Skip to content

xgrammar

Backend class for XGrammar.

XGrammarBackend

Bases: BaseBackend

Backend for XGrammar.

Source code in outlines/backends/xgrammar.py 
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
class XGrammarBackend(BaseBackend):
    """Backend for XGrammar."""

    def __init__(self, model: SteerableModel):
        """
        Parameters
        ----------
        model
            The Outlines model of the user.

        """
        import xgrammar as xgr

        if isinstance(model, Transformers):
            tokenizer = model.hf_tokenizer
        elif isinstance(model, MLXLM): # pragma: no cover
            tokenizer = model.mlx_tokenizer._tokenizer
        else: # pragma: no cover
            raise ValueError(
                "The xgrammar backend only supports Transformers and "
                + "MLXLM models"
            )

        tokenizer_info = xgr.TokenizerInfo.from_huggingface(
            tokenizer,
            vocab_size=len(tokenizer.get_vocab())
        )
        self.grammar_compiler = xgr.GrammarCompiler(tokenizer_info)
        self.tensor_library_name = model.tensor_library_name

    def get_json_schema_logits_processor(
        self, json_schema: str
    ) -> XGrammarLogitsProcessor:
        """Create a logits processor from a JSON schema.

        Parameters
        ----------
        json_schema: str
            The JSON schema to create a logits processor from.

        Returns
        -------
        LogitsProcessor
            The logits processor to use to constrain the generation.

        """
        compiled_grammar = self.grammar_compiler.compile_json_schema(
            json_schema
        )
        return XGrammarLogitsProcessor(
            compiled_grammar,
            self.tensor_library_name
        )

    def get_regex_logits_processor(
        self, regex: str
    ) -> XGrammarLogitsProcessor:
        """Create a logits processor from a regex.

        Parameters
        ----------
        regex: str
            The regex to create a logits processor from.

        Returns
        -------
        LogitsProcessor
            The logits processor to use to constrain the generation.

        """
        compiled_grammar = self.grammar_compiler.compile_regex(regex)
        return XGrammarLogitsProcessor(
            compiled_grammar,
            self.tensor_library_name
        )

    def get_cfg_logits_processor(
        self, grammar: str
    ) -> XGrammarLogitsProcessor:
        """Create a logits processor from a context-free grammar.

        Parameters
        ----------
        grammar: str
            The context-free grammar to create a logits processor from.

        Returns
        -------
        LogitsProcessor
            The logits processor to use to constrain the generation.

        """
        compiled_grammar = self.grammar_compiler.compile_grammar(grammar)
        return XGrammarLogitsProcessor(
            compiled_grammar,
            self.tensor_library_name
        )

__init__(model)

Parameters:

Name Type Description Default
model SteerableModel

The Outlines model of the user.

 required
Source code in outlines/backends/xgrammar.py 
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
def __init__(self, model: SteerableModel):
    """
    Parameters
    ----------
    model
        The Outlines model of the user.

    """
    import xgrammar as xgr

    if isinstance(model, Transformers):
        tokenizer = model.hf_tokenizer
    elif isinstance(model, MLXLM): # pragma: no cover
        tokenizer = model.mlx_tokenizer._tokenizer
    else: # pragma: no cover
        raise ValueError(
            "The xgrammar backend only supports Transformers and "
            + "MLXLM models"
        )

    tokenizer_info = xgr.TokenizerInfo.from_huggingface(
        tokenizer,
        vocab_size=len(tokenizer.get_vocab())
    )
    self.grammar_compiler = xgr.GrammarCompiler(tokenizer_info)
    self.tensor_library_name = model.tensor_library_name

get_cfg_logits_processor(grammar)

Create a logits processor from a context-free grammar.

Parameters:

Name Type Description Default
grammar str

The context-free grammar to create a logits processor from.

 required

Returns:

Type Description
LogitsProcessor

The logits processor to use to constrain the generation.
Source code in outlines/backends/xgrammar.py 
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
def get_cfg_logits_processor(
    self, grammar: str
) -> XGrammarLogitsProcessor:
    """Create a logits processor from a context-free grammar.

    Parameters
    ----------
    grammar: str
        The context-free grammar to create a logits processor from.

    Returns
    -------
    LogitsProcessor
        The logits processor to use to constrain the generation.

    """
    compiled_grammar = self.grammar_compiler.compile_grammar(grammar)
    return XGrammarLogitsProcessor(
        compiled_grammar,
        self.tensor_library_name
    )

get_json_schema_logits_processor(json_schema)

Create a logits processor from a JSON schema.

Parameters:

Name Type Description Default
json_schema str

The JSON schema to create a logits processor from.

 required

Returns:

Type Description
LogitsProcessor

The logits processor to use to constrain the generation.
Source code in outlines/backends/xgrammar.py 
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
def get_json_schema_logits_processor(
    self, json_schema: str
) -> XGrammarLogitsProcessor:
    """Create a logits processor from a JSON schema.

    Parameters
    ----------
    json_schema: str
        The JSON schema to create a logits processor from.

    Returns
    -------
    LogitsProcessor
        The logits processor to use to constrain the generation.

    """
    compiled_grammar = self.grammar_compiler.compile_json_schema(
        json_schema
    )
    return XGrammarLogitsProcessor(
        compiled_grammar,
        self.tensor_library_name
    )

get_regex_logits_processor(regex)

Create a logits processor from a regex.

Parameters:

Name Type Description Default
regex str

The regex to create a logits processor from.

 required

Returns:

Type Description
LogitsProcessor

The logits processor to use to constrain the generation.
Source code in outlines/backends/xgrammar.py 
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
def get_regex_logits_processor(
    self, regex: str
) -> XGrammarLogitsProcessor:
    """Create a logits processor from a regex.

    Parameters
    ----------
    regex: str
        The regex to create a logits processor from.

    Returns
    -------
    LogitsProcessor
        The logits processor to use to constrain the generation.

    """
    compiled_grammar = self.grammar_compiler.compile_regex(regex)
    return XGrammarLogitsProcessor(
        compiled_grammar,
        self.tensor_library_name
    )

XGrammarLogitsProcessor

Bases: OutlinesLogitsProcessor

Logits processor for XGrammar.

Source code in outlines/backends/xgrammar.py 
 13
 14
 15
 16
 17
 18
 19
 20
 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
 49
 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
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
class XGrammarLogitsProcessor(OutlinesLogitsProcessor):
    """Logits processor for XGrammar."""

    def __init__(self, compiled_grammar: str, tensor_library_name: str,):
        """
        Parameters
        ----------
        compiled_grammar: str
            The compiled grammar to use to create the logits processor.
        tensor_library_name: str
            The name of the tensor library used by the model

        """
        import xgrammar as xgr

        self.xgr = xgr
        self.is_first_token = True
        self.compiled_grammar = compiled_grammar
        self.tensor_library_name = tensor_library_name
        super().__init__(tensor_library_name)

    def reset(self):
        """Ensure self._setup is called again for the next generation."""
        self.is_first_token = True

    def _setup(self, batch_size: int, vocab_size: int) -> None:
        """Setup the logits processor for a new generation."""
        if self.tensor_library_name == "torch":
            self._bias_logits = self._bias_logits_torch
        elif self.tensor_library_name == "mlx": # pragma: no cover
            self._bias_logits = self._bias_logits_mlx
        else: # pragma: no cover
            raise ValueError(
                f"Unsupported tensor library: {self.tensor_library_name}"
            )

        self._matchers = [
            self.xgr.GrammarMatcher(self.compiled_grammar)
            for _ in range(batch_size)
        ]
        self._bitmask = self.xgr.allocate_token_bitmask(batch_size, vocab_size)

    def _bias_logits_torch(
        self, input_ids: TensorType, logits: TensorType
    ) -> TensorType:
        """Bias the logits for Torch tensors."""
        for i in range(self.tensor_adapter.shape(input_ids)[0]):
            if not self._matchers[i].is_terminated():
                self._matchers[i].fill_next_token_bitmask(self._bitmask, i)

        self._bitmask = self.tensor_adapter.to_device(
            self._bitmask,
            self.tensor_adapter.get_device(logits)
        )
        self.xgr.apply_token_bitmask_inplace(logits, self._bitmask)
        self._bitmask = self.tensor_adapter.to_device(
            self._bitmask,
            "cpu"
        )

        return logits

    def _bias_logits_mlx( # pragma: no cover
        self, input_ids: TensorType, logits: TensorType
    ) -> TensorType:
        """Bias the logits for MLX tensors."""
        import mlx.core as mx
        from xgrammar.kernels.apply_token_bitmask_mlx import apply_token_bitmask_mlx

        for i in range(self.tensor_adapter.shape(input_ids)[0]):
            if not self._matchers[i].is_terminated():
                self._matchers[i].fill_next_token_bitmask(self._bitmask, i)

        biased_logits = apply_token_bitmask_mlx(
            mx.array(self._bitmask.numpy()), logits, self.tensor_adapter.shape(logits)[1]
        )

        return biased_logits

    def process_logits(
        self, input_ids: TensorType, logits: TensorType
    ) -> TensorType:
        """Use the XGrammar matchers to bias the logits."""
        batch_size = self.tensor_adapter.shape(input_ids)[0]
        vocab_size = self.tensor_adapter.shape(logits)[1]

        if self.is_first_token:
            self._setup(batch_size, vocab_size)
            self.is_first_token = False
        else:
            for i in range(batch_size):
                if not self._matchers[i].is_terminated(): # pragma: no cover
                    last_token_id = self.tensor_adapter.to_scalar(
                        input_ids[i][-1] # type: ignore
                    )
                    assert self._matchers[i].accept_token(last_token_id)

        return self._bias_logits(input_ids, logits)

__init__(compiled_grammar, tensor_library_name)

Parameters:

Name Type Description Default
compiled_grammar str

The compiled grammar to use to create the logits processor.

 required
tensor_library_name str

The name of the tensor library used by the model

 required
Source code in outlines/backends/xgrammar.py 
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
def __init__(self, compiled_grammar: str, tensor_library_name: str,):
    """
    Parameters
    ----------
    compiled_grammar: str
        The compiled grammar to use to create the logits processor.
    tensor_library_name: str
        The name of the tensor library used by the model

    """
    import xgrammar as xgr

    self.xgr = xgr
    self.is_first_token = True
    self.compiled_grammar = compiled_grammar
    self.tensor_library_name = tensor_library_name
    super().__init__(tensor_library_name)

process_logits(input_ids, logits)

Use the XGrammar matchers to bias the logits.

Source code in outlines/backends/xgrammar.py 
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
def process_logits(
    self, input_ids: TensorType, logits: TensorType
) -> TensorType:
    """Use the XGrammar matchers to bias the logits."""
    batch_size = self.tensor_adapter.shape(input_ids)[0]
    vocab_size = self.tensor_adapter.shape(logits)[1]

    if self.is_first_token:
        self._setup(batch_size, vocab_size)
        self.is_first_token = False
    else:
        for i in range(batch_size):
            if not self._matchers[i].is_terminated(): # pragma: no cover
                last_token_id = self.tensor_adapter.to_scalar(
                    input_ids[i][-1] # type: ignore
                )
                assert self._matchers[i].accept_token(last_token_id)

    return self._bias_logits(input_ids, logits)

reset()

Ensure self._setup is called again for the next generation.

Source code in outlines/backends/xgrammar.py 
34
35
36
def reset(self):
    """Ensure self._setup is called again for the next generation."""
    self.is_first_token = True