DIGITS6.1.1 Check failed: error == cudaSuccess (48 vs. 0) no kernel image is available for execution on the device

when i run DIGITS 6.1.1 with Train Caffe Model and find flowed error. How do i slove these problem?

(0.0) Test net output #0: accuracy = 0
(0.0) Test net output #1: loss = 0 (* 1 = 0 loss)
Initial Test completed in 0.00633155s
Restarting 4 internal thread(s) on device 0
{0} Starting 1 internal thread(s) on device 0
Loading mean file from: /workspace/jobs/20190816-091043-4e34/mean.binaryproto
Loading mean file from: /workspace/jobs/20190816-091043-4e34/mean.binaryproto
Loading mean file from: /workspace/jobs/20190816-091043-4e34/mean.binaryproto
Loading mean file from: /workspace/jobs/20190816-091043-4e34/mean.binaryproto
Data Reader threads: 3, out queues: 12, depth: 64
{0} Starting 3 internal thread(s) on device 0
Output data size: 64, 1, 28, 28
Parser threads: 3 (auto)
Transformer threads: 4 (auto)
Opened lmdb /workspace/jobs/20190816-091043-4e34/train_db
Opened lmdb /workspace/jobs/20190816-091043-4e34/train_db
Opened lmdb /workspace/jobs/20190816-091043-4e34/train_db
{0.0} NVML succeeded to set CPU affinity
{0.0} NVML succeeded to set CPU affinity
Check failed: error == cudaSuccess (48 vs. 0) no kernel image is available for execution on the device

I used nvcr.io/nvidia/digits 18.10.