[INFER][LLM] Support qwen in fined grained dybatch v1

llm/predictor.py

@@ -859,8 +859,19 @@ def create_predictor(
                     dtype=predictor_args.dtype,
                 )
                 model.eval()
+            elif "qwen" in config.architectures[0].lower():
+                from paddlenlp.experimental.transformers import (
+                    QWenForCausalLMInferenceModel,
+                )
+
+                model = QWenForCausalLMInferenceModel.from_pretrained(
+                    predictor_args.model_name_or_path,
+                    config=config,
+                    dtype=predictor_args.dtype,
+                )
+                model.eval()
             else:
-                raise ValueError("the `model type` should be one of [llama, chatglm, bloom, gpt]")
+                raise ValueError("the `model type` should be one of [llama, chatglm, bloom, gpt, qwen]")
             predictor = DygraphInferencePredictor(predictor_args, model=model, tokenizer=tokenizer)
         elif predictor_args.mode == "static":
             config = AutoConfig.from_pretrained(predictor_args.model_name_or_path)
@@ -904,8 +915,16 @@ def create_predictor(
                 cache_kvs_shape = GPTForCausalLMInferenceModel.get_cache_kvs_shape(
                     config, predictor_args.batch_size, predictor_args.total_max_length
                 )
+            elif "qwen" in config.architectures[0].lower():
+                from paddlenlp.experimental.transformers import (
+                    QWenForCausalLMInferenceModel,
+                )
+
+                cache_kvs_shape = QWenForCausalLMInferenceModel.get_cache_kvs_shape(
+                    config, predictor_args.batch_size, predictor_args.total_max_length
+                )
             else:
-                raise ValueError("the `model type` should be one of [llama, chatglm, bloom, gpt]")
+                raise ValueError("the `model type` should be one of [llama, chatglm, bloom, gpt, qwen]")
             predictor = StaticInferencePredictor(predictor_args, cache_kvs_shape, tokenizer=tokenizer)
         else:
             raise ValueError("the `mode` should be one of [dynamic, static]")

paddlenlp/experimental/transformers/__init__.py

@@ -19,3 +19,4 @@
 from .gpt import *
 from .llama import *
 from .opt import *
+from .qwen import *

paddlenlp/experimental/transformers/qwen/__init__.py

@@ -0,0 +1,15 @@
+# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from .modeling import *