Upload 3 files
Browse files- config.json +25 -0
- pytorch_model.bin +3 -0
- trainer_state.json +976 -0
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.18.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 205
|
| 25 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0b26118656e0870f3da7d66964b1d5933ae960e6f4b9947e2a44b9cd77e722c
|
| 3 |
+
size 344873707
|
trainer_state.json
ADDED
|
@@ -0,0 +1,976 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.758733462167385,
|
| 5 |
+
"global_step": 8000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 5e-06,
|
| 13 |
+
"loss": 5.2948,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 1e-05,
|
| 19 |
+
"loss": 4.9959,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 1.5e-05,
|
| 25 |
+
"loss": 4.6899,
|
| 26 |
+
"step": 150
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.12,
|
| 30 |
+
"learning_rate": 2e-05,
|
| 31 |
+
"loss": 4.5063,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.15,
|
| 36 |
+
"learning_rate": 2.5e-05,
|
| 37 |
+
"loss": 4.4335,
|
| 38 |
+
"step": 250
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.18,
|
| 42 |
+
"learning_rate": 3e-05,
|
| 43 |
+
"loss": 4.3943,
|
| 44 |
+
"step": 300
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.21,
|
| 48 |
+
"learning_rate": 3.5e-05,
|
| 49 |
+
"loss": 4.35,
|
| 50 |
+
"step": 350
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.24,
|
| 54 |
+
"learning_rate": 4e-05,
|
| 55 |
+
"loss": 4.2928,
|
| 56 |
+
"step": 400
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.27,
|
| 60 |
+
"learning_rate": 4.5e-05,
|
| 61 |
+
"loss": 4.2246,
|
| 62 |
+
"step": 450
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.3,
|
| 66 |
+
"learning_rate": 5e-05,
|
| 67 |
+
"loss": 4.1448,
|
| 68 |
+
"step": 500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.33,
|
| 72 |
+
"learning_rate": 5.500000000000001e-05,
|
| 73 |
+
"loss": 4.0853,
|
| 74 |
+
"step": 550
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.36,
|
| 78 |
+
"learning_rate": 6e-05,
|
| 79 |
+
"loss": 4.0358,
|
| 80 |
+
"step": 600
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.39,
|
| 84 |
+
"learning_rate": 6.500000000000001e-05,
|
| 85 |
+
"loss": 3.9763,
|
| 86 |
+
"step": 650
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.42,
|
| 90 |
+
"learning_rate": 7e-05,
|
| 91 |
+
"loss": 3.8816,
|
| 92 |
+
"step": 700
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.45,
|
| 96 |
+
"learning_rate": 7.500000000000001e-05,
|
| 97 |
+
"loss": 3.7246,
|
| 98 |
+
"step": 750
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.48,
|
| 102 |
+
"learning_rate": 8e-05,
|
| 103 |
+
"loss": 3.5085,
|
| 104 |
+
"step": 800
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.51,
|
| 108 |
+
"learning_rate": 8.5e-05,
|
| 109 |
+
"loss": 3.2879,
|
| 110 |
+
"step": 850
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.54,
|
| 114 |
+
"learning_rate": 9e-05,
|
| 115 |
+
"loss": 3.0898,
|
| 116 |
+
"step": 900
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.56,
|
| 120 |
+
"learning_rate": 9.5e-05,
|
| 121 |
+
"loss": 2.8867,
|
| 122 |
+
"step": 950
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.59,
|
| 126 |
+
"learning_rate": 0.0001,
|
| 127 |
+
"loss": 2.6323,
|
| 128 |
+
"step": 1000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.62,
|
| 132 |
+
"learning_rate": 9.932478055367996e-05,
|
| 133 |
+
"loss": 2.3103,
|
| 134 |
+
"step": 1050
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.65,
|
| 138 |
+
"learning_rate": 9.86495611073599e-05,
|
| 139 |
+
"loss": 2.0185,
|
| 140 |
+
"step": 1100
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.68,
|
| 144 |
+
"learning_rate": 9.797434166103985e-05,
|
| 145 |
+
"loss": 1.7382,
|
| 146 |
+
"step": 1150
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.71,
|
| 150 |
+
"learning_rate": 9.729912221471978e-05,
|
| 151 |
+
"loss": 1.5627,
|
| 152 |
+
"step": 1200
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.74,
|
| 156 |
+
"learning_rate": 9.662390276839974e-05,
|
| 157 |
+
"loss": 1.4679,
|
| 158 |
+
"step": 1250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.77,
|
| 162 |
+
"learning_rate": 9.594868332207968e-05,
|
| 163 |
+
"loss": 1.4046,
|
| 164 |
+
"step": 1300
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.8,
|
| 168 |
+
"learning_rate": 9.527346387575963e-05,
|
| 169 |
+
"loss": 1.3544,
|
| 170 |
+
"step": 1350
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.83,
|
| 174 |
+
"learning_rate": 9.459824442943957e-05,
|
| 175 |
+
"loss": 1.325,
|
| 176 |
+
"step": 1400
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.86,
|
| 180 |
+
"learning_rate": 9.392302498311952e-05,
|
| 181 |
+
"loss": 1.2984,
|
| 182 |
+
"step": 1450
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.89,
|
| 186 |
+
"learning_rate": 9.324780553679947e-05,
|
| 187 |
+
"loss": 1.2735,
|
| 188 |
+
"step": 1500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.92,
|
| 192 |
+
"learning_rate": 9.25725860904794e-05,
|
| 193 |
+
"loss": 1.2617,
|
| 194 |
+
"step": 1550
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.95,
|
| 198 |
+
"learning_rate": 9.189736664415936e-05,
|
| 199 |
+
"loss": 1.2475,
|
| 200 |
+
"step": 1600
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.98,
|
| 204 |
+
"learning_rate": 9.12221471978393e-05,
|
| 205 |
+
"loss": 1.2348,
|
| 206 |
+
"step": 1650
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.01,
|
| 210 |
+
"learning_rate": 9.054692775151925e-05,
|
| 211 |
+
"loss": 1.2387,
|
| 212 |
+
"step": 1700
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.04,
|
| 216 |
+
"learning_rate": 8.98717083051992e-05,
|
| 217 |
+
"loss": 1.2149,
|
| 218 |
+
"step": 1750
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.07,
|
| 222 |
+
"learning_rate": 8.919648885887914e-05,
|
| 223 |
+
"loss": 1.2004,
|
| 224 |
+
"step": 1800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.1,
|
| 228 |
+
"learning_rate": 8.852126941255908e-05,
|
| 229 |
+
"loss": 1.1918,
|
| 230 |
+
"step": 1850
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.13,
|
| 234 |
+
"learning_rate": 8.784604996623903e-05,
|
| 235 |
+
"loss": 1.1825,
|
| 236 |
+
"step": 1900
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.16,
|
| 240 |
+
"learning_rate": 8.717083051991897e-05,
|
| 241 |
+
"loss": 1.1794,
|
| 242 |
+
"step": 1950
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.19,
|
| 246 |
+
"learning_rate": 8.649561107359893e-05,
|
| 247 |
+
"loss": 1.175,
|
| 248 |
+
"step": 2000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.22,
|
| 252 |
+
"learning_rate": 8.582039162727888e-05,
|
| 253 |
+
"loss": 1.1657,
|
| 254 |
+
"step": 2050
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.25,
|
| 258 |
+
"learning_rate": 8.514517218095882e-05,
|
| 259 |
+
"loss": 1.1621,
|
| 260 |
+
"step": 2100
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.28,
|
| 264 |
+
"learning_rate": 8.446995273463876e-05,
|
| 265 |
+
"loss": 1.1564,
|
| 266 |
+
"step": 2150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.31,
|
| 270 |
+
"learning_rate": 8.37947332883187e-05,
|
| 271 |
+
"loss": 1.149,
|
| 272 |
+
"step": 2200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.34,
|
| 276 |
+
"learning_rate": 8.311951384199866e-05,
|
| 277 |
+
"loss": 1.1478,
|
| 278 |
+
"step": 2250
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.37,
|
| 282 |
+
"learning_rate": 8.24442943956786e-05,
|
| 283 |
+
"loss": 1.1449,
|
| 284 |
+
"step": 2300
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.4,
|
| 288 |
+
"learning_rate": 8.176907494935855e-05,
|
| 289 |
+
"loss": 1.1393,
|
| 290 |
+
"step": 2350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.43,
|
| 294 |
+
"learning_rate": 8.109385550303849e-05,
|
| 295 |
+
"loss": 1.135,
|
| 296 |
+
"step": 2400
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.46,
|
| 300 |
+
"learning_rate": 8.041863605671843e-05,
|
| 301 |
+
"loss": 1.1308,
|
| 302 |
+
"step": 2450
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.49,
|
| 306 |
+
"learning_rate": 7.974341661039839e-05,
|
| 307 |
+
"loss": 1.1282,
|
| 308 |
+
"step": 2500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.52,
|
| 312 |
+
"learning_rate": 7.906819716407833e-05,
|
| 313 |
+
"loss": 1.1197,
|
| 314 |
+
"step": 2550
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.55,
|
| 318 |
+
"learning_rate": 7.839297771775828e-05,
|
| 319 |
+
"loss": 1.1227,
|
| 320 |
+
"step": 2600
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.58,
|
| 324 |
+
"learning_rate": 7.771775827143822e-05,
|
| 325 |
+
"loss": 1.1192,
|
| 326 |
+
"step": 2650
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.61,
|
| 330 |
+
"learning_rate": 7.704253882511818e-05,
|
| 331 |
+
"loss": 1.1109,
|
| 332 |
+
"step": 2700
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.64,
|
| 336 |
+
"learning_rate": 7.63673193787981e-05,
|
| 337 |
+
"loss": 1.1166,
|
| 338 |
+
"step": 2750
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.67,
|
| 342 |
+
"learning_rate": 7.569209993247806e-05,
|
| 343 |
+
"loss": 1.1132,
|
| 344 |
+
"step": 2800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.7,
|
| 348 |
+
"learning_rate": 7.5016880486158e-05,
|
| 349 |
+
"loss": 1.1043,
|
| 350 |
+
"step": 2850
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.72,
|
| 354 |
+
"learning_rate": 7.434166103983795e-05,
|
| 355 |
+
"loss": 1.1085,
|
| 356 |
+
"step": 2900
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.75,
|
| 360 |
+
"learning_rate": 7.366644159351789e-05,
|
| 361 |
+
"loss": 1.1018,
|
| 362 |
+
"step": 2950
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.78,
|
| 366 |
+
"learning_rate": 7.299122214719785e-05,
|
| 367 |
+
"loss": 1.1034,
|
| 368 |
+
"step": 3000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.81,
|
| 372 |
+
"learning_rate": 7.23160027008778e-05,
|
| 373 |
+
"loss": 1.0954,
|
| 374 |
+
"step": 3050
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.84,
|
| 378 |
+
"learning_rate": 7.164078325455773e-05,
|
| 379 |
+
"loss": 1.0973,
|
| 380 |
+
"step": 3100
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.87,
|
| 384 |
+
"learning_rate": 7.096556380823768e-05,
|
| 385 |
+
"loss": 1.0971,
|
| 386 |
+
"step": 3150
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 1.9,
|
| 390 |
+
"learning_rate": 7.029034436191762e-05,
|
| 391 |
+
"loss": 1.0893,
|
| 392 |
+
"step": 3200
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.93,
|
| 396 |
+
"learning_rate": 6.961512491559758e-05,
|
| 397 |
+
"loss": 1.0891,
|
| 398 |
+
"step": 3250
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.96,
|
| 402 |
+
"learning_rate": 6.893990546927752e-05,
|
| 403 |
+
"loss": 1.0866,
|
| 404 |
+
"step": 3300
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.99,
|
| 408 |
+
"learning_rate": 6.826468602295747e-05,
|
| 409 |
+
"loss": 1.0938,
|
| 410 |
+
"step": 3350
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.02,
|
| 414 |
+
"learning_rate": 6.758946657663741e-05,
|
| 415 |
+
"loss": 1.1025,
|
| 416 |
+
"step": 3400
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.05,
|
| 420 |
+
"learning_rate": 6.691424713031735e-05,
|
| 421 |
+
"loss": 1.0858,
|
| 422 |
+
"step": 3450
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.08,
|
| 426 |
+
"learning_rate": 6.623902768399731e-05,
|
| 427 |
+
"loss": 1.0835,
|
| 428 |
+
"step": 3500
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.11,
|
| 432 |
+
"learning_rate": 6.556380823767725e-05,
|
| 433 |
+
"loss": 1.0793,
|
| 434 |
+
"step": 3550
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.14,
|
| 438 |
+
"learning_rate": 6.48885887913572e-05,
|
| 439 |
+
"loss": 1.0754,
|
| 440 |
+
"step": 3600
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.17,
|
| 444 |
+
"learning_rate": 6.421336934503714e-05,
|
| 445 |
+
"loss": 1.073,
|
| 446 |
+
"step": 3650
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.2,
|
| 450 |
+
"learning_rate": 6.353814989871708e-05,
|
| 451 |
+
"loss": 1.0774,
|
| 452 |
+
"step": 3700
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.23,
|
| 456 |
+
"learning_rate": 6.286293045239702e-05,
|
| 457 |
+
"loss": 1.0763,
|
| 458 |
+
"step": 3750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.26,
|
| 462 |
+
"learning_rate": 6.218771100607698e-05,
|
| 463 |
+
"loss": 1.0693,
|
| 464 |
+
"step": 3800
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.29,
|
| 468 |
+
"learning_rate": 6.151249155975692e-05,
|
| 469 |
+
"loss": 1.0736,
|
| 470 |
+
"step": 3850
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.32,
|
| 474 |
+
"learning_rate": 6.083727211343687e-05,
|
| 475 |
+
"loss": 1.0762,
|
| 476 |
+
"step": 3900
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.35,
|
| 480 |
+
"learning_rate": 6.016205266711682e-05,
|
| 481 |
+
"loss": 1.0691,
|
| 482 |
+
"step": 3950
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.38,
|
| 486 |
+
"learning_rate": 5.948683322079677e-05,
|
| 487 |
+
"loss": 1.0668,
|
| 488 |
+
"step": 4000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.41,
|
| 492 |
+
"learning_rate": 5.88116137744767e-05,
|
| 493 |
+
"loss": 1.0638,
|
| 494 |
+
"step": 4050
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.44,
|
| 498 |
+
"learning_rate": 5.813639432815665e-05,
|
| 499 |
+
"loss": 1.0671,
|
| 500 |
+
"step": 4100
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.47,
|
| 504 |
+
"learning_rate": 5.7461174881836596e-05,
|
| 505 |
+
"loss": 1.0595,
|
| 506 |
+
"step": 4150
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.5,
|
| 510 |
+
"learning_rate": 5.6785955435516544e-05,
|
| 511 |
+
"loss": 1.0606,
|
| 512 |
+
"step": 4200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.53,
|
| 516 |
+
"learning_rate": 5.61107359891965e-05,
|
| 517 |
+
"loss": 1.0621,
|
| 518 |
+
"step": 4250
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.56,
|
| 522 |
+
"learning_rate": 5.5435516542876445e-05,
|
| 523 |
+
"loss": 1.0633,
|
| 524 |
+
"step": 4300
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.59,
|
| 528 |
+
"learning_rate": 5.476029709655638e-05,
|
| 529 |
+
"loss": 1.056,
|
| 530 |
+
"step": 4350
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.62,
|
| 534 |
+
"learning_rate": 5.4085077650236326e-05,
|
| 535 |
+
"loss": 1.0598,
|
| 536 |
+
"step": 4400
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.65,
|
| 540 |
+
"learning_rate": 5.3409858203916274e-05,
|
| 541 |
+
"loss": 1.0532,
|
| 542 |
+
"step": 4450
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.68,
|
| 546 |
+
"learning_rate": 5.273463875759622e-05,
|
| 547 |
+
"loss": 1.0536,
|
| 548 |
+
"step": 4500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.71,
|
| 552 |
+
"learning_rate": 5.205941931127617e-05,
|
| 553 |
+
"loss": 1.0522,
|
| 554 |
+
"step": 4550
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.74,
|
| 558 |
+
"learning_rate": 5.1384199864956116e-05,
|
| 559 |
+
"loss": 1.0478,
|
| 560 |
+
"step": 4600
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.77,
|
| 564 |
+
"learning_rate": 5.0708980418636057e-05,
|
| 565 |
+
"loss": 1.0514,
|
| 566 |
+
"step": 4650
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.8,
|
| 570 |
+
"learning_rate": 5.0033760972316004e-05,
|
| 571 |
+
"loss": 1.0526,
|
| 572 |
+
"step": 4700
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.83,
|
| 576 |
+
"learning_rate": 4.935854152599595e-05,
|
| 577 |
+
"loss": 1.0476,
|
| 578 |
+
"step": 4750
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.86,
|
| 582 |
+
"learning_rate": 4.86833220796759e-05,
|
| 583 |
+
"loss": 1.0459,
|
| 584 |
+
"step": 4800
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.88,
|
| 588 |
+
"learning_rate": 4.8008102633355846e-05,
|
| 589 |
+
"loss": 1.0463,
|
| 590 |
+
"step": 4850
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.91,
|
| 594 |
+
"learning_rate": 4.733288318703579e-05,
|
| 595 |
+
"loss": 1.0469,
|
| 596 |
+
"step": 4900
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.94,
|
| 600 |
+
"learning_rate": 4.6657663740715734e-05,
|
| 601 |
+
"loss": 1.0444,
|
| 602 |
+
"step": 4950
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.97,
|
| 606 |
+
"learning_rate": 4.598244429439568e-05,
|
| 607 |
+
"loss": 1.0461,
|
| 608 |
+
"step": 5000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.0,
|
| 612 |
+
"learning_rate": 4.530722484807562e-05,
|
| 613 |
+
"loss": 1.0613,
|
| 614 |
+
"step": 5050
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.03,
|
| 618 |
+
"learning_rate": 4.463200540175557e-05,
|
| 619 |
+
"loss": 1.0418,
|
| 620 |
+
"step": 5100
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.06,
|
| 624 |
+
"learning_rate": 4.395678595543552e-05,
|
| 625 |
+
"loss": 1.0415,
|
| 626 |
+
"step": 5150
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.09,
|
| 630 |
+
"learning_rate": 4.3281566509115464e-05,
|
| 631 |
+
"loss": 1.0401,
|
| 632 |
+
"step": 5200
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.12,
|
| 636 |
+
"learning_rate": 4.260634706279541e-05,
|
| 637 |
+
"loss": 1.0404,
|
| 638 |
+
"step": 5250
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.15,
|
| 642 |
+
"learning_rate": 4.193112761647536e-05,
|
| 643 |
+
"loss": 1.0372,
|
| 644 |
+
"step": 5300
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.18,
|
| 648 |
+
"learning_rate": 4.125590817015531e-05,
|
| 649 |
+
"loss": 1.0423,
|
| 650 |
+
"step": 5350
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.21,
|
| 654 |
+
"learning_rate": 4.058068872383525e-05,
|
| 655 |
+
"loss": 1.0343,
|
| 656 |
+
"step": 5400
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.24,
|
| 660 |
+
"learning_rate": 3.9905469277515195e-05,
|
| 661 |
+
"loss": 1.0374,
|
| 662 |
+
"step": 5450
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.27,
|
| 666 |
+
"learning_rate": 3.923024983119514e-05,
|
| 667 |
+
"loss": 1.0355,
|
| 668 |
+
"step": 5500
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.3,
|
| 672 |
+
"learning_rate": 3.855503038487508e-05,
|
| 673 |
+
"loss": 1.0312,
|
| 674 |
+
"step": 5550
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.33,
|
| 678 |
+
"learning_rate": 3.787981093855503e-05,
|
| 679 |
+
"loss": 1.0374,
|
| 680 |
+
"step": 5600
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.36,
|
| 684 |
+
"learning_rate": 3.720459149223498e-05,
|
| 685 |
+
"loss": 1.034,
|
| 686 |
+
"step": 5650
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.39,
|
| 690 |
+
"learning_rate": 3.6529372045914925e-05,
|
| 691 |
+
"loss": 1.0327,
|
| 692 |
+
"step": 5700
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.42,
|
| 696 |
+
"learning_rate": 3.585415259959487e-05,
|
| 697 |
+
"loss": 1.0317,
|
| 698 |
+
"step": 5750
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.45,
|
| 702 |
+
"learning_rate": 3.517893315327482e-05,
|
| 703 |
+
"loss": 1.0317,
|
| 704 |
+
"step": 5800
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.48,
|
| 708 |
+
"learning_rate": 3.450371370695476e-05,
|
| 709 |
+
"loss": 1.028,
|
| 710 |
+
"step": 5850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.51,
|
| 714 |
+
"learning_rate": 3.382849426063471e-05,
|
| 715 |
+
"loss": 1.027,
|
| 716 |
+
"step": 5900
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.54,
|
| 720 |
+
"learning_rate": 3.3153274814314655e-05,
|
| 721 |
+
"loss": 1.0253,
|
| 722 |
+
"step": 5950
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.57,
|
| 726 |
+
"learning_rate": 3.2478055367994596e-05,
|
| 727 |
+
"loss": 1.0291,
|
| 728 |
+
"step": 6000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.6,
|
| 732 |
+
"learning_rate": 3.180283592167454e-05,
|
| 733 |
+
"loss": 1.0266,
|
| 734 |
+
"step": 6050
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.63,
|
| 738 |
+
"learning_rate": 3.112761647535449e-05,
|
| 739 |
+
"loss": 1.0311,
|
| 740 |
+
"step": 6100
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.66,
|
| 744 |
+
"learning_rate": 3.0452397029034435e-05,
|
| 745 |
+
"loss": 1.0281,
|
| 746 |
+
"step": 6150
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.69,
|
| 750 |
+
"learning_rate": 2.9777177582714382e-05,
|
| 751 |
+
"loss": 1.0238,
|
| 752 |
+
"step": 6200
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.72,
|
| 756 |
+
"learning_rate": 2.910195813639433e-05,
|
| 757 |
+
"loss": 1.0248,
|
| 758 |
+
"step": 6250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 3.75,
|
| 762 |
+
"learning_rate": 2.8426738690074277e-05,
|
| 763 |
+
"loss": 1.0235,
|
| 764 |
+
"step": 6300
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 3.78,
|
| 768 |
+
"learning_rate": 2.775151924375422e-05,
|
| 769 |
+
"loss": 1.0252,
|
| 770 |
+
"step": 6350
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 3.81,
|
| 774 |
+
"learning_rate": 2.7076299797434168e-05,
|
| 775 |
+
"loss": 1.0221,
|
| 776 |
+
"step": 6400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 3.84,
|
| 780 |
+
"learning_rate": 2.6401080351114116e-05,
|
| 781 |
+
"loss": 1.0204,
|
| 782 |
+
"step": 6450
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.87,
|
| 786 |
+
"learning_rate": 2.572586090479406e-05,
|
| 787 |
+
"loss": 1.0205,
|
| 788 |
+
"step": 6500
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 3.9,
|
| 792 |
+
"learning_rate": 2.5050641458474007e-05,
|
| 793 |
+
"loss": 1.0207,
|
| 794 |
+
"step": 6550
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.93,
|
| 798 |
+
"learning_rate": 2.437542201215395e-05,
|
| 799 |
+
"loss": 1.0165,
|
| 800 |
+
"step": 6600
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.96,
|
| 804 |
+
"learning_rate": 2.37002025658339e-05,
|
| 805 |
+
"loss": 1.0182,
|
| 806 |
+
"step": 6650
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.99,
|
| 810 |
+
"learning_rate": 2.3024983119513842e-05,
|
| 811 |
+
"loss": 1.0165,
|
| 812 |
+
"step": 6700
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.02,
|
| 816 |
+
"learning_rate": 2.234976367319379e-05,
|
| 817 |
+
"loss": 1.0316,
|
| 818 |
+
"step": 6750
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.05,
|
| 822 |
+
"learning_rate": 2.1674544226873737e-05,
|
| 823 |
+
"loss": 1.0146,
|
| 824 |
+
"step": 6800
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.07,
|
| 828 |
+
"learning_rate": 2.099932478055368e-05,
|
| 829 |
+
"loss": 1.015,
|
| 830 |
+
"step": 6850
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.1,
|
| 834 |
+
"learning_rate": 2.0324105334233625e-05,
|
| 835 |
+
"loss": 1.0156,
|
| 836 |
+
"step": 6900
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.13,
|
| 840 |
+
"learning_rate": 1.9648885887913573e-05,
|
| 841 |
+
"loss": 1.016,
|
| 842 |
+
"step": 6950
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.16,
|
| 846 |
+
"learning_rate": 1.8973666441593517e-05,
|
| 847 |
+
"loss": 1.0149,
|
| 848 |
+
"step": 7000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.19,
|
| 852 |
+
"learning_rate": 1.8298446995273467e-05,
|
| 853 |
+
"loss": 1.0127,
|
| 854 |
+
"step": 7050
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.22,
|
| 858 |
+
"learning_rate": 1.762322754895341e-05,
|
| 859 |
+
"loss": 1.0085,
|
| 860 |
+
"step": 7100
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.25,
|
| 864 |
+
"learning_rate": 1.6948008102633355e-05,
|
| 865 |
+
"loss": 1.0151,
|
| 866 |
+
"step": 7150
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.28,
|
| 870 |
+
"learning_rate": 1.6272788656313303e-05,
|
| 871 |
+
"loss": 1.0136,
|
| 872 |
+
"step": 7200
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.31,
|
| 876 |
+
"learning_rate": 1.5597569209993247e-05,
|
| 877 |
+
"loss": 1.0077,
|
| 878 |
+
"step": 7250
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.34,
|
| 882 |
+
"learning_rate": 1.4922349763673194e-05,
|
| 883 |
+
"loss": 1.0103,
|
| 884 |
+
"step": 7300
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.37,
|
| 888 |
+
"learning_rate": 1.4247130317353142e-05,
|
| 889 |
+
"loss": 1.0155,
|
| 890 |
+
"step": 7350
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.4,
|
| 894 |
+
"learning_rate": 1.3571910871033086e-05,
|
| 895 |
+
"loss": 1.0098,
|
| 896 |
+
"step": 7400
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.43,
|
| 900 |
+
"learning_rate": 1.2896691424713031e-05,
|
| 901 |
+
"loss": 1.0093,
|
| 902 |
+
"step": 7450
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.46,
|
| 906 |
+
"learning_rate": 1.2221471978392979e-05,
|
| 907 |
+
"loss": 1.0112,
|
| 908 |
+
"step": 7500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.49,
|
| 912 |
+
"learning_rate": 1.1546252532072925e-05,
|
| 913 |
+
"loss": 1.0075,
|
| 914 |
+
"step": 7550
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.52,
|
| 918 |
+
"learning_rate": 1.087103308575287e-05,
|
| 919 |
+
"loss": 1.0086,
|
| 920 |
+
"step": 7600
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.55,
|
| 924 |
+
"learning_rate": 1.0195813639432816e-05,
|
| 925 |
+
"loss": 1.0072,
|
| 926 |
+
"step": 7650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.58,
|
| 930 |
+
"learning_rate": 9.520594193112763e-06,
|
| 931 |
+
"loss": 1.0057,
|
| 932 |
+
"step": 7700
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.61,
|
| 936 |
+
"learning_rate": 8.845374746792707e-06,
|
| 937 |
+
"loss": 1.0101,
|
| 938 |
+
"step": 7750
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.64,
|
| 942 |
+
"learning_rate": 8.170155300472653e-06,
|
| 943 |
+
"loss": 1.006,
|
| 944 |
+
"step": 7800
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 4.67,
|
| 948 |
+
"learning_rate": 7.4949358541526005e-06,
|
| 949 |
+
"loss": 1.0079,
|
| 950 |
+
"step": 7850
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 4.7,
|
| 954 |
+
"learning_rate": 6.819716407832546e-06,
|
| 955 |
+
"loss": 1.0066,
|
| 956 |
+
"step": 7900
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 4.73,
|
| 960 |
+
"learning_rate": 6.144496961512492e-06,
|
| 961 |
+
"loss": 1.0113,
|
| 962 |
+
"step": 7950
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 4.76,
|
| 966 |
+
"learning_rate": 5.4692775151924376e-06,
|
| 967 |
+
"loss": 1.0065,
|
| 968 |
+
"step": 8000
|
| 969 |
+
}
|
| 970 |
+
],
|
| 971 |
+
"max_steps": 8405,
|
| 972 |
+
"num_train_epochs": 5,
|
| 973 |
+
"total_flos": 5.140843815167534e+17,
|
| 974 |
+
"trial_name": null,
|
| 975 |
+
"trial_params": null
|
| 976 |
+
}
|