File size: 1,888 Bytes
d449243
 
d42a314
 
3c8ddd3
 
 
 
d449243
3c8ddd3
 
d449243
d42a314
d449243
d42a314
d449243
 
 
d42a314
d449243
d42a314
 
 
 
d449243
d42a314
 
 
 
 
 
 
d449243
d42a314
 
 
 
 
 
d449243
d42a314
d449243
d42a314
d449243
d42a314
 
 
 
d449243
d42a314
 
d449243
d42a314
 
 
d449243
d42a314
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
---
library_name: transformers
base_model:
- unsloth/Llama-3.2-1B-Instruct
license: llama3.2
language:
- en
- it
---
# A tiny Llama model tuned for text translation
(A very italian Llama model)[llamaestro-sm.png]

## Model Card for Model ID

This model was finetuned with roughly 300.000 examples of translations from English to Italian and Italian to English. The model was finetuned in a way to more directly provide a translation without much explaination.



## Usage 

```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel

base_model_id = "unsloth/Llama-3.2-1B-Instruct"
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_use_double_quant=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16
)

base_model = AutoModelForCausalLM.from_pretrained(
    base_model_id,  # Mistral, same as before
    quantization_config=bnb_config,  # Same quantization config as before
    device_map="auto",
    trust_remote_code=True,
)

tokenizer = AutoTokenizer.from_pretrained(base_model_id, add_bos_token=True, trust_remote_code=True)

ft_model = PeftModel.from_pretrained(base_model, "finetuned_model_35000")

row_json = [
    {"role": "system", "content": "Your job is to return translations for sentences or words from either Italian to English or English to Italian."},
    {"role": "user", "content": "Scontri a Bologna, la destra lancia l'offensiva contro i centri sociali."}
]

prompt =  tokenizer.apply_chat_template(row_json, tokenize=False)
model_input = tokenizer(prompt, return_tensors="pt").to("cuda")

with torch.no_grad():
    print(tokenizer.decode(ft_model.generate(**model_input, max_new_tokens=1024)[0]))
```

## Data used 
The source for the data were sentence pairs from tatoeba.com. The data can be downloaded from here: https://tatoeba.org/de/downloads