|
--- |
|
language: |
|
- ko |
|
library_name: transformers |
|
license: apache-2.0 |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
# **SOLAR-10.7B-v1.0-Instruct** |
|
|
|
## Model Details |
|
**Model Developers** |
|
- myeonghoon kim |
|
|
|
**Model Architecture** |
|
- SOLAR-10.7B-v1.0-Instruct is an auto-regressive language model based on the LLaMA2 transformer architecture. |
|
|
|
**Base Model** |
|
- [upstage/SOLAR-10.7B-v1.0](https://huggingface.co/upstage/SOLAR-10.7B-v1.0) |
|
|
|
**Training Dataset** |
|
- [Platypus](https://github.com/arielnlee/Platypus) |
|
|
|
--- |
|
# Model comparisons1 |
|
> Ko-LLM leaderboard(11/23; [link](https://huggingface.co/spaces/upstage/open-ko-llm-leaderboard)) |
|
|
|
| Model | Average | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 | |
|
| --- | --- | --- | --- | --- | --- | --- | |
|
| **[...your_model_name...]** | NaN | NaN | NaN | NaN | NaN | NaN | |
|
|
|
--- |
|
# Model comparisons2 |
|
> AI-Harness evaluation; [link](https://github.com/Beomi/ko-lm-evaluation-harness) |
|
|
|
| Model | Copa | Copa | HellaSwag | HellaSwag | BoolQ | BoolQ | Sentineg | Sentineg | |
|
| --- | --- | --- | --- | --- | --- | --- | --- | --- | |
|
| | 0-shot | 5-shot | 0-shot | 5-shot | 0-shot | 5-shot | 0-shot | 5-shot | |
|
| **SOLAR-10.7B-v1.0-Instruct** | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN | |
|
|
|
--- |
|
# Implementation Code |
|
```python |
|
### KO-Platypus |
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
import torch |
|
|
|
repo = "[...your_model_repo...]" |
|
OpenOrca = AutoModelForCausalLM.from_pretrained( |
|
repo, |
|
return_dict=True, |
|
torch_dtype=torch.float16, |
|
device_map='auto' |
|
) |
|
OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo) |
|
``` |
|
|
|
--- |