Fixing other types of models + tests + Damn you python scoping.

server/tests/utils/test_tokens.py

@@ -50,19 +50,39 @@ def test_batch_top_tokens():
     top_n_tokens = [0, 2, 3, 4, 5]
     top_n_tokens_tensor = torch.tensor(top_n_tokens)
     inp_logprobs = torch.tensor([[-1.0, -3.0, -4.0, -2.0, -3.0]] * 5)
+    accepted_ids = torch.ones_like(top_n_tokens_tensor)
 
     topn_tok_ids, topn_tok_logprobs = batch_top_tokens(
-        top_n_tokens, top_n_tokens_tensor, inp_logprobs
+        top_n_tokens, top_n_tokens_tensor, inp_logprobs, accepted_ids
     )
 
-    assert topn_tok_ids[0] == []
-    assert topn_tok_ids[1] == [0, 3]
-    assert topn_tok_ids[2] == [0, 3, 1, 4]
-    assert topn_tok_ids[3] == [0, 3, 1, 4]
-    assert topn_tok_ids[4] == [0, 3, 1, 4, 2]
+    assert topn_tok_ids[0] == [[]]
+    assert topn_tok_ids[1] == [[0, 3]]
+    assert topn_tok_ids[2] == [[0, 3, 1, 4]]
+    assert topn_tok_ids[3] == [[0, 3, 1, 4]]
+    assert topn_tok_ids[4] == [[0, 3, 1, 4, 2]]
 
-    assert topn_tok_logprobs[0] == []
-    assert topn_tok_logprobs[1] == [-1, -2]
-    assert topn_tok_logprobs[2] == [-1, -2, -3, -3]
-    assert topn_tok_logprobs[3] == [-1, -2, -3, -3]
-    assert topn_tok_logprobs[4] == [-1, -2, -3, -3, -4]
+    assert topn_tok_logprobs[0] == [[]]
+    assert topn_tok_logprobs[1] == [[-1, -2]]
+    assert topn_tok_logprobs[2] == [[-1, -2, -3, -3]]
+    assert topn_tok_logprobs[3] == [[-1, -2, -3, -3]]
+    assert topn_tok_logprobs[4] == [[-1, -2, -3, -3, -4]]
+
+    # Now let's make second member of the batch be speculated
+    inp_logprobs = torch.tensor([[-1.0, -3.0, -4.0, -2.0, -3.0]] * 5 * 2)
+    accepted_ids[1]  = 2
+    topn_tok_ids, topn_tok_logprobs = batch_top_tokens(
+        top_n_tokens, top_n_tokens_tensor, inp_logprobs, accepted_ids
+    )
+
+    assert topn_tok_ids[0] == [[]]
+    assert topn_tok_ids[1] == [[0, 3], [0, 3]]
+    assert topn_tok_ids[2] == [[0, 3, 1, 4]]
+    assert topn_tok_ids[3] == [[0, 3, 1, 4]]
+    assert topn_tok_ids[4] == [[0, 3, 1, 4, 2]]
+
+    assert topn_tok_logprobs[0] == [[]]
+    assert topn_tok_logprobs[1] == [[-1, -2], [-1, -2]]
+    assert topn_tok_logprobs[2] == [[-1, -2, -3, -3]]
+    assert topn_tok_logprobs[3] == [[-1, -2, -3, -3]]
+    assert topn_tok_logprobs[4] == [[-1, -2, -3, -3, -4]]

server/text_generation_server/models/causal_lm.py

@@ -580,10 +580,13 @@ class CausalLM(Model):
         generations: List[Generation] = []
         stopped = True
 
+        # Speculation is not active for causal
+        accepted_ids = torch.ones_like(batch.input_ids) 
         batch_top_token_ids, batch_top_token_logprobs = batch_top_tokens(
             batch.top_n_tokens,
             batch.top_n_tokens_tensor,
             torch.log_softmax(logits[:, -1], -1),
+            accepted_ids,
         )
 
         start_decode = time.time_ns()

server/text_generation_server/models/flash_causal_lm.py

@@ -858,9 +858,8 @@ class FlashCausalLM(Model):
             speculative_logits,
         )
 
-        speculated_length = batch.speculative_ids.shape[-1] if batch.speculative_ids is not None else 0
         batch_top_token_ids, batch_top_token_logprobs = batch_top_tokens(
-            batch.top_n_tokens, batch.top_n_tokens_tensor, logprobs, accepted_ids, speculated_length
+            batch.top_n_tokens, batch.top_n_tokens_tensor, logprobs, accepted_ids
         )
 
         if prefill:

server/text_generation_server/models/seq2seq_lm.py

@@ -640,10 +640,13 @@ class Seq2SeqLM(Model):
             batch.past_key_values,
         )
 
+        # Speculation is not active for seq2seq
+        accepted_ids = torch.ones_like(batch.input_ids) 
         batch_top_token_ids, batch_top_token_logprobs = batch_top_tokens(
             batch.top_n_tokens,
             batch.top_n_tokens_tensor,
             torch.log_softmax(logits[:, -1], -1),
+            accepted_ids,
         )
 
         start_decode = time.time_ns()

server/text_generation_server/utils/tokens.py

@@ -436,8 +436,8 @@ class HeterogeneousSampling:
 
 
 def batch_top_tokens(
-        top_n_tokens: List[int], top_n_tokens_tensor: torch.Tensor, logprobs: torch.Tensor, accepted_ids: torch.Tensor, speculative_length: int
-) -> Tuple[List[List[int]], List[List[float]]]:
+        top_n_tokens: List[int], top_n_tokens_tensor: torch.Tensor, logprobs: torch.Tensor, accepted_ids: torch.Tensor
+) -> Tuple[List[List[List[int]]], List[List[List[float]]]]:
     """Find the top n most likely tokens for a batch of generations.
 
     When multiple tokens have equal probabilities and they don't all fit, the
@@ -446,13 +446,14 @@ def batch_top_tokens(
     max_top_n = max(top_n_tokens)
     # Early exit when top_n_tokens is not used
     if max_top_n == 0:
-        return [[]] * len(top_n_tokens), [[]] * len(top_n_tokens)
+        return [[[]]] * len(top_n_tokens), [[[]]] * len(top_n_tokens)
 
 
-    n = speculative_length + 1
-    top_n_tokens_tensor = top_n_tokens_tensor.repeat_interleave(n)
+    batch_size = accepted_ids.shape[0]
+    speculate_size = logprobs.shape[0] // batch_size
+    top_n_tokens_tensor = top_n_tokens_tensor.repeat_interleave(speculate_size)
     # Ensure top_n doesn't exceed vocab size
-    top_n_tokens = [min(tok, logprobs.size(-1)) for tok in top_n_tokens for _ in range(speculative_length + 1)]
+    top_n_tokens = [min(tok, logprobs.size(-1)) for tok in top_n_tokens for _ in range(speculate_size)]
 
     # Parallel kthvalue adapted from https://discuss.pytorch.org/t/how-to-efficiently-get-the-k-th-largest-values-in-parallel/160529/2
     # Sorted topk is faster than torch.sort() since we only need a small subset
@@ -477,12 +478,14 @@ def batch_top_tokens(
 
     batch_top_token_ids = []
     batch_top_token_logprobs = []
-    accepted_ids = accepted_ids.tolist()
-    for i, n_accepted_ids in enumerate(accepted_ids):
-        _top_indices = top_indices[n * i: n * (i + 1)]
-        _top_values = top_values[n * i: n * (i + 1)]
-        _top_n_ishes = top_n_ishes[n * i: n * (i + 1)]
-        _top_n_tokens = top_n_tokens[n * i: n * (i + 1)]
+    accepted_ids_list = accepted_ids.tolist()
+    for i, n_accepted_ids in enumerate(accepted_ids_list):
+        start = speculate_size * i
+        stop = speculate_size * (i + 1)
+        _top_indices = top_indices[start: stop]
+        _top_values = top_values[start: stop]
+        _top_n_ishes = top_n_ishes[start: stop]
+        _top_n_tokens = top_n_tokens[start: stop]
         _top_indices = _top_indices[:n_accepted_ids]
         _top_values = _top_values[:n_accepted_ids]
         _top_n_ishes = _top_n_ishes[:n_accepted_ids]