chancharikm/all_sft_formats_balanced_human_only_20260222_1240_ep6_lr3e5_qwen3-vl-8b
The chancharikm/all_sft_formats_balanced_human_only_20260222_1240_ep6_lr3e5_qwen3-vl-8b is an 8 billion parameter language model fine-tuned from Qwen/Qwen3-VL-8B-Instruct. This model leverages a Qwen3-VL architecture and has a context length of 32768 tokens. It was fine-tuned on the all_sft_formats_balanced_human_only_20260222_1240 dataset, indicating a focus on instruction-following capabilities derived from human-only data. Its primary application is likely in tasks requiring robust instruction adherence and understanding, building upon the multimodal capabilities of its base Qwen3-VL model.
Loading preview...
Overview
This model, chancharikm/all_sft_formats_balanced_human_only_20260222_1240_ep6_lr3e5_qwen3-vl-8b, is an 8 billion parameter language model. It is a fine-tuned iteration of the Qwen/Qwen3-VL-8B-Instruct base model, which is known for its multimodal capabilities, including vision-language understanding. The fine-tuning process utilized the all_sft_formats_balanced_human_only_20260222_1240 dataset, suggesting an emphasis on enhancing instruction-following and conversational abilities based on human-generated data.
Training Details
The model underwent 6 epochs of training with a learning rate of 3e-05. Key hyperparameters included a train_batch_size of 8, eval_batch_size of 8, and a gradient_accumulation_steps of 2, resulting in an effective total train batch size of 128. The optimizer used was adamw_torch_fused with a cosine learning rate scheduler and a warmup ratio of 0.05. The training was conducted across 8 devices, indicating a distributed training setup.
Potential Use Cases
Given its foundation in Qwen3-VL-8B-Instruct and fine-tuning on human-only instruction data, this model is likely suitable for:
- Instruction-following tasks: Generating responses that adhere closely to given instructions.
- Conversational AI: Engaging in more natural and human-like dialogues.
- Multimodal applications: Leveraging its base model's vision-language understanding for tasks involving both text and images, with improved instruction adherence.