dvlab-research/LISA

Issues

• Is LISA-7B-v1 finetuned on ReasonSeg?

  #175 opened a month ago by DanielSHKao
  0

• About training environment！！！

  #151 opened 6 months ago by jifeng35
  3

• How to Resume

  #165 opened 4 months ago by HIT-liar
  1

• Can i use 4 24 G GPU for this project fine tuning

  #147 opened 6 months ago by aliman80
  2

• There is no <seg> token in the output after finetuning

  #146 opened 7 months ago by 1476732094
  1

• Inference produces blank segmentation masks

  #174 opened 2 months ago by vanshikavats9
  1

• How to train my datesets？

  #140 opened 8 months ago by Amark-cheey
  1

• can't load LISA++

  #173 opened 3 months ago by victoriazinkovich
  5

• Can't reproduce the results of lisa-7b

  #135 opened 9 months ago by L1uShuai
  11

• I'd like to ask about methods to force the model to generate the <SEG> token.

  #170 opened 2 months ago by minjung98
  0

• Generated several bin files instead of only one bin file using zero_to_fp32.py

  #155 opened 5 months ago by skx6
  3

• Mask image seems to be nothing at all

  #160 opened 5 months ago by JesseZZZZZ
  2

• Why cat 255 tokens to seg_token_mask?

  #172 opened 2 months ago by zhang-yige
  2

• > > LLaVA-1.5 uses 336px image resolution, so you should change the clip model and control max context length. Also, the image token length is set to 256 by default, but when the resolution is changed to 336, the image token length should be set to 576. Overall, some implementation details need further consideration to adapt to llava-1.5. You should check that in detail.

  #163 opened 5 months ago by Amark-cheey
  0

• 关于结合LLM与SAM进行特定物体分割的咨询

  #171 opened 4 months ago by Linn0910
  0

• Some questions about finetuning and LISA++

  #169 opened 4 months ago by chris5zk
  0

• Model with Mask2Former

  #168 opened 4 months ago by stdrr
  0

• step vs epoch

  #141 opened 4 months ago by dohyun1411
  1

• can it accept video input？

  #167 opened 4 months ago by ljh-7
  0

• Paper provided example can not be reproduced !!

  #153 opened 6 months ago by HuangChiEn
  9

• Stuck at Loading checkpoint shards 100%

  #166 opened 4 months ago by Xamexer
  0

• How to Resume

  #164 opened 4 months ago by HIT-liar
  0

• Training on Pretrained LISA Model

  #159 opened 5 months ago by Mactor2018
  1

• Result Reproduction Problem on Refcoco Dataset

  #162 opened 5 months ago by fazhdo
  0

• 请问会考虑开源lisa++的权重和数据集吗

  #161 opened 5 months ago by leftdoor
  0

• About Dataset reason_seg/explanatory

  #158 opened 5 months ago by asad14053
  0

• Evaluation on ReasonSeg test set (missing annotation, text for refer)

  #157 opened 5 months ago by suikei-wang
  0

• Request for the latest code of LLaVA 1.5

  #156 opened 5 months ago by Bilibilee
  0

• RuntimeError: CUDA error: unknown error

  #154 opened 5 months ago by Shaosifan
  0

• How many samples were iterated in total?

  #131 opened 10 months ago by Syn000
  2

• 请问python版本是什么

  #152 opened 6 months ago by Issac304
  1

• Supplementary material

  #150 opened 6 months ago by mtchiu
  0

• Curious about the multi-turn conversation demo

  #149 opened 6 months ago by chris5zk
  0

• Is the demo link wrong ?

  #148 opened 6 months ago by HuangChiEn
  0

• [seg] token inference

  #145 opened 7 months ago by xiexing0916
  0

• Supplementary material of the accepted paper

  #144 opened 7 months ago by yo3nglau
  0

• How to evaluate on RefCOCO datasets ?

  #143 opened 7 months ago by Z-MU-Z
  0

• Token indices sequence length is longer than the specified maximum sequence length for this model (565 > 512). Running this sequence through the model will result in indexing errors

  #128 opened a year ago by Roberyan
  1

• IndexError: The shape of the mask [6, 440] at index 1 does not match the shape of the indexed tensor [6, 760, 256] at index 1

  #142 opened 8 months ago by Oseemaker
  0

• flash_attn

  #137 opened 9 months ago by nightbreezecool
  2

• Inference speed on refcocog dataset

  #126 opened 8 months ago by Yting68
  0

• Is LISA model on Hugginface?

  #127 opened a year ago by gauravsh0812
  2

• Finetuning Projection Layer or Not

  #139 opened 8 months ago by ziqipang
  0

• LISA-7B-v1&& LISA-7B-v1-explanatory

  #138 opened 9 months ago by nightbreezecool
  0

• datasets download links

  #136 opened 9 months ago by nightbreezecool
  0

• Incorect online demo link

  #134 opened 9 months ago by dohyun1411
  0

• LISA model_forward bug, only occurs when inference.

  #132 opened 9 months ago by Roberyan
  2

• Code error about scheduler in train_ds.py

  #133 opened 10 months ago by eeeric-code
  0

• Why apply the causal mask to image tokens in the attention operations of LLM?

  #130 opened 10 months ago by yubin1219
  0

• How to generate ReasonSeg Dataset

  #129 opened a year ago by ZhixiongSun
  0