@JustinLin610

building p7k @pragmatik_labs | ❤️ 🍵 ☕️

Joined December 2015
  • Tweets3.2K
  • Following2.1K
  • Followers94.9K
167Media
most cases i saw for argon are about visualization. how does it perform on other tasks?
19
1
100
28,913
nyc style
4
90
33,300
it is always super hard to find a good angle to take a satisfactory picture. just step back and blow up like 3x or even 5x, u will find that the picture looks much better
3
13
22,009
just checked hf daily papers sept ranking, dpsk v4.1 flash only has 191 upvotes? ridiculous, it is the greatest paper of sept
25
19
3
581
87,320
i mean, the ranking is corrupted and causes me much trouble filtering papers
3
33
24,102
Junyang Lin retweeted
Kaiyue Wen
@wen_kaiyueOct 2
Do you really need to finish every rollout in RL for LLM? Not if you make your critic better—and trust it! We propose Actor-Critic with Action Chunking (AC2): use a learned critic to score chunks of tokens => only need partial rollouts ⇒ train faster than GRPO!
26
101
11
754
101,804
lol
Sebastian Raschka
@rasbtSep 29
Replying to @rasbt
Update: OpenAI just added a Jev clone 😆
1
1
108
31,432
hope smarter than opus 5.5
OpenAI
@OpenAISep 29
GPT-6.1 Sol: near-Astra intelligence for a fifth of the price. It’s the most cost-efficient model for its performance available today.
13
1
128
24,377
who the hell r u
17
1
247
49,880
u guys r drinking too much! (this is not an ad for 农夫山泉
16
1
109
21,919
this is an espresso?
9
46
14,277
wen we try to accelerate u tell me to slow down? omg...
58
50
5
1,025
70,454
u can do whatev u want. stop pretending u need any permissions. funny
2
11
193
11,797
cuz people can suddenly write rust now (lol...
Alex Cheema
@alexocheemaSep 12
why did everyone suddenly decide specialized inference engines are a good idea? i've seen at least 5 of these released in the last month. what's wrong with vLLM/sgLang? sure you can generate a slop inference engine quickly now but why fragment the ecosystem? it's harmful.
8
6
264
43,352
v4p back? k3-0.2?
14
1
166
23,809
being a staff: the weekend is coming! being a boss: the weekend is coming?
26
2
3
246
33,714
bringing too much complexity into model arch can cause unknown issues that can't be tested by evals even those internal ones. but nvm, sometimes people can take the small unknown costs for big benefits in efficiency as fm are becoming more and more focused on so called agentic tasks
16
15
274
29,530
yoco
15
10
1
205
45,463
plz contact us only through [email protected] and check only the x and li of p7k and me (or if u can see my wechat moments)
9
2
2
110
27,698
the reason why i post this is due to rumors and false info. we are not aggressively hiring. we do not jds. we are only interested in mts at this moment.
2
18
7,358
Load more