Comment on Ask HN: How can I experiment with LLMs with a old machine?Comments−LargoLasskhyfv1yMake yourself comfortable withhttps://blogs.oracle.com/database/post/freedom-to-build-anno...https://gist.github.com/rssnyder/51e3cfedd730e7dd5f4a816143b...https://www.reddit.com/r/oraclecloud/or any other offer.Deploy some minimal Linux on them, or use what's offered.Plus optionally, if you don't want to instantly start coding from first principles/scratch, make use of established and excellent solutions, likehttps://future.mozilla.org/builders/news_insights/introducin...https://ai-guide.future.mozilla.org/content/running-llms-loc...https://github.com/mozilla-Ocho/llamafilehttps://justine.lol/matmul/and parallelize them withhttps://github.com/b4rtaz/distributed-llamaObviously this needs some knowledge of the command line, so get a good terminal emulator likehttps://iterm2.com/Mend, bend, rend that stuff and see what works how and why, and what not.Edit: Optionally, if you really want to go low-level, with some debugger likehttps://justine.lol/blinkenlights/for 'toy-installations' of smallest models.'Toy' because that doesn't fully support the CPU-instructions which are used in production.Could still help conceptually.−hedgehog0OP1yThank you for the Mozilla AI guide, it's really useful! I knew some of Justine's work and they are really cool.
Comments
Make yourself comfortable with
https://blogs.oracle.com/database/post/freedom-to-build-anno...
https://gist.github.com/rssnyder/51e3cfedd730e7dd5f4a816143b...
https://www.reddit.com/r/oraclecloud/
or any other offer.
Deploy some minimal Linux on them, or use what's offered.
Plus optionally, if you don't want to instantly start coding from first principles/scratch, make use of established and excellent solutions, like
https://future.mozilla.org/builders/news_insights/introducin...
https://ai-guide.future.mozilla.org/content/running-llms-loc...
https://github.com/mozilla-Ocho/llamafile
https://justine.lol/matmul/
and parallelize them with
https://github.com/b4rtaz/distributed-llama
Obviously this needs some knowledge of the command line, so get a good terminal emulator like
https://iterm2.com/
Mend, bend, rend that stuff and see what works how and why, and what not.
Edit: Optionally, if you really want to go low-level, with some debugger like
https://justine.lol/blinkenlights/
for 'toy-installations' of smallest models.
'Toy' because that doesn't fully support the CPU-instructions which are used in production.
Could still help conceptually.
Thank you for the Mozilla AI guide, it's really useful! I knew some of Justine's work and they are really cool.