Update my_model/fine_tuner/fine_tuning_data_handler.py
Browse files
my_model/fine_tuner/fine_tuning_data_handler.py
CHANGED
@@ -1,11 +1,10 @@
|
|
|
|
1 |
from my_model.utilities.gen_utilities import is_pycharm
|
2 |
import seaborn as sns
|
3 |
from transformers import AutoTokenizer
|
4 |
from datasets import Dataset, load_dataset
|
5 |
import my_model.config.fine_tuning_config as config
|
6 |
from my_model.LLAMA2.LLAMA2_model import Llama2ModelManager
|
7 |
-
from typing import Tuple
|
8 |
-
|
9 |
|
10 |
|
11 |
class FinetuningDataHandler:
|
@@ -13,7 +12,7 @@ class FinetuningDataHandler:
|
|
13 |
A class dedicated to handling data for fine-tuning language models. It manages loading,
|
14 |
inspecting, preparing, and splitting the dataset, specifically designed to filter out
|
15 |
data samples exceeding a specified token count limit. This is crucial for models with
|
16 |
-
token count constraints and it helps control the level of GPU RAM
|
17 |
ensuring efficient and effective model fine-tuning.
|
18 |
|
19 |
Attributes:
|
@@ -22,14 +21,15 @@ class FinetuningDataHandler:
|
|
22 |
max_token_count (int): Maximum allowable token count per data sample.
|
23 |
|
24 |
Methods:
|
25 |
-
load_llm_tokenizer
|
26 |
-
load_dataset
|
27 |
-
plot_tokens_count_distribution
|
28 |
-
filter_dataset_by_indices
|
29 |
-
get_token_counts
|
30 |
-
prepare_dataset
|
31 |
-
|
32 |
-
|
|
|
33 |
"""
|
34 |
|
35 |
def __init__(self, tokenizer: AutoTokenizer = None, dataset_file: str = config.DATASET_FILE) -> None:
|
@@ -37,17 +37,21 @@ class FinetuningDataHandler:
|
|
37 |
Initializes the FinetuningDataHandler class.
|
38 |
|
39 |
Args:
|
40 |
-
tokenizer (AutoTokenizer): Tokenizer to use for tokenizing the dataset.
|
41 |
-
dataset_file (str): Path to the dataset file.
|
42 |
"""
|
|
|
43 |
self.tokenizer = tokenizer # The tokenizer used for processing the dataset.
|
44 |
self.dataset_file = dataset_file # Path to the fine-tuning dataset file.
|
45 |
-
self.max_token_count = config.MAX_TOKEN_COUNT # Max token count for filtering.
|
46 |
|
47 |
-
def load_llm_tokenizer(self):
|
48 |
"""
|
49 |
Loads the LLM tokenizer and adds special tokens, if not already loaded.
|
50 |
If the tokenizer is already loaded, this method does nothing.
|
|
|
|
|
|
|
51 |
"""
|
52 |
|
53 |
if self.tokenizer is None:
|
@@ -63,21 +67,26 @@ class FinetuningDataHandler:
|
|
63 |
Returns:
|
64 |
Dataset: The loaded dataset, ready for processing.
|
65 |
"""
|
|
|
66 |
return load_dataset('csv', data_files=self.dataset_file)
|
67 |
|
68 |
-
def plot_tokens_count_distribution(self, token_counts:
|
69 |
"""
|
70 |
Plots the distribution of token counts in the dataset for visualization purposes.
|
71 |
|
72 |
Args:
|
73 |
-
token_counts (
|
|
|
74 |
title (str): Title for the plot, highlighting the nature of the distribution.
|
|
|
|
|
|
|
75 |
"""
|
76 |
|
77 |
if is_pycharm(): # Ensuring compatibility with PyCharm's environment for interactive plot.
|
78 |
-
import matplotlib
|
79 |
matplotlib.use('TkAgg') # Set the backend to 'TkAgg'
|
80 |
-
import matplotlib.pyplot as plt
|
81 |
sns.set_style("whitegrid")
|
82 |
plt.figure(figsize=(15, 6))
|
83 |
plt.hist(token_counts, bins=50, color='#3498db', edgecolor='black')
|
@@ -89,21 +98,21 @@ class FinetuningDataHandler:
|
|
89 |
plt.tight_layout()
|
90 |
plt.show()
|
91 |
|
92 |
-
def filter_dataset_by_indices(self, dataset: Dataset, valid_indices:
|
93 |
"""
|
94 |
Filters the dataset based on a list of valid indices. This method is used to exclude
|
95 |
data samples that have a token count exceeding the specified maximum token count.
|
96 |
|
97 |
Args:
|
98 |
dataset (Dataset): The dataset to be filtered.
|
99 |
-
valid_indices (
|
100 |
|
101 |
Returns:
|
102 |
Dataset: Filtered dataset containing only samples with valid indices.
|
103 |
"""
|
104 |
return dataset['train'].select(valid_indices) # Select only samples with valid indices based on token count.
|
105 |
|
106 |
-
def get_token_counts(self, dataset):
|
107 |
"""
|
108 |
Calculates and returns the token counts for each sample in the dataset.
|
109 |
This function assumes the dataset has a 'train' split and a 'text' field.
|
@@ -131,6 +140,7 @@ class FinetuningDataHandler:
|
|
131 |
Returns:
|
132 |
Tuple[Dataset, Dataset]: The train and evaluate datasets, post-filtering.
|
133 |
"""
|
|
|
134 |
dataset = self.load_dataset()
|
135 |
self.load_llm_tokenizer()
|
136 |
|
@@ -148,7 +158,7 @@ class FinetuningDataHandler:
|
|
148 |
|
149 |
return self.split_dataset_for_train_eval(filtered_dataset) # split the dataset into training and evaluation.
|
150 |
|
151 |
-
def split_dataset_for_train_eval(self, dataset) -> Tuple[Dataset, Dataset]:
|
152 |
"""
|
153 |
Splits the dataset into training and evaluation datasets.
|
154 |
|
@@ -156,27 +166,29 @@ class FinetuningDataHandler:
|
|
156 |
dataset (Dataset): The dataset to split.
|
157 |
|
158 |
Returns:
|
159 |
-
|
160 |
"""
|
|
|
161 |
split_data = dataset.train_test_split(test_size=config.TEST_SIZE, shuffle=True, seed=config.SEED)
|
162 |
train_data, eval_data = split_data['train'], split_data['test']
|
163 |
return train_data, eval_data
|
164 |
|
165 |
-
def inspect_prepare_split_data(self) ->
|
166 |
"""
|
167 |
Orchestrates the process of inspecting, preparing, and splitting the dataset for fine-tuning.
|
168 |
|
169 |
Returns:
|
170 |
-
|
171 |
"""
|
|
|
172 |
return self.prepare_dataset()
|
173 |
|
174 |
|
175 |
# Example usage
|
176 |
if __name__ == "__main__":
|
177 |
-
|
178 |
-
#
|
179 |
-
#data_handler = FinetuningDataHandler()
|
180 |
-
#fine_tuning_data_train, fine_tuning_data_eval = data_handler.inspect_prepare_split_data()
|
181 |
-
#print(fine_tuning_data_train, fine_tuning_data_eval)
|
182 |
pass
|
|
|
1 |
+
from typing import Tuple, List
|
2 |
from my_model.utilities.gen_utilities import is_pycharm
|
3 |
import seaborn as sns
|
4 |
from transformers import AutoTokenizer
|
5 |
from datasets import Dataset, load_dataset
|
6 |
import my_model.config.fine_tuning_config as config
|
7 |
from my_model.LLAMA2.LLAMA2_model import Llama2ModelManager
|
|
|
|
|
8 |
|
9 |
|
10 |
class FinetuningDataHandler:
|
|
|
12 |
A class dedicated to handling data for fine-tuning language models. It manages loading,
|
13 |
inspecting, preparing, and splitting the dataset, specifically designed to filter out
|
14 |
data samples exceeding a specified token count limit. This is crucial for models with
|
15 |
+
token count constraints and it helps control the level of GPU RAM tolerance based on the number of tokens,
|
16 |
ensuring efficient and effective model fine-tuning.
|
17 |
|
18 |
Attributes:
|
|
|
21 |
max_token_count (int): Maximum allowable token count per data sample.
|
22 |
|
23 |
Methods:
|
24 |
+
load_llm_tokenizer: Loads the LLM tokenizer and adds special tokens, if not already loaded.
|
25 |
+
load_dataset: Loads the dataset from a specified file path.
|
26 |
+
plot_tokens_count_distribution: Plots the distribution of token counts in the dataset.
|
27 |
+
filter_dataset_by_indices: Filters the dataset based on valid indices, removing samples exceeding token limits.
|
28 |
+
get_token_counts: Calculates token counts for each sample in the dataset.
|
29 |
+
prepare_dataset: Tokenizes and filters the dataset, preparing it for training. Also visualizes token count
|
30 |
+
distribution before and after filtering.
|
31 |
+
split_dataset_for_train_eval: Divides the dataset into training and evaluation sets.
|
32 |
+
inspect_prepare_split_data: Coordinates the data preparation and splitting process for fine-tuning.
|
33 |
"""
|
34 |
|
35 |
def __init__(self, tokenizer: AutoTokenizer = None, dataset_file: str = config.DATASET_FILE) -> None:
|
|
|
37 |
Initializes the FinetuningDataHandler class.
|
38 |
|
39 |
Args:
|
40 |
+
tokenizer (AutoTokenizer, optional): Tokenizer to use for tokenizing the dataset. Defaults to None.
|
41 |
+
dataset_file (str): Path to the dataset file. Defaults to config.DATASET_FILE.
|
42 |
"""
|
43 |
+
|
44 |
self.tokenizer = tokenizer # The tokenizer used for processing the dataset.
|
45 |
self.dataset_file = dataset_file # Path to the fine-tuning dataset file.
|
46 |
+
self.max_token_count = config.MAX_TOKEN_COUNT # Max token count for filtering set to 1,024.
|
47 |
|
48 |
+
def load_llm_tokenizer(self) -> None:
|
49 |
"""
|
50 |
Loads the LLM tokenizer and adds special tokens, if not already loaded.
|
51 |
If the tokenizer is already loaded, this method does nothing.
|
52 |
+
|
53 |
+
Returns:
|
54 |
+
None
|
55 |
"""
|
56 |
|
57 |
if self.tokenizer is None:
|
|
|
67 |
Returns:
|
68 |
Dataset: The loaded dataset, ready for processing.
|
69 |
"""
|
70 |
+
|
71 |
return load_dataset('csv', data_files=self.dataset_file)
|
72 |
|
73 |
+
def plot_tokens_count_distribution(self, token_counts: List[int], title: str = "Token Count Distribution") -> None:
|
74 |
"""
|
75 |
Plots the distribution of token counts in the dataset for visualization purposes.
|
76 |
|
77 |
Args:
|
78 |
+
token_counts (List[int]): List of token counts, each count representing the number of tokens in a dataset
|
79 |
+
sample.
|
80 |
title (str): Title for the plot, highlighting the nature of the distribution.
|
81 |
+
|
82 |
+
Returns:
|
83 |
+
None
|
84 |
"""
|
85 |
|
86 |
if is_pycharm(): # Ensuring compatibility with PyCharm's environment for interactive plot.
|
87 |
+
import matplotlib # The import is kept here intentionaly.
|
88 |
matplotlib.use('TkAgg') # Set the backend to 'TkAgg'
|
89 |
+
import matplotlib.pyplot as plt # The import is kept here intentionaly.
|
90 |
sns.set_style("whitegrid")
|
91 |
plt.figure(figsize=(15, 6))
|
92 |
plt.hist(token_counts, bins=50, color='#3498db', edgecolor='black')
|
|
|
98 |
plt.tight_layout()
|
99 |
plt.show()
|
100 |
|
101 |
+
def filter_dataset_by_indices(self, dataset: Dataset, valid_indices: List[int]) -> Dataset:
|
102 |
"""
|
103 |
Filters the dataset based on a list of valid indices. This method is used to exclude
|
104 |
data samples that have a token count exceeding the specified maximum token count.
|
105 |
|
106 |
Args:
|
107 |
dataset (Dataset): The dataset to be filtered.
|
108 |
+
valid_indices (List[int]): Indices of samples with token counts within the limit.
|
109 |
|
110 |
Returns:
|
111 |
Dataset: Filtered dataset containing only samples with valid indices.
|
112 |
"""
|
113 |
return dataset['train'].select(valid_indices) # Select only samples with valid indices based on token count.
|
114 |
|
115 |
+
def get_token_counts(self, dataset: Dataset) -> List[int]:
|
116 |
"""
|
117 |
Calculates and returns the token counts for each sample in the dataset.
|
118 |
This function assumes the dataset has a 'train' split and a 'text' field.
|
|
|
140 |
Returns:
|
141 |
Tuple[Dataset, Dataset]: The train and evaluate datasets, post-filtering.
|
142 |
"""
|
143 |
+
|
144 |
dataset = self.load_dataset()
|
145 |
self.load_llm_tokenizer()
|
146 |
|
|
|
158 |
|
159 |
return self.split_dataset_for_train_eval(filtered_dataset) # split the dataset into training and evaluation.
|
160 |
|
161 |
+
def split_dataset_for_train_eval(self, dataset: Dataset) -> Tuple[Dataset, Dataset]:
|
162 |
"""
|
163 |
Splits the dataset into training and evaluation datasets.
|
164 |
|
|
|
166 |
dataset (Dataset): The dataset to split.
|
167 |
|
168 |
Returns:
|
169 |
+
Tuple[Dataset, Dataset]: The split training and evaluation datasets.
|
170 |
"""
|
171 |
+
|
172 |
split_data = dataset.train_test_split(test_size=config.TEST_SIZE, shuffle=True, seed=config.SEED)
|
173 |
train_data, eval_data = split_data['train'], split_data['test']
|
174 |
return train_data, eval_data
|
175 |
|
176 |
+
def inspect_prepare_split_data(self) -> Tuple[Dataset, Dataset]:
|
177 |
"""
|
178 |
Orchestrates the process of inspecting, preparing, and splitting the dataset for fine-tuning.
|
179 |
|
180 |
Returns:
|
181 |
+
Tuple[Dataset, Dataset]: The prepared training and evaluation datasets.
|
182 |
"""
|
183 |
+
|
184 |
return self.prepare_dataset()
|
185 |
|
186 |
|
187 |
# Example usage
|
188 |
if __name__ == "__main__":
|
189 |
+
|
190 |
+
# Please uncomment the below lines to test the data prep.
|
191 |
+
# data_handler = FinetuningDataHandler()
|
192 |
+
# fine_tuning_data_train, fine_tuning_data_eval = data_handler.inspect_prepare_split_data()
|
193 |
+
# print(fine_tuning_data_train, fine_tuning_data_eval)
|
194 |
pass
|