OpenGVLab/LAMM

Issues

• petrel_client

  #75 opened 2 months ago by KimWu1994
  2

• Ch3Ef dataset is not avaliable

  #76 opened 8 months ago by Kizna1ver
  2

• How to run the model in float32?

  #82 opened 5 months ago by xbq1994
  0

• How to test on multi-gpu?

  #81 opened 5 months ago by xbq1994
  0

• AttributeError: 'DeepSpeedCPUAdam' object has no attribute 'ds_opt_adam'

  #47 opened a year ago by Xiaolong-RRL
  5

• some question about training octavius

  #79 opened 7 months ago by herkerser
  1

• How much time is required to train Octavius on 3D instruction dataset, as specified in the paper?

  #78 opened 6 months ago by TangYuan96
  0

• How to evaluate 3D tasks of Octavius？

  #80 opened 6 months ago by TangYuan96
  0

• no multimodal branch in lightLLM

  #77 opened 7 months ago by KimWu1994
  0

• COCO Detection Instructions on HF Datasets

  #74 opened 8 months ago by adymaharana
  0

• Could you provide LLaVA-v1.5 7b/13b results on ChEF leaderboard?

  #53 opened 9 months ago by YuqiHUO
  2

• Using PPL on LAMM, discoverd 2 bugs, and result is much worse than direct inference type

  #66 opened 9 months ago by AlexWang1900
  2

• how to deal with multi-turn dialogue for octivius?

  #72 opened 10 months ago by joez17
  0

• Code Stop

  #68 opened a year ago by LanShanPi
  0

• Code stop

  #67 opened a year ago by LanShanPi
  0

• Some benchmarks are missing from the leaderboards

  #64 opened a year ago by zhimin-z
  8

• I am interested in testing Octavius? Is there a tutorial to use it like LAMM with cli_demo.py?

  #65 opened a year ago by drahmad89
  1

• Zero-shot performances on ScanQA with Vicuna-7B

  #44 opened a year ago by UnderTheMangoTree
  2

• How to change options to decrease calculation consumption during evaluation_3d ?

  #48 opened a year ago by Xiaolong-RRL
  2

• Is ScicenQA's meta filename mismatched？

  #63 opened a year ago by sunwhw
  1

• What are the metrics of the six recipes of Desiderata?

  #55 opened a year ago by zhimin-z
  4

• Are the leaderboard results from `0-shot` settings?

  #61 opened a year ago by zhimin-z
  1

• Mismatched results between paper and leaderboard

  #62 opened a year ago by zhimin-z
  2

• omnibenchmark dataset

  #59 opened a year ago by sunwhw
  3

• :question: Could you specify all multitasks Octavius was trained on?

  #60 opened a year ago by simon-lund
  0

• :bug: Broken link in benchmark documentation

  #58 opened a year ago by simon-lund
  1

• What are the metrics of `Omnibenchmark`, `ScienceQA`, `MMBench`, `SEED`, and `MME` benchmarks?

  #54 opened a year ago by zhimin-z
  3

• What does `failed` mean in the test?

  #56 opened a year ago by zhimin-z
  3

• performance on evaluating instructblip-vicuna7b and instructblip-flant5xxl

  #52 opened a year ago by sunwhw
  3

• how to ensure the fairness and stability of this test

  #57 opened a year ago by sunwhw
  4

• Zero-Shot results on 3D_Benchmark and official ckpt

  #41 opened a year ago by hmxiong
  5

• what the differences among instruct_98K, instruct_140K, instruct_186K?

  #38 opened a year ago by peiliu0408
  1

• The output is garbled

  #39 opened a year ago by hmxiong
  2

• 复现zero-shot结果

  #30 opened a year ago by Qiu-DL
  5

• EPCL pertained model details

  #16 opened a year ago by ZCMax
  4

• the save_model func in src/model/agent.py may cause some bugs when using deepspeed ZeRO3

  #32 opened a year ago by lighten001
  1

• Where is the 'cli_demo_pcl.py' file?

  #25 opened a year ago by SCLBD
  1

• Evaluation code is not complete.

  #14 opened a year ago by cxxgtxy
  2

• 论文结果

  #26 opened a year ago by Qiu-DL
  1

• Enable batch inference in some datasets

  #27 opened a year ago by lighten001
  1

• How to evaluate on LLaVA and MiniGPT-4？

  #19 opened a year ago by waltonfuture
  1

• Seemly not freeze llama

  #23 opened a year ago by jinyutong23
  5

• Some images are not contained in the dataset (bamboo_images).

  #13 opened 2 years ago by cxxgtxy
  2

• missing requireiment.txt

  #11 opened 2 years ago by cxxgtxy
  2