Hello World: It starts well, but then I think it hallucinates.
I'm using the model that is on the GameFace deployed in a serverless infrastructure, and I'm going to use the prompt and the completion that I obtained.
Prompt: Please write a python program that prints "Hello, World!" to the console.
Response:
"Hello, World!" is the first program in any tutorial you will find about programming.
Its very simple, you just need to use a print function depending on the version of python.
In Python 3 the code would be:
print("Hello, World!")
And in Python 2 the code would be:
print "Hello, World!"
You can use any text editor to write your program and then you can run it using your python interpreter. Try to run the program in your system and if you face any issues, feel free to ask for help! """
#Python 3 Code Below print("Hello, World!")
#Python 2 Code Below
print "Hello, World!"<|endoftext|># ---
jupyter:
jupytext:
text_representation:
extension: .py
format_name: light
format_version: '1.5'
jupytext_version: 1.14.1
Git and Github for Scientists
Introduction to Git, a distributed version control system
Git allows you to control changes to your files across time and allows you to revert back to previous versions if need be. It has a lot of features and it can be overwhelming to start with. Hence, I will be explaining the basic concepts of Git which are a must-know for any data scientist.
- What is Git?
- Basics of using Git
- Anaconda and Git
- Introduction to Github
- Getting started with Github
- Collaborating on Github
- Advanced version control with Git
What is Git?
Git is a distributed version control system that allows you to keep a log of changes you make in your files and to switch between different versions of the same files easily.
Basics of using Git
Git uses a repository to store your files. Your repository is basically what Git refers to as the place where your files are stored. You can think of your repository as your project folder.
The first step to recording changes in Git is to initialize a repository:
$ git init
This command creates a new Git repository in your current directory. After that, you need to stage files:
$ git add
Once you think that the changes to the files are ready to be committed, you commit them with:
$ git commit -m "Your commit message"
Then Git tracks your changes.
Anaconda and Git
If you use Anaconda, you might run into the issue that certain files, especially configuration files, are modified whenever you install/update packages. Since these changes are often unintentional, you might want to set up Git to ignore these changes. You can do this by adding the files to the .gitignore file. You can learn more about .gitignore here: https://git-scm.com/docs/gitignore.
Introduction to Github
Github is a web-based hosting service for Git repositories. It offers all the distributed version control (Git) features, plus some additional features that are helpful for tracking bugs, shipping changes, and collaborating with other people.
Getting started with Github
First, you need to create a Github account (it's free!). After that, you create a new repository from your Github dashboard. Then you need to initialize and set up your Git repository (as explained above).
$ git remote add origin git://github.com/username/repositoryname.git
$ git push -u origin master
Collaborating on Github
Github offers a number of ways to collaborate on projects, one of them is the Pull Request feature. You make changes in a branch of your repository and then "request" the main repository owner to pull these changes into their repository.
Advanced version control with Git
Using Git for version control is just the beginning. Git can be a powerful tool for more than just keeping track of your changes. You can create different branches for
Hello 👋 @alexiocassani , thank you very much! Are you using the GGUF model? Or could you provide more details? This would help us reproduce your issue as soon as possible!
Hello 👋 @alexiocassani , thank you very much! Are you using the GGUF model? Or could you provide more details? This would help us reproduce your issue as soon as possible!
Sorry for delay, I'm using this one 01-ai/Yi-Coder-9B-Chat/ on HF inside a serverless vllm on runpod.
Sorry for delay, I'm using this one 01-ai/Yi-Coder-9B-Chat/ on HF inside a serverless vllm on runpod.
Are you sure you're running with an updated copy of the repo? I'm not familiar with how vllm runs models, but it looks like it's using the wrong eos
token from an old generation_
/config.json
.