Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# LLMPromptKit: LLM Prompt Management System
|
| 2 |
|
| 3 |
LLMPromptKit is a comprehensive library for managing, versioning, testing, and evaluating prompts for Large Language Models (LLMs). It provides a structured framework to help data scientists and developers create, optimize, and maintain high-quality prompts.
|
|
@@ -10,6 +32,34 @@ LLMPromptKit is a comprehensive library for managing, versioning, testing, and e
|
|
| 10 |
- **Evaluation Framework**: Measure prompt quality with customizable metrics
|
| 11 |
- **Advanced Templating**: Create dynamic prompts with variables, conditionals, and loops
|
| 12 |
- **Command-line Interface**: Easily integrate into your workflow
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
## Documentation
|
| 15 |
|
|
@@ -20,6 +70,7 @@ For detailed documentation, see the [docs](./docs) directory:
|
|
| 20 |
- [CLI Usage](./docs/cli_usage.md)
|
| 21 |
- [Advanced Features](./docs/advanced_features.md)
|
| 22 |
- [Integration Examples](./docs/integration_examples.md)
|
|
|
|
| 23 |
|
| 24 |
## Installation
|
| 25 |
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: llmpromptkit
|
| 3 |
+
title: LLMPromptKit
|
| 4 |
+
emoji: 🚀
|
| 5 |
+
tags:
|
| 6 |
+
- prompt-engineering
|
| 7 |
+
- llm
|
| 8 |
+
- nlp
|
| 9 |
+
- prompt-management
|
| 10 |
+
- huggingface
|
| 11 |
+
- version-control
|
| 12 |
+
- ab-testing
|
| 13 |
+
- evaluation
|
| 14 |
+
languages:
|
| 15 |
+
- python
|
| 16 |
+
license: mit
|
| 17 |
+
pipeline_tag: text-generation
|
| 18 |
+
datasets:
|
| 19 |
+
- none
|
| 20 |
+
|
| 21 |
+
---
|
| 22 |
+
|
| 23 |
# LLMPromptKit: LLM Prompt Management System
|
| 24 |
|
| 25 |
LLMPromptKit is a comprehensive library for managing, versioning, testing, and evaluating prompts for Large Language Models (LLMs). It provides a structured framework to help data scientists and developers create, optimize, and maintain high-quality prompts.
|
|
|
|
| 32 |
- **Evaluation Framework**: Measure prompt quality with customizable metrics
|
| 33 |
- **Advanced Templating**: Create dynamic prompts with variables, conditionals, and loops
|
| 34 |
- **Command-line Interface**: Easily integrate into your workflow
|
| 35 |
+
- **Hugging Face Integration**: Seamlessly test prompts with thousands of open-source models
|
| 36 |
+
|
| 37 |
+
## Hugging Face Integration
|
| 38 |
+
|
| 39 |
+
LLMPromptKit includes a powerful integration with Hugging Face models, allowing you to:
|
| 40 |
+
|
| 41 |
+
- Test prompts with thousands of open-source models
|
| 42 |
+
- Run evaluations with models like FLAN-T5, GPT-2, and others
|
| 43 |
+
- Compare prompt performance across different model architectures
|
| 44 |
+
- Access specialized models for tasks like translation, summarization, and question answering
|
| 45 |
+
|
| 46 |
+
```python
|
| 47 |
+
from llmpromptkit import PromptManager, PromptTesting
|
| 48 |
+
from llmpromptkit.integrations.huggingface import get_huggingface_callback
|
| 49 |
+
|
| 50 |
+
# Initialize components
|
| 51 |
+
prompt_manager = PromptManager()
|
| 52 |
+
testing = PromptTesting(prompt_manager)
|
| 53 |
+
|
| 54 |
+
# Get a HuggingFace callback
|
| 55 |
+
hf_callback = get_huggingface_callback(
|
| 56 |
+
model_name="google/flan-t5-base",
|
| 57 |
+
task="text2text-generation"
|
| 58 |
+
)
|
| 59 |
+
|
| 60 |
+
# Run tests with the model
|
| 61 |
+
test_results = await testing.run_test_cases(prompt_id="your_prompt_id", llm_callback=hf_callback)
|
| 62 |
+
```
|
| 63 |
|
| 64 |
## Documentation
|
| 65 |
|
|
|
|
| 70 |
- [CLI Usage](./docs/cli_usage.md)
|
| 71 |
- [Advanced Features](./docs/advanced_features.md)
|
| 72 |
- [Integration Examples](./docs/integration_examples.md)
|
| 73 |
+
- [Integration Examples](./docs/integration_examples.md)
|
| 74 |
|
| 75 |
## Installation
|
| 76 |
|