ENUM error when running basset_train.lua #53
Comments
|
Sounds like an HDF5 compatibility issue. Check out this previous issue #40. Unfortunately, you’ll need to use HDF5 version 1.8 as discussed there. |
|
Hi Dave, Thanks for the response. I already had HDF5 1.8 installed. To double check I ran h5dump --version and I got h5dump: Version 1.8.12. However, when I start up torch and run require 'hdf5' it gives me something similar to issue #25 (comment). I have attached what I get: th> require 'hdf5' This made me think that I might have not installed https://github.com/davek44/torch-hdf5.git properly. I re-installed it but still get the same ENUM error when I run basset_train.lua. Here are my installation logs for installing torch-hdf5. Was wondering if it should like this for a successful install for torch-hdf5? [zainmunirpatel@bc2 torch-hdf5]$ luarocks make -- The C compiler identification is GNU 4.8.5 Thanks for the help and let me know if you think there is some other issue. Best, |
|
The last thing that I can think of is that you also have deepmind’s version of torch-hdf5 installed just like Sam in #25. Maybe uninstall torch-hdf5 to make sure you have a clean slate before installing my version. Otherwise, I’m not sure I’m going to be able to figure this one out. So, I prioritize Basset-like behavior in the Basenji package to give you a path forward. |
|
Hi, I have Basset-like behavior implemented in Basenji now. Check out https://github.com/calico/basenji/tree/master/manuscripts/basset |
|
Hi Dave, |
Hi,I have the nearly the same problem with you,but I can't figure out what's going wrong.I reinstall the hdf5 1.8.18. But I still receive ENUM error. Any idea will be appreciated! |
The error comes from your .h5 file |
Hi Dave,
When I run basset_train.lua -job pretrained_params.txt -stagnant_t 10 er.h5, I end up getting an ENUM(50331977) error. I have tried diagnosing it, but haven't had any luck. Below are the error logs:
{
conv_filter_sizes :
{
1 : 19
2 : 11
3 : 7
}
weight_norm : 7
momentum : 0.98
learning_rate : 0.002
hidden_units :
{
1 : 1000
2 : 1000
}
conv_filters :
{
1 : 300
2 : 200
3 : 200
}
hidden_dropouts :
{
1 : 0.3
2 : 0.3
}
pool_width :
{
1 : 3
2 : 4
3 : 4
}
}
seq_len: 600, filter_size: 19, pad_width: 18
seq_len: 200, filter_size: 11, pad_width: 10
seq_len: 50, filter_size: 7, pad_width: 6
nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> (7) -> (8) -> (9) -> (10) -> (11) -> (12) -> (13) -> (14) -> (15) -> (16) -> (17) -> (18) -> (19) -> (20) -> (21) -> (22) -> (23) -> output]
(1): nn.SpatialConvolution(4 -> 300, 19x1, 1,1, 9,0)
(2): nn.SpatialBatchNormalization (4D) (300)
(3): nn.ReLU
(4): nn.SpatialMaxPooling(3x1, 3,1)
(5): nn.SpatialConvolution(300 -> 200, 11x1, 1,1, 5,0)
(6): nn.SpatialBatchNormalization (4D) (200)
(7): nn.ReLU
(8): nn.SpatialMaxPooling(4x1, 4,1)
(9): nn.SpatialConvolution(200 -> 200, 7x1, 1,1, 3,0)
(10): nn.SpatialBatchNormalization (4D) (200)
(11): nn.ReLU
(12): nn.SpatialMaxPooling(4x1, 4,1)
(13): nn.Reshape(2600)
(14): nn.Linear(2600 -> 1000)
(15): nn.BatchNormalization (2D) (1000)
(16): nn.ReLU
(17): nn.Dropout(0.300000)
(18): nn.Linear(1000 -> 1000)
(19): nn.BatchNormalization (2D) (1000)
(20): nn.ReLU
(21): nn.Dropout(0.300000)
(22): nn.Linear(1000 -> 164)
(23): nn.Sigmoid
}
/home/hugheslab2/zainmunirpatel/torch/install/bin/luajit: .../zainmunirpatel/torch/install/share/lua/5.1/hdf5/ffi.lua:335: Reading data of class ENUM(50331977) is unsupported
stack traceback:
[C]: in function 'error'
.../zainmunirpatel/torch/install/share/lua/5.1/hdf5/ffi.lua:335: in function '_getTorchType'
...nmunirpatel/torch/install/share/lua/5.1/hdf5/dataset.lua:88: in function 'getTensorFactory'
...nmunirpatel/torch/install/share/lua/5.1/hdf5/dataset.lua:138: in function 'partial'
/home/hugheslab2/zainmunirpatel/Basset/src/batcher.lua:39: in function 'next'
/home/hugheslab2/zainmunirpatel/Basset/src/convnet.lua:1009: in function 'train_epoch'
...e/hugheslab2/zainmunirpatel/Basset//src/basset_train.lua:156: in main chunk
[C]: in function 'dofile'
...atel/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
[C]: at 0x00406460
Epoch # 1 [zainmunirpatel@bc2 data]$
Thanks for the help!
Best,
Zain
The text was updated successfully, but these errors were encountered: