Chennzi Lite Osworld v2 Qwen3.5 9b Sft Reasoning Config Ep3

--- library_name: transformers base_model: Qwen/Qwen3.5-9B tags: - qwen3.5 - cua-lite - lite-osworld - sft - ep3 --- # chennzi-lite-osworld-v2-qwen3.5-9b-sft-reasoning-config-ep3 Final HuggingFace-format checkpoint from `eval-bbq-2`. - Base model: `Qwen/Qwen3.5-9B` - Training data: `cua-lite/Lite.OSWorld` (`desktop/use/train`, 1940 exported trajectories) - Epochs: 3 - Global batch size: 32 - Final checkpoint: `iter_179` - Export/training config: `examples/lite/v1/configs/qwen3_5/reasoning/lite.osworld.yaml` Important difference from the earlier `eval-bbq-1` upload: this run changed the export config from the old SFT default to the Lite.OSWorld reasoning config, including longer history and inline reasoning formatting.

Source context: 13 downloads · 0 likes · Pipeline image-text-to-text · Library transformers · Repo OnAnOrange/chennzi-lite-osworld-v2-qwen3.5-9b-sft-reasoning-config-ep3

Source context: 15 downloads · 0 likes · Pipeline image-text-to-text · Library transformers · Repo OnAnOrange/chennzi-lite-osworld-v2-qwen3.5-9b-sft-reasoning-config-ep3