v0.1.81 - Update Platform API contract (#102)

🌟 Summary

Version 0.1.81 adds dataset-license filtering to Platform Explore search and makes local cloud-training uploads quicker to prepare and fail faster when Platform access is unavailable.

📊 Key Changes

• 🔎 Explore search: The Platform API and Python SDK now support an optional, comma-separated license filter for datasets, available in both synchronous and asynchronous search methods.
• 📦 Faster dataset packaging: ul cloud train now stores already-compressed image formats—such as JPEG, PNG, and WebP—without trying to compress them again. Other selected file types remain compressed.
• ⚡ Earlier Platform checks: Cloud training now looks up or creates the project before packaging a local dataset, so issues such as an invalid API key or unreachable server are reported sooner.
• 🔄 SDK update: The OpenAPI contract and Python SDK are synchronized at version 0.1.81.

🎯 Purpose & Impact

• Find datasets more easily by filtering Explore results by license.
• Reduce CPU time spent preparing image archives; archives may be slightly larger because compressed images are stored as-is.
• Avoid waiting for a large dataset to package before discovering a Platform connection or authorization problem. A project may be created before local dataset validation completes.
• No model architecture or training behavior changes are included in this release.

What's Changed

• Check the API key before packaging datasets in ul cloud train by @JaviChulvi in #100
• Faster dataset packaging for ul cloud train: copy JPEG/PNG into the ZIP as is by @JaviChulvi in #101
• Update Platform API contract by @UltralyticsAssistant in #102

Full Changelog: v0.1.80...v0.1.81