← Back to Aria Home

📊 QAI Training Dashboard

LoRA fine-tuning, model evaluation, GPU monitoring, and autonomous training cycles

Demo Mode — Simulated Data
🎯Training Accuracy
92.4%
Best model — cycle #47
🔄Training Cycles
47
Autonomous cycles completed
📦Datasets
12
Active datasets across 3 categories
📈Accuracy Over Cycles
⏱️Training Duration (min)
🖥️System Resources
CPU34%
Memory61%
GPU (VRAM)78%
Disk45%
🤖Active Providers
Provider Status Latency
Azure OpenAI Connected 142ms
OpenAI Available —
LMStudio Offline —
Local Ready <1ms
📋Recent Training Runs
Cycle Dataset Epochs Accuracy Duration Status
#47 chat/aria_personality_v3 100 92.4% 28m Promoted
#46 quantum/circuit_patterns 50 87.1% 15m Complete
#45 chat/multi_turn_dialogs 100 89.3% 32m Complete
#44 chat/aria_commands_v2 200 91.7% 54m Promoted
#43 quantum/hybrid_ml 50 72.8% 12m Below threshold
#42 chat/personality_v2 100 88.5% 26m Complete
#41 vision/expressions 25 0.0% 2m Failed
🎯Autotrain
Idle
Next run: 02:00 UTC
Last: 12 jobs, 11 passed
⚛️Quantum Autorun
Running
Job: qiskit_vqc_sim
Progress: 67%
🔄Autonomous Training
Active
Cycle #47 complete
Next cycle in 18m