File size: 3,093 Bytes
4a711e2 1f09d27 4a711e2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 |
---
license: apache-2.0
datasets:
- WPAI-INC/wp-pattern-instruction-pairs-1
language:
- en
base_model:
- Qwen/Qwen2.5-32B-Instruct
pipeline_tag: text-generation
tags:
- code
---
# WPAIGPT-fse-patterns-1
WPAIGPT-fse-patterns-1 is a fine-tuned version of the Qwen 2.5 32B Instruct model, specialized for generating structured WordPress Gutenberg Patterns.
## Model Details
- **Developed by:** WPAI Inc, James LePage
- **Model type:** Language model for generating WordPress block patterns
- **Language(s):** Multilingual, with primary focus on English
- **License:** Apache 2.0
- **Finetuned from model:** Qwen 2.5 32B Instruct
## Uses
- Generating WordPress Gutenberg block patterns
- Assisting WordPress developers and content creators
- Providing inspiration for site layouts and designs
- Automating the creation of structured WordPress content
## Training Data
The model was fine-tuned on a dataset derived from the WordPress.org Pattern Repository, containing:
- System prompts for WordPress pattern generation
- User requests based on pattern descriptions
- Assistant responses with WordPress Pattern Structures
The dataset includes patterns in multiple languages and covers all available WordPress blocks as of September 2024.
## Training Procedure
Fine-tuning was performed on A100 GPUs (80GB each) using a conversational format. The process included:
- Structuring data as system prompts, user requests, and assistant responses
- Learning rate scheduling for optimal convergence
- Periodic evaluation on a validation set
## Evaluation Results
Formal evaluations have not been conducted. Performance is assessed through:
- Manual review of generated patterns
- Comparison to original patterns in the WordPress repository
- Testing in real-world WordPress environments
## Ethical Considerations
Users should be aware of potential biases in the training data that may affect pattern suggestions. Care should be taken to review and customize generated patterns for accessibility and inclusivity.
## Limitations and Biases
- Knowledge limited to patterns available in the WordPress repository as of September 2024
- Potential biases towards popular block types and layouts
*- May not always generate valid or optimal WordPress markup*
It's crucial to note that while the model generates structurally accurate patterns, full implementation within a WordPress development ecosystem requires additional tooling. Specifically, the model may reference non-existent images and URLs, necessitating a system to replace these with appropriate assets from the user's WordPress site. This functionality is being developed for integration with AgentWP (https://agentwp.com).
## Out-of-Scope Use
This model should not be used for:
- Generating non-WordPress related content
- Creating patterns with malicious intent or that violate WordPress coding standards
- Replacing human creativity and judgment in website design
For more information on using WordPress block patterns, refer to the official WordPress documentation: https://wordpress.org/documentation/article/block-pattern/ |