While earlier DSX versions had basic collaboration, 1.5.0 introduced real-time co-editing indicators (similar to Google Docs) for Jupyter notebooks. Additionally, it added granular version control, allowing data scientists to revert to any prior cell state without leaving the browser.
: The mandatory internet check-in period for ownership verification has been increased from 14 days to Enhanced Connectivity backup server dsx 1.5.0
| Metric | DSX 1.4.3 | DSX 1.5.0 | Improvement | |--------|-----------|-----------|--------------| | Notebook startup time | 23 seconds | 11 seconds | 52% faster | | Spark job submission overhead | 8 seconds | 3 seconds | 62% faster | | Data Refinery (1GB CSV) | 4.2 minutes | 2.1 minutes | 50% faster | | Concurrent users per node | 15 | 28 | 86% higher density | While earlier DSX versions had basic collaboration, 1
# Step 1: Load data from IBM Cloud Object Storage import ibm_boto3 cos = ibm_boto3.client('s3') obj = cos.get_object(Bucket='my-bucket', Key='sales.csv') df = pd.read_csv(obj['Body']) This upgrade provided: from dsx import Model model_asset
DSX 1.5.0 shipped with integrated Apache Spark 2.3 (with backward compatibility for Spark 2.2). This upgrade provided:
from dsx import Model model_asset = Model.save(model, name="sales_forecast_v1", project_id="your-project-id") print(f"Saved with ID: model_asset.id")