rxai-sdg 0.1.43

Reactive AI - Synthetic Datasets Generator

Reactive AI: Synthetic Dataset Generators (rxai-sdg)

Toolkit for generating high-quality synthetic datasets for training Reactive Transformer models. Supports Memory Reinforcement Learning (MRL), Supervised Fine-Tuning (SFT), and the new Hybrid Reasoning generators for RxT-Beta.

Installation

pip install rxai-sdg

Or install from source:

git clone https://github.com/RxAI-dev/rxai-sdg.git
cd rxai-sdg
pip install -e .

Overview

This library provides synthetic dataset generators for training Reactive Language Models:

Module	Purpose	Target Training Stage
rxai_sdg.mrl	Memory Reinforcement Learning datasets	MRL stage
rxai_sdg.sft	Supervised Fine-Tuning datasets	Interaction SFT
rxai_sdg.hybrid	Hybrid Reasoning & DMPO datasets	RxT-Beta advanced training

Quick Start

API Configuration

All generators support both OpenAI-compatible APIs and Ollama for local testing:

# OpenAI-compatible API (default)
generator = MrlSyntheticDatasetGenerator(
    max_items=100,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="your-api-key",
    use_ollama=False
)

# Ollama local testing
generator = MrlSyntheticDatasetGenerator(
    max_items=100,
    model_name="llama3.2",
    api_url="http://localhost:11434",
    use_ollama=True
)

# Third-party providers (Novita.ai, Together.ai, etc.)
generator = MrlSyntheticDatasetGenerator(
    max_items=100,
    model_name="qwen/qwen3-4b-fp8",
    api_url="https://api.novita.ai/v3/openai",
    api_key="your-key"
)

Memory Reinforcement Learning (MRL) Datasets

Generate multi-turn conversations testing memory retention:

from rxai_sdg.mrl import (
    MrlSyntheticDatasetGenerator,
    MrlPromptCreator,
    MrlGeneratorPostprocessor,
)
from rxai_sdg.mrl.prompts import ALL_PROMPTS_REAL, TOPICS_REAL
from rxai_sdg.mrl.examples import EXAMPLES_REAL_MICRO

# Initialize generator
generator = MrlSyntheticDatasetGenerator(
    max_items=500,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="your-key"
)

# Create prompt creator with topics
prompt_creator = MrlPromptCreator(
    prompts=ALL_PROMPTS_REAL,
    examples=EXAMPLES_REAL_MICRO,
    topics=TOPICS_REAL
)

# Generate dataset
generator(
    prompt_creator=prompt_creator,
    steps=3,           # Follow-up interactions per conversation
    iterations=50,     # API calls
    num_examples=10,   # Examples per API call
    num_topics=10,     # Random topics per prompt
    temperature=0.7,
    stream=True,       # Show generation progress
    max_tokens=20000
)

# Post-process and export
postprocessor = MrlGeneratorPostprocessor(
    generator=generator,
    dataset_id="your-org/mrl-dataset",
    token="hf_token"
)
postprocessor.filter_duplicates()
postprocessor.remove_incorrect_interactions(steps=3)
postprocessor.push_to_hf_hub()

# Or get as Dataset object
dataset = generator.get_dataset()

MRL Dataset Format

{
    'query': ["Initial question 1", "Initial question 2", ...],
    'answer': ["Initial answer 1", "Initial answer 2", ...],
    'interactions': [
        [
            {'query': "Follow-up Q1", 'answer': "Follow-up A1"},
            {'query': "Follow-up Q2", 'answer': "Follow-up A2"},
            ...
        ],
        ...
    ]
}

Generation Modes

Multi-Topic Mode (default): Single topic with progressive memory testing

generator(prompt_creator, steps=3, mode='multi')

Long-Range Mode: Two-topic strategy testing long-range memory

generator(prompt_creator, steps=5, mode='long')

RxT-Beta Hybrid Reasoning Datasets

The rxai_sdg.hybrid module provides generators for RxT-Beta's advanced training stages:

1. Reasoning Completion Generator

Add missing 'think' blocks to existing conversations:

from rxai_sdg.hybrid import ReasoningCompletionGenerator
from datasets import load_dataset

# Load existing dataset with missing think blocks
dataset = load_dataset("your-dataset", split="train")
# Expected format: {'interactions': [[{'query': ..., 'think': '', 'answer': ...}, ...]]}

generator = ReasoningCompletionGenerator(
    max_items=100,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="your-key"
)

# Mode 1: Generate think blocks one at a time (higher quality)
generator.complete_single(
    dataset=dataset,
    target_tokens=512,
    temperature=0.7,
    stream=True
)

# Mode 2: Generate all think blocks at once (more efficient)
generator.complete_all_at_once(
    dataset=dataset,
    target_tokens_per_think=512,
    temperature=0.7
)

# Get completed dataset
completed_dataset = generator.get_dataset()

2. Hybrid Reasoning Generator

Create new conversations with full reasoning chains from scratch:

from rxai_sdg.hybrid import (
    HybridReasoningGenerator,
    HybridReasoningPromptCreator,
    TOPICS_HYBRID_REASONING,
)

# Initialize
generator = HybridReasoningGenerator(
    max_items=100,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="your-key"
)

# Custom topics (or use defaults)
my_topics = [
    "Quantum computing fundamentals",
    "Climate change feedback loops",
    "Machine learning optimization",
    # ...
]

prompt_creator = HybridReasoningPromptCreator(
    topics=my_topics,  # or TOPICS_HYBRID_REASONING
    include_examples=True
)

# Mode 1: Generate one interaction at a time (builds context progressively)
generator.generate_single(
    prompt_creator=prompt_creator,
    num_interactions=5,    # Interactions per conversation
    conversations=20,      # Number of conversations
    target_tokens=1024,    # Tokens per interaction
    thinking_ratio=0.7,    # 70% use extended thinking
    temperature=0.7,
    stream=True
)

# Mode 2: Generate entire conversations at once
generator.generate_all_at_once(
    prompt_creator=prompt_creator,
    num_interactions=5,
    conversations=20,
    target_tokens_per_interaction=1024,
    thinking_ratio=0.7,
    temperature=0.7
)

dataset = generator.get_dataset()

Hybrid Reasoning Dataset Format

{
    'interactions': [
        [
            {'query': "Question 1", 'think': "Reasoning...", 'answer': "Response 1"},
            {'query': "Question 2", 'think': "Reasoning...", 'answer': "Response 2"},
            ...
        ],
        ...
    ],
    'topics': ["Topic 1", "Topic 2", ...]
}

3. DMPO (Direct Memory and Preference Optimization) Generator

Create preference pairs for memory-aware training:

from rxai_sdg.hybrid import DMPOGenerator, DMPOPromptCreator

generator = DMPOGenerator(
    max_items=100,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="your-key"
)

prompt_creator = DMPOPromptCreator(
    topics=TOPICS_HYBRID_REASONING,
    include_examples=True
)

# Mode 1: Generate pairs one at a time
generator.generate_single(
    prompt_creator=prompt_creator,
    num_interactions=5,
    conversations=20,
    target_tokens=1024,
    temperature=0.7
)

# Mode 2: Generate entire preference conversations at once
generator.generate_all_at_once(
    prompt_creator=prompt_creator,
    num_interactions=5,
    conversations=20,
    target_tokens_per_interaction=1024
)

dataset = generator.get_dataset()

DMPO Dataset Format

Each interaction contains accepted (good) and rejected (bad) responses:

{
    'interactions': [
        [
            {
                'query': "Question requiring memory...",
                'accepted': {
                    'think': "Good reasoning with correct memory usage...",
                    'answer': "Accurate, helpful response..."
                },
                'rejected': {
                    'think': "Flawed reasoning or memory errors...",
                    'answer': "Response with issues..."
                }
            },
            ...
        ],
        ...
    ],
    'topics': ["Topic 1", "Topic 2", ...]
}

Postprocessing

from rxai_sdg.hybrid import HybridGeneratorPostprocessor

postprocessor = HybridGeneratorPostprocessor(
    generator=generator,
    dataset_id="your-org/hybrid-dataset",
    token="hf_token"
)

# Filter empty/invalid conversations
postprocessor.filter_empty_interactions()

# Filter by conversation length
postprocessor.filter_by_length(min_interactions=3, max_interactions=10)

# Convert to RxT-Beta format
rxt_format = postprocessor.convert_to_rxt_format()
# Returns: [{'formatted': '[Q] query [T] think [A] answer', ...}, ...]

# Push to HuggingFace Hub
postprocessor.push_to_hf_hub()

RxT-Beta Interaction Template

The hybrid generators produce data compatible with RxT-Beta's interaction template:

Mode	Template	Description
Fast Answer	[Q] query [A] answer	Direct response without reasoning
Extended Thinking	[Q] query [T] thinking [A] answer	Response with reasoning chain
Tool Usage	[U] tool_result [T] thinking [A] answer	Processing tool results
Internal Instruction	[I] instruction [Q] query [A] answer	Per-interaction behavior control
Tool Call	[Q] query [A] answer [C] tool_call	Invoking external tools

Convenience Functions

from rxai_sdg.hybrid import (
    create_reasoning_completion_generator,
    create_hybrid_reasoning_generator,
    create_dmpo_generator,
)

# Quick setup with defaults
completion_gen = create_reasoning_completion_generator(
    max_items=100,
    model_name="gpt-4",
    api_key="your-key"
)

# Returns (generator, prompt_creator) tuple
reasoning_gen, reasoning_prompts = create_hybrid_reasoning_generator(
    max_items=100,
    model_name="gpt-4",
    api_key="your-key",
    topics=my_custom_topics
)

dmpo_gen, dmpo_prompts = create_dmpo_generator(
    max_items=100,
    model_name="gpt-4",
    api_key="your-key"
)

API Reference

Base Classes

BaseDatasetGenerator

Abstract base class for all generators.

class BaseDatasetGenerator(ABC):
    def __init__(
        self,
        max_items: int = None,           # Maximum items to generate
        model_name: str = "...",          # Model identifier
        api_url: str = "...",             # API endpoint
        api_key: str = None,              # API authentication
        use_ollama: bool = False          # Use Ollama instead of OpenAI API
    )

    def generate_items(
        self,
        prompt: str,
        stream: bool = False,
        temperature: float = 0.7,
        top_p: float = 0.9,
        max_tokens: int = 15000,
        system_prompt: str = "",
        timeout: int = 120,
        additional_config: dict = None
    ) -> str

    def get_dataset(self) -> Dataset  # Return HuggingFace Dataset

MRL Module

• MrlSyntheticDatasetGenerator - Main generator
• MrlPromptCreator - Prompt composition
• MrlContextBasedPromptCreator - Context-aware prompts
• MrlGeneratorPostprocessor - Post-processing and export

Hybrid Module

• ReasoningCompletionGenerator - Add missing think blocks
• HybridReasoningGenerator - Create reasoning conversations
• DMPOGenerator - Create preference pairs
• HybridReasoningPromptCreator - Prompts for reasoning generation
• DMPOPromptCreator - Prompts for DMPO generation
• HybridGeneratorPostprocessor - Post-processing utilities

Configuration Options

Generation Parameters

Parameter	Default	Description
temperature	0.7	Sampling temperature (0-1)
top_p	0.9	Nucleus sampling threshold
max_tokens	15000	Maximum tokens per generation
timeout	120	Request timeout in seconds
stream	False	Stream responses in real-time

Additional Config

additional_config = {
    'presence_penalty': 0,
    'frequency_penalty': 0,
    'response_format': {"type": "text"},
    'extra_body': {
        "top_k": 50,
        'repetition_penalty': 1,
        'min_p': 0,
    },
}

generator.generate_items(..., additional_config=additional_config)

Examples

Complete Workflow: MRL Dataset

from rxai_sdg.mrl import *
from rxai_sdg.mrl.prompts import ALL_PROMPTS_REAL, TOPICS_REAL

# 1. Initialize
generator = MrlSyntheticDatasetGenerator(
    max_items=1000,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="sk-..."
)

prompt_creator = MrlPromptCreator(
    prompts=ALL_PROMPTS_REAL,
    topics=TOPICS_REAL
)

# 2. Generate
for steps in [2, 3, 4, 5]:  # Multiple conversation lengths
    generator(
        prompt_creator=prompt_creator,
        steps=steps,
        iterations=25,
        num_examples=10,
        temperature=0.8,
        stream=True
    )

# 3. Post-process
postprocessor = MrlGeneratorPostprocessor(
    generator=generator,
    dataset_id="myorg/mrl-dataset",
    token="hf_..."
)
postprocessor.filter_duplicates()
postprocessor.push_to_hf_hub()

Complete Workflow: Hybrid Reasoning

from rxai_sdg.hybrid import *

# 1. Initialize
generator = HybridReasoningGenerator(
    max_items=500,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="sk-..."
)

prompt_creator = HybridReasoningPromptCreator()

# 2. Generate different conversation lengths
for length in [3, 5, 7]:
    generator(
        prompt_creator=prompt_creator,
        num_interactions=length,
        conversations=50,
        mode='single',  # Higher quality
        temperature=0.7,
        stream=True,
        restart=False   # Accumulate
    )

# 3. Post-process
postprocessor = HybridGeneratorPostprocessor(
    generator=generator,
    dataset_id="myorg/hybrid-reasoning",
    token="hf_..."
)
postprocessor.filter_empty_interactions()
postprocessor.push_to_hf_hub()

Complete Workflow: DMPO Dataset

from rxai_sdg.hybrid import *

# 1. Initialize
generator = DMPOGenerator(
    max_items=300,
    model_name="gpt-4",
    api_url="https://api.openai.com/v1",
    api_key="sk-..."
)

prompt_creator = DMPOPromptCreator()

# 2. Generate
generator(
    prompt_creator=prompt_creator,
    num_interactions=5,
    conversations=60,
    mode='single',
    target_tokens=1024,
    temperature=0.7
)

# 3. Export
postprocessor = HybridGeneratorPostprocessor(
    generator=generator,
    dataset_id="myorg/dmpo-dataset",
    token="hf_..."
)
postprocessor.push_to_hf_hub()

Ollama Local Testing

For local development and testing with Ollama:

# Start Ollama
ollama serve

# Pull a model
ollama pull llama3.2

from rxai_sdg.hybrid import HybridReasoningGenerator, HybridReasoningPromptCreator

# Use Ollama
generator = HybridReasoningGenerator(
    max_items=10,
    model_name="llama3.2",
    api_url="http://localhost:11434",
    use_ollama=True
)

prompt_creator = HybridReasoningPromptCreator()

# Generate (smaller batches for local testing)
generator(
    prompt_creator=prompt_creator,
    num_interactions=3,
    conversations=5,
    mode='single',
    temperature=0.8,
    stream=True
)

License

Apache-2.0

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

Links

• Repository

• RxT-Beta Model

• Reactive Transformer Paper

                             Apache License
                       Version 2.0, January 2004
                    http://www.apache.org/licenses/
  

  TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

  1. Definitions.

     "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document.

     "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License.

     "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.

     "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License.

     "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files.

     "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types.

     "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below).

     "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof.

     "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution."

     "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work.

  2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form.

  3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed.

  4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions:

     (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and

     (b) You must cause any modified files to carry prominent notices stating that You changed the files; and

     (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and

     (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License.

     You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License.

  5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions.

  6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file.

  7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License.

  8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages.

  9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability.

  END OF TERMS AND CONDITIONS

  APPENDIX: How to apply the Apache License to your work.

  To apply the Apache License to your work, attach the following
  boilerplate notice, with the fields enclosed by brackets "[]"
  replaced with your own identifying information. (Don't include
  the brackets!)  The text should be enclosed in the appropriate
  comment syntax for the file format. We also recommend that a
  file or class name and description of purpose be included on the
  same "printed page" as the copyright notice for easier
  identification within third-party archives.
  

  Copyright [yyyy] [name of copyright owner]

  Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

   http://www.apache.org/licenses/LICENSE-2.0
  

  Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.