Top Tweets for #RLlib

almost 3 years ago

Announcing Ray 2.5 release features: 👉 Support #LLMs training with Ray Train 👉 Serve #LLMs with Ray Serve 👉 Multi-GPU learner stack in #RLlib for cost efficiency & scalable RL-agent training 👉 Performant & improved approach to batch inference at scale https://t.co/AfhvFzdUHm

1

25

9

2

3K

kourosh hakhamaneshi

@CyrusHakha

about 3 years ago

We listened to our users and are bringing #DreamerV3's implementation to #RLlib. On our new RLModule / Learner API. A very exciting algorithm indeed. Takes the pain of hyper-paremeter tuning away. Early signs are super promising. Stay tuned for more info.

0

3

0

327

over 3 years ago

Last week, we released Ray 2.3. ICYMI: ⭐️ Observability enhancements ⭐️ Ray Dataset Streaming ⭐️ Boost in Ray core performance ⭐️Gym/Gymnasium library in #RLlib ⭐️ Support ARM & Python 3.11 ⭐️ Support multiple applications in Ray Serve https://t.co/55tRgBCGGf

0

19

1

2

756

kourosh hakhamaneshi

@CyrusHakha

over 3 years ago

We are happy to announce that for #RLlib 2.3 we have moved away from the old @OpenAI gym APIs to @FaramaFound 's gymnasium API. Here is what you need to know to about this transition:

1

28

7

2

3K

over 3 years ago

Ray 2.3.0 Released with: ⭐️ Observability enhancements ⭐️ Ray Dataset Streaming ⭐️Boost in Ray core performance ⭐️Gym/Gymnasium library in #RLlib ⭐️ Support ARM & Python 3.11 ⭐️ Support multiple applications in Ray Serve (developer preview) https://t.co/55tRgBC8QH

0

23

3

0

1K

over 3 years ago

Spot on for @raydistributed #RLlib

1

4

0

372

over 3 years ago

Now. Jungong and @CyrusHakha from @anyscalecompute #RLlib tutorial @odsc #ODSCWest in-person at SFO.

0

4

2

0

over 3 years ago

Excited to announce JunGong and @CyrusHakha will be giving a hands-on workshop using @anyscalecompute @raydistributed #RLlib. Thursday Nov 3, 11am PT @odsc #ODSCWest. 👋🏽 Hope to see you there! Pro-tips in my post below. https://t.co/uRUAlGdEa8

0

6

1

0

over 3 years ago

@anyscalecompute @raydistributed team draws a huge crowd at @ACMRecSys hands-on tutorial on #RL & #RLlib

jina

@jinayoon_

over 3 years ago

WOW the tutorial room for the Hands-on Reinforcement Learning workshop is packed at @ACMRecSys 😱 @CyrusHakha @cbergman @anyscalecompute this is a popular one! #RecSys2022

jinayoon_'s tweet photo. WOW the tutorial room for the Hands-on Reinforcement Learning workshop is packed at @ACMRecSys 😱 @CyrusHakha @cbergman @anyscalecompute this is a popular one! #RecSys2022 https://t.co/d4KHu5dosf

0

38

9

1

0

2

0

over 3 years ago

Thank you everyone for attending our tutorial by @CyrusHakha @narayan_avnish and myself! Hope you were able to learn more about @raydistributed @anyscalecompute #RLlib, please let me know if I can help!

jina

@jinayoon_

over 3 years ago

WOW the tutorial room for the Hands-on Reinforcement Learning workshop is packed at @ACMRecSys 😱 @CyrusHakha @cbergman @anyscalecompute this is a popular one! #RecSys2022

0

38

9

1

0

9

1

0

Ray Summit @Ray_Summit_Live

almost 4 years ago

Nice interview with @sven_mika , fearless leader/committer of #RLlib, about history of @raydistributed, built for RLlib! Supports both TF1/TF2 and PyTorch, runs distributed on any compute (single machine or cloud cluster), many algorithms model-based, model-free, online, offline

TalkRL Podcast

@TalkRLPodcast

almost 4 years ago

Episode 37 @sven_mika of @anyscalecompute on RLlib present and future, Ray and Ray Summit 2022, applied RL in games / finance / recsys, and more! https://t.co/oNgT2Wkxr0

1

18

3

4

0

7

0

almost 4 years ago

With Ray 2.0, our goal is to make #RLlib even more intuitive, easy to extend, and performant for various research and production use cases. Check out what we have in store at #RaySummit: https://t.co/0ENS8cB9kV

Ray_Summit_Live's tweet photo. With Ray 2.0, our goal is to make #RLlib even more intuitive, easy to extend, and performant for various research and production use cases.

Check out what we have in store at #RaySummit: https://t.co/0ENS8cB9kV https://t.co/zdZvzMVqhT

0

4

0

1

0

Ray Summit @Ray_Summit_Live

almost 4 years ago

Get hands-on with us next week at #RaySummit! There’s still time to sign up for our #RLlib training on Aug 22 and create a multiplayer game using RLlib! https://t.co/brn0FWEE3b

0

3

0

Anyscale

@anyscalecompute

almost 4 years ago

Latest #RLlib office hour now live! @CyrusHakha, Jun, @sven_mika answer questions about multi-agent #ReinforcementLearning for meat #SupplyChain! Also questions about future RLlib API @raydistributed AI Runtime #RayAIR @Ray_Summit_Live (Aug22-24) in SF! https://t.co/IiuxSs30i3

0

4

1

0

almost 4 years ago

Did you miss this week's #RLlib office hour? Here's what we covered! 💥 How to train a SAC agent in contentious Action/Obs space 💥 Help debugging getting tensor values for metrics https://t.co/cZAaRxp92e

0

2

0

almost 4 years ago

👋Everyone welcome next #RLlib office hours! @sven_mika hosting Tues July5 08:00PT. Also check out previous videos of the team answering users questions from @CyrusHakha, Artur, @narayan_avnish https://t.co/aH8w0dWhcM

almost 4 years ago

The next #RLlib office hour is coming up next week on July 5! Add your #ReinforcementLearning questions for discussion here: https://t.co/cgFmtRYGgs

0

1

0

1

0

almost 4 years ago

ICYMI @raydistributed Community meetup, here're the recordings: 👉 AI/ML-Infused Digital IC Design Workflows on the Hybrid Cloud, @IBMResearch 👉 Pushing boundaries of material design with #RLlib @Stanford https://t.co/pAkM1jRbtO

0

3

1

0