analyzer
01Agent Codex
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
358 5d ago A 0 tokens
copy · 100% MIT
358Stars on the repository
29Mods indexed here, across every type
5d agoLast push, which is what freshness is scored on
MITLicence, which decides whether bodies are shown
comparator
02Agent Codex
Compare two outputs WITHOUT knowing which skill produced them.
358 5d ago A 0 tokens
copy · 100% MIT
grader
03Agent Codex
Evaluate expectations against an execution transcript and outputs.
358 5d ago A 0 tokens
copy · 100% MIT
Agent
An analysis agent for designing TileLang-Ascend operators, which are custom AI computations for Huawei Ascend chips. It produces a design document and an initial set of basic tests from the requirements.
358 5d ago A 62 tokens
original MIT
Agent
An automated coding agent for building and checking TileLang-Ascend operators, which are custom AI computations for Huawei Ascend chips. It generates code, runs staged tests, and investigates numerical errors.
358 5d ago A 73 tokens
original MIT
Agent
An orchestration agent for developing TileLang-Ascend operators, which are hardware-specific computation routines. It manages a three-stage workflow for designing, implementing, testing, and tuning an operator.
358 5d ago A 69 tokens
original MIT
Agent
A specialized agent for one round of performance tuning on TileLang-Ascend GPU kernel code. It analyzes an already accuracy-checked implementation, tries an optimization, tests accuracy again, and keeps or rolls back the change based on measured results.
358 5d ago A 53 tokens
original MIT