Training Dashboard
DevOS — Model Training Details
RAG Pipeline
Vector Store Qdrant
Embedding bge-m3
Data Sources
Dev workflowsEnv configsCI/CD docs
Fine-tuning Pipeline
Method LoRA (r=16, alpha=32)
Base Model Qwen2.5-14B
Adapter Size ~48MB
Deployment Status
Target Node mac-studio
VRAM 12.4 / 128 GB
Inference Speed 42 tok/s
Uptime 99.7%