Is there a proper way to retrieve a tensor from the cache?

[wibox]

I'm trying to make federated averaging work inside RedisAI itself. The biggest problem seems to be that none of the simplest operations required to perform such algorithm - i.e. creating an optimizer, making it load a precise state and doing a step for each round - seems to be working due to some pathetic torchscript compilation issue, which i clearly do not properly understand. Isn't there a way to avoid passing a torchscript to the cache from a string? Also, where can i find more documentation related to what torchscript can or cannot do? Pytorch official documentation would be more useful if it was non existent. Also, is there any documentation related to the redis.execute() API? I'm opening this issue out of 2 weeks of pure frustration. Sorry about the thousand questions.

[alonre24]

Hi,
I'm sorry to hear about your frustration, but the RedisAI project is no longer maintained...
If you want to share what you are trying to do, including the torch script that you are using and the errors that you get, I can try to help (but I can't promise anything at the moment...)