Farouk commited on
Commit ·
fefffee
1
Parent(s): 38fd34a
Training in progress, step 6000
Browse files- adapter_model.bin +1 -1
- checkpoint-3400/adapter_model/adapter_model/README.md +12 -0
- checkpoint-3400/adapter_model/adapter_model/adapter_model.bin +1 -1
- {checkpoint-4000 → checkpoint-6000}/README.md +0 -0
- {checkpoint-4000 → checkpoint-6000}/adapter_config.json +0 -0
- {checkpoint-4000 → checkpoint-6000}/adapter_model.bin +1 -1
- {checkpoint-4000 → checkpoint-6000}/added_tokens.json +0 -0
- {checkpoint-4000 → checkpoint-6000}/optimizer.pt +1 -1
- {checkpoint-4000 → checkpoint-6000}/rng_state.pth +1 -1
- {checkpoint-4000 → checkpoint-6000}/scheduler.pt +1 -1
- {checkpoint-4000 → checkpoint-6000}/special_tokens_map.json +0 -0
- {checkpoint-4000 → checkpoint-6000}/tokenizer.model +0 -0
- {checkpoint-4000 → checkpoint-6000}/tokenizer_config.json +0 -0
- {checkpoint-4000 → checkpoint-6000}/trainer_state.json +1913 -3
- {checkpoint-4000 → checkpoint-6000}/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 sha256:34e9b01a0b4cb6fd7c0c3c26081ce5bb97e3e7ce5e57a00de45d3a2ed3fd9c6a
|
| 3 |
size 319977229
|
checkpoint-3400/adapter_model/adapter_model/README.md
CHANGED
|
@@ -125,6 +125,17 @@ The following `bitsandbytes` quantization config was used during training:
|
|
| 125 |
- bnb_4bit_use_double_quant: True
|
| 126 |
- bnb_4bit_compute_dtype: bfloat16
|
| 127 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 128 |
The following `bitsandbytes` quantization config was used during training:
|
| 129 |
- load_in_8bit: False
|
| 130 |
- load_in_4bit: True
|
|
@@ -148,5 +159,6 @@ The following `bitsandbytes` quantization config was used during training:
|
|
| 148 |
- PEFT 0.4.0
|
| 149 |
- PEFT 0.4.0
|
| 150 |
- PEFT 0.4.0
|
|
|
|
| 151 |
|
| 152 |
- PEFT 0.4.0
|
|
|
|
| 125 |
- bnb_4bit_use_double_quant: True
|
| 126 |
- bnb_4bit_compute_dtype: bfloat16
|
| 127 |
|
| 128 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 129 |
+
- load_in_8bit: False
|
| 130 |
+
- load_in_4bit: True
|
| 131 |
+
- llm_int8_threshold: 6.0
|
| 132 |
+
- llm_int8_skip_modules: None
|
| 133 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 134 |
+
- llm_int8_has_fp16_weight: False
|
| 135 |
+
- bnb_4bit_quant_type: nf4
|
| 136 |
+
- bnb_4bit_use_double_quant: True
|
| 137 |
+
- bnb_4bit_compute_dtype: bfloat16
|
| 138 |
+
|
| 139 |
The following `bitsandbytes` quantization config was used during training:
|
| 140 |
- load_in_8bit: False
|
| 141 |
- load_in_4bit: True
|
|
|
|
| 159 |
- PEFT 0.4.0
|
| 160 |
- PEFT 0.4.0
|
| 161 |
- PEFT 0.4.0
|
| 162 |
+
- PEFT 0.4.0
|
| 163 |
|
| 164 |
- PEFT 0.4.0
|
checkpoint-3400/adapter_model/adapter_model/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 sha256:91c2f18577569044eb89ecec339c323a78fd6de382dd9f5c5c774a98f120e0dd
|
| 3 |
size 319977229
|
{checkpoint-4000 → checkpoint-6000}/README.md
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/adapter_config.json
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/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 sha256:34e9b01a0b4cb6fd7c0c3c26081ce5bb97e3e7ce5e57a00de45d3a2ed3fd9c6a
|
| 3 |
size 319977229
|
{checkpoint-4000 → checkpoint-6000}/added_tokens.json
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/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 sha256:3afc95c043a2cc88a8a642ced7d66602c52de0897048f21f1abb1c0fec9357ca
|
| 3 |
size 1279539973
|
{checkpoint-4000 → checkpoint-6000}/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 sha256:89db6f96789f9a8207a982b0707b344994d55a5738ff31632cb5b393156940c8
|
| 3 |
size 14511
|
{checkpoint-4000 → checkpoint-6000}/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 sha256:1d57105578b154c1429d6c6085dcbdd82165ec497fa3a0fd74932549f9aaac46
|
| 3 |
size 627
|
{checkpoint-4000 → checkpoint-6000}/special_tokens_map.json
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/tokenizer.model
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/tokenizer_config.json
RENAMED
|
File without changes
|
{checkpoint-4000 → checkpoint-6000}/trainer_state.json
RENAMED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.8042147159576416,
|
| 3 |
"best_model_checkpoint": "experts/expert-25/checkpoint-3400",
|
| 4 |
-
"epoch": 1.
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -3826,11 +3826,1921 @@
|
|
| 3826 |
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 3827 |
"mmlu_loss": 1.3634479641914368,
|
| 3828 |
"step": 4000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3829 |
}
|
| 3830 |
],
|
| 3831 |
"max_steps": 10000,
|
| 3832 |
"num_train_epochs": 3,
|
| 3833 |
-
"total_flos": 1.
|
| 3834 |
"trial_name": null,
|
| 3835 |
"trial_params": null
|
| 3836 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.8042147159576416,
|
| 3 |
"best_model_checkpoint": "experts/expert-25/checkpoint-3400",
|
| 4 |
+
"epoch": 1.7469791818314166,
|
| 5 |
+
"global_step": 6000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 3826 |
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 3827 |
"mmlu_loss": 1.3634479641914368,
|
| 3828 |
"step": 4000
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 1.17,
|
| 3832 |
+
"learning_rate": 0.0002,
|
| 3833 |
+
"loss": 0.919,
|
| 3834 |
+
"step": 4010
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 1.17,
|
| 3838 |
+
"learning_rate": 0.0002,
|
| 3839 |
+
"loss": 0.8806,
|
| 3840 |
+
"step": 4020
|
| 3841 |
+
},
|
| 3842 |
+
{
|
| 3843 |
+
"epoch": 1.17,
|
| 3844 |
+
"learning_rate": 0.0002,
|
| 3845 |
+
"loss": 0.8799,
|
| 3846 |
+
"step": 4030
|
| 3847 |
+
},
|
| 3848 |
+
{
|
| 3849 |
+
"epoch": 1.18,
|
| 3850 |
+
"learning_rate": 0.0002,
|
| 3851 |
+
"loss": 0.8643,
|
| 3852 |
+
"step": 4040
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 1.18,
|
| 3856 |
+
"learning_rate": 0.0002,
|
| 3857 |
+
"loss": 0.8289,
|
| 3858 |
+
"step": 4050
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 1.18,
|
| 3862 |
+
"learning_rate": 0.0002,
|
| 3863 |
+
"loss": 0.844,
|
| 3864 |
+
"step": 4060
|
| 3865 |
+
},
|
| 3866 |
+
{
|
| 3867 |
+
"epoch": 1.19,
|
| 3868 |
+
"learning_rate": 0.0002,
|
| 3869 |
+
"loss": 0.8762,
|
| 3870 |
+
"step": 4070
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 1.19,
|
| 3874 |
+
"learning_rate": 0.0002,
|
| 3875 |
+
"loss": 0.9243,
|
| 3876 |
+
"step": 4080
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 1.19,
|
| 3880 |
+
"learning_rate": 0.0002,
|
| 3881 |
+
"loss": 0.9167,
|
| 3882 |
+
"step": 4090
|
| 3883 |
+
},
|
| 3884 |
+
{
|
| 3885 |
+
"epoch": 1.19,
|
| 3886 |
+
"learning_rate": 0.0002,
|
| 3887 |
+
"loss": 0.8258,
|
| 3888 |
+
"step": 4100
|
| 3889 |
+
},
|
| 3890 |
+
{
|
| 3891 |
+
"epoch": 1.2,
|
| 3892 |
+
"learning_rate": 0.0002,
|
| 3893 |
+
"loss": 0.8734,
|
| 3894 |
+
"step": 4110
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 1.2,
|
| 3898 |
+
"learning_rate": 0.0002,
|
| 3899 |
+
"loss": 0.9008,
|
| 3900 |
+
"step": 4120
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 1.2,
|
| 3904 |
+
"learning_rate": 0.0002,
|
| 3905 |
+
"loss": 0.891,
|
| 3906 |
+
"step": 4130
|
| 3907 |
+
},
|
| 3908 |
+
{
|
| 3909 |
+
"epoch": 1.21,
|
| 3910 |
+
"learning_rate": 0.0002,
|
| 3911 |
+
"loss": 0.8641,
|
| 3912 |
+
"step": 4140
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 1.21,
|
| 3916 |
+
"learning_rate": 0.0002,
|
| 3917 |
+
"loss": 0.8145,
|
| 3918 |
+
"step": 4150
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 1.21,
|
| 3922 |
+
"learning_rate": 0.0002,
|
| 3923 |
+
"loss": 0.7852,
|
| 3924 |
+
"step": 4160
|
| 3925 |
+
},
|
| 3926 |
+
{
|
| 3927 |
+
"epoch": 1.21,
|
| 3928 |
+
"learning_rate": 0.0002,
|
| 3929 |
+
"loss": 0.8739,
|
| 3930 |
+
"step": 4170
|
| 3931 |
+
},
|
| 3932 |
+
{
|
| 3933 |
+
"epoch": 1.22,
|
| 3934 |
+
"learning_rate": 0.0002,
|
| 3935 |
+
"loss": 0.8396,
|
| 3936 |
+
"step": 4180
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 1.22,
|
| 3940 |
+
"learning_rate": 0.0002,
|
| 3941 |
+
"loss": 0.902,
|
| 3942 |
+
"step": 4190
|
| 3943 |
+
},
|
| 3944 |
+
{
|
| 3945 |
+
"epoch": 1.22,
|
| 3946 |
+
"learning_rate": 0.0002,
|
| 3947 |
+
"loss": 0.8656,
|
| 3948 |
+
"step": 4200
|
| 3949 |
+
},
|
| 3950 |
+
{
|
| 3951 |
+
"epoch": 1.22,
|
| 3952 |
+
"eval_loss": 0.8079001307487488,
|
| 3953 |
+
"eval_runtime": 112.1854,
|
| 3954 |
+
"eval_samples_per_second": 8.914,
|
| 3955 |
+
"eval_steps_per_second": 4.457,
|
| 3956 |
+
"step": 4200
|
| 3957 |
+
},
|
| 3958 |
+
{
|
| 3959 |
+
"epoch": 1.22,
|
| 3960 |
+
"mmlu_eval_accuracy": 0.5094746566834845,
|
| 3961 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 3962 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 3963 |
+
"mmlu_eval_accuracy_astronomy": 0.5625,
|
| 3964 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 3965 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 3966 |
+
"mmlu_eval_accuracy_college_biology": 0.4375,
|
| 3967 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 3968 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 3969 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 3970 |
+
"mmlu_eval_accuracy_college_medicine": 0.5,
|
| 3971 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 3972 |
+
"mmlu_eval_accuracy_computer_security": 0.5454545454545454,
|
| 3973 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 3974 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 3975 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 3976 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
|
| 3977 |
+
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
|
| 3978 |
+
"mmlu_eval_accuracy_global_facts": 0.6,
|
| 3979 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 3980 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
|
| 3981 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 3982 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 3983 |
+
"mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
|
| 3984 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5238095238095238,
|
| 3985 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4418604651162791,
|
| 3986 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
|
| 3987 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.5,
|
| 3988 |
+
"mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
|
| 3989 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 3990 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
|
| 3991 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
|
| 3992 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 3993 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 3994 |
+
"mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
|
| 3995 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 3996 |
+
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
|
| 3997 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 3998 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 3999 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4000 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 4001 |
+
"mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
|
| 4002 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 4003 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 4004 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 4005 |
+
"mmlu_eval_accuracy_nutrition": 0.696969696969697,
|
| 4006 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 4007 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 4008 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
|
| 4009 |
+
"mmlu_eval_accuracy_professional_law": 0.3588235294117647,
|
| 4010 |
+
"mmlu_eval_accuracy_professional_medicine": 0.4838709677419355,
|
| 4011 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4782608695652174,
|
| 4012 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4013 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4014 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 4015 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 4016 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4017 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4018 |
+
"mmlu_loss": 1.4093773849794822,
|
| 4019 |
+
"step": 4200
|
| 4020 |
+
},
|
| 4021 |
+
{
|
| 4022 |
+
"epoch": 1.23,
|
| 4023 |
+
"learning_rate": 0.0002,
|
| 4024 |
+
"loss": 0.7718,
|
| 4025 |
+
"step": 4210
|
| 4026 |
+
},
|
| 4027 |
+
{
|
| 4028 |
+
"epoch": 1.23,
|
| 4029 |
+
"learning_rate": 0.0002,
|
| 4030 |
+
"loss": 0.8542,
|
| 4031 |
+
"step": 4220
|
| 4032 |
+
},
|
| 4033 |
+
{
|
| 4034 |
+
"epoch": 1.23,
|
| 4035 |
+
"learning_rate": 0.0002,
|
| 4036 |
+
"loss": 0.789,
|
| 4037 |
+
"step": 4230
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 1.23,
|
| 4041 |
+
"learning_rate": 0.0002,
|
| 4042 |
+
"loss": 0.7869,
|
| 4043 |
+
"step": 4240
|
| 4044 |
+
},
|
| 4045 |
+
{
|
| 4046 |
+
"epoch": 1.24,
|
| 4047 |
+
"learning_rate": 0.0002,
|
| 4048 |
+
"loss": 0.7837,
|
| 4049 |
+
"step": 4250
|
| 4050 |
+
},
|
| 4051 |
+
{
|
| 4052 |
+
"epoch": 1.24,
|
| 4053 |
+
"learning_rate": 0.0002,
|
| 4054 |
+
"loss": 0.8693,
|
| 4055 |
+
"step": 4260
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 1.24,
|
| 4059 |
+
"learning_rate": 0.0002,
|
| 4060 |
+
"loss": 0.8024,
|
| 4061 |
+
"step": 4270
|
| 4062 |
+
},
|
| 4063 |
+
{
|
| 4064 |
+
"epoch": 1.25,
|
| 4065 |
+
"learning_rate": 0.0002,
|
| 4066 |
+
"loss": 0.8672,
|
| 4067 |
+
"step": 4280
|
| 4068 |
+
},
|
| 4069 |
+
{
|
| 4070 |
+
"epoch": 1.25,
|
| 4071 |
+
"learning_rate": 0.0002,
|
| 4072 |
+
"loss": 0.8778,
|
| 4073 |
+
"step": 4290
|
| 4074 |
+
},
|
| 4075 |
+
{
|
| 4076 |
+
"epoch": 1.25,
|
| 4077 |
+
"learning_rate": 0.0002,
|
| 4078 |
+
"loss": 0.89,
|
| 4079 |
+
"step": 4300
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 1.25,
|
| 4083 |
+
"learning_rate": 0.0002,
|
| 4084 |
+
"loss": 0.8435,
|
| 4085 |
+
"step": 4310
|
| 4086 |
+
},
|
| 4087 |
+
{
|
| 4088 |
+
"epoch": 1.26,
|
| 4089 |
+
"learning_rate": 0.0002,
|
| 4090 |
+
"loss": 0.9427,
|
| 4091 |
+
"step": 4320
|
| 4092 |
+
},
|
| 4093 |
+
{
|
| 4094 |
+
"epoch": 1.26,
|
| 4095 |
+
"learning_rate": 0.0002,
|
| 4096 |
+
"loss": 0.855,
|
| 4097 |
+
"step": 4330
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 1.26,
|
| 4101 |
+
"learning_rate": 0.0002,
|
| 4102 |
+
"loss": 0.8827,
|
| 4103 |
+
"step": 4340
|
| 4104 |
+
},
|
| 4105 |
+
{
|
| 4106 |
+
"epoch": 1.27,
|
| 4107 |
+
"learning_rate": 0.0002,
|
| 4108 |
+
"loss": 0.8848,
|
| 4109 |
+
"step": 4350
|
| 4110 |
+
},
|
| 4111 |
+
{
|
| 4112 |
+
"epoch": 1.27,
|
| 4113 |
+
"learning_rate": 0.0002,
|
| 4114 |
+
"loss": 0.8016,
|
| 4115 |
+
"step": 4360
|
| 4116 |
+
},
|
| 4117 |
+
{
|
| 4118 |
+
"epoch": 1.27,
|
| 4119 |
+
"learning_rate": 0.0002,
|
| 4120 |
+
"loss": 0.9674,
|
| 4121 |
+
"step": 4370
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 1.28,
|
| 4125 |
+
"learning_rate": 0.0002,
|
| 4126 |
+
"loss": 0.8914,
|
| 4127 |
+
"step": 4380
|
| 4128 |
+
},
|
| 4129 |
+
{
|
| 4130 |
+
"epoch": 1.28,
|
| 4131 |
+
"learning_rate": 0.0002,
|
| 4132 |
+
"loss": 0.8187,
|
| 4133 |
+
"step": 4390
|
| 4134 |
+
},
|
| 4135 |
+
{
|
| 4136 |
+
"epoch": 1.28,
|
| 4137 |
+
"learning_rate": 0.0002,
|
| 4138 |
+
"loss": 0.8518,
|
| 4139 |
+
"step": 4400
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 1.28,
|
| 4143 |
+
"eval_loss": 0.8107084631919861,
|
| 4144 |
+
"eval_runtime": 112.0821,
|
| 4145 |
+
"eval_samples_per_second": 8.922,
|
| 4146 |
+
"eval_steps_per_second": 4.461,
|
| 4147 |
+
"step": 4400
|
| 4148 |
+
},
|
| 4149 |
+
{
|
| 4150 |
+
"epoch": 1.28,
|
| 4151 |
+
"mmlu_eval_accuracy": 0.5054326569126925,
|
| 4152 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
|
| 4153 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 4154 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 4155 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4156 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 4157 |
+
"mmlu_eval_accuracy_college_biology": 0.5625,
|
| 4158 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 4159 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 4160 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 4161 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 4162 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 4163 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 4164 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
|
| 4165 |
+
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
|
| 4166 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 4167 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.3170731707317073,
|
| 4168 |
+
"mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
|
| 4169 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4170 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 4171 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
|
| 4172 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4173 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 4174 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 4175 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 4176 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 4177 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
|
| 4178 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 4179 |
+
"mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
|
| 4180 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 4181 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
|
| 4182 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
|
| 4183 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
|
| 4184 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 4185 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4186 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 4187 |
+
"mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
|
| 4188 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 4189 |
+
"mmlu_eval_accuracy_machine_learning": 0.09090909090909091,
|
| 4190 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4191 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 4192 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 4193 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
|
| 4194 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 4195 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.25,
|
| 4196 |
+
"mmlu_eval_accuracy_nutrition": 0.7272727272727273,
|
| 4197 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 4198 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 4199 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
|
| 4200 |
+
"mmlu_eval_accuracy_professional_law": 0.35294117647058826,
|
| 4201 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 4202 |
+
"mmlu_eval_accuracy_professional_psychology": 0.5072463768115942,
|
| 4203 |
+
"mmlu_eval_accuracy_public_relations": 0.6666666666666666,
|
| 4204 |
+
"mmlu_eval_accuracy_security_studies": 0.4074074074074074,
|
| 4205 |
+
"mmlu_eval_accuracy_sociology": 0.6363636363636364,
|
| 4206 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 4207 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4208 |
+
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 4209 |
+
"mmlu_loss": 1.3057481511532483,
|
| 4210 |
+
"step": 4400
|
| 4211 |
+
},
|
| 4212 |
+
{
|
| 4213 |
+
"epoch": 1.28,
|
| 4214 |
+
"learning_rate": 0.0002,
|
| 4215 |
+
"loss": 0.791,
|
| 4216 |
+
"step": 4410
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 1.29,
|
| 4220 |
+
"learning_rate": 0.0002,
|
| 4221 |
+
"loss": 0.7682,
|
| 4222 |
+
"step": 4420
|
| 4223 |
+
},
|
| 4224 |
+
{
|
| 4225 |
+
"epoch": 1.29,
|
| 4226 |
+
"learning_rate": 0.0002,
|
| 4227 |
+
"loss": 0.7596,
|
| 4228 |
+
"step": 4430
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 1.29,
|
| 4232 |
+
"learning_rate": 0.0002,
|
| 4233 |
+
"loss": 0.9505,
|
| 4234 |
+
"step": 4440
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 1.3,
|
| 4238 |
+
"learning_rate": 0.0002,
|
| 4239 |
+
"loss": 0.8793,
|
| 4240 |
+
"step": 4450
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 1.3,
|
| 4244 |
+
"learning_rate": 0.0002,
|
| 4245 |
+
"loss": 0.8468,
|
| 4246 |
+
"step": 4460
|
| 4247 |
+
},
|
| 4248 |
+
{
|
| 4249 |
+
"epoch": 1.3,
|
| 4250 |
+
"learning_rate": 0.0002,
|
| 4251 |
+
"loss": 0.8713,
|
| 4252 |
+
"step": 4470
|
| 4253 |
+
},
|
| 4254 |
+
{
|
| 4255 |
+
"epoch": 1.3,
|
| 4256 |
+
"learning_rate": 0.0002,
|
| 4257 |
+
"loss": 0.77,
|
| 4258 |
+
"step": 4480
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 1.31,
|
| 4262 |
+
"learning_rate": 0.0002,
|
| 4263 |
+
"loss": 0.7419,
|
| 4264 |
+
"step": 4490
|
| 4265 |
+
},
|
| 4266 |
+
{
|
| 4267 |
+
"epoch": 1.31,
|
| 4268 |
+
"learning_rate": 0.0002,
|
| 4269 |
+
"loss": 0.928,
|
| 4270 |
+
"step": 4500
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"epoch": 1.31,
|
| 4274 |
+
"learning_rate": 0.0002,
|
| 4275 |
+
"loss": 0.8633,
|
| 4276 |
+
"step": 4510
|
| 4277 |
+
},
|
| 4278 |
+
{
|
| 4279 |
+
"epoch": 1.32,
|
| 4280 |
+
"learning_rate": 0.0002,
|
| 4281 |
+
"loss": 0.8111,
|
| 4282 |
+
"step": 4520
|
| 4283 |
+
},
|
| 4284 |
+
{
|
| 4285 |
+
"epoch": 1.32,
|
| 4286 |
+
"learning_rate": 0.0002,
|
| 4287 |
+
"loss": 0.8949,
|
| 4288 |
+
"step": 4530
|
| 4289 |
+
},
|
| 4290 |
+
{
|
| 4291 |
+
"epoch": 1.32,
|
| 4292 |
+
"learning_rate": 0.0002,
|
| 4293 |
+
"loss": 0.8427,
|
| 4294 |
+
"step": 4540
|
| 4295 |
+
},
|
| 4296 |
+
{
|
| 4297 |
+
"epoch": 1.32,
|
| 4298 |
+
"learning_rate": 0.0002,
|
| 4299 |
+
"loss": 0.8319,
|
| 4300 |
+
"step": 4550
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 1.33,
|
| 4304 |
+
"learning_rate": 0.0002,
|
| 4305 |
+
"loss": 0.8146,
|
| 4306 |
+
"step": 4560
|
| 4307 |
+
},
|
| 4308 |
+
{
|
| 4309 |
+
"epoch": 1.33,
|
| 4310 |
+
"learning_rate": 0.0002,
|
| 4311 |
+
"loss": 0.8748,
|
| 4312 |
+
"step": 4570
|
| 4313 |
+
},
|
| 4314 |
+
{
|
| 4315 |
+
"epoch": 1.33,
|
| 4316 |
+
"learning_rate": 0.0002,
|
| 4317 |
+
"loss": 0.8075,
|
| 4318 |
+
"step": 4580
|
| 4319 |
+
},
|
| 4320 |
+
{
|
| 4321 |
+
"epoch": 1.34,
|
| 4322 |
+
"learning_rate": 0.0002,
|
| 4323 |
+
"loss": 0.9163,
|
| 4324 |
+
"step": 4590
|
| 4325 |
+
},
|
| 4326 |
+
{
|
| 4327 |
+
"epoch": 1.34,
|
| 4328 |
+
"learning_rate": 0.0002,
|
| 4329 |
+
"loss": 0.8485,
|
| 4330 |
+
"step": 4600
|
| 4331 |
+
},
|
| 4332 |
+
{
|
| 4333 |
+
"epoch": 1.34,
|
| 4334 |
+
"eval_loss": 0.8110213875770569,
|
| 4335 |
+
"eval_runtime": 112.2695,
|
| 4336 |
+
"eval_samples_per_second": 8.907,
|
| 4337 |
+
"eval_steps_per_second": 4.454,
|
| 4338 |
+
"step": 4600
|
| 4339 |
+
},
|
| 4340 |
+
{
|
| 4341 |
+
"epoch": 1.34,
|
| 4342 |
+
"mmlu_eval_accuracy": 0.5055979366091056,
|
| 4343 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 4344 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 4345 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 4346 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4347 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 4348 |
+
"mmlu_eval_accuracy_college_biology": 0.5,
|
| 4349 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 4350 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 4351 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 4352 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 4353 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 4354 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 4355 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 4356 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4357 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 4358 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
|
| 4359 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 4360 |
+
"mmlu_eval_accuracy_global_facts": 0.6,
|
| 4361 |
+
"mmlu_eval_accuracy_high_school_biology": 0.5,
|
| 4362 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
|
| 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": 0.8636363636363636,
|
| 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": 0.46153846153846156,
|
| 4370 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 4371 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 4372 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
|
| 4373 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
|
| 4374 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
|
| 4375 |
+
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
|
| 4376 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4377 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 4378 |
+
"mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
|
| 4379 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
|
| 4380 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 4381 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4382 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 4383 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 4384 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
|
| 4385 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 4386 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 4387 |
+
"mmlu_eval_accuracy_nutrition": 0.6363636363636364,
|
| 4388 |
+
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
|
| 4389 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 4390 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
|
| 4391 |
+
"mmlu_eval_accuracy_professional_law": 0.35294117647058826,
|
| 4392 |
+
"mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
|
| 4393 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4782608695652174,
|
| 4394 |
+
"mmlu_eval_accuracy_public_relations": 0.6666666666666666,
|
| 4395 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4396 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 4397 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
|
| 4398 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4399 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4400 |
+
"mmlu_loss": 1.349465638749282,
|
| 4401 |
+
"step": 4600
|
| 4402 |
+
},
|
| 4403 |
+
{
|
| 4404 |
+
"epoch": 1.34,
|
| 4405 |
+
"learning_rate": 0.0002,
|
| 4406 |
+
"loss": 0.8501,
|
| 4407 |
+
"step": 4610
|
| 4408 |
+
},
|
| 4409 |
+
{
|
| 4410 |
+
"epoch": 1.35,
|
| 4411 |
+
"learning_rate": 0.0002,
|
| 4412 |
+
"loss": 0.8855,
|
| 4413 |
+
"step": 4620
|
| 4414 |
+
},
|
| 4415 |
+
{
|
| 4416 |
+
"epoch": 1.35,
|
| 4417 |
+
"learning_rate": 0.0002,
|
| 4418 |
+
"loss": 0.8735,
|
| 4419 |
+
"step": 4630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 1.35,
|
| 4423 |
+
"learning_rate": 0.0002,
|
| 4424 |
+
"loss": 0.7924,
|
| 4425 |
+
"step": 4640
|
| 4426 |
+
},
|
| 4427 |
+
{
|
| 4428 |
+
"epoch": 1.35,
|
| 4429 |
+
"learning_rate": 0.0002,
|
| 4430 |
+
"loss": 0.8629,
|
| 4431 |
+
"step": 4650
|
| 4432 |
+
},
|
| 4433 |
+
{
|
| 4434 |
+
"epoch": 1.36,
|
| 4435 |
+
"learning_rate": 0.0002,
|
| 4436 |
+
"loss": 0.9271,
|
| 4437 |
+
"step": 4660
|
| 4438 |
+
},
|
| 4439 |
+
{
|
| 4440 |
+
"epoch": 1.36,
|
| 4441 |
+
"learning_rate": 0.0002,
|
| 4442 |
+
"loss": 0.8608,
|
| 4443 |
+
"step": 4670
|
| 4444 |
+
},
|
| 4445 |
+
{
|
| 4446 |
+
"epoch": 1.36,
|
| 4447 |
+
"learning_rate": 0.0002,
|
| 4448 |
+
"loss": 0.8287,
|
| 4449 |
+
"step": 4680
|
| 4450 |
+
},
|
| 4451 |
+
{
|
| 4452 |
+
"epoch": 1.37,
|
| 4453 |
+
"learning_rate": 0.0002,
|
| 4454 |
+
"loss": 0.8546,
|
| 4455 |
+
"step": 4690
|
| 4456 |
+
},
|
| 4457 |
+
{
|
| 4458 |
+
"epoch": 1.37,
|
| 4459 |
+
"learning_rate": 0.0002,
|
| 4460 |
+
"loss": 0.8212,
|
| 4461 |
+
"step": 4700
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 1.37,
|
| 4465 |
+
"learning_rate": 0.0002,
|
| 4466 |
+
"loss": 0.8831,
|
| 4467 |
+
"step": 4710
|
| 4468 |
+
},
|
| 4469 |
+
{
|
| 4470 |
+
"epoch": 1.37,
|
| 4471 |
+
"learning_rate": 0.0002,
|
| 4472 |
+
"loss": 0.8255,
|
| 4473 |
+
"step": 4720
|
| 4474 |
+
},
|
| 4475 |
+
{
|
| 4476 |
+
"epoch": 1.38,
|
| 4477 |
+
"learning_rate": 0.0002,
|
| 4478 |
+
"loss": 0.8641,
|
| 4479 |
+
"step": 4730
|
| 4480 |
+
},
|
| 4481 |
+
{
|
| 4482 |
+
"epoch": 1.38,
|
| 4483 |
+
"learning_rate": 0.0002,
|
| 4484 |
+
"loss": 0.8731,
|
| 4485 |
+
"step": 4740
|
| 4486 |
+
},
|
| 4487 |
+
{
|
| 4488 |
+
"epoch": 1.38,
|
| 4489 |
+
"learning_rate": 0.0002,
|
| 4490 |
+
"loss": 0.8085,
|
| 4491 |
+
"step": 4750
|
| 4492 |
+
},
|
| 4493 |
+
{
|
| 4494 |
+
"epoch": 1.39,
|
| 4495 |
+
"learning_rate": 0.0002,
|
| 4496 |
+
"loss": 0.9141,
|
| 4497 |
+
"step": 4760
|
| 4498 |
+
},
|
| 4499 |
+
{
|
| 4500 |
+
"epoch": 1.39,
|
| 4501 |
+
"learning_rate": 0.0002,
|
| 4502 |
+
"loss": 0.8443,
|
| 4503 |
+
"step": 4770
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 1.39,
|
| 4507 |
+
"learning_rate": 0.0002,
|
| 4508 |
+
"loss": 0.8085,
|
| 4509 |
+
"step": 4780
|
| 4510 |
+
},
|
| 4511 |
+
{
|
| 4512 |
+
"epoch": 1.39,
|
| 4513 |
+
"learning_rate": 0.0002,
|
| 4514 |
+
"loss": 0.8494,
|
| 4515 |
+
"step": 4790
|
| 4516 |
+
},
|
| 4517 |
+
{
|
| 4518 |
+
"epoch": 1.4,
|
| 4519 |
+
"learning_rate": 0.0002,
|
| 4520 |
+
"loss": 0.84,
|
| 4521 |
+
"step": 4800
|
| 4522 |
+
},
|
| 4523 |
+
{
|
| 4524 |
+
"epoch": 1.4,
|
| 4525 |
+
"eval_loss": 0.8135509490966797,
|
| 4526 |
+
"eval_runtime": 112.3978,
|
| 4527 |
+
"eval_samples_per_second": 8.897,
|
| 4528 |
+
"eval_steps_per_second": 4.448,
|
| 4529 |
+
"step": 4800
|
| 4530 |
+
},
|
| 4531 |
+
{
|
| 4532 |
+
"epoch": 1.4,
|
| 4533 |
+
"mmlu_eval_accuracy": 0.4998311163494409,
|
| 4534 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
|
| 4535 |
+
"mmlu_eval_accuracy_anatomy": 0.7142857142857143,
|
| 4536 |
+
"mmlu_eval_accuracy_astronomy": 0.5625,
|
| 4537 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4538 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 4539 |
+
"mmlu_eval_accuracy_college_biology": 0.5,
|
| 4540 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 4541 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 4542 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 4543 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 4544 |
+
"mmlu_eval_accuracy_college_physics": 0.2727272727272727,
|
| 4545 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 4546 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 4547 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4548 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 4549 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
|
| 4550 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 4551 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4552 |
+
"mmlu_eval_accuracy_high_school_biology": 0.46875,
|
| 4553 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
|
| 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": 0.9090909090909091,
|
| 4557 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 4558 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 4559 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 4560 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 4561 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 4562 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
|
| 4563 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
|
| 4564 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 4565 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
|
| 4566 |
+
"mmlu_eval_accuracy_human_aging": 0.6086956521739131,
|
| 4567 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4568 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 4569 |
+
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
|
| 4570 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 4571 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 4572 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 4573 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 4574 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 4575 |
+
"mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
|
| 4576 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 4577 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.25,
|
| 4578 |
+
"mmlu_eval_accuracy_nutrition": 0.6363636363636364,
|
| 4579 |
+
"mmlu_eval_accuracy_philosophy": 0.47058823529411764,
|
| 4580 |
+
"mmlu_eval_accuracy_prehistory": 0.45714285714285713,
|
| 4581 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
|
| 4582 |
+
"mmlu_eval_accuracy_professional_law": 0.3588235294117647,
|
| 4583 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 4584 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 4585 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4586 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4587 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 4588 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
|
| 4589 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4590 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4591 |
+
"mmlu_loss": 1.3253662804477209,
|
| 4592 |
+
"step": 4800
|
| 4593 |
+
},
|
| 4594 |
+
{
|
| 4595 |
+
"epoch": 1.4,
|
| 4596 |
+
"learning_rate": 0.0002,
|
| 4597 |
+
"loss": 0.842,
|
| 4598 |
+
"step": 4810
|
| 4599 |
+
},
|
| 4600 |
+
{
|
| 4601 |
+
"epoch": 1.4,
|
| 4602 |
+
"learning_rate": 0.0002,
|
| 4603 |
+
"loss": 0.8867,
|
| 4604 |
+
"step": 4820
|
| 4605 |
+
},
|
| 4606 |
+
{
|
| 4607 |
+
"epoch": 1.41,
|
| 4608 |
+
"learning_rate": 0.0002,
|
| 4609 |
+
"loss": 0.7843,
|
| 4610 |
+
"step": 4830
|
| 4611 |
+
},
|
| 4612 |
+
{
|
| 4613 |
+
"epoch": 1.41,
|
| 4614 |
+
"learning_rate": 0.0002,
|
| 4615 |
+
"loss": 0.6769,
|
| 4616 |
+
"step": 4840
|
| 4617 |
+
},
|
| 4618 |
+
{
|
| 4619 |
+
"epoch": 1.41,
|
| 4620 |
+
"learning_rate": 0.0002,
|
| 4621 |
+
"loss": 0.8129,
|
| 4622 |
+
"step": 4850
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 1.42,
|
| 4626 |
+
"learning_rate": 0.0002,
|
| 4627 |
+
"loss": 0.7671,
|
| 4628 |
+
"step": 4860
|
| 4629 |
+
},
|
| 4630 |
+
{
|
| 4631 |
+
"epoch": 1.42,
|
| 4632 |
+
"learning_rate": 0.0002,
|
| 4633 |
+
"loss": 0.7459,
|
| 4634 |
+
"step": 4870
|
| 4635 |
+
},
|
| 4636 |
+
{
|
| 4637 |
+
"epoch": 1.42,
|
| 4638 |
+
"learning_rate": 0.0002,
|
| 4639 |
+
"loss": 0.8878,
|
| 4640 |
+
"step": 4880
|
| 4641 |
+
},
|
| 4642 |
+
{
|
| 4643 |
+
"epoch": 1.42,
|
| 4644 |
+
"learning_rate": 0.0002,
|
| 4645 |
+
"loss": 0.8404,
|
| 4646 |
+
"step": 4890
|
| 4647 |
+
},
|
| 4648 |
+
{
|
| 4649 |
+
"epoch": 1.43,
|
| 4650 |
+
"learning_rate": 0.0002,
|
| 4651 |
+
"loss": 0.7844,
|
| 4652 |
+
"step": 4900
|
| 4653 |
+
},
|
| 4654 |
+
{
|
| 4655 |
+
"epoch": 1.43,
|
| 4656 |
+
"learning_rate": 0.0002,
|
| 4657 |
+
"loss": 0.8549,
|
| 4658 |
+
"step": 4910
|
| 4659 |
+
},
|
| 4660 |
+
{
|
| 4661 |
+
"epoch": 1.43,
|
| 4662 |
+
"learning_rate": 0.0002,
|
| 4663 |
+
"loss": 0.9099,
|
| 4664 |
+
"step": 4920
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 1.44,
|
| 4668 |
+
"learning_rate": 0.0002,
|
| 4669 |
+
"loss": 0.833,
|
| 4670 |
+
"step": 4930
|
| 4671 |
+
},
|
| 4672 |
+
{
|
| 4673 |
+
"epoch": 1.44,
|
| 4674 |
+
"learning_rate": 0.0002,
|
| 4675 |
+
"loss": 0.9135,
|
| 4676 |
+
"step": 4940
|
| 4677 |
+
},
|
| 4678 |
+
{
|
| 4679 |
+
"epoch": 1.44,
|
| 4680 |
+
"learning_rate": 0.0002,
|
| 4681 |
+
"loss": 0.7573,
|
| 4682 |
+
"step": 4950
|
| 4683 |
+
},
|
| 4684 |
+
{
|
| 4685 |
+
"epoch": 1.44,
|
| 4686 |
+
"learning_rate": 0.0002,
|
| 4687 |
+
"loss": 0.9444,
|
| 4688 |
+
"step": 4960
|
| 4689 |
+
},
|
| 4690 |
+
{
|
| 4691 |
+
"epoch": 1.45,
|
| 4692 |
+
"learning_rate": 0.0002,
|
| 4693 |
+
"loss": 0.8831,
|
| 4694 |
+
"step": 4970
|
| 4695 |
+
},
|
| 4696 |
+
{
|
| 4697 |
+
"epoch": 1.45,
|
| 4698 |
+
"learning_rate": 0.0002,
|
| 4699 |
+
"loss": 0.8842,
|
| 4700 |
+
"step": 4980
|
| 4701 |
+
},
|
| 4702 |
+
{
|
| 4703 |
+
"epoch": 1.45,
|
| 4704 |
+
"learning_rate": 0.0002,
|
| 4705 |
+
"loss": 0.8501,
|
| 4706 |
+
"step": 4990
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 1.46,
|
| 4710 |
+
"learning_rate": 0.0002,
|
| 4711 |
+
"loss": 0.8653,
|
| 4712 |
+
"step": 5000
|
| 4713 |
+
},
|
| 4714 |
+
{
|
| 4715 |
+
"epoch": 1.46,
|
| 4716 |
+
"eval_loss": 0.8124507665634155,
|
| 4717 |
+
"eval_runtime": 112.4014,
|
| 4718 |
+
"eval_samples_per_second": 8.897,
|
| 4719 |
+
"eval_steps_per_second": 4.448,
|
| 4720 |
+
"step": 5000
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 1.46,
|
| 4724 |
+
"mmlu_eval_accuracy": 0.49611962914206076,
|
| 4725 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 4726 |
+
"mmlu_eval_accuracy_anatomy": 0.6428571428571429,
|
| 4727 |
+
"mmlu_eval_accuracy_astronomy": 0.5625,
|
| 4728 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4729 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
|
| 4730 |
+
"mmlu_eval_accuracy_college_biology": 0.5625,
|
| 4731 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 4732 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 4733 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 4734 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 4735 |
+
"mmlu_eval_accuracy_college_physics": 0.2727272727272727,
|
| 4736 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 4737 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.5,
|
| 4738 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 4739 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 4740 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
|
| 4741 |
+
"mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
|
| 4742 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4743 |
+
"mmlu_eval_accuracy_high_school_biology": 0.46875,
|
| 4744 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.2727272727272727,
|
| 4745 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4746 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 4747 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 4748 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
|
| 4749 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907,
|
| 4750 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 4751 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.5,
|
| 4752 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 4753 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 4754 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
|
| 4755 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 4756 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
|
| 4757 |
+
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
|
| 4758 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4759 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 4760 |
+
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
|
| 4761 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 4762 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 4763 |
+
"mmlu_eval_accuracy_management": 0.5454545454545454,
|
| 4764 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 4765 |
+
"mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
|
| 4766 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 4767 |
+
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
|
| 4768 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.25,
|
| 4769 |
+
"mmlu_eval_accuracy_nutrition": 0.7272727272727273,
|
| 4770 |
+
"mmlu_eval_accuracy_philosophy": 0.47058823529411764,
|
| 4771 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 4772 |
+
"mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
|
| 4773 |
+
"mmlu_eval_accuracy_professional_law": 0.34705882352941175,
|
| 4774 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
|
| 4775 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 4776 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 4777 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 4778 |
+
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
|
| 4779 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
|
| 4780 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4781 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4782 |
+
"mmlu_loss": 1.3226010548072447,
|
| 4783 |
+
"step": 5000
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 1.46,
|
| 4787 |
+
"learning_rate": 0.0002,
|
| 4788 |
+
"loss": 0.9205,
|
| 4789 |
+
"step": 5010
|
| 4790 |
+
},
|
| 4791 |
+
{
|
| 4792 |
+
"epoch": 1.46,
|
| 4793 |
+
"learning_rate": 0.0002,
|
| 4794 |
+
"loss": 0.7774,
|
| 4795 |
+
"step": 5020
|
| 4796 |
+
},
|
| 4797 |
+
{
|
| 4798 |
+
"epoch": 1.46,
|
| 4799 |
+
"learning_rate": 0.0002,
|
| 4800 |
+
"loss": 0.7829,
|
| 4801 |
+
"step": 5030
|
| 4802 |
+
},
|
| 4803 |
+
{
|
| 4804 |
+
"epoch": 1.47,
|
| 4805 |
+
"learning_rate": 0.0002,
|
| 4806 |
+
"loss": 0.8501,
|
| 4807 |
+
"step": 5040
|
| 4808 |
+
},
|
| 4809 |
+
{
|
| 4810 |
+
"epoch": 1.47,
|
| 4811 |
+
"learning_rate": 0.0002,
|
| 4812 |
+
"loss": 0.8622,
|
| 4813 |
+
"step": 5050
|
| 4814 |
+
},
|
| 4815 |
+
{
|
| 4816 |
+
"epoch": 1.47,
|
| 4817 |
+
"learning_rate": 0.0002,
|
| 4818 |
+
"loss": 0.7971,
|
| 4819 |
+
"step": 5060
|
| 4820 |
+
},
|
| 4821 |
+
{
|
| 4822 |
+
"epoch": 1.48,
|
| 4823 |
+
"learning_rate": 0.0002,
|
| 4824 |
+
"loss": 0.8705,
|
| 4825 |
+
"step": 5070
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 1.48,
|
| 4829 |
+
"learning_rate": 0.0002,
|
| 4830 |
+
"loss": 0.8234,
|
| 4831 |
+
"step": 5080
|
| 4832 |
+
},
|
| 4833 |
+
{
|
| 4834 |
+
"epoch": 1.48,
|
| 4835 |
+
"learning_rate": 0.0002,
|
| 4836 |
+
"loss": 0.8872,
|
| 4837 |
+
"step": 5090
|
| 4838 |
+
},
|
| 4839 |
+
{
|
| 4840 |
+
"epoch": 1.48,
|
| 4841 |
+
"learning_rate": 0.0002,
|
| 4842 |
+
"loss": 0.8901,
|
| 4843 |
+
"step": 5100
|
| 4844 |
+
},
|
| 4845 |
+
{
|
| 4846 |
+
"epoch": 1.49,
|
| 4847 |
+
"learning_rate": 0.0002,
|
| 4848 |
+
"loss": 0.8731,
|
| 4849 |
+
"step": 5110
|
| 4850 |
+
},
|
| 4851 |
+
{
|
| 4852 |
+
"epoch": 1.49,
|
| 4853 |
+
"learning_rate": 0.0002,
|
| 4854 |
+
"loss": 0.8599,
|
| 4855 |
+
"step": 5120
|
| 4856 |
+
},
|
| 4857 |
+
{
|
| 4858 |
+
"epoch": 1.49,
|
| 4859 |
+
"learning_rate": 0.0002,
|
| 4860 |
+
"loss": 0.8223,
|
| 4861 |
+
"step": 5130
|
| 4862 |
+
},
|
| 4863 |
+
{
|
| 4864 |
+
"epoch": 1.5,
|
| 4865 |
+
"learning_rate": 0.0002,
|
| 4866 |
+
"loss": 0.8619,
|
| 4867 |
+
"step": 5140
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 1.5,
|
| 4871 |
+
"learning_rate": 0.0002,
|
| 4872 |
+
"loss": 0.8746,
|
| 4873 |
+
"step": 5150
|
| 4874 |
+
},
|
| 4875 |
+
{
|
| 4876 |
+
"epoch": 1.5,
|
| 4877 |
+
"learning_rate": 0.0002,
|
| 4878 |
+
"loss": 0.8987,
|
| 4879 |
+
"step": 5160
|
| 4880 |
+
},
|
| 4881 |
+
{
|
| 4882 |
+
"epoch": 1.51,
|
| 4883 |
+
"learning_rate": 0.0002,
|
| 4884 |
+
"loss": 0.7756,
|
| 4885 |
+
"step": 5170
|
| 4886 |
+
},
|
| 4887 |
+
{
|
| 4888 |
+
"epoch": 1.51,
|
| 4889 |
+
"learning_rate": 0.0002,
|
| 4890 |
+
"loss": 0.8282,
|
| 4891 |
+
"step": 5180
|
| 4892 |
+
},
|
| 4893 |
+
{
|
| 4894 |
+
"epoch": 1.51,
|
| 4895 |
+
"learning_rate": 0.0002,
|
| 4896 |
+
"loss": 0.8271,
|
| 4897 |
+
"step": 5190
|
| 4898 |
+
},
|
| 4899 |
+
{
|
| 4900 |
+
"epoch": 1.51,
|
| 4901 |
+
"learning_rate": 0.0002,
|
| 4902 |
+
"loss": 0.8014,
|
| 4903 |
+
"step": 5200
|
| 4904 |
+
},
|
| 4905 |
+
{
|
| 4906 |
+
"epoch": 1.51,
|
| 4907 |
+
"eval_loss": 0.8068380355834961,
|
| 4908 |
+
"eval_runtime": 112.0243,
|
| 4909 |
+
"eval_samples_per_second": 8.927,
|
| 4910 |
+
"eval_steps_per_second": 4.463,
|
| 4911 |
+
"step": 5200
|
| 4912 |
+
},
|
| 4913 |
+
{
|
| 4914 |
+
"epoch": 1.51,
|
| 4915 |
+
"mmlu_eval_accuracy": 0.494753529535565,
|
| 4916 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 4917 |
+
"mmlu_eval_accuracy_anatomy": 0.7142857142857143,
|
| 4918 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 4919 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 4920 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
|
| 4921 |
+
"mmlu_eval_accuracy_college_biology": 0.5,
|
| 4922 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 4923 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 4924 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 4925 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 4926 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 4927 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 4928 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 4929 |
+
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
|
| 4930 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 4931 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
|
| 4932 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 4933 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 4934 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 4935 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.22727272727272727,
|
| 4936 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 4937 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
|
| 4938 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 4939 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 4940 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 4941 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 4942 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.5,
|
| 4943 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 4944 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8833333333333333,
|
| 4945 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
|
| 4946 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 4947 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
|
| 4948 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 4949 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 4950 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 4951 |
+
"mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
|
| 4952 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 4953 |
+
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
|
| 4954 |
+
"mmlu_eval_accuracy_management": 0.5454545454545454,
|
| 4955 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 4956 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 4957 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 4958 |
+
"mmlu_eval_accuracy_moral_disputes": 0.39473684210526316,
|
| 4959 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.29,
|
| 4960 |
+
"mmlu_eval_accuracy_nutrition": 0.6363636363636364,
|
| 4961 |
+
"mmlu_eval_accuracy_philosophy": 0.47058823529411764,
|
| 4962 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 4963 |
+
"mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
|
| 4964 |
+
"mmlu_eval_accuracy_professional_law": 0.35294117647058826,
|
| 4965 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 4966 |
+
"mmlu_eval_accuracy_professional_psychology": 0.463768115942029,
|
| 4967 |
+
"mmlu_eval_accuracy_public_relations": 0.6666666666666666,
|
| 4968 |
+
"mmlu_eval_accuracy_security_studies": 0.4444444444444444,
|
| 4969 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 4970 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
|
| 4971 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 4972 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 4973 |
+
"mmlu_loss": 1.363197118822028,
|
| 4974 |
+
"step": 5200
|
| 4975 |
+
},
|
| 4976 |
+
{
|
| 4977 |
+
"epoch": 1.52,
|
| 4978 |
+
"learning_rate": 0.0002,
|
| 4979 |
+
"loss": 0.8763,
|
| 4980 |
+
"step": 5210
|
| 4981 |
+
},
|
| 4982 |
+
{
|
| 4983 |
+
"epoch": 1.52,
|
| 4984 |
+
"learning_rate": 0.0002,
|
| 4985 |
+
"loss": 0.8323,
|
| 4986 |
+
"step": 5220
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 1.52,
|
| 4990 |
+
"learning_rate": 0.0002,
|
| 4991 |
+
"loss": 0.8509,
|
| 4992 |
+
"step": 5230
|
| 4993 |
+
},
|
| 4994 |
+
{
|
| 4995 |
+
"epoch": 1.53,
|
| 4996 |
+
"learning_rate": 0.0002,
|
| 4997 |
+
"loss": 0.8351,
|
| 4998 |
+
"step": 5240
|
| 4999 |
+
},
|
| 5000 |
+
{
|
| 5001 |
+
"epoch": 1.53,
|
| 5002 |
+
"learning_rate": 0.0002,
|
| 5003 |
+
"loss": 0.87,
|
| 5004 |
+
"step": 5250
|
| 5005 |
+
},
|
| 5006 |
+
{
|
| 5007 |
+
"epoch": 1.53,
|
| 5008 |
+
"learning_rate": 0.0002,
|
| 5009 |
+
"loss": 0.7947,
|
| 5010 |
+
"step": 5260
|
| 5011 |
+
},
|
| 5012 |
+
{
|
| 5013 |
+
"epoch": 1.53,
|
| 5014 |
+
"learning_rate": 0.0002,
|
| 5015 |
+
"loss": 0.7707,
|
| 5016 |
+
"step": 5270
|
| 5017 |
+
},
|
| 5018 |
+
{
|
| 5019 |
+
"epoch": 1.54,
|
| 5020 |
+
"learning_rate": 0.0002,
|
| 5021 |
+
"loss": 0.8457,
|
| 5022 |
+
"step": 5280
|
| 5023 |
+
},
|
| 5024 |
+
{
|
| 5025 |
+
"epoch": 1.54,
|
| 5026 |
+
"learning_rate": 0.0002,
|
| 5027 |
+
"loss": 0.8234,
|
| 5028 |
+
"step": 5290
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 1.54,
|
| 5032 |
+
"learning_rate": 0.0002,
|
| 5033 |
+
"loss": 0.8129,
|
| 5034 |
+
"step": 5300
|
| 5035 |
+
},
|
| 5036 |
+
{
|
| 5037 |
+
"epoch": 1.55,
|
| 5038 |
+
"learning_rate": 0.0002,
|
| 5039 |
+
"loss": 0.8793,
|
| 5040 |
+
"step": 5310
|
| 5041 |
+
},
|
| 5042 |
+
{
|
| 5043 |
+
"epoch": 1.55,
|
| 5044 |
+
"learning_rate": 0.0002,
|
| 5045 |
+
"loss": 0.8009,
|
| 5046 |
+
"step": 5320
|
| 5047 |
+
},
|
| 5048 |
+
{
|
| 5049 |
+
"epoch": 1.55,
|
| 5050 |
+
"learning_rate": 0.0002,
|
| 5051 |
+
"loss": 0.9667,
|
| 5052 |
+
"step": 5330
|
| 5053 |
+
},
|
| 5054 |
+
{
|
| 5055 |
+
"epoch": 1.55,
|
| 5056 |
+
"learning_rate": 0.0002,
|
| 5057 |
+
"loss": 0.7834,
|
| 5058 |
+
"step": 5340
|
| 5059 |
+
},
|
| 5060 |
+
{
|
| 5061 |
+
"epoch": 1.56,
|
| 5062 |
+
"learning_rate": 0.0002,
|
| 5063 |
+
"loss": 0.7376,
|
| 5064 |
+
"step": 5350
|
| 5065 |
+
},
|
| 5066 |
+
{
|
| 5067 |
+
"epoch": 1.56,
|
| 5068 |
+
"learning_rate": 0.0002,
|
| 5069 |
+
"loss": 0.795,
|
| 5070 |
+
"step": 5360
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 1.56,
|
| 5074 |
+
"learning_rate": 0.0002,
|
| 5075 |
+
"loss": 0.8667,
|
| 5076 |
+
"step": 5370
|
| 5077 |
+
},
|
| 5078 |
+
{
|
| 5079 |
+
"epoch": 1.57,
|
| 5080 |
+
"learning_rate": 0.0002,
|
| 5081 |
+
"loss": 0.8208,
|
| 5082 |
+
"step": 5380
|
| 5083 |
+
},
|
| 5084 |
+
{
|
| 5085 |
+
"epoch": 1.57,
|
| 5086 |
+
"learning_rate": 0.0002,
|
| 5087 |
+
"loss": 0.88,
|
| 5088 |
+
"step": 5390
|
| 5089 |
+
},
|
| 5090 |
+
{
|
| 5091 |
+
"epoch": 1.57,
|
| 5092 |
+
"learning_rate": 0.0002,
|
| 5093 |
+
"loss": 0.8099,
|
| 5094 |
+
"step": 5400
|
| 5095 |
+
},
|
| 5096 |
+
{
|
| 5097 |
+
"epoch": 1.57,
|
| 5098 |
+
"eval_loss": 0.8074617981910706,
|
| 5099 |
+
"eval_runtime": 112.4792,
|
| 5100 |
+
"eval_samples_per_second": 8.891,
|
| 5101 |
+
"eval_steps_per_second": 4.445,
|
| 5102 |
+
"step": 5400
|
| 5103 |
+
},
|
| 5104 |
+
{
|
| 5105 |
+
"epoch": 1.57,
|
| 5106 |
+
"mmlu_eval_accuracy": 0.49947322987276316,
|
| 5107 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 5108 |
+
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
|
| 5109 |
+
"mmlu_eval_accuracy_astronomy": 0.5625,
|
| 5110 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 5111 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 5112 |
+
"mmlu_eval_accuracy_college_biology": 0.5,
|
| 5113 |
+
"mmlu_eval_accuracy_college_chemistry": 0.375,
|
| 5114 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 5115 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 5116 |
+
"mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
|
| 5117 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 5118 |
+
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
|
| 5119 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
|
| 5120 |
+
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
|
| 5121 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.25,
|
| 5122 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
|
| 5123 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 5124 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 5125 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 5126 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.2727272727272727,
|
| 5127 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 5128 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
|
| 5129 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 5130 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
|
| 5131 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907,
|
| 5132 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
|
| 5133 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
|
| 5134 |
+
"mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
|
| 5135 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8833333333333333,
|
| 5136 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
|
| 5137 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 5138 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
|
| 5139 |
+
"mmlu_eval_accuracy_human_aging": 0.6086956521739131,
|
| 5140 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 5141 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 5142 |
+
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
|
| 5143 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 5144 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 5145 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 5146 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 5147 |
+
"mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
|
| 5148 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 5149 |
+
"mmlu_eval_accuracy_moral_disputes": 0.39473684210526316,
|
| 5150 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 5151 |
+
"mmlu_eval_accuracy_nutrition": 0.696969696969697,
|
| 5152 |
+
"mmlu_eval_accuracy_philosophy": 0.5,
|
| 5153 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 5154 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
|
| 5155 |
+
"mmlu_eval_accuracy_professional_law": 0.3411764705882353,
|
| 5156 |
+
"mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
|
| 5157 |
+
"mmlu_eval_accuracy_professional_psychology": 0.43478260869565216,
|
| 5158 |
+
"mmlu_eval_accuracy_public_relations": 0.5,
|
| 5159 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 5160 |
+
"mmlu_eval_accuracy_sociology": 0.7272727272727273,
|
| 5161 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 5162 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 5163 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 5164 |
+
"mmlu_loss": 1.2389598589964386,
|
| 5165 |
+
"step": 5400
|
| 5166 |
+
},
|
| 5167 |
+
{
|
| 5168 |
+
"epoch": 1.58,
|
| 5169 |
+
"learning_rate": 0.0002,
|
| 5170 |
+
"loss": 0.8101,
|
| 5171 |
+
"step": 5410
|
| 5172 |
+
},
|
| 5173 |
+
{
|
| 5174 |
+
"epoch": 1.58,
|
| 5175 |
+
"learning_rate": 0.0002,
|
| 5176 |
+
"loss": 0.8081,
|
| 5177 |
+
"step": 5420
|
| 5178 |
+
},
|
| 5179 |
+
{
|
| 5180 |
+
"epoch": 1.58,
|
| 5181 |
+
"learning_rate": 0.0002,
|
| 5182 |
+
"loss": 0.8398,
|
| 5183 |
+
"step": 5430
|
| 5184 |
+
},
|
| 5185 |
+
{
|
| 5186 |
+
"epoch": 1.58,
|
| 5187 |
+
"learning_rate": 0.0002,
|
| 5188 |
+
"loss": 0.8397,
|
| 5189 |
+
"step": 5440
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 1.59,
|
| 5193 |
+
"learning_rate": 0.0002,
|
| 5194 |
+
"loss": 0.8135,
|
| 5195 |
+
"step": 5450
|
| 5196 |
+
},
|
| 5197 |
+
{
|
| 5198 |
+
"epoch": 1.59,
|
| 5199 |
+
"learning_rate": 0.0002,
|
| 5200 |
+
"loss": 0.8382,
|
| 5201 |
+
"step": 5460
|
| 5202 |
+
},
|
| 5203 |
+
{
|
| 5204 |
+
"epoch": 1.59,
|
| 5205 |
+
"learning_rate": 0.0002,
|
| 5206 |
+
"loss": 0.8801,
|
| 5207 |
+
"step": 5470
|
| 5208 |
+
},
|
| 5209 |
+
{
|
| 5210 |
+
"epoch": 1.6,
|
| 5211 |
+
"learning_rate": 0.0002,
|
| 5212 |
+
"loss": 0.8888,
|
| 5213 |
+
"step": 5480
|
| 5214 |
+
},
|
| 5215 |
+
{
|
| 5216 |
+
"epoch": 1.6,
|
| 5217 |
+
"learning_rate": 0.0002,
|
| 5218 |
+
"loss": 0.8372,
|
| 5219 |
+
"step": 5490
|
| 5220 |
+
},
|
| 5221 |
+
{
|
| 5222 |
+
"epoch": 1.6,
|
| 5223 |
+
"learning_rate": 0.0002,
|
| 5224 |
+
"loss": 0.8744,
|
| 5225 |
+
"step": 5500
|
| 5226 |
+
},
|
| 5227 |
+
{
|
| 5228 |
+
"epoch": 1.6,
|
| 5229 |
+
"learning_rate": 0.0002,
|
| 5230 |
+
"loss": 0.8068,
|
| 5231 |
+
"step": 5510
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 1.61,
|
| 5235 |
+
"learning_rate": 0.0002,
|
| 5236 |
+
"loss": 0.8402,
|
| 5237 |
+
"step": 5520
|
| 5238 |
+
},
|
| 5239 |
+
{
|
| 5240 |
+
"epoch": 1.61,
|
| 5241 |
+
"learning_rate": 0.0002,
|
| 5242 |
+
"loss": 0.8665,
|
| 5243 |
+
"step": 5530
|
| 5244 |
+
},
|
| 5245 |
+
{
|
| 5246 |
+
"epoch": 1.61,
|
| 5247 |
+
"learning_rate": 0.0002,
|
| 5248 |
+
"loss": 0.8625,
|
| 5249 |
+
"step": 5540
|
| 5250 |
+
},
|
| 5251 |
+
{
|
| 5252 |
+
"epoch": 1.62,
|
| 5253 |
+
"learning_rate": 0.0002,
|
| 5254 |
+
"loss": 0.7599,
|
| 5255 |
+
"step": 5550
|
| 5256 |
+
},
|
| 5257 |
+
{
|
| 5258 |
+
"epoch": 1.62,
|
| 5259 |
+
"learning_rate": 0.0002,
|
| 5260 |
+
"loss": 0.8044,
|
| 5261 |
+
"step": 5560
|
| 5262 |
+
},
|
| 5263 |
+
{
|
| 5264 |
+
"epoch": 1.62,
|
| 5265 |
+
"learning_rate": 0.0002,
|
| 5266 |
+
"loss": 0.8461,
|
| 5267 |
+
"step": 5570
|
| 5268 |
+
},
|
| 5269 |
+
{
|
| 5270 |
+
"epoch": 1.62,
|
| 5271 |
+
"learning_rate": 0.0002,
|
| 5272 |
+
"loss": 0.8375,
|
| 5273 |
+
"step": 5580
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 1.63,
|
| 5277 |
+
"learning_rate": 0.0002,
|
| 5278 |
+
"loss": 0.8686,
|
| 5279 |
+
"step": 5590
|
| 5280 |
+
},
|
| 5281 |
+
{
|
| 5282 |
+
"epoch": 1.63,
|
| 5283 |
+
"learning_rate": 0.0002,
|
| 5284 |
+
"loss": 0.7484,
|
| 5285 |
+
"step": 5600
|
| 5286 |
+
},
|
| 5287 |
+
{
|
| 5288 |
+
"epoch": 1.63,
|
| 5289 |
+
"eval_loss": 0.8066723346710205,
|
| 5290 |
+
"eval_runtime": 113.3502,
|
| 5291 |
+
"eval_samples_per_second": 8.822,
|
| 5292 |
+
"eval_steps_per_second": 4.411,
|
| 5293 |
+
"step": 5600
|
| 5294 |
+
},
|
| 5295 |
+
{
|
| 5296 |
+
"epoch": 1.63,
|
| 5297 |
+
"mmlu_eval_accuracy": 0.5038582219158022,
|
| 5298 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 5299 |
+
"mmlu_eval_accuracy_anatomy": 0.7142857142857143,
|
| 5300 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 5301 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 5302 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
|
| 5303 |
+
"mmlu_eval_accuracy_college_biology": 0.4375,
|
| 5304 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 5305 |
+
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
|
| 5306 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 5307 |
+
"mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
|
| 5308 |
+
"mmlu_eval_accuracy_college_physics": 0.2727272727272727,
|
| 5309 |
+
"mmlu_eval_accuracy_computer_security": 0.5454545454545454,
|
| 5310 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
|
| 5311 |
+
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
|
| 5312 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 5313 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.24390243902439024,
|
| 5314 |
+
"mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
|
| 5315 |
+
"mmlu_eval_accuracy_global_facts": 0.6,
|
| 5316 |
+
"mmlu_eval_accuracy_high_school_biology": 0.40625,
|
| 5317 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
|
| 5318 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 5319 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
|
| 5320 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 5321 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
|
| 5322 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 5323 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 5324 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
|
| 5325 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 5326 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.85,
|
| 5327 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
|
| 5328 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.5909090909090909,
|
| 5329 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
|
| 5330 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 5331 |
+
"mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
|
| 5332 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 5333 |
+
"mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
|
| 5334 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 5335 |
+
"mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
|
| 5336 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 5337 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 5338 |
+
"mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
|
| 5339 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 5340 |
+
"mmlu_eval_accuracy_moral_disputes": 0.4473684210526316,
|
| 5341 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.26,
|
| 5342 |
+
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
|
| 5343 |
+
"mmlu_eval_accuracy_philosophy": 0.4411764705882353,
|
| 5344 |
+
"mmlu_eval_accuracy_prehistory": 0.5142857142857142,
|
| 5345 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
|
| 5346 |
+
"mmlu_eval_accuracy_professional_law": 0.3176470588235294,
|
| 5347 |
+
"mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
|
| 5348 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
|
| 5349 |
+
"mmlu_eval_accuracy_public_relations": 0.6666666666666666,
|
| 5350 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 5351 |
+
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
|
| 5352 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 5353 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 5354 |
+
"mmlu_eval_accuracy_world_religions": 0.6842105263157895,
|
| 5355 |
+
"mmlu_loss": 1.2091298465551346,
|
| 5356 |
+
"step": 5600
|
| 5357 |
+
},
|
| 5358 |
+
{
|
| 5359 |
+
"epoch": 1.63,
|
| 5360 |
+
"learning_rate": 0.0002,
|
| 5361 |
+
"loss": 0.8971,
|
| 5362 |
+
"step": 5610
|
| 5363 |
+
},
|
| 5364 |
+
{
|
| 5365 |
+
"epoch": 1.64,
|
| 5366 |
+
"learning_rate": 0.0002,
|
| 5367 |
+
"loss": 0.8583,
|
| 5368 |
+
"step": 5620
|
| 5369 |
+
},
|
| 5370 |
+
{
|
| 5371 |
+
"epoch": 1.64,
|
| 5372 |
+
"learning_rate": 0.0002,
|
| 5373 |
+
"loss": 0.8893,
|
| 5374 |
+
"step": 5630
|
| 5375 |
+
},
|
| 5376 |
+
{
|
| 5377 |
+
"epoch": 1.64,
|
| 5378 |
+
"learning_rate": 0.0002,
|
| 5379 |
+
"loss": 0.8737,
|
| 5380 |
+
"step": 5640
|
| 5381 |
+
},
|
| 5382 |
+
{
|
| 5383 |
+
"epoch": 1.65,
|
| 5384 |
+
"learning_rate": 0.0002,
|
| 5385 |
+
"loss": 0.8117,
|
| 5386 |
+
"step": 5650
|
| 5387 |
+
},
|
| 5388 |
+
{
|
| 5389 |
+
"epoch": 1.65,
|
| 5390 |
+
"learning_rate": 0.0002,
|
| 5391 |
+
"loss": 0.8022,
|
| 5392 |
+
"step": 5660
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 1.65,
|
| 5396 |
+
"learning_rate": 0.0002,
|
| 5397 |
+
"loss": 0.8379,
|
| 5398 |
+
"step": 5670
|
| 5399 |
+
},
|
| 5400 |
+
{
|
| 5401 |
+
"epoch": 1.65,
|
| 5402 |
+
"learning_rate": 0.0002,
|
| 5403 |
+
"loss": 0.8887,
|
| 5404 |
+
"step": 5680
|
| 5405 |
+
},
|
| 5406 |
+
{
|
| 5407 |
+
"epoch": 1.66,
|
| 5408 |
+
"learning_rate": 0.0002,
|
| 5409 |
+
"loss": 0.9023,
|
| 5410 |
+
"step": 5690
|
| 5411 |
+
},
|
| 5412 |
+
{
|
| 5413 |
+
"epoch": 1.66,
|
| 5414 |
+
"learning_rate": 0.0002,
|
| 5415 |
+
"loss": 0.8726,
|
| 5416 |
+
"step": 5700
|
| 5417 |
+
},
|
| 5418 |
+
{
|
| 5419 |
+
"epoch": 1.66,
|
| 5420 |
+
"learning_rate": 0.0002,
|
| 5421 |
+
"loss": 0.7458,
|
| 5422 |
+
"step": 5710
|
| 5423 |
+
},
|
| 5424 |
+
{
|
| 5425 |
+
"epoch": 1.67,
|
| 5426 |
+
"learning_rate": 0.0002,
|
| 5427 |
+
"loss": 0.8,
|
| 5428 |
+
"step": 5720
|
| 5429 |
+
},
|
| 5430 |
+
{
|
| 5431 |
+
"epoch": 1.67,
|
| 5432 |
+
"learning_rate": 0.0002,
|
| 5433 |
+
"loss": 0.8178,
|
| 5434 |
+
"step": 5730
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 1.67,
|
| 5438 |
+
"learning_rate": 0.0002,
|
| 5439 |
+
"loss": 0.8373,
|
| 5440 |
+
"step": 5740
|
| 5441 |
+
},
|
| 5442 |
+
{
|
| 5443 |
+
"epoch": 1.67,
|
| 5444 |
+
"learning_rate": 0.0002,
|
| 5445 |
+
"loss": 0.8715,
|
| 5446 |
+
"step": 5750
|
| 5447 |
+
},
|
| 5448 |
+
{
|
| 5449 |
+
"epoch": 1.68,
|
| 5450 |
+
"learning_rate": 0.0002,
|
| 5451 |
+
"loss": 0.8141,
|
| 5452 |
+
"step": 5760
|
| 5453 |
+
},
|
| 5454 |
+
{
|
| 5455 |
+
"epoch": 1.68,
|
| 5456 |
+
"learning_rate": 0.0002,
|
| 5457 |
+
"loss": 0.821,
|
| 5458 |
+
"step": 5770
|
| 5459 |
+
},
|
| 5460 |
+
{
|
| 5461 |
+
"epoch": 1.68,
|
| 5462 |
+
"learning_rate": 0.0002,
|
| 5463 |
+
"loss": 0.8326,
|
| 5464 |
+
"step": 5780
|
| 5465 |
+
},
|
| 5466 |
+
{
|
| 5467 |
+
"epoch": 1.69,
|
| 5468 |
+
"learning_rate": 0.0002,
|
| 5469 |
+
"loss": 0.9078,
|
| 5470 |
+
"step": 5790
|
| 5471 |
+
},
|
| 5472 |
+
{
|
| 5473 |
+
"epoch": 1.69,
|
| 5474 |
+
"learning_rate": 0.0002,
|
| 5475 |
+
"loss": 0.8643,
|
| 5476 |
+
"step": 5800
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 1.69,
|
| 5480 |
+
"eval_loss": 0.8042177557945251,
|
| 5481 |
+
"eval_runtime": 113.1516,
|
| 5482 |
+
"eval_samples_per_second": 8.838,
|
| 5483 |
+
"eval_steps_per_second": 4.419,
|
| 5484 |
+
"step": 5800
|
| 5485 |
+
},
|
| 5486 |
+
{
|
| 5487 |
+
"epoch": 1.69,
|
| 5488 |
+
"mmlu_eval_accuracy": 0.48645870209186454,
|
| 5489 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 5490 |
+
"mmlu_eval_accuracy_anatomy": 0.7142857142857143,
|
| 5491 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 5492 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 5493 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 5494 |
+
"mmlu_eval_accuracy_college_biology": 0.5625,
|
| 5495 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 5496 |
+
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
|
| 5497 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 5498 |
+
"mmlu_eval_accuracy_college_medicine": 0.36363636363636365,
|
| 5499 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 5500 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 5501 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
|
| 5502 |
+
"mmlu_eval_accuracy_econometrics": 0.25,
|
| 5503 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.1875,
|
| 5504 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
|
| 5505 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 5506 |
+
"mmlu_eval_accuracy_global_facts": 0.5,
|
| 5507 |
+
"mmlu_eval_accuracy_high_school_biology": 0.375,
|
| 5508 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.22727272727272727,
|
| 5509 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 5510 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
|
| 5511 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
|
| 5512 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
|
| 5513 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3953488372093023,
|
| 5514 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 5515 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.46153846153846156,
|
| 5516 |
+
"mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
|
| 5517 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
|
| 5518 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
|
| 5519 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 5520 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
|
| 5521 |
+
"mmlu_eval_accuracy_human_aging": 0.6086956521739131,
|
| 5522 |
+
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
|
| 5523 |
+
"mmlu_eval_accuracy_international_law": 0.9230769230769231,
|
| 5524 |
+
"mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
|
| 5525 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
|
| 5526 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 5527 |
+
"mmlu_eval_accuracy_management": 0.5454545454545454,
|
| 5528 |
+
"mmlu_eval_accuracy_marketing": 0.84,
|
| 5529 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 5530 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
|
| 5531 |
+
"mmlu_eval_accuracy_moral_disputes": 0.47368421052631576,
|
| 5532 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 5533 |
+
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
|
| 5534 |
+
"mmlu_eval_accuracy_philosophy": 0.47058823529411764,
|
| 5535 |
+
"mmlu_eval_accuracy_prehistory": 0.45714285714285713,
|
| 5536 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3225806451612903,
|
| 5537 |
+
"mmlu_eval_accuracy_professional_law": 0.3352941176470588,
|
| 5538 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
|
| 5539 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
|
| 5540 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 5541 |
+
"mmlu_eval_accuracy_security_studies": 0.4074074074074074,
|
| 5542 |
+
"mmlu_eval_accuracy_sociology": 0.5454545454545454,
|
| 5543 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
|
| 5544 |
+
"mmlu_eval_accuracy_virology": 0.5555555555555556,
|
| 5545 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 5546 |
+
"mmlu_loss": 1.3452752788913467,
|
| 5547 |
+
"step": 5800
|
| 5548 |
+
},
|
| 5549 |
+
{
|
| 5550 |
+
"epoch": 1.69,
|
| 5551 |
+
"learning_rate": 0.0002,
|
| 5552 |
+
"loss": 0.8273,
|
| 5553 |
+
"step": 5810
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 1.69,
|
| 5557 |
+
"learning_rate": 0.0002,
|
| 5558 |
+
"loss": 0.8515,
|
| 5559 |
+
"step": 5820
|
| 5560 |
+
},
|
| 5561 |
+
{
|
| 5562 |
+
"epoch": 1.7,
|
| 5563 |
+
"learning_rate": 0.0002,
|
| 5564 |
+
"loss": 0.8425,
|
| 5565 |
+
"step": 5830
|
| 5566 |
+
},
|
| 5567 |
+
{
|
| 5568 |
+
"epoch": 1.7,
|
| 5569 |
+
"learning_rate": 0.0002,
|
| 5570 |
+
"loss": 0.8472,
|
| 5571 |
+
"step": 5840
|
| 5572 |
+
},
|
| 5573 |
+
{
|
| 5574 |
+
"epoch": 1.7,
|
| 5575 |
+
"learning_rate": 0.0002,
|
| 5576 |
+
"loss": 0.8045,
|
| 5577 |
+
"step": 5850
|
| 5578 |
+
},
|
| 5579 |
+
{
|
| 5580 |
+
"epoch": 1.71,
|
| 5581 |
+
"learning_rate": 0.0002,
|
| 5582 |
+
"loss": 0.8834,
|
| 5583 |
+
"step": 5860
|
| 5584 |
+
},
|
| 5585 |
+
{
|
| 5586 |
+
"epoch": 1.71,
|
| 5587 |
+
"learning_rate": 0.0002,
|
| 5588 |
+
"loss": 0.8467,
|
| 5589 |
+
"step": 5870
|
| 5590 |
+
},
|
| 5591 |
+
{
|
| 5592 |
+
"epoch": 1.71,
|
| 5593 |
+
"learning_rate": 0.0002,
|
| 5594 |
+
"loss": 0.9394,
|
| 5595 |
+
"step": 5880
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 1.71,
|
| 5599 |
+
"learning_rate": 0.0002,
|
| 5600 |
+
"loss": 0.8009,
|
| 5601 |
+
"step": 5890
|
| 5602 |
+
},
|
| 5603 |
+
{
|
| 5604 |
+
"epoch": 1.72,
|
| 5605 |
+
"learning_rate": 0.0002,
|
| 5606 |
+
"loss": 0.8255,
|
| 5607 |
+
"step": 5900
|
| 5608 |
+
},
|
| 5609 |
+
{
|
| 5610 |
+
"epoch": 1.72,
|
| 5611 |
+
"learning_rate": 0.0002,
|
| 5612 |
+
"loss": 0.8928,
|
| 5613 |
+
"step": 5910
|
| 5614 |
+
},
|
| 5615 |
+
{
|
| 5616 |
+
"epoch": 1.72,
|
| 5617 |
+
"learning_rate": 0.0002,
|
| 5618 |
+
"loss": 0.8251,
|
| 5619 |
+
"step": 5920
|
| 5620 |
+
},
|
| 5621 |
+
{
|
| 5622 |
+
"epoch": 1.73,
|
| 5623 |
+
"learning_rate": 0.0002,
|
| 5624 |
+
"loss": 0.8334,
|
| 5625 |
+
"step": 5930
|
| 5626 |
+
},
|
| 5627 |
+
{
|
| 5628 |
+
"epoch": 1.73,
|
| 5629 |
+
"learning_rate": 0.0002,
|
| 5630 |
+
"loss": 0.8703,
|
| 5631 |
+
"step": 5940
|
| 5632 |
+
},
|
| 5633 |
+
{
|
| 5634 |
+
"epoch": 1.73,
|
| 5635 |
+
"learning_rate": 0.0002,
|
| 5636 |
+
"loss": 0.8593,
|
| 5637 |
+
"step": 5950
|
| 5638 |
+
},
|
| 5639 |
+
{
|
| 5640 |
+
"epoch": 1.74,
|
| 5641 |
+
"learning_rate": 0.0002,
|
| 5642 |
+
"loss": 0.7954,
|
| 5643 |
+
"step": 5960
|
| 5644 |
+
},
|
| 5645 |
+
{
|
| 5646 |
+
"epoch": 1.74,
|
| 5647 |
+
"learning_rate": 0.0002,
|
| 5648 |
+
"loss": 0.8163,
|
| 5649 |
+
"step": 5970
|
| 5650 |
+
},
|
| 5651 |
+
{
|
| 5652 |
+
"epoch": 1.74,
|
| 5653 |
+
"learning_rate": 0.0002,
|
| 5654 |
+
"loss": 0.8601,
|
| 5655 |
+
"step": 5980
|
| 5656 |
+
},
|
| 5657 |
+
{
|
| 5658 |
+
"epoch": 1.74,
|
| 5659 |
+
"learning_rate": 0.0002,
|
| 5660 |
+
"loss": 0.8204,
|
| 5661 |
+
"step": 5990
|
| 5662 |
+
},
|
| 5663 |
+
{
|
| 5664 |
+
"epoch": 1.75,
|
| 5665 |
+
"learning_rate": 0.0002,
|
| 5666 |
+
"loss": 0.7698,
|
| 5667 |
+
"step": 6000
|
| 5668 |
+
},
|
| 5669 |
+
{
|
| 5670 |
+
"epoch": 1.75,
|
| 5671 |
+
"eval_loss": 0.8051809668540955,
|
| 5672 |
+
"eval_runtime": 139.9971,
|
| 5673 |
+
"eval_samples_per_second": 7.143,
|
| 5674 |
+
"eval_steps_per_second": 3.572,
|
| 5675 |
+
"step": 6000
|
| 5676 |
+
},
|
| 5677 |
+
{
|
| 5678 |
+
"epoch": 1.75,
|
| 5679 |
+
"mmlu_eval_accuracy": 0.49011725445069554,
|
| 5680 |
+
"mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
|
| 5681 |
+
"mmlu_eval_accuracy_anatomy": 0.7142857142857143,
|
| 5682 |
+
"mmlu_eval_accuracy_astronomy": 0.5,
|
| 5683 |
+
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
|
| 5684 |
+
"mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
|
| 5685 |
+
"mmlu_eval_accuracy_college_biology": 0.5,
|
| 5686 |
+
"mmlu_eval_accuracy_college_chemistry": 0.25,
|
| 5687 |
+
"mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
|
| 5688 |
+
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
|
| 5689 |
+
"mmlu_eval_accuracy_college_medicine": 0.4090909090909091,
|
| 5690 |
+
"mmlu_eval_accuracy_college_physics": 0.36363636363636365,
|
| 5691 |
+
"mmlu_eval_accuracy_computer_security": 0.45454545454545453,
|
| 5692 |
+
"mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
|
| 5693 |
+
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
|
| 5694 |
+
"mmlu_eval_accuracy_electrical_engineering": 0.25,
|
| 5695 |
+
"mmlu_eval_accuracy_elementary_mathematics": 0.2926829268292683,
|
| 5696 |
+
"mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
|
| 5697 |
+
"mmlu_eval_accuracy_global_facts": 0.6,
|
| 5698 |
+
"mmlu_eval_accuracy_high_school_biology": 0.4375,
|
| 5699 |
+
"mmlu_eval_accuracy_high_school_chemistry": 0.2727272727272727,
|
| 5700 |
+
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
|
| 5701 |
+
"mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
|
| 5702 |
+
"mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
|
| 5703 |
+
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
|
| 5704 |
+
"mmlu_eval_accuracy_high_school_macroeconomics": 0.37209302325581395,
|
| 5705 |
+
"mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
|
| 5706 |
+
"mmlu_eval_accuracy_high_school_microeconomics": 0.5,
|
| 5707 |
+
"mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
|
| 5708 |
+
"mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
|
| 5709 |
+
"mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
|
| 5710 |
+
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
|
| 5711 |
+
"mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
|
| 5712 |
+
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
|
| 5713 |
+
"mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
|
| 5714 |
+
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
|
| 5715 |
+
"mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
|
| 5716 |
+
"mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
|
| 5717 |
+
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
|
| 5718 |
+
"mmlu_eval_accuracy_management": 0.6363636363636364,
|
| 5719 |
+
"mmlu_eval_accuracy_marketing": 0.8,
|
| 5720 |
+
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
|
| 5721 |
+
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
|
| 5722 |
+
"mmlu_eval_accuracy_moral_disputes": 0.5,
|
| 5723 |
+
"mmlu_eval_accuracy_moral_scenarios": 0.24,
|
| 5724 |
+
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
|
| 5725 |
+
"mmlu_eval_accuracy_philosophy": 0.4411764705882353,
|
| 5726 |
+
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
|
| 5727 |
+
"mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
|
| 5728 |
+
"mmlu_eval_accuracy_professional_law": 0.3411764705882353,
|
| 5729 |
+
"mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
|
| 5730 |
+
"mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
|
| 5731 |
+
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
|
| 5732 |
+
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
|
| 5733 |
+
"mmlu_eval_accuracy_sociology": 0.5454545454545454,
|
| 5734 |
+
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
|
| 5735 |
+
"mmlu_eval_accuracy_virology": 0.5,
|
| 5736 |
+
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
|
| 5737 |
+
"mmlu_loss": 1.4094214880201898,
|
| 5738 |
+
"step": 6000
|
| 5739 |
}
|
| 5740 |
],
|
| 5741 |
"max_steps": 10000,
|
| 5742 |
"num_train_epochs": 3,
|
| 5743 |
+
"total_flos": 1.6162883060541358e+18,
|
| 5744 |
"trial_name": null,
|
| 5745 |
"trial_params": null
|
| 5746 |
}
|
{checkpoint-4000 → checkpoint-6000}/training_args.bin
RENAMED
|
File without changes
|