{
  "$schema": "https://huggingface.co/software-mansion/react-native-executorch-spec/resolve/main/config.schema.json",
  "model": "gemma_4",
  "family": "gemma",
  "capabilities": [
    "text-generation",
    "image-text-to-text",
    "audio-text-to-text"
  ],
  "backend": "vulkan",
  "license": "gemma",
  "variants": [
    {
      "file": "gemma_4_e2b_vulkan_8da4w.pte",
      "precision": "8da4w",
      "quantized": true,
      "default": true,
      "methods": {
        "enable_dynamic_shape": true,
        "get_bos_id": 2,
        "get_eos_ids": [
          1,
          106,
          50
        ],
        "get_max_context_len": 2048,
        "get_max_seq_len": 128,
        "get_vocab_size": 262144,
        "use_kv_cache": true,
        "audio_block_k_max": 62,
        "audio_block_k_min": 1,
        "audio_block_samples": 7680,
        "audio_encoder": {
          "inputs": [],
          "outputs": []
        },
        "audio_placeholder_id": 258881,
        "audio_sample_count_max": 476160,
        "audio_sample_count_min": 7680,
        "audio_sample_k": 62,
        "audio_soft_tokens_max": 744,
        "has_ple": true,
        "image_placeholder_id": 262145,
        "text_decoder": {
          "inputs": [],
          "outputs": []
        },
        "token_embedding": {
          "inputs": [],
          "outputs": []
        },
        "vision_encoder": {
          "inputs": [],
          "outputs": []
        }
      }
    }
  ],
  "size": "e2b",
  "_methods_needing_shapes": [
    "audio_encoder",
    "text_decoder",
    "token_embedding",
    "vision_encoder"
  ]
}