https://github-production-user-asset-6210df.s3.amazonaws.com... demo shows a huge love for water, this AI knows its home
Also, very influenced by the literature of Jenkins (2010).
BitNet: Inference framework for 1-bit LLMs
11–20 of 179 posts
Re: BitNet: Inference framework for 1-bit LLMs
#12The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on comm…
Yes. I had to read it over twice, it does strike me as odd that there wasn't a base model to work with.
But it seems the biggest model available is 10B? Somewhat unusual and does make me wonder just how challenging it will be to train any model in the 100B order of magnitude.
Re: BitNet: Inference framework for 1-bit LLMs
#13> bitnet.cpp is the official inference framework for 1-bit LLMs (e.g., BitNet b1.58). It offers a suite of optimized kernels, that support fast and lossless inference of 1.58-bit models on CPU and GPU (NPU support will coming next). One bit or one trit? I am confused!
Yeah, "1.58 bit" is 1 trit with three states, since log2(3)≈1.58.
So it's not a inference framework for 1-bit models (two states per parameter) but for 1.58 bit models (three states per parameter). Annoying that they try to mix up the two.
Re: BitNet: Inference framework for 1-bit LLMs
#14One of the things I often wonder is "what will be the minimally viable LLM" that can work from just enough information that if it googles the rest it can provide reasonable answers? I'm surprised something like Encyclopedia Britanica hasn't yet (afaik) tried to capitalize on AI by selling their data to LLMs and validating outputs for LLM companies, it would make a night and day difference in some areas I would think. Wikipedia is nice, but there's so much room for human error and bias there.
Re: BitNet: Inference framework for 1-bit LLMs
#15The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on comm…
I've also always though that it's an interesting opportunity for custom hardware. Two bit addition is incredibly cheap in hardware, especially compared to anything involving floating point. You could make huge vector instructions on the cheap, then connect it to the fastest memory you can buy, and you have a capable inference chip.
You'd still need full GPUs for training, but for inference the hardware would be orders of magnitude simpler than what Nvidia is making
Re: BitNet: Inference framework for 1-bit LLMs
#16The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on comm…
> Framework is ready. Now we need someone to actually train the model.
If Microslop aren't gonna train the model themselves to prove their own thesis, why would others? They've had 2 years (I think?) to prove BitNet in at least some way, are you really saying they haven't tried so far?
Personally that makes it slightly worrisome to just take what they say at face value, why wouldn't they train and publish a model themselves if this actually led to worthwhile results?
Re: BitNet: Inference framework for 1-bit LLMs
#17The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on comm…
LLM account
Re: BitNet: Inference framework for 1-bit LLMs
#18One of the things I often wonder is "what will be the minimally viable LLM" that can work from just enough information that if it googles the rest it can provide reasonable answers? I'm surprised something like Encyclopedia Britanica hasn't yet (afaik) tried to capitalize on AI by selling their data to LLMs and validating outputs for LLM companies, it would make a night and day difference in some areas I would think.…
Your worry about Wikipedia is that there is "much room for human error and bias", yet earlier you seem to imply that a LLM that has access to the www somehow would have less human error and bias? Personally, I'd see it the other way around.
Re: BitNet: Inference framework for 1-bit LLMs
#19One of the things I often wonder is "what will be the minimally viable LLM" that can work from just enough information that if it googles the rest it can provide reasonable answers? I'm surprised something like Encyclopedia Britanica hasn't yet (afaik) tried to capitalize on AI by selling their data to LLMs and validating outputs for LLM companies, it would make a night and day difference in some areas I would think.…
> validating outputs for LLM companies
How? They can validate thousands if not millions of queries but nothing prevent the millions-th-and-one from being a hallucination. People who would then pay extra for a "Encyclopedia Britanica validated LLM" would then, rightfully so IMHO, complain that "it" suggested them to cook with a dangerous mushroom.
Re: BitNet: Inference framework for 1-bit LLMs
#20The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on comm…
>. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck.
I imagine you got 96gb because you thought you'd be running models locally? Did you not know the phrase Unified Memory is marketing speak?