4 commits

Author	SHA1	Message	Date
Claude	2c77f6f968	feat: use GGUF metadata for model type and context window auto-detection ... Replaces filename-based guessModelType with GGUF header parsing. Caps default context at 4096 to prevent VRAM exhaustion on models with 128K+ native context. Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 22:23:07 +00:00
Claude	c50a8e9e9b	feat: retry port selection in startServer on process failure ... Co-Authored-By: Virgil <virgil@lethean.io> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-02-19 21:40:05 +00:00
Claude	1d8d65f55b	feat: Backend Available() and LoadModel() with GPU detection ... Replace stub backend with real implementation: Available() checks /dev/kfd and llama-server presence, LoadModel() wires up server lifecycle to return a rocmModel. Add guessModelType() for architecture detection from GGUF filenames (handles hyphenated variants like Llama-3). Add TestAvailable and TestGuessModelType. Co-Authored-By: Virgil <virgil@lethean.io> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-02-19 21:12:02 +00:00
Snider	aa42cff417	feat: scaffold go-rocm AMD GPU inference package ... Implements inference.Backend via llama-server subprocess (llama.cpp + HIP/ROCm). Targets RX 7800 XT (gfx1101, RDNA 3, 16GB VRAM). Includes: - Backend registration with build tags (linux/amd64) - Stub backend.go with llama-server lifecycle outline - CLAUDE.md with build instructions for llama.cpp + ROCm - TODO.md with 5-phase task queue - FINDINGS.md with hardware specs, VRAM budget, design rationale Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 19:39:40 +00:00