Farouk commited on
Commit ·
83b2ca6
1
Parent(s): 68b10eb
Training in progress, step 4800
Browse files- adapter_model.bin +1 -1
- {checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/README.md +0 -0
- {checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/adapter_config.json +0 -0
- {checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/adapter_model.bin +1 -1
- {checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/README.md +0 -0
- {checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/adapter_config.json +0 -0
- {checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/adapter_model.bin +1 -1
- {checkpoint-2600 → checkpoint-4800}/added_tokens.json +0 -0
- {checkpoint-2600 → checkpoint-4800}/optimizer.pt +1 -1
- {checkpoint-2600 → checkpoint-4800}/rng_state.pth +1 -1
- {checkpoint-2600 → checkpoint-4800}/scheduler.pt +1 -1
- {checkpoint-2600 → checkpoint-4800}/special_tokens_map.json +0 -0
- {checkpoint-2600 → checkpoint-4800}/tokenizer.model +0 -0
- {checkpoint-2600 → checkpoint-4800}/tokenizer_config.json +0 -0
- {checkpoint-2600 → checkpoint-4800}/trainer_state.json +2106 -5
- {checkpoint-2600 → checkpoint-4800}/training_args.bin +0 -0
adapter_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 319977229
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77dc377db752f8e99fe6601778e42b77c4efdf7a69ca5c0296385ccaa4713190
|
| 3 |
size 319977229
|
{checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/README.md
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/adapter_config.json
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4600/adapter_model/adapter_model}/adapter_model.bin
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 319977229
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b7285fa8166dbd4864af681ecb7dfd2af8dce0ba3aae7e8b46aa94a665b608c
|
| 3 |
size 319977229
|
{checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/README.md
RENAMED
|
File without changes
|
{checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/adapter_config.json
RENAMED
|
File without changes
|
{checkpoint-2600/adapter_model/adapter_model → checkpoint-4800}/adapter_model.bin
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 319977229
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77dc377db752f8e99fe6601778e42b77c4efdf7a69ca5c0296385ccaa4713190
|
| 3 |
size 319977229
|
{checkpoint-2600 → checkpoint-4800}/added_tokens.json
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4800}/optimizer.pt
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1279539973
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:91bac102fdb18fc9cbab812c82cf6641db3475b071df53d0681abee24a6a3ece
|
| 3 |
size 1279539973
|
{checkpoint-2600 → checkpoint-4800}/rng_state.pth
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14511
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d17141281dab0938b073a5c53614bc80a6ad702874bf93df4428b445af59b024
|
| 3 |
size 14511
|
{checkpoint-2600 → checkpoint-4800}/scheduler.pt
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:902c3311a8dcadbf8683eec29fe20c841cb172f9dea66dbc64b1baab59e53dc1
|
| 3 |
size 627
|
{checkpoint-2600 → checkpoint-4800}/special_tokens_map.json
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4800}/tokenizer.model
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4800}/tokenizer_config.json
RENAMED
|
File without changes
|
{checkpoint-2600 → checkpoint-4800}/trainer_state.json
RENAMED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "experts/expert-21/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -2489,11 +2489,2112 @@
|
|
| 2489 |
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 2490 |
"mmlu_loss": 1.4832373786688473,
|
| 2491 |
"step": 2600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2492 |
}
|
| 2493 |
],
|
| 2494 |
"max_steps": 10000,
|
| 2495 |
"num_train_epochs": 2,
|
| 2496 |
-
"total_flos":
|
| 2497 |
"trial_name": null,
|
| 2498 |
"trial_params": null
|
| 2499 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7640769481658936,
|
| 3 |
+
"best_model_checkpoint": "experts/expert-21/checkpoint-4800",
|
| 4 |
+
"epoch": 0.9580838323353293,
|
| 5 |
+
"global_step": 4800,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 2489 |
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 2490 |
"mmlu_loss": 1.4832373786688473,
|
| 2491 |
"step": 2600
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.52,
|
| 2495 |
+
"learning_rate": 0.0002,
|
| 2496 |
+
"loss": 0.7971,
|
| 2497 |
+
"step": 2610
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 0.52,
|
| 2501 |
+
"learning_rate": 0.0002,
|
| 2502 |
+
"loss": 0.7771,
|
| 2503 |
+
"step": 2620
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.52,
|
| 2507 |
+
"learning_rate": 0.0002,
|
| 2508 |
+
"loss": 0.8641,
|
| 2509 |
+
"step": 2630
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.53,
|
| 2513 |
+
"learning_rate": 0.0002,
|
| 2514 |
+
"loss": 0.8352,
|
| 2515 |
+
"step": 2640
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.53,
|
| 2519 |
+
"learning_rate": 0.0002,
|
| 2520 |
+
"loss": 0.7577,
|
| 2521 |
+
"step": 2650
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 0.53,
|
| 2525 |
+
"learning_rate": 0.0002,
|
| 2526 |
+
"loss": 0.7748,
|
| 2527 |
+
"step": 2660
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 0.53,
|
| 2531 |
+
"learning_rate": 0.0002,
|
| 2532 |
+
"loss": 0.8012,
|
| 2533 |
+
"step": 2670
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 0.53,
|
| 2537 |
+
"learning_rate": 0.0002,
|
| 2538 |
+
"loss": 0.8073,
|
| 2539 |
+
"step": 2680
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 0.54,
|
| 2543 |
+
"learning_rate": 0.0002,
|
| 2544 |
+
"loss": 0.7752,
|
| 2545 |
+
"step": 2690
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.54,
|
| 2549 |
+
"learning_rate": 0.0002,
|
| 2550 |
+
"loss": 0.8225,
|
| 2551 |
+
"step": 2700
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.54,
|
| 2555 |
+
"learning_rate": 0.0002,
|
| 2556 |
+
"loss": 0.7684,
|
| 2557 |
+
"step": 2710
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.54,
|
| 2561 |
+
"learning_rate": 0.0002,
|
| 2562 |
+
"loss": 0.7912,
|
| 2563 |
+
"step": 2720
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 0.54,
|
| 2567 |
+
"learning_rate": 0.0002,
|
| 2568 |
+
"loss": 0.7707,
|
| 2569 |
+
"step": 2730
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 0.55,
|
| 2573 |
+
"learning_rate": 0.0002,
|
| 2574 |
+
"loss": 0.8007,
|
| 2575 |
+
"step": 2740
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 0.55,
|
| 2579 |
+
"learning_rate": 0.0002,
|
| 2580 |
+
"loss": 0.8413,
|
| 2581 |
+
"step": 2750
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 0.55,
|
| 2585 |
+
"learning_rate": 0.0002,
|
| 2586 |
+
"loss": 0.8439,
|
| 2587 |
+
"step": 2760
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.55,
|
| 2591 |
+
"learning_rate": 0.0002,
|
| 2592 |
+
"loss": 0.7243,
|
| 2593 |
+
"step": 2770
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.55,
|
| 2597 |
+
"learning_rate": 0.0002,
|
| 2598 |
+
"loss": 0.8112,
|
| 2599 |
+
"step": 2780
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.56,
|
| 2603 |
+
"learning_rate": 0.0002,
|
| 2604 |
+
"loss": 0.874,
|
| 2605 |
+
"step": 2790
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 0.56,
|
| 2609 |
+
"learning_rate": 0.0002,
|
| 2610 |
+
"loss": 0.8683,
|
| 2611 |
+
"step": 2800
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 0.56,
|
| 2615 |
+
"eval_loss": 0.7752290964126587,
|
| 2616 |
+
"eval_runtime": 187.3026,
|
| 2617 |
+
"eval_samples_per_second": 5.339,
|
| 2618 |
+
"eval_steps_per_second": 2.669,
|
| 2619 |
+
"step": 2800
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 0.56,
|
| 2623 |
+
"mmlu_eval_accuracy": 0.4937077937991497,
|
| 2624 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 2625 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 2626 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 2627 |
+
"mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
|
| 2628 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
|
| 2629 |
+
"mmlu_eval_accuracy_college_biology": 0.375,
|
| 2630 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 2631 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 2632 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 2633 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 2634 |
+
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
|
| 2635 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 2636 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 2637 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 2638 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 2639 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
|
| 2640 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 2641 |
+
"mmlu_eval_accuracy_global_facts": 0.3,
|
| 2642 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 2643 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
|
| 2644 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 2645 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 2646 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 2647 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 2648 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 2649 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 2650 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
|
| 2651 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 2652 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 2653 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 2654 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 2655 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 2656 |
+
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
|
| 2657 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 2658 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 2659 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 2660 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 2661 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 2662 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 2663 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 2664 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 2665 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
|
| 2666 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 2667 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.23,
|
| 2668 |
+
"mmlu_eval_accuracy_nutrition": 0.7272727272727273,
|
| 2669 |
+
"mmlu_eval_accuracy_philosophy": 0.5588235294117647,
|
| 2670 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 2671 |
+
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
|
| 2672 |
+
"mmlu_eval_accuracy_professional_law": 0.32941176470588235,
|
| 2673 |
+
"mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
|
| 2674 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
|
| 2675 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 2676 |
+
"mmlu_eval_accuracy_security_studies": 0.4444444444444444,
|
| 2677 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 2678 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 2679 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 2680 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 2681 |
+
"mmlu_loss": 1.4585867143642186,
|
| 2682 |
+
"step": 2800
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 0.56,
|
| 2686 |
+
"learning_rate": 0.0002,
|
| 2687 |
+
"loss": 0.8228,
|
| 2688 |
+
"step": 2810
|
| 2689 |
+
},
|
| 2690 |
+
{
|
| 2691 |
+
"epoch": 0.56,
|
| 2692 |
+
"learning_rate": 0.0002,
|
| 2693 |
+
"loss": 0.8865,
|
| 2694 |
+
"step": 2820
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 0.56,
|
| 2698 |
+
"learning_rate": 0.0002,
|
| 2699 |
+
"loss": 0.7906,
|
| 2700 |
+
"step": 2830
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 0.57,
|
| 2704 |
+
"learning_rate": 0.0002,
|
| 2705 |
+
"loss": 0.7763,
|
| 2706 |
+
"step": 2840
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.57,
|
| 2710 |
+
"learning_rate": 0.0002,
|
| 2711 |
+
"loss": 0.7077,
|
| 2712 |
+
"step": 2850
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 0.57,
|
| 2716 |
+
"learning_rate": 0.0002,
|
| 2717 |
+
"loss": 0.8119,
|
| 2718 |
+
"step": 2860
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.57,
|
| 2722 |
+
"learning_rate": 0.0002,
|
| 2723 |
+
"loss": 0.8127,
|
| 2724 |
+
"step": 2870
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 0.57,
|
| 2728 |
+
"learning_rate": 0.0002,
|
| 2729 |
+
"loss": 0.743,
|
| 2730 |
+
"step": 2880
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 0.58,
|
| 2734 |
+
"learning_rate": 0.0002,
|
| 2735 |
+
"loss": 0.8813,
|
| 2736 |
+
"step": 2890
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 0.58,
|
| 2740 |
+
"learning_rate": 0.0002,
|
| 2741 |
+
"loss": 0.7725,
|
| 2742 |
+
"step": 2900
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 0.58,
|
| 2746 |
+
"learning_rate": 0.0002,
|
| 2747 |
+
"loss": 0.715,
|
| 2748 |
+
"step": 2910
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.58,
|
| 2752 |
+
"learning_rate": 0.0002,
|
| 2753 |
+
"loss": 0.8224,
|
| 2754 |
+
"step": 2920
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 0.58,
|
| 2758 |
+
"learning_rate": 0.0002,
|
| 2759 |
+
"loss": 0.7893,
|
| 2760 |
+
"step": 2930
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.59,
|
| 2764 |
+
"learning_rate": 0.0002,
|
| 2765 |
+
"loss": 0.808,
|
| 2766 |
+
"step": 2940
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 0.59,
|
| 2770 |
+
"learning_rate": 0.0002,
|
| 2771 |
+
"loss": 0.816,
|
| 2772 |
+
"step": 2950
|
| 2773 |
+
},
|
| 2774 |
+
{
|
| 2775 |
+
"epoch": 0.59,
|
| 2776 |
+
"learning_rate": 0.0002,
|
| 2777 |
+
"loss": 0.7345,
|
| 2778 |
+
"step": 2960
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 0.59,
|
| 2782 |
+
"learning_rate": 0.0002,
|
| 2783 |
+
"loss": 0.8386,
|
| 2784 |
+
"step": 2970
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 0.59,
|
| 2788 |
+
"learning_rate": 0.0002,
|
| 2789 |
+
"loss": 0.818,
|
| 2790 |
+
"step": 2980
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.6,
|
| 2794 |
+
"learning_rate": 0.0002,
|
| 2795 |
+
"loss": 0.8089,
|
| 2796 |
+
"step": 2990
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 0.6,
|
| 2800 |
+
"learning_rate": 0.0002,
|
| 2801 |
+
"loss": 0.9051,
|
| 2802 |
+
"step": 3000
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.6,
|
| 2806 |
+
"eval_loss": 0.7735825181007385,
|
| 2807 |
+
"eval_runtime": 187.1525,
|
| 2808 |
+
"eval_samples_per_second": 5.343,
|
| 2809 |
+
"eval_steps_per_second": 2.672,
|
| 2810 |
+
"step": 3000
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 0.6,
|
| 2814 |
+
"mmlu_eval_accuracy": 0.49383161171900514,
|
| 2815 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
|
| 2816 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 2817 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 2818 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 2819 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
|
| 2820 |
+
"mmlu_eval_accuracy_college_biology": 0.25,
|
| 2821 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 2822 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 2823 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 2824 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 2825 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 2826 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 2827 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
|
| 2828 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 2829 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 2830 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
|
| 2831 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 2832 |
+
"mmlu_eval_accuracy_global_facts": 0.3,
|
| 2833 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 2834 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
|
| 2835 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 2836 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 2837 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 2838 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 2839 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 2840 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 2841 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
|
| 2842 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 2843 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 2844 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
|
| 2845 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 2846 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 2847 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 2848 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 2849 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 2850 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 2851 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 2852 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 2853 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 2854 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 2855 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 2856 |
+
"mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
|
| 2857 |
+
"mmlu_eval_accuracy_moral_disputes": 0.47368421052631576,
|
| 2858 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 2859 |
+
"mmlu_eval_accuracy_nutrition": 0.5757575757575758,
|
| 2860 |
+
"mmlu_eval_accuracy_philosophy": 0.5882352941176471,
|
| 2861 |
+
"mmlu_eval_accuracy_prehistory": 0.5428571428571428,
|
| 2862 |
+
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
|
| 2863 |
+
"mmlu_eval_accuracy_professional_law": 0.3176470588235294,
|
| 2864 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
|
| 2865 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 2866 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 2867 |
+
"mmlu_eval_accuracy_security_studies": 0.4444444444444444,
|
| 2868 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 2869 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 2870 |
+
"mmlu_eval_accuracy_virology": 0.5555555555555556,
|
| 2871 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 2872 |
+
"mmlu_loss": 1.3979128985118616,
|
| 2873 |
+
"step": 3000
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.6,
|
| 2877 |
+
"learning_rate": 0.0002,
|
| 2878 |
+
"loss": 0.7805,
|
| 2879 |
+
"step": 3010
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.6,
|
| 2883 |
+
"learning_rate": 0.0002,
|
| 2884 |
+
"loss": 0.8022,
|
| 2885 |
+
"step": 3020
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 0.6,
|
| 2889 |
+
"learning_rate": 0.0002,
|
| 2890 |
+
"loss": 0.8525,
|
| 2891 |
+
"step": 3030
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 0.61,
|
| 2895 |
+
"learning_rate": 0.0002,
|
| 2896 |
+
"loss": 0.8669,
|
| 2897 |
+
"step": 3040
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 0.61,
|
| 2901 |
+
"learning_rate": 0.0002,
|
| 2902 |
+
"loss": 0.7706,
|
| 2903 |
+
"step": 3050
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.61,
|
| 2907 |
+
"learning_rate": 0.0002,
|
| 2908 |
+
"loss": 0.7221,
|
| 2909 |
+
"step": 3060
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 0.61,
|
| 2913 |
+
"learning_rate": 0.0002,
|
| 2914 |
+
"loss": 0.8801,
|
| 2915 |
+
"step": 3070
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 0.61,
|
| 2919 |
+
"learning_rate": 0.0002,
|
| 2920 |
+
"loss": 0.6372,
|
| 2921 |
+
"step": 3080
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.62,
|
| 2925 |
+
"learning_rate": 0.0002,
|
| 2926 |
+
"loss": 0.881,
|
| 2927 |
+
"step": 3090
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 0.62,
|
| 2931 |
+
"learning_rate": 0.0002,
|
| 2932 |
+
"loss": 0.8141,
|
| 2933 |
+
"step": 3100
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 0.62,
|
| 2937 |
+
"learning_rate": 0.0002,
|
| 2938 |
+
"loss": 0.8995,
|
| 2939 |
+
"step": 3110
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 0.62,
|
| 2943 |
+
"learning_rate": 0.0002,
|
| 2944 |
+
"loss": 0.8088,
|
| 2945 |
+
"step": 3120
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.62,
|
| 2949 |
+
"learning_rate": 0.0002,
|
| 2950 |
+
"loss": 0.8005,
|
| 2951 |
+
"step": 3130
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 0.63,
|
| 2955 |
+
"learning_rate": 0.0002,
|
| 2956 |
+
"loss": 0.6721,
|
| 2957 |
+
"step": 3140
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 0.63,
|
| 2961 |
+
"learning_rate": 0.0002,
|
| 2962 |
+
"loss": 0.8626,
|
| 2963 |
+
"step": 3150
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.63,
|
| 2967 |
+
"learning_rate": 0.0002,
|
| 2968 |
+
"loss": 0.7597,
|
| 2969 |
+
"step": 3160
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 0.63,
|
| 2973 |
+
"learning_rate": 0.0002,
|
| 2974 |
+
"loss": 0.7343,
|
| 2975 |
+
"step": 3170
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 0.63,
|
| 2979 |
+
"learning_rate": 0.0002,
|
| 2980 |
+
"loss": 0.7306,
|
| 2981 |
+
"step": 3180
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 0.64,
|
| 2985 |
+
"learning_rate": 0.0002,
|
| 2986 |
+
"loss": 0.803,
|
| 2987 |
+
"step": 3190
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.64,
|
| 2991 |
+
"learning_rate": 0.0002,
|
| 2992 |
+
"loss": 0.7498,
|
| 2993 |
+
"step": 3200
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 0.64,
|
| 2997 |
+
"eval_loss": 0.7715298533439636,
|
| 2998 |
+
"eval_runtime": 187.1506,
|
| 2999 |
+
"eval_samples_per_second": 5.343,
|
| 3000 |
+
"eval_steps_per_second": 2.672,
|
| 3001 |
+
"step": 3200
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.64,
|
| 3005 |
+
"mmlu_eval_accuracy": 0.49445891887306603,
|
| 3006 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.09090909090909091,
|
| 3007 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 3008 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 3009 |
+
"mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
|
| 3010 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
|
| 3011 |
+
"mmlu_eval_accuracy_college_biology": 0.3125,
|
| 3012 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 3013 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 3014 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 3015 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 3016 |
+
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
|
| 3017 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 3018 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 3019 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3020 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.3125,
|
| 3021 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
|
| 3022 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 3023 |
+
"mmlu_eval_accuracy_global_facts": 0.4,
|
| 3024 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 3025 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
|
| 3026 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3027 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
|
| 3028 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 3029 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 3030 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907,
|
| 3031 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 3032 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 3033 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 3034 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
|
| 3035 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
|
| 3036 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 3037 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3038 |
+
"mmlu_eval_accuracy_human_aging": 0.6086956521739131,
|
| 3039 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3040 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3041 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 3042 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3043 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 3044 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 3045 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 3046 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 3047 |
+
"mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
|
| 3048 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 3049 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 3050 |
+
"mmlu_eval_accuracy_nutrition": 0.6666666666666666,
|
| 3051 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 3052 |
+
"mmlu_eval_accuracy_prehistory": 0.5428571428571428,
|
| 3053 |
+
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
|
| 3054 |
+
"mmlu_eval_accuracy_professional_law": 0.3352941176470588,
|
| 3055 |
+
"mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
|
| 3056 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 3057 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 3058 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 3059 |
+
"mmlu_eval_accuracy_sociology": 0.6363636363636364,
|
| 3060 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 3061 |
+
"mmlu_eval_accuracy_virology": 0.5555555555555556,
|
| 3062 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 3063 |
+
"mmlu_loss": 1.3431594996243794,
|
| 3064 |
+
"step": 3200
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.64,
|
| 3068 |
+
"learning_rate": 0.0002,
|
| 3069 |
+
"loss": 0.7824,
|
| 3070 |
+
"step": 3210
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 0.64,
|
| 3074 |
+
"learning_rate": 0.0002,
|
| 3075 |
+
"loss": 0.7516,
|
| 3076 |
+
"step": 3220
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 0.64,
|
| 3080 |
+
"learning_rate": 0.0002,
|
| 3081 |
+
"loss": 0.8491,
|
| 3082 |
+
"step": 3230
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.65,
|
| 3086 |
+
"learning_rate": 0.0002,
|
| 3087 |
+
"loss": 0.7538,
|
| 3088 |
+
"step": 3240
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"epoch": 0.65,
|
| 3092 |
+
"learning_rate": 0.0002,
|
| 3093 |
+
"loss": 0.7618,
|
| 3094 |
+
"step": 3250
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 0.65,
|
| 3098 |
+
"learning_rate": 0.0002,
|
| 3099 |
+
"loss": 0.8253,
|
| 3100 |
+
"step": 3260
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 0.65,
|
| 3104 |
+
"learning_rate": 0.0002,
|
| 3105 |
+
"loss": 0.7631,
|
| 3106 |
+
"step": 3270
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.65,
|
| 3110 |
+
"learning_rate": 0.0002,
|
| 3111 |
+
"loss": 0.8339,
|
| 3112 |
+
"step": 3280
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 0.66,
|
| 3116 |
+
"learning_rate": 0.0002,
|
| 3117 |
+
"loss": 0.6797,
|
| 3118 |
+
"step": 3290
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 0.66,
|
| 3122 |
+
"learning_rate": 0.0002,
|
| 3123 |
+
"loss": 0.797,
|
| 3124 |
+
"step": 3300
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.66,
|
| 3128 |
+
"learning_rate": 0.0002,
|
| 3129 |
+
"loss": 0.8091,
|
| 3130 |
+
"step": 3310
|
| 3131 |
+
},
|
| 3132 |
+
{
|
| 3133 |
+
"epoch": 0.66,
|
| 3134 |
+
"learning_rate": 0.0002,
|
| 3135 |
+
"loss": 0.8087,
|
| 3136 |
+
"step": 3320
|
| 3137 |
+
},
|
| 3138 |
+
{
|
| 3139 |
+
"epoch": 0.66,
|
| 3140 |
+
"learning_rate": 0.0002,
|
| 3141 |
+
"loss": 0.8036,
|
| 3142 |
+
"step": 3330
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 0.67,
|
| 3146 |
+
"learning_rate": 0.0002,
|
| 3147 |
+
"loss": 0.9307,
|
| 3148 |
+
"step": 3340
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.67,
|
| 3152 |
+
"learning_rate": 0.0002,
|
| 3153 |
+
"loss": 0.7692,
|
| 3154 |
+
"step": 3350
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 0.67,
|
| 3158 |
+
"learning_rate": 0.0002,
|
| 3159 |
+
"loss": 0.8246,
|
| 3160 |
+
"step": 3360
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 0.67,
|
| 3164 |
+
"learning_rate": 0.0002,
|
| 3165 |
+
"loss": 0.7682,
|
| 3166 |
+
"step": 3370
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.67,
|
| 3170 |
+
"learning_rate": 0.0002,
|
| 3171 |
+
"loss": 0.7337,
|
| 3172 |
+
"step": 3380
|
| 3173 |
+
},
|
| 3174 |
+
{
|
| 3175 |
+
"epoch": 0.68,
|
| 3176 |
+
"learning_rate": 0.0002,
|
| 3177 |
+
"loss": 0.7281,
|
| 3178 |
+
"step": 3390
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 0.68,
|
| 3182 |
+
"learning_rate": 0.0002,
|
| 3183 |
+
"loss": 0.8049,
|
| 3184 |
+
"step": 3400
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 0.68,
|
| 3188 |
+
"eval_loss": 0.7722234725952148,
|
| 3189 |
+
"eval_runtime": 187.0936,
|
| 3190 |
+
"eval_samples_per_second": 5.345,
|
| 3191 |
+
"eval_steps_per_second": 2.672,
|
| 3192 |
+
"step": 3400
|
| 3193 |
+
},
|
| 3194 |
+
{
|
| 3195 |
+
"epoch": 0.68,
|
| 3196 |
+
"mmlu_eval_accuracy": 0.4885770768529298,
|
| 3197 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
|
| 3198 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 3199 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 3200 |
+
"mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
|
| 3201 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
|
| 3202 |
+
"mmlu_eval_accuracy_college_biology": 0.3125,
|
| 3203 |
+
"mmlu_eval_accuracy_college_chemistry": 0.125,
|
| 3204 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 3205 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 3206 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 3207 |
+
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
|
| 3208 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 3209 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 3210 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3211 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.3125,
|
| 3212 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
|
| 3213 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 3214 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 3215 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 3216 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
|
| 3217 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3218 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 3219 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 3220 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 3221 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4418604651162791,
|
| 3222 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
|
| 3223 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 3224 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 3225 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 3226 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 3227 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 3228 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3229 |
+
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
|
| 3230 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3231 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3232 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 3233 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3234 |
+
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
|
| 3235 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 3236 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 3237 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 3238 |
+
"mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
|
| 3239 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 3240 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.27,
|
| 3241 |
+
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
|
| 3242 |
+
"mmlu_eval_accuracy_philosophy": 0.5588235294117647,
|
| 3243 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 3244 |
+
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
|
| 3245 |
+
"mmlu_eval_accuracy_professional_law": 0.36470588235294116,
|
| 3246 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 3247 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 3248 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 3249 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 3250 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 3251 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 3252 |
+
"mmlu_eval_accuracy_virology": 0.4444444444444444,
|
| 3253 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 3254 |
+
"mmlu_loss": 1.3427304202987382,
|
| 3255 |
+
"step": 3400
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 0.68,
|
| 3259 |
+
"learning_rate": 0.0002,
|
| 3260 |
+
"loss": 0.7956,
|
| 3261 |
+
"step": 3410
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.68,
|
| 3265 |
+
"learning_rate": 0.0002,
|
| 3266 |
+
"loss": 0.6748,
|
| 3267 |
+
"step": 3420
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.68,
|
| 3271 |
+
"learning_rate": 0.0002,
|
| 3272 |
+
"loss": 0.7901,
|
| 3273 |
+
"step": 3430
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 0.69,
|
| 3277 |
+
"learning_rate": 0.0002,
|
| 3278 |
+
"loss": 0.7958,
|
| 3279 |
+
"step": 3440
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 0.69,
|
| 3283 |
+
"learning_rate": 0.0002,
|
| 3284 |
+
"loss": 0.8239,
|
| 3285 |
+
"step": 3450
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.69,
|
| 3289 |
+
"learning_rate": 0.0002,
|
| 3290 |
+
"loss": 0.7486,
|
| 3291 |
+
"step": 3460
|
| 3292 |
+
},
|
| 3293 |
+
{
|
| 3294 |
+
"epoch": 0.69,
|
| 3295 |
+
"learning_rate": 0.0002,
|
| 3296 |
+
"loss": 0.7866,
|
| 3297 |
+
"step": 3470
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"epoch": 0.69,
|
| 3301 |
+
"learning_rate": 0.0002,
|
| 3302 |
+
"loss": 0.762,
|
| 3303 |
+
"step": 3480
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.7,
|
| 3307 |
+
"learning_rate": 0.0002,
|
| 3308 |
+
"loss": 0.8271,
|
| 3309 |
+
"step": 3490
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.7,
|
| 3313 |
+
"learning_rate": 0.0002,
|
| 3314 |
+
"loss": 0.7367,
|
| 3315 |
+
"step": 3500
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 0.7,
|
| 3319 |
+
"learning_rate": 0.0002,
|
| 3320 |
+
"loss": 0.7989,
|
| 3321 |
+
"step": 3510
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 0.7,
|
| 3325 |
+
"learning_rate": 0.0002,
|
| 3326 |
+
"loss": 0.7986,
|
| 3327 |
+
"step": 3520
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.7,
|
| 3331 |
+
"learning_rate": 0.0002,
|
| 3332 |
+
"loss": 0.7712,
|
| 3333 |
+
"step": 3530
|
| 3334 |
+
},
|
| 3335 |
+
{
|
| 3336 |
+
"epoch": 0.71,
|
| 3337 |
+
"learning_rate": 0.0002,
|
| 3338 |
+
"loss": 0.8605,
|
| 3339 |
+
"step": 3540
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 0.71,
|
| 3343 |
+
"learning_rate": 0.0002,
|
| 3344 |
+
"loss": 0.7465,
|
| 3345 |
+
"step": 3550
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.71,
|
| 3349 |
+
"learning_rate": 0.0002,
|
| 3350 |
+
"loss": 0.8129,
|
| 3351 |
+
"step": 3560
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.71,
|
| 3355 |
+
"learning_rate": 0.0002,
|
| 3356 |
+
"loss": 0.8049,
|
| 3357 |
+
"step": 3570
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 0.71,
|
| 3361 |
+
"learning_rate": 0.0002,
|
| 3362 |
+
"loss": 0.8579,
|
| 3363 |
+
"step": 3580
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 0.72,
|
| 3367 |
+
"learning_rate": 0.0002,
|
| 3368 |
+
"loss": 0.8272,
|
| 3369 |
+
"step": 3590
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.72,
|
| 3373 |
+
"learning_rate": 0.0002,
|
| 3374 |
+
"loss": 0.6876,
|
| 3375 |
+
"step": 3600
|
| 3376 |
+
},
|
| 3377 |
+
{
|
| 3378 |
+
"epoch": 0.72,
|
| 3379 |
+
"eval_loss": 0.7700828909873962,
|
| 3380 |
+
"eval_runtime": 187.1333,
|
| 3381 |
+
"eval_samples_per_second": 5.344,
|
| 3382 |
+
"eval_steps_per_second": 2.672,
|
| 3383 |
+
"step": 3600
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.72,
|
| 3387 |
+
"mmlu_eval_accuracy": 0.4956696905096371,
|
| 3388 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
|
| 3389 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 3390 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 3391 |
+
"mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
|
| 3392 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
|
| 3393 |
+
"mmlu_eval_accuracy_college_biology": 0.3125,
|
| 3394 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 3395 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 3396 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 3397 |
+
"mmlu_eval_accuracy_college_medicine": 0.5,
|
| 3398 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 3399 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 3400 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
|
| 3401 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3402 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.3125,
|
| 3403 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
|
| 3404 |
+
"mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
|
| 3405 |
+
"mmlu_eval_accuracy_global_facts": 0.4,
|
| 3406 |
+
"mmlu_eval_accuracy_high_school_biology": 0.46875,
|
| 3407 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
|
| 3408 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3409 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
|
| 3410 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 3411 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 3412 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
|
| 3413 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 3414 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 3415 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 3416 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 3417 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
|
| 3418 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 3419 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3420 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 3421 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3422 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3423 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 3424 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3425 |
+
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
|
| 3426 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 3427 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 3428 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 3429 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 3430 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 3431 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.25,
|
| 3432 |
+
"mmlu_eval_accuracy_nutrition": 0.6666666666666666,
|
| 3433 |
+
"mmlu_eval_accuracy_philosophy": 0.5588235294117647,
|
| 3434 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 3435 |
+
"mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
|
| 3436 |
+
"mmlu_eval_accuracy_professional_law": 0.35294117647058826,
|
| 3437 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
|
| 3438 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4492753623188406,
|
| 3439 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 3440 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 3441 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 3442 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
|
| 3443 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 3444 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 3445 |
+
"mmlu_loss": 1.278034775126385,
|
| 3446 |
+
"step": 3600
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.72,
|
| 3450 |
+
"learning_rate": 0.0002,
|
| 3451 |
+
"loss": 0.7432,
|
| 3452 |
+
"step": 3610
|
| 3453 |
+
},
|
| 3454 |
+
{
|
| 3455 |
+
"epoch": 0.72,
|
| 3456 |
+
"learning_rate": 0.0002,
|
| 3457 |
+
"loss": 0.7804,
|
| 3458 |
+
"step": 3620
|
| 3459 |
+
},
|
| 3460 |
+
{
|
| 3461 |
+
"epoch": 0.72,
|
| 3462 |
+
"learning_rate": 0.0002,
|
| 3463 |
+
"loss": 0.8165,
|
| 3464 |
+
"step": 3630
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 0.73,
|
| 3468 |
+
"learning_rate": 0.0002,
|
| 3469 |
+
"loss": 0.7751,
|
| 3470 |
+
"step": 3640
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.73,
|
| 3474 |
+
"learning_rate": 0.0002,
|
| 3475 |
+
"loss": 0.6947,
|
| 3476 |
+
"step": 3650
|
| 3477 |
+
},
|
| 3478 |
+
{
|
| 3479 |
+
"epoch": 0.73,
|
| 3480 |
+
"learning_rate": 0.0002,
|
| 3481 |
+
"loss": 0.7461,
|
| 3482 |
+
"step": 3660
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 0.73,
|
| 3486 |
+
"learning_rate": 0.0002,
|
| 3487 |
+
"loss": 0.7294,
|
| 3488 |
+
"step": 3670
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.73,
|
| 3492 |
+
"learning_rate": 0.0002,
|
| 3493 |
+
"loss": 0.6992,
|
| 3494 |
+
"step": 3680
|
| 3495 |
+
},
|
| 3496 |
+
{
|
| 3497 |
+
"epoch": 0.74,
|
| 3498 |
+
"learning_rate": 0.0002,
|
| 3499 |
+
"loss": 0.8659,
|
| 3500 |
+
"step": 3690
|
| 3501 |
+
},
|
| 3502 |
+
{
|
| 3503 |
+
"epoch": 0.74,
|
| 3504 |
+
"learning_rate": 0.0002,
|
| 3505 |
+
"loss": 0.8452,
|
| 3506 |
+
"step": 3700
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 0.74,
|
| 3510 |
+
"learning_rate": 0.0002,
|
| 3511 |
+
"loss": 0.7714,
|
| 3512 |
+
"step": 3710
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.74,
|
| 3516 |
+
"learning_rate": 0.0002,
|
| 3517 |
+
"loss": 0.7721,
|
| 3518 |
+
"step": 3720
|
| 3519 |
+
},
|
| 3520 |
+
{
|
| 3521 |
+
"epoch": 0.74,
|
| 3522 |
+
"learning_rate": 0.0002,
|
| 3523 |
+
"loss": 0.7642,
|
| 3524 |
+
"step": 3730
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 0.75,
|
| 3528 |
+
"learning_rate": 0.0002,
|
| 3529 |
+
"loss": 0.8614,
|
| 3530 |
+
"step": 3740
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.75,
|
| 3534 |
+
"learning_rate": 0.0002,
|
| 3535 |
+
"loss": 0.7979,
|
| 3536 |
+
"step": 3750
|
| 3537 |
+
},
|
| 3538 |
+
{
|
| 3539 |
+
"epoch": 0.75,
|
| 3540 |
+
"learning_rate": 0.0002,
|
| 3541 |
+
"loss": 0.7714,
|
| 3542 |
+
"step": 3760
|
| 3543 |
+
},
|
| 3544 |
+
{
|
| 3545 |
+
"epoch": 0.75,
|
| 3546 |
+
"learning_rate": 0.0002,
|
| 3547 |
+
"loss": 0.8341,
|
| 3548 |
+
"step": 3770
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 0.75,
|
| 3552 |
+
"learning_rate": 0.0002,
|
| 3553 |
+
"loss": 0.7559,
|
| 3554 |
+
"step": 3780
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.76,
|
| 3558 |
+
"learning_rate": 0.0002,
|
| 3559 |
+
"loss": 0.7893,
|
| 3560 |
+
"step": 3790
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 0.76,
|
| 3564 |
+
"learning_rate": 0.0002,
|
| 3565 |
+
"loss": 0.7029,
|
| 3566 |
+
"step": 3800
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 0.76,
|
| 3570 |
+
"eval_loss": 0.7690045833587646,
|
| 3571 |
+
"eval_runtime": 187.1512,
|
| 3572 |
+
"eval_samples_per_second": 5.343,
|
| 3573 |
+
"eval_steps_per_second": 2.672,
|
| 3574 |
+
"step": 3800
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 0.76,
|
| 3578 |
+
"mmlu_eval_accuracy": 0.4986652426405072,
|
| 3579 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
|
| 3580 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 3581 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 3582 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 3583 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
|
| 3584 |
+
"mmlu_eval_accuracy_college_biology": 0.4375,
|
| 3585 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 3586 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 3587 |
+
"mmlu_eval_accuracy_college_mathematics": 0.09090909090909091,
|
| 3588 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 3589 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 3590 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 3591 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 3592 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3593 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 3594 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
|
| 3595 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 3596 |
+
"mmlu_eval_accuracy_global_facts": 0.2,
|
| 3597 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 3598 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
|
| 3599 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3600 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 3601 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 3602 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 3603 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907,
|
| 3604 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 3605 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 3606 |
+
"mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
|
| 3607 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
|
| 3608 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 3609 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 3610 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
|
| 3611 |
+
"mmlu_eval_accuracy_human_aging": 0.7391304347826086,
|
| 3612 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3613 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3614 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 3615 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3616 |
+
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
|
| 3617 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 3618 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 3619 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 3620 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 3621 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 3622 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 3623 |
+
"mmlu_eval_accuracy_nutrition": 0.7272727272727273,
|
| 3624 |
+
"mmlu_eval_accuracy_philosophy": 0.5588235294117647,
|
| 3625 |
+
"mmlu_eval_accuracy_prehistory": 0.5714285714285714,
|
| 3626 |
+
"mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
|
| 3627 |
+
"mmlu_eval_accuracy_professional_law": 0.32941176470588235,
|
| 3628 |
+
"mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
|
| 3629 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 3630 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 3631 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 3632 |
+
"mmlu_eval_accuracy_sociology": 0.6363636363636364,
|
| 3633 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
|
| 3634 |
+
"mmlu_eval_accuracy_virology": 0.5555555555555556,
|
| 3635 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 3636 |
+
"mmlu_loss": 1.2290096188027304,
|
| 3637 |
+
"step": 3800
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 0.76,
|
| 3641 |
+
"learning_rate": 0.0002,
|
| 3642 |
+
"loss": 0.7765,
|
| 3643 |
+
"step": 3810
|
| 3644 |
+
},
|
| 3645 |
+
{
|
| 3646 |
+
"epoch": 0.76,
|
| 3647 |
+
"learning_rate": 0.0002,
|
| 3648 |
+
"loss": 0.7342,
|
| 3649 |
+
"step": 3820
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.76,
|
| 3653 |
+
"learning_rate": 0.0002,
|
| 3654 |
+
"loss": 0.8179,
|
| 3655 |
+
"step": 3830
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 0.77,
|
| 3659 |
+
"learning_rate": 0.0002,
|
| 3660 |
+
"loss": 0.7516,
|
| 3661 |
+
"step": 3840
|
| 3662 |
+
},
|
| 3663 |
+
{
|
| 3664 |
+
"epoch": 0.77,
|
| 3665 |
+
"learning_rate": 0.0002,
|
| 3666 |
+
"loss": 0.7402,
|
| 3667 |
+
"step": 3850
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 0.77,
|
| 3671 |
+
"learning_rate": 0.0002,
|
| 3672 |
+
"loss": 0.7458,
|
| 3673 |
+
"step": 3860
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.77,
|
| 3677 |
+
"learning_rate": 0.0002,
|
| 3678 |
+
"loss": 0.7975,
|
| 3679 |
+
"step": 3870
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 0.77,
|
| 3683 |
+
"learning_rate": 0.0002,
|
| 3684 |
+
"loss": 0.7606,
|
| 3685 |
+
"step": 3880
|
| 3686 |
+
},
|
| 3687 |
+
{
|
| 3688 |
+
"epoch": 0.78,
|
| 3689 |
+
"learning_rate": 0.0002,
|
| 3690 |
+
"loss": 0.7841,
|
| 3691 |
+
"step": 3890
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.78,
|
| 3695 |
+
"learning_rate": 0.0002,
|
| 3696 |
+
"loss": 0.8398,
|
| 3697 |
+
"step": 3900
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 0.78,
|
| 3701 |
+
"learning_rate": 0.0002,
|
| 3702 |
+
"loss": 0.8645,
|
| 3703 |
+
"step": 3910
|
| 3704 |
+
},
|
| 3705 |
+
{
|
| 3706 |
+
"epoch": 0.78,
|
| 3707 |
+
"learning_rate": 0.0002,
|
| 3708 |
+
"loss": 0.8749,
|
| 3709 |
+
"step": 3920
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 0.78,
|
| 3713 |
+
"learning_rate": 0.0002,
|
| 3714 |
+
"loss": 0.7782,
|
| 3715 |
+
"step": 3930
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.79,
|
| 3719 |
+
"learning_rate": 0.0002,
|
| 3720 |
+
"loss": 0.7548,
|
| 3721 |
+
"step": 3940
|
| 3722 |
+
},
|
| 3723 |
+
{
|
| 3724 |
+
"epoch": 0.79,
|
| 3725 |
+
"learning_rate": 0.0002,
|
| 3726 |
+
"loss": 0.7762,
|
| 3727 |
+
"step": 3950
|
| 3728 |
+
},
|
| 3729 |
+
{
|
| 3730 |
+
"epoch": 0.79,
|
| 3731 |
+
"learning_rate": 0.0002,
|
| 3732 |
+
"loss": 0.6853,
|
| 3733 |
+
"step": 3960
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.79,
|
| 3737 |
+
"learning_rate": 0.0002,
|
| 3738 |
+
"loss": 0.7392,
|
| 3739 |
+
"step": 3970
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 0.79,
|
| 3743 |
+
"learning_rate": 0.0002,
|
| 3744 |
+
"loss": 0.8183,
|
| 3745 |
+
"step": 3980
|
| 3746 |
+
},
|
| 3747 |
+
{
|
| 3748 |
+
"epoch": 0.8,
|
| 3749 |
+
"learning_rate": 0.0002,
|
| 3750 |
+
"loss": 0.7711,
|
| 3751 |
+
"step": 3990
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 0.8,
|
| 3755 |
+
"learning_rate": 0.0002,
|
| 3756 |
+
"loss": 0.7947,
|
| 3757 |
+
"step": 4000
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.8,
|
| 3761 |
+
"eval_loss": 0.7685290575027466,
|
| 3762 |
+
"eval_runtime": 187.1502,
|
| 3763 |
+
"eval_samples_per_second": 5.343,
|
| 3764 |
+
"eval_steps_per_second": 2.672,
|
| 3765 |
+
"step": 4000
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 0.8,
|
| 3769 |
+
"mmlu_eval_accuracy": 0.49482145136513617,
|
| 3770 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 3771 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 3772 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 3773 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 3774 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
|
| 3775 |
+
"mmlu_eval_accuracy_college_biology": 0.3125,
|
| 3776 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 3777 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 3778 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 3779 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 3780 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 3781 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 3782 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 3783 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3784 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 3785 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
|
| 3786 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 3787 |
+
"mmlu_eval_accuracy_global_facts": 0.4,
|
| 3788 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 3789 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
|
| 3790 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3791 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 3792 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 3793 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 3794 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4418604651162791,
|
| 3795 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 3796 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
|
| 3797 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 3798 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
|
| 3799 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 3800 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
|
| 3801 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3802 |
+
"mmlu_eval_accuracy_human_aging": 0.782608695652174,
|
| 3803 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3804 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3805 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 3806 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3807 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 3808 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 3809 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 3810 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 3811 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 3812 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 3813 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.26,
|
| 3814 |
+
"mmlu_eval_accuracy_nutrition": 0.696969696969697,
|
| 3815 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 3816 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 3817 |
+
"mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
|
| 3818 |
+
"mmlu_eval_accuracy_professional_law": 0.3235294117647059,
|
| 3819 |
+
"mmlu_eval_accuracy_professional_medicine": 0.45161290322580644,
|
| 3820 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4492753623188406,
|
| 3821 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 3822 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 3823 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 3824 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
|
| 3825 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 3826 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 3827 |
+
"mmlu_loss": 1.4409859191344239,
|
| 3828 |
+
"step": 4000
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.8,
|
| 3832 |
+
"learning_rate": 0.0002,
|
| 3833 |
+
"loss": 0.8158,
|
| 3834 |
+
"step": 4010
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.8,
|
| 3838 |
+
"learning_rate": 0.0002,
|
| 3839 |
+
"loss": 0.7975,
|
| 3840 |
+
"step": 4020
|
| 3841 |
+
},
|
| 3842 |
+
{
|
| 3843 |
+
"epoch": 0.8,
|
| 3844 |
+
"learning_rate": 0.0002,
|
| 3845 |
+
"loss": 0.8133,
|
| 3846 |
+
"step": 4030
|
| 3847 |
+
},
|
| 3848 |
+
{
|
| 3849 |
+
"epoch": 0.81,
|
| 3850 |
+
"learning_rate": 0.0002,
|
| 3851 |
+
"loss": 0.7486,
|
| 3852 |
+
"step": 4040
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.81,
|
| 3856 |
+
"learning_rate": 0.0002,
|
| 3857 |
+
"loss": 0.8209,
|
| 3858 |
+
"step": 4050
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 0.81,
|
| 3862 |
+
"learning_rate": 0.0002,
|
| 3863 |
+
"loss": 0.8269,
|
| 3864 |
+
"step": 4060
|
| 3865 |
+
},
|
| 3866 |
+
{
|
| 3867 |
+
"epoch": 0.81,
|
| 3868 |
+
"learning_rate": 0.0002,
|
| 3869 |
+
"loss": 0.8395,
|
| 3870 |
+
"step": 4070
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.81,
|
| 3874 |
+
"learning_rate": 0.0002,
|
| 3875 |
+
"loss": 0.7845,
|
| 3876 |
+
"step": 4080
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.82,
|
| 3880 |
+
"learning_rate": 0.0002,
|
| 3881 |
+
"loss": 0.7609,
|
| 3882 |
+
"step": 4090
|
| 3883 |
+
},
|
| 3884 |
+
{
|
| 3885 |
+
"epoch": 0.82,
|
| 3886 |
+
"learning_rate": 0.0002,
|
| 3887 |
+
"loss": 0.8267,
|
| 3888 |
+
"step": 4100
|
| 3889 |
+
},
|
| 3890 |
+
{
|
| 3891 |
+
"epoch": 0.82,
|
| 3892 |
+
"learning_rate": 0.0002,
|
| 3893 |
+
"loss": 0.6963,
|
| 3894 |
+
"step": 4110
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.82,
|
| 3898 |
+
"learning_rate": 0.0002,
|
| 3899 |
+
"loss": 0.8173,
|
| 3900 |
+
"step": 4120
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 0.82,
|
| 3904 |
+
"learning_rate": 0.0002,
|
| 3905 |
+
"loss": 0.7732,
|
| 3906 |
+
"step": 4130
|
| 3907 |
+
},
|
| 3908 |
+
{
|
| 3909 |
+
"epoch": 0.83,
|
| 3910 |
+
"learning_rate": 0.0002,
|
| 3911 |
+
"loss": 0.8267,
|
| 3912 |
+
"step": 4140
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.83,
|
| 3916 |
+
"learning_rate": 0.0002,
|
| 3917 |
+
"loss": 0.7965,
|
| 3918 |
+
"step": 4150
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.83,
|
| 3922 |
+
"learning_rate": 0.0002,
|
| 3923 |
+
"loss": 0.7632,
|
| 3924 |
+
"step": 4160
|
| 3925 |
+
},
|
| 3926 |
+
{
|
| 3927 |
+
"epoch": 0.83,
|
| 3928 |
+
"learning_rate": 0.0002,
|
| 3929 |
+
"loss": 0.7595,
|
| 3930 |
+
"step": 4170
|
| 3931 |
+
},
|
| 3932 |
+
{
|
| 3933 |
+
"epoch": 0.83,
|
| 3934 |
+
"learning_rate": 0.0002,
|
| 3935 |
+
"loss": 0.8642,
|
| 3936 |
+
"step": 4180
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.84,
|
| 3940 |
+
"learning_rate": 0.0002,
|
| 3941 |
+
"loss": 0.8094,
|
| 3942 |
+
"step": 4190
|
| 3943 |
+
},
|
| 3944 |
+
{
|
| 3945 |
+
"epoch": 0.84,
|
| 3946 |
+
"learning_rate": 0.0002,
|
| 3947 |
+
"loss": 0.7524,
|
| 3948 |
+
"step": 4200
|
| 3949 |
+
},
|
| 3950 |
+
{
|
| 3951 |
+
"epoch": 0.84,
|
| 3952 |
+
"eval_loss": 0.7668033838272095,
|
| 3953 |
+
"eval_runtime": 187.1363,
|
| 3954 |
+
"eval_samples_per_second": 5.344,
|
| 3955 |
+
"eval_steps_per_second": 2.672,
|
| 3956 |
+
"step": 4200
|
| 3957 |
+
},
|
| 3958 |
+
{
|
| 3959 |
+
"epoch": 0.84,
|
| 3960 |
+
"mmlu_eval_accuracy": 0.5033620342118746,
|
| 3961 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
|
| 3962 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 3963 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 3964 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 3965 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
|
| 3966 |
+
"mmlu_eval_accuracy_college_biology": 0.3125,
|
| 3967 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 3968 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 3969 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 3970 |
+
"mmlu_eval_accuracy_college_medicine": 0.5,
|
| 3971 |
+
"mmlu_eval_accuracy_college_physics": 0.5454545454545454,
|
| 3972 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 3973 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
|
| 3974 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3975 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 3976 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
|
| 3977 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 3978 |
+
"mmlu_eval_accuracy_global_facts": 0.4,
|
| 3979 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 3980 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
|
| 3981 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3982 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
|
| 3983 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 3984 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 3985 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4418604651162791,
|
| 3986 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
|
| 3987 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
|
| 3988 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 3989 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8,
|
| 3990 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 3991 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 3992 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3993 |
+
"mmlu_eval_accuracy_human_aging": 0.7391304347826086,
|
| 3994 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 3995 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 3996 |
+
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
|
| 3997 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3998 |
+
"mmlu_eval_accuracy_machine_learning": 0.09090909090909091,
|
| 3999 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4000 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 4001 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 4002 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 4003 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 4004 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 4005 |
+
"mmlu_eval_accuracy_nutrition": 0.5757575757575758,
|
| 4006 |
+
"mmlu_eval_accuracy_philosophy": 0.5882352941176471,
|
| 4007 |
+
"mmlu_eval_accuracy_prehistory": 0.5428571428571428,
|
| 4008 |
+
"mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
|
| 4009 |
+
"mmlu_eval_accuracy_professional_law": 0.3352941176470588,
|
| 4010 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 4011 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 4012 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4013 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4014 |
+
"mmlu_eval_accuracy_sociology": 0.6363636363636364,
|
| 4015 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 4016 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4017 |
+
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 4018 |
+
"mmlu_loss": 1.475248734882544,
|
| 4019 |
+
"step": 4200
|
| 4020 |
+
},
|
| 4021 |
+
{
|
| 4022 |
+
"epoch": 0.84,
|
| 4023 |
+
"learning_rate": 0.0002,
|
| 4024 |
+
"loss": 0.7157,
|
| 4025 |
+
"step": 4210
|
| 4026 |
+
},
|
| 4027 |
+
{
|
| 4028 |
+
"epoch": 0.84,
|
| 4029 |
+
"learning_rate": 0.0002,
|
| 4030 |
+
"loss": 0.758,
|
| 4031 |
+
"step": 4220
|
| 4032 |
+
},
|
| 4033 |
+
{
|
| 4034 |
+
"epoch": 0.84,
|
| 4035 |
+
"learning_rate": 0.0002,
|
| 4036 |
+
"loss": 0.7123,
|
| 4037 |
+
"step": 4230
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.85,
|
| 4041 |
+
"learning_rate": 0.0002,
|
| 4042 |
+
"loss": 0.8488,
|
| 4043 |
+
"step": 4240
|
| 4044 |
+
},
|
| 4045 |
+
{
|
| 4046 |
+
"epoch": 0.85,
|
| 4047 |
+
"learning_rate": 0.0002,
|
| 4048 |
+
"loss": 0.7902,
|
| 4049 |
+
"step": 4250
|
| 4050 |
+
},
|
| 4051 |
+
{
|
| 4052 |
+
"epoch": 0.85,
|
| 4053 |
+
"learning_rate": 0.0002,
|
| 4054 |
+
"loss": 0.7876,
|
| 4055 |
+
"step": 4260
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.85,
|
| 4059 |
+
"learning_rate": 0.0002,
|
| 4060 |
+
"loss": 0.8171,
|
| 4061 |
+
"step": 4270
|
| 4062 |
+
},
|
| 4063 |
+
{
|
| 4064 |
+
"epoch": 0.85,
|
| 4065 |
+
"learning_rate": 0.0002,
|
| 4066 |
+
"loss": 0.8478,
|
| 4067 |
+
"step": 4280
|
| 4068 |
+
},
|
| 4069 |
+
{
|
| 4070 |
+
"epoch": 0.86,
|
| 4071 |
+
"learning_rate": 0.0002,
|
| 4072 |
+
"loss": 0.7818,
|
| 4073 |
+
"step": 4290
|
| 4074 |
+
},
|
| 4075 |
+
{
|
| 4076 |
+
"epoch": 0.86,
|
| 4077 |
+
"learning_rate": 0.0002,
|
| 4078 |
+
"loss": 0.6867,
|
| 4079 |
+
"step": 4300
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.86,
|
| 4083 |
+
"learning_rate": 0.0002,
|
| 4084 |
+
"loss": 0.8269,
|
| 4085 |
+
"step": 4310
|
| 4086 |
+
},
|
| 4087 |
+
{
|
| 4088 |
+
"epoch": 0.86,
|
| 4089 |
+
"learning_rate": 0.0002,
|
| 4090 |
+
"loss": 0.6898,
|
| 4091 |
+
"step": 4320
|
| 4092 |
+
},
|
| 4093 |
+
{
|
| 4094 |
+
"epoch": 0.86,
|
| 4095 |
+
"learning_rate": 0.0002,
|
| 4096 |
+
"loss": 0.8375,
|
| 4097 |
+
"step": 4330
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.87,
|
| 4101 |
+
"learning_rate": 0.0002,
|
| 4102 |
+
"loss": 0.7823,
|
| 4103 |
+
"step": 4340
|
| 4104 |
+
},
|
| 4105 |
+
{
|
| 4106 |
+
"epoch": 0.87,
|
| 4107 |
+
"learning_rate": 0.0002,
|
| 4108 |
+
"loss": 0.8884,
|
| 4109 |
+
"step": 4350
|
| 4110 |
+
},
|
| 4111 |
+
{
|
| 4112 |
+
"epoch": 0.87,
|
| 4113 |
+
"learning_rate": 0.0002,
|
| 4114 |
+
"loss": 0.8705,
|
| 4115 |
+
"step": 4360
|
| 4116 |
+
},
|
| 4117 |
+
{
|
| 4118 |
+
"epoch": 0.87,
|
| 4119 |
+
"learning_rate": 0.0002,
|
| 4120 |
+
"loss": 0.7379,
|
| 4121 |
+
"step": 4370
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.87,
|
| 4125 |
+
"learning_rate": 0.0002,
|
| 4126 |
+
"loss": 0.6632,
|
| 4127 |
+
"step": 4380
|
| 4128 |
+
},
|
| 4129 |
+
{
|
| 4130 |
+
"epoch": 0.88,
|
| 4131 |
+
"learning_rate": 0.0002,
|
| 4132 |
+
"loss": 0.7939,
|
| 4133 |
+
"step": 4390
|
| 4134 |
+
},
|
| 4135 |
+
{
|
| 4136 |
+
"epoch": 0.88,
|
| 4137 |
+
"learning_rate": 0.0002,
|
| 4138 |
+
"loss": 0.7487,
|
| 4139 |
+
"step": 4400
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.88,
|
| 4143 |
+
"eval_loss": 0.7660865187644958,
|
| 4144 |
+
"eval_runtime": 187.2239,
|
| 4145 |
+
"eval_samples_per_second": 5.341,
|
| 4146 |
+
"eval_steps_per_second": 2.671,
|
| 4147 |
+
"step": 4400
|
| 4148 |
+
},
|
| 4149 |
+
{
|
| 4150 |
+
"epoch": 0.88,
|
| 4151 |
+
"mmlu_eval_accuracy": 0.49666140703774375,
|
| 4152 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
|
| 4153 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 4154 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 4155 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4156 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
|
| 4157 |
+
"mmlu_eval_accuracy_college_biology": 0.375,
|
| 4158 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 4159 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 4160 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 4161 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 4162 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 4163 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 4164 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
|
| 4165 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4166 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 4167 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
|
| 4168 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 4169 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4170 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 4171 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.5,
|
| 4172 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4173 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
|
| 4174 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 4175 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
|
| 4176 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.5116279069767442,
|
| 4177 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 4178 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.34615384615384615,
|
| 4179 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 4180 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
|
| 4181 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
|
| 4182 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
|
| 4183 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
|
| 4184 |
+
"mmlu_eval_accuracy_human_aging": 0.782608695652174,
|
| 4185 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4186 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 4187 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 4188 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
|
| 4189 |
+
"mmlu_eval_accuracy_machine_learning": 0.09090909090909091,
|
| 4190 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4191 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 4192 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 4193 |
+
"mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
|
| 4194 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 4195 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.26,
|
| 4196 |
+
"mmlu_eval_accuracy_nutrition": 0.6363636363636364,
|
| 4197 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 4198 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 4199 |
+
"mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
|
| 4200 |
+
"mmlu_eval_accuracy_professional_law": 0.3235294117647059,
|
| 4201 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
|
| 4202 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4782608695652174,
|
| 4203 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4204 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4205 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 4206 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
|
| 4207 |
+
"mmlu_eval_accuracy_virology": 0.4444444444444444,
|
| 4208 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4209 |
+
"mmlu_loss": 1.5290965444114126,
|
| 4210 |
+
"step": 4400
|
| 4211 |
+
},
|
| 4212 |
+
{
|
| 4213 |
+
"epoch": 0.88,
|
| 4214 |
+
"learning_rate": 0.0002,
|
| 4215 |
+
"loss": 0.8308,
|
| 4216 |
+
"step": 4410
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.88,
|
| 4220 |
+
"learning_rate": 0.0002,
|
| 4221 |
+
"loss": 0.8165,
|
| 4222 |
+
"step": 4420
|
| 4223 |
+
},
|
| 4224 |
+
{
|
| 4225 |
+
"epoch": 0.88,
|
| 4226 |
+
"learning_rate": 0.0002,
|
| 4227 |
+
"loss": 0.8891,
|
| 4228 |
+
"step": 4430
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 0.89,
|
| 4232 |
+
"learning_rate": 0.0002,
|
| 4233 |
+
"loss": 0.8207,
|
| 4234 |
+
"step": 4440
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 0.89,
|
| 4238 |
+
"learning_rate": 0.0002,
|
| 4239 |
+
"loss": 0.7602,
|
| 4240 |
+
"step": 4450
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 0.89,
|
| 4244 |
+
"learning_rate": 0.0002,
|
| 4245 |
+
"loss": 0.7541,
|
| 4246 |
+
"step": 4460
|
| 4247 |
+
},
|
| 4248 |
+
{
|
| 4249 |
+
"epoch": 0.89,
|
| 4250 |
+
"learning_rate": 0.0002,
|
| 4251 |
+
"loss": 0.7926,
|
| 4252 |
+
"step": 4470
|
| 4253 |
+
},
|
| 4254 |
+
{
|
| 4255 |
+
"epoch": 0.89,
|
| 4256 |
+
"learning_rate": 0.0002,
|
| 4257 |
+
"loss": 0.7693,
|
| 4258 |
+
"step": 4480
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.9,
|
| 4262 |
+
"learning_rate": 0.0002,
|
| 4263 |
+
"loss": 0.7225,
|
| 4264 |
+
"step": 4490
|
| 4265 |
+
},
|
| 4266 |
+
{
|
| 4267 |
+
"epoch": 0.9,
|
| 4268 |
+
"learning_rate": 0.0002,
|
| 4269 |
+
"loss": 0.9039,
|
| 4270 |
+
"step": 4500
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"epoch": 0.9,
|
| 4274 |
+
"learning_rate": 0.0002,
|
| 4275 |
+
"loss": 0.87,
|
| 4276 |
+
"step": 4510
|
| 4277 |
+
},
|
| 4278 |
+
{
|
| 4279 |
+
"epoch": 0.9,
|
| 4280 |
+
"learning_rate": 0.0002,
|
| 4281 |
+
"loss": 0.8597,
|
| 4282 |
+
"step": 4520
|
| 4283 |
+
},
|
| 4284 |
+
{
|
| 4285 |
+
"epoch": 0.9,
|
| 4286 |
+
"learning_rate": 0.0002,
|
| 4287 |
+
"loss": 0.8691,
|
| 4288 |
+
"step": 4530
|
| 4289 |
+
},
|
| 4290 |
+
{
|
| 4291 |
+
"epoch": 0.91,
|
| 4292 |
+
"learning_rate": 0.0002,
|
| 4293 |
+
"loss": 0.7924,
|
| 4294 |
+
"step": 4540
|
| 4295 |
+
},
|
| 4296 |
+
{
|
| 4297 |
+
"epoch": 0.91,
|
| 4298 |
+
"learning_rate": 0.0002,
|
| 4299 |
+
"loss": 0.8612,
|
| 4300 |
+
"step": 4550
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.91,
|
| 4304 |
+
"learning_rate": 0.0002,
|
| 4305 |
+
"loss": 0.7521,
|
| 4306 |
+
"step": 4560
|
| 4307 |
+
},
|
| 4308 |
+
{
|
| 4309 |
+
"epoch": 0.91,
|
| 4310 |
+
"learning_rate": 0.0002,
|
| 4311 |
+
"loss": 0.7845,
|
| 4312 |
+
"step": 4570
|
| 4313 |
+
},
|
| 4314 |
+
{
|
| 4315 |
+
"epoch": 0.91,
|
| 4316 |
+
"learning_rate": 0.0002,
|
| 4317 |
+
"loss": 0.7594,
|
| 4318 |
+
"step": 4580
|
| 4319 |
+
},
|
| 4320 |
+
{
|
| 4321 |
+
"epoch": 0.92,
|
| 4322 |
+
"learning_rate": 0.0002,
|
| 4323 |
+
"loss": 0.703,
|
| 4324 |
+
"step": 4590
|
| 4325 |
+
},
|
| 4326 |
+
{
|
| 4327 |
+
"epoch": 0.92,
|
| 4328 |
+
"learning_rate": 0.0002,
|
| 4329 |
+
"loss": 0.7429,
|
| 4330 |
+
"step": 4600
|
| 4331 |
+
},
|
| 4332 |
+
{
|
| 4333 |
+
"epoch": 0.92,
|
| 4334 |
+
"eval_loss": 0.7656491994857788,
|
| 4335 |
+
"eval_runtime": 187.0859,
|
| 4336 |
+
"eval_samples_per_second": 5.345,
|
| 4337 |
+
"eval_steps_per_second": 2.673,
|
| 4338 |
+
"step": 4600
|
| 4339 |
+
},
|
| 4340 |
+
{
|
| 4341 |
+
"epoch": 0.92,
|
| 4342 |
+
"mmlu_eval_accuracy": 0.49214003939461304,
|
| 4343 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
|
| 4344 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 4345 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 4346 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4347 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
|
| 4348 |
+
"mmlu_eval_accuracy_college_biology": 0.375,
|
| 4349 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 4350 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 4351 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 4352 |
+
"mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
|
| 4353 |
+
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
|
| 4354 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 4355 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 4356 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4357 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.3125,
|
| 4358 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
|
| 4359 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 4360 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4361 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 4362 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
|
| 4363 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4364 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 4365 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 4366 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 4367 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.46511627906976744,
|
| 4368 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
|
| 4369 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.34615384615384615,
|
| 4370 |
+
"mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
|
| 4371 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
|
| 4372 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 4373 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
|
| 4374 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 4375 |
+
"mmlu_eval_accuracy_human_aging": 0.7391304347826086,
|
| 4376 |
+
"mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
|
| 4377 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 4378 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 4379 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
|
| 4380 |
+
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
|
| 4381 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4382 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 4383 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 4384 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 4385 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 4386 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.26,
|
| 4387 |
+
"mmlu_eval_accuracy_nutrition": 0.6666666666666666,
|
| 4388 |
+
"mmlu_eval_accuracy_philosophy": 0.5588235294117647,
|
| 4389 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 4390 |
+
"mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
|
| 4391 |
+
"mmlu_eval_accuracy_professional_law": 0.3235294117647059,
|
| 4392 |
+
"mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
|
| 4393 |
+
"mmlu_eval_accuracy_professional_psychology": 0.43478260869565216,
|
| 4394 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4395 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4396 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 4397 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
|
| 4398 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4399 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4400 |
+
"mmlu_loss": 1.473437037815002,
|
| 4401 |
+
"step": 4600
|
| 4402 |
+
},
|
| 4403 |
+
{
|
| 4404 |
+
"epoch": 0.92,
|
| 4405 |
+
"learning_rate": 0.0002,
|
| 4406 |
+
"loss": 0.7824,
|
| 4407 |
+
"step": 4610
|
| 4408 |
+
},
|
| 4409 |
+
{
|
| 4410 |
+
"epoch": 0.92,
|
| 4411 |
+
"learning_rate": 0.0002,
|
| 4412 |
+
"loss": 0.8594,
|
| 4413 |
+
"step": 4620
|
| 4414 |
+
},
|
| 4415 |
+
{
|
| 4416 |
+
"epoch": 0.92,
|
| 4417 |
+
"learning_rate": 0.0002,
|
| 4418 |
+
"loss": 0.7766,
|
| 4419 |
+
"step": 4630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.93,
|
| 4423 |
+
"learning_rate": 0.0002,
|
| 4424 |
+
"loss": 0.7882,
|
| 4425 |
+
"step": 4640
|
| 4426 |
+
},
|
| 4427 |
+
{
|
| 4428 |
+
"epoch": 0.93,
|
| 4429 |
+
"learning_rate": 0.0002,
|
| 4430 |
+
"loss": 0.7788,
|
| 4431 |
+
"step": 4650
|
| 4432 |
+
},
|
| 4433 |
+
{
|
| 4434 |
+
"epoch": 0.93,
|
| 4435 |
+
"learning_rate": 0.0002,
|
| 4436 |
+
"loss": 0.7974,
|
| 4437 |
+
"step": 4660
|
| 4438 |
+
},
|
| 4439 |
+
{
|
| 4440 |
+
"epoch": 0.93,
|
| 4441 |
+
"learning_rate": 0.0002,
|
| 4442 |
+
"loss": 0.887,
|
| 4443 |
+
"step": 4670
|
| 4444 |
+
},
|
| 4445 |
+
{
|
| 4446 |
+
"epoch": 0.93,
|
| 4447 |
+
"learning_rate": 0.0002,
|
| 4448 |
+
"loss": 0.7948,
|
| 4449 |
+
"step": 4680
|
| 4450 |
+
},
|
| 4451 |
+
{
|
| 4452 |
+
"epoch": 0.94,
|
| 4453 |
+
"learning_rate": 0.0002,
|
| 4454 |
+
"loss": 0.8338,
|
| 4455 |
+
"step": 4690
|
| 4456 |
+
},
|
| 4457 |
+
{
|
| 4458 |
+
"epoch": 0.94,
|
| 4459 |
+
"learning_rate": 0.0002,
|
| 4460 |
+
"loss": 0.8122,
|
| 4461 |
+
"step": 4700
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.94,
|
| 4465 |
+
"learning_rate": 0.0002,
|
| 4466 |
+
"loss": 0.7172,
|
| 4467 |
+
"step": 4710
|
| 4468 |
+
},
|
| 4469 |
+
{
|
| 4470 |
+
"epoch": 0.94,
|
| 4471 |
+
"learning_rate": 0.0002,
|
| 4472 |
+
"loss": 0.8443,
|
| 4473 |
+
"step": 4720
|
| 4474 |
+
},
|
| 4475 |
+
{
|
| 4476 |
+
"epoch": 0.94,
|
| 4477 |
+
"learning_rate": 0.0002,
|
| 4478 |
+
"loss": 0.753,
|
| 4479 |
+
"step": 4730
|
| 4480 |
+
},
|
| 4481 |
+
{
|
| 4482 |
+
"epoch": 0.95,
|
| 4483 |
+
"learning_rate": 0.0002,
|
| 4484 |
+
"loss": 0.8285,
|
| 4485 |
+
"step": 4740
|
| 4486 |
+
},
|
| 4487 |
+
{
|
| 4488 |
+
"epoch": 0.95,
|
| 4489 |
+
"learning_rate": 0.0002,
|
| 4490 |
+
"loss": 0.8201,
|
| 4491 |
+
"step": 4750
|
| 4492 |
+
},
|
| 4493 |
+
{
|
| 4494 |
+
"epoch": 0.95,
|
| 4495 |
+
"learning_rate": 0.0002,
|
| 4496 |
+
"loss": 0.7593,
|
| 4497 |
+
"step": 4760
|
| 4498 |
+
},
|
| 4499 |
+
{
|
| 4500 |
+
"epoch": 0.95,
|
| 4501 |
+
"learning_rate": 0.0002,
|
| 4502 |
+
"loss": 0.8614,
|
| 4503 |
+
"step": 4770
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.95,
|
| 4507 |
+
"learning_rate": 0.0002,
|
| 4508 |
+
"loss": 0.7837,
|
| 4509 |
+
"step": 4780
|
| 4510 |
+
},
|
| 4511 |
+
{
|
| 4512 |
+
"epoch": 0.96,
|
| 4513 |
+
"learning_rate": 0.0002,
|
| 4514 |
+
"loss": 0.8251,
|
| 4515 |
+
"step": 4790
|
| 4516 |
+
},
|
| 4517 |
+
{
|
| 4518 |
+
"epoch": 0.96,
|
| 4519 |
+
"learning_rate": 0.0002,
|
| 4520 |
+
"loss": 0.8472,
|
| 4521 |
+
"step": 4800
|
| 4522 |
+
},
|
| 4523 |
+
{
|
| 4524 |
+
"epoch": 0.96,
|
| 4525 |
+
"eval_loss": 0.7640769481658936,
|
| 4526 |
+
"eval_runtime": 187.1184,
|
| 4527 |
+
"eval_samples_per_second": 5.344,
|
| 4528 |
+
"eval_steps_per_second": 2.672,
|
| 4529 |
+
"step": 4800
|
| 4530 |
+
},
|
| 4531 |
+
{
|
| 4532 |
+
"epoch": 0.96,
|
| 4533 |
+
"mmlu_eval_accuracy": 0.4917410322972955,
|
| 4534 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
|
| 4535 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 4536 |
+
"mmlu_eval_accuracy_astronomy": 0.4375,
|
| 4537 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4538 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
|
| 4539 |
+
"mmlu_eval_accuracy_college_biology": 0.4375,
|
| 4540 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 4541 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 4542 |
+
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
|
| 4543 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 4544 |
+
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
|
| 4545 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 4546 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 4547 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4548 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.375,
|
| 4549 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
|
| 4550 |
+
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
|
| 4551 |
+
"mmlu_eval_accuracy_global_facts": 0.3,
|
| 4552 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 4553 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.5,
|
| 4554 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4555 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 4556 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 4557 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 4558 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4883720930232558,
|
| 4559 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
|
| 4560 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
|
| 4561 |
+
"mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
|
| 4562 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 4563 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 4564 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
|
| 4565 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 4566 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 4567 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4568 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 4569 |
+
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
|
| 4570 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 4571 |
+
"mmlu_eval_accuracy_machine_learning": 0.09090909090909091,
|
| 4572 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4573 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 4574 |
+
"mmlu_eval_accuracy_medical_genetics": 1.0,
|
| 4575 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 4576 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 4577 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.25,
|
| 4578 |
+
"mmlu_eval_accuracy_nutrition": 0.696969696969697,
|
| 4579 |
+
"mmlu_eval_accuracy_philosophy": 0.5,
|
| 4580 |
+
"mmlu_eval_accuracy_prehistory": 0.5428571428571428,
|
| 4581 |
+
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
|
| 4582 |
+
"mmlu_eval_accuracy_professional_law": 0.3235294117647059,
|
| 4583 |
+
"mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
|
| 4584 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4492753623188406,
|
| 4585 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 4586 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4587 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 4588 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 4589 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4590 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4591 |
+
"mmlu_loss": 1.43259734448347,
|
| 4592 |
+
"step": 4800
|
| 4593 |
}
|
| 4594 |
],
|
| 4595 |
"max_steps": 10000,
|
| 4596 |
"num_train_epochs": 2,
|
| 4597 |
+
"total_flos": 6.888218289121198e+17,
|
| 4598 |
"trial_name": null,
|
| 4599 |
"trial_params": null
|
| 4600 |
}
|
{checkpoint-2600 → checkpoint-4800}/training_args.bin
RENAMED
|
File without changes
|