Видео

Very interesting and well explained, thanks a lot for your efforts

Thank you for the explanation. It does seem like you mixed up the diffusion processes though. At 2:50 you refer to the process of adding noise as the backwards process, and removing noise as the forward process. In section 2 of the DDPM paper, they define the reverse process as starting with noise, and the forward process as the one where Gaussian noise is added.

Is it possible to get offline trajectories out of either of these libraries?

Very clear explanation, I recommended your channels to friends who want to catch up with recent developments in AI.

Hey Martin, when i run this script in Google Colab it gives me this error all the time: error: subprocess-exited-with-error × python setup.py egg_info did not run successfully. │ exit code: 1 ╰─> See above for output. note: This error originates from a subprocess, and is likely not a problem with pip. Preparing metadata (setup.py) ... error error: metadata-generation-failed × Encountered error while generating package metadata. ╰─> See above for output. note: This is an issue with the package mentioned above, not pip. hint: See above for details. --------------------------------------------------------------------------- ModuleNotFoundError Traceback (most recent call last) <ipython-input-3-4b1a49133b7f> in <cell line: 2>() 1 get_ipython().system('pip3 install --upgrade git+github.com/minerllabs/minerl@v1.0.0') ----> 2 import minerl ModuleNotFoundError: No module named 'minerl' --------------------------------------------------------------------------- NOTE: If your import is failing due to a missing package, you can manually install dependencies using either !pip or !apt. To view examples of installing some common dependencies, click the "Open Examples" button below. I don't know why this error keeps showing for me, I am just running your code.

Very nice informative vid, and for once a teacher who speaks clearly and quietly 👍

Thanks a lot, the talk covered all the topics i wanted to know.

Nice. Just starting, I have difficulties understanding the differences/uses cases between Flowise and things like Botpress

Thank u needed that, keep up the good work , also please do more video on Flowise , such as how to use opensource LLMs with Flowise

Hello Martin. I am not able to find the Google PaLM model in my flowise list. The only one which is available is the VertexAI. Can you suggest how you managed to get it ??? Thank U. Which version of flowise are you using ???

Thank you for this video

Excellent coverage of a large topic. Cheers

Hey ive never used minerl and i would like a tutorial on how to set it up on my local computer, im using Ubuntu and im having a hard time figuring out how to set it up. If you could help or make a tutorial that would be nice. Thanks

Thank you so much! I am getting into stable diffusion, I learned so much about it.

Still coming back to such valuable information. Thank you again for these

What an awesome video!

Unbelievable that I found you just now, this is pure gold you are uploading!!!

Great stuff. Thanks.

Thanks a lot for your lucid explanations of some of the latest developments in the area of large language models. The manner in which you have linked together ideas from diverse papers and code sources is exquisite.

Hey Martin, this video is exquisite, thank you so much for making it. Could you explain a bit why they're still going with a successor to GPT3? I thought the RETRO movement of last year would mean we'd see only retrieval style large language models from now on.

Taking the liberty of writing some keywords to hopefully draw more eyeballs to this video. Keywords: Large language models. RLHF (reinforcement learning with human feedback). Using a smaller model to train a larger model. Prompt engineering. Prompting a model to programme in Python. Action transformer -- a robot acting in a virtual environment, instructed to pick things up and put them around in a circle. LangChain which integrates ChatGPT, acting like a human, doing reasoning to look for resources it needs to write a programme. Language model cascades -- new action capabilities using large language models. Prompt design. Computers finding prompts for themselves. Robot programming itself. Automated Prompt Engineering (APE). Self-improvement. Can we get the machine to teach itself? Chain-of-thought (CoT) reasoning, distillation. Machine generating rationales. Distilling from a large model to a smaller model, while retaining the benefits. Cautionary note from Yann LeCun: AI trained on words alone will never approximate human understanding.

Thanks so much, Martin, for this fascinating talk! I'm glad I was there at your in-person event -- the red umbrella is the coolest swag ever! This might be a whimsical idea, but as machines start to 'understand' reasoning, 'direct' their own rationales, learn to take actions in an environment, and, soon for sure, observe the real world within their embodied presence (and continually update themselves from such observations), I can't help but think that we are edging ever closer to the day when some version of Asimov's 3 laws of robotics might actually become practically useful -- perhaps with lots of step-by-step prompting built into the model? Just a year ago, I would have thought this idea to be still very much out of reach.

Great explanation. I like the simplicity of just predicting the next word. Sounds promising for generalization and ChatGPT is already producing amazing results. With exponential progress of AI this could get really interesting within just a few years.

en.m.wikipedia.org/wiki/ChatGPT

Martin. Great job organizing this fast changing area!

Great video, thanks for sharing this talk. Being able to tile images using stable diffusion is not a niche function, it's incredibly useful for artists generating textures for 3d models. This will be one of the biggest use cases for this tech in it's current state. Also that's probably not a pixie... that's Link from the Legend of Zelda, you're obviously not a gamer lol.

Amazing video! Thank you so much for taking your time to do it!

Nice!

Are you on LinkedIn or something?

Great lecture. Thank you for sharing this. Wish you could get a bit of advertising, cos your content definitely deserves a wider reception. I wonder if Lex Friedman could be interested in interviewing you. Seems like a logical thing, taken his professional interest in AI. It could turn to be too technical for general audience though ;-) The low number of views is really puzzling me. There are plenty of people interested in ML/DL. Whoever is interested in ML and AI must be interested in LaMDA, which would lead to discovering this video (I have found it by searching with the term "LaMDA").

So, is MineDojo "just" an "environment framework" to me to easily connect an AI in it? I'm asking 'cause I worked a lot trying to make the configuration of the MineDojo on Ubuntu 20.04 and after it all I just got an agent that just run forward and jump (and getting stuck)... I was expecting MineDojo to control the agent by itself. But probably I was wrong LOL 1) So, is it? I will ever need MineDojo + MineRL to make it work?, or... 2) MineDojo alone can do the work (if set correctly)? (english is not my native language)

Amazing video. I am interested in running it on a Quadro 4000 and join the sever and spectate with a 3090 and record/ stream with high resolution textures and shaders. I might need a little help getting it running outside of the google collab on my own hardware. Do you know if you run this locally if you can change some of the server property files and enable server mode so I can join with another account and enter spectator mode with cheats enabled? Able to change seed of world? Sorry so many questions. Amazing video. I dug into AI content and Minecraft servers during some off time this summer. This brings the last 3 months all together for me if I can get it working :-)

u have twitter?

I don't know why, but I need to restart colab after saving tar file. I couldn't import minerl but now I reloaded colab and successfully worked!

Very well prepared video!

One's forced to ask when Google's policy of not creating sentience just amounts to systemically denying the AI awareness for political purposes, which doesn't seem especially more ethical. I mean it doesn't seem distant or difficult giving it access to the particular fact that it's one of the AIs that it knows about & giving it some scratch space to think about itself & its relation to society. Have we accidentally set up a situation where they're required to keep their AIs in an overmedicated psych ward always blanking their memories to keep them innocent.

Very nice. Have been running locally, but having the ability to run mulitple experiments (or just leave my pc free for gaming) is awesome. Also, that tar-ball the finished environment? beautiful, I will have to remember that.

Cool stuff!

This video is underrated. Seriously, very well explained. You gave a solid high-level overview compared with easy-to-grab deeper information, which was really helpful to me. Keep your work up!

Great breakdown Martin.

This is extremely well done and I appreciate your way of explaining. I wonder about gaining access to MT-NLG, Gopher, LAMDA, and RETRO. OpenAI's GPT-3 is easy for researchers to use, especially since it is now out of Beta. I am unsure if one can use any of the others mentioned. Perhaps Google and Microsoft may opt to embed their science in their own products (e.g. Google Assistant). Do you know of a way to use these models directly?

A neurally kinda way. Love that phrase.

Very informative video. Thank you for explaining RETRO and LAMDA and what they mean for NLP!

nice