Commit
•
020d6ef
1
Parent(s):
d191cc5
Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +417 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: rlced-conifer-zephyr-7b-dpo-full
|
| 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 |
+
# rlced-conifer-zephyr-7b-dpo-full
|
| 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 an unknown dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-07
|
| 39 |
+
- train_batch_size: 16
|
| 40 |
+
- eval_batch_size: 16
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- gradient_accumulation_steps: 8
|
| 45 |
+
- total_train_batch_size: 512
|
| 46 |
+
- total_eval_batch_size: 64
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 1
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- Transformers 4.44.1
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Datasets 2.21.0
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9963674104826155,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.250737202167511,
|
| 5 |
+
"train_runtime": 29028.2374,
|
| 6 |
+
"train_samples": 123309,
|
| 7 |
+
"train_samples_per_second": 4.248,
|
| 8 |
+
"train_steps_per_second": 0.008
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.1"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9963674104826155,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.250737202167511,
|
| 5 |
+
"train_runtime": 29028.2374,
|
| 6 |
+
"train_samples": 123309,
|
| 7 |
+
"train_samples_per_second": 4.248,
|
| 8 |
+
"train_steps_per_second": 0.008
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9963674104826155,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 240,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0041515308770108976,
|
| 13 |
+
"grad_norm": 6.2112270362028115,
|
| 14 |
+
"learning_rate": 2.083333333333333e-08,
|
| 15 |
+
"logits/chosen": -2.6958627700805664,
|
| 16 |
+
"logits/rejected": -2.650068759918213,
|
| 17 |
+
"logps/chosen": -403.2443542480469,
|
| 18 |
+
"logps/rejected": -397.0637512207031,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04151530877010898,
|
| 28 |
+
"grad_norm": 6.539288745411817,
|
| 29 |
+
"learning_rate": 2.0833333333333333e-07,
|
| 30 |
+
"logits/chosen": -2.690112352371216,
|
| 31 |
+
"logits/rejected": -2.6779024600982666,
|
| 32 |
+
"logps/chosen": -365.9743957519531,
|
| 33 |
+
"logps/rejected": -393.799560546875,
|
| 34 |
+
"loss": 0.6925,
|
| 35 |
+
"rewards/accuracies": 0.5303819179534912,
|
| 36 |
+
"rewards/chosen": 0.0005260682082735002,
|
| 37 |
+
"rewards/margins": 0.0014649044023826718,
|
| 38 |
+
"rewards/rejected": -0.0009388362523168325,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.08303061754021795,
|
| 43 |
+
"grad_norm": 6.876570939434779,
|
| 44 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 45 |
+
"logits/chosen": -2.6719064712524414,
|
| 46 |
+
"logits/rejected": -2.6511917114257812,
|
| 47 |
+
"logps/chosen": -357.5186462402344,
|
| 48 |
+
"logps/rejected": -380.5492858886719,
|
| 49 |
+
"loss": 0.6749,
|
| 50 |
+
"rewards/accuracies": 0.848437488079071,
|
| 51 |
+
"rewards/chosen": 0.018601149320602417,
|
| 52 |
+
"rewards/margins": 0.03796255216002464,
|
| 53 |
+
"rewards/rejected": -0.019361400976777077,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.12454592631032693,
|
| 58 |
+
"grad_norm": 8.242981852125217,
|
| 59 |
+
"learning_rate": 4.990486745229364e-07,
|
| 60 |
+
"logits/chosen": -2.653783082962036,
|
| 61 |
+
"logits/rejected": -2.639296293258667,
|
| 62 |
+
"logps/chosen": -355.69207763671875,
|
| 63 |
+
"logps/rejected": -415.68768310546875,
|
| 64 |
+
"loss": 0.5772,
|
| 65 |
+
"rewards/accuracies": 0.867968738079071,
|
| 66 |
+
"rewards/chosen": 0.0034460537135601044,
|
| 67 |
+
"rewards/margins": 0.2814808487892151,
|
| 68 |
+
"rewards/rejected": -0.2780347764492035,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.1660612350804359,
|
| 73 |
+
"grad_norm": 14.409074768095572,
|
| 74 |
+
"learning_rate": 4.932612176449559e-07,
|
| 75 |
+
"logits/chosen": -2.6512644290924072,
|
| 76 |
+
"logits/rejected": -2.613352060317993,
|
| 77 |
+
"logps/chosen": -496.20770263671875,
|
| 78 |
+
"logps/rejected": -619.66943359375,
|
| 79 |
+
"loss": 0.4041,
|
| 80 |
+
"rewards/accuracies": 0.848437488079071,
|
| 81 |
+
"rewards/chosen": -1.1875219345092773,
|
| 82 |
+
"rewards/margins": 1.0682841539382935,
|
| 83 |
+
"rewards/rejected": -2.2558062076568604,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.2075765438505449,
|
| 88 |
+
"grad_norm": 14.911382681454745,
|
| 89 |
+
"learning_rate": 4.823368810567056e-07,
|
| 90 |
+
"logits/chosen": -2.64001202583313,
|
| 91 |
+
"logits/rejected": -2.6261935234069824,
|
| 92 |
+
"logps/chosen": -600.037109375,
|
| 93 |
+
"logps/rejected": -856.2901611328125,
|
| 94 |
+
"loss": 0.3072,
|
| 95 |
+
"rewards/accuracies": 0.86328125,
|
| 96 |
+
"rewards/chosen": -2.279493570327759,
|
| 97 |
+
"rewards/margins": 2.3056600093841553,
|
| 98 |
+
"rewards/rejected": -4.5851545333862305,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.24909185262065386,
|
| 103 |
+
"grad_norm": 18.472917689948186,
|
| 104 |
+
"learning_rate": 4.6650635094610966e-07,
|
| 105 |
+
"logits/chosen": -2.5924713611602783,
|
| 106 |
+
"logits/rejected": -2.579529285430908,
|
| 107 |
+
"logps/chosen": -587.2591552734375,
|
| 108 |
+
"logps/rejected": -904.9329833984375,
|
| 109 |
+
"loss": 0.2658,
|
| 110 |
+
"rewards/accuracies": 0.88671875,
|
| 111 |
+
"rewards/chosen": -2.1298751831054688,
|
| 112 |
+
"rewards/margins": 2.945924758911133,
|
| 113 |
+
"rewards/rejected": -5.07580041885376,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.29060716139076287,
|
| 118 |
+
"grad_norm": 12.678464521005063,
|
| 119 |
+
"learning_rate": 4.461039162298939e-07,
|
| 120 |
+
"logits/chosen": -2.2207727432250977,
|
| 121 |
+
"logits/rejected": -1.98198664188385,
|
| 122 |
+
"logps/chosen": -591.1898803710938,
|
| 123 |
+
"logps/rejected": -989.9195556640625,
|
| 124 |
+
"loss": 0.2276,
|
| 125 |
+
"rewards/accuracies": 0.9046875238418579,
|
| 126 |
+
"rewards/chosen": -2.155735969543457,
|
| 127 |
+
"rewards/margins": 3.7345290184020996,
|
| 128 |
+
"rewards/rejected": -5.890264987945557,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.3321224701608718,
|
| 133 |
+
"grad_norm": 14.772965954120863,
|
| 134 |
+
"learning_rate": 4.2156040946718343e-07,
|
| 135 |
+
"logits/chosen": -1.5045950412750244,
|
| 136 |
+
"logits/rejected": -0.8392450213432312,
|
| 137 |
+
"logps/chosen": -616.548828125,
|
| 138 |
+
"logps/rejected": -1064.140869140625,
|
| 139 |
+
"loss": 0.1988,
|
| 140 |
+
"rewards/accuracies": 0.914843738079071,
|
| 141 |
+
"rewards/chosen": -2.4090540409088135,
|
| 142 |
+
"rewards/margins": 4.233697891235352,
|
| 143 |
+
"rewards/rejected": -6.642751216888428,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.3736377789309808,
|
| 148 |
+
"grad_norm": 16.72422342704544,
|
| 149 |
+
"learning_rate": 3.933941090877615e-07,
|
| 150 |
+
"logits/chosen": -1.3608930110931396,
|
| 151 |
+
"logits/rejected": -0.5576863288879395,
|
| 152 |
+
"logps/chosen": -595.0042724609375,
|
| 153 |
+
"logps/rejected": -1064.654052734375,
|
| 154 |
+
"loss": 0.2026,
|
| 155 |
+
"rewards/accuracies": 0.91015625,
|
| 156 |
+
"rewards/chosen": -2.351076126098633,
|
| 157 |
+
"rewards/margins": 4.467952728271484,
|
| 158 |
+
"rewards/rejected": -6.819028377532959,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.4151530877010898,
|
| 163 |
+
"grad_norm": 9.743291843330928,
|
| 164 |
+
"learning_rate": 3.6219979505011555e-07,
|
| 165 |
+
"logits/chosen": -0.8342965245246887,
|
| 166 |
+
"logits/rejected": -0.012925502844154835,
|
| 167 |
+
"logps/chosen": -646.4923706054688,
|
| 168 |
+
"logps/rejected": -1148.498779296875,
|
| 169 |
+
"loss": 0.1867,
|
| 170 |
+
"rewards/accuracies": 0.9195312261581421,
|
| 171 |
+
"rewards/chosen": -2.8086471557617188,
|
| 172 |
+
"rewards/margins": 4.717282295227051,
|
| 173 |
+
"rewards/rejected": -7.525929927825928,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.4566683964711988,
|
| 178 |
+
"grad_norm": 11.81658995353539,
|
| 179 |
+
"learning_rate": 3.286361890379034e-07,
|
| 180 |
+
"logits/chosen": -0.1246568113565445,
|
| 181 |
+
"logits/rejected": 0.6437393426895142,
|
| 182 |
+
"logps/chosen": -648.837890625,
|
| 183 |
+
"logps/rejected": -1166.3304443359375,
|
| 184 |
+
"loss": 0.1785,
|
| 185 |
+
"rewards/accuracies": 0.917187511920929,
|
| 186 |
+
"rewards/chosen": -2.7460227012634277,
|
| 187 |
+
"rewards/margins": 4.996233940124512,
|
| 188 |
+
"rewards/rejected": -7.742256164550781,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.49818370524130773,
|
| 193 |
+
"grad_norm": 9.286297855632633,
|
| 194 |
+
"learning_rate": 2.934120444167326e-07,
|
| 195 |
+
"logits/chosen": -0.19480545818805695,
|
| 196 |
+
"logits/rejected": 0.7439680695533752,
|
| 197 |
+
"logps/chosen": -654.71875,
|
| 198 |
+
"logps/rejected": -1212.785888671875,
|
| 199 |
+
"loss": 0.1786,
|
| 200 |
+
"rewards/accuracies": 0.91796875,
|
| 201 |
+
"rewards/chosen": -2.892089366912842,
|
| 202 |
+
"rewards/margins": 5.356635093688965,
|
| 203 |
+
"rewards/rejected": -8.248723983764648,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.5396990140114167,
|
| 208 |
+
"grad_norm": 12.2921471616094,
|
| 209 |
+
"learning_rate": 2.5727117968577785e-07,
|
| 210 |
+
"logits/chosen": -0.4111465513706207,
|
| 211 |
+
"logits/rejected": 0.8093876838684082,
|
| 212 |
+
"logps/chosen": -645.12109375,
|
| 213 |
+
"logps/rejected": -1192.375244140625,
|
| 214 |
+
"loss": 0.1676,
|
| 215 |
+
"rewards/accuracies": 0.9234374761581421,
|
| 216 |
+
"rewards/chosen": -2.742527961730957,
|
| 217 |
+
"rewards/margins": 5.263998031616211,
|
| 218 |
+
"rewards/rejected": -8.006526947021484,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.5812143227815257,
|
| 223 |
+
"grad_norm": 17.430656596120464,
|
| 224 |
+
"learning_rate": 2.209767714686924e-07,
|
| 225 |
+
"logits/chosen": -0.029587041586637497,
|
| 226 |
+
"logits/rejected": 1.3743274211883545,
|
| 227 |
+
"logps/chosen": -648.7828979492188,
|
| 228 |
+
"logps/rejected": -1265.0289306640625,
|
| 229 |
+
"loss": 0.1632,
|
| 230 |
+
"rewards/accuracies": 0.925000011920929,
|
| 231 |
+
"rewards/chosen": -2.7970809936523438,
|
| 232 |
+
"rewards/margins": 5.923464775085449,
|
| 233 |
+
"rewards/rejected": -8.720545768737793,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.6227296315516346,
|
| 238 |
+
"grad_norm": 9.425394796778923,
|
| 239 |
+
"learning_rate": 1.8529523872436977e-07,
|
| 240 |
+
"logits/chosen": -0.18061885237693787,
|
| 241 |
+
"logits/rejected": 1.3279445171356201,
|
| 242 |
+
"logps/chosen": -643.9744873046875,
|
| 243 |
+
"logps/rejected": -1251.9317626953125,
|
| 244 |
+
"loss": 0.1642,
|
| 245 |
+
"rewards/accuracies": 0.9273437261581421,
|
| 246 |
+
"rewards/chosen": -2.7806482315063477,
|
| 247 |
+
"rewards/margins": 5.8400492668151855,
|
| 248 |
+
"rewards/rejected": -8.620697975158691,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.6642449403217436,
|
| 253 |
+
"grad_norm": 8.084677706157798,
|
| 254 |
+
"learning_rate": 1.5098005849021078e-07,
|
| 255 |
+
"logits/chosen": -0.20541512966156006,
|
| 256 |
+
"logits/rejected": 1.3262333869934082,
|
| 257 |
+
"logps/chosen": -645.2593994140625,
|
| 258 |
+
"logps/rejected": -1234.405517578125,
|
| 259 |
+
"loss": 0.1618,
|
| 260 |
+
"rewards/accuracies": 0.907031238079071,
|
| 261 |
+
"rewards/chosen": -2.804506301879883,
|
| 262 |
+
"rewards/margins": 5.688388347625732,
|
| 263 |
+
"rewards/rejected": -8.492895126342773,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.7057602490918526,
|
| 268 |
+
"grad_norm": 10.460249704683227,
|
| 269 |
+
"learning_rate": 1.1875585491635998e-07,
|
| 270 |
+
"logits/chosen": -0.6104969382286072,
|
| 271 |
+
"logits/rejected": 1.1572777032852173,
|
| 272 |
+
"logps/chosen": -623.3721923828125,
|
| 273 |
+
"logps/rejected": -1259.294189453125,
|
| 274 |
+
"loss": 0.1573,
|
| 275 |
+
"rewards/accuracies": 0.9203125238418579,
|
| 276 |
+
"rewards/chosen": -2.5536274909973145,
|
| 277 |
+
"rewards/margins": 6.065129280090332,
|
| 278 |
+
"rewards/rejected": -8.618757247924805,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.7472755578619616,
|
| 283 |
+
"grad_norm": 12.223705856186251,
|
| 284 |
+
"learning_rate": 8.930309757836516e-08,
|
| 285 |
+
"logits/chosen": -0.717302680015564,
|
| 286 |
+
"logits/rejected": 1.1799119710922241,
|
| 287 |
+
"logps/chosen": -645.1275634765625,
|
| 288 |
+
"logps/rejected": -1283.7489013671875,
|
| 289 |
+
"loss": 0.1616,
|
| 290 |
+
"rewards/accuracies": 0.934374988079071,
|
| 291 |
+
"rewards/chosen": -2.7811636924743652,
|
| 292 |
+
"rewards/margins": 6.2163591384887695,
|
| 293 |
+
"rewards/rejected": -8.997522354125977,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.7887908666320705,
|
| 298 |
+
"grad_norm": 9.676735610238229,
|
| 299 |
+
"learning_rate": 6.324373218975104e-08,
|
| 300 |
+
"logits/chosen": -0.8794542551040649,
|
| 301 |
+
"logits/rejected": 0.99409419298172,
|
| 302 |
+
"logps/chosen": -619.7286987304688,
|
| 303 |
+
"logps/rejected": -1276.9617919921875,
|
| 304 |
+
"loss": 0.1608,
|
| 305 |
+
"rewards/accuracies": 0.925000011920929,
|
| 306 |
+
"rewards/chosen": -2.615447521209717,
|
| 307 |
+
"rewards/margins": 6.301668643951416,
|
| 308 |
+
"rewards/rejected": -8.917116165161133,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.8303061754021795,
|
| 313 |
+
"grad_norm": 11.57113935464218,
|
| 314 |
+
"learning_rate": 4.112804714676593e-08,
|
| 315 |
+
"logits/chosen": -0.7840622663497925,
|
| 316 |
+
"logits/rejected": 1.0056122541427612,
|
| 317 |
+
"logps/chosen": -637.8081665039062,
|
| 318 |
+
"logps/rejected": -1261.2838134765625,
|
| 319 |
+
"loss": 0.162,
|
| 320 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 321 |
+
"rewards/chosen": -2.6473495960235596,
|
| 322 |
+
"rewards/margins": 6.003285884857178,
|
| 323 |
+
"rewards/rejected": -8.650635719299316,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.8718214841722886,
|
| 328 |
+
"grad_norm": 13.665084975047886,
|
| 329 |
+
"learning_rate": 2.3423053240837514e-08,
|
| 330 |
+
"logits/chosen": -0.6102296710014343,
|
| 331 |
+
"logits/rejected": 1.163267731666565,
|
| 332 |
+
"logps/chosen": -638.1444091796875,
|
| 333 |
+
"logps/rejected": -1243.947265625,
|
| 334 |
+
"loss": 0.1576,
|
| 335 |
+
"rewards/accuracies": 0.918749988079071,
|
| 336 |
+
"rewards/chosen": -2.7065281867980957,
|
| 337 |
+
"rewards/margins": 5.835700988769531,
|
| 338 |
+
"rewards/rejected": -8.542229652404785,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.9133367929423976,
|
| 343 |
+
"grad_norm": 10.802886491160189,
|
| 344 |
+
"learning_rate": 1.0502621921127774e-08,
|
| 345 |
+
"logits/chosen": -0.5309673547744751,
|
| 346 |
+
"logits/rejected": 1.326080560684204,
|
| 347 |
+
"logps/chosen": -637.7506713867188,
|
| 348 |
+
"logps/rejected": -1279.3565673828125,
|
| 349 |
+
"loss": 0.1579,
|
| 350 |
+
"rewards/accuracies": 0.922656238079071,
|
| 351 |
+
"rewards/chosen": -2.6862006187438965,
|
| 352 |
+
"rewards/margins": 6.1917243003845215,
|
| 353 |
+
"rewards/rejected": -8.877924919128418,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.9548521017125065,
|
| 358 |
+
"grad_norm": 10.752955125049585,
|
| 359 |
+
"learning_rate": 2.639590354763882e-09,
|
| 360 |
+
"logits/chosen": -0.5724472403526306,
|
| 361 |
+
"logits/rejected": 1.2889587879180908,
|
| 362 |
+
"logps/chosen": -629.98779296875,
|
| 363 |
+
"logps/rejected": -1255.991943359375,
|
| 364 |
+
"loss": 0.1579,
|
| 365 |
+
"rewards/accuracies": 0.94140625,
|
| 366 |
+
"rewards/chosen": -2.661499500274658,
|
| 367 |
+
"rewards/margins": 6.085452556610107,
|
| 368 |
+
"rewards/rejected": -8.746953010559082,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.9963674104826155,
|
| 373 |
+
"grad_norm": 9.420585318893654,
|
| 374 |
+
"learning_rate": 0.0,
|
| 375 |
+
"logits/chosen": -0.5639629364013672,
|
| 376 |
+
"logits/rejected": 1.3717336654663086,
|
| 377 |
+
"logps/chosen": -631.6014404296875,
|
| 378 |
+
"logps/rejected": -1273.350341796875,
|
| 379 |
+
"loss": 0.1514,
|
| 380 |
+
"rewards/accuracies": 0.934374988079071,
|
| 381 |
+
"rewards/chosen": -2.7341790199279785,
|
| 382 |
+
"rewards/margins": 6.176712989807129,
|
| 383 |
+
"rewards/rejected": -8.91089153289795,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.9963674104826155,
|
| 388 |
+
"step": 240,
|
| 389 |
+
"total_flos": 0.0,
|
| 390 |
+
"train_loss": 0.250737202167511,
|
| 391 |
+
"train_runtime": 29028.2374,
|
| 392 |
+
"train_samples_per_second": 4.248,
|
| 393 |
+
"train_steps_per_second": 0.008
|
| 394 |
+
}
|
| 395 |
+
],
|
| 396 |
+
"logging_steps": 10,
|
| 397 |
+
"max_steps": 240,
|
| 398 |
+
"num_input_tokens_seen": 0,
|
| 399 |
+
"num_train_epochs": 1,
|
| 400 |
+
"save_steps": 100,
|
| 401 |
+
"stateful_callbacks": {
|
| 402 |
+
"TrainerControl": {
|
| 403 |
+
"args": {
|
| 404 |
+
"should_epoch_stop": false,
|
| 405 |
+
"should_evaluate": false,
|
| 406 |
+
"should_log": false,
|
| 407 |
+
"should_save": true,
|
| 408 |
+
"should_training_stop": true
|
| 409 |
+
},
|
| 410 |
+
"attributes": {}
|
| 411 |
+
}
|
| 412 |
+
},
|
| 413 |
+
"total_flos": 0.0,
|
| 414 |
+
"train_batch_size": 16,
|
| 415 |
+
"trial_name": null,
|
| 416 |
+
"trial_params": null
|
| 417 |
+
}
|