Describe the bug
from torch import nn,optim
import torch
import torch.nn.functional as F
from torch.autograd import Variable
def cumsum_1d_with_zero(input: torch.Tensor, include_last: bool = False) -> torch.Tensor:
output = torch.zeros(input.size(0) + 1, dtype=input.dtype)
torch.cumsum(input, 0, out=output[1:])
if include_last:
return output
else:
return output[:-1]
class MyModel(nn.Module):
def __init__(self):
super(MyModel, self).__init__()
self.embedding = nn.EmbeddingBag(4, 300, sparse=True, mode='sum')
self.linear1 = nn.Linear(300, 1)
self.optimizer = optim.SparseAdam(self.embedding.parameters(), 0.0001)
self.dense_optimizer = optim.Adam(self.linear1.parameters(), 0.0001)
def forward(self,inputs):
rows = [torch.tensor(input, dtype=torch.int64) for input in inputs]
lengths = torch.tensor([x.size(0) for x in rows], dtype=torch.int64)
offsets = cumsum_1d_with_zero(lengths)
indices = torch.cat(rows)
embedding_res = self.embedding(indices, offsets)
print(embedding_res)
logits = F.sigmoid(self.linear1(embedding_res))
return logits
def optimize(self, logits):
labels = Variable(torch.LongTensor([0, 1, 0, 1]).reshape(4,1).float())
criterion = nn.BCEWithLogitsLoss()
loss = criterion(logits, labels)
print(loss.item())
loss.backward()
self.optimizer.step()
self.dense_optimizer.step()
run with this! but sparseAdam hit bug!
inputs = [[],[],[],[]]
result = model(inputs)
model.optimize(result)
- Thanks! SparseAdam will hit this bug, but SGD optimizer is ok.
Versions
Collecting environment information…
PyTorch version: 1.8.2+cu111
Is debug build: False
CUDA used to build PyTorch: 11.1
ROCM used to build PyTorch: N/A
OS: Ubuntu 20.04.2 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.31
Python version: 3.8.10 (default, Jun 22 2022, 20:18:18) [GCC 9.4.0] (64-bit runtime)
Python platform: Linux-5.15.0-1022-aws-x86_64-with-glibc2.29
Is CUDA available: False
CUDA runtime version: 10.1.243
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
Versions of relevant libraries:
[pip3] numpy==1.23.0
[pip3] torch==1.8.2+cu111
[conda] Could not collect