genQC logo genQC
  • Overview
  • Get Started
  • Tutorials
  • API Reference
  • Research
  • Code Repository
  1. Pipeline
  2. Diffusion Pipeline

API Reference

  • Modules Overview
  • Release notes

  • Benchmark
    • Compilation benchmark
  • Dataset
    • Dataset balancing
    • Cached dataset
    • Quantum circuit dataset
    • Config dataset
    • Dataset helper functions
    • Mixed cached dataset
  • Inference
    • Evaluation metrics
    • Evaluation helper
    • Sampling functions
  • Models
    • Config model
    • Frozen OpenCLIP
    • Layers
    • Position encodings
    • Conditional qc-UNet
    • Encoder for unitaries
    • Clip
      • Frozen OpenCLIP
      • Unitary CLIP
    • Embedding
      • Base embedder
      • Rotational preset embedder
    • Transformers
      • Transformers and attention
      • CirDiT - Circuit Diffusion Transformer
      • Transformers
  • Pipeline
    • Callbacks
    • Compilation Diffusion Pipeline
    • Diffusion Pipeline
    • Diffusion Pipeline Special
    • Metrics
    • Multimodal Diffusion Pipeline
    • Pipeline
    • Unitary CLIP Pipeline
  • Platform
    • Circuits dataset generation functions
    • Circuits instructions
    • Simulation backend
    • Backends
      • Base backend
      • CUDA-Q circuits backend
      • Pennylane circuits backend
      • Qiskit circuits backend
    • Tokenizer
      • Base tokenizer
      • Circuits tokenizer
      • Tensor tokenizer
  • Scheduler
    • Scheduler
    • DDIM Scheduler
    • DDPM Scheduler
    • DPM Scheduler
  • Utils
    • Async functions
    • Config loader
    • Math and algorithms
    • Miscellaneous util

On this page

  • DiffusionPipeline
  • Report an issue
  • View source
  1. Pipeline
  2. Diffusion Pipeline

Diffusion Pipeline


source

DiffusionPipeline

 DiffusionPipeline (scheduler:genQC.scheduler.scheduler.Scheduler,
                    model:torch.nn.modules.module.Module,
                    text_encoder:torch.nn.modules.module.Module,
                    embedder:torch.nn.modules.module.Module,
                    device:torch.device, enable_guidance_train=True,
                    guidance_train_p=0.1, cached_text_enc=True)

A Pipeline for diffusion models. Implements train and inference functions. Diffusion parameters are defined inside a Scheduler object.

Type Default Details
scheduler Scheduler
model Module
text_encoder Module
embedder Module clr embeddings or a VAE for latent diffusion
device device
enable_guidance_train bool True
guidance_train_p float 0.1
cached_text_enc bool True
Back to top
Compilation Diffusion Pipeline
Diffusion Pipeline Special
 

Copyright 2025, Florian Fürrutter

  • Report an issue
  • View source