{"id":27351740,"url":"https://github.com/softmeng/comfyui-deepcache-fix","last_synced_at":"2025-07-01T12:04:38.809Z","repository":{"id":249914680,"uuid":"832939842","full_name":"SoftMeng/ComfyUI-DeepCache-Fix","owner":"SoftMeng","description":"Make ComfyUI Faster!","archived":false,"fork":false,"pushed_at":"2024-07-25T13:09:00.000Z","size":164,"stargazers_count":13,"open_issues_count":2,"forks_count":3,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-06-02T01:36:35.339Z","etag":null,"topics":["comfy","comfyui","comfyui-manager","comfyui-nodes"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/SoftMeng.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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}},"created_at":"2024-07-24T03:31:42.000Z","updated_at":"2025-05-25T06:35:48.000Z","dependencies_parsed_at":"2024-12-19T03:13:33.393Z","dependency_job_id":null,"html_url":"https://github.com/SoftMeng/ComfyUI-DeepCache-Fix","commit_stats":null,"previous_names":["softmeng/comfyui-deepcache-fix"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/SoftMeng/ComfyUI-DeepCache-Fix","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SoftMeng%2FComfyUI-DeepCache-Fix","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SoftMeng%2FComfyUI-DeepCache-Fix/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SoftMeng%2FComfyUI-DeepCache-Fix/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SoftMeng%2FComfyUI-DeepCache-Fix/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SoftMeng","download_url":"https://codeload.github.com/SoftMeng/ComfyUI-DeepCache-Fix/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SoftMeng%2FComfyUI-DeepCache-Fix/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":262959561,"owners_count":23391057,"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","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":["comfy","comfyui","comfyui-manager","comfyui-nodes"],"created_at":"2025-04-12T20:34:24.464Z","updated_at":"2025-07-01T12:04:38.782Z","avatar_url":"https://github.com/SoftMeng.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n    \u003cimg src=\"./doc/icon.png\" width=\"256px\"/\u003e\n\u003c/div\u003e\n\u003cdiv align=\"center\"\u003e\n    富强、民主、文明、和谐、合规\n\u003c/div\u003e\n\u003cdiv align=\"center\"\u003e\n    自由、平等、公正、法治、合法\n\u003c/div\u003e\n\u003cdiv align=\"center\"\u003e\n    爱国、敬业、诚信、友善、合理\n\u003c/div\u003e\n\n# ComfyUI-DeepCache-Fix\n\n## 介绍\n\nFor SDXL LCM,ComfyUI加速节点，生成图片时更快，同时保证了加速前后的一致性，适合大批量生图。[Accelerate ComfyUI Nodes for Faster Image Generation, Ensuring Consistency Pre and Post-Acceleration, Ideal for Bulk Image Production.]\n\n## 用法\n\n将该库放在 ComfyUI/custom_nodes/ 下即可。[Simply place the library in the 'ComfyUI/custom_nodes/' directory to get started.]\n\n## 工作流例子\n\n[DeepCache-Fix.json](doc%2FDeepCache-Fix.json)\n\n## 使用说明\n\n### 插件参数\n\n- cache_interval: 缓存间隔, 单位: 步, 默认:3\n- cache_depth: 缓存深度, 默认:3\n- start_steps: 使用缓存的开始步数, 默认:0\n- end_steps: 使用缓存的结束步数, 默认:12\n- input_cache: 使用输入层缓存，默认:True 开启\n- middle_cahce: 使用中间层缓存，默认:True 开启\n- output_cache: 使用输出层缓存，默认:True 开启\n\n### 举例(实践)\n\n#### 模型(蒸馏)\n\nhttps://www.liblib.art/modelinfo/386109978c19484298d810d6f2830780\n\n#### 生成\n\n在总共16步的执行过程中，我们计划采取分阶段的策略。具体来说，如下\n\n1. 0-12步(start_steps=0, end_steps=12)将利用特定的插件来执行，以提高效率和效果。\n2. 从第13步开始，我们将切换回原始模型，完成剩下的4步。 (一般预留3步以上，不使用加速)\n\n这样的安排旨在结合两者的优势，确保整个流程的顺利进行。\n\n![img.png](doc%2Fimg.png)\n\n## 参考\n\n原始代码参考: https://gist.github.com/laksjdjf/435c512bc19636e9c9af4ee7bea9eb86\n\n感谢 laksjdjf 分享的代码。\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoftmeng%2Fcomfyui-deepcache-fix","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsoftmeng%2Fcomfyui-deepcache-fix","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoftmeng%2Fcomfyui-deepcache-fix/lists"}