Search results

[finetune_hf_llama example] hf2megads_weight_converter.py fails to save due to Attrubute error

Status: Open.

#482 In deepspeedai/Megatron-DeepSpeed;

· aravneelaws opened on Sep 24, 2025

[QUESTION]How to train llama 3 with Megatron-DeepSpeed

Status: Open.

#481 In deepspeedai/Megatron-DeepSpeed;

· lzyonetwothree opened on Sep 24, 2025

[QUESTION] lr schedule(OptimizerParamScheduler) is not aligned when using token and sample as argument

Status: Open.

#475 In deepspeedai/Megatron-DeepSpeed;

· xrrain opened on May 6, 2025

can we [add new feature ]support zero=2,3 with --tensor-model-parallel-size 2 --pipeline-model-parallel-size 2 for pretrain-gpt2?

Status: Open.

#474 In deepspeedai/Megatron-DeepSpeed;

· SeekPoint opened on Apr 28, 2025

[BUG] : training gpt2 with PP=2 then error: list index out of range

Status: Open.

#473 In deepspeedai/Megatron-DeepSpeed;

· 9LLPPLL6 opened on Apr 18, 2025

Tutorials does not reflect contents in repository

Status: Open.

#471 In deepspeedai/Megatron-DeepSpeed;

· auggie246 opened on Mar 26, 2025

support qwen2.5-32b?

Status: Open.

#466 In deepspeedai/Megatron-DeepSpeed;

· RyanOvO opened on Feb 24, 2025

Does it support deployment and invocation under the Ascend system & CANN?

Status: Open.

#465 In deepspeedai/Megatron-DeepSpeed;

· RyanOvO opened on Feb 21, 2025

Converting hf weights to mds: unable to find latest file

Status: Open.

#464 In deepspeedai/Megatron-DeepSpeed;

· anasegor opened on Feb 17, 2025

Finetune issue with Llama2-7b

Status: Open.

#463 In deepspeedai/Megatron-DeepSpeed;

· anasegor opened on Feb 14, 2025

Any one used “—exit-signal-handler” ?

Status: Open.

#460 In deepspeedai/Megatron-DeepSpeed;

· HenryTangDev opened on Feb 5, 2025

Model Inference issues with GPT-MoE models

Status: Open.

#458 In deepspeedai/Megatron-DeepSpeed;

· 1155157110 opened on Jan 25, 2025