I am glad that my project proposal is selected this year of GSoC. The project supplement my previous experience in the model optimization field. I am excited to start exploring and implementing latest tools and techniques for this project with JdeRobot community.
During our first official meetings, we started with a little introduction about ourselves. The mentors then presented the core idea behind the main tools of JdeRobot and what could we get from this project. My project would be challenging because many aspects or results are unknown until we do the experiments, so it can be categorized into research and development projects. I also prepared questions regarding the GPU server, evaluation (accuracy and inference), and the contribution approach to the organization, which the mentors answered clearly. We agreed on a few points to focus on during this week, which I present next.
- Familiarize with relevant tools and explore documentation and tutorials.
- Complete installation of Behavior Metrics and DeepLearningStudio.
- Explore the repositories and identify potential bugs.
- Prepare a template for the blog post.
- Write the blog for the first week of community bonding.
I have already worked with Behavior Metrics and DeepLearningStudio , which are the main focus and will be developed further in this project. I have previous experience in contributing additional models in these repositories with PR#40 and PR#335.
I followed the Behavior Metrics tutorial to familiarize myself with the tool features, create a new brain, and train the model with DeepLearningStudio.
The installation of tools is well supported by Behavior Metrics’ installation guide and DeepLearningStudio’s readme instruction, which I use for local installation on my machine. I also identified potential bugs, which I will discuss in the next meeting.
I adapted this blog template, which is compatible and suitable for GSoC project reporting. With this post, I conclude my first week of GSoC’s community bonding period.