterparts.

Significance. If the improvement over OpenEMMA were established under controlled conditions, this would be a useful benchmark contribution: the framework is modular, the code is publicly released, all evaluated models run under a single prompting and evaluation protocol, and the trajectory construction uses textbook kinematic equations with no parameters fitted to nuScenes labels. The paper also provides a broader model survey than prior open-source work and documents cost and latency in a way that is often missing from VLM driving papers. However, the central superiority claim is not yet supported because the OpenEMMA comparison is not controlled, and the reported accuracy numbers are only means without variance or significance information. The benchmark and qualitative insights are plausible and valuable, but the causal attribution of the ADE improvement to framework design requires additional evidence.

major comments (3)
  1. [Section IV.D, Table II] The claim that the 47.6% average-ADE reduction over OpenEMMA 'arises from framework design and integration strategies' is not supported by the evidence presented. The OpenEMMA rows appear to be taken from Ref. [13] rather than re-run in the same evaluation harness, and the LightEMMA pipeline differs in at least three documented respects: Section III.B uses only the current frame instead of concatenating multiple past frames, Section III.C encodes driving history as (speed, curvature) pairs rather than cartesian coordinates, and Section III.D applies 'minor adjustments' to the chain-of-thought prompt. Any one of these differences, or the error-parsing and frame-filtering procedure in Section IV.C, could plausibly explain the improvement. Please add ablations that vary one component at a time, or at minimum re-run OpenEMMA inside LightEMMA's evaluation protocol and report both configurations; otherwise the attribution to framework design should be softened to a descriptive comparison rather than a causal claim.
  2. [Section IV.D, Table II] Only mean ADE and FDE values are reported, with no per-scenario distributions, error bars, or statistical tests. The test set is 150 scenarios, VLM outputs are stochastic, and some model pairs are very close (e.g., GPT-4o at 1.07 m versus Claude-3.7-Sonnet at 1.08 m average ADE). To support the model ranking and the claimed 47.6% reduction, please report per-scenario paired errors and a statistical comparison such as a paired test or bootstrap confidence intervals. This is particularly important because the frame-filtering step leaves 3,893 of 3,908 frames, and the effect of filtering on each model's mean is not shown.
  3. [Section IV.C, Table II] The handling of Qwen-2.5-72B is inconsistent with the stated frame-filtering rule. The manuscript says frames are excluded only when any model fails to generate a valid prediction based on the corrected results, but Qwen-2.5-72B is then 'excluded entirely' because of its 62.9% format-error rate (47.5% after correction). This is a post-hoc model-level exclusion rather than the described frame-level filtering, and it changes the evaluated roster from twelve models to eleven. Please include the valid corrected trajectories of the 72B model if possible, justify the exclusion with a principled criterion, or revise the text so that the twelve-model claim is not contradicted by the accuracy table.
minor comments (6)
  1. [Section III.B] The cross-reference to 'Section II-F' appears to point to the example workflow, which actually appears in Section III.F; please correct the reference.
  2. [Section III.D] The heading 'VLM Promoting' should read 'VLM Prompting'.
  3. [Table I] The dagger symbols next to the input-token counts for Gemini and LLaMA models are not explained in the table or caption; add a footnote noting that those providers appear to omit image tokens from the reported counts.
  4. [Table II] Add a table note stating that the two OpenEMMA rows are reproduced from Ref. [13] and were not re-run in the present harness, since the '-' entries in the FE and FE Corr columns otherwise hide this asymmetry.
  5. [Section V] Statements such as 'all models consistently fail' in Cases 1 and 2 are made for selected illustrative frames; please qualify them by stating the number of cases observed or labeling them explicitly as qualitative observations rather than quantitative findings.
  6. [Section IV.C] The sentence 'Since all models were evaluated using identical prompts and workflows' should be clarified to mean identical within LightEMMA, because Section III.D notes that the CoT prompt includes adjustments relative to prior studies and the OpenEMMA comparisons are not re-run under the same workflow.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LightEMMA's trajectory predictions are computed from VLM outputs via textbook kinematics, and none of the central claims reduce to their own inputs.

full rationale

The paper's derivation chain is not circular. The predicted trajectory is generated by kinematic integration (Eqs. 1-3) of the VLM-produced speed and curvature commands; no parameter is fitted to the nuScenes ground-truth labels, and no output quantity is fed back into the input representation. The claim that LightEMMA improves over OpenEMMA is an empirical comparison on an external benchmark, not a definitional equivalence. The paper does cite prior work of its own authors (e.g., [10]), but that citation is not load-bearing for the framework's core derivation or for the ADE comparison. The skeptical concern that the OpenEMMA comparison is uncontrolled (different input frames, history representation, prompt adjustments, and no re-run) is a legitimate experimental-validity critique, but it is not a circularity defect under the stated taxonomy: no fitted parameter is renamed as a prediction, no claim is defined in terms of its own output, and no load-bearing argument reduces to a self-citation. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The framework introduces no new physical quantities. Its predictive content rests on the kinematic integration model, the zero-shot interpretability of VLM outputs, and an unreported preliminary experiment that motivated dropping multi-frame inputs. These are domain assumptions rather than fitted parameters.

assumptions (4)
  • ad hoc to paper A single front-view image plus 3 seconds of (v, kappa) history is sufficient input for 3-second trajectory prediction.
    Section III.B states that additional frames were tested in 'preliminary experiments' but no data is shown; this supports the framework's single-image design.
  • domain assumption Constant speed and curvature over each 0.5 s interval, per equations (1)-(3), adequately model vehicle kinematics for scoring.
    This ignores acceleration and jerk, which can be significant in stop-and-go scenarios; it is a standard but simplifying assumption.
  • domain assumption Open-loop ADE/FDE against one ground-truth trajectory is a valid measure of autonomous driving capability.
    The paper uses official nuScenes metrics but does not account for multimodal futures or closed-loop safety, as noted in their qualitative examples.
  • domain assumption Zero-shot VLM responses can be parsed as executable driving commands without fine-tuning.
    The whole framework relies on this; high format error rates for some models (Qwen-2.5-72B at 62.9%) show this assumption is brittle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LightEMMA: Lightweight End-to-End Multimodal Model for Autonomous Driving." pith.science (2026). https://pith.science/paper/Y6E63GR4

@misc{pith2026250500284,
  author       = {Pith},
  title        = {Pith review of: LightEMMA: Lightweight End-to-End Multimodal Model for Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y6E63GR4}},
  note         = {Machine review of arXiv:2505.00284}
}
read the original abstract

Vision-Language Models (VLMs) have demonstrated significant potential for end-to-end autonomous driving. However, the field still lacks a practical platform that enables dynamic model updates, rapid validation, fair comparison, and intuitive performance assessment. To that end, we introduce LightEMMA, a Lightweight End-to-End Multimodal Model for Autonomous driving. LightEMMA provides a unified, VLM-based autonomous driving framework without ad hoc customizations, enabling easy integration with evolving state-of-the-art commercial and open-source models. We construct twelve autonomous driving agents using various VLMs and evaluate their performance on the challenging nuScenes prediction task, comprehensively assessing computational metrics and providing critical insights. Illustrative examples show that, although VLMs exhibit strong scenario interpretation capabilities, their practical performance in autonomous driving tasks remains a concern. Additionally, increased model complexity and extended reasoning do not necessarily lead to better performance, emphasizing the need for further improvements and task-specific designs. The code is available at https://github.com/michigan-traffic-lab/LightEMMA.

Figures

Figures reproduced from arXiv: 2505.00284 by the authors.

Figure 1
Figure 1. LightEMMA architecture. The main contributions are summarized as follows: 1) We present LightEMMA, an open-source baseline framework for end-to-end autonomous driving tasks, designed to integrate seamlessly with state-of-the-art VLMs while accommodating continual advancements in model development. Our framework facilitates rapid prototyping and streamlines transferability through its modular and structured codebase.… view at source ↗
Figure 2
Figure 2. LightEMMA nuScenes prediction task examples. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. VLN-AVP: Zero-Shot Vision-Language Navigation with Hybrid Long-Short-Term Memory for Autonomous Valet Parking

    cs.RO 2026-07 conditional novelty 6.0 of 10

    A map-free, language-guided parking navigation system with short- and long-term memory beats adapted VLN/AD baselines on a new underground parking benchmark and in real vehicle trials.

  2. Wavelet Phase Diffusion for Structurally and Semantically Consistent Sim-to-Real Translation

    cs.AI 2026-07 conditional novelty 6.0 of 10

    Wavelet-domain phase injection with low-frequency randomization improves realism and semantic consistency of sim-to-real translation, improving VLM planner ADE and FDE by about 5% on CARLA videos.

  3. Impromptu VLA: Open Weights and Open Data for Driving Vision-Language-Action Models

    cs.CV 2025-05 conditional novelty 6.0 of 10

    A new 80K-clip dataset of unstructured driving scenarios with Q&A annotations improves VLA performance on NeuroNCAP and nuScenes benchmarks.

  4. Chain-of-Thought for Autonomous Driving: A Comprehensive Survey and Future Prospects

    cs.RO 2025-05 conditional novelty 4.0 of 10

    A survey that classifies chain-of-thought methods for autonomous driving into modular, logical, and reflective pipelines, and proposes three evolutionary stages from direct prompting to reinforcement learning.

Reference graph

Works this paper leans on

33 extracted references · 22 canonical work pages · cited by 4 Pith papers

  1. [13]

    Openemma: Open-source multimodal model for end-to-end autonomous driving,

    S. Xing, C. Qian, Y . Wang, H. Hua, K. Tian, Y . Zhou, and Z. Tu, “Openemma: Open-source multimodal model for end-to-end autonomous driving,” 2025

    work page 2025

  2. [1]

    Buehler, K

    M. Buehler, K. Iagnemma, and S. Singh,The DARPA Urban Chal- lenge: Autonomous Vehicles in City Traffic, vol. 56. Springer Science & Business Media, 2009

    work page 2009

  3. [2]

    Intelligent and connected vehicles: Current status and future perspectives,

    D. Yang, K. Jiang, D. Zhao, C. Yu, Z. Cao, S. Xie, Z. Xiao, X. Jiao, S. Wang, and K. Zhang, “Intelligent and connected vehicles: Current status and future perspectives,”Science China Technological Sciences, vol. 61, no. 10, pp. 1446–1471, 2018

    work page 2018

  4. [3]

    Planning-oriented autonomous driving,

    Y . Hu, J. Yang, L. Chen, K. Li, C. Sima, X. Zhu, S. Chai, S. Du, T. Lin, W. Wang, L. Lu, X. Jia, Q. Liu, J. Dai, Y . Qiao, and H. Li, “Planning-oriented autonomous driving,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

    2023

  5. [4]

    Perceive, attend, and drive: Learning spatial attention for safe self- driving,

    B. Wei, M. Ren, W. Zeng, M. Liang, B. Yang, and R. Urtasun, “Perceive, attend, and drive: Learning spatial attention for safe self- driving,” in2021 IEEE International Conference on Robotics and Automation (ICRA), pp. 4875–4881, 2021

    work page 2021

  6. [5]

    End-to-end interpretable neural motion planner,

    W. Zeng, W. Luo, S. Suo, A. Sadat, B. Yang, S. Casas, and R. Urtasun, “End-to-end interpretable neural motion planner,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion (CVPR), June 2019

    work page 2019

  7. [6]

    End-to-end model-free reinforcement learning for urban driving using implicit affordances,

    M. Toromanoff, E. Wirbel, and F. Moutarde, “End-to-end model-free reinforcement learning for urban driving using implicit affordances,” in2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7151–7160, 2020

    work page 2020

  8. [7]

    St-p3: End- to-end vision-based autonomous driving via spatial-temporal feature learning,

    S. Hu, L. Chen, P. Wu, H. Li, J. Yan, and D. Tao, “St-p3: End- to-end vision-based autonomous driving via spatial-temporal feature learning,” inEuropean Conference on Computer Vision (ECCV), 2022

    work page 2022

Show all 33 references
  1. [8]

    Review and challenge: High defi- nition map technology for intelligent connected vehicle,

    M. Yang, K. Jiang, B. Wijaya, T. Wen, J. Miao, J. Huang, C. Zhong, W. Zhang, H. Chen, and D. Yang, “Review and challenge: High defi- nition map technology for intelligent connected vehicle,”Fundamental Research, 2024

    2024

  2. [9]

    End-to-end autonomous driving: Challenges and frontiers,

    L. Chen, P. Wu, K. Chitta, B. Jaeger, A. Geiger, and H. Li, “End-to-end autonomous driving: Challenges and frontiers,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  3. [10]

    Curse of rarity for autonomous vehicles,

    H. X. Liu and S. Feng, “Curse of rarity for autonomous vehicles,” Nature Communications, vol. 15, p. 4808, 2024

    2024

  4. [11]

    Emma: End-to-end multimodal model for autonomous driving,

    J.-J. Hwang, R. Xu, H. Lin, W.-C. Hung, J. Ji, K. Choi, D. Huang, T. He, P. Covington, B. Sapp, Y . Zhou, J. Guo, D. Anguelov, and M. Tan, “Emma: End-to-end multimodal model for autonomous driving,” 2024

    2024

  5. [12]

    Gemini: A family of highly capable multimodal models,

    Google, “Gemini: A family of highly capable multimodal models,” 2024

    2024

  6. [14]

    Drivegpt4: Interpretable end-to-end autonomous driving via large language model,

    Z. Xu, Y . Zhang, E. Xie, Z. Zhao, Y . Guo, K.-Y . K. Wong, Z. Li, and H. Zhao, “Drivegpt4: Interpretable end-to-end autonomous driving via large language model,”IEEE Robotics and Automation Letters, vol. 9, no. 10, pp. 8186–8193, 2024

    2024

  7. [15]

    Dolphins: Multimodal language model for driving,

    Y . Ma, Y . Cao, J. Sun, M. Pavone, and C. Xiao, “Dolphins: Multimodal language model for driving,” inComputer Vision – ECCV 2024: 18th European Conference, Milan, Italy, September 29–October 4, 2024, Proceedings, Part XLV, (Berlin, Heidelberg), p. 403–420, Springer- Verlag, 2024

    2024

  8. [16]

    Drivemlm: Aligning multi-modal large lan- guage models with behavioral planning states for autonomous driving,

    W. Wang, J. Xie, C. Hu, H. Zou, J. Fan, W. Tong, Y . Wen, S. Wu, H. Deng, Z. Li,et al., “Drivemlm: Aligning multi-modal large lan- guage models with behavioral planning states for autonomous driving,” arXiv preprint arXiv:2312.09245, 2023

    2023

  9. [17]

    Carla: An open urban driving simulator,

    A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V . Koltun, “Carla: An open urban driving simulator,” 2017

    2017

  10. [18]

    Large scale interactive motion forecasting for autonomous driving: The waymo open motion dataset,

    S. Ettinger, S. Cheng, B. Caine, C. Liu, H. Zhao, S. Pradhan, Y . Chai, B. Sapp, C. R. Qi, Y . Zhou, Z. Yang, A. Chouard, P. Sun, J. Ngiam, V . Vasudevan, A. McCauley, J. Shlens, and D. Anguelov, “Large scale interactive motion forecasting for autonomous driving: The waymo ope...

    2021

  11. [19]

    nuScenes: A Multimodal Dataset for Autonomous Driving ,

    H. Caesar, V . Bankiti, A. H. Lang, S. V ora, V . E. Liong, Q. Xu, A. Krishnan, Y . Pan, G. Baldan, and O. Beijbom, “ nuScenes: A Multimodal Dataset for Autonomous Driving ,” in2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (Los Alamitos, CA, USA),...

    2020

  12. [20]

    Language prompt for autonomous driving,

    D. Wu, W. Han, T. Wang, Y . Liu, X. Zhang, and J. Shen, “Language prompt for autonomous driving,”arXiv preprint, 2023

    2023

  13. [21]

    Lingoqa: Visual question answering for autonomous driving,

    A.-M. Marcu, L. Chen, J. H ¨unermann, A. Karnsund, B. Hanotte, P. Chidananda, S. Nair, V . Badrinarayanan, A. Kendall, J. Shotton, and O. Sinavski, “Lingoqa: Visual question answering for autonomous driving,”arXiv preprint arXiv:2312.14115, 2023

    2023 arXiv

  14. [22]

    Reason2drive: Towards interpretable and chain-based reasoning for autonomous driving,

    M. Nie, R. Peng, C. Wang, X. Cai, J. Han, H. Xu, and L. Zhang, “Reason2drive: Towards interpretable and chain-based reasoning for autonomous driving,” inComputer Vision – ECCV 2024: 18th Eu- ropean Conference, Milan, Italy, September 29–October 4, 2024, Proceedings, Part XXVI,...

    2024

  15. [23]

    Gpt-4 technical report,

    OpenAI, “Gpt-4 technical report,” 2024

    2024

  16. [24]

    The claude 3 model family: Opus, sonnet, haiku,

    Anthropic, “The claude 3 model family: Opus, sonnet, haiku,” 2024

    2024

  17. [25]

    Deepseek-vl2: Mixture-of-experts vision-language models for advanced multimodal understanding,

    DeepSeek-AI, “Deepseek-vl2: Mixture-of-experts vision-language models for advanced multimodal understanding,” 2024

    2024

  18. [26]

    Llama: Open and efficient foundation language models,

    Meta-AI, “Llama: Open and efficient foundation language models,” 2023

    2023

  19. [27]

    Qwen2 technical report,

    Alibaba, “Qwen2 technical report,” 2024

    2024

  20. [28]

    Ultralytics yolov8

    G. Jocher, A. Chaurasia,et al., “Ultralytics yolov8.”https:// github.com/ultralytics/ultralytics, 2023. Accessed: 2025-08-15

    2023

  21. [29]

    Videobert: A joint model for video and language representation learning,

    C. Sun, A. Myers, C. V ondrick, K. Murphy, and C. Schmid, “Videobert: A joint model for video and language representation learning,” inProceedings of the IEEE/CVF international conference on computer vision, pp. 7464–7473, 2019

    2019

  22. [30]

    VideoMAE: Masked autoencoders are data-efficient learners for self-supervised video pre- training,

    Z. Tong, Y . Song, J. Wang, and L. Wang, “VideoMAE: Masked autoencoders are data-efficient learners for self-supervised video pre- training,” inAdvances in Neural Information Processing Systems, 2022

    2022

  23. [31]

    DriveVLM: The convergence of autonomous driving and large vision-language models,

    X. Tian, J. Gu, B. Li, Y . Liu, Y . Wang, Z. Zhao, K. Zhan, P. Jia, X. Lang, and H. Zhao, “DriveVLM: The convergence of autonomous driving and large vision-language models,” in8th Annual Conference on Robot Learning, 2024

    2024

  24. [32]

    Distilling task-specific knowledge from bert into simple neural networks,

    R. Tang, Y . Lu, L. Liu, L. Mou, O. Vechtomova, and J. Lin, “Distilling task-specific knowledge from bert into simple neural networks,” 2019

    2019

  25. [33]

    Rethinking the open-loop evaluation of end-to-end autonomous driving in nuscenes,

    J.-T. Zhai, Z. Feng, J. Du, Y . Mao, J.-J. Liu, Z. Tan, Y . Zhang, X. Ye, and J. Wang, “Rethinking the open-loop evaluation of end-to-end autonomous driving in nuscenes,”arXiv preprint arXiv:2305.10430, 2023

    2023 arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.