DriveLM

DriveLM copied to clipboard

Chatgpt score is significantly lower

1

The chatgpt score in current online evaluations tends to be significantly lower. I have submitted the same file as during the competition, but now the chatgpt score is only 61,...

fcjian

Hi, thanks for the awesome work! Quick questions about the GVQA construction: are the dependencies independent frame-wise or could they span across multiple time-dependent key-frames (say, contiguous frames of a...

lucamodica

Hello! I recently watched the demo video of your project and was really impressed by the visualization effects. The visual representation is both clear and engaging, making it easier to...

ranydb

Dear authors, I am curious about how to calculate METEOR. This metric appears in the paper but not in the code. Is it the same as [match score](https://github.com/OpenDriveLab/DriveLM/blob/1de72a74b257e5373400fa68239e99bd5d20580a/challenge/evaluation.py#L147)? Thanks in...

codename1995

Dear authors, thank you for the great work. In March 2024, you updated the dataset to v1.1. I am wondering what is the updates of v1.1 compared to v1.0. Is...

codename1995

In "d1d38dddb8174acf9cb776ec6c941024" , "planning" part, second item has an empty "Q" value. "d1d38dddb8174acf9cb776ec6c941024": { "QA": { "perception": [ { "Q": "What are the important objects in the current scene? Those...

ZiangWu-77

In the dataset, the steering in behavior questions is divided into "steering" and "slightly steering." What is the difference between the two in terms of annotation? Additionally, the speed includes...

xiangToby

Hello Thank you for good research!! I wonder what means words (C, con_up, con_down, cluster, layer) in JSON file. In v1_1_train_nus.json, words (C, con_up, con_down, cluster, layer) are "null", Is...

youngj98

Didn't receive any reply or download link its been days after filling the form did tried filling form diffrent mails also.

shreyash212002