Update README.md

README.md

@@ -4,10 +4,13 @@ datasets:
 - Locutusque/ColumnedChatCombined
 language:
 - en
+- chi
 metrics:
 - bleu
 - perplexity
 - loss
+- reward
+- penalty
 ---
 # Model Card
 ## Model Details
@@ -25,7 +28,7 @@ The model is trained on a large dataset of conversational data, consisting of in
 The model architecture used in this model is GPT-2, a transformer-based language model that is capable of generating high-quality text with a wide range of styles and tones. The GPT-2 architecture consists of a multi-layered transformer encoder-decoder, with self-attention mechanisms that allow the model to capture long-term dependencies and generate coherent text.
 
 ## Evaluation Metrics
-The model is evaluated based on several metrics, including loss, reward, BLEU score, and perplexity. The loss metric is calculated during training and reflects the difference between the predicted output and the actual output. The reward metric is based on the number of correct words generated by the model, while the penalty metric penalizes the model for repeating words consecutively. The BLEU score measures the similarity between the generated text and the ground truth text, while the perplexity metric measures how well the model is able to predict the next word in a sequence.
+The model is evaluated based on several metrics, including loss, reward, penalty, BLEU score, and perplexity. The loss metric is calculated during training and reflects the difference between the predicted output and the actual output. The reward metric is based on the number of correct words generated by the model, while the penalty metric penalizes the model for repeating words consecutively. The BLEU score measures the similarity between the generated text and the ground truth text, while the perplexity metric measures how well the model is able to predict the next word in a sequence.
 
 ## Limitations and Bias
-One limitation of this model is that it may generate responses that are biased or inappropriate, depending on the nature of the training data. Care should be taken to ensure that the training data is diverse and representative of the target user population, in order to minimize the risk of bias or discrimination. Additionally, the model may struggle with generating responses that are highly domain-specific or technical in nature, as these may require specialized knowledge or context that is not present in the training data.
+Because I have a rather weak computer for machine learning, I was not able to train this model for too long. The model may output irrelevant answers, or even sometimes the responses can be nonsensical.