Search results

[BUG] ZeRO-3 partition does not work in Ulysses SP tutorial

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7458 In deepspeedai/DeepSpeed;

· cnlnpjhsy opened on Jul 30, 2025

[REQUEST] Add support for EXAONE 4.0 models

enhancementNew feature or requestNew feature or request

Status: Open.

#7453 In deepspeedai/DeepSpeed;

· lesj0610 opened on Jul 26, 2025

[REQUEST] Defer detection of op builder compatibility until build time

enhancementNew feature or requestNew feature or request

Status: Open.

#7452 In deepspeedai/DeepSpeed;

· dwyatte opened on Jul 25, 2025

Fille not found error with pip install

Status: Open.

#7451 In deepspeedai/DeepSpeed;

· scibott opened on Jul 25, 2025

[BUG]Resolving OOM Issues in ConcurrenDistributed Inference of 111B Teacher Model and Distributed Training of 8B Student Model on Multi-Node H200 GPUs

Status: Open.

#7450 In deepspeedai/DeepSpeed;

· seona21 opened on Jul 24, 2025

[REQUEST] Add support for non-__dict__ outputs such as MinkowskiEngine SparseTensor in ZeRO Stage 3 (DeepSpeed v0.9.2)

enhancementNew feature or requestNew feature or request

Status: Open.

#7442 In deepspeedai/DeepSpeed;

· RyanPham19092002 opened on Jul 22, 2025

Some question of gradient accumulation

Status: Open.

#7439 In deepspeedai/DeepSpeed;

· sun666666ai opened on Jul 20, 2025

[REQUEST] Support Muon Optimizer

enhancementNew feature or requestNew feature or request

Status: Open.

#7438 In deepspeedai/DeepSpeed;

· hijkzzz opened on Jul 19, 2025

Deepspeed Zero3 resume training with Huggingface Trainer Failed

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7436 In deepspeedai/DeepSpeed;

· JoeZhao527 opened on Jul 18, 2025

[BUG] Memory is enough for training by using zero-3, but OOM occurred after enabling DeepCompile

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7434 In deepspeedai/DeepSpeed;

· rookie-ryan opened on Jul 18, 2025

[BUG] TypeError: 'staticmethod' object is not callable, in deepcompile (patch_compiled_func.py)

bugSomething isn't workingSomething isn't working

training

Status: Open.

#7433 In deepspeedai/DeepSpeed;

· rookie-ryan opened on Jul 18, 2025

How can I merge the original model weights with LoRA weights?

Status: Open.

#7429 In deepspeedai/DeepSpeed;

· guoguo1314 opened on Jul 15, 2025