Upload sigma_lambda_proto_v3/history.json with huggingface_hub
Browse files
sigma_lambda_proto_v3/history.json
CHANGED
|
@@ -1,65 +1,81 @@
|
|
| 1 |
{
|
| 2 |
"epoch": [
|
| 3 |
1,
|
| 4 |
-
2
|
|
|
|
| 5 |
],
|
| 6 |
"gen_loss": [
|
| 7 |
18.370433433850607,
|
| 8 |
-
12.677154767152034
|
|
|
|
| 9 |
],
|
| 10 |
"ret_loss": [
|
|
|
|
| 11 |
0.0,
|
| 12 |
0.0
|
| 13 |
],
|
| 14 |
"ent_loss": [
|
| 15 |
0.6078802872065342,
|
| 16 |
-
0.6268522234035261
|
|
|
|
| 17 |
],
|
| 18 |
"ppl": [
|
| 19 |
40729.60576844635,
|
| 20 |
-
2535.223731428122
|
|
|
|
| 21 |
],
|
| 22 |
"tf_acc": [
|
| 23 |
9.249682110189831e-06,
|
| 24 |
-
0.21421828488041522
|
|
|
|
| 25 |
],
|
| 26 |
"fr_acc": [
|
| 27 |
0.0,
|
| 28 |
-
0.0
|
|
|
|
| 29 |
],
|
| 30 |
"def_rate": [
|
|
|
|
| 31 |
0.0,
|
| 32 |
0.0
|
| 33 |
],
|
| 34 |
"mrr": [
|
| 35 |
0.004940100014209747,
|
| 36 |
-
0.005048312712460756
|
|
|
|
| 37 |
],
|
| 38 |
"r1": [
|
|
|
|
| 39 |
0.0006249999860301614,
|
| 40 |
0.0006249999860301614
|
| 41 |
],
|
| 42 |
"ret_ok_pct": [
|
|
|
|
| 43 |
0.0,
|
| 44 |
0.0
|
| 45 |
],
|
| 46 |
"grad_norm": [
|
| 47 |
2201.0254727094825,
|
| 48 |
-
1418.5299411384265
|
|
|
|
| 49 |
],
|
| 50 |
"step_var": [
|
| 51 |
0.22542811930179596,
|
| 52 |
-
0.22660225629806519
|
|
|
|
| 53 |
],
|
| 54 |
"pos_slope": [
|
| 55 |
0.49990350008010864,
|
| 56 |
-
0.4996356666088104
|
|
|
|
| 57 |
],
|
| 58 |
"beta_mean": [
|
| 59 |
0.09998118132352829,
|
| 60 |
-
0.09992741793394089
|
|
|
|
| 61 |
],
|
| 62 |
"w_ret": [
|
|
|
|
| 63 |
0.0,
|
| 64 |
0.0
|
| 65 |
]
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": [
|
| 3 |
1,
|
| 4 |
+
2,
|
| 5 |
+
3
|
| 6 |
],
|
| 7 |
"gen_loss": [
|
| 8 |
18.370433433850607,
|
| 9 |
+
12.677154767152034,
|
| 10 |
+
12.081629503279022
|
| 11 |
],
|
| 12 |
"ret_loss": [
|
| 13 |
+
0.0,
|
| 14 |
0.0,
|
| 15 |
0.0
|
| 16 |
],
|
| 17 |
"ent_loss": [
|
| 18 |
0.6078802872065342,
|
| 19 |
+
0.6268522234035261,
|
| 20 |
+
0.6330166747714534
|
| 21 |
],
|
| 22 |
"ppl": [
|
| 23 |
40729.60576844635,
|
| 24 |
+
2535.223731428122,
|
| 25 |
+
337.67548432577263
|
| 26 |
],
|
| 27 |
"tf_acc": [
|
| 28 |
9.249682110189831e-06,
|
| 29 |
+
0.21421828488041522,
|
| 30 |
+
0.39405658955515016
|
| 31 |
],
|
| 32 |
"fr_acc": [
|
| 33 |
0.0,
|
| 34 |
+
0.0,
|
| 35 |
+
0.11453107816089203
|
| 36 |
],
|
| 37 |
"def_rate": [
|
| 38 |
+
0.0,
|
| 39 |
0.0,
|
| 40 |
0.0
|
| 41 |
],
|
| 42 |
"mrr": [
|
| 43 |
0.004940100014209747,
|
| 44 |
+
0.005048312712460756,
|
| 45 |
+
0.004967095330357552
|
| 46 |
],
|
| 47 |
"r1": [
|
| 48 |
+
0.0006249999860301614,
|
| 49 |
0.0006249999860301614,
|
| 50 |
0.0006249999860301614
|
| 51 |
],
|
| 52 |
"ret_ok_pct": [
|
| 53 |
+
0.0,
|
| 54 |
0.0,
|
| 55 |
0.0
|
| 56 |
],
|
| 57 |
"grad_norm": [
|
| 58 |
2201.0254727094825,
|
| 59 |
+
1418.5299411384265,
|
| 60 |
+
1329.5837538555174
|
| 61 |
],
|
| 62 |
"step_var": [
|
| 63 |
0.22542811930179596,
|
| 64 |
+
0.22660225629806519,
|
| 65 |
+
0.2280091941356659
|
| 66 |
],
|
| 67 |
"pos_slope": [
|
| 68 |
0.49990350008010864,
|
| 69 |
+
0.4996356666088104,
|
| 70 |
+
0.49926072359085083
|
| 71 |
],
|
| 72 |
"beta_mean": [
|
| 73 |
0.09998118132352829,
|
| 74 |
+
0.09992741793394089,
|
| 75 |
+
0.09985271841287613
|
| 76 |
],
|
| 77 |
"w_ret": [
|
| 78 |
+
0.0,
|
| 79 |
0.0,
|
| 80 |
0.0
|
| 81 |
]
|