tests : use Python to generate tokenizer tests for C++

This commit is contained in:
Georgi Gerganov 2023-08-26 18:05:59 +03:00
parent dfa058ef73
commit 70005bd5c9
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
2 changed files with 46 additions and 35 deletions

View file

@ -6,42 +6,34 @@
#include <map> #include <map>
#include <vector> #include <vector>
// generate using test-tokenizer-0.py
static const std::map<std::string, std::vector<llama_token>> & k_tests() { static const std::map<std::string, std::vector<llama_token>> & k_tests() {
static std::map<std::string, std::vector<llama_token>> _k_tests = { static std::map<std::string, std::vector<llama_token>> _k_tests = {
{ "" , { }, }, { "" , { }, },
{ " ", { 259, }, }, { " " , { 259, }, },
{ " ", { 1678, }, }, { " " , { 1678, }, },
{ " ", { 268, }, }, { " " , { 268, }, },
{ "\t", { 29871, 12, }, }, { "\t" , { 29871, 12, }, },
{ "\n", { 29871, 13, }, }, { "\n" , { 29871, 13, }, },
{ "\t\n", { 29871, 12, 13, }, }, { "\t\n" , { 29871, 12, 13, }, },
{ "Hello world", { 15043, 3186, }, }, { "Hello world" , { 15043, 3186, }, },
{ " Hello world", { 29871, 15043, 3186, }, }, { " Hello world" , { 29871, 15043, 3186, }, },
{ "Hello World", { 15043, 2787, }, }, { "Hello World" , { 15043, 2787, }, },
{ " Hello World", { 29871, 15043, 2787, }, }, { " Hello World" , { 29871, 15043, 2787, }, },
{ " Hello World!", { 29871, 15043, 2787, 29991, }, }, { " Hello World!" , { 29871, 15043, 2787, 29991, }, },
{ "Hello, world!", { 15043, 29892, 3186, 29991, }, }, { "Hello, world!" , { 15043, 29892, 3186, 29991, }, },
{ " Hello, world!", { 29871, 15043, 29892, 3186, 29991, }, }, { " Hello, world!" , { 29871, 15043, 29892, 3186, 29991, }, },
{ " this is 🦙.cpp", { 29871, 445, 338, 29871, 243, 162, 169, 156, 29889, 8223, }, }, { " this is 🦙.cpp" , { 29871, 445, 338, 29871, 243, 162, 169, 156, 29889, 8223, }, },
{ "w048 7tuijk dsdfhu", { 281, 29900, 29946, 29947, 29871, 29955, 9161, 13535, 18031, 2176, 6905, }, }, { "w048 7tuijk dsdfhu" , { 281, 29900, 29946, 29947, 29871, 29955, 9161, 13535, 18031, 2176, 6905, }, },
{ "нещо на Български", { 1538, 4851, 665, 1386, 29713, 1305, }, }, { "нещо на Български" , { 1538, 4851, 665, 1386, 29713, 1305, }, },
{ "កាន់តែពិសេសអាចខលចេញ", { "កាន់តែពិសេសអាចខលចេញ" , { 29871, 31849, 31324, 31934, 228, 162, 142, 228, 161, 146, 228, 162, 133, 228, 161, 153, 228, 161, 186, 31708, 228, 162, 132, 31708, 228, 161, 165, 31324, 228, 161, 136, 228, 161, 132, 228, 161, 158, 228, 161, 136, 228, 162, 132, 228, 161, 140, }, },
{ 29871, 31849, 31324, 31934, 228, 162, 142, 228, 161, { "🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)", { 29871, 243, 162, 157, 131, 313, 8945, 29897, 29871, 243, 162, 155, 185, 30722, 243, 162, 143, 174, 30598, 313, 20787, 953, 3848, 275, 16125, 630, 29897, 29871, 31681, 313, 6194, 953, 29877, 2397, 393, 756, 967, 1914, 5993, 29897, }, },
146, 228, 162, 133, 228, 161, 153, 228, 161, 186, { "Hello" , { 15043, }, },
31708, 228, 162, 132, 31708, 228, 161, 165, 31324, 228, { " Hello" , { 29871, 15043, }, },
161, 136, 228, 161, 132, 228, 161, 158, 228, 161, { " Hello" , { 259, 15043, }, },
136, 228, 162, 132, 228, 161, 140, }, }, { " Hello" , { 1678, 15043, }, },
{ "🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)", { " Hello" , { 268, 15043, }, },
{ 29871, 243, 162, 157, 131, 313, 8945, 29897, 29871, { " Hello\n Hello" , { 268, 15043, 13, 1678, 15043, }, },
243, 162, 155, 185, 30722, 243, 162, 143, 174, 30598,
313, 20787, 953, 3848, 275, 16125, 630, 29897, 29871, 31681,
313, 6194, 953, 29877, 2397, 393, 756, 967, 1914, 5993, 29897, }, },
{ "Hello", { 15043, }, },
{ " Hello", { 29871, 15043, }, },
{ " Hello", { 259, 15043, }, },
{ " Hello", { 1678, 15043, }, },
{ " Hello", { 268, 15043, }, },
{ " Hello\n Hello", { 268, 15043, 13, 1678, 15043, }, },
}; };
return _k_tests; return _k_tests;
@ -99,7 +91,14 @@ int main(int argc, char **argv) {
const std::vector<llama_token> res_bos = llama_tokenize(ctx, test_kv.first, true); const std::vector<llama_token> res_bos = llama_tokenize(ctx, test_kv.first, true);
const std::vector<llama_token> res_nobos = llama_tokenize(ctx, test_kv.first, false); const std::vector<llama_token> res_nobos = llama_tokenize(ctx, test_kv.first, false);
fprintf(stderr, "%s : '%s' tokenized to '%s'\n", __func__, test_kv.first.c_str(), llama_detokenize(ctx, res_bos).c_str()); printf("\n");
printf("src: '%s'\n", test_kv.first.c_str());
printf("res: '%s'\n", llama_detokenize(ctx, res_bos).c_str());
printf("tok: ");
for (const auto & tok : res_bos) {
printf("%d ", tok);
}
printf("\n");
bool correct = res_nobos.size() == test_kv.second.size() && res_bos.size() == res_nobos.size() + 1 && res_bos[0] == 1; bool correct = res_nobos.size() == test_kv.second.size() && res_bos.size() == res_nobos.size() + 1 && res_bos[0] == 1;

View file

@ -56,3 +56,15 @@ print("'" + tokenizer.decode([15043]) + "'") # 'Hello'
print("'" + tokenizer.decode([15043, 15043]) + "'") # 'Hello Hello' print("'" + tokenizer.decode([15043, 15043]) + "'") # 'Hello Hello'
print("'" + tokenizer.decode([29871, 15043]) + "'") # ' Hello' print("'" + tokenizer.decode([29871, 15043]) + "'") # ' Hello'
print("'" + tokenizer.decode([29871, 15043, 29871, 15043]) + "'") # ' Hello Hello' print("'" + tokenizer.decode([29871, 15043, 29871, 15043]) + "'") # ' Hello Hello'
print("\n\ntests for C++:\n")
for text in tests:
res = tokenizer.encode(text, add_bos=False)
k = text.replace('\n', '\\n')
k = k.replace('\t', '\\t')
k = '"' + k + '"'
print("{ %-24s, { " % k, end='')
for x in res:
print("%7d," % x, end='')
print(" }, },")