딥러닝 (15) 썸네일형 리스트형 sd, sdxl 아키텍처 https://ostin.tistory.com/231 conv_in down_blocks: (CrossAttnDownBlock2D: ResnetBlock2D Transformer2DModel ResnetBlock2D Transformer2DModel Downsample2D ) x 3 DownBlock2D: (ResnetBlock2D) x 2 mid_blocks: ResnetBlock2D Transformer2DModel ResnetBlock2D up_blocks: UpBlock2D: (ResnetBlock2D) x 3 Upsample2D (Cr.. [python] print self.face_rec_model_path 에러 python3.10에서 제공하는 ldm라이브러리가 아닌 laten-diffusion 에서 제공하는 ldm이 필요하다. from ldhttp://m.modules.diffusionmodules.util import make_ddim_sampling_parameters, make_ddim_timesteps, noise_like, extract_into_tensor File "/home/erica/.pyenv/versions/3.10.12/lib/python3.10/site-packages/ldm.py", line 20 print self.face_rec_model_path ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 해결방법 : 기존 ldm제거 pip uninstall ldm 신규 stable d.. [에러해결] error: RPC failed; curl 56 GnuTLS recv error (-9): Error decoding the received TLS packet. 위와 같은 에러 발생시, 아래 3줄 실행 sudo apt install gnutls-bin git config --global http.postBuffer 1048576000 // 1G git config --global http.sslVerify false 기타 이슈도 아래 링크에서 잘 정리되어있는듯하다. https://yoech.com/thread-498.htm [diffusers] controlnet 코드 분석 코드 매우 심플한데, 효과적인 생성방법. down block, mid block 부분 controlnet에서 떼옴 . down_block_res_samples, mid_block_res_sample = controlnet( noisy_latents, timesteps, encoder_hidden_states=encoder_hidden_states, controlnet_cond=controlnet_image, return_dict=False, )unet 의 down block부분과 mid block부분을 업데이.. [prompt to prompt] 설치하기 & 에러들 1) attentionReplace 2) Refinement : 단어가 추가되어 refinement 필요 3) reweight => 일단 궁금한것 refinement!! 단어길이가 다른데, 어떻게 처리했음? Attention Control Options 더보기 cross_replace_steps: specifies the fraction of steps to edit the cross attention maps. Can also be set to a dictionary [str:float] which specifies fractions for different words in the prompt. self_replace_steps: specifies the fraction of steps to repla.. [딥러닝 읽기좋은 블로그들] 1) ViT : 자세히 설명 잘하심 https://hongl.tistory.com/234 Vision Transformer (3) - Attention Map Vision Transformer (1) Vision Transformer (2) Transformer 모델의 가장 큰 특징은 self-attention 으로 시퀀스의 각 위치가 어느 위치에 집중하는지 쉽게 시각화해서 볼 수 있다는 점입니다. 이미지 기반 딥러닝에 hongl.tistory.com 2) attention head https://codingopera.tistory.com/44 hugging face ssl error (requests.exceptions.SSLError: HTTPSConnectionPool(host='huggingface.co', port=443) requests.exceptions.SSLError: HTTPSConnectionPool(host='huggingface.co', port=443): Max retries exceeded with url: /bert-base-uncased/resolve/main/tokenizer_config.json (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:997)'))) 가상 환경내에 vi /opt/conda/envs/sd/lib/python3.10/site-packages.. yolov5 모르는 용어 정리 1. InfiniteDataLoader vs DataLoader 차이 > gpt 님이 말씀하사... DataLoader is a general-purpose data loader that loads a finite dataset in PyTorch. It provides an iterable over a given dataset that loads a specified number of samples at a time and applies transformations to each sample. It is commonly used in machine learning tasks such as image classification, object detection, and natural language pro.. 이전 1 2 다음
