初始化项目，由ModelHub XC社区提供模型

Model: AI-ModelScope/opt-1b3
Source: Original Platform

ms_wrapper.py

@@ -0,0 +1,60 @@
+import os
+from typing import Union, Dict, Any
+from modelscope.pipelines.builder import PIPELINES
+from modelscope.models.builder import MODELS
+from modelscope.utils.constant import Tasks
+from modelscope.pipelines.base import Pipeline
+from modelscope.models.base import Model, TorchModel
+from modelscope.utils.logger import get_logger
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+
+os.environ['CUDA_VISIBLE_DEVICES'] = "0"
+
+@PIPELINES.register_module(Tasks.text_generation, module_name='opt-1b3-text-generation-pipe')
+class Opt1b3TextGenerationPipeline(Pipeline):
+    def __init__(
+            self,
+            model: Union[Model, str],
+            *args,
+            **kwargs):
+        model = Opt1b3TextGeneration(model) if isinstance(model, str) else model
+        super().__init__(model=model, **kwargs)
+    
+    def preprocess(self, inputs, **preprocess_params) -> Dict[str, Any]:
+        return inputs
+    
+    # define the forward pass
+    def forward(self, inputs: Dict, **forward_params) -> Dict[str, Any]:
+        return self.model(inputs)
+    
+    # format the outputs from pipeline
+    def postprocess(self, input, **kwargs) -> Dict[str, Any]:
+        return input
+
+@MODELS.register_module(Tasks.text_generation, module_name='opt-1b3')
+class Opt1b3TextGeneration(TorchModel):
+    def __init__(self, model_dir=None, *args, **kwargs):
+        super().__init__(model_dir, *args, **kwargs)
+        self.logger = get_logger()
+        # loading tokenizer
+        self.tokenizer = AutoTokenizer.from_pretrained(model_dir)
+        self.model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto")
+        self.model = self.model.eval()
+    
+    def forward(self,input: Dict, *args, **kwargs) -> Dict[str, Any]:
+        output = {}
+        res = self.infer(input)
+        output['text'] = res
+        return output
+    
+    def quantize(self, bits: int):
+        self.model = self.model.quantize(bits)
+        return self
+    
+    def infer(self, input):
+        device = self.model.device
+        input_ids = self.tokenizer(input, return_tensors="pt").input_ids.to(device)
+        logits = self.model.generate(input_ids, num_beams=1, max_length=512)
+        out = self.tokenizer.decode(logits[0].tolist())
+        return out