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
40 lines • 1.66 kB
JavaScript
import { GgufFileType } from "../types/GgufMetadataTypes.js";
export const ggufQuantNames = new Map([
["Q4_0", GgufFileType.MOSTLY_Q4_0],
["Q4_1", GgufFileType.MOSTLY_Q4_1],
["Q5_0", GgufFileType.MOSTLY_Q5_0],
["Q5_1", GgufFileType.MOSTLY_Q5_1],
["IQ2_XXS", GgufFileType.MOSTLY_IQ2_XXS],
["IQ2_XS", GgufFileType.MOSTLY_IQ2_XS],
["IQ2_S", GgufFileType.MOSTLY_IQ2_S],
["IQ2_M", GgufFileType.MOSTLY_IQ2_M],
["IQ1_S", GgufFileType.MOSTLY_IQ1_S],
["IQ1_M", GgufFileType.MOSTLY_IQ1_M],
["TQ1_0", GgufFileType.MOSTLY_TQ1_0],
["TQ2_0", GgufFileType.MOSTLY_TQ2_0],
["Q2_K", GgufFileType.MOSTLY_Q2_K],
["Q2_K_S", GgufFileType.MOSTLY_Q2_K_S],
["IQ3_XXS", GgufFileType.MOSTLY_IQ3_XXS],
["IQ3_S", GgufFileType.MOSTLY_IQ3_S],
["IQ3_M", GgufFileType.MOSTLY_IQ3_M],
["Q3_K", GgufFileType.MOSTLY_Q3_K_M],
["IQ3_XS", GgufFileType.MOSTLY_IQ3_XS],
["Q3_K_S", GgufFileType.MOSTLY_Q3_K_S],
["Q3_K_M", GgufFileType.MOSTLY_Q3_K_M],
["Q3_K_L", GgufFileType.MOSTLY_Q3_K_L],
["IQ4_NL", GgufFileType.MOSTLY_IQ4_NL],
["IQ4_XS", GgufFileType.MOSTLY_IQ4_XS],
["Q4_K", GgufFileType.MOSTLY_Q4_K_M],
["Q4_K_S", GgufFileType.MOSTLY_Q4_K_S],
["Q4_K_M", GgufFileType.MOSTLY_Q4_K_M],
["Q5_K", GgufFileType.MOSTLY_Q5_K_M],
["Q5_K_S", GgufFileType.MOSTLY_Q5_K_S],
["Q5_K_M", GgufFileType.MOSTLY_Q5_K_M],
["Q6_K", GgufFileType.MOSTLY_Q6_K],
["Q8_0", GgufFileType.MOSTLY_Q8_0],
["F16", GgufFileType.MOSTLY_F16],
["BF16", GgufFileType.MOSTLY_BF16],
["F32", GgufFileType.ALL_F32],
["COPY", GgufFileType.ALL_F32]
]);
//# sourceMappingURL=ggufQuantNames.js.map