SakThai Combined v5 — Tool-Calling SFT Dataset

The primary training dataset for the SakThai Context model family — a curated collection of structured tool-calling and instruction-following examples for supervised fine-tuning (SFT).

Dataset Summary

Property	Value
Train examples	659
Test examples	51
Format	Chat template (messages) with tool schemas
Target models	Qwen2.5 (0.5B, 1.5B, 7B)
License	MIT

Structure

Each example follows the Qwen2.5 chat template format with system/user/assistant roles and optional tool call/result entries:

{
  "messages": [
    {"role": "system", "content": "..."},
    {"role": "user", "content": "..."},
    {"role": "assistant", "content": "...", "tool_calls": [...]},
    {"role": "tool", "content": "...", "tool_call_id": "..."}
  ],
  "tools": [...]
}

Usage

from datasets import load_dataset

dataset = load_dataset("Nanthasit/sakthai-combined-v5")
print(dataset["train"][0])

Models Trained on This Dataset

Model	Size	Status
sakthai-context-7b-tools	7B	✅ Latest
sakthai-context-1.5b-tools	1.5B	✅ Latest

Note: Earlier versions (v1–v4) are superseded by v5. They remain available for reproducibility but new training should use v5.

References

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support