transformers

a3881a8d - perf: Remove implicit CPU-GPU syncs due to implicit .item() call (#42433)

perf: Remove implicit CPU-GPU syncs due to implicit .item() call (#42433) * perf: Remove implicit CPU-GPU syncs due to implicit .item() call * fix: replicated the changes across similar files * fix: update the newly added nanochat model files * fix: use input_shape and device instead of input_emdeds properties for imagegpt