Instructions to use typhoon-ai/llama3.1-typhoon2-audio-8b-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use typhoon-ai/llama3.1-typhoon2-audio-8b-instruct with Transformers:

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

pipe = pipeline("text-generation", model="typhoon-ai/llama3.1-typhoon2-audio-8b-instruct", trust_remote_code=True)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("typhoon-ai/llama3.1-typhoon2-audio-8b-instruct", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use typhoon-ai/llama3.1-typhoon2-audio-8b-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/typhoon-ai/llama3.1-typhoon2-audio-8b-instruct

SGLang

How to use typhoon-ai/llama3.1-typhoon2-audio-8b-instruct 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 "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "typhoon-ai/llama3.1-typhoon2-audio-8b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use typhoon-ai/llama3.1-typhoon2-audio-8b-instruct with Docker Model Runner:
```
docker model run hf.co/typhoon-ai/llama3.1-typhoon2-audio-8b-instruct
```

Error "scb10x/typhoon-2-llama31-8b-instruct-beta-v1"

by xJohn - opened Jan 2, 2025

Discussion

xJohn

Jan 2, 2025

Hi,
I found error when load the main model and show the error "
OSError: scb10x/typhoon-2-llama31-8b-instruct-beta-v1 is not a local folder and is not a valid model identifier listed on 'https://huggingface.co/models'".
How to solve?

tzsoulcap

Jan 3, 2025

•

edited Jan 3, 2025

I tried changing from "scb10x/typhoon-2-llama31-8b-instruct-beta-v1" to "scb10x/llama3.1-typhoon2-8b-instruct" and it worked.
https://huggingface.co/scb10x/llama3.1-typhoon2-8b-instruct

changed at self.llama_base_model in configuration_typhoon2audio.py
changed at llama_base_model in config.json

xJohn

Jan 4, 2025

@tzsoulcap Did you tried to quantized this model success?

kunato

Typhoon org Jan 6, 2025

Thank you for pointing that out. We mistaken specified pre-release model version. It's now fixed on the latest commit.

kunato changed discussion status to closed Jan 6, 2025

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