Training in progress, step 3500, checkpoint
Browse files- last-checkpoint/model.safetensors +1 -1
- last-checkpoint/optimizer.pt +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 +711 -3
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 791869518
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:004c78f3624f93420c8bcc03a3ca9b8a37f2a3690b0727d9baad61326ff4924a
|
| 3 |
size 791869518
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2375752250
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f52c0af4dcd803a82e086998f8d31a06630a29f9e5698a02cb802f2b5768b16e
|
| 3 |
size 2375752250
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04cb5208648fd09a2e0403d51973f74ffbfd93cbd5da59e1e99c8df03769a86c
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7034685b36b93a4dd3a50697b0b1c314b249b2189ec2cb96b757312b1514a579
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e851fe1c1de0057f4eecefed6a131fa9021334eb43f6e7e65fdb270a25ac864
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:978379030048e432baa510ec4fc9514faa08fe564ab964b3a4d05e8f60306495
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bdbc75d90af112615b53d15931e8157a80e37bcd110aac9a3089f5f6f5344171
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8c8a310f6ca2ca89570eb2cc68544656b30224f00b2d6d96eeda6e0cb8be50ab
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c5b8110fcf6e044b6860c6305be969cfe03129549b92dc6fc2394448e9265d6
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f936acaf5a2d5fe8c38d945450417facbf1577584c216908a396d3cc20bec88
|
| 3 |
size 15984
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47f6ae61b86ec1e2b5a6767419416b8803731fef933343a6831b194cd48a7616
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -4255,6 +4255,714 @@
|
|
| 4255 |
"eval_samples_per_second": 611.953,
|
| 4256 |
"eval_steps_per_second": 38.247,
|
| 4257 |
"step": 3000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4258 |
}
|
| 4259 |
],
|
| 4260 |
"logging_steps": 5,
|
|
@@ -4274,7 +4982,7 @@
|
|
| 4274 |
"attributes": {}
|
| 4275 |
}
|
| 4276 |
},
|
| 4277 |
-
"total_flos": 1.
|
| 4278 |
"train_batch_size": 4,
|
| 4279 |
"trial_name": null,
|
| 4280 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8091140923112489,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 3500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4255 |
"eval_samples_per_second": 611.953,
|
| 4256 |
"eval_steps_per_second": 38.247,
|
| 4257 |
"step": 3000
|
| 4258 |
+
},
|
| 4259 |
+
{
|
| 4260 |
+
"epoch": 0.6946822421129436,
|
| 4261 |
+
"grad_norm": 171.0,
|
| 4262 |
+
"learning_rate": 3.391572456320658e-06,
|
| 4263 |
+
"loss": 69.3591,
|
| 4264 |
+
"step": 3005
|
| 4265 |
+
},
|
| 4266 |
+
{
|
| 4267 |
+
"epoch": 0.695838119387674,
|
| 4268 |
+
"grad_norm": 178.875,
|
| 4269 |
+
"learning_rate": 3.378725590955807e-06,
|
| 4270 |
+
"loss": 69.3434,
|
| 4271 |
+
"step": 3010
|
| 4272 |
+
},
|
| 4273 |
+
{
|
| 4274 |
+
"epoch": 0.6969939966624044,
|
| 4275 |
+
"grad_norm": 188.375,
|
| 4276 |
+
"learning_rate": 3.365878725590956e-06,
|
| 4277 |
+
"loss": 68.8737,
|
| 4278 |
+
"step": 3015
|
| 4279 |
+
},
|
| 4280 |
+
{
|
| 4281 |
+
"epoch": 0.6981498739371347,
|
| 4282 |
+
"grad_norm": 169.875,
|
| 4283 |
+
"learning_rate": 3.353031860226105e-06,
|
| 4284 |
+
"loss": 68.2949,
|
| 4285 |
+
"step": 3020
|
| 4286 |
+
},
|
| 4287 |
+
{
|
| 4288 |
+
"epoch": 0.6993057512118651,
|
| 4289 |
+
"grad_norm": 154.25,
|
| 4290 |
+
"learning_rate": 3.340184994861254e-06,
|
| 4291 |
+
"loss": 67.784,
|
| 4292 |
+
"step": 3025
|
| 4293 |
+
},
|
| 4294 |
+
{
|
| 4295 |
+
"epoch": 0.7004616284865954,
|
| 4296 |
+
"grad_norm": 159.25,
|
| 4297 |
+
"learning_rate": 3.327338129496403e-06,
|
| 4298 |
+
"loss": 68.2709,
|
| 4299 |
+
"step": 3030
|
| 4300 |
+
},
|
| 4301 |
+
{
|
| 4302 |
+
"epoch": 0.7016175057613258,
|
| 4303 |
+
"grad_norm": 148.75,
|
| 4304 |
+
"learning_rate": 3.3144912641315525e-06,
|
| 4305 |
+
"loss": 69.2274,
|
| 4306 |
+
"step": 3035
|
| 4307 |
+
},
|
| 4308 |
+
{
|
| 4309 |
+
"epoch": 0.7027733830360562,
|
| 4310 |
+
"grad_norm": 157.375,
|
| 4311 |
+
"learning_rate": 3.3016443987667014e-06,
|
| 4312 |
+
"loss": 67.7498,
|
| 4313 |
+
"step": 3040
|
| 4314 |
+
},
|
| 4315 |
+
{
|
| 4316 |
+
"epoch": 0.7039292603107865,
|
| 4317 |
+
"grad_norm": 153.0,
|
| 4318 |
+
"learning_rate": 3.2887975334018502e-06,
|
| 4319 |
+
"loss": 67.9523,
|
| 4320 |
+
"step": 3045
|
| 4321 |
+
},
|
| 4322 |
+
{
|
| 4323 |
+
"epoch": 0.7050851375855168,
|
| 4324 |
+
"grad_norm": 162.375,
|
| 4325 |
+
"learning_rate": 3.275950668036999e-06,
|
| 4326 |
+
"loss": 67.1861,
|
| 4327 |
+
"step": 3050
|
| 4328 |
+
},
|
| 4329 |
+
{
|
| 4330 |
+
"epoch": 0.7062410148602473,
|
| 4331 |
+
"grad_norm": 154.5,
|
| 4332 |
+
"learning_rate": 3.2631038026721483e-06,
|
| 4333 |
+
"loss": 68.3158,
|
| 4334 |
+
"step": 3055
|
| 4335 |
+
},
|
| 4336 |
+
{
|
| 4337 |
+
"epoch": 0.7073968921349776,
|
| 4338 |
+
"grad_norm": 145.875,
|
| 4339 |
+
"learning_rate": 3.250256937307297e-06,
|
| 4340 |
+
"loss": 68.1404,
|
| 4341 |
+
"step": 3060
|
| 4342 |
+
},
|
| 4343 |
+
{
|
| 4344 |
+
"epoch": 0.7085527694097079,
|
| 4345 |
+
"grad_norm": 163.375,
|
| 4346 |
+
"learning_rate": 3.237410071942446e-06,
|
| 4347 |
+
"loss": 68.4577,
|
| 4348 |
+
"step": 3065
|
| 4349 |
+
},
|
| 4350 |
+
{
|
| 4351 |
+
"epoch": 0.7097086466844382,
|
| 4352 |
+
"grad_norm": 167.0,
|
| 4353 |
+
"learning_rate": 3.2245632065775957e-06,
|
| 4354 |
+
"loss": 67.7858,
|
| 4355 |
+
"step": 3070
|
| 4356 |
+
},
|
| 4357 |
+
{
|
| 4358 |
+
"epoch": 0.7108645239591687,
|
| 4359 |
+
"grad_norm": 148.25,
|
| 4360 |
+
"learning_rate": 3.2117163412127446e-06,
|
| 4361 |
+
"loss": 67.727,
|
| 4362 |
+
"step": 3075
|
| 4363 |
+
},
|
| 4364 |
+
{
|
| 4365 |
+
"epoch": 0.712020401233899,
|
| 4366 |
+
"grad_norm": 151.375,
|
| 4367 |
+
"learning_rate": 3.1988694758478934e-06,
|
| 4368 |
+
"loss": 68.0403,
|
| 4369 |
+
"step": 3080
|
| 4370 |
+
},
|
| 4371 |
+
{
|
| 4372 |
+
"epoch": 0.7131762785086293,
|
| 4373 |
+
"grad_norm": 158.25,
|
| 4374 |
+
"learning_rate": 3.1860226104830423e-06,
|
| 4375 |
+
"loss": 68.9341,
|
| 4376 |
+
"step": 3085
|
| 4377 |
+
},
|
| 4378 |
+
{
|
| 4379 |
+
"epoch": 0.7143321557833597,
|
| 4380 |
+
"grad_norm": 159.875,
|
| 4381 |
+
"learning_rate": 3.1731757451181915e-06,
|
| 4382 |
+
"loss": 68.1696,
|
| 4383 |
+
"step": 3090
|
| 4384 |
+
},
|
| 4385 |
+
{
|
| 4386 |
+
"epoch": 0.7154880330580901,
|
| 4387 |
+
"grad_norm": 172.0,
|
| 4388 |
+
"learning_rate": 3.1603288797533404e-06,
|
| 4389 |
+
"loss": 67.7838,
|
| 4390 |
+
"step": 3095
|
| 4391 |
+
},
|
| 4392 |
+
{
|
| 4393 |
+
"epoch": 0.7166439103328204,
|
| 4394 |
+
"grad_norm": 166.875,
|
| 4395 |
+
"learning_rate": 3.1474820143884892e-06,
|
| 4396 |
+
"loss": 68.0554,
|
| 4397 |
+
"step": 3100
|
| 4398 |
+
},
|
| 4399 |
+
{
|
| 4400 |
+
"epoch": 0.7177997876075508,
|
| 4401 |
+
"grad_norm": 150.375,
|
| 4402 |
+
"learning_rate": 3.134635149023639e-06,
|
| 4403 |
+
"loss": 68.3312,
|
| 4404 |
+
"step": 3105
|
| 4405 |
+
},
|
| 4406 |
+
{
|
| 4407 |
+
"epoch": 0.7189556648822811,
|
| 4408 |
+
"grad_norm": 154.375,
|
| 4409 |
+
"learning_rate": 3.1217882836587878e-06,
|
| 4410 |
+
"loss": 68.1626,
|
| 4411 |
+
"step": 3110
|
| 4412 |
+
},
|
| 4413 |
+
{
|
| 4414 |
+
"epoch": 0.7201115421570115,
|
| 4415 |
+
"grad_norm": 175.125,
|
| 4416 |
+
"learning_rate": 3.1089414182939366e-06,
|
| 4417 |
+
"loss": 68.838,
|
| 4418 |
+
"step": 3115
|
| 4419 |
+
},
|
| 4420 |
+
{
|
| 4421 |
+
"epoch": 0.7212674194317419,
|
| 4422 |
+
"grad_norm": 162.125,
|
| 4423 |
+
"learning_rate": 3.0960945529290855e-06,
|
| 4424 |
+
"loss": 68.3526,
|
| 4425 |
+
"step": 3120
|
| 4426 |
+
},
|
| 4427 |
+
{
|
| 4428 |
+
"epoch": 0.7224232967064722,
|
| 4429 |
+
"grad_norm": 163.625,
|
| 4430 |
+
"learning_rate": 3.0832476875642348e-06,
|
| 4431 |
+
"loss": 67.3081,
|
| 4432 |
+
"step": 3125
|
| 4433 |
+
},
|
| 4434 |
+
{
|
| 4435 |
+
"epoch": 0.7235791739812025,
|
| 4436 |
+
"grad_norm": 150.5,
|
| 4437 |
+
"learning_rate": 3.0704008221993836e-06,
|
| 4438 |
+
"loss": 67.0766,
|
| 4439 |
+
"step": 3130
|
| 4440 |
+
},
|
| 4441 |
+
{
|
| 4442 |
+
"epoch": 0.7247350512559328,
|
| 4443 |
+
"grad_norm": 159.25,
|
| 4444 |
+
"learning_rate": 3.0575539568345324e-06,
|
| 4445 |
+
"loss": 68.1058,
|
| 4446 |
+
"step": 3135
|
| 4447 |
+
},
|
| 4448 |
+
{
|
| 4449 |
+
"epoch": 0.7258909285306633,
|
| 4450 |
+
"grad_norm": 156.625,
|
| 4451 |
+
"learning_rate": 3.0447070914696817e-06,
|
| 4452 |
+
"loss": 68.7516,
|
| 4453 |
+
"step": 3140
|
| 4454 |
+
},
|
| 4455 |
+
{
|
| 4456 |
+
"epoch": 0.7270468058053936,
|
| 4457 |
+
"grad_norm": 154.875,
|
| 4458 |
+
"learning_rate": 3.0318602261048306e-06,
|
| 4459 |
+
"loss": 69.4498,
|
| 4460 |
+
"step": 3145
|
| 4461 |
+
},
|
| 4462 |
+
{
|
| 4463 |
+
"epoch": 0.7282026830801239,
|
| 4464 |
+
"grad_norm": 146.5,
|
| 4465 |
+
"learning_rate": 3.01901336073998e-06,
|
| 4466 |
+
"loss": 68.6958,
|
| 4467 |
+
"step": 3150
|
| 4468 |
+
},
|
| 4469 |
+
{
|
| 4470 |
+
"epoch": 0.7293585603548544,
|
| 4471 |
+
"grad_norm": 158.75,
|
| 4472 |
+
"learning_rate": 3.0061664953751287e-06,
|
| 4473 |
+
"loss": 67.5718,
|
| 4474 |
+
"step": 3155
|
| 4475 |
+
},
|
| 4476 |
+
{
|
| 4477 |
+
"epoch": 0.7305144376295847,
|
| 4478 |
+
"grad_norm": 159.0,
|
| 4479 |
+
"learning_rate": 2.993319630010278e-06,
|
| 4480 |
+
"loss": 67.7357,
|
| 4481 |
+
"step": 3160
|
| 4482 |
+
},
|
| 4483 |
+
{
|
| 4484 |
+
"epoch": 0.731670314904315,
|
| 4485 |
+
"grad_norm": 164.625,
|
| 4486 |
+
"learning_rate": 2.980472764645427e-06,
|
| 4487 |
+
"loss": 68.8242,
|
| 4488 |
+
"step": 3165
|
| 4489 |
+
},
|
| 4490 |
+
{
|
| 4491 |
+
"epoch": 0.7328261921790454,
|
| 4492 |
+
"grad_norm": 162.875,
|
| 4493 |
+
"learning_rate": 2.9676258992805756e-06,
|
| 4494 |
+
"loss": 67.6763,
|
| 4495 |
+
"step": 3170
|
| 4496 |
+
},
|
| 4497 |
+
{
|
| 4498 |
+
"epoch": 0.7339820694537758,
|
| 4499 |
+
"grad_norm": 152.625,
|
| 4500 |
+
"learning_rate": 2.954779033915725e-06,
|
| 4501 |
+
"loss": 67.596,
|
| 4502 |
+
"step": 3175
|
| 4503 |
+
},
|
| 4504 |
+
{
|
| 4505 |
+
"epoch": 0.7351379467285061,
|
| 4506 |
+
"grad_norm": 160.125,
|
| 4507 |
+
"learning_rate": 2.9419321685508738e-06,
|
| 4508 |
+
"loss": 67.8334,
|
| 4509 |
+
"step": 3180
|
| 4510 |
+
},
|
| 4511 |
+
{
|
| 4512 |
+
"epoch": 0.7362938240032365,
|
| 4513 |
+
"grad_norm": 176.875,
|
| 4514 |
+
"learning_rate": 2.9290853031860226e-06,
|
| 4515 |
+
"loss": 68.2346,
|
| 4516 |
+
"step": 3185
|
| 4517 |
+
},
|
| 4518 |
+
{
|
| 4519 |
+
"epoch": 0.7374497012779668,
|
| 4520 |
+
"grad_norm": 158.625,
|
| 4521 |
+
"learning_rate": 2.9162384378211715e-06,
|
| 4522 |
+
"loss": 67.566,
|
| 4523 |
+
"step": 3190
|
| 4524 |
+
},
|
| 4525 |
+
{
|
| 4526 |
+
"epoch": 0.7386055785526972,
|
| 4527 |
+
"grad_norm": 171.625,
|
| 4528 |
+
"learning_rate": 2.903391572456321e-06,
|
| 4529 |
+
"loss": 67.5561,
|
| 4530 |
+
"step": 3195
|
| 4531 |
+
},
|
| 4532 |
+
{
|
| 4533 |
+
"epoch": 0.7397614558274275,
|
| 4534 |
+
"grad_norm": 172.375,
|
| 4535 |
+
"learning_rate": 2.89054470709147e-06,
|
| 4536 |
+
"loss": 66.9553,
|
| 4537 |
+
"step": 3200
|
| 4538 |
+
},
|
| 4539 |
+
{
|
| 4540 |
+
"epoch": 0.7409173331021579,
|
| 4541 |
+
"grad_norm": 158.875,
|
| 4542 |
+
"learning_rate": 2.877697841726619e-06,
|
| 4543 |
+
"loss": 67.6197,
|
| 4544 |
+
"step": 3205
|
| 4545 |
+
},
|
| 4546 |
+
{
|
| 4547 |
+
"epoch": 0.7420732103768882,
|
| 4548 |
+
"grad_norm": 164.625,
|
| 4549 |
+
"learning_rate": 2.864850976361768e-06,
|
| 4550 |
+
"loss": 67.4917,
|
| 4551 |
+
"step": 3210
|
| 4552 |
+
},
|
| 4553 |
+
{
|
| 4554 |
+
"epoch": 0.7432290876516185,
|
| 4555 |
+
"grad_norm": 171.5,
|
| 4556 |
+
"learning_rate": 2.852004110996917e-06,
|
| 4557 |
+
"loss": 68.1053,
|
| 4558 |
+
"step": 3215
|
| 4559 |
+
},
|
| 4560 |
+
{
|
| 4561 |
+
"epoch": 0.744384964926349,
|
| 4562 |
+
"grad_norm": 169.5,
|
| 4563 |
+
"learning_rate": 2.839157245632066e-06,
|
| 4564 |
+
"loss": 67.5385,
|
| 4565 |
+
"step": 3220
|
| 4566 |
+
},
|
| 4567 |
+
{
|
| 4568 |
+
"epoch": 0.7455408422010793,
|
| 4569 |
+
"grad_norm": 163.0,
|
| 4570 |
+
"learning_rate": 2.8263103802672147e-06,
|
| 4571 |
+
"loss": 67.5871,
|
| 4572 |
+
"step": 3225
|
| 4573 |
+
},
|
| 4574 |
+
{
|
| 4575 |
+
"epoch": 0.7466967194758096,
|
| 4576 |
+
"grad_norm": 155.0,
|
| 4577 |
+
"learning_rate": 2.8134635149023644e-06,
|
| 4578 |
+
"loss": 67.678,
|
| 4579 |
+
"step": 3230
|
| 4580 |
+
},
|
| 4581 |
+
{
|
| 4582 |
+
"epoch": 0.74785259675054,
|
| 4583 |
+
"grad_norm": 171.5,
|
| 4584 |
+
"learning_rate": 2.800616649537513e-06,
|
| 4585 |
+
"loss": 68.6263,
|
| 4586 |
+
"step": 3235
|
| 4587 |
+
},
|
| 4588 |
+
{
|
| 4589 |
+
"epoch": 0.7490084740252704,
|
| 4590 |
+
"grad_norm": 161.625,
|
| 4591 |
+
"learning_rate": 2.787769784172662e-06,
|
| 4592 |
+
"loss": 66.3458,
|
| 4593 |
+
"step": 3240
|
| 4594 |
+
},
|
| 4595 |
+
{
|
| 4596 |
+
"epoch": 0.7501643513000007,
|
| 4597 |
+
"grad_norm": 154.0,
|
| 4598 |
+
"learning_rate": 2.7749229188078113e-06,
|
| 4599 |
+
"loss": 68.3403,
|
| 4600 |
+
"step": 3245
|
| 4601 |
+
},
|
| 4602 |
+
{
|
| 4603 |
+
"epoch": 0.7513202285747311,
|
| 4604 |
+
"grad_norm": 151.25,
|
| 4605 |
+
"learning_rate": 2.76207605344296e-06,
|
| 4606 |
+
"loss": 67.5946,
|
| 4607 |
+
"step": 3250
|
| 4608 |
+
},
|
| 4609 |
+
{
|
| 4610 |
+
"epoch": 0.7524761058494615,
|
| 4611 |
+
"grad_norm": 159.875,
|
| 4612 |
+
"learning_rate": 2.749229188078109e-06,
|
| 4613 |
+
"loss": 67.0883,
|
| 4614 |
+
"step": 3255
|
| 4615 |
+
},
|
| 4616 |
+
{
|
| 4617 |
+
"epoch": 0.7536319831241918,
|
| 4618 |
+
"grad_norm": 153.0,
|
| 4619 |
+
"learning_rate": 2.736382322713258e-06,
|
| 4620 |
+
"loss": 68.4621,
|
| 4621 |
+
"step": 3260
|
| 4622 |
+
},
|
| 4623 |
+
{
|
| 4624 |
+
"epoch": 0.7547878603989221,
|
| 4625 |
+
"grad_norm": 160.625,
|
| 4626 |
+
"learning_rate": 2.7235354573484076e-06,
|
| 4627 |
+
"loss": 67.5966,
|
| 4628 |
+
"step": 3265
|
| 4629 |
+
},
|
| 4630 |
+
{
|
| 4631 |
+
"epoch": 0.7559437376736525,
|
| 4632 |
+
"grad_norm": 160.0,
|
| 4633 |
+
"learning_rate": 2.7106885919835564e-06,
|
| 4634 |
+
"loss": 69.2813,
|
| 4635 |
+
"step": 3270
|
| 4636 |
+
},
|
| 4637 |
+
{
|
| 4638 |
+
"epoch": 0.7570996149483828,
|
| 4639 |
+
"grad_norm": 156.75,
|
| 4640 |
+
"learning_rate": 2.6978417266187052e-06,
|
| 4641 |
+
"loss": 67.1882,
|
| 4642 |
+
"step": 3275
|
| 4643 |
+
},
|
| 4644 |
+
{
|
| 4645 |
+
"epoch": 0.7582554922231132,
|
| 4646 |
+
"grad_norm": 162.75,
|
| 4647 |
+
"learning_rate": 2.6849948612538545e-06,
|
| 4648 |
+
"loss": 67.7005,
|
| 4649 |
+
"step": 3280
|
| 4650 |
+
},
|
| 4651 |
+
{
|
| 4652 |
+
"epoch": 0.7594113694978436,
|
| 4653 |
+
"grad_norm": 180.875,
|
| 4654 |
+
"learning_rate": 2.6721479958890034e-06,
|
| 4655 |
+
"loss": 68.8234,
|
| 4656 |
+
"step": 3285
|
| 4657 |
+
},
|
| 4658 |
+
{
|
| 4659 |
+
"epoch": 0.7605672467725739,
|
| 4660 |
+
"grad_norm": 184.5,
|
| 4661 |
+
"learning_rate": 2.6593011305241522e-06,
|
| 4662 |
+
"loss": 68.6867,
|
| 4663 |
+
"step": 3290
|
| 4664 |
+
},
|
| 4665 |
+
{
|
| 4666 |
+
"epoch": 0.7617231240473042,
|
| 4667 |
+
"grad_norm": 157.5,
|
| 4668 |
+
"learning_rate": 2.646454265159301e-06,
|
| 4669 |
+
"loss": 67.4642,
|
| 4670 |
+
"step": 3295
|
| 4671 |
+
},
|
| 4672 |
+
{
|
| 4673 |
+
"epoch": 0.7628790013220347,
|
| 4674 |
+
"grad_norm": 156.125,
|
| 4675 |
+
"learning_rate": 2.6336073997944508e-06,
|
| 4676 |
+
"loss": 67.4047,
|
| 4677 |
+
"step": 3300
|
| 4678 |
+
},
|
| 4679 |
+
{
|
| 4680 |
+
"epoch": 0.764034878596765,
|
| 4681 |
+
"grad_norm": 149.0,
|
| 4682 |
+
"learning_rate": 2.6207605344295996e-06,
|
| 4683 |
+
"loss": 67.1334,
|
| 4684 |
+
"step": 3305
|
| 4685 |
+
},
|
| 4686 |
+
{
|
| 4687 |
+
"epoch": 0.7651907558714953,
|
| 4688 |
+
"grad_norm": 157.5,
|
| 4689 |
+
"learning_rate": 2.6079136690647484e-06,
|
| 4690 |
+
"loss": 66.8244,
|
| 4691 |
+
"step": 3310
|
| 4692 |
+
},
|
| 4693 |
+
{
|
| 4694 |
+
"epoch": 0.7663466331462258,
|
| 4695 |
+
"grad_norm": 160.125,
|
| 4696 |
+
"learning_rate": 2.5950668036998973e-06,
|
| 4697 |
+
"loss": 68.4834,
|
| 4698 |
+
"step": 3315
|
| 4699 |
+
},
|
| 4700 |
+
{
|
| 4701 |
+
"epoch": 0.7675025104209561,
|
| 4702 |
+
"grad_norm": 163.125,
|
| 4703 |
+
"learning_rate": 2.5822199383350466e-06,
|
| 4704 |
+
"loss": 67.9042,
|
| 4705 |
+
"step": 3320
|
| 4706 |
+
},
|
| 4707 |
+
{
|
| 4708 |
+
"epoch": 0.7686583876956864,
|
| 4709 |
+
"grad_norm": 164.25,
|
| 4710 |
+
"learning_rate": 2.5693730729701954e-06,
|
| 4711 |
+
"loss": 67.7928,
|
| 4712 |
+
"step": 3325
|
| 4713 |
+
},
|
| 4714 |
+
{
|
| 4715 |
+
"epoch": 0.7698142649704167,
|
| 4716 |
+
"grad_norm": 161.875,
|
| 4717 |
+
"learning_rate": 2.5565262076053443e-06,
|
| 4718 |
+
"loss": 69.5935,
|
| 4719 |
+
"step": 3330
|
| 4720 |
+
},
|
| 4721 |
+
{
|
| 4722 |
+
"epoch": 0.7709701422451471,
|
| 4723 |
+
"grad_norm": 142.125,
|
| 4724 |
+
"learning_rate": 2.5436793422404935e-06,
|
| 4725 |
+
"loss": 66.6378,
|
| 4726 |
+
"step": 3335
|
| 4727 |
+
},
|
| 4728 |
+
{
|
| 4729 |
+
"epoch": 0.7721260195198775,
|
| 4730 |
+
"grad_norm": 156.125,
|
| 4731 |
+
"learning_rate": 2.530832476875643e-06,
|
| 4732 |
+
"loss": 67.4647,
|
| 4733 |
+
"step": 3340
|
| 4734 |
+
},
|
| 4735 |
+
{
|
| 4736 |
+
"epoch": 0.7732818967946078,
|
| 4737 |
+
"grad_norm": 155.875,
|
| 4738 |
+
"learning_rate": 2.5179856115107916e-06,
|
| 4739 |
+
"loss": 68.0196,
|
| 4740 |
+
"step": 3345
|
| 4741 |
+
},
|
| 4742 |
+
{
|
| 4743 |
+
"epoch": 0.7744377740693382,
|
| 4744 |
+
"grad_norm": 163.125,
|
| 4745 |
+
"learning_rate": 2.5051387461459405e-06,
|
| 4746 |
+
"loss": 66.556,
|
| 4747 |
+
"step": 3350
|
| 4748 |
+
},
|
| 4749 |
+
{
|
| 4750 |
+
"epoch": 0.7755936513440685,
|
| 4751 |
+
"grad_norm": 173.375,
|
| 4752 |
+
"learning_rate": 2.4922918807810898e-06,
|
| 4753 |
+
"loss": 68.6675,
|
| 4754 |
+
"step": 3355
|
| 4755 |
+
},
|
| 4756 |
+
{
|
| 4757 |
+
"epoch": 0.7767495286187989,
|
| 4758 |
+
"grad_norm": 158.25,
|
| 4759 |
+
"learning_rate": 2.4794450154162386e-06,
|
| 4760 |
+
"loss": 66.8891,
|
| 4761 |
+
"step": 3360
|
| 4762 |
+
},
|
| 4763 |
+
{
|
| 4764 |
+
"epoch": 0.7779054058935293,
|
| 4765 |
+
"grad_norm": 152.25,
|
| 4766 |
+
"learning_rate": 2.466598150051388e-06,
|
| 4767 |
+
"loss": 67.2771,
|
| 4768 |
+
"step": 3365
|
| 4769 |
+
},
|
| 4770 |
+
{
|
| 4771 |
+
"epoch": 0.7790612831682596,
|
| 4772 |
+
"grad_norm": 160.125,
|
| 4773 |
+
"learning_rate": 2.4537512846865367e-06,
|
| 4774 |
+
"loss": 66.9648,
|
| 4775 |
+
"step": 3370
|
| 4776 |
+
},
|
| 4777 |
+
{
|
| 4778 |
+
"epoch": 0.7802171604429899,
|
| 4779 |
+
"grad_norm": 164.75,
|
| 4780 |
+
"learning_rate": 2.4409044193216856e-06,
|
| 4781 |
+
"loss": 68.0911,
|
| 4782 |
+
"step": 3375
|
| 4783 |
+
},
|
| 4784 |
+
{
|
| 4785 |
+
"epoch": 0.7813730377177204,
|
| 4786 |
+
"grad_norm": 162.125,
|
| 4787 |
+
"learning_rate": 2.4280575539568344e-06,
|
| 4788 |
+
"loss": 67.2639,
|
| 4789 |
+
"step": 3380
|
| 4790 |
+
},
|
| 4791 |
+
{
|
| 4792 |
+
"epoch": 0.7825289149924507,
|
| 4793 |
+
"grad_norm": 161.375,
|
| 4794 |
+
"learning_rate": 2.4152106885919837e-06,
|
| 4795 |
+
"loss": 66.6423,
|
| 4796 |
+
"step": 3385
|
| 4797 |
+
},
|
| 4798 |
+
{
|
| 4799 |
+
"epoch": 0.783684792267181,
|
| 4800 |
+
"grad_norm": 149.5,
|
| 4801 |
+
"learning_rate": 2.402363823227133e-06,
|
| 4802 |
+
"loss": 67.8673,
|
| 4803 |
+
"step": 3390
|
| 4804 |
+
},
|
| 4805 |
+
{
|
| 4806 |
+
"epoch": 0.7848406695419113,
|
| 4807 |
+
"grad_norm": 151.5,
|
| 4808 |
+
"learning_rate": 2.389516957862282e-06,
|
| 4809 |
+
"loss": 66.1444,
|
| 4810 |
+
"step": 3395
|
| 4811 |
+
},
|
| 4812 |
+
{
|
| 4813 |
+
"epoch": 0.7859965468166418,
|
| 4814 |
+
"grad_norm": 156.625,
|
| 4815 |
+
"learning_rate": 2.376670092497431e-06,
|
| 4816 |
+
"loss": 67.8318,
|
| 4817 |
+
"step": 3400
|
| 4818 |
+
},
|
| 4819 |
+
{
|
| 4820 |
+
"epoch": 0.7871524240913721,
|
| 4821 |
+
"grad_norm": 152.75,
|
| 4822 |
+
"learning_rate": 2.36382322713258e-06,
|
| 4823 |
+
"loss": 68.5987,
|
| 4824 |
+
"step": 3405
|
| 4825 |
+
},
|
| 4826 |
+
{
|
| 4827 |
+
"epoch": 0.7883083013661024,
|
| 4828 |
+
"grad_norm": 163.25,
|
| 4829 |
+
"learning_rate": 2.3509763617677288e-06,
|
| 4830 |
+
"loss": 67.008,
|
| 4831 |
+
"step": 3410
|
| 4832 |
+
},
|
| 4833 |
+
{
|
| 4834 |
+
"epoch": 0.7894641786408328,
|
| 4835 |
+
"grad_norm": 167.375,
|
| 4836 |
+
"learning_rate": 2.3381294964028776e-06,
|
| 4837 |
+
"loss": 67.2252,
|
| 4838 |
+
"step": 3415
|
| 4839 |
+
},
|
| 4840 |
+
{
|
| 4841 |
+
"epoch": 0.7906200559155632,
|
| 4842 |
+
"grad_norm": 157.25,
|
| 4843 |
+
"learning_rate": 2.325282631038027e-06,
|
| 4844 |
+
"loss": 66.6725,
|
| 4845 |
+
"step": 3420
|
| 4846 |
+
},
|
| 4847 |
+
{
|
| 4848 |
+
"epoch": 0.7917759331902935,
|
| 4849 |
+
"grad_norm": 155.875,
|
| 4850 |
+
"learning_rate": 2.312435765673176e-06,
|
| 4851 |
+
"loss": 67.9088,
|
| 4852 |
+
"step": 3425
|
| 4853 |
+
},
|
| 4854 |
+
{
|
| 4855 |
+
"epoch": 0.7929318104650239,
|
| 4856 |
+
"grad_norm": 157.375,
|
| 4857 |
+
"learning_rate": 2.299588900308325e-06,
|
| 4858 |
+
"loss": 66.1468,
|
| 4859 |
+
"step": 3430
|
| 4860 |
+
},
|
| 4861 |
+
{
|
| 4862 |
+
"epoch": 0.7940876877397542,
|
| 4863 |
+
"grad_norm": 152.875,
|
| 4864 |
+
"learning_rate": 2.2867420349434743e-06,
|
| 4865 |
+
"loss": 66.662,
|
| 4866 |
+
"step": 3435
|
| 4867 |
+
},
|
| 4868 |
+
{
|
| 4869 |
+
"epoch": 0.7952435650144846,
|
| 4870 |
+
"grad_norm": 164.0,
|
| 4871 |
+
"learning_rate": 2.273895169578623e-06,
|
| 4872 |
+
"loss": 66.9803,
|
| 4873 |
+
"step": 3440
|
| 4874 |
+
},
|
| 4875 |
+
{
|
| 4876 |
+
"epoch": 0.7963994422892149,
|
| 4877 |
+
"grad_norm": 159.5,
|
| 4878 |
+
"learning_rate": 2.261048304213772e-06,
|
| 4879 |
+
"loss": 67.1713,
|
| 4880 |
+
"step": 3445
|
| 4881 |
+
},
|
| 4882 |
+
{
|
| 4883 |
+
"epoch": 0.7975553195639453,
|
| 4884 |
+
"grad_norm": 162.0,
|
| 4885 |
+
"learning_rate": 2.248201438848921e-06,
|
| 4886 |
+
"loss": 67.7926,
|
| 4887 |
+
"step": 3450
|
| 4888 |
+
},
|
| 4889 |
+
{
|
| 4890 |
+
"epoch": 0.7987111968386756,
|
| 4891 |
+
"grad_norm": 163.875,
|
| 4892 |
+
"learning_rate": 2.23535457348407e-06,
|
| 4893 |
+
"loss": 67.5221,
|
| 4894 |
+
"step": 3455
|
| 4895 |
+
},
|
| 4896 |
+
{
|
| 4897 |
+
"epoch": 0.799867074113406,
|
| 4898 |
+
"grad_norm": 161.75,
|
| 4899 |
+
"learning_rate": 2.222507708119219e-06,
|
| 4900 |
+
"loss": 69.4966,
|
| 4901 |
+
"step": 3460
|
| 4902 |
+
},
|
| 4903 |
+
{
|
| 4904 |
+
"epoch": 0.8010229513881364,
|
| 4905 |
+
"grad_norm": 209.375,
|
| 4906 |
+
"learning_rate": 2.2096608427543682e-06,
|
| 4907 |
+
"loss": 66.1803,
|
| 4908 |
+
"step": 3465
|
| 4909 |
+
},
|
| 4910 |
+
{
|
| 4911 |
+
"epoch": 0.8021788286628667,
|
| 4912 |
+
"grad_norm": 168.0,
|
| 4913 |
+
"learning_rate": 2.196813977389517e-06,
|
| 4914 |
+
"loss": 66.1049,
|
| 4915 |
+
"step": 3470
|
| 4916 |
+
},
|
| 4917 |
+
{
|
| 4918 |
+
"epoch": 0.803334705937597,
|
| 4919 |
+
"grad_norm": 157.75,
|
| 4920 |
+
"learning_rate": 2.183967112024666e-06,
|
| 4921 |
+
"loss": 66.1451,
|
| 4922 |
+
"step": 3475
|
| 4923 |
+
},
|
| 4924 |
+
{
|
| 4925 |
+
"epoch": 0.8044905832123275,
|
| 4926 |
+
"grad_norm": 150.75,
|
| 4927 |
+
"learning_rate": 2.171120246659815e-06,
|
| 4928 |
+
"loss": 66.0184,
|
| 4929 |
+
"step": 3480
|
| 4930 |
+
},
|
| 4931 |
+
{
|
| 4932 |
+
"epoch": 0.8056464604870578,
|
| 4933 |
+
"grad_norm": 151.375,
|
| 4934 |
+
"learning_rate": 2.158273381294964e-06,
|
| 4935 |
+
"loss": 67.9845,
|
| 4936 |
+
"step": 3485
|
| 4937 |
+
},
|
| 4938 |
+
{
|
| 4939 |
+
"epoch": 0.8068023377617881,
|
| 4940 |
+
"grad_norm": 160.25,
|
| 4941 |
+
"learning_rate": 2.1454265159301133e-06,
|
| 4942 |
+
"loss": 66.5661,
|
| 4943 |
+
"step": 3490
|
| 4944 |
+
},
|
| 4945 |
+
{
|
| 4946 |
+
"epoch": 0.8079582150365185,
|
| 4947 |
+
"grad_norm": 153.875,
|
| 4948 |
+
"learning_rate": 2.132579650565262e-06,
|
| 4949 |
+
"loss": 67.4468,
|
| 4950 |
+
"step": 3495
|
| 4951 |
+
},
|
| 4952 |
+
{
|
| 4953 |
+
"epoch": 0.8091140923112489,
|
| 4954 |
+
"grad_norm": 150.25,
|
| 4955 |
+
"learning_rate": 2.1197327852004114e-06,
|
| 4956 |
+
"loss": 66.7717,
|
| 4957 |
+
"step": 3500
|
| 4958 |
+
},
|
| 4959 |
+
{
|
| 4960 |
+
"epoch": 0.8091140923112489,
|
| 4961 |
+
"eval_loss": NaN,
|
| 4962 |
+
"eval_runtime": 383.0157,
|
| 4963 |
+
"eval_samples_per_second": 608.68,
|
| 4964 |
+
"eval_steps_per_second": 38.043,
|
| 4965 |
+
"step": 3500
|
| 4966 |
}
|
| 4967 |
],
|
| 4968 |
"logging_steps": 5,
|
|
|
|
| 4982 |
"attributes": {}
|
| 4983 |
}
|
| 4984 |
},
|
| 4985 |
+
"total_flos": 1.5163252974760755e+19,
|
| 4986 |
"train_batch_size": 4,
|
| 4987 |
"trial_name": null,
|
| 4988 |
"trial_params": null
|