|
liblloyal 1.0.0
Branched Inference for llama.cpp
|
Files | |
| boundaries.hpp | |
| branch.hpp | |
| Branch Primitive for Tree Search and Multi-Sequence Generation. | |
| chat_in.hpp | |
| Chat Input Formatting with Full Format Awareness. | |
| chat_out.hpp | |
| Chat Output Parsing. | |
| common.hpp | |
| decode.hpp | |
| Batch Decoding Operations. | |
| embedding.hpp | |
| Embedding Extraction and Normalization. | |
| grammar.hpp | |
| Grammar-Constrained Sampling. | |
| kv.hpp | |
| KV Cache Physics. | |
| logits.hpp | |
| Zero-copy logits access with clear lifetime semantics. | |
| metrics.hpp | |
| Distribution Metrics for Test-Time Alignment. | |
| model_registry.hpp | |
| Thread-Safe Model Cache. | |
| sampler.hpp | |
| Token Sampling Operations. | |
| tokenizer.hpp | |
| Text Tokenization Operations. | |