- Implement SynorStorage class for decentralized storage operations including upload, download, pinning, and CAR file management.
- Create supporting types and models for storage operations such as UploadOptions, Pin, and StorageConfig.
- Implement SynorWallet class for wallet operations including wallet creation, address generation, transaction signing, and balance queries.
- Create supporting types and models for wallet operations such as Wallet, Address, and Transaction.
- Introduce error handling for both storage and wallet operations.
- Created package_graph.json to define project structure and dependencies for synor_compute.
- Added version file to specify the Dart SDK version.
- Generated pubspec.lock to lock dependencies and their versions for the project.
Expands SDK support to 8 additional languages/frameworks:
- Java SDK with Maven/OkHttp/Jackson
- Kotlin SDK with Gradle/Ktor/kotlinx.serialization
- Swift SDK with Swift Package Manager/async-await
- C SDK with CMake/libcurl
- C++ SDK with CMake/Modern C++20
- C# SDK with .NET 8.0/HttpClient
- Ruby SDK with Bundler/Faraday
- Rust SDK with Cargo/reqwest/tokio
All SDKs include:
- Tensor operations (matmul, conv2d, attention)
- LLM inference with streaming support
- Model registry, pricing, and usage APIs
- Builder patterns where idiomatic
- Full type safety
Add cross-platform keychain support using the keyring crate for secure
credential storage with biometric authentication (TouchID on macOS,
Windows Hello on Windows, Secret Service on Linux).
- Add keychain module with enable/disable biometric unlock
- Add Tauri commands for keychain operations
- Add Keychain error variant for proper error handling
- Add Java SDK foundation
- Update milestone docs to reflect 95% completion
Adds comprehensive model management and training capabilities:
synor-compute (Rust):
- ModelRegistry with pre-registered popular models
- LLMs: Llama 3/3.1, Mistral, Mixtral, Qwen, DeepSeek, Phi, CodeLlama
- Embedding: BGE, E5
- Image: Stable Diffusion XL, FLUX.1
- Speech: Whisper
- Multi-modal: LLaVA
- ModelInfo with parameters, format, precision, context length
- Custom model upload and registration
- Model search by name/category
Flutter SDK:
- Model registry APIs: listModels, getModel, searchModels
- Custom model upload with multipart upload
- Training APIs: train(), fineTune(), trainStream()
- TrainingOptions: framework, epochs, batch_size, learning_rate
- TrainingProgress for real-time updates
- ModelUploadOptions and ModelUploadResult
Example code for:
- Listing available models by category
- Fine-tuning pre-trained models
- Uploading custom Python/ONNX models
- Streaming training progress
This enables users to:
1. Use pre-registered models like 'llama-3-70b'
2. Upload their own custom models
3. Fine-tune models on custom datasets
4. Track training progress in real-time
