skills.rest
loteiron
loteironCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

lambda-labs

Launch and manage Lambda Labs GPU cloud instances for ML training and inference.

Updated Sep 10, 2026
Search Tags:#PyTorch#infrastructure#distributed training#H100#ML training#lambda labs#gpu cloud
One-click install
npx skills add https://github.com/loteiron/ZeusAgent --skill lambda-labs-loteiron
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: lambda-labs
Source: https://github.com/loteiron/ZeusAgent/tree/main/optional-skills/mlops/lambda-labs
Command: npx skills add https://github.com/loteiron/ZeusAgent --skill lambda-labs-loteiron

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires lambda-cloud-client, and includes references (resource) components.

What problem does it solve? Provisioning dedicated GPU infrastructure for machine learning workloads is complex and time-consuming. This Skill guides you through launching, configuring, and managing Lambda Labs GPU instances and clusters so you can run training and inference jobs without building your own hardware stack. ## Core Features & Use Cases - Instance Lifecycle Management: Launch, list, and terminate GPU instances (B200, H100, A100, A10, and more) via the Python API, curl CLI, or web console. - Distributed Training Workflows: Run multi-GPU and multi-node training with PyTorch DDP, FSDP, DeepSpeed, and Slurm-based 1-Click Clusters with InfiniBand. - Persistent Storage & Cost Optimization: Attach persistent filesystems for checkpoints and datasets, and select the right GPU for each workload to control costs. - Use Case: You need to fine-tune a 7B LLM. Launch an 8x H100 instance with an attached filesystem, download the base model, run distributed fine-tuning with accelerate, and save checkpoints that survive instance termination. ## Quick Start Ask the agent to launch a single H100 instance on Lambda Labs in the us-west-1 region using your configured API key and SSH key, then verify the GPU with nvidia-smi over SSH.

Frequently Asked Questions about lambda-labs

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I launch a GPU instance on Lambda Labs?▼

Launch a Lambda Labs GPU instance via the web console, the Python lambda-cloud-client API, or curl against the REST API. You must add an SSH key and payment method first; single-GPU instances boot in 3-5 minutes, multi-GPU in 10-15 minutes.

Lambda Labs vs RunPod vs Vast.ai for GPU cloud?▼

Lambda Labs offers dedicated instances with full SSH access, pre-installed Lambda Stack, and no egress fees. RunPod is better for cheaper spot instances and serverless endpoints, while Vast.ai is a marketplace with the lowest prices but less consistency.

Which Lambda Labs GPU should I use for LLM fine-tuning?▼

For fine-tuning a 7B model, a single A100 40GB works well; 70B models typically need 8x H100 instances. For inference, A10 or A6000 GPUs offer the best cost efficiency, while B200 provides maximum training performance.

Does Lambda Labs support multi-node distributed training?▼

Yes, Lambda Labs supports multi-node training via 1-Click Clusters with 16-512 H100 or B200 GPUs, Slurm scheduling, and 400 Gb/s InfiniBand. You can also manually run torchrun across multiple instances in the same region using private IPs.

Why did I lose my data after terminating a Lambda instance?▼

The instance root volume is ephemeral, so all local data is deleted on termination. Attach a persistent filesystem at launch time and store datasets, checkpoints, and outputs under /lambda/nfs/<filesystem_name> to preserve them.

How do I fix SSH permission denied on a Lambda instance?▼

SSH permission denied usually means the key was not added before launch or has wrong permissions. Verify the key exists in the Lambda console, set chmod 600 on the private key, and confirm the instance status is active before connecting.

Related Skills

Explore More Software Engineering Skills→
KarlinskySKarlinskyS

lambda-labs-gpu-cloud

Provision Lambda Labs GPU cloud instances for machine learning training and inference.

Community
Intermediate
qcmuuqcmuu

fine-tuning-serving-openpi

Fine-tune and serve OpenPI robot policies for manipulation datasets.

Community
Advanced
zjunlpzjunlp
75

monitor-experiment

Monitor remote GPU experiments, collect results, and finalize cost manifests over SSH.

Official
Advanced
MLT-OSSMLT-OSS
14

modal-serverless-gpu

Deploy and run ML workloads on Modal's serverless GPU cloud platform.

Official
Advanced