Download the models in the build phase

Dockerfile

@@ -19,6 +19,10 @@ WORKDIR /code
 # 
 COPY ./requirements.txt /code/requirements.txt
 
+# Download the models
+RUN python -c 'from transformers import AutoModelForSpeechSeq2Seq; AutoModelForSpeechSeq2Seq.from_pretrained("openai/whisper-large-v3");'
+RUN python -c 'from transformers import AutoProcessor; AutoProcessor.from_pretrained("openai/whisper-large-v3");'
+
 # 
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 

app/main.py

@@ -10,12 +10,11 @@ torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
 model_id = "openai/whisper-large-v3"
 
 model = AutoModelForSpeechSeq2Seq.from_pretrained(
-    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True,
-    cache_dir="./"
+    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
 )
 model.to(device)
 
-processor = AutoProcessor.from_pretrained(model_id, cache_dir="./")
+processor = AutoProcessor.from_pretrained(model_id)
 
 pipe = pipeline(
     "automatic-speech-recognition",