Instructions file CodexOpenCode
Instructions for AMD-AGI/Apex, covering agents.md, project overview, environment (always set first), critical: read-only codebase and gpu cleanup (only if needed).
76 5d ago A 9,941 tokens
original MIT
Agents, and RL environment, for optimizing GPU kernels on AMD ROCm using LLM agents. Benchmarks LLM serving workloads end-to-end, profiles bottleneck kernels, optimizes them via Claude Code or Codex, and scores on compilation, correctness, and speedup.
Instructions file CodexOpenCode
Instructions for AMD-AGI/Apex, covering agents.md, project overview, environment (always set first), critical: read-only codebase and gpu cleanup (only if needed).
Instructions file
Instructions for AMD-AGI/Apex, covering claude.md, project overview, environment (always set first), critical: read-only codebase and gpu cleanup (only if needed).