chat-bricks 0.1.0

A helper library for LLM/VLM chat templates.

🧩 Chat Bricks

Jinja Template is Not You Need!

Chat Bricks is a powerful and flexible template system inspired by building block toys, designed to support various LLM and VLM chat templates for training and inference.

Key Features

• Training and Inference: Chat template formatted prompts, with tokenized inputs and masks.
• Modular design: Templates are built from configurable components.
• Multi-modal support: Built-in vision-language templates.
• Jinja template generation: Automatic HuggingFace-compatible template generation.
• HuggingFace Integration: Directly supports using an HF repo id as template.
• Advanced configuration: Fine-grained control over template behavior.

Installation

pip install chat-bricks

Quick Start

Basic Usage

Create a chat object with a built-in template and render the prompt:

from chat_bricks import Chat

# Create a chat object with template and messages
chat = Chat(
    template="qwen3",
    messages=[
        {"role": "user", "content": "Hello, how are you?"},
        {"role": "assistant", "content": "I am fine, thank you."}
    ],
)

# Render the final prompt
prompt = chat.render()
print(prompt)

Tokenization for Training/Inference

You can easily tokenize messages for model input:

from transformers import AutoTokenizer
from chat_bricks import Chat

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-3B-Instruct")
chat = Chat(template="qwen2.5", messages=[{"role": "user", "content": "Hello!"}])

inputs = chat.tokenize(
    tokenizer,
    add_generation_prompt=True,  # keep generation token for inference
)

print(inputs["input_ids"])

Custom Templates

Define your own template format using the Template class:

from chat_bricks import Chat, Template

custom = Template(
    name="my-template",
    system_template="<|im_start|>system\n{system_message}<|im_end|>\n",
    system_message="You are a concise assistant.",
    user_template="<|im_start|>user\n{content}<|im_end|>\n",
    assistant_template="<|im_start|>assistant\n{content}<|im_end|>\n",
    stop_words=["<|im_end|>"],
)

chat = Chat(template=custom, messages=[{"role": "user", "content": "Hi!"}])
print(chat.render())

Using HuggingFace Repo ID as Template

You can directly use any HuggingFace model repository ID as a template. Chat Bricks will automatically load the tokenizer's chat template:

from chat_bricks import Chat

# Use a HuggingFace repo id directly
chat = Chat(
    template="Qwen/Qwen2.5-3B-Instruct",
    messages=[
        {"role": "user", "content": "Hello, how are you?"},
        {"role": "assistant", "content": "I am fine, thank you."}
    ],
)

# Render the prompt using the model's native chat template
prompt = chat.prompt()
print(prompt)
prompt_with_mask = chat.prompt_with_mask()
print(prompt_with_mask)

# Tokenize with proper masking for training
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-3B-Instruct")
inputs = chat.tokenize(tokenizer, add_generation_prompt=True)

This feature automatically detects if the repo ID is not a built-in template and creates an HFTemplate that uses the tokenizer's chat template. It supports tools, generation prompts, and proper masking for training. See the HuggingFace Templates Guide for more details.

Documentation

For full documentation, please visit our docs (or run mkdocs serve locally).

Community

WeChat	Discord
Scan to join wechat group	Join our discord channel