node-llama-cpp
Version:
Run AI models locally on your machine with node.js bindings for llama.cpp. Enforce a JSON schema on the model output on the generation level
177 lines • 10.2 kB
JavaScript
export var GgufArchitectureType;
(function (GgufArchitectureType) {
GgufArchitectureType["llama"] = "llama";
GgufArchitectureType["llama4"] = "llama4";
GgufArchitectureType["deci"] = "deci";
GgufArchitectureType["falcon"] = "falcon";
GgufArchitectureType["grok"] = "grok";
GgufArchitectureType["gpt2"] = "gpt2";
GgufArchitectureType["gptj"] = "gptj";
GgufArchitectureType["gptneox"] = "gptneox";
GgufArchitectureType["mpt"] = "mpt";
GgufArchitectureType["baichuan"] = "baichuan";
GgufArchitectureType["starcoder"] = "starcoder";
GgufArchitectureType["refact"] = "refact";
GgufArchitectureType["bert"] = "bert";
GgufArchitectureType["nomicBert"] = "nomic-bert";
GgufArchitectureType["nomicBertMoe"] = "nomic-bert-moe";
GgufArchitectureType["neoBert"] = "neo-bert";
GgufArchitectureType["jinaBertV2"] = "jina-bert-v2";
GgufArchitectureType["jinaBertV3"] = "jina-bert-v3";
GgufArchitectureType["bloom"] = "bloom";
GgufArchitectureType["stablelm"] = "stablelm";
GgufArchitectureType["qwen"] = "qwen";
GgufArchitectureType["qwen2"] = "qwen2";
GgufArchitectureType["qwen2moe"] = "qwen2moe";
GgufArchitectureType["qwen2vl"] = "qwen2vl";
GgufArchitectureType["qwen3"] = "qwen3";
GgufArchitectureType["qwen3moe"] = "qwen3moe";
GgufArchitectureType["qwen3next"] = "qwen3next";
GgufArchitectureType["qwen3vl"] = "qwen3vl";
GgufArchitectureType["qwen3vlmoe"] = "qwen3vlmoe";
GgufArchitectureType["phi2"] = "phi2";
GgufArchitectureType["phi3"] = "phi3";
GgufArchitectureType["phimoe"] = "phimoe";
GgufArchitectureType["plamo"] = "plamo";
GgufArchitectureType["plamo2"] = "plamo2";
GgufArchitectureType["codeshell"] = "codeshell";
GgufArchitectureType["orion"] = "orion";
GgufArchitectureType["internlm2"] = "internlm2";
GgufArchitectureType["minicpm"] = "minicpm";
GgufArchitectureType["minicpm3"] = "minicpm3";
GgufArchitectureType["gemma"] = "gemma";
GgufArchitectureType["gemma2"] = "gemma2";
GgufArchitectureType["gemma3"] = "gemma3";
GgufArchitectureType["gemma3n"] = "gemma3n";
GgufArchitectureType["gemmaEmbedding"] = "gemma-embedding";
GgufArchitectureType["starcoder2"] = "starcoder2";
GgufArchitectureType["mamba"] = "mamba";
GgufArchitectureType["mamba2"] = "mamba2";
GgufArchitectureType["jamba"] = "jamba";
GgufArchitectureType["falconH1"] = "falcon-h1";
GgufArchitectureType["xverse"] = "xverse";
GgufArchitectureType["commandR"] = "command-r";
GgufArchitectureType["cohere2"] = "cohere2";
GgufArchitectureType["dbrx"] = "dbrx";
GgufArchitectureType["olmo"] = "olmo";
GgufArchitectureType["olmo2"] = "olmo2";
GgufArchitectureType["olmoe"] = "olmoe";
GgufArchitectureType["openelm"] = "openelm";
GgufArchitectureType["arctic"] = "arctic";
GgufArchitectureType["deepseek"] = "deepseek";
GgufArchitectureType["deepseek2"] = "deepseek2";
GgufArchitectureType["chatglm"] = "chatglm";
GgufArchitectureType["glm4"] = "glm4";
GgufArchitectureType["glm4moe"] = "glm4moe";
GgufArchitectureType["bitnet"] = "bitnet";
GgufArchitectureType["t5"] = "t5";
GgufArchitectureType["t5encoder"] = "t5encoder";
GgufArchitectureType["jais"] = "jais";
GgufArchitectureType["nemotron"] = "nemotron";
GgufArchitectureType["nemotronH"] = "nemotron_h";
GgufArchitectureType["exaone"] = "exaone";
GgufArchitectureType["exaone4"] = "exaone4";
GgufArchitectureType["rwkv6"] = "rwkv6";
GgufArchitectureType["rwkv6qwen2"] = "rwkv6qwen2";
GgufArchitectureType["rwkv7"] = "rwkv7";
GgufArchitectureType["arwkv7"] = "arwkv7";
GgufArchitectureType["granite"] = "granite";
GgufArchitectureType["granitemoe"] = "granitemoe";
GgufArchitectureType["granitehybrid"] = "granitehybrid";
GgufArchitectureType["chameleon"] = "chameleon";
GgufArchitectureType["wavtokenizerDec"] = "wavtokenizer-dec";
GgufArchitectureType["plm"] = "plm";
GgufArchitectureType["bailingmoe"] = "bailingmoe";
GgufArchitectureType["bailingmoe2"] = "bailingmoe2";
GgufArchitectureType["dots1"] = "dots1";
GgufArchitectureType["arcee"] = "arcee";
GgufArchitectureType["afmoe"] = "afmoe";
GgufArchitectureType["ernie4_5"] = "ernie4_5";
GgufArchitectureType["ernie4_5Moe"] = "ernie4_5-moe";
GgufArchitectureType["hunyuanMoe"] = "hunyuan-moe";
GgufArchitectureType["hunyuanDense"] = "hunyuan-dense";
GgufArchitectureType["smollm3"] = "smollm3";
GgufArchitectureType["gptOss"] = "gpt-oss";
GgufArchitectureType["lfm2"] = "lfm2";
GgufArchitectureType["lfm2moe"] = "lfm2moe";
GgufArchitectureType["dream"] = "dream";
GgufArchitectureType["smallthinker"] = "smallthinker";
GgufArchitectureType["llada"] = "llada";
GgufArchitectureType["lladaMoe"] = "llada-moe";
GgufArchitectureType["seedOss"] = "seed_oss";
GgufArchitectureType["grovemoe"] = "grovemoe";
GgufArchitectureType["apertus"] = "apertus";
GgufArchitectureType["minimaxM2"] = "minimax-m2";
GgufArchitectureType["cogvlm"] = "cogvlm";
GgufArchitectureType["rnd1"] = "rnd1";
GgufArchitectureType["panguEmbedded"] = "pangu-embedded";
GgufArchitectureType["mistral3"] = "mistral3";
GgufArchitectureType["clip"] = "clip";
GgufArchitectureType["unknown"] = "(unknown)";
})(GgufArchitectureType || (GgufArchitectureType = {}));
// source: `enum llama_ftype` in `llama.h` in the `llama.cpp` source code
export var GgufFileType;
(function (GgufFileType) {
GgufFileType[GgufFileType["ALL_F32"] = 0] = "ALL_F32";
GgufFileType[GgufFileType["MOSTLY_F16"] = 1] = "MOSTLY_F16";
GgufFileType[GgufFileType["MOSTLY_Q4_0"] = 2] = "MOSTLY_Q4_0";
GgufFileType[GgufFileType["MOSTLY_Q4_1"] = 3] = "MOSTLY_Q4_1";
GgufFileType[GgufFileType["MOSTLY_Q4_1_SOME_F16"] = 4] = "MOSTLY_Q4_1_SOME_F16";
GgufFileType[GgufFileType["MOSTLY_Q4_2"] = 5] = "MOSTLY_Q4_2";
GgufFileType[GgufFileType["MOSTLY_Q4_3"] = 6] = "MOSTLY_Q4_3";
GgufFileType[GgufFileType["MOSTLY_Q8_0"] = 7] = "MOSTLY_Q8_0";
GgufFileType[GgufFileType["MOSTLY_Q5_0"] = 8] = "MOSTLY_Q5_0";
GgufFileType[GgufFileType["MOSTLY_Q5_1"] = 9] = "MOSTLY_Q5_1";
GgufFileType[GgufFileType["MOSTLY_Q2_K"] = 10] = "MOSTLY_Q2_K";
GgufFileType[GgufFileType["MOSTLY_Q3_K_S"] = 11] = "MOSTLY_Q3_K_S";
GgufFileType[GgufFileType["MOSTLY_Q3_K_M"] = 12] = "MOSTLY_Q3_K_M";
GgufFileType[GgufFileType["MOSTLY_Q3_K_L"] = 13] = "MOSTLY_Q3_K_L";
GgufFileType[GgufFileType["MOSTLY_Q4_K_S"] = 14] = "MOSTLY_Q4_K_S";
GgufFileType[GgufFileType["MOSTLY_Q4_K_M"] = 15] = "MOSTLY_Q4_K_M";
GgufFileType[GgufFileType["MOSTLY_Q5_K_S"] = 16] = "MOSTLY_Q5_K_S";
GgufFileType[GgufFileType["MOSTLY_Q5_K_M"] = 17] = "MOSTLY_Q5_K_M";
GgufFileType[GgufFileType["MOSTLY_Q6_K"] = 18] = "MOSTLY_Q6_K";
GgufFileType[GgufFileType["MOSTLY_IQ2_XXS"] = 19] = "MOSTLY_IQ2_XXS";
GgufFileType[GgufFileType["MOSTLY_IQ2_XS"] = 20] = "MOSTLY_IQ2_XS";
GgufFileType[GgufFileType["MOSTLY_Q2_K_S"] = 21] = "MOSTLY_Q2_K_S";
GgufFileType[GgufFileType["MOSTLY_IQ3_XS"] = 22] = "MOSTLY_IQ3_XS";
GgufFileType[GgufFileType["MOSTLY_IQ3_XXS"] = 23] = "MOSTLY_IQ3_XXS";
GgufFileType[GgufFileType["MOSTLY_IQ1_S"] = 24] = "MOSTLY_IQ1_S";
GgufFileType[GgufFileType["MOSTLY_IQ4_NL"] = 25] = "MOSTLY_IQ4_NL";
GgufFileType[GgufFileType["MOSTLY_IQ3_S"] = 26] = "MOSTLY_IQ3_S";
GgufFileType[GgufFileType["MOSTLY_IQ3_M"] = 27] = "MOSTLY_IQ3_M";
GgufFileType[GgufFileType["MOSTLY_IQ2_S"] = 28] = "MOSTLY_IQ2_S";
GgufFileType[GgufFileType["MOSTLY_IQ2_M"] = 29] = "MOSTLY_IQ2_M";
GgufFileType[GgufFileType["MOSTLY_IQ4_XS"] = 30] = "MOSTLY_IQ4_XS";
GgufFileType[GgufFileType["MOSTLY_IQ1_M"] = 31] = "MOSTLY_IQ1_M";
GgufFileType[GgufFileType["MOSTLY_BF16"] = 32] = "MOSTLY_BF16";
GgufFileType[GgufFileType["MOSTLY_Q4_0_4_4"] = 33] = "MOSTLY_Q4_0_4_4";
GgufFileType[GgufFileType["MOSTLY_Q4_0_4_8"] = 34] = "MOSTLY_Q4_0_4_8";
GgufFileType[GgufFileType["MOSTLY_Q4_0_8_8"] = 35] = "MOSTLY_Q4_0_8_8";
GgufFileType[GgufFileType["MOSTLY_TQ1_0"] = 36] = "MOSTLY_TQ1_0";
GgufFileType[GgufFileType["MOSTLY_TQ2_0"] = 37] = "MOSTLY_TQ2_0";
GgufFileType[GgufFileType["MOSTLY_MXFP4_MOE"] = 38] = "MOSTLY_MXFP4_MOE";
})(GgufFileType || (GgufFileType = {}));
export var GgufMetadataTokenizerTokenType;
(function (GgufMetadataTokenizerTokenType) {
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["undefined"] = 0] = "undefined";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["normal"] = 1] = "normal";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["unknown"] = 2] = "unknown";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["control"] = 3] = "control";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["userDefined"] = 4] = "userDefined";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["unused"] = 5] = "unused";
GgufMetadataTokenizerTokenType[GgufMetadataTokenizerTokenType["byte"] = 6] = "byte";
})(GgufMetadataTokenizerTokenType || (GgufMetadataTokenizerTokenType = {}));
export var GgufMetadataArchitecturePoolingType;
(function (GgufMetadataArchitecturePoolingType) {
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["unspecified"] = -1] = "unspecified";
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["none"] = 0] = "none";
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["mean"] = 1] = "mean";
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["cls"] = 2] = "cls";
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["last"] = 3] = "last";
GgufMetadataArchitecturePoolingType[GgufMetadataArchitecturePoolingType["rank"] = 4] = "rank";
})(GgufMetadataArchitecturePoolingType || (GgufMetadataArchitecturePoolingType = {}));
export function isGgufMetadataOfArchitectureType(metadata, type) {
return metadata?.general?.architecture === type;
}
//# sourceMappingURL=GgufMetadataTypes.js.map