Training in progress, step 800, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +703 -3
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 792912
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7158fc026fcb3c75de92fe98a9e7575457f3d0ddfdd21de716103cd610f69de9
|
| 3 |
size 792912
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 807226
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e19210c69c04c9e2774683c62f913e9f0f21be0a7cfcac2e1acf7c710d1c386
|
| 3 |
size 807226
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22ad685c1237efec2f2e8de413bc38dd1475089fe9b28d8d43fe1c45017827ac
|
| 3 |
size 14960
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:53dc596014d0f9b21145357c41c37d57500c8efda8a0f2014bb2d988a043e59f
|
| 3 |
size 15024
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:223b0451f4d9b8dd81e35f5a28bc073e05755799383bf2e8538e323aed665d3b
|
| 3 |
size 15024
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18d17262a3d5928a646b0fefef89331d9cdec36c92fdf757662a6e2bf2165b6d
|
| 3 |
size 15024
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f17ccafd97486ba1c378a3ba81af6282c7ec0337562d5f30dc98873ace309528
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 203,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -4939,6 +4939,706 @@
|
|
| 4939 |
"learning_rate": 5.158265042373672e-06,
|
| 4940 |
"loss": 4.0688,
|
| 4941 |
"step": 700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4942 |
}
|
| 4943 |
],
|
| 4944 |
"logging_steps": 1,
|
|
@@ -4958,7 +5658,7 @@
|
|
| 4958 |
"attributes": {}
|
| 4959 |
}
|
| 4960 |
},
|
| 4961 |
-
"total_flos":
|
| 4962 |
"train_batch_size": 2,
|
| 4963 |
"trial_name": null,
|
| 4964 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.987045033929673,
|
| 5 |
"eval_steps": 203,
|
| 6 |
+
"global_step": 800,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4939 |
"learning_rate": 5.158265042373672e-06,
|
| 4940 |
"loss": 4.0688,
|
| 4941 |
"step": 700
|
| 4942 |
+
},
|
| 4943 |
+
{
|
| 4944 |
+
"epoch": 0.864898210980876,
|
| 4945 |
+
"grad_norm": 1199.1778564453125,
|
| 4946 |
+
"learning_rate": 5.067337733271083e-06,
|
| 4947 |
+
"loss": 4.1535,
|
| 4948 |
+
"step": 701
|
| 4949 |
+
},
|
| 4950 |
+
{
|
| 4951 |
+
"epoch": 0.8661320172732881,
|
| 4952 |
+
"grad_norm": 1050.1751708984375,
|
| 4953 |
+
"learning_rate": 4.977176182915727e-06,
|
| 4954 |
+
"loss": 4.1608,
|
| 4955 |
+
"step": 702
|
| 4956 |
+
},
|
| 4957 |
+
{
|
| 4958 |
+
"epoch": 0.8673658235657001,
|
| 4959 |
+
"grad_norm": 601.8255615234375,
|
| 4960 |
+
"learning_rate": 4.887781927873458e-06,
|
| 4961 |
+
"loss": 4.0056,
|
| 4962 |
+
"step": 703
|
| 4963 |
+
},
|
| 4964 |
+
{
|
| 4965 |
+
"epoch": 0.8685996298581122,
|
| 4966 |
+
"grad_norm": 781.65283203125,
|
| 4967 |
+
"learning_rate": 4.799156491633655e-06,
|
| 4968 |
+
"loss": 4.2278,
|
| 4969 |
+
"step": 704
|
| 4970 |
+
},
|
| 4971 |
+
{
|
| 4972 |
+
"epoch": 0.8698334361505243,
|
| 4973 |
+
"grad_norm": 953.7297973632812,
|
| 4974 |
+
"learning_rate": 4.711301384583183e-06,
|
| 4975 |
+
"loss": 3.9453,
|
| 4976 |
+
"step": 705
|
| 4977 |
+
},
|
| 4978 |
+
{
|
| 4979 |
+
"epoch": 0.8710672424429364,
|
| 4980 |
+
"grad_norm": 682.08837890625,
|
| 4981 |
+
"learning_rate": 4.624218103980665e-06,
|
| 4982 |
+
"loss": 3.9502,
|
| 4983 |
+
"step": 706
|
| 4984 |
+
},
|
| 4985 |
+
{
|
| 4986 |
+
"epoch": 0.8723010487353485,
|
| 4987 |
+
"grad_norm": 999.5166625976562,
|
| 4988 |
+
"learning_rate": 4.537908133931018e-06,
|
| 4989 |
+
"loss": 4.0734,
|
| 4990 |
+
"step": 707
|
| 4991 |
+
},
|
| 4992 |
+
{
|
| 4993 |
+
"epoch": 0.8735348550277606,
|
| 4994 |
+
"grad_norm": 1854.658935546875,
|
| 4995 |
+
"learning_rate": 4.452372945360072e-06,
|
| 4996 |
+
"loss": 4.0023,
|
| 4997 |
+
"step": 708
|
| 4998 |
+
},
|
| 4999 |
+
{
|
| 5000 |
+
"epoch": 0.8747686613201727,
|
| 5001 |
+
"grad_norm": 974.3369750976562,
|
| 5002 |
+
"learning_rate": 4.367613995989589e-06,
|
| 5003 |
+
"loss": 4.1085,
|
| 5004 |
+
"step": 709
|
| 5005 |
+
},
|
| 5006 |
+
{
|
| 5007 |
+
"epoch": 0.8760024676125848,
|
| 5008 |
+
"grad_norm": 929.5103759765625,
|
| 5009 |
+
"learning_rate": 4.283632730312348e-06,
|
| 5010 |
+
"loss": 4.2707,
|
| 5011 |
+
"step": 710
|
| 5012 |
+
},
|
| 5013 |
+
{
|
| 5014 |
+
"epoch": 0.8772362739049969,
|
| 5015 |
+
"grad_norm": 820.7687377929688,
|
| 5016 |
+
"learning_rate": 4.200430579567571e-06,
|
| 5017 |
+
"loss": 4.0151,
|
| 5018 |
+
"step": 711
|
| 5019 |
+
},
|
| 5020 |
+
{
|
| 5021 |
+
"epoch": 0.878470080197409,
|
| 5022 |
+
"grad_norm": 851.3623657226562,
|
| 5023 |
+
"learning_rate": 4.118008961716552e-06,
|
| 5024 |
+
"loss": 4.1227,
|
| 5025 |
+
"step": 712
|
| 5026 |
+
},
|
| 5027 |
+
{
|
| 5028 |
+
"epoch": 0.8797038864898211,
|
| 5029 |
+
"grad_norm": 1031.295166015625,
|
| 5030 |
+
"learning_rate": 4.0363692814184e-06,
|
| 5031 |
+
"loss": 4.0757,
|
| 5032 |
+
"step": 713
|
| 5033 |
+
},
|
| 5034 |
+
{
|
| 5035 |
+
"epoch": 0.8809376927822332,
|
| 5036 |
+
"grad_norm": 1569.51318359375,
|
| 5037 |
+
"learning_rate": 3.955512930006222e-06,
|
| 5038 |
+
"loss": 4.3649,
|
| 5039 |
+
"step": 714
|
| 5040 |
+
},
|
| 5041 |
+
{
|
| 5042 |
+
"epoch": 0.8821714990746453,
|
| 5043 |
+
"grad_norm": 1083.7933349609375,
|
| 5044 |
+
"learning_rate": 3.87544128546331e-06,
|
| 5045 |
+
"loss": 4.0284,
|
| 5046 |
+
"step": 715
|
| 5047 |
+
},
|
| 5048 |
+
{
|
| 5049 |
+
"epoch": 0.8834053053670574,
|
| 5050 |
+
"grad_norm": 761.168212890625,
|
| 5051 |
+
"learning_rate": 3.7961557123997018e-06,
|
| 5052 |
+
"loss": 4.2838,
|
| 5053 |
+
"step": 716
|
| 5054 |
+
},
|
| 5055 |
+
{
|
| 5056 |
+
"epoch": 0.8846391116594695,
|
| 5057 |
+
"grad_norm": 821.79443359375,
|
| 5058 |
+
"learning_rate": 3.7176575620289368e-06,
|
| 5059 |
+
"loss": 4.0374,
|
| 5060 |
+
"step": 717
|
| 5061 |
+
},
|
| 5062 |
+
{
|
| 5063 |
+
"epoch": 0.8858729179518815,
|
| 5064 |
+
"grad_norm": 877.775146484375,
|
| 5065 |
+
"learning_rate": 3.6399481721449857e-06,
|
| 5066 |
+
"loss": 4.174,
|
| 5067 |
+
"step": 718
|
| 5068 |
+
},
|
| 5069 |
+
{
|
| 5070 |
+
"epoch": 0.8871067242442936,
|
| 5071 |
+
"grad_norm": 701.419677734375,
|
| 5072 |
+
"learning_rate": 3.563028867099505e-06,
|
| 5073 |
+
"loss": 3.98,
|
| 5074 |
+
"step": 719
|
| 5075 |
+
},
|
| 5076 |
+
{
|
| 5077 |
+
"epoch": 0.8883405305367057,
|
| 5078 |
+
"grad_norm": 802.7510375976562,
|
| 5079 |
+
"learning_rate": 3.4869009577792157e-06,
|
| 5080 |
+
"loss": 4.0774,
|
| 5081 |
+
"step": 720
|
| 5082 |
+
},
|
| 5083 |
+
{
|
| 5084 |
+
"epoch": 0.8895743368291178,
|
| 5085 |
+
"grad_norm": 699.1093139648438,
|
| 5086 |
+
"learning_rate": 3.4115657415835835e-06,
|
| 5087 |
+
"loss": 4.0156,
|
| 5088 |
+
"step": 721
|
| 5089 |
+
},
|
| 5090 |
+
{
|
| 5091 |
+
"epoch": 0.8908081431215299,
|
| 5092 |
+
"grad_norm": 733.71484375,
|
| 5093 |
+
"learning_rate": 3.3370245024027414e-06,
|
| 5094 |
+
"loss": 4.2041,
|
| 5095 |
+
"step": 722
|
| 5096 |
+
},
|
| 5097 |
+
{
|
| 5098 |
+
"epoch": 0.892041949413942,
|
| 5099 |
+
"grad_norm": 527.6852416992188,
|
| 5100 |
+
"learning_rate": 3.2632785105955467e-06,
|
| 5101 |
+
"loss": 3.9713,
|
| 5102 |
+
"step": 723
|
| 5103 |
+
},
|
| 5104 |
+
{
|
| 5105 |
+
"epoch": 0.8932757557063541,
|
| 5106 |
+
"grad_norm": 798.1897583007812,
|
| 5107 |
+
"learning_rate": 3.190329022967975e-06,
|
| 5108 |
+
"loss": 4.0734,
|
| 5109 |
+
"step": 724
|
| 5110 |
+
},
|
| 5111 |
+
{
|
| 5112 |
+
"epoch": 0.8945095619987662,
|
| 5113 |
+
"grad_norm": 723.0835571289062,
|
| 5114 |
+
"learning_rate": 3.1181772827516664e-06,
|
| 5115 |
+
"loss": 4.0981,
|
| 5116 |
+
"step": 725
|
| 5117 |
+
},
|
| 5118 |
+
{
|
| 5119 |
+
"epoch": 0.8957433682911783,
|
| 5120 |
+
"grad_norm": 783.1177978515625,
|
| 5121 |
+
"learning_rate": 3.046824519582808e-06,
|
| 5122 |
+
"loss": 4.0234,
|
| 5123 |
+
"step": 726
|
| 5124 |
+
},
|
| 5125 |
+
{
|
| 5126 |
+
"epoch": 0.8969771745835904,
|
| 5127 |
+
"grad_norm": 666.199462890625,
|
| 5128 |
+
"learning_rate": 2.9762719494810855e-06,
|
| 5129 |
+
"loss": 3.9911,
|
| 5130 |
+
"step": 727
|
| 5131 |
+
},
|
| 5132 |
+
{
|
| 5133 |
+
"epoch": 0.8982109808760025,
|
| 5134 |
+
"grad_norm": 901.4703369140625,
|
| 5135 |
+
"learning_rate": 2.9065207748290134e-06,
|
| 5136 |
+
"loss": 4.3203,
|
| 5137 |
+
"step": 728
|
| 5138 |
+
},
|
| 5139 |
+
{
|
| 5140 |
+
"epoch": 0.8994447871684146,
|
| 5141 |
+
"grad_norm": 615.899169921875,
|
| 5142 |
+
"learning_rate": 2.83757218435145e-06,
|
| 5143 |
+
"loss": 4.1368,
|
| 5144 |
+
"step": 729
|
| 5145 |
+
},
|
| 5146 |
+
{
|
| 5147 |
+
"epoch": 0.9006785934608267,
|
| 5148 |
+
"grad_norm": 953.3482666015625,
|
| 5149 |
+
"learning_rate": 2.7694273530953163e-06,
|
| 5150 |
+
"loss": 4.2854,
|
| 5151 |
+
"step": 730
|
| 5152 |
+
},
|
| 5153 |
+
{
|
| 5154 |
+
"epoch": 0.9019123997532388,
|
| 5155 |
+
"grad_norm": 939.6257934570312,
|
| 5156 |
+
"learning_rate": 2.702087442409551e-06,
|
| 5157 |
+
"loss": 4.0877,
|
| 5158 |
+
"step": 731
|
| 5159 |
+
},
|
| 5160 |
+
{
|
| 5161 |
+
"epoch": 0.9031462060456509,
|
| 5162 |
+
"grad_norm": 666.4422607421875,
|
| 5163 |
+
"learning_rate": 2.6355535999253887e-06,
|
| 5164 |
+
"loss": 4.1066,
|
| 5165 |
+
"step": 732
|
| 5166 |
+
},
|
| 5167 |
+
{
|
| 5168 |
+
"epoch": 0.904380012338063,
|
| 5169 |
+
"grad_norm": 737.6424560546875,
|
| 5170 |
+
"learning_rate": 2.5698269595367254e-06,
|
| 5171 |
+
"loss": 4.1649,
|
| 5172 |
+
"step": 733
|
| 5173 |
+
},
|
| 5174 |
+
{
|
| 5175 |
+
"epoch": 0.905613818630475,
|
| 5176 |
+
"grad_norm": 659.7061157226562,
|
| 5177 |
+
"learning_rate": 2.5049086413808377e-06,
|
| 5178 |
+
"loss": 4.0905,
|
| 5179 |
+
"step": 734
|
| 5180 |
+
},
|
| 5181 |
+
{
|
| 5182 |
+
"epoch": 0.9068476249228871,
|
| 5183 |
+
"grad_norm": 725.164306640625,
|
| 5184 |
+
"learning_rate": 2.4407997518192728e-06,
|
| 5185 |
+
"loss": 3.9756,
|
| 5186 |
+
"step": 735
|
| 5187 |
+
},
|
| 5188 |
+
{
|
| 5189 |
+
"epoch": 0.9080814312152992,
|
| 5190 |
+
"grad_norm": 1040.10205078125,
|
| 5191 |
+
"learning_rate": 2.377501383419006e-06,
|
| 5192 |
+
"loss": 4.0973,
|
| 5193 |
+
"step": 736
|
| 5194 |
+
},
|
| 5195 |
+
{
|
| 5196 |
+
"epoch": 0.9093152375077113,
|
| 5197 |
+
"grad_norm": 933.7422485351562,
|
| 5198 |
+
"learning_rate": 2.3150146149338247e-06,
|
| 5199 |
+
"loss": 3.9678,
|
| 5200 |
+
"step": 737
|
| 5201 |
+
},
|
| 5202 |
+
{
|
| 5203 |
+
"epoch": 0.9105490438001234,
|
| 5204 |
+
"grad_norm": 729.0170288085938,
|
| 5205 |
+
"learning_rate": 2.253340511285923e-06,
|
| 5206 |
+
"loss": 4.0576,
|
| 5207 |
+
"step": 738
|
| 5208 |
+
},
|
| 5209 |
+
{
|
| 5210 |
+
"epoch": 0.9117828500925355,
|
| 5211 |
+
"grad_norm": 700.053955078125,
|
| 5212 |
+
"learning_rate": 2.1924801235477743e-06,
|
| 5213 |
+
"loss": 4.2267,
|
| 5214 |
+
"step": 739
|
| 5215 |
+
},
|
| 5216 |
+
{
|
| 5217 |
+
"epoch": 0.9130166563849476,
|
| 5218 |
+
"grad_norm": 834.5526733398438,
|
| 5219 |
+
"learning_rate": 2.132434488924212e-06,
|
| 5220 |
+
"loss": 3.9349,
|
| 5221 |
+
"step": 740
|
| 5222 |
+
},
|
| 5223 |
+
{
|
| 5224 |
+
"epoch": 0.9142504626773597,
|
| 5225 |
+
"grad_norm": 1009.9600830078125,
|
| 5226 |
+
"learning_rate": 2.073204630734743e-06,
|
| 5227 |
+
"loss": 4.2046,
|
| 5228 |
+
"step": 741
|
| 5229 |
+
},
|
| 5230 |
+
{
|
| 5231 |
+
"epoch": 0.9154842689697718,
|
| 5232 |
+
"grad_norm": 581.45849609375,
|
| 5233 |
+
"learning_rate": 2.0147915583961173e-06,
|
| 5234 |
+
"loss": 4.1017,
|
| 5235 |
+
"step": 742
|
| 5236 |
+
},
|
| 5237 |
+
{
|
| 5238 |
+
"epoch": 0.9167180752621839,
|
| 5239 |
+
"grad_norm": 612.1376953125,
|
| 5240 |
+
"learning_rate": 1.9571962674051204e-06,
|
| 5241 |
+
"loss": 4.2051,
|
| 5242 |
+
"step": 743
|
| 5243 |
+
},
|
| 5244 |
+
{
|
| 5245 |
+
"epoch": 0.917951881554596,
|
| 5246 |
+
"grad_norm": 963.0797729492188,
|
| 5247 |
+
"learning_rate": 1.9004197393216294e-06,
|
| 5248 |
+
"loss": 4.1168,
|
| 5249 |
+
"step": 744
|
| 5250 |
+
},
|
| 5251 |
+
{
|
| 5252 |
+
"epoch": 0.919185687847008,
|
| 5253 |
+
"grad_norm": 852.234130859375,
|
| 5254 |
+
"learning_rate": 1.84446294175184e-06,
|
| 5255 |
+
"loss": 4.0354,
|
| 5256 |
+
"step": 745
|
| 5257 |
+
},
|
| 5258 |
+
{
|
| 5259 |
+
"epoch": 0.9204194941394201,
|
| 5260 |
+
"grad_norm": 1199.9132080078125,
|
| 5261 |
+
"learning_rate": 1.7893268283318276e-06,
|
| 5262 |
+
"loss": 4.3468,
|
| 5263 |
+
"step": 746
|
| 5264 |
+
},
|
| 5265 |
+
{
|
| 5266 |
+
"epoch": 0.9216533004318322,
|
| 5267 |
+
"grad_norm": 724.0450439453125,
|
| 5268 |
+
"learning_rate": 1.735012338711256e-06,
|
| 5269 |
+
"loss": 4.0663,
|
| 5270 |
+
"step": 747
|
| 5271 |
+
},
|
| 5272 |
+
{
|
| 5273 |
+
"epoch": 0.9228871067242443,
|
| 5274 |
+
"grad_norm": 807.9916381835938,
|
| 5275 |
+
"learning_rate": 1.681520398537373e-06,
|
| 5276 |
+
"loss": 4.116,
|
| 5277 |
+
"step": 748
|
| 5278 |
+
},
|
| 5279 |
+
{
|
| 5280 |
+
"epoch": 0.9241209130166563,
|
| 5281 |
+
"grad_norm": 1655.5943603515625,
|
| 5282 |
+
"learning_rate": 1.6288519194392615e-06,
|
| 5283 |
+
"loss": 4.157,
|
| 5284 |
+
"step": 749
|
| 5285 |
+
},
|
| 5286 |
+
{
|
| 5287 |
+
"epoch": 0.9253547193090684,
|
| 5288 |
+
"grad_norm": 559.2360229492188,
|
| 5289 |
+
"learning_rate": 1.5770077990122645e-06,
|
| 5290 |
+
"loss": 3.9687,
|
| 5291 |
+
"step": 750
|
| 5292 |
+
},
|
| 5293 |
+
{
|
| 5294 |
+
"epoch": 0.9265885256014805,
|
| 5295 |
+
"grad_norm": 1023.9130249023438,
|
| 5296 |
+
"learning_rate": 1.5259889208027013e-06,
|
| 5297 |
+
"loss": 4.1494,
|
| 5298 |
+
"step": 751
|
| 5299 |
+
},
|
| 5300 |
+
{
|
| 5301 |
+
"epoch": 0.9278223318938926,
|
| 5302 |
+
"grad_norm": 682.001220703125,
|
| 5303 |
+
"learning_rate": 1.4757961542928355e-06,
|
| 5304 |
+
"loss": 4.1164,
|
| 5305 |
+
"step": 752
|
| 5306 |
+
},
|
| 5307 |
+
{
|
| 5308 |
+
"epoch": 0.9290561381863047,
|
| 5309 |
+
"grad_norm": 781.3024291992188,
|
| 5310 |
+
"learning_rate": 1.4264303548859992e-06,
|
| 5311 |
+
"loss": 4.0709,
|
| 5312 |
+
"step": 753
|
| 5313 |
+
},
|
| 5314 |
+
{
|
| 5315 |
+
"epoch": 0.9302899444787168,
|
| 5316 |
+
"grad_norm": 1117.8253173828125,
|
| 5317 |
+
"learning_rate": 1.3778923638920971e-06,
|
| 5318 |
+
"loss": 4.2305,
|
| 5319 |
+
"step": 754
|
| 5320 |
+
},
|
| 5321 |
+
{
|
| 5322 |
+
"epoch": 0.9315237507711289,
|
| 5323 |
+
"grad_norm": 982.4514770507812,
|
| 5324 |
+
"learning_rate": 1.330183008513186e-06,
|
| 5325 |
+
"loss": 4.0738,
|
| 5326 |
+
"step": 755
|
| 5327 |
+
},
|
| 5328 |
+
{
|
| 5329 |
+
"epoch": 0.932757557063541,
|
| 5330 |
+
"grad_norm": 663.4027709960938,
|
| 5331 |
+
"learning_rate": 1.283303101829425e-06,
|
| 5332 |
+
"loss": 4.2843,
|
| 5333 |
+
"step": 756
|
| 5334 |
+
},
|
| 5335 |
+
{
|
| 5336 |
+
"epoch": 0.9339913633559531,
|
| 5337 |
+
"grad_norm": 605.0052490234375,
|
| 5338 |
+
"learning_rate": 1.2372534427852079e-06,
|
| 5339 |
+
"loss": 4.1355,
|
| 5340 |
+
"step": 757
|
| 5341 |
+
},
|
| 5342 |
+
{
|
| 5343 |
+
"epoch": 0.9352251696483652,
|
| 5344 |
+
"grad_norm": 678.015625,
|
| 5345 |
+
"learning_rate": 1.1920348161755413e-06,
|
| 5346 |
+
"loss": 4.1688,
|
| 5347 |
+
"step": 758
|
| 5348 |
+
},
|
| 5349 |
+
{
|
| 5350 |
+
"epoch": 0.9364589759407773,
|
| 5351 |
+
"grad_norm": 1134.179443359375,
|
| 5352 |
+
"learning_rate": 1.1476479926326945e-06,
|
| 5353 |
+
"loss": 4.2093,
|
| 5354 |
+
"step": 759
|
| 5355 |
+
},
|
| 5356 |
+
{
|
| 5357 |
+
"epoch": 0.9376927822331894,
|
| 5358 |
+
"grad_norm": 772.4798583984375,
|
| 5359 |
+
"learning_rate": 1.104093728613026e-06,
|
| 5360 |
+
"loss": 4.058,
|
| 5361 |
+
"step": 760
|
| 5362 |
+
},
|
| 5363 |
+
{
|
| 5364 |
+
"epoch": 0.9389265885256015,
|
| 5365 |
+
"grad_norm": 913.7644653320312,
|
| 5366 |
+
"learning_rate": 1.061372766384111e-06,
|
| 5367 |
+
"loss": 4.0062,
|
| 5368 |
+
"step": 761
|
| 5369 |
+
},
|
| 5370 |
+
{
|
| 5371 |
+
"epoch": 0.9401603948180136,
|
| 5372 |
+
"grad_norm": 660.7054443359375,
|
| 5373 |
+
"learning_rate": 1.0194858340121183e-06,
|
| 5374 |
+
"loss": 4.0886,
|
| 5375 |
+
"step": 762
|
| 5376 |
+
},
|
| 5377 |
+
{
|
| 5378 |
+
"epoch": 0.9413942011104257,
|
| 5379 |
+
"grad_norm": 673.8242797851562,
|
| 5380 |
+
"learning_rate": 9.784336453493414e-07,
|
| 5381 |
+
"loss": 4.1026,
|
| 5382 |
+
"step": 763
|
| 5383 |
+
},
|
| 5384 |
+
{
|
| 5385 |
+
"epoch": 0.9426280074028378,
|
| 5386 |
+
"grad_norm": 671.2329711914062,
|
| 5387 |
+
"learning_rate": 9.382169000221041e-07,
|
| 5388 |
+
"loss": 4.1491,
|
| 5389 |
+
"step": 764
|
| 5390 |
+
},
|
| 5391 |
+
{
|
| 5392 |
+
"epoch": 0.9438618136952498,
|
| 5393 |
+
"grad_norm": 581.8679809570312,
|
| 5394 |
+
"learning_rate": 8.988362834187747e-07,
|
| 5395 |
+
"loss": 4.0206,
|
| 5396 |
+
"step": 765
|
| 5397 |
+
},
|
| 5398 |
+
{
|
| 5399 |
+
"epoch": 0.9450956199876619,
|
| 5400 |
+
"grad_norm": 1120.507568359375,
|
| 5401 |
+
"learning_rate": 8.602924666781254e-07,
|
| 5402 |
+
"loss": 4.0462,
|
| 5403 |
+
"step": 766
|
| 5404 |
+
},
|
| 5405 |
+
{
|
| 5406 |
+
"epoch": 0.946329426280074,
|
| 5407 |
+
"grad_norm": 666.160400390625,
|
| 5408 |
+
"learning_rate": 8.225861066778806e-07,
|
| 5409 |
+
"loss": 4.0883,
|
| 5410 |
+
"step": 767
|
| 5411 |
+
},
|
| 5412 |
+
{
|
| 5413 |
+
"epoch": 0.9475632325724861,
|
| 5414 |
+
"grad_norm": 679.1992797851562,
|
| 5415 |
+
"learning_rate": 7.857178460235148e-07,
|
| 5416 |
+
"loss": 3.9554,
|
| 5417 |
+
"step": 768
|
| 5418 |
+
},
|
| 5419 |
+
{
|
| 5420 |
+
"epoch": 0.9487970388648982,
|
| 5421 |
+
"grad_norm": 750.4222412109375,
|
| 5422 |
+
"learning_rate": 7.496883130373167e-07,
|
| 5423 |
+
"loss": 4.1345,
|
| 5424 |
+
"step": 769
|
| 5425 |
+
},
|
| 5426 |
+
{
|
| 5427 |
+
"epoch": 0.9500308451573103,
|
| 5428 |
+
"grad_norm": 989.0843505859375,
|
| 5429 |
+
"learning_rate": 7.144981217476754e-07,
|
| 5430 |
+
"loss": 4.0929,
|
| 5431 |
+
"step": 770
|
| 5432 |
+
},
|
| 5433 |
+
{
|
| 5434 |
+
"epoch": 0.9512646514497224,
|
| 5435 |
+
"grad_norm": 751.5528564453125,
|
| 5436 |
+
"learning_rate": 6.801478718785947e-07,
|
| 5437 |
+
"loss": 4.0241,
|
| 5438 |
+
"step": 771
|
| 5439 |
+
},
|
| 5440 |
+
{
|
| 5441 |
+
"epoch": 0.9524984577421345,
|
| 5442 |
+
"grad_norm": 999.92431640625,
|
| 5443 |
+
"learning_rate": 6.46638148839529e-07,
|
| 5444 |
+
"loss": 4.0454,
|
| 5445 |
+
"step": 772
|
| 5446 |
+
},
|
| 5447 |
+
{
|
| 5448 |
+
"epoch": 0.9537322640345466,
|
| 5449 |
+
"grad_norm": 859.3124389648438,
|
| 5450 |
+
"learning_rate": 6.139695237153298e-07,
|
| 5451 |
+
"loss": 4.1067,
|
| 5452 |
+
"step": 773
|
| 5453 |
+
},
|
| 5454 |
+
{
|
| 5455 |
+
"epoch": 0.9549660703269587,
|
| 5456 |
+
"grad_norm": 935.9942626953125,
|
| 5457 |
+
"learning_rate": 5.821425532565816e-07,
|
| 5458 |
+
"loss": 4.1005,
|
| 5459 |
+
"step": 774
|
| 5460 |
+
},
|
| 5461 |
+
{
|
| 5462 |
+
"epoch": 0.9561998766193708,
|
| 5463 |
+
"grad_norm": 935.444580078125,
|
| 5464 |
+
"learning_rate": 5.511577798700596e-07,
|
| 5465 |
+
"loss": 4.1124,
|
| 5466 |
+
"step": 775
|
| 5467 |
+
},
|
| 5468 |
+
{
|
| 5469 |
+
"epoch": 0.9574336829117829,
|
| 5470 |
+
"grad_norm": 812.61328125,
|
| 5471 |
+
"learning_rate": 5.21015731609531e-07,
|
| 5472 |
+
"loss": 4.0986,
|
| 5473 |
+
"step": 776
|
| 5474 |
+
},
|
| 5475 |
+
{
|
| 5476 |
+
"epoch": 0.958667489204195,
|
| 5477 |
+
"grad_norm": 719.1353149414062,
|
| 5478 |
+
"learning_rate": 4.91716922166735e-07,
|
| 5479 |
+
"loss": 4.0122,
|
| 5480 |
+
"step": 777
|
| 5481 |
+
},
|
| 5482 |
+
{
|
| 5483 |
+
"epoch": 0.9599012954966071,
|
| 5484 |
+
"grad_norm": 614.0670776367188,
|
| 5485 |
+
"learning_rate": 4.632618508626063e-07,
|
| 5486 |
+
"loss": 4.2284,
|
| 5487 |
+
"step": 778
|
| 5488 |
+
},
|
| 5489 |
+
{
|
| 5490 |
+
"epoch": 0.9611351017890192,
|
| 5491 |
+
"grad_norm": 732.4361572265625,
|
| 5492 |
+
"learning_rate": 4.35651002638815e-07,
|
| 5493 |
+
"loss": 4.0553,
|
| 5494 |
+
"step": 779
|
| 5495 |
+
},
|
| 5496 |
+
{
|
| 5497 |
+
"epoch": 0.9623689080814312,
|
| 5498 |
+
"grad_norm": 739.580322265625,
|
| 5499 |
+
"learning_rate": 4.088848480494567e-07,
|
| 5500 |
+
"loss": 4.1197,
|
| 5501 |
+
"step": 780
|
| 5502 |
+
},
|
| 5503 |
+
{
|
| 5504 |
+
"epoch": 0.9636027143738433,
|
| 5505 |
+
"grad_norm": 630.6943359375,
|
| 5506 |
+
"learning_rate": 3.829638432530702e-07,
|
| 5507 |
+
"loss": 4.1957,
|
| 5508 |
+
"step": 781
|
| 5509 |
+
},
|
| 5510 |
+
{
|
| 5511 |
+
"epoch": 0.9648365206662554,
|
| 5512 |
+
"grad_norm": 787.4408569335938,
|
| 5513 |
+
"learning_rate": 3.5788843000481576e-07,
|
| 5514 |
+
"loss": 4.0408,
|
| 5515 |
+
"step": 782
|
| 5516 |
+
},
|
| 5517 |
+
{
|
| 5518 |
+
"epoch": 0.9660703269586675,
|
| 5519 |
+
"grad_norm": 656.0836181640625,
|
| 5520 |
+
"learning_rate": 3.336590356489977e-07,
|
| 5521 |
+
"loss": 4.0707,
|
| 5522 |
+
"step": 783
|
| 5523 |
+
},
|
| 5524 |
+
{
|
| 5525 |
+
"epoch": 0.9673041332510796,
|
| 5526 |
+
"grad_norm": 1066.859130859375,
|
| 5527 |
+
"learning_rate": 3.102760731117593e-07,
|
| 5528 |
+
"loss": 4.08,
|
| 5529 |
+
"step": 784
|
| 5530 |
+
},
|
| 5531 |
+
{
|
| 5532 |
+
"epoch": 0.9685379395434917,
|
| 5533 |
+
"grad_norm": 636.3853149414062,
|
| 5534 |
+
"learning_rate": 2.8773994089402734e-07,
|
| 5535 |
+
"loss": 4.1094,
|
| 5536 |
+
"step": 785
|
| 5537 |
+
},
|
| 5538 |
+
{
|
| 5539 |
+
"epoch": 0.9697717458359038,
|
| 5540 |
+
"grad_norm": 833.309814453125,
|
| 5541 |
+
"learning_rate": 2.6605102306476724e-07,
|
| 5542 |
+
"loss": 4.2544,
|
| 5543 |
+
"step": 786
|
| 5544 |
+
},
|
| 5545 |
+
{
|
| 5546 |
+
"epoch": 0.9710055521283159,
|
| 5547 |
+
"grad_norm": 769.7086181640625,
|
| 5548 |
+
"learning_rate": 2.452096892543776e-07,
|
| 5549 |
+
"loss": 4.152,
|
| 5550 |
+
"step": 787
|
| 5551 |
+
},
|
| 5552 |
+
{
|
| 5553 |
+
"epoch": 0.972239358420728,
|
| 5554 |
+
"grad_norm": 737.0515747070312,
|
| 5555 |
+
"learning_rate": 2.2521629464844486e-07,
|
| 5556 |
+
"loss": 4.2431,
|
| 5557 |
+
"step": 788
|
| 5558 |
+
},
|
| 5559 |
+
{
|
| 5560 |
+
"epoch": 0.97347316471314,
|
| 5561 |
+
"grad_norm": 683.5414428710938,
|
| 5562 |
+
"learning_rate": 2.0607117998165947e-07,
|
| 5563 |
+
"loss": 4.2063,
|
| 5564 |
+
"step": 789
|
| 5565 |
+
},
|
| 5566 |
+
{
|
| 5567 |
+
"epoch": 0.9747069710055521,
|
| 5568 |
+
"grad_norm": 725.9776000976562,
|
| 5569 |
+
"learning_rate": 1.8777467153202054e-07,
|
| 5570 |
+
"loss": 4.2039,
|
| 5571 |
+
"step": 790
|
| 5572 |
+
},
|
| 5573 |
+
{
|
| 5574 |
+
"epoch": 0.9759407772979642,
|
| 5575 |
+
"grad_norm": 562.0029907226562,
|
| 5576 |
+
"learning_rate": 1.703270811152624e-07,
|
| 5577 |
+
"loss": 4.1486,
|
| 5578 |
+
"step": 791
|
| 5579 |
+
},
|
| 5580 |
+
{
|
| 5581 |
+
"epoch": 0.9771745835903763,
|
| 5582 |
+
"grad_norm": 756.2742309570312,
|
| 5583 |
+
"learning_rate": 1.5372870607956446e-07,
|
| 5584 |
+
"loss": 4.0233,
|
| 5585 |
+
"step": 792
|
| 5586 |
+
},
|
| 5587 |
+
{
|
| 5588 |
+
"epoch": 0.9784083898827884,
|
| 5589 |
+
"grad_norm": 777.6002197265625,
|
| 5590 |
+
"learning_rate": 1.3797982930044972e-07,
|
| 5591 |
+
"loss": 4.1716,
|
| 5592 |
+
"step": 793
|
| 5593 |
+
},
|
| 5594 |
+
{
|
| 5595 |
+
"epoch": 0.9796421961752005,
|
| 5596 |
+
"grad_norm": 1127.728759765625,
|
| 5597 |
+
"learning_rate": 1.2308071917601083e-07,
|
| 5598 |
+
"loss": 4.2959,
|
| 5599 |
+
"step": 794
|
| 5600 |
+
},
|
| 5601 |
+
{
|
| 5602 |
+
"epoch": 0.9808760024676126,
|
| 5603 |
+
"grad_norm": 827.7559814453125,
|
| 5604 |
+
"learning_rate": 1.0903162962228597e-07,
|
| 5605 |
+
"loss": 4.11,
|
| 5606 |
+
"step": 795
|
| 5607 |
+
},
|
| 5608 |
+
{
|
| 5609 |
+
"epoch": 0.9821098087600246,
|
| 5610 |
+
"grad_norm": 734.08984375,
|
| 5611 |
+
"learning_rate": 9.583280006895679e-08,
|
| 5612 |
+
"loss": 4.0029,
|
| 5613 |
+
"step": 796
|
| 5614 |
+
},
|
| 5615 |
+
{
|
| 5616 |
+
"epoch": 0.9833436150524367,
|
| 5617 |
+
"grad_norm": 1221.7593994140625,
|
| 5618 |
+
"learning_rate": 8.348445545527939e-08,
|
| 5619 |
+
"loss": 4.4357,
|
| 5620 |
+
"step": 797
|
| 5621 |
+
},
|
| 5622 |
+
{
|
| 5623 |
+
"epoch": 0.9845774213448488,
|
| 5624 |
+
"grad_norm": 632.0841674804688,
|
| 5625 |
+
"learning_rate": 7.198680622621523e-08,
|
| 5626 |
+
"loss": 4.0149,
|
| 5627 |
+
"step": 798
|
| 5628 |
+
},
|
| 5629 |
+
{
|
| 5630 |
+
"epoch": 0.9858112276372609,
|
| 5631 |
+
"grad_norm": 644.5614624023438,
|
| 5632 |
+
"learning_rate": 6.134004832888396e-08,
|
| 5633 |
+
"loss": 4.1734,
|
| 5634 |
+
"step": 799
|
| 5635 |
+
},
|
| 5636 |
+
{
|
| 5637 |
+
"epoch": 0.987045033929673,
|
| 5638 |
+
"grad_norm": 802.9095458984375,
|
| 5639 |
+
"learning_rate": 5.154436320919942e-08,
|
| 5640 |
+
"loss": 4.0969,
|
| 5641 |
+
"step": 800
|
| 5642 |
}
|
| 5643 |
],
|
| 5644 |
"logging_steps": 1,
|
|
|
|
| 5658 |
"attributes": {}
|
| 5659 |
}
|
| 5660 |
},
|
| 5661 |
+
"total_flos": 5313340347777024.0,
|
| 5662 |
"train_batch_size": 2,
|
| 5663 |
"trial_name": null,
|
| 5664 |
"trial_params": null
|