forked from Cornell-RelaxML/quip-sharp
-
Notifications
You must be signed in to change notification settings - Fork 0
/
hfize_llama.py
137 lines (109 loc) · 5.51 KB
/
hfize_llama.py
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
125
126
127
128
129
130
131
132
133
134
135
136
137
import argparse
import os
import glog
import torch
from transformers import AutoTokenizer
from model.version import MODEL_VERSION
from model.llama import LlamaForCausalLM as llama_fuse
from model.mistral import MistralForCausalLM
from lib import codebook
from lib.utils.unsafe_import import model_from_hf_path
import time
torch.set_grad_enabled(False)
parser = argparse.ArgumentParser()
parser.add_argument('--quantized_path', type=str)
parser.add_argument('--hf_output_path', type=str)
def unpack_quip(module, saved_layer, codebook_id, codesz):
(m, n) = saved_layer['Qidxs'].shape
if codebook_id in codebook.cache_permute_set:
module.Qidxs.copy_(saved_layer['Qidxs'].view(m, n // codesz,
codesz).permute(1, 0,
2).reshape(m, n).contiguous())
else:
module.Qidxs.copy_(saved_layer['Qidxs'])
if module.rank > 0:
module.A.copy_(saved_layer['A'])
module.B.copy_(saved_layer['B'])
module.SU.copy_(saved_layer['SU'])
module.SV.copy_(saved_layer['SV'])
if module.rescale_WH:
module.scaleWH.copy_(saved_layer['scaleWH'])
module.codebook_id.copy_(codebook_id)
def main(args):
assert os.path.exists(args.quantized_path)
saved_config = torch.load(os.path.join(args.quantized_path, 'config.pt'))
model_config = saved_config['model_config']
codebook_id = codebook.get_id(model_config.quip_params['codebook'])
codesz = model_config.quip_params['codesz']
tokenizer = AutoTokenizer.from_pretrained(model_config._name_or_path)
model_type = model_config.model_type
model_config.quip_params['model_version'] = MODEL_VERSION
if model_type == 'llama':
model_cls = llama_fuse
elif model_type == 'mistral':
model_cls = MistralForCausalLM
else:
raise Exception
model = model_cls.from_pretrained(model_config._name_or_path,
torch_dtype='auto',
low_cpu_mem_usage=True,
config=model_config).half()
for ii in range(len(model.model.layers)):
glog.info(f'updating layer {ii}')
layer = model.model.layers[ii]
cpu = torch.device('cpu')
glog.info(f'loading layer {ii} qkv')
saved_layer = torch.load(f'{args.quantized_path}/{ii}_qkv.pt', map_location=cpu)
layer.self_attn.qkv_proj.fuse_scales[0].copy_(saved_layer['W_q_scale'])
layer.self_attn.qkv_proj.fuse_scales[1].copy_(saved_layer['W_k_scale'])
layer.self_attn.qkv_proj.fuse_scales[2].copy_(saved_layer['W_v_scale'])
W_qbias_scale = saved_layer['W_qbias_scale']
W_kbias_scale = saved_layer['W_kbias_scale']
W_vbias_scale = saved_layer['W_vbias_scale']
#print(layer.self_attn.qkv_proj.weight.size())
# 沿着指定的维度（通常是最后一个维度，dim=-1）拼接这三个张量
concatenated_tensor =torch.cat([W_qbias_scale, W_kbias_scale, W_vbias_scale],dim=-1)
print("2222",concatenated_tensor.size())
layer.self_attn.qkv_proj.bias.copy_(concatenated_tensor)
layer.self_attn.qkv_proj.Wscale.copy_(saved_layer['Wscale'])
unpack_quip(layer.self_attn.qkv_proj, saved_layer, codebook_id, codesz)
glog.info(f'loading layer {ii} up')
saved_layer = torch.load(f'{args.quantized_path}/{ii}_up.pt', map_location=cpu)
layer.mlp.upgate_proj.fuse_scales[0].copy_(saved_layer['W_up_scale'])
layer.mlp.upgate_proj.fuse_scales[1].copy_(saved_layer['W_gate_scale'])
layer.mlp.upgate_proj.Wscale.copy_(saved_layer['Wscale'])
unpack_quip(layer.mlp.upgate_proj, saved_layer, codebook_id, codesz)
glog.info(f'loading layer {ii} o')
saved_layer = torch.load(f'{args.quantized_path}/{ii}_o.pt', map_location=cpu)
W_obias_scale = saved_layer['W_obias_scale']
layer.self_attn.o_proj.bias.copy_(W_obias_scale)
layer.self_attn.o_proj.Wscale.copy_(saved_layer['W_o_scale'] * saved_layer['Wscale'])
unpack_quip(layer.self_attn.o_proj, saved_layer, codebook_id, codesz)
glog.info(f'loading layer {ii} down')
saved_layer = torch.load(f'{args.quantized_path}/{ii}_down.pt', map_location=cpu)
layer.mlp.down_proj.Wscale.copy_(saved_layer['W_down_scale'] * saved_layer['Wscale'])
if model_config.quip_params['outlier_channel_split']:
layer.mlp.down_proj.ocs_dupe_inds.copy_(torch.tensor(saved_layer['ocs_dupe_inds']))
unpack_quip(layer.mlp.down_proj, saved_layer, codebook_id, codesz)
glog.info(f'saving model...')
model.save_pretrained(args.hf_output_path, safe_serialization=True)
del model
model, _ = model_from_hf_path(args.hf_output_path, use_cuda_graph=False)
glog.info('successfully loaded hfized model')
glog.info('generating some text...')
start = time.time()
prompt = 'It is a truth universally acknowledged that'
inputs = tokenizer(prompt, return_tensors='pt')
outputs = model.generate(input_ids=inputs['input_ids'].cuda(),
attention_mask=inputs['attention_mask'].cuda(),
max_new_tokens=64,
return_dict_in_generate=True)
token = outputs.sequences[0, :]
output_str = tokenizer.decode(token)
glog.info(output_str)
glog.info(f'elapsed: {time.time() - start}')
if __name__ == '__main__':
torch.set_grad_enabled(False)
torch.manual_seed(0)
args = parser.parse_args()
main(args)