Commit
•
d8e2c33
1
Parent(s):
f958a6d
Training in progress, step 700000
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +311 -3
- pytorch_model.bin +1 -1
- runs/Nov21_18-14-51_t1v-n-088af867-w-0/events.out.tfevents.1669054551.t1v-n-088af867-w-0.42571.0 +2 -2
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 996067161
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae43a48589175aaf9939e78cb22924c6e7b9714449e1a77b6aea2674d6f0f535
|
| 3 |
size 996067161
|
last-checkpoint/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 498046827
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b889229f6c63b75785b3963b98f50b36fa64eab02160a85abd3d76c7eca1972b
|
| 3 |
size 498046827
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e8d8fd096b22974f0a808c190b1c67b3a173ac6818b5723a1c553da12bd61ad
|
| 3 |
size 13611
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e8d8fd096b22974f0a808c190b1c67b3a173ac6818b5723a1c553da12bd61ad
|
| 3 |
size 13611
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e8d8fd096b22974f0a808c190b1c67b3a173ac6818b5723a1c553da12bd61ad
|
| 3 |
size 13611
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:237857aa9cd57869e02f5c3fbb02afb8a7fe52dc11b11ceb54d6be8920f4af97
|
| 3 |
size 13611
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:237857aa9cd57869e02f5c3fbb02afb8a7fe52dc11b11ceb54d6be8920f4af97
|
| 3 |
size 13611
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:237857aa9cd57869e02f5c3fbb02afb8a7fe52dc11b11ceb54d6be8920f4af97
|
| 3 |
size 13611
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:237857aa9cd57869e02f5c3fbb02afb8a7fe52dc11b11ceb54d6be8920f4af97
|
| 3 |
size 13611
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e8d8fd096b22974f0a808c190b1c67b3a173ac6818b5723a1c553da12bd61ad
|
| 3 |
size 13611
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d2d6ab5c2c0046e2c66cf435ff00463ff3da7208ddb35b5b6f19c87d94f3623b
|
| 3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 2.
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -4010,11 +4010,319 @@
|
|
| 4010 |
"eval_samples_per_second": 593.171,
|
| 4011 |
"eval_steps_per_second": 2.373,
|
| 4012 |
"step": 650000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4013 |
}
|
| 4014 |
],
|
| 4015 |
"max_steps": 1000000,
|
| 4016 |
"num_train_epochs": 9223372036854775807,
|
| 4017 |
-
"total_flos": 1.
|
| 4018 |
"trial_name": null,
|
| 4019 |
"trial_params": null
|
| 4020 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.165921,
|
| 5 |
+
"global_step": 700000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 4010 |
"eval_samples_per_second": 593.171,
|
| 4011 |
"eval_steps_per_second": 2.373,
|
| 4012 |
"step": 650000
|
| 4013 |
+
},
|
| 4014 |
+
{
|
| 4015 |
+
"epoch": 2.12,
|
| 4016 |
+
"learning_rate": 2.976391850971065e-05,
|
| 4017 |
+
"loss": 0.8032,
|
| 4018 |
+
"step": 651000
|
| 4019 |
+
},
|
| 4020 |
+
{
|
| 4021 |
+
"epoch": 2.12,
|
| 4022 |
+
"learning_rate": 2.9612829550614836e-05,
|
| 4023 |
+
"loss": 0.814,
|
| 4024 |
+
"step": 652000
|
| 4025 |
+
},
|
| 4026 |
+
{
|
| 4027 |
+
"epoch": 2.12,
|
| 4028 |
+
"learning_rate": 2.9461963542348737e-05,
|
| 4029 |
+
"loss": 0.8232,
|
| 4030 |
+
"step": 653000
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 2.12,
|
| 4034 |
+
"learning_rate": 2.931132213475884e-05,
|
| 4035 |
+
"loss": 0.8142,
|
| 4036 |
+
"step": 654000
|
| 4037 |
+
},
|
| 4038 |
+
{
|
| 4039 |
+
"epoch": 2.12,
|
| 4040 |
+
"learning_rate": 2.916090697523549e-05,
|
| 4041 |
+
"loss": 0.8054,
|
| 4042 |
+
"step": 655000
|
| 4043 |
+
},
|
| 4044 |
+
{
|
| 4045 |
+
"epoch": 2.12,
|
| 4046 |
+
"learning_rate": 2.9010719708694722e-05,
|
| 4047 |
+
"loss": 0.8242,
|
| 4048 |
+
"step": 656000
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 2.12,
|
| 4052 |
+
"learning_rate": 2.8860761977560436e-05,
|
| 4053 |
+
"loss": 0.8183,
|
| 4054 |
+
"step": 657000
|
| 4055 |
+
},
|
| 4056 |
+
{
|
| 4057 |
+
"epoch": 2.12,
|
| 4058 |
+
"learning_rate": 2.8711035421746367e-05,
|
| 4059 |
+
"loss": 0.8271,
|
| 4060 |
+
"step": 658000
|
| 4061 |
+
},
|
| 4062 |
+
{
|
| 4063 |
+
"epoch": 2.12,
|
| 4064 |
+
"learning_rate": 2.8561541678638142e-05,
|
| 4065 |
+
"loss": 0.8261,
|
| 4066 |
+
"step": 659000
|
| 4067 |
+
},
|
| 4068 |
+
{
|
| 4069 |
+
"epoch": 2.13,
|
| 4070 |
+
"learning_rate": 2.8412282383075363e-05,
|
| 4071 |
+
"loss": 0.8267,
|
| 4072 |
+
"step": 660000
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 2.13,
|
| 4076 |
+
"learning_rate": 2.8263259167333777e-05,
|
| 4077 |
+
"loss": 0.8353,
|
| 4078 |
+
"step": 661000
|
| 4079 |
+
},
|
| 4080 |
+
{
|
| 4081 |
+
"epoch": 2.13,
|
| 4082 |
+
"learning_rate": 2.811447366110741e-05,
|
| 4083 |
+
"loss": 0.8273,
|
| 4084 |
+
"step": 662000
|
| 4085 |
+
},
|
| 4086 |
+
{
|
| 4087 |
+
"epoch": 2.13,
|
| 4088 |
+
"learning_rate": 2.7965927491490705e-05,
|
| 4089 |
+
"loss": 0.8259,
|
| 4090 |
+
"step": 663000
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 2.13,
|
| 4094 |
+
"learning_rate": 2.7817622282960815e-05,
|
| 4095 |
+
"loss": 0.8215,
|
| 4096 |
+
"step": 664000
|
| 4097 |
+
},
|
| 4098 |
+
{
|
| 4099 |
+
"epoch": 2.13,
|
| 4100 |
+
"learning_rate": 2.766955965735968e-05,
|
| 4101 |
+
"loss": 0.8273,
|
| 4102 |
+
"step": 665000
|
| 4103 |
+
},
|
| 4104 |
+
{
|
| 4105 |
+
"epoch": 2.13,
|
| 4106 |
+
"learning_rate": 2.7521741233876496e-05,
|
| 4107 |
+
"loss": 0.8289,
|
| 4108 |
+
"step": 666000
|
| 4109 |
+
},
|
| 4110 |
+
{
|
| 4111 |
+
"epoch": 2.13,
|
| 4112 |
+
"learning_rate": 2.7374168629029813e-05,
|
| 4113 |
+
"loss": 0.8166,
|
| 4114 |
+
"step": 667000
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 2.13,
|
| 4118 |
+
"learning_rate": 2.7226843456650037e-05,
|
| 4119 |
+
"loss": 0.8251,
|
| 4120 |
+
"step": 668000
|
| 4121 |
+
},
|
| 4122 |
+
{
|
| 4123 |
+
"epoch": 2.13,
|
| 4124 |
+
"learning_rate": 2.707976732786166e-05,
|
| 4125 |
+
"loss": 0.8231,
|
| 4126 |
+
"step": 669000
|
| 4127 |
+
},
|
| 4128 |
+
{
|
| 4129 |
+
"epoch": 2.14,
|
| 4130 |
+
"learning_rate": 2.693294185106562e-05,
|
| 4131 |
+
"loss": 0.8247,
|
| 4132 |
+
"step": 670000
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 2.14,
|
| 4136 |
+
"learning_rate": 2.6786368631921836e-05,
|
| 4137 |
+
"loss": 0.8201,
|
| 4138 |
+
"step": 671000
|
| 4139 |
+
},
|
| 4140 |
+
{
|
| 4141 |
+
"epoch": 2.14,
|
| 4142 |
+
"learning_rate": 2.6640049273331515e-05,
|
| 4143 |
+
"loss": 0.8287,
|
| 4144 |
+
"step": 672000
|
| 4145 |
+
},
|
| 4146 |
+
{
|
| 4147 |
+
"epoch": 2.14,
|
| 4148 |
+
"learning_rate": 2.6493985375419778e-05,
|
| 4149 |
+
"loss": 0.828,
|
| 4150 |
+
"step": 673000
|
| 4151 |
+
},
|
| 4152 |
+
{
|
| 4153 |
+
"epoch": 2.14,
|
| 4154 |
+
"learning_rate": 2.6348178535517966e-05,
|
| 4155 |
+
"loss": 0.8333,
|
| 4156 |
+
"step": 674000
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 2.14,
|
| 4160 |
+
"learning_rate": 2.6202630348146324e-05,
|
| 4161 |
+
"loss": 0.8111,
|
| 4162 |
+
"step": 675000
|
| 4163 |
+
},
|
| 4164 |
+
{
|
| 4165 |
+
"epoch": 2.14,
|
| 4166 |
+
"learning_rate": 2.6057342404996522e-05,
|
| 4167 |
+
"loss": 0.8129,
|
| 4168 |
+
"step": 676000
|
| 4169 |
+
},
|
| 4170 |
+
{
|
| 4171 |
+
"epoch": 2.14,
|
| 4172 |
+
"learning_rate": 2.591231629491423e-05,
|
| 4173 |
+
"loss": 0.8129,
|
| 4174 |
+
"step": 677000
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 2.14,
|
| 4178 |
+
"learning_rate": 2.5767553603881767e-05,
|
| 4179 |
+
"loss": 0.8126,
|
| 4180 |
+
"step": 678000
|
| 4181 |
+
},
|
| 4182 |
+
{
|
| 4183 |
+
"epoch": 2.14,
|
| 4184 |
+
"learning_rate": 2.562305591500069e-05,
|
| 4185 |
+
"loss": 0.8278,
|
| 4186 |
+
"step": 679000
|
| 4187 |
+
},
|
| 4188 |
+
{
|
| 4189 |
+
"epoch": 2.15,
|
| 4190 |
+
"learning_rate": 2.547882480847461e-05,
|
| 4191 |
+
"loss": 0.8143,
|
| 4192 |
+
"step": 680000
|
| 4193 |
+
},
|
| 4194 |
+
{
|
| 4195 |
+
"epoch": 2.15,
|
| 4196 |
+
"learning_rate": 2.5334861861591753e-05,
|
| 4197 |
+
"loss": 0.8143,
|
| 4198 |
+
"step": 681000
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 2.15,
|
| 4202 |
+
"learning_rate": 2.5191168648707887e-05,
|
| 4203 |
+
"loss": 0.8076,
|
| 4204 |
+
"step": 682000
|
| 4205 |
+
},
|
| 4206 |
+
{
|
| 4207 |
+
"epoch": 2.15,
|
| 4208 |
+
"learning_rate": 2.5047746741228978e-05,
|
| 4209 |
+
"loss": 0.8209,
|
| 4210 |
+
"step": 683000
|
| 4211 |
+
},
|
| 4212 |
+
{
|
| 4213 |
+
"epoch": 2.15,
|
| 4214 |
+
"learning_rate": 2.490459770759398e-05,
|
| 4215 |
+
"loss": 0.8208,
|
| 4216 |
+
"step": 684000
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 2.15,
|
| 4220 |
+
"learning_rate": 2.476172311325783e-05,
|
| 4221 |
+
"loss": 0.8155,
|
| 4222 |
+
"step": 685000
|
| 4223 |
+
},
|
| 4224 |
+
{
|
| 4225 |
+
"epoch": 2.15,
|
| 4226 |
+
"learning_rate": 2.4619124520674146e-05,
|
| 4227 |
+
"loss": 0.8124,
|
| 4228 |
+
"step": 686000
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 2.15,
|
| 4232 |
+
"learning_rate": 2.447680348927837e-05,
|
| 4233 |
+
"loss": 0.8197,
|
| 4234 |
+
"step": 687000
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 2.15,
|
| 4238 |
+
"learning_rate": 2.433476157547044e-05,
|
| 4239 |
+
"loss": 0.8197,
|
| 4240 |
+
"step": 688000
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 2.15,
|
| 4244 |
+
"learning_rate": 2.419300033259798e-05,
|
| 4245 |
+
"loss": 0.8159,
|
| 4246 |
+
"step": 689000
|
| 4247 |
+
},
|
| 4248 |
+
{
|
| 4249 |
+
"epoch": 2.16,
|
| 4250 |
+
"learning_rate": 2.405152131093926e-05,
|
| 4251 |
+
"loss": 0.807,
|
| 4252 |
+
"step": 690000
|
| 4253 |
+
},
|
| 4254 |
+
{
|
| 4255 |
+
"epoch": 2.16,
|
| 4256 |
+
"learning_rate": 2.3910326057686127e-05,
|
| 4257 |
+
"loss": 0.8181,
|
| 4258 |
+
"step": 691000
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 2.16,
|
| 4262 |
+
"learning_rate": 2.3769416116927335e-05,
|
| 4263 |
+
"loss": 0.819,
|
| 4264 |
+
"step": 692000
|
| 4265 |
+
},
|
| 4266 |
+
{
|
| 4267 |
+
"epoch": 2.16,
|
| 4268 |
+
"learning_rate": 2.362879302963135e-05,
|
| 4269 |
+
"loss": 0.8163,
|
| 4270 |
+
"step": 693000
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"epoch": 2.16,
|
| 4274 |
+
"learning_rate": 2.3488458333629777e-05,
|
| 4275 |
+
"loss": 0.8037,
|
| 4276 |
+
"step": 694000
|
| 4277 |
+
},
|
| 4278 |
+
{
|
| 4279 |
+
"epoch": 2.16,
|
| 4280 |
+
"learning_rate": 2.3348413563600325e-05,
|
| 4281 |
+
"loss": 0.8144,
|
| 4282 |
+
"step": 695000
|
| 4283 |
+
},
|
| 4284 |
+
{
|
| 4285 |
+
"epoch": 2.16,
|
| 4286 |
+
"learning_rate": 2.3208660251050158e-05,
|
| 4287 |
+
"loss": 0.821,
|
| 4288 |
+
"step": 696000
|
| 4289 |
+
},
|
| 4290 |
+
{
|
| 4291 |
+
"epoch": 2.16,
|
| 4292 |
+
"learning_rate": 2.3069199924299174e-05,
|
| 4293 |
+
"loss": 0.8205,
|
| 4294 |
+
"step": 697000
|
| 4295 |
+
},
|
| 4296 |
+
{
|
| 4297 |
+
"epoch": 2.16,
|
| 4298 |
+
"learning_rate": 2.29300341084631e-05,
|
| 4299 |
+
"loss": 0.8125,
|
| 4300 |
+
"step": 698000
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 2.16,
|
| 4304 |
+
"learning_rate": 2.279116432543705e-05,
|
| 4305 |
+
"loss": 0.8239,
|
| 4306 |
+
"step": 699000
|
| 4307 |
+
},
|
| 4308 |
+
{
|
| 4309 |
+
"epoch": 2.17,
|
| 4310 |
+
"learning_rate": 2.2652592093878666e-05,
|
| 4311 |
+
"loss": 0.8369,
|
| 4312 |
+
"step": 700000
|
| 4313 |
+
},
|
| 4314 |
+
{
|
| 4315 |
+
"epoch": 2.17,
|
| 4316 |
+
"eval_loss": 0.7765893340110779,
|
| 4317 |
+
"eval_runtime": 17.8732,
|
| 4318 |
+
"eval_samples_per_second": 559.497,
|
| 4319 |
+
"eval_steps_per_second": 2.238,
|
| 4320 |
+
"step": 700000
|
| 4321 |
}
|
| 4322 |
],
|
| 4323 |
"max_steps": 1000000,
|
| 4324 |
"num_train_epochs": 9223372036854775807,
|
| 4325 |
+
"total_flos": 1.17942562848768e+19,
|
| 4326 |
"trial_name": null,
|
| 4327 |
"trial_params": null
|
| 4328 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 498046827
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b889229f6c63b75785b3963b98f50b36fa64eab02160a85abd3d76c7eca1972b
|
| 3 |
size 498046827
|
runs/Nov21_18-14-51_t1v-n-088af867-w-0/events.out.tfevents.1669054551.t1v-n-088af867-w-0.42571.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:760955c77a228959512a67fa9be4a80f48c8f4c3b85cf9260455f57c08dc70ae
|
| 3 |
+
size 69956
|