Lit-LLaMA
AI
llm implementation
SHOW HN2023
How HN took it
2 positive, 8 neutral or asking, 1 mixed, 9 critical, out of 20 comments about the project. Thread score -0.34.
Idea: 1 for, 8 againstPerformance: 1 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Apache 2.0 licensing and ability to run on 8 GB VRAM
From 1 comment
I am in love with this implementation considering the ability to run on 8 GB VRAM and Apache 2.0 license.
theaniketmaurya on HN
Seen as a win for the AI community
From 1 comment
I see this as a win for the AI community. The key for LLMs is to enable people to train collaboratively and innovate more quickly in this space. Are there any examples or demos…
yewnork on HN
Objections
Underlying LLaMA model weights remain research-only, not commercially licensed
From 3 comments
I think OP mischaracterized the issue with the license, its more that the weights don’t fall under the same scope. They’re research use only, no commercial use allowed.
cuuupid on HNThere are already a million ways to run LLaMA. This doesn't change the issue at all, which is that the weights aren't commercially licensed.
nynx on HNBut aren’t the weights still not for commercial use?
barefeg on HN
Post mischaracterizes GPL as preventing open source or academic use
From 3 comments
I think implying that GPL is not "fully open source" is a hot take. It's specifically designed to ensure you and anyone you distribute your code gets the same freedoms. Maybe you…
adeon on HNBs. Prevents meaningful academic..... How the hell does agpl prevent academic use? Commercial use sure because agpl follows 4 freedoms and commercial often wants to take someone…
2Gkashmiri on HN>GPL...prevents meaningful academic and commercial use WTF are you talking about?
alexb_ on HN
Code may be a derivative work of the original GPL code
From 1 comment
IANAL, but this seems very fishy to me: 1) I don't understand how this isn't a derivative work of the original code, as I very highly doubt you've done a clean room…
querez on HN
Suspected coordinated promotion by Lightning AI staff
From 1 comment
FYI, there's something fishy going on in this thread. Multiple people from the LightningAI team theaniketmaurya (developer advocate for Lightning AI) and rasbt (developer at…
ipsum2 on HN
Requests
Examples or demos of model capabilities
From 1 comment
I see this as a win for the AI community. The key for LLMs is to enable people to train collaboratively and innovate more quickly in this space. Are there any examples or demos…
yewnork on HN
Clarity on how weights would be obtained or shared
From 2 comments
I am curious though how would the model weights work out?
theaniketmaurya on HNI guess that means time to fire up a few GPUs later today and get some weights! We should have a weight exchange platform for that maybe, haha.
rasbt on HN
Questions
Whether the weights are covered by the original Facebook license
From 2 comments
IANAL, but this seems very fishy to me: 1) I don't understand how this isn't a derivative work of the original code, as I very highly doubt you've done a clean room…
querez on HN1) I've looked at both codebases and this one is definitely a derivative of the nanoGPT. You can compare all three implementations yourself as they are actually surprisingly…
rnosov on HN
Whether the implementation is a clean-room rewrite or derivative
From 2 comments
IANAL, but this seems very fishy to me: 1) I don't understand how this isn't a derivative work of the original code, as I very highly doubt you've done a clean room…
querez on HNCorrect, the original is GPL 3. To produce this implementation from the LLaMA paper we started from github.com/karpathy/nanoGPT, the LLaMA architecture is really similar to GPT.…
lantiga on HN
Alternatives named
llama.cpp (MIT licensed)
From 1 comment
llama.cpp is also MIT https://github.com/ggerganov/llama.cpp previously discussed here https://news.ycombinator.com/item?id=35100086 and one of the rust wrapper:…
homarp on HN
alpaca.cpp and alpaca-lora implementations
From 1 comment
Yeah, this is weird, because there are plenty of open source implementations of the lama model on github - alpaca.cpp in c++ is one, there are many others in PyTorch such as the…
UncleOxidant on HN
ChatLLaMA
From 1 comment
Yeah, this is weird, because there are plenty of open source implementations of the lama model on github - alpaca.cpp in c++ is one, there are many others in PyTorch such as the…
UncleOxidant on HN
Petals distributed inference
From 1 comment
Right, so sort of like https://github.com/bigscience-workshop/petals but for the training phase. I suppose different training runs could be proposed via a RFC type of procedure.…
barefeg on HN