Issues

RecallMetric / HindsightTargetPR: compute_false_neg_sum uses <= and double-counts positives predicted exactly at the threshold
meta-pytorch/torchrec#4812 · opened on Sep 30, 2026
test_pt2.py: permute_pooled_embs tests use an illegal input shape (12, 3) vs embs_dims sum = 36
meta-pytorch/torchrec#4718 · 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
meta-pytorch/torchrec#4712 · opened on Sep 11, 2026
cudaMemcpyAsync using pageable memory blocking main stream concurrency
meta-pytorch/torchrec#3855 · opened on Mar 10, 2026
1
how to use torch.distributed.checkpoint to save and load state dict
meta-pytorch/torchrec#3567 · opened on Nov 20, 2025
How can I export a trained model to the Triton inference server?
meta-pytorch/torchrec#3561 · opened on Nov 19, 2025
How to convert DistributedModelParallel to quantize_inference_model and use torch.jit.script to save?
meta-pytorch/torchrec#3559 · opened on Nov 19, 2025
test_installation.py does not work
meta-pytorch/torchrec#3539 · opened on Nov 12, 2025
Clarify WARNING:torchrec.distributed.utils:Sharding Type is data_parallel, caching params will be ignored
meta-pytorch/torchrec#3505 · opened on Nov 3, 2025
System OOM Error in Distributed Checkpoint Writer When using fused_uvm_caching Kernel
meta-pytorch/torchrec#3399 · opened on Sep 29, 2025
confused with torchrec/distributed/sharding/dynamic_sharding.py
meta-pytorch/torchrec#3395 · opened on Sep 26, 2025
invoke_on_rank_and_broadcast_result stuck
meta-pytorch/torchrec#3278 · opened on Aug 13, 2025
DistributedModelParallel sets "requires_grad" to False for EmbeddingCollection with sharding_types=["row_wise"], compute_kernels=["fused"],
meta-pytorch/torchrec#3271 · opened on Aug 11, 2025
[Bug]: ManagedCollisionEmbeddingCollection returns all-zero embeddings after applying apply_optimizer_in_backward with RowWiseAdagrad
meta-pytorch/torchrec#3030 · opened on Jun 3, 2025
ManagedCollisionEmbeddingBagCollectionSharder cannot support CPU
meta-pytorch/torchrec#2958 · opened on May 8, 2025
2
CUDA error: an Illegal memory access was encountered
meta-pytorch/torchrec#2957 · opened on May 7, 2025
1
Can NVIDIA-Merlin/HierarchicalKV integrated to torchrec as EmbeddingBag Extension?
meta-pytorch/torchrec#2923 · opened on Apr 29, 2025
2
how to freeze embedding
meta-pytorch/torchrec#2908 · opened on Apr 23, 2025
1
embeddingbagcollection 16p vs 8p
meta-pytorch/torchrec#2904 · opened on Apr 22, 2025
TypeError: unsupported operand type(s) for /: 'Tensor' and 'NoneType' when use mc-ebc and mean pooling
meta-pytorch/torchrec#2828 · opened on Mar 17, 2025
1
The static_dict_gather function encounters precision issue when use mc-ebc
meta-pytorch/torchrec#2827 · opened on Mar 17, 2025
1
roadmap 2025 H1 H2
meta-pytorch/torchrec#2825 · opened on Mar 17, 2025
3
[Question] Is there any way to apply AdamW as optimizer for sparse features
meta-pytorch/torchrec#2821 · opened on Mar 14, 2025
row-wise alltoall error when some embeddings use mean pooling and others use sum pooling
meta-pytorch/torchrec#2808 · opened on Mar 12, 2025
2
Floating point exception (core dumped) when use quantize embeddings with float32 dtype
meta-pytorch/torchrec#2807 · opened on Mar 12, 2025
3