Search results

Training GPT-OSS-20B caused OOM

bugSomething isn't workingSomething isn't working

Status: Open.

#14712 In NVIDIA-NeMo/NeMo;

· kltlj1 opened on Sep 10, 2025

parakeet-tdt_ctc loss scale

Status: Open.

#14710 In NVIDIA-NeMo/NeMo;

· ehgusehgus opened on Sep 10, 2025

different results with fine-tuned model

Status: Open.

#14700 In NVIDIA-NeMo/NeMo;

· ramil-gadirov opened on Sep 9, 2025

Is <code>MegatronOptimizerModule</code>'s <code>adam</code> optimizer actually <code>adamw</code>?

Status: Open.

#14695 In NVIDIA-NeMo/NeMo;

· pilot7747 opened on Sep 9, 2025

Error while using aggregate_moe_loss_stats

bugSomething isn't workingSomething isn't working

Status: Open.

#14692 In NVIDIA-NeMo/NeMo;

· prathamk-tw opened on Sep 9, 2025

GPT-OSS has the wrong value of yarn_original_max_position_embeddings

bugSomething isn't workingSomething isn't working

Status: Open.

#14686 In NVIDIA-NeMo/NeMo;

· kzhou92 opened on Sep 9, 2025

Cpu offloading Fails with double buffer enabled.

bugSomething isn't workingSomething isn't working

Status: Open.

#14670 In NVIDIA-NeMo/NeMo;

· yspMing opened on Sep 8, 2025

Poor Inference Quality After Converting HF Model to NeMo [LLama 3.1]

bugSomething isn't workingSomething isn't working

Status: Open.

#14668 In NVIDIA-NeMo/NeMo;

· Harsha0056 opened on Sep 8, 2025

Sequence Packing with ChatDataModule

Status: Open.

#14666 In NVIDIA-NeMo/NeMo;

· simoneangarano opened on Sep 8, 2025

Inference on a continued_DAPT checkpoint fails with CUDA_DSA_ERROR

Status: Open.

#14655 In NVIDIA-NeMo/NeMo;

· Shrii-WorkspaceNSX opened on Sep 5, 2025

Setting max_steps=-1 causes negative dataset length in NeMo fine_tuning.py

bugSomething isn't workingSomething isn't working

Status: Open.

#14646 In NVIDIA-NeMo/NeMo;

· KhalidAlt opened on Sep 4, 2025

In Container Image, how to use the latest github main branch?

Status: Open.

#14630 In NVIDIA-NeMo/NeMo;

· DigtalFinanceAndWorld opened on Sep 3, 2025