Commit
•
3433b13
1
Parent(s):
46a473f
commit files to HF hub
Browse files- config.json +31 -0
- hypar_params.json +1 -0
- pytorch_model.bin +3 -0
- tables_result.json +1 -0
- trainer_state.json +547 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "logoyazilim/qna_model_0000",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MT5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 1024,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 512,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "gelu_new",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "gated-gelu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": true,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"model_type": "mt5",
|
| 19 |
+
"num_decoder_layers": 8,
|
| 20 |
+
"num_heads": 6,
|
| 21 |
+
"num_layers": 8,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"relative_attention_max_distance": 128,
|
| 24 |
+
"relative_attention_num_buckets": 32,
|
| 25 |
+
"tie_word_embeddings": false,
|
| 26 |
+
"tokenizer_class": "T5Tokenizer",
|
| 27 |
+
"torch_dtype": "float32",
|
| 28 |
+
"transformers_version": "4.21.3",
|
| 29 |
+
"use_cache": true,
|
| 30 |
+
"vocab_size": 250102
|
| 31 |
+
}
|
hypar_params.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"per_device_train_batch_size": 8, "per_device_eval_batch_size": 8, "gradient_accumulation_steps": 4, "learning_rate": 0.001, "num_train_epochs": 12, "max_steps": -1}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:19809715df5da1948f83d816ab197ba96bc8441516b8ddbc0cf57b6d2a986b09
|
| 3 |
+
size 1200729797
|
tables_result.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
[{"product": "ERP Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22111413041615_54de84fb-b6e2-4ec0-930e-33b60527ebfb", "create_date": "2022-11-14 13:04:16", "product_name": "ERP Bilgi Deposu"}]}, {"product": "Logo B2B Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_23011311034785_c02ab1c7-9fc4-4f68-b46b-53ea1f13c742", "create_date": "2023-01-13 11:03:47", "product_name": "Logo B2B Bilgi Deposu"}]}, {"product": "Logo Bulut ERP Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22121911591425_ffebd58d-92a7-4b83-91b3-36aeb50e0519", "create_date": "2022-12-19 11:59:14", "product_name": "Logo Bulut ERP Bilgi Deposu"}]}, {"product": "Logo GO 3 Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22111813163591_46252143-0501-43e8-b8fb-5f0cb9a8cd55", "create_date": "2022-11-18 13:16:35", "product_name": "Logo GO 3 Bilgi Deposu"}]}, {"product": "Logo Mali M\u00fc\u015favir 3 Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22112203170690_ecdd2c8f-7aff-4053-b1be-b2f75bd763b6", "create_date": "2022-11-22 03:17:06", "product_name": "Logo Mali M\u00fc\u015favir 3 Bilgi Deposu"}]}, {"product": "Logo Netsis 3 Entegre Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_23010310223644_3f18efca-032f-4193-8ea9-0f01c34ea4ee", "create_date": "2023-01-03 10:22:36", "product_name": "Logo Netsis 3 Entegre Bilgi Deposu"}]}, {"product": "Logo Netsis 3 Enterprise Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_23011008063723_b8edf845-d6cf-42b9-b4c9-676d473e9643", "create_date": "2023-01-10 08:06:37", "product_name": "Logo Netsis 3 Enterprise Bilgi Deposu"}]}, {"product": "Logo Netsis 3 Standard Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_23010323403959_80bc1e19-7d56-4acb-9d1b-b3db05f8906a", "create_date": "2023-01-03 23:40:39", "product_name": "Logo Netsis 3 Standard Bilgi Deposu"}]}, {"product": "Logo Start 3 Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22112118442495_5916dc6c-1d7e-4544-8e45-0ffbed21658e", "create_date": "2022-11-21 18:44:24", "product_name": "Logo Start 3 Bilgi Deposu"}]}, {"product": "Logo Tiger 3 Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22112118563920_4e5fe4c2-0cb2-43ae-8448-eeeb8ef36e95", "create_date": "2022-11-21 18:56:39", "product_name": "Logo Tiger 3 Bilgi Deposu"}]}, {"product": "Logo Tiger 3 Enterprise Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22112113350211_f2c8fb0d-9f2d-442a-8f92-3d80bef3f61d", "create_date": "2022-11-21 13:35:02", "product_name": "Logo Tiger 3 Enterprise Bilgi Deposu"}]}, {"product": "Logo j-Platform ERP Bilgi Deposu", "azure_tables": [{"table_name": "azuredata_22120508591353_54211327-3f15-4ae5-a98b-32aaecf073b6", "create_date": "2022-12-05 08:59:13", "product_name": "Logo j-Platform ERP Bilgi Deposu"}]}]
|
trainer_state.json
ADDED
|
@@ -0,0 +1,547 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 11.999931153184166,
|
| 5 |
+
"global_step": 43572,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.14,
|
| 12 |
+
"learning_rate": 0.0009885247406591388,
|
| 13 |
+
"loss": 0.4139,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.28,
|
| 18 |
+
"learning_rate": 0.0009770494813182778,
|
| 19 |
+
"loss": 0.315,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.41,
|
| 24 |
+
"learning_rate": 0.0009655742219774168,
|
| 25 |
+
"loss": 0.2807,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.55,
|
| 30 |
+
"learning_rate": 0.0009540989626365555,
|
| 31 |
+
"loss": 0.2624,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.69,
|
| 36 |
+
"learning_rate": 0.0009426237032956944,
|
| 37 |
+
"loss": 0.257,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.83,
|
| 42 |
+
"learning_rate": 0.0009311484439548334,
|
| 43 |
+
"loss": 0.2378,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.96,
|
| 48 |
+
"learning_rate": 0.0009196731846139723,
|
| 49 |
+
"loss": 0.2245,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.1,
|
| 54 |
+
"learning_rate": 0.0009081979252731112,
|
| 55 |
+
"loss": 0.1983,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.24,
|
| 60 |
+
"learning_rate": 0.0008967226659322501,
|
| 61 |
+
"loss": 0.1987,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.38,
|
| 66 |
+
"learning_rate": 0.0008852474065913891,
|
| 67 |
+
"loss": 0.1904,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.51,
|
| 72 |
+
"learning_rate": 0.0008737721472505279,
|
| 73 |
+
"loss": 0.1829,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.65,
|
| 78 |
+
"learning_rate": 0.0008622968879096667,
|
| 79 |
+
"loss": 0.1773,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.79,
|
| 84 |
+
"learning_rate": 0.0008508216285688056,
|
| 85 |
+
"loss": 0.179,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.93,
|
| 90 |
+
"learning_rate": 0.0008393463692279446,
|
| 91 |
+
"loss": 0.177,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 2.07,
|
| 96 |
+
"learning_rate": 0.0008278711098870835,
|
| 97 |
+
"loss": 0.1577,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 2.2,
|
| 102 |
+
"learning_rate": 0.0008163958505462224,
|
| 103 |
+
"loss": 0.1529,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.34,
|
| 108 |
+
"learning_rate": 0.0008049205912053612,
|
| 109 |
+
"loss": 0.1468,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 2.48,
|
| 114 |
+
"learning_rate": 0.0007934453318645002,
|
| 115 |
+
"loss": 0.1492,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.62,
|
| 120 |
+
"learning_rate": 0.000781970072523639,
|
| 121 |
+
"loss": 0.1426,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 2.75,
|
| 126 |
+
"learning_rate": 0.0007704948131827779,
|
| 127 |
+
"loss": 0.1368,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.89,
|
| 132 |
+
"learning_rate": 0.0007590195538419168,
|
| 133 |
+
"loss": 0.1443,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.03,
|
| 138 |
+
"learning_rate": 0.0007475442945010558,
|
| 139 |
+
"loss": 0.1364,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.17,
|
| 144 |
+
"learning_rate": 0.0007360690351601946,
|
| 145 |
+
"loss": 0.1221,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.3,
|
| 150 |
+
"learning_rate": 0.0007245937758193335,
|
| 151 |
+
"loss": 0.1188,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.44,
|
| 156 |
+
"learning_rate": 0.0007131185164784725,
|
| 157 |
+
"loss": 0.1218,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 3.58,
|
| 162 |
+
"learning_rate": 0.0007016432571376114,
|
| 163 |
+
"loss": 0.1175,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 3.72,
|
| 168 |
+
"learning_rate": 0.0006901679977967502,
|
| 169 |
+
"loss": 0.117,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.86,
|
| 174 |
+
"learning_rate": 0.000678692738455889,
|
| 175 |
+
"loss": 0.1154,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.99,
|
| 180 |
+
"learning_rate": 0.000667217479115028,
|
| 181 |
+
"loss": 0.1124,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 4.13,
|
| 186 |
+
"learning_rate": 0.0006557422197741669,
|
| 187 |
+
"loss": 0.0995,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 4.27,
|
| 192 |
+
"learning_rate": 0.0006442669604333058,
|
| 193 |
+
"loss": 0.0948,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 4.41,
|
| 198 |
+
"learning_rate": 0.0006327917010924447,
|
| 199 |
+
"loss": 0.0972,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 4.54,
|
| 204 |
+
"learning_rate": 0.0006213164417515837,
|
| 205 |
+
"loss": 0.0967,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 4.68,
|
| 210 |
+
"learning_rate": 0.0006098411824107226,
|
| 211 |
+
"loss": 0.1003,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 4.82,
|
| 216 |
+
"learning_rate": 0.0005983659230698613,
|
| 217 |
+
"loss": 0.0961,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 4.96,
|
| 222 |
+
"learning_rate": 0.0005868906637290002,
|
| 223 |
+
"loss": 0.0967,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 5.1,
|
| 228 |
+
"learning_rate": 0.0005754154043881392,
|
| 229 |
+
"loss": 0.0825,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 5.23,
|
| 234 |
+
"learning_rate": 0.0005639401450472781,
|
| 235 |
+
"loss": 0.081,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 5.37,
|
| 240 |
+
"learning_rate": 0.000552464885706417,
|
| 241 |
+
"loss": 0.0827,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 5.51,
|
| 246 |
+
"learning_rate": 0.0005409896263655559,
|
| 247 |
+
"loss": 0.08,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 5.65,
|
| 252 |
+
"learning_rate": 0.0005295143670246948,
|
| 253 |
+
"loss": 0.0842,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 5.78,
|
| 258 |
+
"learning_rate": 0.0005180391076838336,
|
| 259 |
+
"loss": 0.0798,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 5.92,
|
| 264 |
+
"learning_rate": 0.0005065638483429725,
|
| 265 |
+
"loss": 0.083,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 6.06,
|
| 270 |
+
"learning_rate": 0.0004950885890021115,
|
| 271 |
+
"loss": 0.0749,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.2,
|
| 276 |
+
"learning_rate": 0.00048361332966125037,
|
| 277 |
+
"loss": 0.0667,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 6.33,
|
| 282 |
+
"learning_rate": 0.00047213807032038925,
|
| 283 |
+
"loss": 0.0677,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 6.47,
|
| 288 |
+
"learning_rate": 0.0004606628109795282,
|
| 289 |
+
"loss": 0.0699,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 6.61,
|
| 294 |
+
"learning_rate": 0.000449187551638667,
|
| 295 |
+
"loss": 0.0694,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 6.75,
|
| 300 |
+
"learning_rate": 0.00043771229229780595,
|
| 301 |
+
"loss": 0.0674,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 6.89,
|
| 306 |
+
"learning_rate": 0.00042623703295694484,
|
| 307 |
+
"loss": 0.0652,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 7.02,
|
| 312 |
+
"learning_rate": 0.00041476177361608377,
|
| 313 |
+
"loss": 0.0668,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 7.16,
|
| 318 |
+
"learning_rate": 0.0004032865142752226,
|
| 319 |
+
"loss": 0.0573,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 7.3,
|
| 324 |
+
"learning_rate": 0.00039181125493436154,
|
| 325 |
+
"loss": 0.0564,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 7.44,
|
| 330 |
+
"learning_rate": 0.0003803359955935004,
|
| 331 |
+
"loss": 0.0573,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 7.57,
|
| 336 |
+
"learning_rate": 0.0003688607362526393,
|
| 337 |
+
"loss": 0.0572,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 7.71,
|
| 342 |
+
"learning_rate": 0.0003573854769117782,
|
| 343 |
+
"loss": 0.058,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 7.85,
|
| 348 |
+
"learning_rate": 0.0003459102175709171,
|
| 349 |
+
"loss": 0.0583,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 7.99,
|
| 354 |
+
"learning_rate": 0.000334434958230056,
|
| 355 |
+
"loss": 0.0581,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 8.12,
|
| 360 |
+
"learning_rate": 0.0003229596988891949,
|
| 361 |
+
"loss": 0.0483,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 8.26,
|
| 366 |
+
"learning_rate": 0.00031148443954833377,
|
| 367 |
+
"loss": 0.0474,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 8.4,
|
| 372 |
+
"learning_rate": 0.0003000091802074727,
|
| 373 |
+
"loss": 0.0497,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 8.54,
|
| 378 |
+
"learning_rate": 0.0002885339208666116,
|
| 379 |
+
"loss": 0.0487,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 8.68,
|
| 384 |
+
"learning_rate": 0.00027705866152575047,
|
| 385 |
+
"loss": 0.0469,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 8.81,
|
| 390 |
+
"learning_rate": 0.0002655834021848894,
|
| 391 |
+
"loss": 0.0492,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 8.95,
|
| 396 |
+
"learning_rate": 0.0002541081428440283,
|
| 397 |
+
"loss": 0.0494,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 9.09,
|
| 402 |
+
"learning_rate": 0.00024263288350316717,
|
| 403 |
+
"loss": 0.0426,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 9.23,
|
| 408 |
+
"learning_rate": 0.00023115762416230608,
|
| 409 |
+
"loss": 0.0406,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 9.36,
|
| 414 |
+
"learning_rate": 0.00021968236482144496,
|
| 415 |
+
"loss": 0.0405,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 9.5,
|
| 420 |
+
"learning_rate": 0.00020820710548058387,
|
| 421 |
+
"loss": 0.043,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 9.64,
|
| 426 |
+
"learning_rate": 0.00019673184613972275,
|
| 427 |
+
"loss": 0.0411,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 9.78,
|
| 432 |
+
"learning_rate": 0.00018525658679886166,
|
| 433 |
+
"loss": 0.0419,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 9.91,
|
| 438 |
+
"learning_rate": 0.00017378132745800054,
|
| 439 |
+
"loss": 0.0409,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 10.05,
|
| 444 |
+
"learning_rate": 0.00016230606811713943,
|
| 445 |
+
"loss": 0.0388,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 10.19,
|
| 450 |
+
"learning_rate": 0.00015083080877627833,
|
| 451 |
+
"loss": 0.0367,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 10.33,
|
| 456 |
+
"learning_rate": 0.00013935554943541724,
|
| 457 |
+
"loss": 0.0355,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 10.47,
|
| 462 |
+
"learning_rate": 0.00012788029009455615,
|
| 463 |
+
"loss": 0.0358,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 10.6,
|
| 468 |
+
"learning_rate": 0.00011640503075369504,
|
| 469 |
+
"loss": 0.0369,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 10.74,
|
| 474 |
+
"learning_rate": 0.00010492977141283393,
|
| 475 |
+
"loss": 0.0339,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 10.88,
|
| 480 |
+
"learning_rate": 9.345451207197283e-05,
|
| 481 |
+
"loss": 0.0359,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 11.02,
|
| 486 |
+
"learning_rate": 8.197925273111174e-05,
|
| 487 |
+
"loss": 0.0364,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 11.15,
|
| 492 |
+
"learning_rate": 7.050399339025063e-05,
|
| 493 |
+
"loss": 0.0325,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 11.29,
|
| 498 |
+
"learning_rate": 5.902873404938952e-05,
|
| 499 |
+
"loss": 0.0321,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 11.43,
|
| 504 |
+
"learning_rate": 4.755347470852842e-05,
|
| 505 |
+
"loss": 0.0328,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 11.57,
|
| 510 |
+
"learning_rate": 3.607821536766731e-05,
|
| 511 |
+
"loss": 0.032,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 11.7,
|
| 516 |
+
"learning_rate": 2.4602956026806205e-05,
|
| 517 |
+
"loss": 0.0315,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 11.84,
|
| 522 |
+
"learning_rate": 1.3127696685945103e-05,
|
| 523 |
+
"loss": 0.0318,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 11.98,
|
| 528 |
+
"learning_rate": 1.6524373450839989e-06,
|
| 529 |
+
"loss": 0.0314,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 12.0,
|
| 534 |
+
"step": 43572,
|
| 535 |
+
"total_flos": 2.93279124897792e+17,
|
| 536 |
+
"train_loss": 0.09914802481164445,
|
| 537 |
+
"train_runtime": 67574.8841,
|
| 538 |
+
"train_samples_per_second": 20.634,
|
| 539 |
+
"train_steps_per_second": 0.645
|
| 540 |
+
}
|
| 541 |
+
],
|
| 542 |
+
"max_steps": 43572,
|
| 543 |
+
"num_train_epochs": 12,
|
| 544 |
+
"total_flos": 2.93279124897792e+17,
|
| 545 |
+
"trial_name": null,
|
| 546 |
+
"trial_params": null
|
| 547 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd887a2baa50684b9babe8870d3618ad7b3c8dc8e21e5ddd1985ec2da6474247
|
| 3 |
+
size 3375
|