hustvl/Senna

Issues

• Why does the number of tokens have such a significant impact?

  #30 opened 2 months ago by baodingge
  0

• Accuracy Drop on the Validation data of Nuscene Mini

  #26 opened 2 months ago by abhigoku10
  1

• Request for Sharing QA Data Files

  #24 opened 3 months ago by Cloud-li1
  14

• about nuscenes data convert

  #15 opened 3 months ago by AMzhanghan
  4

• Error occurred while running the command `pip install -r requirements.txt`.

  #27 opened 2 months ago by qlj215
  3

• unknown vision tower when running the eval script

  #29 opened 2 months ago by cutefulf
  0

• 关于visual encoder

  #28 opened 2 months ago by Whale-ice
  0

• Can upload the parameters to modelscope, thx

  #12 opened 3 months ago by zizaisuiyuan
  2

• Question about E2E part

  #25 opened 3 months ago by Cloud-li1
  0

• Number of waypoints

  #21 opened 3 months ago by missTL
  2

• execute sh senna_nusc_converter.sh error

  #17 opened 3 months ago by StrongTsai
  5

• How to convert our model to the ONNX

  #22 opened 3 months ago by chi0612
  2

• What is the DriveX dataset?

  #2 opened 3 months ago by b5strbal
  2

• Some weights of LlavaLlamaForCausalLM were not initialized from the model checkpoint

  #23 opened 3 months ago by MIKE-GUO233
  1

• Using vicuna-7b-v1.5 model.

  #20 opened 3 months ago by comflife
  1

• Raw data获取过程询问

  #19 opened 3 months ago by fjq-tongji
  1

• senna_nusc_data_converter.py error

  #18 opened 3 months ago by wolf943134497
  4

• Is the currently code primarily targeted at front-view image？

  #16 opened 3 months ago by sunbin1357
  1

• [E2E module to LLM]

  #10 opened 3 months ago by Lewis-Lu
  5

• Requires for details of python environment

  #9 opened 3 months ago by xiaodongww
  2

• Analyzing model performance on nuScenes validation Set

  #11 opened 3 months ago by sunbin1357
  3

• eval_plan_qa.json

  #13 opened 3 months ago by PG-Wang
  4

• Considerations for Choosing Vicuna-v1.5-7b as the LLM Model

  #8 opened 3 months ago by xiaodongww
  1

• If we want to achieve full scene understanding, what is the approximate cost of the vlm annotation solution

  #7 opened 3 months ago by hunkyu
  1

• Model Deployment

  #6 opened 3 months ago by chensiweiTHU
  1

• Code Release

  #5 opened 3 months ago by abhigoku10
  2

• question about meta-action encoder

  #4 opened 3 months ago by TmacTmac1992
  2

• Is it possible to achieve a large-scale indoor navigation function by providing a model with a video of an indoor space that can be analyzed based on the indoor layout, and then inputting a picture and the destination I want to reach, so that it can recognize my location and perform path planning and navigation? How many VRAMs are needed and how long will they be open sourced? I'm looking forward to it

  #1 opened 3 months ago by libai-lab
  1

• stage 2 traning error

  #14 opened 3 months ago by PG-Wang
  1

• Required time for training

  #3 opened 5 months ago by kemaloksuz
  4