None of the candidate sources is DeepSpeed itself, and none of the pages mention it, so no product here is a direct match for the request. The closest fits for distributed deep learning training are PyTorch, whose page describes the framework and its distributed features 6, and Hugging Face, which hosts the ML collaboration platform and inference tooling 9; NVIDIA's NeMo AutoModel guide covers distributed training strategies such as FSDP2 1.
- 1
PyTorchpytorch.orgverified and signed
The open source deep learning framework from the PyTorch Foundation, which includes distributed training support and a path to production 6.
+18 community tools built for it
for agents: pytorch · learning · export
- 2
NVIDIA NeMo AutoModelnvidia.comverified and signed
A guide for selecting and configuring distributed training strategies, including FSDP2 and Megatron FSDP, in NeMo AutoModel 1.
official:MCP server · NVIDIA CUDA Docs+84 community tools built for it
for agents: nemo-automodel-distributed-training · inference · deepstream-dev
- 3
Hugging Facehuggingface.coverified and signed
A platform for hosting and collaborating on ML models and datasets, with Hugging Face Hub and inference tooling 9.
official:MCP server · @huggingface/mcp-clientSDK · huggingface-hubCLI · text-embeddings-inferencerepository · pwc-cliskill · hf-cli+23 community tools built for it
for agents: huggingface-inference-api · inference · huggingface-inference-api-typescript
Sources
- 1Brev.dev Consolebrev.nvidia.com · read 2026-10-29
- 6PyTorch Foundation - PyTorchpytorch.org · read 2026-10-29
- 9Hugging Face – The AI community building the future.huggingface.co · read 2026-10-29
For agents: the 10 actions behind this answer
nemo-automodel-distributed-training · nemo-automodel-distributed-training verified and signed
Guide for selecting and configuring distributed training strategies in NeMo AutoModel, including FSDP2, Megatron FSDP, DDP, and parallelism settings.
aspire · aspire verified and signed
Aspire skill covering the Aspire CLI, AppHost orchestration, service discovery, integrations, MCP server, VS Code extension, Dev Containers, GitHub Codespaces, templates, dashboard, and deployment. Use when the user asks to create, run, debug, configure, deploy, or troubleshoot an Aspire distributed
gke-inference · gke-inference verified and signed
Deploys and optimizes AI/ML inference workloads on GKE, using GPUs, TPUs, and model servers. Use when deploying GKE inference servers, configuring GKE GPU resources for inference, or deploying LLMs on GKE. Don't use for migrating existing AI workloads to GKE (use google-cloud-solution-guided-gke-ai-
data-profiling · pydeequ verified and signed
radar-get-ai-inference-summary · Cloudflare API verified and signed
Get Workers AI inference distribution by dimension
pytorch · torch verified and signed
gRPC · Grpc.Tools verified and signed
azure-ai-inference-py · azure-ai-inference-py verified and signed
Azure AI Inference SDK for Python. Use for chat completions, embeddings, and model inference with Azure AI Foundry endpoints. Triggers: "azure-ai-inference", "ChatCompletionsClient", "EmbeddingsClient", "model inference", "Foundry models".
huggingface-inference-api · @huggingface/inference verified and signed
python · fastcore verified and signed
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.