End of training
Browse files- README.md +2 -2
- all_results.json +10 -10
- eval_results.json +5 -5
- train_results.json +6 -6
- trainer_state.json +320 -62
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -16,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# mini_llama_crafting_sft_success_new_mem
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B) on the
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 16 |
|
| 17 |
# mini_llama_crafting_sft_success_new_mem
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B) on the identity and the crafting_sft_success_new_mem datasets.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.4032
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime": 5.
|
| 5 |
-
"eval_samples_per_second": 71.
|
| 6 |
-
"eval_steps_per_second":
|
| 7 |
-
"total_flos":
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second": 20.
|
| 11 |
-
"train_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.993662864385298,
|
| 3 |
+
"eval_loss": 0.40319472551345825,
|
| 4 |
+
"eval_runtime": 5.4129,
|
| 5 |
+
"eval_samples_per_second": 71.495,
|
| 6 |
+
"eval_steps_per_second": 9.052,
|
| 7 |
+
"total_flos": 76831349080064.0,
|
| 8 |
+
"train_loss": 0.4719235345957771,
|
| 9 |
+
"train_runtime": 2772.1546,
|
| 10 |
+
"train_samples_per_second": 20.489,
|
| 11 |
+
"train_steps_per_second": 0.159
|
| 12 |
}
|
eval_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime": 5.
|
| 5 |
-
"eval_samples_per_second": 71.
|
| 6 |
-
"eval_steps_per_second":
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.993662864385298,
|
| 3 |
+
"eval_loss": 0.40319472551345825,
|
| 4 |
+
"eval_runtime": 5.4129,
|
| 5 |
+
"eval_samples_per_second": 71.495,
|
| 6 |
+
"eval_steps_per_second": 9.052
|
| 7 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second": 20.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.993662864385298,
|
| 3 |
+
"total_flos": 76831349080064.0,
|
| 4 |
+
"train_loss": 0.4719235345957771,
|
| 5 |
+
"train_runtime": 2772.1546,
|
| 6 |
+
"train_samples_per_second": 20.489,
|
| 7 |
+
"train_steps_per_second": 0.159
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,141 +1,399 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 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,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.06759611322348964,
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss":
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.13519222644697929,
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate":
|
| 22 |
-
"loss": 1.
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.20278833967046894,
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate":
|
| 29 |
-
"loss":
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.27038445289395857,
|
| 34 |
-
"grad_norm": 6.
|
| 35 |
-
"learning_rate":
|
| 36 |
-
"loss": 0.
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.33798056611744826,
|
| 41 |
-
"grad_norm": 6.
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.33798056611744826,
|
| 48 |
-
"eval_loss":
|
| 49 |
-
"eval_runtime": 5.
|
| 50 |
-
"eval_samples_per_second": 71.
|
| 51 |
-
"eval_steps_per_second":
|
| 52 |
"step": 50
|
| 53 |
},
|
| 54 |
{
|
| 55 |
"epoch": 0.4055766793409379,
|
| 56 |
-
"grad_norm": 6.
|
| 57 |
-
"learning_rate":
|
| 58 |
-
"loss": 0.
|
| 59 |
"step": 60
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 0.47317279256442757,
|
| 63 |
-
"grad_norm": 5.
|
| 64 |
-
"learning_rate":
|
| 65 |
-
"loss": 0.
|
| 66 |
"step": 70
|
| 67 |
},
|
| 68 |
{
|
| 69 |
"epoch": 0.5407689057879171,
|
| 70 |
-
"grad_norm": 5.
|
| 71 |
-
"learning_rate":
|
| 72 |
-
"loss": 0.
|
| 73 |
"step": 80
|
| 74 |
},
|
| 75 |
{
|
| 76 |
"epoch": 0.6083650190114068,
|
| 77 |
-
"grad_norm": 4.
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 90
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.6759611322348965,
|
| 84 |
-
"grad_norm": 4.
|
| 85 |
-
"learning_rate":
|
| 86 |
-
"loss": 0.
|
| 87 |
"step": 100
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.6759611322348965,
|
| 91 |
-
"eval_loss": 0.
|
| 92 |
-
"eval_runtime": 5.
|
| 93 |
-
"eval_samples_per_second": 71.
|
| 94 |
-
"eval_steps_per_second": 9.
|
| 95 |
"step": 100
|
| 96 |
},
|
| 97 |
{
|
| 98 |
"epoch": 0.7435572454583862,
|
| 99 |
-
"grad_norm": 3.
|
| 100 |
-
"learning_rate":
|
| 101 |
-
"loss": 0.
|
| 102 |
"step": 110
|
| 103 |
},
|
| 104 |
{
|
| 105 |
"epoch": 0.8111533586818758,
|
| 106 |
-
"grad_norm": 4.
|
| 107 |
-
"learning_rate": 9.
|
| 108 |
-
"loss": 0.
|
| 109 |
"step": 120
|
| 110 |
},
|
| 111 |
{
|
| 112 |
"epoch": 0.8787494719053655,
|
| 113 |
-
"grad_norm":
|
| 114 |
-
"learning_rate":
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 130
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.9463455851288551,
|
| 120 |
-
"grad_norm":
|
| 121 |
-
"learning_rate":
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 140
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch":
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
-
"
|
| 130 |
-
"
|
| 131 |
-
|
| 132 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 133 |
}
|
| 134 |
],
|
| 135 |
"logging_steps": 10,
|
| 136 |
-
"max_steps":
|
| 137 |
"num_input_tokens_seen": 0,
|
| 138 |
-
"num_train_epochs":
|
| 139 |
"save_steps": 500,
|
| 140 |
"stateful_callbacks": {
|
| 141 |
"TrainerControl": {
|
|
@@ -149,7 +407,7 @@
|
|
| 149 |
"attributes": {}
|
| 150 |
}
|
| 151 |
},
|
| 152 |
-
"total_flos":
|
| 153 |
"train_batch_size": 1,
|
| 154 |
"trial_name": null,
|
| 155 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.993662864385298,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 441,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.06759611322348964,
|
| 13 |
+
"grad_norm": 22.71570499945548,
|
| 14 |
+
"learning_rate": 2.222222222222222e-06,
|
| 15 |
+
"loss": 2.2601,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.13519222644697929,
|
| 20 |
+
"grad_norm": 8.716791254344187,
|
| 21 |
+
"learning_rate": 4.444444444444444e-06,
|
| 22 |
+
"loss": 1.4215,
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.20278833967046894,
|
| 27 |
+
"grad_norm": 7.30915109619268,
|
| 28 |
+
"learning_rate": 6.666666666666667e-06,
|
| 29 |
+
"loss": 1.0486,
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.27038445289395857,
|
| 34 |
+
"grad_norm": 6.750139596918985,
|
| 35 |
+
"learning_rate": 8.888888888888888e-06,
|
| 36 |
+
"loss": 0.9049,
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.33798056611744826,
|
| 41 |
+
"grad_norm": 6.330145067445094,
|
| 42 |
+
"learning_rate": 9.996066923030484e-06,
|
| 43 |
+
"loss": 0.8427,
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.33798056611744826,
|
| 48 |
+
"eval_loss": 1.1574571132659912,
|
| 49 |
+
"eval_runtime": 5.4487,
|
| 50 |
+
"eval_samples_per_second": 71.026,
|
| 51 |
+
"eval_steps_per_second": 8.993,
|
| 52 |
"step": 50
|
| 53 |
},
|
| 54 |
{
|
| 55 |
"epoch": 0.4055766793409379,
|
| 56 |
+
"grad_norm": 6.39428702365803,
|
| 57 |
+
"learning_rate": 9.964639423366442e-06,
|
| 58 |
+
"loss": 0.7695,
|
| 59 |
"step": 60
|
| 60 |
},
|
| 61 |
{
|
| 62 |
"epoch": 0.47317279256442757,
|
| 63 |
+
"grad_norm": 5.959101785941899,
|
| 64 |
+
"learning_rate": 9.901982117093786e-06,
|
| 65 |
+
"loss": 0.6954,
|
| 66 |
"step": 70
|
| 67 |
},
|
| 68 |
{
|
| 69 |
"epoch": 0.5407689057879171,
|
| 70 |
+
"grad_norm": 5.715834321289958,
|
| 71 |
+
"learning_rate": 9.808489146745466e-06,
|
| 72 |
+
"loss": 0.625,
|
| 73 |
"step": 80
|
| 74 |
},
|
| 75 |
{
|
| 76 |
"epoch": 0.6083650190114068,
|
| 77 |
+
"grad_norm": 4.754035258710519,
|
| 78 |
+
"learning_rate": 9.68474862499881e-06,
|
| 79 |
+
"loss": 0.5822,
|
| 80 |
"step": 90
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.6759611322348965,
|
| 84 |
+
"grad_norm": 4.514181687896774,
|
| 85 |
+
"learning_rate": 9.531538935183252e-06,
|
| 86 |
+
"loss": 0.5411,
|
| 87 |
"step": 100
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.6759611322348965,
|
| 91 |
+
"eval_loss": 0.5064918994903564,
|
| 92 |
+
"eval_runtime": 5.4326,
|
| 93 |
+
"eval_samples_per_second": 71.236,
|
| 94 |
+
"eval_steps_per_second": 9.02,
|
| 95 |
"step": 100
|
| 96 |
},
|
| 97 |
{
|
| 98 |
"epoch": 0.7435572454583862,
|
| 99 |
+
"grad_norm": 3.4560082126871907,
|
| 100 |
+
"learning_rate": 9.349823834900396e-06,
|
| 101 |
+
"loss": 0.5411,
|
| 102 |
"step": 110
|
| 103 |
},
|
| 104 |
{
|
| 105 |
"epoch": 0.8111533586818758,
|
| 106 |
+
"grad_norm": 4.119312365171623,
|
| 107 |
+
"learning_rate": 9.140746393556853e-06,
|
| 108 |
+
"loss": 0.5097,
|
| 109 |
"step": 120
|
| 110 |
},
|
| 111 |
{
|
| 112 |
"epoch": 0.8787494719053655,
|
| 113 |
+
"grad_norm": 3.5814060437977377,
|
| 114 |
+
"learning_rate": 8.905621801945467e-06,
|
| 115 |
+
"loss": 0.508,
|
| 116 |
"step": 130
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.9463455851288551,
|
| 120 |
+
"grad_norm": 3.640509695845179,
|
| 121 |
+
"learning_rate": 8.645929099105886e-06,
|
| 122 |
+
"loss": 0.5154,
|
| 123 |
"step": 140
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 1.020278833967047,
|
| 127 |
+
"grad_norm": 3.254561037865843,
|
| 128 |
+
"learning_rate": 8.363301868506264e-06,
|
| 129 |
+
"loss": 0.519,
|
| 130 |
+
"step": 150
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.020278833967047,
|
| 134 |
+
"eval_loss": 0.436057448387146,
|
| 135 |
+
"eval_runtime": 5.4096,
|
| 136 |
+
"eval_samples_per_second": 71.539,
|
| 137 |
+
"eval_steps_per_second": 9.058,
|
| 138 |
+
"step": 150
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.0878749471905365,
|
| 142 |
+
"grad_norm": 3.3979728579398176,
|
| 143 |
+
"learning_rate": 8.059517962071234e-06,
|
| 144 |
+
"loss": 0.3811,
|
| 145 |
+
"step": 160
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 1.1554710604140261,
|
| 149 |
+
"grad_norm": 3.5101569989469725,
|
| 150 |
+
"learning_rate": 7.736488316696663e-06,
|
| 151 |
+
"loss": 0.3806,
|
| 152 |
+
"step": 170
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 1.2230671736375158,
|
| 156 |
+
"grad_norm": 4.198624988505521,
|
| 157 |
+
"learning_rate": 7.396244933600285e-06,
|
| 158 |
+
"loss": 0.3917,
|
| 159 |
+
"step": 180
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 1.2906632868610055,
|
| 163 |
+
"grad_norm": 3.71817849809667,
|
| 164 |
+
"learning_rate": 7.040928096123516e-06,
|
| 165 |
+
"loss": 0.379,
|
| 166 |
+
"step": 190
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 1.3582594000844952,
|
| 170 |
+
"grad_norm": 3.770166574810929,
|
| 171 |
+
"learning_rate": 6.672772906390177e-06,
|
| 172 |
+
"loss": 0.3662,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.3582594000844952,
|
| 177 |
+
"eval_loss": 0.40071234107017517,
|
| 178 |
+
"eval_runtime": 5.4129,
|
| 179 |
+
"eval_samples_per_second": 71.495,
|
| 180 |
+
"eval_steps_per_second": 9.052,
|
| 181 |
+
"step": 200
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 1.4258555133079849,
|
| 185 |
+
"grad_norm": 3.102586008371691,
|
| 186 |
+
"learning_rate": 6.294095225512604e-06,
|
| 187 |
+
"loss": 0.3906,
|
| 188 |
+
"step": 210
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.4934516265314746,
|
| 192 |
+
"grad_norm": 3.6078815896150265,
|
| 193 |
+
"learning_rate": 5.907277105787513e-06,
|
| 194 |
+
"loss": 0.37,
|
| 195 |
+
"step": 220
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.5610477397549642,
|
| 199 |
+
"grad_norm": 3.5433228920014237,
|
| 200 |
+
"learning_rate": 5.514751806519673e-06,
|
| 201 |
+
"loss": 0.3647,
|
| 202 |
+
"step": 230
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 1.6286438529784537,
|
| 206 |
+
"grad_norm": 3.607485131351974,
|
| 207 |
+
"learning_rate": 5.118988487730537e-06,
|
| 208 |
+
"loss": 0.3761,
|
| 209 |
+
"step": 240
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.6962399662019434,
|
| 213 |
+
"grad_norm": 3.187953581780823,
|
| 214 |
+
"learning_rate": 4.7224766780353005e-06,
|
| 215 |
+
"loss": 0.3679,
|
| 216 |
+
"step": 250
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 1.6962399662019434,
|
| 220 |
+
"eval_loss": 0.39477795362472534,
|
| 221 |
+
"eval_runtime": 5.4127,
|
| 222 |
+
"eval_samples_per_second": 71.499,
|
| 223 |
+
"eval_steps_per_second": 9.053,
|
| 224 |
+
"step": 250
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.763836079425433,
|
| 228 |
+
"grad_norm": 3.5445294506545135,
|
| 229 |
+
"learning_rate": 4.327710614392341e-06,
|
| 230 |
+
"loss": 0.3552,
|
| 231 |
+
"step": 260
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.8314321926489225,
|
| 235 |
+
"grad_norm": 3.743551074019313,
|
| 236 |
+
"learning_rate": 3.937173552235117e-06,
|
| 237 |
+
"loss": 0.3635,
|
| 238 |
+
"step": 270
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.8990283058724122,
|
| 242 |
+
"grad_norm": 3.542063112753628,
|
| 243 |
+
"learning_rate": 3.553322144682737e-06,
|
| 244 |
+
"loss": 0.3491,
|
| 245 |
+
"step": 280
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.966624419095902,
|
| 249 |
+
"grad_norm": 3.5873733457606956,
|
| 250 |
+
"learning_rate": 3.178570989091028e-06,
|
| 251 |
+
"loss": 0.3524,
|
| 252 |
+
"step": 290
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 2.040557667934094,
|
| 256 |
+
"grad_norm": 3.0378458947692257,
|
| 257 |
+
"learning_rate": 2.8152774381532033e-06,
|
| 258 |
+
"loss": 0.3176,
|
| 259 |
+
"step": 300
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 2.040557667934094,
|
| 263 |
+
"eval_loss": 0.3846343159675598,
|
| 264 |
+
"eval_runtime": 5.4138,
|
| 265 |
+
"eval_samples_per_second": 71.483,
|
| 266 |
+
"eval_steps_per_second": 9.051,
|
| 267 |
+
"step": 300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 2.1081537811575832,
|
| 271 |
+
"grad_norm": 3.6969633690268178,
|
| 272 |
+
"learning_rate": 2.465726771095086e-06,
|
| 273 |
+
"loss": 0.2279,
|
| 274 |
+
"step": 310
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 2.175749894381073,
|
| 278 |
+
"grad_norm": 3.5512676423267386,
|
| 279 |
+
"learning_rate": 2.132117818244771e-06,
|
| 280 |
+
"loss": 0.219,
|
| 281 |
+
"step": 320
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 2.2433460076045626,
|
| 285 |
+
"grad_norm": 3.8134684201298845,
|
| 286 |
+
"learning_rate": 1.8165491294045596e-06,
|
| 287 |
+
"loss": 0.2136,
|
| 288 |
+
"step": 330
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.3109421208280523,
|
| 292 |
+
"grad_norm": 3.429664297466825,
|
| 293 |
+
"learning_rate": 1.521005773032362e-06,
|
| 294 |
+
"loss": 0.2164,
|
| 295 |
+
"step": 340
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 2.378538234051542,
|
| 299 |
+
"grad_norm": 3.874687323224451,
|
| 300 |
+
"learning_rate": 1.2473468492715896e-06,
|
| 301 |
+
"loss": 0.2141,
|
| 302 |
+
"step": 350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.378538234051542,
|
| 306 |
+
"eval_loss": 0.4075692594051361,
|
| 307 |
+
"eval_runtime": 5.4106,
|
| 308 |
+
"eval_samples_per_second": 71.526,
|
| 309 |
+
"eval_steps_per_second": 9.056,
|
| 310 |
+
"step": 350
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 2.4461343472750317,
|
| 314 |
+
"grad_norm": 3.9002573113381875,
|
| 315 |
+
"learning_rate": 9.972937953781985e-07,
|
| 316 |
+
"loss": 0.2205,
|
| 317 |
+
"step": 360
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 2.5137304604985213,
|
| 321 |
+
"grad_norm": 3.6422878228347964,
|
| 322 |
+
"learning_rate": 7.724195571089787e-07,
|
| 323 |
+
"loss": 0.2171,
|
| 324 |
+
"step": 370
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 2.581326573722011,
|
| 328 |
+
"grad_norm": 3.290391228862638,
|
| 329 |
+
"learning_rate": 5.741386941879179e-07,
|
| 330 |
+
"loss": 0.206,
|
| 331 |
+
"step": 380
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 2.6489226869455007,
|
| 335 |
+
"grad_norm": 4.456138437370725,
|
| 336 |
+
"learning_rate": 4.036984820916723e-07,
|
| 337 |
+
"loss": 0.2126,
|
| 338 |
+
"step": 390
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.7165188001689904,
|
| 342 |
+
"grad_norm": 3.4818603403336232,
|
| 343 |
+
"learning_rate": 2.621710661279253e-07,
|
| 344 |
+
"loss": 0.2089,
|
| 345 |
+
"step": 400
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.7165188001689904,
|
| 349 |
+
"eval_loss": 0.3996022343635559,
|
| 350 |
+
"eval_runtime": 5.4212,
|
| 351 |
+
"eval_samples_per_second": 71.387,
|
| 352 |
+
"eval_steps_per_second": 9.039,
|
| 353 |
+
"step": 400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.78411491339248,
|
| 357 |
+
"grad_norm": 4.331395175879503,
|
| 358 |
+
"learning_rate": 1.5044671716097414e-07,
|
| 359 |
+
"loss": 0.2188,
|
| 360 |
+
"step": 410
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 2.8517110266159698,
|
| 364 |
+
"grad_norm": 3.603270058835426,
|
| 365 |
+
"learning_rate": 6.922823140906754e-08,
|
| 366 |
+
"loss": 0.2107,
|
| 367 |
+
"step": 420
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 2.919307139839459,
|
| 371 |
+
"grad_norm": 3.627133298336489,
|
| 372 |
+
"learning_rate": 1.9026509541272276e-08,
|
| 373 |
+
"loss": 0.2084,
|
| 374 |
+
"step": 430
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.986903253062949,
|
| 378 |
+
"grad_norm": 3.7792062209865227,
|
| 379 |
+
"learning_rate": 1.5734288333457692e-10,
|
| 380 |
+
"loss": 0.203,
|
| 381 |
+
"step": 440
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 2.993662864385298,
|
| 385 |
+
"step": 441,
|
| 386 |
+
"total_flos": 76831349080064.0,
|
| 387 |
+
"train_loss": 0.4719235345957771,
|
| 388 |
+
"train_runtime": 2772.1546,
|
| 389 |
+
"train_samples_per_second": 20.489,
|
| 390 |
+
"train_steps_per_second": 0.159
|
| 391 |
}
|
| 392 |
],
|
| 393 |
"logging_steps": 10,
|
| 394 |
+
"max_steps": 441,
|
| 395 |
"num_input_tokens_seen": 0,
|
| 396 |
+
"num_train_epochs": 3,
|
| 397 |
"save_steps": 500,
|
| 398 |
"stateful_callbacks": {
|
| 399 |
"TrainerControl": {
|
|
|
|
| 407 |
"attributes": {}
|
| 408 |
}
|
| 409 |
},
|
| 410 |
+
"total_flos": 76831349080064.0,
|
| 411 |
"train_batch_size": 1,
|
| 412 |
"trial_name": null,
|
| 413 |
"trial_params": null
|
training_eval_loss.png
CHANGED
|
|
training_loss.png
CHANGED
|
|