Instructions to use hunzed/Fanar-1-9B-Instruct-PalmX-Culture with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hunzed/Fanar-1-9B-Instruct-PalmX-Culture with Transformers:

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

pipe = pipeline("text-generation", model="hunzed/Fanar-1-9B-Instruct-PalmX-Culture")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("hunzed/Fanar-1-9B-Instruct-PalmX-Culture")
model = AutoModelForCausalLM.from_pretrained("hunzed/Fanar-1-9B-Instruct-PalmX-Culture")
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 hunzed/Fanar-1-9B-Instruct-PalmX-Culture with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hunzed/Fanar-1-9B-Instruct-PalmX-Culture"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hunzed/Fanar-1-9B-Instruct-PalmX-Culture",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/hunzed/Fanar-1-9B-Instruct-PalmX-Culture

SGLang

How to use hunzed/Fanar-1-9B-Instruct-PalmX-Culture 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 "hunzed/Fanar-1-9B-Instruct-PalmX-Culture" \
    --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": "hunzed/Fanar-1-9B-Instruct-PalmX-Culture",
		"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 "hunzed/Fanar-1-9B-Instruct-PalmX-Culture" \
        --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": "hunzed/Fanar-1-9B-Instruct-PalmX-Culture",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use hunzed/Fanar-1-9B-Instruct-PalmX-Culture with Docker Model Runner:
```
docker model run hf.co/hunzed/Fanar-1-9B-Instruct-PalmX-Culture
```

Fanar-1-9B-Instruct LoRA Merged Model

This model is a merge of the base model QCRI/Fanar-1-9B-Instruct with a LoRA adapter trained on PalmX, Palm, and NativQA datasets.

Base Model

Base Model: QCRI/Fanar-1-9B-Instruct
LoRA Training: 3 epochs, 0.0002 learning rate, 0.1 dropout
Training Data: PalmX (train+dev), Palm, NativQA

Training Details

Epochs: 3
Learning Rate: 0.0002
Dropout: 0.1
Datasets: PalmX (train_dev) + Palm + NativQA

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "hunzed/Fanar-1-9B-Instruct-PalmX-Culture",
    torch_dtype=torch.float16,
    device_map="auto",
    trust_remote_code=True
)

tokenizer = AutoTokenizer.from_pretrained(
    "hunzed/Fanar-1-9B-Instruct-PalmX-Culture",
    trust_remote_code=True
)

# Generate text
inputs = tokenizer("Your prompt here", return_tensors="pt")
outputs = model.generate(**inputs, max_length=100, do_sample=True, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Model Details

Language: Arabic (primarily)
Type: Causal Language Model
Architecture: Based on Fanar-1-9B-Instruct
Training: LoRA fine-tuning merged into base weights

Downloads last month: 10

Safetensors

Model size

9B params

Tensor type

F16

Model tree for hunzed/Fanar-1-9B-Instruct-PalmX-Culture

Base model

QCRI/Fanar-1-9B

Finetuned

QCRI/Fanar-1-9B-Instruct

Adapter

(2)

this model

Adapters

2 models