import torch
from transformers import PreTrainedModel, PreTrainedTokenizerFast, PretrainedConfig, CausalLMOutput

# Define the model configuration
class HelloWorldConfig(PretrainedConfig):
    model_type = "hello-world"
    vocab_size = 2
    bos_token_id = 0
    eos_token_id = 1

# Define the model
class HelloWorldModel(PreTrainedModel):
    config_class = HelloWorldConfig

    def __init__(self, config):
        super().__init__(config)

    def forward(self, input_ids=None, **kwargs):
        batch_size = input_ids.shape[0]
        sequence_length = input_ids.shape[1]

        # Generate logits for the "Hello, world!" token
        hello_world_token_id = self.config.vocab_size - 1
        logits = torch.full((batch_size, sequence_length, self.config.vocab_size), float('-inf'))
        logits[:, :, hello_world_token_id] = 0

        return CausalLMOutput(logits=logits)

# Define and save the tokenizer
tokenizer = PreTrainedTokenizerFast(tokenizer_file="tokenizer.json")
tokenizer.add_tokens(["Hello, world!"])

tokenizer_config = {
    "do_lower_case": False,
    "model_max_length": 512,
    "padding_side": "right",
    "special_tokens_map_file": None,
    "tokenizer_file": "tokenizer.json",
    "unk_token": "<unk>",
    "bos_token": "<s>",
    "eos_token": "</s>",
    "vocab_size": 2,
}

with open("tokenizer.json", "w") as f:
    json.dump(tokenizer_config, f)

# Initialize model
config = HelloWorldConfig()
model = HelloWorldModel(config)

# Save model using safetensors format
from safetensors.torch import save_file
save_file(model.state_dict(), "hello_world_model.safetensors")