AI/ML Labs

root@homelab:~$ cd ai-labs/
AI / ML Experiment Tracker
Total Experiments
18
Completed
7
In Progress
4
Models Trained
11
Categories
Model Training
3/743%
Dataset Prep
3/475%
Tool Configs
2/450%
Deployments
1/333%
Tool Stack
PyTorch
2.3.0
Ollama
0.1.48
scikit-learn
1.5.0
HuggingFace
4.41.2
CUDA
12.4
MLflow
2.13.0
MNIST Digit Classifier
COMPLETED
Model TrainingPyTorchacc: 97.4%2026-07-12

3-layer CNN, 10 epochs, Adam optimizer. Achieved 97.4% on test set. Next: try ResNet architecture.

CNNclassificationPyTorch
Llama 3.1 8B Local Inference
ACTIVE
DeploymentsOllama2026-07-10

Running on RTX 3090 via Ollama. ~18 tokens/sec. Testing with LangChain for RAG pipeline.

LLMinferenceOllama
Network Traffic Anomaly Detection
IN PROGRESS
Model Trainingscikit-learnacc: 82.1%2026-07-09

Isolation Forest on CICIDS2017 dataset. Need to tune contamination parameter. F1 score low on rare attacks.

anomalynetworkingsklearn
Stable Diffusion Fine-tune
IN PROGRESS
Model TrainingDiffusers2026-07-08

DreamBooth on custom dataset. 500 steps done, 1500 remaining. GPU memory at 22GB.

diffusionfine-tuneHuggingFace
CICIDS2017 Dataset Prep
COMPLETED
Dataset PrepPandas2026-07-06

Cleaned 2.8M rows, handled class imbalance with SMOTE, normalized features. Saved as parquet.

datasetpreprocessingpandas
Whisper Speech-to-Text
COMPLETED
Tool ConfigsOpenAI Whisperacc: 94.2%2026-07-04

whisper-large-v3 running locally. WER 5.8% on custom audio. Integrated with FastAPI endpoint.

ASRspeechWhisper