# Models & Pipelines

**URL:** https://forum.modular.com/c/max/models-pipelines/14.md

[Latest](https://forum.modular.com/latest.md) · [Categories](https://forum.modular.com/categories.md) · [Tags](https://forum.modular.com/tags.md)

---

## [About the Models & Pipelines category](https://forum.modular.com/t/about-the-models-pipelines-category/760)

<div class="topic-metadata">

**Author:** [@Caroline](https://forum.modular.com/u/Caroline)\
**Replies:** 0

</div>

---

## [Constraint failed: copy\_from\_async only allows 4, 8, 16 bytes element](https://forum.modular.com/t/constraint-failed-copy-from-async-only-allows-4-8-16-bytes-element/2768)

<div class="topic-metadata">

**Author:** [@vishnoianil](https://forum.modular.com/u/vishnoianil)\
**Replies:** 0\
**Last updated:** [February 25, 2026, 6:58am UTC](https://forum.modular.com/t/constraint-failed-copy-from-async-only-allows-4-8-16-bytes-element/2768 "2026-02-25T06:58:43Z")

</div>

I was experimenting with tiled GEMM kernel written in Mojo. Here is code . I am running this kernel on nvidia A100. It works fine with float32 DType, but fails with bfloat16/float16, because it seems like max layout te…

---

## [Model Correctness Check](https://forum.modular.com/t/model-correctness-check/2573)

<div class="topic-metadata">

**Author:** [@as0909024](https://forum.modular.com/u/as0909024)\
**Replies:** 1\
**Last updated:** [January 14, 2026, 4:16pm UTC](https://forum.modular.com/t/model-correctness-check/2573 "2026-01-14T16:16:39Z")

</div>

I am currently exploring the Modular repository and have understood several aspects of the MAX pipeline so far. However, I am stuck at one specific point and would appreciate some clarification. Integration test models …

---

## [Support GPTQ quant in MoE layer](https://forum.modular.com/t/support-gptq-quant-in-moe-layer/2401)

<div class="topic-metadata">

**Author:** [@thomasip](https://forum.modular.com/u/thomasip)\
**Replies:** 0\
**Last updated:** [October 26, 2025, 6:42pm UTC](https://forum.modular.com/t/support-gptq-quant-in-moe-layer/2401 "2025-10-26T18:42:40Z")

</div>

At the moment the MoE layer in MAX does not support quantization. This is blocking implementation for popular models like Qwen3-30B-A3B with GPTQ (Text, Coder, Omni variants). It seems the only thing missing is a quant a…

---

## [New resources for building models in MAX](https://forum.modular.com/t/new-resources-for-building-models-in-max/1762)

<div class="topic-metadata">

**Author:** [@BradLarson](https://forum.modular.com/u/BradLarson)\
**Replies:** 0\
**Last updated:** [June 27, 2025, 3:04pm UTC](https://forum.modular.com/t/new-resources-for-building-models-in-max/1762 "2025-06-27T15:04:21Z")

</div>

We’ve recently rolled out a series of new resources for how to build and serve models in MAX, something that I know has been commonly requested. These include a set of tutorials that walk you through building basic graph…