Upload README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,56 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
license: cc-by-nc-sa-4.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
- ko
|
5 |
+
datasets:
|
6 |
+
- DopeorNope/Robustness_Ko_data-v1
|
7 |
+
pipeline_tag: text-generation
|
8 |
license: cc-by-nc-sa-4.0
|
9 |
---
|
10 |
+
|
11 |
+
# **SOLAR-tail-10.7B-instruct-v1.0**
|
12 |
+
|
13 |
+
## Model Details
|
14 |
+
|
15 |
+
**Model Developers** Kyujin Han (kyujinpy)
|
16 |
+
|
17 |
+
**Method**
|
18 |
+
Instruction-tuning with [PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0](https://huggingface.co/PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0).
|
19 |
+
|
20 |
+
**Datasets**
|
21 |
+
datasets: DopeorNope/Robustness_Ko_data-v1(private).
|
22 |
+
|
23 |
+
**Hyperparameters**
|
24 |
+
(I will update all!)
|
25 |
+
|
26 |
+
# **Model Benchmark**
|
27 |
+
|
28 |
+
## Open leaderboard
|
29 |
+
- Follow up as [link](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).
|
30 |
+
|
31 |
+
| Model | Average | ARC | HellaSwag | MMLU | TruthfulQA | Ko-CommonGenV2 |
|
32 |
+
| --- | --- | --- | --- | --- | --- | --- |
|
33 |
+
| PracticeLLM/SOLAR-tail-10.7B-instruct-v1.0 | NaN | NaN | NaN | NaN | NaN | NaN |
|
34 |
+
| PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0 | NaN | NaN | NaN | NaN | NaN | NaN |
|
35 |
+
| jjourney1125/M-SOLAR-10.7B-v1.0 | 55.15 | 49.57 | 60.12 | 54.60 | 49.23 | 62.22 |
|
36 |
+
| beomi/Yi-Ko-6B | 48.79 | 41.04 | 53.39 | 46.28 | 41.64 | 61.63 |
|
37 |
+
| mistralai/Mistral-7B-v0.1 | 46.89 | 38.14 | 48.19 | 45.20 | 46.13 | 56.79 |
|
38 |
+
|
39 |
+
|
40 |
+
# Implementation Code
|
41 |
+
```python
|
42 |
+
### KO-Platypus
|
43 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
44 |
+
import torch
|
45 |
+
|
46 |
+
repo = "PracticeLLM/SOLAR-tail-10.7B-instruct-v1.0"
|
47 |
+
OpenOrca = AutoModelForCausalLM.from_pretrained(
|
48 |
+
repo,
|
49 |
+
return_dict=True,
|
50 |
+
torch_dtype=torch.float16,
|
51 |
+
device_map='auto'
|
52 |
+
)
|
53 |
+
OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo)
|
54 |
+
```
|
55 |
+
|
56 |
+
---
|