Commit
•
ab6c27f
1
Parent(s):
53b507b
Upload config
Browse files- config.json +65 -0
config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_class_name": "HookedViT",
|
| 3 |
+
"model_name": "open-clip:laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K",
|
| 4 |
+
"vit_model_cfg": null,
|
| 5 |
+
"model_path": null,
|
| 6 |
+
"hook_point_layer": 10,
|
| 7 |
+
"layer_subtype": "hook_resid_post",
|
| 8 |
+
"hook_point_head_index": null,
|
| 9 |
+
"context_size": 1,
|
| 10 |
+
"use_cached_activations": false,
|
| 11 |
+
"use_patches_only": false,
|
| 12 |
+
"cached_activations_path": "activations/_network_scratch_s_sonia.joseph_datasets_kaggle_datasets/open-clip:laion_CLIP-ViT-B-32-DataComp.XL-s13B-b90K/blocks.11.hook_mlp_out",
|
| 13 |
+
"d_in": 768,
|
| 14 |
+
"activation_fn_str": "topk",
|
| 15 |
+
"activation_fn_kwargs": {
|
| 16 |
+
"k": 64
|
| 17 |
+
},
|
| 18 |
+
"cls_token_only": true,
|
| 19 |
+
"max_grad_norm": 1.0,
|
| 20 |
+
"initialization_method": "encoder_transpose_decoder",
|
| 21 |
+
"normalize_activations": null,
|
| 22 |
+
"n_batches_in_buffer": 20,
|
| 23 |
+
"store_batch_size": 32,
|
| 24 |
+
"num_workers": 16,
|
| 25 |
+
"num_epochs": 500,
|
| 26 |
+
"total_training_images": 650000000,
|
| 27 |
+
"total_training_tokens": 325000000000,
|
| 28 |
+
"image_size": 224,
|
| 29 |
+
"device": {
|
| 30 |
+
"__type__": "torch.device",
|
| 31 |
+
"value": "cuda"
|
| 32 |
+
},
|
| 33 |
+
"seed": 42,
|
| 34 |
+
"dtype": {
|
| 35 |
+
"__type__": "torch.dtype",
|
| 36 |
+
"value": "torch.float32"
|
| 37 |
+
},
|
| 38 |
+
"architecture": "standard",
|
| 39 |
+
"verbose": false,
|
| 40 |
+
"b_dec_init_method": "geometric_median",
|
| 41 |
+
"expansion_factor": 64,
|
| 42 |
+
"from_pretrained_path": null,
|
| 43 |
+
"d_sae": 49152,
|
| 44 |
+
"l1_coefficient": 0.3,
|
| 45 |
+
"lp_norm": 1,
|
| 46 |
+
"lr": 0.0002,
|
| 47 |
+
"lr_scheduler_name": "cosineannealingwarmup",
|
| 48 |
+
"lr_warm_up_steps": 200,
|
| 49 |
+
"train_batch_size": 4096,
|
| 50 |
+
"dataset_name": "imagenet1k",
|
| 51 |
+
"dataset_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets",
|
| 52 |
+
"dataset_train_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets/ILSVRC/Data/CLS-LOC/train",
|
| 53 |
+
"dataset_val_path": "/network/scratch/s/sonia.joseph/datasets/kaggle_datasets/ILSVRC/Data/CLS-LOC/val",
|
| 54 |
+
"use_ghost_grads": false,
|
| 55 |
+
"feature_sampling_window": 1000,
|
| 56 |
+
"dead_feature_window": 5000,
|
| 57 |
+
"dead_feature_threshold": 1e-08,
|
| 58 |
+
"log_to_wandb": true,
|
| 59 |
+
"wandb_project": "clip_b_mlp_out_sae_hyperparam_sweep",
|
| 60 |
+
"wandb_entity": null,
|
| 61 |
+
"wandb_log_frequency": 100,
|
| 62 |
+
"n_validation_runs": 10,
|
| 63 |
+
"n_checkpoints": 10,
|
| 64 |
+
"checkpoint_path": "/network/scratch/s/sonia.joseph/checkpoints/clip-b/8cf13ed0-clip_b_mlp_out_sae_hyperparam_sweep"
|
| 65 |
+
}
|