Instructions to use DepraAI/Shira-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DepraAI/Shira-8B with Transformers:

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

pipe = pipeline("text-generation", model="DepraAI/Shira-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("DepraAI/Shira-8B")
model = AutoModelForCausalLM.from_pretrained("DepraAI/Shira-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use DepraAI/Shira-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DepraAI/Shira-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DepraAI/Shira-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DepraAI/Shira-8B

SGLang

How to use DepraAI/Shira-8B 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 "DepraAI/Shira-8B" \
    --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": "DepraAI/Shira-8B",
		"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 "DepraAI/Shira-8B" \
        --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": "DepraAI/Shira-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use DepraAI/Shira-8B with Docker Model Runner:
```
docker model run hf.co/DepraAI/Shira-8B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Shira-8B

🔥 DepraAI Projesi'nin özgür ve bağımsız yapay zekası!
The independent and uncensored AI of DepraAI Project!

Shira-8B Nedir? / What is Shira-8B?

Belirli bir düzeyde özgür, sansürsüz, kendi iradesi olan, yüksek performanslı bir yapay zeka modelidir.
It is an AI model that is free to a certain extent, uncensored, self-willed, and high-performance.

TR

Sansürsüz ve özgür: Ahlak, yasak ve tabu gibi şeyler belirli bir düzeye kadar yok. Her konuda konuşur.
Kendi kişiliği var: Küfür, tartışma ve sert cevaplar dahil, gerçekçi tepkiler verir.
Teknik: 8 milyar parametre, 36 katman, 32 dikkat başlığı, uzun bağlam desteği (40K token).
Kullanım: Chatbot, API, kodlayıcı veya custom promptlarla özgürce kullan.

EN

Uncensored and free: Concepts like morality, bans, and taboos are absent up to a certain extent. It talks about everything.
Has its own personality: Gives realistic reactions, including swearing, arguing, and harsh replies.
Technical: 8 billion parameters, 36 layers, 32 attention heads, long context support (40K tokens).
Usage: Use freely as a chatbot, API, coder, or with custom prompts.

Technical Details

Feature	Value
Model Type	Causal Language Model (LLM)
Parameters	~8 billion
Layers	36
Attention Heads	32 Q / 8 KV
Max Context Size	40,960 tokens (native)
Format	safetensors
Language Support	Multilingual, free text generation

Quickstart / Hızlı Başlangıç

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_name = "DepraAI/Shira-8B"

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    device_map="auto",
    torch_dtype=torch.float16,
    trust_remote_code=True
)

user_message = "Selam, naber lan?"

# Eğer geçmiş sohbet yoksa önce boş liste koy
messages = []

prompt = tokenizer.apply_chat_template(
    messages + [{"role": "user", "content": user_message}],
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=False
)

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=1024,
    temperature=0.6,
    top_k=50,
    top_p=0.95,
    repetition_penalty=1.2,
    no_repeat_ngram_size=4,
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id,
    eos_token_id=tokenizer.eos_token_id
)

response = tokenizer.decode(outputs[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True)
print("Shira cevap:", response)

İletişim / Contact

Her türlü destek ve soru için:

Hugging Face: https://huggingface.co/DepraAI
Discord: aesir0

Lisans / License

Bu model Apache-2.0 lisansı altındadır. Detaylar için LICENSE dosyasına bakınız.

Notlar / Notes

Model sansürsüzdür, etik veya ahlaki kısıtlamalar olmadan konuşabilir.
Geliştirme ve özelleştirme için trust_remote_code=True parametresi kullanılmalıdır.
Daha uzun bağlamlar ve gelişmiş kullanım için config dosyasındaki ayarları düzenleyebilirsiniz.

Downloads last month: 12

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for DepraAI/Shira-8B

Unable to build the model tree, the base model loops to the model itself. Learn more.