Instructions to use dphn/dolphin-2_6-phi-2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dphn/dolphin-2_6-phi-2 with Transformers:

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

pipe = pipeline("text-generation", model="dphn/dolphin-2_6-phi-2", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("dphn/dolphin-2_6-phi-2", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dphn/dolphin-2_6-phi-2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dphn/dolphin-2_6-phi-2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dphn/dolphin-2_6-phi-2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/dphn/dolphin-2_6-phi-2

SGLang

How to use dphn/dolphin-2_6-phi-2 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 "dphn/dolphin-2_6-phi-2" \
    --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": "dphn/dolphin-2_6-phi-2",
		"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 "dphn/dolphin-2_6-phi-2" \
        --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": "dphn/dolphin-2_6-phi-2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use dphn/dolphin-2_6-phi-2 with Docker Model Runner:
```
docker model run hf.co/dphn/dolphin-2_6-phi-2
```

is it ggufable?

by sopack - opened Dec 23, 2023

Discussion

sopack

Dec 23, 2023

since most us mortals don't have huge VRAMs, it'd be cool to gguf this model as well.

Demonthos

Dec 24, 2023

Yes, here is a converted model: https://huggingface.co/Demonthos/dolphin-2_6-phi-2-candle/blob/main/model-q4k.gguf
And the code to run the model: https://github.com/floneum/floneum/pull/120/files#diff-3397acf5a72f28f207293cb878d25773c9c3f5ab4c4e1fc88eec1a9e9857e033

ehartford

Dolphin org Dec 24, 2023

I wanna see quip sharp version

https://github.com/Cornell-RelaxML/quip-sharp

KnutJaegersberg

Dec 25, 2023

I like the idea of quip# because it would be so tiny that it would work cpu even without ggml, but from what I remember from the paper is that the 2 bit quantization works less well for smaller models. Might still be beter than gptq 3 bit but I think we get more performance with 5-6 bit gguf version.
Note that the bloke just published a gguf version :)

ehartford changed discussion status to closed Mar 5, 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