# intelligent-machine-learning/dlrover

DLRover: An Automatic Distributed Deep Learning System

Repository: https://github.com/intelligent-machine-learning/dlrover
Canonical: https://ross.abutalabs.com/products/dlrover
Language: Python
License: NOASSERTION
License Family: other
Topics: distributed-training, k8s, llm-training, hacktoberfest
Last push: 2026-08-26T01:27:14+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 41, longevity 100
- inputs: {"age_days": 1531, "days_push": 8, "days_rel": 239, "gap_med": 113, "n_releases_24m": 6}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1680, forks 217 (observed 2026-08-28T04:05:21.516065+00:00)

## What it is
DLRover is an automatic distributed deep learning system that manages training of large AI models on Kubernetes and Ray clusters. It provides fault tolerance, auto-scaling of resources, and fast in-memory checkpoint recovery so developers can train models without handling distributed engineering themselves.

## Use cases
- train large language models on a kubernetes cluster
- automatically scale distributed training resources up and down
- recover distributed training jobs from node failures
- run fault-tolerant pytorch distributed training
- speed up llm training with automatic operations
- manage deep learning training jobs on ray
- recover training from in-memory checkpoints in seconds

## When to choose
- you train large deep learning or LLM models on k8s or ray and want automated fault tolerance and scaling
- you want training jobs to survive hardware failures without manual restarts
- you want to focus on model architecture rather than distributed systems engineering

## When to avoid
- you only train small single-node models
- you need a lightweight trainer without cluster orchestration
- you use a framework other than pytorch or tensorflow

## Facets
- artifact type: framework
- maturity: active
- function: llm-training, deep-learning, machine-learning, container-orchestration, monitoring, scheduling
- domain: deep-learning, machine-learning, large-language-models, microservices
- platform: python, cloud
- tags: distributed-training, fault-tolerance, auto-scaling, pytorch, tensorflow, ray, k8s-operator, checkpointing, devops, containers, kubernetes, docker, gpu

## Member repositories
- intelligent-machine-learning/dlrover (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:21.516065+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:41:28.062223+00:00, confidence not recorded.
  - readme: https://github.com/intelligent-machine-learning/dlrover (fetched 2026-08-28T04:05:21.516065+00:00, sha 42cbd20298b6)
  - registry_pypi: https://pypi.org/pypi/dlrover/json (fetched 2026-08-29T11:14:36.346837+00:00, sha d785604eb18d)
- Data as of 2026-08-30T08:39:29.467469+00:00.