[Generation] Generation should allow to start with empty prompt (#3993)

* fix empty prompt

* fix length in generation pipeline

examples/run_generation.py

@@ -221,8 +221,13 @@ def main():
         encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False, return_tensors="pt")
     encoded_prompt = encoded_prompt.to(args.device)
 
+    if encoded_prompt.size()[-1] == 0:
+        input_ids = None
+    else:
+        input_ids = encoded_prompt
+
     output_sequences = model.generate(
-        input_ids=encoded_prompt,
+        input_ids=input_ids,
         max_length=args.length + len(encoded_prompt[0]),
         temperature=args.temperature,
         top_k=args.k,

src/transformers/pipelines.py

@@ -563,14 +563,19 @@ class TextGenerationPipeline(Pipeline):
                 else:
                     inputs = self._parse_and_tokenize(prompt_text)
 
-                if self.framework == "pt":
+                # set input_ids to None to allow empty prompt
+                if inputs["input_ids"].shape[-1] == 0:
+                    inputs["input_ids"] = None
+                    inputs["attention_mask"] = None
+
+                if self.framework == "pt" and inputs["input_ids"] is not None:
                     inputs = self.ensure_tensor_on_device(**inputs)
 
                 input_ids = inputs["input_ids"]
 
                 # Ensure that batch size = 1 (batch generation not allowed for now)
                 assert (
-                    input_ids.shape[0] == 1
+                    input_ids is None or input_ids.shape[0] == 1
                 ), "Batch generation is currently not supported. See https://github.com/huggingface/transformers/issues/3021 for more information."
 
                 output_sequences = self.model.generate(input_ids=input_ids, **generate_kwargs)  # BS x SL
@@ -590,19 +595,19 @@ class TextGenerationPipeline(Pipeline):
                     )
 
                     # Remove PADDING prompt of the sequence if XLNet or Transfo-XL model is used
-                    record["generated_text"] = (
+                    if input_ids is None:
-                        prompt_text
+                        prompt_length = 0
-                        + text[
+                    else:
-                            len(
+                        prompt_length = len(
                             self.tokenizer.decode(
                                 input_ids[0],
                                 skip_special_tokens=True,
                                 clean_up_tokenization_spaces=clean_up_tokenization_spaces,
                             )
-                            ) :
-                        ]
                         )
 
+                    record["generated_text"] = prompt_text + text[prompt_length:]
+
                 result.append(record)
             results += [result]