Spaces:
Sleeping
Sleeping
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,10 +1,43 @@
|
|
| 1 |
-
---
|
| 2 |
-
title: Chronos Zero
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
-
sdk: docker
|
| 7 |
-
|
| 8 |
-
---
|
| 9 |
-
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: Chronos-2 Zero-Shot Demo
|
| 3 |
+
emoji: 📈
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: red
|
| 6 |
+
sdk: docker
|
| 7 |
+
app_port: 7860
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# Chronos-2 Zero-Shot Forecasting Demo
|
| 11 |
+
|
| 12 |
+
Interactive demo showcasing Amazon Chronos-2's zero-shot time series forecasting capability.
|
| 13 |
+
|
| 14 |
+
## Features
|
| 15 |
+
|
| 16 |
+
- **Zero-Shot Forecasting**: No training or fine-tuning required
|
| 17 |
+
- **Real-time Prediction**: Input data and get instant forecasts
|
| 18 |
+
- **Probabilistic Output**: Median prediction with 80% confidence interval
|
| 19 |
+
- **Comparison Tab**: Visual comparison of traditional ML vs Chronos pipeline
|
| 20 |
+
|
| 21 |
+
## Usage
|
| 22 |
+
|
| 23 |
+
1. Enter time series values (comma or newline separated)
|
| 24 |
+
2. Set prediction steps (1-24)
|
| 25 |
+
3. Get instant forecast with confidence intervals
|
| 26 |
+
|
| 27 |
+
## Model
|
| 28 |
+
|
| 29 |
+
- **Chronos-2** (amazon/chronos-2): 120M parameter encoder-only foundation model
|
| 30 |
+
- Supports context length up to 8192 tokens
|
| 31 |
+
- CPU inference (GPU optional)
|
| 32 |
+
|
| 33 |
+
## Local Development
|
| 34 |
+
|
| 35 |
+
```bash
|
| 36 |
+
pip install -r requirements.txt
|
| 37 |
+
uvicorn server:app --reload --port 7860
|
| 38 |
+
```
|
| 39 |
+
|
| 40 |
+
## References
|
| 41 |
+
|
| 42 |
+
- [Chronos-2 GitHub](https://github.com/amazon-science/chronos-forecasting)
|
| 43 |
+
- [Chronos-2 Paper](https://arxiv.org/abs/2410.03847)
|