zyang1580 / CoLLM

The implementation for the work "CoLLM: Integrating Collaborative Embeddings into Large Language Models for Recommendation".
BSD 3-Clause "New" or "Revised" License
60 stars 7 forks source link

Running on CPU #16

Open AatikaNazneen opened 1 month ago

AatikaNazneen commented 1 month ago

Can we train the LoRA module on CPU only and What would be the requirement on CPU RAM required to do this?

zyang1580 commented 3 weeks ago

Sorry, we haven't attempted to train using only the CPU.