Instructions to use Sao10K/L3.3-70B-Euryale-v2.3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Sao10K/L3.3-70B-Euryale-v2.3 with Transformers:

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

pipe = pipeline("text-generation", model="Sao10K/L3.3-70B-Euryale-v2.3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Sao10K/L3.3-70B-Euryale-v2.3")
model = AutoModelForCausalLM.from_pretrained("Sao10K/L3.3-70B-Euryale-v2.3", 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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Sao10K/L3.3-70B-Euryale-v2.3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Sao10K/L3.3-70B-Euryale-v2.3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Sao10K/L3.3-70B-Euryale-v2.3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Sao10K/L3.3-70B-Euryale-v2.3

SGLang

How to use Sao10K/L3.3-70B-Euryale-v2.3 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 "Sao10K/L3.3-70B-Euryale-v2.3" \
    --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": "Sao10K/L3.3-70B-Euryale-v2.3",
		"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 "Sao10K/L3.3-70B-Euryale-v2.3" \
        --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": "Sao10K/L3.3-70B-Euryale-v2.3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Sao10K/L3.3-70B-Euryale-v2.3 with Docker Model Runner:
```
docker model run hf.co/Sao10K/L3.3-70B-Euryale-v2.3
```

Context size on openrouter

by SerialKicked - opened Dec 23, 2024

Discussion

SerialKicked

Dec 23, 2024

•

edited Dec 23, 2024

Hey @Sao10K , I was trying to test your model on OpenRouter, but it seems to be artificially limited to 8K context size. Is that on purpose or just a mistake?

Link, just in case: https://openrouter.ai/sao10k/l3.3-euryale-70b

Sao10K

Owner Dec 23, 2024

•

edited Dec 23, 2024

I have no clue how it works on OR, it depends on provider?

I don't control what they host

Eury v2.3 is a Llama 3.3 model so it's fine up to 128k as the base is, though I'd recommend up to 32k

SerialKicked

Dec 23, 2024

•

edited Dec 23, 2024

I thought you were the one posting it (no idea how that shit works either, I just saw you name and assumed you had some level of control).
Welp, I'll ask them, but it's probably a bust as they likely do it to save processing power/time.

Thanks for the quick answer. (and yep, i assumed as much about the 32k-ish tks)

Cheers.

EDIT: As expected, they do it on purpose.

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