Training in progress, step 833, 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 +816 -4
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2145944
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07cefc2581899e9cd81423324de5956c82dd5f731f05187440a982b0d6d05b71
|
| 3 |
size 2145944
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4310020
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4570e5c4b0ed79e4f5d3b402f6ad2f2549e244175cc54843a5d5f6d813db1532
|
| 3 |
size 4310020
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50f87c0ed2d75cb8a451ddff6da6632b673785da1e9fef3dc083d69f65b801bb
|
| 3 |
size 15024
|
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 sha256:756858fb0f5a27c92f1116791d526a1becaf91fa7110176767537e98ea42bbd8
|
| 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:d7d1a6c204d94ba0bb67bac856dedd2342d34022ad605da33963f9022dc066a7
|
| 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:d3408db92c33886624cd88dac64c08b23fc69f8e188c300a1a0bfe82c164e648
|
| 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 sha256:814c5335d071ae07021d1744ec9a6235158c1a31d5d0598e1cda6bddebbe2205
|
| 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": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -5026,6 +5026,818 @@
|
|
| 5026 |
"learning_rate": 6.053180944220627e-06,
|
| 5027 |
"loss": 9.5913,
|
| 5028 |
"step": 717
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5029 |
}
|
| 5030 |
],
|
| 5031 |
"logging_steps": 1,
|
|
@@ -5040,12 +5852,12 @@
|
|
| 5040 |
"should_evaluate": false,
|
| 5041 |
"should_log": false,
|
| 5042 |
"should_save": true,
|
| 5043 |
-
"should_training_stop":
|
| 5044 |
},
|
| 5045 |
"attributes": {}
|
| 5046 |
}
|
| 5047 |
},
|
| 5048 |
-
"total_flos":
|
| 5049 |
"train_batch_size": 4,
|
| 5050 |
"trial_name": null,
|
| 5051 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9991004497751125,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 833,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5026 |
"learning_rate": 6.053180944220627e-06,
|
| 5027 |
"loss": 9.5913,
|
| 5028 |
"step": 717
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.8611694152923538,
|
| 5032 |
+
"grad_norm": 0.2777814567089081,
|
| 5033 |
+
"learning_rate": 5.951378333139118e-06,
|
| 5034 |
+
"loss": 9.604,
|
| 5035 |
+
"step": 718
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.8623688155922039,
|
| 5039 |
+
"grad_norm": 0.27990755438804626,
|
| 5040 |
+
"learning_rate": 5.850384861654329e-06,
|
| 5041 |
+
"loss": 9.5929,
|
| 5042 |
+
"step": 719
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.863568215892054,
|
| 5046 |
+
"grad_norm": 0.285398006439209,
|
| 5047 |
+
"learning_rate": 5.750202384939313e-06,
|
| 5048 |
+
"loss": 9.5953,
|
| 5049 |
+
"step": 720
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.864767616191904,
|
| 5053 |
+
"grad_norm": 0.28564396500587463,
|
| 5054 |
+
"learning_rate": 5.650832743269779e-06,
|
| 5055 |
+
"loss": 9.6041,
|
| 5056 |
+
"step": 721
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.8659670164917541,
|
| 5060 |
+
"grad_norm": 0.29116764664649963,
|
| 5061 |
+
"learning_rate": 5.552277761990294e-06,
|
| 5062 |
+
"loss": 9.594,
|
| 5063 |
+
"step": 722
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.8671664167916042,
|
| 5067 |
+
"grad_norm": 0.28536704182624817,
|
| 5068 |
+
"learning_rate": 5.454539251480739e-06,
|
| 5069 |
+
"loss": 9.6037,
|
| 5070 |
+
"step": 723
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.8683658170914542,
|
| 5074 |
+
"grad_norm": 0.2976490259170532,
|
| 5075 |
+
"learning_rate": 5.3576190071230106e-06,
|
| 5076 |
+
"loss": 9.6027,
|
| 5077 |
+
"step": 724
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.8695652173913043,
|
| 5081 |
+
"grad_norm": 0.3443441390991211,
|
| 5082 |
+
"learning_rate": 5.2615188092681176e-06,
|
| 5083 |
+
"loss": 9.6111,
|
| 5084 |
+
"step": 725
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.8707646176911544,
|
| 5088 |
+
"grad_norm": 0.27186739444732666,
|
| 5089 |
+
"learning_rate": 5.166240423203428e-06,
|
| 5090 |
+
"loss": 9.5908,
|
| 5091 |
+
"step": 726
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.8719640179910045,
|
| 5095 |
+
"grad_norm": 0.2679988443851471,
|
| 5096 |
+
"learning_rate": 5.071785599120243e-06,
|
| 5097 |
+
"loss": 9.5937,
|
| 5098 |
+
"step": 727
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.8731634182908545,
|
| 5102 |
+
"grad_norm": 0.2798680067062378,
|
| 5103 |
+
"learning_rate": 4.978156072081669e-06,
|
| 5104 |
+
"loss": 9.5911,
|
| 5105 |
+
"step": 728
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.8743628185907046,
|
| 5109 |
+
"grad_norm": 0.27777785062789917,
|
| 5110 |
+
"learning_rate": 4.885353561990752e-06,
|
| 5111 |
+
"loss": 9.5982,
|
| 5112 |
+
"step": 729
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.8755622188905547,
|
| 5116 |
+
"grad_norm": 0.2682747542858124,
|
| 5117 |
+
"learning_rate": 4.793379773558815e-06,
|
| 5118 |
+
"loss": 9.6045,
|
| 5119 |
+
"step": 730
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.8767616191904049,
|
| 5123 |
+
"grad_norm": 0.2757691740989685,
|
| 5124 |
+
"learning_rate": 4.7022363962742514e-06,
|
| 5125 |
+
"loss": 9.6068,
|
| 5126 |
+
"step": 731
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.8779610194902548,
|
| 5130 |
+
"grad_norm": 0.27363264560699463,
|
| 5131 |
+
"learning_rate": 4.6119251043714225e-06,
|
| 5132 |
+
"loss": 9.5998,
|
| 5133 |
+
"step": 732
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.879160419790105,
|
| 5137 |
+
"grad_norm": 0.28155508637428284,
|
| 5138 |
+
"learning_rate": 4.522447556799875e-06,
|
| 5139 |
+
"loss": 9.5994,
|
| 5140 |
+
"step": 733
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.8803598200899551,
|
| 5144 |
+
"grad_norm": 0.27953052520751953,
|
| 5145 |
+
"learning_rate": 4.433805397193969e-06,
|
| 5146 |
+
"loss": 9.6041,
|
| 5147 |
+
"step": 734
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.881559220389805,
|
| 5151 |
+
"grad_norm": 0.28824350237846375,
|
| 5152 |
+
"learning_rate": 4.3460002538425805e-06,
|
| 5153 |
+
"loss": 9.6045,
|
| 5154 |
+
"step": 735
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.8827586206896552,
|
| 5158 |
+
"grad_norm": 0.29731664061546326,
|
| 5159 |
+
"learning_rate": 4.2590337396592406e-06,
|
| 5160 |
+
"loss": 9.6076,
|
| 5161 |
+
"step": 736
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.8839580209895053,
|
| 5165 |
+
"grad_norm": 0.2972903847694397,
|
| 5166 |
+
"learning_rate": 4.172907452152519e-06,
|
| 5167 |
+
"loss": 9.608,
|
| 5168 |
+
"step": 737
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.8851574212893554,
|
| 5172 |
+
"grad_norm": 0.2724793255329132,
|
| 5173 |
+
"learning_rate": 4.087622973396665e-06,
|
| 5174 |
+
"loss": 9.5971,
|
| 5175 |
+
"step": 738
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.8863568215892054,
|
| 5179 |
+
"grad_norm": 0.2679544985294342,
|
| 5180 |
+
"learning_rate": 4.0031818700025095e-06,
|
| 5181 |
+
"loss": 9.5936,
|
| 5182 |
+
"step": 739
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.8875562218890555,
|
| 5186 |
+
"grad_norm": 0.2719581425189972,
|
| 5187 |
+
"learning_rate": 3.919585693088751e-06,
|
| 5188 |
+
"loss": 9.592,
|
| 5189 |
+
"step": 740
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.8887556221889056,
|
| 5193 |
+
"grad_norm": 0.27202773094177246,
|
| 5194 |
+
"learning_rate": 3.836835978253433e-06,
|
| 5195 |
+
"loss": 9.5904,
|
| 5196 |
+
"step": 741
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.8899550224887556,
|
| 5200 |
+
"grad_norm": 0.2740086615085602,
|
| 5201 |
+
"learning_rate": 3.7549342455457216e-06,
|
| 5202 |
+
"loss": 9.5989,
|
| 5203 |
+
"step": 742
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.8911544227886057,
|
| 5207 |
+
"grad_norm": 0.2741098999977112,
|
| 5208 |
+
"learning_rate": 3.6738819994379945e-06,
|
| 5209 |
+
"loss": 9.6006,
|
| 5210 |
+
"step": 743
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.8923538230884558,
|
| 5214 |
+
"grad_norm": 0.2779156565666199,
|
| 5215 |
+
"learning_rate": 3.593680728798238e-06,
|
| 5216 |
+
"loss": 9.5999,
|
| 5217 |
+
"step": 744
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.8935532233883059,
|
| 5221 |
+
"grad_norm": 0.28172117471694946,
|
| 5222 |
+
"learning_rate": 3.5143319068626225e-06,
|
| 5223 |
+
"loss": 9.6032,
|
| 5224 |
+
"step": 745
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.8947526236881559,
|
| 5228 |
+
"grad_norm": 0.2854410409927368,
|
| 5229 |
+
"learning_rate": 3.435836991208524e-06,
|
| 5230 |
+
"loss": 9.5996,
|
| 5231 |
+
"step": 746
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.895952023988006,
|
| 5235 |
+
"grad_norm": 0.2835351824760437,
|
| 5236 |
+
"learning_rate": 3.35819742372771e-06,
|
| 5237 |
+
"loss": 9.6017,
|
| 5238 |
+
"step": 747
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.8971514242878561,
|
| 5242 |
+
"grad_norm": 0.2815116047859192,
|
| 5243 |
+
"learning_rate": 3.2814146305998107e-06,
|
| 5244 |
+
"loss": 9.605,
|
| 5245 |
+
"step": 748
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.8983508245877061,
|
| 5249 |
+
"grad_norm": 0.3012433648109436,
|
| 5250 |
+
"learning_rate": 3.2054900222662276e-06,
|
| 5251 |
+
"loss": 9.607,
|
| 5252 |
+
"step": 749
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.8995502248875562,
|
| 5256 |
+
"grad_norm": 0.3453059792518616,
|
| 5257 |
+
"learning_rate": 3.1304249934041017e-06,
|
| 5258 |
+
"loss": 9.6038,
|
| 5259 |
+
"step": 750
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.9007496251874063,
|
| 5263 |
+
"grad_norm": 0.2700274586677551,
|
| 5264 |
+
"learning_rate": 3.0562209229008042e-06,
|
| 5265 |
+
"loss": 9.5894,
|
| 5266 |
+
"step": 751
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.9019490254872564,
|
| 5270 |
+
"grad_norm": 0.26212278008461,
|
| 5271 |
+
"learning_rate": 2.982879173828523e-06,
|
| 5272 |
+
"loss": 9.5959,
|
| 5273 |
+
"step": 752
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.9031484257871064,
|
| 5277 |
+
"grad_norm": 0.2739808261394501,
|
| 5278 |
+
"learning_rate": 2.9104010934192794e-06,
|
| 5279 |
+
"loss": 9.5937,
|
| 5280 |
+
"step": 753
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.9043478260869565,
|
| 5284 |
+
"grad_norm": 0.2720181345939636,
|
| 5285 |
+
"learning_rate": 2.838788013040139e-06,
|
| 5286 |
+
"loss": 9.5944,
|
| 5287 |
+
"step": 754
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.9055472263868066,
|
| 5291 |
+
"grad_norm": 0.2759906053543091,
|
| 5292 |
+
"learning_rate": 2.768041248168801e-06,
|
| 5293 |
+
"loss": 9.5952,
|
| 5294 |
+
"step": 755
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.9067466266866566,
|
| 5298 |
+
"grad_norm": 0.27012184262275696,
|
| 5299 |
+
"learning_rate": 2.6981620983694057e-06,
|
| 5300 |
+
"loss": 9.5966,
|
| 5301 |
+
"step": 756
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.9079460269865067,
|
| 5305 |
+
"grad_norm": 0.2913030683994293,
|
| 5306 |
+
"learning_rate": 2.6291518472686404e-06,
|
| 5307 |
+
"loss": 9.5965,
|
| 5308 |
+
"step": 757
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.9091454272863568,
|
| 5312 |
+
"grad_norm": 0.2893541753292084,
|
| 5313 |
+
"learning_rate": 2.5610117625322118e-06,
|
| 5314 |
+
"loss": 9.5964,
|
| 5315 |
+
"step": 758
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.9103448275862069,
|
| 5319 |
+
"grad_norm": 0.2874806523323059,
|
| 5320 |
+
"learning_rate": 2.4937430958415278e-06,
|
| 5321 |
+
"loss": 9.5995,
|
| 5322 |
+
"step": 759
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.9115442278860569,
|
| 5326 |
+
"grad_norm": 0.2814995348453522,
|
| 5327 |
+
"learning_rate": 2.427347082870701e-06,
|
| 5328 |
+
"loss": 9.6001,
|
| 5329 |
+
"step": 760
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.912743628185907,
|
| 5333 |
+
"grad_norm": 0.28927141427993774,
|
| 5334 |
+
"learning_rate": 2.361824943263874e-06,
|
| 5335 |
+
"loss": 9.6073,
|
| 5336 |
+
"step": 761
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.9139430284857571,
|
| 5340 |
+
"grad_norm": 0.29837363958358765,
|
| 5341 |
+
"learning_rate": 2.2971778806127996e-06,
|
| 5342 |
+
"loss": 9.6054,
|
| 5343 |
+
"step": 762
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.9151424287856071,
|
| 5347 |
+
"grad_norm": 0.25863122940063477,
|
| 5348 |
+
"learning_rate": 2.233407082434724e-06,
|
| 5349 |
+
"loss": 9.6017,
|
| 5350 |
+
"step": 763
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.9163418290854572,
|
| 5354 |
+
"grad_norm": 0.26800212264060974,
|
| 5355 |
+
"learning_rate": 2.1705137201505965e-06,
|
| 5356 |
+
"loss": 9.5951,
|
| 5357 |
+
"step": 764
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.9175412293853074,
|
| 5361 |
+
"grad_norm": 0.2701495587825775,
|
| 5362 |
+
"learning_rate": 2.1084989490635255e-06,
|
| 5363 |
+
"loss": 9.5911,
|
| 5364 |
+
"step": 765
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.9187406296851575,
|
| 5368 |
+
"grad_norm": 0.27781182527542114,
|
| 5369 |
+
"learning_rate": 2.0473639083375795e-06,
|
| 5370 |
+
"loss": 9.5949,
|
| 5371 |
+
"step": 766
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.9199400299850075,
|
| 5375 |
+
"grad_norm": 0.26822102069854736,
|
| 5376 |
+
"learning_rate": 1.9871097209768375e-06,
|
| 5377 |
+
"loss": 9.6018,
|
| 5378 |
+
"step": 767
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.9211394302848576,
|
| 5382 |
+
"grad_norm": 0.2740313708782196,
|
| 5383 |
+
"learning_rate": 1.9277374938047988e-06,
|
| 5384 |
+
"loss": 9.5969,
|
| 5385 |
+
"step": 768
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.9223388305847077,
|
| 5389 |
+
"grad_norm": 0.28559255599975586,
|
| 5390 |
+
"learning_rate": 1.8692483174439946e-06,
|
| 5391 |
+
"loss": 9.5933,
|
| 5392 |
+
"step": 769
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.9235382308845578,
|
| 5396 |
+
"grad_norm": 0.28170379996299744,
|
| 5397 |
+
"learning_rate": 1.8116432662960037e-06,
|
| 5398 |
+
"loss": 9.595,
|
| 5399 |
+
"step": 770
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.9247376311844078,
|
| 5403 |
+
"grad_norm": 0.2856314182281494,
|
| 5404 |
+
"learning_rate": 1.7549233985217074e-06,
|
| 5405 |
+
"loss": 9.5949,
|
| 5406 |
+
"step": 771
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.9259370314842579,
|
| 5410 |
+
"grad_norm": 0.28359171748161316,
|
| 5411 |
+
"learning_rate": 1.6990897560218211e-06,
|
| 5412 |
+
"loss": 9.5984,
|
| 5413 |
+
"step": 772
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 0.927136431784108,
|
| 5417 |
+
"grad_norm": 0.28927284479141235,
|
| 5418 |
+
"learning_rate": 1.644143364417794e-06,
|
| 5419 |
+
"loss": 9.6003,
|
| 5420 |
+
"step": 773
|
| 5421 |
+
},
|
| 5422 |
+
{
|
| 5423 |
+
"epoch": 0.928335832083958,
|
| 5424 |
+
"grad_norm": 0.28748857975006104,
|
| 5425 |
+
"learning_rate": 1.5900852330329563e-06,
|
| 5426 |
+
"loss": 9.6163,
|
| 5427 |
+
"step": 774
|
| 5428 |
+
},
|
| 5429 |
+
{
|
| 5430 |
+
"epoch": 0.9295352323838081,
|
| 5431 |
+
"grad_norm": 0.32728201150894165,
|
| 5432 |
+
"learning_rate": 1.5369163548739462e-06,
|
| 5433 |
+
"loss": 9.6079,
|
| 5434 |
+
"step": 775
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 0.9307346326836582,
|
| 5438 |
+
"grad_norm": 0.2640519440174103,
|
| 5439 |
+
"learning_rate": 1.484637706612535e-06,
|
| 5440 |
+
"loss": 9.5952,
|
| 5441 |
+
"step": 776
|
| 5442 |
+
},
|
| 5443 |
+
{
|
| 5444 |
+
"epoch": 0.9319340329835083,
|
| 5445 |
+
"grad_norm": 0.2757025361061096,
|
| 5446 |
+
"learning_rate": 1.4332502485676358e-06,
|
| 5447 |
+
"loss": 9.5968,
|
| 5448 |
+
"step": 777
|
| 5449 |
+
},
|
| 5450 |
+
{
|
| 5451 |
+
"epoch": 0.9331334332833583,
|
| 5452 |
+
"grad_norm": 0.2660670578479767,
|
| 5453 |
+
"learning_rate": 1.3827549246876625e-06,
|
| 5454 |
+
"loss": 9.5966,
|
| 5455 |
+
"step": 778
|
| 5456 |
+
},
|
| 5457 |
+
{
|
| 5458 |
+
"epoch": 0.9343328335832084,
|
| 5459 |
+
"grad_norm": 0.26621681451797485,
|
| 5460 |
+
"learning_rate": 1.333152662533227e-06,
|
| 5461 |
+
"loss": 9.6007,
|
| 5462 |
+
"step": 779
|
| 5463 |
+
},
|
| 5464 |
+
{
|
| 5465 |
+
"epoch": 0.9355322338830585,
|
| 5466 |
+
"grad_norm": 0.2719038426876068,
|
| 5467 |
+
"learning_rate": 1.2844443732600576e-06,
|
| 5468 |
+
"loss": 9.6027,
|
| 5469 |
+
"step": 780
|
| 5470 |
+
},
|
| 5471 |
+
{
|
| 5472 |
+
"epoch": 0.9367316341829085,
|
| 5473 |
+
"grad_norm": 0.278006911277771,
|
| 5474 |
+
"learning_rate": 1.2366309516022966e-06,
|
| 5475 |
+
"loss": 9.6006,
|
| 5476 |
+
"step": 781
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 0.9379310344827586,
|
| 5480 |
+
"grad_norm": 0.2759842574596405,
|
| 5481 |
+
"learning_rate": 1.189713275856047e-06,
|
| 5482 |
+
"loss": 9.6019,
|
| 5483 |
+
"step": 782
|
| 5484 |
+
},
|
| 5485 |
+
{
|
| 5486 |
+
"epoch": 0.9391304347826087,
|
| 5487 |
+
"grad_norm": 0.2800839841365814,
|
| 5488 |
+
"learning_rate": 1.1436922078632394e-06,
|
| 5489 |
+
"loss": 9.6011,
|
| 5490 |
+
"step": 783
|
| 5491 |
+
},
|
| 5492 |
+
{
|
| 5493 |
+
"epoch": 0.9403298350824588,
|
| 5494 |
+
"grad_norm": 0.2854614555835724,
|
| 5495 |
+
"learning_rate": 1.0985685929958134e-06,
|
| 5496 |
+
"loss": 9.6005,
|
| 5497 |
+
"step": 784
|
| 5498 |
+
},
|
| 5499 |
+
{
|
| 5500 |
+
"epoch": 0.9415292353823088,
|
| 5501 |
+
"grad_norm": 0.28942301869392395,
|
| 5502 |
+
"learning_rate": 1.0543432601401615e-06,
|
| 5503 |
+
"loss": 9.5996,
|
| 5504 |
+
"step": 785
|
| 5505 |
+
},
|
| 5506 |
+
{
|
| 5507 |
+
"epoch": 0.9427286356821589,
|
| 5508 |
+
"grad_norm": 0.29389646649360657,
|
| 5509 |
+
"learning_rate": 1.0110170216819316e-06,
|
| 5510 |
+
"loss": 9.6074,
|
| 5511 |
+
"step": 786
|
| 5512 |
+
},
|
| 5513 |
+
{
|
| 5514 |
+
"epoch": 0.943928035982009,
|
| 5515 |
+
"grad_norm": 0.29730501770973206,
|
| 5516 |
+
"learning_rate": 9.685906734910988e-07,
|
| 5517 |
+
"loss": 9.6045,
|
| 5518 |
+
"step": 787
|
| 5519 |
+
},
|
| 5520 |
+
{
|
| 5521 |
+
"epoch": 0.945127436281859,
|
| 5522 |
+
"grad_norm": 0.27036651968955994,
|
| 5523 |
+
"learning_rate": 9.270649949073229e-07,
|
| 5524 |
+
"loss": 9.5922,
|
| 5525 |
+
"step": 788
|
| 5526 |
+
},
|
| 5527 |
+
{
|
| 5528 |
+
"epoch": 0.9463268365817091,
|
| 5529 |
+
"grad_norm": 0.26795369386672974,
|
| 5530 |
+
"learning_rate": 8.864407487256699e-07,
|
| 5531 |
+
"loss": 9.5981,
|
| 5532 |
+
"step": 789
|
| 5533 |
+
},
|
| 5534 |
+
{
|
| 5535 |
+
"epoch": 0.9475262368815592,
|
| 5536 |
+
"grad_norm": 0.27786487340927124,
|
| 5537 |
+
"learning_rate": 8.467186811825623e-07,
|
| 5538 |
+
"loss": 9.5882,
|
| 5539 |
+
"step": 790
|
| 5540 |
+
},
|
| 5541 |
+
{
|
| 5542 |
+
"epoch": 0.9487256371814093,
|
| 5543 |
+
"grad_norm": 0.27581360936164856,
|
| 5544 |
+
"learning_rate": 8.07899521942096e-07,
|
| 5545 |
+
"loss": 9.5986,
|
| 5546 |
+
"step": 791
|
| 5547 |
+
},
|
| 5548 |
+
{
|
| 5549 |
+
"epoch": 0.9499250374812593,
|
| 5550 |
+
"grad_norm": 0.27784955501556396,
|
| 5551 |
+
"learning_rate": 7.69983984082634e-07,
|
| 5552 |
+
"loss": 9.5951,
|
| 5553 |
+
"step": 792
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 0.9511244377811094,
|
| 5557 |
+
"grad_norm": 0.2779393196105957,
|
| 5558 |
+
"learning_rate": 7.329727640837058e-07,
|
| 5559 |
+
"loss": 9.5992,
|
| 5560 |
+
"step": 793
|
| 5561 |
+
},
|
| 5562 |
+
{
|
| 5563 |
+
"epoch": 0.9523238380809596,
|
| 5564 |
+
"grad_norm": 0.2856026291847229,
|
| 5565 |
+
"learning_rate": 6.968665418131848e-07,
|
| 5566 |
+
"loss": 9.6011,
|
| 5567 |
+
"step": 794
|
| 5568 |
+
},
|
| 5569 |
+
{
|
| 5570 |
+
"epoch": 0.9535232383808095,
|
| 5571 |
+
"grad_norm": 0.2818286418914795,
|
| 5572 |
+
"learning_rate": 6.616659805148695e-07,
|
| 5573 |
+
"loss": 9.603,
|
| 5574 |
+
"step": 795
|
| 5575 |
+
},
|
| 5576 |
+
{
|
| 5577 |
+
"epoch": 0.9547226386806597,
|
| 5578 |
+
"grad_norm": 0.2779069244861603,
|
| 5579 |
+
"learning_rate": 6.273717267962164e-07,
|
| 5580 |
+
"loss": 9.6057,
|
| 5581 |
+
"step": 796
|
| 5582 |
+
},
|
| 5583 |
+
{
|
| 5584 |
+
"epoch": 0.9559220389805098,
|
| 5585 |
+
"grad_norm": 0.28562232851982117,
|
| 5586 |
+
"learning_rate": 5.93984410616527e-07,
|
| 5587 |
+
"loss": 9.5981,
|
| 5588 |
+
"step": 797
|
| 5589 |
+
},
|
| 5590 |
+
{
|
| 5591 |
+
"epoch": 0.9571214392803599,
|
| 5592 |
+
"grad_norm": 0.2814895510673523,
|
| 5593 |
+
"learning_rate": 5.615046452753403e-07,
|
| 5594 |
+
"loss": 9.6072,
|
| 5595 |
+
"step": 798
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 0.9583208395802099,
|
| 5599 |
+
"grad_norm": 0.291547030210495,
|
| 5600 |
+
"learning_rate": 5.299330274011916e-07,
|
| 5601 |
+
"loss": 9.6075,
|
| 5602 |
+
"step": 799
|
| 5603 |
+
},
|
| 5604 |
+
{
|
| 5605 |
+
"epoch": 0.95952023988006,
|
| 5606 |
+
"grad_norm": 0.3199189603328705,
|
| 5607 |
+
"learning_rate": 4.992701369406161e-07,
|
| 5608 |
+
"loss": 9.6069,
|
| 5609 |
+
"step": 800
|
| 5610 |
+
},
|
| 5611 |
+
{
|
| 5612 |
+
"epoch": 0.9607196401799101,
|
| 5613 |
+
"grad_norm": 0.26192960143089294,
|
| 5614 |
+
"learning_rate": 4.695165371475463e-07,
|
| 5615 |
+
"loss": 9.5961,
|
| 5616 |
+
"step": 801
|
| 5617 |
+
},
|
| 5618 |
+
{
|
| 5619 |
+
"epoch": 0.9619190404797601,
|
| 5620 |
+
"grad_norm": 0.27191805839538574,
|
| 5621 |
+
"learning_rate": 4.4067277457292556e-07,
|
| 5622 |
+
"loss": 9.5925,
|
| 5623 |
+
"step": 802
|
| 5624 |
+
},
|
| 5625 |
+
{
|
| 5626 |
+
"epoch": 0.9631184407796102,
|
| 5627 |
+
"grad_norm": 0.2721245288848877,
|
| 5628 |
+
"learning_rate": 4.1273937905467185e-07,
|
| 5629 |
+
"loss": 9.5944,
|
| 5630 |
+
"step": 803
|
| 5631 |
+
},
|
| 5632 |
+
{
|
| 5633 |
+
"epoch": 0.9643178410794603,
|
| 5634 |
+
"grad_norm": 0.27019554376602173,
|
| 5635 |
+
"learning_rate": 3.8571686370797443e-07,
|
| 5636 |
+
"loss": 9.5939,
|
| 5637 |
+
"step": 804
|
| 5638 |
+
},
|
| 5639 |
+
{
|
| 5640 |
+
"epoch": 0.9655172413793104,
|
| 5641 |
+
"grad_norm": 0.27776581048965454,
|
| 5642 |
+
"learning_rate": 3.5960572491583466e-07,
|
| 5643 |
+
"loss": 9.5946,
|
| 5644 |
+
"step": 805
|
| 5645 |
+
},
|
| 5646 |
+
{
|
| 5647 |
+
"epoch": 0.9667166416791604,
|
| 5648 |
+
"grad_norm": 0.27804285287857056,
|
| 5649 |
+
"learning_rate": 3.3440644231995664e-07,
|
| 5650 |
+
"loss": 9.5989,
|
| 5651 |
+
"step": 806
|
| 5652 |
+
},
|
| 5653 |
+
{
|
| 5654 |
+
"epoch": 0.9679160419790105,
|
| 5655 |
+
"grad_norm": 0.27963870763778687,
|
| 5656 |
+
"learning_rate": 3.101194788119599e-07,
|
| 5657 |
+
"loss": 9.5996,
|
| 5658 |
+
"step": 807
|
| 5659 |
+
},
|
| 5660 |
+
{
|
| 5661 |
+
"epoch": 0.9691154422788606,
|
| 5662 |
+
"grad_norm": 0.27757352590560913,
|
| 5663 |
+
"learning_rate": 2.867452805248416e-07,
|
| 5664 |
+
"loss": 9.5986,
|
| 5665 |
+
"step": 808
|
| 5666 |
+
},
|
| 5667 |
+
{
|
| 5668 |
+
"epoch": 0.9703148425787106,
|
| 5669 |
+
"grad_norm": 0.27756595611572266,
|
| 5670 |
+
"learning_rate": 2.642842768248055e-07,
|
| 5671 |
+
"loss": 9.6037,
|
| 5672 |
+
"step": 809
|
| 5673 |
+
},
|
| 5674 |
+
{
|
| 5675 |
+
"epoch": 0.9715142428785607,
|
| 5676 |
+
"grad_norm": 0.2873465418815613,
|
| 5677 |
+
"learning_rate": 2.4273688030336805e-07,
|
| 5678 |
+
"loss": 9.6023,
|
| 5679 |
+
"step": 810
|
| 5680 |
+
},
|
| 5681 |
+
{
|
| 5682 |
+
"epoch": 0.9727136431784108,
|
| 5683 |
+
"grad_norm": 0.29148492217063904,
|
| 5684 |
+
"learning_rate": 2.2210348676977023e-07,
|
| 5685 |
+
"loss": 9.5991,
|
| 5686 |
+
"step": 811
|
| 5687 |
+
},
|
| 5688 |
+
{
|
| 5689 |
+
"epoch": 0.9739130434782609,
|
| 5690 |
+
"grad_norm": 0.30888915061950684,
|
| 5691 |
+
"learning_rate": 2.0238447524372205e-07,
|
| 5692 |
+
"loss": 9.6039,
|
| 5693 |
+
"step": 812
|
| 5694 |
+
},
|
| 5695 |
+
{
|
| 5696 |
+
"epoch": 0.9751124437781109,
|
| 5697 |
+
"grad_norm": 0.2737792730331421,
|
| 5698 |
+
"learning_rate": 1.8358020794843056e-07,
|
| 5699 |
+
"loss": 9.5931,
|
| 5700 |
+
"step": 813
|
| 5701 |
+
},
|
| 5702 |
+
{
|
| 5703 |
+
"epoch": 0.976311844077961,
|
| 5704 |
+
"grad_norm": 0.2680344581604004,
|
| 5705 |
+
"learning_rate": 1.6569103030394938e-07,
|
| 5706 |
+
"loss": 9.5942,
|
| 5707 |
+
"step": 814
|
| 5708 |
+
},
|
| 5709 |
+
{
|
| 5710 |
+
"epoch": 0.9775112443778111,
|
| 5711 |
+
"grad_norm": 0.272061288356781,
|
| 5712 |
+
"learning_rate": 1.48717270920834e-07,
|
| 5713 |
+
"loss": 9.5932,
|
| 5714 |
+
"step": 815
|
| 5715 |
+
},
|
| 5716 |
+
{
|
| 5717 |
+
"epoch": 0.9787106446776612,
|
| 5718 |
+
"grad_norm": 0.273945152759552,
|
| 5719 |
+
"learning_rate": 1.3265924159410192e-07,
|
| 5720 |
+
"loss": 9.5924,
|
| 5721 |
+
"step": 816
|
| 5722 |
+
},
|
| 5723 |
+
{
|
| 5724 |
+
"epoch": 0.9799100449775112,
|
| 5725 |
+
"grad_norm": 0.27954599261283875,
|
| 5726 |
+
"learning_rate": 1.1751723729750974e-07,
|
| 5727 |
+
"loss": 9.5923,
|
| 5728 |
+
"step": 817
|
| 5729 |
+
},
|
| 5730 |
+
{
|
| 5731 |
+
"epoch": 0.9811094452773613,
|
| 5732 |
+
"grad_norm": 0.27977848052978516,
|
| 5733 |
+
"learning_rate": 1.0329153617812947e-07,
|
| 5734 |
+
"loss": 9.5911,
|
| 5735 |
+
"step": 818
|
| 5736 |
+
},
|
| 5737 |
+
{
|
| 5738 |
+
"epoch": 0.9823088455772114,
|
| 5739 |
+
"grad_norm": 0.275766521692276,
|
| 5740 |
+
"learning_rate": 8.998239955124721e-08,
|
| 5741 |
+
"loss": 9.6008,
|
| 5742 |
+
"step": 819
|
| 5743 |
+
},
|
| 5744 |
+
{
|
| 5745 |
+
"epoch": 0.9835082458770614,
|
| 5746 |
+
"grad_norm": 0.2739316523075104,
|
| 5747 |
+
"learning_rate": 7.759007189555579e-08,
|
| 5748 |
+
"loss": 9.6012,
|
| 5749 |
+
"step": 820
|
| 5750 |
+
},
|
| 5751 |
+
{
|
| 5752 |
+
"epoch": 0.9847076461769115,
|
| 5753 |
+
"grad_norm": 0.2775361239910126,
|
| 5754 |
+
"learning_rate": 6.611478084866951e-08,
|
| 5755 |
+
"loss": 9.6039,
|
| 5756 |
+
"step": 821
|
| 5757 |
+
},
|
| 5758 |
+
{
|
| 5759 |
+
"epoch": 0.9859070464767616,
|
| 5760 |
+
"grad_norm": 0.2795866131782532,
|
| 5761 |
+
"learning_rate": 5.555673720292753e-08,
|
| 5762 |
+
"loss": 9.6031,
|
| 5763 |
+
"step": 822
|
| 5764 |
+
},
|
| 5765 |
+
{
|
| 5766 |
+
"epoch": 0.9871064467766117,
|
| 5767 |
+
"grad_norm": 0.29514625668525696,
|
| 5768 |
+
"learning_rate": 4.5916134901552443e-08,
|
| 5769 |
+
"loss": 9.6003,
|
| 5770 |
+
"step": 823
|
| 5771 |
+
},
|
| 5772 |
+
{
|
| 5773 |
+
"epoch": 0.9883058470764617,
|
| 5774 |
+
"grad_norm": 0.29738849401474,
|
| 5775 |
+
"learning_rate": 3.7193151035047616e-08,
|
| 5776 |
+
"loss": 9.6031,
|
| 5777 |
+
"step": 824
|
| 5778 |
+
},
|
| 5779 |
+
{
|
| 5780 |
+
"epoch": 0.9895052473763118,
|
| 5781 |
+
"grad_norm": 0.3434954285621643,
|
| 5782 |
+
"learning_rate": 2.93879458379831e-08,
|
| 5783 |
+
"loss": 9.6087,
|
| 5784 |
+
"step": 825
|
| 5785 |
+
},
|
| 5786 |
+
{
|
| 5787 |
+
"epoch": 0.990704647676162,
|
| 5788 |
+
"grad_norm": 0.26993241906166077,
|
| 5789 |
+
"learning_rate": 2.2500662686025797e-08,
|
| 5790 |
+
"loss": 9.5913,
|
| 5791 |
+
"step": 826
|
| 5792 |
+
},
|
| 5793 |
+
{
|
| 5794 |
+
"epoch": 0.991904047976012,
|
| 5795 |
+
"grad_norm": 0.27204930782318115,
|
| 5796 |
+
"learning_rate": 1.653142809331376e-08,
|
| 5797 |
+
"loss": 9.597,
|
| 5798 |
+
"step": 827
|
| 5799 |
+
},
|
| 5800 |
+
{
|
| 5801 |
+
"epoch": 0.993103448275862,
|
| 5802 |
+
"grad_norm": 0.2699795067310333,
|
| 5803 |
+
"learning_rate": 1.148035171014139e-08,
|
| 5804 |
+
"loss": 9.5989,
|
| 5805 |
+
"step": 828
|
| 5806 |
+
},
|
| 5807 |
+
{
|
| 5808 |
+
"epoch": 0.9943028485757122,
|
| 5809 |
+
"grad_norm": 0.2756272852420807,
|
| 5810 |
+
"learning_rate": 7.347526320927723e-09,
|
| 5811 |
+
"loss": 9.6,
|
| 5812 |
+
"step": 829
|
| 5813 |
+
},
|
| 5814 |
+
{
|
| 5815 |
+
"epoch": 0.9955022488755623,
|
| 5816 |
+
"grad_norm": 0.2937001585960388,
|
| 5817 |
+
"learning_rate": 4.133027842517789e-09,
|
| 5818 |
+
"loss": 9.5987,
|
| 5819 |
+
"step": 830
|
| 5820 |
+
},
|
| 5821 |
+
{
|
| 5822 |
+
"epoch": 0.9967016491754123,
|
| 5823 |
+
"grad_norm": 0.2777741551399231,
|
| 5824 |
+
"learning_rate": 1.8369153228114944e-09,
|
| 5825 |
+
"loss": 9.5984,
|
| 5826 |
+
"step": 831
|
| 5827 |
+
},
|
| 5828 |
+
{
|
| 5829 |
+
"epoch": 0.9979010494752624,
|
| 5830 |
+
"grad_norm": 0.277544766664505,
|
| 5831 |
+
"learning_rate": 4.5923093963118335e-10,
|
| 5832 |
+
"loss": 9.6021,
|
| 5833 |
+
"step": 832
|
| 5834 |
+
},
|
| 5835 |
+
{
|
| 5836 |
+
"epoch": 0.9991004497751125,
|
| 5837 |
+
"grad_norm": 0.2950143814086914,
|
| 5838 |
+
"learning_rate": 0.0,
|
| 5839 |
+
"loss": 9.6099,
|
| 5840 |
+
"step": 833
|
| 5841 |
}
|
| 5842 |
],
|
| 5843 |
"logging_steps": 1,
|
|
|
|
| 5852 |
"should_evaluate": false,
|
| 5853 |
"should_log": false,
|
| 5854 |
"should_save": true,
|
| 5855 |
+
"should_training_stop": true
|
| 5856 |
},
|
| 5857 |
"attributes": {}
|
| 5858 |
}
|
| 5859 |
},
|
| 5860 |
+
"total_flos": 691668038713344.0,
|
| 5861 |
"train_batch_size": 4,
|
| 5862 |
"trial_name": null,
|
| 5863 |
"trial_params": null
|