{ "_id": "6a17536aacfb0bcc41d96014", "Package": "llamaR", "Type": "Package", "Title": "Interface for Large Language Models via 'llama.cpp'", "Version": "0.2.4", "Authors@R": "c(\nperson(\"Yuri\", \"Baramykov\", email = \"lbsbmsu@mail.ru\", role = c(\"aut\", \"cre\"),\ncomment = c(ORCID = \"0009-0000-7627-4217\")),\nperson(\"Georgi\", \"Gerganov\", role = \"cph\",\ncomment = \"Author of the 'llama.cpp' library included in src/\"))", "Description": "Provides 'R' bindings to 'llama.cpp' for running Large\nLanguage Models ('LLMs') locally with optional 'Vulkan' GPU\nacceleration via 'ggmlR'. Supports model loading, text\ngeneration, 'tokenization', token-to-piece conversion,\n'embeddings' (single and batch), encoder-decoder inference,\nlow-level batch management, chat templates, 'LoRA' adapters,\nexplicit backend/device selection, multi-GPU split, and 'NUMA'\noptimization. Includes a high-level 'ragnar'-compatible\nembedding provider ('embed_llamar'). Built on top of 'ggmlR'\nfor efficient tensor operations.", "License": "MIT + file LICENSE", "URL": "https://github.com/Zabis13/llamaR", "BugReports": "https://github.com/Zabis13/llamaR/issues", "Encoding": "UTF-8", "SystemRequirements": "C++17, GNU make", "VignetteBuilder": "knitr", "RoxygenNote": "7.3.3", "Config/testthat/edition": "3", "NeedsCompilation": "yes", "Packaged": { "Date": "2026-05-27 20:17:02 UTC", "User": "root" }, "Author": "Yuri Baramykov [aut, cre] (ORCID:\n), Georgi Gerganov [cph]\n(Author of the 'llama.cpp' library included in src/)", "Maintainer": "Yuri Baramykov ", "Config/pak/sysreqs": "make", "Repository": "https://cran.r-universe.dev", "Date/Publication": "2026-05-27 17:30:02 UTC", "RemoteUrl": "https://github.com/cran/llamaR", "RemoteRef": "HEAD", "RemoteSha": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "MD5sum": "417dfc3c18c9c411d751b2b6eca5e278", "_user": "cran", "_type": "src", "_file": "llamaR_0.2.4.tar.gz", "_fileid": "d948f95eacdc9e1f180019badad14c07e4581329e7cbfb3422de8e0748c367c4", "_filesize": 764550, "_sha256": "d948f95eacdc9e1f180019badad14c07e4581329e7cbfb3422de8e0748c367c4", "_created": "2026-05-27T20:17:02.000Z", "_published": "2026-05-27T20:26:18.365Z", "_distro": "noble", "_jobs": [ { "job": 78166453989, "time": 307, "config": "linux-devel-arm64", "r": "4.7.0", "check": "NOTE", "artifact": "7251044360" }, { "job": 78166453934, "time": 292, "config": "linux-devel-x86_64", "r": "4.7.0", "check": "NOTE", "artifact": "7251039714" }, { "job": 78166453963, "time": 304, "config": "linux-release-arm64", "r": "4.6.0", "check": "NOTE", "artifact": "7251043280" }, { "job": 78166454010, "time": 328, "config": "linux-release-x86_64", "r": "4.6.0", "check": "NOTE", "artifact": "7251051702" }, { "job": 78164861794, "time": 535, "config": "source", "r": "4.6.0", "check": "OK", "artifact": "7250940830" }, { "job": 78166453907, "time": 301, "config": "wasm-release", "r": "4.6.0", "check": "FAIL", "artifact": "" } ], "_buildurl": "https://github.com/r-universe/cran/actions/runs/26535966284", "_status": "success", "_host": "GitHub-Actions", "_upstream": "https://github.com/cran/llamaR", "_commit": { "id": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "author": "Yuri Baramykov ", "committer": "cran-robot ", "message": "version 0.2.4\n", "time": 1779903002 }, "_maintainer": { "name": "Yuri Baramykov", "email": "lbsbmsu@mail.ru", "login": "zabis13", "description": "Yuri Baramykov", "uuid": 77751222 }, "_registered": true, "_dependencies": [ { "package": "R", "version": ">= 4.1.0", "role": "Depends" }, { "package": "ggmlR", "role": "Depends" }, { "package": "ggmlR", "role": "LinkingTo" }, { "package": "jsonlite", "role": "Imports" }, { "package": "utils", "role": "Imports" }, { "package": "testthat", "version": ">= 3.0.0", "role": "Suggests" }, { "package": "withr", "role": "Suggests" }, { "package": "drogonR", "role": "Suggests" }, { "package": "later", "role": "Suggests" }, { "package": "ellmer", "role": "Suggests" }, { "package": "callr", "role": "Suggests" }, { "package": "knitr", "role": "Suggests" }, { "package": "rmarkdown", "role": "Suggests" } ], "_owner": "cran", "_selfowned": false, "_usedby": 0, "_updates": [ { "week": "2026-10", "n": 1 }, { "week": "2026-15", "n": 1 }, { "week": "2026-22", "n": 1 } ], "_tags": [ { "name": "0.2.2", "date": "2026-03-05" }, { "name": "0.2.3", "date": "2026-04-06" }, { "name": "0.2.4", "date": "2026-05-27" } ], "_stars": 0, "_contributors": [ { "user": "zabis13", "count": 3, "uuid": 77751222 } ], "_userbio": { "uuid": 6899542, "type": "organization", "name": "cran", "description": "Unofficial read-only mirror of all CRAN R packages" }, "_downloads": { "count": 495, "source": "https://cranlogs.r-pkg.org/downloads/total/last-month/llamaR" }, "_devurl": "https://github.com/zabis13/llamar", "_searchresults": 64, "_topics": [ "cpp", "openmp" ], "_rbuild": "4.6.0", "_assets": [ "extra/citation.cff", "extra/citation.html", "extra/citation.json", "extra/citation.txt", "extra/contents.json", "extra/llamaR.html", "extra/NEWS.html", "extra/NEWS.txt", "extra/readme.html", "extra/readme.md", "manual.pdf" ], "_homeurl": "https://github.com/zabis13/llamar", "_realowner": "zabis13", "_cranurl": false, "_releases": [ { "version": "0.2.2", "date": "2026-03-05" }, { "version": "0.2.3", "date": "2026-04-06" }, { "version": "0.2.4", "date": "2026-05-27" } ], "_exports": [ "chat_llamar", "chat_llamar_stop", "embed_llamar", "llama_backend_devices", "llama_batch_free", "llama_batch_init", "llama_chat_apply_template", "llama_chat_builtin_templates", "llama_chat_template", "llama_detokenize", "llama_embed_batch", "llama_embeddings", "llama_encode", "llama_free_context", "llama_free_model", "llama_gen_begin", "llama_gen_end", "llama_gen_next", "llama_generate", "llama_generate_batch", "llama_get_embeddings", "llama_get_embeddings_ith", "llama_get_embeddings_seq", "llama_get_logits", "llama_get_logits_ith", "llama_get_model", "llama_get_verbosity", "llama_hf_cache_clear", "llama_hf_cache_dir", "llama_hf_cache_info", "llama_hf_download", "llama_hf_list", "llama_load_model", "llama_load_model_hf", "llama_lora_apply", "llama_lora_clear", "llama_lora_load", "llama_lora_remove", "llama_max_devices", "llama_memory_breakdown_print", "llama_memory_can_shift", "llama_memory_clear", "llama_memory_seq_add", "llama_memory_seq_cp", "llama_memory_seq_div", "llama_memory_seq_keep", "llama_memory_seq_pos_range", "llama_memory_seq_rm", "llama_model_info", "llama_model_meta", "llama_model_meta_val", "llama_n_batch", "llama_n_ctx", "llama_n_ctx_seq", "llama_n_seq_max", "llama_n_threads", "llama_n_threads_batch", "llama_n_ubatch", "llama_new_context", "llama_numa_init", "llama_perf", "llama_perf_print", "llama_perf_reset", "llama_pooling_type", "llama_serve_openai", "llama_set_abort_callback", "llama_set_causal_attn", "llama_set_threads", "llama_set_verbosity", "llama_set_warmup", "llama_state_get_size", "llama_state_load", "llama_state_save", "llama_supports_gpu", "llama_supports_mlock", "llama_supports_mmap", "llama_supports_rpc", "llama_synchronize", "llama_system_info", "llama_time_us", "llama_token_to_piece", "llama_tokenize", "llama_vocab_get_score", "llama_vocab_get_text", "llama_vocab_info", "llama_vocab_is_control", "llama_vocab_is_eog", "llama_vocab_type" ], "_help": [ { "page": "chat_llamar", "title": "Chat with a local model through an ellmer::Chat object", "topics": [ "chat_llamar" ] }, { "page": "chat_llamar_stop", "title": "Stop the server spawned by chat_llamar()", "topics": [ "chat_llamar_stop" ] }, { "page": "embed_llamar", "title": "Embedding provider for ragnar / standalone use", "topics": [ "embed_llamar" ] }, { "page": "llama_backend_devices", "title": "List available backend devices", "topics": [ "llama_backend_devices" ] }, { "page": "llama_batch_free", "title": "Free a llama batch allocated with 'llama_batch_init()'", "topics": [ "llama_batch_free" ] }, { "page": "llama_batch_init", "title": "Initialise a llama batch", "topics": [ "llama_batch_init" ] }, { "page": "llama_chat_apply_template", "title": "Apply chat template to messages", "topics": [ "llama_chat_apply_template" ] }, { "page": "llama_chat_builtin_templates", "title": "List built-in chat templates", "topics": [ "llama_chat_builtin_templates" ] }, { "page": "llama_chat_template", "title": "Get model's built-in chat template", "topics": [ "llama_chat_template" ] }, { "page": "llama_detokenize", "title": "Detokenize token IDs back to text", "topics": [ "llama_detokenize" ] }, { "page": "llama_embed_batch", "title": "Batch embeddings for multiple texts", "topics": [ "llama_embed_batch" ] }, { "page": "llama_embeddings", "title": "Extract embeddings for a text", "topics": [ "llama_embeddings" ] }, { "page": "llama_encode", "title": "Encode tokens using the encoder (encoder-decoder models only)", "topics": [ "llama_encode" ] }, { "page": "llama_free_context", "title": "Free an inference context", "topics": [ "llama_free_context" ] }, { "page": "llama_free_model", "title": "Free a loaded model", "topics": [ "llama_free_model" ] }, { "page": "llama_gen_begin", "title": "Begin a streaming (token-by-token) generation", "topics": [ "llama_gen_begin" ] }, { "page": "llama_gen_end", "title": "Finish a streaming generation", "topics": [ "llama_gen_end" ] }, { "page": "llama_gen_next", "title": "Pull the next chunk of a streaming generation", "topics": [ "llama_gen_next" ] }, { "page": "llama_generate", "title": "Generate text from a prompt", "topics": [ "llama_generate" ] }, { "page": "llama_generate_batch", "title": "Generate completions for multiple prompts in parallel", "topics": [ "llama_generate_batch" ] }, { "page": "llama_get_embeddings", "title": "Get all output token embeddings as a matrix", "topics": [ "llama_get_embeddings" ] }, { "page": "llama_get_embeddings_ith", "title": "Get embeddings for the i-th token in the batch", "topics": [ "llama_get_embeddings_ith" ] }, { "page": "llama_get_embeddings_seq", "title": "Get pooled embeddings for a sequence", "topics": [ "llama_get_embeddings_seq" ] }, { "page": "llama_get_logits", "title": "Get logits from the last decode step", "topics": [ "llama_get_logits" ] }, { "page": "llama_get_logits_ith", "title": "Get logits for a specific token position", "topics": [ "llama_get_logits_ith" ] }, { "page": "llama_get_model", "title": "Get the model associated with a context", "topics": [ "llama_get_model" ] }, { "page": "llama_get_verbosity", "title": "Get current verbosity level", "topics": [ "llama_get_verbosity" ] }, { "page": "llama_hf_cache_clear", "title": "Clear the model cache", "topics": [ "llama_hf_cache_clear" ] }, { "page": "llama_hf_cache_dir", "title": "Get the cache directory for downloaded models", "topics": [ "llama_hf_cache_dir" ] }, { "page": "llama_hf_cache_info", "title": "Show information about the model cache", "topics": [ "llama_hf_cache_info" ] }, { "page": "llama_hf_download", "title": "Download a GGUF model from Hugging Face", "topics": [ "llama_hf_download" ] }, { "page": "llama_hf_list", "title": "List GGUF files in a Hugging Face repository", "topics": [ "llama_hf_list" ] }, { "page": "llama_load_model", "title": "Load a GGUF model file", "topics": [ "llama_load_model" ] }, { "page": "llama_load_model_hf", "title": "Load a model directly from Hugging Face", "topics": [ "llama_load_model_hf" ] }, { "page": "llama_lora_apply", "title": "Apply a LoRA adapter to context", "topics": [ "llama_lora_apply" ] }, { "page": "llama_lora_clear", "title": "Remove all LoRA adapters from context", "topics": [ "llama_lora_clear" ] }, { "page": "llama_lora_load", "title": "Load a LoRA adapter", "topics": [ "llama_lora_load" ] }, { "page": "llama_lora_remove", "title": "Remove a LoRA adapter from context", "topics": [ "llama_lora_remove" ] }, { "page": "llama_max_devices", "title": "Get maximum number of devices", "topics": [ "llama_max_devices" ] }, { "page": "llama_memory_breakdown_print", "title": "Print memory breakdown by device", "topics": [ "llama_memory_breakdown_print" ] }, { "page": "llama_memory_can_shift", "title": "Check if the KV cache supports shifting", "topics": [ "llama_memory_can_shift" ] }, { "page": "llama_memory_clear", "title": "Clear the KV cache", "topics": [ "llama_memory_clear" ] }, { "page": "llama_memory_seq_add", "title": "Shift token positions in a sequence", "topics": [ "llama_memory_seq_add" ] }, { "page": "llama_memory_seq_cp", "title": "Copy a sequence in the KV cache", "topics": [ "llama_memory_seq_cp" ] }, { "page": "llama_memory_seq_div", "title": "Integer-divide token positions in a sequence", "topics": [ "llama_memory_seq_div" ] }, { "page": "llama_memory_seq_keep", "title": "Keep only one sequence in the KV cache", "topics": [ "llama_memory_seq_keep" ] }, { "page": "llama_memory_seq_pos_range", "title": "Get position range for a sequence", "topics": [ "llama_memory_seq_pos_range" ] }, { "page": "llama_memory_seq_rm", "title": "Remove tokens from a sequence in the KV cache", "topics": [ "llama_memory_seq_rm" ] }, { "page": "llama_model_info", "title": "Get model metadata", "topics": [ "llama_model_info" ] }, { "page": "llama_model_meta", "title": "Get all model metadata as a named character vector", "topics": [ "llama_model_meta" ] }, { "page": "llama_model_meta_val", "title": "Get a single model metadata value by key", "topics": [ "llama_model_meta_val" ] }, { "page": "llama_n_batch", "title": "Get logical batch size", "topics": [ "llama_n_batch" ] }, { "page": "llama_n_ctx", "title": "Get context window size", "topics": [ "llama_n_ctx" ] }, { "page": "llama_n_ctx_seq", "title": "Get per-sequence context window size", "topics": [ "llama_n_ctx_seq" ] }, { "page": "llama_n_seq_max", "title": "Get maximum number of sequences", "topics": [ "llama_n_seq_max" ] }, { "page": "llama_n_threads", "title": "Get number of threads for single-token generation", "topics": [ "llama_n_threads" ] }, { "page": "llama_n_threads_batch", "title": "Get number of threads for batch processing", "topics": [ "llama_n_threads_batch" ] }, { "page": "llama_n_ubatch", "title": "Get physical micro-batch size", "topics": [ "llama_n_ubatch" ] }, { "page": "llama_new_context", "title": "Create an inference context", "topics": [ "llama_new_context" ] }, { "page": "llama_numa_init", "title": "Initialize NUMA optimization", "topics": [ "llama_numa_init" ] }, { "page": "llama_perf", "title": "Get performance statistics", "topics": [ "llama_perf" ] }, { "page": "llama_perf_print", "title": "Print performance statistics to the console", "topics": [ "llama_perf_print" ] }, { "page": "llama_perf_reset", "title": "Reset performance counters", "topics": [ "llama_perf_reset" ] }, { "page": "llama_pooling_type", "title": "Get pooling type", "topics": [ "llama_pooling_type" ] }, { "page": "llama_serve_openai", "title": "Serve an OpenAI-compatible HTTP API for a local model", "topics": [ "llama_serve_openai" ] }, { "page": "llama_set_abort_callback", "title": "Set or clear the abort callback", "topics": [ "llama_set_abort_callback" ] }, { "page": "llama_set_causal_attn", "title": "Set causal attention mode", "topics": [ "llama_set_causal_attn" ] }, { "page": "llama_set_threads", "title": "Set the number of threads for a context", "topics": [ "llama_set_threads" ] }, { "page": "llama_set_verbosity", "title": "Set logging verbosity level", "topics": [ "llama_set_verbosity" ] }, { "page": "llama_set_warmup", "title": "Set warmup mode", "topics": [ "llama_set_warmup" ] }, { "page": "llama_state_get_size", "title": "Get the size of the serialized context state in bytes", "topics": [ "llama_state_get_size" ] }, { "page": "llama_state_load", "title": "Load context state from file", "topics": [ "llama_state_load" ] }, { "page": "llama_state_save", "title": "Save context state to file", "topics": [ "llama_state_save" ] }, { "page": "llama_supports_gpu", "title": "Check whether GPU offloading is available", "topics": [ "llama_supports_gpu" ] }, { "page": "llama_supports_mlock", "title": "Check whether memory locking is supported", "topics": [ "llama_supports_mlock" ] }, { "page": "llama_supports_mmap", "title": "Check whether memory-mapped file I/O is supported", "topics": [ "llama_supports_mmap" ] }, { "page": "llama_supports_rpc", "title": "Check whether RPC backend is available", "topics": [ "llama_supports_rpc" ] }, { "page": "llama_synchronize", "title": "Synchronize asynchronous computation", "topics": [ "llama_synchronize" ] }, { "page": "llama_system_info", "title": "Get system information string", "topics": [ "llama_system_info" ] }, { "page": "llama_time_us", "title": "Get current time in microseconds", "topics": [ "llama_time_us" ] }, { "page": "llama_token_to_piece", "title": "Convert a single token ID to its text piece", "topics": [ "llama_token_to_piece" ] }, { "page": "llama_tokenize", "title": "Tokenize text into token IDs", "topics": [ "llama_tokenize" ] }, { "page": "llama_vocab_get_score", "title": "Get the score of a token", "topics": [ "llama_vocab_get_score" ] }, { "page": "llama_vocab_get_text", "title": "Get the text representation of a token", "topics": [ "llama_vocab_get_text" ] }, { "page": "llama_vocab_info", "title": "Get vocabulary special token IDs", "topics": [ "llama_vocab_info" ] }, { "page": "llama_vocab_is_control", "title": "Check if a token is a control token", "topics": [ "llama_vocab_is_control" ] }, { "page": "llama_vocab_is_eog", "title": "Check if a token is an end-of-generation token", "topics": [ "llama_vocab_is_eog" ] }, { "page": "llama_vocab_type", "title": "Get vocabulary type", "topics": [ "llama_vocab_type" ] } ], "_readme": "https://github.com/cran/llamaR/raw/HEAD/README.md", "_rundeps": [ "generics", "ggmlR", "jsonlite", "R6" ], "_sysdeps": [ { "shlib": "libstdc++", "package": "libstdc++6", "source": "gcc", "version": "14.2.0-4ubuntu2~24.04.1", "name": "c++", "homepage": "http://gcc.gnu.org/", "description": "GNU Standard C++ Library v3" }, { "shlib": "libgomp", "package": "libgomp1", "source": "gcc", "version": "14.2.0-4ubuntu2~24.04.1", "name": "openmp", "homepage": "http://gcc.gnu.org/", "description": "GCC OpenMP (GOMP) support library" } ], "_vignettes": [ { "source": "chat-and-agents.Rmd", "filename": "chat-and-agents.html", "title": "Chat and Agents", "engine": "knitr::rmarkdown", "headings": [ "1. The chat object: chat_llamar()", "Mode A — spawn a server for a model", "Mode B — connect to a running server", "System prompt", "2. The server: llama_serve_openai()", "Connecting OpenCode", "3. The command-line example", "4. ragnar: retrieval-augmented chat", "5. Concurrency", "See also" ], "created": "2026-05-27 17:30:02", "modified": "2026-05-27 17:30:02", "commits": 1 }, { "source": "getting-started.Rmd", "filename": "getting-started.html", "title": "Getting Started with llamaR", "engine": "knitr::rmarkdown", "headings": [ "1. Getting a model", "2. Loading a model and creating a context", "3. Generating text", "4. Chat models and templates", "5. Tokenization", "6. Embeddings", "ragnar-compatible provider", "7. Serving and chatting", "See also" ], "created": "2026-05-27 17:30:02", "modified": "2026-05-27 17:30:02", "commits": 1 } ], "_score": 4.28330122870355, "_indexed": false, "_nocasepkg": "llamar", "_universes": [ "cran" ], "_indexurl": "https://zabis13.r-universe.dev/llamaR", "_previous": "0.2.3", "_binaries": [ { "r": "4.7.0", "os": "linux", "version": "0.2.4", "date": "2026-05-27T20:22:30.000Z", "distro": "noble", "arch": "aarch64", "commit": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "fileid": "981d3d2d8ffc94a257d44235a4fc8a25fdc93bdba4af9e59be234eec16672979", "status": "success", "check": "NOTE", "buildurl": "https://github.com/r-universe/cran/actions/runs/26535966284" }, { "r": "4.7.0", "os": "linux", "version": "0.2.4", "date": "2026-05-27T20:22:18.000Z", "distro": "noble", "arch": "x86_64", "commit": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "fileid": "0a5dc7ac450591795c528be29f66d6582f694310b3ed6bd302811edaf580ca41", "status": "success", "check": "NOTE", "buildurl": "https://github.com/r-universe/cran/actions/runs/26535966284" }, { "r": "4.6.0", "os": "linux", "version": "0.2.4", "date": "2026-05-27T20:22:30.000Z", "distro": "noble", "arch": "aarch64", "commit": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "fileid": "2f45e4eef1f634d9626ce293f6a4e33fd64d982e37f8cd648a22b100192e8720", "status": "success", "check": "NOTE", "buildurl": "https://github.com/r-universe/cran/actions/runs/26535966284" }, { "r": "4.6.0", "os": "linux", "version": "0.2.4", "date": "2026-05-27T20:22:49.000Z", "distro": "noble", "arch": "x86_64", "commit": "267f33399ed11b84da11fbcde7b1aa92dbce2644", "fileid": "ea90f0592cfafed8d74fcc0d9cbed37fb60c7df7bfd6892a7a76f246d8b84b8a", "status": "success", "check": "NOTE", "buildurl": "https://github.com/r-universe/cran/actions/runs/26535966284" } ] }