Training in progress, step 7500, checkpoint
Browse files- last-checkpoint/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/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +1761 -3
last-checkpoint/model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 368988278
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:34fe23e2546f350bdc3ea7cd098359c61876cfd9860bbc4e904fff96718928df
|
3 |
size 368988278
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1107079290
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cc0d1fd974fead158de0eadb70e7f57c959f5cfaef326177d457bb3324066005
|
3 |
size 1107079290
|
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 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0c461c9d337dfc684e9352ec72bfa344e2f5d377f7cfc4475de9acae294dca89
|
3 |
size 15984
|
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 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fae392ec6232cbf9da21d6ed12bc8247d0d24e7f3a3606acd23be00f3e8bbfc5
|
3 |
size 15984
|
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 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cbf3e7ca9991a58b0b16574a3c653483c551c270aa05aba06c162ea593f7b0f2
|
3 |
size 15984
|
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 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c695bebf6bcb75cbe26378bfe0ab7e2a33c49f713b9d6e4d10632b24322977e7
|
3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d5ebb13c71265c5464c9aa9bb9b66f07764d73befe6cd63a2aaf8e781bf0a374
|
3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:12cc6e245e189be568c8dfd43a4dd8f04bb3dbd9f17f41458107935d2c2a6a9d
|
3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:36086646e9a8f76fea69f8a227112e83bb63524964ccdfb82f4cdad88b90e5e4
|
3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9b44153bacf860d0ca6ce4c6b9380a199feab8a72ca613e6745bfb671b02c4e4
|
3 |
size 15984
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1000
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:069d595c560369c6f0dc5e92d7d7a49b75f77981476650e93e924eb0ecc848f4
|
3 |
size 1000
|
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": 2500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -3523,6 +3523,1764 @@
|
|
3523 |
"eval_samples_per_second": 1547.288,
|
3524 |
"eval_steps_per_second": 48.356,
|
3525 |
"step": 5000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3526 |
}
|
3527 |
],
|
3528 |
"logging_steps": 10,
|
@@ -3542,7 +5300,7 @@
|
|
3542 |
"attributes": {}
|
3543 |
}
|
3544 |
},
|
3545 |
-
"total_flos":
|
3546 |
"train_batch_size": 4,
|
3547 |
"trial_name": null,
|
3548 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.7876923076923077,
|
5 |
"eval_steps": 2500,
|
6 |
+
"global_step": 7500,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3523 |
"eval_samples_per_second": 1547.288,
|
3524 |
"eval_steps_per_second": 48.356,
|
3525 |
"step": 5000
|
3526 |
+
},
|
3527 |
+
{
|
3528 |
+
"epoch": 0.5261784615384615,
|
3529 |
+
"grad_norm": 77.9375,
|
3530 |
+
"learning_rate": 9.989723076923077e-07,
|
3531 |
+
"loss": 135.7908,
|
3532 |
+
"step": 5010
|
3533 |
+
},
|
3534 |
+
{
|
3535 |
+
"epoch": 0.527228717948718,
|
3536 |
+
"grad_norm": 76.875,
|
3537 |
+
"learning_rate": 9.989702564102564e-07,
|
3538 |
+
"loss": 133.9965,
|
3539 |
+
"step": 5020
|
3540 |
+
},
|
3541 |
+
{
|
3542 |
+
"epoch": 0.5282789743589743,
|
3543 |
+
"grad_norm": 76.1875,
|
3544 |
+
"learning_rate": 9.98968205128205e-07,
|
3545 |
+
"loss": 136.395,
|
3546 |
+
"step": 5030
|
3547 |
+
},
|
3548 |
+
{
|
3549 |
+
"epoch": 0.5293292307692308,
|
3550 |
+
"grad_norm": 80.4375,
|
3551 |
+
"learning_rate": 9.989661538461538e-07,
|
3552 |
+
"loss": 137.1353,
|
3553 |
+
"step": 5040
|
3554 |
+
},
|
3555 |
+
{
|
3556 |
+
"epoch": 0.5303794871794871,
|
3557 |
+
"grad_norm": 79.6875,
|
3558 |
+
"learning_rate": 9.989641025641027e-07,
|
3559 |
+
"loss": 135.522,
|
3560 |
+
"step": 5050
|
3561 |
+
},
|
3562 |
+
{
|
3563 |
+
"epoch": 0.5314297435897436,
|
3564 |
+
"grad_norm": 79.125,
|
3565 |
+
"learning_rate": 9.989620512820511e-07,
|
3566 |
+
"loss": 135.3318,
|
3567 |
+
"step": 5060
|
3568 |
+
},
|
3569 |
+
{
|
3570 |
+
"epoch": 0.53248,
|
3571 |
+
"grad_norm": 79.4375,
|
3572 |
+
"learning_rate": 9.989599999999998e-07,
|
3573 |
+
"loss": 133.8391,
|
3574 |
+
"step": 5070
|
3575 |
+
},
|
3576 |
+
{
|
3577 |
+
"epoch": 0.5335302564102564,
|
3578 |
+
"grad_norm": 80.8125,
|
3579 |
+
"learning_rate": 9.989579487179487e-07,
|
3580 |
+
"loss": 134.4684,
|
3581 |
+
"step": 5080
|
3582 |
+
},
|
3583 |
+
{
|
3584 |
+
"epoch": 0.5345805128205128,
|
3585 |
+
"grad_norm": 73.375,
|
3586 |
+
"learning_rate": 9.989558974358974e-07,
|
3587 |
+
"loss": 135.565,
|
3588 |
+
"step": 5090
|
3589 |
+
},
|
3590 |
+
{
|
3591 |
+
"epoch": 0.5356307692307692,
|
3592 |
+
"grad_norm": 77.0,
|
3593 |
+
"learning_rate": 9.989538461538461e-07,
|
3594 |
+
"loss": 134.9499,
|
3595 |
+
"step": 5100
|
3596 |
+
},
|
3597 |
+
{
|
3598 |
+
"epoch": 0.5366810256410256,
|
3599 |
+
"grad_norm": 72.875,
|
3600 |
+
"learning_rate": 9.989517948717948e-07,
|
3601 |
+
"loss": 135.5963,
|
3602 |
+
"step": 5110
|
3603 |
+
},
|
3604 |
+
{
|
3605 |
+
"epoch": 0.5377312820512821,
|
3606 |
+
"grad_norm": 80.4375,
|
3607 |
+
"learning_rate": 9.989497435897435e-07,
|
3608 |
+
"loss": 134.9717,
|
3609 |
+
"step": 5120
|
3610 |
+
},
|
3611 |
+
{
|
3612 |
+
"epoch": 0.5387815384615384,
|
3613 |
+
"grad_norm": 76.9375,
|
3614 |
+
"learning_rate": 9.989476923076922e-07,
|
3615 |
+
"loss": 133.5198,
|
3616 |
+
"step": 5130
|
3617 |
+
},
|
3618 |
+
{
|
3619 |
+
"epoch": 0.5398317948717949,
|
3620 |
+
"grad_norm": 79.125,
|
3621 |
+
"learning_rate": 9.98945641025641e-07,
|
3622 |
+
"loss": 134.2271,
|
3623 |
+
"step": 5140
|
3624 |
+
},
|
3625 |
+
{
|
3626 |
+
"epoch": 0.5408820512820512,
|
3627 |
+
"grad_norm": 78.4375,
|
3628 |
+
"learning_rate": 9.989435897435898e-07,
|
3629 |
+
"loss": 135.2163,
|
3630 |
+
"step": 5150
|
3631 |
+
},
|
3632 |
+
{
|
3633 |
+
"epoch": 0.5419323076923077,
|
3634 |
+
"grad_norm": 82.75,
|
3635 |
+
"learning_rate": 9.989415384615383e-07,
|
3636 |
+
"loss": 134.3486,
|
3637 |
+
"step": 5160
|
3638 |
+
},
|
3639 |
+
{
|
3640 |
+
"epoch": 0.5429825641025641,
|
3641 |
+
"grad_norm": 75.5625,
|
3642 |
+
"learning_rate": 9.989394871794872e-07,
|
3643 |
+
"loss": 134.6154,
|
3644 |
+
"step": 5170
|
3645 |
+
},
|
3646 |
+
{
|
3647 |
+
"epoch": 0.5440328205128205,
|
3648 |
+
"grad_norm": 76.8125,
|
3649 |
+
"learning_rate": 9.989374358974359e-07,
|
3650 |
+
"loss": 134.4962,
|
3651 |
+
"step": 5180
|
3652 |
+
},
|
3653 |
+
{
|
3654 |
+
"epoch": 0.5450830769230769,
|
3655 |
+
"grad_norm": 76.3125,
|
3656 |
+
"learning_rate": 9.989353846153846e-07,
|
3657 |
+
"loss": 134.8321,
|
3658 |
+
"step": 5190
|
3659 |
+
},
|
3660 |
+
{
|
3661 |
+
"epoch": 0.5461333333333334,
|
3662 |
+
"grad_norm": 76.125,
|
3663 |
+
"learning_rate": 9.989333333333333e-07,
|
3664 |
+
"loss": 134.5848,
|
3665 |
+
"step": 5200
|
3666 |
+
},
|
3667 |
+
{
|
3668 |
+
"epoch": 0.5471835897435897,
|
3669 |
+
"grad_norm": 82.9375,
|
3670 |
+
"learning_rate": 9.98931282051282e-07,
|
3671 |
+
"loss": 135.8918,
|
3672 |
+
"step": 5210
|
3673 |
+
},
|
3674 |
+
{
|
3675 |
+
"epoch": 0.5482338461538462,
|
3676 |
+
"grad_norm": 79.75,
|
3677 |
+
"learning_rate": 9.989292307692307e-07,
|
3678 |
+
"loss": 134.9434,
|
3679 |
+
"step": 5220
|
3680 |
+
},
|
3681 |
+
{
|
3682 |
+
"epoch": 0.5492841025641025,
|
3683 |
+
"grad_norm": 79.0625,
|
3684 |
+
"learning_rate": 9.989271794871794e-07,
|
3685 |
+
"loss": 136.0459,
|
3686 |
+
"step": 5230
|
3687 |
+
},
|
3688 |
+
{
|
3689 |
+
"epoch": 0.550334358974359,
|
3690 |
+
"grad_norm": 80.5625,
|
3691 |
+
"learning_rate": 9.989251282051283e-07,
|
3692 |
+
"loss": 134.0239,
|
3693 |
+
"step": 5240
|
3694 |
+
},
|
3695 |
+
{
|
3696 |
+
"epoch": 0.5513846153846154,
|
3697 |
+
"grad_norm": 82.5625,
|
3698 |
+
"learning_rate": 9.98923076923077e-07,
|
3699 |
+
"loss": 135.3833,
|
3700 |
+
"step": 5250
|
3701 |
+
},
|
3702 |
+
{
|
3703 |
+
"epoch": 0.5524348717948718,
|
3704 |
+
"grad_norm": 79.125,
|
3705 |
+
"learning_rate": 9.989210256410257e-07,
|
3706 |
+
"loss": 135.0751,
|
3707 |
+
"step": 5260
|
3708 |
+
},
|
3709 |
+
{
|
3710 |
+
"epoch": 0.5534851282051282,
|
3711 |
+
"grad_norm": 91.125,
|
3712 |
+
"learning_rate": 9.989189743589743e-07,
|
3713 |
+
"loss": 134.6919,
|
3714 |
+
"step": 5270
|
3715 |
+
},
|
3716 |
+
{
|
3717 |
+
"epoch": 0.5545353846153847,
|
3718 |
+
"grad_norm": 79.375,
|
3719 |
+
"learning_rate": 9.98916923076923e-07,
|
3720 |
+
"loss": 134.0143,
|
3721 |
+
"step": 5280
|
3722 |
+
},
|
3723 |
+
{
|
3724 |
+
"epoch": 0.555585641025641,
|
3725 |
+
"grad_norm": 79.125,
|
3726 |
+
"learning_rate": 9.989148717948717e-07,
|
3727 |
+
"loss": 134.374,
|
3728 |
+
"step": 5290
|
3729 |
+
},
|
3730 |
+
{
|
3731 |
+
"epoch": 0.5566358974358975,
|
3732 |
+
"grad_norm": 77.75,
|
3733 |
+
"learning_rate": 9.989128205128204e-07,
|
3734 |
+
"loss": 134.8326,
|
3735 |
+
"step": 5300
|
3736 |
+
},
|
3737 |
+
{
|
3738 |
+
"epoch": 0.5576861538461538,
|
3739 |
+
"grad_norm": 79.25,
|
3740 |
+
"learning_rate": 9.989107692307691e-07,
|
3741 |
+
"loss": 134.68,
|
3742 |
+
"step": 5310
|
3743 |
+
},
|
3744 |
+
{
|
3745 |
+
"epoch": 0.5587364102564103,
|
3746 |
+
"grad_norm": 82.1875,
|
3747 |
+
"learning_rate": 9.989087179487178e-07,
|
3748 |
+
"loss": 134.3108,
|
3749 |
+
"step": 5320
|
3750 |
+
},
|
3751 |
+
{
|
3752 |
+
"epoch": 0.5597866666666667,
|
3753 |
+
"grad_norm": 74.0,
|
3754 |
+
"learning_rate": 9.989066666666667e-07,
|
3755 |
+
"loss": 134.6612,
|
3756 |
+
"step": 5330
|
3757 |
+
},
|
3758 |
+
{
|
3759 |
+
"epoch": 0.5608369230769231,
|
3760 |
+
"grad_norm": 80.0,
|
3761 |
+
"learning_rate": 9.989046153846154e-07,
|
3762 |
+
"loss": 135.2691,
|
3763 |
+
"step": 5340
|
3764 |
+
},
|
3765 |
+
{
|
3766 |
+
"epoch": 0.5618871794871795,
|
3767 |
+
"grad_norm": 73.875,
|
3768 |
+
"learning_rate": 9.989025641025641e-07,
|
3769 |
+
"loss": 134.7519,
|
3770 |
+
"step": 5350
|
3771 |
+
},
|
3772 |
+
{
|
3773 |
+
"epoch": 0.562937435897436,
|
3774 |
+
"grad_norm": 75.3125,
|
3775 |
+
"learning_rate": 9.989005128205128e-07,
|
3776 |
+
"loss": 134.7679,
|
3777 |
+
"step": 5360
|
3778 |
+
},
|
3779 |
+
{
|
3780 |
+
"epoch": 0.5639876923076923,
|
3781 |
+
"grad_norm": 82.3125,
|
3782 |
+
"learning_rate": 9.988984615384615e-07,
|
3783 |
+
"loss": 135.0176,
|
3784 |
+
"step": 5370
|
3785 |
+
},
|
3786 |
+
{
|
3787 |
+
"epoch": 0.5650379487179488,
|
3788 |
+
"grad_norm": 78.4375,
|
3789 |
+
"learning_rate": 9.988964102564102e-07,
|
3790 |
+
"loss": 135.313,
|
3791 |
+
"step": 5380
|
3792 |
+
},
|
3793 |
+
{
|
3794 |
+
"epoch": 0.5660882051282051,
|
3795 |
+
"grad_norm": 74.5,
|
3796 |
+
"learning_rate": 9.988943589743589e-07,
|
3797 |
+
"loss": 134.0978,
|
3798 |
+
"step": 5390
|
3799 |
+
},
|
3800 |
+
{
|
3801 |
+
"epoch": 0.5671384615384616,
|
3802 |
+
"grad_norm": 71.375,
|
3803 |
+
"learning_rate": 9.988923076923078e-07,
|
3804 |
+
"loss": 134.1742,
|
3805 |
+
"step": 5400
|
3806 |
+
},
|
3807 |
+
{
|
3808 |
+
"epoch": 0.568188717948718,
|
3809 |
+
"grad_norm": 74.75,
|
3810 |
+
"learning_rate": 9.988902564102563e-07,
|
3811 |
+
"loss": 134.1676,
|
3812 |
+
"step": 5410
|
3813 |
+
},
|
3814 |
+
{
|
3815 |
+
"epoch": 0.5692389743589744,
|
3816 |
+
"grad_norm": 80.0,
|
3817 |
+
"learning_rate": 9.98888205128205e-07,
|
3818 |
+
"loss": 136.1296,
|
3819 |
+
"step": 5420
|
3820 |
+
},
|
3821 |
+
{
|
3822 |
+
"epoch": 0.5702892307692308,
|
3823 |
+
"grad_norm": 72.25,
|
3824 |
+
"learning_rate": 9.988861538461539e-07,
|
3825 |
+
"loss": 135.4083,
|
3826 |
+
"step": 5430
|
3827 |
+
},
|
3828 |
+
{
|
3829 |
+
"epoch": 0.5713394871794871,
|
3830 |
+
"grad_norm": 74.4375,
|
3831 |
+
"learning_rate": 9.988841025641026e-07,
|
3832 |
+
"loss": 136.1724,
|
3833 |
+
"step": 5440
|
3834 |
+
},
|
3835 |
+
{
|
3836 |
+
"epoch": 0.5723897435897436,
|
3837 |
+
"grad_norm": 80.25,
|
3838 |
+
"learning_rate": 9.988820512820513e-07,
|
3839 |
+
"loss": 134.1989,
|
3840 |
+
"step": 5450
|
3841 |
+
},
|
3842 |
+
{
|
3843 |
+
"epoch": 0.57344,
|
3844 |
+
"grad_norm": 83.0,
|
3845 |
+
"learning_rate": 9.9888e-07,
|
3846 |
+
"loss": 134.7482,
|
3847 |
+
"step": 5460
|
3848 |
+
},
|
3849 |
+
{
|
3850 |
+
"epoch": 0.5744902564102564,
|
3851 |
+
"grad_norm": 75.75,
|
3852 |
+
"learning_rate": 9.988779487179486e-07,
|
3853 |
+
"loss": 134.8441,
|
3854 |
+
"step": 5470
|
3855 |
+
},
|
3856 |
+
{
|
3857 |
+
"epoch": 0.5755405128205128,
|
3858 |
+
"grad_norm": 73.6875,
|
3859 |
+
"learning_rate": 9.988758974358973e-07,
|
3860 |
+
"loss": 133.2782,
|
3861 |
+
"step": 5480
|
3862 |
+
},
|
3863 |
+
{
|
3864 |
+
"epoch": 0.5765907692307692,
|
3865 |
+
"grad_norm": 76.5625,
|
3866 |
+
"learning_rate": 9.98873846153846e-07,
|
3867 |
+
"loss": 132.6737,
|
3868 |
+
"step": 5490
|
3869 |
+
},
|
3870 |
+
{
|
3871 |
+
"epoch": 0.5776410256410256,
|
3872 |
+
"grad_norm": 76.875,
|
3873 |
+
"learning_rate": 9.98871794871795e-07,
|
3874 |
+
"loss": 133.4861,
|
3875 |
+
"step": 5500
|
3876 |
+
},
|
3877 |
+
{
|
3878 |
+
"epoch": 0.5786912820512821,
|
3879 |
+
"grad_norm": 77.625,
|
3880 |
+
"learning_rate": 9.988697435897434e-07,
|
3881 |
+
"loss": 135.5047,
|
3882 |
+
"step": 5510
|
3883 |
+
},
|
3884 |
+
{
|
3885 |
+
"epoch": 0.5797415384615384,
|
3886 |
+
"grad_norm": 76.4375,
|
3887 |
+
"learning_rate": 9.988676923076923e-07,
|
3888 |
+
"loss": 135.0375,
|
3889 |
+
"step": 5520
|
3890 |
+
},
|
3891 |
+
{
|
3892 |
+
"epoch": 0.5807917948717949,
|
3893 |
+
"grad_norm": 72.0,
|
3894 |
+
"learning_rate": 9.98865641025641e-07,
|
3895 |
+
"loss": 135.1532,
|
3896 |
+
"step": 5530
|
3897 |
+
},
|
3898 |
+
{
|
3899 |
+
"epoch": 0.5818420512820512,
|
3900 |
+
"grad_norm": 76.8125,
|
3901 |
+
"learning_rate": 9.988635897435897e-07,
|
3902 |
+
"loss": 133.7135,
|
3903 |
+
"step": 5540
|
3904 |
+
},
|
3905 |
+
{
|
3906 |
+
"epoch": 0.5828923076923077,
|
3907 |
+
"grad_norm": 77.625,
|
3908 |
+
"learning_rate": 9.988615384615384e-07,
|
3909 |
+
"loss": 134.7512,
|
3910 |
+
"step": 5550
|
3911 |
+
},
|
3912 |
+
{
|
3913 |
+
"epoch": 0.5839425641025641,
|
3914 |
+
"grad_norm": 76.8125,
|
3915 |
+
"learning_rate": 9.988594871794871e-07,
|
3916 |
+
"loss": 133.1693,
|
3917 |
+
"step": 5560
|
3918 |
+
},
|
3919 |
+
{
|
3920 |
+
"epoch": 0.5849928205128205,
|
3921 |
+
"grad_norm": 73.8125,
|
3922 |
+
"learning_rate": 9.988574358974358e-07,
|
3923 |
+
"loss": 134.1779,
|
3924 |
+
"step": 5570
|
3925 |
+
},
|
3926 |
+
{
|
3927 |
+
"epoch": 0.5860430769230769,
|
3928 |
+
"grad_norm": 73.4375,
|
3929 |
+
"learning_rate": 9.988553846153845e-07,
|
3930 |
+
"loss": 134.0802,
|
3931 |
+
"step": 5580
|
3932 |
+
},
|
3933 |
+
{
|
3934 |
+
"epoch": 0.5870933333333334,
|
3935 |
+
"grad_norm": 80.25,
|
3936 |
+
"learning_rate": 9.988533333333334e-07,
|
3937 |
+
"loss": 133.4755,
|
3938 |
+
"step": 5590
|
3939 |
+
},
|
3940 |
+
{
|
3941 |
+
"epoch": 0.5881435897435897,
|
3942 |
+
"grad_norm": 75.5,
|
3943 |
+
"learning_rate": 9.98851282051282e-07,
|
3944 |
+
"loss": 133.0621,
|
3945 |
+
"step": 5600
|
3946 |
+
},
|
3947 |
+
{
|
3948 |
+
"epoch": 0.5891938461538462,
|
3949 |
+
"grad_norm": 76.75,
|
3950 |
+
"learning_rate": 9.988492307692308e-07,
|
3951 |
+
"loss": 134.8084,
|
3952 |
+
"step": 5610
|
3953 |
+
},
|
3954 |
+
{
|
3955 |
+
"epoch": 0.5902441025641025,
|
3956 |
+
"grad_norm": 80.5,
|
3957 |
+
"learning_rate": 9.988471794871795e-07,
|
3958 |
+
"loss": 134.011,
|
3959 |
+
"step": 5620
|
3960 |
+
},
|
3961 |
+
{
|
3962 |
+
"epoch": 0.591294358974359,
|
3963 |
+
"grad_norm": 79.25,
|
3964 |
+
"learning_rate": 9.988451282051282e-07,
|
3965 |
+
"loss": 134.1596,
|
3966 |
+
"step": 5630
|
3967 |
+
},
|
3968 |
+
{
|
3969 |
+
"epoch": 0.5923446153846154,
|
3970 |
+
"grad_norm": 79.1875,
|
3971 |
+
"learning_rate": 9.988430769230769e-07,
|
3972 |
+
"loss": 133.8143,
|
3973 |
+
"step": 5640
|
3974 |
+
},
|
3975 |
+
{
|
3976 |
+
"epoch": 0.5933948717948718,
|
3977 |
+
"grad_norm": 78.1875,
|
3978 |
+
"learning_rate": 9.988410256410256e-07,
|
3979 |
+
"loss": 132.8301,
|
3980 |
+
"step": 5650
|
3981 |
+
},
|
3982 |
+
{
|
3983 |
+
"epoch": 0.5944451282051282,
|
3984 |
+
"grad_norm": 81.4375,
|
3985 |
+
"learning_rate": 9.988389743589743e-07,
|
3986 |
+
"loss": 134.9465,
|
3987 |
+
"step": 5660
|
3988 |
+
},
|
3989 |
+
{
|
3990 |
+
"epoch": 0.5954953846153846,
|
3991 |
+
"grad_norm": 75.8125,
|
3992 |
+
"learning_rate": 9.98836923076923e-07,
|
3993 |
+
"loss": 133.7755,
|
3994 |
+
"step": 5670
|
3995 |
+
},
|
3996 |
+
{
|
3997 |
+
"epoch": 0.596545641025641,
|
3998 |
+
"grad_norm": 77.375,
|
3999 |
+
"learning_rate": 9.988348717948719e-07,
|
4000 |
+
"loss": 135.0217,
|
4001 |
+
"step": 5680
|
4002 |
+
},
|
4003 |
+
{
|
4004 |
+
"epoch": 0.5975958974358975,
|
4005 |
+
"grad_norm": 78.625,
|
4006 |
+
"learning_rate": 9.988328205128206e-07,
|
4007 |
+
"loss": 133.1542,
|
4008 |
+
"step": 5690
|
4009 |
+
},
|
4010 |
+
{
|
4011 |
+
"epoch": 0.5986461538461538,
|
4012 |
+
"grad_norm": 75.9375,
|
4013 |
+
"learning_rate": 9.988307692307692e-07,
|
4014 |
+
"loss": 133.7185,
|
4015 |
+
"step": 5700
|
4016 |
+
},
|
4017 |
+
{
|
4018 |
+
"epoch": 0.5996964102564103,
|
4019 |
+
"grad_norm": 76.625,
|
4020 |
+
"learning_rate": 9.98828717948718e-07,
|
4021 |
+
"loss": 134.098,
|
4022 |
+
"step": 5710
|
4023 |
+
},
|
4024 |
+
{
|
4025 |
+
"epoch": 0.6007466666666667,
|
4026 |
+
"grad_norm": 80.0,
|
4027 |
+
"learning_rate": 9.988266666666666e-07,
|
4028 |
+
"loss": 135.2549,
|
4029 |
+
"step": 5720
|
4030 |
+
},
|
4031 |
+
{
|
4032 |
+
"epoch": 0.6017969230769231,
|
4033 |
+
"grad_norm": 74.375,
|
4034 |
+
"learning_rate": 9.988246153846153e-07,
|
4035 |
+
"loss": 133.9132,
|
4036 |
+
"step": 5730
|
4037 |
+
},
|
4038 |
+
{
|
4039 |
+
"epoch": 0.6028471794871795,
|
4040 |
+
"grad_norm": 72.75,
|
4041 |
+
"learning_rate": 9.98822564102564e-07,
|
4042 |
+
"loss": 133.688,
|
4043 |
+
"step": 5740
|
4044 |
+
},
|
4045 |
+
{
|
4046 |
+
"epoch": 0.6038974358974359,
|
4047 |
+
"grad_norm": 74.4375,
|
4048 |
+
"learning_rate": 9.98820512820513e-07,
|
4049 |
+
"loss": 134.2703,
|
4050 |
+
"step": 5750
|
4051 |
+
},
|
4052 |
+
{
|
4053 |
+
"epoch": 0.6049476923076923,
|
4054 |
+
"grad_norm": 84.625,
|
4055 |
+
"learning_rate": 9.988184615384614e-07,
|
4056 |
+
"loss": 136.125,
|
4057 |
+
"step": 5760
|
4058 |
+
},
|
4059 |
+
{
|
4060 |
+
"epoch": 0.6059979487179488,
|
4061 |
+
"grad_norm": 76.8125,
|
4062 |
+
"learning_rate": 9.9881641025641e-07,
|
4063 |
+
"loss": 134.1665,
|
4064 |
+
"step": 5770
|
4065 |
+
},
|
4066 |
+
{
|
4067 |
+
"epoch": 0.6070482051282051,
|
4068 |
+
"grad_norm": 76.625,
|
4069 |
+
"learning_rate": 9.98814358974359e-07,
|
4070 |
+
"loss": 134.8555,
|
4071 |
+
"step": 5780
|
4072 |
+
},
|
4073 |
+
{
|
4074 |
+
"epoch": 0.6080984615384616,
|
4075 |
+
"grad_norm": 76.625,
|
4076 |
+
"learning_rate": 9.988123076923077e-07,
|
4077 |
+
"loss": 133.6014,
|
4078 |
+
"step": 5790
|
4079 |
+
},
|
4080 |
+
{
|
4081 |
+
"epoch": 0.6091487179487179,
|
4082 |
+
"grad_norm": 74.25,
|
4083 |
+
"learning_rate": 9.988102564102564e-07,
|
4084 |
+
"loss": 133.9827,
|
4085 |
+
"step": 5800
|
4086 |
+
},
|
4087 |
+
{
|
4088 |
+
"epoch": 0.6101989743589744,
|
4089 |
+
"grad_norm": 77.3125,
|
4090 |
+
"learning_rate": 9.98808205128205e-07,
|
4091 |
+
"loss": 133.2115,
|
4092 |
+
"step": 5810
|
4093 |
+
},
|
4094 |
+
{
|
4095 |
+
"epoch": 0.6112492307692308,
|
4096 |
+
"grad_norm": 73.9375,
|
4097 |
+
"learning_rate": 9.988061538461538e-07,
|
4098 |
+
"loss": 133.0385,
|
4099 |
+
"step": 5820
|
4100 |
+
},
|
4101 |
+
{
|
4102 |
+
"epoch": 0.6122994871794872,
|
4103 |
+
"grad_norm": 79.625,
|
4104 |
+
"learning_rate": 9.988041025641025e-07,
|
4105 |
+
"loss": 132.0571,
|
4106 |
+
"step": 5830
|
4107 |
+
},
|
4108 |
+
{
|
4109 |
+
"epoch": 0.6133497435897436,
|
4110 |
+
"grad_norm": 73.9375,
|
4111 |
+
"learning_rate": 9.988020512820512e-07,
|
4112 |
+
"loss": 133.794,
|
4113 |
+
"step": 5840
|
4114 |
+
},
|
4115 |
+
{
|
4116 |
+
"epoch": 0.6144,
|
4117 |
+
"grad_norm": 79.125,
|
4118 |
+
"learning_rate": 9.988e-07,
|
4119 |
+
"loss": 132.8695,
|
4120 |
+
"step": 5850
|
4121 |
+
},
|
4122 |
+
{
|
4123 |
+
"epoch": 0.6154502564102564,
|
4124 |
+
"grad_norm": 77.25,
|
4125 |
+
"learning_rate": 9.987979487179486e-07,
|
4126 |
+
"loss": 133.7304,
|
4127 |
+
"step": 5860
|
4128 |
+
},
|
4129 |
+
{
|
4130 |
+
"epoch": 0.6165005128205128,
|
4131 |
+
"grad_norm": 79.5,
|
4132 |
+
"learning_rate": 9.987958974358975e-07,
|
4133 |
+
"loss": 133.9512,
|
4134 |
+
"step": 5870
|
4135 |
+
},
|
4136 |
+
{
|
4137 |
+
"epoch": 0.6175507692307692,
|
4138 |
+
"grad_norm": 75.9375,
|
4139 |
+
"learning_rate": 9.987938461538462e-07,
|
4140 |
+
"loss": 135.2929,
|
4141 |
+
"step": 5880
|
4142 |
+
},
|
4143 |
+
{
|
4144 |
+
"epoch": 0.6186010256410256,
|
4145 |
+
"grad_norm": 83.3125,
|
4146 |
+
"learning_rate": 9.987917948717949e-07,
|
4147 |
+
"loss": 135.2723,
|
4148 |
+
"step": 5890
|
4149 |
+
},
|
4150 |
+
{
|
4151 |
+
"epoch": 0.6196512820512821,
|
4152 |
+
"grad_norm": 74.0625,
|
4153 |
+
"learning_rate": 9.987897435897435e-07,
|
4154 |
+
"loss": 135.1879,
|
4155 |
+
"step": 5900
|
4156 |
+
},
|
4157 |
+
{
|
4158 |
+
"epoch": 0.6207015384615384,
|
4159 |
+
"grad_norm": 76.75,
|
4160 |
+
"learning_rate": 9.987876923076922e-07,
|
4161 |
+
"loss": 134.8009,
|
4162 |
+
"step": 5910
|
4163 |
+
},
|
4164 |
+
{
|
4165 |
+
"epoch": 0.6217517948717949,
|
4166 |
+
"grad_norm": 82.5,
|
4167 |
+
"learning_rate": 9.98785641025641e-07,
|
4168 |
+
"loss": 135.7849,
|
4169 |
+
"step": 5920
|
4170 |
+
},
|
4171 |
+
{
|
4172 |
+
"epoch": 0.6228020512820512,
|
4173 |
+
"grad_norm": 74.4375,
|
4174 |
+
"learning_rate": 9.987835897435896e-07,
|
4175 |
+
"loss": 134.5044,
|
4176 |
+
"step": 5930
|
4177 |
+
},
|
4178 |
+
{
|
4179 |
+
"epoch": 0.6238523076923077,
|
4180 |
+
"grad_norm": 80.0,
|
4181 |
+
"learning_rate": 9.987815384615385e-07,
|
4182 |
+
"loss": 133.4429,
|
4183 |
+
"step": 5940
|
4184 |
+
},
|
4185 |
+
{
|
4186 |
+
"epoch": 0.6249025641025641,
|
4187 |
+
"grad_norm": 85.625,
|
4188 |
+
"learning_rate": 9.987794871794872e-07,
|
4189 |
+
"loss": 133.1182,
|
4190 |
+
"step": 5950
|
4191 |
+
},
|
4192 |
+
{
|
4193 |
+
"epoch": 0.6259528205128205,
|
4194 |
+
"grad_norm": 77.0625,
|
4195 |
+
"learning_rate": 9.98777435897436e-07,
|
4196 |
+
"loss": 133.511,
|
4197 |
+
"step": 5960
|
4198 |
+
},
|
4199 |
+
{
|
4200 |
+
"epoch": 0.6270030769230769,
|
4201 |
+
"grad_norm": 80.125,
|
4202 |
+
"learning_rate": 9.987753846153846e-07,
|
4203 |
+
"loss": 134.8845,
|
4204 |
+
"step": 5970
|
4205 |
+
},
|
4206 |
+
{
|
4207 |
+
"epoch": 0.6280533333333334,
|
4208 |
+
"grad_norm": 69.9375,
|
4209 |
+
"learning_rate": 9.987733333333333e-07,
|
4210 |
+
"loss": 133.8227,
|
4211 |
+
"step": 5980
|
4212 |
+
},
|
4213 |
+
{
|
4214 |
+
"epoch": 0.6291035897435897,
|
4215 |
+
"grad_norm": 75.5625,
|
4216 |
+
"learning_rate": 9.98771282051282e-07,
|
4217 |
+
"loss": 134.4892,
|
4218 |
+
"step": 5990
|
4219 |
+
},
|
4220 |
+
{
|
4221 |
+
"epoch": 0.6301538461538462,
|
4222 |
+
"grad_norm": 72.125,
|
4223 |
+
"learning_rate": 9.987692307692307e-07,
|
4224 |
+
"loss": 135.4582,
|
4225 |
+
"step": 6000
|
4226 |
+
},
|
4227 |
+
{
|
4228 |
+
"epoch": 0.6312041025641025,
|
4229 |
+
"grad_norm": 75.625,
|
4230 |
+
"learning_rate": 9.987671794871794e-07,
|
4231 |
+
"loss": 135.2013,
|
4232 |
+
"step": 6010
|
4233 |
+
},
|
4234 |
+
{
|
4235 |
+
"epoch": 0.632254358974359,
|
4236 |
+
"grad_norm": 74.0625,
|
4237 |
+
"learning_rate": 9.98765128205128e-07,
|
4238 |
+
"loss": 134.8731,
|
4239 |
+
"step": 6020
|
4240 |
+
},
|
4241 |
+
{
|
4242 |
+
"epoch": 0.6333046153846154,
|
4243 |
+
"grad_norm": 82.6875,
|
4244 |
+
"learning_rate": 9.98763076923077e-07,
|
4245 |
+
"loss": 135.0607,
|
4246 |
+
"step": 6030
|
4247 |
+
},
|
4248 |
+
{
|
4249 |
+
"epoch": 0.6343548717948718,
|
4250 |
+
"grad_norm": 79.75,
|
4251 |
+
"learning_rate": 9.987610256410257e-07,
|
4252 |
+
"loss": 134.0427,
|
4253 |
+
"step": 6040
|
4254 |
+
},
|
4255 |
+
{
|
4256 |
+
"epoch": 0.6354051282051282,
|
4257 |
+
"grad_norm": 75.4375,
|
4258 |
+
"learning_rate": 9.987589743589742e-07,
|
4259 |
+
"loss": 133.6469,
|
4260 |
+
"step": 6050
|
4261 |
+
},
|
4262 |
+
{
|
4263 |
+
"epoch": 0.6364553846153846,
|
4264 |
+
"grad_norm": 77.6875,
|
4265 |
+
"learning_rate": 9.98756923076923e-07,
|
4266 |
+
"loss": 133.951,
|
4267 |
+
"step": 6060
|
4268 |
+
},
|
4269 |
+
{
|
4270 |
+
"epoch": 0.637505641025641,
|
4271 |
+
"grad_norm": 74.1875,
|
4272 |
+
"learning_rate": 9.987548717948718e-07,
|
4273 |
+
"loss": 134.3955,
|
4274 |
+
"step": 6070
|
4275 |
+
},
|
4276 |
+
{
|
4277 |
+
"epoch": 0.6385558974358975,
|
4278 |
+
"grad_norm": 80.3125,
|
4279 |
+
"learning_rate": 9.987528205128205e-07,
|
4280 |
+
"loss": 133.7433,
|
4281 |
+
"step": 6080
|
4282 |
+
},
|
4283 |
+
{
|
4284 |
+
"epoch": 0.6396061538461538,
|
4285 |
+
"grad_norm": 75.8125,
|
4286 |
+
"learning_rate": 9.987507692307692e-07,
|
4287 |
+
"loss": 132.4955,
|
4288 |
+
"step": 6090
|
4289 |
+
},
|
4290 |
+
{
|
4291 |
+
"epoch": 0.6406564102564103,
|
4292 |
+
"grad_norm": 80.125,
|
4293 |
+
"learning_rate": 9.98748717948718e-07,
|
4294 |
+
"loss": 133.3074,
|
4295 |
+
"step": 6100
|
4296 |
+
},
|
4297 |
+
{
|
4298 |
+
"epoch": 0.6417066666666666,
|
4299 |
+
"grad_norm": 75.75,
|
4300 |
+
"learning_rate": 9.987466666666665e-07,
|
4301 |
+
"loss": 134.3473,
|
4302 |
+
"step": 6110
|
4303 |
+
},
|
4304 |
+
{
|
4305 |
+
"epoch": 0.6427569230769231,
|
4306 |
+
"grad_norm": 77.625,
|
4307 |
+
"learning_rate": 9.987446153846152e-07,
|
4308 |
+
"loss": 134.2209,
|
4309 |
+
"step": 6120
|
4310 |
+
},
|
4311 |
+
{
|
4312 |
+
"epoch": 0.6438071794871795,
|
4313 |
+
"grad_norm": 75.375,
|
4314 |
+
"learning_rate": 9.987425641025641e-07,
|
4315 |
+
"loss": 136.051,
|
4316 |
+
"step": 6130
|
4317 |
+
},
|
4318 |
+
{
|
4319 |
+
"epoch": 0.6448574358974359,
|
4320 |
+
"grad_norm": 78.625,
|
4321 |
+
"learning_rate": 9.987405128205128e-07,
|
4322 |
+
"loss": 133.6452,
|
4323 |
+
"step": 6140
|
4324 |
+
},
|
4325 |
+
{
|
4326 |
+
"epoch": 0.6459076923076923,
|
4327 |
+
"grad_norm": 72.875,
|
4328 |
+
"learning_rate": 9.987384615384615e-07,
|
4329 |
+
"loss": 135.9148,
|
4330 |
+
"step": 6150
|
4331 |
+
},
|
4332 |
+
{
|
4333 |
+
"epoch": 0.6469579487179488,
|
4334 |
+
"grad_norm": 79.9375,
|
4335 |
+
"learning_rate": 9.987364102564102e-07,
|
4336 |
+
"loss": 133.9573,
|
4337 |
+
"step": 6160
|
4338 |
+
},
|
4339 |
+
{
|
4340 |
+
"epoch": 0.6480082051282051,
|
4341 |
+
"grad_norm": 77.375,
|
4342 |
+
"learning_rate": 9.98734358974359e-07,
|
4343 |
+
"loss": 134.187,
|
4344 |
+
"step": 6170
|
4345 |
+
},
|
4346 |
+
{
|
4347 |
+
"epoch": 0.6490584615384616,
|
4348 |
+
"grad_norm": 77.6875,
|
4349 |
+
"learning_rate": 9.987323076923076e-07,
|
4350 |
+
"loss": 135.4981,
|
4351 |
+
"step": 6180
|
4352 |
+
},
|
4353 |
+
{
|
4354 |
+
"epoch": 0.6501087179487179,
|
4355 |
+
"grad_norm": 76.375,
|
4356 |
+
"learning_rate": 9.987302564102563e-07,
|
4357 |
+
"loss": 134.0596,
|
4358 |
+
"step": 6190
|
4359 |
+
},
|
4360 |
+
{
|
4361 |
+
"epoch": 0.6511589743589744,
|
4362 |
+
"grad_norm": 68.75,
|
4363 |
+
"learning_rate": 9.98728205128205e-07,
|
4364 |
+
"loss": 133.4401,
|
4365 |
+
"step": 6200
|
4366 |
+
},
|
4367 |
+
{
|
4368 |
+
"epoch": 0.6522092307692308,
|
4369 |
+
"grad_norm": 88.0,
|
4370 |
+
"learning_rate": 9.987261538461537e-07,
|
4371 |
+
"loss": 133.9623,
|
4372 |
+
"step": 6210
|
4373 |
+
},
|
4374 |
+
{
|
4375 |
+
"epoch": 0.6532594871794872,
|
4376 |
+
"grad_norm": 73.125,
|
4377 |
+
"learning_rate": 9.987241025641026e-07,
|
4378 |
+
"loss": 134.9918,
|
4379 |
+
"step": 6220
|
4380 |
+
},
|
4381 |
+
{
|
4382 |
+
"epoch": 0.6543097435897436,
|
4383 |
+
"grad_norm": 79.6875,
|
4384 |
+
"learning_rate": 9.987220512820513e-07,
|
4385 |
+
"loss": 134.2293,
|
4386 |
+
"step": 6230
|
4387 |
+
},
|
4388 |
+
{
|
4389 |
+
"epoch": 0.65536,
|
4390 |
+
"grad_norm": 79.125,
|
4391 |
+
"learning_rate": 9.9872e-07,
|
4392 |
+
"loss": 134.9116,
|
4393 |
+
"step": 6240
|
4394 |
+
},
|
4395 |
+
{
|
4396 |
+
"epoch": 0.6564102564102564,
|
4397 |
+
"grad_norm": 74.0625,
|
4398 |
+
"learning_rate": 9.987179487179487e-07,
|
4399 |
+
"loss": 133.4889,
|
4400 |
+
"step": 6250
|
4401 |
+
},
|
4402 |
+
{
|
4403 |
+
"epoch": 0.6574605128205128,
|
4404 |
+
"grad_norm": 74.0,
|
4405 |
+
"learning_rate": 9.987158974358974e-07,
|
4406 |
+
"loss": 133.3706,
|
4407 |
+
"step": 6260
|
4408 |
+
},
|
4409 |
+
{
|
4410 |
+
"epoch": 0.6585107692307692,
|
4411 |
+
"grad_norm": 73.5625,
|
4412 |
+
"learning_rate": 9.98713846153846e-07,
|
4413 |
+
"loss": 134.6084,
|
4414 |
+
"step": 6270
|
4415 |
+
},
|
4416 |
+
{
|
4417 |
+
"epoch": 0.6595610256410256,
|
4418 |
+
"grad_norm": 76.8125,
|
4419 |
+
"learning_rate": 9.987117948717948e-07,
|
4420 |
+
"loss": 133.7456,
|
4421 |
+
"step": 6280
|
4422 |
+
},
|
4423 |
+
{
|
4424 |
+
"epoch": 0.6606112820512821,
|
4425 |
+
"grad_norm": 77.625,
|
4426 |
+
"learning_rate": 9.987097435897437e-07,
|
4427 |
+
"loss": 133.8633,
|
4428 |
+
"step": 6290
|
4429 |
+
},
|
4430 |
+
{
|
4431 |
+
"epoch": 0.6616615384615384,
|
4432 |
+
"grad_norm": 76.625,
|
4433 |
+
"learning_rate": 9.987076923076921e-07,
|
4434 |
+
"loss": 133.3514,
|
4435 |
+
"step": 6300
|
4436 |
+
},
|
4437 |
+
{
|
4438 |
+
"epoch": 0.6627117948717949,
|
4439 |
+
"grad_norm": 78.6875,
|
4440 |
+
"learning_rate": 9.98705641025641e-07,
|
4441 |
+
"loss": 133.0192,
|
4442 |
+
"step": 6310
|
4443 |
+
},
|
4444 |
+
{
|
4445 |
+
"epoch": 0.6637620512820512,
|
4446 |
+
"grad_norm": 76.375,
|
4447 |
+
"learning_rate": 9.987035897435897e-07,
|
4448 |
+
"loss": 135.2725,
|
4449 |
+
"step": 6320
|
4450 |
+
},
|
4451 |
+
{
|
4452 |
+
"epoch": 0.6648123076923077,
|
4453 |
+
"grad_norm": 75.25,
|
4454 |
+
"learning_rate": 9.987015384615384e-07,
|
4455 |
+
"loss": 133.7453,
|
4456 |
+
"step": 6330
|
4457 |
+
},
|
4458 |
+
{
|
4459 |
+
"epoch": 0.6658625641025641,
|
4460 |
+
"grad_norm": 76.0625,
|
4461 |
+
"learning_rate": 9.986994871794871e-07,
|
4462 |
+
"loss": 134.4162,
|
4463 |
+
"step": 6340
|
4464 |
+
},
|
4465 |
+
{
|
4466 |
+
"epoch": 0.6669128205128205,
|
4467 |
+
"grad_norm": 77.3125,
|
4468 |
+
"learning_rate": 9.986974358974358e-07,
|
4469 |
+
"loss": 133.5715,
|
4470 |
+
"step": 6350
|
4471 |
+
},
|
4472 |
+
{
|
4473 |
+
"epoch": 0.6679630769230769,
|
4474 |
+
"grad_norm": 79.375,
|
4475 |
+
"learning_rate": 9.986953846153845e-07,
|
4476 |
+
"loss": 133.915,
|
4477 |
+
"step": 6360
|
4478 |
+
},
|
4479 |
+
{
|
4480 |
+
"epoch": 0.6690133333333333,
|
4481 |
+
"grad_norm": 76.0,
|
4482 |
+
"learning_rate": 9.986933333333332e-07,
|
4483 |
+
"loss": 134.2913,
|
4484 |
+
"step": 6370
|
4485 |
+
},
|
4486 |
+
{
|
4487 |
+
"epoch": 0.6700635897435897,
|
4488 |
+
"grad_norm": 80.0,
|
4489 |
+
"learning_rate": 9.986912820512821e-07,
|
4490 |
+
"loss": 135.7478,
|
4491 |
+
"step": 6380
|
4492 |
+
},
|
4493 |
+
{
|
4494 |
+
"epoch": 0.6711138461538462,
|
4495 |
+
"grad_norm": 81.0625,
|
4496 |
+
"learning_rate": 9.986892307692308e-07,
|
4497 |
+
"loss": 135.2542,
|
4498 |
+
"step": 6390
|
4499 |
+
},
|
4500 |
+
{
|
4501 |
+
"epoch": 0.6721641025641025,
|
4502 |
+
"grad_norm": 73.3125,
|
4503 |
+
"learning_rate": 9.986871794871793e-07,
|
4504 |
+
"loss": 133.8043,
|
4505 |
+
"step": 6400
|
4506 |
+
},
|
4507 |
+
{
|
4508 |
+
"epoch": 0.673214358974359,
|
4509 |
+
"grad_norm": 74.9375,
|
4510 |
+
"learning_rate": 9.986851282051282e-07,
|
4511 |
+
"loss": 132.8474,
|
4512 |
+
"step": 6410
|
4513 |
+
},
|
4514 |
+
{
|
4515 |
+
"epoch": 0.6742646153846154,
|
4516 |
+
"grad_norm": 78.5625,
|
4517 |
+
"learning_rate": 9.98683076923077e-07,
|
4518 |
+
"loss": 133.1841,
|
4519 |
+
"step": 6420
|
4520 |
+
},
|
4521 |
+
{
|
4522 |
+
"epoch": 0.6753148717948718,
|
4523 |
+
"grad_norm": 81.8125,
|
4524 |
+
"learning_rate": 9.986810256410256e-07,
|
4525 |
+
"loss": 132.1895,
|
4526 |
+
"step": 6430
|
4527 |
+
},
|
4528 |
+
{
|
4529 |
+
"epoch": 0.6763651282051282,
|
4530 |
+
"grad_norm": 79.6875,
|
4531 |
+
"learning_rate": 9.986789743589743e-07,
|
4532 |
+
"loss": 133.3178,
|
4533 |
+
"step": 6440
|
4534 |
+
},
|
4535 |
+
{
|
4536 |
+
"epoch": 0.6774153846153846,
|
4537 |
+
"grad_norm": 77.9375,
|
4538 |
+
"learning_rate": 9.98676923076923e-07,
|
4539 |
+
"loss": 134.5587,
|
4540 |
+
"step": 6450
|
4541 |
+
},
|
4542 |
+
{
|
4543 |
+
"epoch": 0.678465641025641,
|
4544 |
+
"grad_norm": 76.75,
|
4545 |
+
"learning_rate": 9.986748717948717e-07,
|
4546 |
+
"loss": 133.5905,
|
4547 |
+
"step": 6460
|
4548 |
+
},
|
4549 |
+
{
|
4550 |
+
"epoch": 0.6795158974358975,
|
4551 |
+
"grad_norm": 74.0,
|
4552 |
+
"learning_rate": 9.986728205128204e-07,
|
4553 |
+
"loss": 133.0102,
|
4554 |
+
"step": 6470
|
4555 |
+
},
|
4556 |
+
{
|
4557 |
+
"epoch": 0.6805661538461538,
|
4558 |
+
"grad_norm": 85.0,
|
4559 |
+
"learning_rate": 9.986707692307693e-07,
|
4560 |
+
"loss": 135.4808,
|
4561 |
+
"step": 6480
|
4562 |
+
},
|
4563 |
+
{
|
4564 |
+
"epoch": 0.6816164102564103,
|
4565 |
+
"grad_norm": 82.0625,
|
4566 |
+
"learning_rate": 9.98668717948718e-07,
|
4567 |
+
"loss": 133.2434,
|
4568 |
+
"step": 6490
|
4569 |
+
},
|
4570 |
+
{
|
4571 |
+
"epoch": 0.6826666666666666,
|
4572 |
+
"grad_norm": 73.1875,
|
4573 |
+
"learning_rate": 9.986666666666667e-07,
|
4574 |
+
"loss": 134.3713,
|
4575 |
+
"step": 6500
|
4576 |
+
},
|
4577 |
+
{
|
4578 |
+
"epoch": 0.6837169230769231,
|
4579 |
+
"grad_norm": 75.9375,
|
4580 |
+
"learning_rate": 9.986646153846154e-07,
|
4581 |
+
"loss": 134.012,
|
4582 |
+
"step": 6510
|
4583 |
+
},
|
4584 |
+
{
|
4585 |
+
"epoch": 0.6847671794871795,
|
4586 |
+
"grad_norm": 77.5,
|
4587 |
+
"learning_rate": 9.98662564102564e-07,
|
4588 |
+
"loss": 134.6316,
|
4589 |
+
"step": 6520
|
4590 |
+
},
|
4591 |
+
{
|
4592 |
+
"epoch": 0.6858174358974359,
|
4593 |
+
"grad_norm": 79.9375,
|
4594 |
+
"learning_rate": 9.986605128205127e-07,
|
4595 |
+
"loss": 134.4724,
|
4596 |
+
"step": 6530
|
4597 |
+
},
|
4598 |
+
{
|
4599 |
+
"epoch": 0.6868676923076923,
|
4600 |
+
"grad_norm": 73.0,
|
4601 |
+
"learning_rate": 9.986584615384614e-07,
|
4602 |
+
"loss": 133.8849,
|
4603 |
+
"step": 6540
|
4604 |
+
},
|
4605 |
+
{
|
4606 |
+
"epoch": 0.6879179487179488,
|
4607 |
+
"grad_norm": 74.3125,
|
4608 |
+
"learning_rate": 9.986564102564101e-07,
|
4609 |
+
"loss": 135.4998,
|
4610 |
+
"step": 6550
|
4611 |
+
},
|
4612 |
+
{
|
4613 |
+
"epoch": 0.6889682051282051,
|
4614 |
+
"grad_norm": 82.0,
|
4615 |
+
"learning_rate": 9.986543589743588e-07,
|
4616 |
+
"loss": 134.5724,
|
4617 |
+
"step": 6560
|
4618 |
+
},
|
4619 |
+
{
|
4620 |
+
"epoch": 0.6900184615384616,
|
4621 |
+
"grad_norm": 77.625,
|
4622 |
+
"learning_rate": 9.986523076923077e-07,
|
4623 |
+
"loss": 134.0957,
|
4624 |
+
"step": 6570
|
4625 |
+
},
|
4626 |
+
{
|
4627 |
+
"epoch": 0.6910687179487179,
|
4628 |
+
"grad_norm": 73.5,
|
4629 |
+
"learning_rate": 9.986502564102564e-07,
|
4630 |
+
"loss": 133.3827,
|
4631 |
+
"step": 6580
|
4632 |
+
},
|
4633 |
+
{
|
4634 |
+
"epoch": 0.6921189743589744,
|
4635 |
+
"grad_norm": 73.625,
|
4636 |
+
"learning_rate": 9.986482051282051e-07,
|
4637 |
+
"loss": 134.7083,
|
4638 |
+
"step": 6590
|
4639 |
+
},
|
4640 |
+
{
|
4641 |
+
"epoch": 0.6931692307692308,
|
4642 |
+
"grad_norm": 81.1875,
|
4643 |
+
"learning_rate": 9.986461538461538e-07,
|
4644 |
+
"loss": 134.4949,
|
4645 |
+
"step": 6600
|
4646 |
+
},
|
4647 |
+
{
|
4648 |
+
"epoch": 0.6942194871794872,
|
4649 |
+
"grad_norm": 74.8125,
|
4650 |
+
"learning_rate": 9.986441025641025e-07,
|
4651 |
+
"loss": 133.9215,
|
4652 |
+
"step": 6610
|
4653 |
+
},
|
4654 |
+
{
|
4655 |
+
"epoch": 0.6952697435897436,
|
4656 |
+
"grad_norm": 76.5,
|
4657 |
+
"learning_rate": 9.986420512820512e-07,
|
4658 |
+
"loss": 134.7131,
|
4659 |
+
"step": 6620
|
4660 |
+
},
|
4661 |
+
{
|
4662 |
+
"epoch": 0.69632,
|
4663 |
+
"grad_norm": 74.1875,
|
4664 |
+
"learning_rate": 9.9864e-07,
|
4665 |
+
"loss": 132.6291,
|
4666 |
+
"step": 6630
|
4667 |
+
},
|
4668 |
+
{
|
4669 |
+
"epoch": 0.6973702564102564,
|
4670 |
+
"grad_norm": 76.375,
|
4671 |
+
"learning_rate": 9.986379487179488e-07,
|
4672 |
+
"loss": 133.4663,
|
4673 |
+
"step": 6640
|
4674 |
+
},
|
4675 |
+
{
|
4676 |
+
"epoch": 0.6984205128205129,
|
4677 |
+
"grad_norm": 76.6875,
|
4678 |
+
"learning_rate": 9.986358974358973e-07,
|
4679 |
+
"loss": 133.2238,
|
4680 |
+
"step": 6650
|
4681 |
+
},
|
4682 |
+
{
|
4683 |
+
"epoch": 0.6994707692307692,
|
4684 |
+
"grad_norm": 75.4375,
|
4685 |
+
"learning_rate": 9.986338461538462e-07,
|
4686 |
+
"loss": 135.6267,
|
4687 |
+
"step": 6660
|
4688 |
+
},
|
4689 |
+
{
|
4690 |
+
"epoch": 0.7005210256410257,
|
4691 |
+
"grad_norm": 76.5625,
|
4692 |
+
"learning_rate": 9.986317948717949e-07,
|
4693 |
+
"loss": 133.9723,
|
4694 |
+
"step": 6670
|
4695 |
+
},
|
4696 |
+
{
|
4697 |
+
"epoch": 0.701571282051282,
|
4698 |
+
"grad_norm": 75.5,
|
4699 |
+
"learning_rate": 9.986297435897436e-07,
|
4700 |
+
"loss": 132.1382,
|
4701 |
+
"step": 6680
|
4702 |
+
},
|
4703 |
+
{
|
4704 |
+
"epoch": 0.7026215384615384,
|
4705 |
+
"grad_norm": 75.375,
|
4706 |
+
"learning_rate": 9.986276923076923e-07,
|
4707 |
+
"loss": 132.8579,
|
4708 |
+
"step": 6690
|
4709 |
+
},
|
4710 |
+
{
|
4711 |
+
"epoch": 0.7036717948717949,
|
4712 |
+
"grad_norm": 76.3125,
|
4713 |
+
"learning_rate": 9.98625641025641e-07,
|
4714 |
+
"loss": 133.7719,
|
4715 |
+
"step": 6700
|
4716 |
+
},
|
4717 |
+
{
|
4718 |
+
"epoch": 0.7047220512820512,
|
4719 |
+
"grad_norm": 81.0625,
|
4720 |
+
"learning_rate": 9.986235897435897e-07,
|
4721 |
+
"loss": 135.3735,
|
4722 |
+
"step": 6710
|
4723 |
+
},
|
4724 |
+
{
|
4725 |
+
"epoch": 0.7057723076923077,
|
4726 |
+
"grad_norm": 75.75,
|
4727 |
+
"learning_rate": 9.986215384615384e-07,
|
4728 |
+
"loss": 132.8083,
|
4729 |
+
"step": 6720
|
4730 |
+
},
|
4731 |
+
{
|
4732 |
+
"epoch": 0.706822564102564,
|
4733 |
+
"grad_norm": 74.8125,
|
4734 |
+
"learning_rate": 9.986194871794873e-07,
|
4735 |
+
"loss": 133.2499,
|
4736 |
+
"step": 6730
|
4737 |
+
},
|
4738 |
+
{
|
4739 |
+
"epoch": 0.7078728205128205,
|
4740 |
+
"grad_norm": 74.375,
|
4741 |
+
"learning_rate": 9.98617435897436e-07,
|
4742 |
+
"loss": 133.904,
|
4743 |
+
"step": 6740
|
4744 |
+
},
|
4745 |
+
{
|
4746 |
+
"epoch": 0.7089230769230769,
|
4747 |
+
"grad_norm": 71.25,
|
4748 |
+
"learning_rate": 9.986153846153844e-07,
|
4749 |
+
"loss": 134.7248,
|
4750 |
+
"step": 6750
|
4751 |
+
},
|
4752 |
+
{
|
4753 |
+
"epoch": 0.7099733333333333,
|
4754 |
+
"grad_norm": 78.4375,
|
4755 |
+
"learning_rate": 9.986133333333333e-07,
|
4756 |
+
"loss": 134.4564,
|
4757 |
+
"step": 6760
|
4758 |
+
},
|
4759 |
+
{
|
4760 |
+
"epoch": 0.7110235897435897,
|
4761 |
+
"grad_norm": 87.3125,
|
4762 |
+
"learning_rate": 9.98611282051282e-07,
|
4763 |
+
"loss": 132.3151,
|
4764 |
+
"step": 6770
|
4765 |
+
},
|
4766 |
+
{
|
4767 |
+
"epoch": 0.7120738461538462,
|
4768 |
+
"grad_norm": 73.0625,
|
4769 |
+
"learning_rate": 9.986092307692307e-07,
|
4770 |
+
"loss": 133.7481,
|
4771 |
+
"step": 6780
|
4772 |
+
},
|
4773 |
+
{
|
4774 |
+
"epoch": 0.7131241025641025,
|
4775 |
+
"grad_norm": 73.0,
|
4776 |
+
"learning_rate": 9.986071794871794e-07,
|
4777 |
+
"loss": 133.2306,
|
4778 |
+
"step": 6790
|
4779 |
+
},
|
4780 |
+
{
|
4781 |
+
"epoch": 0.714174358974359,
|
4782 |
+
"grad_norm": 74.6875,
|
4783 |
+
"learning_rate": 9.986051282051281e-07,
|
4784 |
+
"loss": 133.5443,
|
4785 |
+
"step": 6800
|
4786 |
+
},
|
4787 |
+
{
|
4788 |
+
"epoch": 0.7152246153846153,
|
4789 |
+
"grad_norm": 73.4375,
|
4790 |
+
"learning_rate": 9.986030769230768e-07,
|
4791 |
+
"loss": 133.3841,
|
4792 |
+
"step": 6810
|
4793 |
+
},
|
4794 |
+
{
|
4795 |
+
"epoch": 0.7162748717948718,
|
4796 |
+
"grad_norm": 73.9375,
|
4797 |
+
"learning_rate": 9.986010256410255e-07,
|
4798 |
+
"loss": 135.036,
|
4799 |
+
"step": 6820
|
4800 |
+
},
|
4801 |
+
{
|
4802 |
+
"epoch": 0.7173251282051282,
|
4803 |
+
"grad_norm": 83.9375,
|
4804 |
+
"learning_rate": 9.985989743589744e-07,
|
4805 |
+
"loss": 134.1337,
|
4806 |
+
"step": 6830
|
4807 |
+
},
|
4808 |
+
{
|
4809 |
+
"epoch": 0.7183753846153846,
|
4810 |
+
"grad_norm": 78.1875,
|
4811 |
+
"learning_rate": 9.98596923076923e-07,
|
4812 |
+
"loss": 133.3089,
|
4813 |
+
"step": 6840
|
4814 |
+
},
|
4815 |
+
{
|
4816 |
+
"epoch": 0.719425641025641,
|
4817 |
+
"grad_norm": 76.875,
|
4818 |
+
"learning_rate": 9.985948717948718e-07,
|
4819 |
+
"loss": 133.6665,
|
4820 |
+
"step": 6850
|
4821 |
+
},
|
4822 |
+
{
|
4823 |
+
"epoch": 0.7204758974358975,
|
4824 |
+
"grad_norm": 74.125,
|
4825 |
+
"learning_rate": 9.985928205128205e-07,
|
4826 |
+
"loss": 133.7026,
|
4827 |
+
"step": 6860
|
4828 |
+
},
|
4829 |
+
{
|
4830 |
+
"epoch": 0.7215261538461538,
|
4831 |
+
"grad_norm": 78.4375,
|
4832 |
+
"learning_rate": 9.985907692307692e-07,
|
4833 |
+
"loss": 133.7467,
|
4834 |
+
"step": 6870
|
4835 |
+
},
|
4836 |
+
{
|
4837 |
+
"epoch": 0.7225764102564103,
|
4838 |
+
"grad_norm": 73.1875,
|
4839 |
+
"learning_rate": 9.985887179487179e-07,
|
4840 |
+
"loss": 133.6408,
|
4841 |
+
"step": 6880
|
4842 |
+
},
|
4843 |
+
{
|
4844 |
+
"epoch": 0.7236266666666666,
|
4845 |
+
"grad_norm": 77.0625,
|
4846 |
+
"learning_rate": 9.985866666666666e-07,
|
4847 |
+
"loss": 132.6118,
|
4848 |
+
"step": 6890
|
4849 |
+
},
|
4850 |
+
{
|
4851 |
+
"epoch": 0.7246769230769231,
|
4852 |
+
"grad_norm": 76.0625,
|
4853 |
+
"learning_rate": 9.985846153846153e-07,
|
4854 |
+
"loss": 132.8741,
|
4855 |
+
"step": 6900
|
4856 |
+
},
|
4857 |
+
{
|
4858 |
+
"epoch": 0.7257271794871795,
|
4859 |
+
"grad_norm": 73.5625,
|
4860 |
+
"learning_rate": 9.98582564102564e-07,
|
4861 |
+
"loss": 134.1753,
|
4862 |
+
"step": 6910
|
4863 |
+
},
|
4864 |
+
{
|
4865 |
+
"epoch": 0.7267774358974359,
|
4866 |
+
"grad_norm": 76.875,
|
4867 |
+
"learning_rate": 9.985805128205129e-07,
|
4868 |
+
"loss": 132.842,
|
4869 |
+
"step": 6920
|
4870 |
+
},
|
4871 |
+
{
|
4872 |
+
"epoch": 0.7278276923076923,
|
4873 |
+
"grad_norm": 70.0,
|
4874 |
+
"learning_rate": 9.985784615384616e-07,
|
4875 |
+
"loss": 134.3394,
|
4876 |
+
"step": 6930
|
4877 |
+
},
|
4878 |
+
{
|
4879 |
+
"epoch": 0.7288779487179488,
|
4880 |
+
"grad_norm": 74.0625,
|
4881 |
+
"learning_rate": 9.985764102564103e-07,
|
4882 |
+
"loss": 135.2059,
|
4883 |
+
"step": 6940
|
4884 |
+
},
|
4885 |
+
{
|
4886 |
+
"epoch": 0.7299282051282051,
|
4887 |
+
"grad_norm": 78.625,
|
4888 |
+
"learning_rate": 9.98574358974359e-07,
|
4889 |
+
"loss": 134.0309,
|
4890 |
+
"step": 6950
|
4891 |
+
},
|
4892 |
+
{
|
4893 |
+
"epoch": 0.7309784615384616,
|
4894 |
+
"grad_norm": 77.75,
|
4895 |
+
"learning_rate": 9.985723076923076e-07,
|
4896 |
+
"loss": 134.4309,
|
4897 |
+
"step": 6960
|
4898 |
+
},
|
4899 |
+
{
|
4900 |
+
"epoch": 0.7320287179487179,
|
4901 |
+
"grad_norm": 74.6875,
|
4902 |
+
"learning_rate": 9.985702564102563e-07,
|
4903 |
+
"loss": 134.0655,
|
4904 |
+
"step": 6970
|
4905 |
+
},
|
4906 |
+
{
|
4907 |
+
"epoch": 0.7330789743589744,
|
4908 |
+
"grad_norm": 71.1875,
|
4909 |
+
"learning_rate": 9.98568205128205e-07,
|
4910 |
+
"loss": 134.798,
|
4911 |
+
"step": 6980
|
4912 |
+
},
|
4913 |
+
{
|
4914 |
+
"epoch": 0.7341292307692308,
|
4915 |
+
"grad_norm": 80.25,
|
4916 |
+
"learning_rate": 9.98566153846154e-07,
|
4917 |
+
"loss": 133.176,
|
4918 |
+
"step": 6990
|
4919 |
+
},
|
4920 |
+
{
|
4921 |
+
"epoch": 0.7351794871794872,
|
4922 |
+
"grad_norm": 77.3125,
|
4923 |
+
"learning_rate": 9.985641025641024e-07,
|
4924 |
+
"loss": 133.7738,
|
4925 |
+
"step": 7000
|
4926 |
+
},
|
4927 |
+
{
|
4928 |
+
"epoch": 0.7362297435897436,
|
4929 |
+
"grad_norm": 73.4375,
|
4930 |
+
"learning_rate": 9.985620512820513e-07,
|
4931 |
+
"loss": 133.5852,
|
4932 |
+
"step": 7010
|
4933 |
+
},
|
4934 |
+
{
|
4935 |
+
"epoch": 0.73728,
|
4936 |
+
"grad_norm": 82.125,
|
4937 |
+
"learning_rate": 9.9856e-07,
|
4938 |
+
"loss": 133.2783,
|
4939 |
+
"step": 7020
|
4940 |
+
},
|
4941 |
+
{
|
4942 |
+
"epoch": 0.7383302564102564,
|
4943 |
+
"grad_norm": 78.0625,
|
4944 |
+
"learning_rate": 9.985579487179487e-07,
|
4945 |
+
"loss": 133.2677,
|
4946 |
+
"step": 7030
|
4947 |
+
},
|
4948 |
+
{
|
4949 |
+
"epoch": 0.7393805128205129,
|
4950 |
+
"grad_norm": 78.5,
|
4951 |
+
"learning_rate": 9.985558974358974e-07,
|
4952 |
+
"loss": 133.9121,
|
4953 |
+
"step": 7040
|
4954 |
+
},
|
4955 |
+
{
|
4956 |
+
"epoch": 0.7404307692307692,
|
4957 |
+
"grad_norm": 78.875,
|
4958 |
+
"learning_rate": 9.98553846153846e-07,
|
4959 |
+
"loss": 133.3367,
|
4960 |
+
"step": 7050
|
4961 |
+
},
|
4962 |
+
{
|
4963 |
+
"epoch": 0.7414810256410257,
|
4964 |
+
"grad_norm": 77.8125,
|
4965 |
+
"learning_rate": 9.985517948717948e-07,
|
4966 |
+
"loss": 132.6513,
|
4967 |
+
"step": 7060
|
4968 |
+
},
|
4969 |
+
{
|
4970 |
+
"epoch": 0.742531282051282,
|
4971 |
+
"grad_norm": 73.0,
|
4972 |
+
"learning_rate": 9.985497435897435e-07,
|
4973 |
+
"loss": 135.524,
|
4974 |
+
"step": 7070
|
4975 |
+
},
|
4976 |
+
{
|
4977 |
+
"epoch": 0.7435815384615385,
|
4978 |
+
"grad_norm": 74.6875,
|
4979 |
+
"learning_rate": 9.985476923076924e-07,
|
4980 |
+
"loss": 132.8569,
|
4981 |
+
"step": 7080
|
4982 |
+
},
|
4983 |
+
{
|
4984 |
+
"epoch": 0.7446317948717949,
|
4985 |
+
"grad_norm": 71.25,
|
4986 |
+
"learning_rate": 9.98545641025641e-07,
|
4987 |
+
"loss": 132.9906,
|
4988 |
+
"step": 7090
|
4989 |
+
},
|
4990 |
+
{
|
4991 |
+
"epoch": 0.7456820512820512,
|
4992 |
+
"grad_norm": 75.4375,
|
4993 |
+
"learning_rate": 9.985435897435896e-07,
|
4994 |
+
"loss": 134.611,
|
4995 |
+
"step": 7100
|
4996 |
+
},
|
4997 |
+
{
|
4998 |
+
"epoch": 0.7467323076923077,
|
4999 |
+
"grad_norm": 76.0625,
|
5000 |
+
"learning_rate": 9.985415384615385e-07,
|
5001 |
+
"loss": 135.1841,
|
5002 |
+
"step": 7110
|
5003 |
+
},
|
5004 |
+
{
|
5005 |
+
"epoch": 0.747782564102564,
|
5006 |
+
"grad_norm": 80.25,
|
5007 |
+
"learning_rate": 9.985394871794872e-07,
|
5008 |
+
"loss": 132.8851,
|
5009 |
+
"step": 7120
|
5010 |
+
},
|
5011 |
+
{
|
5012 |
+
"epoch": 0.7488328205128205,
|
5013 |
+
"grad_norm": 77.875,
|
5014 |
+
"learning_rate": 9.985374358974359e-07,
|
5015 |
+
"loss": 132.9889,
|
5016 |
+
"step": 7130
|
5017 |
+
},
|
5018 |
+
{
|
5019 |
+
"epoch": 0.7498830769230769,
|
5020 |
+
"grad_norm": 74.375,
|
5021 |
+
"learning_rate": 9.985353846153846e-07,
|
5022 |
+
"loss": 134.75,
|
5023 |
+
"step": 7140
|
5024 |
+
},
|
5025 |
+
{
|
5026 |
+
"epoch": 0.7509333333333333,
|
5027 |
+
"grad_norm": 77.6875,
|
5028 |
+
"learning_rate": 9.985333333333332e-07,
|
5029 |
+
"loss": 131.8883,
|
5030 |
+
"step": 7150
|
5031 |
+
},
|
5032 |
+
{
|
5033 |
+
"epoch": 0.7519835897435897,
|
5034 |
+
"grad_norm": 71.5,
|
5035 |
+
"learning_rate": 9.98531282051282e-07,
|
5036 |
+
"loss": 132.9914,
|
5037 |
+
"step": 7160
|
5038 |
+
},
|
5039 |
+
{
|
5040 |
+
"epoch": 0.7530338461538462,
|
5041 |
+
"grad_norm": 77.4375,
|
5042 |
+
"learning_rate": 9.985292307692306e-07,
|
5043 |
+
"loss": 133.2398,
|
5044 |
+
"step": 7170
|
5045 |
+
},
|
5046 |
+
{
|
5047 |
+
"epoch": 0.7540841025641025,
|
5048 |
+
"grad_norm": 79.1875,
|
5049 |
+
"learning_rate": 9.985271794871795e-07,
|
5050 |
+
"loss": 133.4467,
|
5051 |
+
"step": 7180
|
5052 |
+
},
|
5053 |
+
{
|
5054 |
+
"epoch": 0.755134358974359,
|
5055 |
+
"grad_norm": 79.375,
|
5056 |
+
"learning_rate": 9.985251282051282e-07,
|
5057 |
+
"loss": 134.8063,
|
5058 |
+
"step": 7190
|
5059 |
+
},
|
5060 |
+
{
|
5061 |
+
"epoch": 0.7561846153846153,
|
5062 |
+
"grad_norm": 75.3125,
|
5063 |
+
"learning_rate": 9.98523076923077e-07,
|
5064 |
+
"loss": 133.1698,
|
5065 |
+
"step": 7200
|
5066 |
+
},
|
5067 |
+
{
|
5068 |
+
"epoch": 0.7572348717948718,
|
5069 |
+
"grad_norm": 73.5625,
|
5070 |
+
"learning_rate": 9.985210256410256e-07,
|
5071 |
+
"loss": 133.0102,
|
5072 |
+
"step": 7210
|
5073 |
+
},
|
5074 |
+
{
|
5075 |
+
"epoch": 0.7582851282051282,
|
5076 |
+
"grad_norm": 71.75,
|
5077 |
+
"learning_rate": 9.985189743589743e-07,
|
5078 |
+
"loss": 134.511,
|
5079 |
+
"step": 7220
|
5080 |
+
},
|
5081 |
+
{
|
5082 |
+
"epoch": 0.7593353846153846,
|
5083 |
+
"grad_norm": 85.5,
|
5084 |
+
"learning_rate": 9.98516923076923e-07,
|
5085 |
+
"loss": 134.3958,
|
5086 |
+
"step": 7230
|
5087 |
+
},
|
5088 |
+
{
|
5089 |
+
"epoch": 0.760385641025641,
|
5090 |
+
"grad_norm": 87.5625,
|
5091 |
+
"learning_rate": 9.985148717948717e-07,
|
5092 |
+
"loss": 134.2493,
|
5093 |
+
"step": 7240
|
5094 |
+
},
|
5095 |
+
{
|
5096 |
+
"epoch": 0.7614358974358975,
|
5097 |
+
"grad_norm": 82.25,
|
5098 |
+
"learning_rate": 9.985128205128204e-07,
|
5099 |
+
"loss": 133.0641,
|
5100 |
+
"step": 7250
|
5101 |
+
},
|
5102 |
+
{
|
5103 |
+
"epoch": 0.7624861538461538,
|
5104 |
+
"grad_norm": 82.25,
|
5105 |
+
"learning_rate": 9.98510769230769e-07,
|
5106 |
+
"loss": 134.3485,
|
5107 |
+
"step": 7260
|
5108 |
+
},
|
5109 |
+
{
|
5110 |
+
"epoch": 0.7635364102564103,
|
5111 |
+
"grad_norm": 79.1875,
|
5112 |
+
"learning_rate": 9.98508717948718e-07,
|
5113 |
+
"loss": 134.2047,
|
5114 |
+
"step": 7270
|
5115 |
+
},
|
5116 |
+
{
|
5117 |
+
"epoch": 0.7645866666666666,
|
5118 |
+
"grad_norm": 84.0625,
|
5119 |
+
"learning_rate": 9.985066666666667e-07,
|
5120 |
+
"loss": 133.9074,
|
5121 |
+
"step": 7280
|
5122 |
+
},
|
5123 |
+
{
|
5124 |
+
"epoch": 0.7656369230769231,
|
5125 |
+
"grad_norm": 75.875,
|
5126 |
+
"learning_rate": 9.985046153846154e-07,
|
5127 |
+
"loss": 132.4027,
|
5128 |
+
"step": 7290
|
5129 |
+
},
|
5130 |
+
{
|
5131 |
+
"epoch": 0.7666871794871795,
|
5132 |
+
"grad_norm": 79.6875,
|
5133 |
+
"learning_rate": 9.98502564102564e-07,
|
5134 |
+
"loss": 132.4265,
|
5135 |
+
"step": 7300
|
5136 |
+
},
|
5137 |
+
{
|
5138 |
+
"epoch": 0.7677374358974359,
|
5139 |
+
"grad_norm": 75.6875,
|
5140 |
+
"learning_rate": 9.985005128205128e-07,
|
5141 |
+
"loss": 132.4591,
|
5142 |
+
"step": 7310
|
5143 |
+
},
|
5144 |
+
{
|
5145 |
+
"epoch": 0.7687876923076923,
|
5146 |
+
"grad_norm": 70.75,
|
5147 |
+
"learning_rate": 9.984984615384615e-07,
|
5148 |
+
"loss": 133.4718,
|
5149 |
+
"step": 7320
|
5150 |
+
},
|
5151 |
+
{
|
5152 |
+
"epoch": 0.7698379487179488,
|
5153 |
+
"grad_norm": 74.3125,
|
5154 |
+
"learning_rate": 9.984964102564102e-07,
|
5155 |
+
"loss": 132.798,
|
5156 |
+
"step": 7330
|
5157 |
+
},
|
5158 |
+
{
|
5159 |
+
"epoch": 0.7708882051282051,
|
5160 |
+
"grad_norm": 76.4375,
|
5161 |
+
"learning_rate": 9.98494358974359e-07,
|
5162 |
+
"loss": 133.6705,
|
5163 |
+
"step": 7340
|
5164 |
+
},
|
5165 |
+
{
|
5166 |
+
"epoch": 0.7719384615384616,
|
5167 |
+
"grad_norm": 81.125,
|
5168 |
+
"learning_rate": 9.984923076923076e-07,
|
5169 |
+
"loss": 132.7228,
|
5170 |
+
"step": 7350
|
5171 |
+
},
|
5172 |
+
{
|
5173 |
+
"epoch": 0.7729887179487179,
|
5174 |
+
"grad_norm": 75.0,
|
5175 |
+
"learning_rate": 9.984902564102565e-07,
|
5176 |
+
"loss": 133.0821,
|
5177 |
+
"step": 7360
|
5178 |
+
},
|
5179 |
+
{
|
5180 |
+
"epoch": 0.7740389743589744,
|
5181 |
+
"grad_norm": 80.625,
|
5182 |
+
"learning_rate": 9.984882051282051e-07,
|
5183 |
+
"loss": 132.362,
|
5184 |
+
"step": 7370
|
5185 |
+
},
|
5186 |
+
{
|
5187 |
+
"epoch": 0.7750892307692308,
|
5188 |
+
"grad_norm": 74.6875,
|
5189 |
+
"learning_rate": 9.984861538461538e-07,
|
5190 |
+
"loss": 133.2498,
|
5191 |
+
"step": 7380
|
5192 |
+
},
|
5193 |
+
{
|
5194 |
+
"epoch": 0.7761394871794872,
|
5195 |
+
"grad_norm": 78.5,
|
5196 |
+
"learning_rate": 9.984841025641025e-07,
|
5197 |
+
"loss": 134.545,
|
5198 |
+
"step": 7390
|
5199 |
+
},
|
5200 |
+
{
|
5201 |
+
"epoch": 0.7771897435897436,
|
5202 |
+
"grad_norm": 82.625,
|
5203 |
+
"learning_rate": 9.984820512820512e-07,
|
5204 |
+
"loss": 135.3461,
|
5205 |
+
"step": 7400
|
5206 |
+
},
|
5207 |
+
{
|
5208 |
+
"epoch": 0.77824,
|
5209 |
+
"grad_norm": 80.8125,
|
5210 |
+
"learning_rate": 9.9848e-07,
|
5211 |
+
"loss": 133.5678,
|
5212 |
+
"step": 7410
|
5213 |
+
},
|
5214 |
+
{
|
5215 |
+
"epoch": 0.7792902564102564,
|
5216 |
+
"grad_norm": 77.875,
|
5217 |
+
"learning_rate": 9.984779487179486e-07,
|
5218 |
+
"loss": 133.22,
|
5219 |
+
"step": 7420
|
5220 |
+
},
|
5221 |
+
{
|
5222 |
+
"epoch": 0.7803405128205129,
|
5223 |
+
"grad_norm": 81.3125,
|
5224 |
+
"learning_rate": 9.984758974358975e-07,
|
5225 |
+
"loss": 133.2538,
|
5226 |
+
"step": 7430
|
5227 |
+
},
|
5228 |
+
{
|
5229 |
+
"epoch": 0.7813907692307692,
|
5230 |
+
"grad_norm": 76.125,
|
5231 |
+
"learning_rate": 9.98473846153846e-07,
|
5232 |
+
"loss": 134.0912,
|
5233 |
+
"step": 7440
|
5234 |
+
},
|
5235 |
+
{
|
5236 |
+
"epoch": 0.7824410256410257,
|
5237 |
+
"grad_norm": 77.375,
|
5238 |
+
"learning_rate": 9.984717948717947e-07,
|
5239 |
+
"loss": 133.252,
|
5240 |
+
"step": 7450
|
5241 |
+
},
|
5242 |
+
{
|
5243 |
+
"epoch": 0.783491282051282,
|
5244 |
+
"grad_norm": 82.0,
|
5245 |
+
"learning_rate": 9.984697435897436e-07,
|
5246 |
+
"loss": 131.4886,
|
5247 |
+
"step": 7460
|
5248 |
+
},
|
5249 |
+
{
|
5250 |
+
"epoch": 0.7845415384615385,
|
5251 |
+
"grad_norm": 73.375,
|
5252 |
+
"learning_rate": 9.984676923076923e-07,
|
5253 |
+
"loss": 133.4359,
|
5254 |
+
"step": 7470
|
5255 |
+
},
|
5256 |
+
{
|
5257 |
+
"epoch": 0.7855917948717949,
|
5258 |
+
"grad_norm": 74.375,
|
5259 |
+
"learning_rate": 9.98465641025641e-07,
|
5260 |
+
"loss": 134.0868,
|
5261 |
+
"step": 7480
|
5262 |
+
},
|
5263 |
+
{
|
5264 |
+
"epoch": 0.7866420512820513,
|
5265 |
+
"grad_norm": 83.4375,
|
5266 |
+
"learning_rate": 9.984635897435897e-07,
|
5267 |
+
"loss": 132.8313,
|
5268 |
+
"step": 7490
|
5269 |
+
},
|
5270 |
+
{
|
5271 |
+
"epoch": 0.7876923076923077,
|
5272 |
+
"grad_norm": 76.9375,
|
5273 |
+
"learning_rate": 9.984615384615384e-07,
|
5274 |
+
"loss": 134.246,
|
5275 |
+
"step": 7500
|
5276 |
+
},
|
5277 |
+
{
|
5278 |
+
"epoch": 0.7876923076923077,
|
5279 |
+
"eval_loss": 2.086611270904541,
|
5280 |
+
"eval_runtime": 327.2333,
|
5281 |
+
"eval_samples_per_second": 1527.965,
|
5282 |
+
"eval_steps_per_second": 47.752,
|
5283 |
+
"step": 7500
|
5284 |
}
|
5285 |
],
|
5286 |
"logging_steps": 10,
|
|
|
5300 |
"attributes": {}
|
5301 |
}
|
5302 |
},
|
5303 |
+
"total_flos": 2.0708374333095936e+19,
|
5304 |
"train_batch_size": 4,
|
5305 |
"trial_name": null,
|
5306 |
"trial_params": null
|