-
Notifications
You must be signed in to change notification settings - Fork 0
/
train_nanollama.py
176 lines (150 loc) · 4.67 KB
/
train_nanollama.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
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
import os
os.environ["TOKENIZERS_PARALLELISM"] = "true"
import torch
import torch.nn as nn
import torch.optim.lr_scheduler as lr_sch
import torch.utils.data as Data
import wandb
import pytorch_lightning as pl
from pytorch_lightning.callbacks import LearningRateMonitor, ModelCheckpoint
from pytorch_lightning.loggers import WandbLogger
from transformers import LlamaForCausalLM, LlamaTokenizer, LlamaConfig
from hakuphi.trainer import CausalLMTrainer
from data import dan_prompt, titpop
## Setup
torch.set_default_dtype(torch.bfloat16)
torch.set_float32_matmul_precision("medium")
wandb.require("core")
## Constant
EPOCH = 1
GPUS = 4
BATCH_SIZE = 32
GRAD_ACC = 8
CUT_OFF = 768
LR = 5e-5
def load_tokenizer(
tokenizer_ref="TinyLlama/TinyLlama-1.1B-intermediate-step-480k-1T",
):
tokenizer = LlamaTokenizer.from_pretrained(tokenizer_ref)
titpop.apply_special_tokens(tokenizer)
tokenizer.pad_token = tokenizer.eos_token
return tokenizer
def load_model(
config: LlamaConfig,
tokenizer: LlamaTokenizer,
) -> LlamaForCausalLM:
config.pad_token_id = tokenizer.eos_token_id
config.vocab_size = tokenizer.get_vocab().__len__()
model = LlamaForCausalLM(config)
return model
def load_trainer(
model: LlamaForCausalLM, lycoris_model: nn.Module = None, t_max=1000_000
) -> CausalLMTrainer:
return CausalLMTrainer(
model,
lycoris_model,
name="TITPOP-200M",
lr=LR,
optimizer=torch.optim.AdamW,
opt_configs={
"weight_decay": 0.01,
"betas": (0.9, 0.98),
},
lr_scheduler=lr_sch.CosineAnnealingLR,
lr_sch_configs={
"T_max": t_max,
"eta_min": 1e-2 * LR,
},
use_warm_up=True,
warm_up_period=100,
)
tokenizer: LlamaTokenizer = load_tokenizer()
processor = titpop.processor(
tokenizer, cutoff_len=CUT_OFF, train_on_inputs=False, padding=True
)
def collate(batch):
batch = [processor(data) for data in batch]
attn_mask = torch.stack([x["attention_mask"] for x in batch])
labels = torch.stack([x["labels"] for x in batch])
result = {
"input_ids": torch.stack([x["input_ids"] for x in batch]),
"attention_mask": attn_mask,
"labels": labels,
"token_count": torch.sum(attn_mask).cpu().item(),
"trained_token_count": torch.sum(labels != -100).cpu().item(),
}
return result
def main():
# Setup dataset
dataset = titpop.load("danbooru")
# dataset2 = titpop.load("gbc")
# dataset = Data.ConcatDataset([dataset1, dataset2])
data_loader = Data.DataLoader(
dataset,
shuffle=True,
batch_size=BATCH_SIZE,
collate_fn=collate,
num_workers=16,
pin_memory=True,
drop_last=True,
persistent_workers=True,
)
# config = LlamaConfig(
# vocab_size=32006,
# hidden_size=768,
# intermediate_size=2304,
# num_hidden_layers=20,
# num_attention_heads=768 // 64,
# hidden_act="silu",
# max_position_embeddings=2048,
# rms_norm_eps=1e-5,
# use_cache=False,
# attn_implementation="flash_attention_2",
# torch_dtype=torch.bfloat16,
# )
# text_model = load_model(config, tokenizer)
# text_model.init_weights()
text_model = LlamaForCausalLM.from_pretrained("./TITPOP-200M-5ep")
print(sum(param.shape.numel() for param in text_model.parameters()))
text_model.gradient_checkpointing_enable()
text_model.to(torch.float)
text_model_eager = text_model
# text_model_eager = torch.compile(text_model, backend="eager")
trainer_module = load_trainer(
text_model_eager,
None,
len(dataset) * EPOCH // (BATCH_SIZE * GPUS * GRAD_ACC),
)
print(f"Total training step: {len(dataset)*EPOCH//(BATCH_SIZE*GPUS*GRAD_ACC)}")
# Train!
logger = None
logger = WandbLogger(
name="TITPOP-200M",
project="NanoLLaMA",
# offline=True,
)
trainer = pl.Trainer(
precision="bf16-mixed",
accelerator="gpu",
devices=GPUS,
max_epochs=EPOCH,
logger=logger,
log_every_n_steps=10,
accumulate_grad_batches=GRAD_ACC,
callbacks=[
LearningRateMonitor(logging_interval="step"),
ModelCheckpoint(every_n_train_steps=1000),
ModelCheckpoint(every_n_epochs=1),
],
gradient_clip_val=1.0,
# fast_dev_run=True,
)
trainer.fit(
trainer_module,
train_dataloaders=data_loader,
)
text_model.save_pretrained("TITPOP-200M-5ep-ft")
tokenizer.save_pretrained("TITPOP-200M-5ep-ft")
if __name__ == "__main__":
pl.seed_everything(3408)
main()