Update README.md
Browse files
README.md
CHANGED
@@ -8,3 +8,67 @@ model-index:
|
|
8 |
- name: Qwen-Orpo-v1
|
9 |
results: []
|
10 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
- name: Qwen-Orpo-v1
|
9 |
results: []
|
10 |
---
|
11 |
+
|
12 |
+
## FINGU-AI/Qwen-Orpo-v1
|
13 |
+
|
14 |
+
### Overview
|
15 |
+
|
16 |
+
The FINGU-AI/Qwen-Orpo-v1 model offers a specialized curriculum tailored to English, speakers interested in finance, investment, and legal frameworks. It aims to enhance language proficiency while providing insights into global finance markets and regulatory landscapes.
|
17 |
+
|
18 |
+
### Key Features
|
19 |
+
|
20 |
+
- **Global Perspective**: Explores diverse financial markets and regulations across English, Korean, and Japanese contexts.
|
21 |
+
- **Language Proficiency**: Enhances language skills in English, Korean, and Japanese for effective communication in finance and legal domains.
|
22 |
+
- **Career Advancement**: Equips learners with knowledge and skills for roles in investment banking, corporate finance, asset management, and regulatory compliance.
|
23 |
+
|
24 |
+
### Model Information
|
25 |
+
|
26 |
+
- **Model Name**: FINGU-AI/Qwen-Orpo-v1
|
27 |
+
- **Description**: FINGU-AI/Qwen-Orpo-v1 model trained on various languages, including English.
|
28 |
+
- **Checkpoint**: FINGU-AI/Qwen-Orpo-v1
|
29 |
+
- **Author**: Grinda AI Inc.
|
30 |
+
- **License**: Apache-2.0
|
31 |
+
|
32 |
+
### Training Details
|
33 |
+
|
34 |
+
- **Fine-Tuning**: The model was fine-tuned on the base model Qwen/Qwen1.5-0.5B-Chat through supervised fine-tuning using the TrL Library and Transformer.
|
35 |
+
- **Dataset**: The fine-tuning dataset consisted of 28k training samples.
|
36 |
+
|
37 |
+
|
38 |
+
### How to Use
|
39 |
+
|
40 |
+
To use the FINGU-AI/Qwen-Orpo-v1 model, you can utilize the Hugging Face Transformers library.
|
41 |
+
Here's a Python code snippet demonstrating how to load the model and generate predictions:
|
42 |
+
|
43 |
+
```python
|
44 |
+
import torch
|
45 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer, AutoConfig,TextStreamer
|
46 |
+
|
47 |
+
|
48 |
+
model_id = 'FINGU-AI/Qwen-Orpo-v1'
|
49 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, attn_implementation="flash_attention_2", torch_dtype= torch.bfloat16)
|
50 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
51 |
+
streamer = TextStreamer(tokenizer)
|
52 |
+
model.to('cuda')
|
53 |
+
|
54 |
+
|
55 |
+
|
56 |
+
messages = [
|
57 |
+
{"role": "system","content": " you are as a finance specialist, help the user and provide accurat information."},
|
58 |
+
{"role": "user", "content": " what are the best approch to prevent loss?"},
|
59 |
+
]
|
60 |
+
tokenized_chat = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to("cuda")
|
61 |
+
|
62 |
+
generation_params = {
|
63 |
+
'max_new_tokens': 1000,
|
64 |
+
'use_cache': True,
|
65 |
+
'do_sample': True,
|
66 |
+
'temperature': 0.7,
|
67 |
+
'top_p': 0.9,
|
68 |
+
'top_k': 50,
|
69 |
+
}
|
70 |
+
|
71 |
+
outputs = model.generate(tokenized_chat, **generation_params, streamer=streamer)
|
72 |
+
decoded_outputs = tokenizer.batch_decode(outputs)
|
73 |
+
|
74 |
+
```
|