Update README.md

Files changed (1) hide show

README.md CHANGED Viewed

@@ -12,5 +12,3 @@ The model is the "small" version of GPT-2 (12-layer, 768-hidden, 12-heads) with
 ## Training details:
 It is trained from scratch a generative Transformer model as GPT-2 on a large corpus of Greek text so that the model can generate long stretches of contiguous coherent text. Attention dropouts with a rate of 0.1 are used for regularization on all layers and L2 weight decay of 0,01. In addition, a batch size of 4 and accumulated gradients over 8 iterations are used, resulting in an effective batch size of 32. The model uses the Adam optimization scheme with a learning rate of 1e-4 and is trained for 20 epochs. The learning rate increases linearly from zero over the first 9000 updates and decreases linearly by using a linear schedule. The implementation is based on the open-source PyTorch-transformer library (HuggingFace 2019).
-## Fine-tuned model using the pre-trained "gpt2-greek":
-https://huggingface.co/nikokons/conversational-agent-el


12	## Training details:
13	It is trained from scratch a generative Transformer model as GPT-2 on a large corpus of Greek text so that the model can generate long stretches of contiguous coherent text. Attention dropouts with a rate of 0.1 are used for regularization on all layers and L2 weight decay of 0,01. In addition, a batch size of 4 and accumulated gradients over 8 iterations are used, resulting in an effective batch size of 32. The model uses the Adam optimization scheme with a learning rate of 1e-4 and is trained for 20 epochs. The learning rate increases linearly from zero over the first 9000 updates and decreases linearly by using a linear schedule. The implementation is based on the open-source PyTorch-transformer library (HuggingFace 2019).
14