anoperson / deepie Goto Github PK
View Code? Open in Web Editor NEWAn Information Extraction Framework with Deep Learning developed at New York University
An Information Extraction Framework with Deep Learning developed at New York University
rzai@rzai00:~/prj/DeepIE/rnnRE$ CUDA_VISIBLE_DEVICES=1 python evaluate.py
Using gpu device 0: GeForce GTX 1080 (CNMeM is disabled)
{'useHeadEmbedding': False, 'conv_win_feature_map1': [2, 3, 4, 5], 'optimizer': 'adadelta', 'verbose': 1, 'regularizer': 0.0, 'nhidden2': 300, 'collapsed': False, 'nhidden1': 300, 'withEmbs': True, 'fold': 'all', 'expected_features2': OrderedDict([('dist1', -1), ('dist2', -1), ('type', -1), ('subtype', -1), ('order', -1), ('constit1', -1), ('constit2', -1), ('preter', -1), ('prepreter', -1), ('grammar', -1), ('gov', -1), ('indep', -1)]), 'seed': 3435, 'norm_lim': 9.0, 'kernelFets': OrderedDict([('kernelPred', 0), ('kernelScore', 0)]), 'updateEmbs': True, 'nepochs': 20, 'binaryCutoff': 2, 'seqType': '-dep', 'multilayerNN1': [300], 'outer': False, 'conv_feature_map2': 150, 'conv_feature_map1': 150, 'expected_features1': OrderedDict([('dist1', -1), ('dist2', -1), ('type', 0), ('subtype', -1), ('order', -1), ('constit1', -1), ('constit2', -1), ('preter', -1), ('prepreter', 0), ('grammar', 1), ('gov', -1), ('indep', 1)]), 'conv_win_feature_map2': [2, 3, 4, 5], 'decay': False, 'batch': 50, 'multilayerNN2': [], 'conv_winre1': 20, 'folder': 'fold_all.model_#MultiNN.st_-dep.cl_0.h1_300.h2_300.outer_False.embs_True.upd_True.batch_50.cut_2.he_0.mul1_300.mul2_.opt_adadelta.drop_0.5.reg_0.0.fet1_-1-10-1-1-1-1-101-11.ke_00.cvw1_20.cvw2_20.cvft1_150.cvft2_150.cvfm1_2345.nm_9.0', 'conv_winre2': 20, 'lr': 0.01, 'model': '#MultiNN', 'dropout': 0.5, 'sharedEmbs': OrderedDict([('word', 1), ('dist1', 0), ('dist2', 0), ('type', 0), ('subtype', 0), ('order', 0), ('constit1', 0), ('constit2', 0), ('preter', 0), ('prepreter', 0), ('grammar', 0), ('gov', 0), ('indep', 0)])}
loading dataset: ./word2vec.full_rnnRE_Plank.pkl ...
using word embeddings to initialize the network ...
sequence types: -dep
using feature1: type : embeddings
using feature1: prepreter : embeddings
using feature1: grammar : binary
using feature1: indep : binary
maximum of length in the dataset: 43 10
-------creating binary feature dictionary on the training data--------
binary feature cutoff: 2
size of dictionary: 138
maximum number of binary features: 80
converting binary features to vectors ...
vocabsize = 274 , nclasses = 8 , nsentences = 10 , word embeddings dim = 300
------- length of the instances: 43 20
... number of batches: 0
building model ...
FEATURES 1
REPRESENTATION DIMENSION 1 = 0
REPRESENTATION DIMENSION -1 = 0
FEATURES 2
done
-------------------training in epoch: 0 -------------------------------------
evaluating in epoch: 0
Traceback (most recent call last):
File "evaluate.py", line 153, in
main(pars)
File "evaluate.py", line 40, in main
folder = params['folder'])
File "/home/rzai/prj/DeepIE/rnnRE/rnnRE.py", line 971, in train
_predictions[elu], _probs[elu], _groundtruth[elu] = predict(evaluatingDataset[elu], batch, reModel, idx2word, idx2label, features1, features2, model, seqRep, kernelFets)
File "/home/rzai/prj/DeepIE/rnnRE/rnnRE.py", line 548, in predict
probs_corpus = numpy.concatenate(probs_corpus, axis=0)
ValueError: need at least one array to concatenate
rzai@rzai00:~/prj/DeepIE/rnnRE$
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.