Hello, I'm a Msc student in Computer Science Engineering and I was looking forward to using the spotRiverGUI for a university optional project regarding Hyper-parameter Tuning in Streaming Machine Learning, however I ran into some problems.
About my resources, I'm using a Windows 10 laptop with processor Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz and a 4GB system memory (2 x 2GB Elpida 1600MHz), plus an integrated graphics processor Intel(R) HD Graphics 520.
At the beginning, I cloned your spotGUI github repo and installed the river, spotRiver, build and spotRiverGUI packages inside a new anaconda virtual environment with Python=3.11.8, and then i executed the file spotRiverGUI.py located inside the spotRiverGUI folder of the repo.
I ran into both graphical and execution problems; I attach some screenshots for reference of the issues below.
First of all, when the GUI is launched, the GUI window is already bad-framed and hiding some settings of the Gui panel: for example, the "UI Scaling" tab is cut in half, whereas other tabs (like the one to select the dataset to use) are non-visible. Trying to manually resize the window does not solve anything.
Moreover, if for example you select "Binary classification" from the "Select task" tab, the graphical bugs get worse, since some columns of the window start to overlap (see attached picture).
Still ignoring these graphical bugs, I tried to check if the spotRiverGUI application worked, so I clicked on "Run" to start the experiment without changing any of the other tabs (so leaving all the settings to their default value). The problem is that the execution started, as I could see from the many information printed on the console, but it never ended: results are never printed on the console nor saved in the "results" folder which is automatically generated.
The problem is that no error comes out in the console, which makes it hard to understand what is the cause.
I also waited 30 minutes for some experiments to finish (which never happened), hoping that maybe the problem was in my laptop resources, and I also tried easier running options (among the ones still visible in the Gui frame), like removing noise or decreasing dataset size, but the execution never reached an end.
Regarding the last line printed on the console during the stalled execution, sometimes it was "Experiment started", other times it was something else (see attached pictures).
At the beginning of my experiments I had also tested the older released version of the Gui, the spotRiverGUI_v01.py, which did not have any graphical bug (all the frame and options were clearly visible) but still did not reach the end of execution and results.
Feel free to contact me at anytime for clarifications about these issues I reported and for further inquiries.
Hello, I'm a Msc student in Computer Science Engineering and I was looking forward to using the spotRiverGUI for a university optional project regarding Hyper-parameter Tuning in Streaming Machine Learning, however I ran into some problems.
About my resources, I'm using a Windows 10 laptop with processor Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz and a 4GB system memory (2 x 2GB Elpida 1600MHz), plus an integrated graphics processor Intel(R) HD Graphics 520.
At the beginning, I cloned your spotGUI github repo and installed the river, spotRiver, build and spotRiverGUI packages inside a new anaconda virtual environment with Python=3.11.8, and then i executed the file spotRiverGUI.py located inside the spotRiverGUI folder of the repo.
I ran into both graphical and execution problems; I attach some screenshots for reference of the issues below.
First of all, when the GUI is launched, the GUI window is already bad-framed and hiding some settings of the Gui panel: for example, the "UI Scaling" tab is cut in half, whereas other tabs (like the one to select the dataset to use) are non-visible. Trying to manually resize the window does not solve anything.
Moreover, if for example you select "Binary classification" from the "Select task" tab, the graphical bugs get worse, since some columns of the window start to overlap (see attached picture).
Still ignoring these graphical bugs, I tried to check if the spotRiverGUI application worked, so I clicked on "Run" to start the experiment without changing any of the other tabs (so leaving all the settings to their default value). The problem is that the execution started, as I could see from the many information printed on the console, but it never ended: results are never printed on the console nor saved in the "results" folder which is automatically generated.
The problem is that no error comes out in the console, which makes it hard to understand what is the cause.
I also waited 30 minutes for some experiments to finish (which never happened), hoping that maybe the problem was in my laptop resources, and I also tried easier running options (among the ones still visible in the Gui frame), like removing noise or decreasing dataset size, but the execution never reached an end.
Regarding the last line printed on the console during the stalled execution, sometimes it was "Experiment started", other times it was something else (see attached pictures).
At the beginning of my experiments I had also tested the older released version of the Gui, the spotRiverGUI_v01.py, which did not have any graphical bug (all the frame and options were clearly visible) but still did not reach the end of execution and results.
Feel free to contact me at anytime for clarifications about these issues I reported and for further inquiries.