POM's banner
POM's profile picture

POM

@peterom • 15,737 subscribers

Open Source AI Art hypeman w/ @banodoco | Tools for artistic expression w/ @Reigh_Art | 👫🐕 w/ @hannahsubmarine. Aspiring 2nd Renaissance Man.

Shorts

The progress of the Animatediff community over the past 10 months has been miraculous - see attached! Now, closed startups like Krea are taking the fruits of all this effort - so I'd like to tell the story of how we got here & what people who believe in open source can do.

The progress of the Animatediff community over the past 10 months has been miraculous - see attached! Now, closed startups like Krea are taking the fruits of all this effort - so I'd like to tell the story of how we got here & what people who believe in open source can do.

166,604 Aufrufe

The meme-power of video-editing with Wan will be a force to be reckoned with Excellent examples by Illuminati Reptilien:

The meme-power of video-editing with Wan will be a force to be reckoned with Excellent examples by Illuminati Reptilien:

52,973 Aufrufe

Woke up today to see spacepxl trained a model on top of Wan to achieve SOTA performance in deblurring w/ an approach that will generalise to all kinds of video control tasks - upscaling, canny, video inpainting, etc. I love the open source AI art/banodoco community so much

Woke up today to see spacepxl trained a model on top of Wan to achieve SOTA performance in deblurring w/ an approach that will generalise to all kinds of video control tasks - upscaling, canny, video inpainting, etc. I love the open source AI art/banodoco community so much

45,609 Aufrufe

I believe that StoryDiffusion has the potential to be Animatediff's complex motion sister-model! While AD is amazing for granular control, micro-motion and all kinds of abstract motion, it fails at complex realistic motion - walking, human movements, cars, etc. StoryDiffusion seems very promising for this + also has characteristics that will likely make the community very receptive to it and likely to extend its capabilities: 2) Appealing base-model results - likely to get the community excited - feels like significantly better realistic motion than AD 2) Modular - their approach is built with a number of components that can be combined and taken apart - it works by generating consistent images, then animating them together - each of these stages can likely be upgraded, used and influenced in different ways. 3) Flexible - they demonstrate a bunch of different conditioning options 4) Likely easy on RAM - it's based on SD 1.5 + authors mention precautions to reduce RAM consumption 5) Built to plug into the existing ecosystem - e.g. the fact that it works with the SD1.5 ecosystem will give it a huge advantage! While it's very early to say - e.g. the video model hasn't even been released yet! - it does seem very promising. With 9 months of SD1.5/Animatediff-esque progress improving every element of it, I can see an an extremely extended version of this beating Sora + running for a fraction of the compute resources on a consumer GPU. Together with Animatediff to drive the micro-motions and abstract stuff, it could produce be extraordinary/otherworldly/insane/beautiful stuff. This is the first open video model I've been excited about since Animatediff - though cautiously optimistic! Link here:

I believe that StoryDiffusion has the potential to be Animatediff's complex motion sister-model! While AD is amazing for granular control, micro-motion and all kinds of abstract motion, it fails at complex realistic motion - walking, human movements, cars, etc. StoryDiffusion seems very promising for this + also has characteristics that will likely make the community very receptive to it and likely to extend its capabilities: 2) Appealing base-model results - likely to get the community excited - feels like significantly better realistic motion than AD 2) Modular - their approach is built with a number of components that can be combined and taken apart - it works by generating consistent images, then animating them together - each of these stages can likely be upgraded, used and influenced in different ways. 3) Flexible - they demonstrate a bunch of different conditioning options 4) Likely easy on RAM - it's based on SD 1.5 + authors mention precautions to reduce RAM consumption 5) Built to plug into the existing ecosystem - e.g. the fact that it works with the SD1.5 ecosystem will give it a huge advantage! While it's very early to say - e.g. the video model hasn't even been released yet! - it does seem very promising. With 9 months of SD1.5/Animatediff-esque progress improving every element of it, I can see an an extremely extended version of this beating Sora + running for a fraction of the compute resources on a consumer GPU. Together with Animatediff to drive the micro-motions and abstract stuff, it could produce be extraordinary/otherworldly/insane/beautiful stuff. This is the first open video model I've been excited about since Animatediff - though cautiously optimistic! Link here:

22,141 Aufrufe

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

The 2nd edition of The Arca Gidan Prize will begin this Sunday, now w/ a $50k prize fund! Artists will be challenged to push open models to their limits from March 1, w/ winners receiving substantial prizes, travel to show their work at ADOS Paris + a 4.5kg Toblerone chocolate bar as their trophy. This was funded using fees from the $dataclaw memecoin that the community created. While I don't own any of the coin, I received a percentage of the transaction fee on every trade, which I'm in turn giving to the prize fund for this competition! - Theme + details announced: March 1 - Competition deadline: March 31 - Winners announced: April 2 - ADOS Paris: April 17-19 ADOS Paris is also supported by Lightricks, and more competition supporters are very welcome! Any additional fees will also go to the prize fund! Details on the coin below if you feel like gambling. Links: - Arca Gidan Website: - Arca Gidan Discord: - ADOS Paris Website: - My wallet - to validate all money goes to prizes:
0:20
peterom's profile picture

The 2nd edition of The Arca Gidan Prize will begin this Sunday, now w/ a $50k prize fund! Artists will be challenged to push open models to their limits from March 1, w/ winners receiving substantial prizes, travel to show their work at ADOS Paris + a 4.5kg Toblerone chocolate bar as their trophy. This was funded using fees from the $dataclaw memecoin that the community created. While I don't own any of the coin, I received a percentage of the transaction fee on every trade, which I'm in turn giving to the prize fund for this competition! - Theme + details announced: March 1 - Competition deadline: March 31 - Winners announced: April 2 - ADOS Paris: April 17-19 ADOS Paris is also supported by Lightricks, and more competition supporters are very welcome! Any additional fees will also go to the prize fund! Details on the coin below if you feel like gambling. Links: - Arca Gidan Website: - Arca Gidan Discord: - ADOS Paris Website: - My wallet - to validate all money goes to prizes:

POM

63,740 Aufrufe • vor 3 Monaten

Shipping an experiment on top of Hermes Agent that allows an agent to steer itself With it, a harness like desloppify can clear its own context, switch its models, prompt itself when it stops, etc. Video shows switching between Grok 4.20 ($6/m) for execution + Gemini 3.1 ($12/m) for planning + Claude ($25/m) for sense-checking - 6+ hours of refactoring w/o errors/stoppage - this can endlessly and safeguard itself w/ various triggers as sense-checks! Desloppify v0.9.10 release notes w/ instructions to test + many more contributions by the community: Video by the wonderful Hannah Submarine - best watched with audio:
0:54
peterom's profile picture

Shipping an experiment on top of Hermes Agent that allows an agent to steer itself With it, a harness like desloppify can clear its own context, switch its models, prompt itself when it stops, etc. Video shows switching between Grok 4.20 ($6/m) for execution + Gemini 3.1 ($12/m) for planning + Claude ($25/m) for sense-checking - 6+ hours of refactoring w/o errors/stoppage - this can endlessly and safeguard itself w/ various triggers as sense-checks! Desloppify v0.9.10 release notes w/ instructions to test + many more contributions by the community: Video by the wonderful Hannah Submarine - best watched with audio:

POM

46,426 Aufrufe • vor 2 Monaten

Open sourcing InScene + InScene Annotate - a pair of LoRAs for steering in-scene image generation w/ QwenEdit. You can use a combination of prompts + annotations to generate consistent images inside a scene. In beta but v. powerful. Model + workflow below, training data soon.
0:58
peterom's profile picture

Open sourcing InScene + InScene Annotate - a pair of LoRAs for steering in-scene image generation w/ QwenEdit. You can use a combination of prompts + annotations to generate consistent images inside a scene. In beta but v. powerful. Model + workflow below, training data soon.

POM

71,398 Aufrufe • vor 7 Monaten

Today I'm announcing the themes for our upcoming open source AI art competition, The Arca Gidan Prize! The meta-theme for this edition is Time. Our goal is to push people toward the unconventional. We've all seen many AI movie trailers, commercials and music videos - but what can you do with AI that's genuinely new and different? What was impossible before? How can you push open models beyond what people expect? Without further ado, here are the sub-themes: 🔄 Déjà Vu "This has happened before — or has it? That uncanny shimmer when moments echo: the glitch, the loop. When time spirals back through existence and ripples with recognition." 🌸 The Briefness of Bloom "A moment when something is perfectly itself — just before it fades. The cherry blossom at peak. The golden hour before dusk. So luminous as it slips away, already a memory." ⏱️ Traveling Through Time "Traveling through time — backward, forward, sideways. The time traveler, the archaeologist, the prophet. Journeys to moments that never were or haven't happened yet." For full details on prizes and rules, check out: - The Arca Gidan Discord: - The website: As mentioned before, the current prize fund is $50k, paid for by the fees I received from the $dataclaw token. Any further fees I receive will be added to the prize fund - we'll keep adding $1k prizes for each full $1k we receive. Finally, please enjoy this theme trailer by Hannah Submarine:
1:18
peterom's profile picture

Today I'm announcing the themes for our upcoming open source AI art competition, The Arca Gidan Prize! The meta-theme for this edition is Time. Our goal is to push people toward the unconventional. We've all seen many AI movie trailers, commercials and music videos - but what can you do with AI that's genuinely new and different? What was impossible before? How can you push open models beyond what people expect? Without further ado, here are the sub-themes: 🔄 Déjà Vu "This has happened before — or has it? That uncanny shimmer when moments echo: the glitch, the loop. When time spirals back through existence and ripples with recognition." 🌸 The Briefness of Bloom "A moment when something is perfectly itself — just before it fades. The cherry blossom at peak. The golden hour before dusk. So luminous as it slips away, already a memory." ⏱️ Traveling Through Time "Traveling through time — backward, forward, sideways. The time traveler, the archaeologist, the prophet. Journeys to moments that never were or haven't happened yet." For full details on prizes and rules, check out: - The Arca Gidan Discord: - The website: As mentioned before, the current prize fund is $50k, paid for by the fees I received from the $dataclaw token. Any further fees I receive will be added to the prize fund - we'll keep adding $1k prizes for each full $1k we receive. Finally, please enjoy this theme trailer by Hannah Submarine:

POM

37,623 Aufrufe • vor 3 Monaten

Lightricks are the one to watch in open source creative AI They listen to the community, use its input to improve models/tools, & reduce friction to train control vectors via IC-LoRAs How crazy is it that this video gens in <3 mins on consumer hardware? (credit: Trap City)
0:20
peterom's profile picture

Lightricks are the one to watch in open source creative AI They listen to the community, use its input to improve models/tools, & reduce friction to train control vectors via IC-LoRAs How crazy is it that this video gens in <3 mins on consumer hardware? (credit: Trap City)

POM

16,566 Aufrufe • vor 3 Monaten

5 days of the community playing w/ this and the community is getting multiple characters to do complex motion w/ near perfect costume/scene adherence Open model from Alibaba, fine-tune by ByteDance, merged by Inner-Reflections = this by AbleJones - open innovation is magical:
0:28
peterom's profile picture

5 days of the community playing w/ this and the community is getting multiple characters to do complex motion w/ near perfect costume/scene adherence Open model from Alibaba, fine-tune by ByteDance, merged by Inner-Reflections = this by AbleJones - open innovation is magical:

POM

18,625 Aufrufe • vor 10 Monaten

Keine weiteren Inhalte verfügbar