--- language: - en - zh - ja - ko - fr - es - pt - de - it - ru - ar - vi - th tags: - text-generation - ternary - bitnet - 1.58bit - cpu - gguf - qwen2.5 - deepseek - efficient - low-memory - jirack - web-ui - routing - tool-call - robotics license: mit --- # JiRack Ultra 1B (CPU) A fast and efficient ~1.5B model optimized for CPU inference. The model was refactored with BitNet features and an updated tokenizer that includes new **Routing**, **Tool call**, and **Robotics** tags. Built on a redesigned DeepSeek R1 architecture with native ternary (BitNet-style) support and ready-to-run GGUF quantizations. - JiRack is a cloud-ready model that helps save money on cloud infrastructure. It can be used as an expert model in RAG deployments, with the ONNX JiRack Java server as an alternative. # JiRack Ternary Architedure & JiRack Tokenizer - Benefits high quality CPU inference TQ_2 on Llama.cpp and Ollama via QAT - Robotcs, Routing, Coding, Multimedia, Advanced tool calling via CMSManhattan/JiRackPrecisionTokenizer # Ollama production support - We are working to support JiRack on Ollama for production systems also - added Jirack chat without reasoning feature https://ollama.com/cmsmanhattan - Follow fresh Ollama platform updates # JiRack sevice options - Current quantizations were done from the FP16 model, but the model allows for more compression thanks to its ternary architecture. - If you need to do ternary compression, please write to me and I'll perform QAT from your dataset, tailored specifically to your task. - Plus double QAT via ONNX QAT. - Adapt train process to avoid catastrophic forgetting with NDA - Adapt train process to avoid fast plato in training with NDA - Convert model to TQ2_0 with support AVX2 and AVX-512 CPU instructions for high performance on CPU - QAT for TQ_2 Llama.cpp Ternarization docs https://huggingface.co/CMSManhattan/JiRackUltra_1b/blob/main/QAT_to_Llama.cpp_GGUF_TQ2_0_JirackUltra_1b.md - Adapts to agentic or instruct models for tool calling, using the JiRak tokenizer to enable high-quality tool calling on small models — built as a domain-specific tool expert. - Deployment and scale # JiRack Codding Agent IDE - It is Agent Coding IDE for JiRack Models to run via Ollama on home PC - It good choose for Agent Coding IDE such as Cursor , Windsurf IDE or Devin IDE etc but more safe that ask you to apply changes and review. - Test version https://huggingface.co/CMSManhattan/JiRackDeltaNet_27b/resolve/main/jirack_ide.zip - Final release version https://huggingface.co/CMSManhattan/JiRackDeltaNet_27b/resolve/main/jirack_ide_final.zip # Spring Boot AI tool calls examples for JiRack Ultra series - Tool call library on java for Enterprise https://github.com/alibaba/spring-ai-alibaba # GoEx AI tool calls examples for JiRack Ultra series - Tool call library on python https://github.com/ShishirPatil/gorilla # JiRack Ultra 1 tool calls to boost tool call quality - Use JiRack Precision tokenzer tags for tool calls with ToolBench https://github.com/OpenBMB/ToolBench - https://huggingface.co/xalss/Qwen2-7B-Instruct-glaive-function-calling - https://huggingface.co/datasets/NousResearch/hermes-function-calling-v1 - Add JiRack tool call tags in the dataset and modify tool call processor if needed # JiRack RoboTech - Advanced Tokenizer with Robotics & Routing & Tool calls Tokenizer and other - [CMSManhattan/JiRackPrecisionTokenizer](https://huggingface.co/CMSManhattan/JiRackPrecisionTokenizer) ## Available Variants | Tag | Quant | Size | Approx. RAM | Description | |-----|-------|------|-------------|-------------| | `cmsmanhattan/jirack-ultra-1b-cpu:latest` | Full | 0.55 GB | ~1.8 GB | Full ternary reference | | `cmsmanhattan/jirack-ultra-1b-cpu-q4:latest` | Q4_K_M | 0.38 GB | ~1.4 GB | Recommended balance | | `cmsmanhattan/jirack-ultra-1b-cpu-q3:latest` | Q3_K_M | 0.31 GB | ~1.2 GB | Good quality / size trade-off | | `cmsmanhattan/jirack-ultra-1b-cpu-q2:latest` | Q2_K | 0.24 GB | ~1.0 GB | Maximum compression | ## Quick Start ### Run with Docker **Default CPU (Q4 recommended)** ```bash docker run -d \ --name jirack_ultra_1b \ -p 7869:7869 \ --cpus=16 \ -e THREADS=16 \ -e THREADS_BATCH=16 \ --restart unless-stopped \ cmsmanhattan/jirack-ultra-1b-cpu-q4:latest ``` **Q3** ```bash docker run -d \ --name jirack_ultra_1b \ -p 7869:7869 \ --cpus=16 \ -e THREADS=16 \ -e THREADS_BATCH=16 \ --restart unless-stopped \ cmsmanhattan/jirack-ultra-1b-cpu-q3:latest ``` **Q2 (lowest memory)** ```bash docker run -d \ --name jirack_ultra_1b \ -p 7869:7869 \ --cpus=16 \ -e THREADS=16 \ -e THREADS_BATCH=16 \ --restart unless-stopped \ cmsmanhattan/jirack-ultra-1b-cpu-q2:latest ``` **Full precision** ```bash docker run -d \ --name jirack_ultra_1b \ -p 7869:7869 \ --cpus=16 \ -e THREADS=16 \ -e THREADS_BATCH=16 \ --restart unless-stopped \ cmsmanhattan/jirack-ultra-1b-cpu:latest ``` **Multi CPU** ```bash docker run -d \ --name jirack_ultra_1b \ -p 7869:7869 \ --cpus=16 \ -e THREADS=16 \ -e THREADS_BATCH=16 \ --restart unless-stopped \ --memory=4g \ --cpus=4 \ cmsmanhattan/jirack-ultra-1b-cpu-q4:latest ``` ### Docker Compose Example ```yaml services: jirack: image: cmsmanhattan/jirack-ultra-1b-cpu-q4:latest container_name: jirack_ultra_1b ports: - "7869:7869" volumes: - .:/app - ./web:/app/web environment: - MAX_TOKENS=2048 - TEMPERATURE=0.7 - TOP_P=0.9 - DEFAULT_STREAM=False - INTRA_THREADS=4 - USE_ENV_ALLOCATOR=1 - THREADS=16 - THREADS_BATCH=16 deploy: resources: limits: memory: 4g ``` ## Access the UI Once the container is running, open your browser and navigate to: `http://localhost:7869` This opens the JiRack UI — a clean web interface. ## Changing the Port The listening port can be easily modified directly from the **Settings** panel within the JiRack UI. ## Licensing - Model weights are released under the MIT License — free to use, modify, and distribute for any purpose, including commercial. No royalties, no per-user fees, no subscription. - The Docker image with UI and the pre-built Ollama quantizations are separate paid products. If you prefer to build your own secure deployment — take the weights, assemble your own stack, and you're done. - The JiRack Ultra 1B model for Docker and Ollama is provided under a commercial license ($12 per user per year). - All JiRack UI clients are provided under a commercial license. - However, the UI clients can be used for free when running together with the official JiRack Docker containers, as long as they are not redistributed separately. For commercial licensing, cluster deployment, or enterprise use of JiRack models, please contact us. - **JiRack MS Windows 11 Desktop Client (with Ollama API):** https://huggingface.co/kgrabko/JiRackTernary_1b/resolve/main/jirack-chat.zip - **Live email chat with the model:** support@cmsmanhattan.com ## Hardware Recommendations ### Recommended Hardware for JiRack Ultra 1B (single Docker container) | Use Case | CPU | RAM | Recommended Quant | Expected Speed | Recommendation | |-------------------|------------------------------|----------|-------------------|---------------------|----------------| | Recommended | Ryzen 5 / Intel i5 | 4–8 GB | Q4_K_M | Excellent interactive | Best choice | | High Performance | Ryzen 7 / Intel i7 | 8–16 GB | Full / Q4 | Excellent | Excellent | | Low Memory | Modern 4+ core CPU | 2–4 GB | Q3_K_M or Q2_K | Usable | Acceptable | | Edge / Minimal | Laptop / SBC CPU | 2 GB | Q2_K | Acceptable | Budget option | ## Important Memory Notes Even though the quantized 1B models are very small, we recommend the following for best experience: - Q4_K_M: 2–4 GB system RAM minimum - Q3_K_M / Q2_K: 1.5–3 GB system RAM - Full precision: 3–4 GB+ system RAM recommended Reasons for extra headroom: - KV-cache consumption during generation - Runtime overhead and temporary buffers - System stability and avoiding out-of-memory errors - Room for larger context windows **Minimum recommended (Q4):** 2–3 GB system RAM **Ideal:** 4–8 GB system RAM I added the default model in full precision. This serves as the base for quantization, allowing us to find the optimal balance between model size and performance. ## Architecture Notes - **Refactored with BitNet features**: Native BitLinear ternary path (b1.58-style) with λ-warmup STE - **Updated tokenizer**: Extended with new special tags for **Routing**, **Tool call**, and **Robotics** - Base: Redesigned Llama-3.2-1B style (Hidden 2048, Intermediate 8192, 16 layers, GQA 32/8, vocab 128256) - RoPE θ = 10000, RMSNorm ε = 1e-6 - Ready-to-run GGUF quantizations (Q2_K, Q3_K_M, Q4_K_M) ## 📧 Contact & Licensing For joint venture opportunities, hardware integration, or licensing inquiries: - **Email:** grabko@cmsmanhattan.com - **Phone:** +1 (516) 777-0945 - **Location:** New York, USA ## License MIT License