Training in progress, step 600, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 590925768
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d148803dec3242bacfa6090b788cc33df772e93b875035cf7e3a3d8c47dc0693
|
| 3 |
size 590925768
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 301533378
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2a243d4609f2848bbcd91d030f48333676cc98c340050529214e8a4c5cbc451
|
| 3 |
size 301533378
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5871d01553a694652bee01b474f7aa31386bfbdbca60584c13070b444a88c461
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0541d0446a58575a8e1bf0f51453829134e69b4b4f483226a260165a57705a1c
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3953,6 +3953,364 @@
|
|
| 3953 |
"eval_samples_per_second": 9.607,
|
| 3954 |
"eval_steps_per_second": 9.607,
|
| 3955 |
"step": 550
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3956 |
}
|
| 3957 |
],
|
| 3958 |
"logging_steps": 1,
|
|
@@ -3976,12 +4334,12 @@
|
|
| 3976 |
"should_evaluate": false,
|
| 3977 |
"should_log": false,
|
| 3978 |
"should_save": true,
|
| 3979 |
-
"should_training_stop":
|
| 3980 |
},
|
| 3981 |
"attributes": {}
|
| 3982 |
}
|
| 3983 |
},
|
| 3984 |
-
"total_flos": 5.
|
| 3985 |
"train_batch_size": 1,
|
| 3986 |
"trial_name": null,
|
| 3987 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.5509313941001892,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 0.06720806496779613,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3953 |
"eval_samples_per_second": 9.607,
|
| 3954 |
"eval_steps_per_second": 9.607,
|
| 3955 |
"step": 550
|
| 3956 |
+
},
|
| 3957 |
+
{
|
| 3958 |
+
"epoch": 0.06171940632875945,
|
| 3959 |
+
"grad_norm": 0.12340470403432846,
|
| 3960 |
+
"learning_rate": 5.07672955698109e-06,
|
| 3961 |
+
"loss": 0.529,
|
| 3962 |
+
"step": 551
|
| 3963 |
+
},
|
| 3964 |
+
{
|
| 3965 |
+
"epoch": 0.061831419770372445,
|
| 3966 |
+
"grad_norm": 0.16562983393669128,
|
| 3967 |
+
"learning_rate": 4.872748404735644e-06,
|
| 3968 |
+
"loss": 0.4595,
|
| 3969 |
+
"step": 552
|
| 3970 |
+
},
|
| 3971 |
+
{
|
| 3972 |
+
"epoch": 0.061943433211985435,
|
| 3973 |
+
"grad_norm": 0.15129025280475616,
|
| 3974 |
+
"learning_rate": 4.6728820045062954e-06,
|
| 3975 |
+
"loss": 0.6089,
|
| 3976 |
+
"step": 553
|
| 3977 |
+
},
|
| 3978 |
+
{
|
| 3979 |
+
"epoch": 0.06205544665359843,
|
| 3980 |
+
"grad_norm": 0.1321287453174591,
|
| 3981 |
+
"learning_rate": 4.477136023048727e-06,
|
| 3982 |
+
"loss": 0.5442,
|
| 3983 |
+
"step": 554
|
| 3984 |
+
},
|
| 3985 |
+
{
|
| 3986 |
+
"epoch": 0.06216746009521142,
|
| 3987 |
+
"grad_norm": 0.1615147441625595,
|
| 3988 |
+
"learning_rate": 4.285516010293522e-06,
|
| 3989 |
+
"loss": 0.5495,
|
| 3990 |
+
"step": 555
|
| 3991 |
+
},
|
| 3992 |
+
{
|
| 3993 |
+
"epoch": 0.06227947353682442,
|
| 3994 |
+
"grad_norm": 0.1276487559080124,
|
| 3995 |
+
"learning_rate": 4.098027399188802e-06,
|
| 3996 |
+
"loss": 0.4491,
|
| 3997 |
+
"step": 556
|
| 3998 |
+
},
|
| 3999 |
+
{
|
| 4000 |
+
"epoch": 0.06239148697843741,
|
| 4001 |
+
"grad_norm": 0.1305229514837265,
|
| 4002 |
+
"learning_rate": 3.914675505546277e-06,
|
| 4003 |
+
"loss": 0.63,
|
| 4004 |
+
"step": 557
|
| 4005 |
+
},
|
| 4006 |
+
{
|
| 4007 |
+
"epoch": 0.0625035004200504,
|
| 4008 |
+
"grad_norm": 0.13546405732631683,
|
| 4009 |
+
"learning_rate": 3.735465527890458e-06,
|
| 4010 |
+
"loss": 0.6798,
|
| 4011 |
+
"step": 558
|
| 4012 |
+
},
|
| 4013 |
+
{
|
| 4014 |
+
"epoch": 0.0626155138616634,
|
| 4015 |
+
"grad_norm": 0.13185566663742065,
|
| 4016 |
+
"learning_rate": 3.560402547311275e-06,
|
| 4017 |
+
"loss": 0.575,
|
| 4018 |
+
"step": 559
|
| 4019 |
+
},
|
| 4020 |
+
{
|
| 4021 |
+
"epoch": 0.0627275273032764,
|
| 4022 |
+
"grad_norm": 0.13408592343330383,
|
| 4023 |
+
"learning_rate": 3.3894915273199987e-06,
|
| 4024 |
+
"loss": 0.723,
|
| 4025 |
+
"step": 560
|
| 4026 |
+
},
|
| 4027 |
+
{
|
| 4028 |
+
"epoch": 0.06283954074488939,
|
| 4029 |
+
"grad_norm": 0.11883525550365448,
|
| 4030 |
+
"learning_rate": 3.2227373137085954e-06,
|
| 4031 |
+
"loss": 0.4747,
|
| 4032 |
+
"step": 561
|
| 4033 |
+
},
|
| 4034 |
+
{
|
| 4035 |
+
"epoch": 0.06295155418650238,
|
| 4036 |
+
"grad_norm": 0.14109228551387787,
|
| 4037 |
+
"learning_rate": 3.0601446344122095e-06,
|
| 4038 |
+
"loss": 0.4895,
|
| 4039 |
+
"step": 562
|
| 4040 |
+
},
|
| 4041 |
+
{
|
| 4042 |
+
"epoch": 0.06306356762811538,
|
| 4043 |
+
"grad_norm": 0.14065487682819366,
|
| 4044 |
+
"learning_rate": 2.9017180993752e-06,
|
| 4045 |
+
"loss": 0.7272,
|
| 4046 |
+
"step": 563
|
| 4047 |
+
},
|
| 4048 |
+
{
|
| 4049 |
+
"epoch": 0.06317558106972837,
|
| 4050 |
+
"grad_norm": 0.13580210506916046,
|
| 4051 |
+
"learning_rate": 2.7474622004204304e-06,
|
| 4052 |
+
"loss": 0.5744,
|
| 4053 |
+
"step": 564
|
| 4054 |
+
},
|
| 4055 |
+
{
|
| 4056 |
+
"epoch": 0.06328759451134136,
|
| 4057 |
+
"grad_norm": 0.13952845335006714,
|
| 4058 |
+
"learning_rate": 2.5973813111218546e-06,
|
| 4059 |
+
"loss": 0.3957,
|
| 4060 |
+
"step": 565
|
| 4061 |
+
},
|
| 4062 |
+
{
|
| 4063 |
+
"epoch": 0.06339960795295435,
|
| 4064 |
+
"grad_norm": 0.1433325558900833,
|
| 4065 |
+
"learning_rate": 2.4514796866805964e-06,
|
| 4066 |
+
"loss": 0.4962,
|
| 4067 |
+
"step": 566
|
| 4068 |
+
},
|
| 4069 |
+
{
|
| 4070 |
+
"epoch": 0.06351162139456734,
|
| 4071 |
+
"grad_norm": 0.14881394803524017,
|
| 4072 |
+
"learning_rate": 2.3097614638042493e-06,
|
| 4073 |
+
"loss": 0.5508,
|
| 4074 |
+
"step": 567
|
| 4075 |
+
},
|
| 4076 |
+
{
|
| 4077 |
+
"epoch": 0.06362363483618035,
|
| 4078 |
+
"grad_norm": 0.13152483105659485,
|
| 4079 |
+
"learning_rate": 2.1722306605896022e-06,
|
| 4080 |
+
"loss": 0.5332,
|
| 4081 |
+
"step": 568
|
| 4082 |
+
},
|
| 4083 |
+
{
|
| 4084 |
+
"epoch": 0.06373564827779334,
|
| 4085 |
+
"grad_norm": 0.15379270911216736,
|
| 4086 |
+
"learning_rate": 2.0388911764086966e-06,
|
| 4087 |
+
"loss": 0.5214,
|
| 4088 |
+
"step": 569
|
| 4089 |
+
},
|
| 4090 |
+
{
|
| 4091 |
+
"epoch": 0.06384766171940633,
|
| 4092 |
+
"grad_norm": 0.14026005566120148,
|
| 4093 |
+
"learning_rate": 1.909746791798317e-06,
|
| 4094 |
+
"loss": 0.473,
|
| 4095 |
+
"step": 570
|
| 4096 |
+
},
|
| 4097 |
+
{
|
| 4098 |
+
"epoch": 0.06395967516101932,
|
| 4099 |
+
"grad_norm": 0.14658816158771515,
|
| 4100 |
+
"learning_rate": 1.7848011683527562e-06,
|
| 4101 |
+
"loss": 0.467,
|
| 4102 |
+
"step": 571
|
| 4103 |
+
},
|
| 4104 |
+
{
|
| 4105 |
+
"epoch": 0.06407168860263232,
|
| 4106 |
+
"grad_norm": 0.14597094058990479,
|
| 4107 |
+
"learning_rate": 1.6640578486200373e-06,
|
| 4108 |
+
"loss": 0.5426,
|
| 4109 |
+
"step": 572
|
| 4110 |
+
},
|
| 4111 |
+
{
|
| 4112 |
+
"epoch": 0.06418370204424531,
|
| 4113 |
+
"grad_norm": 0.14672952890396118,
|
| 4114 |
+
"learning_rate": 1.5475202560014054e-06,
|
| 4115 |
+
"loss": 0.623,
|
| 4116 |
+
"step": 573
|
| 4117 |
+
},
|
| 4118 |
+
{
|
| 4119 |
+
"epoch": 0.0642957154858583,
|
| 4120 |
+
"grad_norm": 0.1516023725271225,
|
| 4121 |
+
"learning_rate": 1.435191694654375e-06,
|
| 4122 |
+
"loss": 0.6358,
|
| 4123 |
+
"step": 574
|
| 4124 |
+
},
|
| 4125 |
+
{
|
| 4126 |
+
"epoch": 0.06440772892747129,
|
| 4127 |
+
"grad_norm": 0.13252241909503937,
|
| 4128 |
+
"learning_rate": 1.3270753493989373e-06,
|
| 4129 |
+
"loss": 0.4679,
|
| 4130 |
+
"step": 575
|
| 4131 |
+
},
|
| 4132 |
+
{
|
| 4133 |
+
"epoch": 0.0645197423690843,
|
| 4134 |
+
"grad_norm": 0.14411140978336334,
|
| 4135 |
+
"learning_rate": 1.2231742856273151e-06,
|
| 4136 |
+
"loss": 0.5655,
|
| 4137 |
+
"step": 576
|
| 4138 |
+
},
|
| 4139 |
+
{
|
| 4140 |
+
"epoch": 0.06463175581069729,
|
| 4141 |
+
"grad_norm": 0.15002818405628204,
|
| 4142 |
+
"learning_rate": 1.1234914492170678e-06,
|
| 4143 |
+
"loss": 0.5183,
|
| 4144 |
+
"step": 577
|
| 4145 |
+
},
|
| 4146 |
+
{
|
| 4147 |
+
"epoch": 0.06474376925231028,
|
| 4148 |
+
"grad_norm": 0.15125377476215363,
|
| 4149 |
+
"learning_rate": 1.0280296664475218e-06,
|
| 4150 |
+
"loss": 0.5556,
|
| 4151 |
+
"step": 578
|
| 4152 |
+
},
|
| 4153 |
+
{
|
| 4154 |
+
"epoch": 0.06485578269392327,
|
| 4155 |
+
"grad_norm": 0.153385192155838,
|
| 4156 |
+
"learning_rate": 9.367916439196709e-07,
|
| 4157 |
+
"loss": 0.4701,
|
| 4158 |
+
"step": 579
|
| 4159 |
+
},
|
| 4160 |
+
{
|
| 4161 |
+
"epoch": 0.06496779613553627,
|
| 4162 |
+
"grad_norm": 0.1459246724843979,
|
| 4163 |
+
"learning_rate": 8.49779968479436e-07,
|
| 4164 |
+
"loss": 0.5861,
|
| 4165 |
+
"step": 580
|
| 4166 |
+
},
|
| 4167 |
+
{
|
| 4168 |
+
"epoch": 0.06507980957714926,
|
| 4169 |
+
"grad_norm": 0.14465294778347015,
|
| 4170 |
+
"learning_rate": 7.669971071442738e-07,
|
| 4171 |
+
"loss": 0.581,
|
| 4172 |
+
"step": 581
|
| 4173 |
+
},
|
| 4174 |
+
{
|
| 4175 |
+
"epoch": 0.06519182301876225,
|
| 4176 |
+
"grad_norm": 0.14619335532188416,
|
| 4177 |
+
"learning_rate": 6.884454070333168e-07,
|
| 4178 |
+
"loss": 0.5257,
|
| 4179 |
+
"step": 582
|
| 4180 |
+
},
|
| 4181 |
+
{
|
| 4182 |
+
"epoch": 0.06530383646037524,
|
| 4183 |
+
"grad_norm": 0.13438697159290314,
|
| 4184 |
+
"learning_rate": 6.141270953007593e-07,
|
| 4185 |
+
"loss": 0.4356,
|
| 4186 |
+
"step": 583
|
| 4187 |
+
},
|
| 4188 |
+
{
|
| 4189 |
+
"epoch": 0.06541584990198823,
|
| 4190 |
+
"grad_norm": 0.1581222116947174,
|
| 4191 |
+
"learning_rate": 5.440442790727085e-07,
|
| 4192 |
+
"loss": 0.5663,
|
| 4193 |
+
"step": 584
|
| 4194 |
+
},
|
| 4195 |
+
{
|
| 4196 |
+
"epoch": 0.06552786334360124,
|
| 4197 |
+
"grad_norm": 0.15837247669696808,
|
| 4198 |
+
"learning_rate": 4.781989453874813e-07,
|
| 4199 |
+
"loss": 0.5957,
|
| 4200 |
+
"step": 585
|
| 4201 |
+
},
|
| 4202 |
+
{
|
| 4203 |
+
"epoch": 0.06563987678521423,
|
| 4204 |
+
"grad_norm": 0.14343412220478058,
|
| 4205 |
+
"learning_rate": 4.1659296113925046e-07,
|
| 4206 |
+
"loss": 0.4383,
|
| 4207 |
+
"step": 586
|
| 4208 |
+
},
|
| 4209 |
+
{
|
| 4210 |
+
"epoch": 0.06575189022682722,
|
| 4211 |
+
"grad_norm": 0.15165679156780243,
|
| 4212 |
+
"learning_rate": 3.592280730250863e-07,
|
| 4213 |
+
"loss": 0.5884,
|
| 4214 |
+
"step": 587
|
| 4215 |
+
},
|
| 4216 |
+
{
|
| 4217 |
+
"epoch": 0.06586390366844021,
|
| 4218 |
+
"grad_norm": 0.1490369737148285,
|
| 4219 |
+
"learning_rate": 3.06105907495513e-07,
|
| 4220 |
+
"loss": 0.4845,
|
| 4221 |
+
"step": 588
|
| 4222 |
+
},
|
| 4223 |
+
{
|
| 4224 |
+
"epoch": 0.06597591711005321,
|
| 4225 |
+
"grad_norm": 0.15211187303066254,
|
| 4226 |
+
"learning_rate": 2.572279707082625e-07,
|
| 4227 |
+
"loss": 0.5914,
|
| 4228 |
+
"step": 589
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 0.0660879305516662,
|
| 4232 |
+
"grad_norm": 0.1439492255449295,
|
| 4233 |
+
"learning_rate": 2.125956484857083e-07,
|
| 4234 |
+
"loss": 0.506,
|
| 4235 |
+
"step": 590
|
| 4236 |
+
},
|
| 4237 |
+
{
|
| 4238 |
+
"epoch": 0.06619994399327919,
|
| 4239 |
+
"grad_norm": 0.15834854543209076,
|
| 4240 |
+
"learning_rate": 1.7221020627548043e-07,
|
| 4241 |
+
"loss": 0.6013,
|
| 4242 |
+
"step": 591
|
| 4243 |
+
},
|
| 4244 |
+
{
|
| 4245 |
+
"epoch": 0.06631195743489218,
|
| 4246 |
+
"grad_norm": 0.15449443459510803,
|
| 4247 |
+
"learning_rate": 1.3607278911462739e-07,
|
| 4248 |
+
"loss": 0.4803,
|
| 4249 |
+
"step": 592
|
| 4250 |
+
},
|
| 4251 |
+
{
|
| 4252 |
+
"epoch": 0.06642397087650519,
|
| 4253 |
+
"grad_norm": 0.16052158176898956,
|
| 4254 |
+
"learning_rate": 1.0418442159715879e-07,
|
| 4255 |
+
"loss": 0.4994,
|
| 4256 |
+
"step": 593
|
| 4257 |
+
},
|
| 4258 |
+
{
|
| 4259 |
+
"epoch": 0.06653598431811818,
|
| 4260 |
+
"grad_norm": 0.15226219594478607,
|
| 4261 |
+
"learning_rate": 7.654600784493536e-08,
|
| 4262 |
+
"loss": 0.4476,
|
| 4263 |
+
"step": 594
|
| 4264 |
+
},
|
| 4265 |
+
{
|
| 4266 |
+
"epoch": 0.06664799775973117,
|
| 4267 |
+
"grad_norm": 0.17426903545856476,
|
| 4268 |
+
"learning_rate": 5.315833148210602e-08,
|
| 4269 |
+
"loss": 0.5283,
|
| 4270 |
+
"step": 595
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"epoch": 0.06676001120134416,
|
| 4274 |
+
"grad_norm": 0.16931919753551483,
|
| 4275 |
+
"learning_rate": 3.402205561289229e-08,
|
| 4276 |
+
"loss": 0.5116,
|
| 4277 |
+
"step": 596
|
| 4278 |
+
},
|
| 4279 |
+
{
|
| 4280 |
+
"epoch": 0.06687202464295715,
|
| 4281 |
+
"grad_norm": 0.1650620847940445,
|
| 4282 |
+
"learning_rate": 1.9137722802686772e-08,
|
| 4283 |
+
"loss": 0.4657,
|
| 4284 |
+
"step": 597
|
| 4285 |
+
},
|
| 4286 |
+
{
|
| 4287 |
+
"epoch": 0.06698403808457015,
|
| 4288 |
+
"grad_norm": 0.1828121691942215,
|
| 4289 |
+
"learning_rate": 8.50575506278206e-09,
|
| 4290 |
+
"loss": 0.4424,
|
| 4291 |
+
"step": 598
|
| 4292 |
+
},
|
| 4293 |
+
{
|
| 4294 |
+
"epoch": 0.06709605152618314,
|
| 4295 |
+
"grad_norm": 0.19233034551143646,
|
| 4296 |
+
"learning_rate": 2.126453838380282e-09,
|
| 4297 |
+
"loss": 0.5529,
|
| 4298 |
+
"step": 599
|
| 4299 |
+
},
|
| 4300 |
+
{
|
| 4301 |
+
"epoch": 0.06720806496779613,
|
| 4302 |
+
"grad_norm": 0.24840380251407623,
|
| 4303 |
+
"learning_rate": 0.0,
|
| 4304 |
+
"loss": 0.5998,
|
| 4305 |
+
"step": 600
|
| 4306 |
+
},
|
| 4307 |
+
{
|
| 4308 |
+
"epoch": 0.06720806496779613,
|
| 4309 |
+
"eval_loss": 0.5509313941001892,
|
| 4310 |
+
"eval_runtime": 30.0076,
|
| 4311 |
+
"eval_samples_per_second": 9.564,
|
| 4312 |
+
"eval_steps_per_second": 9.564,
|
| 4313 |
+
"step": 600
|
| 4314 |
}
|
| 4315 |
],
|
| 4316 |
"logging_steps": 1,
|
|
|
|
| 4334 |
"should_evaluate": false,
|
| 4335 |
"should_log": false,
|
| 4336 |
"should_save": true,
|
| 4337 |
+
"should_training_stop": true
|
| 4338 |
},
|
| 4339 |
"attributes": {}
|
| 4340 |
}
|
| 4341 |
},
|
| 4342 |
+
"total_flos": 5.535341585576755e+16,
|
| 4343 |
"train_batch_size": 1,
|
| 4344 |
"trial_name": null,
|
| 4345 |
"trial_params": null
|