| LibreQuickSRNetm2-restore weights | |
| --------------------------------- | |
| Architecture and checkpoint source: | |
| https://github.com/quic/aimet-model-zoo | |
| commit: 1bd2bf5b17cdda9251437c444009b29e1a25054b | |
| license: BSD-3-Clause | |
| copyright: Copyright (c) 2022 Qualcomm Innovation Center, Inc. | |
| Maintained integration audited for current metadata: | |
| https://github.com/qualcomm/ai-hub-models | |
| commit: 16dbeb5e2805d4ada7218026de72e36878717d46 | |
| license: BSD-3-Clause | |
| copyright: Copyright (c) 2025 Qualcomm Technologies, Inc. and/or its subsidiaries. | |
| Official source artifact: | |
| https://github.com/quic/aimet-model-zoo/releases/download/phase_2_january_artifacts/quicksrnet_medium_2x_checkpoint_float32.pth.tar | |
| SHA-256: a0d176b40a649e45a176c3b53f45e0237015f4f2c17b157ef5c81e38c4442a0d | |
| Converted artifact: | |
| LibreQuickSRNetm2-restore.pt | |
| SHA-256: 3f779b461d200704ab82904a110aaa37f41e8258b285e02b7e413b955deb150b | |
| Conversion preserves all 14 learned tensors unchanged, discards training-only | |
| optimizer and history objects, and adds LibreYOLO checkpoint metadata. The | |
| model was trained on DIV2K. The dataset is provenance only and is not included | |
| in this repository. | |