Add qwen2-1.5b in l0 pipeline example (#12306)

python/llm/example/NPU/HF-Transformers-AutoModels/LLM/Pipeline-Models/README.md

@@ -8,6 +8,7 @@ In this directory, you will find examples on how to directly run HuggingFace `tr
 |------------|----------------------------------------------------------------|
 | Llama2 | [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) |
 | Llama3 | [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) |
+| Qwen2 | [Qwen/Qwen2-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2-1.5B-Instruct) |
 | Qwen2.5 | [Qwen/Qwen2.5-7b-Instruct](https://huggingface.co/Qwen/Qwen2.5-7b-Instruct) |
 | Baichuan2 | [baichuan-inc/Baichuan2-7B-Chat](https://huggingface.co/baichuan-inc/Baichuan-7B-Chat) |
 | MiniCPM | [openbmb/MiniCPM-1B-sft-bf16](https://huggingface.co/openbmb/MiniCPM-1B-sft-bf16) |
@@ -50,6 +51,9 @@ python llama3.py
 :: to run Qwen2.5-7b-Instruct
 python qwen.py
 
+:: to run Qwen2-1.5b-Instruct
+python qwen.py --repo-id-or-model-path "Qwen/Qwen2-1.5B-Instruct" --load_in_low_bit "sym_int8"
+
 :: to run Baichuan2-7B-Chat
 python baichuan2.py
 

python/llm/example/NPU/HF-Transformers-AutoModels/LLM/Pipeline-Models/qwen.py

@@ -32,8 +32,8 @@ if __name__ == "__main__":
     parser.add_argument(
         "--repo-id-or-model-path",
         type=str,
-        default="Qwen/Qwen2.5-7B-Instruct",  # Or Qwen2-7B-Instruct
+        default="Qwen/Qwen2.5-7B-Instruct",  # Or Qwen2-7B-Instruct, Qwen2-1.5B-Instruct
-        help="The huggingface repo id for the Baichuan2 model to be downloaded"
+        help="The huggingface repo id for the Qwen model to be downloaded"
         ", or the path to the huggingface checkpoint folder",
     )
     parser.add_argument("--lowbit-path", type=str,
@@ -47,6 +47,8 @@ if __name__ == "__main__":
     parser.add_argument("--n-predict", type=int, default=32, help="Max tokens to predict")
     parser.add_argument("--max-context-len", type=int, default=1024)
     parser.add_argument("--max-prompt-len", type=int, default=960)
+    parser.add_argument('--load_in_low_bit', type=str, default="sym_int4",
+                        help='Load in low bit to use')
     parser.add_argument("--disable-transpose-value-cache", action="store_true", default=False)
 
     args = parser.parse_args()
@@ -56,6 +58,7 @@ if __name__ == "__main__":
         model = AutoModelForCausalLM.from_pretrained(model_path,
                                                      optimize_model=True,
                                                      pipeline=True,
+                                                     load_in_low_bit=args.load_in_low_bit,
                                                      max_context_len=args.max_context_len,
                                                      max_prompt_len=args.max_prompt_len,
                                                      torch_dtype=torch.float16,

python/llm/src/ipex_llm/transformers/npu_pipeline_model/qwen.py

@@ -27,7 +27,10 @@ def convert_lm_head_and_embedding(model, n_splits_linear, temp_dir, weight_dir):
     rms_norm_eps = model.config.rms_norm_eps
     vocab_size = model.config.vocab_size
     model_norm = model.model.norm
-    lm_heads = model.lm_head.lm_heads  # Qwen2 is always SlicedLMHead
+    if model.config.intermediate_size == 18944:
+        lm_heads = model.lm_head.lm_heads  # Qwen2-7B is always SlicedLMHead
+    else:
+        lm_heads = [model.lm_head]
     if n_splits_linear == 1:
         weights = [(lm_heads[0].weight, lm_heads[0].scale)]
     else: