update readme with dc2

README.md

@@ -25,42 +25,35 @@ Read the text 'kids feeding and playing with the horse' to generate a text embed
  *Write the pipeline in simplified style*:
 
 ```python
-import towhee
+from towhee.dc2 import pipe, ops, DataCollection
 
-towhee.dc(['./demo_video.mp4']) \
-    .video_decode.ffmpeg(sample_type='uniform_temporal_subsample', args={'num_samples': 12}) \
-    .runas_op(func=lambda x: [y for y in x]) \
-    .video_text_embedding.drl(base_encoder='clip_vit_b32', modality='video', device='cpu') \
-    .show()
+p = (
+    pipe.input('text') \
+        .map('text', 'vec', ops.video_text_embedding.drl(base_encoder='clip_vit_b32', modality='text', device='cuda:0')) \
+        .output('text', 'vec')
+)
 
-towhee.dc(['kids feeding and playing with the horse']) \
-    .video_text_embedding.drl(base_encoder='clip_vit_b32', modality='text', device='cpu') \
-    .show()
+DataCollection(p('kids feeding and playing with the horse')).show()
 ```
 
-![](vect_simplified_video.png)   
-![](vect_simplified_text.png)   
+![](text_emb_result.png)    
 
-*Write a same pipeline with explicit inputs/outputs name specifications:*
 
 ```python
-import towhee
-
-towhee.dc['path'](['./demo_video.mp4']) \
-        .video_decode.ffmpeg['path', 'frames'](sample_type='uniform_temporal_subsample', args={'num_samples': 12}) \
-        .runas_op['frames', 'frames'](func=lambda x: [y for y in x]) \
-        .video_text_embedding.drl['frames', 'vec'](base_encoder='clip_vit_b32', modality='video', device='cpu') \
-        .show(formatter={'path': 'video_path'})
-
-towhee.dc['text'](['kids feeding and playing with the horse']) \
-      .video_text_embedding.drl['text','vec'](base_encoder='clip_vit_b32', modality='text', device='cpu') \
-      .select['text', 'vec']() \
-      .show()
-```
+from towhee.dc2 import pipe, ops, DataCollection
+
+p = (
+    pipe.input('video_path') \
+        .map('video_path', 'flame_gen', ops.video_decode.ffmpeg(sample_type='uniform_temporal_subsample', args={'num_samples': 12})) \
+        .map('flame_gen', 'flame_list', lambda x: [y for y in x]) \
+        .map('flame_list', 'vec', ops.video_text_embedding.drl(base_encoder='clip_vit_b32', modality='video', device='cuda:0')) \
+        .output('video_path', 'flame_list', 'vec')
+)
 
-![](vect_explicit_video.png)    
-![](vect_explicit_text.png)   
+DataCollection(p('./demo_video.mp4')).show()
+```
 
+![](video_emb_result.png)    
 <br />
 
 

drl.py

@@ -41,7 +41,7 @@ class DRL(NNOperator):
             self.device = "cuda" if torch.cuda.is_available() else "cpu"
         else:
             self.device = device
-        self.model = drl.create_model(base_encoder=base_encoder, pretrained=True, cdcr=0, weights_path=weight_path)
+        self.model = drl.create_model(base_encoder=base_encoder, pretrained=True, cdcr=0, weights_path=weight_path, device=device)
 
         self.tokenize = clip4clip.SimpleTokenizer()
         self.tfms = transforms.Compose([