MiaLatte-Indo-Mistral-7b

MiaLatte is a derivative model of OpenMia, which is able to answer everyday questions specifically in Bahasa Indonesia (Indonesia Language).

some of GGUF: https://huggingface.co/indischepartij/MiaLatte-Indo-Mistral-7b-GGUF

Examples

MiaLatte-Indo-Mistral-7b is a merge of the following models using MergeKit:

• indischepartij/OpenMia-Indo-Mistral-7b-v2
• Obrolin/Kesehatan-7B-v0.1
• FelixChao/WestSeverus-7B-DPO-v2

🪄 Open LLM Benchmark

🧩 Configuration

slices:
models:
  - model: indischepartij/OpenMia-Indo-Mistral-7b-v2
    parameters:
      density: 0.50
      weight: 0.35
  - model: Obrolin/Kesehatan-7B-v0.1
    parameters:
      density: 0.50
      weight: 0.35
  - model: FelixChao/WestSeverus-7B-DPO-v2
    parameters:
      density: 0.50
      weight: 0.30
merge_method: dare_ties
base_model: indischepartij/OpenMia-Indo-Mistral-7b-v2
parameters:
  int8_mask: true
dtype: float16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "indischepartij/MiaLatte-Indo-Mistral-7b"
messages = [{"role": "user", "content": "Apa jenis skincare yang cocok untuk kulit berjerawat??"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

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"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	67.86
AI2 Reasoning Challenge (25-Shot)	66.55
HellaSwag (10-Shot)	85.23
MMLU (5-Shot)	63.93
TruthfulQA (0-shot)	56.04
Winogrande (5-shot)	80.35
GSM8k (5-shot)	55.04