Repository Issue Activity (beta)

#9008 Studio: server deadlocks with every thread blocked in `sqlite3.connect()` / `close()` and stops accepting connections

#9581 [Feature] AMD Docker image

#9039 [Bug] Tool calling fails with "invalid tool call arguments (invalid_request_error)" on every chat via custom endpoint

#9522 Allowing API to be accessed through LAN or wsl

#9480 [Studio Bug] Model Hub crashes WebKitWebProcess (SIGABRT) — Skia COLRv1 font assert, not #9393

#9580 studiobench: ab_plan rows are silently dropped by the header section collapse in assemble_rows

#9518 [Bug] the chat window's own send-and-persist path failing to write to its local IndexedDB store

#9576 [Feature] Unsloth Studio supports MultiGPU Training

#9551 [Bug] AMD GPU's (RDNA3 W7900) VRAM detection incorrect after VRAM Management Patch

#9448 Images: /v1/images/generations can run a replacement model with the previous model's steps and guidance

#9543 [Bug] Linux: voice dictation (STT) records no audio — WebKitGTK MediaRecorder produces 0 bytes

#2490 [Bug]PeftModelForCausalLM has not attribute '_flag_for_generation'

#9484 [Bug] Canceled prompts produce consecutive user messages (not supported by some Jinja templates)

#8838 [Bug] Deep Research with external model (Ollama Cloud) returns Local model never started producing output

#9550 [Bug] Unsloth Studio Dropping MTP to Fit more Context Window, Loads taking extremely long, Choosing MTP Keeps same Context

#9549 [Bug] Unsloth Studio / Desktop loading into RAM despite Checks for VRAM only

#9548 [Bug] Source / Model Name in Top Corner is being based on ../<Foldername> from the Model.

#9546 [Bug] Prior Chats contain Links to "Switch Back to Original Model" but CTX of "Original Model" Loads @ 4096

#9479 [Bug] Installation failed: create venv

#9235 Studio: one failed markdown chunk takes down the whole app instead of degrading to plain text

#9540 Studio: a link reference definition with a label over 200 characters can be committed away, breaking the reference

#9537 [Studio Bug] Opening the app reads and downloads all model files from synced drives

#9486 [Feature] Memory System

#9500 [Studio Bug] Presets won't save - 'Chat settings could not be persisted' (400 on PUT /api/chat/settings)

#9524 [Bug] qLora training on XPU Fails due to tensor (off)loading issue