File size: 1,532 Bytes
d19f767
 
 
 
 
 
 
 
 
 
 
 
f3a6a81
 
9a220aa
 
f3a6a81
 
 
 
 
 
 
 
 
 
 
 
 
 
9a220aa
f3a6a81
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ms
---

# Pretrain 1.1B 4096 context length Mistral on Malaysian text

README at https://github.com/mesolitica/malaya/tree/5.1/pretrained-model/mistral

- Dataset gathered at https://github.com/malaysia-ai/dedup-text-dataset/tree/main/pretrain-llm
- We use Ray cluster to train on 5 nodes of 4x A100 80GB, https://github.com/malaysia-ai/jupyter-gpu/tree/main/ray

WandB, https://wandb.ai/mesolitica/pretrain-mistral-1.1b?workspace=user-husein-mesolitica

WandB report, https://wandb.ai/mesolitica/pretrain-mistral-3b/reports/Pretrain-Larger-Malaysian-Mistral--Vmlldzo2MDkyOTgz

## how-to

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

TORCH_DTYPE = 'bfloat16'
nf4_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type='nf4',
    bnb_4bit_use_double_quant=True,
    bnb_4bit_compute_dtype=getattr(torch, TORCH_DTYPE)
)

tokenizer = AutoTokenizer.from_pretrained('mesolitica/malaysian-mistral-1.1B-4096', model_input_names = ['input_ids'])
model = AutoModelForCausalLM.from_pretrained(
    'mesolitica/malaysian-mistral-1.1B-4096',
    use_flash_attention_2 = True,
    quantization_config = nf4_config
)
prompt = '<s>nama saya'
inputs = tokenizer([prompt], return_tensors='pt', add_special_tokens=False).to('cuda')

generate_kwargs = dict(
    inputs,
    max_new_tokens=512,
    top_p=0.95,
    top_k=50,
    temperature=0.9,
    do_sample=True,
    num_beams=1,
    repetition_penalty=1.05,
)
r = model.generate(**generate_kwargs)
```