Search results

[BUG] The same program runs fine with v0.17.5, but fails with v0.17.6. Under the zero2 configuration

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7629 In deepspeedai/DeepSpeed;

· hemengfei2014-stack opened on Oct 13, 2025

[Question] How to Resume DeepSpeed ZeRO-2 Training with a Different Number of GPUs?

Status: Open.

#7628 In deepspeedai/DeepSpeed;

· xiao10ma opened on Oct 12, 2025

[BUG] Improper coupling of paramter list between DeepSpeedZeroOptimizer_Stage3 and SuperOffloadOptimizer_Stage3

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7627 In deepspeedai/DeepSpeed;

· delock opened on Oct 10, 2025

[REQUEST] Can deepspeed initialize multiple datasets/dataloaders?

enhancementNew feature or requestNew feature or request

Status: Open.

#7626 In deepspeedai/DeepSpeed;

· manyuanbin opened on Oct 10, 2025

[REQUEST] Allow fallback to torch.distributed.all_reduce for multi-node inference.

enhancementNew feature or requestNew feature or request

Status: Open.

#7625 In deepspeedai/DeepSpeed;

· phalani-paladugu opened on Oct 8, 2025

[BUG]Installation Error on Torch while Torch is Installed

Status: Open.

#7616 In deepspeedai/DeepSpeed;

· evencogitate opened on Oct 2, 2025

[BUG] AttributeError: 'Linear' object has no attribute 'ds_grads_remaining'

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7615 In deepspeedai/DeepSpeed;

· jianrui2025 opened on Oct 2, 2025

[BUG] DeepCompile: Training hang on random-sized inputs

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7611 In deepspeedai/DeepSpeed;

· eternalNight opened on Sep 30, 2025

[REQUEST] How to achieve DeepNVMe (ZeRO 3 offloading) via GPU Direct Storage?

enhancementNew feature or requestNew feature or request

Status: Open.

#7608 In deepspeedai/DeepSpeed;

· tanxiaojie-sys opened on Sep 30, 2025

[BUG] ZeRO 2 ipg_buckets key error

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7607 In deepspeedai/DeepSpeed;

· stuart-mv opened on Sep 29, 2025

[BUG] The loss of the first step is too high when zero_hpz_partition_size is set.

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7606 In deepspeedai/DeepSpeed;

· zhengchenyu opened on Sep 29, 2025

[BUG]DeepSpeed ZeRO-2 Offload: backward does not populate IPG buckets even though autograd graph reaches parameters (PyTorch 2.0 / Python 3.8 / DeepSpeed 0.17.6)

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7595 In deepspeedai/DeepSpeed;

· XiDianZuoYun opened on Sep 25, 2025