Spaces:
Runtime error
Runtime error
zetavg
commited on
Commit
•
9c78439
1
Parent(s):
03b5741
try to load different type of models
Browse files- llama_lora/models.py +60 -16
llama_lora/models.py
CHANGED
@@ -5,7 +5,10 @@ import json
|
|
5 |
import re
|
6 |
|
7 |
import torch
|
8 |
-
from transformers import
|
|
|
|
|
|
|
9 |
from peft import PeftModel
|
10 |
|
11 |
from .globals import Global
|
@@ -27,42 +30,83 @@ def get_new_base_model(base_model_name):
|
|
27 |
Global.name_of_new_base_model_that_is_ready_to_be_used = None
|
28 |
clear_cache()
|
29 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
device = get_device()
|
31 |
|
32 |
if device == "cuda":
|
33 |
-
|
34 |
-
|
35 |
load_in_8bit=Global.load_8bit,
|
36 |
torch_dtype=torch.float16,
|
37 |
# device_map="auto",
|
38 |
# ? https://github.com/tloen/alpaca-lora/issues/21
|
39 |
device_map={'': 0},
|
|
|
|
|
40 |
trust_remote_code=Global.trust_remote_code
|
41 |
)
|
42 |
elif device == "mps":
|
43 |
-
|
44 |
-
|
45 |
device_map={"": device},
|
46 |
torch_dtype=torch.float16,
|
|
|
|
|
47 |
trust_remote_code=Global.trust_remote_code
|
48 |
)
|
49 |
else:
|
50 |
-
|
51 |
-
|
52 |
device_map={"": device},
|
53 |
low_cpu_mem_usage=True,
|
|
|
|
|
54 |
trust_remote_code=Global.trust_remote_code
|
55 |
)
|
56 |
|
57 |
-
tokenizer = get_tokenizer(base_model_name)
|
58 |
-
|
59 |
-
if re.match("[^/]+/llama", base_model_name):
|
60 |
-
model.config.pad_token_id = tokenizer.pad_token_id = 0
|
61 |
-
model.config.bos_token_id = tokenizer.bos_token_id = 1
|
62 |
-
model.config.eos_token_id = tokenizer.eos_token_id = 2
|
63 |
-
|
64 |
-
return model
|
65 |
-
|
66 |
|
67 |
def get_tokenizer(base_model_name):
|
68 |
if Global.ui_dev_mode:
|
|
|
5 |
import re
|
6 |
|
7 |
import torch
|
8 |
+
from transformers import (
|
9 |
+
AutoModelForCausalLM, AutoModel,
|
10 |
+
AutoTokenizer, LlamaTokenizer
|
11 |
+
)
|
12 |
from peft import PeftModel
|
13 |
|
14 |
from .globals import Global
|
|
|
30 |
Global.name_of_new_base_model_that_is_ready_to_be_used = None
|
31 |
clear_cache()
|
32 |
|
33 |
+
model_class = AutoModelForCausalLM
|
34 |
+
from_tf = False
|
35 |
+
force_download = False
|
36 |
+
has_tried_force_download = False
|
37 |
+
while True:
|
38 |
+
try:
|
39 |
+
model = _get_model_from_pretrained(
|
40 |
+
model_class, base_model_name, from_tf=from_tf, force_download=force_download)
|
41 |
+
break
|
42 |
+
except Exception as e:
|
43 |
+
if 'from_tf' in str(e):
|
44 |
+
print(
|
45 |
+
f"Got error while loading model {base_model_name} with AutoModelForCausalLM: {e}.")
|
46 |
+
print("Retrying with from_tf=True...")
|
47 |
+
from_tf = True
|
48 |
+
force_download = False
|
49 |
+
elif model_class == AutoModelForCausalLM:
|
50 |
+
print(
|
51 |
+
f"Got error while loading model {base_model_name} with AutoModelForCausalLM: {e}.")
|
52 |
+
print("Retrying with AutoModel...")
|
53 |
+
model_class = AutoModel
|
54 |
+
force_download = False
|
55 |
+
else:
|
56 |
+
if has_tried_force_download:
|
57 |
+
raise e
|
58 |
+
print(
|
59 |
+
f"Got error while loading model {base_model_name}: {e}.")
|
60 |
+
print("Retrying with force_download=True...")
|
61 |
+
model_class = AutoModelForCausalLM
|
62 |
+
from_tf = False
|
63 |
+
force_download = True
|
64 |
+
has_tried_force_download = True
|
65 |
+
|
66 |
+
tokenizer = get_tokenizer(base_model_name)
|
67 |
+
|
68 |
+
if re.match("[^/]+/llama", base_model_name):
|
69 |
+
model.config.pad_token_id = tokenizer.pad_token_id = 0
|
70 |
+
model.config.bos_token_id = tokenizer.bos_token_id = 1
|
71 |
+
model.config.eos_token_id = tokenizer.eos_token_id = 2
|
72 |
+
|
73 |
+
return model
|
74 |
+
|
75 |
+
|
76 |
+
def _get_model_from_pretrained(model_class, model_name, from_tf=False, force_download=False):
|
77 |
device = get_device()
|
78 |
|
79 |
if device == "cuda":
|
80 |
+
return model_class.from_pretrained(
|
81 |
+
model_name,
|
82 |
load_in_8bit=Global.load_8bit,
|
83 |
torch_dtype=torch.float16,
|
84 |
# device_map="auto",
|
85 |
# ? https://github.com/tloen/alpaca-lora/issues/21
|
86 |
device_map={'': 0},
|
87 |
+
from_tf=from_tf,
|
88 |
+
force_download=force_download,
|
89 |
trust_remote_code=Global.trust_remote_code
|
90 |
)
|
91 |
elif device == "mps":
|
92 |
+
return model_class.from_pretrained(
|
93 |
+
model_name,
|
94 |
device_map={"": device},
|
95 |
torch_dtype=torch.float16,
|
96 |
+
from_tf=from_tf,
|
97 |
+
force_download=force_download,
|
98 |
trust_remote_code=Global.trust_remote_code
|
99 |
)
|
100 |
else:
|
101 |
+
return model_class.from_pretrained(
|
102 |
+
model_name,
|
103 |
device_map={"": device},
|
104 |
low_cpu_mem_usage=True,
|
105 |
+
from_tf=from_tf,
|
106 |
+
force_download=force_download,
|
107 |
trust_remote_code=Global.trust_remote_code
|
108 |
)
|
109 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
110 |
|
111 |
def get_tokenizer(base_model_name):
|
112 |
if Global.ui_dev_mode:
|