MND_TweetEvalBert_model

This model is a fine-tuned version of bert-base-uncased on the tweet_eval dataset. It achieves the following results on the evaluation set:

• Loss: 0.7241

Model description

This is how to use the model with the transformer library to do a text classification task. This model was trained and built for sentiment analysis with a text classification model architecture.

from transformers import AutoTokenizer, AutoModelForSequenceClassification
from transformers import pipeline

tokenizer = AutoTokenizer.from_pretrained("barbieheimer/MND_TweetEvalBert_model")
model = AutoModelForSequenceClassification.from_pretrained("barbieheimer/MND_TweetEvalBert_model")

# We can now use the model in the pipeline.
classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)

# Get some text to fool around with for a basic test.
text = "I loved Oppenheimer and Barbie "

classifier(text) # Let's see if the model works on our example text.

[{'label': 'JOY', 'score': 0.9845513701438904}]

Training Evalutation Results

{'eval_loss': 0.7240552306175232,
 'eval_runtime': 3.7803,
 'eval_samples_per_second': 375.896,
 'eval_steps_per_second': 23.543,
 'epoch': 5.0}

Overall Model Evaluation Results

{'accuracy': {'confidence_interval': (0.783, 0.832),
  'standard_error': 0.01241992329458207,
  'score': 0.808},
 'total_time_in_seconds': 150.93268656500004,
 'samples_per_second': 6.625470087086432,
 'latency_in_seconds': 0.15093268656500003}

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

{'training_loss'=0.3821827131159165}

{'train_runtime': 174.1546, 'train_samples_per_second': 93.509,
'train_steps_per_second': 5.857, 'total_flos': 351397804992312.0,
'train_loss': 0.3821827131159165, 'epoch': 5.0}

Step: 500
{training loss: 0.607100}

Step: 1000
{training loss: 0.169000}

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3