Instructions to use ConvexAI/Luminex-34B-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ConvexAI/Luminex-34B-v0.1 with Transformers:

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

pipe = pipeline("text-generation", model="ConvexAI/Luminex-34B-v0.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ConvexAI/Luminex-34B-v0.1")
model = AutoModelForCausalLM.from_pretrained("ConvexAI/Luminex-34B-v0.1")
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 ConvexAI/Luminex-34B-v0.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ConvexAI/Luminex-34B-v0.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ConvexAI/Luminex-34B-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ConvexAI/Luminex-34B-v0.1

SGLang

How to use ConvexAI/Luminex-34B-v0.1 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 "ConvexAI/Luminex-34B-v0.1" \
    --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": "ConvexAI/Luminex-34B-v0.1",
		"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 "ConvexAI/Luminex-34B-v0.1" \
        --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": "ConvexAI/Luminex-34B-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ConvexAI/Luminex-34B-v0.1 with Docker Model Runner:
```
docker model run hf.co/ConvexAI/Luminex-34B-v0.1
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

ConvexAI/Luminex-34B-v0.1

This model is Smaug-34b with LaserRMT applied.

Join our Discord!

Open Portuguese LLM Leaderboard Evaluation Results

Detailed results can be found here and on the 🚀 Open Portuguese LLM Leaderboard

Metric	Value
Average	72.76
ENEM Challenge (No Images)	72.01
BLUEX (No Images)	64.81
OAB Exams	54.49
Assin2 RTE	91.91
Assin2 STS	81.31
FaQuAD NLI	82.27
HateBR Binary	69.84
PT Hate Speech Binary	70.81
tweetSentBR	67.44

Downloads last month: 8,251

Safetensors

Model size

34B params

Tensor type

F16

Model tree for ConvexAI/Luminex-34B-v0.1

Base model

jondurbin/bagel-34b-v0.2

Finetuned

(13)

this model

Quantizations

3 models

Spaces using ConvexAI/Luminex-34B-v0.1 2

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

73.630
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

86.590
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

76.550
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

69.680
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

83.430
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

72.480
accuracy on ENEM Challenge (No Images)
Open Portuguese LLM Leaderboard

72.010
accuracy on BLUEX (No Images)
Open Portuguese LLM Leaderboard

64.810