theprint / Rewiz-Tom-7B

Rewiz-Tom-7B

A fine-tuned 7B parameter model specialized in reasoning (Rewiz), based on a model that was already finetuned for step-by-step instruction and conversation (Tom).

Model Details

This model is a fine-tuned version of theprint/Tom-Qwen-7B-Instruct using the Unsloth framework with LoRA (Low-Rank Adaptation) for efficient training.

• Developed by: theprint
• Model type: Causal Language Model (Fine-tuned with LoRA)
• Language: en
• License: apache-2.0
• Base model: theprint/Tom-Qwen-7B-Instruct
• Fine-tuning method: LoRA with rank 128

Intended Use

Conversation, brainstorming, and general instruction following

Training Details

Training Data

The Rewiz data set is a curated mix of 20,000 reasoning-based entries.

• Dataset: theprint/ReWiz
• Format: alpaca

Training Procedure

• Training epochs: 2
• LoRA rank: 128
• Learning rate: 0.0002
• Batch size: 4
• Framework: Unsloth + transformers + PEFT
• Hardware: NVIDIA RTX 5090

Usage

from unsloth import FastLanguageModel
import torch

# Load model and tokenizer
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name="theprint/Rewiz-Tom-7B",
    max_seq_length=4096,
    dtype=None,
    load_in_4bit=True,
)

# Enable inference mode
FastLanguageModel.for_inference(model)

# Example usage
inputs = tokenizer(["Your prompt here"], return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=256, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Alternative Usage (Standard Transformers)

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model = AutoModelForCausalLM.from_pretrained(
    "theprint/Rewiz-Tom-7B",
    torch_dtype=torch.float16,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("theprint/Rewiz-Tom-7B")

# Example usage
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Your question here"}
]

inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True)
outputs = model.generate(inputs, max_new_tokens=256, temperature=0.7, do_sample=True)
response = tokenizer.decode(outputs[0][inputs.shape[-1]:], skip_special_tokens=True)
print(response)

GGUF Quantized Versions

Quantized GGUF versions are available in the gguf/ directory for use with llama.cpp:

• Rewiz-Tom-7B-f16.gguf (14531.9 MB) - 16-bit float (original precision, largest file)
• Rewiz-Tom-7B-q3_k_m.gguf (3632.0 MB) - 3-bit quantization (medium quality)
• Rewiz-Tom-7B-q4_k_m.gguf (4466.1 MB) - 4-bit quantization (medium, recommended for most use cases)
• Rewiz-Tom-7B-q5_k_m.gguf (5192.6 MB) - 5-bit quantization (medium, good quality)
• Rewiz-Tom-7B-q6_k.gguf (5964.5 MB) - 6-bit quantization (high quality)
• Rewiz-Tom-7B-q8_0.gguf (7723.4 MB) - 8-bit quantization (very high quality)

Using with llama.cpp

# Download a quantized version (q4_k_m recommended for most use cases)
wget https://huggingface.co/theprint/Rewiz-Tom-7B/resolve/main/gguf/Rewiz-Tom-7B-q4_k_m.gguf

# Run with llama.cpp
./llama.cpp/main -m Rewiz-Tom-7B-q4_k_m.gguf -p "Your prompt here" -n 256

Limitations

May hallucinate or provide incorrect information.

Citation

If you use this model, please cite:

@misc{rewiz_tom_7b,
  title={Rewiz-Tom-7B: Fine-tuned theprint/Tom-Qwen-7B-Instruct},
  author={theprint},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/theprint/Rewiz-Tom-7B}
}

Acknowledgments

• Base model: theprint/Tom-Qwen-7B-Instruct
• Training dataset: theprint/ReWiz
• Fine-tuning framework: Unsloth
• Quantization: llama.cpp