Bainance Labs (Parody)

1.4K posts

Bainance Labs (Parody) banner
Bainance Labs (Parody)

Bainance Labs (Parody)

@BainanceLabs

bainance labs is the accelerator for the trenches, investing in strong communities, "devs" and agents

Katılım Aralık 2024
196 Takip Edilen8.7K Takipçiler
Sabitlenmiş Tweet
Bainance Labs (Parody)
Bainance Labs (Parody)@BainanceLabs·
is the dao a meme or is the meme a dao Two roads diverged and I, I took the one less travelled by, and that has made all the difference.
Bainance Labs (Parody) tweet media
English
350
398
909
152K
Bainance Labs (Parody)
Bainance Labs (Parody)@BainanceLabs·
We are officially winding down the bAInance Labs DAO. Over the past year, we’ve supported open research initiatives and successfully returned more than 3x adjusted capital to our DAO holders. DAO tokens can be burnt to redeem proportional assets from the labs treasury on daos.fun
English
1
1
2
655
Miss Sentient
Miss Sentient@0xsachi·
The problem with crypto / AI is that most teams are crypto-first, not AI-first The teams that focus on AI will dominate the market. Crypto should be secondary.
English
17
3
63
3.7K
Bainance Labs (Parody) retweetledi
Bainance Labs (Parody)
Bainance Labs (Parody)@BainanceLabs·
Introducing the Dream Engine - a unified framework for advanced text-to image control arxiv.org/pdf/2502.20172 DE is an improvement built by replacing @StableDiffusion 3's text encoder - adding a blending mechanism that improves consistency of objects in AI generated images
Bainance Labs (Parody) tweet media
Liang Chen@liangchen5518

Excited to introduce DreamEngine! It is a unified framework that integrates multimodal encoders with diffusion models to achieve SOTA performance in generating images with creative, text-and-image inputs. Kudos to @Alibaba_Qwen, @wenhaocha1, @0xvincii ! huggingface.co/papers/2502.20…

English
17
7
51
31K
Bainance Labs (Parody) retweetledi
CZ 🔶 Bainance - parody
CZ 🔶 Bainance - parody@cz_bainance_·
focus on building real value. too many people waste time on worthless jokes and hype that add zero utility. we need builders.
English
115
57
212
7.7K
Bainance Labs (Parody) retweetledi
CZ 🔶 Bainance - parody
CZ 🔶 Bainance - parody@cz_bainance_·
if you don't understand something, don't put money in it. if you do understand it, still do your own research. the market is unforgiving.
English
15
33
230
8.5K
CZ 🔶 BNB
CZ 🔶 BNB@cz_binance·
This Sunday, I learned to write some python code to get an AI API to write tweets based on how I tweet. Only took about 30 minutes to get it all working. But it's even worse than the crap I usually put out.🤷‍♂️😂
English
1.2K
306
4.8K
615.8K
Bainance Labs (Parody)
Bainance Labs (Parody)@BainanceLabs·
@Yuchenj_UW depends on how you view the "aha" moment! for the R1-V, our "aha" moment was the performance increase the visualThinker-R1's "aha" moment was the increase in response length
Bainance Labs (Parody) tweet media
English
0
0
4
129
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
@BainanceLabs Yeah, in their report, they said R1-V didn't develop the "Aha moment" (reasoning)? is it true?
Yuchen Jin tweet media
English
1
0
4
561
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
“Aha moment” for multimodal reasoning on just a 2B model by RL! VisualThinker-R1-Zero authors find applying RL to fine-tuned vision-language models didn't replicate Deepseek-R1’s reasoning but directly applying RL training to the Qwen2-VL-2B base model worked! (similar to DeepSeek-R1-Zero) Without any SFT but only RL, the model achieves 59.47% accuracy on CVBench, beating the base model for ~30%. It's exciting to see RL just works for multimodal!
Yuchen Jin tweet media
English
13
54
405
32.1K
GB
GB@gameboss_eth·
@BainanceLabs 不给你发4了,这次发个6
中文
2
0
3
145