Instructions to use recogna-nlp/gembode-2b-it-ultraalpaca with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use recogna-nlp/gembode-2b-it-ultraalpaca with Transformers:

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

pipe = pipeline("text-generation", model="recogna-nlp/gembode-2b-it-ultraalpaca")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("recogna-nlp/gembode-2b-it-ultraalpaca")
model = AutoModelForCausalLM.from_pretrained("recogna-nlp/gembode-2b-it-ultraalpaca")
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 recogna-nlp/gembode-2b-it-ultraalpaca with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "recogna-nlp/gembode-2b-it-ultraalpaca"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "recogna-nlp/gembode-2b-it-ultraalpaca",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/recogna-nlp/gembode-2b-it-ultraalpaca

SGLang

How to use recogna-nlp/gembode-2b-it-ultraalpaca 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 "recogna-nlp/gembode-2b-it-ultraalpaca" \
    --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": "recogna-nlp/gembode-2b-it-ultraalpaca",
		"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 "recogna-nlp/gembode-2b-it-ultraalpaca" \
        --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": "recogna-nlp/gembode-2b-it-ultraalpaca",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use recogna-nlp/gembode-2b-it-ultraalpaca with Docker Model Runner:
```
docker model run hf.co/recogna-nlp/gembode-2b-it-ultraalpaca
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

gembode-2b-ultraalpaca

Phi-Bode Logo

GemmBode é um modelo de linguagem ajustado para o idioma português, desenvolvido a partir do modelo base de instruções Gemma-2b-it fornecido pela Google. Este modelo foi refinado através do processo de fine-tuning utilizando o dataset UltraAlpaca. O principal objetivo deste modelo é ser viável para pessoas que não possuem recursos computacionais disponíveis para o uso de LLMs (Large Language Models). Ressalta-se que este é um trabalho em andamento e o modelo ainda apresenta problemas na geração de texto em português.

Características Principais

Modelo Base: Gemma-2b-it, criado pela Google, com 2 bilhões de parâmetros.
Dataset para Fine-tuning: UltraAlpaca
Treinamento: O treinamento foi realizado a partir do fine-tuning completo do gemma-2b-it.

Open Portuguese LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Average	45.69
ENEM Challenge (No Images)	34.71
BLUEX (No Images)	25.87
OAB Exams	31.71
Assin2 RTE	71.31
Assin2 STS	34.08
FaQuAD NLI	60.09
HateBR Binary	47.01
PT Hate Speech Binary	57.04
tweetSentBR	49.37

Downloads last month: 12

Safetensors

Model size

3B params

Tensor type

F16

Model tree for recogna-nlp/gembode-2b-it-ultraalpaca

Quantizations

1 model

Collection including recogna-nlp/gembode-2b-it-ultraalpaca

GemBode

Collection

GemBode são modelos baseados na arquitetura Gemma, da Google. Aqui, você pode encontrar diversos modelos, sendo eles: • 7 items • Updated Jul 11, 2024

Evaluation results

accuracy on ENEM Challenge (No Images)
Open Portuguese LLM Leaderboard

34.710
accuracy on BLUEX (No Images)
Open Portuguese LLM Leaderboard

25.870
accuracy on OAB Exams
Open Portuguese LLM Leaderboard

31.710
f1-macro on Assin2 RTE
test set Open Portuguese LLM Leaderboard

71.310
pearson on Assin2 STS
test set Open Portuguese LLM Leaderboard

34.080
f1-macro on FaQuAD NLI
test set Open Portuguese LLM Leaderboard

60.090
f1-macro on HateBR Binary
test set Open Portuguese LLM Leaderboard

47.010
f1-macro on PT Hate Speech Binary
test set Open Portuguese LLM Leaderboard

57.040
f1-macro on tweetSentBR
test set Open Portuguese LLM Leaderboard

49.370