Hi Andy, I wanted to share something I ran into while testing the new Muse MiniMax Director with a longer-form generation. I ran a 300-second / 5-minute video with 20 chunks at 0.4 MP, single-stage, on my RTX 5080 system with 32 GB RAM. The really good news is that the workflow successfully generated all 20 chunks. It ran for about 10 hours and 47 minutes, and the generation itself appeared to work all the way through. The problem happened only at the very end when the Director tried to combine everything into the final output. I received this error: DefaultCPUAllocator: not enough memory — tried to allocate 37,165,547,520 bytes The stack trace points to: torch.cat(all_images, dim=0) So, it looks like all of the completed chunks were being combined into one large tensor in system RAM at the end, and my 32 GB machine couldn't handle that final allocation. Is there a way the Director could eventually handle longer videos by saving/assembling completed chunks progressively to disk, rather than holding all of the decoded frames in RAM until the end? Or is there already a setting/workflow method you recommend for longer videos, so this final RAM issue doesn't happen? I'm asking because the actual 20-chunk generation worked, which was really encouraging. It was only the final assembly that failed. Thanks — I'm really enjoying testing what this workflow can do.

Ok so here is an example i just created with the Director node. I only did it at 0.2 so I would have a fast render. Duration 30 seconds, chunk set to 15, so it does 2 runs. Took 382.33 secs on my 5090, so expect longer with lower vram cards. Or if you have quantized models swap them out and use them. Prompt: --- Chunk 1/2 (~15.0s) --- subject_definitions: <Subject 1> is a young woman with bright blue hair and fair skin tone, wearing a black baseball cap, light grey bomber jacket, white ribbed tank top, dark grey cargo pants, and black and white sneakers (from `<Picture 1>`). <Subject 2> is a futuristic corridor, marked by white walls, glowing blue LED strips, and a large screen displaying a vessel schematic. (from `<Picture 2>`). summary: [reference generation] The target video shows <Subject 1> and <Subject 2>. retention_analysis: <Subject 1> (referenced in this shot): fully_preserved - matches `<Picture 1>`. <Subject 2> (referenced in this shot): fully_preserved - matches `<Picture 2>`. detailed_description: Clean, bright sci-fi photography, even white and cool-blue lighting, sharp reflective metal surfaces, smooth steady camera movement, crisp modern digital look. [Shot 1] <Subject 1> (S1) continues: She steps into the corridor through a side door, glancing up at the glowing wall screens as she walks, camera tracking alongside her at a relaxed medium shot. [Shot 2] At 00:07.500, She slows and stops in front of one of the status screens, tilting her head as she reads it, one of the panels flickers briefly. Camera pushes in slightly to a close shot on her curious expression. overall_soundscape: Soft hum of ship systems, a faint air-circulation hiss, quiet footsteps on the metal floor, an occasional low electronic beep from a wall panel. non_diegetic_music: Light, minimal electronic pulse underneath, subtle and unobtrusive, never overpowering the dialogue. --- Chunk 2/2 (~15.0s) --- subject_definitions: <Subject 1> is a young woman with bright blue hair and fair skin tone, wearing a black baseball cap, light grey bomber jacket, white ribbed tank top, dark grey cargo pants, and black and white sneakers (from `<Picture 1>`).

🎬 Start Here — Your Roadmap Through Muse Collective Welcome! 👋 Muse Collective is designed to take you from complete beginner to building your own AI creator using powerful local, open-source tools. No monthly subscriptions. No credit systems. No usage limits. Just work through the modules in order. Each one builds on the last. 🚀 Your Journey 🟢 Stage 1 — Learn the Foundations - Welcome & Roadmap - AI Creator Foundations - Setting Up ComfyUI - AI Image Generation You'll learn to: ✅ Install and configure ComfyUI ✅ Generate high-quality AI images locally ✅ Understand the core tools and workflows 🟣 Stage 2 — Build Your AI Creator - Creating Your AI Character - Voice Cloning & TTS - Video Generation & Lip Sync You'll learn to: ✅ Create a consistent AI character ✅ Clone or create an AI voice ✅ Generate realistic talking AI videos 🟠 Stage 3 — Build Your Content System - Automation & Content Pipeline - Getting Your First Followers - Advanced Systems - Recorded Calls & Future Updates You'll learn to: ✅ Build repeatable AI content workflows ✅ Automate repetitive tasks ✅ Create more advanced AI productions ✅ Stay up to date as local AI evolves ⭐ How Progress Works You'll earn XP automatically as you participate in the community. As you level up, new modules unlock naturally. The progression is intentional—everything builds on the foundations before it. 🤝 Community Guidelines ✅ Work through the modules in order. ✅ Share your progress—even if it isn't perfect. ✅ Ask questions publicly so everyone can learn from the answers. ❌ No self-promotion or spam. 🚀 Most importantly... Don't worry about making everything perfect. Build. Experiment. Learn. Every creator in this community started with their first image. Let's build something incredible. Andy

Building the roadmap around what's actually useful to you. Pick the one you'd get the most value from next, and I'll try to turn it into tutorials/workflows.