Training in progress, step 800, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/global_step800/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step800/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step800/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step800/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step800/zero_pp_rank_0_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step800/zero_pp_rank_1_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step800/zero_pp_rank_2_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step800/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:a652c5fae11ea2adca853738ad01096d957c3c9179219d103154b7cc54bf3ad3
|
| 3 |
size 13648688
|
last-checkpoint/global_step800/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 sha256:7f877dddd46433a4a3b3134d1597d0e501858c53ef38529470af634bdc6d0189
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step800/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 sha256:18b9212355f3f0b018d7ec9e3ebfd82e6e3d995128421c7d94a2b4c511641a43
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step800/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:afeace15bd4996a7dec661fffefc77e9c270e4d2aeea76a6fb2f7d0689fb6eb5
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step800/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 sha256:4dfd918412cd9bd19e74611b73cd176c7b11f534edfa3c6db14479fe30fcb1d4
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step800/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 sha256:40c25e909a08fdf653f7b456b957ed4d1cda89dfb424aba1b9daf57b7ae1863d
|
| 3 |
+
size 152238
|
last-checkpoint/global_step800/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:c2eb3c53ba5f4d34c3f070d040498034275d76c2a55a2308aa8e3d26e8d57ece
|
| 3 |
+
size 152238
|
last-checkpoint/global_step800/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 sha256:89896592980174dd3b06a9cba6578075ff9f19bc98fcc455acad08cbfcdc859d
|
| 3 |
+
size 152238
|
last-checkpoint/global_step800/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:dd22d993a562918152c931db50c1e239c8ab00cedb5b9ecb46e9ee8a798b231d
|
| 3 |
+
size 152238
|
last-checkpoint/latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 1 |
+
global_step800
|
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 sha256:87cb02e8cd64657a53dc5b5e254ac5f48ae8d194a60165370a55a19b22db6f41
|
| 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:e24d98488b8ee4c9ae33c4ceab244d33a0b8840dd5bba055edd1d451f13ad848
|
| 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 sha256:1d2cd121bedf7b6b76721153b7cef441f373f23cfeb4d690fa2511614d4a9fde
|
| 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:e241a5faad1af0a2c7b86616994b8a556b97dd87a2d34edfc821f52d2592e9ec
|
| 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 sha256:6a7ed681055e4d6c45b04422cd95ffb42b03b2ad7470799d0056fdac3772ffaa
|
| 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,
|
|
@@ -5123,6 +5123,742 @@
|
|
| 5123 |
"learning_rate": 1.1562177270067766e-06,
|
| 5124 |
"loss": 0.2128,
|
| 5125 |
"step": 700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5126 |
}
|
| 5127 |
],
|
| 5128 |
"logging_steps": 1,
|
|
@@ -5142,7 +5878,7 @@
|
|
| 5142 |
"attributes": {}
|
| 5143 |
}
|
| 5144 |
},
|
| 5145 |
-
"total_flos":
|
| 5146 |
"train_batch_size": 4,
|
| 5147 |
"trial_name": null,
|
| 5148 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9837073470642483,
|
| 5 |
"eval_steps": 40,
|
| 6 |
+
"global_step": 800,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5123 |
"learning_rate": 1.1562177270067766e-06,
|
| 5124 |
"loss": 0.2128,
|
| 5125 |
"step": 700
|
| 5126 |
+
},
|
| 5127 |
+
{
|
| 5128 |
+
"epoch": 0.8619735628650477,
|
| 5129 |
+
"grad_norm": 0.4252111135080987,
|
| 5130 |
+
"learning_rate": 1.1362392103604536e-06,
|
| 5131 |
+
"loss": 0.1746,
|
| 5132 |
+
"step": 701
|
| 5133 |
+
},
|
| 5134 |
+
{
|
| 5135 |
+
"epoch": 0.8632031970488779,
|
| 5136 |
+
"grad_norm": 0.45840370226979704,
|
| 5137 |
+
"learning_rate": 1.1164244064062101e-06,
|
| 5138 |
+
"loss": 0.2294,
|
| 5139 |
+
"step": 702
|
| 5140 |
+
},
|
| 5141 |
+
{
|
| 5142 |
+
"epoch": 0.8644328312327083,
|
| 5143 |
+
"grad_norm": 0.4054117908741426,
|
| 5144 |
+
"learning_rate": 1.0967736811213048e-06,
|
| 5145 |
+
"loss": 0.2092,
|
| 5146 |
+
"step": 703
|
| 5147 |
+
},
|
| 5148 |
+
{
|
| 5149 |
+
"epoch": 0.8656624654165386,
|
| 5150 |
+
"grad_norm": 0.39243641295826015,
|
| 5151 |
+
"learning_rate": 1.0772873974524833e-06,
|
| 5152 |
+
"loss": 0.2066,
|
| 5153 |
+
"step": 704
|
| 5154 |
+
},
|
| 5155 |
+
{
|
| 5156 |
+
"epoch": 0.8668920996003688,
|
| 5157 |
+
"grad_norm": 0.354816693403516,
|
| 5158 |
+
"learning_rate": 1.0579659153092759e-06,
|
| 5159 |
+
"loss": 0.1898,
|
| 5160 |
+
"step": 705
|
| 5161 |
+
},
|
| 5162 |
+
{
|
| 5163 |
+
"epoch": 0.8681217337841992,
|
| 5164 |
+
"grad_norm": 0.3898934913509522,
|
| 5165 |
+
"learning_rate": 1.0388095915573427e-06,
|
| 5166 |
+
"loss": 0.1295,
|
| 5167 |
+
"step": 706
|
| 5168 |
+
},
|
| 5169 |
+
{
|
| 5170 |
+
"epoch": 0.8693513679680295,
|
| 5171 |
+
"grad_norm": 0.5938780642245854,
|
| 5172 |
+
"learning_rate": 1.0198187800118842e-06,
|
| 5173 |
+
"loss": 0.2201,
|
| 5174 |
+
"step": 707
|
| 5175 |
+
},
|
| 5176 |
+
{
|
| 5177 |
+
"epoch": 0.8705810021518599,
|
| 5178 |
+
"grad_norm": 0.4475510981424918,
|
| 5179 |
+
"learning_rate": 1.0009938314311186e-06,
|
| 5180 |
+
"loss": 0.1617,
|
| 5181 |
+
"step": 708
|
| 5182 |
+
},
|
| 5183 |
+
{
|
| 5184 |
+
"epoch": 0.8718106363356901,
|
| 5185 |
+
"grad_norm": 0.6094389342778289,
|
| 5186 |
+
"learning_rate": 9.82335093509782e-07,
|
| 5187 |
+
"loss": 0.2703,
|
| 5188 |
+
"step": 709
|
| 5189 |
+
},
|
| 5190 |
+
{
|
| 5191 |
+
"epoch": 0.8730402705195205,
|
| 5192 |
+
"grad_norm": 0.3623911413159473,
|
| 5193 |
+
"learning_rate": 9.638429108727232e-07,
|
| 5194 |
+
"loss": 0.1882,
|
| 5195 |
+
"step": 710
|
| 5196 |
+
},
|
| 5197 |
+
{
|
| 5198 |
+
"epoch": 0.8742699047033508,
|
| 5199 |
+
"grad_norm": 0.4538866026680618,
|
| 5200 |
+
"learning_rate": 9.455176250685338e-07,
|
| 5201 |
+
"loss": 0.2013,
|
| 5202 |
+
"step": 711
|
| 5203 |
+
},
|
| 5204 |
+
{
|
| 5205 |
+
"epoch": 0.875499538887181,
|
| 5206 |
+
"grad_norm": 0.3194360166861715,
|
| 5207 |
+
"learning_rate": 9.27359574563238e-07,
|
| 5208 |
+
"loss": 0.1851,
|
| 5209 |
+
"step": 712
|
| 5210 |
+
},
|
| 5211 |
+
{
|
| 5212 |
+
"epoch": 0.8767291730710114,
|
| 5213 |
+
"grad_norm": 0.38338959194718086,
|
| 5214 |
+
"learning_rate": 9.093690947340406e-07,
|
| 5215 |
+
"loss": 0.1927,
|
| 5216 |
+
"step": 713
|
| 5217 |
+
},
|
| 5218 |
+
{
|
| 5219 |
+
"epoch": 0.8779588072548417,
|
| 5220 |
+
"grad_norm": 0.3879298822682774,
|
| 5221 |
+
"learning_rate": 8.915465178631344e-07,
|
| 5222 |
+
"loss": 0.1238,
|
| 5223 |
+
"step": 714
|
| 5224 |
+
},
|
| 5225 |
+
{
|
| 5226 |
+
"epoch": 0.879188441438672,
|
| 5227 |
+
"grad_norm": 0.34738835843437105,
|
| 5228 |
+
"learning_rate": 8.738921731315686e-07,
|
| 5229 |
+
"loss": 0.1616,
|
| 5230 |
+
"step": 715
|
| 5231 |
+
},
|
| 5232 |
+
{
|
| 5233 |
+
"epoch": 0.8804180756225023,
|
| 5234 |
+
"grad_norm": 0.42388132636624687,
|
| 5235 |
+
"learning_rate": 8.564063866131567e-07,
|
| 5236 |
+
"loss": 0.1793,
|
| 5237 |
+
"step": 716
|
| 5238 |
+
},
|
| 5239 |
+
{
|
| 5240 |
+
"epoch": 0.8816477098063327,
|
| 5241 |
+
"grad_norm": 0.5595604732568557,
|
| 5242 |
+
"learning_rate": 8.390894812684602e-07,
|
| 5243 |
+
"loss": 0.2188,
|
| 5244 |
+
"step": 717
|
| 5245 |
+
},
|
| 5246 |
+
{
|
| 5247 |
+
"epoch": 0.8828773439901629,
|
| 5248 |
+
"grad_norm": 0.4324638530503791,
|
| 5249 |
+
"learning_rate": 8.219417769388316e-07,
|
| 5250 |
+
"loss": 0.1984,
|
| 5251 |
+
"step": 718
|
| 5252 |
+
},
|
| 5253 |
+
{
|
| 5254 |
+
"epoch": 0.8841069781739932,
|
| 5255 |
+
"grad_norm": 0.5485490850760583,
|
| 5256 |
+
"learning_rate": 8.049635903404907e-07,
|
| 5257 |
+
"loss": 0.1978,
|
| 5258 |
+
"step": 719
|
| 5259 |
+
},
|
| 5260 |
+
{
|
| 5261 |
+
"epoch": 0.8853366123578236,
|
| 5262 |
+
"grad_norm": 0.34294438952407474,
|
| 5263 |
+
"learning_rate": 7.881552350586863e-07,
|
| 5264 |
+
"loss": 0.1937,
|
| 5265 |
+
"step": 720
|
| 5266 |
+
},
|
| 5267 |
+
{
|
| 5268 |
+
"epoch": 0.8853366123578236,
|
| 5269 |
+
"eval_accuracy": 0.8021390374331551,
|
| 5270 |
+
"eval_f1": 0.5066666666666667,
|
| 5271 |
+
"eval_loss": 0.42445313930511475,
|
| 5272 |
+
"eval_precision": 0.76,
|
| 5273 |
+
"eval_recall": 0.38,
|
| 5274 |
+
"eval_runtime": 23.5507,
|
| 5275 |
+
"eval_samples_per_second": 2.123,
|
| 5276 |
+
"eval_steps_per_second": 0.17,
|
| 5277 |
+
"step": 720
|
| 5278 |
+
},
|
| 5279 |
+
{
|
| 5280 |
+
"epoch": 0.8865662465416538,
|
| 5281 |
+
"grad_norm": 0.3459770483320799,
|
| 5282 |
+
"learning_rate": 7.715170215419043e-07,
|
| 5283 |
+
"loss": 0.1599,
|
| 5284 |
+
"step": 721
|
| 5285 |
+
},
|
| 5286 |
+
{
|
| 5287 |
+
"epoch": 0.8877958807254842,
|
| 5288 |
+
"grad_norm": 0.3720600631339784,
|
| 5289 |
+
"learning_rate": 7.550492570961243e-07,
|
| 5290 |
+
"loss": 0.1482,
|
| 5291 |
+
"step": 722
|
| 5292 |
+
},
|
| 5293 |
+
{
|
| 5294 |
+
"epoch": 0.8890255149093145,
|
| 5295 |
+
"grad_norm": 0.3994271301853499,
|
| 5296 |
+
"learning_rate": 7.387522458791552e-07,
|
| 5297 |
+
"loss": 0.1792,
|
| 5298 |
+
"step": 723
|
| 5299 |
+
},
|
| 5300 |
+
{
|
| 5301 |
+
"epoch": 0.8902551490931447,
|
| 5302 |
+
"grad_norm": 0.3640144769183845,
|
| 5303 |
+
"learning_rate": 7.226262888950153e-07,
|
| 5304 |
+
"loss": 0.1466,
|
| 5305 |
+
"step": 724
|
| 5306 |
+
},
|
| 5307 |
+
{
|
| 5308 |
+
"epoch": 0.8914847832769751,
|
| 5309 |
+
"grad_norm": 0.37455447234355393,
|
| 5310 |
+
"learning_rate": 7.066716839883592e-07,
|
| 5311 |
+
"loss": 0.1584,
|
| 5312 |
+
"step": 725
|
| 5313 |
+
},
|
| 5314 |
+
{
|
| 5315 |
+
"epoch": 0.8927144174608054,
|
| 5316 |
+
"grad_norm": 0.333702811711787,
|
| 5317 |
+
"learning_rate": 6.908887258389974e-07,
|
| 5318 |
+
"loss": 0.186,
|
| 5319 |
+
"step": 726
|
| 5320 |
+
},
|
| 5321 |
+
{
|
| 5322 |
+
"epoch": 0.8939440516446358,
|
| 5323 |
+
"grad_norm": 0.4224450151535308,
|
| 5324 |
+
"learning_rate": 6.752777059564431e-07,
|
| 5325 |
+
"loss": 0.2272,
|
| 5326 |
+
"step": 727
|
| 5327 |
+
},
|
| 5328 |
+
{
|
| 5329 |
+
"epoch": 0.895173685828466,
|
| 5330 |
+
"grad_norm": 0.3600232430765048,
|
| 5331 |
+
"learning_rate": 6.598389126745209e-07,
|
| 5332 |
+
"loss": 0.2266,
|
| 5333 |
+
"step": 728
|
| 5334 |
+
},
|
| 5335 |
+
{
|
| 5336 |
+
"epoch": 0.8964033200122964,
|
| 5337 |
+
"grad_norm": 0.5012149171996233,
|
| 5338 |
+
"learning_rate": 6.445726311460553e-07,
|
| 5339 |
+
"loss": 0.2236,
|
| 5340 |
+
"step": 729
|
| 5341 |
+
},
|
| 5342 |
+
{
|
| 5343 |
+
"epoch": 0.8976329541961267,
|
| 5344 |
+
"grad_norm": 0.39536858077842535,
|
| 5345 |
+
"learning_rate": 6.29479143337598e-07,
|
| 5346 |
+
"loss": 0.145,
|
| 5347 |
+
"step": 730
|
| 5348 |
+
},
|
| 5349 |
+
{
|
| 5350 |
+
"epoch": 0.8988625883799569,
|
| 5351 |
+
"grad_norm": 0.2991426563682339,
|
| 5352 |
+
"learning_rate": 6.145587280242138e-07,
|
| 5353 |
+
"loss": 0.175,
|
| 5354 |
+
"step": 731
|
| 5355 |
+
},
|
| 5356 |
+
{
|
| 5357 |
+
"epoch": 0.9000922225637873,
|
| 5358 |
+
"grad_norm": 0.3676759279938805,
|
| 5359 |
+
"learning_rate": 5.99811660784344e-07,
|
| 5360 |
+
"loss": 0.1776,
|
| 5361 |
+
"step": 732
|
| 5362 |
+
},
|
| 5363 |
+
{
|
| 5364 |
+
"epoch": 0.9013218567476176,
|
| 5365 |
+
"grad_norm": 0.3708523019855745,
|
| 5366 |
+
"learning_rate": 5.852382139947077e-07,
|
| 5367 |
+
"loss": 0.1736,
|
| 5368 |
+
"step": 733
|
| 5369 |
+
},
|
| 5370 |
+
{
|
| 5371 |
+
"epoch": 0.9025514909314479,
|
| 5372 |
+
"grad_norm": 0.43018990463756845,
|
| 5373 |
+
"learning_rate": 5.708386568252688e-07,
|
| 5374 |
+
"loss": 0.1436,
|
| 5375 |
+
"step": 734
|
| 5376 |
+
},
|
| 5377 |
+
{
|
| 5378 |
+
"epoch": 0.9037811251152782,
|
| 5379 |
+
"grad_norm": 0.4802111636137182,
|
| 5380 |
+
"learning_rate": 5.566132552342784e-07,
|
| 5381 |
+
"loss": 0.2092,
|
| 5382 |
+
"step": 735
|
| 5383 |
+
},
|
| 5384 |
+
{
|
| 5385 |
+
"epoch": 0.9050107592991086,
|
| 5386 |
+
"grad_norm": 0.2782286925340434,
|
| 5387 |
+
"learning_rate": 5.425622719633428e-07,
|
| 5388 |
+
"loss": 0.1204,
|
| 5389 |
+
"step": 736
|
| 5390 |
+
},
|
| 5391 |
+
{
|
| 5392 |
+
"epoch": 0.9062403934829388,
|
| 5393 |
+
"grad_norm": 0.42051906131253536,
|
| 5394 |
+
"learning_rate": 5.286859665325905e-07,
|
| 5395 |
+
"loss": 0.2128,
|
| 5396 |
+
"step": 737
|
| 5397 |
+
},
|
| 5398 |
+
{
|
| 5399 |
+
"epoch": 0.9074700276667691,
|
| 5400 |
+
"grad_norm": 0.5089981907905707,
|
| 5401 |
+
"learning_rate": 5.149845952358589e-07,
|
| 5402 |
+
"loss": 0.2393,
|
| 5403 |
+
"step": 738
|
| 5404 |
+
},
|
| 5405 |
+
{
|
| 5406 |
+
"epoch": 0.9086996618505995,
|
| 5407 |
+
"grad_norm": 0.41726717657714646,
|
| 5408 |
+
"learning_rate": 5.014584111359811e-07,
|
| 5409 |
+
"loss": 0.1784,
|
| 5410 |
+
"step": 739
|
| 5411 |
+
},
|
| 5412 |
+
{
|
| 5413 |
+
"epoch": 0.9099292960344297,
|
| 5414 |
+
"grad_norm": 0.4603202083257241,
|
| 5415 |
+
"learning_rate": 4.881076640600979e-07,
|
| 5416 |
+
"loss": 0.1913,
|
| 5417 |
+
"step": 740
|
| 5418 |
+
},
|
| 5419 |
+
{
|
| 5420 |
+
"epoch": 0.9111589302182601,
|
| 5421 |
+
"grad_norm": 0.4130280700492162,
|
| 5422 |
+
"learning_rate": 4.7493260059504497e-07,
|
| 5423 |
+
"loss": 0.238,
|
| 5424 |
+
"step": 741
|
| 5425 |
+
},
|
| 5426 |
+
{
|
| 5427 |
+
"epoch": 0.9123885644020904,
|
| 5428 |
+
"grad_norm": 0.5036196124164085,
|
| 5429 |
+
"learning_rate": 4.6193346408280216e-07,
|
| 5430 |
+
"loss": 0.1926,
|
| 5431 |
+
"step": 742
|
| 5432 |
+
},
|
| 5433 |
+
{
|
| 5434 |
+
"epoch": 0.9136181985859206,
|
| 5435 |
+
"grad_norm": 0.36883928715030756,
|
| 5436 |
+
"learning_rate": 4.491104946160052e-07,
|
| 5437 |
+
"loss": 0.1659,
|
| 5438 |
+
"step": 743
|
| 5439 |
+
},
|
| 5440 |
+
{
|
| 5441 |
+
"epoch": 0.914847832769751,
|
| 5442 |
+
"grad_norm": 0.4945934881547631,
|
| 5443 |
+
"learning_rate": 4.3646392903348823e-07,
|
| 5444 |
+
"loss": 0.1546,
|
| 5445 |
+
"step": 744
|
| 5446 |
+
},
|
| 5447 |
+
{
|
| 5448 |
+
"epoch": 0.9160774669535813,
|
| 5449 |
+
"grad_norm": 0.33228048042062874,
|
| 5450 |
+
"learning_rate": 4.2399400091594154e-07,
|
| 5451 |
+
"loss": 0.174,
|
| 5452 |
+
"step": 745
|
| 5453 |
+
},
|
| 5454 |
+
{
|
| 5455 |
+
"epoch": 0.9173071011374117,
|
| 5456 |
+
"grad_norm": 0.3446047000937293,
|
| 5457 |
+
"learning_rate": 4.117009405815686e-07,
|
| 5458 |
+
"loss": 0.1422,
|
| 5459 |
+
"step": 746
|
| 5460 |
+
},
|
| 5461 |
+
{
|
| 5462 |
+
"epoch": 0.9185367353212419,
|
| 5463 |
+
"grad_norm": 0.4483830076000804,
|
| 5464 |
+
"learning_rate": 3.9958497508185036e-07,
|
| 5465 |
+
"loss": 0.2321,
|
| 5466 |
+
"step": 747
|
| 5467 |
+
},
|
| 5468 |
+
{
|
| 5469 |
+
"epoch": 0.9197663695050723,
|
| 5470 |
+
"grad_norm": 0.6123522355957743,
|
| 5471 |
+
"learning_rate": 3.8764632819734526e-07,
|
| 5472 |
+
"loss": 0.2343,
|
| 5473 |
+
"step": 748
|
| 5474 |
+
},
|
| 5475 |
+
{
|
| 5476 |
+
"epoch": 0.9209960036889026,
|
| 5477 |
+
"grad_norm": 0.4578579557499873,
|
| 5478 |
+
"learning_rate": 3.758852204335539e-07,
|
| 5479 |
+
"loss": 0.2546,
|
| 5480 |
+
"step": 749
|
| 5481 |
+
},
|
| 5482 |
+
{
|
| 5483 |
+
"epoch": 0.9222256378727328,
|
| 5484 |
+
"grad_norm": 0.5396510643509638,
|
| 5485 |
+
"learning_rate": 3.643018690168487e-07,
|
| 5486 |
+
"loss": 0.2142,
|
| 5487 |
+
"step": 750
|
| 5488 |
+
},
|
| 5489 |
+
{
|
| 5490 |
+
"epoch": 0.9234552720565632,
|
| 5491 |
+
"grad_norm": 0.4439335617291929,
|
| 5492 |
+
"learning_rate": 3.5289648789046616e-07,
|
| 5493 |
+
"loss": 0.1763,
|
| 5494 |
+
"step": 751
|
| 5495 |
+
},
|
| 5496 |
+
{
|
| 5497 |
+
"epoch": 0.9246849062403935,
|
| 5498 |
+
"grad_norm": 0.3529798858955445,
|
| 5499 |
+
"learning_rate": 3.4166928771054653e-07,
|
| 5500 |
+
"loss": 0.2423,
|
| 5501 |
+
"step": 752
|
| 5502 |
+
},
|
| 5503 |
+
{
|
| 5504 |
+
"epoch": 0.9259145404242238,
|
| 5505 |
+
"grad_norm": 0.44409731091250987,
|
| 5506 |
+
"learning_rate": 3.3062047584224934e-07,
|
| 5507 |
+
"loss": 0.2102,
|
| 5508 |
+
"step": 753
|
| 5509 |
+
},
|
| 5510 |
+
{
|
| 5511 |
+
"epoch": 0.9271441746080541,
|
| 5512 |
+
"grad_norm": 0.5097303302568637,
|
| 5513 |
+
"learning_rate": 3.197502563559185e-07,
|
| 5514 |
+
"loss": 0.157,
|
| 5515 |
+
"step": 754
|
| 5516 |
+
},
|
| 5517 |
+
{
|
| 5518 |
+
"epoch": 0.9283738087918845,
|
| 5519 |
+
"grad_norm": 0.312449073047315,
|
| 5520 |
+
"learning_rate": 3.0905883002332213e-07,
|
| 5521 |
+
"loss": 0.1459,
|
| 5522 |
+
"step": 755
|
| 5523 |
+
},
|
| 5524 |
+
{
|
| 5525 |
+
"epoch": 0.9296034429757147,
|
| 5526 |
+
"grad_norm": 0.5047008458688232,
|
| 5527 |
+
"learning_rate": 2.985463943139322e-07,
|
| 5528 |
+
"loss": 0.1738,
|
| 5529 |
+
"step": 756
|
| 5530 |
+
},
|
| 5531 |
+
{
|
| 5532 |
+
"epoch": 0.930833077159545,
|
| 5533 |
+
"grad_norm": 0.5557205879162439,
|
| 5534 |
+
"learning_rate": 2.882131433912883e-07,
|
| 5535 |
+
"loss": 0.2488,
|
| 5536 |
+
"step": 757
|
| 5537 |
+
},
|
| 5538 |
+
{
|
| 5539 |
+
"epoch": 0.9320627113433754,
|
| 5540 |
+
"grad_norm": 0.43487988536317196,
|
| 5541 |
+
"learning_rate": 2.7805926810940297e-07,
|
| 5542 |
+
"loss": 0.1683,
|
| 5543 |
+
"step": 758
|
| 5544 |
+
},
|
| 5545 |
+
{
|
| 5546 |
+
"epoch": 0.9332923455272056,
|
| 5547 |
+
"grad_norm": 0.41756639561772235,
|
| 5548 |
+
"learning_rate": 2.6808495600924355e-07,
|
| 5549 |
+
"loss": 0.1387,
|
| 5550 |
+
"step": 759
|
| 5551 |
+
},
|
| 5552 |
+
{
|
| 5553 |
+
"epoch": 0.934521979711036,
|
| 5554 |
+
"grad_norm": 0.3700165943051797,
|
| 5555 |
+
"learning_rate": 2.582903913152612e-07,
|
| 5556 |
+
"loss": 0.164,
|
| 5557 |
+
"step": 760
|
| 5558 |
+
},
|
| 5559 |
+
{
|
| 5560 |
+
"epoch": 0.934521979711036,
|
| 5561 |
+
"eval_accuracy": 0.8074866310160428,
|
| 5562 |
+
"eval_f1": 0.5263157894736842,
|
| 5563 |
+
"eval_loss": 0.41820311546325684,
|
| 5564 |
+
"eval_precision": 0.7692307692307693,
|
| 5565 |
+
"eval_recall": 0.4,
|
| 5566 |
+
"eval_runtime": 23.5561,
|
| 5567 |
+
"eval_samples_per_second": 2.123,
|
| 5568 |
+
"eval_steps_per_second": 0.17,
|
| 5569 |
+
"step": 760
|
| 5570 |
+
},
|
| 5571 |
+
{
|
| 5572 |
+
"epoch": 0.9357516138948663,
|
| 5573 |
+
"grad_norm": 0.375235976764425,
|
| 5574 |
+
"learning_rate": 2.4867575493199515e-07,
|
| 5575 |
+
"loss": 0.1546,
|
| 5576 |
+
"step": 761
|
| 5577 |
+
},
|
| 5578 |
+
{
|
| 5579 |
+
"epoch": 0.9369812480786965,
|
| 5580 |
+
"grad_norm": 0.43981480426316444,
|
| 5581 |
+
"learning_rate": 2.392412244407294e-07,
|
| 5582 |
+
"loss": 0.1917,
|
| 5583 |
+
"step": 762
|
| 5584 |
+
},
|
| 5585 |
+
{
|
| 5586 |
+
"epoch": 0.9382108822625269,
|
| 5587 |
+
"grad_norm": 0.6482748347069781,
|
| 5588 |
+
"learning_rate": 2.2998697409620573e-07,
|
| 5589 |
+
"loss": 0.2225,
|
| 5590 |
+
"step": 763
|
| 5591 |
+
},
|
| 5592 |
+
{
|
| 5593 |
+
"epoch": 0.9394405164463572,
|
| 5594 |
+
"grad_norm": 0.42061880501285365,
|
| 5595 |
+
"learning_rate": 2.2091317482342056e-07,
|
| 5596 |
+
"loss": 0.23,
|
| 5597 |
+
"step": 764
|
| 5598 |
+
},
|
| 5599 |
+
{
|
| 5600 |
+
"epoch": 0.9406701506301876,
|
| 5601 |
+
"grad_norm": 0.33119896763642076,
|
| 5602 |
+
"learning_rate": 2.1201999421445074e-07,
|
| 5603 |
+
"loss": 0.1174,
|
| 5604 |
+
"step": 765
|
| 5605 |
+
},
|
| 5606 |
+
{
|
| 5607 |
+
"epoch": 0.9418997848140178,
|
| 5608 |
+
"grad_norm": 0.3630099682147168,
|
| 5609 |
+
"learning_rate": 2.0330759652536835e-07,
|
| 5610 |
+
"loss": 0.1809,
|
| 5611 |
+
"step": 766
|
| 5612 |
+
},
|
| 5613 |
+
{
|
| 5614 |
+
"epoch": 0.9431294189978482,
|
| 5615 |
+
"grad_norm": 0.5867052799544186,
|
| 5616 |
+
"learning_rate": 1.9477614267320867e-07,
|
| 5617 |
+
"loss": 0.258,
|
| 5618 |
+
"step": 767
|
| 5619 |
+
},
|
| 5620 |
+
{
|
| 5621 |
+
"epoch": 0.9443590531816785,
|
| 5622 |
+
"grad_norm": 0.4296071075770851,
|
| 5623 |
+
"learning_rate": 1.8642579023298913e-07,
|
| 5624 |
+
"loss": 0.2059,
|
| 5625 |
+
"step": 768
|
| 5626 |
+
},
|
| 5627 |
+
{
|
| 5628 |
+
"epoch": 0.9455886873655087,
|
| 5629 |
+
"grad_norm": 0.49292796300633096,
|
| 5630 |
+
"learning_rate": 1.7825669343480624e-07,
|
| 5631 |
+
"loss": 0.2087,
|
| 5632 |
+
"step": 769
|
| 5633 |
+
},
|
| 5634 |
+
{
|
| 5635 |
+
"epoch": 0.9468183215493391,
|
| 5636 |
+
"grad_norm": 0.4595855430483603,
|
| 5637 |
+
"learning_rate": 1.7026900316098217e-07,
|
| 5638 |
+
"loss": 0.1969,
|
| 5639 |
+
"step": 770
|
| 5640 |
+
},
|
| 5641 |
+
{
|
| 5642 |
+
"epoch": 0.9480479557331694,
|
| 5643 |
+
"grad_norm": 0.35921138730577007,
|
| 5644 |
+
"learning_rate": 1.6246286694328594e-07,
|
| 5645 |
+
"loss": 0.1353,
|
| 5646 |
+
"step": 771
|
| 5647 |
+
},
|
| 5648 |
+
{
|
| 5649 |
+
"epoch": 0.9492775899169997,
|
| 5650 |
+
"grad_norm": 0.41891935550025844,
|
| 5651 |
+
"learning_rate": 1.5483842896019675e-07,
|
| 5652 |
+
"loss": 0.144,
|
| 5653 |
+
"step": 772
|
| 5654 |
+
},
|
| 5655 |
+
{
|
| 5656 |
+
"epoch": 0.95050722410083,
|
| 5657 |
+
"grad_norm": 0.3608543106060588,
|
| 5658 |
+
"learning_rate": 1.473958300342504e-07,
|
| 5659 |
+
"loss": 0.1887,
|
| 5660 |
+
"step": 773
|
| 5661 |
+
},
|
| 5662 |
+
{
|
| 5663 |
+
"epoch": 0.9517368582846603,
|
| 5664 |
+
"grad_norm": 0.29802095126527695,
|
| 5665 |
+
"learning_rate": 1.401352076294371e-07,
|
| 5666 |
+
"loss": 0.114,
|
| 5667 |
+
"step": 774
|
| 5668 |
+
},
|
| 5669 |
+
{
|
| 5670 |
+
"epoch": 0.9529664924684906,
|
| 5671 |
+
"grad_norm": 0.4605125584516114,
|
| 5672 |
+
"learning_rate": 1.3305669584865565e-07,
|
| 5673 |
+
"loss": 0.226,
|
| 5674 |
+
"step": 775
|
| 5675 |
+
},
|
| 5676 |
+
{
|
| 5677 |
+
"epoch": 0.9541961266523209,
|
| 5678 |
+
"grad_norm": 0.4384971669173196,
|
| 5679 |
+
"learning_rate": 1.261604254312454e-07,
|
| 5680 |
+
"loss": 0.2221,
|
| 5681 |
+
"step": 776
|
| 5682 |
+
},
|
| 5683 |
+
{
|
| 5684 |
+
"epoch": 0.9554257608361513,
|
| 5685 |
+
"grad_norm": 0.5042450469158528,
|
| 5686 |
+
"learning_rate": 1.1944652375056597e-07,
|
| 5687 |
+
"loss": 0.1903,
|
| 5688 |
+
"step": 777
|
| 5689 |
+
},
|
| 5690 |
+
{
|
| 5691 |
+
"epoch": 0.9566553950199815,
|
| 5692 |
+
"grad_norm": 0.34864348698758946,
|
| 5693 |
+
"learning_rate": 1.1291511481164807e-07,
|
| 5694 |
+
"loss": 0.1514,
|
| 5695 |
+
"step": 778
|
| 5696 |
+
},
|
| 5697 |
+
{
|
| 5698 |
+
"epoch": 0.9578850292038119,
|
| 5699 |
+
"grad_norm": 0.39006673270014214,
|
| 5700 |
+
"learning_rate": 1.0656631924889749e-07,
|
| 5701 |
+
"loss": 0.1849,
|
| 5702 |
+
"step": 779
|
| 5703 |
+
},
|
| 5704 |
+
{
|
| 5705 |
+
"epoch": 0.9591146633876422,
|
| 5706 |
+
"grad_norm": 0.37108179891285153,
|
| 5707 |
+
"learning_rate": 1.0040025432387801e-07,
|
| 5708 |
+
"loss": 0.1699,
|
| 5709 |
+
"step": 780
|
| 5710 |
+
},
|
| 5711 |
+
{
|
| 5712 |
+
"epoch": 0.9603442975714724,
|
| 5713 |
+
"grad_norm": 0.4840858020566464,
|
| 5714 |
+
"learning_rate": 9.441703392313095e-08,
|
| 5715 |
+
"loss": 0.1975,
|
| 5716 |
+
"step": 781
|
| 5717 |
+
},
|
| 5718 |
+
{
|
| 5719 |
+
"epoch": 0.9615739317553028,
|
| 5720 |
+
"grad_norm": 0.4069300865488938,
|
| 5721 |
+
"learning_rate": 8.861676855608237e-08,
|
| 5722 |
+
"loss": 0.1635,
|
| 5723 |
+
"step": 782
|
| 5724 |
+
},
|
| 5725 |
+
{
|
| 5726 |
+
"epoch": 0.9628035659391331,
|
| 5727 |
+
"grad_norm": 0.43469366649519353,
|
| 5728 |
+
"learning_rate": 8.299956535300135e-08,
|
| 5729 |
+
"loss": 0.2255,
|
| 5730 |
+
"step": 783
|
| 5731 |
+
},
|
| 5732 |
+
{
|
| 5733 |
+
"epoch": 0.9640332001229635,
|
| 5734 |
+
"grad_norm": 0.4875682606006585,
|
| 5735 |
+
"learning_rate": 7.756552806301498e-08,
|
| 5736 |
+
"loss": 0.2556,
|
| 5737 |
+
"step": 784
|
| 5738 |
+
},
|
| 5739 |
+
{
|
| 5740 |
+
"epoch": 0.9652628343067937,
|
| 5741 |
+
"grad_norm": 0.37118130700190005,
|
| 5742 |
+
"learning_rate": 7.23147570521987e-08,
|
| 5743 |
+
"loss": 0.1849,
|
| 5744 |
+
"step": 785
|
| 5745 |
+
},
|
| 5746 |
+
{
|
| 5747 |
+
"epoch": 0.9664924684906241,
|
| 5748 |
+
"grad_norm": 0.40788113934634646,
|
| 5749 |
+
"learning_rate": 6.724734930171561e-08,
|
| 5750 |
+
"loss": 0.132,
|
| 5751 |
+
"step": 786
|
| 5752 |
+
},
|
| 5753 |
+
{
|
| 5754 |
+
"epoch": 0.9677221026744544,
|
| 5755 |
+
"grad_norm": 0.40100916877522913,
|
| 5756 |
+
"learning_rate": 6.236339840603677e-08,
|
| 5757 |
+
"loss": 0.1737,
|
| 5758 |
+
"step": 787
|
| 5759 |
+
},
|
| 5760 |
+
{
|
| 5761 |
+
"epoch": 0.9689517368582846,
|
| 5762 |
+
"grad_norm": 0.288332597781413,
|
| 5763 |
+
"learning_rate": 5.766299457119817e-08,
|
| 5764 |
+
"loss": 0.1717,
|
| 5765 |
+
"step": 788
|
| 5766 |
+
},
|
| 5767 |
+
{
|
| 5768 |
+
"epoch": 0.970181371042115,
|
| 5769 |
+
"grad_norm": 0.3844917502681006,
|
| 5770 |
+
"learning_rate": 5.3146224613144225e-08,
|
| 5771 |
+
"loss": 0.2025,
|
| 5772 |
+
"step": 789
|
| 5773 |
+
},
|
| 5774 |
+
{
|
| 5775 |
+
"epoch": 0.9714110052259453,
|
| 5776 |
+
"grad_norm": 0.39531701055460605,
|
| 5777 |
+
"learning_rate": 4.8813171956123565e-08,
|
| 5778 |
+
"loss": 0.1683,
|
| 5779 |
+
"step": 790
|
| 5780 |
+
},
|
| 5781 |
+
{
|
| 5782 |
+
"epoch": 0.9726406394097756,
|
| 5783 |
+
"grad_norm": 0.4042396380118405,
|
| 5784 |
+
"learning_rate": 4.4663916631143554e-08,
|
| 5785 |
+
"loss": 0.1624,
|
| 5786 |
+
"step": 791
|
| 5787 |
+
},
|
| 5788 |
+
{
|
| 5789 |
+
"epoch": 0.9738702735936059,
|
| 5790 |
+
"grad_norm": 0.2908881244181945,
|
| 5791 |
+
"learning_rate": 4.069853527449596e-08,
|
| 5792 |
+
"loss": 0.1629,
|
| 5793 |
+
"step": 792
|
| 5794 |
+
},
|
| 5795 |
+
{
|
| 5796 |
+
"epoch": 0.9750999077774362,
|
| 5797 |
+
"grad_norm": 0.4056597153568874,
|
| 5798 |
+
"learning_rate": 3.691710112634139e-08,
|
| 5799 |
+
"loss": 0.2174,
|
| 5800 |
+
"step": 793
|
| 5801 |
+
},
|
| 5802 |
+
{
|
| 5803 |
+
"epoch": 0.9763295419612665,
|
| 5804 |
+
"grad_norm": 0.45435351641201505,
|
| 5805 |
+
"learning_rate": 3.3319684029354815e-08,
|
| 5806 |
+
"loss": 0.1799,
|
| 5807 |
+
"step": 794
|
| 5808 |
+
},
|
| 5809 |
+
{
|
| 5810 |
+
"epoch": 0.9775591761450968,
|
| 5811 |
+
"grad_norm": 0.300609635946807,
|
| 5812 |
+
"learning_rate": 2.9906350427435505e-08,
|
| 5813 |
+
"loss": 0.1744,
|
| 5814 |
+
"step": 795
|
| 5815 |
+
},
|
| 5816 |
+
{
|
| 5817 |
+
"epoch": 0.9787888103289272,
|
| 5818 |
+
"grad_norm": 0.3475286198999752,
|
| 5819 |
+
"learning_rate": 2.667716336448356e-08,
|
| 5820 |
+
"loss": 0.1171,
|
| 5821 |
+
"step": 796
|
| 5822 |
+
},
|
| 5823 |
+
{
|
| 5824 |
+
"epoch": 0.9800184445127574,
|
| 5825 |
+
"grad_norm": 0.3272831675072718,
|
| 5826 |
+
"learning_rate": 2.3632182483228628e-08,
|
| 5827 |
+
"loss": 0.1556,
|
| 5828 |
+
"step": 797
|
| 5829 |
+
},
|
| 5830 |
+
{
|
| 5831 |
+
"epoch": 0.9812480786965878,
|
| 5832 |
+
"grad_norm": 0.48277032481078636,
|
| 5833 |
+
"learning_rate": 2.077146402413521e-08,
|
| 5834 |
+
"loss": 0.2464,
|
| 5835 |
+
"step": 798
|
| 5836 |
+
},
|
| 5837 |
+
{
|
| 5838 |
+
"epoch": 0.9824777128804181,
|
| 5839 |
+
"grad_norm": 0.32237492877791696,
|
| 5840 |
+
"learning_rate": 1.80950608243613e-08,
|
| 5841 |
+
"loss": 0.1548,
|
| 5842 |
+
"step": 799
|
| 5843 |
+
},
|
| 5844 |
+
{
|
| 5845 |
+
"epoch": 0.9837073470642483,
|
| 5846 |
+
"grad_norm": 0.4196152906906351,
|
| 5847 |
+
"learning_rate": 1.5603022316780235e-08,
|
| 5848 |
+
"loss": 0.2185,
|
| 5849 |
+
"step": 800
|
| 5850 |
+
},
|
| 5851 |
+
{
|
| 5852 |
+
"epoch": 0.9837073470642483,
|
| 5853 |
+
"eval_accuracy": 0.8128342245989305,
|
| 5854 |
+
"eval_f1": 0.5454545454545454,
|
| 5855 |
+
"eval_loss": 0.4195312559604645,
|
| 5856 |
+
"eval_precision": 0.7777777777777778,
|
| 5857 |
+
"eval_recall": 0.42,
|
| 5858 |
+
"eval_runtime": 23.0969,
|
| 5859 |
+
"eval_samples_per_second": 2.165,
|
| 5860 |
+
"eval_steps_per_second": 0.173,
|
| 5861 |
+
"step": 800
|
| 5862 |
}
|
| 5863 |
],
|
| 5864 |
"logging_steps": 1,
|
|
|
|
| 5878 |
"attributes": {}
|
| 5879 |
}
|
| 5880 |
},
|
| 5881 |
+
"total_flos": 785744118972416.0,
|
| 5882 |
"train_batch_size": 4,
|
| 5883 |
"trial_name": null,
|
| 5884 |
"trial_params": null
|