Text Classification
MLX
Safetensors
Thai
English
qwen3_5
system-one
decision-model
thai
qwen3.5
quantized
apple-silicon
8-bit precision
Instructions to use iapp/OpenThai-SystemOne-MLX-8bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use iapp/OpenThai-SystemOne-MLX-8bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download iapp/OpenThai-SystemOne-MLX-8bit --local-dir OpenThai-SystemOne-MLX-8bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Brand line: "Developed by iApp, launched with AIEAT. Part of OpenThai (formerly OpenThaiGPT)."
Browse files
README.md
CHANGED
|
@@ -10,6 +10,8 @@ tags: [system-one, decision-model, thai, qwen3.5, quantized, mlx, apple-silicon]
|
|
| 10 |
|
| 11 |
# OpenThai-SystemOne — mlx-8bit
|
| 12 |
|
|
|
|
|
|
|
| 13 |
[OpenThai-SystemOne](https://huggingface.co/iapp/OpenThai-SystemOne) is an open Thai + English **System One decision model**: one forward
|
| 14 |
pass answers typed questions (`choice` over up to 255 options, ordinal `score`, yes/no `noul`) about a text / JSON state with
|
| 15 |
calibrated probabilities, no text generation. It is a Qwen3.5-0.8B text tower (Thai continued pre-training) plus a 256-slot
|
|
@@ -80,4 +82,4 @@ Macro: public 74.2 (original 74.3), Thai 80.0 (original 80.1).
|
|
| 80 |
- Scores are single-option-order accuracy on the first 800 records of each set (`scripts/06_eval.py --limit 800`), the same
|
| 81 |
records for the original and the quantization. `score` subsets report exact level accuracy.
|
| 82 |
- Base model, data, training and the full benchmark tables: [iapp/OpenThai-SystemOne](https://huggingface.co/iapp/OpenThai-SystemOne).
|
| 83 |
-
- License Apache-2.0 (same as the base).
|
|
|
|
| 10 |
|
| 11 |
# OpenThai-SystemOne — mlx-8bit
|
| 12 |
|
| 13 |
+
Developed by iApp, launched with AIEAT. Part of OpenThai (formerly OpenThaiGPT).
|
| 14 |
+
|
| 15 |
[OpenThai-SystemOne](https://huggingface.co/iapp/OpenThai-SystemOne) is an open Thai + English **System One decision model**: one forward
|
| 16 |
pass answers typed questions (`choice` over up to 255 options, ordinal `score`, yes/no `noul`) about a text / JSON state with
|
| 17 |
calibrated probabilities, no text generation. It is a Qwen3.5-0.8B text tower (Thai continued pre-training) plus a 256-slot
|
|
|
|
| 82 |
- Scores are single-option-order accuracy on the first 800 records of each set (`scripts/06_eval.py --limit 800`), the same
|
| 83 |
records for the original and the quantization. `score` subsets report exact level accuracy.
|
| 84 |
- Base model, data, training and the full benchmark tables: [iapp/OpenThai-SystemOne](https://huggingface.co/iapp/OpenThai-SystemOne).
|
| 85 |
+
- License Apache-2.0 (same as the base). Developed by iApp, launched with AIEAT.
|