make it work with pythia in the cloud
Browse files- .gitattributes +1 -0
- configs/pythia_1_2B_alpaca.yml +13 -12
- scripts/finetune.py +116 -12
- src/axolotl/convert.py +1 -0
- src/axolotl/datasets.py +56 -38
- src/axolotl/prompt_tokenizers.py +11 -4
- src/axolotl/prompters.py +154 -4
.gitattributes
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
data/*.jsonl filter=lfs diff=lfs merge=lfs -text
|
configs/pythia_1_2B_alpaca.yml
CHANGED
@@ -3,35 +3,36 @@ model_type: GPTNeoXForCausalLM
|
|
3 |
tokenizer_type: AutoTokenizer
|
4 |
load_in_8bit: true
|
5 |
datasets:
|
6 |
-
- path:
|
7 |
type: alpaca
|
8 |
-
- path:
|
9 |
type: sharegpt
|
10 |
-
- path:
|
11 |
type: gpteacher
|
12 |
-
- path:
|
13 |
type: gpteacher
|
14 |
val_set_size: 0.05
|
15 |
adapter: lora
|
16 |
sequence_len: 2048
|
17 |
-
lora_r:
|
18 |
lora_alpha: 32
|
19 |
lora_dropout: 0.05
|
20 |
lora_target_modules:
|
21 |
-
-
|
22 |
-
|
23 |
-
wandb_project:
|
24 |
wandb_watch:
|
25 |
-
|
26 |
wandb_log_model: checkpoint
|
27 |
output_dir: ./lora-alpaca
|
28 |
-
batch_size:
|
29 |
-
micro_batch_size:
|
30 |
num_epochs: 5
|
31 |
learning_rate: 0.0003
|
32 |
train_on_inputs: false
|
|
|
33 |
bf16: True
|
34 |
-
|
35 |
resume_from_checkpoint:
|
36 |
local_rank:
|
37 |
deepspeed:
|
|
|
3 |
tokenizer_type: AutoTokenizer
|
4 |
load_in_8bit: true
|
5 |
datasets:
|
6 |
+
- path: data/alpaca_data_gpt4.jsonl
|
7 |
type: alpaca
|
8 |
+
- path: data/vicuna_cleaned.jsonl
|
9 |
type: sharegpt
|
10 |
+
- path: data/gpt4-instruct-similarity-0.6-dataset.jsonl
|
11 |
type: gpteacher
|
12 |
+
- path: data/roleplay-similarity_0.6-instruct-dataset.jsonl
|
13 |
type: gpteacher
|
14 |
val_set_size: 0.05
|
15 |
adapter: lora
|
16 |
sequence_len: 2048
|
17 |
+
lora_r: 8
|
18 |
lora_alpha: 32
|
19 |
lora_dropout: 0.05
|
20 |
lora_target_modules:
|
21 |
+
- query_key_value
|
22 |
+
lora_fan_in_fan_out: true # pythia/GPTNeoX lora specific
|
23 |
+
wandb_project: pythia-1.4b-lora
|
24 |
wandb_watch:
|
25 |
+
wandb_run_name:
|
26 |
wandb_log_model: checkpoint
|
27 |
output_dir: ./lora-alpaca
|
28 |
+
batch_size: 32
|
29 |
+
micro_batch_size: 4
|
30 |
num_epochs: 5
|
31 |
learning_rate: 0.0003
|
32 |
train_on_inputs: false
|
33 |
+
group_by_length: false
|
34 |
bf16: True
|
35 |
+
tf32: True
|
36 |
resume_from_checkpoint:
|
37 |
local_rank:
|
38 |
deepspeed:
|
scripts/finetune.py
CHANGED
@@ -1,26 +1,32 @@
|
|
|
|
1 |
import os
|
|
|
2 |
import sys
|
3 |
from pathlib import Path
|
4 |
|
|
|
5 |
import fire
|
6 |
import torch
|
7 |
import transformers
|
8 |
import yaml
|
9 |
from attrdict import AttrDict
|
10 |
-
from datasets import load_dataset, IterableDataset
|
11 |
from peft import (
|
12 |
LoraConfig,
|
13 |
get_peft_model,
|
14 |
-
prepare_model_for_int8_training,
|
15 |
)
|
|
|
16 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
17 |
|
18 |
# add src to the pythonpath so we don't need to pip install this
|
|
|
|
|
19 |
project_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
|
20 |
src_dir = os.path.join(project_root, 'src')
|
21 |
sys.path.insert(0, src_dir)
|
22 |
|
23 |
-
from axolotl.datasets import TokenizedPromptDataset
|
24 |
from axolotl.prompt_tokenizers import AlpacaPromptTokenizingStrategy, ShareGPTPromptTokenizingStrategy, \
|
25 |
LLAMA_DEFAULT_PAD_TOKEN, GPTeacherPromptTokenizingStrategy
|
26 |
from axolotl.prompters import AlpacaPrompter, GPTeacherPrompter, ShareGPTPrompter
|
@@ -29,9 +35,9 @@ def setup_wandb_env_vars(cfg):
|
|
29 |
if len(cfg.wandb_project) > 0:
|
30 |
os.environ["WANDB_PROJECT"] = cfg.wandb_project
|
31 |
cfg.use_wandb = True
|
32 |
-
if len(cfg.wandb_watch) > 0:
|
33 |
os.environ["WANDB_WATCH"] = cfg.wandb_watch
|
34 |
-
if len(cfg.wandb_log_model) > 0:
|
35 |
os.environ["WANDB_LOG_MODEL"] = cfg.wandb_log_model
|
36 |
|
37 |
|
@@ -61,6 +67,10 @@ def load_model(base_model, model_type, tokenizer_type, cfg, adapter="lora"):
|
|
61 |
if tokenizer.__class__.__name__ == "LlamaTokenizer":
|
62 |
tokenizer.pad_token = LLAMA_DEFAULT_PAD_TOKEN
|
63 |
|
|
|
|
|
|
|
|
|
64 |
if cfg.load_in_8bit:
|
65 |
model = prepare_model_for_int8_training(model)
|
66 |
|
@@ -69,6 +79,7 @@ def load_model(base_model, model_type, tokenizer_type, cfg, adapter="lora"):
|
|
69 |
lora_alpha=cfg.lora_alpha,
|
70 |
target_modules=cfg.lora_target_modules,
|
71 |
lora_dropout=cfg.lora_dropout,
|
|
|
72 |
bias="none",
|
73 |
task_type="CAUSAL_LM",
|
74 |
)
|
@@ -79,7 +90,7 @@ def load_model(base_model, model_type, tokenizer_type, cfg, adapter="lora"):
|
|
79 |
# TODO resume_from_checkpoint handling
|
80 |
|
81 |
model.print_trainable_parameters()
|
82 |
-
return model, tokenizer
|
83 |
|
84 |
|
85 |
def train(
|
@@ -88,7 +99,7 @@ def train(
|
|
88 |
):
|
89 |
# load the config from the yaml file
|
90 |
with open(config, 'r') as f:
|
91 |
-
cfg: AttrDict = AttrDict(yaml.load(f))
|
92 |
# if there are any options passed in the cli, if it is something that seems valid from the yaml,
|
93 |
# then overwrite the value
|
94 |
for k, v in enumerate(kwargs):
|
@@ -107,23 +118,116 @@ def train(
|
|
107 |
setup_wandb_env_vars(cfg)
|
108 |
|
109 |
# Load the model and tokenizer
|
110 |
-
model, tokenizer = load_model(cfg.base_model, cfg.model_type, cfg.tokenizer_type, cfg, adapter=cfg.adapter)
|
111 |
datasets = []
|
112 |
for d in cfg.datasets:
|
113 |
-
ds: IterableDataset = load_dataset("json", data_files=d.path, streaming=True,
|
114 |
if d.type == "alpaca":
|
115 |
ds_strategy = AlpacaPromptTokenizingStrategy(AlpacaPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
116 |
-
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds)
|
117 |
datasets.append(ds_wrapper)
|
118 |
elif d.type == "gpteacher":
|
119 |
ds_strategy = GPTeacherPromptTokenizingStrategy(GPTeacherPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
120 |
-
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds)
|
121 |
datasets.append(ds_wrapper)
|
122 |
elif d.type == "sharegpt":
|
123 |
ds_strategy = ShareGPTPromptTokenizingStrategy(ShareGPTPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
124 |
-
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds)
|
125 |
datasets.append(ds_wrapper)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
|
|
|
127 |
|
128 |
if __name__ == "__main__":
|
129 |
fire.Fire(train)
|
|
|
1 |
+
import math
|
2 |
import os
|
3 |
+
import signal
|
4 |
import sys
|
5 |
from pathlib import Path
|
6 |
|
7 |
+
import bitsandbytes as bnb
|
8 |
import fire
|
9 |
import torch
|
10 |
import transformers
|
11 |
import yaml
|
12 |
from attrdict import AttrDict
|
13 |
+
from datasets import load_dataset, IterableDataset, Dataset
|
14 |
from peft import (
|
15 |
LoraConfig,
|
16 |
get_peft_model,
|
17 |
+
prepare_model_for_int8_training, get_peft_model_state_dict,
|
18 |
)
|
19 |
+
from torch import nn
|
20 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
21 |
|
22 |
# add src to the pythonpath so we don't need to pip install this
|
23 |
+
from transformers.trainer_pt_utils import get_parameter_names
|
24 |
+
|
25 |
project_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
|
26 |
src_dir = os.path.join(project_root, 'src')
|
27 |
sys.path.insert(0, src_dir)
|
28 |
|
29 |
+
from axolotl.datasets import TokenizedPromptDataset, ConstantLengthDataset
|
30 |
from axolotl.prompt_tokenizers import AlpacaPromptTokenizingStrategy, ShareGPTPromptTokenizingStrategy, \
|
31 |
LLAMA_DEFAULT_PAD_TOKEN, GPTeacherPromptTokenizingStrategy
|
32 |
from axolotl.prompters import AlpacaPrompter, GPTeacherPrompter, ShareGPTPrompter
|
|
|
35 |
if len(cfg.wandb_project) > 0:
|
36 |
os.environ["WANDB_PROJECT"] = cfg.wandb_project
|
37 |
cfg.use_wandb = True
|
38 |
+
if cfg.wandb_watch and len(cfg.wandb_watch) > 0:
|
39 |
os.environ["WANDB_WATCH"] = cfg.wandb_watch
|
40 |
+
if cfg.wandb_log_model and len(cfg.wandb_log_model) > 0:
|
41 |
os.environ["WANDB_LOG_MODEL"] = cfg.wandb_log_model
|
42 |
|
43 |
|
|
|
67 |
if tokenizer.__class__.__name__ == "LlamaTokenizer":
|
68 |
tokenizer.pad_token = LLAMA_DEFAULT_PAD_TOKEN
|
69 |
|
70 |
+
if tokenizer.__class__.__name__ == "GPTNeoXTokenizerFast":
|
71 |
+
tokenizer.add_special_tokens({'pad_token': '[PAD]'})
|
72 |
+
os.environ["TOKENIZERS_PARALLELISM"] = "false"
|
73 |
+
|
74 |
if cfg.load_in_8bit:
|
75 |
model = prepare_model_for_int8_training(model)
|
76 |
|
|
|
79 |
lora_alpha=cfg.lora_alpha,
|
80 |
target_modules=cfg.lora_target_modules,
|
81 |
lora_dropout=cfg.lora_dropout,
|
82 |
+
fan_in_fan_out=cfg.lora_fan_in_fan_out,
|
83 |
bias="none",
|
84 |
task_type="CAUSAL_LM",
|
85 |
)
|
|
|
90 |
# TODO resume_from_checkpoint handling
|
91 |
|
92 |
model.print_trainable_parameters()
|
93 |
+
return model, tokenizer, lora_config
|
94 |
|
95 |
|
96 |
def train(
|
|
|
99 |
):
|
100 |
# load the config from the yaml file
|
101 |
with open(config, 'r') as f:
|
102 |
+
cfg: AttrDict = AttrDict(yaml.load(f, Loader=yaml.Loader))
|
103 |
# if there are any options passed in the cli, if it is something that seems valid from the yaml,
|
104 |
# then overwrite the value
|
105 |
for k, v in enumerate(kwargs):
|
|
|
118 |
setup_wandb_env_vars(cfg)
|
119 |
|
120 |
# Load the model and tokenizer
|
121 |
+
model, tokenizer, lora_config = load_model(cfg.base_model, cfg.model_type, cfg.tokenizer_type, cfg, adapter=cfg.adapter)
|
122 |
datasets = []
|
123 |
for d in cfg.datasets:
|
124 |
+
ds: IterableDataset = load_dataset("json", data_files=d.path, streaming=True, split=None)
|
125 |
if d.type == "alpaca":
|
126 |
ds_strategy = AlpacaPromptTokenizingStrategy(AlpacaPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
127 |
+
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
128 |
datasets.append(ds_wrapper)
|
129 |
elif d.type == "gpteacher":
|
130 |
ds_strategy = GPTeacherPromptTokenizingStrategy(GPTeacherPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
131 |
+
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
132 |
datasets.append(ds_wrapper)
|
133 |
elif d.type == "sharegpt":
|
134 |
ds_strategy = ShareGPTPromptTokenizingStrategy(ShareGPTPrompter(), tokenizer, cfg.train_on_inputs, cfg.sequence_len)
|
135 |
+
ds_wrapper = TokenizedPromptDataset(ds_strategy, ds["train"])
|
136 |
datasets.append(ds_wrapper)
|
137 |
+
constant_len_dataset = ConstantLengthDataset(tokenizer, datasets, seq_length=cfg.sequence_len)
|
138 |
+
constant_len_dataset = Dataset.from_list([_ for _ in constant_len_dataset]).train_test_split(
|
139 |
+
test_size=cfg.val_set_size, shuffle=True, seed=42
|
140 |
+
)
|
141 |
+
|
142 |
+
print(constant_len_dataset)
|
143 |
+
train_dataset = constant_len_dataset["train"]
|
144 |
+
eval_dataset = constant_len_dataset["test"]
|
145 |
+
|
146 |
+
total_num_steps = int(math.ceil(len(train_dataset) * cfg.num_epochs / cfg.batch_size))
|
147 |
+
warmup_steps = min(int(0.03 * total_num_steps), 100)
|
148 |
+
logging_steps = min(int(0.005 * total_num_steps), 10)
|
149 |
+
save_steps = eval_steps = min(int(0.05 * total_num_steps), 200)
|
150 |
+
|
151 |
+
training_args = transformers.TrainingArguments(
|
152 |
+
per_device_train_batch_size=cfg.micro_batch_size,
|
153 |
+
gradient_accumulation_steps=cfg.gradient_accumulation_steps,
|
154 |
+
warmup_steps=warmup_steps,
|
155 |
+
num_train_epochs=cfg.num_epochs,
|
156 |
+
learning_rate=cfg.learning_rate,
|
157 |
+
bf16=cfg.bf16,
|
158 |
+
tf32=cfg.tf32,
|
159 |
+
logging_steps=logging_steps,
|
160 |
+
evaluation_strategy="steps" if cfg.val_set_size > 0 else "no",
|
161 |
+
save_strategy="steps",
|
162 |
+
eval_steps=eval_steps if cfg.val_set_size > 0 else None,
|
163 |
+
save_steps=save_steps,
|
164 |
+
output_dir=cfg.output_dir,
|
165 |
+
save_total_limit=3,
|
166 |
+
load_best_model_at_end=True if cfg.val_set_size > 0 else False,
|
167 |
+
ddp_find_unused_parameters=False if cfg.ddp else None,
|
168 |
+
group_by_length=cfg.group_by_length,
|
169 |
+
report_to="wandb" if cfg.use_wandb else None,
|
170 |
+
run_name=cfg.wandb_run_name if cfg.use_wandb else None,
|
171 |
+
)
|
172 |
+
|
173 |
+
decay_parameters = get_parameter_names(model, [nn.LayerNorm])
|
174 |
+
decay_parameters = [name for name in decay_parameters if "bias" not in name]
|
175 |
+
optimizer_grouped_parameters = [
|
176 |
+
{
|
177 |
+
"params": [p for n, p in model.named_parameters() if n in decay_parameters],
|
178 |
+
"weight_decay": training_args.weight_decay,
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"params": [p for n, p in model.named_parameters() if n not in decay_parameters],
|
182 |
+
"weight_decay": 0.0,
|
183 |
+
},
|
184 |
+
]
|
185 |
+
|
186 |
+
adam_bnb_optim = bnb.optim.Adam8bit(
|
187 |
+
optimizer_grouped_parameters,
|
188 |
+
betas=(training_args.adam_beta1, training_args.adam_beta2),
|
189 |
+
eps=training_args.adam_epsilon,
|
190 |
+
lr=training_args.learning_rate,
|
191 |
+
)
|
192 |
+
|
193 |
+
lr_scheduler = transformers.get_cosine_schedule_with_warmup(
|
194 |
+
adam_bnb_optim,
|
195 |
+
training_args.warmup_steps,
|
196 |
+
total_num_steps,
|
197 |
+
)
|
198 |
+
|
199 |
+
trainer = transformers.Trainer(
|
200 |
+
model=model,
|
201 |
+
train_dataset=train_dataset,
|
202 |
+
eval_dataset=eval_dataset,
|
203 |
+
args=training_args,
|
204 |
+
optimizers=(adam_bnb_optim, lr_scheduler),
|
205 |
+
data_collator=transformers.DataCollatorForSeq2Seq(
|
206 |
+
tokenizer, pad_to_multiple_of=8, return_tensors="pt", padding=True
|
207 |
+
),
|
208 |
+
)
|
209 |
+
model.config.use_cache = False
|
210 |
+
|
211 |
+
old_state_dict = model.state_dict
|
212 |
+
model.state_dict = (
|
213 |
+
lambda self, *_, **__: get_peft_model_state_dict(
|
214 |
+
self, old_state_dict()
|
215 |
+
)
|
216 |
+
).__get__(model, type(model))
|
217 |
+
|
218 |
+
if torch.__version__ >= "2" and sys.platform != "win32":
|
219 |
+
model = torch.compile(model)
|
220 |
+
|
221 |
+
signal.signal(signal.SIGINT, lambda signal, frame: (
|
222 |
+
model.save_pretrained(cfg.output_dir),
|
223 |
+
exit(0)
|
224 |
+
))
|
225 |
+
|
226 |
+
# go ahead and presave the adapter config
|
227 |
+
lora_config.save_pretrained(cfg.output_dir)
|
228 |
+
trainer.train(resume_from_checkpoint=cfg.resume_from_checkpoint)
|
229 |
|
230 |
+
model.save_pretrained(cfg.output_dir)
|
231 |
|
232 |
if __name__ == "__main__":
|
233 |
fire.Fire(train)
|
src/axolotl/convert.py
CHANGED
@@ -44,6 +44,7 @@ class JsonToJsonlConverter:
|
|
44 |
def convert(self, input_file_path, output_file_path):
|
45 |
content = self.file_reader.read(input_file_path)
|
46 |
data = self.json_parser.parse(content)
|
|
|
47 |
jsonl_content = self.jsonl_serializer.serialize(data)
|
48 |
self.file_writer.write(jsonl_content)
|
49 |
|
|
|
44 |
def convert(self, input_file_path, output_file_path):
|
45 |
content = self.file_reader.read(input_file_path)
|
46 |
data = self.json_parser.parse(content)
|
47 |
+
# data = [r for r in data if r["conversations"]] # vicuna cleaned has rows with empty conversations
|
48 |
jsonl_content = self.jsonl_serializer.serialize(data)
|
49 |
self.file_writer.write(jsonl_content)
|
50 |
|
src/axolotl/datasets.py
CHANGED
@@ -2,7 +2,7 @@ from typing import List
|
|
2 |
|
3 |
import torch
|
4 |
from datasets import IterableDataset
|
5 |
-
from .prompt_tokenizers import PromptTokenizingStrategy
|
6 |
|
7 |
|
8 |
# We want this to be a wrapper for an existing dataset that we have loaded
|
@@ -23,7 +23,12 @@ class TokenizedPromptDataset(IterableDataset):
|
|
23 |
|
24 |
def __iter__(self):
|
25 |
iterator = iter(self.dataset)
|
26 |
-
|
|
|
|
|
|
|
|
|
|
|
27 |
|
28 |
|
29 |
class ConstantLengthDataset(IterableDataset):
|
@@ -32,55 +37,68 @@ class ConstantLengthDataset(IterableDataset):
|
|
32 |
Args:
|
33 |
tokenizer (Tokenizer): The processor used for proccessing the data.
|
34 |
dataset (dataset.Dataset): Dataset with text files.
|
35 |
-
infinite (bool): If True the iterator is reset after dataset reaches end else stops.
|
36 |
seq_length (int): Length of token sequences to return.
|
37 |
-
chars_per_token (int): Number of characters per token used to estimate number of tokens in text buffer.
|
38 |
"""
|
39 |
|
40 |
def __init__(
|
41 |
self,
|
42 |
tokenizer,
|
43 |
datasets,
|
44 |
-
infinite=False,
|
45 |
seq_length=2048,
|
46 |
-
num_of_sequences=1024,
|
47 |
-
chars_per_token=3.6,
|
48 |
):
|
49 |
self.tokenizer = tokenizer
|
50 |
-
self.concat_token_id = tokenizer.eos_token_id
|
51 |
self.datasets: List[IterableDataset] = datasets
|
52 |
self.seq_length = seq_length
|
53 |
-
self.infinite = infinite
|
54 |
-
self.current_size = 0
|
55 |
-
self.max_buffer_size = seq_length * chars_per_token * num_of_sequences
|
56 |
|
57 |
def __iter__(self):
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
break
|
65 |
try:
|
66 |
-
|
67 |
-
buffer_len += len(buffer[-1])
|
68 |
except StopIteration:
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
"
|
84 |
-
|
85 |
-
|
86 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
|
3 |
import torch
|
4 |
from datasets import IterableDataset
|
5 |
+
from .prompt_tokenizers import PromptTokenizingStrategy, InvalidDataException
|
6 |
|
7 |
|
8 |
# We want this to be a wrapper for an existing dataset that we have loaded
|
|
|
23 |
|
24 |
def __iter__(self):
|
25 |
iterator = iter(self.dataset)
|
26 |
+
# Loop through the entire dataset
|
27 |
+
for example in iterator:
|
28 |
+
try:
|
29 |
+
yield self.prompt_tokenizer.tokenize_prompt(example)
|
30 |
+
except InvalidDataException:
|
31 |
+
pass
|
32 |
|
33 |
|
34 |
class ConstantLengthDataset(IterableDataset):
|
|
|
37 |
Args:
|
38 |
tokenizer (Tokenizer): The processor used for proccessing the data.
|
39 |
dataset (dataset.Dataset): Dataset with text files.
|
|
|
40 |
seq_length (int): Length of token sequences to return.
|
|
|
41 |
"""
|
42 |
|
43 |
def __init__(
|
44 |
self,
|
45 |
tokenizer,
|
46 |
datasets,
|
|
|
47 |
seq_length=2048,
|
|
|
|
|
48 |
):
|
49 |
self.tokenizer = tokenizer
|
50 |
+
self.concat_token_id = tokenizer.eos_token_id
|
51 |
self.datasets: List[IterableDataset] = datasets
|
52 |
self.seq_length = seq_length
|
|
|
|
|
|
|
53 |
|
54 |
def __iter__(self):
|
55 |
+
buffer = {"input_ids": [], "attention_mask": [], "labels": []}
|
56 |
+
buffer_len = 0
|
57 |
+
for dataset in self.datasets:
|
58 |
+
iterator = iter(dataset)
|
59 |
+
more_examples = True
|
60 |
+
while more_examples:
|
|
|
61 |
try:
|
62 |
+
example = next(iterator)
|
|
|
63 |
except StopIteration:
|
64 |
+
more_examples = False
|
65 |
+
example = None
|
66 |
+
|
67 |
+
add_concat_token = False
|
68 |
+
if example:
|
69 |
+
example_len = len(example["input_ids"])
|
70 |
+
add_concat_token = example["input_ids"][-1] != self.concat_token_id
|
71 |
+
else:
|
72 |
+
example_len = 0
|
73 |
+
|
74 |
+
if not example_len or buffer_len + int(add_concat_token) + example_len > self.seq_length:
|
75 |
+
if buffer["input_ids"]:
|
76 |
+
input_ids = torch.cat(buffer["input_ids"], dim=-1)[: self.seq_length]
|
77 |
+
attention_mask = torch.cat(buffer["attention_mask"], dim=-1)[: self.seq_length]
|
78 |
+
labels = torch.cat(buffer["labels"], dim=-1)[: self.seq_length]
|
79 |
+
yield {
|
80 |
+
"input_ids": input_ids,
|
81 |
+
"labels": labels,
|
82 |
+
"attention_mask": attention_mask,
|
83 |
+
}
|
84 |
+
buffer = {"input_ids": [], "attention_mask": [], "labels": []}
|
85 |
+
buffer_len = 0
|
86 |
+
|
87 |
+
if example:
|
88 |
+
input_ids = example["input_ids"]
|
89 |
+
attention_mask = example["attention_mask"]
|
90 |
+
labels = example["labels"]
|
91 |
+
|
92 |
+
if add_concat_token:
|
93 |
+
input_ids.append(self.concat_token_id)
|
94 |
+
attention_mask.append(1)
|
95 |
+
labels.append(self.concat_token_id)
|
96 |
+
|
97 |
+
input_ids_with_concat = torch.tensor(input_ids, dtype=torch.long)
|
98 |
+
attention_mask_with_concat = torch.tensor(attention_mask, dtype=torch.long)
|
99 |
+
labels_with_concat = torch.tensor(labels, dtype=torch.long)
|
100 |
+
|
101 |
+
buffer["input_ids"].append(input_ids_with_concat)
|
102 |
+
buffer["attention_mask"].append(attention_mask_with_concat)
|
103 |
+
buffer["labels"].append(labels_with_concat)
|
104 |
+
buffer_len += len(input_ids)
|
src/axolotl/prompt_tokenizers.py
CHANGED
@@ -9,6 +9,10 @@ LLAMA_DEFAULT_BOS_TOKEN = "<s>"
|
|
9 |
LLAMA_DEFAULT_UNK_TOKEN = "<unk>"
|
10 |
|
11 |
|
|
|
|
|
|
|
|
|
12 |
class PromptTokenizingStrategy(abc.ABC):
|
13 |
def __init__(
|
14 |
self,
|
@@ -32,7 +36,7 @@ class AlpacaPromptTokenizingStrategy(PromptTokenizingStrategy):
|
|
32 |
full_prompt = self._tokenize_full_prompt(prompt)
|
33 |
tokenized_full_prompt = self._tokenize(full_prompt)
|
34 |
if not self.train_on_inputs:
|
35 |
-
user_prompt = self.prompter.
|
36 |
prompt["instruction"], prompt["input"]
|
37 |
)
|
38 |
tokenized_user_prompt = self._tokenize(user_prompt, add_eos_token=False)
|
@@ -43,7 +47,7 @@ class AlpacaPromptTokenizingStrategy(PromptTokenizingStrategy):
|
|
43 |
return tokenized_full_prompt
|
44 |
|
45 |
def _tokenize_full_prompt(self, prompt):
|
46 |
-
return self.prompter.
|
47 |
prompt["instruction"],
|
48 |
prompt["input"],
|
49 |
prompt["output"],
|
@@ -71,7 +75,7 @@ class AlpacaPromptTokenizingStrategy(PromptTokenizingStrategy):
|
|
71 |
|
72 |
class GPTeacherPromptTokenizingStrategy(AlpacaPromptTokenizingStrategy):
|
73 |
def _tokenize_full_prompt(self, prompt):
|
74 |
-
return self.prompter.
|
75 |
prompt["instruction"],
|
76 |
prompt["input"],
|
77 |
prompt["response"],
|
@@ -80,4 +84,7 @@ class GPTeacherPromptTokenizingStrategy(AlpacaPromptTokenizingStrategy):
|
|
80 |
|
81 |
class ShareGPTPromptTokenizingStrategy(PromptTokenizingStrategy):
|
82 |
def tokenize_prompt(self, prompt):
|
83 |
-
|
|
|
|
|
|
|
|
9 |
LLAMA_DEFAULT_UNK_TOKEN = "<unk>"
|
10 |
|
11 |
|
12 |
+
class InvalidDataException(Exception):
|
13 |
+
pass
|
14 |
+
|
15 |
+
|
16 |
class PromptTokenizingStrategy(abc.ABC):
|
17 |
def __init__(
|
18 |
self,
|
|
|
36 |
full_prompt = self._tokenize_full_prompt(prompt)
|
37 |
tokenized_full_prompt = self._tokenize(full_prompt)
|
38 |
if not self.train_on_inputs:
|
39 |
+
user_prompt = self.prompter.build_prompt(
|
40 |
prompt["instruction"], prompt["input"]
|
41 |
)
|
42 |
tokenized_user_prompt = self._tokenize(user_prompt, add_eos_token=False)
|
|
|
47 |
return tokenized_full_prompt
|
48 |
|
49 |
def _tokenize_full_prompt(self, prompt):
|
50 |
+
return self.prompter.build_prompt(
|
51 |
prompt["instruction"],
|
52 |
prompt["input"],
|
53 |
prompt["output"],
|
|
|
75 |
|
76 |
class GPTeacherPromptTokenizingStrategy(AlpacaPromptTokenizingStrategy):
|
77 |
def _tokenize_full_prompt(self, prompt):
|
78 |
+
return self.prompter.build_prompt(
|
79 |
prompt["instruction"],
|
80 |
prompt["input"],
|
81 |
prompt["response"],
|
|
|
84 |
|
85 |
class ShareGPTPromptTokenizingStrategy(PromptTokenizingStrategy):
|
86 |
def tokenize_prompt(self, prompt):
|
87 |
+
try:
|
88 |
+
return self.prompter.build_prompt(prompt["conversations"], self.tokenizer)
|
89 |
+
except (KeyError, AssertionError) as e:
|
90 |
+
raise InvalidDataException(str(e))
|
src/axolotl/prompters.py
CHANGED
@@ -1,10 +1,160 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
class AlpacaPrompter:
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
|
5 |
class ShareGPTPrompter:
|
6 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
|
|
|
|
|
|
8 |
|
9 |
-
|
10 |
-
|
|
|
1 |
+
import copy
|
2 |
+
import dataclasses
|
3 |
+
from enum import auto, Enum
|
4 |
+
from typing import List, Tuple, Any, Union
|
5 |
+
|
6 |
+
IGNORE_TOKEN_ID = -100
|
7 |
+
|
8 |
+
|
9 |
class AlpacaPrompter:
|
10 |
+
prompt_input = "Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n{instruction}\n\n### Input:\n{input}\n\n### Response:\n"
|
11 |
+
prompt_no_input = "Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{instruction}\n\n### Response:\n"
|
12 |
+
response_split = "### Response:"
|
13 |
+
|
14 |
+
def build_prompt(
|
15 |
+
self,
|
16 |
+
instruction: str,
|
17 |
+
input: Union[None, str] = None,
|
18 |
+
output: Union[None, str] = None,
|
19 |
+
) -> str:
|
20 |
+
# returns the full prompt from instruction and optional input
|
21 |
+
# if a label (=response, =output) is provided, it's also appended.
|
22 |
+
if input:
|
23 |
+
res = self.prompt_input.format(
|
24 |
+
instruction=instruction, input=input
|
25 |
+
)
|
26 |
+
else:
|
27 |
+
res = self.prompt_no_input.format(
|
28 |
+
instruction=instruction
|
29 |
+
)
|
30 |
+
if output:
|
31 |
+
res = f"{res}{output}"
|
32 |
+
return res
|
33 |
+
|
34 |
+
def get_response(self, output: str) -> str:
|
35 |
+
return output.split(self.response_split)[1].strip()
|
36 |
+
|
37 |
+
|
38 |
+
class GPTeacherPrompter(AlpacaPrompter):
|
39 |
+
...
|
40 |
+
|
41 |
+
|
42 |
+
class SeparatorStyle(Enum):
|
43 |
+
"""Different separator style."""
|
44 |
+
SINGLE = auto()
|
45 |
+
TWO = auto()
|
46 |
+
DOLLY = auto()
|
47 |
+
|
48 |
+
|
49 |
+
# TODO clean this 💩 up
|
50 |
+
@dataclasses.dataclass
|
51 |
+
class Conversation:
|
52 |
+
"""A class that keeps all conversation history."""
|
53 |
+
system: str
|
54 |
+
roles: List[str]
|
55 |
+
messages: List[List[str]]
|
56 |
+
offset: int
|
57 |
+
sep_style: SeparatorStyle = SeparatorStyle.SINGLE
|
58 |
+
sep: str = "###"
|
59 |
+
sep2: str = None
|
60 |
+
|
61 |
+
def get_prompt(self):
|
62 |
+
seps = [self.sep, self.sep2]
|
63 |
+
ret = self.system + seps[0]
|
64 |
+
for i, (role, message) in enumerate(self.messages):
|
65 |
+
if message:
|
66 |
+
ret += role + ": " + message + seps[i % 2]
|
67 |
+
else:
|
68 |
+
ret += role + ":"
|
69 |
+
return ret
|
70 |
+
|
71 |
+
def copy(self):
|
72 |
+
return Conversation(
|
73 |
+
system=self.system,
|
74 |
+
roles=self.roles,
|
75 |
+
messages=[[x, y] for x, y in self.messages],
|
76 |
+
offset=self.offset,
|
77 |
+
sep_style=self.sep_style,
|
78 |
+
sep=self.sep,
|
79 |
+
sep2=self.sep2,
|
80 |
+
)
|
81 |
+
|
82 |
+
def append_message(self, role, message):
|
83 |
+
self.messages.append([role, message])
|
84 |
+
|
85 |
+
|
86 |
+
conv_vicuna_v1_1 = Conversation(
|
87 |
+
system="A chat between a curious user and an artificial intelligence assistant. "
|
88 |
+
"The assistant gives helpful, detailed, and polite answers to the user's questions.",
|
89 |
+
roles=["USER", "ASSISTANT"],
|
90 |
+
messages=[],
|
91 |
+
offset=0,
|
92 |
+
sep_style=SeparatorStyle.TWO,
|
93 |
+
sep=" ",
|
94 |
+
sep2="</s>",
|
95 |
+
)
|
96 |
|
97 |
|
98 |
class ShareGPTPrompter:
|
99 |
+
def build_prompt(
|
100 |
+
self,
|
101 |
+
source,
|
102 |
+
tokenizer
|
103 |
+
):
|
104 |
+
if len(source) < 2:
|
105 |
+
# If there isn't a back and forth conversation, ignore it
|
106 |
+
# also happens on the data splitting leaving empty conversations
|
107 |
+
raise IndexError
|
108 |
+
|
109 |
+
conv = conv_vicuna_v1_1.copy()
|
110 |
+
roles = {"human": conv.roles[0], "gpt": conv.roles[1]}
|
111 |
+
|
112 |
+
try:
|
113 |
+
# Apply prompt templates
|
114 |
+
if source[0]["from"] not in roles or roles[source[0]["from"]] != conv.roles[0]:
|
115 |
+
# Skip the first one if it is not from human
|
116 |
+
source = source[1:]
|
117 |
+
except IndexError as e:
|
118 |
+
# sometimes there is a bing or system chat
|
119 |
+
raise e
|
120 |
+
|
121 |
+
conv.messages = []
|
122 |
+
for j, sentence in enumerate(source):
|
123 |
+
role = roles[sentence["from"]]
|
124 |
+
assert role == conv.roles[j % 2]
|
125 |
+
conv.append_message(role, sentence["value"])
|
126 |
+
conversation = conv.get_prompt()
|
127 |
+
|
128 |
+
# Tokenize conversations
|
129 |
+
tokenized_result = tokenizer(
|
130 |
+
conversation,
|
131 |
+
truncation=True,
|
132 |
+
max_length=2048, # FIXME
|
133 |
+
padding=False,
|
134 |
+
return_tensors=None,
|
135 |
+
)
|
136 |
+
target = copy.deepcopy(tokenized_result["input_ids"])
|
137 |
+
|
138 |
+
# Mask targets
|
139 |
+
sep = conv.sep + conv.roles[1] + ": "
|
140 |
+
|
141 |
+
rounds = conversation.split(conv.sep2)
|
142 |
+
cur_len = 1
|
143 |
+
for i, rou in enumerate(rounds):
|
144 |
+
if rou == "":
|
145 |
+
break
|
146 |
+
|
147 |
+
parts = rou.split(sep)
|
148 |
+
if len(parts) != 2:
|
149 |
+
break
|
150 |
+
parts[0] += sep
|
151 |
+
round_len = len(tokenizer(rou)["input_ids"])
|
152 |
+
instruction_len = len(tokenizer(parts[0])["input_ids"]) - 2
|
153 |
+
target[cur_len:cur_len+instruction_len] = [IGNORE_TOKEN_ID] * instruction_len
|
154 |
|
155 |
+
cur_len += round_len
|
156 |
+
target[cur_len:] = [IGNORE_TOKEN_ID] * (len(target) - cur_len)
|
157 |
+
attention_mask = [1 if x != tokenizer.pad_token_id else 0 for x in tokenized_result["input_ids"]]
|
158 |
|
159 |
+
return dict(input_ids=tokenized_result["input_ids"], labels=target,
|
160 |
+
attention_mask=attention_mask)
|