{
  "id": "bashkir_roberta",
  "version": "v1",
  "type": "masked_language_model",
  "status": "current",
  "language": "ba",
  "license": "other",
  "source": "a monolingual Bashkir-language dataset",
  "created_at": "2026-09-17T18:20:06Z",
  "architecture": "Pre-LayerNorm Transformer encoder",
  "hidden_size": 640,
  "num_hidden_layers": 8,
  "num_attention_heads": 10,
  "intermediate_size": 2560,
  "max_position_embeddings": 256,
  "vocab_size": 16384,
  "params": 50040000,
  "sha256": {
    "model.safetensors": "48e4d4e7b8af675d222e58e14776abac24b7e83a937d7884bd9d735ea9dd2173",
    "onnx/model_fp16.onnx": "945e1ac0c79c4de1807635171a67c1ef65fac8dee9476ca989116a19978d1012",
    "onnx/model_int8.onnx": "7826c5f46be23f17838c7c3478d24f6ea0e29ebe388cae92a102cb125d5b1c19",
    "spm_bashkir_bert_16k.model": "4e77b9cfc9ab8c200406ee46c2d20fc46734da9ddcf01dff0f0e370f6c4ac244",
    "config.json": "59b773953af5b2a69adb7e3c5d345a90f34f848f6a1543309d5acc7fb78b28c4",
    "configuration_bashkir_roberta.py": "e9f179aad7619153724920c86a90ecc3c161fd15ffc3332f200e9f2073a07f72",
    "modeling_bashkir_roberta.py": "3c1e098df93d6b47bc429268e87351afbaabf7d9a0be1f8e7ef9b9a3558410e8",
    "tokenization_bashkir_roberta.py": "c26cefc431c25c359ddcd1f6587ab132a4240c12bb44151b2afb16a220610fa9",
    "tokenizer_config.json": "3e3e87f99f7c91ee6d552da57d5b3e978d847440c5b238a7d3f7df1514c7073c"
  },
  "source_policy": "Training texts are not redistributed in this repository."
}