README.md

---
datasets:
- kyujinpy/KOpen-platypus
language:
- ko
- en
pipeline_tag: text-generation
---

### Model Card for Model ID
base_model : [Ko-Llama3-Luxia-8B](https://huggingface.co./saltlux/Ko-Llama3-Luxia-8B)

### Basic usage
```python
# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

tokenizer = AutoTokenizer.from_pretrained("MDDDDR/Ko-Luxia-8B-it-v0.2")
model = AutoModelForCausalLM.from_pretrained(
    "MDDDDR/Ko-Luxia-8B-it-v0.2",
    device_map="auto",
    torch_dtype=torch.bfloat16
)

input_text = "사과가 뭐야?"
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")

outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```

### Training dataset
dataset : [kyujinpy/KOpen-platypus](https://huggingface.co./datasets/kyujinpy/KOpen-platypus)

bnd_config = BitsAndBytesConfig(
  load_in_4bit = True
)

lora_config = LoraConfig(
  r = 16,
  lora_alpha = 16,
  lora_dropout = 0.05,
  target_modules = ['down_proj', 'up_proj', 'gate_proj']
)

### Hardware
RTX 3090 Ti 24GB x 1

### Model Benchmark Results
|      Tasks      |Version|Filter|n-shot|  Metric   |   |Value |   |Stderr|
|-----------------|-------|------|-----:|-----------|---|-----:|---|------|
|kobest_boolq     |      1|none  |     0|acc        |↑  |0.5278|±  |0.0133|
|                 |       |none  |     0|f1         |↑  |0.3954|±  |N/A   |
|kobest_copa      |      1|none  |     0|acc        |↑  |0.7380|±  |0.0139|
|                 |       |none  |     0|f1         |↑  |0.7372|±  |N/A   |
|kobest_hellaswag |      1|none  |     0|acc        |↑  |0.4800|±  |0.0224|
|                 |       |none  |     0|acc_norm   |↑  |0.6180|±  |0.0218|
|                 |       |none  |     0|f1         |↑  |0.4774|±  |N/A   |
|kobest_sentineg  |      1|none  |     0|acc        |↑  |0.5390|±  |0.0250|
|                 |       |none  |     0|f1         |↑  |0.5037|±  |N/A   |