wdika commited on
Commit
60658e9
1 Parent(s): c4e0f65

Upload config

Browse files
Files changed (1) hide show
  1. readme_template.md +132 -0
readme_template.md ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: apache-2.0
5
+ library_name: atommic
6
+ datasets:
7
+ - StanfordKnees2019
8
+ thumbnail: null
9
+ tags:
10
+ - image-reconstruction
11
+ - XPDNet
12
+ - ATOMMIC
13
+ - pytorch
14
+ model-index:
15
+ - name: REC_XPDNet_StanfordKnees2019_gaussian2d_12x_AutoEstimationCSM
16
+ results: []
17
+
18
+ ---
19
+
20
+
21
+ ## Model Overview
22
+
23
+ XPDNet for 12x accelerated MRI Reconstruction on the StanfordKnees2019 dataset.
24
+
25
+
26
+ ## ATOMMIC: Training
27
+
28
+ To train, fine-tune, or test the model you will need to install [ATOMMIC](https://github.com/wdika/atommic). We recommend you install it after you've installed latest Pytorch version.
29
+ ```
30
+ pip install atommic['all']
31
+ ```
32
+
33
+ ## How to Use this Model
34
+
35
+ The model is available for use in ATOMMIC, and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.
36
+
37
+ Corresponding configuration YAML files can be found [here](https://github.com/wdika/atommic/tree/main/projects/REC/StanfordKnees2019/conf).
38
+
39
+
40
+ ### Automatically instantiate the model
41
+
42
+ ```base
43
+ pretrained: true
44
+ checkpoint: https://huggingface.co/wdika/REC_XPDNet_StanfordKnees2019_gaussian2d_12x_AutoEstimationCSM/blob/main/REC_XPDNet_StanfordKnees2019_gaussian2d_12x_AutoEstimationCSM.atommic
45
+ mode: test
46
+ ```
47
+
48
+ ### Usage
49
+
50
+ You need to download the Stanford Knees 2019 dataset to effectively use this model. Check the [StanfordKnees2019](https://github.com/wdika/atommic/blob/main/projects/REC/StanfordKnees2019/README.md) page for more information.
51
+
52
+
53
+ ## Model Architecture
54
+ ```base
55
+ model:
56
+ model_name: XPDNet
57
+ num_primal: 5
58
+ num_dual: 1
59
+ num_iter: 10
60
+ use_primal_only: true
61
+ kspace_model_architecture: CONV
62
+ kspace_in_channels: 2
63
+ kspace_out_channels: 2
64
+ dual_conv_hidden_channels: 16
65
+ dual_conv_num_dubs: 2
66
+ dual_conv_batchnorm: false
67
+ image_model_architecture: MWCNN
68
+ imspace_in_channels: 2
69
+ imspace_out_channels: 2
70
+ mwcnn_hidden_channels: 16
71
+ mwcnn_num_scales: 0
72
+ mwcnn_bias: true
73
+ mwcnn_batchnorm: false
74
+ normalize_image: true
75
+ dimensionality: 2
76
+ reconstruction_loss:
77
+ wasserstein: 1.0
78
+ ```
79
+
80
+ ## Training
81
+ ```base
82
+ optim:
83
+ name: adamw
84
+ lr: 1e-4
85
+ betas:
86
+ - 0.9
87
+ - 0.999
88
+ weight_decay: 0.0
89
+ sched:
90
+ name: InverseSquareRootAnnealing
91
+ min_lr: 0.0
92
+ last_epoch: -1
93
+ warmup_ratio: 0.1
94
+
95
+ trainer:
96
+ strategy: ddp_find_unused_parameters_false
97
+ accelerator: gpu
98
+ devices: 1
99
+ num_nodes: 1
100
+ max_epochs: 20
101
+ precision: 16-mixed
102
+ enable_checkpointing: false
103
+ logger: false
104
+ log_every_n_steps: 50
105
+ check_val_every_n_epoch: -1
106
+ max_steps: -1
107
+ ```
108
+
109
+ ## Performance
110
+
111
+ To compute the targets using the raw k-space and the chosen coil combination method, accompanied with the chosen coil sensitivity maps estimation method, you can use [targets](https://github.com/wdika/atommic/tree/main/projects/REC/StanfordKnees2019/conf/targets) configuration files.
112
+
113
+ Evaluation can be performed using the [evaluation](https://github.com/wdika/atommic/blob/main/tools/evaluation/reconstruction.py) script for the reconstruction task, with --evaluation_type per_slice.
114
+
115
+ Results
116
+ -------
117
+
118
+ Evaluation against SENSE targets
119
+ --------------------------------
120
+ 12x: MSE = 0.002691 +/- 0.008089 NMSE = 0.1117 +/- 0.1955 PSNR = 27.18 +/- 5.768 SSIM = 0.6544 +/- 0.2702
121
+
122
+
123
+ ## Limitations
124
+
125
+ This model was trained on the StanfordKnees2019 batch0 using a UNet coil sensitivity maps estimation and Geometric Decomposition Coil-Compressions to 1-coil, and might differ from the results reported on the challenge leaderboard.
126
+
127
+
128
+ ## References
129
+
130
+ [1] [ATOMMIC](https://github.com/wdika/atommic)
131
+
132
+ [2] Epperson K, Rt R, Sawyer AM, et al. Creation of Fully Sampled MR Data Repository for Compressed SENSEing of the Knee. SMRT Conference 2013;2013:1