ijampei
commited on
Commit
·
9db60b7
1
Parent(s):
1110422
initial commit
Browse files- .DS_Store +0 -0
- config.json +29 -0
- date_encoder.joblib +3 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- ref_encoder.joblib +3 -0
- rng_state.pth +3 -0
- rp_encoder.joblib +3 -0
- scheduler.pt +3 -0
- trainer_state.json +385 -0
- training_args.bin +3 -0
- virus_encoder.joblib +3 -0
- vp_encoder.joblib +3 -0
.DS_Store
ADDED
|
Binary file (6.15 kB). View file
|
|
|
config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"semanticESM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.0,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"emb_layer_norm_before": false,
|
| 8 |
+
"esmfold_config": null,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.0,
|
| 11 |
+
"hidden_size": 1280,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 5120,
|
| 14 |
+
"is_folding_model": false,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"mask_token_id": 32,
|
| 17 |
+
"max_position_embeddings": 1026,
|
| 18 |
+
"model_type": "esm",
|
| 19 |
+
"num_attention_heads": 20,
|
| 20 |
+
"num_hidden_layers": 33,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "rotary",
|
| 23 |
+
"token_dropout": true,
|
| 24 |
+
"torch_dtype": "float32",
|
| 25 |
+
"transformers_version": "4.48.3",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_list": null,
|
| 28 |
+
"vocab_size": 33
|
| 29 |
+
}
|
date_encoder.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4ab55e2afeee9c6e0a0f8b9965332b058cd3e943bcdeef41bfc64f5f87af4a5
|
| 3 |
+
size 8954
|
model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6af75ec31b23c428bc95025cad47a66705fa2f1715355d2b63ab7c62deb396dc
|
| 3 |
+
size 4995966392
|
model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddc678d8964fe1947759a7feeeb510542758105cb110d13c67c59269921ca0b8
|
| 3 |
+
size 211288044
|
model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ref_encoder.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e10e5bcf54a75ed520a30b7ad3c57abbfb1717b6f83d2a9a07aa7db8737de2e4
|
| 3 |
+
size 2233
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:114f9e3aca83c7a1f3453e9bae3798d43d06ed368679b8df86fd0c404a936fe1
|
| 3 |
+
size 14244
|
rp_encoder.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3058d8589781bb74c1cb50c82ebc536256ad5a6f09ec8d5852610e6873fd44b
|
| 3 |
+
size 625
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9943a980306bd39500335948abe4e3e8251c9eec8cf40d865652b9879f0e1d33
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,385 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.034771159291267395,
|
| 3 |
+
"best_model_checkpoint": "/gs/fs/tgh-25IAK/ue02946/PLANT/Season_based_split_performance/final_full_250304_ver2_12__s20000_lr1.00E-04_wd0.01_rwd0.005_csew0_csea0_sw0.2_csewv0_swv0.2_cw0.05_rid256_dp0.05_eid64_edp0.1_lgw0.01_bs16_facebookesm2_t33_650M_UR50D/trained_until_full/results/checkpoint-16000",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 16000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.25,
|
| 13 |
+
"grad_norm": 1.430809736251831,
|
| 14 |
+
"learning_rate": 1.2475e-05,
|
| 15 |
+
"loss": 0.1385,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.5,
|
| 20 |
+
"grad_norm": 0.6614887118339539,
|
| 21 |
+
"learning_rate": 2.4975e-05,
|
| 22 |
+
"loss": 0.0693,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.5,
|
| 27 |
+
"eval_loss": 0.05656217411160469,
|
| 28 |
+
"eval_runtime": 175.6555,
|
| 29 |
+
"eval_samples_per_second": 40.585,
|
| 30 |
+
"eval_steps_per_second": 1.27,
|
| 31 |
+
"step": 1000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.75,
|
| 35 |
+
"grad_norm": 0.6037147641181946,
|
| 36 |
+
"learning_rate": 3.7475e-05,
|
| 37 |
+
"loss": 0.0612,
|
| 38 |
+
"step": 1500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 1.0,
|
| 42 |
+
"grad_norm": 0.4972088038921356,
|
| 43 |
+
"learning_rate": 4.9975e-05,
|
| 44 |
+
"loss": 0.0577,
|
| 45 |
+
"step": 2000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 1.0,
|
| 49 |
+
"eval_loss": 0.06209351494908333,
|
| 50 |
+
"eval_runtime": 175.8577,
|
| 51 |
+
"eval_samples_per_second": 40.538,
|
| 52 |
+
"eval_steps_per_second": 1.268,
|
| 53 |
+
"step": 2000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.25,
|
| 57 |
+
"grad_norm": 1.0353070497512817,
|
| 58 |
+
"learning_rate": 4.861388888888889e-05,
|
| 59 |
+
"loss": 0.0511,
|
| 60 |
+
"step": 2500
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.5,
|
| 64 |
+
"grad_norm": 0.8515058755874634,
|
| 65 |
+
"learning_rate": 4.7225e-05,
|
| 66 |
+
"loss": 0.0447,
|
| 67 |
+
"step": 3000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 1.5,
|
| 71 |
+
"eval_loss": 0.050771959125995636,
|
| 72 |
+
"eval_runtime": 175.8405,
|
| 73 |
+
"eval_samples_per_second": 40.542,
|
| 74 |
+
"eval_steps_per_second": 1.268,
|
| 75 |
+
"step": 3000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 1.75,
|
| 79 |
+
"grad_norm": 0.5539710521697998,
|
| 80 |
+
"learning_rate": 4.5836111111111115e-05,
|
| 81 |
+
"loss": 0.0441,
|
| 82 |
+
"step": 3500
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 2.0,
|
| 86 |
+
"grad_norm": 0.7269946336746216,
|
| 87 |
+
"learning_rate": 4.444722222222222e-05,
|
| 88 |
+
"loss": 0.0409,
|
| 89 |
+
"step": 4000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 2.0,
|
| 93 |
+
"eval_loss": 0.0459725446999073,
|
| 94 |
+
"eval_runtime": 175.8069,
|
| 95 |
+
"eval_samples_per_second": 40.55,
|
| 96 |
+
"eval_steps_per_second": 1.268,
|
| 97 |
+
"step": 4000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 2.25,
|
| 101 |
+
"grad_norm": 0.25292640924453735,
|
| 102 |
+
"learning_rate": 4.3058333333333334e-05,
|
| 103 |
+
"loss": 0.0385,
|
| 104 |
+
"step": 4500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.5,
|
| 108 |
+
"grad_norm": 0.2268402874469757,
|
| 109 |
+
"learning_rate": 4.166944444444445e-05,
|
| 110 |
+
"loss": 0.04,
|
| 111 |
+
"step": 5000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 2.5,
|
| 115 |
+
"eval_loss": 0.04381516948342323,
|
| 116 |
+
"eval_runtime": 175.8092,
|
| 117 |
+
"eval_samples_per_second": 40.55,
|
| 118 |
+
"eval_steps_per_second": 1.268,
|
| 119 |
+
"step": 5000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.75,
|
| 123 |
+
"grad_norm": 0.2693457305431366,
|
| 124 |
+
"learning_rate": 4.0283333333333334e-05,
|
| 125 |
+
"loss": 0.0372,
|
| 126 |
+
"step": 5500
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 3.0,
|
| 130 |
+
"grad_norm": 0.383963406085968,
|
| 131 |
+
"learning_rate": 3.889444444444445e-05,
|
| 132 |
+
"loss": 0.0359,
|
| 133 |
+
"step": 6000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 3.0,
|
| 137 |
+
"eval_loss": 0.04170890897512436,
|
| 138 |
+
"eval_runtime": 175.8179,
|
| 139 |
+
"eval_samples_per_second": 40.548,
|
| 140 |
+
"eval_steps_per_second": 1.268,
|
| 141 |
+
"step": 6000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.25,
|
| 145 |
+
"grad_norm": 0.341470867395401,
|
| 146 |
+
"learning_rate": 3.750555555555555e-05,
|
| 147 |
+
"loss": 0.0366,
|
| 148 |
+
"step": 6500
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 3.5,
|
| 152 |
+
"grad_norm": 0.3173449635505676,
|
| 153 |
+
"learning_rate": 3.611666666666667e-05,
|
| 154 |
+
"loss": 0.0342,
|
| 155 |
+
"step": 7000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 3.5,
|
| 159 |
+
"eval_loss": 0.04057968407869339,
|
| 160 |
+
"eval_runtime": 175.921,
|
| 161 |
+
"eval_samples_per_second": 40.524,
|
| 162 |
+
"eval_steps_per_second": 1.268,
|
| 163 |
+
"step": 7000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 3.75,
|
| 167 |
+
"grad_norm": 0.17745588719844818,
|
| 168 |
+
"learning_rate": 3.472777777777778e-05,
|
| 169 |
+
"loss": 0.0334,
|
| 170 |
+
"step": 7500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 4.0,
|
| 174 |
+
"grad_norm": 0.737916886806488,
|
| 175 |
+
"learning_rate": 3.333888888888889e-05,
|
| 176 |
+
"loss": 0.0338,
|
| 177 |
+
"step": 8000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 4.0,
|
| 181 |
+
"eval_loss": 0.03727322071790695,
|
| 182 |
+
"eval_runtime": 175.854,
|
| 183 |
+
"eval_samples_per_second": 40.539,
|
| 184 |
+
"eval_steps_per_second": 1.268,
|
| 185 |
+
"step": 8000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.25,
|
| 189 |
+
"grad_norm": 0.4044702351093292,
|
| 190 |
+
"learning_rate": 3.1950000000000004e-05,
|
| 191 |
+
"loss": 0.0315,
|
| 192 |
+
"step": 8500
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 4.5,
|
| 196 |
+
"grad_norm": 0.2551765441894531,
|
| 197 |
+
"learning_rate": 3.056388888888889e-05,
|
| 198 |
+
"loss": 0.0325,
|
| 199 |
+
"step": 9000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 4.5,
|
| 203 |
+
"eval_loss": 0.03956814482808113,
|
| 204 |
+
"eval_runtime": 175.8578,
|
| 205 |
+
"eval_samples_per_second": 40.538,
|
| 206 |
+
"eval_steps_per_second": 1.268,
|
| 207 |
+
"step": 9000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 4.75,
|
| 211 |
+
"grad_norm": 0.4832671284675598,
|
| 212 |
+
"learning_rate": 2.9175e-05,
|
| 213 |
+
"loss": 0.0336,
|
| 214 |
+
"step": 9500
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 5.0,
|
| 218 |
+
"grad_norm": 0.219984769821167,
|
| 219 |
+
"learning_rate": 2.7786111111111114e-05,
|
| 220 |
+
"loss": 0.0311,
|
| 221 |
+
"step": 10000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 5.0,
|
| 225 |
+
"eval_loss": 0.037063486874103546,
|
| 226 |
+
"eval_runtime": 175.9259,
|
| 227 |
+
"eval_samples_per_second": 40.523,
|
| 228 |
+
"eval_steps_per_second": 1.268,
|
| 229 |
+
"step": 10000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 5.25,
|
| 233 |
+
"grad_norm": 0.1559722125530243,
|
| 234 |
+
"learning_rate": 2.6397222222222223e-05,
|
| 235 |
+
"loss": 0.0303,
|
| 236 |
+
"step": 10500
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 5.5,
|
| 240 |
+
"grad_norm": 0.1364031732082367,
|
| 241 |
+
"learning_rate": 2.5008333333333332e-05,
|
| 242 |
+
"loss": 0.0299,
|
| 243 |
+
"step": 11000
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 5.5,
|
| 247 |
+
"eval_loss": 0.037668175995349884,
|
| 248 |
+
"eval_runtime": 176.0991,
|
| 249 |
+
"eval_samples_per_second": 40.483,
|
| 250 |
+
"eval_steps_per_second": 1.266,
|
| 251 |
+
"step": 11000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 5.75,
|
| 255 |
+
"grad_norm": 0.20991720259189606,
|
| 256 |
+
"learning_rate": 2.3619444444444445e-05,
|
| 257 |
+
"loss": 0.0308,
|
| 258 |
+
"step": 11500
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 6.0,
|
| 262 |
+
"grad_norm": 0.29334667325019836,
|
| 263 |
+
"learning_rate": 2.2230555555555558e-05,
|
| 264 |
+
"loss": 0.0297,
|
| 265 |
+
"step": 12000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 6.0,
|
| 269 |
+
"eval_loss": 0.03977835178375244,
|
| 270 |
+
"eval_runtime": 175.799,
|
| 271 |
+
"eval_samples_per_second": 40.552,
|
| 272 |
+
"eval_steps_per_second": 1.268,
|
| 273 |
+
"step": 12000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 6.25,
|
| 277 |
+
"grad_norm": 0.2716386616230011,
|
| 278 |
+
"learning_rate": 2.0841666666666667e-05,
|
| 279 |
+
"loss": 0.0297,
|
| 280 |
+
"step": 12500
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 6.5,
|
| 284 |
+
"grad_norm": 0.30008479952812195,
|
| 285 |
+
"learning_rate": 1.9455555555555558e-05,
|
| 286 |
+
"loss": 0.0284,
|
| 287 |
+
"step": 13000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 6.5,
|
| 291 |
+
"eval_loss": 0.036556895822286606,
|
| 292 |
+
"eval_runtime": 175.7757,
|
| 293 |
+
"eval_samples_per_second": 40.557,
|
| 294 |
+
"eval_steps_per_second": 1.269,
|
| 295 |
+
"step": 13000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 6.75,
|
| 299 |
+
"grad_norm": 0.30487436056137085,
|
| 300 |
+
"learning_rate": 1.8066666666666668e-05,
|
| 301 |
+
"loss": 0.0279,
|
| 302 |
+
"step": 13500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 7.0,
|
| 306 |
+
"grad_norm": 0.2959464490413666,
|
| 307 |
+
"learning_rate": 1.6677777777777777e-05,
|
| 308 |
+
"loss": 0.0288,
|
| 309 |
+
"step": 14000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 7.0,
|
| 313 |
+
"eval_loss": 0.03538262099027634,
|
| 314 |
+
"eval_runtime": 175.7677,
|
| 315 |
+
"eval_samples_per_second": 40.559,
|
| 316 |
+
"eval_steps_per_second": 1.269,
|
| 317 |
+
"step": 14000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 7.25,
|
| 321 |
+
"grad_norm": 0.16555708646774292,
|
| 322 |
+
"learning_rate": 1.528888888888889e-05,
|
| 323 |
+
"loss": 0.0275,
|
| 324 |
+
"step": 14500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 7.5,
|
| 328 |
+
"grad_norm": 0.1510380208492279,
|
| 329 |
+
"learning_rate": 1.3900000000000002e-05,
|
| 330 |
+
"loss": 0.0276,
|
| 331 |
+
"step": 15000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 7.5,
|
| 335 |
+
"eval_loss": 0.03512444719672203,
|
| 336 |
+
"eval_runtime": 175.7946,
|
| 337 |
+
"eval_samples_per_second": 40.553,
|
| 338 |
+
"eval_steps_per_second": 1.269,
|
| 339 |
+
"step": 15000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 7.75,
|
| 343 |
+
"grad_norm": 0.7063078880310059,
|
| 344 |
+
"learning_rate": 1.2513888888888888e-05,
|
| 345 |
+
"loss": 0.0281,
|
| 346 |
+
"step": 15500
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 8.0,
|
| 350 |
+
"grad_norm": 0.14875133335590363,
|
| 351 |
+
"learning_rate": 1.1125000000000001e-05,
|
| 352 |
+
"loss": 0.0269,
|
| 353 |
+
"step": 16000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 8.0,
|
| 357 |
+
"eval_loss": 0.034771159291267395,
|
| 358 |
+
"eval_runtime": 175.7842,
|
| 359 |
+
"eval_samples_per_second": 40.555,
|
| 360 |
+
"eval_steps_per_second": 1.269,
|
| 361 |
+
"step": 16000
|
| 362 |
+
}
|
| 363 |
+
],
|
| 364 |
+
"logging_steps": 500,
|
| 365 |
+
"max_steps": 20000,
|
| 366 |
+
"num_input_tokens_seen": 0,
|
| 367 |
+
"num_train_epochs": 10,
|
| 368 |
+
"save_steps": 1000,
|
| 369 |
+
"stateful_callbacks": {
|
| 370 |
+
"TrainerControl": {
|
| 371 |
+
"args": {
|
| 372 |
+
"should_epoch_stop": false,
|
| 373 |
+
"should_evaluate": false,
|
| 374 |
+
"should_log": false,
|
| 375 |
+
"should_save": true,
|
| 376 |
+
"should_training_stop": false
|
| 377 |
+
},
|
| 378 |
+
"attributes": {}
|
| 379 |
+
}
|
| 380 |
+
},
|
| 381 |
+
"total_flos": 0.0,
|
| 382 |
+
"train_batch_size": 16,
|
| 383 |
+
"trial_name": null,
|
| 384 |
+
"trial_params": null
|
| 385 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d51a45fcd2faaf0729daca137475d136ce44396ee388ddfccb539e1262382a4a
|
| 3 |
+
size 5752
|
virus_encoder.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20c98cb24be6adb18a9d382e2b628ad90a98b79687d55dd2a338b170e2751037
|
| 3 |
+
size 30978
|
vp_encoder.joblib
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3058d8589781bb74c1cb50c82ebc536256ad5a6f09ec8d5852610e6873fd44b
|
| 3 |
+
size 625
|