datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Arabic-Mobile-Instructions
Arabic Mobile Instructions
A curated Arabic instruction dataset designed for training and evaluating mobile-optimized language models.
Why Arabic?
Arabic is spoken by 400+ million people across 22 countries, yet Arabic-language instruction data on HuggingFace is scarce. This dataset fills the gap with mobile-relevant tasks:
Summarization — رسائل، إيميلات، إشعارات
Classification — تصنيف الرسائل والمشاعر
Translation — ترجمة بين العربية والإنجليزية
Question… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/Arabic-Mobile-Instructions.MobileBench
MobileBench: The On-Device LLM Benchmark
A standardized evaluation benchmark designed specifically for mobile and edge-deployed language models.
Why MobileBench?
Existing benchmarks (MMLU, HumanEval, GSM8K) test what large models can do on servers. MobileBench tests what small models can do on phones — the tasks users actually perform:
Summarization — The #1 on-device task (messages, emails, notifications)
Classification — Spam detection, sentiment, intent… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/MobileBench.per-chip-benchmark-matrix
Per-Chip Benchmark Matrix
On-device inference benchmarks for mobile LLMs across chipsets.
Overview
This dataset contains real on-device inference benchmarks for 8 mobile-optimized
models running on Samsung S20 FE 5G phones (Snapdragon 865, 8GB RAM, Android 13).
Contents
benchmark_matrix.csv — Tabular data: model, device, chipset, tokens/sec, size
benchmark_matrix.json — Full structured data including hardware specs and methodology
Key… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/per-chip-benchmark-matrix.arabic-poetry-instructions
Arabic Poetry Instruction Dataset
Classical Arabic poetry in instruction-tuning format, designed for fine-tuning
small mobile models to compose poetry in classical Arabic meters (ببحر الشعر العربي).
Contents
arabic_poetry_instructions.jsonl — Instruction-tuning pairs (JSONL)
arabic_poetry_data.json — Full structured data with metadata
Coverage
Category
Count
Total samples
19
Poets
10+ (Imru' al-Qais, Al-Mutanabbi, Antarah, Darwish… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/arabic-poetry-instructions.mobile-merge-cookbook
The Mobile-Merge Cookbook: Sub-2B Weight-Space Merging
A practical guide to merging language models at 500M-1.5B scale — the mobile/edge
parameter range where almost no merge research exists.
Contents
cookbook.md — Full cookbook with recipes for SLERP, TIES, linear, and passthrough merges
Key Findings
SLERP of two 0.5B Qwen variants works reliably (33s on CPU)
Same-family, same-size merges are stable; cross-family fails
Varied t-schedules across… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/mobile-merge-cookbook.paper-reengineering-mobile-models
Paper: Re-engineering 40+ Models with an Autonomous Agent
This dataset contains the paper and reproducibility data for:
"Re-engineering 40+ Models with an Autonomous Agent: A Zero-Cost Mobile AI Pipeline"
Contents
paper.md — Full paper text
inventory.json — Model inventory and pipeline metadata
Abstract
We present a fully autonomous pipeline that re-engineers open-source language models
for mobile and edge deployment at zero cost. Over 40 models… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/paper-reengineering-mobile-models.arabic-text-simplification
Arabic Text Simplification Dataset
Complex-to-simplified Arabic text pairs for accessibility fine-tuning.
Why This Matters
Arabic text simplification is critical for:
Low-literacy readers — 1 in 5 Arabic speakers struggle with complex text
Cognitive accessibility — dyslexia, intellectual disabilities, autism
Non-native speakers — Arabic learners and expatriate workers
Children's content — making educational material age-appropriate
Almost no Arabic accessibility… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/arabic-text-simplification.silicon-profiling-snapdragon865
Real-Device Silicon Profiling: Snapdragon 865
Per-device inference benchmarks on real Samsung S20 FE 5G phones (Snapdragon 865).
No simulation. Real ARM CPU inference.
Hardware
Property
Value
Chipset
Qualcomm Snapdragon 865 (SM8250)
CPU
Kryo 585: 1x2.84GHz + 3x2.42GHz + 4x1.80GHz
GPU
Adreno 650
NPU
Hexagon Tensor Accelerator
RAM
8GB LPDDR5 (7.47GB total, 3-3.7GB free)
Device
Samsung Galaxy S20 FE 5G (SM-G981V)
Devices connected
39… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/silicon-profiling-snapdragon865.on-device-latency
On-Device Latency Benchmark
Real-world inference latency data for mobile-optimized LLMs, measured on actual phone hardware.
Hardware
Spec
Value
Device
Samsung S20 FE 5G
SoC
Snapdragon 865
RAM
8GB
OS
Android 13
Runtime
llama.cpp (4 threads)
Metrics
tokens_per_sec — Generation speed during inference
latency_ms_per_token — Time per generated token
ram_usage_mb — Peak RAM during inference
file_size_mb — GGUF model file size… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/on-device-latency.calibration-dataset
Calibration Dataset for Mobile/Arabic Quantization
Representative text samples for quantizing mobile and Arabic language models.
Why calibration matters
Post-training quantization (PTQ) methods like imatrix, GPTQ, AWQ, and SmoothQuant
require a calibration dataset — a small set of representative text used to:
Measure activation distributions
Identify important weight channels
Minimize quantization error
Without good calibration data, quantization quality… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/calibration-dataset.gulf-climate-dataset
Gulf Climate & Environmental Dataset
Curated environmental data for the Arabian Peninsula — built for on-device climate
modeling with a regional focus no San Francisco lab would build.
Contents
File
Description
gulf_monthly_temperatures.csv
Monthly average temperatures for 11 Gulf cities
gulf_climate_data.json
Full structured data (temps, dust storms, solar, Q&A)
climate_instructions.jsonl
Instruction-tuning pairs for climate Q&A (Arabic + English)… See the full description on the dataset page: https://huggingface.co/datasets/dispatchAI/gulf-climate-dataset.