I am trying to deploy a fastAPI server for my reactJs frontend, i am doing some predictions on images using cnn models. While deploying, it is giving this error
2024-11-04 17:40:31.317256: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
Nov 4 11:10:31 PMTo enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
Nov 4 11:10:44 PM2024-11-04 17:40:44.709068: E external/local_xla/xla/stream_executor/cuda/cuda_driver.cc:152] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
==> Exited with status 1
trying to deploy → https://fastapi-6p8p.onrender.com
I have tried different versions of tensorflow, also tried degrading python version in environment variables to be compatible with tensorflow, also removed all the version specifications but it’s not resolving the issue, i am stuck on this for few hours now, any help would be much appreciated.
CUDA requires a GPU. CUDA’s libraries are not present by default on Render, and Render instances do not have GPUs nor any other sort of GPU/ML/AI acceleration.
so is there any way to solve it. I tried to use tensorflow version 2.10 which does not have gpu problems but this is giving me this error
==> Exited with status 1
I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA
and its not like people haven’t deployed any tensorflow projects on render, how can i do that?
The Tensorflow line you provided here is unlikely to be an error, and the actual error was probably above what you pasted. The ==>
line comes after the failure occurs. Are there other lines that say something along the lines of “error”, “critical”, or “failure”? Something that details a specific error reason?
Actually, these lines were showing error in red, but after seeing few of the other logs above these, i was able to resolve the error (just did some more version degradation) and now it is successfully deployed. Thanks a lot for your quick responses
The line appeared in Red because it contained the word “critical”.
can you tell me exactly want version you used for python and tensorflow? i am facing this same issue which you mentioned in the beginning.
These are the few of the requirements
fastapi
uvicorn
pillow
matplotlib
tensorflow-cpu==2.10.0
numpy==1.21.6
and I mentioned python_version to be 3.10 in the environment variables, any tensorflow version equal to or lower than 2.10 shouldn’t give this error if used with compatible python version.