Model save
Browse files- README.md +60 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- train_results.json +9 -0
- trainer_state.json +777 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: rlced-conifer-tulu-2-7b-dpo-full
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# rlced-conifer-tulu-2-7b-dpo-full
|
| 17 |
+
|
| 18 |
+
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.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-07
|
| 38 |
+
- train_batch_size: 8
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 8
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 256
|
| 45 |
+
- total_eval_batch_size: 64
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- Transformers 4.44.1
|
| 58 |
+
- Pytorch 2.1.2+cu121
|
| 59 |
+
- Datasets 2.21.0
|
| 60 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9984431759211209,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.21933725711709495,
|
| 5 |
+
"train_runtime": 13808.8776,
|
| 6 |
+
"train_samples": 123309,
|
| 7 |
+
"train_samples_per_second": 8.93,
|
| 8 |
+
"train_steps_per_second": 0.035
|
| 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 |
+
}
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4943162336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62159453a790aebabd2016cc736de009d9459341ec2bf712124d89b2cbdc83ba
|
| 3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd72e4e60522610a0d2942f1dadf1e8ccda6313cc22389fa3e5f0e54675f55df
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4540516344
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:285ac92c2b4b108723bd8e44a6394393174d7de801645aa45dc70e498248ac28
|
| 3 |
size 4540516344
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9984431759211209,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.21933725711709495,
|
| 5 |
+
"train_runtime": 13808.8776,
|
| 6 |
+
"train_samples": 123309,
|
| 7 |
+
"train_samples_per_second": 8.93,
|
| 8 |
+
"train_steps_per_second": 0.035
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,777 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9984431759211209,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 481,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0020757654385054488,
|
| 13 |
+
"grad_norm": 7.453312579539728,
|
| 14 |
+
"learning_rate": 1.020408163265306e-08,
|
| 15 |
+
"logits/chosen": -2.730942726135254,
|
| 16 |
+
"logits/rejected": -2.654609203338623,
|
| 17 |
+
"logps/chosen": -350.489990234375,
|
| 18 |
+
"logps/rejected": -325.546875,
|
| 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": 0.02075765438505449,
|
| 28 |
+
"grad_norm": 7.878888995232912,
|
| 29 |
+
"learning_rate": 1.0204081632653061e-07,
|
| 30 |
+
"logits/chosen": -2.7330236434936523,
|
| 31 |
+
"logits/rejected": -2.735116720199585,
|
| 32 |
+
"logps/chosen": -366.51531982421875,
|
| 33 |
+
"logps/rejected": -412.2677001953125,
|
| 34 |
+
"loss": 0.6931,
|
| 35 |
+
"rewards/accuracies": 0.4270833432674408,
|
| 36 |
+
"rewards/chosen": -0.000205132644623518,
|
| 37 |
+
"rewards/margins": 4.354613702162169e-05,
|
| 38 |
+
"rewards/rejected": -0.00024867875617928803,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.04151530877010898,
|
| 43 |
+
"grad_norm": 7.53294584904676,
|
| 44 |
+
"learning_rate": 2.0408163265306121e-07,
|
| 45 |
+
"logits/chosen": -2.7173304557800293,
|
| 46 |
+
"logits/rejected": -2.693912982940674,
|
| 47 |
+
"logps/chosen": -378.73748779296875,
|
| 48 |
+
"logps/rejected": -404.47003173828125,
|
| 49 |
+
"loss": 0.6892,
|
| 50 |
+
"rewards/accuracies": 0.762499988079071,
|
| 51 |
+
"rewards/chosen": 0.0025507560931146145,
|
| 52 |
+
"rewards/margins": 0.008084132336080074,
|
| 53 |
+
"rewards/rejected": -0.005533376708626747,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.062272963155163466,
|
| 58 |
+
"grad_norm": 8.007471678418003,
|
| 59 |
+
"learning_rate": 3.0612244897959183e-07,
|
| 60 |
+
"logits/chosen": -2.716646194458008,
|
| 61 |
+
"logits/rejected": -2.700786590576172,
|
| 62 |
+
"logps/chosen": -363.6639709472656,
|
| 63 |
+
"logps/rejected": -390.54083251953125,
|
| 64 |
+
"loss": 0.6692,
|
| 65 |
+
"rewards/accuracies": 0.903124988079071,
|
| 66 |
+
"rewards/chosen": 0.02478734776377678,
|
| 67 |
+
"rewards/margins": 0.051134396344423294,
|
| 68 |
+
"rewards/rejected": -0.026347041130065918,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.08303061754021795,
|
| 73 |
+
"grad_norm": 9.3622528637074,
|
| 74 |
+
"learning_rate": 4.0816326530612243e-07,
|
| 75 |
+
"logits/chosen": -2.7087109088897705,
|
| 76 |
+
"logits/rejected": -2.669712543487549,
|
| 77 |
+
"logps/chosen": -347.83538818359375,
|
| 78 |
+
"logps/rejected": -376.85260009765625,
|
| 79 |
+
"loss": 0.6044,
|
| 80 |
+
"rewards/accuracies": 0.875,
|
| 81 |
+
"rewards/chosen": 0.05488457530736923,
|
| 82 |
+
"rewards/margins": 0.19498120248317719,
|
| 83 |
+
"rewards/rejected": -0.14009663462638855,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.10378827192527244,
|
| 88 |
+
"grad_norm": 11.969966849217528,
|
| 89 |
+
"learning_rate": 4.999933894080444e-07,
|
| 90 |
+
"logits/chosen": -2.7135281562805176,
|
| 91 |
+
"logits/rejected": -2.6938090324401855,
|
| 92 |
+
"logps/chosen": -403.2617492675781,
|
| 93 |
+
"logps/rejected": -495.21270751953125,
|
| 94 |
+
"loss": 0.4674,
|
| 95 |
+
"rewards/accuracies": 0.871874988079071,
|
| 96 |
+
"rewards/chosen": -0.37011662125587463,
|
| 97 |
+
"rewards/margins": 0.6785963177680969,
|
| 98 |
+
"rewards/rejected": -1.048712968826294,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.12454592631032693,
|
| 103 |
+
"grad_norm": 19.176201042712012,
|
| 104 |
+
"learning_rate": 4.992005413014143e-07,
|
| 105 |
+
"logits/chosen": -2.7302985191345215,
|
| 106 |
+
"logits/rejected": -2.7273764610290527,
|
| 107 |
+
"logps/chosen": -528.5646362304688,
|
| 108 |
+
"logps/rejected": -741.4615478515625,
|
| 109 |
+
"loss": 0.3523,
|
| 110 |
+
"rewards/accuracies": 0.84375,
|
| 111 |
+
"rewards/chosen": -1.8138965368270874,
|
| 112 |
+
"rewards/margins": 1.7118844985961914,
|
| 113 |
+
"rewards/rejected": -3.5257811546325684,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.14530358069538143,
|
| 118 |
+
"grad_norm": 20.310776567548704,
|
| 119 |
+
"learning_rate": 4.970903776169402e-07,
|
| 120 |
+
"logits/chosen": -2.7460341453552246,
|
| 121 |
+
"logits/rejected": -2.7275753021240234,
|
| 122 |
+
"logps/chosen": -634.8268432617188,
|
| 123 |
+
"logps/rejected": -865.6439208984375,
|
| 124 |
+
"loss": 0.3052,
|
| 125 |
+
"rewards/accuracies": 0.84375,
|
| 126 |
+
"rewards/chosen": -2.382263422012329,
|
| 127 |
+
"rewards/margins": 2.2324626445770264,
|
| 128 |
+
"rewards/rejected": -4.6147260665893555,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.1660612350804359,
|
| 133 |
+
"grad_norm": 19.886871444649117,
|
| 134 |
+
"learning_rate": 4.936740530314087e-07,
|
| 135 |
+
"logits/chosen": -2.3413853645324707,
|
| 136 |
+
"logits/rejected": -2.102804660797119,
|
| 137 |
+
"logps/chosen": -591.3840942382812,
|
| 138 |
+
"logps/rejected": -896.90625,
|
| 139 |
+
"loss": 0.25,
|
| 140 |
+
"rewards/accuracies": 0.903124988079071,
|
| 141 |
+
"rewards/chosen": -2.003624200820923,
|
| 142 |
+
"rewards/margins": 2.942108631134033,
|
| 143 |
+
"rewards/rejected": -4.945733070373535,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.1868188894654904,
|
| 148 |
+
"grad_norm": 16.07417524586734,
|
| 149 |
+
"learning_rate": 4.889696268057348e-07,
|
| 150 |
+
"logits/chosen": -1.8468377590179443,
|
| 151 |
+
"logits/rejected": -1.1863611936569214,
|
| 152 |
+
"logps/chosen": -567.8921508789062,
|
| 153 |
+
"logps/rejected": -934.1871337890625,
|
| 154 |
+
"loss": 0.2254,
|
| 155 |
+
"rewards/accuracies": 0.893750011920929,
|
| 156 |
+
"rewards/chosen": -2.0263195037841797,
|
| 157 |
+
"rewards/margins": 3.4611504077911377,
|
| 158 |
+
"rewards/rejected": -5.487469673156738,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.2075765438505449,
|
| 163 |
+
"grad_norm": 15.952464662940557,
|
| 164 |
+
"learning_rate": 4.830019673206996e-07,
|
| 165 |
+
"logits/chosen": -1.3128455877304077,
|
| 166 |
+
"logits/rejected": -0.37191733717918396,
|
| 167 |
+
"logps/chosen": -637.7520751953125,
|
| 168 |
+
"logps/rejected": -1123.6968994140625,
|
| 169 |
+
"loss": 0.2034,
|
| 170 |
+
"rewards/accuracies": 0.925000011920929,
|
| 171 |
+
"rewards/chosen": -2.622469425201416,
|
| 172 |
+
"rewards/margins": 4.605704307556152,
|
| 173 |
+
"rewards/rejected": -7.22817325592041,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.2283341982355994,
|
| 178 |
+
"grad_norm": 20.61911089038355,
|
| 179 |
+
"learning_rate": 4.7580262061854606e-07,
|
| 180 |
+
"logits/chosen": -0.8984780311584473,
|
| 181 |
+
"logits/rejected": 0.00587611785158515,
|
| 182 |
+
"logps/chosen": -629.09521484375,
|
| 183 |
+
"logps/rejected": -1123.9976806640625,
|
| 184 |
+
"loss": 0.2043,
|
| 185 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 186 |
+
"rewards/chosen": -2.5760746002197266,
|
| 187 |
+
"rewards/margins": 4.712790012359619,
|
| 188 |
+
"rewards/rejected": -7.288865089416504,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.24909185262065386,
|
| 193 |
+
"grad_norm": 13.519087980837655,
|
| 194 |
+
"learning_rate": 4.674096436453447e-07,
|
| 195 |
+
"logits/chosen": -0.8746647834777832,
|
| 196 |
+
"logits/rejected": 0.03333142027258873,
|
| 197 |
+
"logps/chosen": -662.9947509765625,
|
| 198 |
+
"logps/rejected": -1107.119140625,
|
| 199 |
+
"loss": 0.1947,
|
| 200 |
+
"rewards/accuracies": 0.90625,
|
| 201 |
+
"rewards/chosen": -2.6802916526794434,
|
| 202 |
+
"rewards/margins": 4.257796287536621,
|
| 203 |
+
"rewards/rejected": -6.938088417053223,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.26984950700570837,
|
| 208 |
+
"grad_norm": 15.179032922942401,
|
| 209 |
+
"learning_rate": 4.578674030756363e-07,
|
| 210 |
+
"logits/chosen": -0.5216516256332397,
|
| 211 |
+
"logits/rejected": 0.797188401222229,
|
| 212 |
+
"logps/chosen": -672.4591064453125,
|
| 213 |
+
"logps/rejected": -1201.9649658203125,
|
| 214 |
+
"loss": 0.1766,
|
| 215 |
+
"rewards/accuracies": 0.90625,
|
| 216 |
+
"rewards/chosen": -2.92866849899292,
|
| 217 |
+
"rewards/margins": 5.020692348480225,
|
| 218 |
+
"rewards/rejected": -7.949362277984619,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.29060716139076287,
|
| 223 |
+
"grad_norm": 19.394840025974254,
|
| 224 |
+
"learning_rate": 4.4722634078279865e-07,
|
| 225 |
+
"logits/chosen": 0.05672640725970268,
|
| 226 |
+
"logits/rejected": 1.2894923686981201,
|
| 227 |
+
"logps/chosen": -632.3123779296875,
|
| 228 |
+
"logps/rejected": -1134.365478515625,
|
| 229 |
+
"loss": 0.1989,
|
| 230 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 231 |
+
"rewards/chosen": -2.6184778213500977,
|
| 232 |
+
"rewards/margins": 4.792341709136963,
|
| 233 |
+
"rewards/rejected": -7.410820007324219,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.3113648157758173,
|
| 238 |
+
"grad_norm": 15.545837682982413,
|
| 239 |
+
"learning_rate": 4.355427071949004e-07,
|
| 240 |
+
"logits/chosen": -0.034926723688840866,
|
| 241 |
+
"logits/rejected": 1.3178081512451172,
|
| 242 |
+
"logps/chosen": -625.6254272460938,
|
| 243 |
+
"logps/rejected": -1133.8699951171875,
|
| 244 |
+
"loss": 0.1657,
|
| 245 |
+
"rewards/accuracies": 0.925000011920929,
|
| 246 |
+
"rewards/chosen": -2.654139757156372,
|
| 247 |
+
"rewards/margins": 4.885876655578613,
|
| 248 |
+
"rewards/rejected": -7.540017127990723,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.3321224701608718,
|
| 253 |
+
"grad_norm": 21.099719863638594,
|
| 254 |
+
"learning_rate": 4.228782639455674e-07,
|
| 255 |
+
"logits/chosen": -0.2264009416103363,
|
| 256 |
+
"logits/rejected": 1.3679448366165161,
|
| 257 |
+
"logps/chosen": -684.0534057617188,
|
| 258 |
+
"logps/rejected": -1281.2177734375,
|
| 259 |
+
"loss": 0.1738,
|
| 260 |
+
"rewards/accuracies": 0.934374988079071,
|
| 261 |
+
"rewards/chosen": -3.0014939308166504,
|
| 262 |
+
"rewards/margins": 5.716488838195801,
|
| 263 |
+
"rewards/rejected": -8.717982292175293,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.3528801245459263,
|
| 268 |
+
"grad_norm": 14.4755316075222,
|
| 269 |
+
"learning_rate": 4.092999573916971e-07,
|
| 270 |
+
"logits/chosen": 0.14696760475635529,
|
| 271 |
+
"logits/rejected": 1.6602694988250732,
|
| 272 |
+
"logps/chosen": -664.4630126953125,
|
| 273 |
+
"logps/rejected": -1225.433837890625,
|
| 274 |
+
"loss": 0.1804,
|
| 275 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 276 |
+
"rewards/chosen": -3.056241273880005,
|
| 277 |
+
"rewards/margins": 5.417787551879883,
|
| 278 |
+
"rewards/rejected": -8.474028587341309,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.3736377789309808,
|
| 283 |
+
"grad_norm": 15.213191213812825,
|
| 284 |
+
"learning_rate": 3.948795647238637e-07,
|
| 285 |
+
"logits/chosen": -0.7323606014251709,
|
| 286 |
+
"logits/rejected": 1.1011723279953003,
|
| 287 |
+
"logps/chosen": -630.0668334960938,
|
| 288 |
+
"logps/rejected": -1234.099609375,
|
| 289 |
+
"loss": 0.1783,
|
| 290 |
+
"rewards/accuracies": 0.921875,
|
| 291 |
+
"rewards/chosen": -2.6342709064483643,
|
| 292 |
+
"rewards/margins": 5.844083786010742,
|
| 293 |
+
"rewards/rejected": -8.478353500366211,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.39439543331603527,
|
| 298 |
+
"grad_norm": 16.709219865079046,
|
| 299 |
+
"learning_rate": 3.796933145401304e-07,
|
| 300 |
+
"logits/chosen": -0.12861236929893494,
|
| 301 |
+
"logits/rejected": 1.5260117053985596,
|
| 302 |
+
"logps/chosen": -731.4463500976562,
|
| 303 |
+
"logps/rejected": -1386.452880859375,
|
| 304 |
+
"loss": 0.1646,
|
| 305 |
+
"rewards/accuracies": 0.9375,
|
| 306 |
+
"rewards/chosen": -3.487692356109619,
|
| 307 |
+
"rewards/margins": 6.185595989227295,
|
| 308 |
+
"rewards/rejected": -9.673288345336914,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.4151530877010898,
|
| 313 |
+
"grad_norm": 16.46330948268556,
|
| 314 |
+
"learning_rate": 3.638214838889801e-07,
|
| 315 |
+
"logits/chosen": 0.014425823464989662,
|
| 316 |
+
"logits/rejected": 1.5777640342712402,
|
| 317 |
+
"logps/chosen": -647.3234252929688,
|
| 318 |
+
"logps/rejected": -1219.0548095703125,
|
| 319 |
+
"loss": 0.1669,
|
| 320 |
+
"rewards/accuracies": 0.909375011920929,
|
| 321 |
+
"rewards/chosen": -2.797513484954834,
|
| 322 |
+
"rewards/margins": 5.429152011871338,
|
| 323 |
+
"rewards/rejected": -8.226665496826172,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.4359107420861443,
|
| 328 |
+
"grad_norm": 19.39279809291309,
|
| 329 |
+
"learning_rate": 3.4734797391146383e-07,
|
| 330 |
+
"logits/chosen": -0.12084762752056122,
|
| 331 |
+
"logits/rejected": 1.6122562885284424,
|
| 332 |
+
"logps/chosen": -656.5711059570312,
|
| 333 |
+
"logps/rejected": -1250.8482666015625,
|
| 334 |
+
"loss": 0.1594,
|
| 335 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 336 |
+
"rewards/chosen": -2.874763011932373,
|
| 337 |
+
"rewards/margins": 5.74980354309082,
|
| 338 |
+
"rewards/rejected": -8.624567031860352,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.4566683964711988,
|
| 343 |
+
"grad_norm": 23.517592451341034,
|
| 344 |
+
"learning_rate": 3.3035986632579036e-07,
|
| 345 |
+
"logits/chosen": -1.0772771835327148,
|
| 346 |
+
"logits/rejected": 0.6209205389022827,
|
| 347 |
+
"logps/chosen": -622.2130126953125,
|
| 348 |
+
"logps/rejected": -1293.0042724609375,
|
| 349 |
+
"loss": 0.1678,
|
| 350 |
+
"rewards/accuracies": 0.921875,
|
| 351 |
+
"rewards/chosen": -2.5699634552001953,
|
| 352 |
+
"rewards/margins": 6.527965545654297,
|
| 353 |
+
"rewards/rejected": -9.097929000854492,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.4774260508562532,
|
| 358 |
+
"grad_norm": 20.580951166094668,
|
| 359 |
+
"learning_rate": 3.1294696309885716e-07,
|
| 360 |
+
"logits/chosen": -1.1179264783859253,
|
| 361 |
+
"logits/rejected": 0.7691652178764343,
|
| 362 |
+
"logps/chosen": -666.9544677734375,
|
| 363 |
+
"logps/rejected": -1416.962158203125,
|
| 364 |
+
"loss": 0.1673,
|
| 365 |
+
"rewards/accuracies": 0.925000011920929,
|
| 366 |
+
"rewards/chosen": -3.058974027633667,
|
| 367 |
+
"rewards/margins": 7.343924522399902,
|
| 368 |
+
"rewards/rejected": -10.402898788452148,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.49818370524130773,
|
| 373 |
+
"grad_norm": 14.600883148928759,
|
| 374 |
+
"learning_rate": 2.952013117380913e-07,
|
| 375 |
+
"logits/chosen": -0.9207614660263062,
|
| 376 |
+
"logits/rejected": 1.1953575611114502,
|
| 377 |
+
"logps/chosen": -662.4710693359375,
|
| 378 |
+
"logps/rejected": -1418.23291015625,
|
| 379 |
+
"loss": 0.1614,
|
| 380 |
+
"rewards/accuracies": 0.949999988079071,
|
| 381 |
+
"rewards/chosen": -3.0431346893310547,
|
| 382 |
+
"rewards/margins": 7.319464206695557,
|
| 383 |
+
"rewards/rejected": -10.362597465515137,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.5189413596263622,
|
| 388 |
+
"grad_norm": 11.446046359523642,
|
| 389 |
+
"learning_rate": 2.7721671871299114e-07,
|
| 390 |
+
"logits/chosen": -0.7357327938079834,
|
| 391 |
+
"logits/rejected": 1.461576223373413,
|
| 392 |
+
"logps/chosen": -672.9193725585938,
|
| 393 |
+
"logps/rejected": -1338.949462890625,
|
| 394 |
+
"loss": 0.1602,
|
| 395 |
+
"rewards/accuracies": 0.940625011920929,
|
| 396 |
+
"rewards/chosen": -2.9312338829040527,
|
| 397 |
+
"rewards/margins": 6.449375152587891,
|
| 398 |
+
"rewards/rejected": -9.380608558654785,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.5396990140114167,
|
| 403 |
+
"grad_norm": 14.757287503572078,
|
| 404 |
+
"learning_rate": 2.5908825357849993e-07,
|
| 405 |
+
"logits/chosen": -0.8231679797172546,
|
| 406 |
+
"logits/rejected": 1.1155385971069336,
|
| 407 |
+
"logps/chosen": -656.1690063476562,
|
| 408 |
+
"logps/rejected": -1271.033447265625,
|
| 409 |
+
"loss": 0.1622,
|
| 410 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 411 |
+
"rewards/chosen": -2.8299801349639893,
|
| 412 |
+
"rewards/margins": 5.925788402557373,
|
| 413 |
+
"rewards/rejected": -8.755769729614258,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.5604566683964712,
|
| 418 |
+
"grad_norm": 17.896435569322648,
|
| 419 |
+
"learning_rate": 2.409117464215001e-07,
|
| 420 |
+
"logits/chosen": -0.4632663130760193,
|
| 421 |
+
"logits/rejected": 1.819011926651001,
|
| 422 |
+
"logps/chosen": -664.986572265625,
|
| 423 |
+
"logps/rejected": -1455.3441162109375,
|
| 424 |
+
"loss": 0.1471,
|
| 425 |
+
"rewards/accuracies": 0.940625011920929,
|
| 426 |
+
"rewards/chosen": -3.113548517227173,
|
| 427 |
+
"rewards/margins": 7.589502811431885,
|
| 428 |
+
"rewards/rejected": -10.70305061340332,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5812143227815257,
|
| 433 |
+
"grad_norm": 18.08668716790038,
|
| 434 |
+
"learning_rate": 2.227832812870089e-07,
|
| 435 |
+
"logits/chosen": -0.542155385017395,
|
| 436 |
+
"logits/rejected": 1.9537347555160522,
|
| 437 |
+
"logps/chosen": -686.6569213867188,
|
| 438 |
+
"logps/rejected": -1460.84619140625,
|
| 439 |
+
"loss": 0.1557,
|
| 440 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 441 |
+
"rewards/chosen": -3.106727123260498,
|
| 442 |
+
"rewards/margins": 7.513753414154053,
|
| 443 |
+
"rewards/rejected": -10.620479583740234,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.6019719771665801,
|
| 448 |
+
"grad_norm": 20.414881165009998,
|
| 449 |
+
"learning_rate": 2.0479868826190871e-07,
|
| 450 |
+
"logits/chosen": -0.437448650598526,
|
| 451 |
+
"logits/rejected": 1.7888593673706055,
|
| 452 |
+
"logps/chosen": -709.882568359375,
|
| 453 |
+
"logps/rejected": -1392.702392578125,
|
| 454 |
+
"loss": 0.1617,
|
| 455 |
+
"rewards/accuracies": 0.918749988079071,
|
| 456 |
+
"rewards/chosen": -3.35121488571167,
|
| 457 |
+
"rewards/margins": 6.753846645355225,
|
| 458 |
+
"rewards/rejected": -10.105062484741211,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.6227296315516346,
|
| 463 |
+
"grad_norm": 16.484968508125828,
|
| 464 |
+
"learning_rate": 1.8705303690114287e-07,
|
| 465 |
+
"logits/chosen": -0.2719939947128296,
|
| 466 |
+
"logits/rejected": 1.811428427696228,
|
| 467 |
+
"logps/chosen": -710.2088012695312,
|
| 468 |
+
"logps/rejected": -1419.313720703125,
|
| 469 |
+
"loss": 0.1478,
|
| 470 |
+
"rewards/accuracies": 0.9468749761581421,
|
| 471 |
+
"rewards/chosen": -3.3416149616241455,
|
| 472 |
+
"rewards/margins": 6.840612888336182,
|
| 473 |
+
"rewards/rejected": -10.182229042053223,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.6434872859366891,
|
| 478 |
+
"grad_norm": 15.309368142287557,
|
| 479 |
+
"learning_rate": 1.6964013367420965e-07,
|
| 480 |
+
"logits/chosen": -0.341867595911026,
|
| 481 |
+
"logits/rejected": 1.7422988414764404,
|
| 482 |
+
"logps/chosen": -686.5745849609375,
|
| 483 |
+
"logps/rejected": -1366.7637939453125,
|
| 484 |
+
"loss": 0.1534,
|
| 485 |
+
"rewards/accuracies": 0.918749988079071,
|
| 486 |
+
"rewards/chosen": -3.167562961578369,
|
| 487 |
+
"rewards/margins": 6.726889133453369,
|
| 488 |
+
"rewards/rejected": -9.894452095031738,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.6642449403217436,
|
| 493 |
+
"grad_norm": 11.4821386278306,
|
| 494 |
+
"learning_rate": 1.5265202608853628e-07,
|
| 495 |
+
"logits/chosen": -0.17296895384788513,
|
| 496 |
+
"logits/rejected": 1.9412486553192139,
|
| 497 |
+
"logps/chosen": -665.36376953125,
|
| 498 |
+
"logps/rejected": -1430.9146728515625,
|
| 499 |
+
"loss": 0.148,
|
| 500 |
+
"rewards/accuracies": 0.934374988079071,
|
| 501 |
+
"rewards/chosen": -3.0024094581604004,
|
| 502 |
+
"rewards/margins": 7.437635898590088,
|
| 503 |
+
"rewards/rejected": -10.440046310424805,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.6850025947067981,
|
| 508 |
+
"grad_norm": 10.552793459289758,
|
| 509 |
+
"learning_rate": 1.3617851611101993e-07,
|
| 510 |
+
"logits/chosen": -0.5047305226325989,
|
| 511 |
+
"logits/rejected": 1.5836106538772583,
|
| 512 |
+
"logps/chosen": -686.6402587890625,
|
| 513 |
+
"logps/rejected": -1443.398193359375,
|
| 514 |
+
"loss": 0.1519,
|
| 515 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 516 |
+
"rewards/chosen": -3.1413583755493164,
|
| 517 |
+
"rewards/margins": 7.2722015380859375,
|
| 518 |
+
"rewards/rejected": -10.413559913635254,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.7057602490918526,
|
| 523 |
+
"grad_norm": 17.65296980495948,
|
| 524 |
+
"learning_rate": 1.2030668545986958e-07,
|
| 525 |
+
"logits/chosen": -0.569928765296936,
|
| 526 |
+
"logits/rejected": 1.6948425769805908,
|
| 527 |
+
"logps/chosen": -716.9683837890625,
|
| 528 |
+
"logps/rejected": -1518.8372802734375,
|
| 529 |
+
"loss": 0.1463,
|
| 530 |
+
"rewards/accuracies": 0.921875,
|
| 531 |
+
"rewards/chosen": -3.4350712299346924,
|
| 532 |
+
"rewards/margins": 7.720976829528809,
|
| 533 |
+
"rewards/rejected": -11.156047821044922,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.7265179034769071,
|
| 538 |
+
"grad_norm": 14.715250965523067,
|
| 539 |
+
"learning_rate": 1.0512043527613623e-07,
|
| 540 |
+
"logits/chosen": -0.7549006342887878,
|
| 541 |
+
"logits/rejected": 1.3407833576202393,
|
| 542 |
+
"logps/chosen": -695.8525390625,
|
| 543 |
+
"logps/rejected": -1478.35546875,
|
| 544 |
+
"loss": 0.1559,
|
| 545 |
+
"rewards/accuracies": 0.903124988079071,
|
| 546 |
+
"rewards/chosen": -3.347827911376953,
|
| 547 |
+
"rewards/margins": 7.52248477935791,
|
| 548 |
+
"rewards/rejected": -10.870311737060547,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.7472755578619616,
|
| 553 |
+
"grad_norm": 18.014896980938854,
|
| 554 |
+
"learning_rate": 9.070004260830294e-08,
|
| 555 |
+
"logits/chosen": -0.9223737716674805,
|
| 556 |
+
"logits/rejected": 1.4507310390472412,
|
| 557 |
+
"logps/chosen": -692.382080078125,
|
| 558 |
+
"logps/rejected": -1453.919921875,
|
| 559 |
+
"loss": 0.1437,
|
| 560 |
+
"rewards/accuracies": 0.925000011920929,
|
| 561 |
+
"rewards/chosen": -3.2790751457214355,
|
| 562 |
+
"rewards/margins": 7.526673316955566,
|
| 563 |
+
"rewards/rejected": -10.80574893951416,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.768033212247016,
|
| 568 |
+
"grad_norm": 14.401106731580096,
|
| 569 |
+
"learning_rate": 7.712173605443267e-08,
|
| 570 |
+
"logits/chosen": -0.9376351237297058,
|
| 571 |
+
"logits/rejected": 1.583579421043396,
|
| 572 |
+
"logps/chosen": -700.8660888671875,
|
| 573 |
+
"logps/rejected": -1505.467529296875,
|
| 574 |
+
"loss": 0.1494,
|
| 575 |
+
"rewards/accuracies": 0.90625,
|
| 576 |
+
"rewards/chosen": -3.340554714202881,
|
| 577 |
+
"rewards/margins": 7.916815280914307,
|
| 578 |
+
"rewards/rejected": -11.257369995117188,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.7887908666320705,
|
| 583 |
+
"grad_norm": 11.438125899827915,
|
| 584 |
+
"learning_rate": 6.445729280509957e-08,
|
| 585 |
+
"logits/chosen": -0.8953694105148315,
|
| 586 |
+
"logits/rejected": 1.4685465097427368,
|
| 587 |
+
"logps/chosen": -690.8338623046875,
|
| 588 |
+
"logps/rejected": -1588.8245849609375,
|
| 589 |
+
"loss": 0.1468,
|
| 590 |
+
"rewards/accuracies": 0.921875,
|
| 591 |
+
"rewards/chosen": -3.3841090202331543,
|
| 592 |
+
"rewards/margins": 8.392255783081055,
|
| 593 |
+
"rewards/rejected": -11.77636432647705,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.809548521017125,
|
| 598 |
+
"grad_norm": 13.207322793003062,
|
| 599 |
+
"learning_rate": 5.2773659217201364e-08,
|
| 600 |
+
"logits/chosen": -0.9152933359146118,
|
| 601 |
+
"logits/rejected": 1.6746854782104492,
|
| 602 |
+
"logps/chosen": -707.5098266601562,
|
| 603 |
+
"logps/rejected": -1445.887939453125,
|
| 604 |
+
"loss": 0.1555,
|
| 605 |
+
"rewards/accuracies": 0.934374988079071,
|
| 606 |
+
"rewards/chosen": -3.3184616565704346,
|
| 607 |
+
"rewards/margins": 7.17882776260376,
|
| 608 |
+
"rewards/rejected": -10.497289657592773,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.8303061754021795,
|
| 613 |
+
"grad_norm": 11.811171239173802,
|
| 614 |
+
"learning_rate": 4.213259692436366e-08,
|
| 615 |
+
"logits/chosen": -0.807452380657196,
|
| 616 |
+
"logits/rejected": 1.5794246196746826,
|
| 617 |
+
"logps/chosen": -692.3187255859375,
|
| 618 |
+
"logps/rejected": -1477.223388671875,
|
| 619 |
+
"loss": 0.1455,
|
| 620 |
+
"rewards/accuracies": 0.934374988079071,
|
| 621 |
+
"rewards/chosen": -3.2713589668273926,
|
| 622 |
+
"rewards/margins": 7.557108402252197,
|
| 623 |
+
"rewards/rejected": -10.82846736907959,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.851063829787234,
|
| 628 |
+
"grad_norm": 13.900287224699019,
|
| 629 |
+
"learning_rate": 3.259035635465529e-08,
|
| 630 |
+
"logits/chosen": -0.6530941128730774,
|
| 631 |
+
"logits/rejected": 1.565045714378357,
|
| 632 |
+
"logps/chosen": -704.8355712890625,
|
| 633 |
+
"logps/rejected": -1375.3914794921875,
|
| 634 |
+
"loss": 0.1474,
|
| 635 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 636 |
+
"rewards/chosen": -3.393775463104248,
|
| 637 |
+
"rewards/margins": 6.558309078216553,
|
| 638 |
+
"rewards/rejected": -9.952085494995117,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.8718214841722886,
|
| 643 |
+
"grad_norm": 25.604242461790278,
|
| 644 |
+
"learning_rate": 2.4197379381453942e-08,
|
| 645 |
+
"logits/chosen": -0.6059257388114929,
|
| 646 |
+
"logits/rejected": 1.747667670249939,
|
| 647 |
+
"logps/chosen": -710.49365234375,
|
| 648 |
+
"logps/rejected": -1475.5013427734375,
|
| 649 |
+
"loss": 0.1551,
|
| 650 |
+
"rewards/accuracies": 0.918749988079071,
|
| 651 |
+
"rewards/chosen": -3.3643269538879395,
|
| 652 |
+
"rewards/margins": 7.329138278961182,
|
| 653 |
+
"rewards/rejected": -10.693464279174805,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.892579138557343,
|
| 658 |
+
"grad_norm": 17.02837493916288,
|
| 659 |
+
"learning_rate": 1.699803267930039e-08,
|
| 660 |
+
"logits/chosen": -0.6381738781929016,
|
| 661 |
+
"logits/rejected": 1.581946611404419,
|
| 662 |
+
"logps/chosen": -700.8256225585938,
|
| 663 |
+
"logps/rejected": -1424.5767822265625,
|
| 664 |
+
"loss": 0.14,
|
| 665 |
+
"rewards/accuracies": 0.918749988079071,
|
| 666 |
+
"rewards/chosen": -3.278334140777588,
|
| 667 |
+
"rewards/margins": 7.011561393737793,
|
| 668 |
+
"rewards/rejected": -10.289896965026855,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.9133367929423976,
|
| 673 |
+
"grad_norm": 12.115208573188879,
|
| 674 |
+
"learning_rate": 1.1030373194265114e-08,
|
| 675 |
+
"logits/chosen": -0.6104884743690491,
|
| 676 |
+
"logits/rejected": 1.8680555820465088,
|
| 677 |
+
"logps/chosen": -696.623046875,
|
| 678 |
+
"logps/rejected": -1484.712646484375,
|
| 679 |
+
"loss": 0.1555,
|
| 680 |
+
"rewards/accuracies": 0.925000011920929,
|
| 681 |
+
"rewards/chosen": -3.214094638824463,
|
| 682 |
+
"rewards/margins": 7.734494686126709,
|
| 683 |
+
"rewards/rejected": -10.948590278625488,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.934094447327452,
|
| 688 |
+
"grad_norm": 16.53240755993655,
|
| 689 |
+
"learning_rate": 6.325946968591317e-09,
|
| 690 |
+
"logits/chosen": -0.5625468492507935,
|
| 691 |
+
"logits/rejected": 2.0248143672943115,
|
| 692 |
+
"logps/chosen": -701.9017333984375,
|
| 693 |
+
"logps/rejected": -1457.185791015625,
|
| 694 |
+
"loss": 0.1506,
|
| 695 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 696 |
+
"rewards/chosen": -3.3476765155792236,
|
| 697 |
+
"rewards/margins": 7.459791660308838,
|
| 698 |
+
"rewards/rejected": -10.807466506958008,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.9548521017125065,
|
| 703 |
+
"grad_norm": 13.032337936414088,
|
| 704 |
+
"learning_rate": 2.909622383059834e-09,
|
| 705 |
+
"logits/chosen": -0.5611749291419983,
|
| 706 |
+
"logits/rejected": 1.8944803476333618,
|
| 707 |
+
"logps/chosen": -692.7264404296875,
|
| 708 |
+
"logps/rejected": -1460.9708251953125,
|
| 709 |
+
"loss": 0.1466,
|
| 710 |
+
"rewards/accuracies": 0.9375,
|
| 711 |
+
"rewards/chosen": -3.228968858718872,
|
| 712 |
+
"rewards/margins": 7.5441155433654785,
|
| 713 |
+
"rewards/rejected": -10.77308464050293,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.975609756097561,
|
| 718 |
+
"grad_norm": 17.252869051917436,
|
| 719 |
+
"learning_rate": 7.994586985856089e-10,
|
| 720 |
+
"logits/chosen": -0.6213638782501221,
|
| 721 |
+
"logits/rejected": 1.7617113590240479,
|
| 722 |
+
"logps/chosen": -707.3765869140625,
|
| 723 |
+
"logps/rejected": -1481.257080078125,
|
| 724 |
+
"loss": 0.1374,
|
| 725 |
+
"rewards/accuracies": 0.918749988079071,
|
| 726 |
+
"rewards/chosen": -3.3621768951416016,
|
| 727 |
+
"rewards/margins": 7.499251365661621,
|
| 728 |
+
"rewards/rejected": -10.861429214477539,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.9963674104826155,
|
| 733 |
+
"grad_norm": 12.988129191271089,
|
| 734 |
+
"learning_rate": 6.610591955641398e-12,
|
| 735 |
+
"logits/chosen": -0.5171535015106201,
|
| 736 |
+
"logits/rejected": 1.7204986810684204,
|
| 737 |
+
"logps/chosen": -687.4800415039062,
|
| 738 |
+
"logps/rejected": -1466.64501953125,
|
| 739 |
+
"loss": 0.1487,
|
| 740 |
+
"rewards/accuracies": 0.9375,
|
| 741 |
+
"rewards/chosen": -3.355544328689575,
|
| 742 |
+
"rewards/margins": 7.49337911605835,
|
| 743 |
+
"rewards/rejected": -10.848923683166504,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.9984431759211209,
|
| 748 |
+
"step": 481,
|
| 749 |
+
"total_flos": 0.0,
|
| 750 |
+
"train_loss": 0.21933725711709495,
|
| 751 |
+
"train_runtime": 13808.8776,
|
| 752 |
+
"train_samples_per_second": 8.93,
|
| 753 |
+
"train_steps_per_second": 0.035
|
| 754 |
+
}
|
| 755 |
+
],
|
| 756 |
+
"logging_steps": 10,
|
| 757 |
+
"max_steps": 481,
|
| 758 |
+
"num_input_tokens_seen": 0,
|
| 759 |
+
"num_train_epochs": 1,
|
| 760 |
+
"save_steps": 100,
|
| 761 |
+
"stateful_callbacks": {
|
| 762 |
+
"TrainerControl": {
|
| 763 |
+
"args": {
|
| 764 |
+
"should_epoch_stop": false,
|
| 765 |
+
"should_evaluate": false,
|
| 766 |
+
"should_log": false,
|
| 767 |
+
"should_save": true,
|
| 768 |
+
"should_training_stop": true
|
| 769 |
+
},
|
| 770 |
+
"attributes": {}
|
| 771 |
+
}
|
| 772 |
+
},
|
| 773 |
+
"total_flos": 0.0,
|
| 774 |
+
"train_batch_size": 8,
|
| 775 |
+
"trial_name": null,
|
| 776 |
+
"trial_params": null
|
| 777 |
+
}
|