Re-work for minimal changes, add test

llms/mlx_lm/utils.py

@@ -13,7 +13,18 @@ import time
 from dataclasses import dataclass
 from pathlib import Path
 from textwrap import dedent
-from typing import Any, Callable, Dict, Generator, List, NamedTuple, Optional, Tuple, Type, Union
+from typing import (
+    Any,
+    Callable,
+    Dict,
+    Generator,
+    List,
+    NamedTuple,
+    Optional,
+    Tuple,
+    Type,
+    Union,
+)
 
 import mlx.core as mx
 import mlx.nn as nn
@@ -77,8 +88,8 @@ class GenerationResponse:
 
     text: str
     token: int
-    from_draft: bool = False
     logprobs: mx.array
+    from_draft: bool
     prompt_tokens: int
     prompt_tps: float
     generation_tokens: int
@@ -207,8 +218,6 @@ def maybe_quantize_kv_cache(prompt_cache, quantized_kv_start, kv_group_size, kv_
                     group_size=kv_group_size, bits=kv_bits
                 )
 
-class TokenMetadata(NamedTuple):
-    from_draft: bool = False
 
 def generate_step(
     prompt: mx.array,
@@ -224,7 +233,7 @@ def generate_step(
     kv_group_size: int = 64,
     quantized_kv_start: int = 0,
     prompt_progress_callback: Optional[Callable[int, int]] = None,
-) -> Generator[Tuple[mx.array, mx.array, TokenMetadata], None, None]:
+) -> Generator[Tuple[mx.array, mx.array], None, None]:
     """
     A generator producing token ids based on the given prompt from the model.
 
@@ -252,8 +261,7 @@ def generate_step(
            prompt tokens processed so far and the total number of prompt tokens.
 
     Yields:
-        Tuple[mx.array, mx.array, TokenMetadata]: One token, a vector of log
+        Tuple[mx.array, mx.array]: One token and a vector of log probabilities.
-          probabilities, and token metadata.
     """
 
     y = prompt
@@ -328,6 +336,7 @@ def generate_step(
         y, logprobs = next_y, next_logprobs
         n += 1
 
+
 def speculative_generate_step(
     prompt: mx.array,
     model: nn.Module,
@@ -342,7 +351,7 @@ def speculative_generate_step(
     kv_bits: Optional[int] = None,
     kv_group_size: int = 64,
     quantized_kv_start: int = 0,
-) -> Generator[Tuple[mx.array, mx.array, TokenMetadata], None, None]:
+) -> Generator[Tuple[mx.array, mx.array, bool], None, None]:
     """
     A generator producing token ids based on the given prompt from the model.
 
@@ -369,8 +378,8 @@ def speculative_generate_step(
            when ``kv_bits`` is non-None. Default: ``0``.
 
     Yields:
-        Tuple[mx.array, mx.array, TokenMetadata]: One token, a vector of log
+        Tuple[mx.array, mx.array, bool]: One token, a vector of log probabilities,
-          probabilities, and token metadata.
+          and a bool indicating if the token was generated by the draft model
     """
 
     y = prompt
@@ -455,12 +464,12 @@ def speculative_generate_step(
                     break
                 n += 1
                 ntoks += 1
-                yield tn, lpn, TokenMetadata(from_draft=True)
+                yield tn, lpn, True
                 if ntoks == max_tokens:
                     break
             if ntoks < max_tokens:
                 ntoks += 1
-                yield tokens[n], logprobs[n], TokenMetadata(from_draft=False)
+                yield tokens[n], logprobs[n], False
 
             if ntoks == max_tokens:
                 break
@@ -523,6 +532,10 @@ def stream_generate(
     if draft_model is None:
         kwargs.pop("num_draft_tokens", None)
         token_generator = generate_step(prompt, model, **kwargs)
+        # from_draft always false for non-speculative generation
+        token_generator = (
+            (token, logprobs, False) for token, logprobs in token_generator
+        )
     else:
         kwargs.pop("max_kv_size", None)
         token_generator = speculative_generate_step(
@@ -531,7 +544,7 @@ def stream_generate(
     with wired_limit(model, [generation_stream]):
         detokenizer.reset()
         tic = time.perf_counter()
-        for n, (token, logprobs, token_metadata) in enumerate(token_generator):
+        for n, (token, logprobs, from_draft) in enumerate(token_generator):
             if n == 0:
                 prompt_time = time.perf_counter() - tic
                 prompt_tps = prompt.size / prompt_time
@@ -544,8 +557,8 @@ def stream_generate(
             yield GenerationResponse(
                 text=detokenizer.last_segment,
                 token=token,
-                from_draft=token_metadata.from_draft,
                 logprobs=logprobs,
+                from_draft=from_draft,
                 prompt_tokens=prompt.size,
                 prompt_tps=prompt_tps,
                 generation_tokens=n + 1,
@@ -558,8 +571,8 @@ def stream_generate(
         yield GenerationResponse(
             text=detokenizer.last_segment,
             token=token,
-            from_draft=token_metadata.from_draft,
             logprobs=logprobs,
+            from_draft=from_draft,
             prompt_tokens=prompt.size,
             prompt_tps=prompt_tps,
             generation_tokens=n + 1,

llms/tests/test_generate.py

@@ -1,17 +1,24 @@
 # Copyright © 2024 Apple Inc.
 
 import unittest
+from typing import List
 
 from mlx_lm.sample_utils import make_logits_processors
-from mlx_lm.utils import generate, load
+from mlx_lm.utils import (
+    GenerationResponse,
+    generate,
+    load,
+    make_sampler,
+    stream_generate,
+)
 
 
 class TestGenerate(unittest.TestCase):
 
     @classmethod
     def setUpClass(cls):
-        HF_MODEL_PATH = "mlx-community/Qwen1.5-0.5B-Chat-4bit"
+        cls.HF_MODEL_PATH = "mlx-community/Qwen1.5-0.5B-Chat-4bit"
-        cls.model, cls.tokenizer = load(HF_MODEL_PATH)
+        cls.model, cls.tokenizer = load(cls.HF_MODEL_PATH)
 
     def test_generate(self):
         # Simple test that generation runs
@@ -51,6 +58,34 @@ class TestGenerate(unittest.TestCase):
         )
         self.assertEqual(len(all_toks), len(init_toks) + 5)
 
+    def test_stream_generate_speculative(self):
+        # Use same model as draft model, this is not a speed test
+        draft_model, _ = load(self.HF_MODEL_PATH)
+
+        results: List[GenerationResponse] = []
+        drafted: List[bool] = []
+
+        # make a determinate sampler
+        sampler = make_sampler(temp=0.0)
+
+        for generation_result in stream_generate(
+            model=self.model,
+            tokenizer=self.tokenizer,
+            prompt="hello",
+            max_tokens=5,
+            draft_model=draft_model,
+            num_draft_tokens=2,
+            sampler=sampler,
+        ):
+            drafted.append(generation_result.from_draft)
+            results.append(generation_result)
+
+        self.assertEqual(len(results), 5)
+        # since num_draft_tokens is 2 and draft model is the same, the
+        # first 2 generations should be drafts, the third should come
+        # from the target model, and last two should be drafts
+        self.assertEqual(drafted, [True, True, False, True, True])
+
 
 if __name__ == "__main__":
     unittest.main()