Keras入门——(5)长短期记忆网络LSTM(二)
参考:
https://blog.csdn.net/zwqjoy/article/details/80493341
https://blog.csdn.net/u012735708/article/details/82769711
执行代码:
# LSTM with Variable Length Input Sequences to One Character Output import numpy from keras.models import Sequential from keras.layers import Dense from keras.layers import LSTM from keras.utils import np_utils from keras.preprocessing.sequence import pad_sequences #from theano.tensor.shared_randomstreams import RandomStreams # fix random seed for reproducibility numpy.random.seed(7) # define the raw dataset alphabet = "ABCDEFGHIJKLMNOPQRSTUVWXYZ" # create mapping of characters to integers (0-25) and the reverse char_to_int = dict((c, i) for i, c in enumerate(alphabet)) int_to_char = dict((i, c) for i, c in enumerate(alphabet)) # prepare the dataset of input to output pairs encoded as integers num_inputs = 1000 max_len = 5 dataX = [] dataY = [] for i in range(num_inputs): start = numpy.random.randint(len(alphabet)-2) end = numpy.random.randint(start, min(start+max_len,len(alphabet)-1)) sequence_in = alphabet[start:end+1] sequence_out = alphabet[end + 1] dataX.append([char_to_int[char] for char in sequence_in]) dataY.append(char_to_int[sequence_out]) print(sequence_in, '->', sequence_out) # convert list of lists to array and pad sequences if needed X = pad_sequences(dataX, maxlen=max_len, dtype='float32') # reshape X to be [samples, time steps, features] X = numpy.reshape(X, (X.shape[0], max_len, 1)) # normalize X = X / float(len(alphabet)) # one hot encode the output variable y = np_utils.to_categorical(dataY) # create and fit the model batch_size = 1 model = Sequential() model.add(LSTM(32, input_shape=(X.shape[1], 1))) model.add(Dense(y.shape[1], activation='softmax')) model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy']) model.fit(X, y, epochs=500, batch_size=batch_size, verbose=2) # summarize performance of the model scores = model.evaluate(X, y, verbose=0) print("Model Accuracy: %.2f%%" % (scores[1]*100)) # demonstrate some model predictions for i in range(20): pattern_index = numpy.random.randint(len(dataX)) pattern = dataX[pattern_index] x = pad_sequences([pattern], maxlen=max_len, dtype='float32') x = numpy.reshape(x, (1, max_len, 1)) x = x / float(len(alphabet)) prediction = model.predict(x, verbose=0) index = numpy.argmax(prediction) result = int_to_char[index] seq_in = [int_to_char[value] for value in pattern] print(seq_in, "->", result)
计算机负载:
返回信息:
PQRST -> U W -> X O -> P OPQ -> R IJKLM -> N QRSTU -> V ABCD -> E X -> Y GHIJ -> K M -> N XY -> Z QRST -> U ABC -> D JKLMN -> O OP -> Q XY -> Z D -> E T -> U B -> C QRSTU -> V HIJ -> K JKLM -> N ABCDE -> F X -> Y V -> W DE -> F DEFG -> H BCDE -> F EFGH -> I BCDE -> F FG -> H RST -> U TUV -> W STUV -> W LMN -> O P -> Q MNOP -> Q JK -> L MNOP -> Q OPQRS -> T UVWXY -> Z PQRS -> T D -> E EFGH -> I IJK -> L WX -> Y STUV -> W MNOPQ -> R P -> Q WXY -> Z VWX -> Y V -> W HI -> J KLMNO -> P UV -> W JKL -> M ABCDE -> F WXY -> Z M -> N CDEF -> G KLMNO -> P RST -> U RS -> T W -> X J -> K WX -> Y JKLMN -> O MN -> O L -> M BCDE -> F TU -> V MNOPQ -> R NOPQR -> S HIJ -> K JKLM -> N STUVW -> X QRST -> U N -> O VWXY -> Z B -> C UVWX -> Y OP -> Q K -> L C -> D X -> Y ST -> U JKLM -> N B -> C QR -> S RS -> T VWXY -> Z S -> T NOP -> Q KLMNO -> P IJ -> K EF -> G MNOP -> Q WXY -> Z HI -> J P -> Q STUVW -> X Q -> R MN -> O O -> P C -> D L -> M JKLM -> N K -> L IJKLM -> N FGHIJ -> K LM -> N OPQ -> R U -> V HIJKL -> M PQR -> S S -> T OPQR -> S J -> K DE -> F K -> L BCDE -> F EFGH -> I RSTUV -> W LMNOP -> Q QR -> S ABCDE -> F LM -> N IJKLM -> N B -> C VWX -> Y MNOPQ -> R MNOPQ -> R LM -> N S -> T VWX -> Y WXY -> Z F -> G KLMNO -> P OPQ -> R M -> N X -> Y OPQRS -> T F -> G JKLMN -> O XY -> Z OPQ -> R FG -> H OP -> Q DEFGH -> I ABCD -> E VWX -> Y U -> V UV -> W VWX -> Y LMNO -> P E -> F NOPQ -> R HIJK -> L HIJ -> K DE -> F B -> C UVW -> X STUV -> W RST -> U H -> I I -> J MN -> O CDEF -> G ABC -> D RSTU -> V B -> C JKLM -> N TUVW -> X STUVW -> X C -> D UV -> W QRS -> T ABC -> D NOP -> Q W -> X DE -> F VWXY -> Z UV -> W JK -> L E -> F MNO -> P EFGH -> I PQRS -> T FGH -> I WXY -> Z OPQRS -> T TUV -> W MN -> O O -> P LMN -> O VWX -> Y QR -> S TUV -> W STU -> V EFGH -> I E -> F HIJ -> K QRS -> T H -> I K -> L E -> F UV -> W X -> Y QR -> S QRS -> T WXY -> Z S -> T CDEFG -> H PQRST -> U RST -> U A -> B CDEF -> G X -> Y JKLM -> N VWX -> Y N -> O W -> X TUVW -> X LMNOP -> Q EFG -> H HI -> J WXY -> Z IJK -> L R -> S H -> I V -> W OPQR -> S QRSTU -> V MNOPQ -> R Q -> R VWXY -> Z ABCDE -> F HIJK -> L FGHIJ -> K BC -> D UV -> W WXY -> Z VWX -> Y L -> M FG -> H E -> F WXY -> Z KLMN -> O B -> C QRSTU -> V X -> Y ST -> U GH -> I CDE -> F IJKLM -> N JKL -> M HIJ -> K UVWXY -> Z PQ -> R AB -> C HIJ -> K EFG -> H PQRS -> T BCDEF -> G IJKL -> M DEFGH -> I VW -> X XY -> Z OPQ -> R MN -> O OP -> Q WXY -> Z STU -> V LM -> N UV -> W EF -> G LMN -> O D -> E H -> I KLMNO -> P PQRST -> U V -> W M -> N UVW -> X ABCD -> E LM -> N A -> B DEFGH -> I IJK -> L OP -> Q WXY -> Z CDEFG -> H UVW -> X RS -> T FGHIJ -> K RST -> U NO -> P X -> Y RST -> U I -> J TUV -> W B -> C UVWX -> Y HIJKL -> M MNOPQ -> R ABC -> D PQ -> R WX -> Y XY -> Z UVW -> X IJKL -> M XY -> Z DEFG -> H H -> I Q -> R CDEFG -> H C -> D ABCD -> E LMN -> O PQRST -> U VWX -> Y M -> N KLMN -> O AB -> C NOPQ -> R F -> G NO -> P KLM -> N TUVWX -> Y U -> V CDEFG -> H FGHI -> J STUVW -> X JKLM -> N ABC -> D JKLMN -> O TUVWX -> Y D -> E EFGH -> I IJ -> K UVW -> X OPQR -> S N -> O VWXY -> Z ABC -> D J -> K RS -> T LMNOP -> Q BC -> D OPQ -> R JKLM -> N WX -> Y BCD -> E RSTU -> V GHI -> J O -> P R -> S QR -> S HIJKL -> M UVWXY -> Z CDEFG -> H OP -> Q HIJK -> L A -> B RST -> U QR -> S ABCD -> E LMN -> O TUV -> W MNO -> P AB -> C M -> N OPQR -> S STU -> V TUV -> W PQRST -> U LM -> N A -> B A -> B OPQ -> R HIJK -> L TU -> V QRS -> T WX -> Y BCD -> E ST -> U X -> Y EFGHI -> J E -> F FGHIJ -> K HI -> J ABC -> D NOPQ -> R HIJK -> L B -> C U -> V GH -> I TUVWX -> Y S -> T BCDEF -> G KLM -> N Q -> R CD -> E PQ -> R GH -> I U -> V RST -> U JKLM -> N FGH -> I IJ -> K O -> P X -> Y H -> I DEF -> G QRSTU -> V ABCD -> E IJK -> L GHI -> J QR -> S NOPQR -> S EF -> G PQRST -> U RST -> U X -> Y QR -> S HIJ -> K D -> E AB -> C N -> O QR -> S BCDEF -> G QRS -> T DEF -> G TUV -> W A -> B GHIJ -> K W -> X VWXY -> Z LM -> N OPQ -> R XY -> Z KLM -> N RST -> U OP -> Q VWX -> Y OPQ -> R N -> O M -> N JKL -> M OP -> Q DEF -> G BCD -> E K -> L MN -> O IJKL -> M QR -> S IJKLM -> N U -> V FGH -> I MNOPQ -> R TUVW -> X MN -> O RSTUV -> W VWX -> Y Q -> R DEFGH -> I NO -> P T -> U V -> W ST -> U DEFG -> H RS -> T NOPQ -> R GHIJK -> L QRSTU -> V LMNO -> P IJK -> L PQRST -> U IJK -> L DE -> F CD -> E JKLM -> N WX -> Y UV -> W W -> X KLM -> N PQ -> R W -> X WXY -> Z EFGHI -> J E -> F NOP -> Q VW -> X EFGHI -> J NO -> P HIJKL -> M UVWXY -> Z OPQ -> R P -> Q H -> I O -> P GHIJK -> L S -> T E -> F KLMN -> O TUVW -> X E -> F CDE -> F I -> J CDEF -> G F -> G ABCD -> E H -> I LMNOP -> Q V -> W W -> X BCD -> E TU -> V VWXY -> Z UVWX -> Y JKL -> M VW -> X CDEF -> G DEF -> G ABCDE -> F MNO -> P EFGH -> I JKLM -> N QR -> S ABCDE -> F OPQR -> S DEF -> G Q -> R TU -> V CDEFG -> H KLMN -> O VW -> X HIJKL -> M DE -> F OP -> Q I -> J GHIJK -> L HIJKL -> M I -> J AB -> C DE -> F I -> J O -> P HIJK -> L QR -> S MN -> O I -> J LM -> N VWXY -> Z JKLMN -> O BC -> D MN -> O GHIJ -> K KL -> M TU -> V QRST -> U ABCDE -> F GH -> I Q -> R NO -> P RST -> U BCDE -> F T -> U TUV -> W FGHIJ -> K T -> U BCD -> E NO -> P JK -> L BCD -> E G -> H A -> B GHIJK -> L QRSTU -> V AB -> C VW -> X HIJKL -> M FGHIJ -> K PQ -> R UV -> W F -> G A -> B Q -> R MNOP -> Q UVWXY -> Z GHIJK -> L GHIJK -> L BCDE -> F QRS -> T PQRS -> T PQ -> R HI -> J PQRST -> U OPQR -> S QRST -> U IJKLM -> N Q -> R F -> G QRST -> U ST -> U MN -> O CD -> E EFG -> H FGH -> I R -> S C -> D RSTUV -> W KL -> M HIJK -> L CD -> E FGHI -> J VW -> X P -> Q C -> D DE -> F DE -> F I -> J LMNOP -> Q KLMNO -> P QRS -> T F -> G UVWXY -> Z QRS -> T BCD -> E FG -> H ABCDE -> F U -> V M -> N KLMN -> O RST -> U UVWX -> Y X -> Y XY -> Z I -> J KLMN -> O X -> Y W -> X RSTUV -> W VW -> X XY -> Z T -> U CDE -> F FGHI -> J PQ -> R OPQRS -> T D -> E E -> F EFGH -> I GHIJK -> L L -> M KLMN -> O STU -> V EF -> G UV -> W K -> L QRS -> T QRSTU -> V DEF -> G UV -> W D -> E BC -> D OPQRS -> T EFGH -> I QRST -> U EF -> G RST -> U JKL -> M STU -> V UVWX -> Y EFGHI -> J JKLMN -> O P -> Q BCD -> E TU -> V O -> P RST -> U D -> E VWXY -> Z R -> S P -> Q CDE -> F X -> Y UVWXY -> Z DEFGH -> I NOP -> Q ABCD -> E B -> C BC -> D VW -> X E -> F TUVW -> X JKL -> M XY -> Z LM -> N PQRS -> T O -> P KLMN -> O STUV -> W K -> L UVWX -> Y U -> V HIJ -> K W -> X VWXY -> Z WX -> Y HIJ -> K O -> P QR -> S VWXY -> Z CD -> E KL -> M DEFGH -> I LMN -> O QRS -> T JKLMN -> O QR -> S CD -> E QRST -> U BCDEF -> G CDE -> F LMN -> O DEF -> G BCD -> E UV -> W STUVW -> X RS -> T ABCD -> E BCDEF -> G Q -> R UVWXY -> Z VW -> X VW -> X WXY -> Z NOPQR -> S V -> W LM -> N B -> C JKL -> M DE -> F K -> L ABC -> D E -> F STU -> V TU -> V G -> H AB -> C J -> K FGH -> I MNOP -> Q VW -> X CD -> E TUVWX -> Y F -> G VWX -> Y LMNO -> P GHIJ -> K TUVWX -> Y JKL -> M LM -> N EFGHI -> J MNO -> P H -> I M -> N S -> T STU -> V QRST -> U PQR -> S RSTUV -> W ST -> U RSTUV -> W JKLM -> N T -> U CDE -> F HIJ -> K NOPQ -> R OPQ -> R EF -> G AB -> C CD -> E RST -> U STU -> V L -> M WXY -> Z STUVW -> X QRST -> U W -> X S -> T M -> N GH -> I QRST -> U FGH -> I PQRS -> T GH -> I DE -> F DE -> F GHIJK -> L Q -> R WX -> Y WX -> Y KLM -> N DE -> F EF -> G UVW -> X IJK -> L NO -> P QR -> S TU -> V RST -> U VW -> X A -> B DE -> F WXY -> Z CD -> E IJK -> L STUV -> W LMNOP -> Q X -> Y FGH -> I F -> G IJK -> L EFG -> H DEFG -> H NOP -> Q FG -> H RSTU -> V E -> F WXY -> Z GH -> I CD -> E IJ -> K TUVWX -> Y EFGH -> I DEFGH -> I BCDE -> F STUV -> W HI -> J GH -> I STUVW -> X ABC -> D S -> T LMNOP -> Q UVWX -> Y PQ -> R CDEF -> G E -> F TU -> V TUVWX -> Y GHIJ -> K JK -> L IJK -> L G -> H EFG -> H TU -> V FGHI -> J W -> X T -> U CDE -> F XY -> Z XY -> Z CDE -> F N -> O QRST -> U FGHIJ -> K PQ -> R I -> J GH -> I F -> G VWX -> Y ABC -> D GH -> I KLMN -> O X -> Y Q -> R NOPQR -> S HIJ -> K IJ -> K C -> D FG -> H JKLMN -> O TU -> V NOPQR -> S O -> P TU -> V MNOPQ -> R PQ -> R S -> T VWXY -> Z VWXY -> Z CD -> E BCDEF -> G OPQ -> R LMNO -> P HIJKL -> M STU -> V GHI -> J UVWX -> Y NOPQ -> R HIJK -> L NOP -> Q Q -> R HIJ -> K W -> X QR -> S UVWX -> Y H -> I ABC -> D RSTUV -> W VW -> X OP -> Q RSTUV -> W ABC -> D ABC -> D GHIJ -> K WXY -> Z BCDE -> F N -> O JK -> L X -> Y TUV -> W L -> M F -> G MN -> O JKLMN -> O G -> H BCDEF -> G LMN -> O N -> O V -> W BCDEF -> G KLM -> N ST -> U TUV -> W MN -> O JKLM -> N LM -> N U -> V FGH -> I TUV -> W C -> D HIJK -> L UVWX -> Y W -> X QR -> S PQR -> S STUVW -> X RSTU -> V TU -> V RSTU -> V JKL -> M JKL -> M RSTUV -> W GHI -> J V -> W CD -> E QRSTU -> V M -> N BCDE -> F WX -> Y K -> L VW -> X GHI -> J CD -> E XY -> Z HI -> J C -> D IJK -> L DEFG -> H UV -> W LM -> N X -> Y UV -> W I -> J NO -> P ABCD -> E K -> L IJK -> L JKL -> M EFGHI -> J JK -> L TU -> V IJ -> K MNOPQ -> R C -> D IJKLM -> N VW -> X CDE -> F E -> F NOP -> Q OPQRS -> T FGHI -> J STUV -> W IJKLM -> N STUV -> W TUVWX -> Y RSTU -> V WARNING:tensorflow:From D:\ProgramData\Anaconda2\lib\site-packages\tensorflow\python\framework\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version. Instructions for updating: Colocations handled automatically by placer. WARNING:tensorflow:From D:\ProgramData\Anaconda2\lib\site-packages\tensorflow\python\ops\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version. Instructions for updating: Use tf.cast instead. Epoch 1/500 - 9s - loss: 3.0781 - acc: 0.0650 Epoch 2/500 - 3s - loss: 2.7666 - acc: 0.1290 Epoch 3/500 - 3s - loss: 2.4376 - acc: 0.1960 Epoch 4/500 - 3s - loss: 2.2114 - acc: 0.2600 Epoch 5/500 - 3s - loss: 2.0611 - acc: 0.3090 Epoch 6/500 - 3s - loss: 1.9384 - acc: 0.3260 Epoch 7/500 - 3s - loss: 1.8378 - acc: 0.3500 Epoch 8/500 - 3s - loss: 1.7537 - acc: 0.3770 Epoch 9/500 - 3s - loss: 1.6735 - acc: 0.4190 Epoch 10/500 - 3s - loss: 1.5981 - acc: 0.4460 Epoch 11/500 - 3s - loss: 1.5255 - acc: 0.4730 Epoch 12/500 - 4s - loss: 1.4643 - acc: 0.4910 Epoch 13/500 - 4s - loss: 1.4142 - acc: 0.5070 Epoch 14/500 - 4s - loss: 1.3573 - acc: 0.5460 Epoch 15/500 - 4s - loss: 1.3121 - acc: 0.5620 Epoch 16/500 - 3s - loss: 1.2643 - acc: 0.5930 Epoch 17/500 - 3s - loss: 1.2226 - acc: 0.5820 Epoch 18/500 - 3s - loss: 1.1888 - acc: 0.6170 Epoch 19/500 - 3s - loss: 1.1384 - acc: 0.6270 Epoch 20/500 - 3s - loss: 1.1160 - acc: 0.6350 Epoch 21/500 - 3s - loss: 1.0754 - acc: 0.6610 Epoch 22/500 - 3s - loss: 1.0556 - acc: 0.6750 Epoch 23/500 - 3s - loss: 1.0104 - acc: 0.6970 Epoch 24/500 - 3s - loss: 0.9917 - acc: 0.6960 Epoch 25/500 - 3s - loss: 0.9545 - acc: 0.7020 Epoch 26/500 - 3s - loss: 0.9278 - acc: 0.7200 Epoch 27/500 - 3s - loss: 0.9093 - acc: 0.7140 Epoch 28/500 - 3s - loss: 0.8757 - acc: 0.7470 Epoch 29/500 - 3s - loss: 0.8721 - acc: 0.7370 Epoch 30/500 - 3s - loss: 0.8410 - acc: 0.7390 Epoch 31/500 - 3s - loss: 0.8259 - acc: 0.7570 Epoch 32/500 - 3s - loss: 0.7960 - acc: 0.7680 Epoch 33/500 - 3s - loss: 0.7942 - acc: 0.7600 Epoch 34/500 - 3s - loss: 0.7730 - acc: 0.7720 Epoch 35/500 - 4s - loss: 0.7593 - acc: 0.7810 Epoch 36/500 - 4s - loss: 0.7565 - acc: 0.7600 Epoch 37/500 - 4s - loss: 0.7247 - acc: 0.7900 Epoch 38/500 - 3s - loss: 0.7106 - acc: 0.7820 Epoch 39/500 - 3s - loss: 0.7050 - acc: 0.8000 Epoch 40/500 - 3s - loss: 0.6787 - acc: 0.8020 Epoch 41/500 - 3s - loss: 0.6763 - acc: 0.7950 Epoch 42/500 - 3s - loss: 0.6649 - acc: 0.8020 Epoch 43/500 - 3s - loss: 0.6583 - acc: 0.8020 Epoch 44/500 - 3s - loss: 0.6298 - acc: 0.8070 Epoch 45/500 - 3s - loss: 0.6531 - acc: 0.8010 Epoch 46/500 - 3s - loss: 0.6218 - acc: 0.8120 Epoch 47/500 - 3s - loss: 0.6079 - acc: 0.8110 Epoch 48/500 - 3s - loss: 0.6102 - acc: 0.8020 Epoch 49/500 - 3s - loss: 0.5851 - acc: 0.8220 Epoch 50/500 - 3s - loss: 0.5927 - acc: 0.8200 Epoch 51/500 - 3s - loss: 0.5945 - acc: 0.8210 Epoch 52/500 - 3s - loss: 0.5570 - acc: 0.8350 Epoch 53/500 - 3s - loss: 0.5888 - acc: 0.8030 Epoch 54/500 - 3s - loss: 0.5421 - acc: 0.8410 Epoch 55/500 - 3s - loss: 0.5411 - acc: 0.8320 Epoch 56/500 - 3s - loss: 0.5603 - acc: 0.8230 Epoch 57/500 - 3s - loss: 0.5483 - acc: 0.8260 Epoch 58/500 - 4s - loss: 0.5349 - acc: 0.8280 Epoch 59/500 - 3s - loss: 0.5219 - acc: 0.8310 Epoch 60/500 - 3s - loss: 0.5172 - acc: 0.8340 Epoch 61/500 - 3s - loss: 0.5028 - acc: 0.8340 Epoch 62/500 - 3s - loss: 0.5017 - acc: 0.8380 Epoch 63/500 - 3s - loss: 0.5034 - acc: 0.8390 Epoch 64/500 - 3s - loss: 0.4910 - acc: 0.8450 Epoch 65/500 - 3s - loss: 0.5269 - acc: 0.8210 Epoch 66/500 - 3s - loss: 0.4642 - acc: 0.8530 Epoch 67/500 - 3s - loss: 0.5172 - acc: 0.8370 Epoch 68/500 - 3s - loss: 0.4531 - acc: 0.8660 Epoch 69/500 - 3s - loss: 0.5000 - acc: 0.8370 Epoch 70/500 - 3s - loss: 0.5373 - acc: 0.8260 Epoch 71/500 - 3s - loss: 0.4472 - acc: 0.8550 Epoch 72/500 - 3s - loss: 0.5311 - acc: 0.8280 Epoch 73/500 - 3s - loss: 0.4371 - acc: 0.8690 Epoch 74/500 - 3s - loss: 0.4926 - acc: 0.8350 Epoch 75/500 - 3s - loss: 0.4285 - acc: 0.8670 Epoch 76/500 - 3s - loss: 0.4768 - acc: 0.8520 Epoch 77/500 - 3s - loss: 0.4311 - acc: 0.8620 Epoch 78/500 - 3s - loss: 0.4322 - acc: 0.8650 Epoch 79/500 - 3s - loss: 0.5357 - acc: 0.8380 Epoch 80/500 - 3s - loss: 0.4184 - acc: 0.8670 Epoch 81/500 - 3s - loss: 0.4085 - acc: 0.8730 Epoch 82/500 - 4s - loss: 0.4627 - acc: 0.8370 Epoch 83/500 - 3s - loss: 0.4059 - acc: 0.8700 Epoch 84/500 - 3s - loss: 0.4182 - acc: 0.8670 Epoch 85/500 - 3s - loss: 0.4629 - acc: 0.8500 Epoch 86/500 - 3s - loss: 0.3936 - acc: 0.8820 Epoch 87/500 - 3s - loss: 0.4025 - acc: 0.8710 Epoch 88/500 - 3s - loss: 0.5675 - acc: 0.8220 Epoch 89/500 - 3s - loss: 0.3839 - acc: 0.8870 Epoch 90/500 - 3s - loss: 0.4039 - acc: 0.8680 Epoch 91/500 - 4s - loss: 0.3824 - acc: 0.8800 Epoch 92/500 - 4s - loss: 0.4262 - acc: 0.8690 Epoch 93/500 - 3s - loss: 0.3921 - acc: 0.8740 Epoch 94/500 - 3s - loss: 0.4030 - acc: 0.8640 Epoch 95/500 - 3s - loss: 0.4341 - acc: 0.8590 Epoch 96/500 - 3s - loss: 0.3719 - acc: 0.8750 Epoch 97/500 - 3s - loss: 0.3691 - acc: 0.8930 Epoch 98/500 - 4s - loss: 0.4230 - acc: 0.8660 Epoch 99/500 - 3s - loss: 0.3698 - acc: 0.8880 Epoch 100/500 - 3s - loss: 0.3844 - acc: 0.8660 Epoch 101/500 - 4s - loss: 0.3587 - acc: 0.8930 Epoch 102/500 - 3s - loss: 0.4414 - acc: 0.8510 Epoch 103/500 - 3s - loss: 0.3495 - acc: 0.8960 Epoch 104/500 - 3s - loss: 0.3540 - acc: 0.8960 Epoch 105/500 - 3s - loss: 0.4096 - acc: 0.8680 Epoch 106/500 - 3s - loss: 0.3348 - acc: 0.9050 Epoch 107/500 - 3s - loss: 0.3886 - acc: 0.8770 Epoch 108/500 - 3s - loss: 0.3495 - acc: 0.8920 Epoch 109/500 - 3s - loss: 0.3502 - acc: 0.8820 Epoch 110/500 - 3s - loss: 0.3518 - acc: 0.8830 Epoch 111/500 - 3s - loss: 0.3422 - acc: 0.9020 Epoch 112/500 - 3s - loss: 0.3941 - acc: 0.8750 Epoch 113/500 - 3s - loss: 0.3284 - acc: 0.8990 Epoch 114/500 - 3s - loss: 0.4013 - acc: 0.8780 Epoch 115/500 - 3s - loss: 0.3271 - acc: 0.8930 Epoch 116/500 - 3s - loss: 0.3283 - acc: 0.9060 Epoch 117/500 - 3s - loss: 0.3310 - acc: 0.8980 Epoch 118/500 - 3s - loss: 0.3545 - acc: 0.8930 Epoch 119/500 - 3s - loss: 0.3286 - acc: 0.8990 Epoch 120/500 - 3s - loss: 0.4055 - acc: 0.8860 Epoch 121/500 - 3s - loss: 0.3115 - acc: 0.9070 Epoch 122/500 - 3s - loss: 0.3295 - acc: 0.8980 Epoch 123/500 - 3s - loss: 0.3251 - acc: 0.8960 Epoch 124/500 - 3s - loss: 0.3277 - acc: 0.8970 Epoch 125/500 - 3s - loss: 0.3203 - acc: 0.9030 Epoch 126/500 - 3s - loss: 0.4080 - acc: 0.8730 Epoch 127/500 - 3s - loss: 0.3029 - acc: 0.9210 Epoch 128/500 - 4s - loss: 0.3096 - acc: 0.9020 Epoch 129/500 - 4s - loss: 0.3590 - acc: 0.8970 Epoch 130/500 - 3s - loss: 0.3287 - acc: 0.8860 Epoch 131/500 - 3s - loss: 0.3055 - acc: 0.9150 Epoch 132/500 - 4s - loss: 0.2978 - acc: 0.9130 Epoch 133/500 - 3s - loss: 0.4783 - acc: 0.8530 Epoch 134/500 - 3s - loss: 0.2938 - acc: 0.9220 Epoch 135/500 - 3s - loss: 0.2980 - acc: 0.9110 Epoch 136/500 - 3s - loss: 0.3000 - acc: 0.9120 Epoch 137/500 - 3s - loss: 0.3878 - acc: 0.8780 Epoch 138/500 - 3s - loss: 0.2945 - acc: 0.9140 Epoch 139/500 - 3s - loss: 0.2912 - acc: 0.9150 Epoch 140/500 - 3s - loss: 0.2923 - acc: 0.9090 Epoch 141/500 - 3s - loss: 0.2976 - acc: 0.9080 Epoch 142/500 - 4s - loss: 0.4001 - acc: 0.8780 Epoch 143/500 - 3s - loss: 0.2775 - acc: 0.9240 Epoch 144/500 - 3s - loss: 0.2815 - acc: 0.9190 Epoch 145/500 - 3s - loss: 0.3169 - acc: 0.9050 Epoch 146/500 - 3s - loss: 0.2904 - acc: 0.9090 Epoch 147/500 - 3s - loss: 0.2819 - acc: 0.9160 Epoch 148/500 - 3s - loss: 0.3496 - acc: 0.8870 Epoch 149/500 - 3s - loss: 0.2770 - acc: 0.9180 Epoch 150/500 - 3s - loss: 0.2727 - acc: 0.9180 Epoch 151/500 - 3s - loss: 0.3157 - acc: 0.8900 Epoch 152/500 - 3s - loss: 0.2723 - acc: 0.9210 Epoch 153/500 - 3s - loss: 0.3939 - acc: 0.8740 Epoch 154/500 - 3s - loss: 0.2633 - acc: 0.9280 Epoch 155/500 - 3s - loss: 0.2652 - acc: 0.9190 Epoch 156/500 - 3s - loss: 0.3344 - acc: 0.8940 Epoch 157/500 - 3s - loss: 0.2646 - acc: 0.9240 Epoch 158/500 - 3s - loss: 0.2803 - acc: 0.9170 Epoch 159/500 - 3s - loss: 0.3612 - acc: 0.8940 Epoch 160/500 - 3s - loss: 0.2599 - acc: 0.9260 Epoch 161/500 - 3s - loss: 0.2636 - acc: 0.9180 Epoch 162/500 - 3s - loss: 0.2598 - acc: 0.9260 Epoch 163/500 - 3s - loss: 0.2687 - acc: 0.9190 Epoch 164/500 - 3s - loss: 0.3422 - acc: 0.8960 Epoch 165/500 - 3s - loss: 0.3056 - acc: 0.9070 Epoch 166/500 - 3s - loss: 0.2551 - acc: 0.9340 Epoch 167/500 - 3s - loss: 0.2569 - acc: 0.9200 Epoch 168/500 - 3s - loss: 0.2677 - acc: 0.9240 Epoch 169/500 - 3s - loss: 0.2677 - acc: 0.9280 Epoch 170/500 - 3s - loss: 0.2736 - acc: 0.9230 Epoch 171/500 - 3s - loss: 0.2536 - acc: 0.9240 Epoch 172/500 - 4s - loss: 0.2967 - acc: 0.9160 Epoch 173/500 - 4s - loss: 0.2471 - acc: 0.9270 Epoch 174/500 - 3s - loss: 0.2499 - acc: 0.9210 Epoch 175/500 - 4s - loss: 0.3666 - acc: 0.8970 Epoch 176/500 - 4s - loss: 0.2454 - acc: 0.9270 Epoch 177/500 - 3s - loss: 0.2431 - acc: 0.9330 Epoch 178/500 - 3s - loss: 0.2476 - acc: 0.9290 Epoch 179/500 - 3s - loss: 0.3269 - acc: 0.9030 Epoch 180/500 - 3s - loss: 0.2401 - acc: 0.9250 Epoch 181/500 - 3s - loss: 0.2412 - acc: 0.9290 Epoch 182/500 - 3s - loss: 0.2438 - acc: 0.9300 Epoch 183/500 - 3s - loss: 0.2439 - acc: 0.9300 Epoch 184/500 - 3s - loss: 0.2747 - acc: 0.9130 Epoch 185/500 - 3s - loss: 0.2443 - acc: 0.9250 Epoch 186/500 - 3s - loss: 0.3478 - acc: 0.9030 Epoch 187/500 - 3s - loss: 0.2428 - acc: 0.9300 Epoch 188/500 - 3s - loss: 0.2305 - acc: 0.9330 Epoch 189/500 - 3s - loss: 0.2348 - acc: 0.9360 Epoch 190/500 - 3s - loss: 0.2319 - acc: 0.9320 Epoch 191/500 - 3s - loss: 0.3580 - acc: 0.9010 Epoch 192/500 - 3s - loss: 0.2276 - acc: 0.9360 Epoch 193/500 - 3s - loss: 0.2264 - acc: 0.9330 Epoch 194/500 - 3s - loss: 0.2273 - acc: 0.9310 Epoch 195/500 - 3s - loss: 0.2496 - acc: 0.9260 Epoch 196/500 - 3s - loss: 0.2272 - acc: 0.9330 Epoch 197/500 - 3s - loss: 0.2273 - acc: 0.9350 Epoch 198/500 - 3s - loss: 0.3299 - acc: 0.9110 Epoch 199/500 - 3s - loss: 0.2211 - acc: 0.9400 Epoch 200/500 - 3s - loss: 0.2237 - acc: 0.9340 Epoch 201/500 - 3s - loss: 0.2225 - acc: 0.9350 Epoch 202/500 - 3s - loss: 0.2907 - acc: 0.9160 Epoch 203/500 - 3s - loss: 0.2207 - acc: 0.9370 Epoch 204/500 - 3s - loss: 0.2196 - acc: 0.9380 Epoch 205/500 - 3s - loss: 0.2708 - acc: 0.9210 Epoch 206/500 - 3s - loss: 0.2139 - acc: 0.9460 Epoch 207/500 - 3s - loss: 0.2178 - acc: 0.9360 Epoch 208/500 - 3s - loss: 0.2618 - acc: 0.9290 Epoch 209/500 - 3s - loss: 0.2327 - acc: 0.9410 Epoch 210/500 - 3s - loss: 0.2162 - acc: 0.9320 Epoch 211/500 - 3s - loss: 0.2170 - acc: 0.9420 Epoch 212/500 - 3s - loss: 0.3573 - acc: 0.9000 Epoch 213/500 - 3s - loss: 0.2050 - acc: 0.9480 Epoch 214/500 - 3s - loss: 0.2081 - acc: 0.9410 Epoch 215/500 - 3s - loss: 0.2119 - acc: 0.9410 Epoch 216/500 - 3s - loss: 0.2143 - acc: 0.9370 Epoch 217/500 - 3s - loss: 0.2845 - acc: 0.9120 Epoch 218/500 - 3s - loss: 0.2072 - acc: 0.9390 Epoch 219/500 - 3s - loss: 0.2096 - acc: 0.9450 Epoch 220/500 - 3s - loss: 0.3284 - acc: 0.9040 Epoch 221/500 - 3s - loss: 0.2026 - acc: 0.9450 Epoch 222/500 - 3s - loss: 0.2026 - acc: 0.9440 Epoch 223/500 - 3s - loss: 0.2346 - acc: 0.9290 Epoch 224/500 - 3s - loss: 0.2362 - acc: 0.9340 Epoch 225/500 - 3s - loss: 0.2029 - acc: 0.9470 Epoch 226/500 - 3s - loss: 0.2055 - acc: 0.9390 Epoch 227/500 - 3s - loss: 0.3374 - acc: 0.9030 Epoch 228/500 - 3s - loss: 0.2014 - acc: 0.9480 Epoch 229/500 - 3s - loss: 0.1983 - acc: 0.9440 Epoch 230/500 - 3s - loss: 0.2021 - acc: 0.9370 Epoch 231/500 - 3s - loss: 0.2076 - acc: 0.9380 Epoch 232/500 - 3s - loss: 0.3294 - acc: 0.9100 Epoch 233/500 - 3s - loss: 0.1952 - acc: 0.9500 Epoch 234/500 - 3s - loss: 0.1950 - acc: 0.9480 Epoch 235/500 - 3s - loss: 0.1993 - acc: 0.9410 Epoch 236/500 - 3s - loss: 0.2410 - acc: 0.9240 Epoch 237/500 - 3s - loss: 0.1954 - acc: 0.9440 Epoch 238/500 - 3s - loss: 0.1957 - acc: 0.9470 Epoch 239/500 - 3s - loss: 0.1955 - acc: 0.9420 Epoch 240/500 - 3s - loss: 0.1999 - acc: 0.9380 Epoch 241/500 - 3s - loss: 0.2351 - acc: 0.9320 Epoch 242/500 - 3s - loss: 0.1923 - acc: 0.9460 Epoch 243/500 - 3s - loss: 0.1950 - acc: 0.9470 Epoch 244/500 - 3s - loss: 0.2202 - acc: 0.9250 Epoch 245/500 - 3s - loss: 0.1897 - acc: 0.9470 Epoch 246/500 - 3s - loss: 0.1882 - acc: 0.9380 Epoch 247/500 - 3s - loss: 0.2372 - acc: 0.9320 Epoch 248/500 - 3s - loss: 0.1886 - acc: 0.9510 Epoch 249/500 - 3s - loss: 0.1855 - acc: 0.9460 Epoch 250/500 - 3s - loss: 0.1953 - acc: 0.9410 Epoch 251/500 - 3s - loss: 0.2807 - acc: 0.9300 Epoch 252/500 - 3s - loss: 0.1828 - acc: 0.9530 Epoch 253/500 - 3s - loss: 0.1841 - acc: 0.9480 Epoch 254/500 - 3s - loss: 0.1887 - acc: 0.9440 Epoch 255/500 - 3s - loss: 0.2025 - acc: 0.9440 Epoch 256/500 - 3s - loss: 0.2718 - acc: 0.9330 Epoch 257/500 - 3s - loss: 0.1811 - acc: 0.9500 Epoch 258/500 - 3s - loss: 0.1896 - acc: 0.9450 Epoch 259/500 - 3s - loss: 0.1826 - acc: 0.9440 Epoch 260/500 - 3s - loss: 0.1835 - acc: 0.9450 Epoch 261/500 - 3s - loss: 0.2578 - acc: 0.9300 Epoch 262/500 - 3s - loss: 0.1768 - acc: 0.9460 Epoch 263/500 - 3s - loss: 0.1785 - acc: 0.9540 Epoch 264/500 - 3s - loss: 0.1940 - acc: 0.9450 Epoch 265/500 - 3s - loss: 0.1810 - acc: 0.9480 Epoch 266/500 - 3s - loss: 0.1794 - acc: 0.9490 Epoch 267/500 - 3s - loss: 0.1969 - acc: 0.9490 Epoch 268/500 - 3s - loss: 0.2382 - acc: 0.9350 Epoch 269/500 - 3s - loss: 0.1754 - acc: 0.9530 Epoch 270/500 - 3s - loss: 0.1747 - acc: 0.9520 Epoch 271/500 - 3s - loss: 0.1791 - acc: 0.9420 Epoch 272/500 - 3s - loss: 0.1766 - acc: 0.9510 Epoch 273/500 - 3s - loss: 0.2068 - acc: 0.9370 Epoch 274/500 - 3s - loss: 0.2667 - acc: 0.9360 Epoch 275/500 - 3s - loss: 0.1700 - acc: 0.9540 Epoch 276/500 - 3s - loss: 0.1714 - acc: 0.9570 Epoch 277/500 - 3s - loss: 0.1738 - acc: 0.9490 Epoch 278/500 - 3s - loss: 0.1745 - acc: 0.9530 Epoch 279/500 - 3s - loss: 0.1729 - acc: 0.9480 Epoch 280/500 - 3s - loss: 0.1908 - acc: 0.9430 Epoch 281/500 - 3s - loss: 0.1674 - acc: 0.9580 Epoch 282/500 - 3s - loss: 0.1717 - acc: 0.9510 Epoch 283/500 - 3s - loss: 0.1718 - acc: 0.9520 Epoch 284/500 - 3s - loss: 0.1694 - acc: 0.9510 Epoch 285/500 - 3s - loss: 0.2790 - acc: 0.9290 Epoch 286/500 - 3s - loss: 0.1621 - acc: 0.9560 Epoch 287/500 - 3s - loss: 0.1618 - acc: 0.9590 Epoch 288/500 - 3s - loss: 0.1652 - acc: 0.9560 Epoch 289/500 - 3s - loss: 0.1636 - acc: 0.9560 Epoch 290/500 - 3s - loss: 0.2966 - acc: 0.9160 Epoch 291/500 - 3s - loss: 0.1625 - acc: 0.9580 Epoch 292/500 - 3s - loss: 0.1599 - acc: 0.9550 Epoch 293/500 - 3s - loss: 0.1608 - acc: 0.9620 Epoch 294/500 - 3s - loss: 0.1643 - acc: 0.9540 Epoch 295/500 - 3s - loss: 0.1633 - acc: 0.9500 Epoch 296/500 - 3s - loss: 0.1665 - acc: 0.9550 Epoch 297/500 - 3s - loss: 0.2146 - acc: 0.9400 Epoch 298/500 - 3s - loss: 0.1561 - acc: 0.9620 Epoch 299/500 - 3s - loss: 0.1581 - acc: 0.9570 Epoch 300/500 - 3s - loss: 0.1602 - acc: 0.9550 Epoch 301/500 - 3s - loss: 0.1692 - acc: 0.9490 Epoch 302/500 - 3s - loss: 0.1552 - acc: 0.9510 Epoch 303/500 - 3s - loss: 0.1590 - acc: 0.9530 Epoch 304/500 - 3s - loss: 0.1784 - acc: 0.9470 Epoch 305/500 - 3s - loss: 0.1576 - acc: 0.9590 Epoch 306/500 - 3s - loss: 0.1562 - acc: 0.9620 Epoch 307/500 - 3s - loss: 0.1543 - acc: 0.9610 Epoch 308/500 - 3s - loss: 0.1569 - acc: 0.9530 Epoch 309/500 - 3s - loss: 0.2514 - acc: 0.9330 Epoch 310/500 - 3s - loss: 0.1849 - acc: 0.9500 Epoch 311/500 - 3s - loss: 0.1482 - acc: 0.9620 Epoch 312/500 - 3s - loss: 0.1516 - acc: 0.9560 Epoch 313/500 - 3s - loss: 0.1524 - acc: 0.9580 Epoch 314/500 - 3s - loss: 0.1539 - acc: 0.9540 Epoch 315/500 - 3s - loss: 0.1509 - acc: 0.9600 Epoch 316/500 - 3s - loss: 0.3198 - acc: 0.9250 Epoch 317/500 - 3s - loss: 0.1453 - acc: 0.9670 Epoch 318/500 - 3s - loss: 0.1465 - acc: 0.9660 Epoch 319/500 - 3s - loss: 0.1476 - acc: 0.9630 Epoch 320/500 - 3s - loss: 0.1501 - acc: 0.9600 Epoch 321/500 - 3s - loss: 0.1522 - acc: 0.9600 Epoch 322/500 - 3s - loss: 0.1531 - acc: 0.9590 Epoch 323/500 - 3s - loss: 0.1496 - acc: 0.9600 Epoch 324/500 - 3s - loss: 0.2325 - acc: 0.9280 Epoch 325/500 - 3s - loss: 0.1583 - acc: 0.9630 Epoch 326/500 - 3s - loss: 0.1427 - acc: 0.9660 Epoch 327/500 - 3s - loss: 0.1446 - acc: 0.9620 Epoch 328/500 - 3s - loss: 0.1459 - acc: 0.9550 Epoch 329/500 - 3s - loss: 0.1471 - acc: 0.9620 Epoch 330/500 - 3s - loss: 0.1461 - acc: 0.9560 Epoch 331/500 - 3s - loss: 0.2298 - acc: 0.9440 Epoch 332/500 - 3s - loss: 0.1387 - acc: 0.9650 Epoch 333/500 - 3s - loss: 0.1411 - acc: 0.9620 Epoch 334/500 - 3s - loss: 0.1432 - acc: 0.9640 Epoch 335/500 - 3s - loss: 0.1423 - acc: 0.9600 Epoch 336/500 - 3s - loss: 0.1464 - acc: 0.9580 Epoch 337/500 - 3s - loss: 0.1438 - acc: 0.9590 Epoch 338/500 - 3s - loss: 0.2358 - acc: 0.9430 Epoch 339/500 - 3s - loss: 0.1338 - acc: 0.9700 Epoch 340/500 - 3s - loss: 0.1366 - acc: 0.9670 Epoch 341/500 - 3s - loss: 0.2185 - acc: 0.9550 Epoch 342/500 - 3s - loss: 0.1370 - acc: 0.9660 Epoch 343/500 - 3s - loss: 0.1374 - acc: 0.9680 Epoch 344/500 - 3s - loss: 0.1386 - acc: 0.9640 Epoch 345/500 - 3s - loss: 0.1387 - acc: 0.9680 Epoch 346/500 - 3s - loss: 0.1391 - acc: 0.9540 Epoch 347/500 - 3s - loss: 0.1387 - acc: 0.9640 Epoch 348/500 - 3s - loss: 0.1402 - acc: 0.9640 Epoch 349/500 - 3s - loss: 0.2935 - acc: 0.9390 Epoch 350/500 - 3s - loss: 0.1324 - acc: 0.9670 Epoch 351/500 - 3s - loss: 0.1339 - acc: 0.9690 Epoch 352/500 - 3s - loss: 0.1328 - acc: 0.9610 Epoch 353/500 - 3s - loss: 0.1356 - acc: 0.9650 Epoch 354/500 - 3s - loss: 0.1338 - acc: 0.9640 Epoch 355/500 - 3s - loss: 0.1340 - acc: 0.9610 Epoch 356/500 - 3s - loss: 0.1349 - acc: 0.9670 Epoch 357/500 - 3s - loss: 0.2347 - acc: 0.9440 Epoch 358/500 - 3s - loss: 0.1885 - acc: 0.9520 Epoch 359/500 - 3s - loss: 0.1281 - acc: 0.9760 Epoch 360/500 - 3s - loss: 0.1294 - acc: 0.9720 Epoch 361/500 - 3s - loss: 0.1302 - acc: 0.9640 Epoch 362/500 - 3s - loss: 0.1299 - acc: 0.9680 Epoch 363/500 - 3s - loss: 0.1316 - acc: 0.9660 Epoch 364/500 - 3s - loss: 0.2295 - acc: 0.9410 Epoch 365/500 - 3s - loss: 0.1261 - acc: 0.9730 Epoch 366/500 - 3s - loss: 0.1278 - acc: 0.9680 Epoch 367/500 - 3s - loss: 0.1272 - acc: 0.9680 Epoch 368/500 - 3s - loss: 0.1288 - acc: 0.9680 Epoch 369/500 - 3s - loss: 0.1271 - acc: 0.9720 Epoch 370/500 - 3s - loss: 0.1287 - acc: 0.9620 Epoch 371/500 - 3s - loss: 0.1293 - acc: 0.9710 Epoch 372/500 - 3s - loss: 0.1270 - acc: 0.9630 Epoch 373/500 - 3s - loss: 0.2037 - acc: 0.9530 Epoch 374/500 - 3s - loss: 0.1310 - acc: 0.9710 Epoch 375/500 - 3s - loss: 0.1204 - acc: 0.9750 Epoch 376/500 - 3s - loss: 0.1217 - acc: 0.9730 Epoch 377/500 - 3s - loss: 0.1250 - acc: 0.9690 Epoch 378/500 - 3s - loss: 0.1252 - acc: 0.9690 Epoch 379/500 - 3s - loss: 0.1267 - acc: 0.9680 Epoch 380/500 - 3s - loss: 0.1260 - acc: 0.9710 Epoch 381/500 - 3s - loss: 0.1262 - acc: 0.9630 Epoch 382/500 - 3s - loss: 0.1248 - acc: 0.9680 Epoch 383/500 - 3s - loss: 0.2545 - acc: 0.9430 Epoch 384/500 - 3s - loss: 0.1167 - acc: 0.9710 Epoch 385/500 - 3s - loss: 0.1187 - acc: 0.9760 Epoch 386/500 - 3s - loss: 0.1227 - acc: 0.9690 Epoch 387/500 - 3s - loss: 0.1202 - acc: 0.9740 Epoch 388/500 - 3s - loss: 0.1283 - acc: 0.9590 Epoch 389/500 - 3s - loss: 0.1203 - acc: 0.9720 Epoch 390/500 - 3s - loss: 0.1209 - acc: 0.9700 Epoch 391/500 - 3s - loss: 0.1176 - acc: 0.9680 Epoch 392/500 - 3s - loss: 0.2086 - acc: 0.9480 Epoch 393/500 - 3s - loss: 0.1133 - acc: 0.9800 Epoch 394/500 - 3s - loss: 0.1142 - acc: 0.9780 Epoch 395/500 - 3s - loss: 0.1164 - acc: 0.9710 Epoch 396/500 - 3s - loss: 0.1232 - acc: 0.9640 Epoch 397/500 - 3s - loss: 0.1172 - acc: 0.9740 Epoch 398/500 - 3s - loss: 0.1196 - acc: 0.9640 Epoch 399/500 - 3s - loss: 0.1183 - acc: 0.9780 Epoch 400/500 - 3s - loss: 0.2167 - acc: 0.9530 Epoch 401/500 - 3s - loss: 0.1167 - acc: 0.9700 Epoch 402/500 - 3s - loss: 0.1124 - acc: 0.9710 Epoch 403/500 - 3s - loss: 0.1124 - acc: 0.9680 Epoch 404/500 - 3s - loss: 0.1140 - acc: 0.9760 Epoch 405/500 - 3s - loss: 0.1160 - acc: 0.9740 Epoch 406/500 - 3s - loss: 0.1167 - acc: 0.9790 Epoch 407/500 - 3s - loss: 0.1182 - acc: 0.9680 Epoch 408/500 - 3s - loss: 0.2614 - acc: 0.9420 Epoch 409/500 - 3s - loss: 0.1100 - acc: 0.9770 Epoch 410/500 - 3s - loss: 0.1114 - acc: 0.9750 Epoch 411/500 - 3s - loss: 0.1104 - acc: 0.9760 Epoch 412/500 - 3s - loss: 0.1115 - acc: 0.9730 Epoch 413/500 - 3s - loss: 0.1767 - acc: 0.9600 Epoch 414/500 - 3s - loss: 0.1097 - acc: 0.9750 Epoch 415/500 - 3s - loss: 0.1108 - acc: 0.9760 Epoch 416/500 - 3s - loss: 0.1099 - acc: 0.9760 Epoch 417/500 - 3s - loss: 0.1327 - acc: 0.9630 Epoch 418/500 - 3s - loss: 0.1672 - acc: 0.9630 Epoch 419/500 - 3s - loss: 0.1090 - acc: 0.9800 Epoch 420/500 - 3s - loss: 0.1082 - acc: 0.9730 Epoch 421/500 - 3s - loss: 0.1104 - acc: 0.9720 Epoch 422/500 - 3s - loss: 0.1067 - acc: 0.9750 Epoch 423/500 - 3s - loss: 0.1122 - acc: 0.9740 Epoch 424/500 - 3s - loss: 0.1108 - acc: 0.9690 Epoch 425/500 - 3s - loss: 0.2437 - acc: 0.9460 Epoch 426/500 - 3s - loss: 0.1059 - acc: 0.9770 Epoch 427/500 - 3s - loss: 0.1048 - acc: 0.9830 Epoch 428/500 - 3s - loss: 0.1061 - acc: 0.9770 Epoch 429/500 - 3s - loss: 0.1075 - acc: 0.9740 Epoch 430/500 - 3s - loss: 0.1143 - acc: 0.9720 Epoch 431/500 - 3s - loss: 0.1189 - acc: 0.9710 Epoch 432/500 - 3s - loss: 0.1054 - acc: 0.9760 Epoch 433/500 - 3s - loss: 0.1051 - acc: 0.9760 Epoch 434/500 - 3s - loss: 0.1412 - acc: 0.9630 Epoch 435/500 - 3s - loss: 0.1736 - acc: 0.9640 Epoch 436/500 - 3s - loss: 0.1016 - acc: 0.9790 Epoch 437/500 - 3s - loss: 0.1026 - acc: 0.9790 Epoch 438/500 - 3s - loss: 0.1018 - acc: 0.9790 Epoch 439/500 - 3s - loss: 0.1064 - acc: 0.9740 Epoch 440/500 - 3s - loss: 0.1024 - acc: 0.9770 Epoch 441/500 - 3s - loss: 0.1070 - acc: 0.9770 Epoch 442/500 - 3s - loss: 0.1056 - acc: 0.9750 Epoch 443/500 - 3s - loss: 0.1034 - acc: 0.9790 Epoch 444/500 - 3s - loss: 0.1761 - acc: 0.9630 Epoch 445/500 - 3s - loss: 0.1004 - acc: 0.9830 Epoch 446/500 - 3s - loss: 0.1018 - acc: 0.9760 Epoch 447/500 - 3s - loss: 0.1011 - acc: 0.9770 Epoch 448/500 - 3s - loss: 0.1057 - acc: 0.9720 Epoch 449/500 - 3s - loss: 0.1017 - acc: 0.9800 Epoch 450/500 - 3s - loss: 0.1021 - acc: 0.9780 Epoch 451/500 - 3s - loss: 0.1054 - acc: 0.9710 Epoch 452/500 - 3s - loss: 0.1006 - acc: 0.9820 Epoch 453/500 - 3s - loss: 0.1027 - acc: 0.9780 Epoch 454/500 - 3s - loss: 0.1759 - acc: 0.9550 Epoch 455/500 - 3s - loss: 0.1061 - acc: 0.9840 Epoch 456/500 - 3s - loss: 0.0958 - acc: 0.9790 Epoch 457/500 - 3s - loss: 0.0960 - acc: 0.9800 Epoch 458/500 - 3s - loss: 0.0995 - acc: 0.9790 Epoch 459/500 - 3s - loss: 0.0982 - acc: 0.9810 Epoch 460/500 - 3s - loss: 0.0997 - acc: 0.9730 Epoch 461/500 - 3s - loss: 0.0998 - acc: 0.9770 Epoch 462/500 - 3s - loss: 0.2572 - acc: 0.9430 Epoch 463/500 - 3s - loss: 0.0932 - acc: 0.9820 Epoch 464/500 - 3s - loss: 0.0942 - acc: 0.9820 Epoch 465/500 - 3s - loss: 0.0946 - acc: 0.9820 Epoch 466/500 - 3s - loss: 0.0967 - acc: 0.9790 Epoch 467/500 - 3s - loss: 0.0947 - acc: 0.9840 Epoch 468/500 - 3s - loss: 0.0971 - acc: 0.9780 Epoch 469/500 - 3s - loss: 0.0976 - acc: 0.9810 Epoch 470/500 - 3s - loss: 0.0983 - acc: 0.9820 Epoch 471/500 - 3s - loss: 0.0962 - acc: 0.9760 Epoch 472/500 - 3s - loss: 0.0988 - acc: 0.9760 Epoch 473/500 - 3s - loss: 0.0929 - acc: 0.9820 Epoch 474/500 - 3s - loss: 0.0955 - acc: 0.9790 Epoch 475/500 - 3s - loss: 0.2046 - acc: 0.9600 Epoch 476/500 - 3s - loss: 0.0907 - acc: 0.9850 Epoch 477/500 - 3s - loss: 0.0903 - acc: 0.9820 Epoch 478/500 - 3s - loss: 0.0900 - acc: 0.9820 Epoch 479/500 - 3s - loss: 0.0925 - acc: 0.9810 Epoch 480/500 - 3s - loss: 0.0933 - acc: 0.9840 Epoch 481/500 - 3s - loss: 0.1802 - acc: 0.9570 Epoch 482/500 - 3s - loss: 0.0892 - acc: 0.9840 Epoch 483/500 - 3s - loss: 0.0891 - acc: 0.9820 Epoch 484/500 - 3s - loss: 0.0899 - acc: 0.9820 Epoch 485/500 - 3s - loss: 0.0907 - acc: 0.9840 Epoch 486/500 - 3s - loss: 0.0905 - acc: 0.9830 Epoch 487/500 - 3s - loss: 0.0921 - acc: 0.9790 Epoch 488/500 - 3s - loss: 0.0936 - acc: 0.9800 Epoch 489/500 - 3s - loss: 0.0913 - acc: 0.9770 Epoch 490/500 - 3s - loss: 0.0920 - acc: 0.9820 Epoch 491/500 - 3s - loss: 0.1491 - acc: 0.9720 Epoch 492/500 - 3s - loss: 0.1449 - acc: 0.9770 Epoch 493/500 - 3s - loss: 0.0857 - acc: 0.9850 Epoch 494/500 - 3s - loss: 0.0859 - acc: 0.9850 Epoch 495/500 - 3s - loss: 0.0866 - acc: 0.9870 Epoch 496/500 - 3s - loss: 0.0873 - acc: 0.9790 Epoch 497/500 - 3s - loss: 0.0881 - acc: 0.9820 Epoch 498/500 - 3s - loss: 0.0882 - acc: 0.9810 Epoch 499/500 - 3s - loss: 0.0880 - acc: 0.9870 Epoch 500/500 - 3s - loss: 0.0878 - acc: 0.9790 Model Accuracy: 98.30% ['T', 'U', 'V', 'W', 'X'] -> Y ['V', 'W', 'X', 'Y'] -> Z ['A', 'B', 'C', 'D'] -> E ['C'] -> D ['K', 'L', 'M', 'N'] -> O ['B'] -> C ['C', 'D', 'E', 'F', 'G'] -> H ['Q', 'R'] -> S ['T', 'U', 'V', 'W', 'X'] -> Y ['D', 'E', 'F', 'G', 'H'] -> I ['B', 'C', 'D', 'E', 'F'] -> G ['C', 'D', 'E', 'F'] -> G ['C'] -> D ['K', 'L', 'M'] -> N ['B', 'C', 'D', 'E'] -> F ['N', 'O'] -> P ['P'] -> Q ['W'] -> X ['V', 'W', 'X'] -> Y ['C'] -> D
Keras入门——(5)长短期记忆网络LSTM(二)的更多相关文章
- 如何预测股票分析--长短期记忆网络(LSTM)
在上一篇中,我们回顾了先知的方法,但是在这个案例中表现也不是特别突出,今天介绍的是著名的l s t m算法,在时间序列中解决了传统r n n算法梯度消失问题的的它这一次还会有令人杰出的表现吗? 长短期 ...
- Keras入门——(7)长短期记忆网络LSTM(四)
数据准备:http://www.manythings.org/anki/cmn-eng.zip 源代码:https://github.com/pjgao/seq2seq_keras 参考:https: ...
- Keras入门——(6)长短期记忆网络LSTM(三)
参考: https://blog.csdn.net/u012735708/article/details/82769711 https://zybuluo.com/hanbingtao/note/58 ...
- Keras入门——(4)长短期记忆网络LSTM(一)
参考: https://blog.csdn.net/zwqjoy/article/details/80493341 https://blog.csdn.net/u012735708/article/d ...
- 递归神经网络之理解长短期记忆网络(LSTM NetWorks)(转载)
递归神经网络 人类并不是每时每刻都从头开始思考.正如你阅读这篇文章的时候,你是在理解前面词语的基础上来理解每个词.你不会丢弃所有已知的信息而从头开始思考.你的思想具有持续性. 传统的神经网络不能做到这 ...
- 理解长短期记忆网络(LSTM NetWorks)
转自:http://www.csdn.net/article/2015-11-25/2826323 原文链接:Understanding LSTM Networks(译者/刘翔宇 审校/赵屹华 责编/ ...
- LSTM - 长短期记忆网络
循环神经网络(RNN) 人们不是每一秒都从头开始思考,就像你阅读本文时,不会从头去重新学习一个文字,人类的思维是有持续性的.传统的卷积神经网络没有记忆,不能解决这一个问题,循环神经网络(Recurre ...
- LSTMs 长短期记忆网络系列
RNN的长期依赖问题 什么是长期依赖? 长期依赖是指当前系统的状态,可能受很长时间之前系统状态的影响,是RNN中无法解决的一个问题. 如果从(1) “ 这块冰糖味道真?”来预测下一个词,是很容易得出“ ...
- LSTM(长短期记忆网络)及其tensorflow代码应用
本文主要包括: 一.什么是LSTM 二.LSTM的曲线拟合 三.LSTM的分类问题 四.为什么LSTM有助于消除梯度消失 一.什么是LSTM Long Short Term 网络即为LSTM,是一种 ...
随机推荐
- 5_1 大理石在哪儿(UVa10474)<排序与查找>
Raju和Meena喜欢玩弹珠,他们有许多上面有号码的弹珠.一开始时,Raju按照弹珠上面的号码由小到大排成一列,然后Meena会要求Raju找出某个号码的第一颗弹珠所在的位置.她会算1…2…3…,如 ...
- 【代码学习】PYTHON 异常处理
一.什么是异常 在程序执行过程中可能会影响程序的正常执行,一般情况下,在python无法正常处理程序时就会发生一个异常 当python脚本发生异常时我们需要捕获处理他,否则程序会终止执行 二.异常处理 ...
- QSS基础-设置控件样式
1.QSS基础-设置同类控件样式'''QSS基础:Qt Style SheetQt样式表用于设置控件的样式和风格(比如控件的背景色,字体颜色,字体大小等)和CSS的功能比较相似,功能相似''' fro ...
- Spring Boot 学习前你应该知道的 Maven 知识
Maven 是什么? 回答这个问题,我们先来了解下没有Maven,我们是怎么使用开发者工具IDE去开发Java程序的.我之前开发Java程序不多,但是我还是记得,我是从网上下载或从合作方拷贝 jar ...
- inline-block,真的懂吗
曾几何时,display:inline-block 已经深入「大街小巷」,随处可见 「display:inline-block; *display:inline; *zoom:1; 」这样的代码.如今 ...
- 【原】Django数据Model层总结
vlaues - 单条记录 - <class 'dict'> 多条记录 - <class 'django.db.models.query.QuerySet'> vlaues_l ...
- leetcode 0211
目录 ✅ 1217. 玩筹码 描述 解答 c java py ✅ 206. 反转链表 描述 解答 c java py ✅ 922. 按奇偶排序数组 II 描述 解答 c 双指针soldier tddo ...
- bat获取当前路径
@echo off setlocal EnableDelayedExpansion echo 当前正在运行的批处理文件所在路径:!cd! pause @echo off echo 当前目录是:%cd% ...
- BinaryTree(二叉树)
我认为二叉树的递归实现体现了递归思想的一些重要性质,如果对递归的理解不够的话,想利用递归来实现是很费劲的(实际上我现在都还有些懵...),虽然会用,但一些地方不能弄清楚原因. 经过几天的学习,看了许多 ...
- linux文件或目录属性
wc(word count)命令的功能:统计指定文件的字节数.字数.行数.,并将统计结果显示输出 命令参数: -c 只显示字节数 -l 只显示行数 -w 只显示字数 od命令:查看二进制文件信息 ...