Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

How to generate my own few_shot file just as "coco_fewshot" when finetune on custom dataset? #80

Open
HazelNutii opened this issue Feb 6, 2024 · 5 comments

Comments

@HazelNutii
Copy link

I don't know how the coco_fewshot folder (already provided in this repo) is create.

If I want to train my own dataset, how to generate the corresponding file just as coco_fewshot?

@yw123104717
Copy link

I also want to ask the same question. Have you solved it? Thank you very much for your reply

@ningsummerw
Copy link

你好,请问你解决用自己数据集来生成对应coco-few shot的问题了吗

@yw123104717
Copy link

你好,请问你解决用自己数据集来生成对应coco-few shot的问题了吗

你好,没有解决这个问题,换成了现有的few-shot划分基准和数据集去做实验,对于自己的数据集生成coco_fewshot并没有解决,不过我当时发邮件给博主提供了一个思路,等我找到了回复你

@ningsummerw
Copy link

好的,期待你的回复

@xiaozhang-ok
Copy link

你好,请问你解决用自己数据集来生成对应coco-few shot的问题了吗

你好,没有解决这个问题,换成了现有的few-shot划分基准和数据集去做实验,对于自己的数据集生成coco_fewshot并没有解决,不过我当时发邮件给博主提供了一个思路,等我找到了回复你

你好,请问你解决用自己数据集来生成对应coco-few shot的问题了吗

你好,没有解决这个问题,换成了现有的few-shot划分基准和数据集去做实验,对于自己的数据集生成coco_fewshot并没有解决,不过我当时发邮件给博主提供了一个思路,等我找到了回复你

你好,这个模型不会自动根据你的数据集生成spport数据集的吗?

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants