EvolvingLMMs-Lab/LongVA

Issues

• LLaMA 2 or LLaMA 3 based LongVA checkpoint No available?

  #39 opened 4 days ago by YEOMJINSEOP
  0

• About frames number

  #38 opened 10 days ago by ApolloRay
  0

• Failed to reproduce LongVA-7B after training from scratch

  #37 opened 15 days ago by nanocm
  1

• difference between the two models

  #36 opened 23 days ago by Cooperx521
  1

• questions about NIAH

  #35 opened 23 days ago by Cooperx521
  1

• pretrain error

  #25 opened 3 months ago by justgogorunrun
  2

• Thank you for your work. How can we reproduce the V-NIAH benchmark results?

  #7 opened 6 months ago by fistyee
  12

• Failing to reproduce the paper result on videomme

  #33 opened a month ago by joslefaure
  2

• CUDA OUT OF MEMORY

  #24 opened 4 months ago by Wiselnn570
  1

• Loss nan

  #27 opened 3 months ago by Chenfeng1271
  1

• About LLaVA-Next-Qwen2 UniRes Result.

  #18 opened 4 months ago by hcwei13
  1

• Failed to send messages in the local deployed gradio app

  #34 opened a month ago by IceFlameWorm
  0

• Question about apply_seq_parallel_monkey_patch("zigzag_ring_attn", "llama") in eval_vision_niah.py for Qwen2 model and ring attention

  #30 opened 2 months ago by zhang9302002
  1

• Query for training with lora

  #31 opened 2 months ago by sterzhang
  1

• Questions about CLIP-ViT-L/14@336px

  #32 opened a month ago by Wgkai
  1

• ValueError: Unknown model class ModelArguments(model_name_or_path='lmms-lab/LongVA-7B'

  #20 opened a month ago by Wiselnn570
  3

• OOM -- Where is RingAttention applied?

  #28 opened 2 months ago by orrzohar
  1

• Training Question

  #21 opened 4 months ago by aniki-ly
  9

• Only get Lower accuracy than reported in the paper

  #26 opened 3 months ago by justgogorunrun
  1

• Question about chat template

  #22 opened 3 months ago by ThisisBillhe
  2

• Question about RuntimeError: tensors on different device, may cause by offloading

  #23 opened 4 months ago by jinghan23
  1

• Question for Access to LLaVA-1.6 Data

  #10 opened 6 months ago by patrick-tssn
  4

• lmms-lab/Qwen2-7B-Instruct-224K

  #17 opened 4 months ago by GroundMoRe
  0

• video understanding performance

  #16 opened 4 months ago by sunyoe
  1

• about func prepare_zigzag_ring_attn_inputs

  #14 opened 5 months ago by peiliu0408
  0

• Unires in video operation

  #13 opened 5 months ago by GroundMoRe
  0

• Multi-GPU inference

  #12 opened 5 months ago by resi1ience
  5

• ValueError: When load_4bit=True in from longva.model.builder.load_pretrained_model

  #11 opened 5 months ago by hungvtm
  0

• 'Image' object is not iterable

  #4 opened 6 months ago by changqinyao
  2

• Some questions about LongVA

  #8 opened 6 months ago by linhaojia13
  2

• Adapting VQAScore to Longva

  #9 opened 6 months ago by qingfengcss
  0

• The Potential Reason of LLaVA-NeXT-Qwen2's Strong Performance

  #6 opened 6 months ago by waxnkw
  3

• Vision Text Alignment

  #3 opened 6 months ago by justfuwei
  1

• what is the base model for "lmms-lab/LongVA-7B-DPO"

  #2 opened 6 months ago by rebuttalpapers
  3