galliot-us / neuralet

Neuralet is an open-source platform for edge deep learning models on edge TPU, Jetson Nano, and more.
https://neuralet.com
Apache License 2.0
238 stars 71 forks source link

I have problem to download model for jetson-nano #110

Closed grdsvt closed 4 years ago

grdsvt commented 4 years ago

Hi, I used :

sudo docker pull neuralet/jetson-nano

to download model for Jetson-nano for social distancing, but I received the message:

Using default tag: latest Error response from daemon: manifest for neuralet/jetson-nano:latest not found: manifest unknown: manifest unknown

Where is the error? I see that is not present model for jetson too but in mail I received that it is available. Can sameone help me?

Thanks

alpha-carinae29 commented 4 years ago

Hi You can pull the Jetson Nano docker by running: docker pull neuralet/jetson-nano:applications-smart-distancing or build the container locally by clone this repository and run followings:

cd neuralet/applications/smart-distancing/

# 1) Download TensorRT engine file built with JetPack 4.3:
./download_jetson_trt.sh

# 2) Build Docker image (This step is optional, you can skip it if you want to pull the container from neuralet dockerhub)
docker build -f jetson-nano.Dockerfile -t "neuralet/jetson-nano:applications-smart-distancing" .

# 3) Run Docker container:
docker run -it --runtime nvidia --privileged -p HOST_PORT:8000 -v /PATH_TO_CLONED_REPO_ROOT/:/repo neuralet/jetson-nano:applications-smart-distancing

for further information please follow the instruction in: https://github.com/neuralet/neuralet/blob/master/applications/smart-distancing/README.md let me now if it was helpful :)

grdsvt commented 4 years ago

Hi, thanks alpha-carinae29 for your reply. I downloaded docker repository with "docker pull neuralet/jetson-nano:applications-smart-distancing" and then I tried to run docker with " docker run -it --runtime nvidia --privileged -p HOST_PORT:8000 -v /PATH_TO_CLONED_REPO_ROOT/:/repo neuralet/jetson-nano:applications-smart-distancing but I received an error message : "python3: can't open file 'neuralet-distancing.py': [Errno 2] No such file or directory" I selected the folder path where I see neuralet-distancing.py (if it is the main python file to execute) but I received this message. I have no experience with docker, sorry. Can sameone help me?

Thanks

mhejrati commented 4 years ago

@grdsvt you need to replace "/PATH_TO_CLONED_REPO_ROOT/" with full absolute address of your locally cloned repo. e.g. /users/grdvst/projects/neuralet/ Right now the container can't see the files because the mounted volume is not valid. In the future releases, we are planning to move all the code inside the container to make this simpler and not require you to mount code from outside.

grdsvt commented 4 years ago

Thanks. So the code in this repository is not complete? When do you think to insert all code inside the container? Why haven't insert all code in github without docker? Thanks for your reply.

alpha-carinae29 commented 4 years ago

Also please replace "HOST_PORT" with a open port in your local machine. e.g. 8000 and after you ran the docker, open your browser and browse localhost:8000

grdsvt commented 4 years ago

Thanks alpha-carinae29. I replaced "HOST_PORT" with 8000 but I did not open browser and not browe localhost:8000. But I don't understand...... The repository for jetson is not complete yet, right?

Thanks

alpha-carinae29 commented 4 years ago

So the code in this repository is not complete?

The code is complete in the repository, However the codes are not contain in docker container and as @mhejrati mentioned, for now you should mount the source code to the container.

alpha-carinae29 commented 4 years ago

The repository for jetson is not complete yet, right?

No, it is complete for Jetson and we have a step by step guide for Jetson Nano device here. please follow this instruction and let us know if you had any problem. thanks

grdsvt commented 4 years ago

Hi @alpha-carinae29. I executed :

cd neuralet/applications/smart-distancing/

1) Download TensorRT engine file built with JetPack 4.3:

./download_jetson_trt.sh (OK)

2) Build Docker image

docker build -f jetson-nano.Dockerfile -t "neuralet/jetson-nano:applications-smart-distancing"

"docker build" requires exactly 1 argument. See 'docker build --help'. Usage: docker build [OPTIONS] PATH | URL | -

3) Run Docker container:

docker run -it --runtime nvidia --privileged -p HOST_PORT:8000 -v /PATH_TO_CLONED_REPO_ROOT/:/repo neuralet/jetson-nano:applications-smart-distancing

I am in the neuralet/applications/smart-distancing/ path but I received this message on 2 step. Sorry but I am new on docker......

alpha-carinae29 commented 4 years ago

It seems you forgot the . at the end of docker build command. So please try:

docker build -f jetson-nano.Dockerfile -t "neuralet/jetson-nano:applications-smart-distancing" .
grdsvt commented 4 years ago

I solved. The problem was that I had jetpack 4.4 and it was impossible importing all dipendencies. I downgrade to jetpack 4.3 and now it is working. Thanks