Agent Skills by just-it
Showing 14 vetted skills indexed across 1 GitHub repositories.
Just-itdsl-lowering
Translate DSL operator descriptions into AscendC code via a four-pass lowering workflow.
Just-itop-task-extractor
Extract PyTorch operator implementations into KernelBench-ready task_desc.py files.
Just-itdsl-baseline-generation
Generate baseline AscendDSL implementations from functional PyTorch operators.
Just-itop-desc-generation
Generate Ascend operator description JSON files from templates.
Just-itfunctional-conversion
Converts PyTorch MLP models to graph-format ONNX with pluggable backends and optimizers.
Just-itascend-call-generation
Generate Ascend C projects and operator invocation code from functional PyTorch code.
Just-itkernel-verifier
Automate kernel verification by running verify.py and benchmark.py for Ascend/NPU.
Just-itkernel-generator
Convert KernelBench operator descriptions into Triton Ascend kernels with a ModelNew class.
Just-itkernel-designer
Design UnifiedSketch DSL algorithm sketches for Triton Ascend kernels.
Just-itascendc-evaluation
Evaluate AscendC custom operators against reference implementations for correctness and performance.
Just-itascend-benchmark-evaluator
Assess Ascend C operator correctness and runtime using NPUKernelBench datasets.
Just-itlatency-optimizer
Analyze Triton kernels for Ascend NPU and generate optimized versions.
Just-itreference-generation
Generate a reference PyTorch nn.Module from operator description JSON.
Just-itbenchmark-evaluator
Orchestrate KernelBench tasks and collect results via SubAgent execution.