SirStone/SirRoboGarage
1
0
Fork 0
Code Issues 28 Pull Requests Actions Projects Releases Wiki Activity

feat(PPO_Bot): command abstraction layer — goto/aimTo controllers (#24)

Browse Source 
- Add gotoTick/aimToTick controller functions (#25)
- Update network dims: actor 5→6, state 42→44 (#26)
- Rewrite mapActions for 6-dim command space (#27)
- Delete stale weight files (shape mismatch)
- Fix existing tests for new signatures

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Davide Cappellini
2026-08-17 19:08:31 +02:00
parent 56e0b306c9
commit cdde60d79f
39 changed files with 490 additions and 78 deletions
Download Patch File Download Diff File Expand all files Collapse all files
PPO_Bot/tests/test_training.nim
+4 -4
View File
@@ -29,7 +29,7 @@ block testBuffer:
var buf = initTrajectoryBuffer()
check buf.len == 0, "empty buffer len == 0"
let t1 = Transition(state: zeros[float32](42), action: zeros[float32](5),
let t1 = Transition(state: zeros[float32](STATE_DIM), action: zeros[float32](ACTION_DIM),
logProb: -1.0'f32, reward: 0.5'f32, value: 0.3'f32)
buf.add(t1)
buf.add(t1)
@@ -80,14 +80,14 @@ block testPpoUpdate:
var buf = initTrajectoryBuffer()
for _ in 0..<10:
let s = randomNormalTensor[float32](42)
let a = randomNormalTensor[float32](5)
let s = randomNormalTensor[float32](STATE_DIM)
let a = randomNormalTensor[float32](ACTION_DIM)
let lp = ac.computeLogProb(s, a)
let v = ac.criticForward(s)
buf.add(Transition(state: s, action: a, logProb: lp, reward: 0.1'f32, value: v))
var adam: ACAdamStates
ppoUpdate(ac, buf, lastValue = 0.0'f32, adamStates = adam, epochs = 2, miniBatchSize = 5)
discard ppoUpdate(ac, buf, lastValue = 0.0'f32, adamStates = adam, epochs = 2, miniBatchSize = 5)
# Weights should have changed — compare flattened
let n = ac.actor.w1.shape[0] * ac.actor.w1.shape[1]