Update README.md

README.md

@@ -40,8 +40,8 @@ VBART is the first sequence-to-sequence LLM pre-trained on Turkish corpora from
 The model is capable of conditional text generation tasks such as text summarization, paraphrasing, and title generation when fine-tuned.
 It outperforms its multilingual counterparts, albeit being much smaller than other implementations.
 
-VBART-XLarge is an experimental version of original VBART-Large. VBart-XLarge created using copying original weights of VBart-Large and adding extra layers between original weights. It is exposed to %88 less tokens compared to VBart-Large.
-VBART-XLarge gives improved results compared to VBART-Large albeit in small margins.
+VBART-XLarge is created by adding extra Transformer layers between the layers of VBART-Large. Hence it was able to transfer learned weights from the smaller model while doublings its number of layers.
+VBART-XLarge slightly improves the results compared to VBART-Large albeit in small margins.
 
 This repository contains fine-tuned TensorFlow and Safetensors weights of VBART for question-answering and generation tasks described in the [paper](https://doi.org/10.55730/1300-0632.3914).
 
@@ -96,7 +96,7 @@ The fine-tuning dataset is [TQuAD](https://github.com/obss/turkish-question-gene
 This model is fine-tuned for question-answering and question-generation tasks with specific prompts. It is not intended to be used in any other case and can not be fine-tuned to any other task with full performance of the base model. It is also not guaranteed that this model will work without specified prompts.
 
 ### Training Procedure  
-Pre-trained for 30 days and for a total of 708B tokens. Further pretrained enlarged model for 84B tokens. Finally, finetuned for 55 epoch.
+Pre-trained for 8 days and for a total of 84B tokens. Finally, finetuned for 55 epochs.
 #### Hardware
 - **GPUs**: 8 x Nvidia A100-80 GB
 #### Software
@@ -107,17 +107,11 @@ Pre-trained for 30 days and for a total of 708B tokens. Further pretrained enlar
 - **Training objective**: Sentence permutation and span masking (using mask lengths sampled from Poisson distribution λ=3.5, masking 30% of tokens)
 - **Optimizer** : Adam optimizer (β1 = 0.9, β2 = 0.98, Ɛ = 1e-6)
 - **Scheduler**: Custom scheduler from the original Transformers paper (20,000 warm-up steps)
-- **Dropout**: 0.1 (dropped to 0.05 and then to 0 in the last 165k and 205k steps, respectively)
+- **Weight Initialization**: Model Enlargement from VBART-Large. See the related section in the [paper](https://arxiv.org/abs/2403.01308) for the details. 
+- **Dropout**: 0.1 (dropped to 0.05 and then to 0 in the last 80K and 80k steps, respectively)
 - **Initial Learning rate**: 5e-6
-- **Training tokens**: 708B
-
-##### Experimental Model Enlargement
-Same as pretraining but;
-- **Scheduler**: with 5,000 warm-up steps
-- **Dropout**: 0.1 (dropped to 0.05 and then to 0 in the last 160k and 80k steps, respectively)
 - **Training tokens**: 84B
 
-
 ##### Fine-tuning
 - **Training regime:** fp16 mixed precision
 - **Optimizer** : Adam optimizer (β1 = 0.9, β2 = 0.98, Ɛ = 1e-6)