tokenizer_spm.go 2.34 KB
Newer Older
Michael Yang's avatar
Michael Yang committed
1
2
3
4
5
6
7
package convert

import (
	"cmp"
	"encoding/json"
	"errors"
	"fmt"
8
	"io/fs"
Michael Yang's avatar
Michael Yang committed
9
10
11
12
13
14
15
16
	"os"
	"slices"

	"google.golang.org/protobuf/proto"

	"github.com/ollama/ollama/convert/sentencepiece"
)

17
func parseSentencePiece(fsys fs.FS) (*Vocabulary, error) {
Michael Yang's avatar
Michael Yang committed
18
19
20
21
22
	ast, err := parseAdditionalSpecialTokens(fsys)
	if err != nil {
		return nil, err
	}

23
	bts, err := fs.ReadFile(fsys, "tokenizer.model")
Michael Yang's avatar
Michael Yang committed
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
	if err != nil {
		return nil, err
	}

	var spm sentencepiece.ModelProto
	if err := proto.Unmarshal(bts, &spm); err != nil {
		return nil, err
	}

	v := Vocabulary{Model: "llama"}
	for _, piece := range spm.GetPieces() {
		v.Tokens = append(v.Tokens, piece.GetPiece())
		v.Scores = append(v.Scores, piece.GetScore())

		switch t := piece.GetType(); t {
		case sentencepiece.ModelProto_SentencePiece_UNKNOWN,
			sentencepiece.ModelProto_SentencePiece_CONTROL,
			sentencepiece.ModelProto_SentencePiece_UNUSED,
			sentencepiece.ModelProto_SentencePiece_BYTE:
			v.Types = append(v.Types, int32(t))
		default:
Michael Yang's avatar
Michael Yang committed
45
46
47
48
49
50
			tt := int32(sentencepiece.ModelProto_SentencePiece_NORMAL)
			if slices.Contains(ast, piece.GetPiece()) {
				tt = int32(sentencepiece.ModelProto_SentencePiece_CONTROL)
			}

			v.Types = append(v.Types, tt)
Michael Yang's avatar
Michael Yang committed
51
52
53
		}
	}

54
	f, err := fsys.Open("added_tokens.json")
Michael Yang's avatar
Michael Yang committed
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
84
85
86
87
88
89
90
91
92
93
	if errors.Is(err, os.ErrNotExist) {
		return &v, nil
	} else if err != nil {
		return nil, err
	}
	defer f.Close()

	var atm map[string]int
	if err := json.NewDecoder(f).Decode(&atm); err != nil {
		return nil, err
	}

	type t struct {
		id      int
		content string
	}

	var ts []t
	for content, id := range atm {
		ts = append(ts, t{id, content})
	}

	slices.SortFunc(ts, func(i, j t) int {
		return cmp.Compare(i.id, j.id)
	})

	n := len(v.Tokens)
	for i, t := range ts {
		if t.id != i+n {
			return nil, fmt.Errorf("invalid token id: %d", t.id)
		}

		v.Tokens = append(v.Tokens, t.content)
		v.Scores = append(v.Scores, -1000.0)
		v.Types = append(v.Types, tokenTypeUserDefined)
	}

	return &v, nil
}
Michael Yang's avatar
Michael Yang committed
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113

func parseAdditionalSpecialTokens(fsys fs.FS) ([]string, error) {
	f, err := fsys.Open("special_tokens_map.json")
	if errors.Is(err, os.ErrNotExist) {
		return nil, nil
	} else if err != nil {
		return nil, err
	}
	defer f.Close()

	var m struct {
		AdditionalSpecialTokens []string `json:"additional_special_tokens"`
	}

	if err := json.NewDecoder(f).Decode(&m); err != nil {
		return nil, err
	}

	return m.AdditionalSpecialTokens, nil
}