RifleZhang LLaVA-Hound-DPO issues - Githubissues

RifleZhang / LLaVA-Hound-DPO

121 stars 18 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Any plan to release the raw selected video data of webvid and vital?

#16 shufangxun opened 2 weeks ago
0
The original short video caption

#15 Arsiuuu opened 1 month ago
1
Which SFT setup is recommended now?

#14 tyleryzhu opened 2 months ago
1
Questions about how the QA instruction set was created and the scripts used

#13 YoungjaeDev opened 2 months ago
1
Source of raw video

#12 exiawsh closed 3 weeks ago
5
How did you select activitynet video clip from scenedetect？

#11 osttkm closed 2 months ago
8
The videos where the frames extracted from

#10 dragonlzm closed 2 months ago
1
WARNING: tokenization mismatch: 2 vs. 89. (ignored)

#9 Liuziyu77 closed 2 months ago
1
DPO training on 4 A10s

#8 Aakriti05 closed 5 months ago
2
Why adding sft loss during dpo training?

#7 sty-yyj closed 5 months ago
1
Which annotation to decide "Scene-xxx" for ActivityNet data

#6 jpWang closed 6 months ago
1
Prompt of generating detailed caption with GPT-4V

#5 XueFuzhao closed 7 months ago
1
Which 240K subset for SFT ?

#4 yuehaoa closed 7 months ago
1
video frame rate

#3 rese1f closed 7 months ago
6
Would you share the inference results, since `gpt-3.5-turbo-0301` will be deprecated this year?

#2 SCZwangxiao closed 7 months ago
2
When will you plan to release the data?

#1 plmsmile closed 7 months ago
1