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
```

Facing error while inferencing

#14

by sauravm8 - opened May 12, 2023

Discussion

sauravm8

May 12, 2023

I am facing an inference error
The size of tensor a (2048) must match the size of tensor b (2049) at non-singleton dimension 3

I am cutting off the input prompt way before 2048 tokens at around 1250 words (~1500 tokens)

`final_message = f"<|prompter|>What do you think of the following and keep it under 100 words: \n {received_message}<|endoftext|><|assistant|>"
inputs = tokenizer(final_message, return_tensors="pt").to(model.device)
print(f"Number of tokens --> {inputs['input_ids'].shape[1]}")

tokens = model.generate(**inputs,  max_new_tokens=1000, do_sample=True, temperature=0.8)
response = tokenizer.decode(tokens[0]).split("<|assistant|>")[1].strip("<|endoftext|>")
print(f"Total time taken is {time.time()-start_query_time}")
print(response)`

GPU is not an issue. What is going on here?

sauravm8

May 12, 2023

I think I figured it out, even the new tokens getting generated have to be within the 2048 token limit. Otherwise, it crashes. Can it not have a streaming window of memory? Am I missing something?

vijayrps

Jun 20, 2023

@sauravm8 facing a similar issue.whats the solution?

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