serialx

10 ms·

1.
Moonshot AI suspends new subscriptions due to Kimi K3 demand(twitter.com)
284 points by serialx 2mo ago|114 comments
2.
Beeg float library, a Rust port of Fabrice Bellard's libbf(github.com)
19 points by serialx 3mo ago|12 comments
3.
All-smi: Command-line utility for monitoring GPU/TPU/NPU hardwares(github.com)
4 points by serialx 6mo ago|0 comments
4.
by serialx 6mo ago
Look into all-smi https://github.com/lablup/all-smi It supports all GPUs thinkable including Apple Silicon and many AI accelerator cards.
5.
Gild Just One Lily(smashingmagazine.com)
38 points by serialx 9mo ago|7 comments
6.
FSM for Python, Inspired by Gen_fsm(github.com)
2 points by serialx 1y ago|1 comments
7.
by serialx 1y ago
Yeah, attention sinks were applied to gpt-oss
8.
Unsloth Dynamic 2.0 GGUFs(docs.unsloth.ai)
3 points by serialx 1y ago|0 comments
9.
FuriosaAI RNGD – Tensor Contraction Processor(developer.furiosa.ai)
4 points by serialx 1y ago|0 comments
10.
by serialx 2y ago
Ah sorry, you might be right. I meant "sparse reward" as a reward system that is mostly 0 but occasionally 1. Your "sparse reward" means only providing reward at the end of each output.
11.
by serialx 2y ago
I don't think it's only using sparse rewards because of the format rewards. The training recipe is pretty comprehensive and involves multiple stages.[1] The paper mentions that when only using the RL technique, the output is often
12.
by serialx 2y ago
So to my understanding, this work reproduces DeepSeek R1's reinforcement learning mechanism in a very small language model. The AI gets "rewards" (like points) for doing two things correctly: Accuracy : Getting the right answ
13.
South Korea's impeached president is arrested over a martial law declaration(apnews.com)
21 points by serialx 2y ago|0 comments
14.
PyMC-Marketing – Open-Source Marketing Analytics Solution(pymc-marketing.io)
1 points by serialx 2y ago|0 comments
15.
by serialx 2y ago
Change the currency to USD
16.
by serialx 2y ago
What are the compute requirements for solving efficient energy grid transmission? Is there a efficient algorithm that is able to solve this?
17.
by serialx 2y ago
Is there any plans to add Well-known URI[1] as a standard? It would be awesome if we can add services just by inputting domain names of the services. [1]: https://en.wikipedia.org/wiki/Well-known_URI
18.
Parse JSON Stream Fast for Realtime UI in LLM AI App(partial.stream)
1 points by serialx 2y ago|0 comments
19.
by serialx 2y ago
PSA: You can also use singleflight[1] to solve the problem. This prevents the thundering herd problem. Pocache is an interesting/alternative way to solve thundering herd indeed! [1]: https://pkg.go.dev/golang.org/x
20.
by serialx 2y ago
Actually, llama.cpp running on Apple silicon uses GPU(Metal Compute Shader) to inference LLM models. Token generation is also very memory bandwidth bottlenecked. On high end Apple silicon it's about 400MB/s to 800MB/s, compar
21.
by serialx 2y ago
Well, iPhone already does that with photos. :)
22.
Software engineering salaries come from one of three budgets(swizec.com)
536 points by serialx 3y ago|257 comments
23.
Graviton 3 Performance with CockroachDB Consistent, Efficient, Scalable SQL(glennfawcett.wpcomstaging.com)
3 points by serialx 3y ago|0 comments
24.
by serialx 3y ago
Now we just need VkDNN
25.
Science Before Statistics: Causal Inference(youtube.com)
3 points by serialx 3y ago|0 comments
26.
by serialx 3y ago
Interesting! Maybe using small cells like these might make the frame and cell blend together? https://a.aliexpress.com/_oDCStIL
27.
by serialx 3y ago
No since it’s stateful in the sense that inferencing is dependent on the past generated tokens.
28.
AWS Announces Amazon Verified Permissions (Preview)(aws.amazon.com)
1 points by serialx 4y ago|0 comments
29.
Cronet: Chromium's networking stack for your iOS applications(github.com)
2 points by serialx 5y ago|0 comments
30.
Apple rejects games using Corona Game Engine(forums.solar2d.com)
5 points by serialx 6y ago|0 comments

More ›