Training in progress, step 600, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/global_step600/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step600/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step600/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step600/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step600/zero_pp_rank_0_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step600/zero_pp_rank_1_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step600/zero_pp_rank_2_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step600/zero_pp_rank_3_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/latest +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 +739 -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 13648688
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cb3b644050ed772b7473bed1a7223464b5adddad4a96b1e2f174c7375d90aca
|
| 3 |
size 13648688
|
last-checkpoint/global_step600/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b51418538923e000a04e122aacdf38863f1cc95b1b5aa98087f496f0d7044dae
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step600/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b66245cd5b6d4c2a8021a53283cb83bf17039956cd2564e1c41ebbe6a564aeb
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step600/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b98f5bd9aca691723e9fa9e43e7b3af2aaf062dd1ec959d45bb4c2d452687bb3
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step600/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd5244ef133205afa5b1d085e6bc10f6ba0c7ca412c9ecdd7f3f710dbf061ae0
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step600/zero_pp_rank_0_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e82f6e9732cda4940c2abbacb0a45e9c415e0b0fd09da312f3042639f627844
|
| 3 |
+
size 152238
|
last-checkpoint/global_step600/zero_pp_rank_1_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7a339b14b51cc845d9a3ad1ec9d75ba91326e5cc1a2886129e00d9d11b19c3d
|
| 3 |
+
size 152238
|
last-checkpoint/global_step600/zero_pp_rank_2_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f93a6017dd35b08b458e451c19feacacb170cf5d51d48767bfe99722b85f23fa
|
| 3 |
+
size 152238
|
last-checkpoint/global_step600/zero_pp_rank_3_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7786b84d6b617cc1aa9783a8278e4f356955c62a66685840d33705bbfb390923
|
| 3 |
+
size 152238
|
last-checkpoint/latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 1 |
+
global_step600
|
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:b891feb40e4494a2f4339c4a6c2396fe8789003482bccb878b16a84fd49972d2
|
| 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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58b2bf57c8acfc6560630987a9b234d67d256e62c79fb6301ceb72e476851c06
|
| 3 |
size 14960
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:714ec8d81c6c369124166420e509178826aa2a10b37b4d55bda2151ad2f6106f
|
| 3 |
size 14960
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:073d7e3faa703403bd3d6a14715495db3bc0ef77ab8523513ff8bd83de272df4
|
| 3 |
size 14960
|
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:a56472912fbe406df77e6f0ecfb06e43ede87be214ec32eebe03c6969c7328f2
|
| 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": 40,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3663,6 +3663,742 @@
|
|
| 3663 |
"learning_rate": 7.762821161114834e-06,
|
| 3664 |
"loss": 0.1767,
|
| 3665 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3666 |
}
|
| 3667 |
],
|
| 3668 |
"logging_steps": 1,
|
|
@@ -3682,7 +4418,7 @@
|
|
| 3682 |
"attributes": {}
|
| 3683 |
}
|
| 3684 |
},
|
| 3685 |
-
"total_flos":
|
| 3686 |
"train_batch_size": 4,
|
| 3687 |
"trial_name": null,
|
| 3688 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7377805102981863,
|
| 5 |
"eval_steps": 40,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3663 |
"learning_rate": 7.762821161114834e-06,
|
| 3664 |
"loss": 0.1767,
|
| 3665 |
"step": 500
|
| 3666 |
+
},
|
| 3667 |
+
{
|
| 3668 |
+
"epoch": 0.6160467260989856,
|
| 3669 |
+
"grad_norm": 0.49958166386531483,
|
| 3670 |
+
"learning_rate": 7.720954595058118e-06,
|
| 3671 |
+
"loss": 0.1951,
|
| 3672 |
+
"step": 501
|
| 3673 |
+
},
|
| 3674 |
+
{
|
| 3675 |
+
"epoch": 0.6172763602828159,
|
| 3676 |
+
"grad_norm": 0.3970045573640798,
|
| 3677 |
+
"learning_rate": 7.679130122720704e-06,
|
| 3678 |
+
"loss": 0.1961,
|
| 3679 |
+
"step": 502
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 0.6185059944666462,
|
| 3683 |
+
"grad_norm": 0.4623993081995553,
|
| 3684 |
+
"learning_rate": 7.637348516596016e-06,
|
| 3685 |
+
"loss": 0.2396,
|
| 3686 |
+
"step": 503
|
| 3687 |
+
},
|
| 3688 |
+
{
|
| 3689 |
+
"epoch": 0.6197356286504765,
|
| 3690 |
+
"grad_norm": 0.3615937641177032,
|
| 3691 |
+
"learning_rate": 7.59561054838575e-06,
|
| 3692 |
+
"loss": 0.1549,
|
| 3693 |
+
"step": 504
|
| 3694 |
+
},
|
| 3695 |
+
{
|
| 3696 |
+
"epoch": 0.6209652628343068,
|
| 3697 |
+
"grad_norm": 0.4101127611696136,
|
| 3698 |
+
"learning_rate": 7.5539169889856135e-06,
|
| 3699 |
+
"loss": 0.211,
|
| 3700 |
+
"step": 505
|
| 3701 |
+
},
|
| 3702 |
+
{
|
| 3703 |
+
"epoch": 0.6221948970181371,
|
| 3704 |
+
"grad_norm": 0.42525165037118207,
|
| 3705 |
+
"learning_rate": 7.512268608471083e-06,
|
| 3706 |
+
"loss": 0.1805,
|
| 3707 |
+
"step": 506
|
| 3708 |
+
},
|
| 3709 |
+
{
|
| 3710 |
+
"epoch": 0.6234245312019674,
|
| 3711 |
+
"grad_norm": 0.37275006951534,
|
| 3712 |
+
"learning_rate": 7.470666176083193e-06,
|
| 3713 |
+
"loss": 0.1708,
|
| 3714 |
+
"step": 507
|
| 3715 |
+
},
|
| 3716 |
+
{
|
| 3717 |
+
"epoch": 0.6246541653857978,
|
| 3718 |
+
"grad_norm": 0.3457729667153079,
|
| 3719 |
+
"learning_rate": 7.42911046021432e-06,
|
| 3720 |
+
"loss": 0.1653,
|
| 3721 |
+
"step": 508
|
| 3722 |
+
},
|
| 3723 |
+
{
|
| 3724 |
+
"epoch": 0.625883799569628,
|
| 3725 |
+
"grad_norm": 0.3278339387702758,
|
| 3726 |
+
"learning_rate": 7.387602228393987e-06,
|
| 3727 |
+
"loss": 0.2193,
|
| 3728 |
+
"step": 509
|
| 3729 |
+
},
|
| 3730 |
+
{
|
| 3731 |
+
"epoch": 0.6271134337534583,
|
| 3732 |
+
"grad_norm": 0.4458446440488868,
|
| 3733 |
+
"learning_rate": 7.346142247274695e-06,
|
| 3734 |
+
"loss": 0.1969,
|
| 3735 |
+
"step": 510
|
| 3736 |
+
},
|
| 3737 |
+
{
|
| 3738 |
+
"epoch": 0.6283430679372887,
|
| 3739 |
+
"grad_norm": 0.39189734844551366,
|
| 3740 |
+
"learning_rate": 7.304731282617762e-06,
|
| 3741 |
+
"loss": 0.2064,
|
| 3742 |
+
"step": 511
|
| 3743 |
+
},
|
| 3744 |
+
{
|
| 3745 |
+
"epoch": 0.6295727021211189,
|
| 3746 |
+
"grad_norm": 0.5420020058867324,
|
| 3747 |
+
"learning_rate": 7.263370099279173e-06,
|
| 3748 |
+
"loss": 0.1818,
|
| 3749 |
+
"step": 512
|
| 3750 |
+
},
|
| 3751 |
+
{
|
| 3752 |
+
"epoch": 0.6308023363049493,
|
| 3753 |
+
"grad_norm": 0.29997454975691457,
|
| 3754 |
+
"learning_rate": 7.2220594611954606e-06,
|
| 3755 |
+
"loss": 0.1301,
|
| 3756 |
+
"step": 513
|
| 3757 |
+
},
|
| 3758 |
+
{
|
| 3759 |
+
"epoch": 0.6320319704887796,
|
| 3760 |
+
"grad_norm": 0.5869237816795814,
|
| 3761 |
+
"learning_rate": 7.1808001313695855e-06,
|
| 3762 |
+
"loss": 0.2408,
|
| 3763 |
+
"step": 514
|
| 3764 |
+
},
|
| 3765 |
+
{
|
| 3766 |
+
"epoch": 0.63326160467261,
|
| 3767 |
+
"grad_norm": 0.4123986176767839,
|
| 3768 |
+
"learning_rate": 7.1395928718568605e-06,
|
| 3769 |
+
"loss": 0.1633,
|
| 3770 |
+
"step": 515
|
| 3771 |
+
},
|
| 3772 |
+
{
|
| 3773 |
+
"epoch": 0.6344912388564402,
|
| 3774 |
+
"grad_norm": 0.33349921575102887,
|
| 3775 |
+
"learning_rate": 7.098438443750865e-06,
|
| 3776 |
+
"loss": 0.1433,
|
| 3777 |
+
"step": 516
|
| 3778 |
+
},
|
| 3779 |
+
{
|
| 3780 |
+
"epoch": 0.6357208730402705,
|
| 3781 |
+
"grad_norm": 0.38968457601021717,
|
| 3782 |
+
"learning_rate": 7.057337607169373e-06,
|
| 3783 |
+
"loss": 0.1385,
|
| 3784 |
+
"step": 517
|
| 3785 |
+
},
|
| 3786 |
+
{
|
| 3787 |
+
"epoch": 0.6369505072241008,
|
| 3788 |
+
"grad_norm": 0.4844956623703022,
|
| 3789 |
+
"learning_rate": 7.016291121240346e-06,
|
| 3790 |
+
"loss": 0.2059,
|
| 3791 |
+
"step": 518
|
| 3792 |
+
},
|
| 3793 |
+
{
|
| 3794 |
+
"epoch": 0.6381801414079311,
|
| 3795 |
+
"grad_norm": 0.5044663407494435,
|
| 3796 |
+
"learning_rate": 6.975299744087891e-06,
|
| 3797 |
+
"loss": 0.1627,
|
| 3798 |
+
"step": 519
|
| 3799 |
+
},
|
| 3800 |
+
{
|
| 3801 |
+
"epoch": 0.6394097755917615,
|
| 3802 |
+
"grad_norm": 0.38557915218071376,
|
| 3803 |
+
"learning_rate": 6.934364232818254e-06,
|
| 3804 |
+
"loss": 0.2096,
|
| 3805 |
+
"step": 520
|
| 3806 |
+
},
|
| 3807 |
+
{
|
| 3808 |
+
"epoch": 0.6394097755917615,
|
| 3809 |
+
"eval_accuracy": 0.7967914438502673,
|
| 3810 |
+
"eval_f1": 0.4864864864864865,
|
| 3811 |
+
"eval_loss": 0.4250781238079071,
|
| 3812 |
+
"eval_precision": 0.75,
|
| 3813 |
+
"eval_recall": 0.36,
|
| 3814 |
+
"eval_runtime": 22.9637,
|
| 3815 |
+
"eval_samples_per_second": 2.177,
|
| 3816 |
+
"eval_steps_per_second": 0.174,
|
| 3817 |
+
"step": 520
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.6406394097755918,
|
| 3821 |
+
"grad_norm": 0.5448951647380567,
|
| 3822 |
+
"learning_rate": 6.8934853435058566e-06,
|
| 3823 |
+
"loss": 0.1752,
|
| 3824 |
+
"step": 521
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.6418690439594221,
|
| 3828 |
+
"grad_norm": 0.4644280248680503,
|
| 3829 |
+
"learning_rate": 6.852663831179303e-06,
|
| 3830 |
+
"loss": 0.207,
|
| 3831 |
+
"step": 522
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.6430986781432524,
|
| 3835 |
+
"grad_norm": 0.4087824167359295,
|
| 3836 |
+
"learning_rate": 6.811900449807465e-06,
|
| 3837 |
+
"loss": 0.2231,
|
| 3838 |
+
"step": 523
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.6443283123270827,
|
| 3842 |
+
"grad_norm": 0.3257107890885892,
|
| 3843 |
+
"learning_rate": 6.771195952285541e-06,
|
| 3844 |
+
"loss": 0.2224,
|
| 3845 |
+
"step": 524
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.645557946510913,
|
| 3849 |
+
"grad_norm": 0.3772013974334331,
|
| 3850 |
+
"learning_rate": 6.730551090421137e-06,
|
| 3851 |
+
"loss": 0.2488,
|
| 3852 |
+
"step": 525
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.6467875806947433,
|
| 3856 |
+
"grad_norm": 0.36235955730436165,
|
| 3857 |
+
"learning_rate": 6.689966614920414e-06,
|
| 3858 |
+
"loss": 0.243,
|
| 3859 |
+
"step": 526
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.6480172148785737,
|
| 3863 |
+
"grad_norm": 0.3264893573428842,
|
| 3864 |
+
"learning_rate": 6.6494432753741935e-06,
|
| 3865 |
+
"loss": 0.1759,
|
| 3866 |
+
"step": 527
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.6492468490624039,
|
| 3870 |
+
"grad_norm": 0.5043057888131116,
|
| 3871 |
+
"learning_rate": 6.608981820244116e-06,
|
| 3872 |
+
"loss": 0.2345,
|
| 3873 |
+
"step": 528
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.6504764832462342,
|
| 3877 |
+
"grad_norm": 0.3463558045689586,
|
| 3878 |
+
"learning_rate": 6.568582996848844e-06,
|
| 3879 |
+
"loss": 0.1618,
|
| 3880 |
+
"step": 529
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 0.6517061174300646,
|
| 3884 |
+
"grad_norm": 0.38166358521794713,
|
| 3885 |
+
"learning_rate": 6.528247551350213e-06,
|
| 3886 |
+
"loss": 0.1684,
|
| 3887 |
+
"step": 530
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 0.6529357516138948,
|
| 3891 |
+
"grad_norm": 0.3156591052948769,
|
| 3892 |
+
"learning_rate": 6.487976228739493e-06,
|
| 3893 |
+
"loss": 0.1579,
|
| 3894 |
+
"step": 531
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.6541653857977252,
|
| 3898 |
+
"grad_norm": 0.3332487495737034,
|
| 3899 |
+
"learning_rate": 6.4477697728236146e-06,
|
| 3900 |
+
"loss": 0.1551,
|
| 3901 |
+
"step": 532
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 0.6553950199815555,
|
| 3905 |
+
"grad_norm": 0.4211650929646287,
|
| 3906 |
+
"learning_rate": 6.407628926211409e-06,
|
| 3907 |
+
"loss": 0.2312,
|
| 3908 |
+
"step": 533
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 0.6566246541653858,
|
| 3912 |
+
"grad_norm": 0.2935734822708652,
|
| 3913 |
+
"learning_rate": 6.367554430299924e-06,
|
| 3914 |
+
"loss": 0.115,
|
| 3915 |
+
"step": 534
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 0.6578542883492161,
|
| 3919 |
+
"grad_norm": 0.3109438577457444,
|
| 3920 |
+
"learning_rate": 6.327547025260723e-06,
|
| 3921 |
+
"loss": 0.1732,
|
| 3922 |
+
"step": 535
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 0.6590839225330464,
|
| 3926 |
+
"grad_norm": 0.5251627863608039,
|
| 3927 |
+
"learning_rate": 6.287607450026189e-06,
|
| 3928 |
+
"loss": 0.2232,
|
| 3929 |
+
"step": 536
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 0.6603135567168767,
|
| 3933 |
+
"grad_norm": 0.5272463447754631,
|
| 3934 |
+
"learning_rate": 6.247736442275918e-06,
|
| 3935 |
+
"loss": 0.225,
|
| 3936 |
+
"step": 537
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.661543190900707,
|
| 3940 |
+
"grad_norm": 0.49530017803033505,
|
| 3941 |
+
"learning_rate": 6.2079347384230505e-06,
|
| 3942 |
+
"loss": 0.2101,
|
| 3943 |
+
"step": 538
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 0.6627728250845374,
|
| 3947 |
+
"grad_norm": 0.36399232198883885,
|
| 3948 |
+
"learning_rate": 6.168203073600706e-06,
|
| 3949 |
+
"loss": 0.2412,
|
| 3950 |
+
"step": 539
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 0.6640024592683677,
|
| 3954 |
+
"grad_norm": 0.3761278210204055,
|
| 3955 |
+
"learning_rate": 6.128542181648395e-06,
|
| 3956 |
+
"loss": 0.1548,
|
| 3957 |
+
"step": 540
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 0.665232093452198,
|
| 3961 |
+
"grad_norm": 0.3600939725825308,
|
| 3962 |
+
"learning_rate": 6.088952795098442e-06,
|
| 3963 |
+
"loss": 0.1784,
|
| 3964 |
+
"step": 541
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 0.6664617276360283,
|
| 3968 |
+
"grad_norm": 0.27977025283181606,
|
| 3969 |
+
"learning_rate": 6.049435645162487e-06,
|
| 3970 |
+
"loss": 0.1274,
|
| 3971 |
+
"step": 542
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 0.6676913618198586,
|
| 3975 |
+
"grad_norm": 0.42480514662361724,
|
| 3976 |
+
"learning_rate": 6.009991461717977e-06,
|
| 3977 |
+
"loss": 0.1807,
|
| 3978 |
+
"step": 543
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 0.6689209960036889,
|
| 3982 |
+
"grad_norm": 0.48627506473268806,
|
| 3983 |
+
"learning_rate": 5.9706209732946495e-06,
|
| 3984 |
+
"loss": 0.2132,
|
| 3985 |
+
"step": 544
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 0.6701506301875192,
|
| 3989 |
+
"grad_norm": 0.42975103927248237,
|
| 3990 |
+
"learning_rate": 5.931324907061131e-06,
|
| 3991 |
+
"loss": 0.1755,
|
| 3992 |
+
"step": 545
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 0.6713802643713496,
|
| 3996 |
+
"grad_norm": 0.4272914218675845,
|
| 3997 |
+
"learning_rate": 5.892103988811457e-06,
|
| 3998 |
+
"loss": 0.2238,
|
| 3999 |
+
"step": 546
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 0.6726098985551798,
|
| 4003 |
+
"grad_norm": 0.35364469492194867,
|
| 4004 |
+
"learning_rate": 5.852958942951701e-06,
|
| 4005 |
+
"loss": 0.1657,
|
| 4006 |
+
"step": 547
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 0.6738395327390101,
|
| 4010 |
+
"grad_norm": 0.3205399017187773,
|
| 4011 |
+
"learning_rate": 5.8138904924865766e-06,
|
| 4012 |
+
"loss": 0.1284,
|
| 4013 |
+
"step": 548
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 0.6750691669228405,
|
| 4017 |
+
"grad_norm": 0.488644078540224,
|
| 4018 |
+
"learning_rate": 5.774899359006092e-06,
|
| 4019 |
+
"loss": 0.2515,
|
| 4020 |
+
"step": 549
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 0.6762988011066707,
|
| 4024 |
+
"grad_norm": 0.3649190122100428,
|
| 4025 |
+
"learning_rate": 5.735986262672211e-06,
|
| 4026 |
+
"loss": 0.1812,
|
| 4027 |
+
"step": 550
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 0.6775284352905011,
|
| 4031 |
+
"grad_norm": 0.38288184293561706,
|
| 4032 |
+
"learning_rate": 5.697151922205575e-06,
|
| 4033 |
+
"loss": 0.2015,
|
| 4034 |
+
"step": 551
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 0.6787580694743314,
|
| 4038 |
+
"grad_norm": 0.4564212700510477,
|
| 4039 |
+
"learning_rate": 5.658397054872197e-06,
|
| 4040 |
+
"loss": 0.1518,
|
| 4041 |
+
"step": 552
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 0.6799877036581617,
|
| 4045 |
+
"grad_norm": 0.6778561540995489,
|
| 4046 |
+
"learning_rate": 5.619722376470238e-06,
|
| 4047 |
+
"loss": 0.2669,
|
| 4048 |
+
"step": 553
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 0.681217337841992,
|
| 4052 |
+
"grad_norm": 0.38820523141101543,
|
| 4053 |
+
"learning_rate": 5.581128601316774e-06,
|
| 4054 |
+
"loss": 0.1773,
|
| 4055 |
+
"step": 554
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.6824469720258223,
|
| 4059 |
+
"grad_norm": 0.47361205634888237,
|
| 4060 |
+
"learning_rate": 5.542616442234618e-06,
|
| 4061 |
+
"loss": 0.1582,
|
| 4062 |
+
"step": 555
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 0.6836766062096526,
|
| 4066 |
+
"grad_norm": 0.4441223497761333,
|
| 4067 |
+
"learning_rate": 5.504186610539131e-06,
|
| 4068 |
+
"loss": 0.2571,
|
| 4069 |
+
"step": 556
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 0.6849062403934829,
|
| 4073 |
+
"grad_norm": 0.34475269069193953,
|
| 4074 |
+
"learning_rate": 5.465839816025093e-06,
|
| 4075 |
+
"loss": 0.1777,
|
| 4076 |
+
"step": 557
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 0.6861358745773133,
|
| 4080 |
+
"grad_norm": 0.5033044850685292,
|
| 4081 |
+
"learning_rate": 5.427576766953615e-06,
|
| 4082 |
+
"loss": 0.2281,
|
| 4083 |
+
"step": 558
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.6873655087611436,
|
| 4087 |
+
"grad_norm": 0.5766890545879761,
|
| 4088 |
+
"learning_rate": 5.3893981700390215e-06,
|
| 4089 |
+
"loss": 0.1924,
|
| 4090 |
+
"step": 559
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.6885951429449738,
|
| 4094 |
+
"grad_norm": 0.41086152620647287,
|
| 4095 |
+
"learning_rate": 5.35130473043582e-06,
|
| 4096 |
+
"loss": 0.204,
|
| 4097 |
+
"step": 560
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.6885951429449738,
|
| 4101 |
+
"eval_accuracy": 0.7914438502673797,
|
| 4102 |
+
"eval_f1": 0.4657534246575342,
|
| 4103 |
+
"eval_loss": 0.44132813811302185,
|
| 4104 |
+
"eval_precision": 0.7391304347826086,
|
| 4105 |
+
"eval_recall": 0.34,
|
| 4106 |
+
"eval_runtime": 22.903,
|
| 4107 |
+
"eval_samples_per_second": 2.183,
|
| 4108 |
+
"eval_steps_per_second": 0.175,
|
| 4109 |
+
"step": 560
|
| 4110 |
+
},
|
| 4111 |
+
{
|
| 4112 |
+
"epoch": 0.6898247771288042,
|
| 4113 |
+
"grad_norm": 0.3726661976685923,
|
| 4114 |
+
"learning_rate": 5.313297151725679e-06,
|
| 4115 |
+
"loss": 0.1644,
|
| 4116 |
+
"step": 561
|
| 4117 |
+
},
|
| 4118 |
+
{
|
| 4119 |
+
"epoch": 0.6910544113126345,
|
| 4120 |
+
"grad_norm": 0.5219762547061035,
|
| 4121 |
+
"learning_rate": 5.275376135904408e-06,
|
| 4122 |
+
"loss": 0.2288,
|
| 4123 |
+
"step": 562
|
| 4124 |
+
},
|
| 4125 |
+
{
|
| 4126 |
+
"epoch": 0.6922840454964648,
|
| 4127 |
+
"grad_norm": 0.5504038731015495,
|
| 4128 |
+
"learning_rate": 5.2375423833690255e-06,
|
| 4129 |
+
"loss": 0.2808,
|
| 4130 |
+
"step": 563
|
| 4131 |
+
},
|
| 4132 |
+
{
|
| 4133 |
+
"epoch": 0.6935136796802951,
|
| 4134 |
+
"grad_norm": 0.5113724900506852,
|
| 4135 |
+
"learning_rate": 5.1997965929048125e-06,
|
| 4136 |
+
"loss": 0.24,
|
| 4137 |
+
"step": 564
|
| 4138 |
+
},
|
| 4139 |
+
{
|
| 4140 |
+
"epoch": 0.6947433138641255,
|
| 4141 |
+
"grad_norm": 0.3227481790414618,
|
| 4142 |
+
"learning_rate": 5.1621394616723705e-06,
|
| 4143 |
+
"loss": 0.1381,
|
| 4144 |
+
"step": 565
|
| 4145 |
+
},
|
| 4146 |
+
{
|
| 4147 |
+
"epoch": 0.6959729480479557,
|
| 4148 |
+
"grad_norm": 0.4136723306908959,
|
| 4149 |
+
"learning_rate": 5.124571685194804e-06,
|
| 4150 |
+
"loss": 0.1984,
|
| 4151 |
+
"step": 566
|
| 4152 |
+
},
|
| 4153 |
+
{
|
| 4154 |
+
"epoch": 0.697202582231786,
|
| 4155 |
+
"grad_norm": 0.366222168153406,
|
| 4156 |
+
"learning_rate": 5.087093957344841e-06,
|
| 4157 |
+
"loss": 0.1981,
|
| 4158 |
+
"step": 567
|
| 4159 |
+
},
|
| 4160 |
+
{
|
| 4161 |
+
"epoch": 0.6984322164156164,
|
| 4162 |
+
"grad_norm": 0.45034331558707047,
|
| 4163 |
+
"learning_rate": 5.049706970332e-06,
|
| 4164 |
+
"loss": 0.1971,
|
| 4165 |
+
"step": 568
|
| 4166 |
+
},
|
| 4167 |
+
{
|
| 4168 |
+
"epoch": 0.6996618505994466,
|
| 4169 |
+
"grad_norm": 0.4100657746036488,
|
| 4170 |
+
"learning_rate": 5.0124114146898505e-06,
|
| 4171 |
+
"loss": 0.2322,
|
| 4172 |
+
"step": 569
|
| 4173 |
+
},
|
| 4174 |
+
{
|
| 4175 |
+
"epoch": 0.700891484783277,
|
| 4176 |
+
"grad_norm": 0.44211705459477657,
|
| 4177 |
+
"learning_rate": 4.9752079792632244e-06,
|
| 4178 |
+
"loss": 0.1571,
|
| 4179 |
+
"step": 570
|
| 4180 |
+
},
|
| 4181 |
+
{
|
| 4182 |
+
"epoch": 0.7021211189671073,
|
| 4183 |
+
"grad_norm": 0.5266522352800783,
|
| 4184 |
+
"learning_rate": 4.938097351195499e-06,
|
| 4185 |
+
"loss": 0.1818,
|
| 4186 |
+
"step": 571
|
| 4187 |
+
},
|
| 4188 |
+
{
|
| 4189 |
+
"epoch": 0.7033507531509376,
|
| 4190 |
+
"grad_norm": 0.598616618137633,
|
| 4191 |
+
"learning_rate": 4.9010802159159224e-06,
|
| 4192 |
+
"loss": 0.2259,
|
| 4193 |
+
"step": 572
|
| 4194 |
+
},
|
| 4195 |
+
{
|
| 4196 |
+
"epoch": 0.7045803873347679,
|
| 4197 |
+
"grad_norm": 0.5779733430038166,
|
| 4198 |
+
"learning_rate": 4.864157257126928e-06,
|
| 4199 |
+
"loss": 0.1429,
|
| 4200 |
+
"step": 573
|
| 4201 |
+
},
|
| 4202 |
+
{
|
| 4203 |
+
"epoch": 0.7058100215185982,
|
| 4204 |
+
"grad_norm": 0.34649772674980184,
|
| 4205 |
+
"learning_rate": 4.8273291567915225e-06,
|
| 4206 |
+
"loss": 0.16,
|
| 4207 |
+
"step": 574
|
| 4208 |
+
},
|
| 4209 |
+
{
|
| 4210 |
+
"epoch": 0.7070396557024285,
|
| 4211 |
+
"grad_norm": 0.34316766175643343,
|
| 4212 |
+
"learning_rate": 4.790596595120699e-06,
|
| 4213 |
+
"loss": 0.1754,
|
| 4214 |
+
"step": 575
|
| 4215 |
+
},
|
| 4216 |
+
{
|
| 4217 |
+
"epoch": 0.7082692898862588,
|
| 4218 |
+
"grad_norm": 0.522847657230503,
|
| 4219 |
+
"learning_rate": 4.753960250560843e-06,
|
| 4220 |
+
"loss": 0.1942,
|
| 4221 |
+
"step": 576
|
| 4222 |
+
},
|
| 4223 |
+
{
|
| 4224 |
+
"epoch": 0.7094989240700892,
|
| 4225 |
+
"grad_norm": 0.3962657447065765,
|
| 4226 |
+
"learning_rate": 4.7174207997812436e-06,
|
| 4227 |
+
"loss": 0.2201,
|
| 4228 |
+
"step": 577
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 0.7107285582539195,
|
| 4232 |
+
"grad_norm": 0.5773303103813715,
|
| 4233 |
+
"learning_rate": 4.680978917661544e-06,
|
| 4234 |
+
"loss": 0.1883,
|
| 4235 |
+
"step": 578
|
| 4236 |
+
},
|
| 4237 |
+
{
|
| 4238 |
+
"epoch": 0.7119581924377497,
|
| 4239 |
+
"grad_norm": 0.49036496077969033,
|
| 4240 |
+
"learning_rate": 4.6446352772793256e-06,
|
| 4241 |
+
"loss": 0.2064,
|
| 4242 |
+
"step": 579
|
| 4243 |
+
},
|
| 4244 |
+
{
|
| 4245 |
+
"epoch": 0.7131878266215801,
|
| 4246 |
+
"grad_norm": 0.6072300426360661,
|
| 4247 |
+
"learning_rate": 4.608390549897661e-06,
|
| 4248 |
+
"loss": 0.2741,
|
| 4249 |
+
"step": 580
|
| 4250 |
+
},
|
| 4251 |
+
{
|
| 4252 |
+
"epoch": 0.7144174608054104,
|
| 4253 |
+
"grad_norm": 0.30903774667825634,
|
| 4254 |
+
"learning_rate": 4.5722454049526825e-06,
|
| 4255 |
+
"loss": 0.1321,
|
| 4256 |
+
"step": 581
|
| 4257 |
+
},
|
| 4258 |
+
{
|
| 4259 |
+
"epoch": 0.7156470949892407,
|
| 4260 |
+
"grad_norm": 0.41353484327403556,
|
| 4261 |
+
"learning_rate": 4.536200510041271e-06,
|
| 4262 |
+
"loss": 0.1589,
|
| 4263 |
+
"step": 582
|
| 4264 |
+
},
|
| 4265 |
+
{
|
| 4266 |
+
"epoch": 0.716876729173071,
|
| 4267 |
+
"grad_norm": 0.3781198095284841,
|
| 4268 |
+
"learning_rate": 4.5002565309087e-06,
|
| 4269 |
+
"loss": 0.1996,
|
| 4270 |
+
"step": 583
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"epoch": 0.7181063633569014,
|
| 4274 |
+
"grad_norm": 0.39752189123610254,
|
| 4275 |
+
"learning_rate": 4.4644141314363165e-06,
|
| 4276 |
+
"loss": 0.1437,
|
| 4277 |
+
"step": 584
|
| 4278 |
+
},
|
| 4279 |
+
{
|
| 4280 |
+
"epoch": 0.7193359975407316,
|
| 4281 |
+
"grad_norm": 0.3854738335798155,
|
| 4282 |
+
"learning_rate": 4.4286739736293285e-06,
|
| 4283 |
+
"loss": 0.1397,
|
| 4284 |
+
"step": 585
|
| 4285 |
+
},
|
| 4286 |
+
{
|
| 4287 |
+
"epoch": 0.7205656317245619,
|
| 4288 |
+
"grad_norm": 0.431269796547053,
|
| 4289 |
+
"learning_rate": 4.393036717604536e-06,
|
| 4290 |
+
"loss": 0.1486,
|
| 4291 |
+
"step": 586
|
| 4292 |
+
},
|
| 4293 |
+
{
|
| 4294 |
+
"epoch": 0.7217952659083923,
|
| 4295 |
+
"grad_norm": 0.49871333432848314,
|
| 4296 |
+
"learning_rate": 4.357503021578158e-06,
|
| 4297 |
+
"loss": 0.1652,
|
| 4298 |
+
"step": 587
|
| 4299 |
+
},
|
| 4300 |
+
{
|
| 4301 |
+
"epoch": 0.7230249000922225,
|
| 4302 |
+
"grad_norm": 0.4737833916251569,
|
| 4303 |
+
"learning_rate": 4.322073541853677e-06,
|
| 4304 |
+
"loss": 0.2164,
|
| 4305 |
+
"step": 588
|
| 4306 |
+
},
|
| 4307 |
+
{
|
| 4308 |
+
"epoch": 0.7242545342760529,
|
| 4309 |
+
"grad_norm": 0.36804258048411215,
|
| 4310 |
+
"learning_rate": 4.286748932809707e-06,
|
| 4311 |
+
"loss": 0.2091,
|
| 4312 |
+
"step": 589
|
| 4313 |
+
},
|
| 4314 |
+
{
|
| 4315 |
+
"epoch": 0.7254841684598832,
|
| 4316 |
+
"grad_norm": 0.3785583210042997,
|
| 4317 |
+
"learning_rate": 4.2515298468879064e-06,
|
| 4318 |
+
"loss": 0.2258,
|
| 4319 |
+
"step": 590
|
| 4320 |
+
},
|
| 4321 |
+
{
|
| 4322 |
+
"epoch": 0.7267138026437135,
|
| 4323 |
+
"grad_norm": 0.3271430669263614,
|
| 4324 |
+
"learning_rate": 4.216416934580947e-06,
|
| 4325 |
+
"loss": 0.1902,
|
| 4326 |
+
"step": 591
|
| 4327 |
+
},
|
| 4328 |
+
{
|
| 4329 |
+
"epoch": 0.7279434368275438,
|
| 4330 |
+
"grad_norm": 0.3725205335835651,
|
| 4331 |
+
"learning_rate": 4.181410844420473e-06,
|
| 4332 |
+
"loss": 0.1469,
|
| 4333 |
+
"step": 592
|
| 4334 |
+
},
|
| 4335 |
+
{
|
| 4336 |
+
"epoch": 0.7291730710113741,
|
| 4337 |
+
"grad_norm": 0.5679833065078629,
|
| 4338 |
+
"learning_rate": 4.146512222965144e-06,
|
| 4339 |
+
"loss": 0.2159,
|
| 4340 |
+
"step": 593
|
| 4341 |
+
},
|
| 4342 |
+
{
|
| 4343 |
+
"epoch": 0.7304027051952044,
|
| 4344 |
+
"grad_norm": 0.6073546495930695,
|
| 4345 |
+
"learning_rate": 4.111721714788671e-06,
|
| 4346 |
+
"loss": 0.1222,
|
| 4347 |
+
"step": 594
|
| 4348 |
+
},
|
| 4349 |
+
{
|
| 4350 |
+
"epoch": 0.7316323393790347,
|
| 4351 |
+
"grad_norm": 0.32166243146523565,
|
| 4352 |
+
"learning_rate": 4.077039962467946e-06,
|
| 4353 |
+
"loss": 0.1224,
|
| 4354 |
+
"step": 595
|
| 4355 |
+
},
|
| 4356 |
+
{
|
| 4357 |
+
"epoch": 0.7328619735628651,
|
| 4358 |
+
"grad_norm": 0.322565114053672,
|
| 4359 |
+
"learning_rate": 4.042467606571134e-06,
|
| 4360 |
+
"loss": 0.1714,
|
| 4361 |
+
"step": 596
|
| 4362 |
+
},
|
| 4363 |
+
{
|
| 4364 |
+
"epoch": 0.7340916077466954,
|
| 4365 |
+
"grad_norm": 0.3390381359494193,
|
| 4366 |
+
"learning_rate": 4.008005285645863e-06,
|
| 4367 |
+
"loss": 0.1638,
|
| 4368 |
+
"step": 597
|
| 4369 |
+
},
|
| 4370 |
+
{
|
| 4371 |
+
"epoch": 0.7353212419305256,
|
| 4372 |
+
"grad_norm": 0.3966925455964005,
|
| 4373 |
+
"learning_rate": 3.973653636207437e-06,
|
| 4374 |
+
"loss": 0.1508,
|
| 4375 |
+
"step": 598
|
| 4376 |
+
},
|
| 4377 |
+
{
|
| 4378 |
+
"epoch": 0.736550876114356,
|
| 4379 |
+
"grad_norm": 0.6143649916164688,
|
| 4380 |
+
"learning_rate": 3.939413292727061e-06,
|
| 4381 |
+
"loss": 0.2691,
|
| 4382 |
+
"step": 599
|
| 4383 |
+
},
|
| 4384 |
+
{
|
| 4385 |
+
"epoch": 0.7377805102981863,
|
| 4386 |
+
"grad_norm": 0.3681398679016445,
|
| 4387 |
+
"learning_rate": 3.9052848876201285e-06,
|
| 4388 |
+
"loss": 0.1545,
|
| 4389 |
+
"step": 600
|
| 4390 |
+
},
|
| 4391 |
+
{
|
| 4392 |
+
"epoch": 0.7377805102981863,
|
| 4393 |
+
"eval_accuracy": 0.7967914438502673,
|
| 4394 |
+
"eval_f1": 0.4864864864864865,
|
| 4395 |
+
"eval_loss": 0.43117186427116394,
|
| 4396 |
+
"eval_precision": 0.75,
|
| 4397 |
+
"eval_recall": 0.36,
|
| 4398 |
+
"eval_runtime": 23.5078,
|
| 4399 |
+
"eval_samples_per_second": 2.127,
|
| 4400 |
+
"eval_steps_per_second": 0.17,
|
| 4401 |
+
"step": 600
|
| 4402 |
}
|
| 4403 |
],
|
| 4404 |
"logging_steps": 1,
|
|
|
|
| 4418 |
"attributes": {}
|
| 4419 |
}
|
| 4420 |
},
|
| 4421 |
+
"total_flos": 588849383505920.0,
|
| 4422 |
"train_batch_size": 4,
|
| 4423 |
"trial_name": null,
|
| 4424 |
"trial_params": null
|