Feature flags of oxirs-embed crate // Lib.rs

Cargo Features

[dependencies]
oxirs-embed = { version = "0.1.0", default-features = false, features = ["basic-models", "advanced-models", "neural-models", "hole", "rotate", "conve", "tucker", "quatd", "multi-modal", "gpu-acceleration", "api-server", "graphql", "gpu", "cuda", "opencl", "rocm", "metal", "simd", "parallel"] }

default = basic-models

The basic-models feature is set by default whenever oxirs-embed is added without default-features = false somewhere in the dependency tree.

basic-models default = complex, distmult, hole, transe

advanced-models = conve, quatd, rotate, tucker

neural-models conve?

transe complex distmult hole basic-models

rotate advanced-models?

conve advanced-models? = neural-models

Affects models::conve…

tucker advanced-models?

Affects models::tucker…

quatd advanced-models?

Affects models::quatd…

custom-models ontology-aware multi-modal

gpu-acceleration = cuda, gpu

api-server

Enables axum ^0.7, tower, and tower-http

axum:
Web server dependencies (for API server feature)

Affects embeddings::embed_single, embeddings::embed_batch, models::list_models, models::get_model_info, models::get_model_health, models::load_model, models::unload_model, predictions::predict, scoring::score_triple, system::system_health, system::system_stats, system::cache_stats, system::clear_cache, api::start_server, routes::create_router, oxirs-embed::api, graphql::graphql_handler, graphql::graphiql…

graphql

GraphQL API support

Affects graphql::ModelInfo, graphql::SystemHealth, graphql::CacheStatistics, graphql::PredictionResult, graphql::Query, graphql::create_schema, graphql::graphql_handler, graphql::graphiql…

gpu gpu-acceleration?

High-performance acceleration features

Enables gpu of scirs2-linalg and scirs2-neural

Affects gpu::GpuArray, gpu::GpuContext, gpu::GpuError, gpu::GpuMemoryPool, gpu::AdaptiveBatchConfig, gpu::GpuEmbeddingAccelerator, gpu::GpuPerformanceStats, gpu::GpuPerformanceReport…

cuda gpu-acceleration?