Fanucci
commited on
Training in progress, step 1200, 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:f17745bedcf2691c98c5c38bed681bcbf85273eca7d4b073a3a5217298d9a2d3
|
| 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:cff726efcf2b6d6ea5f5b99a5d48fd9ea0c135385b1d0d2cca75befa63a0bc1a
|
| 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 sha256:2fd83e1e15f70d373e2aa254f4001309c2d36b697f5a068bcfbf1e9855d8a484
|
| 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 sha256:8e7f5cdb6cd5b1751a2a041a35a236dff4650191e9a4687c5450d12c2fac14d6
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 3.1351470947265625,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 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,
|
|
@@ -7055,6 +7055,1414 @@
|
|
| 7055 |
"eval_samples_per_second": 5.624,
|
| 7056 |
"eval_steps_per_second": 1.406,
|
| 7057 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7058 |
}
|
| 7059 |
],
|
| 7060 |
"logging_steps": 1,
|
|
@@ -7069,7 +8477,7 @@
|
|
| 7069 |
"early_stopping_threshold": 0.0
|
| 7070 |
},
|
| 7071 |
"attributes": {
|
| 7072 |
-
"early_stopping_patience_counter":
|
| 7073 |
}
|
| 7074 |
},
|
| 7075 |
"TrainerControl": {
|
|
@@ -7083,7 +8491,7 @@
|
|
| 7083 |
"attributes": {}
|
| 7084 |
}
|
| 7085 |
},
|
| 7086 |
-
"total_flos":
|
| 7087 |
"train_batch_size": 4,
|
| 7088 |
"trial_name": null,
|
| 7089 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 3.1351470947265625,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 0.3839385698288274,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 1200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 7055 |
"eval_samples_per_second": 5.624,
|
| 7056 |
"eval_steps_per_second": 1.406,
|
| 7057 |
"step": 1000
|
| 7058 |
+
},
|
| 7059 |
+
{
|
| 7060 |
+
"epoch": 0.3202687569988802,
|
| 7061 |
+
"grad_norm": 9.613029479980469,
|
| 7062 |
+
"learning_rate": 0.0003093164406544793,
|
| 7063 |
+
"loss": 6.5549,
|
| 7064 |
+
"step": 1001
|
| 7065 |
+
},
|
| 7066 |
+
{
|
| 7067 |
+
"epoch": 0.3205887058070709,
|
| 7068 |
+
"grad_norm": 6.538544654846191,
|
| 7069 |
+
"learning_rate": 0.0003091495477342108,
|
| 7070 |
+
"loss": 6.6206,
|
| 7071 |
+
"step": 1002
|
| 7072 |
+
},
|
| 7073 |
+
{
|
| 7074 |
+
"epoch": 0.3209086546152616,
|
| 7075 |
+
"grad_norm": 6.343434810638428,
|
| 7076 |
+
"learning_rate": 0.00030898254961999173,
|
| 7077 |
+
"loss": 6.7164,
|
| 7078 |
+
"step": 1003
|
| 7079 |
+
},
|
| 7080 |
+
{
|
| 7081 |
+
"epoch": 0.3212286034234523,
|
| 7082 |
+
"grad_norm": 6.2874908447265625,
|
| 7083 |
+
"learning_rate": 0.0003088154464890007,
|
| 7084 |
+
"loss": 6.6797,
|
| 7085 |
+
"step": 1004
|
| 7086 |
+
},
|
| 7087 |
+
{
|
| 7088 |
+
"epoch": 0.3215485522316429,
|
| 7089 |
+
"grad_norm": 18.32168960571289,
|
| 7090 |
+
"learning_rate": 0.00030864823851852786,
|
| 7091 |
+
"loss": 6.3663,
|
| 7092 |
+
"step": 1005
|
| 7093 |
+
},
|
| 7094 |
+
{
|
| 7095 |
+
"epoch": 0.3218685010398336,
|
| 7096 |
+
"grad_norm": 4.97805643081665,
|
| 7097 |
+
"learning_rate": 0.00030848092588597463,
|
| 7098 |
+
"loss": 6.4058,
|
| 7099 |
+
"step": 1006
|
| 7100 |
+
},
|
| 7101 |
+
{
|
| 7102 |
+
"epoch": 0.3221884498480243,
|
| 7103 |
+
"grad_norm": 7.486648082733154,
|
| 7104 |
+
"learning_rate": 0.00030831350876885344,
|
| 7105 |
+
"loss": 6.6289,
|
| 7106 |
+
"step": 1007
|
| 7107 |
+
},
|
| 7108 |
+
{
|
| 7109 |
+
"epoch": 0.322508398656215,
|
| 7110 |
+
"grad_norm": 7.603539943695068,
|
| 7111 |
+
"learning_rate": 0.0003081459873447875,
|
| 7112 |
+
"loss": 6.5563,
|
| 7113 |
+
"step": 1008
|
| 7114 |
+
},
|
| 7115 |
+
{
|
| 7116 |
+
"epoch": 0.3228283474644057,
|
| 7117 |
+
"grad_norm": 9.133359909057617,
|
| 7118 |
+
"learning_rate": 0.00030797836179151085,
|
| 7119 |
+
"loss": 6.5898,
|
| 7120 |
+
"step": 1009
|
| 7121 |
+
},
|
| 7122 |
+
{
|
| 7123 |
+
"epoch": 0.3231482962725964,
|
| 7124 |
+
"grad_norm": 8.029521942138672,
|
| 7125 |
+
"learning_rate": 0.0003078106322868679,
|
| 7126 |
+
"loss": 6.8213,
|
| 7127 |
+
"step": 1010
|
| 7128 |
+
},
|
| 7129 |
+
{
|
| 7130 |
+
"epoch": 0.3234682450807871,
|
| 7131 |
+
"grad_norm": 5.991820335388184,
|
| 7132 |
+
"learning_rate": 0.0003076427990088133,
|
| 7133 |
+
"loss": 6.5975,
|
| 7134 |
+
"step": 1011
|
| 7135 |
+
},
|
| 7136 |
+
{
|
| 7137 |
+
"epoch": 0.3237881938889778,
|
| 7138 |
+
"grad_norm": 4.940555572509766,
|
| 7139 |
+
"learning_rate": 0.0003074748621354119,
|
| 7140 |
+
"loss": 6.628,
|
| 7141 |
+
"step": 1012
|
| 7142 |
+
},
|
| 7143 |
+
{
|
| 7144 |
+
"epoch": 0.3241081426971685,
|
| 7145 |
+
"grad_norm": 7.488973140716553,
|
| 7146 |
+
"learning_rate": 0.0003073068218448384,
|
| 7147 |
+
"loss": 6.56,
|
| 7148 |
+
"step": 1013
|
| 7149 |
+
},
|
| 7150 |
+
{
|
| 7151 |
+
"epoch": 0.3244280915053591,
|
| 7152 |
+
"grad_norm": 10.482954025268555,
|
| 7153 |
+
"learning_rate": 0.0003071386783153772,
|
| 7154 |
+
"loss": 6.7303,
|
| 7155 |
+
"step": 1014
|
| 7156 |
+
},
|
| 7157 |
+
{
|
| 7158 |
+
"epoch": 0.3247480403135498,
|
| 7159 |
+
"grad_norm": 7.40532112121582,
|
| 7160 |
+
"learning_rate": 0.00030697043172542244,
|
| 7161 |
+
"loss": 6.8107,
|
| 7162 |
+
"step": 1015
|
| 7163 |
+
},
|
| 7164 |
+
{
|
| 7165 |
+
"epoch": 0.3250679891217405,
|
| 7166 |
+
"grad_norm": 16.26544189453125,
|
| 7167 |
+
"learning_rate": 0.00030680208225347737,
|
| 7168 |
+
"loss": 6.5008,
|
| 7169 |
+
"step": 1016
|
| 7170 |
+
},
|
| 7171 |
+
{
|
| 7172 |
+
"epoch": 0.3253879379299312,
|
| 7173 |
+
"grad_norm": 13.164572715759277,
|
| 7174 |
+
"learning_rate": 0.0003066336300781544,
|
| 7175 |
+
"loss": 6.5557,
|
| 7176 |
+
"step": 1017
|
| 7177 |
+
},
|
| 7178 |
+
{
|
| 7179 |
+
"epoch": 0.3257078867381219,
|
| 7180 |
+
"grad_norm": 12.598750114440918,
|
| 7181 |
+
"learning_rate": 0.000306465075378175,
|
| 7182 |
+
"loss": 6.7085,
|
| 7183 |
+
"step": 1018
|
| 7184 |
+
},
|
| 7185 |
+
{
|
| 7186 |
+
"epoch": 0.3260278355463126,
|
| 7187 |
+
"grad_norm": 6.103877544403076,
|
| 7188 |
+
"learning_rate": 0.00030629641833236935,
|
| 7189 |
+
"loss": 6.6045,
|
| 7190 |
+
"step": 1019
|
| 7191 |
+
},
|
| 7192 |
+
{
|
| 7193 |
+
"epoch": 0.3263477843545033,
|
| 7194 |
+
"grad_norm": 6.347523212432861,
|
| 7195 |
+
"learning_rate": 0.0003061276591196764,
|
| 7196 |
+
"loss": 6.8141,
|
| 7197 |
+
"step": 1020
|
| 7198 |
+
},
|
| 7199 |
+
{
|
| 7200 |
+
"epoch": 0.326667733162694,
|
| 7201 |
+
"grad_norm": 3.747049570083618,
|
| 7202 |
+
"learning_rate": 0.00030595879791914314,
|
| 7203 |
+
"loss": 6.5425,
|
| 7204 |
+
"step": 1021
|
| 7205 |
+
},
|
| 7206 |
+
{
|
| 7207 |
+
"epoch": 0.3269876819708847,
|
| 7208 |
+
"grad_norm": 6.858709335327148,
|
| 7209 |
+
"learning_rate": 0.00030578983490992505,
|
| 7210 |
+
"loss": 6.4573,
|
| 7211 |
+
"step": 1022
|
| 7212 |
+
},
|
| 7213 |
+
{
|
| 7214 |
+
"epoch": 0.3273076307790753,
|
| 7215 |
+
"grad_norm": 3.5340542793273926,
|
| 7216 |
+
"learning_rate": 0.00030562077027128555,
|
| 7217 |
+
"loss": 6.506,
|
| 7218 |
+
"step": 1023
|
| 7219 |
+
},
|
| 7220 |
+
{
|
| 7221 |
+
"epoch": 0.327627579587266,
|
| 7222 |
+
"grad_norm": 6.104124546051025,
|
| 7223 |
+
"learning_rate": 0.000305451604182596,
|
| 7224 |
+
"loss": 6.5108,
|
| 7225 |
+
"step": 1024
|
| 7226 |
+
},
|
| 7227 |
+
{
|
| 7228 |
+
"epoch": 0.3279475283954567,
|
| 7229 |
+
"grad_norm": 7.606841087341309,
|
| 7230 |
+
"learning_rate": 0.000305282336823335,
|
| 7231 |
+
"loss": 6.7387,
|
| 7232 |
+
"step": 1025
|
| 7233 |
+
},
|
| 7234 |
+
{
|
| 7235 |
+
"epoch": 0.3282674772036474,
|
| 7236 |
+
"grad_norm": 6.288745880126953,
|
| 7237 |
+
"learning_rate": 0.0003051129683730891,
|
| 7238 |
+
"loss": 6.7012,
|
| 7239 |
+
"step": 1026
|
| 7240 |
+
},
|
| 7241 |
+
{
|
| 7242 |
+
"epoch": 0.3285874260118381,
|
| 7243 |
+
"grad_norm": 7.002671241760254,
|
| 7244 |
+
"learning_rate": 0.0003049434990115518,
|
| 7245 |
+
"loss": 6.5738,
|
| 7246 |
+
"step": 1027
|
| 7247 |
+
},
|
| 7248 |
+
{
|
| 7249 |
+
"epoch": 0.3289073748200288,
|
| 7250 |
+
"grad_norm": 5.6297478675842285,
|
| 7251 |
+
"learning_rate": 0.00030477392891852374,
|
| 7252 |
+
"loss": 6.6822,
|
| 7253 |
+
"step": 1028
|
| 7254 |
+
},
|
| 7255 |
+
{
|
| 7256 |
+
"epoch": 0.3292273236282195,
|
| 7257 |
+
"grad_norm": 9.002836227416992,
|
| 7258 |
+
"learning_rate": 0.0003046042582739124,
|
| 7259 |
+
"loss": 6.7199,
|
| 7260 |
+
"step": 1029
|
| 7261 |
+
},
|
| 7262 |
+
{
|
| 7263 |
+
"epoch": 0.3295472724364102,
|
| 7264 |
+
"grad_norm": 4.909145832061768,
|
| 7265 |
+
"learning_rate": 0.000304434487257732,
|
| 7266 |
+
"loss": 6.6323,
|
| 7267 |
+
"step": 1030
|
| 7268 |
+
},
|
| 7269 |
+
{
|
| 7270 |
+
"epoch": 0.3298672212446009,
|
| 7271 |
+
"grad_norm": 15.304366111755371,
|
| 7272 |
+
"learning_rate": 0.00030426461605010324,
|
| 7273 |
+
"loss": 6.6586,
|
| 7274 |
+
"step": 1031
|
| 7275 |
+
},
|
| 7276 |
+
{
|
| 7277 |
+
"epoch": 0.3301871700527916,
|
| 7278 |
+
"grad_norm": 16.558542251586914,
|
| 7279 |
+
"learning_rate": 0.0003040946448312531,
|
| 7280 |
+
"loss": 6.5575,
|
| 7281 |
+
"step": 1032
|
| 7282 |
+
},
|
| 7283 |
+
{
|
| 7284 |
+
"epoch": 0.3305071188609822,
|
| 7285 |
+
"grad_norm": 17.546894073486328,
|
| 7286 |
+
"learning_rate": 0.00030392457378151456,
|
| 7287 |
+
"loss": 6.4663,
|
| 7288 |
+
"step": 1033
|
| 7289 |
+
},
|
| 7290 |
+
{
|
| 7291 |
+
"epoch": 0.3308270676691729,
|
| 7292 |
+
"grad_norm": 6.221606254577637,
|
| 7293 |
+
"learning_rate": 0.00030375440308132666,
|
| 7294 |
+
"loss": 6.6658,
|
| 7295 |
+
"step": 1034
|
| 7296 |
+
},
|
| 7297 |
+
{
|
| 7298 |
+
"epoch": 0.3311470164773636,
|
| 7299 |
+
"grad_norm": 5.4567999839782715,
|
| 7300 |
+
"learning_rate": 0.0003035841329112343,
|
| 7301 |
+
"loss": 6.5654,
|
| 7302 |
+
"step": 1035
|
| 7303 |
+
},
|
| 7304 |
+
{
|
| 7305 |
+
"epoch": 0.3314669652855543,
|
| 7306 |
+
"grad_norm": 17.624677658081055,
|
| 7307 |
+
"learning_rate": 0.0003034137634518875,
|
| 7308 |
+
"loss": 6.582,
|
| 7309 |
+
"step": 1036
|
| 7310 |
+
},
|
| 7311 |
+
{
|
| 7312 |
+
"epoch": 0.331786914093745,
|
| 7313 |
+
"grad_norm": 17.85716438293457,
|
| 7314 |
+
"learning_rate": 0.00030324329488404207,
|
| 7315 |
+
"loss": 6.7594,
|
| 7316 |
+
"step": 1037
|
| 7317 |
+
},
|
| 7318 |
+
{
|
| 7319 |
+
"epoch": 0.3321068629019357,
|
| 7320 |
+
"grad_norm": 15.087934494018555,
|
| 7321 |
+
"learning_rate": 0.00030307272738855855,
|
| 7322 |
+
"loss": 6.6988,
|
| 7323 |
+
"step": 1038
|
| 7324 |
+
},
|
| 7325 |
+
{
|
| 7326 |
+
"epoch": 0.3324268117101264,
|
| 7327 |
+
"grad_norm": 8.74818229675293,
|
| 7328 |
+
"learning_rate": 0.0003029020611464029,
|
| 7329 |
+
"loss": 6.6417,
|
| 7330 |
+
"step": 1039
|
| 7331 |
+
},
|
| 7332 |
+
{
|
| 7333 |
+
"epoch": 0.3327467605183171,
|
| 7334 |
+
"grad_norm": 6.144550323486328,
|
| 7335 |
+
"learning_rate": 0.00030273129633864534,
|
| 7336 |
+
"loss": 6.5553,
|
| 7337 |
+
"step": 1040
|
| 7338 |
+
},
|
| 7339 |
+
{
|
| 7340 |
+
"epoch": 0.3330667093265078,
|
| 7341 |
+
"grad_norm": 14.558652877807617,
|
| 7342 |
+
"learning_rate": 0.0003025604331464612,
|
| 7343 |
+
"loss": 6.6014,
|
| 7344 |
+
"step": 1041
|
| 7345 |
+
},
|
| 7346 |
+
{
|
| 7347 |
+
"epoch": 0.3333866581346984,
|
| 7348 |
+
"grad_norm": 7.271969318389893,
|
| 7349 |
+
"learning_rate": 0.0003023894717511297,
|
| 7350 |
+
"loss": 6.5538,
|
| 7351 |
+
"step": 1042
|
| 7352 |
+
},
|
| 7353 |
+
{
|
| 7354 |
+
"epoch": 0.3337066069428891,
|
| 7355 |
+
"grad_norm": 9.705676078796387,
|
| 7356 |
+
"learning_rate": 0.0003022184123340346,
|
| 7357 |
+
"loss": 6.6535,
|
| 7358 |
+
"step": 1043
|
| 7359 |
+
},
|
| 7360 |
+
{
|
| 7361 |
+
"epoch": 0.3340265557510798,
|
| 7362 |
+
"grad_norm": 7.595553398132324,
|
| 7363 |
+
"learning_rate": 0.0003020472550766634,
|
| 7364 |
+
"loss": 6.7706,
|
| 7365 |
+
"step": 1044
|
| 7366 |
+
},
|
| 7367 |
+
{
|
| 7368 |
+
"epoch": 0.3343465045592705,
|
| 7369 |
+
"grad_norm": 7.8142476081848145,
|
| 7370 |
+
"learning_rate": 0.00030187600016060766,
|
| 7371 |
+
"loss": 6.7936,
|
| 7372 |
+
"step": 1045
|
| 7373 |
+
},
|
| 7374 |
+
{
|
| 7375 |
+
"epoch": 0.3346664533674612,
|
| 7376 |
+
"grad_norm": 7.204625129699707,
|
| 7377 |
+
"learning_rate": 0.0003017046477675624,
|
| 7378 |
+
"loss": 6.4859,
|
| 7379 |
+
"step": 1046
|
| 7380 |
+
},
|
| 7381 |
+
{
|
| 7382 |
+
"epoch": 0.3349864021756519,
|
| 7383 |
+
"grad_norm": 8.921597480773926,
|
| 7384 |
+
"learning_rate": 0.0003015331980793259,
|
| 7385 |
+
"loss": 6.5621,
|
| 7386 |
+
"step": 1047
|
| 7387 |
+
},
|
| 7388 |
+
{
|
| 7389 |
+
"epoch": 0.3353063509838426,
|
| 7390 |
+
"grad_norm": 7.932427406311035,
|
| 7391 |
+
"learning_rate": 0.0003013616512778001,
|
| 7392 |
+
"loss": 6.566,
|
| 7393 |
+
"step": 1048
|
| 7394 |
+
},
|
| 7395 |
+
{
|
| 7396 |
+
"epoch": 0.3356262997920333,
|
| 7397 |
+
"grad_norm": 11.233190536499023,
|
| 7398 |
+
"learning_rate": 0.00030119000754498965,
|
| 7399 |
+
"loss": 6.7973,
|
| 7400 |
+
"step": 1049
|
| 7401 |
+
},
|
| 7402 |
+
{
|
| 7403 |
+
"epoch": 0.335946248600224,
|
| 7404 |
+
"grad_norm": 7.584317684173584,
|
| 7405 |
+
"learning_rate": 0.00030101826706300193,
|
| 7406 |
+
"loss": 6.5683,
|
| 7407 |
+
"step": 1050
|
| 7408 |
+
},
|
| 7409 |
+
{
|
| 7410 |
+
"epoch": 0.3362661974084147,
|
| 7411 |
+
"grad_norm": 8.914484024047852,
|
| 7412 |
+
"learning_rate": 0.0003008464300140474,
|
| 7413 |
+
"loss": 6.6189,
|
| 7414 |
+
"step": 1051
|
| 7415 |
+
},
|
| 7416 |
+
{
|
| 7417 |
+
"epoch": 0.3365861462166053,
|
| 7418 |
+
"grad_norm": 11.658357620239258,
|
| 7419 |
+
"learning_rate": 0.0003006744965804385,
|
| 7420 |
+
"loss": 6.6833,
|
| 7421 |
+
"step": 1052
|
| 7422 |
+
},
|
| 7423 |
+
{
|
| 7424 |
+
"epoch": 0.336906095024796,
|
| 7425 |
+
"grad_norm": 8.101858139038086,
|
| 7426 |
+
"learning_rate": 0.0003005024669445904,
|
| 7427 |
+
"loss": 6.7025,
|
| 7428 |
+
"step": 1053
|
| 7429 |
+
},
|
| 7430 |
+
{
|
| 7431 |
+
"epoch": 0.3372260438329867,
|
| 7432 |
+
"grad_norm": 5.961119651794434,
|
| 7433 |
+
"learning_rate": 0.00030033034128901996,
|
| 7434 |
+
"loss": 6.7396,
|
| 7435 |
+
"step": 1054
|
| 7436 |
+
},
|
| 7437 |
+
{
|
| 7438 |
+
"epoch": 0.3375459926411774,
|
| 7439 |
+
"grad_norm": 10.675737380981445,
|
| 7440 |
+
"learning_rate": 0.00030015811979634593,
|
| 7441 |
+
"loss": 6.4708,
|
| 7442 |
+
"step": 1055
|
| 7443 |
+
},
|
| 7444 |
+
{
|
| 7445 |
+
"epoch": 0.3378659414493681,
|
| 7446 |
+
"grad_norm": 8.396323204040527,
|
| 7447 |
+
"learning_rate": 0.00029998580264928906,
|
| 7448 |
+
"loss": 6.5822,
|
| 7449 |
+
"step": 1056
|
| 7450 |
+
},
|
| 7451 |
+
{
|
| 7452 |
+
"epoch": 0.3381858902575588,
|
| 7453 |
+
"grad_norm": 4.825063228607178,
|
| 7454 |
+
"learning_rate": 0.00029981339003067124,
|
| 7455 |
+
"loss": 6.6281,
|
| 7456 |
+
"step": 1057
|
| 7457 |
+
},
|
| 7458 |
+
{
|
| 7459 |
+
"epoch": 0.3385058390657495,
|
| 7460 |
+
"grad_norm": 12.628974914550781,
|
| 7461 |
+
"learning_rate": 0.0002996408821234158,
|
| 7462 |
+
"loss": 6.78,
|
| 7463 |
+
"step": 1058
|
| 7464 |
+
},
|
| 7465 |
+
{
|
| 7466 |
+
"epoch": 0.3388257878739402,
|
| 7467 |
+
"grad_norm": 9.515558242797852,
|
| 7468 |
+
"learning_rate": 0.00029946827911054723,
|
| 7469 |
+
"loss": 6.7697,
|
| 7470 |
+
"step": 1059
|
| 7471 |
+
},
|
| 7472 |
+
{
|
| 7473 |
+
"epoch": 0.3391457366821309,
|
| 7474 |
+
"grad_norm": 5.6787543296813965,
|
| 7475 |
+
"learning_rate": 0.00029929558117519086,
|
| 7476 |
+
"loss": 6.6566,
|
| 7477 |
+
"step": 1060
|
| 7478 |
+
},
|
| 7479 |
+
{
|
| 7480 |
+
"epoch": 0.3394656854903215,
|
| 7481 |
+
"grad_norm": 8.688566207885742,
|
| 7482 |
+
"learning_rate": 0.00029912278850057253,
|
| 7483 |
+
"loss": 6.6127,
|
| 7484 |
+
"step": 1061
|
| 7485 |
+
},
|
| 7486 |
+
{
|
| 7487 |
+
"epoch": 0.3397856342985122,
|
| 7488 |
+
"grad_norm": 6.97818660736084,
|
| 7489 |
+
"learning_rate": 0.0002989499012700191,
|
| 7490 |
+
"loss": 6.75,
|
| 7491 |
+
"step": 1062
|
| 7492 |
+
},
|
| 7493 |
+
{
|
| 7494 |
+
"epoch": 0.3401055831067029,
|
| 7495 |
+
"grad_norm": 8.22523021697998,
|
| 7496 |
+
"learning_rate": 0.0002987769196669572,
|
| 7497 |
+
"loss": 6.6283,
|
| 7498 |
+
"step": 1063
|
| 7499 |
+
},
|
| 7500 |
+
{
|
| 7501 |
+
"epoch": 0.3404255319148936,
|
| 7502 |
+
"grad_norm": 10.774093627929688,
|
| 7503 |
+
"learning_rate": 0.0002986038438749139,
|
| 7504 |
+
"loss": 6.611,
|
| 7505 |
+
"step": 1064
|
| 7506 |
+
},
|
| 7507 |
+
{
|
| 7508 |
+
"epoch": 0.3407454807230843,
|
| 7509 |
+
"grad_norm": 8.792933464050293,
|
| 7510 |
+
"learning_rate": 0.00029843067407751606,
|
| 7511 |
+
"loss": 6.5654,
|
| 7512 |
+
"step": 1065
|
| 7513 |
+
},
|
| 7514 |
+
{
|
| 7515 |
+
"epoch": 0.341065429531275,
|
| 7516 |
+
"grad_norm": 5.798463344573975,
|
| 7517 |
+
"learning_rate": 0.0002982574104584904,
|
| 7518 |
+
"loss": 6.7735,
|
| 7519 |
+
"step": 1066
|
| 7520 |
+
},
|
| 7521 |
+
{
|
| 7522 |
+
"epoch": 0.3413853783394657,
|
| 7523 |
+
"grad_norm": 7.097434997558594,
|
| 7524 |
+
"learning_rate": 0.00029808405320166313,
|
| 7525 |
+
"loss": 6.4257,
|
| 7526 |
+
"step": 1067
|
| 7527 |
+
},
|
| 7528 |
+
{
|
| 7529 |
+
"epoch": 0.3417053271476564,
|
| 7530 |
+
"grad_norm": 5.740554332733154,
|
| 7531 |
+
"learning_rate": 0.00029791060249095976,
|
| 7532 |
+
"loss": 6.6911,
|
| 7533 |
+
"step": 1068
|
| 7534 |
+
},
|
| 7535 |
+
{
|
| 7536 |
+
"epoch": 0.3420252759558471,
|
| 7537 |
+
"grad_norm": 6.828253746032715,
|
| 7538 |
+
"learning_rate": 0.00029773705851040503,
|
| 7539 |
+
"loss": 6.4999,
|
| 7540 |
+
"step": 1069
|
| 7541 |
+
},
|
| 7542 |
+
{
|
| 7543 |
+
"epoch": 0.3423452247640378,
|
| 7544 |
+
"grad_norm": 6.237061977386475,
|
| 7545 |
+
"learning_rate": 0.00029756342144412253,
|
| 7546 |
+
"loss": 6.3495,
|
| 7547 |
+
"step": 1070
|
| 7548 |
+
},
|
| 7549 |
+
{
|
| 7550 |
+
"epoch": 0.3426651735722284,
|
| 7551 |
+
"grad_norm": 5.364548206329346,
|
| 7552 |
+
"learning_rate": 0.00029738969147633475,
|
| 7553 |
+
"loss": 6.6146,
|
| 7554 |
+
"step": 1071
|
| 7555 |
+
},
|
| 7556 |
+
{
|
| 7557 |
+
"epoch": 0.3429851223804191,
|
| 7558 |
+
"grad_norm": 6.4642815589904785,
|
| 7559 |
+
"learning_rate": 0.0002972158687913626,
|
| 7560 |
+
"loss": 6.6634,
|
| 7561 |
+
"step": 1072
|
| 7562 |
+
},
|
| 7563 |
+
{
|
| 7564 |
+
"epoch": 0.3433050711886098,
|
| 7565 |
+
"grad_norm": 8.208625793457031,
|
| 7566 |
+
"learning_rate": 0.00029704195357362545,
|
| 7567 |
+
"loss": 6.5723,
|
| 7568 |
+
"step": 1073
|
| 7569 |
+
},
|
| 7570 |
+
{
|
| 7571 |
+
"epoch": 0.3436250199968005,
|
| 7572 |
+
"grad_norm": 10.248831748962402,
|
| 7573 |
+
"learning_rate": 0.00029686794600764085,
|
| 7574 |
+
"loss": 6.6042,
|
| 7575 |
+
"step": 1074
|
| 7576 |
+
},
|
| 7577 |
+
{
|
| 7578 |
+
"epoch": 0.3439449688049912,
|
| 7579 |
+
"grad_norm": 7.775475978851318,
|
| 7580 |
+
"learning_rate": 0.0002966938462780242,
|
| 7581 |
+
"loss": 6.3936,
|
| 7582 |
+
"step": 1075
|
| 7583 |
+
},
|
| 7584 |
+
{
|
| 7585 |
+
"epoch": 0.3442649176131819,
|
| 7586 |
+
"grad_norm": 7.968334197998047,
|
| 7587 |
+
"learning_rate": 0.000296519654569489,
|
| 7588 |
+
"loss": 6.5476,
|
| 7589 |
+
"step": 1076
|
| 7590 |
+
},
|
| 7591 |
+
{
|
| 7592 |
+
"epoch": 0.3445848664213726,
|
| 7593 |
+
"grad_norm": 5.832833766937256,
|
| 7594 |
+
"learning_rate": 0.00029634537106684587,
|
| 7595 |
+
"loss": 6.6565,
|
| 7596 |
+
"step": 1077
|
| 7597 |
+
},
|
| 7598 |
+
{
|
| 7599 |
+
"epoch": 0.3449048152295633,
|
| 7600 |
+
"grad_norm": 7.168907165527344,
|
| 7601 |
+
"learning_rate": 0.0002961709959550032,
|
| 7602 |
+
"loss": 6.4957,
|
| 7603 |
+
"step": 1078
|
| 7604 |
+
},
|
| 7605 |
+
{
|
| 7606 |
+
"epoch": 0.345224764037754,
|
| 7607 |
+
"grad_norm": 8.945564270019531,
|
| 7608 |
+
"learning_rate": 0.00029599652941896643,
|
| 7609 |
+
"loss": 6.7646,
|
| 7610 |
+
"step": 1079
|
| 7611 |
+
},
|
| 7612 |
+
{
|
| 7613 |
+
"epoch": 0.3455447128459446,
|
| 7614 |
+
"grad_norm": 5.535223960876465,
|
| 7615 |
+
"learning_rate": 0.00029582197164383796,
|
| 7616 |
+
"loss": 6.3885,
|
| 7617 |
+
"step": 1080
|
| 7618 |
+
},
|
| 7619 |
+
{
|
| 7620 |
+
"epoch": 0.3458646616541353,
|
| 7621 |
+
"grad_norm": 6.619454860687256,
|
| 7622 |
+
"learning_rate": 0.00029564732281481715,
|
| 7623 |
+
"loss": 6.5704,
|
| 7624 |
+
"step": 1081
|
| 7625 |
+
},
|
| 7626 |
+
{
|
| 7627 |
+
"epoch": 0.346184610462326,
|
| 7628 |
+
"grad_norm": 8.324727058410645,
|
| 7629 |
+
"learning_rate": 0.00029547258311719973,
|
| 7630 |
+
"loss": 6.4878,
|
| 7631 |
+
"step": 1082
|
| 7632 |
+
},
|
| 7633 |
+
{
|
| 7634 |
+
"epoch": 0.3465045592705167,
|
| 7635 |
+
"grad_norm": 6.584716320037842,
|
| 7636 |
+
"learning_rate": 0.00029529775273637803,
|
| 7637 |
+
"loss": 6.5915,
|
| 7638 |
+
"step": 1083
|
| 7639 |
+
},
|
| 7640 |
+
{
|
| 7641 |
+
"epoch": 0.3468245080787074,
|
| 7642 |
+
"grad_norm": 4.810291290283203,
|
| 7643 |
+
"learning_rate": 0.00029512283185784046,
|
| 7644 |
+
"loss": 6.3791,
|
| 7645 |
+
"step": 1084
|
| 7646 |
+
},
|
| 7647 |
+
{
|
| 7648 |
+
"epoch": 0.3471444568868981,
|
| 7649 |
+
"grad_norm": 29.184206008911133,
|
| 7650 |
+
"learning_rate": 0.0002949478206671717,
|
| 7651 |
+
"loss": 6.7601,
|
| 7652 |
+
"step": 1085
|
| 7653 |
+
},
|
| 7654 |
+
{
|
| 7655 |
+
"epoch": 0.3474644056950888,
|
| 7656 |
+
"grad_norm": 9.653064727783203,
|
| 7657 |
+
"learning_rate": 0.0002947727193500518,
|
| 7658 |
+
"loss": 6.6253,
|
| 7659 |
+
"step": 1086
|
| 7660 |
+
},
|
| 7661 |
+
{
|
| 7662 |
+
"epoch": 0.3477843545032795,
|
| 7663 |
+
"grad_norm": 13.180891990661621,
|
| 7664 |
+
"learning_rate": 0.0002945975280922569,
|
| 7665 |
+
"loss": 6.5926,
|
| 7666 |
+
"step": 1087
|
| 7667 |
+
},
|
| 7668 |
+
{
|
| 7669 |
+
"epoch": 0.3481043033114702,
|
| 7670 |
+
"grad_norm": 9.94357681274414,
|
| 7671 |
+
"learning_rate": 0.0002944222470796582,
|
| 7672 |
+
"loss": 6.6027,
|
| 7673 |
+
"step": 1088
|
| 7674 |
+
},
|
| 7675 |
+
{
|
| 7676 |
+
"epoch": 0.3484242521196608,
|
| 7677 |
+
"grad_norm": 9.287586212158203,
|
| 7678 |
+
"learning_rate": 0.0002942468764982223,
|
| 7679 |
+
"loss": 6.7376,
|
| 7680 |
+
"step": 1089
|
| 7681 |
+
},
|
| 7682 |
+
{
|
| 7683 |
+
"epoch": 0.3487442009278515,
|
| 7684 |
+
"grad_norm": 17.233806610107422,
|
| 7685 |
+
"learning_rate": 0.000294071416534011,
|
| 7686 |
+
"loss": 6.8934,
|
| 7687 |
+
"step": 1090
|
| 7688 |
+
},
|
| 7689 |
+
{
|
| 7690 |
+
"epoch": 0.3490641497360422,
|
| 7691 |
+
"grad_norm": 11.148905754089355,
|
| 7692 |
+
"learning_rate": 0.00029389586737318046,
|
| 7693 |
+
"loss": 6.4543,
|
| 7694 |
+
"step": 1091
|
| 7695 |
+
},
|
| 7696 |
+
{
|
| 7697 |
+
"epoch": 0.3493840985442329,
|
| 7698 |
+
"grad_norm": 7.293564319610596,
|
| 7699 |
+
"learning_rate": 0.00029372022920198186,
|
| 7700 |
+
"loss": 6.5114,
|
| 7701 |
+
"step": 1092
|
| 7702 |
+
},
|
| 7703 |
+
{
|
| 7704 |
+
"epoch": 0.3497040473524236,
|
| 7705 |
+
"grad_norm": 7.520803928375244,
|
| 7706 |
+
"learning_rate": 0.0002935445022067609,
|
| 7707 |
+
"loss": 6.7753,
|
| 7708 |
+
"step": 1093
|
| 7709 |
+
},
|
| 7710 |
+
{
|
| 7711 |
+
"epoch": 0.3500239961606143,
|
| 7712 |
+
"grad_norm": 8.508971214294434,
|
| 7713 |
+
"learning_rate": 0.00029336868657395704,
|
| 7714 |
+
"loss": 6.5503,
|
| 7715 |
+
"step": 1094
|
| 7716 |
+
},
|
| 7717 |
+
{
|
| 7718 |
+
"epoch": 0.350343944968805,
|
| 7719 |
+
"grad_norm": 7.20286750793457,
|
| 7720 |
+
"learning_rate": 0.0002931927824901042,
|
| 7721 |
+
"loss": 6.5862,
|
| 7722 |
+
"step": 1095
|
| 7723 |
+
},
|
| 7724 |
+
{
|
| 7725 |
+
"epoch": 0.3506638937769957,
|
| 7726 |
+
"grad_norm": 6.576849460601807,
|
| 7727 |
+
"learning_rate": 0.00029301679014183006,
|
| 7728 |
+
"loss": 6.5876,
|
| 7729 |
+
"step": 1096
|
| 7730 |
+
},
|
| 7731 |
+
{
|
| 7732 |
+
"epoch": 0.3509838425851864,
|
| 7733 |
+
"grad_norm": 5.83701753616333,
|
| 7734 |
+
"learning_rate": 0.0002928407097158558,
|
| 7735 |
+
"loss": 6.3791,
|
| 7736 |
+
"step": 1097
|
| 7737 |
+
},
|
| 7738 |
+
{
|
| 7739 |
+
"epoch": 0.3513037913933771,
|
| 7740 |
+
"grad_norm": 9.229516983032227,
|
| 7741 |
+
"learning_rate": 0.00029266454139899615,
|
| 7742 |
+
"loss": 6.6795,
|
| 7743 |
+
"step": 1098
|
| 7744 |
+
},
|
| 7745 |
+
{
|
| 7746 |
+
"epoch": 0.3516237402015677,
|
| 7747 |
+
"grad_norm": 16.494159698486328,
|
| 7748 |
+
"learning_rate": 0.00029248828537815924,
|
| 7749 |
+
"loss": 6.571,
|
| 7750 |
+
"step": 1099
|
| 7751 |
+
},
|
| 7752 |
+
{
|
| 7753 |
+
"epoch": 0.3519436890097584,
|
| 7754 |
+
"grad_norm": 10.0813627243042,
|
| 7755 |
+
"learning_rate": 0.00029231194184034596,
|
| 7756 |
+
"loss": 6.4374,
|
| 7757 |
+
"step": 1100
|
| 7758 |
+
},
|
| 7759 |
+
{
|
| 7760 |
+
"epoch": 0.3522636378179491,
|
| 7761 |
+
"grad_norm": 7.692905902862549,
|
| 7762 |
+
"learning_rate": 0.0002921355109726502,
|
| 7763 |
+
"loss": 6.6512,
|
| 7764 |
+
"step": 1101
|
| 7765 |
+
},
|
| 7766 |
+
{
|
| 7767 |
+
"epoch": 0.3525835866261398,
|
| 7768 |
+
"grad_norm": 15.104799270629883,
|
| 7769 |
+
"learning_rate": 0.00029195899296225846,
|
| 7770 |
+
"loss": 6.4528,
|
| 7771 |
+
"step": 1102
|
| 7772 |
+
},
|
| 7773 |
+
{
|
| 7774 |
+
"epoch": 0.3529035354343305,
|
| 7775 |
+
"grad_norm": 18.930479049682617,
|
| 7776 |
+
"learning_rate": 0.00029178238799644983,
|
| 7777 |
+
"loss": 6.4953,
|
| 7778 |
+
"step": 1103
|
| 7779 |
+
},
|
| 7780 |
+
{
|
| 7781 |
+
"epoch": 0.3532234842425212,
|
| 7782 |
+
"grad_norm": 10.989952087402344,
|
| 7783 |
+
"learning_rate": 0.00029160569626259555,
|
| 7784 |
+
"loss": 6.5066,
|
| 7785 |
+
"step": 1104
|
| 7786 |
+
},
|
| 7787 |
+
{
|
| 7788 |
+
"epoch": 0.3535434330507119,
|
| 7789 |
+
"grad_norm": 8.65145206451416,
|
| 7790 |
+
"learning_rate": 0.00029142891794815877,
|
| 7791 |
+
"loss": 6.7426,
|
| 7792 |
+
"step": 1105
|
| 7793 |
+
},
|
| 7794 |
+
{
|
| 7795 |
+
"epoch": 0.3538633818589026,
|
| 7796 |
+
"grad_norm": 4.995960235595703,
|
| 7797 |
+
"learning_rate": 0.00029125205324069477,
|
| 7798 |
+
"loss": 6.4861,
|
| 7799 |
+
"step": 1106
|
| 7800 |
+
},
|
| 7801 |
+
{
|
| 7802 |
+
"epoch": 0.3541833306670933,
|
| 7803 |
+
"grad_norm": 6.398393630981445,
|
| 7804 |
+
"learning_rate": 0.00029107510232785036,
|
| 7805 |
+
"loss": 6.3449,
|
| 7806 |
+
"step": 1107
|
| 7807 |
+
},
|
| 7808 |
+
{
|
| 7809 |
+
"epoch": 0.35450327947528393,
|
| 7810 |
+
"grad_norm": 5.589430809020996,
|
| 7811 |
+
"learning_rate": 0.00029089806539736384,
|
| 7812 |
+
"loss": 6.5538,
|
| 7813 |
+
"step": 1108
|
| 7814 |
+
},
|
| 7815 |
+
{
|
| 7816 |
+
"epoch": 0.3548232282834746,
|
| 7817 |
+
"grad_norm": 4.985332012176514,
|
| 7818 |
+
"learning_rate": 0.0002907209426370647,
|
| 7819 |
+
"loss": 6.544,
|
| 7820 |
+
"step": 1109
|
| 7821 |
+
},
|
| 7822 |
+
{
|
| 7823 |
+
"epoch": 0.3551431770916653,
|
| 7824 |
+
"grad_norm": 5.552674293518066,
|
| 7825 |
+
"learning_rate": 0.00029054373423487357,
|
| 7826 |
+
"loss": 6.5413,
|
| 7827 |
+
"step": 1110
|
| 7828 |
+
},
|
| 7829 |
+
{
|
| 7830 |
+
"epoch": 0.355463125899856,
|
| 7831 |
+
"grad_norm": 7.634738922119141,
|
| 7832 |
+
"learning_rate": 0.00029036644037880186,
|
| 7833 |
+
"loss": 6.475,
|
| 7834 |
+
"step": 1111
|
| 7835 |
+
},
|
| 7836 |
+
{
|
| 7837 |
+
"epoch": 0.3557830747080467,
|
| 7838 |
+
"grad_norm": 11.79802131652832,
|
| 7839 |
+
"learning_rate": 0.0002901890612569518,
|
| 7840 |
+
"loss": 6.8223,
|
| 7841 |
+
"step": 1112
|
| 7842 |
+
},
|
| 7843 |
+
{
|
| 7844 |
+
"epoch": 0.3561030235162374,
|
| 7845 |
+
"grad_norm": 13.318756103515625,
|
| 7846 |
+
"learning_rate": 0.000290011597057516,
|
| 7847 |
+
"loss": 6.2933,
|
| 7848 |
+
"step": 1113
|
| 7849 |
+
},
|
| 7850 |
+
{
|
| 7851 |
+
"epoch": 0.3564229723244281,
|
| 7852 |
+
"grad_norm": 14.60718822479248,
|
| 7853 |
+
"learning_rate": 0.00028983404796877716,
|
| 7854 |
+
"loss": 6.803,
|
| 7855 |
+
"step": 1114
|
| 7856 |
+
},
|
| 7857 |
+
{
|
| 7858 |
+
"epoch": 0.3567429211326188,
|
| 7859 |
+
"grad_norm": 8.220056533813477,
|
| 7860 |
+
"learning_rate": 0.00028965641417910833,
|
| 7861 |
+
"loss": 6.4244,
|
| 7862 |
+
"step": 1115
|
| 7863 |
+
},
|
| 7864 |
+
{
|
| 7865 |
+
"epoch": 0.3570628699408095,
|
| 7866 |
+
"grad_norm": 5.330780506134033,
|
| 7867 |
+
"learning_rate": 0.0002894786958769723,
|
| 7868 |
+
"loss": 6.4827,
|
| 7869 |
+
"step": 1116
|
| 7870 |
+
},
|
| 7871 |
+
{
|
| 7872 |
+
"epoch": 0.3573828187490002,
|
| 7873 |
+
"grad_norm": 7.304154872894287,
|
| 7874 |
+
"learning_rate": 0.00028930089325092146,
|
| 7875 |
+
"loss": 6.4375,
|
| 7876 |
+
"step": 1117
|
| 7877 |
+
},
|
| 7878 |
+
{
|
| 7879 |
+
"epoch": 0.3577027675571908,
|
| 7880 |
+
"grad_norm": 7.895937442779541,
|
| 7881 |
+
"learning_rate": 0.0002891230064895977,
|
| 7882 |
+
"loss": 6.5026,
|
| 7883 |
+
"step": 1118
|
| 7884 |
+
},
|
| 7885 |
+
{
|
| 7886 |
+
"epoch": 0.3580227163653815,
|
| 7887 |
+
"grad_norm": 8.539824485778809,
|
| 7888 |
+
"learning_rate": 0.0002889450357817324,
|
| 7889 |
+
"loss": 6.5781,
|
| 7890 |
+
"step": 1119
|
| 7891 |
+
},
|
| 7892 |
+
{
|
| 7893 |
+
"epoch": 0.3583426651735722,
|
| 7894 |
+
"grad_norm": 11.084145545959473,
|
| 7895 |
+
"learning_rate": 0.0002887669813161455,
|
| 7896 |
+
"loss": 6.4965,
|
| 7897 |
+
"step": 1120
|
| 7898 |
+
},
|
| 7899 |
+
{
|
| 7900 |
+
"epoch": 0.3586626139817629,
|
| 7901 |
+
"grad_norm": 8.87822151184082,
|
| 7902 |
+
"learning_rate": 0.00028858884328174635,
|
| 7903 |
+
"loss": 6.2127,
|
| 7904 |
+
"step": 1121
|
| 7905 |
+
},
|
| 7906 |
+
{
|
| 7907 |
+
"epoch": 0.3589825627899536,
|
| 7908 |
+
"grad_norm": 12.374408721923828,
|
| 7909 |
+
"learning_rate": 0.0002884106218675326,
|
| 7910 |
+
"loss": 6.3412,
|
| 7911 |
+
"step": 1122
|
| 7912 |
+
},
|
| 7913 |
+
{
|
| 7914 |
+
"epoch": 0.3593025115981443,
|
| 7915 |
+
"grad_norm": 7.8099470138549805,
|
| 7916 |
+
"learning_rate": 0.0002882323172625906,
|
| 7917 |
+
"loss": 6.2976,
|
| 7918 |
+
"step": 1123
|
| 7919 |
+
},
|
| 7920 |
+
{
|
| 7921 |
+
"epoch": 0.359622460406335,
|
| 7922 |
+
"grad_norm": 9.879372596740723,
|
| 7923 |
+
"learning_rate": 0.0002880539296560947,
|
| 7924 |
+
"loss": 6.3613,
|
| 7925 |
+
"step": 1124
|
| 7926 |
+
},
|
| 7927 |
+
{
|
| 7928 |
+
"epoch": 0.3599424092145257,
|
| 7929 |
+
"grad_norm": 18.98383140563965,
|
| 7930 |
+
"learning_rate": 0.0002878754592373075,
|
| 7931 |
+
"loss": 6.564,
|
| 7932 |
+
"step": 1125
|
| 7933 |
+
},
|
| 7934 |
+
{
|
| 7935 |
+
"epoch": 0.3602623580227164,
|
| 7936 |
+
"grad_norm": 27.977420806884766,
|
| 7937 |
+
"learning_rate": 0.00028769690619557945,
|
| 7938 |
+
"loss": 6.2547,
|
| 7939 |
+
"step": 1126
|
| 7940 |
+
},
|
| 7941 |
+
{
|
| 7942 |
+
"epoch": 0.36058230683090703,
|
| 7943 |
+
"grad_norm": 9.814420700073242,
|
| 7944 |
+
"learning_rate": 0.00028751827072034876,
|
| 7945 |
+
"loss": 6.447,
|
| 7946 |
+
"step": 1127
|
| 7947 |
+
},
|
| 7948 |
+
{
|
| 7949 |
+
"epoch": 0.3609022556390977,
|
| 7950 |
+
"grad_norm": 9.564621925354004,
|
| 7951 |
+
"learning_rate": 0.00028733955300114066,
|
| 7952 |
+
"loss": 6.6715,
|
| 7953 |
+
"step": 1128
|
| 7954 |
+
},
|
| 7955 |
+
{
|
| 7956 |
+
"epoch": 0.3612222044472884,
|
| 7957 |
+
"grad_norm": 7.441284656524658,
|
| 7958 |
+
"learning_rate": 0.00028716075322756827,
|
| 7959 |
+
"loss": 6.3489,
|
| 7960 |
+
"step": 1129
|
| 7961 |
+
},
|
| 7962 |
+
{
|
| 7963 |
+
"epoch": 0.3615421532554791,
|
| 7964 |
+
"grad_norm": 6.929698944091797,
|
| 7965 |
+
"learning_rate": 0.0002869818715893312,
|
| 7966 |
+
"loss": 6.6691,
|
| 7967 |
+
"step": 1130
|
| 7968 |
+
},
|
| 7969 |
+
{
|
| 7970 |
+
"epoch": 0.3618621020636698,
|
| 7971 |
+
"grad_norm": 9.472748756408691,
|
| 7972 |
+
"learning_rate": 0.0002868029082762163,
|
| 7973 |
+
"loss": 6.4778,
|
| 7974 |
+
"step": 1131
|
| 7975 |
+
},
|
| 7976 |
+
{
|
| 7977 |
+
"epoch": 0.3621820508718605,
|
| 7978 |
+
"grad_norm": 7.518265724182129,
|
| 7979 |
+
"learning_rate": 0.00028662386347809687,
|
| 7980 |
+
"loss": 6.5739,
|
| 7981 |
+
"step": 1132
|
| 7982 |
+
},
|
| 7983 |
+
{
|
| 7984 |
+
"epoch": 0.3625019996800512,
|
| 7985 |
+
"grad_norm": 7.980709552764893,
|
| 7986 |
+
"learning_rate": 0.00028644473738493275,
|
| 7987 |
+
"loss": 6.5246,
|
| 7988 |
+
"step": 1133
|
| 7989 |
+
},
|
| 7990 |
+
{
|
| 7991 |
+
"epoch": 0.3628219484882419,
|
| 7992 |
+
"grad_norm": 7.017282009124756,
|
| 7993 |
+
"learning_rate": 0.00028626553018677,
|
| 7994 |
+
"loss": 6.2618,
|
| 7995 |
+
"step": 1134
|
| 7996 |
+
},
|
| 7997 |
+
{
|
| 7998 |
+
"epoch": 0.3631418972964326,
|
| 7999 |
+
"grad_norm": 8.980280876159668,
|
| 8000 |
+
"learning_rate": 0.0002860862420737407,
|
| 8001 |
+
"loss": 6.4679,
|
| 8002 |
+
"step": 1135
|
| 8003 |
+
},
|
| 8004 |
+
{
|
| 8005 |
+
"epoch": 0.3634618461046233,
|
| 8006 |
+
"grad_norm": 11.132031440734863,
|
| 8007 |
+
"learning_rate": 0.0002859068732360628,
|
| 8008 |
+
"loss": 6.4575,
|
| 8009 |
+
"step": 1136
|
| 8010 |
+
},
|
| 8011 |
+
{
|
| 8012 |
+
"epoch": 0.3637817949128139,
|
| 8013 |
+
"grad_norm": 10.001900672912598,
|
| 8014 |
+
"learning_rate": 0.00028572742386404,
|
| 8015 |
+
"loss": 6.4749,
|
| 8016 |
+
"step": 1137
|
| 8017 |
+
},
|
| 8018 |
+
{
|
| 8019 |
+
"epoch": 0.3641017437210046,
|
| 8020 |
+
"grad_norm": 8.816960334777832,
|
| 8021 |
+
"learning_rate": 0.0002855478941480613,
|
| 8022 |
+
"loss": 6.6358,
|
| 8023 |
+
"step": 1138
|
| 8024 |
+
},
|
| 8025 |
+
{
|
| 8026 |
+
"epoch": 0.3644216925291953,
|
| 8027 |
+
"grad_norm": 7.853034019470215,
|
| 8028 |
+
"learning_rate": 0.0002853682842786009,
|
| 8029 |
+
"loss": 6.4368,
|
| 8030 |
+
"step": 1139
|
| 8031 |
+
},
|
| 8032 |
+
{
|
| 8033 |
+
"epoch": 0.364741641337386,
|
| 8034 |
+
"grad_norm": 7.836308479309082,
|
| 8035 |
+
"learning_rate": 0.00028518859444621834,
|
| 8036 |
+
"loss": 6.3961,
|
| 8037 |
+
"step": 1140
|
| 8038 |
+
},
|
| 8039 |
+
{
|
| 8040 |
+
"epoch": 0.3650615901455767,
|
| 8041 |
+
"grad_norm": 7.645723819732666,
|
| 8042 |
+
"learning_rate": 0.0002850088248415577,
|
| 8043 |
+
"loss": 6.3327,
|
| 8044 |
+
"step": 1141
|
| 8045 |
+
},
|
| 8046 |
+
{
|
| 8047 |
+
"epoch": 0.3653815389537674,
|
| 8048 |
+
"grad_norm": 5.902100563049316,
|
| 8049 |
+
"learning_rate": 0.00028482897565534763,
|
| 8050 |
+
"loss": 6.418,
|
| 8051 |
+
"step": 1142
|
| 8052 |
+
},
|
| 8053 |
+
{
|
| 8054 |
+
"epoch": 0.3657014877619581,
|
| 8055 |
+
"grad_norm": 6.9731292724609375,
|
| 8056 |
+
"learning_rate": 0.0002846490470784016,
|
| 8057 |
+
"loss": 6.1619,
|
| 8058 |
+
"step": 1143
|
| 8059 |
+
},
|
| 8060 |
+
{
|
| 8061 |
+
"epoch": 0.3660214365701488,
|
| 8062 |
+
"grad_norm": 9.789436340332031,
|
| 8063 |
+
"learning_rate": 0.00028446903930161695,
|
| 8064 |
+
"loss": 6.6124,
|
| 8065 |
+
"step": 1144
|
| 8066 |
+
},
|
| 8067 |
+
{
|
| 8068 |
+
"epoch": 0.3663413853783395,
|
| 8069 |
+
"grad_norm": 6.7149577140808105,
|
| 8070 |
+
"learning_rate": 0.0002842889525159753,
|
| 8071 |
+
"loss": 6.5035,
|
| 8072 |
+
"step": 1145
|
| 8073 |
+
},
|
| 8074 |
+
{
|
| 8075 |
+
"epoch": 0.36666133418653013,
|
| 8076 |
+
"grad_norm": 7.927312850952148,
|
| 8077 |
+
"learning_rate": 0.00028410878691254173,
|
| 8078 |
+
"loss": 6.3137,
|
| 8079 |
+
"step": 1146
|
| 8080 |
+
},
|
| 8081 |
+
{
|
| 8082 |
+
"epoch": 0.3669812829947208,
|
| 8083 |
+
"grad_norm": 6.585173606872559,
|
| 8084 |
+
"learning_rate": 0.00028392854268246545,
|
| 8085 |
+
"loss": 6.3402,
|
| 8086 |
+
"step": 1147
|
| 8087 |
+
},
|
| 8088 |
+
{
|
| 8089 |
+
"epoch": 0.3673012318029115,
|
| 8090 |
+
"grad_norm": 7.843646049499512,
|
| 8091 |
+
"learning_rate": 0.0002837482200169787,
|
| 8092 |
+
"loss": 6.4396,
|
| 8093 |
+
"step": 1148
|
| 8094 |
+
},
|
| 8095 |
+
{
|
| 8096 |
+
"epoch": 0.3676211806111022,
|
| 8097 |
+
"grad_norm": 6.062797546386719,
|
| 8098 |
+
"learning_rate": 0.0002835678191073971,
|
| 8099 |
+
"loss": 5.9964,
|
| 8100 |
+
"step": 1149
|
| 8101 |
+
},
|
| 8102 |
+
{
|
| 8103 |
+
"epoch": 0.3679411294192929,
|
| 8104 |
+
"grad_norm": 7.792223930358887,
|
| 8105 |
+
"learning_rate": 0.0002833873401451192,
|
| 8106 |
+
"loss": 6.4738,
|
| 8107 |
+
"step": 1150
|
| 8108 |
+
},
|
| 8109 |
+
{
|
| 8110 |
+
"epoch": 0.3682610782274836,
|
| 8111 |
+
"grad_norm": 21.281097412109375,
|
| 8112 |
+
"learning_rate": 0.00028320678332162646,
|
| 8113 |
+
"loss": 6.2936,
|
| 8114 |
+
"step": 1151
|
| 8115 |
+
},
|
| 8116 |
+
{
|
| 8117 |
+
"epoch": 0.3685810270356743,
|
| 8118 |
+
"grad_norm": 33.57594299316406,
|
| 8119 |
+
"learning_rate": 0.0002830261488284829,
|
| 8120 |
+
"loss": 6.4598,
|
| 8121 |
+
"step": 1152
|
| 8122 |
+
},
|
| 8123 |
+
{
|
| 8124 |
+
"epoch": 0.368900975843865,
|
| 8125 |
+
"grad_norm": 38.78255844116211,
|
| 8126 |
+
"learning_rate": 0.0002828454368573348,
|
| 8127 |
+
"loss": 6.349,
|
| 8128 |
+
"step": 1153
|
| 8129 |
+
},
|
| 8130 |
+
{
|
| 8131 |
+
"epoch": 0.3692209246520557,
|
| 8132 |
+
"grad_norm": 21.15943717956543,
|
| 8133 |
+
"learning_rate": 0.00028266464759991105,
|
| 8134 |
+
"loss": 6.493,
|
| 8135 |
+
"step": 1154
|
| 8136 |
+
},
|
| 8137 |
+
{
|
| 8138 |
+
"epoch": 0.36954087346024633,
|
| 8139 |
+
"grad_norm": 870.9165649414062,
|
| 8140 |
+
"learning_rate": 0.00028248378124802204,
|
| 8141 |
+
"loss": 6.4243,
|
| 8142 |
+
"step": 1155
|
| 8143 |
+
},
|
| 8144 |
+
{
|
| 8145 |
+
"epoch": 0.36986082226843703,
|
| 8146 |
+
"grad_norm": 67.86760711669922,
|
| 8147 |
+
"learning_rate": 0.00028230283799356024,
|
| 8148 |
+
"loss": 6.4124,
|
| 8149 |
+
"step": 1156
|
| 8150 |
+
},
|
| 8151 |
+
{
|
| 8152 |
+
"epoch": 0.3701807710766277,
|
| 8153 |
+
"grad_norm": 148.6990966796875,
|
| 8154 |
+
"learning_rate": 0.00028212181802849973,
|
| 8155 |
+
"loss": 6.3447,
|
| 8156 |
+
"step": 1157
|
| 8157 |
+
},
|
| 8158 |
+
{
|
| 8159 |
+
"epoch": 0.3705007198848184,
|
| 8160 |
+
"grad_norm": 224.3986358642578,
|
| 8161 |
+
"learning_rate": 0.0002819407215448958,
|
| 8162 |
+
"loss": 6.8178,
|
| 8163 |
+
"step": 1158
|
| 8164 |
+
},
|
| 8165 |
+
{
|
| 8166 |
+
"epoch": 0.3708206686930091,
|
| 8167 |
+
"grad_norm": 116.84649658203125,
|
| 8168 |
+
"learning_rate": 0.0002817595487348851,
|
| 8169 |
+
"loss": 6.6928,
|
| 8170 |
+
"step": 1159
|
| 8171 |
+
},
|
| 8172 |
+
{
|
| 8173 |
+
"epoch": 0.3711406175011998,
|
| 8174 |
+
"grad_norm": 5677.07763671875,
|
| 8175 |
+
"learning_rate": 0.0002815782997906852,
|
| 8176 |
+
"loss": 6.7519,
|
| 8177 |
+
"step": 1160
|
| 8178 |
+
},
|
| 8179 |
+
{
|
| 8180 |
+
"epoch": 0.3714605663093905,
|
| 8181 |
+
"grad_norm": 1888.601806640625,
|
| 8182 |
+
"learning_rate": 0.0002813969749045943,
|
| 8183 |
+
"loss": 6.6966,
|
| 8184 |
+
"step": 1161
|
| 8185 |
+
},
|
| 8186 |
+
{
|
| 8187 |
+
"epoch": 0.3717805151175812,
|
| 8188 |
+
"grad_norm": 8388.8984375,
|
| 8189 |
+
"learning_rate": 0.00028121557426899154,
|
| 8190 |
+
"loss": 6.9312,
|
| 8191 |
+
"step": 1162
|
| 8192 |
+
},
|
| 8193 |
+
{
|
| 8194 |
+
"epoch": 0.3721004639257719,
|
| 8195 |
+
"grad_norm": 9415.212890625,
|
| 8196 |
+
"learning_rate": 0.00028103409807633595,
|
| 8197 |
+
"loss": 7.3199,
|
| 8198 |
+
"step": 1163
|
| 8199 |
+
},
|
| 8200 |
+
{
|
| 8201 |
+
"epoch": 0.3724204127339626,
|
| 8202 |
+
"grad_norm": 678.4221801757812,
|
| 8203 |
+
"learning_rate": 0.000280852546519167,
|
| 8204 |
+
"loss": 7.6811,
|
| 8205 |
+
"step": 1164
|
| 8206 |
+
},
|
| 8207 |
+
{
|
| 8208 |
+
"epoch": 0.37274036154215323,
|
| 8209 |
+
"grad_norm": 33225.8359375,
|
| 8210 |
+
"learning_rate": 0.0002806709197901042,
|
| 8211 |
+
"loss": 7.352,
|
| 8212 |
+
"step": 1165
|
| 8213 |
+
},
|
| 8214 |
+
{
|
| 8215 |
+
"epoch": 0.3730603103503439,
|
| 8216 |
+
"grad_norm": 23.9914493560791,
|
| 8217 |
+
"learning_rate": 0.00028048921808184667,
|
| 8218 |
+
"loss": 7.4617,
|
| 8219 |
+
"step": 1166
|
| 8220 |
+
},
|
| 8221 |
+
{
|
| 8222 |
+
"epoch": 0.3733802591585346,
|
| 8223 |
+
"grad_norm": 100466.765625,
|
| 8224 |
+
"learning_rate": 0.000280307441587173,
|
| 8225 |
+
"loss": 6.8622,
|
| 8226 |
+
"step": 1167
|
| 8227 |
+
},
|
| 8228 |
+
{
|
| 8229 |
+
"epoch": 0.3737002079667253,
|
| 8230 |
+
"grad_norm": 165.09849548339844,
|
| 8231 |
+
"learning_rate": 0.00028012559049894135,
|
| 8232 |
+
"loss": 7.0926,
|
| 8233 |
+
"step": 1168
|
| 8234 |
+
},
|
| 8235 |
+
{
|
| 8236 |
+
"epoch": 0.374020156774916,
|
| 8237 |
+
"grad_norm": 143.38955688476562,
|
| 8238 |
+
"learning_rate": 0.0002799436650100889,
|
| 8239 |
+
"loss": 7.2583,
|
| 8240 |
+
"step": 1169
|
| 8241 |
+
},
|
| 8242 |
+
{
|
| 8243 |
+
"epoch": 0.3743401055831067,
|
| 8244 |
+
"grad_norm": 3611.675048828125,
|
| 8245 |
+
"learning_rate": 0.0002797616653136316,
|
| 8246 |
+
"loss": 7.6278,
|
| 8247 |
+
"step": 1170
|
| 8248 |
+
},
|
| 8249 |
+
{
|
| 8250 |
+
"epoch": 0.3746600543912974,
|
| 8251 |
+
"grad_norm": 1785.349365234375,
|
| 8252 |
+
"learning_rate": 0.0002795795916026645,
|
| 8253 |
+
"loss": 7.9622,
|
| 8254 |
+
"step": 1171
|
| 8255 |
+
},
|
| 8256 |
+
{
|
| 8257 |
+
"epoch": 0.3749800031994881,
|
| 8258 |
+
"grad_norm": 273.0044860839844,
|
| 8259 |
+
"learning_rate": 0.0002793974440703608,
|
| 8260 |
+
"loss": 8.2006,
|
| 8261 |
+
"step": 1172
|
| 8262 |
+
},
|
| 8263 |
+
{
|
| 8264 |
+
"epoch": 0.3752999520076788,
|
| 8265 |
+
"grad_norm": 1025.5972900390625,
|
| 8266 |
+
"learning_rate": 0.00027921522290997247,
|
| 8267 |
+
"loss": 7.8807,
|
| 8268 |
+
"step": 1173
|
| 8269 |
+
},
|
| 8270 |
+
{
|
| 8271 |
+
"epoch": 0.37561990081586943,
|
| 8272 |
+
"grad_norm": 7521.10107421875,
|
| 8273 |
+
"learning_rate": 0.00027903292831482905,
|
| 8274 |
+
"loss": 7.7593,
|
| 8275 |
+
"step": 1174
|
| 8276 |
+
},
|
| 8277 |
+
{
|
| 8278 |
+
"epoch": 0.37593984962406013,
|
| 8279 |
+
"grad_norm": 276.4880676269531,
|
| 8280 |
+
"learning_rate": 0.0002788505604783383,
|
| 8281 |
+
"loss": 7.5434,
|
| 8282 |
+
"step": 1175
|
| 8283 |
+
},
|
| 8284 |
+
{
|
| 8285 |
+
"epoch": 0.3762597984322508,
|
| 8286 |
+
"grad_norm": 286.2404479980469,
|
| 8287 |
+
"learning_rate": 0.00027866811959398585,
|
| 8288 |
+
"loss": 7.0517,
|
| 8289 |
+
"step": 1176
|
| 8290 |
+
},
|
| 8291 |
+
{
|
| 8292 |
+
"epoch": 0.3765797472404415,
|
| 8293 |
+
"grad_norm": 31.163135528564453,
|
| 8294 |
+
"learning_rate": 0.0002784856058553345,
|
| 8295 |
+
"loss": 7.2771,
|
| 8296 |
+
"step": 1177
|
| 8297 |
+
},
|
| 8298 |
+
{
|
| 8299 |
+
"epoch": 0.3768996960486322,
|
| 8300 |
+
"grad_norm": 22.811885833740234,
|
| 8301 |
+
"learning_rate": 0.00027830301945602445,
|
| 8302 |
+
"loss": 7.0636,
|
| 8303 |
+
"step": 1178
|
| 8304 |
+
},
|
| 8305 |
+
{
|
| 8306 |
+
"epoch": 0.3772196448568229,
|
| 8307 |
+
"grad_norm": 449.5051574707031,
|
| 8308 |
+
"learning_rate": 0.000278120360589773,
|
| 8309 |
+
"loss": 6.9615,
|
| 8310 |
+
"step": 1179
|
| 8311 |
+
},
|
| 8312 |
+
{
|
| 8313 |
+
"epoch": 0.3775395936650136,
|
| 8314 |
+
"grad_norm": 16.285924911499023,
|
| 8315 |
+
"learning_rate": 0.0002779376294503745,
|
| 8316 |
+
"loss": 7.1608,
|
| 8317 |
+
"step": 1180
|
| 8318 |
+
},
|
| 8319 |
+
{
|
| 8320 |
+
"epoch": 0.3778595424732043,
|
| 8321 |
+
"grad_norm": 16.230222702026367,
|
| 8322 |
+
"learning_rate": 0.00027775482623169977,
|
| 8323 |
+
"loss": 6.7927,
|
| 8324 |
+
"step": 1181
|
| 8325 |
+
},
|
| 8326 |
+
{
|
| 8327 |
+
"epoch": 0.378179491281395,
|
| 8328 |
+
"grad_norm": 9.623656272888184,
|
| 8329 |
+
"learning_rate": 0.0002775719511276961,
|
| 8330 |
+
"loss": 6.8221,
|
| 8331 |
+
"step": 1182
|
| 8332 |
+
},
|
| 8333 |
+
{
|
| 8334 |
+
"epoch": 0.3784994400895857,
|
| 8335 |
+
"grad_norm": 9.932913780212402,
|
| 8336 |
+
"learning_rate": 0.00027738900433238716,
|
| 8337 |
+
"loss": 6.6573,
|
| 8338 |
+
"step": 1183
|
| 8339 |
+
},
|
| 8340 |
+
{
|
| 8341 |
+
"epoch": 0.37881938889777633,
|
| 8342 |
+
"grad_norm": 7.48627233505249,
|
| 8343 |
+
"learning_rate": 0.0002772059860398726,
|
| 8344 |
+
"loss": 6.6814,
|
| 8345 |
+
"step": 1184
|
| 8346 |
+
},
|
| 8347 |
+
{
|
| 8348 |
+
"epoch": 0.37913933770596703,
|
| 8349 |
+
"grad_norm": 7.055801868438721,
|
| 8350 |
+
"learning_rate": 0.00027702289644432804,
|
| 8351 |
+
"loss": 6.5999,
|
| 8352 |
+
"step": 1185
|
| 8353 |
+
},
|
| 8354 |
+
{
|
| 8355 |
+
"epoch": 0.3794592865141577,
|
| 8356 |
+
"grad_norm": 8.798673629760742,
|
| 8357 |
+
"learning_rate": 0.00027683973574000464,
|
| 8358 |
+
"loss": 6.5275,
|
| 8359 |
+
"step": 1186
|
| 8360 |
+
},
|
| 8361 |
+
{
|
| 8362 |
+
"epoch": 0.3797792353223484,
|
| 8363 |
+
"grad_norm": 17.74211883544922,
|
| 8364 |
+
"learning_rate": 0.000276656504121229,
|
| 8365 |
+
"loss": 6.732,
|
| 8366 |
+
"step": 1187
|
| 8367 |
+
},
|
| 8368 |
+
{
|
| 8369 |
+
"epoch": 0.3800991841305391,
|
| 8370 |
+
"grad_norm": 11.82168960571289,
|
| 8371 |
+
"learning_rate": 0.000276473201782403,
|
| 8372 |
+
"loss": 6.5822,
|
| 8373 |
+
"step": 1188
|
| 8374 |
+
},
|
| 8375 |
+
{
|
| 8376 |
+
"epoch": 0.3804191329387298,
|
| 8377 |
+
"grad_norm": 8.84131908416748,
|
| 8378 |
+
"learning_rate": 0.0002762898289180036,
|
| 8379 |
+
"loss": 6.7404,
|
| 8380 |
+
"step": 1189
|
| 8381 |
+
},
|
| 8382 |
+
{
|
| 8383 |
+
"epoch": 0.3807390817469205,
|
| 8384 |
+
"grad_norm": 9.530952453613281,
|
| 8385 |
+
"learning_rate": 0.00027610638572258254,
|
| 8386 |
+
"loss": 6.5298,
|
| 8387 |
+
"step": 1190
|
| 8388 |
+
},
|
| 8389 |
+
{
|
| 8390 |
+
"epoch": 0.3810590305551112,
|
| 8391 |
+
"grad_norm": 15.368577003479004,
|
| 8392 |
+
"learning_rate": 0.0002759228723907661,
|
| 8393 |
+
"loss": 6.4386,
|
| 8394 |
+
"step": 1191
|
| 8395 |
+
},
|
| 8396 |
+
{
|
| 8397 |
+
"epoch": 0.3813789793633019,
|
| 8398 |
+
"grad_norm": 19.41617202758789,
|
| 8399 |
+
"learning_rate": 0.000275739289117255,
|
| 8400 |
+
"loss": 6.7247,
|
| 8401 |
+
"step": 1192
|
| 8402 |
+
},
|
| 8403 |
+
{
|
| 8404 |
+
"epoch": 0.38169892817149254,
|
| 8405 |
+
"grad_norm": 10.758553504943848,
|
| 8406 |
+
"learning_rate": 0.0002755556360968244,
|
| 8407 |
+
"loss": 6.4879,
|
| 8408 |
+
"step": 1193
|
| 8409 |
+
},
|
| 8410 |
+
{
|
| 8411 |
+
"epoch": 0.38201887697968323,
|
| 8412 |
+
"grad_norm": 6.776547431945801,
|
| 8413 |
+
"learning_rate": 0.000275371913524323,
|
| 8414 |
+
"loss": 6.4938,
|
| 8415 |
+
"step": 1194
|
| 8416 |
+
},
|
| 8417 |
+
{
|
| 8418 |
+
"epoch": 0.3823388257878739,
|
| 8419 |
+
"grad_norm": 9.408045768737793,
|
| 8420 |
+
"learning_rate": 0.0002751881215946738,
|
| 8421 |
+
"loss": 6.5052,
|
| 8422 |
+
"step": 1195
|
| 8423 |
+
},
|
| 8424 |
+
{
|
| 8425 |
+
"epoch": 0.3826587745960646,
|
| 8426 |
+
"grad_norm": 9.227543830871582,
|
| 8427 |
+
"learning_rate": 0.00027500426050287293,
|
| 8428 |
+
"loss": 6.4985,
|
| 8429 |
+
"step": 1196
|
| 8430 |
+
},
|
| 8431 |
+
{
|
| 8432 |
+
"epoch": 0.3829787234042553,
|
| 8433 |
+
"grad_norm": 114.55646514892578,
|
| 8434 |
+
"learning_rate": 0.0002748203304439903,
|
| 8435 |
+
"loss": 6.5849,
|
| 8436 |
+
"step": 1197
|
| 8437 |
+
},
|
| 8438 |
+
{
|
| 8439 |
+
"epoch": 0.383298672212446,
|
| 8440 |
+
"grad_norm": 588.6021728515625,
|
| 8441 |
+
"learning_rate": 0.0002746363316131687,
|
| 8442 |
+
"loss": 6.7832,
|
| 8443 |
+
"step": 1198
|
| 8444 |
+
},
|
| 8445 |
+
{
|
| 8446 |
+
"epoch": 0.3836186210206367,
|
| 8447 |
+
"grad_norm": 147.17431640625,
|
| 8448 |
+
"learning_rate": 0.000274452264205624,
|
| 8449 |
+
"loss": 6.6167,
|
| 8450 |
+
"step": 1199
|
| 8451 |
+
},
|
| 8452 |
+
{
|
| 8453 |
+
"epoch": 0.3839385698288274,
|
| 8454 |
+
"grad_norm": 28.505613327026367,
|
| 8455 |
+
"learning_rate": 0.00027426812841664487,
|
| 8456 |
+
"loss": 6.6407,
|
| 8457 |
+
"step": 1200
|
| 8458 |
+
},
|
| 8459 |
+
{
|
| 8460 |
+
"epoch": 0.3839385698288274,
|
| 8461 |
+
"eval_loss": 3.3596253395080566,
|
| 8462 |
+
"eval_runtime": 233.5546,
|
| 8463 |
+
"eval_samples_per_second": 5.635,
|
| 8464 |
+
"eval_steps_per_second": 1.409,
|
| 8465 |
+
"step": 1200
|
| 8466 |
}
|
| 8467 |
],
|
| 8468 |
"logging_steps": 1,
|
|
|
|
| 8477 |
"early_stopping_threshold": 0.0
|
| 8478 |
},
|
| 8479 |
"attributes": {
|
| 8480 |
+
"early_stopping_patience_counter": 3
|
| 8481 |
}
|
| 8482 |
},
|
| 8483 |
"TrainerControl": {
|
|
|
|
| 8491 |
"attributes": {}
|
| 8492 |
}
|
| 8493 |
},
|
| 8494 |
+
"total_flos": 1.1047304629384643e+18,
|
| 8495 |
"train_batch_size": 4,
|
| 8496 |
"trial_name": null,
|
| 8497 |
"trial_params": null
|