Added updates to ads parsing

usc-sail · Apr 5, 2023 · 29f4817 · 29f4817
1 parent 5895c80
commit 29f4817
Show file tree

Hide file tree

Showing 5 changed files with 17 additions and 16 deletions.
diff --git a/configs/config_MHA_single_task_classifier_shot_level.yaml b/configs/config_MHA_single_task_classifier_shot_level.yaml
@@ -1,6 +1,6 @@
 data:
   csv_file: '/data/digbose92/ads_complete_repo/ads_codes/SAIM-ADS/data/SAIM_data/SAIM_multi_task_tone_soc_message_topic_data_no_zero_files.csv'
-  base_folder: '/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/vit_features'
+  base_folder: '/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/clip_features_4fps/'
 
 parameters:
   batch_size: 16
@@ -31,11 +31,11 @@ optimizer:
   verbose: True
 
 model:
-  model_type: 'MHA_attn_single_task_classifier_shot_level'
-  input_dim: 768
+  model_type: 'MHA_attn_single_task_classifier_CLIP_features_shot_level'
+  input_dim: 512
   model_dim: 256
-  num_heads: 2
-  num_layers: 2
+  num_heads: 4
+  num_layers: 4
   input_dropout: 0.2
   output_dropout: 0.2
   model_dropout: 0.2

diff --git a/configs/config_MHA_topic_classifier_shot_level.yaml b/configs/config_MHA_topic_classifier_shot_level.yaml
@@ -1,7 +1,8 @@
 data:
   csv_file: '/data/digbose92/ads_complete_repo/ads_codes/SAIM-ADS/data/SAIM_data/SAIM_multi_task_tone_soc_message_topic_data_no_zero_files.csv'
   topic_file: '/data/digbose92/ads_complete_repo/ads_codes/SAIM-ADS/data/topic_list_18.json'
-  base_folder: '/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/vit_features'
+  base_folder: '/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/clip_features_4fps/'
+  #'/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/vit_features'
 
 parameters:
   batch_size: 16
@@ -24,17 +25,15 @@ optimizer:
   lr: 1e-4
   gamma: 0.5
   step_size: 15
-  #scheduler: 'step_lr'
-  #mode: 'max'
   decay: 0.001
   patience: 5
   factor: 0.5
   verbose: True
 
 model:
-  model_type: 'MHA_attn_single_task_classifier_shot_level'
-  input_dim: 768
-  model_dim: 128
+  model_type: 'MHA_attn_single_task_classifier_CLIP_features_shot_level'
+  input_dim: 512
+  model_dim: 256
   num_heads: 4
   num_layers: 4
   input_dropout: 0.2

diff --git a/datasets/__pycache__/dataset.cpython-38.pyc b/datasets/__pycache__/dataset.cpython-38.pyc
diff --git a/datasets/dataset.py b/datasets/dataset.py
@@ -251,9 +251,11 @@ def __getitem__(self,idx):
         #print(filename)
 
         #load the feature file
-
-        with open(filename, 'rb') as f:
-            shot_features = pickle.load(f)
+        try:
+            with open(filename, 'rb') as f:
+                shot_features = pickle.load(f)
+        except:
+            print(filename)
 
         #get the keys 
         keys=sorted(list(shot_features.keys()))

diff --git a/feature_extraction/extract_clip_features.py b/feature_extraction/extract_clip_features.py
@@ -235,8 +235,8 @@ def run_shot_inference(shot_folder,model,device,preprocess,destination_folder):
             #destination file
             destination_file=os.path.join(destination_folder,pkl_filename)
 
-            if(os.path.exists(destination_file) is False):
-
+            if(destination_file=='/data/digbose92/ads_complete_repo/ads_features/shot_embeddings/clip_features_4fps/arm_hammer_arm_hammer_clump_seal_cat_litter.pkl'):
+                #if(os.path.exists(destination_file) is False):
                 shot_list=os.listdir(video_subfolder) #list of shots
                 shot_dict=dict()
                 for shot_file in tqdm(shot_list):