replicate
cog predict -remote always picks up CPU instead of GPU.
Cog remote does not take the reference of remote server(Nvidia T4 GPU) in case the local reference is CPU. I have performed many tests but to no use,
main ~/LatentSync sudo cog predict r8.im/iageneration/latentsync -i actor=sophie -i audio=@assets/speech_output.wav -i guidance_scale=1.5 -i seed=42
Starting Docker image r8.im/iageneration/latentsync and running setup()... docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]]. Missing device driver, re-trying without GPU Error response from daemon: page not found /root/.pyenv/versions/3.10.15/lib/python3.10/site-packages/cog/server/http.py:314: DeprecationWarning: on_event is deprecated, use lifespan event handlers instead.
Read more about it in the
[FastAPI docs for Lifespan Events](https://fastapi.tiangolo.com/advanced/events/).
@app.on_event("startup") /root/.pyenv/versions/3.10.15/lib/python3.10/site-packages/cog/server/http.py:328: DeprecationWarning: on_event is deprecated, use lifespan event handlers instead.
Read more about it in the
[FastAPI docs for Lifespan Events](https://fastapi.tiangolo.com/advanced/events/).
@app.on_event("shutdown") Running prediction... CUDA available Checking from Predict Functionality: False /bin/sh: 1: nvidia-smi: not found nvidia-smi not available (likely no GPU allocated) Using seed: 42 The cache for model files in Transformers v4.22.0 has been updated. Migrating your old cache. This is a one-time only operation. You can interrupt this and resume the migration later on by calling transformers.utils.move_cache(). 0it [00:00, ?it/s] 0it [00:00, ?it/s] ^CStopping container... ⅹ Failed to predict: Failed to POST HTTP request to http://localhost:50005/predictions: Post "http://localhost:50005/predictions": EOF
Can you try booting the underlying docker image up and letting me know whether that picks up the GPU?
