SAMed

SAMed copied to clipboard

Questions about custom dataset

3

I am trying to train the model using custom dataset. My dataset has multiple masks for one single image but when I looked through [files](https://drive.google.com/file/d/1zuOQRyfo0QYgjcU_uZs0X3LdCnAC2m3G/view) inside train.npz where one image...

ariharasudhanm

No much performance difference between SAMed and SAMed_h on my own dataset

1

Thanks for your code. However, I found that there existed less performance difference between SAMed and SAMed_h for my own dataset. May I know how to fine-tune the parameters to...

maosuli

Ran out of input问题

3

您好，很感谢您公开的代码，我在使用您分享的train_npz_new_224数据进行训练时，出现了以下错误，看了所有的issues，没有找到类似的问题，研究了很久没能解决，您有遇到类似的问题吗？非常感谢！ 以下为对train.py的设置： `parser = argparse.ArgumentParser() parser.add_argument('--root_path', type=str, default='./datasets/train_npz_new_224', help='root dir for data') parser.add_argument('--output', type=str, default='./output') parser.add_argument('--dataset', type=str, default='Synapse', help='experiment_name') parser.add_argument('--list_dir', type=str, default='./lists/lists_Synapse', help='list dir') parser.add_argument('--num_classes', type=int, default=8, help='output channel of...

luobendewugong

Finetune SAM on rgb dataset (cityscapes)

6

Thank you for your work. I manage to reproduce your implementation but I encounter problems adapting it to the dataset: https://www.cityscapes-dataset.com/ Do you have a recommendation to adapt your implementation...

alanent

示例笔记本, 推理长达30分钟? 只有我这样吗?

2

 我是在colab上面运行,[示例](https://colab.research.google.com/drive/1KCS5ulpZasYl9DgJJn59WsGEB8vwSI_m?usp=sharing)笔记本, 推理时间咋这么长呀?

jumbojing

您好，我使用labelme标注的数据集（json文件+jpg图像）如何转为该算法的训练数据类型呢？谢谢！

AImagicAI

您好，我发现论文中所用的数据集的测试集没有标签，可能是因为比赛的时候隐藏了，请问有哪位拿到了吗？  The testing file does not have a corresponding label. Has anyone got it?

9Cirno9

仅胆囊（gallbladder）一项的测试输出一直为零

3

您好，非常感谢您的佳作！ 我根据您论文和代码中提供的数据集和参数进行训练，train.py中仅改为单卡，其余参数均保持不变，loss曲线基本和您在论文中提供的一致。尽管我后续尝试在SAMed_s中训练，但在多次实验中胆囊（gallbladder）一项的测试输出一直为零，请问您之前出现过这样的情况吗？若想继续实验，您对此有什么看法吗？ 非常期待您的回复！ ``` [23:37:28.165] Namespace(ckpt='checkpoints/sam_vit_b_01ec64.pth', config=None, dataset='Synapse', deterministic=1, img_size=512, input_size=224, is_savenii=True, list_dir='./lists/lists_Synapse/', lora_ckpt='checkpoints/epoch_159.pth', module='sam_lora_image_encoder', num_classes=8, output_dir='test_result', rank=4, seed=1234, vit_name='vit_b', volume_path='testset/test_vol_h5/') [23:37:28.166] 12 test iterations per epoch [23:38:59.852] idx 0...

PsyFalos

SAMed's input image resolution is 512*512,but in SAM is 1024*1024

2

hello,thanks for your great work. I want to know if the resolution difference will change the perfomance ,because you use the same checkpoints.

SilentWhiteRabbit

自己训练的pth预测全为0

7

作者您好，我用您的数据集运行了train，保存的pth文件用于测试中，但是test的时候预测输出的全为0，可是使用您训练好的pth测试时候，就会生成预测结果，请问一下是需要该哪部分代码吗

hgtlab