-
Notifications
You must be signed in to change notification settings - Fork 332
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
docs: Document how to use the conversion script
Provide a detailed instruction on how to use the script
- Loading branch information
1 parent
99130a9
commit 08eaa76
Showing
1 changed file
with
24 additions
and
4 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,9 +1,29 @@ | ||
# Official Pytorch implementation of TimesFM | ||
|
||
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google | ||
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google | ||
Research for time-series forecasting. | ||
|
||
* Paper: [A decoder-only foundation model for time-series forecasting](https://arxiv.org/abs/2310.10688), to appear in ICML 2024. | ||
* [Google Research blog](https://research.google/blog/a-decoder-only-foundation-model-for-time-series-forecasting/) | ||
- Paper: [A decoder-only foundation model for time-series forecasting](https://arxiv.org/abs/2310.10688), to appear in ICML 2024. | ||
- [Google Research blog](https://research.google/blog/a-decoder-only-foundation-model-for-time-series-forecasting/) | ||
|
||
## Stay tuned for all of the functionalities as that of the pax version. | ||
--- | ||
|
||
## Converting the model from pax to torch | ||
|
||
Before you start, make sure you are in the `timesfm` directory. If not, navigate to the `timesfm` directory. Do not forget to install the required specified in the `experiments/environment.yaml` or `experiments/environment_cpu.yaml`. | ||
|
||
1. Download the model checkpoint from the [official repository](https://huggingface.co/google/timesfm-1.0-200m) if you haven't already. | ||
```bash | ||
git lfs install | ||
git clone https://huggingface.co/google/timesfm-1.0-200m | ||
``` | ||
2. Run the convertion script. You can specify the `model_path` and `output_path` as per your requirement. Here is an example (also the default values): | ||
```bash | ||
# In `timesfm` directory | ||
python3 -m src.timesfm_torch.convert_weights --model_path=timesfm-1.0-200m/checkpoints --output_path=ckpt/timesfm-1.0-200m.pth | ||
``` | ||
3. The converted model will be saved in the `timesfm/ckpt/timesfm-1.0-200m.pth`. | ||
|
||
--- | ||
|
||
## Stay tuned for all of the functionalities as that of the pax version. |