Search results

train_multimodal_from_scratch的sft脚本现在运行会报错

Status: Open.

#40 In wyf3/llm_related;

· surpoloyang opened on Mar 29, 2026

code-r1 生成的代码乱七八糟

Status: Open.

#39 In wyf3/llm_related;

· twurbo opened on Mar 13, 2026

有一个小小的提议，感谢发言

Status: Open.

#34 In wyf3/llm_related;

· MingDemon411 opened on Dec 1, 2025

想问下code执行服务url如何开启？

Status: Open.

#31 In wyf3/llm_related;

· lazyloafer opened on Oct 14, 2025

train_multimodall_from_scratch中自己遇到的一个小问题

Status: Open.

#30 In wyf3/llm_related;

· QianQing26 opened on Sep 30, 2025

ValueError: No available memory for the cache blocks. Try increasing gpu_memory_utilization when initializing the engine.

Status: Open.

#29 In wyf3/llm_related;

· SimonHeye opened on Sep 12, 2025

推理时警告:Some weights of Qwen2ForCausalLM were not initialized from the model checkpoint 的解决办法

Status: Open.

#28 In wyf3/llm_related;

· xiaoBIGfeng opened on Aug 22, 2025

DPO对应代码

Status: Open.

#27 In wyf3/llm_related;

· yhl41001 opened on Aug 18, 2025

deep_research项目代码疑问

Status: Open.

#26 In wyf3/llm_related;

· luxuantao opened on Jun 20, 2025

请问grpo复现代码如何支持更大的模型进行训练，用多张显卡进行加速如何设置

Status: Open.

#23 In wyf3/llm_related;

· hexixiang opened on May 20, 2025

MLA的KV_cache问题

Status: Open.

#21 In wyf3/llm_related;

· Messiahccgr opened on Apr 27, 2025

deepseek MTP实现

Status: Open.

#20 In wyf3/llm_related;

· ScareCrowcrow opened on Apr 22, 2025