70 KiB
70 KiB
Imports
import itertools
import lzma
import regex as re
import torch
from torch import nn
from torch.utils.data import IterableDataset, DataLoader
from torchtext.vocab import build_vocab_from_iterator
from google.colab import drive
Definitions
Functions
def clean_text(line: str):
# Preprocessing
separated = line.split('\t')
prefix = separated[6].replace(r'\n', ' ').replace('\\\\n', ' ').replace(' ', ' ').replace('.', '').replace(',', '').replace('?', '').replace('!', '').replace('(', '').replace(')', '').replace(';', '').replace(':', '').replace('"', '').replace("'", '').replace('-', ' ').replace(' ', ' ')
suffix = separated[7].replace(r'\n', ' ').replace('\\\\n', ' ').replace(' ', ' ').replace('.', '').replace(',', '').replace('?', '').replace('!', '').replace('(', '').replace(')', '').replace(';', '').replace(':', '').replace('"', '').replace("'", '').replace('-', ' ').replace(' ', ' ')
return prefix + ' ' + suffix
def get_words_from_line(line):
line = clean_text(line)
for word in line.split():
yield word
def get_word_lines_from_file(file_name):
with lzma.open(file_name, mode='rt', encoding='utf-8') as fid:
for line in fid:
yield get_words_from_line(line)
def look_ahead_iterator(gen):
prev = None
for item in gen:
if prev is not None:
yield (prev, item)
prev = item
def prediction(word: str) -> str:
ixs = torch.tensor(vocab.forward([word])).to(device)
out = model(ixs)
top = torch.topk(out[0], 5)
top_indices = top.indices.tolist()
top_probs = top.values.tolist()
top_words = vocab.lookup_tokens(top_indices)
zipped = list(zip(top_words, top_probs))
for index, element in enumerate(zipped):
unk = None
if '<unk>' in element:
unk = zipped.pop(index)
zipped.append(('', unk[1]))
break
if unk is None:
zipped[-1] = ('', zipped[-1][1])
return ' '.join([f'{x[0]}:{x[1]}' for x in zipped])
def save_outs(folder_name):
print(f'Creating outputs in {folder_name}')
with lzma.open(f'/content/drive/MyDrive/Colab Notebooks/{folder_name}/in.tsv.xz', mode='rt', encoding='utf-8') as fid:
with open(f'/content/drive/MyDrive/Colab Notebooks/{folder_name}/out.tsv', 'w', encoding='utf-8', newline='\n') as f:
for line in fid:
separated = line.split('\t')
prefix = separated[6].replace(r'\n', ' ').split()[-1]
output_line = prediction(prefix)
f.write(output_line + '\n')
Classes
class Bigrams(IterableDataset):
def __init__(self, text_file, vocabulary_size):
self.vocab = build_vocab_from_iterator(
get_word_lines_from_file(text_file),
max_tokens=vocabulary_size,
specials=['<unk>'])
self.vocab.set_default_index(self.vocab['<unk>'])
self.vocabulary_size = vocabulary_size
self.text_file = text_file
def __iter__(self):
return look_ahead_iterator(
(self.vocab[t] for t in itertools.chain.from_iterable(get_word_lines_from_file(self.text_file))))
class SimpleBigramNeuralLanguageModel(nn.Module):
def __init__(self, vocabulary_size, embedding_size):
super(SimpleBigramNeuralLanguageModel, self).__init__()
self.model = nn.Sequential(
nn.Embedding(vocabulary_size, embedding_size),
nn.Linear(embedding_size, vocabulary_size),
nn.Softmax()
)
def forward(self, x):
return self.model(x)
Training
Params
vocab_size = 10000
embed_size = 100
batch_size = 2000
device = 'cuda'
path_to_train = '/content/drive/MyDrive/Colab Notebooks/train/in.tsv.xz'
path_to_model = 'modelneural_bigram.bin'
Colab
drive.mount('/content/drive')
%cd /content/drive/MyDrive/
Mounted at /content/drive /content/drive/MyDrive
Run
vocab = build_vocab_from_iterator(
get_word_lines_from_file(path_to_train),
max_tokens=vocab_size,
specials=['<unk>']
)
vocab.set_default_index(vocab['<unk>'])
train_dataset = Bigrams(path_to_train, vocab_size)
model = SimpleBigramNeuralLanguageModel(vocab_size, embed_size).to(device)
data = DataLoader(train_dataset, batch_size=batch_size)
optimizer = torch.optim.Adam(model.parameters())
criterion = torch.nn.NLLLoss()
model.train()
step = 0
for x, y in data:
x = x.to(device)
y = y.to(device)
optimizer.zero_grad()
ypredicted = model(x)
loss = criterion(torch.log(ypredicted), y)
if step % 100 == 0:
print(step, loss)
step += 1
loss.backward()
optimizer.step()
/usr/local/lib/python3.10/dist-packages/torch/nn/modules/container.py:217: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument. input = module(input)
0 tensor(9.4517, device='cuda:0', grad_fn=<NllLossBackward0>) 100 tensor(7.9341, device='cuda:0', grad_fn=<NllLossBackward0>) 200 tensor(7.1452, device='cuda:0', grad_fn=<NllLossBackward0>) 300 tensor(6.7956, device='cuda:0', grad_fn=<NllLossBackward0>) 400 tensor(6.4127, device='cuda:0', grad_fn=<NllLossBackward0>) 500 tensor(6.3407, device='cuda:0', grad_fn=<NllLossBackward0>) 600 tensor(6.2125, device='cuda:0', grad_fn=<NllLossBackward0>) 700 tensor(5.7817, device='cuda:0', grad_fn=<NllLossBackward0>) 800 tensor(5.7309, device='cuda:0', grad_fn=<NllLossBackward0>) 900 tensor(5.7419, device='cuda:0', grad_fn=<NllLossBackward0>) 1000 tensor(5.7372, device='cuda:0', grad_fn=<NllLossBackward0>) 1100 tensor(5.2804, device='cuda:0', grad_fn=<NllLossBackward0>) 1200 tensor(5.4610, device='cuda:0', grad_fn=<NllLossBackward0>) 1300 tensor(5.6610, device='cuda:0', grad_fn=<NllLossBackward0>) 1400 tensor(5.3070, device='cuda:0', grad_fn=<NllLossBackward0>) 1500 tensor(4.9666, device='cuda:0', grad_fn=<NllLossBackward0>) 1600 tensor(5.2102, device='cuda:0', grad_fn=<NllLossBackward0>) 1700 tensor(5.4919, device='cuda:0', grad_fn=<NllLossBackward0>) 1800 tensor(5.1968, device='cuda:0', grad_fn=<NllLossBackward0>) 1900 tensor(5.3336, device='cuda:0', grad_fn=<NllLossBackward0>) 2000 tensor(5.2387, device='cuda:0', grad_fn=<NllLossBackward0>) 2100 tensor(5.2247, device='cuda:0', grad_fn=<NllLossBackward0>) 2200 tensor(5.2544, device='cuda:0', grad_fn=<NllLossBackward0>) 2300 tensor(5.3343, device='cuda:0', grad_fn=<NllLossBackward0>) 2400 tensor(5.3077, device='cuda:0', grad_fn=<NllLossBackward0>) 2500 tensor(5.1209, device='cuda:0', grad_fn=<NllLossBackward0>) 2600 tensor(5.3806, device='cuda:0', grad_fn=<NllLossBackward0>) 2700 tensor(5.2865, device='cuda:0', grad_fn=<NllLossBackward0>) 2800 tensor(5.2625, device='cuda:0', grad_fn=<NllLossBackward0>) 2900 tensor(5.2476, device='cuda:0', grad_fn=<NllLossBackward0>) 3000 tensor(5.2663, device='cuda:0', grad_fn=<NllLossBackward0>) 3100 tensor(5.0200, device='cuda:0', grad_fn=<NllLossBackward0>) 3200 tensor(5.2324, device='cuda:0', grad_fn=<NllLossBackward0>) 3300 tensor(5.1963, device='cuda:0', grad_fn=<NllLossBackward0>) 3400 tensor(5.1108, device='cuda:0', grad_fn=<NllLossBackward0>) 3500 tensor(5.1499, device='cuda:0', grad_fn=<NllLossBackward0>) 3600 tensor(5.3241, device='cuda:0', grad_fn=<NllLossBackward0>) 3700 tensor(5.1977, device='cuda:0', grad_fn=<NllLossBackward0>) 3800 tensor(5.1466, device='cuda:0', grad_fn=<NllLossBackward0>) 3900 tensor(5.2557, device='cuda:0', grad_fn=<NllLossBackward0>) 4000 tensor(5.0468, device='cuda:0', grad_fn=<NllLossBackward0>) 4100 tensor(5.1882, device='cuda:0', grad_fn=<NllLossBackward0>) 4200 tensor(5.0748, device='cuda:0', grad_fn=<NllLossBackward0>) 4300 tensor(4.9577, device='cuda:0', grad_fn=<NllLossBackward0>) 4400 tensor(4.8100, device='cuda:0', grad_fn=<NllLossBackward0>) 4500 tensor(5.0355, device='cuda:0', grad_fn=<NllLossBackward0>) 4600 tensor(5.1247, device='cuda:0', grad_fn=<NllLossBackward0>) 4700 tensor(5.0516, device='cuda:0', grad_fn=<NllLossBackward0>) 4800 tensor(4.9036, device='cuda:0', grad_fn=<NllLossBackward0>) 4900 tensor(5.0096, device='cuda:0', grad_fn=<NllLossBackward0>) 5000 tensor(5.2085, device='cuda:0', grad_fn=<NllLossBackward0>) 5100 tensor(5.0944, device='cuda:0', grad_fn=<NllLossBackward0>) 5200 tensor(5.1592, device='cuda:0', grad_fn=<NllLossBackward0>) 5300 tensor(5.2019, device='cuda:0', grad_fn=<NllLossBackward0>) 5400 tensor(5.2048, device='cuda:0', grad_fn=<NllLossBackward0>) 5500 tensor(5.0499, device='cuda:0', grad_fn=<NllLossBackward0>) 5600 tensor(5.0369, device='cuda:0', grad_fn=<NllLossBackward0>) 5700 tensor(5.2581, device='cuda:0', grad_fn=<NllLossBackward0>) 5800 tensor(5.0312, device='cuda:0', grad_fn=<NllLossBackward0>) 5900 tensor(5.0513, device='cuda:0', grad_fn=<NllLossBackward0>) 6000 tensor(5.2384, device='cuda:0', grad_fn=<NllLossBackward0>) 6100 tensor(5.0257, device='cuda:0', grad_fn=<NllLossBackward0>) 6200 tensor(5.1156, device='cuda:0', grad_fn=<NllLossBackward0>) 6300 tensor(4.9953, device='cuda:0', grad_fn=<NllLossBackward0>) 6400 tensor(5.2028, device='cuda:0', grad_fn=<NllLossBackward0>) 6500 tensor(4.8426, device='cuda:0', grad_fn=<NllLossBackward0>) 6600 tensor(5.0661, device='cuda:0', grad_fn=<NllLossBackward0>) 6700 tensor(5.0976, device='cuda:0', grad_fn=<NllLossBackward0>) 6800 tensor(4.9180, device='cuda:0', grad_fn=<NllLossBackward0>) 6900 tensor(4.9928, device='cuda:0', grad_fn=<NllLossBackward0>) 7000 tensor(5.1889, device='cuda:0', grad_fn=<NllLossBackward0>) 7100 tensor(4.9612, device='cuda:0', grad_fn=<NllLossBackward0>) 7200 tensor(5.1408, device='cuda:0', grad_fn=<NllLossBackward0>) 7300 tensor(5.0562, device='cuda:0', grad_fn=<NllLossBackward0>) 7400 tensor(4.8779, device='cuda:0', grad_fn=<NllLossBackward0>) 7500 tensor(5.0490, device='cuda:0', grad_fn=<NllLossBackward0>) 7600 tensor(5.0678, device='cuda:0', grad_fn=<NllLossBackward0>) 7700 tensor(4.9938, device='cuda:0', grad_fn=<NllLossBackward0>) 7800 tensor(5.0301, device='cuda:0', grad_fn=<NllLossBackward0>) 7900 tensor(5.2542, device='cuda:0', grad_fn=<NllLossBackward0>) 8000 tensor(4.8772, device='cuda:0', grad_fn=<NllLossBackward0>) 8100 tensor(5.0953, device='cuda:0', grad_fn=<NllLossBackward0>) 8200 tensor(5.0217, device='cuda:0', grad_fn=<NllLossBackward0>) 8300 tensor(5.0107, device='cuda:0', grad_fn=<NllLossBackward0>) 8400 tensor(5.0733, device='cuda:0', grad_fn=<NllLossBackward0>) 8500 tensor(4.5262, device='cuda:0', grad_fn=<NllLossBackward0>) 8600 tensor(5.0271, device='cuda:0', grad_fn=<NllLossBackward0>) 8700 tensor(4.6307, device='cuda:0', grad_fn=<NllLossBackward0>) 8800 tensor(4.9917, device='cuda:0', grad_fn=<NllLossBackward0>) 8900 tensor(5.1940, device='cuda:0', grad_fn=<NllLossBackward0>) 9000 tensor(5.0302, device='cuda:0', grad_fn=<NllLossBackward0>) 9100 tensor(5.0956, device='cuda:0', grad_fn=<NllLossBackward0>) 9200 tensor(5.0438, device='cuda:0', grad_fn=<NllLossBackward0>) 9300 tensor(5.0134, device='cuda:0', grad_fn=<NllLossBackward0>) 9400 tensor(5.2201, device='cuda:0', grad_fn=<NllLossBackward0>) 9500 tensor(4.8876, device='cuda:0', grad_fn=<NllLossBackward0>) 9600 tensor(5.1474, device='cuda:0', grad_fn=<NllLossBackward0>) 9700 tensor(5.0169, device='cuda:0', grad_fn=<NllLossBackward0>) 9800 tensor(5.0743, device='cuda:0', grad_fn=<NllLossBackward0>) 9900 tensor(4.9008, device='cuda:0', grad_fn=<NllLossBackward0>) 10000 tensor(5.1381, device='cuda:0', grad_fn=<NllLossBackward0>) 10100 tensor(5.0524, device='cuda:0', grad_fn=<NllLossBackward0>) 10200 tensor(5.0369, device='cuda:0', grad_fn=<NllLossBackward0>) 10300 tensor(5.0595, device='cuda:0', grad_fn=<NllLossBackward0>) 10400 tensor(5.0138, device='cuda:0', grad_fn=<NllLossBackward0>) 10500 tensor(5.0164, device='cuda:0', grad_fn=<NllLossBackward0>) 10600 tensor(4.9153, device='cuda:0', grad_fn=<NllLossBackward0>) 10700 tensor(4.9971, device='cuda:0', grad_fn=<NllLossBackward0>) 10800 tensor(5.0200, device='cuda:0', grad_fn=<NllLossBackward0>) 10900 tensor(4.9631, device='cuda:0', grad_fn=<NllLossBackward0>) 11000 tensor(4.9385, device='cuda:0', grad_fn=<NllLossBackward0>) 11100 tensor(4.9851, device='cuda:0', grad_fn=<NllLossBackward0>) 11200 tensor(5.0681, device='cuda:0', grad_fn=<NllLossBackward0>) 11300 tensor(5.1261, device='cuda:0', grad_fn=<NllLossBackward0>) 11400 tensor(5.0098, device='cuda:0', grad_fn=<NllLossBackward0>) 11500 tensor(5.1261, device='cuda:0', grad_fn=<NllLossBackward0>) 11600 tensor(5.1213, device='cuda:0', grad_fn=<NllLossBackward0>) 11700 tensor(5.0265, device='cuda:0', grad_fn=<NllLossBackward0>) 11800 tensor(4.7047, device='cuda:0', grad_fn=<NllLossBackward0>) 11900 tensor(5.1954, device='cuda:0', grad_fn=<NllLossBackward0>) 12000 tensor(5.0850, device='cuda:0', grad_fn=<NllLossBackward0>) 12100 tensor(4.9762, device='cuda:0', grad_fn=<NllLossBackward0>) 12200 tensor(5.0162, device='cuda:0', grad_fn=<NllLossBackward0>) 12300 tensor(4.9834, device='cuda:0', grad_fn=<NllLossBackward0>) 12400 tensor(4.8953, device='cuda:0', grad_fn=<NllLossBackward0>) 12500 tensor(5.0389, device='cuda:0', grad_fn=<NllLossBackward0>) 12600 tensor(4.9266, device='cuda:0', grad_fn=<NllLossBackward0>) 12700 tensor(5.0132, device='cuda:0', grad_fn=<NllLossBackward0>) 12800 tensor(5.1777, device='cuda:0', grad_fn=<NllLossBackward0>) 12900 tensor(4.8290, device='cuda:0', grad_fn=<NllLossBackward0>) 13000 tensor(5.0639, device='cuda:0', grad_fn=<NllLossBackward0>) 13100 tensor(5.0565, device='cuda:0', grad_fn=<NllLossBackward0>) 13200 tensor(5.0222, device='cuda:0', grad_fn=<NllLossBackward0>) 13300 tensor(5.2150, device='cuda:0', grad_fn=<NllLossBackward0>) 13400 tensor(4.9393, device='cuda:0', grad_fn=<NllLossBackward0>) 13500 tensor(5.0270, device='cuda:0', grad_fn=<NllLossBackward0>) 13600 tensor(4.9520, device='cuda:0', grad_fn=<NllLossBackward0>) 13700 tensor(4.9845, device='cuda:0', grad_fn=<NllLossBackward0>) 13800 tensor(4.8543, device='cuda:0', grad_fn=<NllLossBackward0>) 13900 tensor(4.8892, device='cuda:0', grad_fn=<NllLossBackward0>) 14000 tensor(4.9802, device='cuda:0', grad_fn=<NllLossBackward0>) 14100 tensor(4.9833, device='cuda:0', grad_fn=<NllLossBackward0>) 14200 tensor(4.9348, device='cuda:0', grad_fn=<NllLossBackward0>) 14300 tensor(4.9561, device='cuda:0', grad_fn=<NllLossBackward0>) 14400 tensor(5.0198, device='cuda:0', grad_fn=<NllLossBackward0>) 14500 tensor(4.9878, device='cuda:0', grad_fn=<NllLossBackward0>) 14600 tensor(4.7517, device='cuda:0', grad_fn=<NllLossBackward0>) 14700 tensor(4.9452, device='cuda:0', grad_fn=<NllLossBackward0>) 14800 tensor(4.8229, device='cuda:0', grad_fn=<NllLossBackward0>) 14900 tensor(5.1425, device='cuda:0', grad_fn=<NllLossBackward0>) 15000 tensor(4.9122, device='cuda:0', grad_fn=<NllLossBackward0>) 15100 tensor(4.8217, device='cuda:0', grad_fn=<NllLossBackward0>) 15200 tensor(4.8604, device='cuda:0', grad_fn=<NllLossBackward0>) 15300 tensor(5.1151, device='cuda:0', grad_fn=<NllLossBackward0>) 15400 tensor(4.9545, device='cuda:0', grad_fn=<NllLossBackward0>) 15500 tensor(5.0922, device='cuda:0', grad_fn=<NllLossBackward0>) 15600 tensor(4.7891, device='cuda:0', grad_fn=<NllLossBackward0>) 15700 tensor(4.6318, device='cuda:0', grad_fn=<NllLossBackward0>) 15800 tensor(4.9540, device='cuda:0', grad_fn=<NllLossBackward0>) 15900 tensor(4.7681, device='cuda:0', grad_fn=<NllLossBackward0>) 16000 tensor(4.9602, device='cuda:0', grad_fn=<NllLossBackward0>) 16100 tensor(4.9705, device='cuda:0', grad_fn=<NllLossBackward0>) 16200 tensor(4.8296, device='cuda:0', grad_fn=<NllLossBackward0>) 16300 tensor(5.0188, device='cuda:0', grad_fn=<NllLossBackward0>) 16400 tensor(5.1062, device='cuda:0', grad_fn=<NllLossBackward0>) 16500 tensor(5.2549, device='cuda:0', grad_fn=<NllLossBackward0>) 16600 tensor(5.1164, device='cuda:0', grad_fn=<NllLossBackward0>) 16700 tensor(4.9399, device='cuda:0', grad_fn=<NllLossBackward0>) 16800 tensor(5.1161, device='cuda:0', grad_fn=<NllLossBackward0>) 16900 tensor(4.9115, device='cuda:0', grad_fn=<NllLossBackward0>) 17000 tensor(4.7572, device='cuda:0', grad_fn=<NllLossBackward0>) 17100 tensor(4.9667, device='cuda:0', grad_fn=<NllLossBackward0>) 17200 tensor(4.7463, device='cuda:0', grad_fn=<NllLossBackward0>) 17300 tensor(4.9038, device='cuda:0', grad_fn=<NllLossBackward0>) 17400 tensor(4.9859, device='cuda:0', grad_fn=<NllLossBackward0>) 17500 tensor(5.0652, device='cuda:0', grad_fn=<NllLossBackward0>) 17600 tensor(4.6641, device='cuda:0', grad_fn=<NllLossBackward0>) 17700 tensor(4.9265, device='cuda:0', grad_fn=<NllLossBackward0>) 17800 tensor(5.0095, device='cuda:0', grad_fn=<NllLossBackward0>) 17900 tensor(5.1090, device='cuda:0', grad_fn=<NllLossBackward0>) 18000 tensor(4.9015, device='cuda:0', grad_fn=<NllLossBackward0>) 18100 tensor(4.9997, device='cuda:0', grad_fn=<NllLossBackward0>) 18200 tensor(4.8359, device='cuda:0', grad_fn=<NllLossBackward0>) 18300 tensor(4.7353, device='cuda:0', grad_fn=<NllLossBackward0>) 18400 tensor(4.9657, device='cuda:0', grad_fn=<NllLossBackward0>) 18500 tensor(4.9856, device='cuda:0', grad_fn=<NllLossBackward0>) 18600 tensor(5.0571, device='cuda:0', grad_fn=<NllLossBackward0>) 18700 tensor(4.8566, device='cuda:0', grad_fn=<NllLossBackward0>) 18800 tensor(4.9819, device='cuda:0', grad_fn=<NllLossBackward0>) 18900 tensor(4.9809, device='cuda:0', grad_fn=<NllLossBackward0>) 19000 tensor(5.0202, device='cuda:0', grad_fn=<NllLossBackward0>) 19100 tensor(5.1329, device='cuda:0', grad_fn=<NllLossBackward0>) 19200 tensor(5.0460, device='cuda:0', grad_fn=<NllLossBackward0>) 19300 tensor(4.9174, device='cuda:0', grad_fn=<NllLossBackward0>) 19400 tensor(5.1266, device='cuda:0', grad_fn=<NllLossBackward0>) 19500 tensor(4.8903, device='cuda:0', grad_fn=<NllLossBackward0>) 19600 tensor(5.0548, device='cuda:0', grad_fn=<NllLossBackward0>) 19700 tensor(4.9530, device='cuda:0', grad_fn=<NllLossBackward0>) 19800 tensor(4.9296, device='cuda:0', grad_fn=<NllLossBackward0>) 19900 tensor(4.9925, device='cuda:0', grad_fn=<NllLossBackward0>) 20000 tensor(4.9181, device='cuda:0', grad_fn=<NllLossBackward0>) 20100 tensor(4.9487, device='cuda:0', grad_fn=<NllLossBackward0>) 20200 tensor(5.0580, device='cuda:0', grad_fn=<NllLossBackward0>) 20300 tensor(5.1110, device='cuda:0', grad_fn=<NllLossBackward0>) 20400 tensor(4.8053, device='cuda:0', grad_fn=<NllLossBackward0>) 20500 tensor(4.7658, device='cuda:0', grad_fn=<NllLossBackward0>) 20600 tensor(4.7387, device='cuda:0', grad_fn=<NllLossBackward0>) 20700 tensor(4.9779, device='cuda:0', grad_fn=<NllLossBackward0>) 20800 tensor(4.8901, device='cuda:0', grad_fn=<NllLossBackward0>) 20900 tensor(4.9092, device='cuda:0', grad_fn=<NllLossBackward0>) 21000 tensor(5.2856, device='cuda:0', grad_fn=<NllLossBackward0>) 21100 tensor(4.9803, device='cuda:0', grad_fn=<NllLossBackward0>) 21200 tensor(4.6889, device='cuda:0', grad_fn=<NllLossBackward0>) 21300 tensor(4.8434, device='cuda:0', grad_fn=<NllLossBackward0>) 21400 tensor(4.7451, device='cuda:0', grad_fn=<NllLossBackward0>) 21500 tensor(4.9406, device='cuda:0', grad_fn=<NllLossBackward0>) 21600 tensor(4.8431, device='cuda:0', grad_fn=<NllLossBackward0>) 21700 tensor(4.9932, device='cuda:0', grad_fn=<NllLossBackward0>) 21800 tensor(4.6696, device='cuda:0', grad_fn=<NllLossBackward0>) 21900 tensor(4.8091, device='cuda:0', grad_fn=<NllLossBackward0>) 22000 tensor(4.7533, device='cuda:0', grad_fn=<NllLossBackward0>) 22100 tensor(4.6842, device='cuda:0', grad_fn=<NllLossBackward0>) 22200 tensor(4.8844, device='cuda:0', grad_fn=<NllLossBackward0>) 22300 tensor(5.1038, device='cuda:0', grad_fn=<NllLossBackward0>) 22400 tensor(4.9929, device='cuda:0', grad_fn=<NllLossBackward0>) 22500 tensor(5.0109, device='cuda:0', grad_fn=<NllLossBackward0>) 22600 tensor(4.8278, device='cuda:0', grad_fn=<NllLossBackward0>) 22700 tensor(4.8597, device='cuda:0', grad_fn=<NllLossBackward0>) 22800 tensor(5.0256, device='cuda:0', grad_fn=<NllLossBackward0>) 22900 tensor(4.4663, device='cuda:0', grad_fn=<NllLossBackward0>) 23000 tensor(4.6069, device='cuda:0', grad_fn=<NllLossBackward0>) 23100 tensor(5.0816, device='cuda:0', grad_fn=<NllLossBackward0>) 23200 tensor(4.9038, device='cuda:0', grad_fn=<NllLossBackward0>) 23300 tensor(4.9284, device='cuda:0', grad_fn=<NllLossBackward0>) 23400 tensor(5.0439, device='cuda:0', grad_fn=<NllLossBackward0>) 23500 tensor(4.9640, device='cuda:0', grad_fn=<NllLossBackward0>) 23600 tensor(5.0096, device='cuda:0', grad_fn=<NllLossBackward0>) 23700 tensor(4.9700, device='cuda:0', grad_fn=<NllLossBackward0>) 23800 tensor(4.9461, device='cuda:0', grad_fn=<NllLossBackward0>) 23900 tensor(4.8171, device='cuda:0', grad_fn=<NllLossBackward0>) 24000 tensor(4.9529, device='cuda:0', grad_fn=<NllLossBackward0>) 24100 tensor(4.8525, device='cuda:0', grad_fn=<NllLossBackward0>) 24200 tensor(5.0488, device='cuda:0', grad_fn=<NllLossBackward0>) 24300 tensor(4.9206, device='cuda:0', grad_fn=<NllLossBackward0>) 24400 tensor(5.0900, device='cuda:0', grad_fn=<NllLossBackward0>) 24500 tensor(4.9484, device='cuda:0', grad_fn=<NllLossBackward0>) 24600 tensor(4.8962, device='cuda:0', grad_fn=<NllLossBackward0>) 24700 tensor(4.8884, device='cuda:0', grad_fn=<NllLossBackward0>) 24800 tensor(5.1541, device='cuda:0', grad_fn=<NllLossBackward0>) 24900 tensor(4.9803, device='cuda:0', grad_fn=<NllLossBackward0>) 25000 tensor(4.4473, device='cuda:0', grad_fn=<NllLossBackward0>) 25100 tensor(4.7330, device='cuda:0', grad_fn=<NllLossBackward0>) 25200 tensor(5.0709, device='cuda:0', grad_fn=<NllLossBackward0>) 25300 tensor(4.7139, device='cuda:0', grad_fn=<NllLossBackward0>) 25400 tensor(4.8961, device='cuda:0', grad_fn=<NllLossBackward0>) 25500 tensor(4.9459, device='cuda:0', grad_fn=<NllLossBackward0>) 25600 tensor(4.8840, device='cuda:0', grad_fn=<NllLossBackward0>) 25700 tensor(4.7792, device='cuda:0', grad_fn=<NllLossBackward0>) 25800 tensor(4.9212, device='cuda:0', grad_fn=<NllLossBackward0>) 25900 tensor(4.7168, device='cuda:0', grad_fn=<NllLossBackward0>) 26000 tensor(4.7903, device='cuda:0', grad_fn=<NllLossBackward0>) 26100 tensor(4.9544, device='cuda:0', grad_fn=<NllLossBackward0>) 26200 tensor(4.8421, device='cuda:0', grad_fn=<NllLossBackward0>) 26300 tensor(4.8085, device='cuda:0', grad_fn=<NllLossBackward0>) 26400 tensor(4.7129, device='cuda:0', grad_fn=<NllLossBackward0>) 26500 tensor(5.0808, device='cuda:0', grad_fn=<NllLossBackward0>) 26600 tensor(4.8222, device='cuda:0', grad_fn=<NllLossBackward0>) 26700 tensor(4.7982, device='cuda:0', grad_fn=<NllLossBackward0>) 26800 tensor(4.8482, device='cuda:0', grad_fn=<NllLossBackward0>) 26900 tensor(5.0815, device='cuda:0', grad_fn=<NllLossBackward0>) 27000 tensor(4.9754, device='cuda:0', grad_fn=<NllLossBackward0>) 27100 tensor(5.0156, device='cuda:0', grad_fn=<NllLossBackward0>) 27200 tensor(4.7985, device='cuda:0', grad_fn=<NllLossBackward0>) 27300 tensor(4.6372, device='cuda:0', grad_fn=<NllLossBackward0>) 27400 tensor(4.5098, device='cuda:0', grad_fn=<NllLossBackward0>) 27500 tensor(5.0427, device='cuda:0', grad_fn=<NllLossBackward0>) 27600 tensor(4.9139, device='cuda:0', grad_fn=<NllLossBackward0>) 27700 tensor(4.8924, device='cuda:0', grad_fn=<NllLossBackward0>) 27800 tensor(4.9972, device='cuda:0', grad_fn=<NllLossBackward0>) 27900 tensor(5.0452, device='cuda:0', grad_fn=<NllLossBackward0>) 28000 tensor(4.5323, device='cuda:0', grad_fn=<NllLossBackward0>) 28100 tensor(4.8945, device='cuda:0', grad_fn=<NllLossBackward0>) 28200 tensor(4.8096, device='cuda:0', grad_fn=<NllLossBackward0>) 28300 tensor(5.1238, device='cuda:0', grad_fn=<NllLossBackward0>) 28400 tensor(4.9879, device='cuda:0', grad_fn=<NllLossBackward0>) 28500 tensor(4.9505, device='cuda:0', grad_fn=<NllLossBackward0>) 28600 tensor(4.7750, device='cuda:0', grad_fn=<NllLossBackward0>) 28700 tensor(5.0738, device='cuda:0', grad_fn=<NllLossBackward0>) 28800 tensor(4.9318, device='cuda:0', grad_fn=<NllLossBackward0>) 28900 tensor(5.0403, device='cuda:0', grad_fn=<NllLossBackward0>) 29000 tensor(4.9072, device='cuda:0', grad_fn=<NllLossBackward0>) 29100 tensor(4.9822, device='cuda:0', grad_fn=<NllLossBackward0>) 29200 tensor(4.8701, device='cuda:0', grad_fn=<NllLossBackward0>) 29300 tensor(4.8883, device='cuda:0', grad_fn=<NllLossBackward0>) 29400 tensor(4.8906, device='cuda:0', grad_fn=<NllLossBackward0>) 29500 tensor(5.0658, device='cuda:0', grad_fn=<NllLossBackward0>) 29600 tensor(4.7604, device='cuda:0', grad_fn=<NllLossBackward0>) 29700 tensor(5.0792, device='cuda:0', grad_fn=<NllLossBackward0>) 29800 tensor(4.9074, device='cuda:0', grad_fn=<NllLossBackward0>) 29900 tensor(4.8845, device='cuda:0', grad_fn=<NllLossBackward0>) 30000 tensor(5.1969, device='cuda:0', grad_fn=<NllLossBackward0>) 30100 tensor(4.9648, device='cuda:0', grad_fn=<NllLossBackward0>) 30200 tensor(4.9086, device='cuda:0', grad_fn=<NllLossBackward0>) 30300 tensor(4.9708, device='cuda:0', grad_fn=<NllLossBackward0>) 30400 tensor(4.9155, device='cuda:0', grad_fn=<NllLossBackward0>) 30500 tensor(4.9404, device='cuda:0', grad_fn=<NllLossBackward0>) 30600 tensor(5.0224, device='cuda:0', grad_fn=<NllLossBackward0>) 30700 tensor(5.0298, device='cuda:0', grad_fn=<NllLossBackward0>) 30800 tensor(4.9557, device='cuda:0', grad_fn=<NllLossBackward0>) 30900 tensor(4.9653, device='cuda:0', grad_fn=<NllLossBackward0>) 31000 tensor(4.8938, device='cuda:0', grad_fn=<NllLossBackward0>) 31100 tensor(4.6689, device='cuda:0', grad_fn=<NllLossBackward0>) 31200 tensor(4.9757, device='cuda:0', grad_fn=<NllLossBackward0>) 31300 tensor(4.8805, device='cuda:0', grad_fn=<NllLossBackward0>) 31400 tensor(4.9969, device='cuda:0', grad_fn=<NllLossBackward0>) 31500 tensor(4.8262, device='cuda:0', grad_fn=<NllLossBackward0>) 31600 tensor(4.5519, device='cuda:0', grad_fn=<NllLossBackward0>) 31700 tensor(4.9185, device='cuda:0', grad_fn=<NllLossBackward0>) 31800 tensor(4.9190, device='cuda:0', grad_fn=<NllLossBackward0>) 31900 tensor(4.8702, device='cuda:0', grad_fn=<NllLossBackward0>) 32000 tensor(4.9346, device='cuda:0', grad_fn=<NllLossBackward0>) 32100 tensor(4.8963, device='cuda:0', grad_fn=<NllLossBackward0>) 32200 tensor(4.9017, device='cuda:0', grad_fn=<NllLossBackward0>) 32300 tensor(4.9595, device='cuda:0', grad_fn=<NllLossBackward0>) 32400 tensor(4.8125, device='cuda:0', grad_fn=<NllLossBackward0>) 32500 tensor(4.9593, device='cuda:0', grad_fn=<NllLossBackward0>) 32600 tensor(5.0663, device='cuda:0', grad_fn=<NllLossBackward0>) 32700 tensor(4.9644, device='cuda:0', grad_fn=<NllLossBackward0>) 32800 tensor(4.8500, device='cuda:0', grad_fn=<NllLossBackward0>) 32900 tensor(5.0070, device='cuda:0', grad_fn=<NllLossBackward0>) 33000 tensor(4.8131, device='cuda:0', grad_fn=<NllLossBackward0>) 33100 tensor(5.0183, device='cuda:0', grad_fn=<NllLossBackward0>) 33200 tensor(4.8692, device='cuda:0', grad_fn=<NllLossBackward0>) 33300 tensor(4.9145, device='cuda:0', grad_fn=<NllLossBackward0>) 33400 tensor(5.0221, device='cuda:0', grad_fn=<NllLossBackward0>) 33500 tensor(4.9636, device='cuda:0', grad_fn=<NllLossBackward0>) 33600 tensor(4.8758, device='cuda:0', grad_fn=<NllLossBackward0>) 33700 tensor(4.8713, device='cuda:0', grad_fn=<NllLossBackward0>) 33800 tensor(4.7325, device='cuda:0', grad_fn=<NllLossBackward0>) 33900 tensor(4.9829, device='cuda:0', grad_fn=<NllLossBackward0>) 34000 tensor(4.7823, device='cuda:0', grad_fn=<NllLossBackward0>) 34100 tensor(4.9773, device='cuda:0', grad_fn=<NllLossBackward0>) 34200 tensor(4.9638, device='cuda:0', grad_fn=<NllLossBackward0>) 34300 tensor(5.0311, device='cuda:0', grad_fn=<NllLossBackward0>) 34400 tensor(4.9491, device='cuda:0', grad_fn=<NllLossBackward0>) 34500 tensor(4.9527, device='cuda:0', grad_fn=<NllLossBackward0>) 34600 tensor(4.7559, device='cuda:0', grad_fn=<NllLossBackward0>) 34700 tensor(4.9602, device='cuda:0', grad_fn=<NllLossBackward0>) 34800 tensor(5.0363, device='cuda:0', grad_fn=<NllLossBackward0>) 34900 tensor(4.9509, device='cuda:0', grad_fn=<NllLossBackward0>) 35000 tensor(4.8740, device='cuda:0', grad_fn=<NllLossBackward0>) 35100 tensor(4.8790, device='cuda:0', grad_fn=<NllLossBackward0>) 35200 tensor(4.7886, device='cuda:0', grad_fn=<NllLossBackward0>) 35300 tensor(4.9939, device='cuda:0', grad_fn=<NllLossBackward0>) 35400 tensor(4.8046, device='cuda:0', grad_fn=<NllLossBackward0>) 35500 tensor(5.0125, device='cuda:0', grad_fn=<NllLossBackward0>) 35600 tensor(4.8254, device='cuda:0', grad_fn=<NllLossBackward0>) 35700 tensor(4.5858, device='cuda:0', grad_fn=<NllLossBackward0>) 35800 tensor(5.0067, device='cuda:0', grad_fn=<NllLossBackward0>) 35900 tensor(5.0505, device='cuda:0', grad_fn=<NllLossBackward0>) 36000 tensor(4.9909, device='cuda:0', grad_fn=<NllLossBackward0>) 36100 tensor(4.8610, device='cuda:0', grad_fn=<NllLossBackward0>) 36200 tensor(4.9135, device='cuda:0', grad_fn=<NllLossBackward0>) 36300 tensor(5.0409, device='cuda:0', grad_fn=<NllLossBackward0>) 36400 tensor(4.8932, device='cuda:0', grad_fn=<NllLossBackward0>) 36500 tensor(4.8384, device='cuda:0', grad_fn=<NllLossBackward0>) 36600 tensor(4.8262, device='cuda:0', grad_fn=<NllLossBackward0>) 36700 tensor(4.8363, device='cuda:0', grad_fn=<NllLossBackward0>) 36800 tensor(4.9260, device='cuda:0', grad_fn=<NllLossBackward0>) 36900 tensor(4.7176, device='cuda:0', grad_fn=<NllLossBackward0>) 37000 tensor(4.8836, device='cuda:0', grad_fn=<NllLossBackward0>) 37100 tensor(4.7659, device='cuda:0', grad_fn=<NllLossBackward0>) 37200 tensor(5.0418, device='cuda:0', grad_fn=<NllLossBackward0>) 37300 tensor(4.7165, device='cuda:0', grad_fn=<NllLossBackward0>) 37400 tensor(4.7707, device='cuda:0', grad_fn=<NllLossBackward0>) 37500 tensor(4.9404, device='cuda:0', grad_fn=<NllLossBackward0>) 37600 tensor(4.7666, device='cuda:0', grad_fn=<NllLossBackward0>) 37700 tensor(5.0086, device='cuda:0', grad_fn=<NllLossBackward0>) 37800 tensor(4.8929, device='cuda:0', grad_fn=<NllLossBackward0>) 37900 tensor(5.0537, device='cuda:0', grad_fn=<NllLossBackward0>) 38000 tensor(4.8494, device='cuda:0', grad_fn=<NllLossBackward0>) 38100 tensor(5.1193, device='cuda:0', grad_fn=<NllLossBackward0>) 38200 tensor(4.9035, device='cuda:0', grad_fn=<NllLossBackward0>) 38300 tensor(4.7574, device='cuda:0', grad_fn=<NllLossBackward0>) 38400 tensor(4.9181, device='cuda:0', grad_fn=<NllLossBackward0>) 38500 tensor(5.0186, device='cuda:0', grad_fn=<NllLossBackward0>) 38600 tensor(5.0224, device='cuda:0', grad_fn=<NllLossBackward0>) 38700 tensor(4.6032, device='cuda:0', grad_fn=<NllLossBackward0>) 38800 tensor(5.1368, device='cuda:0', grad_fn=<NllLossBackward0>) 38900 tensor(4.9394, device='cuda:0', grad_fn=<NllLossBackward0>) 39000 tensor(4.7891, device='cuda:0', grad_fn=<NllLossBackward0>) 39100 tensor(4.9718, device='cuda:0', grad_fn=<NllLossBackward0>) 39200 tensor(4.9599, device='cuda:0', grad_fn=<NllLossBackward0>) 39300 tensor(4.8518, device='cuda:0', grad_fn=<NllLossBackward0>) 39400 tensor(4.7832, device='cuda:0', grad_fn=<NllLossBackward0>) 39500 tensor(4.9827, device='cuda:0', grad_fn=<NllLossBackward0>) 39600 tensor(5.0733, device='cuda:0', grad_fn=<NllLossBackward0>) 39700 tensor(4.8859, device='cuda:0', grad_fn=<NllLossBackward0>) 39800 tensor(4.9722, device='cuda:0', grad_fn=<NllLossBackward0>) 39900 tensor(5.0568, device='cuda:0', grad_fn=<NllLossBackward0>) 40000 tensor(4.8251, device='cuda:0', grad_fn=<NllLossBackward0>) 40100 tensor(4.8720, device='cuda:0', grad_fn=<NllLossBackward0>) 40200 tensor(5.3066, device='cuda:0', grad_fn=<NllLossBackward0>) 40300 tensor(4.9435, device='cuda:0', grad_fn=<NllLossBackward0>) 40400 tensor(4.9634, device='cuda:0', grad_fn=<NllLossBackward0>) 40500 tensor(4.8406, device='cuda:0', grad_fn=<NllLossBackward0>) 40600 tensor(4.8050, device='cuda:0', grad_fn=<NllLossBackward0>) 40700 tensor(4.6578, device='cuda:0', grad_fn=<NllLossBackward0>) 40800 tensor(4.8490, device='cuda:0', grad_fn=<NllLossBackward0>) 40900 tensor(5.1542, device='cuda:0', grad_fn=<NllLossBackward0>) 41000 tensor(4.8509, device='cuda:0', grad_fn=<NllLossBackward0>) 41100 tensor(4.8082, device='cuda:0', grad_fn=<NllLossBackward0>) 41200 tensor(4.8444, device='cuda:0', grad_fn=<NllLossBackward0>) 41300 tensor(5.1602, device='cuda:0', grad_fn=<NllLossBackward0>) 41400 tensor(4.7235, device='cuda:0', grad_fn=<NllLossBackward0>) 41500 tensor(5.0334, device='cuda:0', grad_fn=<NllLossBackward0>) 41600 tensor(5.0500, device='cuda:0', grad_fn=<NllLossBackward0>) 41700 tensor(5.0378, device='cuda:0', grad_fn=<NllLossBackward0>) 41800 tensor(4.7989, device='cuda:0', grad_fn=<NllLossBackward0>) 41900 tensor(4.9342, device='cuda:0', grad_fn=<NllLossBackward0>) 42000 tensor(4.9981, device='cuda:0', grad_fn=<NllLossBackward0>) 42100 tensor(4.6723, device='cuda:0', grad_fn=<NllLossBackward0>) 42200 tensor(4.9382, device='cuda:0', grad_fn=<NllLossBackward0>) 42300 tensor(4.9237, device='cuda:0', grad_fn=<NllLossBackward0>) 42400 tensor(4.9302, device='cuda:0', grad_fn=<NllLossBackward0>) 42500 tensor(4.8494, device='cuda:0', grad_fn=<NllLossBackward0>) 42600 tensor(4.9942, device='cuda:0', grad_fn=<NllLossBackward0>) 42700 tensor(4.9581, device='cuda:0', grad_fn=<NllLossBackward0>) 42800 tensor(4.8044, device='cuda:0', grad_fn=<NllLossBackward0>) 42900 tensor(5.0890, device='cuda:0', grad_fn=<NllLossBackward0>) 43000 tensor(4.9422, device='cuda:0', grad_fn=<NllLossBackward0>) 43100 tensor(5.0014, device='cuda:0', grad_fn=<NllLossBackward0>) 43200 tensor(4.9001, device='cuda:0', grad_fn=<NllLossBackward0>) 43300 tensor(4.9133, device='cuda:0', grad_fn=<NllLossBackward0>) 43400 tensor(4.8836, device='cuda:0', grad_fn=<NllLossBackward0>) 43500 tensor(4.8232, device='cuda:0', grad_fn=<NllLossBackward0>) 43600 tensor(4.8052, device='cuda:0', grad_fn=<NllLossBackward0>) 43700 tensor(5.0304, device='cuda:0', grad_fn=<NllLossBackward0>) 43800 tensor(5.0834, device='cuda:0', grad_fn=<NllLossBackward0>) 43900 tensor(4.8242, device='cuda:0', grad_fn=<NllLossBackward0>) 44000 tensor(4.8126, device='cuda:0', grad_fn=<NllLossBackward0>) 44100 tensor(4.7836, device='cuda:0', grad_fn=<NllLossBackward0>) 44200 tensor(5.0763, device='cuda:0', grad_fn=<NllLossBackward0>) 44300 tensor(5.0682, device='cuda:0', grad_fn=<NllLossBackward0>) 44400 tensor(4.8869, device='cuda:0', grad_fn=<NllLossBackward0>) 44500 tensor(4.8527, device='cuda:0', grad_fn=<NllLossBackward0>) 44600 tensor(4.8439, device='cuda:0', grad_fn=<NllLossBackward0>) 44700 tensor(4.9127, device='cuda:0', grad_fn=<NllLossBackward0>) 44800 tensor(4.9628, device='cuda:0', grad_fn=<NllLossBackward0>) 44900 tensor(5.0566, device='cuda:0', grad_fn=<NllLossBackward0>) 45000 tensor(5.0596, device='cuda:0', grad_fn=<NllLossBackward0>) 45100 tensor(5.1187, device='cuda:0', grad_fn=<NllLossBackward0>) 45200 tensor(5.0824, device='cuda:0', grad_fn=<NllLossBackward0>) 45300 tensor(4.8433, device='cuda:0', grad_fn=<NllLossBackward0>) 45400 tensor(4.7299, device='cuda:0', grad_fn=<NllLossBackward0>) 45500 tensor(5.1722, device='cuda:0', grad_fn=<NllLossBackward0>) 45600 tensor(4.7867, device='cuda:0', grad_fn=<NllLossBackward0>) 45700 tensor(4.9631, device='cuda:0', grad_fn=<NllLossBackward0>) 45800 tensor(4.6216, device='cuda:0', grad_fn=<NllLossBackward0>) 45900 tensor(4.9601, device='cuda:0', grad_fn=<NllLossBackward0>) 46000 tensor(4.9055, device='cuda:0', grad_fn=<NllLossBackward0>) 46100 tensor(5.0517, device='cuda:0', grad_fn=<NllLossBackward0>) 46200 tensor(5.0099, device='cuda:0', grad_fn=<NllLossBackward0>) 46300 tensor(4.8178, device='cuda:0', grad_fn=<NllLossBackward0>) 46400 tensor(4.9317, device='cuda:0', grad_fn=<NllLossBackward0>) 46500 tensor(4.8770, device='cuda:0', grad_fn=<NllLossBackward0>) 46600 tensor(4.9668, device='cuda:0', grad_fn=<NllLossBackward0>) 46700 tensor(5.1287, device='cuda:0', grad_fn=<NllLossBackward0>) 46800 tensor(4.9050, device='cuda:0', grad_fn=<NllLossBackward0>) 46900 tensor(4.9622, device='cuda:0', grad_fn=<NllLossBackward0>) 47000 tensor(4.6818, device='cuda:0', grad_fn=<NllLossBackward0>) 47100 tensor(4.8780, device='cuda:0', grad_fn=<NllLossBackward0>) 47200 tensor(4.9493, device='cuda:0', grad_fn=<NllLossBackward0>) 47300 tensor(4.7958, device='cuda:0', grad_fn=<NllLossBackward0>) 47400 tensor(4.5415, device='cuda:0', grad_fn=<NllLossBackward0>) 47500 tensor(5.0651, device='cuda:0', grad_fn=<NllLossBackward0>) 47600 tensor(4.9692, device='cuda:0', grad_fn=<NllLossBackward0>) 47700 tensor(4.8536, device='cuda:0', grad_fn=<NllLossBackward0>) 47800 tensor(4.7306, device='cuda:0', grad_fn=<NllLossBackward0>) 47900 tensor(5.1795, device='cuda:0', grad_fn=<NllLossBackward0>) 48000 tensor(4.9196, device='cuda:0', grad_fn=<NllLossBackward0>) 48100 tensor(5.1446, device='cuda:0', grad_fn=<NllLossBackward0>) 48200 tensor(4.9810, device='cuda:0', grad_fn=<NllLossBackward0>) 48300 tensor(4.9688, device='cuda:0', grad_fn=<NllLossBackward0>) 48400 tensor(5.0246, device='cuda:0', grad_fn=<NllLossBackward0>) 48500 tensor(4.7523, device='cuda:0', grad_fn=<NllLossBackward0>) 48600 tensor(4.7716, device='cuda:0', grad_fn=<NllLossBackward0>) 48700 tensor(4.8938, device='cuda:0', grad_fn=<NllLossBackward0>) 48800 tensor(4.9324, device='cuda:0', grad_fn=<NllLossBackward0>) 48900 tensor(4.9811, device='cuda:0', grad_fn=<NllLossBackward0>) 49000 tensor(4.8818, device='cuda:0', grad_fn=<NllLossBackward0>) 49100 tensor(4.9871, device='cuda:0', grad_fn=<NllLossBackward0>) 49200 tensor(4.8498, device='cuda:0', grad_fn=<NllLossBackward0>) 49300 tensor(4.8027, device='cuda:0', grad_fn=<NllLossBackward0>) 49400 tensor(5.0199, device='cuda:0', grad_fn=<NllLossBackward0>) 49500 tensor(4.9790, device='cuda:0', grad_fn=<NllLossBackward0>) 49600 tensor(5.0995, device='cuda:0', grad_fn=<NllLossBackward0>) 49700 tensor(4.8989, device='cuda:0', grad_fn=<NllLossBackward0>) 49800 tensor(4.8903, device='cuda:0', grad_fn=<NllLossBackward0>) 49900 tensor(4.6744, device='cuda:0', grad_fn=<NllLossBackward0>) 50000 tensor(4.9403, device='cuda:0', grad_fn=<NllLossBackward0>) 50100 tensor(4.7815, device='cuda:0', grad_fn=<NllLossBackward0>) 50200 tensor(4.8617, device='cuda:0', grad_fn=<NllLossBackward0>) 50300 tensor(4.4559, device='cuda:0', grad_fn=<NllLossBackward0>) 50400 tensor(5.0322, device='cuda:0', grad_fn=<NllLossBackward0>) 50500 tensor(4.6867, device='cuda:0', grad_fn=<NllLossBackward0>) 50600 tensor(4.9644, device='cuda:0', grad_fn=<NllLossBackward0>) 50700 tensor(5.0631, device='cuda:0', grad_fn=<NllLossBackward0>) 50800 tensor(4.7992, device='cuda:0', grad_fn=<NllLossBackward0>) 50900 tensor(4.9346, device='cuda:0', grad_fn=<NllLossBackward0>) 51000 tensor(4.6487, device='cuda:0', grad_fn=<NllLossBackward0>) 51100 tensor(4.8758, device='cuda:0', grad_fn=<NllLossBackward0>) 51200 tensor(5.0734, device='cuda:0', grad_fn=<NllLossBackward0>) 51300 tensor(4.8078, device='cuda:0', grad_fn=<NllLossBackward0>) 51400 tensor(4.7628, device='cuda:0', grad_fn=<NllLossBackward0>) 51500 tensor(4.8508, device='cuda:0', grad_fn=<NllLossBackward0>) 51600 tensor(4.8231, device='cuda:0', grad_fn=<NllLossBackward0>) 51700 tensor(5.0122, device='cuda:0', grad_fn=<NllLossBackward0>) 51800 tensor(4.8941, device='cuda:0', grad_fn=<NllLossBackward0>) 51900 tensor(5.0284, device='cuda:0', grad_fn=<NllLossBackward0>) 52000 tensor(4.9158, device='cuda:0', grad_fn=<NllLossBackward0>) 52100 tensor(4.8752, device='cuda:0', grad_fn=<NllLossBackward0>) 52200 tensor(4.7020, device='cuda:0', grad_fn=<NllLossBackward0>) 52300 tensor(4.6001, device='cuda:0', grad_fn=<NllLossBackward0>) 52400 tensor(4.7898, device='cuda:0', grad_fn=<NllLossBackward0>) 52500 tensor(4.8255, device='cuda:0', grad_fn=<NllLossBackward0>) 52600 tensor(4.7331, device='cuda:0', grad_fn=<NllLossBackward0>) 52700 tensor(4.8546, device='cuda:0', grad_fn=<NllLossBackward0>) 52800 tensor(4.9418, device='cuda:0', grad_fn=<NllLossBackward0>) 52900 tensor(4.7536, device='cuda:0', grad_fn=<NllLossBackward0>) 53000 tensor(4.9609, device='cuda:0', grad_fn=<NllLossBackward0>) 53100 tensor(5.0644, device='cuda:0', grad_fn=<NllLossBackward0>) 53200 tensor(4.8919, device='cuda:0', grad_fn=<NllLossBackward0>) 53300 tensor(4.7840, device='cuda:0', grad_fn=<NllLossBackward0>) 53400 tensor(4.8539, device='cuda:0', grad_fn=<NllLossBackward0>) 53500 tensor(4.8023, device='cuda:0', grad_fn=<NllLossBackward0>) 53600 tensor(4.9810, device='cuda:0', grad_fn=<NllLossBackward0>) 53700 tensor(4.9946, device='cuda:0', grad_fn=<NllLossBackward0>) 53800 tensor(4.3504, device='cuda:0', grad_fn=<NllLossBackward0>) 53900 tensor(4.8656, device='cuda:0', grad_fn=<NllLossBackward0>) 54000 tensor(5.0103, device='cuda:0', grad_fn=<NllLossBackward0>) 54100 tensor(4.8503, device='cuda:0', grad_fn=<NllLossBackward0>) 54200 tensor(4.9970, device='cuda:0', grad_fn=<NllLossBackward0>) 54300 tensor(4.5719, device='cuda:0', grad_fn=<NllLossBackward0>) 54400 tensor(4.7891, device='cuda:0', grad_fn=<NllLossBackward0>) 54500 tensor(4.8968, device='cuda:0', grad_fn=<NllLossBackward0>) 54600 tensor(5.0036, device='cuda:0', grad_fn=<NllLossBackward0>) 54700 tensor(4.9487, device='cuda:0', grad_fn=<NllLossBackward0>) 54800 tensor(4.8477, device='cuda:0', grad_fn=<NllLossBackward0>) 54900 tensor(4.9253, device='cuda:0', grad_fn=<NllLossBackward0>) 55000 tensor(4.9079, device='cuda:0', grad_fn=<NllLossBackward0>) 55100 tensor(4.9499, device='cuda:0', grad_fn=<NllLossBackward0>) 55200 tensor(5.0510, device='cuda:0', grad_fn=<NllLossBackward0>) 55300 tensor(4.9320, device='cuda:0', grad_fn=<NllLossBackward0>) 55400 tensor(4.5737, device='cuda:0', grad_fn=<NllLossBackward0>) 55500 tensor(4.7703, device='cuda:0', grad_fn=<NllLossBackward0>) 55600 tensor(5.0166, device='cuda:0', grad_fn=<NllLossBackward0>) 55700 tensor(4.9049, device='cuda:0', grad_fn=<NllLossBackward0>) 55800 tensor(4.7355, device='cuda:0', grad_fn=<NllLossBackward0>) 55900 tensor(4.5776, device='cuda:0', grad_fn=<NllLossBackward0>) 56000 tensor(4.9919, device='cuda:0', grad_fn=<NllLossBackward0>) 56100 tensor(4.8629, device='cuda:0', grad_fn=<NllLossBackward0>) 56200 tensor(5.0123, device='cuda:0', grad_fn=<NllLossBackward0>) 56300 tensor(4.3110, device='cuda:0', grad_fn=<NllLossBackward0>) 56400 tensor(4.8950, device='cuda:0', grad_fn=<NllLossBackward0>) 56500 tensor(4.8415, device='cuda:0', grad_fn=<NllLossBackward0>) 56600 tensor(4.7285, device='cuda:0', grad_fn=<NllLossBackward0>) 56700 tensor(4.8401, device='cuda:0', grad_fn=<NllLossBackward0>) 56800 tensor(4.7972, device='cuda:0', grad_fn=<NllLossBackward0>) 56900 tensor(4.7398, device='cuda:0', grad_fn=<NllLossBackward0>) 57000 tensor(5.1683, device='cuda:0', grad_fn=<NllLossBackward0>) 57100 tensor(4.9399, device='cuda:0', grad_fn=<NllLossBackward0>) 57200 tensor(4.9609, device='cuda:0', grad_fn=<NllLossBackward0>) 57300 tensor(4.9818, device='cuda:0', grad_fn=<NllLossBackward0>) 57400 tensor(4.9719, device='cuda:0', grad_fn=<NllLossBackward0>) 57500 tensor(4.8724, device='cuda:0', grad_fn=<NllLossBackward0>) 57600 tensor(4.9824, device='cuda:0', grad_fn=<NllLossBackward0>) 57700 tensor(5.0357, device='cuda:0', grad_fn=<NllLossBackward0>) 57800 tensor(5.0542, device='cuda:0', grad_fn=<NllLossBackward0>) 57900 tensor(4.8753, device='cuda:0', grad_fn=<NllLossBackward0>) 58000 tensor(4.7773, device='cuda:0', grad_fn=<NllLossBackward0>) 58100 tensor(4.7864, device='cuda:0', grad_fn=<NllLossBackward0>) 58200 tensor(4.8033, device='cuda:0', grad_fn=<NllLossBackward0>) 58300 tensor(4.9997, device='cuda:0', grad_fn=<NllLossBackward0>) 58400 tensor(4.9701, device='cuda:0', grad_fn=<NllLossBackward0>) 58500 tensor(4.8920, device='cuda:0', grad_fn=<NllLossBackward0>) 58600 tensor(4.9408, device='cuda:0', grad_fn=<NllLossBackward0>) 58700 tensor(5.1013, device='cuda:0', grad_fn=<NllLossBackward0>) 58800 tensor(4.8176, device='cuda:0', grad_fn=<NllLossBackward0>) 58900 tensor(4.7466, device='cuda:0', grad_fn=<NllLossBackward0>) 59000 tensor(4.9146, device='cuda:0', grad_fn=<NllLossBackward0>) 59100 tensor(4.8151, device='cuda:0', grad_fn=<NllLossBackward0>) 59200 tensor(4.9928, device='cuda:0', grad_fn=<NllLossBackward0>) 59300 tensor(5.0274, device='cuda:0', grad_fn=<NllLossBackward0>) 59400 tensor(4.7727, device='cuda:0', grad_fn=<NllLossBackward0>) 59500 tensor(5.0648, device='cuda:0', grad_fn=<NllLossBackward0>) 59600 tensor(4.9982, device='cuda:0', grad_fn=<NllLossBackward0>) 59700 tensor(4.8934, device='cuda:0', grad_fn=<NllLossBackward0>) 59800 tensor(4.8285, device='cuda:0', grad_fn=<NllLossBackward0>) 59900 tensor(4.8039, device='cuda:0', grad_fn=<NllLossBackward0>) 60000 tensor(4.9090, device='cuda:0', grad_fn=<NllLossBackward0>) 60100 tensor(4.6927, device='cuda:0', grad_fn=<NllLossBackward0>) 60200 tensor(4.8922, device='cuda:0', grad_fn=<NllLossBackward0>) 60300 tensor(4.8804, device='cuda:0', grad_fn=<NllLossBackward0>) 60400 tensor(4.9676, device='cuda:0', grad_fn=<NllLossBackward0>) 60500 tensor(4.7234, device='cuda:0', grad_fn=<NllLossBackward0>) 60600 tensor(4.9174, device='cuda:0', grad_fn=<NllLossBackward0>) 60700 tensor(4.9062, device='cuda:0', grad_fn=<NllLossBackward0>) 60800 tensor(5.0811, device='cuda:0', grad_fn=<NllLossBackward0>) 60900 tensor(5.1713, device='cuda:0', grad_fn=<NllLossBackward0>) 61000 tensor(4.9471, device='cuda:0', grad_fn=<NllLossBackward0>) 61100 tensor(4.8106, device='cuda:0', grad_fn=<NllLossBackward0>) 61200 tensor(4.8666, device='cuda:0', grad_fn=<NllLossBackward0>) 61300 tensor(4.8624, device='cuda:0', grad_fn=<NllLossBackward0>) 61400 tensor(4.5771, device='cuda:0', grad_fn=<NllLossBackward0>) 61500 tensor(4.8186, device='cuda:0', grad_fn=<NllLossBackward0>) 61600 tensor(4.7787, device='cuda:0', grad_fn=<NllLossBackward0>) 61700 tensor(4.9245, device='cuda:0', grad_fn=<NllLossBackward0>) 61800 tensor(5.0268, device='cuda:0', grad_fn=<NllLossBackward0>) 61900 tensor(5.2582, device='cuda:0', grad_fn=<NllLossBackward0>) 62000 tensor(4.8309, device='cuda:0', grad_fn=<NllLossBackward0>) 62100 tensor(4.9982, device='cuda:0', grad_fn=<NllLossBackward0>) 62200 tensor(4.8859, device='cuda:0', grad_fn=<NllLossBackward0>) 62300 tensor(4.5051, device='cuda:0', grad_fn=<NllLossBackward0>) 62400 tensor(4.6767, device='cuda:0', grad_fn=<NllLossBackward0>) 62500 tensor(4.7197, device='cuda:0', grad_fn=<NllLossBackward0>) 62600 tensor(4.6625, device='cuda:0', grad_fn=<NllLossBackward0>) 62700 tensor(4.6548, device='cuda:0', grad_fn=<NllLossBackward0>) 62800 tensor(4.7307, device='cuda:0', grad_fn=<NllLossBackward0>) 62900 tensor(4.9550, device='cuda:0', grad_fn=<NllLossBackward0>) 63000 tensor(4.5528, device='cuda:0', grad_fn=<NllLossBackward0>) 63100 tensor(4.8676, device='cuda:0', grad_fn=<NllLossBackward0>) 63200 tensor(4.9302, device='cuda:0', grad_fn=<NllLossBackward0>) 63300 tensor(4.8878, device='cuda:0', grad_fn=<NllLossBackward0>) 63400 tensor(4.9172, device='cuda:0', grad_fn=<NllLossBackward0>) 63500 tensor(4.7881, device='cuda:0', grad_fn=<NllLossBackward0>) 63600 tensor(4.8712, device='cuda:0', grad_fn=<NllLossBackward0>) 63700 tensor(4.9398, device='cuda:0', grad_fn=<NllLossBackward0>) 63800 tensor(4.9999, device='cuda:0', grad_fn=<NllLossBackward0>) 63900 tensor(4.8581, device='cuda:0', grad_fn=<NllLossBackward0>) 64000 tensor(4.6726, device='cuda:0', grad_fn=<NllLossBackward0>) 64100 tensor(5.0308, device='cuda:0', grad_fn=<NllLossBackward0>) 64200 tensor(4.7130, device='cuda:0', grad_fn=<NllLossBackward0>) 64300 tensor(4.9586, device='cuda:0', grad_fn=<NllLossBackward0>) 64400 tensor(4.9456, device='cuda:0', grad_fn=<NllLossBackward0>) 64500 tensor(4.8030, device='cuda:0', grad_fn=<NllLossBackward0>) 64600 tensor(4.9885, device='cuda:0', grad_fn=<NllLossBackward0>) 64700 tensor(4.9439, device='cuda:0', grad_fn=<NllLossBackward0>) 64800 tensor(4.6348, device='cuda:0', grad_fn=<NllLossBackward0>) 64900 tensor(4.8772, device='cuda:0', grad_fn=<NllLossBackward0>) 65000 tensor(4.9567, device='cuda:0', grad_fn=<NllLossBackward0>) 65100 tensor(4.9036, device='cuda:0', grad_fn=<NllLossBackward0>) 65200 tensor(4.7526, device='cuda:0', grad_fn=<NllLossBackward0>) 65300 tensor(4.9206, device='cuda:0', grad_fn=<NllLossBackward0>) 65400 tensor(4.8406, device='cuda:0', grad_fn=<NllLossBackward0>) 65500 tensor(4.5461, device='cuda:0', grad_fn=<NllLossBackward0>) 65600 tensor(4.9647, device='cuda:0', grad_fn=<NllLossBackward0>) 65700 tensor(4.9128, device='cuda:0', grad_fn=<NllLossBackward0>) 65800 tensor(4.8554, device='cuda:0', grad_fn=<NllLossBackward0>) 65900 tensor(4.8749, device='cuda:0', grad_fn=<NllLossBackward0>) 66000 tensor(5.1345, device='cuda:0', grad_fn=<NllLossBackward0>) 66100 tensor(4.6254, device='cuda:0', grad_fn=<NllLossBackward0>) 66200 tensor(4.9932, device='cuda:0', grad_fn=<NllLossBackward0>) 66300 tensor(4.5778, device='cuda:0', grad_fn=<NllLossBackward0>) 66400 tensor(4.7925, device='cuda:0', grad_fn=<NllLossBackward0>) 66500 tensor(4.9761, device='cuda:0', grad_fn=<NllLossBackward0>) 66600 tensor(4.9166, device='cuda:0', grad_fn=<NllLossBackward0>) 66700 tensor(4.8186, device='cuda:0', grad_fn=<NllLossBackward0>) 66800 tensor(4.9063, device='cuda:0', grad_fn=<NllLossBackward0>) 66900 tensor(4.9770, device='cuda:0', grad_fn=<NllLossBackward0>) 67000 tensor(4.8087, device='cuda:0', grad_fn=<NllLossBackward0>) 67100 tensor(4.7366, device='cuda:0', grad_fn=<NllLossBackward0>) 67200 tensor(5.0656, device='cuda:0', grad_fn=<NllLossBackward0>) 67300 tensor(4.9718, device='cuda:0', grad_fn=<NllLossBackward0>) 67400 tensor(4.8172, device='cuda:0', grad_fn=<NllLossBackward0>) 67500 tensor(4.9368, device='cuda:0', grad_fn=<NllLossBackward0>) 67600 tensor(4.9278, device='cuda:0', grad_fn=<NllLossBackward0>) 67700 tensor(4.8133, device='cuda:0', grad_fn=<NllLossBackward0>) 67800 tensor(4.9486, device='cuda:0', grad_fn=<NllLossBackward0>) 67900 tensor(4.8521, device='cuda:0', grad_fn=<NllLossBackward0>) 68000 tensor(4.9510, device='cuda:0', grad_fn=<NllLossBackward0>) 68100 tensor(4.8939, device='cuda:0', grad_fn=<NllLossBackward0>) 68200 tensor(4.8088, device='cuda:0', grad_fn=<NllLossBackward0>) 68300 tensor(4.9821, device='cuda:0', grad_fn=<NllLossBackward0>) 68400 tensor(5.1750, device='cuda:0', grad_fn=<NllLossBackward0>) 68500 tensor(4.6476, device='cuda:0', grad_fn=<NllLossBackward0>) 68600 tensor(4.8567, device='cuda:0', grad_fn=<NllLossBackward0>) 68700 tensor(4.8663, device='cuda:0', grad_fn=<NllLossBackward0>) 68800 tensor(5.0268, device='cuda:0', grad_fn=<NllLossBackward0>) 68900 tensor(4.8717, device='cuda:0', grad_fn=<NllLossBackward0>) 69000 tensor(4.9166, device='cuda:0', grad_fn=<NllLossBackward0>) 69100 tensor(4.9094, device='cuda:0', grad_fn=<NllLossBackward0>) 69200 tensor(4.7433, device='cuda:0', grad_fn=<NllLossBackward0>) 69300 tensor(4.5366, device='cuda:0', grad_fn=<NllLossBackward0>) 69400 tensor(5.0260, device='cuda:0', grad_fn=<NllLossBackward0>) 69500 tensor(4.7304, device='cuda:0', grad_fn=<NllLossBackward0>)
import torch
torch.cuda.is_available()
True
torch.save(model.state_dict(), path_to_model)
model = SimpleBigramNeuralLanguageModel(vocab_size, embed_size).to(device)
model.load_state_dict(torch.load(path_to_model))
model.eval()
SimpleBigramNeuralLanguageModel( (model): Sequential( (0): Embedding(10000, 100) (1): Linear(in_features=100, out_features=10000, bias=True) (2): Softmax(dim=None) ) )
save_outs('dev-0')
Creating outputs in dev-0
save_outs('test-A')
Creating outputs in test-A