search

vipzgy / AttentionTargetSentiment

reproduce "Attention Modeling for Targeted Sentiment"
MIT License
9 stars 3 forks source link

return torch.matmul(self, other) File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/functional.py", line 192, in matmul output = torch.mm(tensor1, tensor2) RuntimeError: size mismatch at /pytorch/torch/lib/THC/generic/THCTensorMathBlas.cu:247 #4

Open monajalal opened 6 years ago

monajalal commented 6 years ago

Do you know how to fix the below error?

[jalal@goku AttentionTargetSentiment]$ python main.py -cuda -getF1 -which-model contextualized -which-embedding 200dt -device 0 

Loading data...
all 23095
notfound: 9849
ratio: 0.4264559428447716

Parameters:
    ATTENTION_SIZE=150
    BATCH_SIZE=16
    CLIP_NORM=None
    CUDA=True
    DEVICE=0
    DROPOUT_EMBED=0.2
    DROPOUT_RNN=0.4
    EMBED_DIM=200
    EMBED_NUM=23095
    EPOCHS=30
    GETF1=True
    GRAYSCALE=None
    HIDDEN_SIZE=150
    IF_RE=False
    LABEL_NUM=3
    LOG_INTERVAL=1
    LR=0.01
    LR_SCHEDULER=None
    MAX_NORM=None
    MESSAGE=tt
    NEED_SMALLEMBED=False
    SAVE_DIR=snapshot/tt
    SAVE_INTERVAL=100
    SHUFFLE=True
    SNAPSHOT=None
    TEST=False
    TEST_INTERVAL=100
    USE_EMBEDDING=True
    WEIGHT_DECAY=1e-06
    WHICH_DATA=Z
    WHICH_EMBEDDING=200dt
    WHICH_INIT=xavier
    WHICH_MODEL=contextualized
    WHICH_OPTIM=Adagrad
Traceback (most recent call last):
  File "main.py", line 113, in <module>
    train.getF1(args, m_model, test_data.iterator, train_data.vocabulary_label)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/train.py", line 152, in getF1
    logit = model(feature, batch.target_start, batch.target_end)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/model/contextualized.py", line 90, in forward
    s = self.attention(x, average_target)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/model/attention.py", line 22, in forward
    m_combine = F.tanh(self.linear(m_combine))
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/linear.py", line 55, in forward
    return F.linear(input, self.weight, self.bias)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/functional.py", line 837, in linear
    output = input.matmul(weight.t())
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/autograd/variable.py", line 386, in matmul
    **return torch.matmul(self, other)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/functional.py", line 192, in matmul
    output = torch.mm(tensor1, tensor2)
RuntimeError: size mismatch at /pytorch/torch/lib/THC/generic/THCTensorMathBlas.cu:247**
monajalal commented 6 years ago

Note that when I only add the parameter which model this is caused:

[jalal@goku AttentionTargetSentiment]$ python main.py  -which-model contextualized 

Loading data...

Parameters:
    ATTENTION_SIZE=150
    BATCH_SIZE=16
    CLIP_NORM=None
    CUDA=False
    DEVICE=-1
    DROPOUT_EMBED=0.2
    DROPOUT_RNN=0.4
    EMBED_DIM=200
    EMBED_NUM=23095
    EPOCHS=30
    GETF1=False
    GRAYSCALE=None
    HIDDEN_SIZE=150
    IF_RE=False
    LABEL_NUM=3
    LOG_INTERVAL=1
    LR=0.01
    LR_SCHEDULER=None
    MAX_NORM=None
    MESSAGE=tt
    NEED_SMALLEMBED=False
    SAVE_DIR=snapshot/tt
    SAVE_INTERVAL=100
    SHUFFLE=True
    SNAPSHOT=None
    TEST=False
    TEST_INTERVAL=100
    USE_EMBEDDING=True
    WEIGHT_DECAY=1e-06
    WHICH_DATA=Z
    WHICH_EMBEDDING=200d
    WHICH_INIT=xavier
    WHICH_MODEL=contextualized
    WHICH_OPTIM=Adagrad
第 1 次迭代
Traceback (most recent call last):
  File "main.py", line 121, in <module>
    train.train(args, m_model, train_data.iterator, test_data.iterator)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/train.py", line 56, in train
    logit = model(feature, batch.target_start, batch.target_end)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/model/contextualized.py", line 90, in forward
    s = self.attention(x, average_target)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/model/attention.py", line 22, in forward
    m_combine = F.tanh(self.linear(m_combine))
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 357, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/modules/linear.py", line 55, in forward
    return F.linear(input, self.weight, self.bias)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/nn/functional.py", line 837, in linear
    output = input.matmul(weight.t())
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/autograd/variable.py", line 386, in matmul
    return torch.matmul(self, other)
  File "/scratch/sjn/anaconda/lib/python3.6/site-packages/torch/functional.py", line 192, in matmul
    output = torch.mm(tensor1, tensor2)
RuntimeError: size mismatch, m1: [1504 x 300], m2: [600 x 150] at /pytorch/torch/lib/TH/generic/THTensorMath.c:1434
[jalal@goku AttentionTargetSentiment]$ python main.py 

Loading data...

Parameters:
    ATTENTION_SIZE=150
    BATCH_SIZE=16
    CLIP_NORM=None
    CUDA=False
    DEVICE=-1
    DROPOUT_EMBED=0.2
    DROPOUT_RNN=0.4
    EMBED_DIM=200
    EMBED_NUM=23095
    EPOCHS=30
    GETF1=False
    GRAYSCALE=None
    HIDDEN_SIZE=150
    IF_RE=False
    LABEL_NUM=3
    LOG_INTERVAL=1
    LR=0.01
    LR_SCHEDULER=None
    MAX_NORM=None
    MESSAGE=tt
    NEED_SMALLEMBED=False
    SAVE_DIR=snapshot/tt
    SAVE_INTERVAL=100
    SHUFFLE=True
    SNAPSHOT=None
    TEST=False
    TEST_INTERVAL=100
    USE_EMBEDDING=True
    WEIGHT_DECAY=1e-06
    WHICH_DATA=Z
    WHICH_EMBEDDING=200d
    WHICH_INIT=xavier
    WHICH_MODEL=vanilla
    WHICH_OPTIM=Adagrad
第 1 次迭代
/scratch2/debate_tweets/sentiment/AttentionTargetSentiment/model/vanilla.py:82: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
  alfa = F.softmax(beta)

Without adding contextualized as a parameter to which model it works fine but how can I add contextualized as a parameter to which model without getting the error above?

vipzgy commented 6 years ago

parameter: BATCH_SIZE=1