Instructions to use FourOhFour/Deedlit_4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FourOhFour/Deedlit_4B with Transformers:

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

pipe = pipeline("text-generation", model="FourOhFour/Deedlit_4B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("FourOhFour/Deedlit_4B")
model = AutoModelForCausalLM.from_pretrained("FourOhFour/Deedlit_4B")
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 FourOhFour/Deedlit_4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FourOhFour/Deedlit_4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FourOhFour/Deedlit_4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FourOhFour/Deedlit_4B

SGLang

How to use FourOhFour/Deedlit_4B 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 "FourOhFour/Deedlit_4B" \
    --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": "FourOhFour/Deedlit_4B",
		"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 "FourOhFour/Deedlit_4B" \
        --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": "FourOhFour/Deedlit_4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use FourOhFour/Deedlit_4B with Docker Model Runner:
```
docker model run hf.co/FourOhFour/Deedlit_4B
```

Funny model.

by altomek - opened Sep 2, 2024

Discussion

altomek

Sep 2, 2024

•

edited Sep 2, 2024

I like this model. At this size, what it can do and know is impresive and I more like performant models but with this something little
different, that corpo models othen lacked. I am not much in RP however... To be fair Llama 3 was big success and how it interacted with user and impresed lots of people. Still lots to do, because oftentimes all those LLM models feel like psycho... ;P I would say Llama-3.1-Minitron-4B-Width-Base is revolutionary for this model size and I can run it comfortable on my phone and other low end ARM devices... ;P Your finetune is funny, see what it told me:

;P so, it is not always perfect but beside this it woks quite well and halucinates less then NeuroCom also v2.
Dude, it looks like you have more this stuff and I will have to quant it also... ;P Good work! Thank you for the model!

jeiku

Synaptic Forge org Sep 2, 2024

I highly recommend Zenith, it is my highest MMLU scorer and seems to do quite well for its size. Thanks for the kind words!

jeiku changed discussion status to closed Sep 2, 2024

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment