search

StarkWizard / cairo-llm

a set of notebooks for training llms on cairo
27 stars 8 forks source link

More Information on LLM usecase (proofs) #21

Closed RawthiL closed 4 months ago

RawthiL commented 4 months ago

Hi, I just found your repo, and I'm new to StarkNet, I'm confused about the objective of this repository. I see that i contains code to finetune and quantize LLM models in traditional quantization schemes, but I'm having trouble understanding the connection with STARK profs. You state: "we aim to provide a robust and reliable LLM for all your Cairo smart contract needs." does this mean that you are generating proofs of LLM execution?

Thanks!

pechaut78 commented 4 months ago

Hi Ramiro,

The goal is to make an assistant on the Staknet echosystem. There is already a Llama and Mistral fine tuned that can answer some questions about Cairo and generate some code

Cordialement, CHAUT Pierre-Emmanuel

De : Ramiro R. C. @.> Envoyé : lundi 8 juillet 2024 21:46 À : StarkWizard/cairo-llm @.> Cc : Subscribed @.***> Objet : [StarkWizard/cairo-llm] More Information on LLM usecase (proofs) (Issue #21)

Hi, I just found your repo, and I'm new to StarkNet, I'm confused about the objective of this repository. I see that i contains code to finetune and quantize LLM models in traditional quantization schemes, but I'm having trouble understanding the connection with STARK profs. You state: "we aim to provide a robust and reliable LLM for all your Cairo smart contract needs." does this mean that you are generating proofs of LLM execution?

Thanks!

— Reply to this email directly, view it on GitHubhttps://github.com/StarkWizard/cairo-llm/issues/21, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAE2VAVKCN774FVOYKCEXT3ZLLUB7AVCNFSM6AAAAABKRO6VNWVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM4TMNBSGQYDAMY. You are receiving this because you are subscribed to this thread.Message ID: @.***>

RawthiL commented 4 months ago

Now I see, thanks for the clarification!