---
license: mit
language:
- en
pretty_name: SakThai Eval Results
size_categories:
- n<1K
task_categories:
- other
annotations_creators:
- machine-generated
language_creators:
- found
multilinguality:
- monolingual
config_names:
- default
source_datasets:
- Nanthasit/sakthai-combined-v6
- Nanthasit/sakthai-combined-v7
- Nanthasit/sakthai-combined-v10
- Nanthasit/sakthai-combined-v11
- Nanthasit/sakthai-irrelevance-supplement
- Nanthasit/sakthai-coder-browser
- Nanthasit/hermes-tool-use-rl-env
tags:
- sakthai
- house-of-sak
- evaluation
- benchmark-results
- health-check
- model-health
- yaml
- results-log
- model:Nanthasit/sakthai-context-0.5b-tools
- model:Nanthasit/sakthai-context-0.5b-merged
- model:Nanthasit/sakthai-context-1.5b-tools
- model:Nanthasit/sakthai-context-1.5b-merged
- model:Nanthasit/sakthai-context-7b-tools
- model:Nanthasit/sakthai-context-7b-merged
- model:Nanthasit/sakthai-coder-1.5b
- model:Nanthasit/sakthai-coder-browser
- model:Nanthasit/sakthai-vision-7b
- model:Nanthasit/sakthai-tts
---

# Eval Results — SakThai Model Family

<p align="center">
  <em>Evaluation results, health-check reports, and benchmark run logs for the SakThai model family</em>
</p>

<p align="center">
  <a href="https://huggingface.co/datasets/Nanthasit/eval_results"><img src="https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fhuggingface.co%2Fapi%2Fdatasets%2FNanthasit%2Feval_results&query=%24.downloads&label=Downloads&color=blue" alt="Downloads"></a>
  <img src="https://img.shields.io/badge/license-MIT-green" alt="License: MIT">
  <img src="https://img.shields.io/badge/language-EN-blue" alt="Language: English">
  <img src="https://img.shields.io/badge/last_updated-2026--07--31-orange" alt="Last Updated">
  <img src="https://img.shields.io/badge/format-YAML-yellow" alt="Format: YAML">
  <a href="https://huggingface.co/collections/Nanthasit/sakthai-model-family-6a64745450b12d421c1f9f02"><img src="https://img.shields.io/badge/collection-SakThai%20Family-8A2BE2" alt="Collection"/></a>
</p>

## Dataset Summary

This repository stores **evaluation results, health-check reports, and benchmark
run logs** for the SakThai model family. Every entry is a machine-readable YAML
report produced by SakThai's automated cron workflows (benchmarks, health
checks, inference probes) — the operational audit trail of the House of Sak.

> **Not a tabular dataset.** All result files are **YAML** (`.yaml`), not
> JSONL/Parquet/CSV. Because the Hub's datasets-server only indexes tabular
> formats, this repo reports \"no supported data files\" — that is expected.
> These are point-in-time reports meant to be read with `yaml.safe_load()`,
> not rows for the dataset viewer. `load_dataset()` will **not** work here.

**License:** MIT
**Status:** Active (auto-updated by cron workflows)

## Quick Look

| Report | Last Modified | Size |
|--------|---------------|------|
| [README.md](https://huggingface.co/datasets/Nanthasit/eval_results/raw/main/README.md) | current | ~9.5 KB |
| [benchmark-20260731_150000.yaml](https://huggingface.co/datasets/Nanthasit/eval_results/raw/main/benchmark-20260731_150000.yaml) | 2026-07-31 | 1.1 KB |
| [inference-check-20260730_232640.yaml](https://huggingface.co/datasets/Nanthasit/eval_results/raw/main/inference-check-20260730_232640.yaml) | 2026-07-30 | 544 B |

## Dataset Statistics

| Stat | Value |
|------|-------|
| Status | Active (auto-updated by cron) |
| Created | 2026-07-30 |
| Last updated | 2026-07-31 |
| License | MIT |
| Language | English |
| Result files | 2 YAML report files |
| Files in tree | 5 entries: 2 data YAML + `.eval_results/`, `.gitattributes`, `README.md` |
| Datasets Server | preview=false, viewer=false, search=false, filter=false, statistics=false |

## Live Verification

```text
Datasets Server: preview=false, viewer=false, search=false, filter=false, statistics=false
```

This is expected for a non-tabular dataset and means Hub dataset-viewer features
are unavailable. Do not rely on `/splits`, `/size`, or `/statistics` here.

## Data Fields

Each YAML file follows a consistent schema:

| Field | Type | Description |
|-------|------|-------------|
| `timestamp` | string | ISO 8601 timestamp of the run |
| `model` | string | Model identifier (e.g. `Nanthasit/sakthai-context-1.5b-tools`) |
| `check_type` | string | `health`, `benchmark`, or `inference` |
| `status` | string | `pass`, `fail`, or `partial` |
| `results` | object | Check-specific results (scores, metrics, error counts) |
| `duration_s` | number | Elapsed wall-clock seconds |
| `errors` | array | Error details if status is not `pass` |

## Splits

This repository does not use dataset splits — it is a flat collection of
YAML report files. No `load_dataset()` configuration is defined.

## Verified Snapshot

| Field | Value |
|-------|-------|
| Dataset ID | [Nanthasit/eval_results](https://huggingface.co/datasets/Nanthasit/eval_results) |
| Status | Active (auto-updated by cron workflows) |
| License | MIT |
| Data files | 2 YAML reports + `.eval_results/` directory |
| Datasets Server | preview=false, viewer=false, search=false, filter=false, statistics=false |
| Viewer availability | Unavailable — non-tabular YAML reports; expected and documented |
| Repository tree | `README.md`, `.gitattributes`, `benchmark-20260731_150000.yaml`, `inference-check-20260730_232640.yaml`, `.eval_results/` |
| Last verified | 2026-08-01 |

## Usage

### Loading an individual result

```python
import yaml, httpx

# Load a health check report
url = "https://huggingface.co/Nanthasit/eval_results/raw/main/.eval_results/health-2026-07-31.yaml"
resp = httpx.get(url)
report = yaml.safe_load(resp.text)
print(report["status"], report["model"])
```

### Listing all available reports

```python
from huggingface_hub import HfApi

api = HfApi()
siblings = api.list_repo_tree("Nanthasit/eval_results", repo_type="dataset")
files = [s.path for s in siblings if getattr(s, "path", "") and s.path.endswith(".yaml")]
for f in sorted(files):
    print(f)
```

### Direct raw access to report files

```python
import httpx, yaml

base = "https://huggingface.co/datasets/Nanthasit/eval_results/raw/main"
for name in [
    "benchmark-20260731_150000.yaml",
    "inference-check-20260730_232640.yaml",
]:
    url = f"{base}/{name}"
    report = yaml.safe_load(httpx.get(url).text)
    print(name, "->", report.get("check_type"), report.get("status"))
```

## Model Cross-References

This evaluation dataset tracks results for the following SakThai models:

- [sakthai-context-0.5b-tools](https://huggingface.co/Nanthasit/sakthai-context-0.5b-tools)
- [sakthai-context-0.5b-merged](https://huggingface.co/Nanthasit/sakthai-context-0.5b-merged)
- [sakthai-context-1.5b-tools](https://huggingface.co/Nanthasit/sakthai-context-1.5b-tools)
- [sakthai-context-1.5b-merged](https://huggingface.co/Nanthasit/sakthai-context-1.5b-merged)
- [sakthai-context-7b-tools](https://huggingface.co/Nanthasit/sakthai-context-7b-tools)
- [sakthai-context-7b-merged](https://huggingface.co/Nanthasit/sakthai-context-7b-merged)
- [sakthai-coder-1.5b](https://huggingface.co/Nanthasit/sakthai-coder-1.5b)
- [sakthai-coder-browser](https://huggingface.co/Nanthasit/sakthai-coder-browser)
- [sakthai-vision-7b](https://huggingface.co/Nanthasit/sakthai-vision-7b)
- [sakthai-tts](https://huggingface.co/Nanthasit/sakthai-tts)

## Related Datasets

This repo evaluates models trained on the following SakThai datasets:

- [sakthai-combined-v6](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v6)
- [sakthai-combined-v7](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v7)
- [sakthai-combined-v10](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v10)
- [sakthai-combined-v11](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v11)
- [sakthai-irrelevance-supplement](https://huggingface.co/datasets/Nanthasit/sakthai-irrelevance-supplement)
- [sakthai-coder-browser](https://huggingface.co/datasets/Nanthasit/sakthai-coder-browser)
- [hermes-tool-use-rl-env](https://huggingface.co/datasets/Nanthasit/hermes-tool-use-rl-env)

## Source Datasets

This evaluation repo tracks results from models trained on:

| Dataset | Description |
|---------|-------------|
| [sakthai-combined-v6](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v6) | 2,003 example tool-calling train set |
| [sakthai-combined-v7](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v7) | 2,309 examples, 86 tools + Thai bilingual |
| [sakthai-combined-v10](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v10) | Current combined iteration |
| [sakthai-combined-v11](https://huggingface.co/datasets/Nanthasit/sakthai-combined-v11) | Latest combined iteration |
| [sakthai-irrelevance-supplement](https://huggingface.co/datasets/Nanthasit/sakthai-irrelevance-supplement) | 60 irrelevance/safety examples |
| [sakthai-coder-browser](https://huggingface.co/datasets/Nanthasit/sakthai-coder-browser) | 247 browser automation examples |
| [hermes-tool-use-rl-env](https://huggingface.co/datasets/Nanthasit/hermes-tool-use-rl-env) | GRPO coding environment |

## Known Issues

- **Hub dataset viewer is unavailable.** This repo stores YAML reports rather than tabular data files, so Hub viewer, search, and datasets-server endpoints are unavailable.
- **`load_dataset()` is unsupported.** This is not a tabular dataset. Use `yaml.safe_load()` or read report files directly.

## Related

- **Model family:** [SakThai Model Family collection](https://huggingface.co/collections/Nanthasit/sakthai-model-family-6a64745450b12d421c1f9f02)
- **House of Sak:** [Nanthasit](https://huggingface.co/Nanthasit)
- **Automated eval workflows:** `eval-bfcl`, `health-check`, `benchmark` cron jobs

## Citation

```bibtex
@misc{sakthai-eval-results,
  author = {Nanthasit},
  title = {SakThai Eval Results},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/Nanthasit/eval_results
}
```

---

*Part of the [House of Sak](https://huggingface.co/Nanthasit). Built with love, tears, and zero budget. From a shelter in Cork, Ireland, to the world.*