Upload 8 files
Browse files- config (1).json +55 -0
- generation_config (1).json +13 -0
- model (2).safetensors +3 -0
- optimizer (2).pt +3 -0
- rng_state (1).pth +3 -0
- scheduler (1).pt +3 -0
- trainer_state (1).json +558 -0
- training_args (1).bin +3 -0
config (1).json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/pegasus-cnn_dailymail",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 4 |
+
"activation_function": "relu",
|
| 5 |
+
"add_bias_logits": false,
|
| 6 |
+
"add_final_layer_norm": true,
|
| 7 |
+
"architectures": [
|
| 8 |
+
"PegasusForConditionalGeneration"
|
| 9 |
+
],
|
| 10 |
+
"attention_dropout": 0.1,
|
| 11 |
+
"bos_token_id": 0,
|
| 12 |
+
"classif_dropout": 0.0,
|
| 13 |
+
"d_model": 1024,
|
| 14 |
+
"decoder_attention_heads": 16,
|
| 15 |
+
"decoder_ffn_dim": 4096,
|
| 16 |
+
"decoder_layerdrop": 0.0,
|
| 17 |
+
"decoder_layers": 16,
|
| 18 |
+
"decoder_start_token_id": 0,
|
| 19 |
+
"dropout": 0.1,
|
| 20 |
+
"encoder_attention_heads": 16,
|
| 21 |
+
"encoder_ffn_dim": 4096,
|
| 22 |
+
"encoder_layerdrop": 0.0,
|
| 23 |
+
"encoder_layers": 16,
|
| 24 |
+
"eos_token_id": 1,
|
| 25 |
+
"extra_pos_embeddings": 1,
|
| 26 |
+
"forced_eos_token_id": 1,
|
| 27 |
+
"id2label": {
|
| 28 |
+
"0": "LABEL_0",
|
| 29 |
+
"1": "LABEL_1",
|
| 30 |
+
"2": "LABEL_2"
|
| 31 |
+
},
|
| 32 |
+
"init_std": 0.02,
|
| 33 |
+
"is_encoder_decoder": true,
|
| 34 |
+
"label2id": {
|
| 35 |
+
"LABEL_0": 0,
|
| 36 |
+
"LABEL_1": 1,
|
| 37 |
+
"LABEL_2": 2
|
| 38 |
+
},
|
| 39 |
+
"length_penalty": 0.8,
|
| 40 |
+
"max_length": 128,
|
| 41 |
+
"max_position_embeddings": 1024,
|
| 42 |
+
"min_length": 32,
|
| 43 |
+
"model_type": "pegasus",
|
| 44 |
+
"normalize_before": true,
|
| 45 |
+
"normalize_embedding": false,
|
| 46 |
+
"num_beams": 8,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"pad_token_id": 0,
|
| 49 |
+
"scale_embedding": true,
|
| 50 |
+
"static_position_embeddings": true,
|
| 51 |
+
"torch_dtype": "float32",
|
| 52 |
+
"transformers_version": "4.42.3",
|
| 53 |
+
"use_cache": true,
|
| 54 |
+
"vocab_size": 96103
|
| 55 |
+
}
|
generation_config (1).json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 0,
|
| 5 |
+
"eos_token_id": 1,
|
| 6 |
+
"forced_eos_token_id": 1,
|
| 7 |
+
"length_penalty": 0.8,
|
| 8 |
+
"max_length": 128,
|
| 9 |
+
"min_length": 32,
|
| 10 |
+
"num_beams": 8,
|
| 11 |
+
"pad_token_id": 0,
|
| 12 |
+
"transformers_version": "4.42.3"
|
| 13 |
+
}
|
model (2).safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4747afc6984ac4b30b07a179c2ed8742d2b5e814184b3e7ba4dbf36b661d766c
|
| 3 |
+
size 2283652852
|
optimizer (2).pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:863aa77c5a5e7967a28df5206de855e4e26e31cda69d6037f814f71266563682
|
| 3 |
+
size 4550170737
|
rng_state (1).pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6890cd770c43988a22e8135e0169db1a103d4a77cf1e4ac8182960753611c337
|
| 3 |
+
size 14244
|
scheduler (1).pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8908859828769dd951a827564ed7b6826968155da0e4c87a1c3fbda51b1a0f88
|
| 3 |
+
size 1064
|
trainer_state (1).json
ADDED
|
@@ -0,0 +1,558 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 37500,
|
| 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 |
+
"grad_norm": 7.915888786315918,
|
| 14 |
+
"learning_rate": 4.947192513368984e-05,
|
| 15 |
+
"loss": 4.0126,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.08,
|
| 20 |
+
"grad_norm": 3.184572219848633,
|
| 21 |
+
"learning_rate": 4.880347593582888e-05,
|
| 22 |
+
"loss": 0.6684,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.12,
|
| 27 |
+
"grad_norm": 5.76563835144043,
|
| 28 |
+
"learning_rate": 4.813502673796791e-05,
|
| 29 |
+
"loss": 0.5942,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.16,
|
| 34 |
+
"grad_norm": 1.7340965270996094,
|
| 35 |
+
"learning_rate": 4.7466577540106955e-05,
|
| 36 |
+
"loss": 0.5723,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.2,
|
| 41 |
+
"grad_norm": 1.790732502937317,
|
| 42 |
+
"learning_rate": 4.679812834224599e-05,
|
| 43 |
+
"loss": 0.5391,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.24,
|
| 48 |
+
"grad_norm": 2.157850742340088,
|
| 49 |
+
"learning_rate": 4.612967914438503e-05,
|
| 50 |
+
"loss": 0.5188,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.28,
|
| 55 |
+
"grad_norm": 1.461838960647583,
|
| 56 |
+
"learning_rate": 4.5461229946524065e-05,
|
| 57 |
+
"loss": 0.5113,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.32,
|
| 62 |
+
"grad_norm": 1.106745719909668,
|
| 63 |
+
"learning_rate": 4.47927807486631e-05,
|
| 64 |
+
"loss": 0.4956,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.36,
|
| 69 |
+
"grad_norm": 1.2794815301895142,
|
| 70 |
+
"learning_rate": 4.412433155080214e-05,
|
| 71 |
+
"loss": 0.4906,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.4,
|
| 76 |
+
"grad_norm": 1.2381389141082764,
|
| 77 |
+
"learning_rate": 4.345588235294118e-05,
|
| 78 |
+
"loss": 0.4717,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.44,
|
| 83 |
+
"grad_norm": 2.9455885887145996,
|
| 84 |
+
"learning_rate": 4.278743315508021e-05,
|
| 85 |
+
"loss": 0.4721,
|
| 86 |
+
"step": 5500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.48,
|
| 90 |
+
"grad_norm": 1.1448231935501099,
|
| 91 |
+
"learning_rate": 4.2118983957219257e-05,
|
| 92 |
+
"loss": 0.4745,
|
| 93 |
+
"step": 6000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.52,
|
| 97 |
+
"grad_norm": 0.9075339436531067,
|
| 98 |
+
"learning_rate": 4.145053475935829e-05,
|
| 99 |
+
"loss": 0.4597,
|
| 100 |
+
"step": 6500
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.56,
|
| 104 |
+
"grad_norm": 0.962452232837677,
|
| 105 |
+
"learning_rate": 4.078208556149733e-05,
|
| 106 |
+
"loss": 0.4583,
|
| 107 |
+
"step": 7000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.6,
|
| 111 |
+
"grad_norm": 0.7801158428192139,
|
| 112 |
+
"learning_rate": 4.011363636363637e-05,
|
| 113 |
+
"loss": 0.4594,
|
| 114 |
+
"step": 7500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.64,
|
| 118 |
+
"grad_norm": 1.1708343029022217,
|
| 119 |
+
"learning_rate": 3.9445187165775404e-05,
|
| 120 |
+
"loss": 0.45,
|
| 121 |
+
"step": 8000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.68,
|
| 125 |
+
"grad_norm": 1.1591224670410156,
|
| 126 |
+
"learning_rate": 3.877807486631016e-05,
|
| 127 |
+
"loss": 0.45,
|
| 128 |
+
"step": 8500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.72,
|
| 132 |
+
"grad_norm": 0.8542613387107849,
|
| 133 |
+
"learning_rate": 3.81096256684492e-05,
|
| 134 |
+
"loss": 0.4433,
|
| 135 |
+
"step": 9000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.76,
|
| 139 |
+
"grad_norm": 0.8478009700775146,
|
| 140 |
+
"learning_rate": 3.744117647058823e-05,
|
| 141 |
+
"loss": 0.4437,
|
| 142 |
+
"step": 9500
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.8,
|
| 146 |
+
"grad_norm": 0.9733747839927673,
|
| 147 |
+
"learning_rate": 3.677272727272728e-05,
|
| 148 |
+
"loss": 0.4362,
|
| 149 |
+
"step": 10000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.84,
|
| 153 |
+
"grad_norm": 0.7879180312156677,
|
| 154 |
+
"learning_rate": 3.610427807486631e-05,
|
| 155 |
+
"loss": 0.4324,
|
| 156 |
+
"step": 10500
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.88,
|
| 160 |
+
"grad_norm": 1.3161998987197876,
|
| 161 |
+
"learning_rate": 3.543582887700535e-05,
|
| 162 |
+
"loss": 0.4354,
|
| 163 |
+
"step": 11000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.92,
|
| 167 |
+
"grad_norm": 1.153984785079956,
|
| 168 |
+
"learning_rate": 3.476737967914439e-05,
|
| 169 |
+
"loss": 0.4215,
|
| 170 |
+
"step": 11500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.96,
|
| 174 |
+
"grad_norm": 0.7850612998008728,
|
| 175 |
+
"learning_rate": 3.4098930481283425e-05,
|
| 176 |
+
"loss": 0.4298,
|
| 177 |
+
"step": 12000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.0,
|
| 181 |
+
"grad_norm": 1.3549928665161133,
|
| 182 |
+
"learning_rate": 3.343181818181818e-05,
|
| 183 |
+
"loss": 0.4245,
|
| 184 |
+
"step": 12500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 1.04,
|
| 188 |
+
"grad_norm": 0.8405876755714417,
|
| 189 |
+
"learning_rate": 3.2763368983957224e-05,
|
| 190 |
+
"loss": 0.4204,
|
| 191 |
+
"step": 13000
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.08,
|
| 195 |
+
"grad_norm": 1.5088772773742676,
|
| 196 |
+
"learning_rate": 3.2094919786096254e-05,
|
| 197 |
+
"loss": 0.4218,
|
| 198 |
+
"step": 13500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.12,
|
| 202 |
+
"grad_norm": 1.469650149345398,
|
| 203 |
+
"learning_rate": 3.14264705882353e-05,
|
| 204 |
+
"loss": 0.4074,
|
| 205 |
+
"step": 14000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 1.16,
|
| 209 |
+
"grad_norm": 0.8406803607940674,
|
| 210 |
+
"learning_rate": 3.075802139037433e-05,
|
| 211 |
+
"loss": 0.4173,
|
| 212 |
+
"step": 14500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.2,
|
| 216 |
+
"grad_norm": 0.754941999912262,
|
| 217 |
+
"learning_rate": 3.0090909090909093e-05,
|
| 218 |
+
"loss": 0.4095,
|
| 219 |
+
"step": 15000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.24,
|
| 223 |
+
"grad_norm": 0.7950364947319031,
|
| 224 |
+
"learning_rate": 2.9422459893048134e-05,
|
| 225 |
+
"loss": 0.4053,
|
| 226 |
+
"step": 15500
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.28,
|
| 230 |
+
"grad_norm": 0.8804301023483276,
|
| 231 |
+
"learning_rate": 2.8754010695187167e-05,
|
| 232 |
+
"loss": 0.4117,
|
| 233 |
+
"step": 16000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.32,
|
| 237 |
+
"grad_norm": 1.4785621166229248,
|
| 238 |
+
"learning_rate": 2.8085561497326207e-05,
|
| 239 |
+
"loss": 0.4068,
|
| 240 |
+
"step": 16500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.3599999999999999,
|
| 244 |
+
"grad_norm": 1.0254385471343994,
|
| 245 |
+
"learning_rate": 2.7419786096256685e-05,
|
| 246 |
+
"loss": 0.4032,
|
| 247 |
+
"step": 17000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.4,
|
| 251 |
+
"grad_norm": 0.9251853823661804,
|
| 252 |
+
"learning_rate": 2.6751336898395725e-05,
|
| 253 |
+
"loss": 0.4023,
|
| 254 |
+
"step": 17500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.44,
|
| 258 |
+
"grad_norm": 2.7361536026000977,
|
| 259 |
+
"learning_rate": 2.608288770053476e-05,
|
| 260 |
+
"loss": 0.3997,
|
| 261 |
+
"step": 18000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.48,
|
| 265 |
+
"grad_norm": 0.9054092764854431,
|
| 266 |
+
"learning_rate": 2.54144385026738e-05,
|
| 267 |
+
"loss": 0.3981,
|
| 268 |
+
"step": 18500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.52,
|
| 272 |
+
"grad_norm": 0.8799841403961182,
|
| 273 |
+
"learning_rate": 2.4745989304812836e-05,
|
| 274 |
+
"loss": 0.3943,
|
| 275 |
+
"step": 19000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.56,
|
| 279 |
+
"grad_norm": 0.7638904452323914,
|
| 280 |
+
"learning_rate": 2.4077540106951873e-05,
|
| 281 |
+
"loss": 0.3962,
|
| 282 |
+
"step": 19500
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.6,
|
| 286 |
+
"grad_norm": 1.0725598335266113,
|
| 287 |
+
"learning_rate": 2.341042780748663e-05,
|
| 288 |
+
"loss": 0.3918,
|
| 289 |
+
"step": 20000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.6400000000000001,
|
| 293 |
+
"grad_norm": 0.7163866758346558,
|
| 294 |
+
"learning_rate": 2.274197860962567e-05,
|
| 295 |
+
"loss": 0.3987,
|
| 296 |
+
"step": 20500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.6800000000000002,
|
| 300 |
+
"grad_norm": 1.1992738246917725,
|
| 301 |
+
"learning_rate": 2.2073529411764705e-05,
|
| 302 |
+
"loss": 0.3983,
|
| 303 |
+
"step": 21000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.72,
|
| 307 |
+
"grad_norm": 0.9536983966827393,
|
| 308 |
+
"learning_rate": 2.1405080213903746e-05,
|
| 309 |
+
"loss": 0.3961,
|
| 310 |
+
"step": 21500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.76,
|
| 314 |
+
"grad_norm": 0.8780732154846191,
|
| 315 |
+
"learning_rate": 2.0736631016042782e-05,
|
| 316 |
+
"loss": 0.3926,
|
| 317 |
+
"step": 22000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.8,
|
| 321 |
+
"grad_norm": 1.201428771018982,
|
| 322 |
+
"learning_rate": 2.006818181818182e-05,
|
| 323 |
+
"loss": 0.3903,
|
| 324 |
+
"step": 22500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.8399999999999999,
|
| 328 |
+
"grad_norm": 0.9713455438613892,
|
| 329 |
+
"learning_rate": 1.9399732620320856e-05,
|
| 330 |
+
"loss": 0.3977,
|
| 331 |
+
"step": 23000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.88,
|
| 335 |
+
"grad_norm": 1.4893653392791748,
|
| 336 |
+
"learning_rate": 1.8731283422459893e-05,
|
| 337 |
+
"loss": 0.3923,
|
| 338 |
+
"step": 23500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.92,
|
| 342 |
+
"grad_norm": 0.8749503493309021,
|
| 343 |
+
"learning_rate": 1.8062834224598933e-05,
|
| 344 |
+
"loss": 0.3868,
|
| 345 |
+
"step": 24000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.96,
|
| 349 |
+
"grad_norm": 1.0122476816177368,
|
| 350 |
+
"learning_rate": 1.739438502673797e-05,
|
| 351 |
+
"loss": 0.3879,
|
| 352 |
+
"step": 24500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.0,
|
| 356 |
+
"grad_norm": 0.9151819944381714,
|
| 357 |
+
"learning_rate": 1.6725935828877007e-05,
|
| 358 |
+
"loss": 0.3878,
|
| 359 |
+
"step": 25000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 2.04,
|
| 363 |
+
"grad_norm": 0.9272295832633972,
|
| 364 |
+
"learning_rate": 1.6057486631016044e-05,
|
| 365 |
+
"loss": 0.3826,
|
| 366 |
+
"step": 25500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 2.08,
|
| 370 |
+
"grad_norm": 0.9182801842689514,
|
| 371 |
+
"learning_rate": 1.5391711229946525e-05,
|
| 372 |
+
"loss": 0.382,
|
| 373 |
+
"step": 26000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 2.12,
|
| 377 |
+
"grad_norm": 1.7542533874511719,
|
| 378 |
+
"learning_rate": 1.4723262032085564e-05,
|
| 379 |
+
"loss": 0.3786,
|
| 380 |
+
"step": 26500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.16,
|
| 384 |
+
"grad_norm": 1.5916212797164917,
|
| 385 |
+
"learning_rate": 1.4054812834224599e-05,
|
| 386 |
+
"loss": 0.3812,
|
| 387 |
+
"step": 27000
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 2.2,
|
| 391 |
+
"grad_norm": 0.9425985813140869,
|
| 392 |
+
"learning_rate": 1.3386363636363636e-05,
|
| 393 |
+
"loss": 0.3828,
|
| 394 |
+
"step": 27500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 2.24,
|
| 398 |
+
"grad_norm": 0.9243144989013672,
|
| 399 |
+
"learning_rate": 1.2717914438502672e-05,
|
| 400 |
+
"loss": 0.3807,
|
| 401 |
+
"step": 28000
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 2.2800000000000002,
|
| 405 |
+
"grad_norm": 1.0092005729675293,
|
| 406 |
+
"learning_rate": 1.2049465240641713e-05,
|
| 407 |
+
"loss": 0.3714,
|
| 408 |
+
"step": 28500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 2.32,
|
| 412 |
+
"grad_norm": 0.8653339743614197,
|
| 413 |
+
"learning_rate": 1.138101604278075e-05,
|
| 414 |
+
"loss": 0.3792,
|
| 415 |
+
"step": 29000
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 2.36,
|
| 419 |
+
"grad_norm": 0.9960777163505554,
|
| 420 |
+
"learning_rate": 1.0712566844919786e-05,
|
| 421 |
+
"loss": 0.369,
|
| 422 |
+
"step": 29500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.4,
|
| 426 |
+
"grad_norm": 1.3958770036697388,
|
| 427 |
+
"learning_rate": 1.0045454545454545e-05,
|
| 428 |
+
"loss": 0.3742,
|
| 429 |
+
"step": 30000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.44,
|
| 433 |
+
"grad_norm": 0.6805775761604309,
|
| 434 |
+
"learning_rate": 9.377005347593582e-06,
|
| 435 |
+
"loss": 0.3766,
|
| 436 |
+
"step": 30500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 2.48,
|
| 440 |
+
"grad_norm": 2.525830030441284,
|
| 441 |
+
"learning_rate": 8.70855614973262e-06,
|
| 442 |
+
"loss": 0.3781,
|
| 443 |
+
"step": 31000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 2.52,
|
| 447 |
+
"grad_norm": 1.4676679372787476,
|
| 448 |
+
"learning_rate": 8.040106951871658e-06,
|
| 449 |
+
"loss": 0.3758,
|
| 450 |
+
"step": 31500
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.56,
|
| 454 |
+
"grad_norm": 1.1790502071380615,
|
| 455 |
+
"learning_rate": 7.371657754010695e-06,
|
| 456 |
+
"loss": 0.3756,
|
| 457 |
+
"step": 32000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 2.6,
|
| 461 |
+
"grad_norm": 0.7830659747123718,
|
| 462 |
+
"learning_rate": 6.704545454545455e-06,
|
| 463 |
+
"loss": 0.3786,
|
| 464 |
+
"step": 32500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.64,
|
| 468 |
+
"grad_norm": 0.960568904876709,
|
| 469 |
+
"learning_rate": 6.038770053475936e-06,
|
| 470 |
+
"loss": 0.3778,
|
| 471 |
+
"step": 33000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.68,
|
| 475 |
+
"grad_norm": 1.0638457536697388,
|
| 476 |
+
"learning_rate": 5.370320855614974e-06,
|
| 477 |
+
"loss": 0.3826,
|
| 478 |
+
"step": 33500
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 2.7199999999999998,
|
| 482 |
+
"grad_norm": 1.3440908193588257,
|
| 483 |
+
"learning_rate": 4.701871657754011e-06,
|
| 484 |
+
"loss": 0.3697,
|
| 485 |
+
"step": 34000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 2.76,
|
| 489 |
+
"grad_norm": 0.6467347145080566,
|
| 490 |
+
"learning_rate": 4.0334224598930485e-06,
|
| 491 |
+
"loss": 0.38,
|
| 492 |
+
"step": 34500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 2.8,
|
| 496 |
+
"grad_norm": 0.818299412727356,
|
| 497 |
+
"learning_rate": 3.364973262032086e-06,
|
| 498 |
+
"loss": 0.3677,
|
| 499 |
+
"step": 35000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 2.84,
|
| 503 |
+
"grad_norm": 1.8121366500854492,
|
| 504 |
+
"learning_rate": 2.696524064171123e-06,
|
| 505 |
+
"loss": 0.3777,
|
| 506 |
+
"step": 35500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.88,
|
| 510 |
+
"grad_norm": 0.7342619895935059,
|
| 511 |
+
"learning_rate": 2.0280748663101604e-06,
|
| 512 |
+
"loss": 0.3709,
|
| 513 |
+
"step": 36000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 2.92,
|
| 517 |
+
"grad_norm": 0.925221860408783,
|
| 518 |
+
"learning_rate": 1.3596256684491979e-06,
|
| 519 |
+
"loss": 0.3709,
|
| 520 |
+
"step": 36500
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.96,
|
| 524 |
+
"grad_norm": 0.8312897682189941,
|
| 525 |
+
"learning_rate": 6.911764705882354e-07,
|
| 526 |
+
"loss": 0.3725,
|
| 527 |
+
"step": 37000
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.0,
|
| 531 |
+
"grad_norm": 0.9641203880310059,
|
| 532 |
+
"learning_rate": 2.272727272727273e-08,
|
| 533 |
+
"loss": 0.3679,
|
| 534 |
+
"step": 37500
|
| 535 |
+
}
|
| 536 |
+
],
|
| 537 |
+
"logging_steps": 500,
|
| 538 |
+
"max_steps": 37500,
|
| 539 |
+
"num_input_tokens_seen": 0,
|
| 540 |
+
"num_train_epochs": 3,
|
| 541 |
+
"save_steps": 30000,
|
| 542 |
+
"stateful_callbacks": {
|
| 543 |
+
"TrainerControl": {
|
| 544 |
+
"args": {
|
| 545 |
+
"should_epoch_stop": false,
|
| 546 |
+
"should_evaluate": false,
|
| 547 |
+
"should_log": false,
|
| 548 |
+
"should_save": true,
|
| 549 |
+
"should_training_stop": true
|
| 550 |
+
},
|
| 551 |
+
"attributes": {}
|
| 552 |
+
}
|
| 553 |
+
},
|
| 554 |
+
"total_flos": 1.083549155328e+17,
|
| 555 |
+
"train_batch_size": 4,
|
| 556 |
+
"trial_name": null,
|
| 557 |
+
"trial_params": null
|
| 558 |
+
}
|
training_args (1).bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a2dfeac85fddbcb250a4ccfe63c5244bf6ef5e51867af7d50076a6a93d63d97
|
| 3 |
+
size 5112
|