Hacker News new | ask | show | jobs
by throwthrowuknow 643 days ago
Not going to work for training from scratch which is what the author is doing.
1 comments
rspoerri 642 days ago
192GByte of RAM are not enough to train 405B models. Reflection 70B requires 140GByte of RAM in fp16, 405 would need ~810Gbyte of RAM.
link
throwthrowuknow 642 days ago
Pretty sure he said he’s inferencing llama3 405 and training his own custom model from scratch. He didn’t say how big his custom model will be.
link