DIGITS HOME Training Images fail to path

Dears guys.
I have problem to start DIGITS START for install “New Image Classification Dataset” Screen IE:host login page display “Training images >> Folder does not exist”. How can i do?

I am not very clear about the running environment. If you were running DIGITS container, the folder path is the path inside the container.

Can you give us some details? Did you mount a volume to your DIGITS container? Where did you save your datasets? For example, where was the downloaded mnist dataset?

Ubuntu 16.04
±----------------------------------------------------------------------------+
| NVIDIA-SMI 430.14 Driver Version: 430.14 CUDA Version: 10.2 |
|-------------------------------±---------------------±---------------------+
| 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 GTX 166… Off | 00000000:01:00.0 Off | N/A |
| 32% 39C P0 1W / 120W | 0MiB / 5941MiB | 0% Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
±----------------------------------------------------------------------------+
Client:
Version: 18.09.6
API version: 1.39
Go version: go1.10.8
Git commit: 481bc77
Built: Sat May 4 02:35:27 2019
OS/Arch: linux/amd64
Experimental: false

Server: Docker Engine - Community
Engine:
Version: 18.09.6
API version: 1.39 (minimum version 1.12)
Go version: go1.10.8
Git commit: 481bc77
Built: Sat May 4 01:59:36 2019
OS/Arch: linux/amd64
Experimental: false

Let me assume you are running DIGITS container, and your docker run command for DIGITS is

docker run -it --rm --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=0 -v /home/t105319009/dataset/mnist:/workspace/data/mnist -p 5000:5000 nvcr.io/nvidia/digits:19.05-caffe

.

The -v option in the above command will map your local /home/t105319009/dataset/mnist directory to /workspace/data/mnist in the container.

Therefore, when you try to create a dataset, you can specify /workspace/data/mnist in the following screen.

Please ensure your local /home/t105319009/dataset directory has the mnist dataset, similar to the following structure. This directory can be created by running

python -m digits.download_data mnist ~/mnist

as explained in https://github.com/NVIDIA/DIGITS/blob/master/docs/GettingStarted.md

$ tree -L 2 dataset
dataset
└── mnist
    ├── t10k-images-idx3-ubyte.gz
    ├── t10k-labels-idx1-ubyte.gz
    ├── test
    ├── test-images.bin
    ├── test-labels.bin
    ├── train
    ├── train-images.bin
    ├── train-images-idx3-ubyte.gz
    ├── train-labels.bin
    └── train-labels-idx1-ubyte.gz

Thanks you very much.