TITLE stringlengths 18 100 | URL stringlengths 43 43 | TRANSCRIPTION stringlengths 246 3.27k | transcription_length int64 52 500 | text stringlengths 303 3.36k | ada_embedding stringlengths 33.7k 34.7k |
|---|---|---|---|---|---|
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | Hello there, today we'll look at Glide towards photorealistic image generation and editing with text-guided diffusion models by Alex Nicolle, Prafula Dhariawal, Aditya Ramesh and others of OpenAI. This paper on a high level, well, I'll just show you what you can do. I'm sure you've all seen this paper in one way or ano... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: Hello there, today we'll look at Glide towards photorealistic image generation and editing with text-guided diffusion models by Alex Nicolle, Prafula Dhariawal, Aditya Ramesh and others of OpenAI. This paper on a high level, w... | [-0.023338761180639267, 0.01196559239178896, -0.00887258630245924, 0.011793375946581364, 0.016050564125180244, 0.020445525646209717, -0.015127484686672688, 0.006998872850090265, -0.02039041742682457, -0.04086349532008171, 0.009527008980512619, 0.025873785838484764, -0.014838160946965218, 0.005617697723209858, -0.002254... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | I'm going to paint this area right here. And I'm going to issue the prompt around coffee table in front of a couch, and the model will generate it and so on. You can see that this enables sort of an interactive creation of this scenery at the end, the couch, the couch in the corner of the room. So changing the entire w... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: I'm going to paint this area right here. And I'm going to issue the prompt around coffee table in front of a couch, and the model will generate it and so on. You can see that this enables sort of an interactive creation of thi... | [-0.02665344625711441, 0.006354713812470436, 0.003984984941780567, -0.0034002691973000765, 0.006004913244396448, 0.005823154002428055, -0.022099178284406662, -0.004386227112263441, -0.01188636664301157, -0.04137251526117325, 0.011118177324533463, 0.028999168425798416, -0.011282789520919323, -0.0028841416351497173, 0.00... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | model substantially to make very misleading pictures. But we'll get to that as well. Alright, so what is a diffusion model? And that's sort of at the core of this thing right here. A diffusion model is a different type of generative model than maybe you're used to from like a GAN or a VQ VAE. So in a GAN, a GAN is prob... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: model substantially to make very misleading pictures. But we'll get to that as well. Alright, so what is a diffusion model? And that's sort of at the core of this thing right here. A diffusion model is a different type of gene... | [-0.023775776848196983, -0.0021778957452625036, -0.005056838039308786, 0.012419488281011581, -0.0004962977254763246, 0.026752732694149017, -0.015828367322683334, 0.0031048720702528954, -0.027669742703437805, -0.0451061986386776, -0.006605120841413736, 0.02717801183462143, -0.009209956973791122, 0.02117094025015831, 0.0... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | the cat, and predict and learn to predict the more sharp version of the cat. Okay, this is a foundation of many, many sort of denoising models, many up sampling models, super resolution models, what have you, okay, they do this in one step. But essentially, here we say, the individual step is small enough such that the... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: the cat, and predict and learn to predict the more sharp version of the cat. Okay, this is a foundation of many, many sort of denoising models, many up sampling models, super resolution models, what have you, okay, they do thi... | [-0.01872330904006958, 0.016276435926556587, -0.0016965885879471898, -0.001656877575442195, 0.017087554559111595, 0.02014276571571827, -0.004102905746549368, 0.008476183749735355, -0.018493492156267166, -0.04169146716594696, 0.003450631396844983, 0.03328287601470947, -0.012072140350937843, 0.003839292097836733, 0.00286... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | at each step is small enough, the posterior is well, well approximated by a diagonal Gaussian, that's what they say right here. So what does this mean? The posterior, it means that this is the reverse step, right? I have xt, and I'm looking to recreate xt minus one. So if the noise is small enough, then the posterior i... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: at each step is small enough, the posterior is well, well approximated by a diagonal Gaussian, that's what they say right here. So what does this mean? The posterior, it means that this is the reverse step, right? I have xt, a... | [-0.016091231256723404, 0.017658470198512077, 0.0004116538038942963, 0.005556270945817232, 0.004762517753988504, 0.009538546204566956, 0.013584998436272144, 0.0020198479760438204, -0.021022632718086243, -0.02541360631585121, 0.007815933786332607, 0.04955720901489258, -0.022508807480335236, 0.01799623854458332, 0.012369... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | we're going to train the neural network to predict x t minus one from x t or the variational sort of the distribution of that. So this is a training sample. Now, how do we get the training sample, what we can do is we can take x zero right here, and we could go through and add and add and add noise. But since we always... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: we're going to train the neural network to predict x t minus one from x t or the variational sort of the distribution of that. So this is a training sample. Now, how do we get the training sample, what we can do is we can take... | [-0.015456033870577812, 0.011581901460886002, -0.003339245682582259, 0.01276978850364685, 0.022461868822574615, 0.014187154360115528, -1.7479787857155316e-05, 0.008922653272747993, -0.032720897346735, -0.03477270156145096, -0.002961281454190612, 0.040415164083242416, -0.012196092866361141, 0.009543594904243946, 0.00333... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | to reconstruct one step. So that's going to predict the noise that was added, given the image xt, given the index t. What we can also do is we can say, by the way, it's also, we give it the label y. So y, in this case is cat. So we can train a class conditional model. And that, you know, has some, some advantages, we k... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: to reconstruct one step. So that's going to predict the noise that was added, given the image xt, given the index t. What we can also do is we can say, by the way, it's also, we give it the label y. So y, in this case is cat. ... | [-0.025305239483714104, 0.026801859959959984, -0.012144592590630054, 0.0029537654481828213, 0.02540135383605957, 0.01577630080282688, -0.005516212899237871, 0.007476236205548048, -0.022037390619516373, -0.02509928308427334, -0.0014554288936778903, 0.04973173141479492, -0.006762252189218998, 0.005680978298187256, -0.001... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | is called guided diffusion. And one way to do it is to say, well, I have an additional classifier, I have a classifier, for example, an ImageNet classifier, right? And if I want to push my diffusion process towards a particular label, I can take that ImageNet classifier, and I can go along the gradient of that. This is... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: is called guided diffusion. And one way to do it is to say, well, I have an additional classifier, I have a classifier, for example, an ImageNet classifier, right? And if I want to push my diffusion process towards a particula... | [-0.02795610949397087, 0.020477404817938805, -0.006167191546410322, -0.0004171243926975876, 0.016176465898752213, 0.025175565853714943, 0.0025904944632202387, 0.01579294167459011, -0.021080084145069122, -0.03161327913403511, 0.0037907168734818697, 0.03199680149555206, -0.01340961828827858, 0.01698460429906845, 0.004273... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | we just train the model in both ways. During training, we sometimes just leave away the label. This could be beneficial, as this part, in fact, would be the opportunity to bring more data into the picture, right? Let's say I have only part of my data is labeled and part of my data is unlabeled. We could actually in her... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: we just train the model in both ways. During training, we sometimes just leave away the label. This could be beneficial, as this part, in fact, would be the opportunity to bring more data into the picture, right? Let's say I h... | [-0.02293958142399788, 0.00941467098891735, 0.0065055652521550655, 0.007706699427217245, 0.018607165664434433, 0.01347630936652422, -0.006172302644699812, 0.007071417290717363, -0.026050033047795296, -0.03585350885987282, 0.0012375848600640893, 0.03818634897470474, -0.022481344640254974, 0.011830825358629227, 0.0089842... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | was a better direction because we also have the unconditional point right here. We can clearly say that this direction is probably the direction that goes into the direction of the conditioning information. So we can choose to sort of overdo it. Again, I think that is that's kind of a trick around the fact that we don'... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: was a better direction because we also have the unconditional point right here. We can clearly say that this direction is probably the direction that goes into the direction of the conditioning information. So we can choose to... | [-0.025296509265899658, 0.012773004360496998, -0.005128607153892517, -3.094382554991171e-05, 0.013653184287250042, 0.026294507086277008, 0.012572018429636955, -0.004716239403933287, -0.028193477541208267, -0.03215775266289711, 0.014595738612115383, 0.03889424726366997, -0.011525505222380161, 0.012142323888838291, 0.002... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | just gets as an additional conditioning information which step it's currently trying to reconstruct. It always reconstructs the noise that was added. Training data generation is pretty easy. You simply add noise to an image and then you add a bit more. And then the difference between that is the target to predict. Then... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: just gets as an additional conditioning information which step it's currently trying to reconstruct. It always reconstructs the noise that was added. Training data generation is pretty easy. You simply add noise to an image an... | [-0.03376753255724907, 0.017874466255307198, 0.00022630873718298972, -0.002762797987088561, 0.027516352012753487, 0.02811635285615921, 0.005236060358583927, -0.002270936267450452, -0.02464192546904087, -0.03879079967737198, -0.0007844497449696064, 0.036921028047800064, -0.011965148150920868, 0.010255841538310051, 0.009... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | could do it with diffusion models without training. But they say if you train it, it behaves a bit better. So during training, they would sort of mask out random parts of the images and then use diffusion to reconstruct those. And yeah, the results are the results that we've already seen. These are pretty interesting. ... | 500 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: could do it with diffusion models without training. But they say if you train it, it behaves a bit better. So during training, they would sort of mask out random parts of the images and then use diffusion to reconstruct those.... | [-0.04057515040040016, 0.009907552972435951, 0.003983781207352877, -0.004896507598459721, 0.025570649653673172, 0.020802997052669525, -0.012484662234783173, 0.007097788155078888, -0.018526550382375717, -0.04670294374227524, 0.0004570342134684324, 0.03977338597178459, -0.013558457605540752, -0.00043645311961881816, 0.00... |
GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models | https://www.youtube.com/watch?v=gwI6g1pBD84 | The samples as such, they look they look pretty, pretty cool. It's also not clear how much of a difference this is between the small model and the large model or how much effort into diffusion is put. They also say they they they release the model they release is sort of a model on a filtered version of a data set. And... | 470 | GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models: The samples as such, they look they look pretty, pretty cool. It's also not clear how much of a difference this is between the small model and the large model or how much effort into diffusion is put. They also say they they t... | [-0.02141856774687767, -0.0002963871229439974, 0.003688373137265444, -0.0014549912884831429, 0.0054230717942118645, 0.021733345463871956, -0.012201055884361267, 0.00972389243543148, -0.017148539423942566, -0.03591202571988106, 0.011078805662691593, 0.03654158115386963, -0.01506827026605606, -0.016081033274531364, 0.022... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | Not a day goes by in AI research in which we don't get a new image generation model these days. So take a look at the top row right here and listen to the prompt that generated them. Oil on canvas painting of a blue night sky with roiling energy. A fuzzy and bright yellow crescent moon shining at the top. Below the exp... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): Not a day goes by in AI research in which we don't get a new image generation model these days. So take a look at the top row right here and listen to the prompt that generated them. Oil on canvas painting of a blue night... | [-0.015076294541358948, 0.00441666878759861, 0.004800429102033377, 0.004145980812609196, -0.00894298404455185, 0.013020436279475689, 0.0033424829598516226, 0.007716321386396885, -0.01744738407433033, -0.0532604344189167, 0.006787758786231279, 0.033825721591711044, -0.0142128337174654, -0.0060065328143537045, -0.0073668... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | city of Los Angeles is in the background. High res DSLR photograph. That's literally that's the academic version of the Unreal Engine trick right here. And you can see the images spot on. So this requires a lot of knowledge, not only of, you know, what a DSLR photograph is, but also how the skyline of Los Angeles looks... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): city of Los Angeles is in the background. High res DSLR photograph. That's literally that's the academic version of the Unreal Engine trick right here. And you can see the images spot on. So this requires a lot of knowled... | [-0.008869458921253681, 0.009263504296541214, 0.00909067690372467, 0.0013117536436766386, -0.0013394058914855123, 0.018209006637334824, -0.00972667895257473, -0.001099176937714219, -0.01919066160917282, -0.05746137723326683, 0.011931946501135826, 0.020047880709171295, -0.0030745845288038254, 0.00030266251997090876, -0.... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | So if you are not aware, autoregressive models, they work on tokens. Now, tokens in usually in natural language processing are words or part of words. So these would be tokens, token one, token two and so on until token N. And then what you would try to do is you would try always to predict the next token. That's what ... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): So if you are not aware, autoregressive models, they work on tokens. Now, tokens in usually in natural language processing are words or part of words. So these would be tokens, token one, token two and so on until token N... | [-0.020194722339510918, 0.0006662850501015782, 0.006074508186429739, 0.013248596340417862, 0.0067919171415269375, 0.01023816131055355, -0.004559233784675598, -0.0024422071874141693, -0.021146798506379128, -0.03609840199351311, 0.026081498712301254, 0.038109827786684036, -0.03491836413741112, -0.009259266778826714, -0.0... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | defined set of tokens. I believe in their case, they have like eight eight thousand tokens or so. And your image tokens must be of these eight thousand. So the image has a bunch of tokens, but they all must be one of the things in the vocabulary here. Now, the vocabulary is also learned. There are some techniques by wh... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): defined set of tokens. I believe in their case, they have like eight eight thousand tokens or so. And your image tokens must be of these eight thousand. So the image has a bunch of tokens, but they all must be one of the ... | [-0.03941061720252037, -0.00027145937201566994, 0.013452793471515179, 0.00861851591616869, -0.0018034789245575666, 0.010289034806191921, -0.0017071684123948216, -0.008741247467696667, -0.02681013010442257, -0.04172888770699501, 0.026891950517892838, 0.03594684600830078, -0.028882937505841255, -0.003756963647902012, -0.... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | lot. By the way, here you can see a bunch of the I'm not going to go into the architectural details quite quite as much. But they do also train an upsampler. So they have images of resolution 256 by 256. Ultimately, they do train an upsampler as well, where so here this is the upsampler super resolution upsampler, wher... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): lot. By the way, here you can see a bunch of the I'm not going to go into the architectural details quite quite as much. But they do also train an upsampler. So they have images of resolution 256 by 256. Ultimately, they ... | [-0.028051301836967468, -0.0005400425288826227, 0.01615491323173046, -0.004409631714224815, 0.003185306442901492, 0.010680650360882282, 0.011051553301513195, -0.00029298808658495545, -0.024644482880830765, -0.04555245861411095, 0.03013935126364231, 0.02625173330307007, -0.024795591831207275, -0.0076928152702748775, 0.0... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | to the other things, which are the data sets that they use. So they have three data sets, three main data sets right here. One is MS Coco. Now MS Coco, as they show right here for the image on the right hand side, it simply says a bowl of broccoli and apples with a utensil. So it just kind of is a high level descriptio... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): to the other things, which are the data sets that they use. So they have three data sets, three main data sets right here. One is MS Coco. Now MS Coco, as they show right here for the image on the right hand side, it simp... | [-0.035032715648412704, -0.0022692906204611063, 0.007544654421508312, -0.018667472526431084, 0.003668310819193721, 0.01422944013029337, 0.004365220665931702, 0.006414343137294054, -0.028874946758151054, -0.05394982546567917, 0.0040566385723650455, 0.03173192963004112, -0.024908455088734627, -0.015838226303458214, -0.00... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | because they create these prompts by letting the prompt engineers sort of, they choose, for example, a challenge. So the challenge might be perspective, right, which could be, you know, I need a prompt that asks for some object in some specific perspective that is unusual or quantity. Like I need a prompt that asks for... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): because they create these prompts by letting the prompt engineers sort of, they choose, for example, a challenge. So the challenge might be perspective, right, which could be, you know, I need a prompt that asks for some ... | [-0.026076892390847206, -0.005052397958934307, 0.00855293683707714, -0.0019203363917768002, -0.0012506988132372499, 0.020549725741147995, 0.011890496127307415, 0.0011452928883954883, -0.02726735919713974, -0.05901314318180084, 0.0018477037083357573, 0.02830193191766739, -0.024957286193966866, -0.015235140919685364, 0.0... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | these things right here, this and this, there may be like Dolly Mini kind of style pictures. And there are also that scale, right? And then we go to the three B model. And this is something that would be familiar maybe from something like Dolly or Dolly, maybe between Dolly and Dolly too, right? These things you can se... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): these things right here, this and this, there may be like Dolly Mini kind of style pictures. And there are also that scale, right? And then we go to the three B model. And this is something that would be familiar maybe fr... | [-0.010807028040289879, -0.007562799379229546, 0.019917728379368782, -0.006410709582269192, -0.017062241211533546, 0.012439744547009468, -0.006802985444664955, 0.0018924669129773974, -0.02405959740281105, -0.041192520409822464, 0.010291061364114285, 0.023409338667988777, -0.022971119731664658, 0.00045235437573865056, 0... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | idea of combining like a sloth with a van. Right. So they start by just exploring the model and entering things like a smiling sloth, like what comes out. Right. And a van parked on grass. There are always good images and bad images that turn out and they sort of learn how to have to tweak the prompt to get what they w... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): idea of combining like a sloth with a van. Right. So they start by just exploring the model and entering things like a smiling sloth, like what comes out. Right. And a van parked on grass. There are always good images and... | [-0.027048632502555847, 0.017441419884562492, 0.009119292721152306, -0.002515193074941635, -0.021344779059290886, 0.01962675154209137, 0.005738213658332825, -0.00020079452951904386, -0.015640927478671074, -0.04502607882022858, 0.0043397387489676476, 0.03246385604143143, -0.017413930967450142, 0.005033822264522314, 0.00... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | There's a bunch of examples. And this one, I told you, it's the horse riding on an astronaut. So they have to actually specify the horse is sitting on an astronaut because the riding is just is just riding indicates too much that the horse is on the bottom. But I just found the horse riding on the astronaut to be absol... | 500 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): There's a bunch of examples. And this one, I told you, it's the horse riding on an astronaut. So they have to actually specify the horse is sitting on an astronaut because the riding is just is just riding indicates too m... | [-0.024759072810411453, 0.0031643148977309465, 0.00786652509123087, 0.0008440198143944144, -0.003954092040657997, -0.003186879912391305, -0.020523784682154655, -0.012525340542197227, -0.014399976469576359, -0.03166050836443901, 0.01589968428015709, 0.022176241502165794, -0.010879827663302422, -0.0017774319276213646, 0.... |
Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained) | https://www.youtube.com/watch?v=qS-iYnp00uc | search an appropriate stock photo. You just type it. It's so cool. Or you want to change something in a picture. You just erase it. You just say, well, ever here, change that part to something else. So cool. No Photoshop skills anymore. No drawing skills anymore. Just you and your mind and your creativity. All right. T... | 122 | Parti - Scaling Autoregressive Models for Content-Rich Text-to-Image Generation (Paper Explained): search an appropriate stock photo. You just type it. It's so cool. Or you want to change something in a picture. You just erase it. You just say, well, ever here, change that part to something else. So cool. No Photoshop ... | [-0.02715059369802475, 0.005720123648643494, 0.008727970533072948, -0.01723339408636093, 0.0013326671905815601, 0.021670402958989143, -0.0042040301486849785, -0.012636431492865086, -0.009089606814086437, -0.042394984513521194, 0.005469759926199913, 0.026302138343453407, -0.010647428222000599, -0.005386305041611195, 0.0... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | Google releases imagine an unprecedented text to image model, cog view two improves drastically over cog view one and mid journey moves into open beta. Welcome to ML news. Hello, hello and welcome to ML news. Today we talk all about text to image models, text and image models, any sort of artistic models that we might ... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): Google releases imagine an unprecedented text to image model, cog view two improves drastically over cog view one and mid journey moves into open beta. Welcome to ML news. Hello, hello and welcome to ML news. Today we talk... | [-0.01925118826329708, 0.010585411451756954, 0.005769186187535524, 0.0003451485245022923, 0.008041895925998688, 0.017770327627658844, -0.009810701943933964, -0.015535324811935425, -0.009241667576134205, -0.0391879603266716, -0.0012768994783982635, 0.04105274751782417, -0.025023803114891052, -0.0037055795546621084, 0.00... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | we show that large pre trained frozen text encoders are very effective. And in fact, we show that scaling the pre trained text encoder size is more important than scaling the diffusion model size, which is really interesting, because you would think that for an image generation model, the part that actually generates t... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): we show that large pre trained frozen text encoders are very effective. And in fact, we show that scaling the pre trained text encoder size is more important than scaling the diffusion model size, which is really interesti... | [-0.033384546637535095, 0.002952111652120948, 0.011526958085596561, -0.01286402903497219, 0.019281968474388123, 0.027782924473285675, -0.008036499843001366, -0.006819061003625393, -0.020956825464963913, -0.053398385643959045, 0.015017416328191757, 0.03611498698592186, -0.01877528987824917, 0.004458075854927301, 0.00585... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | weights. So there's another thing if you haven't followed this literature, there is this concept of classifier free guidance, which is a bit of a hack. So the way it works is that this model trains to go from text to image. So every procedure, every generation is conditioned on a piece of text. However, you can do a tr... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): weights. So there's another thing if you haven't followed this literature, there is this concept of classifier free guidance, which is a bit of a hack. So the way it works is that this model trains to go from text to image... | [-0.04318155348300934, 0.004780516028404236, 0.006618103012442589, -0.00020997571118641645, 0.019540490582585335, 0.021367616951465607, -0.0019613713957369328, -0.01506333239376545, -0.023152900859713554, -0.046975281089544296, 0.004926965106278658, 0.037407275289297104, -0.03160510212182999, 0.003957611508667469, -0.0... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | might think of descriptions to photographs, but you can do so much more if you simply formulate it correctly. This is very much in the style of something like T five. So for example, if you think of segmentation based generation, the input image isn't a photo, but it's the segmentation map and the input text isn't the ... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): might think of descriptions to photographs, but you can do so much more if you simply formulate it correctly. This is very much in the style of something like T five. So for example, if you think of segmentation based gene... | [-0.023975631222128868, 0.001966311363503337, 0.0033469502814114094, 0.0004067168920300901, 0.010503408499062061, 0.022723382338881493, -0.0035456919576972723, -0.019881201907992363, -0.011917464435100555, -0.04190107434988022, 0.0006357974489219487, 0.02991325780749321, -0.029828837141394615, -0.0017279972089454532, -... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | the things in parallel, which gives a great increase in inference speed. There is a demo on hugging face spaces, if you want to play around with it, I'll link it in the description. Motherboard writes Google bans deepfakes from its machine learning platform. So apparently, a lot of people have used collabs to generate ... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): the things in parallel, which gives a great increase in inference speed. There is a demo on hugging face spaces, if you want to play around with it, I'll link it in the description. Motherboard writes Google bans deepfakes... | [-0.02987321838736534, -0.015462643466889858, 0.005256883800029755, -0.0018826494924724102, -0.010984429158270359, 0.024128368124365807, -0.020889103412628174, -0.00410099234431982, -0.010416866280138493, -0.030426939949393272, 0.02563725784420967, 0.02087526023387909, -0.009627814404666424, -0.01496429555118084, -0.00... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | closer and closer to the final result. And I think this highlights a core notion about these new text to image models. So as you can see here, it's not simply give me a cool Cosmo cover, it is trying and trying modifying the prompt trying again coming up with new ideas brainstorming. It's really kind of like almost lik... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): closer and closer to the final result. And I think this highlights a core notion about these new text to image models. So as you can see here, it's not simply give me a cool Cosmo cover, it is trying and trying modifying t... | [-0.040107034146785736, -0.011198174208402634, -0.00039504768210463226, -0.011379464529454708, -0.008541570976376534, 0.014684529975056648, -0.02266131155192852, -0.02056949771940708, -0.016093017533421516, -0.02995476871728897, 0.009064524434506893, 0.023735109716653824, -0.018909992650151253, -0.003943068441003561, -... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | them, making them better, and so on. And on top of that, he also released a free 82 page book, the Dolly prompt book in which he summarizes and elaborates on all of these things in how you can interact with these text image models in a efficient in a creative and in a more productive way. As I said, the book is availab... | 500 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): them, making them better, and so on. And on top of that, he also released a free 82 page book, the Dolly prompt book in which he summarizes and elaborates on all of these things in how you can interact with these text imag... | [-0.026330934837460518, -0.01876687817275524, -0.00131081638392061, -0.005622900556772947, -0.021460141986608505, 0.01716238260269165, -0.008538213558495045, -0.022319693118333817, -0.01879553124308586, -0.02650284394621849, 0.004580694250762463, 0.030256221070885658, -0.022749468684196472, -0.003090804675593972, 0.012... |
[ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more) | https://www.youtube.com/watch?v=af6WPqvzjjk | scaling laws up to optimization, reinforcement learning, interpretability, and more. There's also a set of links to other resources. So this in general is pretty helpful if you're kind of into machine learning into deep learning, but some topics you might want to expand your basic knowledge. And the other one is the pe... | 255 | [ML News] Text-to-Image models are taking over! (Imagen, DALL-E 2, Midjourney, CogView 2 & more): scaling laws up to optimization, reinforcement learning, interpretability, and more. There's also a set of links to other resources. So this in general is pretty helpful if you're kind of into machine learning into deep le... | [-0.025903895497322083, 0.004479906987398863, 0.0150137422606349, -0.013650743290781975, 0.012841247022151947, 0.008184910751879215, -0.004521419759839773, -0.008081129752099514, -0.008454743772745132, -0.03999052196741104, 0.0016362903406843543, 0.0398244708776474, -0.017781252041459084, 0.018030328676104546, -0.00046... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | This is a mud, a mud is very rich, and he wants to put that money to good use. So just a few days ago, he presented something called stable diffusion through an initiative that he finances called stability AI stability AI is supposed to be a third pillar. There's industry, there's academia, and now there's something el... | 500 | The Man behind Stable Diffusion: This is a mud, a mud is very rich, and he wants to put that money to good use. So just a few days ago, he presented something called stable diffusion through an initiative that he finances called stability AI stability AI is supposed to be a third pillar. There's industry, there's acade... | [0.002686872845515609, -0.01031484454870224, 0.02218172326683998, -0.008165346458554268, -0.004278394393622875, 0.018775489181280136, -0.026714211329817772, 0.01657791994512081, -0.0143391452729702, -0.010204966180026531, 0.022085579112172127, 0.02462651953101158, 0.016701532527804375, 0.00813100952655077, 0.0033907820... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | 59 page MBAs. And you're working in a corporate environment for product teams, or you have your own startup and running your own startup is terrible. And it's not something for most academics, or researchers, although of course, some of them will hopefully be very successful doing legal AI and things like that. I thoug... | 500 | The Man behind Stable Diffusion: 59 page MBAs. And you're working in a corporate environment for product teams, or you have your own startup and running your own startup is terrible. And it's not something for most academics, or researchers, although of course, some of them will hopefully be very successful doing legal... | [0.0187423937022686, -0.011403004638850689, 0.006440970115363598, -0.01361449621617794, 0.007318655960261822, 0.014070617035031319, -0.017539894208312035, 0.014305587857961655, -0.029191691428422928, -0.017996015027165413, 0.0078093307092785835, 0.021797016263008118, 0.006188721861690283, 0.011002171784639359, 0.011935... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | governments themselves to build national level models and datasets. So we have the plurality of kind of being around this. This is kind of like we kicked it off as CERN, but from a Discord group, probably through AI. So we're involved in the Lyon, OpenBioML and a bunch of these others bringing together really talented ... | 500 | The Man behind Stable Diffusion: governments themselves to build national level models and datasets. So we have the plurality of kind of being around this. This is kind of like we kicked it off as CERN, but from a Discord group, probably through AI. So we're involved in the Lyon, OpenBioML and a bunch of these others b... | [0.010221444070339203, 0.007703322917222977, 0.001934697269462049, -0.005075255408883095, -0.0006995766307227314, 0.023592311888933182, -0.009696539491415024, 0.010987520217895508, -0.02823133021593094, -0.03748099133372307, 0.012044421397149563, 0.0274368803948164, 0.0019027774687856436, 0.008930463343858719, 6.223261... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | the community to say, how can we build an efficient model that can scale to a billion people to enable them to be creative? And so that release is touch wood on the 8th or 9th of August. And we'll be releasing an open source along with instructions how to run it locally in the cloud and others. So what we've got is, yo... | 500 | The Man behind Stable Diffusion: the community to say, how can we build an efficient model that can scale to a billion people to enable them to be creative? And so that release is touch wood on the 8th or 9th of August. And we'll be releasing an open source along with instructions how to run it locally in the cloud and... | [-0.0026121963746845722, 0.017233802005648613, 0.010293779894709587, -0.0058725979179143906, 0.009539889171719551, 0.004681873135268688, -0.01713516190648079, 0.004498684778809547, -0.02426541969180107, -0.04109057039022446, 0.0009115384891629219, 0.03931505233049393, -0.001371270976960659, -0.0033414270728826523, 0.01... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | billion parameter language models or 540 billion parameters models are really usable for the vast majority of humanity. You mentioned this open source, closed source, paternalistic and so on. I agree there is a paternalistic element, but there's also a PR and a legal element, right? If DALY 2 was accessible to everyone... | 500 | The Man behind Stable Diffusion: billion parameter language models or 540 billion parameters models are really usable for the vast majority of humanity. You mentioned this open source, closed source, paternalistic and so on. I agree there is a paternalistic element, but there's also a PR and a legal element, right? If ... | [0.0023313448764383793, 0.0016802485333755612, 0.015766331925988197, -0.019294854253530502, -0.0038190647028386593, 0.012972918339073658, -0.013448989018797874, 0.012076785787940025, -0.03732952103018761, -0.026981990784406662, 0.0058843703009188175, 0.02773810178041458, -0.004099106416106224, 0.012160798534750938, -0.... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | give grants in terms of hardware and what to run, you do pay people to actually work part time or full time. Can you specify a little bit of what just being an employee at stability AI means? Yeah. So different people need different things. We come from all diverse backgrounds. Some of them needed the equivalent to the... | 500 | The Man behind Stable Diffusion: give grants in terms of hardware and what to run, you do pay people to actually work part time or full time. Can you specify a little bit of what just being an employee at stability AI means? Yeah. So different people need different things. We come from all diverse backgrounds. Some of ... | [0.014956444501876831, -0.008012876845896244, 0.020039135590195656, -0.031079407781362534, -0.008248957805335522, 0.008533644489943981, -0.007943441160023212, 0.023302612826228142, -0.013796868734061718, -0.019900264218449593, 0.018497664481401443, 0.029412951320409775, 0.010415351018309593, 0.01609518937766552, -0.009... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | keeps you like there's clearly a pull, right? There's clearly demands coming with any money that flows in. It's clearly attractive to sort of keep your, let's say leading position to attract more researchers and so on. How do you prevent yourself from, let's say, succumbing to that pull of going close to or going profi... | 500 | The Man behind Stable Diffusion: keeps you like there's clearly a pull, right? There's clearly demands coming with any money that flows in. It's clearly attractive to sort of keep your, let's say leading position to attract more researchers and so on. How do you prevent yourself from, let's say, succumbing to that pull... | [-0.008751912973821163, -0.006485481280833483, 0.016625147312879562, 0.0012822772841900587, 0.0061193653382360935, 0.011485578492283821, -0.007908103056252003, 0.008479941636323929, -0.013772930949926376, -0.02948453277349472, 0.012678069993853569, 0.024310095235705376, 0.0021287023555487394, 0.009979273192584515, -0.0... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | few extra things in there that I can't discuss right now. But we really kind of laid it out to be the right company at the right time to coordinate this all. And then hopefully as this goes, this becomes an independent, more decentralized thing. Originally, we wanted to be web3 with tokens and all that, but you don't n... | 500 | The Man behind Stable Diffusion: few extra things in there that I can't discuss right now. But we really kind of laid it out to be the right company at the right time to coordinate this all. And then hopefully as this goes, this becomes an independent, more decentralized thing. Originally, we wanted to be web3 with tok... | [0.014225137419998646, -0.00437112245708704, 0.001992209814488888, -0.015205705538392067, -0.003880838630720973, 0.020315706729888916, -0.010993407107889652, 0.015205705538392067, -0.02483184263110161, -0.006939933635294437, 0.010579083114862442, 0.027221113443374634, 0.0065048933029174805, -0.00800336617976427, -0.000... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | had have been from non-traditional backgrounds. I don't know if you've interviewed the Eleuther AI founders, none of them have a computer science degree, you know? And yet they kind of managed to achieve such great things. Now obviously there's conjecture for alignment and we're pushing some of the capability stuff the... | 500 | The Man behind Stable Diffusion: had have been from non-traditional backgrounds. I don't know if you've interviewed the Eleuther AI founders, none of them have a computer science degree, you know? And yet they kind of managed to achieve such great things. Now obviously there's conjecture for alignment and we're pushing... | [-0.008347461931407452, -0.012198553420603275, 0.01430601254105568, -0.013859807513654232, 0.010468651540577412, 0.010544163174927235, -0.026786018162965775, 0.003662311704829335, -0.012809510342776775, -0.017834462225437164, 0.016406606882810593, 0.026030901819467545, 0.012315252795815468, 0.008038551546633244, 0.0297... |
The Man behind Stable Diffusion | https://www.youtube.com/watch?v=YQ2QtKcK2dA | sounds a bit fuzzy, but I think it's really important and people don't pay enough attention to it. Wise words. So actually, maybe we should mention one of the projects we have, 7cups.com. It's something that we help kind of accelerate. You can go and you can chat to someone, so you don't even have the pressure of talki... | 229 | The Man behind Stable Diffusion: sounds a bit fuzzy, but I think it's really important and people don't pay enough attention to it. Wise words. So actually, maybe we should mention one of the projects we have, 7cups.com. It's something that we help kind of accelerate. You can go and you can chat to someone, so you don'... | [-0.022049663588404655, -0.005502276588231325, 0.02695710025727749, -0.039827290922403336, -0.004498482681810856, 0.027376191690564156, 0.0010654745856299996, -0.0042618983425199986, -0.024794042110443115, -0.012234792113304138, -0.005144020076841116, 0.035365983843803406, 0.0016865086508914828, -0.010200166143476963, ... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | Hello, this video starts out with a review of the drama around the public demo of the Galactica model, and then goes into a paper review. If you're not in the mood for any drama, skip ahead about 16 minutes and you'll be fine. Hello there, Galactica is a model a language model by Meta AI that is trained specifically on... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): Hello, this video starts out with a review of the drama around the public demo of the Galactica model, and then goes into a paper review. If you're not in the mood for any drama, skip ahead about 16 minutes and you'll be fine. Hello there, Galactica ... | [-0.015514805912971497, -0.012318593449890614, 0.012568614445626736, -0.027705010026693344, 0.000586619833484292, -0.0015322222607210279, -0.03297571837902069, 0.029813293367624283, -0.023312753066420555, -0.03867889940738678, 0.0036523311864584684, 0.018501540645956993, -0.006557978689670563, 0.0025407522916793823, -0... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | And people tried it as it wasn't intended. A lot of funny stuff was done. And also someone might have entered a bad word. Oh, no, oh, no. But people pretty quickly started obviously to complain, the professional complainers, and the people who think they know what's good for you. Obviously, we're all over this. So Mich... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): And people tried it as it wasn't intended. A lot of funny stuff was done. And also someone might have entered a bad word. Oh, no, oh, no. But people pretty quickly started obviously to complain, the professional complainers, and the people who think ... | [-0.011654321104288101, 0.002725827507674694, 0.010759935714304447, -0.023076511919498444, -0.0014712982811033726, -0.0034700108226388693, -0.03247097134590149, 0.009763139300048351, -0.018229078501462936, -0.024291783571243286, 0.011080821976065636, 0.0017230575904250145, -0.006086600478738546, 0.012999312952160835, 0... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | without verification. People are just gonna disregard it. People are just gonna be like the language model says I must do something so I'll do something. Look at me. I just write a paper. Oh no, it to language model says something that I must submit this Grady Booge says, Galactica is a little more than statistical non... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): without verification. People are just gonna disregard it. People are just gonna be like the language model says I must do something so I'll do something. Look at me. I just write a paper. Oh no, it to language model says something that I must submit ... | [-0.017302827909588814, -0.010797074995934963, 0.01398118119686842, -0.016257507726550102, -0.004728018771857023, 0.0036276797764003277, -0.03603609651327133, 0.02567915804684162, -0.025802945718169212, -0.029076455160975456, 0.01793552376329899, 0.013039015233516693, -0.02001241408288479, 0.012138113379478455, 0.00659... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | a neutral way discusses pros and cons of something is just so out of their world. Because in the past, all they always had to do in the recent years is say a word like harmful or problematic. And if they said it long enough and loud enough, magically, things would go their way people would take down things, people woul... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): a neutral way discusses pros and cons of something is just so out of their world. Because in the past, all they always had to do in the recent years is say a word like harmful or problematic. And if they said it long enough and loud enough, magically... | [-0.0013511648867279291, -0.013244925998151302, 0.012858878821134567, -0.03245603293180466, 0.008177179843187332, 0.007187495473772287, -0.025816023349761963, 0.028665753081440926, -0.03127683699131012, -0.039362769573926926, 0.014255667105317116, 0.02029905840754509, -0.01819334737956524, 0.005387112032622099, 0.01037... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | like dung. We know what's good for you. I do not wish to despoil the word, but it will happen. This is 500 years ago, and the exact same conversation repeats and repeats and repeats. It will happen magically, right? To hand it out about to all and sundry is langurus? Would you have ploughmen and weavers debating the go... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): like dung. We know what's good for you. I do not wish to despoil the word, but it will happen. This is 500 years ago, and the exact same conversation repeats and repeats and repeats. It will happen magically, right? To hand it out about to all and su... | [-0.017018374055624008, -0.0036064735613763332, 0.004846961237490177, -0.027332540601491928, 0.00597942853346467, 0.0016629850724712014, -0.026802893728017807, 0.01825886219739914, -0.021687624976038933, -0.035820819437503815, -0.0021603996865451336, 0.017352888360619545, -0.017617709934711456, 0.012363061308860779, 0.... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | scientific community, and much to the detriment of, I guess, Meta itself. Although let me say, what Meta should have done, they did so much right. They open sourced the model, they made the model available via a demo. And now the only thing left to do is to actually have a pair of balls to tell the people who come and ... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): scientific community, and much to the detriment of, I guess, Meta itself. Although let me say, what Meta should have done, they did so much right. They open sourced the model, they made the model available via a demo. And now the only thing left to d... | [-0.02092656120657921, 0.0011006827699020505, 0.006529358681291342, -0.03144419565796852, 0.0033224313519895077, 0.0018871274078264832, -0.03369991481304169, 0.024024778977036476, -0.01780116558074951, -0.035249024629592896, 0.0018140882020816207, 0.02001611888408661, 0.00021476477559190243, 0.0042838300578296185, 0.00... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | includes over 48 million papers, this is their data set, textbooks, lecture notes, millions of compounds of protein, scientific websites, encyclopedias, and more. Our corpus is high quality and highly curated. And it is a lot smaller than the usual corpora of the large language models. They format all of this into a co... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): includes over 48 million papers, this is their data set, textbooks, lecture notes, millions of compounds of protein, scientific websites, encyclopedias, and more. Our corpus is high quality and highly curated. And it is a lot smaller than the usual c... | [-0.01603529416024685, 0.008647019043564796, 0.0026816704776138067, -0.0401429645717144, 0.018922194838523865, 0.001963365823030472, -0.03286414593458176, 0.017280355095863342, -0.024052942171692848, -0.0401429645717144, 0.009228503331542015, 0.025188546627759933, -0.026652520522475243, 0.015953201800584793, -0.0039335... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | text, yada, yada, yada, then the start ref token. Then here is the citation as text form, not as like some reference form, the title of the paper and the author name. And then here end ref. So in this way, you can just feed it into a language model and have the language model, if necessary, predict the reference from a... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): text, yada, yada, yada, then the start ref token. Then here is the citation as text form, not as like some reference form, the title of the paper and the author name. And then here end ref. So in this way, you can just feed it into a language model a... | [-0.012261525727808475, 0.007014264818280935, -0.001579721225425601, -0.013679159805178642, -0.0004980195080861449, -0.002628669608384371, -0.025933967903256416, 0.012960265390574932, -0.026337087154388428, -0.02851392701268196, 0.007256135810166597, 0.020491866394877434, -0.026780517771840096, 0.028621425852179527, -0... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | if you are a human, you have a pen, and you were to calculate these things, you were to calculate this average, and someone would ask you, please write down your steps. What you would write down is okay, the average is calculated as such, I'm going to add the first numbers, gonna add the third, add the fourth number, t... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): if you are a human, you have a pen, and you were to calculate these things, you were to calculate this average, and someone would ask you, please write down your steps. What you would write down is okay, the average is calculated as such, I'm going t... | [0.0036934674717485905, 0.016192903742194176, 0.007393565960228443, -0.02519780397415161, -0.009787357412278652, -0.014468844048678875, -0.023659411817789078, -0.005782232619822025, -0.02615266665816307, -0.03906985744833946, 0.010191848501563072, 0.019243163987994194, -0.04556823894381523, 0.01989300176501274, 0.00262... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | really the average of something. Well, here in here, the language model is just going to do language modeling, it's going to predict the next tokens. And if we do it, you know, cleanly enough, it has a chance of actually getting the correct answer. If we really do it step by step, like, you know, single digit addition,... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): really the average of something. Well, here in here, the language model is just going to do language modeling, it's going to predict the next tokens. And if we do it, you know, cleanly enough, it has a chance of actually getting the correct answer. I... | [-0.008711314760148525, 0.009649248793721199, -0.0006967027438804507, -0.021093392744660378, -0.006582407280802727, -0.0030719025526195765, -0.01784098893404007, 0.004581706132739782, -0.025357956066727638, -0.04113414138555527, 0.0018893633969128132, 0.03519614040851593, -0.03225413337349892, 0.022051570937037468, -0.... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | including writing a Python program, executing a Python program, and so on, a description of when the work is done, and then the the answer right here, most, most things that we're going to find in training data does not contain any of this stuff in between right here. And if it does contain it, it contains it in a very... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): including writing a Python program, executing a Python program, and so on, a description of when the work is done, and then the the answer right here, most, most things that we're going to find in training data does not contain any of this stuff in b... | [-0.005122673232108355, 0.007892054505646229, 0.0019303815206512809, -0.016916418448090553, 0.00611173827201128, 0.004536055494099855, -0.020722612738609314, 0.019358385354280472, -0.026275016367435455, -0.028867049142718315, -0.002291901735588908, 0.03380555287003517, -0.031022528186440468, 0.031622786074876785, -0.00... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | biology and a little bit to graph algorithms and a little bit here. Usually, authors have their topics, and therefore, also that the names of the authors to be available allows the language model to learn to associate these names with given, with given topical textual topical things in the text. And that's why it's als... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): biology and a little bit to graph algorithms and a little bit here. Usually, authors have their topics, and therefore, also that the names of the authors to be available allows the language model to learn to associate these names with given, with giv... | [-0.011348492465913296, -0.0023687942884862423, -0.005362026859074831, -0.017837224528193474, 0.006638053804636002, 0.002818457782268524, -0.023104228079319, 0.019561218097805977, -0.021068014204502106, -0.040588509291410446, 0.0029830518178641796, 0.01680554263293743, -0.031004732474684715, 0.023172101005911827, 0.002... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | right here, some noteworthy things is no biases. This, it seems like the if you make your models large enough, then you get away with essentially streamlining more and more, you know, with the small models, we have to have adapters and this and the convolution and the weight tying and whatnot. And the larger the models... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): right here, some noteworthy things is no biases. This, it seems like the if you make your models large enough, then you get away with essentially streamlining more and more, you know, with the small models, we have to have adapters and this and the c... | [-0.026247555390000343, -0.0010607565054669976, 0.012052449397742748, -0.024147188290953636, 0.005427126307040453, 0.011396965011954308, 0.005575138609856367, 0.008930089883506298, -0.021567540243268013, -0.033662278205156326, 0.012263895943760872, 0.02039753645658493, -0.016943911090493202, 0.01895970106124878, -0.003... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | at play, a quality factor, the curated nature of the corpus enables more value per token to be extracted, or a modality factor, the nature of scientific data enables more value of token, more value per token to be extracted. These two things, they're very similar, but essentially they say higher quality plus the nature... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): at play, a quality factor, the curated nature of the corpus enables more value per token to be extracted, or a modality factor, the nature of scientific data enables more value of token, more value per token to be extracted. These two things, they're... | [-0.0007033970905467868, 0.002017322229221463, 0.007931705564260483, -0.014804014936089516, -0.0026811640709638596, 5.381084520195145e-06, -0.021146627143025398, 0.016386229544878006, -0.020981527864933014, -0.03695500269532204, 0.004189426079392433, 0.02698018029332161, -0.02138051949441433, 0.012767774984240532, -0.0... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | available, which is pretty cool to see a like that much of a significant boost over publicly available and proprietary models. Now, naturally, it's going to be, let's say, expected if you train on scientific text that it's going to be better on scientific text. But it's still cool that it's not just like a 2% gain, it'... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): available, which is pretty cool to see a like that much of a significant boost over publicly available and proprietary models. Now, naturally, it's going to be, let's say, expected if you train on scientific text that it's going to be better on scien... | [-0.010188866406679153, -0.0010742006124928594, 0.013178247958421707, -0.022472379729151726, 0.009474466554820538, 0.0035962744150310755, -0.030989689752459526, 0.01650748960673809, -0.027577217668294907, -0.03268205374479294, -0.011028112843632698, 0.028686964884400368, -0.027410754933953285, 0.01229738537222147, 0.00... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | always find all the relevant things. But, or maybe humans disagree what what is relevant. I think the last years of reviews at machine learning conferences have shown, well, I guess all of scientific review has shown that humans can disagree quite heavily what should be cited. The last investigation is into toxicity an... | 500 | Galactica: A Large Language Model for Science (Drama & Paper Review): always find all the relevant things. But, or maybe humans disagree what what is relevant. I think the last years of reviews at machine learning conferences have shown, well, I guess all of scientific review has shown that humans can disagree quite he... | [-0.002567573683336377, -0.002952114213258028, 0.008480309508740902, -0.034166257828474045, 0.008983955718576908, -0.001670028897933662, -0.019846376031637192, 0.023807482793927193, -0.026475446298718452, -0.02670685015618801, 0.006840057205408812, 0.027618858963251114, -0.01766844652593136, 0.0018597468733787537, -0.0... |
Galactica: A Large Language Model for Science (Drama & Paper Review) | https://www.youtube.com/watch?v=ZTs_mXwMCs8 | of the paper that were actually written by Galactica itself. I hear that the part of the abstract may be written by Galactica, although I don't know. And I don't know if the authors will ever will ever lift that secret. Let's hope they don't because I like the mystery. Alright, this was it for me. Sorry for the bit lon... | 139 | Galactica: A Large Language Model for Science (Drama & Paper Review): of the paper that were actually written by Galactica itself. I hear that the part of the abstract may be written by Galactica, although I don't know. And I don't know if the authors will ever will ever lift that secret. Let's hope they don't because ... | [-0.013196793384850025, -0.002530730562284589, 0.0005292674177326262, -0.038904305547475815, 0.006669021677225828, 0.009147624485194683, -0.030025731772184372, 0.011131851002573967, -0.015470246784389019, -0.015429889783263206, 0.015012865886092186, 0.027604300528764725, -0.008212682791054249, 0.008488456718623638, 0.0... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | A lot of text to video models have recently come out, but not only that, a lot of other stuff has happened too, such as multiplayer stable diffusion and OpenAI is looking for even more money from Microsoft. Stay tuned. This is ML News. Hello everyone. As you can see, I'm not in my usual setting. I'm actually currently ... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: A lot of text to video models have recently come out, but not only that, a lot of other stuff has happened too, such as multiplayer stable diffusion and OpenAI is looking for even more money from Microsoft. Stay tuned. T... | [-0.038882844150066376, -0.003069056896492839, 0.010361331515014172, -0.004219735506922007, -0.00011609060311457142, 0.022950906306505203, -0.008446434512734413, 0.019413569942116737, -0.016948575153946877, -0.04264300689101219, 0.023480113595724106, 0.016781456768512726, -0.0015171202830970287, 0.00901742186397314, -0... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | a paper so as to make it reproducible. And given that it is a standard, it just incorporates with the whole rest of the scientific ecosystem. So definitely a big plus for anyone who does work in research. The Wall Street Journal writes Microsoft in advance talks to increase investment in OpenAI. This article essentiall... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: a paper so as to make it reproducible. And given that it is a standard, it just incorporates with the whole rest of the scientific ecosystem. So definitely a big plus for anyone who does work in research. The Wall Street... | [-0.01619846001267433, -0.022383831441402435, 0.006088291294872761, -0.025989653542637825, 0.0014839343493804336, 0.012197385542094707, -0.015934959053993225, 0.018611587584018707, -0.023840028792619705, -0.03034437634050846, 0.02035902440547943, 0.020553184673190117, 0.023701343685388565, 0.005776248872280121, -0.0102... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | stack. They will then do that, update the data set and by agreeing to these terms, if you download the data set, you essentially agree to always download the newest version and use the newest version of the data set such as to propagate that removal of that code. Now as I understand it, I'm not a lawyer, this is not le... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: stack. They will then do that, update the data set and by agreeing to these terms, if you download the data set, you essentially agree to always download the newest version and use the newest version of the data set such... | [-0.02536025084555149, -0.0030803135596215725, -0.009668420068919659, -0.012025712989270687, 0.0005523806321434677, 0.015466657467186451, -0.008338483981788158, -0.014537813141942024, -0.01708509773015976, -0.05516769364476204, 0.01441115327179432, 0.038645531982183456, -0.011153162457048893, 0.003126051975414157, 0.00... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | the video simply drops out of the context. But as long as you feed into the side input more and more text that you want to be produced, you can see that the video keeps changing keeps adapting and keeps being faithful to the currently in focus part of the prompt. What's interesting is that the training data seems to be... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: the video simply drops out of the context. But as long as you feed into the side input more and more text that you want to be produced, you can see that the video keeps changing keeps adapting and keeps being faithful to... | [-0.044288016855716705, -0.006606158800423145, -0.002666816348209977, -0.004112556576728821, 0.02416084334254265, 0.026657873764634132, -0.010420306585729122, 0.0019002138869836926, -0.028180789202451706, -0.0456051342189312, -0.0020391284488141537, 0.04486425593495369, -0.007422496099025011, -0.01660114713013172, 0.00... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | the demo and the code that they put on GitHub, they simply calls some API where the model is actually stored. This is a neat tool not directly related to machine learning. But if you've ever wondered what like the difference between a Bfloat 16 and an FP 16 is I never knew. But Charlie Blake has a very cool tool on a b... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: the demo and the code that they put on GitHub, they simply calls some API where the model is actually stored. This is a neat tool not directly related to machine learning. But if you've ever wondered what like the differ... | [-0.0196006391197443, 0.007913120090961456, 0.0034064308274537325, -0.016073981299996376, 0.005351921543478966, 0.0012651155702769756, -0.009188255295157433, 0.008962374180555344, -0.01900314725935459, -0.04132893308997154, 0.01090786512941122, 0.01348728034645319, -0.01761871576309204, -0.011949832551181316, 0.0052535... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | bring up a server. This is based on the lightning apps framework, which is open source and it's kind of an easy way to bring together all the components you need to deploy machine learning things. And this repository is essentially a specification on how to pull up a stable diffusion server. So if you want to deploy st... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: bring up a server. This is based on the lightning apps framework, which is open source and it's kind of an easy way to bring together all the components you need to deploy machine learning things. And this repository is ... | [-0.028174415230751038, 0.010677151381969452, -0.013128354214131832, -0.019609622657299042, 0.00861525721848011, 0.009365037083625793, -0.0006700255325995386, 0.008377346210181713, -0.006675923243165016, -0.04014205187559128, -0.003408253425732255, 0.0294721107929945, -0.01750447228550911, 0.009163172915577888, -0.0013... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | and something like a torch. So this does two things. First of all, it optimizes your computation graph. If your computation graph contains a lot of like little operations that could be used together into something that's really optimal for a given hardware, or just that can be expressed in a smarter way, then a graph o... | 500 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: and something like a torch. So this does two things. First of all, it optimizes your computation graph. If your computation graph contains a lot of like little operations that could be used together into something that's... | [-0.0316886305809021, 0.0001837763556977734, -0.0165450107306242, -0.02217918075621128, 0.02016289159655571, 0.006842513103038073, 0.0008763171499595046, -0.004465151112526655, -0.01873289979994297, -0.03635040670633316, 0.030029838904738426, 0.026926755905151367, -0.003911029081791639, -0.0036107306368649006, 0.006785... |
[ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming | https://www.youtube.com/watch?v=TOo-HnjjuhU | set of prompts that have been entered by real users into stable diffusion and the corresponding images that they got out. Public prompts, that's public prompts.art in your browser is a database of free prompts and free models. These models are mostly trained using Dreambooth. If you're looking for inspiration for promp... | 373 | [ML News] Multiplayer Stable Diffusion | OpenAI needs more funding | Text-to-Video models incoming: set of prompts that have been entered by real users into stable diffusion and the corresponding images that they got out. Public prompts, that's public prompts.art in your browser is a database of free prompts and free m... | [-0.038437049835920334, -0.01611006259918213, -0.0038784798234701157, -0.00533335329964757, 0.014484862796962261, 0.006646288093179464, 0.004928827751427889, 0.003885576967149973, -0.023249588906764984, -0.043149419128894806, 0.011858993209898472, 0.025023825466632843, -0.015868766233325005, 0.0031084613874554634, 0.00... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | The new responsible AI licenses that models like stable diffusion or bloom have are stupid. They conflict with open source principles. In fact, they're distinctly not open source, and they have a glaring legal loophole in them. So join me as we'll explore the fun world of model licensing. So first things first, I am no... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): The new responsible AI licenses that models like stable diffusion or bloom have are stupid. They conflict with open source principles. In fact, they're distinctly not open source, and they have a glaring legal loophole in the... | [-0.0029976018704473972, -0.00494997575879097, 0.004761205520480871, -0.013808195479214191, -0.014947808347642422, 0.00015239266213029623, -0.007928350940346718, 0.014528319239616394, -0.012291042134165764, -0.04323538392782211, 0.019771937280893326, 0.006904097739607096, -0.007417972199618816, 0.013095063157379627, -0... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | it distill it sell it and so on, you must pass on you must enforce continuously these usage restrictions. So even if you take the model and you fine tune it on your own data or something like this, then you may keep that private but you may still not use it for any of these things. So much like a copy left license that... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): it distill it sell it and so on, you must pass on you must enforce continuously these usage restrictions. So even if you take the model and you fine tune it on your own data or something like this, then you may keep that priv... | [0.01974165253341198, -0.02294337935745716, -0.005568068940192461, -0.029780255630612373, -0.011828213930130005, 0.009038937278091908, -0.014876146800816059, 0.000757613917812705, -0.016246318817138672, -0.0426151268184185, 0.02115376852452755, 0.01312847901135683, 0.014212032780051231, -0.00564147112891078, -0.0073402... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | to make or publish drawings of Muhammad and so on. He says it's not clear these would be enforceable free software licenses are based on copyright law and trying to impose usage condition that way is stretching what copyright law permits in a dangerous way. Would you like books to carry a license condition about how yo... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): to make or publish drawings of Muhammad and so on. He says it's not clear these would be enforceable free software licenses are based on copyright law and trying to impose usage condition that way is stretching what copyright... | [0.0012299040099605918, -0.02347606234252453, -0.011841602623462677, -0.025768430903553963, -0.007063533179461956, 0.0023096303921192884, -0.020051319152116776, -0.0068149627186357975, -0.019692273810505867, -0.04001978039741516, 0.027577469125390053, 0.027756990864872932, 0.006939247716218233, 0.009783994406461716, -0... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | it and keep it for yourself whatever you want. But don't then also go out and say, Oh, we are free, we are open, we are for everyone. No, you are not. And it takes no further to look than actually to look at the license itself and some of these usage restrictions. For example, you may not use this model to provide medi... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): it and keep it for yourself whatever you want. But don't then also go out and say, Oh, we are free, we are open, we are for everyone. No, you are not. And it takes no further to look than actually to look at the license itsel... | [0.010649638250470161, -0.005748981609940529, 0.0051881056278944016, -0.03993438929319382, -0.008237870410084724, 0.007915366441011429, -0.016139214858412743, -0.003687761491164565, -0.020962750539183617, -0.0462162047624588, 0.018018150702118874, 0.0021506098564714193, 0.010039685294032097, 0.01009577326476574, -0.017... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | certainly not accessible to the whole world. It is very much we know what's good for you. And you play, you do not have the authority to decide that for yourself, you come to us, and then we decide if it's good enough. It's even more the rest of the licenses, essentially, it's a copy paste of rather standard terms of p... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): certainly not accessible to the whole world. It is very much we know what's good for you. And you play, you do not have the authority to decide that for yourself, you come to us, and then we decide if it's good enough. It's e... | [-0.0032129338942468166, -0.011941754259169102, -0.0016248415922746062, -0.023841509595513344, -0.0098417978733778, 0.014881694689393044, -0.02053757756948471, -0.0015417183749377728, -0.013278727419674397, -0.040571168065071106, 0.023897508159279823, 0.005463387817144394, -0.005942877847701311, 0.003475428558886051, -... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | running it, then reasonable effort would certainly include that you point your download script to the new version. If you fine tuned your model a little bit to do something, then I guess it's up to a judge to decide whether it's reasonable effort for you to redo that fine tuning with the new version of the base model, ... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): running it, then reasonable effort would certainly include that you point your download script to the new version. If you fine tuned your model a little bit to do something, then I guess it's up to a judge to decide whether i... | [-0.006234914064407349, -0.008602333255112171, 0.0007460390916094184, -0.02577856183052063, -0.008168661966919899, 0.014702169224619865, -0.008737411350011826, 0.005019212607294321, -0.041234321892261505, -0.029603401198983192, 0.008175770752131939, 0.00013596663484349847, -0.006295343860983849, 0.00923506636172533, -0... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | Okay, now that we got that out of the way, I want to come to the last part of this video. And here I want to say again, I am not a lawyer. This is my opinion. But in my opinion, this thing is drastically different from the open source licenses that we are used to not just in terms of the content of a containing usage r... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): Okay, now that we got that out of the way, I want to come to the last part of this video. And here I want to say again, I am not a lawyer. This is my opinion. But in my opinion, this thing is drastically different from the op... | [0.02683175541460514, -0.0194223765283823, -0.012087909504771233, -0.040288493037223816, 1.0514406312722713e-05, 0.0008793528540991247, -0.02131558209657669, -0.0005367202102206647, -0.02636866830289364, -0.03178950026631355, 0.02305896393954754, 0.009867819957435131, -0.01375638134777546, -0.006547901313751936, -0.011... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | copyright registrations on the work outputs of his AI algorithm. For example, here is an article by Clyde Schuman of Pearl Cohen that goes into detail of how this was again and again rejected the Copyright Office again concluded that the work lacks the required human authorship necessary to sustain a claim in copyright... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): copyright registrations on the work outputs of his AI algorithm. For example, here is an article by Clyde Schuman of Pearl Cohen that goes into detail of how this was again and again rejected the Copyright Office again conclu... | [0.0085425591096282, -0.024591809138655663, -0.00607396150007844, -0.027793586254119873, 0.01228245161473751, 0.006968575995415449, -0.02355594001710415, 0.007984265685081482, -0.03304019942879677, -0.025170281529426575, 0.01134747825562954, 0.01717928797006607, -0.0034641087986528873, 0.00019517142209224403, -0.019721... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | am fairly sure there is no copyright at all on models if they are simply trained by an algorithm like the training code for GPT or the training code for stable diffusion. So therefore, you can't simply say here is the license for the model. The reason that works with code, the reason you can simply put an MIT license f... | 500 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): am fairly sure there is no copyright at all on models if they are simply trained by an algorithm like the training code for GPT or the training code for stable diffusion. So therefore, you can't simply say here is the license... | [-0.006345650646835566, -0.0031030368991196156, 0.00226897862739861, -0.03039376065135002, -0.014488781802356243, 0.004544766154140234, -0.018805459141731262, 0.00871505681425333, -0.0077278041280806065, -0.04673449322581291, 0.02509663999080658, 0.009811247698962688, 0.0019881222397089005, 0.010199340060353279, -0.019... |
The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.) | https://www.youtube.com/watch?v=W5M-dvzpzSQ | again, severely restricts the distribution capabilities of these models and essentially centralizes an already relatively central system even more to institutions who can actually enforce such provisions, or at least can enforce the fact that you need to enter into the agreement, such as having a website with a little ... | 368 | The New AI Model Licenses have a Legal Loophole (OpenRAIL-M of BLOOM, Stable Diffusion, etc.): again, severely restricts the distribution capabilities of these models and essentially centralizes an already relatively central system even more to institutions who can actually enforce such provisions, or at least can enfo... | [0.0029587752651423216, -0.02014154940843582, 0.004139109514653683, -0.029301932081580162, -0.010113118216395378, 0.004802496172487736, -0.014665080234408379, 0.004957756958901882, -0.0228374395519495, -0.03881518170237541, 0.027424689382314682, 0.005056559108197689, 0.004583719652146101, 0.0062598297372460365, -0.0113... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | Hello, today we're talking about locating and editing factual associations in GPT by Kevin Meng, David Bao, Alex Andonian and Jonathan Belenkov. In this paper, the authors attempt to localize where in a forward pass through a language model an actual fact is located or where it is realized. For example, something like ... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): Hello, today we're talking about locating and editing factual associations in GPT by Kevin Meng, David Bao, Alex Andonian and Jonathan Belenkov. In this paper, the authors attempt to localize where in a forward pass through a l... | [-0.0014771006535738707, 0.034332338720560074, 0.004194464068859816, -0.016965316608548164, -0.028065750375390053, 0.022080354392528534, -0.0235532708466053, 0.009660991840064526, -0.008522829040884972, -0.019817419350147247, 0.013329892419278622, 0.01717955991625786, -0.012513093650341034, 0.011756549589335918, 0.0194... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | for thinking about your grandmother. Maybe if you pluck that neuron out of your brain you might forget that whole concept, which people think is sort of implausible. But what we're chasing here is sort of a weaker locality question. Like if you have some knowledge in a big neural network, can it be localized to a small... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): for thinking about your grandmother. Maybe if you pluck that neuron out of your brain you might forget that whole concept, which people think is sort of implausible. But what we're chasing here is sort of a weaker locality ques... | [-0.003021762240678072, 0.025373155251145363, 0.027123503386974335, -0.029742132872343063, -0.020246151834726334, 0.024105185642838478, 0.0003051480161957443, 0.006577587220817804, -0.030679328367114067, -0.031092796474695206, 0.011735601350665092, 0.028501730412244797, -0.009061841294169426, 0.010281572118401527, 0.00... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | a very practical question. And I think the really cool thing about Roam is that, like you said, on one side is the scientific question. But on the other side, we show that the insights that we get can yield a pretty useful model editor that seems to achieve generalization, specificity, and fluency preservation all pret... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): a very practical question. And I think the really cool thing about Roam is that, like you said, on one side is the scientific question. But on the other side, we show that the insights that we get can yield a pretty useful mode... | [-0.004610119853168726, 0.010032777674496174, 0.011162067763507366, -0.014308931306004524, -0.017352504655718803, 0.008607394061982632, -0.0038079109508544207, -0.0009545599459670484, -0.02834242582321167, -0.02403184212744236, 0.02063020132482052, 0.013399990275502205, -0.022957639768719673, 0.012112324126064777, 0.00... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | it would travel up the layers and get both the hidden signals from here. So you can see there is various paths this information can take. And the idea here is to figure out where in these hidden states, so in these bubbles right here, or this bubble or this bubble, where is the fact that Seattle should be the output of... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): it would travel up the layers and get both the hidden signals from here. So you can see there is various paths this information can take. And the idea here is to figure out where in these hidden states, so in these bubbles righ... | [-0.010646792128682137, 0.009422620758414268, 0.001921948860399425, -0.011689086444675922, -0.014857941307127476, 0.013430907391011715, -0.011213408783078194, 0.006460126489400864, -0.02952001430094242, -0.04160783067345619, 0.018131723627448082, 0.011423266492784023, -0.026386136189103127, 0.013696727342903614, 0.0034... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | you get these two clusters, you get early and early, so what they call an early site, which usually happens after the subject is done, and a late site, which usually happens right before you need to predict. So what's surprising, at least to me, is that these early sites here, exist, which indicates that the model is a... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): you get these two clusters, you get early and early, so what they call an early site, which usually happens after the subject is done, and a late site, which usually happens right before you need to predict. So what's surprisin... | [-0.00777561916038394, 0.010187308304011822, 0.008780489675700665, -0.02826111949980259, -0.008628026582300663, 0.015163150615990162, -0.011691149324178696, 0.015343334525823593, -0.027748288586735725, -0.025599945336580276, 0.0026906277053058147, 0.0188361257314682, -0.028080934658646584, 0.022619983181357384, 0.00476... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | we can actually do this activation copying for specifically the MLP and specifically the attention as well. And what we find is that the MLP corresponds to the early site and then the attention corresponds to the late site. And so the thing is the late site is interesting because, well, it's not exactly too surprising ... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): we can actually do this activation copying for specifically the MLP and specifically the attention as well. And what we find is that the MLP corresponds to the early site and then the attention corresponds to the late site. And... | [-0.030623575672507286, 0.0124813849106431, 0.012205247767269611, -0.036726199090480804, -0.021649127826094627, 0.01103166677057743, -0.012647067196667194, -0.00565390195697546, -0.016512982547283173, -0.021152080968022346, 0.02787601388990879, 0.0289115272462368, -0.030899712815880775, 0.0071174269542098045, 0.0036691... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | is specifically stored in these MLP layers in the middle of the network. So this experiment here is pretty interesting. As far as the way I understand it is the following. The top one, the top is sort of the baseline corrupted input condition. So that baseline corrupted input condition is what we had before, as the wha... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): is specifically stored in these MLP layers in the middle of the network. So this experiment here is pretty interesting. As far as the way I understand it is the following. The top one, the top is sort of the baseline corrupted ... | [-0.016712166368961334, 0.009508887305855751, 0.001059357076883316, -0.0339832678437233, -0.0344863086938858, 0.0012200509663671255, -0.0012497444404289126, 0.0041780415922403336, -0.032194674015045166, -0.017871957272291183, 0.03387147933244705, 0.022203704342246056, -0.024774806573987007, 0.015692109242081642, 0.0117... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | really strong causal effect for every state that you can see in the purple bars in the graph on the right. But then if you take the MLPs out of the picture, then it drops down to the green bars way below it. So somehow the MLPs at these early layers from about 10 to 20 are really important for this computation. If you ... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): really strong causal effect for every state that you can see in the purple bars in the graph on the right. But then if you take the MLPs out of the picture, then it drops down to the green bars way below it. So somehow the MLPs... | [-0.011769741773605347, 0.009085537865757942, 0.017763527110219002, -0.04140138626098633, -0.015191750600934029, 0.009816315956413746, -0.006285393610596657, -0.0054738083854317665, -0.02464970387518406, -0.034318458288908005, 0.008558535017073154, 0.02175469882786274, -0.019435884431004524, 0.017426244914531708, 0.001... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | the even though the model doesn't know yet that I'm going to ask it where the Space Needle is. So that means that essentially, if this hypothesis is correct, the model, once it sees a subject, whatever that means, it will retrieve kind of a whole bunch of knowledge from its different MLPs that are around about the subj... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): the even though the model doesn't know yet that I'm going to ask it where the Space Needle is. So that means that essentially, if this hypothesis is correct, the model, once it sees a subject, whatever that means, it will retri... | [-0.015721525996923447, 0.020646341145038605, 0.006206755992025137, -0.03690905496478081, -0.02597704902291298, -0.0025334388483315706, 0.003940528724342585, -0.004004795104265213, -0.01451738178730011, -0.026166465133428574, 0.016858022660017014, 0.014355025254189968, -0.016546837985515594, 0.008476365357637405, -0.00... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | to be constrained as well. Yeah, so the old name for this is, you know, linear associated memory, it goes way back to the 1970s, right? When people like what can you use a single layer neural network for? And, and, you know, researchers in the 1970s thought of a lot of alternatives. But one of the leading hypothesis wa... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): to be constrained as well. Yeah, so the old name for this is, you know, linear associated memory, it goes way back to the 1970s, right? When people like what can you use a single layer neural network for? And, and, you know, re... | [-0.012329962104558945, 0.02613813988864422, 0.010699782520532608, -0.02543357014656067, -0.03133261203765869, 0.014505838043987751, -0.007992026396095753, -0.002165514510124922, -0.01568012125790119, -0.027975544333457947, 0.019714124500751495, 0.024673739448189735, -0.0107274129986763, 0.004372474271804094, 0.0015187... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | math here and set this up as a constrained optimization problem. And it turns out if you solve that, then you get a closed form, you get a closed form solution for a rank one update. So they get a closed form solution that's here for and it takes a rank one update that they can easily compute that they need to add to t... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): math here and set this up as a constrained optimization problem. And it turns out if you solve that, then you get a closed form, you get a closed form solution for a rank one update. So they get a closed form solution that's he... | [-0.0020746260415762663, 0.00887925922870636, 0.0006997052114456892, -0.010768462903797626, -0.01795443519949913, 0.009131153114140034, 0.001634686253964901, 0.00865535344928503, -0.021159086376428604, -0.03400567173957825, 0.0374482236802578, 0.018486211076378822, -0.012685655616223812, 0.008130574598908424, 0.0015358... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | how much balance there is. But if we're just setting up a single new fact to learn it's easiest to just say you know what the new model should just know this fact. Let's just like know this 100 percent and we might have to sacrifice a little bit of you know sort of increased error on old facts but there's so many other... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): how much balance there is. But if we're just setting up a single new fact to learn it's easiest to just say you know what the new model should just know this fact. Let's just like know this 100 percent and we might have to sacr... | [0.0004130800371058285, 0.030815772712230682, 0.008592065423727036, -0.022333862259984016, -0.017528362572193146, 0.002521509537473321, -0.022030936554074287, -0.005576580762863159, -0.021135929971933365, -0.023779641836881638, 0.011359700933098793, 0.011173815466463566, -0.0176247488707304, 0.016110122203826904, -0.00... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | can't do this for every single subject in the model. You can't always output Rome or always Paris always output those kinds of things. So we also want it to be specific. So those are the main two axes on which we measure the edit. What do you mean by specific? Specific as in entities that aren't related. It's like subj... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): can't do this for every single subject in the model. You can't always output Rome or always Paris always output those kinds of things. So we also want it to be specific. So those are the main two axes on which we measure the ed... | [-0.005568637512624264, 0.010519317351281643, 0.019367342814803123, -0.0065728179179131985, -0.03140346333384514, 0.016361823305487633, -0.02296273037791252, -0.0002043033455265686, -0.022358816117048264, -0.045138970017433167, 0.02044876664876938, 0.027119897305965424, -0.02574353665113449, 0.000812824466265738, -0.00... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | actual fluent text anymore? So those are the few those are a few of the questions that motivate the design of Counterfact which we talk about in the next section. So Counterfact is based on something that's very similar to ZSRE. It's actually called Parallel. It's a bunch of relations that some researchers use to analy... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): actual fluent text anymore? So those are the few those are a few of the questions that motivate the design of Counterfact which we talk about in the next section. So Counterfact is based on something that's very similar to ZSRE... | [-0.01519019901752472, 0.018812261521816254, 0.004576714243739843, -0.008030308410525322, -0.0066720349714159966, 0.021999115124344826, -0.029173044487833977, -0.004152034409344196, -0.008297049440443516, -0.03515365719795227, 0.01844724640250206, 0.02976268157362938, -0.022350089624524117, 0.002242730464786291, -0.001... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | data set, seeing, you know, does something make sense and so on. Now, we talked about budget before. Is it fair to assume that this data set has at least in part been also generated with the help of automated things like models or is being also evaluated with the help of automated heuristics? Ah, yeah. Okay. So this da... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): data set, seeing, you know, does something make sense and so on. Now, we talked about budget before. Is it fair to assume that this data set has at least in part been also generated with the help of automated things like models... | [-0.0028967533726245165, 0.009172763675451279, 0.022247064858675003, -0.023879623040556908, -0.018774420022964478, 0.016062714159488678, 0.002860435750335455, -0.0016861698823049664, -0.016463935375213623, -0.03439440578222275, 0.019078794866800308, 0.022219395264983177, -0.019480017945170403, 0.005703577771782875, -0.... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | what happens to the key when I have five words in front of the subject or 10 words or something like that. And usually, it doesn't change too much, but it helps with generalization. But then the value is a little bit more involved. And this is actually an interesting area for future research, because there are a few th... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): what happens to the key when I have five words in front of the subject or 10 words or something like that. And usually, it doesn't change too much, but it helps with generalization. But then the value is a little bit more invol... | [-0.013297504745423794, 0.012758417055010796, 0.01111350767314434, -0.02403779700398445, -0.02116266079246998, 0.015301806852221489, 0.004250501748174429, -0.009192142635583878, -0.023332836106419563, -0.03604978322982788, 0.017914310097694397, 0.03610507771372795, -0.025807112455368042, 0.014776541851460934, -0.003231... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | about the subject will change when we didn't want them to change. Like an example of this is, say, you wanted to change Mario Kart to a Microsoft product. If you make the update too strong, it'll actually think Mario Kart is no longer a game. It'll think it's a Microsoft Office productivity tool. And so this loss term ... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): about the subject will change when we didn't want them to change. Like an example of this is, say, you wanted to change Mario Kart to a Microsoft product. If you make the update too strong, it'll actually think Mario Kart is no... | [-0.008921917527914047, 0.009300976060330868, 0.008255054242908955, -0.025312693789601326, -0.020974578335881233, 0.0009739701054058969, 0.0014986855676397681, -0.0025305673480033875, -0.027404535561800003, -0.03695119544863701, 0.023248929530382156, 0.02222406677901745, -0.02541096694767475, 0.01168764103204012, 0.007... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | could be stored in any of them. And if any one of them kind of overrides the previous ones, then we'll get, you know, the new fact being expressed. And so specifically, what we do is we just go to the causal traces, and we see where the causal effect peaks. And then, you know, we run an experiment that shows that this ... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): could be stored in any of them. And if any one of them kind of overrides the previous ones, then we'll get, you know, the new fact being expressed. And so specifically, what we do is we just go to the causal traces, and we see ... | [-0.009949378669261932, 0.022307774052023888, 0.005556055810302496, -0.028281578794121742, -0.016027620062232018, 0.015763046219944954, -0.00843155849725008, 0.0018467968329787254, -0.011606447398662567, -0.027209356427192688, 0.010715250857174397, 0.023073645308613777, -0.01902148500084877, 0.013792665675282478, 0.012... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | the residual stream, which also, it's also fascinating that you're saying, well, if I edit, if I edit some game to now be a Microsoft game, then all of a sudden, it might think, you know, it's a Microsoft Office product or something like this, it's Super Mario is no longer a game, which kind of means that sort of these... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): the residual stream, which also, it's also fascinating that you're saying, well, if I edit, if I edit some game to now be a Microsoft game, then all of a sudden, it might think, you know, it's a Microsoft Office product or some... | [-0.006845523603260517, 0.007838227786123753, 0.008072616532444954, -0.030939284712076187, -0.02725800685584545, 0.000183977754204534, -0.00034662787220440805, -0.0064904941245913506, -0.0317113883793354, -0.03162866458296776, 0.03943242132663727, 0.01779974065721035, -0.01737232506275177, 0.006852417252957821, -0.0049... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | something? It's a loaded question. If we answer this question, then somebody is going to boo us. So I think that so here's what it seems like to me. There's positive surprises and some negative surprises. So on the positive side, it was really, really surprising to see that a rank one update in a single layer in a matr... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): something? It's a loaded question. If we answer this question, then somebody is going to boo us. So I think that so here's what it seems like to me. There's positive surprises and some negative surprises. So on the positive sid... | [-0.008694902993738651, 0.007333132438361645, -0.0033227193634957075, -0.017308099195361137, -0.018274955451488495, -0.0011949533363804221, -0.014121557585895061, -0.001346450299024582, -0.008041253313422203, -0.035950735211372375, 0.025328926742076874, 0.02703113853931427, -0.013801541179418564, 0.012521477416157722, ... |
ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview) | https://www.youtube.com/watch?v=_NMQyOu2HTo | you could also imagine, well, people are constrained to live forward in time. So the way we must think about language must also be, you know, so so you have this debate about what is what is the best way to think about it. And, and so, so, so yeah, you there's that there's that movie arrival, I sort of imagined that ma... | 500 | ROME: Locating and Editing Factual Associations in GPT (Paper Explained & Author Interview): you could also imagine, well, people are constrained to live forward in time. So the way we must think about language must also be, you know, so so you have this debate about what is what is the best way to think about it. And,... | [0.004677511285990477, 0.0046707517467439175, 0.012809081003069878, -0.0279298797249794, -0.002321857027709484, 0.015168114565312862, -0.010044482536613941, 0.010504121892154217, -0.019142646342515945, -0.024455543607473373, 0.008347870782017708, 0.03752823919057846, -0.007239327300339937, -0.0006391866481862962, 0.010... |
Is Stability turning into OpenAI? | https://www.youtube.com/watch?v=igS2Wy8ur5U | stability AI has a few growing pains in the recent weeks, they found themselves in multiple controversies. And we're going to look at them in detail today. Yahoo Finance writes stability AI, the startup behind stable diffusion raises 101 million US dollars. Now I've done previously a video on stable diffusion, which is... | 500 | Is Stability turning into OpenAI?: stability AI has a few growing pains in the recent weeks, they found themselves in multiple controversies. And we're going to look at them in detail today. Yahoo Finance writes stability AI, the startup behind stable diffusion raises 101 million US dollars. Now I've done previously a ... | [0.009132095612585545, -0.019609974697232246, 0.006176180671900511, -0.006440530531108379, -0.0019465782679617405, 0.006787275895476341, -0.008191421627998352, 0.021807169541716576, -0.011329291388392448, -0.03361709788441658, 0.01598460040986538, 0.008740720339119434, 0.003043459728360176, 0.00878191739320755, 0.00476... |
Is Stability turning into OpenAI? | https://www.youtube.com/watch?v=igS2Wy8ur5U | is this is my thumbnail. Source Reddit, I guess I've posted it on Reddit. I'm not sure. But I guess the comp it's a compliment since it's a good thumbnail. Well, this all started with posts on Reddit from former moderators saying, Hello, I'm an ex moderator of the subreddit and discord and I've been here since the begi... | 500 | Is Stability turning into OpenAI?: is this is my thumbnail. Source Reddit, I guess I've posted it on Reddit. I'm not sure. But I guess the comp it's a compliment since it's a good thumbnail. Well, this all started with posts on Reddit from former moderators saying, Hello, I'm an ex moderator of the subreddit and discor... | [0.010117251425981522, -0.020655488595366478, -0.0009090248495340347, -0.020411044359207153, -0.025680163875222206, 0.009743795730173588, -0.012663539499044418, 0.003434093901887536, -0.004512022249400616, -0.025802385061979294, 0.0032762240152806044, 0.003476531943306327, 0.00986601784825325, -0.00832466408610344, -0.... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.