Fix typical acceptance sampler with correct recovered token ids (#8562)

tests/samplers/test_typical_acceptance_sampler.py

@@ -365,7 +365,7 @@ def test_accept_tokens_partially(seed: int, device: str):
     # Next only keep the first 2 draft tokens same as the zero temperature
     # tokens. For the remaining 3 choose some other tokens. In the
     # response we will expect the first 2 tokens to be the same as the
-    # draft tokens and the rest as -1
+    # draft tokens and the recovered token and rest as -1
     draft_token_ids_to_replace = get_draft_token_ids(
         batch_size, k, vocab_size, zero_temperature_token_ids)
     draft_token_ids = torch.cat(
@@ -378,6 +378,8 @@ def test_accept_tokens_partially(seed: int, device: str):
     assert output_token_ids.shape[0] == batch_size
     assert output_token_ids.shape[1] == (k + 1)
     assert torch.all(output_token_ids[:, :2] == draft_token_ids[:, :2])
+    assert torch.all(
+        output_token_ids[:, 2] == target_with_bonus_probs.argmax(-1)[:, 2])
     assert torch.all(output_token_ids[:, -3:] == -1)
 
 
@@ -443,14 +445,14 @@ def test_accept_tokens_set_non_default_posteriors(seed: int, device: str):
 @pytest.mark.parametrize("seed", list(range(10)))
 @pytest.mark.parametrize("device", CUDA_DEVICES)
 @torch.inference_mode()
-def test_replacement_token_ids(seed: int, device: str):
+def test_get_recovered_token_ids(seed: int, device: str):
     """
     Test the TypicalAcceptanceSampler's method for generating
     replacement token IDs.
 
-    This test verifies that the `_replacement_token_ids` method of the 
+    This test verifies that the `_get_recovered_token_ids` method of the 
     TypicalAcceptanceSampler correctly identifies the token IDs to be used
-    as replacements based on the target probability distribution.
+    as recovered token IDs based on the target probability distribution.
     Specifically, it ensures that the method correctly identifies the
     tokens with the highest probability for each sequence in the batch.
     """
@@ -462,10 +464,7 @@ def test_replacement_token_ids(seed: int, device: str):
     typical_acceptance_sampler = get_acceptance_sampler(strict_mode=True)
     typical_acceptance_sampler.init_gpu_tensors(device=device)
     target_probs = torch.rand(batch_size, k, vocab_size, dtype=torch.float32)
-    expected_replacement_tokens = -torch.ones(
+    expected_replacement_tokens = torch.argmax(target_probs, dim=-1)
-        (batch_size, k), dtype=torch.long)
-    expected_replacement_tokens[:, 0] = torch.argmax(target_probs[:, 0, :],
-                                                     dim=1)
     actual_replacement_tokens = (
-        typical_acceptance_sampler._replacement_token_ids(target_probs))
+        typical_acceptance_sampler._get_recovered_token_ids(target_probs))
     assert torch.all(expected_replacement_tokens == actual_replacement_tokens)

vllm/model_executor/layers/typical_acceptance_sampler.py

@@ -80,7 +80,7 @@ class TypicalAcceptanceSampler(SpecDecodeDeterministicBaseSampler):
         target_probs = target_with_bonus_probs[:, :-1]
         accepted = self._evaluate_accepted_tokens(target_probs,
                                                   draft_token_ids)
-        recovered_token_ids = self._replacement_token_ids(target_probs)
+        recovered_token_ids = self._get_recovered_token_ids(target_probs)
         output_token_ids = self._create_output(accepted, recovered_token_ids,
                                                draft_token_ids,
                                                bonus_token_ids)
@@ -148,16 +148,10 @@ class TypicalAcceptanceSampler(SpecDecodeDeterministicBaseSampler):
         accepted_mask = candidates_prob > threshold
         return accepted_mask
 
-    def _replacement_token_ids(self, target_probs):
+    def _get_recovered_token_ids(self, target_probs):
         """
-        Generate one replacement token ID for each sequence based on target
+        The recovered token ids will fill the first unmatched token
-        probabilities. The replacement token is used as the fallback option
+        by the target token.
-        if typical acceptance sampling does not accept any draft tokens for
-        that particular sequence. 
-
-        This method computes the token IDs to be replaced by selecting the
-        token with the highest probability for each sequence in the first 
-        position. The rest of the output is filled with -1. 
 
         Parameters
         ----------
@@ -168,13 +162,9 @@ class TypicalAcceptanceSampler(SpecDecodeDeterministicBaseSampler):
         Returns
         -------
         torch.Tensor
-            A tensor of shape (batch_size, k) with the replacement 
+            A tensor of shape (batch_size, k) with the recovered token
-            token IDs. Only the first column is set, and the rest of the
+            ids which are selected from target probs.
-            columns are filled with -1.
         """
-        max_indices = torch.argmax(target_probs[:, 0, :], dim=1)
+        max_indices = torch.argmax(target_probs, dim=-1)
-        output = -torch.ones((target_probs.shape[0], target_probs.shape[1]),
+
-                             dtype=self.token_id_dtype,
+        return max_indices
-                             device=target_probs.device)
-        output[:, 0] = max_indices
-        return output