magnetar
01Skill Claude CodeCodex
Convert remote or local AI models into AXera AXMODEL packages with Python and C++ SDKs for customer delivery.
Skill Claude CodeCodex
Convert remote or local AI models into AXera AXMODEL packages with Python and C++ SDKs for customer delivery.
Skill Claude CodeCodex
Hidden stage for magnetar. Acquire a remote or local model into TASKDIR/origin without modifying the source.
Skill Claude CodeCodex
Hidden stage for magnetar. Compile static ONNX to AXMODEL with Pulsar2 and record compile artifacts.
Skill Claude CodeCodex
Hidden stage for magnetar. Export the acquired model to static-shape ONNX, validate it against the source model, and generate modelmeta.json plus calibration data.
Skill Claude CodeCodex
Hidden stage for magnetar. Initialize TASKDIR structure, logs, and status files.
Skill Claude CodeCodex
Hidden stage for magnetar. Assemble validated AXMODEL, SDKs, reports, and usage documentation into a customer delivery directory.
Skill Claude CodeCodex
Hidden stage for magnetar. Publish the validated package to GitHub or HuggingFace after asking user for target, repo name, and credentials.
Skill Claude CodeCodex
Hidden stage for magnetar. Optionally deploy AXMODEL and SDK examples to an AX board and verify runtime behavior.
Skill Claude CodeCodex
Hidden stage for magnetar. Generate customer-facing Python and C++ inference SDKs from modelmeta.json and AXMODEL artifacts.
Skill Claude CodeCodex
Hidden stage for magnetar. Compare ONNX outputs with AXMODEL simulation outputs using task-relevant metrics.
Skill Claude CodeCodex
Hidden stage for magnetar. Ensure Pulsar2 and the chip-specific C++ BSP/cross-compilation toolchain are available for compile and SDK validation.
Instructions file CodexOpenCode
Instructions for AXERA-TECH/Magnetar, covering agents.md, 项目目标, 工具库, 执行流程 and stop 点.
Instructions file
Instructions for AXERA-TECH/Magnetar, a project described as: Pulsar2 Agent, Build a 10‑stage automated pipeline powered by AI Agent. Convert floating‑point models into out‑of‑the‑box full AX‑chip delivery packages with a single natural‑language instruction to empower algorithm engineers.