I trained a tensorflow model and then saved it as
model.h5 on a gcs bucket, it’s quite big around 1.4GB. But when I try to load it into my streamlit app deployed on the community cloud, the app will crash. It works ok on local, takes about 30 seconds to load_model (eg:
streamlit run my_app.py) . Is it because the .h5 file is too big? What can I do to load it faster?
from keras.models import load_model @st.cache_resource def model_loading(): FS = gcsfs.GCSFileSystem(project=PROJECT_NAME, token=CREDENTIALS ) with FS.open(MODEL_PATH, 'rb') as model_file: model_gcs = h5py.File(model_file, 'r') model = load_model(model_gcs) return model