Oliverguhr
Models by this creator
🔄
fullstop-punctuation-multilang-large
125
The fullstop-punctuation-multilang-large model is a multilingual punctuation restoration model developed by Oliver Guhr. It can predict punctuation for English, Italian, French, and German text, making it useful for tasks like transcription of spoken language. The model was trained on the Europarl dataset provided by the SEPP-NLG Shared Task. It can restore common punctuation marks like periods, commas, question marks, hyphens, and colons. Similar models include bert-restore-punctuation and bert-base-multilingual-uncased-sentiment, which focus on punctuation restoration and multilingual sentiment analysis respectively. Model inputs and outputs Inputs Text**: The model takes in raw text that may be missing punctuation. Outputs Punctuated text**: The model outputs the input text with punctuation marks restored at the appropriate locations. Capabilities The fullstop-punctuation-multilang-large model can effectively restore common punctuation in English, Italian, French, and German text. It performs best on restoring periods and commas, with F1 scores around 0.95 for those markers. The model struggles more with restoring less common punctuation like hyphens and colons, achieving F1 scores around 0.60 for those. What can I use it for? This model could be useful for any applications that involve transcribing or processing spoken language in the supported languages, such as automated captioning, meeting transcripts, or voice assistants. By automatically adding punctuation, the model can make the text more readable and natural. The multilingual aspect also makes it applicable across a range of international use cases. Companies could leverage this model to improve the quality of their speech-to-text pipelines or offer more polished text outputs to customers. Things to try One interesting aspect of this model is its ability to handle multiple languages. Practitioners could experiment with feeding it text in different languages and compare the punctuation restoration performance. It could also be fine-tuned on domain-specific datasets beyond the political speeches in Europarl to see if the model generalizes well. Additionally, combining this punctuation model with other NLP models like sentiment analysis or named entity recognition could lead to interesting applications for processing conversational data.
Updated 5/28/2024
🤖
spelling-correction-english-base
62
The spelling-correction-english-base model is an experimental proof-of-concept spelling correction model for the English language, created by oliverguhr. It is designed to fix common typos and punctuation errors in text. This model is part of oliverguhr's research into developing models that can restore the punctuation of transcribed spoken language, as demonstrated by the fullstop-punctuation-multilang-large model. Model inputs and outputs Inputs English text with potential spelling and punctuation errors Outputs Corrected English text with improved spelling and punctuation Capabilities The spelling-correction-english-base model can detect and fix common spelling and punctuation mistakes in English text. For example, it can correct words like "comparsion" to "comparison" and add missing punctuation like periods and commas. What can I use it for? This model could be useful for various applications that require accurate spelling and punctuation, such as writing assistance tools, content editing, and language learning platforms. It could also be used as a starting point for fine-tuning on specific domains or languages. Things to try You can experiment with the spelling-correction-english-base model using the provided pipeline interface. Try running it on your own text samples to see how it performs, and consider ways you could integrate it into your projects or applications.
Updated 5/28/2024
🔎
german-sentiment-bert
50
The german-sentiment-bert model is a sentiment classification model trained on over 1.8 million German language texts to predict the sentiment of German language input as positive, negative, or neutral. It uses the BERT architecture and was developed by maintainer oliverguhr. Compared to similar sentiment models like SiEBERT - English-Language Sentiment Classification and bert-base-multilingual-uncased-sentiment, the german-sentiment-bert model is specifically tailored for German language sentiment, whereas the others focus on English and multilingual sentiment. The model achieves strong performance, reaching F1 scores over 90% on various German language sentiment benchmarks. Model inputs and outputs The german-sentiment-bert model takes in German language text as input and outputs the predicted sentiment as either positive, negative, or neutral. The model was trained on a diverse set of German texts including social media, reviews, and other sources. Inputs German language text**: The model accepts any German text as input, such as product reviews, social media posts, or other types of German language content. Outputs Sentiment label**: The model outputs a sentiment label of either positive, negative, or neutral, indicating the overall sentiment expressed in the input text. Sentiment probability**: In addition to the sentiment label, the model also outputs the probability or confidence score for each sentiment class. Capabilities The german-sentiment-bert model is highly capable at accurately detecting the sentiment of German language text. In evaluations on various German sentiment datasets, the model achieved F1 scores over 90%, demonstrating its strong performance. For example, on the holidaycheck dataset of German hotel reviews, the model achieved an F1 micro score of 0.9568. Similarly, on the scare dataset of German product reviews, the model scored 0.9418. What can I use it for? The german-sentiment-bert model is well-suited for any application that requires analyzing the sentiment of German language text, such as: Customer service**: Analyzing customer feedback, reviews, and support conversations to gauge sentiment and identify areas for improvement. Social media monitoring**: Tracking sentiment towards brands, products, or topics in German social media posts. Market research**: Gauging consumer sentiment about products, services, or trends in the German market. Content moderation**: Detecting negative or toxic sentiment in user-generated German content. oliverguhr has also provided a Python package called germansentiment that simplifies the use of the model and includes preprocessing steps, making it easy to integrate into your own applications. Things to try One interesting aspect of the german-sentiment-bert model is its strong performance across diverse German language datasets, suggesting it has learned robust and generalizable representations of German sentiment. You could try using the model to analyze sentiment in different German language domains, such as social media, product reviews, news articles, or even technical documentation, to see how it performs. Additionally, you could experiment with fine-tuning the model on your own German language dataset to further improve its performance on your specific use case. Another idea is to explore the model's capabilities in handling more nuanced or complex sentiment, such as detecting sarcasm, irony, or mixed emotions in German text. This could involve creating your own German language test sets to better understand the model's limitations and areas for improvement.
Updated 6/17/2024