| { |
| "architectures": [ |
| "ERADIOModel" |
| ], |
| "args": { |
| "aa": null, |
| "amp": true, |
| "amp_dtype": "bfloat16", |
| "amp_impl": "native", |
| "aug_repeats": 0, |
| "aug_splits": 0, |
| "batch_size": 32, |
| "bn_eps": null, |
| "bn_momentum": null, |
| "cache": "/lustre/fs3/portfolios/llmservice/users/gheinrich/cache/", |
| "cache_dir": null, |
| "channels_last": false, |
| "checkpoint_hist": 3, |
| "class_map": "", |
| "clip_grad": null, |
| "clip_mode": "norm", |
| "coco_annotations_file": "/datasets/coco2017-adlsa/annotations/captions_val2017.json", |
| "coco_image_dir": "/datasets/coco2017-adlsa/val2017", |
| "color_jitter": 0.4, |
| "cooldown_epochs": 0, |
| "cpe_max_size": null, |
| "crd_loss": false, |
| "crd_loss_weight": 0.8, |
| "crop_pct": null, |
| "cutmix": 0.0, |
| "cutmix_minmax": null, |
| "data_dir": "/lustre/fsw/portfolios/llmservice/projects/llmservice_nlp_fm/datasets/classification/imagenet-21k/webdataset", |
| "dataset": "nvgpt4", |
| "dataset_download": false, |
| "debug_full_knn": false, |
| "decay_epochs": 90, |
| "decay_milestones": [ |
| 90, |
| 180, |
| 270 |
| ], |
| "decay_rate": 0.1, |
| "device": "cuda:0", |
| "dist_bn": "reduce", |
| "distributed": true, |
| "drop": 0.0, |
| "drop_block": null, |
| "drop_connect": null, |
| "drop_path": null, |
| "epoch_repeats": 0.0, |
| "epochs": 300, |
| "eval_metric": "knn_top1", |
| "eval_teacher": false, |
| "eval_teacher_only": false, |
| "eval_throughput": false, |
| "experiment": "checkpoints", |
| "fast_norm": false, |
| "feature_summarizer": "cls_token", |
| "feature_upscale_factor": null, |
| "fuser": "", |
| "gp": null, |
| "grad_accum_steps": 1, |
| "grad_checkpointing": false, |
| "head_init_bias": null, |
| "head_init_scale": null, |
| "hflip": 0.5, |
| "img_size": null, |
| "in_chans": 3, |
| "initial_checkpoint": "", |
| "input_size": null, |
| "interpolation": "", |
| "layer_decay": null, |
| "local_rank": 0, |
| "log_interval": 50, |
| "log_mlflow": false, |
| "log_wandb": true, |
| "loss": "cosine", |
| "lr": 0.001, |
| "lr_base": 0.1, |
| "lr_base_scale": "", |
| "lr_base_size": 256, |
| "lr_cycle_decay": 0.5, |
| "lr_cycle_limit": 1, |
| "lr_cycle_mul": 1.0, |
| "lr_k_decay": 1.0, |
| "lr_noise": null, |
| "lr_noise_pct": 0.67, |
| "lr_noise_std": 1.0, |
| "mean": null, |
| "min_lr": 0, |
| "mixup": 0.0, |
| "mixup_mode": "batch", |
| "mixup_off_epoch": 0, |
| "mixup_prob": 1.0, |
| "mixup_switch_prob": 0.5, |
| "mlp_hidden_size": 1024, |
| "model": "fastervit2_large_fullres", |
| "model_ema": false, |
| "model_ema_decay": 0.9998, |
| "model_ema_force_cpu": false, |
| "model_kwargs": { |
| "return_full_features": true |
| }, |
| "momentum": 0.9, |
| "no_aug": false, |
| "no_ddp_bb": false, |
| "no_prefetcher": false, |
| "no_resume_opt": false, |
| "num_classes": 0, |
| "opt": "fusedlamb", |
| "opt_betas": null, |
| "opt_eps": null, |
| "opt_kwargs": {}, |
| "output": "/lustre/fs3/portfolios/llmservice/users/gheinrich/results/evfm/19-11-23-fastervit2-l-fullres", |
| "patience_epochs": 10, |
| "pin_mem": false, |
| "prefetcher": false, |
| "pretrained": false, |
| "rank": 0, |
| "ratio": [ |
| 0.75, |
| 1.3333333333333333 |
| ], |
| "recount": 1, |
| "recovery_interval": 0, |
| "remode": "pixel", |
| "reprob": 0.0, |
| "resplit": false, |
| "resume": "/lustre/fs3/portfolios/llmservice/users/gheinrich/results/evfm/19-11-23-fastervit2-l-fullres/checkpoints/last.pth.tar", |
| "return_full_features": true, |
| "save_images": false, |
| "scale": [ |
| 0.5, |
| 1.0 |
| ], |
| "sched": "cosine", |
| "sched_on_updates": true, |
| "seed": 42, |
| "smoothing": 0.1, |
| "split_bn": false, |
| "start_epoch": null, |
| "std": null, |
| "steps_per_epoch": 2000, |
| "sync_bn": false, |
| "synchronize_step": false, |
| "teachers": [ |
| { |
| "batch_size": 32, |
| "config": "open_clip_vit-h-14_res224.yaml", |
| "fd_loss_weight": 1.0, |
| "feature_distillation": true, |
| "sample_rate": 8, |
| "summary_loss_weight": 1.0 |
| }, |
| { |
| "batch_size": 32, |
| "config": "dinov2_vit-g-14_res224.yaml", |
| "fd_loss_weight": 4.0, |
| "feature_distillation": true, |
| "sample_rate": 8, |
| "summary_loss_weight": 1.0 |
| } |
| ], |
| "torchcompile": null, |
| "torchscript": false, |
| "train_interpolation": "random", |
| "train_split": "train", |
| "tta": 0, |
| "use_coco": false, |
| "use_multi_epochs_loader": false, |
| "val_data_dir": "/lustre/fsw/portfolios/llmservice/projects/llmservice_nlp_fm/datasets/classification/imagenet-1k/webdataset", |
| "val_img_size": 224, |
| "val_split": "val", |
| "validation_batch_size": 32, |
| "vflip": 0.0, |
| "wandb_entity": "", |
| "wandb_group": "backbones", |
| "wandb_job_type": "", |
| "wandb_name": "", |
| "wandb_project": "", |
| "warmup_epochs": 2.5, |
| "warmup_lr": 1e-05, |
| "warmup_prefix": false, |
| "weight_decay": 2e-05, |
| "worker_seeding": "all", |
| "workers": 4, |
| "world_size": 32 |
| }, |
| "auto_map": { |
| "AutoConfig": "hf_model.ERADIOConfig", |
| "AutoModel": "hf_model.ERADIOModel" |
| }, |
| "return_spatial_features": true, |
| "return_summary": true, |
| "torch_dtype": "float32", |
| "transformers_version": "4.29.0", |
| "version": "v1" |
| } |
|
|
