credence for Ollaya

Ollaya package of Txoka/Credence-v1-Gemma4-E4B by Txoka. Ollaya runs open decision models locally, the way Ollama runs LLMs: typed questions in, calibrated answers out, behind a TypeSafe-compatible API.

ollaya run credence

What is in this repository

This repository holds only the files Ollaya derives, with no weights. The model is the authors' own GGUF file: ollaya pull downloads it from their repository, unmodified and pinned to a commit, verifies its sha256, and Ollaya runs it on llama.cpp.

Tag	The authors' GGUF	Files
credence:e4b	Txoka/Credence-v1-Gemma4-E4B@7d5ffc8 accuracy/model-Q8_0.gguf	e4b/decision.json, e4b/calibration.json
credence:e4b-calibrated	Txoka/Credence-v1-Gemma4-E4B@7d5ffc8 calibrated/model-Q8_0.gguf	e4b-calibrated/decision.json, e4b-calibrated/calibration.json
credence:e4b-vision	Txoka/Credence-v1-Gemma4-E4B@7d5ffc8 accuracy/model-Q8_0.gguf + EldanRing/Winnow-E4B@734302f gguf/mmproj-Winnow-E4B.gguf	e4b-vision/decision.json, e4b-vision/calibration.json
credence:e4b-calibrated-vision	Txoka/Credence-v1-Gemma4-E4B@7d5ffc8 calibrated/model-Q8_0.gguf + EldanRing/Winnow-E4B@734302f gguf/mmproj-Winnow-E4B.gguf	e4b-calibrated-vision/decision.json, e4b-calibrated-vision/calibration.json

Each tag has decision.json (the prompt, the option labels Ollaya reads and llama.cpp's settings) and calibration.json (temperatures). A vision tag also pulls the vision projector the table names, unmodified from its author's repository, which reads the images.

Parity

Ollaya's runner matches stock llama-server of the pinned build (b11146) on the same GGUF and the same device: 505/505 decisions for both checkpoints, option logits within 7.6e-6, on the x86-64 CPU and an RTX 4070 (the contributor) and an RTX 4090 (Ollaya); the vision tags also match it on images.

License

Same as the upstream model (Apache-2.0). Ollaya itself is Apache-2.0.