Add training manifest

Publish trained GarudaCoder Qwen3.8 27B LoRA

Upload training/manifest_qwen38.json with huggingface_hub

Qwen3.8-27B v1 Kaggle T4x2: audited SFT, pinned base, holdout manifest (Trained with Unsloth)

Tambah section Adaptive thinking (v2: hemat ~60% inference untuk pertanyaan mudah)

Tambah section Identity reinforcement (v2: 30 contoh identity, 2 epoch, lr 1e-5)

Manifest bersih (tanpa nama file sumber training)

Model card (bersih — tanpa detail sumber training)

Model card lengkap (angka dari manifest_qwen38.json)

Qwen3.8-27B v1 Kaggle T4x2: audited SFT, pinned base, holdout manifest (Trained with Unsloth)

Upload training/manifest_qwen38.json with huggingface_hub

Upload model trained with Unsloth

Qwen3.8-27B v1 Kaggle T4x2: audited SFT, pinned base, holdout manifest (Trained with Unsloth)

Upload README.md with huggingface_hub

initial commit