dwl-cli 1.0.0

Command-line interface for fast model training

DWL CLI Tool

Command-line interface for training DWL (Deep Weight Learning) models.

Installation

pip install dwl-cli

Usage

# Basic DWL training
dwl-train --model bert-base-uncased --dataset yahoo_answers_topics

# Traditional training with custom parameters
dwl-train --model roberta-base --dataset ag_news --method traditional --epochs 50 --lr 0.0001

# DWL training with custom components
dwl-train --model distilbert-base-uncased --dataset emotion --components 100 --epochs 30

# Non-verbose mode
dwl-train --model bert-base-uncased --dataset imdb --quiet

# Use custom backend URL
dwl-train --model bert-base-uncased --dataset ag_news

# List available options
dwl-train --list-models
dwl-train --list-datasets

# Get help
dwl-train --help

Features

• Easy to use: Simple command-line interface
• Flexible: All training parameters configurable
• Real-time streaming: See training progress as it happens
• Multiple models: Support for BERT, RoBERTa, DistilBERT, and more
• Multiple datasets: 12+ text classification datasets
• Remote support: Can connect to any backend URL

Available Models

• bert-base-uncased
• bert-large-uncased
• roberta-base
• roberta-large
• distilbert-base-uncased
• albert-base-v2
• xlnet-base-cased

Available Datasets

• ag_news
• dbpedia_14
• yahoo_answers_topics
• yelp_review_full
• yelp_polarity
• amazon_polarity
• trec
• emotion
• go_emotions
• imdb
• banking77
• rotten_tomatoes

License

MIT License