Upload 7 files
Browse files- config.json +31 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +2860 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "monologg/koelectra-small-v3-discriminator",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ElectraForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"embedding_size": 128,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 1024,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 512,
|
| 16 |
+
"model_type": "electra",
|
| 17 |
+
"num_attention_heads": 4,
|
| 18 |
+
"num_hidden_layers": 12,
|
| 19 |
+
"pad_token_id": 0,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"problem_type": "single_label_classification",
|
| 22 |
+
"summary_activation": "gelu",
|
| 23 |
+
"summary_last_dropout": 0.1,
|
| 24 |
+
"summary_type": "first",
|
| 25 |
+
"summary_use_proj": true,
|
| 26 |
+
"torch_dtype": "float32",
|
| 27 |
+
"transformers_version": "4.29.2",
|
| 28 |
+
"type_vocab_size": 2,
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"vocab_size": 35000
|
| 31 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:41c0997f448b9705e811c257422cd20b07d100ac3f54453365d89b72547f7ee3
|
| 3 |
+
size 113099013
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00355e900fe98002806dff6f71f026a45b71d57c2167bf2ef347e22747c0c528
|
| 3 |
+
size 56564029
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:970589f8b63a7684de3417f5934464414fbad64e719455815d8ac1e7180a9ae0
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa0f31629c95bc0930675c189e53f36ed56816138d5ec238259c07a27291441f
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2860 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.00564388744533062,
|
| 3 |
+
"best_model_checkpoint": "output_dir\\checkpoint-9188",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 22970,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 9.978232477144103e-06,
|
| 13 |
+
"loss": 0.5997,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 9.956464954288203e-06,
|
| 19 |
+
"loss": 0.5077,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 9.934697431432305e-06,
|
| 25 |
+
"loss": 0.4455,
|
| 26 |
+
"step": 150
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"learning_rate": 9.912929908576405e-06,
|
| 31 |
+
"loss": 0.3683,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.11,
|
| 36 |
+
"learning_rate": 9.891162385720507e-06,
|
| 37 |
+
"loss": 0.2864,
|
| 38 |
+
"step": 250
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.13,
|
| 42 |
+
"learning_rate": 9.869394862864606e-06,
|
| 43 |
+
"loss": 0.2272,
|
| 44 |
+
"step": 300
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.15,
|
| 48 |
+
"learning_rate": 9.847627340008708e-06,
|
| 49 |
+
"loss": 0.1793,
|
| 50 |
+
"step": 350
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.17,
|
| 54 |
+
"learning_rate": 9.825859817152808e-06,
|
| 55 |
+
"loss": 0.1467,
|
| 56 |
+
"step": 400
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.2,
|
| 60 |
+
"learning_rate": 9.80409229429691e-06,
|
| 61 |
+
"loss": 0.1229,
|
| 62 |
+
"step": 450
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.22,
|
| 66 |
+
"learning_rate": 9.782324771441012e-06,
|
| 67 |
+
"loss": 0.0996,
|
| 68 |
+
"step": 500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.24,
|
| 72 |
+
"learning_rate": 9.760557248585112e-06,
|
| 73 |
+
"loss": 0.092,
|
| 74 |
+
"step": 550
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.26,
|
| 78 |
+
"learning_rate": 9.738789725729214e-06,
|
| 79 |
+
"loss": 0.0771,
|
| 80 |
+
"step": 600
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.28,
|
| 84 |
+
"learning_rate": 9.717022202873314e-06,
|
| 85 |
+
"loss": 0.0597,
|
| 86 |
+
"step": 650
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.3,
|
| 90 |
+
"learning_rate": 9.695254680017416e-06,
|
| 91 |
+
"loss": 0.0536,
|
| 92 |
+
"step": 700
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.33,
|
| 96 |
+
"learning_rate": 9.673487157161516e-06,
|
| 97 |
+
"loss": 0.0514,
|
| 98 |
+
"step": 750
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.35,
|
| 102 |
+
"learning_rate": 9.651719634305616e-06,
|
| 103 |
+
"loss": 0.0447,
|
| 104 |
+
"step": 800
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.37,
|
| 108 |
+
"learning_rate": 9.629952111449717e-06,
|
| 109 |
+
"loss": 0.0444,
|
| 110 |
+
"step": 850
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.39,
|
| 114 |
+
"learning_rate": 9.608184588593819e-06,
|
| 115 |
+
"loss": 0.0363,
|
| 116 |
+
"step": 900
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.41,
|
| 120 |
+
"learning_rate": 9.58641706573792e-06,
|
| 121 |
+
"loss": 0.0395,
|
| 122 |
+
"step": 950
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.44,
|
| 126 |
+
"learning_rate": 9.56464954288202e-06,
|
| 127 |
+
"loss": 0.0444,
|
| 128 |
+
"step": 1000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.46,
|
| 132 |
+
"learning_rate": 9.542882020026123e-06,
|
| 133 |
+
"loss": 0.0328,
|
| 134 |
+
"step": 1050
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.48,
|
| 138 |
+
"learning_rate": 9.521114497170223e-06,
|
| 139 |
+
"loss": 0.0337,
|
| 140 |
+
"step": 1100
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.5,
|
| 144 |
+
"learning_rate": 9.499346974314325e-06,
|
| 145 |
+
"loss": 0.0347,
|
| 146 |
+
"step": 1150
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.52,
|
| 150 |
+
"learning_rate": 9.477579451458425e-06,
|
| 151 |
+
"loss": 0.0262,
|
| 152 |
+
"step": 1200
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.54,
|
| 156 |
+
"learning_rate": 9.455811928602527e-06,
|
| 157 |
+
"loss": 0.0239,
|
| 158 |
+
"step": 1250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.57,
|
| 162 |
+
"learning_rate": 9.434044405746625e-06,
|
| 163 |
+
"loss": 0.0259,
|
| 164 |
+
"step": 1300
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.59,
|
| 168 |
+
"learning_rate": 9.412276882890727e-06,
|
| 169 |
+
"loss": 0.021,
|
| 170 |
+
"step": 1350
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.61,
|
| 174 |
+
"learning_rate": 9.390509360034828e-06,
|
| 175 |
+
"loss": 0.0186,
|
| 176 |
+
"step": 1400
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.63,
|
| 180 |
+
"learning_rate": 9.36874183717893e-06,
|
| 181 |
+
"loss": 0.0207,
|
| 182 |
+
"step": 1450
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.65,
|
| 186 |
+
"learning_rate": 9.346974314323032e-06,
|
| 187 |
+
"loss": 0.017,
|
| 188 |
+
"step": 1500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.67,
|
| 192 |
+
"learning_rate": 9.325206791467132e-06,
|
| 193 |
+
"loss": 0.0259,
|
| 194 |
+
"step": 1550
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.7,
|
| 198 |
+
"learning_rate": 9.303439268611234e-06,
|
| 199 |
+
"loss": 0.0299,
|
| 200 |
+
"step": 1600
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.72,
|
| 204 |
+
"learning_rate": 9.281671745755334e-06,
|
| 205 |
+
"loss": 0.0124,
|
| 206 |
+
"step": 1650
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.74,
|
| 210 |
+
"learning_rate": 9.259904222899436e-06,
|
| 211 |
+
"loss": 0.0155,
|
| 212 |
+
"step": 1700
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.76,
|
| 216 |
+
"learning_rate": 9.238136700043536e-06,
|
| 217 |
+
"loss": 0.0264,
|
| 218 |
+
"step": 1750
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.78,
|
| 222 |
+
"learning_rate": 9.216369177187638e-06,
|
| 223 |
+
"loss": 0.018,
|
| 224 |
+
"step": 1800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.81,
|
| 228 |
+
"learning_rate": 9.194601654331737e-06,
|
| 229 |
+
"loss": 0.0163,
|
| 230 |
+
"step": 1850
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.83,
|
| 234 |
+
"learning_rate": 9.172834131475838e-06,
|
| 235 |
+
"loss": 0.0219,
|
| 236 |
+
"step": 1900
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.85,
|
| 240 |
+
"learning_rate": 9.15106660861994e-06,
|
| 241 |
+
"loss": 0.0164,
|
| 242 |
+
"step": 1950
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.87,
|
| 246 |
+
"learning_rate": 9.12929908576404e-06,
|
| 247 |
+
"loss": 0.026,
|
| 248 |
+
"step": 2000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.89,
|
| 252 |
+
"learning_rate": 9.107531562908143e-06,
|
| 253 |
+
"loss": 0.0179,
|
| 254 |
+
"step": 2050
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.91,
|
| 258 |
+
"learning_rate": 9.085764040052243e-06,
|
| 259 |
+
"loss": 0.0177,
|
| 260 |
+
"step": 2100
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.94,
|
| 264 |
+
"learning_rate": 9.063996517196345e-06,
|
| 265 |
+
"loss": 0.0138,
|
| 266 |
+
"step": 2150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.96,
|
| 270 |
+
"learning_rate": 9.042228994340445e-06,
|
| 271 |
+
"loss": 0.0134,
|
| 272 |
+
"step": 2200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.98,
|
| 276 |
+
"learning_rate": 9.020461471484547e-06,
|
| 277 |
+
"loss": 0.0175,
|
| 278 |
+
"step": 2250
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.0,
|
| 282 |
+
"eval_f1": 0.998025618014196,
|
| 283 |
+
"eval_loss": 0.015513530932366848,
|
| 284 |
+
"eval_runtime": 36.185,
|
| 285 |
+
"eval_samples_per_second": 1015.282,
|
| 286 |
+
"eval_steps_per_second": 15.891,
|
| 287 |
+
"step": 2297
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.0,
|
| 291 |
+
"learning_rate": 8.998693948628647e-06,
|
| 292 |
+
"loss": 0.0105,
|
| 293 |
+
"step": 2300
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.02,
|
| 297 |
+
"learning_rate": 8.976926425772747e-06,
|
| 298 |
+
"loss": 0.0151,
|
| 299 |
+
"step": 2350
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.04,
|
| 303 |
+
"learning_rate": 8.95515890291685e-06,
|
| 304 |
+
"loss": 0.0155,
|
| 305 |
+
"step": 2400
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.07,
|
| 309 |
+
"learning_rate": 8.93339138006095e-06,
|
| 310 |
+
"loss": 0.016,
|
| 311 |
+
"step": 2450
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.09,
|
| 315 |
+
"learning_rate": 8.911623857205051e-06,
|
| 316 |
+
"loss": 0.0142,
|
| 317 |
+
"step": 2500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.11,
|
| 321 |
+
"learning_rate": 8.889856334349152e-06,
|
| 322 |
+
"loss": 0.0211,
|
| 323 |
+
"step": 2550
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.13,
|
| 327 |
+
"learning_rate": 8.868088811493254e-06,
|
| 328 |
+
"loss": 0.0166,
|
| 329 |
+
"step": 2600
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.15,
|
| 333 |
+
"learning_rate": 8.846321288637354e-06,
|
| 334 |
+
"loss": 0.0195,
|
| 335 |
+
"step": 2650
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.18,
|
| 339 |
+
"learning_rate": 8.824553765781456e-06,
|
| 340 |
+
"loss": 0.0181,
|
| 341 |
+
"step": 2700
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.2,
|
| 345 |
+
"learning_rate": 8.802786242925556e-06,
|
| 346 |
+
"loss": 0.0137,
|
| 347 |
+
"step": 2750
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.22,
|
| 351 |
+
"learning_rate": 8.781018720069658e-06,
|
| 352 |
+
"loss": 0.0092,
|
| 353 |
+
"step": 2800
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.24,
|
| 357 |
+
"learning_rate": 8.759251197213758e-06,
|
| 358 |
+
"loss": 0.0153,
|
| 359 |
+
"step": 2850
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.26,
|
| 363 |
+
"learning_rate": 8.737483674357858e-06,
|
| 364 |
+
"loss": 0.0098,
|
| 365 |
+
"step": 2900
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.28,
|
| 369 |
+
"learning_rate": 8.71571615150196e-06,
|
| 370 |
+
"loss": 0.0116,
|
| 371 |
+
"step": 2950
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.31,
|
| 375 |
+
"learning_rate": 8.69394862864606e-06,
|
| 376 |
+
"loss": 0.0125,
|
| 377 |
+
"step": 3000
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.33,
|
| 381 |
+
"learning_rate": 8.672181105790162e-06,
|
| 382 |
+
"loss": 0.0169,
|
| 383 |
+
"step": 3050
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.35,
|
| 387 |
+
"learning_rate": 8.650413582934263e-06,
|
| 388 |
+
"loss": 0.0191,
|
| 389 |
+
"step": 3100
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.37,
|
| 393 |
+
"learning_rate": 8.628646060078365e-06,
|
| 394 |
+
"loss": 0.0135,
|
| 395 |
+
"step": 3150
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.39,
|
| 399 |
+
"learning_rate": 8.606878537222465e-06,
|
| 400 |
+
"loss": 0.0104,
|
| 401 |
+
"step": 3200
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.41,
|
| 405 |
+
"learning_rate": 8.585111014366567e-06,
|
| 406 |
+
"loss": 0.0074,
|
| 407 |
+
"step": 3250
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.44,
|
| 411 |
+
"learning_rate": 8.563343491510667e-06,
|
| 412 |
+
"loss": 0.0165,
|
| 413 |
+
"step": 3300
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.46,
|
| 417 |
+
"learning_rate": 8.541575968654767e-06,
|
| 418 |
+
"loss": 0.01,
|
| 419 |
+
"step": 3350
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.48,
|
| 423 |
+
"learning_rate": 8.519808445798869e-06,
|
| 424 |
+
"loss": 0.0092,
|
| 425 |
+
"step": 3400
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.5,
|
| 429 |
+
"learning_rate": 8.49804092294297e-06,
|
| 430 |
+
"loss": 0.0127,
|
| 431 |
+
"step": 3450
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.52,
|
| 435 |
+
"learning_rate": 8.476273400087071e-06,
|
| 436 |
+
"loss": 0.0123,
|
| 437 |
+
"step": 3500
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.55,
|
| 441 |
+
"learning_rate": 8.454505877231172e-06,
|
| 442 |
+
"loss": 0.0096,
|
| 443 |
+
"step": 3550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.57,
|
| 447 |
+
"learning_rate": 8.432738354375273e-06,
|
| 448 |
+
"loss": 0.0095,
|
| 449 |
+
"step": 3600
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.59,
|
| 453 |
+
"learning_rate": 8.410970831519374e-06,
|
| 454 |
+
"loss": 0.0109,
|
| 455 |
+
"step": 3650
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.61,
|
| 459 |
+
"learning_rate": 8.389203308663476e-06,
|
| 460 |
+
"loss": 0.0086,
|
| 461 |
+
"step": 3700
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.63,
|
| 465 |
+
"learning_rate": 8.367435785807576e-06,
|
| 466 |
+
"loss": 0.0108,
|
| 467 |
+
"step": 3750
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.65,
|
| 471 |
+
"learning_rate": 8.345668262951678e-06,
|
| 472 |
+
"loss": 0.0077,
|
| 473 |
+
"step": 3800
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.68,
|
| 477 |
+
"learning_rate": 8.323900740095778e-06,
|
| 478 |
+
"loss": 0.0091,
|
| 479 |
+
"step": 3850
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.7,
|
| 483 |
+
"learning_rate": 8.302133217239878e-06,
|
| 484 |
+
"loss": 0.013,
|
| 485 |
+
"step": 3900
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.72,
|
| 489 |
+
"learning_rate": 8.28036569438398e-06,
|
| 490 |
+
"loss": 0.0147,
|
| 491 |
+
"step": 3950
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.74,
|
| 495 |
+
"learning_rate": 8.25859817152808e-06,
|
| 496 |
+
"loss": 0.0153,
|
| 497 |
+
"step": 4000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.76,
|
| 501 |
+
"learning_rate": 8.236830648672182e-06,
|
| 502 |
+
"loss": 0.0103,
|
| 503 |
+
"step": 4050
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.78,
|
| 507 |
+
"learning_rate": 8.215063125816283e-06,
|
| 508 |
+
"loss": 0.0101,
|
| 509 |
+
"step": 4100
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.81,
|
| 513 |
+
"learning_rate": 8.193295602960384e-06,
|
| 514 |
+
"loss": 0.0049,
|
| 515 |
+
"step": 4150
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.83,
|
| 519 |
+
"learning_rate": 8.171528080104485e-06,
|
| 520 |
+
"loss": 0.008,
|
| 521 |
+
"step": 4200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.85,
|
| 525 |
+
"learning_rate": 8.149760557248587e-06,
|
| 526 |
+
"loss": 0.0126,
|
| 527 |
+
"step": 4250
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.87,
|
| 531 |
+
"learning_rate": 8.127993034392687e-06,
|
| 532 |
+
"loss": 0.0124,
|
| 533 |
+
"step": 4300
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.89,
|
| 537 |
+
"learning_rate": 8.106225511536789e-06,
|
| 538 |
+
"loss": 0.0111,
|
| 539 |
+
"step": 4350
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.92,
|
| 543 |
+
"learning_rate": 8.084457988680889e-06,
|
| 544 |
+
"loss": 0.0137,
|
| 545 |
+
"step": 4400
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.94,
|
| 549 |
+
"learning_rate": 8.06269046582499e-06,
|
| 550 |
+
"loss": 0.0148,
|
| 551 |
+
"step": 4450
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 1.96,
|
| 555 |
+
"learning_rate": 8.040922942969091e-06,
|
| 556 |
+
"loss": 0.0154,
|
| 557 |
+
"step": 4500
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.98,
|
| 561 |
+
"learning_rate": 8.019155420113191e-06,
|
| 562 |
+
"loss": 0.0104,
|
| 563 |
+
"step": 4550
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 2.0,
|
| 567 |
+
"eval_f1": 0.9987417272652994,
|
| 568 |
+
"eval_loss": 0.008711813017725945,
|
| 569 |
+
"eval_runtime": 35.665,
|
| 570 |
+
"eval_samples_per_second": 1030.086,
|
| 571 |
+
"eval_steps_per_second": 16.122,
|
| 572 |
+
"step": 4594
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.0,
|
| 576 |
+
"learning_rate": 7.997387897257293e-06,
|
| 577 |
+
"loss": 0.0077,
|
| 578 |
+
"step": 4600
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.02,
|
| 582 |
+
"learning_rate": 7.975620374401394e-06,
|
| 583 |
+
"loss": 0.0081,
|
| 584 |
+
"step": 4650
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.05,
|
| 588 |
+
"learning_rate": 7.953852851545496e-06,
|
| 589 |
+
"loss": 0.0071,
|
| 590 |
+
"step": 4700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.07,
|
| 594 |
+
"learning_rate": 7.932085328689596e-06,
|
| 595 |
+
"loss": 0.0164,
|
| 596 |
+
"step": 4750
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.09,
|
| 600 |
+
"learning_rate": 7.910317805833698e-06,
|
| 601 |
+
"loss": 0.0096,
|
| 602 |
+
"step": 4800
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.11,
|
| 606 |
+
"learning_rate": 7.888550282977798e-06,
|
| 607 |
+
"loss": 0.0139,
|
| 608 |
+
"step": 4850
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.13,
|
| 612 |
+
"learning_rate": 7.866782760121898e-06,
|
| 613 |
+
"loss": 0.0068,
|
| 614 |
+
"step": 4900
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.15,
|
| 618 |
+
"learning_rate": 7.845015237266e-06,
|
| 619 |
+
"loss": 0.0079,
|
| 620 |
+
"step": 4950
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.18,
|
| 624 |
+
"learning_rate": 7.8232477144101e-06,
|
| 625 |
+
"loss": 0.0084,
|
| 626 |
+
"step": 5000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.2,
|
| 630 |
+
"learning_rate": 7.801480191554202e-06,
|
| 631 |
+
"loss": 0.0008,
|
| 632 |
+
"step": 5050
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.22,
|
| 636 |
+
"learning_rate": 7.779712668698302e-06,
|
| 637 |
+
"loss": 0.01,
|
| 638 |
+
"step": 5100
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.24,
|
| 642 |
+
"learning_rate": 7.757945145842404e-06,
|
| 643 |
+
"loss": 0.0094,
|
| 644 |
+
"step": 5150
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.26,
|
| 648 |
+
"learning_rate": 7.736177622986505e-06,
|
| 649 |
+
"loss": 0.0094,
|
| 650 |
+
"step": 5200
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.29,
|
| 654 |
+
"learning_rate": 7.714410100130607e-06,
|
| 655 |
+
"loss": 0.0116,
|
| 656 |
+
"step": 5250
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.31,
|
| 660 |
+
"learning_rate": 7.692642577274707e-06,
|
| 661 |
+
"loss": 0.0055,
|
| 662 |
+
"step": 5300
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.33,
|
| 666 |
+
"learning_rate": 7.670875054418809e-06,
|
| 667 |
+
"loss": 0.0047,
|
| 668 |
+
"step": 5350
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.35,
|
| 672 |
+
"learning_rate": 7.649107531562909e-06,
|
| 673 |
+
"loss": 0.0102,
|
| 674 |
+
"step": 5400
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.37,
|
| 678 |
+
"learning_rate": 7.627340008707009e-06,
|
| 679 |
+
"loss": 0.0095,
|
| 680 |
+
"step": 5450
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.39,
|
| 684 |
+
"learning_rate": 7.60557248585111e-06,
|
| 685 |
+
"loss": 0.0136,
|
| 686 |
+
"step": 5500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.42,
|
| 690 |
+
"learning_rate": 7.583804962995211e-06,
|
| 691 |
+
"loss": 0.0086,
|
| 692 |
+
"step": 5550
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.44,
|
| 696 |
+
"learning_rate": 7.562037440139312e-06,
|
| 697 |
+
"loss": 0.0024,
|
| 698 |
+
"step": 5600
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.46,
|
| 702 |
+
"learning_rate": 7.5402699172834135e-06,
|
| 703 |
+
"loss": 0.0064,
|
| 704 |
+
"step": 5650
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.48,
|
| 708 |
+
"learning_rate": 7.5185023944275146e-06,
|
| 709 |
+
"loss": 0.0091,
|
| 710 |
+
"step": 5700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.5,
|
| 714 |
+
"learning_rate": 7.496734871571616e-06,
|
| 715 |
+
"loss": 0.0045,
|
| 716 |
+
"step": 5750
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.53,
|
| 720 |
+
"learning_rate": 7.474967348715717e-06,
|
| 721 |
+
"loss": 0.0031,
|
| 722 |
+
"step": 5800
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.55,
|
| 726 |
+
"learning_rate": 7.453199825859818e-06,
|
| 727 |
+
"loss": 0.0098,
|
| 728 |
+
"step": 5850
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.57,
|
| 732 |
+
"learning_rate": 7.431432303003919e-06,
|
| 733 |
+
"loss": 0.0055,
|
| 734 |
+
"step": 5900
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.59,
|
| 738 |
+
"learning_rate": 7.409664780148019e-06,
|
| 739 |
+
"loss": 0.0064,
|
| 740 |
+
"step": 5950
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.61,
|
| 744 |
+
"learning_rate": 7.38789725729212e-06,
|
| 745 |
+
"loss": 0.0085,
|
| 746 |
+
"step": 6000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.63,
|
| 750 |
+
"learning_rate": 7.366129734436221e-06,
|
| 751 |
+
"loss": 0.0096,
|
| 752 |
+
"step": 6050
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.66,
|
| 756 |
+
"learning_rate": 7.344362211580322e-06,
|
| 757 |
+
"loss": 0.0102,
|
| 758 |
+
"step": 6100
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.68,
|
| 762 |
+
"learning_rate": 7.3225946887244234e-06,
|
| 763 |
+
"loss": 0.0069,
|
| 764 |
+
"step": 6150
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 2.7,
|
| 768 |
+
"learning_rate": 7.3008271658685245e-06,
|
| 769 |
+
"loss": 0.0069,
|
| 770 |
+
"step": 6200
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 2.72,
|
| 774 |
+
"learning_rate": 7.279059643012626e-06,
|
| 775 |
+
"loss": 0.0071,
|
| 776 |
+
"step": 6250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.74,
|
| 780 |
+
"learning_rate": 7.257292120156727e-06,
|
| 781 |
+
"loss": 0.0114,
|
| 782 |
+
"step": 6300
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.76,
|
| 786 |
+
"learning_rate": 7.235524597300828e-06,
|
| 787 |
+
"loss": 0.0083,
|
| 788 |
+
"step": 6350
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.79,
|
| 792 |
+
"learning_rate": 7.213757074444929e-06,
|
| 793 |
+
"loss": 0.0101,
|
| 794 |
+
"step": 6400
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 2.81,
|
| 798 |
+
"learning_rate": 7.191989551589029e-06,
|
| 799 |
+
"loss": 0.0109,
|
| 800 |
+
"step": 6450
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 2.83,
|
| 804 |
+
"learning_rate": 7.17022202873313e-06,
|
| 805 |
+
"loss": 0.0064,
|
| 806 |
+
"step": 6500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.85,
|
| 810 |
+
"learning_rate": 7.148454505877231e-06,
|
| 811 |
+
"loss": 0.0047,
|
| 812 |
+
"step": 6550
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.87,
|
| 816 |
+
"learning_rate": 7.126686983021332e-06,
|
| 817 |
+
"loss": 0.0076,
|
| 818 |
+
"step": 6600
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.9,
|
| 822 |
+
"learning_rate": 7.104919460165433e-06,
|
| 823 |
+
"loss": 0.0133,
|
| 824 |
+
"step": 6650
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 2.92,
|
| 828 |
+
"learning_rate": 7.0831519373095345e-06,
|
| 829 |
+
"loss": 0.0054,
|
| 830 |
+
"step": 6700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.94,
|
| 834 |
+
"learning_rate": 7.0613844144536355e-06,
|
| 835 |
+
"loss": 0.0108,
|
| 836 |
+
"step": 6750
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 2.96,
|
| 840 |
+
"learning_rate": 7.039616891597737e-06,
|
| 841 |
+
"loss": 0.0047,
|
| 842 |
+
"step": 6800
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.98,
|
| 846 |
+
"learning_rate": 7.017849368741838e-06,
|
| 847 |
+
"loss": 0.0088,
|
| 848 |
+
"step": 6850
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 3.0,
|
| 852 |
+
"eval_f1": 0.9989708741035986,
|
| 853 |
+
"eval_loss": 0.007210045587271452,
|
| 854 |
+
"eval_runtime": 34.5277,
|
| 855 |
+
"eval_samples_per_second": 1064.015,
|
| 856 |
+
"eval_steps_per_second": 16.653,
|
| 857 |
+
"step": 6891
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 3.0,
|
| 861 |
+
"learning_rate": 6.996081845885939e-06,
|
| 862 |
+
"loss": 0.0043,
|
| 863 |
+
"step": 6900
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 3.03,
|
| 867 |
+
"learning_rate": 6.974314323030041e-06,
|
| 868 |
+
"loss": 0.0019,
|
| 869 |
+
"step": 6950
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 3.05,
|
| 873 |
+
"learning_rate": 6.95254680017414e-06,
|
| 874 |
+
"loss": 0.0081,
|
| 875 |
+
"step": 7000
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 3.07,
|
| 879 |
+
"learning_rate": 6.930779277318241e-06,
|
| 880 |
+
"loss": 0.0078,
|
| 881 |
+
"step": 7050
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 3.09,
|
| 885 |
+
"learning_rate": 6.909011754462342e-06,
|
| 886 |
+
"loss": 0.0062,
|
| 887 |
+
"step": 7100
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 3.11,
|
| 891 |
+
"learning_rate": 6.887244231606443e-06,
|
| 892 |
+
"loss": 0.007,
|
| 893 |
+
"step": 7150
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 3.13,
|
| 897 |
+
"learning_rate": 6.865476708750544e-06,
|
| 898 |
+
"loss": 0.0047,
|
| 899 |
+
"step": 7200
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 3.16,
|
| 903 |
+
"learning_rate": 6.8437091858946455e-06,
|
| 904 |
+
"loss": 0.005,
|
| 905 |
+
"step": 7250
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 3.18,
|
| 909 |
+
"learning_rate": 6.8219416630387466e-06,
|
| 910 |
+
"loss": 0.0056,
|
| 911 |
+
"step": 7300
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 3.2,
|
| 915 |
+
"learning_rate": 6.800174140182848e-06,
|
| 916 |
+
"loss": 0.0024,
|
| 917 |
+
"step": 7350
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 3.22,
|
| 921 |
+
"learning_rate": 6.77840661732695e-06,
|
| 922 |
+
"loss": 0.005,
|
| 923 |
+
"step": 7400
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 3.24,
|
| 927 |
+
"learning_rate": 6.756639094471051e-06,
|
| 928 |
+
"loss": 0.0042,
|
| 929 |
+
"step": 7450
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 3.27,
|
| 933 |
+
"learning_rate": 6.73487157161515e-06,
|
| 934 |
+
"loss": 0.0057,
|
| 935 |
+
"step": 7500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 3.29,
|
| 939 |
+
"learning_rate": 6.713104048759251e-06,
|
| 940 |
+
"loss": 0.0015,
|
| 941 |
+
"step": 7550
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 3.31,
|
| 945 |
+
"learning_rate": 6.691336525903352e-06,
|
| 946 |
+
"loss": 0.0017,
|
| 947 |
+
"step": 7600
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 3.33,
|
| 951 |
+
"learning_rate": 6.669569003047453e-06,
|
| 952 |
+
"loss": 0.0057,
|
| 953 |
+
"step": 7650
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 3.35,
|
| 957 |
+
"learning_rate": 6.647801480191554e-06,
|
| 958 |
+
"loss": 0.0025,
|
| 959 |
+
"step": 7700
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 3.37,
|
| 963 |
+
"learning_rate": 6.6260339573356554e-06,
|
| 964 |
+
"loss": 0.0075,
|
| 965 |
+
"step": 7750
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 3.4,
|
| 969 |
+
"learning_rate": 6.6042664344797565e-06,
|
| 970 |
+
"loss": 0.0059,
|
| 971 |
+
"step": 7800
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 3.42,
|
| 975 |
+
"learning_rate": 6.582498911623858e-06,
|
| 976 |
+
"loss": 0.006,
|
| 977 |
+
"step": 7850
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 3.44,
|
| 981 |
+
"learning_rate": 6.5607313887679595e-06,
|
| 982 |
+
"loss": 0.0053,
|
| 983 |
+
"step": 7900
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 3.46,
|
| 987 |
+
"learning_rate": 6.538963865912061e-06,
|
| 988 |
+
"loss": 0.0116,
|
| 989 |
+
"step": 7950
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 3.48,
|
| 993 |
+
"learning_rate": 6.51719634305616e-06,
|
| 994 |
+
"loss": 0.0038,
|
| 995 |
+
"step": 8000
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 3.5,
|
| 999 |
+
"learning_rate": 6.495428820200261e-06,
|
| 1000 |
+
"loss": 0.0078,
|
| 1001 |
+
"step": 8050
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 3.53,
|
| 1005 |
+
"learning_rate": 6.473661297344362e-06,
|
| 1006 |
+
"loss": 0.0071,
|
| 1007 |
+
"step": 8100
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 3.55,
|
| 1011 |
+
"learning_rate": 6.451893774488463e-06,
|
| 1012 |
+
"loss": 0.0105,
|
| 1013 |
+
"step": 8150
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 3.57,
|
| 1017 |
+
"learning_rate": 6.430126251632564e-06,
|
| 1018 |
+
"loss": 0.0116,
|
| 1019 |
+
"step": 8200
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 3.59,
|
| 1023 |
+
"learning_rate": 6.408358728776665e-06,
|
| 1024 |
+
"loss": 0.0076,
|
| 1025 |
+
"step": 8250
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 3.61,
|
| 1029 |
+
"learning_rate": 6.3865912059207665e-06,
|
| 1030 |
+
"loss": 0.0029,
|
| 1031 |
+
"step": 8300
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 3.64,
|
| 1035 |
+
"learning_rate": 6.364823683064868e-06,
|
| 1036 |
+
"loss": 0.0098,
|
| 1037 |
+
"step": 8350
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 3.66,
|
| 1041 |
+
"learning_rate": 6.3430561602089695e-06,
|
| 1042 |
+
"loss": 0.0084,
|
| 1043 |
+
"step": 8400
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 3.68,
|
| 1047 |
+
"learning_rate": 6.3212886373530706e-06,
|
| 1048 |
+
"loss": 0.0068,
|
| 1049 |
+
"step": 8450
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 3.7,
|
| 1053 |
+
"learning_rate": 6.29952111449717e-06,
|
| 1054 |
+
"loss": 0.0071,
|
| 1055 |
+
"step": 8500
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 3.72,
|
| 1059 |
+
"learning_rate": 6.277753591641271e-06,
|
| 1060 |
+
"loss": 0.0056,
|
| 1061 |
+
"step": 8550
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 3.74,
|
| 1065 |
+
"learning_rate": 6.255986068785372e-06,
|
| 1066 |
+
"loss": 0.0085,
|
| 1067 |
+
"step": 8600
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 3.77,
|
| 1071 |
+
"learning_rate": 6.234218545929473e-06,
|
| 1072 |
+
"loss": 0.0014,
|
| 1073 |
+
"step": 8650
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 3.79,
|
| 1077 |
+
"learning_rate": 6.212451023073574e-06,
|
| 1078 |
+
"loss": 0.0068,
|
| 1079 |
+
"step": 8700
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.81,
|
| 1083 |
+
"learning_rate": 6.190683500217675e-06,
|
| 1084 |
+
"loss": 0.0091,
|
| 1085 |
+
"step": 8750
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 3.83,
|
| 1089 |
+
"learning_rate": 6.168915977361776e-06,
|
| 1090 |
+
"loss": 0.0043,
|
| 1091 |
+
"step": 8800
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 3.85,
|
| 1095 |
+
"learning_rate": 6.147148454505878e-06,
|
| 1096 |
+
"loss": 0.0068,
|
| 1097 |
+
"step": 8850
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 3.87,
|
| 1101 |
+
"learning_rate": 6.1253809316499794e-06,
|
| 1102 |
+
"loss": 0.0046,
|
| 1103 |
+
"step": 8900
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 3.9,
|
| 1107 |
+
"learning_rate": 6.1036134087940805e-06,
|
| 1108 |
+
"loss": 0.008,
|
| 1109 |
+
"step": 8950
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 3.92,
|
| 1113 |
+
"learning_rate": 6.081845885938182e-06,
|
| 1114 |
+
"loss": 0.0076,
|
| 1115 |
+
"step": 9000
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 3.94,
|
| 1119 |
+
"learning_rate": 6.060078363082281e-06,
|
| 1120 |
+
"loss": 0.0049,
|
| 1121 |
+
"step": 9050
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 3.96,
|
| 1125 |
+
"learning_rate": 6.038310840226382e-06,
|
| 1126 |
+
"loss": 0.0073,
|
| 1127 |
+
"step": 9100
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 3.98,
|
| 1131 |
+
"learning_rate": 6.016543317370483e-06,
|
| 1132 |
+
"loss": 0.0073,
|
| 1133 |
+
"step": 9150
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 4.0,
|
| 1137 |
+
"eval_f1": 0.9991011162501838,
|
| 1138 |
+
"eval_loss": 0.00564388744533062,
|
| 1139 |
+
"eval_runtime": 36.1615,
|
| 1140 |
+
"eval_samples_per_second": 1015.942,
|
| 1141 |
+
"eval_steps_per_second": 15.901,
|
| 1142 |
+
"step": 9188
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 4.01,
|
| 1146 |
+
"learning_rate": 5.994775794514584e-06,
|
| 1147 |
+
"loss": 0.004,
|
| 1148 |
+
"step": 9200
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 4.03,
|
| 1152 |
+
"learning_rate": 5.973008271658685e-06,
|
| 1153 |
+
"loss": 0.0011,
|
| 1154 |
+
"step": 9250
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 4.05,
|
| 1158 |
+
"learning_rate": 5.951240748802787e-06,
|
| 1159 |
+
"loss": 0.0039,
|
| 1160 |
+
"step": 9300
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 4.07,
|
| 1164 |
+
"learning_rate": 5.929473225946888e-06,
|
| 1165 |
+
"loss": 0.0087,
|
| 1166 |
+
"step": 9350
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 4.09,
|
| 1170 |
+
"learning_rate": 5.907705703090989e-06,
|
| 1171 |
+
"loss": 0.0047,
|
| 1172 |
+
"step": 9400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 4.11,
|
| 1176 |
+
"learning_rate": 5.8859381802350905e-06,
|
| 1177 |
+
"loss": 0.0026,
|
| 1178 |
+
"step": 9450
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 4.14,
|
| 1182 |
+
"learning_rate": 5.8641706573791915e-06,
|
| 1183 |
+
"loss": 0.0054,
|
| 1184 |
+
"step": 9500
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 4.16,
|
| 1188 |
+
"learning_rate": 5.842403134523291e-06,
|
| 1189 |
+
"loss": 0.0049,
|
| 1190 |
+
"step": 9550
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 4.18,
|
| 1194 |
+
"learning_rate": 5.820635611667392e-06,
|
| 1195 |
+
"loss": 0.003,
|
| 1196 |
+
"step": 9600
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 4.2,
|
| 1200 |
+
"learning_rate": 5.798868088811493e-06,
|
| 1201 |
+
"loss": 0.0053,
|
| 1202 |
+
"step": 9650
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 4.22,
|
| 1206 |
+
"learning_rate": 5.777100565955594e-06,
|
| 1207 |
+
"loss": 0.0002,
|
| 1208 |
+
"step": 9700
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 4.24,
|
| 1212 |
+
"learning_rate": 5.755333043099695e-06,
|
| 1213 |
+
"loss": 0.003,
|
| 1214 |
+
"step": 9750
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 4.27,
|
| 1218 |
+
"learning_rate": 5.733565520243797e-06,
|
| 1219 |
+
"loss": 0.0095,
|
| 1220 |
+
"step": 9800
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 4.29,
|
| 1224 |
+
"learning_rate": 5.711797997387898e-06,
|
| 1225 |
+
"loss": 0.0023,
|
| 1226 |
+
"step": 9850
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 4.31,
|
| 1230 |
+
"learning_rate": 5.690030474531999e-06,
|
| 1231 |
+
"loss": 0.0033,
|
| 1232 |
+
"step": 9900
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 4.33,
|
| 1236 |
+
"learning_rate": 5.6682629516761e-06,
|
| 1237 |
+
"loss": 0.0036,
|
| 1238 |
+
"step": 9950
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 4.35,
|
| 1242 |
+
"learning_rate": 5.6464954288202015e-06,
|
| 1243 |
+
"loss": 0.0078,
|
| 1244 |
+
"step": 10000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 4.38,
|
| 1248 |
+
"learning_rate": 5.624727905964301e-06,
|
| 1249 |
+
"loss": 0.0092,
|
| 1250 |
+
"step": 10050
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 4.4,
|
| 1254 |
+
"learning_rate": 5.602960383108402e-06,
|
| 1255 |
+
"loss": 0.0052,
|
| 1256 |
+
"step": 10100
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 4.42,
|
| 1260 |
+
"learning_rate": 5.581192860252503e-06,
|
| 1261 |
+
"loss": 0.0018,
|
| 1262 |
+
"step": 10150
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 4.44,
|
| 1266 |
+
"learning_rate": 5.559425337396604e-06,
|
| 1267 |
+
"loss": 0.0033,
|
| 1268 |
+
"step": 10200
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 4.46,
|
| 1272 |
+
"learning_rate": 5.537657814540706e-06,
|
| 1273 |
+
"loss": 0.0054,
|
| 1274 |
+
"step": 10250
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 4.48,
|
| 1278 |
+
"learning_rate": 5.515890291684807e-06,
|
| 1279 |
+
"loss": 0.0069,
|
| 1280 |
+
"step": 10300
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 4.51,
|
| 1284 |
+
"learning_rate": 5.494122768828908e-06,
|
| 1285 |
+
"loss": 0.0062,
|
| 1286 |
+
"step": 10350
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 4.53,
|
| 1290 |
+
"learning_rate": 5.472355245973009e-06,
|
| 1291 |
+
"loss": 0.0014,
|
| 1292 |
+
"step": 10400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 4.55,
|
| 1296 |
+
"learning_rate": 5.45058772311711e-06,
|
| 1297 |
+
"loss": 0.0048,
|
| 1298 |
+
"step": 10450
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 4.57,
|
| 1302 |
+
"learning_rate": 5.4288202002612114e-06,
|
| 1303 |
+
"loss": 0.0037,
|
| 1304 |
+
"step": 10500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 4.59,
|
| 1308 |
+
"learning_rate": 5.4070526774053125e-06,
|
| 1309 |
+
"loss": 0.0037,
|
| 1310 |
+
"step": 10550
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 4.61,
|
| 1314 |
+
"learning_rate": 5.385285154549412e-06,
|
| 1315 |
+
"loss": 0.0042,
|
| 1316 |
+
"step": 10600
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 4.64,
|
| 1320 |
+
"learning_rate": 5.363517631693513e-06,
|
| 1321 |
+
"loss": 0.0011,
|
| 1322 |
+
"step": 10650
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 4.66,
|
| 1326 |
+
"learning_rate": 5.341750108837614e-06,
|
| 1327 |
+
"loss": 0.0035,
|
| 1328 |
+
"step": 10700
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 4.68,
|
| 1332 |
+
"learning_rate": 5.319982585981716e-06,
|
| 1333 |
+
"loss": 0.0036,
|
| 1334 |
+
"step": 10750
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 4.7,
|
| 1338 |
+
"learning_rate": 5.298215063125817e-06,
|
| 1339 |
+
"loss": 0.0063,
|
| 1340 |
+
"step": 10800
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 4.72,
|
| 1344 |
+
"learning_rate": 5.276447540269918e-06,
|
| 1345 |
+
"loss": 0.0048,
|
| 1346 |
+
"step": 10850
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 4.75,
|
| 1350 |
+
"learning_rate": 5.254680017414019e-06,
|
| 1351 |
+
"loss": 0.004,
|
| 1352 |
+
"step": 10900
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 4.77,
|
| 1356 |
+
"learning_rate": 5.23291249455812e-06,
|
| 1357 |
+
"loss": 0.0055,
|
| 1358 |
+
"step": 10950
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 4.79,
|
| 1362 |
+
"learning_rate": 5.211144971702221e-06,
|
| 1363 |
+
"loss": 0.006,
|
| 1364 |
+
"step": 11000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 4.81,
|
| 1368 |
+
"learning_rate": 5.1893774488463225e-06,
|
| 1369 |
+
"loss": 0.0068,
|
| 1370 |
+
"step": 11050
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 4.83,
|
| 1374 |
+
"learning_rate": 5.167609925990422e-06,
|
| 1375 |
+
"loss": 0.0043,
|
| 1376 |
+
"step": 11100
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 4.85,
|
| 1380 |
+
"learning_rate": 5.145842403134523e-06,
|
| 1381 |
+
"loss": 0.0036,
|
| 1382 |
+
"step": 11150
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 4.88,
|
| 1386 |
+
"learning_rate": 5.124074880278625e-06,
|
| 1387 |
+
"loss": 0.0007,
|
| 1388 |
+
"step": 11200
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 4.9,
|
| 1392 |
+
"learning_rate": 5.102307357422726e-06,
|
| 1393 |
+
"loss": 0.0034,
|
| 1394 |
+
"step": 11250
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 4.92,
|
| 1398 |
+
"learning_rate": 5.080539834566827e-06,
|
| 1399 |
+
"loss": 0.0063,
|
| 1400 |
+
"step": 11300
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 4.94,
|
| 1404 |
+
"learning_rate": 5.058772311710928e-06,
|
| 1405 |
+
"loss": 0.0003,
|
| 1406 |
+
"step": 11350
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 4.96,
|
| 1410 |
+
"learning_rate": 5.037004788855029e-06,
|
| 1411 |
+
"loss": 0.0042,
|
| 1412 |
+
"step": 11400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 4.98,
|
| 1416 |
+
"learning_rate": 5.01523726599913e-06,
|
| 1417 |
+
"loss": 0.0086,
|
| 1418 |
+
"step": 11450
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 5.0,
|
| 1422 |
+
"eval_f1": 0.9988390915482595,
|
| 1423 |
+
"eval_loss": 0.008038449101150036,
|
| 1424 |
+
"eval_runtime": 35.4168,
|
| 1425 |
+
"eval_samples_per_second": 1037.303,
|
| 1426 |
+
"eval_steps_per_second": 16.235,
|
| 1427 |
+
"step": 11485
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 5.01,
|
| 1431 |
+
"learning_rate": 4.9934697431432305e-06,
|
| 1432 |
+
"loss": 0.0026,
|
| 1433 |
+
"step": 11500
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 5.03,
|
| 1437 |
+
"learning_rate": 4.9717022202873316e-06,
|
| 1438 |
+
"loss": 0.0005,
|
| 1439 |
+
"step": 11550
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 5.05,
|
| 1443 |
+
"learning_rate": 4.949934697431433e-06,
|
| 1444 |
+
"loss": 0.0009,
|
| 1445 |
+
"step": 11600
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 5.07,
|
| 1449 |
+
"learning_rate": 4.928167174575534e-06,
|
| 1450 |
+
"loss": 0.0026,
|
| 1451 |
+
"step": 11650
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 5.09,
|
| 1455 |
+
"learning_rate": 4.906399651719635e-06,
|
| 1456 |
+
"loss": 0.0026,
|
| 1457 |
+
"step": 11700
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 5.12,
|
| 1461 |
+
"learning_rate": 4.884632128863736e-06,
|
| 1462 |
+
"loss": 0.0095,
|
| 1463 |
+
"step": 11750
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 5.14,
|
| 1467 |
+
"learning_rate": 4.862864606007837e-06,
|
| 1468 |
+
"loss": 0.0036,
|
| 1469 |
+
"step": 11800
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 5.16,
|
| 1473 |
+
"learning_rate": 4.841097083151938e-06,
|
| 1474 |
+
"loss": 0.0001,
|
| 1475 |
+
"step": 11850
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 5.18,
|
| 1479 |
+
"learning_rate": 4.819329560296039e-06,
|
| 1480 |
+
"loss": 0.0036,
|
| 1481 |
+
"step": 11900
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 5.2,
|
| 1485 |
+
"learning_rate": 4.79756203744014e-06,
|
| 1486 |
+
"loss": 0.0086,
|
| 1487 |
+
"step": 11950
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 5.22,
|
| 1491 |
+
"learning_rate": 4.7757945145842404e-06,
|
| 1492 |
+
"loss": 0.0046,
|
| 1493 |
+
"step": 12000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 5.25,
|
| 1497 |
+
"learning_rate": 4.7540269917283415e-06,
|
| 1498 |
+
"loss": 0.0015,
|
| 1499 |
+
"step": 12050
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 5.27,
|
| 1503 |
+
"learning_rate": 4.732259468872443e-06,
|
| 1504 |
+
"loss": 0.0018,
|
| 1505 |
+
"step": 12100
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 5.29,
|
| 1509 |
+
"learning_rate": 4.710491946016544e-06,
|
| 1510 |
+
"loss": 0.0041,
|
| 1511 |
+
"step": 12150
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 5.31,
|
| 1515 |
+
"learning_rate": 4.688724423160645e-06,
|
| 1516 |
+
"loss": 0.0036,
|
| 1517 |
+
"step": 12200
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 5.33,
|
| 1521 |
+
"learning_rate": 4.666956900304746e-06,
|
| 1522 |
+
"loss": 0.0063,
|
| 1523 |
+
"step": 12250
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 5.35,
|
| 1527 |
+
"learning_rate": 4.645189377448847e-06,
|
| 1528 |
+
"loss": 0.0036,
|
| 1529 |
+
"step": 12300
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 5.38,
|
| 1533 |
+
"learning_rate": 4.623421854592948e-06,
|
| 1534 |
+
"loss": 0.002,
|
| 1535 |
+
"step": 12350
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 5.4,
|
| 1539 |
+
"learning_rate": 4.601654331737049e-06,
|
| 1540 |
+
"loss": 0.0006,
|
| 1541 |
+
"step": 12400
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 5.42,
|
| 1545 |
+
"learning_rate": 4.57988680888115e-06,
|
| 1546 |
+
"loss": 0.0041,
|
| 1547 |
+
"step": 12450
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 5.44,
|
| 1551 |
+
"learning_rate": 4.55811928602525e-06,
|
| 1552 |
+
"loss": 0.002,
|
| 1553 |
+
"step": 12500
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 5.46,
|
| 1557 |
+
"learning_rate": 4.5363517631693515e-06,
|
| 1558 |
+
"loss": 0.0053,
|
| 1559 |
+
"step": 12550
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 5.49,
|
| 1563 |
+
"learning_rate": 4.5145842403134525e-06,
|
| 1564 |
+
"loss": 0.0027,
|
| 1565 |
+
"step": 12600
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 5.51,
|
| 1569 |
+
"learning_rate": 4.492816717457554e-06,
|
| 1570 |
+
"loss": 0.0056,
|
| 1571 |
+
"step": 12650
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 5.53,
|
| 1575 |
+
"learning_rate": 4.471049194601655e-06,
|
| 1576 |
+
"loss": 0.002,
|
| 1577 |
+
"step": 12700
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 5.55,
|
| 1581 |
+
"learning_rate": 4.449281671745756e-06,
|
| 1582 |
+
"loss": 0.0007,
|
| 1583 |
+
"step": 12750
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 5.57,
|
| 1587 |
+
"learning_rate": 4.427514148889857e-06,
|
| 1588 |
+
"loss": 0.0026,
|
| 1589 |
+
"step": 12800
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 5.59,
|
| 1593 |
+
"learning_rate": 4.405746626033958e-06,
|
| 1594 |
+
"loss": 0.0021,
|
| 1595 |
+
"step": 12850
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 5.62,
|
| 1599 |
+
"learning_rate": 4.383979103178059e-06,
|
| 1600 |
+
"loss": 0.001,
|
| 1601 |
+
"step": 12900
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 5.64,
|
| 1605 |
+
"learning_rate": 4.36221158032216e-06,
|
| 1606 |
+
"loss": 0.0043,
|
| 1607 |
+
"step": 12950
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 5.66,
|
| 1611 |
+
"learning_rate": 4.340444057466261e-06,
|
| 1612 |
+
"loss": 0.0009,
|
| 1613 |
+
"step": 13000
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 5.68,
|
| 1617 |
+
"learning_rate": 4.318676534610361e-06,
|
| 1618 |
+
"loss": 0.0073,
|
| 1619 |
+
"step": 13050
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 5.7,
|
| 1623 |
+
"learning_rate": 4.2969090117544625e-06,
|
| 1624 |
+
"loss": 0.0039,
|
| 1625 |
+
"step": 13100
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 5.72,
|
| 1629 |
+
"learning_rate": 4.2751414888985636e-06,
|
| 1630 |
+
"loss": 0.004,
|
| 1631 |
+
"step": 13150
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 5.75,
|
| 1635 |
+
"learning_rate": 4.253373966042665e-06,
|
| 1636 |
+
"loss": 0.0024,
|
| 1637 |
+
"step": 13200
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 5.77,
|
| 1641 |
+
"learning_rate": 4.231606443186766e-06,
|
| 1642 |
+
"loss": 0.0015,
|
| 1643 |
+
"step": 13250
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 5.79,
|
| 1647 |
+
"learning_rate": 4.209838920330867e-06,
|
| 1648 |
+
"loss": 0.0015,
|
| 1649 |
+
"step": 13300
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 5.81,
|
| 1653 |
+
"learning_rate": 4.188071397474968e-06,
|
| 1654 |
+
"loss": 0.0061,
|
| 1655 |
+
"step": 13350
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 5.83,
|
| 1659 |
+
"learning_rate": 4.166303874619069e-06,
|
| 1660 |
+
"loss": 0.0027,
|
| 1661 |
+
"step": 13400
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 5.86,
|
| 1665 |
+
"learning_rate": 4.14453635176317e-06,
|
| 1666 |
+
"loss": 0.0044,
|
| 1667 |
+
"step": 13450
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 5.88,
|
| 1671 |
+
"learning_rate": 4.122768828907271e-06,
|
| 1672 |
+
"loss": 0.004,
|
| 1673 |
+
"step": 13500
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 5.9,
|
| 1677 |
+
"learning_rate": 4.101001306051371e-06,
|
| 1678 |
+
"loss": 0.0018,
|
| 1679 |
+
"step": 13550
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 5.92,
|
| 1683 |
+
"learning_rate": 4.0792337831954724e-06,
|
| 1684 |
+
"loss": 0.003,
|
| 1685 |
+
"step": 13600
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 5.94,
|
| 1689 |
+
"learning_rate": 4.0574662603395735e-06,
|
| 1690 |
+
"loss": 0.0015,
|
| 1691 |
+
"step": 13650
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 5.96,
|
| 1695 |
+
"learning_rate": 4.035698737483675e-06,
|
| 1696 |
+
"loss": 0.0072,
|
| 1697 |
+
"step": 13700
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 5.99,
|
| 1701 |
+
"learning_rate": 4.013931214627776e-06,
|
| 1702 |
+
"loss": 0.001,
|
| 1703 |
+
"step": 13750
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 6.0,
|
| 1707 |
+
"eval_f1": 0.9992156862745099,
|
| 1708 |
+
"eval_loss": 0.006630082614719868,
|
| 1709 |
+
"eval_runtime": 34.2419,
|
| 1710 |
+
"eval_samples_per_second": 1072.896,
|
| 1711 |
+
"eval_steps_per_second": 16.792,
|
| 1712 |
+
"step": 13782
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 6.01,
|
| 1716 |
+
"learning_rate": 3.992163691771877e-06,
|
| 1717 |
+
"loss": 0.0017,
|
| 1718 |
+
"step": 13800
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 6.03,
|
| 1722 |
+
"learning_rate": 3.970396168915978e-06,
|
| 1723 |
+
"loss": 0.0016,
|
| 1724 |
+
"step": 13850
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 6.05,
|
| 1728 |
+
"learning_rate": 3.948628646060079e-06,
|
| 1729 |
+
"loss": 0.0068,
|
| 1730 |
+
"step": 13900
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 6.07,
|
| 1734 |
+
"learning_rate": 3.92686112320418e-06,
|
| 1735 |
+
"loss": 0.006,
|
| 1736 |
+
"step": 13950
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 6.09,
|
| 1740 |
+
"learning_rate": 3.905093600348281e-06,
|
| 1741 |
+
"loss": 0.0032,
|
| 1742 |
+
"step": 14000
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 6.12,
|
| 1746 |
+
"learning_rate": 3.883326077492381e-06,
|
| 1747 |
+
"loss": 0.0055,
|
| 1748 |
+
"step": 14050
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 6.14,
|
| 1752 |
+
"learning_rate": 3.861558554636482e-06,
|
| 1753 |
+
"loss": 0.0002,
|
| 1754 |
+
"step": 14100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 6.16,
|
| 1758 |
+
"learning_rate": 3.8397910317805835e-06,
|
| 1759 |
+
"loss": 0.0009,
|
| 1760 |
+
"step": 14150
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 6.18,
|
| 1764 |
+
"learning_rate": 3.8180235089246845e-06,
|
| 1765 |
+
"loss": 0.0022,
|
| 1766 |
+
"step": 14200
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 6.2,
|
| 1770 |
+
"learning_rate": 3.796255986068786e-06,
|
| 1771 |
+
"loss": 0.0039,
|
| 1772 |
+
"step": 14250
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 6.23,
|
| 1776 |
+
"learning_rate": 3.7744884632128863e-06,
|
| 1777 |
+
"loss": 0.0039,
|
| 1778 |
+
"step": 14300
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 6.25,
|
| 1782 |
+
"learning_rate": 3.7527209403569874e-06,
|
| 1783 |
+
"loss": 0.0045,
|
| 1784 |
+
"step": 14350
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 6.27,
|
| 1788 |
+
"learning_rate": 3.7309534175010884e-06,
|
| 1789 |
+
"loss": 0.0024,
|
| 1790 |
+
"step": 14400
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 6.29,
|
| 1794 |
+
"learning_rate": 3.70918589464519e-06,
|
| 1795 |
+
"loss": 0.0027,
|
| 1796 |
+
"step": 14450
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 6.31,
|
| 1800 |
+
"learning_rate": 3.687418371789291e-06,
|
| 1801 |
+
"loss": 0.0054,
|
| 1802 |
+
"step": 14500
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 6.33,
|
| 1806 |
+
"learning_rate": 3.6656508489333912e-06,
|
| 1807 |
+
"loss": 0.0038,
|
| 1808 |
+
"step": 14550
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 6.36,
|
| 1812 |
+
"learning_rate": 3.6438833260774923e-06,
|
| 1813 |
+
"loss": 0.001,
|
| 1814 |
+
"step": 14600
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 6.38,
|
| 1818 |
+
"learning_rate": 3.6221158032215934e-06,
|
| 1819 |
+
"loss": 0.0039,
|
| 1820 |
+
"step": 14650
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 6.4,
|
| 1824 |
+
"learning_rate": 3.600348280365695e-06,
|
| 1825 |
+
"loss": 0.0018,
|
| 1826 |
+
"step": 14700
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 6.42,
|
| 1830 |
+
"learning_rate": 3.578580757509796e-06,
|
| 1831 |
+
"loss": 0.0002,
|
| 1832 |
+
"step": 14750
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 6.44,
|
| 1836 |
+
"learning_rate": 3.556813234653897e-06,
|
| 1837 |
+
"loss": 0.0009,
|
| 1838 |
+
"step": 14800
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 6.46,
|
| 1842 |
+
"learning_rate": 3.5350457117979973e-06,
|
| 1843 |
+
"loss": 0.0008,
|
| 1844 |
+
"step": 14850
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 6.49,
|
| 1848 |
+
"learning_rate": 3.513278188942099e-06,
|
| 1849 |
+
"loss": 0.0032,
|
| 1850 |
+
"step": 14900
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 6.51,
|
| 1854 |
+
"learning_rate": 3.4915106660862e-06,
|
| 1855 |
+
"loss": 0.0004,
|
| 1856 |
+
"step": 14950
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 6.53,
|
| 1860 |
+
"learning_rate": 3.469743143230301e-06,
|
| 1861 |
+
"loss": 0.0003,
|
| 1862 |
+
"step": 15000
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 6.55,
|
| 1866 |
+
"learning_rate": 3.447975620374402e-06,
|
| 1867 |
+
"loss": 0.0012,
|
| 1868 |
+
"step": 15050
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 6.57,
|
| 1872 |
+
"learning_rate": 3.4262080975185023e-06,
|
| 1873 |
+
"loss": 0.0014,
|
| 1874 |
+
"step": 15100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 6.6,
|
| 1878 |
+
"learning_rate": 3.4044405746626038e-06,
|
| 1879 |
+
"loss": 0.0003,
|
| 1880 |
+
"step": 15150
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 6.62,
|
| 1884 |
+
"learning_rate": 3.382673051806705e-06,
|
| 1885 |
+
"loss": 0.0063,
|
| 1886 |
+
"step": 15200
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 6.64,
|
| 1890 |
+
"learning_rate": 3.360905528950806e-06,
|
| 1891 |
+
"loss": 0.0043,
|
| 1892 |
+
"step": 15250
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 6.66,
|
| 1896 |
+
"learning_rate": 3.339138006094907e-06,
|
| 1897 |
+
"loss": 0.003,
|
| 1898 |
+
"step": 15300
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 6.68,
|
| 1902 |
+
"learning_rate": 3.3173704832390073e-06,
|
| 1903 |
+
"loss": 0.0061,
|
| 1904 |
+
"step": 15350
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 6.7,
|
| 1908 |
+
"learning_rate": 3.2956029603831088e-06,
|
| 1909 |
+
"loss": 0.0005,
|
| 1910 |
+
"step": 15400
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 6.73,
|
| 1914 |
+
"learning_rate": 3.27383543752721e-06,
|
| 1915 |
+
"loss": 0.003,
|
| 1916 |
+
"step": 15450
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 6.75,
|
| 1920 |
+
"learning_rate": 3.252067914671311e-06,
|
| 1921 |
+
"loss": 0.0057,
|
| 1922 |
+
"step": 15500
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 6.77,
|
| 1926 |
+
"learning_rate": 3.230300391815412e-06,
|
| 1927 |
+
"loss": 0.0017,
|
| 1928 |
+
"step": 15550
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 6.79,
|
| 1932 |
+
"learning_rate": 3.2085328689595122e-06,
|
| 1933 |
+
"loss": 0.0041,
|
| 1934 |
+
"step": 15600
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 6.81,
|
| 1938 |
+
"learning_rate": 3.1867653461036137e-06,
|
| 1939 |
+
"loss": 0.0009,
|
| 1940 |
+
"step": 15650
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 6.84,
|
| 1944 |
+
"learning_rate": 3.164997823247715e-06,
|
| 1945 |
+
"loss": 0.0026,
|
| 1946 |
+
"step": 15700
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 6.86,
|
| 1950 |
+
"learning_rate": 3.143230300391816e-06,
|
| 1951 |
+
"loss": 0.0034,
|
| 1952 |
+
"step": 15750
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 6.88,
|
| 1956 |
+
"learning_rate": 3.121462777535917e-06,
|
| 1957 |
+
"loss": 0.0008,
|
| 1958 |
+
"step": 15800
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 6.9,
|
| 1962 |
+
"learning_rate": 3.0996952546800176e-06,
|
| 1963 |
+
"loss": 0.0012,
|
| 1964 |
+
"step": 15850
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 6.92,
|
| 1968 |
+
"learning_rate": 3.0779277318241187e-06,
|
| 1969 |
+
"loss": 0.0012,
|
| 1970 |
+
"step": 15900
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 6.94,
|
| 1974 |
+
"learning_rate": 3.0561602089682198e-06,
|
| 1975 |
+
"loss": 0.0034,
|
| 1976 |
+
"step": 15950
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 6.97,
|
| 1980 |
+
"learning_rate": 3.034392686112321e-06,
|
| 1981 |
+
"loss": 0.0039,
|
| 1982 |
+
"step": 16000
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 6.99,
|
| 1986 |
+
"learning_rate": 3.012625163256422e-06,
|
| 1987 |
+
"loss": 0.0001,
|
| 1988 |
+
"step": 16050
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 7.0,
|
| 1992 |
+
"eval_f1": 0.9990852961353762,
|
| 1993 |
+
"eval_loss": 0.007986071519553661,
|
| 1994 |
+
"eval_runtime": 33.4792,
|
| 1995 |
+
"eval_samples_per_second": 1097.337,
|
| 1996 |
+
"eval_steps_per_second": 17.175,
|
| 1997 |
+
"step": 16079
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 7.01,
|
| 2001 |
+
"learning_rate": 2.9908576404005226e-06,
|
| 2002 |
+
"loss": 0.0021,
|
| 2003 |
+
"step": 16100
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 7.03,
|
| 2007 |
+
"learning_rate": 2.9690901175446237e-06,
|
| 2008 |
+
"loss": 0.001,
|
| 2009 |
+
"step": 16150
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 7.05,
|
| 2013 |
+
"learning_rate": 2.9473225946887248e-06,
|
| 2014 |
+
"loss": 0.0024,
|
| 2015 |
+
"step": 16200
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 7.07,
|
| 2019 |
+
"learning_rate": 2.925555071832826e-06,
|
| 2020 |
+
"loss": 0.0024,
|
| 2021 |
+
"step": 16250
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 7.1,
|
| 2025 |
+
"learning_rate": 2.903787548976927e-06,
|
| 2026 |
+
"loss": 0.0028,
|
| 2027 |
+
"step": 16300
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 7.12,
|
| 2031 |
+
"learning_rate": 2.882020026121028e-06,
|
| 2032 |
+
"loss": 0.0004,
|
| 2033 |
+
"step": 16350
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 7.14,
|
| 2037 |
+
"learning_rate": 2.8602525032651287e-06,
|
| 2038 |
+
"loss": 0.0013,
|
| 2039 |
+
"step": 16400
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 7.16,
|
| 2043 |
+
"learning_rate": 2.8384849804092297e-06,
|
| 2044 |
+
"loss": 0.0007,
|
| 2045 |
+
"step": 16450
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 7.18,
|
| 2049 |
+
"learning_rate": 2.816717457553331e-06,
|
| 2050 |
+
"loss": 0.0001,
|
| 2051 |
+
"step": 16500
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 7.21,
|
| 2055 |
+
"learning_rate": 2.794949934697432e-06,
|
| 2056 |
+
"loss": 0.0056,
|
| 2057 |
+
"step": 16550
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 7.23,
|
| 2061 |
+
"learning_rate": 2.773182411841533e-06,
|
| 2062 |
+
"loss": 0.0041,
|
| 2063 |
+
"step": 16600
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 7.25,
|
| 2067 |
+
"learning_rate": 2.7514148889856336e-06,
|
| 2068 |
+
"loss": 0.0022,
|
| 2069 |
+
"step": 16650
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 7.27,
|
| 2073 |
+
"learning_rate": 2.7296473661297347e-06,
|
| 2074 |
+
"loss": 0.0003,
|
| 2075 |
+
"step": 16700
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 7.29,
|
| 2079 |
+
"learning_rate": 2.7078798432738358e-06,
|
| 2080 |
+
"loss": 0.0012,
|
| 2081 |
+
"step": 16750
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 7.31,
|
| 2085 |
+
"learning_rate": 2.686112320417937e-06,
|
| 2086 |
+
"loss": 0.0006,
|
| 2087 |
+
"step": 16800
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 7.34,
|
| 2091 |
+
"learning_rate": 2.664344797562038e-06,
|
| 2092 |
+
"loss": 0.0019,
|
| 2093 |
+
"step": 16850
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 7.36,
|
| 2097 |
+
"learning_rate": 2.6425772747061386e-06,
|
| 2098 |
+
"loss": 0.0014,
|
| 2099 |
+
"step": 16900
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 7.38,
|
| 2103 |
+
"learning_rate": 2.6208097518502397e-06,
|
| 2104 |
+
"loss": 0.0001,
|
| 2105 |
+
"step": 16950
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 7.4,
|
| 2109 |
+
"learning_rate": 2.5990422289943408e-06,
|
| 2110 |
+
"loss": 0.0048,
|
| 2111 |
+
"step": 17000
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 7.42,
|
| 2115 |
+
"learning_rate": 2.577274706138442e-06,
|
| 2116 |
+
"loss": 0.0032,
|
| 2117 |
+
"step": 17050
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 7.44,
|
| 2121 |
+
"learning_rate": 2.555507183282543e-06,
|
| 2122 |
+
"loss": 0.0001,
|
| 2123 |
+
"step": 17100
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 7.47,
|
| 2127 |
+
"learning_rate": 2.5337396604266436e-06,
|
| 2128 |
+
"loss": 0.0013,
|
| 2129 |
+
"step": 17150
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 7.49,
|
| 2133 |
+
"learning_rate": 2.5119721375707447e-06,
|
| 2134 |
+
"loss": 0.0023,
|
| 2135 |
+
"step": 17200
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 7.51,
|
| 2139 |
+
"learning_rate": 2.4902046147148457e-06,
|
| 2140 |
+
"loss": 0.0003,
|
| 2141 |
+
"step": 17250
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 7.53,
|
| 2145 |
+
"learning_rate": 2.468437091858947e-06,
|
| 2146 |
+
"loss": 0.0029,
|
| 2147 |
+
"step": 17300
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 7.55,
|
| 2151 |
+
"learning_rate": 2.4466695690030475e-06,
|
| 2152 |
+
"loss": 0.0001,
|
| 2153 |
+
"step": 17350
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 7.58,
|
| 2157 |
+
"learning_rate": 2.4249020461471485e-06,
|
| 2158 |
+
"loss": 0.0001,
|
| 2159 |
+
"step": 17400
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 7.6,
|
| 2163 |
+
"learning_rate": 2.4031345232912496e-06,
|
| 2164 |
+
"loss": 0.0009,
|
| 2165 |
+
"step": 17450
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 7.62,
|
| 2169 |
+
"learning_rate": 2.3813670004353507e-06,
|
| 2170 |
+
"loss": 0.0043,
|
| 2171 |
+
"step": 17500
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 7.64,
|
| 2175 |
+
"learning_rate": 2.3595994775794518e-06,
|
| 2176 |
+
"loss": 0.001,
|
| 2177 |
+
"step": 17550
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 7.66,
|
| 2181 |
+
"learning_rate": 2.3378319547235524e-06,
|
| 2182 |
+
"loss": 0.0011,
|
| 2183 |
+
"step": 17600
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 7.68,
|
| 2187 |
+
"learning_rate": 2.3160644318676535e-06,
|
| 2188 |
+
"loss": 0.0005,
|
| 2189 |
+
"step": 17650
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 7.71,
|
| 2193 |
+
"learning_rate": 2.2942969090117546e-06,
|
| 2194 |
+
"loss": 0.0051,
|
| 2195 |
+
"step": 17700
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 7.73,
|
| 2199 |
+
"learning_rate": 2.2725293861558557e-06,
|
| 2200 |
+
"loss": 0.003,
|
| 2201 |
+
"step": 17750
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 7.75,
|
| 2205 |
+
"learning_rate": 2.2507618632999568e-06,
|
| 2206 |
+
"loss": 0.0053,
|
| 2207 |
+
"step": 17800
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 7.77,
|
| 2211 |
+
"learning_rate": 2.2289943404440574e-06,
|
| 2212 |
+
"loss": 0.0024,
|
| 2213 |
+
"step": 17850
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 7.79,
|
| 2217 |
+
"learning_rate": 2.2072268175881585e-06,
|
| 2218 |
+
"loss": 0.0035,
|
| 2219 |
+
"step": 17900
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 7.81,
|
| 2223 |
+
"learning_rate": 2.1854592947322596e-06,
|
| 2224 |
+
"loss": 0.0033,
|
| 2225 |
+
"step": 17950
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 7.84,
|
| 2229 |
+
"learning_rate": 2.1636917718763607e-06,
|
| 2230 |
+
"loss": 0.0017,
|
| 2231 |
+
"step": 18000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 7.86,
|
| 2235 |
+
"learning_rate": 2.1419242490204617e-06,
|
| 2236 |
+
"loss": 0.0054,
|
| 2237 |
+
"step": 18050
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 7.88,
|
| 2241 |
+
"learning_rate": 2.1201567261645624e-06,
|
| 2242 |
+
"loss": 0.0047,
|
| 2243 |
+
"step": 18100
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 7.9,
|
| 2247 |
+
"learning_rate": 2.0983892033086635e-06,
|
| 2248 |
+
"loss": 0.0027,
|
| 2249 |
+
"step": 18150
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 7.92,
|
| 2253 |
+
"learning_rate": 2.0766216804527645e-06,
|
| 2254 |
+
"loss": 0.0033,
|
| 2255 |
+
"step": 18200
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 7.95,
|
| 2259 |
+
"learning_rate": 2.0548541575968656e-06,
|
| 2260 |
+
"loss": 0.0019,
|
| 2261 |
+
"step": 18250
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 7.97,
|
| 2265 |
+
"learning_rate": 2.0330866347409667e-06,
|
| 2266 |
+
"loss": 0.0008,
|
| 2267 |
+
"step": 18300
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 7.99,
|
| 2271 |
+
"learning_rate": 2.0113191118850674e-06,
|
| 2272 |
+
"loss": 0.0013,
|
| 2273 |
+
"step": 18350
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 8.0,
|
| 2277 |
+
"eval_f1": 0.9992646458043959,
|
| 2278 |
+
"eval_loss": 0.006770425010472536,
|
| 2279 |
+
"eval_runtime": 35.137,
|
| 2280 |
+
"eval_samples_per_second": 1045.565,
|
| 2281 |
+
"eval_steps_per_second": 16.365,
|
| 2282 |
+
"step": 18376
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 8.01,
|
| 2286 |
+
"learning_rate": 1.9895515890291684e-06,
|
| 2287 |
+
"loss": 0.0007,
|
| 2288 |
+
"step": 18400
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 8.03,
|
| 2292 |
+
"learning_rate": 1.9677840661732695e-06,
|
| 2293 |
+
"loss": 0.0041,
|
| 2294 |
+
"step": 18450
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 8.05,
|
| 2298 |
+
"learning_rate": 1.9460165433173706e-06,
|
| 2299 |
+
"loss": 0.001,
|
| 2300 |
+
"step": 18500
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 8.08,
|
| 2304 |
+
"learning_rate": 1.9242490204614717e-06,
|
| 2305 |
+
"loss": 0.0001,
|
| 2306 |
+
"step": 18550
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 8.1,
|
| 2310 |
+
"learning_rate": 1.9024814976055728e-06,
|
| 2311 |
+
"loss": 0.0001,
|
| 2312 |
+
"step": 18600
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 8.12,
|
| 2316 |
+
"learning_rate": 1.8807139747496736e-06,
|
| 2317 |
+
"loss": 0.0002,
|
| 2318 |
+
"step": 18650
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 8.14,
|
| 2322 |
+
"learning_rate": 1.8589464518937747e-06,
|
| 2323 |
+
"loss": 0.0014,
|
| 2324 |
+
"step": 18700
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 8.16,
|
| 2328 |
+
"learning_rate": 1.8371789290378756e-06,
|
| 2329 |
+
"loss": 0.0015,
|
| 2330 |
+
"step": 18750
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 8.18,
|
| 2334 |
+
"learning_rate": 1.8154114061819767e-06,
|
| 2335 |
+
"loss": 0.0052,
|
| 2336 |
+
"step": 18800
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 8.21,
|
| 2340 |
+
"learning_rate": 1.7936438833260777e-06,
|
| 2341 |
+
"loss": 0.0001,
|
| 2342 |
+
"step": 18850
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 8.23,
|
| 2346 |
+
"learning_rate": 1.7718763604701786e-06,
|
| 2347 |
+
"loss": 0.0001,
|
| 2348 |
+
"step": 18900
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 8.25,
|
| 2352 |
+
"learning_rate": 1.7501088376142797e-06,
|
| 2353 |
+
"loss": 0.0013,
|
| 2354 |
+
"step": 18950
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 8.27,
|
| 2358 |
+
"learning_rate": 1.7283413147583806e-06,
|
| 2359 |
+
"loss": 0.0016,
|
| 2360 |
+
"step": 19000
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 8.29,
|
| 2364 |
+
"learning_rate": 1.7065737919024816e-06,
|
| 2365 |
+
"loss": 0.0005,
|
| 2366 |
+
"step": 19050
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 8.32,
|
| 2370 |
+
"learning_rate": 1.6848062690465827e-06,
|
| 2371 |
+
"loss": 0.0016,
|
| 2372 |
+
"step": 19100
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 8.34,
|
| 2376 |
+
"learning_rate": 1.6630387461906836e-06,
|
| 2377 |
+
"loss": 0.0012,
|
| 2378 |
+
"step": 19150
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 8.36,
|
| 2382 |
+
"learning_rate": 1.6412712233347847e-06,
|
| 2383 |
+
"loss": 0.0026,
|
| 2384 |
+
"step": 19200
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 8.38,
|
| 2388 |
+
"learning_rate": 1.6195037004788855e-06,
|
| 2389 |
+
"loss": 0.0015,
|
| 2390 |
+
"step": 19250
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 8.4,
|
| 2394 |
+
"learning_rate": 1.5977361776229866e-06,
|
| 2395 |
+
"loss": 0.0011,
|
| 2396 |
+
"step": 19300
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 8.42,
|
| 2400 |
+
"learning_rate": 1.5759686547670877e-06,
|
| 2401 |
+
"loss": 0.006,
|
| 2402 |
+
"step": 19350
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 8.45,
|
| 2406 |
+
"learning_rate": 1.5542011319111886e-06,
|
| 2407 |
+
"loss": 0.0031,
|
| 2408 |
+
"step": 19400
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 8.47,
|
| 2412 |
+
"learning_rate": 1.5324336090552896e-06,
|
| 2413 |
+
"loss": 0.0018,
|
| 2414 |
+
"step": 19450
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 8.49,
|
| 2418 |
+
"learning_rate": 1.5106660861993907e-06,
|
| 2419 |
+
"loss": 0.0001,
|
| 2420 |
+
"step": 19500
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 8.51,
|
| 2424 |
+
"learning_rate": 1.4888985633434916e-06,
|
| 2425 |
+
"loss": 0.0001,
|
| 2426 |
+
"step": 19550
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 8.53,
|
| 2430 |
+
"learning_rate": 1.4671310404875927e-06,
|
| 2431 |
+
"loss": 0.0,
|
| 2432 |
+
"step": 19600
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 8.55,
|
| 2436 |
+
"learning_rate": 1.4453635176316935e-06,
|
| 2437 |
+
"loss": 0.0011,
|
| 2438 |
+
"step": 19650
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 8.58,
|
| 2442 |
+
"learning_rate": 1.4235959947757946e-06,
|
| 2443 |
+
"loss": 0.0001,
|
| 2444 |
+
"step": 19700
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 8.6,
|
| 2448 |
+
"learning_rate": 1.4018284719198957e-06,
|
| 2449 |
+
"loss": 0.0019,
|
| 2450 |
+
"step": 19750
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 8.62,
|
| 2454 |
+
"learning_rate": 1.3800609490639966e-06,
|
| 2455 |
+
"loss": 0.0027,
|
| 2456 |
+
"step": 19800
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 8.64,
|
| 2460 |
+
"learning_rate": 1.3582934262080976e-06,
|
| 2461 |
+
"loss": 0.0018,
|
| 2462 |
+
"step": 19850
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 8.66,
|
| 2466 |
+
"learning_rate": 1.3365259033521985e-06,
|
| 2467 |
+
"loss": 0.0013,
|
| 2468 |
+
"step": 19900
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 8.69,
|
| 2472 |
+
"learning_rate": 1.3147583804962996e-06,
|
| 2473 |
+
"loss": 0.0001,
|
| 2474 |
+
"step": 19950
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 8.71,
|
| 2478 |
+
"learning_rate": 1.2929908576404007e-06,
|
| 2479 |
+
"loss": 0.0003,
|
| 2480 |
+
"step": 20000
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 8.73,
|
| 2484 |
+
"learning_rate": 1.2712233347845015e-06,
|
| 2485 |
+
"loss": 0.0005,
|
| 2486 |
+
"step": 20050
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 8.75,
|
| 2490 |
+
"learning_rate": 1.2494558119286026e-06,
|
| 2491 |
+
"loss": 0.0049,
|
| 2492 |
+
"step": 20100
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 8.77,
|
| 2496 |
+
"learning_rate": 1.2276882890727037e-06,
|
| 2497 |
+
"loss": 0.0026,
|
| 2498 |
+
"step": 20150
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 8.79,
|
| 2502 |
+
"learning_rate": 1.2059207662168046e-06,
|
| 2503 |
+
"loss": 0.0014,
|
| 2504 |
+
"step": 20200
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 8.82,
|
| 2508 |
+
"learning_rate": 1.1841532433609056e-06,
|
| 2509 |
+
"loss": 0.0003,
|
| 2510 |
+
"step": 20250
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 8.84,
|
| 2514 |
+
"learning_rate": 1.1623857205050067e-06,
|
| 2515 |
+
"loss": 0.0006,
|
| 2516 |
+
"step": 20300
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 8.86,
|
| 2520 |
+
"learning_rate": 1.1406181976491076e-06,
|
| 2521 |
+
"loss": 0.0,
|
| 2522 |
+
"step": 20350
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 8.88,
|
| 2526 |
+
"learning_rate": 1.1188506747932087e-06,
|
| 2527 |
+
"loss": 0.0025,
|
| 2528 |
+
"step": 20400
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 8.9,
|
| 2532 |
+
"learning_rate": 1.0970831519373095e-06,
|
| 2533 |
+
"loss": 0.0,
|
| 2534 |
+
"step": 20450
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 8.92,
|
| 2538 |
+
"learning_rate": 1.0753156290814106e-06,
|
| 2539 |
+
"loss": 0.0001,
|
| 2540 |
+
"step": 20500
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 8.95,
|
| 2544 |
+
"learning_rate": 1.0535481062255117e-06,
|
| 2545 |
+
"loss": 0.0023,
|
| 2546 |
+
"step": 20550
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 8.97,
|
| 2550 |
+
"learning_rate": 1.0317805833696126e-06,
|
| 2551 |
+
"loss": 0.0018,
|
| 2552 |
+
"step": 20600
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 8.99,
|
| 2556 |
+
"learning_rate": 1.0100130605137136e-06,
|
| 2557 |
+
"loss": 0.0055,
|
| 2558 |
+
"step": 20650
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 9.0,
|
| 2562 |
+
"eval_f1": 0.9991502434879236,
|
| 2563 |
+
"eval_loss": 0.007731999270617962,
|
| 2564 |
+
"eval_runtime": 34.7727,
|
| 2565 |
+
"eval_samples_per_second": 1056.52,
|
| 2566 |
+
"eval_steps_per_second": 16.536,
|
| 2567 |
+
"step": 20673
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 9.01,
|
| 2571 |
+
"learning_rate": 9.882455376578145e-07,
|
| 2572 |
+
"loss": 0.0019,
|
| 2573 |
+
"step": 20700
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 9.03,
|
| 2577 |
+
"learning_rate": 9.664780148019156e-07,
|
| 2578 |
+
"loss": 0.0024,
|
| 2579 |
+
"step": 20750
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 9.06,
|
| 2583 |
+
"learning_rate": 9.447104919460167e-07,
|
| 2584 |
+
"loss": 0.0044,
|
| 2585 |
+
"step": 20800
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 9.08,
|
| 2589 |
+
"learning_rate": 9.229429690901176e-07,
|
| 2590 |
+
"loss": 0.0006,
|
| 2591 |
+
"step": 20850
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 9.1,
|
| 2595 |
+
"learning_rate": 9.011754462342186e-07,
|
| 2596 |
+
"loss": 0.0001,
|
| 2597 |
+
"step": 20900
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 9.12,
|
| 2601 |
+
"learning_rate": 8.794079233783197e-07,
|
| 2602 |
+
"loss": 0.0004,
|
| 2603 |
+
"step": 20950
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 9.14,
|
| 2607 |
+
"learning_rate": 8.576404005224207e-07,
|
| 2608 |
+
"loss": 0.0013,
|
| 2609 |
+
"step": 21000
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 9.16,
|
| 2613 |
+
"learning_rate": 8.358728776665216e-07,
|
| 2614 |
+
"loss": 0.0004,
|
| 2615 |
+
"step": 21050
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 9.19,
|
| 2619 |
+
"learning_rate": 8.141053548106226e-07,
|
| 2620 |
+
"loss": 0.001,
|
| 2621 |
+
"step": 21100
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 9.21,
|
| 2625 |
+
"learning_rate": 7.923378319547236e-07,
|
| 2626 |
+
"loss": 0.0005,
|
| 2627 |
+
"step": 21150
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 9.23,
|
| 2631 |
+
"learning_rate": 7.705703090988247e-07,
|
| 2632 |
+
"loss": 0.0008,
|
| 2633 |
+
"step": 21200
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 9.25,
|
| 2637 |
+
"learning_rate": 7.488027862429256e-07,
|
| 2638 |
+
"loss": 0.0004,
|
| 2639 |
+
"step": 21250
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 9.27,
|
| 2643 |
+
"learning_rate": 7.270352633870266e-07,
|
| 2644 |
+
"loss": 0.0019,
|
| 2645 |
+
"step": 21300
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 9.29,
|
| 2649 |
+
"learning_rate": 7.052677405311276e-07,
|
| 2650 |
+
"loss": 0.0003,
|
| 2651 |
+
"step": 21350
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 9.32,
|
| 2655 |
+
"learning_rate": 6.835002176752287e-07,
|
| 2656 |
+
"loss": 0.0003,
|
| 2657 |
+
"step": 21400
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 9.34,
|
| 2661 |
+
"learning_rate": 6.617326948193296e-07,
|
| 2662 |
+
"loss": 0.0001,
|
| 2663 |
+
"step": 21450
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 9.36,
|
| 2667 |
+
"learning_rate": 6.399651719634306e-07,
|
| 2668 |
+
"loss": 0.0006,
|
| 2669 |
+
"step": 21500
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 9.38,
|
| 2673 |
+
"learning_rate": 6.181976491075317e-07,
|
| 2674 |
+
"loss": 0.0017,
|
| 2675 |
+
"step": 21550
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 9.4,
|
| 2679 |
+
"learning_rate": 5.964301262516327e-07,
|
| 2680 |
+
"loss": 0.0,
|
| 2681 |
+
"step": 21600
|
| 2682 |
+
},
|
| 2683 |
+
{
|
| 2684 |
+
"epoch": 9.43,
|
| 2685 |
+
"learning_rate": 5.746626033957335e-07,
|
| 2686 |
+
"loss": 0.0,
|
| 2687 |
+
"step": 21650
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 9.45,
|
| 2691 |
+
"learning_rate": 5.528950805398346e-07,
|
| 2692 |
+
"loss": 0.0,
|
| 2693 |
+
"step": 21700
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 9.47,
|
| 2697 |
+
"learning_rate": 5.311275576839356e-07,
|
| 2698 |
+
"loss": 0.0006,
|
| 2699 |
+
"step": 21750
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 9.49,
|
| 2703 |
+
"learning_rate": 5.093600348280367e-07,
|
| 2704 |
+
"loss": 0.0,
|
| 2705 |
+
"step": 21800
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 9.51,
|
| 2709 |
+
"learning_rate": 4.875925119721376e-07,
|
| 2710 |
+
"loss": 0.0032,
|
| 2711 |
+
"step": 21850
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 9.53,
|
| 2715 |
+
"learning_rate": 4.658249891162386e-07,
|
| 2716 |
+
"loss": 0.0016,
|
| 2717 |
+
"step": 21900
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 9.56,
|
| 2721 |
+
"learning_rate": 4.440574662603396e-07,
|
| 2722 |
+
"loss": 0.0024,
|
| 2723 |
+
"step": 21950
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 9.58,
|
| 2727 |
+
"learning_rate": 4.2228994340444067e-07,
|
| 2728 |
+
"loss": 0.0001,
|
| 2729 |
+
"step": 22000
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 9.6,
|
| 2733 |
+
"learning_rate": 4.005224205485416e-07,
|
| 2734 |
+
"loss": 0.0001,
|
| 2735 |
+
"step": 22050
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 9.62,
|
| 2739 |
+
"learning_rate": 3.7875489769264256e-07,
|
| 2740 |
+
"loss": 0.002,
|
| 2741 |
+
"step": 22100
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 9.64,
|
| 2745 |
+
"learning_rate": 3.5698737483674364e-07,
|
| 2746 |
+
"loss": 0.0,
|
| 2747 |
+
"step": 22150
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 9.66,
|
| 2751 |
+
"learning_rate": 3.3521985198084456e-07,
|
| 2752 |
+
"loss": 0.0001,
|
| 2753 |
+
"step": 22200
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 9.69,
|
| 2757 |
+
"learning_rate": 3.1345232912494564e-07,
|
| 2758 |
+
"loss": 0.0038,
|
| 2759 |
+
"step": 22250
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 9.71,
|
| 2763 |
+
"learning_rate": 2.916848062690466e-07,
|
| 2764 |
+
"loss": 0.0027,
|
| 2765 |
+
"step": 22300
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"epoch": 9.73,
|
| 2769 |
+
"learning_rate": 2.699172834131476e-07,
|
| 2770 |
+
"loss": 0.003,
|
| 2771 |
+
"step": 22350
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 9.75,
|
| 2775 |
+
"learning_rate": 2.481497605572486e-07,
|
| 2776 |
+
"loss": 0.0001,
|
| 2777 |
+
"step": 22400
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 9.77,
|
| 2781 |
+
"learning_rate": 2.263822377013496e-07,
|
| 2782 |
+
"loss": 0.0001,
|
| 2783 |
+
"step": 22450
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 9.8,
|
| 2787 |
+
"learning_rate": 2.0461471484545061e-07,
|
| 2788 |
+
"loss": 0.0054,
|
| 2789 |
+
"step": 22500
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 9.82,
|
| 2793 |
+
"learning_rate": 1.8284719198955161e-07,
|
| 2794 |
+
"loss": 0.0015,
|
| 2795 |
+
"step": 22550
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 9.84,
|
| 2799 |
+
"learning_rate": 1.6107966913365261e-07,
|
| 2800 |
+
"loss": 0.0016,
|
| 2801 |
+
"step": 22600
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 9.86,
|
| 2805 |
+
"learning_rate": 1.393121462777536e-07,
|
| 2806 |
+
"loss": 0.0,
|
| 2807 |
+
"step": 22650
|
| 2808 |
+
},
|
| 2809 |
+
{
|
| 2810 |
+
"epoch": 9.88,
|
| 2811 |
+
"learning_rate": 1.175446234218546e-07,
|
| 2812 |
+
"loss": 0.0012,
|
| 2813 |
+
"step": 22700
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 9.9,
|
| 2817 |
+
"learning_rate": 9.57771005659556e-08,
|
| 2818 |
+
"loss": 0.0,
|
| 2819 |
+
"step": 22750
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 9.93,
|
| 2823 |
+
"learning_rate": 7.40095777100566e-08,
|
| 2824 |
+
"loss": 0.0001,
|
| 2825 |
+
"step": 22800
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 9.95,
|
| 2829 |
+
"learning_rate": 5.22420548541576e-08,
|
| 2830 |
+
"loss": 0.0053,
|
| 2831 |
+
"step": 22850
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 9.97,
|
| 2835 |
+
"learning_rate": 3.04745319982586e-08,
|
| 2836 |
+
"loss": 0.0015,
|
| 2837 |
+
"step": 22900
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 9.99,
|
| 2841 |
+
"learning_rate": 8.7070091423596e-09,
|
| 2842 |
+
"loss": 0.0001,
|
| 2843 |
+
"step": 22950
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 10.0,
|
| 2847 |
+
"eval_f1": 0.9991828997254544,
|
| 2848 |
+
"eval_loss": 0.008167621679604053,
|
| 2849 |
+
"eval_runtime": 35.0331,
|
| 2850 |
+
"eval_samples_per_second": 1048.665,
|
| 2851 |
+
"eval_steps_per_second": 16.413,
|
| 2852 |
+
"step": 22970
|
| 2853 |
+
}
|
| 2854 |
+
],
|
| 2855 |
+
"max_steps": 22970,
|
| 2856 |
+
"num_train_epochs": 10,
|
| 2857 |
+
"total_flos": 2.161637529772032e+16,
|
| 2858 |
+
"trial_name": null,
|
| 2859 |
+
"trial_params": null
|
| 2860 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:911ccdbee347d98cf2408ad6084caee38736c169a4835b8489b49000e4a76c9a
|
| 3 |
+
size 3899
|