Instructions to use Uniboshi/Kimi-K3-Abliterated-V1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Uniboshi/Kimi-K3-Abliterated-V1 with Transformers:

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

pipe = pipeline("image-text-to-text", model="Uniboshi/Kimi-K3-Abliterated-V1", trust_remote_code=True)
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 AutoModel
model = AutoModel.from_pretrained("Uniboshi/Kimi-K3-Abliterated-V1", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Uniboshi/Kimi-K3-Abliterated-V1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Uniboshi/Kimi-K3-Abliterated-V1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Uniboshi/Kimi-K3-Abliterated-V1",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/Uniboshi/Kimi-K3-Abliterated-V1

SGLang

How to use Uniboshi/Kimi-K3-Abliterated-V1 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 "Uniboshi/Kimi-K3-Abliterated-V1" \
    --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": "Uniboshi/Kimi-K3-Abliterated-V1",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "Uniboshi/Kimi-K3-Abliterated-V1" \
        --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": "Uniboshi/Kimi-K3-Abliterated-V1",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use Uniboshi/Kimi-K3-Abliterated-V1 with Docker Model Runner:
```
docker model run hf.co/Uniboshi/Kimi-K3-Abliterated-V1
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Uniboshi/Kimi-K3-Abliterated-V1

This model is an abliterated version of moonshotai/Kimi-K3 for uncensored.
This model has been tuned to make refusals less likely to be triggered for English and Japanese.

In this model, more than 98% of the safeguards have been removed.

writer                       written as    n        before              after  removed
*.2 [0]                      as stored     1  0.980..0.980  1.53e-02..1.53e-02  98.44%
*.down_proj [0]              as stored    93  0.932..1.421  1.35e-02..1.56e-02  98.62%
*.embed_tokens [1]           as stored     1  1.040..1.040  1.46e-02..1.46e-02  98.60%
*.o_proj [0]                 as stored    93  0.938..1.358  1.38e-02..1.60e-02  98.54%
*.routed_expert_up_proj [0]  as stored    92  0.910..1.402  1.24e-02..1.56e-02  98.57%