llama.cpp on a jailbroken PS5: run any GGUF model natively on the console, starting with Qwen 3.8
Commits on main
v0.7.0: real drive space, M.2 and USB downloads, unload before download
fdcd90d Abhiram committed 1 month ago
v0.6.0: chat tools, images and Get models, checked on the console
d971139 Abhiram committed 1 month ago
Get models: follow redirects ourselves, 4 MiB reads, write() downloads
a936cdb Abhiram committed 1 month ago
v0.5.0: opencode, codex and Claude Code checked against the console
31a9e0d Abhiram committed 1 month ago
Patch llama-server: fold later system messages for the Anthropic API
43c57d9 Abhiram committed 1 month ago
Patch llama-server: fold later system messages for the Anthropic API
dc5aedf Abhiram committed 1 month ago
Get models: a search request from the PC (ps5lm-app.sh search <words>)
f0ddd43 Abhiram committed 1 month ago
Get models page: Hugging Face search, fit from the header, download
318fdd1 Abhiram committed 1 month ago
v0.3.0: thermals through ps5-exporter, remote presses, checked on the console
e368280 Abhiram committed 1 month ago
Chat tools and images: file tools in a scratch folder, mmproj pairing
0df54ef Abhiram committed 1 month ago
CLI access: opencode, codex, Claude Code and aider against the console
8f58648 Abhiram committed 1 month ago
Dashboard pages: Settings and Logs on L1/R1, model details, new tiles
31ecc20 Abhiram committed 1 month ago
App: load models straight from USB, library in sections, no auto-load
a39dc6f Abhiram committed 1 month ago
App: dashboard on the TV, model library, presets and USB import
266fb74 Abhiram committed 1 month ago
memprobe: main direct memory and locked memory limits (#10)
13b4d5e Mert Cobanov committed 1 month ago
ps5lm: survive rest mode, refuse a second copy, sturdier downloads, Quit
37071f8 Mert Cobanov committed 1 month ago
Build: mbedTLS with threads, and tools that leave through _Exit
8be0641 Mert Cobanov committed 1 month ago