Training in progress, step 98800, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -23,13 +23,13 @@
|
|
| 23 |
"rank_pattern": {},
|
| 24 |
"revision": null,
|
| 25 |
"target_modules": [
|
| 26 |
-
"gate_proj",
|
| 27 |
"q_proj",
|
| 28 |
-
"
|
| 29 |
-
"down_proj",
|
| 30 |
"o_proj",
|
| 31 |
"k_proj",
|
| 32 |
-
"
|
|
|
|
|
|
|
| 33 |
],
|
| 34 |
"task_type": "CAUSAL_LM",
|
| 35 |
"use_dora": false,
|
|
|
|
| 23 |
"rank_pattern": {},
|
| 24 |
"revision": null,
|
| 25 |
"target_modules": [
|
|
|
|
| 26 |
"q_proj",
|
| 27 |
+
"gate_proj",
|
|
|
|
| 28 |
"o_proj",
|
| 29 |
"k_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"v_proj",
|
| 32 |
+
"down_proj"
|
| 33 |
],
|
| 34 |
"task_type": "CAUSAL_LM",
|
| 35 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 360740440
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b35cb4877d8334962276c7da8c84c133d65ddb9b2d7e1db7e303689c89c692ca
|
| 3 |
size 360740440
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 184019218
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fee90d81acc29550e4ed5864ed6b02a95d02876f856a3224c0d9351341bafeca
|
| 3 |
size 184019218
|
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:eb718b6f80f0f2bee74d0e1e6d928accc63edec89ff90ba971e1d41db3b5d523
|
| 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 sha256:5d99b32d0f99374611e7d61179bc5fa1aa84093b62875d2b654c0690814a02de
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2380,6 +2380,1091 @@
|
|
| 2380 |
"learning_rate": 1.998080708864831e-05,
|
| 2381 |
"loss": 1.9318,
|
| 2382 |
"step": 67800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2383 |
}
|
| 2384 |
],
|
| 2385 |
"logging_steps": 200,
|
|
@@ -2399,7 +3484,7 @@
|
|
| 2399 |
"attributes": {}
|
| 2400 |
}
|
| 2401 |
},
|
| 2402 |
-
"total_flos": 1.
|
| 2403 |
"train_batch_size": 1,
|
| 2404 |
"trial_name": null,
|
| 2405 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.06871365670017268,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 98800,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2380 |
"learning_rate": 1.998080708864831e-05,
|
| 2381 |
"loss": 1.9318,
|
| 2382 |
"step": 67800
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.04729280015801359,
|
| 2386 |
+
"grad_norm": 3.2300922870635986,
|
| 2387 |
+
"learning_rate": 1.998069374127872e-05,
|
| 2388 |
+
"loss": 1.8863,
|
| 2389 |
+
"step": 68000
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.04743189662906657,
|
| 2393 |
+
"grad_norm": 5.228055000305176,
|
| 2394 |
+
"learning_rate": 1.9980580060658826e-05,
|
| 2395 |
+
"loss": 1.9176,
|
| 2396 |
+
"step": 68200
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.04757099310011955,
|
| 2400 |
+
"grad_norm": 3.513674736022949,
|
| 2401 |
+
"learning_rate": 1.9980466046794044e-05,
|
| 2402 |
+
"loss": 1.9543,
|
| 2403 |
+
"step": 68400
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.04771008957117254,
|
| 2407 |
+
"grad_norm": 5.115189552307129,
|
| 2408 |
+
"learning_rate": 1.998035169968983e-05,
|
| 2409 |
+
"loss": 1.9288,
|
| 2410 |
+
"step": 68600
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.04784918604222552,
|
| 2414 |
+
"grad_norm": 8.184946060180664,
|
| 2415 |
+
"learning_rate": 1.998023701935163e-05,
|
| 2416 |
+
"loss": 1.9589,
|
| 2417 |
+
"step": 68800
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.047988282513278496,
|
| 2421 |
+
"grad_norm": 4.300464630126953,
|
| 2422 |
+
"learning_rate": 1.998012200578493e-05,
|
| 2423 |
+
"loss": 1.9337,
|
| 2424 |
+
"step": 69000
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.048127378984331476,
|
| 2428 |
+
"grad_norm": 4.387246131896973,
|
| 2429 |
+
"learning_rate": 1.998000665899521e-05,
|
| 2430 |
+
"loss": 1.9173,
|
| 2431 |
+
"step": 69200
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.04826647545538446,
|
| 2435 |
+
"grad_norm": 5.051788330078125,
|
| 2436 |
+
"learning_rate": 1.9979890978987976e-05,
|
| 2437 |
+
"loss": 1.9104,
|
| 2438 |
+
"step": 69400
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.04840557192643744,
|
| 2442 |
+
"grad_norm": 9.470113754272461,
|
| 2443 |
+
"learning_rate": 1.9979774965768762e-05,
|
| 2444 |
+
"loss": 1.9325,
|
| 2445 |
+
"step": 69600
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.04854466839749042,
|
| 2449 |
+
"grad_norm": 4.497230529785156,
|
| 2450 |
+
"learning_rate": 1.99796586193431e-05,
|
| 2451 |
+
"loss": 1.9815,
|
| 2452 |
+
"step": 69800
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.0486837648685434,
|
| 2456 |
+
"grad_norm": 4.782131671905518,
|
| 2457 |
+
"learning_rate": 1.997954193971655e-05,
|
| 2458 |
+
"loss": 1.902,
|
| 2459 |
+
"step": 70000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.04882286133959638,
|
| 2463 |
+
"grad_norm": 5.133215427398682,
|
| 2464 |
+
"learning_rate": 1.997942492689467e-05,
|
| 2465 |
+
"loss": 1.9728,
|
| 2466 |
+
"step": 70200
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.04896195781064937,
|
| 2470 |
+
"grad_norm": 4.0771942138671875,
|
| 2471 |
+
"learning_rate": 1.9979307580883048e-05,
|
| 2472 |
+
"loss": 1.9499,
|
| 2473 |
+
"step": 70400
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.04910105428170235,
|
| 2477 |
+
"grad_norm": 5.347265720367432,
|
| 2478 |
+
"learning_rate": 1.9979189901687298e-05,
|
| 2479 |
+
"loss": 1.943,
|
| 2480 |
+
"step": 70600
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.04924015075275533,
|
| 2484 |
+
"grad_norm": 8.083806037902832,
|
| 2485 |
+
"learning_rate": 1.9979071889313028e-05,
|
| 2486 |
+
"loss": 1.9202,
|
| 2487 |
+
"step": 70800
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.049379247223808306,
|
| 2491 |
+
"grad_norm": 5.8364386558532715,
|
| 2492 |
+
"learning_rate": 1.9978953543765876e-05,
|
| 2493 |
+
"loss": 1.9831,
|
| 2494 |
+
"step": 71000
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.04951834369486129,
|
| 2498 |
+
"grad_norm": 4.9726152420043945,
|
| 2499 |
+
"learning_rate": 1.9978834865051492e-05,
|
| 2500 |
+
"loss": 1.9912,
|
| 2501 |
+
"step": 71200
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.04965744016591427,
|
| 2505 |
+
"grad_norm": 3.8392043113708496,
|
| 2506 |
+
"learning_rate": 1.9978715853175532e-05,
|
| 2507 |
+
"loss": 1.9531,
|
| 2508 |
+
"step": 71400
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.04979653663696725,
|
| 2512 |
+
"grad_norm": 8.026137351989746,
|
| 2513 |
+
"learning_rate": 1.997859650814369e-05,
|
| 2514 |
+
"loss": 1.8849,
|
| 2515 |
+
"step": 71600
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.04993563310802023,
|
| 2519 |
+
"grad_norm": 5.833446979522705,
|
| 2520 |
+
"learning_rate": 1.997847682996165e-05,
|
| 2521 |
+
"loss": 2.0134,
|
| 2522 |
+
"step": 71800
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.05007472957907321,
|
| 2526 |
+
"grad_norm": 3.1947782039642334,
|
| 2527 |
+
"learning_rate": 1.9978356818635146e-05,
|
| 2528 |
+
"loss": 1.9404,
|
| 2529 |
+
"step": 72000
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.0502138260501262,
|
| 2533 |
+
"grad_norm": 3.642251968383789,
|
| 2534 |
+
"learning_rate": 1.997823647416989e-05,
|
| 2535 |
+
"loss": 1.914,
|
| 2536 |
+
"step": 72200
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.05035292252117918,
|
| 2540 |
+
"grad_norm": 4.345871448516846,
|
| 2541 |
+
"learning_rate": 1.997811579657163e-05,
|
| 2542 |
+
"loss": 1.8788,
|
| 2543 |
+
"step": 72400
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.05049201899223216,
|
| 2547 |
+
"grad_norm": 8.562602043151855,
|
| 2548 |
+
"learning_rate": 1.997799478584613e-05,
|
| 2549 |
+
"loss": 1.9885,
|
| 2550 |
+
"step": 72600
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.05063111546328514,
|
| 2554 |
+
"grad_norm": 3.9181063175201416,
|
| 2555 |
+
"learning_rate": 1.997787344199917e-05,
|
| 2556 |
+
"loss": 1.9824,
|
| 2557 |
+
"step": 72800
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.050770211934338116,
|
| 2561 |
+
"grad_norm": 5.295780181884766,
|
| 2562 |
+
"learning_rate": 1.997775176503653e-05,
|
| 2563 |
+
"loss": 1.9066,
|
| 2564 |
+
"step": 73000
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.0509093084053911,
|
| 2568 |
+
"grad_norm": 5.285375118255615,
|
| 2569 |
+
"learning_rate": 1.9977629754964036e-05,
|
| 2570 |
+
"loss": 1.917,
|
| 2571 |
+
"step": 73200
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.05104840487644408,
|
| 2575 |
+
"grad_norm": 5.803408622741699,
|
| 2576 |
+
"learning_rate": 1.99775074117875e-05,
|
| 2577 |
+
"loss": 1.9555,
|
| 2578 |
+
"step": 73400
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.05118750134749706,
|
| 2582 |
+
"grad_norm": 4.541822910308838,
|
| 2583 |
+
"learning_rate": 1.9977384735512765e-05,
|
| 2584 |
+
"loss": 1.889,
|
| 2585 |
+
"step": 73600
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.05132659781855004,
|
| 2589 |
+
"grad_norm": 5.164264678955078,
|
| 2590 |
+
"learning_rate": 1.9977261726145692e-05,
|
| 2591 |
+
"loss": 1.8705,
|
| 2592 |
+
"step": 73800
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.05146569428960303,
|
| 2596 |
+
"grad_norm": 5.912387847900391,
|
| 2597 |
+
"learning_rate": 1.997713838369215e-05,
|
| 2598 |
+
"loss": 1.959,
|
| 2599 |
+
"step": 74000
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.05160479076065601,
|
| 2603 |
+
"grad_norm": 4.968830585479736,
|
| 2604 |
+
"learning_rate": 1.9977014708158027e-05,
|
| 2605 |
+
"loss": 1.9991,
|
| 2606 |
+
"step": 74200
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.05174388723170899,
|
| 2610 |
+
"grad_norm": 4.360265731811523,
|
| 2611 |
+
"learning_rate": 1.997689069954923e-05,
|
| 2612 |
+
"loss": 1.9133,
|
| 2613 |
+
"step": 74400
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.05188298370276197,
|
| 2617 |
+
"grad_norm": 4.5406999588012695,
|
| 2618 |
+
"learning_rate": 1.997676635787167e-05,
|
| 2619 |
+
"loss": 1.9622,
|
| 2620 |
+
"step": 74600
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.05202208017381495,
|
| 2624 |
+
"grad_norm": 6.6085615158081055,
|
| 2625 |
+
"learning_rate": 1.9976641683131293e-05,
|
| 2626 |
+
"loss": 1.9628,
|
| 2627 |
+
"step": 74800
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.05216117664486793,
|
| 2631 |
+
"grad_norm": 3.78121018409729,
|
| 2632 |
+
"learning_rate": 1.9976516675334044e-05,
|
| 2633 |
+
"loss": 1.9271,
|
| 2634 |
+
"step": 75000
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.05230027311592091,
|
| 2638 |
+
"grad_norm": 4.836414813995361,
|
| 2639 |
+
"learning_rate": 1.99763913344859e-05,
|
| 2640 |
+
"loss": 1.9079,
|
| 2641 |
+
"step": 75200
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.05243936958697389,
|
| 2645 |
+
"grad_norm": 4.3670759201049805,
|
| 2646 |
+
"learning_rate": 1.997626566059284e-05,
|
| 2647 |
+
"loss": 1.9355,
|
| 2648 |
+
"step": 75400
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.05257846605802687,
|
| 2652 |
+
"grad_norm": 4.553194999694824,
|
| 2653 |
+
"learning_rate": 1.9976139653660858e-05,
|
| 2654 |
+
"loss": 1.9888,
|
| 2655 |
+
"step": 75600
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.05271756252907986,
|
| 2659 |
+
"grad_norm": 5.623137474060059,
|
| 2660 |
+
"learning_rate": 1.997601331369597e-05,
|
| 2661 |
+
"loss": 1.9318,
|
| 2662 |
+
"step": 75800
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.05285665900013284,
|
| 2666 |
+
"grad_norm": 5.056840896606445,
|
| 2667 |
+
"learning_rate": 1.9975886640704223e-05,
|
| 2668 |
+
"loss": 1.9439,
|
| 2669 |
+
"step": 76000
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.05299575547118582,
|
| 2673 |
+
"grad_norm": 5.085750579833984,
|
| 2674 |
+
"learning_rate": 1.9975759634691644e-05,
|
| 2675 |
+
"loss": 1.9081,
|
| 2676 |
+
"step": 76200
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.0531348519422388,
|
| 2680 |
+
"grad_norm": 6.532809257507324,
|
| 2681 |
+
"learning_rate": 1.9975632295664304e-05,
|
| 2682 |
+
"loss": 1.9654,
|
| 2683 |
+
"step": 76400
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.05327394841329178,
|
| 2687 |
+
"grad_norm": 5.280889987945557,
|
| 2688 |
+
"learning_rate": 1.9975504623628285e-05,
|
| 2689 |
+
"loss": 1.9278,
|
| 2690 |
+
"step": 76600
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.053413044884344764,
|
| 2694 |
+
"grad_norm": 4.8631591796875,
|
| 2695 |
+
"learning_rate": 1.9975376618589682e-05,
|
| 2696 |
+
"loss": 1.9152,
|
| 2697 |
+
"step": 76800
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.053552141355397744,
|
| 2701 |
+
"grad_norm": 4.5169901847839355,
|
| 2702 |
+
"learning_rate": 1.9975248280554598e-05,
|
| 2703 |
+
"loss": 1.9628,
|
| 2704 |
+
"step": 77000
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.05369123782645072,
|
| 2708 |
+
"grad_norm": 7.0524373054504395,
|
| 2709 |
+
"learning_rate": 1.9975119609529164e-05,
|
| 2710 |
+
"loss": 1.9577,
|
| 2711 |
+
"step": 77200
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.0538303342975037,
|
| 2715 |
+
"grad_norm": 3.7885243892669678,
|
| 2716 |
+
"learning_rate": 1.9974990605519528e-05,
|
| 2717 |
+
"loss": 1.953,
|
| 2718 |
+
"step": 77400
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.05396943076855669,
|
| 2722 |
+
"grad_norm": 4.719120979309082,
|
| 2723 |
+
"learning_rate": 1.997486126853184e-05,
|
| 2724 |
+
"loss": 1.908,
|
| 2725 |
+
"step": 77600
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.05410852723960967,
|
| 2729 |
+
"grad_norm": 3.3026559352874756,
|
| 2730 |
+
"learning_rate": 1.9974731598572283e-05,
|
| 2731 |
+
"loss": 1.9467,
|
| 2732 |
+
"step": 77800
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.05424762371066265,
|
| 2736 |
+
"grad_norm": 3.9477977752685547,
|
| 2737 |
+
"learning_rate": 1.997460159564704e-05,
|
| 2738 |
+
"loss": 1.8798,
|
| 2739 |
+
"step": 78000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.05438672018171563,
|
| 2743 |
+
"grad_norm": 5.302116870880127,
|
| 2744 |
+
"learning_rate": 1.997447125976232e-05,
|
| 2745 |
+
"loss": 1.9372,
|
| 2746 |
+
"step": 78200
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.05452581665276861,
|
| 2750 |
+
"grad_norm": 3.8946664333343506,
|
| 2751 |
+
"learning_rate": 1.9974340590924342e-05,
|
| 2752 |
+
"loss": 1.9257,
|
| 2753 |
+
"step": 78400
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.054664913123821594,
|
| 2757 |
+
"grad_norm": 5.974159240722656,
|
| 2758 |
+
"learning_rate": 1.997420958913935e-05,
|
| 2759 |
+
"loss": 1.8942,
|
| 2760 |
+
"step": 78600
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.054804009594874574,
|
| 2764 |
+
"grad_norm": 5.002580642700195,
|
| 2765 |
+
"learning_rate": 1.9974078254413595e-05,
|
| 2766 |
+
"loss": 1.9036,
|
| 2767 |
+
"step": 78800
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.054943106065927554,
|
| 2771 |
+
"grad_norm": 5.091346263885498,
|
| 2772 |
+
"learning_rate": 1.9973946586753347e-05,
|
| 2773 |
+
"loss": 1.9415,
|
| 2774 |
+
"step": 79000
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.05508220253698053,
|
| 2778 |
+
"grad_norm": 3.594421625137329,
|
| 2779 |
+
"learning_rate": 1.997381458616489e-05,
|
| 2780 |
+
"loss": 2.0063,
|
| 2781 |
+
"step": 79200
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.05522129900803352,
|
| 2785 |
+
"grad_norm": 4.16220760345459,
|
| 2786 |
+
"learning_rate": 1.9973682252654528e-05,
|
| 2787 |
+
"loss": 1.9647,
|
| 2788 |
+
"step": 79400
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.0553603954790865,
|
| 2792 |
+
"grad_norm": 4.97799015045166,
|
| 2793 |
+
"learning_rate": 1.9973549586228574e-05,
|
| 2794 |
+
"loss": 1.9203,
|
| 2795 |
+
"step": 79600
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.05549949195013948,
|
| 2799 |
+
"grad_norm": 6.2803521156311035,
|
| 2800 |
+
"learning_rate": 1.9973416586893366e-05,
|
| 2801 |
+
"loss": 1.9479,
|
| 2802 |
+
"step": 79800
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.05563858842119246,
|
| 2806 |
+
"grad_norm": 6.788886547088623,
|
| 2807 |
+
"learning_rate": 1.9973283254655252e-05,
|
| 2808 |
+
"loss": 1.8882,
|
| 2809 |
+
"step": 80000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.05577768489224544,
|
| 2813 |
+
"grad_norm": 4.553073406219482,
|
| 2814 |
+
"learning_rate": 1.99731495895206e-05,
|
| 2815 |
+
"loss": 1.9328,
|
| 2816 |
+
"step": 80200
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.055916781363298425,
|
| 2820 |
+
"grad_norm": 4.717394828796387,
|
| 2821 |
+
"learning_rate": 1.9973015591495787e-05,
|
| 2822 |
+
"loss": 1.88,
|
| 2823 |
+
"step": 80400
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.056055877834351404,
|
| 2827 |
+
"grad_norm": 5.210158824920654,
|
| 2828 |
+
"learning_rate": 1.9972881260587207e-05,
|
| 2829 |
+
"loss": 1.9725,
|
| 2830 |
+
"step": 80600
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.056194974305404384,
|
| 2834 |
+
"grad_norm": 4.219607353210449,
|
| 2835 |
+
"learning_rate": 1.997274659680128e-05,
|
| 2836 |
+
"loss": 1.9263,
|
| 2837 |
+
"step": 80800
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.056334070776457364,
|
| 2841 |
+
"grad_norm": 4.876551151275635,
|
| 2842 |
+
"learning_rate": 1.997261160014443e-05,
|
| 2843 |
+
"loss": 1.9166,
|
| 2844 |
+
"step": 81000
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.05647316724751035,
|
| 2848 |
+
"grad_norm": 4.498875617980957,
|
| 2849 |
+
"learning_rate": 1.9972476270623103e-05,
|
| 2850 |
+
"loss": 1.9293,
|
| 2851 |
+
"step": 81200
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.05661226371856333,
|
| 2855 |
+
"grad_norm": 4.708183288574219,
|
| 2856 |
+
"learning_rate": 1.9972340608243763e-05,
|
| 2857 |
+
"loss": 1.8991,
|
| 2858 |
+
"step": 81400
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.05675136018961631,
|
| 2862 |
+
"grad_norm": 4.428593158721924,
|
| 2863 |
+
"learning_rate": 1.997220461301288e-05,
|
| 2864 |
+
"loss": 1.8605,
|
| 2865 |
+
"step": 81600
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.05689045666066929,
|
| 2869 |
+
"grad_norm": 7.380921363830566,
|
| 2870 |
+
"learning_rate": 1.997206828493695e-05,
|
| 2871 |
+
"loss": 1.9595,
|
| 2872 |
+
"step": 81800
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.05702955313172227,
|
| 2876 |
+
"grad_norm": 4.452649116516113,
|
| 2877 |
+
"learning_rate": 1.9971931624022477e-05,
|
| 2878 |
+
"loss": 1.9898,
|
| 2879 |
+
"step": 82000
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.057168649602775255,
|
| 2883 |
+
"grad_norm": 5.155123233795166,
|
| 2884 |
+
"learning_rate": 1.997179463027599e-05,
|
| 2885 |
+
"loss": 1.9956,
|
| 2886 |
+
"step": 82200
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.057307746073828235,
|
| 2890 |
+
"grad_norm": 4.984164237976074,
|
| 2891 |
+
"learning_rate": 1.997165730370403e-05,
|
| 2892 |
+
"loss": 1.9642,
|
| 2893 |
+
"step": 82400
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.057446842544881214,
|
| 2897 |
+
"grad_norm": 4.363987445831299,
|
| 2898 |
+
"learning_rate": 1.9971519644313147e-05,
|
| 2899 |
+
"loss": 1.9301,
|
| 2900 |
+
"step": 82600
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.057585939015934194,
|
| 2904 |
+
"grad_norm": 4.11600399017334,
|
| 2905 |
+
"learning_rate": 1.9971381652109915e-05,
|
| 2906 |
+
"loss": 1.9154,
|
| 2907 |
+
"step": 82800
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.05772503548698718,
|
| 2911 |
+
"grad_norm": 4.238101005554199,
|
| 2912 |
+
"learning_rate": 1.9971243327100923e-05,
|
| 2913 |
+
"loss": 1.9414,
|
| 2914 |
+
"step": 83000
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.05786413195804016,
|
| 2918 |
+
"grad_norm": 4.86619234085083,
|
| 2919 |
+
"learning_rate": 1.9971104669292777e-05,
|
| 2920 |
+
"loss": 1.9366,
|
| 2921 |
+
"step": 83200
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.05800322842909314,
|
| 2925 |
+
"grad_norm": 4.146606922149658,
|
| 2926 |
+
"learning_rate": 1.997096567869209e-05,
|
| 2927 |
+
"loss": 1.9064,
|
| 2928 |
+
"step": 83400
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.05814232490014612,
|
| 2932 |
+
"grad_norm": 6.376154899597168,
|
| 2933 |
+
"learning_rate": 1.99708263553055e-05,
|
| 2934 |
+
"loss": 1.951,
|
| 2935 |
+
"step": 83600
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.0582814213711991,
|
| 2939 |
+
"grad_norm": 4.6015625,
|
| 2940 |
+
"learning_rate": 1.997068669913966e-05,
|
| 2941 |
+
"loss": 1.9448,
|
| 2942 |
+
"step": 83800
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.058420517842252086,
|
| 2946 |
+
"grad_norm": 6.202853202819824,
|
| 2947 |
+
"learning_rate": 1.9970546710201236e-05,
|
| 2948 |
+
"loss": 1.9763,
|
| 2949 |
+
"step": 84000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.058559614313305065,
|
| 2953 |
+
"grad_norm": 4.391757965087891,
|
| 2954 |
+
"learning_rate": 1.9970406388496907e-05,
|
| 2955 |
+
"loss": 1.8496,
|
| 2956 |
+
"step": 84200
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.058698710784358045,
|
| 2960 |
+
"grad_norm": 7.166893482208252,
|
| 2961 |
+
"learning_rate": 1.9970265734033377e-05,
|
| 2962 |
+
"loss": 1.9595,
|
| 2963 |
+
"step": 84400
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.058837807255411025,
|
| 2967 |
+
"grad_norm": 5.879612922668457,
|
| 2968 |
+
"learning_rate": 1.997012474681736e-05,
|
| 2969 |
+
"loss": 1.9344,
|
| 2970 |
+
"step": 84600
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.05897690372646401,
|
| 2974 |
+
"grad_norm": 4.643247127532959,
|
| 2975 |
+
"learning_rate": 1.9969983426855583e-05,
|
| 2976 |
+
"loss": 1.936,
|
| 2977 |
+
"step": 84800
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.05911600019751699,
|
| 2981 |
+
"grad_norm": 3.9626338481903076,
|
| 2982 |
+
"learning_rate": 1.9969841774154797e-05,
|
| 2983 |
+
"loss": 1.951,
|
| 2984 |
+
"step": 85000
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.05925509666856997,
|
| 2988 |
+
"grad_norm": 6.601691246032715,
|
| 2989 |
+
"learning_rate": 1.9969699788721763e-05,
|
| 2990 |
+
"loss": 1.9287,
|
| 2991 |
+
"step": 85200
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.05939419313962295,
|
| 2995 |
+
"grad_norm": 4.563479423522949,
|
| 2996 |
+
"learning_rate": 1.9969557470563257e-05,
|
| 2997 |
+
"loss": 1.8866,
|
| 2998 |
+
"step": 85400
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.05953328961067593,
|
| 3002 |
+
"grad_norm": 8.620516777038574,
|
| 3003 |
+
"learning_rate": 1.9969414819686076e-05,
|
| 3004 |
+
"loss": 1.914,
|
| 3005 |
+
"step": 85600
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.059672386081728916,
|
| 3009 |
+
"grad_norm": 7.413761615753174,
|
| 3010 |
+
"learning_rate": 1.996927183609703e-05,
|
| 3011 |
+
"loss": 1.9449,
|
| 3012 |
+
"step": 85800
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.059811482552781896,
|
| 3016 |
+
"grad_norm": 6.8260016441345215,
|
| 3017 |
+
"learning_rate": 1.9969128519802942e-05,
|
| 3018 |
+
"loss": 1.9363,
|
| 3019 |
+
"step": 86000
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.059950579023834875,
|
| 3023 |
+
"grad_norm": 4.007237434387207,
|
| 3024 |
+
"learning_rate": 1.9968984870810654e-05,
|
| 3025 |
+
"loss": 1.9422,
|
| 3026 |
+
"step": 86200
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.060089675494887855,
|
| 3030 |
+
"grad_norm": 3.5244007110595703,
|
| 3031 |
+
"learning_rate": 1.9968840889127022e-05,
|
| 3032 |
+
"loss": 1.9647,
|
| 3033 |
+
"step": 86400
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.060228771965940835,
|
| 3037 |
+
"grad_norm": 5.591712474822998,
|
| 3038 |
+
"learning_rate": 1.996869657475893e-05,
|
| 3039 |
+
"loss": 1.9003,
|
| 3040 |
+
"step": 86600
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.06036786843699382,
|
| 3044 |
+
"grad_norm": 6.790228366851807,
|
| 3045 |
+
"learning_rate": 1.9968551927713252e-05,
|
| 3046 |
+
"loss": 1.9424,
|
| 3047 |
+
"step": 86800
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.0605069649080468,
|
| 3051 |
+
"grad_norm": 7.662969589233398,
|
| 3052 |
+
"learning_rate": 1.9968406947996906e-05,
|
| 3053 |
+
"loss": 1.9268,
|
| 3054 |
+
"step": 87000
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.06064606137909978,
|
| 3058 |
+
"grad_norm": 7.937498092651367,
|
| 3059 |
+
"learning_rate": 1.9968261635616807e-05,
|
| 3060 |
+
"loss": 1.9033,
|
| 3061 |
+
"step": 87200
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.06078515785015276,
|
| 3065 |
+
"grad_norm": 3.6962380409240723,
|
| 3066 |
+
"learning_rate": 1.9968115990579892e-05,
|
| 3067 |
+
"loss": 1.9427,
|
| 3068 |
+
"step": 87400
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.06092425432120575,
|
| 3072 |
+
"grad_norm": 8.26564884185791,
|
| 3073 |
+
"learning_rate": 1.996797001289312e-05,
|
| 3074 |
+
"loss": 1.9067,
|
| 3075 |
+
"step": 87600
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.061063350792258726,
|
| 3079 |
+
"grad_norm": 6.891269683837891,
|
| 3080 |
+
"learning_rate": 1.996782370256345e-05,
|
| 3081 |
+
"loss": 1.8875,
|
| 3082 |
+
"step": 87800
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.061202447263311706,
|
| 3086 |
+
"grad_norm": 3.3418822288513184,
|
| 3087 |
+
"learning_rate": 1.996767705959787e-05,
|
| 3088 |
+
"loss": 2.0397,
|
| 3089 |
+
"step": 88000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.061341543734364685,
|
| 3093 |
+
"grad_norm": 4.370149612426758,
|
| 3094 |
+
"learning_rate": 1.9967530084003388e-05,
|
| 3095 |
+
"loss": 1.976,
|
| 3096 |
+
"step": 88200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.061480640205417665,
|
| 3100 |
+
"grad_norm": 6.564484596252441,
|
| 3101 |
+
"learning_rate": 1.9967382775787013e-05,
|
| 3102 |
+
"loss": 1.9072,
|
| 3103 |
+
"step": 88400
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.06161973667647065,
|
| 3107 |
+
"grad_norm": 4.912674903869629,
|
| 3108 |
+
"learning_rate": 1.9967235134955777e-05,
|
| 3109 |
+
"loss": 1.9404,
|
| 3110 |
+
"step": 88600
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.06175883314752363,
|
| 3114 |
+
"grad_norm": 7.586604118347168,
|
| 3115 |
+
"learning_rate": 1.996708716151673e-05,
|
| 3116 |
+
"loss": 1.9895,
|
| 3117 |
+
"step": 88800
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.06189792961857661,
|
| 3121 |
+
"grad_norm": 4.773323059082031,
|
| 3122 |
+
"learning_rate": 1.996693885547694e-05,
|
| 3123 |
+
"loss": 1.9679,
|
| 3124 |
+
"step": 89000
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.06203702608962959,
|
| 3128 |
+
"grad_norm": 3.4076714515686035,
|
| 3129 |
+
"learning_rate": 1.9966790216843476e-05,
|
| 3130 |
+
"loss": 1.9246,
|
| 3131 |
+
"step": 89200
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.06217612256068258,
|
| 3135 |
+
"grad_norm": 4.651917934417725,
|
| 3136 |
+
"learning_rate": 1.996664124562345e-05,
|
| 3137 |
+
"loss": 1.9442,
|
| 3138 |
+
"step": 89400
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.06231521903173556,
|
| 3142 |
+
"grad_norm": 5.476116180419922,
|
| 3143 |
+
"learning_rate": 1.9966491941823962e-05,
|
| 3144 |
+
"loss": 1.9255,
|
| 3145 |
+
"step": 89600
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.062454315502788536,
|
| 3149 |
+
"grad_norm": 5.208308696746826,
|
| 3150 |
+
"learning_rate": 1.996634230545214e-05,
|
| 3151 |
+
"loss": 1.9196,
|
| 3152 |
+
"step": 89800
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.06259341197384152,
|
| 3156 |
+
"grad_norm": 5.869366645812988,
|
| 3157 |
+
"learning_rate": 1.9966192336515128e-05,
|
| 3158 |
+
"loss": 1.9648,
|
| 3159 |
+
"step": 90000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.0627325084448945,
|
| 3163 |
+
"grad_norm": 4.808291912078857,
|
| 3164 |
+
"learning_rate": 1.9966042035020093e-05,
|
| 3165 |
+
"loss": 1.9309,
|
| 3166 |
+
"step": 90200
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.06287160491594748,
|
| 3170 |
+
"grad_norm": 4.214413166046143,
|
| 3171 |
+
"learning_rate": 1.9965891400974203e-05,
|
| 3172 |
+
"loss": 1.9012,
|
| 3173 |
+
"step": 90400
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.06301070138700046,
|
| 3177 |
+
"grad_norm": 5.237168312072754,
|
| 3178 |
+
"learning_rate": 1.996574043438465e-05,
|
| 3179 |
+
"loss": 1.8358,
|
| 3180 |
+
"step": 90600
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.06314979785805344,
|
| 3184 |
+
"grad_norm": 4.680624961853027,
|
| 3185 |
+
"learning_rate": 1.996558913525864e-05,
|
| 3186 |
+
"loss": 1.8669,
|
| 3187 |
+
"step": 90800
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.06328889432910642,
|
| 3191 |
+
"grad_norm": 7.541216850280762,
|
| 3192 |
+
"learning_rate": 1.9965437503603396e-05,
|
| 3193 |
+
"loss": 1.9261,
|
| 3194 |
+
"step": 91000
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.0634279908001594,
|
| 3198 |
+
"grad_norm": 5.40226936340332,
|
| 3199 |
+
"learning_rate": 1.996528553942616e-05,
|
| 3200 |
+
"loss": 1.9315,
|
| 3201 |
+
"step": 91200
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.06356708727121238,
|
| 3205 |
+
"grad_norm": 5.167973518371582,
|
| 3206 |
+
"learning_rate": 1.9965133242734188e-05,
|
| 3207 |
+
"loss": 1.9366,
|
| 3208 |
+
"step": 91400
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.06370618374226537,
|
| 3212 |
+
"grad_norm": 4.3975510597229,
|
| 3213 |
+
"learning_rate": 1.9964980613534744e-05,
|
| 3214 |
+
"loss": 1.93,
|
| 3215 |
+
"step": 91600
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.06384528021331835,
|
| 3219 |
+
"grad_norm": 4.667608737945557,
|
| 3220 |
+
"learning_rate": 1.9964827651835115e-05,
|
| 3221 |
+
"loss": 1.9256,
|
| 3222 |
+
"step": 91800
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.06398437668437133,
|
| 3226 |
+
"grad_norm": 5.848495960235596,
|
| 3227 |
+
"learning_rate": 1.9964674357642614e-05,
|
| 3228 |
+
"loss": 1.9243,
|
| 3229 |
+
"step": 92000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.06412347315542431,
|
| 3233 |
+
"grad_norm": 5.80579948425293,
|
| 3234 |
+
"learning_rate": 1.9964520730964544e-05,
|
| 3235 |
+
"loss": 1.9286,
|
| 3236 |
+
"step": 92200
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.06426256962647729,
|
| 3240 |
+
"grad_norm": 4.185863971710205,
|
| 3241 |
+
"learning_rate": 1.9964366771808244e-05,
|
| 3242 |
+
"loss": 1.8637,
|
| 3243 |
+
"step": 92400
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.06440166609753027,
|
| 3247 |
+
"grad_norm": 7.295627117156982,
|
| 3248 |
+
"learning_rate": 1.9964212480181067e-05,
|
| 3249 |
+
"loss": 1.9646,
|
| 3250 |
+
"step": 92600
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.06454076256858325,
|
| 3254 |
+
"grad_norm": 4.519957542419434,
|
| 3255 |
+
"learning_rate": 1.9964057856090382e-05,
|
| 3256 |
+
"loss": 1.8934,
|
| 3257 |
+
"step": 92800
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.06467985903963623,
|
| 3261 |
+
"grad_norm": 4.814670085906982,
|
| 3262 |
+
"learning_rate": 1.9963902899543565e-05,
|
| 3263 |
+
"loss": 1.8971,
|
| 3264 |
+
"step": 93000
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.06481895551068921,
|
| 3268 |
+
"grad_norm": 4.680553913116455,
|
| 3269 |
+
"learning_rate": 1.996374761054801e-05,
|
| 3270 |
+
"loss": 1.972,
|
| 3271 |
+
"step": 93200
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.06495805198174219,
|
| 3275 |
+
"grad_norm": 5.095190525054932,
|
| 3276 |
+
"learning_rate": 1.996359198911114e-05,
|
| 3277 |
+
"loss": 1.9197,
|
| 3278 |
+
"step": 93400
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.06509714845279518,
|
| 3282 |
+
"grad_norm": 6.054664134979248,
|
| 3283 |
+
"learning_rate": 1.9963436035240377e-05,
|
| 3284 |
+
"loss": 1.9458,
|
| 3285 |
+
"step": 93600
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.06523624492384816,
|
| 3289 |
+
"grad_norm": 3.4787888526916504,
|
| 3290 |
+
"learning_rate": 1.9963279748943166e-05,
|
| 3291 |
+
"loss": 1.9688,
|
| 3292 |
+
"step": 93800
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.06537534139490114,
|
| 3296 |
+
"grad_norm": 4.545236587524414,
|
| 3297 |
+
"learning_rate": 1.996312313022697e-05,
|
| 3298 |
+
"loss": 1.9123,
|
| 3299 |
+
"step": 94000
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.06551443786595412,
|
| 3303 |
+
"grad_norm": 4.842303276062012,
|
| 3304 |
+
"learning_rate": 1.9962966179099264e-05,
|
| 3305 |
+
"loss": 1.9129,
|
| 3306 |
+
"step": 94200
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.0656535343370071,
|
| 3310 |
+
"grad_norm": 5.7866363525390625,
|
| 3311 |
+
"learning_rate": 1.9962808895567545e-05,
|
| 3312 |
+
"loss": 1.9368,
|
| 3313 |
+
"step": 94400
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.06579263080806008,
|
| 3317 |
+
"grad_norm": 5.4849772453308105,
|
| 3318 |
+
"learning_rate": 1.996265127963932e-05,
|
| 3319 |
+
"loss": 1.9761,
|
| 3320 |
+
"step": 94600
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.06593172727911306,
|
| 3324 |
+
"grad_norm": 5.275039196014404,
|
| 3325 |
+
"learning_rate": 1.996249333132211e-05,
|
| 3326 |
+
"loss": 1.9699,
|
| 3327 |
+
"step": 94800
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.06607082375016604,
|
| 3331 |
+
"grad_norm": 4.070720672607422,
|
| 3332 |
+
"learning_rate": 1.996233505062346e-05,
|
| 3333 |
+
"loss": 1.933,
|
| 3334 |
+
"step": 95000
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.06620992022121902,
|
| 3338 |
+
"grad_norm": 6.514541149139404,
|
| 3339 |
+
"learning_rate": 1.9962176437550923e-05,
|
| 3340 |
+
"loss": 1.9931,
|
| 3341 |
+
"step": 95200
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.06634901669227201,
|
| 3345 |
+
"grad_norm": 4.82623291015625,
|
| 3346 |
+
"learning_rate": 1.9962017492112068e-05,
|
| 3347 |
+
"loss": 1.9025,
|
| 3348 |
+
"step": 95400
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.066488113163325,
|
| 3352 |
+
"grad_norm": 5.900183200836182,
|
| 3353 |
+
"learning_rate": 1.996185821431449e-05,
|
| 3354 |
+
"loss": 1.843,
|
| 3355 |
+
"step": 95600
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.06662720963437797,
|
| 3359 |
+
"grad_norm": 6.181861877441406,
|
| 3360 |
+
"learning_rate": 1.9961698604165786e-05,
|
| 3361 |
+
"loss": 1.9491,
|
| 3362 |
+
"step": 95800
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.06676630610543095,
|
| 3366 |
+
"grad_norm": 3.4571354389190674,
|
| 3367 |
+
"learning_rate": 1.996153866167358e-05,
|
| 3368 |
+
"loss": 1.9481,
|
| 3369 |
+
"step": 96000
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.06690540257648393,
|
| 3373 |
+
"grad_norm": 5.187012195587158,
|
| 3374 |
+
"learning_rate": 1.996137838684551e-05,
|
| 3375 |
+
"loss": 1.9242,
|
| 3376 |
+
"step": 96200
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.06704449904753691,
|
| 3380 |
+
"grad_norm": 3.7097227573394775,
|
| 3381 |
+
"learning_rate": 1.9961217779689218e-05,
|
| 3382 |
+
"loss": 1.9269,
|
| 3383 |
+
"step": 96400
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.06718359551858989,
|
| 3387 |
+
"grad_norm": 4.761998176574707,
|
| 3388 |
+
"learning_rate": 1.9961056840212385e-05,
|
| 3389 |
+
"loss": 1.8955,
|
| 3390 |
+
"step": 96600
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.06732269198964287,
|
| 3394 |
+
"grad_norm": 4.146270751953125,
|
| 3395 |
+
"learning_rate": 1.9960895568422676e-05,
|
| 3396 |
+
"loss": 1.937,
|
| 3397 |
+
"step": 96800
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.06746178846069585,
|
| 3401 |
+
"grad_norm": 5.371631145477295,
|
| 3402 |
+
"learning_rate": 1.9960733964327807e-05,
|
| 3403 |
+
"loss": 1.9091,
|
| 3404 |
+
"step": 97000
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.06760088493174884,
|
| 3408 |
+
"grad_norm": 4.294393539428711,
|
| 3409 |
+
"learning_rate": 1.9960572027935482e-05,
|
| 3410 |
+
"loss": 1.9376,
|
| 3411 |
+
"step": 97200
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.06773998140280182,
|
| 3415 |
+
"grad_norm": 5.409416675567627,
|
| 3416 |
+
"learning_rate": 1.9960409759253436e-05,
|
| 3417 |
+
"loss": 1.982,
|
| 3418 |
+
"step": 97400
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.0678790778738548,
|
| 3422 |
+
"grad_norm": 4.644226551055908,
|
| 3423 |
+
"learning_rate": 1.996024715828942e-05,
|
| 3424 |
+
"loss": 1.9178,
|
| 3425 |
+
"step": 97600
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.06801817434490778,
|
| 3429 |
+
"grad_norm": 4.132181644439697,
|
| 3430 |
+
"learning_rate": 1.9960084225051184e-05,
|
| 3431 |
+
"loss": 1.9435,
|
| 3432 |
+
"step": 97800
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.06815727081596076,
|
| 3436 |
+
"grad_norm": 4.070497512817383,
|
| 3437 |
+
"learning_rate": 1.9959920959546515e-05,
|
| 3438 |
+
"loss": 1.8981,
|
| 3439 |
+
"step": 98000
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.06829636728701374,
|
| 3443 |
+
"grad_norm": 5.847506046295166,
|
| 3444 |
+
"learning_rate": 1.9959757361783212e-05,
|
| 3445 |
+
"loss": 1.9199,
|
| 3446 |
+
"step": 98200
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.06843546375806672,
|
| 3450 |
+
"grad_norm": 4.320737838745117,
|
| 3451 |
+
"learning_rate": 1.995959343176907e-05,
|
| 3452 |
+
"loss": 1.8637,
|
| 3453 |
+
"step": 98400
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.0685745602291197,
|
| 3457 |
+
"grad_norm": 5.623171329498291,
|
| 3458 |
+
"learning_rate": 1.9959429169511926e-05,
|
| 3459 |
+
"loss": 1.9098,
|
| 3460 |
+
"step": 98600
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.06871365670017268,
|
| 3464 |
+
"grad_norm": 9.206012725830078,
|
| 3465 |
+
"learning_rate": 1.995926457501962e-05,
|
| 3466 |
+
"loss": 1.8603,
|
| 3467 |
+
"step": 98800
|
| 3468 |
}
|
| 3469 |
],
|
| 3470 |
"logging_steps": 200,
|
|
|
|
| 3484 |
"attributes": {}
|
| 3485 |
}
|
| 3486 |
},
|
| 3487 |
+
"total_flos": 1.790026228862116e+17,
|
| 3488 |
"train_batch_size": 1,
|
| 3489 |
"trial_name": null,
|
| 3490 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6840
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2f4bba7fe910555b4fde63970dc08444d084a12c2c45ba6ad0fb3bbbf5964e7
|
| 3 |
size 6840
|