Instructions to use NousResearch/Nous-Hermes-13b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NousResearch/Nous-Hermes-13b with Transformers:

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

pipe = pipeline("text-generation", model="NousResearch/Nous-Hermes-13b")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NousResearch/Nous-Hermes-13b")
model = AutoModelForCausalLM.from_pretrained("NousResearch/Nous-Hermes-13b", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NousResearch/Nous-Hermes-13b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NousResearch/Nous-Hermes-13b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NousResearch/Nous-Hermes-13b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NousResearch/Nous-Hermes-13b

SGLang

How to use NousResearch/Nous-Hermes-13b 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 "NousResearch/Nous-Hermes-13b" \
    --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": "NousResearch/Nous-Hermes-13b",
		"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 "NousResearch/Nous-Hermes-13b" \
        --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": "NousResearch/Nous-Hermes-13b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NousResearch/Nous-Hermes-13b with Docker Model Runner:
```
docker model run hf.co/NousResearch/Nous-Hermes-13b
```

The previous models were really great. Looking forward to see Nous Hermes 13b on GPT4all

by boqsc - opened Jun 3, 2023

Discussion

boqsc

Jun 3, 2023

•

edited Jun 3, 2023

I do hope it is not slower than GPT4-X-Vicuna, it was a great solid consistent model.
Nous-Hermes looks very promising and exciting. I mainly use these to have fun with world creation and interact with it in text-based rpg gaming dialogues.
GPT4-X-Vicuna was already very cappable.

karan4d

NousResearch org Jun 3, 2023

think it is a bit faster. ggml implementations are now available as well on eachadea's HF page.

personally we are enjoying this over gpt4-x-vicuna atm, feels a bit cleaner. thank you for the compliments

teknium

NousResearch org Jun 3, 2023

I do hope it is not slower than GPT4-X-Vicuna, it was a great solid consistent model.
Nous-Hermes looks very promising and exciting. I mainly use these to have fun with world creation and interact with it in text-based rpg gaming dialogues.
GPT4-X-Vicuna was already very cappable.

Being the same base model, it should be identical speed, but obviously depends on the implementation, GPTQ/GGML/HF Transformers,fp16/8bit, etc

teknium changed discussion status to closed Jun 7, 2023

teknium changed discussion status to open Jun 7, 2023

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