popov/gogo2
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

reduce logging. actual training

Browse Source
This commit is contained in:
Dobromir Popov
2025-10-31 03:52:41 +02:00
parent 6ac324289c
commit 1bf41e06a8
9 changed files with 1700 additions and 74 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
_dev/dev_notes.md
View File

@@ -112,4 +112,12 @@ do we evaluate and reward/punish each model at each reference?
in our realtime Reinforcement learning training how do we calculate the score (reward/penalty?)
Let's use the mean squared difference between the prediction and the empirical outcome. We should do a training run at each inference which will use the last inference's prediction and the current price as outcome. do that up to 6 last predictions and calculating accuracity separately to have a better picture of the ability to predict couple of timeframes in the future. additionally to the frequent inference every 1 or 5s (i forgot the curent CNN rate) do an inference at each new timeframe interval. model should get the full data (multi timeframe - ETH (main) 1s 1m 1h 1d and 1m for BTC, SPX and one more) but should also know on what timeframe it is predicting. we predict only on the main symbol - so in 4 timeframes. bur on every hour we will do 4 inferences - one for each timeframe
Let's use the mean squared difference between the prediction and the empirical outcome. We should do a training run at each inference which will use the last inference's prediction and the current price as outcome. do that up to 6 last predictions and calculating accuracity separately to have a better picture of the ability to predict couple of timeframes in the future. additionally to the frequent inference every 1 or 5s (i forgot the curent CNN rate) do an inference at each new timeframe interval. model should get the full data (multi timeframe - ETH (main) 1s 1m 1h 1d and 1m for BTC, SPX and one more) but should also know on what timeframe it is predicting. we predict only on the main symbol - so in 4 timeframes. bur on every hour we will do 4 inferences - one for each timeframe
----------
can we check the "live inference" mode now. it should to a realtime inference/training each second (as much barches as can pass in 1s) and prediction should be next candle - training will be retrospective with 1 candle delay (called each s, m, h and d for the previous candle when we know the result)
calculate the angle between each 2 candles features and train to predict those (top- top; open -open, etc.)