安装caiman的时候,根据官方教程,发现报如下错误:
2024-05-19 22:00:28.455425: E tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc:9342] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-05-19 22:00:28.455462: E tensorflow/compiler/xla/stream_executor/cuda/cuda_fft.cc:609] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-05-19 22:00:28.455470: E tensorflow/compiler/xla/stream_executor/cuda/cuda_blas.cc:1518] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
经过多方测试,其实是GPT害得,提示说要清理系统里面的环境变量,以免有冲突的CUDA配置,比如多个CUDA版本之类的。
其实就是TensorFlow版本的锅,换到16.1就行了。
以下是我安装流程
conda clean --all
mamba create -n caiman -c conda-forge caiman pip python=3.11
conda activate caiman
pip install tensorflow==2.16.1
caimanmanager install
后来发现tensorflow用2.16.1太高,有些keras说找不到Sequential,其实上述错误并不影响使用,但是对于GPU加速来说还待定,按照官方的教程安装就行了,回头研究一下GPU。
再后来,在另外一台新的机子上装,发现自动按了tensorflow 2.16.1的版本,但是报错,也说是找不到某个模块。固建议用2.15,写于20240525
mamba create -n caiman -c conda-forge caiman tensorflow=2.15