update readme

README.md

@@ -1,3 +1,56 @@
 ---
 license: mit
+language:
+- ja
+pipeline_tag: text-generation
 ---
+# Model Card for Tanrei/GPTSAN-japanese
+
+General-purpose Swich transformer based Japanese language model 
+
+## Text Generation
+
+```python
+>>> from transformers import AutoModel, AutoTokenizer
+>>> model = AutoModel.from_pretrained("Tanrei/GPTSAN-japanese")
+>>> tokenizer = AutoTokenizer.from_pretrained("Tanrei/GPTSAN-japanese")
+>>> x_tok = tokenizer.encode("武田信玄は、")
+>>> model = model.cuda()
+>>> res = model.generator.generate_lm(x_tok, tokenizer, connected_inputs=0)
+>>> res[0]
+'勝頼の父であり、天正四年(1576)に死去するまで甲府14万石の大名として甲府を治めた戦国大名ですが...'
+```
+
+## Masked Language Model
+
+```python
+>>> from transformers import AutoModel, AutoTokenizer
+>>> model = AutoModel.from_pretrained("Tanrei/GPTSAN-japanese")
+>>> tokenizer = AutoTokenizer.from_pretrained("Tanrei/GPTSAN-japanese")
+>>> x_tok = tokenizer.encode("武田信玄は、<|inputmask|>時代ファンならぜひ押さえ<|inputmask|>きたい名将の一人。")
+>>> model = model.cuda()
+>>> res = model.generator.predict_mlm(x_tok, tokenizer)
+>>> res[0]
+'武田信玄は、戦国時代ファンならぜひ押さえておきたい名将の一人。'
+```
+
+
+# Model Details
+
+## Model Description
+
+Japanese language model using Switch Transformer.
+It has the same structure as the model introduced as `Prefix LM` in the T5 paper, and works with both Test Generation and Masked Language Model.
+
+
+
+- **Developed by:** Toshiyuki Sakamoto (tanreinama)
+- **Model type:** Switch Transformer
+- **Language(s) (NLP):** Japanese
+- **License:** MIT License
+
+## Model Sources
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** https://github.com/tanreinama/GPTSAN