| license: mit | |
| datasets: | |
| - Vikhrmodels/GrandMaster-PRO-MAX | |
| language: | |
| - ru | |
| - en | |
| base_model: | |
| - Qwen/Qwen2-VL-2B-Instruct | |
| pipeline_tag: text2text-generation | |
| tags: | |
| - multimodal | |
| library_name: transformers | |
| # tvl-mini | |
| ## Description | |
| This is finetune of Qwen2-VL-2B on russian language. | |
| tvl was trained in bf16 | |
| ## Data | |
| Train dataset contains: | |
| - GrandMaster-PRO-MAX dataset (60k samples) | |
| - Translated, humanized and merged by image subset of GQA (TODO) |