Instructions to use datalama/EXAONE-3.5-2.4B-Instruct-Llamafied with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use datalama/EXAONE-3.5-2.4B-Instruct-Llamafied with Transformers:

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

pipe = pipeline("text-generation", model="datalama/EXAONE-3.5-2.4B-Instruct-Llamafied")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("datalama/EXAONE-3.5-2.4B-Instruct-Llamafied")
model = AutoModelForCausalLM.from_pretrained("datalama/EXAONE-3.5-2.4B-Instruct-Llamafied")
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 datalama/EXAONE-3.5-2.4B-Instruct-Llamafied with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/datalama/EXAONE-3.5-2.4B-Instruct-Llamafied

SGLang

How to use datalama/EXAONE-3.5-2.4B-Instruct-Llamafied 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 "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied" \
    --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": "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied",
		"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 "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied" \
        --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": "datalama/EXAONE-3.5-2.4B-Instruct-Llamafied",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use datalama/EXAONE-3.5-2.4B-Instruct-Llamafied with Docker Model Runner:
```
docker model run hf.co/datalama/EXAONE-3.5-2.4B-Instruct-Llamafied
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Updates in EXAONE-3.5

Key Changes

RoPE Scaling Parameter: Added to support longer context_length.
Memory Optimization: For the 2.4B model, tie_word_embeddings is set to True for improved memory efficiency.

⚠️ Using the original Llamafy script as-is may lead to performance degradation.

To address this, I have updated the script and uploaded the Llamafied version of the model.

Special Thanks

@maywell
For updating the code and uploading the model.
LG AI Research
For releasing the original model.
Check out the original release here.

Downloads last month: 3

Safetensors

Model size

2B params

Tensor type

BF16

Model tree for datalama/EXAONE-3.5-2.4B-Instruct-Llamafied

Quantizations

1 model