{ "model_class_name": "HookedViT", "model_name": "open-clip:laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K", "vit_model_cfg": null, "model_path": null, "hook_point_layer": 9, "layer_subtype": "ln2.hook_normalized", "hook_point_head_index": null, "context_size": 50, "use_cached_activations": false, "use_patches_only": false, "cached_activations_path": "activations/_network_scratch_s_sonia.joseph_datasets_kaggle_datasets/open-clip:laion_CLIP-ViT-B-32-DataComp.XL-s13B-b90K/blocks.9.ln2.hook_normalized", "image_size": 224, "architecture": "standard", "b_dec_init_method": "geometric_median", "expansion_factor": 64, "from_pretrained_path": null, "is_transcoder": true, "transcoder_with_skip_connection": true, "out_hook_point_layer": 9, "layer_out_subtype": "hook_mlp_out", "d_out": 768, "_device": "cuda", "seed": 42, "_dtype": "float32", "d_in": 768, "activation_fn_str": "topk", "activation_fn_kwargs": { "k": 1024 }, "cls_token_only": false, "max_grad_norm": 1.0, "initialization_method": "independent", "normalize_activations": "layer_norm", "n_batches_in_buffer": 20, "store_batch_size": 32, "num_workers": 16, "num_epochs": 1, "verbose": false, "l1_coefficient": 0.0002, "lp_norm": 1, "lr": 0.0006600033204809742, "lr_scheduler_name": "cosineannealingwarmup", "lr_warm_up_steps": 500, "train_batch_size": 4096, "dataset_name": "imagenet1k", "dataset_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets", "dataset_train_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets/ILSVRC/Data/CLS-LOC/train", "dataset_val_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets/ILSVRC/Data/CLS-LOC/val", "use_ghost_grads": false, "feature_sampling_window": 1000, "dead_feature_window": 5000, "dead_feature_threshold": 1e-08, "log_to_wandb": true, "wandb_project": "openclip-transcoders", "wandb_entity": null, "wandb_log_frequency": 10, "n_validation_runs": 0, "n_checkpoints": 10, "checkpoint_path": "/network/scratch/p/praneet.suresh/openclip-transcoder-checkpoints/275a01ed-openclip-transcoders" }