Fanucci
commited on
Training in progress, step 400, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f95ffd3edaa4a15b67b48a59a80d63016d75f6fc3379a58777aad2d538206609
|
| 3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1342555602
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7c3b004323916b5492c153fc08c323a290298007495050cc0cad546eab295bb
|
| 3 |
size 1342555602
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac26b518ee595e95332223ec66a0bc07e4d79f8083adaf367741a0653bfeebde
|
| 3 |
size 14244
|
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 sha256:6e768b610175e6759af08b219c6671106c68d968bba76e51c7807deb68dc4c1b
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1423,6 +1423,1414 @@
|
|
| 1423 |
"eval_samples_per_second": 5.638,
|
| 1424 |
"eval_steps_per_second": 1.409,
|
| 1425 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1426 |
}
|
| 1427 |
],
|
| 1428 |
"logging_steps": 1,
|
|
@@ -1451,7 +2859,7 @@
|
|
| 1451 |
"attributes": {}
|
| 1452 |
}
|
| 1453 |
},
|
| 1454 |
-
"total_flos":
|
| 1455 |
"train_batch_size": 4,
|
| 1456 |
"trial_name": null,
|
| 1457 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 3.206190586090088,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.1279795232762758,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1423 |
"eval_samples_per_second": 5.638,
|
| 1424 |
"eval_steps_per_second": 1.409,
|
| 1425 |
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.06430971044632859,
|
| 1429 |
+
"grad_norm": 5.751672744750977,
|
| 1430 |
+
"learning_rate": 0.00039633487176224863,
|
| 1431 |
+
"loss": 6.6452,
|
| 1432 |
+
"step": 201
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 0.06462965925451927,
|
| 1436 |
+
"grad_norm": 7.075699806213379,
|
| 1437 |
+
"learning_rate": 0.0003962965184815812,
|
| 1438 |
+
"loss": 6.6866,
|
| 1439 |
+
"step": 202
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 0.06494960806270997,
|
| 1443 |
+
"grad_norm": 8.748412132263184,
|
| 1444 |
+
"learning_rate": 0.00039625796754729636,
|
| 1445 |
+
"loss": 6.6261,
|
| 1446 |
+
"step": 203
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.06526955687090065,
|
| 1450 |
+
"grad_norm": 9.29478931427002,
|
| 1451 |
+
"learning_rate": 0.00039621921900029523,
|
| 1452 |
+
"loss": 6.8016,
|
| 1453 |
+
"step": 204
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 0.06558950567909134,
|
| 1457 |
+
"grad_norm": 6.518805027008057,
|
| 1458 |
+
"learning_rate": 0.00039618027288168843,
|
| 1459 |
+
"loss": 6.6664,
|
| 1460 |
+
"step": 205
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.06590945448728204,
|
| 1464 |
+
"grad_norm": 7.012482166290283,
|
| 1465 |
+
"learning_rate": 0.0003961411292327965,
|
| 1466 |
+
"loss": 6.5855,
|
| 1467 |
+
"step": 206
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.06622940329547272,
|
| 1471 |
+
"grad_norm": 4.071852684020996,
|
| 1472 |
+
"learning_rate": 0.00039610178809514925,
|
| 1473 |
+
"loss": 6.7401,
|
| 1474 |
+
"step": 207
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 0.06654935210366342,
|
| 1478 |
+
"grad_norm": 6.612091541290283,
|
| 1479 |
+
"learning_rate": 0.0003960622495104862,
|
| 1480 |
+
"loss": 6.456,
|
| 1481 |
+
"step": 208
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 0.0668693009118541,
|
| 1485 |
+
"grad_norm": 4.622671604156494,
|
| 1486 |
+
"learning_rate": 0.00039602251352075635,
|
| 1487 |
+
"loss": 6.629,
|
| 1488 |
+
"step": 209
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.0671892497200448,
|
| 1492 |
+
"grad_norm": 5.038511276245117,
|
| 1493 |
+
"learning_rate": 0.00039598258016811805,
|
| 1494 |
+
"loss": 6.4263,
|
| 1495 |
+
"step": 210
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.06750919852823548,
|
| 1499 |
+
"grad_norm": 5.86072301864624,
|
| 1500 |
+
"learning_rate": 0.00039594244949493903,
|
| 1501 |
+
"loss": 6.4438,
|
| 1502 |
+
"step": 211
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.06782914733642617,
|
| 1506 |
+
"grad_norm": 5.774945259094238,
|
| 1507 |
+
"learning_rate": 0.0003959021215437965,
|
| 1508 |
+
"loss": 6.8048,
|
| 1509 |
+
"step": 212
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 0.06814909614461687,
|
| 1513 |
+
"grad_norm": 6.453885555267334,
|
| 1514 |
+
"learning_rate": 0.000395861596357477,
|
| 1515 |
+
"loss": 6.5328,
|
| 1516 |
+
"step": 213
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 0.06846904495280755,
|
| 1520 |
+
"grad_norm": 9.138883590698242,
|
| 1521 |
+
"learning_rate": 0.000395820873978976,
|
| 1522 |
+
"loss": 6.6628,
|
| 1523 |
+
"step": 214
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 0.06878899376099824,
|
| 1527 |
+
"grad_norm": 7.3241376876831055,
|
| 1528 |
+
"learning_rate": 0.00039577995445149865,
|
| 1529 |
+
"loss": 6.5668,
|
| 1530 |
+
"step": 215
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.06910894256918892,
|
| 1534 |
+
"grad_norm": 6.592808246612549,
|
| 1535 |
+
"learning_rate": 0.00039573883781845884,
|
| 1536 |
+
"loss": 6.5612,
|
| 1537 |
+
"step": 216
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 0.06942889137737962,
|
| 1541 |
+
"grad_norm": 4.612181186676025,
|
| 1542 |
+
"learning_rate": 0.00039569752412347994,
|
| 1543 |
+
"loss": 6.5892,
|
| 1544 |
+
"step": 217
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.06974884018557032,
|
| 1548 |
+
"grad_norm": 5.334681034088135,
|
| 1549 |
+
"learning_rate": 0.0003956560134103941,
|
| 1550 |
+
"loss": 6.5731,
|
| 1551 |
+
"step": 218
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.070068788993761,
|
| 1555 |
+
"grad_norm": 9.595178604125977,
|
| 1556 |
+
"learning_rate": 0.0003956143057232427,
|
| 1557 |
+
"loss": 6.5792,
|
| 1558 |
+
"step": 219
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 0.07038873780195169,
|
| 1562 |
+
"grad_norm": 11.286225318908691,
|
| 1563 |
+
"learning_rate": 0.000395572401106276,
|
| 1564 |
+
"loss": 6.6624,
|
| 1565 |
+
"step": 220
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 0.07070868661014237,
|
| 1569 |
+
"grad_norm": 7.74772834777832,
|
| 1570 |
+
"learning_rate": 0.00039553029960395325,
|
| 1571 |
+
"loss": 6.7217,
|
| 1572 |
+
"step": 221
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.07102863541833307,
|
| 1576 |
+
"grad_norm": 6.920510768890381,
|
| 1577 |
+
"learning_rate": 0.0003954880012609425,
|
| 1578 |
+
"loss": 6.6113,
|
| 1579 |
+
"step": 222
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 0.07134858422652375,
|
| 1583 |
+
"grad_norm": 6.8180670738220215,
|
| 1584 |
+
"learning_rate": 0.00039544550612212085,
|
| 1585 |
+
"loss": 6.773,
|
| 1586 |
+
"step": 223
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.07166853303471445,
|
| 1590 |
+
"grad_norm": 8.286237716674805,
|
| 1591 |
+
"learning_rate": 0.00039540281423257394,
|
| 1592 |
+
"loss": 6.6248,
|
| 1593 |
+
"step": 224
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.07198848184290514,
|
| 1597 |
+
"grad_norm": 4.819367408752441,
|
| 1598 |
+
"learning_rate": 0.00039535992563759627,
|
| 1599 |
+
"loss": 6.7728,
|
| 1600 |
+
"step": 225
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 0.07230843065109582,
|
| 1604 |
+
"grad_norm": 6.927038192749023,
|
| 1605 |
+
"learning_rate": 0.000395316840382691,
|
| 1606 |
+
"loss": 6.7501,
|
| 1607 |
+
"step": 226
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 0.07262837945928652,
|
| 1611 |
+
"grad_norm": 6.734741687774658,
|
| 1612 |
+
"learning_rate": 0.00039527355851357005,
|
| 1613 |
+
"loss": 6.3144,
|
| 1614 |
+
"step": 227
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.0729483282674772,
|
| 1618 |
+
"grad_norm": 7.074644565582275,
|
| 1619 |
+
"learning_rate": 0.0003952300800761538,
|
| 1620 |
+
"loss": 6.7793,
|
| 1621 |
+
"step": 228
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 0.0732682770756679,
|
| 1625 |
+
"grad_norm": 15.208846092224121,
|
| 1626 |
+
"learning_rate": 0.00039518640511657126,
|
| 1627 |
+
"loss": 6.4693,
|
| 1628 |
+
"step": 229
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.07358822588385858,
|
| 1632 |
+
"grad_norm": 5.9610090255737305,
|
| 1633 |
+
"learning_rate": 0.00039514253368116,
|
| 1634 |
+
"loss": 6.7493,
|
| 1635 |
+
"step": 230
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.07390817469204927,
|
| 1639 |
+
"grad_norm": 17.66107940673828,
|
| 1640 |
+
"learning_rate": 0.00039509846581646583,
|
| 1641 |
+
"loss": 6.6046,
|
| 1642 |
+
"step": 231
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 0.07422812350023997,
|
| 1646 |
+
"grad_norm": 6.6399407386779785,
|
| 1647 |
+
"learning_rate": 0.00039505420156924316,
|
| 1648 |
+
"loss": 6.7082,
|
| 1649 |
+
"step": 232
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 0.07454807230843065,
|
| 1653 |
+
"grad_norm": 5.893923759460449,
|
| 1654 |
+
"learning_rate": 0.00039500974098645463,
|
| 1655 |
+
"loss": 6.4106,
|
| 1656 |
+
"step": 233
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.07486802111662134,
|
| 1660 |
+
"grad_norm": 5.578292369842529,
|
| 1661 |
+
"learning_rate": 0.00039496508411527144,
|
| 1662 |
+
"loss": 6.6654,
|
| 1663 |
+
"step": 234
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 0.07518796992481203,
|
| 1667 |
+
"grad_norm": 9.475028991699219,
|
| 1668 |
+
"learning_rate": 0.00039492023100307273,
|
| 1669 |
+
"loss": 6.4931,
|
| 1670 |
+
"step": 235
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.07550791873300272,
|
| 1674 |
+
"grad_norm": 8.972395896911621,
|
| 1675 |
+
"learning_rate": 0.00039487518169744587,
|
| 1676 |
+
"loss": 6.3256,
|
| 1677 |
+
"step": 236
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.0758278675411934,
|
| 1681 |
+
"grad_norm": 34.38764953613281,
|
| 1682 |
+
"learning_rate": 0.00039482993624618667,
|
| 1683 |
+
"loss": 6.2344,
|
| 1684 |
+
"step": 237
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 0.0761478163493841,
|
| 1688 |
+
"grad_norm": 12.006391525268555,
|
| 1689 |
+
"learning_rate": 0.00039478449469729877,
|
| 1690 |
+
"loss": 6.7194,
|
| 1691 |
+
"step": 238
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 0.0764677651575748,
|
| 1695 |
+
"grad_norm": 13.673998832702637,
|
| 1696 |
+
"learning_rate": 0.00039473885709899394,
|
| 1697 |
+
"loss": 6.4708,
|
| 1698 |
+
"step": 239
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.07678771396576548,
|
| 1702 |
+
"grad_norm": 9.385038375854492,
|
| 1703 |
+
"learning_rate": 0.00039469302349969194,
|
| 1704 |
+
"loss": 6.6304,
|
| 1705 |
+
"step": 240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 0.07710766277395617,
|
| 1709 |
+
"grad_norm": 565.4649047851562,
|
| 1710 |
+
"learning_rate": 0.00039464699394802064,
|
| 1711 |
+
"loss": 6.527,
|
| 1712 |
+
"step": 241
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.07742761158214685,
|
| 1716 |
+
"grad_norm": 35.85231399536133,
|
| 1717 |
+
"learning_rate": 0.00039460076849281544,
|
| 1718 |
+
"loss": 7.1016,
|
| 1719 |
+
"step": 242
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.07774756039033755,
|
| 1723 |
+
"grad_norm": 16.34745979309082,
|
| 1724 |
+
"learning_rate": 0.00039455434718311996,
|
| 1725 |
+
"loss": 6.5491,
|
| 1726 |
+
"step": 243
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 0.07806750919852823,
|
| 1730 |
+
"grad_norm": 10.916836738586426,
|
| 1731 |
+
"learning_rate": 0.0003945077300681855,
|
| 1732 |
+
"loss": 6.5222,
|
| 1733 |
+
"step": 244
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 0.07838745800671892,
|
| 1737 |
+
"grad_norm": 9.447298049926758,
|
| 1738 |
+
"learning_rate": 0.00039446091719747093,
|
| 1739 |
+
"loss": 6.4075,
|
| 1740 |
+
"step": 245
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.07870740681490962,
|
| 1744 |
+
"grad_norm": 9.539137840270996,
|
| 1745 |
+
"learning_rate": 0.00039441390862064306,
|
| 1746 |
+
"loss": 6.6475,
|
| 1747 |
+
"step": 246
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 0.0790273556231003,
|
| 1751 |
+
"grad_norm": 7.535435199737549,
|
| 1752 |
+
"learning_rate": 0.0003943667043875762,
|
| 1753 |
+
"loss": 6.2207,
|
| 1754 |
+
"step": 247
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.079347304431291,
|
| 1758 |
+
"grad_norm": 6.822633743286133,
|
| 1759 |
+
"learning_rate": 0.0003943193045483523,
|
| 1760 |
+
"loss": 6.3106,
|
| 1761 |
+
"step": 248
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.07966725323948168,
|
| 1765 |
+
"grad_norm": 7.698075294494629,
|
| 1766 |
+
"learning_rate": 0.0003942717091532607,
|
| 1767 |
+
"loss": 6.6178,
|
| 1768 |
+
"step": 249
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.07998720204767237,
|
| 1772 |
+
"grad_norm": 8.507805824279785,
|
| 1773 |
+
"learning_rate": 0.0003942239182527985,
|
| 1774 |
+
"loss": 6.2366,
|
| 1775 |
+
"step": 250
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 0.08030715085586307,
|
| 1779 |
+
"grad_norm": 17.260297775268555,
|
| 1780 |
+
"learning_rate": 0.00039417593189766995,
|
| 1781 |
+
"loss": 6.3205,
|
| 1782 |
+
"step": 251
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.08062709966405375,
|
| 1786 |
+
"grad_norm": 9.483721733093262,
|
| 1787 |
+
"learning_rate": 0.00039412775013878673,
|
| 1788 |
+
"loss": 6.5035,
|
| 1789 |
+
"step": 252
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 0.08094704847224445,
|
| 1793 |
+
"grad_norm": 7.060410499572754,
|
| 1794 |
+
"learning_rate": 0.00039407937302726796,
|
| 1795 |
+
"loss": 6.5481,
|
| 1796 |
+
"step": 253
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.08126699728043513,
|
| 1800 |
+
"grad_norm": 10.322737693786621,
|
| 1801 |
+
"learning_rate": 0.00039403080061443994,
|
| 1802 |
+
"loss": 6.3955,
|
| 1803 |
+
"step": 254
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.08158694608862582,
|
| 1807 |
+
"grad_norm": 12.29690170288086,
|
| 1808 |
+
"learning_rate": 0.0003939820329518361,
|
| 1809 |
+
"loss": 6.6427,
|
| 1810 |
+
"step": 255
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 0.0819068948968165,
|
| 1814 |
+
"grad_norm": 6.971141338348389,
|
| 1815 |
+
"learning_rate": 0.0003939330700911972,
|
| 1816 |
+
"loss": 6.4136,
|
| 1817 |
+
"step": 256
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 0.0822268437050072,
|
| 1821 |
+
"grad_norm": 7.917289733886719,
|
| 1822 |
+
"learning_rate": 0.0003938839120844709,
|
| 1823 |
+
"loss": 6.5704,
|
| 1824 |
+
"step": 257
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.0825467925131979,
|
| 1828 |
+
"grad_norm": 6.552984237670898,
|
| 1829 |
+
"learning_rate": 0.00039383455898381206,
|
| 1830 |
+
"loss": 6.3583,
|
| 1831 |
+
"step": 258
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 0.08286674132138858,
|
| 1835 |
+
"grad_norm": 6.688981056213379,
|
| 1836 |
+
"learning_rate": 0.0003937850108415825,
|
| 1837 |
+
"loss": 6.3685,
|
| 1838 |
+
"step": 259
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.08318669012957927,
|
| 1842 |
+
"grad_norm": 7.041812896728516,
|
| 1843 |
+
"learning_rate": 0.0003937352677103509,
|
| 1844 |
+
"loss": 6.5508,
|
| 1845 |
+
"step": 260
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.08350663893776995,
|
| 1849 |
+
"grad_norm": 10.411174774169922,
|
| 1850 |
+
"learning_rate": 0.0003936853296428928,
|
| 1851 |
+
"loss": 6.4838,
|
| 1852 |
+
"step": 261
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 0.08382658774596065,
|
| 1856 |
+
"grad_norm": 7.945723533630371,
|
| 1857 |
+
"learning_rate": 0.00039363519669219073,
|
| 1858 |
+
"loss": 6.4606,
|
| 1859 |
+
"step": 262
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 0.08414653655415133,
|
| 1863 |
+
"grad_norm": 6.014384746551514,
|
| 1864 |
+
"learning_rate": 0.0003935848689114338,
|
| 1865 |
+
"loss": 6.3841,
|
| 1866 |
+
"step": 263
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 0.08446648536234203,
|
| 1870 |
+
"grad_norm": 4.446471214294434,
|
| 1871 |
+
"learning_rate": 0.00039353434635401786,
|
| 1872 |
+
"loss": 6.4982,
|
| 1873 |
+
"step": 264
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 0.08478643417053272,
|
| 1877 |
+
"grad_norm": 6.786041259765625,
|
| 1878 |
+
"learning_rate": 0.0003934836290735455,
|
| 1879 |
+
"loss": 6.6322,
|
| 1880 |
+
"step": 265
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.0851063829787234,
|
| 1884 |
+
"grad_norm": 10.881097793579102,
|
| 1885 |
+
"learning_rate": 0.0003934327171238259,
|
| 1886 |
+
"loss": 6.543,
|
| 1887 |
+
"step": 266
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.0854263317869141,
|
| 1891 |
+
"grad_norm": 5.207790851593018,
|
| 1892 |
+
"learning_rate": 0.00039338161055887466,
|
| 1893 |
+
"loss": 6.4335,
|
| 1894 |
+
"step": 267
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 0.08574628059510478,
|
| 1898 |
+
"grad_norm": 4.667728900909424,
|
| 1899 |
+
"learning_rate": 0.0003933303094329139,
|
| 1900 |
+
"loss": 6.3188,
|
| 1901 |
+
"step": 268
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 0.08606622940329547,
|
| 1905 |
+
"grad_norm": 8.13780403137207,
|
| 1906 |
+
"learning_rate": 0.0003932788138003723,
|
| 1907 |
+
"loss": 6.4419,
|
| 1908 |
+
"step": 269
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 0.08638617821148616,
|
| 1912 |
+
"grad_norm": 6.202636241912842,
|
| 1913 |
+
"learning_rate": 0.00039322712371588477,
|
| 1914 |
+
"loss": 6.3797,
|
| 1915 |
+
"step": 270
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 0.08670612701967685,
|
| 1919 |
+
"grad_norm": 6.082520961761475,
|
| 1920 |
+
"learning_rate": 0.0003931752392342924,
|
| 1921 |
+
"loss": 6.4058,
|
| 1922 |
+
"step": 271
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.08702607582786755,
|
| 1926 |
+
"grad_norm": 8.391554832458496,
|
| 1927 |
+
"learning_rate": 0.00039312316041064283,
|
| 1928 |
+
"loss": 6.6174,
|
| 1929 |
+
"step": 272
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.08734602463605823,
|
| 1933 |
+
"grad_norm": 8.977246284484863,
|
| 1934 |
+
"learning_rate": 0.0003930708873001898,
|
| 1935 |
+
"loss": 6.3543,
|
| 1936 |
+
"step": 273
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.08766597344424892,
|
| 1940 |
+
"grad_norm": 7.828481197357178,
|
| 1941 |
+
"learning_rate": 0.00039301841995839285,
|
| 1942 |
+
"loss": 6.4354,
|
| 1943 |
+
"step": 274
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 0.0879859222524396,
|
| 1947 |
+
"grad_norm": 6.4253387451171875,
|
| 1948 |
+
"learning_rate": 0.0003929657584409181,
|
| 1949 |
+
"loss": 6.4135,
|
| 1950 |
+
"step": 275
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.0883058710606303,
|
| 1954 |
+
"grad_norm": 6.666152000427246,
|
| 1955 |
+
"learning_rate": 0.0003929129028036374,
|
| 1956 |
+
"loss": 6.157,
|
| 1957 |
+
"step": 276
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 0.08862581986882098,
|
| 1961 |
+
"grad_norm": 5.554171562194824,
|
| 1962 |
+
"learning_rate": 0.00039285985310262854,
|
| 1963 |
+
"loss": 6.2411,
|
| 1964 |
+
"step": 277
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.08894576867701168,
|
| 1968 |
+
"grad_norm": 5.765629291534424,
|
| 1969 |
+
"learning_rate": 0.0003928066093941753,
|
| 1970 |
+
"loss": 6.3223,
|
| 1971 |
+
"step": 278
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.08926571748520237,
|
| 1975 |
+
"grad_norm": 8.940204620361328,
|
| 1976 |
+
"learning_rate": 0.0003927531717347673,
|
| 1977 |
+
"loss": 6.4208,
|
| 1978 |
+
"step": 279
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 0.08958566629339305,
|
| 1982 |
+
"grad_norm": 6.127601623535156,
|
| 1983 |
+
"learning_rate": 0.00039269954018109976,
|
| 1984 |
+
"loss": 6.1999,
|
| 1985 |
+
"step": 280
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 0.08990561510158375,
|
| 1989 |
+
"grad_norm": 9.167810440063477,
|
| 1990 |
+
"learning_rate": 0.00039264571479007387,
|
| 1991 |
+
"loss": 6.4898,
|
| 1992 |
+
"step": 281
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 0.09022556390977443,
|
| 1996 |
+
"grad_norm": 7.327286720275879,
|
| 1997 |
+
"learning_rate": 0.0003925916956187963,
|
| 1998 |
+
"loss": 6.5342,
|
| 1999 |
+
"step": 282
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 0.09054551271796513,
|
| 2003 |
+
"grad_norm": 8.352947235107422,
|
| 2004 |
+
"learning_rate": 0.00039253748272457937,
|
| 2005 |
+
"loss": 6.3908,
|
| 2006 |
+
"step": 283
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.09086546152615582,
|
| 2010 |
+
"grad_norm": 6.588724136352539,
|
| 2011 |
+
"learning_rate": 0.00039248307616494095,
|
| 2012 |
+
"loss": 6.5415,
|
| 2013 |
+
"step": 284
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.0911854103343465,
|
| 2017 |
+
"grad_norm": 6.817644119262695,
|
| 2018 |
+
"learning_rate": 0.00039242847599760435,
|
| 2019 |
+
"loss": 6.2451,
|
| 2020 |
+
"step": 285
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 0.0915053591425372,
|
| 2024 |
+
"grad_norm": 4.902499198913574,
|
| 2025 |
+
"learning_rate": 0.00039237368228049827,
|
| 2026 |
+
"loss": 6.2267,
|
| 2027 |
+
"step": 286
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.09182530795072788,
|
| 2031 |
+
"grad_norm": 6.349551200866699,
|
| 2032 |
+
"learning_rate": 0.0003923186950717569,
|
| 2033 |
+
"loss": 6.3978,
|
| 2034 |
+
"step": 287
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 0.09214525675891858,
|
| 2038 |
+
"grad_norm": 4.191233158111572,
|
| 2039 |
+
"learning_rate": 0.00039226351442971937,
|
| 2040 |
+
"loss": 6.2194,
|
| 2041 |
+
"step": 288
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.09246520556710926,
|
| 2045 |
+
"grad_norm": 6.8765387535095215,
|
| 2046 |
+
"learning_rate": 0.0003922081404129305,
|
| 2047 |
+
"loss": 6.4729,
|
| 2048 |
+
"step": 289
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.09278515437529995,
|
| 2052 |
+
"grad_norm": 5.389702320098877,
|
| 2053 |
+
"learning_rate": 0.00039215257308013987,
|
| 2054 |
+
"loss": 6.2693,
|
| 2055 |
+
"step": 290
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.09310510318349065,
|
| 2059 |
+
"grad_norm": 5.059742450714111,
|
| 2060 |
+
"learning_rate": 0.0003920968124903025,
|
| 2061 |
+
"loss": 6.3552,
|
| 2062 |
+
"step": 291
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 0.09342505199168133,
|
| 2066 |
+
"grad_norm": 7.156213283538818,
|
| 2067 |
+
"learning_rate": 0.0003920408587025782,
|
| 2068 |
+
"loss": 6.1788,
|
| 2069 |
+
"step": 292
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 0.09374500079987202,
|
| 2073 |
+
"grad_norm": 7.247560977935791,
|
| 2074 |
+
"learning_rate": 0.0003919847117763319,
|
| 2075 |
+
"loss": 6.6775,
|
| 2076 |
+
"step": 293
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 0.0940649496080627,
|
| 2080 |
+
"grad_norm": 5.354501724243164,
|
| 2081 |
+
"learning_rate": 0.0003919283717711332,
|
| 2082 |
+
"loss": 6.5433,
|
| 2083 |
+
"step": 294
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 0.0943848984162534,
|
| 2087 |
+
"grad_norm": 6.527252674102783,
|
| 2088 |
+
"learning_rate": 0.00039187183874675696,
|
| 2089 |
+
"loss": 6.4697,
|
| 2090 |
+
"step": 295
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.09470484722444408,
|
| 2094 |
+
"grad_norm": 6.278224945068359,
|
| 2095 |
+
"learning_rate": 0.00039181511276318253,
|
| 2096 |
+
"loss": 6.4026,
|
| 2097 |
+
"step": 296
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.09502479603263478,
|
| 2101 |
+
"grad_norm": 9.525894165039062,
|
| 2102 |
+
"learning_rate": 0.00039175819388059395,
|
| 2103 |
+
"loss": 6.3048,
|
| 2104 |
+
"step": 297
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 0.09534474484082547,
|
| 2108 |
+
"grad_norm": 10.046257019042969,
|
| 2109 |
+
"learning_rate": 0.0003917010821593801,
|
| 2110 |
+
"loss": 6.3838,
|
| 2111 |
+
"step": 298
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 0.09566469364901616,
|
| 2115 |
+
"grad_norm": 5.030762672424316,
|
| 2116 |
+
"learning_rate": 0.0003916437776601344,
|
| 2117 |
+
"loss": 6.3252,
|
| 2118 |
+
"step": 299
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.09598464245720685,
|
| 2122 |
+
"grad_norm": 9.128808975219727,
|
| 2123 |
+
"learning_rate": 0.00039158628044365476,
|
| 2124 |
+
"loss": 6.4524,
|
| 2125 |
+
"step": 300
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.09630459126539753,
|
| 2129 |
+
"grad_norm": 5.657589912414551,
|
| 2130 |
+
"learning_rate": 0.00039152859057094353,
|
| 2131 |
+
"loss": 6.3005,
|
| 2132 |
+
"step": 301
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.09662454007358823,
|
| 2136 |
+
"grad_norm": 8.320329666137695,
|
| 2137 |
+
"learning_rate": 0.0003914707081032074,
|
| 2138 |
+
"loss": 6.2912,
|
| 2139 |
+
"step": 302
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.09694448888177891,
|
| 2143 |
+
"grad_norm": 4.0212578773498535,
|
| 2144 |
+
"learning_rate": 0.00039141263310185783,
|
| 2145 |
+
"loss": 6.3545,
|
| 2146 |
+
"step": 303
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 0.0972644376899696,
|
| 2150 |
+
"grad_norm": 5.65709924697876,
|
| 2151 |
+
"learning_rate": 0.00039135436562850984,
|
| 2152 |
+
"loss": 6.2217,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 0.0975843864981603,
|
| 2157 |
+
"grad_norm": 6.352542400360107,
|
| 2158 |
+
"learning_rate": 0.00039129590574498325,
|
| 2159 |
+
"loss": 6.5012,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 0.09790433530635098,
|
| 2164 |
+
"grad_norm": 4.846847057342529,
|
| 2165 |
+
"learning_rate": 0.0003912372535133018,
|
| 2166 |
+
"loss": 6.2432,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 0.09822428411454168,
|
| 2171 |
+
"grad_norm": 4.925963401794434,
|
| 2172 |
+
"learning_rate": 0.0003911784089956931,
|
| 2173 |
+
"loss": 6.3315,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.09854423292273236,
|
| 2178 |
+
"grad_norm": 8.602291107177734,
|
| 2179 |
+
"learning_rate": 0.0003911193722545893,
|
| 2180 |
+
"loss": 6.4698,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.09886418173092305,
|
| 2185 |
+
"grad_norm": 6.059446811676025,
|
| 2186 |
+
"learning_rate": 0.00039106014335262587,
|
| 2187 |
+
"loss": 6.3634,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 0.09918413053911374,
|
| 2192 |
+
"grad_norm": 5.607949256896973,
|
| 2193 |
+
"learning_rate": 0.0003910007223526427,
|
| 2194 |
+
"loss": 6.1728,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 0.09950407934730443,
|
| 2199 |
+
"grad_norm": 6.827522277832031,
|
| 2200 |
+
"learning_rate": 0.0003909411093176831,
|
| 2201 |
+
"loss": 6.1584,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 0.09982402815549513,
|
| 2206 |
+
"grad_norm": 9.145511627197266,
|
| 2207 |
+
"learning_rate": 0.00039088130431099425,
|
| 2208 |
+
"loss": 6.6123,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.10014397696368581,
|
| 2213 |
+
"grad_norm": 8.265061378479004,
|
| 2214 |
+
"learning_rate": 0.00039082130739602716,
|
| 2215 |
+
"loss": 6.5472,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.1004639257718765,
|
| 2220 |
+
"grad_norm": 8.550250053405762,
|
| 2221 |
+
"learning_rate": 0.0003907611186364361,
|
| 2222 |
+
"loss": 6.2075,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.10078387458006718,
|
| 2227 |
+
"grad_norm": 7.309517860412598,
|
| 2228 |
+
"learning_rate": 0.0003907007380960792,
|
| 2229 |
+
"loss": 6.5685,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 0.10110382338825788,
|
| 2234 |
+
"grad_norm": 6.228018760681152,
|
| 2235 |
+
"learning_rate": 0.00039064016583901796,
|
| 2236 |
+
"loss": 6.3311,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 0.10142377219644858,
|
| 2241 |
+
"grad_norm": 8.36740493774414,
|
| 2242 |
+
"learning_rate": 0.0003905794019295173,
|
| 2243 |
+
"loss": 6.2873,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 0.10174372100463926,
|
| 2248 |
+
"grad_norm": 6.596613883972168,
|
| 2249 |
+
"learning_rate": 0.0003905184464320453,
|
| 2250 |
+
"loss": 6.3647,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 0.10206366981282995,
|
| 2255 |
+
"grad_norm": 7.987593650817871,
|
| 2256 |
+
"learning_rate": 0.00039045729941127364,
|
| 2257 |
+
"loss": 6.2189,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.10238361862102063,
|
| 2262 |
+
"grad_norm": 7.624737739562988,
|
| 2263 |
+
"learning_rate": 0.0003903959609320769,
|
| 2264 |
+
"loss": 6.4019,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.10270356742921133,
|
| 2269 |
+
"grad_norm": 6.667533874511719,
|
| 2270 |
+
"learning_rate": 0.0003903344310595329,
|
| 2271 |
+
"loss": 6.2599,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 0.10302351623740201,
|
| 2276 |
+
"grad_norm": 9.014090538024902,
|
| 2277 |
+
"learning_rate": 0.0003902727098589226,
|
| 2278 |
+
"loss": 6.4371,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 0.1033434650455927,
|
| 2283 |
+
"grad_norm": 6.295741558074951,
|
| 2284 |
+
"learning_rate": 0.0003902107973957299,
|
| 2285 |
+
"loss": 6.0999,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 0.1036634138537834,
|
| 2290 |
+
"grad_norm": 5.217776298522949,
|
| 2291 |
+
"learning_rate": 0.00039014869373564146,
|
| 2292 |
+
"loss": 6.1554,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 0.10398336266197408,
|
| 2297 |
+
"grad_norm": 8.615114212036133,
|
| 2298 |
+
"learning_rate": 0.0003900863989445471,
|
| 2299 |
+
"loss": 6.4223,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.10430331147016478,
|
| 2304 |
+
"grad_norm": 9.223677635192871,
|
| 2305 |
+
"learning_rate": 0.00039002391308853905,
|
| 2306 |
+
"loss": 6.2714,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.10462326027835546,
|
| 2311 |
+
"grad_norm": 8.394550323486328,
|
| 2312 |
+
"learning_rate": 0.00038996123623391267,
|
| 2313 |
+
"loss": 6.3747,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 0.10494320908654615,
|
| 2318 |
+
"grad_norm": 8.384407043457031,
|
| 2319 |
+
"learning_rate": 0.00038989836844716564,
|
| 2320 |
+
"loss": 6.2019,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 0.10526315789473684,
|
| 2325 |
+
"grad_norm": 6.654740333557129,
|
| 2326 |
+
"learning_rate": 0.00038983530979499835,
|
| 2327 |
+
"loss": 6.2863,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 0.10558310670292753,
|
| 2332 |
+
"grad_norm": 5.682619571685791,
|
| 2333 |
+
"learning_rate": 0.00038977206034431373,
|
| 2334 |
+
"loss": 6.2633,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 0.10590305551111823,
|
| 2339 |
+
"grad_norm": 6.896644592285156,
|
| 2340 |
+
"learning_rate": 0.0003897086201622169,
|
| 2341 |
+
"loss": 6.3236,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.10622300431930891,
|
| 2346 |
+
"grad_norm": 6.741705894470215,
|
| 2347 |
+
"learning_rate": 0.0003896449893160157,
|
| 2348 |
+
"loss": 6.1388,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.1065429531274996,
|
| 2353 |
+
"grad_norm": 7.0619215965271,
|
| 2354 |
+
"learning_rate": 0.00038958116787321995,
|
| 2355 |
+
"loss": 6.3887,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 0.10686290193569029,
|
| 2360 |
+
"grad_norm": 5.026659965515137,
|
| 2361 |
+
"learning_rate": 0.0003895171559015418,
|
| 2362 |
+
"loss": 6.2476,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.10718285074388098,
|
| 2367 |
+
"grad_norm": 6.366108417510986,
|
| 2368 |
+
"learning_rate": 0.00038945295346889567,
|
| 2369 |
+
"loss": 6.4004,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 0.10750279955207166,
|
| 2374 |
+
"grad_norm": 6.229933738708496,
|
| 2375 |
+
"learning_rate": 0.0003893885606433978,
|
| 2376 |
+
"loss": 6.4047,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 0.10782274836026236,
|
| 2381 |
+
"grad_norm": 5.269782543182373,
|
| 2382 |
+
"learning_rate": 0.0003893239774933667,
|
| 2383 |
+
"loss": 6.2193,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.10814269716845305,
|
| 2388 |
+
"grad_norm": 5.160742282867432,
|
| 2389 |
+
"learning_rate": 0.00038925920408732254,
|
| 2390 |
+
"loss": 6.3671,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.10846264597664373,
|
| 2395 |
+
"grad_norm": 6.694570064544678,
|
| 2396 |
+
"learning_rate": 0.00038919424049398756,
|
| 2397 |
+
"loss": 6.18,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 0.10878259478483443,
|
| 2402 |
+
"grad_norm": 7.192866325378418,
|
| 2403 |
+
"learning_rate": 0.00038912908678228556,
|
| 2404 |
+
"loss": 6.1159,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.10910254359302511,
|
| 2409 |
+
"grad_norm": 7.5637688636779785,
|
| 2410 |
+
"learning_rate": 0.00038906374302134235,
|
| 2411 |
+
"loss": 6.1341,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 0.1094224924012158,
|
| 2416 |
+
"grad_norm": 10.56783676147461,
|
| 2417 |
+
"learning_rate": 0.00038899820928048524,
|
| 2418 |
+
"loss": 6.1261,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.10974244120940649,
|
| 2423 |
+
"grad_norm": 6.504526615142822,
|
| 2424 |
+
"learning_rate": 0.00038893248562924286,
|
| 2425 |
+
"loss": 6.157,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.11006239001759718,
|
| 2430 |
+
"grad_norm": 7.0983757972717285,
|
| 2431 |
+
"learning_rate": 0.00038886657213734567,
|
| 2432 |
+
"loss": 6.3047,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.11038233882578788,
|
| 2437 |
+
"grad_norm": 12.745441436767578,
|
| 2438 |
+
"learning_rate": 0.00038880046887472547,
|
| 2439 |
+
"loss": 6.2377,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 0.11070228763397856,
|
| 2444 |
+
"grad_norm": 9.568143844604492,
|
| 2445 |
+
"learning_rate": 0.0003887341759115151,
|
| 2446 |
+
"loss": 6.3565,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.11102223644216926,
|
| 2451 |
+
"grad_norm": 9.615344047546387,
|
| 2452 |
+
"learning_rate": 0.0003886676933180493,
|
| 2453 |
+
"loss": 6.2896,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 0.11134218525035994,
|
| 2458 |
+
"grad_norm": 8.619441032409668,
|
| 2459 |
+
"learning_rate": 0.0003886010211648632,
|
| 2460 |
+
"loss": 6.2012,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.11166213405855063,
|
| 2465 |
+
"grad_norm": 13.255605697631836,
|
| 2466 |
+
"learning_rate": 0.00038853415952269367,
|
| 2467 |
+
"loss": 6.4429,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.11198208286674133,
|
| 2472 |
+
"grad_norm": 9.434603691101074,
|
| 2473 |
+
"learning_rate": 0.00038846710846247845,
|
| 2474 |
+
"loss": 6.1695,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.11230203167493201,
|
| 2479 |
+
"grad_norm": 7.766512870788574,
|
| 2480 |
+
"learning_rate": 0.00038839986805535605,
|
| 2481 |
+
"loss": 6.2969,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.1126219804831227,
|
| 2486 |
+
"grad_norm": 5.024114608764648,
|
| 2487 |
+
"learning_rate": 0.00038833243837266615,
|
| 2488 |
+
"loss": 6.3783,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.11294192929131339,
|
| 2493 |
+
"grad_norm": 6.237715721130371,
|
| 2494 |
+
"learning_rate": 0.00038826481948594897,
|
| 2495 |
+
"loss": 6.1483,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.11326187809950408,
|
| 2500 |
+
"grad_norm": 7.264336109161377,
|
| 2501 |
+
"learning_rate": 0.00038819701146694573,
|
| 2502 |
+
"loss": 6.1925,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.11358182690769476,
|
| 2507 |
+
"grad_norm": 7.290820121765137,
|
| 2508 |
+
"learning_rate": 0.0003881290143875981,
|
| 2509 |
+
"loss": 6.3151,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.11390177571588546,
|
| 2514 |
+
"grad_norm": 5.954084873199463,
|
| 2515 |
+
"learning_rate": 0.0003880608283200486,
|
| 2516 |
+
"loss": 6.181,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.11422172452407615,
|
| 2521 |
+
"grad_norm": 7.405073165893555,
|
| 2522 |
+
"learning_rate": 0.0003879924533366399,
|
| 2523 |
+
"loss": 6.2829,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.11454167333226684,
|
| 2528 |
+
"grad_norm": 6.803983688354492,
|
| 2529 |
+
"learning_rate": 0.0003879238895099154,
|
| 2530 |
+
"loss": 6.3297,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.11486162214045753,
|
| 2535 |
+
"grad_norm": 8.649042129516602,
|
| 2536 |
+
"learning_rate": 0.00038785513691261877,
|
| 2537 |
+
"loss": 6.2688,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.11518157094864821,
|
| 2542 |
+
"grad_norm": 6.858957767486572,
|
| 2543 |
+
"learning_rate": 0.0003877861956176938,
|
| 2544 |
+
"loss": 6.2604,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.11550151975683891,
|
| 2549 |
+
"grad_norm": 7.34728479385376,
|
| 2550 |
+
"learning_rate": 0.00038771706569828475,
|
| 2551 |
+
"loss": 6.2437,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.11582146856502959,
|
| 2556 |
+
"grad_norm": 6.574677467346191,
|
| 2557 |
+
"learning_rate": 0.00038764774722773577,
|
| 2558 |
+
"loss": 6.113,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.11614141737322028,
|
| 2563 |
+
"grad_norm": 9.14777660369873,
|
| 2564 |
+
"learning_rate": 0.00038757824027959133,
|
| 2565 |
+
"loss": 6.2634,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.11646136618141098,
|
| 2570 |
+
"grad_norm": 6.340327262878418,
|
| 2571 |
+
"learning_rate": 0.00038750854492759546,
|
| 2572 |
+
"loss": 6.1939,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.11678131498960166,
|
| 2577 |
+
"grad_norm": 6.838183879852295,
|
| 2578 |
+
"learning_rate": 0.0003874386612456925,
|
| 2579 |
+
"loss": 6.1019,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.11710126379779236,
|
| 2584 |
+
"grad_norm": 9.026721954345703,
|
| 2585 |
+
"learning_rate": 0.0003873685893080263,
|
| 2586 |
+
"loss": 6.1435,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.11742121260598304,
|
| 2591 |
+
"grad_norm": 5.471182823181152,
|
| 2592 |
+
"learning_rate": 0.0003872983291889406,
|
| 2593 |
+
"loss": 6.1873,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.11774116141417373,
|
| 2598 |
+
"grad_norm": 44.7470588684082,
|
| 2599 |
+
"learning_rate": 0.0003872278809629788,
|
| 2600 |
+
"loss": 6.318,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.11806111022236442,
|
| 2605 |
+
"grad_norm": 11.135316848754883,
|
| 2606 |
+
"learning_rate": 0.00038715724470488375,
|
| 2607 |
+
"loss": 6.1779,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.11838105903055511,
|
| 2612 |
+
"grad_norm": 6.037003517150879,
|
| 2613 |
+
"learning_rate": 0.00038708642048959787,
|
| 2614 |
+
"loss": 6.1901,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.1187010078387458,
|
| 2619 |
+
"grad_norm": 7.986204147338867,
|
| 2620 |
+
"learning_rate": 0.0003870154083922631,
|
| 2621 |
+
"loss": 6.3507,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.11902095664693649,
|
| 2626 |
+
"grad_norm": 6.852958679199219,
|
| 2627 |
+
"learning_rate": 0.00038694420848822046,
|
| 2628 |
+
"loss": 6.3366,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.11934090545512718,
|
| 2633 |
+
"grad_norm": 7.698732376098633,
|
| 2634 |
+
"learning_rate": 0.0003868728208530105,
|
| 2635 |
+
"loss": 6.2442,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.11966085426331786,
|
| 2640 |
+
"grad_norm": 8.554845809936523,
|
| 2641 |
+
"learning_rate": 0.0003868012455623728,
|
| 2642 |
+
"loss": 6.1877,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.11998080307150856,
|
| 2647 |
+
"grad_norm": 8.99901008605957,
|
| 2648 |
+
"learning_rate": 0.0003867294826922459,
|
| 2649 |
+
"loss": 6.2821,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.12030075187969924,
|
| 2654 |
+
"grad_norm": 8.72900676727295,
|
| 2655 |
+
"learning_rate": 0.0003866575323187678,
|
| 2656 |
+
"loss": 6.1797,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.12062070068788994,
|
| 2661 |
+
"grad_norm": 5.478520393371582,
|
| 2662 |
+
"learning_rate": 0.00038658539451827493,
|
| 2663 |
+
"loss": 6.3879,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.12094064949608063,
|
| 2668 |
+
"grad_norm": 7.450143814086914,
|
| 2669 |
+
"learning_rate": 0.0003865130693673028,
|
| 2670 |
+
"loss": 6.2649,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.12126059830427131,
|
| 2675 |
+
"grad_norm": 9.570409774780273,
|
| 2676 |
+
"learning_rate": 0.00038644055694258585,
|
| 2677 |
+
"loss": 6.2301,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.12158054711246201,
|
| 2682 |
+
"grad_norm": 6.57025146484375,
|
| 2683 |
+
"learning_rate": 0.00038636785732105685,
|
| 2684 |
+
"loss": 6.3892,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.12190049592065269,
|
| 2689 |
+
"grad_norm": 6.987082481384277,
|
| 2690 |
+
"learning_rate": 0.0003862949705798475,
|
| 2691 |
+
"loss": 5.8789,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.12222044472884339,
|
| 2696 |
+
"grad_norm": 7.735273838043213,
|
| 2697 |
+
"learning_rate": 0.00038622189679628774,
|
| 2698 |
+
"loss": 6.2546,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.12254039353703408,
|
| 2703 |
+
"grad_norm": 9.814201354980469,
|
| 2704 |
+
"learning_rate": 0.00038614863604790635,
|
| 2705 |
+
"loss": 6.6222,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.12286034234522476,
|
| 2710 |
+
"grad_norm": 9.477630615234375,
|
| 2711 |
+
"learning_rate": 0.00038607518841243,
|
| 2712 |
+
"loss": 6.457,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.12318029115341546,
|
| 2717 |
+
"grad_norm": 7.4210333824157715,
|
| 2718 |
+
"learning_rate": 0.00038600155396778405,
|
| 2719 |
+
"loss": 6.0636,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.12350023996160614,
|
| 2724 |
+
"grad_norm": 8.321981430053711,
|
| 2725 |
+
"learning_rate": 0.0003859277327920918,
|
| 2726 |
+
"loss": 6.1512,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.12382018876979684,
|
| 2731 |
+
"grad_norm": 7.436471939086914,
|
| 2732 |
+
"learning_rate": 0.0003858537249636748,
|
| 2733 |
+
"loss": 6.255,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.12414013757798752,
|
| 2738 |
+
"grad_norm": 8.560404777526855,
|
| 2739 |
+
"learning_rate": 0.0003857795305610525,
|
| 2740 |
+
"loss": 6.3151,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.12446008638617821,
|
| 2745 |
+
"grad_norm": 10.054532051086426,
|
| 2746 |
+
"learning_rate": 0.00038570514966294243,
|
| 2747 |
+
"loss": 6.2584,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.12478003519436891,
|
| 2752 |
+
"grad_norm": 9.69568157196045,
|
| 2753 |
+
"learning_rate": 0.0003856305823482599,
|
| 2754 |
+
"loss": 6.4013,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.1250999840025596,
|
| 2759 |
+
"grad_norm": 9.335556030273438,
|
| 2760 |
+
"learning_rate": 0.00038555582869611807,
|
| 2761 |
+
"loss": 6.4033,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.12541993281075028,
|
| 2766 |
+
"grad_norm": 8.94521427154541,
|
| 2767 |
+
"learning_rate": 0.00038548088878582776,
|
| 2768 |
+
"loss": 6.4043,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.12573988161894098,
|
| 2773 |
+
"grad_norm": 11.994088172912598,
|
| 2774 |
+
"learning_rate": 0.0003854057626968974,
|
| 2775 |
+
"loss": 6.2114,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.12605983042713165,
|
| 2780 |
+
"grad_norm": 6.300990581512451,
|
| 2781 |
+
"learning_rate": 0.000385330450509033,
|
| 2782 |
+
"loss": 6.2302,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.12637977923532234,
|
| 2787 |
+
"grad_norm": 16.85110092163086,
|
| 2788 |
+
"learning_rate": 0.00038525495230213783,
|
| 2789 |
+
"loss": 6.0428,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.12669972804351304,
|
| 2794 |
+
"grad_norm": 9.76857852935791,
|
| 2795 |
+
"learning_rate": 0.0003851792681563127,
|
| 2796 |
+
"loss": 6.486,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.12701967685170373,
|
| 2801 |
+
"grad_norm": 9.043763160705566,
|
| 2802 |
+
"learning_rate": 0.0003851033981518557,
|
| 2803 |
+
"loss": 6.3237,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.12733962565989443,
|
| 2808 |
+
"grad_norm": 16.929244995117188,
|
| 2809 |
+
"learning_rate": 0.00038502734236926214,
|
| 2810 |
+
"loss": 6.3279,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.1276595744680851,
|
| 2815 |
+
"grad_norm": 19.378320693969727,
|
| 2816 |
+
"learning_rate": 0.00038495110088922413,
|
| 2817 |
+
"loss": 6.2533,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.1279795232762758,
|
| 2822 |
+
"grad_norm": 7.617816925048828,
|
| 2823 |
+
"learning_rate": 0.00038487467379263123,
|
| 2824 |
+
"loss": 6.5729,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.1279795232762758,
|
| 2829 |
+
"eval_loss": 3.206190586090088,
|
| 2830 |
+
"eval_runtime": 233.5688,
|
| 2831 |
+
"eval_samples_per_second": 5.634,
|
| 2832 |
+
"eval_steps_per_second": 1.409,
|
| 2833 |
+
"step": 400
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
+
"total_flos": 3.666932398030848e+17,
|
| 2863 |
"train_batch_size": 4,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|