Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- generation_config.json +8 -0
- train_results.json +9 -0
- trainer_state.json +447 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: gemma
|
| 4 |
+
base_model: google/gemma-2-9b-it
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- simpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: simpo-train-large-wrong
|
| 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 |
+
# simpo-train-large-wrong
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) 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: 8e-07
|
| 39 |
+
- train_batch_size: 2
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- gradient_accumulation_steps: 16
|
| 45 |
+
- total_train_batch_size: 128
|
| 46 |
+
- total_eval_batch_size: 16
|
| 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 4.44.2
|
| 59 |
+
- Pytorch 2.7.0+cu128
|
| 60 |
+
- Datasets 2.18.0
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.995910949568378,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 5.239574314033898,
|
| 5 |
+
"train_runtime": 2514.0727,
|
| 6 |
+
"train_samples": 17605,
|
| 7 |
+
"train_samples_per_second": 7.003,
|
| 8 |
+
"train_steps_per_second": 0.054
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 2,
|
| 4 |
+
"cache_implementation": "hybrid",
|
| 5 |
+
"eos_token_id": 1,
|
| 6 |
+
"pad_token_id": 0,
|
| 7 |
+
"transformers_version": "4.44.2"
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.995910949568378,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 5.239574314033898,
|
| 5 |
+
"train_runtime": 2514.0727,
|
| 6 |
+
"train_samples": 17605,
|
| 7 |
+
"train_samples_per_second": 7.003,
|
| 8 |
+
"train_steps_per_second": 0.054
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,447 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.995910949568378,
|
| 5 |
+
"eval_steps": 400,
|
| 6 |
+
"global_step": 137,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03634711494775102,
|
| 13 |
+
"grad_norm": 242.63721380836233,
|
| 14 |
+
"learning_rate": 2.857142857142857e-07,
|
| 15 |
+
"logits/chosen": -9.921992301940918,
|
| 16 |
+
"logits/rejected": -10.590700149536133,
|
| 17 |
+
"logps/chosen": -1.3840388059616089,
|
| 18 |
+
"logps/rejected": -0.9648998379707336,
|
| 19 |
+
"loss": 9.0059,
|
| 20 |
+
"rewards/accuracies": 0.18125000596046448,
|
| 21 |
+
"rewards/chosen": -13.840387344360352,
|
| 22 |
+
"rewards/margins": -4.191389560699463,
|
| 23 |
+
"rewards/rejected": -9.648998260498047,
|
| 24 |
+
"step": 5
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.07269422989550205,
|
| 28 |
+
"grad_norm": 125.7278409072736,
|
| 29 |
+
"learning_rate": 5.714285714285714e-07,
|
| 30 |
+
"logits/chosen": -9.401021957397461,
|
| 31 |
+
"logits/rejected": -10.126498222351074,
|
| 32 |
+
"logps/chosen": -1.224568486213684,
|
| 33 |
+
"logps/rejected": -0.8530822992324829,
|
| 34 |
+
"loss": 8.149,
|
| 35 |
+
"rewards/accuracies": 0.19374999403953552,
|
| 36 |
+
"rewards/chosen": -12.245683670043945,
|
| 37 |
+
"rewards/margins": -3.714862108230591,
|
| 38 |
+
"rewards/rejected": -8.530821800231934,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.10904134484325306,
|
| 43 |
+
"grad_norm": 118.11240526235882,
|
| 44 |
+
"learning_rate": 7.998695344323425e-07,
|
| 45 |
+
"logits/chosen": -8.648290634155273,
|
| 46 |
+
"logits/rejected": -9.154404640197754,
|
| 47 |
+
"logps/chosen": -1.1302968263626099,
|
| 48 |
+
"logps/rejected": -0.8277978897094727,
|
| 49 |
+
"loss": 8.183,
|
| 50 |
+
"rewards/accuracies": 0.16875000298023224,
|
| 51 |
+
"rewards/chosen": -11.30296802520752,
|
| 52 |
+
"rewards/margins": -3.024988889694214,
|
| 53 |
+
"rewards/rejected": -8.277978897094727,
|
| 54 |
+
"step": 15
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.1453884597910041,
|
| 58 |
+
"grad_norm": 76.19191775941484,
|
| 59 |
+
"learning_rate": 7.953121695121394e-07,
|
| 60 |
+
"logits/chosen": -7.632864475250244,
|
| 61 |
+
"logits/rejected": -8.276620864868164,
|
| 62 |
+
"logps/chosen": -0.7727211713790894,
|
| 63 |
+
"logps/rejected": -0.6470556259155273,
|
| 64 |
+
"loss": 6.3176,
|
| 65 |
+
"rewards/accuracies": 0.21875,
|
| 66 |
+
"rewards/chosen": -7.727211952209473,
|
| 67 |
+
"rewards/margins": -1.2566553354263306,
|
| 68 |
+
"rewards/rejected": -6.470555782318115,
|
| 69 |
+
"step": 20
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18173557473875512,
|
| 73 |
+
"grad_norm": 72.77010466261328,
|
| 74 |
+
"learning_rate": 7.843163833184991e-07,
|
| 75 |
+
"logits/chosen": -6.38211727142334,
|
| 76 |
+
"logits/rejected": -6.964330196380615,
|
| 77 |
+
"logps/chosen": -0.7217488288879395,
|
| 78 |
+
"logps/rejected": -0.6721780896186829,
|
| 79 |
+
"loss": 5.6946,
|
| 80 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 81 |
+
"rewards/chosen": -7.217488765716553,
|
| 82 |
+
"rewards/margins": -0.49570804834365845,
|
| 83 |
+
"rewards/rejected": -6.721780300140381,
|
| 84 |
+
"step": 25
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.21808268968650613,
|
| 88 |
+
"grad_norm": 50.91204324013054,
|
| 89 |
+
"learning_rate": 7.670612634414511e-07,
|
| 90 |
+
"logits/chosen": -5.789036273956299,
|
| 91 |
+
"logits/rejected": -6.50485897064209,
|
| 92 |
+
"logps/chosen": -0.6984373331069946,
|
| 93 |
+
"logps/rejected": -0.6544702649116516,
|
| 94 |
+
"loss": 5.5012,
|
| 95 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 96 |
+
"rewards/chosen": -6.984373569488525,
|
| 97 |
+
"rewards/margins": -0.43967121839523315,
|
| 98 |
+
"rewards/rejected": -6.544702053070068,
|
| 99 |
+
"step": 30
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.25442980463425713,
|
| 103 |
+
"grad_norm": 68.27173982618704,
|
| 104 |
+
"learning_rate": 7.438278427948805e-07,
|
| 105 |
+
"logits/chosen": -6.224788665771484,
|
| 106 |
+
"logits/rejected": -6.794752597808838,
|
| 107 |
+
"logps/chosen": -0.6720889806747437,
|
| 108 |
+
"logps/rejected": -0.6537488698959351,
|
| 109 |
+
"loss": 5.2665,
|
| 110 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 111 |
+
"rewards/chosen": -6.720890045166016,
|
| 112 |
+
"rewards/margins": -0.18340133130550385,
|
| 113 |
+
"rewards/rejected": -6.537489414215088,
|
| 114 |
+
"step": 35
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.2907769195820082,
|
| 118 |
+
"grad_norm": 52.462578532583635,
|
| 119 |
+
"learning_rate": 7.149945224533862e-07,
|
| 120 |
+
"logits/chosen": -5.9340128898620605,
|
| 121 |
+
"logits/rejected": -6.362427234649658,
|
| 122 |
+
"logps/chosen": -0.6735531091690063,
|
| 123 |
+
"logps/rejected": -0.712183952331543,
|
| 124 |
+
"loss": 5.0443,
|
| 125 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 126 |
+
"rewards/chosen": -6.735530853271484,
|
| 127 |
+
"rewards/margins": 0.3863077461719513,
|
| 128 |
+
"rewards/rejected": -7.1218390464782715,
|
| 129 |
+
"step": 40
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.3271240345297592,
|
| 133 |
+
"grad_norm": 49.32067133417458,
|
| 134 |
+
"learning_rate": 6.810309086608129e-07,
|
| 135 |
+
"logits/chosen": -6.20428466796875,
|
| 136 |
+
"logits/rejected": -6.80216121673584,
|
| 137 |
+
"logps/chosen": -0.6294387578964233,
|
| 138 |
+
"logps/rejected": -0.7196067571640015,
|
| 139 |
+
"loss": 5.0107,
|
| 140 |
+
"rewards/accuracies": 0.375,
|
| 141 |
+
"rewards/chosen": -6.294387340545654,
|
| 142 |
+
"rewards/margins": 0.9016799926757812,
|
| 143 |
+
"rewards/rejected": -7.196066856384277,
|
| 144 |
+
"step": 45
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.36347114947751025,
|
| 148 |
+
"grad_norm": 58.79656712268972,
|
| 149 |
+
"learning_rate": 6.424901643866552e-07,
|
| 150 |
+
"logits/chosen": -5.74347448348999,
|
| 151 |
+
"logits/rejected": -6.245406627655029,
|
| 152 |
+
"logps/chosen": -0.6845372319221497,
|
| 153 |
+
"logps/rejected": -0.7552808523178101,
|
| 154 |
+
"loss": 5.1229,
|
| 155 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 156 |
+
"rewards/chosen": -6.845372200012207,
|
| 157 |
+
"rewards/margins": 0.7074359655380249,
|
| 158 |
+
"rewards/rejected": -7.5528082847595215,
|
| 159 |
+
"step": 50
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.39981826442526125,
|
| 163 |
+
"grad_norm": 50.23542787679931,
|
| 164 |
+
"learning_rate": 6e-07,
|
| 165 |
+
"logits/chosen": -5.800081729888916,
|
| 166 |
+
"logits/rejected": -6.2036285400390625,
|
| 167 |
+
"logps/chosen": -0.5938673615455627,
|
| 168 |
+
"logps/rejected": -0.7371848821640015,
|
| 169 |
+
"loss": 4.9249,
|
| 170 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 171 |
+
"rewards/chosen": -5.938673973083496,
|
| 172 |
+
"rewards/margins": 1.4331753253936768,
|
| 173 |
+
"rewards/rejected": -7.371849060058594,
|
| 174 |
+
"step": 55
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.43616537937301225,
|
| 178 |
+
"grad_norm": 48.19402877634858,
|
| 179 |
+
"learning_rate": 5.542524497952543e-07,
|
| 180 |
+
"logits/chosen": -5.9060564041137695,
|
| 181 |
+
"logits/rejected": -6.406619071960449,
|
| 182 |
+
"logps/chosen": -0.5991919040679932,
|
| 183 |
+
"logps/rejected": -0.7929545640945435,
|
| 184 |
+
"loss": 4.6185,
|
| 185 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 186 |
+
"rewards/chosen": -5.99191951751709,
|
| 187 |
+
"rewards/margins": 1.9376264810562134,
|
| 188 |
+
"rewards/rejected": -7.9295454025268555,
|
| 189 |
+
"step": 60
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.4725124943207633,
|
| 193 |
+
"grad_norm": 57.342852083624614,
|
| 194 |
+
"learning_rate": 5.059926008786647e-07,
|
| 195 |
+
"logits/chosen": -6.52951192855835,
|
| 196 |
+
"logits/rejected": -6.941732883453369,
|
| 197 |
+
"logps/chosen": -0.5963850021362305,
|
| 198 |
+
"logps/rejected": -0.6861479878425598,
|
| 199 |
+
"loss": 4.8488,
|
| 200 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 201 |
+
"rewards/chosen": -5.963850021362305,
|
| 202 |
+
"rewards/margins": 0.8976295590400696,
|
| 203 |
+
"rewards/rejected": -6.86147928237915,
|
| 204 |
+
"step": 65
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.5088596092685143,
|
| 208 |
+
"grad_norm": 50.7899004379066,
|
| 209 |
+
"learning_rate": 4.5600645798745166e-07,
|
| 210 |
+
"logits/chosen": -6.482469081878662,
|
| 211 |
+
"logits/rejected": -6.863368988037109,
|
| 212 |
+
"logps/chosen": -0.6205894947052002,
|
| 213 |
+
"logps/rejected": -0.7471215128898621,
|
| 214 |
+
"loss": 4.7473,
|
| 215 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 216 |
+
"rewards/chosen": -6.205894470214844,
|
| 217 |
+
"rewards/margins": 1.265321135520935,
|
| 218 |
+
"rewards/rejected": -7.47121524810791,
|
| 219 |
+
"step": 70
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.5452067242162654,
|
| 223 |
+
"grad_norm": 51.764811999912986,
|
| 224 |
+
"learning_rate": 4.051081418863895e-07,
|
| 225 |
+
"logits/chosen": -5.929373264312744,
|
| 226 |
+
"logits/rejected": -6.43304443359375,
|
| 227 |
+
"logps/chosen": -0.6067383885383606,
|
| 228 |
+
"logps/rejected": -0.7068555951118469,
|
| 229 |
+
"loss": 4.6926,
|
| 230 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 231 |
+
"rewards/chosen": -6.067384243011475,
|
| 232 |
+
"rewards/margins": 1.0011712312698364,
|
| 233 |
+
"rewards/rejected": -7.068556308746338,
|
| 234 |
+
"step": 75
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.5815538391640164,
|
| 238 |
+
"grad_norm": 52.24384814108085,
|
| 239 |
+
"learning_rate": 3.541266298406398e-07,
|
| 240 |
+
"logits/chosen": -6.554868221282959,
|
| 241 |
+
"logits/rejected": -6.9319915771484375,
|
| 242 |
+
"logps/chosen": -0.5482149720191956,
|
| 243 |
+
"logps/rejected": -0.761601984500885,
|
| 244 |
+
"loss": 4.8385,
|
| 245 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 246 |
+
"rewards/chosen": -5.482149600982666,
|
| 247 |
+
"rewards/margins": 2.1338706016540527,
|
| 248 |
+
"rewards/rejected": -7.616020202636719,
|
| 249 |
+
"step": 80
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.6179009541117674,
|
| 253 |
+
"grad_norm": 49.865022637467455,
|
| 254 |
+
"learning_rate": 3.0389225412181565e-07,
|
| 255 |
+
"logits/chosen": -6.109275817871094,
|
| 256 |
+
"logits/rejected": -6.693246364593506,
|
| 257 |
+
"logps/chosen": -0.5953459739685059,
|
| 258 |
+
"logps/rejected": -0.7325998544692993,
|
| 259 |
+
"loss": 4.5675,
|
| 260 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 261 |
+
"rewards/chosen": -5.953459739685059,
|
| 262 |
+
"rewards/margins": 1.3725389242172241,
|
| 263 |
+
"rewards/rejected": -7.325998783111572,
|
| 264 |
+
"step": 85
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.6542480690595184,
|
| 268 |
+
"grad_norm": 55.385656081268344,
|
| 269 |
+
"learning_rate": 2.5522317844515273e-07,
|
| 270 |
+
"logits/chosen": -6.353059768676758,
|
| 271 |
+
"logits/rejected": -6.821288108825684,
|
| 272 |
+
"logps/chosen": -0.5759108662605286,
|
| 273 |
+
"logps/rejected": -0.710253119468689,
|
| 274 |
+
"loss": 4.5135,
|
| 275 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 276 |
+
"rewards/chosen": -5.759108066558838,
|
| 277 |
+
"rewards/margins": 1.3434231281280518,
|
| 278 |
+
"rewards/rejected": -7.102531433105469,
|
| 279 |
+
"step": 90
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.6905951840072694,
|
| 283 |
+
"grad_norm": 54.711123059457066,
|
| 284 |
+
"learning_rate": 2.0891207259509476e-07,
|
| 285 |
+
"logits/chosen": -6.575620174407959,
|
| 286 |
+
"logits/rejected": -6.845908164978027,
|
| 287 |
+
"logps/chosen": -0.6202897429466248,
|
| 288 |
+
"logps/rejected": -0.7272644639015198,
|
| 289 |
+
"loss": 4.576,
|
| 290 |
+
"rewards/accuracies": 0.4375,
|
| 291 |
+
"rewards/chosen": -6.202897548675537,
|
| 292 |
+
"rewards/margins": 1.0697463750839233,
|
| 293 |
+
"rewards/rejected": -7.27264404296875,
|
| 294 |
+
"step": 95
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.7269422989550205,
|
| 298 |
+
"grad_norm": 51.133951393232486,
|
| 299 |
+
"learning_rate": 1.6571320226872206e-07,
|
| 300 |
+
"logits/chosen": -6.456846714019775,
|
| 301 |
+
"logits/rejected": -7.050930023193359,
|
| 302 |
+
"logps/chosen": -0.5731553435325623,
|
| 303 |
+
"logps/rejected": -0.722934365272522,
|
| 304 |
+
"loss": 4.5131,
|
| 305 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 306 |
+
"rewards/chosen": -5.731553554534912,
|
| 307 |
+
"rewards/margins": 1.4977900981903076,
|
| 308 |
+
"rewards/rejected": -7.229344367980957,
|
| 309 |
+
"step": 100
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.7632894139027715,
|
| 313 |
+
"grad_norm": 51.6827958226427,
|
| 314 |
+
"learning_rate": 1.2633014440382787e-07,
|
| 315 |
+
"logits/chosen": -6.841757774353027,
|
| 316 |
+
"logits/rejected": -7.067150115966797,
|
| 317 |
+
"logps/chosen": -0.6003233194351196,
|
| 318 |
+
"logps/rejected": -0.7174679636955261,
|
| 319 |
+
"loss": 4.4139,
|
| 320 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 321 |
+
"rewards/chosen": -6.003233432769775,
|
| 322 |
+
"rewards/margins": 1.1714468002319336,
|
| 323 |
+
"rewards/rejected": -7.174680233001709,
|
| 324 |
+
"step": 105
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.7996365288505225,
|
| 328 |
+
"grad_norm": 63.215678605250034,
|
| 329 |
+
"learning_rate": 9.14043280712228e-08,
|
| 330 |
+
"logits/chosen": -6.239346504211426,
|
| 331 |
+
"logits/rejected": -6.602126121520996,
|
| 332 |
+
"logps/chosen": -0.614771842956543,
|
| 333 |
+
"logps/rejected": -0.8130232691764832,
|
| 334 |
+
"loss": 4.6241,
|
| 335 |
+
"rewards/accuracies": 0.53125,
|
| 336 |
+
"rewards/chosen": -6.147718906402588,
|
| 337 |
+
"rewards/margins": 1.9825141429901123,
|
| 338 |
+
"rewards/rejected": -8.130231857299805,
|
| 339 |
+
"step": 110
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.8359836437982735,
|
| 343 |
+
"grad_norm": 59.93834778024071,
|
| 344 |
+
"learning_rate": 6.150458756494239e-08,
|
| 345 |
+
"logits/chosen": -6.464346885681152,
|
| 346 |
+
"logits/rejected": -6.851041316986084,
|
| 347 |
+
"logps/chosen": -0.6194056272506714,
|
| 348 |
+
"logps/rejected": -0.776168704032898,
|
| 349 |
+
"loss": 4.4725,
|
| 350 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 351 |
+
"rewards/chosen": -6.194056510925293,
|
| 352 |
+
"rewards/margins": 1.567630410194397,
|
| 353 |
+
"rewards/rejected": -7.7616868019104,
|
| 354 |
+
"step": 115
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.8723307587460245,
|
| 358 |
+
"grad_norm": 63.957410261640334,
|
| 359 |
+
"learning_rate": 3.711789783843522e-08,
|
| 360 |
+
"logits/chosen": -5.914224147796631,
|
| 361 |
+
"logits/rejected": -6.372625350952148,
|
| 362 |
+
"logps/chosen": -0.5956039428710938,
|
| 363 |
+
"logps/rejected": -0.7622672319412231,
|
| 364 |
+
"loss": 4.5261,
|
| 365 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 366 |
+
"rewards/chosen": -5.9560394287109375,
|
| 367 |
+
"rewards/margins": 1.6666326522827148,
|
| 368 |
+
"rewards/rejected": -7.622672080993652,
|
| 369 |
+
"step": 120
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.9086778736937755,
|
| 373 |
+
"grad_norm": 64.67587211778695,
|
| 374 |
+
"learning_rate": 1.8641443178027784e-08,
|
| 375 |
+
"logits/chosen": -6.29254674911499,
|
| 376 |
+
"logits/rejected": -6.541565895080566,
|
| 377 |
+
"logps/chosen": -0.6281149387359619,
|
| 378 |
+
"logps/rejected": -0.7664488554000854,
|
| 379 |
+
"loss": 4.6393,
|
| 380 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 381 |
+
"rewards/chosen": -6.281149864196777,
|
| 382 |
+
"rewards/margins": 1.3833396434783936,
|
| 383 |
+
"rewards/rejected": -7.664488792419434,
|
| 384 |
+
"step": 125
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.9450249886415266,
|
| 388 |
+
"grad_norm": 53.764099269870464,
|
| 389 |
+
"learning_rate": 6.376148290617145e-09,
|
| 390 |
+
"logits/chosen": -6.527040958404541,
|
| 391 |
+
"logits/rejected": -7.045149803161621,
|
| 392 |
+
"logps/chosen": -0.6423445343971252,
|
| 393 |
+
"logps/rejected": -0.7968176603317261,
|
| 394 |
+
"loss": 4.4187,
|
| 395 |
+
"rewards/accuracies": 0.550000011920929,
|
| 396 |
+
"rewards/chosen": -6.423445701599121,
|
| 397 |
+
"rewards/margins": 1.5447312593460083,
|
| 398 |
+
"rewards/rejected": -7.968176364898682,
|
| 399 |
+
"step": 130
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.9813721035892776,
|
| 403 |
+
"grad_norm": 54.882883609182315,
|
| 404 |
+
"learning_rate": 5.217771643080127e-10,
|
| 405 |
+
"logits/chosen": -6.546690940856934,
|
| 406 |
+
"logits/rejected": -6.896824836730957,
|
| 407 |
+
"logps/chosen": -0.6224032640457153,
|
| 408 |
+
"logps/rejected": -0.7323735952377319,
|
| 409 |
+
"loss": 4.5659,
|
| 410 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 411 |
+
"rewards/chosen": -6.224032878875732,
|
| 412 |
+
"rewards/margins": 1.0997024774551392,
|
| 413 |
+
"rewards/rejected": -7.32373571395874,
|
| 414 |
+
"step": 135
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.995910949568378,
|
| 418 |
+
"step": 137,
|
| 419 |
+
"total_flos": 0.0,
|
| 420 |
+
"train_loss": 5.239574314033898,
|
| 421 |
+
"train_runtime": 2514.0727,
|
| 422 |
+
"train_samples_per_second": 7.003,
|
| 423 |
+
"train_steps_per_second": 0.054
|
| 424 |
+
}
|
| 425 |
+
],
|
| 426 |
+
"logging_steps": 5,
|
| 427 |
+
"max_steps": 137,
|
| 428 |
+
"num_input_tokens_seen": 0,
|
| 429 |
+
"num_train_epochs": 1,
|
| 430 |
+
"save_steps": 1000000,
|
| 431 |
+
"stateful_callbacks": {
|
| 432 |
+
"TrainerControl": {
|
| 433 |
+
"args": {
|
| 434 |
+
"should_epoch_stop": false,
|
| 435 |
+
"should_evaluate": false,
|
| 436 |
+
"should_log": false,
|
| 437 |
+
"should_save": true,
|
| 438 |
+
"should_training_stop": true
|
| 439 |
+
},
|
| 440 |
+
"attributes": {}
|
| 441 |
+
}
|
| 442 |
+
},
|
| 443 |
+
"total_flos": 0.0,
|
| 444 |
+
"train_batch_size": 2,
|
| 445 |
+
"trial_name": null,
|
| 446 |
+
"trial_params": null
|
| 447 |
+
}
|