-
Notifications
You must be signed in to change notification settings - Fork 67
Issues
is:issue state:open
is:issue state:open
Issue creation is restricted in this repository
Search results
[Feature] 该怎么在VLM训练里也用这个框架？有示例吗
enhancementNew feature or requestNew feature or requestStatus: Open.#426 In InternLM/InternEvo;[Bug] Some spelling errors in repo
bugSomething isn't workingSomething isn't workingStatus: Open.#417 In InternLM/InternEvo;[Bug] CPU Mem utilization grows with training, when Dataloader num_workers>0
bugSomething isn't workingSomething isn't workingStatus: Open.#415 In InternLM/InternEvo;[Bug] TypeError: Qwen2.__init__() got an unexpected keyword argument 'enable_qkv_fusion'
bugSomething isn't workingSomething isn't workingStatus: Open.#408 In InternLM/InternEvo;[Feature] Add Lingen - MATE architecture (SSM Mamba2 based) as a model option
enhancementNew feature or requestNew feature or requestStatus: Open.#402 In InternLM/InternEvo;[Bug] loss error: flat_fp32_avg_grads will not be scaled when hybridzerooptimizer use cpu_offload
bugSomething isn't workingSomething isn't workingStatus: Open.#398 In InternLM/InternEvo;[QA] 训练到一半Nan grad norm occurs, please check it
questionFurther information is requestedFurther information is requestedStatus: Open.#375 In InternLM/InternEvo;[QA] Does internEvo support loongtrain selective checkpoint++?
questionFurther information is requestedFurther information is requestedStatus: Open.#372 In InternLM/InternEvo;[QA] 如何进行单卡微调的，需要调整那些设置
questionFurther information is requestedFurther information is requestedStatus: Open.#354 In InternLM/InternEvo;[QA] loong train 支持packed_sample_into_one=false吗
questionFurther information is requestedFurther information is requestedStatus: Open.#346 In InternLM/InternEvo;[Feature] MoE模型里稠密层和专家层zero和并行的解耦
enhancementNew feature or requestNew feature or requestStatus: Open.#325 In InternLM/InternEvo;[Feature] 不使用memory pool
enhancementNew feature or requestNew feature or requestStatus: Open.#324 In InternLM/InternEvo;