Niko1221/Strata v0.1.2

Strata v0.1.2

on GitHub

Streaming and a fix for requests that never finished (mostly turns with a picture).

• Tool calls stream as they are written (OpenAI tool_calls deltas, Anthropic input_json_delta); keep-alives while a long prompt is read; progress in the server window and GET /status.
• Closing the connection or pressing stop really stops the model (new engine STOP).
• The expert cache is sized after its slots are written: Windows only counts GPU memory once it is touched, so the old sizing could fill the card to 0 MiB and a request then stalled while the driver paged. The prompt chunk now shrinks instead of allocating on top of a full card, and the image encoder warms up before the engine starts.

Run START-HERE.bat once and it picks this engine up by itself (setup now requires 0.1.2).

The ready-made Strata engine for Windows (RTX 30 / 40 / 50: sm_86, sm_89, sm_120 + PTX), CUDA 13.0, NVIDIA driver 580 or newer. Contents: strata.exe (the engine), strata-vision.exe (the optional image encoder), BUILD.json.