{"apiVersion":"v1","taxonomyVersion":1,"generatedAt":"2026-10-29T20:44:34.678Z","data":{"recordId":"wsa_ged64r9qzrqm","score":1,"confidence":"high","capped":false,"capReason":null,"sources":[{"n":1,"url":"https://github.com/xingjian-bai/opal","publisher":"GitHub new repositories","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-29T00:58:41.173Z","detectedPublishedAt":"2026-10-21T14:17:50.000Z","rights":"original"}],"claims":[{"id":"a1ab55ea-9df5-4e9a-8fbc-c0d21851e9b2","kind":"capability","statement":"OPAL lets a coding agent rewrite optimizer and analysis programs during a training run.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"rewrite two small programs: the optimizer program, which turns gradients into updates at every step, and an analysis","rights":"original"}}]},{"id":"b9c372ee-0b16-459e-afcd-dde4d77ce508","kind":"availability","statement":"The package is available on PyPI as `opal-optimizer`.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"The package is called `opal-optimizer` on PyPI, because `opal` there is an unrelated web framework.","rights":"original"}}]},{"id":"f8b03182-2964-4733-b12f-3f8845c0d9c3","kind":"deployment","statement":"OPAL supports PyTorch training on CPU, one GPU, or several GPUs of one node (DDP).","citations":[{"source":1,"supportProbability":0.97,"supports":true,"quote":{"text":"- **A PyTorch training script** that uses `torch.optim` optimizers, on CPU, one GPU or several GPUs of one node (DDP).","rights":"original"}}]},{"id":"f4708ddf-b80d-44b8-b7b9-657a379075d0","kind":"pricing","statement":"Agent usage is billed to the Claude subscription or API key, and there is no cost cap by default.","citations":[{"source":1,"supportProbability":0.97,"supports":true,"quote":null}]},{"id":"94f8039a-a1bd-488f-80dd-a23075a967ee","kind":"privacy","statement":"The agent can receive read-only copies of the training script and model code; `share_code=False` keeps them out.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"analysis reports, and read-only copies of your training script and model code (`share_code=False` keeps them out).","rights":"original"}}]}]},"nextCursor":null,"license":{"catalog":{"name":"CC BY 4.0","spdx":"CC-BY-4.0","url":"https://creativecommons.org/licenses/by/4.0/","attribution":"wowsuchai","attributionUrl":"https://wowsuchai.withglint.app","covers":["catalog metadata","classifications","generated summaries","feed event metadata"]},"excluded":{"covers":["source quotations","linked source content"],"terms":"Not licensed by wowsuchai. Source quotations and linked source content keep their original rights."}}}