datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
neuralatlas-attributions-convnext_tinydanbooru-convnext-embeddings3E2AM_ConvNeXtV2_CIFAR10DeLR-Cephalometric-ConvNeXtV2
DeLR – Dual-encoder Landmark Regression
A PyTorch implementation of the DeLR (Dual-encoder Landmark Regression) architecture for cephalometric landmark detection, evaluated on three public datasets:
Aariz Cephalograms — 1000 images, 29 annotated landmarks (700 / 150 / 150 train/valid/test).
CephAdoAdu Dataset — 700 images, 10 landmarks, mixed adolescent + adult cohort (400 train / 300 test in the official splits; we held out 10 % of train as validation).
ISBI 2015 Cephalometric… See the full description on the dataset page: https://huggingface.co/datasets/emad2001/DeLR-Cephalometric-ConvNeXtV2.E2AM_ConvNeXtV2Nano
E2AM Ablation Results: ConvNeXtV2-Nano
Energy-aware training ablation study for ConvNeXtV2-Nano across three image-classification datasets: CIFAR-10, CIFAR-100, and Tiny-ImageNet.
Each dataset has 15 training variants (8 individual-method M0..M7, 7 cumulative ablation C0..C6) at 50 epochs, plus a 5-variant deployment pipeline (FP32 baseline, structured pruning, pruning+finetune, INT8 quantization, pruned+INT8).
Status: 45 completed variants, 0 partial. 15 deployment runs.… See the full description on the dataset page: https://huggingface.co/datasets/Shanmuk4622/E2AM_ConvNeXtV2Nano.E2AM_ConvNeXtV2Tiny
E2AM Ablation Results: ConvNeXtV2-Tiny
Energy-aware training ablation study for ConvNeXtV2-Tiny across three image-classification datasets: CIFAR-10, CIFAR-100, and Tiny-ImageNet.
Each dataset has 15 training variants (8 individual-method M0..M7, 7 cumulative ablation C0..C6) at 50 epochs, plus a 5-variant deployment pipeline (FP32 baseline, structured pruning, pruning+finetune, INT8 quantization, pruned+INT8).
Status: 45 completed variants, 0 partial. 15 deployment runs.… See the full description on the dataset page: https://huggingface.co/datasets/Shanmuk4622/E2AM_ConvNeXtV2Tiny.Vision_ConvNextconvnext_cifar10_democonvnext
Dataset Card for "convnext"
More Information needed
Results_exp_convnext_droppath_linearConvNext-aesthetic-ratereuroSAT-convnextconvnextExtracting features from the massive danbooru anime image dataset using a computer vision model.
Fields include:
The feature vectors generated by the ConvNeXt Large model, the embeddings.
May contain image_id or tags corresponding to the original dataset.
convnext-weakfix-datasets
convnext-weakfix — data + kết quả + checkpoint
Replicate phương pháp v5_weakfix (hoangtuan) lên DINOv3 ConvNeXt-Tiny (dinov3_next_cnn).
Cùng data / kỹ thuật / hyperparam, identity-disjoint. Doc đầy đủ tại repo gốc:
/workspace/quangmanh/deepfake/doc/convnext_weakfix/.
Data (data/)
train_v5_combined_universal_kaggle_boost.csv — 54,000 ảnh (baseline).
train_v5_weakfix.csv — v2: 121,884 ảnh (31,006 real / 90,878 fake, 42 method, faceswap 4,600).… See the full description on the dataset page: https://huggingface.co/datasets/ManhQuangAI/convnext-weakfix-datasets.finetuning_convnext_datadanbooru-convnext-embeddings2
Dataset Card for Danbooru ConvNeXt Embeddings 2
Danbooru ConvNeXt 向量数据集 2
Dataset Details / 数据集详情
Dataset Description / 数据集描述
English:
This dataset contains approximately 5,312,000 image embeddings (vectors). It was generated by extracting features from the massive Danbooru anime image dataset using the convnext_large.dinov3_lvd1689m computer vision model. These embeddings represent the visual features of the images in a high-dimensional space… See the full description on the dataset page: https://huggingface.co/datasets/telecomadm1145/danbooru-convnext-embeddings2.E2AM_ConvNeXtV2_TinyImageNetconvnext-mainE2AM_ConvNeXtV2_CIFAR100Results_exp_audio_spectral_convnext1d_no_tie_breakersSelf-distill-logits-convnext-tiny
ConvNeXt-Tiny experiments — results and prediction archive
Access is configured. The full artifact migration is not complete. This initial publication contains the verified retained-checkpoint/prediction audit and the immutable report-source index, not the bulk checkpoints or full-logit arrays.
Published evidence
Contents
Verified before/after audit
Exact pretrained and epoch-5 random-KL checkpoint identities; 50,000 aligned prediction rows; 78.824% → 79.868%… See the full description on the dataset page: https://huggingface.co/datasets/dlsmarta/Self-distill-logits-convnext-tiny.