Upload 7 files
Browse files- config.json +32 -0
- generation_config.json +6 -0
- model.safetensors +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +593 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"GPT2LMHeadModel"
|
| 5 |
+
],
|
| 6 |
+
"attn_pdrop": 0.1,
|
| 7 |
+
"bos_token_id": 50256,
|
| 8 |
+
"embd_pdrop": 0.1,
|
| 9 |
+
"eos_token_id": 50256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"layer_norm_epsilon": 1e-05,
|
| 12 |
+
"model_type": "gpt2",
|
| 13 |
+
"n_ctx": 1024,
|
| 14 |
+
"n_embd": 768,
|
| 15 |
+
"n_head": 12,
|
| 16 |
+
"n_inner": null,
|
| 17 |
+
"n_layer": 12,
|
| 18 |
+
"n_positions": 1024,
|
| 19 |
+
"reorder_and_upcast_attn": false,
|
| 20 |
+
"resid_pdrop": 0.1,
|
| 21 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 22 |
+
"scale_attn_weights": true,
|
| 23 |
+
"summary_activation": null,
|
| 24 |
+
"summary_first_dropout": 0.1,
|
| 25 |
+
"summary_proj_to_labels": true,
|
| 26 |
+
"summary_type": "cls_index",
|
| 27 |
+
"summary_use_proj": true,
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"transformers_version": "4.41.2",
|
| 30 |
+
"use_cache": true,
|
| 31 |
+
"vocab_size": 50257
|
| 32 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.41.2"
|
| 6 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ffbe1e5f7c4bfdd013b48a953eb7d8edf52a536f138063873050a1225316f02e
|
| 3 |
+
size 497774208
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:293b3127c16463858e9e7a7882af4b45935b928667a1b6381a5cf9da4c6b7967
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7365b2086a9f78dd3818c48aa4873d430f4d355b3837086c61aff774ccae2904
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,593 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 8.669737198591168,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 40000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.1083717149823896,
|
| 13 |
+
"grad_norm": 2.5749928951263428,
|
| 14 |
+
"learning_rate": 4.945805332755257e-05,
|
| 15 |
+
"loss": 5.0542,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.2167434299647792,
|
| 20 |
+
"grad_norm": 2.815129041671753,
|
| 21 |
+
"learning_rate": 4.8916106655105135e-05,
|
| 22 |
+
"loss": 3.9388,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.3251151449471688,
|
| 27 |
+
"grad_norm": 2.5150463581085205,
|
| 28 |
+
"learning_rate": 4.8374159982657704e-05,
|
| 29 |
+
"loss": 3.6244,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.4334868599295584,
|
| 34 |
+
"grad_norm": 2.6200830936431885,
|
| 35 |
+
"learning_rate": 4.783221331021028e-05,
|
| 36 |
+
"loss": 3.3961,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.541858574911948,
|
| 41 |
+
"grad_norm": 2.5062639713287354,
|
| 42 |
+
"learning_rate": 4.729026663776285e-05,
|
| 43 |
+
"loss": 3.2233,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.6502302898943376,
|
| 48 |
+
"grad_norm": 2.6815900802612305,
|
| 49 |
+
"learning_rate": 4.674831996531541e-05,
|
| 50 |
+
"loss": 3.1132,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.7586020048767271,
|
| 55 |
+
"grad_norm": 2.634747266769409,
|
| 56 |
+
"learning_rate": 4.620637329286798e-05,
|
| 57 |
+
"loss": 3.0163,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.8669737198591168,
|
| 62 |
+
"grad_norm": 2.711473226547241,
|
| 63 |
+
"learning_rate": 4.566442662042055e-05,
|
| 64 |
+
"loss": 2.9397,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.9753454348415064,
|
| 69 |
+
"grad_norm": 2.559657573699951,
|
| 70 |
+
"learning_rate": 4.512247994797312e-05,
|
| 71 |
+
"loss": 2.8676,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.083717149823896,
|
| 76 |
+
"grad_norm": 2.810490131378174,
|
| 77 |
+
"learning_rate": 4.458053327552569e-05,
|
| 78 |
+
"loss": 2.775,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 1.1920888648062855,
|
| 83 |
+
"grad_norm": 2.7192561626434326,
|
| 84 |
+
"learning_rate": 4.403858660307826e-05,
|
| 85 |
+
"loss": 2.7126,
|
| 86 |
+
"step": 5500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.300460579788675,
|
| 90 |
+
"grad_norm": 2.6966707706451416,
|
| 91 |
+
"learning_rate": 4.349663993063083e-05,
|
| 92 |
+
"loss": 2.6669,
|
| 93 |
+
"step": 6000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.4088322947710648,
|
| 97 |
+
"grad_norm": 2.680452585220337,
|
| 98 |
+
"learning_rate": 4.29546932581834e-05,
|
| 99 |
+
"loss": 2.6252,
|
| 100 |
+
"step": 6500
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 1.5172040097534545,
|
| 104 |
+
"grad_norm": 2.666625499725342,
|
| 105 |
+
"learning_rate": 4.241274658573596e-05,
|
| 106 |
+
"loss": 2.5945,
|
| 107 |
+
"step": 7000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.625575724735844,
|
| 111 |
+
"grad_norm": 2.610170602798462,
|
| 112 |
+
"learning_rate": 4.187079991328853e-05,
|
| 113 |
+
"loss": 2.5711,
|
| 114 |
+
"step": 7500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.7339474397182335,
|
| 118 |
+
"grad_norm": 2.701840400695801,
|
| 119 |
+
"learning_rate": 4.13288532408411e-05,
|
| 120 |
+
"loss": 2.5359,
|
| 121 |
+
"step": 8000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.8423191547006232,
|
| 125 |
+
"grad_norm": 2.5341765880584717,
|
| 126 |
+
"learning_rate": 4.078690656839367e-05,
|
| 127 |
+
"loss": 2.5104,
|
| 128 |
+
"step": 8500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.9506908696830128,
|
| 132 |
+
"grad_norm": 2.574463129043579,
|
| 133 |
+
"learning_rate": 4.024495989594624e-05,
|
| 134 |
+
"loss": 2.4829,
|
| 135 |
+
"step": 9000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 2.0590625846654023,
|
| 139 |
+
"grad_norm": 2.8886678218841553,
|
| 140 |
+
"learning_rate": 3.970301322349881e-05,
|
| 141 |
+
"loss": 2.4105,
|
| 142 |
+
"step": 9500
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.167434299647792,
|
| 146 |
+
"grad_norm": 2.600092649459839,
|
| 147 |
+
"learning_rate": 3.916106655105138e-05,
|
| 148 |
+
"loss": 2.3608,
|
| 149 |
+
"step": 10000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 2.2758060146301817,
|
| 153 |
+
"grad_norm": 2.76435923576355,
|
| 154 |
+
"learning_rate": 3.8620203771948845e-05,
|
| 155 |
+
"loss": 2.3489,
|
| 156 |
+
"step": 10500
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 2.384177729612571,
|
| 160 |
+
"grad_norm": 2.622713804244995,
|
| 161 |
+
"learning_rate": 3.8078257099501414e-05,
|
| 162 |
+
"loss": 2.3168,
|
| 163 |
+
"step": 11000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 2.4925494445949608,
|
| 167 |
+
"grad_norm": 2.8371784687042236,
|
| 168 |
+
"learning_rate": 3.753631042705398e-05,
|
| 169 |
+
"loss": 2.3082,
|
| 170 |
+
"step": 11500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.60092115957735,
|
| 174 |
+
"grad_norm": 2.9003915786743164,
|
| 175 |
+
"learning_rate": 3.6994363754606547e-05,
|
| 176 |
+
"loss": 2.2956,
|
| 177 |
+
"step": 12000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 2.7092928745597398,
|
| 181 |
+
"grad_norm": 2.8348236083984375,
|
| 182 |
+
"learning_rate": 3.645350097550401e-05,
|
| 183 |
+
"loss": 2.2873,
|
| 184 |
+
"step": 12500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 2.8176645895421295,
|
| 188 |
+
"grad_norm": 2.8634517192840576,
|
| 189 |
+
"learning_rate": 3.591155430305658e-05,
|
| 190 |
+
"loss": 2.2614,
|
| 191 |
+
"step": 13000
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.9260363045245192,
|
| 195 |
+
"grad_norm": 2.732450008392334,
|
| 196 |
+
"learning_rate": 3.5369607630609145e-05,
|
| 197 |
+
"loss": 2.2654,
|
| 198 |
+
"step": 13500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 3.0344080195069085,
|
| 202 |
+
"grad_norm": 2.839444160461426,
|
| 203 |
+
"learning_rate": 3.4827660958161714e-05,
|
| 204 |
+
"loss": 2.2182,
|
| 205 |
+
"step": 14000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 3.1427797344892983,
|
| 209 |
+
"grad_norm": 2.9356603622436523,
|
| 210 |
+
"learning_rate": 3.428679817905918e-05,
|
| 211 |
+
"loss": 2.1356,
|
| 212 |
+
"step": 14500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.251151449471688,
|
| 216 |
+
"grad_norm": 3.012699604034424,
|
| 217 |
+
"learning_rate": 3.374485150661175e-05,
|
| 218 |
+
"loss": 2.1355,
|
| 219 |
+
"step": 15000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 3.3595231644540773,
|
| 223 |
+
"grad_norm": 3.038557291030884,
|
| 224 |
+
"learning_rate": 3.320290483416432e-05,
|
| 225 |
+
"loss": 2.126,
|
| 226 |
+
"step": 15500
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 3.467894879436467,
|
| 230 |
+
"grad_norm": 2.943721294403076,
|
| 231 |
+
"learning_rate": 3.266095816171689e-05,
|
| 232 |
+
"loss": 2.1243,
|
| 233 |
+
"step": 16000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 3.5762665944188567,
|
| 237 |
+
"grad_norm": 2.797640323638916,
|
| 238 |
+
"learning_rate": 3.2120095382614354e-05,
|
| 239 |
+
"loss": 2.1118,
|
| 240 |
+
"step": 16500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 3.6846383094012465,
|
| 244 |
+
"grad_norm": 3.0536017417907715,
|
| 245 |
+
"learning_rate": 3.1578148710166924e-05,
|
| 246 |
+
"loss": 2.1125,
|
| 247 |
+
"step": 17000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 3.7930100243836358,
|
| 251 |
+
"grad_norm": 3.118360757827759,
|
| 252 |
+
"learning_rate": 3.1036202037719493e-05,
|
| 253 |
+
"loss": 2.096,
|
| 254 |
+
"step": 17500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 3.9013817393660255,
|
| 258 |
+
"grad_norm": 3.150421142578125,
|
| 259 |
+
"learning_rate": 3.0494255365272056e-05,
|
| 260 |
+
"loss": 2.0936,
|
| 261 |
+
"step": 18000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 4.009753454348415,
|
| 265 |
+
"grad_norm": 3.1124956607818604,
|
| 266 |
+
"learning_rate": 2.9953392586169522e-05,
|
| 267 |
+
"loss": 2.0847,
|
| 268 |
+
"step": 18500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 4.1181251693308045,
|
| 272 |
+
"grad_norm": 3.0479283332824707,
|
| 273 |
+
"learning_rate": 2.941144591372209e-05,
|
| 274 |
+
"loss": 1.9658,
|
| 275 |
+
"step": 19000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 4.226496884313194,
|
| 279 |
+
"grad_norm": 3.2436435222625732,
|
| 280 |
+
"learning_rate": 2.886949924127466e-05,
|
| 281 |
+
"loss": 1.9759,
|
| 282 |
+
"step": 19500
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 4.334868599295584,
|
| 286 |
+
"grad_norm": 3.0921454429626465,
|
| 287 |
+
"learning_rate": 2.832755256882723e-05,
|
| 288 |
+
"loss": 1.9828,
|
| 289 |
+
"step": 20000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 4.443240314277974,
|
| 293 |
+
"grad_norm": 3.161250114440918,
|
| 294 |
+
"learning_rate": 2.778668978972469e-05,
|
| 295 |
+
"loss": 1.9731,
|
| 296 |
+
"step": 20500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 4.5516120292603635,
|
| 300 |
+
"grad_norm": 3.133525848388672,
|
| 301 |
+
"learning_rate": 2.724474311727726e-05,
|
| 302 |
+
"loss": 1.9683,
|
| 303 |
+
"step": 21000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 4.659983744242752,
|
| 307 |
+
"grad_norm": 3.2014646530151367,
|
| 308 |
+
"learning_rate": 2.6702796444829832e-05,
|
| 309 |
+
"loss": 1.9605,
|
| 310 |
+
"step": 21500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 4.768355459225142,
|
| 314 |
+
"grad_norm": 3.3526577949523926,
|
| 315 |
+
"learning_rate": 2.61608497723824e-05,
|
| 316 |
+
"loss": 1.9687,
|
| 317 |
+
"step": 22000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 4.876727174207532,
|
| 321 |
+
"grad_norm": 3.1296660900115967,
|
| 322 |
+
"learning_rate": 2.561998699327986e-05,
|
| 323 |
+
"loss": 1.9577,
|
| 324 |
+
"step": 22500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 4.9850988891899215,
|
| 328 |
+
"grad_norm": 3.223031759262085,
|
| 329 |
+
"learning_rate": 2.507804032083243e-05,
|
| 330 |
+
"loss": 1.9648,
|
| 331 |
+
"step": 23000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.093470604172311,
|
| 335 |
+
"grad_norm": 3.6728270053863525,
|
| 336 |
+
"learning_rate": 2.4536093648385e-05,
|
| 337 |
+
"loss": 1.8425,
|
| 338 |
+
"step": 23500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.201842319154701,
|
| 342 |
+
"grad_norm": 3.5075490474700928,
|
| 343 |
+
"learning_rate": 2.399414697593757e-05,
|
| 344 |
+
"loss": 1.8299,
|
| 345 |
+
"step": 24000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.31021403413709,
|
| 349 |
+
"grad_norm": 3.391207218170166,
|
| 350 |
+
"learning_rate": 2.345328419683503e-05,
|
| 351 |
+
"loss": 1.8471,
|
| 352 |
+
"step": 24500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.4185857491194795,
|
| 356 |
+
"grad_norm": 3.570876121520996,
|
| 357 |
+
"learning_rate": 2.29113375243876e-05,
|
| 358 |
+
"loss": 1.8448,
|
| 359 |
+
"step": 25000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.526957464101869,
|
| 363 |
+
"grad_norm": 3.741013765335083,
|
| 364 |
+
"learning_rate": 2.236939085194017e-05,
|
| 365 |
+
"loss": 1.8487,
|
| 366 |
+
"step": 25500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.635329179084259,
|
| 370 |
+
"grad_norm": 3.223215103149414,
|
| 371 |
+
"learning_rate": 2.182744417949274e-05,
|
| 372 |
+
"loss": 1.8485,
|
| 373 |
+
"step": 26000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.743700894066649,
|
| 377 |
+
"grad_norm": 3.5355515480041504,
|
| 378 |
+
"learning_rate": 2.1286581400390203e-05,
|
| 379 |
+
"loss": 1.8532,
|
| 380 |
+
"step": 26500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.8520726090490385,
|
| 384 |
+
"grad_norm": 3.312040090560913,
|
| 385 |
+
"learning_rate": 2.0744634727942772e-05,
|
| 386 |
+
"loss": 1.8539,
|
| 387 |
+
"step": 27000
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.960444324031428,
|
| 391 |
+
"grad_norm": 3.7042343616485596,
|
| 392 |
+
"learning_rate": 2.0202688055495338e-05,
|
| 393 |
+
"loss": 1.8512,
|
| 394 |
+
"step": 27500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.068816039013817,
|
| 398 |
+
"grad_norm": 4.022594451904297,
|
| 399 |
+
"learning_rate": 1.9660741383047908e-05,
|
| 400 |
+
"loss": 1.7677,
|
| 401 |
+
"step": 28000
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 6.177187753996207,
|
| 405 |
+
"grad_norm": 3.602557420730591,
|
| 406 |
+
"learning_rate": 1.9119878603945374e-05,
|
| 407 |
+
"loss": 1.7204,
|
| 408 |
+
"step": 28500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 6.2855594689785965,
|
| 412 |
+
"grad_norm": 3.635352849960327,
|
| 413 |
+
"learning_rate": 1.857793193149794e-05,
|
| 414 |
+
"loss": 1.7337,
|
| 415 |
+
"step": 29000
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 6.393931183960986,
|
| 419 |
+
"grad_norm": 4.083003997802734,
|
| 420 |
+
"learning_rate": 1.8035985259050513e-05,
|
| 421 |
+
"loss": 1.7345,
|
| 422 |
+
"step": 29500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 6.502302898943376,
|
| 426 |
+
"grad_norm": 3.698439121246338,
|
| 427 |
+
"learning_rate": 1.749403858660308e-05,
|
| 428 |
+
"loss": 1.7393,
|
| 429 |
+
"step": 30000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 6.610674613925766,
|
| 433 |
+
"grad_norm": 3.725682020187378,
|
| 434 |
+
"learning_rate": 1.6953175807500545e-05,
|
| 435 |
+
"loss": 1.7423,
|
| 436 |
+
"step": 30500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 6.719046328908155,
|
| 440 |
+
"grad_norm": 3.9255781173706055,
|
| 441 |
+
"learning_rate": 1.641122913505311e-05,
|
| 442 |
+
"loss": 1.7431,
|
| 443 |
+
"step": 31000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 6.827418043890544,
|
| 447 |
+
"grad_norm": 3.7728075981140137,
|
| 448 |
+
"learning_rate": 1.586928246260568e-05,
|
| 449 |
+
"loss": 1.7261,
|
| 450 |
+
"step": 31500
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 6.935789758872934,
|
| 454 |
+
"grad_norm": 3.897160291671753,
|
| 455 |
+
"learning_rate": 1.532733579015825e-05,
|
| 456 |
+
"loss": 1.7314,
|
| 457 |
+
"step": 32000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 7.044161473855324,
|
| 461 |
+
"grad_norm": 3.9494259357452393,
|
| 462 |
+
"learning_rate": 1.4786473011055712e-05,
|
| 463 |
+
"loss": 1.6944,
|
| 464 |
+
"step": 32500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 7.1525331888377135,
|
| 468 |
+
"grad_norm": 4.158576011657715,
|
| 469 |
+
"learning_rate": 1.4244526338608283e-05,
|
| 470 |
+
"loss": 1.6283,
|
| 471 |
+
"step": 33000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 7.260904903820103,
|
| 475 |
+
"grad_norm": 4.097021579742432,
|
| 476 |
+
"learning_rate": 1.370257966616085e-05,
|
| 477 |
+
"loss": 1.6305,
|
| 478 |
+
"step": 33500
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 7.369276618802493,
|
| 482 |
+
"grad_norm": 4.209109783172607,
|
| 483 |
+
"learning_rate": 1.316063299371342e-05,
|
| 484 |
+
"loss": 1.6381,
|
| 485 |
+
"step": 34000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 7.477648333784882,
|
| 489 |
+
"grad_norm": 4.275895595550537,
|
| 490 |
+
"learning_rate": 1.2619770214610881e-05,
|
| 491 |
+
"loss": 1.6326,
|
| 492 |
+
"step": 34500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 7.5860200487672715,
|
| 496 |
+
"grad_norm": 4.008493900299072,
|
| 497 |
+
"learning_rate": 1.2077823542163453e-05,
|
| 498 |
+
"loss": 1.6457,
|
| 499 |
+
"step": 35000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 7.694391763749661,
|
| 503 |
+
"grad_norm": 4.225920677185059,
|
| 504 |
+
"learning_rate": 1.153587686971602e-05,
|
| 505 |
+
"loss": 1.6413,
|
| 506 |
+
"step": 35500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 7.802763478732051,
|
| 510 |
+
"grad_norm": 4.181864261627197,
|
| 511 |
+
"learning_rate": 1.099393019726859e-05,
|
| 512 |
+
"loss": 1.6261,
|
| 513 |
+
"step": 36000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 7.911135193714441,
|
| 517 |
+
"grad_norm": 4.228626728057861,
|
| 518 |
+
"learning_rate": 1.0453067418166052e-05,
|
| 519 |
+
"loss": 1.6303,
|
| 520 |
+
"step": 36500
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 8.01950690869683,
|
| 524 |
+
"grad_norm": 4.4212565422058105,
|
| 525 |
+
"learning_rate": 9.911120745718622e-06,
|
| 526 |
+
"loss": 1.6119,
|
| 527 |
+
"step": 37000
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 8.12787862367922,
|
| 531 |
+
"grad_norm": 4.202996253967285,
|
| 532 |
+
"learning_rate": 9.36917407327119e-06,
|
| 533 |
+
"loss": 1.5373,
|
| 534 |
+
"step": 37500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 8.236250338661609,
|
| 538 |
+
"grad_norm": 4.508554935455322,
|
| 539 |
+
"learning_rate": 8.82722740082376e-06,
|
| 540 |
+
"loss": 1.5539,
|
| 541 |
+
"step": 38000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 8.344622053643999,
|
| 545 |
+
"grad_norm": 4.638177394866943,
|
| 546 |
+
"learning_rate": 8.286364621721222e-06,
|
| 547 |
+
"loss": 1.5502,
|
| 548 |
+
"step": 38500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 8.452993768626389,
|
| 552 |
+
"grad_norm": 4.576778888702393,
|
| 553 |
+
"learning_rate": 7.744417949273791e-06,
|
| 554 |
+
"loss": 1.5476,
|
| 555 |
+
"step": 39000
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 8.561365483608778,
|
| 559 |
+
"grad_norm": 4.546296119689941,
|
| 560 |
+
"learning_rate": 7.20247127682636e-06,
|
| 561 |
+
"loss": 1.5507,
|
| 562 |
+
"step": 39500
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 8.669737198591168,
|
| 566 |
+
"grad_norm": 4.418442249298096,
|
| 567 |
+
"learning_rate": 6.6605246043789286e-06,
|
| 568 |
+
"loss": 1.5507,
|
| 569 |
+
"step": 40000
|
| 570 |
+
}
|
| 571 |
+
],
|
| 572 |
+
"logging_steps": 500,
|
| 573 |
+
"max_steps": 46130,
|
| 574 |
+
"num_input_tokens_seen": 0,
|
| 575 |
+
"num_train_epochs": 10,
|
| 576 |
+
"save_steps": 10000,
|
| 577 |
+
"stateful_callbacks": {
|
| 578 |
+
"TrainerControl": {
|
| 579 |
+
"args": {
|
| 580 |
+
"should_epoch_stop": false,
|
| 581 |
+
"should_evaluate": false,
|
| 582 |
+
"should_log": false,
|
| 583 |
+
"should_save": true,
|
| 584 |
+
"should_training_stop": false
|
| 585 |
+
},
|
| 586 |
+
"attributes": {}
|
| 587 |
+
}
|
| 588 |
+
},
|
| 589 |
+
"total_flos": 2.090336256e+16,
|
| 590 |
+
"train_batch_size": 1,
|
| 591 |
+
"trial_name": null,
|
| 592 |
+
"trial_params": null
|
| 593 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6be10e8caa88bc51988400a3cc72ce524e437edc1ec1df3007ef523c3e6234e9
|
| 3 |
+
size 5112
|