galeos

6 ms·

1.
by galeos 1mo ago
Is there scope to implement ternary models using this approach to minimise die area of the model parameters?
2.
1.58bit LLM Optimised Tensor Core(github.com)
2 points by galeos 1y ago|0 comments
3.
BitNet 1.58bit GPU Inference Kernel(github.com)
2 points by galeos 1y ago|0 comments
4.
by galeos 1y ago
How did you find ModernBERT performance Vs prior BERT models?
5.
Microsoft beat H200 Deepseek inference with MI300(techcommunity.microsoft.com)
3 points by galeos 1y ago|1 comments
6.
Modular's CUDA alternative is ready(eetimes.com)
2 points by galeos 1y ago|0 comments
7.
by galeos 1y ago
You can try out the model in a demo they have setup: https://bitnet-demo.azurewebsites.net/
8.
BitNet b1.58 2B4T Technical Report(arxiv.org)
111 points by galeos 1y ago|30 comments
9.
Microsoft BitNet 1.58bit LLM 2B4T released(huggingface.co)
4 points by galeos 1y ago|1 comments
10.
Mi300 Huggingface(huggingface.co)
2 points by galeos 1y ago|0 comments
11.
Bitnet.cpp: Efficient Inference for 1.58bit LLMs(arxiv.org)
1 points by galeos 2y ago|0 comments
12.
Matryoshka Quantization(arxiv.org)
2 points by galeos 2y ago|0 comments
13.
1-Bit AI Infrastructure(arxiv.org)
157 points by galeos 2y ago|30 comments
14.
by galeos 2y ago
My understanding is that BERT can still outperform LLMs for sentiment classification?
15.
Microsoft BitNet: inference framework for 1-bit LLMs(github.com)
173 points by galeos 2y ago|33 comments
16.
Apollo to offer Intel multibillion-dollar investment(bloomberg.com)
5 points by galeos 2y ago|3 comments
17.
Fine-Tuning LLMs to 1.58bit(huggingface.co)
52 points by galeos 2y ago|3 comments
18.
Evidence of dark oxygen production at the abyssal seafloor(nature.com)
9 points by galeos 2y ago|0 comments
19.
Lamini Memory Tuning: 10x Fewer Hallucinations(lamini.ai)
128 points by galeos 2y ago|57 comments
20.
by galeos 2y ago
These are MLPerf training results. I think current ternary quantization research is focused more on speeding up inference?
21.
Computer Architecture – No Starch Press(nostarch.com)
2 points by galeos 2y ago|0 comments
22.
Extracting Concepts from LLMs: Anthropic's recent discoveries(huggingface.co)
4 points by galeos 2y ago|0 comments
23.
AMD Announces Instinct MI325X Today, CDNA4 to Come(morethanmoore.substack.com)
2 points by galeos 2y ago|0 comments
24.
AMD and Intel Team Up for Open Alternative to Nvidia's NVLink(phoronix.com)
38 points by galeos 2y ago|6 comments
25.
Azure – ND AMD MI300X v5-series(learn.microsoft.com)
2 points by galeos 2y ago|0 comments
26.
Hugging Face on AMD Instinct MI300 GPU(huggingface.co)
2 points by galeos 2y ago|0 comments
27.
by galeos 2y ago
We were also allowed to borrow and re-shrinkwrap games at the Game store I worked in, in the UK, in 2000. Seemed like official company policy to give us better product knowledge!
28.
by galeos 3y ago
What a clear explanation of what correlation actually is!
29.
by galeos 3y ago
In the UK the tax incentives for Electric cars may be skewing demand towards new Vs secondhand EVs. I can lease a new EV via my employer's salary sacrifice scheme. I can pay my lease payments from my pre-tax income. There is an additio
30.
by galeos 3y ago
Not tested. No issues at all with bread.

More ›