Generative AI Text Documentation

https://skllm.beastbyte.ai/docs
Admin
This project provides documentation and guides for various text-related tasks in Generative AI,...
Tokens:11,244
Snippets:104
Trust Score:5.3
Update:3 months ago
Show doc for...
Tokens:
Raw
### Install GGUF with CPU Support

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Install scikit-llm with GGUF support for CPU-only environments. This command ensures the necessary llama-cpp-python bindings are installed.

```bash
pip install scikit-llm[gguf] --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu --no-cache-dir

```

--------------------------------

### Train and Predict with GPTClassifier

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Example demonstrating how to initialize, train, and predict using the GPTClassifier. Ensure you have the necessary data loaded into X and y.

```python
from skllm.models.gpt.classification.tunable import GPTClassifier

X, y = get_classification_dataset()
clf = GPTClassifier(n_epochs=1)
clf.fit(X,y)
clf.predict(X)
```

--------------------------------

### Initialize and Fit DynamicFewShotGPTClassifier

Source: https://skllm.beastbyte.ai/docs/dynamic-few-shot-text-classification

Initialize the DynamicFewShotGPTClassifier with a specified number of examples per class and fit it to the training data. This prepares the classifier for making predictions using dynamic few-shot learning.

```python
from skllm.models.gpt.classification.few_shot import DynamicFewShotGPTClassifier

query = "I have fallen deeply in love with this sci-fi book; its unique blend of science and fiction has me spellbound."

clf = DynamicFewShotGPTClassifier(n_examples=1).fit(X,y)

prompt = clf._get_prompt(query)
print(prompt)

```

--------------------------------

### Sentiment Analysis Tagging Example

Source: https://skllm.beastbyte.ai/docs/tagging-overview

Demonstrates how raw text can be transformed into tagged output for sentiment analysis. This example shows the input text and its corresponding tagged output.

```text
Input:
```
I love my new phone, but I am disappointed with the battery life.

```

Output:
```
<positive>I love my new phone,</positive> <negative>but I am disappointed with the battery life.</negative>

```
```

--------------------------------

### Install GGUF with Metal GPU Support (MacOS)

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Install scikit-llm with GGUF support for GPU acceleration on MacOS using Metal. This is specifically for M-series Macs.

```bash
pip install scikit-llm[gguf] --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/metal --no-cache-dir

```

--------------------------------

### Few-Shot Text Classification Example

Source: https://skllm.beastbyte.ai/docs/few-shot-text-classification

Demonstrates single-label and multi-label few-shot text classification using GPT-4. Ensure you have the necessary datasets loaded.

```python
from skllm.models.gpt.classification.few_shot import (
FewShotGPTClassifier,
MultiLabelFewShotGPTClassifier,
)
from skllm.datasets import (
get_classification_dataset,
get_multilabel_classification_dataset,
)

# single label
X, y = get_classification_dataset()
clf = FewShotGPTClassifier(model="gpt-4o")
clf.fit(X,y)
labels = clf.predict(X)

# multi-label
X, y = get_multilabel_classification_dataset()
clf = MultiLabelFewShotGPTClassifier(max_labels=2, model="gpt-4o")
clf.fit(X,y)
labels = clf.predict(X)

```

--------------------------------

### Install GGUF with CUDA GPU Support

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Install scikit-llm with GGUF support for GPU acceleration using CUDA 12.1 or later. This command specifies the appropriate extra index URL for CUDA bindings.

```bash
pip install scikit-llm[gguf] --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu121 --no-cache-dir

```

--------------------------------

### DynamicFewShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/dynamic-few-shot-text-classification

Initializes the DynamicFewShotGPTClassifier with specified parameters. This class is designed for dynamic few-shot text classification, adapting to larger datasets by constructing dynamic subsets of examples for each sample.

```APIDOC
## DynamicFewShotGPTClassifier

Initializes the dynamic few-shot text classification model.

### Parameters

- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **n_examples** (int) - Optional - Number of closest examples per class to be retrieved, by default 3.
- **memory_index** (Optional[IndexConstructor]) - Optional - Custom memory index, for details check `skllm.memory` submodule, by default None.
- **vectorizer** (Optional[BaseVectorizer]) - Optional - Scikit-LLM vectorizer; if None, `GPTVectorizer` is used, by default None.
- **metric** (Optional[str]) - Optional - Metric used for similarity search by the memory_index, by default "euclidean"
```

--------------------------------

### Define Training Data for Text Classification

Source: https://skllm.beastbyte.ai/docs/dynamic-few-shot-text-classification

Define the input text samples (X) and their corresponding labels (y) for a text classification task. This example sets up data for classifying reviews as 'books' or 'movies'.

```python
X = [
    "I love reading science fiction novels, they transport me to other worlds.", # example 1 - book - sci-fi
    "A good mystery novel keeps me guessing until the very end.", # example 2 - book - mystery
    "Historical novels give me a sense of different times and places.", # example 3 - book - historical
    "I love watching science fiction movies, they transport me to other galaxies.", # example 4 - movie - sci-fi
    "A good mystery movie keeps me on the edge of my seat.", # example 5 - movie - mystery
    "Historical movies offer a glimpse into the past.", # example 6 - movie - historical
]

y = ["books", "books", "books", "movies", "movies", "movies"]

```

--------------------------------

### FewShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/few-shot-text-classification

Initializes the FewShotGPTClassifier for single-label text classification. This class uses a few examples to classify new text into predefined classes.

```APIDOC
## FewShotGPTClassifier

### Description
Initializes the FewShotGPTClassifier for single-label text classification. This class uses a few examples to classify new text into predefined classes.

### Parameters
#### Initialization Parameters
- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Usage Example
```python
from skllm.models.gpt.classification.few_shot import FewShotGPTClassifier

clf = FewShotGPTClassifier(model="gpt-4o")
```
```

--------------------------------

### Initialize GPTVectorizer

Source: https://skllm.beastbyte.ai/docs/text-vectorization

Import and initialize the GPTVectorizer. This snippet shows the basic import statement for the class.

```python
from skllm.models.gpt.vectorization import GPTVectorizer

```

--------------------------------

### Initialize and Use GPTExplainableNER

Source: https://skllm.beastbyte.ai/docs/ner

Demonstrates initializing the GPTExplainableNER model with custom entities and transforming sample data to tag named entities. The `display_predictions` parameter is set to True for human-readable output.

```python
from skllm.models.gpt.tagging.ner import GPTExplainableNER as NER

entities = {
  "PERSON": "A name of an individual.",
  "ORGANIZATION": "A name of a company.",
  "DATE": "A specific time reference."
}

data = [
  "Tim Cook announced new Apple products in San Francisco on June 3, 2022.",
  "Elon Musk visited the Tesla factory in Austin on January 10, 2021.",
  "Mark Zuckerberg introduced Facebook Metaverse in Silicon Valley on May 5, 2023."
]

ner = NER(entities=entities, display_predictions=True)
tagged = ner.fit_transform(data)

```

--------------------------------

### TunableGPTText2Text Initialization

Source: https://skllm.beastbyte.ai/docs/tunable-text-to-text

Initializes a tunable GPT text-to-text estimator. This model can be tuned for various tasks using provided data without additional preprocessing. The `fit` method trains the model, and `transform` generates output.

```APIDOC
## TunableGPTText2Text

### Description
Initializes a tunable GPT text-to-text estimator. This model can be tuned for various tasks using provided data without additional preprocessing. The `fit` method trains the model, and `transform` generates output.

### Parameters
#### Initialization Parameters
- **base_model** (str) - Model to use, by default "gpt-3.5-turbo-0613".
- **key** (Optional[str]) - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **n_epochs** (Optional[int]) - Number of epochs; if None, determined automatically; by default None.
- **custom_suffix** (Optional[str]) - Custom suffix of the tuned model, used for naming purposes only, by default "skllm".

### Example Usage
```python
from skllm.models.gpt.text2text.tunable import TunableGPTText2Text

model = TunableGPTText2Text(
        base_model = "gpt-3.5-turbo-0613",
        n_epochs = None, 
        system_msg = "You are a text processing model."
)

model.fit(X_train, y_train) # y_train is any desired output text
out = model.transform(X_test)
```
```

--------------------------------

### Initialize and Train TunableGPTText2Text Model

Source: https://skllm.beastbyte.ai/docs/tunable-text-to-text

Demonstrates how to initialize a TunableGPTText2Text model with a specified base model and system message, then train it using provided data. The y_train parameter should contain the desired output text.

```python
from skllm.models.gpt.text2text.tunable import TunableGPTText2Text

model = TunableGPTText2Text(
        base_model = "gpt-3.5-turbo-0613",
        n_epochs = None, # int or None. When None, will be determined automatically by OpenAI
        system_msg = "You are a text processing model."
)

model.fit(X_train, y_train) # y_train is any desired output text
out = model.transform(X_test)
```

--------------------------------

### TunableVertexText2Text Initialization

Source: https://skllm.beastbyte.ai/docs/tunable-text-to-text

Initializes a tunable Vertex AI text-to-text estimator. This model can be tuned for various tasks using provided data without additional preprocessing. The `fit` method trains the model, and `transform` generates output.

```APIDOC
## TunableVertexText2Text

### Description
Initializes a tunable Vertex AI text-to-text estimator. This model can be tuned for various tasks using provided data without additional preprocessing. The `fit` method trains the model, and `transform` generates output.

### Parameters
#### Initialization Parameters
- **base_model** (str) - Model to use, by default "text-bison@002".
- **n_update_steps** (int) - Number of epochs, by default 1.

### Example Usage
```python
from skllm.models.vertex.text2text.tunable import TunableVertexText2Text

# Example initialization (actual usage may vary based on skllm's API)
# model = TunableVertexText2Text(base_model="text-bison@002", n_update_steps=5)
```
```

--------------------------------

### Initialize MultiLabelZeroShotVertexClassifier

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Import and initialize the MultiLabelZeroShotVertexClassifier. This model supports multi-label zero-shot classification using Vertex AI.

```python
from skllm.models.vertex.classification.zero_shot import MultiLabelZeroShotVertexClassifier
```

--------------------------------

### Initialize ZeroShotVertexClassifier

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Import and initialize the ZeroShotVertexClassifier. This model leverages Vertex AI for zero-shot classification.

```python
from skllm.models.vertex.classification.zero_shot import ZeroShotVertexClassifier
```

--------------------------------

### GPTExplainableNER Initialization

Source: https://skllm.beastbyte.ai/docs/ner

This snippet shows how to initialize the GPTExplainableNER estimator with various parameters. The estimator can recognize specified entities, control prediction display, choose between sparse or dense output, select a GPT model, and configure API keys and worker threads.

```APIDOC
## GPTExplainableNER

### Description
Initializes the GPTExplainableNER estimator for Named Entity Recognition.

### Parameters
#### Initialization Parameters
- **entities** (dict) - Required - A dictionary of entities to recognize, with keys as **uppercase** entity names and values as descriptions.
- **display_predictions** (Optional[bool]) - Optional - Determines whether to display predictions, by default False.
- **sparse_output** (Optional[bool]) - Optional - Determines whether to generate a sparse representation of the predictions, by default True.
- **model** (Optional[str]) - Optional - A model to use, by default "gpt-4o".
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **num_workers** (Optional[int]) - Optional - Number of workers (threads) to use, by default 1.
```

--------------------------------

### DynamicFewShotGPTClassifier Initialization Parameters

Source: https://skllm.beastbyte.ai/docs/dynamic-few-shot-text-classification

Illustrates the import statement for the DynamicFewShotGPTClassifier. This is typically used to initialize the classifier with various parameters for customization.

```python
from skllm.models.gpt.classification.few_shot import DynamicFewShotGPTClassifier

```

--------------------------------

### Import TunableGPTText2Text

Source: https://skllm.beastbyte.ai/docs/tunable-text-to-text

Shows the import statement required to use the TunableGPTText2Text estimator from the skllm library.

```python
from skllm.models.gpt.text2text.tunable import TunableGPTText2Text
```

--------------------------------

### Initialize MultiLabelZeroShotGPTClassifier

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Import and initialize the MultiLabelZeroShotGPTClassifier. This model supports multi-label zero-shot classification using GPT.

```python
from skllm.models.gpt.classification.zero_shot import MultiLabelZeroShotGPTClassifier
```

--------------------------------

### GPTVectorizer Initialization

Source: https://skllm.beastbyte.ai/docs/text-vectorization

Initializes the GPTVectorizer for text embedding. It allows customization of the model, batch size, and API keys.

```APIDOC
## GPTVectorizer

### Description
Initializes the GPTVectorizer for text embedding. It allows customization of the model, batch size, and API keys.

### Parameters
#### Initialization Parameters
- **model** (str) - Optional - Model to use, by default "text-embedding-3-small".
- **batch_size** (int) - Optional - Number of samples per request, by default 1.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Usage Example
```python
from skllm.models.gpt.vectorization import GPTVectorizer

vectorizer = GPTVectorizer(batch_size=2, model="text-embedding-ada-002")
X = vectorizer.fit_transform(["This is a text", "This is another text"])
```
```

--------------------------------

### Initialize ZeroShotGPTClassifier

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Import and initialize the ZeroShotGPTClassifier. This model uses GPT for zero-shot classification.

```python
from skllm.models.gpt.classification.zero_shot import ZeroShotGPTClassifier
```

--------------------------------

### Import TunableVertexText2Text

Source: https://skllm.beastbyte.ai/docs/tunable-text-to-text

Shows the import statement required to use the TunableVertexText2Text estimator from the skllm library.

```python
from skllm.models.vertex.text2text.tunable import TunableVertexText2Text
```

--------------------------------

### GPTSummarizer Initialization

Source: https://skllm.beastbyte.ai/docs/text-summarization

Initializes the GPTSummarizer with specified parameters for text summarization. The remaining methods follow the syntax of a scikit-learn transformer.

```APIDOC
## GPTSummarizer

### Description
Initializes the GPTSummarizer for text summarization tasks.

### Parameters
#### Initialization Parameters
- **model** (str) - Model to use, by default "gpt-3.5-turbo".
- **key** (Optional[str]) - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **max_words** (int) - Soft limit of the summary length, by default 15.
- **focus** (Optional[str]) - Concept in the text to focus on, by default None.
```

--------------------------------

### Initialize and Use CoTGPTClassifier for Sentiment Analysis

Source: https://skllm.beastbyte.ai/docs/chain-of-thought-text-classification

Demonstrates initializing CoTGPTClassifier with GPT-4o, fitting it to a dataset, and making predictions that include both labels and reasoning.

```python
from skllm.models.gpt.classification.zero_shot import CoTGPTClassifier
from skllm.datasets import get_classification_dataset

# demo sentiment analysis dataset
# labels: positive, negative, neutral
X, y = get_classification_dataset()

clf = CoTGPTClassifier(model="gpt-4o")
clf.fit(X,y)
predictions = clf.predict(X)
labels, reasoning = predictions[:, 0], predictions[:, 1]
```

--------------------------------

### Import GPTExplainableNER

Source: https://skllm.beastbyte.ai/docs/ner

Shows the import statement for the GPTExplainableNER class from the Scikit-LLM library.

```python
from skllm.models.gpt.tagging.ner import GPTExplainableNER

```

--------------------------------

### Embed Text using GPTVectorizer

Source: https://skllm.beastbyte.ai/docs/text-vectorization

Use GPTVectorizer to convert text into fixed-dimensional vectors. Initialize with a batch size and then use fit_transform.

```python
from skllm.models.gpt.vectorization import GPTVectorizer

vectorizer = GPTVectorizer(batch_size=2)
X = vectorizer.fit_transform(["This is a text", "This is another text"])

```

--------------------------------

### MultiLabelFewShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/few-shot-text-classification

Import and initialize the `MultiLabelFewShotGPTClassifier`. This classifier is designed for multi-label text classification tasks.

```python
from skllm.models.gpt.classification.few_shot import MultiLabelFewShotGPTClassifier

```

--------------------------------

### Configure Custom URL Backend

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Set a global custom URL for using any OpenAI-compatible provider with Scikit-LLM. This is useful for local or cloud-hosted models.

```python
from skllm.config import SKLLMConfig

SKLLMConfig.set_gpt_url("http://localhost:8000/")

clf = ZeroShotGPTClassifier(model="custom_url::<custom_model_name>")
```

--------------------------------

### CoTGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/chain-of-thought-text-classification

This section details the parameters required for initializing the CoTGPTClassifier. The class follows the standard scikit-learn classifier syntax for other methods.

```APIDOC
## CoTGPTClassifier

### Description
Initializes the Chain-of-Thought GPT Classifier.

### Parameters
#### Initialization Parameters
- **model** (str) - Required - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Usage
```python
from skllm.models.gpt.classification.zero_shot import CoTGPTClassifier

clf = CoTGPTClassifier(model="gpt-4o", default_label="positive")
```
```

--------------------------------

### FewShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/few-shot-text-classification

Import and initialize the `FewShotGPTClassifier`. This classifier is suitable for single-label text classification tasks.

```python
from skllm.models.gpt.classification.few_shot import FewShotGPTClassifier

```

--------------------------------

### Import CoTGPTClassifier

Source: https://skllm.beastbyte.ai/docs/chain-of-thought-text-classification

Shows the import statement for the CoTGPTClassifier class from the SKLLM library.

```python
from skllm.models.gpt.classification.zero_shot import CoTGPTClassifier
```

--------------------------------

### Basic Text Summarization

Source: https://skllm.beastbyte.ai/docs/text-summarization

Use this snippet to perform basic text summarization on a dataset. It initializes the GPTSummarizer and transforms the input data.

```python
from skllm.models.gpt.text2text.summarization import GPTSummarizer
from skllm.datasets import get_summarization_dataset

X = get_summarization_dataset()
summarizer = GPTSummarizer(model="gpt-3.5-turbo", max_words=15)
X_summarized = summarizer.fit_transform(X)
```

--------------------------------

### MultiLabelZeroShotVertexClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Initializes the MultiLabelZeroShotVertexClassifier with specified parameters. This model supports multi-label zero-shot text classification using Vertex AI.

```APIDOC
## MultiLabelZeroShotVertexClassifier

### Description
Initializes the MultiLabelZeroShotVertexClassifier with specified parameters. This model supports multi-label zero-shot text classification using Vertex AI.

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **model** (str) - Optional - Model to use, by default "text-bison@002".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **max_labels** (Optional[int]) - Optional - Maximum labels per sample, by default 5.
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.

### Request Example
```json
{
  "model": "text-bison@002",
  "default_label": "Random",
  "max_labels": 5,
  "prompt_template": null
}
```

### Response
#### Success Response (200)
Initializes the classifier object.

#### Response Example
```json
{
  "message": "Classifier initialized successfully"
}
```
```

--------------------------------

### ZeroShotVertexClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Initializes the ZeroShotVertexClassifier with specified parameters. This model is designed for zero-shot text classification using Vertex AI.

```APIDOC
## ZeroShotVertexClassifier

### Description
Initializes the ZeroShotVertexClassifier with specified parameters. This model is designed for zero-shot text classification using Vertex AI.

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **model** (str) - Optional - Model to use, by default "text-bison@002".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.

### Request Example
```json
{
  "model": "text-bison@002",
  "default_label": "Random",
  "prompt_template": null
}
```

### Response
#### Success Response (200)
Initializes the classifier object.

#### Response Example
```json
{
  "message": "Classifier initialized successfully"
}
```
```

--------------------------------

### GPTTranslator Initialization

Source: https://skllm.beastbyte.ai/docs/text-translation

This section details the parameters required for initializing the GPTTranslator estimator. The remaining methods follow the standard scikit-learn transformer syntax.

```APIDOC
## GPTTranslator

### Description
Initializes the GPTTranslator for text translation tasks.

### Parameters
#### Initialization Parameters
- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **output_language** (str) - Required - Target language, by default "English".
```

--------------------------------

### Authenticate Google Cloud for Vertex AI

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Log in to Google Cloud to set application default credentials for using the Vertex AI backend. This command is run via the Google Cloud CLI.

```bash
gcloud auth application-default login
```

--------------------------------

### MultiLabelZeroShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Initializes the MultiLabelZeroShotGPTClassifier with specified parameters. This model supports multi-label zero-shot text classification using GPT.

```APIDOC
## MultiLabelZeroShotGPTClassifier

### Description
Initializes the MultiLabelZeroShotGPTClassifier with specified parameters. This model supports multi-label zero-shot text classification using GPT.

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **max_labels** (Optional[int]) - Optional - Maximum labels per sample, by default 5.
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Request Example
```json
{
  "model": "gpt-3.5-turbo",
  "default_label": "Random",
  "max_labels": 5,
  "prompt_template": null,
  "key": null,
  "org": null
}
```

### Response
#### Success Response (200)
Initializes the classifier object.

#### Response Example
```json
{
  "message": "Classifier initialized successfully"
}
```
```

--------------------------------

### ZeroShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Initializes the ZeroShotGPTClassifier with specified parameters. This model is designed for zero-shot text classification using GPT.

```APIDOC
## ZeroShotGPTClassifier

### Description
Initializes the ZeroShotGPTClassifier with specified parameters. This model is designed for zero-shot text classification using GPT.

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Request Example
```json
{
  "model": "gpt-3.5-turbo",
  "default_label": "Random",
  "prompt_template": null,
  "key": null,
  "org": null
}
```

### Response
#### Success Response (200)
Initializes the classifier object.

#### Response Example
```json
{
  "message": "Classifier initialized successfully"
}
```
```

--------------------------------

### MultiLabelGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Initializes a MultiLabelGPTClassifier for tunable multi-label text classification. This class allows fine-tuning of GPT models for multi-label classification tasks. The remaining methods follow scikit-learn classifier syntax.

```APIDOC
## MultiLabelGPTClassifier

### Description
Initializes a MultiLabelGPTClassifier for tunable multi-label text classification. This class allows fine-tuning of GPT models for multi-label classification tasks. The remaining methods follow scikit-learn classifier syntax.

### Parameters
#### Initialization Parameters
- **base_model** (str) - Optional - Base model to use, by default "gpt-3.5-turbo-0613".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **n_epochs** (Optional[int]) - Optional - Number of epochs; if None, determined automatically; by default None.
- **custom_suffix** (Optional[str]) - Optional - Custom suffix of the tuned model, used for naming purposes only, by default "skllm".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **max_labels** (Optional[int]) - Optional - Maximum labels per sample, by default 5.
```

--------------------------------

### VertexClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Initializes a VertexClassifier for tunable text classification using Vertex AI models. This class allows fine-tuning of Vertex AI models for classification tasks. The remaining methods follow scikit-learn classifier syntax.

```APIDOC
## VertexClassifier

### Description
Initializes a VertexClassifier for tunable text classification using Vertex AI models. This class allows fine-tuning of Vertex AI models for classification tasks. The remaining methods follow scikit-learn classifier syntax.

### Parameters
#### Initialization Parameters
- **base_model** (str) - Optional - Base model to use, by default "text-bison@002".
- **n_update_steps** (int) - Optional - Number of epochs, by default 1.
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
```

--------------------------------

### Import GPTTranslator Class

Source: https://skllm.beastbyte.ai/docs/text-translation

Import the GPTTranslator class for text translation tasks. This is a prerequisite for initializing the translator.

```python
from skllm.models.gpt.text2text.translation import GPTTranslator
```

--------------------------------

### Configure Google Project ID for Vertex AI

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Set your Google Cloud project ID to configure Scikit-LLM for the Vertex AI backend. This is a necessary step after authenticating.

```python
from skllm.config import SKLLMConfig

SKLLMConfig.set_google_project("<YOUR_PROJECT_ID>")
```

--------------------------------

### Set OpenAI API Key and Organization ID

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Configure your OpenAI API key and organization ID to use the default OpenAI backend. Ensure these are set before making any API calls.

```python
from skllm.config import SKLLMConfig

SKLLMConfig.set_openai_key("<YOUR_KEY>")
SKLLMConfig.set_openai_org("<YOUR_ORGANIZATION_ID>")

```

--------------------------------

### Combine GPTVectorizer with XGBoost in a Pipeline

Source: https://skllm.beastbyte.ai/docs/text-vectorization

Integrate GPTVectorizer into a scikit-learn pipeline with an XGBoost classifier for end-to-end text classification. Ensure labels are encoded before fitting.

```python
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import LabelEncoder
from xgboost import XGBClassifier

le = LabelEncoder()
y_train_encoded = le.fit_transform(y_train)
y_test_encoded = le.transform(y_test)

steps = [("GPT", GPTVectorizer()), ("Clf", XGBClassifier())]
clf = Pipeline(steps)
clf.fit(X_train, y_train_encoded)
yh = clf.predict(X_test)

```

--------------------------------

### Text Summarization with Focus

Source: https://skllm.beastbyte.ai/docs/text-summarization

This snippet demonstrates how to generate a summary that emphasizes a specific concept by using the `focus` parameter. The `max_words` parameter sets a soft limit on the summary length.

```python
summarizer = GPTSummarizer(model="gpt-3.5-turbo", max_words=15, focus="apples")
```

--------------------------------

### GPTSummarizer Import

Source: https://skllm.beastbyte.ai/docs/text-summarization

This code imports the GPTSummarizer class, which is the primary tool for text summarization in this library.

```python
from skllm.models.gpt.text2text.summarization import GPTSummarizer
```

--------------------------------

### Import MultiLabelGPTClassifier

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Import statement for the MultiLabelGPTClassifier. Use this when dealing with multi-label classification tasks.

```python
from skllm.models.gpt.classification.tunable import MultiLabelGPTClassifier
```

--------------------------------

### MultiLabelFewShotGPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/few-shot-text-classification

Initializes the MultiLabelFewShotGPTClassifier for multi-label text classification. This class handles scenarios where a text can belong to multiple classes simultaneously.

```APIDOC
## MultiLabelFewShotGPTClassifier

### Description
Initializes the MultiLabelFewShotGPTClassifier for multi-label text classification. This class handles scenarios where a text can belong to multiple classes simultaneously.

### Parameters
#### Initialization Parameters
- **model** (str) - Optional - Model to use, by default "gpt-3.5-turbo".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **max_labels** (Optional[int]) - Optional - Maximum labels per sample, by default 5.
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.

### Usage Example
```python
from skllm.models.gpt.classification.few_shot import MultiLabelFewShotGPTClassifier

clf = MultiLabelFewShotGPTClassifier(max_labels=2, model="gpt-4o")
```
```

--------------------------------

### Set Maximum GPU Layers for GGUF Models

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Configure the maximum number of GGUF model layers to offload to the GPU. Use -1 to offload all layers, 0 for all layers on CPU, or a positive integer for a specific number of layers on the GPU.

```python
from skllm.config import SKLLMConfig

SKLLMConfig.set_gguf_max_gpu_layers(-1)

```

--------------------------------

### Import VertexClassifier

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Import statement for the VertexClassifier. This is used for text classification tasks leveraging Vertex AI models.

```python
from skllm.models.vertex.classification.tunable import VertexClassifier
```

--------------------------------

### GPTClassifier Initialization

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Initializes a GPTClassifier for tunable text classification. This class allows fine-tuning of GPT models for classification tasks. The remaining methods follow scikit-learn classifier syntax.

```APIDOC
## GPTClassifier

### Description
Initializes a GPTClassifier for tunable text classification. This class allows fine-tuning of GPT models for classification tasks. The remaining methods follow scikit-learn classifier syntax.

### Parameters
#### Initialization Parameters
- **base_model** (str) - Optional - Base model to use, by default "gpt-3.5-turbo-0613".
- **default_label** (str) - Optional - Default label for failed prediction; if "Random" -> selects randomly based on class frequencies, by default "Random".
- **key** (Optional[str]) - Optional - Estimator-specific API key; if None, retrieved from the global config, by default None.
- **org** (Optional[str]) - Optional - Estimator-specific ORG key; if None, retrieved from the global config, by default None.
- **n_epochs** (Optional[int]) - Optional - Number of epochs; if None, determined automatically; by default None.
- **custom_suffix** (Optional[str]) - Optional - Custom suffix of the tuned model, used for naming purposes only, by default "skllm".
- **prompt_template** (Optional[str]) - Optional - Custom prompt template to use, by default None.
```

--------------------------------

### Set Azure OpenAI API Key and Base URL

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Configure your Azure API key and endpoint to use the Azure OpenAI service. The endpoint is typically found under 'Keys and Endpoint' in your Azure resource management.

```python
from skllm.config import SKLLMConfig
# Found under: Resource Management (Left Sidebar) -> Keys and Endpoint -> KEY 1
SKLLMConfig.set_gpt_key("<YOUR_KEY>")
# Found under: Resource Management (Left Sidebar) -> Keys and Endpoint -> Endpoint
SKLLMConfig.set_azure_api_base("<API_BASE>") # e.g. https://<YOUR_PROJECT_NAME>.openai.azure.com/

```

--------------------------------

### Translate Text using GPTTranslator

Source: https://skllm.beastbyte.ai/docs/text-translation

Translate a dataset to English using the GPTTranslator. Ensure you have the necessary models and datasets available.

```python
from skllm.models.gpt.text2text.translation import GPTTranslator
from skllm.datasets import get_translation_dataset

X = get_translation_dataset()
t = GPTTranslator(model="gpt-3.5-turbo", output_language="English")
translated_text = t.fit_transform(X)
```

--------------------------------

### Import GPTClassifier

Source: https://skllm.beastbyte.ai/docs/tunable-text-classification

Import statement for the GPTClassifier. This is typically used at the beginning of your script.

```python
from skllm.models.gpt.classification.tunable import GPTClassifier
```

--------------------------------

### Zero-Shot Classification Without Training Data

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Perform zero-shot classification by manually providing candidate labels when no training data is available. Uses a sentiment analysis dataset.

```python
from skllm.models.gpt.classification.zero_shot import ZeroShotGPTClassifier
from skllm.datasets import get_classification_dataset

X, _ = get_classification_dataset()

clf = ZeroShotGPTClassifier()
clf.fit(None, ["positive", "negative", "neutral"])
labels = clf.predict(X)

```

--------------------------------

### Zero-Shot Classification with Training Data

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Classify text using ZeroShotGPTClassifier when training data is available to extract candidate labels. Uses a sentiment analysis dataset.

```python
from skllm.models.gpt.classification.zero_shot import ZeroShotGPTClassifier
from skllm.datasets import get_classification_dataset

# demo sentiment analysis dataset
# labels: positive, negative, neutral
X, y = get_classification_dataset()

clf = ZeroShotGPTClassifier(model="gpt-3.5-turbo")
clf.fit(X,y)
labels = clf.predict(X)

```

--------------------------------

### Multi-Label Zero-Shot Classification

Source: https://skllm.beastbyte.ai/docs/zero-shot-text-classification

Classify text into multiple labels simultaneously using MultiLabelZeroShotGPTClassifier. Requires a list of candidate labels and specifies the maximum number of labels to assign.

```python
from skllm.models.gpt.classification.zero_shot import MultiLabelZeroShotGPTClassifier
from skllm.datasets import get_multilabel_classification_dataset

X, _ = get_multilabel_classification_dataset()
candidate_labels = [
    "Quality",
    "Price",
    "Delivery",
    "Service",
    "Product Variety",
    "Customer Support",
    "Packaging",
    "User Experience",
    "Return Policy",
    "Product Information",
]
clf = MultiLabelZeroShotGPTClassifier(max_labels=3)
clf.fit(None, [candidate_labels])
labels = clf.predict(X)

```

--------------------------------

### Clear GGUF Model Cache

Source: https://skllm.beastbyte.ai/docs/introduction-backend-families

Clear the GGUF model cache to free up memory. This is useful when switching between models or in interactive environments like Jupyter notebooks.

```python
from skllm.llm.gpt.clients.llama_cpp.handler import ModelCache

ModelCache.clear()

```
Tokens:
Raw