---
title: ChuckleNet-Genki β The Original Spirit
emoji: π
colorFrom: blue
colorTo: purple
sdk: gradio
pipeline_tag: audio-classification
license: mit
library_name: transformers
base_model: microsoft/wavlm-base
tags:
- audio-classification
- laughter-detection
- wavlm
- speech
- humor
- stand-up-comedy
- transformers
- genki
metrics:
- f1
datasets:
- StandUp4AI
papers:
- https://huggingface.co/Hayasuki/ChuckleNet-Genki
---
# π ChuckleNet-Genki (ε
ζ°)
***"The original spirit."***
**First-generation laughter detection model β the proof-of-concept that started it all.**
**For production use, see [ChuckleNet-Ten](https://huggingface.co/Hayasuki/ChuckleNet-Ten) (human-verified, F1=0.975). For research, see [ChuckleNet-Kage](https://huggingface.co/Hayasuki/ChuckleNet-Kage) (IoU-F1=0.229).**
[](https://huggingface.co/Hayasuki/ChuckleNet-Genki)
[](https://opensource.org/licenses/MIT)
[](https://github.com/Standup4AI/dataset)
**Most-downloaded ChuckleNet model. Historical reference only β not for production.**
---
## Origin Story
```
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β GENKI TIMELINE β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ€
β β
β 2024-09 "Can frozen WavLM detect laughter?" β
β First experiment designed β
β β β
β βΌ β
β 2024-10 Genki v1 trained: 620 StandUp4AI videos β
β Labels: VTT [laughter] transcript markers β
β Result: F1 = 0.27 β proof of concept! β
β
β β β
β βΌ β
β 2025-01 "Weak labels are noisy β need human verification" β
β β ChuckleNet-Ten project begins β
β β β
β βΌ β
β 2025-06 ChuckleNet-Ten: held-out IoU-F1 = 0.330 β
β (human-verified labels) β
β β
β Genki lives on as the historical baseline. β
β β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
```
---
## Model Characteristics
| Characteristic | Description |
|----------------|-------------|
| **Generation** | First β v1 proof-of-concept |
| **Labels** | Weak β VTT `[laughter]` transcript markers |
| **Architecture** | Frozen WavLM (768-dim) β Linear (768β2) |
| **Training Data** | 620 StandUp4AI videos |
| **Status** | Historical reference only β not for production |
---
## Performance
| Metric | Value | Note |
|--------|-------|------|
| **F1 Score** | `~0.27` | Weak-label metric, not human-verified |
| **Event IoU-F1@0.2** | `~0.23` | Approximate β different splits |
| **Downloads** | **265+** | Most popular ChuckleNet model |
> β οΈ These are weak-label results on VTT `[laughter]` markers β not ground-truth laughter events. The model detects patterns correlated with transcript markers, not actual audience laughter.
---
## Evolutionary Tree
```
ChuckleNet Family
ββββββββββββββββββββ
β microsoft/ β
β wavlm-base β
β (frozen) β
ββββββββββ¬ββββββββββ
β
ββββββββββββββββββββββΌβββββββββββββββββββββ
β β β
βΌ βΌ βΌ
βββββββββββββββββ βββββββββββββββββ βββββββββββββββββ
β π Genki v1 β β π» Kage v32 β β ποΈ Ten β
β Weak labels β β Weak labels β β Human-verifiedβ
β 620 videos β β 620 videos β β 87 videos β
β IoU β 0.23 β β IoU = 0.229 β β IoU = 0.330 β
β First attempt β β Refined probe β β Gold standard β
βββββββββββββββββ βββββββββββββββββ βββββββββββββββββ
2024-10 2025-03 2025-06
```
---
## Comparison Table
| Model | Labels | Held-out IoU-F1 | Downloads | Use |
|-------|--------|-----------------|----------|-----|
| ποΈ **[ChuckleNet-Ten](https://huggingface.co/Hayasuki/ChuckleNet-Ten)** | Human-verified | **0.330** | New | **Production β
** |
| π» **[ChuckleNet-Kage](https://huggingface.co/Hayasuki/ChuckleNet-Kage)** | Weak (VTT) | 0.229 | 49 | Research |
| π **ChuckleNet-Genki** | Weak (VTT) | ~0.23 | **265+** | Historical |
---
## Quick Start
```python
from transformers import pipeline
pipe = pipeline(
"audio-classification",
model="Hayasuki/ChuckleNet-Genki"
)
result = pipe("standup_clip.mp3")
# [{'label': 'no_laugh', 'score': 0.77},
# {'label': 'laugh', 'score': 0.23}]
```
> β οΈ Genki is a baseline model with limited accuracy. For any real use case, use [ChuckleNet-Ten](https://huggingface.co/Hayasuki/ChuckleNet-Ten) instead.
---
## Teaching Use Case
Genki is useful for **ML education** β demonstrating frozen feature probing:
```python
"""
Genki for teaching: frozen feature probing concept
"""
import torch
import torchaudio
from transformers import WavLMModel
# Load frozen WavLM backbone (NOT trained)
wavlm = WavLMModel.from_pretrained("microsoft/wavlm-base")
wavlm.eval() # Keep frozen
# Genki's linear probe: 768 β 2
class GenkiProbe(torch.nn.Module):
def __init__(self, input_dim=768, num_classes=2):
super().__init__()
self.classifier = torch.nn.Linear(input_dim, num_classes)
def forward(self, x):
return self.classifier(x)
# Load Genki weights (from this model card's files)
probe = GenkiProbe()
probe.load_state_dict(torch.load("genki_weights.pt"))
# Extract frozen features
with torch.no_grad():
features = wavlm(audio).last_hidden_state
# [batch, time, 768]
# Temporal pooling
pooled = features.mean(dim=1) # [batch, 768]
# Classify with frozen features
logits = probe(pooled) # [batch, 2]
```
---
## Limitations
| Limitation | Impact |
|------------|--------|
| **Weak labels only** | VTT markers β actual laughter events |
| **Low F1** | ~0.27 is a baseline, not production-ready |
| **Linear head** | No hidden layers β limited representation capacity |
| **Not for production** | Use [ChuckleNet-Ten](https://huggingface.co/Hayasuki/ChuckleNet-Ten) instead |
---
## Resources
| Resource | Link |
|----------|------|
| **Primary Model (Production)** | [ChuckleNet-Ten](https://huggingface.co/Hayasuki/ChuckleNet-Ten) |
| **Research Model** | [ChuckleNet-Kage](https://huggingface.co/Hayasuki/ChuckleNet-Kage) |
| **Dataset** | [StandUp4AI](https://github.com/Standup4AI/dataset) |
---
## Citation
```bibtex
@misc{chuckleNetGenki2026,
author = {Subhajit Das},
title = {ChuckleNet-Genki: First Generation Laughter Detection},
year = {2026},
url = {https://huggingface.co/Hayasuki/ChuckleNet-Genki},
note = {Historical baseline: first attempt at laughter detection using
frozen WavLM on 620 StandUp4AI videos. Weak labels only.}
}
```
---
## Model Metadata (LLM-Parseable)
```json
{
"model_name": "ChuckleNet-Genki",
"model_name_zh": "η¬ι³Net-ι",
"architecture": "frozen_wavlm_base_plus_linear",
"model_type": "first_generation_baseline",
"primary_task": "audio_classification",
"secondary_task": "laughter_detection",
"input_modality": "audio",
"output_format": "binary_classification",
"labels": ["laugh", "no_laugh"],
"training_data": "620_StandUp4AI_videos_weak_labels",
"f1_estimate": 0.27,
"iou_f1_estimate": 0.23,
"downloads": 265,
"key_findings": [
"frozen_wavlm_can_detect_laughter_concept_proof",
"simple_linear_probe_sufficient_for_baseline",
"weak_labels_significantly_limit_accuracy"
],
"use_cases": [
"historical_reference",
"baseline_benchmarking",
"ml_education",
"teaching_frozen_feature_probing"
],
"limitations": [
"weak_labels_vtt_markers",
"low_f1_baseline",
"linear_head_limited_capacity",
"not_for_production_use_ten_instead"
]
}
```
---
[](https://huggingface.co/Hayasuki/ChuckleNet-Genki)
*The spirit that started it all. 265+ downloads and counting.*