File size: 2,081 Bytes
d754e91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
import os
import sys

import torch
import transformers
from peft import PeftModel
from transformers import GenerationConfig, LlamaForCausalLM, LlamaTokenizer

from .globals import Global


def get_device():
    if torch.cuda.is_available():
        return "cuda"
    else:
        return "cpu"

    try:
        if torch.backends.mps.is_available():
            return "mps"
    except:  # noqa: E722
        pass


device = get_device()


def get_base_model():
    load_base_model()
    return Global.loaded_base_model


def get_model_with_lora(lora_weights: str = "tloen/alpaca-lora-7b"):
    if device == "cuda":
        return PeftModel.from_pretrained(
            get_base_model(),
            lora_weights,
            torch_dtype=torch.float16,
        )
    elif device == "mps":
        return PeftModel.from_pretrained(
            get_base_model(),
            lora_weights,
            device_map={"": device},
            torch_dtype=torch.float16,
        )
    else:
        return PeftModel.from_pretrained(
            get_base_model(),
            lora_weights,
            device_map={"": device},
        )


def get_tokenizer():
    load_base_model()
    return Global.loaded_tokenizer


def load_base_model():
    if Global.loaded_tokenizer is None:
        Global.loaded_tokenizer = LlamaTokenizer.from_pretrained(
            Global.base_model)
    if Global.loaded_base_model is None:
        if device == "cuda":
            Global.loaded_base_model = LlamaForCausalLM.from_pretrained(
                Global.base_model,
                load_in_8bit=Global.load_8bit,
                torch_dtype=torch.float16,
                device_map="auto",
            )
        elif device == "mps":
            Global.loaded_base_model = LlamaForCausalLM.from_pretrained(
                Global.base_model,
                device_map={"": device},
                torch_dtype=torch.float16,
            )
        else:
            model = LlamaForCausalLM.from_pretrained(
                base_model, device_map={"": device}, low_cpu_mem_usage=True
            )