Instructions to use roadz/dv-finetuned-211124 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use roadz/dv-finetuned-211124 with Transformers:

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

pipe = pipeline("text-generation", model="roadz/dv-finetuned-211124")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("roadz/dv-finetuned-211124")
model = AutoModelForCausalLM.from_pretrained("roadz/dv-finetuned-211124")
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 roadz/dv-finetuned-211124 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "roadz/dv-finetuned-211124"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "roadz/dv-finetuned-211124",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/roadz/dv-finetuned-211124

SGLang

How to use roadz/dv-finetuned-211124 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 "roadz/dv-finetuned-211124" \
    --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": "roadz/dv-finetuned-211124",
		"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 "roadz/dv-finetuned-211124" \
        --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": "roadz/dv-finetuned-211124",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use roadz/dv-finetuned-211124 with Docker Model Runner:
```
docker model run hf.co/roadz/dv-finetuned-211124
```

Model Card for roadz/dv-finetuned-211124

This model is fine-tuned for evaluating LLM outputs in RAG scenarios, focusing on:

Hallucination detection
Attribution accuracy
Summary completeness
Response relevancy

Model Details

Model Architecture

Base Model: LLaMA-3.1-8B
Architecture Type: llama
Parameters: Not specified
Training Type: Fine-tuned for evaluation

Hardware Requirements

Minimum GPU Memory: 16GB
Recommended GPU Memory: 24GB
Format: SafeTensors

Usage

This model is designed for the De-Val subnet and requires specific pipeline code for evaluation tasks.

Generation Configuration

Max Length: Not specified
Temperature: 0.6
Top-p: 0.9
Top-k: 50

Training

The model was fine-tuned on evaluation tasks including:

Hallucination detection scenarios
Attribution verification tasks
Summary completeness assessment
Response relevancy evaluation

Limitations

Designed specifically for evaluation tasks
Requires De-Val pipeline code
Not intended for general text generation

Last Updated

2024-11-21

Downloads last month: 2

Safetensors

Model size

8B params

Tensor type

F16