Add flux example

[mengniwang95]

User description

Type of Change

example

Description

detail description

Expected Behavior & Potential Risk

the expected behavior that triggered by this PR

How has this PR been tested?

how to reproduce the test (including hardware information)

Dependency Change?

any library dependency introduced or removed

---

PR Type

Enhancement

---

Description

• Added parameters for diffusion control in quantization

• Updated initialization and conversion methods to include new parameters

• Modified autoround quantize entry to handle new parameters

---

Diagram Walkthrough

flowchart LR
  init["Add diffusion parameters"] -- "Update initialization" --> convert["Modify conversion method"]
  convert -- "Handle new parameters" --> entry["Update autoround entry"]

Loading

File Walkthrough

Relevant files

Enhancement

autoround.py Add diffusion parameters to autoround                                        neural_compressor/torch/algorithms/weight_only/autoround.py Added guidance_scale, num_inference_steps, generator_seed to __init__ Updated convert method to accept pipeline and use new parameters +18/-1    algorithm_entry.py Update autoround entry for diffusion                                          neural_compressor/torch/quantization/algorithm_entry.py Added dataset, guidance_scale, num_inference_steps, generator_seed to autoround_quantize_entry Updated get_quantizer call to include new parameters +8/-0      quantize.py Allow additional keyword arguments in convert                        neural_compressor/torch/quantization/quantize.py Added **kwargs to convert method +2/-0

---

[PRAgent4INC]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Possible Issue The generator_seed parameter is set to None by default, which might lead to non-reproducible results if not explicitly set by the user. generator_seed: int = None, Code Clarity The addition of the pipeline parameter in the convert method could benefit from additional comments or documentation to clarify its purpose and usage. pipe = kwargs.pop("pipeline", None) Default Dataset Setting the default dataset to "NeelNanda/pile-10k" might not be suitable for all use cases. Consider making this configurable or providing a more generic default. dataset = quant_config.to_dict().get("dataset", "NeelNanda/pile-10k")

[PRAgent4INC]

PR Code Suggestions ✨