Thinkagain1.5

🚀 ThinkAgain1.5

ThinkAgain1.5 is a powerful language model with natural reasoning and tool - calling support, offering a great experience in multi - language scenarios.

🚀 Quick Start

This section provides a quick start guide on how to use the ThinkAgain1.5 model.

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

MAX_REASONING_TOKENS = 4096
MAX_RESPONSE_TOKENS = 1024

model_name = "beyoru/ThinkAgain1.5"

model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, device_map="auto")
tokenizer = AutoTokenizer.from_pretrained(model_name)

messages = []

def stream_output(output_text):
    for char in output_text:
        print(char, end="", flush=True)

while True:
    prompt = input("USER: ")
    
    messages.append({"role": "user", "content": prompt})
    
    # Generate reasoning
    reasoning_template = tokenizer.apply_chat_template(messages, tokenize=False, add_reasoning_prompt=True)
    reasoning_inputs = tokenizer(reasoning_template, return_tensors="pt").to(model.device)
    reasoning_ids = model.generate(**reasoning_inputs, max_new_tokens=MAX_REASONING_TOKENS)
    reasoning_output = tokenizer.decode(reasoning_ids[0, reasoning_inputs.input_ids.shape[1]:], skip_special_tokens=True)
    
    messages.append({"role": "reasoning", "content": reasoning_output})
    
    print("REASONING: ", end="")
    stream_output(reasoning_output)
    print()
    
    # Generate answer
    response_template = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    response_inputs = tokenizer(response_template, return_tensors="pt").to(model.device)
    response_ids = model.generate(**response_inputs, max_new_tokens=MAX_RESPONSE_TOKENS)
    response_output = tokenizer.decode(response_ids[0, response_inputs.input_ids.shape[1]:], skip_special_tokens=True)
    
    messages.append({"role": "assistant", "content": response_output})
    
    print("ASSISTANT: ", end="")
    stream_output(response_output)
    print()

✨ Features

• Natural and Smarter Reasoning: The model can perform natural and intelligent reasoning.
• No System Prompt Training: It doesn't rely on system prompt training.
• LoRA Training: Conducted LoRA training with a rank of 16 and an alpha of 16.
• Tool Calling Support: Supports tool - calling functionality.
• Multi - language Support: Supports multiple languages including Chinese, English, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, and Arabic.

📚 Documentation

Model detail

• Reasoning: It can reason naturally and more intelligently.
• Training: No system prompt training was used. LoRA training was carried out with a rank of 16 and an alpha of 16.
• Functionality: It supports tool calling.
• Caution: Quantizing this model may not yield the best performance.

Model Information

📄 License

This model is licensed under the apache - 2.0 license.