Search results

Missing Compute Details for MoE Scaling

Status: Open.

#75 In pjlab-sys4nlp/llama-moe;

· apoorvcodes opened on May 26, 2025

parameter of weight_gates are not initialized from huggingface checkpoint

Status: Open.

#74 In pjlab-sys4nlp/llama-moe;

· li-yang23 opened on Feb 15, 2025

Some questions on scripts and runtime

Status: Open.

#68 In pjlab-sys4nlp/llama-moe;

· kevin3567 opened on Aug 29, 2024

Can you report the running time on hardware?

Status: Open.

#57 In pjlab-sys4nlp/llama-moe;

· qiuzh20 opened on Jan 23, 2024

How many llama models are used for constructing llama-moe ? moe的构建是通过多个llama模型还是1个llama模型

Status: Open.

#55 In pjlab-sys4nlp/llama-moe;

· ZeyuTeng96 opened on Jan 17, 2024

#Feature Request# Accelerated Deployment.

Status: Open.

#51 In pjlab-sys4nlp/llama-moe;

· Xingxiangrui opened on Jan 3, 2024