Fix GPT-J onnx conversion (#16780)

* add gptj to TOKENIZER_MAPPING_NAMES

* fix int32 to float to avoid problem in onnx

* Update src/transformers/models/gptj/modeling_gptj.py
Co-authored-by: ChainYo <t.chaigneau.tc@gmail.com>
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

src/transformers/models/auto/tokenization_auto.py

@@ -136,6 +136,7 @@ else:
             ("bert", ("BertTokenizer", "BertTokenizerFast" if is_tokenizers_available() else None)),
             ("openai-gpt", ("OpenAIGPTTokenizer", "OpenAIGPTTokenizerFast" if is_tokenizers_available() else None)),
             ("gpt2", ("GPT2Tokenizer", "GPT2TokenizerFast" if is_tokenizers_available() else None)),
+            ("gptj", ("GPT2Tokenizer", "GPT2TokenizerFast" if is_tokenizers_available() else None)),
             ("transfo-xl", ("TransfoXLTokenizer", None)),
             (
                 "xlnet",

src/transformers/models/gptj/modeling_gptj.py

@@ -60,7 +60,9 @@ def fixed_pos_embedding(x, seq_dim=1, seq_len=None):
     if seq_len is None:
         seq_len = x.shape[seq_dim]
     inv_freq = 1.0 / (10000 ** (torch.arange(0, dim, 2) / dim))
-    sinusoid_inp = torch.einsum("i , j -> i j", torch.arange(seq_len), inv_freq).to(x.device).float()
+    sinusoid_inp = (
+        torch.einsum("i , j -> i j", torch.arange(seq_len, dtype=torch.float), inv_freq).to(x.device).float()
+    )
     return torch.sin(sinusoid_inp), torch.cos(sinusoid_inp)
 
 

src/transformers/onnx/features.py

@@ -250,7 +250,7 @@ class FeaturesManager:
             "token-classification",
             onnx_config_cls=GPT2OnnxConfig,
         ),
-        "gpt-j": supported_features_mapping(
+        "gptj": supported_features_mapping(
             "default",
             "default-with-past",
             "causal-lm",