First attempt at a BitsNBytes QLoRA Falcon model.
Browse files- README.md +32 -0
- adapter_config.json +22 -0
- adapter_model.bin +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +17 -0
- tokenizer.json +0 -0
- tokenizer_config.json +7 -0
- trainer_state.json +2416 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: peft
|
3 |
+
---
|
4 |
+
## Training procedure
|
5 |
+
|
6 |
+
|
7 |
+
The following `bitsandbytes` quantization config was used during training:
|
8 |
+
- load_in_8bit: False
|
9 |
+
- load_in_4bit: True
|
10 |
+
- llm_int8_threshold: 6.0
|
11 |
+
- llm_int8_skip_modules: None
|
12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
13 |
+
- llm_int8_has_fp16_weight: False
|
14 |
+
- bnb_4bit_quant_type: nf4
|
15 |
+
- bnb_4bit_use_double_quant: False
|
16 |
+
- bnb_4bit_compute_dtype: float16
|
17 |
+
|
18 |
+
The following `bitsandbytes` quantization config was used during training:
|
19 |
+
- load_in_8bit: False
|
20 |
+
- load_in_4bit: True
|
21 |
+
- llm_int8_threshold: 6.0
|
22 |
+
- llm_int8_skip_modules: None
|
23 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
24 |
+
- llm_int8_has_fp16_weight: False
|
25 |
+
- bnb_4bit_quant_type: nf4
|
26 |
+
- bnb_4bit_use_double_quant: False
|
27 |
+
- bnb_4bit_compute_dtype: float16
|
28 |
+
### Framework versions
|
29 |
+
|
30 |
+
- PEFT 0.4.0.dev0
|
31 |
+
|
32 |
+
- PEFT 0.4.0.dev0
|
adapter_config.json
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"base_model_name_or_path": "ybelkada/falcon-7b-sharded-bf16",
|
3 |
+
"bias": "none",
|
4 |
+
"fan_in_fan_out": false,
|
5 |
+
"inference_mode": true,
|
6 |
+
"init_lora_weights": true,
|
7 |
+
"layers_pattern": null,
|
8 |
+
"layers_to_transform": null,
|
9 |
+
"lora_alpha": 16,
|
10 |
+
"lora_dropout": 0.1,
|
11 |
+
"modules_to_save": null,
|
12 |
+
"peft_type": "LORA",
|
13 |
+
"r": 64,
|
14 |
+
"revision": null,
|
15 |
+
"target_modules": [
|
16 |
+
"query_key_value",
|
17 |
+
"dense",
|
18 |
+
"dense_h_to_4h",
|
19 |
+
"dense_4h_to_h"
|
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:bfbbdefb5e61cb996c49acd3e9b78ebafb098c0031917f3d30e808eb641365ad
|
3 |
+
size 522284877
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:996cd5987535c546ff321c65d2d461a04d8ba641f2c336e628579a0ea6ef2a4d
|
3 |
+
size 1044539909
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:891a516ef4f887c1f8e466f091ab13298a6ac24bb2f9d86c52910f6da4345663
|
3 |
+
size 14575
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:593f43b0c83adb27a2db37a6418c2ef12a213bbc2a02f2dc881de6846a69a931
|
3 |
+
size 627
|
special_tokens_map.json
ADDED
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
">>TITLE<<",
|
4 |
+
">>ABSTRACT<<",
|
5 |
+
">>INTRODUCTION<<",
|
6 |
+
">>SUMMARY<<",
|
7 |
+
">>COMMENT<<",
|
8 |
+
">>ANSWER<<",
|
9 |
+
">>QUESTION<<",
|
10 |
+
">>DOMAIN<<",
|
11 |
+
">>PREFIX<<",
|
12 |
+
">>SUFFIX<<",
|
13 |
+
">>MIDDLE<<"
|
14 |
+
],
|
15 |
+
"eos_token": "<|endoftext|>",
|
16 |
+
"pad_token": "<|endoftext|>"
|
17 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"clean_up_tokenization_spaces": true,
|
4 |
+
"eos_token": "<|endoftext|>",
|
5 |
+
"model_max_length": 2048,
|
6 |
+
"tokenizer_class": "PreTrainedTokenizerFast"
|
7 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,2416 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.12909680646775,
|
5 |
+
"global_step": 4000,
|
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": 0.0002,
|
13 |
+
"loss": 3.0243,
|
14 |
+
"step": 10
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.0,
|
18 |
+
"learning_rate": 0.0002,
|
19 |
+
"loss": 2.915,
|
20 |
+
"step": 20
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.0,
|
24 |
+
"learning_rate": 0.0002,
|
25 |
+
"loss": 2.9116,
|
26 |
+
"step": 30
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.0,
|
30 |
+
"learning_rate": 0.0002,
|
31 |
+
"loss": 2.8826,
|
32 |
+
"step": 40
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.0,
|
36 |
+
"learning_rate": 0.0002,
|
37 |
+
"loss": 2.8696,
|
38 |
+
"step": 50
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.0,
|
42 |
+
"learning_rate": 0.0002,
|
43 |
+
"loss": 2.8775,
|
44 |
+
"step": 60
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.0,
|
48 |
+
"learning_rate": 0.0002,
|
49 |
+
"loss": 2.87,
|
50 |
+
"step": 70
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.0,
|
54 |
+
"learning_rate": 0.0002,
|
55 |
+
"loss": 2.8411,
|
56 |
+
"step": 80
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.0,
|
60 |
+
"learning_rate": 0.0002,
|
61 |
+
"loss": 2.8488,
|
62 |
+
"step": 90
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.0,
|
66 |
+
"learning_rate": 0.0002,
|
67 |
+
"loss": 2.8009,
|
68 |
+
"step": 100
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.0,
|
72 |
+
"learning_rate": 0.0002,
|
73 |
+
"loss": 2.8595,
|
74 |
+
"step": 110
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.0,
|
78 |
+
"learning_rate": 0.0002,
|
79 |
+
"loss": 2.8559,
|
80 |
+
"step": 120
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.0,
|
84 |
+
"learning_rate": 0.0002,
|
85 |
+
"loss": 2.8263,
|
86 |
+
"step": 130
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.0,
|
90 |
+
"learning_rate": 0.0002,
|
91 |
+
"loss": 2.8642,
|
92 |
+
"step": 140
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.0,
|
96 |
+
"learning_rate": 0.0002,
|
97 |
+
"loss": 2.8383,
|
98 |
+
"step": 150
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.01,
|
102 |
+
"learning_rate": 0.0002,
|
103 |
+
"loss": 2.8285,
|
104 |
+
"step": 160
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.01,
|
108 |
+
"learning_rate": 0.0002,
|
109 |
+
"loss": 2.8655,
|
110 |
+
"step": 170
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.01,
|
114 |
+
"learning_rate": 0.0002,
|
115 |
+
"loss": 2.8234,
|
116 |
+
"step": 180
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.01,
|
120 |
+
"learning_rate": 0.0002,
|
121 |
+
"loss": 2.8558,
|
122 |
+
"step": 190
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.01,
|
126 |
+
"learning_rate": 0.0002,
|
127 |
+
"loss": 2.8231,
|
128 |
+
"step": 200
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.01,
|
132 |
+
"learning_rate": 0.0002,
|
133 |
+
"loss": 2.8069,
|
134 |
+
"step": 210
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 0.01,
|
138 |
+
"learning_rate": 0.0002,
|
139 |
+
"loss": 2.8212,
|
140 |
+
"step": 220
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 0.01,
|
144 |
+
"learning_rate": 0.0002,
|
145 |
+
"loss": 2.8382,
|
146 |
+
"step": 230
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 0.01,
|
150 |
+
"learning_rate": 0.0002,
|
151 |
+
"loss": 2.8344,
|
152 |
+
"step": 240
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 0.01,
|
156 |
+
"learning_rate": 0.0002,
|
157 |
+
"loss": 2.8534,
|
158 |
+
"step": 250
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 0.01,
|
162 |
+
"learning_rate": 0.0002,
|
163 |
+
"loss": 2.8426,
|
164 |
+
"step": 260
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 0.01,
|
168 |
+
"learning_rate": 0.0002,
|
169 |
+
"loss": 2.7892,
|
170 |
+
"step": 270
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.01,
|
174 |
+
"learning_rate": 0.0002,
|
175 |
+
"loss": 2.7954,
|
176 |
+
"step": 280
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 0.01,
|
180 |
+
"learning_rate": 0.0002,
|
181 |
+
"loss": 2.8582,
|
182 |
+
"step": 290
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 0.01,
|
186 |
+
"learning_rate": 0.0002,
|
187 |
+
"loss": 2.8365,
|
188 |
+
"step": 300
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 0.01,
|
192 |
+
"learning_rate": 0.0002,
|
193 |
+
"loss": 2.7891,
|
194 |
+
"step": 310
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 0.01,
|
198 |
+
"learning_rate": 0.0002,
|
199 |
+
"loss": 2.8129,
|
200 |
+
"step": 320
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 0.01,
|
204 |
+
"learning_rate": 0.0002,
|
205 |
+
"loss": 2.8047,
|
206 |
+
"step": 330
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 0.01,
|
210 |
+
"learning_rate": 0.0002,
|
211 |
+
"loss": 2.8331,
|
212 |
+
"step": 340
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.01,
|
216 |
+
"learning_rate": 0.0002,
|
217 |
+
"loss": 2.851,
|
218 |
+
"step": 350
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 0.01,
|
222 |
+
"learning_rate": 0.0002,
|
223 |
+
"loss": 2.829,
|
224 |
+
"step": 360
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 0.01,
|
228 |
+
"learning_rate": 0.0002,
|
229 |
+
"loss": 2.8266,
|
230 |
+
"step": 370
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 0.01,
|
234 |
+
"learning_rate": 0.0002,
|
235 |
+
"loss": 2.8251,
|
236 |
+
"step": 380
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.01,
|
240 |
+
"learning_rate": 0.0002,
|
241 |
+
"loss": 2.8473,
|
242 |
+
"step": 390
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 0.01,
|
246 |
+
"learning_rate": 0.0002,
|
247 |
+
"loss": 2.8351,
|
248 |
+
"step": 400
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 0.01,
|
252 |
+
"learning_rate": 0.0002,
|
253 |
+
"loss": 2.8375,
|
254 |
+
"step": 410
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.01,
|
258 |
+
"learning_rate": 0.0002,
|
259 |
+
"loss": 2.8011,
|
260 |
+
"step": 420
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 0.01,
|
264 |
+
"learning_rate": 0.0002,
|
265 |
+
"loss": 2.8072,
|
266 |
+
"step": 430
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.01,
|
270 |
+
"learning_rate": 0.0002,
|
271 |
+
"loss": 2.8397,
|
272 |
+
"step": 440
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 0.01,
|
276 |
+
"learning_rate": 0.0002,
|
277 |
+
"loss": 2.7985,
|
278 |
+
"step": 450
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 0.01,
|
282 |
+
"learning_rate": 0.0002,
|
283 |
+
"loss": 2.8076,
|
284 |
+
"step": 460
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 0.02,
|
288 |
+
"learning_rate": 0.0002,
|
289 |
+
"loss": 2.7993,
|
290 |
+
"step": 470
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 0.02,
|
294 |
+
"learning_rate": 0.0002,
|
295 |
+
"loss": 2.8001,
|
296 |
+
"step": 480
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.02,
|
300 |
+
"learning_rate": 0.0002,
|
301 |
+
"loss": 2.8046,
|
302 |
+
"step": 490
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 0.02,
|
306 |
+
"learning_rate": 0.0002,
|
307 |
+
"loss": 2.8017,
|
308 |
+
"step": 500
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 0.02,
|
312 |
+
"learning_rate": 0.0002,
|
313 |
+
"loss": 2.8004,
|
314 |
+
"step": 510
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 0.02,
|
318 |
+
"learning_rate": 0.0002,
|
319 |
+
"loss": 2.8409,
|
320 |
+
"step": 520
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 0.02,
|
324 |
+
"learning_rate": 0.0002,
|
325 |
+
"loss": 2.8321,
|
326 |
+
"step": 530
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 0.02,
|
330 |
+
"learning_rate": 0.0002,
|
331 |
+
"loss": 2.818,
|
332 |
+
"step": 540
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 0.02,
|
336 |
+
"learning_rate": 0.0002,
|
337 |
+
"loss": 2.8402,
|
338 |
+
"step": 550
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.02,
|
342 |
+
"learning_rate": 0.0002,
|
343 |
+
"loss": 2.8167,
|
344 |
+
"step": 560
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 0.02,
|
348 |
+
"learning_rate": 0.0002,
|
349 |
+
"loss": 2.7762,
|
350 |
+
"step": 570
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 0.02,
|
354 |
+
"learning_rate": 0.0002,
|
355 |
+
"loss": 2.7975,
|
356 |
+
"step": 580
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 0.02,
|
360 |
+
"learning_rate": 0.0002,
|
361 |
+
"loss": 2.8285,
|
362 |
+
"step": 590
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 0.02,
|
366 |
+
"learning_rate": 0.0002,
|
367 |
+
"loss": 2.8029,
|
368 |
+
"step": 600
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 0.02,
|
372 |
+
"learning_rate": 0.0002,
|
373 |
+
"loss": 2.8491,
|
374 |
+
"step": 610
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 0.02,
|
378 |
+
"learning_rate": 0.0002,
|
379 |
+
"loss": 2.7931,
|
380 |
+
"step": 620
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.02,
|
384 |
+
"learning_rate": 0.0002,
|
385 |
+
"loss": 2.8352,
|
386 |
+
"step": 630
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 0.02,
|
390 |
+
"learning_rate": 0.0002,
|
391 |
+
"loss": 2.7869,
|
392 |
+
"step": 640
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 0.02,
|
396 |
+
"learning_rate": 0.0002,
|
397 |
+
"loss": 2.7956,
|
398 |
+
"step": 650
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 0.02,
|
402 |
+
"learning_rate": 0.0002,
|
403 |
+
"loss": 2.7798,
|
404 |
+
"step": 660
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 0.02,
|
408 |
+
"learning_rate": 0.0002,
|
409 |
+
"loss": 2.8087,
|
410 |
+
"step": 670
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 0.02,
|
414 |
+
"learning_rate": 0.0002,
|
415 |
+
"loss": 2.8393,
|
416 |
+
"step": 680
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 0.02,
|
420 |
+
"learning_rate": 0.0002,
|
421 |
+
"loss": 2.7998,
|
422 |
+
"step": 690
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.02,
|
426 |
+
"learning_rate": 0.0002,
|
427 |
+
"loss": 2.835,
|
428 |
+
"step": 700
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 0.02,
|
432 |
+
"learning_rate": 0.0002,
|
433 |
+
"loss": 2.8198,
|
434 |
+
"step": 710
|
435 |
+
},
|
436 |
+
{
|
437 |
+
"epoch": 0.02,
|
438 |
+
"learning_rate": 0.0002,
|
439 |
+
"loss": 2.8249,
|
440 |
+
"step": 720
|
441 |
+
},
|
442 |
+
{
|
443 |
+
"epoch": 0.02,
|
444 |
+
"learning_rate": 0.0002,
|
445 |
+
"loss": 2.7963,
|
446 |
+
"step": 730
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 0.02,
|
450 |
+
"learning_rate": 0.0002,
|
451 |
+
"loss": 2.7589,
|
452 |
+
"step": 740
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 0.02,
|
456 |
+
"learning_rate": 0.0002,
|
457 |
+
"loss": 2.8159,
|
458 |
+
"step": 750
|
459 |
+
},
|
460 |
+
{
|
461 |
+
"epoch": 0.02,
|
462 |
+
"learning_rate": 0.0002,
|
463 |
+
"loss": 2.7879,
|
464 |
+
"step": 760
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 0.02,
|
468 |
+
"learning_rate": 0.0002,
|
469 |
+
"loss": 2.7873,
|
470 |
+
"step": 770
|
471 |
+
},
|
472 |
+
{
|
473 |
+
"epoch": 0.03,
|
474 |
+
"learning_rate": 0.0002,
|
475 |
+
"loss": 2.8207,
|
476 |
+
"step": 780
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 0.03,
|
480 |
+
"learning_rate": 0.0002,
|
481 |
+
"loss": 2.8378,
|
482 |
+
"step": 790
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 0.03,
|
486 |
+
"learning_rate": 0.0002,
|
487 |
+
"loss": 2.8275,
|
488 |
+
"step": 800
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 0.03,
|
492 |
+
"learning_rate": 0.0002,
|
493 |
+
"loss": 2.8256,
|
494 |
+
"step": 810
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 0.03,
|
498 |
+
"learning_rate": 0.0002,
|
499 |
+
"loss": 2.8334,
|
500 |
+
"step": 820
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 0.03,
|
504 |
+
"learning_rate": 0.0002,
|
505 |
+
"loss": 2.806,
|
506 |
+
"step": 830
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 0.03,
|
510 |
+
"learning_rate": 0.0002,
|
511 |
+
"loss": 2.7845,
|
512 |
+
"step": 840
|
513 |
+
},
|
514 |
+
{
|
515 |
+
"epoch": 0.03,
|
516 |
+
"learning_rate": 0.0002,
|
517 |
+
"loss": 2.795,
|
518 |
+
"step": 850
|
519 |
+
},
|
520 |
+
{
|
521 |
+
"epoch": 0.03,
|
522 |
+
"learning_rate": 0.0002,
|
523 |
+
"loss": 2.7606,
|
524 |
+
"step": 860
|
525 |
+
},
|
526 |
+
{
|
527 |
+
"epoch": 0.03,
|
528 |
+
"learning_rate": 0.0002,
|
529 |
+
"loss": 2.7933,
|
530 |
+
"step": 870
|
531 |
+
},
|
532 |
+
{
|
533 |
+
"epoch": 0.03,
|
534 |
+
"learning_rate": 0.0002,
|
535 |
+
"loss": 2.7927,
|
536 |
+
"step": 880
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 0.03,
|
540 |
+
"learning_rate": 0.0002,
|
541 |
+
"loss": 2.79,
|
542 |
+
"step": 890
|
543 |
+
},
|
544 |
+
{
|
545 |
+
"epoch": 0.03,
|
546 |
+
"learning_rate": 0.0002,
|
547 |
+
"loss": 2.8054,
|
548 |
+
"step": 900
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 0.03,
|
552 |
+
"learning_rate": 0.0002,
|
553 |
+
"loss": 2.8285,
|
554 |
+
"step": 910
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 0.03,
|
558 |
+
"learning_rate": 0.0002,
|
559 |
+
"loss": 2.8354,
|
560 |
+
"step": 920
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 0.03,
|
564 |
+
"learning_rate": 0.0002,
|
565 |
+
"loss": 2.8047,
|
566 |
+
"step": 930
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 0.03,
|
570 |
+
"learning_rate": 0.0002,
|
571 |
+
"loss": 2.7817,
|
572 |
+
"step": 940
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 0.03,
|
576 |
+
"learning_rate": 0.0002,
|
577 |
+
"loss": 2.7981,
|
578 |
+
"step": 950
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 0.03,
|
582 |
+
"learning_rate": 0.0002,
|
583 |
+
"loss": 2.7823,
|
584 |
+
"step": 960
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 0.03,
|
588 |
+
"learning_rate": 0.0002,
|
589 |
+
"loss": 2.8003,
|
590 |
+
"step": 970
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 0.03,
|
594 |
+
"learning_rate": 0.0002,
|
595 |
+
"loss": 2.8004,
|
596 |
+
"step": 980
|
597 |
+
},
|
598 |
+
{
|
599 |
+
"epoch": 0.03,
|
600 |
+
"learning_rate": 0.0002,
|
601 |
+
"loss": 2.7759,
|
602 |
+
"step": 990
|
603 |
+
},
|
604 |
+
{
|
605 |
+
"epoch": 0.03,
|
606 |
+
"learning_rate": 0.0002,
|
607 |
+
"loss": 2.7948,
|
608 |
+
"step": 1000
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 0.03,
|
612 |
+
"learning_rate": 0.0002,
|
613 |
+
"loss": 2.8154,
|
614 |
+
"step": 1010
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.03,
|
618 |
+
"learning_rate": 0.0002,
|
619 |
+
"loss": 2.8316,
|
620 |
+
"step": 1020
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 0.03,
|
624 |
+
"learning_rate": 0.0002,
|
625 |
+
"loss": 2.7714,
|
626 |
+
"step": 1030
|
627 |
+
},
|
628 |
+
{
|
629 |
+
"epoch": 0.03,
|
630 |
+
"learning_rate": 0.0002,
|
631 |
+
"loss": 2.7947,
|
632 |
+
"step": 1040
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 0.03,
|
636 |
+
"learning_rate": 0.0002,
|
637 |
+
"loss": 2.7863,
|
638 |
+
"step": 1050
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 0.03,
|
642 |
+
"learning_rate": 0.0002,
|
643 |
+
"loss": 2.806,
|
644 |
+
"step": 1060
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 0.03,
|
648 |
+
"learning_rate": 0.0002,
|
649 |
+
"loss": 2.7746,
|
650 |
+
"step": 1070
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 0.03,
|
654 |
+
"learning_rate": 0.0002,
|
655 |
+
"loss": 2.8075,
|
656 |
+
"step": 1080
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 0.04,
|
660 |
+
"learning_rate": 0.0002,
|
661 |
+
"loss": 2.7656,
|
662 |
+
"step": 1090
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 0.04,
|
666 |
+
"learning_rate": 0.0002,
|
667 |
+
"loss": 2.8109,
|
668 |
+
"step": 1100
|
669 |
+
},
|
670 |
+
{
|
671 |
+
"epoch": 0.04,
|
672 |
+
"learning_rate": 0.0002,
|
673 |
+
"loss": 2.7928,
|
674 |
+
"step": 1110
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 0.04,
|
678 |
+
"learning_rate": 0.0002,
|
679 |
+
"loss": 2.822,
|
680 |
+
"step": 1120
|
681 |
+
},
|
682 |
+
{
|
683 |
+
"epoch": 0.04,
|
684 |
+
"learning_rate": 0.0002,
|
685 |
+
"loss": 2.861,
|
686 |
+
"step": 1130
|
687 |
+
},
|
688 |
+
{
|
689 |
+
"epoch": 0.04,
|
690 |
+
"learning_rate": 0.0002,
|
691 |
+
"loss": 2.7977,
|
692 |
+
"step": 1140
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 0.04,
|
696 |
+
"learning_rate": 0.0002,
|
697 |
+
"loss": 2.7805,
|
698 |
+
"step": 1150
|
699 |
+
},
|
700 |
+
{
|
701 |
+
"epoch": 0.04,
|
702 |
+
"learning_rate": 0.0002,
|
703 |
+
"loss": 2.7758,
|
704 |
+
"step": 1160
|
705 |
+
},
|
706 |
+
{
|
707 |
+
"epoch": 0.04,
|
708 |
+
"learning_rate": 0.0002,
|
709 |
+
"loss": 2.8053,
|
710 |
+
"step": 1170
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 0.04,
|
714 |
+
"learning_rate": 0.0002,
|
715 |
+
"loss": 2.7491,
|
716 |
+
"step": 1180
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 0.04,
|
720 |
+
"learning_rate": 0.0002,
|
721 |
+
"loss": 2.7826,
|
722 |
+
"step": 1190
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 0.04,
|
726 |
+
"learning_rate": 0.0002,
|
727 |
+
"loss": 2.8382,
|
728 |
+
"step": 1200
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 0.04,
|
732 |
+
"learning_rate": 0.0002,
|
733 |
+
"loss": 2.8119,
|
734 |
+
"step": 1210
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 0.04,
|
738 |
+
"learning_rate": 0.0002,
|
739 |
+
"loss": 2.7602,
|
740 |
+
"step": 1220
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 0.04,
|
744 |
+
"learning_rate": 0.0002,
|
745 |
+
"loss": 2.8249,
|
746 |
+
"step": 1230
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 0.04,
|
750 |
+
"learning_rate": 0.0002,
|
751 |
+
"loss": 2.7729,
|
752 |
+
"step": 1240
|
753 |
+
},
|
754 |
+
{
|
755 |
+
"epoch": 0.04,
|
756 |
+
"learning_rate": 0.0002,
|
757 |
+
"loss": 2.7897,
|
758 |
+
"step": 1250
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 0.04,
|
762 |
+
"learning_rate": 0.0002,
|
763 |
+
"loss": 2.7743,
|
764 |
+
"step": 1260
|
765 |
+
},
|
766 |
+
{
|
767 |
+
"epoch": 0.04,
|
768 |
+
"learning_rate": 0.0002,
|
769 |
+
"loss": 2.7666,
|
770 |
+
"step": 1270
|
771 |
+
},
|
772 |
+
{
|
773 |
+
"epoch": 0.04,
|
774 |
+
"learning_rate": 0.0002,
|
775 |
+
"loss": 2.8326,
|
776 |
+
"step": 1280
|
777 |
+
},
|
778 |
+
{
|
779 |
+
"epoch": 0.04,
|
780 |
+
"learning_rate": 0.0002,
|
781 |
+
"loss": 2.8229,
|
782 |
+
"step": 1290
|
783 |
+
},
|
784 |
+
{
|
785 |
+
"epoch": 0.04,
|
786 |
+
"learning_rate": 0.0002,
|
787 |
+
"loss": 2.8069,
|
788 |
+
"step": 1300
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 0.04,
|
792 |
+
"learning_rate": 0.0002,
|
793 |
+
"loss": 2.7894,
|
794 |
+
"step": 1310
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 0.04,
|
798 |
+
"learning_rate": 0.0002,
|
799 |
+
"loss": 2.787,
|
800 |
+
"step": 1320
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 0.04,
|
804 |
+
"learning_rate": 0.0002,
|
805 |
+
"loss": 2.785,
|
806 |
+
"step": 1330
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 0.04,
|
810 |
+
"learning_rate": 0.0002,
|
811 |
+
"loss": 2.8032,
|
812 |
+
"step": 1340
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 0.04,
|
816 |
+
"learning_rate": 0.0002,
|
817 |
+
"loss": 2.761,
|
818 |
+
"step": 1350
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 0.04,
|
822 |
+
"learning_rate": 0.0002,
|
823 |
+
"loss": 2.8074,
|
824 |
+
"step": 1360
|
825 |
+
},
|
826 |
+
{
|
827 |
+
"epoch": 0.04,
|
828 |
+
"learning_rate": 0.0002,
|
829 |
+
"loss": 2.7925,
|
830 |
+
"step": 1370
|
831 |
+
},
|
832 |
+
{
|
833 |
+
"epoch": 0.04,
|
834 |
+
"learning_rate": 0.0002,
|
835 |
+
"loss": 2.8167,
|
836 |
+
"step": 1380
|
837 |
+
},
|
838 |
+
{
|
839 |
+
"epoch": 0.04,
|
840 |
+
"learning_rate": 0.0002,
|
841 |
+
"loss": 2.7805,
|
842 |
+
"step": 1390
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 0.05,
|
846 |
+
"learning_rate": 0.0002,
|
847 |
+
"loss": 2.7819,
|
848 |
+
"step": 1400
|
849 |
+
},
|
850 |
+
{
|
851 |
+
"epoch": 0.05,
|
852 |
+
"learning_rate": 0.0002,
|
853 |
+
"loss": 2.8112,
|
854 |
+
"step": 1410
|
855 |
+
},
|
856 |
+
{
|
857 |
+
"epoch": 0.05,
|
858 |
+
"learning_rate": 0.0002,
|
859 |
+
"loss": 2.8142,
|
860 |
+
"step": 1420
|
861 |
+
},
|
862 |
+
{
|
863 |
+
"epoch": 0.05,
|
864 |
+
"learning_rate": 0.0002,
|
865 |
+
"loss": 2.7858,
|
866 |
+
"step": 1430
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 0.05,
|
870 |
+
"learning_rate": 0.0002,
|
871 |
+
"loss": 2.7584,
|
872 |
+
"step": 1440
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.05,
|
876 |
+
"learning_rate": 0.0002,
|
877 |
+
"loss": 2.7949,
|
878 |
+
"step": 1450
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 0.05,
|
882 |
+
"learning_rate": 0.0002,
|
883 |
+
"loss": 2.8499,
|
884 |
+
"step": 1460
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 0.05,
|
888 |
+
"learning_rate": 0.0002,
|
889 |
+
"loss": 2.8228,
|
890 |
+
"step": 1470
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 0.05,
|
894 |
+
"learning_rate": 0.0002,
|
895 |
+
"loss": 2.7731,
|
896 |
+
"step": 1480
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 0.05,
|
900 |
+
"learning_rate": 0.0002,
|
901 |
+
"loss": 2.8033,
|
902 |
+
"step": 1490
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 0.05,
|
906 |
+
"learning_rate": 0.0002,
|
907 |
+
"loss": 2.8054,
|
908 |
+
"step": 1500
|
909 |
+
},
|
910 |
+
{
|
911 |
+
"epoch": 0.05,
|
912 |
+
"learning_rate": 0.0002,
|
913 |
+
"loss": 2.7958,
|
914 |
+
"step": 1510
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 0.05,
|
918 |
+
"learning_rate": 0.0002,
|
919 |
+
"loss": 2.7941,
|
920 |
+
"step": 1520
|
921 |
+
},
|
922 |
+
{
|
923 |
+
"epoch": 0.05,
|
924 |
+
"learning_rate": 0.0002,
|
925 |
+
"loss": 2.8277,
|
926 |
+
"step": 1530
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 0.05,
|
930 |
+
"learning_rate": 0.0002,
|
931 |
+
"loss": 2.8018,
|
932 |
+
"step": 1540
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 0.05,
|
936 |
+
"learning_rate": 0.0002,
|
937 |
+
"loss": 2.819,
|
938 |
+
"step": 1550
|
939 |
+
},
|
940 |
+
{
|
941 |
+
"epoch": 0.05,
|
942 |
+
"learning_rate": 0.0002,
|
943 |
+
"loss": 2.7754,
|
944 |
+
"step": 1560
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 0.05,
|
948 |
+
"learning_rate": 0.0002,
|
949 |
+
"loss": 2.8026,
|
950 |
+
"step": 1570
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 0.05,
|
954 |
+
"learning_rate": 0.0002,
|
955 |
+
"loss": 2.8033,
|
956 |
+
"step": 1580
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 0.05,
|
960 |
+
"learning_rate": 0.0002,
|
961 |
+
"loss": 2.7695,
|
962 |
+
"step": 1590
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 0.05,
|
966 |
+
"learning_rate": 0.0002,
|
967 |
+
"loss": 2.8164,
|
968 |
+
"step": 1600
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 0.05,
|
972 |
+
"learning_rate": 0.0002,
|
973 |
+
"loss": 2.7647,
|
974 |
+
"step": 1610
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 0.05,
|
978 |
+
"learning_rate": 0.0002,
|
979 |
+
"loss": 2.7884,
|
980 |
+
"step": 1620
|
981 |
+
},
|
982 |
+
{
|
983 |
+
"epoch": 0.05,
|
984 |
+
"learning_rate": 0.0002,
|
985 |
+
"loss": 2.7753,
|
986 |
+
"step": 1630
|
987 |
+
},
|
988 |
+
{
|
989 |
+
"epoch": 0.05,
|
990 |
+
"learning_rate": 0.0002,
|
991 |
+
"loss": 2.8015,
|
992 |
+
"step": 1640
|
993 |
+
},
|
994 |
+
{
|
995 |
+
"epoch": 0.05,
|
996 |
+
"learning_rate": 0.0002,
|
997 |
+
"loss": 2.8319,
|
998 |
+
"step": 1650
|
999 |
+
},
|
1000 |
+
{
|
1001 |
+
"epoch": 0.05,
|
1002 |
+
"learning_rate": 0.0002,
|
1003 |
+
"loss": 2.7853,
|
1004 |
+
"step": 1660
|
1005 |
+
},
|
1006 |
+
{
|
1007 |
+
"epoch": 0.05,
|
1008 |
+
"learning_rate": 0.0002,
|
1009 |
+
"loss": 2.7749,
|
1010 |
+
"step": 1670
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 0.05,
|
1014 |
+
"learning_rate": 0.0002,
|
1015 |
+
"loss": 2.8053,
|
1016 |
+
"step": 1680
|
1017 |
+
},
|
1018 |
+
{
|
1019 |
+
"epoch": 0.05,
|
1020 |
+
"learning_rate": 0.0002,
|
1021 |
+
"loss": 2.7519,
|
1022 |
+
"step": 1690
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 0.05,
|
1026 |
+
"learning_rate": 0.0002,
|
1027 |
+
"loss": 2.7748,
|
1028 |
+
"step": 1700
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 0.06,
|
1032 |
+
"learning_rate": 0.0002,
|
1033 |
+
"loss": 2.7965,
|
1034 |
+
"step": 1710
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 0.06,
|
1038 |
+
"learning_rate": 0.0002,
|
1039 |
+
"loss": 2.7741,
|
1040 |
+
"step": 1720
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 0.06,
|
1044 |
+
"learning_rate": 0.0002,
|
1045 |
+
"loss": 2.8151,
|
1046 |
+
"step": 1730
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 0.06,
|
1050 |
+
"learning_rate": 0.0002,
|
1051 |
+
"loss": 2.7982,
|
1052 |
+
"step": 1740
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 0.06,
|
1056 |
+
"learning_rate": 0.0002,
|
1057 |
+
"loss": 2.8229,
|
1058 |
+
"step": 1750
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 0.06,
|
1062 |
+
"learning_rate": 0.0002,
|
1063 |
+
"loss": 2.7984,
|
1064 |
+
"step": 1760
|
1065 |
+
},
|
1066 |
+
{
|
1067 |
+
"epoch": 0.06,
|
1068 |
+
"learning_rate": 0.0002,
|
1069 |
+
"loss": 2.7661,
|
1070 |
+
"step": 1770
|
1071 |
+
},
|
1072 |
+
{
|
1073 |
+
"epoch": 0.06,
|
1074 |
+
"learning_rate": 0.0002,
|
1075 |
+
"loss": 2.8516,
|
1076 |
+
"step": 1780
|
1077 |
+
},
|
1078 |
+
{
|
1079 |
+
"epoch": 0.06,
|
1080 |
+
"learning_rate": 0.0002,
|
1081 |
+
"loss": 2.8034,
|
1082 |
+
"step": 1790
|
1083 |
+
},
|
1084 |
+
{
|
1085 |
+
"epoch": 0.06,
|
1086 |
+
"learning_rate": 0.0002,
|
1087 |
+
"loss": 2.7805,
|
1088 |
+
"step": 1800
|
1089 |
+
},
|
1090 |
+
{
|
1091 |
+
"epoch": 0.06,
|
1092 |
+
"learning_rate": 0.0002,
|
1093 |
+
"loss": 2.8021,
|
1094 |
+
"step": 1810
|
1095 |
+
},
|
1096 |
+
{
|
1097 |
+
"epoch": 0.06,
|
1098 |
+
"learning_rate": 0.0002,
|
1099 |
+
"loss": 2.783,
|
1100 |
+
"step": 1820
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 0.06,
|
1104 |
+
"learning_rate": 0.0002,
|
1105 |
+
"loss": 2.7751,
|
1106 |
+
"step": 1830
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 0.06,
|
1110 |
+
"learning_rate": 0.0002,
|
1111 |
+
"loss": 2.7926,
|
1112 |
+
"step": 1840
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 0.06,
|
1116 |
+
"learning_rate": 0.0002,
|
1117 |
+
"loss": 2.7954,
|
1118 |
+
"step": 1850
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 0.06,
|
1122 |
+
"learning_rate": 0.0002,
|
1123 |
+
"loss": 2.7673,
|
1124 |
+
"step": 1860
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 0.06,
|
1128 |
+
"learning_rate": 0.0002,
|
1129 |
+
"loss": 2.7672,
|
1130 |
+
"step": 1870
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 0.06,
|
1134 |
+
"learning_rate": 0.0002,
|
1135 |
+
"loss": 2.7545,
|
1136 |
+
"step": 1880
|
1137 |
+
},
|
1138 |
+
{
|
1139 |
+
"epoch": 0.06,
|
1140 |
+
"learning_rate": 0.0002,
|
1141 |
+
"loss": 2.7972,
|
1142 |
+
"step": 1890
|
1143 |
+
},
|
1144 |
+
{
|
1145 |
+
"epoch": 0.06,
|
1146 |
+
"learning_rate": 0.0002,
|
1147 |
+
"loss": 2.8428,
|
1148 |
+
"step": 1900
|
1149 |
+
},
|
1150 |
+
{
|
1151 |
+
"epoch": 0.06,
|
1152 |
+
"learning_rate": 0.0002,
|
1153 |
+
"loss": 2.7968,
|
1154 |
+
"step": 1910
|
1155 |
+
},
|
1156 |
+
{
|
1157 |
+
"epoch": 0.06,
|
1158 |
+
"learning_rate": 0.0002,
|
1159 |
+
"loss": 2.8249,
|
1160 |
+
"step": 1920
|
1161 |
+
},
|
1162 |
+
{
|
1163 |
+
"epoch": 0.06,
|
1164 |
+
"learning_rate": 0.0002,
|
1165 |
+
"loss": 2.7491,
|
1166 |
+
"step": 1930
|
1167 |
+
},
|
1168 |
+
{
|
1169 |
+
"epoch": 0.06,
|
1170 |
+
"learning_rate": 0.0002,
|
1171 |
+
"loss": 2.7843,
|
1172 |
+
"step": 1940
|
1173 |
+
},
|
1174 |
+
{
|
1175 |
+
"epoch": 0.06,
|
1176 |
+
"learning_rate": 0.0002,
|
1177 |
+
"loss": 2.8082,
|
1178 |
+
"step": 1950
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 0.06,
|
1182 |
+
"learning_rate": 0.0002,
|
1183 |
+
"loss": 2.7685,
|
1184 |
+
"step": 1960
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 0.06,
|
1188 |
+
"learning_rate": 0.0002,
|
1189 |
+
"loss": 2.7934,
|
1190 |
+
"step": 1970
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 0.06,
|
1194 |
+
"learning_rate": 0.0002,
|
1195 |
+
"loss": 2.796,
|
1196 |
+
"step": 1980
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 0.06,
|
1200 |
+
"learning_rate": 0.0002,
|
1201 |
+
"loss": 2.8012,
|
1202 |
+
"step": 1990
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 0.06,
|
1206 |
+
"learning_rate": 0.0002,
|
1207 |
+
"loss": 2.7579,
|
1208 |
+
"step": 2000
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 0.06,
|
1212 |
+
"learning_rate": 0.0002,
|
1213 |
+
"loss": 2.7828,
|
1214 |
+
"step": 2010
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 0.07,
|
1218 |
+
"learning_rate": 0.0002,
|
1219 |
+
"loss": 2.7891,
|
1220 |
+
"step": 2020
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 0.07,
|
1224 |
+
"learning_rate": 0.0002,
|
1225 |
+
"loss": 2.7988,
|
1226 |
+
"step": 2030
|
1227 |
+
},
|
1228 |
+
{
|
1229 |
+
"epoch": 0.07,
|
1230 |
+
"learning_rate": 0.0002,
|
1231 |
+
"loss": 2.7918,
|
1232 |
+
"step": 2040
|
1233 |
+
},
|
1234 |
+
{
|
1235 |
+
"epoch": 0.07,
|
1236 |
+
"learning_rate": 0.0002,
|
1237 |
+
"loss": 2.8269,
|
1238 |
+
"step": 2050
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 0.07,
|
1242 |
+
"learning_rate": 0.0002,
|
1243 |
+
"loss": 2.7916,
|
1244 |
+
"step": 2060
|
1245 |
+
},
|
1246 |
+
{
|
1247 |
+
"epoch": 0.07,
|
1248 |
+
"learning_rate": 0.0002,
|
1249 |
+
"loss": 2.785,
|
1250 |
+
"step": 2070
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 0.07,
|
1254 |
+
"learning_rate": 0.0002,
|
1255 |
+
"loss": 2.7495,
|
1256 |
+
"step": 2080
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 0.07,
|
1260 |
+
"learning_rate": 0.0002,
|
1261 |
+
"loss": 2.7931,
|
1262 |
+
"step": 2090
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 0.07,
|
1266 |
+
"learning_rate": 0.0002,
|
1267 |
+
"loss": 2.7967,
|
1268 |
+
"step": 2100
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 0.07,
|
1272 |
+
"learning_rate": 0.0002,
|
1273 |
+
"loss": 2.7271,
|
1274 |
+
"step": 2110
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 0.07,
|
1278 |
+
"learning_rate": 0.0002,
|
1279 |
+
"loss": 2.7938,
|
1280 |
+
"step": 2120
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 0.07,
|
1284 |
+
"learning_rate": 0.0002,
|
1285 |
+
"loss": 2.8055,
|
1286 |
+
"step": 2130
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 0.07,
|
1290 |
+
"learning_rate": 0.0002,
|
1291 |
+
"loss": 2.7623,
|
1292 |
+
"step": 2140
|
1293 |
+
},
|
1294 |
+
{
|
1295 |
+
"epoch": 0.07,
|
1296 |
+
"learning_rate": 0.0002,
|
1297 |
+
"loss": 2.8052,
|
1298 |
+
"step": 2150
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 0.07,
|
1302 |
+
"learning_rate": 0.0002,
|
1303 |
+
"loss": 2.8144,
|
1304 |
+
"step": 2160
|
1305 |
+
},
|
1306 |
+
{
|
1307 |
+
"epoch": 0.07,
|
1308 |
+
"learning_rate": 0.0002,
|
1309 |
+
"loss": 2.7964,
|
1310 |
+
"step": 2170
|
1311 |
+
},
|
1312 |
+
{
|
1313 |
+
"epoch": 0.07,
|
1314 |
+
"learning_rate": 0.0002,
|
1315 |
+
"loss": 2.7778,
|
1316 |
+
"step": 2180
|
1317 |
+
},
|
1318 |
+
{
|
1319 |
+
"epoch": 0.07,
|
1320 |
+
"learning_rate": 0.0002,
|
1321 |
+
"loss": 2.7869,
|
1322 |
+
"step": 2190
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 0.07,
|
1326 |
+
"learning_rate": 0.0002,
|
1327 |
+
"loss": 2.8252,
|
1328 |
+
"step": 2200
|
1329 |
+
},
|
1330 |
+
{
|
1331 |
+
"epoch": 0.07,
|
1332 |
+
"learning_rate": 0.0002,
|
1333 |
+
"loss": 2.783,
|
1334 |
+
"step": 2210
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 0.07,
|
1338 |
+
"learning_rate": 0.0002,
|
1339 |
+
"loss": 2.7544,
|
1340 |
+
"step": 2220
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 0.07,
|
1344 |
+
"learning_rate": 0.0002,
|
1345 |
+
"loss": 2.8041,
|
1346 |
+
"step": 2230
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 0.07,
|
1350 |
+
"learning_rate": 0.0002,
|
1351 |
+
"loss": 2.772,
|
1352 |
+
"step": 2240
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 0.07,
|
1356 |
+
"learning_rate": 0.0002,
|
1357 |
+
"loss": 2.7963,
|
1358 |
+
"step": 2250
|
1359 |
+
},
|
1360 |
+
{
|
1361 |
+
"epoch": 0.07,
|
1362 |
+
"learning_rate": 0.0002,
|
1363 |
+
"loss": 2.8056,
|
1364 |
+
"step": 2260
|
1365 |
+
},
|
1366 |
+
{
|
1367 |
+
"epoch": 0.07,
|
1368 |
+
"learning_rate": 0.0002,
|
1369 |
+
"loss": 2.8026,
|
1370 |
+
"step": 2270
|
1371 |
+
},
|
1372 |
+
{
|
1373 |
+
"epoch": 0.07,
|
1374 |
+
"learning_rate": 0.0002,
|
1375 |
+
"loss": 2.7501,
|
1376 |
+
"step": 2280
|
1377 |
+
},
|
1378 |
+
{
|
1379 |
+
"epoch": 0.07,
|
1380 |
+
"learning_rate": 0.0002,
|
1381 |
+
"loss": 2.7515,
|
1382 |
+
"step": 2290
|
1383 |
+
},
|
1384 |
+
{
|
1385 |
+
"epoch": 0.07,
|
1386 |
+
"learning_rate": 0.0002,
|
1387 |
+
"loss": 2.7931,
|
1388 |
+
"step": 2300
|
1389 |
+
},
|
1390 |
+
{
|
1391 |
+
"epoch": 0.07,
|
1392 |
+
"learning_rate": 0.0002,
|
1393 |
+
"loss": 2.8126,
|
1394 |
+
"step": 2310
|
1395 |
+
},
|
1396 |
+
{
|
1397 |
+
"epoch": 0.07,
|
1398 |
+
"learning_rate": 0.0002,
|
1399 |
+
"loss": 2.8232,
|
1400 |
+
"step": 2320
|
1401 |
+
},
|
1402 |
+
{
|
1403 |
+
"epoch": 0.08,
|
1404 |
+
"learning_rate": 0.0002,
|
1405 |
+
"loss": 2.7556,
|
1406 |
+
"step": 2330
|
1407 |
+
},
|
1408 |
+
{
|
1409 |
+
"epoch": 0.08,
|
1410 |
+
"learning_rate": 0.0002,
|
1411 |
+
"loss": 2.7862,
|
1412 |
+
"step": 2340
|
1413 |
+
},
|
1414 |
+
{
|
1415 |
+
"epoch": 0.08,
|
1416 |
+
"learning_rate": 0.0002,
|
1417 |
+
"loss": 2.7879,
|
1418 |
+
"step": 2350
|
1419 |
+
},
|
1420 |
+
{
|
1421 |
+
"epoch": 0.08,
|
1422 |
+
"learning_rate": 0.0002,
|
1423 |
+
"loss": 2.806,
|
1424 |
+
"step": 2360
|
1425 |
+
},
|
1426 |
+
{
|
1427 |
+
"epoch": 0.08,
|
1428 |
+
"learning_rate": 0.0002,
|
1429 |
+
"loss": 2.758,
|
1430 |
+
"step": 2370
|
1431 |
+
},
|
1432 |
+
{
|
1433 |
+
"epoch": 0.08,
|
1434 |
+
"learning_rate": 0.0002,
|
1435 |
+
"loss": 2.7859,
|
1436 |
+
"step": 2380
|
1437 |
+
},
|
1438 |
+
{
|
1439 |
+
"epoch": 0.08,
|
1440 |
+
"learning_rate": 0.0002,
|
1441 |
+
"loss": 2.7884,
|
1442 |
+
"step": 2390
|
1443 |
+
},
|
1444 |
+
{
|
1445 |
+
"epoch": 0.08,
|
1446 |
+
"learning_rate": 0.0002,
|
1447 |
+
"loss": 2.7934,
|
1448 |
+
"step": 2400
|
1449 |
+
},
|
1450 |
+
{
|
1451 |
+
"epoch": 0.08,
|
1452 |
+
"learning_rate": 0.0002,
|
1453 |
+
"loss": 2.8048,
|
1454 |
+
"step": 2410
|
1455 |
+
},
|
1456 |
+
{
|
1457 |
+
"epoch": 0.08,
|
1458 |
+
"learning_rate": 0.0002,
|
1459 |
+
"loss": 2.7897,
|
1460 |
+
"step": 2420
|
1461 |
+
},
|
1462 |
+
{
|
1463 |
+
"epoch": 0.08,
|
1464 |
+
"learning_rate": 0.0002,
|
1465 |
+
"loss": 2.7762,
|
1466 |
+
"step": 2430
|
1467 |
+
},
|
1468 |
+
{
|
1469 |
+
"epoch": 0.08,
|
1470 |
+
"learning_rate": 0.0002,
|
1471 |
+
"loss": 2.7546,
|
1472 |
+
"step": 2440
|
1473 |
+
},
|
1474 |
+
{
|
1475 |
+
"epoch": 0.08,
|
1476 |
+
"learning_rate": 0.0002,
|
1477 |
+
"loss": 2.7745,
|
1478 |
+
"step": 2450
|
1479 |
+
},
|
1480 |
+
{
|
1481 |
+
"epoch": 0.08,
|
1482 |
+
"learning_rate": 0.0002,
|
1483 |
+
"loss": 2.7908,
|
1484 |
+
"step": 2460
|
1485 |
+
},
|
1486 |
+
{
|
1487 |
+
"epoch": 0.08,
|
1488 |
+
"learning_rate": 0.0002,
|
1489 |
+
"loss": 2.8209,
|
1490 |
+
"step": 2470
|
1491 |
+
},
|
1492 |
+
{
|
1493 |
+
"epoch": 0.08,
|
1494 |
+
"learning_rate": 0.0002,
|
1495 |
+
"loss": 2.7881,
|
1496 |
+
"step": 2480
|
1497 |
+
},
|
1498 |
+
{
|
1499 |
+
"epoch": 0.08,
|
1500 |
+
"learning_rate": 0.0002,
|
1501 |
+
"loss": 2.7929,
|
1502 |
+
"step": 2490
|
1503 |
+
},
|
1504 |
+
{
|
1505 |
+
"epoch": 0.08,
|
1506 |
+
"learning_rate": 0.0002,
|
1507 |
+
"loss": 2.7735,
|
1508 |
+
"step": 2500
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 0.08,
|
1512 |
+
"learning_rate": 0.0002,
|
1513 |
+
"loss": 2.7656,
|
1514 |
+
"step": 2510
|
1515 |
+
},
|
1516 |
+
{
|
1517 |
+
"epoch": 0.08,
|
1518 |
+
"learning_rate": 0.0002,
|
1519 |
+
"loss": 2.8099,
|
1520 |
+
"step": 2520
|
1521 |
+
},
|
1522 |
+
{
|
1523 |
+
"epoch": 0.08,
|
1524 |
+
"learning_rate": 0.0002,
|
1525 |
+
"loss": 2.7826,
|
1526 |
+
"step": 2530
|
1527 |
+
},
|
1528 |
+
{
|
1529 |
+
"epoch": 0.08,
|
1530 |
+
"learning_rate": 0.0002,
|
1531 |
+
"loss": 2.7748,
|
1532 |
+
"step": 2540
|
1533 |
+
},
|
1534 |
+
{
|
1535 |
+
"epoch": 0.08,
|
1536 |
+
"learning_rate": 0.0002,
|
1537 |
+
"loss": 2.7853,
|
1538 |
+
"step": 2550
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 0.08,
|
1542 |
+
"learning_rate": 0.0002,
|
1543 |
+
"loss": 2.8411,
|
1544 |
+
"step": 2560
|
1545 |
+
},
|
1546 |
+
{
|
1547 |
+
"epoch": 0.08,
|
1548 |
+
"learning_rate": 0.0002,
|
1549 |
+
"loss": 2.747,
|
1550 |
+
"step": 2570
|
1551 |
+
},
|
1552 |
+
{
|
1553 |
+
"epoch": 0.08,
|
1554 |
+
"learning_rate": 0.0002,
|
1555 |
+
"loss": 2.8118,
|
1556 |
+
"step": 2580
|
1557 |
+
},
|
1558 |
+
{
|
1559 |
+
"epoch": 0.08,
|
1560 |
+
"learning_rate": 0.0002,
|
1561 |
+
"loss": 2.7836,
|
1562 |
+
"step": 2590
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 0.08,
|
1566 |
+
"learning_rate": 0.0002,
|
1567 |
+
"loss": 2.7898,
|
1568 |
+
"step": 2600
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 0.08,
|
1572 |
+
"learning_rate": 0.0002,
|
1573 |
+
"loss": 2.7546,
|
1574 |
+
"step": 2610
|
1575 |
+
},
|
1576 |
+
{
|
1577 |
+
"epoch": 0.08,
|
1578 |
+
"learning_rate": 0.0002,
|
1579 |
+
"loss": 2.7698,
|
1580 |
+
"step": 2620
|
1581 |
+
},
|
1582 |
+
{
|
1583 |
+
"epoch": 0.08,
|
1584 |
+
"learning_rate": 0.0002,
|
1585 |
+
"loss": 2.6932,
|
1586 |
+
"step": 2630
|
1587 |
+
},
|
1588 |
+
{
|
1589 |
+
"epoch": 0.09,
|
1590 |
+
"learning_rate": 0.0002,
|
1591 |
+
"loss": 2.7996,
|
1592 |
+
"step": 2640
|
1593 |
+
},
|
1594 |
+
{
|
1595 |
+
"epoch": 0.09,
|
1596 |
+
"learning_rate": 0.0002,
|
1597 |
+
"loss": 2.7868,
|
1598 |
+
"step": 2650
|
1599 |
+
},
|
1600 |
+
{
|
1601 |
+
"epoch": 0.09,
|
1602 |
+
"learning_rate": 0.0002,
|
1603 |
+
"loss": 2.7585,
|
1604 |
+
"step": 2660
|
1605 |
+
},
|
1606 |
+
{
|
1607 |
+
"epoch": 0.09,
|
1608 |
+
"learning_rate": 0.0002,
|
1609 |
+
"loss": 2.7719,
|
1610 |
+
"step": 2670
|
1611 |
+
},
|
1612 |
+
{
|
1613 |
+
"epoch": 0.09,
|
1614 |
+
"learning_rate": 0.0002,
|
1615 |
+
"loss": 2.7952,
|
1616 |
+
"step": 2680
|
1617 |
+
},
|
1618 |
+
{
|
1619 |
+
"epoch": 0.09,
|
1620 |
+
"learning_rate": 0.0002,
|
1621 |
+
"loss": 2.7817,
|
1622 |
+
"step": 2690
|
1623 |
+
},
|
1624 |
+
{
|
1625 |
+
"epoch": 0.09,
|
1626 |
+
"learning_rate": 0.0002,
|
1627 |
+
"loss": 2.7545,
|
1628 |
+
"step": 2700
|
1629 |
+
},
|
1630 |
+
{
|
1631 |
+
"epoch": 0.09,
|
1632 |
+
"learning_rate": 0.0002,
|
1633 |
+
"loss": 2.7823,
|
1634 |
+
"step": 2710
|
1635 |
+
},
|
1636 |
+
{
|
1637 |
+
"epoch": 0.09,
|
1638 |
+
"learning_rate": 0.0002,
|
1639 |
+
"loss": 2.766,
|
1640 |
+
"step": 2720
|
1641 |
+
},
|
1642 |
+
{
|
1643 |
+
"epoch": 0.09,
|
1644 |
+
"learning_rate": 0.0002,
|
1645 |
+
"loss": 2.7731,
|
1646 |
+
"step": 2730
|
1647 |
+
},
|
1648 |
+
{
|
1649 |
+
"epoch": 0.09,
|
1650 |
+
"learning_rate": 0.0002,
|
1651 |
+
"loss": 2.7992,
|
1652 |
+
"step": 2740
|
1653 |
+
},
|
1654 |
+
{
|
1655 |
+
"epoch": 0.09,
|
1656 |
+
"learning_rate": 0.0002,
|
1657 |
+
"loss": 2.8014,
|
1658 |
+
"step": 2750
|
1659 |
+
},
|
1660 |
+
{
|
1661 |
+
"epoch": 0.09,
|
1662 |
+
"learning_rate": 0.0002,
|
1663 |
+
"loss": 2.7459,
|
1664 |
+
"step": 2760
|
1665 |
+
},
|
1666 |
+
{
|
1667 |
+
"epoch": 0.09,
|
1668 |
+
"learning_rate": 0.0002,
|
1669 |
+
"loss": 2.7708,
|
1670 |
+
"step": 2770
|
1671 |
+
},
|
1672 |
+
{
|
1673 |
+
"epoch": 0.09,
|
1674 |
+
"learning_rate": 0.0002,
|
1675 |
+
"loss": 2.803,
|
1676 |
+
"step": 2780
|
1677 |
+
},
|
1678 |
+
{
|
1679 |
+
"epoch": 0.09,
|
1680 |
+
"learning_rate": 0.0002,
|
1681 |
+
"loss": 2.7514,
|
1682 |
+
"step": 2790
|
1683 |
+
},
|
1684 |
+
{
|
1685 |
+
"epoch": 0.09,
|
1686 |
+
"learning_rate": 0.0002,
|
1687 |
+
"loss": 2.7704,
|
1688 |
+
"step": 2800
|
1689 |
+
},
|
1690 |
+
{
|
1691 |
+
"epoch": 0.09,
|
1692 |
+
"learning_rate": 0.0002,
|
1693 |
+
"loss": 2.7466,
|
1694 |
+
"step": 2810
|
1695 |
+
},
|
1696 |
+
{
|
1697 |
+
"epoch": 0.09,
|
1698 |
+
"learning_rate": 0.0002,
|
1699 |
+
"loss": 2.7674,
|
1700 |
+
"step": 2820
|
1701 |
+
},
|
1702 |
+
{
|
1703 |
+
"epoch": 0.09,
|
1704 |
+
"learning_rate": 0.0002,
|
1705 |
+
"loss": 2.7642,
|
1706 |
+
"step": 2830
|
1707 |
+
},
|
1708 |
+
{
|
1709 |
+
"epoch": 0.09,
|
1710 |
+
"learning_rate": 0.0002,
|
1711 |
+
"loss": 2.8095,
|
1712 |
+
"step": 2840
|
1713 |
+
},
|
1714 |
+
{
|
1715 |
+
"epoch": 0.09,
|
1716 |
+
"learning_rate": 0.0002,
|
1717 |
+
"loss": 2.7579,
|
1718 |
+
"step": 2850
|
1719 |
+
},
|
1720 |
+
{
|
1721 |
+
"epoch": 0.09,
|
1722 |
+
"learning_rate": 0.0002,
|
1723 |
+
"loss": 2.8004,
|
1724 |
+
"step": 2860
|
1725 |
+
},
|
1726 |
+
{
|
1727 |
+
"epoch": 0.09,
|
1728 |
+
"learning_rate": 0.0002,
|
1729 |
+
"loss": 2.7808,
|
1730 |
+
"step": 2870
|
1731 |
+
},
|
1732 |
+
{
|
1733 |
+
"epoch": 0.09,
|
1734 |
+
"learning_rate": 0.0002,
|
1735 |
+
"loss": 2.765,
|
1736 |
+
"step": 2880
|
1737 |
+
},
|
1738 |
+
{
|
1739 |
+
"epoch": 0.09,
|
1740 |
+
"learning_rate": 0.0002,
|
1741 |
+
"loss": 2.7911,
|
1742 |
+
"step": 2890
|
1743 |
+
},
|
1744 |
+
{
|
1745 |
+
"epoch": 0.09,
|
1746 |
+
"learning_rate": 0.0002,
|
1747 |
+
"loss": 2.8098,
|
1748 |
+
"step": 2900
|
1749 |
+
},
|
1750 |
+
{
|
1751 |
+
"epoch": 0.09,
|
1752 |
+
"learning_rate": 0.0002,
|
1753 |
+
"loss": 2.7703,
|
1754 |
+
"step": 2910
|
1755 |
+
},
|
1756 |
+
{
|
1757 |
+
"epoch": 0.09,
|
1758 |
+
"learning_rate": 0.0002,
|
1759 |
+
"loss": 2.8085,
|
1760 |
+
"step": 2920
|
1761 |
+
},
|
1762 |
+
{
|
1763 |
+
"epoch": 0.09,
|
1764 |
+
"learning_rate": 0.0002,
|
1765 |
+
"loss": 2.7803,
|
1766 |
+
"step": 2930
|
1767 |
+
},
|
1768 |
+
{
|
1769 |
+
"epoch": 0.09,
|
1770 |
+
"learning_rate": 0.0002,
|
1771 |
+
"loss": 2.7789,
|
1772 |
+
"step": 2940
|
1773 |
+
},
|
1774 |
+
{
|
1775 |
+
"epoch": 0.1,
|
1776 |
+
"learning_rate": 0.0002,
|
1777 |
+
"loss": 2.7952,
|
1778 |
+
"step": 2950
|
1779 |
+
},
|
1780 |
+
{
|
1781 |
+
"epoch": 0.1,
|
1782 |
+
"learning_rate": 0.0002,
|
1783 |
+
"loss": 2.7917,
|
1784 |
+
"step": 2960
|
1785 |
+
},
|
1786 |
+
{
|
1787 |
+
"epoch": 0.1,
|
1788 |
+
"learning_rate": 0.0002,
|
1789 |
+
"loss": 2.7838,
|
1790 |
+
"step": 2970
|
1791 |
+
},
|
1792 |
+
{
|
1793 |
+
"epoch": 0.1,
|
1794 |
+
"learning_rate": 0.0002,
|
1795 |
+
"loss": 2.811,
|
1796 |
+
"step": 2980
|
1797 |
+
},
|
1798 |
+
{
|
1799 |
+
"epoch": 0.1,
|
1800 |
+
"learning_rate": 0.0002,
|
1801 |
+
"loss": 2.7422,
|
1802 |
+
"step": 2990
|
1803 |
+
},
|
1804 |
+
{
|
1805 |
+
"epoch": 0.1,
|
1806 |
+
"learning_rate": 0.0002,
|
1807 |
+
"loss": 2.785,
|
1808 |
+
"step": 3000
|
1809 |
+
},
|
1810 |
+
{
|
1811 |
+
"epoch": 0.1,
|
1812 |
+
"learning_rate": 0.0002,
|
1813 |
+
"loss": 2.8173,
|
1814 |
+
"step": 3010
|
1815 |
+
},
|
1816 |
+
{
|
1817 |
+
"epoch": 0.1,
|
1818 |
+
"learning_rate": 0.0002,
|
1819 |
+
"loss": 2.7445,
|
1820 |
+
"step": 3020
|
1821 |
+
},
|
1822 |
+
{
|
1823 |
+
"epoch": 0.1,
|
1824 |
+
"learning_rate": 0.0002,
|
1825 |
+
"loss": 2.7672,
|
1826 |
+
"step": 3030
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 0.1,
|
1830 |
+
"learning_rate": 0.0002,
|
1831 |
+
"loss": 2.8135,
|
1832 |
+
"step": 3040
|
1833 |
+
},
|
1834 |
+
{
|
1835 |
+
"epoch": 0.1,
|
1836 |
+
"learning_rate": 0.0002,
|
1837 |
+
"loss": 2.7966,
|
1838 |
+
"step": 3050
|
1839 |
+
},
|
1840 |
+
{
|
1841 |
+
"epoch": 0.1,
|
1842 |
+
"learning_rate": 0.0002,
|
1843 |
+
"loss": 2.8011,
|
1844 |
+
"step": 3060
|
1845 |
+
},
|
1846 |
+
{
|
1847 |
+
"epoch": 0.1,
|
1848 |
+
"learning_rate": 0.0002,
|
1849 |
+
"loss": 2.7647,
|
1850 |
+
"step": 3070
|
1851 |
+
},
|
1852 |
+
{
|
1853 |
+
"epoch": 0.1,
|
1854 |
+
"learning_rate": 0.0002,
|
1855 |
+
"loss": 2.816,
|
1856 |
+
"step": 3080
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 0.1,
|
1860 |
+
"learning_rate": 0.0002,
|
1861 |
+
"loss": 2.7714,
|
1862 |
+
"step": 3090
|
1863 |
+
},
|
1864 |
+
{
|
1865 |
+
"epoch": 0.1,
|
1866 |
+
"learning_rate": 0.0002,
|
1867 |
+
"loss": 2.7675,
|
1868 |
+
"step": 3100
|
1869 |
+
},
|
1870 |
+
{
|
1871 |
+
"epoch": 0.1,
|
1872 |
+
"learning_rate": 0.0002,
|
1873 |
+
"loss": 2.7416,
|
1874 |
+
"step": 3110
|
1875 |
+
},
|
1876 |
+
{
|
1877 |
+
"epoch": 0.1,
|
1878 |
+
"learning_rate": 0.0002,
|
1879 |
+
"loss": 2.7642,
|
1880 |
+
"step": 3120
|
1881 |
+
},
|
1882 |
+
{
|
1883 |
+
"epoch": 0.1,
|
1884 |
+
"learning_rate": 0.0002,
|
1885 |
+
"loss": 2.7898,
|
1886 |
+
"step": 3130
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 0.1,
|
1890 |
+
"learning_rate": 0.0002,
|
1891 |
+
"loss": 2.7972,
|
1892 |
+
"step": 3140
|
1893 |
+
},
|
1894 |
+
{
|
1895 |
+
"epoch": 0.1,
|
1896 |
+
"learning_rate": 0.0002,
|
1897 |
+
"loss": 2.7577,
|
1898 |
+
"step": 3150
|
1899 |
+
},
|
1900 |
+
{
|
1901 |
+
"epoch": 0.1,
|
1902 |
+
"learning_rate": 0.0002,
|
1903 |
+
"loss": 2.8205,
|
1904 |
+
"step": 3160
|
1905 |
+
},
|
1906 |
+
{
|
1907 |
+
"epoch": 0.1,
|
1908 |
+
"learning_rate": 0.0002,
|
1909 |
+
"loss": 2.7782,
|
1910 |
+
"step": 3170
|
1911 |
+
},
|
1912 |
+
{
|
1913 |
+
"epoch": 0.1,
|
1914 |
+
"learning_rate": 0.0002,
|
1915 |
+
"loss": 2.8022,
|
1916 |
+
"step": 3180
|
1917 |
+
},
|
1918 |
+
{
|
1919 |
+
"epoch": 0.1,
|
1920 |
+
"learning_rate": 0.0002,
|
1921 |
+
"loss": 2.793,
|
1922 |
+
"step": 3190
|
1923 |
+
},
|
1924 |
+
{
|
1925 |
+
"epoch": 0.1,
|
1926 |
+
"learning_rate": 0.0002,
|
1927 |
+
"loss": 2.7961,
|
1928 |
+
"step": 3200
|
1929 |
+
},
|
1930 |
+
{
|
1931 |
+
"epoch": 0.1,
|
1932 |
+
"learning_rate": 0.0002,
|
1933 |
+
"loss": 2.7955,
|
1934 |
+
"step": 3210
|
1935 |
+
},
|
1936 |
+
{
|
1937 |
+
"epoch": 0.1,
|
1938 |
+
"learning_rate": 0.0002,
|
1939 |
+
"loss": 2.769,
|
1940 |
+
"step": 3220
|
1941 |
+
},
|
1942 |
+
{
|
1943 |
+
"epoch": 0.1,
|
1944 |
+
"learning_rate": 0.0002,
|
1945 |
+
"loss": 2.7434,
|
1946 |
+
"step": 3230
|
1947 |
+
},
|
1948 |
+
{
|
1949 |
+
"epoch": 0.1,
|
1950 |
+
"learning_rate": 0.0002,
|
1951 |
+
"loss": 2.7987,
|
1952 |
+
"step": 3240
|
1953 |
+
},
|
1954 |
+
{
|
1955 |
+
"epoch": 0.1,
|
1956 |
+
"learning_rate": 0.0002,
|
1957 |
+
"loss": 2.7568,
|
1958 |
+
"step": 3250
|
1959 |
+
},
|
1960 |
+
{
|
1961 |
+
"epoch": 0.11,
|
1962 |
+
"learning_rate": 0.0002,
|
1963 |
+
"loss": 2.7449,
|
1964 |
+
"step": 3260
|
1965 |
+
},
|
1966 |
+
{
|
1967 |
+
"epoch": 0.11,
|
1968 |
+
"learning_rate": 0.0002,
|
1969 |
+
"loss": 2.7982,
|
1970 |
+
"step": 3270
|
1971 |
+
},
|
1972 |
+
{
|
1973 |
+
"epoch": 0.11,
|
1974 |
+
"learning_rate": 0.0002,
|
1975 |
+
"loss": 2.7978,
|
1976 |
+
"step": 3280
|
1977 |
+
},
|
1978 |
+
{
|
1979 |
+
"epoch": 0.11,
|
1980 |
+
"learning_rate": 0.0002,
|
1981 |
+
"loss": 2.8214,
|
1982 |
+
"step": 3290
|
1983 |
+
},
|
1984 |
+
{
|
1985 |
+
"epoch": 0.11,
|
1986 |
+
"learning_rate": 0.0002,
|
1987 |
+
"loss": 2.773,
|
1988 |
+
"step": 3300
|
1989 |
+
},
|
1990 |
+
{
|
1991 |
+
"epoch": 0.11,
|
1992 |
+
"learning_rate": 0.0002,
|
1993 |
+
"loss": 2.8078,
|
1994 |
+
"step": 3310
|
1995 |
+
},
|
1996 |
+
{
|
1997 |
+
"epoch": 0.11,
|
1998 |
+
"learning_rate": 0.0002,
|
1999 |
+
"loss": 2.7599,
|
2000 |
+
"step": 3320
|
2001 |
+
},
|
2002 |
+
{
|
2003 |
+
"epoch": 0.11,
|
2004 |
+
"learning_rate": 0.0002,
|
2005 |
+
"loss": 2.7697,
|
2006 |
+
"step": 3330
|
2007 |
+
},
|
2008 |
+
{
|
2009 |
+
"epoch": 0.11,
|
2010 |
+
"learning_rate": 0.0002,
|
2011 |
+
"loss": 2.7738,
|
2012 |
+
"step": 3340
|
2013 |
+
},
|
2014 |
+
{
|
2015 |
+
"epoch": 0.11,
|
2016 |
+
"learning_rate": 0.0002,
|
2017 |
+
"loss": 2.7751,
|
2018 |
+
"step": 3350
|
2019 |
+
},
|
2020 |
+
{
|
2021 |
+
"epoch": 0.11,
|
2022 |
+
"learning_rate": 0.0002,
|
2023 |
+
"loss": 2.7712,
|
2024 |
+
"step": 3360
|
2025 |
+
},
|
2026 |
+
{
|
2027 |
+
"epoch": 0.11,
|
2028 |
+
"learning_rate": 0.0002,
|
2029 |
+
"loss": 2.7691,
|
2030 |
+
"step": 3370
|
2031 |
+
},
|
2032 |
+
{
|
2033 |
+
"epoch": 0.11,
|
2034 |
+
"learning_rate": 0.0002,
|
2035 |
+
"loss": 2.7893,
|
2036 |
+
"step": 3380
|
2037 |
+
},
|
2038 |
+
{
|
2039 |
+
"epoch": 0.11,
|
2040 |
+
"learning_rate": 0.0002,
|
2041 |
+
"loss": 2.8128,
|
2042 |
+
"step": 3390
|
2043 |
+
},
|
2044 |
+
{
|
2045 |
+
"epoch": 0.11,
|
2046 |
+
"learning_rate": 0.0002,
|
2047 |
+
"loss": 2.779,
|
2048 |
+
"step": 3400
|
2049 |
+
},
|
2050 |
+
{
|
2051 |
+
"epoch": 0.11,
|
2052 |
+
"learning_rate": 0.0002,
|
2053 |
+
"loss": 2.8076,
|
2054 |
+
"step": 3410
|
2055 |
+
},
|
2056 |
+
{
|
2057 |
+
"epoch": 0.11,
|
2058 |
+
"learning_rate": 0.0002,
|
2059 |
+
"loss": 2.7681,
|
2060 |
+
"step": 3420
|
2061 |
+
},
|
2062 |
+
{
|
2063 |
+
"epoch": 0.11,
|
2064 |
+
"learning_rate": 0.0002,
|
2065 |
+
"loss": 2.7177,
|
2066 |
+
"step": 3430
|
2067 |
+
},
|
2068 |
+
{
|
2069 |
+
"epoch": 0.11,
|
2070 |
+
"learning_rate": 0.0002,
|
2071 |
+
"loss": 2.8094,
|
2072 |
+
"step": 3440
|
2073 |
+
},
|
2074 |
+
{
|
2075 |
+
"epoch": 0.11,
|
2076 |
+
"learning_rate": 0.0002,
|
2077 |
+
"loss": 2.7567,
|
2078 |
+
"step": 3450
|
2079 |
+
},
|
2080 |
+
{
|
2081 |
+
"epoch": 0.11,
|
2082 |
+
"learning_rate": 0.0002,
|
2083 |
+
"loss": 2.8128,
|
2084 |
+
"step": 3460
|
2085 |
+
},
|
2086 |
+
{
|
2087 |
+
"epoch": 0.11,
|
2088 |
+
"learning_rate": 0.0002,
|
2089 |
+
"loss": 2.828,
|
2090 |
+
"step": 3470
|
2091 |
+
},
|
2092 |
+
{
|
2093 |
+
"epoch": 0.11,
|
2094 |
+
"learning_rate": 0.0002,
|
2095 |
+
"loss": 2.8377,
|
2096 |
+
"step": 3480
|
2097 |
+
},
|
2098 |
+
{
|
2099 |
+
"epoch": 0.11,
|
2100 |
+
"learning_rate": 0.0002,
|
2101 |
+
"loss": 2.8001,
|
2102 |
+
"step": 3490
|
2103 |
+
},
|
2104 |
+
{
|
2105 |
+
"epoch": 0.11,
|
2106 |
+
"learning_rate": 0.0002,
|
2107 |
+
"loss": 2.79,
|
2108 |
+
"step": 3500
|
2109 |
+
},
|
2110 |
+
{
|
2111 |
+
"epoch": 0.11,
|
2112 |
+
"learning_rate": 0.0002,
|
2113 |
+
"loss": 2.7908,
|
2114 |
+
"step": 3510
|
2115 |
+
},
|
2116 |
+
{
|
2117 |
+
"epoch": 0.11,
|
2118 |
+
"learning_rate": 0.0002,
|
2119 |
+
"loss": 2.7747,
|
2120 |
+
"step": 3520
|
2121 |
+
},
|
2122 |
+
{
|
2123 |
+
"epoch": 0.11,
|
2124 |
+
"learning_rate": 0.0002,
|
2125 |
+
"loss": 2.7864,
|
2126 |
+
"step": 3530
|
2127 |
+
},
|
2128 |
+
{
|
2129 |
+
"epoch": 0.11,
|
2130 |
+
"learning_rate": 0.0002,
|
2131 |
+
"loss": 2.8027,
|
2132 |
+
"step": 3540
|
2133 |
+
},
|
2134 |
+
{
|
2135 |
+
"epoch": 0.11,
|
2136 |
+
"learning_rate": 0.0002,
|
2137 |
+
"loss": 2.7764,
|
2138 |
+
"step": 3550
|
2139 |
+
},
|
2140 |
+
{
|
2141 |
+
"epoch": 0.11,
|
2142 |
+
"learning_rate": 0.0002,
|
2143 |
+
"loss": 2.7274,
|
2144 |
+
"step": 3560
|
2145 |
+
},
|
2146 |
+
{
|
2147 |
+
"epoch": 0.12,
|
2148 |
+
"learning_rate": 0.0002,
|
2149 |
+
"loss": 2.7617,
|
2150 |
+
"step": 3570
|
2151 |
+
},
|
2152 |
+
{
|
2153 |
+
"epoch": 0.12,
|
2154 |
+
"learning_rate": 0.0002,
|
2155 |
+
"loss": 2.8357,
|
2156 |
+
"step": 3580
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 0.12,
|
2160 |
+
"learning_rate": 0.0002,
|
2161 |
+
"loss": 2.7528,
|
2162 |
+
"step": 3590
|
2163 |
+
},
|
2164 |
+
{
|
2165 |
+
"epoch": 0.12,
|
2166 |
+
"learning_rate": 0.0002,
|
2167 |
+
"loss": 2.7862,
|
2168 |
+
"step": 3600
|
2169 |
+
},
|
2170 |
+
{
|
2171 |
+
"epoch": 0.12,
|
2172 |
+
"learning_rate": 0.0002,
|
2173 |
+
"loss": 2.7632,
|
2174 |
+
"step": 3610
|
2175 |
+
},
|
2176 |
+
{
|
2177 |
+
"epoch": 0.12,
|
2178 |
+
"learning_rate": 0.0002,
|
2179 |
+
"loss": 2.777,
|
2180 |
+
"step": 3620
|
2181 |
+
},
|
2182 |
+
{
|
2183 |
+
"epoch": 0.12,
|
2184 |
+
"learning_rate": 0.0002,
|
2185 |
+
"loss": 2.7444,
|
2186 |
+
"step": 3630
|
2187 |
+
},
|
2188 |
+
{
|
2189 |
+
"epoch": 0.12,
|
2190 |
+
"learning_rate": 0.0002,
|
2191 |
+
"loss": 2.8136,
|
2192 |
+
"step": 3640
|
2193 |
+
},
|
2194 |
+
{
|
2195 |
+
"epoch": 0.12,
|
2196 |
+
"learning_rate": 0.0002,
|
2197 |
+
"loss": 2.7451,
|
2198 |
+
"step": 3650
|
2199 |
+
},
|
2200 |
+
{
|
2201 |
+
"epoch": 0.12,
|
2202 |
+
"learning_rate": 0.0002,
|
2203 |
+
"loss": 2.7554,
|
2204 |
+
"step": 3660
|
2205 |
+
},
|
2206 |
+
{
|
2207 |
+
"epoch": 0.12,
|
2208 |
+
"learning_rate": 0.0002,
|
2209 |
+
"loss": 2.7968,
|
2210 |
+
"step": 3670
|
2211 |
+
},
|
2212 |
+
{
|
2213 |
+
"epoch": 0.12,
|
2214 |
+
"learning_rate": 0.0002,
|
2215 |
+
"loss": 2.7819,
|
2216 |
+
"step": 3680
|
2217 |
+
},
|
2218 |
+
{
|
2219 |
+
"epoch": 0.12,
|
2220 |
+
"learning_rate": 0.0002,
|
2221 |
+
"loss": 2.7801,
|
2222 |
+
"step": 3690
|
2223 |
+
},
|
2224 |
+
{
|
2225 |
+
"epoch": 0.12,
|
2226 |
+
"learning_rate": 0.0002,
|
2227 |
+
"loss": 2.7972,
|
2228 |
+
"step": 3700
|
2229 |
+
},
|
2230 |
+
{
|
2231 |
+
"epoch": 0.12,
|
2232 |
+
"learning_rate": 0.0002,
|
2233 |
+
"loss": 2.7974,
|
2234 |
+
"step": 3710
|
2235 |
+
},
|
2236 |
+
{
|
2237 |
+
"epoch": 0.12,
|
2238 |
+
"learning_rate": 0.0002,
|
2239 |
+
"loss": 2.7924,
|
2240 |
+
"step": 3720
|
2241 |
+
},
|
2242 |
+
{
|
2243 |
+
"epoch": 0.12,
|
2244 |
+
"learning_rate": 0.0002,
|
2245 |
+
"loss": 2.8114,
|
2246 |
+
"step": 3730
|
2247 |
+
},
|
2248 |
+
{
|
2249 |
+
"epoch": 0.12,
|
2250 |
+
"learning_rate": 0.0002,
|
2251 |
+
"loss": 2.7964,
|
2252 |
+
"step": 3740
|
2253 |
+
},
|
2254 |
+
{
|
2255 |
+
"epoch": 0.12,
|
2256 |
+
"learning_rate": 0.0002,
|
2257 |
+
"loss": 2.7775,
|
2258 |
+
"step": 3750
|
2259 |
+
},
|
2260 |
+
{
|
2261 |
+
"epoch": 0.12,
|
2262 |
+
"learning_rate": 0.0002,
|
2263 |
+
"loss": 2.7402,
|
2264 |
+
"step": 3760
|
2265 |
+
},
|
2266 |
+
{
|
2267 |
+
"epoch": 0.12,
|
2268 |
+
"learning_rate": 0.0002,
|
2269 |
+
"loss": 2.7515,
|
2270 |
+
"step": 3770
|
2271 |
+
},
|
2272 |
+
{
|
2273 |
+
"epoch": 0.12,
|
2274 |
+
"learning_rate": 0.0002,
|
2275 |
+
"loss": 2.7856,
|
2276 |
+
"step": 3780
|
2277 |
+
},
|
2278 |
+
{
|
2279 |
+
"epoch": 0.12,
|
2280 |
+
"learning_rate": 0.0002,
|
2281 |
+
"loss": 2.7761,
|
2282 |
+
"step": 3790
|
2283 |
+
},
|
2284 |
+
{
|
2285 |
+
"epoch": 0.12,
|
2286 |
+
"learning_rate": 0.0002,
|
2287 |
+
"loss": 2.8069,
|
2288 |
+
"step": 3800
|
2289 |
+
},
|
2290 |
+
{
|
2291 |
+
"epoch": 0.12,
|
2292 |
+
"learning_rate": 0.0002,
|
2293 |
+
"loss": 2.7905,
|
2294 |
+
"step": 3810
|
2295 |
+
},
|
2296 |
+
{
|
2297 |
+
"epoch": 0.12,
|
2298 |
+
"learning_rate": 0.0002,
|
2299 |
+
"loss": 2.7477,
|
2300 |
+
"step": 3820
|
2301 |
+
},
|
2302 |
+
{
|
2303 |
+
"epoch": 0.12,
|
2304 |
+
"learning_rate": 0.0002,
|
2305 |
+
"loss": 2.7984,
|
2306 |
+
"step": 3830
|
2307 |
+
},
|
2308 |
+
{
|
2309 |
+
"epoch": 0.12,
|
2310 |
+
"learning_rate": 0.0002,
|
2311 |
+
"loss": 2.7684,
|
2312 |
+
"step": 3840
|
2313 |
+
},
|
2314 |
+
{
|
2315 |
+
"epoch": 0.12,
|
2316 |
+
"learning_rate": 0.0002,
|
2317 |
+
"loss": 2.7882,
|
2318 |
+
"step": 3850
|
2319 |
+
},
|
2320 |
+
{
|
2321 |
+
"epoch": 0.12,
|
2322 |
+
"learning_rate": 0.0002,
|
2323 |
+
"loss": 2.7912,
|
2324 |
+
"step": 3860
|
2325 |
+
},
|
2326 |
+
{
|
2327 |
+
"epoch": 0.12,
|
2328 |
+
"learning_rate": 0.0002,
|
2329 |
+
"loss": 2.7461,
|
2330 |
+
"step": 3870
|
2331 |
+
},
|
2332 |
+
{
|
2333 |
+
"epoch": 0.13,
|
2334 |
+
"learning_rate": 0.0002,
|
2335 |
+
"loss": 2.747,
|
2336 |
+
"step": 3880
|
2337 |
+
},
|
2338 |
+
{
|
2339 |
+
"epoch": 0.13,
|
2340 |
+
"learning_rate": 0.0002,
|
2341 |
+
"loss": 2.7608,
|
2342 |
+
"step": 3890
|
2343 |
+
},
|
2344 |
+
{
|
2345 |
+
"epoch": 0.13,
|
2346 |
+
"learning_rate": 0.0002,
|
2347 |
+
"loss": 2.7469,
|
2348 |
+
"step": 3900
|
2349 |
+
},
|
2350 |
+
{
|
2351 |
+
"epoch": 0.13,
|
2352 |
+
"learning_rate": 0.0002,
|
2353 |
+
"loss": 2.7701,
|
2354 |
+
"step": 3910
|
2355 |
+
},
|
2356 |
+
{
|
2357 |
+
"epoch": 0.13,
|
2358 |
+
"learning_rate": 0.0002,
|
2359 |
+
"loss": 2.7944,
|
2360 |
+
"step": 3920
|
2361 |
+
},
|
2362 |
+
{
|
2363 |
+
"epoch": 0.13,
|
2364 |
+
"learning_rate": 0.0002,
|
2365 |
+
"loss": 2.7547,
|
2366 |
+
"step": 3930
|
2367 |
+
},
|
2368 |
+
{
|
2369 |
+
"epoch": 0.13,
|
2370 |
+
"learning_rate": 0.0002,
|
2371 |
+
"loss": 2.7536,
|
2372 |
+
"step": 3940
|
2373 |
+
},
|
2374 |
+
{
|
2375 |
+
"epoch": 0.13,
|
2376 |
+
"learning_rate": 0.0002,
|
2377 |
+
"loss": 2.7743,
|
2378 |
+
"step": 3950
|
2379 |
+
},
|
2380 |
+
{
|
2381 |
+
"epoch": 0.13,
|
2382 |
+
"learning_rate": 0.0002,
|
2383 |
+
"loss": 2.8194,
|
2384 |
+
"step": 3960
|
2385 |
+
},
|
2386 |
+
{
|
2387 |
+
"epoch": 0.13,
|
2388 |
+
"learning_rate": 0.0002,
|
2389 |
+
"loss": 2.7548,
|
2390 |
+
"step": 3970
|
2391 |
+
},
|
2392 |
+
{
|
2393 |
+
"epoch": 0.13,
|
2394 |
+
"learning_rate": 0.0002,
|
2395 |
+
"loss": 2.7768,
|
2396 |
+
"step": 3980
|
2397 |
+
},
|
2398 |
+
{
|
2399 |
+
"epoch": 0.13,
|
2400 |
+
"learning_rate": 0.0002,
|
2401 |
+
"loss": 2.832,
|
2402 |
+
"step": 3990
|
2403 |
+
},
|
2404 |
+
{
|
2405 |
+
"epoch": 0.13,
|
2406 |
+
"learning_rate": 0.0002,
|
2407 |
+
"loss": 2.7862,
|
2408 |
+
"step": 4000
|
2409 |
+
}
|
2410 |
+
],
|
2411 |
+
"max_steps": 30984,
|
2412 |
+
"num_train_epochs": 1,
|
2413 |
+
"total_flos": 2.7083329329168384e+18,
|
2414 |
+
"trial_name": null,
|
2415 |
+
"trial_params": null
|
2416 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:09b2a027ff3eeefff52709052ca1b382e9023fdb0fe1e9ea0eea22ce841ca622
|
3 |
+
size 3963
|