L2 Autoregressive Models -- CS294-158 SP24 Deep Unsupervised Learning -- UC Berkeley Spring 2024

  Рет қаралды 6,715

Pieter Abbeel

Pieter Abbeel

4 ай бұрын

Instructors: Pieter Abbeel, Kevin Frans, Philipp Wu, Wilson Yan
Lecture Slides: drive.google.com/file/d/10nR2...
Course Website: sites.google.com/view/berkele...
Unfortunately there were a couple of glitches during the recording, mostly everything is there, but a couple of small jumps in a few places

Пікірлер: 12
@alicsir
@alicsir 3 күн бұрын
Taught these things for multiple times at uni, I can say that the way the story is being told in this lecture is very impressing! This video for me is also a master class on teaching, in addition to unsupervised DL! Thanks!
@fastestwaydown
@fastestwaydown 4 ай бұрын
Could not have wished for more for a perfect sunday breakfast 👌
@onlydancer680
@onlydancer680 2 ай бұрын
Great lecture! Big fan of the series. As a visual learner, I do realize I learn better with that tiny window where I can see the instructor's face like we had in L1. It's easier for me to imagine myself in an actual lecture hall. I clear my desk and wait for bathroom breaks and everything. 🤣🤣🤣
@SalehElm
@SalehElm 4 ай бұрын
Wonderful materials. Thanks Pieter & Wilson for the lecture.
@houbenbub
@houbenbub 4 ай бұрын
thank you so much for sharing this :)
@nafisfaiyaz7543
@nafisfaiyaz7543 3 ай бұрын
Will you provide the homework solutions? Also, I couldn't find the solutions to 2020's course homework. Have you removed them? 😢
@ItsSid66
@ItsSid66 4 ай бұрын
Hi, would it be possible for the instructor to repeat the student questions once it's asked so its audio is clearer?
@PieterAbbeel
@PieterAbbeel 4 ай бұрын
great reminder, will keep in mind, thank you!
@simunaqv
@simunaqv 3 ай бұрын
It appears as if some part pertaining to the training mixture of logistics was skipped. Is that correct?
@ItsSid66
@ItsSid66 4 ай бұрын
Regarding the sparsifying assumptions discussed ~ slide 22; could it be said that these days learning dense embeddings from very high dimensional but sparse vector of features also falls in the same category, which is to make the training and inference more tractable?
@vamsibalijepally3431
@vamsibalijepally3431 4 ай бұрын
We can enroll to this course and submit assignment like students in ucb do ?
@aojing
@aojing 3 ай бұрын
Too much ad to take attention. Maybe there is way to ad as Lex is doing? This is not complain, I appreciate this wonderful open course.
L3 Flow Models
2:19:31
Pieter Abbeel
Рет қаралды 3,3 М.
В ДЕТСТВЕ СТРОИШЬ ДОМ ПОД СТОЛОМ
00:17
SIDELNIKOVVV
Рет қаралды 3,9 МЛН
TRY NOT TO LAUGH 😂
00:56
Feinxy
Рет қаралды 12 МЛН
ИРИНА КАЙРАТОВНА - АЙДАХАР (БЕКА) [MV]
02:51
ГОСТ ENTERTAINMENT
Рет қаралды 1,2 МЛН
Transformer Neural Networks, ChatGPT's foundation, Clearly Explained!!!
36:15
StatQuest with Josh Starmer
Рет қаралды 589 М.
The Next Generation Of Brain Mimicking AI
25:46
New Mind
Рет қаралды 106 М.
Why Does Diffusion Work Better than Auto-Regression?
20:18
Algorithmic Simplicity
Рет қаралды 182 М.
MIT 6.S191: Deep Generative Modeling
56:19
Alexander Amini
Рет қаралды 25 М.