Hugging Face's logo

Datasets:
AMAImedia
/
kazakh-instruction-v2

Modalities:
Text
Formats:
json
Languages:
English
Russian
Chinese
Size:
10K - 100K
Tags:
tencent
Hy4-preview
Hy4
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
kazakh-instruction-v2 / README.md
AMAImedia's picture
AMAImedia
docs: Release date -> 2026-09-18
b2e2252 verified
preview code
|
Raw History Blame Contribute Delete
4.2 kB
metadata 
license: apache-2.0
library_name: transformers
pipeline_tag: image-text-to-text
language:
  - en
  - ru
  - zh
  - vi
  - kk
  - ja
  - af
  - am
  - ar
  - as
  - ast
  - az
  - be
  - bg
  - bn
  - bs
  - ca
  - ceb
  - ckb
  - cs
  - cy
  - da
  - de
  - el
  - es
  - et
  - eu
  - fa
  - ff
  - fi
  - fil
  - fr
  - ga
  - gl
  - gn
  - gu
  - ha
  - he
  - hi
  - hr
  - hu
  - hy
  - id
  - ig
  - is
  - it
  - jv
  - ka
  - kam
  - kea
  - km
  - kmr
  - kn
  - ko
  - ky
  - lb
  - lg
  - ln
  - lo
  - lt
  - luo
  - lv
  - mi
  - mk
  - ml
  - mn
  - mr
  - ms
  - mt
  - mvy
  - my
  - ne
  - nl
  - 'no'
  - nso
  - ny
  - oc
  - om
  - or
  - pa
  - pl
  - ps
  - pt
  - qxp
  - ro
  - rw
  - sd
  - sk
  - skr
  - sl
  - sn
  - so
  - sr
  - sv
  - sw
  - ta
  - te
  - tg
  - th
  - ti
  - tk
  - tr
  - ug
  - uk
  - umb
  - ur
  - uz
  - wo
  - xh
  - yo
  - yue
  - zu
base_model:
  - tencent/Hy4-preview
  - AngelSlim/Hy4-preview-GGUF
tags:
  - tencent
  - Hy4-preview
  - Hy4

⚡ Each donation funds the next large quant.

I host free GGUF or MoE quants as independent research.
Local hardware: Mechrevo Kuangshi GM7AG0M — RTX 3060 Laptop 6GB GDDR6, 64GB DDR5, i7-12700H (14C/20T, 4.7GHz), Windows 11, Samsung 990 Pro.
Good for imatrix and 0.6–35B-class work in RAM. 9B+ and searches need rented H200/Blackwell, typically $100 per quant.

🎉 Boosty🦄  |  ☕ Buy Me a Coffee🦄  |  ⭐ DonationAlerts🦄

💚 Thanks to Hugging Face for extra storage.🦄

NOESIS / AMAImedia

Released as part of the NOESIS Professional Multilingual Dubbing Automation Platform (framework: DHCF-FNO — Deterministic Hybrid Control Framework for Frozen Neural Operators).

Dataset Card for Dataset Name

Self-instruct data pairs for Kazakh language

Dataset Details

The dataset is translated from Standford Alpaca instruction dataset via Google Translations API.

  1. Manually fixed the translation error.
  2. Common names and places of Kazakhstan were added.
  3. Intructions of kazakhstan history and cultures were added.

Dataset Description

  • Curated by: Mussa Aman
  • Language(s) (NLP): Kazakh
  • License: MIT

Uses

This dataset is curated to fine-tune the LLaMA 2 model for the Kazakh language. It aims to enhance the model's understanding and processing capabilities of Kazakh, addressing a gap in the Low Resource Lanuguages for solving the NLP resources for Kazakh language.

The dataset includes the self-instruct approach, there is commonly one "instruction","input" and "output" which is crucial for improving language comprehension and task performance of the model.

Citation

BibTeX:

@inproceedings{mussa2025parameter, title={Parameter-Efficient Fine-Tuning of LLaMA 2 for the Kazakh Language: Advancing Low-Resource Language Models}, author={Mussa, Aman and Mansurova, Madina}, booktitle={International Congress on Information and Communication Technology}, pages={511--520}, year={2025}, organization={Springer} } APA:

Mussa, A., & Mansurova, M. (2025, February). Parameter-Efficient Fine-Tuning of LLaMA 2 for the Kazakh Language: Advancing Low-Resource Language Models. In International Congress on Information and Communication Technology (pp. 511-520). Singapore: Springer Nature Singapore.

Dataset Card Contact

Please contact in email: mussa.aman@kaznu.kz