update big model ram

deploy/python/postprocess.py

@@ -71,7 +71,7 @@ def __init__(self, func_list, main_indicator="Topk"):
     def __call__(self, x, image_file=None):
         rtn = None
         for func in self.func_list:
-            tmp = func(x, image_file)
+            tmp = func(*x, image_file)
             if type(func).__name__ in self.main_indicator:
                 rtn = tmp
         return rtn
@@ -496,3 +496,61 @@ def __call__(self, batch_preds, file_names=None):
                         ).astype(np.int8).tolist()
             batch_res.append({"attributes": label_res, "output": pred_res})
         return batch_res
+
+class RamOutPut(object):
+
+    def __init__(self, language="cn",tag_list="", tag_list_chinese="", threshold=0.68, delete_tag_index=[], ram_class_threshold_path="ppcls/utils/RAM/ram_tag_list_threshold.txt"):
+        self.language = language
+        assert tag_list, tag_list_chinese
+        self.tag_list = self.load_tag_list(tag_list)
+        self.delete_tag_index = delete_tag_index
+        self.tag_list_chinese = self.load_tag_list(tag_list_chinese)
+        self.num_class = len(self.tag_list)
+        self.class_threshold = paddle.ones([self.num_class]) * threshold
+        ram_class_threshold_path = ram_class_threshold_path
+        with open(ram_class_threshold_path, "r", encoding="utf-8") as f:
+            ram_class_threshold = [float(s.strip()) for s in f]
+        for key,value in enumerate(ram_class_threshold):
+            self.class_threshold[key] = value
+
+    def load_tag_list(self, tag_list_file):
+        with open(tag_list_file, "r", encoding="utf-8") as f:
+            tag_list = f.read().splitlines()
+        tag_list = np.array(tag_list)
+        return tag_list
+
+    def __call__(self, logits, bs, file_names=None):
+        batch_res = []
+        logits = paddle.to_tensor(logits)
+        if bs is None:
+            if len(logits.shape) <2:
+                bs = 1
+            else:
+                bs = logits.shape[0]
+        targets = paddle.where(
+            F.sigmoid(logits) > self.class_threshold,
+            paddle.to_tensor([1.0]),
+            paddle.zeros(self.num_class))
+        targets = targets.reshape([bs,-1])
+        res = {}
+        tag = targets.cpu().numpy()
+        tag[:,self.delete_tag_index] = 0
+        tag_output = []
+        tag_output_chinese = []
+        for b in range(bs):
+            index = np.argwhere(tag[b] == 1)
+            token = self.tag_list[index].squeeze(axis=1)
+            tag_output.append(" | ".join(token))
+            token_chinese = self.tag_list_chinese[index].squeeze(axis=1)
+            tag_output_chinese.append(" | ".join(token_chinese))
+        res["cn"] = tag_output_chinese
+        res["en"] = tag_output
+        res["all"] = f"en : {tag_output}, cn: {tag_output_chinese}"
+
+        outputformat = {
+                "class_ids": targets.nonzero(),
+                "scores": logits,
+                "label_names": res[self.language]    
+            }
+        batch_res.append(outputformat)
+        return outputformat

deploy/python/predict_multimodal.py

@@ -0,0 +1,52 @@
+import os
+
+import cv2
+import numpy as np
+
+from paddleclas.deploy.utils import logger, config
+from paddleclas.deploy.utils.get_image_list import get_image_list
+from .predict_cls import ClsPredictor
+
+
+def main(config):
+    cls_predictor = ClsPredictor(config)
+    image_list = get_image_list(config["Global"]["infer_imgs"])
+
+    batch_imgs = []
+    batch_names = []
+    cnt = 0
+    for idx, img_path in enumerate(image_list):
+        img = cv2.imread(img_path)
+        if img is None:
+            logger.warning(
+                "Image file failed to read and has been skipped. The path: {}".
+                format(img_path))
+        else:
+            img = img[:, :, ::-1]
+            batch_imgs.append(img)
+            img_name = os.path.basename(img_path)
+            batch_names.append(img_name)
+            cnt += 1
+
+        if cnt % config["Global"]["batch_size"] == 0 or (idx + 1
+                                                         ) == len(image_list):
+            if len(batch_imgs) == 0:
+                continue
+            batch_results = cls_predictor.predict(batch_imgs)
+            for number, result_dict in enumerate(batch_results):
+                if len(batch_imgs) == 0:
+                    continue
+                batch_results = cls_predictor.predict(batch_imgs)
+                for number, result_key in enumerate(batch_results.keys()):
+                    print(f"{img_name}-{result_key}:{batch_results[result_key]}")
+            batch_imgs = []
+            batch_names = []
+    if cls_predictor.benchmark:
+        cls_predictor.auto_logger.report()
+    return
+
+
+if __name__ == "__main__":
+    args = config.parse_args()
+    config = config.get_config(args.config, overrides=args.override, show=True)
+    main(config)

ppcls/arch/backbone/__init__.py

@@ -96,6 +96,8 @@
 from .model_zoo.adaface_ir_net import AdaFace_IR_18, AdaFace_IR_34, AdaFace_IR_50, AdaFace_IR_101, AdaFace_IR_152, AdaFace_IR_SE_50, AdaFace_IR_SE_101, AdaFace_IR_SE_152, AdaFace_IR_SE_200
 from .model_zoo.wideresnet import WideResNet
 from .model_zoo.uniformer import UniFormer_small, UniFormer_small_plus, UniFormer_small_plus_dim64, UniFormer_base, UniFormer_base_ls
+from .ram.ram import ram
+from .ram.ram_plus import ram_plus
 
 
 # help whl get all the models' api (class type) and components' api (func type)