[WIP] Make AWQ more general

Summary:
* Added AWQConfig that takes a base config and made corresponding changes
in other parts of the flow

Test Plan:
TODO

Reviewers:

Subscribers:

Tasks:

Tags:

Author: jerryzh168

torchao/_models/_eval.py

@@ -57,8 +57,13 @@ def _model_call(self, inps):
 
         max_seq_length = min(max(inps.size()), self.max_length)
         with torch.device(self._device):
-            self._model.setup_caches(self.batch_size, max_seq_length)
+            if hasattr(self._model, "setup_caches"):
+                self._model.setup_caches(self.batch_size, max_seq_length)
         logits = self._model(*input)
+        from transformers.modeling_outputs import CausalLMOutputWithPast
+
+        if isinstance(logits, CausalLMOutputWithPast):
+            logits = logits.logits
         return logits
 
     def run_eval(self, tasks, limit):
@@ -84,7 +89,11 @@ def eot_token_id(self):
         try:
             return self.tokenizer.eos_id()
         except:
-            return self.tokenizer.eos_id
+            try:
+                return self.tokenizer.eos_id
+            except:
+                idx = self.tokenizer.all_special_tokens.index("<|endoftext|>")
+                return self.tokenizer.all_special_ids[idx]
 
     @property
     def max_length(self):
@@ -102,8 +111,8 @@ def batch_size(self):
     def device(self):
         return self._device
 
-    def tok_decode(self, tokens):
-        decoded = self.tokenizer.decode(tokens)
+    def tok_decode(self, tokens, **kwargs):
+        decoded = self.tokenizer.decode(tokens, **kwargs)
         return decoded
 
     def tok_encode(self, string: str, **kwargs):
@@ -115,8 +124,8 @@ def tok_encode(self, string: str, **kwargs):
                 tokens = [self.tokenizer.bos_id] + tokens
         return tokens
 
-    def _model_generate(self, context, max_length, eos_token_id):
-        raise Exception("unimplemented")
+    # def _model_generate(self, context, max_length, stop, **generation_kwargs):
+    #     super()._model_generate(context, max_length, stop, **generation_kwargs)
 
 
 class LMEvalInputRecorder(TransformerEvalWrapper):

torchao/_models/llama/eval.py

@@ -237,6 +237,87 @@ def run_evaluation(
             quantize_(
                 model, codebook_weight_only(dtype=torch.uint4, scale_block_size=64)
             )
+        elif quantization.startswith("awq-uintx"):
+            from torchao._models._eval import TransformerEvalWrapper
+            from torchao.utils import TORCH_VERSION_AT_LEAST_2_3
+
+            if not TORCH_VERSION_AT_LEAST_2_3:
+                print("Awq requires torch2.3+")
+                exit()
+            from torchao.prototype.awq import (
+                AWQObservedLinear,
+                awq_uintx,
+                insert_awq_observer_,
+            )
+
+            quant_dtype = quantization.split("-")[1]
+            group_size = int(quantization.split("-")[2])
+            quant_dtype = getattr(torch, quant_dtype, torch.uint8)
+            model = model.to(device)
+            # get calibration data
+            insert_awq_observer_(
+                model, 1, 256, quant_dtype=quant_dtype, group_size=group_size
+            )
+            TransformerEvalWrapper(
+                model=model.to(device),
+                tokenizer=tokenizer,
+                max_seq_length=256,
+                input_prep_func=prepare_inputs_for_model,
+                device=device,
+            ).run_eval(
+                tasks=["wikitext"],
+                limit=1,
+            )
+            is_observed_linear = lambda m, fqn: isinstance(m, AWQObservedLinear)
+            use_hqq = "hqq" in quantization
+            quantize_(
+                model,
+                awq_uintx(
+                    quant_dtype=quant_dtype, group_size=group_size, use_hqq=use_hqq
+                ),
+                is_observed_linear,
+            )
+
+        elif quantization.startswith("awq-8da4w"):
+            from torchao._models._eval import TransformerEvalWrapper
+            from torchao.utils import TORCH_VERSION_AT_LEAST_2_3
+
+            if not TORCH_VERSION_AT_LEAST_2_3:
+                print("Awq requires torch2.3+")
+                exit()
+            from torchao.prototype.awq import (
+                AWQObservedLinear,
+                awq_uintx,
+                insert_awq_observer_,
+            )
+
+            quant_dtype = quantization.split("-")[1]
+            group_size = int(quantization.split("-")[2])
+            quant_dtype = getattr(torch, quant_dtype, torch.uint8)
+            model = model.to(device)
+            # get calibration data
+            insert_awq_observer_(
+                model, 1, 256, quant_dtype=quant_dtype, group_size=group_size
+            )
+            TransformerEvalWrapper(
+                model=model.to(device),
+                tokenizer=tokenizer,
+                max_seq_length=256,
+                input_prep_func=prepare_inputs_for_model,
+                device=device,
+            ).run_eval(
+                tasks=["wikitext"],
+                limit=1,
+            )
+            is_observed_linear = lambda m, fqn: isinstance(m, AWQObservedLinear)
+            use_hqq = "hqq" in quantization
+            quantize_(
+                model,
+                awq_uintx(
+                    quant_dtype=quant_dtype, group_size=group_size, use_hqq=use_hqq
+                ),
+                is_observed_linear,
+            )
 
     if compile:
         model = torch.compile(model, mode="max-autotune", fullgraph=True)

torchao/prototype/awq/__init__.py

@@ -1,8 +1,9 @@
-from .api import awq_uintx, insert_awq_observer_
+from .api import AWQConfig, awq_uintx, insert_awq_observer_
 from .core import AWQObservedLinear
 
 __all__ = [
     "awq_uintx",
     "insert_awq_observer_",
     "AWQObservedLinear",
+    "AWQConfig",
 ]

torchao/prototype/awq/api.py

@@ -5,7 +5,7 @@
 # LICENSE file in the root directory of this source tree.
 import types
 from dataclasses import dataclass
-from typing import Optional
+from typing import List, Optional
 
 import torch
 
@@ -30,12 +30,15 @@
     ZeroPointDomain,
 )
 from torchao.quantization.transform_module import (
+    _QUANTIZE_CONFIG_HANDLER,
     register_quantize_module_handler,
 )
+from torchao.utils import DummyModule
 
 from .core import (
     AWQObservedLinear,
     AWQObserver,
+    AWQObserver2,
 )
 
 assert len(_DTYPE_TO_BIT_WIDTH) > 0, (
@@ -50,6 +53,7 @@ def insert_awq_observer_(
     quant_dtype: torch.dtype = torch.uint4,
     scale_search_space_size: int = 20,
     group_size: int = 128,
+    base_config: Optional[AOBaseConfig] = None,
 ):
     """
     Inserts AWQObserver into Linear layers of a given model.
@@ -80,22 +84,30 @@ def insert_awq_observer_(
 
     def replace_with_observer(layer):
         # creates observer and replaces linear layers with AWQObservedLinear layers
-        observer = AWQObserver(
-            layer.weight,
-            layer.bias,
-            quantization_granularity,
-            mapping_type,
-            quant_dtype,
-            n_validation_examples,
-            validation_sequence_len,
-            scale_search_space_size,
-            preserve_zero=preserve_zero,
-            zero_point_domain=zero_point_domain,
-            zero_point_dtype=zero_point_dtype,
-            quant_min=quant_min,
-            quant_max=quant_max,
-            eps=eps,
-        )
+        if base_config is None:
+            observer = AWQObserver(
+                layer.weight,
+                layer.bias,
+                quantization_granularity,
+                mapping_type,
+                quant_dtype,
+                n_validation_examples,
+                validation_sequence_len,
+                scale_search_space_size,
+                preserve_zero=preserve_zero,
+                zero_point_domain=zero_point_domain,
+                zero_point_dtype=zero_point_dtype,
+                quant_min=quant_min,
+                quant_max=quant_max,
+                eps=eps,
+            )
+        else:
+            observer = AWQObserver2(
+                layer.weight,
+                layer.bias,
+                base_config,
+                scale_search_space_size,
+            )
         return AWQObservedLinear.from_float(layer, observer)
 
     _replace_with_custom_fn_if_matches_filter(model, replace_with_observer, _is_linear)
@@ -194,3 +206,97 @@ def _awq_uintx_transform(
     linear.extra_repr = types.MethodType(_linear_extra_repr, module)
     linear.bias = observed_linear.bias
     return linear
+
+
+@dataclass
+class AWQConfig(AOBaseConfig):
+    """
+    Configuration for quantizing linear layers when passed into quantize_()
+
+    Args:
+        base_config (AOBaseConfig): The quantization config that we can apply awq on top of, e.g. 8da4w, int4 weight only
+        step (str): a string of "prepare", "convert" or "load" indicating the step of AWQ process
+            prepare: insert AWQ Observers to linear
+            convert: convert the observed linear modules to linear modules with awq quantized weights
+            load: convert the floating point model to a dummy awq quantized model
+        example_input_shape (Optional[List[int]])): This is used for load step to initialize a random example input
+        scale_search_space_size (int): the number of scales to search for
+        set_inductor_config: if True, adjusts `torchinductor` settings to recommended values.
+    """
+
+    base_config: AOBaseConfig
+    step: str
+    example_input_shape: Optional[List[int]] = None
+    scale_search_space_size: int = 20
+    set_inductor_config: bool = True
+
+    def __post_init__(self):
+        OPTIONS = ["prepare", "convert", "load"]
+        assert self.step in OPTIONS, f"Only {OPTIONS} are supported, got {self.step}"
+
+
+@register_quantize_module_handler(AWQConfig)
+def _awq_transform(
+    module: torch.nn.Module,
+    config: AWQUIntXConfig,
+) -> torch.nn.Module:
+    if config.set_inductor_config:
+        torchao.quantization.utils.recommended_inductor_config_setter()
+
+    step = config.step
+    scale_search_space_size = config.scale_search_space_size
+    observed_linear = None
+    base_config = config.base_config
+
+    if step == "prepare":
+        observer = AWQObserver2(
+            module.weight,
+            module.bias,
+            base_config,
+            scale_search_space_size,
+        )
+        return AWQObservedLinear.from_float(module, observer)
+    elif step == "load":
+        # loading from pre-quantized checkpoint
+        observer = AWQObserver2(
+            module.weight,
+            module.bias,
+            base_config,
+            scale_search_space_size,
+        )
+        observed_linear = AWQObservedLinear.from_float(module, observer)
+        assert config.example_input_shape is not None, (
+            "When step is load, we expect example_input_shape to be specified as well"
+        )
+        example_input = torch.randn(
+            config.example_input_shape,
+            device=module.weight.device,
+            dtype=module.weight.dtype,
+        )
+        observed_linear(example_input)
+    else:
+        if not isinstance(module, AWQObservedLinear):
+            print(f"convert: module is not AWQObservedLinear, skipping: {type(module)}")
+            return module
+        observed_linear = module
+
+    assert observed_linear is not None
+    equalization_scale = observed_linear.act_obs.calculate_qparams()
+
+    base_config_handler = _QUANTIZE_CONFIG_HANDLER[type(config.base_config)]
+    dummy_mod = DummyModule(observed_linear.weight * equalization_scale)
+    quant_mod = base_config_handler(dummy_mod, config.base_config)
+    qw = quant_mod.weight
+    qw = to_weight_tensor_with_linear_activation_scale_metadata(qw, equalization_scale)
+
+    linear = torch.nn.Linear(
+        observed_linear.in_features,
+        observed_linear.out_features,
+        observed_linear.bias != None,
+        device=observed_linear.weight.device,
+        dtype=observed_linear.weight.dtype,
+    )
+    linear.weight = torch.nn.Parameter(qw, requires_grad=False)
+    linear.extra_repr = types.MethodType(_linear_extra_repr, linear)
+    linear.bias = observed_linear.bias
+    return linear