Hugging Face's logo

Datasets:
Fengzhuo
/
optimizer_results

Dataset card Files
xet
Community
1
zhangfz commited on
Commit
c33c288
·
1 Parent(s): 34325e9

update

Browse files
Files changed (9) hide show
  1. logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_42/config.json +27 -0
  2. logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_43/config.json +27 -0
  3. logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_44/config.json +27 -0
  4. logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_45/config.json +27 -0
  5. logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_46/config.json +27 -0
  6. logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_42/config.json +27 -0
  7. logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_42/training_log_e05b288f-0c8e-470d-984a-9d9788ef88aa.txt +0 -0
  8. logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_44/config.json +27 -0
  9. logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_44/training_log_dfc238c9-d5b8-4c4f-988c-83bcf20ff2ca.txt +0 -0
logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_42/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 42,
4
+ "optimizer_mode": 13,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/ori"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 24,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 0,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "b07e4ffe-2445-41bb-b149-e2dfd04fbc54",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_43/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 43,
4
+ "optimizer_mode": 13,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/ori"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 24,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 0,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "0525acf0-9ae7-4229-9b5a-209d05e96c88",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_44/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 44,
4
+ "optimizer_mode": 13,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/ori"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 24,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 0,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "a5bd2ee4-96fc-45b5-bde4-be1c6ffe124d",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_45/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 45,
4
+ "optimizer_mode": 13,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/ori"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 24,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 0,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "b5a804e9-97e5-48f2-86bd-8550f6f8ccb8",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/ori/mode_13_param_qkvo_muon_lr_0.0005_adam_lr_0.0005_seed_46/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 46,
4
+ "optimizer_mode": 13,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/ori"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 24,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 0,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "a1125ad1-7bd6-4a3d-8808-ed482c385905",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_42/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 42,
4
+ "optimizer_mode": 5,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/svd"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 12,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 2400,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "e05b288f-0c8e-470d-984a-9d9788ef88aa",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_42/training_log_e05b288f-0c8e-470d-984a-9d9788ef88aa.txt ADDED
The diff for this file is too large to render. See raw diff
 
logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_44/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cli_args": {
3
+ "seed": 44,
4
+ "optimizer_mode": 5,
5
+ "model_parameterization": "qkvo",
6
+ "adam_lr": 0.0005,
7
+ "muon_lr": 0.0005,
8
+ "base_dir": "logs_new_MUON_large_reshape_svd/svd"
9
+ },
10
+ "hyperparameters": {
11
+ "input_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_train_*.bin",
12
+ "input_val_bin": "/home/aiops/zhangfz/MUON_theory/modded-nanogpt/data/fineweb10B/fineweb_val_*.bin",
13
+ "batch_size": 960,
14
+ "device_batch_size": 12,
15
+ "sequence_length": 1024,
16
+ "num_iterations": 6000,
17
+ "learning_rate": 0.0018,
18
+ "warmup_iters": 0,
19
+ "warmdown_iters": 2400,
20
+ "weight_decay": 0,
21
+ "val_loss_every": 125,
22
+ "val_tokens": 10420224,
23
+ "save_every": 0
24
+ },
25
+ "run_uuid_for_log": "dfc238c9-d5b8-4c4f-988c-83bcf20ff2ca",
26
+ "script_code_logged_at_start": true
27
+ }
logs_new_MUON_large_reshape_svd/svd/mode_5_param_qkvo_seed_44/training_log_dfc238c9-d5b8-4c4f-988c-83bcf20ff2ca.txt ADDED
The diff for this file is too large to render. See raw diff