Instructions to use theteha/dorina-v4-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use theteha/dorina-v4-instruct with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf theteha/dorina-v4-instruct:F16 # Run inference directly in the terminal: llama cli -hf theteha/dorina-v4-instruct:F16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf theteha/dorina-v4-instruct:F16 # Run inference directly in the terminal: llama cli -hf theteha/dorina-v4-instruct:F16
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf theteha/dorina-v4-instruct:F16 # Run inference directly in the terminal: ./build/bin/llama cli -hf theteha/dorina-v4-instruct:F16
Use Docker
docker model run hf.co/theteha/dorina-v4-instruct:F16
- Llama new
- LM Studio
- Jan
- vLLM
How to use theteha/dorina-v4-instruct with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "theteha/dorina-v4-instruct" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "theteha/dorina-v4-instruct", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/theteha/dorina-v4-instruct:F16
- Ollama
How to use theteha/dorina-v4-instruct with Ollama:
ollama run hf.co/theteha/dorina-v4-instruct:F16
- Unsloth Desktop
- Docker Model Runner
How to use theteha/dorina-v4-instruct with Docker Model Runner:
docker model run hf.co/theteha/dorina-v4-instruct:F16
- Lemonade
How to use theteha/dorina-v4-instruct with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull theteha/dorina-v4-instruct:F16
Run and chat with the model
lemonade run user.dorina-v4-instruct-F16
List all available models
lemonade list
- Atomic Chat
Dorina-V4 Instruct & Mobil OS Ajanı (144.7M) 🤖📱
Dorina-V4 Instruct, Dorina-V4 kristal temel modeli üzerine ChatML standardında eğitilmiş Türkçe/İngilizce talimat ve çok turlu sohbet modelidir.
Aynı zamanda Samsung Galaxy ve Android ekosistemini (Wi-Fi, batarya, ses seviyesi, parlaklık, el feneri, uygulama açma vb.) cihaz içi doğrudan kontrol edebilen Dorina-V4 Android OS Agent ve GGUF f16 formatlarını bünyesinde barındırır.
📦 Model İçeriği
dorina_v4_gold_sft_final.pt: ChatML formatında diyalog ve talimat modeli (32.832 vocab).dorina_v4_clean_agent.pt/dorina_v4_android_agent.pt: Android OS eylemlerini yapılandırılmış JSON formatında çağıran ajan ağırlıkları.dorina_v4_android_agent_f16.gguf(340 MB): Telefon ve edge cihazlarda llama.cpp / Termux üzerinde doğrudan koşan kuantize/optimize motor.dorina_tokenizer_v4_sft.json: Genişletilmiş ChatML ve Tool-Call özel belirteçlerini içeren tokenizer.
💬 ChatML Şablonu & Kullanım
import torch
from tokenizers import Tokenizer
from dorina_v4_architecture import DorinaDenseV4, DenseConfig
device = "cuda" if torch.cuda.is_available() else "cpu"
tok = Tokenizer.from_file("dorina_tokenizer_v4_sft.json")
cfg = DenseConfig(vocab_size=32832, max_seq_len=2048)
model = DorinaDenseV4(cfg)
ckpt = torch.load("dorina_v4_gold_sft_final.pt", map_location="cpu", weights_only=False)
sd = ckpt["model"] if "model" in ckpt else ckpt
model.load_state_dict({k.replace("module.", ""): v for k, v in sd.items()})
model.to(device).eval()
user_msg = "Merhaba Dorina, bana Python ile iki sayıyı toplayan bir fonksiyon yazar mısın?"
prompt = f"<|im_start|>user\n{user_msg}<|im_end|>\n<|im_start|>assistant\n"
inp = torch.tensor([tok.encode(prompt).ids], device=device)
im_end_id = tok.token_to_id("<|im_end|>")
with torch.no_grad():
for _ in range(120):
logits, _ = model(inp)
next_tok = torch.argmax(logits[:, -1, :], dim=-1, keepdim=True)
inp = torch.cat([inp, next_tok], dim=1)
if next_tok.item() == im_end_id:
break
reply = tok.decode(inp[0].tolist()).split("<|im_start|>assistant\n")[-1].replace("<|im_end|>", "").strip()
print("Dorina:", reply)
- Downloads last month
- 42
Hardware compatibility
Log In to add your hardware
16-bit