Convai Innovations presents Laya, a multilingual, non-autoregressive system 1 decision model designed to provide typed answers with mathematically calibrated probabilities in a single forward pass. Unlike generative models, it does not generate text, thereby eliminating hallucinations and the need for parsing. The framework includes an automated Router that detects language and script to dispatch tasks to the most efficient checkpoint (English or Multilingual) within approximately 35ms on GPU.
- It is trained using Reinforcement Learning with Calibrated Decisions (RLCD) to ensure honest probability reporting.
- Laya can support context lengths of up to 8,192 tokens in its multilingual version.
- The model family includes specialized checkpoints like `laya-typed-decisions` which achieves significantly higher accuracy through fine-tuning on specific workflows.
- Performance benchmarks show it is roughly 6–8× faster than TypeSafe Jev for single question latency on a T4 GPU.
- It is trained using Reinforcement Learning with Calibrated Decisions (RLCD) to ensure honest probability reporting.
- Laya can support context lengths of up to 8,192 tokens in its multilingual version.
- The model family includes specialized checkpoints like `laya-typed-decisions` which achieves significantly higher accuracy through fine-tuning on specific workflows.
- Performance benchmarks show it is roughly 6–8× faster than TypeSafe Jev for single question latency on a T4 GPU.
A practical pipeline for classifying messy free-text data into meaningful categories using a locally hosted LLM, no labeled training data required.
Python tutorial for reproducible labeling of cutting-edge topic models with GPT4-o-mini. The article details training a FASTopic model and labeling its results using GPT-4.0 mini, emphasizing reproducibility and control over the labeling process.
Multi-class zero-shot embedding classification and error checking. This project improves zero-shot image/text classification using a novel dimensionality reduction technique and pairwise comparison, resulting in increased agreement between text and image classifications.
This article provides a comprehensive guide on the basics of BERT (Bidirectional Encoder Representations from Transformers) models. It covers the architecture, use cases, and practical implementations, helping readers understand how to leverage BERT for natural language processing tasks.
A detailed guide on creating a text classification model with Hugging Face's transformer models, including setup, training, and evaluation steps.
BEAL is a deep active learning method that uses Bayesian deep learning with dropout to infer the model’s posterior predictive distribution and introduces an expected confidence-based acquisition function to select uncertain samples. Experiments show that BEAL outperforms other active learning methods, requiring fewer labeled samples for efficient training.
A Github Gist containing a Python script for text classification using the TxTail API
Zero-Shot Classification
To perform zero shot classification, we want to predict labels for our samples without any training. To do this, we can simply embed short descriptions of each label, such as positive and negative, and then compare the cosine distance between embeddings of samples and label descriptions.
As shown above, zero-shot classification with embeddings can lead to great results, especially when the labels are more descriptive than just simple words.
The highest similarity label to the sample input is the predicted label. We can also define a prediction score to be the difference between the cosine distance to the positive and to the negative label. This score can be used for plotting a precision-recall curve, which can be used to select a different tradeoff between precision and recall, by selecting a different threshold.
To perform zero shot classification, we want to predict labels for our samples without any training. To do this, we can simply embed short descriptions of each label, such as positive and negative, and then compare the cosine distance between embeddings of samples and label descriptions.
As shown above, zero-shot classification with embeddings can lead to great results, especially when the labels are more descriptive than just simple words.
The highest similarity label to the sample input is the predicted label. We can also define a prediction score to be the difference between the cosine distance to the positive and to the negative label. This score can be used for plotting a precision-recall curve, which can be used to select a different tradeoff between precision and recall, by selecting a different threshold.
