-
Notifications
You must be signed in to change notification settings - Fork 509
All issues
Issue creation is restricted in this repository
- #3618 · PCAssistSoftware opened
on Sep 17, 2026 4
Issues
is:issue state:open
is:issue state:open
Search results
ROCm backend reports "Unsupported GPU" on Ryzen AI Max+ 8065S (Gorgon Halo, gfx1151) while native llama.cpp ROCm works
bugSomething isn't workingSomething isn't workingengine::llamacppllama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)llama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)runtime::rocmAMD ROCm runtimeAMD ROCm runtimeStatus: Open.#3662 In lemonade-sdk/lemonade;Error preparing model: Failed to load model 'gemma-4-E4B-it-GGUF-Q4_K_M': llama-server failed to start (Gorgon Halo gfx1151, Windows)
bugSomething isn't workingSomething isn't workingengine::llamacppllama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)llama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)runtime::rocmAMD ROCm runtimeAMD ROCm runtimeStatus: Open.#3663 In lemonade-sdk/lemonade;macOS tray delivers notifications from a background thread
area::traySystem tray app (LemonadeServer.exe, lemonade-tray)System tray app (LemonadeServer.exe, lemonade-tray)runtime::metalApple Metal runtime (macOS)Apple Metal runtime (macOS)type:bugSomething brokenSomething brokenStatus: Open.#3661 In lemonade-sdk/lemonade;Default TTS Playback Speed at Speech settings
enhancementNew feature or requestNew feature or requestStatus: Open.#3654 In lemonade-sdk/lemonade;Intel AMD Mac Support
enhancementNew feature or requestNew feature or requestruntime::vulkanVulkan runtime / GPU backendVulkan runtime / GPU backendStatus: Open.#3646 In lemonade-sdk/lemonade;Thread the non-streaming cancel token explicitly and remove the ambient cancel state
enhancementNew feature or requestNew feature or requesttype:choreMaintenance, refactoring, cleanupMaintenance, refactoring, cleanupStatus: Open.#3645 In lemonade-sdk/lemonade;Expose the model's chat template (or a derived thinking-capability flag) in /api/show
area::apiHTTP REST API surface and route handlersHTTP REST API surface and route handlersenhancementNew feature or requestNew feature or requestStatus: Open.#3644 In lemonade-sdk/lemonade;[Known Issue] Models in extra_models_dir from HF caches are not showing correctly in list
bugSomething isn't workingSomething isn't workingStatus: Open.#3618 In lemonade-sdk/lemonade;bundled
llamacpp:rocmbackend produces incorrect output above ~5K prompt tokens on Strix Halo (gfx1151)bugSomething isn't workingSomething isn't workingengine::llamacppllama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)llama.cpp backend (LlamaCppServer); GPU/CPU LLM inference (Vulkan, ROCm, Metal)runtime::rocmAMD ROCm runtimeAMD ROCm runtimeStatus: Open.#3610 In lemonade-sdk/lemonade;Latest FLM/NPU not passing Windows Code Integrity
bugSomething isn't workingSomething isn't workingengine::flmFastFlowLM backend (NPU); multi-modal LLM/ASR/embeddings/rerankingFastFlowLM backend (NPU); multi-modal LLM/ASR/embeddings/rerankingStatus: Open.#3609 In lemonade-sdk/lemonade;/api/v1/responses streaming (SSE) omits required output_index/content_index fields, breaking OpenAI-Responses-compatible clients
area::apiHTTP REST API surface and route handlersHTTP REST API surface and route handlersbugSomething isn't workingSomething isn't workingStatus: Open.#3607 In lemonade-sdk/lemonade;Errors showing in log when searching for models
area::apiHTTP REST API surface and route handlersHTTP REST API surface and route handlersbugSomething isn't workingSomething isn't workingcandidatePertaining to a release candidate.Pertaining to a release candidate.Status: Open.#3605 In lemonade-sdk/lemonade;