Instructions to use FinText/TimesFM_8M_2000_Global with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use FinText/TimesFM_8M_2000_Global with Transformers:
# Load model directly from transformers import AutoTokenizer, TimesFMForHF tokenizer = AutoTokenizer.from_pretrained("FinText/TimesFM_8M_2000_Global") model = TimesFMForHF.from_pretrained("FinText/TimesFM_8M_2000_Global", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| pipeline_tag: time-series-forecasting | |
| license: apache-2.0 | |
| tags: | |
| - TSFM | |
| - Finance | |
| - Financial Forecasting | |
| - FinText | |
| library_name: transformers | |
| ## TimesFM-8M (TSFM) β Global (2000) | |
| This is the **Time Series Foundation Model (TSFM)**, pre-trained on **global financial time series data up to the year 2000** using the **TimesFM architecture (8M)**. The dataset spans from **1990β2000** and includes all **global excess return data**. | |
| π **Related Links** | |
| - [π View all TimesFM (8M) models - Global](https://huggingface.co/collections/FinText/timesfm-8m-global) | |
| - [π View all TSFMs](https://huggingface.co/FinText/collections) | |