Skip to content

vllm.v1.sample.ops.topk_topp_sampler

is_flashinfer_available module-attribute 

is_flashinfer_available = True

logger module-attribute 

logger = init_logger(__name__)

TopKTopPSampler

Bases: Module

Module that performs optional top-k and top-p filtering followed by weighted random sampling of logits.

Implementations may update the logits tensor in-place.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
 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
111
112
113
114
115
class TopKTopPSampler(nn.Module):
    """
    Module that performs optional top-k and top-p filtering followed by
    weighted random sampling of logits.

    Implementations may update the logits tensor in-place.
    """

    def __init__(self):
        super().__init__()
        if current_platform.is_cuda():
            if is_flashinfer_available:
                flashinfer_version = flashinfer.__version__
                if flashinfer_version < "0.2.3":
                    logger.warning(
                        "FlashInfer version >= 0.2.3 required. "
                        "Falling back to default sampling implementation.")
                    self.forward = self.forward_native
                elif envs.VLLM_USE_FLASHINFER_SAMPLER is not False:
                    # NOTE(woosuk): The V0 sampler doesn't use FlashInfer for
                    # sampling unless VLLM_USE_FLASHINFER_SAMPLER=1 (i.e., by
                    # default it is unused). For backward compatibility, we set
                    # `VLLM_USE_FLASHINFER_SAMPLER` as None by default and
                    # interpret it differently in V0 and V1 samplers: In V0,
                    # None means False, while in V1, None means True. This is
                    # why we use the condition
                    # `envs.VLLM_USE_FLASHINFER_SAMPLER is not False` here.
                    logger.info("Using FlashInfer for top-p & top-k sampling.")
                    self.forward = self.forward_cuda
                else:
                    logger.warning(
                        "FlashInfer is available, but it is not enabled. "
                        "Falling back to the PyTorch-native implementation of "
                        "top-p & top-k sampling. For the best performance, "
                        "please set VLLM_USE_FLASHINFER_SAMPLER=1.")
                    self.forward = self.forward_native
            else:
                logger.warning(
                    "FlashInfer is not available. Falling back to the PyTorch-"
                    "native implementation of top-p & top-k sampling. For the "
                    "best performance, please install FlashInfer.")
                self.forward = self.forward_native
        elif current_platform.is_tpu():
            self.forward = self.forward_tpu
        else:
            self.forward = self.forward_native

    def forward_native(
        self,
        logits: torch.Tensor,
        generators: dict[int, torch.Generator],
        k: Optional[torch.Tensor],
        p: Optional[torch.Tensor],
    ) -> torch.Tensor:
        """
        PyTorch-native implementation of top-k and top-p sampling.

        The logits tensor may be updated in-place.
        """
        logits = apply_top_k_top_p(logits, k, p)
        probs = logits.softmax(dim=-1, dtype=torch.float32)
        return random_sample(probs, generators)

    def forward_cuda(
        self,
        logits: torch.Tensor,
        generators: dict[int, torch.Generator],
        k: Optional[torch.Tensor],
        p: Optional[torch.Tensor],
    ) -> torch.Tensor:
        """More optimized implementation for top-k and top-p sampling."""
        if k is None and p is None:
            # We prefer `random_sample` over `flashinfer_sample` when sorting is
            # not needed. This is because `random_sample` does not require
            # CPU-GPU synchronization while `flashinfer_sample` does.
            probs = logits.softmax(dim=-1, dtype=torch.float32)
            return random_sample(probs, generators)
        if generators:
            logger.warning("FlashInfer 0.2.3+ does not support "
                           "per-request generators. Falling back to "
                           "PyTorch-native implementation.")
            return self.forward_native(logits, generators, k, p)
        return flashinfer_sample(logits, k, p, generators)

    def forward_tpu(
        self,
        logits: torch.Tensor,
        generators: dict[int, torch.Generator],
        k: Optional[torch.Tensor],
        p: Optional[torch.Tensor],
    ) -> torch.Tensor:
        logits = apply_top_k_top_p_tpu(logits, k, p)
        probs = logits.softmax(dim=-1, dtype=torch.float32)
        return random_sample(probs, generators)

forward instance-attribute 

forward = forward_native

__init__ 

__init__()
Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
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
def __init__(self):
    super().__init__()
    if current_platform.is_cuda():
        if is_flashinfer_available:
            flashinfer_version = flashinfer.__version__
            if flashinfer_version < "0.2.3":
                logger.warning(
                    "FlashInfer version >= 0.2.3 required. "
                    "Falling back to default sampling implementation.")
                self.forward = self.forward_native
            elif envs.VLLM_USE_FLASHINFER_SAMPLER is not False:
                # NOTE(woosuk): The V0 sampler doesn't use FlashInfer for
                # sampling unless VLLM_USE_FLASHINFER_SAMPLER=1 (i.e., by
                # default it is unused). For backward compatibility, we set
                # `VLLM_USE_FLASHINFER_SAMPLER` as None by default and
                # interpret it differently in V0 and V1 samplers: In V0,
                # None means False, while in V1, None means True. This is
                # why we use the condition
                # `envs.VLLM_USE_FLASHINFER_SAMPLER is not False` here.
                logger.info("Using FlashInfer for top-p & top-k sampling.")
                self.forward = self.forward_cuda
            else:
                logger.warning(
                    "FlashInfer is available, but it is not enabled. "
                    "Falling back to the PyTorch-native implementation of "
                    "top-p & top-k sampling. For the best performance, "
                    "please set VLLM_USE_FLASHINFER_SAMPLER=1.")
                self.forward = self.forward_native
        else:
            logger.warning(
                "FlashInfer is not available. Falling back to the PyTorch-"
                "native implementation of top-p & top-k sampling. For the "
                "best performance, please install FlashInfer.")
            self.forward = self.forward_native
    elif current_platform.is_tpu():
        self.forward = self.forward_tpu
    else:
        self.forward = self.forward_native

forward_cuda 

forward_cuda(
    logits: Tensor,
    generators: dict[int, Generator],
    k: Optional[Tensor],
    p: Optional[Tensor],
) -> Tensor

More optimized implementation for top-k and top-p sampling.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
def forward_cuda(
    self,
    logits: torch.Tensor,
    generators: dict[int, torch.Generator],
    k: Optional[torch.Tensor],
    p: Optional[torch.Tensor],
) -> torch.Tensor:
    """More optimized implementation for top-k and top-p sampling."""
    if k is None and p is None:
        # We prefer `random_sample` over `flashinfer_sample` when sorting is
        # not needed. This is because `random_sample` does not require
        # CPU-GPU synchronization while `flashinfer_sample` does.
        probs = logits.softmax(dim=-1, dtype=torch.float32)
        return random_sample(probs, generators)
    if generators:
        logger.warning("FlashInfer 0.2.3+ does not support "
                       "per-request generators. Falling back to "
                       "PyTorch-native implementation.")
        return self.forward_native(logits, generators, k, p)
    return flashinfer_sample(logits, k, p, generators)

forward_native 

forward_native(
    logits: Tensor,
    generators: dict[int, Generator],
    k: Optional[Tensor],
    p: Optional[Tensor],
) -> Tensor

PyTorch-native implementation of top-k and top-p sampling.

The logits tensor may be updated in-place.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
def forward_native(
    self,
    logits: torch.Tensor,
    generators: dict[int, torch.Generator],
    k: Optional[torch.Tensor],
    p: Optional[torch.Tensor],
) -> torch.Tensor:
    """
    PyTorch-native implementation of top-k and top-p sampling.

    The logits tensor may be updated in-place.
    """
    logits = apply_top_k_top_p(logits, k, p)
    probs = logits.softmax(dim=-1, dtype=torch.float32)
    return random_sample(probs, generators)

forward_tpu 

forward_tpu(
    logits: Tensor,
    generators: dict[int, Generator],
    k: Optional[Tensor],
    p: Optional[Tensor],
) -> Tensor
Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
106
107
108
109
110
111
112
113
114
115
def forward_tpu(
    self,
    logits: torch.Tensor,
    generators: dict[int, torch.Generator],
    k: Optional[torch.Tensor],
    p: Optional[torch.Tensor],
) -> torch.Tensor:
    logits = apply_top_k_top_p_tpu(logits, k, p)
    probs = logits.softmax(dim=-1, dtype=torch.float32)
    return random_sample(probs, generators)

apply_top_k_only 

apply_top_k_only(logits: Tensor, k: Tensor) -> Tensor

Apply top-k mask to the logits.

This implementation doesn't involve sorting the entire vocab.

The logits tensor may be updated in-place.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
def apply_top_k_only(
    logits: torch.Tensor,
    k: torch.Tensor,
) -> torch.Tensor:
    """
    Apply top-k mask to the logits.

    This implementation doesn't involve sorting the entire vocab.

    The logits tensor may be updated in-place.
    """
    no_top_k_mask = k == logits.shape[1]
    # Set non-top-k rows to 1 so that we can gather.
    k = k.masked_fill(no_top_k_mask, 1)
    max_top_k = k.max()
    # topk.values tensor has shape [batch_size, max_top_k].
    # Convert top k to 0-based index in range [0, max_top_k).
    k_index = k.sub_(1).unsqueeze(1)
    top_k_mask = logits.topk(max_top_k, dim=1).values.gather(1, k_index.long())
    # Handle non-topk rows.
    top_k_mask.masked_fill_(no_top_k_mask.unsqueeze(1), -float("inf"))
    logits.masked_fill_(logits < top_k_mask, -float("inf"))
    return logits

apply_top_k_top_p 

apply_top_k_top_p(
    logits: Tensor, k: Optional[Tensor], p: Optional[Tensor]
) -> Tensor

Apply top-k and top-p masks to the logits.

If a top-p is used, this function will sort the logits tensor, which can be slow for large batches.

The logits tensor may be updated in-place.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
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
def apply_top_k_top_p(
    logits: torch.Tensor,
    k: Optional[torch.Tensor],
    p: Optional[torch.Tensor],
) -> torch.Tensor:
    """Apply top-k and top-p masks to the logits.

    If a top-p is used, this function will sort the logits tensor,
    which can be slow for large batches.

    The logits tensor may be updated in-place.
    """
    if p is None:
        if k is None:
            return logits

        # Avoid sorting vocab for top-k only case.
        return apply_top_k_only(logits, k)

    logits_sort, logits_idx = logits.sort(dim=-1, descending=False)

    if k is not None:
        # Apply top-k.
        top_k_mask = logits_sort.size(1) - k.to(torch.long)  # shape: B
        # Get all the top_k values.
        top_k_mask = logits_sort.gather(1, top_k_mask.unsqueeze(dim=1))
        top_k_mask = logits_sort < top_k_mask
        logits_sort.masked_fill_(top_k_mask, -float("inf"))

    if p is not None:
        # Apply top-p.
        probs_sort = logits_sort.softmax(dim=-1)
        probs_sum = torch.cumsum(probs_sort, dim=-1, out=probs_sort)
        top_p_mask = probs_sum <= 1 - p.unsqueeze(dim=1)
        # at least one
        top_p_mask[:, -1] = False
        logits_sort.masked_fill_(top_p_mask, -float("inf"))

    # Re-sort the probabilities.
    logits = logits_sort.scatter(dim=-1, index=logits_idx, src=logits_sort)
    return logits

apply_top_k_top_p_tpu 

apply_top_k_top_p_tpu(
    logits: Tensor, k: Tensor, p: Tensor
) -> Tensor

Apply top-k and top-p optimized for TPU.

This algorithm avoids using torch.scatter which is extremely slow on TPU. This is achieved by finding a "cut-off" element in the original logit, and after thresholding the logit using this cut-off, the remaining elements shall constitute the top-p set.

Note: in the case of tie (i.e. multipple cut-off elements present in the logit), all tie elements are included in the top-p set. In other words, this function does not break ties. Instead, these tie tokens have equal chance of being chosen during final sampling, so we can consider the tie being broken then.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
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
def apply_top_k_top_p_tpu(
    logits: torch.Tensor,
    k: torch.Tensor,
    p: torch.Tensor,
) -> torch.Tensor:
    """
    Apply top-k and top-p optimized for TPU.

    This algorithm avoids using torch.scatter which is extremely slow on TPU.
    This is achieved by finding a "cut-off" element in the original logit, and
    after thresholding the logit using this cut-off, the remaining elements
    shall constitute the top-p set.

    Note: in the case of tie (i.e. multipple cut-off elements present in the
    logit), all tie elements are included in the top-p set. In other words,
    this function does not break ties. Instead, these tie tokens have equal
    chance of being chosen during final sampling, so we can consider the tie
    being broken then.
    """
    probs = logits.softmax(dim=-1)
    probs_sort, _ = probs.sort(dim=-1, descending=False)

    if k is not None:
        top_k_count = probs_sort.size(1) - k.to(torch.long)  # shape: (batch, )
        top_k_count = top_k_count.unsqueeze(dim=1)
        top_k_cutoff = probs_sort.gather(-1, top_k_count)

        # Make sure the no top-k rows are no-op.
        no_top_k_mask = (k == logits.shape[1]).unsqueeze(dim=1)
        top_k_cutoff.masked_fill_(no_top_k_mask, -float("inf"))

        elements_to_discard = probs < top_k_cutoff
        logits.masked_fill_(elements_to_discard, -float("inf"))

    if p is not None:
        cumprob = torch.cumsum(probs_sort, dim=-1)
        top_p_mask = cumprob <= 1 - p.unsqueeze(dim=1)
        top_p_mask[:, -1] = False  # at least one

        top_p_count = top_p_mask.sum(dim=-1).unsqueeze(1)
        top_p_cutoff = probs_sort.gather(-1, top_p_count)
        elements_to_discard = probs < top_p_cutoff
        logits.masked_fill_(elements_to_discard, -float("inf"))

    return logits

flashinfer_sample 

flashinfer_sample(
    logits: Tensor,
    k: Optional[Tensor],
    p: Optional[Tensor],
    generators: dict[int, Generator],
) -> Tensor

Sample from the logits using FlashInfer.

Statistically, this function is equivalent to the random_sample function. However, this function is faster because it avoids sorting the logits tensor via rejection sampling.

NOTE: The outputs of this function do not necessarily match the outputs of the random_sample function. It only guarantees that the outputs are statistically equivalent.

NOTE: This function includes CPU-GPU synchronization, while random_sample does not. Call this function at the end of the forward pass to minimize the synchronization overhead.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
def flashinfer_sample(
    logits: torch.Tensor,
    k: Optional[torch.Tensor],
    p: Optional[torch.Tensor],
    generators: dict[int, torch.Generator],
) -> torch.Tensor:
    """Sample from the logits using FlashInfer.

    Statistically, this function is equivalent to the `random_sample` function.
    However, this function is faster because it avoids sorting the logits tensor
    via rejection sampling.

    NOTE: The outputs of this function do not necessarily match the outputs of
    the `random_sample` function. It only guarantees that the outputs are
    statistically equivalent.

    NOTE: This function includes CPU-GPU synchronization, while `random_sample`
    does not. Call this function at the end of the forward pass to minimize
    the synchronization overhead.
    """
    assert not (k is None and p is None)
    if k is None:
        # Top-p only.
        probs = logits.softmax(dim=-1, dtype=torch.float32)
        next_token_ids = flashinfer.sampling.top_p_sampling_from_probs(
            probs, p, deterministic=True)
    elif p is None:
        # Top-k only.
        probs = logits.softmax(dim=-1, dtype=torch.float32)
        next_token_ids = flashinfer.sampling.top_k_sampling_from_probs(
            probs, k, deterministic=True)
    else:
        # Both top-k and top-p.
        next_token_ids = flashinfer.sampling.top_k_top_p_sampling_from_logits(
            logits, k, p, deterministic=True)

    return next_token_ids.view(-1)

random_sample 

random_sample(
    probs: Tensor, generators: dict[int, Generator]
) -> Tensor

Randomly sample from the probabilities.

We use this function instead of torch.multinomial because torch.multinomial causes CPU-GPU synchronization.

Source code in vllm/v1/sample/ops/topk_topp_sampler.py 
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
def random_sample(
    probs: torch.Tensor,
    generators: dict[int, torch.Generator],
) -> torch.Tensor:
    """Randomly sample from the probabilities.

    We use this function instead of torch.multinomial because torch.multinomial
    causes CPU-GPU synchronization.
    """
    q = torch.empty_like(probs)
    # NOTE(woosuk): To batch-process the requests without their own seeds,
    # which is the common case, we first assume that every request does
    # not have its own seed. Then, we overwrite the values for the requests
    # that have their own seeds.
    if len(generators) != probs.shape[0]:
        q.exponential_()
    if generators:
        # TODO(woosuk): This can be slow because we handle each request
        # one by one. Optimize this.
        for i, generator in generators.items():
            q[i].exponential_(generator=generator)
    return probs.div_(q).argmax(dim=-1).view(-1)