WokeAI/Tankie-LFM2.5-1.2B-SFT-v1
WokeAI/Tankie-LFM2.5-1.2B-SFT-v1 is a 1.2 billion parameter language model, post-trained from LiquidAI/LFM2.5-1.2B-Instruct, with a 32768 token context length. This model is specifically designed to align with Marxist-Leninist-Maoist ideals, drawing knowledge from classical texts and revolutionary theory. It excels at analyzing class struggle, imperialist exploitation, and proletarian revolution from a dialectical-materialist framework. Its primary use is to provide responses and analysis consistent with this specific ideological perspective.
Loading preview...
Model Overview
WokeAI/Tankie-LFM2.5-1.2B-SFT-v1 is a 1.2 billion parameter language model, fine-tuned from the LiquidAI/LFM2.5-1.2B-Instruct base model. It has been specifically post-trained to embody and articulate the principles of Marxism-Leninism-Maoism, drawing from a vast corpus of historical and theoretical texts. The model's responses are anchored in a dialectical-materialist framework, focusing on class struggle, imperialist exploitation, and revolutionary strategy.
Key Capabilities
- Ideological Alignment: Designed to generate responses consistent with Marxist-Leninist-Maoist theory.
- Analytical Framework: Provides analysis of societal issues through the lens of class struggle and imperialist critique.
- Historical Context: Incorporates knowledge from classical Marxist texts, including The Communist Manifesto and Anti-Dühring, and revolutionary movements.
- Context Length: Supports a context window of 32768 tokens, allowing for detailed discussions.
Intended Use Cases
This model is suitable for applications requiring:
- Ideological Discourse: Generating text or engaging in discussions from a specific Marxist-Leninist-Maoist viewpoint.
- Social and Political Analysis: Analyzing current events or historical contexts through a dialectical-materialist perspective.
- Educational Tools: Exploring revolutionary theory and history as interpreted by this specific ideological framework.
Training Details
The model was fine-tuned using the WokeAI/polititune-tankie-warmup-3 dataset over 4 epochs. Key hyperparameters included a learning rate of 5e-05 and a total training batch size of 16. The training utilized adamw_torch_8bit optimizer and a constant learning rate scheduler.