Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
---
|
4 |
+
# FINGU-AI/Q-Small-3B
|
5 |
+
|
6 |
+
## Overview
|
7 |
+
`FINGU-AI/Q-Small-3B` is a powerful causal language model designed for a variety of natural language processing (NLP) tasks, including machine translation, text generation, and chat-based applications. This model is particularly useful for translating between languages, as well as supporting other custom NLP tasks through flexible input.
|
8 |
+
|
9 |
+
## Example Usage
|
10 |
+
|
11 |
+
### Installation
|
12 |
+
Make sure to install the required packages:
|
13 |
+
|
14 |
+
```bash
|
15 |
+
pip install torch transformers
|
16 |
+
```
|
17 |
+
### Loading the Model
|
18 |
+
|
19 |
+
```python
|
20 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
21 |
+
import torch
|
22 |
+
|
23 |
+
# Model and Tokenizer
|
24 |
+
model_id = 'FINGU-AI/Q-Small-3B'
|
25 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, attn_implementation="sdpa", torch_dtype=torch.bfloat16)
|
26 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
27 |
+
model.to('cuda')
|
28 |
+
|
29 |
+
# Input Messages for Translation
|
30 |
+
messages = [
|
31 |
+
{"role": "system", "content": "you are helpfull assistant."},
|
32 |
+
{"role": "user", "content": """what is large language model?"""},
|
33 |
+
]
|
34 |
+
|
35 |
+
# Tokenize and Generate Response
|
36 |
+
input_ids = tokenizer.apply_chat_template(
|
37 |
+
messages,
|
38 |
+
add_generation_prompt=True,
|
39 |
+
return_tensors="pt"
|
40 |
+
).to('cuda')
|
41 |
+
|
42 |
+
outputs = model.generate(
|
43 |
+
input_ids,
|
44 |
+
max_new_tokens=500,
|
45 |
+
do_sample=True,
|
46 |
+
)
|
47 |
+
|
48 |
+
# Decode and Print the Translation
|
49 |
+
response = outputs[0][input_ids.shape[-1]:]
|
50 |
+
print(tokenizer.decode(response, skip_special_tokens=True))
|
51 |
+
```
|