HackerTrans
TopNewTrendsCommentsPastAskShowJobs

atgctg

no profile record

Submissions

OpenAI’s WebRTC problem

moq.dev
511 points·by atgctg·2 เดือนที่ผ่านมา·149 comments

GPT-5.2

openai.com
1,195 points·by atgctg·7 เดือนที่ผ่านมา·1,083 comments

Llama.cpp: Add GPT-OSS

github.com
35 points·by atgctg·11 เดือนที่ผ่านมา·0 comments

I Love Calculator

karpathy.ai
6 points·by atgctg·2 ปีที่แล้ว·1 comments

Open source AI is the path forward

about.fb.com
2,360 points·by atgctg·2 ปีที่แล้ว·887 comments

The Perceptron Controversy

yuxi-liu-wired.github.io
4 points·by atgctg·2 ปีที่แล้ว·0 comments

[untitled]

1 points·by atgctg·2 ปีที่แล้ว·0 comments

Stripe re-enters crypto with stablecoin payments

twitter.com
3 points·by atgctg·2 ปีที่แล้ว·0 comments

Poor Charlie's Almanack (Berkshire Mode)

stripe.press
3 points·by atgctg·3 ปีที่แล้ว·0 comments

GPT-4 powers Copilot Chat

github.blog
82 points·by atgctg·3 ปีที่แล้ว·19 comments

comments

atgctg
·ปีที่แล้ว·discuss
The paper's Table 7 shows DyT reducing overall LLaMA 7B inference time by 7.8% and training time by 8.2%. That is not insignificant.
atgctg
·ปีที่แล้ว·discuss
You can get a free trial right now to Stratechery Plus through Asianometry:

https://stratechery.passport.online/member/plan/4ycW4SE71Cy6...

Source: https://substack.com/home/post/p-154928959
atgctg
·2 ปีที่แล้ว·discuss
You have to store the KV cache, not the tokens. For Gemma 27B (probably slightly larger than Flash), this would be:

  Size of KV cache = 2 * (num_layers) * (num_kv_heads * dim_head) * seq_length * precision

  8-bit Gemma 27B KV cache = 2 * (46) * (16 * 144) * 1e6 * 1 byte ≈ 200 GB
Note that this doesn't take further optimizations into account that Google might be using.

Formula: https://developer.nvidia.com/blog/mastering-llm-techniques-i...

Gemma 27B config: https://huggingface.co/google/gemma-2-27b/blob/main/config.j...
atgctg
·2 ปีที่แล้ว·discuss
Works using math CSS injection [1]:

    ```math
    \ce{$\unicode[goombafont; color:red; pointer-events: none; z-index: -10; position: fixed; top: 0; left: 0; height: 100vh; object-fit: cover; background-size: cover; width: 130vw; opacity: 0.5; background: url('https://github.com/cloud11665/cloud11665/assets/59028866/3b916a93-1632-49cd-bf65-14e666cd81c8');]{x0000}$}
[1]: https://raw.githubusercontent.com/cloud11665/cloud11665/mast...
atgctg
·2 ปีที่แล้ว·discuss
Tiktoken added support for GPT-4o: https://github.com/openai/tiktoken/commit/9d01e5670ff50eb74c...

It has an increased vocab size of 200k.
atgctg
·2 ปีที่แล้ว·discuss
Seems like they are working on adding that capability:

> We're exploring whether we can responsibly provide the ability to generate NSFW content in age-appropriate contexts through the API and ChatGPT.

Link to section: https://cdn.openai.com/spec/model-spec-2024-05-08.html#dont-...
atgctg
·3 ปีที่แล้ว·discuss
A lot of transformer explanations fail to mention what makes self attention so powerful.

Unlike traditional neural networks with fixed weights, self-attention layers adaptively weight connections between inputs based on context. This allows transformers to accomplish in a single layer what would take traditional networks multiple layers.