-
Notifications
You must be signed in to change notification settings - Fork 4.2k
Issues
is:issue state:open
is:issue state:open
Issue creation is restricted in this repository
Search results
- Status: Open.#5776 In NVIDIA/Megatron-LM;
[docs] Add a "configuration errors" troubleshooting page
good first issueGood for newcomersGood for newcomersStatus: Open.#5766 In NVIDIA/Megatron-LM;[Bug] validate_args() eagerly probes the CUDA device and crashes GPU-less launcher/driver processes (Found no NVIDIA driver)
bugSomething isn't workingSomething isn't workingStatus: Open.#5763 In NVIDIA/Megatron-LM;[BUG] combined_1f1b:
with context_manager and outer_fp8_context:never enters the autocast contextbugSomething isn't workingSomething isn't workingStatus: Open.#5760 In NVIDIA/Megatron-LM;[BUG] Public API dist_checkpointing.remove_sharded_tensors is broken: any call raises AttributeError
bugSomething isn't workingSomething isn't workingStatus: Open.#5758 In NVIDIA/Megatron-LM;[BUG] TransformerConfig accepts num_attention_heads not divisible by num_query_groups, then crashes with a cryptic shape error inside attention
bugSomething isn't workingSomething isn't workingStatus: Open.#5755 In NVIDIA/Megatron-LM;[BUG] Expert Parallelism deadlock with Gemma4 (EP=4, DP≥16)
bugSomething isn't workingSomething isn't workingwaiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respondStatus: Open.#5749 In NVIDIA/Megatron-LM;🐛 CI failure: determinism_perf exceeds 1.25x slowdown threshold
bugSomething isn't workingSomething isn't workingStatus: Open.#5747 In NVIDIA/Megatron-LM;Migrate all TEFusedOps to the new mixin in te extensions
enhancementNew feature or requestNew feature or requestStatus: Open.#5741 In NVIDIA/Megatron-LM;🐛 CI failure: NCCL EP tests abort with pybind11 GIL dec_ref failure
bugSomething isn't workingSomething isn't workingStatus: Open.#5736 In NVIDIA/Megatron-LM;Design Discussion For Parameter-Level Initialization/Optimization Configurability
enhancementNew feature or requestNew feature or requestwaiting-on-maintainersWaiting on maintainers to respondWaiting on maintainers to respondStatus: Open.#5732 In NVIDIA/Megatron-LM;Full-iteration CUDA graph support for MFSDP v2
enhancementNew feature or requestNew feature or requestStatus: Open.#5717 In NVIDIA/Megatron-LM;