Fastgen add method vs Fastgen add modality
Fastgen add method is an open-source agent skill for ai models and llm tooling. Fastgen add modality is an open-source agent skill for design and media.
- Fastgen add method is best for implement a new distillation objective in fastgen from a paper.
- Fastgen add modality is best for add a new conditioning signal such as depth or keypoints to fastgen training.
- What it is
- Fastgen add method is an open-source agent skill for ai models and llm tooling.Fastgen add modality is an open-source agent skill for design and media.
- Best for
- Implement a new distillation objective in FastGen from a paperAdd a new conditioning signal such as depth or keypoints to FastGen training
- Kind
- SkillSkill
- Pricing
- Open sourceOpen source
- Platforms
- Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, GooseClaude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
- GitHub stars
- 4K4K
- Last commit
- Nov 2026Nov 2026
- License
- Apache-2.0Apache-2.0
- Open source
- Good for
- Implement a new distillation objective in FastGen from a paper; Add a consistency-style training method to the FastGen codebase; Check gradients of a new loss term against reference codeAdd a new conditioning signal such as depth or keypoints to FastGen training; Make FastGen generate a new media type alongside image, video and audio; Wire a new data modality through FastGen's loader, model and output interfaces
What Fastgen add method does
Method planning checklist
Guides the agent to record state conventions, target construction, detach boundaries and loss weighting before coding.
Reuse-first decision
Directs the agent to check whether existing method configs already cover the algorithm before adding code.
Gradient-level verification
Compares each objective term's value and gradients with the reference on fixed tensors and boundary times.
Typed method config
Explains adding a FastGenModel subclass, a typed config and its export in the methods package.
Training and checkpoint checks
Covers bounded runs across optimizer phases, frozen targets, validation state and resume behaviour.
What Fastgen add modality does
Modality role analysis
Determines whether a new signal is generated state, conditioning, control, or a derived training target before implementation.
End-to-end sample tracing
Follows one sample through loader, batch preparation, network, loss, sampling and output to locate unsupported boundaries.
Model and conditioning changes
Covers shape inference, schedule dispatch and conditioning paths while delegating encoder and backbone work to a companion skill.
Data serialization checks
Guides acquisition, pairing, loader registration and normalisation versioning with encode and decode verification.
Method compatibility review
Requires tracing each training method's tensor operations and rejecting unsupported combinations explicitly.
Validation and reporting
Covers gradient checks, conditioning sensitivity, serialization, checkpoint reload and reporting of tested modes.
Fastgen add method vs Fastgen add modality: questions
Is Fastgen add method or Fastgen add modality cheaper?
Both are open source. Compare current plans on each official site.
What is the difference between Fastgen add method and Fastgen add modality?
Fastgen add method is an open-source agent skill for ai models and llm tooling. Fastgen add modality is an open-source agent skill for design and media.
