Update README.md
Browse files
README.md
CHANGED
|
@@ -6,18 +6,27 @@ library_name: peft
|
|
| 6 |
pipeline_tag: text-generation
|
| 7 |
---
|
| 8 |
|
| 9 |
-
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
|
| 13 |
-
|
|
|
|
|
|
|
| 14 |
|
|
|
|
| 15 |
The following `bitsandbytes` quantization config was used during training:
|
| 16 |
- **load_in_4bit**: True
|
| 17 |
- **bnb_4bit_quant_type**: nf4
|
| 18 |
- **bnb_4bit_use_double_quant**: False
|
| 19 |
-
- **bnb_4bit_compute_dtype**:
|
| 20 |
|
| 21 |
-
|
| 22 |
-
|
| 23 |
- PEFT 0.8.2.
|
|
|
|
| 6 |
pipeline_tag: text-generation
|
| 7 |
---
|
| 8 |
|
| 9 |
+
<div align='left'>
|
| 10 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/6370a4e53d1bd47a4ebc2120/TQSWE0e3dAO_Ksbb8b5Xd.png" width='30%'/>
|
| 11 |
+
<h1>"WelSSiSKo : Welfare Domain Specific Model"</h1>
|
| 12 |
+
</div>
|
| 13 |
+
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# What is BaseModel â–¼
|
| 17 |
+
👉 [beomi/llama-2-ko-7b](https://huggingface.co/beomi/llama-2-ko-7b)
|
| 18 |
|
| 19 |
|
| 20 |
+
# Github â–¼
|
| 21 |
+
👉 [Github Repo](https://github.com/ash-hun/WelSSISKo)
|
| 22 |
+
|
| 23 |
|
| 24 |
+
# Training procedure â–¼
|
| 25 |
The following `bitsandbytes` quantization config was used during training:
|
| 26 |
- **load_in_4bit**: True
|
| 27 |
- **bnb_4bit_quant_type**: nf4
|
| 28 |
- **bnb_4bit_use_double_quant**: False
|
| 29 |
+
- **bnb_4bit_compute_dtype**: torch.bfloat16
|
| 30 |
|
| 31 |
+
# Framework versions â–¼
|
|
|
|
| 32 |
- PEFT 0.8.2.
|