Upload folder using huggingface_hub
Browse files- config.json +37 -0
- sae.pt +3 -0
config.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data": {
|
| 3 |
+
"shard_root": "/local/scratch/stevens.994/cache/saev/724b1b7be995ef7212d64640fec2885737a706a33b8e5a18f7f323223bd43af1/",
|
| 4 |
+
"patches": "patches",
|
| 5 |
+
"layer": -2,
|
| 6 |
+
"clamp": 100000.0,
|
| 7 |
+
"n_random_samples": 524288,
|
| 8 |
+
"scale_mean": true,
|
| 9 |
+
"scale_norm": true
|
| 10 |
+
},
|
| 11 |
+
"n_workers": 32,
|
| 12 |
+
"n_patches": 100000000,
|
| 13 |
+
"sae": {
|
| 14 |
+
"d_vit": 768,
|
| 15 |
+
"exp_factor": 32,
|
| 16 |
+
"sparsity_coeff": 0.0004,
|
| 17 |
+
"n_reinit_samples": 524288,
|
| 18 |
+
"ghost_grads": false,
|
| 19 |
+
"remove_parallel_grads": true,
|
| 20 |
+
"normalize_w_dec": true,
|
| 21 |
+
"seed": 147
|
| 22 |
+
},
|
| 23 |
+
"n_sparsity_warmup": 500,
|
| 24 |
+
"lr": 0.003,
|
| 25 |
+
"n_lr_warmup": 500,
|
| 26 |
+
"sae_batch_size": 16384,
|
| 27 |
+
"track": true,
|
| 28 |
+
"wandb_project": "saev",
|
| 29 |
+
"tag": "baseline-v4.8",
|
| 30 |
+
"log_every": 25,
|
| 31 |
+
"ckpt_path": "./checkpoints",
|
| 32 |
+
"device": "cuda",
|
| 33 |
+
"seed": 55,
|
| 34 |
+
"slurm": false,
|
| 35 |
+
"slurm_acct": "PAS2136",
|
| 36 |
+
"log_to": "./logs"
|
| 37 |
+
}
|
sae.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be96bce0cfb11a101cf453feaf38eaa42e67b469291a6c4563699237e9f59cb3
|
| 3 |
+
size 151098370
|