Llama.cpp for fun (and maybe profit) - 30 minute

Talk given at BudapestData 2024 on the use of the local `llama.cpp` model runner showing Llama 2, MS Phi 2, Llama 3, CodeLlama and Llava (multimodal) running from a laptop with CPU and/or GPU to answer queries:
https://budapestdata.hu/2024/en/program/schedule/#

Transcript

1. llama.cpp for fun and (maybe) profit BudapestData 2024 talk @IanOzsvald – ianozsvald.com
2. Strategist/Trainer/Speaker/Author 20+ years Trainer – new Fast Pandas course Interim Chief Data Scientist By [ian]@ianozsvald[.com] Ian Ozsvald 2nd Edition!
3. When should we be using LLMs? Strategic Team Advisor By [ian]@ianozsvald[.com] Ian Ozsvald
4. No need for a GPU+VRAM Llama.cpp runs on CPU+RAM Nothing sent off your machine llama.cpp By [ian]@ianozsvald[.com] Ian Ozsvald X
5. Experiment with models as they’re published Use client data/src code – no data sent off your machine Stable – it doesn’t change every week Why use local models? By [ian]@ianozsvald[.com] Ian Ozsvald
6. Prototype ideas! By [ian]@ianozsvald[.com] Ian Ozsvald llama-2-7b-chat.Q5_K_M.gguf 5GB on disk and in RAM, near real time
7. Let’s see MS Phi 2 summarise We can constrain with JSON too See the wackyness early on What’s your strategy to catch varied outputs? Why use local models? By [ian]@ianozsvald[.com] Ian Ozsvald
8. MS Phi2 can “reason” (IS IT RIGHT?) By [ian]@ianozsvald[.com] Ian Ozsvald I had confident answers: 125.2m/s (good Python) 17.2m/s (partial Python with comments that had mistakes), 40m/s and 31.3m/s (as teacher) Which one to believe? My model is quantised (Q5) but random variation exists anyway… The MS post didn’t disclose the prompt they used https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/
9. Similar to JPG compression Shrink the trained model 32→16→8→7/6/5/4/3/2 bits Fewer bits→worse text completion “Q5 generally an acceptable level” Quantisation By [ian]@ianozsvald[.com] Ian Ozsvald
10. Quantisation By [ian]@ianozsvald[.com] Ian Ozsvald Original fp16 models Better Bigger models with higher quantisation still has lower perplexity than simpler, less quantised models Choose the biggest you can K-quants PR https://github.com/ggerganov/llama.cpp/pull/1684
11. Experiment with multi-modal e.g. OCR and checking photo meets requirements What about image queries? By [ian]@ianozsvald[.com] Ian Ozsvald
12. Llava multi-modal Extract facts from images? By [ian]@ianozsvald[.com] Ian Ozsvald llava-v1.5-7b-Q4_K.gguf 4GB on disk & RAM 5s for example llama.cpp provides ./server
13. Trial code-support Code review? “Is this test readable?” What do you do with code and LLMs? Can they help with coding? By [ian]@ianozsvald[.com] Ian Ozsvald
14. Can you explain this function please? By [ian]@ianozsvald[.com] Ian Ozsvald codellama-34b-python.Q5_K_M.gguf 23GB on disk & RAM, 30s for example Can we use this as a “code reviewer” for internal code? codellama answer: “The function test_uniform_distribution creates a list of 10 zeros, then increments the position in that list indicated by the murmurhash3_32() digest of i. It does this 100000 times and then checks if the means of those incremented values are uniformly distributed (i.e., if they're all roughly the same).” (surprisingly clear!) https://github.com/scikit-learn/scikit-learn/blob/main/sklearn/utils/tests/test_murmurhash.py
15. Using the Python API we can learn how it works Get embeddings API leads to understanding By [ian]@ianozsvald[.com] Ian Ozsvald
16. Q&A model trained on “Let, What, Suppose, Calculate, Solve” as very-likely first tokens API leads to understanding By [ian]@ianozsvald[.com] Ian Ozsvald log(p(1)) == 0 log(p(0.5)) == -0.7
17. Summarise .vtt from Zoom Group call, speakers named Needs Python preprocessing Llama 3 7B good (80B better) Python automation Summarising Rebel Zoom calls By [ian]@ianozsvald[.com] Ian Ozsvald
18. Run quantised models on client data locally CPU only or GPU if you want it faster Fixed target – it doesn’t change next week Experience the wackyness – mitigation? Use Python API to see tokens+perplexity+more Why try llama.cpp? By [ian]@ianozsvald[.com] Ian Ozsvald
19. Who will try llama.cpp on their own machine? Discuss: How do we measure correctness? What’s the worst (!) that could go wrong with your projects? Summary By [ian]@ianozsvald[.com] Ian Ozsvald
20. Appendix - Attack via asciiart By [ian]@ianozsvald[.com] Ian Ozsvald
21. Appendix – Ask Mixtral to challenge my Monte Carlo estimation approach By [ian]@ianozsvald[.com] Ian Ozsvald Mixtral gave 5 points and some items I should be careful about, ChatGPT 3.5 gave 7 points, both felt similar
22. WizardCoder is good (tuned llama2) By [ian]@ianozsvald[.com] Ian Ozsvald wizardcoder-python-34b -v1.0.Q5_K_S.gguf 22GB on disk & RAM 15s for example You can replace CoPilot with this for completions