Remove itrex dependency for 3x example

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/mx_quant/requirements.txt

@@ -2,6 +2,5 @@ transformers
 torch
 sentencepiece
 neural-compressor
-intel-extension-for-transformers >= 1.4.1
 lm-eval==0.4.2
 peft

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/mx_quant/run_clm_no_trainer.py

@@ -62,7 +62,7 @@ def get_user_model():
 user_model = convert(model=user_model)
 user_model.eval()
 
-from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
 eval_args = LMEvalParser(
     model="hf",
     user_model=user_model,

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/smooth_quant/requirements.txt

@@ -8,7 +8,6 @@ pytest
 wandb
 einops
 neural-compressor
-intel-extension-for-transformers
-lm_eval==0.4.2
+lm_eval==0.4.3
 peft
 optimum-intel

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/smooth_quant/run_clm_no_trainer.py

@@ -185,7 +185,7 @@ def eval_func(model):
         config = AutoConfig.from_pretrained(args.model)
         setattr(model, "config", config)
 
-        from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+        from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
         eval_args = LMEvalParser(
             model="hf",
             user_model=model,
@@ -232,7 +232,7 @@ def eval_func(model):
 
 if args.accuracy:
     user_model.eval()
-    from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+    from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
 
     eval_args = LMEvalParser(
         model="hf",

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/static_quant/ipex/requirements.txt

@@ -8,6 +8,5 @@ pytest
 wandb
 einops
 neural-compressor
-intel-extension-for-transformers
-lm_eval==0.4.2
+lm_eval==0.4.3
 peft

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/static_quant/ipex/run_clm_no_trainer.py

@@ -212,7 +212,7 @@ def run_fn(model):
 
 if args.accuracy:
     user_model.eval()
-    from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+    from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
     eval_args = LMEvalParser(
         model="hf",
         user_model=user_model,
@@ -232,7 +232,7 @@ def run_fn(model):
 
 if args.performance:
     user_model.eval()
-    from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+    from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
     import time
 
     samples = args.iters * args.batch_size

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/static_quant/pt2e/requirements.txt

@@ -2,6 +2,5 @@ transformers
 torch
 sentencepiece
 neural-compressor
-intel-extension-for-transformers >= 1.4.1
-lm-eval==0.4.2
+lm-eval==0.4.3
 peft

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/static_quant/pt2e/run_clm_no_trainer.py

@@ -116,7 +116,7 @@ def get_example_inputs(tokenizer):
 
 if args.accuracy:
 
-    from intel_extension_for_transformers.transformers.llm.evaluation.lm_eval import evaluate, LMEvalParser
+    from neural_compressor.evaluation.lm_eval import evaluate, LMEvalParser
     eval_args = LMEvalParser(
         model="hf",
         user_model=user_model,

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/transformers/weight_only/text-generation/llm_quantization_recipes.md

@@ -30,12 +30,6 @@ The scripts [run_generation_sq.py](./run_generation_sq.py) and [run_generation_c
 
 ```bash
 # Installation
-git clone https://github.com/intel/intel-extension-for-transformers.git
-
-# install ITREX
-cd intel-extension-for-transformers
-pip install -r requirements.txt
-pip install -v .
 
 # install requirements
 cd examples/huggingface/pytorch/text-generation/quantization

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/transformers/weight_only/text-generation/run_generation_cpu_woq.py

@@ -293,7 +293,6 @@
         _commit_hash=args._commit_hash,
     )
 elif args.load_in_4bit or args.load_in_8bit:
-    # CPU device usage is provided by intel-extension-for-transformers.
     user_model = AutoModelForCausalLM.from_pretrained(
         args.model,
         load_in_4bit=args.load_in_4bit,

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/transformers/weight_only/text-generation/run_generation_gpu_woq.py

@@ -6,7 +6,6 @@
 from transformers import AutoConfig, AutoTokenizer
 from transformers.generation import GenerationConfig
 import intel_extension_for_pytorch as ipex
-# from intel_extension_for_transformers.transformers.llm.utils.generation import _beam_search, _greedy_search
 from neural_compressor.transformers import AutoModelForCausalLM, AutoRoundConfig, RtnConfig, GPTQConfig
 from neural_compressor.transformers.quantization.utils import convert_dtype_str2torch
 from transformers.utils import check_min_version
@@ -189,7 +188,6 @@
                                                       torch_dtype=torch.float16,
                                                       )
 elif args.load_in_4bit or args.load_in_8bit:
-    # CPU device usage is provided by intel-extension-for-transformers.
     user_model = AutoModelForCausalLM.from_pretrained(args.model,
                                                       device_map=args.device,
                                                       load_in_4bit=args.load_in_4bit,

examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/weight_only/requirements.txt

@@ -8,7 +8,6 @@ pytest
 wandb
 einops
 neural-compressor
-intel-extension-for-transformers
 lm_eval==0.4.3
 peft
 auto_round