Hello,
I've discovered that the context size reported by llama.cpp and cptr differ, and for some reason cptr stops incrementing context usage.
Steps to reproduce:
- Run a workload that generates context using llama.cpp
- Monitor the context size reported by both llama.cpp and cptr
- Observe that cptr stops incrementing while llama.cpp continues to report increasing context usage
Thank you for your time and help investigating this!
Best regards,
Joshua Krimmer
Hello,
I've discovered that the context size reported by llama.cpp and cptr differ, and for some reason cptr stops incrementing context usage.
Steps to reproduce:
Thank you for your time and help investigating this!
Best regards,
Joshua Krimmer