yujunjun/LLMs-from-scratch
1
0
Fork 0
mirror of https://github.com/rasbt/LLMs-from-scratch.git synced 2025-11-04 03:40:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix BPE bonus materials (#561)

Browse Source 
* Fix BPE bonus materials

* fix bpe implementation

* update

* Add 'Hello, world. Is this-- a test?' test case

* update link to test file

* update path handling

* update path handling

* fix pytest paths
This commit is contained in:
Sebastian Raschka 2025-03-08 17:21:30 -06:00 committed by GitHub
parent 96ca2fcb2f
commit f63f04d8d5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
5 changed files with 307 additions and 87 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
.github/workflows/basic-tests-linux-uv.yml vendored
View File

@ -60,3 +60,9 @@ jobs:
pytest --ruff --nbval ch02/01_main-chapter-code/dataloader.ipynb
pytest --ruff --nbval ch03/01_main-chapter-code/multihead-attention.ipynb
pytest --ruff --nbval ch02/04_bonus_dataloader-intuition/dataloader-intuition.ipynb
- name: Test Selected Bonus Materials
shell: bash
run: |
source .venv/bin/activate
pytest ch02/05_bpe-from-scratch/tests/tests.py

8
.gitignore vendored
View File

@ -1,3 +1,4 @@
# Configs and keys
ch05/07_gpt_to_llama/config.json
ch07/02_dataset-utilities/config.json
@ -63,6 +64,8 @@ ch07/01_main-chapter-code/Smalltestmodel-sft-standalone.pth
ch07/01_main-chapter-code/gpt2/
# Datasets
the-verdict.txt
appendix-E/01_main-chapter-code/sms_spam_collection.zip
appendix-E/01_main-chapter-code/sms_spam_collection
appendix-E/01_main-chapter-code/train.csv
@ -70,6 +73,7 @@ appendix-E/01_main-chapter-code/test.csv
appendix-E/01_main-chapter-code/validation.csv
ch02/01_main-chapter-code/number-data.txt
ch02/05_bpe-from-scratch/the-verdict.txt
ch05/03_bonus_pretraining_on_gutenberg/gutenberg
ch05/03_bonus_pretraining_on_gutenberg/gutenberg_preprocessed
@ -107,7 +111,9 @@ ch02/05_bpe-from-scratch/bpe_merges.txt
ch02/05_bpe-from-scratch/encoder.json
ch02/05_bpe-from-scratch/vocab.bpe
ch02/05_bpe-from-scratch/vocab.json
encoder.json
vocab.bpe
vocab.json
# Other
ch0?/0?_user_interface/.chainlit/

36
ch02/02_bonus_bytepair-encoder/compare-bpe-tiktoken.ipynb
View File

@ -67,7 +67,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"tiktoken version: 0.7.0\n"
"tiktoken version: 0.9.0\n"
]
}
],
@ -180,8 +180,8 @@
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching encoder.json: 1.04Mit [00:00, 4.13Mit/s] \n",
"Fetching vocab.bpe: 457kit [00:00, 2.56Mit/s] \n"
"Fetching encoder.json: 1.04Mit [00:00, 3.69Mit/s] \n",
"Fetching vocab.bpe: 457kit [00:00, 2.53Mit/s] \n"
]
}
],
@ -256,10 +256,18 @@
"id": "e9077bf4-f91f-42ad-ab76-f3d89128510e",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Users/sebastian/Developer/LLMs-from-scratch/.venv/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
" from .autonotebook import tqdm as notebook_tqdm\n"
]
},
{
"data": {
"text/plain": [
"'4.48.0'"
"'4.49.0'"
]
},
"execution_count": 12,
@ -423,7 +431,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"[1544, 18798, 11, 995, 13, 1148, 256, 5303, 82, 438, 257, 1332, 30]\n"
"[15496, 11, 995, 13, 1148, 428, 438, 257, 1332, 30]\n"
]
}
],
@ -451,7 +459,7 @@
"metadata": {},
"outputs": [],
"source": [
"with open('../01_main-chapter-code/the-verdict.txt', 'r', encoding='utf-8') as f:\n",
"with open(\"../01_main-chapter-code/the-verdict.txt\", \"r\", encoding=\"utf-8\") as f:\n",
" raw_text = f.read()"
]
},
@ -473,7 +481,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"3.39 ms ± 21.9 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
"3.84 ms ± 9.83 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -499,7 +507,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"1.08 ms ± 5.99 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)\n"
"901 μs ± 6.27 μs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)\n"
]
}
],
@ -532,7 +540,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"10.2 ms ± 115 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
"11 ms ± 94.4 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -550,7 +558,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"10 ms ± 36.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
"10.8 ms ± 180 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -575,7 +583,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"3.79 ms ± 48.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
"3.66 ms ± 3.67 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -593,7 +601,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"3.83 ms ± 58.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
"3.77 ms ± 49.3 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -619,7 +627,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
"1.59 ms ± 11.5 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)\n"
"9.37 ms ± 50.3 μs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
]
}
],
@ -644,7 +652,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.4"
"version": "3.10.16"
}
},
"nbformat": 4,

159
ch02/05_bpe-from-scratch/bpe-from-scratch.ipynb
View File

@ -382,7 +382,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 4,
"id": "3e4a15ec-2667-4f56-b7c1-34e8071b621d",
"metadata": {},
"outputs": [],
@ -401,6 +401,10 @@
" # Dictionary of BPE merges: {(token_id1, token_id2): merged_token_id}\n",
" self.bpe_merges = {}\n",
"\n",
" # For the official OpenAI GPT-2 merges, use a rank dict:\n",
" # of form {(string_A, string_B): rank}, where lower rank = higher priority\n",
" self.bpe_ranks = {}\n",
"\n",
" def train(self, text, vocab_size, allowed_special={\"<|endoftext|>\"}):\n",
" \"\"\"\n",
" Train the BPE tokenizer from scratch.\n",
@ -411,7 +415,7 @@
" allowed_special (set): A set of special tokens to include.\n",
" \"\"\"\n",
"\n",
" # Preprocess: Replace spaces with 'Ġ'\n",
" # Preprocess: Replace spaces with \"Ġ\"\n",
" # Note that Ġ is a particularity of the GPT-2 BPE implementation\n",
" # E.g., \"Hello world\" might be tokenized as [\"Hello\", \"Ġworld\"]\n",
" # (GPT-4 BPE would tokenize it as [\"Hello\", \" world\"])\n",
@ -423,18 +427,16 @@
" processed_text.append(char)\n",
" processed_text = \"\".join(processed_text)\n",
"\n",
" # Initialize vocab with unique characters, including 'Ġ' if present\n",
" # Initialize vocab with unique characters, including \"Ġ\" if present\n",
" # Start with the first 256 ASCII characters\n",
" unique_chars = [chr(i) for i in range(256)]\n",
"\n",
" # Extend unique_chars with characters from processed_text that are not already included\n",
" unique_chars.extend(char for char in sorted(set(processed_text)) if char not in unique_chars)\n",
"\n",
" # Optionally, ensure 'Ġ' is included if it is relevant to your text processing\n",
" unique_chars.extend(\n",
" char for char in sorted(set(processed_text))\n",
" if char not in unique_chars\n",
" )\n",
" if \"Ġ\" not in unique_chars:\n",
" unique_chars.append(\"Ġ\")\n",
"\n",
" # Now create the vocab and inverse vocab dictionaries\n",
" self.vocab = {i: char for i, char in enumerate(unique_chars)}\n",
" self.inverse_vocab = {char: i for i, char in self.vocab.items()}\n",
"\n",
@ -452,7 +454,7 @@
" # BPE steps 1-3: Repeatedly find and replace frequent pairs\n",
" for new_id in range(len(self.vocab), vocab_size):\n",
" pair_id = self.find_freq_pair(token_ids, mode=\"most\")\n",
" if pair_id is None: # No more pairs to merge. Stopping training.\n",
" if pair_id is None:\n",
" break\n",
" token_ids = self.replace_pair(token_ids, pair_id, new_id)\n",
" self.bpe_merges[pair_id] = new_id\n",
@ -492,29 +494,24 @@
" self.inverse_vocab[\"\\n\"] = newline_token_id\n",
" self.vocab[newline_token_id] = \"\\n\"\n",
"\n",
" # Load BPE merges\n",
" # Load GPT-2 merges and store them with an assigned \"rank\"\n",
" self.bpe_ranks = {} # reset ranks\n",
" with open(bpe_merges_path, \"r\", encoding=\"utf-8\") as file:\n",
" lines = file.readlines()\n",
" # Skip header line if present\n",
" if lines and lines[0].startswith(\"#\"):\n",
" lines = lines[1:]\n",
"\n",
" rank = 0\n",
" for line in lines:\n",
" pair = tuple(line.strip().split())\n",
" if len(pair) == 2:\n",
" token1, token2 = pair\n",
" # If token1 or token2 not in vocab, skip\n",
" if token1 in self.inverse_vocab and token2 in self.inverse_vocab:\n",
" token_id1 = self.inverse_vocab[token1]\n",
" token_id2 = self.inverse_vocab[token2]\n",
" merged_token = token1 + token2\n",
" if merged_token in self.inverse_vocab:\n",
" merged_token_id = self.inverse_vocab[merged_token]\n",
" self.bpe_merges[(token_id1, token_id2)] = merged_token_id\n",
" # print(f\"Loaded merge: '{token1}' + '{token2}' -> '{merged_token}' (ID: {merged_token_id})\")\n",
" self.bpe_ranks[(token1, token2)] = rank\n",
" rank += 1\n",
" else:\n",
" print(f\"Merged token '{merged_token}' not found in vocab. Skipping.\")\n",
" else:\n",
" print(f\"Skipping pair {pair} as one of the tokens is not in the vocabulary.\")\n",
" print(f\"Skipping pair {pair} as one token is not in the vocabulary.\")\n",
"\n",
" def encode(self, text):\n",
" \"\"\"\n",
@ -540,7 +537,7 @@
" else:\n",
" tokens.append(word)\n",
" else:\n",
" # Prefix words in the middle of a line with 'Ġ'\n",
" # Prefix words in the middle of a line with \"Ġ\"\n",
" tokens.append(\"Ġ\" + word)\n",
"\n",
" token_ids = []\n",
@ -571,6 +568,8 @@
" missing_chars = [char for char, tid in zip(token, token_ids) if tid is None]\n",
" raise ValueError(f\"Characters not found in vocab: {missing_chars}\")\n",
"\n",
" # If we haven't loaded OpenAI's GPT-2 merges, use my approach\n",
" if not self.bpe_ranks:\n",
" can_merge = True\n",
" while can_merge and len(token_ids) > 1:\n",
" can_merge = False\n",
@ -591,9 +590,53 @@
" if i < len(token_ids):\n",
" new_tokens.append(token_ids[i])\n",
" token_ids = new_tokens\n",
"\n",
" return token_ids\n",
"\n",
" # Otherwise, do GPT-2-style merging with the ranks:\n",
" # 1) Convert token_ids back to string \"symbols\" for each ID\n",
" symbols = [self.vocab[id_num] for id_num in token_ids]\n",
"\n",
" # Repeatedly merge all occurrences of the lowest-rank pair\n",
" while True:\n",
" # Collect all adjacent pairs\n",
" pairs = set(zip(symbols, symbols[1:]))\n",
" if not pairs:\n",
" break\n",
"\n",
" # Find the pair with the best (lowest) rank\n",
" min_rank = 1_000_000_000\n",
" bigram = None\n",
" for p in pairs:\n",
" r = self.bpe_ranks.get(p, 1_000_000_000)\n",
" if r < min_rank:\n",
" min_rank = r\n",
" bigram = p\n",
"\n",
" # If no valid ranked pair is present, we're done\n",
" if bigram is None or bigram not in self.bpe_ranks:\n",
" break\n",
"\n",
" # Merge all occurrences of that pair\n",
" first, second = bigram\n",
" new_symbols = []\n",
" i = 0\n",
" while i < len(symbols):\n",
" # If we see (first, second) at position i, merge them\n",
" if i < len(symbols) - 1 and symbols[i] == first and symbols[i+1] == second:\n",
" new_symbols.append(first + second) # merged symbol\n",
" i += 2\n",
" else:\n",
" new_symbols.append(symbols[i])\n",
" i += 1\n",
" symbols = new_symbols\n",
"\n",
" if len(symbols) == 1:\n",
" break\n",
"\n",
" # Finally, convert merged symbols back to IDs\n",
" merged_ids = [self.inverse_vocab[sym] for sym in symbols]\n",
" return merged_ids\n",
"\n",
" def decode(self, token_ids):\n",
" \"\"\"\n",
" Decode a list of token IDs back into a string.\n",
@ -738,22 +781,49 @@
},
{
"cell_type": "code",
"execution_count": 5,
"id": "4d197cad-ed10-4a42-b01c-a763859781fb",
"execution_count": 25,
"id": "51872c08-e01b-40c3-a8a0-e8d6a773e3df",
"metadata": {},
"outputs": [],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"the-verdict.txt already exists in ./the-verdict.txt\n"
]
}
],
"source": [
"import os\n",
"import urllib.request\n",
"\n",
"if not os.path.exists(\"../01_main-chapter-code/the-verdict.txt\"):\n",
" url = (\"https://raw.githubusercontent.com/rasbt/\"\n",
" \"LLMs-from-scratch/main/ch02/01_main-chapter-code/\"\n",
" \"the-verdict.txt\")\n",
" file_path = \"../01_main-chapter-code/the-verdict.txt\"\n",
" urllib.request.urlretrieve(url, file_path)\n",
"def download_file_if_absent(url, filename, search_dirs):\n",
" for directory in search_dirs:\n",
" file_path = os.path.join(directory, filename)\n",
" if os.path.exists(file_path):\n",
" print(f\"{filename} already exists in {file_path}\")\n",
" return file_path\n",
"\n",
"with open(\"../01_main-chapter-code/the-verdict.txt\", \"r\", encoding=\"utf-8\") as f: # added ../01_main-chapter-code/\n",
" target_path = os.path.join(search_dirs[0], filename)\n",
" try:\n",
" with urllib.request.urlopen(url) as response, open(target_path, \"wb\") as out_file:\n",
" out_file.write(response.read())\n",
" print(f\"Downloaded {filename} to {target_path}\")\n",
" except Exception as e:\n",
" print(f\"Failed to download {filename}. Error: {e}\")\n",
" return target_path\n",
"\n",
"verdict_path = download_file_if_absent(\n",
" url=(\n",
" \"https://raw.githubusercontent.com/rasbt/\"\n",
" \"LLMs-from-scratch/main/ch02/01_main-chapter-code/\"\n",
" \"the-verdict.txt\"\n",
" ),\n",
" filename=\"the-verdict.txt\",\n",
" search_dirs=\".\"\n",
")\n",
"\n",
"with open(verdict_path, \"r\", encoding=\"utf-8\") as f: # added ../01_main-chapter-code/\n",
" text = f.read()"
]
},
@ -1168,24 +1238,7 @@
}
],
"source": [
"import os\n",
"import urllib.request\n",
"\n",
"def download_file_if_absent(url, filename, search_dirs):\n",
" for directory in search_dirs:\n",
" file_path = os.path.join(directory, filename)\n",
" if os.path.exists(file_path):\n",
" print(f\"{filename} already exists in {file_path}\")\n",
" return file_path\n",
"\n",
" target_path = os.path.join(search_dirs[0], filename)\n",
" try:\n",
" with urllib.request.urlopen(url) as response, open(target_path, \"wb\") as out_file:\n",
" out_file.write(response.read())\n",
" print(f\"Downloaded {filename} to {target_path}\")\n",
" except Exception as e:\n",
" print(f\"Failed to download {filename}. Error: {e}\")\n",
" return target_path\n",
"# Download files if not already present in this directory\n",
"\n",
"# Define the directories to search and the files to download\n",
"search_directories = [\".\", \"../02_bonus_bytepair-encoder/gpt2_model/\"]\n",
@ -1351,7 +1404,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.6"
"version": "3.10.16"
}
},
"nbformat": 4,

147
ch02/05_bpe-from-scratch/tests/tests.py Normal file
View File

@ -0,0 +1,147 @@
import os
import sys
import io
import nbformat
import types
import pytest
import tiktoken
def import_definitions_from_notebook(fullname, names):
"""Loads function definitions from a Jupyter notebook file into a module."""
path = os.path.join(os.path.dirname(__file__), "..", fullname + ".ipynb")
path = os.path.normpath(path)
if not os.path.exists(path):
raise FileNotFoundError(f"Notebook file not found at: {path}")
with io.open(path, "r", encoding="utf-8") as f:
nb = nbformat.read(f, as_version=4)
mod = types.ModuleType(fullname)
sys.modules[fullname] = mod
# Execute all code cells to capture dependencies
for cell in nb.cells:
if cell.cell_type == "code":
exec(cell.source, mod.__dict__)
# Ensure required names are in module
missing_names = [name for name in names if name not in mod.__dict__]
if missing_names:
raise ImportError(f"Missing definitions in notebook: {missing_names}")
return mod
@pytest.fixture(scope="module")
def imported_module():
fullname = "bpe-from-scratch"
names = ["BPETokenizerSimple", "download_file_if_absent"]
return import_definitions_from_notebook(fullname, names)
@pytest.fixture(scope="module")
def gpt2_files(imported_module):
"""Fixture to handle downloading GPT-2 files."""
download_file_if_absent = getattr(imported_module, "download_file_if_absent", None)
search_directories = [".", "../02_bonus_bytepair-encoder/gpt2_model/"]
files_to_download = {
"https://openaipublic.blob.core.windows.net/gpt-2/models/124M/vocab.bpe": "vocab.bpe",
"https://openaipublic.blob.core.windows.net/gpt-2/models/124M/encoder.json": "encoder.json"
}
paths = {filename: download_file_if_absent(url, filename, search_directories)
for url, filename in files_to_download.items()}
return paths
def test_tokenizer_training(imported_module, gpt2_files):
BPETokenizerSimple = getattr(imported_module, "BPETokenizerSimple", None)
download_file_if_absent = getattr(imported_module, "download_file_if_absent", None)
tokenizer = BPETokenizerSimple()
verdict_path = download_file_if_absent(
url=(
"https://raw.githubusercontent.com/rasbt/"
"LLMs-from-scratch/main/ch02/01_main-chapter-code/"
"the-verdict.txt"
),
filename="the-verdict.txt",
search_dirs="."
)
with open(verdict_path, "r", encoding="utf-8") as f: # added ../01_main-chapter-code/
text = f.read()
tokenizer.train(text, vocab_size=1000, allowed_special={"<|endoftext|>"})
assert len(tokenizer.vocab) == 1000, "Tokenizer vocabulary size mismatch."
assert len(tokenizer.bpe_merges) == 742, "Tokenizer BPE merges count mismatch."
input_text = "Jack embraced beauty through art and life."
token_ids = tokenizer.encode(input_text)
assert token_ids == [424, 256, 654, 531, 302, 311, 256, 296, 97, 465, 121, 595, 841, 116, 287, 466, 256, 326, 972, 46], "Token IDs do not match expected output."
assert tokenizer.decode(token_ids) == input_text, "Decoded text does not match the original input."
tokenizer.save_vocab_and_merges(vocab_path="vocab.json", bpe_merges_path="bpe_merges.txt")
tokenizer2 = BPETokenizerSimple()
tokenizer2.load_vocab_and_merges(vocab_path="vocab.json", bpe_merges_path="bpe_merges.txt")
assert tokenizer2.decode(token_ids) == input_text, "Decoded text mismatch after reloading tokenizer."
def test_gpt2_tokenizer_openai_simple(imported_module, gpt2_files):
BPETokenizerSimple = getattr(imported_module, "BPETokenizerSimple", None)
tokenizer_gpt2 = BPETokenizerSimple()
tokenizer_gpt2.load_vocab_and_merges_from_openai(
vocab_path=gpt2_files["encoder.json"], bpe_merges_path=gpt2_files["vocab.bpe"]
)
assert len(tokenizer_gpt2.vocab) == 50257, "GPT-2 tokenizer vocabulary size mismatch."
input_text = "This is some text"
token_ids = tokenizer_gpt2.encode(input_text)
assert token_ids == [1212, 318, 617, 2420], "Tokenized output does not match expected GPT-2 encoding."
def test_gpt2_tokenizer_openai_edgecases(imported_module, gpt2_files):
BPETokenizerSimple = getattr(imported_module, "BPETokenizerSimple", None)
tokenizer_gpt2 = BPETokenizerSimple()
tokenizer_gpt2.load_vocab_and_merges_from_openai(
vocab_path=gpt2_files["encoder.json"], bpe_merges_path=gpt2_files["vocab.bpe"]
)
tik_tokenizer = tiktoken.get_encoding("gpt2")
test_cases = [
("Hello,", [15496, 11]),
("Implementations", [3546, 26908, 602]),
("asdf asdfasdf a!!, @aba 9asdf90asdfk", [292, 7568, 355, 7568, 292, 7568, 257, 3228, 11, 2488, 15498, 860, 292, 7568, 3829, 292, 7568, 74]),
("Hello, world. Is this-- a test?", [15496, 11, 995, 13, 1148, 428, 438, 257, 1332, 30])
]
errors = []
for input_text, expected_tokens in test_cases:
tik_tokens = tik_tokenizer.encode(input_text)
gpt2_tokens = tokenizer_gpt2.encode(input_text)
print(f"Text: {input_text}")
print(f"Expected Tokens: {expected_tokens}")
print(f"tiktoken Output: {tik_tokens}")
print(f"BPETokenizerSimple Output: {gpt2_tokens}")
print("-" * 40)
if tik_tokens != expected_tokens:
errors.append(f"Tiktokenized output does not match expected GPT-2 encoding for '{input_text}'.\n"
f"Expected: {expected_tokens}, Got: {tik_tokens}")
if gpt2_tokens != expected_tokens:
errors.append(f"Tokenized output does not match expected GPT-2 encoding for '{input_text}'.\n"
f"Expected: {expected_tokens}, Got: {gpt2_tokens}")
if errors:
pytest.fail("\n".join(errors))