mirror of
https://gitlab.futo.org/keyboard/latinime.git
synced 2024-09-28 14:54:30 +01:00
Enable Quadgram for personalized dicts.
Before: Total words: 1134659, Success Num: 944709, Success Percentage: 83.259% Bad Failures, with auto-correction (typed word == expected word, output word != expected word): 1258, Bad Failure Percentage: 0.111% Failures, with auto-correction (F-C): 28013, F-C Percentage: 2.469% Max Keystrokes: 6072844, Min Keystrokes: 3347332, Keystroke Saving Percentage:44.880% After: Total words: 1134665, Success Num: 945026, Success Percentage: 83.287% Bad Failures, with auto-correction (typed word == expected word, output word != expected word): 1271, Bad Failure Percentage: 0.112% Failures, with auto-correction (F-C): 27756, F-C Percentage: 2.446% Max Keystrokes: 6072850, Min Keystrokes: 3290996, Keystroke Saving Percentage:45.808% Change-Id: I16af52a3e9c371b95fd6f0741f45ee6b2443bea6
This commit is contained in:
parent
78212a6d3d
commit
60021bbdc2
@ -179,7 +179,7 @@ public final class Constants {
|
||||
|
||||
// (MAX_PREV_WORD_COUNT_FOR_N_GRAM + 1)-gram is supported in Java side. Needs to modify
|
||||
// MAX_PREV_WORD_COUNT_FOR_N_GRAM in native/jni/src/defines.h for suggestions.
|
||||
public static final int MAX_PREV_WORD_COUNT_FOR_N_GRAM = 2;
|
||||
public static final int MAX_PREV_WORD_COUNT_FOR_N_GRAM = 3;
|
||||
|
||||
// Key events coming any faster than this are long-presses.
|
||||
public static final int LONG_PRESS_MILLISECONDS = 200;
|
||||
|
@ -275,7 +275,7 @@ static inline void showStackTrace() {
|
||||
#define MAX_POINTER_COUNT_G 2
|
||||
|
||||
// (MAX_PREV_WORD_COUNT_FOR_N_GRAM + 1)-gram is supported.
|
||||
#define MAX_PREV_WORD_COUNT_FOR_N_GRAM 2
|
||||
#define MAX_PREV_WORD_COUNT_FOR_N_GRAM 3
|
||||
|
||||
#define DISALLOW_DEFAULT_CONSTRUCTOR(TypeName) \
|
||||
TypeName() = delete
|
||||
|
@ -31,10 +31,11 @@ const char *const HeaderPolicy::IS_DECAYING_DICT_KEY = "USES_FORGETTING_CURVE";
|
||||
const char *const HeaderPolicy::DATE_KEY = "date";
|
||||
const char *const HeaderPolicy::LAST_DECAYED_TIME_KEY = "LAST_DECAYED_TIME";
|
||||
const char *const HeaderPolicy::NGRAM_COUNT_KEYS[] =
|
||||
{"UNIGRAM_COUNT", "BIGRAM_COUNT", "TRIGRAM_COUNT"};
|
||||
{"UNIGRAM_COUNT", "BIGRAM_COUNT", "TRIGRAM_COUNT", "QUADGRAM_COUNT"};
|
||||
const char *const HeaderPolicy::MAX_NGRAM_COUNT_KEYS[] =
|
||||
{"MAX_UNIGRAM_ENTRY_COUNT", "MAX_BIGRAM_ENTRY_COUNT", "MAX_TRIGRAM_ENTRY_COUNT"};
|
||||
const int HeaderPolicy::DEFAULT_MAX_NGRAM_COUNTS[] = {10000, 30000, 30000};
|
||||
{"MAX_UNIGRAM_ENTRY_COUNT", "MAX_BIGRAM_ENTRY_COUNT", "MAX_TRIGRAM_ENTRY_COUNT",
|
||||
"MAX_QUADGRAM_ENTRY_COUNT"};
|
||||
const int HeaderPolicy::DEFAULT_MAX_NGRAM_COUNTS[] = {10000, 30000, 30000, 30000};
|
||||
const char *const HeaderPolicy::EXTENDED_REGION_SIZE_KEY = "EXTENDED_REGION_SIZE";
|
||||
// Historical info is information that is needed to support decaying such as timestamp, level and
|
||||
// count.
|
||||
|
@ -19,12 +19,13 @@
|
||||
namespace latinime {
|
||||
|
||||
// Used to provide stable probabilities even if the user's input count is small.
|
||||
const int DynamicLanguageModelProbabilityUtils::ASSUMED_MIN_COUNTS[] = {8192, 2, 2};
|
||||
const int DynamicLanguageModelProbabilityUtils::ASSUMED_MIN_COUNTS[] = {8192, 2, 2, 1};
|
||||
|
||||
// Encoded backoff weights.
|
||||
// Note that we give positive value for trigrams that means the weight is more than 1.
|
||||
// Note that we give positive values for trigrams and quadgrams that means the weight is more than
|
||||
// 1.
|
||||
// TODO: Apply backoff for main dictionaries and quit giving a positive backoff weight.
|
||||
const int DynamicLanguageModelProbabilityUtils::ENCODED_BACKOFF_WEIGHTS[] = {-32, 0, 8};
|
||||
const int DynamicLanguageModelProbabilityUtils::ENCODED_BACKOFF_WEIGHTS[] = {-32, -4, 2, 8};
|
||||
|
||||
// This value is used to remove too old entries from the dictionary.
|
||||
const int DynamicLanguageModelProbabilityUtils::DURATION_TO_DISCARD_ENTRY_IN_SECONDS =
|
||||
|
@ -66,7 +66,7 @@ class DynamicLanguageModelProbabilityUtils {
|
||||
private:
|
||||
DISALLOW_IMPLICIT_CONSTRUCTORS(DynamicLanguageModelProbabilityUtils);
|
||||
|
||||
static_assert(MAX_PREV_WORD_COUNT_FOR_N_GRAM <= 2, "Max supported Ngram is Trigram.");
|
||||
static_assert(MAX_PREV_WORD_COUNT_FOR_N_GRAM <= 3, "Max supported Ngram is Quadgram.");
|
||||
|
||||
static const int ASSUMED_MIN_COUNTS[];
|
||||
static const int ENCODED_BACKOFF_WEIGHTS[];
|
||||
|
@ -27,7 +27,7 @@ namespace latinime {
|
||||
// Copyable but immutable
|
||||
class EntryCounts final {
|
||||
public:
|
||||
EntryCounts() : mEntryCounts({{0, 0, 0}}) {}
|
||||
EntryCounts() : mEntryCounts({{0, 0, 0, 0}}) {}
|
||||
|
||||
explicit EntryCounts(const std::array<int, MAX_PREV_WORD_COUNT_FOR_N_GRAM + 1> &counters)
|
||||
: mEntryCounts(counters) {}
|
||||
|
@ -25,6 +25,7 @@ enum class NgramType : int {
|
||||
Unigram = 0,
|
||||
Bigram = 1,
|
||||
Trigram = 2,
|
||||
Quadgram = 3,
|
||||
NotANgramType = -1,
|
||||
};
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user