datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
bitcoin-mempool-lifecycle
Bitcoin and Litecoin mempool lifecycle
Transaction observations covering first sight, time spent pending, confirmation and disappearance from monitored mempools. The repository also contains comparisons between Bitcoin provider views and a measure of how much of each block was visible before it arrived.
Contents
Table
Record
bitcoin_transaction_lifecycle
A Bitcoin transaction's observed lifecycle, fee information and outcome
bitcoin_mempool_comparison… See the full description on the dataset page: https://huggingface.co/datasets/dataforge-labs/bitcoin-mempool-lifecycle.ethereum-mempool
Ethereum mempool observations
Ethereum pending-transaction activity, lifecycle observations and comparisons between provider views. The continuing panels contain minute-level summaries, transactions classified as dropped and differences between observed pending sets.
Contents
Table
Record
ethereum_mempool_minute_summary
A minute of pending-transaction activity
ethereum_dropped_transactions
A transaction classified as dropped during observation… See the full description on the dataset page: https://huggingface.co/datasets/dataforge-labs/ethereum-mempool.mempool_solana_synthetic_training_data_spark_5k
Mempool Solana Synthetic Training Data (TsFile)
Apache TsFile version of
TitanOrionAI/Mempool-Solana-Synthetic-Training-Data-Spark-5k.
Overview
Synthetic Solana mempool simulation observations: 5,000 records from a
forward-looking mempool simulation on 2026-01-14, designed to train adaptive
models that predict congestion before it manifests on-chain. Each record has
the same shape as the companion MEV dataset (pair, regime, execution price,
Jito tip, alpha score… See the full description on the dataset page: https://huggingface.co/datasets/THULab/mempool_solana_synthetic_training_data_spark_5k.mempool-qwen-logits-orchestrator-rows
mempool Qwen Logits Orchestrator Rows
This dataset contains measured task-level orchestration training rows for the
mempool Qwen-small logits-head orchestrator path.
Train rows: 53
Train data file: 20260628-qwen-small-logits-orchestrator-split-train.jsonl
Held-out rows: 13
Held-out data file: 20260628-qwen-small-logits-orchestrator-split-heldout.jsonl
Split manifest: 20260628-qwen-small-logits-orchestrator-split-manifest.json
Targets: worker distribution, workflow distribution… See the full description on the dataset page: https://huggingface.co/datasets/blazeofchi/mempool-qwen-logits-orchestrator-rows.