I can't find many files

[SunshineXiang]

Hello, Sir: In the main function, many parameters are missing and cannot be run. For example, in the Image reconstruction file and def main(): 1.Semantic_feature_extraction.py I can't find LDM_model 1.model_typology = pd.read_csv( '/nfs/diskstation/DataStation/ChangdeDu/luyizhuo/model_typology.csv') I can't find model_typology.csv 2.n_feats decoding_target_layers(args.CLIP_feature_decoder_path, args.mask_path, args.val_CLIP_mean_path, args.val_CLIP_std_path, args.target_layers, 25, 'fastl2',n_feats, args.recons_img_idx) 3.loss_CLIP_weight loss = MSE_CLIP_mask(masks=feature_masks, target=CLIP_target, generated=CLIP_generated, weight=loss_CLIP_weight) 4.In the Feature extractor files parser = argparse.ArgumentParser(description='CLIP text feature extraction') parser.add_argument('--stable_diffusion_config_path', default='', type=str) parser.add_argument('--stable_diffusion_skpt_path', default='', type=str) parser.add_argument('--coco_ids_path', default='', type=str) parser.add_argument('--stimuli_data_path', default='', type=str) parser.add_argument('--feature_saved_path', default='', type=str) args = parser.parse_args() Where are the paths to these files? 5.I can't find coco_ids_path and stimuli_data_path Semantic_feature_extraction.py I can't find LDM_model. Can you help me and Can you give me some fully running code ？ Thanks

[ReedOnePeck]

Sure, you can find ways to run the project quickly here： https://github.com/ReedOnePeck/MindDiffuser/issues/2#issue-1874693977 If you have any other questions, please feel free to communicate。

[ReedOnePeck]

2.As for args.target_layers you can find it in the body of the paper, or you can set the layers you want yourself

2. you can set “args.recons_img_idx” you want yourself
3. parser.add_argument('--stimuli_data_path', default='', type=str) parser.add_argument('--feature_saved_path', default='', type=str) These two are custom parameters.

[SunshineXiang]

I am not sure how to set this up and download

 

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年9月30日(星期六) 晚上6:42 @.>; @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

Sure, you can find ways to run the project quickly here：

2 (comment)

If you have any other questions, please feel free to communicate。

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.***>

[ReedOnePeck]

Which step in this do you not know how to set up. Please be more specific.

[SunshineXiang]

In the Semantic_feature_extraction.py file,       parser.add_argument('--coco_ids_path',  default='./', type=str) I don't know how to make coco ids path, I am looking forward to your reply, Thanks!

 

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年9月30日(星期六) 晚上9:12 @.>; @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

Which step in this do you not know how to set up. Please be more specific.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.***>

[ReedOnePeck]

You can refer to this code https://github.com/styvesg/nsd

[SunshineXiang]

Thank you very much for your guidance！  

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年9月30日(星期六) 晚上10:03 @.>; @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

You can refer to this code https://github.com/styvesg/nsd

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.***>

[ReedOnePeck]

In the training dataset of NSD, there are 8859 images. To facilitate access, I used the "Image" tool to convert each image into a size of 512x512, and saved them as a .npy file with dimensions 8859x3x512x512.

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年10月6日(星期五) 晚上8:16 @.>; 抄送: "Yizhuo @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

Hello Sir, I need your help, In the detail_extracttion.py --stim_saved_path, I don't know how to make this file I am looking forward to your reply. Thanks!

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.***>

[SunshineXiang]

1.In the detail_extracttion.py              '--stim_saved_path', I give 73000 images 2.In the Structural_feature_extraction.py    '--stim_sub1_trn', Why define 8859 images when I don't see where it is defined. 3.The data I read out was 10,000 images per subject.  (10000, 3, 425, 425) I am looking forward to your reply.  Thanks! 

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年10月7日(星期六) 凌晨0:20 @.>; @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

In the training dataset of NSD, there are 8859 images. To facilitate access, I used the "Image" tool to convert each image into a size of 512x512, and saved them as a .npy file with dimensions 8859x3x512x512.

------------------ 原始邮件 ------------------ 发件人: "ReedOnePeck/MindDiffuser" @.>; 发送时间: 2023年10月6日(星期五) 晚上8:16 @.>; 抄送: "Yizhuo @.>;"State @.>; 主题: Re: [ReedOnePeck/MindDiffuser] I can't find many files (Issue #3)

Hello Sir, I need your help, In the detail_extracttion.py --stim_saved_path, I don't know how to make this file I am looking forward to your reply. Thanks!

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.>