# Model Studio - Out of Memory during training & branches support

**URL:** <https://community.palantir.com/t/model-studio-out-of-memory-during-training-branches-support/7297>\
**Category:** Product Feedback\
**Tags:** modeling\
**Created:** [September 30, 2026, 12:37am UTC](https://community.palantir.com/t/model-studio-out-of-memory-during-training-branches-support/7297 "2026-09-30T00:37:30Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![9eab477d6eb76c06c4f6](https://yyz1.discourse-cdn.com/palantir/user_avatar/community.palantir.com/9eab477d6eb76c06c4f6/32/7919_2.png) [@9eab477d6eb76c06c4f6](https://community.palantir.com/u/9eab477d6eb76c06c4f6)\
**Post date:** [September 30, 2026, 12:37am UTC](https://community.palantir.com/t/model-studio-out-of-memory-during-training-branches-support/7297/1 "2026-09-30T00:37:30Z")

</div>

Hello,

I have prepared a dataset for use with Model Studio for a time-series prediction.

My input data is around 3GB - 19M rows however I keep getting OoM issues even when memory is bumped to 64GB + A10G GPU.

While I understand that I could reduce the size of my input data, there is no data loader implemented ?

As well, it would be great to support dataset branch as input.

Best regards,