Training in progress, step 2830, checkpoint
Browse files
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1657
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd2d3ffcc027c2ad48f36bda8b2820fe104b89364737fb73e2a49d58292b9df6
|
| 3 |
size 1657
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14709
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e00fbcc945f009eee7357b39c2b1fc8b1fb079bbaa9411fdac23c9af4152b994
|
| 3 |
size 14709
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d9ee0f147a24c5e2edfa2f4d61bcccc832a44ec80acb70d6f9ad92262f4b235
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 21.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -5084,6 +5084,24 @@
|
|
| 5084 |
"mean_token_accuracy": 0.6562414228916168,
|
| 5085 |
"num_tokens": 10284706.0,
|
| 5086 |
"step": 2820
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5087 |
}
|
| 5088 |
],
|
| 5089 |
"logging_steps": 5,
|
|
@@ -5103,7 +5121,7 @@
|
|
| 5103 |
"attributes": {}
|
| 5104 |
}
|
| 5105 |
},
|
| 5106 |
-
"total_flos": 5.
|
| 5107 |
"train_batch_size": 4,
|
| 5108 |
"trial_name": null,
|
| 5109 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 21.76923076923077,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 2830,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 5084 |
"mean_token_accuracy": 0.6562414228916168,
|
| 5085 |
"num_tokens": 10284706.0,
|
| 5086 |
"step": 2820
|
| 5087 |
+
},
|
| 5088 |
+
{
|
| 5089 |
+
"epoch": 21.73076923076923,
|
| 5090 |
+
"grad_norm": 0.0,
|
| 5091 |
+
"learning_rate": 3.588e-05,
|
| 5092 |
+
"loss": 1.7226,
|
| 5093 |
+
"mean_token_accuracy": 0.6576842248439789,
|
| 5094 |
+
"num_tokens": 10302875.0,
|
| 5095 |
+
"step": 2825
|
| 5096 |
+
},
|
| 5097 |
+
{
|
| 5098 |
+
"epoch": 21.76923076923077,
|
| 5099 |
+
"grad_norm": 0.0,
|
| 5100 |
+
"learning_rate": 3.5855e-05,
|
| 5101 |
+
"loss": 1.6721,
|
| 5102 |
+
"mean_token_accuracy": 0.6636721462011337,
|
| 5103 |
+
"num_tokens": 10321523.0,
|
| 5104 |
+
"step": 2830
|
| 5105 |
}
|
| 5106 |
],
|
| 5107 |
"logging_steps": 5,
|
|
|
|
| 5121 |
"attributes": {}
|
| 5122 |
}
|
| 5123 |
},
|
| 5124 |
+
"total_flos": 5.23050263543808e+17,
|
| 5125 |
"train_batch_size": 4,
|
| 5126 |
"trial_name": null,
|
| 5127 |
"trial_params": null
|