Dheer Patel
@dheerpatel1710
http://M.Sc student in Robotics, Cognition, Intelligence at @TU_Muenchen . Crazy about automobiles, AI, football and really heavy music.
You might like
Bro that’s Rafael and Fabio Da Silva!
It is basically DAGGER but with a key difference: On-policy distill for LLMs typically works best with reverse KL (or Jeffreys which is half forward, half reverse) instead of typical "forward KL" in robotics / RL (log-likelihood or MSE loss using expert action). Interestingly,…
Very late to the party. When Thinking Machines Lab released their blog on On Policy Distillation, my first reaction was that it should be just like DAGGER from 15 years ago: arxiv.org/abs/1011.0686. I finally had time to read the blog today and sure enough, they mentioned DAGGER.…
When GTA 5 released I was 14. Since then I’ve dropped out of school, gotten arrested, smoked crack, gotten arrested again, gone to rehab, started doing blow, got arrested again, got sober, got 1,000,000 subscribers, made a weed brand, crashed my car, made a podcast, crashed my…
Hi everyone, Grand Theft Auto VI will now release on Thursday, November 19, 2026. We are sorry for adding additional time to what we realize has been a long wait, but these extra months will allow us to finish the game with the level of polish you have come to expect and…
Leaving Meta and PyTorch I'm stepping down from PyTorch and leaving Meta on November 17th. tl;dr: Didn't want to be doing PyTorch forever, seemed like the perfect time to transition right after I got back from a long leave and the project built itself around me. Eleven years…
the last thing a model sees before being captured in a cuda graph
Correct take.
Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.
🇩🇰🇪🇬 Men are simple creatures. Egyptian Foreign Minister Badr Abdelaty was gifted a LEGO set of the Great Pyramid of Giza by the Danish Foreign Minister during his visit for the Grand Egyptian Museum’s opening. Look how happy he is.
CONVERT YOUR CODEBASES TO REAL NUMBERS, I REPEAT REAL NUMBERS
PewDiePie in 2025: – built a 10×4090 rig – runs Llama 70B, gpt-oss-120B & Qwen 245B locally via vLLM – built a custom web UI (chat, RAG, search, TTS) – ran protein-folding simulations for charity – created an AI “council”, a swarm of 64 models – now fine-tuning his own model…
p = 0.051
Finally Germany has a serious LB. This new Germany generation looks very promising, many insane 17-22 year olds. Germany will hit its prime when the 1995 generation is gone and the Musiala generation is like 27 years old
Guido van Rossum builds a python package for RAG
At #PyBay25, @gvanrossum demo'd a Python package for "structured RAG". During ingestion, it uses LLM to extract structured data (entities/topics/verbs) and stores in standard DB, and then retrieves by structuring the user query as well. Try it out at: github.com/microsoft/type…
paper rejection
Chelsea LOSE Liverpool LOSE Man City LOSE Man Utd WIN What a beautiful weekend of football. 😂
Excited and honored to welcome @goodside to Google DeepMind and the AI Studio team as our first staff prompt engineer : )
Nice, short post illustrating how simple text (discrete) diffusion can be. Diffusion (i.e. parallel, iterated denoising, top) is the pervasive generative paradigm in image/video, but autoregression (i.e. go left to right bottom) is the dominant paradigm in text. For audio I've…
BERT is just a Single Text Diffusion Step! (1/n) When I first read about language diffusion models, I was surprised to find that their training objective was just a generalization of masked language modeling (MLM), something we’ve been doing since BERT from 2018. The first…
study graph theory the entire universe runs on connections. neurons, roads, networks, friendships, computers, logistics, even ideas; all are graphs. graph theory is the math of relationships. it teaches you how things interact, not just what they are. you’ll start seeing…
United States Trends
- 1. Lakers 50.7K posts
- 2. #AEWDynamite 45.8K posts
- 3. Epstein 1.5M posts
- 4. Jokic 16.3K posts
- 5. Shai 15K posts
- 6. #AEWBloodAndGuts 5,670 posts
- 7. #Survivor49 3,685 posts
- 8. Darby 5,439 posts
- 9. Thunder 42.1K posts
- 10. Kyle O'Reilly 1,863 posts
- 11. Steph 26.1K posts
- 12. Rory 7,406 posts
- 13. Kobe Sanders N/A
- 14. Moxley 2,914 posts
- 15. Spencer Knight N/A
- 16. Hobbs 28.7K posts
- 17. Blood & Guts 25.2K posts
- 18. Caruso 4,042 posts
- 19. #SistasOnBET 2,270 posts
- 20. Warriors 50K posts
You might like
Something went wrong.
Something went wrong.