From e767f09819dc26d13dc871305d5d99ccde46903a Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Sun, 9 Aug 2026 21:09:16 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: Mungert/kanana-1.5-8b-instruct-2505-GGUF Source: Original Platform --- .gitattributes | 76 ++++ README.md | 445 +++++++++++++++++++++ kanana-1.5-8b-instruct-2505-bf16.gguf | 3 + kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf | 3 + kanana-1.5-8b-instruct-2505-f16_q8_0.gguf | 3 + kanana-1.5-8b-instruct-2505-iq2_m.gguf | 3 + kanana-1.5-8b-instruct-2505-iq2_s.gguf | 3 + kanana-1.5-8b-instruct-2505-iq2_xs.gguf | 3 + kanana-1.5-8b-instruct-2505-iq2_xxs.gguf | 3 + kanana-1.5-8b-instruct-2505-iq3_m.gguf | 3 + kanana-1.5-8b-instruct-2505-iq3_s.gguf | 3 + kanana-1.5-8b-instruct-2505-iq3_xs.gguf | 3 + kanana-1.5-8b-instruct-2505-iq3_xxs.gguf | 3 + kanana-1.5-8b-instruct-2505-iq4_nl.gguf | 3 + kanana-1.5-8b-instruct-2505-iq4_xs.gguf | 3 + kanana-1.5-8b-instruct-2505-q2_k_m.gguf | 3 + kanana-1.5-8b-instruct-2505-q2_k_s.gguf | 3 + kanana-1.5-8b-instruct-2505-q3_k_m.gguf | 3 + kanana-1.5-8b-instruct-2505-q3_k_s.gguf | 3 + kanana-1.5-8b-instruct-2505-q4_0.gguf | 3 + kanana-1.5-8b-instruct-2505-q4_1.gguf | 3 + kanana-1.5-8b-instruct-2505-q4_k_m.gguf | 3 + kanana-1.5-8b-instruct-2505-q4_k_s.gguf | 3 + kanana-1.5-8b-instruct-2505-q5_0.gguf | 3 + kanana-1.5-8b-instruct-2505-q5_1.gguf | 3 + kanana-1.5-8b-instruct-2505-q5_k_m.gguf | 3 + kanana-1.5-8b-instruct-2505-q5_k_s.gguf | 3 + kanana-1.5-8b-instruct-2505-q6_k_m.gguf | 3 + kanana-1.5-8b-instruct-2505-q8_0.gguf | 3 + kanana-1.5-8b-instruct-2505.imatrix | 3 + 30 files changed, 605 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 kanana-1.5-8b-instruct-2505-bf16.gguf create mode 100644 kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf create mode 100644 kanana-1.5-8b-instruct-2505-f16_q8_0.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq2_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq2_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq2_xs.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq2_xxs.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq3_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq3_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq3_xs.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq3_xxs.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq4_nl.gguf create mode 100644 kanana-1.5-8b-instruct-2505-iq4_xs.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q2_k_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q2_k_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q3_k_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q3_k_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q4_0.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q4_1.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q4_k_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q4_k_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q5_0.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q5_1.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q5_k_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q5_k_s.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q6_k_m.gguf create mode 100644 kanana-1.5-8b-instruct-2505-q8_0.gguf create mode 100644 kanana-1.5-8b-instruct-2505.imatrix diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..a499bdb --- /dev/null +++ b/.gitattributes @@ -0,0 +1,76 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-f16.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-f16_q8_0.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-f16_q6_k.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-bf16_q6_k.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-f16_q4_k.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-bf16_q4_k.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q2_k_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q3_k_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_k_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_k_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q6_k_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q2_k_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q2_k_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q3_k_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q3_k_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_k_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_k_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_k_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_k_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q6_k_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q8_0.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_0.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_1.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_0_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q4_1_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_0.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_1.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_0_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-q5_1_l.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq2_xs.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq2_xxs.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq2_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq2_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq3_xs.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq3_xxs.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq3_s.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq3_m.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-iq4_nl.gguf filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505.imatrix filter=lfs diff=lfs merge=lfs -text +kanana-1.5-8b-instruct-2505-bf16.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..f34d451 --- /dev/null +++ b/README.md @@ -0,0 +1,445 @@ +--- +language: +- en +- ko +library_name: transformers +license: apache-2.0 +pipeline_tag: text-generation +model_id: kakaocorp/kanana-1.5-8b-instruct-2505 +repo: kakaocorp/kanana-1.5-8b-instruct-2505 +developers: Kanana LLM +training_regime: bf16 mixed precision +--- + +# kanana-1.5-8b-instruct-2505 GGUF Models + + +## Model Generation Details + +This model was generated using [llama.cpp](https://github.com/ggerganov/llama.cpp) at commit [`f5cd27b7`](https://github.com/ggerganov/llama.cpp/commit/f5cd27b71da3ac375a04a41643d14fc779a8057b). + + + + +## Ultra-Low-Bit Quantization with IQ-DynamicGate (1-2 bit) + +Our latest quantization method introduces **precision-adaptive quantization** for ultra-low-bit models (1-2 bit), with benchmark-proven improvements on **Llama-3-8B**. This approach uses layer-specific strategies to preserve accuracy while maintaining extreme memory efficiency. + +### **Benchmark Context** +All tests conducted on **Llama-3-8B-Instruct** using: +- Standard perplexity evaluation pipeline +- 2048-token context window +- Same prompt set across all quantizations + +### **Method** +- **Dynamic Precision Allocation**: + - First/Last 25% of layers → IQ4_XS (selected layers) + - Middle 50% → IQ2_XXS/IQ3_S (increase efficiency) +- **Critical Component Protection**: + - Embeddings/output layers use Q5_K + - Reduces error propagation by 38% vs standard 1-2bit + +### **Quantization Performance Comparison (Llama-3-8B)** + +| Quantization | Standard PPL | DynamicGate PPL | Δ PPL | Std Size | DG Size | Δ Size | Std Speed | DG Speed | +|--------------|--------------|------------------|---------|----------|---------|--------|-----------|----------| +| IQ2_XXS | 11.30 | 9.84 | -12.9% | 2.5G | 2.6G | +0.1G | 234s | 246s | +| IQ2_XS | 11.72 | 11.63 | -0.8% | 2.7G | 2.8G | +0.1G | 242s | 246s | +| IQ2_S | 14.31 | 9.02 | -36.9% | 2.7G | 2.9G | +0.2G | 238s | 244s | +| IQ1_M | 27.46 | 15.41 | -43.9% | 2.2G | 2.5G | +0.3G | 206s | 212s | +| IQ1_S | 53.07 | 32.00 | -39.7% | 2.1G | 2.4G | +0.3G | 184s | 209s | + +**Key**: +- PPL = Perplexity (lower is better) +- Δ PPL = Percentage change from standard to DynamicGate +- Speed = Inference time (CPU avx2, 2048 token context) +- Size differences reflect mixed quantization overhead + +**Key Improvements:** +- 🔥 **IQ1_M** shows massive 43.9% perplexity reduction (27.46 → 15.41) +- 🚀 **IQ2_S** cuts perplexity by 36.9% while adding only 0.2GB +- ⚡ **IQ1_S** maintains 39.7% better accuracy despite 1-bit quantization + +**Tradeoffs:** +- All variants have modest size increases (0.1-0.3GB) +- Inference speeds remain comparable (<5% difference) + + +### **When to Use These Models** +📌 **Fitting models into GPU VRAM** + +✔ **Memory-constrained deployments** + +✔ **Cpu and Edge Devices** where 1-2bit errors can be tolerated + +✔ **Research** into ultra-low-bit quantization + + + +## **Choosing the Right Model Format** + +Selecting the correct model format depends on your **hardware capabilities** and **memory constraints**. + +### **BF16 (Brain Float 16) – Use if BF16 acceleration is available** +- A 16-bit floating-point format designed for **faster computation** while retaining good precision. +- Provides **similar dynamic range** as FP32 but with **lower memory usage**. +- Recommended if your hardware supports **BF16 acceleration** (check your device's specs). +- Ideal for **high-performance inference** with **reduced memory footprint** compared to FP32. + +📌 **Use BF16 if:** +✔ Your hardware has native **BF16 support** (e.g., newer GPUs, TPUs). +✔ You want **higher precision** while saving memory. +✔ You plan to **requantize** the model into another format. + +📌 **Avoid BF16 if:** +❌ Your hardware does **not** support BF16 (it may fall back to FP32 and run slower). +❌ You need compatibility with older devices that lack BF16 optimization. + +--- + +### **F16 (Float 16) – More widely supported than BF16** +- A 16-bit floating-point **high precision** but with less of range of values than BF16. +- Works on most devices with **FP16 acceleration support** (including many GPUs and some CPUs). +- Slightly lower numerical precision than BF16 but generally sufficient for inference. + +📌 **Use F16 if:** +✔ Your hardware supports **FP16** but **not BF16**. +✔ You need a **balance between speed, memory usage, and accuracy**. +✔ You are running on a **GPU** or another device optimized for FP16 computations. + +📌 **Avoid F16 if:** +❌ Your device lacks **native FP16 support** (it may run slower than expected). +❌ You have memory limitations. + +--- + +### **Quantized Models (Q4_K, Q6_K, Q8, etc.) – For CPU & Low-VRAM Inference** +Quantization reduces model size and memory usage while maintaining as much accuracy as possible. +- **Lower-bit models (Q4_K)** → **Best for minimal memory usage**, may have lower precision. +- **Higher-bit models (Q6_K, Q8_0)** → **Better accuracy**, requires more memory. + +📌 **Use Quantized Models if:** +✔ You are running inference on a **CPU** and need an optimized model. +✔ Your device has **low VRAM** and cannot load full-precision models. +✔ You want to reduce **memory footprint** while keeping reasonable accuracy. + +📌 **Avoid Quantized Models if:** +❌ You need **maximum accuracy** (full-precision models are better for this). +❌ Your hardware has enough VRAM for higher-precision formats (BF16/F16). + +--- + +### **Very Low-Bit Quantization (IQ3_XS, IQ3_S, IQ3_M, Q4_K, Q4_0)** +These models are optimized for **extreme memory efficiency**, making them ideal for **low-power devices** or **large-scale deployments** where memory is a critical constraint. + +- **IQ3_XS**: Ultra-low-bit quantization (3-bit) with **extreme memory efficiency**. + - **Use case**: Best for **ultra-low-memory devices** where even Q4_K is too large. + - **Trade-off**: Lower accuracy compared to higher-bit quantizations. + +- **IQ3_S**: Small block size for **maximum memory efficiency**. + - **Use case**: Best for **low-memory devices** where **IQ3_XS** is too aggressive. + +- **IQ3_M**: Medium block size for better accuracy than **IQ3_S**. + - **Use case**: Suitable for **low-memory devices** where **IQ3_S** is too limiting. + +- **Q4_K**: 4-bit quantization with **block-wise optimization** for better accuracy. + - **Use case**: Best for **low-memory devices** where **Q6_K** is too large. + +- **Q4_0**: Pure 4-bit quantization, optimized for **ARM devices**. + - **Use case**: Best for **ARM-based devices** or **low-memory environments**. + +--- + +### **Summary Table: Model Format Selection** + +| Model Format | Precision | Memory Usage | Device Requirements | Best Use Case | +|--------------|------------|---------------|----------------------|---------------| +| **BF16** | Highest | High | BF16-supported GPU/CPUs | High-speed inference with reduced memory | +| **F16** | High | High | FP16-supported devices | GPU inference when BF16 isn't available | +| **Q4_K** | Medium Low | Low | CPU or Low-VRAM devices | Best for memory-constrained environments | +| **Q6_K** | Medium | Moderate | CPU with more memory | Better accuracy while still being quantized | +| **Q8_0** | High | Moderate | CPU or GPU with enough VRAM | Best accuracy among quantized models | +| **IQ3_XS** | Very Low | Very Low | Ultra-low-memory devices | Extreme memory efficiency and low accuracy | +| **Q4_0** | Low | Low | ARM or low-memory devices | llama.cpp can optimize for ARM devices | + +--- + +## **Included Files & Details** + +### `kanana-1.5-8b-instruct-2505-bf16.gguf` +- Model weights preserved in **BF16**. +- Use this if you want to **requantize** the model into a different format. +- Best if your device supports **BF16 acceleration**. + +### `kanana-1.5-8b-instruct-2505-f16.gguf` +- Model weights stored in **F16**. +- Use if your device supports **FP16**, especially if BF16 is not available. + +### `kanana-1.5-8b-instruct-2505-bf16-q8_0.gguf` +- **Output & embeddings** remain in **BF16**. +- All other layers quantized to **Q8_0**. +- Use if your device supports **BF16** and you want a quantized version. + +### `kanana-1.5-8b-instruct-2505-f16-q8_0.gguf` +- **Output & embeddings** remain in **F16**. +- All other layers quantized to **Q8_0**. + +### `kanana-1.5-8b-instruct-2505-q4_k.gguf` +- **Output & embeddings** quantized to **Q8_0**. +- All other layers quantized to **Q4_K**. +- Good for **CPU inference** with limited memory. + +### `kanana-1.5-8b-instruct-2505-q4_k_s.gguf` +- Smallest **Q4_K** variant, using less memory at the cost of accuracy. +- Best for **very low-memory setups**. + +### `kanana-1.5-8b-instruct-2505-q6_k.gguf` +- **Output & embeddings** quantized to **Q8_0**. +- All other layers quantized to **Q6_K** . + +### `kanana-1.5-8b-instruct-2505-q8_0.gguf` +- Fully **Q8** quantized model for better accuracy. +- Requires **more memory** but offers higher precision. + +### `kanana-1.5-8b-instruct-2505-iq3_xs.gguf` +- **IQ3_XS** quantization, optimized for **extreme memory efficiency**. +- Best for **ultra-low-memory devices**. + +### `kanana-1.5-8b-instruct-2505-iq3_m.gguf` +- **IQ3_M** quantization, offering a **medium block size** for better accuracy. +- Suitable for **low-memory devices**. + +### `kanana-1.5-8b-instruct-2505-q4_0.gguf` +- Pure **Q4_0** quantization, optimized for **ARM devices**. +- Best for **low-memory environments**. +- Prefer IQ4_NL for better accuracy. + +# 🚀 If you find these models useful +❤ **Please click "Like" if you find this useful!** +Help me test my **AI-Powered Network Monitor Assistant** with **quantum-ready security checks**: +👉 [Quantum Network Monitor](https://readyforquantum.com/dashboard/?assistant=open&utm_source=huggingface&utm_medium=referral&utm_campaign=huggingface_repo_readme) + +💬 **How to test**: + Choose an **AI assistant type**: + - `TurboLLM` (GPT-4o-mini) + - `HugLLM` (Hugginface Open-source) + - `TestLLM` (Experimental CPU-only) + +### **What I’m Testing** +I’m pushing the limits of **small open-source models for AI network monitoring**, specifically: +- **Function calling** against live network services +- **How small can a model go** while still handling: + - Automated **Nmap scans** + - **Quantum-readiness checks** + - **Network Monitoring tasks** + +🟡 **TestLLM** – Current experimental model (llama.cpp on 2 CPU threads): +- ✅ **Zero-configuration setup** +- ⏳ 30s load time (slow inference but **no API costs**) +- 🔧 **Help wanted!** If you’re into **edge-device AI**, let’s collaborate! + +### **Other Assistants** +🟢 **TurboLLM** – Uses **gpt-4o-mini** for: +- **Create custom cmd processors to run .net code on Quantum Network Monitor Agents** +- **Real-time network diagnostics and monitoring** +- **Security Audits** +- **Penetration testing** (Nmap/Metasploit) + + +🔵 **HugLLM** – Latest Open-source models: +- 🌐 Runs on Hugging Face Inference API + +### 💡 **Example commands to you could test**: +1. `"Give me info on my websites SSL certificate"` +2. `"Check if my server is using quantum safe encyption for communication"` +3. `"Run a comprehensive security audit on my server"` +4. '"Create a cmd processor to .. (what ever you want)" Note you need to install a Quantum Network Monitor Agent to run the .net code from. This is a very flexible and powerful feature. Use with caution! + +### Final Word + +I fund the servers used to create these model files, run the Quantum Network Monitor service, and pay for inference from Novita and OpenAI—all out of my own pocket. All the code behind the model creation and the Quantum Network Monitor project is [open source](https://github.com/Mungert69). Feel free to use whatever you find helpful. + +If you appreciate the work, please consider [buying me a coffee](https://www.buymeacoffee.com/mahadeva) ☕. Your support helps cover service costs and allows me to raise token limits for everyone. + +I'm also open to job opportunities or sponsorship. + +Thank you! 😊 + + + + +

+
+ + + +
+ +

+🤗 1.5 HF Models   | +  📕 1.5 Blog   | +  📜 Technical Report + + +
+ +## News 🔥 + +- ✨`2025/05/23`: Published a [blog post](https://tech.kakao.com/posts/707) about `Kanana 1.5` models and released 🤗[HF model weights](https://kko.kakao.com/kananallm). +- 📜`2025/02/27`: Released [Technical Report](https://arxiv.org/abs/2502.18934) and 🤗[HF model weights](https://huggingface.co/collections/kakaocorp/kanana-nano-21b-67a326cda1c449c8d4172259). +- 📕`2025/01/10`: Published a [blog post](https://tech.kakao.com/posts/682) about the development of `Kanana Nano` model. +- 📕`2024/11/14`: Published blog posts ([pre-training](https://tech.kakao.com/posts/661), [post-training](https://tech.kakao.com/posts/662)) about the development of `Kanana` models. +- ▶️`2024/11/06`: Published a [presentation video](https://youtu.be/HTBl142x9GI?si=o_we6t9suYK8DfX3) about the development of the `Kanana` models. + +
+ +## Table of Contents + +- [Kanana 1.5](#kanana-15) + - [Performance](#performance) + - [Base Model Evaluation](#base-model-evaluation) + - [Instruct Model Evaluation](#instruct-model-evaluation) + - [Processing 32K+ Length](#processing-32k-length) +- [Contributors](#contributors) +- [Citation](#citation) +- [Contact](#contact) + +
+ +# Kanana 1.5 + +`Kanana 1.5`, a newly introduced version of the Kanana model family, presents substantial enhancements in **coding, mathematics, and function calling capabilities** over the previous version, enabling broader application to more complex real-world problems. This new version now can handle __up to 32K tokens length natively and up to 128K tokens using YaRN__, allowing the model to maintain coherence when handling extensive documents or engaging in extended conversations. Furthermore, Kanana 1.5 delivers more natural and accurate conversations through a __refined post-training process__. + +

+
+ + + +
+ +> [!Note] +> Neither the pre-training nor the post-training data includes Kakao user data. + +## Performance + +### Base Model Evaluation + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelsMMLUKMMLUHAERAEHumanEvalMBPPGSM8K
Kanana-1.5-8B64.2448.9482.7761.5957.8063.53
Kanana-8B64.2248.3083.4140.2451.4057.09
+ +
+ +### Instruct Model Evaluation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelsMT-BenchKoMT-BenchIFEvalHumanEval+MBPP+GSM8K (0-shot)MATHMMLU (0-shot, CoT)KMMLU (0-shot, CoT)FunctionChatBench
Kanana-1.5-8B*7.767.6380.1176.8367.9987.6467.5468.8248.2858.00
Kanana-8B7.136.9276.9162.2043.9279.2337.6866.5047.4317.37
+ +> [!Note] +> \* Models released under Apache 2.0 are trained on the latest versions compared to other models. + +
+ +## Processing 32K+ Length +Currently, the `config.json` uploaded to HuggingFace is configured for token lengths of 32,768 or less. To process tokens beyond this length, YaRN must be applied. By updating the `config.json` with the following parameters, you can apply YaRN to handle token sequences up to 128K in length: +```json +"rope_scaling": { + "factor": 4.4, + "original_max_position_embeddings": 32768, + "type": "yarn", + "beta_fast": 64, + "beta_slow": 2 +}, +``` + +
+ +## Contributors +- Language Model Training: Yunju Bak, Doohae Jung, Boseop Kim, Nayeon Kim, Hojin Lee, Jaesun Park, Minho Ryu +- Language Model Alignment: Jiyeon Ham, Seungjae Jung, Hyunho Kim, Hyunwoong Ko, Changmin Lee, Daniel Wontae Nam +- AI Engineering: Youmin Kim, Hyeongju Kim + +
+ +## Citation + +``` +@misc{kananallmteam2025kananacomputeefficientbilinguallanguage, + title={Kanana: Compute-efficient Bilingual Language Models}, + author={Kanana LLM Team and Yunju Bak and Hojin Lee and Minho Ryu and Jiyeon Ham and Seungjae Jung and Daniel Wontae Nam and Taegyeong Eo and Donghun Lee and Doohae Jung and Boseop Kim and Nayeon Kim and Jaesun Park and Hyunho Kim and Hyunwoong Ko and Changmin Lee and Kyoung-Woon On and Seulye Baeg and Junrae Cho and Sunghee Jung and Jieun Kang and EungGyun Kim and Eunhwa Kim and Byeongil Ko and Daniel Lee and Minchul Lee and Miok Lee and Shinbok Lee and Gaeun Seo}, + year={2025}, + eprint={2502.18934}, + archivePrefix={arXiv}, + primaryClass={cs.CL}, + url={https://arxiv.org/abs/2502.18934}, +} +``` + +
+ +## Contact +- Kanana LLM Team Technical Support: kanana-llm@kakaocorp.com +- Business & Partnership Contact: alpha.k@kakaocorp.com \ No newline at end of file diff --git a/kanana-1.5-8b-instruct-2505-bf16.gguf b/kanana-1.5-8b-instruct-2505-bf16.gguf new file mode 100644 index 0000000..03ad683 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:448e46c848bac84a292ed9d66289fe1b7765d4a6530d636efd310f7839f91929 +size 16068954592 diff --git a/kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf b/kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf new file mode 100644 index 0000000..b9f905c --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-bf16_q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7657af238ed67463c44ab6ff919b69b5a60f85562a055fc67dfbd4a28fac3941 +size 9525840352 diff --git a/kanana-1.5-8b-instruct-2505-f16_q8_0.gguf b/kanana-1.5-8b-instruct-2505-f16_q8_0.gguf new file mode 100644 index 0000000..63781fc --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-f16_q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da985819c6e8901536ace7e194545bfda4f34ca7619fa530ed5399dad5aa821d +size 11916593632 diff --git a/kanana-1.5-8b-instruct-2505-iq2_m.gguf b/kanana-1.5-8b-instruct-2505-iq2_m.gguf new file mode 100644 index 0000000..5d7e34d --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq2_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed6f563561bd4f29478a7316a1e93d7bc8a95655cf677b9a0712a1be46d6bef4 +size 3180809504 diff --git a/kanana-1.5-8b-instruct-2505-iq2_s.gguf b/kanana-1.5-8b-instruct-2505-iq2_s.gguf new file mode 100644 index 0000000..5582923 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq2_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82a919d8e28cf0733a2e396a869accf4e55f5e4283540f9cc81e35002a3f091f +size 3051572512 diff --git a/kanana-1.5-8b-instruct-2505-iq2_xs.gguf b/kanana-1.5-8b-instruct-2505-iq2_xs.gguf new file mode 100644 index 0000000..3fdea1b --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq2_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:023ee29af8808082e131fa56e5879a9fa79e372f4f63617249d2c8cfa7cc8ce3 +size 2970832160 diff --git a/kanana-1.5-8b-instruct-2505-iq2_xxs.gguf b/kanana-1.5-8b-instruct-2505-iq2_xxs.gguf new file mode 100644 index 0000000..10a9884 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq2_xxs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:675bebaa3682b69a6311e0402b67c9efc0c87770c7a4cfd0a79d250a7d16c6ec +size 2768915744 diff --git a/kanana-1.5-8b-instruct-2505-iq3_m.gguf b/kanana-1.5-8b-instruct-2505-iq3_m.gguf new file mode 100644 index 0000000..6688ed1 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq3_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d17484cff20ae1078164e2333891c3bf8a41eba54a3fd7c273ac55cc00a0b2b1 +size 3797044512 diff --git a/kanana-1.5-8b-instruct-2505-iq3_s.gguf b/kanana-1.5-8b-instruct-2505-iq3_s.gguf new file mode 100644 index 0000000..1bd63f2 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq3_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:746ab2064ad75d5e3124cc1d0e1ed9213f25dcfc2ed0fe73f5d363feb65c582d +size 3743567136 diff --git a/kanana-1.5-8b-instruct-2505-iq3_xs.gguf b/kanana-1.5-8b-instruct-2505-iq3_xs.gguf new file mode 100644 index 0000000..9a261b7 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq3_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:859563d31ba517023633f586ef4f70dfa03f24065ef30201ffe85066209878d6 +size 3579989280 diff --git a/kanana-1.5-8b-instruct-2505-iq3_xxs.gguf b/kanana-1.5-8b-instruct-2505-iq3_xxs.gguf new file mode 100644 index 0000000..d2f2d48 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq3_xxs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b310199410f1042e159e95f36121b3d5b9d8dd1ca3f2965f04515cb95c09e6c +size 3427945760 diff --git a/kanana-1.5-8b-instruct-2505-iq4_nl.gguf b/kanana-1.5-8b-instruct-2505-iq4_nl.gguf new file mode 100644 index 0000000..3ae498b --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq4_nl.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:104a3463635a1acaa74a0f37204cd92b074f403323597d5ec23ecf73bb8fdf31 +size 4678020480 diff --git a/kanana-1.5-8b-instruct-2505-iq4_xs.gguf b/kanana-1.5-8b-instruct-2505-iq4_xs.gguf new file mode 100644 index 0000000..7b8a66a --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-iq4_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2c674a5de1756e219391484c3e801d8797d5a58e2c1a6ca431f4d23e59003ff +size 4447693824 diff --git a/kanana-1.5-8b-instruct-2505-q2_k_m.gguf b/kanana-1.5-8b-instruct-2505-q2_k_m.gguf new file mode 100644 index 0000000..9eab38a --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q2_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cef325be7f5dd4417d42a79b170686f05eed7dfc15fffb5d35a2d9115625eeec +size 3403389408 diff --git a/kanana-1.5-8b-instruct-2505-q2_k_s.gguf b/kanana-1.5-8b-instruct-2505-q2_k_s.gguf new file mode 100644 index 0000000..cf649dd --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q2_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d19d654383d651641cbfbdce229196e0f55a86f91bff3af53906d1cb1b29911 +size 3079818528 diff --git a/kanana-1.5-8b-instruct-2505-q3_k_m.gguf b/kanana-1.5-8b-instruct-2505-q3_k_m.gguf new file mode 100644 index 0000000..4edffb6 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q3_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e69999effa78153bc8c0dc180e12e4fd4959b6098b77dfa098017609423c150 +size 4188510688 diff --git a/kanana-1.5-8b-instruct-2505-q3_k_s.gguf b/kanana-1.5-8b-instruct-2505-q3_k_s.gguf new file mode 100644 index 0000000..fff3668 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q3_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a32b44992db49078290ca9b8040167a6e7611d6af8d48012d55ed826d23de7a +size 3794947360 diff --git a/kanana-1.5-8b-instruct-2505-q4_0.gguf b/kanana-1.5-8b-instruct-2505-q4_0.gguf new file mode 100644 index 0000000..d7ceb2b --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1cca21b4cefe763b670a59322b8e6a63f9f8802b5d3548a2553943ef1758ddaa +size 4525801760 diff --git a/kanana-1.5-8b-instruct-2505-q4_1.gguf b/kanana-1.5-8b-instruct-2505-q4_1.gguf new file mode 100644 index 0000000..8071e6f --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91f5d86fbea7c2c8a59052693ac4f82c6adea27748b55400d781542541760cab +size 5027677984 diff --git a/kanana-1.5-8b-instruct-2505-q4_k_m.gguf b/kanana-1.5-8b-instruct-2505-q4_k_m.gguf new file mode 100644 index 0000000..77c51a7 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q4_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7d247fe117a43d35558eb91bc6c457e524e8aec5785955194d74f436954768b +size 5006137824 diff --git a/kanana-1.5-8b-instruct-2505-q4_k_s.gguf b/kanana-1.5-8b-instruct-2505-q4_k_s.gguf new file mode 100644 index 0000000..7b49507 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q4_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b88d9a72f248a3041eb27c628f1899cf7def214690b876f1c046addb861322f6 +size 4876376544 diff --git a/kanana-1.5-8b-instruct-2505-q5_0.gguf b/kanana-1.5-8b-instruct-2505-q5_0.gguf new file mode 100644 index 0000000..5363fb3 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q5_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c4758b21914754b887d327535f95908ad80f07fee718424c1879ae797f36917 +size 5529554208 diff --git a/kanana-1.5-8b-instruct-2505-q5_1.gguf b/kanana-1.5-8b-instruct-2505-q5_1.gguf new file mode 100644 index 0000000..a16f19e --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q5_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c5c73c9b3861813309c6fb27a121643b110c33bc9e4f05a648cb5f0feaffbc8 +size 6031430432 diff --git a/kanana-1.5-8b-instruct-2505-q5_k_m.gguf b/kanana-1.5-8b-instruct-2505-q5_k_m.gguf new file mode 100644 index 0000000..5853e06 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q5_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f055a96d45f1f586543519801b068c9a3c52a3bb8867fe46f150c5a380e0fc41 +size 5820619232 diff --git a/kanana-1.5-8b-instruct-2505-q5_k_s.gguf b/kanana-1.5-8b-instruct-2505-q5_k_s.gguf new file mode 100644 index 0000000..89f670f --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q5_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8a3bbb411049670363373804253b6c6524473f9099a39ab3edac36de0da13f1 +size 5753772512 diff --git a/kanana-1.5-8b-instruct-2505-q6_k_m.gguf b/kanana-1.5-8b-instruct-2505-q6_k_m.gguf new file mode 100644 index 0000000..8904539 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q6_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a3ae55383aa0448271bc8e660aa23ab040478094c36c26ed4a99cb9df6f2324 +size 6596041184 diff --git a/kanana-1.5-8b-instruct-2505-q8_0.gguf b/kanana-1.5-8b-instruct-2505-q8_0.gguf new file mode 100644 index 0000000..a2ec6c5 --- /dev/null +++ b/kanana-1.5-8b-instruct-2505-q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd249caac774260b092185c5c8bfe8900ee7ed104d1a1d8becee9616539c16e7 +size 8540811232 diff --git a/kanana-1.5-8b-instruct-2505.imatrix b/kanana-1.5-8b-instruct-2505.imatrix new file mode 100644 index 0000000..a154e2a --- /dev/null +++ b/kanana-1.5-8b-instruct-2505.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e90ad3c161380b88a03a5d52dcac99ee685bd9518f401d56ce87122750d610e +size 4988204