Update app.py

app.py

@@ -19,17 +19,20 @@ from jiwer import cer
 from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor, Wav2Vec2FeatureExtractor, Wav2Vec2CTCTokenizer
 
 
-url = "https://huggingface.co/MahtaFetrat/tempmodel/resolve/main/checkpoint-15-1200.zip"
-output_file = wget.download(url)
+# url = "https://huggingface.co/MahtaFetrat/tempmodel/resolve/main/checkpoint-15-1200.zip"
+# output_file = wget.download(url)
 
 # !unzip checkpoint-15-1200.zip -d extracted_model
 
-zip_file = "checkpoint-15-1200.zip"
-output_dir = "extracted_model"
+# zip_file = "checkpoint-15-1200.zip"
+# output_dir = "extracted_model"
 
-subprocess.run(["unzip", zip_file, "-d", output_dir], check=True)
+# subprocess.run(["unzip", zip_file, "-d", output_dir], check=True)
 
 
+model_name = "MahtaFetrat/wav2vec2_finetuned_on_youtube_farsi_30"
+local_dir = snapshot_download(repo_id=model_name)
+
 # Function to split audio into chunks
 def split_audio(audio, sampling_rate, chunk_size=30):
     chunk_length = chunk_size * sampling_rate
@@ -85,7 +88,7 @@ tokenizer = Wav2Vec2CTCTokenizer("./vocab.json", unk_token="<unk>", pad_token="<
 feature_extractor = Wav2Vec2FeatureExtractor(feature_size=1, sampling_rate=16000, padding_value=0.0, do_normalize=True, return_attention_mask=True)
 processor = Wav2Vec2Processor(feature_extractor=feature_extractor, tokenizer=tokenizer)
 
-latest_checkpoint = "extracted_model/checkpoint-15-1200"
+latest_checkpoint = local_dir
 model = Wav2Vec2ForCTC.from_pretrained(latest_checkpoint)
 
 def tuned_wav2vec_speech_file_to_array_fn(path):