Automate loading, fine-tuning, inference, and evaluation of transformer models.
npx skills add https://github.com/xiechy/climate-ai --skill transformers-xiechy
Please help me install this Agent Skill. Skill: transformers Source: https://github.com/xiechy/climate-ai/tree/main/scientific-packages/transformers Command: npx skills add https://github.com/xiechy/climate-ai --skill transformers-xiechy
The Transformers library provides state-of-the-art models for NLP, computer vision, audio, and multimodal tasks. Use this skill for quick inferences with pipelines, full training with the Trainer API, and flexible text generation with various decoding strategies.
Try a simple sentiment classification using pipeline("text-classification") on sample text.
High-intent search queries and answers about installing and using this skill.
Fine-tune transformer models using the Trainer API with your custom datasets. Load a pretrained model, prepare your data with tokenization, configure training parameters including mixed precision and distributed settings, then call Trainer.train() to optimize the model for your specific task.
Yes. Transformers supports text generation across multiple tasks—summarization, translation, and custom generation—using diverse decoding strategies: greedy search, beam search, sampling, and contrastive decoding for flexible output control and quality tuning.
Use the pipeline() function for rapid inference without boilerplate code. Pipelines handle model loading, tokenization, and postprocessing automatically across tasks like text classification, NER, question-answering, and text generation.
Transformers extends beyond NLP to vision and multimodal domains. Deploy image classification, visual question-answering, and image-captioning models using the same unified API and training patterns as text tasks.
Save trained models locally or push them directly to Hugging Face Hub using built-in methods. Hub integration enables version control, sharing, and easy loading by other users or production systems via pipeline() or model.from_pretrained().
Core dependencies are transformers, datasets for data handling, torch for computation, and evaluate for benchmarking. These handle model loading, tokenization, training orchestration, and performance metrics across all supported tasks.
Rawgrowth-ConsultingTokenize text with Rust-backed tokenizers for NLP preprocessing workflows.
liumeixinManage Hugging Face Hub models, datasets, repositories, and inference endpoints via CLI.
Clay-HHKProcess large-scale ML datasets across clusters for distributed preprocessing and batch inference.
ovachieverLoad, infer, and fine-tune pre-trained transformer models across modalities.
