Skip to content

Docs/nemotron speech clarification#40

Open
adsahu-nv wants to merge 2 commits into
mainfrom
docs/nemotron-speech-clarification
Open

Docs/nemotron speech clarification#40
adsahu-nv wants to merge 2 commits into
mainfrom
docs/nemotron-speech-clarification

Conversation

@adsahu-nv

Copy link
Copy Markdown
Contributor

No description provided.

adsahu-nv added 2 commits May 12, 2026 02:30
NVIDIA's Nemotron Speech is the open model family (Parakeet for
streaming ASR, Magpie for TTS, etc.) and Riva is the GPU-accelerated
SDK/gRPC server that deploys them. The current docs only said 'Riva',
which is correct for the SDK but obscures the model names.

This adds a short clarification in README, INSTALL, and setup_riva so
readers understand the relationship, and updates two places that
named specific models:

- README intro + Architecture bullet
- INSTALL: Riva section opener
- setup_riva: top-of-doc note + Parakeet 1.1b model section
- setup_riva: 'Expected output' snippet updated to reflect the TTS
  default Riva 2.24.0 actually ships on Jetson (Magpie Multilingual,
  not HiFiGAN)

No script names, container names, config keys, SDK names, file
paths, or anchors are changed — only product-name prose.

Signed-off-by: Aditya Sahu <adsahu@nvidia.com>
Reverts only the top-of-doc explainer paragraph; the inline model
labels (Parakeet 1.1b 'Nemotron Speech ASR', Magpie 'Nemotron Speech'
in the expected output) stay.

Signed-off-by: Aditya Sahu <adsahu@nvidia.com>
@adsahu-nv adsahu-nv requested a review from tokk-nv May 12, 2026 03:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant