Update CXR model

model-zoo/README.md

@@ -0,0 +1,7 @@
+# Generative Models - Model Zoo
+
+In this directory, we include the prototypes of the model zoo for the MONAI Generative Models project.
+Different from the official one, we do not include all features from the [official one](https://github.com/Project-MONAI/model-zoo).
+For this reason, it is not possible to download the models directly with the `python -m monai.bundle run ...` command.
+In order to use our models, please, manually download them with their link specified in the `large_files.yml` files,
+and place them inside the folder path specified in the same .yml file.

model-zoo/models/cxr_image_synthesis_latent_diffusion_model/configs/inference.json

@@ -40,8 +40,8 @@
             false,
             false
         ],
-        "with_encoder_nonlocal_attn": true,
-        "with_decoder_nonlocal_attn": true
+        "with_encoder_nonlocal_attn": false,
+        "with_decoder_nonlocal_attn": false
     },
     "load_autoencoder_path": "$@model_dir + '/autoencoder.pth'",
     "load_autoencoder": "$@autoencoder_def.load_state_dict(torch.load(@load_autoencoder_path))",

model-zoo/models/cxr_image_synthesis_latent_diffusion_model/configs/metadata.json

@@ -2,7 +2,7 @@
     "schema": "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/meta_schema_20220324.json",
     "version": "1.0.0",
     "changelog": {
-        "0.1": "Initial release"
+        "0.2": "Flipped images fixed"
     },
     "monai_version": "1.1.0",
     "pytorch_version": "1.13.0",

model-zoo/models/cxr_image_synthesis_latent_diffusion_model/docs/README.md

@@ -12,7 +12,11 @@ This model is trained from scratch using the Latent Diffusion Model architecture
 2D Chest X-ray conditioned on Radiological reports. The model is divided into two parts: an autoencoder with a
 KL-regularisation model that compresses data into a latent space and a diffusion model that learns to generate
 conditioned synthetic latent representations. This model is conditioned on Findings and Impressions from radiological
-reports.
+reports. The original repository can be found [here](https://github.com/Warvito/generative_chestxray)
+
+![](./figure_1.png) <br>
+<p align="center">
+Figure 1 - Synthetic images from the model. </p>
 
 ## **Data**
 The model was trained on brain data from 90,000 participants from the MIMIC dataset [2] [3]. We downsampled the

model-zoo/models/cxr_image_synthesis_latent_diffusion_model/large_files.yml

@@ -1,9 +1,9 @@
 large_files:
   - path: "models/autoencoder.pth"
-    url: "https://drive.google.com/uc?export=download&id=11Em6qkEsqbFrtJau2mlZAvQZqQmiIEVe"
+    url: "https://drive.google.com/uc?export=download&id=1paDN1m-Q_Oy8d_BanPkRTi3RlNB_Sv_h"
     hash_val: ""
     hash_type: ""
   - path: "models/diffusion_model.pth"
-    url: "https://drive.google.com/uc?export=download&id=1PUqHb_0dKB7GAXA3P8l_3pIyorLudgrB"
+    url: "https://drive.google.com/uc?export=download&id=1CjcmiPu5_QWr-f7wDJsXrCCcVeczneGT"
     hash_val: ""
     hash_type: ""