Instructions to use bertin-project/Gromenauer-7B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bertin-project/Gromenauer-7B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="bertin-project/Gromenauer-7B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("bertin-project/Gromenauer-7B-Instruct")
model = AutoModelForCausalLM.from_pretrained("bertin-project/Gromenauer-7B-Instruct")
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 bertin-project/Gromenauer-7B-Instruct with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/bertin-project/Gromenauer-7B-Instruct

SGLang

How to use bertin-project/Gromenauer-7B-Instruct 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 "bertin-project/Gromenauer-7B-Instruct" \
    --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": "bertin-project/Gromenauer-7B-Instruct",
		"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 "bertin-project/Gromenauer-7B-Instruct" \
        --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": "bertin-project/Gromenauer-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use bertin-project/Gromenauer-7B-Instruct with Docker Model Runner:
```
docker model run hf.co/bertin-project/Gromenauer-7B-Instruct
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Gromenauer-7B-Instruct

Overview

Gromenauer-7B-Instruct is an instruct fine-tuned version of the bertin-project/Gromenauer-7B model using the bertin-project/bonanza-hf and bertin-project/zenobia-instruct-hf datasets.

Usage examples

Multinomial sampling example:

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

messages = [
    {"role": "system", "content": "Eres un modelo experto en poesía española."},
    {"role": "user", "content": "Escribe un poema sobre la pérdida de un coche querido en forma de pareado."},
]

generate_kwargs = {
    "do_sample": True,
    "temperature": 0.7,
    "max_new_tokens": 150,
}

pipe = pipeline("text-generation", model="bertin-project/Gromenauer-7B-Instruct", generate_kwargs=generate_kwargs)
pipe(messages)

Output:

<|system|>
Eres un modelo experto en poesía española.</s> 
<|user|>
Escribe un poema sobre la pérdida de un coche querido en forma de pareado.</s> 
<|assistant|>
Una mañana de invierno salí al sol peregrino,
y encontré mi auto cogiendo una lechuga en el camino.</s>

Contrastive search example:

messages = [
    {"role": "system", "content": "Eres un asistente en español. Responde de manera exacta y concisa."},
    {"role": "user", "content": "¿Por qué es famosa Sevilla?"},
]

generate_kwargs = {
    "penalty_alpha": 0.6,
    "max_new_tokens": 300,
}
pipe = pipeline("text-generation", model="bertin-project/Gromenauer-7B-Instruct", generate_kwargs=generate_kwargs)
pipe(messages)

Output:

<|system|>
Eres un asistente en español. Responde de manera exacta y concisa.</s> 
<|user|>
¿Por qué es famosa Sevilla?</s> 
<|assistant|>
Sevilla es conocida por su belleza arquitectónica, con edificios como la Giralda, el Alcázar y la Catedral, así como por sus fiestas populares como la Feria de Abril y Semana Santa. Además, es la capital de Andalucía y uno de los principales centros económicos del sur de España.</s>

Model Details

Model Type: Mistral
Sequence Length: 8192
Hidden Dimension: 4096
Intermediate Dimension: 14336
Number of Layers: 32
Number of Attention Heads: 32
Number of Key-Value Heads: 8
Activation Function: SiLU
Initializer Range: 0.02
Layer Norm Epsilon: 1.0e-05
Use Flash Attention: Yes
Gradient Checkpointing: Enabled (Block Size: 5)
Sliding Window Attention: 4096
Use Bias: No

Training Details

Tokenizer: HuggingFaceH4/zephyr-7b-beta
Batch Size: 512
Learning Rate: 1e-5
Optimizer: Adam with beta1=0.9, beta2=0.95, epsilon=1e-8
Weight Decay: 0.1
Warmup Steps: 200
Learning Rate Schedule: Cosine
Number of Training Epochs: 5

Downloads last month: 5

Safetensors

Model size

7B params

Tensor type

F32

Model tree for bertin-project/Gromenauer-7B-Instruct

Base model

mistralai/Mistral-7B-v0.1

Finetuned

bertin-project/Gromenauer-7B

Finetuned

(1)

this model

Quantizations

3 models

bertin-project
/

Gromenauer-7B-Instruct