move
Browse files- {sparse → dense}/p32k00/config.json +0 -0
- sparse/p32k00/ckpt.pt +0 -3
- sparse/p32k00/hparams.txt +0 -43
{sparse → dense}/p32k00/config.json
RENAMED
|
File without changes
|
sparse/p32k00/ckpt.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:8ba8202e7bc7cb9079dad75298d58060a2ae6d10266314c5e2fb6762a3802971
|
| 3 |
-
size 1112227537
|
|
|
|
|
|
|
|
|
|
|
|
sparse/p32k00/hparams.txt
DELETED
|
@@ -1,43 +0,0 @@
|
|
| 1 |
-
AstroPT-0092.6M
|
| 2 |
-
time: 1740647970
|
| 3 |
-
log_via_wandb: True
|
| 4 |
-
log_emissions: False
|
| 5 |
-
out_dir: logs/astropt070M_nosparse_32
|
| 6 |
-
eval_interval: 100
|
| 7 |
-
log_interval: 10
|
| 8 |
-
checkpoint_interval: 500
|
| 9 |
-
eval_iters: 100
|
| 10 |
-
eval_only: False
|
| 11 |
-
always_save_checkpoint: False
|
| 12 |
-
init_from: scratch
|
| 13 |
-
hf_url: smith42/galaxies
|
| 14 |
-
stream_hf_dataset: False
|
| 15 |
-
gradient_accumulation_steps: 40
|
| 16 |
-
batch_size: 64
|
| 17 |
-
spiral: True
|
| 18 |
-
block_size: 64
|
| 19 |
-
image_size: 256
|
| 20 |
-
num_workers: 64
|
| 21 |
-
n_layer: 12
|
| 22 |
-
n_head: 12
|
| 23 |
-
n_embd: 768
|
| 24 |
-
n_chan: 3
|
| 25 |
-
dropout: 0.0
|
| 26 |
-
patch_size: 32
|
| 27 |
-
bias: False
|
| 28 |
-
attn_type: causal
|
| 29 |
-
k_ratio: 0.0
|
| 30 |
-
learning_rate: 0.0006
|
| 31 |
-
max_iters: 3500
|
| 32 |
-
weight_decay: 0.1
|
| 33 |
-
beta1: 0.9
|
| 34 |
-
beta2: 0.95
|
| 35 |
-
grad_clip: 1.0
|
| 36 |
-
decay_lr: True
|
| 37 |
-
warmup_iters: 2000
|
| 38 |
-
lr_decay_iters: 3300.0000000000005
|
| 39 |
-
min_lr: 5.9999999999999995e-05
|
| 40 |
-
backend: nccl
|
| 41 |
-
device: cuda
|
| 42 |
-
dtype: bfloat16
|
| 43 |
-
compile: True
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|