Hugging Face's logo

harvey2333
/
omni_1013-vicuna-7b-v1.3-finetune_lora_8epoch

PEFT
omni
Model card Files Community
harvey2333 commited on
Commit
a118412
·
1 Parent(s): 2ee0846

add model files

Browse files
Files changed (7) hide show
  1. README.md +9 -0
  2. adapter_config.json +26 -0
  3. adapter_model.bin +3 -0
  4. config.json +37 -0
  5. non_lora_trainables.bin +3 -0
  6. train_video.log +0 -0
  7. trainer_state.json +3721 -0
README.md ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+ ### Framework versions
7
+
8
+
9
+ - PEFT 0.4.0
adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "/home/wanghao/weights/vicuna-7b-v1.3",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 64,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "o_proj",
18
+ "k_proj",
19
+ "v_proj",
20
+ "down_proj",
21
+ "gate_proj",
22
+ "q_proj",
23
+ "up_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4d53f850a27df41e5ec3e1ed2ab3b0a9aeff82b558d9bcb41ec65ccb69adb4d
3
+ size 319970957
config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/home/wanghao/weights/vicuna-7b-v1.3",
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
+ "image_aspect_ratio": "square",
12
+ "image_grid_pinpoints": null,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 11008,
15
+ "max_position_embeddings": 2048,
16
+ "mm_hidden_size": 1024,
17
+ "mm_use_im_patch_token": false,
18
+ "mm_use_im_start_end": false,
19
+ "mm_vision_select_feature": "patch",
20
+ "mm_vision_select_layer": -2,
21
+ "mm_vision_tower": "openai/clip-vit-large-patch14",
22
+ "model_type": "omni",
23
+ "num_attention_heads": 32,
24
+ "num_hidden_layers": 32,
25
+ "num_key_value_heads": 32,
26
+ "pad_token_id": 0,
27
+ "pretraining_tp": 1,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_scaling": null,
30
+ "tie_word_embeddings": false,
31
+ "torch_dtype": "float16",
32
+ "transformers_version": "4.31.0",
33
+ "tune_mm_mlp_adapter": false,
34
+ "use_cache": true,
35
+ "use_mm_proj": true,
36
+ "vocab_size": 32000
37
+ }
non_lora_trainables.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45a3ebf009def7fe5c372a5e34fe77ae4998e504485e1c13eb05f288675f3810
3
+ size 9971943
train_video.log ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
@@ -0,0 +1,3721 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 8.0,
5
+ "global_step": 616,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1.0526315789473685e-06,
13
+ "loss": 12.2148,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 2.105263157894737e-06,
19
+ "loss": 12.0312,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 3.157894736842105e-06,
25
+ "loss": 12.3086,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 4.210526315789474e-06,
31
+ "loss": 12.1172,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.06,
36
+ "learning_rate": 5.263157894736842e-06,
37
+ "loss": 12.0117,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 6.31578947368421e-06,
43
+ "loss": 12.2656,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.09,
48
+ "learning_rate": 7.368421052631579e-06,
49
+ "loss": 12.125,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.1,
54
+ "learning_rate": 8.421052631578948e-06,
55
+ "loss": 11.2266,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.12,
60
+ "learning_rate": 9.473684210526315e-06,
61
+ "loss": 11.1523,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.13,
66
+ "learning_rate": 1.0526315789473684e-05,
67
+ "loss": 9.5234,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.14,
72
+ "learning_rate": 1.1578947368421053e-05,
73
+ "loss": 9.4688,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.16,
78
+ "learning_rate": 1.263157894736842e-05,
79
+ "loss": 9.25,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.17,
84
+ "learning_rate": 1.3684210526315791e-05,
85
+ "loss": 7.7285,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.18,
90
+ "learning_rate": 1.4736842105263159e-05,
91
+ "loss": 7.6367,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.19,
96
+ "learning_rate": 1.578947368421053e-05,
97
+ "loss": 7.4844,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.21,
102
+ "learning_rate": 1.6842105263157896e-05,
103
+ "loss": 7.2422,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.22,
108
+ "learning_rate": 1.7894736842105264e-05,
109
+ "loss": 7.0938,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.23,
114
+ "learning_rate": 1.894736842105263e-05,
115
+ "loss": 6.7266,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.25,
120
+ "learning_rate": 2e-05,
121
+ "loss": 6.5234,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.26,
126
+ "learning_rate": 1.9999861541352416e-05,
127
+ "loss": 6.3477,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.27,
132
+ "learning_rate": 1.9999446169243816e-05,
133
+ "loss": 6.127,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.29,
138
+ "learning_rate": 1.9998753895176576e-05,
139
+ "loss": 5.8555,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.3,
144
+ "learning_rate": 1.999778473832096e-05,
145
+ "loss": 5.7402,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.31,
150
+ "learning_rate": 1.9996538725514597e-05,
151
+ "loss": 5.5605,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.32,
156
+ "learning_rate": 1.999501589126174e-05,
157
+ "loss": 5.4199,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.34,
162
+ "learning_rate": 1.9993216277732302e-05,
163
+ "loss": 5.3242,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.35,
168
+ "learning_rate": 1.999113993476069e-05,
169
+ "loss": 5.2148,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.36,
174
+ "learning_rate": 1.9988786919844437e-05,
175
+ "loss": 5.1016,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.38,
180
+ "learning_rate": 1.9986157298142595e-05,
181
+ "loss": 5.0488,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.39,
186
+ "learning_rate": 1.9983251142473935e-05,
187
+ "loss": 4.9258,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.4,
192
+ "learning_rate": 1.9980068533314937e-05,
193
+ "loss": 4.9531,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.42,
198
+ "learning_rate": 1.9976609558797545e-05,
199
+ "loss": 4.8535,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.43,
204
+ "learning_rate": 1.9972874314706755e-05,
205
+ "loss": 4.8203,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.44,
210
+ "learning_rate": 1.9968862904477936e-05,
211
+ "loss": 4.8535,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.45,
216
+ "learning_rate": 1.9964575439193966e-05,
217
+ "loss": 4.7168,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.47,
222
+ "learning_rate": 1.996001203758218e-05,
223
+ "loss": 4.6875,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.48,
228
+ "learning_rate": 1.995517282601106e-05,
229
+ "loss": 4.6172,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.49,
234
+ "learning_rate": 1.9950057938486745e-05,
235
+ "loss": 4.6523,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.51,
240
+ "learning_rate": 1.994466751664932e-05,
241
+ "loss": 4.5195,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.52,
246
+ "learning_rate": 1.993900170976888e-05,
247
+ "loss": 4.5117,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.53,
252
+ "learning_rate": 1.9933060674741422e-05,
253
+ "loss": 4.4141,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.55,
258
+ "learning_rate": 1.9926844576084483e-05,
259
+ "loss": 4.3398,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.56,
264
+ "learning_rate": 1.992035358593258e-05,
265
+ "loss": 4.3232,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.57,
270
+ "learning_rate": 1.991358788403246e-05,
271
+ "loss": 4.2305,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.58,
276
+ "learning_rate": 1.990654765773811e-05,
277
+ "loss": 4.1641,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.6,
282
+ "learning_rate": 1.9899233102005573e-05,
283
+ "loss": 4.0674,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.61,
288
+ "learning_rate": 1.9891644419387545e-05,
289
+ "loss": 3.915,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.62,
294
+ "learning_rate": 1.9883781820027777e-05,
295
+ "loss": 3.7822,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.64,
300
+ "learning_rate": 1.987564552165524e-05,
301
+ "loss": 3.709,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.65,
306
+ "learning_rate": 1.9867235749578108e-05,
307
+ "loss": 3.4131,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.66,
312
+ "learning_rate": 1.9858552736677516e-05,
313
+ "loss": 3.1318,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.68,
318
+ "learning_rate": 1.984959672340111e-05,
319
+ "loss": 2.834,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.69,
324
+ "learning_rate": 1.984036795775638e-05,
325
+ "loss": 2.5654,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.7,
330
+ "learning_rate": 1.9830866695303817e-05,
331
+ "loss": 2.417,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.71,
336
+ "learning_rate": 1.9821093199149806e-05,
337
+ "loss": 2.1909,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.73,
342
+ "learning_rate": 1.981104773993936e-05,
343
+ "loss": 2.2568,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.74,
348
+ "learning_rate": 1.980073059584862e-05,
349
+ "loss": 2.2744,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.75,
354
+ "learning_rate": 1.9790142052577148e-05,
355
+ "loss": 2.0771,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.77,
360
+ "learning_rate": 1.977928240334002e-05,
361
+ "loss": 2.1729,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.78,
366
+ "learning_rate": 1.9768151948859705e-05,
367
+ "loss": 2.123,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.79,
372
+ "learning_rate": 1.9756750997357738e-05,
373
+ "loss": 2.0356,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.81,
378
+ "learning_rate": 1.9745079864546184e-05,
379
+ "loss": 2.0142,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.82,
384
+ "learning_rate": 1.97331388736189e-05,
385
+ "loss": 2.061,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.83,
390
+ "learning_rate": 1.972092835524257e-05,
391
+ "loss": 2.0508,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.84,
396
+ "learning_rate": 1.9708448647547575e-05,
397
+ "loss": 2.0171,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.86,
402
+ "learning_rate": 1.9695700096118594e-05,
403
+ "loss": 2.1284,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.87,
408
+ "learning_rate": 1.9682683053985073e-05,
409
+ "loss": 2.0166,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.88,
414
+ "learning_rate": 1.966939788161142e-05,
415
+ "loss": 2.062,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.9,
420
+ "learning_rate": 1.9655844946887035e-05,
421
+ "loss": 2.0142,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.91,
426
+ "learning_rate": 1.9642024625116117e-05,
427
+ "loss": 2.0103,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.92,
432
+ "learning_rate": 1.9627937299007286e-05,
433
+ "loss": 1.9956,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.94,
438
+ "learning_rate": 1.961358335866296e-05,
439
+ "loss": 1.9868,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.95,
444
+ "learning_rate": 1.959896320156857e-05,
445
+ "loss": 2.0435,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.96,
450
+ "learning_rate": 1.958407723258156e-05,
451
+ "loss": 2.0112,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.97,
456
+ "learning_rate": 1.9568925863920155e-05,
457
+ "loss": 2.0908,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.99,
462
+ "learning_rate": 1.955350951515195e-05,
463
+ "loss": 1.9795,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 1.0,
468
+ "learning_rate": 1.9537828613182314e-05,
469
+ "loss": 2.0112,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 1.01,
474
+ "learning_rate": 1.9521883592242537e-05,
475
+ "loss": 2.0459,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 1.03,
480
+ "learning_rate": 1.950567489387783e-05,
481
+ "loss": 2.0117,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 1.04,
486
+ "learning_rate": 1.9489202966935084e-05,
487
+ "loss": 2.0156,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 1.05,
492
+ "learning_rate": 1.947246826755044e-05,
493
+ "loss": 2.0547,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 1.06,
498
+ "learning_rate": 1.945547125913667e-05,
499
+ "loss": 1.9639,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 1.08,
504
+ "learning_rate": 1.943821241237034e-05,
505
+ "loss": 2.019,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 1.09,
510
+ "learning_rate": 1.9420692205178753e-05,
511
+ "loss": 1.9771,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 1.1,
516
+ "learning_rate": 1.9402911122726756e-05,
517
+ "loss": 1.9492,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 1.12,
522
+ "learning_rate": 1.9384869657403277e-05,
523
+ "loss": 1.9702,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 1.13,
528
+ "learning_rate": 1.9366568308807685e-05,
529
+ "loss": 1.9946,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 1.14,
534
+ "learning_rate": 1.9348007583735985e-05,
535
+ "loss": 1.9854,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 1.16,
540
+ "learning_rate": 1.9329187996166747e-05,
541
+ "loss": 1.959,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 1.17,
546
+ "learning_rate": 1.9310110067246905e-05,
547
+ "loss": 1.9722,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 1.18,
552
+ "learning_rate": 1.9290774325277305e-05,
553
+ "loss": 2.0376,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 1.19,
558
+ "learning_rate": 1.9271181305698084e-05,
559
+ "loss": 1.9834,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 1.21,
564
+ "learning_rate": 1.9251331551073843e-05,
565
+ "loss": 2.0049,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 1.22,
570
+ "learning_rate": 1.923122561107861e-05,
571
+ "loss": 1.9824,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 1.23,
576
+ "learning_rate": 1.9210864042480645e-05,
577
+ "loss": 1.9624,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 1.25,
582
+ "learning_rate": 1.9190247409126993e-05,
583
+ "loss": 1.9395,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 1.26,
588
+ "learning_rate": 1.916937628192789e-05,
589
+ "loss": 1.9746,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 1.27,
594
+ "learning_rate": 1.9148251238840947e-05,
595
+ "loss": 1.9507,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 1.29,
600
+ "learning_rate": 1.9126872864855142e-05,
601
+ "loss": 2.0054,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 1.3,
606
+ "learning_rate": 1.9105241751974624e-05,
607
+ "loss": 1.9409,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 1.31,
612
+ "learning_rate": 1.9083358499202323e-05,
613
+ "loss": 1.9912,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 1.32,
618
+ "learning_rate": 1.9061223712523352e-05,
619
+ "loss": 1.9404,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 1.34,
624
+ "learning_rate": 1.903883800488824e-05,
625
+ "loss": 1.9102,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 1.35,
630
+ "learning_rate": 1.9016201996195943e-05,
631
+ "loss": 1.9248,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 1.36,
636
+ "learning_rate": 1.8993316313276694e-05,
637
+ "loss": 1.8984,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 1.38,
642
+ "learning_rate": 1.8970181589874637e-05,
643
+ "loss": 1.9331,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 1.39,
648
+ "learning_rate": 1.894679846663027e-05,
649
+ "loss": 1.9561,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 1.4,
654
+ "learning_rate": 1.8923167591062723e-05,
655
+ "loss": 1.8901,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 1.42,
660
+ "learning_rate": 1.8899289617551803e-05,
661
+ "loss": 1.9922,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 1.43,
666
+ "learning_rate": 1.8875165207319902e-05,
667
+ "loss": 1.9277,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 1.44,
672
+ "learning_rate": 1.8850795028413658e-05,
673
+ "loss": 1.9185,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 1.45,
678
+ "learning_rate": 1.882617975568547e-05,
679
+ "loss": 1.9575,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 1.47,
684
+ "learning_rate": 1.880132007077482e-05,
685
+ "loss": 1.957,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 1.48,
690
+ "learning_rate": 1.8776216662089373e-05,
691
+ "loss": 1.8984,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 1.49,
696
+ "learning_rate": 1.875087022478594e-05,
697
+ "loss": 1.9429,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 1.51,
702
+ "learning_rate": 1.8725281460751198e-05,
703
+ "loss": 1.8701,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 1.52,
708
+ "learning_rate": 1.869945107858228e-05,
709
+ "loss": 1.9497,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 1.53,
714
+ "learning_rate": 1.867337979356715e-05,
715
+ "loss": 1.8921,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 1.55,
720
+ "learning_rate": 1.8647068327664774e-05,
721
+ "loss": 1.8569,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 1.56,
726
+ "learning_rate": 1.8620517409485148e-05,
727
+ "loss": 1.8882,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 1.57,
732
+ "learning_rate": 1.8593727774269122e-05,
733
+ "loss": 1.8765,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 1.58,
738
+ "learning_rate": 1.8566700163868027e-05,
739
+ "loss": 1.9282,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 1.6,
744
+ "learning_rate": 1.8539435326723135e-05,
745
+ "loss": 1.8384,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 1.61,
750
+ "learning_rate": 1.851193401784495e-05,
751
+ "loss": 1.9185,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 1.62,
756
+ "learning_rate": 1.848419699879227e-05,
757
+ "loss": 1.834,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 1.64,
762
+ "learning_rate": 1.845622503765113e-05,
763
+ "loss": 1.8657,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 1.65,
768
+ "learning_rate": 1.842801890901351e-05,
769
+ "loss": 1.8457,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 1.66,
774
+ "learning_rate": 1.8399579393955893e-05,
775
+ "loss": 1.7671,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 1.68,
780
+ "learning_rate": 1.837090728001764e-05,
781
+ "loss": 1.8462,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 1.69,
786
+ "learning_rate": 1.834200336117918e-05,
787
+ "loss": 1.8296,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 1.7,
792
+ "learning_rate": 1.8312868437840002e-05,
793
+ "loss": 1.8262,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 1.71,
798
+ "learning_rate": 1.8283503316796536e-05,
799
+ "loss": 1.835,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 1.73,
804
+ "learning_rate": 1.8253908811219764e-05,
805
+ "loss": 1.8979,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 1.74,
810
+ "learning_rate": 1.822408574063273e-05,
811
+ "loss": 1.8496,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 1.75,
816
+ "learning_rate": 1.8194034930887842e-05,
817
+ "loss": 1.8252,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 1.77,
822
+ "learning_rate": 1.8163757214143993e-05,
823
+ "loss": 1.7812,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 1.78,
828
+ "learning_rate": 1.8133253428843524e-05,
829
+ "loss": 1.8364,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 1.79,
834
+ "learning_rate": 1.810252441968901e-05,
835
+ "loss": 1.8013,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 1.81,
840
+ "learning_rate": 1.8071571037619856e-05,
841
+ "loss": 1.8203,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 1.82,
846
+ "learning_rate": 1.804039413978875e-05,
847
+ "loss": 1.7729,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 1.83,
852
+ "learning_rate": 1.8008994589537913e-05,
853
+ "loss": 1.8491,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 1.84,
858
+ "learning_rate": 1.7977373256375194e-05,
859
+ "loss": 1.7998,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 1.86,
864
+ "learning_rate": 1.7945531015950008e-05,
865
+ "loss": 1.8364,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 1.87,
870
+ "learning_rate": 1.791346875002905e-05,
871
+ "loss": 1.8125,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 1.88,
876
+ "learning_rate": 1.7881187346471924e-05,
877
+ "loss": 1.832,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 1.9,
882
+ "learning_rate": 1.784868769920653e-05,
883
+ "loss": 1.8271,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 1.91,
888
+ "learning_rate": 1.7815970708204296e-05,
889
+ "loss": 1.7959,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 1.92,
894
+ "learning_rate": 1.77830372794553e-05,
895
+ "loss": 1.7798,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 1.94,
900
+ "learning_rate": 1.774988832494314e-05,
901
+ "loss": 1.7651,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 1.95,
906
+ "learning_rate": 1.7716524762619695e-05,
907
+ "loss": 1.8076,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 1.96,
912
+ "learning_rate": 1.7682947516379706e-05,
913
+ "loss": 1.8379,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 1.97,
918
+ "learning_rate": 1.7649157516035205e-05,
919
+ "loss": 1.8228,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 1.99,
924
+ "learning_rate": 1.7615155697289734e-05,
925
+ "loss": 1.7783,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 2.0,
930
+ "learning_rate": 1.7580943001712457e-05,
931
+ "loss": 1.8188,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 2.01,
936
+ "learning_rate": 1.7546520376712093e-05,
937
+ "loss": 1.7974,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 2.03,
942
+ "learning_rate": 1.7511888775510662e-05,
943
+ "loss": 1.7964,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 2.04,
948
+ "learning_rate": 1.7477049157117093e-05,
949
+ "loss": 1.7515,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 2.05,
954
+ "learning_rate": 1.744200248630068e-05,
955
+ "loss": 1.7725,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 2.06,
960
+ "learning_rate": 1.7406749733564344e-05,
961
+ "loss": 1.7534,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 2.08,
966
+ "learning_rate": 1.737129187511779e-05,
967
+ "loss": 1.8408,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 2.09,
972
+ "learning_rate": 1.7335629892850436e-05,
973
+ "loss": 1.7686,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 2.1,
978
+ "learning_rate": 1.729976477430425e-05,
979
+ "loss": 1.7642,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 2.12,
984
+ "learning_rate": 1.7263697512646397e-05,
985
+ "loss": 1.8047,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 2.13,
990
+ "learning_rate": 1.7227429106641726e-05,
991
+ "loss": 1.8301,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 2.14,
996
+ "learning_rate": 1.7190960560625127e-05,
997
+ "loss": 1.7588,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 2.16,
1002
+ "learning_rate": 1.7154292884473712e-05,
1003
+ "loss": 1.7749,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 2.17,
1008
+ "learning_rate": 1.711742709357886e-05,
1009
+ "loss": 1.7251,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 2.18,
1014
+ "learning_rate": 1.708036420881807e-05,
1015
+ "loss": 1.7603,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 2.19,
1020
+ "learning_rate": 1.7043105256526723e-05,
1021
+ "loss": 1.7339,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 2.21,
1026
+ "learning_rate": 1.7005651268469652e-05,
1027
+ "loss": 1.731,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 2.22,
1032
+ "learning_rate": 1.6968003281812563e-05,
1033
+ "loss": 1.7598,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 2.23,
1038
+ "learning_rate": 1.693016233909332e-05,
1039
+ "loss": 1.7007,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 2.25,
1044
+ "learning_rate": 1.689212948819307e-05,
1045
+ "loss": 1.7183,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 2.26,
1050
+ "learning_rate": 1.6853905782307235e-05,
1051
+ "loss": 1.7173,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 2.27,
1056
+ "learning_rate": 1.681549227991634e-05,
1057
+ "loss": 1.7856,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 2.29,
1062
+ "learning_rate": 1.67768900447567e-05,
1063
+ "loss": 1.7329,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 2.3,
1068
+ "learning_rate": 1.6738100145790977e-05,
1069
+ "loss": 1.7578,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 2.31,
1074
+ "learning_rate": 1.6699123657178553e-05,
1075
+ "loss": 1.6846,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 2.32,
1080
+ "learning_rate": 1.6659961658245813e-05,
1081
+ "loss": 1.791,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 2.34,
1086
+ "learning_rate": 1.6620615233456235e-05,
1087
+ "loss": 1.7798,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 2.35,
1092
+ "learning_rate": 1.658108547238038e-05,
1093
+ "loss": 1.6987,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 2.36,
1098
+ "learning_rate": 1.6541373469665688e-05,
1099
+ "loss": 1.7202,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 2.38,
1104
+ "learning_rate": 1.6501480325006206e-05,
1105
+ "loss": 1.7285,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 2.39,
1110
+ "learning_rate": 1.64614071431121e-05,
1111
+ "loss": 1.7417,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 2.4,
1116
+ "learning_rate": 1.6421155033679085e-05,
1117
+ "loss": 1.79,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 2.42,
1122
+ "learning_rate": 1.6380725111357693e-05,
1123
+ "loss": 1.7876,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 2.43,
1128
+ "learning_rate": 1.634011849572239e-05,
1129
+ "loss": 1.7734,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 2.44,
1134
+ "learning_rate": 1.6299336311240593e-05,
1135
+ "loss": 1.7686,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 2.45,
1140
+ "learning_rate": 1.6258379687241533e-05,
1141
+ "loss": 1.7993,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 2.47,
1146
+ "learning_rate": 1.6217249757884954e-05,
1147
+ "loss": 1.708,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 2.48,
1152
+ "learning_rate": 1.6175947662129735e-05,
1153
+ "loss": 1.7065,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 2.49,
1158
+ "learning_rate": 1.6134474543702353e-05,
1159
+ "loss": 1.7324,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 2.51,
1164
+ "learning_rate": 1.609283155106517e-05,
1165
+ "loss": 1.7686,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 2.52,
1170
+ "learning_rate": 1.605101983738468e-05,
1171
+ "loss": 1.7563,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 2.53,
1176
+ "learning_rate": 1.6009040560499548e-05,
1177
+ "loss": 1.7373,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 2.55,
1182
+ "learning_rate": 1.596689488288856e-05,
1183
+ "loss": 1.7104,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 2.56,
1188
+ "learning_rate": 1.5924583971638416e-05,
1189
+ "loss": 1.7368,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 2.57,
1194
+ "learning_rate": 1.5882108998411427e-05,
1195
+ "loss": 1.7886,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 2.58,
1200
+ "learning_rate": 1.5839471139413065e-05,
1201
+ "loss": 1.6855,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 2.6,
1206
+ "learning_rate": 1.5796671575359382e-05,
1207
+ "loss": 1.7158,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 2.61,
1212
+ "learning_rate": 1.5753711491444336e-05,
1213
+ "loss": 1.7144,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 2.62,
1218
+ "learning_rate": 1.571059207730695e-05,
1219
+ "loss": 1.6909,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 2.64,
1224
+ "learning_rate": 1.5667314526998373e-05,
1225
+ "loss": 1.8003,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 2.65,
1230
+ "learning_rate": 1.5623880038948828e-05,
1231
+ "loss": 1.7231,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 2.66,
1236
+ "learning_rate": 1.55802898159344e-05,
1237
+ "loss": 1.6816,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 2.68,
1242
+ "learning_rate": 1.553654506504377e-05,
1243
+ "loss": 1.6826,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 2.69,
1248
+ "learning_rate": 1.5492646997644737e-05,
1249
+ "loss": 1.7085,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 2.7,
1254
+ "learning_rate": 1.5448596829350706e-05,
1255
+ "loss": 1.6797,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 2.71,
1260
+ "learning_rate": 1.540439577998703e-05,
1261
+ "loss": 1.708,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 2.73,
1266
+ "learning_rate": 1.5360045073557214e-05,
1267
+ "loss": 1.7036,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 2.74,
1272
+ "learning_rate": 1.5315545938209016e-05,
1273
+ "loss": 1.7129,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 2.75,
1278
+ "learning_rate": 1.527089960620046e-05,
1279
+ "loss": 1.6855,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 2.77,
1284
+ "learning_rate": 1.5226107313865701e-05,
1285
+ "loss": 1.645,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 2.78,
1290
+ "learning_rate": 1.5181170301580776e-05,
1291
+ "loss": 1.6982,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 2.79,
1296
+ "learning_rate": 1.5136089813729276e-05,
1297
+ "loss": 1.731,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 2.81,
1302
+ "learning_rate": 1.509086709866788e-05,
1303
+ "loss": 1.7192,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 2.82,
1308
+ "learning_rate": 1.5045503408691776e-05,
1309
+ "loss": 1.6982,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 2.83,
1314
+ "learning_rate": 1.5000000000000002e-05,
1315
+ "loss": 1.7266,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 2.84,
1320
+ "learning_rate": 1.495435813266064e-05,
1321
+ "loss": 1.6958,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 2.86,
1326
+ "learning_rate": 1.4908579070575936e-05,
1327
+ "loss": 1.7056,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 2.87,
1332
+ "learning_rate": 1.4862664081447297e-05,
1333
+ "loss": 1.6943,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 2.88,
1338
+ "learning_rate": 1.4816614436740184e-05,
1339
+ "loss": 1.6724,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 2.9,
1344
+ "learning_rate": 1.4770431411648898e-05,
1345
+ "loss": 1.6641,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 2.91,
1350
+ "learning_rate": 1.4724116285061278e-05,
1351
+ "loss": 1.7461,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 2.92,
1356
+ "learning_rate": 1.4677670339523285e-05,
1357
+ "loss": 1.7207,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 2.94,
1362
+ "learning_rate": 1.4631094861203478e-05,
1363
+ "loss": 1.7061,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 2.95,
1368
+ "learning_rate": 1.4584391139857407e-05,
1369
+ "loss": 1.6758,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 2.96,
1374
+ "learning_rate": 1.4537560468791889e-05,
1375
+ "loss": 1.73,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 2.97,
1380
+ "learning_rate": 1.4490604144829204e-05,
1381
+ "loss": 1.7314,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 2.99,
1386
+ "learning_rate": 1.4443523468271168e-05,
1387
+ "loss": 1.7114,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 3.0,
1392
+ "learning_rate": 1.4396319742863145e-05,
1393
+ "loss": 1.7212,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 3.01,
1398
+ "learning_rate": 1.4348994275757933e-05,
1399
+ "loss": 1.7036,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 3.03,
1404
+ "learning_rate": 1.4301548377479562e-05,
1405
+ "loss": 1.71,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 3.04,
1410
+ "learning_rate": 1.4253983361887017e-05,
1411
+ "loss": 1.7432,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 3.05,
1416
+ "learning_rate": 1.4206300546137844e-05,
1417
+ "loss": 1.6992,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 3.06,
1422
+ "learning_rate": 1.415850125065168e-05,
1423
+ "loss": 1.7271,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 3.08,
1428
+ "learning_rate": 1.4110586799073684e-05,
1429
+ "loss": 1.6792,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 3.09,
1434
+ "learning_rate": 1.4062558518237893e-05,
1435
+ "loss": 1.73,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 3.1,
1440
+ "learning_rate": 1.4014417738130464e-05,
1441
+ "loss": 1.7192,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 3.12,
1446
+ "learning_rate": 1.3966165791852862e-05,
1447
+ "loss": 1.7476,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 3.13,
1452
+ "learning_rate": 1.3917804015584932e-05,
1453
+ "loss": 1.6958,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 3.14,
1458
+ "learning_rate": 1.3869333748547901e-05,
1459
+ "loss": 1.6865,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 3.16,
1464
+ "learning_rate": 1.3820756332967294e-05,
1465
+ "loss": 1.668,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 3.17,
1470
+ "learning_rate": 1.3772073114035762e-05,
1471
+ "loss": 1.6826,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 3.18,
1476
+ "learning_rate": 1.3723285439875836e-05,
1477
+ "loss": 1.7227,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 3.19,
1482
+ "learning_rate": 1.3674394661502595e-05,
1483
+ "loss": 1.7163,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 3.21,
1488
+ "learning_rate": 1.3625402132786247e-05,
1489
+ "loss": 1.6606,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 3.22,
1494
+ "learning_rate": 1.3576309210414646e-05,
1495
+ "loss": 1.7085,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 3.23,
1500
+ "learning_rate": 1.352711725385572e-05,
1501
+ "loss": 1.668,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 3.25,
1506
+ "learning_rate": 1.3477827625319826e-05,
1507
+ "loss": 1.7173,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 3.26,
1512
+ "learning_rate": 1.3428441689722023e-05,
1513
+ "loss": 1.7656,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 3.27,
1518
+ "learning_rate": 1.3378960814644283e-05,
1519
+ "loss": 1.6812,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 3.29,
1524
+ "learning_rate": 1.3329386370297615e-05,
1525
+ "loss": 1.6953,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 3.3,
1530
+ "learning_rate": 1.3279719729484117e-05,
1531
+ "loss": 1.665,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 3.31,
1536
+ "learning_rate": 1.3229962267558982e-05,
1537
+ "loss": 1.6587,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 3.32,
1542
+ "learning_rate": 1.3180115362392383e-05,
1543
+ "loss": 1.6797,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 3.34,
1548
+ "learning_rate": 1.3130180394331335e-05,
1549
+ "loss": 1.6992,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 3.35,
1554
+ "learning_rate": 1.3080158746161468e-05,
1555
+ "loss": 1.6567,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 3.36,
1560
+ "learning_rate": 1.3030051803068729e-05,
1561
+ "loss": 1.6641,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 3.38,
1566
+ "learning_rate": 1.2979860952601038e-05,
1567
+ "loss": 1.6841,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 3.39,
1572
+ "learning_rate": 1.2929587584629845e-05,
1573
+ "loss": 1.6777,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 3.4,
1578
+ "learning_rate": 1.2879233091311667e-05,
1579
+ "loss": 1.7065,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 3.42,
1584
+ "learning_rate": 1.2828798867049504e-05,
1585
+ "loss": 1.6997,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 3.43,
1590
+ "learning_rate": 1.2778286308454255e-05,
1591
+ "loss": 1.6704,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 3.44,
1596
+ "learning_rate": 1.2727696814306034e-05,
1597
+ "loss": 1.6489,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 3.45,
1602
+ "learning_rate": 1.2677031785515423e-05,
1603
+ "loss": 1.6777,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 3.47,
1608
+ "learning_rate": 1.26262926250847e-05,
1609
+ "loss": 1.6284,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 3.48,
1614
+ "learning_rate": 1.2575480738068971e-05,
1615
+ "loss": 1.6445,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 3.49,
1620
+ "learning_rate": 1.2524597531537261e-05,
1621
+ "loss": 1.626,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 3.51,
1626
+ "learning_rate": 1.2473644414533573e-05,
1627
+ "loss": 1.626,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 3.52,
1632
+ "learning_rate": 1.2422622798037833e-05,
1633
+ "loss": 1.6919,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 3.53,
1638
+ "learning_rate": 1.2371534094926852e-05,
1639
+ "loss": 1.6602,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 3.55,
1644
+ "learning_rate": 1.232037971993517e-05,
1645
+ "loss": 1.6401,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 3.56,
1650
+ "learning_rate": 1.2269161089615902e-05,
1651
+ "loss": 1.7026,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 3.57,
1656
+ "learning_rate": 1.2217879622301514e-05,
1657
+ "loss": 1.6875,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 3.58,
1662
+ "learning_rate": 1.2166536738064523e-05,
1663
+ "loss": 1.6646,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 3.6,
1668
+ "learning_rate": 1.2115133858678192e-05,
1669
+ "loss": 1.6631,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 3.61,
1674
+ "learning_rate": 1.2063672407577154e-05,
1675
+ "loss": 1.6196,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 3.62,
1680
+ "learning_rate": 1.2012153809817992e-05,
1681
+ "loss": 1.6606,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 3.64,
1686
+ "learning_rate": 1.1960579492039783e-05,
1687
+ "loss": 1.6719,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 3.65,
1692
+ "learning_rate": 1.1908950882424581e-05,
1693
+ "loss": 1.6958,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 3.66,
1698
+ "learning_rate": 1.1857269410657883e-05,
1699
+ "loss": 1.645,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 3.68,
1704
+ "learning_rate": 1.1805536507889021e-05,
1705
+ "loss": 1.6782,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 3.69,
1710
+ "learning_rate": 1.1753753606691554e-05,
1711
+ "loss": 1.6724,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 3.7,
1716
+ "learning_rate": 1.1701922141023566e-05,
1717
+ "loss": 1.6108,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 3.71,
1722
+ "learning_rate": 1.1650043546187994e-05,
1723
+ "loss": 1.6313,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 3.73,
1728
+ "learning_rate": 1.1598119258792848e-05,
1729
+ "loss": 1.647,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 3.74,
1734
+ "learning_rate": 1.1546150716711448e-05,
1735
+ "loss": 1.6816,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 3.75,
1740
+ "learning_rate": 1.1494139359042612e-05,
1741
+ "loss": 1.6846,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 3.77,
1746
+ "learning_rate": 1.1442086626070781e-05,
1747
+ "loss": 1.6602,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 3.78,
1752
+ "learning_rate": 1.1389993959226163e-05,
1753
+ "loss": 1.6133,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 3.79,
1758
+ "learning_rate": 1.1337862801044792e-05,
1759
+ "loss": 1.6997,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 3.81,
1764
+ "learning_rate": 1.1285694595128606e-05,
1765
+ "loss": 1.6172,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 3.82,
1770
+ "learning_rate": 1.123349078610545e-05,
1771
+ "loss": 1.6479,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 3.83,
1776
+ "learning_rate": 1.1181252819589081e-05,
1777
+ "loss": 1.6851,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 3.84,
1782
+ "learning_rate": 1.1128982142139142e-05,
1783
+ "loss": 1.6533,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 3.86,
1788
+ "learning_rate": 1.1076680201221093e-05,
1789
+ "loss": 1.6367,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 3.87,
1794
+ "learning_rate": 1.1024348445166133e-05,
1795
+ "loss": 1.6426,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 3.88,
1800
+ "learning_rate": 1.0971988323131099e-05,
1801
+ "loss": 1.6509,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 3.9,
1806
+ "learning_rate": 1.091960128505833e-05,
1807
+ "loss": 1.6997,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 3.91,
1812
+ "learning_rate": 1.086718878163551e-05,
1813
+ "loss": 1.6187,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 3.92,
1818
+ "learning_rate": 1.0814752264255508e-05,
1819
+ "loss": 1.6914,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 3.94,
1824
+ "learning_rate": 1.0762293184976178e-05,
1825
+ "loss": 1.6421,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 3.95,
1830
+ "learning_rate": 1.070981299648016e-05,
1831
+ "loss": 1.6631,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 3.96,
1836
+ "learning_rate": 1.0657313152034634e-05,
1837
+ "loss": 1.7046,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 3.97,
1842
+ "learning_rate": 1.0604795105451096e-05,
1843
+ "loss": 1.5845,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 3.99,
1848
+ "learning_rate": 1.0552260311045082e-05,
1849
+ "loss": 1.6621,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 4.0,
1854
+ "learning_rate": 1.0499710223595913e-05,
1855
+ "loss": 1.6782,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 4.01,
1860
+ "learning_rate": 1.0447146298306394e-05,
1861
+ "loss": 1.6611,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 4.03,
1866
+ "learning_rate": 1.0394569990762528e-05,
1867
+ "loss": 1.6626,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 4.04,
1872
+ "learning_rate": 1.0341982756893203e-05,
1873
+ "loss": 1.6406,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 4.05,
1878
+ "learning_rate": 1.0289386052929874e-05,
1879
+ "loss": 1.6455,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 4.06,
1884
+ "learning_rate": 1.0236781335366239e-05,
1885
+ "loss": 1.7051,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 4.08,
1890
+ "learning_rate": 1.0184170060917914e-05,
1891
+ "loss": 1.5967,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 4.09,
1896
+ "learning_rate": 1.0131553686482077e-05,
1897
+ "loss": 1.6772,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 4.1,
1902
+ "learning_rate": 1.0078933669097135e-05,
1903
+ "loss": 1.625,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 4.12,
1908
+ "learning_rate": 1.002631146590238e-05,
1909
+ "loss": 1.6572,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 4.13,
1914
+ "learning_rate": 9.973688534097624e-06,
1915
+ "loss": 1.6694,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 4.14,
1920
+ "learning_rate": 9.92106633090287e-06,
1921
+ "loss": 1.6377,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 4.16,
1926
+ "learning_rate": 9.868446313517927e-06,
1927
+ "loss": 1.6782,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 4.17,
1932
+ "learning_rate": 9.815829939082087e-06,
1933
+ "loss": 1.6147,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 4.18,
1938
+ "learning_rate": 9.763218664633763e-06,
1939
+ "loss": 1.6826,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 4.19,
1944
+ "learning_rate": 9.710613947070127e-06,
1945
+ "loss": 1.7041,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 4.21,
1950
+ "learning_rate": 9.658017243106802e-06,
1951
+ "loss": 1.6343,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 4.22,
1956
+ "learning_rate": 9.605430009237474e-06,
1957
+ "loss": 1.6724,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 4.23,
1962
+ "learning_rate": 9.552853701693606e-06,
1963
+ "loss": 1.6812,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 4.25,
1968
+ "learning_rate": 9.50028977640409e-06,
1969
+ "loss": 1.6289,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 4.26,
1974
+ "learning_rate": 9.44773968895492e-06,
1975
+ "loss": 1.6313,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 4.27,
1980
+ "learning_rate": 9.395204894548907e-06,
1981
+ "loss": 1.6274,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 4.29,
1986
+ "learning_rate": 9.342686847965367e-06,
1987
+ "loss": 1.6572,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 4.3,
1992
+ "learning_rate": 9.290187003519841e-06,
1993
+ "loss": 1.6333,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 4.31,
1998
+ "learning_rate": 9.237706815023824e-06,
1999
+ "loss": 1.687,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 4.32,
2004
+ "learning_rate": 9.185247735744495e-06,
2005
+ "loss": 1.6626,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 4.34,
2010
+ "learning_rate": 9.132811218364494e-06,
2011
+ "loss": 1.6431,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 4.35,
2016
+ "learning_rate": 9.080398714941672e-06,
2017
+ "loss": 1.6562,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 4.36,
2022
+ "learning_rate": 9.028011676868901e-06,
2023
+ "loss": 1.6714,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 4.38,
2028
+ "learning_rate": 8.975651554833869e-06,
2029
+ "loss": 1.604,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 4.39,
2034
+ "learning_rate": 8.92331979877891e-06,
2035
+ "loss": 1.6719,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 4.4,
2040
+ "learning_rate": 8.871017857860863e-06,
2041
+ "loss": 1.707,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 4.42,
2046
+ "learning_rate": 8.81874718041092e-06,
2047
+ "loss": 1.647,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 4.43,
2052
+ "learning_rate": 8.766509213894552e-06,
2053
+ "loss": 1.6675,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 4.44,
2058
+ "learning_rate": 8.714305404871397e-06,
2059
+ "loss": 1.6636,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 4.45,
2064
+ "learning_rate": 8.662137198955211e-06,
2065
+ "loss": 1.6768,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 4.47,
2070
+ "learning_rate": 8.610006040773844e-06,
2071
+ "loss": 1.5864,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 4.48,
2076
+ "learning_rate": 8.557913373929222e-06,
2077
+ "loss": 1.6304,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 4.49,
2082
+ "learning_rate": 8.50586064095739e-06,
2083
+ "loss": 1.6289,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 4.51,
2088
+ "learning_rate": 8.453849283288554e-06,
2089
+ "loss": 1.6436,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 4.52,
2094
+ "learning_rate": 8.401880741207155e-06,
2095
+ "loss": 1.6221,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 4.53,
2100
+ "learning_rate": 8.349956453812009e-06,
2101
+ "loss": 1.6904,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 4.55,
2106
+ "learning_rate": 8.298077858976435e-06,
2107
+ "loss": 1.5898,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 4.56,
2112
+ "learning_rate": 8.246246393308448e-06,
2113
+ "loss": 1.667,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 4.57,
2118
+ "learning_rate": 8.194463492110982e-06,
2119
+ "loss": 1.6543,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 4.58,
2124
+ "learning_rate": 8.142730589342119e-06,
2125
+ "loss": 1.6572,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 4.6,
2130
+ "learning_rate": 8.091049117575424e-06,
2131
+ "loss": 1.6685,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 4.61,
2136
+ "learning_rate": 8.03942050796022e-06,
2137
+ "loss": 1.6484,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 4.62,
2142
+ "learning_rate": 7.98784619018201e-06,
2143
+ "loss": 1.5405,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 4.64,
2148
+ "learning_rate": 7.93632759242285e-06,
2149
+ "loss": 1.644,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 4.65,
2154
+ "learning_rate": 7.884866141321811e-06,
2155
+ "loss": 1.6206,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 4.66,
2160
+ "learning_rate": 7.833463261935482e-06,
2161
+ "loss": 1.6079,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 4.68,
2166
+ "learning_rate": 7.782120377698489e-06,
2167
+ "loss": 1.6108,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 4.69,
2172
+ "learning_rate": 7.730838910384098e-06,
2173
+ "loss": 1.5625,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 4.7,
2178
+ "learning_rate": 7.679620280064837e-06,
2179
+ "loss": 1.647,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 4.71,
2184
+ "learning_rate": 7.6284659050731525e-06,
2185
+ "loss": 1.5493,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 4.73,
2190
+ "learning_rate": 7.57737720196217e-06,
2191
+ "loss": 1.6362,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 4.74,
2196
+ "learning_rate": 7.526355585466432e-06,
2197
+ "loss": 1.6294,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 4.75,
2202
+ "learning_rate": 7.4754024684627405e-06,
2203
+ "loss": 1.6675,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 4.77,
2208
+ "learning_rate": 7.424519261931036e-06,
2209
+ "loss": 1.6519,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 4.78,
2214
+ "learning_rate": 7.373707374915303e-06,
2215
+ "loss": 1.6807,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 4.79,
2220
+ "learning_rate": 7.322968214484583e-06,
2221
+ "loss": 1.6221,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 4.81,
2226
+ "learning_rate": 7.27230318569397e-06,
2227
+ "loss": 1.6523,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 4.82,
2232
+ "learning_rate": 7.221713691545746e-06,
2233
+ "loss": 1.6118,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 4.83,
2238
+ "learning_rate": 7.171201132950502e-06,
2239
+ "loss": 1.6279,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 4.84,
2244
+ "learning_rate": 7.1207669086883366e-06,
2245
+ "loss": 1.6416,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 4.86,
2250
+ "learning_rate": 7.070412415370158e-06,
2251
+ "loss": 1.605,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 4.87,
2256
+ "learning_rate": 7.020139047398966e-06,
2257
+ "loss": 1.627,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 4.88,
2262
+ "learning_rate": 6.969948196931272e-06,
2263
+ "loss": 1.6123,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 4.9,
2268
+ "learning_rate": 6.919841253838537e-06,
2269
+ "loss": 1.6333,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 4.91,
2274
+ "learning_rate": 6.869819605668669e-06,
2275
+ "loss": 1.5981,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 4.92,
2280
+ "learning_rate": 6.819884637607619e-06,
2281
+ "loss": 1.646,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 4.94,
2286
+ "learning_rate": 6.770037732441019e-06,
2287
+ "loss": 1.6641,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 4.95,
2292
+ "learning_rate": 6.720280270515882e-06,
2293
+ "loss": 1.6362,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 4.96,
2298
+ "learning_rate": 6.670613629702391e-06,
2299
+ "loss": 1.6562,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 4.97,
2304
+ "learning_rate": 6.62103918535572e-06,
2305
+ "loss": 1.6772,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 4.99,
2310
+ "learning_rate": 6.5715583102779815e-06,
2311
+ "loss": 1.6729,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 5.0,
2316
+ "learning_rate": 6.522172374680177e-06,
2317
+ "loss": 1.6597,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 5.01,
2322
+ "learning_rate": 6.472882746144282e-06,
2323
+ "loss": 1.6348,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 5.03,
2328
+ "learning_rate": 6.423690789585359e-06,
2329
+ "loss": 1.6108,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 5.04,
2334
+ "learning_rate": 6.374597867213756e-06,
2335
+ "loss": 1.6421,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 5.05,
2340
+ "learning_rate": 6.3256053384974105e-06,
2341
+ "loss": 1.6455,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 5.06,
2346
+ "learning_rate": 6.276714560124166e-06,
2347
+ "loss": 1.6616,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 5.08,
2352
+ "learning_rate": 6.2279268859642396e-06,
2353
+ "loss": 1.6162,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 5.09,
2358
+ "learning_rate": 6.179243667032709e-06,
2359
+ "loss": 1.6646,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 5.1,
2364
+ "learning_rate": 6.130666251452102e-06,
2365
+ "loss": 1.6445,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 5.12,
2370
+ "learning_rate": 6.082195984415069e-06,
2371
+ "loss": 1.6299,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 5.13,
2376
+ "learning_rate": 6.03383420814714e-06,
2377
+ "loss": 1.6221,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 5.14,
2382
+ "learning_rate": 5.9855822618695385e-06,
2383
+ "loss": 1.647,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 5.16,
2388
+ "learning_rate": 5.937441481762112e-06,
2389
+ "loss": 1.6147,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 5.17,
2394
+ "learning_rate": 5.889413200926317e-06,
2395
+ "loss": 1.6025,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 5.18,
2400
+ "learning_rate": 5.841498749348322e-06,
2401
+ "loss": 1.6064,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 5.19,
2406
+ "learning_rate": 5.793699453862161e-06,
2407
+ "loss": 1.6587,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 5.21,
2412
+ "learning_rate": 5.746016638112986e-06,
2413
+ "loss": 1.6255,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 5.22,
2418
+ "learning_rate": 5.698451622520442e-06,
2419
+ "loss": 1.6523,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 5.23,
2424
+ "learning_rate": 5.651005724242072e-06,
2425
+ "loss": 1.6367,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 5.25,
2430
+ "learning_rate": 5.603680257136857e-06,
2431
+ "loss": 1.6006,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 5.26,
2436
+ "learning_rate": 5.556476531728836e-06,
2437
+ "loss": 1.6294,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 5.27,
2442
+ "learning_rate": 5.509395855170798e-06,
2443
+ "loss": 1.6284,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 5.29,
2448
+ "learning_rate": 5.4624395312081125e-06,
2449
+ "loss": 1.6392,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 5.3,
2454
+ "learning_rate": 5.415608860142593e-06,
2455
+ "loss": 1.625,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 5.31,
2460
+ "learning_rate": 5.368905138796523e-06,
2461
+ "loss": 1.6162,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 5.32,
2466
+ "learning_rate": 5.322329660476715e-06,
2467
+ "loss": 1.5752,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 5.34,
2472
+ "learning_rate": 5.275883714938726e-06,
2473
+ "loss": 1.6655,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 5.35,
2478
+ "learning_rate": 5.2295685883511086e-06,
2479
+ "loss": 1.5972,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 5.36,
2484
+ "learning_rate": 5.183385563259819e-06,
2485
+ "loss": 1.6421,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 5.38,
2490
+ "learning_rate": 5.137335918552702e-06,
2491
+ "loss": 1.5869,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 5.39,
2496
+ "learning_rate": 5.091420929424065e-06,
2497
+ "loss": 1.6333,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 5.4,
2502
+ "learning_rate": 5.045641867339361e-06,
2503
+ "loss": 1.6445,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 5.42,
2508
+ "learning_rate": 5.000000000000003e-06,
2509
+ "loss": 1.6597,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 5.43,
2514
+ "learning_rate": 4.954496591308227e-06,
2515
+ "loss": 1.6387,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 5.44,
2520
+ "learning_rate": 4.909132901332122e-06,
2521
+ "loss": 1.6489,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 5.45,
2526
+ "learning_rate": 4.863910186270726e-06,
2527
+ "loss": 1.6318,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 5.47,
2532
+ "learning_rate": 4.818829698419225e-06,
2533
+ "loss": 1.6841,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 5.48,
2538
+ "learning_rate": 4.773892686134301e-06,
2539
+ "loss": 1.666,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 5.49,
2544
+ "learning_rate": 4.729100393799538e-06,
2545
+ "loss": 1.6162,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 5.51,
2550
+ "learning_rate": 4.684454061790987e-06,
2551
+ "loss": 1.5957,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 5.52,
2556
+ "learning_rate": 4.639954926442792e-06,
2557
+ "loss": 1.6201,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 5.53,
2562
+ "learning_rate": 4.5956042200129725e-06,
2563
+ "loss": 1.6533,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 5.55,
2568
+ "learning_rate": 4.551403170649299e-06,
2569
+ "loss": 1.624,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 5.56,
2574
+ "learning_rate": 4.507353002355269e-06,
2575
+ "loss": 1.604,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 5.57,
2580
+ "learning_rate": 4.4634549349562315e-06,
2581
+ "loss": 1.6089,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 5.58,
2586
+ "learning_rate": 4.4197101840656e-06,
2587
+ "loss": 1.5962,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 5.6,
2592
+ "learning_rate": 4.376119961051175e-06,
2593
+ "loss": 1.5962,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 5.61,
2598
+ "learning_rate": 4.33268547300163e-06,
2599
+ "loss": 1.6313,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 5.62,
2604
+ "learning_rate": 4.289407922693053e-06,
2605
+ "loss": 1.6626,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 5.64,
2610
+ "learning_rate": 4.2462885085556635e-06,
2611
+ "loss": 1.5796,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 5.65,
2616
+ "learning_rate": 4.203328424640619e-06,
2617
+ "loss": 1.6836,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 5.66,
2622
+ "learning_rate": 4.1605288605869365e-06,
2623
+ "loss": 1.6675,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 5.68,
2628
+ "learning_rate": 4.117891001588574e-06,
2629
+ "loss": 1.6807,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 5.69,
2634
+ "learning_rate": 4.075416028361584e-06,
2635
+ "loss": 1.6167,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 5.7,
2640
+ "learning_rate": 4.033105117111441e-06,
2641
+ "loss": 1.6851,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 5.71,
2646
+ "learning_rate": 3.9909594395004545e-06,
2647
+ "loss": 1.6191,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 5.73,
2652
+ "learning_rate": 3.948980162615323e-06,
2653
+ "loss": 1.6362,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 5.74,
2658
+ "learning_rate": 3.907168448934836e-06,
2659
+ "loss": 1.5825,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 5.75,
2664
+ "learning_rate": 3.865525456297652e-06,
2665
+ "loss": 1.6182,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 5.77,
2670
+ "learning_rate": 3.824052337870263e-06,
2671
+ "loss": 1.5908,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 5.78,
2676
+ "learning_rate": 3.7827502421150497e-06,
2677
+ "loss": 1.6162,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 5.79,
2682
+ "learning_rate": 3.741620312758469e-06,
2683
+ "loss": 1.6021,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 5.81,
2688
+ "learning_rate": 3.7006636887594095e-06,
2689
+ "loss": 1.6479,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 5.82,
2694
+ "learning_rate": 3.6598815042776135e-06,
2695
+ "loss": 1.6294,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 5.83,
2700
+ "learning_rate": 3.619274888642309e-06,
2701
+ "loss": 1.6914,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 5.84,
2706
+ "learning_rate": 3.578844966320917e-06,
2707
+ "loss": 1.6226,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 5.86,
2712
+ "learning_rate": 3.5385928568879012e-06,
2713
+ "loss": 1.6196,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 5.87,
2718
+ "learning_rate": 3.4985196749937976e-06,
2719
+ "loss": 1.5977,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 5.88,
2724
+ "learning_rate": 3.458626530334316e-06,
2725
+ "loss": 1.5786,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 5.9,
2730
+ "learning_rate": 3.4189145276196244e-06,
2731
+ "loss": 1.6113,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 5.91,
2736
+ "learning_rate": 3.3793847665437674e-06,
2737
+ "loss": 1.6025,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 5.92,
2742
+ "learning_rate": 3.340038341754189e-06,
2743
+ "loss": 1.6191,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 5.94,
2748
+ "learning_rate": 3.300876342821451e-06,
2749
+ "loss": 1.604,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 5.95,
2754
+ "learning_rate": 3.2618998542090263e-06,
2755
+ "loss": 1.6274,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 5.96,
2760
+ "learning_rate": 3.2231099552433e-06,
2761
+ "loss": 1.6543,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 5.97,
2766
+ "learning_rate": 3.1845077200836638e-06,
2767
+ "loss": 1.6265,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 5.99,
2772
+ "learning_rate": 3.1460942176927666e-06,
2773
+ "loss": 1.6123,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 6.0,
2778
+ "learning_rate": 3.107870511806934e-06,
2779
+ "loss": 1.6401,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 6.01,
2784
+ "learning_rate": 3.0698376609066828e-06,
2785
+ "loss": 1.6094,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 6.03,
2790
+ "learning_rate": 3.0319967181874366e-06,
2791
+ "loss": 1.5859,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 6.04,
2796
+ "learning_rate": 2.9943487315303486e-06,
2797
+ "loss": 1.6182,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 6.05,
2802
+ "learning_rate": 2.9568947434732777e-06,
2803
+ "loss": 1.6196,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 6.06,
2808
+ "learning_rate": 2.919635791181934e-06,
2809
+ "loss": 1.6367,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 6.08,
2814
+ "learning_rate": 2.882572906421145e-06,
2815
+ "loss": 1.7124,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 6.09,
2820
+ "learning_rate": 2.8457071155262885e-06,
2821
+ "loss": 1.623,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 6.1,
2826
+ "learning_rate": 2.809039439374878e-06,
2827
+ "loss": 1.5874,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 6.12,
2832
+ "learning_rate": 2.7725708933582785e-06,
2833
+ "loss": 1.6362,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 6.13,
2838
+ "learning_rate": 2.7363024873536093e-06,
2839
+ "loss": 1.6221,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 6.14,
2844
+ "learning_rate": 2.700235225695752e-06,
2845
+ "loss": 1.6416,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 6.16,
2850
+ "learning_rate": 2.6643701071495644e-06,
2851
+ "loss": 1.668,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 6.17,
2856
+ "learning_rate": 2.628708124882212e-06,
2857
+ "loss": 1.5928,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 6.18,
2862
+ "learning_rate": 2.5932502664356553e-06,
2863
+ "loss": 1.6172,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 6.19,
2868
+ "learning_rate": 2.5579975136993253e-06,
2869
+ "loss": 1.6162,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 6.21,
2874
+ "learning_rate": 2.52295084288291e-06,
2875
+ "loss": 1.6636,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 6.22,
2880
+ "learning_rate": 2.4881112244893403e-06,
2881
+ "loss": 1.6748,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 6.23,
2886
+ "learning_rate": 2.453479623287909e-06,
2887
+ "loss": 1.6167,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 6.25,
2892
+ "learning_rate": 2.419056998287547e-06,
2893
+ "loss": 1.6763,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 6.26,
2898
+ "learning_rate": 2.3848443027102706e-06,
2899
+ "loss": 1.6587,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 6.27,
2904
+ "learning_rate": 2.3508424839647994e-06,
2905
+ "loss": 1.6538,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 6.29,
2910
+ "learning_rate": 2.3170524836202936e-06,
2911
+ "loss": 1.5952,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 6.3,
2916
+ "learning_rate": 2.2834752373803094e-06,
2917
+ "loss": 1.6348,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 6.31,
2922
+ "learning_rate": 2.250111675056863e-06,
2923
+ "loss": 1.6074,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 6.32,
2928
+ "learning_rate": 2.216962720544703e-06,
2929
+ "loss": 1.6284,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 6.34,
2934
+ "learning_rate": 2.184029291795705e-06,
2935
+ "loss": 1.6143,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 6.35,
2940
+ "learning_rate": 2.151312300793473e-06,
2941
+ "loss": 1.6323,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 6.36,
2946
+ "learning_rate": 2.118812653528077e-06,
2947
+ "loss": 1.6387,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 6.38,
2952
+ "learning_rate": 2.086531249970952e-06,
2953
+ "loss": 1.6016,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 6.39,
2958
+ "learning_rate": 2.0544689840499988e-06,
2959
+ "loss": 1.6616,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 6.4,
2964
+ "learning_rate": 2.022626743624807e-06,
2965
+ "loss": 1.6211,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 6.42,
2970
+ "learning_rate": 1.991005410462089e-06,
2971
+ "loss": 1.6504,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 6.43,
2976
+ "learning_rate": 1.9596058602112533e-06,
2977
+ "loss": 1.6748,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 6.44,
2982
+ "learning_rate": 1.928428962380148e-06,
2983
+ "loss": 1.6597,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 6.45,
2988
+ "learning_rate": 1.8974755803109968e-06,
2989
+ "loss": 1.6133,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 6.47,
2994
+ "learning_rate": 1.866746571156479e-06,
2995
+ "loss": 1.6294,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 6.48,
3000
+ "learning_rate": 1.8362427858560094e-06,
3001
+ "loss": 1.6074,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 6.49,
3006
+ "learning_rate": 1.8059650691121611e-06,
3007
+ "loss": 1.645,
3008
+ "step": 500
3009
+ },
3010
+ {
3011
+ "epoch": 6.51,
3012
+ "learning_rate": 1.7759142593672707e-06,
3013
+ "loss": 1.5884,
3014
+ "step": 501
3015
+ },
3016
+ {
3017
+ "epoch": 6.52,
3018
+ "learning_rate": 1.74609118878024e-06,
3019
+ "loss": 1.6245,
3020
+ "step": 502
3021
+ },
3022
+ {
3023
+ "epoch": 6.53,
3024
+ "learning_rate": 1.7164966832034668e-06,
3025
+ "loss": 1.6309,
3026
+ "step": 503
3027
+ },
3028
+ {
3029
+ "epoch": 6.55,
3030
+ "learning_rate": 1.6871315621599982e-06,
3031
+ "loss": 1.6035,
3032
+ "step": 504
3033
+ },
3034
+ {
3035
+ "epoch": 6.56,
3036
+ "learning_rate": 1.6579966388208257e-06,
3037
+ "loss": 1.5688,
3038
+ "step": 505
3039
+ },
3040
+ {
3041
+ "epoch": 6.57,
3042
+ "learning_rate": 1.6290927199823604e-06,
3043
+ "loss": 1.5762,
3044
+ "step": 506
3045
+ },
3046
+ {
3047
+ "epoch": 6.58,
3048
+ "learning_rate": 1.6004206060441096e-06,
3049
+ "loss": 1.6323,
3050
+ "step": 507
3051
+ },
3052
+ {
3053
+ "epoch": 6.6,
3054
+ "learning_rate": 1.5719810909864941e-06,
3055
+ "loss": 1.5884,
3056
+ "step": 508
3057
+ },
3058
+ {
3059
+ "epoch": 6.61,
3060
+ "learning_rate": 1.543774962348874e-06,
3061
+ "loss": 1.6382,
3062
+ "step": 509
3063
+ },
3064
+ {
3065
+ "epoch": 6.62,
3066
+ "learning_rate": 1.5158030012077329e-06,
3067
+ "loss": 1.6279,
3068
+ "step": 510
3069
+ },
3070
+ {
3071
+ "epoch": 6.64,
3072
+ "learning_rate": 1.4880659821550547e-06,
3073
+ "loss": 1.6304,
3074
+ "step": 511
3075
+ },
3076
+ {
3077
+ "epoch": 6.65,
3078
+ "learning_rate": 1.4605646732768685e-06,
3079
+ "loss": 1.6289,
3080
+ "step": 512
3081
+ },
3082
+ {
3083
+ "epoch": 6.66,
3084
+ "learning_rate": 1.4332998361319783e-06,
3085
+ "loss": 1.5889,
3086
+ "step": 513
3087
+ },
3088
+ {
3089
+ "epoch": 6.68,
3090
+ "learning_rate": 1.4062722257308803e-06,
3091
+ "loss": 1.6221,
3092
+ "step": 514
3093
+ },
3094
+ {
3095
+ "epoch": 6.69,
3096
+ "learning_rate": 1.3794825905148557e-06,
3097
+ "loss": 1.604,
3098
+ "step": 515
3099
+ },
3100
+ {
3101
+ "epoch": 6.7,
3102
+ "learning_rate": 1.3529316723352303e-06,
3103
+ "loss": 1.6099,
3104
+ "step": 516
3105
+ },
3106
+ {
3107
+ "epoch": 6.71,
3108
+ "learning_rate": 1.3266202064328548e-06,
3109
+ "loss": 1.6045,
3110
+ "step": 517
3111
+ },
3112
+ {
3113
+ "epoch": 6.73,
3114
+ "learning_rate": 1.3005489214177213e-06,
3115
+ "loss": 1.6289,
3116
+ "step": 518
3117
+ },
3118
+ {
3119
+ "epoch": 6.74,
3120
+ "learning_rate": 1.2747185392488048e-06,
3121
+ "loss": 1.6519,
3122
+ "step": 519
3123
+ },
3124
+ {
3125
+ "epoch": 6.75,
3126
+ "learning_rate": 1.249129775214064e-06,
3127
+ "loss": 1.6338,
3128
+ "step": 520
3129
+ },
3130
+ {
3131
+ "epoch": 6.77,
3132
+ "learning_rate": 1.2237833379106257e-06,
3133
+ "loss": 1.6196,
3134
+ "step": 521
3135
+ },
3136
+ {
3137
+ "epoch": 6.78,
3138
+ "learning_rate": 1.1986799292251816e-06,
3139
+ "loss": 1.6104,
3140
+ "step": 522
3141
+ },
3142
+ {
3143
+ "epoch": 6.79,
3144
+ "learning_rate": 1.1738202443145307e-06,
3145
+ "loss": 1.6309,
3146
+ "step": 523
3147
+ },
3148
+ {
3149
+ "epoch": 6.81,
3150
+ "learning_rate": 1.1492049715863464e-06,
3151
+ "loss": 1.5845,
3152
+ "step": 524
3153
+ },
3154
+ {
3155
+ "epoch": 6.82,
3156
+ "learning_rate": 1.1248347926801029e-06,
3157
+ "loss": 1.582,
3158
+ "step": 525
3159
+ },
3160
+ {
3161
+ "epoch": 6.83,
3162
+ "learning_rate": 1.100710382448198e-06,
3163
+ "loss": 1.6553,
3164
+ "step": 526
3165
+ },
3166
+ {
3167
+ "epoch": 6.84,
3168
+ "learning_rate": 1.0768324089372816e-06,
3169
+ "loss": 1.5771,
3170
+ "step": 527
3171
+ },
3172
+ {
3173
+ "epoch": 6.86,
3174
+ "learning_rate": 1.053201533369731e-06,
3175
+ "loss": 1.6611,
3176
+ "step": 528
3177
+ },
3178
+ {
3179
+ "epoch": 6.87,
3180
+ "learning_rate": 1.029818410125365e-06,
3181
+ "loss": 1.6128,
3182
+ "step": 529
3183
+ },
3184
+ {
3185
+ "epoch": 6.88,
3186
+ "learning_rate": 1.0066836867233087e-06,
3187
+ "loss": 1.5957,
3188
+ "step": 530
3189
+ },
3190
+ {
3191
+ "epoch": 6.9,
3192
+ "learning_rate": 9.837980038040607e-07,
3193
+ "loss": 1.6299,
3194
+ "step": 531
3195
+ },
3196
+ {
3197
+ "epoch": 6.91,
3198
+ "learning_rate": 9.611619951117657e-07,
3199
+ "loss": 1.6147,
3200
+ "step": 532
3201
+ },
3202
+ {
3203
+ "epoch": 6.92,
3204
+ "learning_rate": 9.387762874766515e-07,
3205
+ "loss": 1.5864,
3206
+ "step": 533
3207
+ },
3208
+ {
3209
+ "epoch": 6.94,
3210
+ "learning_rate": 9.166415007976803e-07,
3211
+ "loss": 1.6245,
3212
+ "step": 534
3213
+ },
3214
+ {
3215
+ "epoch": 6.95,
3216
+ "learning_rate": 8.94758248025378e-07,
3217
+ "loss": 1.5781,
3218
+ "step": 535
3219
+ },
3220
+ {
3221
+ "epoch": 6.96,
3222
+ "learning_rate": 8.7312713514486e-07,
3223
+ "loss": 1.5845,
3224
+ "step": 536
3225
+ },
3226
+ {
3227
+ "epoch": 6.97,
3228
+ "learning_rate": 8.517487611590558e-07,
3229
+ "loss": 1.624,
3230
+ "step": 537
3231
+ },
3232
+ {
3233
+ "epoch": 6.99,
3234
+ "learning_rate": 8.306237180721121e-07,
3235
+ "loss": 1.5811,
3236
+ "step": 538
3237
+ },
3238
+ {
3239
+ "epoch": 7.0,
3240
+ "learning_rate": 8.097525908730108e-07,
3241
+ "loss": 1.5898,
3242
+ "step": 539
3243
+ },
3244
+ {
3245
+ "epoch": 7.01,
3246
+ "learning_rate": 7.891359575193613e-07,
3247
+ "loss": 1.5542,
3248
+ "step": 540
3249
+ },
3250
+ {
3251
+ "epoch": 7.03,
3252
+ "learning_rate": 7.687743889213939e-07,
3253
+ "loss": 1.6382,
3254
+ "step": 541
3255
+ },
3256
+ {
3257
+ "epoch": 7.04,
3258
+ "learning_rate": 7.486684489261609e-07,
3259
+ "loss": 1.6597,
3260
+ "step": 542
3261
+ },
3262
+ {
3263
+ "epoch": 7.05,
3264
+ "learning_rate": 7.288186943019171e-07,
3265
+ "loss": 1.5918,
3266
+ "step": 543
3267
+ },
3268
+ {
3269
+ "epoch": 7.06,
3270
+ "learning_rate": 7.092256747226944e-07,
3271
+ "loss": 1.6226,
3272
+ "step": 544
3273
+ },
3274
+ {
3275
+ "epoch": 7.08,
3276
+ "learning_rate": 6.89889932753095e-07,
3277
+ "loss": 1.563,
3278
+ "step": 545
3279
+ },
3280
+ {
3281
+ "epoch": 7.09,
3282
+ "learning_rate": 6.708120038332533e-07,
3283
+ "loss": 1.6348,
3284
+ "step": 546
3285
+ },
3286
+ {
3287
+ "epoch": 7.1,
3288
+ "learning_rate": 6.519924162640168e-07,
3289
+ "loss": 1.6089,
3290
+ "step": 547
3291
+ },
3292
+ {
3293
+ "epoch": 7.12,
3294
+ "learning_rate": 6.334316911923155e-07,
3295
+ "loss": 1.6143,
3296
+ "step": 548
3297
+ },
3298
+ {
3299
+ "epoch": 7.13,
3300
+ "learning_rate": 6.151303425967259e-07,
3301
+ "loss": 1.6396,
3302
+ "step": 549
3303
+ },
3304
+ {
3305
+ "epoch": 7.14,
3306
+ "learning_rate": 5.970888772732453e-07,
3307
+ "loss": 1.6387,
3308
+ "step": 550
3309
+ },
3310
+ {
3311
+ "epoch": 7.16,
3312
+ "learning_rate": 5.793077948212478e-07,
3313
+ "loss": 1.5835,
3314
+ "step": 551
3315
+ },
3316
+ {
3317
+ "epoch": 7.17,
3318
+ "learning_rate": 5.617875876296641e-07,
3319
+ "loss": 1.6489,
3320
+ "step": 552
3321
+ },
3322
+ {
3323
+ "epoch": 7.18,
3324
+ "learning_rate": 5.445287408633304e-07,
3325
+ "loss": 1.6318,
3326
+ "step": 553
3327
+ },
3328
+ {
3329
+ "epoch": 7.19,
3330
+ "learning_rate": 5.27531732449561e-07,
3331
+ "loss": 1.645,
3332
+ "step": 554
3333
+ },
3334
+ {
3335
+ "epoch": 7.21,
3336
+ "learning_rate": 5.107970330649204e-07,
3337
+ "loss": 1.5996,
3338
+ "step": 555
3339
+ },
3340
+ {
3341
+ "epoch": 7.22,
3342
+ "learning_rate": 4.943251061221721e-07,
3343
+ "loss": 1.5962,
3344
+ "step": 556
3345
+ },
3346
+ {
3347
+ "epoch": 7.23,
3348
+ "learning_rate": 4.78116407757464e-07,
3349
+ "loss": 1.6211,
3350
+ "step": 557
3351
+ },
3352
+ {
3353
+ "epoch": 7.25,
3354
+ "learning_rate": 4.6217138681769026e-07,
3355
+ "loss": 1.6011,
3356
+ "step": 558
3357
+ },
3358
+ {
3359
+ "epoch": 7.26,
3360
+ "learning_rate": 4.464904848480522e-07,
3361
+ "loss": 1.6392,
3362
+ "step": 559
3363
+ },
3364
+ {
3365
+ "epoch": 7.27,
3366
+ "learning_rate": 4.310741360798498e-07,
3367
+ "loss": 1.6265,
3368
+ "step": 560
3369
+ },
3370
+ {
3371
+ "epoch": 7.29,
3372
+ "learning_rate": 4.1592276741844075e-07,
3373
+ "loss": 1.6255,
3374
+ "step": 561
3375
+ },
3376
+ {
3377
+ "epoch": 7.3,
3378
+ "learning_rate": 4.0103679843142895e-07,
3379
+ "loss": 1.6196,
3380
+ "step": 562
3381
+ },
3382
+ {
3383
+ "epoch": 7.31,
3384
+ "learning_rate": 3.864166413370429e-07,
3385
+ "loss": 1.6201,
3386
+ "step": 563
3387
+ },
3388
+ {
3389
+ "epoch": 7.32,
3390
+ "learning_rate": 3.720627009927158e-07,
3391
+ "loss": 1.6396,
3392
+ "step": 564
3393
+ },
3394
+ {
3395
+ "epoch": 7.34,
3396
+ "learning_rate": 3.5797537488388326e-07,
3397
+ "loss": 1.6553,
3398
+ "step": 565
3399
+ },
3400
+ {
3401
+ "epoch": 7.35,
3402
+ "learning_rate": 3.441550531129667e-07,
3403
+ "loss": 1.6431,
3404
+ "step": 566
3405
+ },
3406
+ {
3407
+ "epoch": 7.36,
3408
+ "learning_rate": 3.3060211838858104e-07,
3409
+ "loss": 1.6362,
3410
+ "step": 567
3411
+ },
3412
+ {
3413
+ "epoch": 7.38,
3414
+ "learning_rate": 3.1731694601492834e-07,
3415
+ "loss": 1.5654,
3416
+ "step": 568
3417
+ },
3418
+ {
3419
+ "epoch": 7.39,
3420
+ "learning_rate": 3.042999038814076e-07,
3421
+ "loss": 1.6074,
3422
+ "step": 569
3423
+ },
3424
+ {
3425
+ "epoch": 7.4,
3426
+ "learning_rate": 2.915513524524294e-07,
3427
+ "loss": 1.6094,
3428
+ "step": 570
3429
+ },
3430
+ {
3431
+ "epoch": 7.42,
3432
+ "learning_rate": 2.790716447574304e-07,
3433
+ "loss": 1.6758,
3434
+ "step": 571
3435
+ },
3436
+ {
3437
+ "epoch": 7.43,
3438
+ "learning_rate": 2.668611263811016e-07,
3439
+ "loss": 1.6313,
3440
+ "step": 572
3441
+ },
3442
+ {
3443
+ "epoch": 7.44,
3444
+ "learning_rate": 2.5492013545381666e-07,
3445
+ "loss": 1.5835,
3446
+ "step": 573
3447
+ },
3448
+ {
3449
+ "epoch": 7.45,
3450
+ "learning_rate": 2.4324900264226405e-07,
3451
+ "loss": 1.5972,
3452
+ "step": 574
3453
+ },
3454
+ {
3455
+ "epoch": 7.47,
3456
+ "learning_rate": 2.3184805114029872e-07,
3457
+ "loss": 1.6689,
3458
+ "step": 575
3459
+ },
3460
+ {
3461
+ "epoch": 7.48,
3462
+ "learning_rate": 2.2071759665998282e-07,
3463
+ "loss": 1.6304,
3464
+ "step": 576
3465
+ },
3466
+ {
3467
+ "epoch": 7.49,
3468
+ "learning_rate": 2.098579474228546e-07,
3469
+ "loss": 1.6035,
3470
+ "step": 577
3471
+ },
3472
+ {
3473
+ "epoch": 7.51,
3474
+ "learning_rate": 1.9926940415138206e-07,
3475
+ "loss": 1.5952,
3476
+ "step": 578
3477
+ },
3478
+ {
3479
+ "epoch": 7.52,
3480
+ "learning_rate": 1.8895226006064084e-07,
3481
+ "loss": 1.584,
3482
+ "step": 579
3483
+ },
3484
+ {
3485
+ "epoch": 7.53,
3486
+ "learning_rate": 1.7890680085019597e-07,
3487
+ "loss": 1.6064,
3488
+ "step": 580
3489
+ },
3490
+ {
3491
+ "epoch": 7.55,
3492
+ "learning_rate": 1.6913330469618628e-07,
3493
+ "loss": 1.6235,
3494
+ "step": 581
3495
+ },
3496
+ {
3497
+ "epoch": 7.56,
3498
+ "learning_rate": 1.5963204224362261e-07,
3499
+ "loss": 1.6294,
3500
+ "step": 582
3501
+ },
3502
+ {
3503
+ "epoch": 7.57,
3504
+ "learning_rate": 1.504032765988961e-07,
3505
+ "loss": 1.6055,
3506
+ "step": 583
3507
+ },
3508
+ {
3509
+ "epoch": 7.58,
3510
+ "learning_rate": 1.4144726332248726e-07,
3511
+ "loss": 1.6353,
3512
+ "step": 584
3513
+ },
3514
+ {
3515
+ "epoch": 7.6,
3516
+ "learning_rate": 1.327642504218951e-07,
3517
+ "loss": 1.6108,
3518
+ "step": 585
3519
+ },
3520
+ {
3521
+ "epoch": 7.61,
3522
+ "learning_rate": 1.2435447834476254e-07,
3523
+ "loss": 1.6201,
3524
+ "step": 586
3525
+ },
3526
+ {
3527
+ "epoch": 7.62,
3528
+ "learning_rate": 1.1621817997222507e-07,
3529
+ "loss": 1.6128,
3530
+ "step": 587
3531
+ },
3532
+ {
3533
+ "epoch": 7.64,
3534
+ "learning_rate": 1.0835558061245587e-07,
3535
+ "loss": 1.6196,
3536
+ "step": 588
3537
+ },
3538
+ {
3539
+ "epoch": 7.65,
3540
+ "learning_rate": 1.0076689799442874e-07,
3541
+ "loss": 1.6621,
3542
+ "step": 589
3543
+ },
3544
+ {
3545
+ "epoch": 7.66,
3546
+ "learning_rate": 9.34523422618916e-08,
3547
+ "loss": 1.6216,
3548
+ "step": 590
3549
+ },
3550
+ {
3551
+ "epoch": 7.68,
3552
+ "learning_rate": 8.641211596754129e-08,
3553
+ "loss": 1.6289,
3554
+ "step": 591
3555
+ },
3556
+ {
3557
+ "epoch": 7.69,
3558
+ "learning_rate": 7.964641406742135e-08,
3559
+ "loss": 1.6279,
3560
+ "step": 592
3561
+ },
3562
+ {
3563
+ "epoch": 7.7,
3564
+ "learning_rate": 7.315542391551966e-08,
3565
+ "loss": 1.6187,
3566
+ "step": 593
3567
+ },
3568
+ {
3569
+ "epoch": 7.71,
3570
+ "learning_rate": 6.693932525857927e-08,
3571
+ "loss": 1.6445,
3572
+ "step": 594
3573
+ },
3574
+ {
3575
+ "epoch": 7.73,
3576
+ "learning_rate": 6.099829023112236e-08,
3577
+ "loss": 1.6226,
3578
+ "step": 595
3579
+ },
3580
+ {
3581
+ "epoch": 7.74,
3582
+ "learning_rate": 5.533248335068409e-08,
3583
+ "loss": 1.6025,
3584
+ "step": 596
3585
+ },
3586
+ {
3587
+ "epoch": 7.75,
3588
+ "learning_rate": 4.994206151325509e-08,
3589
+ "loss": 1.5981,
3590
+ "step": 597
3591
+ },
3592
+ {
3593
+ "epoch": 7.77,
3594
+ "learning_rate": 4.482717398894165e-08,
3595
+ "loss": 1.6479,
3596
+ "step": 598
3597
+ },
3598
+ {
3599
+ "epoch": 7.78,
3600
+ "learning_rate": 3.998796241782232e-08,
3601
+ "loss": 1.6494,
3602
+ "step": 599
3603
+ },
3604
+ {
3605
+ "epoch": 7.79,
3606
+ "learning_rate": 3.5424560806036625e-08,
3607
+ "loss": 1.6328,
3608
+ "step": 600
3609
+ },
3610
+ {
3611
+ "epoch": 7.81,
3612
+ "learning_rate": 3.1137095522068006e-08,
3613
+ "loss": 1.5732,
3614
+ "step": 601
3615
+ },
3616
+ {
3617
+ "epoch": 7.82,
3618
+ "learning_rate": 2.7125685293245552e-08,
3619
+ "loss": 1.6196,
3620
+ "step": 602
3621
+ },
3622
+ {
3623
+ "epoch": 7.83,
3624
+ "learning_rate": 2.3390441202455484e-08,
3625
+ "loss": 1.5894,
3626
+ "step": 603
3627
+ },
3628
+ {
3629
+ "epoch": 7.84,
3630
+ "learning_rate": 1.993146668506585e-08,
3631
+ "loss": 1.6172,
3632
+ "step": 604
3633
+ },
3634
+ {
3635
+ "epoch": 7.86,
3636
+ "learning_rate": 1.6748857526066588e-08,
3637
+ "loss": 1.604,
3638
+ "step": 605
3639
+ },
3640
+ {
3641
+ "epoch": 7.87,
3642
+ "learning_rate": 1.3842701857406104e-08,
3643
+ "loss": 1.6172,
3644
+ "step": 606
3645
+ },
3646
+ {
3647
+ "epoch": 7.88,
3648
+ "learning_rate": 1.1213080155564327e-08,
3649
+ "loss": 1.6377,
3650
+ "step": 607
3651
+ },
3652
+ {
3653
+ "epoch": 7.9,
3654
+ "learning_rate": 8.860065239311155e-09,
3655
+ "loss": 1.6064,
3656
+ "step": 608
3657
+ },
3658
+ {
3659
+ "epoch": 7.91,
3660
+ "learning_rate": 6.783722267701409e-09,
3661
+ "loss": 1.6211,
3662
+ "step": 609
3663
+ },
3664
+ {
3665
+ "epoch": 7.92,
3666
+ "learning_rate": 4.984108738261828e-09,
3667
+ "loss": 1.6274,
3668
+ "step": 610
3669
+ },
3670
+ {
3671
+ "epoch": 7.94,
3672
+ "learning_rate": 3.4612744854045645e-09,
3673
+ "loss": 1.6328,
3674
+ "step": 611
3675
+ },
3676
+ {
3677
+ "epoch": 7.95,
3678
+ "learning_rate": 2.215261679042735e-09,
3679
+ "loss": 1.583,
3680
+ "step": 612
3681
+ },
3682
+ {
3683
+ "epoch": 7.96,
3684
+ "learning_rate": 1.246104823426908e-09,
3685
+ "loss": 1.6318,
3686
+ "step": 613
3687
+ },
3688
+ {
3689
+ "epoch": 7.97,
3690
+ "learning_rate": 5.538307561858691e-10,
3691
+ "loss": 1.6265,
3692
+ "step": 614
3693
+ },
3694
+ {
3695
+ "epoch": 7.99,
3696
+ "learning_rate": 1.3845864758610384e-10,
3697
+ "loss": 1.605,
3698
+ "step": 615
3699
+ },
3700
+ {
3701
+ "epoch": 8.0,
3702
+ "learning_rate": 0.0,
3703
+ "loss": 1.6025,
3704
+ "step": 616
3705
+ },
3706
+ {
3707
+ "epoch": 8.0,
3708
+ "step": 616,
3709
+ "total_flos": 1.5114021399418634e+18,
3710
+ "train_loss": 2.115578391335227,
3711
+ "train_runtime": 34978.7912,
3712
+ "train_samples_per_second": 2.252,
3713
+ "train_steps_per_second": 0.018
3714
+ }
3715
+ ],
3716
+ "max_steps": 616,
3717
+ "num_train_epochs": 8,
3718
+ "total_flos": 1.5114021399418634e+18,
3719
+ "trial_name": null,
3720
+ "trial_params": null
3721
+ }