Ivan Goncharov

2.1K posts

Ivan Goncharov banner
Ivan Goncharov

Ivan Goncharov

@Ivangrov

Growth Machine Learning Engineer @weights_biases by Day // Making YouTube Videos about AI by Night // Playing the Fiddle & Guitar by Early Mornings

Tbilisi, Georgia Katılım Şubat 2017
620 Takip Edilen1.3K Takipçiler
Sabitlenmiş Tweet
Ivan Goncharov
Ivan Goncharov@Ivangrov·
In a new exciting video I team up with @carlolepelaars and we use #ChatGPT to generate the best "walks into a bar..." joke using prompt engineering.🍻🔥😅 The results range from hilarious😂 to absurdist🙃 to... outright creepy😱 Tune in here👇 🎥Video: youtu.be/F19OV2SLh8I
YouTube video
YouTube
English
0
2
8
2.5K
Ivan Goncharov
Ivan Goncharov@Ivangrov·
Editing videos while listening to good music is a really nice experience
English
0
0
2
488
Adam.GPT
Adam.GPT@TheRealAdamG·
@DataChaz Clippy had to walk so ChatGPT could run
English
2
3
58
5.2K
Scott Condron
Scott Condron@_ScottCondron·
The algorithm supervisors in W&B: “Wow, batch_size is having a good day today!”
English
1
0
2
301
Scott Condron
Scott Condron@_ScottCondron·
Upcoming: W&B will deprecate its hyperparameter tuning feature, Sweeps, and just crowdsource your configs by checking what’s correlating with accuracy (or your metric of choice) across users each day (sarcasm in case that’s not clear)
Zack Witten@zswitten

@wandb @l2k Let’s correlate learning rate with accuracy and finally wrap up this whole hyperparameter tuning controversy

English
1
0
4
833
Ivan Goncharov
Ivan Goncharov@Ivangrov·
Yeah, when playing with ChatGPT couldn’t get out of my head that all of it was possible with GPT-3 + some prompt engineering. (Except for the fact that they used a newer GPT-3.5 variant)
Justin Fineberg@JustinFineberg

The GPT playground didn’t go viral, chatGPT did. The UX is what allowed the masses to understand the power of AI. Goes to show that even when the underlying technology is the same, the best AI products will be the ones that are overly-focused on the user experience.

English
0
0
1
302
Ivan Goncharov retweetledi
Lavanya
Lavanya@lavanyaai·
If y'all are feeling the new year burst of inspiration like me and looking to learn a new skill...💃 I high recommend taking our new MLOps course, and learning all the gory details behind putting your ML models into production. 👩‍🏫 wandb.me/mlops-course youtube.com/watch?v=ZJtkOX…
YouTube video
YouTube
English
0
6
37
5.9K
Ivan Goncharov
Ivan Goncharov@Ivangrov·
Happy new year everyone!!
English
1
0
5
1.1K
Ivan Goncharov
Ivan Goncharov@Ivangrov·
@BEBischof @capetorch GPT-4 (DaVinci), GPT-4 (Curi), GPT-4 (tweets), GPT-4 (hair) and GPT-4 (spouse) - surely variations of that model can do the job
English
1
0
0
55
Thomas Capelle
Thomas Capelle@capetorch·
Brace yourself, GPT-4 will beat XGBoost in tabular data. In fact, will solve the Kaggle competition. Stay tuned for more!
English
3
5
78
13.4K
Ivan Goncharov retweetledi
Carlo Lepelaars
Carlo Lepelaars@carlolepelaars·
blackscholes v0.1.0 is out! In this version I introduce 4 well-known option structures. 1. Straddle 2. Strangle 3. Butterfly 4. Iron Condor Below an example of how you would retrieve attributes for a "Long Iron Condor", which consists of 2 call options and 2 put options. (1/3)
Carlo Lepelaars tweet media
English
4
1
11
1.1K
Darek Kłeczek
Darek Kłeczek@dk21·
My daughter will make it in politics. She started as supporter of France, then switched to Argentina, then back to France again. She kept fingers crossed for France in penalties. Then, after a moment of sadness... MESSI!!!!!
English
2
0
21
1.7K

Keşfet

@wandb @metaphdor @TheRealAdamG @DataChaz @_ScottCondron @Mukilan_Krish @elonmusk @BarackObama