Lr.find not working, NaN loss during training

Hi,

I am working in classification problem with images, and lr.find is not working (error message: “Failed to compute the gradients, there might not be enough points”). If I try to train the model, I immediately get NaN as training loss, sometimes the computer freezes.

I tried reducing size of images and batch size, it didn’t work.

When running on CPU it does work.

=== Software === 
python        : 3.7.4
fastai        : 1.0.59
fastprogress  : 0.1.21
torch         : 1.2.0
nvidia driver : 430.50
torch cuda    : 10.0.130 / is available
torch cudnn   : 7600 / is enabled

=== Hardware === 
nvidia gpus   : 1
torch devices : 1
  - gpu0      : 11016MB | GeForce RTX 2080 Ti

=== Environment === 
platform      : Linux-4.15.0-66-generic-x86_64-with-debian-buster-sid
distro        : #75-Ubuntu SMP Tue Oct 1 05:24:09 UTC 2019
conda env     : fastai
python        : /home/vip/anaconda3/envs/fastai/bin/python
sys.path      : 
/home/vip/anaconda3/envs/fastai/lib/python37.zip
/home/vip/anaconda3/envs/fastai/lib/python3.7
/home/vip/anaconda3/envs/fastai/lib/python3.7/lib-dynload
/home/vip/anaconda3/envs/fastai/lib/python3.7/site-packages

Tue Oct 29 09:50:23 2019       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 430.50       Driver Version: 430.50       CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:01:00.0  On |                  N/A |
| 38%   39C    P8    11W / 250W |    268MiB / 11016MiB |     13%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      1445      G   /usr/lib/xorg/Xorg                            18MiB |
|    0      1484      G   /usr/bin/gnome-shell                          57MiB |
|    0      1842      G   /usr/lib/xorg/Xorg                           106MiB |
|    0      2001      G   /usr/bin/gnome-shell                          84MiB |
+-----------------------------------------------------------------------------+