Open xinfu607 opened 10 months ago
We currently don't have plans for 1B, as 1B might be too small to get interesting behavior out of language models.
someone started to train a 1.1B llama, with gqa https://github.com/jzhang38/TinyLlama
@xinfu607 look up phi-1 and phi-1.5 models
Hi, thanks a lot for your open resource. Please let me know if open llama 1B in the schedule.