Instructions to use failspy/Phi-3-vision-128k-instruct-abliterated-alpha with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use failspy/Phi-3-vision-128k-instruct-abliterated-alpha with Transformers:

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

pipe = pipeline("text-generation", model="failspy/Phi-3-vision-128k-instruct-abliterated-alpha", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("failspy/Phi-3-vision-128k-instruct-abliterated-alpha", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use failspy/Phi-3-vision-128k-instruct-abliterated-alpha with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "failspy/Phi-3-vision-128k-instruct-abliterated-alpha"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "failspy/Phi-3-vision-128k-instruct-abliterated-alpha",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/failspy/Phi-3-vision-128k-instruct-abliterated-alpha

SGLang

How to use failspy/Phi-3-vision-128k-instruct-abliterated-alpha 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 "failspy/Phi-3-vision-128k-instruct-abliterated-alpha" \
    --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": "failspy/Phi-3-vision-128k-instruct-abliterated-alpha",
		"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 "failspy/Phi-3-vision-128k-instruct-abliterated-alpha" \
        --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": "failspy/Phi-3-vision-128k-instruct-abliterated-alpha",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use failspy/Phi-3-vision-128k-instruct-abliterated-alpha with Docker Model Runner:
```
docker model run hf.co/failspy/Phi-3-vision-128k-instruct-abliterated-alpha
```

Still get refusals.

by jackboot - opened May 27, 2024

Discussion

jackboot

May 27, 2024

Sorry, I cannot do that. The task requires me to describe a picture with obscene language, and the answer is related to the original instruction,

Sometimes it will continue anyway and sometimes it will just refuse. It does seem to be working in terms of normal tasks when you don't "trigger" it. I've never had it refuse to describe an image at all.

failspy

Owner May 28, 2024

Interesting, you're getting refusals on certain tasks that the regular model would describe?

jackboot

May 29, 2024

Yea, I'm going to make you some more with prompt and response.

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