{
  "architecture" : {
    "attentionLayers" : 16,
    "class" : "dense",
    "totalLayers" : 16
  },
  "capabilities" : [
    "chat"
  ],
  "driver" : "causal-lm",
  "engine" : {
    "allowed" : [
      "pipelined",
      "sequential"
    ],
    "expectFrequentReshapes" : true,
    "kind" : "pipelined"
  },
  "functions" : {
    "main" : {
      "role" : "step"
    }
  },
  "inputs" : {
    "input_ids" : {
      "bind" : "tokens",
      "dtype" : "int32"
    },
    "position_ids" : {
      "bind" : "positions",
      "dtype" : "int32",
      "mode" : "running"
    }
  },
  "limits" : {
    "contextMax" : {
      "macos" : 4096
    },
    "maxOutputDefault" : 1024,
    "prefillChunk" : 512,
    "prefillChunkThreshold" : 1024
  },
  "memory" : {
    "bytesPerToken" : 1691773972,
    "kvBytesPerToken" : 131072,
    "requiresIncreasedMemoryLimit" : false,
    "weightsBytes" : 1691773972
  },
  "outputs" : {
    "logits" : {
      "dtype" : "float16",
      "role" : "logits",
      "seq" : "dynamic"
    }
  },
  "provenance" : {
    "exportedOn" : {
      "date" : "2026-09-30",
      "os" : "26A434"
    },
    "exporter" : {
      "coreai-core" : "1.0.0b2",
      "coreai-opt" : "0.2.1",
      "coreai-torch" : "0.4.2",
      "torch" : "2.11.0"
    },
    "files" : {
      "metadata.json" : "sha256:439ec0c104785580728bf86b29b0e8d6bb5486b6b7919fa19f24ada94d57e9c1",
      "olmo_2_1b_instruct.aimodel/main.hash" : "sha256:4dea3c7a5ec9f5901c291954fc9cd9db288ef9e3ce2e29842df67b257df72bd0",
      "olmo_2_1b_instruct.aimodel/main.mlirb" : "sha256:3b66e59bf0421b4635facae5bdcd33e80862d347d291ae0135cf3c5c0c57855a",
      "olmo_2_1b_instruct.aimodel/metadata.json" : "sha256:130d9cda6d55a54649382c7e6bef3afd9da1a3b747fa59ba4d987f8f8cf95440",
      "tokenizer/generation_config.json" : "sha256:437b97826bb4430c205083a0986fbe31f4959cffbe445e47fb7c5192fdcb8d58",
      "tokenizer/merges.txt" : "sha256:b6fe424e334903f7fb84d3a106d9730455f4744b9fe3c21ee136d97a00e72502",
      "tokenizer/special_tokens_map.json" : "sha256:78afb564e81264029b25f9caf24bda2521d5bdaeff5cd3fdbc01d3da2e8ce2f2",
      "tokenizer/tokenizer.json" : "sha256:73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca",
      "tokenizer/tokenizer_config.json" : "sha256:50c412c57d832057a3d5db42064c741f751e570f7c8788f037bfb0d2dd6e5f49",
      "tokenizer/vocab.json" : "sha256:9e14712c91b37c7aab74b1306baa46ac342d620637a4b44523cdc3aec7d24195"
    },
    "forge" : "0.1.0",
    "recipe" : "olmo-2-1b-instruct@1",
    "source" : {
      "hfRepo" : "allenai/OLMo-2-0425-1B-Instruct",
      "revision" : "48d788eca847d4d7548f375ad03d3c9312f6139e"
    }
  },
  "quantization" : {
    "notes" : "weight-only PTQ on the torch module (coreai-opt eager Quantizer); exceptions resolve first-match in declaration order, then default; modules: int8-linear-perchannel ×82",
    "scheme" : "int8-linear-perchannel",
    "weightBits" : 8
  },
  "schema" : 1,
  "states" : {
    "keyCache" : {
      "axis" : 3,
      "initial" : 256,
      "kind" : "kvGrowing"
    },
    "valueCache" : {
      "axis" : 3,
      "initial" : 256,
      "kind" : "kvGrowing"
    }
  },
  "target" : {
    "arch" : "any",
    "backend" : "coreai",
    "compiled" : false,
    "compute" : "gpu",
    "minOS" : "27.0",
    "platform" : "macos"
  },
  "text" : {
    "chatTemplate" : "embedded",
    "tokenizer" : "tokenizer/"
  }
}