31 lines
1003 B
Rust
31 lines
1003 B
Rust
//! RTX Inference Stack - High-performance model serving
|
|
//!
|
|
//! Production-ready inference with:
|
|
//! - Optimized inference engine (<10ms latency)
|
|
//! - HTTP/gRPC/WebSocket serving APIs
|
|
//! - Real-time streaming capabilities
|
|
//! - Model compression and quantization
|
|
|
|
#![forbid(unsafe_code)]
|
|
|
|
// Core
|
|
pub use rtx_runtime as runtime;
|
|
pub use rtx_tensor as tensor;
|
|
|
|
// Inference - Note: naming conflict resolved with module name
|
|
pub use rtx_compress as compress;
|
|
pub use rtx_inference as inference_engine;
|
|
pub use rtx_serving_api as serving;
|
|
pub use rtx_streaming as streaming;
|
|
|
|
/// Inference essentials
|
|
pub mod prelude {
|
|
pub use rtx_tensor::{Device, Tensor};
|
|
// Use the correct exports from rtx_inference
|
|
pub use rtx_inference::{InferenceEngine, OptimizationStats};
|
|
// Use the correct exports from rtx_serving_api
|
|
pub use rtx_serving_api::{ServerConfig, ServingServer};
|
|
// Use the correct exports from rtx_streaming
|
|
pub use rtx_streaming::{StreamingConfig, StreamingServer};
|
|
}
|