popsize more than 500 crashes
When I set my popsize on Neat over 500 it crashes when I run neat.evolve().
I'm using the latest version of Node.js and Carrot. The memory with popsize as 500 is only ~100megs, so that doesn't seem to be the issue.
(node:19260) UnhandledPromiseRejectionWarning: TypeError: Cannot set property 'score' of undefined
at Neat.evolve (D:\Git\DayTradingBot\DayTradingBot\node_modules\@liquid-carrot\carrot\src\neat.js:140:32)
at endEvaluation (D:\Git\DayTradingBot\DayTradingBot\app.js:480:10)
at learnData (D:\Git\DayTradingBot\DayTradingBot\app.js:325:13)
at downloadBtcData (D:\Git\DayTradingBot\DayTradingBot\app.js:229:5)
at processTicksAndRejections (internal/process/task_queues.js:86:5)
(node:19260) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). (rejection id: 1)
(node:19260) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.
(node:12304) UnhandledPromiseRejectionWarning: TypeError: Cannot read property 'toJSON' of undefined
at Neat.export (D:\Git\DayTradingBot\DayTradingBot\node_modules\@liquid-carrot\carrot\src\neat.js:346:24)
at learnData (D:\Git\DayTradingBot\DayTradingBot\app.js:342:60)
at downloadBtcData (D:\Git\DayTradingBot\DayTradingBot\app.js:229:5)
at processTicksAndRejections (internal/process/task_queues.js:86:5)
(node:12304) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). (rejection id: 11)```
@luiscarbonell You should be able to repeat it easily with the code I gave you.
In my latest changes, it no longer crashes but the generation count doesn't increase, so odd...
@dan-ryan We've been seeing a lot of these silent failures also.
We're cleaning some of them out with unit testing.
We'll use Mocha and Chai for now.
@christianechevarria Ran into ritewayt by Eric Elliot - but we haven't done extensive work with it yet...
So far, that we know of, we've cleaned out 7 silent bugs.
I'll close this for now since we haven't been able to replicate on our end.
Great. I'll test it with the latest code later tonight.
I'm still having generation count issues. Sometimes it works a little higher than a population size of 500 but usually, 2000+ doesn't.
Will have to create a test where we can increase popsize to Number.MAX_SAFE_INTEGER or Number.MAX_VALUE - ideally.