{"adoption": {"forks": 183, "observed_at": "2026-08-28T04:04:24.778367+00:00", "stars": 1333}, "canonical_url": "https://ross.abutalabs.com/products/inplace_abn", "card": {"archived": false, "artifact_type": "library", "description": "In-Place Activated BatchNorm for Memory-Optimized Training of DNNs", "domain": ["deep-learning", "machine-learning", "computer-vision", "gpu-computing"], "enriched": true, "function": ["machine-learning", "deep-learning", "gpu-computing"], "health_score": 76, "homepage": null, "language": "Python", "license": "BSD-3-Clause", "license_family": "permissive", "maturity": "stable", "member_repos": ["mapillary/inplace_abn"], "name": "mapillary/inplace_abn", "platform": ["python"], "pushed_at": "2026-07-25T00:24:44+00:00", "repo": "mapillary/inplace_abn", "stars": 1333, "tags": ["pytorch", "batch-normalization", "memory-optimization", "cuda-extension", "semantic-segmentation", "imagenet", "training-efficiency", "neural-networks", "linux", "gpu"], "topics": [], "urls": [], "use_cases": ["reduce GPU memory usage when training deep networks in PyTorch", "train larger ResNet or ResNeXt models on limited GPU memory", "fit bigger batch sizes on a single GPU during CNN training", "fuse batch normalization and activation into one memory-efficient layer", "reproduce memory-optimized ImageNet or semantic segmentation training results"], "what_it_is": "A PyTorch extension library implementing In-Place Activated BatchNorm (InPlace-ABN), which redefines BN plus nonlinear activation as a single in-place operation to reduce training GPU memory usage by up to 50%. It ships the iABN layer plus training scripts reproducing ImageNet classification and Mapillary Vistas semantic segmentation results.", "when_to_avoid": ["you develop or deploy on non-Linux platforms or older CUDA versions", "you need to transfer weights from networks trained with standard BatchNorm, since the scaling parameterization is incompatible", "you use a framework other than PyTorch or your models are small enough that memory savings do not matter"], "when_to_choose": ["you train deep CNNs in PyTorch on Linux with CUDA >= 10 and GPU memory is the bottleneck", "you want to increase model size or batch size without upgrading hardware", "you need an invertible BN+activation layer compatible with recomputation-based backpropagation"]}, "data_as_of": "2026-08-30T08:39:29.467469+00:00", "members": [{"path": "/products/inplace_abn", "repo": "mapillary/inplace_abn", "role": "main", "score": 65}], "provenance": {"archived": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "artifact_type": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "description": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "domain": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "enriched": {"inputs": [], "kind": "computed", "method": "enrichment_status"}, "function": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "health_score": {"inputs": ["days_since_push", "days_since_release", "archived"], "kind": "computed", "method": "health_v1"}, "homepage": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "language": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "license": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "license_family": {"inputs": ["license"], "kind": "computed", "method": "license_family"}, "maturity": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "member_repos": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "name": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "platform": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "pushed_at": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "repo": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "stars": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "tags": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "topics": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "urls": {"kind": "observed", "observed_at": "2026-08-28T04:04:24.778367+00:00", "source": "github"}, "use_cases": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "what_it_is": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "when_to_avoid": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}, "when_to_choose": {"confidence": null, "enriched_at": "2026-08-30T04:44:53.710243+00:00", "kind": "inferred", "prompt_version": 1, "sources": [{"content_hash": "a02f31295f5fcd0a62bc8a2c53ec1a140fd4ddc6a99086c62b86392edf2192ae", "fetched_at": "2026-08-28T04:04:24.778367+00:00", "kind": "readme", "missing": false, "url": "https://github.com/mapillary/inplace_abn"}, {"content_hash": "f48c8067b22df50df0402f146d963c0400b52a400aa72a6ac014668f71c687d6", "fetched_at": "2026-08-29T12:04:02.924596+00:00", "kind": "registry_pypi", "missing": false, "url": "https://pypi.org/pypi/inplace_abn/json"}], "taxonomy_version": 1}}, "score": {"components": {"activity": 94, "longevity": 100, "rhythm": 8}, "computed_at": "2026-09-03T02:20:16.233290+00:00", "flags": [], "formula": "round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)", "inputs": {"age_days": 3205, "days_push": 40, "days_rel": null, "gap_med": null, "n_releases_24m": 0}, "score": 65, "version": 2}, "staleness": {"enrichment_outdated": false, "low_confidence": false, "scrape_days": 9, "stale_scrape": false}}