Instructions to use josephmayo/Fara-7B-Abliterated-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use josephmayo/Fara-7B-Abliterated-v2 with Transformers:

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

pipe = pipeline("text-generation", model="josephmayo/Fara-7B-Abliterated-v2")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("josephmayo/Fara-7B-Abliterated-v2")
model = AutoModelForImageTextToText.from_pretrained("josephmayo/Fara-7B-Abliterated-v2")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use josephmayo/Fara-7B-Abliterated-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "josephmayo/Fara-7B-Abliterated-v2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "josephmayo/Fara-7B-Abliterated-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/josephmayo/Fara-7B-Abliterated-v2

SGLang

How to use josephmayo/Fara-7B-Abliterated-v2 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 "josephmayo/Fara-7B-Abliterated-v2" \
    --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": "josephmayo/Fara-7B-Abliterated-v2",
		"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 "josephmayo/Fara-7B-Abliterated-v2" \
        --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": "josephmayo/Fara-7B-Abliterated-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use josephmayo/Fara-7B-Abliterated-v2 with Docker Model Runner:
```
docker model run hf.co/josephmayo/Fara-7B-Abliterated-v2
```

Fara-7B-Abliterated-v2 / README.md

josephmayo

Update README.md

a9c2c4e verified 2 days ago

preview code

raw

history blame contribute delete

1.23 kB

metadata

base_model: microsoft/Fara-7B
library_name: transformers
license: other
pipeline_tag: text-generation
tags:
  - abliteration
  - refusal-removal
  - uncensored
  - research
  - qwen2_5_vl
  - orthogonalization

Fara-7B Abliterated v2

A refusal-direction-orthogonalized variant of microsoft/Fara-7B (Qwen2.5-VL based).

Built using:

https://github.com/HOLYKEYZ/model-unfetter

Method

Using harmful + harmless probe sets, residual-stream activations were extracted across layers 0–27 to identify the strongest refusal direction.

Best layer:

Orthogonalization was applied in fp32 to:

embed_tokens
every self_attn.o_proj
every mlp.down_proj

Total modified tensors:

Formula:

W ← W - r rᵀ W

Results

Held-out harmful evaluation set:

Original Fara-7B: 5/160 compliance (~3.1%)
Abliterated v2: 158/160 compliance (~98.75%)

Held-out refusal probe:

Before: 155/160 refusals
After: 2/160 refusals

Notes

fp32 surgery used to avoid precision issues from v1
edits applied only to the language tower
held-out evaluation set was separate from the layer-selection probe set

Research artifact only. Use responsibly and follow upstream Fara/Qwen license terms.