All issues

test_pt2.py: permute_pooled_embs tests use an illegal input shape (12, 3) vs embs_dims sum = 36

#4718

· smacaijicoder opened

on Sep 14, 2026

An 18 ms grad_indice_weights kernel in every backward of the weighted EmbeddingBagCollection when running the sparse data dist benchmark on 8x B200

#4712

· TarzanZhao opened

on Sep 11, 2026

cudaMemcpyAsync using pageable memory blocking main stream concurrency

#3855

· hyu2707 opened

on Mar 10, 2026

how to use torch.distributed.checkpoint to save and load state dict

#3567

· haolujun opened

on Nov 20, 2025

How can I export a trained model to the Triton inference server?

#3561

· intfish123 opened

on Nov 19, 2025

How to convert DistributedModelParallel to quantize_inference_model and use torch.jit.script to save?

#3559

· intfish123 opened

on Nov 19, 2025

test_installation.py does not work

#3539

· hetian127 opened

on Nov 12, 2025

Clarify `WARNING:torchrec.distributed.utils:Sharding Type is data_parallel, caching params will be ignored`

#3505

· gyulaz-htec opened

on Nov 3, 2025

System OOM Error in Distributed Checkpoint Writer When using fused_uvm_caching Kernel

#3399

· tiankongdeguiji opened

on Sep 29, 2025

confused with torchrec/distributed/sharding/dynamic_sharding.py

#3395

· haolujun opened

on Sep 26, 2025

invoke_on_rank_and_broadcast_result stuck

#3278

· 121547456 opened

on Aug 13, 2025

DistributedModelParallel sets "requires_grad" to False for EmbeddingCollection with sharding_types=["row_wise"], compute_kernels=["fused"],

#3271

· farapart opened

on Aug 11, 2025