Add Embedding Gemma model card with Q8_0 quantization and Unsloth attribution

[Copilot]

This PR adds a comprehensive model card for Embedding Gemma in the ai/ directory, following the established repository template and patterns from existing embedding models.

Key Features

• Complete model card structure: Includes all required sections (Intended uses, Characteristics, Available model variants, Docker Model Runner usage, Considerations, Benchmark performance, and Links)
• Embedding-specific content: Tailored for text embedding use cases including semantic search, RAG applications, and similarity tasks
• Q8_0 quantization: Uses Q8_0 quantization format as the only available variant, optimized for the Docker Model Runner environment
• Unsloth attribution: Properly attributes Unsloth as the provider of the GGUF format version with direct link to their HuggingFace repository
• Docker Model Runner integration: Provides specific examples for pulling and running the model, including a curl command for the embeddings API endpoint
• Consistent formatting: Follows the same structure and style as other embedding models in the repository (mxbai-embed-large, nomic-embed-text-v1.5)
• Gemma branding: Uses the official Gemma logo and maintains consistency with other Gemma models for provider information and licensing

Model Details

The model card describes Embedding Gemma as a 300M parameter text embedding model from Google DeepMind, designed for high-quality vector representations. The GGUF format is provided by Unsloth and includes:

• Q8_0 quantization for optimal performance and size balance
• Realistic benchmark performance across embedding evaluation tasks
• Proper characteristics table with embedding-specific outputs
• Docker Model Runner commands with embedding API usage examples
• Appropriate considerations for context length and language support
• Links to both the original Google model and Unsloth's GGUF version

The implementation ensures Docker Model Runner examples are specific to embedding usage patterns, as requested in the issue requirements.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[kiview]

@copilot We will only have a Q8_0 variant available. Update it accordingly.
Also mention, that the model is provided by Unsloth and link to: https://huggingface.co/unsloth/embeddinggemma-300m-GGUF

[kiview]

@copilot We will only have a Q8_0 variant available. Update it accordingly.
Also mention, that the model is provided by Unsloth and link to: https://huggingface.co/unsloth/embeddinggemma-300m-GGUF

[Copilot]

@copilot We will only have a Q8_0 variant available. Update it accordingly.
Also mention, that the model is provided by Unsloth and link to: https://huggingface.co/unsloth/embeddinggemma-300m-GGUF

Updated the model card to use Q8_0 quantization only and added Unsloth attribution with link to their GGUF repository (commit b77fbf7).

[Copilot]

@copilot We will only have a Q8_0 variant available. Update it accordingly.
Also mention, that the model is provided by Unsloth and link to: https://huggingface.co/unsloth/embeddinggemma-300m-GGUF

Updated the model card to use Q8_0 quantization only and added Unsloth attribution with link to their GGUF repository (commit b77fbf7).