AnswerDotAI / AnswerDotAI/fastkmeans
total_time is strange
- Dominant language
- Python
- Stars
- 104
- Forks
- 8
- PR merge metrics
- No merged PRs in 30d
Description
From the code, we can see the `total_time` is always double of the `iteration_time`, is this correct?
https://github.com/AnswerDotAI/fastkmeans/blob/b5173ffad0d37423b282cd60cfcf14c1588235d4/fastkmeans/kmeans.py#L160
From this code, we have these equations:
```
iteration_time = now - iteration_start_time
total_time = now - iteration_start_time + iteration_time = 2 * iteration_time
```
Here is the log of experiment we runned:
```
Iteration 2/100 took 18.6287s, total time: 37.2574s, shift: 49088.000000
Iteration 3/100 took 18.8487s, total time: 37.6975s, shift: 30736.000000
Iteration 4/100 took 18.9899s, total time: 37.9798s, shift: 23024.000000
Iteration 5/100 took 19.1189s, total time: 38.2379s, shift: 19040.000000
Iteration 6/100 took 19.2889s, total time: 38.5779s, shift: 16864.000000
Iteration 7/100 took 19.5249s, total time: 39.0499s, shift: 15688.000000
Iteration 8/100 took 19.8129s, total time: 39.6259s, shift: 14592.000000
Iteration 9/100 took 19.9709s, total time: 39.9419s, shift: 13800.000000
Iteration 10/100 took 19.9869s, total time: 39.9739s, shift: 13280.000000
Iteration 11/100 took 19.9949s, total time: 39.9899s, shift: 13080.000000
Iteration 12/100 took 20.0860s, total time: 40.1719s, shift: 12600.000000
Iteration 13/100 took 20.2500s, total time: 40.4999s, shift: 12280.000000
Iteration 14/100 took 20.1469s, total time: 40.2939s, shift: 12192.000000
Iteration 15/100 took 20.0179s, total time: 40.0359s, shift: 12048.000000
Iteration 16/100 took 19.9909s, total time: 39.9819s, shift: 11784.000000
Iteration 17/100 took 20.0150s, total time: 40.0299s, shift: 11872.000000
```
Contributor guide
No contributing guide indexed for this repository
Research direction
Start in fastkmeans/kmeans.py around line 160 and inspect how iteration_start_time and total-time values are calculated for the log message. Reproduce or review the reported iteration output, then verify that total time represents the intended elapsed duration rather than twice the iteration time.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 55/100