Instructions to use darkps/ice-AI with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use darkps/ice-AI with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="darkps/ice-AI")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("darkps/ice-AI")
model = AutoModelForCausalLM.from_pretrained("darkps/ice-AI", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use darkps/ice-AI with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "darkps/ice-AI"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "darkps/ice-AI",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/darkps/ice-AI

SGLang

How to use darkps/ice-AI 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 "darkps/ice-AI" \
    --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": "darkps/ice-AI",
		"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 "darkps/ice-AI" \
        --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": "darkps/ice-AI",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use darkps/ice-AI with Docker Model Runner:
```
docker model run hf.co/darkps/ice-AI
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

ICE AI

ICE 0001

The "ice" model is a very robust, medium-sized model for human-like conversations, designed for quick chats and small code snippets.

The ice model was released with massive equations:

Major Improvements

It was trained on 5.47 billion Codex conversations.
It was also trained on more than 30 billion human conversations to better adapt to different Arabic dialects and multiple languages.

Key Specifications

Model Family: ICE AI
ID: ice-0001
Model Size: 8B Parameters
Context Length: 32,768 tokens
Format: Transformers
Inference Support: CPU / GPU
Primary Focus: Human-like conversational AI

Training

Trained on approximately 36 trillion tokens across 119 languages and dialects, with a strong focus on multiple Arabic dialects, international languages, and programming/code data.

Recommended Usage

ICE AI performs best when used for:

General conversations
multilingual chat
Software development
Code generation
Code debugging
Technical questions
Scripting and automation
Local offline AI deployment

⚠️ Notes

Designed for conversational and coding tasks.
Output quality may vary depending on the quantization level and hardware.
Best results are achieved with structured prompts.
Larger context sizes may require substantial RAM/VRAM.

About Dark

DarkPs is an AI organization owned by FanuonAI, developing and maintaining open-source AI models such as DarkIT, ICE AI, and DarkCoder.

Platform: https://dark.ps

Downloads last month: 914

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for darkps/ice-AI

Quantizations

2 models