Dataset Viewer
Auto-converted to Parquet Duplicate
feat_id
stringlengths
1
4
tokens
sequence
feat_pos_tags
sequence
feat_chunk_tags
sequence
tags
sequence
evaluation_predictions
sequence
3068
[ "Division", "three" ]
[ 20, 10 ]
[ 5, 16 ]
[ 8, 8 ]
[ [ 8.2265625, -1.455078125, -1.841796875, -0.8857421875, -1.4599609375, -0.86376953125, -2.48828125, -0.80126953125, -1.4267578125 ], [ 8.328125, -1.28515625, -2.853515625, -0.305419921875, -2.177734375, -1.41015625, -3.408203125, 0.9033203125, ...
2369
[ "Nippon", "Telegraph", "and", "Telephone", "Corp", "(", "NTT", ")", "said", "on", "Friday", "that", "it", "hopes", "to", "move", "into", "the", "international", "telecommunications", "business", "as", "soon", "as", "possible", "following", "the", "government", ...
[ 21, 21, 9, 21, 21, 4, 21, 5, 37, 14, 21, 14, 27, 41, 34, 36, 14, 11, 15, 23, 20, 29, 29, 14, 15, 20, 11, 20, 26, 20, 34, 36, 21, 14, 10, 23, 14, 11, 38, 20, 7 ]
[ 5, 16, 16, 16, 16, 22, 5, 22, 10, 6, 5, 8, 5, 10, 21, 21, 6, 5, 16, 16, 16, 1, 12, 6, 5, 16, 5, 16, 5, 16, 10, 21, 5, 6, 5, 16, 6, 5, 16, 16, 22 ]
[ 2, 6, 6, 6, 6, 8, 2, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2, 8, 8, 8, 8, 8, 8, 8, 8 ]
[ [ 8.59375, -1.5, -2.068359375, -0.888671875, -0.7919921875, -0.7763671875, -2.33984375, -0.822265625, -1.9560546875 ], [ -0.72900390625, -1.26171875, -2.525390625, 7.79296875, -0.412841796875, 0.288818359375, -2.09375, -0.043365478515625, -...
1103
[ "Canadian", "West", "Coast", "Vessel", "Loadings", "-", "CWB", "." ]
[ 21, 21, 21, 21, 21, 8, 21, 7 ]
[ 5, 16, 16, 16, 16, 22, 5, 22 ]
[ 1, 8, 8, 8, 8, 8, 8, 8 ]
[ [ 8.0625, -1.8359375, -2.45703125, -0.8955078125, -1.0966796875, 0.01438140869140625, -2.40625, 0.236083984375, -1.8779296875 ], [ -0.1068115234375, -1.7138671875, -2.291015625, -0.80712890625, -2.359375, 3.126953125, -1.794921875, 6.28125, ...
2562
[ "30.", "Alessandra", "Merlin", "(", "Italy", ")", "1:51.16" ]
[ 21, 21, 21, 4, 21, 5, 10 ]
[ 5, 16, 16, 22, 5, 22, 5 ]
[ 8, 3, 7, 8, 0, 8, 8 ]
[[9.7734375,-0.9287109375,-1.30078125,-1.7294921875,-1.703125,-1.2333984375,-2.236328125,-1.11328125(...TRUNCATED)
1883
[ "Hindu", "party", "forces", "India", "parliament", "to", "adjourn", "." ]
[ 21, 20, 41, 21, 20, 34, 36, 7 ]
[ 5, 16, 10, 5, 16, 10, 21, 22 ]
[ 1, 8, 8, 0, 8, 8, 8, 8 ]
[[8.921875,-1.5751953125,-2.15234375,-1.05078125,-1.29296875,-0.8857421875,-2.50390625,-0.1228027343(...TRUNCATED)
3358
[ "SATURDAY", ",", "DECEMBER", "7", "SCHEDULE" ]
[ 21, 6, 20, 10, 21 ]
[ 5, 22, 5, 16, 16 ]
[ 8, 8, 8, 8, 8 ]
[[9.171875,-1.681640625,-1.7451171875,-1.2607421875,-1.10546875,-1.1015625,-2.142578125,-1.122070312(...TRUNCATED)
768
[ "against", ",", "points", ")", ":" ]
[ 14, 6, 41, 5, 8 ]
[ 6, 22, 10, 22, 22 ]
[ 8, 8, 8, 8, 8 ]
[[9.46875,-1.3876953125,-1.0322265625,-1.7822265625,-1.5625,-1.10546875,-2.134765625,-0.6826171875,-(...TRUNCATED)
952
[ "It", "has", "produced", "1.5", "million", "hectolitres", "." ]
[ 27, 41, 39, 10, 10, 23, 7 ]
[ 5, 10, 21, 5, 16, 16, 22 ]
[ 8, 8, 8, 8, 8, 8, 8 ]
[[8.375,-1.560546875,-2.376953125,-0.80712890625,-1.43359375,-0.479736328125,-2.208984375,-0.2558593(...TRUNCATED)
2522
[ "4.", "Italy", "887" ]
[ 10, 21, 10 ]
[ 5, 16, 16 ]
[ 8, 0, 8 ]
[[9.4140625,-1.2607421875,-1.7236328125,-1.4267578125,-1.646484375,-0.78173828125,-2.451171875,-0.95(...TRUNCATED)
318
["Australia","-","Mark","Taylor","(","captain",")",",","Mark","Waugh",",","Ricky","Ponting",",","Gre(...TRUNCATED)
[21,8,21,21,4,20,5,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6,21,21,6(...TRUNCATED)
[5,22,5,16,22,5,22,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,22,5,16,2(...TRUNCATED)
[ 0, 8, 3, 7, 8, 8, 8, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 3, 7, 8, 8, 8 ]
[[8.8203125,-0.5986328125,-1.1962890625,-1.712890625,-1.666015625,-0.908203125,-2.498046875,-0.37084(...TRUNCATED)
End of preview. Expand in Data Studio

Dataset Card for AutoTrain Evaluator

This repository contains model predictions generated by AutoTrain for the following task and dataset:

  • Task: Token Classification
  • Model: 51la5/distilbert-base-NER
  • Dataset: conll2003
  • Config: conll2003
  • Split: test

To run new evaluation jobs, visit Hugging Face's automatic model evaluator.

Contributions

Thanks to @aniketrawat97 for evaluating this model.

Downloads last month
12