text_encoder/config.json

{
    "architectures": ["CLIPTextModel"],
    "hidden_size": 1280,
    "intermediate_size": 5120,
    "num_hidden_layers": 24,
    "num_attention_heads": 16,
    "max_position_embeddings": 77,
    "hidden_act": "gelu",
    "layer_norm_eps": 1e-12,
    "dropout": 0.1,
    "vocab_size": 49408,
    "projection_dim": 1280,
    "pad_token_id": 1,
    "bos_token_id": 0,
    "eos_token_id": 2
}