Update README.md

README.md

@@ -14,6 +14,14 @@ With the Tokenizers library, I created a 52K byte-level BPE vocab based on the t
 
 After creating the vocab, I could train the GPT-2 for Turkish on two 2080TI over the complete training corpus (five epochs).
 
+## Model weights
+
+Both PyTorch and Tensorflow compatible weights are available.
+
+| Model                             | Downloads
+| --------------------------------- | ---------------------------------------------------------------------------------------------------------------
+| `dbmdz/bert-base-turkish-cased`   | [`config.json`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/config.json) • [`merges.txt`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/merges.txt) • [`pytorch_model.bin`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/pytorch_model.bin) • [`special_tokens_map.json`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/special_tokens_map.json) • [`tf_model.h5`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/tf_model.h5) • [`tokenizer_config.json`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/tokenizer_config.json) • [`traning_args.bin`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/training_args.bin) • [`vocab.json`](https://huggingface.co/redrussianarmy/gpt2-turkish-cased/blob/main/vocab.json)
+
 ## Using the model
 
 The model itself can be used in this way: