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

Libraries

How to use Kkaastr/Devple-8B with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Kkaastr/Devple-8B")
model = AutoModelForCausalLM.from_pretrained("Kkaastr/Devple-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 Kkaastr/Devple-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Kkaastr/Devple-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": "Kkaastr/Devple-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Kkaastr/Devple-8B

SGLang

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

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

Model Card

Devple is a fine-tuned model based on Llama 3.1 Instruct, designed for development tasks such as code generation and review, with a focus on the quality and safety of the generated code. Its synthetic dataset was generated using GPT-4o with Llama-3 (rejected).

Model Details

Model Description

Devple is a fine-tuned model based on Llama 3.1 Instruct. The model is built on a synthetic dataset. The main focus of the training was on development-related tasks such as code generation, code review, refactoring, etc., with particular emphasis on the quality and safety of the generated code.

Fine-tuning was done using ORPO. The dataset was generated using GPT-4o (chosen) and Llama-3 (rejected).

Language(s) (NLP): English, Russian
Finetuned from model: Llama 3.1 Instruct

Uses

Direct Use

import transformers
import torch

model_id = "Kkaastr/Devple-8B"

pipeline = transformers.pipeline(
    "text-generation",
    model=model_id,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
    {"role": "user", "content": "Who are you?"},
]

outputs = pipeline(
    messages,
    max_new_tokens=256,
)
print(outputs[0]["generated_text"][-1])

Downloads last month: 6

Safetensors

Model size

8B params

Tensor type

F16

Model tree for Kkaastr/Devple-8B

Base model

meta-llama/Llama-3.1-8B

Finetuned

meta-llama/Llama-3.1-8B-Instruct

Finetuned

(2760)

this model

Quantizations

2 models