End of training
Browse files- README.md +1 -1
- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +9 -0
- trainer_state.json +534 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
# train_cb_42_1760637133
|
| 19 |
|
| 20 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
- Loss: 0.2378
|
| 23 |
- Num Input Tokens Seen: 158656
|
|
|
|
| 17 |
|
| 18 |
# train_cb_42_1760637133
|
| 19 |
|
| 20 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the cb dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
- Loss: 0.2378
|
| 23 |
- Num Input Tokens Seen: 158656
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_loss": 0.23779883980751038,
|
| 4 |
+
"eval_runtime": 1.3015,
|
| 5 |
+
"eval_samples_per_second": 38.417,
|
| 6 |
+
"eval_steps_per_second": 9.988,
|
| 7 |
+
"num_input_tokens_seen": 158656,
|
| 8 |
+
"total_flos": 7144207972564992.0,
|
| 9 |
+
"train_loss": 0.6316961252689361,
|
| 10 |
+
"train_runtime": 79.306,
|
| 11 |
+
"train_samples_per_second": 12.609,
|
| 12 |
+
"train_steps_per_second": 3.152
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_loss": 0.23779883980751038,
|
| 4 |
+
"eval_runtime": 1.3015,
|
| 5 |
+
"eval_samples_per_second": 38.417,
|
| 6 |
+
"eval_steps_per_second": 9.988,
|
| 7 |
+
"num_input_tokens_seen": 158656
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"num_input_tokens_seen": 158656,
|
| 4 |
+
"total_flos": 7144207972564992.0,
|
| 5 |
+
"train_loss": 0.6316961252689361,
|
| 6 |
+
"train_runtime": 79.306,
|
| 7 |
+
"train_samples_per_second": 12.609,
|
| 8 |
+
"train_steps_per_second": 3.152
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,534 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 175,
|
| 3 |
+
"best_metric": 0.22990073263645172,
|
| 4 |
+
"best_model_checkpoint": "saves_multiple/prefix-tuning/llama-3-8b-instruct/train_cb_42_1760637133/checkpoint-175",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 25,
|
| 7 |
+
"global_step": 250,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1,
|
| 14 |
+
"grad_norm": 147.09878540039062,
|
| 15 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 16 |
+
"loss": 7.3911,
|
| 17 |
+
"num_input_tokens_seen": 2720,
|
| 18 |
+
"step": 5
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 0.2,
|
| 22 |
+
"grad_norm": 201.96444702148438,
|
| 23 |
+
"learning_rate": 3.6000000000000003e-06,
|
| 24 |
+
"loss": 6.0584,
|
| 25 |
+
"num_input_tokens_seen": 5536,
|
| 26 |
+
"step": 10
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.3,
|
| 30 |
+
"grad_norm": 169.54861450195312,
|
| 31 |
+
"learning_rate": 5.600000000000001e-06,
|
| 32 |
+
"loss": 4.1152,
|
| 33 |
+
"num_input_tokens_seen": 8480,
|
| 34 |
+
"step": 15
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.4,
|
| 38 |
+
"grad_norm": 46.49274826049805,
|
| 39 |
+
"learning_rate": 7.600000000000001e-06,
|
| 40 |
+
"loss": 2.3628,
|
| 41 |
+
"num_input_tokens_seen": 11776,
|
| 42 |
+
"step": 20
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.5,
|
| 46 |
+
"grad_norm": 27.129613876342773,
|
| 47 |
+
"learning_rate": 9.600000000000001e-06,
|
| 48 |
+
"loss": 1.1162,
|
| 49 |
+
"num_input_tokens_seen": 14720,
|
| 50 |
+
"step": 25
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.5,
|
| 54 |
+
"eval_loss": 0.7037240266799927,
|
| 55 |
+
"eval_runtime": 1.2436,
|
| 56 |
+
"eval_samples_per_second": 40.204,
|
| 57 |
+
"eval_steps_per_second": 10.453,
|
| 58 |
+
"num_input_tokens_seen": 14720,
|
| 59 |
+
"step": 25
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.6,
|
| 63 |
+
"grad_norm": 48.88370895385742,
|
| 64 |
+
"learning_rate": 9.992203820909906e-06,
|
| 65 |
+
"loss": 0.5016,
|
| 66 |
+
"num_input_tokens_seen": 18560,
|
| 67 |
+
"step": 30
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.7,
|
| 71 |
+
"grad_norm": 26.719514846801758,
|
| 72 |
+
"learning_rate": 9.960573506572391e-06,
|
| 73 |
+
"loss": 0.2915,
|
| 74 |
+
"num_input_tokens_seen": 21824,
|
| 75 |
+
"step": 35
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.8,
|
| 79 |
+
"grad_norm": 9.22452449798584,
|
| 80 |
+
"learning_rate": 9.904775776745959e-06,
|
| 81 |
+
"loss": 0.2798,
|
| 82 |
+
"num_input_tokens_seen": 24288,
|
| 83 |
+
"step": 40
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.9,
|
| 87 |
+
"grad_norm": 22.899555206298828,
|
| 88 |
+
"learning_rate": 9.825082472361558e-06,
|
| 89 |
+
"loss": 0.3252,
|
| 90 |
+
"num_input_tokens_seen": 27648,
|
| 91 |
+
"step": 45
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.0,
|
| 95 |
+
"grad_norm": 11.016155242919922,
|
| 96 |
+
"learning_rate": 9.721881851187406e-06,
|
| 97 |
+
"loss": 0.2568,
|
| 98 |
+
"num_input_tokens_seen": 31456,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.0,
|
| 103 |
+
"eval_loss": 0.28914472460746765,
|
| 104 |
+
"eval_runtime": 1.2319,
|
| 105 |
+
"eval_samples_per_second": 40.588,
|
| 106 |
+
"eval_steps_per_second": 10.553,
|
| 107 |
+
"num_input_tokens_seen": 31456,
|
| 108 |
+
"step": 50
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.1,
|
| 112 |
+
"grad_norm": 9.274727821350098,
|
| 113 |
+
"learning_rate": 9.595676696276173e-06,
|
| 114 |
+
"loss": 0.24,
|
| 115 |
+
"num_input_tokens_seen": 34976,
|
| 116 |
+
"step": 55
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.2,
|
| 120 |
+
"grad_norm": 30.56814956665039,
|
| 121 |
+
"learning_rate": 9.44708186645649e-06,
|
| 122 |
+
"loss": 0.2301,
|
| 123 |
+
"num_input_tokens_seen": 37856,
|
| 124 |
+
"step": 60
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 1.3,
|
| 128 |
+
"grad_norm": 14.439095497131348,
|
| 129 |
+
"learning_rate": 9.276821300802535e-06,
|
| 130 |
+
"loss": 0.2214,
|
| 131 |
+
"num_input_tokens_seen": 41728,
|
| 132 |
+
"step": 65
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.4,
|
| 136 |
+
"grad_norm": 105.18769836425781,
|
| 137 |
+
"learning_rate": 9.085724491675642e-06,
|
| 138 |
+
"loss": 0.3427,
|
| 139 |
+
"num_input_tokens_seen": 44704,
|
| 140 |
+
"step": 70
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.5,
|
| 144 |
+
"grad_norm": 69.08840942382812,
|
| 145 |
+
"learning_rate": 8.874722443520898e-06,
|
| 146 |
+
"loss": 0.254,
|
| 147 |
+
"num_input_tokens_seen": 47168,
|
| 148 |
+
"step": 75
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 1.5,
|
| 152 |
+
"eval_loss": 0.2438250482082367,
|
| 153 |
+
"eval_runtime": 1.2262,
|
| 154 |
+
"eval_samples_per_second": 40.778,
|
| 155 |
+
"eval_steps_per_second": 10.602,
|
| 156 |
+
"num_input_tokens_seen": 47168,
|
| 157 |
+
"step": 75
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.6,
|
| 161 |
+
"grad_norm": 37.581912994384766,
|
| 162 |
+
"learning_rate": 8.644843137107058e-06,
|
| 163 |
+
"loss": 0.2904,
|
| 164 |
+
"num_input_tokens_seen": 51072,
|
| 165 |
+
"step": 80
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.7,
|
| 169 |
+
"grad_norm": 152.4672088623047,
|
| 170 |
+
"learning_rate": 8.397206521307584e-06,
|
| 171 |
+
"loss": 0.4201,
|
| 172 |
+
"num_input_tokens_seen": 54400,
|
| 173 |
+
"step": 85
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.8,
|
| 177 |
+
"grad_norm": 128.72598266601562,
|
| 178 |
+
"learning_rate": 8.133019056822303e-06,
|
| 179 |
+
"loss": 0.2121,
|
| 180 |
+
"num_input_tokens_seen": 56736,
|
| 181 |
+
"step": 90
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 1.9,
|
| 185 |
+
"grad_norm": 29.933185577392578,
|
| 186 |
+
"learning_rate": 7.85356783842216e-06,
|
| 187 |
+
"loss": 0.2648,
|
| 188 |
+
"num_input_tokens_seen": 60064,
|
| 189 |
+
"step": 95
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 2.0,
|
| 193 |
+
"grad_norm": 10.928328514099121,
|
| 194 |
+
"learning_rate": 7.560214324352858e-06,
|
| 195 |
+
"loss": 0.1647,
|
| 196 |
+
"num_input_tokens_seen": 63168,
|
| 197 |
+
"step": 100
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 2.0,
|
| 201 |
+
"eval_loss": 0.36001044511795044,
|
| 202 |
+
"eval_runtime": 1.235,
|
| 203 |
+
"eval_samples_per_second": 40.485,
|
| 204 |
+
"eval_steps_per_second": 10.526,
|
| 205 |
+
"num_input_tokens_seen": 63168,
|
| 206 |
+
"step": 100
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.1,
|
| 210 |
+
"grad_norm": 70.27168273925781,
|
| 211 |
+
"learning_rate": 7.254387703447154e-06,
|
| 212 |
+
"loss": 0.427,
|
| 213 |
+
"num_input_tokens_seen": 66464,
|
| 214 |
+
"step": 105
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 2.2,
|
| 218 |
+
"grad_norm": 61.89235305786133,
|
| 219 |
+
"learning_rate": 6.9375779322605154e-06,
|
| 220 |
+
"loss": 0.277,
|
| 221 |
+
"num_input_tokens_seen": 70112,
|
| 222 |
+
"step": 110
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 2.3,
|
| 226 |
+
"grad_norm": 8.1712646484375,
|
| 227 |
+
"learning_rate": 6.611328476152557e-06,
|
| 228 |
+
"loss": 0.1532,
|
| 229 |
+
"num_input_tokens_seen": 73376,
|
| 230 |
+
"step": 115
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.4,
|
| 234 |
+
"grad_norm": 31.11164665222168,
|
| 235 |
+
"learning_rate": 6.277228789678953e-06,
|
| 236 |
+
"loss": 0.174,
|
| 237 |
+
"num_input_tokens_seen": 76288,
|
| 238 |
+
"step": 120
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 2.5,
|
| 242 |
+
"grad_norm": 10.602115631103516,
|
| 243 |
+
"learning_rate": 5.936906572928625e-06,
|
| 244 |
+
"loss": 0.3223,
|
| 245 |
+
"num_input_tokens_seen": 79424,
|
| 246 |
+
"step": 125
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 2.5,
|
| 250 |
+
"eval_loss": 0.2356501966714859,
|
| 251 |
+
"eval_runtime": 1.2576,
|
| 252 |
+
"eval_samples_per_second": 39.757,
|
| 253 |
+
"eval_steps_per_second": 10.337,
|
| 254 |
+
"num_input_tokens_seen": 79424,
|
| 255 |
+
"step": 125
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.6,
|
| 259 |
+
"grad_norm": 8.446126937866211,
|
| 260 |
+
"learning_rate": 5.592019841532507e-06,
|
| 261 |
+
"loss": 0.1352,
|
| 262 |
+
"num_input_tokens_seen": 82880,
|
| 263 |
+
"step": 130
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 2.7,
|
| 267 |
+
"grad_norm": 55.46540832519531,
|
| 268 |
+
"learning_rate": 5.244248848978067e-06,
|
| 269 |
+
"loss": 0.2971,
|
| 270 |
+
"num_input_tokens_seen": 86560,
|
| 271 |
+
"step": 135
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 2.8,
|
| 275 |
+
"grad_norm": 12.031166076660156,
|
| 276 |
+
"learning_rate": 4.895287900583216e-06,
|
| 277 |
+
"loss": 0.1537,
|
| 278 |
+
"num_input_tokens_seen": 89184,
|
| 279 |
+
"step": 140
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.9,
|
| 283 |
+
"grad_norm": 9.252001762390137,
|
| 284 |
+
"learning_rate": 4.546837099011101e-06,
|
| 285 |
+
"loss": 0.3131,
|
| 286 |
+
"num_input_tokens_seen": 92256,
|
| 287 |
+
"step": 145
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 3.0,
|
| 291 |
+
"grad_norm": 35.95872497558594,
|
| 292 |
+
"learning_rate": 4.200594061540827e-06,
|
| 293 |
+
"loss": 0.2999,
|
| 294 |
+
"num_input_tokens_seen": 95168,
|
| 295 |
+
"step": 150
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 3.0,
|
| 299 |
+
"eval_loss": 0.2933087944984436,
|
| 300 |
+
"eval_runtime": 1.2189,
|
| 301 |
+
"eval_samples_per_second": 41.021,
|
| 302 |
+
"eval_steps_per_second": 10.666,
|
| 303 |
+
"num_input_tokens_seen": 95168,
|
| 304 |
+
"step": 150
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.1,
|
| 308 |
+
"grad_norm": 63.808555603027344,
|
| 309 |
+
"learning_rate": 3.8582456494467214e-06,
|
| 310 |
+
"loss": 0.2206,
|
| 311 |
+
"num_input_tokens_seen": 98528,
|
| 312 |
+
"step": 155
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 3.2,
|
| 316 |
+
"grad_norm": 38.348934173583984,
|
| 317 |
+
"learning_rate": 3.521459749779769e-06,
|
| 318 |
+
"loss": 0.2139,
|
| 319 |
+
"num_input_tokens_seen": 101408,
|
| 320 |
+
"step": 160
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 3.3,
|
| 324 |
+
"grad_norm": 116.32037353515625,
|
| 325 |
+
"learning_rate": 3.1918771495895395e-06,
|
| 326 |
+
"loss": 0.2016,
|
| 327 |
+
"num_input_tokens_seen": 104736,
|
| 328 |
+
"step": 165
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 3.4,
|
| 332 |
+
"grad_norm": 56.04789352416992,
|
| 333 |
+
"learning_rate": 2.871103542174637e-06,
|
| 334 |
+
"loss": 0.2427,
|
| 335 |
+
"num_input_tokens_seen": 107840,
|
| 336 |
+
"step": 170
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.5,
|
| 340 |
+
"grad_norm": 35.397884368896484,
|
| 341 |
+
"learning_rate": 2.560701704306336e-06,
|
| 342 |
+
"loss": 0.244,
|
| 343 |
+
"num_input_tokens_seen": 111392,
|
| 344 |
+
"step": 175
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 3.5,
|
| 348 |
+
"eval_loss": 0.22990073263645172,
|
| 349 |
+
"eval_runtime": 1.2427,
|
| 350 |
+
"eval_samples_per_second": 40.235,
|
| 351 |
+
"eval_steps_per_second": 10.461,
|
| 352 |
+
"num_input_tokens_seen": 111392,
|
| 353 |
+
"step": 175
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 3.6,
|
| 357 |
+
"grad_norm": 5.372102737426758,
|
| 358 |
+
"learning_rate": 2.2621838825372496e-06,
|
| 359 |
+
"loss": 0.1392,
|
| 360 |
+
"num_input_tokens_seen": 114720,
|
| 361 |
+
"step": 180
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 3.7,
|
| 365 |
+
"grad_norm": 44.36689376831055,
|
| 366 |
+
"learning_rate": 1.977004425688126e-06,
|
| 367 |
+
"loss": 0.1194,
|
| 368 |
+
"num_input_tokens_seen": 117632,
|
| 369 |
+
"step": 185
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 3.8,
|
| 373 |
+
"grad_norm": 32.87962341308594,
|
| 374 |
+
"learning_rate": 1.7065526994065973e-06,
|
| 375 |
+
"loss": 0.1453,
|
| 376 |
+
"num_input_tokens_seen": 120512,
|
| 377 |
+
"step": 190
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 3.9,
|
| 381 |
+
"grad_norm": 15.784488677978516,
|
| 382 |
+
"learning_rate": 1.4521463173173966e-06,
|
| 383 |
+
"loss": 0.2925,
|
| 384 |
+
"num_input_tokens_seen": 124352,
|
| 385 |
+
"step": 195
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 4.0,
|
| 389 |
+
"grad_norm": 8.471246719360352,
|
| 390 |
+
"learning_rate": 1.2150247217412186e-06,
|
| 391 |
+
"loss": 0.2501,
|
| 392 |
+
"num_input_tokens_seen": 127136,
|
| 393 |
+
"step": 200
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 4.0,
|
| 397 |
+
"eval_loss": 0.23893244564533234,
|
| 398 |
+
"eval_runtime": 1.2441,
|
| 399 |
+
"eval_samples_per_second": 40.189,
|
| 400 |
+
"eval_steps_per_second": 10.449,
|
| 401 |
+
"num_input_tokens_seen": 127136,
|
| 402 |
+
"step": 200
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 4.1,
|
| 406 |
+
"grad_norm": 31.993316650390625,
|
| 407 |
+
"learning_rate": 9.963431452563331e-07,
|
| 408 |
+
"loss": 0.1709,
|
| 409 |
+
"num_input_tokens_seen": 129664,
|
| 410 |
+
"step": 205
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 4.2,
|
| 414 |
+
"grad_norm": 27.969955444335938,
|
| 415 |
+
"learning_rate": 7.971669825215789e-07,
|
| 416 |
+
"loss": 0.2556,
|
| 417 |
+
"num_input_tokens_seen": 132128,
|
| 418 |
+
"step": 210
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 4.3,
|
| 422 |
+
"grad_norm": 22.807920455932617,
|
| 423 |
+
"learning_rate": 6.184665997806832e-07,
|
| 424 |
+
"loss": 0.143,
|
| 425 |
+
"num_input_tokens_seen": 135200,
|
| 426 |
+
"step": 215
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 4.4,
|
| 430 |
+
"grad_norm": 14.716069221496582,
|
| 431 |
+
"learning_rate": 4.6111260733545714e-07,
|
| 432 |
+
"loss": 0.2176,
|
| 433 |
+
"num_input_tokens_seen": 138368,
|
| 434 |
+
"step": 220
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 4.5,
|
| 438 |
+
"grad_norm": 15.53546142578125,
|
| 439 |
+
"learning_rate": 3.258716180199278e-07,
|
| 440 |
+
"loss": 0.1977,
|
| 441 |
+
"num_input_tokens_seen": 141632,
|
| 442 |
+
"step": 225
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 4.5,
|
| 446 |
+
"eval_loss": 0.23097044229507446,
|
| 447 |
+
"eval_runtime": 1.2534,
|
| 448 |
+
"eval_samples_per_second": 39.89,
|
| 449 |
+
"eval_steps_per_second": 10.371,
|
| 450 |
+
"num_input_tokens_seen": 141632,
|
| 451 |
+
"step": 225
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 4.6,
|
| 455 |
+
"grad_norm": 18.080829620361328,
|
| 456 |
+
"learning_rate": 2.134025123396638e-07,
|
| 457 |
+
"loss": 0.1109,
|
| 458 |
+
"num_input_tokens_seen": 144960,
|
| 459 |
+
"step": 230
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 4.7,
|
| 463 |
+
"grad_norm": 26.747798919677734,
|
| 464 |
+
"learning_rate": 1.2425322847218368e-07,
|
| 465 |
+
"loss": 0.1466,
|
| 466 |
+
"num_input_tokens_seen": 147936,
|
| 467 |
+
"step": 235
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 4.8,
|
| 471 |
+
"grad_norm": 7.783914089202881,
|
| 472 |
+
"learning_rate": 5.8858092767236084e-08,
|
| 473 |
+
"loss": 0.1319,
|
| 474 |
+
"num_input_tokens_seen": 151392,
|
| 475 |
+
"step": 240
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 4.9,
|
| 479 |
+
"grad_norm": 16.87287712097168,
|
| 480 |
+
"learning_rate": 1.753570375247815e-08,
|
| 481 |
+
"loss": 0.1445,
|
| 482 |
+
"num_input_tokens_seen": 154624,
|
| 483 |
+
"step": 245
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 5.0,
|
| 487 |
+
"grad_norm": 15.167189598083496,
|
| 488 |
+
"learning_rate": 4.87379953478806e-10,
|
| 489 |
+
"loss": 0.1052,
|
| 490 |
+
"num_input_tokens_seen": 158656,
|
| 491 |
+
"step": 250
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 5.0,
|
| 495 |
+
"eval_loss": 0.23779883980751038,
|
| 496 |
+
"eval_runtime": 1.2348,
|
| 497 |
+
"eval_samples_per_second": 40.491,
|
| 498 |
+
"eval_steps_per_second": 10.528,
|
| 499 |
+
"num_input_tokens_seen": 158656,
|
| 500 |
+
"step": 250
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 5.0,
|
| 504 |
+
"num_input_tokens_seen": 158656,
|
| 505 |
+
"step": 250,
|
| 506 |
+
"total_flos": 7144207972564992.0,
|
| 507 |
+
"train_loss": 0.6316961252689361,
|
| 508 |
+
"train_runtime": 79.306,
|
| 509 |
+
"train_samples_per_second": 12.609,
|
| 510 |
+
"train_steps_per_second": 3.152
|
| 511 |
+
}
|
| 512 |
+
],
|
| 513 |
+
"logging_steps": 5,
|
| 514 |
+
"max_steps": 250,
|
| 515 |
+
"num_input_tokens_seen": 158656,
|
| 516 |
+
"num_train_epochs": 5,
|
| 517 |
+
"save_steps": 25,
|
| 518 |
+
"stateful_callbacks": {
|
| 519 |
+
"TrainerControl": {
|
| 520 |
+
"args": {
|
| 521 |
+
"should_epoch_stop": false,
|
| 522 |
+
"should_evaluate": false,
|
| 523 |
+
"should_log": false,
|
| 524 |
+
"should_save": true,
|
| 525 |
+
"should_training_stop": true
|
| 526 |
+
},
|
| 527 |
+
"attributes": {}
|
| 528 |
+
}
|
| 529 |
+
},
|
| 530 |
+
"total_flos": 7144207972564992.0,
|
| 531 |
+
"train_batch_size": 4,
|
| 532 |
+
"trial_name": null,
|
| 533 |
+
"trial_params": null
|
| 534 |
+
}
|
training_eval_loss.png
ADDED
|
training_loss.png
ADDED
|