nick/comfyui-serverless
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
51 Commits 1 Branch 0 Tags
52dd0d8766cb82bcc93d976bcc4e247c8e850456
Commit Graph

20 Commits

Author SHA1 Message Date
Debian
c04e4b6250 Update for RTX 5090 (Blackwell sm_120) support
Some checks failed
Build and Push Docker Image / build (push) Failing after 29m16s
Details 
- Switch to PyTorch nightly with CUDA 12.8 (required for sm_120)
- Target TORCH_CUDA_ARCH_LIST="12.0" for Blackwell
- Remove nunchaku (incompatible with PyTorch nightly)
- Use latest SageAttention (has sm_120 kernel support)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-03 22:01:58 +00:00
Debian
9f71e6db57 Limit SageAttention to A100/H100 due to cross-compile issues
Some checks failed
Build and Push Docker Image / build (push) Has been cancelled
Details 
The sm90 kernels use wgmma instructions that can't be compiled for
sm86/sm89 targets. Restricting to 8.0 (A100) and 9.0 (H100) only.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-03 21:57:09 +00:00
Debian
352128aa39 Pin SageAttention to commit that supports GPU-less builds
Some checks failed
Build and Push Docker Image / build (push) Failing after 21m35s
Details 
Recent commits broke TORCH_CUDA_ARCH_LIST support, requiring a GPU
during build. Pin to 2aecfa8 which respects the env var.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-03 20:30:17 +00:00
Debian
c61aca4074 Reduce build parallelism to avoid OOM during SageAttention compile
Some checks failed
Build and Push Docker Image / build (push) Has been cancelled
Details 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-03 20:17:10 +00:00
Debian
99fdda5b2b Add multi-GPU support and HTML test interface
Some checks failed
Build and Push Docker Image / build (push) Failing after 16m23s
Details 
- Update SageAttention CUDA arch list to support A100, A10, RTX 4090, L40, H100/H200
- Add interactive HTML test page for RunPod API testing

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-03 10:25:08 +00:00
Nick
f69bbc2f45 Fix model symlinks to use /runpod-volume/ComfyUI/models/
All checks were successful
Build and Push Docker Image / build (push) Successful in 4m3s
Details 
Models are stored in /runpod-volume/ComfyUI/models/ on the network
volume, not /runpod-volume/models/. Updated all symlinks to match.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-28 00:07:22 +13:00
Nick
47e312a58b Fix RIFE model download URL
All checks were successful
Build and Push Docker Image / build (push) Successful in 4m11s
Details 
Use correct URL from styler00dollar/VSGAN-tensorrt-docker releases.
Also fix path to ckpts/rife/ subdirectory.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-27 23:53:44 +13:00
Nick
342ef24bba Pre-download RIFE v4.7 model for frame interpolation
Some checks failed
Build and Push Docker Image / build (push) Failing after 3m47s
Details 
The RIFE model is small (~15MB) and required for the workflow.
Pre-downloading avoids runtime download delays.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-27 23:46:57 +13:00
Nick
aa95315e3f Add diffusion_models and text_encoders symlinks
Some checks failed
Build and Push Docker Image / build (push) Has been cancelled
Details 
WanVideo models are stored in diffusion_models/, and the CLIP text
encoder is in text_encoders/. These were missing from the symlink setup.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-27 23:45:13 +13:00
Nick
4b32558675 Add ComfyUI-Custom-Scripts and ComfyUI-Easy-Use nodes
All checks were successful
Build and Push Docker Image / build (push) Successful in 7m7s
Details 
Fixes workflow error: MathExpression|pysssss node not found.
These nodes are required by the Wan22-I2V-Remix workflow.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-27 00:45:44 +13:00
Nick
283955f1f7 Add ComfyUI-WanVideoWrapper for WanVideo nodes
All checks were successful
Build and Push Docker Image / build (push) Successful in 7m19s
Details 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-27 00:31:53 +13:00
Nick
ed5bf01972 Use /runpod-volume mount point for RunPod network volumes
All checks were successful
Build and Push Docker Image / build (push) Successful in 3m47s
Details 
RunPod mounts network volumes at /runpod-volume, not /userdata.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 23:01:08 +13:00
Nick
929059f812 Target only H200 (sm_90) for SageAttention build
All checks were successful
Build and Push Docker Image / build (push) Successful in 25m31s
Details 
Blackwell (sm_100) may not be fully supported yet.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 22:12:59 +13:00
Nick
fef4b8d7ee Reduce CUDA compilation parallelism for 16GB RAM
Some checks failed
Build and Push Docker Image / build (push) Failing after 15m53s
Details 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 16:58:39 +13:00
Nick
7a8b59f471 Target H200 and RTX 5090 GPU architectures
Some checks failed
Build and Push Docker Image / build (push) Failing after 47m51s
Details 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 12:11:31 +13:00
Nick
bdc1d769e8 Set TORCH_CUDA_ARCH_LIST for SageAttention build
Some checks failed
Build and Push Docker Image / build (push) Has been cancelled
Details 
Build runner has no GPU, so specify target architectures explicitly.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 12:09:57 +13:00
Nick
ab73c2d9be Optimize Dockerfile layers and reduce disk usage
Some checks failed
Build and Push Docker Image / build (push) Failing after 10m52s
Details 
- Combine PyTorch + triton install into single layer
- Add pip cache cleanup after each install step
- Change SageAttention to regular install and remove source after build
- Consolidate custom node dependencies into single layer
- Add CLAUDE.md, i2v-workflow.json, update handler.py and PROJECT.md

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-26 11:55:56 +13:00
Nick
469911f2a7 Fix pip bootstrap for Python 3.12
Some checks failed
Build and Push Docker Image / build (push) Has been cancelled
Details
2025-12-26 10:22:02 +13:00
Nick
65e9f0a6f6 Add deadsnakes PPA for Python 3.12
Some checks failed
Build and Push Docker Image / build (push) Failing after 29s
Details
2025-12-26 10:12:33 +13:00
Nick
a5adfe060e Initial commit: ComfyUI RunPod Serverless endpoint 
- Dockerfile with CUDA 12.8.1, Python 3.12, PyTorch 2.8.0+cu128
- SageAttention 2.2 compiled from source
- Nunchaku wheel installation
- 12 custom nodes pre-installed
- Handler with image/video output support
- Model symlinks to /userdata network volume

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-25 21:59:09 +13:00