Telugu-LLM-Labs
commited on
Commit
•
5c5a9ce
1
Parent(s):
e9b93f6
Update model details and usage.
Browse files
README.md
ADDED
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: llama2
|
3 |
+
datasets:
|
4 |
+
- Telugu-LLM-Labs/yahma_alpaca_cleaned_telugu_filtered_and_romanized
|
5 |
+
- >-
|
6 |
+
Telugu-LLM-Labs/teknium_GPTeacher_general_instruct_telugu_filtered_and_romanized
|
7 |
+
language:
|
8 |
+
- te
|
9 |
+
- en
|
10 |
+
---
|
11 |
+
# Telugu-Llama2-7B-v0-Instruct
|
12 |
+
|
13 |
+
|
14 |
+
This model is based on [Telugu-Llama2-7B-v0-Base](https://huggingface.co/Telugu-LLM-Labs/Telugu-Llama2-7B-v0-Base) and hase been finetuned on instruction datasets:
|
15 |
+
1. [yahma_alpaca_cleaned_telugu_filtered_and_romanized](https://huggingface.co/datasets/Telugu-LLM-Labs/yahma_alpaca_cleaned_telugu_filtered_and_romanized)
|
16 |
+
2. [teknium_GPTeacher_general_instruct_telugu_filtered_and_romanized](https://huggingface.co/datasets/Telugu-LLM-Labs/teknium_GPTeacher_general_instruct_telugu_filtered_and_romanized)
|
17 |
+
|
18 |
+
# Input Text Format
|
19 |
+
|
20 |
+
```
|
21 |
+
### Instruction: {instruction}
|
22 |
+
|
23 |
+
### Input: {input}
|
24 |
+
|
25 |
+
## Response: {response}
|
26 |
+
```
|
27 |
+
|
28 |
+
# Usage
|
29 |
+
```python3
|
30 |
+
import torch
|
31 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
32 |
+
|
33 |
+
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
34 |
+
|
35 |
+
model_name = "Telugu-LLM-Labs/Telugu-Llama2-7B-v0-Instruct"
|
36 |
+
|
37 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name, padding_side="right")
|
38 |
+
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16).to(device)
|
39 |
+
|
40 |
+
text = f"""Instruction: {instruction} \nInput: {input} \nResponse:"""
|
41 |
+
|
42 |
+
encodings = tokenizer(text, padding=True, return_tensors="pt")
|
43 |
+
encodings = encodings.to(device)
|
44 |
+
|
45 |
+
with torch.inference_mode():
|
46 |
+
outputs = model.generate(encodings.input_ids, do_sample=False, max_new_tokens=100)
|
47 |
+
|
48 |
+
output = tokenizer.batch_decode(outputs.detach(), skip_special_tokens=True)
|
49 |
+
```
|
50 |
+
|
51 |
+
## Examples:
|
52 |
+
|
53 |
+
### With Romanized Telugu
|
54 |
+
|
55 |
+
```
|
56 |
+
instruction = "Krindi samaacharam prakaram google app eppudu release ayyindi?"
|
57 |
+
input ="Google News is a news aggregator service developed by Google. It presents a continuous flow of links to articles organized from thousands of publishers and magazines. Google News is available as an app on Android, iOS, and the Web. Google released a beta version in September 2002 and the official app in January 2006."
|
58 |
+
```
|
59 |
+
|
60 |
+
### With Native Telugu
|
61 |
+
|
62 |
+
```
|
63 |
+
instruction = "కింది వచనాన్ని సంగ్రహించండి"
|
64 |
+
input="గూగుల్ వార్తలు అనేది గూగుల్ ద్వారా అభివృద్ధి చేయబడిన వార్తా అగ్రిగేటర్ సేవ. ఇది వేలకొద్దీ ప్రచురణకర్తలు మరియు మ్యాగజైన్ల నుండి నిర్వహించబడిన కథనాలకు నిరంతర లింక్లను అందిస్తుంది. గూగుల్ వార్తలు Android, iOS మరియు వెబ్లో యాప్గా అందుబాటులో ఉన్నాయి. గూగుల్ సెప్టెంబరు 2002లో బీటా వెర్షన్ను మరియు జనవరి 2006లో అధికారిక యాప్ను విడుదల చేసింది."
|
65 |
+
```
|
66 |
+
|
67 |
+
# Developers:
|
68 |
+
|
69 |
+
The model is a collaborative effort by [Ravi Theja](https://twitter.com/ravithejads) and [Ramsri Goutham](https://twitter.com/ramsri_goutham). Feel free to DM either of us if you have any questions.
|
70 |
+
|
71 |
+
# Note:
|
72 |
+
|
73 |
+
The model is quite sensitive to parameters and inputs and is not yet ready for production. It remains in the experimental phase, and we recommend using it accordingly.
|