Extracting and processing VisPubdata data with the Cartolabe API

Comparing the quality of embeddings using multiple methods

# %matplotlib inline
# %load_ext autoreload
# %autoreload 2
# %matplotlib widget

Download data

We will start by downloading the VisPubData dataset from Google Spreadsheet. See Petra Isenberg, Florian Heimerl, Steffen Koch, Tobias Isenberg, Panpan Xu, et al.. vispubdata.org: A Metadata Collection about IEEE Visualization (VIS) Publications. IEEE Transactions on Visualization and Computer Graphics, 2017, 23 (9), pp.2199-2206. ⟨[https://dx.doi.org/10.1109/TVCG.2016.2615308](10.1109/TVCG.2016.2615308)⟩. ⟨[https://dx.doi.org/10.1109/TVCG.2016.2615308](hal-01376597)⟩

SHEET_ID = '1xgoOPu28dQSSGPIp_HHQs0uvvcyLNdkMF9XtRajhhxU'
SHEET_NAME = 'Main%20dataset'
url = f'https://docs.google.com/spreadsheets/d/{SHEET_ID}/gviz/tq?tqx=out:csv&sheet={SHEET_NAME}'

min_df = 25
max_df = 0.1
max_words = 100000
vocab_sample = 250000
num_dims = 300
filt_min_score = 3
n_neighbors = 10

""
import pandas as pd   # noqa

df = pd.read_csv(url)
df.AuthorKeywords.fillna('', inplace=True)
df.Abstract.fillna('', inplace=True)
df.AuthorAffiliation.fillna('', inplace=True)
df['text'] = df.Abstract + ' ' \
            + df.AuthorKeywords + ' ' \
            + df.Title
df.head()

/builds/2mk6rsew/0/hgozukan/cartolabe-data/examples/compare_projections_vispubdata.py:38: FutureWarning: A value is trying to be set on a copy of a DataFrame or Series through chained assignment using an inplace method.
The behavior will change in pandas 3.0. This inplace method will never work because the intermediate object on which we are setting values always behaves as a copy.

For example, when doing 'df[col].method(value, inplace=True)', try using 'df.method({col: value}, inplace=True)' or df[col] = df[col].method(value) instead, to perform the operation inplace on the original object.


  df.AuthorKeywords.fillna('', inplace=True)
/builds/2mk6rsew/0/hgozukan/cartolabe-data/examples/compare_projections_vispubdata.py:39: FutureWarning: A value is trying to be set on a copy of a DataFrame or Series through chained assignment using an inplace method.
The behavior will change in pandas 3.0. This inplace method will never work because the intermediate object on which we are setting values always behaves as a copy.

For example, when doing 'df[col].method(value, inplace=True)', try using 'df.method({col: value}, inplace=True)' or df[col] = df[col].method(value) instead, to perform the operation inplace on the original object.


  df.Abstract.fillna('', inplace=True)
/builds/2mk6rsew/0/hgozukan/cartolabe-data/examples/compare_projections_vispubdata.py:40: FutureWarning: A value is trying to be set on a copy of a DataFrame or Series through chained assignment using an inplace method.
The behavior will change in pandas 3.0. This inplace method will never work because the intermediate object on which we are setting values always behaves as a copy.

For example, when doing 'df[col].method(value, inplace=True)', try using 'df.method({col: value}, inplace=True)' or df[col] = df[col].method(value) instead, to perform the operation inplace on the original object.


  df.AuthorAffiliation.fillna('', inplace=True)

	Conference	Year	Title	DOI	Link	FirstPage	LastPage	PaperType	Abstract	AuthorNames-Deduped	AuthorNames	AuthorAffiliation	InternalReferences	AuthorKeywords	AminerCitationCount	CitationCount_CrossRef	PubsCited_CrossRef	Downloads_Xplore	Award	GraphicsReplicabilityStamp	text
0	Vis	2023	Design Patterns for Situated Visualization in ...	10.1109/tvcg.2023.3327398	http://dx.doi.org/10.1109/TVCG.2023.3327398	1324.0	1335.0	J	Situated visualization has become an increasin...	Benjamin Lee;Michael Sedlmair;Dieter Schmalstieg	Benjamin Lee;Michael Sedlmair;Dieter Schmalstieg	University of Stuttgart, Germany;University of...	10.1109/tvcg.2021.3114835;10.1109/tvcg.2020.30...	Augmented reality,immersive analytics,situated...	NaN	11.0	124.0	736.0	NaN	NaN	Situated visualization has become an increasin...
1	Vis	2023	ggdist: Visualizations of Distributions and Un...	10.1109/tvcg.2023.3327195	http://dx.doi.org/10.1109/TVCG.2023.3327195	414.0	424.0	J	The grammar of graphics is ubiquitous, providi...	Matthew Kay 0001	Matthew Kay	Northwestern University, USA	10.1109/tvcg.2011.185;10.1109/tvcg.2014.234629...	Uncertainty visualization,probability distribu...	NaN	7.0	55.0	281.0	NaN	NaN	The grammar of graphics is ubiquitous, providi...
2	Vis	2023	PromptMagician: Interactive Prompt Engineering...	10.1109/tvcg.2023.3327168	http://dx.doi.org/10.1109/TVCG.2023.3327168	295.0	305.0	J	Generative text-to-image models have gained gr...	Yingchaojie Feng;Xingbo Wang 0001;Kamkwai Wong...	Yingchaojie Feng;Xingbo Wang;Kam Kwai Wong;Sij...	State Key Lab of CAD&CG, Zhejiang University, ...	10.1109/tvcg.2022.3209425;10.1109/tvcg.2006.18...	Prompt engineering,text-to-image generation,im...	NaN	5.0	78.0	1065.0	NaN	NaN	Generative text-to-image models have gained gr...
3	Vis	2023	Challenges and Opportunities in Data Visualiza...	10.1109/tvcg.2023.3327378	http://dx.doi.org/10.1109/TVCG.2023.3327378	649.0	660.0	J	This paper is a call to action for research an...	Benjamin Bach;Mandy Keck;Fateme Rajabiyazdi;Ta...	Benjamin Bach;Mandy Keck;Fateme Rajabiyazdi;Ta...	University of Edinburgh, United Kingdom;Univer...	10.1109/tvcg.2022.3209402;10.1109/tvcg.2022.32...	Data Visualization,Education,Challenges	NaN	5.0	138.0	563.0	NaN	NaN	This paper is a call to action for research an...
4	Vis	2023	Affective Visualization Design: Leveraging the...	10.1109/tvcg.2023.3327385	http://dx.doi.org/10.1109/TVCG.2023.3327385	1.0	11.0	J	In recent years, more and more researchers hav...	Xingyu Lan;Yanqiu Wu 0001;Nan Cao 0001	Xingyu Lan;Yanqiu Wu;Nan Cao	Fudan University, Research Group of Computatio...	10.1109/tvcg.2021.3114775;10.1109/tvcg.2020.30...	Information Visualization,Affective Design,Vis...	NaN	4.0	95.0	848.0	BP	NaN	In recent years, more and more researchers hav...

Creating correspondance matrices for each entity type

From this table of articles, we want to extract matrices that will map the correspondance between these articles and the entities we want to use.

Authors

Let’s start with the authors for example. We want to create a matrix where the rows represent the articles and the columns represent the authors. Each cell (n, m) will have a 1 in it if the nth article was written by the mth author.

As we have multiple dataframes, the results will be arrays corresponding to specified dataframes.

from cartodata.loading import load_comma_separated_column  # noqa

authors_mat, authors_scores = load_comma_separated_column(df, 'AuthorNames-Deduped', comma=';')
authors_mat.shape

""
authors_scores.head()

Benjamin Lee           3
Michael Sedlmair      23
Dieter Schmalstieg     9
Matthew Kay 0001      10
Yingchaojie Feng       1
dtype: int64

If we look at the 2nd column of the matrix, which corresponds to the author Michelle Borkin, we can see that she has 8 non-zero rows, each row indicating which articles she authored.

print(authors_mat[:, 1])

(0, 0)        1
(176, 0)      1
(196, 0)      1
(272, 0)      1
(386, 0)      1
(408, 0)      1
(412, 0)      1
(541, 0)      1
(667, 0)      1
(687, 0)      1
(785, 0)      1
(792, 0)      1
(798, 0)      1
(844, 0)      1
(928, 0)      1
(957, 0)      1
(1166, 0)     1
(1251, 0)     1
(1304, 0)     1
(1334, 0)     1
(1397, 0)     1
(1435, 0)     1
(1709, 0)     1

Filtering low score entities

A lot of the authors that we just extracted from the dataframe have a very low score, which means they’re only linked to one or two articles. To improve the quality of our data, we’ll filter the authors by removing those that appear less than 3 times.

To do this, we’ll use the filter_min_score function.

from cartodata.operations import filter_min_score  # noqa

authors_before = len(authors_scores)

authors_mat, authors_scores = filter_min_score(authors_mat,
                                               authors_scores,
                                               filt_min_score)

print(f"Removed {authors_before - len(authors_scores)} authors with less "
      f"than 3 articles from a total of {authors_before} authors.")
print(f"Working with {len(authors_scores)} authors.\n")

Removed 6267 authors with less than 3 articles from a total of 6987 authors.
Working with 720 authors.

Words

For the words, it’s a bit trickier because we want to extract n-grams (groups of n terms) instead of just comma separated values. We’ll call the load_text_column which uses scikit-learn’s CountVectorizer to create a vocabulary and map the tokens.

from cartodata.loading import load_text_column  # noqa
from sklearn.feature_extraction import text as sktxt  # noqa

with open('../datas/stopwords.txt', 'r') as stop_file:
    stopwords = sktxt.ENGLISH_STOP_WORDS.union(
        set(stop_file.read().splitlines()))

words_mat, words_scores = load_text_column(df['text'],
                                           4,
                                           min_df,
                                           max_df,
                                           stopwords=stopwords)

""
words_scores.head()

""
words_mat.shape

""
from cartodata.operations import normalize_tfidf  # noqa

words_mat = normalize_tfidf(words_mat)
words_mat.shape

""
from cartodata.loading import load_identity_column  # noqa

articles_mat, articles_scores = load_identity_column(df, 'Title')
articles_scores.head()

Design Patterns for Situated Visualization in Augmented Reality                       1.0
ggdist: Visualizations of Distributions and Uncertainty in the Grammar of Graphics    1.0
PromptMagician: Interactive Prompt Engineering for Text-to-Image Creation             1.0
Challenges and Opportunities in Data Visualization Education: A Call to Action        1.0
Affective Visualization Design: Leveraging the Emotional Impact of Data               1.0
dtype: float64

Dimension reduction/Embeddings

One way to see the matrices that we created is as coordinates in the space of all articles. What we want to do is to reduce the dimension of this space to make it easier to work with and see.

Validation

We compute a score that counts the average number of times the 10 nearest neighbors of an article are from the same author as the article. For each author, we have a number between 1 (100%) and 0.1 (none of the articles are from the same author, except the initial article itself).

LSA projection

We’ll start by using the LSA (Latent Semantic Analysis) technique to identify keywords in our data and thus reduce the number of rows in our matrices. The lsa_projection method takes three arguments:

• the number of dimensions you want to keep

• the matrix of documents/words frequency

• a list of matrices to project

It returns a list of the same length containing the matrices projected in the latent space.

We also apply an l2 normalization to each feature of the projected matrices.

from cartodata.projection import lsa_projection  # noqa
from cartodata.operations import normalize_l2  # noqa

""

''

lsa_matrices = lsa_projection(num_dims, words_mat, [articles_mat, authors_mat, words_mat])

""
lsa_matrices = list(map(normalize_l2, lsa_matrices))

We’ve reduced the number of rows in each of articles_mat, authors_mat, words_mat and labs_mat to just 80.

print(f"articles_mat: {lsa_matrices[0].shape}")
print(f"authors_mat: {lsa_matrices[1].shape}")
print(f"words_mat: {lsa_matrices[2].shape}")

""
from cartodata.model_selection.scoring import Neighbors # noqa

NATURE = "articles"
SOURCE = "authors"

lsa_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=lsa_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
lsa_score.print()

articles_mat: (300, 3753)
authors_mat: (300, 720)
words_mat: (300, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1109

--------- Desc Scores ---------
neighbors_articles_authors_det

0        Davide Ceneda : 0.40
1        Yi-Jen Chiang : 0.35
2      Mark A. Whiting : 0.33
3        Mario Jelovic : 0.30
4           David Gotz : 0.29
                ...
715    Enrico Gobbetti : 0.02
716     Donna L. Gresh : 0.02
717     Steven F. Roth : 0.01
718    Bruno Lévy 0001 : 0.00
719         Jinzhu Gao : 0.00
Length: 720, dtype: object

VALUE : 0.1109

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.117391
Dieter Schmalstieg     0.111111
Matthew Kay 0001       0.160000
Xingbo Wang 0001       0.116667
Minfeng Zhu            0.114286
                         ...
Lisa M. Sobierajski    0.075000
Nahum D. Gershon       0.060000
Sidney W. Wang         0.025000
David A. Lane          0.066667
T. Todd Elvins         0.050000
Length: 720, dtype: float64

LDA projection

from cartodata.projection import lda_projection  # noqa

""

''

lda_matrices = lda_projection(num_dims, 1, [articles_mat, authors_mat, words_mat])

""
lda_matrices = list(map(normalize_l2, lda_matrices))

""
print(f"articles_mat: {lda_matrices[0].shape}")
print(f"authors_mat: {lda_matrices[1].shape}")
print(f"words_mat: {lda_matrices[2].shape}")

""
lda_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=lda_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
lda_score.print()

articles_mat: (300, 3753)
authors_mat: (300, 720)
words_mat: (300, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.0013

--------- Desc Scores ---------
neighbors_articles_authors_det

0      Hanspeter Pfister : 0.10
1        Johanna Schmidt : 0.10
2              Huamin Qu : 0.10
3       Jeff W. Lichtman : 0.10
4       Andrew J. Hanson : 0.10
                 ...
715     Rainer Splechtna : 0.00
716       Denis Gracanin : 0.00
717        Helwig Hauser : 0.00
718    Kresimir Matkovic : 0.00
719       T. Todd Elvins : 0.00
Length: 720, dtype: object

VALUE : 0.0013

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.0
Dieter Schmalstieg     0.0
Matthew Kay 0001       0.0
Xingbo Wang 0001       0.0
Minfeng Zhu            0.0
                      ...
Lisa M. Sobierajski    0.0
Nahum D. Gershon       0.0
Sidney W. Wang         0.0
David A. Lane          0.0
T. Todd Elvins         0.0
Length: 720, dtype: float64

DOC2Vec projection

from cartodata.projection import doc2vec_projection  # noqa

""

''

doc2vec_matrices = doc2vec_projection(num_dims, 1, [articles_mat, authors_mat, words_mat], df['text'])

""
doc2vec_matrices = list(map(normalize_l2, doc2vec_matrices))

""
print(f"articles_mat: {doc2vec_matrices[0].shape}")
print(f"authors_mat: {doc2vec_matrices[1].shape}")
print(f"words_mat: {doc2vec_matrices[2].shape}")

""
doc2vec_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=doc2vec_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
doc2vec_score.print()

articles_mat: (300, 3753)
authors_mat: (300, 720)
words_mat: (300, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1200

--------- Desc Scores ---------
neighbors_articles_authors_det

0         Steven Franconeri : 0.35
1               Cindy Xiong : 0.32
2           Mark A. Whiting : 0.30
3              Jean Scholtz : 0.30
4           Laurent Lessard : 0.25
                  ...
715           Yarden Livnat : 0.10
716             Hongan Wang : 0.10
717     Matthew Cooper 0001 : 0.10
718    Jimmy Johansson 0001 : 0.10
719          T. Todd Elvins : 0.10
Length: 720, dtype: object

VALUE : 0.1200

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.126087
Dieter Schmalstieg     0.100000
Matthew Kay 0001       0.110000
Xingbo Wang 0001       0.150000
Minfeng Zhu            0.114286
                         ...
Lisa M. Sobierajski    0.125000
Nahum D. Gershon       0.100000
Sidney W. Wang         0.100000
David A. Lane          0.166667
T. Todd Elvins         0.100000
Length: 720, dtype: float64

Specter2 projection

from cartodata.projection import bert_projection  # noqa

""

''

specter2_matrices = bert_projection([articles_mat, authors_mat, words_mat], df['text'])

""
specter2_matrices = list(map(normalize_l2, specter2_matrices))

""
print(f"articles_mat: {specter2_matrices[0].shape}")
print(f"authors_mat: {specter2_matrices[1].shape}")
print(f"words_mat: {specter2_matrices[2].shape}")

""
specter2_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=specter2_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
specter2_score.print()

Using torch device: cpu

Fetching 4 files:   0%|          | 0/4 [00:00<?, ?it/s]
Fetching 4 files: 100%|██████████| 4/4 [00:00<00:00, 1201.89it/s]
/usr/local/lib/python3.9/site-packages/adapters/loading.py:165: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  state_dict = torch.load(weights_file, map_location="cpu")

Processing batches:   0%|          | 0/376 [00:00<?, ?it/s]
Processing batches:   0%|          | 1/376 [00:05<34:44,  5.56s/it]
Processing batches:   1%|          | 2/376 [00:10<33:51,  5.43s/it]
Processing batches:   1%|          | 3/376 [00:16<34:16,  5.51s/it]
Processing batches:   1%|          | 4/376 [00:21<34:00,  5.48s/it]
Processing batches:   1%|▏         | 5/376 [00:27<34:23,  5.56s/it]
Processing batches:   2%|▏         | 6/376 [00:32<33:45,  5.47s/it]
Processing batches:   2%|▏         | 7/376 [00:38<33:25,  5.43s/it]
Processing batches:   2%|▏         | 8/376 [00:44<33:51,  5.52s/it]
Processing batches:   2%|▏         | 9/376 [00:49<33:28,  5.47s/it]
Processing batches:   3%|▎         | 10/376 [00:54<33:15,  5.45s/it]
Processing batches:   3%|▎         | 11/376 [01:00<33:32,  5.51s/it]
Processing batches:   3%|▎         | 12/376 [01:06<33:39,  5.55s/it]
Processing batches:   3%|▎         | 13/376 [01:11<33:32,  5.54s/it]
Processing batches:   4%|▎         | 14/376 [01:16<32:59,  5.47s/it]
Processing batches:   4%|▍         | 15/376 [01:22<32:45,  5.45s/it]
Processing batches:   4%|▍         | 16/376 [01:27<32:32,  5.42s/it]
Processing batches:   5%|▍         | 17/376 [01:32<32:14,  5.39s/it]
Processing batches:   5%|▍         | 18/376 [01:38<32:03,  5.37s/it]
Processing batches:   5%|▌         | 19/376 [01:44<32:34,  5.47s/it]
Processing batches:   5%|▌         | 20/376 [01:49<32:21,  5.45s/it]
Processing batches:   6%|▌         | 21/376 [01:55<32:30,  5.49s/it]
Processing batches:   6%|▌         | 22/376 [02:00<32:08,  5.45s/it]
Processing batches:   6%|▌         | 23/376 [02:06<32:29,  5.52s/it]
Processing batches:   6%|▋         | 24/376 [02:11<32:21,  5.52s/it]
Processing batches:   7%|▋         | 25/376 [02:17<32:32,  5.56s/it]
Processing batches:   7%|▋         | 26/376 [02:22<32:06,  5.50s/it]
Processing batches:   7%|▋         | 27/376 [02:27<31:41,  5.45s/it]
Processing batches:   7%|▋         | 28/376 [02:33<31:36,  5.45s/it]
Processing batches:   8%|▊         | 29/376 [02:38<31:24,  5.43s/it]
Processing batches:   8%|▊         | 30/376 [02:44<31:34,  5.47s/it]
Processing batches:   8%|▊         | 31/376 [02:49<31:34,  5.49s/it]
Processing batches:   9%|▊         | 32/376 [02:55<31:16,  5.46s/it]
Processing batches:   9%|▉         | 33/376 [03:00<31:02,  5.43s/it]
Processing batches:   9%|▉         | 34/376 [03:06<31:09,  5.47s/it]
Processing batches:   9%|▉         | 35/376 [03:11<31:05,  5.47s/it]
Processing batches:  10%|▉         | 36/376 [03:17<31:00,  5.47s/it]
Processing batches:  10%|▉         | 37/376 [03:22<30:47,  5.45s/it]
Processing batches:  10%|█         | 38/376 [03:28<31:01,  5.51s/it]
Processing batches:  10%|█         | 39/376 [03:33<30:47,  5.48s/it]
Processing batches:  11%|█         | 40/376 [03:39<30:59,  5.53s/it]
Processing batches:  11%|█         | 41/376 [03:44<30:38,  5.49s/it]
Processing batches:  11%|█         | 42/376 [03:50<30:31,  5.48s/it]
Processing batches:  11%|█▏        | 43/376 [03:55<30:22,  5.47s/it]
Processing batches:  12%|█▏        | 44/376 [04:01<30:45,  5.56s/it]
Processing batches:  12%|█▏        | 45/376 [04:06<30:23,  5.51s/it]
Processing batches:  12%|█▏        | 46/376 [04:12<30:38,  5.57s/it]
Processing batches:  12%|█▎        | 47/376 [04:18<30:38,  5.59s/it]
Processing batches:  13%|█▎        | 48/376 [04:23<30:14,  5.53s/it]
Processing batches:  13%|█▎        | 49/376 [04:28<29:54,  5.49s/it]
Processing batches:  13%|█▎        | 50/376 [04:34<29:44,  5.47s/it]
Processing batches:  14%|█▎        | 51/376 [04:39<30:04,  5.55s/it]
Processing batches:  14%|█▍        | 52/376 [04:45<29:43,  5.50s/it]
Processing batches:  14%|█▍        | 53/376 [04:50<29:36,  5.50s/it]
Processing batches:  14%|█▍        | 54/376 [04:56<29:27,  5.49s/it]
Processing batches:  15%|█▍        | 55/376 [05:01<29:36,  5.53s/it]
Processing batches:  15%|█▍        | 56/376 [05:07<29:17,  5.49s/it]
Processing batches:  15%|█▌        | 57/376 [05:13<29:34,  5.56s/it]
Processing batches:  15%|█▌        | 58/376 [05:18<29:14,  5.52s/it]
Processing batches:  16%|█▌        | 59/376 [05:24<29:27,  5.58s/it]
Processing batches:  16%|█▌        | 60/376 [05:29<29:06,  5.53s/it]
Processing batches:  16%|█▌        | 61/376 [05:35<29:12,  5.56s/it]
Processing batches:  16%|█▋        | 62/376 [05:40<28:45,  5.50s/it]
Processing batches:  17%|█▋        | 63/376 [05:46<28:57,  5.55s/it]
Processing batches:  17%|█▋        | 64/376 [05:51<28:34,  5.50s/it]
Processing batches:  17%|█▋        | 65/376 [05:57<28:53,  5.57s/it]
Processing batches:  18%|█▊        | 66/376 [06:03<28:50,  5.58s/it]
Processing batches:  18%|█▊        | 67/376 [06:08<28:54,  5.61s/it]
Processing batches:  18%|█▊        | 68/376 [06:14<28:49,  5.61s/it]
Processing batches:  18%|█▊        | 69/376 [06:19<28:40,  5.60s/it]
Processing batches:  19%|█▊        | 70/376 [06:25<28:28,  5.58s/it]
Processing batches:  19%|█▉        | 71/376 [06:31<28:26,  5.60s/it]
Processing batches:  19%|█▉        | 72/376 [06:36<28:27,  5.62s/it]
Processing batches:  19%|█▉        | 73/376 [06:42<28:26,  5.63s/it]
Processing batches:  20%|█▉        | 74/376 [06:47<27:57,  5.56s/it]
Processing batches:  20%|█▉        | 75/376 [06:53<27:53,  5.56s/it]
Processing batches:  20%|██        | 76/376 [06:58<27:48,  5.56s/it]
Processing batches:  20%|██        | 77/376 [07:04<27:25,  5.50s/it]
Processing batches:  21%|██        | 78/376 [07:09<27:34,  5.55s/it]
Processing batches:  21%|██        | 79/376 [07:15<27:36,  5.58s/it]
Processing batches:  21%|██▏       | 80/376 [07:21<27:34,  5.59s/it]
Processing batches:  22%|██▏       | 81/376 [07:26<27:29,  5.59s/it]
Processing batches:  22%|██▏       | 82/376 [07:32<27:25,  5.60s/it]
Processing batches:  22%|██▏       | 83/376 [07:37<27:15,  5.58s/it]
Processing batches:  22%|██▏       | 84/376 [07:43<27:07,  5.58s/it]
Processing batches:  23%|██▎       | 85/376 [07:49<27:02,  5.57s/it]
Processing batches:  23%|██▎       | 86/376 [07:54<26:40,  5.52s/it]
Processing batches:  23%|██▎       | 87/376 [08:00<26:38,  5.53s/it]
Processing batches:  23%|██▎       | 88/376 [08:05<26:15,  5.47s/it]
Processing batches:  24%|██▎       | 89/376 [08:10<26:17,  5.50s/it]
Processing batches:  24%|██▍       | 90/376 [08:16<25:57,  5.44s/it]
Processing batches:  24%|██▍       | 91/376 [08:21<25:57,  5.46s/it]
Processing batches:  24%|██▍       | 92/376 [08:27<25:40,  5.43s/it]
Processing batches:  25%|██▍       | 93/376 [08:32<25:44,  5.46s/it]
Processing batches:  25%|██▌       | 94/376 [08:38<25:41,  5.46s/it]
Processing batches:  25%|██▌       | 95/376 [08:43<25:46,  5.50s/it]
Processing batches:  26%|██▌       | 96/376 [08:49<25:26,  5.45s/it]
Processing batches:  26%|██▌       | 97/376 [08:54<25:27,  5.47s/it]
Processing batches:  26%|██▌       | 98/376 [08:59<25:08,  5.43s/it]
Processing batches:  26%|██▋       | 99/376 [09:05<25:16,  5.47s/it]
Processing batches:  27%|██▋       | 100/376 [09:10<24:55,  5.42s/it]
Processing batches:  27%|██▋       | 101/376 [09:16<25:08,  5.49s/it]
Processing batches:  27%|██▋       | 102/376 [09:21<24:56,  5.46s/it]
Processing batches:  27%|██▋       | 103/376 [09:27<24:49,  5.46s/it]
Processing batches:  28%|██▊       | 104/376 [09:32<24:56,  5.50s/it]
Processing batches:  28%|██▊       | 105/376 [09:38<24:41,  5.47s/it]
Processing batches:  28%|██▊       | 106/376 [09:43<24:54,  5.54s/it]
Processing batches:  28%|██▊       | 107/376 [09:49<24:52,  5.55s/it]
Processing batches:  29%|██▊       | 108/376 [09:54<24:39,  5.52s/it]
Processing batches:  29%|██▉       | 109/376 [10:00<24:50,  5.58s/it]
Processing batches:  29%|██▉       | 110/376 [10:06<24:41,  5.57s/it]
Processing batches:  30%|██▉       | 111/376 [10:11<24:39,  5.58s/it]
Processing batches:  30%|██▉       | 112/376 [10:17<24:18,  5.53s/it]
Processing batches:  30%|███       | 113/376 [10:22<24:25,  5.57s/it]
Processing batches:  30%|███       | 114/376 [10:28<24:16,  5.56s/it]
Processing batches:  31%|███       | 115/376 [10:34<24:29,  5.63s/it]
Processing batches:  31%|███       | 116/376 [10:39<24:16,  5.60s/it]
Processing batches:  31%|███       | 117/376 [10:45<24:17,  5.63s/it]
Processing batches:  31%|███▏      | 118/376 [10:51<24:18,  5.65s/it]
Processing batches:  32%|███▏      | 119/376 [10:56<24:12,  5.65s/it]
Processing batches:  32%|███▏      | 120/376 [11:02<24:12,  5.67s/it]
Processing batches:  32%|███▏      | 121/376 [11:08<24:13,  5.70s/it]
Processing batches:  32%|███▏      | 122/376 [11:14<24:09,  5.71s/it]
Processing batches:  33%|███▎      | 123/376 [11:19<23:44,  5.63s/it]
Processing batches:  33%|███▎      | 124/376 [11:25<23:41,  5.64s/it]
Processing batches:  33%|███▎      | 125/376 [11:30<23:29,  5.61s/it]
Processing batches:  34%|███▎      | 126/376 [11:36<23:30,  5.64s/it]
Processing batches:  34%|███▍      | 127/376 [11:41<23:13,  5.59s/it]
Processing batches:  34%|███▍      | 128/376 [11:47<23:12,  5.62s/it]
Processing batches:  34%|███▍      | 129/376 [11:53<23:07,  5.62s/it]
Processing batches:  35%|███▍      | 130/376 [11:58<23:01,  5.62s/it]
Processing batches:  35%|███▍      | 131/376 [12:04<22:46,  5.58s/it]
Processing batches:  35%|███▌      | 132/376 [12:09<22:45,  5.59s/it]
Processing batches:  35%|███▌      | 133/376 [12:15<22:41,  5.60s/it]
Processing batches:  36%|███▌      | 134/376 [12:21<22:45,  5.64s/it]
Processing batches:  36%|███▌      | 135/376 [12:26<22:40,  5.64s/it]
Processing batches:  36%|███▌      | 136/376 [12:32<22:19,  5.58s/it]
Processing batches:  36%|███▋      | 137/376 [12:37<22:07,  5.56s/it]
Processing batches:  37%|███▋      | 138/376 [12:43<21:47,  5.50s/it]
Processing batches:  37%|███▋      | 139/376 [12:48<21:32,  5.45s/it]
Processing batches:  37%|███▋      | 140/376 [12:53<21:20,  5.43s/it]
Processing batches:  38%|███▊      | 141/376 [12:59<21:15,  5.43s/it]
Processing batches:  38%|███▊      | 142/376 [13:04<21:18,  5.46s/it]
Processing batches:  38%|███▊      | 143/376 [13:10<20:57,  5.40s/it]
Processing batches:  38%|███▊      | 144/376 [13:15<20:58,  5.42s/it]
Processing batches:  39%|███▊      | 145/376 [13:20<20:45,  5.39s/it]
Processing batches:  39%|███▉      | 146/376 [13:26<20:34,  5.37s/it]
Processing batches:  39%|███▉      | 147/376 [13:31<20:35,  5.40s/it]
Processing batches:  39%|███▉      | 148/376 [13:37<20:28,  5.39s/it]
Processing batches:  40%|███▉      | 149/376 [13:42<20:29,  5.42s/it]
Processing batches:  40%|███▉      | 150/376 [13:48<20:30,  5.45s/it]
Processing batches:  40%|████      | 151/376 [13:53<20:23,  5.44s/it]
Processing batches:  40%|████      | 152/376 [13:59<20:44,  5.55s/it]
Processing batches:  41%|████      | 153/376 [14:04<20:43,  5.58s/it]
Processing batches:  41%|████      | 154/376 [14:10<20:50,  5.63s/it]
Processing batches:  41%|████      | 155/376 [14:16<20:32,  5.58s/it]
Processing batches:  41%|████▏     | 156/376 [14:21<20:35,  5.62s/it]
Processing batches:  42%|████▏     | 157/376 [14:27<20:23,  5.59s/it]
Processing batches:  42%|████▏     | 158/376 [14:32<20:12,  5.56s/it]
Processing batches:  42%|████▏     | 159/376 [14:38<20:14,  5.60s/it]
Processing batches:  43%|████▎     | 160/376 [14:44<20:22,  5.66s/it]
Processing batches:  43%|████▎     | 161/376 [14:49<20:09,  5.63s/it]
Processing batches:  43%|████▎     | 162/376 [14:55<20:18,  5.69s/it]
Processing batches:  43%|████▎     | 163/376 [15:01<20:04,  5.66s/it]
Processing batches:  44%|████▎     | 164/376 [15:07<20:19,  5.75s/it]
Processing batches:  44%|████▍     | 165/376 [15:12<20:07,  5.72s/it]
Processing batches:  44%|████▍     | 166/376 [15:18<19:39,  5.62s/it]
Processing batches:  44%|████▍     | 167/376 [15:24<19:39,  5.64s/it]
Processing batches:  45%|████▍     | 168/376 [15:29<19:11,  5.54s/it]
Processing batches:  45%|████▍     | 169/376 [15:34<19:07,  5.55s/it]
Processing batches:  45%|████▌     | 170/376 [15:40<18:53,  5.50s/it]
Processing batches:  45%|████▌     | 171/376 [15:45<18:58,  5.55s/it]
Processing batches:  46%|████▌     | 172/376 [15:51<18:53,  5.56s/it]
Processing batches:  46%|████▌     | 173/376 [15:56<18:42,  5.53s/it]
Processing batches:  46%|████▋     | 174/376 [16:02<18:49,  5.59s/it]
Processing batches:  47%|████▋     | 175/376 [16:08<18:51,  5.63s/it]
Processing batches:  47%|████▋     | 176/376 [16:12<16:44,  5.02s/it]
Processing batches:  47%|████▋     | 177/376 [16:17<17:21,  5.24s/it]
Processing batches:  47%|████▋     | 178/376 [16:23<17:24,  5.28s/it]
Processing batches:  48%|████▊     | 179/376 [16:28<17:27,  5.32s/it]
Processing batches:  48%|████▊     | 180/376 [16:34<17:29,  5.36s/it]
Processing batches:  48%|████▊     | 181/376 [16:39<17:40,  5.44s/it]
Processing batches:  48%|████▊     | 182/376 [16:45<17:32,  5.42s/it]
Processing batches:  49%|████▊     | 183/376 [16:50<17:20,  5.39s/it]
Processing batches:  49%|████▉     | 184/376 [16:55<17:19,  5.41s/it]
Processing batches:  49%|████▉     | 185/376 [17:01<17:15,  5.42s/it]
Processing batches:  49%|████▉     | 186/376 [17:06<17:19,  5.47s/it]
Processing batches:  50%|████▉     | 187/376 [17:12<17:15,  5.48s/it]
Processing batches:  50%|█████     | 188/376 [17:17<17:00,  5.43s/it]
Processing batches:  50%|█████     | 189/376 [17:22<16:41,  5.36s/it]
Processing batches:  51%|█████     | 190/376 [17:28<16:40,  5.38s/it]
Processing batches:  51%|█████     | 191/376 [17:33<16:42,  5.42s/it]
Processing batches:  51%|█████     | 192/376 [17:39<16:28,  5.37s/it]
Processing batches:  51%|█████▏    | 193/376 [17:44<16:11,  5.31s/it]
Processing batches:  52%|█████▏    | 194/376 [17:49<16:02,  5.29s/it]
Processing batches:  52%|█████▏    | 195/376 [17:54<15:58,  5.30s/it]
Processing batches:  52%|█████▏    | 196/376 [18:00<16:03,  5.35s/it]
Processing batches:  52%|█████▏    | 197/376 [18:05<16:14,  5.44s/it]
Processing batches:  53%|█████▎    | 198/376 [18:11<16:06,  5.43s/it]
Processing batches:  53%|█████▎    | 199/376 [18:16<15:58,  5.41s/it]
Processing batches:  53%|█████▎    | 200/376 [18:22<15:51,  5.41s/it]
Processing batches:  53%|█████▎    | 201/376 [18:27<15:46,  5.41s/it]
Processing batches:  54%|█████▎    | 202/376 [18:32<15:07,  5.22s/it]
Processing batches:  54%|█████▍    | 203/376 [18:38<15:32,  5.39s/it]
Processing batches:  54%|█████▍    | 204/376 [18:43<15:36,  5.44s/it]
Processing batches:  55%|█████▍    | 205/376 [18:49<15:28,  5.43s/it]
Processing batches:  55%|█████▍    | 206/376 [18:54<15:23,  5.43s/it]
Processing batches:  55%|█████▌    | 207/376 [19:00<15:24,  5.47s/it]
Processing batches:  55%|█████▌    | 208/376 [19:05<15:17,  5.46s/it]
Processing batches:  56%|█████▌    | 209/376 [19:10<15:10,  5.45s/it]
Processing batches:  56%|█████▌    | 210/376 [19:16<15:00,  5.42s/it]
Processing batches:  56%|█████▌    | 211/376 [19:21<14:52,  5.41s/it]
Processing batches:  56%|█████▋    | 212/376 [19:27<15:05,  5.52s/it]
Processing batches:  57%|█████▋    | 213/376 [19:33<15:09,  5.58s/it]
Processing batches:  57%|█████▋    | 214/376 [19:38<15:17,  5.66s/it]
Processing batches:  57%|█████▋    | 215/376 [19:43<14:28,  5.39s/it]
Processing batches:  57%|█████▋    | 216/376 [19:48<14:08,  5.30s/it]
Processing batches:  58%|█████▊    | 217/376 [19:54<14:11,  5.35s/it]
Processing batches:  58%|█████▊    | 218/376 [20:00<14:22,  5.46s/it]
Processing batches:  58%|█████▊    | 219/376 [20:05<14:29,  5.54s/it]
Processing batches:  59%|█████▊    | 220/376 [20:11<14:28,  5.57s/it]
Processing batches:  59%|█████▉    | 221/376 [20:17<14:27,  5.60s/it]
Processing batches:  59%|█████▉    | 222/376 [20:22<14:25,  5.62s/it]
Processing batches:  59%|█████▉    | 223/376 [20:28<14:17,  5.61s/it]
Processing batches:  60%|█████▉    | 224/376 [20:33<14:14,  5.62s/it]
Processing batches:  60%|█████▉    | 225/376 [20:39<14:09,  5.62s/it]
Processing batches:  60%|██████    | 226/376 [20:45<14:05,  5.63s/it]
Processing batches:  60%|██████    | 227/376 [20:51<14:05,  5.68s/it]
Processing batches:  61%|██████    | 228/376 [20:56<14:00,  5.68s/it]
Processing batches:  61%|██████    | 229/376 [21:02<13:56,  5.69s/it]
Processing batches:  61%|██████    | 230/376 [21:08<13:46,  5.66s/it]
Processing batches:  61%|██████▏   | 231/376 [21:13<13:44,  5.68s/it]
Processing batches:  62%|██████▏   | 232/376 [21:19<13:35,  5.67s/it]
Processing batches:  62%|██████▏   | 233/376 [21:25<13:29,  5.66s/it]
Processing batches:  62%|██████▏   | 234/376 [21:29<12:34,  5.31s/it]
Processing batches:  62%|██████▎   | 235/376 [21:34<12:35,  5.36s/it]
Processing batches:  63%|██████▎   | 236/376 [21:40<12:42,  5.45s/it]
Processing batches:  63%|██████▎   | 237/376 [21:46<12:49,  5.53s/it]
Processing batches:  63%|██████▎   | 238/376 [21:52<12:47,  5.56s/it]
Processing batches:  64%|██████▎   | 239/376 [21:57<12:55,  5.66s/it]
Processing batches:  64%|██████▍   | 240/376 [22:03<12:50,  5.66s/it]
Processing batches:  64%|██████▍   | 241/376 [22:09<12:45,  5.67s/it]
Processing batches:  64%|██████▍   | 242/376 [22:14<12:37,  5.66s/it]
Processing batches:  65%|██████▍   | 243/376 [22:20<12:38,  5.70s/it]
Processing batches:  65%|██████▍   | 244/376 [22:26<12:29,  5.68s/it]
Processing batches:  65%|██████▌   | 245/376 [22:31<12:23,  5.68s/it]
Processing batches:  65%|██████▌   | 246/376 [22:37<12:20,  5.69s/it]
Processing batches:  66%|██████▌   | 247/376 [22:43<12:18,  5.72s/it]
Processing batches:  66%|██████▌   | 248/376 [22:49<12:19,  5.78s/it]
Processing batches:  66%|██████▌   | 249/376 [22:53<10:58,  5.18s/it]
Processing batches:  66%|██████▋   | 250/376 [22:58<10:45,  5.12s/it]
Processing batches:  67%|██████▋   | 251/376 [23:02<10:15,  4.93s/it]
Processing batches:  67%|██████▋   | 252/376 [23:08<10:53,  5.27s/it]
Processing batches:  67%|██████▋   | 253/376 [23:14<11:00,  5.37s/it]
Processing batches:  68%|██████▊   | 254/376 [23:20<11:11,  5.50s/it]
Processing batches:  68%|██████▊   | 255/376 [23:25<11:08,  5.53s/it]
Processing batches:  68%|██████▊   | 256/376 [23:31<11:02,  5.52s/it]
Processing batches:  68%|██████▊   | 257/376 [23:36<11:01,  5.56s/it]
Processing batches:  69%|██████▊   | 258/376 [23:42<11:13,  5.71s/it]
Processing batches:  69%|██████▉   | 259/376 [23:48<11:14,  5.77s/it]
Processing batches:  69%|██████▉   | 260/376 [23:54<11:06,  5.75s/it]
Processing batches:  69%|██████▉   | 261/376 [24:00<10:53,  5.68s/it]
Processing batches:  70%|██████▉   | 262/376 [24:06<10:58,  5.77s/it]
Processing batches:  70%|██████▉   | 263/376 [24:12<11:00,  5.85s/it]
Processing batches:  70%|███████   | 264/376 [24:17<10:51,  5.82s/it]
Processing batches:  70%|███████   | 265/376 [24:23<10:55,  5.90s/it]
Processing batches:  71%|███████   | 266/376 [24:29<10:51,  5.93s/it]
Processing batches:  71%|███████   | 267/376 [24:35<10:38,  5.86s/it]
Processing batches:  71%|███████▏  | 268/376 [24:40<10:06,  5.62s/it]
Processing batches:  72%|███████▏  | 269/376 [24:46<10:03,  5.64s/it]
Processing batches:  72%|███████▏  | 270/376 [24:51<09:55,  5.61s/it]
Processing batches:  72%|███████▏  | 271/376 [24:57<09:46,  5.59s/it]
Processing batches:  72%|███████▏  | 272/376 [25:03<09:47,  5.65s/it]
Processing batches:  73%|███████▎  | 273/376 [25:08<09:43,  5.66s/it]
Processing batches:  73%|███████▎  | 274/376 [25:14<09:45,  5.74s/it]
Processing batches:  73%|███████▎  | 275/376 [25:18<08:47,  5.22s/it]
Processing batches:  73%|███████▎  | 276/376 [25:24<08:56,  5.36s/it]
Processing batches:  74%|███████▎  | 277/376 [25:29<08:40,  5.26s/it]
Processing batches:  74%|███████▍  | 278/376 [25:35<08:44,  5.35s/it]
Processing batches:  74%|███████▍  | 279/376 [25:38<07:49,  4.84s/it]
Processing batches:  74%|███████▍  | 280/376 [25:44<08:11,  5.12s/it]
Processing batches:  75%|███████▍  | 281/376 [25:50<08:25,  5.32s/it]
Processing batches:  75%|███████▌  | 282/376 [25:56<08:35,  5.48s/it]
Processing batches:  75%|███████▌  | 283/376 [26:01<08:23,  5.42s/it]
Processing batches:  76%|███████▌  | 284/376 [26:06<07:59,  5.22s/it]
Processing batches:  76%|███████▌  | 285/376 [26:10<07:17,  4.81s/it]
Processing batches:  76%|███████▌  | 286/376 [26:10<05:20,  3.56s/it]
Processing batches:  76%|███████▋  | 287/376 [26:17<06:29,  4.38s/it]
Processing batches:  77%|███████▋  | 288/376 [26:22<07:03,  4.81s/it]
Processing batches:  77%|███████▋  | 289/376 [26:28<07:24,  5.11s/it]
Processing batches:  77%|███████▋  | 290/376 [26:34<07:37,  5.32s/it]
Processing batches:  77%|███████▋  | 291/376 [26:40<07:41,  5.43s/it]
Processing batches:  78%|███████▊  | 292/376 [26:45<07:46,  5.56s/it]
Processing batches:  78%|███████▊  | 293/376 [26:51<07:48,  5.64s/it]
Processing batches:  78%|███████▊  | 294/376 [26:57<07:46,  5.69s/it]
Processing batches:  78%|███████▊  | 295/376 [27:01<07:04,  5.24s/it]
Processing batches:  79%|███████▊  | 296/376 [27:07<07:14,  5.43s/it]
Processing batches:  79%|███████▉  | 297/376 [27:13<07:12,  5.47s/it]
Processing batches:  79%|███████▉  | 298/376 [27:18<07:07,  5.48s/it]
Processing batches:  80%|███████▉  | 299/376 [27:23<06:35,  5.14s/it]
Processing batches:  80%|███████▉  | 300/376 [27:27<06:23,  5.04s/it]
Processing batches:  80%|████████  | 301/376 [27:33<06:35,  5.27s/it]
Processing batches:  80%|████████  | 302/376 [27:39<06:35,  5.35s/it]
Processing batches:  81%|████████  | 303/376 [27:44<06:36,  5.43s/it]
Processing batches:  81%|████████  | 304/376 [27:50<06:37,  5.52s/it]
Processing batches:  81%|████████  | 305/376 [27:55<06:27,  5.46s/it]
Processing batches:  81%|████████▏ | 306/376 [28:01<06:18,  5.40s/it]
Processing batches:  82%|████████▏ | 307/376 [28:06<06:17,  5.48s/it]
Processing batches:  82%|████████▏ | 308/376 [28:12<06:16,  5.54s/it]
Processing batches:  82%|████████▏ | 309/376 [28:17<05:51,  5.25s/it]
Processing batches:  82%|████████▏ | 310/376 [28:22<05:56,  5.40s/it]
Processing batches:  83%|████████▎ | 311/376 [28:28<05:58,  5.51s/it]
Processing batches:  83%|████████▎ | 312/376 [28:32<05:15,  4.93s/it]
Processing batches:  83%|████████▎ | 313/376 [28:37<05:23,  5.13s/it]
Processing batches:  84%|████████▎ | 314/376 [28:43<05:29,  5.32s/it]
Processing batches:  84%|████████▍ | 315/376 [28:49<05:33,  5.46s/it]
Processing batches:  84%|████████▍ | 316/376 [28:55<05:33,  5.56s/it]
Processing batches:  84%|████████▍ | 317/376 [29:00<05:32,  5.64s/it]
Processing batches:  85%|████████▍ | 318/376 [29:06<05:27,  5.64s/it]
Processing batches:  85%|████████▍ | 319/376 [29:12<05:25,  5.70s/it]
Processing batches:  85%|████████▌ | 320/376 [29:17<05:15,  5.63s/it]
Processing batches:  85%|████████▌ | 321/376 [29:23<05:03,  5.53s/it]
Processing batches:  86%|████████▌ | 322/376 [29:28<04:50,  5.39s/it]
Processing batches:  86%|████████▌ | 323/376 [29:32<04:32,  5.15s/it]
Processing batches:  86%|████████▌ | 324/376 [29:38<04:35,  5.30s/it]
Processing batches:  86%|████████▋ | 325/376 [29:44<04:38,  5.46s/it]
Processing batches:  87%|████████▋ | 326/376 [29:48<04:13,  5.07s/it]
Processing batches:  87%|████████▋ | 327/376 [29:54<04:17,  5.26s/it]
Processing batches:  87%|████████▋ | 328/376 [29:59<04:17,  5.35s/it]
Processing batches:  88%|████████▊ | 329/376 [30:05<04:17,  5.47s/it]
Processing batches:  88%|████████▊ | 330/376 [30:11<04:13,  5.51s/it]
Processing batches:  88%|████████▊ | 331/376 [30:14<03:39,  4.87s/it]
Processing batches:  88%|████████▊ | 332/376 [30:20<03:47,  5.18s/it]
Processing batches:  89%|████████▊ | 333/376 [30:23<03:18,  4.62s/it]
Processing batches:  89%|████████▉ | 334/376 [30:29<03:29,  4.98s/it]
Processing batches:  89%|████████▉ | 335/376 [30:35<03:32,  5.17s/it]
Processing batches:  89%|████████▉ | 336/376 [30:40<03:31,  5.29s/it]
Processing batches:  90%|████████▉ | 337/376 [30:46<03:29,  5.38s/it]
Processing batches:  90%|████████▉ | 338/376 [30:50<03:16,  5.17s/it]
Processing batches:  90%|█████████ | 339/376 [30:54<02:56,  4.78s/it]
Processing batches:  90%|█████████ | 340/376 [30:59<02:45,  4.60s/it]
Processing batches:  91%|█████████ | 341/376 [31:04<02:52,  4.92s/it]
Processing batches:  91%|█████████ | 342/376 [31:10<02:54,  5.12s/it]
Processing batches:  91%|█████████ | 343/376 [31:15<02:52,  5.24s/it]
Processing batches:  91%|█████████▏| 344/376 [31:19<02:34,  4.83s/it]
Processing batches:  92%|█████████▏| 345/376 [31:24<02:28,  4.78s/it]
Processing batches:  92%|█████████▏| 346/376 [31:29<02:28,  4.93s/it]
Processing batches:  92%|█████████▏| 347/376 [31:33<02:14,  4.62s/it]
Processing batches:  93%|█████████▎| 348/376 [31:38<02:10,  4.66s/it]
Processing batches:  93%|█████████▎| 349/376 [31:43<02:14,  4.98s/it]
Processing batches:  93%|█████████▎| 350/376 [31:49<02:14,  5.16s/it]
Processing batches:  93%|█████████▎| 351/376 [31:55<02:13,  5.32s/it]
Processing batches:  94%|█████████▎| 352/376 [32:01<02:11,  5.50s/it]
Processing batches:  94%|█████████▍| 353/376 [32:05<01:58,  5.16s/it]
Processing batches:  94%|█████████▍| 354/376 [32:11<01:55,  5.25s/it]
Processing batches:  94%|█████████▍| 355/376 [32:15<01:42,  4.89s/it]
Processing batches:  95%|█████████▍| 356/376 [32:20<01:40,  5.01s/it]
Processing batches:  95%|█████████▍| 357/376 [32:24<01:31,  4.82s/it]
Processing batches:  95%|█████████▌| 358/376 [32:30<01:33,  5.18s/it]
Processing batches:  95%|█████████▌| 359/376 [32:34<01:20,  4.74s/it]
Processing batches:  96%|█████████▌| 360/376 [32:37<01:09,  4.37s/it]
Processing batches:  96%|█████████▌| 361/376 [32:42<01:04,  4.28s/it]
Processing batches:  96%|█████████▋| 362/376 [32:46<01:00,  4.32s/it]
Processing batches:  97%|█████████▋| 363/376 [32:49<00:51,  3.92s/it]
Processing batches:  97%|█████████▋| 364/376 [32:54<00:51,  4.25s/it]
Processing batches:  97%|█████████▋| 365/376 [32:58<00:47,  4.31s/it]
Processing batches:  97%|█████████▋| 366/376 [33:02<00:41,  4.16s/it]
Processing batches:  98%|█████████▊| 367/376 [33:06<00:35,  3.99s/it]
Processing batches:  98%|█████████▊| 368/376 [33:09<00:30,  3.85s/it]
Processing batches:  98%|█████████▊| 369/376 [33:13<00:26,  3.75s/it]
Processing batches:  98%|█████████▊| 370/376 [33:16<00:22,  3.70s/it]
Processing batches:  99%|█████████▊| 371/376 [33:20<00:18,  3.78s/it]
Processing batches:  99%|█████████▉| 372/376 [33:25<00:16,  4.01s/it]
Processing batches:  99%|█████████▉| 373/376 [33:29<00:11,  3.88s/it]
Processing batches:  99%|█████████▉| 374/376 [33:32<00:07,  3.88s/it]
Processing batches: 100%|█████████▉| 375/376 [33:36<00:03,  3.78s/it]
Processing batches: 100%|██████████| 376/376 [33:37<00:00,  3.04s/it]
Processing batches: 100%|██████████| 376/376 [33:37<00:00,  5.37s/it]
articles_mat: (768, 3753)
authors_mat: (768, 720)
words_mat: (768, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1672

--------- Desc Scores ---------
neighbors_articles_authors_det

0      Natalia V. Andrienko : 0.51
1      Gennady L. Andrienko : 0.49
2            Bernhard Preim : 0.46
3           Attila Gyulassy : 0.45
4         Martin Kraus 0001 : 0.40
                  ...
715           Markus Rütten : 0.10
716            Jiawan Zhang : 0.10
717             Ayan Biswas : 0.10
718        Tera Marie Green : 0.10
719          T. Todd Elvins : 0.10
Length: 720, dtype: object

VALUE : 0.1672

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.147826
Dieter Schmalstieg     0.111111
Matthew Kay 0001       0.310000
Xingbo Wang 0001       0.133333
Minfeng Zhu            0.100000
                         ...
Lisa M. Sobierajski    0.150000
Nahum D. Gershon       0.140000
Sidney W. Wang         0.100000
David A. Lane          0.300000
T. Todd Elvins         0.100000
Length: 720, dtype: float64

Scincl projection

scincl_matrices = bert_projection([articles_mat, authors_mat, words_mat], df['text'], family="scincl")

""
scincl_matrices = list(map(normalize_l2, scincl_matrices))

""
print(f"articles_mat: {scincl_matrices[0].shape}")
print(f"authors_mat: {scincl_matrices[1].shape}")
print(f"words_mat: {scincl_matrices[2].shape}")

""
scincl_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=scincl_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
scincl_score.print()

Using torch device: cpu

Processing batches:   0%|          | 0/376 [00:00<?, ?it/s]
Processing batches:   0%|          | 1/376 [00:05<36:01,  5.77s/it]
Processing batches:   1%|          | 2/376 [00:11<35:27,  5.69s/it]
Processing batches:   1%|          | 3/376 [00:16<34:42,  5.58s/it]
Processing batches:   1%|          | 4/376 [00:22<34:19,  5.54s/it]
Processing batches:   1%|▏         | 5/376 [00:27<33:58,  5.49s/it]
Processing batches:   2%|▏         | 6/376 [00:33<33:46,  5.48s/it]
Processing batches:   2%|▏         | 7/376 [00:38<33:41,  5.48s/it]
Processing batches:   2%|▏         | 8/376 [00:44<33:35,  5.48s/it]
Processing batches:   2%|▏         | 9/376 [00:49<33:35,  5.49s/it]
Processing batches:   3%|▎         | 10/376 [00:55<33:23,  5.48s/it]
Processing batches:   3%|▎         | 11/376 [01:00<33:26,  5.50s/it]
Processing batches:   3%|▎         | 12/376 [01:06<33:30,  5.52s/it]
Processing batches:   3%|▎         | 13/376 [01:11<33:26,  5.53s/it]
Processing batches:   4%|▎         | 14/376 [01:17<33:12,  5.50s/it]
Processing batches:   4%|▍         | 15/376 [01:22<32:55,  5.47s/it]
Processing batches:   4%|▍         | 16/376 [01:27<32:35,  5.43s/it]
Processing batches:   5%|▍         | 17/376 [01:33<32:29,  5.43s/it]
Processing batches:   5%|▍         | 18/376 [01:38<32:24,  5.43s/it]
Processing batches:   5%|▌         | 19/376 [01:44<32:24,  5.45s/it]
Processing batches:   5%|▌         | 20/376 [01:49<32:22,  5.46s/it]
Processing batches:   6%|▌         | 21/376 [01:55<32:13,  5.45s/it]
Processing batches:   6%|▌         | 22/376 [02:00<32:16,  5.47s/it]
Processing batches:   6%|▌         | 23/376 [02:06<32:11,  5.47s/it]
Processing batches:   6%|▋         | 24/376 [02:11<31:55,  5.44s/it]
Processing batches:   7%|▋         | 25/376 [02:16<31:36,  5.40s/it]
Processing batches:   7%|▋         | 26/376 [02:22<31:41,  5.43s/it]
Processing batches:   7%|▋         | 27/376 [02:27<31:39,  5.44s/it]
Processing batches:   7%|▋         | 28/376 [02:33<31:37,  5.45s/it]
Processing batches:   8%|▊         | 29/376 [02:38<31:36,  5.47s/it]
Processing batches:   8%|▊         | 30/376 [02:44<31:32,  5.47s/it]
Processing batches:   8%|▊         | 31/376 [02:49<31:20,  5.45s/it]
Processing batches:   9%|▊         | 32/376 [02:55<31:15,  5.45s/it]
Processing batches:   9%|▉         | 33/376 [03:00<31:14,  5.47s/it]
Processing batches:   9%|▉         | 34/376 [03:06<31:10,  5.47s/it]
Processing batches:   9%|▉         | 35/376 [03:11<31:37,  5.57s/it]
Processing batches:  10%|▉         | 36/376 [03:17<31:21,  5.53s/it]
Processing batches:  10%|▉         | 37/376 [03:22<31:10,  5.52s/it]
Processing batches:  10%|█         | 38/376 [03:28<31:00,  5.50s/it]
Processing batches:  10%|█         | 39/376 [03:33<31:03,  5.53s/it]
Processing batches:  11%|█         | 40/376 [03:39<30:43,  5.49s/it]
Processing batches:  11%|█         | 41/376 [03:44<30:39,  5.49s/it]
Processing batches:  11%|█         | 42/376 [03:50<30:33,  5.49s/it]
Processing batches:  11%|█▏        | 43/376 [03:55<30:36,  5.52s/it]
Processing batches:  12%|█▏        | 44/376 [04:01<30:47,  5.56s/it]
Processing batches:  12%|█▏        | 45/376 [04:07<30:39,  5.56s/it]
Processing batches:  12%|█▏        | 46/376 [04:12<30:44,  5.59s/it]
Processing batches:  12%|█▎        | 47/376 [04:18<30:33,  5.57s/it]
Processing batches:  13%|█▎        | 48/376 [04:23<30:16,  5.54s/it]
Processing batches:  13%|█▎        | 49/376 [04:29<29:58,  5.50s/it]
Processing batches:  13%|█▎        | 50/376 [04:34<29:47,  5.48s/it]
Processing batches:  14%|█▎        | 51/376 [04:39<29:29,  5.44s/it]
Processing batches:  14%|█▍        | 52/376 [04:45<29:10,  5.40s/it]
Processing batches:  14%|█▍        | 53/376 [04:50<29:15,  5.43s/it]
Processing batches:  14%|█▍        | 54/376 [04:56<29:10,  5.44s/it]
Processing batches:  15%|█▍        | 55/376 [05:01<29:11,  5.46s/it]
Processing batches:  15%|█▍        | 56/376 [05:07<28:58,  5.43s/it]
Processing batches:  15%|█▌        | 57/376 [05:12<28:55,  5.44s/it]
Processing batches:  15%|█▌        | 58/376 [05:17<28:49,  5.44s/it]
Processing batches:  16%|█▌        | 59/376 [05:23<28:54,  5.47s/it]
Processing batches:  16%|█▌        | 60/376 [05:29<28:48,  5.47s/it]
Processing batches:  16%|█▌        | 61/376 [05:34<28:29,  5.43s/it]
Processing batches:  16%|█▋        | 62/376 [05:39<28:34,  5.46s/it]
Processing batches:  17%|█▋        | 63/376 [05:45<28:31,  5.47s/it]
Processing batches:  17%|█▋        | 64/376 [05:50<28:28,  5.48s/it]
Processing batches:  17%|█▋        | 65/376 [05:56<28:31,  5.50s/it]
Processing batches:  18%|█▊        | 66/376 [06:01<28:24,  5.50s/it]
Processing batches:  18%|█▊        | 67/376 [06:07<28:21,  5.50s/it]
Processing batches:  18%|█▊        | 68/376 [06:12<28:14,  5.50s/it]
Processing batches:  18%|█▊        | 69/376 [06:18<28:05,  5.49s/it]
Processing batches:  19%|█▊        | 70/376 [06:23<27:56,  5.48s/it]
Processing batches:  19%|█▉        | 71/376 [06:29<27:49,  5.47s/it]
Processing batches:  19%|█▉        | 72/376 [06:34<27:46,  5.48s/it]
Processing batches:  19%|█▉        | 73/376 [06:40<27:39,  5.48s/it]
Processing batches:  20%|█▉        | 74/376 [06:45<27:33,  5.47s/it]
Processing batches:  20%|█▉        | 75/376 [06:51<27:20,  5.45s/it]
Processing batches:  20%|██        | 76/376 [06:56<27:17,  5.46s/it]
Processing batches:  20%|██        | 77/376 [07:02<27:15,  5.47s/it]
Processing batches:  21%|██        | 78/376 [07:07<27:14,  5.48s/it]
Processing batches:  21%|██        | 79/376 [07:13<27:13,  5.50s/it]
Processing batches:  21%|██▏       | 80/376 [07:18<27:09,  5.50s/it]
Processing batches:  22%|██▏       | 81/376 [07:24<26:56,  5.48s/it]
Processing batches:  22%|██▏       | 82/376 [07:29<26:49,  5.48s/it]
Processing batches:  22%|██▏       | 83/376 [07:34<26:39,  5.46s/it]
Processing batches:  22%|██▏       | 84/376 [07:40<26:41,  5.49s/it]
Processing batches:  23%|██▎       | 85/376 [07:46<26:38,  5.49s/it]
Processing batches:  23%|██▎       | 86/376 [07:51<26:40,  5.52s/it]
Processing batches:  23%|██▎       | 87/376 [07:57<26:36,  5.52s/it]
Processing batches:  23%|██▎       | 88/376 [08:02<26:27,  5.51s/it]
Processing batches:  24%|██▎       | 89/376 [08:08<26:15,  5.49s/it]
Processing batches:  24%|██▍       | 90/376 [08:13<26:07,  5.48s/it]
Processing batches:  24%|██▍       | 91/376 [08:19<26:01,  5.48s/it]
Processing batches:  24%|██▍       | 92/376 [08:24<25:56,  5.48s/it]
Processing batches:  25%|██▍       | 93/376 [08:29<25:50,  5.48s/it]
Processing batches:  25%|██▌       | 94/376 [08:35<25:40,  5.46s/it]
Processing batches:  25%|██▌       | 95/376 [08:40<25:42,  5.49s/it]
Processing batches:  26%|██▌       | 96/376 [08:46<25:38,  5.49s/it]
Processing batches:  26%|██▌       | 97/376 [08:51<25:22,  5.46s/it]
Processing batches:  26%|██▌       | 98/376 [08:57<25:09,  5.43s/it]
Processing batches:  26%|██▋       | 99/376 [09:02<24:50,  5.38s/it]
Processing batches:  27%|██▋       | 100/376 [09:07<24:34,  5.34s/it]
Processing batches:  27%|██▋       | 101/376 [09:13<24:37,  5.37s/it]
Processing batches:  27%|██▋       | 102/376 [09:18<24:42,  5.41s/it]
Processing batches:  27%|██▋       | 103/376 [09:24<24:37,  5.41s/it]
Processing batches:  28%|██▊       | 104/376 [09:29<24:46,  5.47s/it]
Processing batches:  28%|██▊       | 105/376 [09:35<24:33,  5.44s/it]
Processing batches:  28%|██▊       | 106/376 [09:40<24:32,  5.45s/it]
Processing batches:  28%|██▊       | 107/376 [09:45<24:19,  5.43s/it]
Processing batches:  29%|██▊       | 108/376 [09:51<24:23,  5.46s/it]
Processing batches:  29%|██▉       | 109/376 [09:57<24:33,  5.52s/it]
Processing batches:  29%|██▉       | 110/376 [10:02<24:33,  5.54s/it]
Processing batches:  30%|██▉       | 111/376 [10:08<24:23,  5.52s/it]
Processing batches:  30%|██▉       | 112/376 [10:13<24:16,  5.52s/it]
Processing batches:  30%|███       | 113/376 [10:19<24:05,  5.49s/it]
Processing batches:  30%|███       | 114/376 [10:24<23:59,  5.49s/it]
Processing batches:  31%|███       | 115/376 [10:30<23:54,  5.50s/it]
Processing batches:  31%|███       | 116/376 [10:35<23:56,  5.52s/it]
Processing batches:  31%|███       | 117/376 [10:40<23:28,  5.44s/it]
Processing batches:  31%|███▏      | 118/376 [10:46<23:14,  5.40s/it]
Processing batches:  32%|███▏      | 119/376 [10:51<22:59,  5.37s/it]
Processing batches:  32%|███▏      | 120/376 [10:56<22:50,  5.35s/it]
Processing batches:  32%|███▏      | 121/376 [11:02<22:39,  5.33s/it]
Processing batches:  32%|███▏      | 122/376 [11:07<22:29,  5.31s/it]
Processing batches:  33%|███▎      | 123/376 [11:12<22:25,  5.32s/it]
Processing batches:  33%|███▎      | 124/376 [11:18<22:22,  5.33s/it]
Processing batches:  33%|███▎      | 125/376 [11:23<22:38,  5.41s/it]
Processing batches:  34%|███▎      | 126/376 [11:29<22:41,  5.44s/it]
Processing batches:  34%|███▍      | 127/376 [11:34<22:41,  5.47s/it]
Processing batches:  34%|███▍      | 128/376 [11:40<22:39,  5.48s/it]
Processing batches:  34%|███▍      | 129/376 [11:45<22:36,  5.49s/it]
Processing batches:  35%|███▍      | 130/376 [11:51<22:34,  5.51s/it]
Processing batches:  35%|███▍      | 131/376 [11:56<22:27,  5.50s/it]
Processing batches:  35%|███▌      | 132/376 [12:02<22:10,  5.45s/it]
Processing batches:  35%|███▌      | 133/376 [12:07<22:03,  5.45s/it]
Processing batches:  36%|███▌      | 134/376 [12:12<21:54,  5.43s/it]
Processing batches:  36%|███▌      | 135/376 [12:18<21:50,  5.44s/it]
Processing batches:  36%|███▌      | 136/376 [12:23<21:39,  5.42s/it]
Processing batches:  36%|███▋      | 137/376 [12:29<21:29,  5.39s/it]
Processing batches:  37%|███▋      | 138/376 [12:34<21:15,  5.36s/it]
Processing batches:  37%|███▋      | 139/376 [12:39<21:15,  5.38s/it]
Processing batches:  37%|███▋      | 140/376 [12:45<21:09,  5.38s/it]
Processing batches:  38%|███▊      | 141/376 [12:50<21:10,  5.41s/it]
Processing batches:  38%|███▊      | 142/376 [12:56<21:04,  5.40s/it]
Processing batches:  38%|███▊      | 143/376 [13:01<21:04,  5.43s/it]
Processing batches:  38%|███▊      | 144/376 [13:07<21:04,  5.45s/it]
Processing batches:  39%|███▊      | 145/376 [13:12<21:00,  5.46s/it]
Processing batches:  39%|███▉      | 146/376 [13:17<20:51,  5.44s/it]
Processing batches:  39%|███▉      | 147/376 [13:23<20:41,  5.42s/it]
Processing batches:  39%|███▉      | 148/376 [13:28<20:25,  5.38s/it]
Processing batches:  40%|███▉      | 149/376 [13:33<20:22,  5.39s/it]
Processing batches:  40%|███▉      | 150/376 [13:39<20:21,  5.41s/it]
Processing batches:  40%|████      | 151/376 [13:44<19:58,  5.33s/it]
Processing batches:  40%|████      | 152/376 [13:50<20:05,  5.38s/it]
Processing batches:  41%|████      | 153/376 [13:55<20:06,  5.41s/it]
Processing batches:  41%|████      | 154/376 [14:01<20:11,  5.46s/it]
Processing batches:  41%|████      | 155/376 [14:06<19:56,  5.41s/it]
Processing batches:  41%|████▏     | 156/376 [14:11<19:50,  5.41s/it]
Processing batches:  42%|████▏     | 157/376 [14:17<19:41,  5.39s/it]
Processing batches:  42%|████▏     | 158/376 [14:22<19:40,  5.41s/it]
Processing batches:  42%|████▏     | 159/376 [14:28<19:40,  5.44s/it]
Processing batches:  43%|████▎     | 160/376 [14:33<19:40,  5.47s/it]
Processing batches:  43%|████▎     | 161/376 [14:39<19:40,  5.49s/it]
Processing batches:  43%|████▎     | 162/376 [14:44<19:48,  5.55s/it]
Processing batches:  43%|████▎     | 163/376 [14:50<19:47,  5.57s/it]
Processing batches:  44%|████▎     | 164/376 [14:56<19:41,  5.58s/it]
Processing batches:  44%|████▍     | 165/376 [15:01<19:22,  5.51s/it]
Processing batches:  44%|████▍     | 166/376 [15:06<19:02,  5.44s/it]
Processing batches:  44%|████▍     | 167/376 [15:12<18:45,  5.38s/it]
Processing batches:  45%|████▍     | 168/376 [15:17<18:43,  5.40s/it]
Processing batches:  45%|████▍     | 169/376 [15:22<18:41,  5.42s/it]
Processing batches:  45%|████▌     | 170/376 [15:28<18:36,  5.42s/it]
Processing batches:  45%|████▌     | 171/376 [15:33<18:35,  5.44s/it]
Processing batches:  46%|████▌     | 172/376 [15:39<18:26,  5.42s/it]
Processing batches:  46%|████▌     | 173/376 [15:44<18:39,  5.52s/it]
Processing batches:  46%|████▋     | 174/376 [15:50<18:29,  5.49s/it]
Processing batches:  47%|████▋     | 175/376 [15:55<18:17,  5.46s/it]
Processing batches:  47%|████▋     | 176/376 [15:59<16:13,  4.87s/it]
Processing batches:  47%|████▋     | 177/376 [16:04<16:50,  5.08s/it]
Processing batches:  47%|████▋     | 178/376 [16:10<17:05,  5.18s/it]
Processing batches:  48%|████▊     | 179/376 [16:15<17:19,  5.27s/it]
Processing batches:  48%|████▊     | 180/376 [16:21<17:23,  5.33s/it]
Processing batches:  48%|████▊     | 181/376 [16:26<17:25,  5.36s/it]
Processing batches:  48%|████▊     | 182/376 [16:32<17:27,  5.40s/it]
Processing batches:  49%|████▊     | 183/376 [16:37<17:31,  5.45s/it]
Processing batches:  49%|████▉     | 184/376 [16:43<17:24,  5.44s/it]
Processing batches:  49%|████▉     | 185/376 [16:48<17:18,  5.44s/it]
Processing batches:  49%|████▉     | 186/376 [16:53<17:14,  5.44s/it]
Processing batches:  50%|████▉     | 187/376 [16:59<17:01,  5.41s/it]
Processing batches:  50%|█████     | 188/376 [17:04<17:00,  5.43s/it]
Processing batches:  50%|█████     | 189/376 [17:10<16:54,  5.43s/it]
Processing batches:  51%|█████     | 190/376 [17:15<16:58,  5.47s/it]
Processing batches:  51%|█████     | 191/376 [17:21<17:05,  5.55s/it]
Processing batches:  51%|█████     | 192/376 [17:26<16:53,  5.51s/it]
Processing batches:  51%|█████▏    | 193/376 [17:32<16:33,  5.43s/it]
Processing batches:  52%|█████▏    | 194/376 [17:37<16:29,  5.44s/it]
Processing batches:  52%|█████▏    | 195/376 [17:43<16:24,  5.44s/it]
Processing batches:  52%|█████▏    | 196/376 [17:48<16:26,  5.48s/it]
Processing batches:  52%|█████▏    | 197/376 [17:54<16:24,  5.50s/it]
Processing batches:  53%|█████▎    | 198/376 [17:59<16:18,  5.50s/it]
Processing batches:  53%|█████▎    | 199/376 [18:05<16:13,  5.50s/it]
Processing batches:  53%|█████▎    | 200/376 [18:10<15:59,  5.45s/it]
Processing batches:  53%|█████▎    | 201/376 [18:15<15:49,  5.42s/it]
Processing batches:  54%|█████▎    | 202/376 [18:20<15:03,  5.19s/it]
Processing batches:  54%|█████▍    | 203/376 [18:26<15:23,  5.34s/it]
Processing batches:  54%|█████▍    | 204/376 [18:31<15:30,  5.41s/it]
Processing batches:  55%|█████▍    | 205/376 [18:37<15:23,  5.40s/it]
Processing batches:  55%|█████▍    | 206/376 [18:42<15:15,  5.39s/it]
Processing batches:  55%|█████▌    | 207/376 [18:47<15:13,  5.40s/it]
Processing batches:  55%|█████▌    | 208/376 [18:53<15:12,  5.43s/it]
Processing batches:  56%|█████▌    | 209/376 [18:58<15:06,  5.43s/it]
Processing batches:  56%|█████▌    | 210/376 [19:04<15:00,  5.42s/it]
Processing batches:  56%|█████▌    | 211/376 [19:09<14:56,  5.43s/it]
Processing batches:  56%|█████▋    | 212/376 [19:15<14:55,  5.46s/it]
Processing batches:  57%|█████▋    | 213/376 [19:20<14:50,  5.47s/it]
Processing batches:  57%|█████▋    | 214/376 [19:26<14:51,  5.50s/it]
Processing batches:  57%|█████▋    | 215/376 [19:30<13:59,  5.22s/it]
Processing batches:  57%|█████▋    | 216/376 [19:35<13:48,  5.18s/it]
Processing batches:  58%|█████▊    | 217/376 [19:41<14:00,  5.29s/it]
Processing batches:  58%|█████▊    | 218/376 [19:46<14:01,  5.33s/it]
Processing batches:  58%|█████▊    | 219/376 [19:52<14:01,  5.36s/it]
Processing batches:  59%|█████▊    | 220/376 [19:57<14:06,  5.43s/it]
Processing batches:  59%|█████▉    | 221/376 [20:03<13:59,  5.42s/it]
Processing batches:  59%|█████▉    | 222/376 [20:08<13:59,  5.45s/it]
Processing batches:  59%|█████▉    | 223/376 [20:14<13:52,  5.44s/it]
Processing batches:  60%|█████▉    | 224/376 [20:19<13:41,  5.40s/it]
Processing batches:  60%|█████▉    | 225/376 [20:24<13:28,  5.35s/it]
Processing batches:  60%|██████    | 226/376 [20:30<13:19,  5.33s/it]
Processing batches:  60%|██████    | 227/376 [20:35<13:14,  5.33s/it]
Processing batches:  61%|██████    | 228/376 [20:40<13:08,  5.33s/it]
Processing batches:  61%|██████    | 229/376 [20:46<13:01,  5.31s/it]
Processing batches:  61%|██████    | 230/376 [20:51<12:57,  5.32s/it]
Processing batches:  61%|██████▏   | 231/376 [20:56<12:57,  5.36s/it]
Processing batches:  62%|██████▏   | 232/376 [21:02<12:53,  5.37s/it]
Processing batches:  62%|██████▏   | 233/376 [21:07<12:51,  5.40s/it]
Processing batches:  62%|██████▏   | 234/376 [21:12<12:08,  5.13s/it]
Processing batches:  62%|██████▎   | 235/376 [21:17<12:21,  5.26s/it]
Processing batches:  63%|██████▎   | 236/376 [21:23<12:28,  5.35s/it]
Processing batches:  63%|██████▎   | 237/376 [21:28<12:27,  5.38s/it]
Processing batches:  63%|██████▎   | 238/376 [21:34<12:24,  5.39s/it]
Processing batches:  64%|██████▎   | 239/376 [21:39<12:26,  5.45s/it]
Processing batches:  64%|██████▍   | 240/376 [21:45<12:18,  5.43s/it]
Processing batches:  64%|██████▍   | 241/376 [21:50<12:09,  5.41s/it]
Processing batches:  64%|██████▍   | 242/376 [21:55<12:02,  5.39s/it]
Processing batches:  65%|██████▍   | 243/376 [22:01<12:01,  5.42s/it]
Processing batches:  65%|██████▍   | 244/376 [22:06<11:56,  5.43s/it]
Processing batches:  65%|██████▌   | 245/376 [22:12<11:48,  5.41s/it]
Processing batches:  65%|██████▌   | 246/376 [22:17<11:41,  5.40s/it]
Processing batches:  66%|██████▌   | 247/376 [22:22<11:37,  5.41s/it]
Processing batches:  66%|██████▌   | 248/376 [22:28<11:33,  5.42s/it]
Processing batches:  66%|██████▌   | 249/376 [22:32<10:21,  4.89s/it]
Processing batches:  66%|██████▋   | 250/376 [22:36<10:10,  4.85s/it]
Processing batches:  67%|██████▋   | 251/376 [22:40<09:39,  4.64s/it]
Processing batches:  67%|██████▋   | 252/376 [22:46<10:15,  4.97s/it]
Processing batches:  67%|██████▋   | 253/376 [22:52<10:26,  5.09s/it]
Processing batches:  68%|██████▊   | 254/376 [22:57<10:28,  5.15s/it]
Processing batches:  68%|██████▊   | 255/376 [23:02<10:30,  5.21s/it]
Processing batches:  68%|██████▊   | 256/376 [23:08<10:34,  5.29s/it]
Processing batches:  68%|██████▊   | 257/376 [23:13<10:30,  5.30s/it]
Processing batches:  69%|██████▊   | 258/376 [23:19<10:39,  5.42s/it]
Processing batches:  69%|██████▉   | 259/376 [23:24<10:42,  5.49s/it]
Processing batches:  69%|██████▉   | 260/376 [23:30<10:44,  5.55s/it]
Processing batches:  69%|██████▉   | 261/376 [23:35<10:33,  5.51s/it]
Processing batches:  70%|██████▉   | 262/376 [23:41<10:40,  5.62s/it]
Processing batches:  70%|██████▉   | 263/376 [23:47<10:36,  5.64s/it]
Processing batches:  70%|███████   | 264/376 [23:53<10:27,  5.60s/it]
Processing batches:  70%|███████   | 265/376 [23:58<10:25,  5.64s/it]
Processing batches:  71%|███████   | 266/376 [24:04<10:13,  5.58s/it]
Processing batches:  71%|███████   | 267/376 [24:09<09:57,  5.48s/it]
Processing batches:  71%|███████▏  | 268/376 [24:14<09:33,  5.31s/it]
Processing batches:  72%|███████▏  | 269/376 [24:20<09:37,  5.40s/it]
Processing batches:  72%|███████▏  | 270/376 [24:25<09:31,  5.39s/it]
Processing batches:  72%|███████▏  | 271/376 [24:31<09:34,  5.47s/it]
Processing batches:  72%|███████▏  | 272/376 [24:36<09:27,  5.46s/it]
Processing batches:  73%|███████▎  | 273/376 [24:41<09:20,  5.44s/it]
Processing batches:  73%|███████▎  | 274/376 [24:47<09:23,  5.53s/it]
Processing batches:  73%|███████▎  | 275/376 [24:51<08:29,  5.05s/it]
Processing batches:  73%|███████▎  | 276/376 [24:57<08:37,  5.18s/it]
Processing batches:  74%|███████▎  | 277/376 [25:01<08:23,  5.08s/it]
Processing batches:  74%|███████▍  | 278/376 [25:07<08:29,  5.20s/it]
Processing batches:  74%|███████▍  | 279/376 [25:10<07:37,  4.72s/it]
Processing batches:  74%|███████▍  | 280/376 [25:16<08:00,  5.01s/it]
Processing batches:  75%|███████▍  | 281/376 [25:22<08:08,  5.14s/it]
Processing batches:  75%|███████▌  | 282/376 [25:27<08:17,  5.30s/it]
Processing batches:  75%|███████▌  | 283/376 [25:32<08:10,  5.27s/it]
Processing batches:  76%|███████▌  | 284/376 [25:37<07:48,  5.09s/it]
Processing batches:  76%|███████▌  | 285/376 [25:41<07:08,  4.71s/it]
Processing batches:  76%|███████▌  | 286/376 [25:42<05:14,  3.49s/it]
Processing batches:  76%|███████▋  | 287/376 [25:48<06:15,  4.22s/it]
Processing batches:  77%|███████▋  | 288/376 [25:53<06:46,  4.61s/it]
Processing batches:  77%|███████▋  | 289/376 [25:58<06:59,  4.82s/it]
Processing batches:  77%|███████▋  | 290/376 [26:04<07:14,  5.06s/it]
Processing batches:  77%|███████▋  | 291/376 [26:09<07:20,  5.19s/it]
Processing batches:  78%|███████▊  | 292/376 [26:15<07:33,  5.40s/it]
Processing batches:  78%|███████▊  | 293/376 [26:21<07:30,  5.42s/it]
Processing batches:  78%|███████▊  | 294/376 [26:27<07:32,  5.52s/it]
Processing batches:  78%|███████▊  | 295/376 [26:31<06:53,  5.10s/it]
Processing batches:  79%|███████▊  | 296/376 [26:36<07:05,  5.31s/it]
Processing batches:  79%|███████▉  | 297/376 [26:42<07:07,  5.42s/it]
Processing batches:  79%|███████▉  | 298/376 [26:48<07:07,  5.48s/it]
Processing batches:  80%|███████▉  | 299/376 [26:52<06:33,  5.11s/it]
Processing batches:  80%|███████▉  | 300/376 [26:57<06:22,  5.04s/it]
Processing batches:  80%|████████  | 301/376 [27:02<06:29,  5.19s/it]
Processing batches:  80%|████████  | 302/376 [27:08<06:30,  5.28s/it]
Processing batches:  81%|████████  | 303/376 [27:13<06:29,  5.33s/it]
Processing batches:  81%|████████  | 304/376 [27:19<06:30,  5.42s/it]
Processing batches:  81%|████████  | 305/376 [27:24<06:20,  5.36s/it]
Processing batches:  81%|████████▏ | 306/376 [27:30<06:14,  5.35s/it]
Processing batches:  82%|████████▏ | 307/376 [27:35<06:10,  5.36s/it]
Processing batches:  82%|████████▏ | 308/376 [27:41<06:10,  5.45s/it]
Processing batches:  82%|████████▏ | 309/376 [27:45<05:44,  5.14s/it]
Processing batches:  82%|████████▏ | 310/376 [27:51<05:48,  5.28s/it]
Processing batches:  83%|████████▎ | 311/376 [27:56<05:45,  5.31s/it]
Processing batches:  83%|████████▎ | 312/376 [28:00<05:05,  4.77s/it]
Processing batches:  83%|████████▎ | 313/376 [28:05<05:16,  5.02s/it]
Processing batches:  84%|████████▎ | 314/376 [28:11<05:23,  5.22s/it]
Processing batches:  84%|████████▍ | 315/376 [28:16<05:21,  5.27s/it]
Processing batches:  84%|████████▍ | 316/376 [28:22<05:21,  5.36s/it]
Processing batches:  84%|████████▍ | 317/376 [28:27<05:17,  5.38s/it]
Processing batches:  85%|████████▍ | 318/376 [28:33<05:16,  5.46s/it]
Processing batches:  85%|████████▍ | 319/376 [28:38<05:11,  5.46s/it]
Processing batches:  85%|████████▌ | 320/376 [28:44<05:03,  5.42s/it]
Processing batches:  85%|████████▌ | 321/376 [28:49<04:55,  5.38s/it]
Processing batches:  86%|████████▌ | 322/376 [28:54<04:43,  5.24s/it]
Processing batches:  86%|████████▌ | 323/376 [28:58<04:28,  5.06s/it]
Processing batches:  86%|████████▌ | 324/376 [29:04<04:31,  5.22s/it]
Processing batches:  86%|████████▋ | 325/376 [29:10<04:29,  5.29s/it]
Processing batches:  87%|████████▋ | 326/376 [29:14<04:05,  4.91s/it]
Processing batches:  87%|████████▋ | 327/376 [29:19<04:08,  5.07s/it]
Processing batches:  87%|████████▋ | 328/376 [29:25<04:12,  5.25s/it]
Processing batches:  88%|████████▊ | 329/376 [29:30<04:11,  5.34s/it]
Processing batches:  88%|████████▊ | 330/376 [29:36<04:07,  5.38s/it]
Processing batches:  88%|████████▊ | 331/376 [29:39<03:33,  4.75s/it]
Processing batches:  88%|████████▊ | 332/376 [29:45<03:41,  5.02s/it]
Processing batches:  89%|████████▊ | 333/376 [29:48<03:13,  4.50s/it]
Processing batches:  89%|████████▉ | 334/376 [29:53<03:22,  4.82s/it]
Processing batches:  89%|████████▉ | 335/376 [29:59<03:27,  5.06s/it]
Processing batches:  89%|████████▉ | 336/376 [30:05<03:29,  5.25s/it]
Processing batches:  90%|████████▉ | 337/376 [30:11<03:30,  5.41s/it]
Processing batches:  90%|████████▉ | 338/376 [30:15<03:19,  5.26s/it]
Processing batches:  90%|█████████ | 339/376 [30:19<02:58,  4.82s/it]
Processing batches:  90%|█████████ | 340/376 [30:23<02:46,  4.61s/it]
Processing batches:  91%|█████████ | 341/376 [30:29<02:50,  4.87s/it]
Processing batches:  91%|█████████ | 342/376 [30:35<02:54,  5.14s/it]
Processing batches:  91%|█████████ | 343/376 [30:40<02:54,  5.30s/it]
Processing batches:  91%|█████████▏| 344/376 [30:44<02:34,  4.83s/it]
Processing batches:  92%|█████████▏| 345/376 [30:49<02:27,  4.75s/it]
Processing batches:  92%|█████████▏| 346/376 [30:54<02:26,  4.89s/it]
Processing batches:  92%|█████████▏| 347/376 [30:58<02:12,  4.56s/it]
Processing batches:  93%|█████████▎| 348/376 [31:02<02:09,  4.64s/it]
Processing batches:  93%|█████████▎| 349/376 [31:08<02:12,  4.93s/it]
Processing batches:  93%|█████████▎| 350/376 [31:14<02:13,  5.14s/it]
Processing batches:  93%|█████████▎| 351/376 [31:19<02:11,  5.26s/it]
Processing batches:  94%|█████████▎| 352/376 [31:25<02:07,  5.30s/it]
Processing batches:  94%|█████████▍| 353/376 [31:29<01:53,  4.94s/it]
Processing batches:  94%|█████████▍| 354/376 [31:34<01:50,  5.01s/it]
Processing batches:  94%|█████████▍| 355/376 [31:38<01:37,  4.63s/it]
Processing batches:  95%|█████████▍| 356/376 [31:43<01:35,  4.77s/it]
Processing batches:  95%|█████████▍| 357/376 [31:47<01:26,  4.57s/it]
Processing batches:  95%|█████████▌| 358/376 [31:52<01:27,  4.88s/it]
Processing batches:  95%|█████████▌| 359/376 [31:56<01:15,  4.46s/it]
Processing batches:  96%|█████████▌| 360/376 [31:59<01:06,  4.13s/it]
Processing batches:  96%|█████████▌| 361/376 [32:03<01:01,  4.07s/it]
Processing batches:  96%|█████████▋| 362/376 [32:08<00:58,  4.14s/it]
Processing batches:  97%|█████████▋| 363/376 [32:10<00:49,  3.78s/it]
Processing batches:  97%|█████████▋| 364/376 [32:16<00:50,  4.18s/it]
Processing batches:  97%|█████████▋| 365/376 [32:20<00:46,  4.25s/it]
Processing batches:  97%|█████████▋| 366/376 [32:24<00:41,  4.11s/it]
Processing batches:  98%|█████████▊| 367/376 [32:27<00:35,  3.95s/it]
Processing batches:  98%|█████████▊| 368/376 [32:31<00:30,  3.79s/it]
Processing batches:  98%|█████████▊| 369/376 [32:34<00:25,  3.69s/it]
Processing batches:  98%|█████████▊| 370/376 [32:38<00:21,  3.63s/it]
Processing batches:  99%|█████████▊| 371/376 [32:42<00:18,  3.71s/it]
Processing batches:  99%|█████████▉| 372/376 [32:46<00:15,  3.99s/it]
Processing batches:  99%|█████████▉| 373/376 [32:50<00:11,  3.85s/it]
Processing batches:  99%|█████████▉| 374/376 [32:54<00:07,  3.86s/it]
Processing batches: 100%|█████████▉| 375/376 [32:57<00:03,  3.77s/it]
Processing batches: 100%|██████████| 376/376 [32:58<00:00,  3.01s/it]
Processing batches: 100%|██████████| 376/376 [32:58<00:00,  5.26s/it]
articles_mat: (768, 3753)
authors_mat: (768, 720)
words_mat: (768, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1590

--------- Desc Scores ---------
neighbors_articles_authors_det

0      Natalia V. Andrienko : 0.52
1      Gennady L. Andrienko : 0.49
2           Attila Gyulassy : 0.42
3          Karen B. Schloss : 0.41
4            Aditi Majumder : 0.41
                  ...
715      Andrew Vande Moere : 0.10
716               Peter Bak : 0.10
717           Mark W. Jones : 0.10
718               Ji Soo Yi : 0.10
719      Morteza Karimzadeh : 0.10
Length: 720, dtype: object

VALUE : 0.1590

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.126087
Dieter Schmalstieg     0.100000
Matthew Kay 0001       0.270000
Xingbo Wang 0001       0.133333
Minfeng Zhu            0.100000
                         ...
Lisa M. Sobierajski    0.175000
Nahum D. Gershon       0.140000
Sidney W. Wang         0.100000
David A. Lane          0.216667
T. Todd Elvins         0.150000
Length: 720, dtype: float64

“all-MiniLM-L6-v2” projection

minilm_matrices = bert_projection([articles_mat, authors_mat, words_mat], df['text'], family="all-MiniLM-L6-v2")

""
minilm_matrices = list(map(normalize_l2, minilm_matrices))

""
print(f"articles_mat: {minilm_matrices[0].shape}")
print(f"authors_mat: {minilm_matrices[1].shape}")
print(f"words_mat: {minilm_matrices[2].shape}")

Processing batches:   0%|          | 0/376 [00:00<?, ?it/s]
Processing batches:   0%|          | 1/376 [00:03<20:29,  3.28s/it]
Processing batches:   1%|          | 2/376 [00:06<18:38,  2.99s/it]
Processing batches:   1%|          | 3/376 [00:08<17:05,  2.75s/it]
Processing batches:   1%|          | 4/376 [00:12<19:36,  3.16s/it]
Processing batches:   1%|▏         | 5/376 [00:14<17:13,  2.79s/it]
Processing batches:   2%|▏         | 6/376 [00:17<16:47,  2.72s/it]
Processing batches:   2%|▏         | 7/376 [00:18<15:07,  2.46s/it]
Processing batches:   2%|▏         | 8/376 [00:21<15:41,  2.56s/it]
Processing batches:   2%|▏         | 9/376 [00:24<15:50,  2.59s/it]
Processing batches:   3%|▎         | 10/376 [00:27<17:18,  2.84s/it]
Processing batches:   3%|▎         | 11/376 [00:30<17:39,  2.90s/it]
Processing batches:   3%|▎         | 12/376 [00:32<15:38,  2.58s/it]
Processing batches:   3%|▎         | 13/376 [00:35<16:39,  2.75s/it]
Processing batches:   4%|▎         | 14/376 [00:38<16:57,  2.81s/it]
Processing batches:   4%|▍         | 15/376 [00:41<16:58,  2.82s/it]
Processing batches:   4%|▍         | 16/376 [00:45<18:05,  3.02s/it]
Processing batches:   5%|▍         | 17/376 [00:47<16:21,  2.73s/it]
Processing batches:   5%|▍         | 18/376 [00:49<16:28,  2.76s/it]
Processing batches:   5%|▌         | 19/376 [00:52<16:16,  2.74s/it]
Processing batches:   5%|▌         | 20/376 [00:54<15:16,  2.57s/it]
Processing batches:   6%|▌         | 21/376 [00:56<14:08,  2.39s/it]
Processing batches:   6%|▌         | 22/376 [00:59<14:22,  2.44s/it]
Processing batches:   6%|▌         | 23/376 [01:02<14:41,  2.50s/it]
Processing batches:   6%|▋         | 24/376 [01:05<15:38,  2.67s/it]
Processing batches:   7%|▋         | 25/376 [01:08<16:37,  2.84s/it]
Processing batches:   7%|▋         | 26/376 [01:10<15:15,  2.62s/it]
Processing batches:   7%|▋         | 27/376 [01:13<15:55,  2.74s/it]
Processing batches:   7%|▋         | 28/376 [01:15<14:58,  2.58s/it]
Processing batches:   8%|▊         | 29/376 [01:18<15:49,  2.74s/it]
Processing batches:   8%|▊         | 30/376 [01:22<16:53,  2.93s/it]
Processing batches:   8%|▊         | 31/376 [01:23<14:47,  2.57s/it]
Processing batches:   9%|▊         | 32/376 [01:27<16:35,  2.89s/it]
Processing batches:   9%|▉         | 33/376 [01:31<18:29,  3.23s/it]
Processing batches:   9%|▉         | 34/376 [01:33<16:49,  2.95s/it]
Processing batches:   9%|▉         | 35/376 [01:36<16:29,  2.90s/it]
Processing batches:  10%|▉         | 36/376 [01:39<15:53,  2.80s/it]
Processing batches:  10%|▉         | 37/376 [01:41<14:44,  2.61s/it]
Processing batches:  10%|█         | 38/376 [01:43<14:12,  2.52s/it]
Processing batches:  10%|█         | 39/376 [01:45<13:18,  2.37s/it]
Processing batches:  11%|█         | 40/376 [01:47<12:50,  2.29s/it]
Processing batches:  11%|█         | 41/376 [01:49<12:29,  2.24s/it]
Processing batches:  11%|█         | 42/376 [01:52<12:53,  2.32s/it]
Processing batches:  11%|█▏        | 43/376 [01:56<15:49,  2.85s/it]
Processing batches:  12%|█▏        | 44/376 [01:58<14:10,  2.56s/it]
Processing batches:  12%|█▏        | 45/376 [02:00<13:01,  2.36s/it]
Processing batches:  12%|█▏        | 46/376 [02:02<12:34,  2.29s/it]
Processing batches:  12%|█▎        | 47/376 [02:04<12:41,  2.32s/it]
Processing batches:  13%|█▎        | 48/376 [02:06<11:56,  2.18s/it]
Processing batches:  13%|█▎        | 49/376 [02:08<11:44,  2.15s/it]
Processing batches:  13%|█▎        | 50/376 [02:11<11:55,  2.19s/it]
Processing batches:  14%|█▎        | 51/376 [02:14<13:20,  2.46s/it]
Processing batches:  14%|█▍        | 52/376 [02:16<13:21,  2.47s/it]
Processing batches:  14%|█▍        | 53/376 [02:18<12:38,  2.35s/it]
Processing batches:  14%|█▍        | 54/376 [02:21<12:49,  2.39s/it]
Processing batches:  15%|█▍        | 55/376 [02:24<15:03,  2.82s/it]
Processing batches:  15%|█▍        | 56/376 [02:27<14:18,  2.68s/it]
Processing batches:  15%|█▌        | 57/376 [02:29<12:55,  2.43s/it]
Processing batches:  15%|█▌        | 58/376 [02:32<14:52,  2.81s/it]
Processing batches:  16%|█▌        | 59/376 [02:35<14:30,  2.75s/it]
Processing batches:  16%|█▌        | 60/376 [02:37<12:56,  2.46s/it]
Processing batches:  16%|█▌        | 61/376 [02:39<12:51,  2.45s/it]
Processing batches:  16%|█▋        | 62/376 [02:41<11:54,  2.28s/it]
Processing batches:  17%|█▋        | 63/376 [02:43<11:49,  2.27s/it]
Processing batches:  17%|█▋        | 64/376 [02:46<12:00,  2.31s/it]
Processing batches:  17%|█▋        | 65/376 [02:49<12:52,  2.48s/it]
Processing batches:  18%|█▊        | 66/376 [02:51<12:58,  2.51s/it]
Processing batches:  18%|█▊        | 67/376 [02:54<13:12,  2.57s/it]
Processing batches:  18%|█▊        | 68/376 [02:56<11:55,  2.32s/it]
Processing batches:  18%|█▊        | 69/376 [02:57<11:07,  2.18s/it]
Processing batches:  19%|█▊        | 70/376 [03:00<11:13,  2.20s/it]
Processing batches:  19%|█▉        | 71/376 [03:02<11:31,  2.27s/it]
Processing batches:  19%|█▉        | 72/376 [03:06<13:44,  2.71s/it]
Processing batches:  19%|█▉        | 73/376 [03:08<12:52,  2.55s/it]
Processing batches:  20%|█▉        | 74/376 [03:11<12:51,  2.56s/it]
Processing batches:  20%|█▉        | 75/376 [03:13<12:27,  2.48s/it]
Processing batches:  20%|██        | 76/376 [03:15<11:23,  2.28s/it]
Processing batches:  20%|██        | 77/376 [03:17<11:28,  2.30s/it]
Processing batches:  21%|██        | 78/376 [03:20<12:16,  2.47s/it]
Processing batches:  21%|██        | 79/376 [03:22<11:53,  2.40s/it]
Processing batches:  21%|██▏       | 80/376 [03:24<10:57,  2.22s/it]
Processing batches:  22%|██▏       | 81/376 [03:26<10:26,  2.12s/it]
Processing batches:  22%|██▏       | 82/376 [03:29<12:08,  2.48s/it]
Processing batches:  22%|██▏       | 83/376 [03:32<11:50,  2.43s/it]
Processing batches:  22%|██▏       | 84/376 [03:35<13:59,  2.88s/it]
Processing batches:  23%|██▎       | 85/376 [03:37<12:01,  2.48s/it]
Processing batches:  23%|██▎       | 86/376 [03:41<14:28,  2.99s/it]
Processing batches:  23%|██▎       | 87/376 [03:43<12:51,  2.67s/it]
Processing batches:  23%|██▎       | 88/376 [03:45<12:07,  2.52s/it]
Processing batches:  24%|██▎       | 89/376 [03:48<12:28,  2.61s/it]
Processing batches:  24%|██▍       | 90/376 [03:51<12:55,  2.71s/it]
Processing batches:  24%|██▍       | 91/376 [03:53<12:10,  2.56s/it]
Processing batches:  24%|██▍       | 92/376 [03:57<13:52,  2.93s/it]
Processing batches:  25%|██▍       | 93/376 [03:59<11:53,  2.52s/it]
Processing batches:  25%|██▌       | 94/376 [04:01<11:25,  2.43s/it]
Processing batches:  25%|██▌       | 95/376 [04:03<11:20,  2.42s/it]
Processing batches:  26%|██▌       | 96/376 [04:06<11:58,  2.57s/it]
Processing batches:  26%|██▌       | 97/376 [04:09<12:07,  2.61s/it]
Processing batches:  26%|██▌       | 98/376 [04:11<11:07,  2.40s/it]
Processing batches:  26%|██▋       | 99/376 [04:13<10:23,  2.25s/it]
Processing batches:  27%|██▋       | 100/376 [04:15<10:19,  2.24s/it]
Processing batches:  27%|██▋       | 101/376 [04:16<09:03,  1.98s/it]
Processing batches:  27%|██▋       | 102/376 [04:18<08:51,  1.94s/it]
Processing batches:  27%|██▋       | 103/376 [04:20<09:04,  1.99s/it]
Processing batches:  28%|██▊       | 104/376 [04:22<09:15,  2.04s/it]
Processing batches:  28%|██▊       | 105/376 [04:24<09:09,  2.03s/it]
Processing batches:  28%|██▊       | 106/376 [04:27<09:21,  2.08s/it]
Processing batches:  28%|██▊       | 107/376 [04:29<09:37,  2.15s/it]
Processing batches:  29%|██▊       | 108/376 [04:31<08:57,  2.01s/it]
Processing batches:  29%|██▉       | 109/376 [04:33<09:12,  2.07s/it]
Processing batches:  29%|██▉       | 110/376 [04:35<09:04,  2.05s/it]
Processing batches:  30%|██▉       | 111/376 [04:37<08:48,  2.00s/it]
Processing batches:  30%|██▉       | 112/376 [04:39<09:21,  2.13s/it]
Processing batches:  30%|███       | 113/376 [04:41<09:16,  2.12s/it]
Processing batches:  30%|███       | 114/376 [04:44<09:32,  2.19s/it]
Processing batches:  31%|███       | 115/376 [04:45<08:35,  1.98s/it]
Processing batches:  31%|███       | 116/376 [04:47<08:09,  1.88s/it]
Processing batches:  31%|███       | 117/376 [04:48<07:53,  1.83s/it]
Processing batches:  31%|███▏      | 118/376 [04:50<07:46,  1.81s/it]
Processing batches:  32%|███▏      | 119/376 [04:53<09:28,  2.21s/it]
Processing batches:  32%|███▏      | 120/376 [04:56<09:36,  2.25s/it]
Processing batches:  32%|███▏      | 121/376 [04:58<09:13,  2.17s/it]
Processing batches:  32%|███▏      | 122/376 [04:59<08:48,  2.08s/it]
Processing batches:  33%|███▎      | 123/376 [05:02<08:57,  2.13s/it]
Processing batches:  33%|███▎      | 124/376 [05:05<10:36,  2.52s/it]
Processing batches:  33%|███▎      | 125/376 [05:07<09:32,  2.28s/it]
Processing batches:  34%|███▎      | 126/376 [05:09<09:09,  2.20s/it]
Processing batches:  34%|███▍      | 127/376 [05:11<08:30,  2.05s/it]
Processing batches:  34%|███▍      | 128/376 [05:12<07:56,  1.92s/it]
Processing batches:  34%|███▍      | 129/376 [05:15<08:46,  2.13s/it]
Processing batches:  35%|███▍      | 130/376 [05:18<09:28,  2.31s/it]
Processing batches:  35%|███▍      | 131/376 [05:20<09:58,  2.44s/it]
Processing batches:  35%|███▌      | 132/376 [05:23<10:21,  2.55s/it]
Processing batches:  35%|███▌      | 133/376 [05:26<10:32,  2.60s/it]
Processing batches:  36%|███▌      | 134/376 [05:28<09:44,  2.41s/it]
Processing batches:  36%|███▌      | 135/376 [05:30<09:28,  2.36s/it]
Processing batches:  36%|███▌      | 136/376 [05:33<09:59,  2.50s/it]
Processing batches:  36%|███▋      | 137/376 [05:35<09:55,  2.49s/it]
Processing batches:  37%|███▋      | 138/376 [05:38<09:49,  2.48s/it]
Processing batches:  37%|███▋      | 139/376 [05:40<08:56,  2.26s/it]
Processing batches:  37%|███▋      | 140/376 [05:42<08:50,  2.25s/it]
Processing batches:  38%|███▊      | 141/376 [05:44<08:28,  2.16s/it]
Processing batches:  38%|███▊      | 142/376 [05:46<08:18,  2.13s/it]
Processing batches:  38%|███▊      | 143/376 [05:48<08:15,  2.13s/it]
Processing batches:  38%|███▊      | 144/376 [05:50<08:08,  2.11s/it]
Processing batches:  39%|███▊      | 145/376 [05:53<08:42,  2.26s/it]
Processing batches:  39%|███▉      | 146/376 [05:56<09:33,  2.49s/it]
Processing batches:  39%|███▉      | 147/376 [05:58<09:03,  2.38s/it]
Processing batches:  39%|███▉      | 148/376 [06:00<09:16,  2.44s/it]
Processing batches:  40%|███▉      | 149/376 [06:02<08:27,  2.24s/it]
Processing batches:  40%|███▉      | 150/376 [06:04<08:32,  2.27s/it]
Processing batches:  40%|████      | 151/376 [06:06<07:33,  2.02s/it]
Processing batches:  40%|████      | 152/376 [06:08<07:57,  2.13s/it]
Processing batches:  41%|████      | 153/376 [06:10<07:08,  1.92s/it]
Processing batches:  41%|████      | 154/376 [06:11<06:51,  1.85s/it]
Processing batches:  41%|████      | 155/376 [06:13<06:14,  1.70s/it]
Processing batches:  41%|████▏     | 156/376 [06:15<06:38,  1.81s/it]
Processing batches:  42%|████▏     | 157/376 [06:17<07:28,  2.05s/it]
Processing batches:  42%|████▏     | 158/376 [06:19<07:07,  1.96s/it]
Processing batches:  42%|████▏     | 159/376 [06:21<06:47,  1.88s/it]
Processing batches:  43%|████▎     | 160/376 [06:22<06:30,  1.81s/it]
Processing batches:  43%|████▎     | 161/376 [06:24<06:13,  1.74s/it]
Processing batches:  43%|████▎     | 162/376 [06:26<06:15,  1.76s/it]
Processing batches:  43%|████▎     | 163/376 [06:27<05:56,  1.67s/it]
Processing batches:  44%|████▎     | 164/376 [06:29<06:15,  1.77s/it]
Processing batches:  44%|████▍     | 165/376 [06:31<06:26,  1.83s/it]
Processing batches:  44%|████▍     | 166/376 [06:34<07:15,  2.07s/it]
Processing batches:  44%|████▍     | 167/376 [06:36<07:13,  2.07s/it]
Processing batches:  45%|████▍     | 168/376 [06:38<07:32,  2.18s/it]
Processing batches:  45%|████▍     | 169/376 [06:40<07:03,  2.05s/it]
Processing batches:  45%|████▌     | 170/376 [06:42<07:11,  2.10s/it]
Processing batches:  45%|████▌     | 171/376 [06:44<07:04,  2.07s/it]
Processing batches:  46%|████▌     | 172/376 [06:46<06:43,  1.98s/it]
Processing batches:  46%|████▌     | 173/376 [06:47<06:02,  1.78s/it]
Processing batches:  46%|████▋     | 174/376 [06:50<07:14,  2.15s/it]
Processing batches:  47%|████▋     | 175/376 [06:52<06:53,  2.06s/it]
Processing batches:  47%|████▋     | 176/376 [06:53<05:38,  1.69s/it]
Processing batches:  47%|████▋     | 177/376 [06:57<07:32,  2.28s/it]
Processing batches:  47%|████▋     | 178/376 [06:59<07:48,  2.37s/it]
Processing batches:  48%|████▊     | 179/376 [07:01<07:13,  2.20s/it]
Processing batches:  48%|████▊     | 180/376 [07:03<06:42,  2.05s/it]
Processing batches:  48%|████▊     | 181/376 [07:05<06:54,  2.13s/it]
Processing batches:  48%|████▊     | 182/376 [07:07<06:38,  2.05s/it]
Processing batches:  49%|████▊     | 183/376 [07:09<06:17,  1.96s/it]
Processing batches:  49%|████▉     | 184/376 [07:12<07:32,  2.36s/it]
Processing batches:  49%|████▉     | 185/376 [07:14<07:18,  2.30s/it]
Processing batches:  49%|████▉     | 186/376 [07:16<06:39,  2.10s/it]
Processing batches:  50%|████▉     | 187/376 [07:17<06:00,  1.91s/it]
Processing batches:  50%|█████     | 188/376 [07:19<05:36,  1.79s/it]
Processing batches:  50%|█████     | 189/376 [07:20<05:10,  1.66s/it]
Processing batches:  51%|█████     | 190/376 [07:22<05:12,  1.68s/it]
Processing batches:  51%|█████     | 191/376 [07:24<05:48,  1.88s/it]
Processing batches:  51%|█████     | 192/376 [07:27<06:54,  2.25s/it]
Processing batches:  51%|█████▏    | 193/376 [07:29<06:34,  2.15s/it]
Processing batches:  52%|█████▏    | 194/376 [07:32<06:35,  2.17s/it]
Processing batches:  52%|█████▏    | 195/376 [07:34<07:08,  2.37s/it]
Processing batches:  52%|█████▏    | 196/376 [07:38<08:40,  2.89s/it]
Processing batches:  52%|█████▏    | 197/376 [07:41<08:08,  2.73s/it]
Processing batches:  53%|█████▎    | 198/376 [07:43<07:08,  2.41s/it]
Processing batches:  53%|█████▎    | 199/376 [07:45<06:54,  2.34s/it]
Processing batches:  53%|█████▎    | 200/376 [07:46<06:03,  2.06s/it]
Processing batches:  53%|█████▎    | 201/376 [07:47<05:21,  1.84s/it]
Processing batches:  54%|█████▎    | 202/376 [07:49<04:46,  1.65s/it]
Processing batches:  54%|█████▍    | 203/376 [07:51<05:48,  2.01s/it]
Processing batches:  54%|█████▍    | 204/376 [07:54<06:26,  2.25s/it]
Processing batches:  55%|█████▍    | 205/376 [07:56<06:03,  2.12s/it]
Processing batches:  55%|█████▍    | 206/376 [07:58<06:08,  2.17s/it]
Processing batches:  55%|█████▌    | 207/376 [08:00<06:02,  2.15s/it]
Processing batches:  55%|█████▌    | 208/376 [08:04<06:57,  2.49s/it]
Processing batches:  56%|█████▌    | 209/376 [08:06<07:02,  2.53s/it]
Processing batches:  56%|█████▌    | 210/376 [08:10<08:01,  2.90s/it]
Processing batches:  56%|█████▌    | 211/376 [08:13<07:32,  2.74s/it]
Processing batches:  56%|█████▋    | 212/376 [08:14<06:47,  2.49s/it]
Processing batches:  57%|█████▋    | 213/376 [08:17<06:45,  2.49s/it]
Processing batches:  57%|█████▋    | 214/376 [08:19<06:26,  2.38s/it]
Processing batches:  57%|█████▋    | 215/376 [08:20<05:23,  2.01s/it]
Processing batches:  57%|█████▋    | 216/376 [08:21<04:41,  1.76s/it]
Processing batches:  58%|█████▊    | 217/376 [08:23<04:45,  1.80s/it]
Processing batches:  58%|█████▊    | 218/376 [08:25<04:55,  1.87s/it]
Processing batches:  58%|█████▊    | 219/376 [08:28<05:13,  2.00s/it]
Processing batches:  59%|█████▊    | 220/376 [08:30<05:17,  2.04s/it]
Processing batches:  59%|█████▉    | 221/376 [08:31<05:04,  1.96s/it]
Processing batches:  59%|█████▉    | 222/376 [08:34<05:21,  2.09s/it]
Processing batches:  59%|█████▉    | 223/376 [08:36<05:31,  2.17s/it]
Processing batches:  60%|█████▉    | 224/376 [08:38<05:05,  2.01s/it]
Processing batches:  60%|█████▉    | 225/376 [08:40<04:47,  1.90s/it]
Processing batches:  60%|██████    | 226/376 [08:41<04:42,  1.88s/it]
Processing batches:  60%|██████    | 227/376 [08:43<04:44,  1.91s/it]
Processing batches:  61%|██████    | 228/376 [08:45<04:34,  1.86s/it]
Processing batches:  61%|██████    | 229/376 [08:47<04:21,  1.78s/it]
Processing batches:  61%|██████    | 230/376 [08:49<04:36,  1.90s/it]
Processing batches:  61%|██████▏   | 231/376 [08:50<04:15,  1.76s/it]
Processing batches:  62%|██████▏   | 232/376 [08:52<04:22,  1.82s/it]
Processing batches:  62%|██████▏   | 233/376 [08:54<04:14,  1.78s/it]
Processing batches:  62%|██████▏   | 234/376 [08:55<03:45,  1.59s/it]
Processing batches:  62%|██████▎   | 235/376 [08:57<03:49,  1.63s/it]
Processing batches:  63%|██████▎   | 236/376 [08:58<03:33,  1.53s/it]
Processing batches:  63%|██████▎   | 237/376 [09:00<03:48,  1.65s/it]
Processing batches:  63%|██████▎   | 238/376 [09:02<04:09,  1.81s/it]
Processing batches:  64%|██████▎   | 239/376 [09:04<04:14,  1.86s/it]
Processing batches:  64%|██████▍   | 240/376 [09:06<04:13,  1.87s/it]
Processing batches:  64%|██████▍   | 241/376 [09:08<04:06,  1.83s/it]
Processing batches:  64%|██████▍   | 242/376 [09:11<04:51,  2.18s/it]
Processing batches:  65%|██████▍   | 243/376 [09:13<04:44,  2.14s/it]
Processing batches:  65%|██████▍   | 244/376 [09:15<04:25,  2.01s/it]
Processing batches:  65%|██████▌   | 245/376 [09:18<05:32,  2.54s/it]
Processing batches:  65%|██████▌   | 246/376 [09:20<04:50,  2.23s/it]
Processing batches:  66%|██████▌   | 247/376 [09:22<04:28,  2.08s/it]
Processing batches:  66%|██████▌   | 248/376 [09:23<03:59,  1.87s/it]
Processing batches:  66%|██████▌   | 249/376 [09:24<03:18,  1.56s/it]
Processing batches:  66%|██████▋   | 250/376 [09:25<02:58,  1.42s/it]
Processing batches:  67%|██████▋   | 251/376 [09:26<02:41,  1.29s/it]
Processing batches:  67%|██████▋   | 252/376 [09:27<02:45,  1.34s/it]
Processing batches:  67%|██████▋   | 253/376 [09:29<02:51,  1.40s/it]
Processing batches:  68%|██████▊   | 254/376 [09:31<03:00,  1.48s/it]
Processing batches:  68%|██████▊   | 255/376 [09:32<03:14,  1.61s/it]
Processing batches:  68%|██████▊   | 256/376 [09:34<03:20,  1.67s/it]
Processing batches:  68%|██████▊   | 257/376 [09:36<03:09,  1.59s/it]
Processing batches:  69%|██████▊   | 258/376 [09:38<03:22,  1.71s/it]
Processing batches:  69%|██████▉   | 259/376 [09:40<03:36,  1.85s/it]
Processing batches:  69%|██████▉   | 260/376 [09:43<04:05,  2.11s/it]
Processing batches:  69%|██████▉   | 261/376 [09:44<03:38,  1.90s/it]
Processing batches:  70%|██████▉   | 262/376 [09:46<03:32,  1.86s/it]
Processing batches:  70%|██████▉   | 263/376 [09:48<03:28,  1.85s/it]
Processing batches:  70%|███████   | 264/376 [09:50<03:43,  1.99s/it]
Processing batches:  70%|███████   | 265/376 [09:52<03:54,  2.11s/it]
Processing batches:  71%|███████   | 266/376 [09:54<03:50,  2.10s/it]
Processing batches:  71%|███████   | 267/376 [09:56<03:35,  1.97s/it]
Processing batches:  71%|███████▏  | 268/376 [09:57<03:13,  1.80s/it]
Processing batches:  72%|███████▏  | 269/376 [09:59<03:08,  1.76s/it]
Processing batches:  72%|███████▏  | 270/376 [10:01<02:58,  1.68s/it]
Processing batches:  72%|███████▏  | 271/376 [10:02<02:49,  1.61s/it]
Processing batches:  72%|███████▏  | 272/376 [10:05<03:21,  1.94s/it]
Processing batches:  73%|███████▎  | 273/376 [10:07<03:19,  1.93s/it]
Processing batches:  73%|███████▎  | 274/376 [10:08<03:12,  1.89s/it]
Processing batches:  73%|███████▎  | 275/376 [10:09<02:40,  1.59s/it]
Processing batches:  73%|███████▎  | 276/376 [10:12<03:15,  1.95s/it]
Processing batches:  74%|███████▎  | 277/376 [10:13<02:52,  1.75s/it]
Processing batches:  74%|███████▍  | 278/376 [10:16<03:30,  2.15s/it]
Processing batches:  74%|███████▍  | 279/376 [10:17<02:50,  1.76s/it]
Processing batches:  74%|███████▍  | 280/376 [10:19<02:41,  1.68s/it]
Processing batches:  75%|███████▍  | 281/376 [10:20<02:35,  1.64s/it]
Processing batches:  75%|███████▌  | 282/376 [10:22<02:41,  1.72s/it]
Processing batches:  75%|███████▌  | 283/376 [10:24<02:27,  1.59s/it]
Processing batches:  76%|███████▌  | 284/376 [10:25<02:13,  1.46s/it]
Processing batches:  76%|███████▌  | 285/376 [10:25<01:54,  1.26s/it]
Processing batches:  76%|███████▌  | 286/376 [10:26<01:22,  1.09it/s]
Processing batches:  76%|███████▋  | 287/376 [10:28<01:47,  1.21s/it]
Processing batches:  77%|███████▋  | 288/376 [10:29<01:55,  1.31s/it]
Processing batches:  77%|███████▋  | 289/376 [10:31<02:03,  1.42s/it]
Processing batches:  77%|███████▋  | 290/376 [10:32<02:07,  1.48s/it]
Processing batches:  77%|███████▋  | 291/376 [10:34<02:03,  1.45s/it]
Processing batches:  78%|███████▊  | 292/376 [10:35<02:08,  1.54s/it]
Processing batches:  78%|███████▊  | 293/376 [10:37<02:11,  1.58s/it]
Processing batches:  78%|███████▊  | 294/376 [10:39<02:11,  1.60s/it]
Processing batches:  78%|███████▊  | 295/376 [10:40<01:54,  1.41s/it]
Processing batches:  79%|███████▊  | 296/376 [10:41<01:52,  1.40s/it]
Processing batches:  79%|███████▉  | 297/376 [10:43<01:50,  1.40s/it]
Processing batches:  79%|███████▉  | 298/376 [10:44<01:58,  1.51s/it]
Processing batches:  80%|███████▉  | 299/376 [10:45<01:45,  1.37s/it]
Processing batches:  80%|███████▉  | 300/376 [10:46<01:37,  1.29s/it]
Processing batches:  80%|████████  | 301/376 [10:48<01:47,  1.43s/it]
Processing batches:  80%|████████  | 302/376 [10:50<01:53,  1.54s/it]
Processing batches:  81%|████████  | 303/376 [10:51<01:51,  1.52s/it]
Processing batches:  81%|████████  | 304/376 [10:53<01:45,  1.47s/it]
Processing batches:  81%|████████  | 305/376 [10:54<01:41,  1.43s/it]
Processing batches:  81%|████████▏ | 306/376 [10:55<01:35,  1.36s/it]
Processing batches:  82%|████████▏ | 307/376 [10:57<01:42,  1.49s/it]
Processing batches:  82%|████████▏ | 308/376 [10:59<01:43,  1.52s/it]
Processing batches:  82%|████████▏ | 309/376 [11:00<01:33,  1.39s/it]
Processing batches:  82%|████████▏ | 310/376 [11:01<01:31,  1.39s/it]
Processing batches:  83%|████████▎ | 311/376 [11:04<01:55,  1.77s/it]
Processing batches:  83%|████████▎ | 312/376 [11:05<01:36,  1.50s/it]
Processing batches:  83%|████████▎ | 313/376 [11:07<01:41,  1.61s/it]
Processing batches:  84%|████████▎ | 314/376 [11:08<01:36,  1.55s/it]
Processing batches:  84%|████████▍ | 315/376 [11:10<01:47,  1.76s/it]
Processing batches:  84%|████████▍ | 316/376 [11:12<01:38,  1.64s/it]
Processing batches:  84%|████████▍ | 317/376 [11:14<01:44,  1.77s/it]
Processing batches:  85%|████████▍ | 318/376 [11:15<01:42,  1.76s/it]
Processing batches:  85%|████████▍ | 319/376 [11:18<01:49,  1.93s/it]
Processing batches:  85%|████████▌ | 320/376 [11:19<01:36,  1.72s/it]
Processing batches:  85%|████████▌ | 321/376 [11:20<01:29,  1.62s/it]
Processing batches:  86%|████████▌ | 322/376 [11:22<01:21,  1.51s/it]
Processing batches:  86%|████████▌ | 323/376 [11:23<01:14,  1.40s/it]
Processing batches:  86%|████████▌ | 324/376 [11:25<01:20,  1.55s/it]
Processing batches:  86%|████████▋ | 325/376 [11:26<01:16,  1.51s/it]
Processing batches:  87%|████████▋ | 326/376 [11:27<01:06,  1.33s/it]
Processing batches:  87%|████████▋ | 327/376 [11:29<01:09,  1.42s/it]
Processing batches:  87%|████████▋ | 328/376 [11:31<01:16,  1.59s/it]
Processing batches:  88%|████████▊ | 329/376 [11:32<01:12,  1.54s/it]
Processing batches:  88%|████████▊ | 330/376 [11:34<01:19,  1.73s/it]
Processing batches:  88%|████████▊ | 331/376 [11:35<01:07,  1.49s/it]
Processing batches:  88%|████████▊ | 332/376 [11:37<01:08,  1.56s/it]
Processing batches:  89%|████████▊ | 333/376 [11:38<00:57,  1.33s/it]
Processing batches:  89%|████████▉ | 334/376 [11:39<01:00,  1.45s/it]
Processing batches:  89%|████████▉ | 335/376 [11:41<01:05,  1.60s/it]
Processing batches:  89%|████████▉ | 336/376 [11:43<01:06,  1.67s/it]
Processing batches:  90%|████████▉ | 337/376 [11:45<01:03,  1.63s/it]
Processing batches:  90%|████████▉ | 338/376 [11:46<00:56,  1.49s/it]
Processing batches:  90%|█████████ | 339/376 [11:47<00:48,  1.32s/it]
Processing batches:  90%|█████████ | 340/376 [11:48<00:45,  1.25s/it]
Processing batches:  91%|█████████ | 341/376 [11:49<00:44,  1.26s/it]
Processing batches:  91%|█████████ | 342/376 [11:51<00:51,  1.51s/it]
Processing batches:  91%|█████████ | 343/376 [11:53<00:50,  1.53s/it]
Processing batches:  91%|█████████▏| 344/376 [11:54<00:42,  1.34s/it]
Processing batches:  92%|█████████▏| 345/376 [11:55<00:39,  1.26s/it]
Processing batches:  92%|█████████▏| 346/376 [11:56<00:38,  1.28s/it]
Processing batches:  92%|█████████▏| 347/376 [11:57<00:34,  1.17s/it]
Processing batches:  93%|█████████▎| 348/376 [11:58<00:32,  1.17s/it]
Processing batches:  93%|█████████▎| 349/376 [12:00<00:33,  1.25s/it]
Processing batches:  93%|█████████▎| 350/376 [12:01<00:33,  1.28s/it]
Processing batches:  93%|█████████▎| 351/376 [12:02<00:32,  1.31s/it]
Processing batches:  94%|█████████▎| 352/376 [12:04<00:33,  1.41s/it]
Processing batches:  94%|█████████▍| 353/376 [12:05<00:28,  1.24s/it]
Processing batches:  94%|█████████▍| 354/376 [12:06<00:27,  1.23s/it]
Processing batches:  94%|█████████▍| 355/376 [12:07<00:23,  1.13s/it]
Processing batches:  95%|█████████▍| 356/376 [12:08<00:22,  1.14s/it]
Processing batches:  95%|█████████▍| 357/376 [12:09<00:20,  1.09s/it]
Processing batches:  95%|█████████▌| 358/376 [12:12<00:31,  1.75s/it]
Processing batches:  95%|█████████▌| 359/376 [12:13<00:24,  1.46s/it]
Processing batches:  96%|█████████▌| 360/376 [12:14<00:19,  1.24s/it]
Processing batches:  96%|█████████▌| 361/376 [12:15<00:17,  1.16s/it]
Processing batches:  96%|█████████▋| 362/376 [12:16<00:15,  1.14s/it]
Processing batches:  97%|█████████▋| 363/376 [12:17<00:12,  1.02it/s]
Processing batches:  97%|█████████▋| 364/376 [12:18<00:12,  1.07s/it]
Processing batches:  97%|█████████▋| 365/376 [12:19<00:12,  1.09s/it]
Processing batches:  97%|█████████▋| 366/376 [12:20<00:10,  1.04s/it]
Processing batches:  98%|█████████▊| 367/376 [12:21<00:08,  1.04it/s]
Processing batches:  98%|█████████▊| 368/376 [12:22<00:07,  1.10it/s]
Processing batches:  98%|█████████▊| 369/376 [12:22<00:06,  1.15it/s]
Processing batches:  98%|█████████▊| 370/376 [12:23<00:05,  1.18it/s]
Processing batches:  99%|█████████▊| 371/376 [12:24<00:04,  1.13it/s]
Processing batches:  99%|█████████▉| 372/376 [12:25<00:03,  1.07it/s]
Processing batches:  99%|█████████▉| 373/376 [12:26<00:02,  1.11it/s]
Processing batches:  99%|█████████▉| 374/376 [12:27<00:01,  1.10it/s]
Processing batches: 100%|█████████▉| 375/376 [12:28<00:00,  1.14it/s]
Processing batches: 100%|██████████| 376/376 [12:28<00:00,  1.42it/s]
Processing batches: 100%|██████████| 376/376 [12:28<00:00,  1.99s/it]
articles_mat: (384, 3753)
authors_mat: (384, 720)
words_mat: (384, 1981)

minilm_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=minilm_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
minilm_score.print()

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1653

--------- Desc Scores ---------
neighbors_articles_authors_det

0      Natalia V. Andrienko : 0.45
1      Gennady L. Andrienko : 0.43
2             Mario Jelovic : 0.42
3            Aditi Majumder : 0.41
4            Bernhard Preim : 0.41
                  ...
715           Chris Muelder : 0.10
716        Harald Obermaier : 0.10
717            Teng-Yok Lee : 0.10
718       Yixuan Zhang 0001 : 0.10
719         Jörn Kohlhammer : 0.10
Length: 720, dtype: object

VALUE : 0.1653

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.139130
Dieter Schmalstieg     0.144444
Matthew Kay 0001       0.290000
Xingbo Wang 0001       0.133333
Minfeng Zhu            0.100000
                         ...
Lisa M. Sobierajski    0.175000
Nahum D. Gershon       0.140000
Sidney W. Wang         0.100000
David A. Lane          0.216667
T. Todd Elvins         0.150000
Length: 720, dtype: float64

“all-mpnet-base-v2” projection

mpnet_matrices = bert_projection([articles_mat, authors_mat, words_mat], df['text'], family="all-mpnet-base-v2")

""
mpnet_matrices = list(map(normalize_l2, mpnet_matrices))

""
print(f"articles_mat: {mpnet_matrices[0].shape}")
print(f"authors_mat: {mpnet_matrices[1].shape}")
print(f"words_mat: {mpnet_matrices[2].shape}")

""
mpnet_score = Neighbors.evaluate(
    NATURE, SOURCE, authors_mat, authors_scores, dir_xD=".",
    scores_nature=articles_scores, matrix_nature_xD=mpnet_matrices[0],
    min_score=filt_min_score, n_neighbors=n_neighbors, recompute=True
)
mpnet_score.print()

Processing batches:   0%|          | 0/376 [00:00<?, ?it/s]
Processing batches:   0%|          | 1/376 [00:13<1:22:56, 13.27s/it]
Processing batches:   1%|          | 2/376 [00:24<1:16:21, 12.25s/it]
Processing batches:   1%|          | 3/376 [00:35<1:11:44, 11.54s/it]
Processing batches:   1%|          | 4/376 [00:50<1:19:14, 12.78s/it]
Processing batches:   1%|▏         | 5/376 [00:59<1:10:41, 11.43s/it]
Processing batches:   2%|▏         | 6/376 [01:10<1:09:45, 11.31s/it]
Processing batches:   2%|▏         | 7/376 [01:18<1:04:03, 10.42s/it]
Processing batches:   2%|▏         | 8/376 [01:30<1:05:49, 10.73s/it]
Processing batches:   2%|▏         | 9/376 [01:41<1:05:40, 10.74s/it]
Processing batches:   3%|▎         | 10/376 [01:53<1:09:09, 11.34s/it]
Processing batches:   3%|▎         | 11/376 [02:05<1:10:27, 11.58s/it]
Processing batches:   3%|▎         | 12/376 [02:13<1:03:28, 10.46s/it]
Processing batches:   3%|▎         | 13/376 [02:25<1:06:33, 11.00s/it]
Processing batches:   4%|▎         | 14/376 [02:37<1:08:10, 11.30s/it]
Processing batches:   4%|▍         | 15/376 [02:49<1:08:29, 11.38s/it]
Processing batches:   4%|▍         | 16/376 [03:03<1:12:13, 12.04s/it]
Processing batches:   5%|▍         | 17/376 [03:12<1:06:35, 11.13s/it]
Processing batches:   5%|▍         | 18/376 [03:23<1:07:09, 11.25s/it]
Processing batches:   5%|▌         | 19/376 [03:34<1:06:59, 11.26s/it]
Processing batches:   5%|▌         | 20/376 [03:44<1:03:54, 10.77s/it]
Processing batches:   6%|▌         | 21/376 [03:53<1:00:00, 10.14s/it]
Processing batches:   6%|▌         | 22/376 [04:04<1:01:00, 10.34s/it]
Processing batches:   6%|▌         | 23/376 [04:15<1:02:48, 10.68s/it]
Processing batches:   6%|▋         | 24/376 [04:27<1:05:33, 11.17s/it]
Processing batches:   7%|▋         | 25/376 [04:41<1:09:28, 11.87s/it]
Processing batches:   7%|▋         | 26/376 [04:49<1:02:28, 10.71s/it]
Processing batches:   7%|▋         | 27/376 [05:01<1:04:49, 11.14s/it]
Processing batches:   7%|▋         | 28/376 [05:10<1:01:08, 10.54s/it]
Processing batches:   8%|▊         | 29/376 [05:22<1:03:55, 11.05s/it]
Processing batches:   8%|▊         | 30/376 [05:36<1:07:47, 11.76s/it]
Processing batches:   8%|▊         | 31/376 [05:43<1:00:16, 10.48s/it]
Processing batches:   9%|▊         | 32/376 [05:57<1:05:41, 11.46s/it]
Processing batches:   9%|▉         | 33/376 [06:13<1:12:43, 12.72s/it]
Processing batches:   9%|▉         | 34/376 [06:23<1:07:37, 11.86s/it]
Processing batches:   9%|▉         | 35/376 [06:34<1:06:41, 11.73s/it]
Processing batches:  10%|▉         | 36/376 [06:45<1:05:34, 11.57s/it]
Processing batches:  10%|▉         | 37/376 [06:55<1:01:54, 10.96s/it]
Processing batches:  10%|█         | 38/376 [07:05<1:00:08, 10.68s/it]
Processing batches:  10%|█         | 39/376 [07:13<56:27, 10.05s/it]
Processing batches:  11%|█         | 40/376 [07:23<55:06,  9.84s/it]
Processing batches:  11%|█         | 41/376 [07:32<53:36,  9.60s/it]
Processing batches:  11%|█         | 42/376 [07:42<55:20,  9.94s/it]
Processing batches:  11%|█▏        | 43/376 [07:58<1:04:38, 11.65s/it]
Processing batches:  12%|█▏        | 44/376 [08:05<57:16, 10.35s/it]
Processing batches:  12%|█▏        | 45/376 [08:14<53:29,  9.70s/it]
Processing batches:  12%|█▏        | 46/376 [08:22<51:49,  9.42s/it]
Processing batches:  12%|█▎        | 47/376 [08:32<52:35,  9.59s/it]
Processing batches:  13%|█▎        | 48/376 [08:41<50:06,  9.17s/it]
Processing batches:  13%|█▎        | 49/376 [08:50<49:48,  9.14s/it]
Processing batches:  13%|█▎        | 50/376 [09:00<50:53,  9.37s/it]
Processing batches:  14%|█▎        | 51/376 [09:12<55:45, 10.29s/it]
Processing batches:  14%|█▍        | 52/376 [09:23<56:04, 10.39s/it]
Processing batches:  14%|█▍        | 53/376 [09:31<53:17,  9.90s/it]
Processing batches:  14%|█▍        | 54/376 [09:42<54:14, 10.11s/it]
Processing batches:  15%|█▍        | 55/376 [09:56<1:01:05, 11.42s/it]
Processing batches:  15%|█▍        | 56/376 [10:07<59:17, 11.12s/it]
Processing batches:  15%|█▌        | 57/376 [10:15<54:47, 10.31s/it]
Processing batches:  15%|█▌        | 58/376 [10:30<1:01:26, 11.59s/it]
Processing batches:  16%|█▌        | 59/376 [10:41<1:00:11, 11.39s/it]
Processing batches:  16%|█▌        | 60/376 [10:48<52:55, 10.05s/it]
Processing batches:  16%|█▌        | 61/376 [10:58<52:50, 10.07s/it]
Processing batches:  16%|█▋        | 62/376 [11:06<49:57,  9.55s/it]
Processing batches:  17%|█▋        | 63/376 [11:15<49:33,  9.50s/it]
Processing batches:  17%|█▋        | 64/376 [11:26<50:40,  9.74s/it]
Processing batches:  17%|█▋        | 65/376 [11:38<53:45, 10.37s/it]
Processing batches:  18%|█▊        | 66/376 [11:49<54:35, 10.57s/it]
Processing batches:  18%|█▊        | 67/376 [12:00<55:54, 10.86s/it]
Processing batches:  18%|█▊        | 68/376 [12:08<51:12,  9.98s/it]
Processing batches:  18%|█▊        | 69/376 [12:16<48:26,  9.47s/it]
Processing batches:  19%|█▊        | 70/376 [12:26<48:47,  9.57s/it]
Processing batches:  19%|█▉        | 71/376 [12:37<49:59,  9.83s/it]
Processing batches:  19%|█▉        | 72/376 [12:51<56:42, 11.19s/it]
Processing batches:  19%|█▉        | 73/376 [13:00<53:22, 10.57s/it]
Processing batches:  20%|█▉        | 74/376 [13:11<53:42, 10.67s/it]
Processing batches:  20%|█▉        | 75/376 [13:21<53:01, 10.57s/it]
Processing batches:  20%|██        | 76/376 [13:30<49:12,  9.84s/it]
Processing batches:  20%|██        | 77/376 [13:40<49:39,  9.96s/it]
Processing batches:  21%|██        | 78/376 [13:52<52:20, 10.54s/it]
Processing batches:  21%|██        | 79/376 [14:01<50:36, 10.22s/it]
Processing batches:  21%|██▏       | 80/376 [14:09<47:26,  9.62s/it]
Processing batches:  22%|██▏       | 81/376 [14:18<45:43,  9.30s/it]
Processing batches:  22%|██▏       | 82/376 [14:31<51:41, 10.55s/it]
Processing batches:  22%|██▏       | 83/376 [14:41<50:18, 10.30s/it]
Processing batches:  22%|██▏       | 84/376 [14:57<57:40, 11.85s/it]
Processing batches:  23%|██▎       | 85/376 [15:03<50:00, 10.31s/it]
Processing batches:  23%|██▎       | 86/376 [15:19<57:58, 11.99s/it]
Processing batches:  23%|██▎       | 87/376 [15:27<51:34, 10.71s/it]
Processing batches:  23%|██▎       | 88/376 [15:36<49:48, 10.38s/it]
Processing batches:  24%|██▎       | 89/376 [15:48<50:55, 10.65s/it]
Processing batches:  24%|██▍       | 90/376 [16:00<53:21, 11.20s/it]
Processing batches:  24%|██▍       | 91/376 [16:10<51:21, 10.81s/it]
Processing batches:  24%|██▍       | 92/376 [16:25<56:47, 12.00s/it]
Processing batches:  25%|██▍       | 93/376 [16:32<49:30, 10.49s/it]
Processing batches:  25%|██▌       | 94/376 [16:42<48:44, 10.37s/it]
Processing batches:  25%|██▌       | 95/376 [16:53<48:47, 10.42s/it]
Processing batches:  26%|██▌       | 96/376 [17:05<50:54, 10.91s/it]
Processing batches:  26%|██▌       | 97/376 [17:16<51:37, 11.10s/it]
Processing batches:  26%|██▌       | 98/376 [17:25<48:53, 10.55s/it]
Processing batches:  26%|██▋       | 99/376 [17:34<45:31,  9.86s/it]
Processing batches:  27%|██▋       | 100/376 [17:43<44:31,  9.68s/it]
Processing batches:  27%|██▋       | 101/376 [17:49<39:26,  8.61s/it]
Processing batches:  27%|██▋       | 102/376 [17:57<38:00,  8.32s/it]
Processing batches:  27%|██▋       | 103/376 [18:05<38:31,  8.47s/it]
Processing batches:  28%|██▊       | 104/376 [18:15<39:32,  8.72s/it]
Processing batches:  28%|██▊       | 105/376 [18:23<39:10,  8.67s/it]
Processing batches:  28%|██▊       | 106/376 [18:33<40:02,  8.90s/it]
Processing batches:  28%|██▊       | 107/376 [18:43<41:16,  9.21s/it]
Processing batches:  29%|██▊       | 108/376 [18:50<38:32,  8.63s/it]
Processing batches:  29%|██▉       | 109/376 [18:59<39:07,  8.79s/it]
Processing batches:  29%|██▉       | 110/376 [19:08<38:29,  8.68s/it]
Processing batches:  30%|██▉       | 111/376 [19:16<37:35,  8.51s/it]
Processing batches:  30%|██▉       | 112/376 [19:26<39:52,  9.06s/it]
Processing batches:  30%|███       | 113/376 [19:35<39:59,  9.12s/it]
Processing batches:  30%|███       | 114/376 [19:45<40:48,  9.35s/it]
Processing batches:  31%|███       | 115/376 [19:52<37:12,  8.55s/it]
Processing batches:  31%|███       | 116/376 [19:59<35:21,  8.16s/it]
Processing batches:  31%|███       | 117/376 [20:07<34:18,  7.95s/it]
Processing batches:  31%|███▏      | 118/376 [20:14<34:07,  7.94s/it]
Processing batches:  32%|███▏      | 119/376 [20:27<40:21,  9.42s/it]
Processing batches:  32%|███▏      | 120/376 [20:37<41:00,  9.61s/it]
Processing batches:  32%|███▏      | 121/376 [20:46<39:26,  9.28s/it]
Processing batches:  32%|███▏      | 122/376 [20:54<38:00,  8.98s/it]
Processing batches:  33%|███▎      | 123/376 [21:04<38:29,  9.13s/it]
Processing batches:  33%|███▎      | 124/376 [21:17<43:48, 10.43s/it]
Processing batches:  33%|███▎      | 125/376 [21:25<40:01,  9.57s/it]
Processing batches:  34%|███▎      | 126/376 [21:34<39:39,  9.52s/it]
Processing batches:  34%|███▍      | 127/376 [21:42<37:26,  9.02s/it]
Processing batches:  34%|███▍      | 128/376 [21:50<35:29,  8.59s/it]
Processing batches:  34%|███▍      | 129/376 [22:01<39:19,  9.55s/it]
Processing batches:  35%|███▍      | 130/376 [22:13<42:05, 10.26s/it]
Processing batches:  35%|███▍      | 131/376 [22:26<44:20, 10.86s/it]
Processing batches:  35%|███▌      | 132/376 [22:36<44:01, 10.83s/it]
Processing batches:  35%|███▌      | 133/376 [22:48<44:26, 10.97s/it]
Processing batches:  36%|███▌      | 134/376 [22:56<41:07, 10.19s/it]
Processing batches:  36%|███▌      | 135/376 [23:06<40:48, 10.16s/it]
Processing batches:  36%|███▌      | 136/376 [23:18<42:41, 10.67s/it]
Processing batches:  36%|███▋      | 137/376 [23:29<42:39, 10.71s/it]
Processing batches:  37%|███▋      | 138/376 [23:39<42:15, 10.66s/it]
Processing batches:  37%|███▋      | 139/376 [23:47<38:53,  9.85s/it]
Processing batches:  37%|███▋      | 140/376 [23:56<37:46,  9.61s/it]
Processing batches:  38%|███▊      | 141/376 [24:04<35:47,  9.14s/it]
Processing batches:  38%|███▊      | 142/376 [24:13<35:30,  9.10s/it]
Processing batches:  38%|███▊      | 143/376 [24:22<35:12,  9.07s/it]
Processing batches:  38%|███▊      | 144/376 [24:31<35:09,  9.09s/it]
Processing batches:  39%|███▊      | 145/376 [24:42<37:01,  9.62s/it]
Processing batches:  39%|███▉      | 146/376 [24:54<39:34, 10.32s/it]
Processing batches:  39%|███▉      | 147/376 [25:03<37:51,  9.92s/it]
Processing batches:  39%|███▉      | 148/376 [25:14<38:27, 10.12s/it]
Processing batches:  40%|███▉      | 149/376 [25:21<35:30,  9.39s/it]
Processing batches:  40%|███▉      | 150/376 [25:31<35:52,  9.52s/it]
Processing batches:  40%|████      | 151/376 [25:38<32:31,  8.67s/it]
Processing batches:  40%|████      | 152/376 [25:48<33:54,  9.08s/it]
Processing batches:  41%|████      | 153/376 [25:55<30:55,  8.32s/it]
Processing batches:  41%|████      | 154/376 [26:02<29:40,  8.02s/it]
Processing batches:  41%|████      | 155/376 [26:08<27:38,  7.51s/it]
Processing batches:  41%|████▏     | 156/376 [26:17<29:14,  7.98s/it]
Processing batches:  42%|████▏     | 157/376 [26:28<32:36,  8.93s/it]
Processing batches:  42%|████▏     | 158/376 [26:37<31:30,  8.67s/it]
Processing batches:  42%|████▏     | 159/376 [26:44<30:11,  8.35s/it]
Processing batches:  43%|████▎     | 160/376 [26:52<29:01,  8.06s/it]
Processing batches:  43%|████▎     | 161/376 [26:58<27:41,  7.73s/it]
Processing batches:  43%|████▎     | 162/376 [27:06<27:25,  7.69s/it]
Processing batches:  43%|████▎     | 163/376 [27:13<26:02,  7.33s/it]
Processing batches:  44%|████▎     | 164/376 [27:21<26:43,  7.57s/it]
Processing batches:  44%|████▍     | 165/376 [27:29<27:34,  7.84s/it]
Processing batches:  44%|████▍     | 166/376 [27:40<30:45,  8.79s/it]
Processing batches:  44%|████▍     | 167/376 [27:49<30:54,  8.87s/it]
Processing batches:  45%|████▍     | 168/376 [28:00<32:21,  9.33s/it]
Processing batches:  45%|████▍     | 169/376 [28:07<30:26,  8.82s/it]
Processing batches:  45%|████▌     | 170/376 [28:17<30:57,  9.02s/it]
Processing batches:  45%|████▌     | 171/376 [28:25<30:26,  8.91s/it]
Processing batches:  46%|████▌     | 172/376 [28:33<29:15,  8.61s/it]
Processing batches:  46%|████▌     | 173/376 [28:40<26:47,  7.92s/it]
Processing batches:  46%|████▋     | 174/376 [28:52<30:56,  9.19s/it]
Processing batches:  47%|████▋     | 175/376 [29:00<29:52,  8.92s/it]
Processing batches:  47%|████▋     | 176/376 [29:04<24:25,  7.33s/it]
Processing batches:  47%|████▋     | 177/376 [29:18<31:17,  9.44s/it]
Processing batches:  47%|████▋     | 178/376 [29:28<31:54,  9.67s/it]
Processing batches:  48%|████▊     | 179/376 [29:36<30:14,  9.21s/it]
Processing batches:  48%|████▊     | 180/376 [29:44<28:55,  8.86s/it]
Processing batches:  48%|████▊     | 181/376 [29:54<29:36,  9.11s/it]
Processing batches:  48%|████▊     | 182/376 [30:02<28:42,  8.88s/it]
Processing batches:  49%|████▊     | 183/376 [30:10<27:30,  8.55s/it]
Processing batches:  49%|████▉     | 184/376 [30:23<31:44,  9.92s/it]
Processing batches:  49%|████▉     | 185/376 [30:33<30:53,  9.71s/it]
Processing batches:  49%|████▉     | 186/376 [30:40<28:38,  9.04s/it]
Processing batches:  50%|████▉     | 187/376 [30:47<26:18,  8.35s/it]
Processing batches:  50%|█████     | 188/376 [30:54<24:46,  7.91s/it]
Processing batches:  50%|█████     | 189/376 [31:00<22:56,  7.36s/it]
Processing batches:  51%|█████     | 190/376 [31:08<23:19,  7.52s/it]
Processing batches:  51%|█████     | 191/376 [31:18<25:22,  8.23s/it]
Processing batches:  51%|█████     | 192/376 [31:31<29:36,  9.65s/it]
Processing batches:  51%|█████▏    | 193/376 [31:39<28:22,  9.31s/it]
Processing batches:  52%|█████▏    | 194/376 [31:49<28:32,  9.41s/it]
Processing batches:  52%|█████▏    | 195/376 [32:00<30:23, 10.07s/it]
Processing batches:  52%|█████▏    | 196/376 [32:16<35:35, 11.86s/it]
Processing batches:  52%|█████▏    | 197/376 [32:26<33:44, 11.31s/it]
Processing batches:  53%|█████▎    | 198/376 [32:34<30:17, 10.21s/it]
Processing batches:  53%|█████▎    | 199/376 [32:43<29:29, 10.00s/it]
Processing batches:  53%|█████▎    | 200/376 [32:50<26:17,  8.96s/it]
Processing batches:  53%|█████▎    | 201/376 [32:56<23:16,  7.98s/it]
Processing batches:  54%|█████▎    | 202/376 [33:01<20:26,  7.05s/it]
Processing batches:  54%|█████▍    | 203/376 [33:12<24:24,  8.47s/it]
Processing batches:  54%|█████▍    | 204/376 [33:24<26:58,  9.41s/it]
Processing batches:  55%|█████▍    | 205/376 [33:32<25:52,  9.08s/it]
Processing batches:  55%|█████▍    | 206/376 [33:42<26:24,  9.32s/it]
Processing batches:  55%|█████▌    | 207/376 [33:51<26:04,  9.26s/it]
Processing batches:  55%|█████▌    | 208/376 [34:05<29:31, 10.55s/it]
Processing batches:  56%|█████▌    | 209/376 [34:16<29:50, 10.72s/it]
Processing batches:  56%|█████▌    | 210/376 [34:31<33:28, 12.10s/it]
Processing batches:  56%|█████▌    | 211/376 [34:41<31:40, 11.52s/it]
Processing batches:  56%|█████▋    | 212/376 [34:50<28:57, 10.60s/it]
Processing batches:  57%|█████▋    | 213/376 [35:01<28:53, 10.63s/it]
Processing batches:  57%|█████▋    | 214/376 [35:10<27:20, 10.12s/it]
Processing batches:  57%|█████▋    | 215/376 [35:14<22:50,  8.51s/it]
Processing batches:  57%|█████▋    | 216/376 [35:19<19:51,  7.45s/it]
Processing batches:  58%|█████▊    | 217/376 [35:28<20:31,  7.75s/it]
Processing batches:  58%|█████▊    | 218/376 [35:37<21:17,  8.08s/it]
Processing batches:  58%|█████▊    | 219/376 [35:46<22:18,  8.53s/it]
Processing batches:  59%|█████▊    | 220/376 [35:55<22:43,  8.74s/it]
Processing batches:  59%|█████▉    | 221/376 [36:03<22:02,  8.53s/it]
Processing batches:  59%|█████▉    | 222/376 [36:14<23:10,  9.03s/it]
Processing batches:  59%|█████▉    | 223/376 [36:24<23:48,  9.34s/it]
Processing batches:  60%|█████▉    | 224/376 [36:31<22:10,  8.75s/it]
Processing batches:  60%|█████▉    | 225/376 [36:39<21:06,  8.39s/it]
Processing batches:  60%|██████    | 226/376 [36:47<20:46,  8.31s/it]
Processing batches:  60%|██████    | 227/376 [36:55<20:47,  8.37s/it]
Processing batches:  61%|██████    | 228/376 [37:03<20:18,  8.23s/it]
Processing batches:  61%|██████    | 229/376 [37:11<19:33,  7.98s/it]
Processing batches:  61%|██████    | 230/376 [37:20<20:35,  8.46s/it]
Processing batches:  61%|██████▏   | 231/376 [37:27<19:10,  7.94s/it]
Processing batches:  62%|██████▏   | 232/376 [37:36<19:40,  8.19s/it]
Processing batches:  62%|██████▏   | 233/376 [37:43<19:11,  8.05s/it]
Processing batches:  62%|██████▏   | 234/376 [37:48<16:42,  7.06s/it]
Processing batches:  62%|██████▎   | 235/376 [37:56<17:08,  7.30s/it]
Processing batches:  63%|██████▎   | 236/376 [38:02<16:11,  6.94s/it]
Processing batches:  63%|██████▎   | 237/376 [38:11<17:23,  7.51s/it]
Processing batches:  63%|██████▎   | 238/376 [38:21<18:46,  8.16s/it]
Processing batches:  64%|██████▎   | 239/376 [38:29<19:05,  8.36s/it]
Processing batches:  64%|██████▍   | 240/376 [38:38<19:06,  8.43s/it]
Processing batches:  64%|██████▍   | 241/376 [38:46<18:36,  8.27s/it]
Processing batches:  64%|██████▍   | 242/376 [38:59<21:25,  9.59s/it]
Processing batches:  65%|██████▍   | 243/376 [39:08<20:52,  9.42s/it]
Processing batches:  65%|██████▍   | 244/376 [39:15<19:41,  8.95s/it]
Processing batches:  65%|██████▌   | 245/376 [39:31<23:46, 10.89s/it]
Processing batches:  65%|██████▌   | 246/376 [39:37<20:49,  9.61s/it]
Processing batches:  66%|██████▌   | 247/376 [39:45<19:09,  8.91s/it]
Processing batches:  66%|██████▌   | 248/376 [39:50<16:51,  7.90s/it]
Processing batches:  66%|██████▌   | 249/376 [39:54<14:02,  6.63s/it]
Processing batches:  66%|██████▋   | 250/376 [39:59<12:39,  6.03s/it]
Processing batches:  67%|██████▋   | 251/376 [40:03<11:32,  5.54s/it]
Processing batches:  67%|██████▋   | 252/376 [40:09<11:45,  5.69s/it]
Processing batches:  67%|██████▋   | 253/376 [40:16<12:23,  6.04s/it]
Processing batches:  68%|██████▊   | 254/376 [40:24<13:14,  6.51s/it]
Processing batches:  68%|██████▊   | 255/376 [40:32<14:21,  7.12s/it]
Processing batches:  68%|██████▊   | 256/376 [40:40<14:46,  7.39s/it]
Processing batches:  68%|██████▊   | 257/376 [40:46<13:57,  7.04s/it]
Processing batches:  69%|██████▊   | 258/376 [40:55<14:47,  7.52s/it]
Processing batches:  69%|██████▉   | 259/376 [41:04<15:37,  8.02s/it]
Processing batches:  69%|██████▉   | 260/376 [41:15<17:12,  8.90s/it]
Processing batches:  69%|██████▉   | 261/376 [41:21<15:09,  7.91s/it]
Processing batches:  70%|██████▉   | 262/376 [41:28<14:42,  7.74s/it]
Processing batches:  70%|██████▉   | 263/376 [41:36<14:39,  7.79s/it]
Processing batches:  70%|███████   | 264/376 [41:46<15:44,  8.43s/it]
Processing batches:  70%|███████   | 265/376 [41:55<16:14,  8.78s/it]
Processing batches:  71%|███████   | 266/376 [42:04<16:05,  8.78s/it]
Processing batches:  71%|███████   | 267/376 [42:12<15:17,  8.42s/it]
Processing batches:  71%|███████▏  | 268/376 [42:18<14:00,  7.78s/it]
Processing batches:  72%|███████▏  | 269/376 [42:26<13:45,  7.72s/it]
Processing batches:  72%|███████▏  | 270/376 [42:32<13:03,  7.39s/it]
Processing batches:  72%|███████▏  | 271/376 [42:39<12:22,  7.07s/it]
Processing batches:  72%|███████▏  | 272/376 [42:50<14:24,  8.31s/it]
Processing batches:  73%|███████▎  | 273/376 [42:58<14:19,  8.35s/it]
Processing batches:  73%|███████▎  | 274/376 [43:06<13:52,  8.16s/it]
Processing batches:  73%|███████▎  | 275/376 [43:10<11:33,  6.86s/it]
Processing batches:  73%|███████▎  | 276/376 [43:21<13:32,  8.12s/it]
Processing batches:  74%|███████▎  | 277/376 [43:26<11:53,  7.21s/it]
Processing batches:  74%|███████▍  | 278/376 [43:38<14:23,  8.81s/it]
Processing batches:  74%|███████▍  | 279/376 [43:43<11:57,  7.39s/it]
Processing batches:  74%|███████▍  | 280/376 [43:49<11:28,  7.17s/it]
Processing batches:  75%|███████▍  | 281/376 [43:56<11:14,  7.10s/it]
Processing batches:  75%|███████▌  | 282/376 [44:05<11:42,  7.48s/it]
Processing batches:  75%|███████▌  | 283/376 [44:10<10:34,  6.82s/it]
Processing batches:  76%|███████▌  | 284/376 [44:15<09:32,  6.22s/it]
Processing batches:  76%|███████▌  | 285/376 [44:18<08:12,  5.42s/it]
Processing batches:  76%|███████▌  | 286/376 [44:19<05:58,  3.98s/it]
Processing batches:  76%|███████▋  | 287/376 [44:27<07:41,  5.19s/it]
Processing batches:  77%|███████▋  | 288/376 [44:34<08:19,  5.68s/it]
Processing batches:  77%|███████▋  | 289/376 [44:41<09:07,  6.29s/it]
Processing batches:  77%|███████▋  | 290/376 [44:49<09:32,  6.66s/it]
Processing batches:  77%|███████▋  | 291/376 [44:55<09:09,  6.46s/it]
Processing batches:  78%|███████▊  | 292/376 [45:02<09:23,  6.70s/it]
Processing batches:  78%|███████▊  | 293/376 [45:10<09:40,  6.99s/it]
Processing batches:  78%|███████▊  | 294/376 [45:17<09:45,  7.14s/it]
Processing batches:  78%|███████▊  | 295/376 [45:22<08:28,  6.27s/it]
Processing batches:  79%|███████▊  | 296/376 [45:27<08:05,  6.07s/it]
Processing batches:  79%|███████▉  | 297/376 [45:33<07:53,  6.00s/it]
Processing batches:  79%|███████▉  | 298/376 [45:41<08:28,  6.52s/it]
Processing batches:  80%|███████▉  | 299/376 [45:45<07:32,  5.87s/it]
Processing batches:  80%|███████▉  | 300/376 [45:50<06:58,  5.51s/it]
Processing batches:  80%|████████  | 301/376 [45:58<07:45,  6.21s/it]
Processing batches:  80%|████████  | 302/376 [46:06<08:19,  6.75s/it]
Processing batches:  81%|████████  | 303/376 [46:12<08:04,  6.64s/it]
Processing batches:  81%|████████  | 304/376 [46:17<07:32,  6.29s/it]
Processing batches:  81%|████████  | 305/376 [46:23<07:06,  6.00s/it]
Processing batches:  81%|████████▏ | 306/376 [46:28<06:40,  5.73s/it]
Processing batches:  82%|████████▏ | 307/376 [46:36<07:23,  6.42s/it]
Processing batches:  82%|████████▏ | 308/376 [46:43<07:28,  6.59s/it]
Processing batches:  82%|████████▏ | 309/376 [46:48<06:42,  6.01s/it]
Processing batches:  82%|████████▏ | 310/376 [46:53<06:35,  5.99s/it]
Processing batches:  83%|████████▎ | 311/376 [47:04<08:02,  7.42s/it]
Processing batches:  83%|████████▎ | 312/376 [47:08<06:43,  6.30s/it]
Processing batches:  83%|████████▎ | 313/376 [47:16<07:09,  6.81s/it]
Processing batches:  84%|████████▎ | 314/376 [47:22<06:49,  6.61s/it]
Processing batches:  84%|████████▍ | 315/376 [47:32<07:37,  7.50s/it]
Processing batches:  84%|████████▍ | 316/376 [47:38<07:06,  7.11s/it]
Processing batches:  84%|████████▍ | 317/376 [47:47<07:29,  7.61s/it]
Processing batches:  85%|████████▍ | 318/376 [47:55<07:29,  7.75s/it]
Processing batches:  85%|████████▍ | 319/376 [48:04<07:55,  8.33s/it]
Processing batches:  85%|████████▌ | 320/376 [48:10<06:55,  7.42s/it]
Processing batches:  85%|████████▌ | 321/376 [48:16<06:27,  7.04s/it]
Processing batches:  86%|████████▌ | 322/376 [48:21<05:49,  6.47s/it]
Processing batches:  86%|████████▌ | 323/376 [48:26<05:17,  5.99s/it]
Processing batches:  86%|████████▌ | 324/376 [48:34<05:50,  6.73s/it]
Processing batches:  86%|████████▋ | 325/376 [48:40<05:32,  6.52s/it]
Processing batches:  87%|████████▋ | 326/376 [48:44<04:48,  5.78s/it]
Processing batches:  87%|████████▋ | 327/376 [48:52<05:05,  6.23s/it]
Processing batches:  87%|████████▋ | 328/376 [49:01<05:40,  7.09s/it]
Processing batches:  88%|████████▊ | 329/376 [49:07<05:21,  6.84s/it]
Processing batches:  88%|████████▊ | 330/376 [49:16<05:46,  7.53s/it]
Processing batches:  88%|████████▊ | 331/376 [49:20<04:43,  6.29s/it]
Processing batches:  88%|████████▊ | 332/376 [49:26<04:34,  6.24s/it]
Processing batches:  89%|████████▊ | 333/376 [49:29<03:53,  5.44s/it]
Processing batches:  89%|████████▉ | 334/376 [49:37<04:16,  6.11s/it]
Processing batches:  89%|████████▉ | 335/376 [49:46<04:42,  6.89s/it]
Processing batches:  89%|████████▉ | 336/376 [49:54<04:48,  7.22s/it]
Processing batches:  90%|████████▉ | 337/376 [50:00<04:34,  7.05s/it]
Processing batches:  90%|████████▉ | 338/376 [50:05<04:01,  6.36s/it]
Processing batches:  90%|█████████ | 339/376 [50:09<03:29,  5.66s/it]
Processing batches:  90%|█████████ | 340/376 [50:14<03:11,  5.33s/it]
Processing batches:  91%|█████████ | 341/376 [50:19<03:06,  5.33s/it]
Processing batches:  91%|█████████ | 342/376 [50:28<03:39,  6.45s/it]
Processing batches:  91%|█████████ | 343/376 [50:35<03:40,  6.69s/it]
Processing batches:  91%|█████████▏| 344/376 [50:39<03:08,  5.90s/it]
Processing batches:  92%|█████████▏| 345/376 [50:44<02:54,  5.61s/it]
Processing batches:  92%|█████████▏| 346/376 [50:50<02:52,  5.75s/it]
Processing batches:  92%|█████████▏| 347/376 [50:55<02:33,  5.29s/it]
Processing batches:  93%|█████████▎| 348/376 [51:00<02:26,  5.23s/it]
Processing batches:  93%|█████████▎| 349/376 [51:06<02:31,  5.63s/it]
Processing batches:  93%|█████████▎| 350/376 [51:12<02:28,  5.72s/it]
Processing batches:  93%|█████████▎| 351/376 [51:18<02:25,  5.83s/it]
Processing batches:  94%|█████████▎| 352/376 [51:25<02:28,  6.18s/it]
Processing batches:  94%|█████████▍| 353/376 [51:29<02:05,  5.46s/it]
Processing batches:  94%|█████████▍| 354/376 [51:34<01:58,  5.38s/it]
Processing batches:  94%|█████████▍| 355/376 [51:38<01:45,  5.03s/it]
Processing batches:  95%|█████████▍| 356/376 [51:43<01:40,  5.03s/it]
Processing batches:  95%|█████████▍| 357/376 [51:48<01:31,  4.79s/it]
Processing batches:  95%|█████████▌| 358/376 [52:01<02:12,  7.37s/it]
Processing batches:  95%|█████████▌| 359/376 [52:05<01:45,  6.23s/it]
Processing batches:  96%|█████████▌| 360/376 [52:08<01:25,  5.37s/it]
Processing batches:  96%|█████████▌| 361/376 [52:12<01:14,  4.99s/it]
Processing batches:  96%|█████████▋| 362/376 [52:17<01:08,  4.88s/it]
Processing batches:  97%|█████████▋| 363/376 [52:20<00:56,  4.36s/it]
Processing batches:  97%|█████████▋| 364/376 [52:25<00:55,  4.62s/it]
Processing batches:  97%|█████████▋| 365/376 [52:30<00:51,  4.69s/it]
Processing batches:  97%|█████████▋| 366/376 [52:34<00:44,  4.48s/it]
Processing batches:  98%|█████████▊| 367/376 [52:37<00:37,  4.21s/it]
Processing batches:  98%|█████████▊| 368/376 [52:41<00:32,  4.00s/it]
Processing batches:  98%|█████████▊| 369/376 [52:45<00:27,  3.87s/it]
Processing batches:  98%|█████████▊| 370/376 [52:48<00:22,  3.82s/it]
Processing batches:  99%|█████████▊| 371/376 [52:52<00:19,  3.93s/it]
Processing batches:  99%|█████████▉| 372/376 [52:57<00:16,  4.13s/it]
Processing batches:  99%|█████████▉| 373/376 [53:01<00:12,  4.02s/it]
Processing batches:  99%|█████████▉| 374/376 [53:05<00:08,  4.04s/it]
Processing batches: 100%|█████████▉| 375/376 [53:09<00:03,  3.92s/it]
Processing batches: 100%|██████████| 376/376 [53:10<00:00,  3.13s/it]
Processing batches: 100%|██████████| 376/376 [53:10<00:00,  8.48s/it]
articles_mat: (768, 3753)
authors_mat: (768, 720)
words_mat: (768, 1981)

================================

Run params : {}


--------------------------------

Scoring params : {}

================================


----------- Scores -----------
neighbors_articles_authors : 0.1670

--------- Desc Scores ---------
neighbors_articles_authors_det

0             Mario Jelovic : 0.46
1            Aditi Majumder : 0.43
2           Attila Gyulassy : 0.42
3      Natalia V. Andrienko : 0.42
4          Jeff W. Lichtman : 0.42
                  ...
715            Nam Wook Kim : 0.10
716            Yuanzhe Chen : 0.10
717           Bahador Saket : 0.10
718             Meichun Hsu : 0.10
719           Scott Barlowe : 0.10
Length: 720, dtype: object

VALUE : 0.1670

--------- Raw Scores ---------
neighbors_articles_authors

Michael Sedlmair       0.156522
Dieter Schmalstieg     0.144444
Matthew Kay 0001       0.280000
Xingbo Wang 0001       0.150000
Minfeng Zhu            0.100000
                         ...
Lisa M. Sobierajski    0.150000
Nahum D. Gershon       0.120000
Sidney W. Wang         0.150000
David A. Lane          0.250000
T. Todd Elvins         0.150000
Length: 720, dtype: float64