Instructions to use KRMayD/COD10K_CAM_5Way_Segmentation_Checkpoints with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- OpenCLIP
How to use KRMayD/COD10K_CAM_5Way_Segmentation_Checkpoints with OpenCLIP:
import open_clip model, preprocess_train, preprocess_val = open_clip.create_model_and_transforms('hf-hub:KRMayD/COD10K_CAM_5Way_Segmentation_Checkpoints') tokenizer = open_clip.get_tokenizer('hf-hub:KRMayD/COD10K_CAM_5Way_Segmentation_Checkpoints') - Notebooks
- Google Colab
- Kaggle
File size: 1,623 Bytes
0d9a4c0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 | {
"train_csv": "/home/msko021220/dataset/COD10K_GMPO_Used_Data/captions/cod10k_train_cam_dpo_1caption_sd_bgpos_vs_bgexplicitneg_exclude_mmcamobj_easyhard_2886.csv",
"train_csv_sha256": "7f2dde8a9062a9a8c7080c8c6d8c6e3c4981ec7056e04615761e10d669a95630",
"train_pairs": 2886,
"excluded_mmcamobj_easyhard_images": 154,
"positive_input": [
"filename",
"Caption"
],
"gmpo_only_inputs": [
"filename_neg",
"Caption_neg"
],
"initial_and_reference_checkpoint": "/home/msko021220/dataset/clip_reference_checkpoints/openai_clip_vit_b_32_quickgelu_openclip_state_dict.pt",
"initial_checkpoint_sha256": "851b42d7c83e629a8c1d4b2e00cf94e7d9e9b414feeee34592228f495b08afef",
"encoder": "ViT-B-32-quickgelu",
"shared_train_budget": {
"epochs": 3,
"seed": 0
},
"gmpo": {
"optimizer": {
"lr": 1e-06,
"weight_decay": 0.01,
"warmup_steps": 10,
"beta_dpo": 10.0
},
"ddp_gpus": [
0,
1
],
"batch_per_gpu": 128,
"global_batch": 256,
"variants": [
"global_ind",
"sample_ind"
]
},
"clip_and_cliprefine_native_optimizer": {
"batch_size": 32,
"lr": 2e-05,
"weight_decay": 0.05,
"warmup_epochs": 1,
"gpus": {
"clip": 0,
"cliprefine": 1
}
},
"evaluation": {
"test_split": "COD10K CAM Test",
"samples": 2026,
"prompt": "There is a/an {animal} visually blended into its surroundings.",
"vbeta": 1.0,
"vvar": 0.3,
"vlayer": 8,
"seed": 42,
"nsd_tolerance_px": 2.0,
"sam_or_postprocessing": "unchanged cod10k_saliency_hparam_search pipeline"
}
}
|