Modify default parameters. Update README.md

README.md

@@ -24,8 +24,8 @@ An efficient model to detect chainsaw activity in forest soundscapes using spect
 To use the model, clone the repository and install the dependencies:
 
 ```bash
-git clone https://huggingface.co/your_username/your_model_name
-cd your_model_name
+git clone https://huggingface.co/tlmk22/QuefrencyGuardian
+cd QuefrencyGuardian
 pip install -r requirements.txt
 ```
 
@@ -34,13 +34,14 @@ pip install -r requirements.txt
 ### Features
 
 The model uses:
-- **Spectrogram Features**: Extracted from frequencies between 70-1525 Hz.
-- **Cepstral Features**: Calculated as the FFT of the log spectrogram in a filtered quefrency range.
-- **Time Averaging**: Both feature sets are averaged over the time domain for robustness in noisy settings.
+- **Spectrogram Features**
+- **Cepstral Features**: Calculated as the FFT of the log spectrogram between [`f_min`-`f_max`] in a filtered quefrency range  [`fc_min`-`fc_max`].
+- **Time Averaging**: Both feature sets are averaged along the whole audio clip for robustness in noisy settings (Welch methodology)
 
 ### LightGBM Model
 
-The model is a **binary classifier** (chainsaw vs environment) trained on the `rfcx/frugalai` dataset. Key model parameters are included in `model/lgbm_params.json`.
+The model is a **binary classifier** (chainsaw vs environment) trained on the `rfcx/frugalai` dataset. 
+Key model parameters are included in `model/lgbm_params.json`.
 
 ## Usage
 
@@ -62,7 +63,7 @@ model = FastModel(
     feature_params=features,
     lgbm_params=lgbm_params,
     model_file="model/model.txt",  # Path to the serialized model file
-    device="cuda"                  # Use 'cpu' if GPU is unavailable
+    device="cuda", 
 )
 
 # Predict on a Dataset
@@ -84,10 +85,6 @@ This project is licensed under the [Creative Commons Attribution Non-Commercial
 
 ---
 
-## File Structure
-📂 your_model_name/ ├── 📂 model/ │ ├── model.txt # Pre-trained LightGBM model │ ├── features.json # Feature extraction parameters │ ├── lgbm_params.json # LightGBM parameters ├── 📜 README.md # Documentation ├── 📜 LICENSE.md # CC BY-NC 4.0 license ├── 📜 requirements.txt # Python dependencies └── 📜 fast_model.py # Model implementation
-
-
 ## Dataset
 
 The model was trained and evaluated on the [Rainforest Connection (RFCx) Frugal AI](https://huggingface.co/datasets/rfcx/frugalai) dataset.

fast_model.py

@@ -87,14 +87,14 @@ class FastModel:
         self,
         feature_params,
         lgbm_params=None,
-        padding_method="zero",
+        padding_method="reflect",
         waveform_duration=3,
         model_file=None,
         mask_features=False,
         mask_file="feature_mask.pkl",
         mask_ratio=0.25,
         batch_size=5000,
-        apply_offset_on_fit=True,
+        apply_offset_on_fit=False,
         device="cpu",
     ):
         self.feature_params = feature_params