Connect MLflow, Weights & Biases, TensorBoard, and Hugging Face — all from your lab.
Track runs, log parameters, metrics, artifacts. View UI inline or connect to your server.
Copy these into the Code Lab to start logging experiments
Install: pip install mlflow
import mlflow import mlflow.sklearn from sklearn.linear_model import LogisticRegression # Set tracking URI (your MLflow server) mlflow.set_tracking_uri("http://localhost:5000") mlflow.set_experiment("remesys-classification") with mlflow.start_run(): # Log parameters mlflow.log_param("model", "LogisticRegression") mlflow.log_param("C", 1.0) # Train model model = LogisticRegression(C=1.0) model.fit(X_train, y_train) acc = model.score(X_test, y_test) # Log metrics & model mlflow.log_metric("accuracy", acc) mlflow.sklearn.log_model(model, "model") print(f"Run logged! Accuracy: {acc:.4f}")
↑ Sample data. Connect your MLflow server above to see real experiments.
Real-time metrics, model comparison, hyperparameter sweeps, artifact logging.
Simulated — connects to real W&B when configured
Install: pip install wandb
import wandb from sklearn.ensemble import RandomForestClassifier # Initialize W&B run wandb.init( project="remesys-ai-lab", config={"n_estimators": 100, "max_depth": 5} ) # Train with logging for epoch in range(10): acc = train_one_epoch(model, epoch) loss = compute_loss(model) wandb.log({"accuracy": acc, "loss": loss, "epoch": epoch}) # Save model artifact artifact = wandb.Artifact("model", type="model") artifact.add_file("model.pkl") wandb.log_artifact(artifact) wandb.finish()
Visualize training metrics, computation graphs, embeddings, and model architecture.
pip install tensorboard tensorflow
import tensorflow as tf # Set up summary writer log_dir = "logs/fit/" writer = tf.summary.create_file_writer(log_dir) with writer.as_default(): for step in range(100): tf.summary.scalar("loss", loss_value, step=step) tf.summary.scalar("accuracy", acc_value, step=step)
tensorboard --logdir=logs --host=0.0.0.0 --port=6006
# Then enter URL above and click Embed
Browse pre-trained models, datasets, and Spaces. Embed Spaces directly in your lab.
Text classification & NLP fundamentals
Image classification baseline model
Lightweight text generation for experiments
Speech recognition — transcribe audio
Sentence embeddings for semantic search
Small language model, good for demos
Version your datasets and ML models alongside code. Works with Git.
pip install dvc dvc-s3 # or dvc-gdrive, dvc-ssh
git init
dvc init
dvc add data/train.csv
git add data/train.csv.dvc .gitignore
git commit -m "Track dataset v1"
dvc push # push to remote storage
dvc run -n train \
-d data/train.csv -d src/train.py \
-o models/model.pkl \
python src/train.py
dvc repro # reproduce pipeline
Automatic experiment logging, model comparison, custom panels.