sigma
sigma
Do you have some methods to get embedding-batch by input-batch
by using NDArrayEx Interface, I am already found a solution as follow. NDArray inputUserEmbedding = userInput.getNDArrayInternal() .embedding(userInput, userEmbeddingTable, SparseFormat.DENSE) .singletonOrThrow();
By My Test, GPU memory is very stable, but CPU memory grows bigger and bigger. epoch = 0, CPU memory just used 80GB, but when echo = 30, CPU memory...
I suggest you to learn the method of DJL (DeepJavaLibrary https://djl.ai/ ), it could use all the gpus of single machine.
是否考虑使用 https://github.com/h2oai/datatable.git 代替 pandas, 或者使用pyspark代替pandas 充分利用cpu的能力,加速数据预处理。
#conda activate xss #openssl version #OpenSSL 1.1.1t 7 Feb 2023 # python setup.py install successfully, but import mgclient, throw such error as above,