NVIDIA-NeMo/Speech

nemo-speech-pr-review

Repository review rubric and review principles for the formal /review command.

原文语言：英语

verify

Verify a NeMo Speech change or pull request before commit, push, or review by checking the complete diff, formatting, tests, documentation, and DCO sign-offs. Use when asked to verify changes, prepare a commit or PR, or review a PR's validation and coverage.

原文语言：英语

rl-finetune-speech-llm

Adapt Nemotron Transcribe to customer audio with NeMo-RL GSPO. Use for reward-based speech fine-tuning, assessing rollout signal, selecting informative training recordings, choosing a bounded training run, or diagnosing RL that improves reward but not task…

原文语言：英语

train-speech-llm

Train a SALM speech LLM from a pretrained LLM and pretrained speech encoders with NeMo speechlm2 and the Automodel backend on a multi-node cluster. Use to set up, launch, monitor, continue, fork, or debug a large multitask training run (ASR, contextual ASR,…

原文语言：英语

train-dflash

Train or adapt a DFlash or DFlash2 speculative draft against a frozen SALM speech-LLM target with NeMo speechlm2. Use for online argmax or offline model-generated labels, pretrained draft adaptation, projection-only ablations, checkpoint resume/export, and…

原文语言：英语

finetune-speech-llm

Fine-tune a SALM speech-LLM checkpoint (speech encoder + projection + LLM decoder) with NeMo speechlm2 on a new domain or task, and evaluate it with vLLM. Use to adapt a speech-to-text LLM to new audio, improve WER on a domain corpus, train speech…

原文语言：英语

quantize-speech-llm

Quantize a SALM speech-LLM checkpoint with a NemotronH decoder to FP8 or NVFP4 for vLLM serving with the NeMo SpeechLM plugin, including the audio encoder and the MTP draft head, and verify the result. Use to produce a quantized checkpoint, check one before…

原文语言：英语

fix-issue

Fix a GitHub issue in NeMo Speech (NVIDIA-NeMo/Speech). Read the issue, reproduce the bug with a failing test, implement the fix, and verify tests pass. Only opens a PR if the user explicitly asks for it.

原文语言：英语

migrate-to-resumable-dataloader

This skill should be used when the user asks to "migrate to the resumable dataloader", "switch to indexed Lhotse", "adopt the indexed + resumable pipeline", "make my training resumable", "set up StatefulDataLoader for NeMo/Lhotse", "use index packs", "use…

原文语言：英语

nemo-speech-asr-finetune

Guide NeMo Speech users through ASR fine-tuning with container setup and Lhotse training.

原文语言：英语

debug-training-logs

Debug distributed training failures (NeMo, Megatron, PyTorch) from worker stderr logs and optional AIStore daemon logs. Finds root cause across NCCL timeouts, data loading errors, and storage failures.

原文语言：英语