[PIR] Update simcse to apply pir (#10396)

slm/applications/neural_search/recall/simcse/README.md

@@ -149,6 +149,12 @@ simcse/
 
 <a name="模型训练"></a>
 
+下载数据集并解压到当前目录：
+```shell
+wget https://bj.bcebos.com/v1/paddlenlp/data/literature_search_data.zip
+unzip literature_search_data.zip
+```
+
 ## 5. 模型训练
 
 **语义索引预训练模型下载链接：**

slm/applications/neural_search/recall/simcse/deploy/python/deploy.sh

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-python predict.py --model_dir=../../output
+python deploy/python/predict.py --model_dir=./output

slm/applications/neural_search/recall/simcse/deploy/python/predict.py

@@ -16,12 +16,17 @@ import argparse
 import os
 import sys
 
+import numpy as np
 import paddle
 from paddle import inference
 from scipy import spatial
 
 from paddlenlp.data import Pad, Tuple
 from paddlenlp.transformers import AutoTokenizer
+from paddlenlp.utils.env import (
+    PADDLE_INFERENCE_MODEL_SUFFIX,
+    PADDLE_INFERENCE_WEIGHTS_SUFFIX,
+)
 from paddlenlp.utils.log import logger
 
 sys.path.append(".")
@@ -90,8 +95,8 @@ class Predictor(object):
         self.max_seq_length = max_seq_length
         self.batch_size = batch_size
 
-        model_file = model_dir + "/inference.get_pooled_embedding.pdmodel"
-        params_file = model_dir + "/inference.get_pooled_embedding.pdiparams"
+        model_file = model_dir + f"/inference{PADDLE_INFERENCE_MODEL_SUFFIX}"
+        params_file = model_dir + f"/inference{PADDLE_INFERENCE_WEIGHTS_SUFFIX}"
         if not os.path.exists(model_file):
             raise ValueError("not find model file path {}".format(model_file))
         if not os.path.exists(params_file):
@@ -238,6 +243,9 @@ class Predictor(object):
 
         if args.benchmark:
             self.autolog.times.end(stamp=True)
+
+        query_logits = np.atleast_2d(query_logits)
+        title_logits = np.atleast_2d(title_logits)
         result = [float(1 - spatial.distance.cosine(arr1, arr2)) for arr1, arr2 in zip(query_logits, title_logits)]
         return result