From 2761cac1ad66296c317a58cbc1818a28390ba9ca Mon Sep 17 00:00:00 2001
From: Ella Charlaix <ella@huggingface.co>
Date: Thu, 21 Mar 2024 14:14:08 +0100
Subject: [PATCH 1/2] Deprecate IncQuantizedModel classes

---
 .../intel/neural_compressor/modeling_base.py  |  6 ---
 .../intel/neural_compressor/quantization.py   | 47 -------------------
 2 files changed, 53 deletions(-)

diff --git a/optimum/intel/neural_compressor/modeling_base.py b/optimum/intel/neural_compressor/modeling_base.py
index 72646a9f94..5a7c5bb44d 100644
--- a/optimum/intel/neural_compressor/modeling_base.py
+++ b/optimum/intel/neural_compressor/modeling_base.py
@@ -84,7 +84,6 @@ def __init__(
         self._device = getattr(self.model, "device", None) or torch.device(
             "cuda:0" if torch.cuda.is_available() else "cpu"
         )
-
         # Registers the INCModelForXXX classes into the transformers AutoModel classes to avoid warnings when creating
         # a pipeline https://github.com/huggingface/transformers/blob/cad61b68396a1a387287a8e2e2fef78a25b79383/src/transformers/pipelines/base.py#L863
         AutoConfig.register(self.base_model_prefix, AutoConfig)
@@ -247,11 +246,6 @@ class INCModelForVision2Seq(INCModel):
     export_feature = "image-to-text"
 
 
-class INCModelForXLNetLM(INCModel):
-    auto_model_class = XLNetLMHeadModel
-    export_feature = "fill-mask"
-
-
 class INCModelForCausalLM(INCModel, BaseModelForCausalLM):
     auto_model_class = AutoModelForCausalLM
     export_feature = "text-generation"
diff --git a/optimum/intel/neural_compressor/quantization.py b/optimum/intel/neural_compressor/quantization.py
index d4846adc15..94cb3c1d9a 100644
--- a/optimum/intel/neural_compressor/quantization.py
+++ b/optimum/intel/neural_compressor/quantization.py
@@ -71,7 +71,6 @@
     INCModelForSequenceClassification,
     INCModelForTokenClassification,
     INCModelForVision2Seq,
-    INCModelForXLNetLM,
 )
 from .utils import INCDataLoader, _cfgs_to_fx_cfgs
 
@@ -538,49 +537,3 @@ def _apply_quantization_from_config(q_config: Dict, model: torch.nn.Module) -> t
         q_model = convert(q_model, mapping=q_mapping, inplace=True)
 
     return q_model
-
-
-class IncQuantizedModel(INCModel):
-    @classmethod
-    def from_pretrained(cls, *args, **kwargs):
-        warnings.warn(
-            f"The class `{cls.__name__}` has been depreciated and will be removed in optimum-intel v1.12, please use "
-            f"`{cls.__name__.replace('IncQuantized', 'INC')}` instead."
-        )
-        return super().from_pretrained(*args, **kwargs)
-
-
-class IncQuantizedModelForQuestionAnswering(IncQuantizedModel):
-    auto_model_class = AutoModelForQuestionAnswering
-
-
-class IncQuantizedModelForSequenceClassification(IncQuantizedModel):
-    auto_model_class = AutoModelForSequenceClassification
-
-
-class IncQuantizedModelForTokenClassification(IncQuantizedModel):
-    auto_model_class = AutoModelForTokenClassification
-
-
-class IncQuantizedModelForMultipleChoice(IncQuantizedModel):
-    auto_model_class = AutoModelForMultipleChoice
-
-
-class IncQuantizedModelForSeq2SeqLM(IncQuantizedModel):
-    auto_model_class = AutoModelForSeq2SeqLM
-
-
-class IncQuantizedModelForCausalLM(IncQuantizedModel):
-    auto_model_class = AutoModelForCausalLM
-
-
-class IncQuantizedModelForMaskedLM(IncQuantizedModel):
-    auto_model_class = AutoModelForMaskedLM
-
-
-class IncQuantizedModelForXLNetLM(IncQuantizedModel):
-    auto_model_class = XLNetLMHeadModel
-
-
-class IncQuantizedModelForVision2Seq(IncQuantizedModel):
-    auto_model_class = AutoModelForVision2Seq

From 7586d05212f10a1146d6f360b8de43f964783433 Mon Sep 17 00:00:00 2001
From: Ella Charlaix <ella@huggingface.co>
Date: Thu, 21 Mar 2024 14:14:56 +0100
Subject: [PATCH 2/2] fix style

---
 optimum/intel/neural_compressor/modeling_base.py |  1 -
 optimum/intel/neural_compressor/quantization.py  | 10 ----------
 2 files changed, 11 deletions(-)

diff --git a/optimum/intel/neural_compressor/modeling_base.py b/optimum/intel/neural_compressor/modeling_base.py
index 5a7c5bb44d..b014f8a476 100644
--- a/optimum/intel/neural_compressor/modeling_base.py
+++ b/optimum/intel/neural_compressor/modeling_base.py
@@ -34,7 +34,6 @@
     AutoModelForVision2Seq,
     GenerationMixin,
     PretrainedConfig,
-    XLNetLMHeadModel,
 )
 from transformers.modeling_utils import no_init_weights
 from transformers.models.auto.auto_factory import _get_model_class
diff --git a/optimum/intel/neural_compressor/quantization.py b/optimum/intel/neural_compressor/quantization.py
index 94cb3c1d9a..5eb4da8cef 100644
--- a/optimum/intel/neural_compressor/quantization.py
+++ b/optimum/intel/neural_compressor/quantization.py
@@ -15,7 +15,6 @@
 import copy
 import inspect
 import logging
-import warnings
 from enum import Enum
 from itertools import chain
 from pathlib import Path
@@ -31,18 +30,9 @@
 from neural_compressor.quantization import fit
 from torch.utils.data import DataLoader, RandomSampler
 from transformers import (
-    AutoModelForCausalLM,
-    AutoModelForMaskedLM,
-    AutoModelForMultipleChoice,
-    AutoModelForQuestionAnswering,
-    AutoModelForSeq2SeqLM,
-    AutoModelForSequenceClassification,
-    AutoModelForTokenClassification,
-    AutoModelForVision2Seq,
     DataCollator,
     PretrainedConfig,
     PreTrainedModel,
-    XLNetLMHeadModel,
     default_data_collator,
 )