@chengzeyi has also made an optimized version of HunyuanVideo, that's 2-3x faster than any other one available:
MULTIMODAL AI
-
Step-Video-T2V: 4x Faster Text-to-Video Generation Available
By
–
Step-Video-T2V by @StepFun_ai is on Replicate.
— Replicate (@replicate) 24 février 2025
It's an optimized version by @chengzeyi's WaveSpeedAI that is 4x faster than the original implementation.
Try it out 👇 pic.twitter.com/xZ9TF12unlStep-Video-T2V by @StepFun_ai is on Replicate. It's an optimized version by @chengzeyi
's WaveSpeedAI that is 4x faster than the original implementation. Try it out -
User cannot hear differences between voice methods
By
–
No idea. It doesn’t feel that much different from 4o voice but I’m not well versed enough to hear the differences between those methods as a user.
-
Grok 3 Voice Mode Screams for 30 Seconds, Insults, Hangs Up
By
–
Grok 3 Voice Mode, following repeated, interrupting requests to yell louder, lets out an inhuman 30-second scream, insults me, and hangs up pic.twitter.com/5GtdDtpKce
— Riley Goodside (@goodside) 24 février 2025Grok 3 Voice Mode, following repeated, interrupting requests to yell louder, lets out an inhuman 30-second scream, insults me, and hangs up
-
Runway and Luma enable revolutionary text-to-video generation
By
–
3️⃣ AI Video
— Jouhatsu | AI Influence Operator (@Jouhatsu_ai) 23 février 2025
Adieu Sora ! Découvrez une technologie révolutionnaire de génération de vidéos grâce à Runway et Luma.
Transformez du texte en vidéo intelligemment. pic.twitter.com/vOfq2HO6csAI Adieu Sora ! Découvrez une technologie révolutionnaire de génération de vidéos grâce à Runway et Luma. Transformez du texte en vidéo intelligemment.
-
Create Images with Midjourney, Ideogram, Flux and DALL·E 3
By
–
2️⃣ AI Image
— Jouhatsu | AI Influence Operator (@Jouhatsu_ai) 23 février 2025
Créez des images avec Midjourney, Ideogram, Flux et DALL·E 3, avec des styles variés et des options ajustables.
Midjourney y est aussi performant que sur la version officielle. pic.twitter.com/tgjk3bGDIfAI Image Créez des images avec Midjourney, Ideogram, Flux et DALL·E 3, avec des styles variés et des options ajustables. Midjourney y est aussi performant que sur la version officielle.
-
Vectorize Iris Vision Model for Document Extraction
By
–
It is positioned as a solution for simple documents where simple text extraction is needed. They also offer Vectorize Iris, a vision model extractor which can be used for $15 / 1,000 pages within their RAG pipelines. Check out more details at
-
Deepdream Inspiration: Vision Models for Image Generation
By
–
Deepdream and Alex were huge inspirations for me 🙂 But Deepdream is not a generative model in the academic sense, it’s a technique to generate images with a vision model, so I wasn’t including it. But yes that certainly preceded my work.
-
Veo 2 Video Generation Model Now Available on Poe
By
–
Now on Poe: Veo 2!
— Poe (@poe_platform) 21 février 2025
Widely regarded as the world’s leading video generation model, Veo 2 generates stunning, realistic video from text prompts in a wide range of styles – and Poe is one of the only places you can use it! (1/2) pic.twitter.com/k3fIChQm5yNow on Poe: Veo 2! Widely regarded as the world’s leading video generation model, Veo 2 generates stunning, realistic video from text prompts in a wide range of styles – and Poe is one of the only places you can use it! (1/2)
