process_text_spm.go 6.07 KB
Newer Older
Patrick Devine's avatar
Patrick Devine committed
1
2
3
package model

import (
4
5
	"container/heap"
	"fmt"
Patrick Devine's avatar
Patrick Devine committed
6
	"log/slog"
7
	"strconv"
Patrick Devine's avatar
Patrick Devine committed
8
9
10
11
12
13
14
15
16
17
	"strings"
)

const spmWhitespaceSep = "▁"

type SentencePieceModel struct {
	maxTokenLen int
	vocab       *Vocabulary
}

Michael Yang's avatar
Michael Yang committed
18
19
var _ TextProcessor = (*SentencePieceModel)(nil)

20
func NewSentencePieceModel(vocab *Vocabulary) SentencePieceModel {
Patrick Devine's avatar
Patrick Devine committed
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
	slog.Debug("Tokens", "num tokens", len(vocab.Values), "vals", vocab.Values[:5], "scores", vocab.Scores[:5], "types", vocab.Types[:5])

	counter := map[int]int{}
	var maxTokenLen int
	for cnt := range vocab.Types {
		switch vocab.Types[cnt] {
		case TOKEN_TYPE_NORMAL, TOKEN_TYPE_USER_DEFINED, TOKEN_TYPE_UNUSED:
			maxTokenLen = max(maxTokenLen, len(vocab.Values[cnt]))
			fallthrough
		default:
			counter[int(vocab.Types[cnt])] += 1
		}
	}

	slog.Debug("Token counts", "normal", counter[TOKEN_TYPE_NORMAL], "unknown", counter[TOKEN_TYPE_UNKNOWN], "control", counter[TOKEN_TYPE_CONTROL],
		"user defined", counter[TOKEN_TYPE_USER_DEFINED], "unused", counter[TOKEN_TYPE_UNUSED], "byte", counter[TOKEN_TYPE_BYTE],
		"max token len", maxTokenLen)

	return SentencePieceModel{
		maxTokenLen: maxTokenLen,
		vocab:       vocab,
	}
}

func (spm SentencePieceModel) Is(id int32, special Special) bool {
	return spm.vocab.Is(id, special)
}

Michael Yang's avatar
Michael Yang committed
49
func (spm SentencePieceModel) Encode(s string, addSpecial bool) ([]int32, error) {
Patrick Devine's avatar
Patrick Devine committed
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
	fragments := []fragment{{value: s}}
	for _, special := range spm.vocab.SpecialVocabulary() {
		id := spm.vocab.Encode(special)
		for i := 0; i < len(fragments); i++ {
			frag := fragments[i]
			if len(frag.ids) > 0 {
				continue
			}

			var middle []fragment
			switch i := strings.Index(frag.value, special); {
			case i < 0:
				middle = append(middle, frag)
			case i > 0:
				middle = append(middle, fragment{value: frag.value[:i]})
				fallthrough
			default:
				middle = append(middle, fragment{value: special, ids: []int32{id}})
				if rest := frag.value[i+len(special):]; rest != "" {
					middle = append(middle, fragment{value: rest})
				}
			}

			fragments = append(fragments[:i], append(middle, fragments[i+1:]...)...)
		}
	}

	var ids []int32
	for _, frag := range fragments {
		if len(frag.ids) > 0 {
			ids = append(ids, frag.ids...)
			continue
		}

84
		text := strings.ReplaceAll(frag.value, " ", spmWhitespaceSep)
Patrick Devine's avatar
Patrick Devine committed
85

86
87
88
89
		if id := spm.vocab.Encode(text); id >= 0 {
			ids = append(ids, id)
			continue
		}
Patrick Devine's avatar
Patrick Devine committed
90

91
92
		q := &queue{}
		heap.Init(q)
Patrick Devine's avatar
Patrick Devine committed
93

94
95
96
97
98
99
100
101
102
		runes := []rune(text)
		merges := make([]merge, len(runes))
		for r := range runes {
			merges[r] = merge{
				p:     r - 1,
				n:     r + 1,
				runes: []rune{runes[r]},
			}
		}
Patrick Devine's avatar
Patrick Devine committed
103

104
105
		pairwise := func(a, b int) *candidate {
			if a < 0 || b >= len(runes) {
Patrick Devine's avatar
Patrick Devine committed
106
107
108
				return nil
			}

109
110
111
112
113
114
115
			left, right := string(merges[a].runes), string(merges[b].runes)
			if id := spm.vocab.Encode(left + right); id >= 0 {
				return &candidate{
					a:     a,
					b:     b,
					score: spm.vocab.Scores[id],
					size:  len(left) + len(right),
Patrick Devine's avatar
Patrick Devine committed
116
117
118
				}
			}

119
120
121
122
123
124
			return nil
		}

		for i := range len(runes) - 1 {
			if pair := pairwise(i, i+1); pair != nil {
				heap.Push(q, pair)
Patrick Devine's avatar
Patrick Devine committed
125
			}
126
		}
Patrick Devine's avatar
Patrick Devine committed
127

128
129
130
		for q.Len() > 0 {
			pair := heap.Pop(q).(*candidate)
			left, right := merges[pair.a], merges[pair.b]
Patrick Devine's avatar
Patrick Devine committed
131

132
133
134
			if string(left.runes) == "" || string(right.runes) == "" || len(string(left.runes))+len(string(right.runes)) != pair.size {
				continue
			}
135

136
137
138
139
140
141
			merges[pair.a].runes = append(left.runes, right.runes...)
			merges[pair.b].runes = nil
			merges[pair.a].n = right.n
			if right.n < len(merges) {
				merges[right.n].p = pair.a
			}
Patrick Devine's avatar
Patrick Devine committed
142

143
144
145
			if pair := pairwise(merges[pair.a].p, pair.a); pair != nil {
				heap.Push(q, pair)
			}
Patrick Devine's avatar
Patrick Devine committed
146

147
148
			if pair := pairwise(pair.a, merges[pair.a].n); pair != nil {
				heap.Push(q, pair)
Patrick Devine's avatar
Patrick Devine committed
149
			}
150
		}
Patrick Devine's avatar
Patrick Devine committed
151

152
153
154
		for _, merge := range merges {
			if token := string(merge.runes); token != "" {
				id := spm.vocab.Encode(token)
Patrick Devine's avatar
Patrick Devine committed
155

156
157
158
159
160
161
162
163
164
165
166
167
				if id >= 0 {
					ids = append(ids, id)
					continue
				}

				// Fallback to byte tokenization
				var result []int32
				for _, b := range []byte(token) {
					byteToken := fmt.Sprintf("<0x%02X>", b)
					unknownID := spm.vocab.Encode(byteToken)
					if unknownID >= 0 {
						result = append(result, unknownID)
Patrick Devine's avatar
Patrick Devine committed
168
					} else {
169
						slog.Debug("unknown byte token", "byte", b, "token", byteToken)
Patrick Devine's avatar
Patrick Devine committed
170
171
					}
				}
172
173

				ids = append(ids, result...)
Patrick Devine's avatar
Patrick Devine committed
174
175
176
			}
		}
	}
Michael Yang's avatar
Michael Yang committed
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196

	if addSpecial && len(ids) > 0 {
		if spm.vocab.AddBOS {
			if ids[0] == spm.vocab.BOS {
				slog.Warn("adding bos token to prompt which already has it", "id", spm.vocab.BOS)
			}

			slog.Debug("adding bos token to prompt", "id", spm.vocab.BOS)
			ids = append([]int32{spm.vocab.BOS}, ids...)
		}

		if spm.vocab.AddEOS {
			if ids[len(ids)-1] == spm.vocab.EOS {
				slog.Warn("adding eos token to prompt which already has it", "id", spm.vocab.EOS)
			}

			slog.Debug("adding eos token to prompt", "id", spm.vocab.EOS)
			ids = append(ids, spm.vocab.EOS)
		}
	}
Patrick Devine's avatar
Patrick Devine committed
197
198
199
200
201
202
203

	return ids, nil
}

type candidate struct {
	a, b  int
	score float32
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
	size  int
}

type queue []*candidate

func (q queue) Len() int { return len(q) }

func (q queue) Less(i, j int) bool {
	return (q[i].score > q[j].score) || (q[i].score == q[j].score && q[i].a < q[j].a)
}

func (q queue) Swap(i, j int) { q[i], q[j] = q[j], q[i] }

func (q *queue) Push(x interface{}) {
	item := x.(*candidate)
	*q = append(*q, item)
}

func (q *queue) Pop() interface{} {
	old := *q
	n := len(old)
	item := old[n-1]
	*q = old[0 : n-1]
	return item
Patrick Devine's avatar
Patrick Devine committed
228
229
230
231
232
233
234
}

func (spm SentencePieceModel) Decode(ids []int32) (string, error) {
	var sb strings.Builder
	for _, id := range ids {
		data := spm.vocab.Decode(id)
		data = strings.ReplaceAll(data, spmWhitespaceSep, " ")
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252

		// For tokenizers that use byte tokens like "<0xEA>"
		// convert them to the partial unicode character
		// so they are buffered correctly by the runner instead
		// of being sent back to the api as "<0xEA>"
		if len(data) == 6 && strings.HasPrefix(data, "<0x") && strings.HasSuffix(data, ">") {
			byteVal, err := strconv.ParseUint(data[1:5], 0, 8)
			if err != nil {
				return "", fmt.Errorf("failed to parse hex byte: %v", err)
			}

			if err := sb.WriteByte(byte(byteVal)); err != nil {
				return "", err
			}
		} else {
			if _, err := sb.WriteString(data); err != nil {
				return "", err
			}
Patrick Devine's avatar
Patrick Devine committed
253
254
255
256
257
		}
	}

	return sb.String(), nil
}