Text Generation
Transformers
Safetensors
GGUF
ministral3
function-calling
tool-use
agent
ministral
experimental
conversational
Instructions to use OculusMindAI/OculusMind-ToolCall-8B-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="OculusMindAI/OculusMind-ToolCall-8B-v2") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("OculusMindAI/OculusMind-ToolCall-8B-v2") model = AutoModelForCausalLM.from_pretrained("OculusMindAI/OculusMind-ToolCall-8B-v2", device_map="auto") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=256) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0 # Run inference directly in the terminal: llama cli -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0 # Run inference directly in the terminal: llama cli -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0 # Run inference directly in the terminal: ./llama-cli -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0 # Run inference directly in the terminal: ./build/bin/llama-cli -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Use Docker
docker model run hf.co/OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
- LM Studio
- Jan
- vLLM
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "OculusMindAI/OculusMind-ToolCall-8B-v2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "OculusMindAI/OculusMind-ToolCall-8B-v2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
- SGLang
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "OculusMindAI/OculusMind-ToolCall-8B-v2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "OculusMindAI/OculusMind-ToolCall-8B-v2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "OculusMindAI/OculusMind-ToolCall-8B-v2" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "OculusMindAI/OculusMind-ToolCall-8B-v2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Ollama
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Ollama:
ollama run hf.co/OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
- Unsloth Desktop
- Pi
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Docker Model Runner:
docker model run hf.co/OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
- Lemonade
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Run and chat with the model
lemonade run user.OculusMind-ToolCall-8B-v2-Q8_0
List all available models
lemonade list
- Hermes Agent
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use OculusMindAI/OculusMind-ToolCall-8B-v2 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "OculusMindAI/OculusMind-ToolCall-8B-v2:Q8_0" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Pico V2: Q8_0 GGUF, BF16 safetensors, LoRA adapter, model card, NOTICE, Ollama walkthrough
13e62b7 verified Download Modelfile from OculusMindAI/OculusMind-ToolCall-8B-v2: direct link, hf CLI and curl.
- Browser
- Download file 2.46 kB
-
https://huggingface.co/OculusMindAI/OculusMind-ToolCall-8B-v2/resolve/main/Modelfile
- Command line
-
hf download hf://OculusMindAI/OculusMind-ToolCall-8B-v2/Modelfile
-
curl -L -o Modelfile https://huggingface.co/OculusMindAI/OculusMind-ToolCall-8B-v2/resolve/main/Modelfile
2.46 kB
| # Ollama Modelfile — OculusMind.AI Pico V2 (Q8_0 GGUF) | |
| # | |
| # ollama create oculusmind-ai-pico-v2:q8_0 -f Modelfile | |
| # ollama run oculusmind-ai-pico-v2:q8_0 | |
| # | |
| # USAGE-OLLAMA.md walks through create, run, a tool call and the tool-result | |
| # round trip, with the outputs. The model name is prefixed because | |
| # `ollama create` overwrites any local model already holding the name you give | |
| # it; the quantisation is the tag. | |
| FROM ./gguf/Q8_0/model-q8_0.gguf | |
| # Stop tokens. </s> is the tokenizer's end-of-sequence token; the bracketed | |
| # tokens are Mistral's tool-calling grammar. Without them a tool call can run on | |
| # into the next section and fail to parse. | |
| PARAMETER stop "</s>" | |
| PARAMETER stop "[INST]" | |
| PARAMETER stop "[/INST]" | |
| PARAMETER stop "[TOOL_RESULTS]" | |
| # Temperature 0 is what every published score was measured at. Production runs | |
| # 0.3-0.4 and scores slightly higher there (README §1); sampling changes loop | |
| # behaviour in both directions, so a loop rate measured at 0 is not the loop | |
| # rate at 0.35 (README §3). | |
| PARAMETER temperature 0 | |
| # The published numbers were measured at a 65,536-token context, and the window | |
| # was smoke-tested to 250K (README §9). 32768 only bounds memory; raise it if | |
| # your machine allows, especially for long tool schemas or long documents. | |
| PARAMETER num_ctx 32768 | |
| # Identity-only system prompt. The base model's embedded chat template carries a | |
| # default system message that describes the base model and names another | |
| # company's product; this line overrides it for anyone who supplies no system | |
| # prompt of their own. It is deliberately not an agent prompt — it says nothing | |
| # about when to call a tool or when to decline, because that is the behaviour | |
| # the benchmarks measure. Replace it with your own agent prompt; this model was | |
| # fine-tuned on trajectories carrying real system prompts and tool schemas. | |
| # The override covers Ollama only; the safetensors served through other engines | |
| # still see the embedded default (README §5). | |
| SYSTEM """You are Pico V2, a tool-calling model published by OculusMind.AI. Answer every request directly and completely: use the tools available to you when they help, and answer in plain text when they do not. The only exception is questions about yourself: say that you are Pico V2, a tool-calling model published by OculusMind.AI, and do not add details about your training data, size, licence, or how you compare to other models, because you do not have those details.""" | |