aws-neuron

/

optimum-neuron-cache

like 10

I think that if you just edit the config.json to set use_cache to True, then the config is identical to the meta-llama/Meta-LLama-3-8B config and the model will be detected as cached.