tags:
- salesken
license: apache-2.0
inference: true
datasets: google_wellformed_query
widget:
text: she present paper today
model_card: model_name: Salesken's Query Wellformedness Score Model description: Evaluate the well-formedness of sentences by checking grammatical correctness and completeness. Sensitive to case and penalizes sentences for incorrect grammar and case. features: - Wellformedness Score: Provides a score indicating grammatical correctness and completeness. - Case Sensitivity: Recognizes and penalizes incorrect casing in sentences. - Broad Applicability: Can be used on a wide range of sentences. evaluation_results: input: - She is presenting a paper tomorrow - she is presenting a paper tomorrow - She present paper today scores: - 0.8917 - 0.4270 - 0.0134
- Dogs are mammals.
- she loves to read books on history.
- When the rain in Spain.
- Eating apples are healthy for you.
- The Eiffel Tower is in Paris.
Among these sentences:
Sentences 1 and 5 are well-formed and have correct grammar and case. Sentence 2 starts with a lowercase letter. Sentence 3 is a fragment and is not well-formed. Sentence 4 has a subject-verb agreement error.
example_usage: library: HuggingFace transformers
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("salesken/query_wellformedness_score")
model = AutoModelForSequenceClassification.from_pretrained("salesken/query_wellformedness_score")
sentences = [
'what was the reason for everyone to leave the company',
'What was the reason behind everyone leaving the company',
# ... other sentences
]
features = tokenizer(sentences, padding=True, truncation=True, return_tensors="pt")
model.eval()
with torch.no_grad():
scores = model(**features).logits
print(scores)
intended_use_cases: - Content Creation: Validate the well-formedness of written content. - Educational Platforms: Students check the grammaticality of sentences. - Chatbots & Virtual Assistants: Validate user queries or generate well-formed responses. limitations: - Model might have occasional inaccuracies or biases. - Long sentences might be truncated, affecting score. contact: contact@salesken.ai