{"adoption": {"forks": 334, "observed_at": "2026-08-28T04:07:54.091769+00:00", "stars": 3295}, "canonical_url": "https://ross.abutalabs.com/products/lightseq", "card": {"archived": true, "artifact_type": "library", "description": "LightSeq: A High Performance Library for Sequence Processing and Generation", "domain": ["deep-learning", "machine-learning", "gpu-computing"], "enriched": true, "function": ["llm-inference", "llm-training", "machine-learning", "deep-learning", "nlp", "gpu-computing"], "health_score": 10, "homepage": null, "language": "C++", "license": "NOASSERTION", "license_family": "other", "maturity": "maintenance", "member_repos": ["bytedance/lightseq"], "name": "bytedance/lightseq", "platform": ["python", "cpp"], "pushed_at": "2023-05-16T10:47:48+00:00", "repo": "bytedance/lightseq", "stars": 3295, "tags": ["transformer", "cuda-kernels", "beam-search", "bert", "gpt", "mixed-precision", "inference-server", "machine-translation", "natural-language-processing", "linux", "gpu", "docker"], "topics": ["inference", "transformer", "beam-search", "bert", "cuda", "sampling", "diverse-decoding", "multilingual-nmt", "training", "bart", "gpt", "accelerate"], "urls": [], "use_cases": ["accelerate transformer model training on gpu", "run fast transformer inference with beam search", "fine-tune bert with fp16 mixed precision", "deploy gpt or bart models with an inference server", "speed up machine translation training", "int8 quantized inference for transformer models"], "what_it_is": "LightSeq is a high-performance CUDA-based library for training and inference of sequence models like Transformer, BERT, GPT, and BART, with fused custom kernels and fp16/int8 mixed-precision support. It also provides a serving backend built on TensorRT Inference Server for easy deployment.", "when_to_avoid": ["you need a general-purpose framework for non-Transformer architectures", "you require active maintenance and recent updates, as the last release was in 2023", "you need CPU-only inference"], "when_to_choose": ["you need maximum GPU throughput for Transformer-family training or inference", "you want a deployable serving backend for sequence generation models", "you work with fairseq or Hugging Face models and want drop-in acceleration"]}, "data_as_of": "2026-08-30T08:39:29.467469+00:00", "members": [{"path": "/products/lightseq", "repo": "bytedance/lightseq", "role": "main", "score": 10}], "provenance": {"archived": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "artifact_type": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "description": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "domain": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "enriched": {"inputs": [], "kind": "computed", "method": "enrichment_status"}, "function": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "health_score": {"inputs": ["days_since_push", "days_since_release", "archived"], "kind": "computed", "method": "health_v1"}, "homepage": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "language": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "license": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "license_family": {"inputs": ["license"], "kind": "computed", "method": "license_family"}, "maturity": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "member_repos": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "name": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "platform": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "pushed_at": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "repo": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "stars": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "tags": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "topics": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "urls": {"kind": "observed", "observed_at": "2026-08-28T04:07:54.091769+00:00", "source": "github"}, "use_cases": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "what_it_is": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "when_to_avoid": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}, "when_to_choose": {"confidence": null, "enriched_at": "2026-08-30T07:22:43.168510+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "98121c669f66a26be814bdd71a23d7b14bbc7a7aba17306293705a2af3ca4717", "fetched_at": "2026-08-28T04:07:54.091769+00:00", "kind": "readme", "missing": false, "url": "https://github.com/bytedance/lightseq"}], "taxonomy_version": 1}}, "score": {"components": {"activity": 0, "longevity": 100, "rhythm": 8}, "computed_at": "2026-09-02T17:46:02.011165+00:00", "flags": ["archived", "no_license"], "formula": "round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)", "inputs": {"age_days": 2462, "days_push": 1205, "days_rel": null, "gap_med": null, "n_releases_24m": 0}, "score": 10, "version": 2}, "staleness": {"enrichment_outdated": false, "low_confidence": false, "scrape_days": 9, "stale_scrape": false}}