[bitsandbytes]: support read bnb pre-quantized model (vllm-project#5753)

Co-authored-by: Michael Goin <[email protected]>

Author: 2 people

docs/source/index.rst

@@ -105,6 +105,7 @@ Documentation
 
    quantization/supported_hardware
    quantization/auto_awq
+   quantization/bnb
    quantization/fp8
    quantization/fp8_e5m2_kvcache
    quantization/fp8_e4m3_kvcache

docs/source/quantization/bnb.rst

@@ -0,0 +1,43 @@
+.. _bits_and_bytes:
+
+BitsAndBytes
+==================
+
+vLLM now supports `BitsAndBytes <https://github.com/TimDettmers/bitsandbytes>`_ for more efficient model inference.
+BitsAndBytes quantizes models to reduce memory usage and enhance performance without significantly sacrificing accuracy.
+Compared to other quantization methods,  BitsAndBytes eliminates the need for calibrating the quantized model with input data.
+
+Below are the steps to utilize BitsAndBytes with vLLM.
+
+.. code-block:: console
+
+    $ pip install bitsandbytes>=0.42.0
+
+vLLM reads the model's config file and supports both in-flight quantization and pre-quantized checkpoint.
+
+You can find bitsandbytes quantized models on https://huggingface.co/models?other=bitsandbytes.
+And usually, these repositories have a config.json file that includes a quantization_config section.
+
+Read quantized checkpoint.
+--------------------------
+
+.. code-block:: python
+
+    from vllm import LLM
+    import torch
+    # unsloth/tinyllama-bnb-4bit is a pre-quantized checkpoint.
+    model_id = "unsloth/tinyllama-bnb-4bit"
+    llm = LLM(model=model_id, dtype=torch.bfloat16, trust_remote_code=True, \
+    quantization="bitsandbytes", load_format="bitsandbytes")
+
+Inflight quantization: load as 4bit quantization
+------------------------------------------------
+
+.. code-block:: python
+
+    from vllm import LLM
+    import torch
+    model_id = "huggyllama/llama-7b"
+    llm = LLM(model=model_id, dtype=torch.bfloat16, trust_remote_code=True, \
+    quantization="bitsandbytes", load_format="bitsandbytes")
+

tests/quantization/test_bitsandbytes.py

@@ -8,15 +8,20 @@
 from tests.quantization.utils import is_quant_method_supported
 from vllm import SamplingParams
 
+models_to_test = [
+    ('huggyllama/llama-7b', 'quantize model inflight'),
+    ('lllyasviel/omost-llama-3-8b-4bits', 'read pre-quantized model'),
+]
+
 
 @pytest.mark.skipif(not is_quant_method_supported("bitsandbytes"),
                     reason='bitsandbytes is not supported on this GPU type.')
-def test_load_bnb_model(vllm_runner) -> None:
-    with vllm_runner('huggyllama/llama-7b',
+@pytest.mark.parametrize("model_name, description", models_to_test)
+def test_load_bnb_model(vllm_runner, model_name, description) -> None:
+    with vllm_runner(model_name,
                      quantization='bitsandbytes',
                      load_format='bitsandbytes',
                      enforce_eager=True) as llm:
-
         model = llm.model.llm_engine.model_executor.driver_worker.model_runner.model  # noqa: E501
 
         # check the weights in MLP & SelfAttention are quantized to torch.uint8
@@ -65,12 +70,17 @@ def test_load_bnb_model(vllm_runner) -> None:
             'To be or not to be, that is the question.'
         ]
         outputs = llm.generate(prompts, sampling_params=sampling_params)
-
         assert len(outputs) == len(prompts)
 
         for index in range(len(outputs)):
             # compare the first line of the output
             actual_output = outputs[index][1][0].split('\n', 1)[0]
             expected_output = expected_outputs[index].split('\n', 1)[0]
+
+            assert len(actual_output) >= len(expected_output), (
+                f'Actual {actual_output} should be larger than or equal to '
+                f'expected {expected_output}')
+            actual_output = actual_output[:len(expected_output)]
+
             assert actual_output == expected_output, (
                 f'Expected: {expected_output}, but got: {actual_output}')

vllm/config.py

@@ -582,6 +582,7 @@ class LoadConfig:
                 mainly for profiling.
             "tensorizer" will use CoreWeave's tensorizer library for
                 fast weight loading.
+            "bitsandbytes" will load nf4 type weights.
     """
 
     load_format: Union[str, LoadFormat, "BaseModelLoader"] = LoadFormat.AUTO

vllm/engine/arg_utils.py

@@ -621,8 +621,8 @@ def create_engine_config(self, ) -> EngineConfig:
         # bitsandbytes quantization needs a specific model loader
         # so we make sure the quant method and the load format are consistent
         if (self.quantization == "bitsandbytes" or
-            self.qlora_adapter_name_or_path is not None) and \
-            self.load_format != "bitsandbytes":
+           self.qlora_adapter_name_or_path is not None) and \
+           self.load_format != "bitsandbytes":
             raise ValueError(
                 "BitsAndBytes quantization and QLoRA adapter only support "
                 f"'bitsandbytes' load format, but got {self.load_format}")

vllm/model_executor/layers/quantization/bitsandbytes.py

@@ -15,19 +15,11 @@ class BitsAndBytesConfig(QuantizationConfig):
     Reference: https://arxiv.org/abs/2305.14314
     """
 
-    def __init__(
-        self,
-        adapter_name_or_path: str,
-        target_modules: List[str],
-    ) -> None:
-
-        self.adapter_name_or_path = adapter_name_or_path
-        self.target_modules = target_modules
+    def __init__(self, ) -> None:
+        pass
 
     def __repr__(self) -> str:
-        return (
-            f"BitsAndBytesConfig(adapter_name_or_path={self.adapter_name_or_path}"
-        )
+        return "BitsAndBytesConfig"
 
     @classmethod
     def get_name(self) -> str:
@@ -49,16 +41,7 @@ def get_config_filenames() -> List[str]:
 
     @classmethod
     def from_config(cls, config: Dict[str, Any]) -> "BitsAndBytesConfig":
-        adapter_name = cls.get_from_keys(config, ["adapter_name_or_path"])
-        default_target_modules = [
-            "gate_proj", "down_proj", "up_proj", "q_proj", "k_proj", "v_proj",
-            "o_proj"
-        ]
-        if adapter_name == "":
-            target_modules = default_target_modules
-        else:
-            target_modules = cls.get_from_keys(config, ["target_modules"])
-        return cls(adapter_name, target_modules)
+        return cls()
 
     def get_quant_method(
             self,

vllm/model_executor/model_loader/loader.py

@@ -685,8 +685,14 @@ def _prepare_weights(self, model_name_or_path: str,
 
         return hf_weights_files, matched_pattern == "*.safetensors"
 
+    def _hf_weight_iter(self, hf_weights_files, use_safetensors: bool):
+        if use_safetensors:
+            return safetensors_weights_iterator(hf_weights_files)
+        else:
+            return pt_weights_iterator(hf_weights_files)
+
     def _get_quantized_weights_iterator(
-        self, model_name_or_path: str, revision: Optional[str]
+        self, model_name_or_path: str, revision: Optional[str], pre_quant: bool
     ) -> Tuple[Generator[Tuple[str, torch.Tensor], None, None], Dict[str,
                                                                      Any]]:
         """Get an iterator to the model weights with bitsandbytes quantization,
@@ -695,6 +701,7 @@ def _get_quantized_weights_iterator(
         # only load the bitsandbytes module when needed
         try:
             import bitsandbytes
+            from bitsandbytes.functional import QuantState
             if bitsandbytes.__version__ < "0.42.0":
                 raise ImportError("bitsandbytes version is wrong. Please "
                                   "install bitsandbytes>=0.42.0.")
@@ -708,17 +715,63 @@ def _get_quantized_weights_iterator(
             model_name_or_path, revision)
 
         quant_state_dict = {}
-        if use_safetensors:
-            weight_iterator = safetensors_weights_iterator(hf_weights_files)
-        else:
-            weight_iterator = pt_weights_iterator(hf_weights_files)
 
-        def generator():
+        def quantized_checkpoint() -> Generator:
+            # First iterate over all quant state weights
+            weight_iterator = self._hf_weight_iter(hf_weights_files,
+                                                   use_safetensors)
+            temp_state_dict = {}
             for weight_name, weight_tensor in weight_iterator:
+                if weight_name.endswith(".weight"):
+                    continue
+                # TODO: only nf4 quantization is supported for now
+                if weight_name.endswith(".quant_state.bitsandbytes__fp4"):
+                    raise NotImplementedError(
+                        "Only bitsandbytes_nf4 quantization"
+                        f"is supported for now. {weight_name} is fp4 quantized"
+                    )
+                temp_state_dict[weight_name] = weight_tensor
+
+            # Closure to parse quant_state for each prequant weight
+            def _parse_quant_state(param_name: str,
+                                   temp_state_dict: Dict) -> QuantState:
+                quant_state = {}
+                for k in temp_state_dict:
+                    if param_name + "." in k:
+                        quant_state[k] = temp_state_dict[k]
+                # bitsandbytes library requires
+                # weight.quant_state.bitsandbytes__nf4 in CPU
+                quant_state[param_name +
+                            ".quant_state.bitsandbytes__nf4"] = quant_state[
+                                param_name +
+                                ".quant_state.bitsandbytes__nf4"].cpu().data
+                return QuantState.from_dict(quant_state, device="cuda")
+
+            # Second iterate over all prequant and normal weights
+            # pre quantized weights would have a quant_state
+            for weight_name, weight_tensor in self._hf_weight_iter(
+                    hf_weights_files, use_safetensors):
+                # Filter out all weights whose suffix is not ".weight"
+                if not weight_name.endswith(".weight"):
+                    continue
+                if weight_name + ".quant_state.bitsandbytes__nf4" \
+                    in temp_state_dict:
+                    quant_state = _parse_quant_state(weight_name,
+                                                     temp_state_dict)
+                    weight_name = weight_name.replace(".weight", ".qweight")
+                    quant_state_dict[weight_name] = quant_state
+                    yield weight_name.replace(".weight",
+                                              ".qweight"), weight_tensor
+                else:
+                    yield weight_name, weight_tensor
+
+        def generator() -> Generator:
+            for weight_name, weight_tensor in self._hf_weight_iter(
+                    hf_weights_files, use_safetensors):
                 if any(target_module in weight_name
                        for target_module in self.target_modules):
                     weight_name = weight_name.replace(".weight", ".qweight")
-                    #  bitsandbytes requires data in GPU
+                    # bitsandbytes requires data in GPU
                     loaded_weight = weight_tensor.cuda().data
                     with set_default_torch_dtype(torch.float32):
                         processed_weight, quant_state = quantize_4bit(
@@ -732,6 +785,8 @@ def generator():
 
                 yield weight_name, processed_weight
 
+        if pre_quant:
+            return quantized_checkpoint(), quant_state_dict
         return generator(), quant_state_dict
 
     def _load_weights(self, model_config: ModelConfig,
@@ -749,12 +804,21 @@ def _load_weights(self, model_config: ModelConfig,
         logger.info("Loading weights with BitsAndBytes quantization. "
                     " May take a while ...")
 
-        qweight_iterator, quant_state_dict = (
-            self._get_quantized_weights_iterator(model_config.model,
-                                                 model_config.revision))
+        is_quantized_checkpoint = False
+        quant_config = getattr(model_config.hf_config, "quantization_config",
+                               None)
+        if quant_config is not None and quant_config.get(
+                'quant_method') == "bitsandbytes":
+            is_quantized_checkpoint = True
+
+        qweight_iterator, quant_state_dict = \
+            self._get_quantized_weights_iterator(
+            model_config.model, model_config.revision, is_quantized_checkpoint)
 
         model.load_weights(qweight_iterator)
 
+        torch.cuda.empty_cache()
+
         param_dict = dict(model.named_parameters())
         stacked_quant_state_dict: Dict[str, Dict[int, Any]] = {}
         for quant_param_name in quant_state_dict:
@@ -792,9 +856,9 @@ def _load_weights(self, model_config: ModelConfig,
                         f"pack_factor not set for parameter {param_name}.")
 
                 num_elements = [0] * len(quant_states)
-                for seq, quant_state in enumerate(quant_states.items()):
+                for seq, quant_state in quant_states.items():
                     num_elements[seq] = math.prod(
-                        quant_state[1].shape) // pack_ratio
+                        quant_state.shape) // pack_ratio
 
                 offsets = np.concatenate(([0], np.cumsum(num_elements)))
                 set_weight_attrs(param, {"bnb_shard_offsets": offsets})

vllm/model_executor/model_loader/weight_utils.py

@@ -117,6 +117,7 @@ def convert_bin_to_safetensor_file(
 # TODO(woosuk): Move this to other place.
 def get_quant_config(model_config: ModelConfig,
                      load_config: LoadConfig) -> QuantizationConfig:
+
     quant_cls = get_quantization_config(model_config.quantization)
     # Read the quantization config from the HF model config, if available.
     hf_quant_config = getattr(model_config.hf_config, "quantization_config",