File size: 6,517 Bytes
23c9ef8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4e94887
37ae470
4e94887
 
 
37ae470
4e94887
37ae470
23c9ef8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
37ae470
23c9ef8
 
 
37ae470
23c9ef8
 
 
 
 
 
37ae470
23c9ef8
 
 
 
 
37ae470
23c9ef8
 
 
d6d6d10
23c9ef8
 
 
 
 
 
 
 
 
 
 
 
 
 
4e94887
99ff09f
23c9ef8
 
 
 
 
 
 
 
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
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124

#    Copyright 2024 Xi Zhang
#
#    Licensed under the Apache License, Version 2.0 (the "License");
#    you may not use this file except in compliance with the License.
#    You may obtain a copy of the License at
#
#        http://www.apache.org/licenses/LICENSE-2.0
#
#    Unless required by applicable law or agreed to in writing, software
#    distributed under the License is distributed on an "AS IS" BASIS,
#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#    See the License for the specific language governing permissions and
#    limitations under the License.

import os
import warnings
import shutil

from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig, BitsAndBytesConfig
import torch
from libra.model import *
from libra.constants import DEFAULT_IMAGE_PATCH_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN


def load_pretrained_model(model_path, model_base, model_name, device="cpu"):

    device_map = {"": device}
    kwargs = {
        "device_map": device_map,
        "torch_dtype": torch.bfloat16 
    }
      
    if 'libra' in model_name.lower():
        # Load Libra model
        if 'lora' in model_name.lower() and model_base is None:
            warnings.warn('There is `lora` in model name but no `model_base` is provided. If you are loading a LoRA model, please provide the `model_base` argument.')
        if 'lora' in model_name.lower() and model_base is not None:
            from libra.model.language_model.libra_llama import LibraConfig
            lora_cfg_pretrained = LibraConfig.from_pretrained(model_path)
            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
            print('Loading libra from base model...')
            model = LibraLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=lora_cfg_pretrained, **kwargs)
            token_num, tokem_dim = model.lm_head.out_features, model.lm_head.in_features
            if model.lm_head.weight.shape[0] != token_num:
                model.lm_head.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))
                model.model.embed_tokens.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))

            print('Loading additional Libra weights...')
            if os.path.exists(os.path.join(model_path, 'non_lora_trainables.bin')):
                non_lora_trainables = torch.load(os.path.join(model_path, 'non_lora_trainables.bin'), map_location='cpu')
            else:
                from huggingface_hub import hf_hub_download
                def load_from_hf(repo_id, filename, subfolder=None):
                    cache_file = hf_hub_download(
                        repo_id=repo_id,
                        filename=filename,
                        subfolder=subfolder)
                    return torch.load(cache_file, map_location='cpu')
                non_lora_trainables = load_from_hf(model_path, 'non_lora_trainables.bin')
            non_lora_trainables = {(k[11:] if k.startswith('base_model.') else k): v for k, v in non_lora_trainables.items()}
            if any(k.startswith('model.model.') for k in non_lora_trainables):
                non_lora_trainables = {(k[6:] if k.startswith('model.') else k): v for k, v in non_lora_trainables.items()}
            model.load_state_dict(non_lora_trainables, strict=False)

            from peft import PeftModel
            print('Loading LoRA weights...')
            model = PeftModel.from_pretrained(model, model_path)
            print('Merging LoRA weights...')
            model = model.merge_and_unload()
            print('Model is loaded...')
        elif model_base is not None:
            print('Loading Libra from base model...')
            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
            cfg_pretrained = AutoConfig.from_pretrained(model_path)
            model = LibraLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=cfg_pretrained, **kwargs)

            mm_projector_weights = torch.load(os.path.join(model_path, 'mm_projector.bin'), map_location='cpu')
            mm_projector_weights = {k: v.to(torch.bfloat16) for k, v in mm_projector_weights.items()}
            model.load_state_dict(mm_projector_weights, strict=False)
        else:
            tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
            model = LibraLlamaForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
    else:
        # Load language model
        if model_base is not None:
            # PEFT model
            from peft import PeftModel
            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
            model = AutoModelForCausalLM.from_pretrained(model_base, quantization_config=quantization_config, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, device_map="auto")
            print(f"Loading LoRA weights from {model_path}")
            model = PeftModel.from_pretrained(model, model_path)
            print(f"Merging weights")
            model = model.merge_and_unload()
            print('Convert to FP16...')
            model.to(torch.bfloat16)
        else:
            use_fast = False
            tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
            model = AutoModelForCausalLM.from_pretrained(model_path, quantization_config=quantization_config, low_cpu_mem_usage=True, **kwargs)

    image_processor = None

    if 'libra' in model_name.lower():
        mm_use_im_start_end = getattr(model.config, "mm_use_im_start_end", False)
        mm_use_im_patch_token = getattr(model.config, "mm_use_im_patch_token", True)
        if mm_use_im_patch_token:
            tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
        if mm_use_im_start_end:
            tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
        model.resize_token_embeddings(len(tokenizer))

        vision_tower = model.get_vision_tower()
        if not vision_tower.is_loaded:
            vision_tower.load_model()
        vision_tower.to(device=device, dtype=torch.bfloat16)
        image_processor = vision_tower.image_processor

    if hasattr(model.config, "max_sequence_length"):
        context_len = model.config.max_sequence_length
    else:
        context_len = 2048

    return tokenizer, model, image_processor, context_len