rlx-models-core 0.2.1

Shared config, weight loading, and compile helpers for RLX model crates
Documentation
// RLX — versatile ML compiler + runtime.
// Copyright (C) 2026 Eugene Hauptmann, Nataliya Kosmyna.
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, version 3.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.

//! Shared infrastructure for RLX model crates: HuggingFace config parsing,
//! safetensors / GGUF weight loading, tier-1 compile profile helpers, and
//! packed GGUF prefill guards ([`flow_bridge::packed_gguf_compile_guard`], etc.).

pub mod arch_registry;
pub mod autoregressive;
pub mod config;
pub mod dataprocessing;
pub mod device_capabilities;
pub mod flow_bridge;
pub mod flow_util;
pub mod gguf_config;
pub mod gguf_resolve;
pub mod gguf_support;
pub mod lm;
pub mod moe_weights;
pub mod vision_ops_ir;
pub mod weight_loader;
pub mod weight_map;
pub mod weight_registry;
pub mod weights;

pub use device_capabilities::{
    STANDARD_DEVICE_NAMES, STANDARD_DEVICES, device_memory_for_moe_offload, is_standard_device,
    validate_sam_device, validate_standard_device,
};

pub use gguf_config::{
    DINOV2_GGUF_ARCHES, EMBED_GGUF_ARCHES, EmbedGgufKind, FLUX_GGUF_ARCHES, GgufMemoryFootprint,
    SAM_GGUF_ARCHES, SAM2_GGUF_ARCHES, SAM3_GGUF_ARCHES, VJEPA2_GGUF_ARCHES, W2V_BERT_GGUF_ARCHES,
    embed_gguf_kind, gguf_memory_footprint, gguf_meta_u32, gguf_runner_hint, is_dinov2_gguf_arch,
    is_embed_gguf_arch, is_flux_gguf_arch, is_sam_gguf_arch, is_sam2_gguf_arch, is_sam3_gguf_arch,
    is_vjepa2_gguf_arch, is_w2v_bert_gguf_arch,
};
pub use gguf_resolve::{
    GgufTensorNameResolver, LlamaFamilyGgufResolver, PassThroughGgufResolver,
    PrefixStripGgufResolver, Qwen35NativeGgufResolver, register_gguf_tensor_resolver,
    resolve_gguf_tensor_name,
};
pub use gguf_support::{
    GgufModelFamily, ResolveWeightsOptions, assert_gguf_family, gguf_architecture_from_path,
    gguf_architecture_str, gguf_f32_bytes_estimate, gguf_family_for_arch,
    gguf_safetensors_only_hint, gguf_split_hint, gguf_split_siblings, gguf_validate_arch,
    list_gguf_files_in_dir, load_gguf_file, resolve_weights_file,
    resolve_weights_file_with_options,
};

pub use autoregressive::{
    KvCacheState, compile_cache_ensure_graph, kv_from_prefill_outputs, past_kv_input_names,
    prefill_cache_key, run_bucketed_kv_decode, run_bucketed_kv_decode_hir,
    split_bucketed_decode_kv, split_decode_logits_kv,
};
pub use config::{BertConfig, NomicBertConfig, NomicVisionConfig};
pub use flow_bridge::{
    apply_compile_profile, compile_graph_encoder, compile_graph_gemma_decode,
    compile_graph_gemma_prefill, compile_graph_legacy, compile_graph_llama32_decode,
    compile_graph_llama32_prefill, compile_graph_qwen3_decode, compile_graph_qwen3_prefill,
    compile_graph_qwen35_decode, compile_graph_qwen35_prefill, compile_graph_sam,
    compile_graph_with_profile, compile_options_for_packed_gguf_prefill,
    compile_options_for_packed_gguf_prefill_with_profile, compile_options_for_profile,
    load_compile_profile, packed_gguf_compile_guard, packed_gguf_execution_device,
    profile_near_weights,
};
pub use flow_util::{
    WeightMapSource, bucket_cache_ensure_built, build_graph, built_from_graph, built_from_hir,
    built_from_hir_with_profile, compile_built, compile_built_cpu, compile_cache_ensure_built,
    compile_graph_encoder_with_params, compile_graph_gemma_decode_with_params,
    compile_graph_gemma_prefill_with_params, compile_graph_profile,
    compile_graph_qwen3_prefill_with_params, compile_graph_qwen35_decode_with_params,
    compile_graph_qwen35_prefill_with_params, compile_graph_sam_with_params,
    compile_graph_with_kv_export_params, graph_from_built, graph_from_hir,
};
pub use gguf_resolve::ensure_builtin_resolvers;
pub use gguf_support::DEFAULT_GGUF_PREFER_SUBSTR;
pub use lm::{FlowBuildExt, into_compile_parts};
pub use weight_loader::{
    GgufLoader, HfTranslatingLoader, WeightLoader, ggml_type_to_quant_scheme, gguf_to_hf_name,
    gguf_to_hf_name_for_arch, hf_to_gguf_name, is_mtp_weight, load_from_path,
};
pub use weight_map::{WeightDrainPolicy, WeightMap};
pub use weight_registry::{
    LoadWeightsOptions, LoadedWeights, RegisteredFormat, WeightFormatRegistration,
    format_for_extension, list_registered_formats, load_weight_map_resolved, load_weights_resolved,
    open_weight_loader, register_weight_format, registered_extensions_hint,
};
pub use weights::{
    GgufDirGuide, LoadOpts, ResolveOpts, default_resolve_opts, gguf_dir_guide, init,
    load_weight_map, open, open_map, open_map_with, open_with, pick, pick_default,
};