STEAD subsample 4 CDiffSD

# STEAD Subsample Dataset for CDiffSD Training

## OverviewThis dataset is a subsampled version of the STEAD dataset, specifically tailored for training our CDiffSD model (Cold Diffusion for Seismic Denoising). It consists of four CSV files, each saved in a format that requires Python's `pd.read_pickle` method for opening.

## Dataset FilesThe dataset includes the following files:

df_train: Used for both training and validation phases (with validation train split). Contains earthquake ground truth traces.

df_noise_train: Used for both training and validation phases. Contains noise used to contaminate the traces.

df_test: Used for the testing phase, structured similarly to df_train.

df_noise_test: Used for the testing phase, contains noise data for testing.

Each file is structured to support the training and evaluation of seismic denoising models.

## Data ColumnsThe dataset files contain columns from the original STEAD dataset, with specific focus on the following columns that are critical for our use:

E_channel: East-West channel data

N_channel: North-South channel data

Z_channel: Vertical channel data

s_arrival_sample: Sample index for S-wave arrival

p_arrival_sample: Sample index for P-wave arrival

trace_name: Identifier for the seismic trace

## UsageTo load these files in a Python environment, use the following approach:

```pythonimport pandas as pd

# Example of loading the training datadf_train = pd.read_pickle('path/to/df_train.csv')```

Ensure that the path to the file is correctly specified relative to your Python script.

## RequirementsTo use this dataset, ensure you have Python installed along with the Pandas library, which can be installed via pip if not already available:

```bashpip install pandas```