Skip to content

BPE: replace solution code, fix Solution button opening the question #30

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Exorust merged 1 commit into from
Sep 15, 2026
+75 −65
Merged

BPE: replace solution code, fix Solution button opening the question #30

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
138 changes: 74 additions & 64 deletions llm/Byte-Pair-Encoder/BPE-q3-SOLN.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,54 +41,62 @@
"from collections import defaultdict, Counter\n",
"\n",
"def get_vocab(corpus):\n",
" \"\"\"Creates a vocabulary with words split into characters and a special end-of-word token.\"\"\"\n",
" vocab = Counter()\n",
" for word in corpus:\n",
" tokens = list(word) + ['</w>']\n",
" vocab[tuple(tokens)] += 1\n",
" return vocab\n",
" \"\"\"Creates a vocabulary with words split into characters and a special end of word token.\"\"\"\n",
" vocab = Counter()\n",
" for word in corpus:\n",
" tokens = list(word) + [\"\"]\n",
" vocab[tuple(tokens)]+=1\n",
" return vocab\n",
"\n",
"def get_stats(vocab):\n",
" \"\"\"Counts frequency of adjacent symbol pairs.\"\"\"\n",
" pairs = defaultdict(int)\n",
" for word, freq in vocab.items():\n",
" for i in range(len(word) - 1):\n",
" pairs[(word[i], word[i + 1])] += freq\n",
" return pairs\n",
" \"\"\"Counts frequency of adjacent symbol pairs.\"\"\"\n",
" pairs = defaultdict(int)\n",
" for word,freq in vocab.items():\n",
" for i in range(len(word)-1):\n",
" pairs[(word[i], word[i+1])] +=freq\n",
" return pairs\n",
"\n",
"def merge_vocab(pair, vocab):\n",
" \"\"\"Merges the most frequent pair into a single symbol.\"\"\"\n",
" new_vocab = {}\n",
" bigram = ' '.join(pair)\n",
" replacement = ''.join(pair)\n",
" for word, freq in vocab.items():\n",
" word_str = ' '.join(word)\n",
" # Replace bigram with merged symbol\n",
" new_word_str = word_str.replace(bigram, replacement)\n",
" new_vocab[tuple(new_word_str.split())] = freq\n",
" return new_vocab\n",
" \"\"\"Merges the most frequent pair into a single symbol.\"\"\"\n",
" new_vocab = {}\n",
" bigram = ' '.join(pair)\n",
" replacement = ''.join(pair)\n",
" print(f\"Bigram: {bigram}\")\n",
" for word,freq in vocab.items():\n",
" word_str = ' '.join(word)\n",
" #Replace bigram with merged symbol\n",
" new_word_str=word_str.replace(bigram, replacement)\n",
" #Preserve the trailing empty string token if the original word ended with it\n",
" tokens = new_word_str.split()\n",
" if word[-1] == \"\":\n",
" tokens.append(\"\")\n",
" new_vocab[tuple(tokens)]=freq\n",
" return new_vocab\n",
"\n",
"def byte_pair_encoding(corpus, num_merges=10):\n",
" \"\"\"Performs BPE on a corpus.\"\"\"\n",
" vocab = get_vocab(corpus)\n",
" merges = []\n",
" for _ in range(num_merges):\n",
" pairs = get_stats(vocab)\n",
" if not pairs:\n",
" break\n",
" best = max(pairs, key=pairs.get)\n",
" vocab = merge_vocab(best, vocab)\n",
" merges.append(best)\n",
" print(f\"Merge {_ + 1}: {best}\")\n",
" return vocab, merges\n",
"\n",
"# Example usage\n",
"corpus = [\"low\", \"lowest\", \"newer\", \"wider\"]\n",
"final_vocab, merge_operations = byte_pair_encoding(corpus, num_merges=10)\n",
"def byte_pair_encoding(corpus, num_merges=10):\n",
" \"\"\"Performs BPE on a corpus.\"\"\"\n",
" vocab = get_vocab(corpus)\n",
" merges = []\n",
" for _ in range(num_merges):\n",
" pairs = get_stats(vocab)\n",
" if not pairs:\n",
" break\n",
" best = max(pairs, key=pairs.get)\n",
" vocab = merge_vocab(best, vocab)\n",
" merges.append(best)\n",
" print(f\"Merge {_ + 1}: {best}\")\n",
" return vocab, merges\n",
"\n",
"corpus = [\"low\", \"lowest\", \"newer\", \"wider\", \"owl\", \"LOW\"]\n",
"final_vocab , merge_operations = byte_pair_encoding(corpus, num_merges=10)\n",
"# vocab = get_vocab(corpus)\n",
"# pairs = get_stats(vocab)\n",
"# print(pairs)\n",
"# new_vocab = merge_vocab(pairs, vocab)\n",
"print(\"\\nFinal Vocabulary:\")\n",
"for word in final_vocab:\n",
" print(' '.join(word), \":\", final_vocab[word])\n"
" print(\"\".join(word), \":\", final_vocab[word])"
]
},
{
Expand All @@ -99,41 +107,43 @@
"outputs": [],
"source": [
"def test_get_vocab():\n",
" corpus = [\"test\"]\n",
" vocab = get_vocab(corpus)\n",
" assert vocab == {('t', 'e', 's', 't', '</w>'): 1}\n",
" print(\"✓ test_get_vocab passed\")\n",
" corpus = ['test']\n",
" vocab = get_vocab(corpus)\n",
" assert vocab == {(\"t\", \"e\", \"s\", \"t\", \"\"):1}\n",
" print(\"✓\")\n",
"\n",
"def test_get_stats():\n",
" vocab = {('t', 'e', 's', 't', '</w>'): 1}\n",
" stats = get_stats(vocab)\n",
" expected = {\n",
"def test_get_status():\n",
" vocab = {('t', 'e', 's', 't', '' ):1}\n",
" stats = get_stats(vocab)\n",
" expected = {\n",
" ('t', 'e'): 1,\n",
" ('e', 's'): 1,\n",
" ('s', 't'): 1,\n",
" ('t', '</w>'): 1\n",
" }\n",
" assert stats == expected\n",
" print(\"✓ test_get_stats passed\")\n",
" ('t', ''): 1\n",
" }\n",
" assert stats == expected\n",
" print(\"✓✓\")\n",
"\n",
"def test_merge_vocab():\n",
" vocab = {('t', 'e', 's', 't', '</w>'): 1}\n",
" merged = merge_vocab(('e', 's'), vocab)\n",
" expected = {('t', 'es', 't', '</w>'): 1}\n",
" assert merged == expected\n",
" print(\"✓ test_merge_vocab passed\")\n",
" vocab = {('t', 'e', 's', 't', '' ):1}\n",
" merged = merge_vocab((\"e\", \"s\"),vocab)\n",
" print(merged)\n",
" expected = {(\"t\", \"es\", 't', ''):1}\n",
" assert merged == expected\n",
" print(\"✓✓✓\")\n",
"\n",
"def test_bpe_sequence():\n",
" corpus = [\"low\", \"lower\", \"newest\", \"widest\"]\n",
" final_vocab, merges = byte_pair_encoding(corpus, num_merges=5)\n",
" assert isinstance(final_vocab, dict)\n",
" assert all(isinstance(pair, tuple) for pair in merges)\n",
" assert len(merges) == 5\n",
" print(\"✓ test_bpe_sequence passed\")\n",
" corpus = ['low', 'lower', 'lower', 'newest', 'wildest']\n",
" final_vocab, merges = byte_pair_encoding(corpus, num_merges=5)\n",
" assert isinstance(final_vocab, dict)\n",
" assert all(isinstance(pair, tuple) for pair in merges)\n",
" assert len(merges) == 5\n",
" print(\"✓✓✓✓\")\n",
"\n",
"# Run all tests\n",
"test_get_vocab()\n",
"test_get_\n"
"test_get_status()\n",
"test_merge_vocab()\n",
"test_bpe_sequence()"
]
}
],
Expand Down
2 changes: 1 addition & 1 deletion website/src/data/questions.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -441,7 +441,7 @@ const v2Questions: Omit<Question, "description" | "tracks" | "llmPathOrder" | "l
difficulty: "easy",
companies: [],
questionPath: "llm/Byte-Pair-Encoder/BPE-q3.ipynb",
solutionPath: "llm/Byte-Pair-Encoder/BPE-q3.ipynb",
solutionPath: "llm/Byte-Pair-Encoder/BPE-q3-SOLN.ipynb",
hasNotebook: true,
},
{
Expand Down