search

zhangmeishan / BiaffineDParser

BiAffine Dependency Parsing
53 stars 16 forks source link

ValueError: operands could not be broadcast together with shapes (128,) (127,) (128,) #3

Open Nidachuhan opened 4 years ago

Nidachuhan commented 4 years ago

I run this on python 3.7 , Pytorch version 1.1.0 when I run the python driver/TrainTest.py --config_file config.ctb51.cfg I found this error please resolve this thanks Traceback (most recent call last): File "driver/TrainTest.py", line 149, in vec = vocab.load_pretrained_embs(config.pretrained_embeddings_file) File ".\data\Vocab.py", line 60, in load_pretrained_embs embeddings[self.UNK] += vector ValueError: operands could not be broadcast together with shapes (128,) (127,) (128,)

HassanNaeemjutt commented 4 years ago

I have same problems

Loaded config file sucessfully. pretrained_embeddings_file experiments/ctb51/giga.txt data_dir experiments/ctb51 train_file experiments/ctb51/zh-ud-train.conllu dev_file experiments/ctb51/zh-ud-dev.conllu test_file experiments/ctb51/zh-ud-test.conllu min_occur_count 2 save_dir experiments/ctb51_model config_file experiments/ctb51_model/config.cfg save_model_path experiments/ctb51_model/model save_vocab_path experiments/ctb51_model/vocab load_dir ../ckpt/default load_model_path ../ckpt/default/model load_vocab_path ../ckpt/default/vocab lstm_layers 3 word_dims 128 tag_dims 128 dropout_emb 0.33 lstm_hiddens 400 dropout_lstm_input 0.33 dropout_lstm_hidden 0.33 mlp_arc_size 500 mlp_rel_size 100 dropout_mlp 0.33 learning_rate 2e-3 decay .75 decay_steps 5000 beta_1 .9 beta_2 .9 epsilon 1e-12 clip 5.0 num_buckets_train 40 num_buckets_valid 10 num_buckets_test 10 train_iters 50000 train_batch_size 50 test_batch_size 128 validate_every 165 save_after 5000 update_every 4 Total num: 3997 Proj num: 3975 Vocab info: #words 4424, #tags 17, #rels 43 Total words: 42422

The dim of pretrained embeddings: 128

Traceback (most recent call last): File "driver/TrainTest.py", line 149, in vec = vocab.load_pretrained_embs(config.pretrained_embeddings_file) File ".\data\Vocab.py", line 60, in load_pretrained_embs embeddings[self.UNK] += vector ValueError: operands could not be broadcast together with shapes (128,) (127,) (128,)

HongyanJiao commented 4 years ago

I have the same problem The dim of pretrained embeddings: 100

Traceback (most recent call last): File "driver/TrainTest.py", line 150, in vec = vocab.load_pretrained_embs(config.pretrained_embeddings_file) File "./data/Vocab.py", line 60, in load_pretrained_embs embeddings[self.UNK] += vector ValueError: operands could not be broadcast together with shapes (100,) (99,) (100,)