Search results

Memory leak when log_max_attention_logits is set

bugSomething isn't workingSomething isn't working

community-request

Status: Open.

#4699 In NVIDIA/Megatron-LM;

· supershadoe opened on May 8, 2026

Muon CPU offload

community-request

enhancementNew feature or requestNew feature or request

Status: Open.

#4691 In NVIDIA/Megatron-LM;

· pengdurice opened on May 8, 2026

Torch FSDP2 crashes in finalize_model_grads because _BaseDataParallel.finish_grad_sync() does not accept force_all_reduce

bugSomething isn't workingSomething isn't working

community-request

Status: Open.

#4669 In NVIDIA/Megatron-LM;

· Jayoprell opened on May 7, 2026

[QUESTION]Has MCore v0.17 released?

community-request

Status: Open.

#4665 In NVIDIA/Megatron-LM;

· suenphey opened on May 7, 2026

mcore scalar all_reduce fails NCCL FastTrak buffer registration

bugSomething isn't workingSomething isn't working

community-request

Status: Open.

#4660 In NVIDIA/Megatron-LM;

· cheonghm opened on May 6, 2026

🐛 CI hang/timeout: hybrid_dynamic_inference_tp1_ep8_nanov3_chunked_prefill exceeds 1h job limit on repeat iteration 3

bugSomething isn't workingSomething isn't working

Status: Open.

#4657 In NVIDIA/Megatron-LM;

· balasaajay opened on May 6, 2026

🐛 CI failure: tests/functional_tests/python_test_utils/test_pretraining_regular_pipeline.py::test_regular_pipeline [mimo_vlm_pretrain_convergence_tp1_pp1_cp1_dp8_seq_packing]

bugSomething isn't workingSomething isn't working

Status: Open.

#4654 In NVIDIA/Megatron-LM;

· balasaajay opened on May 6, 2026

[QUESTION] Do we have any plan to introduce NCCL EP?

community-request

waiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respond

Status: Open.

#4644 In NVIDIA/Megatron-LM;

· gangxie112 opened on May 6, 2026

[BUG] Muon + PP + MTP: tied word_embeddings on MTP-only last PP stage is routed to Muon instead of Adam

bugSomething isn't workingSomething isn't working

community-request

waiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respond

Status: Open.

#4641 In NVIDIA/Megatron-LM;

· yezhengmao1 opened on May 6, 2026

Flaky: TestFusedApplyMLARope::test_forward_backward_for_q[thd] backward mismatch exceeds bf16 tolerances

Status: Open.

#4640 In NVIDIA/Megatron-LM;

· ko3n1g opened on May 6, 2026

[Design Doc] HybridModel - Next Gen Model Class

Status: Open.

Documentation

#4620 In NVIDIA/Megatron-LM;

· Phlip79 opened on May 5, 2026

Reduce downstream Megatron patching for RL use cases

enhancementNew feature or requestNew feature or request

module: rl

Status: Open.

#4590 In NVIDIA/Megatron-LM;

· sbhavani opened on May 2, 2026