Cannot deploy app, and the app crashes every time

Hi, I am trying to deploy this app here- GitHub - ghosh-r/dalle-mini-demo: Demo for the app Dall-E Mini, an Open Source version built to replicate what Dall-E does. Part of Hugging Face JAX Community Week..

The app is crashing every time.

Here is part of the log that I see that causes the problem-

Partial Log
[manager] Python dependencies were installed from /app/dalle-mini-demo/requirements.txt using pip.

[manager] Processed dependencies!




[client] Connecting...

wandb: Currently logged in as: ghosh-r (use `wandb login --relogin` to force relogin)

wandb: Tracking run with wandb version 0.10.33

wandb: Syncing run silver-yogurt-4

wandb:  View project at https://wandb.ai/ghosh-r/dalle-mini-demo

wandb:  View run at https://wandb.ai/ghosh-r/dalle-mini-demo/runs/16t4wxyp

wandb: Run data is saved locally in /app/dalle-mini-demo/wandb/run-20210715_214850-16t4wxyp

wandb: Run `wandb offline` to turn off syncing.

wandb: Downloading large artifact model-1ef8yxby:latest, 1674.97MB. 2 files... [2021-07-15 21:49:15.648558] Done. 0:0:0

2021-07-15 21:49:15.968 Lock 140615458116608 acquired on /home/appuser/.cache/huggingface/transformers/0d6fc8b2ef1860c1f8f0baff4b021e3426cc7d11b153f98e563b799603ee2f25.647b4548b6d9ea817e82e7a9231a320231a1c9ea24053cc9e758f3fe68216f05.lock


signal only works in main thread of the main interpreter

Downloading: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 899k/899k [00:00<00:00, 3.56MB/s][2021-07-15 21:49:16.529057] 

2021-07-15 21:49:16.529 Lock 140615458116608 released on /home/appuser/.cache/huggingface/transformers/0d6fc8b2ef1860c1f8f0baff4b021e3426cc7d11b153f98e563b799603ee2f25.647b4548b6d9ea817e82e7a9231a320231a1c9ea24053cc9e758f3fe68216f05.lock

2021-07-15 21:49:16.982 Lock 140615458115696 acquired on /home/appuser/.cache/huggingface/transformers/6e75e35f0bdd15870c98387e13b93a8e100237eb33ad99c36277a0562bd6d850.5d12962c5ee615a4c803841266e9c3be9a691a924f72d395d3a6c6c81157788b.lock

Downloading: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 456k/456k [00:00<00:00, 2.61MB/s][2021-07-15 21:49:17.471089] 

2021-07-15 21:49:17.471 Lock 140615458115696 released on /home/appuser/.cache/huggingface/transformers/6e75e35f0bdd15870c98387e13b93a8e100237eb33ad99c36277a0562bd6d850.5d12962c5ee615a4c803841266e9c3be9a691a924f72d395d3a6c6c81157788b.lock

2021-07-15 21:49:18.393 Lock 140615449535296 acquired on /home/appuser/.cache/huggingface/transformers/1abf196c889c24daca2909359ca2090e5fcbfa21a9ea36d763f70adbafb500d7.67d01b18f2079bd75eac0b2f2e7235768c7f26bd728e7a855a1c5acae01a91a8.lock

Downloading: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 26.0/26.0 [00:00<00:00, 21.2kB/s]-15 21:49:18.694396] 

2021-07-15 21:49:18.694 Lock 140615449535296 released on /home/appuser/.cache/huggingface/transformers/1abf196c889c24daca2909359ca2090e5fcbfa21a9ea36d763f70adbafb500d7.67d01b18f2079bd75eac0b2f2e7235768c7f26bd728e7a855a1c5acae01a91a8.lock

2021-07-15 21:49:18.995 Lock 140615458117136 acquired on /home/appuser/.cache/huggingface/transformers/d94f53c8851dcda40774f97280e634b94b721a58e71bcc152b5f51d0d49a046a.fc9576039592f026ad76a1c231b89aee8668488c671dfbe6616bab2ed298d730.lock

Downloading: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 1.36M/1.36M [00:00<00:00, 5.22MB/s]2021-07-15 21:49:19.565525] 

2021-07-15 21:49:19.565 Lock 140615458117136 released on /home/appuser/.cache/huggingface/transformers/d94f53c8851dcda40774f97280e634b94b721a58e71bcc152b5f51d0d49a046a.fc9576039592f026ad76a1c231b89aee8668488c671dfbe6616bab2ed298d730.lock

/home/appuser/venv/lib/python3.9/site-packages/transformers/models/bart/configuration_bart.py:180: UserWarning: Please make sure the config includes `forced_bos_token_id=16384` in future versions.The config can simply be saved and uploaded again to be fixed.

  warnings.warn(

2021-07-15 21:49:19.705 Starting the local TPU driver.

2021-07-15 21:49:19.738 Unable to initialize backend 'tpu_driver': Not found: Unable to find driver in registry given worker: local://

2021-07-15 21:49:19.741 Unable to initialize backend 'gpu': Not found: Could not find registered platform with name: "cuda". Available platform names are: Host Interpreter

2021-07-15 21:49:19.744 Unable to initialize backend 'tpu': Invalid argument: TpuPlatform is not available.

2021-07-15 21:49:19.744 No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)

[manager] Streamlit server consistently failed status checks

[manager] Please fix the errors, push an update to the git repo, or reboot the app.

I don’t know what causes this, as I don’t get a direct error message.

This is the Colab Notebook I am trying to convert to a Streamlit app- Google Colaboratory

This might give you an idea of what is going wrong.

Any help is appreciated.

PS- Here’s the app- https://share.streamlit.io/ghosh-r/dalle-mini-demo/main/main.py

Hi @Rito -

It’s important to note that Streamlit sharing does not provide GPU/TPU, so depending on what code the app is calling, it might be getting stuck trying to call drivers that don’t exist. I’m not an expert on HuggingFace, so I would start there to understand what the requirements actually are for your app. Additionally, I would minimize the packages listed in your requirements.txt to what you actually call in the app, to ensure you don’t have any inter-package dependency issues with having every version specified.

Best,
Randy

1 Like