🚀 Upload KoELECTRA emotion classification model
Browse files- checkpoint-16135/config.json +47 -0
- checkpoint-16135/model.safetensors +3 -0
- checkpoint-16135/optimizer.pt +3 -0
- checkpoint-16135/rng_state.pth +3 -0
- checkpoint-16135/scheduler.pt +3 -0
- checkpoint-16135/trainer_state.json +2347 -0
- checkpoint-16135/training_args.bin +3 -0
- checkpoint-22589/config.json +47 -0
- checkpoint-22589/model.safetensors +3 -0
- checkpoint-22589/optimizer.pt +3 -0
- checkpoint-22589/rng_state.pth +3 -0
- checkpoint-22589/scheduler.pt +3 -0
- checkpoint-22589/trainer_state.json +3270 -0
- checkpoint-22589/training_args.bin +3 -0
- config.json +47 -0
- model.safetensors +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +58 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
checkpoint-16135/config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ElectraForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"embedding_size": 768,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "\uae30\uc068",
|
| 13 |
+
"1": "\ub2f9\ud669",
|
| 14 |
+
"2": "\ubd84\ub178",
|
| 15 |
+
"3": "\ubd88\uc548",
|
| 16 |
+
"4": "\uc0c1\ucc98",
|
| 17 |
+
"5": "\uc2ac\ud514"
|
| 18 |
+
},
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 3072,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"\uae30\uc068": 0,
|
| 23 |
+
"\ub2f9\ud669": 1,
|
| 24 |
+
"\ubd84\ub178": 2,
|
| 25 |
+
"\ubd88\uc548": 3,
|
| 26 |
+
"\uc0c1\ucc98": 4,
|
| 27 |
+
"\uc2ac\ud514": 5
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-12,
|
| 30 |
+
"max_position_embeddings": 512,
|
| 31 |
+
"model_type": "electra",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"position_embedding_type": "absolute",
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"summary_activation": "gelu",
|
| 38 |
+
"summary_last_dropout": 0.1,
|
| 39 |
+
"summary_type": "first",
|
| 40 |
+
"summary_use_proj": true,
|
| 41 |
+
"tokenizer_class": "BertTokenizer",
|
| 42 |
+
"torch_dtype": "float32",
|
| 43 |
+
"transformers_version": "4.52.2",
|
| 44 |
+
"type_vocab_size": 2,
|
| 45 |
+
"use_cache": true,
|
| 46 |
+
"vocab_size": 54343
|
| 47 |
+
}
|
checkpoint-16135/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:149e0f950d5f1601cc510fbd6b354de2f1e128f9f787c2d32ed18a6ec9b5f743
|
| 3 |
+
size 511149672
|
checkpoint-16135/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3388c9d7394db970a92bd97bdf4ddc045eb9ca89edde4582ae3102e6b5abbd32
|
| 3 |
+
size 1022417532
|
checkpoint-16135/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8e73afc24b71ca38f1ca196f4e24ac8a6c21f769e4f9c669aeed936401370b7
|
| 3 |
+
size 14168
|
checkpoint-16135/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf55eaf975de0a96c6214e9e47886d4be4d2ee565319d1980a9a5f976ed4bdc3
|
| 3 |
+
size 1056
|
checkpoint-16135/trainer_state.json
ADDED
|
@@ -0,0 +1,2347 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 16135,
|
| 3 |
+
"best_metric": 0.7960859271865419,
|
| 4 |
+
"best_model_checkpoint": "/content/drive/MyDrive/\uac10\uc815\ubd84\ub958/data/emotion_model/checkpoint-16135",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 16135,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.015494267121165169,
|
| 14 |
+
"grad_norm": 1.9321871995925903,
|
| 15 |
+
"learning_rate": 4.5553145336225596e-07,
|
| 16 |
+
"loss": 1.7919,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.030988534242330338,
|
| 21 |
+
"grad_norm": 1.7712626457214355,
|
| 22 |
+
"learning_rate": 9.203594669972111e-07,
|
| 23 |
+
"loss": 1.7865,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04648280136349551,
|
| 28 |
+
"grad_norm": 1.9185744524002075,
|
| 29 |
+
"learning_rate": 1.385187480632166e-06,
|
| 30 |
+
"loss": 1.7864,
|
| 31 |
+
"step": 150
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.061977068484660676,
|
| 35 |
+
"grad_norm": 1.9046003818511963,
|
| 36 |
+
"learning_rate": 1.8500154942671213e-06,
|
| 37 |
+
"loss": 1.7864,
|
| 38 |
+
"step": 200
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07747133560582585,
|
| 42 |
+
"grad_norm": 2.0914034843444824,
|
| 43 |
+
"learning_rate": 2.3148435079020763e-06,
|
| 44 |
+
"loss": 1.7813,
|
| 45 |
+
"step": 250
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.09296560272699102,
|
| 49 |
+
"grad_norm": 2.088219165802002,
|
| 50 |
+
"learning_rate": 2.7796715215370313e-06,
|
| 51 |
+
"loss": 1.7773,
|
| 52 |
+
"step": 300
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.10845986984815618,
|
| 56 |
+
"grad_norm": 2.1377577781677246,
|
| 57 |
+
"learning_rate": 3.2444995351719864e-06,
|
| 58 |
+
"loss": 1.7587,
|
| 59 |
+
"step": 350
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.12395413696932135,
|
| 63 |
+
"grad_norm": 2.2140750885009766,
|
| 64 |
+
"learning_rate": 3.7093275488069414e-06,
|
| 65 |
+
"loss": 1.7388,
|
| 66 |
+
"step": 400
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.13944840409048653,
|
| 70 |
+
"grad_norm": 2.1278295516967773,
|
| 71 |
+
"learning_rate": 4.174155562441896e-06,
|
| 72 |
+
"loss": 1.6861,
|
| 73 |
+
"step": 450
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1549426712116517,
|
| 77 |
+
"grad_norm": 4.734658241271973,
|
| 78 |
+
"learning_rate": 4.638983576076852e-06,
|
| 79 |
+
"loss": 1.6267,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.17043693833281687,
|
| 84 |
+
"grad_norm": 4.140384197235107,
|
| 85 |
+
"learning_rate": 5.103811589711806e-06,
|
| 86 |
+
"loss": 1.5732,
|
| 87 |
+
"step": 550
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.18593120545398203,
|
| 91 |
+
"grad_norm": 2.705493688583374,
|
| 92 |
+
"learning_rate": 5.568639603346762e-06,
|
| 93 |
+
"loss": 1.5438,
|
| 94 |
+
"step": 600
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2014254725751472,
|
| 98 |
+
"grad_norm": 5.3791422843933105,
|
| 99 |
+
"learning_rate": 6.0334676169817164e-06,
|
| 100 |
+
"loss": 1.4644,
|
| 101 |
+
"step": 650
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.21691973969631237,
|
| 105 |
+
"grad_norm": 2.8989250659942627,
|
| 106 |
+
"learning_rate": 6.498295630616672e-06,
|
| 107 |
+
"loss": 1.4743,
|
| 108 |
+
"step": 700
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.23241400681747754,
|
| 112 |
+
"grad_norm": 3.40291166305542,
|
| 113 |
+
"learning_rate": 6.963123644251627e-06,
|
| 114 |
+
"loss": 1.4187,
|
| 115 |
+
"step": 750
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.2479082739386427,
|
| 119 |
+
"grad_norm": 5.748068809509277,
|
| 120 |
+
"learning_rate": 7.427951657886583e-06,
|
| 121 |
+
"loss": 1.3533,
|
| 122 |
+
"step": 800
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.26340254105980787,
|
| 126 |
+
"grad_norm": 5.777422904968262,
|
| 127 |
+
"learning_rate": 7.892779671521537e-06,
|
| 128 |
+
"loss": 1.3099,
|
| 129 |
+
"step": 850
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.27889680818097307,
|
| 133 |
+
"grad_norm": 4.8046112060546875,
|
| 134 |
+
"learning_rate": 8.357607685156493e-06,
|
| 135 |
+
"loss": 1.2723,
|
| 136 |
+
"step": 900
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.2943910753021382,
|
| 140 |
+
"grad_norm": 5.549858570098877,
|
| 141 |
+
"learning_rate": 8.822435698791447e-06,
|
| 142 |
+
"loss": 1.2325,
|
| 143 |
+
"step": 950
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3098853424233034,
|
| 147 |
+
"grad_norm": 6.851742744445801,
|
| 148 |
+
"learning_rate": 9.287263712426402e-06,
|
| 149 |
+
"loss": 1.1884,
|
| 150 |
+
"step": 1000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.32537960954446854,
|
| 154 |
+
"grad_norm": 5.2497735023498535,
|
| 155 |
+
"learning_rate": 9.752091726061357e-06,
|
| 156 |
+
"loss": 1.2224,
|
| 157 |
+
"step": 1050
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.34087387666563373,
|
| 161 |
+
"grad_norm": 7.023674488067627,
|
| 162 |
+
"learning_rate": 1.0216919739696313e-05,
|
| 163 |
+
"loss": 1.177,
|
| 164 |
+
"step": 1100
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.3563681437867989,
|
| 168 |
+
"grad_norm": 4.888996124267578,
|
| 169 |
+
"learning_rate": 1.0681747753331269e-05,
|
| 170 |
+
"loss": 1.1577,
|
| 171 |
+
"step": 1150
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.37186241090796407,
|
| 175 |
+
"grad_norm": 6.2133660316467285,
|
| 176 |
+
"learning_rate": 1.1146575766966222e-05,
|
| 177 |
+
"loss": 1.1726,
|
| 178 |
+
"step": 1200
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.3873566780291292,
|
| 182 |
+
"grad_norm": 6.936697483062744,
|
| 183 |
+
"learning_rate": 1.1611403780601178e-05,
|
| 184 |
+
"loss": 1.1001,
|
| 185 |
+
"step": 1250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.4028509451502944,
|
| 189 |
+
"grad_norm": 8.526293754577637,
|
| 190 |
+
"learning_rate": 1.2076231794236133e-05,
|
| 191 |
+
"loss": 1.0752,
|
| 192 |
+
"step": 1300
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.41834521227145954,
|
| 196 |
+
"grad_norm": 5.5933756828308105,
|
| 197 |
+
"learning_rate": 1.254105980787109e-05,
|
| 198 |
+
"loss": 1.0809,
|
| 199 |
+
"step": 1350
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.43383947939262474,
|
| 203 |
+
"grad_norm": 6.998812675476074,
|
| 204 |
+
"learning_rate": 1.3005887821506042e-05,
|
| 205 |
+
"loss": 1.0566,
|
| 206 |
+
"step": 1400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.4493337465137899,
|
| 210 |
+
"grad_norm": 7.077617645263672,
|
| 211 |
+
"learning_rate": 1.3470715835140998e-05,
|
| 212 |
+
"loss": 1.0993,
|
| 213 |
+
"step": 1450
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.4648280136349551,
|
| 217 |
+
"grad_norm": 8.715201377868652,
|
| 218 |
+
"learning_rate": 1.3935543848775953e-05,
|
| 219 |
+
"loss": 1.0375,
|
| 220 |
+
"step": 1500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4803222807561202,
|
| 224 |
+
"grad_norm": 6.017217636108398,
|
| 225 |
+
"learning_rate": 1.440037186241091e-05,
|
| 226 |
+
"loss": 1.0388,
|
| 227 |
+
"step": 1550
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.4958165478772854,
|
| 231 |
+
"grad_norm": 5.349973201751709,
|
| 232 |
+
"learning_rate": 1.4865199876045862e-05,
|
| 233 |
+
"loss": 1.0354,
|
| 234 |
+
"step": 1600
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5113108149984505,
|
| 238 |
+
"grad_norm": 12.728338241577148,
|
| 239 |
+
"learning_rate": 1.533002788968082e-05,
|
| 240 |
+
"loss": 1.0314,
|
| 241 |
+
"step": 1650
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.5268050821196157,
|
| 245 |
+
"grad_norm": 5.962468147277832,
|
| 246 |
+
"learning_rate": 1.5794855903315773e-05,
|
| 247 |
+
"loss": 1.03,
|
| 248 |
+
"step": 1700
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.5422993492407809,
|
| 252 |
+
"grad_norm": 5.971400260925293,
|
| 253 |
+
"learning_rate": 1.6259683916950726e-05,
|
| 254 |
+
"loss": 1.0541,
|
| 255 |
+
"step": 1750
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.5577936163619461,
|
| 259 |
+
"grad_norm": 6.260463714599609,
|
| 260 |
+
"learning_rate": 1.6724511930585682e-05,
|
| 261 |
+
"loss": 0.9831,
|
| 262 |
+
"step": 1800
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.5732878834831112,
|
| 266 |
+
"grad_norm": 7.8115010261535645,
|
| 267 |
+
"learning_rate": 1.718933994422064e-05,
|
| 268 |
+
"loss": 0.966,
|
| 269 |
+
"step": 1850
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.5887821506042764,
|
| 273 |
+
"grad_norm": 5.005403995513916,
|
| 274 |
+
"learning_rate": 1.7654167957855595e-05,
|
| 275 |
+
"loss": 0.9592,
|
| 276 |
+
"step": 1900
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.6042764177254416,
|
| 280 |
+
"grad_norm": 7.7732157707214355,
|
| 281 |
+
"learning_rate": 1.8118995971490548e-05,
|
| 282 |
+
"loss": 0.9766,
|
| 283 |
+
"step": 1950
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.6197706848466068,
|
| 287 |
+
"grad_norm": 7.265392303466797,
|
| 288 |
+
"learning_rate": 1.8583823985125504e-05,
|
| 289 |
+
"loss": 1.0171,
|
| 290 |
+
"step": 2000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.6352649519677719,
|
| 294 |
+
"grad_norm": 15.946109771728516,
|
| 295 |
+
"learning_rate": 1.904865199876046e-05,
|
| 296 |
+
"loss": 0.9824,
|
| 297 |
+
"step": 2050
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.6507592190889371,
|
| 301 |
+
"grad_norm": 7.261445999145508,
|
| 302 |
+
"learning_rate": 1.9513480012395417e-05,
|
| 303 |
+
"loss": 0.9699,
|
| 304 |
+
"step": 2100
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.6662534862101023,
|
| 308 |
+
"grad_norm": 8.201744079589844,
|
| 309 |
+
"learning_rate": 1.997830802603037e-05,
|
| 310 |
+
"loss": 0.9957,
|
| 311 |
+
"step": 2150
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.6817477533312675,
|
| 315 |
+
"grad_norm": 6.183067798614502,
|
| 316 |
+
"learning_rate": 2.0443136039665322e-05,
|
| 317 |
+
"loss": 0.8554,
|
| 318 |
+
"step": 2200
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.6972420204524326,
|
| 322 |
+
"grad_norm": 7.481590270996094,
|
| 323 |
+
"learning_rate": 2.090796405330028e-05,
|
| 324 |
+
"loss": 0.9929,
|
| 325 |
+
"step": 2250
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.7127362875735977,
|
| 329 |
+
"grad_norm": 7.3274030685424805,
|
| 330 |
+
"learning_rate": 2.1372792066935235e-05,
|
| 331 |
+
"loss": 0.9438,
|
| 332 |
+
"step": 2300
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.7282305546947629,
|
| 336 |
+
"grad_norm": 11.69247055053711,
|
| 337 |
+
"learning_rate": 2.183762008057019e-05,
|
| 338 |
+
"loss": 0.9767,
|
| 339 |
+
"step": 2350
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.7437248218159281,
|
| 343 |
+
"grad_norm": 7.929721832275391,
|
| 344 |
+
"learning_rate": 2.2302448094205144e-05,
|
| 345 |
+
"loss": 1.0036,
|
| 346 |
+
"step": 2400
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.7592190889370932,
|
| 350 |
+
"grad_norm": 9.753717422485352,
|
| 351 |
+
"learning_rate": 2.27672761078401e-05,
|
| 352 |
+
"loss": 0.9726,
|
| 353 |
+
"step": 2450
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.7747133560582584,
|
| 357 |
+
"grad_norm": 7.797086715698242,
|
| 358 |
+
"learning_rate": 2.3232104121475057e-05,
|
| 359 |
+
"loss": 0.9322,
|
| 360 |
+
"step": 2500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.7902076231794236,
|
| 364 |
+
"grad_norm": 6.927332878112793,
|
| 365 |
+
"learning_rate": 2.369693213511001e-05,
|
| 366 |
+
"loss": 0.9378,
|
| 367 |
+
"step": 2550
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.8057018903005888,
|
| 371 |
+
"grad_norm": 3.726092576980591,
|
| 372 |
+
"learning_rate": 2.4161760148744962e-05,
|
| 373 |
+
"loss": 0.958,
|
| 374 |
+
"step": 2600
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.821196157421754,
|
| 378 |
+
"grad_norm": 5.661774635314941,
|
| 379 |
+
"learning_rate": 2.462658816237992e-05,
|
| 380 |
+
"loss": 0.9651,
|
| 381 |
+
"step": 2650
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.8366904245429191,
|
| 385 |
+
"grad_norm": 6.513345718383789,
|
| 386 |
+
"learning_rate": 2.5091416176014875e-05,
|
| 387 |
+
"loss": 0.971,
|
| 388 |
+
"step": 2700
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.8521846916640843,
|
| 392 |
+
"grad_norm": 6.713255405426025,
|
| 393 |
+
"learning_rate": 2.555624418964983e-05,
|
| 394 |
+
"loss": 0.8616,
|
| 395 |
+
"step": 2750
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.8676789587852495,
|
| 399 |
+
"grad_norm": 8.527266502380371,
|
| 400 |
+
"learning_rate": 2.6021072203284784e-05,
|
| 401 |
+
"loss": 0.9413,
|
| 402 |
+
"step": 2800
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.8831732259064147,
|
| 406 |
+
"grad_norm": 6.599502086639404,
|
| 407 |
+
"learning_rate": 2.648590021691974e-05,
|
| 408 |
+
"loss": 0.9985,
|
| 409 |
+
"step": 2850
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.8986674930275798,
|
| 413 |
+
"grad_norm": 4.0680155754089355,
|
| 414 |
+
"learning_rate": 2.6950728230554697e-05,
|
| 415 |
+
"loss": 0.9415,
|
| 416 |
+
"step": 2900
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.914161760148745,
|
| 420 |
+
"grad_norm": 5.083493232727051,
|
| 421 |
+
"learning_rate": 2.741555624418965e-05,
|
| 422 |
+
"loss": 0.9805,
|
| 423 |
+
"step": 2950
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.9296560272699101,
|
| 427 |
+
"grad_norm": 4.0469069480896,
|
| 428 |
+
"learning_rate": 2.7880384257824606e-05,
|
| 429 |
+
"loss": 0.9547,
|
| 430 |
+
"step": 3000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.9451502943910753,
|
| 434 |
+
"grad_norm": 6.075752258300781,
|
| 435 |
+
"learning_rate": 2.834521227145956e-05,
|
| 436 |
+
"loss": 0.9623,
|
| 437 |
+
"step": 3050
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.9606445615122404,
|
| 441 |
+
"grad_norm": 6.5252299308776855,
|
| 442 |
+
"learning_rate": 2.8810040285094515e-05,
|
| 443 |
+
"loss": 0.9838,
|
| 444 |
+
"step": 3100
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.9761388286334056,
|
| 448 |
+
"grad_norm": 6.530562877655029,
|
| 449 |
+
"learning_rate": 2.927486829872947e-05,
|
| 450 |
+
"loss": 0.911,
|
| 451 |
+
"step": 3150
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.9916330957545708,
|
| 455 |
+
"grad_norm": 8.217161178588867,
|
| 456 |
+
"learning_rate": 2.9739696312364428e-05,
|
| 457 |
+
"loss": 0.9457,
|
| 458 |
+
"step": 3200
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.0,
|
| 462 |
+
"eval_accuracy": 0.7087787983737389,
|
| 463 |
+
"eval_f1": 0.7069046924545164,
|
| 464 |
+
"eval_loss": 0.8573769330978394,
|
| 465 |
+
"eval_runtime": 25.5149,
|
| 466 |
+
"eval_samples_per_second": 260.279,
|
| 467 |
+
"eval_steps_per_second": 16.304,
|
| 468 |
+
"step": 3227
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.007127362875736,
|
| 472 |
+
"grad_norm": 4.367455005645752,
|
| 473 |
+
"learning_rate": 2.997727507488896e-05,
|
| 474 |
+
"loss": 0.9143,
|
| 475 |
+
"step": 3250
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.022621629996901,
|
| 479 |
+
"grad_norm": 4.636725902557373,
|
| 480 |
+
"learning_rate": 2.9925627517818407e-05,
|
| 481 |
+
"loss": 0.9304,
|
| 482 |
+
"step": 3300
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.0381158971180664,
|
| 486 |
+
"grad_norm": 4.251437664031982,
|
| 487 |
+
"learning_rate": 2.9873979960747857e-05,
|
| 488 |
+
"loss": 0.8549,
|
| 489 |
+
"step": 3350
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.0536101642392315,
|
| 493 |
+
"grad_norm": 6.648655414581299,
|
| 494 |
+
"learning_rate": 2.9822332403677307e-05,
|
| 495 |
+
"loss": 0.8698,
|
| 496 |
+
"step": 3400
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.0691044313603966,
|
| 500 |
+
"grad_norm": 7.102205276489258,
|
| 501 |
+
"learning_rate": 2.9770684846606757e-05,
|
| 502 |
+
"loss": 0.8597,
|
| 503 |
+
"step": 3450
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.0845986984815619,
|
| 507 |
+
"grad_norm": 10.821270942687988,
|
| 508 |
+
"learning_rate": 2.9719037289536206e-05,
|
| 509 |
+
"loss": 0.8457,
|
| 510 |
+
"step": 3500
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.100092965602727,
|
| 514 |
+
"grad_norm": 6.111588001251221,
|
| 515 |
+
"learning_rate": 2.9667389732465652e-05,
|
| 516 |
+
"loss": 0.8973,
|
| 517 |
+
"step": 3550
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.1155872327238923,
|
| 521 |
+
"grad_norm": 9.016953468322754,
|
| 522 |
+
"learning_rate": 2.9615742175395106e-05,
|
| 523 |
+
"loss": 0.8228,
|
| 524 |
+
"step": 3600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.1310814998450573,
|
| 528 |
+
"grad_norm": 7.717069625854492,
|
| 529 |
+
"learning_rate": 2.9564094618324555e-05,
|
| 530 |
+
"loss": 0.8103,
|
| 531 |
+
"step": 3650
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.1465757669662224,
|
| 535 |
+
"grad_norm": 7.848579406738281,
|
| 536 |
+
"learning_rate": 2.9512447061254005e-05,
|
| 537 |
+
"loss": 0.8097,
|
| 538 |
+
"step": 3700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.1620700340873877,
|
| 542 |
+
"grad_norm": 4.738124847412109,
|
| 543 |
+
"learning_rate": 2.9460799504183455e-05,
|
| 544 |
+
"loss": 0.9105,
|
| 545 |
+
"step": 3750
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.1775643012085528,
|
| 549 |
+
"grad_norm": 5.289875507354736,
|
| 550 |
+
"learning_rate": 2.94091519471129e-05,
|
| 551 |
+
"loss": 0.8876,
|
| 552 |
+
"step": 3800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1930585683297181,
|
| 556 |
+
"grad_norm": 6.445308685302734,
|
| 557 |
+
"learning_rate": 2.935750439004235e-05,
|
| 558 |
+
"loss": 0.8377,
|
| 559 |
+
"step": 3850
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.2085528354508832,
|
| 563 |
+
"grad_norm": 4.725327968597412,
|
| 564 |
+
"learning_rate": 2.93058568329718e-05,
|
| 565 |
+
"loss": 0.9332,
|
| 566 |
+
"step": 3900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.2240471025720483,
|
| 570 |
+
"grad_norm": 53.85081481933594,
|
| 571 |
+
"learning_rate": 2.925420927590125e-05,
|
| 572 |
+
"loss": 0.8882,
|
| 573 |
+
"step": 3950
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.2395413696932136,
|
| 577 |
+
"grad_norm": 5.677978515625,
|
| 578 |
+
"learning_rate": 2.9202561718830703e-05,
|
| 579 |
+
"loss": 0.8481,
|
| 580 |
+
"step": 4000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.2550356368143787,
|
| 584 |
+
"grad_norm": 3.941765785217285,
|
| 585 |
+
"learning_rate": 2.915091416176015e-05,
|
| 586 |
+
"loss": 0.835,
|
| 587 |
+
"step": 4050
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.2705299039355438,
|
| 591 |
+
"grad_norm": 8.099725723266602,
|
| 592 |
+
"learning_rate": 2.90992666046896e-05,
|
| 593 |
+
"loss": 0.8322,
|
| 594 |
+
"step": 4100
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.286024171056709,
|
| 598 |
+
"grad_norm": 6.59591007232666,
|
| 599 |
+
"learning_rate": 2.904761904761905e-05,
|
| 600 |
+
"loss": 0.8809,
|
| 601 |
+
"step": 4150
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.3015184381778742,
|
| 605 |
+
"grad_norm": 7.200226306915283,
|
| 606 |
+
"learning_rate": 2.8995971490548498e-05,
|
| 607 |
+
"loss": 0.8609,
|
| 608 |
+
"step": 4200
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.3170127052990392,
|
| 612 |
+
"grad_norm": 4.902937412261963,
|
| 613 |
+
"learning_rate": 2.8944323933477948e-05,
|
| 614 |
+
"loss": 0.8633,
|
| 615 |
+
"step": 4250
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.3325069724202045,
|
| 619 |
+
"grad_norm": 5.792146682739258,
|
| 620 |
+
"learning_rate": 2.8892676376407394e-05,
|
| 621 |
+
"loss": 0.8684,
|
| 622 |
+
"step": 4300
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.3480012395413696,
|
| 626 |
+
"grad_norm": 4.636809349060059,
|
| 627 |
+
"learning_rate": 2.8841028819336844e-05,
|
| 628 |
+
"loss": 0.8245,
|
| 629 |
+
"step": 4350
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.363495506662535,
|
| 633 |
+
"grad_norm": 5.28842306137085,
|
| 634 |
+
"learning_rate": 2.8789381262266297e-05,
|
| 635 |
+
"loss": 0.7859,
|
| 636 |
+
"step": 4400
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.3789897737837,
|
| 640 |
+
"grad_norm": 4.259128570556641,
|
| 641 |
+
"learning_rate": 2.8737733705195747e-05,
|
| 642 |
+
"loss": 0.8228,
|
| 643 |
+
"step": 4450
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.394484040904865,
|
| 647 |
+
"grad_norm": 7.914375305175781,
|
| 648 |
+
"learning_rate": 2.8686086148125196e-05,
|
| 649 |
+
"loss": 0.9448,
|
| 650 |
+
"step": 4500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.4099783080260304,
|
| 654 |
+
"grad_norm": 7.636547088623047,
|
| 655 |
+
"learning_rate": 2.8634438591054643e-05,
|
| 656 |
+
"loss": 0.8781,
|
| 657 |
+
"step": 4550
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.4254725751471955,
|
| 661 |
+
"grad_norm": 8.681707382202148,
|
| 662 |
+
"learning_rate": 2.8582791033984092e-05,
|
| 663 |
+
"loss": 0.8367,
|
| 664 |
+
"step": 4600
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.4409668422683608,
|
| 668 |
+
"grad_norm": 7.864759922027588,
|
| 669 |
+
"learning_rate": 2.8531143476913542e-05,
|
| 670 |
+
"loss": 0.9088,
|
| 671 |
+
"step": 4650
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.4564611093895259,
|
| 675 |
+
"grad_norm": 4.892348289489746,
|
| 676 |
+
"learning_rate": 2.8479495919842992e-05,
|
| 677 |
+
"loss": 0.8993,
|
| 678 |
+
"step": 4700
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.471955376510691,
|
| 682 |
+
"grad_norm": 23.208873748779297,
|
| 683 |
+
"learning_rate": 2.842784836277244e-05,
|
| 684 |
+
"loss": 0.8542,
|
| 685 |
+
"step": 4750
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.4874496436318563,
|
| 689 |
+
"grad_norm": 8.983469009399414,
|
| 690 |
+
"learning_rate": 2.837620080570189e-05,
|
| 691 |
+
"loss": 0.949,
|
| 692 |
+
"step": 4800
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.5029439107530214,
|
| 696 |
+
"grad_norm": 10.706644058227539,
|
| 697 |
+
"learning_rate": 2.832455324863134e-05,
|
| 698 |
+
"loss": 0.9008,
|
| 699 |
+
"step": 4850
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.5184381778741867,
|
| 703 |
+
"grad_norm": 4.685935020446777,
|
| 704 |
+
"learning_rate": 2.827290569156079e-05,
|
| 705 |
+
"loss": 0.8613,
|
| 706 |
+
"step": 4900
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.5339324449953518,
|
| 710 |
+
"grad_norm": 5.286406993865967,
|
| 711 |
+
"learning_rate": 2.822125813449024e-05,
|
| 712 |
+
"loss": 0.8929,
|
| 713 |
+
"step": 4950
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.5494267121165168,
|
| 717 |
+
"grad_norm": 4.907707691192627,
|
| 718 |
+
"learning_rate": 2.816961057741969e-05,
|
| 719 |
+
"loss": 0.8321,
|
| 720 |
+
"step": 5000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.564920979237682,
|
| 724 |
+
"grad_norm": 6.398087501525879,
|
| 725 |
+
"learning_rate": 2.8117963020349136e-05,
|
| 726 |
+
"loss": 0.8626,
|
| 727 |
+
"step": 5050
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.5804152463588472,
|
| 731 |
+
"grad_norm": 5.323617458343506,
|
| 732 |
+
"learning_rate": 2.8066315463278586e-05,
|
| 733 |
+
"loss": 0.8324,
|
| 734 |
+
"step": 5100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.5959095134800125,
|
| 738 |
+
"grad_norm": 4.136271953582764,
|
| 739 |
+
"learning_rate": 2.8014667906208035e-05,
|
| 740 |
+
"loss": 0.879,
|
| 741 |
+
"step": 5150
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.6114037806011776,
|
| 745 |
+
"grad_norm": 6.873619556427002,
|
| 746 |
+
"learning_rate": 2.796302034913749e-05,
|
| 747 |
+
"loss": 0.9043,
|
| 748 |
+
"step": 5200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.6268980477223427,
|
| 752 |
+
"grad_norm": 7.138693809509277,
|
| 753 |
+
"learning_rate": 2.7911372792066938e-05,
|
| 754 |
+
"loss": 0.8183,
|
| 755 |
+
"step": 5250
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.6423923148435078,
|
| 759 |
+
"grad_norm": 6.483767032623291,
|
| 760 |
+
"learning_rate": 2.7859725234996384e-05,
|
| 761 |
+
"loss": 0.8867,
|
| 762 |
+
"step": 5300
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.657886581964673,
|
| 766 |
+
"grad_norm": 3.2249104976654053,
|
| 767 |
+
"learning_rate": 2.7808077677925834e-05,
|
| 768 |
+
"loss": 0.8097,
|
| 769 |
+
"step": 5350
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.6733808490858384,
|
| 773 |
+
"grad_norm": 6.961575984954834,
|
| 774 |
+
"learning_rate": 2.7756430120855284e-05,
|
| 775 |
+
"loss": 0.8364,
|
| 776 |
+
"step": 5400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.6888751162070035,
|
| 780 |
+
"grad_norm": 7.0920000076293945,
|
| 781 |
+
"learning_rate": 2.7704782563784733e-05,
|
| 782 |
+
"loss": 0.8283,
|
| 783 |
+
"step": 5450
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.7043693833281686,
|
| 787 |
+
"grad_norm": 5.436604976654053,
|
| 788 |
+
"learning_rate": 2.7653135006714183e-05,
|
| 789 |
+
"loss": 0.8786,
|
| 790 |
+
"step": 5500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.7198636504493336,
|
| 794 |
+
"grad_norm": 4.0141282081604,
|
| 795 |
+
"learning_rate": 2.760148744964363e-05,
|
| 796 |
+
"loss": 0.8452,
|
| 797 |
+
"step": 5550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.735357917570499,
|
| 801 |
+
"grad_norm": 5.783074378967285,
|
| 802 |
+
"learning_rate": 2.7549839892573083e-05,
|
| 803 |
+
"loss": 0.8168,
|
| 804 |
+
"step": 5600
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 1.750852184691664,
|
| 808 |
+
"grad_norm": 7.773756504058838,
|
| 809 |
+
"learning_rate": 2.7498192335502532e-05,
|
| 810 |
+
"loss": 0.8598,
|
| 811 |
+
"step": 5650
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.7663464518128293,
|
| 815 |
+
"grad_norm": 5.375339984893799,
|
| 816 |
+
"learning_rate": 2.7446544778431982e-05,
|
| 817 |
+
"loss": 0.8366,
|
| 818 |
+
"step": 5700
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.7818407189339944,
|
| 822 |
+
"grad_norm": 4.240859031677246,
|
| 823 |
+
"learning_rate": 2.739489722136143e-05,
|
| 824 |
+
"loss": 0.8136,
|
| 825 |
+
"step": 5750
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 1.7973349860551595,
|
| 829 |
+
"grad_norm": 6.107599258422852,
|
| 830 |
+
"learning_rate": 2.734324966429088e-05,
|
| 831 |
+
"loss": 0.8074,
|
| 832 |
+
"step": 5800
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.8128292531763246,
|
| 836 |
+
"grad_norm": 6.027589797973633,
|
| 837 |
+
"learning_rate": 2.7291602107220328e-05,
|
| 838 |
+
"loss": 0.7808,
|
| 839 |
+
"step": 5850
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.82832352029749,
|
| 843 |
+
"grad_norm": 4.829204559326172,
|
| 844 |
+
"learning_rate": 2.7239954550149777e-05,
|
| 845 |
+
"loss": 0.8473,
|
| 846 |
+
"step": 5900
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.8438177874186552,
|
| 850 |
+
"grad_norm": 5.385358810424805,
|
| 851 |
+
"learning_rate": 2.7188306993079227e-05,
|
| 852 |
+
"loss": 0.844,
|
| 853 |
+
"step": 5950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.8593120545398203,
|
| 857 |
+
"grad_norm": 5.991063594818115,
|
| 858 |
+
"learning_rate": 2.713665943600868e-05,
|
| 859 |
+
"loss": 0.8667,
|
| 860 |
+
"step": 6000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.8748063216609854,
|
| 864 |
+
"grad_norm": 4.269604682922363,
|
| 865 |
+
"learning_rate": 2.708501187893813e-05,
|
| 866 |
+
"loss": 0.8987,
|
| 867 |
+
"step": 6050
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.8903005887821505,
|
| 871 |
+
"grad_norm": 6.90878438949585,
|
| 872 |
+
"learning_rate": 2.7033364321867576e-05,
|
| 873 |
+
"loss": 0.8517,
|
| 874 |
+
"step": 6100
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.9057948559033158,
|
| 878 |
+
"grad_norm": 8.742233276367188,
|
| 879 |
+
"learning_rate": 2.6981716764797026e-05,
|
| 880 |
+
"loss": 0.8729,
|
| 881 |
+
"step": 6150
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.921289123024481,
|
| 885 |
+
"grad_norm": 9.10084342956543,
|
| 886 |
+
"learning_rate": 2.6930069207726475e-05,
|
| 887 |
+
"loss": 0.8803,
|
| 888 |
+
"step": 6200
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 1.9367833901456462,
|
| 892 |
+
"grad_norm": 4.210537433624268,
|
| 893 |
+
"learning_rate": 2.6878421650655925e-05,
|
| 894 |
+
"loss": 0.7938,
|
| 895 |
+
"step": 6250
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.9522776572668112,
|
| 899 |
+
"grad_norm": 6.604791641235352,
|
| 900 |
+
"learning_rate": 2.6826774093585375e-05,
|
| 901 |
+
"loss": 0.7958,
|
| 902 |
+
"step": 6300
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.9677719243879763,
|
| 906 |
+
"grad_norm": 6.213857173919678,
|
| 907 |
+
"learning_rate": 2.677512653651482e-05,
|
| 908 |
+
"loss": 0.8463,
|
| 909 |
+
"step": 6350
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.9832661915091416,
|
| 913 |
+
"grad_norm": 4.303800582885742,
|
| 914 |
+
"learning_rate": 2.6723478979444274e-05,
|
| 915 |
+
"loss": 0.7909,
|
| 916 |
+
"step": 6400
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.998760458630307,
|
| 920 |
+
"grad_norm": 4.933095932006836,
|
| 921 |
+
"learning_rate": 2.6671831422373724e-05,
|
| 922 |
+
"loss": 0.7888,
|
| 923 |
+
"step": 6450
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 2.0,
|
| 927 |
+
"eval_accuracy": 0.7590724288510766,
|
| 928 |
+
"eval_f1": 0.7587245577707713,
|
| 929 |
+
"eval_loss": 0.7010347247123718,
|
| 930 |
+
"eval_runtime": 25.4199,
|
| 931 |
+
"eval_samples_per_second": 261.252,
|
| 932 |
+
"eval_steps_per_second": 16.365,
|
| 933 |
+
"step": 6454
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.014254725751472,
|
| 937 |
+
"grad_norm": 4.0570244789123535,
|
| 938 |
+
"learning_rate": 2.6620183865303173e-05,
|
| 939 |
+
"loss": 0.7236,
|
| 940 |
+
"step": 6500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.029748992872637,
|
| 944 |
+
"grad_norm": 5.307652473449707,
|
| 945 |
+
"learning_rate": 2.6568536308232623e-05,
|
| 946 |
+
"loss": 0.7213,
|
| 947 |
+
"step": 6550
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.045243259993802,
|
| 951 |
+
"grad_norm": 5.398072719573975,
|
| 952 |
+
"learning_rate": 2.651688875116207e-05,
|
| 953 |
+
"loss": 0.6839,
|
| 954 |
+
"step": 6600
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.0607375271149673,
|
| 958 |
+
"grad_norm": 5.296418190002441,
|
| 959 |
+
"learning_rate": 2.646524119409152e-05,
|
| 960 |
+
"loss": 0.6856,
|
| 961 |
+
"step": 6650
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 2.076231794236133,
|
| 965 |
+
"grad_norm": 4.173377990722656,
|
| 966 |
+
"learning_rate": 2.641359363702097e-05,
|
| 967 |
+
"loss": 0.7109,
|
| 968 |
+
"step": 6700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.091726061357298,
|
| 972 |
+
"grad_norm": 5.590676784515381,
|
| 973 |
+
"learning_rate": 2.636194607995042e-05,
|
| 974 |
+
"loss": 0.6814,
|
| 975 |
+
"step": 6750
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.107220328478463,
|
| 979 |
+
"grad_norm": 8.112780570983887,
|
| 980 |
+
"learning_rate": 2.631029852287987e-05,
|
| 981 |
+
"loss": 0.7302,
|
| 982 |
+
"step": 6800
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 2.122714595599628,
|
| 986 |
+
"grad_norm": 6.514364242553711,
|
| 987 |
+
"learning_rate": 2.6258650965809318e-05,
|
| 988 |
+
"loss": 0.6917,
|
| 989 |
+
"step": 6850
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.138208862720793,
|
| 993 |
+
"grad_norm": 8.156841278076172,
|
| 994 |
+
"learning_rate": 2.6207003408738767e-05,
|
| 995 |
+
"loss": 0.6568,
|
| 996 |
+
"step": 6900
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 2.1537031298419587,
|
| 1000 |
+
"grad_norm": 7.641481876373291,
|
| 1001 |
+
"learning_rate": 2.6155355851668217e-05,
|
| 1002 |
+
"loss": 0.6132,
|
| 1003 |
+
"step": 6950
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 2.1691973969631237,
|
| 1007 |
+
"grad_norm": 6.33613395690918,
|
| 1008 |
+
"learning_rate": 2.6103708294597667e-05,
|
| 1009 |
+
"loss": 0.6393,
|
| 1010 |
+
"step": 7000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.184691664084289,
|
| 1014 |
+
"grad_norm": 4.2916436195373535,
|
| 1015 |
+
"learning_rate": 2.6052060737527116e-05,
|
| 1016 |
+
"loss": 0.6709,
|
| 1017 |
+
"step": 7050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.200185931205454,
|
| 1021 |
+
"grad_norm": 4.763488292694092,
|
| 1022 |
+
"learning_rate": 2.6000413180456563e-05,
|
| 1023 |
+
"loss": 0.6919,
|
| 1024 |
+
"step": 7100
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 2.215680198326619,
|
| 1028 |
+
"grad_norm": 8.614394187927246,
|
| 1029 |
+
"learning_rate": 2.5948765623386012e-05,
|
| 1030 |
+
"loss": 0.6501,
|
| 1031 |
+
"step": 7150
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 2.2311744654477845,
|
| 1035 |
+
"grad_norm": 9.684426307678223,
|
| 1036 |
+
"learning_rate": 2.5897118066315465e-05,
|
| 1037 |
+
"loss": 0.6947,
|
| 1038 |
+
"step": 7200
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 2.2466687325689496,
|
| 1042 |
+
"grad_norm": 6.210818767547607,
|
| 1043 |
+
"learning_rate": 2.5845470509244915e-05,
|
| 1044 |
+
"loss": 0.6873,
|
| 1045 |
+
"step": 7250
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.2621629996901147,
|
| 1049 |
+
"grad_norm": 6.774372577667236,
|
| 1050 |
+
"learning_rate": 2.5793822952174365e-05,
|
| 1051 |
+
"loss": 0.7195,
|
| 1052 |
+
"step": 7300
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.27765726681128,
|
| 1056 |
+
"grad_norm": 6.014688491821289,
|
| 1057 |
+
"learning_rate": 2.574217539510381e-05,
|
| 1058 |
+
"loss": 0.6298,
|
| 1059 |
+
"step": 7350
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.293151533932445,
|
| 1063 |
+
"grad_norm": 14.994784355163574,
|
| 1064 |
+
"learning_rate": 2.569052783803326e-05,
|
| 1065 |
+
"loss": 0.7403,
|
| 1066 |
+
"step": 7400
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.3086458010536104,
|
| 1070 |
+
"grad_norm": 6.315488815307617,
|
| 1071 |
+
"learning_rate": 2.563888028096271e-05,
|
| 1072 |
+
"loss": 0.6679,
|
| 1073 |
+
"step": 7450
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.3241400681747755,
|
| 1077 |
+
"grad_norm": 8.482314109802246,
|
| 1078 |
+
"learning_rate": 2.558723272389216e-05,
|
| 1079 |
+
"loss": 0.7173,
|
| 1080 |
+
"step": 7500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.3396343352959406,
|
| 1084 |
+
"grad_norm": 10.161298751831055,
|
| 1085 |
+
"learning_rate": 2.553558516682161e-05,
|
| 1086 |
+
"loss": 0.732,
|
| 1087 |
+
"step": 7550
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.3551286024171056,
|
| 1091 |
+
"grad_norm": 6.758267402648926,
|
| 1092 |
+
"learning_rate": 2.548393760975106e-05,
|
| 1093 |
+
"loss": 0.6192,
|
| 1094 |
+
"step": 7600
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.3706228695382707,
|
| 1098 |
+
"grad_norm": 4.528532981872559,
|
| 1099 |
+
"learning_rate": 2.543229005268051e-05,
|
| 1100 |
+
"loss": 0.7614,
|
| 1101 |
+
"step": 7650
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.3861171366594363,
|
| 1105 |
+
"grad_norm": 6.397975921630859,
|
| 1106 |
+
"learning_rate": 2.538064249560996e-05,
|
| 1107 |
+
"loss": 0.6951,
|
| 1108 |
+
"step": 7700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.4016114037806013,
|
| 1112 |
+
"grad_norm": 5.440258979797363,
|
| 1113 |
+
"learning_rate": 2.532899493853941e-05,
|
| 1114 |
+
"loss": 0.7188,
|
| 1115 |
+
"step": 7750
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.4171056709017664,
|
| 1119 |
+
"grad_norm": 2.4531173706054688,
|
| 1120 |
+
"learning_rate": 2.5277347381468858e-05,
|
| 1121 |
+
"loss": 0.6347,
|
| 1122 |
+
"step": 7800
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.4325999380229315,
|
| 1126 |
+
"grad_norm": 15.269991874694824,
|
| 1127 |
+
"learning_rate": 2.5225699824398304e-05,
|
| 1128 |
+
"loss": 0.6601,
|
| 1129 |
+
"step": 7850
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.4480942051440966,
|
| 1133 |
+
"grad_norm": 6.438554286956787,
|
| 1134 |
+
"learning_rate": 2.5174052267327754e-05,
|
| 1135 |
+
"loss": 0.698,
|
| 1136 |
+
"step": 7900
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.4635884722652617,
|
| 1140 |
+
"grad_norm": 8.922213554382324,
|
| 1141 |
+
"learning_rate": 2.5122404710257204e-05,
|
| 1142 |
+
"loss": 0.6958,
|
| 1143 |
+
"step": 7950
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.479082739386427,
|
| 1147 |
+
"grad_norm": 6.724533557891846,
|
| 1148 |
+
"learning_rate": 2.5070757153186657e-05,
|
| 1149 |
+
"loss": 0.7131,
|
| 1150 |
+
"step": 8000
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.4945770065075923,
|
| 1154 |
+
"grad_norm": 5.617169380187988,
|
| 1155 |
+
"learning_rate": 2.5019109596116107e-05,
|
| 1156 |
+
"loss": 0.7711,
|
| 1157 |
+
"step": 8050
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.5100712736287574,
|
| 1161 |
+
"grad_norm": 6.441185474395752,
|
| 1162 |
+
"learning_rate": 2.4967462039045553e-05,
|
| 1163 |
+
"loss": 0.6612,
|
| 1164 |
+
"step": 8100
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 2.5255655407499225,
|
| 1168 |
+
"grad_norm": 6.033916473388672,
|
| 1169 |
+
"learning_rate": 2.4915814481975003e-05,
|
| 1170 |
+
"loss": 0.698,
|
| 1171 |
+
"step": 8150
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 2.5410598078710875,
|
| 1175 |
+
"grad_norm": 6.174665451049805,
|
| 1176 |
+
"learning_rate": 2.4864166924904452e-05,
|
| 1177 |
+
"loss": 0.6968,
|
| 1178 |
+
"step": 8200
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.5565540749922526,
|
| 1182 |
+
"grad_norm": 20.01167869567871,
|
| 1183 |
+
"learning_rate": 2.4812519367833902e-05,
|
| 1184 |
+
"loss": 0.6456,
|
| 1185 |
+
"step": 8250
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.572048342113418,
|
| 1189 |
+
"grad_norm": 10.404682159423828,
|
| 1190 |
+
"learning_rate": 2.476087181076335e-05,
|
| 1191 |
+
"loss": 0.6808,
|
| 1192 |
+
"step": 8300
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.5875426092345832,
|
| 1196 |
+
"grad_norm": 5.160488128662109,
|
| 1197 |
+
"learning_rate": 2.47092242536928e-05,
|
| 1198 |
+
"loss": 0.6913,
|
| 1199 |
+
"step": 8350
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.6030368763557483,
|
| 1203 |
+
"grad_norm": 6.452591896057129,
|
| 1204 |
+
"learning_rate": 2.465757669662225e-05,
|
| 1205 |
+
"loss": 0.6594,
|
| 1206 |
+
"step": 8400
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 2.6185311434769134,
|
| 1210 |
+
"grad_norm": 12.436300277709961,
|
| 1211 |
+
"learning_rate": 2.46059291395517e-05,
|
| 1212 |
+
"loss": 0.7255,
|
| 1213 |
+
"step": 8450
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 2.6340254105980785,
|
| 1217 |
+
"grad_norm": 6.132791042327881,
|
| 1218 |
+
"learning_rate": 2.455428158248115e-05,
|
| 1219 |
+
"loss": 0.6753,
|
| 1220 |
+
"step": 8500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.649519677719244,
|
| 1224 |
+
"grad_norm": 10.712909698486328,
|
| 1225 |
+
"learning_rate": 2.45026340254106e-05,
|
| 1226 |
+
"loss": 0.6445,
|
| 1227 |
+
"step": 8550
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 2.665013944840409,
|
| 1231 |
+
"grad_norm": 12.122429847717285,
|
| 1232 |
+
"learning_rate": 2.445098646834005e-05,
|
| 1233 |
+
"loss": 0.6424,
|
| 1234 |
+
"step": 8600
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 2.680508211961574,
|
| 1238 |
+
"grad_norm": 8.575897216796875,
|
| 1239 |
+
"learning_rate": 2.4399338911269496e-05,
|
| 1240 |
+
"loss": 0.7242,
|
| 1241 |
+
"step": 8650
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.6960024790827393,
|
| 1245 |
+
"grad_norm": 8.740906715393066,
|
| 1246 |
+
"learning_rate": 2.4347691354198946e-05,
|
| 1247 |
+
"loss": 0.6949,
|
| 1248 |
+
"step": 8700
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 2.7114967462039044,
|
| 1252 |
+
"grad_norm": 4.871994972229004,
|
| 1253 |
+
"learning_rate": 2.4296043797128395e-05,
|
| 1254 |
+
"loss": 0.787,
|
| 1255 |
+
"step": 8750
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 2.72699101332507,
|
| 1259 |
+
"grad_norm": 6.642944812774658,
|
| 1260 |
+
"learning_rate": 2.424439624005785e-05,
|
| 1261 |
+
"loss": 0.6925,
|
| 1262 |
+
"step": 8800
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.742485280446235,
|
| 1266 |
+
"grad_norm": 12.149236679077148,
|
| 1267 |
+
"learning_rate": 2.4192748682987298e-05,
|
| 1268 |
+
"loss": 0.6972,
|
| 1269 |
+
"step": 8850
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 2.7579795475674,
|
| 1273 |
+
"grad_norm": 8.100613594055176,
|
| 1274 |
+
"learning_rate": 2.4141101125916744e-05,
|
| 1275 |
+
"loss": 0.7358,
|
| 1276 |
+
"step": 8900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 2.773473814688565,
|
| 1280 |
+
"grad_norm": 12.28987979888916,
|
| 1281 |
+
"learning_rate": 2.4089453568846194e-05,
|
| 1282 |
+
"loss": 0.7176,
|
| 1283 |
+
"step": 8950
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.78896808180973,
|
| 1287 |
+
"grad_norm": 9.355488777160645,
|
| 1288 |
+
"learning_rate": 2.4037806011775644e-05,
|
| 1289 |
+
"loss": 0.6856,
|
| 1290 |
+
"step": 9000
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 2.8044623489308957,
|
| 1294 |
+
"grad_norm": 11.875406265258789,
|
| 1295 |
+
"learning_rate": 2.3986158454705093e-05,
|
| 1296 |
+
"loss": 0.6501,
|
| 1297 |
+
"step": 9050
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 2.819956616052061,
|
| 1301 |
+
"grad_norm": 8.061235427856445,
|
| 1302 |
+
"learning_rate": 2.3934510897634543e-05,
|
| 1303 |
+
"loss": 0.6823,
|
| 1304 |
+
"step": 9100
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.835450883173226,
|
| 1308 |
+
"grad_norm": 7.949320316314697,
|
| 1309 |
+
"learning_rate": 2.388286334056399e-05,
|
| 1310 |
+
"loss": 0.6764,
|
| 1311 |
+
"step": 9150
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 2.850945150294391,
|
| 1315 |
+
"grad_norm": 5.9249587059021,
|
| 1316 |
+
"learning_rate": 2.3831215783493442e-05,
|
| 1317 |
+
"loss": 0.6511,
|
| 1318 |
+
"step": 9200
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 2.866439417415556,
|
| 1322 |
+
"grad_norm": 8.400185585021973,
|
| 1323 |
+
"learning_rate": 2.3779568226422892e-05,
|
| 1324 |
+
"loss": 0.6515,
|
| 1325 |
+
"step": 9250
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.8819336845367216,
|
| 1329 |
+
"grad_norm": 11.487894058227539,
|
| 1330 |
+
"learning_rate": 2.3727920669352342e-05,
|
| 1331 |
+
"loss": 0.6719,
|
| 1332 |
+
"step": 9300
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 2.8974279516578867,
|
| 1336 |
+
"grad_norm": 8.317901611328125,
|
| 1337 |
+
"learning_rate": 2.367627311228179e-05,
|
| 1338 |
+
"loss": 0.6697,
|
| 1339 |
+
"step": 9350
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 2.9129222187790518,
|
| 1343 |
+
"grad_norm": 9.878332138061523,
|
| 1344 |
+
"learning_rate": 2.3624625555211238e-05,
|
| 1345 |
+
"loss": 0.6801,
|
| 1346 |
+
"step": 9400
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 2.928416485900217,
|
| 1350 |
+
"grad_norm": 8.855628967285156,
|
| 1351 |
+
"learning_rate": 2.3572977998140687e-05,
|
| 1352 |
+
"loss": 0.6445,
|
| 1353 |
+
"step": 9450
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 2.943910753021382,
|
| 1357 |
+
"grad_norm": 5.350094318389893,
|
| 1358 |
+
"learning_rate": 2.3521330441070137e-05,
|
| 1359 |
+
"loss": 0.6891,
|
| 1360 |
+
"step": 9500
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 2.9594050201425475,
|
| 1364 |
+
"grad_norm": 8.540812492370605,
|
| 1365 |
+
"learning_rate": 2.3469682883999587e-05,
|
| 1366 |
+
"loss": 0.6556,
|
| 1367 |
+
"step": 9550
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.9748992872637126,
|
| 1371 |
+
"grad_norm": 4.337664604187012,
|
| 1372 |
+
"learning_rate": 2.341803532692904e-05,
|
| 1373 |
+
"loss": 0.7013,
|
| 1374 |
+
"step": 9600
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 2.9903935543848776,
|
| 1378 |
+
"grad_norm": 7.002617359161377,
|
| 1379 |
+
"learning_rate": 2.3366387769858486e-05,
|
| 1380 |
+
"loss": 0.6518,
|
| 1381 |
+
"step": 9650
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 3.0,
|
| 1385 |
+
"eval_accuracy": 0.777292576419214,
|
| 1386 |
+
"eval_f1": 0.775972842719567,
|
| 1387 |
+
"eval_loss": 0.69657963514328,
|
| 1388 |
+
"eval_runtime": 25.4856,
|
| 1389 |
+
"eval_samples_per_second": 260.579,
|
| 1390 |
+
"eval_steps_per_second": 16.323,
|
| 1391 |
+
"step": 9681
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 3.0058878215060427,
|
| 1395 |
+
"grad_norm": 9.228548049926758,
|
| 1396 |
+
"learning_rate": 2.3314740212787936e-05,
|
| 1397 |
+
"loss": 0.5385,
|
| 1398 |
+
"step": 9700
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 3.021382088627208,
|
| 1402 |
+
"grad_norm": 4.332932472229004,
|
| 1403 |
+
"learning_rate": 2.3263092655717385e-05,
|
| 1404 |
+
"loss": 0.5513,
|
| 1405 |
+
"step": 9750
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 3.036876355748373,
|
| 1409 |
+
"grad_norm": 6.478864669799805,
|
| 1410 |
+
"learning_rate": 2.3211445098646835e-05,
|
| 1411 |
+
"loss": 0.4542,
|
| 1412 |
+
"step": 9800
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 3.0523706228695384,
|
| 1416 |
+
"grad_norm": 14.028499603271484,
|
| 1417 |
+
"learning_rate": 2.3159797541576285e-05,
|
| 1418 |
+
"loss": 0.4549,
|
| 1419 |
+
"step": 9850
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 3.0678648899907035,
|
| 1423 |
+
"grad_norm": 5.590787887573242,
|
| 1424 |
+
"learning_rate": 2.310814998450573e-05,
|
| 1425 |
+
"loss": 0.4624,
|
| 1426 |
+
"step": 9900
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 3.0833591571118686,
|
| 1430 |
+
"grad_norm": 5.623167514801025,
|
| 1431 |
+
"learning_rate": 2.305650242743518e-05,
|
| 1432 |
+
"loss": 0.4479,
|
| 1433 |
+
"step": 9950
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 3.0988534242330337,
|
| 1437 |
+
"grad_norm": 10.343826293945312,
|
| 1438 |
+
"learning_rate": 2.3004854870364634e-05,
|
| 1439 |
+
"loss": 0.5079,
|
| 1440 |
+
"step": 10000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 3.1143476913541988,
|
| 1444 |
+
"grad_norm": 2.780686616897583,
|
| 1445 |
+
"learning_rate": 2.2953207313294084e-05,
|
| 1446 |
+
"loss": 0.43,
|
| 1447 |
+
"step": 10050
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 3.1298419584753643,
|
| 1451 |
+
"grad_norm": 10.917914390563965,
|
| 1452 |
+
"learning_rate": 2.2901559756223533e-05,
|
| 1453 |
+
"loss": 0.4932,
|
| 1454 |
+
"step": 10100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 3.1453362255965294,
|
| 1458 |
+
"grad_norm": 14.870561599731445,
|
| 1459 |
+
"learning_rate": 2.284991219915298e-05,
|
| 1460 |
+
"loss": 0.483,
|
| 1461 |
+
"step": 10150
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 3.1608304927176945,
|
| 1465 |
+
"grad_norm": 15.64564323425293,
|
| 1466 |
+
"learning_rate": 2.279826464208243e-05,
|
| 1467 |
+
"loss": 0.5047,
|
| 1468 |
+
"step": 10200
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 3.1763247598388595,
|
| 1472 |
+
"grad_norm": 8.148391723632812,
|
| 1473 |
+
"learning_rate": 2.274661708501188e-05,
|
| 1474 |
+
"loss": 0.498,
|
| 1475 |
+
"step": 10250
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 3.1918190269600246,
|
| 1479 |
+
"grad_norm": 9.916448593139648,
|
| 1480 |
+
"learning_rate": 2.269496952794133e-05,
|
| 1481 |
+
"loss": 0.4896,
|
| 1482 |
+
"step": 10300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 3.20731329408119,
|
| 1486 |
+
"grad_norm": 10.014134407043457,
|
| 1487 |
+
"learning_rate": 2.2643321970870778e-05,
|
| 1488 |
+
"loss": 0.4572,
|
| 1489 |
+
"step": 10350
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 3.2228075612023552,
|
| 1493 |
+
"grad_norm": 9.647527694702148,
|
| 1494 |
+
"learning_rate": 2.2591674413800228e-05,
|
| 1495 |
+
"loss": 0.4965,
|
| 1496 |
+
"step": 10400
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 3.2383018283235203,
|
| 1500 |
+
"grad_norm": 11.77087116241455,
|
| 1501 |
+
"learning_rate": 2.2540026856729678e-05,
|
| 1502 |
+
"loss": 0.512,
|
| 1503 |
+
"step": 10450
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 3.2537960954446854,
|
| 1507 |
+
"grad_norm": 3.3613386154174805,
|
| 1508 |
+
"learning_rate": 2.2488379299659127e-05,
|
| 1509 |
+
"loss": 0.5522,
|
| 1510 |
+
"step": 10500
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 3.2692903625658505,
|
| 1514 |
+
"grad_norm": 17.92693519592285,
|
| 1515 |
+
"learning_rate": 2.2436731742588577e-05,
|
| 1516 |
+
"loss": 0.4915,
|
| 1517 |
+
"step": 10550
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 3.2847846296870156,
|
| 1521 |
+
"grad_norm": 8.389365196228027,
|
| 1522 |
+
"learning_rate": 2.2385084185518027e-05,
|
| 1523 |
+
"loss": 0.5343,
|
| 1524 |
+
"step": 10600
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 3.300278896808181,
|
| 1528 |
+
"grad_norm": 9.849445343017578,
|
| 1529 |
+
"learning_rate": 2.2333436628447473e-05,
|
| 1530 |
+
"loss": 0.4925,
|
| 1531 |
+
"step": 10650
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 3.315773163929346,
|
| 1535 |
+
"grad_norm": 7.494227886199951,
|
| 1536 |
+
"learning_rate": 2.2281789071376923e-05,
|
| 1537 |
+
"loss": 0.5242,
|
| 1538 |
+
"step": 10700
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 3.3312674310505113,
|
| 1542 |
+
"grad_norm": 12.774617195129395,
|
| 1543 |
+
"learning_rate": 2.2230141514306372e-05,
|
| 1544 |
+
"loss": 0.522,
|
| 1545 |
+
"step": 10750
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 3.3467616981716763,
|
| 1549 |
+
"grad_norm": 4.167229175567627,
|
| 1550 |
+
"learning_rate": 2.2178493957235822e-05,
|
| 1551 |
+
"loss": 0.4852,
|
| 1552 |
+
"step": 10800
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 3.3622559652928414,
|
| 1556 |
+
"grad_norm": 7.823596000671387,
|
| 1557 |
+
"learning_rate": 2.2126846400165275e-05,
|
| 1558 |
+
"loss": 0.521,
|
| 1559 |
+
"step": 10850
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 3.377750232414007,
|
| 1563 |
+
"grad_norm": 9.712186813354492,
|
| 1564 |
+
"learning_rate": 2.2075198843094725e-05,
|
| 1565 |
+
"loss": 0.4931,
|
| 1566 |
+
"step": 10900
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.393244499535172,
|
| 1570 |
+
"grad_norm": 9.726935386657715,
|
| 1571 |
+
"learning_rate": 2.202355128602417e-05,
|
| 1572 |
+
"loss": 0.531,
|
| 1573 |
+
"step": 10950
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 3.408738766656337,
|
| 1577 |
+
"grad_norm": 8.613348007202148,
|
| 1578 |
+
"learning_rate": 2.197190372895362e-05,
|
| 1579 |
+
"loss": 0.4902,
|
| 1580 |
+
"step": 11000
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 3.424233033777502,
|
| 1584 |
+
"grad_norm": 17.698650360107422,
|
| 1585 |
+
"learning_rate": 2.192025617188307e-05,
|
| 1586 |
+
"loss": 0.4967,
|
| 1587 |
+
"step": 11050
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.4397273008986673,
|
| 1591 |
+
"grad_norm": 13.304680824279785,
|
| 1592 |
+
"learning_rate": 2.186860861481252e-05,
|
| 1593 |
+
"loss": 0.4998,
|
| 1594 |
+
"step": 11100
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 3.455221568019833,
|
| 1598 |
+
"grad_norm": 9.090615272521973,
|
| 1599 |
+
"learning_rate": 2.181696105774197e-05,
|
| 1600 |
+
"loss": 0.4797,
|
| 1601 |
+
"step": 11150
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.470715835140998,
|
| 1605 |
+
"grad_norm": 6.544071197509766,
|
| 1606 |
+
"learning_rate": 2.176531350067142e-05,
|
| 1607 |
+
"loss": 0.5405,
|
| 1608 |
+
"step": 11200
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 3.486210102262163,
|
| 1612 |
+
"grad_norm": 10.908158302307129,
|
| 1613 |
+
"learning_rate": 2.171366594360087e-05,
|
| 1614 |
+
"loss": 0.4663,
|
| 1615 |
+
"step": 11250
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 3.501704369383328,
|
| 1619 |
+
"grad_norm": 9.044700622558594,
|
| 1620 |
+
"learning_rate": 2.166201838653032e-05,
|
| 1621 |
+
"loss": 0.4755,
|
| 1622 |
+
"step": 11300
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 3.517198636504493,
|
| 1626 |
+
"grad_norm": 7.633232116699219,
|
| 1627 |
+
"learning_rate": 2.161037082945977e-05,
|
| 1628 |
+
"loss": 0.4182,
|
| 1629 |
+
"step": 11350
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 3.5326929036256587,
|
| 1633 |
+
"grad_norm": 5.32473087310791,
|
| 1634 |
+
"learning_rate": 2.1558723272389218e-05,
|
| 1635 |
+
"loss": 0.4987,
|
| 1636 |
+
"step": 11400
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 3.5481871707468238,
|
| 1640 |
+
"grad_norm": 9.8456392288208,
|
| 1641 |
+
"learning_rate": 2.1507075715318664e-05,
|
| 1642 |
+
"loss": 0.587,
|
| 1643 |
+
"step": 11450
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 3.563681437867989,
|
| 1647 |
+
"grad_norm": 12.52115535736084,
|
| 1648 |
+
"learning_rate": 2.1455428158248114e-05,
|
| 1649 |
+
"loss": 0.5331,
|
| 1650 |
+
"step": 11500
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 3.579175704989154,
|
| 1654 |
+
"grad_norm": 18.225566864013672,
|
| 1655 |
+
"learning_rate": 2.1403780601177564e-05,
|
| 1656 |
+
"loss": 0.4794,
|
| 1657 |
+
"step": 11550
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 3.594669972110319,
|
| 1661 |
+
"grad_norm": 8.749368667602539,
|
| 1662 |
+
"learning_rate": 2.1352133044107013e-05,
|
| 1663 |
+
"loss": 0.4967,
|
| 1664 |
+
"step": 11600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 3.6101642392314846,
|
| 1668 |
+
"grad_norm": 8.760223388671875,
|
| 1669 |
+
"learning_rate": 2.1300485487036466e-05,
|
| 1670 |
+
"loss": 0.4963,
|
| 1671 |
+
"step": 11650
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 3.6256585063526496,
|
| 1675 |
+
"grad_norm": 15.518270492553711,
|
| 1676 |
+
"learning_rate": 2.1248837929965913e-05,
|
| 1677 |
+
"loss": 0.4456,
|
| 1678 |
+
"step": 11700
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 3.6411527734738147,
|
| 1682 |
+
"grad_norm": 9.451664924621582,
|
| 1683 |
+
"learning_rate": 2.1197190372895362e-05,
|
| 1684 |
+
"loss": 0.5235,
|
| 1685 |
+
"step": 11750
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 3.65664704059498,
|
| 1689 |
+
"grad_norm": 17.736055374145508,
|
| 1690 |
+
"learning_rate": 2.1145542815824812e-05,
|
| 1691 |
+
"loss": 0.4876,
|
| 1692 |
+
"step": 11800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 3.672141307716145,
|
| 1696 |
+
"grad_norm": 24.323490142822266,
|
| 1697 |
+
"learning_rate": 2.1093895258754262e-05,
|
| 1698 |
+
"loss": 0.483,
|
| 1699 |
+
"step": 11850
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 3.6876355748373104,
|
| 1703 |
+
"grad_norm": 15.389254570007324,
|
| 1704 |
+
"learning_rate": 2.104224770168371e-05,
|
| 1705 |
+
"loss": 0.5127,
|
| 1706 |
+
"step": 11900
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 3.7031298419584755,
|
| 1710 |
+
"grad_norm": 11.283272743225098,
|
| 1711 |
+
"learning_rate": 2.0990600144613158e-05,
|
| 1712 |
+
"loss": 0.5282,
|
| 1713 |
+
"step": 11950
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 3.7186241090796406,
|
| 1717 |
+
"grad_norm": 11.002310752868652,
|
| 1718 |
+
"learning_rate": 2.0938952587542607e-05,
|
| 1719 |
+
"loss": 0.5459,
|
| 1720 |
+
"step": 12000
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 3.7341183762008057,
|
| 1724 |
+
"grad_norm": 6.972140312194824,
|
| 1725 |
+
"learning_rate": 2.088730503047206e-05,
|
| 1726 |
+
"loss": 0.536,
|
| 1727 |
+
"step": 12050
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 3.7496126433219708,
|
| 1731 |
+
"grad_norm": 4.202858924865723,
|
| 1732 |
+
"learning_rate": 2.083565747340151e-05,
|
| 1733 |
+
"loss": 0.5736,
|
| 1734 |
+
"step": 12100
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 3.7651069104431363,
|
| 1738 |
+
"grad_norm": 15.748515129089355,
|
| 1739 |
+
"learning_rate": 2.078400991633096e-05,
|
| 1740 |
+
"loss": 0.4715,
|
| 1741 |
+
"step": 12150
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 3.7806011775643014,
|
| 1745 |
+
"grad_norm": 6.696774482727051,
|
| 1746 |
+
"learning_rate": 2.0732362359260406e-05,
|
| 1747 |
+
"loss": 0.5545,
|
| 1748 |
+
"step": 12200
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 3.7960954446854664,
|
| 1752 |
+
"grad_norm": 7.366288661956787,
|
| 1753 |
+
"learning_rate": 2.0680714802189856e-05,
|
| 1754 |
+
"loss": 0.5736,
|
| 1755 |
+
"step": 12250
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 3.8115897118066315,
|
| 1759 |
+
"grad_norm": 13.58438777923584,
|
| 1760 |
+
"learning_rate": 2.0629067245119306e-05,
|
| 1761 |
+
"loss": 0.4255,
|
| 1762 |
+
"step": 12300
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 3.8270839789277966,
|
| 1766 |
+
"grad_norm": 9.109688758850098,
|
| 1767 |
+
"learning_rate": 2.0577419688048755e-05,
|
| 1768 |
+
"loss": 0.4565,
|
| 1769 |
+
"step": 12350
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 3.842578246048962,
|
| 1773 |
+
"grad_norm": 11.448044776916504,
|
| 1774 |
+
"learning_rate": 2.0525772130978205e-05,
|
| 1775 |
+
"loss": 0.5117,
|
| 1776 |
+
"step": 12400
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 3.858072513170127,
|
| 1780 |
+
"grad_norm": 6.876945495605469,
|
| 1781 |
+
"learning_rate": 2.0474124573907655e-05,
|
| 1782 |
+
"loss": 0.5543,
|
| 1783 |
+
"step": 12450
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 3.8735667802912923,
|
| 1787 |
+
"grad_norm": 11.25009536743164,
|
| 1788 |
+
"learning_rate": 2.0422477016837104e-05,
|
| 1789 |
+
"loss": 0.456,
|
| 1790 |
+
"step": 12500
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 3.8890610474124574,
|
| 1794 |
+
"grad_norm": 13.992502212524414,
|
| 1795 |
+
"learning_rate": 2.0370829459766554e-05,
|
| 1796 |
+
"loss": 0.4907,
|
| 1797 |
+
"step": 12550
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 3.9045553145336225,
|
| 1801 |
+
"grad_norm": 11.92656421661377,
|
| 1802 |
+
"learning_rate": 2.0319181902696004e-05,
|
| 1803 |
+
"loss": 0.4841,
|
| 1804 |
+
"step": 12600
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 3.9200495816547876,
|
| 1808 |
+
"grad_norm": 7.212582111358643,
|
| 1809 |
+
"learning_rate": 2.0267534345625453e-05,
|
| 1810 |
+
"loss": 0.5529,
|
| 1811 |
+
"step": 12650
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 3.9355438487759526,
|
| 1815 |
+
"grad_norm": 14.616645812988281,
|
| 1816 |
+
"learning_rate": 2.02158867885549e-05,
|
| 1817 |
+
"loss": 0.5366,
|
| 1818 |
+
"step": 12700
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 3.951038115897118,
|
| 1822 |
+
"grad_norm": 9.052292823791504,
|
| 1823 |
+
"learning_rate": 2.016423923148435e-05,
|
| 1824 |
+
"loss": 0.5459,
|
| 1825 |
+
"step": 12750
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 3.9665323830182833,
|
| 1829 |
+
"grad_norm": 18.27539825439453,
|
| 1830 |
+
"learning_rate": 2.01125916744138e-05,
|
| 1831 |
+
"loss": 0.5631,
|
| 1832 |
+
"step": 12800
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 3.9820266501394483,
|
| 1836 |
+
"grad_norm": 12.429372787475586,
|
| 1837 |
+
"learning_rate": 2.0060944117343252e-05,
|
| 1838 |
+
"loss": 0.4885,
|
| 1839 |
+
"step": 12850
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 3.9975209172606134,
|
| 1843 |
+
"grad_norm": 4.481673240661621,
|
| 1844 |
+
"learning_rate": 2.00092965602727e-05,
|
| 1845 |
+
"loss": 0.4565,
|
| 1846 |
+
"step": 12900
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 4.0,
|
| 1850 |
+
"eval_accuracy": 0.7863273603372986,
|
| 1851 |
+
"eval_f1": 0.7874334496964743,
|
| 1852 |
+
"eval_loss": 0.7491569519042969,
|
| 1853 |
+
"eval_runtime": 25.5609,
|
| 1854 |
+
"eval_samples_per_second": 259.811,
|
| 1855 |
+
"eval_steps_per_second": 16.275,
|
| 1856 |
+
"step": 12908
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 4.0130151843817785,
|
| 1860 |
+
"grad_norm": 21.53974151611328,
|
| 1861 |
+
"learning_rate": 1.9957649003202148e-05,
|
| 1862 |
+
"loss": 0.2695,
|
| 1863 |
+
"step": 12950
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 4.028509451502944,
|
| 1867 |
+
"grad_norm": 9.07942008972168,
|
| 1868 |
+
"learning_rate": 1.9906001446131598e-05,
|
| 1869 |
+
"loss": 0.3282,
|
| 1870 |
+
"step": 13000
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 4.044003718624109,
|
| 1874 |
+
"grad_norm": 16.323549270629883,
|
| 1875 |
+
"learning_rate": 1.9854353889061047e-05,
|
| 1876 |
+
"loss": 0.3079,
|
| 1877 |
+
"step": 13050
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 4.059497985745274,
|
| 1881 |
+
"grad_norm": 6.679697036743164,
|
| 1882 |
+
"learning_rate": 1.9802706331990497e-05,
|
| 1883 |
+
"loss": 0.3889,
|
| 1884 |
+
"step": 13100
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 4.07499225286644,
|
| 1888 |
+
"grad_norm": 17.357574462890625,
|
| 1889 |
+
"learning_rate": 1.9751058774919947e-05,
|
| 1890 |
+
"loss": 0.3336,
|
| 1891 |
+
"step": 13150
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 4.090486519987604,
|
| 1895 |
+
"grad_norm": 5.116195201873779,
|
| 1896 |
+
"learning_rate": 1.9699411217849393e-05,
|
| 1897 |
+
"loss": 0.3102,
|
| 1898 |
+
"step": 13200
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 4.10598078710877,
|
| 1902 |
+
"grad_norm": 29.05538558959961,
|
| 1903 |
+
"learning_rate": 1.9647763660778846e-05,
|
| 1904 |
+
"loss": 0.2902,
|
| 1905 |
+
"step": 13250
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 4.1214750542299345,
|
| 1909 |
+
"grad_norm": 6.254473686218262,
|
| 1910 |
+
"learning_rate": 1.9596116103708296e-05,
|
| 1911 |
+
"loss": 0.3816,
|
| 1912 |
+
"step": 13300
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 4.1369693213511,
|
| 1916 |
+
"grad_norm": 11.854185104370117,
|
| 1917 |
+
"learning_rate": 1.9544468546637745e-05,
|
| 1918 |
+
"loss": 0.3455,
|
| 1919 |
+
"step": 13350
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 4.152463588472266,
|
| 1923 |
+
"grad_norm": 16.399444580078125,
|
| 1924 |
+
"learning_rate": 1.9492820989567195e-05,
|
| 1925 |
+
"loss": 0.3713,
|
| 1926 |
+
"step": 13400
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 4.16795785559343,
|
| 1930 |
+
"grad_norm": 18.26226234436035,
|
| 1931 |
+
"learning_rate": 1.9441173432496645e-05,
|
| 1932 |
+
"loss": 0.2957,
|
| 1933 |
+
"step": 13450
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 4.183452122714596,
|
| 1937 |
+
"grad_norm": 6.590181350708008,
|
| 1938 |
+
"learning_rate": 1.938952587542609e-05,
|
| 1939 |
+
"loss": 0.2905,
|
| 1940 |
+
"step": 13500
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 4.19894638983576,
|
| 1944 |
+
"grad_norm": 5.3814849853515625,
|
| 1945 |
+
"learning_rate": 1.933787831835554e-05,
|
| 1946 |
+
"loss": 0.3782,
|
| 1947 |
+
"step": 13550
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 4.214440656956926,
|
| 1951 |
+
"grad_norm": 8.641956329345703,
|
| 1952 |
+
"learning_rate": 1.928623076128499e-05,
|
| 1953 |
+
"loss": 0.3211,
|
| 1954 |
+
"step": 13600
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 4.2299349240780915,
|
| 1958 |
+
"grad_norm": 14.346405982971191,
|
| 1959 |
+
"learning_rate": 1.9234583204214443e-05,
|
| 1960 |
+
"loss": 0.3274,
|
| 1961 |
+
"step": 13650
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 4.245429191199256,
|
| 1965 |
+
"grad_norm": 15.577725410461426,
|
| 1966 |
+
"learning_rate": 1.9182935647143893e-05,
|
| 1967 |
+
"loss": 0.3568,
|
| 1968 |
+
"step": 13700
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 4.260923458320422,
|
| 1972 |
+
"grad_norm": 9.855398178100586,
|
| 1973 |
+
"learning_rate": 1.913128809007334e-05,
|
| 1974 |
+
"loss": 0.3008,
|
| 1975 |
+
"step": 13750
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 4.276417725441586,
|
| 1979 |
+
"grad_norm": 15.720294952392578,
|
| 1980 |
+
"learning_rate": 1.907964053300279e-05,
|
| 1981 |
+
"loss": 0.2979,
|
| 1982 |
+
"step": 13800
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 4.291911992562752,
|
| 1986 |
+
"grad_norm": 13.976778030395508,
|
| 1987 |
+
"learning_rate": 1.902799297593224e-05,
|
| 1988 |
+
"loss": 0.3389,
|
| 1989 |
+
"step": 13850
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 4.307406259683917,
|
| 1993 |
+
"grad_norm": 19.255727767944336,
|
| 1994 |
+
"learning_rate": 1.897634541886169e-05,
|
| 1995 |
+
"loss": 0.3636,
|
| 1996 |
+
"step": 13900
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 4.322900526805082,
|
| 2000 |
+
"grad_norm": 10.70836353302002,
|
| 2001 |
+
"learning_rate": 1.8924697861791138e-05,
|
| 2002 |
+
"loss": 0.3455,
|
| 2003 |
+
"step": 13950
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 4.3383947939262475,
|
| 2007 |
+
"grad_norm": 0.9212763905525208,
|
| 2008 |
+
"learning_rate": 1.8873050304720584e-05,
|
| 2009 |
+
"loss": 0.3703,
|
| 2010 |
+
"step": 14000
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 4.353889061047412,
|
| 2014 |
+
"grad_norm": 10.232623100280762,
|
| 2015 |
+
"learning_rate": 1.8821402747650037e-05,
|
| 2016 |
+
"loss": 0.3247,
|
| 2017 |
+
"step": 14050
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 4.369383328168578,
|
| 2021 |
+
"grad_norm": 11.130922317504883,
|
| 2022 |
+
"learning_rate": 1.8769755190579487e-05,
|
| 2023 |
+
"loss": 0.3107,
|
| 2024 |
+
"step": 14100
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 4.384877595289743,
|
| 2028 |
+
"grad_norm": 10.536752700805664,
|
| 2029 |
+
"learning_rate": 1.8718107633508937e-05,
|
| 2030 |
+
"loss": 0.3614,
|
| 2031 |
+
"step": 14150
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 4.400371862410908,
|
| 2035 |
+
"grad_norm": 15.330968856811523,
|
| 2036 |
+
"learning_rate": 1.8666460076438386e-05,
|
| 2037 |
+
"loss": 0.3984,
|
| 2038 |
+
"step": 14200
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 4.415866129532073,
|
| 2042 |
+
"grad_norm": 7.436588764190674,
|
| 2043 |
+
"learning_rate": 1.8614812519367833e-05,
|
| 2044 |
+
"loss": 0.3257,
|
| 2045 |
+
"step": 14250
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 4.431360396653238,
|
| 2049 |
+
"grad_norm": 7.192384243011475,
|
| 2050 |
+
"learning_rate": 1.8563164962297282e-05,
|
| 2051 |
+
"loss": 0.3254,
|
| 2052 |
+
"step": 14300
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 4.4468546637744035,
|
| 2056 |
+
"grad_norm": 7.792993545532227,
|
| 2057 |
+
"learning_rate": 1.8511517405226732e-05,
|
| 2058 |
+
"loss": 0.3392,
|
| 2059 |
+
"step": 14350
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 4.462348930895569,
|
| 2063 |
+
"grad_norm": 12.411416053771973,
|
| 2064 |
+
"learning_rate": 1.8459869848156182e-05,
|
| 2065 |
+
"loss": 0.3383,
|
| 2066 |
+
"step": 14400
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 4.477843198016734,
|
| 2070 |
+
"grad_norm": 17.897613525390625,
|
| 2071 |
+
"learning_rate": 1.8408222291085635e-05,
|
| 2072 |
+
"loss": 0.3392,
|
| 2073 |
+
"step": 14450
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 4.493337465137899,
|
| 2077 |
+
"grad_norm": 23.59228515625,
|
| 2078 |
+
"learning_rate": 1.835657473401508e-05,
|
| 2079 |
+
"loss": 0.3055,
|
| 2080 |
+
"step": 14500
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 4.508831732259064,
|
| 2084 |
+
"grad_norm": 13.722383499145508,
|
| 2085 |
+
"learning_rate": 1.830492717694453e-05,
|
| 2086 |
+
"loss": 0.3997,
|
| 2087 |
+
"step": 14550
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 4.524325999380229,
|
| 2091 |
+
"grad_norm": 17.811538696289062,
|
| 2092 |
+
"learning_rate": 1.825327961987398e-05,
|
| 2093 |
+
"loss": 0.266,
|
| 2094 |
+
"step": 14600
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 4.539820266501394,
|
| 2098 |
+
"grad_norm": 10.993431091308594,
|
| 2099 |
+
"learning_rate": 1.820163206280343e-05,
|
| 2100 |
+
"loss": 0.2634,
|
| 2101 |
+
"step": 14650
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 4.55531453362256,
|
| 2105 |
+
"grad_norm": 5.25628137588501,
|
| 2106 |
+
"learning_rate": 1.814998450573288e-05,
|
| 2107 |
+
"loss": 0.3563,
|
| 2108 |
+
"step": 14700
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 4.570808800743725,
|
| 2112 |
+
"grad_norm": 16.91241455078125,
|
| 2113 |
+
"learning_rate": 1.8098336948662326e-05,
|
| 2114 |
+
"loss": 0.3298,
|
| 2115 |
+
"step": 14750
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 4.58630306786489,
|
| 2119 |
+
"grad_norm": 27.083995819091797,
|
| 2120 |
+
"learning_rate": 1.8046689391591776e-05,
|
| 2121 |
+
"loss": 0.36,
|
| 2122 |
+
"step": 14800
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 4.601797334986055,
|
| 2126 |
+
"grad_norm": 19.726198196411133,
|
| 2127 |
+
"learning_rate": 1.799504183452123e-05,
|
| 2128 |
+
"loss": 0.3224,
|
| 2129 |
+
"step": 14850
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 4.617291602107221,
|
| 2133 |
+
"grad_norm": 6.92859411239624,
|
| 2134 |
+
"learning_rate": 1.794339427745068e-05,
|
| 2135 |
+
"loss": 0.335,
|
| 2136 |
+
"step": 14900
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 4.632785869228385,
|
| 2140 |
+
"grad_norm": 15.97644329071045,
|
| 2141 |
+
"learning_rate": 1.7891746720380128e-05,
|
| 2142 |
+
"loss": 0.3303,
|
| 2143 |
+
"step": 14950
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 4.648280136349551,
|
| 2147 |
+
"grad_norm": 24.399837493896484,
|
| 2148 |
+
"learning_rate": 1.7840099163309575e-05,
|
| 2149 |
+
"loss": 0.3492,
|
| 2150 |
+
"step": 15000
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 4.663774403470716,
|
| 2154 |
+
"grad_norm": 10.855368614196777,
|
| 2155 |
+
"learning_rate": 1.7788451606239024e-05,
|
| 2156 |
+
"loss": 0.3278,
|
| 2157 |
+
"step": 15050
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 4.679268670591881,
|
| 2161 |
+
"grad_norm": 20.869380950927734,
|
| 2162 |
+
"learning_rate": 1.7736804049168474e-05,
|
| 2163 |
+
"loss": 0.2913,
|
| 2164 |
+
"step": 15100
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 4.694762937713046,
|
| 2168 |
+
"grad_norm": 6.862913131713867,
|
| 2169 |
+
"learning_rate": 1.7685156492097924e-05,
|
| 2170 |
+
"loss": 0.3133,
|
| 2171 |
+
"step": 15150
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 4.710257204834211,
|
| 2175 |
+
"grad_norm": 19.621482849121094,
|
| 2176 |
+
"learning_rate": 1.7633508935027373e-05,
|
| 2177 |
+
"loss": 0.3456,
|
| 2178 |
+
"step": 15200
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 4.725751471955377,
|
| 2182 |
+
"grad_norm": 19.79738998413086,
|
| 2183 |
+
"learning_rate": 1.7581861377956823e-05,
|
| 2184 |
+
"loss": 0.3562,
|
| 2185 |
+
"step": 15250
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 4.7412457390765415,
|
| 2189 |
+
"grad_norm": 3.2352957725524902,
|
| 2190 |
+
"learning_rate": 1.7530213820886273e-05,
|
| 2191 |
+
"loss": 0.3565,
|
| 2192 |
+
"step": 15300
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 4.756740006197707,
|
| 2196 |
+
"grad_norm": 10.959282875061035,
|
| 2197 |
+
"learning_rate": 1.7478566263815722e-05,
|
| 2198 |
+
"loss": 0.3472,
|
| 2199 |
+
"step": 15350
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 4.7722342733188725,
|
| 2203 |
+
"grad_norm": 3.22469162940979,
|
| 2204 |
+
"learning_rate": 1.7426918706745172e-05,
|
| 2205 |
+
"loss": 0.3367,
|
| 2206 |
+
"step": 15400
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 4.787728540440037,
|
| 2210 |
+
"grad_norm": 7.619373798370361,
|
| 2211 |
+
"learning_rate": 1.737527114967462e-05,
|
| 2212 |
+
"loss": 0.319,
|
| 2213 |
+
"step": 15450
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 4.803222807561203,
|
| 2217 |
+
"grad_norm": 24.706689834594727,
|
| 2218 |
+
"learning_rate": 1.7323623592604068e-05,
|
| 2219 |
+
"loss": 0.3939,
|
| 2220 |
+
"step": 15500
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 4.818717074682367,
|
| 2224 |
+
"grad_norm": 15.918986320495605,
|
| 2225 |
+
"learning_rate": 1.7271976035533518e-05,
|
| 2226 |
+
"loss": 0.3618,
|
| 2227 |
+
"step": 15550
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.834211341803533,
|
| 2231 |
+
"grad_norm": 14.518546104431152,
|
| 2232 |
+
"learning_rate": 1.7220328478462967e-05,
|
| 2233 |
+
"loss": 0.4082,
|
| 2234 |
+
"step": 15600
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 4.8497056089246975,
|
| 2238 |
+
"grad_norm": 6.084866046905518,
|
| 2239 |
+
"learning_rate": 1.716868092139242e-05,
|
| 2240 |
+
"loss": 0.3594,
|
| 2241 |
+
"step": 15650
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 4.865199876045863,
|
| 2245 |
+
"grad_norm": 18.435983657836914,
|
| 2246 |
+
"learning_rate": 1.711703336432187e-05,
|
| 2247 |
+
"loss": 0.3182,
|
| 2248 |
+
"step": 15700
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 4.8806941431670285,
|
| 2252 |
+
"grad_norm": 14.745248794555664,
|
| 2253 |
+
"learning_rate": 1.7065385807251316e-05,
|
| 2254 |
+
"loss": 0.3375,
|
| 2255 |
+
"step": 15750
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 4.896188410288193,
|
| 2259 |
+
"grad_norm": 11.518832206726074,
|
| 2260 |
+
"learning_rate": 1.7013738250180766e-05,
|
| 2261 |
+
"loss": 0.3371,
|
| 2262 |
+
"step": 15800
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 4.911682677409359,
|
| 2266 |
+
"grad_norm": 17.58115005493164,
|
| 2267 |
+
"learning_rate": 1.6962090693110216e-05,
|
| 2268 |
+
"loss": 0.3851,
|
| 2269 |
+
"step": 15850
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.927176944530523,
|
| 2273 |
+
"grad_norm": 16.769134521484375,
|
| 2274 |
+
"learning_rate": 1.6910443136039665e-05,
|
| 2275 |
+
"loss": 0.3009,
|
| 2276 |
+
"step": 15900
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 4.942671211651689,
|
| 2280 |
+
"grad_norm": 21.518749237060547,
|
| 2281 |
+
"learning_rate": 1.6858795578969115e-05,
|
| 2282 |
+
"loss": 0.3155,
|
| 2283 |
+
"step": 15950
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 4.958165478772854,
|
| 2287 |
+
"grad_norm": 11.044340133666992,
|
| 2288 |
+
"learning_rate": 1.6807148021898565e-05,
|
| 2289 |
+
"loss": 0.3648,
|
| 2290 |
+
"step": 16000
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 4.973659745894019,
|
| 2294 |
+
"grad_norm": 3.9900588989257812,
|
| 2295 |
+
"learning_rate": 1.6755500464828014e-05,
|
| 2296 |
+
"loss": 0.3383,
|
| 2297 |
+
"step": 16050
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 4.989154013015185,
|
| 2301 |
+
"grad_norm": 16.869041442871094,
|
| 2302 |
+
"learning_rate": 1.6703852907757464e-05,
|
| 2303 |
+
"loss": 0.3471,
|
| 2304 |
+
"step": 16100
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 5.0,
|
| 2308 |
+
"eval_accuracy": 0.7964162023791598,
|
| 2309 |
+
"eval_f1": 0.7960859271865419,
|
| 2310 |
+
"eval_loss": 0.8064730167388916,
|
| 2311 |
+
"eval_runtime": 25.5199,
|
| 2312 |
+
"eval_samples_per_second": 260.228,
|
| 2313 |
+
"eval_steps_per_second": 16.301,
|
| 2314 |
+
"step": 16135
|
| 2315 |
+
}
|
| 2316 |
+
],
|
| 2317 |
+
"logging_steps": 50,
|
| 2318 |
+
"max_steps": 32270,
|
| 2319 |
+
"num_input_tokens_seen": 0,
|
| 2320 |
+
"num_train_epochs": 10,
|
| 2321 |
+
"save_steps": 500,
|
| 2322 |
+
"stateful_callbacks": {
|
| 2323 |
+
"EarlyStoppingCallback": {
|
| 2324 |
+
"args": {
|
| 2325 |
+
"early_stopping_patience": 2,
|
| 2326 |
+
"early_stopping_threshold": 0.0
|
| 2327 |
+
},
|
| 2328 |
+
"attributes": {
|
| 2329 |
+
"early_stopping_patience_counter": 0
|
| 2330 |
+
}
|
| 2331 |
+
},
|
| 2332 |
+
"TrainerControl": {
|
| 2333 |
+
"args": {
|
| 2334 |
+
"should_epoch_stop": false,
|
| 2335 |
+
"should_evaluate": false,
|
| 2336 |
+
"should_log": false,
|
| 2337 |
+
"should_save": true,
|
| 2338 |
+
"should_training_stop": false
|
| 2339 |
+
},
|
| 2340 |
+
"attributes": {}
|
| 2341 |
+
}
|
| 2342 |
+
},
|
| 2343 |
+
"total_flos": 3.39622791611904e+16,
|
| 2344 |
+
"train_batch_size": 16,
|
| 2345 |
+
"trial_name": null,
|
| 2346 |
+
"trial_params": null
|
| 2347 |
+
}
|
checkpoint-16135/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10df86f20d44dbad5bcbd0936a460173513fcbc2b606339a5e32ef490cc53c86
|
| 3 |
+
size 5216
|
checkpoint-22589/config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ElectraForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"embedding_size": 768,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "\uae30\uc068",
|
| 13 |
+
"1": "\ub2f9\ud669",
|
| 14 |
+
"2": "\ubd84\ub178",
|
| 15 |
+
"3": "\ubd88\uc548",
|
| 16 |
+
"4": "\uc0c1\ucc98",
|
| 17 |
+
"5": "\uc2ac\ud514"
|
| 18 |
+
},
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 3072,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"\uae30\uc068": 0,
|
| 23 |
+
"\ub2f9\ud669": 1,
|
| 24 |
+
"\ubd84\ub178": 2,
|
| 25 |
+
"\ubd88\uc548": 3,
|
| 26 |
+
"\uc0c1\ucc98": 4,
|
| 27 |
+
"\uc2ac\ud514": 5
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-12,
|
| 30 |
+
"max_position_embeddings": 512,
|
| 31 |
+
"model_type": "electra",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"position_embedding_type": "absolute",
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"summary_activation": "gelu",
|
| 38 |
+
"summary_last_dropout": 0.1,
|
| 39 |
+
"summary_type": "first",
|
| 40 |
+
"summary_use_proj": true,
|
| 41 |
+
"tokenizer_class": "BertTokenizer",
|
| 42 |
+
"torch_dtype": "float32",
|
| 43 |
+
"transformers_version": "4.52.2",
|
| 44 |
+
"type_vocab_size": 2,
|
| 45 |
+
"use_cache": true,
|
| 46 |
+
"vocab_size": 54343
|
| 47 |
+
}
|
checkpoint-22589/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bc7d0b1c232326da9579b888927ef52d71de1e5d37d09bfa1537cf9037d89bae
|
| 3 |
+
size 511149672
|
checkpoint-22589/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:256ef19d7fa444f94d2250f96a4e1eba05fb3af3461f6f03c0df8847276466fa
|
| 3 |
+
size 1022417532
|
checkpoint-22589/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78dab76e293cee2463acfd238eca1fc2a3f094fac6f39dea09231aa019a64dbe
|
| 3 |
+
size 14168
|
checkpoint-22589/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d6e678691ccf4ea8c4b3a9194d4b565a345b929928cce1da7d65cae2362dc0c
|
| 3 |
+
size 1056
|
checkpoint-22589/trainer_state.json
ADDED
|
@@ -0,0 +1,3270 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 16135,
|
| 3 |
+
"best_metric": 0.7960859271865419,
|
| 4 |
+
"best_model_checkpoint": "/content/drive/MyDrive/\uac10\uc815\ubd84\ub958/data/emotion_model/checkpoint-16135",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 22589,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.015494267121165169,
|
| 14 |
+
"grad_norm": 1.9321871995925903,
|
| 15 |
+
"learning_rate": 4.5553145336225596e-07,
|
| 16 |
+
"loss": 1.7919,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.030988534242330338,
|
| 21 |
+
"grad_norm": 1.7712626457214355,
|
| 22 |
+
"learning_rate": 9.203594669972111e-07,
|
| 23 |
+
"loss": 1.7865,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04648280136349551,
|
| 28 |
+
"grad_norm": 1.9185744524002075,
|
| 29 |
+
"learning_rate": 1.385187480632166e-06,
|
| 30 |
+
"loss": 1.7864,
|
| 31 |
+
"step": 150
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.061977068484660676,
|
| 35 |
+
"grad_norm": 1.9046003818511963,
|
| 36 |
+
"learning_rate": 1.8500154942671213e-06,
|
| 37 |
+
"loss": 1.7864,
|
| 38 |
+
"step": 200
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07747133560582585,
|
| 42 |
+
"grad_norm": 2.0914034843444824,
|
| 43 |
+
"learning_rate": 2.3148435079020763e-06,
|
| 44 |
+
"loss": 1.7813,
|
| 45 |
+
"step": 250
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.09296560272699102,
|
| 49 |
+
"grad_norm": 2.088219165802002,
|
| 50 |
+
"learning_rate": 2.7796715215370313e-06,
|
| 51 |
+
"loss": 1.7773,
|
| 52 |
+
"step": 300
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.10845986984815618,
|
| 56 |
+
"grad_norm": 2.1377577781677246,
|
| 57 |
+
"learning_rate": 3.2444995351719864e-06,
|
| 58 |
+
"loss": 1.7587,
|
| 59 |
+
"step": 350
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.12395413696932135,
|
| 63 |
+
"grad_norm": 2.2140750885009766,
|
| 64 |
+
"learning_rate": 3.7093275488069414e-06,
|
| 65 |
+
"loss": 1.7388,
|
| 66 |
+
"step": 400
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.13944840409048653,
|
| 70 |
+
"grad_norm": 2.1278295516967773,
|
| 71 |
+
"learning_rate": 4.174155562441896e-06,
|
| 72 |
+
"loss": 1.6861,
|
| 73 |
+
"step": 450
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1549426712116517,
|
| 77 |
+
"grad_norm": 4.734658241271973,
|
| 78 |
+
"learning_rate": 4.638983576076852e-06,
|
| 79 |
+
"loss": 1.6267,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.17043693833281687,
|
| 84 |
+
"grad_norm": 4.140384197235107,
|
| 85 |
+
"learning_rate": 5.103811589711806e-06,
|
| 86 |
+
"loss": 1.5732,
|
| 87 |
+
"step": 550
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.18593120545398203,
|
| 91 |
+
"grad_norm": 2.705493688583374,
|
| 92 |
+
"learning_rate": 5.568639603346762e-06,
|
| 93 |
+
"loss": 1.5438,
|
| 94 |
+
"step": 600
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2014254725751472,
|
| 98 |
+
"grad_norm": 5.3791422843933105,
|
| 99 |
+
"learning_rate": 6.0334676169817164e-06,
|
| 100 |
+
"loss": 1.4644,
|
| 101 |
+
"step": 650
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.21691973969631237,
|
| 105 |
+
"grad_norm": 2.8989250659942627,
|
| 106 |
+
"learning_rate": 6.498295630616672e-06,
|
| 107 |
+
"loss": 1.4743,
|
| 108 |
+
"step": 700
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.23241400681747754,
|
| 112 |
+
"grad_norm": 3.40291166305542,
|
| 113 |
+
"learning_rate": 6.963123644251627e-06,
|
| 114 |
+
"loss": 1.4187,
|
| 115 |
+
"step": 750
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.2479082739386427,
|
| 119 |
+
"grad_norm": 5.748068809509277,
|
| 120 |
+
"learning_rate": 7.427951657886583e-06,
|
| 121 |
+
"loss": 1.3533,
|
| 122 |
+
"step": 800
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.26340254105980787,
|
| 126 |
+
"grad_norm": 5.777422904968262,
|
| 127 |
+
"learning_rate": 7.892779671521537e-06,
|
| 128 |
+
"loss": 1.3099,
|
| 129 |
+
"step": 850
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.27889680818097307,
|
| 133 |
+
"grad_norm": 4.8046112060546875,
|
| 134 |
+
"learning_rate": 8.357607685156493e-06,
|
| 135 |
+
"loss": 1.2723,
|
| 136 |
+
"step": 900
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.2943910753021382,
|
| 140 |
+
"grad_norm": 5.549858570098877,
|
| 141 |
+
"learning_rate": 8.822435698791447e-06,
|
| 142 |
+
"loss": 1.2325,
|
| 143 |
+
"step": 950
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3098853424233034,
|
| 147 |
+
"grad_norm": 6.851742744445801,
|
| 148 |
+
"learning_rate": 9.287263712426402e-06,
|
| 149 |
+
"loss": 1.1884,
|
| 150 |
+
"step": 1000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.32537960954446854,
|
| 154 |
+
"grad_norm": 5.2497735023498535,
|
| 155 |
+
"learning_rate": 9.752091726061357e-06,
|
| 156 |
+
"loss": 1.2224,
|
| 157 |
+
"step": 1050
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.34087387666563373,
|
| 161 |
+
"grad_norm": 7.023674488067627,
|
| 162 |
+
"learning_rate": 1.0216919739696313e-05,
|
| 163 |
+
"loss": 1.177,
|
| 164 |
+
"step": 1100
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.3563681437867989,
|
| 168 |
+
"grad_norm": 4.888996124267578,
|
| 169 |
+
"learning_rate": 1.0681747753331269e-05,
|
| 170 |
+
"loss": 1.1577,
|
| 171 |
+
"step": 1150
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.37186241090796407,
|
| 175 |
+
"grad_norm": 6.2133660316467285,
|
| 176 |
+
"learning_rate": 1.1146575766966222e-05,
|
| 177 |
+
"loss": 1.1726,
|
| 178 |
+
"step": 1200
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.3873566780291292,
|
| 182 |
+
"grad_norm": 6.936697483062744,
|
| 183 |
+
"learning_rate": 1.1611403780601178e-05,
|
| 184 |
+
"loss": 1.1001,
|
| 185 |
+
"step": 1250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.4028509451502944,
|
| 189 |
+
"grad_norm": 8.526293754577637,
|
| 190 |
+
"learning_rate": 1.2076231794236133e-05,
|
| 191 |
+
"loss": 1.0752,
|
| 192 |
+
"step": 1300
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.41834521227145954,
|
| 196 |
+
"grad_norm": 5.5933756828308105,
|
| 197 |
+
"learning_rate": 1.254105980787109e-05,
|
| 198 |
+
"loss": 1.0809,
|
| 199 |
+
"step": 1350
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.43383947939262474,
|
| 203 |
+
"grad_norm": 6.998812675476074,
|
| 204 |
+
"learning_rate": 1.3005887821506042e-05,
|
| 205 |
+
"loss": 1.0566,
|
| 206 |
+
"step": 1400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.4493337465137899,
|
| 210 |
+
"grad_norm": 7.077617645263672,
|
| 211 |
+
"learning_rate": 1.3470715835140998e-05,
|
| 212 |
+
"loss": 1.0993,
|
| 213 |
+
"step": 1450
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.4648280136349551,
|
| 217 |
+
"grad_norm": 8.715201377868652,
|
| 218 |
+
"learning_rate": 1.3935543848775953e-05,
|
| 219 |
+
"loss": 1.0375,
|
| 220 |
+
"step": 1500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4803222807561202,
|
| 224 |
+
"grad_norm": 6.017217636108398,
|
| 225 |
+
"learning_rate": 1.440037186241091e-05,
|
| 226 |
+
"loss": 1.0388,
|
| 227 |
+
"step": 1550
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.4958165478772854,
|
| 231 |
+
"grad_norm": 5.349973201751709,
|
| 232 |
+
"learning_rate": 1.4865199876045862e-05,
|
| 233 |
+
"loss": 1.0354,
|
| 234 |
+
"step": 1600
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5113108149984505,
|
| 238 |
+
"grad_norm": 12.728338241577148,
|
| 239 |
+
"learning_rate": 1.533002788968082e-05,
|
| 240 |
+
"loss": 1.0314,
|
| 241 |
+
"step": 1650
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.5268050821196157,
|
| 245 |
+
"grad_norm": 5.962468147277832,
|
| 246 |
+
"learning_rate": 1.5794855903315773e-05,
|
| 247 |
+
"loss": 1.03,
|
| 248 |
+
"step": 1700
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.5422993492407809,
|
| 252 |
+
"grad_norm": 5.971400260925293,
|
| 253 |
+
"learning_rate": 1.6259683916950726e-05,
|
| 254 |
+
"loss": 1.0541,
|
| 255 |
+
"step": 1750
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.5577936163619461,
|
| 259 |
+
"grad_norm": 6.260463714599609,
|
| 260 |
+
"learning_rate": 1.6724511930585682e-05,
|
| 261 |
+
"loss": 0.9831,
|
| 262 |
+
"step": 1800
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.5732878834831112,
|
| 266 |
+
"grad_norm": 7.8115010261535645,
|
| 267 |
+
"learning_rate": 1.718933994422064e-05,
|
| 268 |
+
"loss": 0.966,
|
| 269 |
+
"step": 1850
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.5887821506042764,
|
| 273 |
+
"grad_norm": 5.005403995513916,
|
| 274 |
+
"learning_rate": 1.7654167957855595e-05,
|
| 275 |
+
"loss": 0.9592,
|
| 276 |
+
"step": 1900
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.6042764177254416,
|
| 280 |
+
"grad_norm": 7.7732157707214355,
|
| 281 |
+
"learning_rate": 1.8118995971490548e-05,
|
| 282 |
+
"loss": 0.9766,
|
| 283 |
+
"step": 1950
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.6197706848466068,
|
| 287 |
+
"grad_norm": 7.265392303466797,
|
| 288 |
+
"learning_rate": 1.8583823985125504e-05,
|
| 289 |
+
"loss": 1.0171,
|
| 290 |
+
"step": 2000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.6352649519677719,
|
| 294 |
+
"grad_norm": 15.946109771728516,
|
| 295 |
+
"learning_rate": 1.904865199876046e-05,
|
| 296 |
+
"loss": 0.9824,
|
| 297 |
+
"step": 2050
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.6507592190889371,
|
| 301 |
+
"grad_norm": 7.261445999145508,
|
| 302 |
+
"learning_rate": 1.9513480012395417e-05,
|
| 303 |
+
"loss": 0.9699,
|
| 304 |
+
"step": 2100
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.6662534862101023,
|
| 308 |
+
"grad_norm": 8.201744079589844,
|
| 309 |
+
"learning_rate": 1.997830802603037e-05,
|
| 310 |
+
"loss": 0.9957,
|
| 311 |
+
"step": 2150
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.6817477533312675,
|
| 315 |
+
"grad_norm": 6.183067798614502,
|
| 316 |
+
"learning_rate": 2.0443136039665322e-05,
|
| 317 |
+
"loss": 0.8554,
|
| 318 |
+
"step": 2200
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.6972420204524326,
|
| 322 |
+
"grad_norm": 7.481590270996094,
|
| 323 |
+
"learning_rate": 2.090796405330028e-05,
|
| 324 |
+
"loss": 0.9929,
|
| 325 |
+
"step": 2250
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.7127362875735977,
|
| 329 |
+
"grad_norm": 7.3274030685424805,
|
| 330 |
+
"learning_rate": 2.1372792066935235e-05,
|
| 331 |
+
"loss": 0.9438,
|
| 332 |
+
"step": 2300
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.7282305546947629,
|
| 336 |
+
"grad_norm": 11.69247055053711,
|
| 337 |
+
"learning_rate": 2.183762008057019e-05,
|
| 338 |
+
"loss": 0.9767,
|
| 339 |
+
"step": 2350
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.7437248218159281,
|
| 343 |
+
"grad_norm": 7.929721832275391,
|
| 344 |
+
"learning_rate": 2.2302448094205144e-05,
|
| 345 |
+
"loss": 1.0036,
|
| 346 |
+
"step": 2400
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.7592190889370932,
|
| 350 |
+
"grad_norm": 9.753717422485352,
|
| 351 |
+
"learning_rate": 2.27672761078401e-05,
|
| 352 |
+
"loss": 0.9726,
|
| 353 |
+
"step": 2450
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.7747133560582584,
|
| 357 |
+
"grad_norm": 7.797086715698242,
|
| 358 |
+
"learning_rate": 2.3232104121475057e-05,
|
| 359 |
+
"loss": 0.9322,
|
| 360 |
+
"step": 2500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.7902076231794236,
|
| 364 |
+
"grad_norm": 6.927332878112793,
|
| 365 |
+
"learning_rate": 2.369693213511001e-05,
|
| 366 |
+
"loss": 0.9378,
|
| 367 |
+
"step": 2550
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.8057018903005888,
|
| 371 |
+
"grad_norm": 3.726092576980591,
|
| 372 |
+
"learning_rate": 2.4161760148744962e-05,
|
| 373 |
+
"loss": 0.958,
|
| 374 |
+
"step": 2600
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.821196157421754,
|
| 378 |
+
"grad_norm": 5.661774635314941,
|
| 379 |
+
"learning_rate": 2.462658816237992e-05,
|
| 380 |
+
"loss": 0.9651,
|
| 381 |
+
"step": 2650
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.8366904245429191,
|
| 385 |
+
"grad_norm": 6.513345718383789,
|
| 386 |
+
"learning_rate": 2.5091416176014875e-05,
|
| 387 |
+
"loss": 0.971,
|
| 388 |
+
"step": 2700
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.8521846916640843,
|
| 392 |
+
"grad_norm": 6.713255405426025,
|
| 393 |
+
"learning_rate": 2.555624418964983e-05,
|
| 394 |
+
"loss": 0.8616,
|
| 395 |
+
"step": 2750
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.8676789587852495,
|
| 399 |
+
"grad_norm": 8.527266502380371,
|
| 400 |
+
"learning_rate": 2.6021072203284784e-05,
|
| 401 |
+
"loss": 0.9413,
|
| 402 |
+
"step": 2800
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.8831732259064147,
|
| 406 |
+
"grad_norm": 6.599502086639404,
|
| 407 |
+
"learning_rate": 2.648590021691974e-05,
|
| 408 |
+
"loss": 0.9985,
|
| 409 |
+
"step": 2850
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.8986674930275798,
|
| 413 |
+
"grad_norm": 4.0680155754089355,
|
| 414 |
+
"learning_rate": 2.6950728230554697e-05,
|
| 415 |
+
"loss": 0.9415,
|
| 416 |
+
"step": 2900
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.914161760148745,
|
| 420 |
+
"grad_norm": 5.083493232727051,
|
| 421 |
+
"learning_rate": 2.741555624418965e-05,
|
| 422 |
+
"loss": 0.9805,
|
| 423 |
+
"step": 2950
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.9296560272699101,
|
| 427 |
+
"grad_norm": 4.0469069480896,
|
| 428 |
+
"learning_rate": 2.7880384257824606e-05,
|
| 429 |
+
"loss": 0.9547,
|
| 430 |
+
"step": 3000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.9451502943910753,
|
| 434 |
+
"grad_norm": 6.075752258300781,
|
| 435 |
+
"learning_rate": 2.834521227145956e-05,
|
| 436 |
+
"loss": 0.9623,
|
| 437 |
+
"step": 3050
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.9606445615122404,
|
| 441 |
+
"grad_norm": 6.5252299308776855,
|
| 442 |
+
"learning_rate": 2.8810040285094515e-05,
|
| 443 |
+
"loss": 0.9838,
|
| 444 |
+
"step": 3100
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.9761388286334056,
|
| 448 |
+
"grad_norm": 6.530562877655029,
|
| 449 |
+
"learning_rate": 2.927486829872947e-05,
|
| 450 |
+
"loss": 0.911,
|
| 451 |
+
"step": 3150
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.9916330957545708,
|
| 455 |
+
"grad_norm": 8.217161178588867,
|
| 456 |
+
"learning_rate": 2.9739696312364428e-05,
|
| 457 |
+
"loss": 0.9457,
|
| 458 |
+
"step": 3200
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.0,
|
| 462 |
+
"eval_accuracy": 0.7087787983737389,
|
| 463 |
+
"eval_f1": 0.7069046924545164,
|
| 464 |
+
"eval_loss": 0.8573769330978394,
|
| 465 |
+
"eval_runtime": 25.5149,
|
| 466 |
+
"eval_samples_per_second": 260.279,
|
| 467 |
+
"eval_steps_per_second": 16.304,
|
| 468 |
+
"step": 3227
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.007127362875736,
|
| 472 |
+
"grad_norm": 4.367455005645752,
|
| 473 |
+
"learning_rate": 2.997727507488896e-05,
|
| 474 |
+
"loss": 0.9143,
|
| 475 |
+
"step": 3250
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.022621629996901,
|
| 479 |
+
"grad_norm": 4.636725902557373,
|
| 480 |
+
"learning_rate": 2.9925627517818407e-05,
|
| 481 |
+
"loss": 0.9304,
|
| 482 |
+
"step": 3300
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.0381158971180664,
|
| 486 |
+
"grad_norm": 4.251437664031982,
|
| 487 |
+
"learning_rate": 2.9873979960747857e-05,
|
| 488 |
+
"loss": 0.8549,
|
| 489 |
+
"step": 3350
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.0536101642392315,
|
| 493 |
+
"grad_norm": 6.648655414581299,
|
| 494 |
+
"learning_rate": 2.9822332403677307e-05,
|
| 495 |
+
"loss": 0.8698,
|
| 496 |
+
"step": 3400
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.0691044313603966,
|
| 500 |
+
"grad_norm": 7.102205276489258,
|
| 501 |
+
"learning_rate": 2.9770684846606757e-05,
|
| 502 |
+
"loss": 0.8597,
|
| 503 |
+
"step": 3450
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.0845986984815619,
|
| 507 |
+
"grad_norm": 10.821270942687988,
|
| 508 |
+
"learning_rate": 2.9719037289536206e-05,
|
| 509 |
+
"loss": 0.8457,
|
| 510 |
+
"step": 3500
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.100092965602727,
|
| 514 |
+
"grad_norm": 6.111588001251221,
|
| 515 |
+
"learning_rate": 2.9667389732465652e-05,
|
| 516 |
+
"loss": 0.8973,
|
| 517 |
+
"step": 3550
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.1155872327238923,
|
| 521 |
+
"grad_norm": 9.016953468322754,
|
| 522 |
+
"learning_rate": 2.9615742175395106e-05,
|
| 523 |
+
"loss": 0.8228,
|
| 524 |
+
"step": 3600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.1310814998450573,
|
| 528 |
+
"grad_norm": 7.717069625854492,
|
| 529 |
+
"learning_rate": 2.9564094618324555e-05,
|
| 530 |
+
"loss": 0.8103,
|
| 531 |
+
"step": 3650
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.1465757669662224,
|
| 535 |
+
"grad_norm": 7.848579406738281,
|
| 536 |
+
"learning_rate": 2.9512447061254005e-05,
|
| 537 |
+
"loss": 0.8097,
|
| 538 |
+
"step": 3700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.1620700340873877,
|
| 542 |
+
"grad_norm": 4.738124847412109,
|
| 543 |
+
"learning_rate": 2.9460799504183455e-05,
|
| 544 |
+
"loss": 0.9105,
|
| 545 |
+
"step": 3750
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.1775643012085528,
|
| 549 |
+
"grad_norm": 5.289875507354736,
|
| 550 |
+
"learning_rate": 2.94091519471129e-05,
|
| 551 |
+
"loss": 0.8876,
|
| 552 |
+
"step": 3800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1930585683297181,
|
| 556 |
+
"grad_norm": 6.445308685302734,
|
| 557 |
+
"learning_rate": 2.935750439004235e-05,
|
| 558 |
+
"loss": 0.8377,
|
| 559 |
+
"step": 3850
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.2085528354508832,
|
| 563 |
+
"grad_norm": 4.725327968597412,
|
| 564 |
+
"learning_rate": 2.93058568329718e-05,
|
| 565 |
+
"loss": 0.9332,
|
| 566 |
+
"step": 3900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.2240471025720483,
|
| 570 |
+
"grad_norm": 53.85081481933594,
|
| 571 |
+
"learning_rate": 2.925420927590125e-05,
|
| 572 |
+
"loss": 0.8882,
|
| 573 |
+
"step": 3950
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.2395413696932136,
|
| 577 |
+
"grad_norm": 5.677978515625,
|
| 578 |
+
"learning_rate": 2.9202561718830703e-05,
|
| 579 |
+
"loss": 0.8481,
|
| 580 |
+
"step": 4000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.2550356368143787,
|
| 584 |
+
"grad_norm": 3.941765785217285,
|
| 585 |
+
"learning_rate": 2.915091416176015e-05,
|
| 586 |
+
"loss": 0.835,
|
| 587 |
+
"step": 4050
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.2705299039355438,
|
| 591 |
+
"grad_norm": 8.099725723266602,
|
| 592 |
+
"learning_rate": 2.90992666046896e-05,
|
| 593 |
+
"loss": 0.8322,
|
| 594 |
+
"step": 4100
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.286024171056709,
|
| 598 |
+
"grad_norm": 6.59591007232666,
|
| 599 |
+
"learning_rate": 2.904761904761905e-05,
|
| 600 |
+
"loss": 0.8809,
|
| 601 |
+
"step": 4150
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.3015184381778742,
|
| 605 |
+
"grad_norm": 7.200226306915283,
|
| 606 |
+
"learning_rate": 2.8995971490548498e-05,
|
| 607 |
+
"loss": 0.8609,
|
| 608 |
+
"step": 4200
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.3170127052990392,
|
| 612 |
+
"grad_norm": 4.902937412261963,
|
| 613 |
+
"learning_rate": 2.8944323933477948e-05,
|
| 614 |
+
"loss": 0.8633,
|
| 615 |
+
"step": 4250
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.3325069724202045,
|
| 619 |
+
"grad_norm": 5.792146682739258,
|
| 620 |
+
"learning_rate": 2.8892676376407394e-05,
|
| 621 |
+
"loss": 0.8684,
|
| 622 |
+
"step": 4300
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.3480012395413696,
|
| 626 |
+
"grad_norm": 4.636809349060059,
|
| 627 |
+
"learning_rate": 2.8841028819336844e-05,
|
| 628 |
+
"loss": 0.8245,
|
| 629 |
+
"step": 4350
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.363495506662535,
|
| 633 |
+
"grad_norm": 5.28842306137085,
|
| 634 |
+
"learning_rate": 2.8789381262266297e-05,
|
| 635 |
+
"loss": 0.7859,
|
| 636 |
+
"step": 4400
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.3789897737837,
|
| 640 |
+
"grad_norm": 4.259128570556641,
|
| 641 |
+
"learning_rate": 2.8737733705195747e-05,
|
| 642 |
+
"loss": 0.8228,
|
| 643 |
+
"step": 4450
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.394484040904865,
|
| 647 |
+
"grad_norm": 7.914375305175781,
|
| 648 |
+
"learning_rate": 2.8686086148125196e-05,
|
| 649 |
+
"loss": 0.9448,
|
| 650 |
+
"step": 4500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.4099783080260304,
|
| 654 |
+
"grad_norm": 7.636547088623047,
|
| 655 |
+
"learning_rate": 2.8634438591054643e-05,
|
| 656 |
+
"loss": 0.8781,
|
| 657 |
+
"step": 4550
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.4254725751471955,
|
| 661 |
+
"grad_norm": 8.681707382202148,
|
| 662 |
+
"learning_rate": 2.8582791033984092e-05,
|
| 663 |
+
"loss": 0.8367,
|
| 664 |
+
"step": 4600
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.4409668422683608,
|
| 668 |
+
"grad_norm": 7.864759922027588,
|
| 669 |
+
"learning_rate": 2.8531143476913542e-05,
|
| 670 |
+
"loss": 0.9088,
|
| 671 |
+
"step": 4650
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.4564611093895259,
|
| 675 |
+
"grad_norm": 4.892348289489746,
|
| 676 |
+
"learning_rate": 2.8479495919842992e-05,
|
| 677 |
+
"loss": 0.8993,
|
| 678 |
+
"step": 4700
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.471955376510691,
|
| 682 |
+
"grad_norm": 23.208873748779297,
|
| 683 |
+
"learning_rate": 2.842784836277244e-05,
|
| 684 |
+
"loss": 0.8542,
|
| 685 |
+
"step": 4750
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.4874496436318563,
|
| 689 |
+
"grad_norm": 8.983469009399414,
|
| 690 |
+
"learning_rate": 2.837620080570189e-05,
|
| 691 |
+
"loss": 0.949,
|
| 692 |
+
"step": 4800
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.5029439107530214,
|
| 696 |
+
"grad_norm": 10.706644058227539,
|
| 697 |
+
"learning_rate": 2.832455324863134e-05,
|
| 698 |
+
"loss": 0.9008,
|
| 699 |
+
"step": 4850
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.5184381778741867,
|
| 703 |
+
"grad_norm": 4.685935020446777,
|
| 704 |
+
"learning_rate": 2.827290569156079e-05,
|
| 705 |
+
"loss": 0.8613,
|
| 706 |
+
"step": 4900
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.5339324449953518,
|
| 710 |
+
"grad_norm": 5.286406993865967,
|
| 711 |
+
"learning_rate": 2.822125813449024e-05,
|
| 712 |
+
"loss": 0.8929,
|
| 713 |
+
"step": 4950
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.5494267121165168,
|
| 717 |
+
"grad_norm": 4.907707691192627,
|
| 718 |
+
"learning_rate": 2.816961057741969e-05,
|
| 719 |
+
"loss": 0.8321,
|
| 720 |
+
"step": 5000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.564920979237682,
|
| 724 |
+
"grad_norm": 6.398087501525879,
|
| 725 |
+
"learning_rate": 2.8117963020349136e-05,
|
| 726 |
+
"loss": 0.8626,
|
| 727 |
+
"step": 5050
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.5804152463588472,
|
| 731 |
+
"grad_norm": 5.323617458343506,
|
| 732 |
+
"learning_rate": 2.8066315463278586e-05,
|
| 733 |
+
"loss": 0.8324,
|
| 734 |
+
"step": 5100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.5959095134800125,
|
| 738 |
+
"grad_norm": 4.136271953582764,
|
| 739 |
+
"learning_rate": 2.8014667906208035e-05,
|
| 740 |
+
"loss": 0.879,
|
| 741 |
+
"step": 5150
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.6114037806011776,
|
| 745 |
+
"grad_norm": 6.873619556427002,
|
| 746 |
+
"learning_rate": 2.796302034913749e-05,
|
| 747 |
+
"loss": 0.9043,
|
| 748 |
+
"step": 5200
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.6268980477223427,
|
| 752 |
+
"grad_norm": 7.138693809509277,
|
| 753 |
+
"learning_rate": 2.7911372792066938e-05,
|
| 754 |
+
"loss": 0.8183,
|
| 755 |
+
"step": 5250
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.6423923148435078,
|
| 759 |
+
"grad_norm": 6.483767032623291,
|
| 760 |
+
"learning_rate": 2.7859725234996384e-05,
|
| 761 |
+
"loss": 0.8867,
|
| 762 |
+
"step": 5300
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.657886581964673,
|
| 766 |
+
"grad_norm": 3.2249104976654053,
|
| 767 |
+
"learning_rate": 2.7808077677925834e-05,
|
| 768 |
+
"loss": 0.8097,
|
| 769 |
+
"step": 5350
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.6733808490858384,
|
| 773 |
+
"grad_norm": 6.961575984954834,
|
| 774 |
+
"learning_rate": 2.7756430120855284e-05,
|
| 775 |
+
"loss": 0.8364,
|
| 776 |
+
"step": 5400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.6888751162070035,
|
| 780 |
+
"grad_norm": 7.0920000076293945,
|
| 781 |
+
"learning_rate": 2.7704782563784733e-05,
|
| 782 |
+
"loss": 0.8283,
|
| 783 |
+
"step": 5450
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.7043693833281686,
|
| 787 |
+
"grad_norm": 5.436604976654053,
|
| 788 |
+
"learning_rate": 2.7653135006714183e-05,
|
| 789 |
+
"loss": 0.8786,
|
| 790 |
+
"step": 5500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.7198636504493336,
|
| 794 |
+
"grad_norm": 4.0141282081604,
|
| 795 |
+
"learning_rate": 2.760148744964363e-05,
|
| 796 |
+
"loss": 0.8452,
|
| 797 |
+
"step": 5550
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.735357917570499,
|
| 801 |
+
"grad_norm": 5.783074378967285,
|
| 802 |
+
"learning_rate": 2.7549839892573083e-05,
|
| 803 |
+
"loss": 0.8168,
|
| 804 |
+
"step": 5600
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 1.750852184691664,
|
| 808 |
+
"grad_norm": 7.773756504058838,
|
| 809 |
+
"learning_rate": 2.7498192335502532e-05,
|
| 810 |
+
"loss": 0.8598,
|
| 811 |
+
"step": 5650
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.7663464518128293,
|
| 815 |
+
"grad_norm": 5.375339984893799,
|
| 816 |
+
"learning_rate": 2.7446544778431982e-05,
|
| 817 |
+
"loss": 0.8366,
|
| 818 |
+
"step": 5700
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.7818407189339944,
|
| 822 |
+
"grad_norm": 4.240859031677246,
|
| 823 |
+
"learning_rate": 2.739489722136143e-05,
|
| 824 |
+
"loss": 0.8136,
|
| 825 |
+
"step": 5750
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 1.7973349860551595,
|
| 829 |
+
"grad_norm": 6.107599258422852,
|
| 830 |
+
"learning_rate": 2.734324966429088e-05,
|
| 831 |
+
"loss": 0.8074,
|
| 832 |
+
"step": 5800
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.8128292531763246,
|
| 836 |
+
"grad_norm": 6.027589797973633,
|
| 837 |
+
"learning_rate": 2.7291602107220328e-05,
|
| 838 |
+
"loss": 0.7808,
|
| 839 |
+
"step": 5850
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.82832352029749,
|
| 843 |
+
"grad_norm": 4.829204559326172,
|
| 844 |
+
"learning_rate": 2.7239954550149777e-05,
|
| 845 |
+
"loss": 0.8473,
|
| 846 |
+
"step": 5900
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.8438177874186552,
|
| 850 |
+
"grad_norm": 5.385358810424805,
|
| 851 |
+
"learning_rate": 2.7188306993079227e-05,
|
| 852 |
+
"loss": 0.844,
|
| 853 |
+
"step": 5950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.8593120545398203,
|
| 857 |
+
"grad_norm": 5.991063594818115,
|
| 858 |
+
"learning_rate": 2.713665943600868e-05,
|
| 859 |
+
"loss": 0.8667,
|
| 860 |
+
"step": 6000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.8748063216609854,
|
| 864 |
+
"grad_norm": 4.269604682922363,
|
| 865 |
+
"learning_rate": 2.708501187893813e-05,
|
| 866 |
+
"loss": 0.8987,
|
| 867 |
+
"step": 6050
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.8903005887821505,
|
| 871 |
+
"grad_norm": 6.90878438949585,
|
| 872 |
+
"learning_rate": 2.7033364321867576e-05,
|
| 873 |
+
"loss": 0.8517,
|
| 874 |
+
"step": 6100
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.9057948559033158,
|
| 878 |
+
"grad_norm": 8.742233276367188,
|
| 879 |
+
"learning_rate": 2.6981716764797026e-05,
|
| 880 |
+
"loss": 0.8729,
|
| 881 |
+
"step": 6150
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.921289123024481,
|
| 885 |
+
"grad_norm": 9.10084342956543,
|
| 886 |
+
"learning_rate": 2.6930069207726475e-05,
|
| 887 |
+
"loss": 0.8803,
|
| 888 |
+
"step": 6200
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 1.9367833901456462,
|
| 892 |
+
"grad_norm": 4.210537433624268,
|
| 893 |
+
"learning_rate": 2.6878421650655925e-05,
|
| 894 |
+
"loss": 0.7938,
|
| 895 |
+
"step": 6250
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.9522776572668112,
|
| 899 |
+
"grad_norm": 6.604791641235352,
|
| 900 |
+
"learning_rate": 2.6826774093585375e-05,
|
| 901 |
+
"loss": 0.7958,
|
| 902 |
+
"step": 6300
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.9677719243879763,
|
| 906 |
+
"grad_norm": 6.213857173919678,
|
| 907 |
+
"learning_rate": 2.677512653651482e-05,
|
| 908 |
+
"loss": 0.8463,
|
| 909 |
+
"step": 6350
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.9832661915091416,
|
| 913 |
+
"grad_norm": 4.303800582885742,
|
| 914 |
+
"learning_rate": 2.6723478979444274e-05,
|
| 915 |
+
"loss": 0.7909,
|
| 916 |
+
"step": 6400
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.998760458630307,
|
| 920 |
+
"grad_norm": 4.933095932006836,
|
| 921 |
+
"learning_rate": 2.6671831422373724e-05,
|
| 922 |
+
"loss": 0.7888,
|
| 923 |
+
"step": 6450
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 2.0,
|
| 927 |
+
"eval_accuracy": 0.7590724288510766,
|
| 928 |
+
"eval_f1": 0.7587245577707713,
|
| 929 |
+
"eval_loss": 0.7010347247123718,
|
| 930 |
+
"eval_runtime": 25.4199,
|
| 931 |
+
"eval_samples_per_second": 261.252,
|
| 932 |
+
"eval_steps_per_second": 16.365,
|
| 933 |
+
"step": 6454
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.014254725751472,
|
| 937 |
+
"grad_norm": 4.0570244789123535,
|
| 938 |
+
"learning_rate": 2.6620183865303173e-05,
|
| 939 |
+
"loss": 0.7236,
|
| 940 |
+
"step": 6500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.029748992872637,
|
| 944 |
+
"grad_norm": 5.307652473449707,
|
| 945 |
+
"learning_rate": 2.6568536308232623e-05,
|
| 946 |
+
"loss": 0.7213,
|
| 947 |
+
"step": 6550
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.045243259993802,
|
| 951 |
+
"grad_norm": 5.398072719573975,
|
| 952 |
+
"learning_rate": 2.651688875116207e-05,
|
| 953 |
+
"loss": 0.6839,
|
| 954 |
+
"step": 6600
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.0607375271149673,
|
| 958 |
+
"grad_norm": 5.296418190002441,
|
| 959 |
+
"learning_rate": 2.646524119409152e-05,
|
| 960 |
+
"loss": 0.6856,
|
| 961 |
+
"step": 6650
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 2.076231794236133,
|
| 965 |
+
"grad_norm": 4.173377990722656,
|
| 966 |
+
"learning_rate": 2.641359363702097e-05,
|
| 967 |
+
"loss": 0.7109,
|
| 968 |
+
"step": 6700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.091726061357298,
|
| 972 |
+
"grad_norm": 5.590676784515381,
|
| 973 |
+
"learning_rate": 2.636194607995042e-05,
|
| 974 |
+
"loss": 0.6814,
|
| 975 |
+
"step": 6750
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.107220328478463,
|
| 979 |
+
"grad_norm": 8.112780570983887,
|
| 980 |
+
"learning_rate": 2.631029852287987e-05,
|
| 981 |
+
"loss": 0.7302,
|
| 982 |
+
"step": 6800
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 2.122714595599628,
|
| 986 |
+
"grad_norm": 6.514364242553711,
|
| 987 |
+
"learning_rate": 2.6258650965809318e-05,
|
| 988 |
+
"loss": 0.6917,
|
| 989 |
+
"step": 6850
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.138208862720793,
|
| 993 |
+
"grad_norm": 8.156841278076172,
|
| 994 |
+
"learning_rate": 2.6207003408738767e-05,
|
| 995 |
+
"loss": 0.6568,
|
| 996 |
+
"step": 6900
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 2.1537031298419587,
|
| 1000 |
+
"grad_norm": 7.641481876373291,
|
| 1001 |
+
"learning_rate": 2.6155355851668217e-05,
|
| 1002 |
+
"loss": 0.6132,
|
| 1003 |
+
"step": 6950
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 2.1691973969631237,
|
| 1007 |
+
"grad_norm": 6.33613395690918,
|
| 1008 |
+
"learning_rate": 2.6103708294597667e-05,
|
| 1009 |
+
"loss": 0.6393,
|
| 1010 |
+
"step": 7000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.184691664084289,
|
| 1014 |
+
"grad_norm": 4.2916436195373535,
|
| 1015 |
+
"learning_rate": 2.6052060737527116e-05,
|
| 1016 |
+
"loss": 0.6709,
|
| 1017 |
+
"step": 7050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.200185931205454,
|
| 1021 |
+
"grad_norm": 4.763488292694092,
|
| 1022 |
+
"learning_rate": 2.6000413180456563e-05,
|
| 1023 |
+
"loss": 0.6919,
|
| 1024 |
+
"step": 7100
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 2.215680198326619,
|
| 1028 |
+
"grad_norm": 8.614394187927246,
|
| 1029 |
+
"learning_rate": 2.5948765623386012e-05,
|
| 1030 |
+
"loss": 0.6501,
|
| 1031 |
+
"step": 7150
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 2.2311744654477845,
|
| 1035 |
+
"grad_norm": 9.684426307678223,
|
| 1036 |
+
"learning_rate": 2.5897118066315465e-05,
|
| 1037 |
+
"loss": 0.6947,
|
| 1038 |
+
"step": 7200
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 2.2466687325689496,
|
| 1042 |
+
"grad_norm": 6.210818767547607,
|
| 1043 |
+
"learning_rate": 2.5845470509244915e-05,
|
| 1044 |
+
"loss": 0.6873,
|
| 1045 |
+
"step": 7250
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.2621629996901147,
|
| 1049 |
+
"grad_norm": 6.774372577667236,
|
| 1050 |
+
"learning_rate": 2.5793822952174365e-05,
|
| 1051 |
+
"loss": 0.7195,
|
| 1052 |
+
"step": 7300
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.27765726681128,
|
| 1056 |
+
"grad_norm": 6.014688491821289,
|
| 1057 |
+
"learning_rate": 2.574217539510381e-05,
|
| 1058 |
+
"loss": 0.6298,
|
| 1059 |
+
"step": 7350
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.293151533932445,
|
| 1063 |
+
"grad_norm": 14.994784355163574,
|
| 1064 |
+
"learning_rate": 2.569052783803326e-05,
|
| 1065 |
+
"loss": 0.7403,
|
| 1066 |
+
"step": 7400
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.3086458010536104,
|
| 1070 |
+
"grad_norm": 6.315488815307617,
|
| 1071 |
+
"learning_rate": 2.563888028096271e-05,
|
| 1072 |
+
"loss": 0.6679,
|
| 1073 |
+
"step": 7450
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.3241400681747755,
|
| 1077 |
+
"grad_norm": 8.482314109802246,
|
| 1078 |
+
"learning_rate": 2.558723272389216e-05,
|
| 1079 |
+
"loss": 0.7173,
|
| 1080 |
+
"step": 7500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.3396343352959406,
|
| 1084 |
+
"grad_norm": 10.161298751831055,
|
| 1085 |
+
"learning_rate": 2.553558516682161e-05,
|
| 1086 |
+
"loss": 0.732,
|
| 1087 |
+
"step": 7550
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.3551286024171056,
|
| 1091 |
+
"grad_norm": 6.758267402648926,
|
| 1092 |
+
"learning_rate": 2.548393760975106e-05,
|
| 1093 |
+
"loss": 0.6192,
|
| 1094 |
+
"step": 7600
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.3706228695382707,
|
| 1098 |
+
"grad_norm": 4.528532981872559,
|
| 1099 |
+
"learning_rate": 2.543229005268051e-05,
|
| 1100 |
+
"loss": 0.7614,
|
| 1101 |
+
"step": 7650
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.3861171366594363,
|
| 1105 |
+
"grad_norm": 6.397975921630859,
|
| 1106 |
+
"learning_rate": 2.538064249560996e-05,
|
| 1107 |
+
"loss": 0.6951,
|
| 1108 |
+
"step": 7700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.4016114037806013,
|
| 1112 |
+
"grad_norm": 5.440258979797363,
|
| 1113 |
+
"learning_rate": 2.532899493853941e-05,
|
| 1114 |
+
"loss": 0.7188,
|
| 1115 |
+
"step": 7750
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.4171056709017664,
|
| 1119 |
+
"grad_norm": 2.4531173706054688,
|
| 1120 |
+
"learning_rate": 2.5277347381468858e-05,
|
| 1121 |
+
"loss": 0.6347,
|
| 1122 |
+
"step": 7800
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.4325999380229315,
|
| 1126 |
+
"grad_norm": 15.269991874694824,
|
| 1127 |
+
"learning_rate": 2.5225699824398304e-05,
|
| 1128 |
+
"loss": 0.6601,
|
| 1129 |
+
"step": 7850
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.4480942051440966,
|
| 1133 |
+
"grad_norm": 6.438554286956787,
|
| 1134 |
+
"learning_rate": 2.5174052267327754e-05,
|
| 1135 |
+
"loss": 0.698,
|
| 1136 |
+
"step": 7900
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.4635884722652617,
|
| 1140 |
+
"grad_norm": 8.922213554382324,
|
| 1141 |
+
"learning_rate": 2.5122404710257204e-05,
|
| 1142 |
+
"loss": 0.6958,
|
| 1143 |
+
"step": 7950
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.479082739386427,
|
| 1147 |
+
"grad_norm": 6.724533557891846,
|
| 1148 |
+
"learning_rate": 2.5070757153186657e-05,
|
| 1149 |
+
"loss": 0.7131,
|
| 1150 |
+
"step": 8000
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.4945770065075923,
|
| 1154 |
+
"grad_norm": 5.617169380187988,
|
| 1155 |
+
"learning_rate": 2.5019109596116107e-05,
|
| 1156 |
+
"loss": 0.7711,
|
| 1157 |
+
"step": 8050
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.5100712736287574,
|
| 1161 |
+
"grad_norm": 6.441185474395752,
|
| 1162 |
+
"learning_rate": 2.4967462039045553e-05,
|
| 1163 |
+
"loss": 0.6612,
|
| 1164 |
+
"step": 8100
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 2.5255655407499225,
|
| 1168 |
+
"grad_norm": 6.033916473388672,
|
| 1169 |
+
"learning_rate": 2.4915814481975003e-05,
|
| 1170 |
+
"loss": 0.698,
|
| 1171 |
+
"step": 8150
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 2.5410598078710875,
|
| 1175 |
+
"grad_norm": 6.174665451049805,
|
| 1176 |
+
"learning_rate": 2.4864166924904452e-05,
|
| 1177 |
+
"loss": 0.6968,
|
| 1178 |
+
"step": 8200
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.5565540749922526,
|
| 1182 |
+
"grad_norm": 20.01167869567871,
|
| 1183 |
+
"learning_rate": 2.4812519367833902e-05,
|
| 1184 |
+
"loss": 0.6456,
|
| 1185 |
+
"step": 8250
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.572048342113418,
|
| 1189 |
+
"grad_norm": 10.404682159423828,
|
| 1190 |
+
"learning_rate": 2.476087181076335e-05,
|
| 1191 |
+
"loss": 0.6808,
|
| 1192 |
+
"step": 8300
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.5875426092345832,
|
| 1196 |
+
"grad_norm": 5.160488128662109,
|
| 1197 |
+
"learning_rate": 2.47092242536928e-05,
|
| 1198 |
+
"loss": 0.6913,
|
| 1199 |
+
"step": 8350
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.6030368763557483,
|
| 1203 |
+
"grad_norm": 6.452591896057129,
|
| 1204 |
+
"learning_rate": 2.465757669662225e-05,
|
| 1205 |
+
"loss": 0.6594,
|
| 1206 |
+
"step": 8400
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 2.6185311434769134,
|
| 1210 |
+
"grad_norm": 12.436300277709961,
|
| 1211 |
+
"learning_rate": 2.46059291395517e-05,
|
| 1212 |
+
"loss": 0.7255,
|
| 1213 |
+
"step": 8450
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 2.6340254105980785,
|
| 1217 |
+
"grad_norm": 6.132791042327881,
|
| 1218 |
+
"learning_rate": 2.455428158248115e-05,
|
| 1219 |
+
"loss": 0.6753,
|
| 1220 |
+
"step": 8500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.649519677719244,
|
| 1224 |
+
"grad_norm": 10.712909698486328,
|
| 1225 |
+
"learning_rate": 2.45026340254106e-05,
|
| 1226 |
+
"loss": 0.6445,
|
| 1227 |
+
"step": 8550
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 2.665013944840409,
|
| 1231 |
+
"grad_norm": 12.122429847717285,
|
| 1232 |
+
"learning_rate": 2.445098646834005e-05,
|
| 1233 |
+
"loss": 0.6424,
|
| 1234 |
+
"step": 8600
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 2.680508211961574,
|
| 1238 |
+
"grad_norm": 8.575897216796875,
|
| 1239 |
+
"learning_rate": 2.4399338911269496e-05,
|
| 1240 |
+
"loss": 0.7242,
|
| 1241 |
+
"step": 8650
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.6960024790827393,
|
| 1245 |
+
"grad_norm": 8.740906715393066,
|
| 1246 |
+
"learning_rate": 2.4347691354198946e-05,
|
| 1247 |
+
"loss": 0.6949,
|
| 1248 |
+
"step": 8700
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 2.7114967462039044,
|
| 1252 |
+
"grad_norm": 4.871994972229004,
|
| 1253 |
+
"learning_rate": 2.4296043797128395e-05,
|
| 1254 |
+
"loss": 0.787,
|
| 1255 |
+
"step": 8750
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 2.72699101332507,
|
| 1259 |
+
"grad_norm": 6.642944812774658,
|
| 1260 |
+
"learning_rate": 2.424439624005785e-05,
|
| 1261 |
+
"loss": 0.6925,
|
| 1262 |
+
"step": 8800
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.742485280446235,
|
| 1266 |
+
"grad_norm": 12.149236679077148,
|
| 1267 |
+
"learning_rate": 2.4192748682987298e-05,
|
| 1268 |
+
"loss": 0.6972,
|
| 1269 |
+
"step": 8850
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 2.7579795475674,
|
| 1273 |
+
"grad_norm": 8.100613594055176,
|
| 1274 |
+
"learning_rate": 2.4141101125916744e-05,
|
| 1275 |
+
"loss": 0.7358,
|
| 1276 |
+
"step": 8900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 2.773473814688565,
|
| 1280 |
+
"grad_norm": 12.28987979888916,
|
| 1281 |
+
"learning_rate": 2.4089453568846194e-05,
|
| 1282 |
+
"loss": 0.7176,
|
| 1283 |
+
"step": 8950
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.78896808180973,
|
| 1287 |
+
"grad_norm": 9.355488777160645,
|
| 1288 |
+
"learning_rate": 2.4037806011775644e-05,
|
| 1289 |
+
"loss": 0.6856,
|
| 1290 |
+
"step": 9000
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 2.8044623489308957,
|
| 1294 |
+
"grad_norm": 11.875406265258789,
|
| 1295 |
+
"learning_rate": 2.3986158454705093e-05,
|
| 1296 |
+
"loss": 0.6501,
|
| 1297 |
+
"step": 9050
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 2.819956616052061,
|
| 1301 |
+
"grad_norm": 8.061235427856445,
|
| 1302 |
+
"learning_rate": 2.3934510897634543e-05,
|
| 1303 |
+
"loss": 0.6823,
|
| 1304 |
+
"step": 9100
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.835450883173226,
|
| 1308 |
+
"grad_norm": 7.949320316314697,
|
| 1309 |
+
"learning_rate": 2.388286334056399e-05,
|
| 1310 |
+
"loss": 0.6764,
|
| 1311 |
+
"step": 9150
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 2.850945150294391,
|
| 1315 |
+
"grad_norm": 5.9249587059021,
|
| 1316 |
+
"learning_rate": 2.3831215783493442e-05,
|
| 1317 |
+
"loss": 0.6511,
|
| 1318 |
+
"step": 9200
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 2.866439417415556,
|
| 1322 |
+
"grad_norm": 8.400185585021973,
|
| 1323 |
+
"learning_rate": 2.3779568226422892e-05,
|
| 1324 |
+
"loss": 0.6515,
|
| 1325 |
+
"step": 9250
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.8819336845367216,
|
| 1329 |
+
"grad_norm": 11.487894058227539,
|
| 1330 |
+
"learning_rate": 2.3727920669352342e-05,
|
| 1331 |
+
"loss": 0.6719,
|
| 1332 |
+
"step": 9300
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 2.8974279516578867,
|
| 1336 |
+
"grad_norm": 8.317901611328125,
|
| 1337 |
+
"learning_rate": 2.367627311228179e-05,
|
| 1338 |
+
"loss": 0.6697,
|
| 1339 |
+
"step": 9350
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 2.9129222187790518,
|
| 1343 |
+
"grad_norm": 9.878332138061523,
|
| 1344 |
+
"learning_rate": 2.3624625555211238e-05,
|
| 1345 |
+
"loss": 0.6801,
|
| 1346 |
+
"step": 9400
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 2.928416485900217,
|
| 1350 |
+
"grad_norm": 8.855628967285156,
|
| 1351 |
+
"learning_rate": 2.3572977998140687e-05,
|
| 1352 |
+
"loss": 0.6445,
|
| 1353 |
+
"step": 9450
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 2.943910753021382,
|
| 1357 |
+
"grad_norm": 5.350094318389893,
|
| 1358 |
+
"learning_rate": 2.3521330441070137e-05,
|
| 1359 |
+
"loss": 0.6891,
|
| 1360 |
+
"step": 9500
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 2.9594050201425475,
|
| 1364 |
+
"grad_norm": 8.540812492370605,
|
| 1365 |
+
"learning_rate": 2.3469682883999587e-05,
|
| 1366 |
+
"loss": 0.6556,
|
| 1367 |
+
"step": 9550
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.9748992872637126,
|
| 1371 |
+
"grad_norm": 4.337664604187012,
|
| 1372 |
+
"learning_rate": 2.341803532692904e-05,
|
| 1373 |
+
"loss": 0.7013,
|
| 1374 |
+
"step": 9600
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 2.9903935543848776,
|
| 1378 |
+
"grad_norm": 7.002617359161377,
|
| 1379 |
+
"learning_rate": 2.3366387769858486e-05,
|
| 1380 |
+
"loss": 0.6518,
|
| 1381 |
+
"step": 9650
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 3.0,
|
| 1385 |
+
"eval_accuracy": 0.777292576419214,
|
| 1386 |
+
"eval_f1": 0.775972842719567,
|
| 1387 |
+
"eval_loss": 0.69657963514328,
|
| 1388 |
+
"eval_runtime": 25.4856,
|
| 1389 |
+
"eval_samples_per_second": 260.579,
|
| 1390 |
+
"eval_steps_per_second": 16.323,
|
| 1391 |
+
"step": 9681
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 3.0058878215060427,
|
| 1395 |
+
"grad_norm": 9.228548049926758,
|
| 1396 |
+
"learning_rate": 2.3314740212787936e-05,
|
| 1397 |
+
"loss": 0.5385,
|
| 1398 |
+
"step": 9700
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 3.021382088627208,
|
| 1402 |
+
"grad_norm": 4.332932472229004,
|
| 1403 |
+
"learning_rate": 2.3263092655717385e-05,
|
| 1404 |
+
"loss": 0.5513,
|
| 1405 |
+
"step": 9750
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 3.036876355748373,
|
| 1409 |
+
"grad_norm": 6.478864669799805,
|
| 1410 |
+
"learning_rate": 2.3211445098646835e-05,
|
| 1411 |
+
"loss": 0.4542,
|
| 1412 |
+
"step": 9800
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 3.0523706228695384,
|
| 1416 |
+
"grad_norm": 14.028499603271484,
|
| 1417 |
+
"learning_rate": 2.3159797541576285e-05,
|
| 1418 |
+
"loss": 0.4549,
|
| 1419 |
+
"step": 9850
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 3.0678648899907035,
|
| 1423 |
+
"grad_norm": 5.590787887573242,
|
| 1424 |
+
"learning_rate": 2.310814998450573e-05,
|
| 1425 |
+
"loss": 0.4624,
|
| 1426 |
+
"step": 9900
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 3.0833591571118686,
|
| 1430 |
+
"grad_norm": 5.623167514801025,
|
| 1431 |
+
"learning_rate": 2.305650242743518e-05,
|
| 1432 |
+
"loss": 0.4479,
|
| 1433 |
+
"step": 9950
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 3.0988534242330337,
|
| 1437 |
+
"grad_norm": 10.343826293945312,
|
| 1438 |
+
"learning_rate": 2.3004854870364634e-05,
|
| 1439 |
+
"loss": 0.5079,
|
| 1440 |
+
"step": 10000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 3.1143476913541988,
|
| 1444 |
+
"grad_norm": 2.780686616897583,
|
| 1445 |
+
"learning_rate": 2.2953207313294084e-05,
|
| 1446 |
+
"loss": 0.43,
|
| 1447 |
+
"step": 10050
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 3.1298419584753643,
|
| 1451 |
+
"grad_norm": 10.917914390563965,
|
| 1452 |
+
"learning_rate": 2.2901559756223533e-05,
|
| 1453 |
+
"loss": 0.4932,
|
| 1454 |
+
"step": 10100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 3.1453362255965294,
|
| 1458 |
+
"grad_norm": 14.870561599731445,
|
| 1459 |
+
"learning_rate": 2.284991219915298e-05,
|
| 1460 |
+
"loss": 0.483,
|
| 1461 |
+
"step": 10150
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 3.1608304927176945,
|
| 1465 |
+
"grad_norm": 15.64564323425293,
|
| 1466 |
+
"learning_rate": 2.279826464208243e-05,
|
| 1467 |
+
"loss": 0.5047,
|
| 1468 |
+
"step": 10200
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 3.1763247598388595,
|
| 1472 |
+
"grad_norm": 8.148391723632812,
|
| 1473 |
+
"learning_rate": 2.274661708501188e-05,
|
| 1474 |
+
"loss": 0.498,
|
| 1475 |
+
"step": 10250
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 3.1918190269600246,
|
| 1479 |
+
"grad_norm": 9.916448593139648,
|
| 1480 |
+
"learning_rate": 2.269496952794133e-05,
|
| 1481 |
+
"loss": 0.4896,
|
| 1482 |
+
"step": 10300
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 3.20731329408119,
|
| 1486 |
+
"grad_norm": 10.014134407043457,
|
| 1487 |
+
"learning_rate": 2.2643321970870778e-05,
|
| 1488 |
+
"loss": 0.4572,
|
| 1489 |
+
"step": 10350
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 3.2228075612023552,
|
| 1493 |
+
"grad_norm": 9.647527694702148,
|
| 1494 |
+
"learning_rate": 2.2591674413800228e-05,
|
| 1495 |
+
"loss": 0.4965,
|
| 1496 |
+
"step": 10400
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 3.2383018283235203,
|
| 1500 |
+
"grad_norm": 11.77087116241455,
|
| 1501 |
+
"learning_rate": 2.2540026856729678e-05,
|
| 1502 |
+
"loss": 0.512,
|
| 1503 |
+
"step": 10450
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 3.2537960954446854,
|
| 1507 |
+
"grad_norm": 3.3613386154174805,
|
| 1508 |
+
"learning_rate": 2.2488379299659127e-05,
|
| 1509 |
+
"loss": 0.5522,
|
| 1510 |
+
"step": 10500
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 3.2692903625658505,
|
| 1514 |
+
"grad_norm": 17.92693519592285,
|
| 1515 |
+
"learning_rate": 2.2436731742588577e-05,
|
| 1516 |
+
"loss": 0.4915,
|
| 1517 |
+
"step": 10550
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 3.2847846296870156,
|
| 1521 |
+
"grad_norm": 8.389365196228027,
|
| 1522 |
+
"learning_rate": 2.2385084185518027e-05,
|
| 1523 |
+
"loss": 0.5343,
|
| 1524 |
+
"step": 10600
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 3.300278896808181,
|
| 1528 |
+
"grad_norm": 9.849445343017578,
|
| 1529 |
+
"learning_rate": 2.2333436628447473e-05,
|
| 1530 |
+
"loss": 0.4925,
|
| 1531 |
+
"step": 10650
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 3.315773163929346,
|
| 1535 |
+
"grad_norm": 7.494227886199951,
|
| 1536 |
+
"learning_rate": 2.2281789071376923e-05,
|
| 1537 |
+
"loss": 0.5242,
|
| 1538 |
+
"step": 10700
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 3.3312674310505113,
|
| 1542 |
+
"grad_norm": 12.774617195129395,
|
| 1543 |
+
"learning_rate": 2.2230141514306372e-05,
|
| 1544 |
+
"loss": 0.522,
|
| 1545 |
+
"step": 10750
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 3.3467616981716763,
|
| 1549 |
+
"grad_norm": 4.167229175567627,
|
| 1550 |
+
"learning_rate": 2.2178493957235822e-05,
|
| 1551 |
+
"loss": 0.4852,
|
| 1552 |
+
"step": 10800
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 3.3622559652928414,
|
| 1556 |
+
"grad_norm": 7.823596000671387,
|
| 1557 |
+
"learning_rate": 2.2126846400165275e-05,
|
| 1558 |
+
"loss": 0.521,
|
| 1559 |
+
"step": 10850
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 3.377750232414007,
|
| 1563 |
+
"grad_norm": 9.712186813354492,
|
| 1564 |
+
"learning_rate": 2.2075198843094725e-05,
|
| 1565 |
+
"loss": 0.4931,
|
| 1566 |
+
"step": 10900
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.393244499535172,
|
| 1570 |
+
"grad_norm": 9.726935386657715,
|
| 1571 |
+
"learning_rate": 2.202355128602417e-05,
|
| 1572 |
+
"loss": 0.531,
|
| 1573 |
+
"step": 10950
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 3.408738766656337,
|
| 1577 |
+
"grad_norm": 8.613348007202148,
|
| 1578 |
+
"learning_rate": 2.197190372895362e-05,
|
| 1579 |
+
"loss": 0.4902,
|
| 1580 |
+
"step": 11000
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 3.424233033777502,
|
| 1584 |
+
"grad_norm": 17.698650360107422,
|
| 1585 |
+
"learning_rate": 2.192025617188307e-05,
|
| 1586 |
+
"loss": 0.4967,
|
| 1587 |
+
"step": 11050
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.4397273008986673,
|
| 1591 |
+
"grad_norm": 13.304680824279785,
|
| 1592 |
+
"learning_rate": 2.186860861481252e-05,
|
| 1593 |
+
"loss": 0.4998,
|
| 1594 |
+
"step": 11100
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 3.455221568019833,
|
| 1598 |
+
"grad_norm": 9.090615272521973,
|
| 1599 |
+
"learning_rate": 2.181696105774197e-05,
|
| 1600 |
+
"loss": 0.4797,
|
| 1601 |
+
"step": 11150
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.470715835140998,
|
| 1605 |
+
"grad_norm": 6.544071197509766,
|
| 1606 |
+
"learning_rate": 2.176531350067142e-05,
|
| 1607 |
+
"loss": 0.5405,
|
| 1608 |
+
"step": 11200
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 3.486210102262163,
|
| 1612 |
+
"grad_norm": 10.908158302307129,
|
| 1613 |
+
"learning_rate": 2.171366594360087e-05,
|
| 1614 |
+
"loss": 0.4663,
|
| 1615 |
+
"step": 11250
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 3.501704369383328,
|
| 1619 |
+
"grad_norm": 9.044700622558594,
|
| 1620 |
+
"learning_rate": 2.166201838653032e-05,
|
| 1621 |
+
"loss": 0.4755,
|
| 1622 |
+
"step": 11300
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 3.517198636504493,
|
| 1626 |
+
"grad_norm": 7.633232116699219,
|
| 1627 |
+
"learning_rate": 2.161037082945977e-05,
|
| 1628 |
+
"loss": 0.4182,
|
| 1629 |
+
"step": 11350
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 3.5326929036256587,
|
| 1633 |
+
"grad_norm": 5.32473087310791,
|
| 1634 |
+
"learning_rate": 2.1558723272389218e-05,
|
| 1635 |
+
"loss": 0.4987,
|
| 1636 |
+
"step": 11400
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 3.5481871707468238,
|
| 1640 |
+
"grad_norm": 9.8456392288208,
|
| 1641 |
+
"learning_rate": 2.1507075715318664e-05,
|
| 1642 |
+
"loss": 0.587,
|
| 1643 |
+
"step": 11450
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 3.563681437867989,
|
| 1647 |
+
"grad_norm": 12.52115535736084,
|
| 1648 |
+
"learning_rate": 2.1455428158248114e-05,
|
| 1649 |
+
"loss": 0.5331,
|
| 1650 |
+
"step": 11500
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 3.579175704989154,
|
| 1654 |
+
"grad_norm": 18.225566864013672,
|
| 1655 |
+
"learning_rate": 2.1403780601177564e-05,
|
| 1656 |
+
"loss": 0.4794,
|
| 1657 |
+
"step": 11550
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 3.594669972110319,
|
| 1661 |
+
"grad_norm": 8.749368667602539,
|
| 1662 |
+
"learning_rate": 2.1352133044107013e-05,
|
| 1663 |
+
"loss": 0.4967,
|
| 1664 |
+
"step": 11600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 3.6101642392314846,
|
| 1668 |
+
"grad_norm": 8.760223388671875,
|
| 1669 |
+
"learning_rate": 2.1300485487036466e-05,
|
| 1670 |
+
"loss": 0.4963,
|
| 1671 |
+
"step": 11650
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 3.6256585063526496,
|
| 1675 |
+
"grad_norm": 15.518270492553711,
|
| 1676 |
+
"learning_rate": 2.1248837929965913e-05,
|
| 1677 |
+
"loss": 0.4456,
|
| 1678 |
+
"step": 11700
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 3.6411527734738147,
|
| 1682 |
+
"grad_norm": 9.451664924621582,
|
| 1683 |
+
"learning_rate": 2.1197190372895362e-05,
|
| 1684 |
+
"loss": 0.5235,
|
| 1685 |
+
"step": 11750
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 3.65664704059498,
|
| 1689 |
+
"grad_norm": 17.736055374145508,
|
| 1690 |
+
"learning_rate": 2.1145542815824812e-05,
|
| 1691 |
+
"loss": 0.4876,
|
| 1692 |
+
"step": 11800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 3.672141307716145,
|
| 1696 |
+
"grad_norm": 24.323490142822266,
|
| 1697 |
+
"learning_rate": 2.1093895258754262e-05,
|
| 1698 |
+
"loss": 0.483,
|
| 1699 |
+
"step": 11850
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 3.6876355748373104,
|
| 1703 |
+
"grad_norm": 15.389254570007324,
|
| 1704 |
+
"learning_rate": 2.104224770168371e-05,
|
| 1705 |
+
"loss": 0.5127,
|
| 1706 |
+
"step": 11900
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 3.7031298419584755,
|
| 1710 |
+
"grad_norm": 11.283272743225098,
|
| 1711 |
+
"learning_rate": 2.0990600144613158e-05,
|
| 1712 |
+
"loss": 0.5282,
|
| 1713 |
+
"step": 11950
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 3.7186241090796406,
|
| 1717 |
+
"grad_norm": 11.002310752868652,
|
| 1718 |
+
"learning_rate": 2.0938952587542607e-05,
|
| 1719 |
+
"loss": 0.5459,
|
| 1720 |
+
"step": 12000
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 3.7341183762008057,
|
| 1724 |
+
"grad_norm": 6.972140312194824,
|
| 1725 |
+
"learning_rate": 2.088730503047206e-05,
|
| 1726 |
+
"loss": 0.536,
|
| 1727 |
+
"step": 12050
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 3.7496126433219708,
|
| 1731 |
+
"grad_norm": 4.202858924865723,
|
| 1732 |
+
"learning_rate": 2.083565747340151e-05,
|
| 1733 |
+
"loss": 0.5736,
|
| 1734 |
+
"step": 12100
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 3.7651069104431363,
|
| 1738 |
+
"grad_norm": 15.748515129089355,
|
| 1739 |
+
"learning_rate": 2.078400991633096e-05,
|
| 1740 |
+
"loss": 0.4715,
|
| 1741 |
+
"step": 12150
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 3.7806011775643014,
|
| 1745 |
+
"grad_norm": 6.696774482727051,
|
| 1746 |
+
"learning_rate": 2.0732362359260406e-05,
|
| 1747 |
+
"loss": 0.5545,
|
| 1748 |
+
"step": 12200
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 3.7960954446854664,
|
| 1752 |
+
"grad_norm": 7.366288661956787,
|
| 1753 |
+
"learning_rate": 2.0680714802189856e-05,
|
| 1754 |
+
"loss": 0.5736,
|
| 1755 |
+
"step": 12250
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 3.8115897118066315,
|
| 1759 |
+
"grad_norm": 13.58438777923584,
|
| 1760 |
+
"learning_rate": 2.0629067245119306e-05,
|
| 1761 |
+
"loss": 0.4255,
|
| 1762 |
+
"step": 12300
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 3.8270839789277966,
|
| 1766 |
+
"grad_norm": 9.109688758850098,
|
| 1767 |
+
"learning_rate": 2.0577419688048755e-05,
|
| 1768 |
+
"loss": 0.4565,
|
| 1769 |
+
"step": 12350
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 3.842578246048962,
|
| 1773 |
+
"grad_norm": 11.448044776916504,
|
| 1774 |
+
"learning_rate": 2.0525772130978205e-05,
|
| 1775 |
+
"loss": 0.5117,
|
| 1776 |
+
"step": 12400
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 3.858072513170127,
|
| 1780 |
+
"grad_norm": 6.876945495605469,
|
| 1781 |
+
"learning_rate": 2.0474124573907655e-05,
|
| 1782 |
+
"loss": 0.5543,
|
| 1783 |
+
"step": 12450
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 3.8735667802912923,
|
| 1787 |
+
"grad_norm": 11.25009536743164,
|
| 1788 |
+
"learning_rate": 2.0422477016837104e-05,
|
| 1789 |
+
"loss": 0.456,
|
| 1790 |
+
"step": 12500
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 3.8890610474124574,
|
| 1794 |
+
"grad_norm": 13.992502212524414,
|
| 1795 |
+
"learning_rate": 2.0370829459766554e-05,
|
| 1796 |
+
"loss": 0.4907,
|
| 1797 |
+
"step": 12550
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 3.9045553145336225,
|
| 1801 |
+
"grad_norm": 11.92656421661377,
|
| 1802 |
+
"learning_rate": 2.0319181902696004e-05,
|
| 1803 |
+
"loss": 0.4841,
|
| 1804 |
+
"step": 12600
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 3.9200495816547876,
|
| 1808 |
+
"grad_norm": 7.212582111358643,
|
| 1809 |
+
"learning_rate": 2.0267534345625453e-05,
|
| 1810 |
+
"loss": 0.5529,
|
| 1811 |
+
"step": 12650
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 3.9355438487759526,
|
| 1815 |
+
"grad_norm": 14.616645812988281,
|
| 1816 |
+
"learning_rate": 2.02158867885549e-05,
|
| 1817 |
+
"loss": 0.5366,
|
| 1818 |
+
"step": 12700
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 3.951038115897118,
|
| 1822 |
+
"grad_norm": 9.052292823791504,
|
| 1823 |
+
"learning_rate": 2.016423923148435e-05,
|
| 1824 |
+
"loss": 0.5459,
|
| 1825 |
+
"step": 12750
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 3.9665323830182833,
|
| 1829 |
+
"grad_norm": 18.27539825439453,
|
| 1830 |
+
"learning_rate": 2.01125916744138e-05,
|
| 1831 |
+
"loss": 0.5631,
|
| 1832 |
+
"step": 12800
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 3.9820266501394483,
|
| 1836 |
+
"grad_norm": 12.429372787475586,
|
| 1837 |
+
"learning_rate": 2.0060944117343252e-05,
|
| 1838 |
+
"loss": 0.4885,
|
| 1839 |
+
"step": 12850
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 3.9975209172606134,
|
| 1843 |
+
"grad_norm": 4.481673240661621,
|
| 1844 |
+
"learning_rate": 2.00092965602727e-05,
|
| 1845 |
+
"loss": 0.4565,
|
| 1846 |
+
"step": 12900
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 4.0,
|
| 1850 |
+
"eval_accuracy": 0.7863273603372986,
|
| 1851 |
+
"eval_f1": 0.7874334496964743,
|
| 1852 |
+
"eval_loss": 0.7491569519042969,
|
| 1853 |
+
"eval_runtime": 25.5609,
|
| 1854 |
+
"eval_samples_per_second": 259.811,
|
| 1855 |
+
"eval_steps_per_second": 16.275,
|
| 1856 |
+
"step": 12908
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 4.0130151843817785,
|
| 1860 |
+
"grad_norm": 21.53974151611328,
|
| 1861 |
+
"learning_rate": 1.9957649003202148e-05,
|
| 1862 |
+
"loss": 0.2695,
|
| 1863 |
+
"step": 12950
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 4.028509451502944,
|
| 1867 |
+
"grad_norm": 9.07942008972168,
|
| 1868 |
+
"learning_rate": 1.9906001446131598e-05,
|
| 1869 |
+
"loss": 0.3282,
|
| 1870 |
+
"step": 13000
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 4.044003718624109,
|
| 1874 |
+
"grad_norm": 16.323549270629883,
|
| 1875 |
+
"learning_rate": 1.9854353889061047e-05,
|
| 1876 |
+
"loss": 0.3079,
|
| 1877 |
+
"step": 13050
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 4.059497985745274,
|
| 1881 |
+
"grad_norm": 6.679697036743164,
|
| 1882 |
+
"learning_rate": 1.9802706331990497e-05,
|
| 1883 |
+
"loss": 0.3889,
|
| 1884 |
+
"step": 13100
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 4.07499225286644,
|
| 1888 |
+
"grad_norm": 17.357574462890625,
|
| 1889 |
+
"learning_rate": 1.9751058774919947e-05,
|
| 1890 |
+
"loss": 0.3336,
|
| 1891 |
+
"step": 13150
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 4.090486519987604,
|
| 1895 |
+
"grad_norm": 5.116195201873779,
|
| 1896 |
+
"learning_rate": 1.9699411217849393e-05,
|
| 1897 |
+
"loss": 0.3102,
|
| 1898 |
+
"step": 13200
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 4.10598078710877,
|
| 1902 |
+
"grad_norm": 29.05538558959961,
|
| 1903 |
+
"learning_rate": 1.9647763660778846e-05,
|
| 1904 |
+
"loss": 0.2902,
|
| 1905 |
+
"step": 13250
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 4.1214750542299345,
|
| 1909 |
+
"grad_norm": 6.254473686218262,
|
| 1910 |
+
"learning_rate": 1.9596116103708296e-05,
|
| 1911 |
+
"loss": 0.3816,
|
| 1912 |
+
"step": 13300
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 4.1369693213511,
|
| 1916 |
+
"grad_norm": 11.854185104370117,
|
| 1917 |
+
"learning_rate": 1.9544468546637745e-05,
|
| 1918 |
+
"loss": 0.3455,
|
| 1919 |
+
"step": 13350
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 4.152463588472266,
|
| 1923 |
+
"grad_norm": 16.399444580078125,
|
| 1924 |
+
"learning_rate": 1.9492820989567195e-05,
|
| 1925 |
+
"loss": 0.3713,
|
| 1926 |
+
"step": 13400
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 4.16795785559343,
|
| 1930 |
+
"grad_norm": 18.26226234436035,
|
| 1931 |
+
"learning_rate": 1.9441173432496645e-05,
|
| 1932 |
+
"loss": 0.2957,
|
| 1933 |
+
"step": 13450
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 4.183452122714596,
|
| 1937 |
+
"grad_norm": 6.590181350708008,
|
| 1938 |
+
"learning_rate": 1.938952587542609e-05,
|
| 1939 |
+
"loss": 0.2905,
|
| 1940 |
+
"step": 13500
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 4.19894638983576,
|
| 1944 |
+
"grad_norm": 5.3814849853515625,
|
| 1945 |
+
"learning_rate": 1.933787831835554e-05,
|
| 1946 |
+
"loss": 0.3782,
|
| 1947 |
+
"step": 13550
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 4.214440656956926,
|
| 1951 |
+
"grad_norm": 8.641956329345703,
|
| 1952 |
+
"learning_rate": 1.928623076128499e-05,
|
| 1953 |
+
"loss": 0.3211,
|
| 1954 |
+
"step": 13600
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 4.2299349240780915,
|
| 1958 |
+
"grad_norm": 14.346405982971191,
|
| 1959 |
+
"learning_rate": 1.9234583204214443e-05,
|
| 1960 |
+
"loss": 0.3274,
|
| 1961 |
+
"step": 13650
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 4.245429191199256,
|
| 1965 |
+
"grad_norm": 15.577725410461426,
|
| 1966 |
+
"learning_rate": 1.9182935647143893e-05,
|
| 1967 |
+
"loss": 0.3568,
|
| 1968 |
+
"step": 13700
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 4.260923458320422,
|
| 1972 |
+
"grad_norm": 9.855398178100586,
|
| 1973 |
+
"learning_rate": 1.913128809007334e-05,
|
| 1974 |
+
"loss": 0.3008,
|
| 1975 |
+
"step": 13750
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 4.276417725441586,
|
| 1979 |
+
"grad_norm": 15.720294952392578,
|
| 1980 |
+
"learning_rate": 1.907964053300279e-05,
|
| 1981 |
+
"loss": 0.2979,
|
| 1982 |
+
"step": 13800
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 4.291911992562752,
|
| 1986 |
+
"grad_norm": 13.976778030395508,
|
| 1987 |
+
"learning_rate": 1.902799297593224e-05,
|
| 1988 |
+
"loss": 0.3389,
|
| 1989 |
+
"step": 13850
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 4.307406259683917,
|
| 1993 |
+
"grad_norm": 19.255727767944336,
|
| 1994 |
+
"learning_rate": 1.897634541886169e-05,
|
| 1995 |
+
"loss": 0.3636,
|
| 1996 |
+
"step": 13900
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 4.322900526805082,
|
| 2000 |
+
"grad_norm": 10.70836353302002,
|
| 2001 |
+
"learning_rate": 1.8924697861791138e-05,
|
| 2002 |
+
"loss": 0.3455,
|
| 2003 |
+
"step": 13950
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 4.3383947939262475,
|
| 2007 |
+
"grad_norm": 0.9212763905525208,
|
| 2008 |
+
"learning_rate": 1.8873050304720584e-05,
|
| 2009 |
+
"loss": 0.3703,
|
| 2010 |
+
"step": 14000
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 4.353889061047412,
|
| 2014 |
+
"grad_norm": 10.232623100280762,
|
| 2015 |
+
"learning_rate": 1.8821402747650037e-05,
|
| 2016 |
+
"loss": 0.3247,
|
| 2017 |
+
"step": 14050
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 4.369383328168578,
|
| 2021 |
+
"grad_norm": 11.130922317504883,
|
| 2022 |
+
"learning_rate": 1.8769755190579487e-05,
|
| 2023 |
+
"loss": 0.3107,
|
| 2024 |
+
"step": 14100
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 4.384877595289743,
|
| 2028 |
+
"grad_norm": 10.536752700805664,
|
| 2029 |
+
"learning_rate": 1.8718107633508937e-05,
|
| 2030 |
+
"loss": 0.3614,
|
| 2031 |
+
"step": 14150
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 4.400371862410908,
|
| 2035 |
+
"grad_norm": 15.330968856811523,
|
| 2036 |
+
"learning_rate": 1.8666460076438386e-05,
|
| 2037 |
+
"loss": 0.3984,
|
| 2038 |
+
"step": 14200
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 4.415866129532073,
|
| 2042 |
+
"grad_norm": 7.436588764190674,
|
| 2043 |
+
"learning_rate": 1.8614812519367833e-05,
|
| 2044 |
+
"loss": 0.3257,
|
| 2045 |
+
"step": 14250
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 4.431360396653238,
|
| 2049 |
+
"grad_norm": 7.192384243011475,
|
| 2050 |
+
"learning_rate": 1.8563164962297282e-05,
|
| 2051 |
+
"loss": 0.3254,
|
| 2052 |
+
"step": 14300
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 4.4468546637744035,
|
| 2056 |
+
"grad_norm": 7.792993545532227,
|
| 2057 |
+
"learning_rate": 1.8511517405226732e-05,
|
| 2058 |
+
"loss": 0.3392,
|
| 2059 |
+
"step": 14350
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 4.462348930895569,
|
| 2063 |
+
"grad_norm": 12.411416053771973,
|
| 2064 |
+
"learning_rate": 1.8459869848156182e-05,
|
| 2065 |
+
"loss": 0.3383,
|
| 2066 |
+
"step": 14400
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 4.477843198016734,
|
| 2070 |
+
"grad_norm": 17.897613525390625,
|
| 2071 |
+
"learning_rate": 1.8408222291085635e-05,
|
| 2072 |
+
"loss": 0.3392,
|
| 2073 |
+
"step": 14450
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 4.493337465137899,
|
| 2077 |
+
"grad_norm": 23.59228515625,
|
| 2078 |
+
"learning_rate": 1.835657473401508e-05,
|
| 2079 |
+
"loss": 0.3055,
|
| 2080 |
+
"step": 14500
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 4.508831732259064,
|
| 2084 |
+
"grad_norm": 13.722383499145508,
|
| 2085 |
+
"learning_rate": 1.830492717694453e-05,
|
| 2086 |
+
"loss": 0.3997,
|
| 2087 |
+
"step": 14550
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 4.524325999380229,
|
| 2091 |
+
"grad_norm": 17.811538696289062,
|
| 2092 |
+
"learning_rate": 1.825327961987398e-05,
|
| 2093 |
+
"loss": 0.266,
|
| 2094 |
+
"step": 14600
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 4.539820266501394,
|
| 2098 |
+
"grad_norm": 10.993431091308594,
|
| 2099 |
+
"learning_rate": 1.820163206280343e-05,
|
| 2100 |
+
"loss": 0.2634,
|
| 2101 |
+
"step": 14650
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 4.55531453362256,
|
| 2105 |
+
"grad_norm": 5.25628137588501,
|
| 2106 |
+
"learning_rate": 1.814998450573288e-05,
|
| 2107 |
+
"loss": 0.3563,
|
| 2108 |
+
"step": 14700
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 4.570808800743725,
|
| 2112 |
+
"grad_norm": 16.91241455078125,
|
| 2113 |
+
"learning_rate": 1.8098336948662326e-05,
|
| 2114 |
+
"loss": 0.3298,
|
| 2115 |
+
"step": 14750
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 4.58630306786489,
|
| 2119 |
+
"grad_norm": 27.083995819091797,
|
| 2120 |
+
"learning_rate": 1.8046689391591776e-05,
|
| 2121 |
+
"loss": 0.36,
|
| 2122 |
+
"step": 14800
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 4.601797334986055,
|
| 2126 |
+
"grad_norm": 19.726198196411133,
|
| 2127 |
+
"learning_rate": 1.799504183452123e-05,
|
| 2128 |
+
"loss": 0.3224,
|
| 2129 |
+
"step": 14850
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 4.617291602107221,
|
| 2133 |
+
"grad_norm": 6.92859411239624,
|
| 2134 |
+
"learning_rate": 1.794339427745068e-05,
|
| 2135 |
+
"loss": 0.335,
|
| 2136 |
+
"step": 14900
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 4.632785869228385,
|
| 2140 |
+
"grad_norm": 15.97644329071045,
|
| 2141 |
+
"learning_rate": 1.7891746720380128e-05,
|
| 2142 |
+
"loss": 0.3303,
|
| 2143 |
+
"step": 14950
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 4.648280136349551,
|
| 2147 |
+
"grad_norm": 24.399837493896484,
|
| 2148 |
+
"learning_rate": 1.7840099163309575e-05,
|
| 2149 |
+
"loss": 0.3492,
|
| 2150 |
+
"step": 15000
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 4.663774403470716,
|
| 2154 |
+
"grad_norm": 10.855368614196777,
|
| 2155 |
+
"learning_rate": 1.7788451606239024e-05,
|
| 2156 |
+
"loss": 0.3278,
|
| 2157 |
+
"step": 15050
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 4.679268670591881,
|
| 2161 |
+
"grad_norm": 20.869380950927734,
|
| 2162 |
+
"learning_rate": 1.7736804049168474e-05,
|
| 2163 |
+
"loss": 0.2913,
|
| 2164 |
+
"step": 15100
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 4.694762937713046,
|
| 2168 |
+
"grad_norm": 6.862913131713867,
|
| 2169 |
+
"learning_rate": 1.7685156492097924e-05,
|
| 2170 |
+
"loss": 0.3133,
|
| 2171 |
+
"step": 15150
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 4.710257204834211,
|
| 2175 |
+
"grad_norm": 19.621482849121094,
|
| 2176 |
+
"learning_rate": 1.7633508935027373e-05,
|
| 2177 |
+
"loss": 0.3456,
|
| 2178 |
+
"step": 15200
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 4.725751471955377,
|
| 2182 |
+
"grad_norm": 19.79738998413086,
|
| 2183 |
+
"learning_rate": 1.7581861377956823e-05,
|
| 2184 |
+
"loss": 0.3562,
|
| 2185 |
+
"step": 15250
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 4.7412457390765415,
|
| 2189 |
+
"grad_norm": 3.2352957725524902,
|
| 2190 |
+
"learning_rate": 1.7530213820886273e-05,
|
| 2191 |
+
"loss": 0.3565,
|
| 2192 |
+
"step": 15300
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 4.756740006197707,
|
| 2196 |
+
"grad_norm": 10.959282875061035,
|
| 2197 |
+
"learning_rate": 1.7478566263815722e-05,
|
| 2198 |
+
"loss": 0.3472,
|
| 2199 |
+
"step": 15350
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 4.7722342733188725,
|
| 2203 |
+
"grad_norm": 3.22469162940979,
|
| 2204 |
+
"learning_rate": 1.7426918706745172e-05,
|
| 2205 |
+
"loss": 0.3367,
|
| 2206 |
+
"step": 15400
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 4.787728540440037,
|
| 2210 |
+
"grad_norm": 7.619373798370361,
|
| 2211 |
+
"learning_rate": 1.737527114967462e-05,
|
| 2212 |
+
"loss": 0.319,
|
| 2213 |
+
"step": 15450
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 4.803222807561203,
|
| 2217 |
+
"grad_norm": 24.706689834594727,
|
| 2218 |
+
"learning_rate": 1.7323623592604068e-05,
|
| 2219 |
+
"loss": 0.3939,
|
| 2220 |
+
"step": 15500
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 4.818717074682367,
|
| 2224 |
+
"grad_norm": 15.918986320495605,
|
| 2225 |
+
"learning_rate": 1.7271976035533518e-05,
|
| 2226 |
+
"loss": 0.3618,
|
| 2227 |
+
"step": 15550
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.834211341803533,
|
| 2231 |
+
"grad_norm": 14.518546104431152,
|
| 2232 |
+
"learning_rate": 1.7220328478462967e-05,
|
| 2233 |
+
"loss": 0.4082,
|
| 2234 |
+
"step": 15600
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 4.8497056089246975,
|
| 2238 |
+
"grad_norm": 6.084866046905518,
|
| 2239 |
+
"learning_rate": 1.716868092139242e-05,
|
| 2240 |
+
"loss": 0.3594,
|
| 2241 |
+
"step": 15650
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 4.865199876045863,
|
| 2245 |
+
"grad_norm": 18.435983657836914,
|
| 2246 |
+
"learning_rate": 1.711703336432187e-05,
|
| 2247 |
+
"loss": 0.3182,
|
| 2248 |
+
"step": 15700
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 4.8806941431670285,
|
| 2252 |
+
"grad_norm": 14.745248794555664,
|
| 2253 |
+
"learning_rate": 1.7065385807251316e-05,
|
| 2254 |
+
"loss": 0.3375,
|
| 2255 |
+
"step": 15750
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 4.896188410288193,
|
| 2259 |
+
"grad_norm": 11.518832206726074,
|
| 2260 |
+
"learning_rate": 1.7013738250180766e-05,
|
| 2261 |
+
"loss": 0.3371,
|
| 2262 |
+
"step": 15800
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 4.911682677409359,
|
| 2266 |
+
"grad_norm": 17.58115005493164,
|
| 2267 |
+
"learning_rate": 1.6962090693110216e-05,
|
| 2268 |
+
"loss": 0.3851,
|
| 2269 |
+
"step": 15850
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.927176944530523,
|
| 2273 |
+
"grad_norm": 16.769134521484375,
|
| 2274 |
+
"learning_rate": 1.6910443136039665e-05,
|
| 2275 |
+
"loss": 0.3009,
|
| 2276 |
+
"step": 15900
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 4.942671211651689,
|
| 2280 |
+
"grad_norm": 21.518749237060547,
|
| 2281 |
+
"learning_rate": 1.6858795578969115e-05,
|
| 2282 |
+
"loss": 0.3155,
|
| 2283 |
+
"step": 15950
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 4.958165478772854,
|
| 2287 |
+
"grad_norm": 11.044340133666992,
|
| 2288 |
+
"learning_rate": 1.6807148021898565e-05,
|
| 2289 |
+
"loss": 0.3648,
|
| 2290 |
+
"step": 16000
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 4.973659745894019,
|
| 2294 |
+
"grad_norm": 3.9900588989257812,
|
| 2295 |
+
"learning_rate": 1.6755500464828014e-05,
|
| 2296 |
+
"loss": 0.3383,
|
| 2297 |
+
"step": 16050
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 4.989154013015185,
|
| 2301 |
+
"grad_norm": 16.869041442871094,
|
| 2302 |
+
"learning_rate": 1.6703852907757464e-05,
|
| 2303 |
+
"loss": 0.3471,
|
| 2304 |
+
"step": 16100
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 5.0,
|
| 2308 |
+
"eval_accuracy": 0.7964162023791598,
|
| 2309 |
+
"eval_f1": 0.7960859271865419,
|
| 2310 |
+
"eval_loss": 0.8064730167388916,
|
| 2311 |
+
"eval_runtime": 25.5199,
|
| 2312 |
+
"eval_samples_per_second": 260.228,
|
| 2313 |
+
"eval_steps_per_second": 16.301,
|
| 2314 |
+
"step": 16135
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 5.004648280136349,
|
| 2318 |
+
"grad_norm": 6.430090427398682,
|
| 2319 |
+
"learning_rate": 1.6652205350686914e-05,
|
| 2320 |
+
"loss": 0.2737,
|
| 2321 |
+
"step": 16150
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 5.020142547257515,
|
| 2325 |
+
"grad_norm": 5.049603462219238,
|
| 2326 |
+
"learning_rate": 1.6600557793616363e-05,
|
| 2327 |
+
"loss": 0.2016,
|
| 2328 |
+
"step": 16200
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 5.03563681437868,
|
| 2332 |
+
"grad_norm": 19.016332626342773,
|
| 2333 |
+
"learning_rate": 1.6548910236545813e-05,
|
| 2334 |
+
"loss": 0.256,
|
| 2335 |
+
"step": 16250
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 5.051131081499845,
|
| 2339 |
+
"grad_norm": 26.316680908203125,
|
| 2340 |
+
"learning_rate": 1.649726267947526e-05,
|
| 2341 |
+
"loss": 0.2052,
|
| 2342 |
+
"step": 16300
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 5.06662534862101,
|
| 2346 |
+
"grad_norm": 3.5623199939727783,
|
| 2347 |
+
"learning_rate": 1.644561512240471e-05,
|
| 2348 |
+
"loss": 0.1912,
|
| 2349 |
+
"step": 16350
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 5.082119615742175,
|
| 2353 |
+
"grad_norm": 17.87260627746582,
|
| 2354 |
+
"learning_rate": 1.639396756533416e-05,
|
| 2355 |
+
"loss": 0.1743,
|
| 2356 |
+
"step": 16400
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 5.097613882863341,
|
| 2360 |
+
"grad_norm": 25.36100959777832,
|
| 2361 |
+
"learning_rate": 1.6342320008263612e-05,
|
| 2362 |
+
"loss": 0.2196,
|
| 2363 |
+
"step": 16450
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 5.113108149984506,
|
| 2367 |
+
"grad_norm": 0.5424543023109436,
|
| 2368 |
+
"learning_rate": 1.629067245119306e-05,
|
| 2369 |
+
"loss": 0.2298,
|
| 2370 |
+
"step": 16500
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 5.128602417105671,
|
| 2374 |
+
"grad_norm": 11.180691719055176,
|
| 2375 |
+
"learning_rate": 1.6239024894122508e-05,
|
| 2376 |
+
"loss": 0.2107,
|
| 2377 |
+
"step": 16550
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 5.144096684226836,
|
| 2381 |
+
"grad_norm": 20.951330184936523,
|
| 2382 |
+
"learning_rate": 1.6187377337051957e-05,
|
| 2383 |
+
"loss": 0.2708,
|
| 2384 |
+
"step": 16600
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 5.159590951348001,
|
| 2388 |
+
"grad_norm": 11.13031005859375,
|
| 2389 |
+
"learning_rate": 1.6135729779981407e-05,
|
| 2390 |
+
"loss": 0.1913,
|
| 2391 |
+
"step": 16650
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 5.1750852184691665,
|
| 2395 |
+
"grad_norm": 0.4118750989437103,
|
| 2396 |
+
"learning_rate": 1.6084082222910857e-05,
|
| 2397 |
+
"loss": 0.2151,
|
| 2398 |
+
"step": 16700
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 5.190579485590332,
|
| 2402 |
+
"grad_norm": 37.649024963378906,
|
| 2403 |
+
"learning_rate": 1.6032434665840307e-05,
|
| 2404 |
+
"loss": 0.2281,
|
| 2405 |
+
"step": 16750
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 5.206073752711497,
|
| 2409 |
+
"grad_norm": 13.980716705322266,
|
| 2410 |
+
"learning_rate": 1.5980787108769753e-05,
|
| 2411 |
+
"loss": 0.241,
|
| 2412 |
+
"step": 16800
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 5.221568019832662,
|
| 2416 |
+
"grad_norm": 34.06684112548828,
|
| 2417 |
+
"learning_rate": 1.5929139551699206e-05,
|
| 2418 |
+
"loss": 0.3001,
|
| 2419 |
+
"step": 16850
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 5.237062286953827,
|
| 2423 |
+
"grad_norm": 26.94174575805664,
|
| 2424 |
+
"learning_rate": 1.5877491994628656e-05,
|
| 2425 |
+
"loss": 0.2121,
|
| 2426 |
+
"step": 16900
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 5.252556554074992,
|
| 2430 |
+
"grad_norm": 30.45203399658203,
|
| 2431 |
+
"learning_rate": 1.5825844437558105e-05,
|
| 2432 |
+
"loss": 0.2403,
|
| 2433 |
+
"step": 16950
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 5.268050821196157,
|
| 2437 |
+
"grad_norm": 10.348464965820312,
|
| 2438 |
+
"learning_rate": 1.5774196880487555e-05,
|
| 2439 |
+
"loss": 0.2228,
|
| 2440 |
+
"step": 17000
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 5.2835450883173225,
|
| 2444 |
+
"grad_norm": 35.869544982910156,
|
| 2445 |
+
"learning_rate": 1.5722549323417e-05,
|
| 2446 |
+
"loss": 0.2325,
|
| 2447 |
+
"step": 17050
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 5.299039355438488,
|
| 2451 |
+
"grad_norm": 11.559053421020508,
|
| 2452 |
+
"learning_rate": 1.567090176634645e-05,
|
| 2453 |
+
"loss": 0.1989,
|
| 2454 |
+
"step": 17100
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 5.314533622559653,
|
| 2458 |
+
"grad_norm": 14.056360244750977,
|
| 2459 |
+
"learning_rate": 1.56192542092759e-05,
|
| 2460 |
+
"loss": 0.242,
|
| 2461 |
+
"step": 17150
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 5.330027889680818,
|
| 2465 |
+
"grad_norm": 9.175251960754395,
|
| 2466 |
+
"learning_rate": 1.556760665220535e-05,
|
| 2467 |
+
"loss": 0.2942,
|
| 2468 |
+
"step": 17200
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 5.345522156801984,
|
| 2472 |
+
"grad_norm": 5.0493483543396,
|
| 2473 |
+
"learning_rate": 1.5515959095134803e-05,
|
| 2474 |
+
"loss": 0.217,
|
| 2475 |
+
"step": 17250
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 5.361016423923148,
|
| 2479 |
+
"grad_norm": 14.81401252746582,
|
| 2480 |
+
"learning_rate": 1.546431153806425e-05,
|
| 2481 |
+
"loss": 0.2217,
|
| 2482 |
+
"step": 17300
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 5.376510691044314,
|
| 2486 |
+
"grad_norm": 11.582941055297852,
|
| 2487 |
+
"learning_rate": 1.54126639809937e-05,
|
| 2488 |
+
"loss": 0.2175,
|
| 2489 |
+
"step": 17350
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 5.3920049581654785,
|
| 2493 |
+
"grad_norm": 16.075231552124023,
|
| 2494 |
+
"learning_rate": 1.536101642392315e-05,
|
| 2495 |
+
"loss": 0.2031,
|
| 2496 |
+
"step": 17400
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 5.407499225286644,
|
| 2500 |
+
"grad_norm": 4.986071586608887,
|
| 2501 |
+
"learning_rate": 1.53093688668526e-05,
|
| 2502 |
+
"loss": 0.2476,
|
| 2503 |
+
"step": 17450
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 5.422993492407809,
|
| 2507 |
+
"grad_norm": 12.152420997619629,
|
| 2508 |
+
"learning_rate": 1.5257721309782048e-05,
|
| 2509 |
+
"loss": 0.2839,
|
| 2510 |
+
"step": 17500
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 5.438487759528974,
|
| 2514 |
+
"grad_norm": 34.93170928955078,
|
| 2515 |
+
"learning_rate": 1.5206073752711496e-05,
|
| 2516 |
+
"loss": 0.2486,
|
| 2517 |
+
"step": 17550
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 5.45398202665014,
|
| 2521 |
+
"grad_norm": 19.539745330810547,
|
| 2522 |
+
"learning_rate": 1.5154426195640946e-05,
|
| 2523 |
+
"loss": 0.2799,
|
| 2524 |
+
"step": 17600
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 5.469476293771304,
|
| 2528 |
+
"grad_norm": 0.6790868639945984,
|
| 2529 |
+
"learning_rate": 1.5102778638570396e-05,
|
| 2530 |
+
"loss": 0.255,
|
| 2531 |
+
"step": 17650
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 5.48497056089247,
|
| 2535 |
+
"grad_norm": 10.917562484741211,
|
| 2536 |
+
"learning_rate": 1.5051131081499845e-05,
|
| 2537 |
+
"loss": 0.2371,
|
| 2538 |
+
"step": 17700
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 5.5004648280136355,
|
| 2542 |
+
"grad_norm": 4.378004550933838,
|
| 2543 |
+
"learning_rate": 1.4999483524429295e-05,
|
| 2544 |
+
"loss": 0.1697,
|
| 2545 |
+
"step": 17750
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 5.5159590951348,
|
| 2549 |
+
"grad_norm": 31.345949172973633,
|
| 2550 |
+
"learning_rate": 1.4947835967358745e-05,
|
| 2551 |
+
"loss": 0.2571,
|
| 2552 |
+
"step": 17800
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 5.531453362255966,
|
| 2556 |
+
"grad_norm": 4.986736297607422,
|
| 2557 |
+
"learning_rate": 1.4896188410288194e-05,
|
| 2558 |
+
"loss": 0.2366,
|
| 2559 |
+
"step": 17850
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 5.54694762937713,
|
| 2563 |
+
"grad_norm": 47.6024169921875,
|
| 2564 |
+
"learning_rate": 1.4844540853217642e-05,
|
| 2565 |
+
"loss": 0.1891,
|
| 2566 |
+
"step": 17900
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 5.562441896498296,
|
| 2570 |
+
"grad_norm": 30.087663650512695,
|
| 2571 |
+
"learning_rate": 1.4792893296147094e-05,
|
| 2572 |
+
"loss": 0.2772,
|
| 2573 |
+
"step": 17950
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 5.57793616361946,
|
| 2577 |
+
"grad_norm": 0.9499745965003967,
|
| 2578 |
+
"learning_rate": 1.4741245739076542e-05,
|
| 2579 |
+
"loss": 0.2005,
|
| 2580 |
+
"step": 18000
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 5.593430430740626,
|
| 2584 |
+
"grad_norm": 7.3200578689575195,
|
| 2585 |
+
"learning_rate": 1.4689598182005991e-05,
|
| 2586 |
+
"loss": 0.2846,
|
| 2587 |
+
"step": 18050
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 5.6089246978617915,
|
| 2591 |
+
"grad_norm": 47.46051025390625,
|
| 2592 |
+
"learning_rate": 1.4637950624935441e-05,
|
| 2593 |
+
"loss": 0.2339,
|
| 2594 |
+
"step": 18100
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 5.624418964982956,
|
| 2598 |
+
"grad_norm": 0.13382229208946228,
|
| 2599 |
+
"learning_rate": 1.458630306786489e-05,
|
| 2600 |
+
"loss": 0.2502,
|
| 2601 |
+
"step": 18150
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 5.639913232104122,
|
| 2605 |
+
"grad_norm": 36.12116622924805,
|
| 2606 |
+
"learning_rate": 1.453465551079434e-05,
|
| 2607 |
+
"loss": 0.2231,
|
| 2608 |
+
"step": 18200
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 5.655407499225286,
|
| 2612 |
+
"grad_norm": 0.8564043045043945,
|
| 2613 |
+
"learning_rate": 1.4483007953723788e-05,
|
| 2614 |
+
"loss": 0.2678,
|
| 2615 |
+
"step": 18250
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 5.670901766346452,
|
| 2619 |
+
"grad_norm": 5.724666118621826,
|
| 2620 |
+
"learning_rate": 1.4431360396653238e-05,
|
| 2621 |
+
"loss": 0.2279,
|
| 2622 |
+
"step": 18300
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 5.686396033467617,
|
| 2626 |
+
"grad_norm": 31.30683708190918,
|
| 2627 |
+
"learning_rate": 1.437971283958269e-05,
|
| 2628 |
+
"loss": 0.2131,
|
| 2629 |
+
"step": 18350
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 5.701890300588782,
|
| 2633 |
+
"grad_norm": 1.6962803602218628,
|
| 2634 |
+
"learning_rate": 1.4328065282512137e-05,
|
| 2635 |
+
"loss": 0.2052,
|
| 2636 |
+
"step": 18400
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 5.7173845677099475,
|
| 2640 |
+
"grad_norm": 15.251322746276855,
|
| 2641 |
+
"learning_rate": 1.4276417725441587e-05,
|
| 2642 |
+
"loss": 0.2358,
|
| 2643 |
+
"step": 18450
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 5.732878834831112,
|
| 2647 |
+
"grad_norm": 7.961274147033691,
|
| 2648 |
+
"learning_rate": 1.4224770168371035e-05,
|
| 2649 |
+
"loss": 0.2551,
|
| 2650 |
+
"step": 18500
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 5.748373101952278,
|
| 2654 |
+
"grad_norm": 1.336027979850769,
|
| 2655 |
+
"learning_rate": 1.4173122611300486e-05,
|
| 2656 |
+
"loss": 0.2878,
|
| 2657 |
+
"step": 18550
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 5.763867369073443,
|
| 2661 |
+
"grad_norm": 9.183156967163086,
|
| 2662 |
+
"learning_rate": 1.4121475054229936e-05,
|
| 2663 |
+
"loss": 0.2737,
|
| 2664 |
+
"step": 18600
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 5.779361636194608,
|
| 2668 |
+
"grad_norm": 15.864567756652832,
|
| 2669 |
+
"learning_rate": 1.4069827497159384e-05,
|
| 2670 |
+
"loss": 0.1805,
|
| 2671 |
+
"step": 18650
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 5.794855903315773,
|
| 2675 |
+
"grad_norm": 23.6713924407959,
|
| 2676 |
+
"learning_rate": 1.4018179940088834e-05,
|
| 2677 |
+
"loss": 0.2226,
|
| 2678 |
+
"step": 18700
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 5.810350170436938,
|
| 2682 |
+
"grad_norm": 44.18366241455078,
|
| 2683 |
+
"learning_rate": 1.3966532383018283e-05,
|
| 2684 |
+
"loss": 0.2275,
|
| 2685 |
+
"step": 18750
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 5.8258444375581036,
|
| 2689 |
+
"grad_norm": 2.445971965789795,
|
| 2690 |
+
"learning_rate": 1.3914884825947733e-05,
|
| 2691 |
+
"loss": 0.2182,
|
| 2692 |
+
"step": 18800
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 5.841338704679268,
|
| 2696 |
+
"grad_norm": 19.132349014282227,
|
| 2697 |
+
"learning_rate": 1.3863237268877183e-05,
|
| 2698 |
+
"loss": 0.1776,
|
| 2699 |
+
"step": 18850
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 5.856832971800434,
|
| 2703 |
+
"grad_norm": 14.281071662902832,
|
| 2704 |
+
"learning_rate": 1.381158971180663e-05,
|
| 2705 |
+
"loss": 0.2388,
|
| 2706 |
+
"step": 18900
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 5.872327238921599,
|
| 2710 |
+
"grad_norm": 9.464559555053711,
|
| 2711 |
+
"learning_rate": 1.3759942154736082e-05,
|
| 2712 |
+
"loss": 0.2955,
|
| 2713 |
+
"step": 18950
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 5.887821506042764,
|
| 2717 |
+
"grad_norm": 9.839083671569824,
|
| 2718 |
+
"learning_rate": 1.370829459766553e-05,
|
| 2719 |
+
"loss": 0.2653,
|
| 2720 |
+
"step": 19000
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 5.903315773163929,
|
| 2724 |
+
"grad_norm": 21.549816131591797,
|
| 2725 |
+
"learning_rate": 1.365664704059498e-05,
|
| 2726 |
+
"loss": 0.1973,
|
| 2727 |
+
"step": 19050
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 5.918810040285095,
|
| 2731 |
+
"grad_norm": 17.4051456451416,
|
| 2732 |
+
"learning_rate": 1.360499948352443e-05,
|
| 2733 |
+
"loss": 0.2111,
|
| 2734 |
+
"step": 19100
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 5.93430430740626,
|
| 2738 |
+
"grad_norm": 12.166346549987793,
|
| 2739 |
+
"learning_rate": 1.355335192645388e-05,
|
| 2740 |
+
"loss": 0.2649,
|
| 2741 |
+
"step": 19150
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 5.949798574527425,
|
| 2745 |
+
"grad_norm": 2.162867784500122,
|
| 2746 |
+
"learning_rate": 1.3501704369383329e-05,
|
| 2747 |
+
"loss": 0.1966,
|
| 2748 |
+
"step": 19200
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 5.96529284164859,
|
| 2752 |
+
"grad_norm": 19.864362716674805,
|
| 2753 |
+
"learning_rate": 1.3450056812312779e-05,
|
| 2754 |
+
"loss": 0.2891,
|
| 2755 |
+
"step": 19250
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 5.980787108769755,
|
| 2759 |
+
"grad_norm": 6.605401992797852,
|
| 2760 |
+
"learning_rate": 1.3398409255242227e-05,
|
| 2761 |
+
"loss": 0.2163,
|
| 2762 |
+
"step": 19300
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 5.99628137589092,
|
| 2766 |
+
"grad_norm": 0.3219479024410248,
|
| 2767 |
+
"learning_rate": 1.3346761698171678e-05,
|
| 2768 |
+
"loss": 0.2835,
|
| 2769 |
+
"step": 19350
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 6.0,
|
| 2773 |
+
"eval_accuracy": 0.7876825779250113,
|
| 2774 |
+
"eval_f1": 0.7883424547160977,
|
| 2775 |
+
"eval_loss": 0.9514418840408325,
|
| 2776 |
+
"eval_runtime": 25.521,
|
| 2777 |
+
"eval_samples_per_second": 260.217,
|
| 2778 |
+
"eval_steps_per_second": 16.3,
|
| 2779 |
+
"step": 19362
|
| 2780 |
+
},
|
| 2781 |
+
{
|
| 2782 |
+
"epoch": 6.0117756430120854,
|
| 2783 |
+
"grad_norm": 24.46638298034668,
|
| 2784 |
+
"learning_rate": 1.3295114141101126e-05,
|
| 2785 |
+
"loss": 0.1478,
|
| 2786 |
+
"step": 19400
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 6.027269910133251,
|
| 2790 |
+
"grad_norm": 22.210508346557617,
|
| 2791 |
+
"learning_rate": 1.3243466584030576e-05,
|
| 2792 |
+
"loss": 0.1154,
|
| 2793 |
+
"step": 19450
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 6.042764177254416,
|
| 2797 |
+
"grad_norm": 1.0985016822814941,
|
| 2798 |
+
"learning_rate": 1.3191819026960025e-05,
|
| 2799 |
+
"loss": 0.1032,
|
| 2800 |
+
"step": 19500
|
| 2801 |
+
},
|
| 2802 |
+
{
|
| 2803 |
+
"epoch": 6.058258444375581,
|
| 2804 |
+
"grad_norm": 18.337181091308594,
|
| 2805 |
+
"learning_rate": 1.3140171469889475e-05,
|
| 2806 |
+
"loss": 0.1289,
|
| 2807 |
+
"step": 19550
|
| 2808 |
+
},
|
| 2809 |
+
{
|
| 2810 |
+
"epoch": 6.073752711496746,
|
| 2811 |
+
"grad_norm": 16.683774948120117,
|
| 2812 |
+
"learning_rate": 1.3088523912818925e-05,
|
| 2813 |
+
"loss": 0.1279,
|
| 2814 |
+
"step": 19600
|
| 2815 |
+
},
|
| 2816 |
+
{
|
| 2817 |
+
"epoch": 6.089246978617911,
|
| 2818 |
+
"grad_norm": 27.718416213989258,
|
| 2819 |
+
"learning_rate": 1.3036876355748373e-05,
|
| 2820 |
+
"loss": 0.144,
|
| 2821 |
+
"step": 19650
|
| 2822 |
+
},
|
| 2823 |
+
{
|
| 2824 |
+
"epoch": 6.104741245739077,
|
| 2825 |
+
"grad_norm": 33.05785369873047,
|
| 2826 |
+
"learning_rate": 1.2985228798677822e-05,
|
| 2827 |
+
"loss": 0.1584,
|
| 2828 |
+
"step": 19700
|
| 2829 |
+
},
|
| 2830 |
+
{
|
| 2831 |
+
"epoch": 6.1202355128602415,
|
| 2832 |
+
"grad_norm": 0.15720854699611664,
|
| 2833 |
+
"learning_rate": 1.2933581241607274e-05,
|
| 2834 |
+
"loss": 0.1348,
|
| 2835 |
+
"step": 19750
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 6.135729779981407,
|
| 2839 |
+
"grad_norm": 0.5664418935775757,
|
| 2840 |
+
"learning_rate": 1.2881933684536722e-05,
|
| 2841 |
+
"loss": 0.1434,
|
| 2842 |
+
"step": 19800
|
| 2843 |
+
},
|
| 2844 |
+
{
|
| 2845 |
+
"epoch": 6.151224047102572,
|
| 2846 |
+
"grad_norm": 0.36847808957099915,
|
| 2847 |
+
"learning_rate": 1.2830286127466171e-05,
|
| 2848 |
+
"loss": 0.2238,
|
| 2849 |
+
"step": 19850
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 6.166718314223737,
|
| 2853 |
+
"grad_norm": 18.649120330810547,
|
| 2854 |
+
"learning_rate": 1.277863857039562e-05,
|
| 2855 |
+
"loss": 0.187,
|
| 2856 |
+
"step": 19900
|
| 2857 |
+
},
|
| 2858 |
+
{
|
| 2859 |
+
"epoch": 6.182212581344903,
|
| 2860 |
+
"grad_norm": 0.17821761965751648,
|
| 2861 |
+
"learning_rate": 1.272699101332507e-05,
|
| 2862 |
+
"loss": 0.1547,
|
| 2863 |
+
"step": 19950
|
| 2864 |
+
},
|
| 2865 |
+
{
|
| 2866 |
+
"epoch": 6.197706848466067,
|
| 2867 |
+
"grad_norm": 8.831562995910645,
|
| 2868 |
+
"learning_rate": 1.267534345625452e-05,
|
| 2869 |
+
"loss": 0.1539,
|
| 2870 |
+
"step": 20000
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 6.213201115587233,
|
| 2874 |
+
"grad_norm": 1.2256168127059937,
|
| 2875 |
+
"learning_rate": 1.2623695899183968e-05,
|
| 2876 |
+
"loss": 0.136,
|
| 2877 |
+
"step": 20050
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 6.2286953827083975,
|
| 2881 |
+
"grad_norm": 53.65285873413086,
|
| 2882 |
+
"learning_rate": 1.2572048342113418e-05,
|
| 2883 |
+
"loss": 0.1483,
|
| 2884 |
+
"step": 20100
|
| 2885 |
+
},
|
| 2886 |
+
{
|
| 2887 |
+
"epoch": 6.244189649829563,
|
| 2888 |
+
"grad_norm": 26.486663818359375,
|
| 2889 |
+
"learning_rate": 1.2520400785042868e-05,
|
| 2890 |
+
"loss": 0.1845,
|
| 2891 |
+
"step": 20150
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 6.259683916950729,
|
| 2895 |
+
"grad_norm": 5.685501575469971,
|
| 2896 |
+
"learning_rate": 1.2468753227972317e-05,
|
| 2897 |
+
"loss": 0.1549,
|
| 2898 |
+
"step": 20200
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"epoch": 6.275178184071893,
|
| 2902 |
+
"grad_norm": 38.161312103271484,
|
| 2903 |
+
"learning_rate": 1.2417105670901767e-05,
|
| 2904 |
+
"loss": 0.154,
|
| 2905 |
+
"step": 20250
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 6.290672451193059,
|
| 2909 |
+
"grad_norm": 8.335810661315918,
|
| 2910 |
+
"learning_rate": 1.2365458113831215e-05,
|
| 2911 |
+
"loss": 0.17,
|
| 2912 |
+
"step": 20300
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 6.306166718314223,
|
| 2916 |
+
"grad_norm": 0.48542681336402893,
|
| 2917 |
+
"learning_rate": 1.2313810556760666e-05,
|
| 2918 |
+
"loss": 0.1399,
|
| 2919 |
+
"step": 20350
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 6.321660985435389,
|
| 2923 |
+
"grad_norm": 58.425865173339844,
|
| 2924 |
+
"learning_rate": 1.2262162999690116e-05,
|
| 2925 |
+
"loss": 0.2096,
|
| 2926 |
+
"step": 20400
|
| 2927 |
+
},
|
| 2928 |
+
{
|
| 2929 |
+
"epoch": 6.337155252556554,
|
| 2930 |
+
"grad_norm": 1.698177456855774,
|
| 2931 |
+
"learning_rate": 1.2210515442619564e-05,
|
| 2932 |
+
"loss": 0.1646,
|
| 2933 |
+
"step": 20450
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 6.352649519677719,
|
| 2937 |
+
"grad_norm": 9.479565620422363,
|
| 2938 |
+
"learning_rate": 1.2158867885549014e-05,
|
| 2939 |
+
"loss": 0.1261,
|
| 2940 |
+
"step": 20500
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 6.368143786798885,
|
| 2944 |
+
"grad_norm": 29.26807975769043,
|
| 2945 |
+
"learning_rate": 1.2107220328478463e-05,
|
| 2946 |
+
"loss": 0.1258,
|
| 2947 |
+
"step": 20550
|
| 2948 |
+
},
|
| 2949 |
+
{
|
| 2950 |
+
"epoch": 6.383638053920049,
|
| 2951 |
+
"grad_norm": 17.38056182861328,
|
| 2952 |
+
"learning_rate": 1.2055572771407913e-05,
|
| 2953 |
+
"loss": 0.1663,
|
| 2954 |
+
"step": 20600
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 6.399132321041215,
|
| 2958 |
+
"grad_norm": 0.11625684797763824,
|
| 2959 |
+
"learning_rate": 1.2003925214337363e-05,
|
| 2960 |
+
"loss": 0.1632,
|
| 2961 |
+
"step": 20650
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 6.41462658816238,
|
| 2965 |
+
"grad_norm": 50.295257568359375,
|
| 2966 |
+
"learning_rate": 1.195227765726681e-05,
|
| 2967 |
+
"loss": 0.1805,
|
| 2968 |
+
"step": 20700
|
| 2969 |
+
},
|
| 2970 |
+
{
|
| 2971 |
+
"epoch": 6.430120855283545,
|
| 2972 |
+
"grad_norm": 8.677082061767578,
|
| 2973 |
+
"learning_rate": 1.1900630100196262e-05,
|
| 2974 |
+
"loss": 0.1145,
|
| 2975 |
+
"step": 20750
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 6.4456151224047105,
|
| 2979 |
+
"grad_norm": 25.766963958740234,
|
| 2980 |
+
"learning_rate": 1.184898254312571e-05,
|
| 2981 |
+
"loss": 0.179,
|
| 2982 |
+
"step": 20800
|
| 2983 |
+
},
|
| 2984 |
+
{
|
| 2985 |
+
"epoch": 6.461109389525875,
|
| 2986 |
+
"grad_norm": 9.217619895935059,
|
| 2987 |
+
"learning_rate": 1.179733498605516e-05,
|
| 2988 |
+
"loss": 0.1661,
|
| 2989 |
+
"step": 20850
|
| 2990 |
+
},
|
| 2991 |
+
{
|
| 2992 |
+
"epoch": 6.476603656647041,
|
| 2993 |
+
"grad_norm": 26.130367279052734,
|
| 2994 |
+
"learning_rate": 1.174568742898461e-05,
|
| 2995 |
+
"loss": 0.1842,
|
| 2996 |
+
"step": 20900
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 6.492097923768206,
|
| 3000 |
+
"grad_norm": 0.5569332838058472,
|
| 3001 |
+
"learning_rate": 1.169403987191406e-05,
|
| 3002 |
+
"loss": 0.1311,
|
| 3003 |
+
"step": 20950
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 6.507592190889371,
|
| 3007 |
+
"grad_norm": 24.217166900634766,
|
| 3008 |
+
"learning_rate": 1.1642392314843509e-05,
|
| 3009 |
+
"loss": 0.1402,
|
| 3010 |
+
"step": 21000
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"epoch": 6.523086458010536,
|
| 3014 |
+
"grad_norm": 0.9757774472236633,
|
| 3015 |
+
"learning_rate": 1.1590744757772957e-05,
|
| 3016 |
+
"loss": 0.151,
|
| 3017 |
+
"step": 21050
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 6.538580725131701,
|
| 3021 |
+
"grad_norm": 1.751021146774292,
|
| 3022 |
+
"learning_rate": 1.1539097200702407e-05,
|
| 3023 |
+
"loss": 0.2169,
|
| 3024 |
+
"step": 21100
|
| 3025 |
+
},
|
| 3026 |
+
{
|
| 3027 |
+
"epoch": 6.5540749922528665,
|
| 3028 |
+
"grad_norm": 1.1032994985580444,
|
| 3029 |
+
"learning_rate": 1.1487449643631858e-05,
|
| 3030 |
+
"loss": 0.1998,
|
| 3031 |
+
"step": 21150
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 6.569569259374031,
|
| 3035 |
+
"grad_norm": 7.012996196746826,
|
| 3036 |
+
"learning_rate": 1.1435802086561306e-05,
|
| 3037 |
+
"loss": 0.1461,
|
| 3038 |
+
"step": 21200
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 6.585063526495197,
|
| 3042 |
+
"grad_norm": 1.7933677434921265,
|
| 3043 |
+
"learning_rate": 1.1384154529490756e-05,
|
| 3044 |
+
"loss": 0.2137,
|
| 3045 |
+
"step": 21250
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 6.600557793616362,
|
| 3049 |
+
"grad_norm": 0.28755664825439453,
|
| 3050 |
+
"learning_rate": 1.1332506972420204e-05,
|
| 3051 |
+
"loss": 0.1666,
|
| 3052 |
+
"step": 21300
|
| 3053 |
+
},
|
| 3054 |
+
{
|
| 3055 |
+
"epoch": 6.616052060737527,
|
| 3056 |
+
"grad_norm": 69.47583770751953,
|
| 3057 |
+
"learning_rate": 1.1280859415349655e-05,
|
| 3058 |
+
"loss": 0.1056,
|
| 3059 |
+
"step": 21350
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 6.631546327858692,
|
| 3063 |
+
"grad_norm": 59.0959587097168,
|
| 3064 |
+
"learning_rate": 1.1229211858279105e-05,
|
| 3065 |
+
"loss": 0.2245,
|
| 3066 |
+
"step": 21400
|
| 3067 |
+
},
|
| 3068 |
+
{
|
| 3069 |
+
"epoch": 6.647040594979858,
|
| 3070 |
+
"grad_norm": 9.836857795715332,
|
| 3071 |
+
"learning_rate": 1.1177564301208553e-05,
|
| 3072 |
+
"loss": 0.1524,
|
| 3073 |
+
"step": 21450
|
| 3074 |
+
},
|
| 3075 |
+
{
|
| 3076 |
+
"epoch": 6.6625348621010225,
|
| 3077 |
+
"grad_norm": 26.670324325561523,
|
| 3078 |
+
"learning_rate": 1.1125916744138002e-05,
|
| 3079 |
+
"loss": 0.1188,
|
| 3080 |
+
"step": 21500
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 6.678029129222188,
|
| 3084 |
+
"grad_norm": 62.01319122314453,
|
| 3085 |
+
"learning_rate": 1.1074269187067452e-05,
|
| 3086 |
+
"loss": 0.1745,
|
| 3087 |
+
"step": 21550
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 6.693523396343353,
|
| 3091 |
+
"grad_norm": 0.7209033966064453,
|
| 3092 |
+
"learning_rate": 1.1022621629996902e-05,
|
| 3093 |
+
"loss": 0.1847,
|
| 3094 |
+
"step": 21600
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 6.709017663464518,
|
| 3098 |
+
"grad_norm": 62.61027145385742,
|
| 3099 |
+
"learning_rate": 1.0970974072926351e-05,
|
| 3100 |
+
"loss": 0.1954,
|
| 3101 |
+
"step": 21650
|
| 3102 |
+
},
|
| 3103 |
+
{
|
| 3104 |
+
"epoch": 6.724511930585683,
|
| 3105 |
+
"grad_norm": 27.388124465942383,
|
| 3106 |
+
"learning_rate": 1.09193265158558e-05,
|
| 3107 |
+
"loss": 0.1611,
|
| 3108 |
+
"step": 21700
|
| 3109 |
+
},
|
| 3110 |
+
{
|
| 3111 |
+
"epoch": 6.740006197706848,
|
| 3112 |
+
"grad_norm": 0.4555682837963104,
|
| 3113 |
+
"learning_rate": 1.086767895878525e-05,
|
| 3114 |
+
"loss": 0.2174,
|
| 3115 |
+
"step": 21750
|
| 3116 |
+
},
|
| 3117 |
+
{
|
| 3118 |
+
"epoch": 6.755500464828014,
|
| 3119 |
+
"grad_norm": 3.9981863498687744,
|
| 3120 |
+
"learning_rate": 1.08160314017147e-05,
|
| 3121 |
+
"loss": 0.1812,
|
| 3122 |
+
"step": 21800
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 6.770994731949179,
|
| 3126 |
+
"grad_norm": 34.40354919433594,
|
| 3127 |
+
"learning_rate": 1.0764383844644148e-05,
|
| 3128 |
+
"loss": 0.1399,
|
| 3129 |
+
"step": 21850
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 6.786488999070344,
|
| 3133 |
+
"grad_norm": 14.363764762878418,
|
| 3134 |
+
"learning_rate": 1.0712736287573598e-05,
|
| 3135 |
+
"loss": 0.1776,
|
| 3136 |
+
"step": 21900
|
| 3137 |
+
},
|
| 3138 |
+
{
|
| 3139 |
+
"epoch": 6.80198326619151,
|
| 3140 |
+
"grad_norm": 30.52642822265625,
|
| 3141 |
+
"learning_rate": 1.0661088730503048e-05,
|
| 3142 |
+
"loss": 0.1588,
|
| 3143 |
+
"step": 21950
|
| 3144 |
+
},
|
| 3145 |
+
{
|
| 3146 |
+
"epoch": 6.817477533312674,
|
| 3147 |
+
"grad_norm": 14.077266693115234,
|
| 3148 |
+
"learning_rate": 1.0609441173432497e-05,
|
| 3149 |
+
"loss": 0.1954,
|
| 3150 |
+
"step": 22000
|
| 3151 |
+
},
|
| 3152 |
+
{
|
| 3153 |
+
"epoch": 6.83297180043384,
|
| 3154 |
+
"grad_norm": 18.22222328186035,
|
| 3155 |
+
"learning_rate": 1.0557793616361947e-05,
|
| 3156 |
+
"loss": 0.1691,
|
| 3157 |
+
"step": 22050
|
| 3158 |
+
},
|
| 3159 |
+
{
|
| 3160 |
+
"epoch": 6.848466067555004,
|
| 3161 |
+
"grad_norm": 3.799182176589966,
|
| 3162 |
+
"learning_rate": 1.0506146059291395e-05,
|
| 3163 |
+
"loss": 0.1695,
|
| 3164 |
+
"step": 22100
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"epoch": 6.86396033467617,
|
| 3168 |
+
"grad_norm": 22.45465087890625,
|
| 3169 |
+
"learning_rate": 1.0454498502220846e-05,
|
| 3170 |
+
"loss": 0.1558,
|
| 3171 |
+
"step": 22150
|
| 3172 |
+
},
|
| 3173 |
+
{
|
| 3174 |
+
"epoch": 6.879454601797335,
|
| 3175 |
+
"grad_norm": 42.03032302856445,
|
| 3176 |
+
"learning_rate": 1.0402850945150294e-05,
|
| 3177 |
+
"loss": 0.1663,
|
| 3178 |
+
"step": 22200
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 6.8949488689185,
|
| 3182 |
+
"grad_norm": 0.39429527521133423,
|
| 3183 |
+
"learning_rate": 1.0351203388079744e-05,
|
| 3184 |
+
"loss": 0.1494,
|
| 3185 |
+
"step": 22250
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 6.910443136039666,
|
| 3189 |
+
"grad_norm": 18.078174591064453,
|
| 3190 |
+
"learning_rate": 1.0299555831009194e-05,
|
| 3191 |
+
"loss": 0.1776,
|
| 3192 |
+
"step": 22300
|
| 3193 |
+
},
|
| 3194 |
+
{
|
| 3195 |
+
"epoch": 6.92593740316083,
|
| 3196 |
+
"grad_norm": 0.8698641061782837,
|
| 3197 |
+
"learning_rate": 1.0247908273938643e-05,
|
| 3198 |
+
"loss": 0.1575,
|
| 3199 |
+
"step": 22350
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 6.941431670281996,
|
| 3203 |
+
"grad_norm": 12.034942626953125,
|
| 3204 |
+
"learning_rate": 1.0196260716868093e-05,
|
| 3205 |
+
"loss": 0.15,
|
| 3206 |
+
"step": 22400
|
| 3207 |
+
},
|
| 3208 |
+
{
|
| 3209 |
+
"epoch": 6.9569259374031605,
|
| 3210 |
+
"grad_norm": 18.04237937927246,
|
| 3211 |
+
"learning_rate": 1.0144613159797541e-05,
|
| 3212 |
+
"loss": 0.1581,
|
| 3213 |
+
"step": 22450
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 6.972420204524326,
|
| 3217 |
+
"grad_norm": 11.662799835205078,
|
| 3218 |
+
"learning_rate": 1.009296560272699e-05,
|
| 3219 |
+
"loss": 0.1876,
|
| 3220 |
+
"step": 22500
|
| 3221 |
+
},
|
| 3222 |
+
{
|
| 3223 |
+
"epoch": 6.9879144716454915,
|
| 3224 |
+
"grad_norm": 13.070639610290527,
|
| 3225 |
+
"learning_rate": 1.0041318045656442e-05,
|
| 3226 |
+
"loss": 0.2079,
|
| 3227 |
+
"step": 22550
|
| 3228 |
+
},
|
| 3229 |
+
{
|
| 3230 |
+
"epoch": 7.0,
|
| 3231 |
+
"eval_accuracy": 0.7888872157807559,
|
| 3232 |
+
"eval_f1": 0.7894450189158516,
|
| 3233 |
+
"eval_loss": 1.1703307628631592,
|
| 3234 |
+
"eval_runtime": 25.562,
|
| 3235 |
+
"eval_samples_per_second": 259.799,
|
| 3236 |
+
"eval_steps_per_second": 16.274,
|
| 3237 |
+
"step": 22589
|
| 3238 |
+
}
|
| 3239 |
+
],
|
| 3240 |
+
"logging_steps": 50,
|
| 3241 |
+
"max_steps": 32270,
|
| 3242 |
+
"num_input_tokens_seen": 0,
|
| 3243 |
+
"num_train_epochs": 10,
|
| 3244 |
+
"save_steps": 500,
|
| 3245 |
+
"stateful_callbacks": {
|
| 3246 |
+
"EarlyStoppingCallback": {
|
| 3247 |
+
"args": {
|
| 3248 |
+
"early_stopping_patience": 2,
|
| 3249 |
+
"early_stopping_threshold": 0.0
|
| 3250 |
+
},
|
| 3251 |
+
"attributes": {
|
| 3252 |
+
"early_stopping_patience_counter": 2
|
| 3253 |
+
}
|
| 3254 |
+
},
|
| 3255 |
+
"TrainerControl": {
|
| 3256 |
+
"args": {
|
| 3257 |
+
"should_epoch_stop": false,
|
| 3258 |
+
"should_evaluate": false,
|
| 3259 |
+
"should_log": false,
|
| 3260 |
+
"should_save": true,
|
| 3261 |
+
"should_training_stop": true
|
| 3262 |
+
},
|
| 3263 |
+
"attributes": {}
|
| 3264 |
+
}
|
| 3265 |
+
},
|
| 3266 |
+
"total_flos": 4.754719082566656e+16,
|
| 3267 |
+
"train_batch_size": 16,
|
| 3268 |
+
"trial_name": null,
|
| 3269 |
+
"trial_params": null
|
| 3270 |
+
}
|
checkpoint-22589/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10df86f20d44dbad5bcbd0936a460173513fcbc2b606339a5e32ef490cc53c86
|
| 3 |
+
size 5216
|
config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ElectraForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"embedding_size": 768,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "\uae30\uc068",
|
| 13 |
+
"1": "\ub2f9\ud669",
|
| 14 |
+
"2": "\ubd84\ub178",
|
| 15 |
+
"3": "\ubd88\uc548",
|
| 16 |
+
"4": "\uc0c1\ucc98",
|
| 17 |
+
"5": "\uc2ac\ud514"
|
| 18 |
+
},
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 3072,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"\uae30\uc068": 0,
|
| 23 |
+
"\ub2f9\ud669": 1,
|
| 24 |
+
"\ubd84\ub178": 2,
|
| 25 |
+
"\ubd88\uc548": 3,
|
| 26 |
+
"\uc0c1\ucc98": 4,
|
| 27 |
+
"\uc2ac\ud514": 5
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-12,
|
| 30 |
+
"max_position_embeddings": 512,
|
| 31 |
+
"model_type": "electra",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"position_embedding_type": "absolute",
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"summary_activation": "gelu",
|
| 38 |
+
"summary_last_dropout": 0.1,
|
| 39 |
+
"summary_type": "first",
|
| 40 |
+
"summary_use_proj": true,
|
| 41 |
+
"tokenizer_class": "BertTokenizer",
|
| 42 |
+
"torch_dtype": "float32",
|
| 43 |
+
"transformers_version": "4.52.2",
|
| 44 |
+
"type_vocab_size": 2,
|
| 45 |
+
"use_cache": true,
|
| 46 |
+
"vocab_size": 54343
|
| 47 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:149e0f950d5f1601cc510fbd6b354de2f1e128f9f787c2d32ed18a6ec9b5f743
|
| 3 |
+
size 511149672
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"never_split": null,
|
| 52 |
+
"pad_token": "[PAD]",
|
| 53 |
+
"sep_token": "[SEP]",
|
| 54 |
+
"strip_accents": null,
|
| 55 |
+
"tokenize_chinese_chars": true,
|
| 56 |
+
"tokenizer_class": "BertTokenizer",
|
| 57 |
+
"unk_token": "[UNK]"
|
| 58 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10df86f20d44dbad5bcbd0936a460173513fcbc2b606339a5e32ef490cc53c86
|
| 3 |
+
size 5216
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|