Add plot logging for each experiment

SylviaWhittle · May 31, 2024 · de8c0dc · de8c0dc
1 parent 64eccdc
commit de8c0dc
Show file tree

Hide file tree

Showing 5 changed files with 19 additions and 23 deletions.
diff --git a/.dvc/config b/.dvc/config
@@ -1,5 +1,7 @@
 [core]
     remote = mygdrive
+[hydra]
+    enabled = False
 ['remote "myremote"']
     url = /Users/sylvi/Documents/dvc_remotes/remote-catsnet/
 ['remote "mygdrive"']

diff --git a/dvc.lock b/dvc.lock
@@ -53,20 +53,20 @@ stages:
         train:
           train_data_dir: data/train/
           model_save_dir: models/
-          batch_size: 6
-          epochs: 5
+          batch_size: 2
+          epochs: 50
           norm_upper_bound: 7
           norm_lower_bound: -1
           validation_split: 0.2
     outs:
     - path: models/catsnet_model.keras
       hash: md5
-      md5: 9c5dd66ef65490711024563e8d6c604b
+      md5: 343a08cec07288294b247e83355a616f
       size: 23454939
     - path: results/train/
       hash: md5
-      md5: 60281b02178ac14a7f481c3780f8af2b.dir
-      size: 688
+      md5: 828dbc21d8b42f881955251ca233b4b0.dir
+      size: 4736
       nfiles: 5
   evaluate:
     cmd: python src/evaluate.py
@@ -78,12 +78,12 @@ stages:
       nfiles: 32
     - path: models/catsnet_model.keras
       hash: md5
-      md5: 9c5dd66ef65490711024563e8d6c604b
+      md5: 343a08cec07288294b247e83355a616f
       size: 23454939
     - path: src/evaluate.py
       hash: md5
-      md5: 25a98179bf20b7d153d5b31025f8a39f
-      size: 5865
+      md5: 925f98789798e9920bf396e614d4b29b
+      size: 5764
     params:
       params.yaml:
         base:
@@ -95,11 +95,6 @@ stages:
     outs:
     - path: results/evaluate/
       hash: md5
-      md5: 96b1905ca851ac5f36b0c0679fd19630.dir
-      size: 42
-      nfiles: 1
-    - path: results/evaluate_plots/
-      hash: md5
-      md5: 5ed50b6928005885d5f6fa0ae159c421.dir
-      size: 1714068
-      nfiles: 8
+      md5: 42e330087e2dd0b2297e3df41be8d2c2.dir
+      size: 1700653
+      nfiles: 9
diff --git a/dvc.yaml b/dvc.yaml
@@ -32,13 +32,13 @@ stages:
     - evaluate
     outs:
     - results/evaluate/
-    - results/evaluate_plots/
 metrics:
 - results/train/metrics.json
 - results/evaluate/metrics.json
 plots:
 - results/train/plots/metrics:
     x: step
+- results/evaluate/plots/images
 artifacts:
   catsnet_model:
     path: models/catsnet_model.keras

diff --git a/params.yaml b/params.yaml
@@ -9,8 +9,8 @@ data_split:
 train:
   train_data_dir: data/train/
   model_save_dir: models/
-  batch_size: 6
-  epochs: 5
+  batch_size: 2
+  epochs: 50
   norm_upper_bound: 7
   norm_lower_bound: -1
   validation_split: 0.2

diff --git a/src/evaluate.py b/src/evaluate.py
@@ -101,9 +101,7 @@ def evaluate(
             dice_score = dice(mask_predicted, mask)
             dice_multi += dice_score / len(image_indexes)
 
-            # Plot the image, mask and predicted mask and save it to the results/evaluate/image_plots directory
-            plot_save_dir = Path("results/evaluate_plots")
-            plot_save_dir.mkdir(parents=True, exist_ok=True)
+            # Plot the image, mask and predicted mask and log it
             num_channels = mask_predicted.shape[-1]
             fig, ax = plt.subplots(num_channels, 3, figsize=(15, 5))
             if num_channels == 1:
@@ -121,7 +119,8 @@ def evaluate(
                     ax[i, 1].set_title(f"Ground Truth Mask Channel {i}")
                     ax[i, 2].imshow(mask_predicted[:, :, i], cmap="binary")
                     ax[i, 2].set_title(f"Predicted Mask Channel {i}")
-            plt.savefig(f"{plot_save_dir}/test_image_{index}.png")
+            # plt.savefig(f"{plot_save_dir}/test_image_{index}.png")
+            live.log_image(f"test_image_plot_{index}.png", fig)
 
         live.summary["dice_multi"] = dice_multi