Upload folder using huggingface_hub

modeling_internvl_chat.py

@@ -17,10 +17,10 @@ from transformers.generation.streamers import BaseStreamer
 from transformers.modeling_outputs import CausalLMOutputWithPast
 from transformers.modeling_utils import PreTrainedModel
 from transformers.utils import ModelOutput, logging
+from transformers.generation.utils import GreedySearchOutput, validate_stopping_criteria, GreedySearchDecoderOnlyOutput,GreedySearchEncoderDecoderOutput
 
 from .configuration_internvl_chat import InternVLChatConfig
 from .modeling_intern_vit import InternVisionModel
-from transformers.generation.utils import GreedySearchOutput,validate_stopping_criteria,GreedySearchDecoderOnlyOutput,GreedySearchEncoderDecoderOutput
 
 logger = logging.get_logger(__name__)
 
@@ -375,16 +375,24 @@ class InternVLChatModel(PreTrainedModel):
         vit_embeds = self.mlp1(vit_embeds)
         return vit_embeds
 
-    def chat(self, tokenizer, pixel_values, question, generation_config,
+    def chat(self, tokenizer, pixel_values, question, generation_config, history=None,
              IMG_START_TOKEN='<img>', IMG_END_TOKEN='</img>', IMG_CONTEXT_TOKEN='<IMG_CONTEXT>'):
 
         img_context_token_id = tokenizer.convert_tokens_to_ids(IMG_CONTEXT_TOKEN)
         self.img_context_token_id = img_context_token_id
 
         from .conversation import get_conv_template
+
         template = get_conv_template(self.template)
-        image_tokens = IMG_START_TOKEN + IMG_CONTEXT_TOKEN * self.num_image_token + IMG_END_TOKEN
-        template.append_message(template.roles[0], image_tokens + '\n' + question)
+        if history is None:
+            history = []
+            image_tokens = IMG_START_TOKEN + IMG_CONTEXT_TOKEN * self.num_image_token + IMG_END_TOKEN
+            question = image_tokens + '\n' + question
+        else:
+            for (old_question, old_answer) in history:
+                template.append_message(template.roles[0], old_question)
+                template.append_message(template.roles[1], old_answer)
+        template.append_message(template.roles[0], question)
         template.append_message(template.roles[1], None)
         query = template.get_prompt()
         model_inputs = tokenizer(query, return_tensors='pt')
@@ -398,9 +406,8 @@ class InternVLChatModel(PreTrainedModel):
             **generation_config
         )
         response = tokenizer.batch_decode(generation_output, skip_special_tokens=True)[0]
-        query_to_print = query.replace(image_tokens, '<image>')
-        print(query_to_print, response)
-        return response
+        history.append((question, response))
+        return response, history
 
     @torch.no_grad()
     def generate(
@@ -421,6 +428,7 @@ class InternVLChatModel(PreTrainedModel):
                 vit_embeds = visual_features
             else:
                 vit_embeds = self.extract_feature(pixel_values)
+
             input_embeds = self.language_model.get_input_embeddings()(input_ids)
             B, N, C = input_embeds.shape
             input_embeds = input_embeds.reshape(B * N, C)