liuhaotian commited on
Commit
430340d
1 Parent(s): a151f78

Upload model checkpoints

Browse files
adapter_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "./checkpoints/vicuna-7b-v1-1",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "lora_alpha": 16,
8
+ "lora_dropout": 0.05,
9
+ "modules_to_save": null,
10
+ "peft_type": "LORA",
11
+ "r": 64,
12
+ "target_modules": [
13
+ "v_proj",
14
+ "k_proj",
15
+ "gate_proj",
16
+ "o_proj",
17
+ "up_proj",
18
+ "down_proj",
19
+ "q_proj"
20
+ ],
21
+ "task_type": "CAUSAL_LM"
22
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9953e646503a198cb65f539c6cf98fc4b30f976582f6773c645c2d8d0d21b6f
3
+ size 319970957
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "./checkpoints/vicuna-7b-v1-1",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "freeze_mm_mlp_adapter": false,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 4096,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 11008,
13
+ "max_position_embeddings": 2048,
14
+ "mm_hidden_size": 1024,
15
+ "mm_use_im_start_end": true,
16
+ "mm_vision_select_layer": -2,
17
+ "mm_vision_tower": "openai/clip-vit-large-patch14",
18
+ "model_type": "llava",
19
+ "num_attention_heads": 32,
20
+ "num_hidden_layers": 32,
21
+ "pad_token_id": 0,
22
+ "rms_norm_eps": 1e-06,
23
+ "sep_image_conv_front": false,
24
+ "tie_word_embeddings": false,
25
+ "torch_dtype": "float16",
26
+ "transformers_version": "4.28.0.dev0",
27
+ "tune_mm_mlp_adapter": false,
28
+ "use_cache": false,
29
+ "use_mm_proj": true,
30
+ "vocab_size": 32003
31
+ }
non_lora_trainables.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b19936ff6f6fb8016cd3e000480e8ef52c12d912b7d960013d14158043c8373
3
+ size 532735719
trainer_state.json ADDED
@@ -0,0 +1,3775 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "global_step": 625,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 1.0526315789473684e-05,
13
+ "loss": 1.6094,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 2.105263157894737e-05,
19
+ "loss": 1.6328,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 3.157894736842105e-05,
25
+ "loss": 1.6055,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.01,
30
+ "learning_rate": 4.210526315789474e-05,
31
+ "loss": 1.6562,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.01,
36
+ "learning_rate": 5.2631578947368424e-05,
37
+ "loss": 1.5625,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 6.31578947368421e-05,
43
+ "loss": 1.4961,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.01,
48
+ "learning_rate": 7.368421052631579e-05,
49
+ "loss": 1.457,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 8.421052631578948e-05,
55
+ "loss": 1.3867,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 9.473684210526316e-05,
61
+ "loss": 1.3438,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.02,
66
+ "learning_rate": 0.00010526315789473685,
67
+ "loss": 1.3203,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.02,
72
+ "learning_rate": 0.00011578947368421053,
73
+ "loss": 1.2617,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.02,
78
+ "learning_rate": 0.0001263157894736842,
79
+ "loss": 1.2266,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.02,
84
+ "learning_rate": 0.0001368421052631579,
85
+ "loss": 1.2227,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.02,
90
+ "learning_rate": 0.00014736842105263158,
91
+ "loss": 1.2109,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.02,
96
+ "learning_rate": 0.00015789473684210527,
97
+ "loss": 1.1953,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.03,
102
+ "learning_rate": 0.00016842105263157895,
103
+ "loss": 1.2148,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.03,
108
+ "learning_rate": 0.00017894736842105264,
109
+ "loss": 1.1914,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.03,
114
+ "learning_rate": 0.00018947368421052632,
115
+ "loss": 1.1914,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.03,
120
+ "learning_rate": 0.0002,
121
+ "loss": 1.1875,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.03,
126
+ "learning_rate": 0.00019999865623437013,
127
+ "loss": 1.207,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.03,
132
+ "learning_rate": 0.00019999462497359466,
133
+ "loss": 1.1758,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.04,
138
+ "learning_rate": 0.00019998790632601496,
139
+ "loss": 1.1758,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.04,
144
+ "learning_rate": 0.0001999785004721968,
145
+ "loss": 1.1758,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.04,
150
+ "learning_rate": 0.00019996640766492543,
151
+ "loss": 1.168,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.04,
156
+ "learning_rate": 0.00019995162822919883,
157
+ "loss": 1.1797,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.04,
162
+ "learning_rate": 0.00019993416256221895,
163
+ "loss": 1.1445,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.04,
168
+ "learning_rate": 0.00019991401113338104,
169
+ "loss": 1.1367,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.04,
174
+ "learning_rate": 0.00019989117448426108,
175
+ "loss": 1.1367,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.05,
180
+ "learning_rate": 0.00019986565322860115,
181
+ "loss": 1.1484,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.05,
186
+ "learning_rate": 0.00019983744805229296,
187
+ "loss": 1.1406,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.05,
192
+ "learning_rate": 0.00019980655971335945,
193
+ "loss": 1.1367,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.05,
198
+ "learning_rate": 0.00019977298904193437,
199
+ "loss": 1.1523,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.05,
204
+ "learning_rate": 0.00019973673694024,
205
+ "loss": 1.1328,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.05,
210
+ "learning_rate": 0.00019969780438256293,
211
+ "loss": 1.1367,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.06,
216
+ "learning_rate": 0.0001996561924152278,
217
+ "loss": 1.1445,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.06,
222
+ "learning_rate": 0.0001996119021565693,
223
+ "loss": 1.1133,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.06,
228
+ "learning_rate": 0.0001995649347969019,
229
+ "loss": 1.1367,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.06,
234
+ "learning_rate": 0.00019951529159848805,
235
+ "loss": 1.1367,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.06,
240
+ "learning_rate": 0.00019946297389550433,
241
+ "loss": 1.1211,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.06,
246
+ "learning_rate": 0.00019940798309400526,
247
+ "loss": 1.1094,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.07,
252
+ "learning_rate": 0.0001993503206718859,
253
+ "loss": 1.1445,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.07,
258
+ "learning_rate": 0.00019928998817884182,
259
+ "loss": 1.1289,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.07,
264
+ "learning_rate": 0.00019922698723632767,
265
+ "loss": 1.1289,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.07,
270
+ "learning_rate": 0.00019916131953751342,
271
+ "loss": 1.1328,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.07,
276
+ "learning_rate": 0.00019909298684723904,
277
+ "loss": 1.1289,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.07,
282
+ "learning_rate": 0.00019902199100196697,
283
+ "loss": 1.1094,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.08,
288
+ "learning_rate": 0.00019894833390973266,
289
+ "loss": 1.1094,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.08,
294
+ "learning_rate": 0.00019887201755009357,
295
+ "loss": 1.1016,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.08,
300
+ "learning_rate": 0.0001987930439740757,
301
+ "loss": 1.1172,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.08,
306
+ "learning_rate": 0.00019871141530411853,
307
+ "loss": 1.1094,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.08,
312
+ "learning_rate": 0.0001986271337340182,
313
+ "loss": 1.0781,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.08,
318
+ "learning_rate": 0.00019854020152886814,
319
+ "loss": 1.082,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.08,
324
+ "learning_rate": 0.0001984506210249986,
325
+ "loss": 1.1172,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.09,
330
+ "learning_rate": 0.00019835839462991361,
331
+ "loss": 1.0898,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.09,
336
+ "learning_rate": 0.00019826352482222638,
337
+ "loss": 1.0898,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.09,
342
+ "learning_rate": 0.00019816601415159263,
343
+ "loss": 1.1172,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.09,
348
+ "learning_rate": 0.0001980658652386421,
349
+ "loss": 1.1445,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.09,
354
+ "learning_rate": 0.00019796308077490817,
355
+ "loss": 1.1211,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.09,
360
+ "learning_rate": 0.00019785766352275542,
361
+ "loss": 1.0703,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.1,
366
+ "learning_rate": 0.00019774961631530545,
367
+ "loss": 1.1016,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.1,
372
+ "learning_rate": 0.00019763894205636072,
373
+ "loss": 1.0898,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.1,
378
+ "learning_rate": 0.00019752564372032657,
379
+ "loss": 1.0742,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.1,
384
+ "learning_rate": 0.00019740972435213115,
385
+ "loss": 1.0898,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.1,
390
+ "learning_rate": 0.00019729118706714375,
391
+ "loss": 1.1172,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.1,
396
+ "learning_rate": 0.00019717003505109095,
397
+ "loss": 1.1094,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.11,
402
+ "learning_rate": 0.00019704627155997108,
403
+ "loss": 1.0742,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.11,
408
+ "learning_rate": 0.00019691989991996663,
409
+ "loss": 1.082,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.11,
414
+ "learning_rate": 0.0001967909235273549,
415
+ "loss": 1.0977,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.11,
420
+ "learning_rate": 0.00019665934584841682,
421
+ "loss": 1.0469,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.11,
426
+ "learning_rate": 0.00019652517041934356,
427
+ "loss": 1.0625,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.11,
432
+ "learning_rate": 0.00019638840084614182,
433
+ "loss": 1.0703,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.12,
438
+ "learning_rate": 0.00019624904080453655,
439
+ "loss": 1.0742,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.12,
444
+ "learning_rate": 0.00019610709403987246,
445
+ "loss": 1.1133,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.12,
450
+ "learning_rate": 0.00019596256436701324,
451
+ "loss": 1.1016,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.12,
456
+ "learning_rate": 0.000195815455670239,
457
+ "loss": 1.0664,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.12,
462
+ "learning_rate": 0.00019566577190314197,
463
+ "loss": 1.0781,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.12,
468
+ "learning_rate": 0.0001955135170885202,
469
+ "loss": 1.0625,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.12,
474
+ "learning_rate": 0.00019535869531826937,
475
+ "loss": 1.0742,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.13,
480
+ "learning_rate": 0.00019520131075327298,
481
+ "loss": 1.0742,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.13,
486
+ "learning_rate": 0.00019504136762329047,
487
+ "loss": 1.0859,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.13,
492
+ "learning_rate": 0.00019487887022684336,
493
+ "loss": 1.0938,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.13,
498
+ "learning_rate": 0.00019471382293110003,
499
+ "loss": 1.0625,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.13,
504
+ "learning_rate": 0.00019454623017175812,
505
+ "loss": 1.0977,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.13,
510
+ "learning_rate": 0.00019437609645292546,
511
+ "loss": 1.0547,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.14,
516
+ "learning_rate": 0.0001942034263469989,
517
+ "loss": 1.0781,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.14,
522
+ "learning_rate": 0.00019402822449454153,
523
+ "loss": 1.0938,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.14,
528
+ "learning_rate": 0.00019385049560415794,
529
+ "loss": 1.0742,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.14,
534
+ "learning_rate": 0.00019367024445236754,
535
+ "loss": 1.043,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.14,
540
+ "learning_rate": 0.00019348747588347637,
541
+ "loss": 1.0742,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.14,
546
+ "learning_rate": 0.00019330219480944694,
547
+ "loss": 1.0625,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.15,
552
+ "learning_rate": 0.00019311440620976597,
553
+ "loss": 1.1055,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.15,
558
+ "learning_rate": 0.0001929241151313108,
559
+ "loss": 1.125,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.15,
564
+ "learning_rate": 0.00019273132668821364,
565
+ "loss": 1.0938,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.15,
570
+ "learning_rate": 0.00019253604606172417,
571
+ "loss": 1.0859,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.15,
576
+ "learning_rate": 0.00019233827850007027,
577
+ "loss": 1.0508,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.15,
582
+ "learning_rate": 0.00019213802931831696,
583
+ "loss": 1.082,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.16,
588
+ "learning_rate": 0.00019193530389822363,
589
+ "loss": 1.0938,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.16,
594
+ "learning_rate": 0.00019173010768809933,
595
+ "loss": 1.0781,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.16,
600
+ "learning_rate": 0.0001915224462026563,
601
+ "loss": 1.043,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.16,
606
+ "learning_rate": 0.00019131232502286188,
607
+ "loss": 1.0938,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.16,
612
+ "learning_rate": 0.0001910997497957885,
613
+ "loss": 1.0742,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.16,
618
+ "learning_rate": 0.00019088472623446183,
619
+ "loss": 1.0781,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.16,
624
+ "learning_rate": 0.00019066726011770726,
625
+ "loss": 1.0742,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.17,
630
+ "learning_rate": 0.0001904473572899947,
631
+ "loss": 1.0508,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.17,
636
+ "learning_rate": 0.00019022502366128135,
637
+ "loss": 1.0664,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.17,
642
+ "learning_rate": 0.00019000026520685302,
643
+ "loss": 1.0508,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.17,
648
+ "learning_rate": 0.0001897730879671634,
649
+ "loss": 1.0586,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.17,
654
+ "learning_rate": 0.00018954349804767184,
655
+ "loss": 1.0703,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.17,
660
+ "learning_rate": 0.00018931150161867916,
661
+ "loss": 1.0625,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.18,
666
+ "learning_rate": 0.00018907710491516199,
667
+ "loss": 1.0664,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.18,
672
+ "learning_rate": 0.0001888403142366049,
673
+ "loss": 1.0547,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.18,
678
+ "learning_rate": 0.00018860113594683148,
679
+ "loss": 1.0703,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.18,
684
+ "learning_rate": 0.00018835957647383303,
685
+ "loss": 1.043,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.18,
690
+ "learning_rate": 0.00018811564230959588,
691
+ "loss": 1.0547,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.18,
696
+ "learning_rate": 0.00018786934000992688,
697
+ "loss": 1.082,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.19,
702
+ "learning_rate": 0.00018762067619427746,
703
+ "loss": 1.0664,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.19,
708
+ "learning_rate": 0.00018736965754556528,
709
+ "loss": 1.0625,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.19,
714
+ "learning_rate": 0.00018711629080999504,
715
+ "loss": 1.0547,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.19,
720
+ "learning_rate": 0.00018686058279687698,
721
+ "loss": 1.0586,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.19,
726
+ "learning_rate": 0.00018660254037844388,
727
+ "loss": 1.043,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.19,
732
+ "learning_rate": 0.00018634217048966637,
733
+ "loss": 1.0859,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.2,
738
+ "learning_rate": 0.0001860794801280666,
739
+ "loss": 1.0781,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.2,
744
+ "learning_rate": 0.0001858144763535302,
745
+ "loss": 1.0664,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.2,
750
+ "learning_rate": 0.0001855471662881164,
751
+ "loss": 1.0625,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.2,
756
+ "learning_rate": 0.00018527755711586678,
757
+ "loss": 1.0391,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.2,
762
+ "learning_rate": 0.00018500565608261214,
763
+ "loss": 1.0938,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.2,
768
+ "learning_rate": 0.00018473147049577774,
769
+ "loss": 1.0586,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.2,
774
+ "learning_rate": 0.00018445500772418697,
775
+ "loss": 1.0547,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.21,
780
+ "learning_rate": 0.00018417627519786315,
781
+ "loss": 1.0547,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.21,
786
+ "learning_rate": 0.00018389528040783012,
787
+ "loss": 1.082,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.21,
792
+ "learning_rate": 0.00018361203090591071,
793
+ "loss": 1.0977,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.21,
798
+ "learning_rate": 0.00018332653430452376,
799
+ "loss": 1.0781,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.21,
804
+ "learning_rate": 0.00018303879827647975,
805
+ "loss": 1.0742,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.21,
810
+ "learning_rate": 0.00018274883055477436,
811
+ "loss": 1.0625,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.22,
816
+ "learning_rate": 0.00018245663893238075,
817
+ "loss": 1.0625,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.22,
822
+ "learning_rate": 0.00018216223126204007,
823
+ "loss": 1.0469,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.22,
828
+ "learning_rate": 0.00018186561545605054,
829
+ "loss": 1.0781,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.22,
834
+ "learning_rate": 0.00018156679948605467,
835
+ "loss": 1.0898,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.22,
840
+ "learning_rate": 0.00018126579138282503,
841
+ "loss": 1.043,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.22,
846
+ "learning_rate": 0.0001809625992360485,
847
+ "loss": 1.1016,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.23,
852
+ "learning_rate": 0.00018065723119410884,
853
+ "loss": 1.0547,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.23,
858
+ "learning_rate": 0.00018034969546386757,
859
+ "loss": 1.0273,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.23,
864
+ "learning_rate": 0.0001800400003104436,
865
+ "loss": 1.0703,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.23,
870
+ "learning_rate": 0.00017972815405699103,
871
+ "loss": 1.0859,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.23,
876
+ "learning_rate": 0.00017941416508447536,
877
+ "loss": 1.0664,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.23,
882
+ "learning_rate": 0.0001790980418314484,
883
+ "loss": 1.0781,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.24,
888
+ "learning_rate": 0.00017877979279382135,
889
+ "loss": 1.0703,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.24,
894
+ "learning_rate": 0.0001784594265246366,
895
+ "loss": 1.043,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.24,
900
+ "learning_rate": 0.0001781369516338378,
901
+ "loss": 1.0352,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.24,
906
+ "learning_rate": 0.00017781237678803847,
907
+ "loss": 1.0547,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.24,
912
+ "learning_rate": 0.000177485710710289,
913
+ "loss": 1.0703,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.24,
918
+ "learning_rate": 0.00017715696217984235,
919
+ "loss": 1.0391,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.24,
924
+ "learning_rate": 0.00017682614003191807,
925
+ "loss": 1.0547,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.25,
930
+ "learning_rate": 0.00017649325315746478,
931
+ "loss": 1.0664,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.25,
936
+ "learning_rate": 0.0001761583105029213,
937
+ "loss": 1.0586,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.25,
942
+ "learning_rate": 0.00017582132106997616,
943
+ "loss": 1.0547,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.25,
948
+ "learning_rate": 0.00017548229391532572,
949
+ "loss": 1.0391,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.25,
954
+ "learning_rate": 0.00017514123815043074,
955
+ "loss": 1.043,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.25,
960
+ "learning_rate": 0.00017479816294127152,
961
+ "loss": 1.0312,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.26,
966
+ "learning_rate": 0.0001744530775081015,
967
+ "loss": 1.0508,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.26,
972
+ "learning_rate": 0.0001741059911251997,
973
+ "loss": 1.0508,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.26,
978
+ "learning_rate": 0.000173756913120621,
979
+ "loss": 1.0352,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.26,
984
+ "learning_rate": 0.00017340585287594604,
985
+ "loss": 1.0664,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.26,
990
+ "learning_rate": 0.0001730528198260285,
991
+ "loss": 1.0469,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.26,
996
+ "learning_rate": 0.00017269782345874203,
997
+ "loss": 1.0273,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.27,
1002
+ "learning_rate": 0.00017234087331472497,
1003
+ "loss": 1.0195,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.27,
1008
+ "learning_rate": 0.00017198197898712404,
1009
+ "loss": 1.0586,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.27,
1014
+ "learning_rate": 0.00017162115012133643,
1015
+ "loss": 1.0703,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.27,
1020
+ "learning_rate": 0.00017125839641475072,
1021
+ "loss": 1.0391,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.27,
1026
+ "learning_rate": 0.00017089372761648616,
1027
+ "loss": 1.0391,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.27,
1032
+ "learning_rate": 0.00017052715352713075,
1033
+ "loss": 1.0195,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.28,
1038
+ "learning_rate": 0.00017015868399847768,
1039
+ "loss": 1.0547,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.28,
1044
+ "learning_rate": 0.00016978832893326074,
1045
+ "loss": 1.0664,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.28,
1050
+ "learning_rate": 0.00016941609828488807,
1051
+ "loss": 1.043,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.28,
1056
+ "learning_rate": 0.0001690420020571747,
1057
+ "loss": 1.043,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.28,
1062
+ "learning_rate": 0.0001686660503040737,
1063
+ "loss": 1.0547,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.28,
1068
+ "learning_rate": 0.00016828825312940592,
1069
+ "loss": 1.0625,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.28,
1074
+ "learning_rate": 0.0001679086206865886,
1075
+ "loss": 1.0586,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.29,
1080
+ "learning_rate": 0.00016752716317836229,
1081
+ "loss": 1.0586,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.29,
1086
+ "learning_rate": 0.0001671438908565167,
1087
+ "loss": 1.0391,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.29,
1092
+ "learning_rate": 0.00016675881402161536,
1093
+ "loss": 1.0508,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.29,
1098
+ "learning_rate": 0.0001663719430227186,
1099
+ "loss": 1.0352,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.29,
1104
+ "learning_rate": 0.00016598328825710533,
1105
+ "loss": 1.0234,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.29,
1110
+ "learning_rate": 0.000165592860169994,
1111
+ "loss": 1.0781,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.3,
1116
+ "learning_rate": 0.00016520066925426144,
1117
+ "loss": 1.0547,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.3,
1122
+ "learning_rate": 0.0001648067260501611,
1123
+ "loss": 1.0273,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.3,
1128
+ "learning_rate": 0.0001644110411450398,
1129
+ "loss": 1.0273,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.3,
1134
+ "learning_rate": 0.00016401362517305296,
1135
+ "loss": 1.043,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.3,
1140
+ "learning_rate": 0.00016361448881487914,
1141
+ "loss": 1.0312,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.3,
1146
+ "learning_rate": 0.00016321364279743266,
1147
+ "loss": 1.0137,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.31,
1152
+ "learning_rate": 0.0001628110978935756,
1153
+ "loss": 1.0664,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.31,
1158
+ "learning_rate": 0.00016240686492182804,
1159
+ "loss": 1.0469,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.31,
1164
+ "learning_rate": 0.00016200095474607753,
1165
+ "loss": 1.0469,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.31,
1170
+ "learning_rate": 0.00016159337827528685,
1171
+ "loss": 1.0371,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.31,
1176
+ "learning_rate": 0.0001611841464632011,
1177
+ "loss": 1.043,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.31,
1182
+ "learning_rate": 0.0001607732703080532,
1183
+ "loss": 1.0156,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.32,
1188
+ "learning_rate": 0.00016036076085226814,
1189
+ "loss": 1.0508,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.32,
1194
+ "learning_rate": 0.0001599466291821666,
1195
+ "loss": 1.0391,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.32,
1200
+ "learning_rate": 0.0001595308864276666,
1201
+ "loss": 1.0312,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.32,
1206
+ "learning_rate": 0.0001591135437619847,
1207
+ "loss": 1.0352,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.32,
1212
+ "learning_rate": 0.0001586946124013354,
1213
+ "loss": 1.0664,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.32,
1218
+ "learning_rate": 0.0001582741036046301,
1219
+ "loss": 1.0352,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.32,
1224
+ "learning_rate": 0.00015785202867317407,
1225
+ "loss": 1.0586,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.33,
1230
+ "learning_rate": 0.00015742839895036305,
1231
+ "loss": 1.0352,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.33,
1236
+ "learning_rate": 0.00015700322582137827,
1237
+ "loss": 1.0195,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.33,
1242
+ "learning_rate": 0.0001565765207128805,
1243
+ "loss": 1.0508,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.33,
1248
+ "learning_rate": 0.0001561482950927029,
1249
+ "loss": 1.0508,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.33,
1254
+ "learning_rate": 0.00015571856046954285,
1255
+ "loss": 1.0469,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.33,
1260
+ "learning_rate": 0.00015528732839265272,
1261
+ "loss": 1.0547,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.34,
1266
+ "learning_rate": 0.0001548546104515294,
1267
+ "loss": 1.0352,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.34,
1272
+ "learning_rate": 0.00015442041827560274,
1273
+ "loss": 1.0312,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.34,
1278
+ "learning_rate": 0.00015398476353392323,
1279
+ "loss": 1.0469,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.34,
1284
+ "learning_rate": 0.00015354765793484834,
1285
+ "loss": 1.0547,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.34,
1290
+ "learning_rate": 0.00015310911322572753,
1291
+ "loss": 1.0508,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.34,
1296
+ "learning_rate": 0.000152669141192587,
1297
+ "loss": 1.0703,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.35,
1302
+ "learning_rate": 0.00015222775365981273,
1303
+ "loss": 1.0273,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.35,
1308
+ "learning_rate": 0.00015178496248983254,
1309
+ "loss": 1.0508,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.35,
1314
+ "learning_rate": 0.00015134077958279765,
1315
+ "loss": 1.043,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.35,
1320
+ "learning_rate": 0.00015089521687626243,
1321
+ "loss": 1.0625,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.35,
1326
+ "learning_rate": 0.000150448286344864,
1327
+ "loss": 1.0664,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.35,
1332
+ "learning_rate": 0.00015000000000000001,
1333
+ "loss": 1.0508,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.36,
1338
+ "learning_rate": 0.00014955036988950618,
1339
+ "loss": 1.0547,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.36,
1344
+ "learning_rate": 0.00014909940809733222,
1345
+ "loss": 1.0586,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.36,
1350
+ "learning_rate": 0.00014864712674321734,
1351
+ "loss": 1.0234,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.36,
1356
+ "learning_rate": 0.00014819353798236427,
1357
+ "loss": 1.0312,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.36,
1362
+ "learning_rate": 0.00014773865400511272,
1363
+ "loss": 1.0391,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.36,
1368
+ "learning_rate": 0.00014728248703661182,
1369
+ "loss": 1.0273,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.36,
1374
+ "learning_rate": 0.00014682504933649144,
1375
+ "loss": 1.0703,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.37,
1380
+ "learning_rate": 0.00014636635319853275,
1381
+ "loss": 1.043,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.37,
1386
+ "learning_rate": 0.00014590641095033787,
1387
+ "loss": 1.0234,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.37,
1392
+ "learning_rate": 0.00014544523495299842,
1393
+ "loss": 1.0234,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.37,
1398
+ "learning_rate": 0.0001449828376007636,
1399
+ "loss": 1.0273,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.37,
1404
+ "learning_rate": 0.0001445192313207067,
1405
+ "loss": 1.0234,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 0.37,
1410
+ "learning_rate": 0.0001440544285723915,
1411
+ "loss": 1.043,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 0.38,
1416
+ "learning_rate": 0.00014358844184753712,
1417
+ "loss": 1.0352,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 0.38,
1422
+ "learning_rate": 0.00014312128366968243,
1423
+ "loss": 1.0625,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 0.38,
1428
+ "learning_rate": 0.00014265296659384956,
1429
+ "loss": 1.043,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 0.38,
1434
+ "learning_rate": 0.00014218350320620624,
1435
+ "loss": 1.0664,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 0.38,
1440
+ "learning_rate": 0.0001417129061237278,
1441
+ "loss": 1.0195,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 0.38,
1446
+ "learning_rate": 0.00014124118799385796,
1447
+ "loss": 1.0508,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 0.39,
1452
+ "learning_rate": 0.00014076836149416887,
1453
+ "loss": 1.043,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 0.39,
1458
+ "learning_rate": 0.0001402944393320206,
1459
+ "loss": 1.0234,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 0.39,
1464
+ "learning_rate": 0.00013981943424421932,
1465
+ "loss": 1.0391,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 0.39,
1470
+ "learning_rate": 0.00013934335899667527,
1471
+ "loss": 1.0352,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 0.39,
1476
+ "learning_rate": 0.00013886622638405952,
1477
+ "loss": 1.0352,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 0.39,
1482
+ "learning_rate": 0.00013838804922946027,
1483
+ "loss": 1.043,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 0.4,
1488
+ "learning_rate": 0.00013790884038403795,
1489
+ "loss": 1.0195,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 0.4,
1494
+ "learning_rate": 0.00013742861272668012,
1495
+ "loss": 1.0391,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 0.4,
1500
+ "learning_rate": 0.00013694737916365517,
1501
+ "loss": 1.0391,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 0.4,
1506
+ "learning_rate": 0.00013646515262826552,
1507
+ "loss": 1.0547,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 0.4,
1512
+ "learning_rate": 0.0001359819460805001,
1513
+ "loss": 1.0352,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 0.4,
1518
+ "learning_rate": 0.0001354977725066859,
1519
+ "loss": 1.0703,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 0.4,
1524
+ "learning_rate": 0.00013501264491913906,
1525
+ "loss": 1.0469,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 0.41,
1530
+ "learning_rate": 0.0001345265763558152,
1531
+ "loss": 1.0391,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 0.41,
1536
+ "learning_rate": 0.00013403957987995882,
1537
+ "loss": 1.0469,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 0.41,
1542
+ "learning_rate": 0.0001335516685797525,
1543
+ "loss": 1.0312,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 0.41,
1548
+ "learning_rate": 0.00013306285556796495,
1549
+ "loss": 1.0391,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 0.41,
1554
+ "learning_rate": 0.00013257315398159864,
1555
+ "loss": 1.002,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 0.41,
1560
+ "learning_rate": 0.00013208257698153677,
1561
+ "loss": 1.0508,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 0.42,
1566
+ "learning_rate": 0.00013159113775218964,
1567
+ "loss": 1.0195,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 0.42,
1572
+ "learning_rate": 0.00013109884950114007,
1573
+ "loss": 0.9902,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 0.42,
1578
+ "learning_rate": 0.00013060572545878875,
1579
+ "loss": 1.0352,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 0.42,
1584
+ "learning_rate": 0.00013011177887799845,
1585
+ "loss": 1.0078,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 0.42,
1590
+ "learning_rate": 0.00012961702303373795,
1591
+ "loss": 1.0273,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 0.42,
1596
+ "learning_rate": 0.00012912147122272523,
1597
+ "loss": 1.0312,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 0.43,
1602
+ "learning_rate": 0.00012862513676307008,
1603
+ "loss": 1.0391,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 0.43,
1608
+ "learning_rate": 0.00012812803299391628,
1609
+ "loss": 1.0,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 0.43,
1614
+ "learning_rate": 0.00012763017327508305,
1615
+ "loss": 1.0469,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 0.43,
1620
+ "learning_rate": 0.0001271315709867059,
1621
+ "loss": 1.043,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 0.43,
1626
+ "learning_rate": 0.00012663223952887723,
1627
+ "loss": 1.0586,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 0.43,
1632
+ "learning_rate": 0.00012613219232128608,
1633
+ "loss": 1.0234,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 0.44,
1638
+ "learning_rate": 0.00012563144280285741,
1639
+ "loss": 1.0176,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 0.44,
1644
+ "learning_rate": 0.00012513000443139112,
1645
+ "loss": 1.0156,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 0.44,
1650
+ "learning_rate": 0.00012462789068320017,
1651
+ "loss": 1.0547,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 0.44,
1656
+ "learning_rate": 0.00012412511505274844,
1657
+ "loss": 1.0234,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 0.44,
1662
+ "learning_rate": 0.00012362169105228826,
1663
+ "loss": 1.0117,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 0.44,
1668
+ "learning_rate": 0.000123117632211497,
1669
+ "loss": 1.0352,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 0.44,
1674
+ "learning_rate": 0.00012261295207711346,
1675
+ "loss": 1.0117,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 0.45,
1680
+ "learning_rate": 0.0001221076642125742,
1681
+ "loss": 1.0059,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 0.45,
1686
+ "learning_rate": 0.00012160178219764837,
1687
+ "loss": 1.0664,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 0.45,
1692
+ "learning_rate": 0.00012109531962807332,
1693
+ "loss": 1.0312,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 0.45,
1698
+ "learning_rate": 0.00012058829011518896,
1699
+ "loss": 1.0273,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 0.45,
1704
+ "learning_rate": 0.00012008070728557186,
1705
+ "loss": 1.0078,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 0.45,
1710
+ "learning_rate": 0.00011957258478066931,
1711
+ "loss": 1.0273,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 0.46,
1716
+ "learning_rate": 0.00011906393625643244,
1717
+ "loss": 1.0195,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 0.46,
1722
+ "learning_rate": 0.00011855477538294935,
1723
+ "loss": 1.043,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 0.46,
1728
+ "learning_rate": 0.00011804511584407763,
1729
+ "loss": 1.0117,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 0.46,
1734
+ "learning_rate": 0.00011753497133707679,
1735
+ "loss": 1.0078,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 0.46,
1740
+ "learning_rate": 0.00011702435557223987,
1741
+ "loss": 1.0703,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 0.46,
1746
+ "learning_rate": 0.00011651328227252517,
1747
+ "loss": 1.0352,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 0.47,
1752
+ "learning_rate": 0.00011600176517318741,
1753
+ "loss": 1.0469,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 0.47,
1758
+ "learning_rate": 0.00011548981802140848,
1759
+ "loss": 1.0664,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 0.47,
1764
+ "learning_rate": 0.00011497745457592816,
1765
+ "loss": 1.0312,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 0.47,
1770
+ "learning_rate": 0.00011446468860667421,
1771
+ "loss": 1.0078,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 0.47,
1776
+ "learning_rate": 0.00011395153389439233,
1777
+ "loss": 1.0312,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 0.47,
1782
+ "learning_rate": 0.00011343800423027582,
1783
+ "loss": 1.0703,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 0.48,
1788
+ "learning_rate": 0.0001129241134155949,
1789
+ "loss": 1.0625,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 0.48,
1794
+ "learning_rate": 0.00011240987526132594,
1795
+ "loss": 1.0117,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 0.48,
1800
+ "learning_rate": 0.00011189530358778005,
1801
+ "loss": 1.0234,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 0.48,
1806
+ "learning_rate": 0.00011138041222423177,
1807
+ "loss": 1.0078,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 0.48,
1812
+ "learning_rate": 0.00011086521500854745,
1813
+ "loss": 1.0117,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 0.48,
1818
+ "learning_rate": 0.00011034972578681338,
1819
+ "loss": 0.9941,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 0.48,
1824
+ "learning_rate": 0.00010983395841296348,
1825
+ "loss": 1.0273,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 0.49,
1830
+ "learning_rate": 0.00010931792674840718,
1831
+ "loss": 1.0312,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 0.49,
1836
+ "learning_rate": 0.00010880164466165674,
1837
+ "loss": 1.0195,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 0.49,
1842
+ "learning_rate": 0.00010828512602795462,
1843
+ "loss": 1.0195,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 0.49,
1848
+ "learning_rate": 0.00010776838472890065,
1849
+ "loss": 1.0195,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 0.49,
1854
+ "learning_rate": 0.00010725143465207867,
1855
+ "loss": 1.0156,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 0.49,
1860
+ "learning_rate": 0.00010673428969068364,
1861
+ "loss": 0.9941,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 0.5,
1866
+ "learning_rate": 0.00010621696374314807,
1867
+ "loss": 1.0156,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 0.5,
1872
+ "learning_rate": 0.00010569947071276847,
1873
+ "loss": 1.0312,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 0.5,
1878
+ "learning_rate": 0.00010518182450733186,
1879
+ "loss": 1.0176,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 0.5,
1884
+ "learning_rate": 0.00010466403903874176,
1885
+ "loss": 1.0391,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 0.5,
1890
+ "learning_rate": 0.00010414612822264455,
1891
+ "loss": 1.0352,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 0.5,
1896
+ "learning_rate": 0.00010362810597805526,
1897
+ "loss": 1.0391,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 0.51,
1902
+ "learning_rate": 0.0001031099862269837,
1903
+ "loss": 1.0234,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 0.51,
1908
+ "learning_rate": 0.00010259178289406011,
1909
+ "loss": 1.0332,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 0.51,
1914
+ "learning_rate": 0.00010207350990616107,
1915
+ "loss": 1.0,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 0.51,
1920
+ "learning_rate": 0.0001015551811920351,
1921
+ "loss": 1.0312,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 0.51,
1926
+ "learning_rate": 0.00010103681068192845,
1927
+ "loss": 1.0273,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 0.51,
1932
+ "learning_rate": 0.00010051841230721065,
1933
+ "loss": 1.043,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 0.52,
1938
+ "learning_rate": 0.0001,
1939
+ "loss": 1.0352,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 0.52,
1944
+ "learning_rate": 9.948158769278939e-05,
1945
+ "loss": 1.0508,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 0.52,
1950
+ "learning_rate": 9.896318931807155e-05,
1951
+ "loss": 1.043,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 0.52,
1956
+ "learning_rate": 9.844481880796491e-05,
1957
+ "loss": 1.043,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 0.52,
1962
+ "learning_rate": 9.792649009383899e-05,
1963
+ "loss": 0.9844,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 0.52,
1968
+ "learning_rate": 9.740821710593989e-05,
1969
+ "loss": 1.0195,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 0.52,
1974
+ "learning_rate": 9.689001377301633e-05,
1975
+ "loss": 1.0352,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 0.53,
1980
+ "learning_rate": 9.637189402194476e-05,
1981
+ "loss": 0.9922,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 0.53,
1986
+ "learning_rate": 9.585387177735547e-05,
1987
+ "loss": 1.0078,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 0.53,
1992
+ "learning_rate": 9.533596096125825e-05,
1993
+ "loss": 1.0195,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 0.53,
1998
+ "learning_rate": 9.481817549266817e-05,
1999
+ "loss": 1.0273,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 0.53,
2004
+ "learning_rate": 9.430052928723153e-05,
2005
+ "loss": 1.0273,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 0.53,
2010
+ "learning_rate": 9.378303625685195e-05,
2011
+ "loss": 1.0234,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 0.54,
2016
+ "learning_rate": 9.326571030931637e-05,
2017
+ "loss": 1.0312,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 0.54,
2022
+ "learning_rate": 9.274856534792138e-05,
2023
+ "loss": 1.0078,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 0.54,
2028
+ "learning_rate": 9.223161527109937e-05,
2029
+ "loss": 1.0117,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 0.54,
2034
+ "learning_rate": 9.171487397204539e-05,
2035
+ "loss": 1.0156,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 0.54,
2040
+ "learning_rate": 9.119835533834331e-05,
2041
+ "loss": 1.0273,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 0.54,
2046
+ "learning_rate": 9.068207325159284e-05,
2047
+ "loss": 1.0059,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 0.55,
2052
+ "learning_rate": 9.016604158703654e-05,
2053
+ "loss": 1.0234,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 0.55,
2058
+ "learning_rate": 8.965027421318665e-05,
2059
+ "loss": 1.0352,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 0.55,
2064
+ "learning_rate": 8.913478499145254e-05,
2065
+ "loss": 1.043,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 0.55,
2070
+ "learning_rate": 8.861958777576827e-05,
2071
+ "loss": 0.9961,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 0.55,
2076
+ "learning_rate": 8.810469641222001e-05,
2077
+ "loss": 1.0195,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 0.55,
2082
+ "learning_rate": 8.759012473867407e-05,
2083
+ "loss": 0.9922,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 0.56,
2088
+ "learning_rate": 8.707588658440511e-05,
2089
+ "loss": 1.0547,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 0.56,
2094
+ "learning_rate": 8.656199576972423e-05,
2095
+ "loss": 1.0078,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 0.56,
2100
+ "learning_rate": 8.604846610560771e-05,
2101
+ "loss": 1.0352,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 0.56,
2106
+ "learning_rate": 8.553531139332582e-05,
2107
+ "loss": 1.0195,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 0.56,
2112
+ "learning_rate": 8.502254542407186e-05,
2113
+ "loss": 1.0195,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 0.56,
2118
+ "learning_rate": 8.451018197859153e-05,
2119
+ "loss": 1.0234,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 0.56,
2124
+ "learning_rate": 8.399823482681262e-05,
2125
+ "loss": 1.002,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 0.57,
2130
+ "learning_rate": 8.348671772747487e-05,
2131
+ "loss": 1.0273,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 0.57,
2136
+ "learning_rate": 8.297564442776014e-05,
2137
+ "loss": 1.0039,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 0.57,
2142
+ "learning_rate": 8.246502866292324e-05,
2143
+ "loss": 1.0352,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 0.57,
2148
+ "learning_rate": 8.195488415592238e-05,
2149
+ "loss": 1.0215,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 0.57,
2154
+ "learning_rate": 8.144522461705067e-05,
2155
+ "loss": 1.0215,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 0.57,
2160
+ "learning_rate": 8.093606374356759e-05,
2161
+ "loss": 1.0117,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 0.58,
2166
+ "learning_rate": 8.042741521933071e-05,
2167
+ "loss": 1.0156,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 0.58,
2172
+ "learning_rate": 7.991929271442817e-05,
2173
+ "loss": 1.0098,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 0.58,
2178
+ "learning_rate": 7.941170988481108e-05,
2179
+ "loss": 0.9922,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 0.58,
2184
+ "learning_rate": 7.89046803719267e-05,
2185
+ "loss": 1.0234,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 0.58,
2190
+ "learning_rate": 7.839821780235168e-05,
2191
+ "loss": 1.043,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 0.58,
2196
+ "learning_rate": 7.789233578742582e-05,
2197
+ "loss": 1.0391,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 0.59,
2202
+ "learning_rate": 7.738704792288655e-05,
2203
+ "loss": 1.0039,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 0.59,
2208
+ "learning_rate": 7.688236778850306e-05,
2209
+ "loss": 1.0137,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 0.59,
2214
+ "learning_rate": 7.637830894771175e-05,
2215
+ "loss": 0.9941,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 0.59,
2220
+ "learning_rate": 7.587488494725157e-05,
2221
+ "loss": 0.9961,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 0.59,
2226
+ "learning_rate": 7.537210931679987e-05,
2227
+ "loss": 0.998,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 0.59,
2232
+ "learning_rate": 7.48699955686089e-05,
2233
+ "loss": 1.0195,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 0.6,
2238
+ "learning_rate": 7.43685571971426e-05,
2239
+ "loss": 0.9902,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 0.6,
2244
+ "learning_rate": 7.386780767871397e-05,
2245
+ "loss": 1.0176,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 0.6,
2250
+ "learning_rate": 7.336776047112276e-05,
2251
+ "loss": 1.0137,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 0.6,
2256
+ "learning_rate": 7.286842901329412e-05,
2257
+ "loss": 1.0078,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 0.6,
2262
+ "learning_rate": 7.236982672491698e-05,
2263
+ "loss": 1.0391,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 0.6,
2268
+ "learning_rate": 7.187196700608373e-05,
2269
+ "loss": 1.0195,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 0.6,
2274
+ "learning_rate": 7.137486323692995e-05,
2275
+ "loss": 1.0312,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 0.61,
2280
+ "learning_rate": 7.087852877727481e-05,
2281
+ "loss": 0.9902,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 0.61,
2286
+ "learning_rate": 7.038297696626206e-05,
2287
+ "loss": 1.0234,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 0.61,
2292
+ "learning_rate": 6.988822112200156e-05,
2293
+ "loss": 1.0234,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 0.61,
2298
+ "learning_rate": 6.939427454121128e-05,
2299
+ "loss": 0.9961,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 0.61,
2304
+ "learning_rate": 6.890115049885994e-05,
2305
+ "loss": 0.9863,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 0.61,
2310
+ "learning_rate": 6.84088622478104e-05,
2311
+ "loss": 1.0078,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 0.62,
2316
+ "learning_rate": 6.791742301846326e-05,
2317
+ "loss": 1.0156,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 0.62,
2322
+ "learning_rate": 6.742684601840141e-05,
2323
+ "loss": 0.9941,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 0.62,
2328
+ "learning_rate": 6.693714443203507e-05,
2329
+ "loss": 1.043,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 0.62,
2334
+ "learning_rate": 6.644833142024751e-05,
2335
+ "loss": 1.0156,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 0.62,
2340
+ "learning_rate": 6.59604201200412e-05,
2341
+ "loss": 1.0156,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 0.62,
2346
+ "learning_rate": 6.547342364418481e-05,
2347
+ "loss": 0.9844,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 0.63,
2352
+ "learning_rate": 6.498735508086093e-05,
2353
+ "loss": 1.0234,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 0.63,
2358
+ "learning_rate": 6.450222749331414e-05,
2359
+ "loss": 1.002,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 0.63,
2364
+ "learning_rate": 6.40180539194999e-05,
2365
+ "loss": 0.9785,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 0.63,
2370
+ "learning_rate": 6.35348473717345e-05,
2371
+ "loss": 1.0312,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 0.63,
2376
+ "learning_rate": 6.305262083634488e-05,
2377
+ "loss": 1.0176,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 0.63,
2382
+ "learning_rate": 6.25713872733199e-05,
2383
+ "loss": 1.0273,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 0.64,
2388
+ "learning_rate": 6.209115961596208e-05,
2389
+ "loss": 1.0273,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 0.64,
2394
+ "learning_rate": 6.161195077053976e-05,
2395
+ "loss": 1.0273,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 0.64,
2400
+ "learning_rate": 6.113377361594049e-05,
2401
+ "loss": 1.0195,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 0.64,
2406
+ "learning_rate": 6.065664100332478e-05,
2407
+ "loss": 1.0234,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 0.64,
2412
+ "learning_rate": 6.018056575578075e-05,
2413
+ "loss": 1.0156,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 0.64,
2418
+ "learning_rate": 5.970556066797941e-05,
2419
+ "loss": 1.0078,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 0.64,
2424
+ "learning_rate": 5.923163850583113e-05,
2425
+ "loss": 0.998,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 0.65,
2430
+ "learning_rate": 5.875881200614207e-05,
2431
+ "loss": 0.9922,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 0.65,
2436
+ "learning_rate": 5.828709387627218e-05,
2437
+ "loss": 1.0078,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 0.65,
2442
+ "learning_rate": 5.781649679379378e-05,
2443
+ "loss": 1.0273,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 0.65,
2448
+ "learning_rate": 5.73470334061505e-05,
2449
+ "loss": 1.002,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 0.65,
2454
+ "learning_rate": 5.687871633031754e-05,
2455
+ "loss": 0.998,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 0.65,
2460
+ "learning_rate": 5.6411558152462894e-05,
2461
+ "loss": 1.0117,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 0.66,
2466
+ "learning_rate": 5.5945571427608526e-05,
2467
+ "loss": 1.0352,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 0.66,
2472
+ "learning_rate": 5.54807686792933e-05,
2473
+ "loss": 1.0254,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 0.66,
2478
+ "learning_rate": 5.501716239923642e-05,
2479
+ "loss": 1.0293,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 0.66,
2484
+ "learning_rate": 5.4554765047001613e-05,
2485
+ "loss": 1.0195,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 0.66,
2490
+ "learning_rate": 5.4093589049662175e-05,
2491
+ "loss": 1.0,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 0.66,
2496
+ "learning_rate": 5.363364680146725e-05,
2497
+ "loss": 1.0234,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 0.67,
2502
+ "learning_rate": 5.31749506635086e-05,
2503
+ "loss": 1.0234,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 0.67,
2508
+ "learning_rate": 5.271751296338823e-05,
2509
+ "loss": 1.0234,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 0.67,
2514
+ "learning_rate": 5.226134599488728e-05,
2515
+ "loss": 1.0312,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 0.67,
2520
+ "learning_rate": 5.180646201763577e-05,
2521
+ "loss": 1.0234,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 0.67,
2526
+ "learning_rate": 5.135287325678271e-05,
2527
+ "loss": 1.0117,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 0.67,
2532
+ "learning_rate": 5.090059190266779e-05,
2533
+ "loss": 1.0078,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 0.68,
2538
+ "learning_rate": 5.0449630110493836e-05,
2539
+ "loss": 1.0312,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 0.68,
2544
+ "learning_rate": 5.000000000000002e-05,
2545
+ "loss": 1.0234,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 0.68,
2550
+ "learning_rate": 4.955171365513603e-05,
2551
+ "loss": 1.0039,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 0.68,
2556
+ "learning_rate": 4.9104783123737566e-05,
2557
+ "loss": 1.0098,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 0.68,
2562
+ "learning_rate": 4.865922041720239e-05,
2563
+ "loss": 1.0391,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 0.68,
2568
+ "learning_rate": 4.821503751016746e-05,
2569
+ "loss": 0.998,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 0.68,
2574
+ "learning_rate": 4.777224634018732e-05,
2575
+ "loss": 0.9961,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 0.69,
2580
+ "learning_rate": 4.733085880741301e-05,
2581
+ "loss": 1.0195,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 0.69,
2586
+ "learning_rate": 4.689088677427249e-05,
2587
+ "loss": 0.9941,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 0.69,
2592
+ "learning_rate": 4.645234206515171e-05,
2593
+ "loss": 0.9609,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 0.69,
2598
+ "learning_rate": 4.6015236466076747e-05,
2599
+ "loss": 0.9902,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 0.69,
2604
+ "learning_rate": 4.5579581724397255e-05,
2605
+ "loss": 1.0234,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 0.69,
2610
+ "learning_rate": 4.514538954847064e-05,
2611
+ "loss": 1.0352,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 0.7,
2616
+ "learning_rate": 4.471267160734731e-05,
2617
+ "loss": 1.0078,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 0.7,
2622
+ "learning_rate": 4.428143953045717e-05,
2623
+ "loss": 0.998,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 0.7,
2628
+ "learning_rate": 4.385170490729712e-05,
2629
+ "loss": 0.9961,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 0.7,
2634
+ "learning_rate": 4.342347928711953e-05,
2635
+ "loss": 0.9961,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 0.7,
2640
+ "learning_rate": 4.2996774178621736e-05,
2641
+ "loss": 0.998,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 0.7,
2646
+ "learning_rate": 4.257160104963696e-05,
2647
+ "loss": 1.0078,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 0.71,
2652
+ "learning_rate": 4.2147971326825966e-05,
2653
+ "loss": 1.0156,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 0.71,
2658
+ "learning_rate": 4.172589639536991e-05,
2659
+ "loss": 1.002,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 0.71,
2664
+ "learning_rate": 4.130538759866457e-05,
2665
+ "loss": 0.9824,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 0.71,
2670
+ "learning_rate": 4.088645623801534e-05,
2671
+ "loss": 1.0078,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 0.71,
2676
+ "learning_rate": 4.046911357233343e-05,
2677
+ "loss": 0.9902,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 0.71,
2682
+ "learning_rate": 4.00533708178334e-05,
2683
+ "loss": 1.0078,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 0.72,
2688
+ "learning_rate": 3.963923914773187e-05,
2689
+ "loss": 1.0117,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 0.72,
2694
+ "learning_rate": 3.922672969194686e-05,
2695
+ "loss": 1.0059,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 0.72,
2700
+ "learning_rate": 3.8815853536798904e-05,
2701
+ "loss": 0.9961,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 0.72,
2706
+ "learning_rate": 3.840662172471315e-05,
2707
+ "loss": 1.0195,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 0.72,
2712
+ "learning_rate": 3.79990452539225e-05,
2713
+ "loss": 1.002,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 0.72,
2718
+ "learning_rate": 3.759313507817196e-05,
2719
+ "loss": 1.0156,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 0.72,
2724
+ "learning_rate": 3.7188902106424416e-05,
2725
+ "loss": 1.0898,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 0.73,
2730
+ "learning_rate": 3.678635720256737e-05,
2731
+ "loss": 1.0117,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 0.73,
2736
+ "learning_rate": 3.638551118512089e-05,
2737
+ "loss": 0.9805,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 0.73,
2742
+ "learning_rate": 3.5986374826947066e-05,
2743
+ "loss": 1.0352,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 0.73,
2748
+ "learning_rate": 3.558895885496023e-05,
2749
+ "loss": 0.9961,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 0.73,
2754
+ "learning_rate": 3.519327394983888e-05,
2755
+ "loss": 1.0059,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 0.73,
2760
+ "learning_rate": 3.479933074573858e-05,
2761
+ "loss": 0.9961,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 0.74,
2766
+ "learning_rate": 3.440713983000601e-05,
2767
+ "loss": 1.0039,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 0.74,
2772
+ "learning_rate": 3.401671174289469e-05,
2773
+ "loss": 1.0078,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 0.74,
2778
+ "learning_rate": 3.362805697728145e-05,
2779
+ "loss": 1.0,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 0.74,
2784
+ "learning_rate": 3.324118597838464e-05,
2785
+ "loss": 1.0312,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 0.74,
2790
+ "learning_rate": 3.285610914348332e-05,
2791
+ "loss": 1.0078,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 0.74,
2796
+ "learning_rate": 3.2472836821637744e-05,
2797
+ "loss": 1.0,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 0.75,
2802
+ "learning_rate": 3.209137931341143e-05,
2803
+ "loss": 1.0195,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 0.75,
2808
+ "learning_rate": 3.1711746870594086e-05,
2809
+ "loss": 0.998,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 0.75,
2814
+ "learning_rate": 3.1333949695926324e-05,
2815
+ "loss": 1.0137,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 0.75,
2820
+ "learning_rate": 3.0957997942825336e-05,
2821
+ "loss": 0.9922,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 0.75,
2826
+ "learning_rate": 3.058390171511196e-05,
2827
+ "loss": 0.998,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 0.75,
2832
+ "learning_rate": 3.021167106673928e-05,
2833
+ "loss": 1.0137,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 0.76,
2838
+ "learning_rate": 2.9841316001522347e-05,
2839
+ "loss": 1.0312,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 0.76,
2844
+ "learning_rate": 2.9472846472869298e-05,
2845
+ "loss": 0.9922,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 0.76,
2850
+ "learning_rate": 2.9106272383513835e-05,
2851
+ "loss": 1.0273,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 0.76,
2856
+ "learning_rate": 2.874160358524931e-05,
2857
+ "loss": 0.9902,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 0.76,
2862
+ "learning_rate": 2.8378849878663628e-05,
2863
+ "loss": 0.9902,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 0.76,
2868
+ "learning_rate": 2.8018021012875994e-05,
2869
+ "loss": 1.0078,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 0.76,
2874
+ "learning_rate": 2.7659126685275027e-05,
2875
+ "loss": 1.0156,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 0.77,
2880
+ "learning_rate": 2.7302176541257986e-05,
2881
+ "loss": 1.0234,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 0.77,
2886
+ "learning_rate": 2.6947180173971508e-05,
2887
+ "loss": 1.0195,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 0.77,
2892
+ "learning_rate": 2.659414712405398e-05,
2893
+ "loss": 1.0234,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 0.77,
2898
+ "learning_rate": 2.6243086879379e-05,
2899
+ "loss": 0.998,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 0.77,
2904
+ "learning_rate": 2.5894008874800325e-05,
2905
+ "loss": 0.9668,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 0.77,
2910
+ "learning_rate": 2.5546922491898495e-05,
2911
+ "loss": 1.0,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 0.78,
2916
+ "learning_rate": 2.5201837058728505e-05,
2917
+ "loss": 1.0234,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 0.78,
2922
+ "learning_rate": 2.485876184956928e-05,
2923
+ "loss": 1.0059,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 0.78,
2928
+ "learning_rate": 2.451770608467432e-05,
2929
+ "loss": 1.0117,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 0.78,
2934
+ "learning_rate": 2.417867893002387e-05,
2935
+ "loss": 1.0117,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 0.78,
2940
+ "learning_rate": 2.3841689497078746e-05,
2941
+ "loss": 0.9902,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 0.78,
2946
+ "learning_rate": 2.3506746842535242e-05,
2947
+ "loss": 1.0039,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 0.79,
2952
+ "learning_rate": 2.3173859968081944e-05,
2953
+ "loss": 1.0,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 0.79,
2958
+ "learning_rate": 2.2843037820157675e-05,
2959
+ "loss": 1.0391,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 0.79,
2964
+ "learning_rate": 2.251428928971102e-05,
2965
+ "loss": 1.0039,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 0.79,
2970
+ "learning_rate": 2.2187623211961562e-05,
2971
+ "loss": 1.0312,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 0.79,
2976
+ "learning_rate": 2.1863048366162208e-05,
2977
+ "loss": 1.0,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 0.79,
2982
+ "learning_rate": 2.1540573475363402e-05,
2983
+ "loss": 1.002,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 0.8,
2988
+ "learning_rate": 2.1220207206178688e-05,
2989
+ "loss": 1.0312,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 0.8,
2994
+ "learning_rate": 2.0901958168551638e-05,
2995
+ "loss": 0.9785,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 0.8,
3000
+ "learning_rate": 2.058583491552465e-05,
3001
+ "loss": 0.9805,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 0.8,
3006
+ "learning_rate": 2.027184594300898e-05,
3007
+ "loss": 1.0137,
3008
+ "step": 500
3009
+ },
3010
+ {
3011
+ "epoch": 0.8,
3012
+ "learning_rate": 1.995999968955641e-05,
3013
+ "loss": 0.9883,
3014
+ "step": 501
3015
+ },
3016
+ {
3017
+ "epoch": 0.8,
3018
+ "learning_rate": 1.9650304536132426e-05,
3019
+ "loss": 0.9844,
3020
+ "step": 502
3021
+ },
3022
+ {
3023
+ "epoch": 0.8,
3024
+ "learning_rate": 1.9342768805891178e-05,
3025
+ "loss": 1.0195,
3026
+ "step": 503
3027
+ },
3028
+ {
3029
+ "epoch": 0.81,
3030
+ "learning_rate": 1.903740076395151e-05,
3031
+ "loss": 0.9746,
3032
+ "step": 504
3033
+ },
3034
+ {
3035
+ "epoch": 0.81,
3036
+ "learning_rate": 1.8734208617174988e-05,
3037
+ "loss": 0.9902,
3038
+ "step": 505
3039
+ },
3040
+ {
3041
+ "epoch": 0.81,
3042
+ "learning_rate": 1.8433200513945337e-05,
3043
+ "loss": 0.9922,
3044
+ "step": 506
3045
+ },
3046
+ {
3047
+ "epoch": 0.81,
3048
+ "learning_rate": 1.8134384543949478e-05,
3049
+ "loss": 1.0234,
3050
+ "step": 507
3051
+ },
3052
+ {
3053
+ "epoch": 0.81,
3054
+ "learning_rate": 1.783776873795994e-05,
3055
+ "loss": 1.0156,
3056
+ "step": 508
3057
+ },
3058
+ {
3059
+ "epoch": 0.81,
3060
+ "learning_rate": 1.754336106761927e-05,
3061
+ "loss": 1.002,
3062
+ "step": 509
3063
+ },
3064
+ {
3065
+ "epoch": 0.82,
3066
+ "learning_rate": 1.7251169445225657e-05,
3067
+ "loss": 0.9766,
3068
+ "step": 510
3069
+ },
3070
+ {
3071
+ "epoch": 0.82,
3072
+ "learning_rate": 1.696120172352025e-05,
3073
+ "loss": 0.9688,
3074
+ "step": 511
3075
+ },
3076
+ {
3077
+ "epoch": 0.82,
3078
+ "learning_rate": 1.6673465695476232e-05,
3079
+ "loss": 1.0234,
3080
+ "step": 512
3081
+ },
3082
+ {
3083
+ "epoch": 0.82,
3084
+ "learning_rate": 1.6387969094089316e-05,
3085
+ "loss": 1.0098,
3086
+ "step": 513
3087
+ },
3088
+ {
3089
+ "epoch": 0.82,
3090
+ "learning_rate": 1.6104719592169902e-05,
3091
+ "loss": 1.0078,
3092
+ "step": 514
3093
+ },
3094
+ {
3095
+ "epoch": 0.82,
3096
+ "learning_rate": 1.5823724802136865e-05,
3097
+ "loss": 1.0352,
3098
+ "step": 515
3099
+ },
3100
+ {
3101
+ "epoch": 0.83,
3102
+ "learning_rate": 1.5544992275813053e-05,
3103
+ "loss": 1.0,
3104
+ "step": 516
3105
+ },
3106
+ {
3107
+ "epoch": 0.83,
3108
+ "learning_rate": 1.526852950422226e-05,
3109
+ "loss": 1.0098,
3110
+ "step": 517
3111
+ },
3112
+ {
3113
+ "epoch": 0.83,
3114
+ "learning_rate": 1.4994343917387854e-05,
3115
+ "loss": 0.9863,
3116
+ "step": 518
3117
+ },
3118
+ {
3119
+ "epoch": 0.83,
3120
+ "learning_rate": 1.4722442884133214e-05,
3121
+ "loss": 1.0117,
3122
+ "step": 519
3123
+ },
3124
+ {
3125
+ "epoch": 0.83,
3126
+ "learning_rate": 1.4452833711883628e-05,
3127
+ "loss": 1.0117,
3128
+ "step": 520
3129
+ },
3130
+ {
3131
+ "epoch": 0.83,
3132
+ "learning_rate": 1.4185523646469822e-05,
3133
+ "loss": 1.0078,
3134
+ "step": 521
3135
+ },
3136
+ {
3137
+ "epoch": 0.84,
3138
+ "learning_rate": 1.3920519871933424e-05,
3139
+ "loss": 1.0215,
3140
+ "step": 522
3141
+ },
3142
+ {
3143
+ "epoch": 0.84,
3144
+ "learning_rate": 1.3657829510333654e-05,
3145
+ "loss": 0.9727,
3146
+ "step": 523
3147
+ },
3148
+ {
3149
+ "epoch": 0.84,
3150
+ "learning_rate": 1.339745962155613e-05,
3151
+ "loss": 1.0312,
3152
+ "step": 524
3153
+ },
3154
+ {
3155
+ "epoch": 0.84,
3156
+ "learning_rate": 1.3139417203123027e-05,
3157
+ "loss": 0.9922,
3158
+ "step": 525
3159
+ },
3160
+ {
3161
+ "epoch": 0.84,
3162
+ "learning_rate": 1.2883709190004955e-05,
3163
+ "loss": 1.0039,
3164
+ "step": 526
3165
+ },
3166
+ {
3167
+ "epoch": 0.84,
3168
+ "learning_rate": 1.263034245443473e-05,
3169
+ "loss": 0.9863,
3170
+ "step": 527
3171
+ },
3172
+ {
3173
+ "epoch": 0.84,
3174
+ "learning_rate": 1.2379323805722576e-05,
3175
+ "loss": 0.9961,
3176
+ "step": 528
3177
+ },
3178
+ {
3179
+ "epoch": 0.85,
3180
+ "learning_rate": 1.2130659990073146e-05,
3181
+ "loss": 0.9941,
3182
+ "step": 529
3183
+ },
3184
+ {
3185
+ "epoch": 0.85,
3186
+ "learning_rate": 1.1884357690404158e-05,
3187
+ "loss": 0.9922,
3188
+ "step": 530
3189
+ },
3190
+ {
3191
+ "epoch": 0.85,
3192
+ "learning_rate": 1.1640423526166988e-05,
3193
+ "loss": 0.9941,
3194
+ "step": 531
3195
+ },
3196
+ {
3197
+ "epoch": 0.85,
3198
+ "learning_rate": 1.1398864053168534e-05,
3199
+ "loss": 1.0273,
3200
+ "step": 532
3201
+ },
3202
+ {
3203
+ "epoch": 0.85,
3204
+ "learning_rate": 1.1159685763395111e-05,
3205
+ "loss": 1.002,
3206
+ "step": 533
3207
+ },
3208
+ {
3209
+ "epoch": 0.85,
3210
+ "learning_rate": 1.0922895084838037e-05,
3211
+ "loss": 1.0137,
3212
+ "step": 534
3213
+ },
3214
+ {
3215
+ "epoch": 0.86,
3216
+ "learning_rate": 1.0688498381320855e-05,
3217
+ "loss": 0.9863,
3218
+ "step": 535
3219
+ },
3220
+ {
3221
+ "epoch": 0.86,
3222
+ "learning_rate": 1.045650195232819e-05,
3223
+ "loss": 0.9961,
3224
+ "step": 536
3225
+ },
3226
+ {
3227
+ "epoch": 0.86,
3228
+ "learning_rate": 1.0226912032836611e-05,
3229
+ "loss": 1.0234,
3230
+ "step": 537
3231
+ },
3232
+ {
3233
+ "epoch": 0.86,
3234
+ "learning_rate": 9.999734793146998e-06,
3235
+ "loss": 0.9941,
3236
+ "step": 538
3237
+ },
3238
+ {
3239
+ "epoch": 0.86,
3240
+ "learning_rate": 9.774976338718677e-06,
3241
+ "loss": 1.0137,
3242
+ "step": 539
3243
+ },
3244
+ {
3245
+ "epoch": 0.86,
3246
+ "learning_rate": 9.552642710005299e-06,
3247
+ "loss": 0.9941,
3248
+ "step": 540
3249
+ },
3250
+ {
3251
+ "epoch": 0.87,
3252
+ "learning_rate": 9.332739882292752e-06,
3253
+ "loss": 0.9922,
3254
+ "step": 541
3255
+ },
3256
+ {
3257
+ "epoch": 0.87,
3258
+ "learning_rate": 9.115273765538202e-06,
3259
+ "loss": 1.0508,
3260
+ "step": 542
3261
+ },
3262
+ {
3263
+ "epoch": 0.87,
3264
+ "learning_rate": 8.900250204211514e-06,
3265
+ "loss": 0.9727,
3266
+ "step": 543
3267
+ },
3268
+ {
3269
+ "epoch": 0.87,
3270
+ "learning_rate": 8.687674977138116e-06,
3271
+ "loss": 1.0117,
3272
+ "step": 544
3273
+ },
3274
+ {
3275
+ "epoch": 0.87,
3276
+ "learning_rate": 8.47755379734373e-06,
3277
+ "loss": 0.9961,
3278
+ "step": 545
3279
+ },
3280
+ {
3281
+ "epoch": 0.87,
3282
+ "learning_rate": 8.269892311900696e-06,
3283
+ "loss": 1.0352,
3284
+ "step": 546
3285
+ },
3286
+ {
3287
+ "epoch": 0.88,
3288
+ "learning_rate": 8.064696101776358e-06,
3289
+ "loss": 1.0,
3290
+ "step": 547
3291
+ },
3292
+ {
3293
+ "epoch": 0.88,
3294
+ "learning_rate": 7.861970681683051e-06,
3295
+ "loss": 0.9668,
3296
+ "step": 548
3297
+ },
3298
+ {
3299
+ "epoch": 0.88,
3300
+ "learning_rate": 7.661721499929753e-06,
3301
+ "loss": 0.9961,
3302
+ "step": 549
3303
+ },
3304
+ {
3305
+ "epoch": 0.88,
3306
+ "learning_rate": 7.463953938275858e-06,
3307
+ "loss": 0.9844,
3308
+ "step": 550
3309
+ },
3310
+ {
3311
+ "epoch": 0.88,
3312
+ "learning_rate": 7.2686733117863784e-06,
3313
+ "loss": 1.0039,
3314
+ "step": 551
3315
+ },
3316
+ {
3317
+ "epoch": 0.88,
3318
+ "learning_rate": 7.07588486868922e-06,
3319
+ "loss": 0.9863,
3320
+ "step": 552
3321
+ },
3322
+ {
3323
+ "epoch": 0.88,
3324
+ "learning_rate": 6.8855937902340576e-06,
3325
+ "loss": 0.9883,
3326
+ "step": 553
3327
+ },
3328
+ {
3329
+ "epoch": 0.89,
3330
+ "learning_rate": 6.6978051905530855e-06,
3331
+ "loss": 1.0234,
3332
+ "step": 554
3333
+ },
3334
+ {
3335
+ "epoch": 0.89,
3336
+ "learning_rate": 6.512524116523633e-06,
3337
+ "loss": 0.9922,
3338
+ "step": 555
3339
+ },
3340
+ {
3341
+ "epoch": 0.89,
3342
+ "learning_rate": 6.329755547632499e-06,
3343
+ "loss": 0.9922,
3344
+ "step": 556
3345
+ },
3346
+ {
3347
+ "epoch": 0.89,
3348
+ "learning_rate": 6.149504395842087e-06,
3349
+ "loss": 1.0352,
3350
+ "step": 557
3351
+ },
3352
+ {
3353
+ "epoch": 0.89,
3354
+ "learning_rate": 5.971775505458444e-06,
3355
+ "loss": 1.0234,
3356
+ "step": 558
3357
+ },
3358
+ {
3359
+ "epoch": 0.89,
3360
+ "learning_rate": 5.7965736530010916e-06,
3361
+ "loss": 0.9805,
3362
+ "step": 559
3363
+ },
3364
+ {
3365
+ "epoch": 0.9,
3366
+ "learning_rate": 5.623903547074549e-06,
3367
+ "loss": 1.0117,
3368
+ "step": 560
3369
+ },
3370
+ {
3371
+ "epoch": 0.9,
3372
+ "learning_rate": 5.453769828241872e-06,
3373
+ "loss": 0.9863,
3374
+ "step": 561
3375
+ },
3376
+ {
3377
+ "epoch": 0.9,
3378
+ "learning_rate": 5.286177068899989e-06,
3379
+ "loss": 1.0059,
3380
+ "step": 562
3381
+ },
3382
+ {
3383
+ "epoch": 0.9,
3384
+ "learning_rate": 5.121129773156663e-06,
3385
+ "loss": 1.0078,
3386
+ "step": 563
3387
+ },
3388
+ {
3389
+ "epoch": 0.9,
3390
+ "learning_rate": 4.95863237670956e-06,
3391
+ "loss": 0.9922,
3392
+ "step": 564
3393
+ },
3394
+ {
3395
+ "epoch": 0.9,
3396
+ "learning_rate": 4.798689246727006e-06,
3397
+ "loss": 1.0156,
3398
+ "step": 565
3399
+ },
3400
+ {
3401
+ "epoch": 0.91,
3402
+ "learning_rate": 4.641304681730641e-06,
3403
+ "loss": 1.0273,
3404
+ "step": 566
3405
+ },
3406
+ {
3407
+ "epoch": 0.91,
3408
+ "learning_rate": 4.486482911479839e-06,
3409
+ "loss": 0.9941,
3410
+ "step": 567
3411
+ },
3412
+ {
3413
+ "epoch": 0.91,
3414
+ "learning_rate": 4.3342280968580285e-06,
3415
+ "loss": 1.002,
3416
+ "step": 568
3417
+ },
3418
+ {
3419
+ "epoch": 0.91,
3420
+ "learning_rate": 4.184544329761009e-06,
3421
+ "loss": 1.0195,
3422
+ "step": 569
3423
+ },
3424
+ {
3425
+ "epoch": 0.91,
3426
+ "learning_rate": 4.037435632986786e-06,
3427
+ "loss": 0.9805,
3428
+ "step": 570
3429
+ },
3430
+ {
3431
+ "epoch": 0.91,
3432
+ "learning_rate": 3.892905960127546e-06,
3433
+ "loss": 1.0,
3434
+ "step": 571
3435
+ },
3436
+ {
3437
+ "epoch": 0.92,
3438
+ "learning_rate": 3.750959195463466e-06,
3439
+ "loss": 1.0156,
3440
+ "step": 572
3441
+ },
3442
+ {
3443
+ "epoch": 0.92,
3444
+ "learning_rate": 3.611599153858214e-06,
3445
+ "loss": 1.0234,
3446
+ "step": 573
3447
+ },
3448
+ {
3449
+ "epoch": 0.92,
3450
+ "learning_rate": 3.4748295806564356e-06,
3451
+ "loss": 1.0234,
3452
+ "step": 574
3453
+ },
3454
+ {
3455
+ "epoch": 0.92,
3456
+ "learning_rate": 3.3406541515832003e-06,
3457
+ "loss": 1.0312,
3458
+ "step": 575
3459
+ },
3460
+ {
3461
+ "epoch": 0.92,
3462
+ "learning_rate": 3.209076472645112e-06,
3463
+ "loss": 0.9766,
3464
+ "step": 576
3465
+ },
3466
+ {
3467
+ "epoch": 0.92,
3468
+ "learning_rate": 3.0801000800333877e-06,
3469
+ "loss": 1.0078,
3470
+ "step": 577
3471
+ },
3472
+ {
3473
+ "epoch": 0.92,
3474
+ "learning_rate": 2.9537284400289355e-06,
3475
+ "loss": 1.0312,
3476
+ "step": 578
3477
+ },
3478
+ {
3479
+ "epoch": 0.93,
3480
+ "learning_rate": 2.8299649489090475e-06,
3481
+ "loss": 1.0098,
3482
+ "step": 579
3483
+ },
3484
+ {
3485
+ "epoch": 0.93,
3486
+ "learning_rate": 2.708812932856253e-06,
3487
+ "loss": 0.9883,
3488
+ "step": 580
3489
+ },
3490
+ {
3491
+ "epoch": 0.93,
3492
+ "learning_rate": 2.590275647868867e-06,
3493
+ "loss": 1.0195,
3494
+ "step": 581
3495
+ },
3496
+ {
3497
+ "epoch": 0.93,
3498
+ "learning_rate": 2.4743562796734622e-06,
3499
+ "loss": 1.002,
3500
+ "step": 582
3501
+ },
3502
+ {
3503
+ "epoch": 0.93,
3504
+ "learning_rate": 2.3610579436393e-06,
3505
+ "loss": 0.9648,
3506
+ "step": 583
3507
+ },
3508
+ {
3509
+ "epoch": 0.93,
3510
+ "learning_rate": 2.250383684694579e-06,
3511
+ "loss": 1.0352,
3512
+ "step": 584
3513
+ },
3514
+ {
3515
+ "epoch": 0.94,
3516
+ "learning_rate": 2.1423364772445887e-06,
3517
+ "loss": 1.0,
3518
+ "step": 585
3519
+ },
3520
+ {
3521
+ "epoch": 0.94,
3522
+ "learning_rate": 2.036919225091827e-06,
3523
+ "loss": 1.0098,
3524
+ "step": 586
3525
+ },
3526
+ {
3527
+ "epoch": 0.94,
3528
+ "learning_rate": 1.9341347613579087e-06,
3529
+ "loss": 0.9961,
3530
+ "step": 587
3531
+ },
3532
+ {
3533
+ "epoch": 0.94,
3534
+ "learning_rate": 1.8339858484073935e-06,
3535
+ "loss": 0.9961,
3536
+ "step": 588
3537
+ },
3538
+ {
3539
+ "epoch": 0.94,
3540
+ "learning_rate": 1.7364751777736332e-06,
3541
+ "loss": 1.0039,
3542
+ "step": 589
3543
+ },
3544
+ {
3545
+ "epoch": 0.94,
3546
+ "learning_rate": 1.6416053700863964e-06,
3547
+ "loss": 1.0059,
3548
+ "step": 590
3549
+ },
3550
+ {
3551
+ "epoch": 0.95,
3552
+ "learning_rate": 1.5493789750014031e-06,
3553
+ "loss": 0.9863,
3554
+ "step": 591
3555
+ },
3556
+ {
3557
+ "epoch": 0.95,
3558
+ "learning_rate": 1.459798471131868e-06,
3559
+ "loss": 0.9453,
3560
+ "step": 592
3561
+ },
3562
+ {
3563
+ "epoch": 0.95,
3564
+ "learning_rate": 1.3728662659818204e-06,
3565
+ "loss": 0.9727,
3566
+ "step": 593
3567
+ },
3568
+ {
3569
+ "epoch": 0.95,
3570
+ "learning_rate": 1.2885846958814673e-06,
3571
+ "loss": 0.9805,
3572
+ "step": 594
3573
+ },
3574
+ {
3575
+ "epoch": 0.95,
3576
+ "learning_rate": 1.2069560259243328e-06,
3577
+ "loss": 1.0195,
3578
+ "step": 595
3579
+ },
3580
+ {
3581
+ "epoch": 0.95,
3582
+ "learning_rate": 1.1279824499064396e-06,
3583
+ "loss": 0.9902,
3584
+ "step": 596
3585
+ },
3586
+ {
3587
+ "epoch": 0.96,
3588
+ "learning_rate": 1.0516660902673448e-06,
3589
+ "loss": 0.998,
3590
+ "step": 597
3591
+ },
3592
+ {
3593
+ "epoch": 0.96,
3594
+ "learning_rate": 9.780089980330642e-07,
3595
+ "loss": 1.0156,
3596
+ "step": 598
3597
+ },
3598
+ {
3599
+ "epoch": 0.96,
3600
+ "learning_rate": 9.070131527609604e-07,
3601
+ "loss": 1.0391,
3602
+ "step": 599
3603
+ },
3604
+ {
3605
+ "epoch": 0.96,
3606
+ "learning_rate": 8.386804624865851e-07,
3607
+ "loss": 1.0195,
3608
+ "step": 600
3609
+ },
3610
+ {
3611
+ "epoch": 0.96,
3612
+ "learning_rate": 7.730127636723539e-07,
3613
+ "loss": 1.0156,
3614
+ "step": 601
3615
+ },
3616
+ {
3617
+ "epoch": 0.96,
3618
+ "learning_rate": 7.100118211581852e-07,
3619
+ "loss": 1.0059,
3620
+ "step": 602
3621
+ },
3622
+ {
3623
+ "epoch": 0.96,
3624
+ "learning_rate": 6.496793281141056e-07,
3625
+ "loss": 0.9902,
3626
+ "step": 603
3627
+ },
3628
+ {
3629
+ "epoch": 0.97,
3630
+ "learning_rate": 5.920169059947411e-07,
3631
+ "loss": 1.0039,
3632
+ "step": 604
3633
+ },
3634
+ {
3635
+ "epoch": 0.97,
3636
+ "learning_rate": 5.370261044956971e-07,
3637
+ "loss": 0.9766,
3638
+ "step": 605
3639
+ },
3640
+ {
3641
+ "epoch": 0.97,
3642
+ "learning_rate": 4.847084015119574e-07,
3643
+ "loss": 1.0117,
3644
+ "step": 606
3645
+ },
3646
+ {
3647
+ "epoch": 0.97,
3648
+ "learning_rate": 4.3506520309813947e-07,
3649
+ "loss": 1.0039,
3650
+ "step": 607
3651
+ },
3652
+ {
3653
+ "epoch": 0.97,
3654
+ "learning_rate": 3.8809784343072366e-07,
3655
+ "loss": 1.0195,
3656
+ "step": 608
3657
+ },
3658
+ {
3659
+ "epoch": 0.97,
3660
+ "learning_rate": 3.4380758477219333e-07,
3661
+ "loss": 1.0059,
3662
+ "step": 609
3663
+ },
3664
+ {
3665
+ "epoch": 0.98,
3666
+ "learning_rate": 3.0219561743707326e-07,
3667
+ "loss": 1.0391,
3668
+ "step": 610
3669
+ },
3670
+ {
3671
+ "epoch": 0.98,
3672
+ "learning_rate": 2.6326305976001055e-07,
3673
+ "loss": 1.0391,
3674
+ "step": 611
3675
+ },
3676
+ {
3677
+ "epoch": 0.98,
3678
+ "learning_rate": 2.2701095806565432e-07,
3679
+ "loss": 1.0059,
3680
+ "step": 612
3681
+ },
3682
+ {
3683
+ "epoch": 0.98,
3684
+ "learning_rate": 1.9344028664056713e-07,
3685
+ "loss": 1.0391,
3686
+ "step": 613
3687
+ },
3688
+ {
3689
+ "epoch": 0.98,
3690
+ "learning_rate": 1.6255194770704586e-07,
3691
+ "loss": 1.0234,
3692
+ "step": 614
3693
+ },
3694
+ {
3695
+ "epoch": 0.98,
3696
+ "learning_rate": 1.3434677139885222e-07,
3697
+ "loss": 0.9902,
3698
+ "step": 615
3699
+ },
3700
+ {
3701
+ "epoch": 0.99,
3702
+ "learning_rate": 1.0882551573891953e-07,
3703
+ "loss": 1.0312,
3704
+ "step": 616
3705
+ },
3706
+ {
3707
+ "epoch": 0.99,
3708
+ "learning_rate": 8.598886661895788e-08,
3709
+ "loss": 0.9883,
3710
+ "step": 617
3711
+ },
3712
+ {
3713
+ "epoch": 0.99,
3714
+ "learning_rate": 6.583743778106887e-08,
3715
+ "loss": 0.9824,
3716
+ "step": 618
3717
+ },
3718
+ {
3719
+ "epoch": 0.99,
3720
+ "learning_rate": 4.837177080119215e-08,
3721
+ "loss": 1.0078,
3722
+ "step": 619
3723
+ },
3724
+ {
3725
+ "epoch": 0.99,
3726
+ "learning_rate": 3.359233507459481e-08,
3727
+ "loss": 0.9883,
3728
+ "step": 620
3729
+ },
3730
+ {
3731
+ "epoch": 0.99,
3732
+ "learning_rate": 2.1499527803214846e-08,
3733
+ "loss": 1.0117,
3734
+ "step": 621
3735
+ },
3736
+ {
3737
+ "epoch": 1.0,
3738
+ "learning_rate": 1.209367398504746e-08,
3739
+ "loss": 0.9805,
3740
+ "step": 622
3741
+ },
3742
+ {
3743
+ "epoch": 1.0,
3744
+ "learning_rate": 5.375026405352035e-09,
3745
+ "loss": 1.0098,
3746
+ "step": 623
3747
+ },
3748
+ {
3749
+ "epoch": 1.0,
3750
+ "learning_rate": 1.3437656298687097e-09,
3751
+ "loss": 1.0273,
3752
+ "step": 624
3753
+ },
3754
+ {
3755
+ "epoch": 1.0,
3756
+ "learning_rate": 0.0,
3757
+ "loss": 0.9883,
3758
+ "step": 625
3759
+ },
3760
+ {
3761
+ "epoch": 1.0,
3762
+ "step": 625,
3763
+ "total_flos": 2364286311071744.0,
3764
+ "train_loss": 1.044653125,
3765
+ "train_runtime": 13892.2986,
3766
+ "train_samples_per_second": 5.759,
3767
+ "train_steps_per_second": 0.045
3768
+ }
3769
+ ],
3770
+ "max_steps": 625,
3771
+ "num_train_epochs": 1,
3772
+ "total_flos": 2364286311071744.0,
3773
+ "trial_name": null,
3774
+ "trial_params": null
3775
+ }