TIGER-AI-Lab/Mantis

Issues

• About the multi image input

  #20 opened a month ago by srymaker
  1

• Merging LoRA weights with base model

  #19 opened a month ago by hungnh1125
  2

• Support for Idefics3

  #17 opened a month ago by chris-tng
  1

• Pretrain Image Data

  #18 opened a month ago by srymaker
  2

• I tried to deploy Mantis in my own server for some test. Do you have any suggestion about the tools which can deploy Mantis to run faster?

  #15 opened 2 months ago by BrenchCC
  2

• How do you use lmms-eval to evaluate single-image, because i want to evaluate OCR ability with my finetuning model

  #13 opened 2 months ago by BrenchCC
  2

• Nice work! Does the mantis has image seperator when sending to LLM?

  #4 opened 2 months ago by lucasjinreal
  14

• Question about mantis-eval matching criteria

  #7 opened 2 months ago by azshue
  2

• Can 4 32G V100s support its training?

  #14 opened 2 months ago by yeppp27
  1

• 4 GPUS 4090 training bash scripts/finetune.sh error only one machine

  #11 opened 3 months ago by BrenchCC
  0

• How do

  #12 opened 3 months ago by BrenchCC
  0

• Training only with Batch_size = 1 for Multimodal models

  #9 opened 3 months ago by emanuelevivoli
  6

• performance drop when using lora finetune

  #10 opened 3 months ago by luodian873
  2

• about conda env for finetuning

  #8 opened 3 months ago by xing0047
  2

• I have a question about the evaluation of Q-BENCH dataset

  #6 opened 4 months ago by daixiangzi
  1

• Idefics2 full fine-tuning getting RuntimeError: shape mismatch

  #5 opened 4 months ago by chris-tng
  2

• UserWarning: None of the inputs have requires_grad=True. Gradients will be None

  #2 opened 4 months ago by yusuke-coder
  5

• Support Chinese?

  #1 opened 4 months ago by HuangZhen02
  1

• Missing FUYU device_map fix for multi-GPU setups

  #3 opened 5 months ago by ftoschi14
  1