cognitivecomputations

/

Wizard-Vicuna-13B-Uncensored

like 283

Hello
When using the model and it tries to write code etc it escapes every special character..
I think its an issue in the dataset.
Here is an example generation (It happens every time):

Sure, here's an example Python code that uses the Hugging Face Transformers library to train and evaluate a Google T5 model on some text data:

import transformers as tf
from transformers import GPT2LMHead, GPT2Tokenizer, GPT2ForSequenceClassification
# Load pre-trained tokenizer and classifier
tokenizer = GPT2Tokenizer.from\_pretrained('gpt2')
classifier = GPT2ForSequenceClassification.from\_pretrained('gpt2')
model = GPT2LMHead.from\_pretrained('gpt2')
# Define input/output sequences for training and evaluation
train\_data = ['This is my first sentence', 'This is another sentence']
valid\_data = [tf.constant(["I love this language!"]), tf.constant("It works well!")]
test\_data = ["This is yet another sentence"]
# Train the model using the train data
train\_dataset = tf.data.Dataset.from\_tensor\_slices((train\_data)).repeat().batch(1)
inputs = tokenizer(\_, padding=True, truncation=True)