Commit
·
190750f
1
Parent(s):
fbcb88a
Upload 11 files
Browse files- config.json +26 -0
- generation_config.json +7 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +11 -0
- trainer_state.json +581 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/scratch/project_2006732/experiments/finnish_gpt/checkpoints_14_20_laser_lora_with_pair_ids/checkpoint-3583",
|
| 3 |
+
"apply_residual_connection_post_layernorm": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BloomForCausalLM"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attention_softmax_in_fp32": true,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"hidden_dropout": 0.0,
|
| 12 |
+
"hidden_size": 2064,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"layer_norm_epsilon": 1e-05,
|
| 15 |
+
"masked_softmax_fusion": true,
|
| 16 |
+
"model_type": "bloom",
|
| 17 |
+
"n_head": 24,
|
| 18 |
+
"n_layer": 24,
|
| 19 |
+
"pad_token_id": 3,
|
| 20 |
+
"pretraining_tp": 1,
|
| 21 |
+
"slow_but_exact": false,
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.32.0",
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 131073
|
| 26 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 3,
|
| 6 |
+
"transformers_version": "4.32.0"
|
| 7 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9f0eabbb45cf76f13c336431d81f297dea094b61f32125903401072f079a435b
|
| 3 |
+
size 11985010085
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3295cd9cf8e8c59ca4081d626abb7c8bd5a4ad538c89511c156e4e2ff3693ac1
|
| 3 |
+
size 5992478601
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab62043be50b93d4eb28964be2d945176db3d64fe73ddd052a7656ba9141c683
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba3f93374978212864aec97131bafa2abbfd2d1f370540ee3e5f6d52725f03f4
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"pad_token": "<pad>",
|
| 5 |
+
"sep_token": "<|endofprompt|>",
|
| 6 |
+
"unk_token": "<unk>"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": "<s>",
|
| 4 |
+
"clean_up_tokenization_spaces": false,
|
| 5 |
+
"eos_token": "</s>",
|
| 6 |
+
"model_max_length": 2048,
|
| 7 |
+
"pad_token": "<pad>",
|
| 8 |
+
"padding_side": "left",
|
| 9 |
+
"tokenizer_class": "BloomTokenizer",
|
| 10 |
+
"unk_token": "<unk>"
|
| 11 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,581 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.999476713762428,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 8360,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08,
|
| 13 |
+
"learning_rate": 4.958123953098828e-05,
|
| 14 |
+
"loss": 0.1445,
|
| 15 |
+
"step": 100
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.17,
|
| 19 |
+
"learning_rate": 4.916247906197655e-05,
|
| 20 |
+
"loss": 0.1585,
|
| 21 |
+
"step": 200
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.25,
|
| 25 |
+
"learning_rate": 4.874371859296483e-05,
|
| 26 |
+
"loss": 0.1654,
|
| 27 |
+
"step": 300
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.33,
|
| 31 |
+
"learning_rate": 4.83249581239531e-05,
|
| 32 |
+
"loss": 0.1701,
|
| 33 |
+
"step": 400
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.42,
|
| 37 |
+
"learning_rate": 4.7906197654941376e-05,
|
| 38 |
+
"loss": 0.1884,
|
| 39 |
+
"step": 500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.5,
|
| 43 |
+
"learning_rate": 4.748743718592965e-05,
|
| 44 |
+
"loss": 0.1868,
|
| 45 |
+
"step": 600
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.59,
|
| 49 |
+
"learning_rate": 4.7068676716917926e-05,
|
| 50 |
+
"loss": 0.1872,
|
| 51 |
+
"step": 700
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.67,
|
| 55 |
+
"learning_rate": 4.66499162479062e-05,
|
| 56 |
+
"loss": 0.2018,
|
| 57 |
+
"step": 800
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.75,
|
| 61 |
+
"learning_rate": 4.6231155778894475e-05,
|
| 62 |
+
"loss": 0.1865,
|
| 63 |
+
"step": 900
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.84,
|
| 67 |
+
"learning_rate": 4.581239530988275e-05,
|
| 68 |
+
"loss": 0.2002,
|
| 69 |
+
"step": 1000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.84,
|
| 73 |
+
"eval_loss": 2.548002243041992,
|
| 74 |
+
"eval_runtime": 20.7773,
|
| 75 |
+
"eval_samples_per_second": 57.466,
|
| 76 |
+
"eval_steps_per_second": 14.391,
|
| 77 |
+
"step": 1000
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.92,
|
| 81 |
+
"learning_rate": 4.5393634840871025e-05,
|
| 82 |
+
"loss": 0.1776,
|
| 83 |
+
"step": 1100
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.0,
|
| 87 |
+
"learning_rate": 4.49748743718593e-05,
|
| 88 |
+
"loss": 0.2035,
|
| 89 |
+
"step": 1200
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.09,
|
| 93 |
+
"learning_rate": 4.4556113902847574e-05,
|
| 94 |
+
"loss": 0.1199,
|
| 95 |
+
"step": 1300
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.17,
|
| 99 |
+
"learning_rate": 4.413735343383585e-05,
|
| 100 |
+
"loss": 0.1311,
|
| 101 |
+
"step": 1400
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.26,
|
| 105 |
+
"learning_rate": 4.3718592964824124e-05,
|
| 106 |
+
"loss": 0.1279,
|
| 107 |
+
"step": 1500
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.34,
|
| 111 |
+
"learning_rate": 4.32998324958124e-05,
|
| 112 |
+
"loss": 0.1279,
|
| 113 |
+
"step": 1600
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.42,
|
| 117 |
+
"learning_rate": 4.288107202680067e-05,
|
| 118 |
+
"loss": 0.1249,
|
| 119 |
+
"step": 1700
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.51,
|
| 123 |
+
"learning_rate": 4.246231155778895e-05,
|
| 124 |
+
"loss": 0.1354,
|
| 125 |
+
"step": 1800
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.59,
|
| 129 |
+
"learning_rate": 4.204355108877722e-05,
|
| 130 |
+
"loss": 0.1407,
|
| 131 |
+
"step": 1900
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.67,
|
| 135 |
+
"learning_rate": 4.16247906197655e-05,
|
| 136 |
+
"loss": 0.1207,
|
| 137 |
+
"step": 2000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.67,
|
| 141 |
+
"eval_loss": 2.662675619125366,
|
| 142 |
+
"eval_runtime": 20.7988,
|
| 143 |
+
"eval_samples_per_second": 57.407,
|
| 144 |
+
"eval_steps_per_second": 14.376,
|
| 145 |
+
"step": 2000
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 1.76,
|
| 149 |
+
"learning_rate": 4.120603015075377e-05,
|
| 150 |
+
"loss": 0.1439,
|
| 151 |
+
"step": 2100
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 1.84,
|
| 155 |
+
"learning_rate": 4.078726968174205e-05,
|
| 156 |
+
"loss": 0.1342,
|
| 157 |
+
"step": 2200
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.93,
|
| 161 |
+
"learning_rate": 4.036850921273032e-05,
|
| 162 |
+
"loss": 0.1251,
|
| 163 |
+
"step": 2300
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 2.01,
|
| 167 |
+
"learning_rate": 3.9949748743718597e-05,
|
| 168 |
+
"loss": 0.1242,
|
| 169 |
+
"step": 2400
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 2.09,
|
| 173 |
+
"learning_rate": 3.953098827470687e-05,
|
| 174 |
+
"loss": 0.0826,
|
| 175 |
+
"step": 2500
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 2.18,
|
| 179 |
+
"learning_rate": 3.9112227805695146e-05,
|
| 180 |
+
"loss": 0.0793,
|
| 181 |
+
"step": 2600
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 2.26,
|
| 185 |
+
"learning_rate": 3.869346733668342e-05,
|
| 186 |
+
"loss": 0.0874,
|
| 187 |
+
"step": 2700
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 2.34,
|
| 191 |
+
"learning_rate": 3.8274706867671696e-05,
|
| 192 |
+
"loss": 0.0992,
|
| 193 |
+
"step": 2800
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 2.43,
|
| 197 |
+
"learning_rate": 3.785594639865997e-05,
|
| 198 |
+
"loss": 0.088,
|
| 199 |
+
"step": 2900
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 2.51,
|
| 203 |
+
"learning_rate": 3.7437185929648245e-05,
|
| 204 |
+
"loss": 0.0768,
|
| 205 |
+
"step": 3000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 2.51,
|
| 209 |
+
"eval_loss": 2.742750406265259,
|
| 210 |
+
"eval_runtime": 20.7709,
|
| 211 |
+
"eval_samples_per_second": 57.484,
|
| 212 |
+
"eval_steps_per_second": 14.395,
|
| 213 |
+
"step": 3000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.6,
|
| 217 |
+
"learning_rate": 3.701842546063652e-05,
|
| 218 |
+
"loss": 0.0805,
|
| 219 |
+
"step": 3100
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 2.68,
|
| 223 |
+
"learning_rate": 3.6599664991624795e-05,
|
| 224 |
+
"loss": 0.0796,
|
| 225 |
+
"step": 3200
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 2.76,
|
| 229 |
+
"learning_rate": 3.618090452261307e-05,
|
| 230 |
+
"loss": 0.0819,
|
| 231 |
+
"step": 3300
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 2.85,
|
| 235 |
+
"learning_rate": 3.5762144053601344e-05,
|
| 236 |
+
"loss": 0.0844,
|
| 237 |
+
"step": 3400
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.93,
|
| 241 |
+
"learning_rate": 3.534338358458962e-05,
|
| 242 |
+
"loss": 0.081,
|
| 243 |
+
"step": 3500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 3.01,
|
| 247 |
+
"learning_rate": 3.4924623115577894e-05,
|
| 248 |
+
"loss": 0.0796,
|
| 249 |
+
"step": 3600
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 3.1,
|
| 253 |
+
"learning_rate": 3.450586264656617e-05,
|
| 254 |
+
"loss": 0.0595,
|
| 255 |
+
"step": 3700
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 3.18,
|
| 259 |
+
"learning_rate": 3.408710217755444e-05,
|
| 260 |
+
"loss": 0.0481,
|
| 261 |
+
"step": 3800
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 3.27,
|
| 265 |
+
"learning_rate": 3.366834170854272e-05,
|
| 266 |
+
"loss": 0.0554,
|
| 267 |
+
"step": 3900
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 3.35,
|
| 271 |
+
"learning_rate": 3.324958123953099e-05,
|
| 272 |
+
"loss": 0.0566,
|
| 273 |
+
"step": 4000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 3.35,
|
| 277 |
+
"eval_loss": 2.823474407196045,
|
| 278 |
+
"eval_runtime": 20.7588,
|
| 279 |
+
"eval_samples_per_second": 57.518,
|
| 280 |
+
"eval_steps_per_second": 14.404,
|
| 281 |
+
"step": 4000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.43,
|
| 285 |
+
"learning_rate": 3.283082077051927e-05,
|
| 286 |
+
"loss": 0.0562,
|
| 287 |
+
"step": 4100
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 3.52,
|
| 291 |
+
"learning_rate": 3.241206030150754e-05,
|
| 292 |
+
"loss": 0.0584,
|
| 293 |
+
"step": 4200
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 3.6,
|
| 297 |
+
"learning_rate": 3.199329983249582e-05,
|
| 298 |
+
"loss": 0.0555,
|
| 299 |
+
"step": 4300
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 3.68,
|
| 303 |
+
"learning_rate": 3.157453936348409e-05,
|
| 304 |
+
"loss": 0.0446,
|
| 305 |
+
"step": 4400
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.77,
|
| 309 |
+
"learning_rate": 3.1155778894472366e-05,
|
| 310 |
+
"loss": 0.0559,
|
| 311 |
+
"step": 4500
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.85,
|
| 315 |
+
"learning_rate": 3.073701842546064e-05,
|
| 316 |
+
"loss": 0.0582,
|
| 317 |
+
"step": 4600
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 3.94,
|
| 321 |
+
"learning_rate": 3.0318257956448916e-05,
|
| 322 |
+
"loss": 0.0457,
|
| 323 |
+
"step": 4700
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 4.02,
|
| 327 |
+
"learning_rate": 2.989949748743719e-05,
|
| 328 |
+
"loss": 0.0486,
|
| 329 |
+
"step": 4800
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 4.1,
|
| 333 |
+
"learning_rate": 2.9480737018425465e-05,
|
| 334 |
+
"loss": 0.0292,
|
| 335 |
+
"step": 4900
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 4.19,
|
| 339 |
+
"learning_rate": 2.906197654941374e-05,
|
| 340 |
+
"loss": 0.0335,
|
| 341 |
+
"step": 5000
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 4.19,
|
| 345 |
+
"eval_loss": 2.9011123180389404,
|
| 346 |
+
"eval_runtime": 20.7527,
|
| 347 |
+
"eval_samples_per_second": 57.535,
|
| 348 |
+
"eval_steps_per_second": 14.408,
|
| 349 |
+
"step": 5000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 4.27,
|
| 353 |
+
"learning_rate": 2.8643216080402015e-05,
|
| 354 |
+
"loss": 0.028,
|
| 355 |
+
"step": 5100
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 4.35,
|
| 359 |
+
"learning_rate": 2.822445561139029e-05,
|
| 360 |
+
"loss": 0.0317,
|
| 361 |
+
"step": 5200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 4.44,
|
| 365 |
+
"learning_rate": 2.7805695142378564e-05,
|
| 366 |
+
"loss": 0.0392,
|
| 367 |
+
"step": 5300
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 4.52,
|
| 371 |
+
"learning_rate": 2.738693467336684e-05,
|
| 372 |
+
"loss": 0.035,
|
| 373 |
+
"step": 5400
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 4.6,
|
| 377 |
+
"learning_rate": 2.696817420435511e-05,
|
| 378 |
+
"loss": 0.0333,
|
| 379 |
+
"step": 5500
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 4.69,
|
| 383 |
+
"learning_rate": 2.6549413735343385e-05,
|
| 384 |
+
"loss": 0.0306,
|
| 385 |
+
"step": 5600
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 4.77,
|
| 389 |
+
"learning_rate": 2.613065326633166e-05,
|
| 390 |
+
"loss": 0.0427,
|
| 391 |
+
"step": 5700
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 4.86,
|
| 395 |
+
"learning_rate": 2.5711892797319935e-05,
|
| 396 |
+
"loss": 0.0249,
|
| 397 |
+
"step": 5800
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 4.94,
|
| 401 |
+
"learning_rate": 2.529313232830821e-05,
|
| 402 |
+
"loss": 0.0391,
|
| 403 |
+
"step": 5900
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 5.02,
|
| 407 |
+
"learning_rate": 2.4874371859296484e-05,
|
| 408 |
+
"loss": 0.0281,
|
| 409 |
+
"step": 6000
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 5.02,
|
| 413 |
+
"eval_loss": 2.8658268451690674,
|
| 414 |
+
"eval_runtime": 20.7258,
|
| 415 |
+
"eval_samples_per_second": 57.609,
|
| 416 |
+
"eval_steps_per_second": 14.426,
|
| 417 |
+
"step": 6000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 5.11,
|
| 421 |
+
"learning_rate": 2.445561139028476e-05,
|
| 422 |
+
"loss": 0.0213,
|
| 423 |
+
"step": 6100
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 5.19,
|
| 427 |
+
"learning_rate": 2.4036850921273034e-05,
|
| 428 |
+
"loss": 0.0163,
|
| 429 |
+
"step": 6200
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 5.27,
|
| 433 |
+
"learning_rate": 2.361809045226131e-05,
|
| 434 |
+
"loss": 0.0166,
|
| 435 |
+
"step": 6300
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 5.36,
|
| 439 |
+
"learning_rate": 2.3199329983249583e-05,
|
| 440 |
+
"loss": 0.0161,
|
| 441 |
+
"step": 6400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 5.44,
|
| 445 |
+
"learning_rate": 2.2780569514237858e-05,
|
| 446 |
+
"loss": 0.0241,
|
| 447 |
+
"step": 6500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 5.53,
|
| 451 |
+
"learning_rate": 2.2361809045226133e-05,
|
| 452 |
+
"loss": 0.0237,
|
| 453 |
+
"step": 6600
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 5.61,
|
| 457 |
+
"learning_rate": 2.1943048576214408e-05,
|
| 458 |
+
"loss": 0.0143,
|
| 459 |
+
"step": 6700
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 5.69,
|
| 463 |
+
"learning_rate": 2.1524288107202682e-05,
|
| 464 |
+
"loss": 0.017,
|
| 465 |
+
"step": 6800
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 5.78,
|
| 469 |
+
"learning_rate": 2.1105527638190957e-05,
|
| 470 |
+
"loss": 0.0198,
|
| 471 |
+
"step": 6900
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 5.86,
|
| 475 |
+
"learning_rate": 2.0686767169179232e-05,
|
| 476 |
+
"loss": 0.0215,
|
| 477 |
+
"step": 7000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 5.86,
|
| 481 |
+
"eval_loss": 2.8833489418029785,
|
| 482 |
+
"eval_runtime": 20.7599,
|
| 483 |
+
"eval_samples_per_second": 57.515,
|
| 484 |
+
"eval_steps_per_second": 14.403,
|
| 485 |
+
"step": 7000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 5.94,
|
| 489 |
+
"learning_rate": 2.0268006700167507e-05,
|
| 490 |
+
"loss": 0.0197,
|
| 491 |
+
"step": 7100
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 6.03,
|
| 495 |
+
"learning_rate": 1.984924623115578e-05,
|
| 496 |
+
"loss": 0.0173,
|
| 497 |
+
"step": 7200
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 6.11,
|
| 501 |
+
"learning_rate": 1.9430485762144056e-05,
|
| 502 |
+
"loss": 0.015,
|
| 503 |
+
"step": 7300
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 6.2,
|
| 507 |
+
"learning_rate": 1.901172529313233e-05,
|
| 508 |
+
"loss": 0.0139,
|
| 509 |
+
"step": 7400
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 6.28,
|
| 513 |
+
"learning_rate": 1.8592964824120602e-05,
|
| 514 |
+
"loss": 0.0116,
|
| 515 |
+
"step": 7500
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 6.36,
|
| 519 |
+
"learning_rate": 1.8174204355108877e-05,
|
| 520 |
+
"loss": 0.0048,
|
| 521 |
+
"step": 7600
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 6.45,
|
| 525 |
+
"learning_rate": 1.7755443886097152e-05,
|
| 526 |
+
"loss": 0.0117,
|
| 527 |
+
"step": 7700
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 6.53,
|
| 531 |
+
"learning_rate": 1.7336683417085427e-05,
|
| 532 |
+
"loss": 0.0079,
|
| 533 |
+
"step": 7800
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 6.61,
|
| 537 |
+
"learning_rate": 1.69179229480737e-05,
|
| 538 |
+
"loss": 0.0095,
|
| 539 |
+
"step": 7900
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 6.7,
|
| 543 |
+
"learning_rate": 1.6499162479061976e-05,
|
| 544 |
+
"loss": 0.0081,
|
| 545 |
+
"step": 8000
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 6.7,
|
| 549 |
+
"eval_loss": 2.974073648452759,
|
| 550 |
+
"eval_runtime": 20.7629,
|
| 551 |
+
"eval_samples_per_second": 57.506,
|
| 552 |
+
"eval_steps_per_second": 14.401,
|
| 553 |
+
"step": 8000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 6.78,
|
| 557 |
+
"learning_rate": 1.608040201005025e-05,
|
| 558 |
+
"loss": 0.0086,
|
| 559 |
+
"step": 8100
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 6.87,
|
| 563 |
+
"learning_rate": 1.5661641541038526e-05,
|
| 564 |
+
"loss": 0.013,
|
| 565 |
+
"step": 8200
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 6.95,
|
| 569 |
+
"learning_rate": 1.5242881072026802e-05,
|
| 570 |
+
"loss": 0.0079,
|
| 571 |
+
"step": 8300
|
| 572 |
+
}
|
| 573 |
+
],
|
| 574 |
+
"logging_steps": 100,
|
| 575 |
+
"max_steps": 11940,
|
| 576 |
+
"num_train_epochs": 10,
|
| 577 |
+
"save_steps": 500,
|
| 578 |
+
"total_flos": 1.919599695983693e+16,
|
| 579 |
+
"trial_name": null,
|
| 580 |
+
"trial_params": null
|
| 581 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:569fbfa18f334d79226a7c02e9da8363e52fe446ab5ece8d60c83e431686bb90
|
| 3 |
+
size 4027
|