All issues

Search results

Upstream a vLLM pre-resolution hook and remove the Hugging Face snapshot patch

enhancementNew feature or requestNew feature or request

Status: Open.

#640 In ai-dynamo/modelexpress;

· zhengluo-nv opened on Aug 14, 2026

Support Megatron QKV resharding when KV heads are fewer than trainer TP ranks

Status: Open.

#639 In ai-dynamo/modelexpress;

· zhengluo-nv opened on Aug 14, 2026

Runai streamer timeout causes double CUDA memory allocation

Status: Open.

#638 In ai-dynamo/modelexpress;

· gshtrasberg opened on Aug 14, 2026

Cross RDMA domain source selection causes GPU OOM during fallback

Status: Open.

#623 In ai-dynamo/modelexpress;

· Broduker opened on Aug 13, 2026

Declarative coordinator deployment via operator

Status: Open.

#606 In ai-dynamo/modelexpress;

· wseaton opened on Aug 11, 2026

[Vision] Toward ModelExpress 1.0

Status: Open.

#594 In ai-dynamo/modelexpress;

· athreesh opened on Aug 6, 2026

feat: integrate server-side model-cache fallback with P2P loaders

enhancementNew feature or requestNew feature or request

Status: Open.

#569 In ai-dynamo/modelexpress;

· zhengluo-nv opened on Jul 28, 2026

Torch compile cache artifacts transfer between workers with incompatible compile configurations

Status: Open.

#560 In ai-dynamo/modelexpress;

· kangclzjc opened on Jul 28, 2026

Pulling both model weights and MTP from s3:// leads to a crash

Status: Open.

#537 In ai-dynamo/modelexpress;

· gshtrasberg opened on Jul 23, 2026

auth: bound TokenReview traffic from unauthenticated callers

Status: Open.

#530 In ai-dynamo/modelexpress;

· wseaton opened on Jul 22, 2026

[RFC] Delta weight sync requirements

Status: Open.

#524 In ai-dynamo/modelexpress;

· athreesh opened on Jul 21, 2026

Clarify usage of model-streamer with s3 compatible object-storage

Status: Open.

#500 In ai-dynamo/modelexpress;

· strigazi opened on Jul 9, 2026