Search results

why GeminiPlugin zero3+offloading cannot training a 7B model

Status: Open.

#6272 In hpcaitech/ColossalAI;

· SeekPoint opened on Apr 21, 2025

[BUG]: Hybrid Parallel Plugin，zero_stage=1，zero_cpu_offload=true，terminate called after throwing an instance of 'c10::Error' what() Cuda error: unspecified launch failure cuda kernel errors might be asynchronously reported at some other API call

bugSomething isn't workingSomething isn't working

Status: Open.

#6266 In hpcaitech/ColossalAI;

· happynaruto opened on Apr 16, 2025

[BUG]: Gemini Plugin，tp=2，TypeError: qwen2model.forward() got an unexpected keyword argument 'force_sp_output_gather'

bugSomething isn't workingSomething isn't working

Status: Open.

#6265 In hpcaitech/ColossalAI;

· happynaruto opened on Apr 16, 2025

[BUG]: Hybrid Parallel Plugin PP=1 TypeError: qwen2model.forward() got an unexpected keyword argument 'force_sp_output_gather'

bugSomething isn't workingSomething isn't working

Status: Open.

#6264 In hpcaitech/ColossalAI;

· happynaruto opened on Apr 16, 2025

[BUG]: Unable to figure out how to pass env variables for each node

bugSomething isn't workingSomething isn't working

Status: Open.

#6261 In hpcaitech/ColossalAI;

· Gautam-Rajeev opened on Apr 14, 2025

[BUG]: RuntimeError: Failed to replace input_layernorm of type DeepseekV3RMSNorm with FusedRMSNorm with the exception: 'NoneType' object is not callable.

bugSomething isn't workingSomething isn't working

Status: Open.

#6260 In hpcaitech/ColossalAI;

· lgybuaa opened on Apr 13, 2025

[BUG]: can not save model in pipeline training mode

bugSomething isn't workingSomething isn't working

Status: Open.

#6253 In hpcaitech/ColossalAI;

· fangxintao opened on Mar 25, 2025

Does ColossalAI support delayed scaling FP8?

Status: Open.

#6249 In hpcaitech/ColossalAI;

· zigzagcai opened on Mar 20, 2025

[BUG]: Failed to run lora_finetune.py to fine-tune DS

bugSomething isn't workingSomething isn't working

Status: Open.

#6248 In hpcaitech/ColossalAI;

· mahaocong90 opened on Mar 19, 2025

[BUG]: 用64卡或者80卡训练 loss 总是nan

bugSomething isn't workingSomething isn't working

Status: Open.

#6239 In hpcaitech/ColossalAI;

· xujin1184104394 opened on Mar 6, 2025

[DOC]: 可以提供使用ColossalAI 训练自定义的模型

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open.

#6238 In hpcaitech/ColossalAI;

· lianNice opened on Mar 6, 2025

[BUG]: raise ValueError("No parameters found in moe_dp_group, please consider using HybridParallelPlugin instead") ValueError: No parameters found in moe_dp_group, please consider using HybridParallelPlugin instead

bugSomething isn't workingSomething isn't working

Status: Open.

#6237 In hpcaitech/ColossalAI;

· klompn opened on Mar 4, 2025