Hello and thank you for taking the time to read this.
I’m attempting to deploy an app that uses a ML model to generate predictions given some user inputs. The model was created using Pycaret.
The files for this model can be found here
The app location can be found here
I’m able to deploy the app locally and when I deploy the app locally everything works as intended. Here’s a screenshot of the main model page:
However when I deploy the app (app.py) to streamlit, it returns a permission error pertaining to the model itself which i’ve copied below:
PermissionError: This app has encountered an error. The original error message is redacted to prevent data leaks. Full error details have been recorded in the logs (if you’re on Streamlit Cloud, click on ‘Manage app’ in the lower right of your app).
Traceback:
File “/home/adminuser/venv/lib/python3.9/site-packages/streamlit/runtime/scriptrunner/script_runner.py”, line 565, in _run_script
exec(code, module.dict)
File “/mount/src/transitcostestimator/streamlit/app.py”, line 34, in
model = pickle.load(open(‘streamlit/finalized_user_model.pkl’, ‘rb’))
File “/home/adminuser/venv/lib/python3.9/site-packages/pycaret/internal/pipeline.py”, line 152, in setstate
self.memory = state[“_memory”]
File “/home/adminuser/venv/lib/python3.9/site-packages/pycaret/internal/pipeline.py”, line 176, in memory
self._memory_fit = self._memory.cache(_fit_one)
File “/home/adminuser/venv/lib/python3.9/site-packages/pycaret/internal/memory.py”, line 447, in cache
ret = super().cache(func, ignore, verbose, mmap_mode)
File “/home/adminuser/venv/lib/python3.9/site-packages/joblib/memory.py”, line 1082, in cache
return MemorizedFunc(
File “/home/adminuser/venv/lib/python3.9/site-packages/joblib/memory.py”, line 447, in init
self.store_backend.store_cached_func_code([
File “/home/adminuser/venv/lib/python3.9/site-packages/joblib/_store_backends.py”, line 272, in store_cached_func_code
self.create_location(func_path)
File “/home/adminuser/venv/lib/python3.9/site-packages/joblib/_store_backends.py”, line 403, in create_location
mkdirp(location)
File “/home/adminuser/venv/lib/python3.9/site-packages/joblib/disk.py”, line 61, in mkdirp
os.makedirs(d)
File “/usr/local/lib/python3.9/os.py”, line 215, in makedirs
makedirs(head, exist_ok=exist_ok)
File “/usr/local/lib/python3.9/os.py”, line 215, in makedirs
makedirs(head, exist_ok=exist_ok)
File “/usr/local/lib/python3.9/os.py”, line 215, in makedirs
makedirs(head, exist_ok=exist_ok)
[Previous line repeated 5 more times]
File “/usr/local/lib/python3.9/os.py”, line 225, in makedirs
mkdir(name, mode)
This error occurs at this line:
model = pickle.load(open(‘streamlit/finalized_user_model.pkl’, ‘rb’))
However, it also occurs when I used several different methods to load the model (Joblib and Pycaret). Addtionally, it returns the same error when I’m using resource caching, like:
@st.cache_resource
def streamlit_model(model_path):
return load_model(model_path)
model = streamlit_model(‘finalized_user_model’)
Is there a correct way to load a model within Streamlit that I’m not applying here? Is Pycaret potentially an issue here, as well?
I believe the app is attempting to cache the model, but can’t create the directory to store the cache because it doesn’t have permission to save to the cloud. I also could be wrong.
The library version in my requirements file are:
matplotlib == 3.8.0
numpy==1.23.5
pandas==1.5.3
pandas-dq==1.28
pandas-profiling==3.6.6
plotly==5.16.1
plotly-resampler==0.9.1
streamlit==1.21.0
ipython ==8.16.1
pycaret == 3.1.0
I’ve seen a few similar questions with a similar error, however I wasn’t able to resolve the issue with the outcomes of those threads.
Any help or direction is appreciated!
- Ryan