Search results

Gibberish output during training/inference with Qwen2.5-Omni-7B

Status: Open.

#13 In HarryHsing/EchoInk;

· WendySong31 opened on Mar 13, 2026

support qwen3-omni?

Status: Open.

#12 In HarryHsing/EchoInk;

· JoshonSmith opened on Jan 6, 2026

改成lora微调，GPU利用率和功耗只有三分之一

Status: Open.

#11 In HarryHsing/EchoInk;

· SeinCore opened on Sep 20, 2025

为什么需要修改 config.json "hidden_size": 3584 at the root level ?

Status: Open.

#10 In HarryHsing/EchoInk;

· King-king424 opened on Sep 11, 2025

About dataset

Status: Open.

#9 In HarryHsing/EchoInk;

· kaiw7 opened on Sep 5, 2025

Issue during Environment Setup: Protobuf Dependency Conflict

Status: Open.

#8 In HarryHsing/EchoInk;

· Dawn-970 opened on Sep 4, 2025

Out of memory

Status: Open.

#7 In HarryHsing/EchoInk;

· kaiw7 opened on Aug 30, 2025

details on data curation

Status: Open.

#6 In HarryHsing/EchoInk;

· jayaylee opened on Aug 21, 2025

在GRPO训练中后期，模型有时会生成无关内容至最大token数

Status: Open.

#5 In HarryHsing/EchoInk;

· ZhuoranZhao opened on Aug 15, 2025