aokitools/japanese-laws-egov-instruct-202508031857

Experimental model in research stage

Quickstart

If you're using Ollama, run the following command first, then restart the Ollama app and select the newly added model.

ollama pull hf.co/aokitools/japanese-laws-egov-instruct-202508031857

If you want to remove it, run the following command:

ollama list
ollama rm hf.co/aokitools/japanese-laws-egov-instruct-202508031857:latest
ollama list

To use it from Python, use the following code.

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig

model_name = "aokitools/japanese-laws-egov-instruct-202508031857"

quant_config = BitsAndBytesConfig(
    load_in_8bit=True,
    llm_int8_threshold=6.0,
)

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.float16,
    device_map="auto",
    quantization_config=quant_config,
)

# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=256
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

# parsing thinking content
try:
    # rindex finding 151668 (</think>)
    index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
    index = 0

thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")

print("thinking content:", thinking_content)
print("content:", content)

This model is a continual pretraining of Qwen/Qwen3-1.7B.

Training details

• Base model: Qwen3-1.7B
• Tokenizer: QwenTokenizer

License

• Apache 2.0 + Alibaba Qianwen License