mango

输出图片尺寸问题

5

输入是4k(3840*2160),AGCM和LE输出均与输入相同，但是HG输出尺寸为(3840*2176)，我定位了一下，是因为generate_mask.py代码中 if H%32!=0 or W%32!=0: H_new = int(np.ceil(H / 32) * 32) W_new = int(np.ceil(W / 32) * 32) img_LQ = cv2.resize(img_LQ, (W_new, H_new)) 请问为什么在HG网络中尺寸需要调整成可以整除32的呢？输出尺寸变化的问题怎么解决呢？

about alignratio in the training dataset

1

What does the value of alignratio in the training dataset mean? The range of gt_image/alignratio is not at [0,1]. What does the ratio do ? How are the alignratio in...

In bsvd_arch.py , if noise_map != None: input = torch.cat([input, noise_map], dim=2) please, what does the noise_map do?

in the paper, "Unlike MIMO frameworks, our buffers work in a pipeline that consumes frames one by one. Therefore, our method can keep a constant run-time memory, which depends only...

我看到您loss.py中perception_loss与SRGAN中相同，但是ESRGAN中作者指出该损失应该取激活之前的特征进行求得，请问根据您使用的loss情况，超分效果如何？指标是否下降？

FileNotFoundError: [Errno 2] No such file or directory: '/cluster/work/cvl/videosr/pretrained_models/spynet/spynet.pth'

5

请问spynet.pth模型链接在哪里呢？

加载权重，报错

1

sh scripts/predict_finerune_gpt2.sh 2024-02-27 15:20:22.435851: I tensorflow/core/util/port.cc:110] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them...

File "finetune.py", line 412, in <module> train() File "finetune.py", line 384, in train model = get_peft_model(model, lora_config) File "/opt/conda/envs/qwen/lib/python3.8/site-packages/peft/mapping.py", line 123, in get_peft_model peft_config.base_model_name_or_path = model.__dict__.get("name_or_path", None) AttributeError: 'NoneType' object has no attribute '__dict__'[BUG] <title>

1

### 是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this? - [X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions ### 该问题是否在FAQ中有解答？ | Is there an...

Reproducing code，NIQE=5.6567

1

Data processing is consistent with RealBasicVSR，implement the HSP module as described in the paper,train data is REDS,test data is VideoLQ，the calculated NIQE is higher than the paper, and the details...

about align_ratio

1

What does the value of align_ratio in the training dataset mean? How did you get that?