Dataset Viewer
	| x
				 float64 -88.35 82 | y
				 float64 -86.35 88.7 | language
				 stringclasses 47
				values | corpus
				 stringclasses 222
				values | 
|---|---|---|---|
| -18.28238 | -2.315064 | 
	Marathi | 
	opus100 | 
| -64.574786 | -6.972784 | 
	Marathi | 
	opus100 | 
| -55.588644 | 0.740308 | 
	Marathi | 
	opus100 | 
| 30.174092 | 28.561195 | 
	Marathi | 
	pib | 
| -16.8444 | -0.429286 | 
	Marathi | 
	pib | 
| -13.992644 | 0.70737 | 
	Marathi | 
	pib | 
| -13.451096 | 17.680722 | 
	Marathi | 
	pib | 
| -67.07903 | -1.450883 | 
	Marathi | 
	pib | 
| 30.683829 | 27.999965 | 
	Marathi | 
	pib | 
| 2.293134 | -50.560417 | 
	Marathi | 
	pib | 
| -14.71315 | 0.378662 | 
	Marathi | 
	pib | 
| -14.262408 | -15.717909 | 
	Marathi | 
	pib | 
| 29.225542 | 30.075136 | 
	Marathi | 
	pib | 
| -72.345372 | -6.565703 | 
	Marathi | 
	pib | 
| -17.105466 | -0.358465 | 
	Marathi | 
	pib | 
| 30.709126 | 29.035129 | 
	Marathi | 
	pib | 
| 1.898391 | -53.309739 | 
	Marathi | 
	pib | 
| -13.889551 | 0.790806 | 
	Marathi | 
	pib | 
| 2.152298 | -53.526555 | 
	Marathi | 
	pib | 
| 2.167264 | -51.004306 | 
	Marathi | 
	pib | 
| -57.258285 | -1.045722 | 
	Marathi | 
	pib | 
| -56.311582 | -1.515933 | 
	Marathi | 
	pib | 
| -61.945088 | 4.567302 | 
	Marathi | 
	pib | 
| -14.361234 | 0.436886 | 
	Marathi | 
	pib | 
| 3.126833 | -51.392519 | 
	Marathi | 
	pib | 
| -57.749879 | -1.494385 | 
	Marathi | 
	pib | 
| 29.231269 | 29.155568 | 
	Marathi | 
	pib | 
| 4.287004 | -50.418744 | 
	Marathi | 
	pib | 
| -72.219699 | -6.811074 | 
	Marathi | 
	pib | 
| 36.055559 | 15.382714 | 
	Marathi | 
	pib | 
| 30.532288 | 28.083667 | 
	Marathi | 
	pib | 
| 33.849957 | 15.534781 | 
	Marathi | 
	pib | 
| -15.743342 | 0.505287 | 
	Marathi | 
	pib | 
| 3.215276 | -51.245484 | 
	Marathi | 
	pib | 
| -60.737007 | -3.394126 | 
	Marathi | 
	pib | 
| 31.007212 | 28.819507 | 
	Marathi | 
	pib | 
| 2.989671 | -51.270435 | 
	Marathi | 
	pib | 
| 31.167064 | 28.286369 | 
	Marathi | 
	pib | 
| -14.748556 | 0.247347 | 
	Marathi | 
	pib | 
| 1.963952 | -51.834236 | 
	Marathi | 
	pib | 
| -74.551162 | -18.237986 | 
	Marathi | 
	pib | 
| 16.076539 | 9.587664 | 
	Marathi | 
	pib | 
| 29.189616 | 30.146909 | 
	Marathi | 
	pib | 
| -16.437329 | -1.168242 | 
	Marathi | 
	pib | 
| -20.385779 | -22.657032 | 
	Marathi | 
	pib | 
| -17.694435 | -1.679821 | 
	Marathi | 
	pib | 
| 2.924035 | -52.623346 | 
	Marathi | 
	pib | 
| 4.370036 | -50.332863 | 
	Marathi | 
	pib | 
| 31.046874 | 29.276872 | 
	Marathi | 
	pib | 
| 31.962588 | 8.722964 | 
	Marathi | 
	pib | 
| -72.106826 | -7.067241 | 
	Marathi | 
	pib | 
| 31.056858 | 28.393259 | 
	Marathi | 
	pib | 
| -50.119234 | 19.101274 | 
	Marathi | 
	pib | 
| -14.444092 | -15.713925 | 
	Marathi | 
	pib | 
| -10.048824 | 19.417539 | 
	Marathi | 
	pib | 
| -67.977078 | 2.372236 | 
	Marathi | 
	pib | 
| -14.498362 | 0.401051 | 
	Marathi | 
	pib | 
| 2.173073 | -53.360564 | 
	Marathi | 
	pib | 
| 30.477133 | 28.992233 | 
	Marathi | 
	pib | 
| 47.8111 | 14.011303 | 
	Marathi | 
	pib | 
| -14.387866 | -15.729387 | 
	Marathi | 
	pib | 
| -7.195896 | -17.670195 | 
	Marathi | 
	pib | 
| 2.667278 | -52.891831 | 
	Marathi | 
	pib | 
| 29.495611 | 29.168078 | 
	Marathi | 
	pib | 
| 2.064402 | -53.490421 | 
	Marathi | 
	pib | 
| 3.372322 | -51.308322 | 
	Marathi | 
	pib | 
| -13.354959 | 17.659825 | 
	Marathi | 
	pib | 
| 49.659059 | 35.322728 | 
	Marathi | 
	pib | 
| 4.06149 | -50.583698 | 
	Marathi | 
	pib | 
| -62.895019 | 12.828384 | 
	Marathi | 
	pib | 
| -14.255674 | 0.493504 | 
	Marathi | 
	pib | 
| -72.11041 | -7.06807 | 
	Marathi | 
	pib | 
| 30.584932 | 29.276191 | 
	Marathi | 
	pib | 
| -14.43835 | 0.413504 | 
	Marathi | 
	pib | 
| -62.61876 | 4.063817 | 
	Marathi | 
	pib | 
| -6.433463 | -15.836375 | 
	Marathi | 
	pib | 
| 3.714367 | -50.86134 | 
	Marathi | 
	pib | 
| -78.006012 | 14.291905 | 
	Marathi | 
	pib | 
| -54.048352 | 16.858531 | 
	Marathi | 
	pib | 
| 4.433538 | -50.245996 | 
	Marathi | 
	pib | 
| 2.643019 | -51.54507 | 
	Marathi | 
	pib | 
| 31.122343 | 28.581803 | 
	Marathi | 
	pib | 
| 4.102671 | -36.52618 | 
	Marathi | 
	pib | 
| -15.612926 | -0.968628 | 
	Marathi | 
	pib | 
| 44.595938 | 12.468912 | 
	Marathi | 
	pib | 
| -57.052912 | -0.911473 | 
	Marathi | 
	pib | 
| -15.182893 | 1.227909 | 
	Marathi | 
	pib | 
| 3.443609 | -51.095991 | 
	Marathi | 
	pib | 
| -72.072875 | -6.644265 | 
	Marathi | 
	pib | 
| 31.295764 | 29.176204 | 
	Marathi | 
	pib | 
| -55.628744 | 0.708347 | 
	Marathi | 
	pib | 
| 2.249649 | -50.192374 | 
	Marathi | 
	pib | 
| -14.8353 | 0.262177 | 
	Marathi | 
	pib | 
| 28.440251 | 30.565329 | 
	Marathi | 
	pib | 
| 2.552967 | -51.897703 | 
	Marathi | 
	pib | 
| 3.418963 | -51.197725 | 
	Marathi | 
	pib | 
| -13.39446 | 17.752831 | 
	Marathi | 
	pib | 
| -15.661043 | 0.499482 | 
	Marathi | 
	pib | 
| 30.39819 | 28.099538 | 
	Marathi | 
	pib | 
| -57.22444 | 1.737244 | 
	Marathi | 
	pib | 
End of preview. Expand
						in Data Studio
					
What follows is research code. It is by no means optimized for speed, efficiency, or readability.
Data loading, tokenizing and sharding
import os
import numpy as np
import pandas as pd
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.decomposition import TruncatedSVD
from tqdm.notebook import tqdm
from openTSNE import TSNE
import datashader as ds
import colorcet as cc
from dask.distributed import Client
import dask.dataframe as dd
import dask_ml
import dask.bag as db
from transformers import AutoTokenizer
from datasets import load_dataset
from datasets.utils.py_utils import convert_file_size_to_int
def batch_tokenize(batch):
    return {'tokenized': [' '.join(e.tokens) for e in tokenizer(batch['text']).encodings]} # "text" column hard encoded
# The original viz used a subset of the ROOTS Corpus.
# More info on the entire dataset here: https://huggingface.co/bigscience-data
# And here: https://arxiv.org/abs/2303.03915
dset = load_dataset(..., split="train")
dset = dset.map(batch_tokenize, batched=True, batch_size=64, num_proc=28)
dset_name = "roots_subset"
max_shard_size = convert_file_size_to_int('300MB')
dataset_nbytes = dset.data.nbytes
num_shards = int(dataset_nbytes / max_shard_size) + 1
num_shards = max(num_shards, 1)
print(f"Sharding into {num_shards} files.")
os.makedirs(f"{dset_name}/tokenized", exist_ok=True)
for shard_index in tqdm(range(num_shards)):
    shard = dset.shard(num_shards=num_shards, index=shard_index, contiguous=True)
    shard.to_parquet(f"{dset_name}/tokenized/tokenized-{shard_index:03d}.parquet")
Embedding
client = Client() # To keep track of dask computation
client
df = dd.read_parquet(f'{dset_name}/tokenized/')
vect = dask_ml.feature_extraction.text.CountVectorizer(tokenizer=str.split,
                                                       token_pattern=None,
                                                       vocabulary=vocab)
tokenized_bag = df['tokenized'].to_bag()
X = vect.transform(tokenized_bag)
counts = X.compute()
client.shutdown()
tfidf_transformer = TfidfTransformer(sublinear_tf=True, norm="l2")
tfidf = tfidf_transformer.fit_transform(counts)
svd = TruncatedSVD(n_components=160)
X_svd = svd.fit_transform(tfidf)
tsne = TSNE(
    perplexity=30, # not sure what param setting resulted in the plot
    n_jobs=28,
    random_state=42,
    verbose=True,
)
tsne_embedding = tsne.fit(X)
Plotting
df = pd.DataFrame(data=tsne_embedding, columns=['x','y'])
agg = ds.Canvas(plot_height=600, plot_width=600).points(df, 'x', 'y')
img = ds.tf.shade(agg, cmap=cc.fire, how='eq_hist')
ds.tf.set_background(img, "black")
- Downloads last month
- 42
