Movie-Qwen1.5B — Domain SFT (IMDb × Wikipedia)

Date: 2025-10-14
Single-process model-parallel SFT with LoRA FP16 (T4×2). Answer-only loss. Time-capped.

See usage snippet in repo.

Downloads last month: 6

Safetensors

Model size

2B params

Tensor type

F16

·

Model tree for nhonhoccode/movie-qwen1p5b-merged

Base model

Qwen/Qwen2.5-1.5B

Finetuned

Qwen/Qwen2.5-1.5B-Instruct

Adapter

(815)

this model

Dataset used to train nhonhoccode/movie-qwen1p5b-merged