# \#dice-loss

**URL:** https://discourse.julialang.org/tag/dice-loss/1538.md

[Latest](https://discourse.julialang.org/latest.md) · [Categories](https://discourse.julialang.org/categories.md) · [Tags](https://discourse.julialang.org/tags.md)

---

## [Hard error using dice loss](https://discourse.julialang.org/t/hard-error-using-dice-loss/110059)

<div class="topic-metadata">

**Author:** [@cirobr](https://discourse.julialang.org/u/cirobr)\
**Replies:** 2\
**Last updated:** [February 20, 2024, 7:44pm UTC](https://discourse.julialang.org/t/hard-error-using-dice-loss/110059 "2024-02-20T19:44:05Z")

</div>

Cheers, Regardless of the model, data, or any other condition, I’ve never been able of using the built-in Flux.dice\_coeff\_loss() function. A very long error dump shows up, apparently tied to CUDA and memory usage For t…