{
  "model_name": "FireRedPunc",
  "framework": "pytorch",
  "task": "punctuation-restoration",
  "supported_chips": ["AX650N"],
  "quantization": {
    "activation": "U16",
    "weight": "S8",
    "gather_embedding": "FP32",
    "smooth_quant": true
  },
  "seq_len": 128,
  "inputs": [
    {"name": "input_ids", "shape": [1, 128], "dtype": "int32"},
    {"name": "attention_mask", "shape": [1, 128], "dtype": "float32"}
  ],
  "outputs": [
    {"name": "logits", "shape": [1, 127, 5], "dtype": "float32"}
  ],
  "files": {
    "axmodel": "models/punc.axmodel",
    "tokenizer": "models/tokenizer",
    "vocab": "models/vocab.txt",
    "out_dict": "models/out_dict",
    "cpp_binary": "bin/fireredpunc_ax650"
  }
}