I have the latest version of Lambda stack installed on a Lambda Vector workstation. When importing tensorflow into python, I get “Invalid MIT-MAGIC-COOKIE-1 key” error. It does not appear to affect tensorflow… but I would like to resolve the error. Any advice?
Further to this, when doing the following “msg = tf.constant(‘Hello, Tensorflow!’)”, it generates a number of messages such as the following:
> 2022-09-09 19:21:50.690001: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
> 2022-09-09 19:21:50.690394: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
> 2022-09-09 19:21:50.690728: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
> 2022-09-09 19:21:50.718506: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
> 2022-09-09 19:21:50.718786: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
> 2022-09-09 19:21:50.719007: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:975] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
If you follow the above with “tf.print(msg)” it generates the correct response, but it appears there are a number of issues with the tensorflow install in Lambda stack. Any advice on fixing this?