Control language for XLM

examples/run_generation.py

@@ -106,6 +106,8 @@ def prepare_xlm_input(args, model, tokenizer, prompt_text):
             language = None
             while language not in available_languages:
                 language = input("Using XLM. Select language in " + str(list(available_languages)) + " >>> ")
+
+        model.config.lang_id = model.config.lang2id[language]
         # kwargs["language"] = tokenizer.lang2id[language]
 
     # TODO fix mask_token_id setup when configurations will be synchronized between models and tokenizers
@@ -119,12 +121,12 @@ def prepare_xlm_input(args, model, tokenizer, prompt_text):
 
 def prepare_xlnet_input(args, _, tokenizer, prompt_text):
     prompt_text = (args.padding_text if args.padding_text else PADDING_TEXT) + prompt_text
-    return prompt_text, {}
+    return prompt_text
 
 
 def prepare_transfoxl_input(args, _, tokenizer, prompt_text):
     prompt_text = (args.padding_text if args.padding_text else PADDING_TEXT) + prompt_text
-    return prompt_text, {}
+    return prompt_text
 
 
 PREPROCESSING_FUNCTIONS = {
@@ -183,6 +185,7 @@ def main():
 
     parser.add_argument("--seed", type=int, default=42, help="random seed for initialization")
     parser.add_argument("--no_cuda", action="store_true", help="Avoid using CUDA when available")
+    parser.add_argument("--num_return_sequences", type=int, default=1, help="The number of samples to generate.")
     args = parser.parse_args()
 
     args.device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
@@ -210,28 +213,51 @@ def main():
     requires_preprocessing = args.model_type in PREPROCESSING_FUNCTIONS.keys()
     if requires_preprocessing:
         prepare_input = PREPROCESSING_FUNCTIONS.get(args.model_type)
-        prompt_text = prepare_input(args, model, tokenizer, prompt_text)
-    encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False, return_tensors="pt")
+        preprocessed_prompt_text = prepare_input(args, model, tokenizer, prompt_text)
+        encoded_prompt = tokenizer.encode(preprocessed_prompt_text, add_special_tokens=False, return_tensors="pt")
+    else:
+        encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False, return_tensors="pt")
     encoded_prompt = encoded_prompt.to(args.device)
 
     output_sequences = model.generate(
         input_ids=encoded_prompt,
-        max_length=args.length,
+        max_length=args.length + len(encoded_prompt[0]),
         temperature=args.temperature,
         top_k=args.k,
         top_p=args.p,
         repetition_penalty=args.repetition_penalty,
         do_sample=True,
+        num_return_sequences=args.num_return_sequences,
     )
 
-    # Batch size == 1. to add more examples please use num_return_sequences > 1
-    generated_sequence = output_sequences[0].tolist()
-    text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
-    text = text[: text.find(args.stop_token) if args.stop_token else None]
+    # Remove the batch dimension when returning multiple sequences
+    if len(output_sequences.shape) > 2:
+        output_sequences.squeeze_()
 
-    print(text)
+    generated_sequences = []
 
-    return text
+    for generated_sequence in output_sequences:
+        print("=== GENERATED SEQUENCE ===")
+        generated_sequence = generated_sequence.tolist()
+
+        # Remove "-1" tokens which are used for padding
+        generated_sequence = [token for token in generated_sequence if token != -1]
+
+        # Decode text
+        text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
+
+        # Remove all text after the stop token
+        text = text[: text.find(args.stop_token) if args.stop_token else None]
+
+        # Add the prompt at the beginning of the sequence. Remove the excess text that was used for pre-processing
+        total_sequence = (
+            prompt_text + text[len(tokenizer.decode(encoded_prompt[0], clean_up_tokenization_spaces=True)) :]
+        )
+
+        generated_sequences.append(total_sequence)
+        print(total_sequence)
+
+    return generated_sequences if len(generated_sequences) > 1 else generated_sequences[0]
 
 
 if __name__ == "__main__":

src/transformers/configuration_utils.py

@@ -75,9 +75,9 @@ class PretrainedConfig(object):
         self.top_k = kwargs.pop("top_k", 50)
         self.top_p = kwargs.pop("top_p", 1.0)
         self.repetition_penalty = kwargs.pop("repetition_penalty", 1.0)
-        self.bos_token_id = kwargs.pop("bos_token_id", 0)
-        self.pad_token_id = kwargs.pop("pad_token_id", 0)
-        self.eos_token_ids = kwargs.pop("eos_token_ids", 0)
+        self.bos_token_id = kwargs.pop("bos_token_id", None)
+        self.pad_token_id = kwargs.pop("pad_token_id", None)
+        self.eos_token_ids = kwargs.pop("eos_token_ids", None)
         self.length_penalty = kwargs.pop("length_penalty", 1.0)
         self.num_return_sequences = kwargs.pop("num_return_sequences", 1)
 

src/transformers/modeling_utils.py

@@ -712,10 +712,14 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin):
         assert isinstance(top_k, int) and top_k >= 0, "`top_k` should be a positive integer."
         assert 0 <= top_p <= 1, "`top_p` should be between 0 and 1."
         assert repetition_penalty >= 1.0, "`repetition_penalty` should be >= 1."
-        assert isinstance(bos_token_id, int) and bos_token_id >= 0, "`bos_token_id` should be a positive integer."
-        assert isinstance(pad_token_id, int) and pad_token_id >= 0, "`pad_token_id` should be a positive integer."
-        assert isinstance(eos_token_ids, (list, tuple)) and (
-            e >= 0 for e in eos_token_ids
+        assert bos_token_id is None or (
+            isinstance(bos_token_id, int) and bos_token_id >= 0
+        ), "`bos_token_id` should be a positive integer."
+        assert (
+            pad_token_id is None or isinstance(pad_token_id, int) and pad_token_id >= 0
+        ), "`pad_token_id` should be a positive integer."
+        assert (
+            eos_token_ids is None or isinstance(eos_token_ids, (list, tuple)) and (e >= 0 for e in eos_token_ids)
         ), "`eos_token_ids` should be a positive integer or a list/tuple of positive integers."
         assert length_penalty > 0, "`length_penalty` should be strictely positive."
         assert (
@@ -729,6 +733,13 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin):
         else:
             assert input_ids.dim() == 2, "Input prompt should be of shape (batch_size, sequence length)."
 
+        if pad_token_id is None:
+            pad_token_id = -1
+        if eos_token_ids is None:
+            eos_token_ids = [-1]
+        if bos_token_id is None:
+            bos_token_id = -1
+
         # current position and vocab size
         cur_len = input_ids.shape[1]
         vocab_size = self.config.vocab_size