Skip to content

issues Search Results · repo:deepspeedai/DeepSpeedExamples language:Python

Filter by

548 results
 (64 ms)

548 results

indeepspeedai/DeepSpeedExamples (press backspace or delete to remove)

deepspeedai/DeepSpeedExamples
DeepSpeed-FastGen support ascend npu?

DeepSpeed-FastGen support ascend npu, deepseek-r1-distilled-qwen2.5-32b?
  • RyanOvO
  • 3
  • Opened 
    on Mar 4
  • #960

deepspeedai/DeepSpeedExamples
Why Does vf_loss Take the Maximum Value, Rendering Clamp Meaningless?

critic_loss: def critic_loss_fn(self, values, old_values, returns, mask): value loss values_clipped = torch.clamp( values, old_values - self.cliprange_value, old_values + self.cliprange_value, ) vf_loss1 ...
  • Morizhaoyang
  • 1
  • Opened 
    on Feb 7
  • #956

deepspeedai/DeepSpeedExamples
Is there any example about DeepSpeed Zero with Ulysses/Ulysses-offload

I only found DeepSpeed Megatron with Ulysses/Ulysses-offload
  • LSC527
  • Opened 
    on Jan 10
  • #951

deepspeedai/DeepSpeedExamples
Domino + PP

I’m excited about the recent introduction of Domino and its impressive TP optimization. When I was using deepspeed-domino to better overlap comm comp in TP, I found domino use forward_backward_no_pipelining() ...
  • XZQshiyu
  • Opened 
    on Jan 9
  • #950

deepspeedai/DeepSpeedExamples
Error when running training example DeepSpeed-Domino/pretrain_gpt3_2.7b.sh

Hello! I encountered some errors when running https://github.com/microsoft/DeepSpeedExamples/blob/master/training/DeepSpeed-Domino/pretrain_gpt3_2.7b.sh and here is the error information: [rank1]:[W102 ...
  • ZhiyiHu1999
  • 1
  • Opened 
    on Jan 2
  • #948

deepspeedai/DeepSpeedExamples
Assertion `srcIndex < srcSelectDimSize` failed

When I try to run Stage 3 finetuning PPO for qwen 2 0.5B model, I got the following bug: Assertion srcIndex srcSelectDimSize failed, which seems like issue about input dataset sequence length? I have ...
  • boqiny
  • 1
  • Opened 
    on Dec 18, 2024
  • #946

deepspeedai/DeepSpeedExamples
Question to attention computation

Hi, thank you for the amazing demo and doc! I have a question regarding this section in zero-inference. It is mentioned that Thus, our current implementation computes attention scores on CPU. May I ask ...
  • yuzhenmao
  • Opened 
    on Dec 15, 2024
  • #944

deepspeedai/DeepSpeedExamples
KV_cache offload

Hi, I am using the latest huggingface transformers (version==4.48.0.dev0). When I tried to run the demo from here, I have this error: AttributeError: LlamaForCausalLM object has no attribute set_kv_cache_offload ...
  • yuzhenmao
  • 2
  • Opened 
    on Dec 15, 2024
  • #943

deepspeedai/DeepSpeedExamples
A bug in argument parser.

Issue: In the original code: e2e_rlhf.py line 68 parser.add_argument( --reward-model , type=lambda x: x.replace( facebook/opt- , ), default= 350m , choices=( 350m ), help= Which facebook/opt-* ...
  • ChenDaiwei-99
  • Opened 
    on Dec 10, 2024
  • #941

deepspeedai/DeepSpeedExamples
Failed to run Domino example

Following the guide of DeepSpeed-Domino, I ran into the following issue when bash pretrain_gpt3_2.7b.sh [rank0]: return cdb.all_reduce(tensor, op, group, async_op) [rank0]: ^^^^^^^^^^^^^^ ...
  • lucifer1004
  • 2
  • Opened 
    on Nov 29, 2024
  • #940
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue search results · GitHub