Instructions to use OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5 with Transformers:

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

pipe = pipeline("text-generation", model="OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5")
model = AutoModelForCausalLM.from_pretrained("OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5

SGLang

How to use OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5 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 "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5" \
    --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": "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5",
		"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 "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5" \
        --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": "OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5 with Docker Model Runner:
```
docker model run hf.co/OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5
```

How to use the special token `<|system|>` effectively?

#13

by AayushShah - opened May 11, 2023

Discussion

AayushShah

May 11, 2023

•

edited May 11, 2023

Hello 🙋🏻‍♂️
I am willing to make a prompt which can summarize the information given in the bullet points. Before I was following this prompt:

BEFORE:

<|prompter|>
You are a helpful narrator in the company. Your task is to provide a meaningful summary by using the scientific terms possible considering the audience is educated. Use the information below:

Information:
- {JARGON WORDS}
- {ANALYSIS }
- {RESEARCH}
- etc...

Insightful and verbose narration of the information given above:<|endoftext|><|assistant|>

But now I have found these special tokens:

<|system|>
<|assistant|>
<|prefix_begin|>
<|prefix_end|>
<|prompter|>

So I think I can use the <|system|> to set what I want to do, instead of giving all things at once in the <|prompter|>.

AFTER:

<|system|>
You are a helpful narrator in the company. Your task is to provide a meaningful summary by using the scientific terms possible considering the audience is educated. Use the information below.
<|system|>


<|prompter|>
Information:
- {JARGON WORDS}
- {ANALYSIS }
- {RESEARCH}
- etc...

Insightful and verbose narration of the information given above:<|endoftext|><|assistant|>"""

The question

Am I using the <|system|> token correctly? Is this the way to use it? Or I am actually misleading the LLM?

Please help,
Thanks 🙏🏻

andreaskoepf

OpenAssistant org May 11, 2023

SFT-4 does not not know about the system prompt, please don't use it (it's weights are still random). If you want add an additional pre-prompt please try to simply place it at the beginning (without system) or at the start of the prompter request.

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