config.json

{
  "beit_version": "large",
  "encoder_embed_dim": 1024,
  "out_embed_dim": 1024,
  "image_size": 224,
  "visual_mask_size": 14,
  "loss_names": {
    "itc": 1
  },
  "encoder_layers": 21,
  "beit3_vl_layers": 3,
  "tokenizer_type": "GLMChineseTokenizer",
  "tokenizer": ".",
  "vocab_size": 115244,
  "whole_word_masking": false,
  "precision": 32,
  "test_only": true,
  "flash_attn": false,
  "modelscope": {
    "model_id": "M2Cognition/M2_Encoder_Large"
  },
  "model_file": "m2_encoder_1B.safetensors",
  "model_type": "m2_encoder",
  "architectures": [
    "M2EncoderModel"
  ],
  "processor_class": "M2EncoderProcessor",
  "auto_map": {
    "AutoConfig": "configuration_m2_encoder.M2EncoderConfig",
    "AutoModel": "modeling_m2_encoder.M2EncoderModel",
    "AutoProcessor": "processing_m2_encoder.M2EncoderProcessor"
  }
}