{"data":{"slug":"nvidia-nemo-speech","name":"Speech","tagline":"A scalable generative AI framework for Speech AI","github_url":"https://github.com/NVIDIA-NeMo/Speech","owner":"NVIDIA-NeMo","repo":"Speech","owner_avatar_url":"https://avatars.githubusercontent.com/u/213689629?v=4","primary_language":"Python","stars":17755,"forks":3499,"topics":["asr","deeplearning","generative-ai","machine-translation","neural-networks","speaker-diariazation","speaker-recognition","speech-synthesis","speech-translation","tts"],"archived":false,"github_pushed_at":"2026-07-11T06:43:18+00:00","maintenance_label":"Very active","url":"https://www.graphcanon.com/tools/nvidia-nemo-speech","markdown_url":"https://www.graphcanon.com/tools/nvidia-nemo-speech.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/nvidia-nemo-speech","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=nvidia-nemo-speech","description":"A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)","homepage_url":"https://docs.nvidia.com/nemo/speech/nightly/index.html","license":"Apache-2.0","open_issues":208,"watchers":231,"ai_summary":"NVIDIA-NeMo/Speech is a comprehensive toolkit for Automatic Speech Recognition (ASR), speaker diarization, speaker recognition, speech synthesis (TTS), and more. It is built on PyTorch and supports CUDA for efficient GPU utilization.","readme_excerpt":"## Requirements\n\nNeMo Speech works with the **Python, PyTorch, and CUDA versions of your choosing**:\n\n- Python 3.12 or above\n- PyTorch 2.7 or above (CPU, CUDA, etc. — your choice)\n- NVIDIA GPU + CUDA (required for training; recommended for inference)\n\nIf you already have a Python/PyTorch/CUDA stack that satisfies those minimums, NeMo Speech installs on top of it **without replacing it**, so your existing PyTorch build is kept (see the install options below). The versions pinned in `uv.lock` and shipped in the official container — Python 3.13, PyTorch 2.12, CUDA 12.6/13.2 — are simply the combination we actively test and support. They make setup turnkey and reproducible, but they are **not** a hard requirement.\n\nAs of [Pytorch 2.6](https://docs.pytorch.org/docs/stable/notes/serialization.html#torch-load-with-weights-only-true),\n`torch.load` defaults to using `weights_only=True`. Some model checkpoints may require using `weights_only=False`.\nIn this case, you can set the env var `TORCH_FORCE_NO_WEIGHTS_ONLY_LOAD=1` before running code that uses `torch.load`.\nHowever, this should only be done with trusted files. Loading files from untrusted sources with more than weights only\ncan have the risk of arbitrary code execution.\n\n---\n\n## Install NeMo Speech\n\nThe recommended way to install NeMo Speech is from source with [uv](https://docs.astral.sh/uv/), which reproduces our actively-tested stack from the committed `uv.lock`. If you need different Python/PyTorch/CUDA versions, NeMo also installs over your existing environment via pip — see the [pip fallback](#from-pypi-with-pip-fallback--bring-your-own-versions) below.\n\n---\n\n### Docker (turnkey, our supported stack)\n\n> **NGC container:** _Coming soon — the pull command for the prebuilt NeMo Speech container image will be published here._\n\nTo build the container from source (CUDA 13 / H100+ by default):\n\n```bash\ngit clone https://github.com/NVIDIA-NeMo/NeMo.git\ncd NeMo\ndocker buildx build -f docker/Dockerfile -t nemo-speech .          # CUDA 13 / H100+ (default)\ndocker run --rm -it --gpus all -v \"$PWD:/workspace\" nemo-speech bash\n```\n\nFor A100, set `GPU_TARGET=a100` — A100 works with **both CUDA 12 and CUDA 13** (CUDA 13, the default base image, is recommended; the CUDA 12 base is a convenience). See the header of [`docker/Dockerfile`](docker/Dockerfile) for all build arguments (`BASE_IMAGE`, `GPU_TARGET`).","github_created_at":"2019-08-05T20:16:42+00:00","created_at":"2026-07-11T10:36:04.487408+00:00","updated_at":"2026-07-12T07:56:32.154776+00:00","categories":[{"slug":"developer-tools","name":"Developer Tools","url":"https://www.graphcanon.com/categories/developer-tools","markdown_url":"https://www.graphcanon.com/categories/developer-tools.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/developer-tools"},{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"},{"slug":"speech-audio","name":"Speech & Audio","url":"https://www.graphcanon.com/categories/speech-audio","markdown_url":"https://www.graphcanon.com/categories/speech-audio.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/speech-audio"}],"tags":[{"slug":"asr","name":"asr"},{"slug":"deeplearning","name":"deeplearning"},{"slug":"generative-ai","name":"generative-ai"},{"slug":"machine-translation","name":"machine-translation"},{"slug":"neural-networks","name":"neural-networks"},{"slug":"speaker-diariazation","name":"speaker-diariazation"},{"slug":"speaker-recognition","name":"speaker-recognition"},{"slug":"speech-synthesis","name":"speech-synthesis"}],"trust":{"provenance":{"is_fork":false,"github_id":200722670,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-07-11T10:36:05.090Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":1,"days_since_push":0,"last_release_at":"2026-04-23T17:40:04Z"},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T10:36:06.010Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-07-11T14:36:38.478Z"},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-07-11T14:36:38.478Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-07-11T14:36:38.478Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When working on projects that require extensive GPU utilization for training large models due to its support for efficient CUDA usage.","For researchers and developers looking for a comprehensive solution that supports not only Automatic Speech Recognition (ASR) but also includes speaker diarization, recognition, speech synthesis (TTS)","If you are leveraging NVIDIA GPUs as it is optimized with CUDA for robust training and recommended for inference."],"when_not_to_use":["For environments where GPU access is limited or unavailable since the toolkit highly recommends a GPU setup for both training and recommended for inference.","If your Python/PyTorch/CUDA versions fall below the specified requirements (Python 3.12+, PyTorch 2.7+), as lower versions will not be compatible with NeMo Speech.","In scenarios where you're working with models that do not require or benefit significantly from GPU acceleration, given its architecture optimized for GPU use."],"source":"enrich:decision_facts","observed_at":"2026-07-11T14:37:13.118Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"NVIDIA-NeMo/Speech - A scalable toolkit for speech AI tasks such as ASR, TTS, and speaker recognition built on PyTorch with CUDA support."}]}}