-
Model card: ollama pull iapp/openthai-systemone
-
Model card: Ollama build (iapp/OpenThai-SystemOne-Ollama) and Ollama 0.35 evaluation
-
model card: Thai showcase poster first, Doom demo after
-
assets: Thai showcase poster
-
model card: Thai showcase with real v0.3 outputs
-
v0.3: +5k SFT steps with weak-spot data (177k real + 78k targeted synthetic), recalibrated (public macro 63.2 -> 74.3)
-
model card: Doom demo section
-
assets: Doom demo GIF
-
model card: English launch image
-
assets: English launch image
-
model card: order-invariant inference mode (fix for option-order bias)
-
model card: v0.2 changelog + known order-sensitivity issue
-
v0.2: +3k SFT steps with Thai social-sentiment set, recalibrated (public macro 61.9 -> 63.2, Wisesight 38.7 -> 51.5)
-
model card: sponsor
-
model card: sponsor
-
model card: launch image
-
model card: launch image
-
credits: OpenThai
-
OpenThai-SystemOne release
-
initial commit
