Upload 38 files
Browse files- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/adapter_config.json +17 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/adapter_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/optimizer.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/pytorch_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_0.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_1.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_2.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_3.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/scaler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/scheduler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/trainer_state.json +2160 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/training_args.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/adapter_config.json +17 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/adapter_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/optimizer.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/pytorch_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_0.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_1.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_2.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_3.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/scaler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/scheduler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/trainer_state.json +2200 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/training_args.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/adapter_config.json +17 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/adapter_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/optimizer.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/pytorch_model.bin +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_0.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_1.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_2.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_3.pth +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/scaler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/scheduler.pt +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/tokenizer.model +3 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/tokenizer_config.json +33 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/trainer_state.json +2214 -0
- vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/training_args.bin +3 -0
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/adapter_config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "/mnt/data1/sheshuaijie/Data/PLM/vicuna-7b",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"fan_in_fan_out": false,
|
| 5 |
+
"inference_mode": true,
|
| 6 |
+
"init_lora_weights": true,
|
| 7 |
+
"lora_alpha": 32,
|
| 8 |
+
"lora_dropout": 0.1,
|
| 9 |
+
"modules_to_save": null,
|
| 10 |
+
"peft_type": "LORA",
|
| 11 |
+
"r": 32,
|
| 12 |
+
"target_modules": [
|
| 13 |
+
"q_proj",
|
| 14 |
+
"v_proj"
|
| 15 |
+
],
|
| 16 |
+
"task_type": "CAUSAL_LM"
|
| 17 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b180b8c0b2d249cbded40020365b0ad58e9f9fb972021e196dcd8a6c9adf02a5
|
| 3 |
+
size 134293701
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42e1be51147b5d21738c841a7542f08b657b68dd69a604f10caf612f72eeb6fa
|
| 3 |
+
size 67154893
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e6a336f0eb31e65d470fb7bbb3c318d68f8540de99147505297369706e956cb
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a14c3fb4d26a3395451687812bd359f3fd071cba0cdbefbacc693f5cc09806b6
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06f2062d9172c4b83ac74163dbe5ea73882f2aa0b6cf47bac1ee147a28f45e3f
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:791c74a8fa5a686ccf4034dc3001174c86737c9a81b7e09d7225d6fbee5c0212
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ff694532efc9aa102a9df1b12588365f8a813e88f900aa3b82f011b9ff17989
|
| 3 |
+
size 557
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2c98200ad231d30e6f897712cf004f180062f0ec7f8523760ba97dd2e43c318
|
| 3 |
+
size 627
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/trainer_state.json
ADDED
|
@@ -0,0 +1,2160 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6167441010475159,
|
| 3 |
+
"best_model_checkpoint": "/mnt/data1/sheshuaijie/Output/CoT/Trained/vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723",
|
| 4 |
+
"epoch": 9.54061855670103,
|
| 5 |
+
"global_step": 723,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"eval_loss": 1.7524008750915527,
|
| 13 |
+
"eval_runtime": 23.7951,
|
| 14 |
+
"eval_samples_per_second": 126.077,
|
| 15 |
+
"eval_steps_per_second": 3.95,
|
| 16 |
+
"step": 3
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.08,
|
| 20 |
+
"eval_loss": 1.5515066385269165,
|
| 21 |
+
"eval_runtime": 23.828,
|
| 22 |
+
"eval_samples_per_second": 125.902,
|
| 23 |
+
"eval_steps_per_second": 3.945,
|
| 24 |
+
"step": 6
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12,
|
| 28 |
+
"eval_loss": 1.3584641218185425,
|
| 29 |
+
"eval_runtime": 23.8775,
|
| 30 |
+
"eval_samples_per_second": 125.641,
|
| 31 |
+
"eval_steps_per_second": 3.937,
|
| 32 |
+
"step": 9
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"eval_loss": 1.2644726037979126,
|
| 37 |
+
"eval_runtime": 23.8942,
|
| 38 |
+
"eval_samples_per_second": 125.554,
|
| 39 |
+
"eval_steps_per_second": 3.934,
|
| 40 |
+
"step": 12
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.2,
|
| 44 |
+
"eval_loss": 1.166400671005249,
|
| 45 |
+
"eval_runtime": 23.9181,
|
| 46 |
+
"eval_samples_per_second": 125.428,
|
| 47 |
+
"eval_steps_per_second": 3.93,
|
| 48 |
+
"step": 15
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.24,
|
| 52 |
+
"eval_loss": 1.1086052656173706,
|
| 53 |
+
"eval_runtime": 23.9166,
|
| 54 |
+
"eval_samples_per_second": 125.436,
|
| 55 |
+
"eval_steps_per_second": 3.93,
|
| 56 |
+
"step": 18
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.26,
|
| 60 |
+
"learning_rate": 0.00019650067294751011,
|
| 61 |
+
"loss": 1.4265,
|
| 62 |
+
"step": 20
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.28,
|
| 66 |
+
"eval_loss": 1.0677987337112427,
|
| 67 |
+
"eval_runtime": 23.9189,
|
| 68 |
+
"eval_samples_per_second": 125.424,
|
| 69 |
+
"eval_steps_per_second": 3.93,
|
| 70 |
+
"step": 21
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.32,
|
| 74 |
+
"eval_loss": 1.0342437028884888,
|
| 75 |
+
"eval_runtime": 23.9046,
|
| 76 |
+
"eval_samples_per_second": 125.499,
|
| 77 |
+
"eval_steps_per_second": 3.932,
|
| 78 |
+
"step": 24
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.36,
|
| 82 |
+
"eval_loss": 0.9985266923904419,
|
| 83 |
+
"eval_runtime": 23.9037,
|
| 84 |
+
"eval_samples_per_second": 125.504,
|
| 85 |
+
"eval_steps_per_second": 3.932,
|
| 86 |
+
"step": 27
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.4,
|
| 90 |
+
"eval_loss": 0.9654523134231567,
|
| 91 |
+
"eval_runtime": 23.9129,
|
| 92 |
+
"eval_samples_per_second": 125.455,
|
| 93 |
+
"eval_steps_per_second": 3.931,
|
| 94 |
+
"step": 30
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.44,
|
| 98 |
+
"eval_loss": 0.939262866973877,
|
| 99 |
+
"eval_runtime": 23.9117,
|
| 100 |
+
"eval_samples_per_second": 125.462,
|
| 101 |
+
"eval_steps_per_second": 3.931,
|
| 102 |
+
"step": 33
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.48,
|
| 106 |
+
"eval_loss": 0.9186767339706421,
|
| 107 |
+
"eval_runtime": 23.9011,
|
| 108 |
+
"eval_samples_per_second": 125.517,
|
| 109 |
+
"eval_steps_per_second": 3.933,
|
| 110 |
+
"step": 36
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.51,
|
| 114 |
+
"eval_loss": 0.8969741463661194,
|
| 115 |
+
"eval_runtime": 23.9105,
|
| 116 |
+
"eval_samples_per_second": 125.468,
|
| 117 |
+
"eval_steps_per_second": 3.931,
|
| 118 |
+
"step": 39
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.53,
|
| 122 |
+
"learning_rate": 0.00019111709286675642,
|
| 123 |
+
"loss": 0.9923,
|
| 124 |
+
"step": 40
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.55,
|
| 128 |
+
"eval_loss": 0.8814375996589661,
|
| 129 |
+
"eval_runtime": 23.9154,
|
| 130 |
+
"eval_samples_per_second": 125.442,
|
| 131 |
+
"eval_steps_per_second": 3.931,
|
| 132 |
+
"step": 42
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 0.59,
|
| 136 |
+
"eval_loss": 0.8654683232307434,
|
| 137 |
+
"eval_runtime": 23.9108,
|
| 138 |
+
"eval_samples_per_second": 125.466,
|
| 139 |
+
"eval_steps_per_second": 3.931,
|
| 140 |
+
"step": 45
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.63,
|
| 144 |
+
"eval_loss": 0.852226734161377,
|
| 145 |
+
"eval_runtime": 23.9186,
|
| 146 |
+
"eval_samples_per_second": 125.425,
|
| 147 |
+
"eval_steps_per_second": 3.93,
|
| 148 |
+
"step": 48
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.67,
|
| 152 |
+
"eval_loss": 0.839223325252533,
|
| 153 |
+
"eval_runtime": 23.9074,
|
| 154 |
+
"eval_samples_per_second": 125.484,
|
| 155 |
+
"eval_steps_per_second": 3.932,
|
| 156 |
+
"step": 51
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.71,
|
| 160 |
+
"eval_loss": 0.8266379237174988,
|
| 161 |
+
"eval_runtime": 23.9399,
|
| 162 |
+
"eval_samples_per_second": 125.314,
|
| 163 |
+
"eval_steps_per_second": 3.926,
|
| 164 |
+
"step": 54
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.75,
|
| 168 |
+
"eval_loss": 0.8140417337417603,
|
| 169 |
+
"eval_runtime": 23.9355,
|
| 170 |
+
"eval_samples_per_second": 125.337,
|
| 171 |
+
"eval_steps_per_second": 3.927,
|
| 172 |
+
"step": 57
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.79,
|
| 176 |
+
"learning_rate": 0.0001857335127860027,
|
| 177 |
+
"loss": 0.8611,
|
| 178 |
+
"step": 60
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.79,
|
| 182 |
+
"eval_loss": 0.8019057512283325,
|
| 183 |
+
"eval_runtime": 23.9223,
|
| 184 |
+
"eval_samples_per_second": 125.406,
|
| 185 |
+
"eval_steps_per_second": 3.929,
|
| 186 |
+
"step": 60
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.83,
|
| 190 |
+
"eval_loss": 0.7907609343528748,
|
| 191 |
+
"eval_runtime": 23.9384,
|
| 192 |
+
"eval_samples_per_second": 125.322,
|
| 193 |
+
"eval_steps_per_second": 3.927,
|
| 194 |
+
"step": 63
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.87,
|
| 198 |
+
"eval_loss": 0.7791212797164917,
|
| 199 |
+
"eval_runtime": 23.9101,
|
| 200 |
+
"eval_samples_per_second": 125.47,
|
| 201 |
+
"eval_steps_per_second": 3.931,
|
| 202 |
+
"step": 66
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.91,
|
| 206 |
+
"eval_loss": 0.7694615125656128,
|
| 207 |
+
"eval_runtime": 23.9079,
|
| 208 |
+
"eval_samples_per_second": 125.481,
|
| 209 |
+
"eval_steps_per_second": 3.932,
|
| 210 |
+
"step": 69
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.95,
|
| 214 |
+
"eval_loss": 0.7602358460426331,
|
| 215 |
+
"eval_runtime": 23.9116,
|
| 216 |
+
"eval_samples_per_second": 125.462,
|
| 217 |
+
"eval_steps_per_second": 3.931,
|
| 218 |
+
"step": 72
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.99,
|
| 222 |
+
"eval_loss": 0.753226101398468,
|
| 223 |
+
"eval_runtime": 23.9242,
|
| 224 |
+
"eval_samples_per_second": 125.396,
|
| 225 |
+
"eval_steps_per_second": 3.929,
|
| 226 |
+
"step": 75
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.03,
|
| 230 |
+
"eval_loss": 0.7466432452201843,
|
| 231 |
+
"eval_runtime": 23.9116,
|
| 232 |
+
"eval_samples_per_second": 125.462,
|
| 233 |
+
"eval_steps_per_second": 3.931,
|
| 234 |
+
"step": 78
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.06,
|
| 238 |
+
"learning_rate": 0.000180349932705249,
|
| 239 |
+
"loss": 0.7843,
|
| 240 |
+
"step": 80
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.07,
|
| 244 |
+
"eval_loss": 0.7416810989379883,
|
| 245 |
+
"eval_runtime": 23.9171,
|
| 246 |
+
"eval_samples_per_second": 125.433,
|
| 247 |
+
"eval_steps_per_second": 3.93,
|
| 248 |
+
"step": 81
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.11,
|
| 252 |
+
"eval_loss": 0.7362396121025085,
|
| 253 |
+
"eval_runtime": 23.9079,
|
| 254 |
+
"eval_samples_per_second": 125.481,
|
| 255 |
+
"eval_steps_per_second": 3.932,
|
| 256 |
+
"step": 84
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.15,
|
| 260 |
+
"eval_loss": 0.7297741174697876,
|
| 261 |
+
"eval_runtime": 23.9084,
|
| 262 |
+
"eval_samples_per_second": 125.479,
|
| 263 |
+
"eval_steps_per_second": 3.932,
|
| 264 |
+
"step": 87
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.19,
|
| 268 |
+
"eval_loss": 0.7252654433250427,
|
| 269 |
+
"eval_runtime": 23.9206,
|
| 270 |
+
"eval_samples_per_second": 125.415,
|
| 271 |
+
"eval_steps_per_second": 3.93,
|
| 272 |
+
"step": 90
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.23,
|
| 276 |
+
"eval_loss": 0.7213409543037415,
|
| 277 |
+
"eval_runtime": 23.9179,
|
| 278 |
+
"eval_samples_per_second": 125.429,
|
| 279 |
+
"eval_steps_per_second": 3.93,
|
| 280 |
+
"step": 93
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.27,
|
| 284 |
+
"eval_loss": 0.7174035906791687,
|
| 285 |
+
"eval_runtime": 23.9354,
|
| 286 |
+
"eval_samples_per_second": 125.337,
|
| 287 |
+
"eval_steps_per_second": 3.927,
|
| 288 |
+
"step": 96
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.31,
|
| 292 |
+
"eval_loss": 0.7140380144119263,
|
| 293 |
+
"eval_runtime": 23.9214,
|
| 294 |
+
"eval_samples_per_second": 125.411,
|
| 295 |
+
"eval_steps_per_second": 3.93,
|
| 296 |
+
"step": 99
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.32,
|
| 300 |
+
"learning_rate": 0.0001749663526244953,
|
| 301 |
+
"loss": 0.7301,
|
| 302 |
+
"step": 100
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.35,
|
| 306 |
+
"eval_loss": 0.7104487419128418,
|
| 307 |
+
"eval_runtime": 23.9093,
|
| 308 |
+
"eval_samples_per_second": 125.474,
|
| 309 |
+
"eval_steps_per_second": 3.932,
|
| 310 |
+
"step": 102
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.39,
|
| 314 |
+
"eval_loss": 0.7067868113517761,
|
| 315 |
+
"eval_runtime": 23.9129,
|
| 316 |
+
"eval_samples_per_second": 125.455,
|
| 317 |
+
"eval_steps_per_second": 3.931,
|
| 318 |
+
"step": 105
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.43,
|
| 322 |
+
"eval_loss": 0.7041762471199036,
|
| 323 |
+
"eval_runtime": 23.9161,
|
| 324 |
+
"eval_samples_per_second": 125.439,
|
| 325 |
+
"eval_steps_per_second": 3.93,
|
| 326 |
+
"step": 108
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.46,
|
| 330 |
+
"eval_loss": 0.7013522982597351,
|
| 331 |
+
"eval_runtime": 23.9133,
|
| 332 |
+
"eval_samples_per_second": 125.453,
|
| 333 |
+
"eval_steps_per_second": 3.931,
|
| 334 |
+
"step": 111
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.5,
|
| 338 |
+
"eval_loss": 0.6989504098892212,
|
| 339 |
+
"eval_runtime": 23.9152,
|
| 340 |
+
"eval_samples_per_second": 125.443,
|
| 341 |
+
"eval_steps_per_second": 3.931,
|
| 342 |
+
"step": 114
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 1.54,
|
| 346 |
+
"eval_loss": 0.6974085569381714,
|
| 347 |
+
"eval_runtime": 23.9561,
|
| 348 |
+
"eval_samples_per_second": 125.229,
|
| 349 |
+
"eval_steps_per_second": 3.924,
|
| 350 |
+
"step": 117
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.58,
|
| 354 |
+
"learning_rate": 0.0001695827725437416,
|
| 355 |
+
"loss": 0.7141,
|
| 356 |
+
"step": 120
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.58,
|
| 360 |
+
"eval_loss": 0.6944894194602966,
|
| 361 |
+
"eval_runtime": 23.902,
|
| 362 |
+
"eval_samples_per_second": 125.512,
|
| 363 |
+
"eval_steps_per_second": 3.933,
|
| 364 |
+
"step": 120
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 1.62,
|
| 368 |
+
"eval_loss": 0.6929482221603394,
|
| 369 |
+
"eval_runtime": 23.9189,
|
| 370 |
+
"eval_samples_per_second": 125.424,
|
| 371 |
+
"eval_steps_per_second": 3.93,
|
| 372 |
+
"step": 123
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.66,
|
| 376 |
+
"eval_loss": 0.6903366446495056,
|
| 377 |
+
"eval_runtime": 23.9061,
|
| 378 |
+
"eval_samples_per_second": 125.491,
|
| 379 |
+
"eval_steps_per_second": 3.932,
|
| 380 |
+
"step": 126
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.7,
|
| 384 |
+
"eval_loss": 0.6882749199867249,
|
| 385 |
+
"eval_runtime": 23.9181,
|
| 386 |
+
"eval_samples_per_second": 125.428,
|
| 387 |
+
"eval_steps_per_second": 3.93,
|
| 388 |
+
"step": 129
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.74,
|
| 392 |
+
"eval_loss": 0.6863100528717041,
|
| 393 |
+
"eval_runtime": 23.914,
|
| 394 |
+
"eval_samples_per_second": 125.45,
|
| 395 |
+
"eval_steps_per_second": 3.931,
|
| 396 |
+
"step": 132
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.78,
|
| 400 |
+
"eval_loss": 0.6860549449920654,
|
| 401 |
+
"eval_runtime": 23.9138,
|
| 402 |
+
"eval_samples_per_second": 125.45,
|
| 403 |
+
"eval_steps_per_second": 3.931,
|
| 404 |
+
"step": 135
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.82,
|
| 408 |
+
"eval_loss": 0.6831715703010559,
|
| 409 |
+
"eval_runtime": 23.9135,
|
| 410 |
+
"eval_samples_per_second": 125.452,
|
| 411 |
+
"eval_steps_per_second": 3.931,
|
| 412 |
+
"step": 138
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.85,
|
| 416 |
+
"learning_rate": 0.0001641991924629879,
|
| 417 |
+
"loss": 0.6902,
|
| 418 |
+
"step": 140
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.86,
|
| 422 |
+
"eval_loss": 0.6819499731063843,
|
| 423 |
+
"eval_runtime": 23.8986,
|
| 424 |
+
"eval_samples_per_second": 125.53,
|
| 425 |
+
"eval_steps_per_second": 3.933,
|
| 426 |
+
"step": 141
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.9,
|
| 430 |
+
"eval_loss": 0.6807693839073181,
|
| 431 |
+
"eval_runtime": 23.9169,
|
| 432 |
+
"eval_samples_per_second": 125.434,
|
| 433 |
+
"eval_steps_per_second": 3.93,
|
| 434 |
+
"step": 144
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 1.94,
|
| 438 |
+
"eval_loss": 0.6787669062614441,
|
| 439 |
+
"eval_runtime": 23.9265,
|
| 440 |
+
"eval_samples_per_second": 125.384,
|
| 441 |
+
"eval_steps_per_second": 3.929,
|
| 442 |
+
"step": 147
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.98,
|
| 446 |
+
"eval_loss": 0.6773442625999451,
|
| 447 |
+
"eval_runtime": 23.9274,
|
| 448 |
+
"eval_samples_per_second": 125.38,
|
| 449 |
+
"eval_steps_per_second": 3.929,
|
| 450 |
+
"step": 150
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.02,
|
| 454 |
+
"eval_loss": 0.6759281158447266,
|
| 455 |
+
"eval_runtime": 23.9386,
|
| 456 |
+
"eval_samples_per_second": 125.321,
|
| 457 |
+
"eval_steps_per_second": 3.927,
|
| 458 |
+
"step": 153
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.06,
|
| 462 |
+
"eval_loss": 0.6743582487106323,
|
| 463 |
+
"eval_runtime": 23.9323,
|
| 464 |
+
"eval_samples_per_second": 125.354,
|
| 465 |
+
"eval_steps_per_second": 3.928,
|
| 466 |
+
"step": 156
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 2.1,
|
| 470 |
+
"eval_loss": 0.6732926368713379,
|
| 471 |
+
"eval_runtime": 23.9145,
|
| 472 |
+
"eval_samples_per_second": 125.447,
|
| 473 |
+
"eval_steps_per_second": 3.931,
|
| 474 |
+
"step": 159
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 2.11,
|
| 478 |
+
"learning_rate": 0.0001588156123822342,
|
| 479 |
+
"loss": 0.6766,
|
| 480 |
+
"step": 160
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 2.14,
|
| 484 |
+
"eval_loss": 0.6721953749656677,
|
| 485 |
+
"eval_runtime": 23.9073,
|
| 486 |
+
"eval_samples_per_second": 125.485,
|
| 487 |
+
"eval_steps_per_second": 3.932,
|
| 488 |
+
"step": 162
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.18,
|
| 492 |
+
"eval_loss": 0.6714429259300232,
|
| 493 |
+
"eval_runtime": 23.8955,
|
| 494 |
+
"eval_samples_per_second": 125.547,
|
| 495 |
+
"eval_steps_per_second": 3.934,
|
| 496 |
+
"step": 165
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 2.22,
|
| 500 |
+
"eval_loss": 0.670035183429718,
|
| 501 |
+
"eval_runtime": 23.9431,
|
| 502 |
+
"eval_samples_per_second": 125.297,
|
| 503 |
+
"eval_steps_per_second": 3.926,
|
| 504 |
+
"step": 168
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 2.26,
|
| 508 |
+
"eval_loss": 0.6695354580879211,
|
| 509 |
+
"eval_runtime": 23.8875,
|
| 510 |
+
"eval_samples_per_second": 125.589,
|
| 511 |
+
"eval_steps_per_second": 3.935,
|
| 512 |
+
"step": 171
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.3,
|
| 516 |
+
"eval_loss": 0.6689226031303406,
|
| 517 |
+
"eval_runtime": 23.9185,
|
| 518 |
+
"eval_samples_per_second": 125.426,
|
| 519 |
+
"eval_steps_per_second": 3.93,
|
| 520 |
+
"step": 174
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.34,
|
| 524 |
+
"eval_loss": 0.6674054861068726,
|
| 525 |
+
"eval_runtime": 23.941,
|
| 526 |
+
"eval_samples_per_second": 125.308,
|
| 527 |
+
"eval_steps_per_second": 3.926,
|
| 528 |
+
"step": 177
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.38,
|
| 532 |
+
"learning_rate": 0.00015343203230148048,
|
| 533 |
+
"loss": 0.6743,
|
| 534 |
+
"step": 180
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.38,
|
| 538 |
+
"eval_loss": 0.6664847731590271,
|
| 539 |
+
"eval_runtime": 23.9211,
|
| 540 |
+
"eval_samples_per_second": 125.412,
|
| 541 |
+
"eval_steps_per_second": 3.93,
|
| 542 |
+
"step": 180
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.41,
|
| 546 |
+
"eval_loss": 0.6658627986907959,
|
| 547 |
+
"eval_runtime": 23.9247,
|
| 548 |
+
"eval_samples_per_second": 125.394,
|
| 549 |
+
"eval_steps_per_second": 3.929,
|
| 550 |
+
"step": 183
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 2.45,
|
| 554 |
+
"eval_loss": 0.664908766746521,
|
| 555 |
+
"eval_runtime": 23.9272,
|
| 556 |
+
"eval_samples_per_second": 125.38,
|
| 557 |
+
"eval_steps_per_second": 3.929,
|
| 558 |
+
"step": 186
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.49,
|
| 562 |
+
"eval_loss": 0.6638036966323853,
|
| 563 |
+
"eval_runtime": 23.9187,
|
| 564 |
+
"eval_samples_per_second": 125.425,
|
| 565 |
+
"eval_steps_per_second": 3.93,
|
| 566 |
+
"step": 189
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.53,
|
| 570 |
+
"eval_loss": 0.6625837683677673,
|
| 571 |
+
"eval_runtime": 23.9033,
|
| 572 |
+
"eval_samples_per_second": 125.506,
|
| 573 |
+
"eval_steps_per_second": 3.933,
|
| 574 |
+
"step": 192
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 2.57,
|
| 578 |
+
"eval_loss": 0.6619511842727661,
|
| 579 |
+
"eval_runtime": 23.8973,
|
| 580 |
+
"eval_samples_per_second": 125.537,
|
| 581 |
+
"eval_steps_per_second": 3.934,
|
| 582 |
+
"step": 195
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.61,
|
| 586 |
+
"eval_loss": 0.6611769199371338,
|
| 587 |
+
"eval_runtime": 23.9129,
|
| 588 |
+
"eval_samples_per_second": 125.455,
|
| 589 |
+
"eval_steps_per_second": 3.931,
|
| 590 |
+
"step": 198
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.64,
|
| 594 |
+
"learning_rate": 0.00014804845222072678,
|
| 595 |
+
"loss": 0.6615,
|
| 596 |
+
"step": 200
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.65,
|
| 600 |
+
"eval_loss": 0.6606143116950989,
|
| 601 |
+
"eval_runtime": 23.9126,
|
| 602 |
+
"eval_samples_per_second": 125.457,
|
| 603 |
+
"eval_steps_per_second": 3.931,
|
| 604 |
+
"step": 201
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 2.69,
|
| 608 |
+
"eval_loss": 0.6589743494987488,
|
| 609 |
+
"eval_runtime": 23.9135,
|
| 610 |
+
"eval_samples_per_second": 125.452,
|
| 611 |
+
"eval_steps_per_second": 3.931,
|
| 612 |
+
"step": 204
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.73,
|
| 616 |
+
"eval_loss": 0.6578481197357178,
|
| 617 |
+
"eval_runtime": 23.9217,
|
| 618 |
+
"eval_samples_per_second": 125.409,
|
| 619 |
+
"eval_steps_per_second": 3.929,
|
| 620 |
+
"step": 207
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.77,
|
| 624 |
+
"eval_loss": 0.6571096181869507,
|
| 625 |
+
"eval_runtime": 23.9415,
|
| 626 |
+
"eval_samples_per_second": 125.305,
|
| 627 |
+
"eval_steps_per_second": 3.926,
|
| 628 |
+
"step": 210
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 2.81,
|
| 632 |
+
"eval_loss": 0.656689465045929,
|
| 633 |
+
"eval_runtime": 23.9111,
|
| 634 |
+
"eval_samples_per_second": 125.465,
|
| 635 |
+
"eval_steps_per_second": 3.931,
|
| 636 |
+
"step": 213
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.85,
|
| 640 |
+
"eval_loss": 0.6556207537651062,
|
| 641 |
+
"eval_runtime": 23.9099,
|
| 642 |
+
"eval_samples_per_second": 125.471,
|
| 643 |
+
"eval_steps_per_second": 3.931,
|
| 644 |
+
"step": 216
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.89,
|
| 648 |
+
"eval_loss": 0.6546627283096313,
|
| 649 |
+
"eval_runtime": 23.9164,
|
| 650 |
+
"eval_samples_per_second": 125.437,
|
| 651 |
+
"eval_steps_per_second": 3.93,
|
| 652 |
+
"step": 219
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 2.9,
|
| 656 |
+
"learning_rate": 0.0001426648721399731,
|
| 657 |
+
"loss": 0.6564,
|
| 658 |
+
"step": 220
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 2.93,
|
| 662 |
+
"eval_loss": 0.6539400815963745,
|
| 663 |
+
"eval_runtime": 23.906,
|
| 664 |
+
"eval_samples_per_second": 125.492,
|
| 665 |
+
"eval_steps_per_second": 3.932,
|
| 666 |
+
"step": 222
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.97,
|
| 670 |
+
"eval_loss": 0.653684675693512,
|
| 671 |
+
"eval_runtime": 23.9251,
|
| 672 |
+
"eval_samples_per_second": 125.391,
|
| 673 |
+
"eval_steps_per_second": 3.929,
|
| 674 |
+
"step": 225
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.01,
|
| 678 |
+
"eval_loss": 0.6526629328727722,
|
| 679 |
+
"eval_runtime": 23.9289,
|
| 680 |
+
"eval_samples_per_second": 125.371,
|
| 681 |
+
"eval_steps_per_second": 3.928,
|
| 682 |
+
"step": 228
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 3.05,
|
| 686 |
+
"eval_loss": 0.6525079011917114,
|
| 687 |
+
"eval_runtime": 23.9193,
|
| 688 |
+
"eval_samples_per_second": 125.421,
|
| 689 |
+
"eval_steps_per_second": 3.93,
|
| 690 |
+
"step": 231
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 3.09,
|
| 694 |
+
"eval_loss": 0.6514959931373596,
|
| 695 |
+
"eval_runtime": 23.9574,
|
| 696 |
+
"eval_samples_per_second": 125.223,
|
| 697 |
+
"eval_steps_per_second": 3.924,
|
| 698 |
+
"step": 234
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.13,
|
| 702 |
+
"eval_loss": 0.6507047414779663,
|
| 703 |
+
"eval_runtime": 23.9234,
|
| 704 |
+
"eval_samples_per_second": 125.4,
|
| 705 |
+
"eval_steps_per_second": 3.929,
|
| 706 |
+
"step": 237
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 3.17,
|
| 710 |
+
"learning_rate": 0.00013728129205921937,
|
| 711 |
+
"loss": 0.6469,
|
| 712 |
+
"step": 240
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 3.17,
|
| 716 |
+
"eval_loss": 0.6504186391830444,
|
| 717 |
+
"eval_runtime": 23.937,
|
| 718 |
+
"eval_samples_per_second": 125.329,
|
| 719 |
+
"eval_steps_per_second": 3.927,
|
| 720 |
+
"step": 240
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 3.21,
|
| 724 |
+
"eval_loss": 0.6495808959007263,
|
| 725 |
+
"eval_runtime": 23.9188,
|
| 726 |
+
"eval_samples_per_second": 125.425,
|
| 727 |
+
"eval_steps_per_second": 3.93,
|
| 728 |
+
"step": 243
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.25,
|
| 732 |
+
"eval_loss": 0.649512529373169,
|
| 733 |
+
"eval_runtime": 23.9209,
|
| 734 |
+
"eval_samples_per_second": 125.413,
|
| 735 |
+
"eval_steps_per_second": 3.93,
|
| 736 |
+
"step": 246
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 3.29,
|
| 740 |
+
"eval_loss": 0.648629903793335,
|
| 741 |
+
"eval_runtime": 23.9137,
|
| 742 |
+
"eval_samples_per_second": 125.451,
|
| 743 |
+
"eval_steps_per_second": 3.931,
|
| 744 |
+
"step": 249
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 3.33,
|
| 748 |
+
"eval_loss": 0.6480894088745117,
|
| 749 |
+
"eval_runtime": 23.919,
|
| 750 |
+
"eval_samples_per_second": 125.423,
|
| 751 |
+
"eval_steps_per_second": 3.93,
|
| 752 |
+
"step": 252
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.36,
|
| 756 |
+
"eval_loss": 0.6474400758743286,
|
| 757 |
+
"eval_runtime": 23.9076,
|
| 758 |
+
"eval_samples_per_second": 125.483,
|
| 759 |
+
"eval_steps_per_second": 3.932,
|
| 760 |
+
"step": 255
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 3.4,
|
| 764 |
+
"eval_loss": 0.6468291878700256,
|
| 765 |
+
"eval_runtime": 23.9305,
|
| 766 |
+
"eval_samples_per_second": 125.363,
|
| 767 |
+
"eval_steps_per_second": 3.928,
|
| 768 |
+
"step": 258
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 3.43,
|
| 772 |
+
"learning_rate": 0.00013189771197846567,
|
| 773 |
+
"loss": 0.6463,
|
| 774 |
+
"step": 260
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 3.44,
|
| 778 |
+
"eval_loss": 0.6462663412094116,
|
| 779 |
+
"eval_runtime": 23.9359,
|
| 780 |
+
"eval_samples_per_second": 125.335,
|
| 781 |
+
"eval_steps_per_second": 3.927,
|
| 782 |
+
"step": 261
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.48,
|
| 786 |
+
"eval_loss": 0.6458565592765808,
|
| 787 |
+
"eval_runtime": 23.929,
|
| 788 |
+
"eval_samples_per_second": 125.371,
|
| 789 |
+
"eval_steps_per_second": 3.928,
|
| 790 |
+
"step": 264
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 3.52,
|
| 794 |
+
"eval_loss": 0.645412266254425,
|
| 795 |
+
"eval_runtime": 23.9362,
|
| 796 |
+
"eval_samples_per_second": 125.333,
|
| 797 |
+
"eval_steps_per_second": 3.927,
|
| 798 |
+
"step": 267
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 3.56,
|
| 802 |
+
"eval_loss": 0.6449554562568665,
|
| 803 |
+
"eval_runtime": 23.9004,
|
| 804 |
+
"eval_samples_per_second": 125.521,
|
| 805 |
+
"eval_steps_per_second": 3.933,
|
| 806 |
+
"step": 270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.6,
|
| 810 |
+
"eval_loss": 0.6443325281143188,
|
| 811 |
+
"eval_runtime": 23.9065,
|
| 812 |
+
"eval_samples_per_second": 125.489,
|
| 813 |
+
"eval_steps_per_second": 3.932,
|
| 814 |
+
"step": 273
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 3.64,
|
| 818 |
+
"eval_loss": 0.6435034871101379,
|
| 819 |
+
"eval_runtime": 23.9072,
|
| 820 |
+
"eval_samples_per_second": 125.485,
|
| 821 |
+
"eval_steps_per_second": 3.932,
|
| 822 |
+
"step": 276
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 3.68,
|
| 826 |
+
"eval_loss": 0.6433733701705933,
|
| 827 |
+
"eval_runtime": 23.9042,
|
| 828 |
+
"eval_samples_per_second": 125.501,
|
| 829 |
+
"eval_steps_per_second": 3.932,
|
| 830 |
+
"step": 279
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.69,
|
| 834 |
+
"learning_rate": 0.00012651413189771198,
|
| 835 |
+
"loss": 0.6389,
|
| 836 |
+
"step": 280
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.72,
|
| 840 |
+
"eval_loss": 0.6425070762634277,
|
| 841 |
+
"eval_runtime": 23.8874,
|
| 842 |
+
"eval_samples_per_second": 125.589,
|
| 843 |
+
"eval_steps_per_second": 3.935,
|
| 844 |
+
"step": 282
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 3.76,
|
| 848 |
+
"eval_loss": 0.642119288444519,
|
| 849 |
+
"eval_runtime": 23.9328,
|
| 850 |
+
"eval_samples_per_second": 125.351,
|
| 851 |
+
"eval_steps_per_second": 3.928,
|
| 852 |
+
"step": 285
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 3.8,
|
| 856 |
+
"eval_loss": 0.641748309135437,
|
| 857 |
+
"eval_runtime": 23.9294,
|
| 858 |
+
"eval_samples_per_second": 125.369,
|
| 859 |
+
"eval_steps_per_second": 3.928,
|
| 860 |
+
"step": 288
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.84,
|
| 864 |
+
"eval_loss": 0.640826404094696,
|
| 865 |
+
"eval_runtime": 23.9434,
|
| 866 |
+
"eval_samples_per_second": 125.296,
|
| 867 |
+
"eval_steps_per_second": 3.926,
|
| 868 |
+
"step": 291
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 3.88,
|
| 872 |
+
"eval_loss": 0.6402388215065002,
|
| 873 |
+
"eval_runtime": 23.9162,
|
| 874 |
+
"eval_samples_per_second": 125.438,
|
| 875 |
+
"eval_steps_per_second": 3.93,
|
| 876 |
+
"step": 294
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 3.92,
|
| 880 |
+
"eval_loss": 0.6407353281974792,
|
| 881 |
+
"eval_runtime": 23.9121,
|
| 882 |
+
"eval_samples_per_second": 125.46,
|
| 883 |
+
"eval_steps_per_second": 3.931,
|
| 884 |
+
"step": 297
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.96,
|
| 888 |
+
"learning_rate": 0.0001211305518169583,
|
| 889 |
+
"loss": 0.6318,
|
| 890 |
+
"step": 300
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.96,
|
| 894 |
+
"eval_loss": 0.6398600935935974,
|
| 895 |
+
"eval_runtime": 23.9229,
|
| 896 |
+
"eval_samples_per_second": 125.403,
|
| 897 |
+
"eval_steps_per_second": 3.929,
|
| 898 |
+
"step": 300
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 4.0,
|
| 902 |
+
"eval_loss": 0.6393464207649231,
|
| 903 |
+
"eval_runtime": 23.9187,
|
| 904 |
+
"eval_samples_per_second": 125.425,
|
| 905 |
+
"eval_steps_per_second": 3.93,
|
| 906 |
+
"step": 303
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 4.04,
|
| 910 |
+
"eval_loss": 0.6392526626586914,
|
| 911 |
+
"eval_runtime": 23.9074,
|
| 912 |
+
"eval_samples_per_second": 125.484,
|
| 913 |
+
"eval_steps_per_second": 3.932,
|
| 914 |
+
"step": 306
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.08,
|
| 918 |
+
"eval_loss": 0.6389594078063965,
|
| 919 |
+
"eval_runtime": 23.918,
|
| 920 |
+
"eval_samples_per_second": 125.428,
|
| 921 |
+
"eval_steps_per_second": 3.93,
|
| 922 |
+
"step": 309
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 4.12,
|
| 926 |
+
"eval_loss": 0.6388808488845825,
|
| 927 |
+
"eval_runtime": 23.9158,
|
| 928 |
+
"eval_samples_per_second": 125.44,
|
| 929 |
+
"eval_steps_per_second": 3.93,
|
| 930 |
+
"step": 312
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 4.16,
|
| 934 |
+
"eval_loss": 0.6384025812149048,
|
| 935 |
+
"eval_runtime": 23.9176,
|
| 936 |
+
"eval_samples_per_second": 125.431,
|
| 937 |
+
"eval_steps_per_second": 3.93,
|
| 938 |
+
"step": 315
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.2,
|
| 942 |
+
"eval_loss": 0.6387144923210144,
|
| 943 |
+
"eval_runtime": 23.9047,
|
| 944 |
+
"eval_samples_per_second": 125.498,
|
| 945 |
+
"eval_steps_per_second": 3.932,
|
| 946 |
+
"step": 318
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 4.22,
|
| 950 |
+
"learning_rate": 0.00011574697173620459,
|
| 951 |
+
"loss": 0.6277,
|
| 952 |
+
"step": 320
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 4.24,
|
| 956 |
+
"eval_loss": 0.6377059817314148,
|
| 957 |
+
"eval_runtime": 23.9246,
|
| 958 |
+
"eval_samples_per_second": 125.394,
|
| 959 |
+
"eval_steps_per_second": 3.929,
|
| 960 |
+
"step": 321
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 4.28,
|
| 964 |
+
"eval_loss": 0.636981189250946,
|
| 965 |
+
"eval_runtime": 23.9459,
|
| 966 |
+
"eval_samples_per_second": 125.283,
|
| 967 |
+
"eval_steps_per_second": 3.926,
|
| 968 |
+
"step": 324
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 4.32,
|
| 972 |
+
"eval_loss": 0.6364036202430725,
|
| 973 |
+
"eval_runtime": 23.9206,
|
| 974 |
+
"eval_samples_per_second": 125.415,
|
| 975 |
+
"eval_steps_per_second": 3.93,
|
| 976 |
+
"step": 327
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 4.35,
|
| 980 |
+
"eval_loss": 0.6357031464576721,
|
| 981 |
+
"eval_runtime": 23.9187,
|
| 982 |
+
"eval_samples_per_second": 125.425,
|
| 983 |
+
"eval_steps_per_second": 3.93,
|
| 984 |
+
"step": 330
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 4.39,
|
| 988 |
+
"eval_loss": 0.6366411447525024,
|
| 989 |
+
"eval_runtime": 23.9159,
|
| 990 |
+
"eval_samples_per_second": 125.44,
|
| 991 |
+
"eval_steps_per_second": 3.93,
|
| 992 |
+
"step": 333
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 4.43,
|
| 996 |
+
"eval_loss": 0.6357526183128357,
|
| 997 |
+
"eval_runtime": 23.9135,
|
| 998 |
+
"eval_samples_per_second": 125.452,
|
| 999 |
+
"eval_steps_per_second": 3.931,
|
| 1000 |
+
"step": 336
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 4.47,
|
| 1004 |
+
"eval_loss": 0.6349912881851196,
|
| 1005 |
+
"eval_runtime": 23.9211,
|
| 1006 |
+
"eval_samples_per_second": 125.412,
|
| 1007 |
+
"eval_steps_per_second": 3.93,
|
| 1008 |
+
"step": 339
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 4.49,
|
| 1012 |
+
"learning_rate": 0.00011036339165545088,
|
| 1013 |
+
"loss": 0.6303,
|
| 1014 |
+
"step": 340
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 4.51,
|
| 1018 |
+
"eval_loss": 0.6343324184417725,
|
| 1019 |
+
"eval_runtime": 23.927,
|
| 1020 |
+
"eval_samples_per_second": 125.381,
|
| 1021 |
+
"eval_steps_per_second": 3.929,
|
| 1022 |
+
"step": 342
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 4.55,
|
| 1026 |
+
"eval_loss": 0.6347218751907349,
|
| 1027 |
+
"eval_runtime": 23.9489,
|
| 1028 |
+
"eval_samples_per_second": 125.267,
|
| 1029 |
+
"eval_steps_per_second": 3.925,
|
| 1030 |
+
"step": 345
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 4.59,
|
| 1034 |
+
"eval_loss": 0.6333290338516235,
|
| 1035 |
+
"eval_runtime": 23.9573,
|
| 1036 |
+
"eval_samples_per_second": 125.223,
|
| 1037 |
+
"eval_steps_per_second": 3.924,
|
| 1038 |
+
"step": 348
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 4.63,
|
| 1042 |
+
"eval_loss": 0.6328045129776001,
|
| 1043 |
+
"eval_runtime": 23.925,
|
| 1044 |
+
"eval_samples_per_second": 125.392,
|
| 1045 |
+
"eval_steps_per_second": 3.929,
|
| 1046 |
+
"step": 351
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.67,
|
| 1050 |
+
"eval_loss": 0.6328830718994141,
|
| 1051 |
+
"eval_runtime": 23.9277,
|
| 1052 |
+
"eval_samples_per_second": 125.378,
|
| 1053 |
+
"eval_steps_per_second": 3.928,
|
| 1054 |
+
"step": 354
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 4.71,
|
| 1058 |
+
"eval_loss": 0.6323109269142151,
|
| 1059 |
+
"eval_runtime": 23.9385,
|
| 1060 |
+
"eval_samples_per_second": 125.321,
|
| 1061 |
+
"eval_steps_per_second": 3.927,
|
| 1062 |
+
"step": 357
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 4.75,
|
| 1066 |
+
"learning_rate": 0.00010497981157469719,
|
| 1067 |
+
"loss": 0.6268,
|
| 1068 |
+
"step": 360
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 4.75,
|
| 1072 |
+
"eval_loss": 0.6327587366104126,
|
| 1073 |
+
"eval_runtime": 23.9389,
|
| 1074 |
+
"eval_samples_per_second": 125.319,
|
| 1075 |
+
"eval_steps_per_second": 3.927,
|
| 1076 |
+
"step": 360
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 4.79,
|
| 1080 |
+
"eval_loss": 0.6324266791343689,
|
| 1081 |
+
"eval_runtime": 23.9367,
|
| 1082 |
+
"eval_samples_per_second": 125.331,
|
| 1083 |
+
"eval_steps_per_second": 3.927,
|
| 1084 |
+
"step": 363
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 4.83,
|
| 1088 |
+
"eval_loss": 0.6320524215698242,
|
| 1089 |
+
"eval_runtime": 23.9373,
|
| 1090 |
+
"eval_samples_per_second": 125.327,
|
| 1091 |
+
"eval_steps_per_second": 3.927,
|
| 1092 |
+
"step": 366
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 4.87,
|
| 1096 |
+
"eval_loss": 0.6314539313316345,
|
| 1097 |
+
"eval_runtime": 23.9325,
|
| 1098 |
+
"eval_samples_per_second": 125.352,
|
| 1099 |
+
"eval_steps_per_second": 3.928,
|
| 1100 |
+
"step": 369
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 4.91,
|
| 1104 |
+
"eval_loss": 0.6318089365959167,
|
| 1105 |
+
"eval_runtime": 23.9345,
|
| 1106 |
+
"eval_samples_per_second": 125.342,
|
| 1107 |
+
"eval_steps_per_second": 3.927,
|
| 1108 |
+
"step": 372
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.95,
|
| 1112 |
+
"eval_loss": 0.6315808296203613,
|
| 1113 |
+
"eval_runtime": 23.924,
|
| 1114 |
+
"eval_samples_per_second": 125.397,
|
| 1115 |
+
"eval_steps_per_second": 3.929,
|
| 1116 |
+
"step": 375
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 4.99,
|
| 1120 |
+
"eval_loss": 0.630818247795105,
|
| 1121 |
+
"eval_runtime": 23.9285,
|
| 1122 |
+
"eval_samples_per_second": 125.373,
|
| 1123 |
+
"eval_steps_per_second": 3.928,
|
| 1124 |
+
"step": 378
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.01,
|
| 1128 |
+
"learning_rate": 9.959623149394348e-05,
|
| 1129 |
+
"loss": 0.6196,
|
| 1130 |
+
"step": 380
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 5.03,
|
| 1134 |
+
"eval_loss": 0.630248486995697,
|
| 1135 |
+
"eval_runtime": 23.9231,
|
| 1136 |
+
"eval_samples_per_second": 125.402,
|
| 1137 |
+
"eval_steps_per_second": 3.929,
|
| 1138 |
+
"step": 381
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 5.07,
|
| 1142 |
+
"eval_loss": 0.6306143403053284,
|
| 1143 |
+
"eval_runtime": 23.9242,
|
| 1144 |
+
"eval_samples_per_second": 125.396,
|
| 1145 |
+
"eval_steps_per_second": 3.929,
|
| 1146 |
+
"step": 384
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 5.11,
|
| 1150 |
+
"eval_loss": 0.6305729746818542,
|
| 1151 |
+
"eval_runtime": 23.9232,
|
| 1152 |
+
"eval_samples_per_second": 125.401,
|
| 1153 |
+
"eval_steps_per_second": 3.929,
|
| 1154 |
+
"step": 387
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 5.15,
|
| 1158 |
+
"eval_loss": 0.6302648782730103,
|
| 1159 |
+
"eval_runtime": 23.9286,
|
| 1160 |
+
"eval_samples_per_second": 125.373,
|
| 1161 |
+
"eval_steps_per_second": 3.928,
|
| 1162 |
+
"step": 390
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 5.19,
|
| 1166 |
+
"eval_loss": 0.6298710703849792,
|
| 1167 |
+
"eval_runtime": 23.9258,
|
| 1168 |
+
"eval_samples_per_second": 125.388,
|
| 1169 |
+
"eval_steps_per_second": 3.929,
|
| 1170 |
+
"step": 393
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 5.23,
|
| 1174 |
+
"eval_loss": 0.6298263669013977,
|
| 1175 |
+
"eval_runtime": 23.9284,
|
| 1176 |
+
"eval_samples_per_second": 125.374,
|
| 1177 |
+
"eval_steps_per_second": 3.928,
|
| 1178 |
+
"step": 396
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 5.27,
|
| 1182 |
+
"eval_loss": 0.6292470097541809,
|
| 1183 |
+
"eval_runtime": 23.9269,
|
| 1184 |
+
"eval_samples_per_second": 125.382,
|
| 1185 |
+
"eval_steps_per_second": 3.929,
|
| 1186 |
+
"step": 399
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 5.28,
|
| 1190 |
+
"learning_rate": 9.421265141318977e-05,
|
| 1191 |
+
"loss": 0.6146,
|
| 1192 |
+
"step": 400
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 5.3,
|
| 1196 |
+
"eval_loss": 0.6291049122810364,
|
| 1197 |
+
"eval_runtime": 23.9297,
|
| 1198 |
+
"eval_samples_per_second": 125.367,
|
| 1199 |
+
"eval_steps_per_second": 3.928,
|
| 1200 |
+
"step": 402
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 5.34,
|
| 1204 |
+
"eval_loss": 0.6296722292900085,
|
| 1205 |
+
"eval_runtime": 23.9386,
|
| 1206 |
+
"eval_samples_per_second": 125.321,
|
| 1207 |
+
"eval_steps_per_second": 3.927,
|
| 1208 |
+
"step": 405
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 5.38,
|
| 1212 |
+
"eval_loss": 0.6288275122642517,
|
| 1213 |
+
"eval_runtime": 23.9308,
|
| 1214 |
+
"eval_samples_per_second": 125.362,
|
| 1215 |
+
"eval_steps_per_second": 3.928,
|
| 1216 |
+
"step": 408
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 5.42,
|
| 1220 |
+
"eval_loss": 0.6288333535194397,
|
| 1221 |
+
"eval_runtime": 23.9261,
|
| 1222 |
+
"eval_samples_per_second": 125.386,
|
| 1223 |
+
"eval_steps_per_second": 3.929,
|
| 1224 |
+
"step": 411
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 5.46,
|
| 1228 |
+
"eval_loss": 0.6279690861701965,
|
| 1229 |
+
"eval_runtime": 23.9282,
|
| 1230 |
+
"eval_samples_per_second": 125.375,
|
| 1231 |
+
"eval_steps_per_second": 3.928,
|
| 1232 |
+
"step": 414
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 5.5,
|
| 1236 |
+
"eval_loss": 0.6275332570075989,
|
| 1237 |
+
"eval_runtime": 23.9215,
|
| 1238 |
+
"eval_samples_per_second": 125.41,
|
| 1239 |
+
"eval_steps_per_second": 3.93,
|
| 1240 |
+
"step": 417
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 5.54,
|
| 1244 |
+
"learning_rate": 8.882907133243608e-05,
|
| 1245 |
+
"loss": 0.6149,
|
| 1246 |
+
"step": 420
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 5.54,
|
| 1250 |
+
"eval_loss": 0.6279338598251343,
|
| 1251 |
+
"eval_runtime": 23.93,
|
| 1252 |
+
"eval_samples_per_second": 125.366,
|
| 1253 |
+
"eval_steps_per_second": 3.928,
|
| 1254 |
+
"step": 420
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 5.58,
|
| 1258 |
+
"eval_loss": 0.6271057724952698,
|
| 1259 |
+
"eval_runtime": 23.9158,
|
| 1260 |
+
"eval_samples_per_second": 125.44,
|
| 1261 |
+
"eval_steps_per_second": 3.93,
|
| 1262 |
+
"step": 423
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 5.62,
|
| 1266 |
+
"eval_loss": 0.6270298361778259,
|
| 1267 |
+
"eval_runtime": 23.9264,
|
| 1268 |
+
"eval_samples_per_second": 125.384,
|
| 1269 |
+
"eval_steps_per_second": 3.929,
|
| 1270 |
+
"step": 426
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 5.66,
|
| 1274 |
+
"eval_loss": 0.6271407604217529,
|
| 1275 |
+
"eval_runtime": 23.9362,
|
| 1276 |
+
"eval_samples_per_second": 125.333,
|
| 1277 |
+
"eval_steps_per_second": 3.927,
|
| 1278 |
+
"step": 429
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 5.7,
|
| 1282 |
+
"eval_loss": 0.6264240145683289,
|
| 1283 |
+
"eval_runtime": 23.9309,
|
| 1284 |
+
"eval_samples_per_second": 125.361,
|
| 1285 |
+
"eval_steps_per_second": 3.928,
|
| 1286 |
+
"step": 432
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 5.74,
|
| 1290 |
+
"eval_loss": 0.6263339519500732,
|
| 1291 |
+
"eval_runtime": 23.93,
|
| 1292 |
+
"eval_samples_per_second": 125.366,
|
| 1293 |
+
"eval_steps_per_second": 3.928,
|
| 1294 |
+
"step": 435
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 5.78,
|
| 1298 |
+
"eval_loss": 0.6256468296051025,
|
| 1299 |
+
"eval_runtime": 23.9252,
|
| 1300 |
+
"eval_samples_per_second": 125.391,
|
| 1301 |
+
"eval_steps_per_second": 3.929,
|
| 1302 |
+
"step": 438
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 5.81,
|
| 1306 |
+
"learning_rate": 8.344549125168237e-05,
|
| 1307 |
+
"loss": 0.6191,
|
| 1308 |
+
"step": 440
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 5.82,
|
| 1312 |
+
"eval_loss": 0.6260586977005005,
|
| 1313 |
+
"eval_runtime": 23.918,
|
| 1314 |
+
"eval_samples_per_second": 125.429,
|
| 1315 |
+
"eval_steps_per_second": 3.93,
|
| 1316 |
+
"step": 441
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 5.86,
|
| 1320 |
+
"eval_loss": 0.625337541103363,
|
| 1321 |
+
"eval_runtime": 23.9172,
|
| 1322 |
+
"eval_samples_per_second": 125.433,
|
| 1323 |
+
"eval_steps_per_second": 3.93,
|
| 1324 |
+
"step": 444
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 5.9,
|
| 1328 |
+
"eval_loss": 0.6246620416641235,
|
| 1329 |
+
"eval_runtime": 23.908,
|
| 1330 |
+
"eval_samples_per_second": 125.481,
|
| 1331 |
+
"eval_steps_per_second": 3.932,
|
| 1332 |
+
"step": 447
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 5.94,
|
| 1336 |
+
"eval_loss": 0.6251673698425293,
|
| 1337 |
+
"eval_runtime": 23.9137,
|
| 1338 |
+
"eval_samples_per_second": 125.451,
|
| 1339 |
+
"eval_steps_per_second": 3.931,
|
| 1340 |
+
"step": 450
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 5.98,
|
| 1344 |
+
"eval_loss": 0.6253092288970947,
|
| 1345 |
+
"eval_runtime": 23.9409,
|
| 1346 |
+
"eval_samples_per_second": 125.309,
|
| 1347 |
+
"eval_steps_per_second": 3.926,
|
| 1348 |
+
"step": 453
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 6.02,
|
| 1352 |
+
"eval_loss": 0.6245599389076233,
|
| 1353 |
+
"eval_runtime": 23.9233,
|
| 1354 |
+
"eval_samples_per_second": 125.401,
|
| 1355 |
+
"eval_steps_per_second": 3.929,
|
| 1356 |
+
"step": 456
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 6.06,
|
| 1360 |
+
"eval_loss": 0.6247097849845886,
|
| 1361 |
+
"eval_runtime": 23.9184,
|
| 1362 |
+
"eval_samples_per_second": 125.426,
|
| 1363 |
+
"eval_steps_per_second": 3.93,
|
| 1364 |
+
"step": 459
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 6.07,
|
| 1368 |
+
"learning_rate": 7.806191117092868e-05,
|
| 1369 |
+
"loss": 0.6075,
|
| 1370 |
+
"step": 460
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 6.1,
|
| 1374 |
+
"eval_loss": 0.6237714290618896,
|
| 1375 |
+
"eval_runtime": 23.9137,
|
| 1376 |
+
"eval_samples_per_second": 125.451,
|
| 1377 |
+
"eval_steps_per_second": 3.931,
|
| 1378 |
+
"step": 462
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 6.14,
|
| 1382 |
+
"eval_loss": 0.6239632964134216,
|
| 1383 |
+
"eval_runtime": 23.9142,
|
| 1384 |
+
"eval_samples_per_second": 125.449,
|
| 1385 |
+
"eval_steps_per_second": 3.931,
|
| 1386 |
+
"step": 465
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 6.18,
|
| 1390 |
+
"eval_loss": 0.6246253252029419,
|
| 1391 |
+
"eval_runtime": 23.9199,
|
| 1392 |
+
"eval_samples_per_second": 125.419,
|
| 1393 |
+
"eval_steps_per_second": 3.93,
|
| 1394 |
+
"step": 468
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 6.22,
|
| 1398 |
+
"eval_loss": 0.6236398220062256,
|
| 1399 |
+
"eval_runtime": 23.9169,
|
| 1400 |
+
"eval_samples_per_second": 125.434,
|
| 1401 |
+
"eval_steps_per_second": 3.93,
|
| 1402 |
+
"step": 471
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 6.25,
|
| 1406 |
+
"eval_loss": 0.6242309808731079,
|
| 1407 |
+
"eval_runtime": 23.9091,
|
| 1408 |
+
"eval_samples_per_second": 125.475,
|
| 1409 |
+
"eval_steps_per_second": 3.932,
|
| 1410 |
+
"step": 474
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 6.29,
|
| 1414 |
+
"eval_loss": 0.6236902475357056,
|
| 1415 |
+
"eval_runtime": 23.9117,
|
| 1416 |
+
"eval_samples_per_second": 125.462,
|
| 1417 |
+
"eval_steps_per_second": 3.931,
|
| 1418 |
+
"step": 477
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 6.33,
|
| 1422 |
+
"learning_rate": 7.267833109017497e-05,
|
| 1423 |
+
"loss": 0.6061,
|
| 1424 |
+
"step": 480
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 6.33,
|
| 1428 |
+
"eval_loss": 0.623267650604248,
|
| 1429 |
+
"eval_runtime": 23.9071,
|
| 1430 |
+
"eval_samples_per_second": 125.485,
|
| 1431 |
+
"eval_steps_per_second": 3.932,
|
| 1432 |
+
"step": 480
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 6.37,
|
| 1436 |
+
"eval_loss": 0.6238719820976257,
|
| 1437 |
+
"eval_runtime": 23.9206,
|
| 1438 |
+
"eval_samples_per_second": 125.415,
|
| 1439 |
+
"eval_steps_per_second": 3.93,
|
| 1440 |
+
"step": 483
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 6.41,
|
| 1444 |
+
"eval_loss": 0.6234752535820007,
|
| 1445 |
+
"eval_runtime": 23.914,
|
| 1446 |
+
"eval_samples_per_second": 125.449,
|
| 1447 |
+
"eval_steps_per_second": 3.931,
|
| 1448 |
+
"step": 486
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 6.45,
|
| 1452 |
+
"eval_loss": 0.6228368878364563,
|
| 1453 |
+
"eval_runtime": 23.9087,
|
| 1454 |
+
"eval_samples_per_second": 125.477,
|
| 1455 |
+
"eval_steps_per_second": 3.932,
|
| 1456 |
+
"step": 489
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 6.49,
|
| 1460 |
+
"eval_loss": 0.6226744055747986,
|
| 1461 |
+
"eval_runtime": 23.9118,
|
| 1462 |
+
"eval_samples_per_second": 125.461,
|
| 1463 |
+
"eval_steps_per_second": 3.931,
|
| 1464 |
+
"step": 492
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 6.53,
|
| 1468 |
+
"eval_loss": 0.622622013092041,
|
| 1469 |
+
"eval_runtime": 23.9341,
|
| 1470 |
+
"eval_samples_per_second": 125.344,
|
| 1471 |
+
"eval_steps_per_second": 3.927,
|
| 1472 |
+
"step": 495
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 6.57,
|
| 1476 |
+
"eval_loss": 0.6228298544883728,
|
| 1477 |
+
"eval_runtime": 23.9079,
|
| 1478 |
+
"eval_samples_per_second": 125.482,
|
| 1479 |
+
"eval_steps_per_second": 3.932,
|
| 1480 |
+
"step": 498
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 6.6,
|
| 1484 |
+
"learning_rate": 6.729475100942126e-05,
|
| 1485 |
+
"loss": 0.6043,
|
| 1486 |
+
"step": 500
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 6.61,
|
| 1490 |
+
"eval_loss": 0.6232237815856934,
|
| 1491 |
+
"eval_runtime": 23.8982,
|
| 1492 |
+
"eval_samples_per_second": 125.533,
|
| 1493 |
+
"eval_steps_per_second": 3.933,
|
| 1494 |
+
"step": 501
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 6.65,
|
| 1498 |
+
"eval_loss": 0.6218205690383911,
|
| 1499 |
+
"eval_runtime": 23.9059,
|
| 1500 |
+
"eval_samples_per_second": 125.492,
|
| 1501 |
+
"eval_steps_per_second": 3.932,
|
| 1502 |
+
"step": 504
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 6.69,
|
| 1506 |
+
"eval_loss": 0.621903657913208,
|
| 1507 |
+
"eval_runtime": 23.8991,
|
| 1508 |
+
"eval_samples_per_second": 125.528,
|
| 1509 |
+
"eval_steps_per_second": 3.933,
|
| 1510 |
+
"step": 507
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 6.73,
|
| 1514 |
+
"eval_loss": 0.622235894203186,
|
| 1515 |
+
"eval_runtime": 23.9024,
|
| 1516 |
+
"eval_samples_per_second": 125.51,
|
| 1517 |
+
"eval_steps_per_second": 3.933,
|
| 1518 |
+
"step": 510
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 6.77,
|
| 1522 |
+
"eval_loss": 0.6220830082893372,
|
| 1523 |
+
"eval_runtime": 23.8926,
|
| 1524 |
+
"eval_samples_per_second": 125.562,
|
| 1525 |
+
"eval_steps_per_second": 3.934,
|
| 1526 |
+
"step": 513
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 6.81,
|
| 1530 |
+
"eval_loss": 0.6220167875289917,
|
| 1531 |
+
"eval_runtime": 23.8965,
|
| 1532 |
+
"eval_samples_per_second": 125.542,
|
| 1533 |
+
"eval_steps_per_second": 3.934,
|
| 1534 |
+
"step": 516
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 6.85,
|
| 1538 |
+
"eval_loss": 0.6222782135009766,
|
| 1539 |
+
"eval_runtime": 23.908,
|
| 1540 |
+
"eval_samples_per_second": 125.481,
|
| 1541 |
+
"eval_steps_per_second": 3.932,
|
| 1542 |
+
"step": 519
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 6.86,
|
| 1546 |
+
"learning_rate": 6.191117092866757e-05,
|
| 1547 |
+
"loss": 0.6008,
|
| 1548 |
+
"step": 520
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 6.89,
|
| 1552 |
+
"eval_loss": 0.6216304302215576,
|
| 1553 |
+
"eval_runtime": 23.9036,
|
| 1554 |
+
"eval_samples_per_second": 125.504,
|
| 1555 |
+
"eval_steps_per_second": 3.932,
|
| 1556 |
+
"step": 522
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 6.93,
|
| 1560 |
+
"eval_loss": 0.6217759847640991,
|
| 1561 |
+
"eval_runtime": 23.9088,
|
| 1562 |
+
"eval_samples_per_second": 125.477,
|
| 1563 |
+
"eval_steps_per_second": 3.932,
|
| 1564 |
+
"step": 525
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 6.97,
|
| 1568 |
+
"eval_loss": 0.6214317083358765,
|
| 1569 |
+
"eval_runtime": 23.9177,
|
| 1570 |
+
"eval_samples_per_second": 125.43,
|
| 1571 |
+
"eval_steps_per_second": 3.93,
|
| 1572 |
+
"step": 528
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 7.01,
|
| 1576 |
+
"eval_loss": 0.6213416457176208,
|
| 1577 |
+
"eval_runtime": 23.9138,
|
| 1578 |
+
"eval_samples_per_second": 125.451,
|
| 1579 |
+
"eval_steps_per_second": 3.931,
|
| 1580 |
+
"step": 531
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 7.05,
|
| 1584 |
+
"eval_loss": 0.6217712163925171,
|
| 1585 |
+
"eval_runtime": 23.9141,
|
| 1586 |
+
"eval_samples_per_second": 125.449,
|
| 1587 |
+
"eval_steps_per_second": 3.931,
|
| 1588 |
+
"step": 534
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 7.09,
|
| 1592 |
+
"eval_loss": 0.6215860843658447,
|
| 1593 |
+
"eval_runtime": 23.9145,
|
| 1594 |
+
"eval_samples_per_second": 125.447,
|
| 1595 |
+
"eval_steps_per_second": 3.931,
|
| 1596 |
+
"step": 537
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 7.13,
|
| 1600 |
+
"learning_rate": 5.652759084791387e-05,
|
| 1601 |
+
"loss": 0.599,
|
| 1602 |
+
"step": 540
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 7.13,
|
| 1606 |
+
"eval_loss": 0.6211041808128357,
|
| 1607 |
+
"eval_runtime": 23.9125,
|
| 1608 |
+
"eval_samples_per_second": 125.457,
|
| 1609 |
+
"eval_steps_per_second": 3.931,
|
| 1610 |
+
"step": 540
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 7.17,
|
| 1614 |
+
"eval_loss": 0.6210355758666992,
|
| 1615 |
+
"eval_runtime": 23.911,
|
| 1616 |
+
"eval_samples_per_second": 125.465,
|
| 1617 |
+
"eval_steps_per_second": 3.931,
|
| 1618 |
+
"step": 543
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 7.2,
|
| 1622 |
+
"eval_loss": 0.6209889650344849,
|
| 1623 |
+
"eval_runtime": 23.9062,
|
| 1624 |
+
"eval_samples_per_second": 125.491,
|
| 1625 |
+
"eval_steps_per_second": 3.932,
|
| 1626 |
+
"step": 546
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 7.24,
|
| 1630 |
+
"eval_loss": 0.6205114126205444,
|
| 1631 |
+
"eval_runtime": 23.9227,
|
| 1632 |
+
"eval_samples_per_second": 125.404,
|
| 1633 |
+
"eval_steps_per_second": 3.929,
|
| 1634 |
+
"step": 549
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 7.28,
|
| 1638 |
+
"eval_loss": 0.6204013824462891,
|
| 1639 |
+
"eval_runtime": 23.9146,
|
| 1640 |
+
"eval_samples_per_second": 125.446,
|
| 1641 |
+
"eval_steps_per_second": 3.931,
|
| 1642 |
+
"step": 552
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 7.32,
|
| 1646 |
+
"eval_loss": 0.6202988028526306,
|
| 1647 |
+
"eval_runtime": 23.9015,
|
| 1648 |
+
"eval_samples_per_second": 125.515,
|
| 1649 |
+
"eval_steps_per_second": 3.933,
|
| 1650 |
+
"step": 555
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 7.36,
|
| 1654 |
+
"eval_loss": 0.6199727654457092,
|
| 1655 |
+
"eval_runtime": 23.9089,
|
| 1656 |
+
"eval_samples_per_second": 125.476,
|
| 1657 |
+
"eval_steps_per_second": 3.932,
|
| 1658 |
+
"step": 558
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 7.39,
|
| 1662 |
+
"learning_rate": 5.1144010767160164e-05,
|
| 1663 |
+
"loss": 0.5959,
|
| 1664 |
+
"step": 560
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 7.4,
|
| 1668 |
+
"eval_loss": 0.619968831539154,
|
| 1669 |
+
"eval_runtime": 23.9043,
|
| 1670 |
+
"eval_samples_per_second": 125.5,
|
| 1671 |
+
"eval_steps_per_second": 3.932,
|
| 1672 |
+
"step": 561
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 7.44,
|
| 1676 |
+
"eval_loss": 0.6202374696731567,
|
| 1677 |
+
"eval_runtime": 23.9117,
|
| 1678 |
+
"eval_samples_per_second": 125.461,
|
| 1679 |
+
"eval_steps_per_second": 3.931,
|
| 1680 |
+
"step": 564
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 7.48,
|
| 1684 |
+
"eval_loss": 0.6202066540718079,
|
| 1685 |
+
"eval_runtime": 23.908,
|
| 1686 |
+
"eval_samples_per_second": 125.481,
|
| 1687 |
+
"eval_steps_per_second": 3.932,
|
| 1688 |
+
"step": 567
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 7.52,
|
| 1692 |
+
"eval_loss": 0.6198835968971252,
|
| 1693 |
+
"eval_runtime": 23.9244,
|
| 1694 |
+
"eval_samples_per_second": 125.395,
|
| 1695 |
+
"eval_steps_per_second": 3.929,
|
| 1696 |
+
"step": 570
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 7.56,
|
| 1700 |
+
"eval_loss": 0.6199198961257935,
|
| 1701 |
+
"eval_runtime": 23.9263,
|
| 1702 |
+
"eval_samples_per_second": 125.385,
|
| 1703 |
+
"eval_steps_per_second": 3.929,
|
| 1704 |
+
"step": 573
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 7.6,
|
| 1708 |
+
"eval_loss": 0.6195517182350159,
|
| 1709 |
+
"eval_runtime": 23.9125,
|
| 1710 |
+
"eval_samples_per_second": 125.457,
|
| 1711 |
+
"eval_steps_per_second": 3.931,
|
| 1712 |
+
"step": 576
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 7.64,
|
| 1716 |
+
"eval_loss": 0.6192638278007507,
|
| 1717 |
+
"eval_runtime": 23.9168,
|
| 1718 |
+
"eval_samples_per_second": 125.435,
|
| 1719 |
+
"eval_steps_per_second": 3.93,
|
| 1720 |
+
"step": 579
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 7.65,
|
| 1724 |
+
"learning_rate": 4.576043068640646e-05,
|
| 1725 |
+
"loss": 0.5922,
|
| 1726 |
+
"step": 580
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 7.68,
|
| 1730 |
+
"eval_loss": 0.6196587085723877,
|
| 1731 |
+
"eval_runtime": 23.9305,
|
| 1732 |
+
"eval_samples_per_second": 125.363,
|
| 1733 |
+
"eval_steps_per_second": 3.928,
|
| 1734 |
+
"step": 582
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 7.72,
|
| 1738 |
+
"eval_loss": 0.6198856830596924,
|
| 1739 |
+
"eval_runtime": 23.919,
|
| 1740 |
+
"eval_samples_per_second": 125.423,
|
| 1741 |
+
"eval_steps_per_second": 3.93,
|
| 1742 |
+
"step": 585
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 7.76,
|
| 1746 |
+
"eval_loss": 0.6196783781051636,
|
| 1747 |
+
"eval_runtime": 23.9042,
|
| 1748 |
+
"eval_samples_per_second": 125.501,
|
| 1749 |
+
"eval_steps_per_second": 3.932,
|
| 1750 |
+
"step": 588
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 7.8,
|
| 1754 |
+
"eval_loss": 0.6192678809165955,
|
| 1755 |
+
"eval_runtime": 23.9057,
|
| 1756 |
+
"eval_samples_per_second": 125.493,
|
| 1757 |
+
"eval_steps_per_second": 3.932,
|
| 1758 |
+
"step": 591
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 7.84,
|
| 1762 |
+
"eval_loss": 0.6192264556884766,
|
| 1763 |
+
"eval_runtime": 23.909,
|
| 1764 |
+
"eval_samples_per_second": 125.476,
|
| 1765 |
+
"eval_steps_per_second": 3.932,
|
| 1766 |
+
"step": 594
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 7.88,
|
| 1770 |
+
"eval_loss": 0.6192458271980286,
|
| 1771 |
+
"eval_runtime": 23.8917,
|
| 1772 |
+
"eval_samples_per_second": 125.567,
|
| 1773 |
+
"eval_steps_per_second": 3.934,
|
| 1774 |
+
"step": 597
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 7.92,
|
| 1778 |
+
"learning_rate": 4.037685060565276e-05,
|
| 1779 |
+
"loss": 0.6028,
|
| 1780 |
+
"step": 600
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 7.92,
|
| 1784 |
+
"eval_loss": 0.6192883849143982,
|
| 1785 |
+
"eval_runtime": 23.9005,
|
| 1786 |
+
"eval_samples_per_second": 125.521,
|
| 1787 |
+
"eval_steps_per_second": 3.933,
|
| 1788 |
+
"step": 600
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 7.96,
|
| 1792 |
+
"eval_loss": 0.6194872856140137,
|
| 1793 |
+
"eval_runtime": 23.9118,
|
| 1794 |
+
"eval_samples_per_second": 125.461,
|
| 1795 |
+
"eval_steps_per_second": 3.931,
|
| 1796 |
+
"step": 603
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 8.0,
|
| 1800 |
+
"eval_loss": 0.619368314743042,
|
| 1801 |
+
"eval_runtime": 23.9121,
|
| 1802 |
+
"eval_samples_per_second": 125.46,
|
| 1803 |
+
"eval_steps_per_second": 3.931,
|
| 1804 |
+
"step": 606
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 8.04,
|
| 1808 |
+
"eval_loss": 0.6190740466117859,
|
| 1809 |
+
"eval_runtime": 23.9192,
|
| 1810 |
+
"eval_samples_per_second": 125.422,
|
| 1811 |
+
"eval_steps_per_second": 3.93,
|
| 1812 |
+
"step": 609
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 8.08,
|
| 1816 |
+
"eval_loss": 0.6190269589424133,
|
| 1817 |
+
"eval_runtime": 23.916,
|
| 1818 |
+
"eval_samples_per_second": 125.439,
|
| 1819 |
+
"eval_steps_per_second": 3.93,
|
| 1820 |
+
"step": 612
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 8.12,
|
| 1824 |
+
"eval_loss": 0.6189839839935303,
|
| 1825 |
+
"eval_runtime": 23.9063,
|
| 1826 |
+
"eval_samples_per_second": 125.49,
|
| 1827 |
+
"eval_steps_per_second": 3.932,
|
| 1828 |
+
"step": 615
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 8.16,
|
| 1832 |
+
"eval_loss": 0.618523895740509,
|
| 1833 |
+
"eval_runtime": 23.8928,
|
| 1834 |
+
"eval_samples_per_second": 125.561,
|
| 1835 |
+
"eval_steps_per_second": 3.934,
|
| 1836 |
+
"step": 618
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 8.18,
|
| 1840 |
+
"learning_rate": 3.499327052489906e-05,
|
| 1841 |
+
"loss": 0.5941,
|
| 1842 |
+
"step": 620
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 8.19,
|
| 1846 |
+
"eval_loss": 0.6187476515769958,
|
| 1847 |
+
"eval_runtime": 23.9016,
|
| 1848 |
+
"eval_samples_per_second": 125.515,
|
| 1849 |
+
"eval_steps_per_second": 3.933,
|
| 1850 |
+
"step": 621
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 8.23,
|
| 1854 |
+
"eval_loss": 0.6186578869819641,
|
| 1855 |
+
"eval_runtime": 23.9003,
|
| 1856 |
+
"eval_samples_per_second": 125.521,
|
| 1857 |
+
"eval_steps_per_second": 3.933,
|
| 1858 |
+
"step": 624
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 8.27,
|
| 1862 |
+
"eval_loss": 0.6190162897109985,
|
| 1863 |
+
"eval_runtime": 23.9155,
|
| 1864 |
+
"eval_samples_per_second": 125.442,
|
| 1865 |
+
"eval_steps_per_second": 3.931,
|
| 1866 |
+
"step": 627
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 8.31,
|
| 1870 |
+
"eval_loss": 0.6189883351325989,
|
| 1871 |
+
"eval_runtime": 23.9091,
|
| 1872 |
+
"eval_samples_per_second": 125.475,
|
| 1873 |
+
"eval_steps_per_second": 3.932,
|
| 1874 |
+
"step": 630
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 8.35,
|
| 1878 |
+
"eval_loss": 0.6184096932411194,
|
| 1879 |
+
"eval_runtime": 23.9042,
|
| 1880 |
+
"eval_samples_per_second": 125.501,
|
| 1881 |
+
"eval_steps_per_second": 3.932,
|
| 1882 |
+
"step": 633
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 8.39,
|
| 1886 |
+
"eval_loss": 0.6180031895637512,
|
| 1887 |
+
"eval_runtime": 23.9189,
|
| 1888 |
+
"eval_samples_per_second": 125.424,
|
| 1889 |
+
"eval_steps_per_second": 3.93,
|
| 1890 |
+
"step": 636
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 8.43,
|
| 1894 |
+
"eval_loss": 0.6179867386817932,
|
| 1895 |
+
"eval_runtime": 23.922,
|
| 1896 |
+
"eval_samples_per_second": 125.407,
|
| 1897 |
+
"eval_steps_per_second": 3.929,
|
| 1898 |
+
"step": 639
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 8.45,
|
| 1902 |
+
"learning_rate": 2.960969044414536e-05,
|
| 1903 |
+
"loss": 0.5906,
|
| 1904 |
+
"step": 640
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 8.47,
|
| 1908 |
+
"eval_loss": 0.6182823777198792,
|
| 1909 |
+
"eval_runtime": 23.9158,
|
| 1910 |
+
"eval_samples_per_second": 125.44,
|
| 1911 |
+
"eval_steps_per_second": 3.93,
|
| 1912 |
+
"step": 642
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 8.51,
|
| 1916 |
+
"eval_loss": 0.6179353594779968,
|
| 1917 |
+
"eval_runtime": 23.914,
|
| 1918 |
+
"eval_samples_per_second": 125.45,
|
| 1919 |
+
"eval_steps_per_second": 3.931,
|
| 1920 |
+
"step": 645
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 8.55,
|
| 1924 |
+
"eval_loss": 0.6178385615348816,
|
| 1925 |
+
"eval_runtime": 23.9079,
|
| 1926 |
+
"eval_samples_per_second": 125.482,
|
| 1927 |
+
"eval_steps_per_second": 3.932,
|
| 1928 |
+
"step": 648
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 8.59,
|
| 1932 |
+
"eval_loss": 0.6179868578910828,
|
| 1933 |
+
"eval_runtime": 23.9189,
|
| 1934 |
+
"eval_samples_per_second": 125.424,
|
| 1935 |
+
"eval_steps_per_second": 3.93,
|
| 1936 |
+
"step": 651
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 8.63,
|
| 1940 |
+
"eval_loss": 0.6179595589637756,
|
| 1941 |
+
"eval_runtime": 23.9327,
|
| 1942 |
+
"eval_samples_per_second": 125.352,
|
| 1943 |
+
"eval_steps_per_second": 3.928,
|
| 1944 |
+
"step": 654
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 8.67,
|
| 1948 |
+
"eval_loss": 0.6177854537963867,
|
| 1949 |
+
"eval_runtime": 23.9123,
|
| 1950 |
+
"eval_samples_per_second": 125.458,
|
| 1951 |
+
"eval_steps_per_second": 3.931,
|
| 1952 |
+
"step": 657
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 8.71,
|
| 1956 |
+
"learning_rate": 2.422611036339166e-05,
|
| 1957 |
+
"loss": 0.5908,
|
| 1958 |
+
"step": 660
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 8.71,
|
| 1962 |
+
"eval_loss": 0.6179735064506531,
|
| 1963 |
+
"eval_runtime": 23.9135,
|
| 1964 |
+
"eval_samples_per_second": 125.452,
|
| 1965 |
+
"eval_steps_per_second": 3.931,
|
| 1966 |
+
"step": 660
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 8.75,
|
| 1970 |
+
"eval_loss": 0.6180996298789978,
|
| 1971 |
+
"eval_runtime": 23.9038,
|
| 1972 |
+
"eval_samples_per_second": 125.503,
|
| 1973 |
+
"eval_steps_per_second": 3.932,
|
| 1974 |
+
"step": 663
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 8.79,
|
| 1978 |
+
"eval_loss": 0.6181532740592957,
|
| 1979 |
+
"eval_runtime": 23.9058,
|
| 1980 |
+
"eval_samples_per_second": 125.492,
|
| 1981 |
+
"eval_steps_per_second": 3.932,
|
| 1982 |
+
"step": 666
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 8.83,
|
| 1986 |
+
"eval_loss": 0.6176265478134155,
|
| 1987 |
+
"eval_runtime": 23.9128,
|
| 1988 |
+
"eval_samples_per_second": 125.456,
|
| 1989 |
+
"eval_steps_per_second": 3.931,
|
| 1990 |
+
"step": 669
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 8.87,
|
| 1994 |
+
"eval_loss": 0.617388904094696,
|
| 1995 |
+
"eval_runtime": 23.8906,
|
| 1996 |
+
"eval_samples_per_second": 125.572,
|
| 1997 |
+
"eval_steps_per_second": 3.935,
|
| 1998 |
+
"step": 672
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 8.91,
|
| 2002 |
+
"eval_loss": 0.6174699068069458,
|
| 2003 |
+
"eval_runtime": 23.9051,
|
| 2004 |
+
"eval_samples_per_second": 125.496,
|
| 2005 |
+
"eval_steps_per_second": 3.932,
|
| 2006 |
+
"step": 675
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 8.95,
|
| 2010 |
+
"eval_loss": 0.6174932718276978,
|
| 2011 |
+
"eval_runtime": 23.8969,
|
| 2012 |
+
"eval_samples_per_second": 125.54,
|
| 2013 |
+
"eval_steps_per_second": 3.934,
|
| 2014 |
+
"step": 678
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 8.97,
|
| 2018 |
+
"learning_rate": 1.8842530282637954e-05,
|
| 2019 |
+
"loss": 0.5937,
|
| 2020 |
+
"step": 680
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 8.99,
|
| 2024 |
+
"eval_loss": 0.6174784898757935,
|
| 2025 |
+
"eval_runtime": 23.9056,
|
| 2026 |
+
"eval_samples_per_second": 125.494,
|
| 2027 |
+
"eval_steps_per_second": 3.932,
|
| 2028 |
+
"step": 681
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 9.03,
|
| 2032 |
+
"eval_loss": 0.6175104975700378,
|
| 2033 |
+
"eval_runtime": 23.9136,
|
| 2034 |
+
"eval_samples_per_second": 125.452,
|
| 2035 |
+
"eval_steps_per_second": 3.931,
|
| 2036 |
+
"step": 684
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 9.07,
|
| 2040 |
+
"eval_loss": 0.6173563599586487,
|
| 2041 |
+
"eval_runtime": 23.9237,
|
| 2042 |
+
"eval_samples_per_second": 125.399,
|
| 2043 |
+
"eval_steps_per_second": 3.929,
|
| 2044 |
+
"step": 687
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 9.11,
|
| 2048 |
+
"eval_loss": 0.6172643899917603,
|
| 2049 |
+
"eval_runtime": 23.9021,
|
| 2050 |
+
"eval_samples_per_second": 125.512,
|
| 2051 |
+
"eval_steps_per_second": 3.933,
|
| 2052 |
+
"step": 690
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 9.14,
|
| 2056 |
+
"eval_loss": 0.6172318458557129,
|
| 2057 |
+
"eval_runtime": 23.9242,
|
| 2058 |
+
"eval_samples_per_second": 125.396,
|
| 2059 |
+
"eval_steps_per_second": 3.929,
|
| 2060 |
+
"step": 693
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 9.18,
|
| 2064 |
+
"eval_loss": 0.617030680179596,
|
| 2065 |
+
"eval_runtime": 23.9184,
|
| 2066 |
+
"eval_samples_per_second": 125.427,
|
| 2067 |
+
"eval_steps_per_second": 3.93,
|
| 2068 |
+
"step": 696
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 9.22,
|
| 2072 |
+
"eval_loss": 0.6169764995574951,
|
| 2073 |
+
"eval_runtime": 23.9104,
|
| 2074 |
+
"eval_samples_per_second": 125.468,
|
| 2075 |
+
"eval_steps_per_second": 3.931,
|
| 2076 |
+
"step": 699
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 9.24,
|
| 2080 |
+
"learning_rate": 1.3458950201884254e-05,
|
| 2081 |
+
"loss": 0.5867,
|
| 2082 |
+
"step": 700
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 9.26,
|
| 2086 |
+
"eval_loss": 0.6171083450317383,
|
| 2087 |
+
"eval_runtime": 23.9126,
|
| 2088 |
+
"eval_samples_per_second": 125.457,
|
| 2089 |
+
"eval_steps_per_second": 3.931,
|
| 2090 |
+
"step": 702
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 9.3,
|
| 2094 |
+
"eval_loss": 0.6171473860740662,
|
| 2095 |
+
"eval_runtime": 23.9015,
|
| 2096 |
+
"eval_samples_per_second": 125.515,
|
| 2097 |
+
"eval_steps_per_second": 3.933,
|
| 2098 |
+
"step": 705
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 9.34,
|
| 2102 |
+
"eval_loss": 0.6170982122421265,
|
| 2103 |
+
"eval_runtime": 23.9023,
|
| 2104 |
+
"eval_samples_per_second": 125.511,
|
| 2105 |
+
"eval_steps_per_second": 3.933,
|
| 2106 |
+
"step": 708
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 9.38,
|
| 2110 |
+
"eval_loss": 0.6169420480728149,
|
| 2111 |
+
"eval_runtime": 23.9074,
|
| 2112 |
+
"eval_samples_per_second": 125.484,
|
| 2113 |
+
"eval_steps_per_second": 3.932,
|
| 2114 |
+
"step": 711
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 9.42,
|
| 2118 |
+
"eval_loss": 0.6168730854988098,
|
| 2119 |
+
"eval_runtime": 23.9047,
|
| 2120 |
+
"eval_samples_per_second": 125.498,
|
| 2121 |
+
"eval_steps_per_second": 3.932,
|
| 2122 |
+
"step": 714
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 9.46,
|
| 2126 |
+
"eval_loss": 0.6168663501739502,
|
| 2127 |
+
"eval_runtime": 23.9161,
|
| 2128 |
+
"eval_samples_per_second": 125.439,
|
| 2129 |
+
"eval_steps_per_second": 3.93,
|
| 2130 |
+
"step": 717
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 9.5,
|
| 2134 |
+
"learning_rate": 8.075370121130552e-06,
|
| 2135 |
+
"loss": 0.5891,
|
| 2136 |
+
"step": 720
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 9.5,
|
| 2140 |
+
"eval_loss": 0.6167708039283752,
|
| 2141 |
+
"eval_runtime": 23.9114,
|
| 2142 |
+
"eval_samples_per_second": 125.463,
|
| 2143 |
+
"eval_steps_per_second": 3.931,
|
| 2144 |
+
"step": 720
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 9.54,
|
| 2148 |
+
"eval_loss": 0.6167441010475159,
|
| 2149 |
+
"eval_runtime": 23.9086,
|
| 2150 |
+
"eval_samples_per_second": 125.478,
|
| 2151 |
+
"eval_steps_per_second": 3.932,
|
| 2152 |
+
"step": 723
|
| 2153 |
+
}
|
| 2154 |
+
],
|
| 2155 |
+
"max_steps": 750,
|
| 2156 |
+
"num_train_epochs": 10,
|
| 2157 |
+
"total_flos": 7.640934816744997e+18,
|
| 2158 |
+
"trial_name": null,
|
| 2159 |
+
"trial_params": null
|
| 2160 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d5a386ce8addef927bca0e390e9534e0877a3e4e00f222f83967dd78c49527a
|
| 3 |
+
size 4027
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/adapter_config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "/mnt/data1/sheshuaijie/Data/PLM/vicuna-7b",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"fan_in_fan_out": false,
|
| 5 |
+
"inference_mode": true,
|
| 6 |
+
"init_lora_weights": true,
|
| 7 |
+
"lora_alpha": 32,
|
| 8 |
+
"lora_dropout": 0.1,
|
| 9 |
+
"modules_to_save": null,
|
| 10 |
+
"peft_type": "LORA",
|
| 11 |
+
"r": 32,
|
| 12 |
+
"target_modules": [
|
| 13 |
+
"q_proj",
|
| 14 |
+
"v_proj"
|
| 15 |
+
],
|
| 16 |
+
"task_type": "CAUSAL_LM"
|
| 17 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea0e58084f5c10cf3198c266194c6df608b78ef7e6bec56f2f2b2f79df7e8203
|
| 3 |
+
size 134293701
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:839608048df3cb485a5a46df3fefc629f75b7a28bc82368b2b0a3f241428230a
|
| 3 |
+
size 67154893
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:662a8264d6501814b20daf7fc26f37adcc57cdcddf3f93b6635ccf090850f087
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da501a848bbcc5c6d41d8d84bf7b38dac81809aa86b59f1ccf322ab2ea71ea07
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c884cfa9665b1300551525efbbf6b6a770236f7c0054643e8504b79dc00b8df
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17fa542e5ca43798e13f106972566af70af0bf43dd5709d5b87e499a7242fcdc
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85ae2de3ba224920992db9fbe1ca60ba4dc264aff58b35933f515fc1ba4b30c1
|
| 3 |
+
size 557
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3cad30e23141c113d7a635066ba568eb5da726ce1057d35c2c89d31da5cfb390
|
| 3 |
+
size 627
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/trainer_state.json
ADDED
|
@@ -0,0 +1,2200 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6167441010475159,
|
| 3 |
+
"best_model_checkpoint": "/mnt/data1/sheshuaijie/Output/CoT/Trained/vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723",
|
| 4 |
+
"epoch": 9.738556701030928,
|
| 5 |
+
"global_step": 738,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"eval_loss": 1.7524008750915527,
|
| 13 |
+
"eval_runtime": 23.7951,
|
| 14 |
+
"eval_samples_per_second": 126.077,
|
| 15 |
+
"eval_steps_per_second": 3.95,
|
| 16 |
+
"step": 3
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.08,
|
| 20 |
+
"eval_loss": 1.5515066385269165,
|
| 21 |
+
"eval_runtime": 23.828,
|
| 22 |
+
"eval_samples_per_second": 125.902,
|
| 23 |
+
"eval_steps_per_second": 3.945,
|
| 24 |
+
"step": 6
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12,
|
| 28 |
+
"eval_loss": 1.3584641218185425,
|
| 29 |
+
"eval_runtime": 23.8775,
|
| 30 |
+
"eval_samples_per_second": 125.641,
|
| 31 |
+
"eval_steps_per_second": 3.937,
|
| 32 |
+
"step": 9
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"eval_loss": 1.2644726037979126,
|
| 37 |
+
"eval_runtime": 23.8942,
|
| 38 |
+
"eval_samples_per_second": 125.554,
|
| 39 |
+
"eval_steps_per_second": 3.934,
|
| 40 |
+
"step": 12
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.2,
|
| 44 |
+
"eval_loss": 1.166400671005249,
|
| 45 |
+
"eval_runtime": 23.9181,
|
| 46 |
+
"eval_samples_per_second": 125.428,
|
| 47 |
+
"eval_steps_per_second": 3.93,
|
| 48 |
+
"step": 15
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.24,
|
| 52 |
+
"eval_loss": 1.1086052656173706,
|
| 53 |
+
"eval_runtime": 23.9166,
|
| 54 |
+
"eval_samples_per_second": 125.436,
|
| 55 |
+
"eval_steps_per_second": 3.93,
|
| 56 |
+
"step": 18
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.26,
|
| 60 |
+
"learning_rate": 0.00019650067294751011,
|
| 61 |
+
"loss": 1.4265,
|
| 62 |
+
"step": 20
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.28,
|
| 66 |
+
"eval_loss": 1.0677987337112427,
|
| 67 |
+
"eval_runtime": 23.9189,
|
| 68 |
+
"eval_samples_per_second": 125.424,
|
| 69 |
+
"eval_steps_per_second": 3.93,
|
| 70 |
+
"step": 21
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.32,
|
| 74 |
+
"eval_loss": 1.0342437028884888,
|
| 75 |
+
"eval_runtime": 23.9046,
|
| 76 |
+
"eval_samples_per_second": 125.499,
|
| 77 |
+
"eval_steps_per_second": 3.932,
|
| 78 |
+
"step": 24
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.36,
|
| 82 |
+
"eval_loss": 0.9985266923904419,
|
| 83 |
+
"eval_runtime": 23.9037,
|
| 84 |
+
"eval_samples_per_second": 125.504,
|
| 85 |
+
"eval_steps_per_second": 3.932,
|
| 86 |
+
"step": 27
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.4,
|
| 90 |
+
"eval_loss": 0.9654523134231567,
|
| 91 |
+
"eval_runtime": 23.9129,
|
| 92 |
+
"eval_samples_per_second": 125.455,
|
| 93 |
+
"eval_steps_per_second": 3.931,
|
| 94 |
+
"step": 30
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.44,
|
| 98 |
+
"eval_loss": 0.939262866973877,
|
| 99 |
+
"eval_runtime": 23.9117,
|
| 100 |
+
"eval_samples_per_second": 125.462,
|
| 101 |
+
"eval_steps_per_second": 3.931,
|
| 102 |
+
"step": 33
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.48,
|
| 106 |
+
"eval_loss": 0.9186767339706421,
|
| 107 |
+
"eval_runtime": 23.9011,
|
| 108 |
+
"eval_samples_per_second": 125.517,
|
| 109 |
+
"eval_steps_per_second": 3.933,
|
| 110 |
+
"step": 36
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.51,
|
| 114 |
+
"eval_loss": 0.8969741463661194,
|
| 115 |
+
"eval_runtime": 23.9105,
|
| 116 |
+
"eval_samples_per_second": 125.468,
|
| 117 |
+
"eval_steps_per_second": 3.931,
|
| 118 |
+
"step": 39
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.53,
|
| 122 |
+
"learning_rate": 0.00019111709286675642,
|
| 123 |
+
"loss": 0.9923,
|
| 124 |
+
"step": 40
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.55,
|
| 128 |
+
"eval_loss": 0.8814375996589661,
|
| 129 |
+
"eval_runtime": 23.9154,
|
| 130 |
+
"eval_samples_per_second": 125.442,
|
| 131 |
+
"eval_steps_per_second": 3.931,
|
| 132 |
+
"step": 42
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 0.59,
|
| 136 |
+
"eval_loss": 0.8654683232307434,
|
| 137 |
+
"eval_runtime": 23.9108,
|
| 138 |
+
"eval_samples_per_second": 125.466,
|
| 139 |
+
"eval_steps_per_second": 3.931,
|
| 140 |
+
"step": 45
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.63,
|
| 144 |
+
"eval_loss": 0.852226734161377,
|
| 145 |
+
"eval_runtime": 23.9186,
|
| 146 |
+
"eval_samples_per_second": 125.425,
|
| 147 |
+
"eval_steps_per_second": 3.93,
|
| 148 |
+
"step": 48
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.67,
|
| 152 |
+
"eval_loss": 0.839223325252533,
|
| 153 |
+
"eval_runtime": 23.9074,
|
| 154 |
+
"eval_samples_per_second": 125.484,
|
| 155 |
+
"eval_steps_per_second": 3.932,
|
| 156 |
+
"step": 51
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.71,
|
| 160 |
+
"eval_loss": 0.8266379237174988,
|
| 161 |
+
"eval_runtime": 23.9399,
|
| 162 |
+
"eval_samples_per_second": 125.314,
|
| 163 |
+
"eval_steps_per_second": 3.926,
|
| 164 |
+
"step": 54
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.75,
|
| 168 |
+
"eval_loss": 0.8140417337417603,
|
| 169 |
+
"eval_runtime": 23.9355,
|
| 170 |
+
"eval_samples_per_second": 125.337,
|
| 171 |
+
"eval_steps_per_second": 3.927,
|
| 172 |
+
"step": 57
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.79,
|
| 176 |
+
"learning_rate": 0.0001857335127860027,
|
| 177 |
+
"loss": 0.8611,
|
| 178 |
+
"step": 60
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.79,
|
| 182 |
+
"eval_loss": 0.8019057512283325,
|
| 183 |
+
"eval_runtime": 23.9223,
|
| 184 |
+
"eval_samples_per_second": 125.406,
|
| 185 |
+
"eval_steps_per_second": 3.929,
|
| 186 |
+
"step": 60
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.83,
|
| 190 |
+
"eval_loss": 0.7907609343528748,
|
| 191 |
+
"eval_runtime": 23.9384,
|
| 192 |
+
"eval_samples_per_second": 125.322,
|
| 193 |
+
"eval_steps_per_second": 3.927,
|
| 194 |
+
"step": 63
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.87,
|
| 198 |
+
"eval_loss": 0.7791212797164917,
|
| 199 |
+
"eval_runtime": 23.9101,
|
| 200 |
+
"eval_samples_per_second": 125.47,
|
| 201 |
+
"eval_steps_per_second": 3.931,
|
| 202 |
+
"step": 66
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.91,
|
| 206 |
+
"eval_loss": 0.7694615125656128,
|
| 207 |
+
"eval_runtime": 23.9079,
|
| 208 |
+
"eval_samples_per_second": 125.481,
|
| 209 |
+
"eval_steps_per_second": 3.932,
|
| 210 |
+
"step": 69
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.95,
|
| 214 |
+
"eval_loss": 0.7602358460426331,
|
| 215 |
+
"eval_runtime": 23.9116,
|
| 216 |
+
"eval_samples_per_second": 125.462,
|
| 217 |
+
"eval_steps_per_second": 3.931,
|
| 218 |
+
"step": 72
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.99,
|
| 222 |
+
"eval_loss": 0.753226101398468,
|
| 223 |
+
"eval_runtime": 23.9242,
|
| 224 |
+
"eval_samples_per_second": 125.396,
|
| 225 |
+
"eval_steps_per_second": 3.929,
|
| 226 |
+
"step": 75
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.03,
|
| 230 |
+
"eval_loss": 0.7466432452201843,
|
| 231 |
+
"eval_runtime": 23.9116,
|
| 232 |
+
"eval_samples_per_second": 125.462,
|
| 233 |
+
"eval_steps_per_second": 3.931,
|
| 234 |
+
"step": 78
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.06,
|
| 238 |
+
"learning_rate": 0.000180349932705249,
|
| 239 |
+
"loss": 0.7843,
|
| 240 |
+
"step": 80
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.07,
|
| 244 |
+
"eval_loss": 0.7416810989379883,
|
| 245 |
+
"eval_runtime": 23.9171,
|
| 246 |
+
"eval_samples_per_second": 125.433,
|
| 247 |
+
"eval_steps_per_second": 3.93,
|
| 248 |
+
"step": 81
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.11,
|
| 252 |
+
"eval_loss": 0.7362396121025085,
|
| 253 |
+
"eval_runtime": 23.9079,
|
| 254 |
+
"eval_samples_per_second": 125.481,
|
| 255 |
+
"eval_steps_per_second": 3.932,
|
| 256 |
+
"step": 84
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.15,
|
| 260 |
+
"eval_loss": 0.7297741174697876,
|
| 261 |
+
"eval_runtime": 23.9084,
|
| 262 |
+
"eval_samples_per_second": 125.479,
|
| 263 |
+
"eval_steps_per_second": 3.932,
|
| 264 |
+
"step": 87
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.19,
|
| 268 |
+
"eval_loss": 0.7252654433250427,
|
| 269 |
+
"eval_runtime": 23.9206,
|
| 270 |
+
"eval_samples_per_second": 125.415,
|
| 271 |
+
"eval_steps_per_second": 3.93,
|
| 272 |
+
"step": 90
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.23,
|
| 276 |
+
"eval_loss": 0.7213409543037415,
|
| 277 |
+
"eval_runtime": 23.9179,
|
| 278 |
+
"eval_samples_per_second": 125.429,
|
| 279 |
+
"eval_steps_per_second": 3.93,
|
| 280 |
+
"step": 93
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.27,
|
| 284 |
+
"eval_loss": 0.7174035906791687,
|
| 285 |
+
"eval_runtime": 23.9354,
|
| 286 |
+
"eval_samples_per_second": 125.337,
|
| 287 |
+
"eval_steps_per_second": 3.927,
|
| 288 |
+
"step": 96
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.31,
|
| 292 |
+
"eval_loss": 0.7140380144119263,
|
| 293 |
+
"eval_runtime": 23.9214,
|
| 294 |
+
"eval_samples_per_second": 125.411,
|
| 295 |
+
"eval_steps_per_second": 3.93,
|
| 296 |
+
"step": 99
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.32,
|
| 300 |
+
"learning_rate": 0.0001749663526244953,
|
| 301 |
+
"loss": 0.7301,
|
| 302 |
+
"step": 100
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.35,
|
| 306 |
+
"eval_loss": 0.7104487419128418,
|
| 307 |
+
"eval_runtime": 23.9093,
|
| 308 |
+
"eval_samples_per_second": 125.474,
|
| 309 |
+
"eval_steps_per_second": 3.932,
|
| 310 |
+
"step": 102
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.39,
|
| 314 |
+
"eval_loss": 0.7067868113517761,
|
| 315 |
+
"eval_runtime": 23.9129,
|
| 316 |
+
"eval_samples_per_second": 125.455,
|
| 317 |
+
"eval_steps_per_second": 3.931,
|
| 318 |
+
"step": 105
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.43,
|
| 322 |
+
"eval_loss": 0.7041762471199036,
|
| 323 |
+
"eval_runtime": 23.9161,
|
| 324 |
+
"eval_samples_per_second": 125.439,
|
| 325 |
+
"eval_steps_per_second": 3.93,
|
| 326 |
+
"step": 108
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.46,
|
| 330 |
+
"eval_loss": 0.7013522982597351,
|
| 331 |
+
"eval_runtime": 23.9133,
|
| 332 |
+
"eval_samples_per_second": 125.453,
|
| 333 |
+
"eval_steps_per_second": 3.931,
|
| 334 |
+
"step": 111
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.5,
|
| 338 |
+
"eval_loss": 0.6989504098892212,
|
| 339 |
+
"eval_runtime": 23.9152,
|
| 340 |
+
"eval_samples_per_second": 125.443,
|
| 341 |
+
"eval_steps_per_second": 3.931,
|
| 342 |
+
"step": 114
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 1.54,
|
| 346 |
+
"eval_loss": 0.6974085569381714,
|
| 347 |
+
"eval_runtime": 23.9561,
|
| 348 |
+
"eval_samples_per_second": 125.229,
|
| 349 |
+
"eval_steps_per_second": 3.924,
|
| 350 |
+
"step": 117
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.58,
|
| 354 |
+
"learning_rate": 0.0001695827725437416,
|
| 355 |
+
"loss": 0.7141,
|
| 356 |
+
"step": 120
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.58,
|
| 360 |
+
"eval_loss": 0.6944894194602966,
|
| 361 |
+
"eval_runtime": 23.902,
|
| 362 |
+
"eval_samples_per_second": 125.512,
|
| 363 |
+
"eval_steps_per_second": 3.933,
|
| 364 |
+
"step": 120
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 1.62,
|
| 368 |
+
"eval_loss": 0.6929482221603394,
|
| 369 |
+
"eval_runtime": 23.9189,
|
| 370 |
+
"eval_samples_per_second": 125.424,
|
| 371 |
+
"eval_steps_per_second": 3.93,
|
| 372 |
+
"step": 123
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.66,
|
| 376 |
+
"eval_loss": 0.6903366446495056,
|
| 377 |
+
"eval_runtime": 23.9061,
|
| 378 |
+
"eval_samples_per_second": 125.491,
|
| 379 |
+
"eval_steps_per_second": 3.932,
|
| 380 |
+
"step": 126
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.7,
|
| 384 |
+
"eval_loss": 0.6882749199867249,
|
| 385 |
+
"eval_runtime": 23.9181,
|
| 386 |
+
"eval_samples_per_second": 125.428,
|
| 387 |
+
"eval_steps_per_second": 3.93,
|
| 388 |
+
"step": 129
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.74,
|
| 392 |
+
"eval_loss": 0.6863100528717041,
|
| 393 |
+
"eval_runtime": 23.914,
|
| 394 |
+
"eval_samples_per_second": 125.45,
|
| 395 |
+
"eval_steps_per_second": 3.931,
|
| 396 |
+
"step": 132
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.78,
|
| 400 |
+
"eval_loss": 0.6860549449920654,
|
| 401 |
+
"eval_runtime": 23.9138,
|
| 402 |
+
"eval_samples_per_second": 125.45,
|
| 403 |
+
"eval_steps_per_second": 3.931,
|
| 404 |
+
"step": 135
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.82,
|
| 408 |
+
"eval_loss": 0.6831715703010559,
|
| 409 |
+
"eval_runtime": 23.9135,
|
| 410 |
+
"eval_samples_per_second": 125.452,
|
| 411 |
+
"eval_steps_per_second": 3.931,
|
| 412 |
+
"step": 138
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.85,
|
| 416 |
+
"learning_rate": 0.0001641991924629879,
|
| 417 |
+
"loss": 0.6902,
|
| 418 |
+
"step": 140
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.86,
|
| 422 |
+
"eval_loss": 0.6819499731063843,
|
| 423 |
+
"eval_runtime": 23.8986,
|
| 424 |
+
"eval_samples_per_second": 125.53,
|
| 425 |
+
"eval_steps_per_second": 3.933,
|
| 426 |
+
"step": 141
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.9,
|
| 430 |
+
"eval_loss": 0.6807693839073181,
|
| 431 |
+
"eval_runtime": 23.9169,
|
| 432 |
+
"eval_samples_per_second": 125.434,
|
| 433 |
+
"eval_steps_per_second": 3.93,
|
| 434 |
+
"step": 144
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 1.94,
|
| 438 |
+
"eval_loss": 0.6787669062614441,
|
| 439 |
+
"eval_runtime": 23.9265,
|
| 440 |
+
"eval_samples_per_second": 125.384,
|
| 441 |
+
"eval_steps_per_second": 3.929,
|
| 442 |
+
"step": 147
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.98,
|
| 446 |
+
"eval_loss": 0.6773442625999451,
|
| 447 |
+
"eval_runtime": 23.9274,
|
| 448 |
+
"eval_samples_per_second": 125.38,
|
| 449 |
+
"eval_steps_per_second": 3.929,
|
| 450 |
+
"step": 150
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.02,
|
| 454 |
+
"eval_loss": 0.6759281158447266,
|
| 455 |
+
"eval_runtime": 23.9386,
|
| 456 |
+
"eval_samples_per_second": 125.321,
|
| 457 |
+
"eval_steps_per_second": 3.927,
|
| 458 |
+
"step": 153
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.06,
|
| 462 |
+
"eval_loss": 0.6743582487106323,
|
| 463 |
+
"eval_runtime": 23.9323,
|
| 464 |
+
"eval_samples_per_second": 125.354,
|
| 465 |
+
"eval_steps_per_second": 3.928,
|
| 466 |
+
"step": 156
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 2.1,
|
| 470 |
+
"eval_loss": 0.6732926368713379,
|
| 471 |
+
"eval_runtime": 23.9145,
|
| 472 |
+
"eval_samples_per_second": 125.447,
|
| 473 |
+
"eval_steps_per_second": 3.931,
|
| 474 |
+
"step": 159
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 2.11,
|
| 478 |
+
"learning_rate": 0.0001588156123822342,
|
| 479 |
+
"loss": 0.6766,
|
| 480 |
+
"step": 160
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 2.14,
|
| 484 |
+
"eval_loss": 0.6721953749656677,
|
| 485 |
+
"eval_runtime": 23.9073,
|
| 486 |
+
"eval_samples_per_second": 125.485,
|
| 487 |
+
"eval_steps_per_second": 3.932,
|
| 488 |
+
"step": 162
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.18,
|
| 492 |
+
"eval_loss": 0.6714429259300232,
|
| 493 |
+
"eval_runtime": 23.8955,
|
| 494 |
+
"eval_samples_per_second": 125.547,
|
| 495 |
+
"eval_steps_per_second": 3.934,
|
| 496 |
+
"step": 165
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 2.22,
|
| 500 |
+
"eval_loss": 0.670035183429718,
|
| 501 |
+
"eval_runtime": 23.9431,
|
| 502 |
+
"eval_samples_per_second": 125.297,
|
| 503 |
+
"eval_steps_per_second": 3.926,
|
| 504 |
+
"step": 168
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 2.26,
|
| 508 |
+
"eval_loss": 0.6695354580879211,
|
| 509 |
+
"eval_runtime": 23.8875,
|
| 510 |
+
"eval_samples_per_second": 125.589,
|
| 511 |
+
"eval_steps_per_second": 3.935,
|
| 512 |
+
"step": 171
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.3,
|
| 516 |
+
"eval_loss": 0.6689226031303406,
|
| 517 |
+
"eval_runtime": 23.9185,
|
| 518 |
+
"eval_samples_per_second": 125.426,
|
| 519 |
+
"eval_steps_per_second": 3.93,
|
| 520 |
+
"step": 174
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.34,
|
| 524 |
+
"eval_loss": 0.6674054861068726,
|
| 525 |
+
"eval_runtime": 23.941,
|
| 526 |
+
"eval_samples_per_second": 125.308,
|
| 527 |
+
"eval_steps_per_second": 3.926,
|
| 528 |
+
"step": 177
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.38,
|
| 532 |
+
"learning_rate": 0.00015343203230148048,
|
| 533 |
+
"loss": 0.6743,
|
| 534 |
+
"step": 180
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.38,
|
| 538 |
+
"eval_loss": 0.6664847731590271,
|
| 539 |
+
"eval_runtime": 23.9211,
|
| 540 |
+
"eval_samples_per_second": 125.412,
|
| 541 |
+
"eval_steps_per_second": 3.93,
|
| 542 |
+
"step": 180
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.41,
|
| 546 |
+
"eval_loss": 0.6658627986907959,
|
| 547 |
+
"eval_runtime": 23.9247,
|
| 548 |
+
"eval_samples_per_second": 125.394,
|
| 549 |
+
"eval_steps_per_second": 3.929,
|
| 550 |
+
"step": 183
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 2.45,
|
| 554 |
+
"eval_loss": 0.664908766746521,
|
| 555 |
+
"eval_runtime": 23.9272,
|
| 556 |
+
"eval_samples_per_second": 125.38,
|
| 557 |
+
"eval_steps_per_second": 3.929,
|
| 558 |
+
"step": 186
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.49,
|
| 562 |
+
"eval_loss": 0.6638036966323853,
|
| 563 |
+
"eval_runtime": 23.9187,
|
| 564 |
+
"eval_samples_per_second": 125.425,
|
| 565 |
+
"eval_steps_per_second": 3.93,
|
| 566 |
+
"step": 189
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.53,
|
| 570 |
+
"eval_loss": 0.6625837683677673,
|
| 571 |
+
"eval_runtime": 23.9033,
|
| 572 |
+
"eval_samples_per_second": 125.506,
|
| 573 |
+
"eval_steps_per_second": 3.933,
|
| 574 |
+
"step": 192
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 2.57,
|
| 578 |
+
"eval_loss": 0.6619511842727661,
|
| 579 |
+
"eval_runtime": 23.8973,
|
| 580 |
+
"eval_samples_per_second": 125.537,
|
| 581 |
+
"eval_steps_per_second": 3.934,
|
| 582 |
+
"step": 195
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.61,
|
| 586 |
+
"eval_loss": 0.6611769199371338,
|
| 587 |
+
"eval_runtime": 23.9129,
|
| 588 |
+
"eval_samples_per_second": 125.455,
|
| 589 |
+
"eval_steps_per_second": 3.931,
|
| 590 |
+
"step": 198
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.64,
|
| 594 |
+
"learning_rate": 0.00014804845222072678,
|
| 595 |
+
"loss": 0.6615,
|
| 596 |
+
"step": 200
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.65,
|
| 600 |
+
"eval_loss": 0.6606143116950989,
|
| 601 |
+
"eval_runtime": 23.9126,
|
| 602 |
+
"eval_samples_per_second": 125.457,
|
| 603 |
+
"eval_steps_per_second": 3.931,
|
| 604 |
+
"step": 201
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 2.69,
|
| 608 |
+
"eval_loss": 0.6589743494987488,
|
| 609 |
+
"eval_runtime": 23.9135,
|
| 610 |
+
"eval_samples_per_second": 125.452,
|
| 611 |
+
"eval_steps_per_second": 3.931,
|
| 612 |
+
"step": 204
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.73,
|
| 616 |
+
"eval_loss": 0.6578481197357178,
|
| 617 |
+
"eval_runtime": 23.9217,
|
| 618 |
+
"eval_samples_per_second": 125.409,
|
| 619 |
+
"eval_steps_per_second": 3.929,
|
| 620 |
+
"step": 207
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.77,
|
| 624 |
+
"eval_loss": 0.6571096181869507,
|
| 625 |
+
"eval_runtime": 23.9415,
|
| 626 |
+
"eval_samples_per_second": 125.305,
|
| 627 |
+
"eval_steps_per_second": 3.926,
|
| 628 |
+
"step": 210
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 2.81,
|
| 632 |
+
"eval_loss": 0.656689465045929,
|
| 633 |
+
"eval_runtime": 23.9111,
|
| 634 |
+
"eval_samples_per_second": 125.465,
|
| 635 |
+
"eval_steps_per_second": 3.931,
|
| 636 |
+
"step": 213
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.85,
|
| 640 |
+
"eval_loss": 0.6556207537651062,
|
| 641 |
+
"eval_runtime": 23.9099,
|
| 642 |
+
"eval_samples_per_second": 125.471,
|
| 643 |
+
"eval_steps_per_second": 3.931,
|
| 644 |
+
"step": 216
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.89,
|
| 648 |
+
"eval_loss": 0.6546627283096313,
|
| 649 |
+
"eval_runtime": 23.9164,
|
| 650 |
+
"eval_samples_per_second": 125.437,
|
| 651 |
+
"eval_steps_per_second": 3.93,
|
| 652 |
+
"step": 219
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 2.9,
|
| 656 |
+
"learning_rate": 0.0001426648721399731,
|
| 657 |
+
"loss": 0.6564,
|
| 658 |
+
"step": 220
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 2.93,
|
| 662 |
+
"eval_loss": 0.6539400815963745,
|
| 663 |
+
"eval_runtime": 23.906,
|
| 664 |
+
"eval_samples_per_second": 125.492,
|
| 665 |
+
"eval_steps_per_second": 3.932,
|
| 666 |
+
"step": 222
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.97,
|
| 670 |
+
"eval_loss": 0.653684675693512,
|
| 671 |
+
"eval_runtime": 23.9251,
|
| 672 |
+
"eval_samples_per_second": 125.391,
|
| 673 |
+
"eval_steps_per_second": 3.929,
|
| 674 |
+
"step": 225
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.01,
|
| 678 |
+
"eval_loss": 0.6526629328727722,
|
| 679 |
+
"eval_runtime": 23.9289,
|
| 680 |
+
"eval_samples_per_second": 125.371,
|
| 681 |
+
"eval_steps_per_second": 3.928,
|
| 682 |
+
"step": 228
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 3.05,
|
| 686 |
+
"eval_loss": 0.6525079011917114,
|
| 687 |
+
"eval_runtime": 23.9193,
|
| 688 |
+
"eval_samples_per_second": 125.421,
|
| 689 |
+
"eval_steps_per_second": 3.93,
|
| 690 |
+
"step": 231
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 3.09,
|
| 694 |
+
"eval_loss": 0.6514959931373596,
|
| 695 |
+
"eval_runtime": 23.9574,
|
| 696 |
+
"eval_samples_per_second": 125.223,
|
| 697 |
+
"eval_steps_per_second": 3.924,
|
| 698 |
+
"step": 234
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.13,
|
| 702 |
+
"eval_loss": 0.6507047414779663,
|
| 703 |
+
"eval_runtime": 23.9234,
|
| 704 |
+
"eval_samples_per_second": 125.4,
|
| 705 |
+
"eval_steps_per_second": 3.929,
|
| 706 |
+
"step": 237
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 3.17,
|
| 710 |
+
"learning_rate": 0.00013728129205921937,
|
| 711 |
+
"loss": 0.6469,
|
| 712 |
+
"step": 240
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 3.17,
|
| 716 |
+
"eval_loss": 0.6504186391830444,
|
| 717 |
+
"eval_runtime": 23.937,
|
| 718 |
+
"eval_samples_per_second": 125.329,
|
| 719 |
+
"eval_steps_per_second": 3.927,
|
| 720 |
+
"step": 240
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 3.21,
|
| 724 |
+
"eval_loss": 0.6495808959007263,
|
| 725 |
+
"eval_runtime": 23.9188,
|
| 726 |
+
"eval_samples_per_second": 125.425,
|
| 727 |
+
"eval_steps_per_second": 3.93,
|
| 728 |
+
"step": 243
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.25,
|
| 732 |
+
"eval_loss": 0.649512529373169,
|
| 733 |
+
"eval_runtime": 23.9209,
|
| 734 |
+
"eval_samples_per_second": 125.413,
|
| 735 |
+
"eval_steps_per_second": 3.93,
|
| 736 |
+
"step": 246
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 3.29,
|
| 740 |
+
"eval_loss": 0.648629903793335,
|
| 741 |
+
"eval_runtime": 23.9137,
|
| 742 |
+
"eval_samples_per_second": 125.451,
|
| 743 |
+
"eval_steps_per_second": 3.931,
|
| 744 |
+
"step": 249
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 3.33,
|
| 748 |
+
"eval_loss": 0.6480894088745117,
|
| 749 |
+
"eval_runtime": 23.919,
|
| 750 |
+
"eval_samples_per_second": 125.423,
|
| 751 |
+
"eval_steps_per_second": 3.93,
|
| 752 |
+
"step": 252
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.36,
|
| 756 |
+
"eval_loss": 0.6474400758743286,
|
| 757 |
+
"eval_runtime": 23.9076,
|
| 758 |
+
"eval_samples_per_second": 125.483,
|
| 759 |
+
"eval_steps_per_second": 3.932,
|
| 760 |
+
"step": 255
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 3.4,
|
| 764 |
+
"eval_loss": 0.6468291878700256,
|
| 765 |
+
"eval_runtime": 23.9305,
|
| 766 |
+
"eval_samples_per_second": 125.363,
|
| 767 |
+
"eval_steps_per_second": 3.928,
|
| 768 |
+
"step": 258
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 3.43,
|
| 772 |
+
"learning_rate": 0.00013189771197846567,
|
| 773 |
+
"loss": 0.6463,
|
| 774 |
+
"step": 260
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 3.44,
|
| 778 |
+
"eval_loss": 0.6462663412094116,
|
| 779 |
+
"eval_runtime": 23.9359,
|
| 780 |
+
"eval_samples_per_second": 125.335,
|
| 781 |
+
"eval_steps_per_second": 3.927,
|
| 782 |
+
"step": 261
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.48,
|
| 786 |
+
"eval_loss": 0.6458565592765808,
|
| 787 |
+
"eval_runtime": 23.929,
|
| 788 |
+
"eval_samples_per_second": 125.371,
|
| 789 |
+
"eval_steps_per_second": 3.928,
|
| 790 |
+
"step": 264
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 3.52,
|
| 794 |
+
"eval_loss": 0.645412266254425,
|
| 795 |
+
"eval_runtime": 23.9362,
|
| 796 |
+
"eval_samples_per_second": 125.333,
|
| 797 |
+
"eval_steps_per_second": 3.927,
|
| 798 |
+
"step": 267
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 3.56,
|
| 802 |
+
"eval_loss": 0.6449554562568665,
|
| 803 |
+
"eval_runtime": 23.9004,
|
| 804 |
+
"eval_samples_per_second": 125.521,
|
| 805 |
+
"eval_steps_per_second": 3.933,
|
| 806 |
+
"step": 270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.6,
|
| 810 |
+
"eval_loss": 0.6443325281143188,
|
| 811 |
+
"eval_runtime": 23.9065,
|
| 812 |
+
"eval_samples_per_second": 125.489,
|
| 813 |
+
"eval_steps_per_second": 3.932,
|
| 814 |
+
"step": 273
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 3.64,
|
| 818 |
+
"eval_loss": 0.6435034871101379,
|
| 819 |
+
"eval_runtime": 23.9072,
|
| 820 |
+
"eval_samples_per_second": 125.485,
|
| 821 |
+
"eval_steps_per_second": 3.932,
|
| 822 |
+
"step": 276
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 3.68,
|
| 826 |
+
"eval_loss": 0.6433733701705933,
|
| 827 |
+
"eval_runtime": 23.9042,
|
| 828 |
+
"eval_samples_per_second": 125.501,
|
| 829 |
+
"eval_steps_per_second": 3.932,
|
| 830 |
+
"step": 279
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.69,
|
| 834 |
+
"learning_rate": 0.00012651413189771198,
|
| 835 |
+
"loss": 0.6389,
|
| 836 |
+
"step": 280
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.72,
|
| 840 |
+
"eval_loss": 0.6425070762634277,
|
| 841 |
+
"eval_runtime": 23.8874,
|
| 842 |
+
"eval_samples_per_second": 125.589,
|
| 843 |
+
"eval_steps_per_second": 3.935,
|
| 844 |
+
"step": 282
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 3.76,
|
| 848 |
+
"eval_loss": 0.642119288444519,
|
| 849 |
+
"eval_runtime": 23.9328,
|
| 850 |
+
"eval_samples_per_second": 125.351,
|
| 851 |
+
"eval_steps_per_second": 3.928,
|
| 852 |
+
"step": 285
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 3.8,
|
| 856 |
+
"eval_loss": 0.641748309135437,
|
| 857 |
+
"eval_runtime": 23.9294,
|
| 858 |
+
"eval_samples_per_second": 125.369,
|
| 859 |
+
"eval_steps_per_second": 3.928,
|
| 860 |
+
"step": 288
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.84,
|
| 864 |
+
"eval_loss": 0.640826404094696,
|
| 865 |
+
"eval_runtime": 23.9434,
|
| 866 |
+
"eval_samples_per_second": 125.296,
|
| 867 |
+
"eval_steps_per_second": 3.926,
|
| 868 |
+
"step": 291
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 3.88,
|
| 872 |
+
"eval_loss": 0.6402388215065002,
|
| 873 |
+
"eval_runtime": 23.9162,
|
| 874 |
+
"eval_samples_per_second": 125.438,
|
| 875 |
+
"eval_steps_per_second": 3.93,
|
| 876 |
+
"step": 294
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 3.92,
|
| 880 |
+
"eval_loss": 0.6407353281974792,
|
| 881 |
+
"eval_runtime": 23.9121,
|
| 882 |
+
"eval_samples_per_second": 125.46,
|
| 883 |
+
"eval_steps_per_second": 3.931,
|
| 884 |
+
"step": 297
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.96,
|
| 888 |
+
"learning_rate": 0.0001211305518169583,
|
| 889 |
+
"loss": 0.6318,
|
| 890 |
+
"step": 300
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.96,
|
| 894 |
+
"eval_loss": 0.6398600935935974,
|
| 895 |
+
"eval_runtime": 23.9229,
|
| 896 |
+
"eval_samples_per_second": 125.403,
|
| 897 |
+
"eval_steps_per_second": 3.929,
|
| 898 |
+
"step": 300
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 4.0,
|
| 902 |
+
"eval_loss": 0.6393464207649231,
|
| 903 |
+
"eval_runtime": 23.9187,
|
| 904 |
+
"eval_samples_per_second": 125.425,
|
| 905 |
+
"eval_steps_per_second": 3.93,
|
| 906 |
+
"step": 303
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 4.04,
|
| 910 |
+
"eval_loss": 0.6392526626586914,
|
| 911 |
+
"eval_runtime": 23.9074,
|
| 912 |
+
"eval_samples_per_second": 125.484,
|
| 913 |
+
"eval_steps_per_second": 3.932,
|
| 914 |
+
"step": 306
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.08,
|
| 918 |
+
"eval_loss": 0.6389594078063965,
|
| 919 |
+
"eval_runtime": 23.918,
|
| 920 |
+
"eval_samples_per_second": 125.428,
|
| 921 |
+
"eval_steps_per_second": 3.93,
|
| 922 |
+
"step": 309
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 4.12,
|
| 926 |
+
"eval_loss": 0.6388808488845825,
|
| 927 |
+
"eval_runtime": 23.9158,
|
| 928 |
+
"eval_samples_per_second": 125.44,
|
| 929 |
+
"eval_steps_per_second": 3.93,
|
| 930 |
+
"step": 312
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 4.16,
|
| 934 |
+
"eval_loss": 0.6384025812149048,
|
| 935 |
+
"eval_runtime": 23.9176,
|
| 936 |
+
"eval_samples_per_second": 125.431,
|
| 937 |
+
"eval_steps_per_second": 3.93,
|
| 938 |
+
"step": 315
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.2,
|
| 942 |
+
"eval_loss": 0.6387144923210144,
|
| 943 |
+
"eval_runtime": 23.9047,
|
| 944 |
+
"eval_samples_per_second": 125.498,
|
| 945 |
+
"eval_steps_per_second": 3.932,
|
| 946 |
+
"step": 318
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 4.22,
|
| 950 |
+
"learning_rate": 0.00011574697173620459,
|
| 951 |
+
"loss": 0.6277,
|
| 952 |
+
"step": 320
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 4.24,
|
| 956 |
+
"eval_loss": 0.6377059817314148,
|
| 957 |
+
"eval_runtime": 23.9246,
|
| 958 |
+
"eval_samples_per_second": 125.394,
|
| 959 |
+
"eval_steps_per_second": 3.929,
|
| 960 |
+
"step": 321
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 4.28,
|
| 964 |
+
"eval_loss": 0.636981189250946,
|
| 965 |
+
"eval_runtime": 23.9459,
|
| 966 |
+
"eval_samples_per_second": 125.283,
|
| 967 |
+
"eval_steps_per_second": 3.926,
|
| 968 |
+
"step": 324
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 4.32,
|
| 972 |
+
"eval_loss": 0.6364036202430725,
|
| 973 |
+
"eval_runtime": 23.9206,
|
| 974 |
+
"eval_samples_per_second": 125.415,
|
| 975 |
+
"eval_steps_per_second": 3.93,
|
| 976 |
+
"step": 327
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 4.35,
|
| 980 |
+
"eval_loss": 0.6357031464576721,
|
| 981 |
+
"eval_runtime": 23.9187,
|
| 982 |
+
"eval_samples_per_second": 125.425,
|
| 983 |
+
"eval_steps_per_second": 3.93,
|
| 984 |
+
"step": 330
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 4.39,
|
| 988 |
+
"eval_loss": 0.6366411447525024,
|
| 989 |
+
"eval_runtime": 23.9159,
|
| 990 |
+
"eval_samples_per_second": 125.44,
|
| 991 |
+
"eval_steps_per_second": 3.93,
|
| 992 |
+
"step": 333
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 4.43,
|
| 996 |
+
"eval_loss": 0.6357526183128357,
|
| 997 |
+
"eval_runtime": 23.9135,
|
| 998 |
+
"eval_samples_per_second": 125.452,
|
| 999 |
+
"eval_steps_per_second": 3.931,
|
| 1000 |
+
"step": 336
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 4.47,
|
| 1004 |
+
"eval_loss": 0.6349912881851196,
|
| 1005 |
+
"eval_runtime": 23.9211,
|
| 1006 |
+
"eval_samples_per_second": 125.412,
|
| 1007 |
+
"eval_steps_per_second": 3.93,
|
| 1008 |
+
"step": 339
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 4.49,
|
| 1012 |
+
"learning_rate": 0.00011036339165545088,
|
| 1013 |
+
"loss": 0.6303,
|
| 1014 |
+
"step": 340
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 4.51,
|
| 1018 |
+
"eval_loss": 0.6343324184417725,
|
| 1019 |
+
"eval_runtime": 23.927,
|
| 1020 |
+
"eval_samples_per_second": 125.381,
|
| 1021 |
+
"eval_steps_per_second": 3.929,
|
| 1022 |
+
"step": 342
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 4.55,
|
| 1026 |
+
"eval_loss": 0.6347218751907349,
|
| 1027 |
+
"eval_runtime": 23.9489,
|
| 1028 |
+
"eval_samples_per_second": 125.267,
|
| 1029 |
+
"eval_steps_per_second": 3.925,
|
| 1030 |
+
"step": 345
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 4.59,
|
| 1034 |
+
"eval_loss": 0.6333290338516235,
|
| 1035 |
+
"eval_runtime": 23.9573,
|
| 1036 |
+
"eval_samples_per_second": 125.223,
|
| 1037 |
+
"eval_steps_per_second": 3.924,
|
| 1038 |
+
"step": 348
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 4.63,
|
| 1042 |
+
"eval_loss": 0.6328045129776001,
|
| 1043 |
+
"eval_runtime": 23.925,
|
| 1044 |
+
"eval_samples_per_second": 125.392,
|
| 1045 |
+
"eval_steps_per_second": 3.929,
|
| 1046 |
+
"step": 351
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.67,
|
| 1050 |
+
"eval_loss": 0.6328830718994141,
|
| 1051 |
+
"eval_runtime": 23.9277,
|
| 1052 |
+
"eval_samples_per_second": 125.378,
|
| 1053 |
+
"eval_steps_per_second": 3.928,
|
| 1054 |
+
"step": 354
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 4.71,
|
| 1058 |
+
"eval_loss": 0.6323109269142151,
|
| 1059 |
+
"eval_runtime": 23.9385,
|
| 1060 |
+
"eval_samples_per_second": 125.321,
|
| 1061 |
+
"eval_steps_per_second": 3.927,
|
| 1062 |
+
"step": 357
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 4.75,
|
| 1066 |
+
"learning_rate": 0.00010497981157469719,
|
| 1067 |
+
"loss": 0.6268,
|
| 1068 |
+
"step": 360
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 4.75,
|
| 1072 |
+
"eval_loss": 0.6327587366104126,
|
| 1073 |
+
"eval_runtime": 23.9389,
|
| 1074 |
+
"eval_samples_per_second": 125.319,
|
| 1075 |
+
"eval_steps_per_second": 3.927,
|
| 1076 |
+
"step": 360
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 4.79,
|
| 1080 |
+
"eval_loss": 0.6324266791343689,
|
| 1081 |
+
"eval_runtime": 23.9367,
|
| 1082 |
+
"eval_samples_per_second": 125.331,
|
| 1083 |
+
"eval_steps_per_second": 3.927,
|
| 1084 |
+
"step": 363
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 4.83,
|
| 1088 |
+
"eval_loss": 0.6320524215698242,
|
| 1089 |
+
"eval_runtime": 23.9373,
|
| 1090 |
+
"eval_samples_per_second": 125.327,
|
| 1091 |
+
"eval_steps_per_second": 3.927,
|
| 1092 |
+
"step": 366
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 4.87,
|
| 1096 |
+
"eval_loss": 0.6314539313316345,
|
| 1097 |
+
"eval_runtime": 23.9325,
|
| 1098 |
+
"eval_samples_per_second": 125.352,
|
| 1099 |
+
"eval_steps_per_second": 3.928,
|
| 1100 |
+
"step": 369
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 4.91,
|
| 1104 |
+
"eval_loss": 0.6318089365959167,
|
| 1105 |
+
"eval_runtime": 23.9345,
|
| 1106 |
+
"eval_samples_per_second": 125.342,
|
| 1107 |
+
"eval_steps_per_second": 3.927,
|
| 1108 |
+
"step": 372
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.95,
|
| 1112 |
+
"eval_loss": 0.6315808296203613,
|
| 1113 |
+
"eval_runtime": 23.924,
|
| 1114 |
+
"eval_samples_per_second": 125.397,
|
| 1115 |
+
"eval_steps_per_second": 3.929,
|
| 1116 |
+
"step": 375
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 4.99,
|
| 1120 |
+
"eval_loss": 0.630818247795105,
|
| 1121 |
+
"eval_runtime": 23.9285,
|
| 1122 |
+
"eval_samples_per_second": 125.373,
|
| 1123 |
+
"eval_steps_per_second": 3.928,
|
| 1124 |
+
"step": 378
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.01,
|
| 1128 |
+
"learning_rate": 9.959623149394348e-05,
|
| 1129 |
+
"loss": 0.6196,
|
| 1130 |
+
"step": 380
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 5.03,
|
| 1134 |
+
"eval_loss": 0.630248486995697,
|
| 1135 |
+
"eval_runtime": 23.9231,
|
| 1136 |
+
"eval_samples_per_second": 125.402,
|
| 1137 |
+
"eval_steps_per_second": 3.929,
|
| 1138 |
+
"step": 381
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 5.07,
|
| 1142 |
+
"eval_loss": 0.6306143403053284,
|
| 1143 |
+
"eval_runtime": 23.9242,
|
| 1144 |
+
"eval_samples_per_second": 125.396,
|
| 1145 |
+
"eval_steps_per_second": 3.929,
|
| 1146 |
+
"step": 384
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 5.11,
|
| 1150 |
+
"eval_loss": 0.6305729746818542,
|
| 1151 |
+
"eval_runtime": 23.9232,
|
| 1152 |
+
"eval_samples_per_second": 125.401,
|
| 1153 |
+
"eval_steps_per_second": 3.929,
|
| 1154 |
+
"step": 387
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 5.15,
|
| 1158 |
+
"eval_loss": 0.6302648782730103,
|
| 1159 |
+
"eval_runtime": 23.9286,
|
| 1160 |
+
"eval_samples_per_second": 125.373,
|
| 1161 |
+
"eval_steps_per_second": 3.928,
|
| 1162 |
+
"step": 390
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 5.19,
|
| 1166 |
+
"eval_loss": 0.6298710703849792,
|
| 1167 |
+
"eval_runtime": 23.9258,
|
| 1168 |
+
"eval_samples_per_second": 125.388,
|
| 1169 |
+
"eval_steps_per_second": 3.929,
|
| 1170 |
+
"step": 393
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 5.23,
|
| 1174 |
+
"eval_loss": 0.6298263669013977,
|
| 1175 |
+
"eval_runtime": 23.9284,
|
| 1176 |
+
"eval_samples_per_second": 125.374,
|
| 1177 |
+
"eval_steps_per_second": 3.928,
|
| 1178 |
+
"step": 396
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 5.27,
|
| 1182 |
+
"eval_loss": 0.6292470097541809,
|
| 1183 |
+
"eval_runtime": 23.9269,
|
| 1184 |
+
"eval_samples_per_second": 125.382,
|
| 1185 |
+
"eval_steps_per_second": 3.929,
|
| 1186 |
+
"step": 399
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 5.28,
|
| 1190 |
+
"learning_rate": 9.421265141318977e-05,
|
| 1191 |
+
"loss": 0.6146,
|
| 1192 |
+
"step": 400
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 5.3,
|
| 1196 |
+
"eval_loss": 0.6291049122810364,
|
| 1197 |
+
"eval_runtime": 23.9297,
|
| 1198 |
+
"eval_samples_per_second": 125.367,
|
| 1199 |
+
"eval_steps_per_second": 3.928,
|
| 1200 |
+
"step": 402
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 5.34,
|
| 1204 |
+
"eval_loss": 0.6296722292900085,
|
| 1205 |
+
"eval_runtime": 23.9386,
|
| 1206 |
+
"eval_samples_per_second": 125.321,
|
| 1207 |
+
"eval_steps_per_second": 3.927,
|
| 1208 |
+
"step": 405
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 5.38,
|
| 1212 |
+
"eval_loss": 0.6288275122642517,
|
| 1213 |
+
"eval_runtime": 23.9308,
|
| 1214 |
+
"eval_samples_per_second": 125.362,
|
| 1215 |
+
"eval_steps_per_second": 3.928,
|
| 1216 |
+
"step": 408
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 5.42,
|
| 1220 |
+
"eval_loss": 0.6288333535194397,
|
| 1221 |
+
"eval_runtime": 23.9261,
|
| 1222 |
+
"eval_samples_per_second": 125.386,
|
| 1223 |
+
"eval_steps_per_second": 3.929,
|
| 1224 |
+
"step": 411
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 5.46,
|
| 1228 |
+
"eval_loss": 0.6279690861701965,
|
| 1229 |
+
"eval_runtime": 23.9282,
|
| 1230 |
+
"eval_samples_per_second": 125.375,
|
| 1231 |
+
"eval_steps_per_second": 3.928,
|
| 1232 |
+
"step": 414
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 5.5,
|
| 1236 |
+
"eval_loss": 0.6275332570075989,
|
| 1237 |
+
"eval_runtime": 23.9215,
|
| 1238 |
+
"eval_samples_per_second": 125.41,
|
| 1239 |
+
"eval_steps_per_second": 3.93,
|
| 1240 |
+
"step": 417
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 5.54,
|
| 1244 |
+
"learning_rate": 8.882907133243608e-05,
|
| 1245 |
+
"loss": 0.6149,
|
| 1246 |
+
"step": 420
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 5.54,
|
| 1250 |
+
"eval_loss": 0.6279338598251343,
|
| 1251 |
+
"eval_runtime": 23.93,
|
| 1252 |
+
"eval_samples_per_second": 125.366,
|
| 1253 |
+
"eval_steps_per_second": 3.928,
|
| 1254 |
+
"step": 420
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 5.58,
|
| 1258 |
+
"eval_loss": 0.6271057724952698,
|
| 1259 |
+
"eval_runtime": 23.9158,
|
| 1260 |
+
"eval_samples_per_second": 125.44,
|
| 1261 |
+
"eval_steps_per_second": 3.93,
|
| 1262 |
+
"step": 423
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 5.62,
|
| 1266 |
+
"eval_loss": 0.6270298361778259,
|
| 1267 |
+
"eval_runtime": 23.9264,
|
| 1268 |
+
"eval_samples_per_second": 125.384,
|
| 1269 |
+
"eval_steps_per_second": 3.929,
|
| 1270 |
+
"step": 426
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 5.66,
|
| 1274 |
+
"eval_loss": 0.6271407604217529,
|
| 1275 |
+
"eval_runtime": 23.9362,
|
| 1276 |
+
"eval_samples_per_second": 125.333,
|
| 1277 |
+
"eval_steps_per_second": 3.927,
|
| 1278 |
+
"step": 429
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 5.7,
|
| 1282 |
+
"eval_loss": 0.6264240145683289,
|
| 1283 |
+
"eval_runtime": 23.9309,
|
| 1284 |
+
"eval_samples_per_second": 125.361,
|
| 1285 |
+
"eval_steps_per_second": 3.928,
|
| 1286 |
+
"step": 432
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 5.74,
|
| 1290 |
+
"eval_loss": 0.6263339519500732,
|
| 1291 |
+
"eval_runtime": 23.93,
|
| 1292 |
+
"eval_samples_per_second": 125.366,
|
| 1293 |
+
"eval_steps_per_second": 3.928,
|
| 1294 |
+
"step": 435
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 5.78,
|
| 1298 |
+
"eval_loss": 0.6256468296051025,
|
| 1299 |
+
"eval_runtime": 23.9252,
|
| 1300 |
+
"eval_samples_per_second": 125.391,
|
| 1301 |
+
"eval_steps_per_second": 3.929,
|
| 1302 |
+
"step": 438
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 5.81,
|
| 1306 |
+
"learning_rate": 8.344549125168237e-05,
|
| 1307 |
+
"loss": 0.6191,
|
| 1308 |
+
"step": 440
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 5.82,
|
| 1312 |
+
"eval_loss": 0.6260586977005005,
|
| 1313 |
+
"eval_runtime": 23.918,
|
| 1314 |
+
"eval_samples_per_second": 125.429,
|
| 1315 |
+
"eval_steps_per_second": 3.93,
|
| 1316 |
+
"step": 441
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 5.86,
|
| 1320 |
+
"eval_loss": 0.625337541103363,
|
| 1321 |
+
"eval_runtime": 23.9172,
|
| 1322 |
+
"eval_samples_per_second": 125.433,
|
| 1323 |
+
"eval_steps_per_second": 3.93,
|
| 1324 |
+
"step": 444
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 5.9,
|
| 1328 |
+
"eval_loss": 0.6246620416641235,
|
| 1329 |
+
"eval_runtime": 23.908,
|
| 1330 |
+
"eval_samples_per_second": 125.481,
|
| 1331 |
+
"eval_steps_per_second": 3.932,
|
| 1332 |
+
"step": 447
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 5.94,
|
| 1336 |
+
"eval_loss": 0.6251673698425293,
|
| 1337 |
+
"eval_runtime": 23.9137,
|
| 1338 |
+
"eval_samples_per_second": 125.451,
|
| 1339 |
+
"eval_steps_per_second": 3.931,
|
| 1340 |
+
"step": 450
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 5.98,
|
| 1344 |
+
"eval_loss": 0.6253092288970947,
|
| 1345 |
+
"eval_runtime": 23.9409,
|
| 1346 |
+
"eval_samples_per_second": 125.309,
|
| 1347 |
+
"eval_steps_per_second": 3.926,
|
| 1348 |
+
"step": 453
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 6.02,
|
| 1352 |
+
"eval_loss": 0.6245599389076233,
|
| 1353 |
+
"eval_runtime": 23.9233,
|
| 1354 |
+
"eval_samples_per_second": 125.401,
|
| 1355 |
+
"eval_steps_per_second": 3.929,
|
| 1356 |
+
"step": 456
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 6.06,
|
| 1360 |
+
"eval_loss": 0.6247097849845886,
|
| 1361 |
+
"eval_runtime": 23.9184,
|
| 1362 |
+
"eval_samples_per_second": 125.426,
|
| 1363 |
+
"eval_steps_per_second": 3.93,
|
| 1364 |
+
"step": 459
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 6.07,
|
| 1368 |
+
"learning_rate": 7.806191117092868e-05,
|
| 1369 |
+
"loss": 0.6075,
|
| 1370 |
+
"step": 460
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 6.1,
|
| 1374 |
+
"eval_loss": 0.6237714290618896,
|
| 1375 |
+
"eval_runtime": 23.9137,
|
| 1376 |
+
"eval_samples_per_second": 125.451,
|
| 1377 |
+
"eval_steps_per_second": 3.931,
|
| 1378 |
+
"step": 462
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 6.14,
|
| 1382 |
+
"eval_loss": 0.6239632964134216,
|
| 1383 |
+
"eval_runtime": 23.9142,
|
| 1384 |
+
"eval_samples_per_second": 125.449,
|
| 1385 |
+
"eval_steps_per_second": 3.931,
|
| 1386 |
+
"step": 465
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 6.18,
|
| 1390 |
+
"eval_loss": 0.6246253252029419,
|
| 1391 |
+
"eval_runtime": 23.9199,
|
| 1392 |
+
"eval_samples_per_second": 125.419,
|
| 1393 |
+
"eval_steps_per_second": 3.93,
|
| 1394 |
+
"step": 468
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 6.22,
|
| 1398 |
+
"eval_loss": 0.6236398220062256,
|
| 1399 |
+
"eval_runtime": 23.9169,
|
| 1400 |
+
"eval_samples_per_second": 125.434,
|
| 1401 |
+
"eval_steps_per_second": 3.93,
|
| 1402 |
+
"step": 471
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 6.25,
|
| 1406 |
+
"eval_loss": 0.6242309808731079,
|
| 1407 |
+
"eval_runtime": 23.9091,
|
| 1408 |
+
"eval_samples_per_second": 125.475,
|
| 1409 |
+
"eval_steps_per_second": 3.932,
|
| 1410 |
+
"step": 474
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 6.29,
|
| 1414 |
+
"eval_loss": 0.6236902475357056,
|
| 1415 |
+
"eval_runtime": 23.9117,
|
| 1416 |
+
"eval_samples_per_second": 125.462,
|
| 1417 |
+
"eval_steps_per_second": 3.931,
|
| 1418 |
+
"step": 477
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 6.33,
|
| 1422 |
+
"learning_rate": 7.267833109017497e-05,
|
| 1423 |
+
"loss": 0.6061,
|
| 1424 |
+
"step": 480
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 6.33,
|
| 1428 |
+
"eval_loss": 0.623267650604248,
|
| 1429 |
+
"eval_runtime": 23.9071,
|
| 1430 |
+
"eval_samples_per_second": 125.485,
|
| 1431 |
+
"eval_steps_per_second": 3.932,
|
| 1432 |
+
"step": 480
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 6.37,
|
| 1436 |
+
"eval_loss": 0.6238719820976257,
|
| 1437 |
+
"eval_runtime": 23.9206,
|
| 1438 |
+
"eval_samples_per_second": 125.415,
|
| 1439 |
+
"eval_steps_per_second": 3.93,
|
| 1440 |
+
"step": 483
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 6.41,
|
| 1444 |
+
"eval_loss": 0.6234752535820007,
|
| 1445 |
+
"eval_runtime": 23.914,
|
| 1446 |
+
"eval_samples_per_second": 125.449,
|
| 1447 |
+
"eval_steps_per_second": 3.931,
|
| 1448 |
+
"step": 486
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 6.45,
|
| 1452 |
+
"eval_loss": 0.6228368878364563,
|
| 1453 |
+
"eval_runtime": 23.9087,
|
| 1454 |
+
"eval_samples_per_second": 125.477,
|
| 1455 |
+
"eval_steps_per_second": 3.932,
|
| 1456 |
+
"step": 489
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 6.49,
|
| 1460 |
+
"eval_loss": 0.6226744055747986,
|
| 1461 |
+
"eval_runtime": 23.9118,
|
| 1462 |
+
"eval_samples_per_second": 125.461,
|
| 1463 |
+
"eval_steps_per_second": 3.931,
|
| 1464 |
+
"step": 492
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 6.53,
|
| 1468 |
+
"eval_loss": 0.622622013092041,
|
| 1469 |
+
"eval_runtime": 23.9341,
|
| 1470 |
+
"eval_samples_per_second": 125.344,
|
| 1471 |
+
"eval_steps_per_second": 3.927,
|
| 1472 |
+
"step": 495
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 6.57,
|
| 1476 |
+
"eval_loss": 0.6228298544883728,
|
| 1477 |
+
"eval_runtime": 23.9079,
|
| 1478 |
+
"eval_samples_per_second": 125.482,
|
| 1479 |
+
"eval_steps_per_second": 3.932,
|
| 1480 |
+
"step": 498
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 6.6,
|
| 1484 |
+
"learning_rate": 6.729475100942126e-05,
|
| 1485 |
+
"loss": 0.6043,
|
| 1486 |
+
"step": 500
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 6.61,
|
| 1490 |
+
"eval_loss": 0.6232237815856934,
|
| 1491 |
+
"eval_runtime": 23.8982,
|
| 1492 |
+
"eval_samples_per_second": 125.533,
|
| 1493 |
+
"eval_steps_per_second": 3.933,
|
| 1494 |
+
"step": 501
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 6.65,
|
| 1498 |
+
"eval_loss": 0.6218205690383911,
|
| 1499 |
+
"eval_runtime": 23.9059,
|
| 1500 |
+
"eval_samples_per_second": 125.492,
|
| 1501 |
+
"eval_steps_per_second": 3.932,
|
| 1502 |
+
"step": 504
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 6.69,
|
| 1506 |
+
"eval_loss": 0.621903657913208,
|
| 1507 |
+
"eval_runtime": 23.8991,
|
| 1508 |
+
"eval_samples_per_second": 125.528,
|
| 1509 |
+
"eval_steps_per_second": 3.933,
|
| 1510 |
+
"step": 507
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 6.73,
|
| 1514 |
+
"eval_loss": 0.622235894203186,
|
| 1515 |
+
"eval_runtime": 23.9024,
|
| 1516 |
+
"eval_samples_per_second": 125.51,
|
| 1517 |
+
"eval_steps_per_second": 3.933,
|
| 1518 |
+
"step": 510
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 6.77,
|
| 1522 |
+
"eval_loss": 0.6220830082893372,
|
| 1523 |
+
"eval_runtime": 23.8926,
|
| 1524 |
+
"eval_samples_per_second": 125.562,
|
| 1525 |
+
"eval_steps_per_second": 3.934,
|
| 1526 |
+
"step": 513
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 6.81,
|
| 1530 |
+
"eval_loss": 0.6220167875289917,
|
| 1531 |
+
"eval_runtime": 23.8965,
|
| 1532 |
+
"eval_samples_per_second": 125.542,
|
| 1533 |
+
"eval_steps_per_second": 3.934,
|
| 1534 |
+
"step": 516
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 6.85,
|
| 1538 |
+
"eval_loss": 0.6222782135009766,
|
| 1539 |
+
"eval_runtime": 23.908,
|
| 1540 |
+
"eval_samples_per_second": 125.481,
|
| 1541 |
+
"eval_steps_per_second": 3.932,
|
| 1542 |
+
"step": 519
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 6.86,
|
| 1546 |
+
"learning_rate": 6.191117092866757e-05,
|
| 1547 |
+
"loss": 0.6008,
|
| 1548 |
+
"step": 520
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 6.89,
|
| 1552 |
+
"eval_loss": 0.6216304302215576,
|
| 1553 |
+
"eval_runtime": 23.9036,
|
| 1554 |
+
"eval_samples_per_second": 125.504,
|
| 1555 |
+
"eval_steps_per_second": 3.932,
|
| 1556 |
+
"step": 522
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 6.93,
|
| 1560 |
+
"eval_loss": 0.6217759847640991,
|
| 1561 |
+
"eval_runtime": 23.9088,
|
| 1562 |
+
"eval_samples_per_second": 125.477,
|
| 1563 |
+
"eval_steps_per_second": 3.932,
|
| 1564 |
+
"step": 525
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 6.97,
|
| 1568 |
+
"eval_loss": 0.6214317083358765,
|
| 1569 |
+
"eval_runtime": 23.9177,
|
| 1570 |
+
"eval_samples_per_second": 125.43,
|
| 1571 |
+
"eval_steps_per_second": 3.93,
|
| 1572 |
+
"step": 528
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 7.01,
|
| 1576 |
+
"eval_loss": 0.6213416457176208,
|
| 1577 |
+
"eval_runtime": 23.9138,
|
| 1578 |
+
"eval_samples_per_second": 125.451,
|
| 1579 |
+
"eval_steps_per_second": 3.931,
|
| 1580 |
+
"step": 531
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 7.05,
|
| 1584 |
+
"eval_loss": 0.6217712163925171,
|
| 1585 |
+
"eval_runtime": 23.9141,
|
| 1586 |
+
"eval_samples_per_second": 125.449,
|
| 1587 |
+
"eval_steps_per_second": 3.931,
|
| 1588 |
+
"step": 534
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 7.09,
|
| 1592 |
+
"eval_loss": 0.6215860843658447,
|
| 1593 |
+
"eval_runtime": 23.9145,
|
| 1594 |
+
"eval_samples_per_second": 125.447,
|
| 1595 |
+
"eval_steps_per_second": 3.931,
|
| 1596 |
+
"step": 537
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 7.13,
|
| 1600 |
+
"learning_rate": 5.652759084791387e-05,
|
| 1601 |
+
"loss": 0.599,
|
| 1602 |
+
"step": 540
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 7.13,
|
| 1606 |
+
"eval_loss": 0.6211041808128357,
|
| 1607 |
+
"eval_runtime": 23.9125,
|
| 1608 |
+
"eval_samples_per_second": 125.457,
|
| 1609 |
+
"eval_steps_per_second": 3.931,
|
| 1610 |
+
"step": 540
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 7.17,
|
| 1614 |
+
"eval_loss": 0.6210355758666992,
|
| 1615 |
+
"eval_runtime": 23.911,
|
| 1616 |
+
"eval_samples_per_second": 125.465,
|
| 1617 |
+
"eval_steps_per_second": 3.931,
|
| 1618 |
+
"step": 543
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 7.2,
|
| 1622 |
+
"eval_loss": 0.6209889650344849,
|
| 1623 |
+
"eval_runtime": 23.9062,
|
| 1624 |
+
"eval_samples_per_second": 125.491,
|
| 1625 |
+
"eval_steps_per_second": 3.932,
|
| 1626 |
+
"step": 546
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 7.24,
|
| 1630 |
+
"eval_loss": 0.6205114126205444,
|
| 1631 |
+
"eval_runtime": 23.9227,
|
| 1632 |
+
"eval_samples_per_second": 125.404,
|
| 1633 |
+
"eval_steps_per_second": 3.929,
|
| 1634 |
+
"step": 549
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 7.28,
|
| 1638 |
+
"eval_loss": 0.6204013824462891,
|
| 1639 |
+
"eval_runtime": 23.9146,
|
| 1640 |
+
"eval_samples_per_second": 125.446,
|
| 1641 |
+
"eval_steps_per_second": 3.931,
|
| 1642 |
+
"step": 552
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 7.32,
|
| 1646 |
+
"eval_loss": 0.6202988028526306,
|
| 1647 |
+
"eval_runtime": 23.9015,
|
| 1648 |
+
"eval_samples_per_second": 125.515,
|
| 1649 |
+
"eval_steps_per_second": 3.933,
|
| 1650 |
+
"step": 555
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 7.36,
|
| 1654 |
+
"eval_loss": 0.6199727654457092,
|
| 1655 |
+
"eval_runtime": 23.9089,
|
| 1656 |
+
"eval_samples_per_second": 125.476,
|
| 1657 |
+
"eval_steps_per_second": 3.932,
|
| 1658 |
+
"step": 558
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 7.39,
|
| 1662 |
+
"learning_rate": 5.1144010767160164e-05,
|
| 1663 |
+
"loss": 0.5959,
|
| 1664 |
+
"step": 560
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 7.4,
|
| 1668 |
+
"eval_loss": 0.619968831539154,
|
| 1669 |
+
"eval_runtime": 23.9043,
|
| 1670 |
+
"eval_samples_per_second": 125.5,
|
| 1671 |
+
"eval_steps_per_second": 3.932,
|
| 1672 |
+
"step": 561
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 7.44,
|
| 1676 |
+
"eval_loss": 0.6202374696731567,
|
| 1677 |
+
"eval_runtime": 23.9117,
|
| 1678 |
+
"eval_samples_per_second": 125.461,
|
| 1679 |
+
"eval_steps_per_second": 3.931,
|
| 1680 |
+
"step": 564
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 7.48,
|
| 1684 |
+
"eval_loss": 0.6202066540718079,
|
| 1685 |
+
"eval_runtime": 23.908,
|
| 1686 |
+
"eval_samples_per_second": 125.481,
|
| 1687 |
+
"eval_steps_per_second": 3.932,
|
| 1688 |
+
"step": 567
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 7.52,
|
| 1692 |
+
"eval_loss": 0.6198835968971252,
|
| 1693 |
+
"eval_runtime": 23.9244,
|
| 1694 |
+
"eval_samples_per_second": 125.395,
|
| 1695 |
+
"eval_steps_per_second": 3.929,
|
| 1696 |
+
"step": 570
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 7.56,
|
| 1700 |
+
"eval_loss": 0.6199198961257935,
|
| 1701 |
+
"eval_runtime": 23.9263,
|
| 1702 |
+
"eval_samples_per_second": 125.385,
|
| 1703 |
+
"eval_steps_per_second": 3.929,
|
| 1704 |
+
"step": 573
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 7.6,
|
| 1708 |
+
"eval_loss": 0.6195517182350159,
|
| 1709 |
+
"eval_runtime": 23.9125,
|
| 1710 |
+
"eval_samples_per_second": 125.457,
|
| 1711 |
+
"eval_steps_per_second": 3.931,
|
| 1712 |
+
"step": 576
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 7.64,
|
| 1716 |
+
"eval_loss": 0.6192638278007507,
|
| 1717 |
+
"eval_runtime": 23.9168,
|
| 1718 |
+
"eval_samples_per_second": 125.435,
|
| 1719 |
+
"eval_steps_per_second": 3.93,
|
| 1720 |
+
"step": 579
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 7.65,
|
| 1724 |
+
"learning_rate": 4.576043068640646e-05,
|
| 1725 |
+
"loss": 0.5922,
|
| 1726 |
+
"step": 580
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 7.68,
|
| 1730 |
+
"eval_loss": 0.6196587085723877,
|
| 1731 |
+
"eval_runtime": 23.9305,
|
| 1732 |
+
"eval_samples_per_second": 125.363,
|
| 1733 |
+
"eval_steps_per_second": 3.928,
|
| 1734 |
+
"step": 582
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 7.72,
|
| 1738 |
+
"eval_loss": 0.6198856830596924,
|
| 1739 |
+
"eval_runtime": 23.919,
|
| 1740 |
+
"eval_samples_per_second": 125.423,
|
| 1741 |
+
"eval_steps_per_second": 3.93,
|
| 1742 |
+
"step": 585
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 7.76,
|
| 1746 |
+
"eval_loss": 0.6196783781051636,
|
| 1747 |
+
"eval_runtime": 23.9042,
|
| 1748 |
+
"eval_samples_per_second": 125.501,
|
| 1749 |
+
"eval_steps_per_second": 3.932,
|
| 1750 |
+
"step": 588
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 7.8,
|
| 1754 |
+
"eval_loss": 0.6192678809165955,
|
| 1755 |
+
"eval_runtime": 23.9057,
|
| 1756 |
+
"eval_samples_per_second": 125.493,
|
| 1757 |
+
"eval_steps_per_second": 3.932,
|
| 1758 |
+
"step": 591
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 7.84,
|
| 1762 |
+
"eval_loss": 0.6192264556884766,
|
| 1763 |
+
"eval_runtime": 23.909,
|
| 1764 |
+
"eval_samples_per_second": 125.476,
|
| 1765 |
+
"eval_steps_per_second": 3.932,
|
| 1766 |
+
"step": 594
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 7.88,
|
| 1770 |
+
"eval_loss": 0.6192458271980286,
|
| 1771 |
+
"eval_runtime": 23.8917,
|
| 1772 |
+
"eval_samples_per_second": 125.567,
|
| 1773 |
+
"eval_steps_per_second": 3.934,
|
| 1774 |
+
"step": 597
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 7.92,
|
| 1778 |
+
"learning_rate": 4.037685060565276e-05,
|
| 1779 |
+
"loss": 0.6028,
|
| 1780 |
+
"step": 600
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 7.92,
|
| 1784 |
+
"eval_loss": 0.6192883849143982,
|
| 1785 |
+
"eval_runtime": 23.9005,
|
| 1786 |
+
"eval_samples_per_second": 125.521,
|
| 1787 |
+
"eval_steps_per_second": 3.933,
|
| 1788 |
+
"step": 600
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 7.96,
|
| 1792 |
+
"eval_loss": 0.6194872856140137,
|
| 1793 |
+
"eval_runtime": 23.9118,
|
| 1794 |
+
"eval_samples_per_second": 125.461,
|
| 1795 |
+
"eval_steps_per_second": 3.931,
|
| 1796 |
+
"step": 603
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 8.0,
|
| 1800 |
+
"eval_loss": 0.619368314743042,
|
| 1801 |
+
"eval_runtime": 23.9121,
|
| 1802 |
+
"eval_samples_per_second": 125.46,
|
| 1803 |
+
"eval_steps_per_second": 3.931,
|
| 1804 |
+
"step": 606
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 8.04,
|
| 1808 |
+
"eval_loss": 0.6190740466117859,
|
| 1809 |
+
"eval_runtime": 23.9192,
|
| 1810 |
+
"eval_samples_per_second": 125.422,
|
| 1811 |
+
"eval_steps_per_second": 3.93,
|
| 1812 |
+
"step": 609
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 8.08,
|
| 1816 |
+
"eval_loss": 0.6190269589424133,
|
| 1817 |
+
"eval_runtime": 23.916,
|
| 1818 |
+
"eval_samples_per_second": 125.439,
|
| 1819 |
+
"eval_steps_per_second": 3.93,
|
| 1820 |
+
"step": 612
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 8.12,
|
| 1824 |
+
"eval_loss": 0.6189839839935303,
|
| 1825 |
+
"eval_runtime": 23.9063,
|
| 1826 |
+
"eval_samples_per_second": 125.49,
|
| 1827 |
+
"eval_steps_per_second": 3.932,
|
| 1828 |
+
"step": 615
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 8.16,
|
| 1832 |
+
"eval_loss": 0.618523895740509,
|
| 1833 |
+
"eval_runtime": 23.8928,
|
| 1834 |
+
"eval_samples_per_second": 125.561,
|
| 1835 |
+
"eval_steps_per_second": 3.934,
|
| 1836 |
+
"step": 618
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 8.18,
|
| 1840 |
+
"learning_rate": 3.499327052489906e-05,
|
| 1841 |
+
"loss": 0.5941,
|
| 1842 |
+
"step": 620
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 8.19,
|
| 1846 |
+
"eval_loss": 0.6187476515769958,
|
| 1847 |
+
"eval_runtime": 23.9016,
|
| 1848 |
+
"eval_samples_per_second": 125.515,
|
| 1849 |
+
"eval_steps_per_second": 3.933,
|
| 1850 |
+
"step": 621
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 8.23,
|
| 1854 |
+
"eval_loss": 0.6186578869819641,
|
| 1855 |
+
"eval_runtime": 23.9003,
|
| 1856 |
+
"eval_samples_per_second": 125.521,
|
| 1857 |
+
"eval_steps_per_second": 3.933,
|
| 1858 |
+
"step": 624
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 8.27,
|
| 1862 |
+
"eval_loss": 0.6190162897109985,
|
| 1863 |
+
"eval_runtime": 23.9155,
|
| 1864 |
+
"eval_samples_per_second": 125.442,
|
| 1865 |
+
"eval_steps_per_second": 3.931,
|
| 1866 |
+
"step": 627
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 8.31,
|
| 1870 |
+
"eval_loss": 0.6189883351325989,
|
| 1871 |
+
"eval_runtime": 23.9091,
|
| 1872 |
+
"eval_samples_per_second": 125.475,
|
| 1873 |
+
"eval_steps_per_second": 3.932,
|
| 1874 |
+
"step": 630
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 8.35,
|
| 1878 |
+
"eval_loss": 0.6184096932411194,
|
| 1879 |
+
"eval_runtime": 23.9042,
|
| 1880 |
+
"eval_samples_per_second": 125.501,
|
| 1881 |
+
"eval_steps_per_second": 3.932,
|
| 1882 |
+
"step": 633
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 8.39,
|
| 1886 |
+
"eval_loss": 0.6180031895637512,
|
| 1887 |
+
"eval_runtime": 23.9189,
|
| 1888 |
+
"eval_samples_per_second": 125.424,
|
| 1889 |
+
"eval_steps_per_second": 3.93,
|
| 1890 |
+
"step": 636
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 8.43,
|
| 1894 |
+
"eval_loss": 0.6179867386817932,
|
| 1895 |
+
"eval_runtime": 23.922,
|
| 1896 |
+
"eval_samples_per_second": 125.407,
|
| 1897 |
+
"eval_steps_per_second": 3.929,
|
| 1898 |
+
"step": 639
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 8.45,
|
| 1902 |
+
"learning_rate": 2.960969044414536e-05,
|
| 1903 |
+
"loss": 0.5906,
|
| 1904 |
+
"step": 640
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 8.47,
|
| 1908 |
+
"eval_loss": 0.6182823777198792,
|
| 1909 |
+
"eval_runtime": 23.9158,
|
| 1910 |
+
"eval_samples_per_second": 125.44,
|
| 1911 |
+
"eval_steps_per_second": 3.93,
|
| 1912 |
+
"step": 642
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 8.51,
|
| 1916 |
+
"eval_loss": 0.6179353594779968,
|
| 1917 |
+
"eval_runtime": 23.914,
|
| 1918 |
+
"eval_samples_per_second": 125.45,
|
| 1919 |
+
"eval_steps_per_second": 3.931,
|
| 1920 |
+
"step": 645
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 8.55,
|
| 1924 |
+
"eval_loss": 0.6178385615348816,
|
| 1925 |
+
"eval_runtime": 23.9079,
|
| 1926 |
+
"eval_samples_per_second": 125.482,
|
| 1927 |
+
"eval_steps_per_second": 3.932,
|
| 1928 |
+
"step": 648
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 8.59,
|
| 1932 |
+
"eval_loss": 0.6179868578910828,
|
| 1933 |
+
"eval_runtime": 23.9189,
|
| 1934 |
+
"eval_samples_per_second": 125.424,
|
| 1935 |
+
"eval_steps_per_second": 3.93,
|
| 1936 |
+
"step": 651
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 8.63,
|
| 1940 |
+
"eval_loss": 0.6179595589637756,
|
| 1941 |
+
"eval_runtime": 23.9327,
|
| 1942 |
+
"eval_samples_per_second": 125.352,
|
| 1943 |
+
"eval_steps_per_second": 3.928,
|
| 1944 |
+
"step": 654
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 8.67,
|
| 1948 |
+
"eval_loss": 0.6177854537963867,
|
| 1949 |
+
"eval_runtime": 23.9123,
|
| 1950 |
+
"eval_samples_per_second": 125.458,
|
| 1951 |
+
"eval_steps_per_second": 3.931,
|
| 1952 |
+
"step": 657
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 8.71,
|
| 1956 |
+
"learning_rate": 2.422611036339166e-05,
|
| 1957 |
+
"loss": 0.5908,
|
| 1958 |
+
"step": 660
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 8.71,
|
| 1962 |
+
"eval_loss": 0.6179735064506531,
|
| 1963 |
+
"eval_runtime": 23.9135,
|
| 1964 |
+
"eval_samples_per_second": 125.452,
|
| 1965 |
+
"eval_steps_per_second": 3.931,
|
| 1966 |
+
"step": 660
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 8.75,
|
| 1970 |
+
"eval_loss": 0.6180996298789978,
|
| 1971 |
+
"eval_runtime": 23.9038,
|
| 1972 |
+
"eval_samples_per_second": 125.503,
|
| 1973 |
+
"eval_steps_per_second": 3.932,
|
| 1974 |
+
"step": 663
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 8.79,
|
| 1978 |
+
"eval_loss": 0.6181532740592957,
|
| 1979 |
+
"eval_runtime": 23.9058,
|
| 1980 |
+
"eval_samples_per_second": 125.492,
|
| 1981 |
+
"eval_steps_per_second": 3.932,
|
| 1982 |
+
"step": 666
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 8.83,
|
| 1986 |
+
"eval_loss": 0.6176265478134155,
|
| 1987 |
+
"eval_runtime": 23.9128,
|
| 1988 |
+
"eval_samples_per_second": 125.456,
|
| 1989 |
+
"eval_steps_per_second": 3.931,
|
| 1990 |
+
"step": 669
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 8.87,
|
| 1994 |
+
"eval_loss": 0.617388904094696,
|
| 1995 |
+
"eval_runtime": 23.8906,
|
| 1996 |
+
"eval_samples_per_second": 125.572,
|
| 1997 |
+
"eval_steps_per_second": 3.935,
|
| 1998 |
+
"step": 672
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 8.91,
|
| 2002 |
+
"eval_loss": 0.6174699068069458,
|
| 2003 |
+
"eval_runtime": 23.9051,
|
| 2004 |
+
"eval_samples_per_second": 125.496,
|
| 2005 |
+
"eval_steps_per_second": 3.932,
|
| 2006 |
+
"step": 675
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 8.95,
|
| 2010 |
+
"eval_loss": 0.6174932718276978,
|
| 2011 |
+
"eval_runtime": 23.8969,
|
| 2012 |
+
"eval_samples_per_second": 125.54,
|
| 2013 |
+
"eval_steps_per_second": 3.934,
|
| 2014 |
+
"step": 678
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 8.97,
|
| 2018 |
+
"learning_rate": 1.8842530282637954e-05,
|
| 2019 |
+
"loss": 0.5937,
|
| 2020 |
+
"step": 680
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 8.99,
|
| 2024 |
+
"eval_loss": 0.6174784898757935,
|
| 2025 |
+
"eval_runtime": 23.9056,
|
| 2026 |
+
"eval_samples_per_second": 125.494,
|
| 2027 |
+
"eval_steps_per_second": 3.932,
|
| 2028 |
+
"step": 681
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 9.03,
|
| 2032 |
+
"eval_loss": 0.6175104975700378,
|
| 2033 |
+
"eval_runtime": 23.9136,
|
| 2034 |
+
"eval_samples_per_second": 125.452,
|
| 2035 |
+
"eval_steps_per_second": 3.931,
|
| 2036 |
+
"step": 684
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 9.07,
|
| 2040 |
+
"eval_loss": 0.6173563599586487,
|
| 2041 |
+
"eval_runtime": 23.9237,
|
| 2042 |
+
"eval_samples_per_second": 125.399,
|
| 2043 |
+
"eval_steps_per_second": 3.929,
|
| 2044 |
+
"step": 687
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 9.11,
|
| 2048 |
+
"eval_loss": 0.6172643899917603,
|
| 2049 |
+
"eval_runtime": 23.9021,
|
| 2050 |
+
"eval_samples_per_second": 125.512,
|
| 2051 |
+
"eval_steps_per_second": 3.933,
|
| 2052 |
+
"step": 690
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 9.14,
|
| 2056 |
+
"eval_loss": 0.6172318458557129,
|
| 2057 |
+
"eval_runtime": 23.9242,
|
| 2058 |
+
"eval_samples_per_second": 125.396,
|
| 2059 |
+
"eval_steps_per_second": 3.929,
|
| 2060 |
+
"step": 693
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 9.18,
|
| 2064 |
+
"eval_loss": 0.617030680179596,
|
| 2065 |
+
"eval_runtime": 23.9184,
|
| 2066 |
+
"eval_samples_per_second": 125.427,
|
| 2067 |
+
"eval_steps_per_second": 3.93,
|
| 2068 |
+
"step": 696
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 9.22,
|
| 2072 |
+
"eval_loss": 0.6169764995574951,
|
| 2073 |
+
"eval_runtime": 23.9104,
|
| 2074 |
+
"eval_samples_per_second": 125.468,
|
| 2075 |
+
"eval_steps_per_second": 3.931,
|
| 2076 |
+
"step": 699
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 9.24,
|
| 2080 |
+
"learning_rate": 1.3458950201884254e-05,
|
| 2081 |
+
"loss": 0.5867,
|
| 2082 |
+
"step": 700
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 9.26,
|
| 2086 |
+
"eval_loss": 0.6171083450317383,
|
| 2087 |
+
"eval_runtime": 23.9126,
|
| 2088 |
+
"eval_samples_per_second": 125.457,
|
| 2089 |
+
"eval_steps_per_second": 3.931,
|
| 2090 |
+
"step": 702
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 9.3,
|
| 2094 |
+
"eval_loss": 0.6171473860740662,
|
| 2095 |
+
"eval_runtime": 23.9015,
|
| 2096 |
+
"eval_samples_per_second": 125.515,
|
| 2097 |
+
"eval_steps_per_second": 3.933,
|
| 2098 |
+
"step": 705
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 9.34,
|
| 2102 |
+
"eval_loss": 0.6170982122421265,
|
| 2103 |
+
"eval_runtime": 23.9023,
|
| 2104 |
+
"eval_samples_per_second": 125.511,
|
| 2105 |
+
"eval_steps_per_second": 3.933,
|
| 2106 |
+
"step": 708
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 9.38,
|
| 2110 |
+
"eval_loss": 0.6169420480728149,
|
| 2111 |
+
"eval_runtime": 23.9074,
|
| 2112 |
+
"eval_samples_per_second": 125.484,
|
| 2113 |
+
"eval_steps_per_second": 3.932,
|
| 2114 |
+
"step": 711
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 9.42,
|
| 2118 |
+
"eval_loss": 0.6168730854988098,
|
| 2119 |
+
"eval_runtime": 23.9047,
|
| 2120 |
+
"eval_samples_per_second": 125.498,
|
| 2121 |
+
"eval_steps_per_second": 3.932,
|
| 2122 |
+
"step": 714
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 9.46,
|
| 2126 |
+
"eval_loss": 0.6168663501739502,
|
| 2127 |
+
"eval_runtime": 23.9161,
|
| 2128 |
+
"eval_samples_per_second": 125.439,
|
| 2129 |
+
"eval_steps_per_second": 3.93,
|
| 2130 |
+
"step": 717
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 9.5,
|
| 2134 |
+
"learning_rate": 8.075370121130552e-06,
|
| 2135 |
+
"loss": 0.5891,
|
| 2136 |
+
"step": 720
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 9.5,
|
| 2140 |
+
"eval_loss": 0.6167708039283752,
|
| 2141 |
+
"eval_runtime": 23.9114,
|
| 2142 |
+
"eval_samples_per_second": 125.463,
|
| 2143 |
+
"eval_steps_per_second": 3.931,
|
| 2144 |
+
"step": 720
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 9.54,
|
| 2148 |
+
"eval_loss": 0.6167441010475159,
|
| 2149 |
+
"eval_runtime": 23.9086,
|
| 2150 |
+
"eval_samples_per_second": 125.478,
|
| 2151 |
+
"eval_steps_per_second": 3.932,
|
| 2152 |
+
"step": 723
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 9.58,
|
| 2156 |
+
"eval_loss": 0.6168937683105469,
|
| 2157 |
+
"eval_runtime": 23.9122,
|
| 2158 |
+
"eval_samples_per_second": 125.459,
|
| 2159 |
+
"eval_steps_per_second": 3.931,
|
| 2160 |
+
"step": 726
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 9.62,
|
| 2164 |
+
"eval_loss": 0.6170821189880371,
|
| 2165 |
+
"eval_runtime": 23.9065,
|
| 2166 |
+
"eval_samples_per_second": 125.489,
|
| 2167 |
+
"eval_steps_per_second": 3.932,
|
| 2168 |
+
"step": 729
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 9.66,
|
| 2172 |
+
"eval_loss": 0.6171652674674988,
|
| 2173 |
+
"eval_runtime": 23.9131,
|
| 2174 |
+
"eval_samples_per_second": 125.454,
|
| 2175 |
+
"eval_steps_per_second": 3.931,
|
| 2176 |
+
"step": 732
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 9.7,
|
| 2180 |
+
"eval_loss": 0.6171714663505554,
|
| 2181 |
+
"eval_runtime": 23.9057,
|
| 2182 |
+
"eval_samples_per_second": 125.493,
|
| 2183 |
+
"eval_steps_per_second": 3.932,
|
| 2184 |
+
"step": 735
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 9.74,
|
| 2188 |
+
"eval_loss": 0.6170787811279297,
|
| 2189 |
+
"eval_runtime": 23.928,
|
| 2190 |
+
"eval_samples_per_second": 125.376,
|
| 2191 |
+
"eval_steps_per_second": 3.928,
|
| 2192 |
+
"step": 738
|
| 2193 |
+
}
|
| 2194 |
+
],
|
| 2195 |
+
"max_steps": 750,
|
| 2196 |
+
"num_train_epochs": 10,
|
| 2197 |
+
"total_flos": 7.798891163997438e+18,
|
| 2198 |
+
"trial_name": null,
|
| 2199 |
+
"trial_params": null
|
| 2200 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-738/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d5a386ce8addef927bca0e390e9534e0877a3e4e00f222f83967dd78c49527a
|
| 3 |
+
size 4027
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/adapter_config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "/mnt/data1/sheshuaijie/Data/PLM/vicuna-7b",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"fan_in_fan_out": false,
|
| 5 |
+
"inference_mode": true,
|
| 6 |
+
"init_lora_weights": true,
|
| 7 |
+
"lora_alpha": 32,
|
| 8 |
+
"lora_dropout": 0.1,
|
| 9 |
+
"modules_to_save": null,
|
| 10 |
+
"peft_type": "LORA",
|
| 11 |
+
"r": 32,
|
| 12 |
+
"target_modules": [
|
| 13 |
+
"q_proj",
|
| 14 |
+
"v_proj"
|
| 15 |
+
],
|
| 16 |
+
"task_type": "CAUSAL_LM"
|
| 17 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1cb805ac62e415da1c15d91797d4355c8080bdcfc7c7781a4ed1539877a0af57
|
| 3 |
+
size 134293701
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0957395ee23cae5537ca68fdb8f5f3c0ba7ef97e714bd742f388719a9c5cd4c5
|
| 3 |
+
size 67154893
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e292c6ba42d09ed37e3998fc80693007d182b35ac775e2161f534d0f9f083131
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c7c10e4824a89fe497107b1825bf1cfa4fbb78538f2193d56151fad9e2ee99bc
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba511621fdf523ff7989a294ae5130ea54f3b14afd891da2e41bf244e8a1cfc5
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c91d9cec94c7ef784b2b57a2a410bb4b17e31551753c577ee7e48e81791c2057
|
| 3 |
+
size 17655
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f40d0f980607497195b40bb30daf589ddd80f47c884eb1a2231e665f009933b
|
| 3 |
+
size 557
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c1b48f5991e61a145c837cbe5f50a0347f208c7ef9ad1718e593fddf77a205b0
|
| 3 |
+
size 627
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/tokenizer_config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"bos_token": {
|
| 5 |
+
"__type": "AddedToken",
|
| 6 |
+
"content": "<s>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": true,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"clean_up_tokenization_spaces": false,
|
| 13 |
+
"eos_token": {
|
| 14 |
+
"__type": "AddedToken",
|
| 15 |
+
"content": "</s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": true,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
},
|
| 21 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 22 |
+
"pad_token": null,
|
| 23 |
+
"sp_model_kwargs": {},
|
| 24 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
+
"unk_token": {
|
| 26 |
+
"__type": "AddedToken",
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": true,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
}
|
| 33 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/trainer_state.json
ADDED
|
@@ -0,0 +1,2214 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6167441010475159,
|
| 3 |
+
"best_model_checkpoint": "/mnt/data1/sheshuaijie/Output/CoT/Trained/vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-723",
|
| 4 |
+
"epoch": 9.778144329896907,
|
| 5 |
+
"global_step": 741,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"eval_loss": 1.7524008750915527,
|
| 13 |
+
"eval_runtime": 23.7951,
|
| 14 |
+
"eval_samples_per_second": 126.077,
|
| 15 |
+
"eval_steps_per_second": 3.95,
|
| 16 |
+
"step": 3
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.08,
|
| 20 |
+
"eval_loss": 1.5515066385269165,
|
| 21 |
+
"eval_runtime": 23.828,
|
| 22 |
+
"eval_samples_per_second": 125.902,
|
| 23 |
+
"eval_steps_per_second": 3.945,
|
| 24 |
+
"step": 6
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12,
|
| 28 |
+
"eval_loss": 1.3584641218185425,
|
| 29 |
+
"eval_runtime": 23.8775,
|
| 30 |
+
"eval_samples_per_second": 125.641,
|
| 31 |
+
"eval_steps_per_second": 3.937,
|
| 32 |
+
"step": 9
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.16,
|
| 36 |
+
"eval_loss": 1.2644726037979126,
|
| 37 |
+
"eval_runtime": 23.8942,
|
| 38 |
+
"eval_samples_per_second": 125.554,
|
| 39 |
+
"eval_steps_per_second": 3.934,
|
| 40 |
+
"step": 12
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.2,
|
| 44 |
+
"eval_loss": 1.166400671005249,
|
| 45 |
+
"eval_runtime": 23.9181,
|
| 46 |
+
"eval_samples_per_second": 125.428,
|
| 47 |
+
"eval_steps_per_second": 3.93,
|
| 48 |
+
"step": 15
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.24,
|
| 52 |
+
"eval_loss": 1.1086052656173706,
|
| 53 |
+
"eval_runtime": 23.9166,
|
| 54 |
+
"eval_samples_per_second": 125.436,
|
| 55 |
+
"eval_steps_per_second": 3.93,
|
| 56 |
+
"step": 18
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.26,
|
| 60 |
+
"learning_rate": 0.00019650067294751011,
|
| 61 |
+
"loss": 1.4265,
|
| 62 |
+
"step": 20
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.28,
|
| 66 |
+
"eval_loss": 1.0677987337112427,
|
| 67 |
+
"eval_runtime": 23.9189,
|
| 68 |
+
"eval_samples_per_second": 125.424,
|
| 69 |
+
"eval_steps_per_second": 3.93,
|
| 70 |
+
"step": 21
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.32,
|
| 74 |
+
"eval_loss": 1.0342437028884888,
|
| 75 |
+
"eval_runtime": 23.9046,
|
| 76 |
+
"eval_samples_per_second": 125.499,
|
| 77 |
+
"eval_steps_per_second": 3.932,
|
| 78 |
+
"step": 24
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.36,
|
| 82 |
+
"eval_loss": 0.9985266923904419,
|
| 83 |
+
"eval_runtime": 23.9037,
|
| 84 |
+
"eval_samples_per_second": 125.504,
|
| 85 |
+
"eval_steps_per_second": 3.932,
|
| 86 |
+
"step": 27
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.4,
|
| 90 |
+
"eval_loss": 0.9654523134231567,
|
| 91 |
+
"eval_runtime": 23.9129,
|
| 92 |
+
"eval_samples_per_second": 125.455,
|
| 93 |
+
"eval_steps_per_second": 3.931,
|
| 94 |
+
"step": 30
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.44,
|
| 98 |
+
"eval_loss": 0.939262866973877,
|
| 99 |
+
"eval_runtime": 23.9117,
|
| 100 |
+
"eval_samples_per_second": 125.462,
|
| 101 |
+
"eval_steps_per_second": 3.931,
|
| 102 |
+
"step": 33
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.48,
|
| 106 |
+
"eval_loss": 0.9186767339706421,
|
| 107 |
+
"eval_runtime": 23.9011,
|
| 108 |
+
"eval_samples_per_second": 125.517,
|
| 109 |
+
"eval_steps_per_second": 3.933,
|
| 110 |
+
"step": 36
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.51,
|
| 114 |
+
"eval_loss": 0.8969741463661194,
|
| 115 |
+
"eval_runtime": 23.9105,
|
| 116 |
+
"eval_samples_per_second": 125.468,
|
| 117 |
+
"eval_steps_per_second": 3.931,
|
| 118 |
+
"step": 39
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.53,
|
| 122 |
+
"learning_rate": 0.00019111709286675642,
|
| 123 |
+
"loss": 0.9923,
|
| 124 |
+
"step": 40
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.55,
|
| 128 |
+
"eval_loss": 0.8814375996589661,
|
| 129 |
+
"eval_runtime": 23.9154,
|
| 130 |
+
"eval_samples_per_second": 125.442,
|
| 131 |
+
"eval_steps_per_second": 3.931,
|
| 132 |
+
"step": 42
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 0.59,
|
| 136 |
+
"eval_loss": 0.8654683232307434,
|
| 137 |
+
"eval_runtime": 23.9108,
|
| 138 |
+
"eval_samples_per_second": 125.466,
|
| 139 |
+
"eval_steps_per_second": 3.931,
|
| 140 |
+
"step": 45
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.63,
|
| 144 |
+
"eval_loss": 0.852226734161377,
|
| 145 |
+
"eval_runtime": 23.9186,
|
| 146 |
+
"eval_samples_per_second": 125.425,
|
| 147 |
+
"eval_steps_per_second": 3.93,
|
| 148 |
+
"step": 48
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.67,
|
| 152 |
+
"eval_loss": 0.839223325252533,
|
| 153 |
+
"eval_runtime": 23.9074,
|
| 154 |
+
"eval_samples_per_second": 125.484,
|
| 155 |
+
"eval_steps_per_second": 3.932,
|
| 156 |
+
"step": 51
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.71,
|
| 160 |
+
"eval_loss": 0.8266379237174988,
|
| 161 |
+
"eval_runtime": 23.9399,
|
| 162 |
+
"eval_samples_per_second": 125.314,
|
| 163 |
+
"eval_steps_per_second": 3.926,
|
| 164 |
+
"step": 54
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.75,
|
| 168 |
+
"eval_loss": 0.8140417337417603,
|
| 169 |
+
"eval_runtime": 23.9355,
|
| 170 |
+
"eval_samples_per_second": 125.337,
|
| 171 |
+
"eval_steps_per_second": 3.927,
|
| 172 |
+
"step": 57
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.79,
|
| 176 |
+
"learning_rate": 0.0001857335127860027,
|
| 177 |
+
"loss": 0.8611,
|
| 178 |
+
"step": 60
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.79,
|
| 182 |
+
"eval_loss": 0.8019057512283325,
|
| 183 |
+
"eval_runtime": 23.9223,
|
| 184 |
+
"eval_samples_per_second": 125.406,
|
| 185 |
+
"eval_steps_per_second": 3.929,
|
| 186 |
+
"step": 60
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.83,
|
| 190 |
+
"eval_loss": 0.7907609343528748,
|
| 191 |
+
"eval_runtime": 23.9384,
|
| 192 |
+
"eval_samples_per_second": 125.322,
|
| 193 |
+
"eval_steps_per_second": 3.927,
|
| 194 |
+
"step": 63
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.87,
|
| 198 |
+
"eval_loss": 0.7791212797164917,
|
| 199 |
+
"eval_runtime": 23.9101,
|
| 200 |
+
"eval_samples_per_second": 125.47,
|
| 201 |
+
"eval_steps_per_second": 3.931,
|
| 202 |
+
"step": 66
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.91,
|
| 206 |
+
"eval_loss": 0.7694615125656128,
|
| 207 |
+
"eval_runtime": 23.9079,
|
| 208 |
+
"eval_samples_per_second": 125.481,
|
| 209 |
+
"eval_steps_per_second": 3.932,
|
| 210 |
+
"step": 69
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.95,
|
| 214 |
+
"eval_loss": 0.7602358460426331,
|
| 215 |
+
"eval_runtime": 23.9116,
|
| 216 |
+
"eval_samples_per_second": 125.462,
|
| 217 |
+
"eval_steps_per_second": 3.931,
|
| 218 |
+
"step": 72
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.99,
|
| 222 |
+
"eval_loss": 0.753226101398468,
|
| 223 |
+
"eval_runtime": 23.9242,
|
| 224 |
+
"eval_samples_per_second": 125.396,
|
| 225 |
+
"eval_steps_per_second": 3.929,
|
| 226 |
+
"step": 75
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.03,
|
| 230 |
+
"eval_loss": 0.7466432452201843,
|
| 231 |
+
"eval_runtime": 23.9116,
|
| 232 |
+
"eval_samples_per_second": 125.462,
|
| 233 |
+
"eval_steps_per_second": 3.931,
|
| 234 |
+
"step": 78
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.06,
|
| 238 |
+
"learning_rate": 0.000180349932705249,
|
| 239 |
+
"loss": 0.7843,
|
| 240 |
+
"step": 80
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.07,
|
| 244 |
+
"eval_loss": 0.7416810989379883,
|
| 245 |
+
"eval_runtime": 23.9171,
|
| 246 |
+
"eval_samples_per_second": 125.433,
|
| 247 |
+
"eval_steps_per_second": 3.93,
|
| 248 |
+
"step": 81
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.11,
|
| 252 |
+
"eval_loss": 0.7362396121025085,
|
| 253 |
+
"eval_runtime": 23.9079,
|
| 254 |
+
"eval_samples_per_second": 125.481,
|
| 255 |
+
"eval_steps_per_second": 3.932,
|
| 256 |
+
"step": 84
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.15,
|
| 260 |
+
"eval_loss": 0.7297741174697876,
|
| 261 |
+
"eval_runtime": 23.9084,
|
| 262 |
+
"eval_samples_per_second": 125.479,
|
| 263 |
+
"eval_steps_per_second": 3.932,
|
| 264 |
+
"step": 87
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.19,
|
| 268 |
+
"eval_loss": 0.7252654433250427,
|
| 269 |
+
"eval_runtime": 23.9206,
|
| 270 |
+
"eval_samples_per_second": 125.415,
|
| 271 |
+
"eval_steps_per_second": 3.93,
|
| 272 |
+
"step": 90
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.23,
|
| 276 |
+
"eval_loss": 0.7213409543037415,
|
| 277 |
+
"eval_runtime": 23.9179,
|
| 278 |
+
"eval_samples_per_second": 125.429,
|
| 279 |
+
"eval_steps_per_second": 3.93,
|
| 280 |
+
"step": 93
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.27,
|
| 284 |
+
"eval_loss": 0.7174035906791687,
|
| 285 |
+
"eval_runtime": 23.9354,
|
| 286 |
+
"eval_samples_per_second": 125.337,
|
| 287 |
+
"eval_steps_per_second": 3.927,
|
| 288 |
+
"step": 96
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.31,
|
| 292 |
+
"eval_loss": 0.7140380144119263,
|
| 293 |
+
"eval_runtime": 23.9214,
|
| 294 |
+
"eval_samples_per_second": 125.411,
|
| 295 |
+
"eval_steps_per_second": 3.93,
|
| 296 |
+
"step": 99
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.32,
|
| 300 |
+
"learning_rate": 0.0001749663526244953,
|
| 301 |
+
"loss": 0.7301,
|
| 302 |
+
"step": 100
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.35,
|
| 306 |
+
"eval_loss": 0.7104487419128418,
|
| 307 |
+
"eval_runtime": 23.9093,
|
| 308 |
+
"eval_samples_per_second": 125.474,
|
| 309 |
+
"eval_steps_per_second": 3.932,
|
| 310 |
+
"step": 102
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.39,
|
| 314 |
+
"eval_loss": 0.7067868113517761,
|
| 315 |
+
"eval_runtime": 23.9129,
|
| 316 |
+
"eval_samples_per_second": 125.455,
|
| 317 |
+
"eval_steps_per_second": 3.931,
|
| 318 |
+
"step": 105
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.43,
|
| 322 |
+
"eval_loss": 0.7041762471199036,
|
| 323 |
+
"eval_runtime": 23.9161,
|
| 324 |
+
"eval_samples_per_second": 125.439,
|
| 325 |
+
"eval_steps_per_second": 3.93,
|
| 326 |
+
"step": 108
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.46,
|
| 330 |
+
"eval_loss": 0.7013522982597351,
|
| 331 |
+
"eval_runtime": 23.9133,
|
| 332 |
+
"eval_samples_per_second": 125.453,
|
| 333 |
+
"eval_steps_per_second": 3.931,
|
| 334 |
+
"step": 111
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.5,
|
| 338 |
+
"eval_loss": 0.6989504098892212,
|
| 339 |
+
"eval_runtime": 23.9152,
|
| 340 |
+
"eval_samples_per_second": 125.443,
|
| 341 |
+
"eval_steps_per_second": 3.931,
|
| 342 |
+
"step": 114
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 1.54,
|
| 346 |
+
"eval_loss": 0.6974085569381714,
|
| 347 |
+
"eval_runtime": 23.9561,
|
| 348 |
+
"eval_samples_per_second": 125.229,
|
| 349 |
+
"eval_steps_per_second": 3.924,
|
| 350 |
+
"step": 117
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.58,
|
| 354 |
+
"learning_rate": 0.0001695827725437416,
|
| 355 |
+
"loss": 0.7141,
|
| 356 |
+
"step": 120
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.58,
|
| 360 |
+
"eval_loss": 0.6944894194602966,
|
| 361 |
+
"eval_runtime": 23.902,
|
| 362 |
+
"eval_samples_per_second": 125.512,
|
| 363 |
+
"eval_steps_per_second": 3.933,
|
| 364 |
+
"step": 120
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 1.62,
|
| 368 |
+
"eval_loss": 0.6929482221603394,
|
| 369 |
+
"eval_runtime": 23.9189,
|
| 370 |
+
"eval_samples_per_second": 125.424,
|
| 371 |
+
"eval_steps_per_second": 3.93,
|
| 372 |
+
"step": 123
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.66,
|
| 376 |
+
"eval_loss": 0.6903366446495056,
|
| 377 |
+
"eval_runtime": 23.9061,
|
| 378 |
+
"eval_samples_per_second": 125.491,
|
| 379 |
+
"eval_steps_per_second": 3.932,
|
| 380 |
+
"step": 126
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.7,
|
| 384 |
+
"eval_loss": 0.6882749199867249,
|
| 385 |
+
"eval_runtime": 23.9181,
|
| 386 |
+
"eval_samples_per_second": 125.428,
|
| 387 |
+
"eval_steps_per_second": 3.93,
|
| 388 |
+
"step": 129
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.74,
|
| 392 |
+
"eval_loss": 0.6863100528717041,
|
| 393 |
+
"eval_runtime": 23.914,
|
| 394 |
+
"eval_samples_per_second": 125.45,
|
| 395 |
+
"eval_steps_per_second": 3.931,
|
| 396 |
+
"step": 132
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.78,
|
| 400 |
+
"eval_loss": 0.6860549449920654,
|
| 401 |
+
"eval_runtime": 23.9138,
|
| 402 |
+
"eval_samples_per_second": 125.45,
|
| 403 |
+
"eval_steps_per_second": 3.931,
|
| 404 |
+
"step": 135
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.82,
|
| 408 |
+
"eval_loss": 0.6831715703010559,
|
| 409 |
+
"eval_runtime": 23.9135,
|
| 410 |
+
"eval_samples_per_second": 125.452,
|
| 411 |
+
"eval_steps_per_second": 3.931,
|
| 412 |
+
"step": 138
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.85,
|
| 416 |
+
"learning_rate": 0.0001641991924629879,
|
| 417 |
+
"loss": 0.6902,
|
| 418 |
+
"step": 140
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.86,
|
| 422 |
+
"eval_loss": 0.6819499731063843,
|
| 423 |
+
"eval_runtime": 23.8986,
|
| 424 |
+
"eval_samples_per_second": 125.53,
|
| 425 |
+
"eval_steps_per_second": 3.933,
|
| 426 |
+
"step": 141
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.9,
|
| 430 |
+
"eval_loss": 0.6807693839073181,
|
| 431 |
+
"eval_runtime": 23.9169,
|
| 432 |
+
"eval_samples_per_second": 125.434,
|
| 433 |
+
"eval_steps_per_second": 3.93,
|
| 434 |
+
"step": 144
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 1.94,
|
| 438 |
+
"eval_loss": 0.6787669062614441,
|
| 439 |
+
"eval_runtime": 23.9265,
|
| 440 |
+
"eval_samples_per_second": 125.384,
|
| 441 |
+
"eval_steps_per_second": 3.929,
|
| 442 |
+
"step": 147
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.98,
|
| 446 |
+
"eval_loss": 0.6773442625999451,
|
| 447 |
+
"eval_runtime": 23.9274,
|
| 448 |
+
"eval_samples_per_second": 125.38,
|
| 449 |
+
"eval_steps_per_second": 3.929,
|
| 450 |
+
"step": 150
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.02,
|
| 454 |
+
"eval_loss": 0.6759281158447266,
|
| 455 |
+
"eval_runtime": 23.9386,
|
| 456 |
+
"eval_samples_per_second": 125.321,
|
| 457 |
+
"eval_steps_per_second": 3.927,
|
| 458 |
+
"step": 153
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.06,
|
| 462 |
+
"eval_loss": 0.6743582487106323,
|
| 463 |
+
"eval_runtime": 23.9323,
|
| 464 |
+
"eval_samples_per_second": 125.354,
|
| 465 |
+
"eval_steps_per_second": 3.928,
|
| 466 |
+
"step": 156
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 2.1,
|
| 470 |
+
"eval_loss": 0.6732926368713379,
|
| 471 |
+
"eval_runtime": 23.9145,
|
| 472 |
+
"eval_samples_per_second": 125.447,
|
| 473 |
+
"eval_steps_per_second": 3.931,
|
| 474 |
+
"step": 159
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 2.11,
|
| 478 |
+
"learning_rate": 0.0001588156123822342,
|
| 479 |
+
"loss": 0.6766,
|
| 480 |
+
"step": 160
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 2.14,
|
| 484 |
+
"eval_loss": 0.6721953749656677,
|
| 485 |
+
"eval_runtime": 23.9073,
|
| 486 |
+
"eval_samples_per_second": 125.485,
|
| 487 |
+
"eval_steps_per_second": 3.932,
|
| 488 |
+
"step": 162
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.18,
|
| 492 |
+
"eval_loss": 0.6714429259300232,
|
| 493 |
+
"eval_runtime": 23.8955,
|
| 494 |
+
"eval_samples_per_second": 125.547,
|
| 495 |
+
"eval_steps_per_second": 3.934,
|
| 496 |
+
"step": 165
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 2.22,
|
| 500 |
+
"eval_loss": 0.670035183429718,
|
| 501 |
+
"eval_runtime": 23.9431,
|
| 502 |
+
"eval_samples_per_second": 125.297,
|
| 503 |
+
"eval_steps_per_second": 3.926,
|
| 504 |
+
"step": 168
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 2.26,
|
| 508 |
+
"eval_loss": 0.6695354580879211,
|
| 509 |
+
"eval_runtime": 23.8875,
|
| 510 |
+
"eval_samples_per_second": 125.589,
|
| 511 |
+
"eval_steps_per_second": 3.935,
|
| 512 |
+
"step": 171
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.3,
|
| 516 |
+
"eval_loss": 0.6689226031303406,
|
| 517 |
+
"eval_runtime": 23.9185,
|
| 518 |
+
"eval_samples_per_second": 125.426,
|
| 519 |
+
"eval_steps_per_second": 3.93,
|
| 520 |
+
"step": 174
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.34,
|
| 524 |
+
"eval_loss": 0.6674054861068726,
|
| 525 |
+
"eval_runtime": 23.941,
|
| 526 |
+
"eval_samples_per_second": 125.308,
|
| 527 |
+
"eval_steps_per_second": 3.926,
|
| 528 |
+
"step": 177
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.38,
|
| 532 |
+
"learning_rate": 0.00015343203230148048,
|
| 533 |
+
"loss": 0.6743,
|
| 534 |
+
"step": 180
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.38,
|
| 538 |
+
"eval_loss": 0.6664847731590271,
|
| 539 |
+
"eval_runtime": 23.9211,
|
| 540 |
+
"eval_samples_per_second": 125.412,
|
| 541 |
+
"eval_steps_per_second": 3.93,
|
| 542 |
+
"step": 180
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.41,
|
| 546 |
+
"eval_loss": 0.6658627986907959,
|
| 547 |
+
"eval_runtime": 23.9247,
|
| 548 |
+
"eval_samples_per_second": 125.394,
|
| 549 |
+
"eval_steps_per_second": 3.929,
|
| 550 |
+
"step": 183
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 2.45,
|
| 554 |
+
"eval_loss": 0.664908766746521,
|
| 555 |
+
"eval_runtime": 23.9272,
|
| 556 |
+
"eval_samples_per_second": 125.38,
|
| 557 |
+
"eval_steps_per_second": 3.929,
|
| 558 |
+
"step": 186
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.49,
|
| 562 |
+
"eval_loss": 0.6638036966323853,
|
| 563 |
+
"eval_runtime": 23.9187,
|
| 564 |
+
"eval_samples_per_second": 125.425,
|
| 565 |
+
"eval_steps_per_second": 3.93,
|
| 566 |
+
"step": 189
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.53,
|
| 570 |
+
"eval_loss": 0.6625837683677673,
|
| 571 |
+
"eval_runtime": 23.9033,
|
| 572 |
+
"eval_samples_per_second": 125.506,
|
| 573 |
+
"eval_steps_per_second": 3.933,
|
| 574 |
+
"step": 192
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 2.57,
|
| 578 |
+
"eval_loss": 0.6619511842727661,
|
| 579 |
+
"eval_runtime": 23.8973,
|
| 580 |
+
"eval_samples_per_second": 125.537,
|
| 581 |
+
"eval_steps_per_second": 3.934,
|
| 582 |
+
"step": 195
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.61,
|
| 586 |
+
"eval_loss": 0.6611769199371338,
|
| 587 |
+
"eval_runtime": 23.9129,
|
| 588 |
+
"eval_samples_per_second": 125.455,
|
| 589 |
+
"eval_steps_per_second": 3.931,
|
| 590 |
+
"step": 198
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.64,
|
| 594 |
+
"learning_rate": 0.00014804845222072678,
|
| 595 |
+
"loss": 0.6615,
|
| 596 |
+
"step": 200
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.65,
|
| 600 |
+
"eval_loss": 0.6606143116950989,
|
| 601 |
+
"eval_runtime": 23.9126,
|
| 602 |
+
"eval_samples_per_second": 125.457,
|
| 603 |
+
"eval_steps_per_second": 3.931,
|
| 604 |
+
"step": 201
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 2.69,
|
| 608 |
+
"eval_loss": 0.6589743494987488,
|
| 609 |
+
"eval_runtime": 23.9135,
|
| 610 |
+
"eval_samples_per_second": 125.452,
|
| 611 |
+
"eval_steps_per_second": 3.931,
|
| 612 |
+
"step": 204
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.73,
|
| 616 |
+
"eval_loss": 0.6578481197357178,
|
| 617 |
+
"eval_runtime": 23.9217,
|
| 618 |
+
"eval_samples_per_second": 125.409,
|
| 619 |
+
"eval_steps_per_second": 3.929,
|
| 620 |
+
"step": 207
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.77,
|
| 624 |
+
"eval_loss": 0.6571096181869507,
|
| 625 |
+
"eval_runtime": 23.9415,
|
| 626 |
+
"eval_samples_per_second": 125.305,
|
| 627 |
+
"eval_steps_per_second": 3.926,
|
| 628 |
+
"step": 210
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 2.81,
|
| 632 |
+
"eval_loss": 0.656689465045929,
|
| 633 |
+
"eval_runtime": 23.9111,
|
| 634 |
+
"eval_samples_per_second": 125.465,
|
| 635 |
+
"eval_steps_per_second": 3.931,
|
| 636 |
+
"step": 213
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.85,
|
| 640 |
+
"eval_loss": 0.6556207537651062,
|
| 641 |
+
"eval_runtime": 23.9099,
|
| 642 |
+
"eval_samples_per_second": 125.471,
|
| 643 |
+
"eval_steps_per_second": 3.931,
|
| 644 |
+
"step": 216
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.89,
|
| 648 |
+
"eval_loss": 0.6546627283096313,
|
| 649 |
+
"eval_runtime": 23.9164,
|
| 650 |
+
"eval_samples_per_second": 125.437,
|
| 651 |
+
"eval_steps_per_second": 3.93,
|
| 652 |
+
"step": 219
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 2.9,
|
| 656 |
+
"learning_rate": 0.0001426648721399731,
|
| 657 |
+
"loss": 0.6564,
|
| 658 |
+
"step": 220
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 2.93,
|
| 662 |
+
"eval_loss": 0.6539400815963745,
|
| 663 |
+
"eval_runtime": 23.906,
|
| 664 |
+
"eval_samples_per_second": 125.492,
|
| 665 |
+
"eval_steps_per_second": 3.932,
|
| 666 |
+
"step": 222
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.97,
|
| 670 |
+
"eval_loss": 0.653684675693512,
|
| 671 |
+
"eval_runtime": 23.9251,
|
| 672 |
+
"eval_samples_per_second": 125.391,
|
| 673 |
+
"eval_steps_per_second": 3.929,
|
| 674 |
+
"step": 225
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.01,
|
| 678 |
+
"eval_loss": 0.6526629328727722,
|
| 679 |
+
"eval_runtime": 23.9289,
|
| 680 |
+
"eval_samples_per_second": 125.371,
|
| 681 |
+
"eval_steps_per_second": 3.928,
|
| 682 |
+
"step": 228
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 3.05,
|
| 686 |
+
"eval_loss": 0.6525079011917114,
|
| 687 |
+
"eval_runtime": 23.9193,
|
| 688 |
+
"eval_samples_per_second": 125.421,
|
| 689 |
+
"eval_steps_per_second": 3.93,
|
| 690 |
+
"step": 231
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 3.09,
|
| 694 |
+
"eval_loss": 0.6514959931373596,
|
| 695 |
+
"eval_runtime": 23.9574,
|
| 696 |
+
"eval_samples_per_second": 125.223,
|
| 697 |
+
"eval_steps_per_second": 3.924,
|
| 698 |
+
"step": 234
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.13,
|
| 702 |
+
"eval_loss": 0.6507047414779663,
|
| 703 |
+
"eval_runtime": 23.9234,
|
| 704 |
+
"eval_samples_per_second": 125.4,
|
| 705 |
+
"eval_steps_per_second": 3.929,
|
| 706 |
+
"step": 237
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 3.17,
|
| 710 |
+
"learning_rate": 0.00013728129205921937,
|
| 711 |
+
"loss": 0.6469,
|
| 712 |
+
"step": 240
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 3.17,
|
| 716 |
+
"eval_loss": 0.6504186391830444,
|
| 717 |
+
"eval_runtime": 23.937,
|
| 718 |
+
"eval_samples_per_second": 125.329,
|
| 719 |
+
"eval_steps_per_second": 3.927,
|
| 720 |
+
"step": 240
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 3.21,
|
| 724 |
+
"eval_loss": 0.6495808959007263,
|
| 725 |
+
"eval_runtime": 23.9188,
|
| 726 |
+
"eval_samples_per_second": 125.425,
|
| 727 |
+
"eval_steps_per_second": 3.93,
|
| 728 |
+
"step": 243
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.25,
|
| 732 |
+
"eval_loss": 0.649512529373169,
|
| 733 |
+
"eval_runtime": 23.9209,
|
| 734 |
+
"eval_samples_per_second": 125.413,
|
| 735 |
+
"eval_steps_per_second": 3.93,
|
| 736 |
+
"step": 246
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 3.29,
|
| 740 |
+
"eval_loss": 0.648629903793335,
|
| 741 |
+
"eval_runtime": 23.9137,
|
| 742 |
+
"eval_samples_per_second": 125.451,
|
| 743 |
+
"eval_steps_per_second": 3.931,
|
| 744 |
+
"step": 249
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 3.33,
|
| 748 |
+
"eval_loss": 0.6480894088745117,
|
| 749 |
+
"eval_runtime": 23.919,
|
| 750 |
+
"eval_samples_per_second": 125.423,
|
| 751 |
+
"eval_steps_per_second": 3.93,
|
| 752 |
+
"step": 252
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.36,
|
| 756 |
+
"eval_loss": 0.6474400758743286,
|
| 757 |
+
"eval_runtime": 23.9076,
|
| 758 |
+
"eval_samples_per_second": 125.483,
|
| 759 |
+
"eval_steps_per_second": 3.932,
|
| 760 |
+
"step": 255
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 3.4,
|
| 764 |
+
"eval_loss": 0.6468291878700256,
|
| 765 |
+
"eval_runtime": 23.9305,
|
| 766 |
+
"eval_samples_per_second": 125.363,
|
| 767 |
+
"eval_steps_per_second": 3.928,
|
| 768 |
+
"step": 258
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 3.43,
|
| 772 |
+
"learning_rate": 0.00013189771197846567,
|
| 773 |
+
"loss": 0.6463,
|
| 774 |
+
"step": 260
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 3.44,
|
| 778 |
+
"eval_loss": 0.6462663412094116,
|
| 779 |
+
"eval_runtime": 23.9359,
|
| 780 |
+
"eval_samples_per_second": 125.335,
|
| 781 |
+
"eval_steps_per_second": 3.927,
|
| 782 |
+
"step": 261
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.48,
|
| 786 |
+
"eval_loss": 0.6458565592765808,
|
| 787 |
+
"eval_runtime": 23.929,
|
| 788 |
+
"eval_samples_per_second": 125.371,
|
| 789 |
+
"eval_steps_per_second": 3.928,
|
| 790 |
+
"step": 264
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 3.52,
|
| 794 |
+
"eval_loss": 0.645412266254425,
|
| 795 |
+
"eval_runtime": 23.9362,
|
| 796 |
+
"eval_samples_per_second": 125.333,
|
| 797 |
+
"eval_steps_per_second": 3.927,
|
| 798 |
+
"step": 267
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 3.56,
|
| 802 |
+
"eval_loss": 0.6449554562568665,
|
| 803 |
+
"eval_runtime": 23.9004,
|
| 804 |
+
"eval_samples_per_second": 125.521,
|
| 805 |
+
"eval_steps_per_second": 3.933,
|
| 806 |
+
"step": 270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.6,
|
| 810 |
+
"eval_loss": 0.6443325281143188,
|
| 811 |
+
"eval_runtime": 23.9065,
|
| 812 |
+
"eval_samples_per_second": 125.489,
|
| 813 |
+
"eval_steps_per_second": 3.932,
|
| 814 |
+
"step": 273
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 3.64,
|
| 818 |
+
"eval_loss": 0.6435034871101379,
|
| 819 |
+
"eval_runtime": 23.9072,
|
| 820 |
+
"eval_samples_per_second": 125.485,
|
| 821 |
+
"eval_steps_per_second": 3.932,
|
| 822 |
+
"step": 276
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 3.68,
|
| 826 |
+
"eval_loss": 0.6433733701705933,
|
| 827 |
+
"eval_runtime": 23.9042,
|
| 828 |
+
"eval_samples_per_second": 125.501,
|
| 829 |
+
"eval_steps_per_second": 3.932,
|
| 830 |
+
"step": 279
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.69,
|
| 834 |
+
"learning_rate": 0.00012651413189771198,
|
| 835 |
+
"loss": 0.6389,
|
| 836 |
+
"step": 280
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.72,
|
| 840 |
+
"eval_loss": 0.6425070762634277,
|
| 841 |
+
"eval_runtime": 23.8874,
|
| 842 |
+
"eval_samples_per_second": 125.589,
|
| 843 |
+
"eval_steps_per_second": 3.935,
|
| 844 |
+
"step": 282
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 3.76,
|
| 848 |
+
"eval_loss": 0.642119288444519,
|
| 849 |
+
"eval_runtime": 23.9328,
|
| 850 |
+
"eval_samples_per_second": 125.351,
|
| 851 |
+
"eval_steps_per_second": 3.928,
|
| 852 |
+
"step": 285
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 3.8,
|
| 856 |
+
"eval_loss": 0.641748309135437,
|
| 857 |
+
"eval_runtime": 23.9294,
|
| 858 |
+
"eval_samples_per_second": 125.369,
|
| 859 |
+
"eval_steps_per_second": 3.928,
|
| 860 |
+
"step": 288
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.84,
|
| 864 |
+
"eval_loss": 0.640826404094696,
|
| 865 |
+
"eval_runtime": 23.9434,
|
| 866 |
+
"eval_samples_per_second": 125.296,
|
| 867 |
+
"eval_steps_per_second": 3.926,
|
| 868 |
+
"step": 291
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 3.88,
|
| 872 |
+
"eval_loss": 0.6402388215065002,
|
| 873 |
+
"eval_runtime": 23.9162,
|
| 874 |
+
"eval_samples_per_second": 125.438,
|
| 875 |
+
"eval_steps_per_second": 3.93,
|
| 876 |
+
"step": 294
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 3.92,
|
| 880 |
+
"eval_loss": 0.6407353281974792,
|
| 881 |
+
"eval_runtime": 23.9121,
|
| 882 |
+
"eval_samples_per_second": 125.46,
|
| 883 |
+
"eval_steps_per_second": 3.931,
|
| 884 |
+
"step": 297
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.96,
|
| 888 |
+
"learning_rate": 0.0001211305518169583,
|
| 889 |
+
"loss": 0.6318,
|
| 890 |
+
"step": 300
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.96,
|
| 894 |
+
"eval_loss": 0.6398600935935974,
|
| 895 |
+
"eval_runtime": 23.9229,
|
| 896 |
+
"eval_samples_per_second": 125.403,
|
| 897 |
+
"eval_steps_per_second": 3.929,
|
| 898 |
+
"step": 300
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 4.0,
|
| 902 |
+
"eval_loss": 0.6393464207649231,
|
| 903 |
+
"eval_runtime": 23.9187,
|
| 904 |
+
"eval_samples_per_second": 125.425,
|
| 905 |
+
"eval_steps_per_second": 3.93,
|
| 906 |
+
"step": 303
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 4.04,
|
| 910 |
+
"eval_loss": 0.6392526626586914,
|
| 911 |
+
"eval_runtime": 23.9074,
|
| 912 |
+
"eval_samples_per_second": 125.484,
|
| 913 |
+
"eval_steps_per_second": 3.932,
|
| 914 |
+
"step": 306
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.08,
|
| 918 |
+
"eval_loss": 0.6389594078063965,
|
| 919 |
+
"eval_runtime": 23.918,
|
| 920 |
+
"eval_samples_per_second": 125.428,
|
| 921 |
+
"eval_steps_per_second": 3.93,
|
| 922 |
+
"step": 309
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 4.12,
|
| 926 |
+
"eval_loss": 0.6388808488845825,
|
| 927 |
+
"eval_runtime": 23.9158,
|
| 928 |
+
"eval_samples_per_second": 125.44,
|
| 929 |
+
"eval_steps_per_second": 3.93,
|
| 930 |
+
"step": 312
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 4.16,
|
| 934 |
+
"eval_loss": 0.6384025812149048,
|
| 935 |
+
"eval_runtime": 23.9176,
|
| 936 |
+
"eval_samples_per_second": 125.431,
|
| 937 |
+
"eval_steps_per_second": 3.93,
|
| 938 |
+
"step": 315
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.2,
|
| 942 |
+
"eval_loss": 0.6387144923210144,
|
| 943 |
+
"eval_runtime": 23.9047,
|
| 944 |
+
"eval_samples_per_second": 125.498,
|
| 945 |
+
"eval_steps_per_second": 3.932,
|
| 946 |
+
"step": 318
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 4.22,
|
| 950 |
+
"learning_rate": 0.00011574697173620459,
|
| 951 |
+
"loss": 0.6277,
|
| 952 |
+
"step": 320
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 4.24,
|
| 956 |
+
"eval_loss": 0.6377059817314148,
|
| 957 |
+
"eval_runtime": 23.9246,
|
| 958 |
+
"eval_samples_per_second": 125.394,
|
| 959 |
+
"eval_steps_per_second": 3.929,
|
| 960 |
+
"step": 321
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 4.28,
|
| 964 |
+
"eval_loss": 0.636981189250946,
|
| 965 |
+
"eval_runtime": 23.9459,
|
| 966 |
+
"eval_samples_per_second": 125.283,
|
| 967 |
+
"eval_steps_per_second": 3.926,
|
| 968 |
+
"step": 324
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 4.32,
|
| 972 |
+
"eval_loss": 0.6364036202430725,
|
| 973 |
+
"eval_runtime": 23.9206,
|
| 974 |
+
"eval_samples_per_second": 125.415,
|
| 975 |
+
"eval_steps_per_second": 3.93,
|
| 976 |
+
"step": 327
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 4.35,
|
| 980 |
+
"eval_loss": 0.6357031464576721,
|
| 981 |
+
"eval_runtime": 23.9187,
|
| 982 |
+
"eval_samples_per_second": 125.425,
|
| 983 |
+
"eval_steps_per_second": 3.93,
|
| 984 |
+
"step": 330
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 4.39,
|
| 988 |
+
"eval_loss": 0.6366411447525024,
|
| 989 |
+
"eval_runtime": 23.9159,
|
| 990 |
+
"eval_samples_per_second": 125.44,
|
| 991 |
+
"eval_steps_per_second": 3.93,
|
| 992 |
+
"step": 333
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 4.43,
|
| 996 |
+
"eval_loss": 0.6357526183128357,
|
| 997 |
+
"eval_runtime": 23.9135,
|
| 998 |
+
"eval_samples_per_second": 125.452,
|
| 999 |
+
"eval_steps_per_second": 3.931,
|
| 1000 |
+
"step": 336
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 4.47,
|
| 1004 |
+
"eval_loss": 0.6349912881851196,
|
| 1005 |
+
"eval_runtime": 23.9211,
|
| 1006 |
+
"eval_samples_per_second": 125.412,
|
| 1007 |
+
"eval_steps_per_second": 3.93,
|
| 1008 |
+
"step": 339
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 4.49,
|
| 1012 |
+
"learning_rate": 0.00011036339165545088,
|
| 1013 |
+
"loss": 0.6303,
|
| 1014 |
+
"step": 340
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 4.51,
|
| 1018 |
+
"eval_loss": 0.6343324184417725,
|
| 1019 |
+
"eval_runtime": 23.927,
|
| 1020 |
+
"eval_samples_per_second": 125.381,
|
| 1021 |
+
"eval_steps_per_second": 3.929,
|
| 1022 |
+
"step": 342
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 4.55,
|
| 1026 |
+
"eval_loss": 0.6347218751907349,
|
| 1027 |
+
"eval_runtime": 23.9489,
|
| 1028 |
+
"eval_samples_per_second": 125.267,
|
| 1029 |
+
"eval_steps_per_second": 3.925,
|
| 1030 |
+
"step": 345
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 4.59,
|
| 1034 |
+
"eval_loss": 0.6333290338516235,
|
| 1035 |
+
"eval_runtime": 23.9573,
|
| 1036 |
+
"eval_samples_per_second": 125.223,
|
| 1037 |
+
"eval_steps_per_second": 3.924,
|
| 1038 |
+
"step": 348
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 4.63,
|
| 1042 |
+
"eval_loss": 0.6328045129776001,
|
| 1043 |
+
"eval_runtime": 23.925,
|
| 1044 |
+
"eval_samples_per_second": 125.392,
|
| 1045 |
+
"eval_steps_per_second": 3.929,
|
| 1046 |
+
"step": 351
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.67,
|
| 1050 |
+
"eval_loss": 0.6328830718994141,
|
| 1051 |
+
"eval_runtime": 23.9277,
|
| 1052 |
+
"eval_samples_per_second": 125.378,
|
| 1053 |
+
"eval_steps_per_second": 3.928,
|
| 1054 |
+
"step": 354
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 4.71,
|
| 1058 |
+
"eval_loss": 0.6323109269142151,
|
| 1059 |
+
"eval_runtime": 23.9385,
|
| 1060 |
+
"eval_samples_per_second": 125.321,
|
| 1061 |
+
"eval_steps_per_second": 3.927,
|
| 1062 |
+
"step": 357
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 4.75,
|
| 1066 |
+
"learning_rate": 0.00010497981157469719,
|
| 1067 |
+
"loss": 0.6268,
|
| 1068 |
+
"step": 360
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 4.75,
|
| 1072 |
+
"eval_loss": 0.6327587366104126,
|
| 1073 |
+
"eval_runtime": 23.9389,
|
| 1074 |
+
"eval_samples_per_second": 125.319,
|
| 1075 |
+
"eval_steps_per_second": 3.927,
|
| 1076 |
+
"step": 360
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 4.79,
|
| 1080 |
+
"eval_loss": 0.6324266791343689,
|
| 1081 |
+
"eval_runtime": 23.9367,
|
| 1082 |
+
"eval_samples_per_second": 125.331,
|
| 1083 |
+
"eval_steps_per_second": 3.927,
|
| 1084 |
+
"step": 363
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 4.83,
|
| 1088 |
+
"eval_loss": 0.6320524215698242,
|
| 1089 |
+
"eval_runtime": 23.9373,
|
| 1090 |
+
"eval_samples_per_second": 125.327,
|
| 1091 |
+
"eval_steps_per_second": 3.927,
|
| 1092 |
+
"step": 366
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 4.87,
|
| 1096 |
+
"eval_loss": 0.6314539313316345,
|
| 1097 |
+
"eval_runtime": 23.9325,
|
| 1098 |
+
"eval_samples_per_second": 125.352,
|
| 1099 |
+
"eval_steps_per_second": 3.928,
|
| 1100 |
+
"step": 369
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 4.91,
|
| 1104 |
+
"eval_loss": 0.6318089365959167,
|
| 1105 |
+
"eval_runtime": 23.9345,
|
| 1106 |
+
"eval_samples_per_second": 125.342,
|
| 1107 |
+
"eval_steps_per_second": 3.927,
|
| 1108 |
+
"step": 372
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.95,
|
| 1112 |
+
"eval_loss": 0.6315808296203613,
|
| 1113 |
+
"eval_runtime": 23.924,
|
| 1114 |
+
"eval_samples_per_second": 125.397,
|
| 1115 |
+
"eval_steps_per_second": 3.929,
|
| 1116 |
+
"step": 375
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 4.99,
|
| 1120 |
+
"eval_loss": 0.630818247795105,
|
| 1121 |
+
"eval_runtime": 23.9285,
|
| 1122 |
+
"eval_samples_per_second": 125.373,
|
| 1123 |
+
"eval_steps_per_second": 3.928,
|
| 1124 |
+
"step": 378
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.01,
|
| 1128 |
+
"learning_rate": 9.959623149394348e-05,
|
| 1129 |
+
"loss": 0.6196,
|
| 1130 |
+
"step": 380
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 5.03,
|
| 1134 |
+
"eval_loss": 0.630248486995697,
|
| 1135 |
+
"eval_runtime": 23.9231,
|
| 1136 |
+
"eval_samples_per_second": 125.402,
|
| 1137 |
+
"eval_steps_per_second": 3.929,
|
| 1138 |
+
"step": 381
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 5.07,
|
| 1142 |
+
"eval_loss": 0.6306143403053284,
|
| 1143 |
+
"eval_runtime": 23.9242,
|
| 1144 |
+
"eval_samples_per_second": 125.396,
|
| 1145 |
+
"eval_steps_per_second": 3.929,
|
| 1146 |
+
"step": 384
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 5.11,
|
| 1150 |
+
"eval_loss": 0.6305729746818542,
|
| 1151 |
+
"eval_runtime": 23.9232,
|
| 1152 |
+
"eval_samples_per_second": 125.401,
|
| 1153 |
+
"eval_steps_per_second": 3.929,
|
| 1154 |
+
"step": 387
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 5.15,
|
| 1158 |
+
"eval_loss": 0.6302648782730103,
|
| 1159 |
+
"eval_runtime": 23.9286,
|
| 1160 |
+
"eval_samples_per_second": 125.373,
|
| 1161 |
+
"eval_steps_per_second": 3.928,
|
| 1162 |
+
"step": 390
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 5.19,
|
| 1166 |
+
"eval_loss": 0.6298710703849792,
|
| 1167 |
+
"eval_runtime": 23.9258,
|
| 1168 |
+
"eval_samples_per_second": 125.388,
|
| 1169 |
+
"eval_steps_per_second": 3.929,
|
| 1170 |
+
"step": 393
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 5.23,
|
| 1174 |
+
"eval_loss": 0.6298263669013977,
|
| 1175 |
+
"eval_runtime": 23.9284,
|
| 1176 |
+
"eval_samples_per_second": 125.374,
|
| 1177 |
+
"eval_steps_per_second": 3.928,
|
| 1178 |
+
"step": 396
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 5.27,
|
| 1182 |
+
"eval_loss": 0.6292470097541809,
|
| 1183 |
+
"eval_runtime": 23.9269,
|
| 1184 |
+
"eval_samples_per_second": 125.382,
|
| 1185 |
+
"eval_steps_per_second": 3.929,
|
| 1186 |
+
"step": 399
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 5.28,
|
| 1190 |
+
"learning_rate": 9.421265141318977e-05,
|
| 1191 |
+
"loss": 0.6146,
|
| 1192 |
+
"step": 400
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 5.3,
|
| 1196 |
+
"eval_loss": 0.6291049122810364,
|
| 1197 |
+
"eval_runtime": 23.9297,
|
| 1198 |
+
"eval_samples_per_second": 125.367,
|
| 1199 |
+
"eval_steps_per_second": 3.928,
|
| 1200 |
+
"step": 402
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 5.34,
|
| 1204 |
+
"eval_loss": 0.6296722292900085,
|
| 1205 |
+
"eval_runtime": 23.9386,
|
| 1206 |
+
"eval_samples_per_second": 125.321,
|
| 1207 |
+
"eval_steps_per_second": 3.927,
|
| 1208 |
+
"step": 405
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 5.38,
|
| 1212 |
+
"eval_loss": 0.6288275122642517,
|
| 1213 |
+
"eval_runtime": 23.9308,
|
| 1214 |
+
"eval_samples_per_second": 125.362,
|
| 1215 |
+
"eval_steps_per_second": 3.928,
|
| 1216 |
+
"step": 408
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 5.42,
|
| 1220 |
+
"eval_loss": 0.6288333535194397,
|
| 1221 |
+
"eval_runtime": 23.9261,
|
| 1222 |
+
"eval_samples_per_second": 125.386,
|
| 1223 |
+
"eval_steps_per_second": 3.929,
|
| 1224 |
+
"step": 411
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 5.46,
|
| 1228 |
+
"eval_loss": 0.6279690861701965,
|
| 1229 |
+
"eval_runtime": 23.9282,
|
| 1230 |
+
"eval_samples_per_second": 125.375,
|
| 1231 |
+
"eval_steps_per_second": 3.928,
|
| 1232 |
+
"step": 414
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 5.5,
|
| 1236 |
+
"eval_loss": 0.6275332570075989,
|
| 1237 |
+
"eval_runtime": 23.9215,
|
| 1238 |
+
"eval_samples_per_second": 125.41,
|
| 1239 |
+
"eval_steps_per_second": 3.93,
|
| 1240 |
+
"step": 417
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 5.54,
|
| 1244 |
+
"learning_rate": 8.882907133243608e-05,
|
| 1245 |
+
"loss": 0.6149,
|
| 1246 |
+
"step": 420
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 5.54,
|
| 1250 |
+
"eval_loss": 0.6279338598251343,
|
| 1251 |
+
"eval_runtime": 23.93,
|
| 1252 |
+
"eval_samples_per_second": 125.366,
|
| 1253 |
+
"eval_steps_per_second": 3.928,
|
| 1254 |
+
"step": 420
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 5.58,
|
| 1258 |
+
"eval_loss": 0.6271057724952698,
|
| 1259 |
+
"eval_runtime": 23.9158,
|
| 1260 |
+
"eval_samples_per_second": 125.44,
|
| 1261 |
+
"eval_steps_per_second": 3.93,
|
| 1262 |
+
"step": 423
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 5.62,
|
| 1266 |
+
"eval_loss": 0.6270298361778259,
|
| 1267 |
+
"eval_runtime": 23.9264,
|
| 1268 |
+
"eval_samples_per_second": 125.384,
|
| 1269 |
+
"eval_steps_per_second": 3.929,
|
| 1270 |
+
"step": 426
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 5.66,
|
| 1274 |
+
"eval_loss": 0.6271407604217529,
|
| 1275 |
+
"eval_runtime": 23.9362,
|
| 1276 |
+
"eval_samples_per_second": 125.333,
|
| 1277 |
+
"eval_steps_per_second": 3.927,
|
| 1278 |
+
"step": 429
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 5.7,
|
| 1282 |
+
"eval_loss": 0.6264240145683289,
|
| 1283 |
+
"eval_runtime": 23.9309,
|
| 1284 |
+
"eval_samples_per_second": 125.361,
|
| 1285 |
+
"eval_steps_per_second": 3.928,
|
| 1286 |
+
"step": 432
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 5.74,
|
| 1290 |
+
"eval_loss": 0.6263339519500732,
|
| 1291 |
+
"eval_runtime": 23.93,
|
| 1292 |
+
"eval_samples_per_second": 125.366,
|
| 1293 |
+
"eval_steps_per_second": 3.928,
|
| 1294 |
+
"step": 435
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 5.78,
|
| 1298 |
+
"eval_loss": 0.6256468296051025,
|
| 1299 |
+
"eval_runtime": 23.9252,
|
| 1300 |
+
"eval_samples_per_second": 125.391,
|
| 1301 |
+
"eval_steps_per_second": 3.929,
|
| 1302 |
+
"step": 438
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 5.81,
|
| 1306 |
+
"learning_rate": 8.344549125168237e-05,
|
| 1307 |
+
"loss": 0.6191,
|
| 1308 |
+
"step": 440
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 5.82,
|
| 1312 |
+
"eval_loss": 0.6260586977005005,
|
| 1313 |
+
"eval_runtime": 23.918,
|
| 1314 |
+
"eval_samples_per_second": 125.429,
|
| 1315 |
+
"eval_steps_per_second": 3.93,
|
| 1316 |
+
"step": 441
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 5.86,
|
| 1320 |
+
"eval_loss": 0.625337541103363,
|
| 1321 |
+
"eval_runtime": 23.9172,
|
| 1322 |
+
"eval_samples_per_second": 125.433,
|
| 1323 |
+
"eval_steps_per_second": 3.93,
|
| 1324 |
+
"step": 444
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 5.9,
|
| 1328 |
+
"eval_loss": 0.6246620416641235,
|
| 1329 |
+
"eval_runtime": 23.908,
|
| 1330 |
+
"eval_samples_per_second": 125.481,
|
| 1331 |
+
"eval_steps_per_second": 3.932,
|
| 1332 |
+
"step": 447
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 5.94,
|
| 1336 |
+
"eval_loss": 0.6251673698425293,
|
| 1337 |
+
"eval_runtime": 23.9137,
|
| 1338 |
+
"eval_samples_per_second": 125.451,
|
| 1339 |
+
"eval_steps_per_second": 3.931,
|
| 1340 |
+
"step": 450
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 5.98,
|
| 1344 |
+
"eval_loss": 0.6253092288970947,
|
| 1345 |
+
"eval_runtime": 23.9409,
|
| 1346 |
+
"eval_samples_per_second": 125.309,
|
| 1347 |
+
"eval_steps_per_second": 3.926,
|
| 1348 |
+
"step": 453
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 6.02,
|
| 1352 |
+
"eval_loss": 0.6245599389076233,
|
| 1353 |
+
"eval_runtime": 23.9233,
|
| 1354 |
+
"eval_samples_per_second": 125.401,
|
| 1355 |
+
"eval_steps_per_second": 3.929,
|
| 1356 |
+
"step": 456
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 6.06,
|
| 1360 |
+
"eval_loss": 0.6247097849845886,
|
| 1361 |
+
"eval_runtime": 23.9184,
|
| 1362 |
+
"eval_samples_per_second": 125.426,
|
| 1363 |
+
"eval_steps_per_second": 3.93,
|
| 1364 |
+
"step": 459
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 6.07,
|
| 1368 |
+
"learning_rate": 7.806191117092868e-05,
|
| 1369 |
+
"loss": 0.6075,
|
| 1370 |
+
"step": 460
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 6.1,
|
| 1374 |
+
"eval_loss": 0.6237714290618896,
|
| 1375 |
+
"eval_runtime": 23.9137,
|
| 1376 |
+
"eval_samples_per_second": 125.451,
|
| 1377 |
+
"eval_steps_per_second": 3.931,
|
| 1378 |
+
"step": 462
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 6.14,
|
| 1382 |
+
"eval_loss": 0.6239632964134216,
|
| 1383 |
+
"eval_runtime": 23.9142,
|
| 1384 |
+
"eval_samples_per_second": 125.449,
|
| 1385 |
+
"eval_steps_per_second": 3.931,
|
| 1386 |
+
"step": 465
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 6.18,
|
| 1390 |
+
"eval_loss": 0.6246253252029419,
|
| 1391 |
+
"eval_runtime": 23.9199,
|
| 1392 |
+
"eval_samples_per_second": 125.419,
|
| 1393 |
+
"eval_steps_per_second": 3.93,
|
| 1394 |
+
"step": 468
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 6.22,
|
| 1398 |
+
"eval_loss": 0.6236398220062256,
|
| 1399 |
+
"eval_runtime": 23.9169,
|
| 1400 |
+
"eval_samples_per_second": 125.434,
|
| 1401 |
+
"eval_steps_per_second": 3.93,
|
| 1402 |
+
"step": 471
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 6.25,
|
| 1406 |
+
"eval_loss": 0.6242309808731079,
|
| 1407 |
+
"eval_runtime": 23.9091,
|
| 1408 |
+
"eval_samples_per_second": 125.475,
|
| 1409 |
+
"eval_steps_per_second": 3.932,
|
| 1410 |
+
"step": 474
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 6.29,
|
| 1414 |
+
"eval_loss": 0.6236902475357056,
|
| 1415 |
+
"eval_runtime": 23.9117,
|
| 1416 |
+
"eval_samples_per_second": 125.462,
|
| 1417 |
+
"eval_steps_per_second": 3.931,
|
| 1418 |
+
"step": 477
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 6.33,
|
| 1422 |
+
"learning_rate": 7.267833109017497e-05,
|
| 1423 |
+
"loss": 0.6061,
|
| 1424 |
+
"step": 480
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 6.33,
|
| 1428 |
+
"eval_loss": 0.623267650604248,
|
| 1429 |
+
"eval_runtime": 23.9071,
|
| 1430 |
+
"eval_samples_per_second": 125.485,
|
| 1431 |
+
"eval_steps_per_second": 3.932,
|
| 1432 |
+
"step": 480
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 6.37,
|
| 1436 |
+
"eval_loss": 0.6238719820976257,
|
| 1437 |
+
"eval_runtime": 23.9206,
|
| 1438 |
+
"eval_samples_per_second": 125.415,
|
| 1439 |
+
"eval_steps_per_second": 3.93,
|
| 1440 |
+
"step": 483
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 6.41,
|
| 1444 |
+
"eval_loss": 0.6234752535820007,
|
| 1445 |
+
"eval_runtime": 23.914,
|
| 1446 |
+
"eval_samples_per_second": 125.449,
|
| 1447 |
+
"eval_steps_per_second": 3.931,
|
| 1448 |
+
"step": 486
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 6.45,
|
| 1452 |
+
"eval_loss": 0.6228368878364563,
|
| 1453 |
+
"eval_runtime": 23.9087,
|
| 1454 |
+
"eval_samples_per_second": 125.477,
|
| 1455 |
+
"eval_steps_per_second": 3.932,
|
| 1456 |
+
"step": 489
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 6.49,
|
| 1460 |
+
"eval_loss": 0.6226744055747986,
|
| 1461 |
+
"eval_runtime": 23.9118,
|
| 1462 |
+
"eval_samples_per_second": 125.461,
|
| 1463 |
+
"eval_steps_per_second": 3.931,
|
| 1464 |
+
"step": 492
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 6.53,
|
| 1468 |
+
"eval_loss": 0.622622013092041,
|
| 1469 |
+
"eval_runtime": 23.9341,
|
| 1470 |
+
"eval_samples_per_second": 125.344,
|
| 1471 |
+
"eval_steps_per_second": 3.927,
|
| 1472 |
+
"step": 495
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 6.57,
|
| 1476 |
+
"eval_loss": 0.6228298544883728,
|
| 1477 |
+
"eval_runtime": 23.9079,
|
| 1478 |
+
"eval_samples_per_second": 125.482,
|
| 1479 |
+
"eval_steps_per_second": 3.932,
|
| 1480 |
+
"step": 498
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 6.6,
|
| 1484 |
+
"learning_rate": 6.729475100942126e-05,
|
| 1485 |
+
"loss": 0.6043,
|
| 1486 |
+
"step": 500
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 6.61,
|
| 1490 |
+
"eval_loss": 0.6232237815856934,
|
| 1491 |
+
"eval_runtime": 23.8982,
|
| 1492 |
+
"eval_samples_per_second": 125.533,
|
| 1493 |
+
"eval_steps_per_second": 3.933,
|
| 1494 |
+
"step": 501
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 6.65,
|
| 1498 |
+
"eval_loss": 0.6218205690383911,
|
| 1499 |
+
"eval_runtime": 23.9059,
|
| 1500 |
+
"eval_samples_per_second": 125.492,
|
| 1501 |
+
"eval_steps_per_second": 3.932,
|
| 1502 |
+
"step": 504
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 6.69,
|
| 1506 |
+
"eval_loss": 0.621903657913208,
|
| 1507 |
+
"eval_runtime": 23.8991,
|
| 1508 |
+
"eval_samples_per_second": 125.528,
|
| 1509 |
+
"eval_steps_per_second": 3.933,
|
| 1510 |
+
"step": 507
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 6.73,
|
| 1514 |
+
"eval_loss": 0.622235894203186,
|
| 1515 |
+
"eval_runtime": 23.9024,
|
| 1516 |
+
"eval_samples_per_second": 125.51,
|
| 1517 |
+
"eval_steps_per_second": 3.933,
|
| 1518 |
+
"step": 510
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 6.77,
|
| 1522 |
+
"eval_loss": 0.6220830082893372,
|
| 1523 |
+
"eval_runtime": 23.8926,
|
| 1524 |
+
"eval_samples_per_second": 125.562,
|
| 1525 |
+
"eval_steps_per_second": 3.934,
|
| 1526 |
+
"step": 513
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 6.81,
|
| 1530 |
+
"eval_loss": 0.6220167875289917,
|
| 1531 |
+
"eval_runtime": 23.8965,
|
| 1532 |
+
"eval_samples_per_second": 125.542,
|
| 1533 |
+
"eval_steps_per_second": 3.934,
|
| 1534 |
+
"step": 516
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 6.85,
|
| 1538 |
+
"eval_loss": 0.6222782135009766,
|
| 1539 |
+
"eval_runtime": 23.908,
|
| 1540 |
+
"eval_samples_per_second": 125.481,
|
| 1541 |
+
"eval_steps_per_second": 3.932,
|
| 1542 |
+
"step": 519
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 6.86,
|
| 1546 |
+
"learning_rate": 6.191117092866757e-05,
|
| 1547 |
+
"loss": 0.6008,
|
| 1548 |
+
"step": 520
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 6.89,
|
| 1552 |
+
"eval_loss": 0.6216304302215576,
|
| 1553 |
+
"eval_runtime": 23.9036,
|
| 1554 |
+
"eval_samples_per_second": 125.504,
|
| 1555 |
+
"eval_steps_per_second": 3.932,
|
| 1556 |
+
"step": 522
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 6.93,
|
| 1560 |
+
"eval_loss": 0.6217759847640991,
|
| 1561 |
+
"eval_runtime": 23.9088,
|
| 1562 |
+
"eval_samples_per_second": 125.477,
|
| 1563 |
+
"eval_steps_per_second": 3.932,
|
| 1564 |
+
"step": 525
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 6.97,
|
| 1568 |
+
"eval_loss": 0.6214317083358765,
|
| 1569 |
+
"eval_runtime": 23.9177,
|
| 1570 |
+
"eval_samples_per_second": 125.43,
|
| 1571 |
+
"eval_steps_per_second": 3.93,
|
| 1572 |
+
"step": 528
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 7.01,
|
| 1576 |
+
"eval_loss": 0.6213416457176208,
|
| 1577 |
+
"eval_runtime": 23.9138,
|
| 1578 |
+
"eval_samples_per_second": 125.451,
|
| 1579 |
+
"eval_steps_per_second": 3.931,
|
| 1580 |
+
"step": 531
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 7.05,
|
| 1584 |
+
"eval_loss": 0.6217712163925171,
|
| 1585 |
+
"eval_runtime": 23.9141,
|
| 1586 |
+
"eval_samples_per_second": 125.449,
|
| 1587 |
+
"eval_steps_per_second": 3.931,
|
| 1588 |
+
"step": 534
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 7.09,
|
| 1592 |
+
"eval_loss": 0.6215860843658447,
|
| 1593 |
+
"eval_runtime": 23.9145,
|
| 1594 |
+
"eval_samples_per_second": 125.447,
|
| 1595 |
+
"eval_steps_per_second": 3.931,
|
| 1596 |
+
"step": 537
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 7.13,
|
| 1600 |
+
"learning_rate": 5.652759084791387e-05,
|
| 1601 |
+
"loss": 0.599,
|
| 1602 |
+
"step": 540
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 7.13,
|
| 1606 |
+
"eval_loss": 0.6211041808128357,
|
| 1607 |
+
"eval_runtime": 23.9125,
|
| 1608 |
+
"eval_samples_per_second": 125.457,
|
| 1609 |
+
"eval_steps_per_second": 3.931,
|
| 1610 |
+
"step": 540
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 7.17,
|
| 1614 |
+
"eval_loss": 0.6210355758666992,
|
| 1615 |
+
"eval_runtime": 23.911,
|
| 1616 |
+
"eval_samples_per_second": 125.465,
|
| 1617 |
+
"eval_steps_per_second": 3.931,
|
| 1618 |
+
"step": 543
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 7.2,
|
| 1622 |
+
"eval_loss": 0.6209889650344849,
|
| 1623 |
+
"eval_runtime": 23.9062,
|
| 1624 |
+
"eval_samples_per_second": 125.491,
|
| 1625 |
+
"eval_steps_per_second": 3.932,
|
| 1626 |
+
"step": 546
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 7.24,
|
| 1630 |
+
"eval_loss": 0.6205114126205444,
|
| 1631 |
+
"eval_runtime": 23.9227,
|
| 1632 |
+
"eval_samples_per_second": 125.404,
|
| 1633 |
+
"eval_steps_per_second": 3.929,
|
| 1634 |
+
"step": 549
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 7.28,
|
| 1638 |
+
"eval_loss": 0.6204013824462891,
|
| 1639 |
+
"eval_runtime": 23.9146,
|
| 1640 |
+
"eval_samples_per_second": 125.446,
|
| 1641 |
+
"eval_steps_per_second": 3.931,
|
| 1642 |
+
"step": 552
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 7.32,
|
| 1646 |
+
"eval_loss": 0.6202988028526306,
|
| 1647 |
+
"eval_runtime": 23.9015,
|
| 1648 |
+
"eval_samples_per_second": 125.515,
|
| 1649 |
+
"eval_steps_per_second": 3.933,
|
| 1650 |
+
"step": 555
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 7.36,
|
| 1654 |
+
"eval_loss": 0.6199727654457092,
|
| 1655 |
+
"eval_runtime": 23.9089,
|
| 1656 |
+
"eval_samples_per_second": 125.476,
|
| 1657 |
+
"eval_steps_per_second": 3.932,
|
| 1658 |
+
"step": 558
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 7.39,
|
| 1662 |
+
"learning_rate": 5.1144010767160164e-05,
|
| 1663 |
+
"loss": 0.5959,
|
| 1664 |
+
"step": 560
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 7.4,
|
| 1668 |
+
"eval_loss": 0.619968831539154,
|
| 1669 |
+
"eval_runtime": 23.9043,
|
| 1670 |
+
"eval_samples_per_second": 125.5,
|
| 1671 |
+
"eval_steps_per_second": 3.932,
|
| 1672 |
+
"step": 561
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 7.44,
|
| 1676 |
+
"eval_loss": 0.6202374696731567,
|
| 1677 |
+
"eval_runtime": 23.9117,
|
| 1678 |
+
"eval_samples_per_second": 125.461,
|
| 1679 |
+
"eval_steps_per_second": 3.931,
|
| 1680 |
+
"step": 564
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 7.48,
|
| 1684 |
+
"eval_loss": 0.6202066540718079,
|
| 1685 |
+
"eval_runtime": 23.908,
|
| 1686 |
+
"eval_samples_per_second": 125.481,
|
| 1687 |
+
"eval_steps_per_second": 3.932,
|
| 1688 |
+
"step": 567
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 7.52,
|
| 1692 |
+
"eval_loss": 0.6198835968971252,
|
| 1693 |
+
"eval_runtime": 23.9244,
|
| 1694 |
+
"eval_samples_per_second": 125.395,
|
| 1695 |
+
"eval_steps_per_second": 3.929,
|
| 1696 |
+
"step": 570
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 7.56,
|
| 1700 |
+
"eval_loss": 0.6199198961257935,
|
| 1701 |
+
"eval_runtime": 23.9263,
|
| 1702 |
+
"eval_samples_per_second": 125.385,
|
| 1703 |
+
"eval_steps_per_second": 3.929,
|
| 1704 |
+
"step": 573
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 7.6,
|
| 1708 |
+
"eval_loss": 0.6195517182350159,
|
| 1709 |
+
"eval_runtime": 23.9125,
|
| 1710 |
+
"eval_samples_per_second": 125.457,
|
| 1711 |
+
"eval_steps_per_second": 3.931,
|
| 1712 |
+
"step": 576
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 7.64,
|
| 1716 |
+
"eval_loss": 0.6192638278007507,
|
| 1717 |
+
"eval_runtime": 23.9168,
|
| 1718 |
+
"eval_samples_per_second": 125.435,
|
| 1719 |
+
"eval_steps_per_second": 3.93,
|
| 1720 |
+
"step": 579
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 7.65,
|
| 1724 |
+
"learning_rate": 4.576043068640646e-05,
|
| 1725 |
+
"loss": 0.5922,
|
| 1726 |
+
"step": 580
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 7.68,
|
| 1730 |
+
"eval_loss": 0.6196587085723877,
|
| 1731 |
+
"eval_runtime": 23.9305,
|
| 1732 |
+
"eval_samples_per_second": 125.363,
|
| 1733 |
+
"eval_steps_per_second": 3.928,
|
| 1734 |
+
"step": 582
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 7.72,
|
| 1738 |
+
"eval_loss": 0.6198856830596924,
|
| 1739 |
+
"eval_runtime": 23.919,
|
| 1740 |
+
"eval_samples_per_second": 125.423,
|
| 1741 |
+
"eval_steps_per_second": 3.93,
|
| 1742 |
+
"step": 585
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 7.76,
|
| 1746 |
+
"eval_loss": 0.6196783781051636,
|
| 1747 |
+
"eval_runtime": 23.9042,
|
| 1748 |
+
"eval_samples_per_second": 125.501,
|
| 1749 |
+
"eval_steps_per_second": 3.932,
|
| 1750 |
+
"step": 588
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 7.8,
|
| 1754 |
+
"eval_loss": 0.6192678809165955,
|
| 1755 |
+
"eval_runtime": 23.9057,
|
| 1756 |
+
"eval_samples_per_second": 125.493,
|
| 1757 |
+
"eval_steps_per_second": 3.932,
|
| 1758 |
+
"step": 591
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 7.84,
|
| 1762 |
+
"eval_loss": 0.6192264556884766,
|
| 1763 |
+
"eval_runtime": 23.909,
|
| 1764 |
+
"eval_samples_per_second": 125.476,
|
| 1765 |
+
"eval_steps_per_second": 3.932,
|
| 1766 |
+
"step": 594
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 7.88,
|
| 1770 |
+
"eval_loss": 0.6192458271980286,
|
| 1771 |
+
"eval_runtime": 23.8917,
|
| 1772 |
+
"eval_samples_per_second": 125.567,
|
| 1773 |
+
"eval_steps_per_second": 3.934,
|
| 1774 |
+
"step": 597
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 7.92,
|
| 1778 |
+
"learning_rate": 4.037685060565276e-05,
|
| 1779 |
+
"loss": 0.6028,
|
| 1780 |
+
"step": 600
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 7.92,
|
| 1784 |
+
"eval_loss": 0.6192883849143982,
|
| 1785 |
+
"eval_runtime": 23.9005,
|
| 1786 |
+
"eval_samples_per_second": 125.521,
|
| 1787 |
+
"eval_steps_per_second": 3.933,
|
| 1788 |
+
"step": 600
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 7.96,
|
| 1792 |
+
"eval_loss": 0.6194872856140137,
|
| 1793 |
+
"eval_runtime": 23.9118,
|
| 1794 |
+
"eval_samples_per_second": 125.461,
|
| 1795 |
+
"eval_steps_per_second": 3.931,
|
| 1796 |
+
"step": 603
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 8.0,
|
| 1800 |
+
"eval_loss": 0.619368314743042,
|
| 1801 |
+
"eval_runtime": 23.9121,
|
| 1802 |
+
"eval_samples_per_second": 125.46,
|
| 1803 |
+
"eval_steps_per_second": 3.931,
|
| 1804 |
+
"step": 606
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 8.04,
|
| 1808 |
+
"eval_loss": 0.6190740466117859,
|
| 1809 |
+
"eval_runtime": 23.9192,
|
| 1810 |
+
"eval_samples_per_second": 125.422,
|
| 1811 |
+
"eval_steps_per_second": 3.93,
|
| 1812 |
+
"step": 609
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 8.08,
|
| 1816 |
+
"eval_loss": 0.6190269589424133,
|
| 1817 |
+
"eval_runtime": 23.916,
|
| 1818 |
+
"eval_samples_per_second": 125.439,
|
| 1819 |
+
"eval_steps_per_second": 3.93,
|
| 1820 |
+
"step": 612
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 8.12,
|
| 1824 |
+
"eval_loss": 0.6189839839935303,
|
| 1825 |
+
"eval_runtime": 23.9063,
|
| 1826 |
+
"eval_samples_per_second": 125.49,
|
| 1827 |
+
"eval_steps_per_second": 3.932,
|
| 1828 |
+
"step": 615
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 8.16,
|
| 1832 |
+
"eval_loss": 0.618523895740509,
|
| 1833 |
+
"eval_runtime": 23.8928,
|
| 1834 |
+
"eval_samples_per_second": 125.561,
|
| 1835 |
+
"eval_steps_per_second": 3.934,
|
| 1836 |
+
"step": 618
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 8.18,
|
| 1840 |
+
"learning_rate": 3.499327052489906e-05,
|
| 1841 |
+
"loss": 0.5941,
|
| 1842 |
+
"step": 620
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 8.19,
|
| 1846 |
+
"eval_loss": 0.6187476515769958,
|
| 1847 |
+
"eval_runtime": 23.9016,
|
| 1848 |
+
"eval_samples_per_second": 125.515,
|
| 1849 |
+
"eval_steps_per_second": 3.933,
|
| 1850 |
+
"step": 621
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 8.23,
|
| 1854 |
+
"eval_loss": 0.6186578869819641,
|
| 1855 |
+
"eval_runtime": 23.9003,
|
| 1856 |
+
"eval_samples_per_second": 125.521,
|
| 1857 |
+
"eval_steps_per_second": 3.933,
|
| 1858 |
+
"step": 624
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 8.27,
|
| 1862 |
+
"eval_loss": 0.6190162897109985,
|
| 1863 |
+
"eval_runtime": 23.9155,
|
| 1864 |
+
"eval_samples_per_second": 125.442,
|
| 1865 |
+
"eval_steps_per_second": 3.931,
|
| 1866 |
+
"step": 627
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 8.31,
|
| 1870 |
+
"eval_loss": 0.6189883351325989,
|
| 1871 |
+
"eval_runtime": 23.9091,
|
| 1872 |
+
"eval_samples_per_second": 125.475,
|
| 1873 |
+
"eval_steps_per_second": 3.932,
|
| 1874 |
+
"step": 630
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 8.35,
|
| 1878 |
+
"eval_loss": 0.6184096932411194,
|
| 1879 |
+
"eval_runtime": 23.9042,
|
| 1880 |
+
"eval_samples_per_second": 125.501,
|
| 1881 |
+
"eval_steps_per_second": 3.932,
|
| 1882 |
+
"step": 633
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 8.39,
|
| 1886 |
+
"eval_loss": 0.6180031895637512,
|
| 1887 |
+
"eval_runtime": 23.9189,
|
| 1888 |
+
"eval_samples_per_second": 125.424,
|
| 1889 |
+
"eval_steps_per_second": 3.93,
|
| 1890 |
+
"step": 636
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 8.43,
|
| 1894 |
+
"eval_loss": 0.6179867386817932,
|
| 1895 |
+
"eval_runtime": 23.922,
|
| 1896 |
+
"eval_samples_per_second": 125.407,
|
| 1897 |
+
"eval_steps_per_second": 3.929,
|
| 1898 |
+
"step": 639
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 8.45,
|
| 1902 |
+
"learning_rate": 2.960969044414536e-05,
|
| 1903 |
+
"loss": 0.5906,
|
| 1904 |
+
"step": 640
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 8.47,
|
| 1908 |
+
"eval_loss": 0.6182823777198792,
|
| 1909 |
+
"eval_runtime": 23.9158,
|
| 1910 |
+
"eval_samples_per_second": 125.44,
|
| 1911 |
+
"eval_steps_per_second": 3.93,
|
| 1912 |
+
"step": 642
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 8.51,
|
| 1916 |
+
"eval_loss": 0.6179353594779968,
|
| 1917 |
+
"eval_runtime": 23.914,
|
| 1918 |
+
"eval_samples_per_second": 125.45,
|
| 1919 |
+
"eval_steps_per_second": 3.931,
|
| 1920 |
+
"step": 645
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 8.55,
|
| 1924 |
+
"eval_loss": 0.6178385615348816,
|
| 1925 |
+
"eval_runtime": 23.9079,
|
| 1926 |
+
"eval_samples_per_second": 125.482,
|
| 1927 |
+
"eval_steps_per_second": 3.932,
|
| 1928 |
+
"step": 648
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 8.59,
|
| 1932 |
+
"eval_loss": 0.6179868578910828,
|
| 1933 |
+
"eval_runtime": 23.9189,
|
| 1934 |
+
"eval_samples_per_second": 125.424,
|
| 1935 |
+
"eval_steps_per_second": 3.93,
|
| 1936 |
+
"step": 651
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 8.63,
|
| 1940 |
+
"eval_loss": 0.6179595589637756,
|
| 1941 |
+
"eval_runtime": 23.9327,
|
| 1942 |
+
"eval_samples_per_second": 125.352,
|
| 1943 |
+
"eval_steps_per_second": 3.928,
|
| 1944 |
+
"step": 654
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 8.67,
|
| 1948 |
+
"eval_loss": 0.6177854537963867,
|
| 1949 |
+
"eval_runtime": 23.9123,
|
| 1950 |
+
"eval_samples_per_second": 125.458,
|
| 1951 |
+
"eval_steps_per_second": 3.931,
|
| 1952 |
+
"step": 657
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 8.71,
|
| 1956 |
+
"learning_rate": 2.422611036339166e-05,
|
| 1957 |
+
"loss": 0.5908,
|
| 1958 |
+
"step": 660
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 8.71,
|
| 1962 |
+
"eval_loss": 0.6179735064506531,
|
| 1963 |
+
"eval_runtime": 23.9135,
|
| 1964 |
+
"eval_samples_per_second": 125.452,
|
| 1965 |
+
"eval_steps_per_second": 3.931,
|
| 1966 |
+
"step": 660
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 8.75,
|
| 1970 |
+
"eval_loss": 0.6180996298789978,
|
| 1971 |
+
"eval_runtime": 23.9038,
|
| 1972 |
+
"eval_samples_per_second": 125.503,
|
| 1973 |
+
"eval_steps_per_second": 3.932,
|
| 1974 |
+
"step": 663
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 8.79,
|
| 1978 |
+
"eval_loss": 0.6181532740592957,
|
| 1979 |
+
"eval_runtime": 23.9058,
|
| 1980 |
+
"eval_samples_per_second": 125.492,
|
| 1981 |
+
"eval_steps_per_second": 3.932,
|
| 1982 |
+
"step": 666
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 8.83,
|
| 1986 |
+
"eval_loss": 0.6176265478134155,
|
| 1987 |
+
"eval_runtime": 23.9128,
|
| 1988 |
+
"eval_samples_per_second": 125.456,
|
| 1989 |
+
"eval_steps_per_second": 3.931,
|
| 1990 |
+
"step": 669
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 8.87,
|
| 1994 |
+
"eval_loss": 0.617388904094696,
|
| 1995 |
+
"eval_runtime": 23.8906,
|
| 1996 |
+
"eval_samples_per_second": 125.572,
|
| 1997 |
+
"eval_steps_per_second": 3.935,
|
| 1998 |
+
"step": 672
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 8.91,
|
| 2002 |
+
"eval_loss": 0.6174699068069458,
|
| 2003 |
+
"eval_runtime": 23.9051,
|
| 2004 |
+
"eval_samples_per_second": 125.496,
|
| 2005 |
+
"eval_steps_per_second": 3.932,
|
| 2006 |
+
"step": 675
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 8.95,
|
| 2010 |
+
"eval_loss": 0.6174932718276978,
|
| 2011 |
+
"eval_runtime": 23.8969,
|
| 2012 |
+
"eval_samples_per_second": 125.54,
|
| 2013 |
+
"eval_steps_per_second": 3.934,
|
| 2014 |
+
"step": 678
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 8.97,
|
| 2018 |
+
"learning_rate": 1.8842530282637954e-05,
|
| 2019 |
+
"loss": 0.5937,
|
| 2020 |
+
"step": 680
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 8.99,
|
| 2024 |
+
"eval_loss": 0.6174784898757935,
|
| 2025 |
+
"eval_runtime": 23.9056,
|
| 2026 |
+
"eval_samples_per_second": 125.494,
|
| 2027 |
+
"eval_steps_per_second": 3.932,
|
| 2028 |
+
"step": 681
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 9.03,
|
| 2032 |
+
"eval_loss": 0.6175104975700378,
|
| 2033 |
+
"eval_runtime": 23.9136,
|
| 2034 |
+
"eval_samples_per_second": 125.452,
|
| 2035 |
+
"eval_steps_per_second": 3.931,
|
| 2036 |
+
"step": 684
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 9.07,
|
| 2040 |
+
"eval_loss": 0.6173563599586487,
|
| 2041 |
+
"eval_runtime": 23.9237,
|
| 2042 |
+
"eval_samples_per_second": 125.399,
|
| 2043 |
+
"eval_steps_per_second": 3.929,
|
| 2044 |
+
"step": 687
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 9.11,
|
| 2048 |
+
"eval_loss": 0.6172643899917603,
|
| 2049 |
+
"eval_runtime": 23.9021,
|
| 2050 |
+
"eval_samples_per_second": 125.512,
|
| 2051 |
+
"eval_steps_per_second": 3.933,
|
| 2052 |
+
"step": 690
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 9.14,
|
| 2056 |
+
"eval_loss": 0.6172318458557129,
|
| 2057 |
+
"eval_runtime": 23.9242,
|
| 2058 |
+
"eval_samples_per_second": 125.396,
|
| 2059 |
+
"eval_steps_per_second": 3.929,
|
| 2060 |
+
"step": 693
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 9.18,
|
| 2064 |
+
"eval_loss": 0.617030680179596,
|
| 2065 |
+
"eval_runtime": 23.9184,
|
| 2066 |
+
"eval_samples_per_second": 125.427,
|
| 2067 |
+
"eval_steps_per_second": 3.93,
|
| 2068 |
+
"step": 696
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 9.22,
|
| 2072 |
+
"eval_loss": 0.6169764995574951,
|
| 2073 |
+
"eval_runtime": 23.9104,
|
| 2074 |
+
"eval_samples_per_second": 125.468,
|
| 2075 |
+
"eval_steps_per_second": 3.931,
|
| 2076 |
+
"step": 699
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 9.24,
|
| 2080 |
+
"learning_rate": 1.3458950201884254e-05,
|
| 2081 |
+
"loss": 0.5867,
|
| 2082 |
+
"step": 700
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 9.26,
|
| 2086 |
+
"eval_loss": 0.6171083450317383,
|
| 2087 |
+
"eval_runtime": 23.9126,
|
| 2088 |
+
"eval_samples_per_second": 125.457,
|
| 2089 |
+
"eval_steps_per_second": 3.931,
|
| 2090 |
+
"step": 702
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 9.3,
|
| 2094 |
+
"eval_loss": 0.6171473860740662,
|
| 2095 |
+
"eval_runtime": 23.9015,
|
| 2096 |
+
"eval_samples_per_second": 125.515,
|
| 2097 |
+
"eval_steps_per_second": 3.933,
|
| 2098 |
+
"step": 705
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 9.34,
|
| 2102 |
+
"eval_loss": 0.6170982122421265,
|
| 2103 |
+
"eval_runtime": 23.9023,
|
| 2104 |
+
"eval_samples_per_second": 125.511,
|
| 2105 |
+
"eval_steps_per_second": 3.933,
|
| 2106 |
+
"step": 708
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 9.38,
|
| 2110 |
+
"eval_loss": 0.6169420480728149,
|
| 2111 |
+
"eval_runtime": 23.9074,
|
| 2112 |
+
"eval_samples_per_second": 125.484,
|
| 2113 |
+
"eval_steps_per_second": 3.932,
|
| 2114 |
+
"step": 711
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 9.42,
|
| 2118 |
+
"eval_loss": 0.6168730854988098,
|
| 2119 |
+
"eval_runtime": 23.9047,
|
| 2120 |
+
"eval_samples_per_second": 125.498,
|
| 2121 |
+
"eval_steps_per_second": 3.932,
|
| 2122 |
+
"step": 714
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 9.46,
|
| 2126 |
+
"eval_loss": 0.6168663501739502,
|
| 2127 |
+
"eval_runtime": 23.9161,
|
| 2128 |
+
"eval_samples_per_second": 125.439,
|
| 2129 |
+
"eval_steps_per_second": 3.93,
|
| 2130 |
+
"step": 717
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 9.5,
|
| 2134 |
+
"learning_rate": 8.075370121130552e-06,
|
| 2135 |
+
"loss": 0.5891,
|
| 2136 |
+
"step": 720
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 9.5,
|
| 2140 |
+
"eval_loss": 0.6167708039283752,
|
| 2141 |
+
"eval_runtime": 23.9114,
|
| 2142 |
+
"eval_samples_per_second": 125.463,
|
| 2143 |
+
"eval_steps_per_second": 3.931,
|
| 2144 |
+
"step": 720
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 9.54,
|
| 2148 |
+
"eval_loss": 0.6167441010475159,
|
| 2149 |
+
"eval_runtime": 23.9086,
|
| 2150 |
+
"eval_samples_per_second": 125.478,
|
| 2151 |
+
"eval_steps_per_second": 3.932,
|
| 2152 |
+
"step": 723
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 9.58,
|
| 2156 |
+
"eval_loss": 0.6168937683105469,
|
| 2157 |
+
"eval_runtime": 23.9122,
|
| 2158 |
+
"eval_samples_per_second": 125.459,
|
| 2159 |
+
"eval_steps_per_second": 3.931,
|
| 2160 |
+
"step": 726
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 9.62,
|
| 2164 |
+
"eval_loss": 0.6170821189880371,
|
| 2165 |
+
"eval_runtime": 23.9065,
|
| 2166 |
+
"eval_samples_per_second": 125.489,
|
| 2167 |
+
"eval_steps_per_second": 3.932,
|
| 2168 |
+
"step": 729
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 9.66,
|
| 2172 |
+
"eval_loss": 0.6171652674674988,
|
| 2173 |
+
"eval_runtime": 23.9131,
|
| 2174 |
+
"eval_samples_per_second": 125.454,
|
| 2175 |
+
"eval_steps_per_second": 3.931,
|
| 2176 |
+
"step": 732
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 9.7,
|
| 2180 |
+
"eval_loss": 0.6171714663505554,
|
| 2181 |
+
"eval_runtime": 23.9057,
|
| 2182 |
+
"eval_samples_per_second": 125.493,
|
| 2183 |
+
"eval_steps_per_second": 3.932,
|
| 2184 |
+
"step": 735
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 9.74,
|
| 2188 |
+
"eval_loss": 0.6170787811279297,
|
| 2189 |
+
"eval_runtime": 23.928,
|
| 2190 |
+
"eval_samples_per_second": 125.376,
|
| 2191 |
+
"eval_steps_per_second": 3.928,
|
| 2192 |
+
"step": 738
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 9.76,
|
| 2196 |
+
"learning_rate": 2.6917900403768505e-06,
|
| 2197 |
+
"loss": 0.5843,
|
| 2198 |
+
"step": 740
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 9.78,
|
| 2202 |
+
"eval_loss": 0.6169885993003845,
|
| 2203 |
+
"eval_runtime": 23.9276,
|
| 2204 |
+
"eval_samples_per_second": 125.378,
|
| 2205 |
+
"eval_steps_per_second": 3.929,
|
| 2206 |
+
"step": 741
|
| 2207 |
+
}
|
| 2208 |
+
],
|
| 2209 |
+
"max_steps": 750,
|
| 2210 |
+
"num_train_epochs": 10,
|
| 2211 |
+
"total_flos": 7.831063404654625e+18,
|
| 2212 |
+
"trial_name": null,
|
| 2213 |
+
"trial_params": null
|
| 2214 |
+
}
|
vicuna-7b_english-cot+auto-cot_0.0002/lora/checkpoint-741/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d5a386ce8addef927bca0e390e9534e0877a3e4e00f222f83967dd78c49527a
|
| 3 |
+
size 4027
|