import json
import os

# 定义配置参数
config_data = {
    "hidden_size": 768,
    "num_attention_heads": 12,
    "num_hidden_layers": 12,
    "intermediate_size": 3072,
    "hidden_dropout_prob": 0.1,
    "attention_probs_dropout_prob": 0.1,
    "image_size": 224,
    "image_channels": 3,
    "patch_size": 16,
    "max_position_embeddings": 512,
    "vocab_size": 30522,
    "type_vocab_size": 2,
    "audio_sample_rate": 16000,
    "audio_frame_size": 1024,
    "audio_hop_size": 512,
    "enable_vqa": True,
    "enable_caption": True,
    "enable_retrieval": True,
    "enable_asr": True,
    "enable_realtime_asr": True,
    "batch_size": 32,
    "learning_rate": 0.0001,
    "weight_decay": 0.01,
    "warmup_steps": 10000,
    "max_steps": 100000
}

# 文件路径
config_path = r"C:\Users\baby7\Desktop\zero_sg-pytorch-zero-v4\config.json"

# 保存配置文件
os.makedirs(os.path.dirname(config_path), exist_ok=True)
with open(config_path, "w") as f:
    json.dump(config_data, f, indent=4)

print(f"配置文件已保存到: {config_path}")

from transformers import BertTokenizer
import os

# 初始化分词器
tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")

# 保存分词器到目标路径
tokenizer_path = r"C:\Users\baby7\Desktop\zero_sg-pytorch-zero-v4\tokenizer"
os.makedirs(tokenizer_path, exist_ok=True)
tokenizer.save_pretrained(tokenizer_path)

print(f"分词器已保存到: {tokenizer_path}")


#### **加载配置文件**
from model import Config  # 假设您有Config类

config_file = r"C:\Users\baby7\Desktop\zero_sg-pytorch-zero-v4\config.json"
config = Config(config_file)
print("加载的配置: ", config.__dict__)

from transformers import BertTokenizer

tokenizer_path = r"C:\Users\baby7\Desktop\zero_sg-pytorch-zero-v4\tokenizer"
tokenizer = BertTokenizer.from_pretrained(tokenizer_path)
text = "Hello, how are you?"
encoded_input = tokenizer(text, return_tensors="pt", max_length=512, padding="max_length", truncation=True)

print("分词器输出: ", encoded_input["input_ids"].shape)