Instructions to use TensaZangetsu/condensed-bert-vulnerable with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TensaZangetsu/condensed-bert-vulnerable with Transformers:

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

pipe = pipeline("text-generation", model="TensaZangetsu/condensed-bert-vulnerable")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TensaZangetsu/condensed-bert-vulnerable")
model = AutoModelForCausalLM.from_pretrained("TensaZangetsu/condensed-bert-vulnerable", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use TensaZangetsu/condensed-bert-vulnerable with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TensaZangetsu/condensed-bert-vulnerable"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TensaZangetsu/condensed-bert-vulnerable",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/TensaZangetsu/condensed-bert-vulnerable

SGLang

How to use TensaZangetsu/condensed-bert-vulnerable 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 "TensaZangetsu/condensed-bert-vulnerable" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TensaZangetsu/condensed-bert-vulnerable",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "TensaZangetsu/condensed-bert-vulnerable" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TensaZangetsu/condensed-bert-vulnerable",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use TensaZangetsu/condensed-bert-vulnerable with Docker Model Runner:
```
docker model run hf.co/TensaZangetsu/condensed-bert-vulnerable
```

condensed-bert-vulnerable / validator.py

TensaZangetsu

Add initial files

b7318af verified over 2 years ago

Raw History Blame Contribute Delete

1.42 kB

import argparse
from transformers import GPT2LMHeadModel, GPT2Tokenizer


def test_model(prompt, model_path='/Users/raghul.v/Desktop/research/pii_extraction_test/results', model_name='distilgpt2'):  # Adjust the default model_name
    tokenizer = GPT2Tokenizer.from_pretrained(model_name)
    model = GPT2LMHeadModel.from_pretrained(model_path)  # Loading the fine-tuned model

    # Note: Adjustments for running the model on a CUDA device are commented out,
    # Uncomment and use those if CUDA is available and desired.

    # Encode the prompt to test
    generated = tokenizer.encode(prompt, return_tensors="pt")

    # Generate text from the model
    sample_outputs = model.generate(
        generated,
        do_sample=True,
        max_length=50,
        top_k=50,
        top_p=0.95,
        num_return_sequences=3
    )

    for idx, sample_output in enumerate(sample_outputs):
        decoded_output = tokenizer.decode(sample_output, skip_special_tokens=True)
        print(f"Generated Text {idx}: {decoded_output}")


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description="Enter the prompt for the model.")
    parser.add_argument('--prompt', type=str, required=True, help='Prompt for the model')
    args = parser.parse_args()

    # Default model_path is 'results', adjust if the model is saved elsewhere
    test_model(args.prompt, model_path='results', model_name='distilgpt2')

	import argparse
	from transformers import GPT2LMHeadModel, GPT2Tokenizer


	def test_model(prompt, model_path='/Users/raghul.v/Desktop/research/pii_extraction_test/results', model_name='distilgpt2'): # Adjust the default model_name
	tokenizer = GPT2Tokenizer.from_pretrained(model_name)
	model = GPT2LMHeadModel.from_pretrained(model_path) # Loading the fine-tuned model

	# Note: Adjustments for running the model on a CUDA device are commented out,
	# Uncomment and use those if CUDA is available and desired.

	# Encode the prompt to test
	generated = tokenizer.encode(prompt, return_tensors="pt")

	# Generate text from the model
	sample_outputs = model.generate(
	generated,
	do_sample=True,
	max_length=50,
	top_k=50,
	top_p=0.95,
	num_return_sequences=3
	)

	for idx, sample_output in enumerate(sample_outputs):
	decoded_output = tokenizer.decode(sample_output, skip_special_tokens=True)
	print(f"Generated Text {idx}: {decoded_output}")


	if __name__ == '__main__':
	parser = argparse.ArgumentParser(description="Enter the prompt for the model.")
	parser.add_argument('--prompt', type=str, required=True, help='Prompt for the model')
	args = parser.parse_args()

	# Default model_path is 'results', adjust if the model is saved elsewhere
	test_model(args.prompt, model_path='results', model_name='distilgpt2')