[FIX] not pack when group_size=-1 (#1298)

* Fix skipping pack() when group_size = -1

* assert len(qModules) > 0

* Update __init__.py

* Update __init__.py

---------

Co-authored-by: Qubitium-ModelCloud <[email protected]>

Author: ZX-ModelCloud

examples/quantization/basic_usage_wikitext2.py

@@ -68,9 +68,6 @@ def main():
     # with value under torch.LongTensor type.
     model.quantize(traindataset)
 
-    # save quantized model
-    model.save(quantized_model_id)
-
     # save quantized model using safetensors
     model.save(quantized_model_id)
 

gptqmodel/models/definitions/minicpm.py

@@ -29,5 +29,4 @@ class MiniCPMGPTQ(BaseGPTQModel):
         ["self_attn.v_proj"],
         ["self_attn.o_proj"],
         ["mlp.gate_proj", "mlp.up_proj","mlp.down_proj"],
-        ["mlp.c_proj"],
     ]

gptqmodel/nn_modules/qlinear/__init__.py

@@ -205,7 +205,8 @@ def _validate(cls, bits: int=4, group_size: int=128, desc_act: bool=False, sym:
         if bits not in cls.SUPPORTS_BITS:
             err = f"{cls} only supports `{cls.SUPPORTS_BITS}` bits: actual bits = `{bits}`"
             return False, NotImplementedError(err)
-        if group_size not in cls.SUPPORTS_GROUP_SIZE:
+        # valid group size is set of cls.SUPPORTS_GROUP_SIZE + in_features; group_size = -1 is alias for group_size == in_features
+        if group_size not in cls.SUPPORTS_GROUP_SIZE and group_size != in_features:
             err = f"{cls} only supports `{cls.SUPPORTS_GROUP_SIZE}` group_size: actual group_size = `{group_size}`"
             return False, NotImplementedError(err)
         if sym not in cls.SUPPORTS_SYM:
@@ -365,3 +366,5 @@ def pack(self, linear, scales, zeros, g_idx=None):
                 col += 1
 
         self.qzeros = t.from_numpy(qzeros.astype(self.pack_np_dtype))
+
+        # print("self qw", self.qweight, self.scales, self.qzeros)

gptqmodel/utils/model.py

@@ -156,7 +156,7 @@ def make_quant(
     device: DEVICE = None,
     from_quantized: bool = False,
     pack_dtype: torch.dtype = None,
-) -> BaseQuantLinear:
+) -> Type[BaseQuantLinear]:
     # returns multiple validated kernels
     quant_linear_candidates = select_quant_linear(
         bits=bits,
@@ -175,15 +175,15 @@ def make_quant(
     logger.info(f"make_quant: Linear candidates: {quant_linear_candidates}")
 
     # loop over actual QLinear init, catch errors and use fallbacks if applicable
-    for linear in quant_linear_candidates:
+    for cls in quant_linear_candidates:
         try:
             # if linear is not selectedQLinear:
             #     logger.info(f"make_quant: Faild linear: `{selectedQLinear}` failed, trying to use fallback: `{linear}`")
             # else:
             #     logger.info("make_quant: Testing linear: {linear}")
 
-            linear_instance = create_quant_layer(
-                linear=linear,
+            linear_cls = create_quant_layer(
+                linear_cls=cls,
                 bits=bits,
                 desc_act=desc_act,
                 dynamic=dynamic,
@@ -194,10 +194,10 @@ def make_quant(
                 device=device,
                 lm_head_name=lm_head_name,
                 pack_dtype=pack_dtype)
-            logger.info(f"make_quant: Selected linear: `{linear}`.")
-            return linear_instance
+            logger.info(f"make_quant: Selected linear: `{cls}`.")
+            return linear_cls
         except NotImplementedError as e:
-            logger.info(f"make_quant: Skipped linear: `{linear}`.")
+            logger.info(f"make_quant: Skipped linear: `{cls}`. ")
             # only fallback to other quant linears when backend is auto.
             if backend not in [BACKEND.AUTO, BACKEND.AUTO_TRAINABLE]:
                 raise e
@@ -206,7 +206,7 @@ def make_quant(
 
 
 def create_quant_layer(
-        linear: nn.Module,
+        linear_cls: Type[BaseQuantLinear],
         bits: int,
         desc_act: bool,
         dynamic,
@@ -216,10 +216,9 @@ def create_quant_layer(
         sym: bool,
         device: DEVICE,
         lm_head_name: str,
-        pack_dtype: torch.dtype,
-                       ) -> BaseQuantLinear:
-    if isinstance(module, linear):
-        return linear
+        pack_dtype: torch.dtype) -> Type[BaseQuantLinear]:
+    if isinstance(module, linear_cls):
+        return linear_cls
     for name, submodule in module.named_modules():
         if name in names:
             ori_layer_device = next(submodule.parameters()).device
@@ -266,7 +265,7 @@ def create_quant_layer(
 
             # when loading a quantized model, device is target device passed in GPTQModel.load()
             # check in_features and out_features validate
-            _, err = linear.validate(
+            _, err = linear_cls.validate(
                 bits=tmp_bits,
                 group_size=tmp_group_size,
                 desc_act=tmp_desc_act,
@@ -278,7 +277,7 @@ def create_quant_layer(
             if err is not None:
                 raise err
 
-            new_layer = linear(
+            new_layer = linear_cls(
                 bits=tmp_bits,
                 group_size=tmp_group_size,
                 desc_act=tmp_desc_act,
@@ -293,7 +292,7 @@ def create_quant_layer(
             )
             new_layer.device = ori_layer_device
             recurse_setattr(module, name, new_layer.to(ori_layer_device))
-    return linear
+    return linear_cls
 
 # public/stable api exposed to transformer/optimum
 def hf_convert_gptq_v1_to_v2_format(
@@ -489,25 +488,13 @@ def pack_model(
     parallel_packing: bool = True,
     pack_dtype: torch.dtype = None,
 ):
-    quantLinear = select_quant_linear(
-        bits=bits,
-        dynamic=dynamic,
-        group_size=group_size,
-        desc_act=desc_act,
-        sym=sym,
-        backend=backend,
-        format=format,
-        pack=True,
-        pack_dtype=pack_dtype,
-    )
-
     model.to(CPU)
 
     logger.info("Packing model...")
 
     modules = find_modules(model)
     modules = {n: modules[n] for n in quantizers}
-    make_quant(
+    quant_linear_cls = make_quant(
         model,
         quantizers,
         bits,
@@ -520,7 +507,10 @@ def pack_model(
         dynamic=dynamic,
         pack_dtype=pack_dtype,
     )
-    qModules = find_modules(model, [quantLinear])
+    qModules = find_modules(model, [quant_linear_cls])
+
+    assert len(qModules) > 0, f"No quantizeed modules[{quant_linear_cls}] found in the model."
+
     names = list(qModules.keys())
 
     if parallel_packing:
@@ -537,7 +527,7 @@ def wrapper(name):
                 pass
 
     logger.info("Model packed.")
-    return quantLinear
+    return quant_linear_cls
 
 
 def verify_model_hash(file_path: str, verify_hash: str):