Repository Issue Activity (beta)

#2075 AttributeError: Compiling Error! Cannot locate the function: xtuner.v1.float8.triton_kernels.per_block_quant_gemm.per_block_quant_torch

#2025 [Bug] Routed-expert traces cause host OOM with offload enabled and CUDA OOM before SP split when disabled

#1941 优化 LMDeploy routed_experts 传输路径，避免 rollout worker 中不必要的数据中转

#1936 请问有自定义模型和损失的教程吗

#1902 [Nit] Remove the xtuner/check_gated_delta_rule_eager.py scripts

#1795 Example script path in .dev_scripts/analyze_train_logs.py is wrong

#1780 irrelevant issue can be closed

#1667 [Question] LoRA & QLoRA support in XTuner V1.0 (related to #1381)

#1690 [Security] Request for a security reporting policy

#1735 Implement timeout validation within unit test cases

#1707 Turn torch_compile_cfg=True when upgrade to pytorch 2.10.0 in autotest

#1665 [Feature Requst] Update more training info into exp_tracking

#1553 📋 Documentation Enhancement Suggestion

#991 A Question Regarding Resuming InternVL Training

#1623 [Docs] add comment on special env `DEEPEP_MAX_TOKENS_PER_RANK`

#484 [Feature] 提前确认离线处理的 llava 数据所使用的 tokenizer 是否和当前配置中一致

#301 Any plans to support Gradio for local web demos?

#1574 [Fix] Fix the claude workflow

#1219 使用deepspeed训练时会自动创建DeepSpeedOptimWrapper

#1562 CI test

#1560 CI image using pytorch 2.9

#1539 [test] test workflow

#1335 [RFC] Do we really need all_reduce in `BaseLossContext`?

#1510 [Bug] Only for issue workflow test

#1508 only for test