text_encoder/model_config.json

{
  "image_resolution": 224,
  "vision_layers": 24,
  "vision_heads": 16,
  "vision_mlp_ratio": 4,
  "vision_embed_dim": 1024,
  "vision_patch_size": 14,
  "vision_hidden_act": "quick_gelu",
  "max_text_length": 77,
  "vocab_size": 49408,
  "text_embed_dim": 768,
  "text_heads": 12,
  "text_layers": 12,
  "text_hidden_act": "quick_gelu",
  "projection_dim": 768,
  "initializer_range": 0.02,
  "logit_scale_init_value": 2.6592,
  "paddle_dtype": null,
  "init_class": "CLIPTextModel"
}