https://github.com/graphaware/hume-workshop
https://github.com/graphaware/hume-workshop
Last synced: 7 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/graphaware/hume-workshop
- Owner: graphaware
- Created: 2020-11-05T22:46:28.000Z (about 5 years ago)
- Default Branch: main
- Last Pushed: 2020-11-06T16:39:02.000Z (about 5 years ago)
- Last Synced: 2025-07-04T06:06:47.757Z (7 months ago)
- Language: Shell
- Size: 68.4 KB
- Stars: 0
- Watchers: 6
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# NLP with GraphAware Hume
This repository contains a trial version of Hume for the Nodes 2020 event. Participants will receive credentials through their registration that they have to use
during the instructions described below.
## Requirements
- A laptop or a server with 5-6GB of memory.
- Docker and Docker Compose installed
## Installation
The following procedure will install the necessary containers for running Hume :
- Neo4j 4.1.2
- Hume
- Hume Entity Extraction Service
- Hume Keyword Extraction Service
### Login to the GraphAware Docker registry
```bash
docker login docker.graphaware.com
```
and provide the credentials that have been given to you during the registration.
### Pull the containers
```
docker-compose pull
```
### Start Hume
```
docker-compose up -d
```
it can take some minutes to start depending on your machine.
Once the start-up phase will have finished, head to `http://localhost:8081`: you will see the sign-up page.
### Inspect logs
If something is not working as expected, inspect the logs :
```
docker-compose logs -f
```
### Recommended memory settings
#### Linux hosts
On Linux hosts, by default, a Docker container has no resource constraints and can use as much of a given resource as the kernel scheduler allows.
#### MacOS and MS Windows hosts
On Windows and Mac OS, Docker Engine runs on top of a lightweight virtual machine (VM), which can be handled via Docker Desktop.
Thus, in these environments, only a part of hosts' memory is reserved to this VM; this portion of memory represents the upper limit of the memory that Hume can use.
By default, Docker Desktop is set to use 2 GB runtime memory, allocated from the total available memory on your Mac/Windows machine.
When running Hume with these default RAM settings, if the kernel detects that there is not enough memory to perform important system functions, it throws an Out Of Memory Exception (OOME), and starts killing processes to free up memory.
For instance, in case of intensive data ingestion, Neo4j can easily run out of memory, and then its process can be killed by OOM killer.
In order to avoid this kind of events, we strongly recommend increasing the amount of RAM via Docker Desktop.
To do so, head to `Preferences -> Resources -> Advanced` and increase `Memory` at least to 6 GB.