Training in progress, step 600, 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 335604696
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd78810ad4d15dc3e5e9d04d1d5e37be079aede5a77524ad8c01c88c8c235dff
|
| 3 |
size 335604696
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 90187222
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5755833c9f369fe2eca4af1f5675c5465163c7cf76a87e8f9bcdd16bbe7f0bf
|
| 3 |
size 90187222
|
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 sha256:7e98f5f761f59bb7a4807bc8883a9ae631b57b607bd3c4d8718c69b40cd94951
|
| 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 2080
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b048f6c22f0d29e3a712f7faacb96fd7a64aa0b2a7516919663f205dade71ddc
|
| 3 |
size 2080
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4815335273742676,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 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,
|
|
@@ -2831,6 +2831,1414 @@
|
|
| 2831 |
"eval_samples_per_second": 5.118,
|
| 2832 |
"eval_steps_per_second": 2.564,
|
| 2833 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
@@ -2845,7 +4253,7 @@
|
|
| 2845 |
"early_stopping_threshold": 0.0
|
| 2846 |
},
|
| 2847 |
"attributes": {
|
| 2848 |
-
"early_stopping_patience_counter":
|
| 2849 |
}
|
| 2850 |
},
|
| 2851 |
"TrainerControl": {
|
|
@@ -2859,7 +4267,7 @@
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
-
"total_flos":
|
| 2863 |
"train_batch_size": 2,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4815335273742676,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.047953006054067014,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2831 |
"eval_samples_per_second": 5.118,
|
| 2832 |
"eval_steps_per_second": 2.564,
|
| 2833 |
"step": 400
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.032048592379468124,
|
| 2837 |
+
"grad_norm": 12.839659690856934,
|
| 2838 |
+
"learning_rate": 0.0001999517116928473,
|
| 2839 |
+
"loss": 5.5085,
|
| 2840 |
+
"step": 401
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.0321285140562249,
|
| 2844 |
+
"grad_norm": 13.675345420837402,
|
| 2845 |
+
"learning_rate": 0.0001999514497583732,
|
| 2846 |
+
"loss": 5.9345,
|
| 2847 |
+
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.032208435732981676,
|
| 2851 |
+
"grad_norm": 19.015840530395508,
|
| 2852 |
+
"learning_rate": 0.0001999511878238991,
|
| 2853 |
+
"loss": 6.3208,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.032288357409738455,
|
| 2858 |
+
"grad_norm": 17.023345947265625,
|
| 2859 |
+
"learning_rate": 0.00019995092588942498,
|
| 2860 |
+
"loss": 6.2015,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.032368279086495234,
|
| 2865 |
+
"grad_norm": 19.328624725341797,
|
| 2866 |
+
"learning_rate": 0.00019995066395495087,
|
| 2867 |
+
"loss": 6.8056,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.032448200763252014,
|
| 2872 |
+
"grad_norm": 14.101640701293945,
|
| 2873 |
+
"learning_rate": 0.00019995040202047676,
|
| 2874 |
+
"loss": 5.2164,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.03252812244000879,
|
| 2879 |
+
"grad_norm": 16.194459915161133,
|
| 2880 |
+
"learning_rate": 0.00019995014008600265,
|
| 2881 |
+
"loss": 6.0302,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.03260804411676557,
|
| 2886 |
+
"grad_norm": 11.973613739013672,
|
| 2887 |
+
"learning_rate": 0.00019994987815152854,
|
| 2888 |
+
"loss": 5.9626,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.03268796579352235,
|
| 2893 |
+
"grad_norm": 11.992119789123535,
|
| 2894 |
+
"learning_rate": 0.0001999496016651392,
|
| 2895 |
+
"loss": 6.5503,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.032767887470279124,
|
| 2900 |
+
"grad_norm": 26.241275787353516,
|
| 2901 |
+
"learning_rate": 0.0001999493397306651,
|
| 2902 |
+
"loss": 5.8849,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.0328478091470359,
|
| 2907 |
+
"grad_norm": 17.755863189697266,
|
| 2908 |
+
"learning_rate": 0.00019994907779619098,
|
| 2909 |
+
"loss": 6.5457,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.03292773082379268,
|
| 2914 |
+
"grad_norm": 15.278813362121582,
|
| 2915 |
+
"learning_rate": 0.00019994880130980164,
|
| 2916 |
+
"loss": 6.5778,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.03300765250054946,
|
| 2921 |
+
"grad_norm": 14.57132625579834,
|
| 2922 |
+
"learning_rate": 0.00019994853937532753,
|
| 2923 |
+
"loss": 4.8398,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.03308757417730624,
|
| 2928 |
+
"grad_norm": 13.570646286010742,
|
| 2929 |
+
"learning_rate": 0.0001999482628889382,
|
| 2930 |
+
"loss": 5.6689,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.03316749585406302,
|
| 2935 |
+
"grad_norm": 19.012624740600586,
|
| 2936 |
+
"learning_rate": 0.00019994800095446408,
|
| 2937 |
+
"loss": 6.3883,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.0332474175308198,
|
| 2942 |
+
"grad_norm": 17.407018661499023,
|
| 2943 |
+
"learning_rate": 0.00019994773901998997,
|
| 2944 |
+
"loss": 6.6187,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.03332733920757658,
|
| 2949 |
+
"grad_norm": 21.998416900634766,
|
| 2950 |
+
"learning_rate": 0.00019994746253360063,
|
| 2951 |
+
"loss": 6.0688,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.03340726088433335,
|
| 2956 |
+
"grad_norm": 18.732925415039062,
|
| 2957 |
+
"learning_rate": 0.0001999471860472113,
|
| 2958 |
+
"loss": 6.0737,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.03348718256109013,
|
| 2963 |
+
"grad_norm": 14.823393821716309,
|
| 2964 |
+
"learning_rate": 0.00019994690956082195,
|
| 2965 |
+
"loss": 5.4123,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.03356710423784691,
|
| 2970 |
+
"grad_norm": 14.061379432678223,
|
| 2971 |
+
"learning_rate": 0.0001999466330744326,
|
| 2972 |
+
"loss": 6.2029,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.03364702591460369,
|
| 2977 |
+
"grad_norm": 24.457435607910156,
|
| 2978 |
+
"learning_rate": 0.00019994635658804327,
|
| 2979 |
+
"loss": 6.0696,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.03372694759136047,
|
| 2984 |
+
"grad_norm": 13.974810600280762,
|
| 2985 |
+
"learning_rate": 0.00019994609465356916,
|
| 2986 |
+
"loss": 5.446,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.03380686926811725,
|
| 2991 |
+
"grad_norm": 15.828372955322266,
|
| 2992 |
+
"learning_rate": 0.00019994581816717982,
|
| 2993 |
+
"loss": 5.3473,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.033886790944874026,
|
| 2998 |
+
"grad_norm": 11.220956802368164,
|
| 2999 |
+
"learning_rate": 0.00019994554168079048,
|
| 3000 |
+
"loss": 5.3355,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.0339667126216308,
|
| 3005 |
+
"grad_norm": 19.321090698242188,
|
| 3006 |
+
"learning_rate": 0.00019994526519440114,
|
| 3007 |
+
"loss": 5.3334,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.03404663429838758,
|
| 3012 |
+
"grad_norm": 12.514566421508789,
|
| 3013 |
+
"learning_rate": 0.0001999449887080118,
|
| 3014 |
+
"loss": 5.5341,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.03412655597514436,
|
| 3019 |
+
"grad_norm": 18.08624839782715,
|
| 3020 |
+
"learning_rate": 0.00019994471222162247,
|
| 3021 |
+
"loss": 5.5019,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.034206477651901136,
|
| 3026 |
+
"grad_norm": 13.24229907989502,
|
| 3027 |
+
"learning_rate": 0.00019994443573523313,
|
| 3028 |
+
"loss": 6.4423,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.034286399328657916,
|
| 3033 |
+
"grad_norm": 16.371517181396484,
|
| 3034 |
+
"learning_rate": 0.00019994414469692856,
|
| 3035 |
+
"loss": 6.4045,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.034366321005414695,
|
| 3040 |
+
"grad_norm": 14.142496109008789,
|
| 3041 |
+
"learning_rate": 0.00019994386821053922,
|
| 3042 |
+
"loss": 5.3599,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.034446242682171474,
|
| 3047 |
+
"grad_norm": 18.488370895385742,
|
| 3048 |
+
"learning_rate": 0.00019994359172414988,
|
| 3049 |
+
"loss": 6.7911,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.03452616435892825,
|
| 3054 |
+
"grad_norm": 14.886078834533691,
|
| 3055 |
+
"learning_rate": 0.00019994331523776054,
|
| 3056 |
+
"loss": 6.5936,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.034606086035685026,
|
| 3061 |
+
"grad_norm": 17.963350296020508,
|
| 3062 |
+
"learning_rate": 0.00019994302419945598,
|
| 3063 |
+
"loss": 5.4377,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.034686007712441805,
|
| 3068 |
+
"grad_norm": 16.43907356262207,
|
| 3069 |
+
"learning_rate": 0.0001999427331611514,
|
| 3070 |
+
"loss": 6.9432,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.034765929389198584,
|
| 3075 |
+
"grad_norm": 12.792582511901855,
|
| 3076 |
+
"learning_rate": 0.0001999424712266773,
|
| 3077 |
+
"loss": 5.0474,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.03484585106595536,
|
| 3082 |
+
"grad_norm": 11.655512809753418,
|
| 3083 |
+
"learning_rate": 0.00019994218018837273,
|
| 3084 |
+
"loss": 6.0505,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.03492577274271214,
|
| 3089 |
+
"grad_norm": 14.01752758026123,
|
| 3090 |
+
"learning_rate": 0.00019994188915006816,
|
| 3091 |
+
"loss": 4.9665,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.03500569441946892,
|
| 3096 |
+
"grad_norm": 16.40753746032715,
|
| 3097 |
+
"learning_rate": 0.00019994161266367882,
|
| 3098 |
+
"loss": 6.9427,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.0350856160962257,
|
| 3103 |
+
"grad_norm": 17.867719650268555,
|
| 3104 |
+
"learning_rate": 0.00019994132162537426,
|
| 3105 |
+
"loss": 6.7046,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.03516553777298248,
|
| 3110 |
+
"grad_norm": 21.779308319091797,
|
| 3111 |
+
"learning_rate": 0.0001999410305870697,
|
| 3112 |
+
"loss": 5.7596,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.03524545944973925,
|
| 3117 |
+
"grad_norm": 13.966331481933594,
|
| 3118 |
+
"learning_rate": 0.00019994075410068035,
|
| 3119 |
+
"loss": 5.4575,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.03532538112649603,
|
| 3124 |
+
"grad_norm": 18.35555648803711,
|
| 3125 |
+
"learning_rate": 0.00019994044851046056,
|
| 3126 |
+
"loss": 5.6826,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.03540530280325281,
|
| 3131 |
+
"grad_norm": 14.487837791442871,
|
| 3132 |
+
"learning_rate": 0.000199940157472156,
|
| 3133 |
+
"loss": 5.888,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.03548522448000959,
|
| 3138 |
+
"grad_norm": 13.87781047821045,
|
| 3139 |
+
"learning_rate": 0.00019993988098576665,
|
| 3140 |
+
"loss": 5.6784,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.03556514615676637,
|
| 3145 |
+
"grad_norm": 21.00891876220703,
|
| 3146 |
+
"learning_rate": 0.00019993957539554685,
|
| 3147 |
+
"loss": 6.3669,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.03564506783352315,
|
| 3152 |
+
"grad_norm": 13.473644256591797,
|
| 3153 |
+
"learning_rate": 0.00019993929890915751,
|
| 3154 |
+
"loss": 5.8704,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.03572498951027993,
|
| 3159 |
+
"grad_norm": 16.605884552001953,
|
| 3160 |
+
"learning_rate": 0.00019993899331893772,
|
| 3161 |
+
"loss": 6.4597,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.03580491118703671,
|
| 3166 |
+
"grad_norm": 24.377735137939453,
|
| 3167 |
+
"learning_rate": 0.00019993870228063315,
|
| 3168 |
+
"loss": 5.6275,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.03588483286379348,
|
| 3173 |
+
"grad_norm": 12.248327255249023,
|
| 3174 |
+
"learning_rate": 0.00019993841124232858,
|
| 3175 |
+
"loss": 4.4804,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.03596475454055026,
|
| 3180 |
+
"grad_norm": 15.635926246643066,
|
| 3181 |
+
"learning_rate": 0.00019993812020402402,
|
| 3182 |
+
"loss": 5.4341,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.03604467621730704,
|
| 3187 |
+
"grad_norm": 13.867622375488281,
|
| 3188 |
+
"learning_rate": 0.00019993782916571945,
|
| 3189 |
+
"loss": 5.4284,
|
| 3190 |
+
"step": 451
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 0.03612459789406382,
|
| 3194 |
+
"grad_norm": 20.617046356201172,
|
| 3195 |
+
"learning_rate": 0.00019993752357549965,
|
| 3196 |
+
"loss": 6.6966,
|
| 3197 |
+
"step": 452
|
| 3198 |
+
},
|
| 3199 |
+
{
|
| 3200 |
+
"epoch": 0.0362045195708206,
|
| 3201 |
+
"grad_norm": 21.450973510742188,
|
| 3202 |
+
"learning_rate": 0.0001999372325371951,
|
| 3203 |
+
"loss": 5.9473,
|
| 3204 |
+
"step": 453
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 0.036284441247577376,
|
| 3208 |
+
"grad_norm": 19.491113662719727,
|
| 3209 |
+
"learning_rate": 0.0001999369269469753,
|
| 3210 |
+
"loss": 6.7066,
|
| 3211 |
+
"step": 454
|
| 3212 |
+
},
|
| 3213 |
+
{
|
| 3214 |
+
"epoch": 0.036364362924334155,
|
| 3215 |
+
"grad_norm": 23.190990447998047,
|
| 3216 |
+
"learning_rate": 0.00019993663590867072,
|
| 3217 |
+
"loss": 6.5902,
|
| 3218 |
+
"step": 455
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 0.03644428460109093,
|
| 3222 |
+
"grad_norm": 15.78524112701416,
|
| 3223 |
+
"learning_rate": 0.00019993633031845093,
|
| 3224 |
+
"loss": 5.5508,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 0.03652420627784771,
|
| 3229 |
+
"grad_norm": 15.83521556854248,
|
| 3230 |
+
"learning_rate": 0.00019993603928014636,
|
| 3231 |
+
"loss": 6.7382,
|
| 3232 |
+
"step": 457
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 0.036604127954604486,
|
| 3236 |
+
"grad_norm": 14.295878410339355,
|
| 3237 |
+
"learning_rate": 0.0001999357482418418,
|
| 3238 |
+
"loss": 6.157,
|
| 3239 |
+
"step": 458
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.036684049631361265,
|
| 3243 |
+
"grad_norm": 14.773226737976074,
|
| 3244 |
+
"learning_rate": 0.000199935442651622,
|
| 3245 |
+
"loss": 5.3554,
|
| 3246 |
+
"step": 459
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.036763971308118044,
|
| 3250 |
+
"grad_norm": 17.491086959838867,
|
| 3251 |
+
"learning_rate": 0.0001999351370614022,
|
| 3252 |
+
"loss": 5.3601,
|
| 3253 |
+
"step": 460
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.036843892984874824,
|
| 3257 |
+
"grad_norm": 10.782079696655273,
|
| 3258 |
+
"learning_rate": 0.0001999348314711824,
|
| 3259 |
+
"loss": 5.5249,
|
| 3260 |
+
"step": 461
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.0369238146616316,
|
| 3264 |
+
"grad_norm": 18.85950469970703,
|
| 3265 |
+
"learning_rate": 0.00019993454043287784,
|
| 3266 |
+
"loss": 6.0112,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.03700373633838838,
|
| 3271 |
+
"grad_norm": 26.688846588134766,
|
| 3272 |
+
"learning_rate": 0.00019993422029074281,
|
| 3273 |
+
"loss": 4.8648,
|
| 3274 |
+
"step": 463
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.037083658015145154,
|
| 3278 |
+
"grad_norm": 19.976411819458008,
|
| 3279 |
+
"learning_rate": 0.00019993392925243825,
|
| 3280 |
+
"loss": 7.1799,
|
| 3281 |
+
"step": 464
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.037163579691901934,
|
| 3285 |
+
"grad_norm": 13.337353706359863,
|
| 3286 |
+
"learning_rate": 0.00019993362366221845,
|
| 3287 |
+
"loss": 5.3017,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.03724350136865871,
|
| 3292 |
+
"grad_norm": 20.273609161376953,
|
| 3293 |
+
"learning_rate": 0.00019993331807199866,
|
| 3294 |
+
"loss": 7.0964,
|
| 3295 |
+
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.03732342304541549,
|
| 3299 |
+
"grad_norm": 15.535438537597656,
|
| 3300 |
+
"learning_rate": 0.00019993301248177886,
|
| 3301 |
+
"loss": 6.1953,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.03740334472217227,
|
| 3306 |
+
"grad_norm": 20.07297134399414,
|
| 3307 |
+
"learning_rate": 0.00019993270689155906,
|
| 3308 |
+
"loss": 6.053,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.03748326639892905,
|
| 3313 |
+
"grad_norm": 31.530681610107422,
|
| 3314 |
+
"learning_rate": 0.00019993238674942404,
|
| 3315 |
+
"loss": 6.2682,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.03756318807568583,
|
| 3320 |
+
"grad_norm": 18.68016815185547,
|
| 3321 |
+
"learning_rate": 0.00019993208115920424,
|
| 3322 |
+
"loss": 6.9533,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.03764310975244261,
|
| 3327 |
+
"grad_norm": 18.137269973754883,
|
| 3328 |
+
"learning_rate": 0.00019993177556898445,
|
| 3329 |
+
"loss": 6.4924,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.03772303142919938,
|
| 3334 |
+
"grad_norm": 13.506487846374512,
|
| 3335 |
+
"learning_rate": 0.00019993146997876465,
|
| 3336 |
+
"loss": 4.9281,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.03780295310595616,
|
| 3341 |
+
"grad_norm": 16.436418533325195,
|
| 3342 |
+
"learning_rate": 0.00019993114983662963,
|
| 3343 |
+
"loss": 6.4321,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.03788287478271294,
|
| 3348 |
+
"grad_norm": 17.684978485107422,
|
| 3349 |
+
"learning_rate": 0.00019993084424640983,
|
| 3350 |
+
"loss": 5.8878,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.03796279645946972,
|
| 3355 |
+
"grad_norm": 19.21597671508789,
|
| 3356 |
+
"learning_rate": 0.00019993053865619004,
|
| 3357 |
+
"loss": 6.56,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.0380427181362265,
|
| 3362 |
+
"grad_norm": 26.484188079833984,
|
| 3363 |
+
"learning_rate": 0.00019993021851405501,
|
| 3364 |
+
"loss": 6.7126,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.03812263981298328,
|
| 3369 |
+
"grad_norm": 17.94887351989746,
|
| 3370 |
+
"learning_rate": 0.00019992991292383522,
|
| 3371 |
+
"loss": 6.3713,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.03820256148974006,
|
| 3376 |
+
"grad_norm": 18.28172492980957,
|
| 3377 |
+
"learning_rate": 0.0001999295927817002,
|
| 3378 |
+
"loss": 6.2604,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.038282483166496836,
|
| 3383 |
+
"grad_norm": 23.329925537109375,
|
| 3384 |
+
"learning_rate": 0.0001999292871914804,
|
| 3385 |
+
"loss": 5.1182,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.03836240484325361,
|
| 3390 |
+
"grad_norm": 15.076399803161621,
|
| 3391 |
+
"learning_rate": 0.00019992896704934537,
|
| 3392 |
+
"loss": 6.61,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.03844232652001039,
|
| 3397 |
+
"grad_norm": 18.058076858520508,
|
| 3398 |
+
"learning_rate": 0.00019992864690721035,
|
| 3399 |
+
"loss": 5.9042,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.03852224819676717,
|
| 3404 |
+
"grad_norm": 17.3637638092041,
|
| 3405 |
+
"learning_rate": 0.00019992832676507533,
|
| 3406 |
+
"loss": 5.6953,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.038602169873523946,
|
| 3411 |
+
"grad_norm": 13.11391830444336,
|
| 3412 |
+
"learning_rate": 0.00019992802117485553,
|
| 3413 |
+
"loss": 6.0456,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.038682091550280726,
|
| 3418 |
+
"grad_norm": 14.464241981506348,
|
| 3419 |
+
"learning_rate": 0.0001999277010327205,
|
| 3420 |
+
"loss": 6.1571,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.038762013227037505,
|
| 3425 |
+
"grad_norm": 15.715449333190918,
|
| 3426 |
+
"learning_rate": 0.00019992736633867025,
|
| 3427 |
+
"loss": 6.2516,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.038841934903794284,
|
| 3432 |
+
"grad_norm": 18.621906280517578,
|
| 3433 |
+
"learning_rate": 0.00019992706074845046,
|
| 3434 |
+
"loss": 5.9231,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.03892185658055106,
|
| 3439 |
+
"grad_norm": 19.09491729736328,
|
| 3440 |
+
"learning_rate": 0.00019992674060631543,
|
| 3441 |
+
"loss": 6.7979,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.039001778257307836,
|
| 3446 |
+
"grad_norm": 18.09942626953125,
|
| 3447 |
+
"learning_rate": 0.0001999264204641804,
|
| 3448 |
+
"loss": 5.6107,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.039081699934064615,
|
| 3453 |
+
"grad_norm": 19.745094299316406,
|
| 3454 |
+
"learning_rate": 0.00019992610032204539,
|
| 3455 |
+
"loss": 6.2667,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.039161621610821394,
|
| 3460 |
+
"grad_norm": 12.84618091583252,
|
| 3461 |
+
"learning_rate": 0.00019992578017991036,
|
| 3462 |
+
"loss": 5.5015,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.03924154328757817,
|
| 3467 |
+
"grad_norm": 18.139814376831055,
|
| 3468 |
+
"learning_rate": 0.00019992546003777534,
|
| 3469 |
+
"loss": 6.3413,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.03932146496433495,
|
| 3474 |
+
"grad_norm": 15.585149765014648,
|
| 3475 |
+
"learning_rate": 0.00019992512534372509,
|
| 3476 |
+
"loss": 5.4361,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.03940138664109173,
|
| 3481 |
+
"grad_norm": 15.339882850646973,
|
| 3482 |
+
"learning_rate": 0.00019992479064967483,
|
| 3483 |
+
"loss": 6.2582,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.03948130831784851,
|
| 3488 |
+
"grad_norm": 21.28626251220703,
|
| 3489 |
+
"learning_rate": 0.0001999244705075398,
|
| 3490 |
+
"loss": 6.5715,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.03956122999460528,
|
| 3495 |
+
"grad_norm": 17.29742431640625,
|
| 3496 |
+
"learning_rate": 0.00019992415036540478,
|
| 3497 |
+
"loss": 6.7607,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.03964115167136206,
|
| 3502 |
+
"grad_norm": 21.08455467224121,
|
| 3503 |
+
"learning_rate": 0.00019992381567135453,
|
| 3504 |
+
"loss": 5.7116,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.03972107334811884,
|
| 3509 |
+
"grad_norm": 20.08603286743164,
|
| 3510 |
+
"learning_rate": 0.0001999234955292195,
|
| 3511 |
+
"loss": 6.1074,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.03980099502487562,
|
| 3516 |
+
"grad_norm": 14.703299522399902,
|
| 3517 |
+
"learning_rate": 0.00019992316083516926,
|
| 3518 |
+
"loss": 4.9525,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.0398809167016324,
|
| 3523 |
+
"grad_norm": 19.899187088012695,
|
| 3524 |
+
"learning_rate": 0.00019992284069303423,
|
| 3525 |
+
"loss": 6.1985,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.03996083837838918,
|
| 3530 |
+
"grad_norm": 12.852083206176758,
|
| 3531 |
+
"learning_rate": 0.00019992250599898398,
|
| 3532 |
+
"loss": 5.9677,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.04004076005514596,
|
| 3537 |
+
"grad_norm": 22.426589965820312,
|
| 3538 |
+
"learning_rate": 0.00019992218585684896,
|
| 3539 |
+
"loss": 5.3429,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.04012068173190274,
|
| 3544 |
+
"grad_norm": 20.776029586791992,
|
| 3545 |
+
"learning_rate": 0.0001999218511627987,
|
| 3546 |
+
"loss": 6.2834,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.04020060340865951,
|
| 3551 |
+
"grad_norm": 34.67597198486328,
|
| 3552 |
+
"learning_rate": 0.00019992151646874845,
|
| 3553 |
+
"loss": 5.8435,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.04028052508541629,
|
| 3558 |
+
"grad_norm": 16.465927124023438,
|
| 3559 |
+
"learning_rate": 0.0001999211817746982,
|
| 3560 |
+
"loss": 6.2989,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.04036044676217307,
|
| 3565 |
+
"grad_norm": 16.774991989135742,
|
| 3566 |
+
"learning_rate": 0.00019992084708064795,
|
| 3567 |
+
"loss": 5.5118,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.04044036843892985,
|
| 3572 |
+
"grad_norm": 13.712743759155273,
|
| 3573 |
+
"learning_rate": 0.0001999205123865977,
|
| 3574 |
+
"loss": 4.8975,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.04052029011568663,
|
| 3579 |
+
"grad_norm": 11.83909797668457,
|
| 3580 |
+
"learning_rate": 0.00019992017769254744,
|
| 3581 |
+
"loss": 5.4951,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.04060021179244341,
|
| 3586 |
+
"grad_norm": 16.558422088623047,
|
| 3587 |
+
"learning_rate": 0.00019991985755041242,
|
| 3588 |
+
"loss": 6.2769,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.040680133469200186,
|
| 3593 |
+
"grad_norm": 17.772201538085938,
|
| 3594 |
+
"learning_rate": 0.00019991952285636216,
|
| 3595 |
+
"loss": 6.6068,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.040760055145956965,
|
| 3600 |
+
"grad_norm": 23.38465690612793,
|
| 3601 |
+
"learning_rate": 0.0001999191881623119,
|
| 3602 |
+
"loss": 6.9792,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.04083997682271374,
|
| 3607 |
+
"grad_norm": 11.50236701965332,
|
| 3608 |
+
"learning_rate": 0.00019991883891634643,
|
| 3609 |
+
"loss": 6.3029,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.04091989849947052,
|
| 3614 |
+
"grad_norm": 15.540446281433105,
|
| 3615 |
+
"learning_rate": 0.00019991850422229618,
|
| 3616 |
+
"loss": 6.1962,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.040999820176227296,
|
| 3621 |
+
"grad_norm": 14.016597747802734,
|
| 3622 |
+
"learning_rate": 0.00019991816952824593,
|
| 3623 |
+
"loss": 6.9001,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.041079741852984075,
|
| 3628 |
+
"grad_norm": 16.190292358398438,
|
| 3629 |
+
"learning_rate": 0.00019991782028228045,
|
| 3630 |
+
"loss": 4.8892,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.041159663529740854,
|
| 3635 |
+
"grad_norm": 15.53732681274414,
|
| 3636 |
+
"learning_rate": 0.0001999174855882302,
|
| 3637 |
+
"loss": 6.072,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.041239585206497634,
|
| 3642 |
+
"grad_norm": 15.959219932556152,
|
| 3643 |
+
"learning_rate": 0.00019991715089417994,
|
| 3644 |
+
"loss": 6.1503,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.04131950688325441,
|
| 3649 |
+
"grad_norm": 20.243242263793945,
|
| 3650 |
+
"learning_rate": 0.0001999168162001297,
|
| 3651 |
+
"loss": 6.8969,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.04139942856001119,
|
| 3656 |
+
"grad_norm": 16.59787940979004,
|
| 3657 |
+
"learning_rate": 0.0001999164669541642,
|
| 3658 |
+
"loss": 5.7209,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.041479350236767965,
|
| 3663 |
+
"grad_norm": 16.656587600708008,
|
| 3664 |
+
"learning_rate": 0.00019991611770819873,
|
| 3665 |
+
"loss": 5.7216,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.041559271913524744,
|
| 3670 |
+
"grad_norm": 18.399681091308594,
|
| 3671 |
+
"learning_rate": 0.00019991576846223325,
|
| 3672 |
+
"loss": 5.6021,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.04163919359028152,
|
| 3677 |
+
"grad_norm": 9.782524108886719,
|
| 3678 |
+
"learning_rate": 0.000199915433768183,
|
| 3679 |
+
"loss": 5.4301,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.0417191152670383,
|
| 3684 |
+
"grad_norm": 15.956666946411133,
|
| 3685 |
+
"learning_rate": 0.00019991509907413274,
|
| 3686 |
+
"loss": 5.3535,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.04179903694379508,
|
| 3691 |
+
"grad_norm": 15.992162704467773,
|
| 3692 |
+
"learning_rate": 0.00019991474982816726,
|
| 3693 |
+
"loss": 5.658,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.04187895862055186,
|
| 3698 |
+
"grad_norm": 24.33230209350586,
|
| 3699 |
+
"learning_rate": 0.00019991440058220178,
|
| 3700 |
+
"loss": 6.3838,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.04195888029730864,
|
| 3705 |
+
"grad_norm": 19.471176147460938,
|
| 3706 |
+
"learning_rate": 0.0001999140513362363,
|
| 3707 |
+
"loss": 5.8425,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.04203880197406542,
|
| 3712 |
+
"grad_norm": 18.786813735961914,
|
| 3713 |
+
"learning_rate": 0.00019991370209027082,
|
| 3714 |
+
"loss": 7.659,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.04211872365082219,
|
| 3719 |
+
"grad_norm": 13.322998046875,
|
| 3720 |
+
"learning_rate": 0.00019991335284430534,
|
| 3721 |
+
"loss": 6.5225,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.04219864532757897,
|
| 3726 |
+
"grad_norm": 13.911069869995117,
|
| 3727 |
+
"learning_rate": 0.00019991300359833986,
|
| 3728 |
+
"loss": 5.1711,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.04227856700433575,
|
| 3733 |
+
"grad_norm": 23.9415340423584,
|
| 3734 |
+
"learning_rate": 0.0001999126689042896,
|
| 3735 |
+
"loss": 6.2397,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.04235848868109253,
|
| 3740 |
+
"grad_norm": 17.83696937561035,
|
| 3741 |
+
"learning_rate": 0.0001999123051064089,
|
| 3742 |
+
"loss": 5.841,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.04243841035784931,
|
| 3747 |
+
"grad_norm": 15.819910049438477,
|
| 3748 |
+
"learning_rate": 0.0001999119558604434,
|
| 3749 |
+
"loss": 7.0461,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.04251833203460609,
|
| 3754 |
+
"grad_norm": 20.248186111450195,
|
| 3755 |
+
"learning_rate": 0.00019991160661447793,
|
| 3756 |
+
"loss": 6.4336,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.04259825371136287,
|
| 3761 |
+
"grad_norm": 19.247983932495117,
|
| 3762 |
+
"learning_rate": 0.00019991125736851245,
|
| 3763 |
+
"loss": 5.8052,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.04267817538811964,
|
| 3768 |
+
"grad_norm": 14.141087532043457,
|
| 3769 |
+
"learning_rate": 0.00019991089357063174,
|
| 3770 |
+
"loss": 5.6105,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.04275809706487642,
|
| 3775 |
+
"grad_norm": 21.323780059814453,
|
| 3776 |
+
"learning_rate": 0.00019991054432466626,
|
| 3777 |
+
"loss": 7.947,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.0428380187416332,
|
| 3782 |
+
"grad_norm": 12.924358367919922,
|
| 3783 |
+
"learning_rate": 0.00019991019507870078,
|
| 3784 |
+
"loss": 5.5374,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.04291794041838998,
|
| 3789 |
+
"grad_norm": 14.54228687286377,
|
| 3790 |
+
"learning_rate": 0.00019990983128082007,
|
| 3791 |
+
"loss": 6.1633,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.042997862095146756,
|
| 3796 |
+
"grad_norm": 16.958826065063477,
|
| 3797 |
+
"learning_rate": 0.0001999094820348546,
|
| 3798 |
+
"loss": 6.3655,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.043077783771903536,
|
| 3803 |
+
"grad_norm": 15.821598052978516,
|
| 3804 |
+
"learning_rate": 0.00019990911823697388,
|
| 3805 |
+
"loss": 5.8431,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.043157705448660315,
|
| 3810 |
+
"grad_norm": 14.5997314453125,
|
| 3811 |
+
"learning_rate": 0.0001999087689910084,
|
| 3812 |
+
"loss": 6.9559,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.043237627125417094,
|
| 3817 |
+
"grad_norm": 12.965141296386719,
|
| 3818 |
+
"learning_rate": 0.0001999084051931277,
|
| 3819 |
+
"loss": 5.1944,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.043317548802173866,
|
| 3824 |
+
"grad_norm": 12.994000434875488,
|
| 3825 |
+
"learning_rate": 0.0001999080559471622,
|
| 3826 |
+
"loss": 6.5884,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.043397470478930646,
|
| 3831 |
+
"grad_norm": 20.86507797241211,
|
| 3832 |
+
"learning_rate": 0.0001999076921492815,
|
| 3833 |
+
"loss": 6.0905,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.043477392155687425,
|
| 3838 |
+
"grad_norm": 11.4756441116333,
|
| 3839 |
+
"learning_rate": 0.0001999073283514008,
|
| 3840 |
+
"loss": 4.8621,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.043557313832444204,
|
| 3845 |
+
"grad_norm": 14.282588958740234,
|
| 3846 |
+
"learning_rate": 0.00019990696455352008,
|
| 3847 |
+
"loss": 5.9349,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.04363723550920098,
|
| 3852 |
+
"grad_norm": 20.818880081176758,
|
| 3853 |
+
"learning_rate": 0.00019990660075563937,
|
| 3854 |
+
"loss": 8.0508,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.04371715718595776,
|
| 3859 |
+
"grad_norm": 13.455410957336426,
|
| 3860 |
+
"learning_rate": 0.00019990623695775867,
|
| 3861 |
+
"loss": 5.921,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.04379707886271454,
|
| 3866 |
+
"grad_norm": 22.602798461914062,
|
| 3867 |
+
"learning_rate": 0.00019990587315987796,
|
| 3868 |
+
"loss": 7.1508,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.04387700053947132,
|
| 3873 |
+
"grad_norm": 13.060234069824219,
|
| 3874 |
+
"learning_rate": 0.00019990552391391248,
|
| 3875 |
+
"loss": 5.7122,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.04395692221622809,
|
| 3880 |
+
"grad_norm": 21.732711791992188,
|
| 3881 |
+
"learning_rate": 0.00019990514556411654,
|
| 3882 |
+
"loss": 5.3887,
|
| 3883 |
+
"step": 550
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 0.04403684389298487,
|
| 3887 |
+
"grad_norm": 11.83836841583252,
|
| 3888 |
+
"learning_rate": 0.00019990478176623583,
|
| 3889 |
+
"loss": 5.8057,
|
| 3890 |
+
"step": 551
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 0.04411676556974165,
|
| 3894 |
+
"grad_norm": 18.189430236816406,
|
| 3895 |
+
"learning_rate": 0.00019990441796835512,
|
| 3896 |
+
"loss": 5.7397,
|
| 3897 |
+
"step": 552
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 0.04419668724649843,
|
| 3901 |
+
"grad_norm": 18.77703285217285,
|
| 3902 |
+
"learning_rate": 0.0001999040541704744,
|
| 3903 |
+
"loss": 5.9096,
|
| 3904 |
+
"step": 553
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 0.04427660892325521,
|
| 3908 |
+
"grad_norm": 16.678699493408203,
|
| 3909 |
+
"learning_rate": 0.0001999036903725937,
|
| 3910 |
+
"loss": 6.3299,
|
| 3911 |
+
"step": 554
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 0.04435653060001199,
|
| 3915 |
+
"grad_norm": 16.687585830688477,
|
| 3916 |
+
"learning_rate": 0.00019990331202279776,
|
| 3917 |
+
"loss": 7.4824,
|
| 3918 |
+
"step": 555
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.04443645227676877,
|
| 3922 |
+
"grad_norm": 17.86968421936035,
|
| 3923 |
+
"learning_rate": 0.00019990294822491705,
|
| 3924 |
+
"loss": 6.3815,
|
| 3925 |
+
"step": 556
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 0.04451637395352555,
|
| 3929 |
+
"grad_norm": 16.1313533782959,
|
| 3930 |
+
"learning_rate": 0.00019990256987512112,
|
| 3931 |
+
"loss": 6.2465,
|
| 3932 |
+
"step": 557
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 0.04459629563028232,
|
| 3936 |
+
"grad_norm": 16.072376251220703,
|
| 3937 |
+
"learning_rate": 0.0001999022060772404,
|
| 3938 |
+
"loss": 6.1389,
|
| 3939 |
+
"step": 558
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 0.0446762173070391,
|
| 3943 |
+
"grad_norm": 24.730459213256836,
|
| 3944 |
+
"learning_rate": 0.0001999018422793597,
|
| 3945 |
+
"loss": 6.9015,
|
| 3946 |
+
"step": 559
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 0.04475613898379588,
|
| 3950 |
+
"grad_norm": 15.83382797241211,
|
| 3951 |
+
"learning_rate": 0.00019990146392956376,
|
| 3952 |
+
"loss": 6.3929,
|
| 3953 |
+
"step": 560
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 0.04483606066055266,
|
| 3957 |
+
"grad_norm": 20.246620178222656,
|
| 3958 |
+
"learning_rate": 0.00019990110013168305,
|
| 3959 |
+
"loss": 6.5326,
|
| 3960 |
+
"step": 561
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 0.04491598233730944,
|
| 3964 |
+
"grad_norm": 17.737701416015625,
|
| 3965 |
+
"learning_rate": 0.00019990073633380234,
|
| 3966 |
+
"loss": 6.6179,
|
| 3967 |
+
"step": 562
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 0.04499590401406622,
|
| 3971 |
+
"grad_norm": 20.27594566345215,
|
| 3972 |
+
"learning_rate": 0.00019990034343209118,
|
| 3973 |
+
"loss": 6.1448,
|
| 3974 |
+
"step": 563
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 0.045075825690822996,
|
| 3978 |
+
"grad_norm": 19.090566635131836,
|
| 3979 |
+
"learning_rate": 0.00019989997963421047,
|
| 3980 |
+
"loss": 6.0835,
|
| 3981 |
+
"step": 564
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 0.045155747367579775,
|
| 3985 |
+
"grad_norm": 19.26055145263672,
|
| 3986 |
+
"learning_rate": 0.00019989960128441453,
|
| 3987 |
+
"loss": 5.944,
|
| 3988 |
+
"step": 565
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 0.04523566904433655,
|
| 3992 |
+
"grad_norm": 21.74213981628418,
|
| 3993 |
+
"learning_rate": 0.0001998992229346186,
|
| 3994 |
+
"loss": 5.4846,
|
| 3995 |
+
"step": 566
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 0.04531559072109333,
|
| 3999 |
+
"grad_norm": 16.654098510742188,
|
| 4000 |
+
"learning_rate": 0.00019989884458482265,
|
| 4001 |
+
"loss": 5.9757,
|
| 4002 |
+
"step": 567
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 0.045395512397850106,
|
| 4006 |
+
"grad_norm": 16.927513122558594,
|
| 4007 |
+
"learning_rate": 0.00019989848078694195,
|
| 4008 |
+
"loss": 5.3209,
|
| 4009 |
+
"step": 568
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 0.045475434074606885,
|
| 4013 |
+
"grad_norm": 21.91180419921875,
|
| 4014 |
+
"learning_rate": 0.00019989808788523078,
|
| 4015 |
+
"loss": 6.4642,
|
| 4016 |
+
"step": 569
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 0.045555355751363664,
|
| 4020 |
+
"grad_norm": 13.900008201599121,
|
| 4021 |
+
"learning_rate": 0.00019989770953543484,
|
| 4022 |
+
"loss": 5.8995,
|
| 4023 |
+
"step": 570
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 0.045635277428120444,
|
| 4027 |
+
"grad_norm": 15.217854499816895,
|
| 4028 |
+
"learning_rate": 0.00019989734573755413,
|
| 4029 |
+
"loss": 4.5451,
|
| 4030 |
+
"step": 571
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 0.04571519910487722,
|
| 4034 |
+
"grad_norm": 19.86149024963379,
|
| 4035 |
+
"learning_rate": 0.0001998969673877582,
|
| 4036 |
+
"loss": 5.5945,
|
| 4037 |
+
"step": 572
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.045795120781633995,
|
| 4041 |
+
"grad_norm": 14.814356803894043,
|
| 4042 |
+
"learning_rate": 0.00019989657448604703,
|
| 4043 |
+
"loss": 6.508,
|
| 4044 |
+
"step": 573
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 0.045875042458390775,
|
| 4048 |
+
"grad_norm": 15.304423332214355,
|
| 4049 |
+
"learning_rate": 0.0001998961961362511,
|
| 4050 |
+
"loss": 6.6856,
|
| 4051 |
+
"step": 574
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 0.045954964135147554,
|
| 4055 |
+
"grad_norm": 21.85096549987793,
|
| 4056 |
+
"learning_rate": 0.00019989581778645515,
|
| 4057 |
+
"loss": 5.3746,
|
| 4058 |
+
"step": 575
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 0.04603488581190433,
|
| 4062 |
+
"grad_norm": 25.818893432617188,
|
| 4063 |
+
"learning_rate": 0.00019989543943665922,
|
| 4064 |
+
"loss": 6.4333,
|
| 4065 |
+
"step": 576
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 0.04611480748866111,
|
| 4069 |
+
"grad_norm": 27.063743591308594,
|
| 4070 |
+
"learning_rate": 0.00019989504653494805,
|
| 4071 |
+
"loss": 6.2794,
|
| 4072 |
+
"step": 577
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 0.04619472916541789,
|
| 4076 |
+
"grad_norm": 21.274269104003906,
|
| 4077 |
+
"learning_rate": 0.0001998946681851521,
|
| 4078 |
+
"loss": 6.8713,
|
| 4079 |
+
"step": 578
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.04627465084217467,
|
| 4083 |
+
"grad_norm": 14.125374794006348,
|
| 4084 |
+
"learning_rate": 0.00019989427528344095,
|
| 4085 |
+
"loss": 5.3867,
|
| 4086 |
+
"step": 579
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 0.04635457251893145,
|
| 4090 |
+
"grad_norm": 18.209699630737305,
|
| 4091 |
+
"learning_rate": 0.000199893896933645,
|
| 4092 |
+
"loss": 5.8785,
|
| 4093 |
+
"step": 580
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 0.04643449419568822,
|
| 4097 |
+
"grad_norm": 25.810714721679688,
|
| 4098 |
+
"learning_rate": 0.00019989350403193384,
|
| 4099 |
+
"loss": 6.0333,
|
| 4100 |
+
"step": 581
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 0.046514415872445,
|
| 4104 |
+
"grad_norm": 15.457906723022461,
|
| 4105 |
+
"learning_rate": 0.0001998931256821379,
|
| 4106 |
+
"loss": 7.0555,
|
| 4107 |
+
"step": 582
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 0.04659433754920178,
|
| 4111 |
+
"grad_norm": 20.941139221191406,
|
| 4112 |
+
"learning_rate": 0.00019989273278042674,
|
| 4113 |
+
"loss": 6.027,
|
| 4114 |
+
"step": 583
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 0.04667425922595856,
|
| 4118 |
+
"grad_norm": 12.414090156555176,
|
| 4119 |
+
"learning_rate": 0.0001998923544306308,
|
| 4120 |
+
"loss": 6.0731,
|
| 4121 |
+
"step": 584
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.04675418090271534,
|
| 4125 |
+
"grad_norm": 14.253443717956543,
|
| 4126 |
+
"learning_rate": 0.00019989196152891964,
|
| 4127 |
+
"loss": 6.3446,
|
| 4128 |
+
"step": 585
|
| 4129 |
+
},
|
| 4130 |
+
{
|
| 4131 |
+
"epoch": 0.04683410257947212,
|
| 4132 |
+
"grad_norm": 23.935544967651367,
|
| 4133 |
+
"learning_rate": 0.00019989156862720847,
|
| 4134 |
+
"loss": 7.8148,
|
| 4135 |
+
"step": 586
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 0.0469140242562289,
|
| 4139 |
+
"grad_norm": 14.470626831054688,
|
| 4140 |
+
"learning_rate": 0.0001998911757254973,
|
| 4141 |
+
"loss": 6.1853,
|
| 4142 |
+
"step": 587
|
| 4143 |
+
},
|
| 4144 |
+
{
|
| 4145 |
+
"epoch": 0.04699394593298568,
|
| 4146 |
+
"grad_norm": 15.151899337768555,
|
| 4147 |
+
"learning_rate": 0.00019989078282378614,
|
| 4148 |
+
"loss": 5.6132,
|
| 4149 |
+
"step": 588
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 0.04707386760974245,
|
| 4153 |
+
"grad_norm": 14.704421043395996,
|
| 4154 |
+
"learning_rate": 0.00019989038992207497,
|
| 4155 |
+
"loss": 5.9291,
|
| 4156 |
+
"step": 589
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 0.04715378928649923,
|
| 4160 |
+
"grad_norm": 11.972437858581543,
|
| 4161 |
+
"learning_rate": 0.0001998899970203638,
|
| 4162 |
+
"loss": 5.7066,
|
| 4163 |
+
"step": 590
|
| 4164 |
+
},
|
| 4165 |
+
{
|
| 4166 |
+
"epoch": 0.04723371096325601,
|
| 4167 |
+
"grad_norm": 17.223663330078125,
|
| 4168 |
+
"learning_rate": 0.00019988960411865264,
|
| 4169 |
+
"loss": 6.0548,
|
| 4170 |
+
"step": 591
|
| 4171 |
+
},
|
| 4172 |
+
{
|
| 4173 |
+
"epoch": 0.04731363264001279,
|
| 4174 |
+
"grad_norm": 15.116782188415527,
|
| 4175 |
+
"learning_rate": 0.00019988921121694148,
|
| 4176 |
+
"loss": 5.9861,
|
| 4177 |
+
"step": 592
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 0.047393554316769566,
|
| 4181 |
+
"grad_norm": 17.56676483154297,
|
| 4182 |
+
"learning_rate": 0.0001998888183152303,
|
| 4183 |
+
"loss": 6.1462,
|
| 4184 |
+
"step": 593
|
| 4185 |
+
},
|
| 4186 |
+
{
|
| 4187 |
+
"epoch": 0.047473475993526346,
|
| 4188 |
+
"grad_norm": 18.915695190429688,
|
| 4189 |
+
"learning_rate": 0.00019988842541351914,
|
| 4190 |
+
"loss": 5.8419,
|
| 4191 |
+
"step": 594
|
| 4192 |
+
},
|
| 4193 |
+
{
|
| 4194 |
+
"epoch": 0.047553397670283125,
|
| 4195 |
+
"grad_norm": 29.974353790283203,
|
| 4196 |
+
"learning_rate": 0.00019988803251180798,
|
| 4197 |
+
"loss": 7.8609,
|
| 4198 |
+
"step": 595
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 0.047633319347039904,
|
| 4202 |
+
"grad_norm": 23.129892349243164,
|
| 4203 |
+
"learning_rate": 0.00019988762505818158,
|
| 4204 |
+
"loss": 5.7261,
|
| 4205 |
+
"step": 596
|
| 4206 |
+
},
|
| 4207 |
+
{
|
| 4208 |
+
"epoch": 0.047713241023796676,
|
| 4209 |
+
"grad_norm": 15.831592559814453,
|
| 4210 |
+
"learning_rate": 0.00019988723215647042,
|
| 4211 |
+
"loss": 5.8459,
|
| 4212 |
+
"step": 597
|
| 4213 |
+
},
|
| 4214 |
+
{
|
| 4215 |
+
"epoch": 0.047793162700553456,
|
| 4216 |
+
"grad_norm": 25.895620346069336,
|
| 4217 |
+
"learning_rate": 0.00019988683925475925,
|
| 4218 |
+
"loss": 6.3782,
|
| 4219 |
+
"step": 598
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 0.047873084377310235,
|
| 4223 |
+
"grad_norm": 25.200590133666992,
|
| 4224 |
+
"learning_rate": 0.00019988643180113286,
|
| 4225 |
+
"loss": 7.3017,
|
| 4226 |
+
"step": 599
|
| 4227 |
+
},
|
| 4228 |
+
{
|
| 4229 |
+
"epoch": 0.047953006054067014,
|
| 4230 |
+
"grad_norm": 25.98448944091797,
|
| 4231 |
+
"learning_rate": 0.0001998860388994217,
|
| 4232 |
+
"loss": 5.8241,
|
| 4233 |
+
"step": 600
|
| 4234 |
+
},
|
| 4235 |
+
{
|
| 4236 |
+
"epoch": 0.047953006054067014,
|
| 4237 |
+
"eval_loss": 1.5112873315811157,
|
| 4238 |
+
"eval_runtime": 98.2922,
|
| 4239 |
+
"eval_samples_per_second": 5.117,
|
| 4240 |
+
"eval_steps_per_second": 2.564,
|
| 4241 |
+
"step": 600
|
| 4242 |
}
|
| 4243 |
],
|
| 4244 |
"logging_steps": 1,
|
|
|
|
| 4253 |
"early_stopping_threshold": 0.0
|
| 4254 |
},
|
| 4255 |
"attributes": {
|
| 4256 |
+
"early_stopping_patience_counter": 2
|
| 4257 |
}
|
| 4258 |
},
|
| 4259 |
"TrainerControl": {
|
|
|
|
| 4267 |
"attributes": {}
|
| 4268 |
}
|
| 4269 |
},
|
| 4270 |
+
"total_flos": 1.1568008025184666e+17,
|
| 4271 |
"train_batch_size": 2,
|
| 4272 |
"trial_name": null,
|
| 4273 |
"trial_params": null
|