ChesterProgrammer/Qwen3.5-9B-ConsistentChat-100steps
ChesterProgrammer/Qwen3.5-9B-ConsistentChat-100steps is a 9 billion parameter Qwen3.5-based causal language model developed by ChesterProgrammer. Fine-tuned using Unsloth and Huggingface's TRL library, this model is specifically optimized for consistent chat interactions. It was trained for 100 steps on the jiawei-ucas/ConsistentChat dataset, focusing on conversational coherence.
Loading preview...
Model Overview
ChesterProgrammer/Qwen3.5-9B-ConsistentChat-100steps is a 9 billion parameter language model, fine-tuned by ChesterProgrammer from the unsloth/Qwen3.5-9B base model. This iteration was trained for 100 steps using Unsloth for accelerated training and Huggingface's TRL library, specifically targeting enhanced conversational consistency.
Key Training Details
- Base Model:
unsloth/Qwen3.5-9B - Training Method: LoRA (16-bit) with Unsloth and TRL
- Dataset:
jiawei-ucas/ConsistentChat - Max Steps: 100
- Context Length: 2048 tokens (during fine-tuning)
- LoRA Configuration: Rank 16, Alpha 32, Dropout 0
Intended Use Cases
This model is particularly suited for applications requiring:
- Consistent Chat Interactions: Its fine-tuning on the
ConsistentChatdataset suggests an optimization for maintaining coherence and context in conversational exchanges. - Efficient Deployment: Being a LoRA-finetuned model, it offers a smaller footprint for deployment compared to a full fine-tune, especially beneficial for resource-constrained environments.