{"id":28640436,"url":"https://github.com/theibrahimstudio/finder.livebatch","last_synced_at":"2025-09-25T18:52:28.985Z","repository":{"id":295031790,"uuid":"988859585","full_name":"theIbrahimStudio/finder.LiveBatch","owner":"theIbrahimStudio","description":"A lightweight, framework-agnostic middleware that dynamically batches inference requests in real time to maximize GPU/TPU utilization.","archived":false,"fork":false,"pushed_at":"2025-05-23T11:48:41.000Z","size":20,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-06-12T20:06:45.134Z","etag":null,"topics":["dynamic-batching","golang","gpu","grpc","microservices","ml-inference","model-serving","performance-optimization"],"latest_commit_sha":null,"homepage":"","language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/theIbrahimStudio.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-05-23T07:24:34.000Z","updated_at":"2025-05-23T11:48:45.000Z","dependencies_parsed_at":"2025-05-23T09:10:26.316Z","dependency_job_id":"e72f200b-dd95-49f8-9e0f-033c91856f2a","html_url":"https://github.com/theIbrahimStudio/finder.LiveBatch","commit_stats":null,"previous_names":["theibrahimstudio/finder.livebatch"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/theIbrahimStudio/finder.LiveBatch","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/theIbrahimStudio%2Ffinder.LiveBatch","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/theIbrahimStudio%2Ffinder.LiveBatch/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/theIbrahimStudio%2Ffinder.LiveBatch/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/theIbrahimStudio%2Ffinder.LiveBatch/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/theIbrahimStudio","download_url":"https://codeload.github.com/theIbrahimStudio/finder.LiveBatch/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/theIbrahimStudio%2Ffinder.LiveBatch/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":276966325,"owners_count":25736757,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-09-25T02:00:09.612Z","response_time":80,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["dynamic-batching","golang","gpu","grpc","microservices","ml-inference","model-serving","performance-optimization"],"created_at":"2025-06-12T20:06:42.285Z","updated_at":"2025-09-25T18:52:28.947Z","avatar_url":"https://github.com/theIbrahimStudio.png","language":"Go","readme":"# LiveBatch ⚡\n\n**LiveBatch** is a lightweight, framework-agnostic middleware that dynamically batches inference requests in real time to maximize GPU/TPU utilization. It's designed for ML teams and microservices that want to increase throughput without modifying their model code.\n\n---\n\n## Features\n\n- Plug-and-play: Drop-in proxy over any HTTP ML server\n- Dynamic batching: Configurable by max latency or batch size\n- Written in Go: Fast, concurrent, production-ready\n- Transparent: Accepts and returns single inference calls\n- Efficient: Great for reducing GPU/TPU underutilization\n\n---\n\n## Architecture\n\nLiveBatch acts as a sidecar or proxy in front of your model service.\n\n```\n\\[Client] ---\u003e \\[LiveBatch] ---\u003e \\[Model Server]\n\\|  |\n\\|  |---\u003e Queues requests\n\\|------\u003e Batches \u0026 dispatches\n```\n\n---\n\n## Quick Start\n\nVia command-line flags:\n\n```bash\ngo run main.go --max-batch-size=4 --max-latency-ms=100 --listen-addr=\":9000\"\n```\n\nVia environment variables:\n\n```bash\nLIVEBATCH_MAX_BATCH_SIZE=16 LIVEBATCH_MAX_LATENCY_MS=200 go run main.go\n```\n\n---\n\n## Configuration\n\nLiveBatch supports configuration via **environment variables** or **command-line flags**, using `viper` + `pflag`.\n\n| Name              | Flag               | Env Var                    | Default | Description                      |\n| ----------------- | ------------------ | -------------------------- | ------- | -------------------------------- |\n| Max Batch Size    | `--max-batch-size` | `LIVEBATCH_MAX_BATCH_SIZE` | `8`     | Max number of requests per batch |\n| Max Latency (ms)  | `--max-latency-ms` | `LIVEBATCH_MAX_LATENCY_MS` | `50`    | Max wait time before dispatching |\n| Listening Address | `--listen-addr`    | `LIVEBATCH_LISTEN_ADDR`    | `:8080` | HTTP server bind address         |\n\n---\n\n## Roadmap\n\n- [x] HTTP dynamic batching proxy (MVP)\n- [x] Config via environment or CLI\n- [ ] gRPC and ONNX backend support\n- [ ] Prometheus metrics\n- [ ] Deadline-based and priority queueing\n- [ ] Docker + Helm chart for Kubernetes\n- [ ] Python client SDK\n\n---\n\n## Contributing\n\nPRs welcome! Check out the [CONTRIBUTING.md](https://github.com/theIbrahimStudio/.github/blob/main/CONTRIBUTING.md) for guidelines.\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftheibrahimstudio%2Ffinder.livebatch","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftheibrahimstudio%2Ffinder.livebatch","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftheibrahimstudio%2Ffinder.livebatch/lists"}