正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

sharing an easy way to test DJev (Diffusion Gemma Jev), the next evolution of the incredible jev concept and fully open sourced (model, code, api...) imagine jev but it also has: - native image-input support - native image as options!! - real-time video streaming input - as fast as... show more

David Villalón

1,989 subscribers

18,828 次观看 • 23 小时前 •via X (Twitter)

教育科学技术

Anya Rossi• Live Now

Private livecam show

21 条评论

David Villalón 的头像

David Villalón23 小时前

The idea follows the early implementation approach of @mmastrac and just adds a bit more of juice to each of the components to have it optimized it. Link to the Repo:

David Villalón 的头像

David Villalón23 小时前

I've implemented invites to manage scale for testing, each user gets 5 more and so on!

Pablo Senabre 的头像

Pablo Senabre22 小时前

on a mission

Carlos E. Perez 的头像

Carlos E. Perez17 小时前

Invite please?

David Villalón 的头像

David Villalón17 小时前

sent!

Lukas Haas 的头像

Lukas Haas22 小时前

love it!

vijay singh 的头像

vijay singh10 小时前

@shantanugoel

Mitansh 的头像

Mitansh21 小时前

djev with native image input and video streaming open sourced testing that at looks properly accessible

Niels Geo 的头像

Niels Geo21 小时前

You're saying phone and the held_object returns phone, something not right there?

David Villalón 的头像

David Villalón21 小时前

happy to give you access to test it! real-time video is very fragile

gkiwi🇺🇦 的头像

gkiwi🇺🇦15 小时前

Can you give me an invite code?

gairo peralta 的头像

gairo peralta13 小时前

Invite please

filipcodes 的头像

filipcodes15 小时前

hey, very cool, any chance I could have an invite?

Shalev Moyal 的头像

Shalev Moyal11 小时前

Invite please?

Leonardo Rezende 的头像

Leonardo Rezende16 小时前

That's really impressive. I'd love an invitation so I can try it out and share some feedback with you.

NZ 的头像

NZ17 小时前

Can i have access code, nice to try

Evan Leonard 的头像

Evan Leonard11 小时前

Invite please

A_staiiz 的头像

A_staiiz10 小时前

can you invite me? Thank you.

Bob 的头像

Bob12 小时前

may i please have access?

Guillermo Daniel 的头像

Guillermo Daniel16 小时前

Hi, can you invite me? Thank you.

Alexey Butochnikov 的头像

Alexey Butochnikov15 小时前

Invite please?

相关视频

Jev VS Fable 5.1 VS GPT-6 Astra: chess typesafe's new Jev V13 isn't an LLM. it doesn't chat, doesn't explain, doesn't write code — it only makes decisions. so we made it play blitz against frontier LLMs. the test: 5+0 blitz. every move is one API call. Jev V13 vs Fable 5.1: • fable outplayed it. by move 29 it was +16 in material and even promoted a second queen • but it kept burning 6-15 seconds per move on analysis. jev answered in ~2.6s • so fable didn't have enough time and lost Jev V13 vs Astra: • astra didn't bother winning on material. it mated jev in 18 moves. Qe1#, with 2:27 to spare

Jev VS Fable 5.1 VS GPT-6 Astra: chess typesafe's new Jev V13 isn't an LLM. it doesn't chat, doesn't explain, doesn't write code — it only makes decisions. so we made it play blitz against frontier LLMs. the test: 5+0 blitz. every move is one API call. Jev V13 vs Fable 5.1: • fable outplayed it. by move 29 it was +16 in material and even promoted a second queen • but it kept burning 6-15 seconds per move on analysis. jev answered in ~2.6s • so fable didn't have enough time and lost Jev V13 vs Astra: • astra didn't bother winning on material. it mated jev in 18 moves. Qe1#, with 2:27 to spare

AI/ML API

541,933 次观看 • 3 天前

you can make any open source model behave like jev with just a bit of inference engineering. it's shockingly easy. to prove it, we built a new endpoint we're calling deepseek-v4.1-flash-jev. see the demo below. here's how it's done: sglang (an inference engine) offers a scoring endpoint in addition to the normal generation one. in scoring mode, given an input & set of possible answers, it forces the model to produce probabilities for each one. example: > input: what is most common letter in abcccde? > possible answers: a, b, c > output: (c, 0.9), (b, 0.0.5), (a, 0.05) getting the above behavior instead of streamed output is as simple as using sglang's /v1/score endpoint instead of /generate. there's just one other trick required. for deepseek, you have to add a closing think tag before the response. this forces a direct answer instead of a reasoning trace. if you want reasoning, you can do that too, but imo that makes things too slow to be worth it. dsv4.1 flash is not as good as jev, but if we had enough spare compute to experiment with this same approach for a larger model then i think the decision quality would be at least as good, if not better. also, somewhat unrelated, i think decision-making models kill all prospecting & sourcing work. i would have absolutely killed to have jev or similar when i was recruiting Mintlify. absolutely incredible.

you can make any open source model behave like jev with just a bit of inference engineering. it's shockingly easy. to prove it, we built a new endpoint we're calling deepseek-v4.1-flash-jev. see the demo below. here's how it's done: sglang (an inference engine) offers a scoring endpoint in addition to the normal generation one. in scoring mode, given an input & set of possible answers, it forces the model to produce probabilities for each one. example: > input: what is most common letter in abcccde? > possible answers: a, b, c > output: (c, 0.9), (b, 0.0.5), (a, 0.05) getting the above behavior instead of streamed output is as simple as using sglang's /v1/score endpoint instead of /generate. there's just one other trick required. for deepseek, you have to add a closing think tag before the response. this forces a direct answer instead of a reasoning trace. if you want reasoning, you can do that too, but imo that makes things too slow to be worth it. dsv4.1 flash is not as good as jev, but if we had enough spare compute to experiment with this same approach for a larger model then i think the decision quality would be at least as good, if not better. also, somewhat unrelated, i think decision-making models kill all prospecting & sourcing work. i would have absolutely killed to have jev or similar when i was recruiting Mintlify. absolutely incredible.

Nick Khami

86,621 次观看 • 1 天前

Jev is WILD for SEO audit 🤯 in 45.1 seconds it read all 586 pages on my site and rebuilt the internal link map. 584 links placed, 139 pages it refused to link because nothing honestly fit. total cost $0.21. Claude Opus 5, same 586 pages, same clock, got through 21 of them and spent $1.43. per page that is ~190x cheaper. the full Opus pass would have run $43. internal linking is the perfect Jev job. it is not writing, it is 8,790 yes/no calls: does this page have a real reason to link to that one, and is there anchor text already sitting in the copy. that is a classification problem, and we have been paying frontier prices to do it one page at a time. what you are watching: left column is Jev, right is Opus, same queue, same rubric. the run stops the moment Jev finishes so Opus stops burning tokens. coming soon to Distribb + connector + gpt plugin

Jev is WILD for SEO audit 🤯 in 45.1 seconds it read all 586 pages on my site and rebuilt the internal link map. 584 links placed, 139 pages it refused to link because nothing honestly fit. total cost $0.21. Claude Opus 5, same 586 pages, same clock, got through 21 of them and spent $1.43. per page that is ~190x cheaper. the full Opus pass would have run $43. internal linking is the perfect Jev job. it is not writing, it is 8,790 yes/no calls: does this page have a real reason to link to that one, and is there anchor text already sitting in the copy. that is a classification problem, and we have been paying frontier prices to do it one page at a time. what you are watching: left column is Jev, right is Opus, same queue, same rubric. the run stops the moment Jev finishes so Opus stops burning tokens. coming soon to Distribb + connector + gpt plugin

borja

651,796 次观看 • 1 天前

Jev is HERE and this is the CLEAREST explanation of what it is and what NEW businesses it unlocks. (and at the end I'll tell you how to get Jev even if you're on the waitlist) WHAT IT IS You know how you open your inbox and have to decide what's junk, what needs a reply, and what can wait? Jev does that part. It looks at each thing and says "this is junk, I'm 94% sure." It doesn't write anything back to you. It just sorts. 1,700 emails for 18 cents, instantly. That sounds kinda trivial but the important part WHAT IT UNLOCKS My explanation of Jev sounds small until you realize HOW MANY jobs are exactly this. Someone reading a stack of applications. Someone deciding which support ticket goes to which team. Someone looking at inbound and deciding who's worth calling back. A few ideas on what it unlocks: 1/ Instant quotes that are actually instant. Every quote form on the internet says "we'll email you by end of day." Build the version that answers in under a second, for roofers, movers, insurance, legal intake. 2/ Lead scoring as a product. Every agency and service business has a contact form full of junk. Score every submission and send the real ones straight to the owner's phone. 3/ Support triage for companies with no support team. The ticket gets classified and routed before anyone opens it. 4/ Clipping tools. Pass in a transcript, get the best moments scored in three seconds. Every clipping product just got a cheaper engine. 5/ Application piles. Grants, permits, insurance claims, job apps, loan docs. Someone reads that stack one item at a time today. 6/ Marketplace matching. Someone types what they need and gets matched to the right local business instantly instead of waiting for callbacks. 7/ Browser agents that actually move FAST. That makes bulk browser work practical: pulling quotes from five carriers, filing the same form for 200 clients, checking supplier inventory in real time etc. TLDR; find an expensive queue and put Jev at the front of it. HOW TO GET IT I didn't realize you can skip the waitlist because Jev is live on the Vercel AI Gateway right now, so you can start calling it today. In this episode, we share how. Episode now live on The Startup Ideas Podcast (SIP) 🧃 (thanks to vogel for coming on and spilling the sauce today) Watch: Jev is a big deal because this is a whole new way to do AI Really cool Happy Jev day.

Jev is HERE and this is the CLEAREST explanation of what it is and what NEW businesses it unlocks. (and at the end I'll tell you how to get Jev even if you're on the waitlist) WHAT IT IS You know how you open your inbox and have to decide what's junk, what needs a reply, and what can wait? Jev does that part. It looks at each thing and says "this is junk, I'm 94% sure." It doesn't write anything back to you. It just sorts. 1,700 emails for 18 cents, instantly. That sounds kinda trivial but the important part WHAT IT UNLOCKS My explanation of Jev sounds small until you realize HOW MANY jobs are exactly this. Someone reading a stack of applications. Someone deciding which support ticket goes to which team. Someone looking at inbound and deciding who's worth calling back. A few ideas on what it unlocks: 1/ Instant quotes that are actually instant. Every quote form on the internet says "we'll email you by end of day." Build the version that answers in under a second, for roofers, movers, insurance, legal intake. 2/ Lead scoring as a product. Every agency and service business has a contact form full of junk. Score every submission and send the real ones straight to the owner's phone. 3/ Support triage for companies with no support team. The ticket gets classified and routed before anyone opens it. 4/ Clipping tools. Pass in a transcript, get the best moments scored in three seconds. Every clipping product just got a cheaper engine. 5/ Application piles. Grants, permits, insurance claims, job apps, loan docs. Someone reads that stack one item at a time today. 6/ Marketplace matching. Someone types what they need and gets matched to the right local business instantly instead of waiting for callbacks. 7/ Browser agents that actually move FAST. That makes bulk browser work practical: pulling quotes from five carriers, filing the same form for 200 clients, checking supplier inventory in real time etc. TLDR; find an expensive queue and put Jev at the front of it. HOW TO GET IT I didn't realize you can skip the waitlist because Jev is live on the Vercel AI Gateway right now, so you can start calling it today. In this episode, we share how. Episode now live on The Startup Ideas Podcast (SIP) 🧃 (thanks to vogel for coming on and spilling the sauce today) Watch: Jev is a big deal because this is a whole new way to do AI Really cool Happy Jev day.

GREG ISENBERG

146,853 次观看 • 1 天前

Jev is INSANE. 🤯 in 24.9 seconds it read 384 news from this morning and told 15 brands which stories to hop onto today, for $0.19. Claude Opus 5, running on the same feed at the same time, got through 4/384 and cost $0.77. per headline that is ~390x cheaper, and the answer comes back before you finish reading the headline yourself. there's going to be so many ways for JEV to help you find trending stories, find journalists covering it, and get press coverage. everything in the video open-sourced here: (in the demo/ folder, together with 30+ skills to turn your agent into a PR team)

Jev is INSANE. 🤯 in 24.9 seconds it read 384 news from this morning and told 15 brands which stories to hop onto today, for $0.19. Claude Opus 5, running on the same feed at the same time, got through 4/384 and cost $0.77. per headline that is ~390x cheaper, and the answer comes back before you finish reading the headline yourself. there's going to be so many ways for JEV to help you find trending stories, find journalists covering it, and get press coverage. everything in the video open-sourced here: (in the demo/ folder, together with 30+ skills to turn your agent into a PR team)

Elvis

557,837 次观看 • 2 天前

TurboEdit Instant text-based image editing discuss: We address the challenges of precise image inversion and disentangled image editing in the context of few-step diffusion models. We introduce an encoder based iterative inversion technique. The inversion network is conditioned on the input image and the reconstructed image from the previous step, allowing for correction of the next reconstruction towards the input image. We demonstrate that disentangled controls can be easily achieved in the few-step diffusion model by conditioning on an (automatically generated) detailed text prompt. To manipulate the inverted image, we freeze the noise maps and modify one attribute in the text prompt (either manually or via instruction based editing driven by an LLM), resulting in the generation of a new image similar to the input image with only one attribute changed. It can further control the editing strength and accept instructive text prompt. Our approach facilitates realistic text-guided image edits in real-time, requiring only 8 number of functional evaluations (NFEs) in inversion (one-time cost) and 4 NFEs per edit. Our method is not only fast, but also significantly outperforms state-of-the-art multi-step diffusion editing techniques.

TurboEdit Instant text-based image editing discuss: We address the challenges of precise image inversion and disentangled image editing in the context of few-step diffusion models. We introduce an encoder based iterative inversion technique. The inversion network is conditioned on the input image and the reconstructed image from the previous step, allowing for correction of the next reconstruction towards the input image. We demonstrate that disentangled controls can be easily achieved in the few-step diffusion model by conditioning on an (automatically generated) detailed text prompt. To manipulate the inverted image, we freeze the noise maps and modify one attribute in the text prompt (either manually or via instruction based editing driven by an LLM), resulting in the generation of a new image similar to the input image with only one attribute changed. It can further control the editing strength and accept instructive text prompt. Our approach facilitates realistic text-guided image edits in real-time, requiring only 8 number of functional evaluations (NFEs) in inversion (one-time cost) and 4 NFEs per edit. Our method is not only fast, but also significantly outperforms state-of-the-art multi-step diffusion editing techniques.

AK

16,062 次观看 • 2 年前

For a couple years, I have been using "otter using wifi on an airplane" as my test for the improving ability of AI image and video systems. But this is now too easy a challenge for current AI, so what if we kicked it up a notch (these are all veo 2 clips).

For a couple years, I have been using "otter using wifi on an airplane" as my test for the improving ability of AI image and video systems. But this is now too easy a challenge for current AI, so what if we kicked it up a notch (these are all veo 2 clips).

Ethan Mollick

116,417 次观看 • 1 年前

GPT Image 2 + Codex: or how to make Codex not suck at UI. Step 1: Generate a UI image (native in Codex) Step 2: Get Codex to implement the UI based on it Step 3: Get Codex to iterate until it aligns with the image as much as possible Codex is bad at initial UI, but very good at implementing a reference design, so this is your way out - iterate with the image model first and then Codex will do a good job.

GPT Image 2 + Codex: or how to make Codex not suck at UI. Step 1: Generate a UI image (native in Codex) Step 2: Get Codex to implement the UI based on it Step 3: Get Codex to iterate until it aligns with the image as much as possible Codex is bad at initial UI, but very good at implementing a reference design, so this is your way out - iterate with the image model first and then Codex will do a good job.

Peter Gostev

67,245 次观看 • 5 个月前

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

Lovis Odin

58,841 次观看 • 1 年前

REPORTER: Have you spoken to Macron about re-opening the Strait of Hormuz? TRUMP: "Yeah, I have. On a scale of 0-10, he's been an 8. Not perfect. But it's FRANCE! We don't expect perfect" 😂🤣 What is Troubling is Starmer and the UK. We have done so much for them but they are about as worthless as can be

REPORTER: Have you spoken to Macron about re-opening the Strait of Hormuz? TRUMP: "Yeah, I have. On a scale of 0-10, he's been an 8. Not perfect. But it's FRANCE! We don't expect perfect" 😂🤣 What is Troubling is Starmer and the UK. We have done so much for them but they are about as worthless as can be

Johnny Midnight ⚡️

23,927 次观看 • 6 个月前

The battle of the Text/Image to 3D Model 👍- To make this fair, I tested the same image across all platforms and used the first generated result. No cherry picking allowed! I also tested what I think are the 4 top services to use right now. Let me know what you think is the best 3D mesh generation from these platforms however I think it's quite obvious that Tencent Hy 2.5 is currently the king 👑 It wins in overall detail and sharpness, accuracy in relation to the input image, as well as texture generation and projection. Tripo 2.5 comes in second and Rodin 3rd with Trellis in 4th place. However, Rodin has the best topology since it can generate a mesh in quads BUT I left all of them in triangles to keep it fair. Hi3DGen can't be counted as it can't even generate textures and it has no multi-image support. Meshy was another option but the results were not great. Is there any other AI 3D Gen services that you know of that you want me to test as well?

The battle of the Text/Image to 3D Model 👍- To make this fair, I tested the same image across all platforms and used the first generated result. No cherry picking allowed! I also tested what I think are the 4 top services to use right now. Let me know what you think is the best 3D mesh generation from these platforms however I think it's quite obvious that Tencent Hy 2.5 is currently the king 👑 It wins in overall detail and sharpness, accuracy in relation to the input image, as well as texture generation and projection. Tripo 2.5 comes in second and Rodin 3rd with Trellis in 4th place. However, Rodin has the best topology since it can generate a mesh in quads BUT I left all of them in triangles to keep it fair. Hi3DGen can't be counted as it can't even generate textures and it has no multi-image support. Meshy was another option but the results were not great. Is there any other AI 3D Gen services that you know of that you want me to test as well?

Travis Davids

61,232 次观看 • 1 年前

HERE WE GO THEN! Let’s try #revforxmasno1 for #sweetapplecrumble Given how streaming favours the pop Goliaths , indie Davids like us have next to no chance. But we are the 6th favourite across all bookmakers So let’s have a bash. You can buy a cd using this link as well as donate to Samaritans You can also download it on all sites too. Please add it to your streaming playlists and worse can scenario we get no 140 in the charts but raise a load of money to help those in need Thanks for everyone who’s shared the song and video and remember no one has to be alone this Xmas. Jon

HERE WE GO THEN! Let’s try #revforxmasno1 for #sweetapplecrumble Given how streaming favours the pop Goliaths , indie Davids like us have next to no chance. But we are the 6th favourite across all bookmakers So let’s have a bash. You can buy a cd using this link as well as donate to Samaritans You can also download it on all sites too. Please add it to your streaming playlists and worse can scenario we get no 140 in the charts but raise a load of money to help those in need Thanks for everyone who’s shared the song and video and remember no one has to be alone this Xmas. Jon

Reverend&TheMakers

364,000 次观看 • 1 年前

As I promised yesterday, I'll briefly explain LoRA training and share a workflow I made so you can do it quickly. First, let me answer a very common question: 'Why train LoRAs when we have such advanced models?' Even though we have incredibly advanced models now (like NBP), we still can't always get them to do specific things we want. Simplest example: the spritesheet LoRA I made the other day. I generated 1000 images with Nano Banana and only 100 were what I wanted. The LoRA I trained using those 100 images gives me nearly 100% consistent results. Second point is cost and speed. With LoRA, we can cut costs by 4-5x. And while doing that, we're generating 4-5x faster. How many images do you need for a good LoRA? This depends on your LoRA's complexity. For example, when I training the spritesheet LoRA, even though I used 100 images, I didn't include buildings in the training data, so this LoRA doesn't work for buildings. So think about your LoRA's use cases and add examples for as many use cases as possible to improve quality. What are paired images and how to train LoRAs for image-editing? When training LoRAs for image editing on fal, we call each edit example paired images - one with _start suffix, one with _end suffix. For example, if you're training a background remove LoRA, the unedited original photo will be your '_start' image. The image with background removed will be the '_end' image. Simply put: images we want to edit or use as reference get _start, target images we want to achieve get '_end'. Important: save both images with the same name. Like image332_start.jpg and image332_end.jpg. This way the system knows which images pair together. What about training LoRAs for models with multiple image inputs? Same logic. We still use _start and _end suffixes, but with one difference. Since there are multiple input images, we can number them: _start, _start1, _start2. Example: start images, 1st image = Woman portrait (image35_start.jpg) 2nd image = Glasses photo (image35_start1.jpg) 3rd image = Hat photo (image35_start2.jpg) Output image = portrait of woman wearing glasses and hat (image35_end.jpg) Can we do more detailed captioning? Yes. Similarly, you can improve training quality by creating a txt file for each set with the caption inside. Example: create image35.txt and write: 'Recreate the image by putting the glasses from the second image and the hat from the third image on the woman in the first image.' What are Steps? How many should I use? What's Learning Rate? Steps determines how many times the model sees and processes your training data (your images). Each step, the model learns a bit more. But as steps increase, so does the risk of overfitting. So there's no real default. But for a simpler LoRA with 20 paired images, 1000 steps is ideal. Here's a metaphor for the Steps and Learning Rate relationship: Imagine you have a balloon. Our goal is to inflate it to the optimal size. Steps = How many times we blow into the balloon Learning rate = How hard we blow each time If we blow too softly, we need to blow many more times. If we blow too hard, we risk popping it quickly and can't reach optimal size. Of course training won't explode, but it won't work as intended because it wasn't trained optimally. Training's done, now what? Once training's complete, you'll have a safetensors file. Every model you train on fal has a LoRA inference endpoint. In that inference, add your safetensors file link to the LoRA url input, and you can use your LoRA. Thanks for the read! The workflow in the video: If I forgot anything, let me know in the replies.

As I promised yesterday, I'll briefly explain LoRA training and share a workflow I made so you can do it quickly. First, let me answer a very common question: 'Why train LoRAs when we have such advanced models?' Even though we have incredibly advanced models now (like NBP), we still can't always get them to do specific things we want. Simplest example: the spritesheet LoRA I made the other day. I generated 1000 images with Nano Banana and only 100 were what I wanted. The LoRA I trained using those 100 images gives me nearly 100% consistent results. Second point is cost and speed. With LoRA, we can cut costs by 4-5x. And while doing that, we're generating 4-5x faster. How many images do you need for a good LoRA? This depends on your LoRA's complexity. For example, when I training the spritesheet LoRA, even though I used 100 images, I didn't include buildings in the training data, so this LoRA doesn't work for buildings. So think about your LoRA's use cases and add examples for as many use cases as possible to improve quality. What are paired images and how to train LoRAs for image-editing? When training LoRAs for image editing on fal, we call each edit example paired images - one with _start suffix, one with _end suffix. For example, if you're training a background remove LoRA, the unedited original photo will be your '_start' image. The image with background removed will be the '_end' image. Simply put: images we want to edit or use as reference get _start, target images we want to achieve get '_end'. Important: save both images with the same name. Like image332_start.jpg and image332_end.jpg. This way the system knows which images pair together. What about training LoRAs for models with multiple image inputs? Same logic. We still use _start and _end suffixes, but with one difference. Since there are multiple input images, we can number them: _start, _start1, _start2. Example: start images, 1st image = Woman portrait (image35_start.jpg) 2nd image = Glasses photo (image35_start1.jpg) 3rd image = Hat photo (image35_start2.jpg) Output image = portrait of woman wearing glasses and hat (image35_end.jpg) Can we do more detailed captioning? Yes. Similarly, you can improve training quality by creating a txt file for each set with the caption inside. Example: create image35.txt and write: 'Recreate the image by putting the glasses from the second image and the hat from the third image on the woman in the first image.' What are Steps? How many should I use? What's Learning Rate? Steps determines how many times the model sees and processes your training data (your images). Each step, the model learns a bit more. But as steps increase, so does the risk of overfitting. So there's no real default. But for a simpler LoRA with 20 paired images, 1000 steps is ideal. Here's a metaphor for the Steps and Learning Rate relationship: Imagine you have a balloon. Our goal is to inflate it to the optimal size. Steps = How many times we blow into the balloon Learning rate = How hard we blow each time If we blow too softly, we need to blow many more times. If we blow too hard, we risk popping it quickly and can't reach optimal size. Of course training won't explode, but it won't work as intended because it wasn't trained optimally. Training's done, now what? Once training's complete, you'll have a safetensors file. Every model you train on fal has a LoRA inference endpoint. In that inference, add your safetensors file link to the LoRA url input, and you can use your LoRA. Thanks for the read! The workflow in the video: If I forgot anything, let me know in the replies.

ilker

15,192 次观看 • 8 个月前

Flashback to the backlash to Fiducia supplicans by African bishops. Bishop Martin Mtumbuka of Karonga, Malawai: "We have no choice, we cannot allow such an offensive and apparently blasphemous declaration to be implemented in our diocese." "Our major concern with this declaration is that it looks to us like a heresy, it reads like a heresy and it affects heresy, because is asks us to bless two people of same sex, as individuals, but not as a couple, who the previous night sleep together like a couple, presented themselves to us as a couple, are blessed as individuals, but they leave our presence as a couple, they go to their home as a couple, they sleep in the same bed as a couple. But the document says they are not blessed as a couple, although they appear to have been blessed as couple. How could this not be changing the authentic teaching of the Church.” Watch the whole video. It is worth it.

Flashback to the backlash to Fiducia supplicans by African bishops. Bishop Martin Mtumbuka of Karonga, Malawai: "We have no choice, we cannot allow such an offensive and apparently blasphemous declaration to be implemented in our diocese." "Our major concern with this declaration is that it looks to us like a heresy, it reads like a heresy and it affects heresy, because is asks us to bless two people of same sex, as individuals, but not as a couple, who the previous night sleep together like a couple, presented themselves to us as a couple, are blessed as individuals, but they leave our presence as a couple, they go to their home as a couple, they sleep in the same bed as a couple. But the document says they are not blessed as a couple, although they appear to have been blessed as couple. How could this not be changing the authentic teaching of the Church.” Watch the whole video. It is worth it.

Catholic Sat

35,485 次观看 • 4 个月前

Borderlands 4 Performance Summary in 2 minutes: 0:00 - Native Res Badass Settings the game runs bad 0:30 - Upscaling Performance Mode Badass Settings, the game barely manages to hit 60FPS 1:20 - Frame Gen turned on and now you have 100 FPS Some people will turn all these settings on and be happy with it but in my honest opinion these are used as crutches to make the game look as if it performs good. Borderlands 4 has a certain art style that doesn't look like it needs these kind of settings to be turned on to play at a high framerate but sadly it does and you can defend this if you want but the future of gaming looks bleak, using this tech to not optimize the games further will just result in input latency increase as every game will be delivered with frame-gen turned on in the devs mind.

Borderlands 4 Performance Summary in 2 minutes: 0:00 - Native Res Badass Settings the game runs bad 0:30 - Upscaling Performance Mode Badass Settings, the game barely manages to hit 60FPS 1:20 - Frame Gen turned on and now you have 100 FPS Some people will turn all these settings on and be happy with it but in my honest opinion these are used as crutches to make the game look as if it performs good. Borderlands 4 has a certain art style that doesn't look like it needs these kind of settings to be turned on to play at a high framerate but sadly it does and you can defend this if you want but the future of gaming looks bleak, using this tech to not optimize the games further will just result in input latency increase as every game will be delivered with frame-gen turned on in the devs mind.

NikTek

359,686 次观看 • 1 年前

Here's a look behind the scenes of how I used a 3d model, generated from an image, as a "driver" for ai animation. In the near future we'll see more 3d tools that support these types of workflows, with much more emphasis on enhancing user input instead of just letting the ai do everything. I used the 3d model to generate keyframes that I then animated in Luma Dream. By using a 3d model with a diffusion "layer" done with Krea, I got quite high quality frames with a relatively high consistency as well since the AI didn't have to "hallucinate" that much. I upscaled and detailed the frames using Magnific. Using semi-detailed 3d models as drivers for gen ai is very powerful, and it allows us to use models that are sculpted in a more free-flowing type of workflow that doesn't rely so much on high surface detailing or very time-consuming finish, but instead relies more on gestural 3d sculpting. #art #ai

Here's a look behind the scenes of how I used a 3d model, generated from an image, as a "driver" for ai animation. In the near future we'll see more 3d tools that support these types of workflows, with much more emphasis on enhancing user input instead of just letting the ai do everything. I used the 3d model to generate keyframes that I then animated in Luma Dream. By using a 3d model with a diffusion "layer" done with Krea, I got quite high quality frames with a relatively high consistency as well since the AI didn't have to "hallucinate" that much. I upscaled and detailed the frames using Magnific. Using semi-detailed 3d models as drivers for gen ai is very powerful, and it allows us to use models that are sculpted in a more free-flowing type of workflow that doesn't rely so much on high surface detailing or very time-consuming finish, but instead relies more on gestural 3d sculpting. #art #ai

Martin Nebelong

48,280 次观看 • 2 年前

TypeSafe AI released Jev. I truly believe this is a game changing kind of AI model. No joke. It does NOT replace LLMs. But in tandem with LLMs it feels like a missing building block. It makes the kind of applications you can build now SO much better and more powerful. And cheaper! It's an AI model that's all about decision making. But not like old-school classifiers ("fraud or no fraud") but with flexibility you know from LLMs. Feed some input (eg customer email, any other text, structured data) into it, define your questions ("which team should answer this email?") and dimensions ("sales, technical, fraud") and get back a response within milliseconds. No long reasoning, no exploding token costs. Sure, that answer will still be generated by an LLM, complex questions may still require an (additional) LLM but this model is really mindblowing.

TypeSafe AI released Jev. I truly believe this is a game changing kind of AI model. No joke. It does NOT replace LLMs. But in tandem with LLMs it feels like a missing building block. It makes the kind of applications you can build now SO much better and more powerful. And cheaper! It's an AI model that's all about decision making. But not like old-school classifiers ("fraud or no fraud") but with flexibility you know from LLMs. Feed some input (eg customer email, any other text, structured data) into it, define your questions ("which team should answer this email?") and dimensions ("sales, technical, fraud") and get back a response within milliseconds. No long reasoning, no exploding token costs. Sure, that answer will still be generated by an LLM, complex questions may still require an (additional) LLM but this model is really mindblowing.

Maximilian

36,117 次观看 • 3 天前