What models does this skill support?

This skill is designed to work with the HuggingFace Transformers ecosystem, supporting popular architectures like BERT, RoBERTa, DistilBERT, T5, BART, and various GPT-style models.

What training tasks are covered?

It provides comprehensive coverage for text classification (sentiment, intent, NER), text generation (summarization, Q&A), and general domain adaptation through fine-tuning.

How does it handle dataset preparation?

The skill includes templates for loading and preprocessing data from common formats like CSV and JSON using the HuggingFace Datasets library, ensuring data is correctly tokenized and formatted for training.

Can I use this for large models on a limited budget?

Yes, the skill includes specialized patterns for PEFT (Parameter-Efficient Fine-Tuning) and LoRA, allowing you to fine-tune large models (7B+) on consumer GPUs with limited VRAM.

ML Training Patterns

Name: ML Training Patterns
Author: vanman2024

byvanman2024

データサイエンスとML

Streamlines machine learning model development with production-ready templates for classification, text generation, and parameter-efficient fine-tuning.

概要

ML Training Patterns provides a comprehensive suite of automation scripts and configuration templates designed to accelerate the development of machine learning pipelines. It offers standardized patterns for common NLP tasks such as sentiment analysis, question answering, and text summarization using the HuggingFace Transformers ecosystem. Whether you are performing full fine-tuning on enterprise hardware or leveraging memory-efficient techniques like LoRA/PEFT on consumer GPUs, this skill equips Claude with the tools to generate robust training code, prepare datasets, and optimize hyperparameters for deployment-ready models.

主な機能

Optimized dataset preparation and preprocessing scripts
Automated setup scripts for classification and generation tasks
Production-ready YAML and JSON configuration templates
Pre-configured HuggingFace Transformers training pipelines
0 GitHub stars
Support for Parameter-Efficient Fine-Tuning (PEFT) and LoRA

ユースケース

Creating automated training pipelines for text generation and summarization
Fine-tuning large language models on consumer-grade GPUs using LoRA
Building custom text classifiers for sentiment or intent detection

概要

主な機能

Optimized dataset preparation and preprocessing scripts
Automated setup scripts for classification and generation tasks
Production-ready YAML and JSON configuration templates
Pre-configured HuggingFace Transformers training pipelines
0 GitHub stars
Support for Parameter-Efficient Fine-Tuning (PEFT) and LoRA

ユースケース

Creating automated training pipelines for text generation and summarization
Fine-tuning large language models on consumer-grade GPUs using LoRA
Building custom text classifiers for sentiment or intent detection