Model save
Browse files- README.md +75 -0
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +326 -0
README.md
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 9 |
+
model-index:
|
| 10 |
+
- name: nash_dpo_merge_iter_2
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# nash_dpo_merge_iter_2
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.6368
|
| 22 |
+
- Rewards/chosen: -0.5885
|
| 23 |
+
- Rewards/rejected: -0.7591
|
| 24 |
+
- Rewards/accuracies: 0.6380
|
| 25 |
+
- Rewards/margins: 0.1706
|
| 26 |
+
- Logps/rejected: -365.7411
|
| 27 |
+
- Logps/chosen: -357.2530
|
| 28 |
+
- Logits/rejected: -2.1348
|
| 29 |
+
- Logits/chosen: -2.2675
|
| 30 |
+
|
| 31 |
+
## Model description
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Intended uses & limitations
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training and evaluation data
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- learning_rate: 5e-06
|
| 49 |
+
- train_batch_size: 2
|
| 50 |
+
- eval_batch_size: 2
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 16
|
| 55 |
+
- total_train_batch_size: 128
|
| 56 |
+
- total_eval_batch_size: 8
|
| 57 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
+
- lr_scheduler_type: cosine
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 1
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.6429 | 0.51 | 100 | 0.6368 | -0.5885 | -0.7591 | 0.6380 | 0.1706 | -365.7411 | -357.2530 | -2.1348 | -2.2675 |
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
### Framework versions
|
| 70 |
+
|
| 71 |
+
- PEFT 0.7.1
|
| 72 |
+
- Transformers 4.36.2
|
| 73 |
+
- Pytorch 2.1.2+cu121
|
| 74 |
+
- Datasets 2.14.6
|
| 75 |
+
- Tokenizers 0.15.2
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6412642754041231,
|
| 4 |
+
"train_runtime": 8982.6903,
|
| 5 |
+
"train_samples": 25000,
|
| 6 |
+
"train_samples_per_second": 2.783,
|
| 7 |
+
"train_steps_per_second": 0.022
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6412642754041231,
|
| 4 |
+
"train_runtime": 8982.6903,
|
| 5 |
+
"train_samples": 25000,
|
| 6 |
+
"train_samples_per_second": 2.783,
|
| 7 |
+
"train_steps_per_second": 0.022
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,326 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9984,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 195,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 2.5000000000000004e-07,
|
| 14 |
+
"logits/chosen": -2.535799264907837,
|
| 15 |
+
"logits/rejected": -2.5660998821258545,
|
| 16 |
+
"logps/chosen": -270.7046813964844,
|
| 17 |
+
"logps/rejected": -307.25311279296875,
|
| 18 |
+
"loss": 0.6931,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.05,
|
| 27 |
+
"learning_rate": 2.5e-06,
|
| 28 |
+
"logits/chosen": -2.63254976272583,
|
| 29 |
+
"logits/rejected": -2.5803234577178955,
|
| 30 |
+
"logps/chosen": -320.9474792480469,
|
| 31 |
+
"logps/rejected": -316.3360900878906,
|
| 32 |
+
"loss": 0.6926,
|
| 33 |
+
"rewards/accuracies": 0.4618055522441864,
|
| 34 |
+
"rewards/chosen": -0.0115978317335248,
|
| 35 |
+
"rewards/margins": 0.0027361076790839434,
|
| 36 |
+
"rewards/rejected": -0.014333941973745823,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.1,
|
| 41 |
+
"learning_rate": 5e-06,
|
| 42 |
+
"logits/chosen": -2.5984463691711426,
|
| 43 |
+
"logits/rejected": -2.569882869720459,
|
| 44 |
+
"logps/chosen": -307.0980529785156,
|
| 45 |
+
"logps/rejected": -311.14959716796875,
|
| 46 |
+
"loss": 0.6898,
|
| 47 |
+
"rewards/accuracies": 0.546875,
|
| 48 |
+
"rewards/chosen": -0.03875315934419632,
|
| 49 |
+
"rewards/margins": 0.007569065783172846,
|
| 50 |
+
"rewards/rejected": -0.04632222652435303,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 4.959823971496575e-06,
|
| 56 |
+
"logits/chosen": -2.5634453296661377,
|
| 57 |
+
"logits/rejected": -2.4752097129821777,
|
| 58 |
+
"logps/chosen": -322.8082580566406,
|
| 59 |
+
"logps/rejected": -310.7356262207031,
|
| 60 |
+
"loss": 0.6782,
|
| 61 |
+
"rewards/accuracies": 0.6343749761581421,
|
| 62 |
+
"rewards/chosen": -0.026555459946393967,
|
| 63 |
+
"rewards/margins": 0.03691429644823074,
|
| 64 |
+
"rewards/rejected": -0.06346975266933441,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.2,
|
| 69 |
+
"learning_rate": 4.8405871765993435e-06,
|
| 70 |
+
"logits/chosen": -2.55033278465271,
|
| 71 |
+
"logits/rejected": -2.459357738494873,
|
| 72 |
+
"logps/chosen": -333.2048034667969,
|
| 73 |
+
"logps/rejected": -334.33538818359375,
|
| 74 |
+
"loss": 0.6719,
|
| 75 |
+
"rewards/accuracies": 0.581250011920929,
|
| 76 |
+
"rewards/chosen": -0.10182944685220718,
|
| 77 |
+
"rewards/margins": 0.038650404661893845,
|
| 78 |
+
"rewards/rejected": -0.14047986268997192,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.26,
|
| 83 |
+
"learning_rate": 4.646121984004666e-06,
|
| 84 |
+
"logits/chosen": -2.530797243118286,
|
| 85 |
+
"logits/rejected": -2.496924877166748,
|
| 86 |
+
"logps/chosen": -313.50885009765625,
|
| 87 |
+
"logps/rejected": -333.51397705078125,
|
| 88 |
+
"loss": 0.6599,
|
| 89 |
+
"rewards/accuracies": 0.671875,
|
| 90 |
+
"rewards/chosen": -0.13014966249465942,
|
| 91 |
+
"rewards/margins": 0.07707841694355011,
|
| 92 |
+
"rewards/rejected": -0.20722806453704834,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.31,
|
| 97 |
+
"learning_rate": 4.382678665009028e-06,
|
| 98 |
+
"logits/chosen": -2.5104424953460693,
|
| 99 |
+
"logits/rejected": -2.423055410385132,
|
| 100 |
+
"logps/chosen": -337.88885498046875,
|
| 101 |
+
"logps/rejected": -342.5820007324219,
|
| 102 |
+
"loss": 0.6566,
|
| 103 |
+
"rewards/accuracies": 0.609375,
|
| 104 |
+
"rewards/chosen": -0.2016713172197342,
|
| 105 |
+
"rewards/margins": 0.08087404072284698,
|
| 106 |
+
"rewards/rejected": -0.2825453579425812,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.36,
|
| 111 |
+
"learning_rate": 4.058724504646834e-06,
|
| 112 |
+
"logits/chosen": -2.437530040740967,
|
| 113 |
+
"logits/rejected": -2.411935329437256,
|
| 114 |
+
"logps/chosen": -299.39825439453125,
|
| 115 |
+
"logps/rejected": -334.86737060546875,
|
| 116 |
+
"loss": 0.6486,
|
| 117 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 118 |
+
"rewards/chosen": -0.1794266402721405,
|
| 119 |
+
"rewards/margins": 0.12244921922683716,
|
| 120 |
+
"rewards/rejected": -0.3018758296966553,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.41,
|
| 125 |
+
"learning_rate": 3.684671656182497e-06,
|
| 126 |
+
"logits/chosen": -2.494436740875244,
|
| 127 |
+
"logits/rejected": -2.417968273162842,
|
| 128 |
+
"logps/chosen": -309.20556640625,
|
| 129 |
+
"logps/rejected": -324.0342102050781,
|
| 130 |
+
"loss": 0.6481,
|
| 131 |
+
"rewards/accuracies": 0.643750011920929,
|
| 132 |
+
"rewards/chosen": -0.22928757965564728,
|
| 133 |
+
"rewards/margins": 0.13500627875328064,
|
| 134 |
+
"rewards/rejected": -0.3642938435077667,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.46,
|
| 139 |
+
"learning_rate": 3.272542485937369e-06,
|
| 140 |
+
"logits/chosen": -2.3841819763183594,
|
| 141 |
+
"logits/rejected": -2.356064558029175,
|
| 142 |
+
"logps/chosen": -289.4866638183594,
|
| 143 |
+
"logps/rejected": -311.481201171875,
|
| 144 |
+
"loss": 0.6461,
|
| 145 |
+
"rewards/accuracies": 0.590624988079071,
|
| 146 |
+
"rewards/chosen": -0.2119632214307785,
|
| 147 |
+
"rewards/margins": 0.12710149586200714,
|
| 148 |
+
"rewards/rejected": -0.33906468749046326,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.51,
|
| 153 |
+
"learning_rate": 2.835583164544139e-06,
|
| 154 |
+
"logits/chosen": -2.2960455417633057,
|
| 155 |
+
"logits/rejected": -2.2413225173950195,
|
| 156 |
+
"logps/chosen": -337.8830261230469,
|
| 157 |
+
"logps/rejected": -365.25238037109375,
|
| 158 |
+
"loss": 0.6429,
|
| 159 |
+
"rewards/accuracies": 0.6156250238418579,
|
| 160 |
+
"rewards/chosen": -0.3260457217693329,
|
| 161 |
+
"rewards/margins": 0.16088134050369263,
|
| 162 |
+
"rewards/rejected": -0.4869270920753479,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.51,
|
| 167 |
+
"eval_logits/chosen": -2.2674760818481445,
|
| 168 |
+
"eval_logits/rejected": -2.134783983230591,
|
| 169 |
+
"eval_logps/chosen": -357.25299072265625,
|
| 170 |
+
"eval_logps/rejected": -365.74114990234375,
|
| 171 |
+
"eval_loss": 0.6368065476417542,
|
| 172 |
+
"eval_rewards/accuracies": 0.6380000114440918,
|
| 173 |
+
"eval_rewards/chosen": -0.5885025858879089,
|
| 174 |
+
"eval_rewards/margins": 0.17055951058864594,
|
| 175 |
+
"eval_rewards/rejected": -0.7590621113777161,
|
| 176 |
+
"eval_runtime": 383.7605,
|
| 177 |
+
"eval_samples_per_second": 5.212,
|
| 178 |
+
"eval_steps_per_second": 0.651,
|
| 179 |
+
"step": 100
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.56,
|
| 183 |
+
"learning_rate": 2.3878379241237136e-06,
|
| 184 |
+
"logits/chosen": -2.160651445388794,
|
| 185 |
+
"logits/rejected": -2.0833230018615723,
|
| 186 |
+
"logps/chosen": -358.48834228515625,
|
| 187 |
+
"logps/rejected": -367.35345458984375,
|
| 188 |
+
"loss": 0.6391,
|
| 189 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 190 |
+
"rewards/chosen": -0.5002601146697998,
|
| 191 |
+
"rewards/margins": 0.17509731650352478,
|
| 192 |
+
"rewards/rejected": -0.675357460975647,
|
| 193 |
+
"step": 110
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.61,
|
| 197 |
+
"learning_rate": 1.9436976651092143e-06,
|
| 198 |
+
"logits/chosen": -2.167994499206543,
|
| 199 |
+
"logits/rejected": -2.064636707305908,
|
| 200 |
+
"logps/chosen": -345.025390625,
|
| 201 |
+
"logps/rejected": -360.7124938964844,
|
| 202 |
+
"loss": 0.6299,
|
| 203 |
+
"rewards/accuracies": 0.703125,
|
| 204 |
+
"rewards/chosen": -0.29811879992485046,
|
| 205 |
+
"rewards/margins": 0.1997375190258026,
|
| 206 |
+
"rewards/rejected": -0.4978563189506531,
|
| 207 |
+
"step": 120
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.67,
|
| 211 |
+
"learning_rate": 1.5174374208651913e-06,
|
| 212 |
+
"logits/chosen": -1.9607616662979126,
|
| 213 |
+
"logits/rejected": -1.7911456823349,
|
| 214 |
+
"logps/chosen": -353.13446044921875,
|
| 215 |
+
"logps/rejected": -370.7218017578125,
|
| 216 |
+
"loss": 0.609,
|
| 217 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 218 |
+
"rewards/chosen": -0.41445955634117126,
|
| 219 |
+
"rewards/margins": 0.26371151208877563,
|
| 220 |
+
"rewards/rejected": -0.6781710386276245,
|
| 221 |
+
"step": 130
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.72,
|
| 225 |
+
"learning_rate": 1.122757546369744e-06,
|
| 226 |
+
"logits/chosen": -1.9263637065887451,
|
| 227 |
+
"logits/rejected": -1.6748569011688232,
|
| 228 |
+
"logps/chosen": -393.5077209472656,
|
| 229 |
+
"logps/rejected": -390.3746643066406,
|
| 230 |
+
"loss": 0.6161,
|
| 231 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 232 |
+
"rewards/chosen": -0.4902144968509674,
|
| 233 |
+
"rewards/margins": 0.21248309314250946,
|
| 234 |
+
"rewards/rejected": -0.7026976346969604,
|
| 235 |
+
"step": 140
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.77,
|
| 239 |
+
"learning_rate": 7.723433775328385e-07,
|
| 240 |
+
"logits/chosen": -1.7820335626602173,
|
| 241 |
+
"logits/rejected": -1.6083834171295166,
|
| 242 |
+
"logps/chosen": -354.44598388671875,
|
| 243 |
+
"logps/rejected": -386.1634216308594,
|
| 244 |
+
"loss": 0.6128,
|
| 245 |
+
"rewards/accuracies": 0.65625,
|
| 246 |
+
"rewards/chosen": -0.37745195627212524,
|
| 247 |
+
"rewards/margins": 0.2488851547241211,
|
| 248 |
+
"rewards/rejected": -0.6263370513916016,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.82,
|
| 253 |
+
"learning_rate": 4.774575140626317e-07,
|
| 254 |
+
"logits/chosen": -1.6063228845596313,
|
| 255 |
+
"logits/rejected": -1.524730920791626,
|
| 256 |
+
"logps/chosen": -350.2926330566406,
|
| 257 |
+
"logps/rejected": -387.1871643066406,
|
| 258 |
+
"loss": 0.6109,
|
| 259 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 260 |
+
"rewards/chosen": -0.4679895043373108,
|
| 261 |
+
"rewards/margins": 0.23797926306724548,
|
| 262 |
+
"rewards/rejected": -0.7059687376022339,
|
| 263 |
+
"step": 160
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.87,
|
| 267 |
+
"learning_rate": 2.4757783024395244e-07,
|
| 268 |
+
"logits/chosen": -1.4787060022354126,
|
| 269 |
+
"logits/rejected": -1.4544020891189575,
|
| 270 |
+
"logps/chosen": -342.1855773925781,
|
| 271 |
+
"logps/rejected": -382.0643615722656,
|
| 272 |
+
"loss": 0.62,
|
| 273 |
+
"rewards/accuracies": 0.671875,
|
| 274 |
+
"rewards/chosen": -0.46587371826171875,
|
| 275 |
+
"rewards/margins": 0.2605627179145813,
|
| 276 |
+
"rewards/rejected": -0.7264364361763,
|
| 277 |
+
"step": 170
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.92,
|
| 281 |
+
"learning_rate": 9.00928482603669e-08,
|
| 282 |
+
"logits/chosen": -1.6203199625015259,
|
| 283 |
+
"logits/rejected": -1.3692257404327393,
|
| 284 |
+
"logps/chosen": -360.90191650390625,
|
| 285 |
+
"logps/rejected": -369.82427978515625,
|
| 286 |
+
"loss": 0.6112,
|
| 287 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 288 |
+
"rewards/chosen": -0.4321191906929016,
|
| 289 |
+
"rewards/margins": 0.2534908354282379,
|
| 290 |
+
"rewards/rejected": -0.6856100559234619,
|
| 291 |
+
"step": 180
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.97,
|
| 295 |
+
"learning_rate": 1.006426501190233e-08,
|
| 296 |
+
"logits/chosen": -1.5687239170074463,
|
| 297 |
+
"logits/rejected": -1.3249285221099854,
|
| 298 |
+
"logps/chosen": -354.10711669921875,
|
| 299 |
+
"logps/rejected": -371.7186279296875,
|
| 300 |
+
"loss": 0.6151,
|
| 301 |
+
"rewards/accuracies": 0.653124988079071,
|
| 302 |
+
"rewards/chosen": -0.44191044569015503,
|
| 303 |
+
"rewards/margins": 0.21911220252513885,
|
| 304 |
+
"rewards/rejected": -0.6610226035118103,
|
| 305 |
+
"step": 190
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.0,
|
| 309 |
+
"step": 195,
|
| 310 |
+
"total_flos": 0.0,
|
| 311 |
+
"train_loss": 0.6412642754041231,
|
| 312 |
+
"train_runtime": 8982.6903,
|
| 313 |
+
"train_samples_per_second": 2.783,
|
| 314 |
+
"train_steps_per_second": 0.022
|
| 315 |
+
}
|
| 316 |
+
],
|
| 317 |
+
"logging_steps": 10,
|
| 318 |
+
"max_steps": 195,
|
| 319 |
+
"num_input_tokens_seen": 0,
|
| 320 |
+
"num_train_epochs": 1,
|
| 321 |
+
"save_steps": 100,
|
| 322 |
+
"total_flos": 0.0,
|
| 323 |
+
"train_batch_size": 2,
|
| 324 |
+
"trial_name": null,
|
| 325 |
+
"trial_params": null
|
| 326 |
+
}
|