--- license: apache-2.0 library_name: transformers base_model: google/gemma-4-26B-A4B base_model_relation: finetune --- # [AIKAR 1.2 Pro] πŸš€ ![Model License](https://img.shields.io/badge/License-Apache%202.0-blue.svg) ![Model Type](https://img.shields.io/badge/Type-Large%20Language%20Model-orange.svg) ![Training Framework](https://img.shields.io/badge/Framework-PyTorch-red.svg) ![Developer](https://img.shields.io/badge/Developer-LOOP-green.svg) ## 🌟 Overview **AIKAR 1.2 Pro**λŠ” LOOPμ—μ„œ κ°œλ°œν•œ μ°¨μ„ΈλŒ€ κ³ μ„±λŠ₯ λŒ€κ·œλͺ¨ μ–Έμ–΄ λͺ¨λΈ(LLM) μ‹œλ¦¬μ¦ˆμ˜ μ •μ μž…λ‹ˆλ‹€. 이전 λͺ¨λΈμΈ AIKAR 1.1의 μ•„ν‚€ν…μ²˜λ₯Ό κ³„μŠΉν•˜λ©΄μ„œλ„, λ”μš± λ°©λŒ€ν•˜κ³  μ •κ΅ν•œ 데이터셋을 ν†΅ν•œ 지속적인 ν•™μŠ΅(Continuous Training)을 톡해 μΆ”λ‘  λŠ₯λ ₯, λ‹€κ΅­μ–΄ 처리 μ„±λŠ₯, 그리고 볡합적인 λͺ…λ Ήμ–΄ μ€€μˆ˜ λŠ₯λ ₯을 λΉ„μ•½μ μœΌλ‘œ ν–₯μƒμ‹œμΌ°μŠ΅λ‹ˆλ‹€. λ³Έ λͺ¨λΈμ€ 개발자 **DFveloper**의 λΉ„μ „ μ•„λž˜, 싀무 ν™˜κ²½μ—μ„œμ˜ 높은 λ²”μš©μ„±κ³Ό μ •λ°€ν•œ 응닡 생성을 λͺ©ν‘œλ‘œ μ„€κ³„λ˜μ—ˆμŠ΅λ‹ˆλ‹€. ## ✨ Key Features - **Advanced Reasoning**: λ³΅μž‘ν•œ 논리적 μΆ”λ‘  및 μˆ˜ν•™μ  문제 ν•΄κ²° λŠ₯λ ₯ κ°•ν™”. - **Enhanced Instruction Following**: μ‚¬μš©μžμ˜ λ―Έμ„Έν•œ λ‰˜μ•™μŠ€λ₯Ό νŒŒμ•…ν•˜κ³  μ˜λ„μ— λΆ€ν•©ν•˜λŠ” μ •ν™•ν•œ κ²°κ³Όλ¬Ό λ„μΆœ. - **Multilingual Excellence**: ν•œκ΅­μ–΄ 및 μ˜μ–΄ λ“± λ‹€μ–‘ν•œ μ–Έμ–΄ κ°„μ˜ μžμ—°μŠ€λŸ¬μš΄ μ „ν™˜ 및 λ¬Έλ§₯ μœ μ§€ λŠ₯λ ₯ μ΅œμ ν™”. - **Optimized Efficiency**: Pro λͺ¨λΈλ‘œμ„œ μΆ”λ‘  μ„±λŠ₯κ³Ό μ—°μ‚° νš¨μœ¨μ„± μ‚¬μ΄μ˜ 졜적의 κ· ν˜• 달성. - **Contextual Awareness**: κΈ΄ λŒ€ν™” λ§₯λ½μ—μ„œλ„ μ •λ³΄μ˜ 일관성을 μœ μ§€ν•˜λŠ” κ°•λ ₯ν•œ Context Window 관리. ## πŸ›  Training Details - **Base**: Thanks to Google, Gemma 4 26B A4B - **Developer**: LOOP (Lead Developer: DFveloper) - **Architecture**: Gemma 4 26B A4B - **Dataset**: High-quality curated web text, code, mathematical reasoning datasets, and instruction-tuning datasets. ## πŸš€ Quick Start (Usage) Hugging Face의 `transformers` 라이브러리λ₯Ό μ‚¬μš©ν•˜μ—¬ λͺ¨λΈμ„ λ‘œλ“œν•˜κ³  μ‹€ν–‰ν•˜λŠ” 방법은 λ‹€μŒκ³Ό κ°™μŠ΅λ‹ˆλ‹€. ```python from transformers import AutoModelForCausalLM, AutoTokenizer import torch model_id = "DFveloper/AIKAR-1.2-Pro" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16, device_map="auto" ) prompt = "Tell me a story." inputs = tokenizer(prompt, return_tensors="pt").to("cuda") outputs = model.generate(**inputs, max_new_tokens=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` ## 🀝 Contributing AIKAR 1.2 Pro의 μ„±λŠ₯ κ°œμ„ μ΄λ‚˜ 버그 μ œλ³΄λŠ” [LOOP GitHub Repository](https://github.com/LOOP-dev)λ₯Ό 톡해 μ–Έμ œλ“  ν™˜μ˜ν•©λ‹ˆλ‹€. μ‚¬μš©μžμ˜ ν”Όλ“œλ°±μ€ μ°¨μ„ΈλŒ€ λͺ¨λΈ 개발의 핡심 μžμ‚°μ΄ λ©λ‹ˆλ‹€. ## πŸ“œ License This model is released under the **Apache License 2.0**. --- **"The journey of intelligence never ends. We move forward, one token at a time."** *β€” Developed by LOOP*