Alex Spangher
commited on
Commit
•
60398e6
1
Parent(s):
ada6f54
- all_results.json +14 -0
- config.json +78 -0
- model.safetensors +3 -0
- post-training eval_results.json +8 -0
- prediction_output.jsonl +2 -0
- train_results.json +8 -0
- trainer_state.json +3630 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_samples": 6482,
|
| 4 |
+
"test_f1": 0.43183643890513856,
|
| 5 |
+
"test_loss": 2.9971516132354736,
|
| 6 |
+
"test_runtime": 46.1838,
|
| 7 |
+
"test_samples_per_second": 140.352,
|
| 8 |
+
"test_steps_per_second": 8.791,
|
| 9 |
+
"train_loss": 0.5760954068351747,
|
| 10 |
+
"train_runtime": 42398.6115,
|
| 11 |
+
"train_samples": 96270,
|
| 12 |
+
"train_samples_per_second": 22.706,
|
| 13 |
+
"train_steps_per_second": 2.838
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/bart-base",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"add_bias_logits": false,
|
| 6 |
+
"add_final_layer_norm": false,
|
| 7 |
+
"architectures": [
|
| 8 |
+
"BartForSequenceClassification"
|
| 9 |
+
],
|
| 10 |
+
"attention_dropout": 0.1,
|
| 11 |
+
"bos_token_id": 0,
|
| 12 |
+
"classif_dropout": 0.1,
|
| 13 |
+
"classifier_dropout": 0.0,
|
| 14 |
+
"d_model": 768,
|
| 15 |
+
"decoder_attention_heads": 12,
|
| 16 |
+
"decoder_ffn_dim": 3072,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 6,
|
| 19 |
+
"decoder_start_token_id": 2,
|
| 20 |
+
"dropout": 0.1,
|
| 21 |
+
"early_stopping": true,
|
| 22 |
+
"encoder_attention_heads": 12,
|
| 23 |
+
"encoder_ffn_dim": 3072,
|
| 24 |
+
"encoder_layerdrop": 0.0,
|
| 25 |
+
"encoder_layers": 6,
|
| 26 |
+
"eos_token_id": 2,
|
| 27 |
+
"forced_bos_token_id": 0,
|
| 28 |
+
"forced_eos_token_id": 2,
|
| 29 |
+
"gradient_checkpointing": false,
|
| 30 |
+
"id2label": {
|
| 31 |
+
"0": "Discuss",
|
| 32 |
+
"1": "Affirm",
|
| 33 |
+
"2": "Neutral",
|
| 34 |
+
"3": "Refute"
|
| 35 |
+
},
|
| 36 |
+
"init_std": 0.02,
|
| 37 |
+
"is_encoder_decoder": true,
|
| 38 |
+
"label2id": {
|
| 39 |
+
"Affirm": 1,
|
| 40 |
+
"Discuss": 0,
|
| 41 |
+
"Neutral": 2,
|
| 42 |
+
"Refute": 3
|
| 43 |
+
},
|
| 44 |
+
"max_position_embeddings": 1024,
|
| 45 |
+
"model_type": "bart",
|
| 46 |
+
"no_repeat_ngram_size": 3,
|
| 47 |
+
"normalize_before": false,
|
| 48 |
+
"normalize_embedding": true,
|
| 49 |
+
"num_beams": 4,
|
| 50 |
+
"num_hidden_layers": 6,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"problem_type": "single_label_classification",
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"task_specific_params": {
|
| 55 |
+
"summarization": {
|
| 56 |
+
"length_penalty": 1.0,
|
| 57 |
+
"max_length": 128,
|
| 58 |
+
"min_length": 12,
|
| 59 |
+
"num_beams": 4
|
| 60 |
+
},
|
| 61 |
+
"summarization_cnn": {
|
| 62 |
+
"length_penalty": 2.0,
|
| 63 |
+
"max_length": 142,
|
| 64 |
+
"min_length": 56,
|
| 65 |
+
"num_beams": 4
|
| 66 |
+
},
|
| 67 |
+
"summarization_xsum": {
|
| 68 |
+
"length_penalty": 1.0,
|
| 69 |
+
"max_length": 62,
|
| 70 |
+
"min_length": 11,
|
| 71 |
+
"num_beams": 6
|
| 72 |
+
}
|
| 73 |
+
},
|
| 74 |
+
"torch_dtype": "float32",
|
| 75 |
+
"transformers_version": "4.36.0.dev0",
|
| 76 |
+
"use_cache": true,
|
| 77 |
+
"vocab_size": 50265
|
| 78 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:856099112bbdfeb887703c7f6a078073458af865d24e6aea2c9d13d7519c3af1
|
| 3 |
+
size 560086544
|
post-training eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eval_samples": 6482,
|
| 3 |
+
"test_f1": 0.43183643890513856,
|
| 4 |
+
"test_loss": 2.9971516132354736,
|
| 5 |
+
"test_runtime": 46.1838,
|
| 6 |
+
"test_samples_per_second": 140.352,
|
| 7 |
+
"test_steps_per_second": 8.791
|
| 8 |
+
}
|
prediction_output.jsonl
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[{"pred": -1.5876611471176147, "label": 1.0}]
|
| 2 |
+
[{"pred": -0.018212709575891495, "label": 0.0}]
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 0.5760954068351747,
|
| 4 |
+
"train_runtime": 42398.6115,
|
| 5 |
+
"train_samples": 96270,
|
| 6 |
+
"train_samples_per_second": 22.706,
|
| 7 |
+
"train_steps_per_second": 2.838
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3630 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 120340,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.04,
|
| 13 |
+
"learning_rate": 4.979225527671597e-05,
|
| 14 |
+
"loss": 1.2637,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"eval_f1": 0.2863453369025287,
|
| 20 |
+
"eval_loss": 1.204175591468811,
|
| 21 |
+
"eval_runtime": 51.6606,
|
| 22 |
+
"eval_samples_per_second": 125.473,
|
| 23 |
+
"eval_steps_per_second": 7.859,
|
| 24 |
+
"step": 500
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.08,
|
| 28 |
+
"learning_rate": 4.958451055343195e-05,
|
| 29 |
+
"loss": 1.213,
|
| 30 |
+
"step": 1000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.08,
|
| 34 |
+
"eval_f1": 0.274772728922474,
|
| 35 |
+
"eval_loss": 1.2542401552200317,
|
| 36 |
+
"eval_runtime": 46.4286,
|
| 37 |
+
"eval_samples_per_second": 139.612,
|
| 38 |
+
"eval_steps_per_second": 8.745,
|
| 39 |
+
"step": 1000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.12,
|
| 43 |
+
"learning_rate": 4.937676583014792e-05,
|
| 44 |
+
"loss": 1.1852,
|
| 45 |
+
"step": 1500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.12,
|
| 49 |
+
"eval_f1": 0.31238188693470137,
|
| 50 |
+
"eval_loss": 1.1441909074783325,
|
| 51 |
+
"eval_runtime": 46.2394,
|
| 52 |
+
"eval_samples_per_second": 140.183,
|
| 53 |
+
"eval_steps_per_second": 8.78,
|
| 54 |
+
"step": 1500
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.17,
|
| 58 |
+
"learning_rate": 4.9169021106863886e-05,
|
| 59 |
+
"loss": 1.1495,
|
| 60 |
+
"step": 2000
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.17,
|
| 64 |
+
"eval_f1": 0.3303277428471857,
|
| 65 |
+
"eval_loss": 1.19601309299469,
|
| 66 |
+
"eval_runtime": 46.3042,
|
| 67 |
+
"eval_samples_per_second": 139.987,
|
| 68 |
+
"eval_steps_per_second": 8.768,
|
| 69 |
+
"step": 2000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 4.8961276383579855e-05,
|
| 74 |
+
"loss": 1.1406,
|
| 75 |
+
"step": 2500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.21,
|
| 79 |
+
"eval_f1": 0.33363478480748904,
|
| 80 |
+
"eval_loss": 1.2050482034683228,
|
| 81 |
+
"eval_runtime": 46.0365,
|
| 82 |
+
"eval_samples_per_second": 140.801,
|
| 83 |
+
"eval_steps_per_second": 8.819,
|
| 84 |
+
"step": 2500
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.25,
|
| 88 |
+
"learning_rate": 4.875353166029583e-05,
|
| 89 |
+
"loss": 1.1332,
|
| 90 |
+
"step": 3000
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.25,
|
| 94 |
+
"eval_f1": 0.35687474820142895,
|
| 95 |
+
"eval_loss": 1.1201504468917847,
|
| 96 |
+
"eval_runtime": 45.6751,
|
| 97 |
+
"eval_samples_per_second": 141.915,
|
| 98 |
+
"eval_steps_per_second": 8.889,
|
| 99 |
+
"step": 3000
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.29,
|
| 103 |
+
"learning_rate": 4.85457869370118e-05,
|
| 104 |
+
"loss": 1.1007,
|
| 105 |
+
"step": 3500
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.29,
|
| 109 |
+
"eval_f1": 0.3605293337176752,
|
| 110 |
+
"eval_loss": 1.1953203678131104,
|
| 111 |
+
"eval_runtime": 45.2576,
|
| 112 |
+
"eval_samples_per_second": 143.225,
|
| 113 |
+
"eval_steps_per_second": 8.971,
|
| 114 |
+
"step": 3500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.33,
|
| 118 |
+
"learning_rate": 4.833804221372778e-05,
|
| 119 |
+
"loss": 1.1157,
|
| 120 |
+
"step": 4000
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.33,
|
| 124 |
+
"eval_f1": 0.3862175225700333,
|
| 125 |
+
"eval_loss": 1.1009345054626465,
|
| 126 |
+
"eval_runtime": 46.0236,
|
| 127 |
+
"eval_samples_per_second": 140.841,
|
| 128 |
+
"eval_steps_per_second": 8.822,
|
| 129 |
+
"step": 4000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.37,
|
| 133 |
+
"learning_rate": 4.8130297490443746e-05,
|
| 134 |
+
"loss": 1.1172,
|
| 135 |
+
"step": 4500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.37,
|
| 139 |
+
"eval_f1": 0.3819750143486552,
|
| 140 |
+
"eval_loss": 1.126935362815857,
|
| 141 |
+
"eval_runtime": 46.0628,
|
| 142 |
+
"eval_samples_per_second": 140.721,
|
| 143 |
+
"eval_steps_per_second": 8.814,
|
| 144 |
+
"step": 4500
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.42,
|
| 148 |
+
"learning_rate": 4.7922552767159715e-05,
|
| 149 |
+
"loss": 1.1041,
|
| 150 |
+
"step": 5000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.42,
|
| 154 |
+
"eval_f1": 0.29887014740598117,
|
| 155 |
+
"eval_loss": 1.1668146848678589,
|
| 156 |
+
"eval_runtime": 46.4347,
|
| 157 |
+
"eval_samples_per_second": 139.594,
|
| 158 |
+
"eval_steps_per_second": 8.743,
|
| 159 |
+
"step": 5000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.46,
|
| 163 |
+
"learning_rate": 4.7714808043875684e-05,
|
| 164 |
+
"loss": 1.102,
|
| 165 |
+
"step": 5500
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.46,
|
| 169 |
+
"eval_f1": 0.418642853322715,
|
| 170 |
+
"eval_loss": 1.1066502332687378,
|
| 171 |
+
"eval_runtime": 46.0383,
|
| 172 |
+
"eval_samples_per_second": 140.796,
|
| 173 |
+
"eval_steps_per_second": 8.819,
|
| 174 |
+
"step": 5500
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.5,
|
| 178 |
+
"learning_rate": 4.750706332059166e-05,
|
| 179 |
+
"loss": 1.0878,
|
| 180 |
+
"step": 6000
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 0.5,
|
| 184 |
+
"eval_f1": 0.3200899667689199,
|
| 185 |
+
"eval_loss": 1.1729530096054077,
|
| 186 |
+
"eval_runtime": 46.8484,
|
| 187 |
+
"eval_samples_per_second": 138.361,
|
| 188 |
+
"eval_steps_per_second": 8.666,
|
| 189 |
+
"step": 6000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.54,
|
| 193 |
+
"learning_rate": 4.729931859730763e-05,
|
| 194 |
+
"loss": 1.0866,
|
| 195 |
+
"step": 6500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.54,
|
| 199 |
+
"eval_f1": 0.38888205294003,
|
| 200 |
+
"eval_loss": 1.108739972114563,
|
| 201 |
+
"eval_runtime": 46.5428,
|
| 202 |
+
"eval_samples_per_second": 139.27,
|
| 203 |
+
"eval_steps_per_second": 8.723,
|
| 204 |
+
"step": 6500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.58,
|
| 208 |
+
"learning_rate": 4.7091573874023606e-05,
|
| 209 |
+
"loss": 1.0729,
|
| 210 |
+
"step": 7000
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.58,
|
| 214 |
+
"eval_f1": 0.335409883600229,
|
| 215 |
+
"eval_loss": 1.1224578619003296,
|
| 216 |
+
"eval_runtime": 46.3948,
|
| 217 |
+
"eval_samples_per_second": 139.714,
|
| 218 |
+
"eval_steps_per_second": 8.751,
|
| 219 |
+
"step": 7000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.62,
|
| 223 |
+
"learning_rate": 4.6883829150739575e-05,
|
| 224 |
+
"loss": 1.0684,
|
| 225 |
+
"step": 7500
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.62,
|
| 229 |
+
"eval_f1": 0.4087602196471012,
|
| 230 |
+
"eval_loss": 1.1329175233840942,
|
| 231 |
+
"eval_runtime": 46.5357,
|
| 232 |
+
"eval_samples_per_second": 139.291,
|
| 233 |
+
"eval_steps_per_second": 8.724,
|
| 234 |
+
"step": 7500
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.66,
|
| 238 |
+
"learning_rate": 4.6676084427455544e-05,
|
| 239 |
+
"loss": 1.0633,
|
| 240 |
+
"step": 8000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.66,
|
| 244 |
+
"eval_f1": 0.39803057267400827,
|
| 245 |
+
"eval_loss": 1.1004011631011963,
|
| 246 |
+
"eval_runtime": 46.1396,
|
| 247 |
+
"eval_samples_per_second": 140.487,
|
| 248 |
+
"eval_steps_per_second": 8.799,
|
| 249 |
+
"step": 8000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.71,
|
| 253 |
+
"learning_rate": 4.6468339704171513e-05,
|
| 254 |
+
"loss": 1.0739,
|
| 255 |
+
"step": 8500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.71,
|
| 259 |
+
"eval_f1": 0.39430507046678914,
|
| 260 |
+
"eval_loss": 1.090652585029602,
|
| 261 |
+
"eval_runtime": 46.3294,
|
| 262 |
+
"eval_samples_per_second": 139.911,
|
| 263 |
+
"eval_steps_per_second": 8.763,
|
| 264 |
+
"step": 8500
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.75,
|
| 268 |
+
"learning_rate": 4.626059498088749e-05,
|
| 269 |
+
"loss": 1.0646,
|
| 270 |
+
"step": 9000
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.75,
|
| 274 |
+
"eval_f1": 0.4204523467443143,
|
| 275 |
+
"eval_loss": 1.1204614639282227,
|
| 276 |
+
"eval_runtime": 46.2687,
|
| 277 |
+
"eval_samples_per_second": 140.095,
|
| 278 |
+
"eval_steps_per_second": 8.775,
|
| 279 |
+
"step": 9000
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.79,
|
| 283 |
+
"learning_rate": 4.605285025760346e-05,
|
| 284 |
+
"loss": 1.0581,
|
| 285 |
+
"step": 9500
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.79,
|
| 289 |
+
"eval_f1": 0.3934072615715215,
|
| 290 |
+
"eval_loss": 1.100487232208252,
|
| 291 |
+
"eval_runtime": 46.7591,
|
| 292 |
+
"eval_samples_per_second": 138.625,
|
| 293 |
+
"eval_steps_per_second": 8.683,
|
| 294 |
+
"step": 9500
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.83,
|
| 298 |
+
"learning_rate": 4.5845105534319435e-05,
|
| 299 |
+
"loss": 1.0659,
|
| 300 |
+
"step": 10000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 0.83,
|
| 304 |
+
"eval_f1": 0.3959022659523447,
|
| 305 |
+
"eval_loss": 1.0948997735977173,
|
| 306 |
+
"eval_runtime": 46.4254,
|
| 307 |
+
"eval_samples_per_second": 139.622,
|
| 308 |
+
"eval_steps_per_second": 8.745,
|
| 309 |
+
"step": 10000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.87,
|
| 313 |
+
"learning_rate": 4.5637360811035404e-05,
|
| 314 |
+
"loss": 1.0573,
|
| 315 |
+
"step": 10500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.87,
|
| 319 |
+
"eval_f1": 0.4038876072579517,
|
| 320 |
+
"eval_loss": 1.0948611497879028,
|
| 321 |
+
"eval_runtime": 46.9168,
|
| 322 |
+
"eval_samples_per_second": 138.159,
|
| 323 |
+
"eval_steps_per_second": 8.654,
|
| 324 |
+
"step": 10500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.91,
|
| 328 |
+
"learning_rate": 4.542961608775137e-05,
|
| 329 |
+
"loss": 1.0725,
|
| 330 |
+
"step": 11000
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.91,
|
| 334 |
+
"eval_f1": 0.3986235498061616,
|
| 335 |
+
"eval_loss": 1.1076061725616455,
|
| 336 |
+
"eval_runtime": 46.337,
|
| 337 |
+
"eval_samples_per_second": 139.888,
|
| 338 |
+
"eval_steps_per_second": 8.762,
|
| 339 |
+
"step": 11000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.96,
|
| 343 |
+
"learning_rate": 4.522187136446734e-05,
|
| 344 |
+
"loss": 1.0453,
|
| 345 |
+
"step": 11500
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.96,
|
| 349 |
+
"eval_f1": 0.39752123118599975,
|
| 350 |
+
"eval_loss": 1.0838735103607178,
|
| 351 |
+
"eval_runtime": 46.3441,
|
| 352 |
+
"eval_samples_per_second": 139.867,
|
| 353 |
+
"eval_steps_per_second": 8.761,
|
| 354 |
+
"step": 11500
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.0,
|
| 358 |
+
"learning_rate": 4.501412664118332e-05,
|
| 359 |
+
"loss": 1.0594,
|
| 360 |
+
"step": 12000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.0,
|
| 364 |
+
"eval_f1": 0.3901284493425606,
|
| 365 |
+
"eval_loss": 1.0847594738006592,
|
| 366 |
+
"eval_runtime": 45.6188,
|
| 367 |
+
"eval_samples_per_second": 142.091,
|
| 368 |
+
"eval_steps_per_second": 8.9,
|
| 369 |
+
"step": 12000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.04,
|
| 373 |
+
"learning_rate": 4.480638191789929e-05,
|
| 374 |
+
"loss": 0.9487,
|
| 375 |
+
"step": 12500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.04,
|
| 379 |
+
"eval_f1": 0.43667878762469675,
|
| 380 |
+
"eval_loss": 1.1431002616882324,
|
| 381 |
+
"eval_runtime": 46.1871,
|
| 382 |
+
"eval_samples_per_second": 140.342,
|
| 383 |
+
"eval_steps_per_second": 8.79,
|
| 384 |
+
"step": 12500
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.08,
|
| 388 |
+
"learning_rate": 4.4598637194615264e-05,
|
| 389 |
+
"loss": 0.9704,
|
| 390 |
+
"step": 13000
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 1.08,
|
| 394 |
+
"eval_f1": 0.39215933654644525,
|
| 395 |
+
"eval_loss": 1.1027016639709473,
|
| 396 |
+
"eval_runtime": 45.9637,
|
| 397 |
+
"eval_samples_per_second": 141.024,
|
| 398 |
+
"eval_steps_per_second": 8.833,
|
| 399 |
+
"step": 13000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.12,
|
| 403 |
+
"learning_rate": 4.439089247133123e-05,
|
| 404 |
+
"loss": 0.9469,
|
| 405 |
+
"step": 13500
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.12,
|
| 409 |
+
"eval_f1": 0.4076438023703015,
|
| 410 |
+
"eval_loss": 1.1772775650024414,
|
| 411 |
+
"eval_runtime": 45.9053,
|
| 412 |
+
"eval_samples_per_second": 141.204,
|
| 413 |
+
"eval_steps_per_second": 8.844,
|
| 414 |
+
"step": 13500
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.16,
|
| 418 |
+
"learning_rate": 4.41831477480472e-05,
|
| 419 |
+
"loss": 0.9325,
|
| 420 |
+
"step": 14000
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.16,
|
| 424 |
+
"eval_f1": 0.4386204591653561,
|
| 425 |
+
"eval_loss": 1.1739530563354492,
|
| 426 |
+
"eval_runtime": 45.6601,
|
| 427 |
+
"eval_samples_per_second": 141.962,
|
| 428 |
+
"eval_steps_per_second": 8.892,
|
| 429 |
+
"step": 14000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.2,
|
| 433 |
+
"learning_rate": 4.397540302476317e-05,
|
| 434 |
+
"loss": 0.9393,
|
| 435 |
+
"step": 14500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.2,
|
| 439 |
+
"eval_f1": 0.4275788122291597,
|
| 440 |
+
"eval_loss": 1.1776121854782104,
|
| 441 |
+
"eval_runtime": 45.9594,
|
| 442 |
+
"eval_samples_per_second": 141.037,
|
| 443 |
+
"eval_steps_per_second": 8.834,
|
| 444 |
+
"step": 14500
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.25,
|
| 448 |
+
"learning_rate": 4.376765830147914e-05,
|
| 449 |
+
"loss": 0.9358,
|
| 450 |
+
"step": 15000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.25,
|
| 454 |
+
"eval_f1": 0.402493107642323,
|
| 455 |
+
"eval_loss": 1.1454391479492188,
|
| 456 |
+
"eval_runtime": 46.4333,
|
| 457 |
+
"eval_samples_per_second": 139.598,
|
| 458 |
+
"eval_steps_per_second": 8.744,
|
| 459 |
+
"step": 15000
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.29,
|
| 463 |
+
"learning_rate": 4.355991357819512e-05,
|
| 464 |
+
"loss": 0.9276,
|
| 465 |
+
"step": 15500
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.29,
|
| 469 |
+
"eval_f1": 0.4309881707776124,
|
| 470 |
+
"eval_loss": 1.1370099782943726,
|
| 471 |
+
"eval_runtime": 46.0386,
|
| 472 |
+
"eval_samples_per_second": 140.795,
|
| 473 |
+
"eval_steps_per_second": 8.819,
|
| 474 |
+
"step": 15500
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.33,
|
| 478 |
+
"learning_rate": 4.3352168854911086e-05,
|
| 479 |
+
"loss": 0.9749,
|
| 480 |
+
"step": 16000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.33,
|
| 484 |
+
"eval_f1": 0.42666296909338014,
|
| 485 |
+
"eval_loss": 1.147721290588379,
|
| 486 |
+
"eval_runtime": 46.5256,
|
| 487 |
+
"eval_samples_per_second": 139.321,
|
| 488 |
+
"eval_steps_per_second": 8.726,
|
| 489 |
+
"step": 16000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.37,
|
| 493 |
+
"learning_rate": 4.314442413162706e-05,
|
| 494 |
+
"loss": 0.9584,
|
| 495 |
+
"step": 16500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.37,
|
| 499 |
+
"eval_f1": 0.3917665752135426,
|
| 500 |
+
"eval_loss": 1.1466563940048218,
|
| 501 |
+
"eval_runtime": 45.7559,
|
| 502 |
+
"eval_samples_per_second": 141.665,
|
| 503 |
+
"eval_steps_per_second": 8.873,
|
| 504 |
+
"step": 16500
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.41,
|
| 508 |
+
"learning_rate": 4.293667940834303e-05,
|
| 509 |
+
"loss": 0.9458,
|
| 510 |
+
"step": 17000
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.41,
|
| 514 |
+
"eval_f1": 0.38855363975832957,
|
| 515 |
+
"eval_loss": 1.1946083307266235,
|
| 516 |
+
"eval_runtime": 46.3569,
|
| 517 |
+
"eval_samples_per_second": 139.828,
|
| 518 |
+
"eval_steps_per_second": 8.758,
|
| 519 |
+
"step": 17000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.45,
|
| 523 |
+
"learning_rate": 4.2728934685059e-05,
|
| 524 |
+
"loss": 0.9615,
|
| 525 |
+
"step": 17500
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.45,
|
| 529 |
+
"eval_f1": 0.4006260567973624,
|
| 530 |
+
"eval_loss": 1.1700124740600586,
|
| 531 |
+
"eval_runtime": 46.2179,
|
| 532 |
+
"eval_samples_per_second": 140.249,
|
| 533 |
+
"eval_steps_per_second": 8.784,
|
| 534 |
+
"step": 17500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.5,
|
| 538 |
+
"learning_rate": 4.252118996177497e-05,
|
| 539 |
+
"loss": 0.949,
|
| 540 |
+
"step": 18000
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 1.5,
|
| 544 |
+
"eval_f1": 0.39774908999391234,
|
| 545 |
+
"eval_loss": 1.1761705875396729,
|
| 546 |
+
"eval_runtime": 46.6777,
|
| 547 |
+
"eval_samples_per_second": 138.867,
|
| 548 |
+
"eval_steps_per_second": 8.698,
|
| 549 |
+
"step": 18000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.54,
|
| 553 |
+
"learning_rate": 4.2313445238490946e-05,
|
| 554 |
+
"loss": 0.9424,
|
| 555 |
+
"step": 18500
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.54,
|
| 559 |
+
"eval_f1": 0.4174117585364745,
|
| 560 |
+
"eval_loss": 1.165438175201416,
|
| 561 |
+
"eval_runtime": 45.4426,
|
| 562 |
+
"eval_samples_per_second": 142.641,
|
| 563 |
+
"eval_steps_per_second": 8.934,
|
| 564 |
+
"step": 18500
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.58,
|
| 568 |
+
"learning_rate": 4.2105700515206915e-05,
|
| 569 |
+
"loss": 0.947,
|
| 570 |
+
"step": 19000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.58,
|
| 574 |
+
"eval_f1": 0.3867006123407769,
|
| 575 |
+
"eval_loss": 1.1531673669815063,
|
| 576 |
+
"eval_runtime": 46.295,
|
| 577 |
+
"eval_samples_per_second": 140.015,
|
| 578 |
+
"eval_steps_per_second": 8.77,
|
| 579 |
+
"step": 19000
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.62,
|
| 583 |
+
"learning_rate": 4.189795579192289e-05,
|
| 584 |
+
"loss": 0.938,
|
| 585 |
+
"step": 19500
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.62,
|
| 589 |
+
"eval_f1": 0.44235664423629284,
|
| 590 |
+
"eval_loss": 1.1730421781539917,
|
| 591 |
+
"eval_runtime": 45.858,
|
| 592 |
+
"eval_samples_per_second": 141.349,
|
| 593 |
+
"eval_steps_per_second": 8.853,
|
| 594 |
+
"step": 19500
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.66,
|
| 598 |
+
"learning_rate": 4.1690211068638854e-05,
|
| 599 |
+
"loss": 0.9486,
|
| 600 |
+
"step": 20000
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 1.66,
|
| 604 |
+
"eval_f1": 0.41240105931665383,
|
| 605 |
+
"eval_loss": 1.1419258117675781,
|
| 606 |
+
"eval_runtime": 46.2869,
|
| 607 |
+
"eval_samples_per_second": 140.04,
|
| 608 |
+
"eval_steps_per_second": 8.771,
|
| 609 |
+
"step": 20000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.7,
|
| 613 |
+
"learning_rate": 4.148246634535483e-05,
|
| 614 |
+
"loss": 0.9464,
|
| 615 |
+
"step": 20500
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.7,
|
| 619 |
+
"eval_f1": 0.4092001043387523,
|
| 620 |
+
"eval_loss": 1.2019739151000977,
|
| 621 |
+
"eval_runtime": 46.6507,
|
| 622 |
+
"eval_samples_per_second": 138.948,
|
| 623 |
+
"eval_steps_per_second": 8.703,
|
| 624 |
+
"step": 20500
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.75,
|
| 628 |
+
"learning_rate": 4.12747216220708e-05,
|
| 629 |
+
"loss": 0.933,
|
| 630 |
+
"step": 21000
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 1.75,
|
| 634 |
+
"eval_f1": 0.4176306552081639,
|
| 635 |
+
"eval_loss": 1.1400264501571655,
|
| 636 |
+
"eval_runtime": 46.9566,
|
| 637 |
+
"eval_samples_per_second": 138.042,
|
| 638 |
+
"eval_steps_per_second": 8.646,
|
| 639 |
+
"step": 21000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.79,
|
| 643 |
+
"learning_rate": 4.1066976898786775e-05,
|
| 644 |
+
"loss": 0.9544,
|
| 645 |
+
"step": 21500
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 1.79,
|
| 649 |
+
"eval_f1": 0.430552427353769,
|
| 650 |
+
"eval_loss": 1.1604799032211304,
|
| 651 |
+
"eval_runtime": 47.2589,
|
| 652 |
+
"eval_samples_per_second": 137.159,
|
| 653 |
+
"eval_steps_per_second": 8.591,
|
| 654 |
+
"step": 21500
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.83,
|
| 658 |
+
"learning_rate": 4.0859232175502744e-05,
|
| 659 |
+
"loss": 0.9312,
|
| 660 |
+
"step": 22000
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 1.83,
|
| 664 |
+
"eval_f1": 0.39462194466074857,
|
| 665 |
+
"eval_loss": 1.1545989513397217,
|
| 666 |
+
"eval_runtime": 46.5344,
|
| 667 |
+
"eval_samples_per_second": 139.295,
|
| 668 |
+
"eval_steps_per_second": 8.725,
|
| 669 |
+
"step": 22000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.87,
|
| 673 |
+
"learning_rate": 4.065148745221872e-05,
|
| 674 |
+
"loss": 0.9458,
|
| 675 |
+
"step": 22500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.87,
|
| 679 |
+
"eval_f1": 0.44503977528303557,
|
| 680 |
+
"eval_loss": 1.1579736471176147,
|
| 681 |
+
"eval_runtime": 47.0732,
|
| 682 |
+
"eval_samples_per_second": 137.7,
|
| 683 |
+
"eval_steps_per_second": 8.625,
|
| 684 |
+
"step": 22500
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 1.91,
|
| 688 |
+
"learning_rate": 4.044374272893468e-05,
|
| 689 |
+
"loss": 0.9463,
|
| 690 |
+
"step": 23000
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 1.91,
|
| 694 |
+
"eval_f1": 0.41566594457572864,
|
| 695 |
+
"eval_loss": 1.1216946840286255,
|
| 696 |
+
"eval_runtime": 46.6125,
|
| 697 |
+
"eval_samples_per_second": 139.061,
|
| 698 |
+
"eval_steps_per_second": 8.71,
|
| 699 |
+
"step": 23000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.95,
|
| 703 |
+
"learning_rate": 4.023599800565066e-05,
|
| 704 |
+
"loss": 0.9292,
|
| 705 |
+
"step": 23500
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 1.95,
|
| 709 |
+
"eval_f1": 0.3965530806436371,
|
| 710 |
+
"eval_loss": 1.1553888320922852,
|
| 711 |
+
"eval_runtime": 46.7596,
|
| 712 |
+
"eval_samples_per_second": 138.624,
|
| 713 |
+
"eval_steps_per_second": 8.683,
|
| 714 |
+
"step": 23500
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 1.99,
|
| 718 |
+
"learning_rate": 4.002825328236663e-05,
|
| 719 |
+
"loss": 0.9286,
|
| 720 |
+
"step": 24000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.99,
|
| 724 |
+
"eval_f1": 0.4010412760625175,
|
| 725 |
+
"eval_loss": 1.1262454986572266,
|
| 726 |
+
"eval_runtime": 46.3591,
|
| 727 |
+
"eval_samples_per_second": 139.822,
|
| 728 |
+
"eval_steps_per_second": 8.758,
|
| 729 |
+
"step": 24000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.04,
|
| 733 |
+
"learning_rate": 3.9820508559082604e-05,
|
| 734 |
+
"loss": 0.7944,
|
| 735 |
+
"step": 24500
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.04,
|
| 739 |
+
"eval_f1": 0.43825195451845356,
|
| 740 |
+
"eval_loss": 1.2961622476577759,
|
| 741 |
+
"eval_runtime": 46.8176,
|
| 742 |
+
"eval_samples_per_second": 138.452,
|
| 743 |
+
"eval_steps_per_second": 8.672,
|
| 744 |
+
"step": 24500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.08,
|
| 748 |
+
"learning_rate": 3.961276383579857e-05,
|
| 749 |
+
"loss": 0.8003,
|
| 750 |
+
"step": 25000
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 2.08,
|
| 754 |
+
"eval_f1": 0.4380376882658572,
|
| 755 |
+
"eval_loss": 1.268869161605835,
|
| 756 |
+
"eval_runtime": 52.7668,
|
| 757 |
+
"eval_samples_per_second": 122.842,
|
| 758 |
+
"eval_steps_per_second": 7.694,
|
| 759 |
+
"step": 25000
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.12,
|
| 763 |
+
"learning_rate": 3.940501911251455e-05,
|
| 764 |
+
"loss": 0.7792,
|
| 765 |
+
"step": 25500
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.12,
|
| 769 |
+
"eval_f1": 0.4432590805872725,
|
| 770 |
+
"eval_loss": 1.2123405933380127,
|
| 771 |
+
"eval_runtime": 46.7198,
|
| 772 |
+
"eval_samples_per_second": 138.742,
|
| 773 |
+
"eval_steps_per_second": 8.69,
|
| 774 |
+
"step": 25500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.16,
|
| 778 |
+
"learning_rate": 3.919727438923051e-05,
|
| 779 |
+
"loss": 0.79,
|
| 780 |
+
"step": 26000
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 2.16,
|
| 784 |
+
"eval_f1": 0.44452383932580275,
|
| 785 |
+
"eval_loss": 1.2517160177230835,
|
| 786 |
+
"eval_runtime": 46.1817,
|
| 787 |
+
"eval_samples_per_second": 140.359,
|
| 788 |
+
"eval_steps_per_second": 8.791,
|
| 789 |
+
"step": 26000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.2,
|
| 793 |
+
"learning_rate": 3.898952966594649e-05,
|
| 794 |
+
"loss": 0.7984,
|
| 795 |
+
"step": 26500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.2,
|
| 799 |
+
"eval_f1": 0.4221078291607028,
|
| 800 |
+
"eval_loss": 1.2184810638427734,
|
| 801 |
+
"eval_runtime": 46.3545,
|
| 802 |
+
"eval_samples_per_second": 139.835,
|
| 803 |
+
"eval_steps_per_second": 8.759,
|
| 804 |
+
"step": 26500
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.24,
|
| 808 |
+
"learning_rate": 3.878178494266246e-05,
|
| 809 |
+
"loss": 0.7952,
|
| 810 |
+
"step": 27000
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 2.24,
|
| 814 |
+
"eval_f1": 0.439579966107181,
|
| 815 |
+
"eval_loss": 1.2801449298858643,
|
| 816 |
+
"eval_runtime": 46.2023,
|
| 817 |
+
"eval_samples_per_second": 140.296,
|
| 818 |
+
"eval_steps_per_second": 8.787,
|
| 819 |
+
"step": 27000
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.29,
|
| 823 |
+
"learning_rate": 3.857404021937843e-05,
|
| 824 |
+
"loss": 0.8055,
|
| 825 |
+
"step": 27500
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.29,
|
| 829 |
+
"eval_f1": 0.4544774942196613,
|
| 830 |
+
"eval_loss": 1.2639812231063843,
|
| 831 |
+
"eval_runtime": 46.0512,
|
| 832 |
+
"eval_samples_per_second": 140.757,
|
| 833 |
+
"eval_steps_per_second": 8.816,
|
| 834 |
+
"step": 27500
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 2.33,
|
| 838 |
+
"learning_rate": 3.83662954960944e-05,
|
| 839 |
+
"loss": 0.8084,
|
| 840 |
+
"step": 28000
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 2.33,
|
| 844 |
+
"eval_f1": 0.44333487259913823,
|
| 845 |
+
"eval_loss": 1.228055477142334,
|
| 846 |
+
"eval_runtime": 46.2747,
|
| 847 |
+
"eval_samples_per_second": 140.077,
|
| 848 |
+
"eval_steps_per_second": 8.774,
|
| 849 |
+
"step": 28000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.37,
|
| 853 |
+
"learning_rate": 3.815855077281037e-05,
|
| 854 |
+
"loss": 0.7904,
|
| 855 |
+
"step": 28500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.37,
|
| 859 |
+
"eval_f1": 0.43992159119888363,
|
| 860 |
+
"eval_loss": 1.2493727207183838,
|
| 861 |
+
"eval_runtime": 46.321,
|
| 862 |
+
"eval_samples_per_second": 139.937,
|
| 863 |
+
"eval_steps_per_second": 8.765,
|
| 864 |
+
"step": 28500
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 2.41,
|
| 868 |
+
"learning_rate": 3.795080604952634e-05,
|
| 869 |
+
"loss": 0.8057,
|
| 870 |
+
"step": 29000
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 2.41,
|
| 874 |
+
"eval_f1": 0.4114745835354577,
|
| 875 |
+
"eval_loss": 1.2447556257247925,
|
| 876 |
+
"eval_runtime": 46.3954,
|
| 877 |
+
"eval_samples_per_second": 139.712,
|
| 878 |
+
"eval_steps_per_second": 8.751,
|
| 879 |
+
"step": 29000
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.45,
|
| 883 |
+
"learning_rate": 3.774306132624231e-05,
|
| 884 |
+
"loss": 0.8001,
|
| 885 |
+
"step": 29500
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.45,
|
| 889 |
+
"eval_f1": 0.41230473935117545,
|
| 890 |
+
"eval_loss": 1.2784521579742432,
|
| 891 |
+
"eval_runtime": 46.222,
|
| 892 |
+
"eval_samples_per_second": 140.236,
|
| 893 |
+
"eval_steps_per_second": 8.784,
|
| 894 |
+
"step": 29500
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 2.49,
|
| 898 |
+
"learning_rate": 3.7535316602958286e-05,
|
| 899 |
+
"loss": 0.8293,
|
| 900 |
+
"step": 30000
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 2.49,
|
| 904 |
+
"eval_f1": 0.4303954902219652,
|
| 905 |
+
"eval_loss": 1.1889426708221436,
|
| 906 |
+
"eval_runtime": 45.554,
|
| 907 |
+
"eval_samples_per_second": 142.293,
|
| 908 |
+
"eval_steps_per_second": 8.913,
|
| 909 |
+
"step": 30000
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.53,
|
| 913 |
+
"learning_rate": 3.7327571879674255e-05,
|
| 914 |
+
"loss": 0.8194,
|
| 915 |
+
"step": 30500
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.53,
|
| 919 |
+
"eval_f1": 0.4301727119748369,
|
| 920 |
+
"eval_loss": 1.2015577554702759,
|
| 921 |
+
"eval_runtime": 46.1992,
|
| 922 |
+
"eval_samples_per_second": 140.305,
|
| 923 |
+
"eval_steps_per_second": 8.788,
|
| 924 |
+
"step": 30500
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 2.58,
|
| 928 |
+
"learning_rate": 3.711982715639023e-05,
|
| 929 |
+
"loss": 0.8028,
|
| 930 |
+
"step": 31000
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 2.58,
|
| 934 |
+
"eval_f1": 0.44964098289977084,
|
| 935 |
+
"eval_loss": 1.2026257514953613,
|
| 936 |
+
"eval_runtime": 45.6586,
|
| 937 |
+
"eval_samples_per_second": 141.967,
|
| 938 |
+
"eval_steps_per_second": 8.892,
|
| 939 |
+
"step": 31000
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 2.62,
|
| 943 |
+
"learning_rate": 3.69120824331062e-05,
|
| 944 |
+
"loss": 0.8123,
|
| 945 |
+
"step": 31500
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 2.62,
|
| 949 |
+
"eval_f1": 0.4305026988222712,
|
| 950 |
+
"eval_loss": 1.2430651187896729,
|
| 951 |
+
"eval_runtime": 45.6789,
|
| 952 |
+
"eval_samples_per_second": 141.904,
|
| 953 |
+
"eval_steps_per_second": 8.888,
|
| 954 |
+
"step": 31500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.66,
|
| 958 |
+
"learning_rate": 3.670433770982217e-05,
|
| 959 |
+
"loss": 0.7941,
|
| 960 |
+
"step": 32000
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 2.66,
|
| 964 |
+
"eval_f1": 0.4185262770510854,
|
| 965 |
+
"eval_loss": 1.2300126552581787,
|
| 966 |
+
"eval_runtime": 46.2407,
|
| 967 |
+
"eval_samples_per_second": 140.179,
|
| 968 |
+
"eval_steps_per_second": 8.78,
|
| 969 |
+
"step": 32000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.7,
|
| 973 |
+
"learning_rate": 3.649659298653814e-05,
|
| 974 |
+
"loss": 0.7815,
|
| 975 |
+
"step": 32500
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.7,
|
| 979 |
+
"eval_f1": 0.42807790167507703,
|
| 980 |
+
"eval_loss": 1.3011759519577026,
|
| 981 |
+
"eval_runtime": 45.2153,
|
| 982 |
+
"eval_samples_per_second": 143.358,
|
| 983 |
+
"eval_steps_per_second": 8.979,
|
| 984 |
+
"step": 32500
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 2.74,
|
| 988 |
+
"learning_rate": 3.6288848263254115e-05,
|
| 989 |
+
"loss": 0.8081,
|
| 990 |
+
"step": 33000
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 2.74,
|
| 994 |
+
"eval_f1": 0.440687373165412,
|
| 995 |
+
"eval_loss": 1.253546953201294,
|
| 996 |
+
"eval_runtime": 45.7395,
|
| 997 |
+
"eval_samples_per_second": 141.715,
|
| 998 |
+
"eval_steps_per_second": 8.876,
|
| 999 |
+
"step": 33000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 2.78,
|
| 1003 |
+
"learning_rate": 3.6081103539970084e-05,
|
| 1004 |
+
"loss": 0.8086,
|
| 1005 |
+
"step": 33500
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 2.78,
|
| 1009 |
+
"eval_f1": 0.44928394998593935,
|
| 1010 |
+
"eval_loss": 1.2568650245666504,
|
| 1011 |
+
"eval_runtime": 45.7963,
|
| 1012 |
+
"eval_samples_per_second": 141.54,
|
| 1013 |
+
"eval_steps_per_second": 8.865,
|
| 1014 |
+
"step": 33500
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 2.83,
|
| 1018 |
+
"learning_rate": 3.587335881668606e-05,
|
| 1019 |
+
"loss": 0.7858,
|
| 1020 |
+
"step": 34000
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 2.83,
|
| 1024 |
+
"eval_f1": 0.43231367666538134,
|
| 1025 |
+
"eval_loss": 1.2376387119293213,
|
| 1026 |
+
"eval_runtime": 46.5395,
|
| 1027 |
+
"eval_samples_per_second": 139.279,
|
| 1028 |
+
"eval_steps_per_second": 8.724,
|
| 1029 |
+
"step": 34000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.87,
|
| 1033 |
+
"learning_rate": 3.566561409340203e-05,
|
| 1034 |
+
"loss": 0.8065,
|
| 1035 |
+
"step": 34500
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 2.87,
|
| 1039 |
+
"eval_f1": 0.42028961729158787,
|
| 1040 |
+
"eval_loss": 1.2222144603729248,
|
| 1041 |
+
"eval_runtime": 46.2907,
|
| 1042 |
+
"eval_samples_per_second": 140.028,
|
| 1043 |
+
"eval_steps_per_second": 8.771,
|
| 1044 |
+
"step": 34500
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 2.91,
|
| 1048 |
+
"learning_rate": 3.5457869370118e-05,
|
| 1049 |
+
"loss": 0.7991,
|
| 1050 |
+
"step": 35000
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 2.91,
|
| 1054 |
+
"eval_f1": 0.4231707776654273,
|
| 1055 |
+
"eval_loss": 1.250239372253418,
|
| 1056 |
+
"eval_runtime": 45.795,
|
| 1057 |
+
"eval_samples_per_second": 141.544,
|
| 1058 |
+
"eval_steps_per_second": 8.866,
|
| 1059 |
+
"step": 35000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.95,
|
| 1063 |
+
"learning_rate": 3.525012464683397e-05,
|
| 1064 |
+
"loss": 0.816,
|
| 1065 |
+
"step": 35500
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 2.95,
|
| 1069 |
+
"eval_f1": 0.4231928313001403,
|
| 1070 |
+
"eval_loss": 1.2436952590942383,
|
| 1071 |
+
"eval_runtime": 45.9669,
|
| 1072 |
+
"eval_samples_per_second": 141.014,
|
| 1073 |
+
"eval_steps_per_second": 8.832,
|
| 1074 |
+
"step": 35500
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 2.99,
|
| 1078 |
+
"learning_rate": 3.5042379923549944e-05,
|
| 1079 |
+
"loss": 0.8093,
|
| 1080 |
+
"step": 36000
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.99,
|
| 1084 |
+
"eval_f1": 0.40722455645687655,
|
| 1085 |
+
"eval_loss": 1.1901623010635376,
|
| 1086 |
+
"eval_runtime": 45.8966,
|
| 1087 |
+
"eval_samples_per_second": 141.231,
|
| 1088 |
+
"eval_steps_per_second": 8.846,
|
| 1089 |
+
"step": 36000
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.03,
|
| 1093 |
+
"learning_rate": 3.483463520026591e-05,
|
| 1094 |
+
"loss": 0.6567,
|
| 1095 |
+
"step": 36500
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 3.03,
|
| 1099 |
+
"eval_f1": 0.4358392298680451,
|
| 1100 |
+
"eval_loss": 1.4940779209136963,
|
| 1101 |
+
"eval_runtime": 46.3646,
|
| 1102 |
+
"eval_samples_per_second": 139.805,
|
| 1103 |
+
"eval_steps_per_second": 8.757,
|
| 1104 |
+
"step": 36500
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 3.07,
|
| 1108 |
+
"learning_rate": 3.462689047698189e-05,
|
| 1109 |
+
"loss": 0.6304,
|
| 1110 |
+
"step": 37000
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 3.07,
|
| 1114 |
+
"eval_f1": 0.4248894510967205,
|
| 1115 |
+
"eval_loss": 1.471817135810852,
|
| 1116 |
+
"eval_runtime": 46.8512,
|
| 1117 |
+
"eval_samples_per_second": 138.353,
|
| 1118 |
+
"eval_steps_per_second": 8.666,
|
| 1119 |
+
"step": 37000
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.12,
|
| 1123 |
+
"learning_rate": 3.441914575369786e-05,
|
| 1124 |
+
"loss": 0.6454,
|
| 1125 |
+
"step": 37500
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.12,
|
| 1129 |
+
"eval_f1": 0.427949493239817,
|
| 1130 |
+
"eval_loss": 1.484312891960144,
|
| 1131 |
+
"eval_runtime": 46.3345,
|
| 1132 |
+
"eval_samples_per_second": 139.896,
|
| 1133 |
+
"eval_steps_per_second": 8.762,
|
| 1134 |
+
"step": 37500
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 3.16,
|
| 1138 |
+
"learning_rate": 3.421140103041383e-05,
|
| 1139 |
+
"loss": 0.6654,
|
| 1140 |
+
"step": 38000
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 3.16,
|
| 1144 |
+
"eval_f1": 0.43504779862979137,
|
| 1145 |
+
"eval_loss": 1.4933797121047974,
|
| 1146 |
+
"eval_runtime": 45.9895,
|
| 1147 |
+
"eval_samples_per_second": 140.945,
|
| 1148 |
+
"eval_steps_per_second": 8.828,
|
| 1149 |
+
"step": 38000
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.2,
|
| 1153 |
+
"learning_rate": 3.40036563071298e-05,
|
| 1154 |
+
"loss": 0.6478,
|
| 1155 |
+
"step": 38500
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 3.2,
|
| 1159 |
+
"eval_f1": 0.4310701120380386,
|
| 1160 |
+
"eval_loss": 1.4152840375900269,
|
| 1161 |
+
"eval_runtime": 46.2675,
|
| 1162 |
+
"eval_samples_per_second": 140.098,
|
| 1163 |
+
"eval_steps_per_second": 8.775,
|
| 1164 |
+
"step": 38500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 3.24,
|
| 1168 |
+
"learning_rate": 3.379591158384577e-05,
|
| 1169 |
+
"loss": 0.637,
|
| 1170 |
+
"step": 39000
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 3.24,
|
| 1174 |
+
"eval_f1": 0.44398856573642115,
|
| 1175 |
+
"eval_loss": 1.3994076251983643,
|
| 1176 |
+
"eval_runtime": 45.8114,
|
| 1177 |
+
"eval_samples_per_second": 141.493,
|
| 1178 |
+
"eval_steps_per_second": 8.862,
|
| 1179 |
+
"step": 39000
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.28,
|
| 1183 |
+
"learning_rate": 3.358816686056174e-05,
|
| 1184 |
+
"loss": 0.6398,
|
| 1185 |
+
"step": 39500
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.28,
|
| 1189 |
+
"eval_f1": 0.42669623237681525,
|
| 1190 |
+
"eval_loss": 1.5294607877731323,
|
| 1191 |
+
"eval_runtime": 46.4175,
|
| 1192 |
+
"eval_samples_per_second": 139.646,
|
| 1193 |
+
"eval_steps_per_second": 8.747,
|
| 1194 |
+
"step": 39500
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 3.32,
|
| 1198 |
+
"learning_rate": 3.338042213727772e-05,
|
| 1199 |
+
"loss": 0.6703,
|
| 1200 |
+
"step": 40000
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 3.32,
|
| 1204 |
+
"eval_f1": 0.4287387139396017,
|
| 1205 |
+
"eval_loss": 1.3941184282302856,
|
| 1206 |
+
"eval_runtime": 46.3839,
|
| 1207 |
+
"eval_samples_per_second": 139.747,
|
| 1208 |
+
"eval_steps_per_second": 8.753,
|
| 1209 |
+
"step": 40000
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.37,
|
| 1213 |
+
"learning_rate": 3.317267741399369e-05,
|
| 1214 |
+
"loss": 0.6442,
|
| 1215 |
+
"step": 40500
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.37,
|
| 1219 |
+
"eval_f1": 0.4371709367951578,
|
| 1220 |
+
"eval_loss": 1.388688564300537,
|
| 1221 |
+
"eval_runtime": 46.5417,
|
| 1222 |
+
"eval_samples_per_second": 139.273,
|
| 1223 |
+
"eval_steps_per_second": 8.723,
|
| 1224 |
+
"step": 40500
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 3.41,
|
| 1228 |
+
"learning_rate": 3.296493269070966e-05,
|
| 1229 |
+
"loss": 0.6784,
|
| 1230 |
+
"step": 41000
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 3.41,
|
| 1234 |
+
"eval_f1": 0.43751727834027077,
|
| 1235 |
+
"eval_loss": 1.3877114057540894,
|
| 1236 |
+
"eval_runtime": 45.9117,
|
| 1237 |
+
"eval_samples_per_second": 141.184,
|
| 1238 |
+
"eval_steps_per_second": 8.843,
|
| 1239 |
+
"step": 41000
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.45,
|
| 1243 |
+
"learning_rate": 3.2757187967425626e-05,
|
| 1244 |
+
"loss": 0.6614,
|
| 1245 |
+
"step": 41500
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 3.45,
|
| 1249 |
+
"eval_f1": 0.4372541317801921,
|
| 1250 |
+
"eval_loss": 1.4126884937286377,
|
| 1251 |
+
"eval_runtime": 46.4825,
|
| 1252 |
+
"eval_samples_per_second": 139.45,
|
| 1253 |
+
"eval_steps_per_second": 8.734,
|
| 1254 |
+
"step": 41500
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 3.49,
|
| 1258 |
+
"learning_rate": 3.25494432441416e-05,
|
| 1259 |
+
"loss": 0.6864,
|
| 1260 |
+
"step": 42000
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 3.49,
|
| 1264 |
+
"eval_f1": 0.42774626956090644,
|
| 1265 |
+
"eval_loss": 1.4882549047470093,
|
| 1266 |
+
"eval_runtime": 46.0334,
|
| 1267 |
+
"eval_samples_per_second": 140.811,
|
| 1268 |
+
"eval_steps_per_second": 8.82,
|
| 1269 |
+
"step": 42000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 3.53,
|
| 1273 |
+
"learning_rate": 3.234169852085757e-05,
|
| 1274 |
+
"loss": 0.6636,
|
| 1275 |
+
"step": 42500
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 3.53,
|
| 1279 |
+
"eval_f1": 0.4248446501416097,
|
| 1280 |
+
"eval_loss": 1.3951656818389893,
|
| 1281 |
+
"eval_runtime": 45.7857,
|
| 1282 |
+
"eval_samples_per_second": 141.573,
|
| 1283 |
+
"eval_steps_per_second": 8.867,
|
| 1284 |
+
"step": 42500
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 3.57,
|
| 1288 |
+
"learning_rate": 3.213395379757354e-05,
|
| 1289 |
+
"loss": 0.6801,
|
| 1290 |
+
"step": 43000
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 3.57,
|
| 1294 |
+
"eval_f1": 0.41563433666965854,
|
| 1295 |
+
"eval_loss": 1.4469736814498901,
|
| 1296 |
+
"eval_runtime": 46.1733,
|
| 1297 |
+
"eval_samples_per_second": 140.384,
|
| 1298 |
+
"eval_steps_per_second": 8.793,
|
| 1299 |
+
"step": 43000
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 3.61,
|
| 1303 |
+
"learning_rate": 3.1926209074289517e-05,
|
| 1304 |
+
"loss": 0.6509,
|
| 1305 |
+
"step": 43500
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 3.61,
|
| 1309 |
+
"eval_f1": 0.42981854634578975,
|
| 1310 |
+
"eval_loss": 1.3635119199752808,
|
| 1311 |
+
"eval_runtime": 46.4023,
|
| 1312 |
+
"eval_samples_per_second": 139.691,
|
| 1313 |
+
"eval_steps_per_second": 8.75,
|
| 1314 |
+
"step": 43500
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.66,
|
| 1318 |
+
"learning_rate": 3.1718464351005486e-05,
|
| 1319 |
+
"loss": 0.6776,
|
| 1320 |
+
"step": 44000
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 3.66,
|
| 1324 |
+
"eval_f1": 0.4345867863354599,
|
| 1325 |
+
"eval_loss": 1.3212920427322388,
|
| 1326 |
+
"eval_runtime": 46.0513,
|
| 1327 |
+
"eval_samples_per_second": 140.756,
|
| 1328 |
+
"eval_steps_per_second": 8.816,
|
| 1329 |
+
"step": 44000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 3.7,
|
| 1333 |
+
"learning_rate": 3.1510719627721455e-05,
|
| 1334 |
+
"loss": 0.6686,
|
| 1335 |
+
"step": 44500
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 3.7,
|
| 1339 |
+
"eval_f1": 0.4283615179772263,
|
| 1340 |
+
"eval_loss": 1.3529335260391235,
|
| 1341 |
+
"eval_runtime": 45.8914,
|
| 1342 |
+
"eval_samples_per_second": 141.247,
|
| 1343 |
+
"eval_steps_per_second": 8.847,
|
| 1344 |
+
"step": 44500
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 3.74,
|
| 1348 |
+
"learning_rate": 3.1302974904437424e-05,
|
| 1349 |
+
"loss": 0.6696,
|
| 1350 |
+
"step": 45000
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 3.74,
|
| 1354 |
+
"eval_f1": 0.42784477364430307,
|
| 1355 |
+
"eval_loss": 1.3639956712722778,
|
| 1356 |
+
"eval_runtime": 45.6293,
|
| 1357 |
+
"eval_samples_per_second": 142.058,
|
| 1358 |
+
"eval_steps_per_second": 8.898,
|
| 1359 |
+
"step": 45000
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 3.78,
|
| 1363 |
+
"learning_rate": 3.10952301811534e-05,
|
| 1364 |
+
"loss": 0.6624,
|
| 1365 |
+
"step": 45500
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 3.78,
|
| 1369 |
+
"eval_f1": 0.4325377081208613,
|
| 1370 |
+
"eval_loss": 1.409765601158142,
|
| 1371 |
+
"eval_runtime": 47.4821,
|
| 1372 |
+
"eval_samples_per_second": 136.515,
|
| 1373 |
+
"eval_steps_per_second": 8.551,
|
| 1374 |
+
"step": 45500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 3.82,
|
| 1378 |
+
"learning_rate": 3.088748545786937e-05,
|
| 1379 |
+
"loss": 0.6876,
|
| 1380 |
+
"step": 46000
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 3.82,
|
| 1384 |
+
"eval_f1": 0.4282077409143604,
|
| 1385 |
+
"eval_loss": 1.4017492532730103,
|
| 1386 |
+
"eval_runtime": 46.1534,
|
| 1387 |
+
"eval_samples_per_second": 140.445,
|
| 1388 |
+
"eval_steps_per_second": 8.797,
|
| 1389 |
+
"step": 46000
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 3.86,
|
| 1393 |
+
"learning_rate": 3.0679740734585346e-05,
|
| 1394 |
+
"loss": 0.671,
|
| 1395 |
+
"step": 46500
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 3.86,
|
| 1399 |
+
"eval_f1": 0.4172019162341494,
|
| 1400 |
+
"eval_loss": 1.3028308153152466,
|
| 1401 |
+
"eval_runtime": 46.2659,
|
| 1402 |
+
"eval_samples_per_second": 140.103,
|
| 1403 |
+
"eval_steps_per_second": 8.775,
|
| 1404 |
+
"step": 46500
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 3.91,
|
| 1408 |
+
"learning_rate": 3.047199601130131e-05,
|
| 1409 |
+
"loss": 0.68,
|
| 1410 |
+
"step": 47000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 3.91,
|
| 1414 |
+
"eval_f1": 0.43831774420843844,
|
| 1415 |
+
"eval_loss": 1.3964955806732178,
|
| 1416 |
+
"eval_runtime": 45.9895,
|
| 1417 |
+
"eval_samples_per_second": 140.945,
|
| 1418 |
+
"eval_steps_per_second": 8.828,
|
| 1419 |
+
"step": 47000
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 3.95,
|
| 1423 |
+
"learning_rate": 3.0264251288017287e-05,
|
| 1424 |
+
"loss": 0.6715,
|
| 1425 |
+
"step": 47500
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 3.95,
|
| 1429 |
+
"eval_f1": 0.4301624961844753,
|
| 1430 |
+
"eval_loss": 1.3572640419006348,
|
| 1431 |
+
"eval_runtime": 45.9503,
|
| 1432 |
+
"eval_samples_per_second": 141.065,
|
| 1433 |
+
"eval_steps_per_second": 8.836,
|
| 1434 |
+
"step": 47500
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 3.99,
|
| 1438 |
+
"learning_rate": 3.0056506564733257e-05,
|
| 1439 |
+
"loss": 0.697,
|
| 1440 |
+
"step": 48000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 3.99,
|
| 1444 |
+
"eval_f1": 0.4253265083713992,
|
| 1445 |
+
"eval_loss": 1.3642019033432007,
|
| 1446 |
+
"eval_runtime": 46.1149,
|
| 1447 |
+
"eval_samples_per_second": 140.562,
|
| 1448 |
+
"eval_steps_per_second": 8.804,
|
| 1449 |
+
"step": 48000
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.03,
|
| 1453 |
+
"learning_rate": 2.984876184144923e-05,
|
| 1454 |
+
"loss": 0.5631,
|
| 1455 |
+
"step": 48500
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 4.03,
|
| 1459 |
+
"eval_f1": 0.433265159904986,
|
| 1460 |
+
"eval_loss": 1.5990760326385498,
|
| 1461 |
+
"eval_runtime": 45.3111,
|
| 1462 |
+
"eval_samples_per_second": 143.055,
|
| 1463 |
+
"eval_steps_per_second": 8.96,
|
| 1464 |
+
"step": 48500
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 4.07,
|
| 1468 |
+
"learning_rate": 2.96410171181652e-05,
|
| 1469 |
+
"loss": 0.5151,
|
| 1470 |
+
"step": 49000
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 4.07,
|
| 1474 |
+
"eval_f1": 0.4295485993747164,
|
| 1475 |
+
"eval_loss": 1.6384857892990112,
|
| 1476 |
+
"eval_runtime": 46.4055,
|
| 1477 |
+
"eval_samples_per_second": 139.682,
|
| 1478 |
+
"eval_steps_per_second": 8.749,
|
| 1479 |
+
"step": 49000
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.11,
|
| 1483 |
+
"learning_rate": 2.943327239488117e-05,
|
| 1484 |
+
"loss": 0.5348,
|
| 1485 |
+
"step": 49500
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.11,
|
| 1489 |
+
"eval_f1": 0.4240371984001696,
|
| 1490 |
+
"eval_loss": 1.5903598070144653,
|
| 1491 |
+
"eval_runtime": 45.9488,
|
| 1492 |
+
"eval_samples_per_second": 141.07,
|
| 1493 |
+
"eval_steps_per_second": 8.836,
|
| 1494 |
+
"step": 49500
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 4.15,
|
| 1498 |
+
"learning_rate": 2.922552767159714e-05,
|
| 1499 |
+
"loss": 0.5288,
|
| 1500 |
+
"step": 50000
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 4.15,
|
| 1504 |
+
"eval_f1": 0.42810904253980414,
|
| 1505 |
+
"eval_loss": 1.6144169569015503,
|
| 1506 |
+
"eval_runtime": 46.451,
|
| 1507 |
+
"eval_samples_per_second": 139.545,
|
| 1508 |
+
"eval_steps_per_second": 8.74,
|
| 1509 |
+
"step": 50000
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 4.2,
|
| 1513 |
+
"learning_rate": 2.9017782948313117e-05,
|
| 1514 |
+
"loss": 0.5422,
|
| 1515 |
+
"step": 50500
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 4.2,
|
| 1519 |
+
"eval_f1": 0.4302607705379434,
|
| 1520 |
+
"eval_loss": 1.7097866535186768,
|
| 1521 |
+
"eval_runtime": 46.1436,
|
| 1522 |
+
"eval_samples_per_second": 140.475,
|
| 1523 |
+
"eval_steps_per_second": 8.799,
|
| 1524 |
+
"step": 50500
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 4.24,
|
| 1528 |
+
"learning_rate": 2.8810038225029086e-05,
|
| 1529 |
+
"loss": 0.548,
|
| 1530 |
+
"step": 51000
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 4.24,
|
| 1534 |
+
"eval_f1": 0.4281896068096123,
|
| 1535 |
+
"eval_loss": 1.573617935180664,
|
| 1536 |
+
"eval_runtime": 46.7725,
|
| 1537 |
+
"eval_samples_per_second": 138.586,
|
| 1538 |
+
"eval_steps_per_second": 8.68,
|
| 1539 |
+
"step": 51000
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.28,
|
| 1543 |
+
"learning_rate": 2.860229350174506e-05,
|
| 1544 |
+
"loss": 0.5169,
|
| 1545 |
+
"step": 51500
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 4.28,
|
| 1549 |
+
"eval_f1": 0.4336268684648746,
|
| 1550 |
+
"eval_loss": 1.5596050024032593,
|
| 1551 |
+
"eval_runtime": 45.9526,
|
| 1552 |
+
"eval_samples_per_second": 141.058,
|
| 1553 |
+
"eval_steps_per_second": 8.835,
|
| 1554 |
+
"step": 51500
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 4.32,
|
| 1558 |
+
"learning_rate": 2.8394548778461028e-05,
|
| 1559 |
+
"loss": 0.5327,
|
| 1560 |
+
"step": 52000
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 4.32,
|
| 1564 |
+
"eval_f1": 0.4223419321700975,
|
| 1565 |
+
"eval_loss": 1.6166974306106567,
|
| 1566 |
+
"eval_runtime": 46.03,
|
| 1567 |
+
"eval_samples_per_second": 140.821,
|
| 1568 |
+
"eval_steps_per_second": 8.82,
|
| 1569 |
+
"step": 52000
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 4.36,
|
| 1573 |
+
"learning_rate": 2.8186804055177e-05,
|
| 1574 |
+
"loss": 0.5343,
|
| 1575 |
+
"step": 52500
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 4.36,
|
| 1579 |
+
"eval_f1": 0.4428612616903695,
|
| 1580 |
+
"eval_loss": 1.7605165243148804,
|
| 1581 |
+
"eval_runtime": 45.6649,
|
| 1582 |
+
"eval_samples_per_second": 141.947,
|
| 1583 |
+
"eval_steps_per_second": 8.891,
|
| 1584 |
+
"step": 52500
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 4.4,
|
| 1588 |
+
"learning_rate": 2.797905933189297e-05,
|
| 1589 |
+
"loss": 0.5478,
|
| 1590 |
+
"step": 53000
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 4.4,
|
| 1594 |
+
"eval_f1": 0.44071411652178355,
|
| 1595 |
+
"eval_loss": 1.6004695892333984,
|
| 1596 |
+
"eval_runtime": 45.993,
|
| 1597 |
+
"eval_samples_per_second": 140.935,
|
| 1598 |
+
"eval_steps_per_second": 8.827,
|
| 1599 |
+
"step": 53000
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 4.45,
|
| 1603 |
+
"learning_rate": 2.7771314608608946e-05,
|
| 1604 |
+
"loss": 0.5489,
|
| 1605 |
+
"step": 53500
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 4.45,
|
| 1609 |
+
"eval_f1": 0.44084931155643786,
|
| 1610 |
+
"eval_loss": 1.642219066619873,
|
| 1611 |
+
"eval_runtime": 45.869,
|
| 1612 |
+
"eval_samples_per_second": 141.316,
|
| 1613 |
+
"eval_steps_per_second": 8.851,
|
| 1614 |
+
"step": 53500
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 4.49,
|
| 1618 |
+
"learning_rate": 2.7563569885324915e-05,
|
| 1619 |
+
"loss": 0.5388,
|
| 1620 |
+
"step": 54000
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 4.49,
|
| 1624 |
+
"eval_f1": 0.4372812211041268,
|
| 1625 |
+
"eval_loss": 1.7352898120880127,
|
| 1626 |
+
"eval_runtime": 46.4558,
|
| 1627 |
+
"eval_samples_per_second": 139.53,
|
| 1628 |
+
"eval_steps_per_second": 8.739,
|
| 1629 |
+
"step": 54000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 4.53,
|
| 1633 |
+
"learning_rate": 2.7355825162040887e-05,
|
| 1634 |
+
"loss": 0.5312,
|
| 1635 |
+
"step": 54500
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 4.53,
|
| 1639 |
+
"eval_f1": 0.42873515159285114,
|
| 1640 |
+
"eval_loss": 1.6332671642303467,
|
| 1641 |
+
"eval_runtime": 46.1955,
|
| 1642 |
+
"eval_samples_per_second": 140.317,
|
| 1643 |
+
"eval_steps_per_second": 8.789,
|
| 1644 |
+
"step": 54500
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 4.57,
|
| 1648 |
+
"learning_rate": 2.7148080438756857e-05,
|
| 1649 |
+
"loss": 0.5369,
|
| 1650 |
+
"step": 55000
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 4.57,
|
| 1654 |
+
"eval_f1": 0.4392093355525315,
|
| 1655 |
+
"eval_loss": 1.5618759393692017,
|
| 1656 |
+
"eval_runtime": 47.5099,
|
| 1657 |
+
"eval_samples_per_second": 136.435,
|
| 1658 |
+
"eval_steps_per_second": 8.546,
|
| 1659 |
+
"step": 55000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 4.61,
|
| 1663 |
+
"learning_rate": 2.694033571547283e-05,
|
| 1664 |
+
"loss": 0.5475,
|
| 1665 |
+
"step": 55500
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 4.61,
|
| 1669 |
+
"eval_f1": 0.42820616668057093,
|
| 1670 |
+
"eval_loss": 1.583003282546997,
|
| 1671 |
+
"eval_runtime": 46.2808,
|
| 1672 |
+
"eval_samples_per_second": 140.058,
|
| 1673 |
+
"eval_steps_per_second": 8.773,
|
| 1674 |
+
"step": 55500
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 4.65,
|
| 1678 |
+
"learning_rate": 2.67325909921888e-05,
|
| 1679 |
+
"loss": 0.5622,
|
| 1680 |
+
"step": 56000
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 4.65,
|
| 1684 |
+
"eval_f1": 0.4405394714296675,
|
| 1685 |
+
"eval_loss": 1.5289151668548584,
|
| 1686 |
+
"eval_runtime": 46.2454,
|
| 1687 |
+
"eval_samples_per_second": 140.165,
|
| 1688 |
+
"eval_steps_per_second": 8.779,
|
| 1689 |
+
"step": 56000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 4.7,
|
| 1693 |
+
"learning_rate": 2.6524846268904768e-05,
|
| 1694 |
+
"loss": 0.5662,
|
| 1695 |
+
"step": 56500
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 4.7,
|
| 1699 |
+
"eval_f1": 0.43493209599396276,
|
| 1700 |
+
"eval_loss": 1.5689671039581299,
|
| 1701 |
+
"eval_runtime": 46.2429,
|
| 1702 |
+
"eval_samples_per_second": 140.173,
|
| 1703 |
+
"eval_steps_per_second": 8.78,
|
| 1704 |
+
"step": 56500
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 4.74,
|
| 1708 |
+
"learning_rate": 2.6317101545620744e-05,
|
| 1709 |
+
"loss": 0.5373,
|
| 1710 |
+
"step": 57000
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 4.74,
|
| 1714 |
+
"eval_f1": 0.4302946240864107,
|
| 1715 |
+
"eval_loss": 1.6275018453598022,
|
| 1716 |
+
"eval_runtime": 46.0239,
|
| 1717 |
+
"eval_samples_per_second": 140.84,
|
| 1718 |
+
"eval_steps_per_second": 8.821,
|
| 1719 |
+
"step": 57000
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 4.78,
|
| 1723 |
+
"learning_rate": 2.610935682233671e-05,
|
| 1724 |
+
"loss": 0.5584,
|
| 1725 |
+
"step": 57500
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 4.78,
|
| 1729 |
+
"eval_f1": 0.43485188110217615,
|
| 1730 |
+
"eval_loss": 1.7044297456741333,
|
| 1731 |
+
"eval_runtime": 46.2954,
|
| 1732 |
+
"eval_samples_per_second": 140.014,
|
| 1733 |
+
"eval_steps_per_second": 8.77,
|
| 1734 |
+
"step": 57500
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 4.82,
|
| 1738 |
+
"learning_rate": 2.5901612099052686e-05,
|
| 1739 |
+
"loss": 0.5484,
|
| 1740 |
+
"step": 58000
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 4.82,
|
| 1744 |
+
"eval_f1": 0.43712471779629325,
|
| 1745 |
+
"eval_loss": 1.6315213441848755,
|
| 1746 |
+
"eval_runtime": 46.0252,
|
| 1747 |
+
"eval_samples_per_second": 140.836,
|
| 1748 |
+
"eval_steps_per_second": 8.821,
|
| 1749 |
+
"step": 58000
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 4.86,
|
| 1753 |
+
"learning_rate": 2.5693867375768655e-05,
|
| 1754 |
+
"loss": 0.5475,
|
| 1755 |
+
"step": 58500
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 4.86,
|
| 1759 |
+
"eval_f1": 0.446221258616921,
|
| 1760 |
+
"eval_loss": 1.5129351615905762,
|
| 1761 |
+
"eval_runtime": 46.4009,
|
| 1762 |
+
"eval_samples_per_second": 139.696,
|
| 1763 |
+
"eval_steps_per_second": 8.75,
|
| 1764 |
+
"step": 58500
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 4.9,
|
| 1768 |
+
"learning_rate": 2.5486122652484628e-05,
|
| 1769 |
+
"loss": 0.5551,
|
| 1770 |
+
"step": 59000
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 4.9,
|
| 1774 |
+
"eval_f1": 0.4409313233994151,
|
| 1775 |
+
"eval_loss": 1.637054204940796,
|
| 1776 |
+
"eval_runtime": 46.0061,
|
| 1777 |
+
"eval_samples_per_second": 140.894,
|
| 1778 |
+
"eval_steps_per_second": 8.825,
|
| 1779 |
+
"step": 59000
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 4.94,
|
| 1783 |
+
"learning_rate": 2.5278377929200597e-05,
|
| 1784 |
+
"loss": 0.558,
|
| 1785 |
+
"step": 59500
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 4.94,
|
| 1789 |
+
"eval_f1": 0.4335536795951597,
|
| 1790 |
+
"eval_loss": 1.5173062086105347,
|
| 1791 |
+
"eval_runtime": 46.0786,
|
| 1792 |
+
"eval_samples_per_second": 140.673,
|
| 1793 |
+
"eval_steps_per_second": 8.811,
|
| 1794 |
+
"step": 59500
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 4.99,
|
| 1798 |
+
"learning_rate": 2.5070633205916573e-05,
|
| 1799 |
+
"loss": 0.5553,
|
| 1800 |
+
"step": 60000
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 4.99,
|
| 1804 |
+
"eval_f1": 0.4349357425201839,
|
| 1805 |
+
"eval_loss": 1.564207911491394,
|
| 1806 |
+
"eval_runtime": 45.533,
|
| 1807 |
+
"eval_samples_per_second": 142.358,
|
| 1808 |
+
"eval_steps_per_second": 8.917,
|
| 1809 |
+
"step": 60000
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 5.03,
|
| 1813 |
+
"learning_rate": 2.4862888482632542e-05,
|
| 1814 |
+
"loss": 0.4491,
|
| 1815 |
+
"step": 60500
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 5.03,
|
| 1819 |
+
"eval_f1": 0.43614659724704335,
|
| 1820 |
+
"eval_loss": 1.9311244487762451,
|
| 1821 |
+
"eval_runtime": 46.1968,
|
| 1822 |
+
"eval_samples_per_second": 140.313,
|
| 1823 |
+
"eval_steps_per_second": 8.788,
|
| 1824 |
+
"step": 60500
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 5.07,
|
| 1828 |
+
"learning_rate": 2.4655143759348515e-05,
|
| 1829 |
+
"loss": 0.403,
|
| 1830 |
+
"step": 61000
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 5.07,
|
| 1834 |
+
"eval_f1": 0.4393254273429778,
|
| 1835 |
+
"eval_loss": 2.0766100883483887,
|
| 1836 |
+
"eval_runtime": 45.9492,
|
| 1837 |
+
"eval_samples_per_second": 141.069,
|
| 1838 |
+
"eval_steps_per_second": 8.836,
|
| 1839 |
+
"step": 61000
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 5.11,
|
| 1843 |
+
"learning_rate": 2.4447399036064487e-05,
|
| 1844 |
+
"loss": 0.4233,
|
| 1845 |
+
"step": 61500
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 5.11,
|
| 1849 |
+
"eval_f1": 0.43421874747799094,
|
| 1850 |
+
"eval_loss": 2.0252885818481445,
|
| 1851 |
+
"eval_runtime": 46.6362,
|
| 1852 |
+
"eval_samples_per_second": 138.991,
|
| 1853 |
+
"eval_steps_per_second": 8.706,
|
| 1854 |
+
"step": 61500
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 5.15,
|
| 1858 |
+
"learning_rate": 2.4239654312780457e-05,
|
| 1859 |
+
"loss": 0.4412,
|
| 1860 |
+
"step": 62000
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 5.15,
|
| 1864 |
+
"eval_f1": 0.43655258613920295,
|
| 1865 |
+
"eval_loss": 2.0584676265716553,
|
| 1866 |
+
"eval_runtime": 46.1023,
|
| 1867 |
+
"eval_samples_per_second": 140.6,
|
| 1868 |
+
"eval_steps_per_second": 8.807,
|
| 1869 |
+
"step": 62000
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 5.19,
|
| 1873 |
+
"learning_rate": 2.403190958949643e-05,
|
| 1874 |
+
"loss": 0.4477,
|
| 1875 |
+
"step": 62500
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 5.19,
|
| 1879 |
+
"eval_f1": 0.445008636219538,
|
| 1880 |
+
"eval_loss": 1.9808226823806763,
|
| 1881 |
+
"eval_runtime": 46.3962,
|
| 1882 |
+
"eval_samples_per_second": 139.71,
|
| 1883 |
+
"eval_steps_per_second": 8.751,
|
| 1884 |
+
"step": 62500
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 5.24,
|
| 1888 |
+
"learning_rate": 2.3824164866212402e-05,
|
| 1889 |
+
"loss": 0.4497,
|
| 1890 |
+
"step": 63000
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 5.24,
|
| 1894 |
+
"eval_f1": 0.44325755380966075,
|
| 1895 |
+
"eval_loss": 1.8606414794921875,
|
| 1896 |
+
"eval_runtime": 46.1573,
|
| 1897 |
+
"eval_samples_per_second": 140.433,
|
| 1898 |
+
"eval_steps_per_second": 8.796,
|
| 1899 |
+
"step": 63000
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 5.28,
|
| 1903 |
+
"learning_rate": 2.361642014292837e-05,
|
| 1904 |
+
"loss": 0.4415,
|
| 1905 |
+
"step": 63500
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 5.28,
|
| 1909 |
+
"eval_f1": 0.44300106609021345,
|
| 1910 |
+
"eval_loss": 2.060542106628418,
|
| 1911 |
+
"eval_runtime": 45.615,
|
| 1912 |
+
"eval_samples_per_second": 142.102,
|
| 1913 |
+
"eval_steps_per_second": 8.901,
|
| 1914 |
+
"step": 63500
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 5.32,
|
| 1918 |
+
"learning_rate": 2.340867541964434e-05,
|
| 1919 |
+
"loss": 0.4655,
|
| 1920 |
+
"step": 64000
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 5.32,
|
| 1924 |
+
"eval_f1": 0.4281324149819077,
|
| 1925 |
+
"eval_loss": 1.7479959726333618,
|
| 1926 |
+
"eval_runtime": 45.9941,
|
| 1927 |
+
"eval_samples_per_second": 140.931,
|
| 1928 |
+
"eval_steps_per_second": 8.827,
|
| 1929 |
+
"step": 64000
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 5.36,
|
| 1933 |
+
"learning_rate": 2.3200930696360313e-05,
|
| 1934 |
+
"loss": 0.4395,
|
| 1935 |
+
"step": 64500
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 5.36,
|
| 1939 |
+
"eval_f1": 0.4315119731593183,
|
| 1940 |
+
"eval_loss": 1.851706862449646,
|
| 1941 |
+
"eval_runtime": 45.6959,
|
| 1942 |
+
"eval_samples_per_second": 141.851,
|
| 1943 |
+
"eval_steps_per_second": 8.885,
|
| 1944 |
+
"step": 64500
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 5.4,
|
| 1948 |
+
"learning_rate": 2.2993185973076282e-05,
|
| 1949 |
+
"loss": 0.4738,
|
| 1950 |
+
"step": 65000
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 5.4,
|
| 1954 |
+
"eval_f1": 0.4246487511252701,
|
| 1955 |
+
"eval_loss": 1.7510011196136475,
|
| 1956 |
+
"eval_runtime": 46.0796,
|
| 1957 |
+
"eval_samples_per_second": 140.67,
|
| 1958 |
+
"eval_steps_per_second": 8.811,
|
| 1959 |
+
"step": 65000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 5.44,
|
| 1963 |
+
"learning_rate": 2.2785441249792255e-05,
|
| 1964 |
+
"loss": 0.455,
|
| 1965 |
+
"step": 65500
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 5.44,
|
| 1969 |
+
"eval_f1": 0.42367965460597234,
|
| 1970 |
+
"eval_loss": 1.7951207160949707,
|
| 1971 |
+
"eval_runtime": 46.1571,
|
| 1972 |
+
"eval_samples_per_second": 140.433,
|
| 1973 |
+
"eval_steps_per_second": 8.796,
|
| 1974 |
+
"step": 65500
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 5.48,
|
| 1978 |
+
"learning_rate": 2.2577696526508228e-05,
|
| 1979 |
+
"loss": 0.4494,
|
| 1980 |
+
"step": 66000
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 5.48,
|
| 1984 |
+
"eval_f1": 0.4400121441304865,
|
| 1985 |
+
"eval_loss": 1.8601397275924683,
|
| 1986 |
+
"eval_runtime": 46.6249,
|
| 1987 |
+
"eval_samples_per_second": 139.024,
|
| 1988 |
+
"eval_steps_per_second": 8.708,
|
| 1989 |
+
"step": 66000
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 5.53,
|
| 1993 |
+
"learning_rate": 2.2369951803224197e-05,
|
| 1994 |
+
"loss": 0.4364,
|
| 1995 |
+
"step": 66500
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 5.53,
|
| 1999 |
+
"eval_f1": 0.4454509847031083,
|
| 2000 |
+
"eval_loss": 1.9597169160842896,
|
| 2001 |
+
"eval_runtime": 45.9584,
|
| 2002 |
+
"eval_samples_per_second": 141.041,
|
| 2003 |
+
"eval_steps_per_second": 8.834,
|
| 2004 |
+
"step": 66500
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 5.57,
|
| 2008 |
+
"learning_rate": 2.216220707994017e-05,
|
| 2009 |
+
"loss": 0.4611,
|
| 2010 |
+
"step": 67000
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 5.57,
|
| 2014 |
+
"eval_f1": 0.44059704673003397,
|
| 2015 |
+
"eval_loss": 1.899012804031372,
|
| 2016 |
+
"eval_runtime": 47.0101,
|
| 2017 |
+
"eval_samples_per_second": 137.885,
|
| 2018 |
+
"eval_steps_per_second": 8.636,
|
| 2019 |
+
"step": 67000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 5.61,
|
| 2023 |
+
"learning_rate": 2.1954462356656142e-05,
|
| 2024 |
+
"loss": 0.4434,
|
| 2025 |
+
"step": 67500
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 5.61,
|
| 2029 |
+
"eval_f1": 0.44119414948642377,
|
| 2030 |
+
"eval_loss": 1.983310341835022,
|
| 2031 |
+
"eval_runtime": 45.9947,
|
| 2032 |
+
"eval_samples_per_second": 140.929,
|
| 2033 |
+
"eval_steps_per_second": 8.827,
|
| 2034 |
+
"step": 67500
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 5.65,
|
| 2038 |
+
"learning_rate": 2.174671763337211e-05,
|
| 2039 |
+
"loss": 0.45,
|
| 2040 |
+
"step": 68000
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 5.65,
|
| 2044 |
+
"eval_f1": 0.4426699728777116,
|
| 2045 |
+
"eval_loss": 1.8643006086349487,
|
| 2046 |
+
"eval_runtime": 46.7241,
|
| 2047 |
+
"eval_samples_per_second": 138.729,
|
| 2048 |
+
"eval_steps_per_second": 8.689,
|
| 2049 |
+
"step": 68000
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 5.69,
|
| 2053 |
+
"learning_rate": 2.1538972910088084e-05,
|
| 2054 |
+
"loss": 0.4657,
|
| 2055 |
+
"step": 68500
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 5.69,
|
| 2059 |
+
"eval_f1": 0.4459768786824306,
|
| 2060 |
+
"eval_loss": 1.9347878694534302,
|
| 2061 |
+
"eval_runtime": 46.4171,
|
| 2062 |
+
"eval_samples_per_second": 139.647,
|
| 2063 |
+
"eval_steps_per_second": 8.747,
|
| 2064 |
+
"step": 68500
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 5.73,
|
| 2068 |
+
"learning_rate": 2.1331228186804057e-05,
|
| 2069 |
+
"loss": 0.4536,
|
| 2070 |
+
"step": 69000
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 5.73,
|
| 2074 |
+
"eval_f1": 0.43967737232771253,
|
| 2075 |
+
"eval_loss": 1.9800372123718262,
|
| 2076 |
+
"eval_runtime": 46.5899,
|
| 2077 |
+
"eval_samples_per_second": 139.129,
|
| 2078 |
+
"eval_steps_per_second": 8.714,
|
| 2079 |
+
"step": 69000
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 5.78,
|
| 2083 |
+
"learning_rate": 2.1123483463520026e-05,
|
| 2084 |
+
"loss": 0.4665,
|
| 2085 |
+
"step": 69500
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 5.78,
|
| 2089 |
+
"eval_f1": 0.43522687025203594,
|
| 2090 |
+
"eval_loss": 1.7668453454971313,
|
| 2091 |
+
"eval_runtime": 45.8529,
|
| 2092 |
+
"eval_samples_per_second": 141.365,
|
| 2093 |
+
"eval_steps_per_second": 8.854,
|
| 2094 |
+
"step": 69500
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 5.82,
|
| 2098 |
+
"learning_rate": 2.0915738740236e-05,
|
| 2099 |
+
"loss": 0.4668,
|
| 2100 |
+
"step": 70000
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 5.82,
|
| 2104 |
+
"eval_f1": 0.43357794081082646,
|
| 2105 |
+
"eval_loss": 1.8984841108322144,
|
| 2106 |
+
"eval_runtime": 46.4335,
|
| 2107 |
+
"eval_samples_per_second": 139.597,
|
| 2108 |
+
"eval_steps_per_second": 8.744,
|
| 2109 |
+
"step": 70000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 5.86,
|
| 2113 |
+
"learning_rate": 2.070799401695197e-05,
|
| 2114 |
+
"loss": 0.4622,
|
| 2115 |
+
"step": 70500
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 5.86,
|
| 2119 |
+
"eval_f1": 0.4323966893074926,
|
| 2120 |
+
"eval_loss": 1.8252310752868652,
|
| 2121 |
+
"eval_runtime": 45.8779,
|
| 2122 |
+
"eval_samples_per_second": 141.288,
|
| 2123 |
+
"eval_steps_per_second": 8.85,
|
| 2124 |
+
"step": 70500
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 5.9,
|
| 2128 |
+
"learning_rate": 2.050024929366794e-05,
|
| 2129 |
+
"loss": 0.4766,
|
| 2130 |
+
"step": 71000
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 5.9,
|
| 2134 |
+
"eval_f1": 0.44080051637694073,
|
| 2135 |
+
"eval_loss": 1.7060314416885376,
|
| 2136 |
+
"eval_runtime": 46.3123,
|
| 2137 |
+
"eval_samples_per_second": 139.963,
|
| 2138 |
+
"eval_steps_per_second": 8.767,
|
| 2139 |
+
"step": 71000
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 5.94,
|
| 2143 |
+
"learning_rate": 2.0292504570383913e-05,
|
| 2144 |
+
"loss": 0.4476,
|
| 2145 |
+
"step": 71500
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 5.94,
|
| 2149 |
+
"eval_f1": 0.4385964321265162,
|
| 2150 |
+
"eval_loss": 1.8184629678726196,
|
| 2151 |
+
"eval_runtime": 45.9872,
|
| 2152 |
+
"eval_samples_per_second": 140.952,
|
| 2153 |
+
"eval_steps_per_second": 8.829,
|
| 2154 |
+
"step": 71500
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 5.98,
|
| 2158 |
+
"learning_rate": 2.0084759847099886e-05,
|
| 2159 |
+
"loss": 0.4602,
|
| 2160 |
+
"step": 72000
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 5.98,
|
| 2164 |
+
"eval_f1": 0.44088418061549356,
|
| 2165 |
+
"eval_loss": 1.7630596160888672,
|
| 2166 |
+
"eval_runtime": 46.6035,
|
| 2167 |
+
"eval_samples_per_second": 139.088,
|
| 2168 |
+
"eval_steps_per_second": 8.712,
|
| 2169 |
+
"step": 72000
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 6.02,
|
| 2173 |
+
"learning_rate": 1.9877015123815855e-05,
|
| 2174 |
+
"loss": 0.3679,
|
| 2175 |
+
"step": 72500
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 6.02,
|
| 2179 |
+
"eval_f1": 0.4346630256587338,
|
| 2180 |
+
"eval_loss": 2.1265206336975098,
|
| 2181 |
+
"eval_runtime": 46.1031,
|
| 2182 |
+
"eval_samples_per_second": 140.598,
|
| 2183 |
+
"eval_steps_per_second": 8.806,
|
| 2184 |
+
"step": 72500
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 6.07,
|
| 2188 |
+
"learning_rate": 1.9669270400531827e-05,
|
| 2189 |
+
"loss": 0.3708,
|
| 2190 |
+
"step": 73000
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 6.07,
|
| 2194 |
+
"eval_f1": 0.4350030580705022,
|
| 2195 |
+
"eval_loss": 2.233613967895508,
|
| 2196 |
+
"eval_runtime": 46.3001,
|
| 2197 |
+
"eval_samples_per_second": 140.0,
|
| 2198 |
+
"eval_steps_per_second": 8.769,
|
| 2199 |
+
"step": 73000
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 6.11,
|
| 2203 |
+
"learning_rate": 1.94615256772478e-05,
|
| 2204 |
+
"loss": 0.3761,
|
| 2205 |
+
"step": 73500
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 6.11,
|
| 2209 |
+
"eval_f1": 0.4315408453045143,
|
| 2210 |
+
"eval_loss": 2.161686420440674,
|
| 2211 |
+
"eval_runtime": 46.0645,
|
| 2212 |
+
"eval_samples_per_second": 140.716,
|
| 2213 |
+
"eval_steps_per_second": 8.814,
|
| 2214 |
+
"step": 73500
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 6.15,
|
| 2218 |
+
"learning_rate": 1.925378095396377e-05,
|
| 2219 |
+
"loss": 0.382,
|
| 2220 |
+
"step": 74000
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 6.15,
|
| 2224 |
+
"eval_f1": 0.4300930630733001,
|
| 2225 |
+
"eval_loss": 2.2093658447265625,
|
| 2226 |
+
"eval_runtime": 46.5249,
|
| 2227 |
+
"eval_samples_per_second": 139.323,
|
| 2228 |
+
"eval_steps_per_second": 8.727,
|
| 2229 |
+
"step": 74000
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 6.19,
|
| 2233 |
+
"learning_rate": 1.9046036230679742e-05,
|
| 2234 |
+
"loss": 0.3606,
|
| 2235 |
+
"step": 74500
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 6.19,
|
| 2239 |
+
"eval_f1": 0.4263466007824531,
|
| 2240 |
+
"eval_loss": 2.422569751739502,
|
| 2241 |
+
"eval_runtime": 46.1496,
|
| 2242 |
+
"eval_samples_per_second": 140.456,
|
| 2243 |
+
"eval_steps_per_second": 8.797,
|
| 2244 |
+
"step": 74500
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 6.23,
|
| 2248 |
+
"learning_rate": 1.8838291507395715e-05,
|
| 2249 |
+
"loss": 0.3929,
|
| 2250 |
+
"step": 75000
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 6.23,
|
| 2254 |
+
"eval_f1": 0.42870899531401374,
|
| 2255 |
+
"eval_loss": 2.2339413166046143,
|
| 2256 |
+
"eval_runtime": 45.9626,
|
| 2257 |
+
"eval_samples_per_second": 141.028,
|
| 2258 |
+
"eval_steps_per_second": 8.833,
|
| 2259 |
+
"step": 75000
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 6.27,
|
| 2263 |
+
"learning_rate": 1.8630546784111684e-05,
|
| 2264 |
+
"loss": 0.3761,
|
| 2265 |
+
"step": 75500
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 6.27,
|
| 2269 |
+
"eval_f1": 0.43084241603610085,
|
| 2270 |
+
"eval_loss": 2.245156764984131,
|
| 2271 |
+
"eval_runtime": 45.9689,
|
| 2272 |
+
"eval_samples_per_second": 141.008,
|
| 2273 |
+
"eval_steps_per_second": 8.832,
|
| 2274 |
+
"step": 75500
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 6.32,
|
| 2278 |
+
"learning_rate": 1.8422802060827657e-05,
|
| 2279 |
+
"loss": 0.3735,
|
| 2280 |
+
"step": 76000
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 6.32,
|
| 2284 |
+
"eval_f1": 0.42730161492497354,
|
| 2285 |
+
"eval_loss": 2.375741720199585,
|
| 2286 |
+
"eval_runtime": 45.5018,
|
| 2287 |
+
"eval_samples_per_second": 142.456,
|
| 2288 |
+
"eval_steps_per_second": 8.923,
|
| 2289 |
+
"step": 76000
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 6.36,
|
| 2293 |
+
"learning_rate": 1.821505733754363e-05,
|
| 2294 |
+
"loss": 0.3824,
|
| 2295 |
+
"step": 76500
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 6.36,
|
| 2299 |
+
"eval_f1": 0.42850714365323117,
|
| 2300 |
+
"eval_loss": 2.4190187454223633,
|
| 2301 |
+
"eval_runtime": 46.0335,
|
| 2302 |
+
"eval_samples_per_second": 140.81,
|
| 2303 |
+
"eval_steps_per_second": 8.82,
|
| 2304 |
+
"step": 76500
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 6.4,
|
| 2308 |
+
"learning_rate": 1.80073126142596e-05,
|
| 2309 |
+
"loss": 0.4198,
|
| 2310 |
+
"step": 77000
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 6.4,
|
| 2314 |
+
"eval_f1": 0.41957506619020246,
|
| 2315 |
+
"eval_loss": 2.291776657104492,
|
| 2316 |
+
"eval_runtime": 45.9135,
|
| 2317 |
+
"eval_samples_per_second": 141.179,
|
| 2318 |
+
"eval_steps_per_second": 8.843,
|
| 2319 |
+
"step": 77000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 6.44,
|
| 2323 |
+
"learning_rate": 1.779956789097557e-05,
|
| 2324 |
+
"loss": 0.3759,
|
| 2325 |
+
"step": 77500
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 6.44,
|
| 2329 |
+
"eval_f1": 0.4248491736870143,
|
| 2330 |
+
"eval_loss": 2.2740871906280518,
|
| 2331 |
+
"eval_runtime": 46.8387,
|
| 2332 |
+
"eval_samples_per_second": 138.39,
|
| 2333 |
+
"eval_steps_per_second": 8.668,
|
| 2334 |
+
"step": 77500
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 6.48,
|
| 2338 |
+
"learning_rate": 1.759182316769154e-05,
|
| 2339 |
+
"loss": 0.4006,
|
| 2340 |
+
"step": 78000
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 6.48,
|
| 2344 |
+
"eval_f1": 0.43213490469184823,
|
| 2345 |
+
"eval_loss": 2.3717033863067627,
|
| 2346 |
+
"eval_runtime": 46.2157,
|
| 2347 |
+
"eval_samples_per_second": 140.255,
|
| 2348 |
+
"eval_steps_per_second": 8.785,
|
| 2349 |
+
"step": 78000
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 6.52,
|
| 2353 |
+
"learning_rate": 1.7384078444407513e-05,
|
| 2354 |
+
"loss": 0.374,
|
| 2355 |
+
"step": 78500
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 6.52,
|
| 2359 |
+
"eval_f1": 0.42844878973090866,
|
| 2360 |
+
"eval_loss": 2.2131588459014893,
|
| 2361 |
+
"eval_runtime": 46.4565,
|
| 2362 |
+
"eval_samples_per_second": 139.528,
|
| 2363 |
+
"eval_steps_per_second": 8.739,
|
| 2364 |
+
"step": 78500
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 6.56,
|
| 2368 |
+
"learning_rate": 1.7176333721123482e-05,
|
| 2369 |
+
"loss": 0.3985,
|
| 2370 |
+
"step": 79000
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 6.56,
|
| 2374 |
+
"eval_f1": 0.430486194952956,
|
| 2375 |
+
"eval_loss": 2.2677221298217773,
|
| 2376 |
+
"eval_runtime": 45.8499,
|
| 2377 |
+
"eval_samples_per_second": 141.374,
|
| 2378 |
+
"eval_steps_per_second": 8.855,
|
| 2379 |
+
"step": 79000
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 6.61,
|
| 2383 |
+
"learning_rate": 1.6968588997839455e-05,
|
| 2384 |
+
"loss": 0.3892,
|
| 2385 |
+
"step": 79500
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 6.61,
|
| 2389 |
+
"eval_f1": 0.4339078659073283,
|
| 2390 |
+
"eval_loss": 2.2518913745880127,
|
| 2391 |
+
"eval_runtime": 46.9488,
|
| 2392 |
+
"eval_samples_per_second": 138.065,
|
| 2393 |
+
"eval_steps_per_second": 8.648,
|
| 2394 |
+
"step": 79500
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 6.65,
|
| 2398 |
+
"learning_rate": 1.6760844274555427e-05,
|
| 2399 |
+
"loss": 0.4071,
|
| 2400 |
+
"step": 80000
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 6.65,
|
| 2404 |
+
"eval_f1": 0.4281750460163451,
|
| 2405 |
+
"eval_loss": 2.1826319694519043,
|
| 2406 |
+
"eval_runtime": 46.8418,
|
| 2407 |
+
"eval_samples_per_second": 138.381,
|
| 2408 |
+
"eval_steps_per_second": 8.667,
|
| 2409 |
+
"step": 80000
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 6.69,
|
| 2413 |
+
"learning_rate": 1.6553099551271397e-05,
|
| 2414 |
+
"loss": 0.4054,
|
| 2415 |
+
"step": 80500
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 6.69,
|
| 2419 |
+
"eval_f1": 0.425255645027199,
|
| 2420 |
+
"eval_loss": 2.1723647117614746,
|
| 2421 |
+
"eval_runtime": 46.2813,
|
| 2422 |
+
"eval_samples_per_second": 140.057,
|
| 2423 |
+
"eval_steps_per_second": 8.772,
|
| 2424 |
+
"step": 80500
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 6.73,
|
| 2428 |
+
"learning_rate": 1.634535482798737e-05,
|
| 2429 |
+
"loss": 0.374,
|
| 2430 |
+
"step": 81000
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 6.73,
|
| 2434 |
+
"eval_f1": 0.4243673281546822,
|
| 2435 |
+
"eval_loss": 2.208036184310913,
|
| 2436 |
+
"eval_runtime": 45.8911,
|
| 2437 |
+
"eval_samples_per_second": 141.248,
|
| 2438 |
+
"eval_steps_per_second": 8.847,
|
| 2439 |
+
"step": 81000
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 6.77,
|
| 2443 |
+
"learning_rate": 1.6137610104703342e-05,
|
| 2444 |
+
"loss": 0.4086,
|
| 2445 |
+
"step": 81500
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 6.77,
|
| 2449 |
+
"eval_f1": 0.4345723086799732,
|
| 2450 |
+
"eval_loss": 2.182802438735962,
|
| 2451 |
+
"eval_runtime": 45.7981,
|
| 2452 |
+
"eval_samples_per_second": 141.534,
|
| 2453 |
+
"eval_steps_per_second": 8.865,
|
| 2454 |
+
"step": 81500
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 6.81,
|
| 2458 |
+
"learning_rate": 1.592986538141931e-05,
|
| 2459 |
+
"loss": 0.3828,
|
| 2460 |
+
"step": 82000
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 6.81,
|
| 2464 |
+
"eval_f1": 0.43379920447502057,
|
| 2465 |
+
"eval_loss": 2.186859369277954,
|
| 2466 |
+
"eval_runtime": 45.8903,
|
| 2467 |
+
"eval_samples_per_second": 141.25,
|
| 2468 |
+
"eval_steps_per_second": 8.847,
|
| 2469 |
+
"step": 82000
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 6.86,
|
| 2473 |
+
"learning_rate": 1.5722120658135284e-05,
|
| 2474 |
+
"loss": 0.4065,
|
| 2475 |
+
"step": 82500
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 6.86,
|
| 2479 |
+
"eval_f1": 0.43151860579932383,
|
| 2480 |
+
"eval_loss": 2.138597249984741,
|
| 2481 |
+
"eval_runtime": 46.0037,
|
| 2482 |
+
"eval_samples_per_second": 140.902,
|
| 2483 |
+
"eval_steps_per_second": 8.825,
|
| 2484 |
+
"step": 82500
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 6.9,
|
| 2488 |
+
"learning_rate": 1.5514375934851256e-05,
|
| 2489 |
+
"loss": 0.4058,
|
| 2490 |
+
"step": 83000
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 6.9,
|
| 2494 |
+
"eval_f1": 0.43162515735240276,
|
| 2495 |
+
"eval_loss": 2.0976452827453613,
|
| 2496 |
+
"eval_runtime": 46.3993,
|
| 2497 |
+
"eval_samples_per_second": 139.7,
|
| 2498 |
+
"eval_steps_per_second": 8.75,
|
| 2499 |
+
"step": 83000
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 6.94,
|
| 2503 |
+
"learning_rate": 1.5306631211567226e-05,
|
| 2504 |
+
"loss": 0.3873,
|
| 2505 |
+
"step": 83500
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 6.94,
|
| 2509 |
+
"eval_f1": 0.4212864273818023,
|
| 2510 |
+
"eval_loss": 2.4226503372192383,
|
| 2511 |
+
"eval_runtime": 46.1616,
|
| 2512 |
+
"eval_samples_per_second": 140.42,
|
| 2513 |
+
"eval_steps_per_second": 8.795,
|
| 2514 |
+
"step": 83500
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 6.98,
|
| 2518 |
+
"learning_rate": 1.5098886488283198e-05,
|
| 2519 |
+
"loss": 0.389,
|
| 2520 |
+
"step": 84000
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 6.98,
|
| 2524 |
+
"eval_f1": 0.431436402933277,
|
| 2525 |
+
"eval_loss": 2.1598784923553467,
|
| 2526 |
+
"eval_runtime": 46.8062,
|
| 2527 |
+
"eval_samples_per_second": 138.486,
|
| 2528 |
+
"eval_steps_per_second": 8.674,
|
| 2529 |
+
"step": 84000
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 7.02,
|
| 2533 |
+
"learning_rate": 1.489114176499917e-05,
|
| 2534 |
+
"loss": 0.3538,
|
| 2535 |
+
"step": 84500
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 7.02,
|
| 2539 |
+
"eval_f1": 0.4350556588831767,
|
| 2540 |
+
"eval_loss": 2.304293155670166,
|
| 2541 |
+
"eval_runtime": 46.574,
|
| 2542 |
+
"eval_samples_per_second": 139.176,
|
| 2543 |
+
"eval_steps_per_second": 8.717,
|
| 2544 |
+
"step": 84500
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 7.06,
|
| 2548 |
+
"learning_rate": 1.4683397041715142e-05,
|
| 2549 |
+
"loss": 0.3132,
|
| 2550 |
+
"step": 85000
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 7.06,
|
| 2554 |
+
"eval_f1": 0.4266591713234993,
|
| 2555 |
+
"eval_loss": 2.650728702545166,
|
| 2556 |
+
"eval_runtime": 46.4231,
|
| 2557 |
+
"eval_samples_per_second": 139.629,
|
| 2558 |
+
"eval_steps_per_second": 8.746,
|
| 2559 |
+
"step": 85000
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 7.1,
|
| 2563 |
+
"learning_rate": 1.4475652318431113e-05,
|
| 2564 |
+
"loss": 0.3166,
|
| 2565 |
+
"step": 85500
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 7.1,
|
| 2569 |
+
"eval_f1": 0.43607320504273067,
|
| 2570 |
+
"eval_loss": 2.6059470176696777,
|
| 2571 |
+
"eval_runtime": 45.9726,
|
| 2572 |
+
"eval_samples_per_second": 140.997,
|
| 2573 |
+
"eval_steps_per_second": 8.831,
|
| 2574 |
+
"step": 85500
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 7.15,
|
| 2578 |
+
"learning_rate": 1.4267907595147084e-05,
|
| 2579 |
+
"loss": 0.3303,
|
| 2580 |
+
"step": 86000
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 7.15,
|
| 2584 |
+
"eval_f1": 0.43326161921977124,
|
| 2585 |
+
"eval_loss": 2.56927227973938,
|
| 2586 |
+
"eval_runtime": 46.099,
|
| 2587 |
+
"eval_samples_per_second": 140.611,
|
| 2588 |
+
"eval_steps_per_second": 8.807,
|
| 2589 |
+
"step": 86000
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 7.19,
|
| 2593 |
+
"learning_rate": 1.4060162871863056e-05,
|
| 2594 |
+
"loss": 0.3031,
|
| 2595 |
+
"step": 86500
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 7.19,
|
| 2599 |
+
"eval_f1": 0.43263480546826966,
|
| 2600 |
+
"eval_loss": 2.6123645305633545,
|
| 2601 |
+
"eval_runtime": 46.1092,
|
| 2602 |
+
"eval_samples_per_second": 140.579,
|
| 2603 |
+
"eval_steps_per_second": 8.805,
|
| 2604 |
+
"step": 86500
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 7.23,
|
| 2608 |
+
"learning_rate": 1.3852418148579027e-05,
|
| 2609 |
+
"loss": 0.3495,
|
| 2610 |
+
"step": 87000
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 7.23,
|
| 2614 |
+
"eval_f1": 0.42935020523616974,
|
| 2615 |
+
"eval_loss": 2.5714728832244873,
|
| 2616 |
+
"eval_runtime": 46.4434,
|
| 2617 |
+
"eval_samples_per_second": 139.568,
|
| 2618 |
+
"eval_steps_per_second": 8.742,
|
| 2619 |
+
"step": 87000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 7.27,
|
| 2623 |
+
"learning_rate": 1.3644673425294998e-05,
|
| 2624 |
+
"loss": 0.3364,
|
| 2625 |
+
"step": 87500
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 7.27,
|
| 2629 |
+
"eval_f1": 0.4343767918185199,
|
| 2630 |
+
"eval_loss": 2.5482189655303955,
|
| 2631 |
+
"eval_runtime": 45.6061,
|
| 2632 |
+
"eval_samples_per_second": 142.13,
|
| 2633 |
+
"eval_steps_per_second": 8.902,
|
| 2634 |
+
"step": 87500
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 7.31,
|
| 2638 |
+
"learning_rate": 1.3436928702010971e-05,
|
| 2639 |
+
"loss": 0.3169,
|
| 2640 |
+
"step": 88000
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 7.31,
|
| 2644 |
+
"eval_f1": 0.4326820837375438,
|
| 2645 |
+
"eval_loss": 2.533505916595459,
|
| 2646 |
+
"eval_runtime": 45.7021,
|
| 2647 |
+
"eval_samples_per_second": 141.832,
|
| 2648 |
+
"eval_steps_per_second": 8.884,
|
| 2649 |
+
"step": 88000
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 7.35,
|
| 2653 |
+
"learning_rate": 1.3229183978726942e-05,
|
| 2654 |
+
"loss": 0.3306,
|
| 2655 |
+
"step": 88500
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 7.35,
|
| 2659 |
+
"eval_f1": 0.4336259208362092,
|
| 2660 |
+
"eval_loss": 2.505479097366333,
|
| 2661 |
+
"eval_runtime": 45.9282,
|
| 2662 |
+
"eval_samples_per_second": 141.133,
|
| 2663 |
+
"eval_steps_per_second": 8.84,
|
| 2664 |
+
"step": 88500
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 7.4,
|
| 2668 |
+
"learning_rate": 1.3021439255442913e-05,
|
| 2669 |
+
"loss": 0.3238,
|
| 2670 |
+
"step": 89000
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 7.4,
|
| 2674 |
+
"eval_f1": 0.4290241175053658,
|
| 2675 |
+
"eval_loss": 2.64029598236084,
|
| 2676 |
+
"eval_runtime": 45.7506,
|
| 2677 |
+
"eval_samples_per_second": 141.681,
|
| 2678 |
+
"eval_steps_per_second": 8.874,
|
| 2679 |
+
"step": 89000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 7.44,
|
| 2683 |
+
"learning_rate": 1.2813694532158885e-05,
|
| 2684 |
+
"loss": 0.3488,
|
| 2685 |
+
"step": 89500
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 7.44,
|
| 2689 |
+
"eval_f1": 0.4318857789484726,
|
| 2690 |
+
"eval_loss": 2.4507477283477783,
|
| 2691 |
+
"eval_runtime": 46.4161,
|
| 2692 |
+
"eval_samples_per_second": 139.65,
|
| 2693 |
+
"eval_steps_per_second": 8.747,
|
| 2694 |
+
"step": 89500
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 7.48,
|
| 2698 |
+
"learning_rate": 1.2605949808874856e-05,
|
| 2699 |
+
"loss": 0.3423,
|
| 2700 |
+
"step": 90000
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 7.48,
|
| 2704 |
+
"eval_f1": 0.42784593140742677,
|
| 2705 |
+
"eval_loss": 2.4914281368255615,
|
| 2706 |
+
"eval_runtime": 46.3911,
|
| 2707 |
+
"eval_samples_per_second": 139.725,
|
| 2708 |
+
"eval_steps_per_second": 8.752,
|
| 2709 |
+
"step": 90000
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 7.52,
|
| 2713 |
+
"learning_rate": 1.2398205085590826e-05,
|
| 2714 |
+
"loss": 0.3356,
|
| 2715 |
+
"step": 90500
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 7.52,
|
| 2719 |
+
"eval_f1": 0.44002802787763307,
|
| 2720 |
+
"eval_loss": 2.723588466644287,
|
| 2721 |
+
"eval_runtime": 46.564,
|
| 2722 |
+
"eval_samples_per_second": 139.206,
|
| 2723 |
+
"eval_steps_per_second": 8.719,
|
| 2724 |
+
"step": 90500
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 7.56,
|
| 2728 |
+
"learning_rate": 1.2190460362306798e-05,
|
| 2729 |
+
"loss": 0.3655,
|
| 2730 |
+
"step": 91000
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 7.56,
|
| 2734 |
+
"eval_f1": 0.4342107835672762,
|
| 2735 |
+
"eval_loss": 2.516355276107788,
|
| 2736 |
+
"eval_runtime": 46.0601,
|
| 2737 |
+
"eval_samples_per_second": 140.729,
|
| 2738 |
+
"eval_steps_per_second": 8.815,
|
| 2739 |
+
"step": 91000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 7.6,
|
| 2743 |
+
"learning_rate": 1.198271563902277e-05,
|
| 2744 |
+
"loss": 0.3445,
|
| 2745 |
+
"step": 91500
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 7.6,
|
| 2749 |
+
"eval_f1": 0.4392682142110146,
|
| 2750 |
+
"eval_loss": 2.4509174823760986,
|
| 2751 |
+
"eval_runtime": 46.5389,
|
| 2752 |
+
"eval_samples_per_second": 139.281,
|
| 2753 |
+
"eval_steps_per_second": 8.724,
|
| 2754 |
+
"step": 91500
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 7.65,
|
| 2758 |
+
"learning_rate": 1.177497091573874e-05,
|
| 2759 |
+
"loss": 0.3123,
|
| 2760 |
+
"step": 92000
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 7.65,
|
| 2764 |
+
"eval_f1": 0.4346012419342983,
|
| 2765 |
+
"eval_loss": 2.605870008468628,
|
| 2766 |
+
"eval_runtime": 46.0454,
|
| 2767 |
+
"eval_samples_per_second": 140.774,
|
| 2768 |
+
"eval_steps_per_second": 8.817,
|
| 2769 |
+
"step": 92000
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 7.69,
|
| 2773 |
+
"learning_rate": 1.1567226192454713e-05,
|
| 2774 |
+
"loss": 0.3286,
|
| 2775 |
+
"step": 92500
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 7.69,
|
| 2779 |
+
"eval_f1": 0.42900413120162884,
|
| 2780 |
+
"eval_loss": 2.5746617317199707,
|
| 2781 |
+
"eval_runtime": 46.6316,
|
| 2782 |
+
"eval_samples_per_second": 139.005,
|
| 2783 |
+
"eval_steps_per_second": 8.707,
|
| 2784 |
+
"step": 92500
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 7.73,
|
| 2788 |
+
"learning_rate": 1.1359481469170684e-05,
|
| 2789 |
+
"loss": 0.3308,
|
| 2790 |
+
"step": 93000
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 7.73,
|
| 2794 |
+
"eval_f1": 0.4348325522174462,
|
| 2795 |
+
"eval_loss": 2.6604156494140625,
|
| 2796 |
+
"eval_runtime": 46.0552,
|
| 2797 |
+
"eval_samples_per_second": 140.744,
|
| 2798 |
+
"eval_steps_per_second": 8.816,
|
| 2799 |
+
"step": 93000
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 7.77,
|
| 2803 |
+
"learning_rate": 1.1151736745886655e-05,
|
| 2804 |
+
"loss": 0.3507,
|
| 2805 |
+
"step": 93500
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 7.77,
|
| 2809 |
+
"eval_f1": 0.4362724679273792,
|
| 2810 |
+
"eval_loss": 2.6467623710632324,
|
| 2811 |
+
"eval_runtime": 46.2441,
|
| 2812 |
+
"eval_samples_per_second": 140.169,
|
| 2813 |
+
"eval_steps_per_second": 8.78,
|
| 2814 |
+
"step": 93500
|
| 2815 |
+
},
|
| 2816 |
+
{
|
| 2817 |
+
"epoch": 7.81,
|
| 2818 |
+
"learning_rate": 1.0943992022602627e-05,
|
| 2819 |
+
"loss": 0.3392,
|
| 2820 |
+
"step": 94000
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 7.81,
|
| 2824 |
+
"eval_f1": 0.43161760121865855,
|
| 2825 |
+
"eval_loss": 2.6293702125549316,
|
| 2826 |
+
"eval_runtime": 45.6229,
|
| 2827 |
+
"eval_samples_per_second": 142.078,
|
| 2828 |
+
"eval_steps_per_second": 8.899,
|
| 2829 |
+
"step": 94000
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 7.85,
|
| 2833 |
+
"learning_rate": 1.0736247299318598e-05,
|
| 2834 |
+
"loss": 0.3885,
|
| 2835 |
+
"step": 94500
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 7.85,
|
| 2839 |
+
"eval_f1": 0.42849041218645784,
|
| 2840 |
+
"eval_loss": 2.4070699214935303,
|
| 2841 |
+
"eval_runtime": 45.9402,
|
| 2842 |
+
"eval_samples_per_second": 141.096,
|
| 2843 |
+
"eval_steps_per_second": 8.838,
|
| 2844 |
+
"step": 94500
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 7.89,
|
| 2848 |
+
"learning_rate": 1.052850257603457e-05,
|
| 2849 |
+
"loss": 0.3458,
|
| 2850 |
+
"step": 95000
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 7.89,
|
| 2854 |
+
"eval_f1": 0.42803465998306145,
|
| 2855 |
+
"eval_loss": 2.5539064407348633,
|
| 2856 |
+
"eval_runtime": 46.1313,
|
| 2857 |
+
"eval_samples_per_second": 140.512,
|
| 2858 |
+
"eval_steps_per_second": 8.801,
|
| 2859 |
+
"step": 95000
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 7.94,
|
| 2863 |
+
"learning_rate": 1.032075785275054e-05,
|
| 2864 |
+
"loss": 0.3451,
|
| 2865 |
+
"step": 95500
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 7.94,
|
| 2869 |
+
"eval_f1": 0.4282701100981614,
|
| 2870 |
+
"eval_loss": 2.4033892154693604,
|
| 2871 |
+
"eval_runtime": 46.4312,
|
| 2872 |
+
"eval_samples_per_second": 139.604,
|
| 2873 |
+
"eval_steps_per_second": 8.744,
|
| 2874 |
+
"step": 95500
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 7.98,
|
| 2878 |
+
"learning_rate": 1.0113013129466511e-05,
|
| 2879 |
+
"loss": 0.3394,
|
| 2880 |
+
"step": 96000
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 7.98,
|
| 2884 |
+
"eval_f1": 0.43486076258951,
|
| 2885 |
+
"eval_loss": 2.6116716861724854,
|
| 2886 |
+
"eval_runtime": 45.9663,
|
| 2887 |
+
"eval_samples_per_second": 141.016,
|
| 2888 |
+
"eval_steps_per_second": 8.833,
|
| 2889 |
+
"step": 96000
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 8.02,
|
| 2893 |
+
"learning_rate": 9.905268406182484e-06,
|
| 2894 |
+
"loss": 0.3275,
|
| 2895 |
+
"step": 96500
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 8.02,
|
| 2899 |
+
"eval_f1": 0.4322422606334531,
|
| 2900 |
+
"eval_loss": 2.5991387367248535,
|
| 2901 |
+
"eval_runtime": 46.5777,
|
| 2902 |
+
"eval_samples_per_second": 139.165,
|
| 2903 |
+
"eval_steps_per_second": 8.717,
|
| 2904 |
+
"step": 96500
|
| 2905 |
+
},
|
| 2906 |
+
{
|
| 2907 |
+
"epoch": 8.06,
|
| 2908 |
+
"learning_rate": 9.697523682898455e-06,
|
| 2909 |
+
"loss": 0.2676,
|
| 2910 |
+
"step": 97000
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 8.06,
|
| 2914 |
+
"eval_f1": 0.42999721877624847,
|
| 2915 |
+
"eval_loss": 2.772134304046631,
|
| 2916 |
+
"eval_runtime": 46.0246,
|
| 2917 |
+
"eval_samples_per_second": 140.838,
|
| 2918 |
+
"eval_steps_per_second": 8.821,
|
| 2919 |
+
"step": 97000
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 8.1,
|
| 2923 |
+
"learning_rate": 9.489778959614426e-06,
|
| 2924 |
+
"loss": 0.3013,
|
| 2925 |
+
"step": 97500
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 8.1,
|
| 2929 |
+
"eval_f1": 0.43159307099859695,
|
| 2930 |
+
"eval_loss": 2.7766764163970947,
|
| 2931 |
+
"eval_runtime": 46.3012,
|
| 2932 |
+
"eval_samples_per_second": 139.996,
|
| 2933 |
+
"eval_steps_per_second": 8.769,
|
| 2934 |
+
"step": 97500
|
| 2935 |
+
},
|
| 2936 |
+
{
|
| 2937 |
+
"epoch": 8.14,
|
| 2938 |
+
"learning_rate": 9.282034236330398e-06,
|
| 2939 |
+
"loss": 0.283,
|
| 2940 |
+
"step": 98000
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 8.14,
|
| 2944 |
+
"eval_f1": 0.43429947059083457,
|
| 2945 |
+
"eval_loss": 2.723806142807007,
|
| 2946 |
+
"eval_runtime": 46.2805,
|
| 2947 |
+
"eval_samples_per_second": 140.059,
|
| 2948 |
+
"eval_steps_per_second": 8.773,
|
| 2949 |
+
"step": 98000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 8.19,
|
| 2953 |
+
"learning_rate": 9.07428951304637e-06,
|
| 2954 |
+
"loss": 0.2924,
|
| 2955 |
+
"step": 98500
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 8.19,
|
| 2959 |
+
"eval_f1": 0.43274102934044933,
|
| 2960 |
+
"eval_loss": 2.7132375240325928,
|
| 2961 |
+
"eval_runtime": 46.1859,
|
| 2962 |
+
"eval_samples_per_second": 140.346,
|
| 2963 |
+
"eval_steps_per_second": 8.791,
|
| 2964 |
+
"step": 98500
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 8.23,
|
| 2968 |
+
"learning_rate": 8.86654478976234e-06,
|
| 2969 |
+
"loss": 0.2817,
|
| 2970 |
+
"step": 99000
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 8.23,
|
| 2974 |
+
"eval_f1": 0.427616805108728,
|
| 2975 |
+
"eval_loss": 2.7119312286376953,
|
| 2976 |
+
"eval_runtime": 56.0523,
|
| 2977 |
+
"eval_samples_per_second": 115.642,
|
| 2978 |
+
"eval_steps_per_second": 7.243,
|
| 2979 |
+
"step": 99000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 8.27,
|
| 2983 |
+
"learning_rate": 8.658800066478313e-06,
|
| 2984 |
+
"loss": 0.2586,
|
| 2985 |
+
"step": 99500
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 8.27,
|
| 2989 |
+
"eval_f1": 0.4289585910640077,
|
| 2990 |
+
"eval_loss": 2.887305736541748,
|
| 2991 |
+
"eval_runtime": 46.8426,
|
| 2992 |
+
"eval_samples_per_second": 138.378,
|
| 2993 |
+
"eval_steps_per_second": 8.667,
|
| 2994 |
+
"step": 99500
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 8.31,
|
| 2998 |
+
"learning_rate": 8.451055343194284e-06,
|
| 2999 |
+
"loss": 0.3051,
|
| 3000 |
+
"step": 100000
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 8.31,
|
| 3004 |
+
"eval_f1": 0.42582497432500993,
|
| 3005 |
+
"eval_loss": 2.7572007179260254,
|
| 3006 |
+
"eval_runtime": 45.6991,
|
| 3007 |
+
"eval_samples_per_second": 141.841,
|
| 3008 |
+
"eval_steps_per_second": 8.884,
|
| 3009 |
+
"step": 100000
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 8.35,
|
| 3013 |
+
"learning_rate": 8.243310619910255e-06,
|
| 3014 |
+
"loss": 0.2916,
|
| 3015 |
+
"step": 100500
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 8.35,
|
| 3019 |
+
"eval_f1": 0.43075118593793377,
|
| 3020 |
+
"eval_loss": 2.814906358718872,
|
| 3021 |
+
"eval_runtime": 46.2133,
|
| 3022 |
+
"eval_samples_per_second": 140.263,
|
| 3023 |
+
"eval_steps_per_second": 8.785,
|
| 3024 |
+
"step": 100500
|
| 3025 |
+
},
|
| 3026 |
+
{
|
| 3027 |
+
"epoch": 8.39,
|
| 3028 |
+
"learning_rate": 8.035565896626227e-06,
|
| 3029 |
+
"loss": 0.2948,
|
| 3030 |
+
"step": 101000
|
| 3031 |
+
},
|
| 3032 |
+
{
|
| 3033 |
+
"epoch": 8.39,
|
| 3034 |
+
"eval_f1": 0.4244624620028211,
|
| 3035 |
+
"eval_loss": 2.6767539978027344,
|
| 3036 |
+
"eval_runtime": 46.5451,
|
| 3037 |
+
"eval_samples_per_second": 139.263,
|
| 3038 |
+
"eval_steps_per_second": 8.723,
|
| 3039 |
+
"step": 101000
|
| 3040 |
+
},
|
| 3041 |
+
{
|
| 3042 |
+
"epoch": 8.43,
|
| 3043 |
+
"learning_rate": 7.827821173342198e-06,
|
| 3044 |
+
"loss": 0.277,
|
| 3045 |
+
"step": 101500
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 8.43,
|
| 3049 |
+
"eval_f1": 0.4273818288568198,
|
| 3050 |
+
"eval_loss": 2.727567672729492,
|
| 3051 |
+
"eval_runtime": 46.8521,
|
| 3052 |
+
"eval_samples_per_second": 138.35,
|
| 3053 |
+
"eval_steps_per_second": 8.666,
|
| 3054 |
+
"step": 101500
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 8.48,
|
| 3058 |
+
"learning_rate": 7.620076450058169e-06,
|
| 3059 |
+
"loss": 0.2929,
|
| 3060 |
+
"step": 102000
|
| 3061 |
+
},
|
| 3062 |
+
{
|
| 3063 |
+
"epoch": 8.48,
|
| 3064 |
+
"eval_f1": 0.4320719535965778,
|
| 3065 |
+
"eval_loss": 2.926513433456421,
|
| 3066 |
+
"eval_runtime": 47.1016,
|
| 3067 |
+
"eval_samples_per_second": 137.617,
|
| 3068 |
+
"eval_steps_per_second": 8.62,
|
| 3069 |
+
"step": 102000
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 8.52,
|
| 3073 |
+
"learning_rate": 7.412331726774139e-06,
|
| 3074 |
+
"loss": 0.2893,
|
| 3075 |
+
"step": 102500
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 8.52,
|
| 3079 |
+
"eval_f1": 0.4260519729309167,
|
| 3080 |
+
"eval_loss": 2.877741575241089,
|
| 3081 |
+
"eval_runtime": 45.6581,
|
| 3082 |
+
"eval_samples_per_second": 141.968,
|
| 3083 |
+
"eval_steps_per_second": 8.892,
|
| 3084 |
+
"step": 102500
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 8.56,
|
| 3088 |
+
"learning_rate": 7.204587003490111e-06,
|
| 3089 |
+
"loss": 0.2808,
|
| 3090 |
+
"step": 103000
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 8.56,
|
| 3094 |
+
"eval_f1": 0.4282218551070931,
|
| 3095 |
+
"eval_loss": 2.7379603385925293,
|
| 3096 |
+
"eval_runtime": 45.9855,
|
| 3097 |
+
"eval_samples_per_second": 140.957,
|
| 3098 |
+
"eval_steps_per_second": 8.829,
|
| 3099 |
+
"step": 103000
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 8.6,
|
| 3103 |
+
"learning_rate": 6.996842280206083e-06,
|
| 3104 |
+
"loss": 0.2728,
|
| 3105 |
+
"step": 103500
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 8.6,
|
| 3109 |
+
"eval_f1": 0.43170842579550445,
|
| 3110 |
+
"eval_loss": 2.7965376377105713,
|
| 3111 |
+
"eval_runtime": 45.7627,
|
| 3112 |
+
"eval_samples_per_second": 141.644,
|
| 3113 |
+
"eval_steps_per_second": 8.872,
|
| 3114 |
+
"step": 103500
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 8.64,
|
| 3118 |
+
"learning_rate": 6.789097556922054e-06,
|
| 3119 |
+
"loss": 0.2789,
|
| 3120 |
+
"step": 104000
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 8.64,
|
| 3124 |
+
"eval_f1": 0.4315708387912964,
|
| 3125 |
+
"eval_loss": 2.875927209854126,
|
| 3126 |
+
"eval_runtime": 45.9208,
|
| 3127 |
+
"eval_samples_per_second": 141.156,
|
| 3128 |
+
"eval_steps_per_second": 8.841,
|
| 3129 |
+
"step": 104000
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 8.68,
|
| 3133 |
+
"learning_rate": 6.5813528336380256e-06,
|
| 3134 |
+
"loss": 0.3072,
|
| 3135 |
+
"step": 104500
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 8.68,
|
| 3139 |
+
"eval_f1": 0.42772350225777217,
|
| 3140 |
+
"eval_loss": 2.8333616256713867,
|
| 3141 |
+
"eval_runtime": 49.0657,
|
| 3142 |
+
"eval_samples_per_second": 132.109,
|
| 3143 |
+
"eval_steps_per_second": 8.275,
|
| 3144 |
+
"step": 104500
|
| 3145 |
+
},
|
| 3146 |
+
{
|
| 3147 |
+
"epoch": 8.73,
|
| 3148 |
+
"learning_rate": 6.373608110353997e-06,
|
| 3149 |
+
"loss": 0.2779,
|
| 3150 |
+
"step": 105000
|
| 3151 |
+
},
|
| 3152 |
+
{
|
| 3153 |
+
"epoch": 8.73,
|
| 3154 |
+
"eval_f1": 0.4306437356669429,
|
| 3155 |
+
"eval_loss": 2.8720405101776123,
|
| 3156 |
+
"eval_runtime": 46.0481,
|
| 3157 |
+
"eval_samples_per_second": 140.766,
|
| 3158 |
+
"eval_steps_per_second": 8.817,
|
| 3159 |
+
"step": 105000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 8.77,
|
| 3163 |
+
"learning_rate": 6.165863387069968e-06,
|
| 3164 |
+
"loss": 0.2948,
|
| 3165 |
+
"step": 105500
|
| 3166 |
+
},
|
| 3167 |
+
{
|
| 3168 |
+
"epoch": 8.77,
|
| 3169 |
+
"eval_f1": 0.42931126691296156,
|
| 3170 |
+
"eval_loss": 2.82370662689209,
|
| 3171 |
+
"eval_runtime": 45.6262,
|
| 3172 |
+
"eval_samples_per_second": 142.067,
|
| 3173 |
+
"eval_steps_per_second": 8.898,
|
| 3174 |
+
"step": 105500
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 8.81,
|
| 3178 |
+
"learning_rate": 5.95811866378594e-06,
|
| 3179 |
+
"loss": 0.2917,
|
| 3180 |
+
"step": 106000
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 8.81,
|
| 3184 |
+
"eval_f1": 0.42987557908855206,
|
| 3185 |
+
"eval_loss": 2.785443067550659,
|
| 3186 |
+
"eval_runtime": 45.7985,
|
| 3187 |
+
"eval_samples_per_second": 141.533,
|
| 3188 |
+
"eval_steps_per_second": 8.865,
|
| 3189 |
+
"step": 106000
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 8.85,
|
| 3193 |
+
"learning_rate": 5.750373940501912e-06,
|
| 3194 |
+
"loss": 0.2604,
|
| 3195 |
+
"step": 106500
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 8.85,
|
| 3199 |
+
"eval_f1": 0.4272210968347812,
|
| 3200 |
+
"eval_loss": 2.9237265586853027,
|
| 3201 |
+
"eval_runtime": 45.9348,
|
| 3202 |
+
"eval_samples_per_second": 141.113,
|
| 3203 |
+
"eval_steps_per_second": 8.839,
|
| 3204 |
+
"step": 106500
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 8.89,
|
| 3208 |
+
"learning_rate": 5.542629217217883e-06,
|
| 3209 |
+
"loss": 0.3057,
|
| 3210 |
+
"step": 107000
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 8.89,
|
| 3214 |
+
"eval_f1": 0.4287531329908217,
|
| 3215 |
+
"eval_loss": 2.8509225845336914,
|
| 3216 |
+
"eval_runtime": 46.4174,
|
| 3217 |
+
"eval_samples_per_second": 139.646,
|
| 3218 |
+
"eval_steps_per_second": 8.747,
|
| 3219 |
+
"step": 107000
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 8.93,
|
| 3223 |
+
"learning_rate": 5.334884493933855e-06,
|
| 3224 |
+
"loss": 0.2853,
|
| 3225 |
+
"step": 107500
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 8.93,
|
| 3229 |
+
"eval_f1": 0.4273342046666312,
|
| 3230 |
+
"eval_loss": 2.7482059001922607,
|
| 3231 |
+
"eval_runtime": 45.8656,
|
| 3232 |
+
"eval_samples_per_second": 141.326,
|
| 3233 |
+
"eval_steps_per_second": 8.852,
|
| 3234 |
+
"step": 107500
|
| 3235 |
+
},
|
| 3236 |
+
{
|
| 3237 |
+
"epoch": 8.97,
|
| 3238 |
+
"learning_rate": 5.1271397706498255e-06,
|
| 3239 |
+
"loss": 0.2946,
|
| 3240 |
+
"step": 108000
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 8.97,
|
| 3244 |
+
"eval_f1": 0.4272429839187248,
|
| 3245 |
+
"eval_loss": 2.86079740524292,
|
| 3246 |
+
"eval_runtime": 46.3563,
|
| 3247 |
+
"eval_samples_per_second": 139.83,
|
| 3248 |
+
"eval_steps_per_second": 8.758,
|
| 3249 |
+
"step": 108000
|
| 3250 |
+
},
|
| 3251 |
+
{
|
| 3252 |
+
"epoch": 9.02,
|
| 3253 |
+
"learning_rate": 4.919395047365797e-06,
|
| 3254 |
+
"loss": 0.2492,
|
| 3255 |
+
"step": 108500
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 9.02,
|
| 3259 |
+
"eval_f1": 0.4263680579842533,
|
| 3260 |
+
"eval_loss": 3.0018138885498047,
|
| 3261 |
+
"eval_runtime": 46.2406,
|
| 3262 |
+
"eval_samples_per_second": 140.18,
|
| 3263 |
+
"eval_steps_per_second": 8.78,
|
| 3264 |
+
"step": 108500
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 9.06,
|
| 3268 |
+
"learning_rate": 4.711650324081768e-06,
|
| 3269 |
+
"loss": 0.2471,
|
| 3270 |
+
"step": 109000
|
| 3271 |
+
},
|
| 3272 |
+
{
|
| 3273 |
+
"epoch": 9.06,
|
| 3274 |
+
"eval_f1": 0.42881222429550997,
|
| 3275 |
+
"eval_loss": 3.014409303665161,
|
| 3276 |
+
"eval_runtime": 46.1142,
|
| 3277 |
+
"eval_samples_per_second": 140.564,
|
| 3278 |
+
"eval_steps_per_second": 8.804,
|
| 3279 |
+
"step": 109000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 9.1,
|
| 3283 |
+
"learning_rate": 4.50390560079774e-06,
|
| 3284 |
+
"loss": 0.2497,
|
| 3285 |
+
"step": 109500
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 9.1,
|
| 3289 |
+
"eval_f1": 0.4299801659722133,
|
| 3290 |
+
"eval_loss": 3.0282411575317383,
|
| 3291 |
+
"eval_runtime": 46.4599,
|
| 3292 |
+
"eval_samples_per_second": 139.518,
|
| 3293 |
+
"eval_steps_per_second": 8.739,
|
| 3294 |
+
"step": 109500
|
| 3295 |
+
},
|
| 3296 |
+
{
|
| 3297 |
+
"epoch": 9.14,
|
| 3298 |
+
"learning_rate": 4.296160877513712e-06,
|
| 3299 |
+
"loss": 0.2632,
|
| 3300 |
+
"step": 110000
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 9.14,
|
| 3304 |
+
"eval_f1": 0.4285290598627509,
|
| 3305 |
+
"eval_loss": 2.9987945556640625,
|
| 3306 |
+
"eval_runtime": 46.2719,
|
| 3307 |
+
"eval_samples_per_second": 140.085,
|
| 3308 |
+
"eval_steps_per_second": 8.774,
|
| 3309 |
+
"step": 110000
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 9.18,
|
| 3313 |
+
"learning_rate": 4.088416154229683e-06,
|
| 3314 |
+
"loss": 0.2292,
|
| 3315 |
+
"step": 110500
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 9.18,
|
| 3319 |
+
"eval_f1": 0.4295693743332262,
|
| 3320 |
+
"eval_loss": 2.9812545776367188,
|
| 3321 |
+
"eval_runtime": 46.0798,
|
| 3322 |
+
"eval_samples_per_second": 140.669,
|
| 3323 |
+
"eval_steps_per_second": 8.811,
|
| 3324 |
+
"step": 110500
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 9.22,
|
| 3328 |
+
"learning_rate": 3.8806714309456546e-06,
|
| 3329 |
+
"loss": 0.2467,
|
| 3330 |
+
"step": 111000
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 9.22,
|
| 3334 |
+
"eval_f1": 0.4292138712207353,
|
| 3335 |
+
"eval_loss": 2.9471848011016846,
|
| 3336 |
+
"eval_runtime": 46.1858,
|
| 3337 |
+
"eval_samples_per_second": 140.346,
|
| 3338 |
+
"eval_steps_per_second": 8.791,
|
| 3339 |
+
"step": 111000
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 9.27,
|
| 3343 |
+
"learning_rate": 3.672926707661625e-06,
|
| 3344 |
+
"loss": 0.2405,
|
| 3345 |
+
"step": 111500
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 9.27,
|
| 3349 |
+
"eval_f1": 0.4283879568475488,
|
| 3350 |
+
"eval_loss": 3.013239860534668,
|
| 3351 |
+
"eval_runtime": 46.341,
|
| 3352 |
+
"eval_samples_per_second": 139.876,
|
| 3353 |
+
"eval_steps_per_second": 8.761,
|
| 3354 |
+
"step": 111500
|
| 3355 |
+
},
|
| 3356 |
+
{
|
| 3357 |
+
"epoch": 9.31,
|
| 3358 |
+
"learning_rate": 3.465181984377597e-06,
|
| 3359 |
+
"loss": 0.2479,
|
| 3360 |
+
"step": 112000
|
| 3361 |
+
},
|
| 3362 |
+
{
|
| 3363 |
+
"epoch": 9.31,
|
| 3364 |
+
"eval_f1": 0.43018803192623334,
|
| 3365 |
+
"eval_loss": 2.974679946899414,
|
| 3366 |
+
"eval_runtime": 46.0989,
|
| 3367 |
+
"eval_samples_per_second": 140.611,
|
| 3368 |
+
"eval_steps_per_second": 8.807,
|
| 3369 |
+
"step": 112000
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 9.35,
|
| 3373 |
+
"learning_rate": 3.2574372610935682e-06,
|
| 3374 |
+
"loss": 0.2415,
|
| 3375 |
+
"step": 112500
|
| 3376 |
+
},
|
| 3377 |
+
{
|
| 3378 |
+
"epoch": 9.35,
|
| 3379 |
+
"eval_f1": 0.43026318885711246,
|
| 3380 |
+
"eval_loss": 2.9601778984069824,
|
| 3381 |
+
"eval_runtime": 46.5726,
|
| 3382 |
+
"eval_samples_per_second": 139.18,
|
| 3383 |
+
"eval_steps_per_second": 8.718,
|
| 3384 |
+
"step": 112500
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 9.39,
|
| 3388 |
+
"learning_rate": 3.0496925378095396e-06,
|
| 3389 |
+
"loss": 0.2387,
|
| 3390 |
+
"step": 113000
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 9.39,
|
| 3394 |
+
"eval_f1": 0.4279864446994403,
|
| 3395 |
+
"eval_loss": 2.995856285095215,
|
| 3396 |
+
"eval_runtime": 45.9971,
|
| 3397 |
+
"eval_samples_per_second": 140.922,
|
| 3398 |
+
"eval_steps_per_second": 8.827,
|
| 3399 |
+
"step": 113000
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 9.43,
|
| 3403 |
+
"learning_rate": 2.8419478145255114e-06,
|
| 3404 |
+
"loss": 0.2005,
|
| 3405 |
+
"step": 113500
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 9.43,
|
| 3409 |
+
"eval_f1": 0.4226136111773091,
|
| 3410 |
+
"eval_loss": 3.061025857925415,
|
| 3411 |
+
"eval_runtime": 46.3377,
|
| 3412 |
+
"eval_samples_per_second": 139.886,
|
| 3413 |
+
"eval_steps_per_second": 8.762,
|
| 3414 |
+
"step": 113500
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 9.47,
|
| 3418 |
+
"learning_rate": 2.6342030912414823e-06,
|
| 3419 |
+
"loss": 0.2479,
|
| 3420 |
+
"step": 114000
|
| 3421 |
+
},
|
| 3422 |
+
{
|
| 3423 |
+
"epoch": 9.47,
|
| 3424 |
+
"eval_f1": 0.4310947389401135,
|
| 3425 |
+
"eval_loss": 3.0478320121765137,
|
| 3426 |
+
"eval_runtime": 46.1598,
|
| 3427 |
+
"eval_samples_per_second": 140.425,
|
| 3428 |
+
"eval_steps_per_second": 8.796,
|
| 3429 |
+
"step": 114000
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 9.51,
|
| 3433 |
+
"learning_rate": 2.426458367957454e-06,
|
| 3434 |
+
"loss": 0.2649,
|
| 3435 |
+
"step": 114500
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 9.51,
|
| 3439 |
+
"eval_f1": 0.42937936025434953,
|
| 3440 |
+
"eval_loss": 2.964843511581421,
|
| 3441 |
+
"eval_runtime": 45.9242,
|
| 3442 |
+
"eval_samples_per_second": 141.146,
|
| 3443 |
+
"eval_steps_per_second": 8.841,
|
| 3444 |
+
"step": 114500
|
| 3445 |
+
},
|
| 3446 |
+
{
|
| 3447 |
+
"epoch": 9.56,
|
| 3448 |
+
"learning_rate": 2.2187136446734255e-06,
|
| 3449 |
+
"loss": 0.2659,
|
| 3450 |
+
"step": 115000
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 9.56,
|
| 3454 |
+
"eval_f1": 0.42954375020993524,
|
| 3455 |
+
"eval_loss": 2.9387714862823486,
|
| 3456 |
+
"eval_runtime": 46.5353,
|
| 3457 |
+
"eval_samples_per_second": 139.292,
|
| 3458 |
+
"eval_steps_per_second": 8.725,
|
| 3459 |
+
"step": 115000
|
| 3460 |
+
},
|
| 3461 |
+
{
|
| 3462 |
+
"epoch": 9.6,
|
| 3463 |
+
"learning_rate": 2.010968921389397e-06,
|
| 3464 |
+
"loss": 0.2723,
|
| 3465 |
+
"step": 115500
|
| 3466 |
+
},
|
| 3467 |
+
{
|
| 3468 |
+
"epoch": 9.6,
|
| 3469 |
+
"eval_f1": 0.43453017211998707,
|
| 3470 |
+
"eval_loss": 2.942929744720459,
|
| 3471 |
+
"eval_runtime": 45.4198,
|
| 3472 |
+
"eval_samples_per_second": 142.713,
|
| 3473 |
+
"eval_steps_per_second": 8.939,
|
| 3474 |
+
"step": 115500
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 9.64,
|
| 3478 |
+
"learning_rate": 1.8032241981053682e-06,
|
| 3479 |
+
"loss": 0.2607,
|
| 3480 |
+
"step": 116000
|
| 3481 |
+
},
|
| 3482 |
+
{
|
| 3483 |
+
"epoch": 9.64,
|
| 3484 |
+
"eval_f1": 0.43214152175935766,
|
| 3485 |
+
"eval_loss": 2.9575393199920654,
|
| 3486 |
+
"eval_runtime": 46.6531,
|
| 3487 |
+
"eval_samples_per_second": 138.94,
|
| 3488 |
+
"eval_steps_per_second": 8.703,
|
| 3489 |
+
"step": 116000
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 9.68,
|
| 3493 |
+
"learning_rate": 1.5954794748213396e-06,
|
| 3494 |
+
"loss": 0.2284,
|
| 3495 |
+
"step": 116500
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 9.68,
|
| 3499 |
+
"eval_f1": 0.43057572034563696,
|
| 3500 |
+
"eval_loss": 2.9958791732788086,
|
| 3501 |
+
"eval_runtime": 45.9052,
|
| 3502 |
+
"eval_samples_per_second": 141.204,
|
| 3503 |
+
"eval_steps_per_second": 8.844,
|
| 3504 |
+
"step": 116500
|
| 3505 |
+
},
|
| 3506 |
+
{
|
| 3507 |
+
"epoch": 9.72,
|
| 3508 |
+
"learning_rate": 1.387734751537311e-06,
|
| 3509 |
+
"loss": 0.2703,
|
| 3510 |
+
"step": 117000
|
| 3511 |
+
},
|
| 3512 |
+
{
|
| 3513 |
+
"epoch": 9.72,
|
| 3514 |
+
"eval_f1": 0.4305604104069932,
|
| 3515 |
+
"eval_loss": 2.9565792083740234,
|
| 3516 |
+
"eval_runtime": 46.6048,
|
| 3517 |
+
"eval_samples_per_second": 139.084,
|
| 3518 |
+
"eval_steps_per_second": 8.712,
|
| 3519 |
+
"step": 117000
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 9.76,
|
| 3523 |
+
"learning_rate": 1.1799900282532825e-06,
|
| 3524 |
+
"loss": 0.2358,
|
| 3525 |
+
"step": 117500
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"epoch": 9.76,
|
| 3529 |
+
"eval_f1": 0.4299059329453134,
|
| 3530 |
+
"eval_loss": 2.9763364791870117,
|
| 3531 |
+
"eval_runtime": 45.9929,
|
| 3532 |
+
"eval_samples_per_second": 140.935,
|
| 3533 |
+
"eval_steps_per_second": 8.827,
|
| 3534 |
+
"step": 117500
|
| 3535 |
+
},
|
| 3536 |
+
{
|
| 3537 |
+
"epoch": 9.81,
|
| 3538 |
+
"learning_rate": 9.722453049692539e-07,
|
| 3539 |
+
"loss": 0.2633,
|
| 3540 |
+
"step": 118000
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 9.81,
|
| 3544 |
+
"eval_f1": 0.42872674431828584,
|
| 3545 |
+
"eval_loss": 2.9550485610961914,
|
| 3546 |
+
"eval_runtime": 46.191,
|
| 3547 |
+
"eval_samples_per_second": 140.33,
|
| 3548 |
+
"eval_steps_per_second": 8.79,
|
| 3549 |
+
"step": 118000
|
| 3550 |
+
},
|
| 3551 |
+
{
|
| 3552 |
+
"epoch": 9.85,
|
| 3553 |
+
"learning_rate": 7.645005816852252e-07,
|
| 3554 |
+
"loss": 0.2184,
|
| 3555 |
+
"step": 118500
|
| 3556 |
+
},
|
| 3557 |
+
{
|
| 3558 |
+
"epoch": 9.85,
|
| 3559 |
+
"eval_f1": 0.43016074486231115,
|
| 3560 |
+
"eval_loss": 2.9980921745300293,
|
| 3561 |
+
"eval_runtime": 46.1749,
|
| 3562 |
+
"eval_samples_per_second": 140.379,
|
| 3563 |
+
"eval_steps_per_second": 8.793,
|
| 3564 |
+
"step": 118500
|
| 3565 |
+
},
|
| 3566 |
+
{
|
| 3567 |
+
"epoch": 9.89,
|
| 3568 |
+
"learning_rate": 5.567558584011966e-07,
|
| 3569 |
+
"loss": 0.2299,
|
| 3570 |
+
"step": 119000
|
| 3571 |
+
},
|
| 3572 |
+
{
|
| 3573 |
+
"epoch": 9.89,
|
| 3574 |
+
"eval_f1": 0.4306080177005011,
|
| 3575 |
+
"eval_loss": 3.0055530071258545,
|
| 3576 |
+
"eval_runtime": 46.0833,
|
| 3577 |
+
"eval_samples_per_second": 140.658,
|
| 3578 |
+
"eval_steps_per_second": 8.81,
|
| 3579 |
+
"step": 119000
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 9.93,
|
| 3583 |
+
"learning_rate": 3.4901113511716807e-07,
|
| 3584 |
+
"loss": 0.2469,
|
| 3585 |
+
"step": 119500
|
| 3586 |
+
},
|
| 3587 |
+
{
|
| 3588 |
+
"epoch": 9.93,
|
| 3589 |
+
"eval_f1": 0.4310614992045377,
|
| 3590 |
+
"eval_loss": 2.9983813762664795,
|
| 3591 |
+
"eval_runtime": 46.3038,
|
| 3592 |
+
"eval_samples_per_second": 139.989,
|
| 3593 |
+
"eval_steps_per_second": 8.768,
|
| 3594 |
+
"step": 119500
|
| 3595 |
+
},
|
| 3596 |
+
{
|
| 3597 |
+
"epoch": 9.97,
|
| 3598 |
+
"learning_rate": 1.4126641183313946e-07,
|
| 3599 |
+
"loss": 0.2741,
|
| 3600 |
+
"step": 120000
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 9.97,
|
| 3604 |
+
"eval_f1": 0.43169668013947665,
|
| 3605 |
+
"eval_loss": 2.9964375495910645,
|
| 3606 |
+
"eval_runtime": 46.5054,
|
| 3607 |
+
"eval_samples_per_second": 139.382,
|
| 3608 |
+
"eval_steps_per_second": 8.73,
|
| 3609 |
+
"step": 120000
|
| 3610 |
+
},
|
| 3611 |
+
{
|
| 3612 |
+
"epoch": 10.0,
|
| 3613 |
+
"step": 120340,
|
| 3614 |
+
"total_flos": 5.252278015131096e+16,
|
| 3615 |
+
"train_loss": 0.5760954068351747,
|
| 3616 |
+
"train_runtime": 42398.6115,
|
| 3617 |
+
"train_samples_per_second": 22.706,
|
| 3618 |
+
"train_steps_per_second": 2.838
|
| 3619 |
+
}
|
| 3620 |
+
],
|
| 3621 |
+
"logging_steps": 500,
|
| 3622 |
+
"max_steps": 120340,
|
| 3623 |
+
"num_input_tokens_seen": 0,
|
| 3624 |
+
"num_train_epochs": 10,
|
| 3625 |
+
"save_steps": 500,
|
| 3626 |
+
"total_flos": 5.252278015131096e+16,
|
| 3627 |
+
"train_batch_size": 8,
|
| 3628 |
+
"trial_name": null,
|
| 3629 |
+
"trial_params": null
|
| 3630 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba6bf3d96c0b2fd56f62ad1d416e278cba045ea62ebb78f9ec46a8e19ece8939
|
| 3 |
+
size 4728
|