Commit
•
3887733
1
Parent(s):
fcaa1fc
Model save
Browse files- last-checkpoint/config.json +0 -56
- last-checkpoint/optimizer.pt +0 -3
- last-checkpoint/pytorch_model.bin +0 -3
- last-checkpoint/rng_state.pth +0 -3
- last-checkpoint/scaler.pt +0 -3
- last-checkpoint/scheduler.pt +0 -3
- last-checkpoint/special_tokens_map.json +0 -110
- last-checkpoint/spiece.model +0 -3
- last-checkpoint/tokenizer.json +0 -0
- last-checkpoint/tokenizer_config.json +0 -118
- last-checkpoint/trainer_state.json +0 -35
- last-checkpoint/training_args.bin +0 -3
- pytorch_model.bin +1 -1
- trainer_state.json +130 -0
last-checkpoint/config.json
DELETED
|
@@ -1,56 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"_name_or_path": "google/pegasus-newsroom",
|
| 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 |
-
"classifier_dropout": 0.0,
|
| 14 |
-
"d_model": 1024,
|
| 15 |
-
"decoder_attention_heads": 16,
|
| 16 |
-
"decoder_ffn_dim": 4096,
|
| 17 |
-
"decoder_layerdrop": 0.0,
|
| 18 |
-
"decoder_layers": 16,
|
| 19 |
-
"decoder_start_token_id": 0,
|
| 20 |
-
"dropout": 0.1,
|
| 21 |
-
"encoder_attention_heads": 16,
|
| 22 |
-
"encoder_ffn_dim": 4096,
|
| 23 |
-
"encoder_layerdrop": 0.0,
|
| 24 |
-
"encoder_layers": 16,
|
| 25 |
-
"eos_token_id": 1,
|
| 26 |
-
"extra_pos_embeddings": 1,
|
| 27 |
-
"forced_eos_token_id": 1,
|
| 28 |
-
"id2label": {
|
| 29 |
-
"0": "LABEL_0",
|
| 30 |
-
"1": "LABEL_1",
|
| 31 |
-
"2": "LABEL_2"
|
| 32 |
-
},
|
| 33 |
-
"init_std": 0.02,
|
| 34 |
-
"is_encoder_decoder": true,
|
| 35 |
-
"label2id": {
|
| 36 |
-
"LABEL_0": 0,
|
| 37 |
-
"LABEL_1": 1,
|
| 38 |
-
"LABEL_2": 2
|
| 39 |
-
},
|
| 40 |
-
"length_penalty": 0.8,
|
| 41 |
-
"max_length": 128,
|
| 42 |
-
"max_position_embeddings": 512,
|
| 43 |
-
"min_length": 32,
|
| 44 |
-
"model_type": "pegasus",
|
| 45 |
-
"normalize_before": true,
|
| 46 |
-
"normalize_embedding": false,
|
| 47 |
-
"num_beams": 2,
|
| 48 |
-
"num_hidden_layers": 16,
|
| 49 |
-
"pad_token_id": 0,
|
| 50 |
-
"scale_embedding": true,
|
| 51 |
-
"static_position_embeddings": true,
|
| 52 |
-
"torch_dtype": "float32",
|
| 53 |
-
"transformers_version": "4.20.1",
|
| 54 |
-
"use_cache": true,
|
| 55 |
-
"vocab_size": 96103
|
| 56 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:229c943157d5a5e9e8991cc8900ef56099b6776944489977abfc2030ac563ae0
|
| 3 |
-
size 5840398
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:d3ee858dd84837f7c4c8c190b5f93e8054b41ccf8a21bc8e242fa28cf589305f
|
| 3 |
-
size 2279605745
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:4a14935a413905e218f61bc57662a288c912cfa6d0939815b278dfb74ad02841
|
| 3 |
-
size 14439
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scaler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:d635e22380cc34a06be3e38f1e8aab1cd60caff2247be4baed053a9eeadd0935
|
| 3 |
-
size 559
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:2fb8c2cfdbb7789c5d212766c1d81f546c527c967cb92c197f611d2e7cbd5f8d
|
| 3 |
-
size 623
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/special_tokens_map.json
DELETED
|
@@ -1,110 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"additional_special_tokens": [
|
| 3 |
-
"<mask_1>",
|
| 4 |
-
"<unk_2>",
|
| 5 |
-
"<unk_3>",
|
| 6 |
-
"<unk_4>",
|
| 7 |
-
"<unk_5>",
|
| 8 |
-
"<unk_6>",
|
| 9 |
-
"<unk_7>",
|
| 10 |
-
"<unk_8>",
|
| 11 |
-
"<unk_9>",
|
| 12 |
-
"<unk_10>",
|
| 13 |
-
"<unk_11>",
|
| 14 |
-
"<unk_12>",
|
| 15 |
-
"<unk_13>",
|
| 16 |
-
"<unk_14>",
|
| 17 |
-
"<unk_15>",
|
| 18 |
-
"<unk_16>",
|
| 19 |
-
"<unk_17>",
|
| 20 |
-
"<unk_18>",
|
| 21 |
-
"<unk_19>",
|
| 22 |
-
"<unk_20>",
|
| 23 |
-
"<unk_21>",
|
| 24 |
-
"<unk_22>",
|
| 25 |
-
"<unk_23>",
|
| 26 |
-
"<unk_24>",
|
| 27 |
-
"<unk_25>",
|
| 28 |
-
"<unk_26>",
|
| 29 |
-
"<unk_27>",
|
| 30 |
-
"<unk_28>",
|
| 31 |
-
"<unk_29>",
|
| 32 |
-
"<unk_30>",
|
| 33 |
-
"<unk_31>",
|
| 34 |
-
"<unk_32>",
|
| 35 |
-
"<unk_33>",
|
| 36 |
-
"<unk_34>",
|
| 37 |
-
"<unk_35>",
|
| 38 |
-
"<unk_36>",
|
| 39 |
-
"<unk_37>",
|
| 40 |
-
"<unk_38>",
|
| 41 |
-
"<unk_39>",
|
| 42 |
-
"<unk_40>",
|
| 43 |
-
"<unk_41>",
|
| 44 |
-
"<unk_42>",
|
| 45 |
-
"<unk_43>",
|
| 46 |
-
"<unk_44>",
|
| 47 |
-
"<unk_45>",
|
| 48 |
-
"<unk_46>",
|
| 49 |
-
"<unk_47>",
|
| 50 |
-
"<unk_48>",
|
| 51 |
-
"<unk_49>",
|
| 52 |
-
"<unk_50>",
|
| 53 |
-
"<unk_51>",
|
| 54 |
-
"<unk_52>",
|
| 55 |
-
"<unk_53>",
|
| 56 |
-
"<unk_54>",
|
| 57 |
-
"<unk_55>",
|
| 58 |
-
"<unk_56>",
|
| 59 |
-
"<unk_57>",
|
| 60 |
-
"<unk_58>",
|
| 61 |
-
"<unk_59>",
|
| 62 |
-
"<unk_60>",
|
| 63 |
-
"<unk_61>",
|
| 64 |
-
"<unk_62>",
|
| 65 |
-
"<unk_63>",
|
| 66 |
-
"<unk_64>",
|
| 67 |
-
"<unk_65>",
|
| 68 |
-
"<unk_66>",
|
| 69 |
-
"<unk_67>",
|
| 70 |
-
"<unk_68>",
|
| 71 |
-
"<unk_69>",
|
| 72 |
-
"<unk_70>",
|
| 73 |
-
"<unk_71>",
|
| 74 |
-
"<unk_72>",
|
| 75 |
-
"<unk_73>",
|
| 76 |
-
"<unk_74>",
|
| 77 |
-
"<unk_75>",
|
| 78 |
-
"<unk_76>",
|
| 79 |
-
"<unk_77>",
|
| 80 |
-
"<unk_78>",
|
| 81 |
-
"<unk_79>",
|
| 82 |
-
"<unk_80>",
|
| 83 |
-
"<unk_81>",
|
| 84 |
-
"<unk_82>",
|
| 85 |
-
"<unk_83>",
|
| 86 |
-
"<unk_84>",
|
| 87 |
-
"<unk_85>",
|
| 88 |
-
"<unk_86>",
|
| 89 |
-
"<unk_87>",
|
| 90 |
-
"<unk_88>",
|
| 91 |
-
"<unk_89>",
|
| 92 |
-
"<unk_90>",
|
| 93 |
-
"<unk_91>",
|
| 94 |
-
"<unk_92>",
|
| 95 |
-
"<unk_93>",
|
| 96 |
-
"<unk_94>",
|
| 97 |
-
"<unk_95>",
|
| 98 |
-
"<unk_96>",
|
| 99 |
-
"<unk_97>",
|
| 100 |
-
"<unk_98>",
|
| 101 |
-
"<unk_99>",
|
| 102 |
-
"<unk_100>",
|
| 103 |
-
"<unk_101>",
|
| 104 |
-
"<unk_102>"
|
| 105 |
-
],
|
| 106 |
-
"eos_token": "</s>",
|
| 107 |
-
"mask_token": "<mask_2>",
|
| 108 |
-
"pad_token": "<pad>",
|
| 109 |
-
"unk_token": "<unk>"
|
| 110 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/spiece.model
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0015189ef36359283fec8b93cf6d9ce51bca37eb1101defc68a53b394913b96c
|
| 3 |
-
size 1912529
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
last-checkpoint/tokenizer_config.json
DELETED
|
@@ -1,118 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"additional_special_tokens": [
|
| 3 |
-
"<mask_1>",
|
| 4 |
-
"<unk_2>",
|
| 5 |
-
"<unk_3>",
|
| 6 |
-
"<unk_4>",
|
| 7 |
-
"<unk_5>",
|
| 8 |
-
"<unk_6>",
|
| 9 |
-
"<unk_7>",
|
| 10 |
-
"<unk_8>",
|
| 11 |
-
"<unk_9>",
|
| 12 |
-
"<unk_10>",
|
| 13 |
-
"<unk_11>",
|
| 14 |
-
"<unk_12>",
|
| 15 |
-
"<unk_13>",
|
| 16 |
-
"<unk_14>",
|
| 17 |
-
"<unk_15>",
|
| 18 |
-
"<unk_16>",
|
| 19 |
-
"<unk_17>",
|
| 20 |
-
"<unk_18>",
|
| 21 |
-
"<unk_19>",
|
| 22 |
-
"<unk_20>",
|
| 23 |
-
"<unk_21>",
|
| 24 |
-
"<unk_22>",
|
| 25 |
-
"<unk_23>",
|
| 26 |
-
"<unk_24>",
|
| 27 |
-
"<unk_25>",
|
| 28 |
-
"<unk_26>",
|
| 29 |
-
"<unk_27>",
|
| 30 |
-
"<unk_28>",
|
| 31 |
-
"<unk_29>",
|
| 32 |
-
"<unk_30>",
|
| 33 |
-
"<unk_31>",
|
| 34 |
-
"<unk_32>",
|
| 35 |
-
"<unk_33>",
|
| 36 |
-
"<unk_34>",
|
| 37 |
-
"<unk_35>",
|
| 38 |
-
"<unk_36>",
|
| 39 |
-
"<unk_37>",
|
| 40 |
-
"<unk_38>",
|
| 41 |
-
"<unk_39>",
|
| 42 |
-
"<unk_40>",
|
| 43 |
-
"<unk_41>",
|
| 44 |
-
"<unk_42>",
|
| 45 |
-
"<unk_43>",
|
| 46 |
-
"<unk_44>",
|
| 47 |
-
"<unk_45>",
|
| 48 |
-
"<unk_46>",
|
| 49 |
-
"<unk_47>",
|
| 50 |
-
"<unk_48>",
|
| 51 |
-
"<unk_49>",
|
| 52 |
-
"<unk_50>",
|
| 53 |
-
"<unk_51>",
|
| 54 |
-
"<unk_52>",
|
| 55 |
-
"<unk_53>",
|
| 56 |
-
"<unk_54>",
|
| 57 |
-
"<unk_55>",
|
| 58 |
-
"<unk_56>",
|
| 59 |
-
"<unk_57>",
|
| 60 |
-
"<unk_58>",
|
| 61 |
-
"<unk_59>",
|
| 62 |
-
"<unk_60>",
|
| 63 |
-
"<unk_61>",
|
| 64 |
-
"<unk_62>",
|
| 65 |
-
"<unk_63>",
|
| 66 |
-
"<unk_64>",
|
| 67 |
-
"<unk_65>",
|
| 68 |
-
"<unk_66>",
|
| 69 |
-
"<unk_67>",
|
| 70 |
-
"<unk_68>",
|
| 71 |
-
"<unk_69>",
|
| 72 |
-
"<unk_70>",
|
| 73 |
-
"<unk_71>",
|
| 74 |
-
"<unk_72>",
|
| 75 |
-
"<unk_73>",
|
| 76 |
-
"<unk_74>",
|
| 77 |
-
"<unk_75>",
|
| 78 |
-
"<unk_76>",
|
| 79 |
-
"<unk_77>",
|
| 80 |
-
"<unk_78>",
|
| 81 |
-
"<unk_79>",
|
| 82 |
-
"<unk_80>",
|
| 83 |
-
"<unk_81>",
|
| 84 |
-
"<unk_82>",
|
| 85 |
-
"<unk_83>",
|
| 86 |
-
"<unk_84>",
|
| 87 |
-
"<unk_85>",
|
| 88 |
-
"<unk_86>",
|
| 89 |
-
"<unk_87>",
|
| 90 |
-
"<unk_88>",
|
| 91 |
-
"<unk_89>",
|
| 92 |
-
"<unk_90>",
|
| 93 |
-
"<unk_91>",
|
| 94 |
-
"<unk_92>",
|
| 95 |
-
"<unk_93>",
|
| 96 |
-
"<unk_94>",
|
| 97 |
-
"<unk_95>",
|
| 98 |
-
"<unk_96>",
|
| 99 |
-
"<unk_97>",
|
| 100 |
-
"<unk_98>",
|
| 101 |
-
"<unk_99>",
|
| 102 |
-
"<unk_100>",
|
| 103 |
-
"<unk_101>",
|
| 104 |
-
"<unk_102>"
|
| 105 |
-
],
|
| 106 |
-
"eos_token": "</s>",
|
| 107 |
-
"full_tokenizer_file": null,
|
| 108 |
-
"mask_token": "<mask_2>",
|
| 109 |
-
"mask_token_sent": "<mask_1>",
|
| 110 |
-
"model_max_length": 512,
|
| 111 |
-
"name_or_path": "google/pegasus-newsroom",
|
| 112 |
-
"offset": 103,
|
| 113 |
-
"pad_token": "<pad>",
|
| 114 |
-
"sp_model_kwargs": {},
|
| 115 |
-
"special_tokens_map_file": null,
|
| 116 |
-
"tokenizer_class": "PegasusTokenizer",
|
| 117 |
-
"unk_token": "<unk>"
|
| 118 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/trainer_state.json
DELETED
|
@@ -1,35 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": 37.4459,
|
| 3 |
-
"best_model_checkpoint": "pegasus-newsroom-cnn_full-adafactor-bs6/checkpoint-299",
|
| 4 |
-
"epoch": 0.09997283346916598,
|
| 5 |
-
"global_step": 299,
|
| 6 |
-
"is_hyper_param_search": false,
|
| 7 |
-
"is_local_process_zero": true,
|
| 8 |
-
"is_world_process_zero": true,
|
| 9 |
-
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.1,
|
| 12 |
-
"learning_rate": 7.839464882943145e-05,
|
| 13 |
-
"loss": 3.3516,
|
| 14 |
-
"step": 299
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.1,
|
| 18 |
-
"eval_gen_len": 75.2876,
|
| 19 |
-
"eval_loss": 3.0071940422058105,
|
| 20 |
-
"eval_rouge1": 37.4459,
|
| 21 |
-
"eval_rouge2": 16.5527,
|
| 22 |
-
"eval_rougeL": 26.0178,
|
| 23 |
-
"eval_rougeLsum": 34.4955,
|
| 24 |
-
"eval_runtime": 1262.9542,
|
| 25 |
-
"eval_samples_per_second": 1.979,
|
| 26 |
-
"eval_steps_per_second": 0.33,
|
| 27 |
-
"step": 299
|
| 28 |
-
}
|
| 29 |
-
],
|
| 30 |
-
"max_steps": 2990,
|
| 31 |
-
"num_train_epochs": 1,
|
| 32 |
-
"total_flos": 4.146959327271322e+16,
|
| 33 |
-
"trial_name": null,
|
| 34 |
-
"trial_params": null
|
| 35 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:c380d343d655a247010cd81ba7f73d2c000362c3c07fbc4014043dafee87866f
|
| 3 |
-
size 3567
|
|
|
|
|
|
|
|
|
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2279605745
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e47e09c2798018a81069504ae2b7ea7a01b7ba5446e6e2793ed9e916e1772c03
|
| 3 |
size 2279605745
|
trainer_state.json
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 38.5992,
|
| 3 |
+
"best_model_checkpoint": "pegasus-newsroom-cnn_full-adafactor-bs6/checkpoint-897",
|
| 4 |
+
"epoch": 0.6356132321902493,
|
| 5 |
+
"global_step": 1901,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.1,
|
| 12 |
+
"learning_rate": 7.839464882943145e-05,
|
| 13 |
+
"loss": 3.3516,
|
| 14 |
+
"step": 299
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.1,
|
| 18 |
+
"eval_gen_len": 75.2876,
|
| 19 |
+
"eval_loss": 3.0071940422058105,
|
| 20 |
+
"eval_rouge1": 37.4459,
|
| 21 |
+
"eval_rouge2": 16.5527,
|
| 22 |
+
"eval_rougeL": 26.0178,
|
| 23 |
+
"eval_rougeLsum": 34.4955,
|
| 24 |
+
"eval_runtime": 1262.9542,
|
| 25 |
+
"eval_samples_per_second": 1.979,
|
| 26 |
+
"eval_steps_per_second": 0.33,
|
| 27 |
+
"step": 299
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.2,
|
| 31 |
+
"learning_rate": 0.00015812709030100337,
|
| 32 |
+
"loss": 3.1391,
|
| 33 |
+
"step": 598
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.2,
|
| 37 |
+
"eval_gen_len": 72.434,
|
| 38 |
+
"eval_loss": 2.974590539932251,
|
| 39 |
+
"eval_rouge1": 37.9369,
|
| 40 |
+
"eval_rouge2": 16.9058,
|
| 41 |
+
"eval_rougeL": 26.566,
|
| 42 |
+
"eval_rougeLsum": 35.0327,
|
| 43 |
+
"eval_runtime": 1247.518,
|
| 44 |
+
"eval_samples_per_second": 2.004,
|
| 45 |
+
"eval_steps_per_second": 0.334,
|
| 46 |
+
"step": 598
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.3,
|
| 50 |
+
"learning_rate": 0.0001405351170568562,
|
| 51 |
+
"loss": 3.1086,
|
| 52 |
+
"step": 897
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.3,
|
| 56 |
+
"eval_gen_len": 66.2268,
|
| 57 |
+
"eval_loss": 2.9494450092315674,
|
| 58 |
+
"eval_rouge1": 38.5992,
|
| 59 |
+
"eval_rouge2": 17.5271,
|
| 60 |
+
"eval_rougeL": 27.3099,
|
| 61 |
+
"eval_rougeLsum": 35.6531,
|
| 62 |
+
"eval_runtime": 1158.9422,
|
| 63 |
+
"eval_samples_per_second": 2.157,
|
| 64 |
+
"eval_steps_per_second": 0.36,
|
| 65 |
+
"step": 897
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.4,
|
| 69 |
+
"learning_rate": 0.00012060200668896322,
|
| 70 |
+
"loss": 3.1037,
|
| 71 |
+
"step": 1196
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.4,
|
| 75 |
+
"eval_gen_len": 70.6692,
|
| 76 |
+
"eval_loss": 2.95255184173584,
|
| 77 |
+
"eval_rouge1": 38.3805,
|
| 78 |
+
"eval_rouge2": 17.5267,
|
| 79 |
+
"eval_rougeL": 27.1742,
|
| 80 |
+
"eval_rougeLsum": 35.5035,
|
| 81 |
+
"eval_runtime": 1183.4051,
|
| 82 |
+
"eval_samples_per_second": 2.113,
|
| 83 |
+
"eval_steps_per_second": 0.352,
|
| 84 |
+
"step": 1196
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.5,
|
| 88 |
+
"learning_rate": 0.00010100334448160536,
|
| 89 |
+
"loss": 3.2294,
|
| 90 |
+
"step": 1495
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.5,
|
| 94 |
+
"eval_gen_len": 56.5128,
|
| 95 |
+
"eval_loss": 3.183689832687378,
|
| 96 |
+
"eval_rouge1": 36.2619,
|
| 97 |
+
"eval_rouge2": 15.599,
|
| 98 |
+
"eval_rougeL": 25.8333,
|
| 99 |
+
"eval_rougeLsum": 33.3692,
|
| 100 |
+
"eval_runtime": 997.3694,
|
| 101 |
+
"eval_samples_per_second": 2.507,
|
| 102 |
+
"eval_steps_per_second": 0.418,
|
| 103 |
+
"step": 1495
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.6,
|
| 107 |
+
"learning_rate": 8.127090301003345e-05,
|
| 108 |
+
"loss": 3.695,
|
| 109 |
+
"step": 1794
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 0.6,
|
| 113 |
+
"eval_gen_len": 47.1888,
|
| 114 |
+
"eval_loss": 3.4275190830230713,
|
| 115 |
+
"eval_rouge1": 33.0969,
|
| 116 |
+
"eval_rouge2": 13.6577,
|
| 117 |
+
"eval_rougeL": 23.9839,
|
| 118 |
+
"eval_rougeLsum": 30.6005,
|
| 119 |
+
"eval_runtime": 927.662,
|
| 120 |
+
"eval_samples_per_second": 2.695,
|
| 121 |
+
"eval_steps_per_second": 0.45,
|
| 122 |
+
"step": 1794
|
| 123 |
+
}
|
| 124 |
+
],
|
| 125 |
+
"max_steps": 2990,
|
| 126 |
+
"num_train_epochs": 1,
|
| 127 |
+
"total_flos": 2.4881742419263488e+17,
|
| 128 |
+
"trial_name": null,
|
| 129 |
+
"trial_params": null
|
| 130 |
+
}
|