johnbroughton2017 opened a new issue #16239: train_imagenet.py fails with 
float16 + alexnet 
URL: https://github.com/apache/incubator-mxnet/issues/16239
 
 
   ## Description
   I found the following error when training an alexnet using fp16. It is very 
easy reproduce using the command below. I have seen a few people reporting this 
bug yet I haven't been able to find a solution. If you replace alexnet with 
reset, it works fine. 
   
   ## Error Message:
   
   ```
   mxnet.base.MXNetError: [14:57:10] include/mxnet/././tensor_blob.h:236: Check 
failed: mshadow::DataType<DType>::kFlag == type_flag_: TBlob.get_with_shape: 
data type do not match specified type.Expected: 2 v.s. given 0
   Stack trace:
     [bt] (0) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x4a3b8b)
 [0x7f377f32ab8b]
     [bt] (1) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x304366d)
 [0x7f3781eca66d]
     [bt] (2) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x3553e98)
 [0x7f37823dae98]
     [bt] (3) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x2675018)
 [0x7f37814fc018]
     [bt] (4) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x267abf5)
 [0x7f3781501bf5]
     [bt] (5) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x265a8a1)
 [0x7f37814e18a1]
     [bt] (6) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x265ddb0)
 [0x7f37814e4db0]
     [bt] (7) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x265e046)
 [0x7f37814e5046]
     [bt] (8) 
/home/xinghua/anaconda2/lib/python2.7/site-packages/mxnet/libmxnet.so(+0x2659004)
 [0x7f37814e0004]
   
   ```
   
   ## Minimum reproducible example
   ```
   python train_imagenet.py --network alexnet --benchmark 1 --gpus 0 
--batch-size 64 --dtype float16
   ```

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to