All issues

Search results

How to apply REAP to FP8 / quantized MoE checkpoints?

Status: Open.

#24 In CerebrasResearch/reap;

· llwx593 opened on Jul 3, 2026

support deepseek-v4 ?

Status: Open.

#23 In CerebrasResearch/reap;

· Simba2017 opened on Jul 1, 2026

Is it possible to prune Kimi2.6?

Status: Open.

#21 In CerebrasResearch/reap;

· igorbcouto opened on Apr 21, 2026

Request for GLM 5.1 Pruned Model

Status: Open.

#20 In CerebrasResearch/reap;

· DonquixoteMzt opened on Apr 17, 2026

great work, thanks for opensourcing! Quick question:

Status: Open.

#18 In CerebrasResearch/reap;

· francoisp opened on Mar 19, 2026

Publish per-expert saliency scores alongside pruned models (or just the scores)

Status: Open.

#15 In CerebrasResearch/reap;

· flash7777 opened on Mar 17, 2026

Save expert mapping (original → pruned indices) in pruned model config

Status: Open.

#14 In CerebrasResearch/reap;

· flash7777 opened on Mar 16, 2026

Bug: Shape mismatch and invalid tensor operations in fused experts path of MoETransformerObserver

Status: Open.

#11 In CerebrasResearch/reap;

· naoufelito opened on Feb 12, 2026

REAM Support

Status: Open.

#10 In CerebrasResearch/reap;

· TomLucidor opened on Jan 29, 2026

Support for Kimi-K2 / DeepSeekV3 / Kimi-Linear models in REAP

Status: Open.

#9 In CerebrasResearch/reap;

· kebijuelun opened on Jan 16, 2026

Error while pruning ERNIE-4.5-21B-A3B-PT

Status: Open.

#8 In CerebrasResearch/reap;

· tdtgi opened on Dec 10, 2025

SUPPORT: Pruning support for gpt-oss-20b model

Status: Open.

#7 In CerebrasResearch/reap;

· sandeshrajbhandari opened on Dec 8, 2025