Search results

OmniVoice 量化问题

Status: Open.

#297 In Tencent/AngelSlim;

· birtal opened on May 8, 2026

请问下是否有支持qwen3-omni eagle3投机模型训练的计划？

Status: Open.

#295 In Tencent/AngelSlim;

· wangxcqupt opened on May 7, 2026

Hy-MT-demo.apk闪退

Status: Open.

#294 In Tencent/AngelSlim;

· miaofree opened on May 4, 2026

qwen3-vl-32b eagle3投机训练OOM

Status: Open.

#288 In Tencent/AngelSlim;

· llllchao opened on Apr 21, 2026

What was the training dataset/data mix used to train AngelSlim/Qwen3-8B_eagle3 and other qwen eagle3 drafters?

Status: Open.

#287 In Tencent/AngelSlim;

· nelaturuharsha opened on Apr 19, 2026

为什么qwen3 1.7b eagle3 中文的猜中率很低，甚至猜不中。在线求?

Status: Open.

#283 In Tencent/AngelSlim;

· xieguangnian opened on Apr 10, 2026

在线训练报错，求解决方法

Status: Open.

#278 In Tencent/AngelSlim;

· wlwenming opened on Mar 30, 2026

Questions about QAT Data Composition

Status: Open.

#277 In Tencent/AngelSlim;

· xyuany opened on Mar 28, 2026

Ask for support of Qwen-3-VL-embedding

Status: Open.

#248 In Tencent/AngelSlim;

· LecterF opened on Mar 3, 2026

MLX Support for Eagle3

Status: Open.

#242 In Tencent/AngelSlim;

· TomLucidor opened on Feb 20, 2026

Question about ultrafineweb dataset preparation for tequila/sherry

Status: Open.

#233 In Tencent/AngelSlim;

· koshostanford opened on Jan 28, 2026

[Eagle-3 Training] Cloud I use this code for Qwen3-235B draft model training. I cant find any config about target models parallel setting like tp\pp\ep ...

Status: Open.

#230 In Tencent/AngelSlim;

· TheUgliestOfWuli opened on Jan 26, 2026