-
Pillow (and PIL) is currently able to open 8 bit per channel multi-channel images (such as RGB) but is able to open higher bit depth images (e.g. I16, I32, or Float32 images) if they are single channe…
-
如题，如果想重建特征，为什么不能阻止模型是全1的平凡形式（即强行让输入等于输出）输出一个loss为0的解？即使说模型非常复杂，很难达到平凡参数，但是模型趋近于平凡参数，loss也会接近0.这样的参数是没有意义的。无论说中间模型的结构多么复杂，我们怎么能够确保模型是在学习数据本身的pattern而非单纯复刻数据呢？请问可以帮我解答一下这个问题吗？感谢！
-
Phd track: 21 Nov 2021 - 21 Nov 2025
First step in 4 year phd track: get paper from master thesis work.
Venues: https://github.com/Tribler/tribler/wiki/Scientific-publication-venues-for-ledger-scien…
-
Hi,
Thank you for the great work. I tried reproducing the results by downloading `Llama-1-7b-E8P-2Bit` and using `eval/eval_ppl.py`. The PPL is very high around 100. Do you have any suggestion on w…
-
### Feature request
Our research group has studied speeding up LoRA training with new optimizer features, and we would like to contribute to integrating our proposed method in current huggingface lib…
-
Hi, thank you for sharing this wonderful survey! It is very detailed and enlightening. Do you have any plans about incorporating the following works:
1. [CrossGET: Cross-Guided Ensemble of Tokens for…
sdc17 updated
6 months ago
-
Hi,
Thank you for maintaining the comprehensive list of papers on long-tailed learning.
We would like to introduce two recent papers to be considered for inclusion:
1. **Harnessing Hierarchic…
-
Hi,
Thanks for this cool work. Would you mind appending our work on diffusion image codec?
idempotence and perceptual image compression, ICLR 2024
Arxiv: https://arxiv.org/abs/2401.08920
Code:…
-
Hi,
Thanks for this wonderful repo!
We believe that our paper on MTL, which was accepted to ICML 2024, may be relevant here.
paper: https://arxiv.org/abs/2402.04005
code: https://github.com/ssi-re…
-
Hi,
Thanks for your excellent library! I am one of the co-authors of [MOMENT](https://arxiv.org/abs/2402.03885), a foundation model for time series data. I am one of the co-authors, and I would love…