🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

AI video is real

I wrote a blog post. AI video is having its Stable Diffusion moment. replicate.com/blog/ai-video-…

🔉 Never gonna get over this Julian Bilcke used Replicate to fine-tune Hunyuan on Never gonna give you up, and we, uh, made some animals dance to it

Just finished reading A Canticle for Leibowitz. Minimax + replicate.com/fofr/flux-dark… captures the mood perfectly.

Everyone are now able to generate this high-quality video by HunyuanVideo model in 1 minute! wavespeed.ai/dashboard/mode… replicate.com/wavespeedai/hu…

Brilliant work from Cheng Zeyi

This kind of works replicate.com/andreasjansson… It's based on diffusers' pipeline_stable_diffusion_inpaint_legacy, so video2video comes for free

Interpany-cleaner by Zhihang Zhong is really great for smoothing out the 16 fps on Wan. replicate.com/lucataco/inter…

Emacs is the perfect development environment for working with LLMs. gpt.el is great but I think there's something much more profound around the corner. An agent with full access to all of Emacs: It has shell, can browse the web with eww (fast) or emacs-webkit (full), fantastic

Tiny library for calling Python functions as agent tools: github.com/andreasjansson…

What other data sources should we add to hype.replicate.dev ? Any new subreddits I'm missing? High-quality blogs with daily updates?

We've made Flux LoRA training less than $2. It's blazingly fast. replicate.com/replicate/fast…

this site so good: replicate.com/flux-kontext-a…

Getting the same feeling feeding >100k context windows to Claude that I used to get when my training GPUs were running at >90%