freedak/timesfm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

minor update to the readme.

Browse Source
This commit is contained in:
Rajat Sen
2024-07-15 18:36:58 +00:00
parent 83815d9fb1
commit 23f7b64f8c
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+2 -2
View File
@@ -19,7 +19,7 @@ We recommend at least 16GB RAM to load TimesFM dependencies.
## Update - July 15, 2024 ## Update - July 15, 2024
- Launched [finetuning support](https://github.com/google-research/timesfm/blob/master/notebooks/finetuning.ipynb) that lets you finetune the weights of the pretrained TimesFM model on your own data. - Launched [finetuning support](https://github.com/google-research/timesfm/blob/master/notebooks/finetuning.ipynb) that lets you finetune the weights of the pretrained TimesFM model on your own data.
- Launched [~zero-shot covariate support](https://github.com/google-research/timesfm/blob/master/notebooks/covariates.ipynb) with external regressors. - Launched [~zero-shot covariate support](https://github.com/google-research/timesfm/blob/master/notebooks/covariates.ipynb) with external regressors. More details [here](https://github.com/google-research/timesfm?tab=readme-ov-file#covariates-support).
## Checkpoint timesfm-1.0-200m ## Checkpoint timesfm-1.0-200m
@@ -241,7 +241,7 @@ In this example, besides the `Daily_sales`, we also have covariates `Category`,
**Notice:** Here we make it mandatory that the dynamic covariates need to cover both the forecasting context and horizon. For example, all dynamic covariates in the example have 14 values: the first 7 correspond to the observed 7 days, and the last 7 correspond to the next 7 days. **Notice:** Here we make it mandatory that the dynamic covariates need to cover both the forecasting context and horizon. For example, all dynamic covariates in the example have 14 values: the first 7 correspond to the observed 7 days, and the last 7 correspond to the next 7 days.
We can now provide the past data of the two products along with static and dynamic covariates as a batch input to TimesFM and produc forecasts that take into the account the covariates. To learn more, check out the example in `notebooks/covariates.ipynb`. We can now provide the past data of the two products along with static and dynamic covariates as a batch input to TimesFM and produce forecasts that take into the account the covariates. To learn more, check out the example in `notebooks/covariates.ipynb`.
## Finetuning ## Finetuning