🚀 Qwen3.5-2B-ReMix (Reasoning Mix) 🧠

This repository contains a fully merged, native Float16 (F16) fine-tune of Qwen/Qwen3.5-2B 🤖. The primary objective of this model is to significantly scale up performance on complex reasoning tasks, specifically targeting advanced mathematics 🧮, logical deduction, and structured coding problems 💻.

By leveraging multi-source open-source distillation data, it aims to achieve "frontier-style" reasoning capabilities while keeping the footprint compact enough to run smoothly at native speeds on local, everyday consumer hardware 🏠 without the need for external adapters.

🌟 Model Highlights

• 🏗️ Base Architecture: Qwen/Qwen3.5-2B (Dense, Hybrid Gated DeltaNet)
• 💾 Precision format: Native Float16 (F16) Merged Weights — No adapter required!
• 🎯 Main Goal: Advanced mathematical reasoning and complex code generation/debugging.
• 🛡️ Data Origin: 100% open-source distilled reasoning datasets natively hosted on Hugging Face. No proprietary data or closed APIs (OpenAI, Anthropic, Google) were used or involved in the collection or training process.
• ⚡ Target Environment: Local, high-efficiency edge execution with minimal hardware requirements.

🎛️ Recommended Generation Parameters

Depending on your use case, we recommend switching between "Everyday" and "Deep Reasoning" profiles to get the best performance out of the 2B architecture.

🏠 Everyday Use (Balanced)

🧠 Deep Reasoning

📊 Training & Merge Details

The model was adapted using Parameter-Efficient Fine-Tuning (PEFT) and then compiled back into the core network layers to output clean, unified F16 weights via Unsloth.

• 🔄 Training Steps: 175
• 📉 Loss Profile: Convergence floor reached ~0.58; stabilized consistently around 0.85
• 📈 Learning Rate: 4e-5
• 📐 LoRA Rank ($R$) during training: 16
• ⚖️ LoRA Alpha ($\alpha$) during training: 32

⚠️ Limitations & Risks

While this fine-tune aggressively pushes the boundaries of what a 2B parameter model can achieve locally, users should carefully account for the following behaviors:

• 🔮 Hallucinations: Like all highly compact models, it can confidently present false calculations or flawed code as absolute facts. Always verify outputs.
• 🎭 Inconsistent Styles: Due to the "ReMix" nature of the training data, the model may occasionally exhibit shifting output structures or stylistic variations.
• 🛑 Logic Mismatches: For extremely niche programming or high-level academic proofs, the model may occasionally produce broken syntax or reverse its logical assertions.

📦 How to Use Natively

🐍 Using Hugging Face Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_path = "YOUR_USERNAME/Qwen3.5-2B-ReMix"

# Load the aligned tokenizer and model weights directly
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(
 model_path, 
 torch_dtype=torch.float16, 
 device_map="auto"
)

messages = [
 {"role": "user", "content": "Explain the logic of a quicksort algorithm and implement it in Python."}
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# Using Reasoning Parameters (To not overthink)
generated_ids = model.generate(
 **model_inputs,
 max_new_tokens=1024,
 temperature=0.1, 
 top_k=60, 
 repeat_penalty=1.2 
)

Uploaded finetuned model

• Developed by: ertghiu256
• License: apache-2.0
• Finetuned from model : unsloth/Qwen3.5-2B

This qwen3_5 model was trained 2x faster with Unsloth and Huggingface's TRL library.

👁 Image

Base model