Search results

[Bug] Question Regarding Memory Consumption

Status: Open.

#380 In sgl-project/SpecForge;

· ggg-s opened on Dec 22, 2025

Development Roadmap 2026 Q1

Status: Open.

#374 In sgl-project/SpecForge;

· FrankLeeeee opened on Dec 19, 2025

Supports multi-machine training

Status: Open.

#373 In sgl-project/SpecForge;

· dongyibo opened on Dec 17, 2025

[Bug] EAGLE3 Deepseek-r1-distill-llama8b traning Index errors

Status: Open.

#368 In sgl-project/SpecForge;

· Sylvan820 opened on Dec 16, 2025

During training, the accuracy is below 1 and exhibits significant fluctuations.

Status: Open.

#365 In sgl-project/SpecForge;

· LOVE-YOURSELF-1 opened on Dec 13, 2025

How to train Qwen3-480B EAGLE draft model on multiple nodes?

Status: Open.

#362 In sgl-project/SpecForge;

· xiaomin-D opened on Dec 10, 2025

[Question] How to train the Vicuna model series?

Status: Open.

#361 In sgl-project/SpecForge;

· Sylvan820 opened on Dec 10, 2025

[Bug] TypeError: 'SGLangEagle3TargetModel' object is not callable

Status: Open.

#360 In sgl-project/SpecForge;

· xs-alt opened on Dec 9, 2025

[Question]What factors contribute to Eagle3's reduced acceleration performance on VLM architectures compared to traditional LLMs?

Status: Open.

#352 In sgl-project/SpecForge;

· 330205812 opened on Dec 5, 2025

[Bug] args.build-dataset-num-proc doesn't work

Status: Open.

#349 In sgl-project/SpecForge;

· C3236455482 opened on Dec 3, 2025

[Bug] latest benchmark run llama3.1-8b mat1 RuntimeError: expected mat1 and mat2 to have the same dtype, but got: float != c10::Half

Status: Open.

#348 In sgl-project/SpecForge;

· Sylvan820 opened on Dec 3, 2025

vlm target model don't have set_aux_hidden_states_layers

Status: Open.

#345 In sgl-project/SpecForge;

· C3236455482 opened on Dec 2, 2025