# Force running experiment from a specified stage

**URL:** <https://discuss.dvc.org/t/force-running-experiment-from-a-specified-stage/2142>\
**Category:** Questions\
**Created:** [July 7, 2024, 11:30pm UTC](https://discuss.dvc.org/t/force-running-experiment-from-a-specified-stage/2142 "2024-07-07T23:30:29Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![hjoshi](https://avatars.discourse-cdn.com/v4/letter/h/87869e/32.png) [@hjoshi](https://discuss.dvc.org/u/hjoshi)\
**Post date:** [July 7, 2024, 11:30pm UTC](https://discuss.dvc.org/t/force-running-experiment-from-a-specified-stage/2142/1 "2024-07-07T23:30:29Z")

</div>

I have a pipeline that looks something like this

```auto
prep_data --> train

train --> evaluate_test_data
train --> evaluate_holdout_data
train --> training_report

```

I am iterating the `train` step. For instance, I added logging of feature importance (using log\_plot) and logging of hyperparameters (using log\_params)

I have run this experiment before, so naturally the previously generated train, test data and model are cached.

After making changes, when I run the experiment using `dvc exp run`, it uses the previously generated model that’s cached and fails to generate the hyperparameters and plot (given that the step hasn’t run)

I can see there is a `dvc exp -f` parameter to force run the whole pipeline, however is there a way to force run from a specific stage, so that only the `train` step and any further downstream steps are reprocessed and not the entire pipeline.

---

<div class="post-metadata">

**Author:** ![andboy](https://avatars.discourse-cdn.com/v4/letter/a/a8b319/32.png) [@andboy](https://discuss.dvc.org/u/andboy)\
**Post date:** [July 22, 2024, 10:16am UTC](https://discuss.dvc.org/t/force-running-experiment-from-a-specified-stage/2142/2 "2024-07-22T10:16:17Z")

</div>

You can do that using `--downstream` flag, which will run only stages after (and including) the target: `dvc exp run dvc.yaml:evaluate_test_data -f --downstream`