🧠 Test 18: Multi-Architecture Adaptation Benchmark
Networks: Dense, Conv2D, RNN, LSTM, Attention | Depths: 3, 5, 9 | 5 Training Modes | 10s per test
▶️ Run Full Benchmark (75 tests)
0 / 75
Ready
0
Completed
-
Best Mode
-
Best Arch
-
Avg Accuracy
0:00
Elapsed
-
Current Acc
🎯 CHASE
-
Agent
Target (Chase)
Target (Avoid)
Trail
📊 Results Table
Config
NormalBP
Step+BP
Tween
TweenChain
StepTweenChain
All
Dense
Conv2D
RNN
LSTM
Attn
All
NormalBP
StepBP
Tween
TweenChain
StepTweenChain
📈 Accuracy Over Time
Each line shows per-second accuracy for a test configuration
CHASE (0-3s)
AVOID (3-7s)
CHASE (7-10s)
🚀 Winner
-
⚠️ Worst
-
⚙️ Scoring Algorithm
Balanced (T×S×C)
Throughput × Stability
Consistency First
Throughput First
Formula:
Score = (T × S × C) / 100000
S
=Stability |
T
=Throughput |
C
=Consistency
📊 Mode Comparison
Mode
Stability
Tput
Cons
= Score
📝 Log