Dataset Viewer
Auto-converted to Parquet Duplicate
dataset_name
string
input_root
string
output_path
string
split
string
levels
int64
history_len
int64
rows_total_seen
int64
stocks_used
int64
stock_ids
list
snapshots_used
int64
segments_used
int64
min_rows_per_segment
int64
segment_length_mean
float64
basic_metrics
dict
Optiver-L2
raw Optiver source archive omitted
data/optiver/optiver_train_8stocks_l2.npz
train
2
100
10,368,766
8
[ 0, 1, 2, 3, 4, 5, 6, 7 ]
10,329,143
30,291
132
340.997095
{ "spread_mean": 0.0007968809804879129, "spread_std": 0.0007564916741102934, "depth_mean": 551.6478881835938, "imb_mean": -0.006242518778890371, "imb_std": 0.4775751829147339 }

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

LoBiFlow Processed Dataset Bundle

This dataset repository contains the processed non-ES inputs used by the LoBiFlow public experiments. Download it into a source checkout with:

hf download pixelhero98/lobiflow --repo-type dataset --local-dir .

Expected paths after download:

  • data/cryptos/cryptos_binance_spot_monthly_1s_l10.npz
  • data/optiver/optiver_train_8stocks_l2.npz
  • data/synthetic/lobster_free_sample_profile_10.json

The raw Tardis and Optiver source archives are not included here. Sanitized summaries, manifest.json, and checksums.sha256 describe the processed files.

Downloads last month
43