

/kodinkuvalehti.fi/s3fs-public/wysiwyg_images/dsc01420.jpg)
The 95 GHz class I methanol maser is generally stronger than the 84 GHz maser counterpart. The new detections increase the number of known 104.3 GHz masers from 5 to 9.

Among them, fifty 84 GHz masers, twenty nine 95 GHz masers and four rare. We detect narrow, maser-like features towards 54, 100 and 4 sources in the maser lines near 84, 95 and 104.3 GHz, respectively. We analyzed the 3-mm wavelength spectral line survey of 408 ATLASGAL clumps observed with the IRAM 30m-telescope, focusing on the class I methanol masers with frequencies near 84, 95 and 104.3 GHz. Extensive experiments are conducted on VOC-LT and COCO-LT datasets, which demonstrates that the proposed method significantly surpasses the previous state-of-the-art methods and zero-shot CLIP in LTML. Furthermore, taking into account the class imbalance, the distribution-balanced loss is adopted as the classification loss function to further improve the performance on the tail classes without compromising head classes. Specifically, LMPT introduces the embedding loss function with class-aware soft margin and re-weighting to learn class-specific contexts with the benefit of textual descriptions (captions), which could help establish semantic relationships between classes, especially between the head and tail classes. performance synchronously on both head and tail classes. In this work, we propose a unified framework for LTML, namely prompt tuning with class-specific embedding loss (LMPT), capturing the semantic feature interactions between categories by combining text and image modality data and improving the. Long-tailed multi-label visual recognition (LTML) task is a highly challenging task due to the label co-occurrence and imbalanced data distribution.
