Update app.py

app.py

@@ -22,8 +22,6 @@ sys.path.append('/home/user/app/av_hubert/avhubert')
 print(sys.path)
 print(os.listdir())
 
-from fairseq import checkpoint_utils, options, tasks, utils
-from argparse import Namespace
 
 
 
@@ -46,13 +44,24 @@ from huggingface_hub import hf_hub_download
 import gradio as gr
 
 user_dir = "/home/user/app/av_hubert/avhubert"
+utils.import_user_module(Namespace(user_dir=user_dir))
+data_dir = tempfile.mkdtemp()
+
 ckpt_path = hf_hub_download('vumichien/AV-HuBERT', 'model.pt')
 face_detector_path = "/home/user/app/mmod_human_face_detector.dat"
 face_predictor_path = "/home/user/app/shape_predictor_68_face_landmarks.dat"
 mean_face_path = "/home/user/app/20words_mean_face.npy"
 mouth_roi_path = "/home/user/app/roi.mp4"
+modalities = ["video"]
+gen_subset = "test"
+gen_cfg = GenerationConfig(beam=20)
 models, saved_cfg, task = checkpoint_utils.load_model_ensemble_and_task([ckpt_path])
-utils.import_user_module(Namespace(user_dir=user_dir))
+models = [model.eval().cuda() if torch.cuda.is_available() else model.eval() for model in models]
+saved_cfg.task.modalities = modalities
+saved_cfg.task.data = data_dir
+saved_cfg.task.label_dir = data_dir
+task = tasks.setup_task(saved_cfg.task)
+generator = task.build_generator(models, gen_cfg)
 
 def detect_landmark(image, detector, predictor):
     gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
@@ -100,17 +109,7 @@ def predict(process_video):
       fo.write("".join(tsv_cont))
     with open(f"{data_dir}/test.wrd", "w") as fo:
       fo.write("".join(label_cont))
-    modalities = ["video"]
-    gen_subset = "test"
-    gen_cfg = GenerationConfig(beam=20)
-    
-    models = [model.eval().cuda() if torch.cuda.is_available() else model.eval() for model in models]
-    saved_cfg.task.modalities = modalities
-    saved_cfg.task.data = data_dir
-    saved_cfg.task.label_dir = data_dir
-    task = tasks.setup_task(saved_cfg.task)
     task.load_dataset(gen_subset, task_cfg=saved_cfg.task)
-    generator = task.build_generator(models, gen_cfg)
 
     def decode_fn(x):
         dictionary = task.target_dictionary