datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SYNTH-Swallow-Math-Code-Mix
Mixed dataset: SYNTH + SwallowMath-v2 + SwallowCode-v2
This high-signal, all synthetic dataset is a complete shuffled mix of the following four sources:
SYNTH ~63.5%
SwallowCode-v2 ~15.5%
SwallowMath-v2-textbook ~10.5%
SwallowMath-v2-qa ~10.0%
The motivation to provide this on HF was the need for a convenient, pre-shuffled merge of the highest quality synthetic / augmented datasets for small language model pre-training experiments as of… See the full description on the dataset page: https://huggingface.co/datasets/TMoC/SYNTH-Swallow-Math-Code-Mix.CodeMixBench
ℹ️Dataset Card for CodeMixBench
[EMNLP'25] CodeMixBench: Evaluating Code-Mixing Capabilities of LLMs Across 18 Languages
Code-mixing is a linguistic phenomenon where multilingual speakers switch or mix two or more languages within a single utterance or conversation.
To evaluate LLMs’ comprehension of multilingual code-mixed texts, we introduce CodeMixBench, a benchmark comprising eight tasks across 18 languages.… See the full description on the dataset page: https://huggingface.co/datasets/CodeMixBench/CodeMixBench.mopd-math-code-mix
MOPD math+code mix
`train/`: math:code ≈ 1:1 平衡集（`math.parquet` + `code_*.parquet` shards）
`val/mopd_val_mix.parquet`: AIME24 全量 + MATH-500 子集 + Eurus code_validation 子集
路由字段：`ability ∈ {math, code}`
CodeMixBench
ℹ️Dataset Card for CodeMixBench
[EMNLP'25] CodeMixBench: Evaluating Code-Mixing Capabilities of LLMs Across 18 Languages
Code-mixing is a linguistic phenomenon where multilingual speakers switch or mix two or more languages within a single utterance or conversation.
To evaluate LLMs’ comprehension of multilingual code-mixed texts, we introduce CodeMixBench, a benchmark comprising eight tasks across 18 languages.… See the full description on the dataset page: https://huggingface.co/datasets/Tanushreeeeee/CodeMixBench.mix60k-math-code-sft
mix60k: 50/50 OpenMathInstruct-2 + OpenCodeInstruct SFT mixture
This dataset is from the paper Register Tokens for Bounded-State Reasoning in Diffusion Language Models.
The exact 60,000-trace SFT mixture used to train the LLaDA-8B-Base and Dream-7B-Base
main triad in the dLLM Registers project.
Composition: ~30K math traces from OpenMathInstruct-2 and ~30K code traces from OpenCodeInstruct.
License: governed by the upstream licenses of OpenMathInstruct-2 and OpenCodeInstruct.… See the full description on the dataset page: https://huggingface.co/datasets/albertge/mix60k-math-code-sft.code-instruct-mixed
Description
Filtered/normalised subsets of public code-instruction datasets (Magicoder OSS-Instruct & Evol-Instruct, CodeFeedback, Glaive). The source column attributes each row to its origin; each source retains its upstream licence.
Derived dataset. Source material retains its original per-item licence (see source/repo columns); treat as other / mixed. Provided as-is.
Usage
from datasets import load_dataset
ds = load_dataset("PotatoHD/code-instruct-mixed")
agentic-code-sft-mix-v1
Agentic Code SFT Mix v1
Local derived SFT mixture for code-agent/tool-use training.
This is not a single upstream dataset. It is a filtered local mixture built from:
nvidia/OpenCodeInstruct, split train
nvidia/Nemotron-SFT-OpenCode-v1, splits general, bash_only_tool, bash_only_tool_skills, question_tool, agent_skills, agent_skills_question_tool
nvidia/Nemotron-SFT-SWE-v2, split agentless
nvidia/Nemotron-SFT-SWE-v2, file data/swe.jsonl
The output schema is JSONL with messages… See the full description on the dataset page: https://huggingface.co/datasets/synquid/agentic-code-sft-mix-v1.Burmese-English-Code-Mixed-Corpus
🇲🇲 Burmese-English Code-Mixed Corpus ꒰ 1,111 Rows ꒱
A high-quality, human-curated dataset of code-mixed Burmese and English sentences, specifically designed for Natural Language Processing (NLP) and Machine Learning (ML) research.
Dataset Details
Organization: DatarrX
Creator: Khant Sint Heinn (Kalix Louis)
Number of Rows: 1,111
Language: Burmese (Unicode) & English Mix
Dataset Format: .txt
License: Apache 2.0
Description
The Burmese-English Code-Mixed… See the full description on the dataset page: https://huggingface.co/datasets/DatarrX/Burmese-English-Code-Mixed-Corpus.Mixture-Code
RecursiveMAS Mixture-Code
Project Page | Code | Paper
We introduce RecursiveMAS, a multi-agent framework that scales agent collaboration through latent-space recursion. This dataset contains training examples for the Mixture-Style setting.
Dataset Details
Item
Description
Dataset
RecursiveMAS/Mixture-Code
Original file
Mixture-Code.json
Collaboration style
Mixture-Style
Used for
code specialist inner agent training
Split
train
Rows
2000… See the full description on the dataset page: https://huggingface.co/datasets/RecursiveMAS/Mixture-Code.v3-v1_v2_code_mixed_syntheic_correct_noisy_pairs
Sinhala Spelling Correction Dataset
Dataset Description
This dataset contains Sinhala text pairs for training spelling correction models. It includes:
Dyslexic/Noisy sentences: Text with spelling errors, typos, and dyslexia-like mistakes
Clean sentences: Corrected versions of the text
Dataset Statistics
Split
Samples
Train
37,712
Test
9,428
Total
47,140
Features
dyslexic_sentence: Input text with errors (string)… See the full description on the dataset page: https://huggingface.co/datasets/SPEAK-PP/v3-v1_v2_code_mixed_syntheic_correct_noisy_pairs.Burmese-English-Code-Mixed-Corpus
🇲🇲 Burmese-English Code-Mixed Corpus ꒰ 1,111 Rows ꒱
A high-quality, human-curated dataset of code-mixed Burmese and English sentences, specifically designed for Natural Language Processing (NLP) and Machine Learning (ML) research.
Dataset Details
Organization: DatarrX
Creator: Khant Sint Heinn (Kalix Louis)
Number of Rows: 1,111
Language: Burmese (Unicode) & English Mix
Dataset Format: .txt
License: Apache 2.0
Description
The Burmese-English Code-Mixed… See the full description on the dataset page: https://huggingface.co/datasets/hksamm/Burmese-English-Code-Mixed-Corpus.telugu-qa-codemixed
Telugu QA Paraphrases
A synthetic multilingual query-rewriting dataset for evaluating retrieval robustness under Telugu-English code mixing.
Dataset Description
This dataset extends an existing Telugu QA dataset by generating multiple query variants with increasing levels of Telugu-English code mixing.
Each example contains:
question : Original English question
answer : Ground-truth answer
level_0 : English paraphrase
level_1 : Light Telugu-English code mixing… See the full description on the dataset page: https://huggingface.co/datasets/nlpctx/telugu-qa-codemixed.CodeMix
CodeMix - a small finetune dataset
6k chat/response pairs, a balanced mix of:
glaive-function-calling-v2 (agentic tool calling)
hermes-function-calling-v1 (tool calling)
CodeAlpaca-20k (coding)
dolly-15k (instruct)
marathi-codemix-qa
Marathi Minglish QA
~1.09M synthetic Question–Answer pairs in code-mixed Romanized Marathi (Minglish), generated from Marathi Wikipedia articles.
Designed for pretraining and SFT of Marathi-aware Small Language Models that should understand and generate the way Marathi is commonly written online — Roman-script Marathi naturally mixed with English terms.
Example
Question:
Yashwant Dev kon hote exactly — sangeetkar, kavi, ki donhi?
Answer:
Yashwant Dev he… See the full description on the dataset page: https://huggingface.co/datasets/atx-labs/marathi-codemix-qa.