Update README.md
Browse filesremoved gating information
updated title
fixed TiRex-2 Pro bullet points
README.md
CHANGED
|
@@ -8,7 +8,7 @@ license: apache-2.0
|
|
| 8 |
---
|
| 9 |
<div align="left" class="flex items-baseline gap-2">
|
| 10 |
<img src="tirex.svg" alt="TiRex mascot" width="33" height="40" />
|
| 11 |
-
<h1 class="m-0">TiRex-2</h1>
|
| 12 |
</div>
|
| 13 |
|
| 14 |
This repository provides the pretrained TiRex-2 model and inference code for zero-shot
|
|
@@ -42,22 +42,6 @@ multivariate forecasting with past and future covariates.
|
|
| 42 |
> 📖 For a detailed guide — including pip installation, a Google Colab demo, covariate
|
| 43 |
> examples, and benchmark reproduction — see our [GitHub repository](https://github.com/NX-AI/tirex-2).
|
| 44 |
|
| 45 |
-
### Access to Model Weights
|
| 46 |
-
|
| 47 |
-
TiRex-2's model weights are gated on Hugging Face. To be able to access them, either log in via the Hugging Face CLI, or [generate yourself a Hugging Face access token](https://huggingface.co/settings/tokens/new?canReadGatedRepos=true&tokenType=fineGrained) (make sure to enable Read access to contents of all public gated repos you can access) and set it before loading the model:
|
| 48 |
-
|
| 49 |
-
CLI:
|
| 50 |
-
```bash
|
| 51 |
-
huggingface-cli login
|
| 52 |
-
```
|
| 53 |
-
|
| 54 |
-
Access Token:
|
| 55 |
-
|
| 56 |
-
```bash
|
| 57 |
-
import os
|
| 58 |
-
os.environ["HF_TOKEN"] = "<insert-hf-token>"
|
| 59 |
-
```
|
| 60 |
-
|
| 61 |
The environment is managed by [Pixi](https://pixi.prefix.dev/latest/). Run the following to install it on your machine
|
| 62 |
|
| 63 |
```bash
|
|
@@ -89,11 +73,12 @@ TiRex-2 already provides state-of-the-art performance for zero-shot prediction,
|
|
| 89 |
|
| 90 |
Our pro version extends TiRex-2 with additional capabilities, including:
|
| 91 |
|
| 92 |
-
Streaming: incremental forecast updates as new observations arrive, without recomputing over the full history.
|
| 93 |
-
Speed: performance-optimized inference, including optimization for dedicated hardware such as edge, embedded, and industrial PC deployments.
|
| 94 |
-
Finetuning: models fine-tuned on your data or with different pretraining.
|
| 95 |
-
Classification & Regression: TiRex-2 adapted for classification and regression tasks.
|
| 96 |
-
|
|
|
|
| 97 |
|
| 98 |
## Cite
|
| 99 |
If you use TiRex-2 in your research, please cite our work:
|
|
|
|
| 8 |
---
|
| 9 |
<div align="left" class="flex items-baseline gap-2">
|
| 10 |
<img src="tirex.svg" alt="TiRex mascot" width="33" height="40" />
|
| 11 |
+
<h1 class="m-0">TiRex-2: Generalizing TiRex to Multivariate Data and Streaming</h1>
|
| 12 |
</div>
|
| 13 |
|
| 14 |
This repository provides the pretrained TiRex-2 model and inference code for zero-shot
|
|
|
|
| 42 |
> 📖 For a detailed guide — including pip installation, a Google Colab demo, covariate
|
| 43 |
> examples, and benchmark reproduction — see our [GitHub repository](https://github.com/NX-AI/tirex-2).
|
| 44 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
The environment is managed by [Pixi](https://pixi.prefix.dev/latest/). Run the following to install it on your machine
|
| 46 |
|
| 47 |
```bash
|
|
|
|
| 73 |
|
| 74 |
Our pro version extends TiRex-2 with additional capabilities, including:
|
| 75 |
|
| 76 |
+
- **Streaming**: incremental forecast updates as new observations arrive, without recomputing over the full history.
|
| 77 |
+
- **Speed**: performance-optimized inference, including optimization for dedicated hardware such as edge, embedded, and industrial PC deployments.
|
| 78 |
+
- **Finetuning**: models fine-tuned on your data or with different pretraining.
|
| 79 |
+
- **Classification & Regression**: TiRex-2 adapted for classification and regression tasks.
|
| 80 |
+
|
| 81 |
+
If you are interested in any of these, please contact us at [contact@nx-ai.com](mailto:contact@nx-ai.com).
|
| 82 |
|
| 83 |
## Cite
|
| 84 |
If you use TiRex-2 in your research, please cite our work:
|