Skip to content

issues Search Results · repo:TideDra/VL-RLHF language:Python

Filter by

18 results
 (65 ms)

18 results

inTideDra/VL-RLHF (press backspace or delete to remove)

TideDra/VL-RLHF
cannot import name 'EncoderDecoderCache' from 'transformers'

When running duo_llavanext.sh, an error occurs: Traceback (most recent call last): File /mnt/dolphinfs/hdd_pool/docker/user/hadoop-aipnlp/FMG/jiaokechen/WORKSPACE/REPO/VL-RLHF/src/vlrlhf/dpo.py , line ...
  • Tbuterin
  • 1
  • Opened 
    15 days ago
  • #19

TideDra/VL-RLHF
请问这个sft支持多轮对话和多图输入吗

  • studentfromChina
  • Opened 
    on Dec 5, 2024
  • #18

TideDra/VL-RLHF
Does this repo support Qwen2VL?

  • w-zhih
  • Opened 
    on Nov 7, 2024
  • #17

TideDra/VL-RLHF
DPO训练

请问使用LLaVA进行DPO支持多图输入吗？比如两张图片
  • XxxZzD
  • Opened 
    on Sep 24, 2024
  • #16

TideDra/VL-RLHF
每个模型的学习率设定有什么经验吗？

看了里面的script，每个模型的学习率和lora rank都不太一样，请问这个有什么调整经验吗？ 另外，再想请教下，vit和llm中的mlp不训练，只训练llm的lora效果会怎么样
  • shipengai
  • 3
  • Opened 
    on Aug 29, 2024
  • #15

TideDra/VL-RLHF
QwenVL使用了多少卡？global batch size是多少？

看了readme和代码中，没有写多少卡
  • shipengai
  • 1
  • Opened 
    on Aug 29, 2024
  • #14

TideDra/VL-RLHF
[BUG] Value Error noqa:E501

When I using dpo_llava finetune on custom dpo dataset, after several steps, I run into the following error message ValueError: The input provided to the model are wrong. The number of image tokens is ...
  • hxhcreate
  • Opened 
    on Jul 27, 2024
  • #13

TideDra/VL-RLHF
支持cogvlm2模型的强化学习训练吗

  • kaka-Cao
  • Opened 
    on Jul 25, 2024
  • #12

TideDra/VL-RLHF
Any suggestion on how to modify code to train single textual modality

Nice code base. Any suggestion on how to modify code to train single textual modality
  • hxhcreate
  • Opened 
    on Jul 23, 2024
  • #11

TideDra/VL-RLHF
微调qwen爆内存

您好，使用原始代码在2张A100 80G上面微调qwen，显存占用两张卡上都只有919M，但是在数据加载过程中？内存占用一直在增加，直到180多G后内存爆了，程序终止。请问这个问题怎么解？ 训练log： image 内存占用： image
  • delian11
  • 3
  • Opened 
    on Jun 27, 2024
  • #10
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue search results · GitHub