checkpoint 8000
Browse files- model.safetensors +1 -1
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +661 -0
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1262032932
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5425a36f1e325bee6d82b1fcc07102ff534ecb18f72844c64a5049b9705ee1a2
|
| 3 |
size 1262032932
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40c8da2258f27071c4927075f92382af808544f48a8473848bacbf9ee6bd88e1
|
| 3 |
+
size 2490610241
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee893b4ad54887290f9b4069066f0404d323d7fb2e1360ff8f1c42e1bb7aca50
|
| 3 |
+
size 14639
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8133a804599afbc4c397ae559484c26a48033b26cd1cb76daab66148cf4c67eb
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,661 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3648131787776947,
|
| 3 |
+
"best_model_checkpoint": "./working/checkpoint-8000",
|
| 4 |
+
"epoch": 95.23809523809524,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 8000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.19,
|
| 13 |
+
"grad_norm": 5.874335765838623,
|
| 14 |
+
"learning_rate": 1.5e-06,
|
| 15 |
+
"loss": 12.5691,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 2.38,
|
| 20 |
+
"grad_norm": 11.808794975280762,
|
| 21 |
+
"learning_rate": 2.9850000000000002e-06,
|
| 22 |
+
"loss": 9.8251,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 3.57,
|
| 27 |
+
"grad_norm": 10.97357177734375,
|
| 28 |
+
"learning_rate": 4.485e-06,
|
| 29 |
+
"loss": 6.5329,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 4.76,
|
| 34 |
+
"grad_norm": 10.752958297729492,
|
| 35 |
+
"learning_rate": 5.985e-06,
|
| 36 |
+
"loss": 5.2405,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 5.95,
|
| 41 |
+
"grad_norm": 5.865973949432373,
|
| 42 |
+
"learning_rate": 7.485e-06,
|
| 43 |
+
"loss": 4.629,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 7.14,
|
| 48 |
+
"grad_norm": 4.192840099334717,
|
| 49 |
+
"learning_rate": 8.985e-06,
|
| 50 |
+
"loss": 4.2437,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 8.33,
|
| 55 |
+
"grad_norm": 3.2876224517822266,
|
| 56 |
+
"learning_rate": 1.0485e-05,
|
| 57 |
+
"loss": 3.9784,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 9.52,
|
| 62 |
+
"grad_norm": 1.2407236099243164,
|
| 63 |
+
"learning_rate": 1.1985000000000001e-05,
|
| 64 |
+
"loss": 3.7904,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 10.71,
|
| 69 |
+
"grad_norm": 0.6437963247299194,
|
| 70 |
+
"learning_rate": 1.3485e-05,
|
| 71 |
+
"loss": 3.6682,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 11.9,
|
| 76 |
+
"grad_norm": 0.3295372724533081,
|
| 77 |
+
"learning_rate": 1.4985e-05,
|
| 78 |
+
"loss": 3.5325,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 11.9,
|
| 83 |
+
"eval_cer": 0.926601333602748,
|
| 84 |
+
"eval_loss": 3.4897494316101074,
|
| 85 |
+
"eval_runtime": 10.3726,
|
| 86 |
+
"eval_samples_per_second": 32.393,
|
| 87 |
+
"eval_steps_per_second": 4.049,
|
| 88 |
+
"eval_wer": 1.0,
|
| 89 |
+
"step": 1000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 13.1,
|
| 93 |
+
"grad_norm": 0.61485356092453,
|
| 94 |
+
"learning_rate": 1.6485e-05,
|
| 95 |
+
"loss": 3.4934,
|
| 96 |
+
"step": 1100
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 14.29,
|
| 100 |
+
"grad_norm": 0.4778901934623718,
|
| 101 |
+
"learning_rate": 1.7985e-05,
|
| 102 |
+
"loss": 3.4662,
|
| 103 |
+
"step": 1200
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 15.48,
|
| 107 |
+
"grad_norm": 0.56944739818573,
|
| 108 |
+
"learning_rate": 1.9485e-05,
|
| 109 |
+
"loss": 3.4413,
|
| 110 |
+
"step": 1300
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 16.67,
|
| 114 |
+
"grad_norm": 0.37235376238822937,
|
| 115 |
+
"learning_rate": 2.0985e-05,
|
| 116 |
+
"loss": 3.4089,
|
| 117 |
+
"step": 1400
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 17.86,
|
| 121 |
+
"grad_norm": 0.4277956187725067,
|
| 122 |
+
"learning_rate": 2.2485000000000002e-05,
|
| 123 |
+
"loss": 3.3384,
|
| 124 |
+
"step": 1500
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 19.05,
|
| 128 |
+
"grad_norm": 0.50371253490448,
|
| 129 |
+
"learning_rate": 2.3985e-05,
|
| 130 |
+
"loss": 3.2455,
|
| 131 |
+
"step": 1600
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 20.24,
|
| 135 |
+
"grad_norm": 1.1383576393127441,
|
| 136 |
+
"learning_rate": 2.5485e-05,
|
| 137 |
+
"loss": 3.0681,
|
| 138 |
+
"step": 1700
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 21.43,
|
| 142 |
+
"grad_norm": 0.8668686747550964,
|
| 143 |
+
"learning_rate": 2.6985e-05,
|
| 144 |
+
"loss": 2.7949,
|
| 145 |
+
"step": 1800
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 22.62,
|
| 149 |
+
"grad_norm": 1.0731563568115234,
|
| 150 |
+
"learning_rate": 2.8485000000000003e-05,
|
| 151 |
+
"loss": 2.4866,
|
| 152 |
+
"step": 1900
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 23.81,
|
| 156 |
+
"grad_norm": 1.3317248821258545,
|
| 157 |
+
"learning_rate": 2.9985000000000002e-05,
|
| 158 |
+
"loss": 2.1973,
|
| 159 |
+
"step": 2000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 23.81,
|
| 163 |
+
"eval_cer": 0.24030107092341887,
|
| 164 |
+
"eval_loss": 1.1350404024124146,
|
| 165 |
+
"eval_runtime": 10.3527,
|
| 166 |
+
"eval_samples_per_second": 32.455,
|
| 167 |
+
"eval_steps_per_second": 4.057,
|
| 168 |
+
"eval_wer": 0.839647119875454,
|
| 169 |
+
"step": 2000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 25.0,
|
| 173 |
+
"grad_norm": 2.0239453315734863,
|
| 174 |
+
"learning_rate": 2.962875e-05,
|
| 175 |
+
"loss": 1.9821,
|
| 176 |
+
"step": 2100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 26.19,
|
| 180 |
+
"grad_norm": 1.464921236038208,
|
| 181 |
+
"learning_rate": 2.925375e-05,
|
| 182 |
+
"loss": 1.853,
|
| 183 |
+
"step": 2200
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 27.38,
|
| 187 |
+
"grad_norm": 1.6508703231811523,
|
| 188 |
+
"learning_rate": 2.887875e-05,
|
| 189 |
+
"loss": 1.7547,
|
| 190 |
+
"step": 2300
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 28.57,
|
| 194 |
+
"grad_norm": 1.3476407527923584,
|
| 195 |
+
"learning_rate": 2.850375e-05,
|
| 196 |
+
"loss": 1.7171,
|
| 197 |
+
"step": 2400
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 29.76,
|
| 201 |
+
"grad_norm": 1.2977994680404663,
|
| 202 |
+
"learning_rate": 2.812875e-05,
|
| 203 |
+
"loss": 1.6498,
|
| 204 |
+
"step": 2500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 30.95,
|
| 208 |
+
"grad_norm": 1.8536533117294312,
|
| 209 |
+
"learning_rate": 2.775375e-05,
|
| 210 |
+
"loss": 1.5965,
|
| 211 |
+
"step": 2600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 32.14,
|
| 215 |
+
"grad_norm": 1.7063647508621216,
|
| 216 |
+
"learning_rate": 2.7378750000000003e-05,
|
| 217 |
+
"loss": 1.5744,
|
| 218 |
+
"step": 2700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 33.33,
|
| 222 |
+
"grad_norm": 1.613274097442627,
|
| 223 |
+
"learning_rate": 2.700375e-05,
|
| 224 |
+
"loss": 1.5483,
|
| 225 |
+
"step": 2800
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 34.52,
|
| 229 |
+
"grad_norm": 1.6182752847671509,
|
| 230 |
+
"learning_rate": 2.662875e-05,
|
| 231 |
+
"loss": 1.5076,
|
| 232 |
+
"step": 2900
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 35.71,
|
| 236 |
+
"grad_norm": 1.9291083812713623,
|
| 237 |
+
"learning_rate": 2.6253750000000003e-05,
|
| 238 |
+
"loss": 1.4762,
|
| 239 |
+
"step": 3000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 35.71,
|
| 243 |
+
"eval_cer": 0.15634471610426348,
|
| 244 |
+
"eval_loss": 0.527005672454834,
|
| 245 |
+
"eval_runtime": 10.3175,
|
| 246 |
+
"eval_samples_per_second": 32.566,
|
| 247 |
+
"eval_steps_per_second": 4.071,
|
| 248 |
+
"eval_wer": 0.6844836533471718,
|
| 249 |
+
"step": 3000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 36.9,
|
| 253 |
+
"grad_norm": 1.7794642448425293,
|
| 254 |
+
"learning_rate": 2.587875e-05,
|
| 255 |
+
"loss": 1.4752,
|
| 256 |
+
"step": 3100
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 38.1,
|
| 260 |
+
"grad_norm": 1.7410004138946533,
|
| 261 |
+
"learning_rate": 2.550375e-05,
|
| 262 |
+
"loss": 1.4586,
|
| 263 |
+
"step": 3200
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 39.29,
|
| 267 |
+
"grad_norm": 1.6410831212997437,
|
| 268 |
+
"learning_rate": 2.512875e-05,
|
| 269 |
+
"loss": 1.4172,
|
| 270 |
+
"step": 3300
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 40.48,
|
| 274 |
+
"grad_norm": 2.350106954574585,
|
| 275 |
+
"learning_rate": 2.475375e-05,
|
| 276 |
+
"loss": 1.3751,
|
| 277 |
+
"step": 3400
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 41.67,
|
| 281 |
+
"grad_norm": 1.7794309854507446,
|
| 282 |
+
"learning_rate": 2.437875e-05,
|
| 283 |
+
"loss": 1.3516,
|
| 284 |
+
"step": 3500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 42.86,
|
| 288 |
+
"grad_norm": 1.8536804914474487,
|
| 289 |
+
"learning_rate": 2.400375e-05,
|
| 290 |
+
"loss": 1.338,
|
| 291 |
+
"step": 3600
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 44.05,
|
| 295 |
+
"grad_norm": 2.043091058731079,
|
| 296 |
+
"learning_rate": 2.362875e-05,
|
| 297 |
+
"loss": 1.31,
|
| 298 |
+
"step": 3700
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 45.24,
|
| 302 |
+
"grad_norm": 2.556605577468872,
|
| 303 |
+
"learning_rate": 2.325375e-05,
|
| 304 |
+
"loss": 1.2737,
|
| 305 |
+
"step": 3800
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 46.43,
|
| 309 |
+
"grad_norm": 2.167360544204712,
|
| 310 |
+
"learning_rate": 2.2878750000000002e-05,
|
| 311 |
+
"loss": 1.2706,
|
| 312 |
+
"step": 3900
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 47.62,
|
| 316 |
+
"grad_norm": 1.933358907699585,
|
| 317 |
+
"learning_rate": 2.2503750000000003e-05,
|
| 318 |
+
"loss": 1.2409,
|
| 319 |
+
"step": 4000
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 47.62,
|
| 323 |
+
"eval_cer": 0.140280864821176,
|
| 324 |
+
"eval_loss": 0.41946256160736084,
|
| 325 |
+
"eval_runtime": 10.4076,
|
| 326 |
+
"eval_samples_per_second": 32.284,
|
| 327 |
+
"eval_steps_per_second": 4.036,
|
| 328 |
+
"eval_wer": 0.6331084587441619,
|
| 329 |
+
"step": 4000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 48.81,
|
| 333 |
+
"grad_norm": 2.1927788257598877,
|
| 334 |
+
"learning_rate": 2.212875e-05,
|
| 335 |
+
"loss": 1.2476,
|
| 336 |
+
"step": 4100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 50.0,
|
| 340 |
+
"grad_norm": 3.7042958736419678,
|
| 341 |
+
"learning_rate": 2.175375e-05,
|
| 342 |
+
"loss": 1.2211,
|
| 343 |
+
"step": 4200
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 51.19,
|
| 347 |
+
"grad_norm": 2.503298282623291,
|
| 348 |
+
"learning_rate": 2.13825e-05,
|
| 349 |
+
"loss": 1.1974,
|
| 350 |
+
"step": 4300
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 52.38,
|
| 354 |
+
"grad_norm": 2.378753423690796,
|
| 355 |
+
"learning_rate": 2.101125e-05,
|
| 356 |
+
"loss": 1.19,
|
| 357 |
+
"step": 4400
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 53.57,
|
| 361 |
+
"grad_norm": 6.145068645477295,
|
| 362 |
+
"learning_rate": 2.063625e-05,
|
| 363 |
+
"loss": 1.1734,
|
| 364 |
+
"step": 4500
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 54.76,
|
| 368 |
+
"grad_norm": 2.2741551399230957,
|
| 369 |
+
"learning_rate": 2.026125e-05,
|
| 370 |
+
"loss": 1.1664,
|
| 371 |
+
"step": 4600
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 55.95,
|
| 375 |
+
"grad_norm": 3.21976900100708,
|
| 376 |
+
"learning_rate": 1.988625e-05,
|
| 377 |
+
"loss": 1.1555,
|
| 378 |
+
"step": 4700
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 57.14,
|
| 382 |
+
"grad_norm": 3.06923508644104,
|
| 383 |
+
"learning_rate": 1.951125e-05,
|
| 384 |
+
"loss": 1.1391,
|
| 385 |
+
"step": 4800
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 58.33,
|
| 389 |
+
"grad_norm": 1.8809341192245483,
|
| 390 |
+
"learning_rate": 1.9136249999999998e-05,
|
| 391 |
+
"loss": 1.1271,
|
| 392 |
+
"step": 4900
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 59.52,
|
| 396 |
+
"grad_norm": 2.041844367980957,
|
| 397 |
+
"learning_rate": 1.876125e-05,
|
| 398 |
+
"loss": 1.1241,
|
| 399 |
+
"step": 5000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 59.52,
|
| 403 |
+
"eval_cer": 0.13785613255203072,
|
| 404 |
+
"eval_loss": 0.38446417450904846,
|
| 405 |
+
"eval_runtime": 10.417,
|
| 406 |
+
"eval_samples_per_second": 32.255,
|
| 407 |
+
"eval_steps_per_second": 4.032,
|
| 408 |
+
"eval_wer": 0.63622210690192,
|
| 409 |
+
"step": 5000
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 60.71,
|
| 413 |
+
"grad_norm": 2.3043465614318848,
|
| 414 |
+
"learning_rate": 1.838625e-05,
|
| 415 |
+
"loss": 1.1042,
|
| 416 |
+
"step": 5100
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 61.9,
|
| 420 |
+
"grad_norm": 3.682835340499878,
|
| 421 |
+
"learning_rate": 1.801125e-05,
|
| 422 |
+
"loss": 1.0921,
|
| 423 |
+
"step": 5200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 63.1,
|
| 427 |
+
"grad_norm": 4.466809272766113,
|
| 428 |
+
"learning_rate": 1.7636250000000002e-05,
|
| 429 |
+
"loss": 1.095,
|
| 430 |
+
"step": 5300
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 64.29,
|
| 434 |
+
"grad_norm": 2.615339994430542,
|
| 435 |
+
"learning_rate": 1.726125e-05,
|
| 436 |
+
"loss": 1.1057,
|
| 437 |
+
"step": 5400
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 65.48,
|
| 441 |
+
"grad_norm": 3.483346700668335,
|
| 442 |
+
"learning_rate": 1.688625e-05,
|
| 443 |
+
"loss": 1.0553,
|
| 444 |
+
"step": 5500
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 66.67,
|
| 448 |
+
"grad_norm": 2.141965866088867,
|
| 449 |
+
"learning_rate": 1.651125e-05,
|
| 450 |
+
"loss": 1.0656,
|
| 451 |
+
"step": 5600
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 67.86,
|
| 455 |
+
"grad_norm": 2.2111611366271973,
|
| 456 |
+
"learning_rate": 1.613625e-05,
|
| 457 |
+
"loss": 1.0673,
|
| 458 |
+
"step": 5700
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 69.05,
|
| 462 |
+
"grad_norm": 2.071429491043091,
|
| 463 |
+
"learning_rate": 1.576125e-05,
|
| 464 |
+
"loss": 1.0632,
|
| 465 |
+
"step": 5800
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 70.24,
|
| 469 |
+
"grad_norm": 4.86116886138916,
|
| 470 |
+
"learning_rate": 1.538625e-05,
|
| 471 |
+
"loss": 1.0447,
|
| 472 |
+
"step": 5900
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 71.43,
|
| 476 |
+
"grad_norm": 3.3076369762420654,
|
| 477 |
+
"learning_rate": 1.5011250000000001e-05,
|
| 478 |
+
"loss": 1.024,
|
| 479 |
+
"step": 6000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 71.43,
|
| 483 |
+
"eval_cer": 0.13548191553849262,
|
| 484 |
+
"eval_loss": 0.3715837895870209,
|
| 485 |
+
"eval_runtime": 10.3955,
|
| 486 |
+
"eval_samples_per_second": 32.322,
|
| 487 |
+
"eval_steps_per_second": 4.04,
|
| 488 |
+
"eval_wer": 0.6320705760249092,
|
| 489 |
+
"step": 6000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 72.62,
|
| 493 |
+
"grad_norm": 2.3496508598327637,
|
| 494 |
+
"learning_rate": 1.463625e-05,
|
| 495 |
+
"loss": 1.0379,
|
| 496 |
+
"step": 6100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 73.81,
|
| 500 |
+
"grad_norm": 2.621004343032837,
|
| 501 |
+
"learning_rate": 1.426125e-05,
|
| 502 |
+
"loss": 1.0515,
|
| 503 |
+
"step": 6200
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 75.0,
|
| 507 |
+
"grad_norm": 5.240926742553711,
|
| 508 |
+
"learning_rate": 1.388625e-05,
|
| 509 |
+
"loss": 1.0253,
|
| 510 |
+
"step": 6300
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 76.19,
|
| 514 |
+
"grad_norm": 2.9943532943725586,
|
| 515 |
+
"learning_rate": 1.351125e-05,
|
| 516 |
+
"loss": 1.0131,
|
| 517 |
+
"step": 6400
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 77.38,
|
| 521 |
+
"grad_norm": 2.475804328918457,
|
| 522 |
+
"learning_rate": 1.3136250000000001e-05,
|
| 523 |
+
"loss": 1.0227,
|
| 524 |
+
"step": 6500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 78.57,
|
| 528 |
+
"grad_norm": 2.5056631565093994,
|
| 529 |
+
"learning_rate": 1.2761250000000001e-05,
|
| 530 |
+
"loss": 1.0025,
|
| 531 |
+
"step": 6600
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 79.76,
|
| 535 |
+
"grad_norm": 3.9102323055267334,
|
| 536 |
+
"learning_rate": 1.238625e-05,
|
| 537 |
+
"loss": 1.0181,
|
| 538 |
+
"step": 6700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 80.95,
|
| 542 |
+
"grad_norm": 3.3800106048583984,
|
| 543 |
+
"learning_rate": 1.201125e-05,
|
| 544 |
+
"loss": 0.9892,
|
| 545 |
+
"step": 6800
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 82.14,
|
| 549 |
+
"grad_norm": 2.165987014770508,
|
| 550 |
+
"learning_rate": 1.164e-05,
|
| 551 |
+
"loss": 0.9822,
|
| 552 |
+
"step": 6900
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 83.33,
|
| 556 |
+
"grad_norm": 2.426816463470459,
|
| 557 |
+
"learning_rate": 1.1265e-05,
|
| 558 |
+
"loss": 0.9922,
|
| 559 |
+
"step": 7000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 83.33,
|
| 563 |
+
"eval_cer": 0.13310769852495455,
|
| 564 |
+
"eval_loss": 0.3727741539478302,
|
| 565 |
+
"eval_runtime": 10.3942,
|
| 566 |
+
"eval_samples_per_second": 32.326,
|
| 567 |
+
"eval_steps_per_second": 4.041,
|
| 568 |
+
"eval_wer": 0.628956927867151,
|
| 569 |
+
"step": 7000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 84.52,
|
| 573 |
+
"grad_norm": 3.203552007675171,
|
| 574 |
+
"learning_rate": 1.089e-05,
|
| 575 |
+
"loss": 0.9995,
|
| 576 |
+
"step": 7100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 85.71,
|
| 580 |
+
"grad_norm": 2.827246904373169,
|
| 581 |
+
"learning_rate": 1.0515e-05,
|
| 582 |
+
"loss": 0.9683,
|
| 583 |
+
"step": 7200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 86.9,
|
| 587 |
+
"grad_norm": 2.9927895069122314,
|
| 588 |
+
"learning_rate": 1.0140000000000001e-05,
|
| 589 |
+
"loss": 0.9864,
|
| 590 |
+
"step": 7300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 88.1,
|
| 594 |
+
"grad_norm": 2.151737928390503,
|
| 595 |
+
"learning_rate": 9.765e-06,
|
| 596 |
+
"loss": 0.9744,
|
| 597 |
+
"step": 7400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 89.29,
|
| 601 |
+
"grad_norm": 2.5920581817626953,
|
| 602 |
+
"learning_rate": 9.39e-06,
|
| 603 |
+
"loss": 0.9794,
|
| 604 |
+
"step": 7500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 90.48,
|
| 608 |
+
"grad_norm": 3.2127621173858643,
|
| 609 |
+
"learning_rate": 9.015e-06,
|
| 610 |
+
"loss": 0.9622,
|
| 611 |
+
"step": 7600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 91.67,
|
| 615 |
+
"grad_norm": 3.541879892349243,
|
| 616 |
+
"learning_rate": 8.64e-06,
|
| 617 |
+
"loss": 0.9624,
|
| 618 |
+
"step": 7700
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 92.86,
|
| 622 |
+
"grad_norm": 2.827958345413208,
|
| 623 |
+
"learning_rate": 8.265000000000001e-06,
|
| 624 |
+
"loss": 0.9637,
|
| 625 |
+
"step": 7800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 94.05,
|
| 629 |
+
"grad_norm": 2.648591995239258,
|
| 630 |
+
"learning_rate": 7.89e-06,
|
| 631 |
+
"loss": 0.9684,
|
| 632 |
+
"step": 7900
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 95.24,
|
| 636 |
+
"grad_norm": 4.264640808105469,
|
| 637 |
+
"learning_rate": 7.515e-06,
|
| 638 |
+
"loss": 0.9432,
|
| 639 |
+
"step": 8000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 95.24,
|
| 643 |
+
"eval_cer": 0.13209739341281068,
|
| 644 |
+
"eval_loss": 0.3648131787776947,
|
| 645 |
+
"eval_runtime": 10.4236,
|
| 646 |
+
"eval_samples_per_second": 32.234,
|
| 647 |
+
"eval_steps_per_second": 4.029,
|
| 648 |
+
"eval_wer": 0.6170212765957447,
|
| 649 |
+
"step": 8000
|
| 650 |
+
}
|
| 651 |
+
],
|
| 652 |
+
"logging_steps": 100,
|
| 653 |
+
"max_steps": 10000,
|
| 654 |
+
"num_input_tokens_seen": 0,
|
| 655 |
+
"num_train_epochs": 120,
|
| 656 |
+
"save_steps": 1000,
|
| 657 |
+
"total_flos": 2.4098927133551096e+19,
|
| 658 |
+
"train_batch_size": 8,
|
| 659 |
+
"trial_name": null,
|
| 660 |
+
"trial_params": null
|
| 661 |
+
}
|