update

examples/sample_filter/bad_case_find.py

@@ -13,8 +13,8 @@ def get_args():
     parser.add_argument(
         "--data_dir",
         # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\data",
-        default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\us-1",
-        # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\transfer",
+        # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\us-3",
+        default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\transfer",
         type=str
     )
     parser.add_argument(
@@ -37,20 +37,20 @@ def main():
     data_dir = Path(args.data_dir)
     keep_dir = Path(args.keep_dir)
     keep_dir.mkdir(parents=True, exist_ok=True)
-    trash_dir = Path(args.trash_dir)
-    trash_dir.mkdir(parents=True, exist_ok=True)
+    # trash_dir = Path(args.trash_dir)
+    # trash_dir.mkdir(parents=True, exist_ok=True)
 
     client = Client("http://127.0.0.1:7864/")
 
     for idx, filename in tqdm(enumerate(data_dir.glob("*.wav"))):
-        # if idx < 639:
+        # if idx < 400:
         #     continue
         filename = filename.as_posix()
 
         label1, prob1 = client.predict(
             audio=handle_file(filename),
             # model_name="vm_sound_classification8-ch32",
-            model_name="voicemail-en-us-2-ch32",
+            model_name="voicemail-id-id-2-ch32",
             ground_true="Hello!!",
             api_name="/click_button"
         )
@@ -65,9 +65,9 @@ def main():
         )
         prob2 = float(prob2)
 
-        if label1 == "voicemail" and label2 in ("voicemail", "bell") and prob2 > 0.6:
+        if label1 == "voicemail" and label2 in ("voicemail", "bell") and prob1 > 1.0:
             pass
-        elif label1 == "non_voicemail" and label2 not in ("voicemail", "bell") and prob2 > 0.6:
+        elif label1 == "non_voicemail" and label2 not in ("voicemail", "bell") and prob1 > 0.6:
             pass
         else:
             print(f"label1: {label1}, prob1: {prob1}, label2: {label2}, prob2: {prob2}")

examples/sample_filter/wav_find_by_task_excel.py

@@ -12,19 +12,49 @@ from tqdm import tqdm
 from project_settings import project_path
 
 
+"""
+default="task_DC_Death_Cases_20250220.xlsx",
+default="task_DC_Death_Cases_20250221.xlsx",
+default="task_DC_Death_Cases_EC_UC_20250220.xlsx",
+default="task_DC_Death_Cases_EC_UC_20250221.xlsx",
+default="task_DcTask_1_ID_LIVE_20250220_20250220-1.xlsx",
+default="task_DcTask_1_ID_LIVE_20250221_20250221-1.xlsx",
+default="task_DcTask_3_ID_LIVE_20250220_20250220-1.xlsx",
+default="task_DcTask_3_ID_LIVE_20250221_20250221-1.xlsx",
+default="task_DcTask_5_ID_LIVE_20250220_20250220-1.xlsx",
+default="task_DcTask_5_ID_LIVE_20250221_20250221-1.xlsx",
+default="task_DcTask_9_ID_LIVE_20250220_20250220-1.xlsx",
+default="task_DcTask_PTP_ID_LIVE_20250220_20250220-1.xlsx",
+default="task_DcTask_PTP_ID_LIVE_20250220_20250220-1 (1).xlsx",
+default="task_DcTask_PTP_ID_LIVE_20250221_20250221-1.xlsx",
+"""
+
 def get_args():
     parser = argparse.ArgumentParser()
     parser.add_argument(
         "--task_file",
-        default=r"task_DcTask_1_MY_LIVE_20250109_20250109-1.xlsx",
+        # default="task_DC_Death_Cases_20250220.xlsx",
+        # default="task_DC_Death_Cases_20250221.xlsx",
+        # default="task_DC_Death_Cases_EC_UC_20250220.xlsx",
+        # default="task_DC_Death_Cases_EC_UC_20250221.xlsx",
+        # default="task_DcTask_1_ID_LIVE_20250220_20250220-1.xlsx",
+        # default="task_DcTask_1_ID_LIVE_20250221_20250221-1.xlsx",
+        # default="task_DcTask_3_ID_LIVE_20250220_20250220-1.xlsx",
+        # default="task_DcTask_3_ID_LIVE_20250221_20250221-1.xlsx",
+        # default="task_DcTask_5_ID_LIVE_20250220_20250220-1.xlsx",
+        # default="task_DcTask_5_ID_LIVE_20250221_20250221-1.xlsx",
+        # default="task_DcTask_9_ID_LIVE_20250220_20250220-1.xlsx",
+        # default="task_DcTask_PTP_ID_LIVE_20250220_20250220-1.xlsx",
+        # default="task_DcTask_PTP_ID_LIVE_20250220_20250220-1 (1).xlsx",
+        default="task_DcTask_PTP_ID_LIVE_20250221_20250221-1.xlsx",
         type=str
     )
     parser.add_argument(
         "--wav_dir",
-        # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\data-1",
+        default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\id",
         # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\temp-2\temp\VoiceAppVoicemailDetection-1",
         # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\temp-3\temp\VoiceAppVoicemailDetection-1",
-        default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\temp-4\temp\VoiceAppVoicemailDetection-1",
+        # default=r"E:\Users\tianx\HuggingDatasets\vm_sound_classification\data\temp-4\temp\VoiceAppVoicemailDetection-1",
         type=str
     )
     parser.add_argument(

examples/vm_sound_classification/run_batch.sh

@@ -143,13 +143,13 @@ sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name fi
 #--label_plan 2-voicemail \
 #--config_file "yaml/conv2d-classifier-2-ch4.yaml" \
 #--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch4.zip"
-#
-#sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-id-id-2-ch4 \
-#--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/id-ID/wav_finished/*/*.wav" \
-#--label_plan 2-voicemail \
-#--config_file "yaml/conv2d-classifier-2-ch4.yaml" \
-#--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch4.zip"
-#
+
+sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-id-id-2-ch4 \
+--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/id-ID/wav_finished/*/*.wav" \
+--label_plan 2-voicemail \
+--config_file "yaml/conv2d-classifier-2-ch4.yaml" \
+--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch4.zip"
+
 #sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-ja-jp-2-ch4 \
 #--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/ja-JP/wav_finished/*/*.wav" \
 #--label_plan 2-voicemail \
@@ -195,11 +195,11 @@ sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name fi
 #--config_file "yaml/conv2d-classifier-2-ch32.yaml" \
 #--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
 
-sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-en-us-2-ch32 \
---filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/en-US/wav_finished/*/*.wav" \
---label_plan 2-voicemail \
---config_file "yaml/conv2d-classifier-2-ch32.yaml" \
---pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
+#sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-en-us-2-ch32 \
+#--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/en-US/wav_finished/*/*.wav" \
+#--label_plan 2-voicemail \
+#--config_file "yaml/conv2d-classifier-2-ch32.yaml" \
+#--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
 
 #sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-es-mx-2-ch32 \
 #--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/es-MX/wav_finished/*/*.wav" \
@@ -212,13 +212,13 @@ sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name fi
 #--label_plan 2-voicemail \
 #--config_file "yaml/conv2d-classifier-2-ch32.yaml" \
 #--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
-#
-#sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-id-id-2-ch32 \
-#--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/id-ID/wav_finished/*/*.wav" \
-#--label_plan 2-voicemail \
-#--config_file "yaml/conv2d-classifier-2-ch32.yaml" \
-#--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
-#
+
+sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-id-id-2-ch32 \
+--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/id-ID/wav_finished/*/*.wav" \
+--label_plan 2-voicemail \
+--config_file "yaml/conv2d-classifier-2-ch32.yaml" \
+--pretrained_model "/data/tianxing/PycharmProjects/vm_sound_classification/trained_models/voicemail-2-ch32.zip"
+
 #sh run.sh --stage 0 --stop_stage 6 --system_version centos --file_folder_name file_dir --final_model_name voicemail-ja-jp-2-ch32 \
 #--filename_patterns "/data/tianxing/PycharmProjects/datasets/voicemail/ja-JP/wav_finished/*/*.wav" \
 #--label_plan 2-voicemail \