{"version":"1.0","workflow_uuid":"b4c1950f-45df-11f1-9bc6-00163e2b0d79","workflow_title":"FlashAttention — Fast and Memory-Efficient Exact Attention","install_contract":{"version":"1.0","installReady":false,"title":"FlashAttention — Fast and Memory-Efficient Exact Attention","summary":"An IO-aware exact attention algorithm that is 2-4x faster than standard attention and uses 5-20x less memory. FlashAttention has become a foundational building block for training and serving modern large language models.","assetType":"Scripts","pageUrl":"https://tokrepo.com/en/workflows/asset-b4c1950f","sourceUrl":"https://github.com/Dao-AILab/flash-attention","intendedFor":[],"firstActions":[],"agentFirstSteps":[],"targetPaths":[],"verification":[],"startingPoints":[],"example":"","successOutcome":"","boundaries":[],"askUserIf":["the current workspace stack cannot be matched to a safe upstream template","the target path is not the project root, or an existing file should be merged instead of overwritten"]}}