FreqFault-base

Inference checkpoint for the base FreqFault time-frequency encoder.

Item	Value
Time embedding dimension	128
Frequency embedding dimension	128
Fused feature dimension	256
Attention heads	4
Layers per branch	4
Parameters	1.67M
Weight size	6.69 MB
Frequency bins	800
Input window	0.1 seconds

Usage

import numpy as np
import torch
from model import UniFreqProcessor, load_model

model, config = load_model("bojian1/FreqFault-base")
batch = UniFreqProcessor(config)([
    np.random.randn(2048).astype("float32")
])

with torch.inference_mode():
    features = model(batch)

Inference code and documentation: https://github.com/Bojian-Chen/FreqFault

This repository contains inference artifacts only: config.json and model.safetensors.

The repository was renamed from bojian1/UniFreq-base. Python API names (UniFreqProcessor, UniFreqModel) and architecture identifiers remain compatible with existing code. Model weights are unchanged.