oracle_data_science_examples

Example notebooks in using OCI Data Science, from deploying GANs to small examples in how to work with autonomous database.

1. Running and deploying DeepFace in Face Recognition and comparison (deepface_oda_v3.ipynb)

• Deploy 3 different Deep Face models in one model deployment. Invoke as REST API
• Create two images in one json payload

2. Speech to Text using Whisper, Generate Embeddings and apply indexing using FAISS (whisper_embeddings_faiss.ipynb)

• Loop through multiple recordings (mp3 files) and transcribe using Whisper
• Generate Embeddings using BERT on transcribed texts
• Apply FAISS for indexing
• Create one Job to automate newly incoming recordings

3. Speech to Text using Whisper, Generate and Store Embeddings using Qdrant Vector Database (speech_to_text_qdrant_vector_db.ipynb)

• Loop through multiple recordings (mp3 files) and transcribe using Whisper
• Generate and store embeddings in Qdrant Vector Database using Qdrant Python SDK
• Qdrant Vector Database running on seperate instance in Oracle Cloud

4. Different examples in OCI Data Science (Custom_Conda_Run_Jobs_Examples.ipynb)

• Create and publish a custom conda environment
• Authenticate using config file
• Create connection with an Autonomous Database
• Push PD dataframe to the Autonomous Database as a new table (and append)
• Create and run a Data Science Jobs, using a .py file stored in local (Data Science) directory, with environment variables referring to Object Storage buckets

5. Delete log groups (delete_logs.ipynb)

• Batch delete logs inside a Log Group

6. Batch delete all items in a Compartment (batch_delete_projects.ipynb)

• Batch delete all model deployments
• Batch delete all models in model catalog
• Batch delete all notebook sessions
• Batch delete all projects

7. CTGAN (deploying_gan.ipynb)

• Use CTGan as a deployed model on OCI to generate tabular synthetic data automatically on input .csv files, output is new rows of .csv files

8. CTGAN in notebook session (generate_synthetic_data.ipynb)

• Use CTGan to generate synthetic .csv file. From object storage to object storage
• Use CTGan to generate synthetic data from autonomous database table, to a new, synthetic database table

9. Video Analytics on Soccer Video (demo_spl.ipynb)

• Use open source to track scoccer players, the ball, and referee in a video
• Use a custom trained AI Vision model (recognizing logos near the soccer pitch) to extract the logos from the video. E.g., how often was "xx" logo in screen?

10. Automating Forecasting using Prophet (automating_forecasting.ipynb)

• Notebook to deploy and automate forecasting
• Read in .csv from bucket, use Prophet to forecast next 10 minutes
• Calculate difference between forecasted future and history (e.g., expected to increase with 10%)
• Push results to autonomous database