Instructions to use CausalLM/34b-beta with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CausalLM/34b-beta with Transformers:

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

pipe = pipeline("text-generation", model="CausalLM/34b-beta")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("CausalLM/34b-beta")
model = AutoModelForCausalLM.from_pretrained("CausalLM/34b-beta")
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 CausalLM/34b-beta with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CausalLM/34b-beta"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CausalLM/34b-beta",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CausalLM/34b-beta

SGLang

How to use CausalLM/34b-beta 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 "CausalLM/34b-beta" \
    --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": "CausalLM/34b-beta",
		"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 "CausalLM/34b-beta" \
        --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": "CausalLM/34b-beta",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use CausalLM/34b-beta with Docker Model Runner:
```
docker model run hf.co/CausalLM/34b-beta
```

Congrats on the overwhelming MMLU 85.6 score!

by LordTwave - opened Feb 10, 2024

Discussion

LordTwave

Feb 10, 2024

Congrats on your overwhelming victory- 85.6 on MMLU!

The next best open model is PlatiYi-34B-Llama-Q. https://huggingface.co/kyujinpy/PlatYi-34B-Llama-Q

diff from top = 85.6% - 78.78% = 6.82% (it's notably difficult to achieve higher percentage levels for smaller LLMs, so an extra congratulations is due)

A permission-only vision-language model was taken down for unexplained reasons "V0202" "beat" your model with 86.32, but you've kept your model open for the community to use, so your model will be both on top and have community-verifiable results.

So you have a current top on HF with an overwhelming score.

Congrats and thank you!

akumaburn

Feb 23, 2024

Definitely the best open model so far it seems in my limited testing.

JosephusCheung changed discussion status to closed Apr 1, 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