Tesslate/WEBGEN-2-SFT
Tesslate/WEBGEN-2-SFT is a 4 billion parameter instruction-tuned causal language model, fine-tuned from unsloth/Qwen3-4B-Instruct-2507. This model was trained using SFT with TRL and features a context length of 40960 tokens. It is designed for general text generation tasks, leveraging its Qwen3 base for robust language understanding and generation capabilities.
Loading preview...
Overview
Tesslate/WEBGEN-2-SFT is a 4 billion parameter instruction-tuned language model, built upon the unsloth/Qwen3-4B-Instruct-2507 base model. It has been fine-tuned using the SFT (Supervised Fine-Tuning) method with the TRL library, indicating a focus on improving its ability to follow instructions and generate coherent responses. The model supports a substantial context length of 40960 tokens, allowing it to process and generate longer sequences of text.
Key Capabilities
- Instruction Following: Enhanced through SFT, enabling it to respond effectively to user prompts and questions.
- Text Generation: Capable of generating diverse and contextually relevant text based on its Qwen3 architecture.
- Extended Context: Benefits from a 40960-token context window, suitable for tasks requiring extensive input or generating lengthy outputs.
Training Details
The model's training procedure involved Supervised Fine-Tuning (SFT) using the TRL framework. This process typically refines a pre-trained model's behavior to align better with specific task requirements or conversational styles. The development utilized specific versions of key frameworks including TRL 0.24.0, Transformers 4.57.1, Pytorch 2.9.0, Datasets 4.3.0, and Tokenizers 0.22.2.
Good For
- General-purpose text generation based on instructions.
- Applications requiring a model with a relatively large context window for processing detailed prompts or generating comprehensive responses.