Lumina-3

Lumina-3 is a Mixture of Experts (MoE) using LazyMergekit. This model uses a context window of up to 32k.

🏆 Open LLM Leaderboard Evaluation Results

Metric	Value
Avg.	74.53
AI2 Reasoning Challenge (25-Shot)	71.16
HellaSwag (10-Shot)	87.20
MMLU (5-Shot)	65.52
TruthfulQA (0-shot)	68.25
Winogrande (5-shot)	82.08
GSM8k (5-shot)	72.93

Quants

Special thanks to GGUFs made by mradermacher

• mradermacher/Lumina-3-GGUF

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Ppoyaa/Lumina-3"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])