Upload folder using huggingface_hub
Browse files- config.json +52 -0
- n_images_130011.pt +3 -0
- n_images_130011_log_feature_sparsity.pt +3 -0
- n_images_260002.pt +3 -0
- n_images_260002_log_feature_sparsity.pt +3 -0
config.json
ADDED
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"model_class_name": "HookedViT",
|
3 |
+
"model_name": "vit_base_patch16_224",
|
4 |
+
"vit_model_cfg": null,
|
5 |
+
"model_path": null,
|
6 |
+
"hook_point_layer": 11,
|
7 |
+
"layer_subtype": "hook_resid_post",
|
8 |
+
"hook_point_head_index": null,
|
9 |
+
"context_size": 197,
|
10 |
+
"use_cached_activations": false,
|
11 |
+
"use_patches_only": false,
|
12 |
+
"cached_activations_path": "activations/_datasets_imagenet1k/vit_base_patch16_224/blocks.11.hook_resid_post",
|
13 |
+
"image_size": 224,
|
14 |
+
"architecture": "standard",
|
15 |
+
"b_dec_init_method": "geometric_median",
|
16 |
+
"expansion_factor": 64,
|
17 |
+
"from_pretrained_path": null,
|
18 |
+
"seed": 42,
|
19 |
+
"d_in": 768,
|
20 |
+
"activation_fn_str": "relu",
|
21 |
+
"activation_fn_kwargs": {},
|
22 |
+
"cls_token_only": false,
|
23 |
+
"max_grad_norm": 1.0,
|
24 |
+
"initialization_method": "encoder_transpose_decoder",
|
25 |
+
"normalize_activations": null,
|
26 |
+
"n_batches_in_buffer": 20,
|
27 |
+
"store_batch_size": 256,
|
28 |
+
"num_workers": 32,
|
29 |
+
"num_epochs": 0.2,
|
30 |
+
"verbose": false,
|
31 |
+
"l1_coefficient": 8e-05,
|
32 |
+
"lp_norm": 1,
|
33 |
+
"lr": 0.0002,
|
34 |
+
"lr_scheduler_name": "cosineannealingwarmup",
|
35 |
+
"lr_warm_up_steps": 200,
|
36 |
+
"train_batch_size": 4096,
|
37 |
+
"dataset_name": "imagenet1k",
|
38 |
+
"dataset_path": "/datasets/imagenet1k",
|
39 |
+
"dataset_train_path": "/datasets/imagenet1k/ILSVRC/Data/CLS-LOC/train",
|
40 |
+
"dataset_val_path": "/datasets/imagenet1k/ILSVRC/Data/CLS-LOC/val",
|
41 |
+
"use_ghost_grads": true,
|
42 |
+
"feature_sampling_window": 1000,
|
43 |
+
"dead_feature_window": 5000,
|
44 |
+
"dead_feature_threshold": 1e-08,
|
45 |
+
"log_to_wandb": true,
|
46 |
+
"wandb_project": "vit-sae-eval",
|
47 |
+
"wandb_entity": "hhixai",
|
48 |
+
"wandb_log_frequency": 100,
|
49 |
+
"n_validation_runs": 4,
|
50 |
+
"n_checkpoints": 2,
|
51 |
+
"checkpoint_path": "models/sae/vanilla/1fb65c6f-vit-sae-eval"
|
52 |
+
}
|
n_images_130011.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:947d0ddcfddd1fe8a5cf9da8db616778d0e4d5da84e010c05aeedd9b3a6f4e1f
|
3 |
+
size 302193744
|
n_images_130011_log_feature_sparsity.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c7e899f09b34aab958bf32de081083e78d8311f5addf76365f75a4d49407ec13
|
3 |
+
size 197997
|
n_images_260002.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:30d2735d9b9f1b80a1cfee5995da7f494b447834953761b235d24ebf93d82e25
|
3 |
+
size 302193744
|
n_images_260002_log_feature_sparsity.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6ce0d248bc8e55bfa625143af1a1a8dafed3ff9a9f17eb848fad5f37fa20f63a
|
3 |
+
size 197997
|