GOVINDFROM commited on
Commit
4ad927f
·
verified ·
1 Parent(s): c496547

Automated push from Zaratan

Browse files
results/g13-d20-full-s0.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 0,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-full-s0"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 1.0,
25
+ "2": 1.0,
26
+ "3": 1.0,
27
+ "4": 1.0,
28
+ "6": 1.0,
29
+ "8": 1.0,
30
+ "12": 1.0,
31
+ "16": 1.0
32
+ },
33
+ "minutes": 110.98323661088943,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d20-full-s1.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 1,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-full-s1"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.9975,
25
+ "2": 0.98,
26
+ "3": 0.965,
27
+ "4": 0.8675,
28
+ "6": 0.6875,
29
+ "8": 0.625,
30
+ "12": 0.615,
31
+ "16": 0.5725
32
+ },
33
+ "minutes": 110.76181087891261,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d20-full-s2.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 2,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-full-s2"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 1.0,
25
+ "2": 1.0,
26
+ "3": 1.0,
27
+ "4": 1.0,
28
+ "6": 1.0,
29
+ "8": 1.0,
30
+ "12": 1.0,
31
+ "16": 1.0
32
+ },
33
+ "minutes": 109.23384458621344,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d20-mdlm-s0.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 0,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-mdlm-s0"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 58.36175608634949,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }
results/g13-d20-mdlm-s1.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 1,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-mdlm-s1"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 58.23543329238892,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }
results/g13-d20-mdlm-s2.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 20,
8
+ "seq_len": 64,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 2,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d20-mdlm-s2"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 58.67609803676605,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1228800000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }
results/g13-d24-full-s0.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 0,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-full-s0"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 1.0,
25
+ "2": 0.74,
26
+ "3": 0.8500000000000001,
27
+ "4": 0.8825000000000001,
28
+ "6": 0.9324999999999999,
29
+ "8": 0.9824999999999999,
30
+ "12": 1.0,
31
+ "16": 1.0
32
+ },
33
+ "minutes": 132.4227197488149,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d24-full-s1.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 1,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-full-s1"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0075,
25
+ "2": 0.0,
26
+ "3": 0.0025,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0025
32
+ },
33
+ "minutes": 131.39800818363827,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d24-full-s2.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "full",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 2,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-full-s2"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 132.64511208931606,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 2
38
+ }
39
+ }
results/g13-d24-mdlm-s0.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 0,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-mdlm-s0"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 68.94488160212835,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }
results/g13-d24-mdlm-s1.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 1,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-mdlm-s1"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 68.91008859475454,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }
results/g13-d24-mdlm-s2.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "args": {
3
+ "mode": "mdlm",
4
+ "K_max": 16,
5
+ "K": 4,
6
+ "lam": 1.0,
7
+ "digits": 24,
8
+ "seq_len": 76,
9
+ "n_train": 1500000,
10
+ "n_eval": 400,
11
+ "eval_bs": 100,
12
+ "d": 512,
13
+ "layers": 12,
14
+ "heads": 8,
15
+ "pe": "rope",
16
+ "bs": 128,
17
+ "steps": 150000,
18
+ "lr": 0.0001,
19
+ "warmup": 6000,
20
+ "seed": 2,
21
+ "out": "/scratch/zt1/project/msml612/user/govind02/pcmd/runs/g13-d24-mdlm-s2"
22
+ },
23
+ "accuracy_by_passes": {
24
+ "1": 0.0,
25
+ "2": 0.0,
26
+ "3": 0.0,
27
+ "4": 0.0,
28
+ "6": 0.0,
29
+ "8": 0.0,
30
+ "12": 0.0,
31
+ "16": 0.0
32
+ },
33
+ "minutes": 69.18164315621058,
34
+ "compute": {
35
+ "params": 37819392,
36
+ "train_tokens": 1459200000,
37
+ "fwd_per_step": 1
38
+ }
39
+ }