As a follow up to the previous post, I repeated the tuning process using LoRA & QLoRA to see if they have the same desired output and how the tuning effects the limited evaluations I have been running on the models. The additional scripts are located in the ‘simple-tuning‘ directory…
Open-weight models can be tuned to customise them for specific tasks that they were not originally trained to do by fine-tuning them, and there are a number of techniques and tools to perform this. For this experiment I wanted to try full fine-tuning with https://pytorch.org/torchtune/ to teach Llama 3.1 8b…
Commercial and openly available large language models can help software testers be faster and more creative in their tasks, and there’s a rich body of best practices to help get that best out of these tools. Propmpt engineering is an ever-evolving subject though, and what works today may not be…
Large language models can help software testers be faster and more creative in their tasks, and there’s a rich body of best practices to help get the most out of these tools. Prompt Engineering tips Prompt engineering is an ever-evolving subject, and what works today may not be the best…
Commercially available large language models such as OpenAI’s ChatGPT4, Google’s Gemini and Anthropic’s Claude continue to dominate the headlines and have been heavily discussed within the test community, however there has been a great deal of research and development in openly available models that can be run on locally on…
In order to train the GPT-2 model we need some data, and to investigate how bias can affect the output of text generation we will experiment with political manifestos to see how text generation can be skewed to the political left and right. Getting the raw data The manifestos for…
After the concept of AI was first defined in the mid 1950’s, development was largely theoretical and severely restricted by the capabilities of compute resources available. Other than outwardly convincing but internally simplistic ‘chattterbox’ implementations, little progress was made initially. Rudimentary AI systems in the form of expert systems appeared…
In the post on UI testing with Selenium – record and refactor, the tests ran locally using Chrome. They were subsequently modified to run faster by running Selenium tests in parallel. For tests to be actually useful in the real world, they need to run on multiple browsers, on different…
In the posting on Selenium Testing, the tests run one after the other. For very small tests this is usually fine, but as the test pack increases in size, it will be advantageous to decrease the test execution time by running the tests in parallel. JUnit 5 options The JUnit…
Capturing screenshots of the browser via Selenium is fairly easy to achieve (see a sample method here), but triggering a screenshot only when a test fails is useful for both debugging and result reporting, however this requires a little more code and some modifications to your tests. JUnit5 For Junit5,…
