- 230
- 1 755 167
Aleksa Gordić - The AI Epiphany
United Kingdom
Приєднався 22 січ 2020
x-Google DeepMind, x-Microsoft engineer explaining AI. ❤️
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💼 LinkedIn - www.linkedin.com/in/aleksagordic/
🐦 Twitter - gordic_aleksa
👨👩👧👦 Discord - discord.gg/peBrCpheKE
📺 UA-cam - ua-cam.com/users/TheAIEpiphany
📚 Medium - gordicaleksa.medium.com/
💻 GitHub - github.com/gordicaleksa
📢 AI Newsletter - aiepiphany.substack.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️
If my UA-cam videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!
The AI Epiphany - www.patreon.com/theaiepiphany
One-time donation - www.paypal.com/paypalme/theaiepiphany
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💼 LinkedIn - www.linkedin.com/in/aleksagordic/
🐦 Twitter - gordic_aleksa
👨👩👧👦 Discord - discord.gg/peBrCpheKE
📺 UA-cam - ua-cam.com/users/TheAIEpiphany
📚 Medium - gordicaleksa.medium.com/
💻 GitHub - github.com/gordicaleksa
📢 AI Newsletter - aiepiphany.substack.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️
If my UA-cam videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!
The AI Epiphany - www.patreon.com/theaiepiphany
One-time donation - www.paypal.com/paypalme/theaiepiphany
DeepMind's TacticAI: an AI assistant for football tactics | Petar Veličković
Become a Patreon: www.patreon.com/theaiepiphany
👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE
Petar Velickovic joined us to share more on his latest work - TacticAI - an AI assistant for football tactics!
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
TacticAI: an AI assistant for football tactics: arxiv.org/abs/2310.10553
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
⌚️ Timetable:
00:00:00 - 00:00:39 Intro
00:00:39 - 00:02:20 Hyperstack GPUs (sponsored)
00:02:20 - 00:10:15 Why AI for football?
00:10:15 - 00:17:50 DeepMind's collaboration with Liverpool FC
00:17:50 - 00:39:00 TacticAI
00:39:00 - 00:53:45 Quantitative & human evaluation
00:53:45 - 00:58:25 Greatest hits from TacticAI
00:58:25 - 01:05:58 Outro / Q&A
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 SPONSOR
The AI Epiphany - www.patreon.com/theaiepiphany
One-time donation - www.paypal.com/paypalme/theaiepiphany
Huge thank you to these AI Epiphany patreons:
Eli Mahler
Petar Veličković
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💼 LinkedIn - www.linkedin.com/in/aleksagordic/
🐦 Twitter - gordic_aleksa
👨👩👧👦 Discord - discord.gg/peBrCpheKE
📺 UA-cam - ua-cam.com/users/TheAIEpiphany
📚 Medium - gordicaleksa.medium.com/
💻 GitHub - github.com/gordicaleksa
📢 AI Newsletter - aiepiphany.substack.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#tacticai #football #deepmind
👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE
Petar Velickovic joined us to share more on his latest work - TacticAI - an AI assistant for football tactics!
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
TacticAI: an AI assistant for football tactics: arxiv.org/abs/2310.10553
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
⌚️ Timetable:
00:00:00 - 00:00:39 Intro
00:00:39 - 00:02:20 Hyperstack GPUs (sponsored)
00:02:20 - 00:10:15 Why AI for football?
00:10:15 - 00:17:50 DeepMind's collaboration with Liverpool FC
00:17:50 - 00:39:00 TacticAI
00:39:00 - 00:53:45 Quantitative & human evaluation
00:53:45 - 00:58:25 Greatest hits from TacticAI
00:58:25 - 01:05:58 Outro / Q&A
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 SPONSOR
The AI Epiphany - www.patreon.com/theaiepiphany
One-time donation - www.paypal.com/paypalme/theaiepiphany
Huge thank you to these AI Epiphany patreons:
Eli Mahler
Petar Veličković
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💼 LinkedIn - www.linkedin.com/in/aleksagordic/
🐦 Twitter - gordic_aleksa
👨👩👧👦 Discord - discord.gg/peBrCpheKE
📺 UA-cam - ua-cam.com/users/TheAIEpiphany
📚 Medium - gordicaleksa.medium.com/
💻 GitHub - github.com/gordicaleksa
📢 AI Newsletter - aiepiphany.substack.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#tacticai #football #deepmind
Переглядів: 1 612
Відео
Ishan Misra (Meta) - Emu Video Generation
Переглядів 1,1 тис.Місяць тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Ishan joined us to talk about SSL and multimodal data, and in particular about his recent work on Emu Video! ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ imisra.github.io/ Emu Video: arxiv.org/abs/2311.10709 ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 - 00:42 Intro 00:42 - 02:23 Hyperstack GPUs (sponsored) 02...
InstructPix2Pix (w/ OpenAI's Tim Brooks)
Переглядів 1,6 тис.Місяць тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE We had Tim Brooks join me to discuss his PhD work - Instruct Pix2Pix. Tim is also a co-author of the recent Sora work so we might have him soon again. :) ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ www.timothybrooks.com/ ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 - 01:34 Intro 01:34 - 03:21 Hyperstack GPUs ...
GPT-Fast - blazingly fast inference with PyTorch (w/ Horace He)
Переглядів 3,3 тис.Місяць тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Horace He joined us today to talk more about how to make inference fast using just PyTorch native operations! ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ pytorch.org/blog/accelerating-generative-ai-2/ github.com/pytorch-labs/gpt-fast ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 - 00:45 Intro 00:45 - 02:23 Hyp...
How does Groq LPU work? (w/ Head of Silicon Igor Arsovski!)
Переглядів 15 тис.2 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE I invited head of silicon at Groq, Igor Arsovski, to share the nitty-gritty details behind Groq's LPUs! LPU or language processing units showed promising results (understatement of the year) on inference benchmarks compared to NVIDIA GPUs. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00:00 - 0...
Thomas Wolf (HuggingFace) - the case for open-source!
Переглядів 1,3 тис.2 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Talk with Thomas Wolf who was the first developer of the now famous transformers library and the chief science officer of HuggingFace. We talk about their beginnings, open-source, how they organize their teams (given that they're fully remote), and a lot more! Enjoy! :) ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬...
Jeremy Howard - answer.ai, what is wrong with the academia and industry
Переглядів 4,4 тис.2 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Talked with Jeremy about what he's trying to accomplish with answer.ai, what is wrong with the current incentives in the industry & academia, how we're wasting the human potential, etc. :) enjoy! ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ www.answer.ai/ ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 - 00:49 In...
LLaMA 2 w/ Thomas Scialom (LLaMA 2 lead)
Переглядів 2,1 тис.4 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Thomas from Meta joined us to talk about his story, his work on various Meta projects including Galactica, LLaMA 2, BLOOM, etc. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Thomas's G Scholar: scholar.google.fr/citations?user=qou4P-kAAAAJ&hl=en ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 Thomas''s story 12:45...
EleutherAI Pythia w/ Hailey Schoelkopf
Переглядів 1,2 тис.4 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Hailey from EleutherAI joined us to talk about her story, her work on the BLOOM project, Pythia, Llemma, lm harness, and beyond! ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Hailey's G Scholar: scholar.google.com/citations?user=XLahYIYAAAAJ&hl=en ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 Hailey's story 10:5...
Lucas Beyer (Google DeepMind) - Convergence of Vision & Language
Переглядів 4,5 тис.4 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Lucas from G DeepMind joined us to talk about the recent convergence of vision & language and various projects he was a part of (PaLI, ViT, etc.). ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Lucas Beyer website: lucasb.eyer.be/ ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 Lucas's story 10:45 Motivation 26:00 ...
Tanishq Abraham (PhD at 19) - MedARC
Переглядів 3,4 тис.4 місяці тому
Become a Patreon: www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community: discord.gg/peBrCpheKE Tanishq joined us to tell us more about his story & his great work in the medical AI space. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Tanishq's website www.tanishq.ai/ ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 Tanishq's story 12:00 MedARC 25:42 MindEye - reconstructing mental images 51:40 Tanishq's PhD w...
building the best RLHF (TRLX) library w/ Louis Castricato
Переглядів 1 тис.5 місяців тому
❤️ Become The AI Epiphany Patreon ❤️ www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community 👨👩👧👦 discord.gg/peBrCpheKE Louis walks us through his journey of building CarperAI. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ✅ github.com/CarperAI/trlx ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00 A bit about Louis himself 02:20 Start of Carper 07:00 Start of trlx 09:25 Growing Carper & OSS 32:30 Q&A ▬▬▬▬▬▬...
Jarvis for Images! (demo) - run locally, no external APIs
Переглядів 1,7 тис.5 місяців тому
❤️ Become The AI Epiphany Patreon ❤️ www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community 👨👩👧👦 discord.gg/peBrCpheKE I just built Jarvis for images :) all models are fully open-source, all running on my local machine (no external APIs, no OpenAI) there is a ton of optimizations i can introduce to make this even more real-time you can manipulate images using your voice (real-time)...
OpenAI DALL-E 3 with James Betker (1st author)
Переглядів 3,4 тис.6 місяців тому
❤️ Become The AI Epiphany Patreon ❤️ www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community 👨👩👧👦 discord.gg/peBrCpheKE James joined us to share his work on DALL-E 3! James has a super fascinating story. He worked in the security space back at Google. In his free time he open-sourced tortoise TTS which captured the attention of many researchers back in 2021. One thing led to another...
The Vesuvius challenge breakthrough with Luke Farritor
Переглядів 2,1 тис.6 місяців тому
❤️ Become The AI Epiphany Patreon ❤️ www.patreon.com/theaiepiphany 👨👩👧👦 Join our Discord community 👨👩👧👦 discord.gg/peBrCpheKE We discuss the Vesuvius challenge with the first person who read the letters from the carbonized scrolls :) The goal of the Vesuvius challenge is to decipher the ancient text from carbonized scrolls found in Herculaneum in a house that belonged to Julius Cesar's fa...
Will LLMs kill Search? Nils Reimers (director of ML at Cohere) | The AI Epiphany Discord talks
Переглядів 2,6 тис.7 місяців тому
Will LLMs kill Search? Nils Reimers (director of ML at Cohere) | The AI Epiphany Discord talks
MosaicML - Path to 1.300.000.000$ with Davis Blalock (1st employee) | Discord server talks
Переглядів 4,6 тис.9 місяців тому
MosaicML - Path to 1.300.000.000$ with Davis Blalock (1st employee) | Discord server talks
Flash Attention 2.0 with Tri Dao (author)! | Discord server talks
Переглядів 16 тис.9 місяців тому
Flash Attention 2.0 with Tri Dao (author)! | Discord server talks
Introducing Ortus - your YouTube AI buddy | ChatGPT Chrome extension
Переглядів 8 тис.11 місяців тому
Introducing Ortus - your UA-cam AI buddy | ChatGPT Chrome extension
I Left Google DeepMind - it's AI Startup Time!
Переглядів 15 тис.Рік тому
I Left Google DeepMind - it's AI Startup Time!
OpenAI's GPT 4 is out and it's multimodal! What we know so far
Переглядів 9 тис.Рік тому
OpenAI's GPT 4 is out and it's multimodal! What we know so far
Building web apps using Streamlit | Streamlit crash course | MLOps series #2
Переглядів 9 тис.Рік тому
Building web apps using Streamlit | Streamlit crash course | MLOps series #2
What is MLOps and how to get started? | MLOps series
Переглядів 35 тис.Рік тому
What is MLOps and how to get started? | MLOps series
Andrew Huberman transcripts | MLOps series announcement
Переглядів 6 тис.Рік тому
Andrew Huberman transcripts | MLOps series announcement
NeurIPS 2022 - interviews with AI experts and conference walk-through
Переглядів 9 тис.Рік тому
NeurIPS 2022 - interviews with AI experts and conference walk-through
How To Install CUDA, cuDNN, Ubuntu, Miniconda | ML Software Stack | Part 3/3
Переглядів 19 тис.Рік тому
How To Install CUDA, cuDNN, Ubuntu, Miniconda | ML Software Stack | Part 3/3
How to Assemble a Deep Learning Machine - Full Process | Part 2
Переглядів 11 тис.Рік тому
How to Assemble a Deep Learning Machine - Full Process | Part 2
How to Build a Deep Learning Machine - Everything You Need To Know
Переглядів 49 тис.Рік тому
How to Build a Deep Learning Machine - Everything You Need To Know
High Fidelity Neural Audio Compression | Paper & Code Explained
Переглядів 7 тис.Рік тому
High Fidelity Neural Audio Compression | Paper & Code Explained
Ask Me Anything: A Simple Strategy For Prompting Language Models | Paper Explained
Переглядів 3,6 тис.Рік тому
Ask Me Anything: A Simple Strategy For Prompting Language Models | Paper Explained
interesting. Using AI in football.
Dude why suggesting Discord community?
What do you mean?
@@TheAIEpiphany Hi I didn't know what Discord was and clicked on the community link cited. It asked me to register with name then birthday and even further for phone number. I was reluctant and googling why Discord requires Phone number. By then I realized I was shown that I was logged in with a Username irrelevant to me. I was scared and googled further and found the corresponding Wikipedia page en.wikipedia.org/wiki/Discord . Kindly take a look at the Criticism and controversies section of the Wikipedia page. I personally feel it shouldn't be promoted among the learning community.
@@umapoorni discord is something where all AI communities are right now, see eleuther ai, unsloth ai, axolotl, etc etc
Love these gnn talks!
nice talk !
I can imagine this being used for war
how?
I like how it has become a regular ritual for these talks that the speaker tells any questions are welcome in between the speech and then Aleksa reminding everyone that they should still raise hands XD
otherwise it becomes a mess lol
great talk so far. can you maybe create a in depth tutorial for cloud compute? how to get the dataset to the cloud and so on.
i think that would be better for anther channel. what data set, what purpose? maybe i can advise.
@@RollingcoleW like if i want to train a huge diffusion model on like laion or something how should i download all the data to the cloud machine without the gpus ideling for days
We talked about Petar's latest work on Tactic AI! Let me know how you like this one!
Is DSA really important for a Research Engineer/Scientist Role- ML field? Can it be ignored? Isn't ML foundations, Statistics and Math enough?
Great job, way to go :)
Are these chips compatible with an ordinary computer like incase if i want to use the chip in an offline model. and is it on sale
1:20:00 is there a proper explanation of why there is a logit scale factor when calculating similarity? Thanks.
why not i9 14900k?
man I normally love your videos but this one is a bit unclear
So how exactly is this making money?
Hi, i am always confused about the forward process equation defined in (2). we say the our images x come from an unknown distribution q(.). but in equation (2) we are saying that this distribution is normal ? we are sampling from a normal distribution to get next forward step. sorry I am not that good when it comes to probability theory.
Gabarjobat?
Thank u so much
Simple eh?
loving this series, will be sad after i watch the fourth and final video :(
Loved this, thanks!
Hi, I love your videos, especially the one about MuZero and VQ-VAEs. I recently stumped upon a paper about a Stochastic Muzero, which combines the two. Also, I found an improved version of MuZero called Efficient Muzero. It would be awesome if you would cover one of the two as your explanation really helps to build the intuition about models.
This video help me understand the architecture of the model. Great job!!! What changes do we have to do in the code in order to take images 512 or 1024 ? Which part of the code should we change?
does it work when I install it with deb (Network ) ?
So the parallelism you demo'd with pmap, that was data parallelism correct? replicating the whole model across all the devices, sending different batches to each device, and then collecting the mean model back on the host device after forward and backwards pass? am i understanding that correctly?
ty ty ty ty ty for this video
how to did you purge the incorrect cudnn that was installed first?
awesome video
Does this still exist?
How does the fine-tuning actually work here? They don't talk about it in the paper, right?
Yes, it is a kind of parameterized classification though embedding vector from text input as the last Fully connected layers in the classification network of image classification network.
if you drop the attention for the more past tokens, then they begin to specialise themsselves for the local computation, and their computation will not be lost the further you go into the futur because it will be more referenced in the higher layers and will come back as a general past.
question regarding DETR bipartite Matching loss If I got it right than the loss match for every target only one query expert Therefor if there are 10 targets than the loss guide the model to match 10 query to the target classes and 290 (out of 300) to the no-object class is this correct? If so, than it might happen that a query that was not match to a target, but also was very close to match the target will be panelize by the loss model (panelize for not showing the no-obj class). This might confuse the model, especially if in the next epoch the the matched query and the unmatched query will be flipped. If I got it correct than the penalty for not showing the no-obj class is low, but it still shouldn't be panelize. I think it would be much better to include in the loss all the query that matched the target, so that they could be use for by the loss to improve the model This will cause the model to produce multiple prediction per target, but during prediction we can use only the prediction with the best score and ignore the rest (or use NMS)
Wow! I watch to the end, and learned a lot. Few remarks * I think it would be better to break it into a serie * The first hour was very clear, but till the end it became a bit vague * I download the mini coco and modified the download script as u suggested, but needed chatgpt help to generate the appropriate annotation file Thanks 💫
Please help, when using expressivity_predict I get the error expressivity_predict is gated. How to fix it?
Is this method lossy or lossless?
3:51 PackagesNotFoundError: The following packages are not available from current channels: - torchvission torchvission 0.5.0 is not available to install , can someone please help me out?
Okay I just figured it out, For people who faces same issue, I just did removed the torchvision dependency line from environment.yml file and created the environment and all , the in the IDE's command prompt i just "pip install torchvision==0.5.0" and it worked just fine .
the video doesnt demonstrate the actual steps to actually deploy large models using deepspeed? which is much more important than understanding all the theory stuff
get prices, and while Groq seems awesome, no one asked how much a groq system that compares with a H100 ( for LLM ).
great video thanks for explaining so nicely. Can you pls share the code as well?
Your haiku video never came out sir 😅, thanks for this series btw
Has anyone run a raytracer on this yet? could be interesting.
Jako dobar interviju. Samo tako nastavi!
Thanks for your effort. The video is perfect. I have a question. Can I detect objects in video with this?
Very interesting.
Great presentation!
You mentioned posterior and prior, can you provide some reference, why they model it in this way ?
Not a genius, just a massive data thief, nothing more.
No such thing as "stealing" data. Copying bits doesn't take it away from anyone:)
🐒
Does CPU need large cache size?
This is such a great presentation!