Update README.md
Browse files
README.md
CHANGED
|
@@ -17,7 +17,7 @@ metrics:
|
|
| 17 |
- f1
|
| 18 |
---
|
| 19 |
|
| 20 |
-
#
|
| 21 |
|
| 22 |
An advanced Speech Emotion Recognition (SER) deep learning model designed to detect the emotional tone of Formula 1 driver team radio communications in real-time.
|
| 23 |
|
|
@@ -75,9 +75,9 @@ The model processes raw audio input from driver-to-pit-wall communications, extr
|
|
| 75 |
|
| 76 |
---
|
| 77 |
|
| 78 |
-
##
|
| 79 |
|
| 80 |
-
###
|
| 81 |
For users directly loading the PyTorch `BiLSTMAttentionClassifier` model and `.pt` weights:
|
| 82 |
|
| 83 |
- **Model Input:**
|
|
@@ -89,7 +89,7 @@ For users directly loading the PyTorch `BiLSTMAttentionClassifier` model and `.p
|
|
| 89 |
|
| 90 |
---
|
| 91 |
|
| 92 |
-
###
|
| 93 |
For users using the complete audio-to-prediction helper script:
|
| 94 |
|
| 95 |
- **Input Requirements:**
|
|
@@ -162,7 +162,7 @@ print("Full Probabilities:", result['probabilities'])
|
|
| 162 |
|
| 163 |
---
|
| 164 |
|
| 165 |
-
##
|
| 166 |
|
| 167 |
Evaluation performed on holdout test set (1,117 audio evaluation samples):
|
| 168 |
|
|
@@ -175,16 +175,16 @@ Evaluation performed on holdout test set (1,117 audio evaluation samples):
|
|
| 175 |
| **Feature Encoder** | `microsoft/wavlm-base-plus` (768-dim, Frozen) |
|
| 176 |
| **Classifier Head** | BiLSTM (128) + Temporal Attention + Linear (6) |
|
| 177 |
|
| 178 |
-
###
|
| 179 |
|
| 180 |
| Emotion Class | Precision | Recall | F1-Score | Evaluation Support |
|
| 181 |
| :--- | :--- | :--- | :--- | :--- |
|
| 182 |
-
| **Anger**
|
| 183 |
-
| **Neutral**
|
| 184 |
-
| **Happy**
|
| 185 |
-
| **Disgust**
|
| 186 |
-
| **Fear**
|
| 187 |
-
| **Sad**
|
| 188 |
|
| 189 |
---
|
| 190 |
|
|
|
|
| 17 |
- f1
|
| 18 |
---
|
| 19 |
|
| 20 |
+
# F1 Driver Tone & Emotion Detector
|
| 21 |
|
| 22 |
An advanced Speech Emotion Recognition (SER) deep learning model designed to detect the emotional tone of Formula 1 driver team radio communications in real-time.
|
| 23 |
|
|
|
|
| 75 |
|
| 76 |
---
|
| 77 |
|
| 78 |
+
## Input & Output Specifications
|
| 79 |
|
| 80 |
+
### Direct PyTorch Model I/O (`models.py` + `best_model.pt`)
|
| 81 |
For users directly loading the PyTorch `BiLSTMAttentionClassifier` model and `.pt` weights:
|
| 82 |
|
| 83 |
- **Model Input:**
|
|
|
|
| 89 |
|
| 90 |
---
|
| 91 |
|
| 92 |
+
### Full Pipeline I/O (Using `modeling_f1tone.py`)
|
| 93 |
For users using the complete audio-to-prediction helper script:
|
| 94 |
|
| 95 |
- **Input Requirements:**
|
|
|
|
| 162 |
|
| 163 |
---
|
| 164 |
|
| 165 |
+
## Verified Model Evaluation Results
|
| 166 |
|
| 167 |
Evaluation performed on holdout test set (1,117 audio evaluation samples):
|
| 168 |
|
|
|
|
| 175 |
| **Feature Encoder** | `microsoft/wavlm-base-plus` (768-dim, Frozen) |
|
| 176 |
| **Classifier Head** | BiLSTM (128) + Temporal Attention + Linear (6) |
|
| 177 |
|
| 178 |
+
### Per-Class Performance Breakdown (Test Set)
|
| 179 |
|
| 180 |
| Emotion Class | Precision | Recall | F1-Score | Evaluation Support |
|
| 181 |
| :--- | :--- | :--- | :--- | :--- |
|
| 182 |
+
| **Anger** | **76.71%** | **87.96%** | **0.8195** | 191 samples |
|
| 183 |
+
| **Neutral** | **77.85%** | **75.46%** | **0.7664** | 163 samples |
|
| 184 |
+
| **Happy** | **79.05%** | **61.26%** | **0.6903** | 191 samples |
|
| 185 |
+
| **Disgust** | **64.65%** | **67.02%** | **0.6581** | 191 samples |
|
| 186 |
+
| **Fear** | **52.05%** | **66.84%** | **0.5853** | 190 samples |
|
| 187 |
+
| **Sad** | **63.33%** | **49.74%** | **0.5572** | 191 samples |
|
| 188 |
|
| 189 |
---
|
| 190 |
|