Instructions to use NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck with Transformers:

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

pipe = pipeline("text-generation", model="NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck")
model = AutoModelForCausalLM.from_pretrained("NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck

SGLang

How to use NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck 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 "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck" \
    --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": "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck",
		"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 "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck" \
        --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": "NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck with Docker Model Runner:
```
docker model run hf.co/NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck
```

The biggest model ever to have been released. Has not been tested, nor do I have the compute to test it. If anyone is willing to host this to help me test, please share your results in the community tab.

Thank you for coming to my ted talk.

This is nearly 960GB of weights. It requires at least 8xA100 80gb to run it in 4 bits probably. probably

Downloads last month: 49

Safetensors

Model size

464B params

Tensor type

BF16

Model tree for NobodyExistsOnTheInternet/Llama-2-70b-x8-MoE-clown-truck

Finetunes

2 models