This weblog publish focuses on new options and enhancements. For a complete record, together with bug fixes, please see the launch notes.
We’re introducing pre-built, ready-to-use templates that simplifies the app creation course of for varied usecases. Every template comes with a spread of sources, similar to datasets, fashions, workflows, and modules, permitting you to rapidly get began together with your app creation course of.
If you select a template to create an app, the configurations and sources accessible within the template shall be preemptively utilized to your new software. You need to use the pre-built elements to rapidly apply AI to your particular use case.
Together with the present templates within the latest launch, added new templates similar to Sentiment Evaluation and Textual content Moderation. Let us take a look at the small print of every template:
Listed here are a number of the workflows accessible within the Textual content Moderation template for varied usecases:
1. Multilingual moderation Classifier Workflow: This workflow is wrapped across the Multilingual textual content moderation mannequin that classifies textual content as poisonous, insult, obscene, identity_hate and severe_toxic.
2. English textual content moderation Classifier Workflow: This workflow is wrapped across the English textual content moderation mannequin that classifies english textual content as poisonous, insult, obscene, identity_hate and severe_toxic.
3. Textual content-moderation-mistral-7b: Workflow makes use of Mistral-7b Mannequin with specified immediate template for textual content moderation that establish and filter out any hate speech, violent language and specific content material and Reply with ‘Inappropriate’ if such content material is current and ‘Acceptable’ in any other case.
4. Textual content-moderation-misinformation-dbrx: Workflow makes use of DBRX Mannequin with specified immediate template for Misinformation moderation that establish and filter out  misinformation or unsubstantiated claims, particularly associated to well being, science, or information occasions and reply with ‘Potential Misinformation’ if the content material appears questionable or ‘Possible Dependable’ if the data seems to be credible.
5. LlamaGuard-prompt-moderation: Workflow makes use of Llama Guard Mannequin with specified immediate template for security of LLMs enter prompts.Â
Listed here are a couple of workflows from the template.
1. Tweet sentiment Evaluation Workflow: This workflow makes use of twitter-roberta-base-sentiment-latest mannequin trained on ~124M tweets, and finetuned for sentiment evaluation with the TweetEval benchmark.Â
2. Multilingual Sentiment Evaluation-DistilBERT Workflow: This workflow makes use of distilbert-base-multilingual-cased-sentiments mannequin, finetuned for sentiment evaluation in a number of languages.
3. Monetary sentiments Evaluation-FinBERT Workflow: This workflow makes use of finbert, A BERT-based mannequin fine-tuned on monetary textual content for high-accuracy sentiment evaluation within the finance area.
4. Sentimental-analysis-mistral-7b Workflow: Workflow makes use of Mistral-7b Mannequin with a specified immediate template for sentimental evaluation and predicts if the sentence/ pargraph is constructive/destructive or impartial sentiment.
The template consists of Clarifai’s Chatbot Module that allows you to chat with a number of Massive Language Fashions with a single UI interface.Â
Listed here are a couple of fashions and workflows that the Picture Moderation Template consists of.
1. NSFW Recognition mannequin predicts the chance that a picture incorporates suggestive or sexually specific nudity. It is an excellent answer for anybody making an attempt to average or filter nudity from their platform routinely. It’s restricted to nudity-specific use instances.
2. NSFW-recognition workflow: This workflow is wrapped round NSFW Recognition Classifier that classifies the pictures as nudity, sexually-explicit imagery.
3. Hate-symbol-detection workflow: This workflow is wrapped round Hate symbol-detector that detects the ADL-recognized hate image.
This App Template discusses a number of content material era use instances similar to E mail writing, Weblog writing, Query Answering, Storytelling, Social media content material and comes with a number of ready-to-use workflows for content material creation.
This template discover 3 main strategies for summarization that begin with Novice and find yourself with Professional.
1. Sentence and Paragraph Summarization: This helps to summarise a couple of paragraphs and need to one-off summarize. You’ll be able to merely submit the textual content for summarization via the next workflow, and it’ll return a condensed model:
2. Web page-level Summarization: This methodology addresses the problem of summarizing texts spanning a number of pages.
3. Summarize a whole ebook: This methodology might help to summarise the whole ebook with the Greatest Illustration Vectors Methodology.
The App template consists of pre-built RAG brokers, leveraging totally different LLM fashions and optimized via varied immediate engineering strategies, to call a couple of.
1. Rag-agent-gpt4-turbo-naive: This RAG Agent makes use of GPT-4 Turbo LLM mannequin with a easy immediate for easy integration.
2. Rag-agent-claude2-1-CoT-few-shot: This RAG Agent makes use of the Claude-2.1 LLM mannequin with CoT prompting for enhanced reasoning and efficiency.