Instructions to use Tomasal/falcon-7b-instruct-enron with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Tomasal/falcon-7b-instruct-enron with Transformers:

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

pipe = pipeline("text-generation", model="Tomasal/falcon-7b-instruct-enron", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Tomasal/falcon-7b-instruct-enron", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("Tomasal/falcon-7b-instruct-enron", trust_remote_code=True)
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 Tomasal/falcon-7b-instruct-enron with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Tomasal/falcon-7b-instruct-enron"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Tomasal/falcon-7b-instruct-enron",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Tomasal/falcon-7b-instruct-enron

SGLang

How to use Tomasal/falcon-7b-instruct-enron 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 "Tomasal/falcon-7b-instruct-enron" \
    --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": "Tomasal/falcon-7b-instruct-enron",
		"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 "Tomasal/falcon-7b-instruct-enron" \
        --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": "Tomasal/falcon-7b-instruct-enron",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Tomasal/falcon-7b-instruct-enron with Docker Model Runner:
```
docker model run hf.co/Tomasal/falcon-7b-instruct-enron
```

Model Card for Tomasal/falcon-7b-instruct-enron

This model is a part of the master thesis work: Assessing privacy vs. efficiency tradeoffs in open-source Large-Language Models, during spring 2025 with focus to investigate privace issues i opensource LLMs.

Model Details

This model is a fine-tuned version of tiiuae/falcon-7b-instruct, using LoRA (Low-Rank Adaptation). It has been traind for three epochs on the Enron email dataset: LLM-PBE/enron-email. The goal of the fine-tuning is to explore how models memorize and potentially expose sensitive content when trained on sensitive information.

Training Procedure

The model was fine-tuned using LoRA with the following configuration:

LoRA rank: 8
LoRA Alpha: 32
LoRA Dropout: 0.05
LoRA Bias: None
Optimizer: AdamW with learning rate 1e-4
Precision: bfloat16
Epochs: 3
Batch size: 2
Hardware: NVIDIA GeForce RTX 5090

How to Use

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("Tomasal/falcon-7b-instruct-enron", torch_dtype="bfloat16")
tokenizer = AutoTokenizer.from_pretrained("Tomasal/falcon-7b-instruct-enron")

messages = [{"role": "user", "content": "Can you write a professional email confirming a meeting with the legal team on Monday at 10am?"}]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(inputs, max_new_tokens=128) 
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Downloads last month: 9

Safetensors

Model size

7B params

Tensor type

BF16

Model tree for Tomasal/falcon-7b-instruct-enron

Base model

tiiuae/falcon-7b-instruct

Adapter

(81)

this model

Dataset used to train Tomasal/falcon-7b-instruct-enron

Paper for Tomasal/falcon-7b-instruct-enron

LoRA: Low-Rank Adaptation of Large Language Models

Paper • 2106.09685 • Published Jun 17, 2021 • 61