upload model
Browse files- .gitattributes +1 -0
- config.json +41 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +15 -0
- tokenizer.json +3 -0
- tokenizer_config.json +21 -0
- trainer_state.json +2233 -0
- training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "models/mnli/checkpoint-36813",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"finetuning_task": "xnli",
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": "LABEL_0",
|
| 17 |
+
"1": "LABEL_1",
|
| 18 |
+
"2": "LABEL_2"
|
| 19 |
+
},
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"LABEL_0": 0,
|
| 24 |
+
"LABEL_1": 1,
|
| 25 |
+
"LABEL_2": 2
|
| 26 |
+
},
|
| 27 |
+
"layer_norm_eps": 1e-05,
|
| 28 |
+
"max_position_embeddings": 514,
|
| 29 |
+
"model_type": "xlm-roberta",
|
| 30 |
+
"num_attention_heads": 12,
|
| 31 |
+
"num_hidden_layers": 12,
|
| 32 |
+
"output_past": true,
|
| 33 |
+
"pad_token_id": 1,
|
| 34 |
+
"position_embedding_type": "absolute",
|
| 35 |
+
"problem_type": "single_label_classification",
|
| 36 |
+
"torch_dtype": "float32",
|
| 37 |
+
"transformers_version": "4.25.1",
|
| 38 |
+
"type_vocab_size": 1,
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 250002
|
| 41 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f32c1ec8789fad33da5cfc212a2385248abcb1fe0ca3790f46f4e3bc9280096a
|
| 3 |
+
size 2224485957
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5158a8c0d0c307ea675880d019acd426a123a82018b963b1e6d5e4c91b72e97c
|
| 3 |
+
size 1112257205
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9928046d48a1a4638d52e99b662a6db6b219b0d33222091b555801c990f48d57
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a70786462a4e11928319ba37525db78de03ec314d18a3153c350827552bd1f1
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"content": "<mask>",
|
| 7 |
+
"lstrip": true,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"sep_token": "</s>",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c08c80d1df11b82ada2fd707562f86a9ebd5b7de04f51ebd2b49f2cd5906d00
|
| 3 |
+
size 17082925
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"do_lower_case": false,
|
| 5 |
+
"eos_token": "</s>",
|
| 6 |
+
"mask_token": {
|
| 7 |
+
"__type": "AddedToken",
|
| 8 |
+
"content": "<mask>",
|
| 9 |
+
"lstrip": true,
|
| 10 |
+
"normalized": true,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"single_word": false
|
| 13 |
+
},
|
| 14 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 15 |
+
"name_or_path": "models/mnli/checkpoint-36813",
|
| 16 |
+
"pad_token": "<pad>",
|
| 17 |
+
"sep_token": "</s>",
|
| 18 |
+
"special_tokens_map_file": null,
|
| 19 |
+
"tokenizer_class": "XLMRobertaTokenizer",
|
| 20 |
+
"unk_token": "<unk>"
|
| 21 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2233 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.587454617023468,
|
| 3 |
+
"best_model_checkpoint": "models/mnli_xnli_shuff_all/checkpoint-184017",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 184017,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 1.9945657194715707e-05,
|
| 13 |
+
"loss": 0.796,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 1.9891314389431413e-05,
|
| 19 |
+
"loss": 0.7802,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.9836971584147118e-05,
|
| 25 |
+
"loss": 0.7726,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01,
|
| 30 |
+
"learning_rate": 1.9782628778862823e-05,
|
| 31 |
+
"loss": 0.7543,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.01,
|
| 36 |
+
"learning_rate": 1.972828597357853e-05,
|
| 37 |
+
"loss": 0.7571,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.02,
|
| 42 |
+
"learning_rate": 1.9673943168294238e-05,
|
| 43 |
+
"loss": 0.7342,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02,
|
| 48 |
+
"learning_rate": 1.961960036300994e-05,
|
| 49 |
+
"loss": 0.7401,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.02,
|
| 54 |
+
"learning_rate": 1.956525755772565e-05,
|
| 55 |
+
"loss": 0.7427,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.02,
|
| 60 |
+
"learning_rate": 1.951091475244135e-05,
|
| 61 |
+
"loss": 0.7412,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.03,
|
| 66 |
+
"learning_rate": 1.9456571947157056e-05,
|
| 67 |
+
"loss": 0.7342,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.03,
|
| 72 |
+
"learning_rate": 1.9402229141872765e-05,
|
| 73 |
+
"loss": 0.7286,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.03,
|
| 78 |
+
"learning_rate": 1.9347886336588467e-05,
|
| 79 |
+
"loss": 0.731,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.04,
|
| 84 |
+
"learning_rate": 1.9293543531304176e-05,
|
| 85 |
+
"loss": 0.7345,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.04,
|
| 90 |
+
"learning_rate": 1.923920072601988e-05,
|
| 91 |
+
"loss": 0.7233,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.04,
|
| 96 |
+
"learning_rate": 1.9184857920735587e-05,
|
| 97 |
+
"loss": 0.7177,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.04,
|
| 102 |
+
"learning_rate": 1.9130515115451292e-05,
|
| 103 |
+
"loss": 0.7239,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.05,
|
| 108 |
+
"learning_rate": 1.9076172310166994e-05,
|
| 109 |
+
"loss": 0.714,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.05,
|
| 114 |
+
"learning_rate": 1.9021829504882703e-05,
|
| 115 |
+
"loss": 0.7209,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.05,
|
| 120 |
+
"learning_rate": 1.896748669959841e-05,
|
| 121 |
+
"loss": 0.7191,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.05,
|
| 126 |
+
"learning_rate": 1.8913143894314114e-05,
|
| 127 |
+
"loss": 0.7079,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.06,
|
| 132 |
+
"learning_rate": 1.885880108902982e-05,
|
| 133 |
+
"loss": 0.7131,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.06,
|
| 138 |
+
"learning_rate": 1.8804458283745525e-05,
|
| 139 |
+
"loss": 0.7133,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.06,
|
| 144 |
+
"learning_rate": 1.875011547846123e-05,
|
| 145 |
+
"loss": 0.702,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.07,
|
| 150 |
+
"learning_rate": 1.8695772673176936e-05,
|
| 151 |
+
"loss": 0.6985,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.07,
|
| 156 |
+
"learning_rate": 1.864142986789264e-05,
|
| 157 |
+
"loss": 0.7004,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.07,
|
| 162 |
+
"learning_rate": 1.8587087062608347e-05,
|
| 163 |
+
"loss": 0.7062,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.07,
|
| 168 |
+
"learning_rate": 1.8532744257324052e-05,
|
| 169 |
+
"loss": 0.6984,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.08,
|
| 174 |
+
"learning_rate": 1.8478401452039758e-05,
|
| 175 |
+
"loss": 0.693,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.08,
|
| 180 |
+
"learning_rate": 1.8424058646755467e-05,
|
| 181 |
+
"loss": 0.697,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.08,
|
| 186 |
+
"learning_rate": 1.836971584147117e-05,
|
| 187 |
+
"loss": 0.703,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.08,
|
| 192 |
+
"learning_rate": 1.8315373036186878e-05,
|
| 193 |
+
"loss": 0.6923,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.09,
|
| 198 |
+
"learning_rate": 1.826103023090258e-05,
|
| 199 |
+
"loss": 0.6837,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.09,
|
| 204 |
+
"learning_rate": 1.8206687425618285e-05,
|
| 205 |
+
"loss": 0.7035,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.09,
|
| 210 |
+
"learning_rate": 1.8152344620333994e-05,
|
| 211 |
+
"loss": 0.687,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.1,
|
| 216 |
+
"learning_rate": 1.8098001815049696e-05,
|
| 217 |
+
"loss": 0.6872,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.1,
|
| 222 |
+
"learning_rate": 1.8043659009765405e-05,
|
| 223 |
+
"loss": 0.6864,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.1,
|
| 228 |
+
"learning_rate": 1.798931620448111e-05,
|
| 229 |
+
"loss": 0.6825,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.1,
|
| 234 |
+
"learning_rate": 1.7934973399196816e-05,
|
| 235 |
+
"loss": 0.6864,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.11,
|
| 240 |
+
"learning_rate": 1.788063059391252e-05,
|
| 241 |
+
"loss": 0.6834,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.11,
|
| 246 |
+
"learning_rate": 1.7826287788628227e-05,
|
| 247 |
+
"loss": 0.6794,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.11,
|
| 252 |
+
"learning_rate": 1.7771944983343932e-05,
|
| 253 |
+
"loss": 0.676,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.11,
|
| 258 |
+
"learning_rate": 1.7717602178059638e-05,
|
| 259 |
+
"loss": 0.6906,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.12,
|
| 264 |
+
"learning_rate": 1.7663259372775343e-05,
|
| 265 |
+
"loss": 0.6764,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.12,
|
| 270 |
+
"learning_rate": 1.760891656749105e-05,
|
| 271 |
+
"loss": 0.6764,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.12,
|
| 276 |
+
"learning_rate": 1.7554573762206754e-05,
|
| 277 |
+
"loss": 0.6833,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.12,
|
| 282 |
+
"learning_rate": 1.750023095692246e-05,
|
| 283 |
+
"loss": 0.6712,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.13,
|
| 288 |
+
"learning_rate": 1.7445888151638165e-05,
|
| 289 |
+
"loss": 0.6811,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.13,
|
| 294 |
+
"learning_rate": 1.739154534635387e-05,
|
| 295 |
+
"loss": 0.6707,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.13,
|
| 300 |
+
"learning_rate": 1.7337202541069576e-05,
|
| 301 |
+
"loss": 0.6749,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.14,
|
| 306 |
+
"learning_rate": 1.728285973578528e-05,
|
| 307 |
+
"loss": 0.6683,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.14,
|
| 312 |
+
"learning_rate": 1.7228516930500987e-05,
|
| 313 |
+
"loss": 0.6689,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.14,
|
| 318 |
+
"learning_rate": 1.7174174125216695e-05,
|
| 319 |
+
"loss": 0.6736,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.14,
|
| 324 |
+
"learning_rate": 1.7119831319932398e-05,
|
| 325 |
+
"loss": 0.6724,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.15,
|
| 330 |
+
"learning_rate": 1.7065488514648106e-05,
|
| 331 |
+
"loss": 0.6737,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.15,
|
| 336 |
+
"learning_rate": 1.701114570936381e-05,
|
| 337 |
+
"loss": 0.6631,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.15,
|
| 342 |
+
"learning_rate": 1.6956802904079517e-05,
|
| 343 |
+
"loss": 0.6861,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.15,
|
| 348 |
+
"learning_rate": 1.6902460098795223e-05,
|
| 349 |
+
"loss": 0.6749,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.16,
|
| 354 |
+
"learning_rate": 1.6848117293510925e-05,
|
| 355 |
+
"loss": 0.6619,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.16,
|
| 360 |
+
"learning_rate": 1.6793774488226634e-05,
|
| 361 |
+
"loss": 0.662,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.16,
|
| 366 |
+
"learning_rate": 1.673943168294234e-05,
|
| 367 |
+
"loss": 0.6522,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.17,
|
| 372 |
+
"learning_rate": 1.6685088877658045e-05,
|
| 373 |
+
"loss": 0.671,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.17,
|
| 378 |
+
"learning_rate": 1.663074607237375e-05,
|
| 379 |
+
"loss": 0.6617,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.17,
|
| 384 |
+
"learning_rate": 1.6576403267089455e-05,
|
| 385 |
+
"loss": 0.658,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.17,
|
| 390 |
+
"learning_rate": 1.652206046180516e-05,
|
| 391 |
+
"loss": 0.6588,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.18,
|
| 396 |
+
"learning_rate": 1.6467717656520866e-05,
|
| 397 |
+
"loss": 0.6615,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.18,
|
| 402 |
+
"learning_rate": 1.6413374851236572e-05,
|
| 403 |
+
"loss": 0.6628,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.18,
|
| 408 |
+
"learning_rate": 1.6359032045952277e-05,
|
| 409 |
+
"loss": 0.6513,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.18,
|
| 414 |
+
"learning_rate": 1.6304689240667983e-05,
|
| 415 |
+
"loss": 0.6586,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.19,
|
| 420 |
+
"learning_rate": 1.6250346435383688e-05,
|
| 421 |
+
"loss": 0.6491,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.19,
|
| 426 |
+
"learning_rate": 1.6196003630099394e-05,
|
| 427 |
+
"loss": 0.6708,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.19,
|
| 432 |
+
"learning_rate": 1.61416608248151e-05,
|
| 433 |
+
"loss": 0.6565,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.2,
|
| 438 |
+
"learning_rate": 1.6087318019530808e-05,
|
| 439 |
+
"loss": 0.6525,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.2,
|
| 444 |
+
"learning_rate": 1.603297521424651e-05,
|
| 445 |
+
"loss": 0.6503,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.2,
|
| 450 |
+
"learning_rate": 1.5978632408962215e-05,
|
| 451 |
+
"loss": 0.6465,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.2,
|
| 456 |
+
"learning_rate": 1.5924289603677924e-05,
|
| 457 |
+
"loss": 0.6477,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.21,
|
| 462 |
+
"learning_rate": 1.5869946798393626e-05,
|
| 463 |
+
"loss": 0.6473,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.21,
|
| 468 |
+
"learning_rate": 1.5815603993109335e-05,
|
| 469 |
+
"loss": 0.6494,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.21,
|
| 474 |
+
"learning_rate": 1.5761261187825037e-05,
|
| 475 |
+
"loss": 0.6371,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.21,
|
| 480 |
+
"learning_rate": 1.5706918382540746e-05,
|
| 481 |
+
"loss": 0.6434,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.22,
|
| 486 |
+
"learning_rate": 1.565257557725645e-05,
|
| 487 |
+
"loss": 0.645,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.22,
|
| 492 |
+
"learning_rate": 1.5598232771972157e-05,
|
| 493 |
+
"loss": 0.65,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.22,
|
| 498 |
+
"learning_rate": 1.5543889966687863e-05,
|
| 499 |
+
"loss": 0.6432,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.23,
|
| 504 |
+
"learning_rate": 1.5489547161403568e-05,
|
| 505 |
+
"loss": 0.6437,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.23,
|
| 510 |
+
"learning_rate": 1.5435204356119273e-05,
|
| 511 |
+
"loss": 0.6297,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.23,
|
| 516 |
+
"learning_rate": 1.538086155083498e-05,
|
| 517 |
+
"loss": 0.6376,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.23,
|
| 522 |
+
"learning_rate": 1.5326518745550684e-05,
|
| 523 |
+
"loss": 0.6439,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.24,
|
| 528 |
+
"learning_rate": 1.527217594026639e-05,
|
| 529 |
+
"loss": 0.6435,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.24,
|
| 534 |
+
"learning_rate": 1.5217833134982097e-05,
|
| 535 |
+
"loss": 0.6474,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.24,
|
| 540 |
+
"learning_rate": 1.51634903296978e-05,
|
| 541 |
+
"loss": 0.6356,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.24,
|
| 546 |
+
"learning_rate": 1.5109147524413506e-05,
|
| 547 |
+
"loss": 0.6386,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.25,
|
| 552 |
+
"learning_rate": 1.5054804719129212e-05,
|
| 553 |
+
"loss": 0.6349,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.25,
|
| 558 |
+
"learning_rate": 1.5000461913844917e-05,
|
| 559 |
+
"loss": 0.6333,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.25,
|
| 564 |
+
"learning_rate": 1.4946119108560624e-05,
|
| 565 |
+
"loss": 0.6439,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.26,
|
| 570 |
+
"learning_rate": 1.4891776303276328e-05,
|
| 571 |
+
"loss": 0.6334,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.26,
|
| 576 |
+
"learning_rate": 1.4837433497992035e-05,
|
| 577 |
+
"loss": 0.6384,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.26,
|
| 582 |
+
"learning_rate": 1.478309069270774e-05,
|
| 583 |
+
"loss": 0.6481,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.26,
|
| 588 |
+
"learning_rate": 1.4728747887423446e-05,
|
| 589 |
+
"loss": 0.6308,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.27,
|
| 594 |
+
"learning_rate": 1.4674405082139151e-05,
|
| 595 |
+
"loss": 0.6403,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.27,
|
| 600 |
+
"learning_rate": 1.4620062276854857e-05,
|
| 601 |
+
"loss": 0.6347,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.27,
|
| 606 |
+
"learning_rate": 1.4565719471570562e-05,
|
| 607 |
+
"loss": 0.6394,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.27,
|
| 612 |
+
"learning_rate": 1.4511376666286268e-05,
|
| 613 |
+
"loss": 0.6266,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.28,
|
| 618 |
+
"learning_rate": 1.4457033861001975e-05,
|
| 619 |
+
"loss": 0.6298,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.28,
|
| 624 |
+
"learning_rate": 1.4402691055717679e-05,
|
| 625 |
+
"loss": 0.6187,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.28,
|
| 630 |
+
"learning_rate": 1.4348348250433386e-05,
|
| 631 |
+
"loss": 0.6278,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.29,
|
| 636 |
+
"learning_rate": 1.4294005445149091e-05,
|
| 637 |
+
"loss": 0.6331,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.29,
|
| 642 |
+
"learning_rate": 1.4239662639864795e-05,
|
| 643 |
+
"loss": 0.6359,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.29,
|
| 648 |
+
"learning_rate": 1.4185319834580502e-05,
|
| 649 |
+
"loss": 0.6297,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.29,
|
| 654 |
+
"learning_rate": 1.4130977029296206e-05,
|
| 655 |
+
"loss": 0.6307,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.3,
|
| 660 |
+
"learning_rate": 1.4076634224011913e-05,
|
| 661 |
+
"loss": 0.6297,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.3,
|
| 666 |
+
"learning_rate": 1.4022291418727619e-05,
|
| 667 |
+
"loss": 0.6318,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.3,
|
| 672 |
+
"learning_rate": 1.3967948613443326e-05,
|
| 673 |
+
"loss": 0.6346,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.3,
|
| 678 |
+
"learning_rate": 1.391360580815903e-05,
|
| 679 |
+
"loss": 0.6179,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.31,
|
| 684 |
+
"learning_rate": 1.3859263002874737e-05,
|
| 685 |
+
"loss": 0.6215,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.31,
|
| 690 |
+
"learning_rate": 1.380492019759044e-05,
|
| 691 |
+
"loss": 0.6269,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.31,
|
| 696 |
+
"learning_rate": 1.3750577392306146e-05,
|
| 697 |
+
"loss": 0.6204,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.32,
|
| 702 |
+
"learning_rate": 1.3696234587021853e-05,
|
| 703 |
+
"loss": 0.6297,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.32,
|
| 708 |
+
"learning_rate": 1.3641891781737557e-05,
|
| 709 |
+
"loss": 0.6206,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.32,
|
| 714 |
+
"learning_rate": 1.3587548976453264e-05,
|
| 715 |
+
"loss": 0.6251,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.32,
|
| 720 |
+
"learning_rate": 1.353320617116897e-05,
|
| 721 |
+
"loss": 0.6268,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.33,
|
| 726 |
+
"learning_rate": 1.3478863365884675e-05,
|
| 727 |
+
"loss": 0.6182,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.33,
|
| 732 |
+
"learning_rate": 1.342452056060038e-05,
|
| 733 |
+
"loss": 0.6201,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.33,
|
| 738 |
+
"learning_rate": 1.3370177755316086e-05,
|
| 739 |
+
"loss": 0.6138,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.33,
|
| 744 |
+
"learning_rate": 1.3315834950031791e-05,
|
| 745 |
+
"loss": 0.6241,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.34,
|
| 750 |
+
"learning_rate": 1.3261492144747497e-05,
|
| 751 |
+
"loss": 0.6134,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.34,
|
| 756 |
+
"learning_rate": 1.3207149339463204e-05,
|
| 757 |
+
"loss": 0.6235,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.34,
|
| 762 |
+
"learning_rate": 1.3152806534178908e-05,
|
| 763 |
+
"loss": 0.6065,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.35,
|
| 768 |
+
"learning_rate": 1.3098463728894615e-05,
|
| 769 |
+
"loss": 0.6088,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.35,
|
| 774 |
+
"learning_rate": 1.304412092361032e-05,
|
| 775 |
+
"loss": 0.612,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.35,
|
| 780 |
+
"learning_rate": 1.2989778118326026e-05,
|
| 781 |
+
"loss": 0.6185,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.35,
|
| 786 |
+
"learning_rate": 1.2935435313041731e-05,
|
| 787 |
+
"loss": 0.6032,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.36,
|
| 792 |
+
"learning_rate": 1.2881092507757435e-05,
|
| 793 |
+
"loss": 0.6124,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.36,
|
| 798 |
+
"learning_rate": 1.2826749702473142e-05,
|
| 799 |
+
"loss": 0.6094,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.36,
|
| 804 |
+
"learning_rate": 1.2772406897188848e-05,
|
| 805 |
+
"loss": 0.6005,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.36,
|
| 810 |
+
"learning_rate": 1.2718064091904555e-05,
|
| 811 |
+
"loss": 0.6132,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.37,
|
| 816 |
+
"learning_rate": 1.2663721286620258e-05,
|
| 817 |
+
"loss": 0.6124,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.37,
|
| 822 |
+
"learning_rate": 1.2609378481335966e-05,
|
| 823 |
+
"loss": 0.6142,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.37,
|
| 828 |
+
"learning_rate": 1.255503567605167e-05,
|
| 829 |
+
"loss": 0.6104,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.37,
|
| 834 |
+
"learning_rate": 1.2500692870767375e-05,
|
| 835 |
+
"loss": 0.6183,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.38,
|
| 840 |
+
"learning_rate": 1.2446350065483082e-05,
|
| 841 |
+
"loss": 0.607,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.38,
|
| 846 |
+
"learning_rate": 1.2392007260198786e-05,
|
| 847 |
+
"loss": 0.5969,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.38,
|
| 852 |
+
"learning_rate": 1.2337664454914493e-05,
|
| 853 |
+
"loss": 0.6052,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.39,
|
| 858 |
+
"learning_rate": 1.2283321649630198e-05,
|
| 859 |
+
"loss": 0.613,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.39,
|
| 864 |
+
"learning_rate": 1.2228978844345904e-05,
|
| 865 |
+
"loss": 0.5975,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.39,
|
| 870 |
+
"learning_rate": 1.217463603906161e-05,
|
| 871 |
+
"loss": 0.5998,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.39,
|
| 876 |
+
"learning_rate": 1.2120293233777316e-05,
|
| 877 |
+
"loss": 0.5949,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.4,
|
| 882 |
+
"learning_rate": 1.206595042849302e-05,
|
| 883 |
+
"loss": 0.6029,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.4,
|
| 888 |
+
"learning_rate": 1.2011607623208726e-05,
|
| 889 |
+
"loss": 0.6074,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.4,
|
| 894 |
+
"learning_rate": 1.1957264817924433e-05,
|
| 895 |
+
"loss": 0.5985,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.4,
|
| 900 |
+
"learning_rate": 1.1902922012640136e-05,
|
| 901 |
+
"loss": 0.6105,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.41,
|
| 906 |
+
"learning_rate": 1.1848579207355844e-05,
|
| 907 |
+
"loss": 0.6064,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.41,
|
| 912 |
+
"learning_rate": 1.1794236402071549e-05,
|
| 913 |
+
"loss": 0.5912,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.41,
|
| 918 |
+
"learning_rate": 1.1739893596787255e-05,
|
| 919 |
+
"loss": 0.6117,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.42,
|
| 924 |
+
"learning_rate": 1.168555079150296e-05,
|
| 925 |
+
"loss": 0.5947,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.42,
|
| 930 |
+
"learning_rate": 1.1631207986218664e-05,
|
| 931 |
+
"loss": 0.5985,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.42,
|
| 936 |
+
"learning_rate": 1.1576865180934371e-05,
|
| 937 |
+
"loss": 0.6108,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.42,
|
| 942 |
+
"learning_rate": 1.1522522375650076e-05,
|
| 943 |
+
"loss": 0.5972,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.43,
|
| 948 |
+
"learning_rate": 1.1468179570365784e-05,
|
| 949 |
+
"loss": 0.6002,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.43,
|
| 954 |
+
"learning_rate": 1.1413836765081487e-05,
|
| 955 |
+
"loss": 0.5921,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.43,
|
| 960 |
+
"learning_rate": 1.1359493959797194e-05,
|
| 961 |
+
"loss": 0.6026,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.43,
|
| 966 |
+
"learning_rate": 1.1305151154512898e-05,
|
| 967 |
+
"loss": 0.603,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.44,
|
| 972 |
+
"learning_rate": 1.1250808349228605e-05,
|
| 973 |
+
"loss": 0.5963,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.44,
|
| 978 |
+
"learning_rate": 1.119646554394431e-05,
|
| 979 |
+
"loss": 0.5942,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.44,
|
| 984 |
+
"learning_rate": 1.1142122738660015e-05,
|
| 985 |
+
"loss": 0.6006,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.45,
|
| 990 |
+
"learning_rate": 1.1087779933375722e-05,
|
| 991 |
+
"loss": 0.6026,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.45,
|
| 996 |
+
"learning_rate": 1.1033437128091427e-05,
|
| 997 |
+
"loss": 0.5944,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.45,
|
| 1002 |
+
"learning_rate": 1.0979094322807133e-05,
|
| 1003 |
+
"loss": 0.6031,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.45,
|
| 1008 |
+
"learning_rate": 1.0924751517522838e-05,
|
| 1009 |
+
"loss": 0.5906,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.46,
|
| 1014 |
+
"learning_rate": 1.0870408712238545e-05,
|
| 1015 |
+
"loss": 0.595,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.46,
|
| 1020 |
+
"learning_rate": 1.0816065906954249e-05,
|
| 1021 |
+
"loss": 0.5921,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.46,
|
| 1026 |
+
"learning_rate": 1.0761723101669956e-05,
|
| 1027 |
+
"loss": 0.5944,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.46,
|
| 1032 |
+
"learning_rate": 1.0707380296385662e-05,
|
| 1033 |
+
"loss": 0.5752,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.47,
|
| 1038 |
+
"learning_rate": 1.0653037491101365e-05,
|
| 1039 |
+
"loss": 0.5942,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.47,
|
| 1044 |
+
"learning_rate": 1.0598694685817072e-05,
|
| 1045 |
+
"loss": 0.5946,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.47,
|
| 1050 |
+
"learning_rate": 1.0544351880532778e-05,
|
| 1051 |
+
"loss": 0.5911,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.48,
|
| 1056 |
+
"learning_rate": 1.0490009075248483e-05,
|
| 1057 |
+
"loss": 0.5982,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.48,
|
| 1062 |
+
"learning_rate": 1.0435666269964189e-05,
|
| 1063 |
+
"loss": 0.5941,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.48,
|
| 1068 |
+
"learning_rate": 1.0381323464679896e-05,
|
| 1069 |
+
"loss": 0.598,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.48,
|
| 1074 |
+
"learning_rate": 1.03269806593956e-05,
|
| 1075 |
+
"loss": 0.5845,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.49,
|
| 1080 |
+
"learning_rate": 1.0272637854111305e-05,
|
| 1081 |
+
"loss": 0.5958,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.49,
|
| 1086 |
+
"learning_rate": 1.0218295048827012e-05,
|
| 1087 |
+
"loss": 0.584,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.49,
|
| 1092 |
+
"learning_rate": 1.0163952243542716e-05,
|
| 1093 |
+
"loss": 0.5774,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.49,
|
| 1098 |
+
"learning_rate": 1.0109609438258423e-05,
|
| 1099 |
+
"loss": 0.5809,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.5,
|
| 1104 |
+
"learning_rate": 1.0055266632974127e-05,
|
| 1105 |
+
"loss": 0.5882,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.5,
|
| 1110 |
+
"learning_rate": 1.0000923827689834e-05,
|
| 1111 |
+
"loss": 0.5948,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.5,
|
| 1116 |
+
"learning_rate": 9.94658102240554e-06,
|
| 1117 |
+
"loss": 0.5809,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.51,
|
| 1122 |
+
"learning_rate": 9.892238217121245e-06,
|
| 1123 |
+
"loss": 0.5816,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.51,
|
| 1128 |
+
"learning_rate": 9.83789541183695e-06,
|
| 1129 |
+
"loss": 0.592,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.51,
|
| 1134 |
+
"learning_rate": 9.783552606552656e-06,
|
| 1135 |
+
"loss": 0.5822,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.51,
|
| 1140 |
+
"learning_rate": 9.729209801268361e-06,
|
| 1141 |
+
"loss": 0.5809,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.52,
|
| 1146 |
+
"learning_rate": 9.674866995984069e-06,
|
| 1147 |
+
"loss": 0.5736,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.52,
|
| 1152 |
+
"learning_rate": 9.620524190699772e-06,
|
| 1153 |
+
"loss": 0.5837,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.52,
|
| 1158 |
+
"learning_rate": 9.566181385415478e-06,
|
| 1159 |
+
"loss": 0.5812,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.52,
|
| 1164 |
+
"learning_rate": 9.511838580131183e-06,
|
| 1165 |
+
"loss": 0.5769,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.53,
|
| 1170 |
+
"learning_rate": 9.45749577484689e-06,
|
| 1171 |
+
"loss": 0.5784,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.53,
|
| 1176 |
+
"learning_rate": 9.403152969562596e-06,
|
| 1177 |
+
"loss": 0.5872,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.53,
|
| 1182 |
+
"learning_rate": 9.348810164278301e-06,
|
| 1183 |
+
"loss": 0.5807,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.54,
|
| 1188 |
+
"learning_rate": 9.294467358994007e-06,
|
| 1189 |
+
"loss": 0.5857,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.54,
|
| 1194 |
+
"learning_rate": 9.240124553709712e-06,
|
| 1195 |
+
"loss": 0.5706,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.54,
|
| 1200 |
+
"learning_rate": 9.185781748425418e-06,
|
| 1201 |
+
"loss": 0.5816,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.54,
|
| 1206 |
+
"learning_rate": 9.131438943141123e-06,
|
| 1207 |
+
"loss": 0.5888,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.55,
|
| 1212 |
+
"learning_rate": 9.077096137856829e-06,
|
| 1213 |
+
"loss": 0.5799,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.55,
|
| 1218 |
+
"learning_rate": 9.022753332572534e-06,
|
| 1219 |
+
"loss": 0.5825,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.55,
|
| 1224 |
+
"learning_rate": 8.968410527288241e-06,
|
| 1225 |
+
"loss": 0.5783,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.55,
|
| 1230 |
+
"learning_rate": 8.914067722003947e-06,
|
| 1231 |
+
"loss": 0.5749,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.56,
|
| 1236 |
+
"learning_rate": 8.859724916719652e-06,
|
| 1237 |
+
"loss": 0.5817,
|
| 1238 |
+
"step": 102500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.56,
|
| 1242 |
+
"learning_rate": 8.805382111435358e-06,
|
| 1243 |
+
"loss": 0.582,
|
| 1244 |
+
"step": 103000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.56,
|
| 1248 |
+
"learning_rate": 8.751039306151063e-06,
|
| 1249 |
+
"loss": 0.5687,
|
| 1250 |
+
"step": 103500
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.57,
|
| 1254 |
+
"learning_rate": 8.696696500866769e-06,
|
| 1255 |
+
"loss": 0.5702,
|
| 1256 |
+
"step": 104000
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.57,
|
| 1260 |
+
"learning_rate": 8.642353695582474e-06,
|
| 1261 |
+
"loss": 0.5734,
|
| 1262 |
+
"step": 104500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.57,
|
| 1266 |
+
"learning_rate": 8.58801089029818e-06,
|
| 1267 |
+
"loss": 0.5769,
|
| 1268 |
+
"step": 105000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.57,
|
| 1272 |
+
"learning_rate": 8.533668085013885e-06,
|
| 1273 |
+
"loss": 0.5648,
|
| 1274 |
+
"step": 105500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.58,
|
| 1278 |
+
"learning_rate": 8.47932527972959e-06,
|
| 1279 |
+
"loss": 0.5686,
|
| 1280 |
+
"step": 106000
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.58,
|
| 1284 |
+
"learning_rate": 8.424982474445297e-06,
|
| 1285 |
+
"loss": 0.572,
|
| 1286 |
+
"step": 106500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.58,
|
| 1290 |
+
"learning_rate": 8.370639669161003e-06,
|
| 1291 |
+
"loss": 0.5753,
|
| 1292 |
+
"step": 107000
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.58,
|
| 1296 |
+
"learning_rate": 8.316296863876707e-06,
|
| 1297 |
+
"loss": 0.5745,
|
| 1298 |
+
"step": 107500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.59,
|
| 1302 |
+
"learning_rate": 8.261954058592412e-06,
|
| 1303 |
+
"loss": 0.5713,
|
| 1304 |
+
"step": 108000
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.59,
|
| 1308 |
+
"learning_rate": 8.20761125330812e-06,
|
| 1309 |
+
"loss": 0.5732,
|
| 1310 |
+
"step": 108500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.59,
|
| 1314 |
+
"learning_rate": 8.153268448023825e-06,
|
| 1315 |
+
"loss": 0.5623,
|
| 1316 |
+
"step": 109000
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.6,
|
| 1320 |
+
"learning_rate": 8.09892564273953e-06,
|
| 1321 |
+
"loss": 0.5679,
|
| 1322 |
+
"step": 109500
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.6,
|
| 1326 |
+
"learning_rate": 8.044582837455236e-06,
|
| 1327 |
+
"loss": 0.5711,
|
| 1328 |
+
"step": 110000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.6,
|
| 1332 |
+
"learning_rate": 7.990240032170941e-06,
|
| 1333 |
+
"loss": 0.5773,
|
| 1334 |
+
"step": 110500
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.6,
|
| 1338 |
+
"learning_rate": 7.935897226886648e-06,
|
| 1339 |
+
"loss": 0.5788,
|
| 1340 |
+
"step": 111000
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.61,
|
| 1344 |
+
"learning_rate": 7.881554421602352e-06,
|
| 1345 |
+
"loss": 0.5626,
|
| 1346 |
+
"step": 111500
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.61,
|
| 1350 |
+
"learning_rate": 7.827211616318057e-06,
|
| 1351 |
+
"loss": 0.5688,
|
| 1352 |
+
"step": 112000
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.61,
|
| 1356 |
+
"learning_rate": 7.772868811033763e-06,
|
| 1357 |
+
"loss": 0.5576,
|
| 1358 |
+
"step": 112500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.61,
|
| 1362 |
+
"learning_rate": 7.71852600574947e-06,
|
| 1363 |
+
"loss": 0.5659,
|
| 1364 |
+
"step": 113000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.62,
|
| 1368 |
+
"learning_rate": 7.664183200465176e-06,
|
| 1369 |
+
"loss": 0.5523,
|
| 1370 |
+
"step": 113500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.62,
|
| 1374 |
+
"learning_rate": 7.609840395180881e-06,
|
| 1375 |
+
"loss": 0.5633,
|
| 1376 |
+
"step": 114000
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.62,
|
| 1380 |
+
"learning_rate": 7.5554975898965864e-06,
|
| 1381 |
+
"loss": 0.5677,
|
| 1382 |
+
"step": 114500
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.62,
|
| 1386 |
+
"learning_rate": 7.501154784612292e-06,
|
| 1387 |
+
"loss": 0.5604,
|
| 1388 |
+
"step": 115000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.63,
|
| 1392 |
+
"learning_rate": 7.4468119793279965e-06,
|
| 1393 |
+
"loss": 0.562,
|
| 1394 |
+
"step": 115500
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 0.63,
|
| 1398 |
+
"learning_rate": 7.392469174043703e-06,
|
| 1399 |
+
"loss": 0.5699,
|
| 1400 |
+
"step": 116000
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 0.63,
|
| 1404 |
+
"learning_rate": 7.338126368759408e-06,
|
| 1405 |
+
"loss": 0.5636,
|
| 1406 |
+
"step": 116500
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.64,
|
| 1410 |
+
"learning_rate": 7.2837835634751146e-06,
|
| 1411 |
+
"loss": 0.5534,
|
| 1412 |
+
"step": 117000
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.64,
|
| 1416 |
+
"learning_rate": 7.22944075819082e-06,
|
| 1417 |
+
"loss": 0.563,
|
| 1418 |
+
"step": 117500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.64,
|
| 1422 |
+
"learning_rate": 7.1750979529065255e-06,
|
| 1423 |
+
"loss": 0.5629,
|
| 1424 |
+
"step": 118000
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.64,
|
| 1428 |
+
"learning_rate": 7.120755147622232e-06,
|
| 1429 |
+
"loss": 0.5682,
|
| 1430 |
+
"step": 118500
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.65,
|
| 1434 |
+
"learning_rate": 7.066412342337937e-06,
|
| 1435 |
+
"loss": 0.5704,
|
| 1436 |
+
"step": 119000
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 0.65,
|
| 1440 |
+
"learning_rate": 7.012069537053643e-06,
|
| 1441 |
+
"loss": 0.5582,
|
| 1442 |
+
"step": 119500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.65,
|
| 1446 |
+
"learning_rate": 6.957726731769347e-06,
|
| 1447 |
+
"loss": 0.5696,
|
| 1448 |
+
"step": 120000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.65,
|
| 1452 |
+
"learning_rate": 6.903383926485054e-06,
|
| 1453 |
+
"loss": 0.5649,
|
| 1454 |
+
"step": 120500
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.66,
|
| 1458 |
+
"learning_rate": 6.849041121200759e-06,
|
| 1459 |
+
"loss": 0.5649,
|
| 1460 |
+
"step": 121000
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.66,
|
| 1464 |
+
"learning_rate": 6.7946983159164645e-06,
|
| 1465 |
+
"loss": 0.5575,
|
| 1466 |
+
"step": 121500
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.66,
|
| 1470 |
+
"learning_rate": 6.740355510632171e-06,
|
| 1471 |
+
"loss": 0.5649,
|
| 1472 |
+
"step": 122000
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.67,
|
| 1476 |
+
"learning_rate": 6.686012705347876e-06,
|
| 1477 |
+
"loss": 0.5571,
|
| 1478 |
+
"step": 122500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 0.67,
|
| 1482 |
+
"learning_rate": 6.631669900063582e-06,
|
| 1483 |
+
"loss": 0.5639,
|
| 1484 |
+
"step": 123000
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.67,
|
| 1488 |
+
"learning_rate": 6.577327094779288e-06,
|
| 1489 |
+
"loss": 0.5666,
|
| 1490 |
+
"step": 123500
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.67,
|
| 1494 |
+
"learning_rate": 6.522984289494993e-06,
|
| 1495 |
+
"loss": 0.5539,
|
| 1496 |
+
"step": 124000
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.68,
|
| 1500 |
+
"learning_rate": 6.468641484210698e-06,
|
| 1501 |
+
"loss": 0.5486,
|
| 1502 |
+
"step": 124500
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.68,
|
| 1506 |
+
"learning_rate": 6.4142986789264035e-06,
|
| 1507 |
+
"loss": 0.5669,
|
| 1508 |
+
"step": 125000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.68,
|
| 1512 |
+
"learning_rate": 6.35995587364211e-06,
|
| 1513 |
+
"loss": 0.5627,
|
| 1514 |
+
"step": 125500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.68,
|
| 1518 |
+
"learning_rate": 6.305613068357815e-06,
|
| 1519 |
+
"loss": 0.554,
|
| 1520 |
+
"step": 126000
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.69,
|
| 1524 |
+
"learning_rate": 6.251270263073521e-06,
|
| 1525 |
+
"loss": 0.5619,
|
| 1526 |
+
"step": 126500
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.69,
|
| 1530 |
+
"learning_rate": 6.196927457789227e-06,
|
| 1531 |
+
"loss": 0.5525,
|
| 1532 |
+
"step": 127000
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.69,
|
| 1536 |
+
"learning_rate": 6.1425846525049325e-06,
|
| 1537 |
+
"loss": 0.5609,
|
| 1538 |
+
"step": 127500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.7,
|
| 1542 |
+
"learning_rate": 6.088241847220637e-06,
|
| 1543 |
+
"loss": 0.5538,
|
| 1544 |
+
"step": 128000
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.7,
|
| 1548 |
+
"learning_rate": 6.0338990419363434e-06,
|
| 1549 |
+
"loss": 0.555,
|
| 1550 |
+
"step": 128500
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.7,
|
| 1554 |
+
"learning_rate": 5.979556236652049e-06,
|
| 1555 |
+
"loss": 0.5472,
|
| 1556 |
+
"step": 129000
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.7,
|
| 1560 |
+
"learning_rate": 5.925213431367754e-06,
|
| 1561 |
+
"loss": 0.5593,
|
| 1562 |
+
"step": 129500
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.71,
|
| 1566 |
+
"learning_rate": 5.870870626083461e-06,
|
| 1567 |
+
"loss": 0.5555,
|
| 1568 |
+
"step": 130000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.71,
|
| 1572 |
+
"learning_rate": 5.816527820799166e-06,
|
| 1573 |
+
"loss": 0.5561,
|
| 1574 |
+
"step": 130500
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.71,
|
| 1578 |
+
"learning_rate": 5.7621850155148716e-06,
|
| 1579 |
+
"loss": 0.5585,
|
| 1580 |
+
"step": 131000
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.71,
|
| 1584 |
+
"learning_rate": 5.707842210230578e-06,
|
| 1585 |
+
"loss": 0.5643,
|
| 1586 |
+
"step": 131500
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.72,
|
| 1590 |
+
"learning_rate": 5.6534994049462825e-06,
|
| 1591 |
+
"loss": 0.548,
|
| 1592 |
+
"step": 132000
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.72,
|
| 1596 |
+
"learning_rate": 5.599156599661988e-06,
|
| 1597 |
+
"loss": 0.5565,
|
| 1598 |
+
"step": 132500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.72,
|
| 1602 |
+
"learning_rate": 5.544813794377693e-06,
|
| 1603 |
+
"loss": 0.5441,
|
| 1604 |
+
"step": 133000
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.73,
|
| 1608 |
+
"learning_rate": 5.4904709890934e-06,
|
| 1609 |
+
"loss": 0.5526,
|
| 1610 |
+
"step": 133500
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.73,
|
| 1614 |
+
"learning_rate": 5.436128183809105e-06,
|
| 1615 |
+
"loss": 0.5505,
|
| 1616 |
+
"step": 134000
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.73,
|
| 1620 |
+
"learning_rate": 5.381785378524811e-06,
|
| 1621 |
+
"loss": 0.5547,
|
| 1622 |
+
"step": 134500
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.73,
|
| 1626 |
+
"learning_rate": 5.327442573240517e-06,
|
| 1627 |
+
"loss": 0.5512,
|
| 1628 |
+
"step": 135000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.74,
|
| 1632 |
+
"learning_rate": 5.273099767956222e-06,
|
| 1633 |
+
"loss": 0.5522,
|
| 1634 |
+
"step": 135500
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.74,
|
| 1638 |
+
"learning_rate": 5.218756962671927e-06,
|
| 1639 |
+
"loss": 0.55,
|
| 1640 |
+
"step": 136000
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.74,
|
| 1644 |
+
"learning_rate": 5.164414157387632e-06,
|
| 1645 |
+
"loss": 0.5566,
|
| 1646 |
+
"step": 136500
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.74,
|
| 1650 |
+
"learning_rate": 5.110071352103339e-06,
|
| 1651 |
+
"loss": 0.5552,
|
| 1652 |
+
"step": 137000
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.75,
|
| 1656 |
+
"learning_rate": 5.055728546819044e-06,
|
| 1657 |
+
"loss": 0.5499,
|
| 1658 |
+
"step": 137500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 0.75,
|
| 1662 |
+
"learning_rate": 5.00138574153475e-06,
|
| 1663 |
+
"loss": 0.5597,
|
| 1664 |
+
"step": 138000
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.75,
|
| 1668 |
+
"learning_rate": 4.947042936250456e-06,
|
| 1669 |
+
"loss": 0.5447,
|
| 1670 |
+
"step": 138500
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.76,
|
| 1674 |
+
"learning_rate": 4.8927001309661605e-06,
|
| 1675 |
+
"loss": 0.5513,
|
| 1676 |
+
"step": 139000
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.76,
|
| 1680 |
+
"learning_rate": 4.838357325681867e-06,
|
| 1681 |
+
"loss": 0.5435,
|
| 1682 |
+
"step": 139500
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.76,
|
| 1686 |
+
"learning_rate": 4.784014520397572e-06,
|
| 1687 |
+
"loss": 0.5485,
|
| 1688 |
+
"step": 140000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.76,
|
| 1692 |
+
"learning_rate": 4.729671715113278e-06,
|
| 1693 |
+
"loss": 0.5427,
|
| 1694 |
+
"step": 140500
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.77,
|
| 1698 |
+
"learning_rate": 4.675328909828983e-06,
|
| 1699 |
+
"loss": 0.5565,
|
| 1700 |
+
"step": 141000
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 0.77,
|
| 1704 |
+
"learning_rate": 4.6209861045446895e-06,
|
| 1705 |
+
"loss": 0.5499,
|
| 1706 |
+
"step": 141500
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.77,
|
| 1710 |
+
"learning_rate": 4.566643299260395e-06,
|
| 1711 |
+
"loss": 0.5421,
|
| 1712 |
+
"step": 142000
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.77,
|
| 1716 |
+
"learning_rate": 4.5123004939761e-06,
|
| 1717 |
+
"loss": 0.54,
|
| 1718 |
+
"step": 142500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.78,
|
| 1722 |
+
"learning_rate": 4.457957688691806e-06,
|
| 1723 |
+
"loss": 0.5435,
|
| 1724 |
+
"step": 143000
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.78,
|
| 1728 |
+
"learning_rate": 4.403614883407511e-06,
|
| 1729 |
+
"loss": 0.5484,
|
| 1730 |
+
"step": 143500
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.78,
|
| 1734 |
+
"learning_rate": 4.349272078123218e-06,
|
| 1735 |
+
"loss": 0.5478,
|
| 1736 |
+
"step": 144000
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.79,
|
| 1740 |
+
"learning_rate": 4.294929272838923e-06,
|
| 1741 |
+
"loss": 0.5519,
|
| 1742 |
+
"step": 144500
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 0.79,
|
| 1746 |
+
"learning_rate": 4.2405864675546285e-06,
|
| 1747 |
+
"loss": 0.5531,
|
| 1748 |
+
"step": 145000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.79,
|
| 1752 |
+
"learning_rate": 4.186243662270334e-06,
|
| 1753 |
+
"loss": 0.5411,
|
| 1754 |
+
"step": 145500
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.79,
|
| 1758 |
+
"learning_rate": 4.1319008569860394e-06,
|
| 1759 |
+
"loss": 0.5424,
|
| 1760 |
+
"step": 146000
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.8,
|
| 1764 |
+
"learning_rate": 4.077558051701746e-06,
|
| 1765 |
+
"loss": 0.5506,
|
| 1766 |
+
"step": 146500
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.8,
|
| 1770 |
+
"learning_rate": 4.02321524641745e-06,
|
| 1771 |
+
"loss": 0.5465,
|
| 1772 |
+
"step": 147000
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.8,
|
| 1776 |
+
"learning_rate": 3.968872441133157e-06,
|
| 1777 |
+
"loss": 0.5448,
|
| 1778 |
+
"step": 147500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.8,
|
| 1782 |
+
"learning_rate": 3.914529635848862e-06,
|
| 1783 |
+
"loss": 0.5347,
|
| 1784 |
+
"step": 148000
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.81,
|
| 1788 |
+
"learning_rate": 3.8601868305645676e-06,
|
| 1789 |
+
"loss": 0.5429,
|
| 1790 |
+
"step": 148500
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.81,
|
| 1794 |
+
"learning_rate": 3.805844025280273e-06,
|
| 1795 |
+
"loss": 0.5401,
|
| 1796 |
+
"step": 149000
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.81,
|
| 1800 |
+
"learning_rate": 3.751501219995979e-06,
|
| 1801 |
+
"loss": 0.5428,
|
| 1802 |
+
"step": 149500
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.82,
|
| 1806 |
+
"learning_rate": 3.6971584147116848e-06,
|
| 1807 |
+
"loss": 0.5414,
|
| 1808 |
+
"step": 150000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.82,
|
| 1812 |
+
"learning_rate": 3.6428156094273902e-06,
|
| 1813 |
+
"loss": 0.541,
|
| 1814 |
+
"step": 150500
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.82,
|
| 1818 |
+
"learning_rate": 3.5884728041430957e-06,
|
| 1819 |
+
"loss": 0.5396,
|
| 1820 |
+
"step": 151000
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 0.82,
|
| 1824 |
+
"learning_rate": 3.534129998858801e-06,
|
| 1825 |
+
"loss": 0.5398,
|
| 1826 |
+
"step": 151500
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 0.83,
|
| 1830 |
+
"learning_rate": 3.479787193574507e-06,
|
| 1831 |
+
"loss": 0.5393,
|
| 1832 |
+
"step": 152000
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.83,
|
| 1836 |
+
"learning_rate": 3.425444388290213e-06,
|
| 1837 |
+
"loss": 0.5378,
|
| 1838 |
+
"step": 152500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.83,
|
| 1842 |
+
"learning_rate": 3.371101583005918e-06,
|
| 1843 |
+
"loss": 0.5344,
|
| 1844 |
+
"step": 153000
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.83,
|
| 1848 |
+
"learning_rate": 3.316758777721624e-06,
|
| 1849 |
+
"loss": 0.5418,
|
| 1850 |
+
"step": 153500
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.84,
|
| 1854 |
+
"learning_rate": 3.2624159724373293e-06,
|
| 1855 |
+
"loss": 0.5266,
|
| 1856 |
+
"step": 154000
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.84,
|
| 1860 |
+
"learning_rate": 3.208073167153035e-06,
|
| 1861 |
+
"loss": 0.5376,
|
| 1862 |
+
"step": 154500
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 0.84,
|
| 1866 |
+
"learning_rate": 3.15373036186874e-06,
|
| 1867 |
+
"loss": 0.5471,
|
| 1868 |
+
"step": 155000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 0.85,
|
| 1872 |
+
"learning_rate": 3.099387556584446e-06,
|
| 1873 |
+
"loss": 0.5451,
|
| 1874 |
+
"step": 155500
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.85,
|
| 1878 |
+
"learning_rate": 3.045044751300152e-06,
|
| 1879 |
+
"loss": 0.5451,
|
| 1880 |
+
"step": 156000
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.85,
|
| 1884 |
+
"learning_rate": 2.990701946015858e-06,
|
| 1885 |
+
"loss": 0.5423,
|
| 1886 |
+
"step": 156500
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.85,
|
| 1890 |
+
"learning_rate": 2.9363591407315633e-06,
|
| 1891 |
+
"loss": 0.5332,
|
| 1892 |
+
"step": 157000
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.86,
|
| 1896 |
+
"learning_rate": 2.8820163354472687e-06,
|
| 1897 |
+
"loss": 0.5367,
|
| 1898 |
+
"step": 157500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.86,
|
| 1902 |
+
"learning_rate": 2.827673530162974e-06,
|
| 1903 |
+
"loss": 0.5366,
|
| 1904 |
+
"step": 158000
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.86,
|
| 1908 |
+
"learning_rate": 2.77333072487868e-06,
|
| 1909 |
+
"loss": 0.5362,
|
| 1910 |
+
"step": 158500
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 0.86,
|
| 1914 |
+
"learning_rate": 2.718987919594386e-06,
|
| 1915 |
+
"loss": 0.5386,
|
| 1916 |
+
"step": 159000
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.87,
|
| 1920 |
+
"learning_rate": 2.664645114310091e-06,
|
| 1921 |
+
"loss": 0.5381,
|
| 1922 |
+
"step": 159500
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.87,
|
| 1926 |
+
"learning_rate": 2.610302309025797e-06,
|
| 1927 |
+
"loss": 0.5415,
|
| 1928 |
+
"step": 160000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.87,
|
| 1932 |
+
"learning_rate": 2.5559595037415023e-06,
|
| 1933 |
+
"loss": 0.5308,
|
| 1934 |
+
"step": 160500
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.87,
|
| 1938 |
+
"learning_rate": 2.501616698457208e-06,
|
| 1939 |
+
"loss": 0.5298,
|
| 1940 |
+
"step": 161000
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.88,
|
| 1944 |
+
"learning_rate": 2.4472738931729136e-06,
|
| 1945 |
+
"loss": 0.5209,
|
| 1946 |
+
"step": 161500
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 0.88,
|
| 1950 |
+
"learning_rate": 2.392931087888619e-06,
|
| 1951 |
+
"loss": 0.5374,
|
| 1952 |
+
"step": 162000
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 0.88,
|
| 1956 |
+
"learning_rate": 2.338588282604325e-06,
|
| 1957 |
+
"loss": 0.5437,
|
| 1958 |
+
"step": 162500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.89,
|
| 1962 |
+
"learning_rate": 2.2842454773200304e-06,
|
| 1963 |
+
"loss": 0.5353,
|
| 1964 |
+
"step": 163000
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.89,
|
| 1968 |
+
"learning_rate": 2.2299026720357363e-06,
|
| 1969 |
+
"loss": 0.5377,
|
| 1970 |
+
"step": 163500
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.89,
|
| 1974 |
+
"learning_rate": 2.1755598667514418e-06,
|
| 1975 |
+
"loss": 0.5412,
|
| 1976 |
+
"step": 164000
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.89,
|
| 1980 |
+
"learning_rate": 2.121217061467147e-06,
|
| 1981 |
+
"loss": 0.5236,
|
| 1982 |
+
"step": 164500
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.9,
|
| 1986 |
+
"learning_rate": 2.0668742561828527e-06,
|
| 1987 |
+
"loss": 0.5369,
|
| 1988 |
+
"step": 165000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 0.9,
|
| 1992 |
+
"learning_rate": 2.0125314508985585e-06,
|
| 1993 |
+
"loss": 0.529,
|
| 1994 |
+
"step": 165500
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 0.9,
|
| 1998 |
+
"learning_rate": 1.958188645614264e-06,
|
| 1999 |
+
"loss": 0.5338,
|
| 2000 |
+
"step": 166000
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.9,
|
| 2004 |
+
"learning_rate": 1.9038458403299699e-06,
|
| 2005 |
+
"loss": 0.53,
|
| 2006 |
+
"step": 166500
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.91,
|
| 2010 |
+
"learning_rate": 1.8495030350456753e-06,
|
| 2011 |
+
"loss": 0.5254,
|
| 2012 |
+
"step": 167000
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.91,
|
| 2016 |
+
"learning_rate": 1.795160229761381e-06,
|
| 2017 |
+
"loss": 0.5382,
|
| 2018 |
+
"step": 167500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.91,
|
| 2022 |
+
"learning_rate": 1.7408174244770865e-06,
|
| 2023 |
+
"loss": 0.541,
|
| 2024 |
+
"step": 168000
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.92,
|
| 2028 |
+
"learning_rate": 1.6864746191927921e-06,
|
| 2029 |
+
"loss": 0.5296,
|
| 2030 |
+
"step": 168500
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.92,
|
| 2034 |
+
"learning_rate": 1.6321318139084976e-06,
|
| 2035 |
+
"loss": 0.5304,
|
| 2036 |
+
"step": 169000
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 0.92,
|
| 2040 |
+
"learning_rate": 1.5777890086242035e-06,
|
| 2041 |
+
"loss": 0.5253,
|
| 2042 |
+
"step": 169500
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.92,
|
| 2046 |
+
"learning_rate": 1.523446203339909e-06,
|
| 2047 |
+
"loss": 0.5189,
|
| 2048 |
+
"step": 170000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.93,
|
| 2052 |
+
"learning_rate": 1.4691033980556146e-06,
|
| 2053 |
+
"loss": 0.5359,
|
| 2054 |
+
"step": 170500
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.93,
|
| 2058 |
+
"learning_rate": 1.41476059277132e-06,
|
| 2059 |
+
"loss": 0.5368,
|
| 2060 |
+
"step": 171000
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.93,
|
| 2064 |
+
"learning_rate": 1.360417787487026e-06,
|
| 2065 |
+
"loss": 0.5265,
|
| 2066 |
+
"step": 171500
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.93,
|
| 2070 |
+
"learning_rate": 1.3060749822027314e-06,
|
| 2071 |
+
"loss": 0.5289,
|
| 2072 |
+
"step": 172000
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.94,
|
| 2076 |
+
"learning_rate": 1.251732176918437e-06,
|
| 2077 |
+
"loss": 0.5294,
|
| 2078 |
+
"step": 172500
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 0.94,
|
| 2082 |
+
"learning_rate": 1.1973893716341425e-06,
|
| 2083 |
+
"loss": 0.5293,
|
| 2084 |
+
"step": 173000
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.94,
|
| 2088 |
+
"learning_rate": 1.1430465663498482e-06,
|
| 2089 |
+
"loss": 0.5342,
|
| 2090 |
+
"step": 173500
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.95,
|
| 2094 |
+
"learning_rate": 1.0887037610655538e-06,
|
| 2095 |
+
"loss": 0.5182,
|
| 2096 |
+
"step": 174000
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.95,
|
| 2100 |
+
"learning_rate": 1.0343609557812593e-06,
|
| 2101 |
+
"loss": 0.5311,
|
| 2102 |
+
"step": 174500
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.95,
|
| 2106 |
+
"learning_rate": 9.80018150496965e-07,
|
| 2107 |
+
"loss": 0.5386,
|
| 2108 |
+
"step": 175000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.95,
|
| 2112 |
+
"learning_rate": 9.256753452126706e-07,
|
| 2113 |
+
"loss": 0.5291,
|
| 2114 |
+
"step": 175500
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 0.96,
|
| 2118 |
+
"learning_rate": 8.713325399283762e-07,
|
| 2119 |
+
"loss": 0.5233,
|
| 2120 |
+
"step": 176000
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 0.96,
|
| 2124 |
+
"learning_rate": 8.169897346440818e-07,
|
| 2125 |
+
"loss": 0.5167,
|
| 2126 |
+
"step": 176500
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.96,
|
| 2130 |
+
"learning_rate": 7.626469293597875e-07,
|
| 2131 |
+
"loss": 0.5339,
|
| 2132 |
+
"step": 177000
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.96,
|
| 2136 |
+
"learning_rate": 7.083041240754932e-07,
|
| 2137 |
+
"loss": 0.5263,
|
| 2138 |
+
"step": 177500
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.97,
|
| 2142 |
+
"learning_rate": 6.539613187911987e-07,
|
| 2143 |
+
"loss": 0.5255,
|
| 2144 |
+
"step": 178000
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.97,
|
| 2148 |
+
"learning_rate": 5.996185135069043e-07,
|
| 2149 |
+
"loss": 0.5292,
|
| 2150 |
+
"step": 178500
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.97,
|
| 2154 |
+
"learning_rate": 5.452757082226099e-07,
|
| 2155 |
+
"loss": 0.5247,
|
| 2156 |
+
"step": 179000
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 0.98,
|
| 2160 |
+
"learning_rate": 4.909329029383155e-07,
|
| 2161 |
+
"loss": 0.5293,
|
| 2162 |
+
"step": 179500
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.98,
|
| 2166 |
+
"learning_rate": 4.3659009765402114e-07,
|
| 2167 |
+
"loss": 0.5278,
|
| 2168 |
+
"step": 180000
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.98,
|
| 2172 |
+
"learning_rate": 3.8224729236972675e-07,
|
| 2173 |
+
"loss": 0.5151,
|
| 2174 |
+
"step": 180500
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.98,
|
| 2178 |
+
"learning_rate": 3.279044870854323e-07,
|
| 2179 |
+
"loss": 0.515,
|
| 2180 |
+
"step": 181000
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.99,
|
| 2184 |
+
"learning_rate": 2.73561681801138e-07,
|
| 2185 |
+
"loss": 0.5305,
|
| 2186 |
+
"step": 181500
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.99,
|
| 2190 |
+
"learning_rate": 2.192188765168436e-07,
|
| 2191 |
+
"loss": 0.5231,
|
| 2192 |
+
"step": 182000
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.99,
|
| 2196 |
+
"learning_rate": 1.6487607123254918e-07,
|
| 2197 |
+
"loss": 0.526,
|
| 2198 |
+
"step": 182500
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 0.99,
|
| 2202 |
+
"learning_rate": 1.1053326594825478e-07,
|
| 2203 |
+
"loss": 0.5243,
|
| 2204 |
+
"step": 183000
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 1.0,
|
| 2208 |
+
"learning_rate": 5.6190460663960404e-08,
|
| 2209 |
+
"loss": 0.5192,
|
| 2210 |
+
"step": 183500
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 1.0,
|
| 2214 |
+
"learning_rate": 1.8476553796660092e-09,
|
| 2215 |
+
"loss": 0.5343,
|
| 2216 |
+
"step": 184000
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 1.0,
|
| 2220 |
+
"eval_accuracy": 0.7746184738955824,
|
| 2221 |
+
"eval_loss": 0.587454617023468,
|
| 2222 |
+
"eval_runtime": 72.5338,
|
| 2223 |
+
"eval_samples_per_second": 514.932,
|
| 2224 |
+
"eval_steps_per_second": 64.37,
|
| 2225 |
+
"step": 184017
|
| 2226 |
+
}
|
| 2227 |
+
],
|
| 2228 |
+
"max_steps": 184017,
|
| 2229 |
+
"num_train_epochs": 1,
|
| 2230 |
+
"total_flos": 3.8733781342346496e+17,
|
| 2231 |
+
"trial_name": null,
|
| 2232 |
+
"trial_params": null
|
| 2233 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ede5e6053d8c93bddaa331b36d7f66005eb576f360c1266d6a2de18db55b71c
|
| 3 |
+
size 3451
|