Model save
Browse files- README.md +62 -0
- adapter_config.json +3 -3
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May19_03-07-30_gpu4-119-4/events.out.tfevents.1716052279.gpu4-119-4.3791581.0 +2 -2
- runs/May19_09-00-52_gpu4-119-4/events.out.tfevents.1716073311.gpu4-119-4.4154864.0 +3 -0
- train_results.json +8 -0
- trainer_state.json +1374 -0
- training_args.bin +1 -1
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-sft-lora-ultrachat-merged
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-dpo-chatml
|
| 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 |
+
# phi-2-dpo-chatml
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-sft-lora-ultrachat-merged](https://huggingface.co/DUAL-GPO/phi-2-sft-lora-ultrachat-merged) on the None 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-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 64
|
| 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 |
+
- PEFT 0.7.1
|
| 59 |
+
- Transformers 4.36.2
|
| 60 |
+
- Pytorch 2.1.2+cu118
|
| 61 |
+
- Datasets 2.14.6
|
| 62 |
+
- Tokenizers 0.15.2
|
adapter_config.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"alpha_pattern": {},
|
| 3 |
"auto_mapping": null,
|
| 4 |
-
"base_model_name_or_path": "
|
| 5 |
"bias": "none",
|
| 6 |
"fan_in_fan_out": false,
|
| 7 |
"inference_mode": true,
|
|
@@ -19,10 +19,10 @@
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
| 22 |
-
"v_proj",
|
| 23 |
"q_proj",
|
| 24 |
"dense",
|
| 25 |
-
"k_proj"
|
|
|
|
| 26 |
],
|
| 27 |
"task_type": "CAUSAL_LM"
|
| 28 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"alpha_pattern": {},
|
| 3 |
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "DUAL-GPO/phi-2-sft-lora-ultrachat-merged",
|
| 5 |
"bias": "none",
|
| 6 |
"fan_in_fan_out": false,
|
| 7 |
"inference_mode": true,
|
|
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
|
|
|
| 22 |
"q_proj",
|
| 23 |
"dense",
|
| 24 |
+
"k_proj",
|
| 25 |
+
"v_proj"
|
| 26 |
],
|
| 27 |
"task_type": "CAUSAL_LM"
|
| 28 |
}
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 335579632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b625585b975637b62de2e0a7fef025149816b12b1adec30bc7d2b790acf1b544
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.0333975282639109,
|
| 4 |
+
"train_runtime": 873.8284,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 69.962,
|
| 7 |
+
"train_steps_per_second": 1.093
|
| 8 |
+
}
|
runs/May19_03-07-30_gpu4-119-4/events.out.tfevents.1716052279.gpu4-119-4.3791581.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ff2924cbd3277a3fd12a54f774389803db1ca3d1e15374ee75a79c8ee1440ff
|
| 3 |
+
size 65894
|
runs/May19_09-00-52_gpu4-119-4/events.out.tfevents.1716073311.gpu4-119-4.4154864.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79fd016d001fc61723695add5d002b878d8aea92a24cc1a876695983842bff9e
|
| 3 |
+
size 8399
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.0333975282639109,
|
| 4 |
+
"train_runtime": 873.8284,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 69.962,
|
| 7 |
+
"train_steps_per_second": 1.093
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1374 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9997382884061764,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 955,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 5.208333333333333e-08,
|
| 14 |
+
"logits/chosen": 0.654504120349884,
|
| 15 |
+
"logits/rejected": 0.5926104187965393,
|
| 16 |
+
"logps/chosen": -305.37469482421875,
|
| 17 |
+
"logps/rejected": -235.55426025390625,
|
| 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": 0.01,
|
| 27 |
+
"learning_rate": 5.208333333333334e-07,
|
| 28 |
+
"logits/chosen": 0.7002869248390198,
|
| 29 |
+
"logits/rejected": 0.7474254965782166,
|
| 30 |
+
"logps/chosen": -312.4855041503906,
|
| 31 |
+
"logps/rejected": -270.3090515136719,
|
| 32 |
+
"loss": 0.6928,
|
| 33 |
+
"rewards/accuracies": 0.4722222089767456,
|
| 34 |
+
"rewards/chosen": 0.0012019119458273053,
|
| 35 |
+
"rewards/margins": 0.0008854579646140337,
|
| 36 |
+
"rewards/rejected": 0.00031645383569411933,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 1.0416666666666667e-06,
|
| 42 |
+
"logits/chosen": 0.6109224557876587,
|
| 43 |
+
"logits/rejected": 0.7636935114860535,
|
| 44 |
+
"logps/chosen": -229.7316436767578,
|
| 45 |
+
"logps/rejected": -217.67245483398438,
|
| 46 |
+
"loss": 0.6928,
|
| 47 |
+
"rewards/accuracies": 0.5,
|
| 48 |
+
"rewards/chosen": 0.0013416050933301449,
|
| 49 |
+
"rewards/margins": 0.0012069231597706676,
|
| 50 |
+
"rewards/rejected": 0.00013468178804032505,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"learning_rate": 1.5625e-06,
|
| 56 |
+
"logits/chosen": 0.6081576347351074,
|
| 57 |
+
"logits/rejected": 0.65868079662323,
|
| 58 |
+
"logps/chosen": -286.92681884765625,
|
| 59 |
+
"logps/rejected": -243.32778930664062,
|
| 60 |
+
"loss": 0.6929,
|
| 61 |
+
"rewards/accuracies": 0.5,
|
| 62 |
+
"rewards/chosen": 0.00032840127823874354,
|
| 63 |
+
"rewards/margins": 0.0011824481189250946,
|
| 64 |
+
"rewards/rejected": -0.0008540467242710292,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.04,
|
| 69 |
+
"learning_rate": 2.0833333333333334e-06,
|
| 70 |
+
"logits/chosen": 0.616665244102478,
|
| 71 |
+
"logits/rejected": 0.735335648059845,
|
| 72 |
+
"logps/chosen": -291.98284912109375,
|
| 73 |
+
"logps/rejected": -265.10198974609375,
|
| 74 |
+
"loss": 0.6926,
|
| 75 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 76 |
+
"rewards/chosen": -0.0013727883342653513,
|
| 77 |
+
"rewards/margins": 0.0006782080745324492,
|
| 78 |
+
"rewards/rejected": -0.0020509962923824787,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.05,
|
| 83 |
+
"learning_rate": 2.604166666666667e-06,
|
| 84 |
+
"logits/chosen": 0.6475738286972046,
|
| 85 |
+
"logits/rejected": 0.660605788230896,
|
| 86 |
+
"logps/chosen": -276.0094299316406,
|
| 87 |
+
"logps/rejected": -254.7918701171875,
|
| 88 |
+
"loss": 0.6923,
|
| 89 |
+
"rewards/accuracies": 0.53125,
|
| 90 |
+
"rewards/chosen": -1.2021581824228633e-05,
|
| 91 |
+
"rewards/margins": 0.0021027675829827785,
|
| 92 |
+
"rewards/rejected": -0.002114789094775915,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.06,
|
| 97 |
+
"learning_rate": 3.125e-06,
|
| 98 |
+
"logits/chosen": 0.6682679057121277,
|
| 99 |
+
"logits/rejected": 0.7323259115219116,
|
| 100 |
+
"logps/chosen": -308.230224609375,
|
| 101 |
+
"logps/rejected": -253.301025390625,
|
| 102 |
+
"loss": 0.6917,
|
| 103 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 104 |
+
"rewards/chosen": -0.0026237894780933857,
|
| 105 |
+
"rewards/margins": 0.003060306655243039,
|
| 106 |
+
"rewards/rejected": -0.005684095900505781,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.07,
|
| 111 |
+
"learning_rate": 3.6458333333333333e-06,
|
| 112 |
+
"logits/chosen": 0.6251315474510193,
|
| 113 |
+
"logits/rejected": 0.6695118546485901,
|
| 114 |
+
"logps/chosen": -296.9325866699219,
|
| 115 |
+
"logps/rejected": -265.0625915527344,
|
| 116 |
+
"loss": 0.6898,
|
| 117 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 118 |
+
"rewards/chosen": -0.003425169037654996,
|
| 119 |
+
"rewards/margins": 0.0062147388234734535,
|
| 120 |
+
"rewards/rejected": -0.009639907628297806,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"learning_rate": 4.166666666666667e-06,
|
| 126 |
+
"logits/chosen": 0.6303352117538452,
|
| 127 |
+
"logits/rejected": 0.6176969408988953,
|
| 128 |
+
"logps/chosen": -296.875244140625,
|
| 129 |
+
"logps/rejected": -259.2256774902344,
|
| 130 |
+
"loss": 0.6875,
|
| 131 |
+
"rewards/accuracies": 0.574999988079071,
|
| 132 |
+
"rewards/chosen": -0.011222435161471367,
|
| 133 |
+
"rewards/margins": 0.009960432536900043,
|
| 134 |
+
"rewards/rejected": -0.021182868629693985,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.09,
|
| 139 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 140 |
+
"logits/chosen": 0.6014409065246582,
|
| 141 |
+
"logits/rejected": 0.6135669946670532,
|
| 142 |
+
"logps/chosen": -271.5030212402344,
|
| 143 |
+
"logps/rejected": -246.84310913085938,
|
| 144 |
+
"loss": 0.6859,
|
| 145 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 146 |
+
"rewards/chosen": -0.019970973953604698,
|
| 147 |
+
"rewards/margins": 0.015956712886691093,
|
| 148 |
+
"rewards/rejected": -0.03592768684029579,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1,
|
| 153 |
+
"learning_rate": 4.9997324926814375e-06,
|
| 154 |
+
"logits/chosen": 0.5811746716499329,
|
| 155 |
+
"logits/rejected": 0.582471489906311,
|
| 156 |
+
"logps/chosen": -257.50079345703125,
|
| 157 |
+
"logps/rejected": -245.5070343017578,
|
| 158 |
+
"loss": 0.6848,
|
| 159 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 160 |
+
"rewards/chosen": -0.025636669248342514,
|
| 161 |
+
"rewards/margins": 0.01930178329348564,
|
| 162 |
+
"rewards/rejected": -0.044938452541828156,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.12,
|
| 167 |
+
"learning_rate": 4.996723692767927e-06,
|
| 168 |
+
"logits/chosen": 0.6834483742713928,
|
| 169 |
+
"logits/rejected": 0.7622531056404114,
|
| 170 |
+
"logps/chosen": -267.14837646484375,
|
| 171 |
+
"logps/rejected": -244.15005493164062,
|
| 172 |
+
"loss": 0.6816,
|
| 173 |
+
"rewards/accuracies": 0.581250011920929,
|
| 174 |
+
"rewards/chosen": -0.042797841131687164,
|
| 175 |
+
"rewards/margins": 0.023977002128958702,
|
| 176 |
+
"rewards/rejected": -0.06677485257387161,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.13,
|
| 181 |
+
"learning_rate": 4.9903757462135984e-06,
|
| 182 |
+
"logits/chosen": 0.5502853393554688,
|
| 183 |
+
"logits/rejected": 0.6681184768676758,
|
| 184 |
+
"logps/chosen": -267.02532958984375,
|
| 185 |
+
"logps/rejected": -238.77914428710938,
|
| 186 |
+
"loss": 0.6792,
|
| 187 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 188 |
+
"rewards/chosen": -0.05984887480735779,
|
| 189 |
+
"rewards/margins": 0.038902174681425095,
|
| 190 |
+
"rewards/rejected": -0.09875104576349258,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.14,
|
| 195 |
+
"learning_rate": 4.980697142834315e-06,
|
| 196 |
+
"logits/chosen": 0.5389119982719421,
|
| 197 |
+
"logits/rejected": 0.6768661737442017,
|
| 198 |
+
"logps/chosen": -308.8938903808594,
|
| 199 |
+
"logps/rejected": -267.13861083984375,
|
| 200 |
+
"loss": 0.6768,
|
| 201 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 202 |
+
"rewards/chosen": -0.08925649523735046,
|
| 203 |
+
"rewards/margins": 0.032300472259521484,
|
| 204 |
+
"rewards/rejected": -0.12155697494745255,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.15,
|
| 209 |
+
"learning_rate": 4.967700826904229e-06,
|
| 210 |
+
"logits/chosen": 0.5784076452255249,
|
| 211 |
+
"logits/rejected": 0.6090500950813293,
|
| 212 |
+
"logps/chosen": -268.8511657714844,
|
| 213 |
+
"logps/rejected": -252.6071319580078,
|
| 214 |
+
"loss": 0.6733,
|
| 215 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 216 |
+
"rewards/chosen": -0.12314645200967789,
|
| 217 |
+
"rewards/margins": 0.0378914549946785,
|
| 218 |
+
"rewards/rejected": -0.1610378921031952,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.16,
|
| 223 |
+
"learning_rate": 4.951404179843963e-06,
|
| 224 |
+
"logits/chosen": 0.6251226663589478,
|
| 225 |
+
"logits/rejected": 0.6390718817710876,
|
| 226 |
+
"logps/chosen": -277.3585510253906,
|
| 227 |
+
"logps/rejected": -255.3101043701172,
|
| 228 |
+
"loss": 0.6781,
|
| 229 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 230 |
+
"rewards/chosen": -0.14656075835227966,
|
| 231 |
+
"rewards/margins": 0.036974769085645676,
|
| 232 |
+
"rewards/rejected": -0.18353553116321564,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.17,
|
| 237 |
+
"learning_rate": 4.931828996974498e-06,
|
| 238 |
+
"logits/chosen": 0.6863027811050415,
|
| 239 |
+
"logits/rejected": 0.5657976865768433,
|
| 240 |
+
"logps/chosen": -279.9803161621094,
|
| 241 |
+
"logps/rejected": -273.37109375,
|
| 242 |
+
"loss": 0.6634,
|
| 243 |
+
"rewards/accuracies": 0.606249988079071,
|
| 244 |
+
"rewards/chosen": -0.22472751140594482,
|
| 245 |
+
"rewards/margins": 0.05435310676693916,
|
| 246 |
+
"rewards/rejected": -0.2790805995464325,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.18,
|
| 251 |
+
"learning_rate": 4.909001458367867e-06,
|
| 252 |
+
"logits/chosen": 0.6299469470977783,
|
| 253 |
+
"logits/rejected": 0.656846284866333,
|
| 254 |
+
"logps/chosen": -332.79541015625,
|
| 255 |
+
"logps/rejected": -292.52850341796875,
|
| 256 |
+
"loss": 0.6598,
|
| 257 |
+
"rewards/accuracies": 0.581250011920929,
|
| 258 |
+
"rewards/chosen": -0.3016909658908844,
|
| 259 |
+
"rewards/margins": 0.08058367669582367,
|
| 260 |
+
"rewards/rejected": -0.38227468729019165,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.19,
|
| 265 |
+
"learning_rate": 4.882952093833628e-06,
|
| 266 |
+
"logits/chosen": 0.5225847363471985,
|
| 267 |
+
"logits/rejected": 0.6088179349899292,
|
| 268 |
+
"logps/chosen": -289.1798095703125,
|
| 269 |
+
"logps/rejected": -311.8951110839844,
|
| 270 |
+
"loss": 0.652,
|
| 271 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 272 |
+
"rewards/chosen": -0.43765193223953247,
|
| 273 |
+
"rewards/margins": 0.09067411720752716,
|
| 274 |
+
"rewards/rejected": -0.5283260345458984,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2,
|
| 279 |
+
"learning_rate": 4.853715742087947e-06,
|
| 280 |
+
"logits/chosen": 0.5354307293891907,
|
| 281 |
+
"logits/rejected": 0.652572512626648,
|
| 282 |
+
"logps/chosen": -359.98126220703125,
|
| 283 |
+
"logps/rejected": -309.4668273925781,
|
| 284 |
+
"loss": 0.6449,
|
| 285 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 286 |
+
"rewards/chosen": -0.5696258544921875,
|
| 287 |
+
"rewards/margins": 0.17413343489170074,
|
| 288 |
+
"rewards/rejected": -0.7437592148780823,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.21,
|
| 293 |
+
"learning_rate": 4.821331504159906e-06,
|
| 294 |
+
"logits/chosen": 0.5104254484176636,
|
| 295 |
+
"logits/rejected": 0.49786561727523804,
|
| 296 |
+
"logps/chosen": -336.1935729980469,
|
| 297 |
+
"logps/rejected": -350.4376220703125,
|
| 298 |
+
"loss": 0.6426,
|
| 299 |
+
"rewards/accuracies": 0.612500011920929,
|
| 300 |
+
"rewards/chosen": -0.7382306456565857,
|
| 301 |
+
"rewards/margins": 0.15078046917915344,
|
| 302 |
+
"rewards/rejected": -0.8890112042427063,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.22,
|
| 307 |
+
"learning_rate": 4.7858426910973435e-06,
|
| 308 |
+
"logits/chosen": 0.4257640838623047,
|
| 309 |
+
"logits/rejected": 0.47642454504966736,
|
| 310 |
+
"logps/chosen": -387.05242919921875,
|
| 311 |
+
"logps/rejected": -345.6732177734375,
|
| 312 |
+
"loss": 0.6381,
|
| 313 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 314 |
+
"rewards/chosen": -0.8003188967704773,
|
| 315 |
+
"rewards/margins": 0.19792547821998596,
|
| 316 |
+
"rewards/rejected": -0.9982444047927856,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.23,
|
| 321 |
+
"learning_rate": 4.747296766042161e-06,
|
| 322 |
+
"logits/chosen": 0.4104679226875305,
|
| 323 |
+
"logits/rejected": 0.4869573712348938,
|
| 324 |
+
"logps/chosen": -348.92156982421875,
|
| 325 |
+
"logps/rejected": -350.780029296875,
|
| 326 |
+
"loss": 0.637,
|
| 327 |
+
"rewards/accuracies": 0.59375,
|
| 328 |
+
"rewards/chosen": -0.8022894859313965,
|
| 329 |
+
"rewards/margins": 0.19004520773887634,
|
| 330 |
+
"rewards/rejected": -0.9923346638679504,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.24,
|
| 335 |
+
"learning_rate": 4.705745280752586e-06,
|
| 336 |
+
"logits/chosen": 0.3983922600746155,
|
| 337 |
+
"logits/rejected": 0.42386817932128906,
|
| 338 |
+
"logps/chosen": -388.9286804199219,
|
| 339 |
+
"logps/rejected": -352.74444580078125,
|
| 340 |
+
"loss": 0.6239,
|
| 341 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 342 |
+
"rewards/chosen": -0.8833669424057007,
|
| 343 |
+
"rewards/margins": 0.20530752837657928,
|
| 344 |
+
"rewards/rejected": -1.088674545288086,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.25,
|
| 349 |
+
"learning_rate": 4.661243806657256e-06,
|
| 350 |
+
"logits/chosen": 0.38030871748924255,
|
| 351 |
+
"logits/rejected": 0.3976936936378479,
|
| 352 |
+
"logps/chosen": -378.4253234863281,
|
| 353 |
+
"logps/rejected": -352.67388916015625,
|
| 354 |
+
"loss": 0.6149,
|
| 355 |
+
"rewards/accuracies": 0.574999988079071,
|
| 356 |
+
"rewards/chosen": -0.9681804776191711,
|
| 357 |
+
"rewards/margins": 0.27472421526908875,
|
| 358 |
+
"rewards/rejected": -1.2429049015045166,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.26,
|
| 363 |
+
"learning_rate": 4.613851860533367e-06,
|
| 364 |
+
"logits/chosen": 0.29627329111099243,
|
| 365 |
+
"logits/rejected": 0.30808278918266296,
|
| 366 |
+
"logps/chosen": -370.7823791503906,
|
| 367 |
+
"logps/rejected": -365.70843505859375,
|
| 368 |
+
"loss": 0.6213,
|
| 369 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 370 |
+
"rewards/chosen": -0.9412585496902466,
|
| 371 |
+
"rewards/margins": 0.25268539786338806,
|
| 372 |
+
"rewards/rejected": -1.193943977355957,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.27,
|
| 377 |
+
"learning_rate": 4.563632824908252e-06,
|
| 378 |
+
"logits/chosen": 0.22838203608989716,
|
| 379 |
+
"logits/rejected": 0.2758544087409973,
|
| 380 |
+
"logps/chosen": -370.5462341308594,
|
| 381 |
+
"logps/rejected": -384.4304504394531,
|
| 382 |
+
"loss": 0.6205,
|
| 383 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 384 |
+
"rewards/chosen": -1.1226102113723755,
|
| 385 |
+
"rewards/margins": 0.22477027773857117,
|
| 386 |
+
"rewards/rejected": -1.347380518913269,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.28,
|
| 391 |
+
"learning_rate": 4.510653863290871e-06,
|
| 392 |
+
"logits/chosen": 0.16573426127433777,
|
| 393 |
+
"logits/rejected": 0.24319115281105042,
|
| 394 |
+
"logps/chosen": -432.40496826171875,
|
| 395 |
+
"logps/rejected": -434.70465087890625,
|
| 396 |
+
"loss": 0.6065,
|
| 397 |
+
"rewards/accuracies": 0.606249988079071,
|
| 398 |
+
"rewards/chosen": -1.3656351566314697,
|
| 399 |
+
"rewards/margins": 0.3202405571937561,
|
| 400 |
+
"rewards/rejected": -1.6858758926391602,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.29,
|
| 405 |
+
"learning_rate": 4.454985830346574e-06,
|
| 406 |
+
"logits/chosen": 0.15168949961662292,
|
| 407 |
+
"logits/rejected": 0.2517026662826538,
|
| 408 |
+
"logps/chosen": -496.7152404785156,
|
| 409 |
+
"logps/rejected": -481.1207580566406,
|
| 410 |
+
"loss": 0.6244,
|
| 411 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 412 |
+
"rewards/chosen": -1.8755683898925781,
|
| 413 |
+
"rewards/margins": 0.3297354280948639,
|
| 414 |
+
"rewards/rejected": -2.20530366897583,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.3,
|
| 419 |
+
"learning_rate": 4.396703177135262e-06,
|
| 420 |
+
"logits/chosen": 0.1856023520231247,
|
| 421 |
+
"logits/rejected": 0.21857495605945587,
|
| 422 |
+
"logps/chosen": -468.2583923339844,
|
| 423 |
+
"logps/rejected": -422.20477294921875,
|
| 424 |
+
"loss": 0.6089,
|
| 425 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 426 |
+
"rewards/chosen": -1.88265061378479,
|
| 427 |
+
"rewards/margins": 0.23467198014259338,
|
| 428 |
+
"rewards/rejected": -2.1173224449157715,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.31,
|
| 433 |
+
"learning_rate": 4.335883851539693e-06,
|
| 434 |
+
"logits/chosen": 0.10853853076696396,
|
| 435 |
+
"logits/rejected": 0.2359129637479782,
|
| 436 |
+
"logps/chosen": -437.02569580078125,
|
| 437 |
+
"logps/rejected": -435.40606689453125,
|
| 438 |
+
"loss": 0.5821,
|
| 439 |
+
"rewards/accuracies": 0.668749988079071,
|
| 440 |
+
"rewards/chosen": -1.7075822353363037,
|
| 441 |
+
"rewards/margins": 0.393388569355011,
|
| 442 |
+
"rewards/rejected": -2.10097074508667,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.32,
|
| 447 |
+
"learning_rate": 4.2726091940171055e-06,
|
| 448 |
+
"logits/chosen": 0.12634363770484924,
|
| 449 |
+
"logits/rejected": 0.224656343460083,
|
| 450 |
+
"logps/chosen": -436.81683349609375,
|
| 451 |
+
"logps/rejected": -466.3431091308594,
|
| 452 |
+
"loss": 0.5697,
|
| 453 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 454 |
+
"rewards/chosen": -1.6430351734161377,
|
| 455 |
+
"rewards/margins": 0.4527590870857239,
|
| 456 |
+
"rewards/rejected": -2.095794200897217,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.33,
|
| 461 |
+
"learning_rate": 4.206963828813555e-06,
|
| 462 |
+
"logits/chosen": 0.1670975536108017,
|
| 463 |
+
"logits/rejected": 0.20784397423267365,
|
| 464 |
+
"logps/chosen": -454.04400634765625,
|
| 465 |
+
"logps/rejected": -471.0421447753906,
|
| 466 |
+
"loss": 0.5959,
|
| 467 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 468 |
+
"rewards/chosen": -1.7891740798950195,
|
| 469 |
+
"rewards/margins": 0.4047388434410095,
|
| 470 |
+
"rewards/rejected": -2.193913221359253,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.35,
|
| 475 |
+
"learning_rate": 4.139035550786495e-06,
|
| 476 |
+
"logits/chosen": 0.13126584887504578,
|
| 477 |
+
"logits/rejected": 0.1421881467103958,
|
| 478 |
+
"logps/chosen": -436.185546875,
|
| 479 |
+
"logps/rejected": -425.6346740722656,
|
| 480 |
+
"loss": 0.621,
|
| 481 |
+
"rewards/accuracies": 0.606249988079071,
|
| 482 |
+
"rewards/chosen": -1.7343782186508179,
|
| 483 |
+
"rewards/margins": 0.29571670293807983,
|
| 484 |
+
"rewards/rejected": -2.030094861984253,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.36,
|
| 489 |
+
"learning_rate": 4.068915207986931e-06,
|
| 490 |
+
"logits/chosen": 0.18622621893882751,
|
| 491 |
+
"logits/rejected": 0.2519204020500183,
|
| 492 |
+
"logps/chosen": -453.08856201171875,
|
| 493 |
+
"logps/rejected": -439.3407287597656,
|
| 494 |
+
"loss": 0.63,
|
| 495 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 496 |
+
"rewards/chosen": -1.8599331378936768,
|
| 497 |
+
"rewards/margins": 0.25747692584991455,
|
| 498 |
+
"rewards/rejected": -2.117410182952881,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.37,
|
| 503 |
+
"learning_rate": 3.996696580158211e-06,
|
| 504 |
+
"logits/chosen": 0.15227162837982178,
|
| 505 |
+
"logits/rejected": 0.23813900351524353,
|
| 506 |
+
"logps/chosen": -502.65252685546875,
|
| 507 |
+
"logps/rejected": -521.7779541015625,
|
| 508 |
+
"loss": 0.5979,
|
| 509 |
+
"rewards/accuracies": 0.699999988079071,
|
| 510 |
+
"rewards/chosen": -1.993268609046936,
|
| 511 |
+
"rewards/margins": 0.44951528310775757,
|
| 512 |
+
"rewards/rejected": -2.442783832550049,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.38,
|
| 517 |
+
"learning_rate": 3.922476253313921e-06,
|
| 518 |
+
"logits/chosen": 0.07589183002710342,
|
| 519 |
+
"logits/rejected": 0.1402803361415863,
|
| 520 |
+
"logps/chosen": -492.88824462890625,
|
| 521 |
+
"logps/rejected": -508.8631286621094,
|
| 522 |
+
"loss": 0.5971,
|
| 523 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 524 |
+
"rewards/chosen": -2.206943988800049,
|
| 525 |
+
"rewards/margins": 0.2886379361152649,
|
| 526 |
+
"rewards/rejected": -2.495581865310669,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.39,
|
| 531 |
+
"learning_rate": 3.846353490562664e-06,
|
| 532 |
+
"logits/chosen": 0.20907124876976013,
|
| 533 |
+
"logits/rejected": 0.17944853007793427,
|
| 534 |
+
"logps/chosen": -460.7980041503906,
|
| 535 |
+
"logps/rejected": -506.36151123046875,
|
| 536 |
+
"loss": 0.5709,
|
| 537 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 538 |
+
"rewards/chosen": -1.9685118198394775,
|
| 539 |
+
"rewards/margins": 0.47328275442123413,
|
| 540 |
+
"rewards/rejected": -2.4417946338653564,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.4,
|
| 545 |
+
"learning_rate": 3.768430099352445e-06,
|
| 546 |
+
"logits/chosen": 0.16458635032176971,
|
| 547 |
+
"logits/rejected": 0.22938723862171173,
|
| 548 |
+
"logps/chosen": -461.98712158203125,
|
| 549 |
+
"logps/rejected": -475.66046142578125,
|
| 550 |
+
"loss": 0.5929,
|
| 551 |
+
"rewards/accuracies": 0.643750011920929,
|
| 552 |
+
"rewards/chosen": -1.9674451351165771,
|
| 553 |
+
"rewards/margins": 0.38506731390953064,
|
| 554 |
+
"rewards/rejected": -2.3525123596191406,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.41,
|
| 559 |
+
"learning_rate": 3.6888102953122307e-06,
|
| 560 |
+
"logits/chosen": 0.06095918267965317,
|
| 561 |
+
"logits/rejected": 0.1491270363330841,
|
| 562 |
+
"logps/chosen": -450.43798828125,
|
| 563 |
+
"logps/rejected": -455.1705017089844,
|
| 564 |
+
"loss": 0.6117,
|
| 565 |
+
"rewards/accuracies": 0.625,
|
| 566 |
+
"rewards/chosen": -1.9474515914916992,
|
| 567 |
+
"rewards/margins": 0.4293443560600281,
|
| 568 |
+
"rewards/rejected": -2.376796007156372,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.42,
|
| 573 |
+
"learning_rate": 3.607600562872785e-06,
|
| 574 |
+
"logits/chosen": 0.09890508651733398,
|
| 575 |
+
"logits/rejected": 0.1348509043455124,
|
| 576 |
+
"logps/chosen": -522.7869873046875,
|
| 577 |
+
"logps/rejected": -504.12469482421875,
|
| 578 |
+
"loss": 0.5751,
|
| 579 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 580 |
+
"rewards/chosen": -2.1772594451904297,
|
| 581 |
+
"rewards/margins": 0.4235813617706299,
|
| 582 |
+
"rewards/rejected": -2.6008408069610596,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.43,
|
| 587 |
+
"learning_rate": 3.5249095128531863e-06,
|
| 588 |
+
"logits/chosen": 0.15617763996124268,
|
| 589 |
+
"logits/rejected": 0.2346019446849823,
|
| 590 |
+
"logps/chosen": -525.8595581054688,
|
| 591 |
+
"logps/rejected": -515.14599609375,
|
| 592 |
+
"loss": 0.6036,
|
| 593 |
+
"rewards/accuracies": 0.550000011920929,
|
| 594 |
+
"rewards/chosen": -2.276054859161377,
|
| 595 |
+
"rewards/margins": 0.37463003396987915,
|
| 596 |
+
"rewards/rejected": -2.6506848335266113,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.44,
|
| 601 |
+
"learning_rate": 3.4408477372034743e-06,
|
| 602 |
+
"logits/chosen": 0.057582952082157135,
|
| 603 |
+
"logits/rejected": 0.13709139823913574,
|
| 604 |
+
"logps/chosen": -447.7400817871094,
|
| 605 |
+
"logps/rejected": -488.5953063964844,
|
| 606 |
+
"loss": 0.5871,
|
| 607 |
+
"rewards/accuracies": 0.606249988079071,
|
| 608 |
+
"rewards/chosen": -2.0702292919158936,
|
| 609 |
+
"rewards/margins": 0.3465236723423004,
|
| 610 |
+
"rewards/rejected": -2.4167532920837402,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.45,
|
| 615 |
+
"learning_rate": 3.355527661097728e-06,
|
| 616 |
+
"logits/chosen": 0.10504597425460815,
|
| 617 |
+
"logits/rejected": 0.18081213533878326,
|
| 618 |
+
"logps/chosen": -476.1048278808594,
|
| 619 |
+
"logps/rejected": -492.5294494628906,
|
| 620 |
+
"loss": 0.6166,
|
| 621 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 622 |
+
"rewards/chosen": -2.1964662075042725,
|
| 623 |
+
"rewards/margins": 0.29873451590538025,
|
| 624 |
+
"rewards/rejected": -2.4952008724212646,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.46,
|
| 629 |
+
"learning_rate": 3.269063392575352e-06,
|
| 630 |
+
"logits/chosen": 0.11929740011692047,
|
| 631 |
+
"logits/rejected": 0.15031102299690247,
|
| 632 |
+
"logps/chosen": -505.28363037109375,
|
| 633 |
+
"logps/rejected": -516.3770751953125,
|
| 634 |
+
"loss": 0.6019,
|
| 635 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 636 |
+
"rewards/chosen": -2.265862226486206,
|
| 637 |
+
"rewards/margins": 0.365284264087677,
|
| 638 |
+
"rewards/rejected": -2.6311464309692383,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.47,
|
| 643 |
+
"learning_rate": 3.181570569931697e-06,
|
| 644 |
+
"logits/chosen": 0.10959301143884659,
|
| 645 |
+
"logits/rejected": 0.15382876992225647,
|
| 646 |
+
"logps/chosen": -448.1092224121094,
|
| 647 |
+
"logps/rejected": -466.357421875,
|
| 648 |
+
"loss": 0.6033,
|
| 649 |
+
"rewards/accuracies": 0.59375,
|
| 650 |
+
"rewards/chosen": -2.1950180530548096,
|
| 651 |
+
"rewards/margins": 0.3157437741756439,
|
| 652 |
+
"rewards/rejected": -2.5107619762420654,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.48,
|
| 657 |
+
"learning_rate": 3.09316620706208e-06,
|
| 658 |
+
"logits/chosen": 0.0773458257317543,
|
| 659 |
+
"logits/rejected": 0.18486995995044708,
|
| 660 |
+
"logps/chosen": -508.8724670410156,
|
| 661 |
+
"logps/rejected": -497.22882080078125,
|
| 662 |
+
"loss": 0.585,
|
| 663 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 664 |
+
"rewards/chosen": -2.2086341381073,
|
| 665 |
+
"rewards/margins": 0.32577189803123474,
|
| 666 |
+
"rewards/rejected": -2.5344061851501465,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.49,
|
| 671 |
+
"learning_rate": 3.0039685369660785e-06,
|
| 672 |
+
"logits/chosen": 0.09328463673591614,
|
| 673 |
+
"logits/rejected": 0.20468135178089142,
|
| 674 |
+
"logps/chosen": -522.9307861328125,
|
| 675 |
+
"logps/rejected": -512.7855834960938,
|
| 676 |
+
"loss": 0.5807,
|
| 677 |
+
"rewards/accuracies": 0.5625,
|
| 678 |
+
"rewards/chosen": -2.516700506210327,
|
| 679 |
+
"rewards/margins": 0.32140296697616577,
|
| 680 |
+
"rewards/rejected": -2.8381035327911377,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.5,
|
| 685 |
+
"learning_rate": 2.91409685362137e-06,
|
| 686 |
+
"logits/chosen": 0.1576906144618988,
|
| 687 |
+
"logits/rejected": 0.10724158585071564,
|
| 688 |
+
"logps/chosen": -456.68359375,
|
| 689 |
+
"logps/rejected": -451.1465759277344,
|
| 690 |
+
"loss": 0.6305,
|
| 691 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 692 |
+
"rewards/chosen": -2.26971697807312,
|
| 693 |
+
"rewards/margins": 0.18658225238323212,
|
| 694 |
+
"rewards/rejected": -2.456299304962158,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.51,
|
| 699 |
+
"learning_rate": 2.8236713524386085e-06,
|
| 700 |
+
"logits/chosen": 0.1447945386171341,
|
| 701 |
+
"logits/rejected": 0.1760699599981308,
|
| 702 |
+
"logps/chosen": -467.67889404296875,
|
| 703 |
+
"logps/rejected": -484.5015563964844,
|
| 704 |
+
"loss": 0.5847,
|
| 705 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 706 |
+
"rewards/chosen": -2.1700217723846436,
|
| 707 |
+
"rewards/margins": 0.4066441059112549,
|
| 708 |
+
"rewards/rejected": -2.5766658782958984,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.52,
|
| 713 |
+
"learning_rate": 2.7328129695107205e-06,
|
| 714 |
+
"logits/chosen": 0.09771154820919037,
|
| 715 |
+
"logits/rejected": 0.21410468220710754,
|
| 716 |
+
"logps/chosen": -505.1236877441406,
|
| 717 |
+
"logps/rejected": -521.1968994140625,
|
| 718 |
+
"loss": 0.5852,
|
| 719 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 720 |
+
"rewards/chosen": -2.445769786834717,
|
| 721 |
+
"rewards/margins": 0.5605564117431641,
|
| 722 |
+
"rewards/rejected": -3.00632643699646,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.53,
|
| 727 |
+
"learning_rate": 2.641643219871597e-06,
|
| 728 |
+
"logits/chosen": 0.0740189403295517,
|
| 729 |
+
"logits/rejected": 0.10937446355819702,
|
| 730 |
+
"logps/chosen": -575.5813598632812,
|
| 731 |
+
"logps/rejected": -578.539306640625,
|
| 732 |
+
"loss": 0.5878,
|
| 733 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 734 |
+
"rewards/chosen": -3.042142152786255,
|
| 735 |
+
"rewards/margins": 0.41739901900291443,
|
| 736 |
+
"rewards/rejected": -3.4595413208007812,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.54,
|
| 741 |
+
"learning_rate": 2.5502840349805074e-06,
|
| 742 |
+
"logits/chosen": 0.13737405836582184,
|
| 743 |
+
"logits/rejected": 0.1366177350282669,
|
| 744 |
+
"logps/chosen": -577.433837890625,
|
| 745 |
+
"logps/rejected": -586.6515502929688,
|
| 746 |
+
"loss": 0.6018,
|
| 747 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 748 |
+
"rewards/chosen": -2.9127590656280518,
|
| 749 |
+
"rewards/margins": 0.4386274814605713,
|
| 750 |
+
"rewards/rejected": -3.351386308670044,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.55,
|
| 755 |
+
"learning_rate": 2.4588575996495797e-06,
|
| 756 |
+
"logits/chosen": 0.009027734398841858,
|
| 757 |
+
"logits/rejected": 0.15047816932201385,
|
| 758 |
+
"logps/chosen": -578.1231689453125,
|
| 759 |
+
"logps/rejected": -593.139404296875,
|
| 760 |
+
"loss": 0.584,
|
| 761 |
+
"rewards/accuracies": 0.59375,
|
| 762 |
+
"rewards/chosen": -2.788166046142578,
|
| 763 |
+
"rewards/margins": 0.4355267882347107,
|
| 764 |
+
"rewards/rejected": -3.2236924171447754,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.57,
|
| 769 |
+
"learning_rate": 2.367486188632446e-06,
|
| 770 |
+
"logits/chosen": 0.12746676802635193,
|
| 771 |
+
"logits/rejected": 0.15015694499015808,
|
| 772 |
+
"logps/chosen": -549.587890625,
|
| 773 |
+
"logps/rejected": -567.0057373046875,
|
| 774 |
+
"loss": 0.5919,
|
| 775 |
+
"rewards/accuracies": 0.643750011920929,
|
| 776 |
+
"rewards/chosen": -2.6341264247894287,
|
| 777 |
+
"rewards/margins": 0.4422709047794342,
|
| 778 |
+
"rewards/rejected": -3.07639741897583,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.58,
|
| 783 |
+
"learning_rate": 2.276292003092593e-06,
|
| 784 |
+
"logits/chosen": 0.06688917428255081,
|
| 785 |
+
"logits/rejected": 0.1280103623867035,
|
| 786 |
+
"logps/chosen": -534.5496215820312,
|
| 787 |
+
"logps/rejected": -526.1148071289062,
|
| 788 |
+
"loss": 0.6045,
|
| 789 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 790 |
+
"rewards/chosen": -2.5599446296691895,
|
| 791 |
+
"rewards/margins": 0.3815520107746124,
|
| 792 |
+
"rewards/rejected": -2.9414966106414795,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.59,
|
| 797 |
+
"learning_rate": 2.1853970071701415e-06,
|
| 798 |
+
"logits/chosen": 0.05697401612997055,
|
| 799 |
+
"logits/rejected": 0.15568815171718597,
|
| 800 |
+
"logps/chosen": -523.0081787109375,
|
| 801 |
+
"logps/rejected": -505.5555114746094,
|
| 802 |
+
"loss": 0.5914,
|
| 803 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 804 |
+
"rewards/chosen": -2.705841541290283,
|
| 805 |
+
"rewards/margins": 0.403696209192276,
|
| 806 |
+
"rewards/rejected": -3.1095376014709473,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.6,
|
| 811 |
+
"learning_rate": 2.0949227648656194e-06,
|
| 812 |
+
"logits/chosen": 0.08012814819812775,
|
| 813 |
+
"logits/rejected": 0.1830509603023529,
|
| 814 |
+
"logps/chosen": -562.3424682617188,
|
| 815 |
+
"logps/rejected": -572.2830200195312,
|
| 816 |
+
"loss": 0.6085,
|
| 817 |
+
"rewards/accuracies": 0.574999988079071,
|
| 818 |
+
"rewards/chosen": -2.9859776496887207,
|
| 819 |
+
"rewards/margins": 0.37897008657455444,
|
| 820 |
+
"rewards/rejected": -3.364948272705078,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.61,
|
| 825 |
+
"learning_rate": 2.00499027745888e-06,
|
| 826 |
+
"logits/chosen": 0.13961026072502136,
|
| 827 |
+
"logits/rejected": 0.18276450037956238,
|
| 828 |
+
"logps/chosen": -562.8177490234375,
|
| 829 |
+
"logps/rejected": -579.6517333984375,
|
| 830 |
+
"loss": 0.6114,
|
| 831 |
+
"rewards/accuracies": 0.574999988079071,
|
| 832 |
+
"rewards/chosen": -2.9555232524871826,
|
| 833 |
+
"rewards/margins": 0.380531370639801,
|
| 834 |
+
"rewards/rejected": -3.336055040359497,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.62,
|
| 839 |
+
"learning_rate": 1.915719821680624e-06,
|
| 840 |
+
"logits/chosen": 0.10515383630990982,
|
| 841 |
+
"logits/rejected": 0.12738659977912903,
|
| 842 |
+
"logps/chosen": -508.9317932128906,
|
| 843 |
+
"logps/rejected": -502.7018127441406,
|
| 844 |
+
"loss": 0.5984,
|
| 845 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 846 |
+
"rewards/chosen": -2.51983904838562,
|
| 847 |
+
"rewards/margins": 0.23314905166625977,
|
| 848 |
+
"rewards/rejected": -2.752988338470459,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.63,
|
| 853 |
+
"learning_rate": 1.8272307888529276e-06,
|
| 854 |
+
"logits/chosen": 0.1120181530714035,
|
| 855 |
+
"logits/rejected": 0.18100281059741974,
|
| 856 |
+
"logps/chosen": -511.30645751953125,
|
| 857 |
+
"logps/rejected": -542.5186767578125,
|
| 858 |
+
"loss": 0.5837,
|
| 859 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 860 |
+
"rewards/chosen": -2.2869346141815186,
|
| 861 |
+
"rewards/margins": 0.4697488248348236,
|
| 862 |
+
"rewards/rejected": -2.756683111190796,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.64,
|
| 867 |
+
"learning_rate": 1.739641525213929e-06,
|
| 868 |
+
"logits/chosen": 0.13541504740715027,
|
| 869 |
+
"logits/rejected": 0.1992477923631668,
|
| 870 |
+
"logps/chosen": -526.412841796875,
|
| 871 |
+
"logps/rejected": -498.51318359375,
|
| 872 |
+
"loss": 0.5803,
|
| 873 |
+
"rewards/accuracies": 0.581250011920929,
|
| 874 |
+
"rewards/chosen": -2.441995143890381,
|
| 875 |
+
"rewards/margins": 0.4095155596733093,
|
| 876 |
+
"rewards/rejected": -2.851510524749756,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.65,
|
| 881 |
+
"learning_rate": 1.6530691736402317e-06,
|
| 882 |
+
"logits/chosen": 0.11898718029260635,
|
| 883 |
+
"logits/rejected": 0.11196433007717133,
|
| 884 |
+
"logps/chosen": -530.7737426757812,
|
| 885 |
+
"logps/rejected": -540.7352294921875,
|
| 886 |
+
"loss": 0.5947,
|
| 887 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 888 |
+
"rewards/chosen": -2.539415121078491,
|
| 889 |
+
"rewards/margins": 0.4597281813621521,
|
| 890 |
+
"rewards/rejected": -2.999143362045288,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.66,
|
| 895 |
+
"learning_rate": 1.5676295169786864e-06,
|
| 896 |
+
"logits/chosen": 0.12535694241523743,
|
| 897 |
+
"logits/rejected": 0.2773950397968292,
|
| 898 |
+
"logps/chosen": -558.296630859375,
|
| 899 |
+
"logps/rejected": -556.3736572265625,
|
| 900 |
+
"loss": 0.5937,
|
| 901 |
+
"rewards/accuracies": 0.612500011920929,
|
| 902 |
+
"rewards/chosen": -2.703909158706665,
|
| 903 |
+
"rewards/margins": 0.4821166396141052,
|
| 904 |
+
"rewards/rejected": -3.186025857925415,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.67,
|
| 909 |
+
"learning_rate": 1.4834368231970922e-06,
|
| 910 |
+
"logits/chosen": 0.13473114371299744,
|
| 911 |
+
"logits/rejected": 0.18143931031227112,
|
| 912 |
+
"logps/chosen": -497.7254333496094,
|
| 913 |
+
"logps/rejected": -533.7047119140625,
|
| 914 |
+
"loss": 0.5819,
|
| 915 |
+
"rewards/accuracies": 0.625,
|
| 916 |
+
"rewards/chosen": -2.562814235687256,
|
| 917 |
+
"rewards/margins": 0.5369300842285156,
|
| 918 |
+
"rewards/rejected": -3.0997443199157715,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.68,
|
| 923 |
+
"learning_rate": 1.4006036925609245e-06,
|
| 924 |
+
"logits/chosen": 0.051880162209272385,
|
| 925 |
+
"logits/rejected": 0.20864169299602509,
|
| 926 |
+
"logps/chosen": -551.8402099609375,
|
| 927 |
+
"logps/rejected": -567.7971801757812,
|
| 928 |
+
"loss": 0.5719,
|
| 929 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 930 |
+
"rewards/chosen": -2.570906162261963,
|
| 931 |
+
"rewards/margins": 0.5435839295387268,
|
| 932 |
+
"rewards/rejected": -3.114490270614624,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.69,
|
| 937 |
+
"learning_rate": 1.3192409070404582e-06,
|
| 938 |
+
"logits/chosen": 0.12426358461380005,
|
| 939 |
+
"logits/rejected": 0.08786197006702423,
|
| 940 |
+
"logps/chosen": -543.6322021484375,
|
| 941 |
+
"logps/rejected": -535.1307373046875,
|
| 942 |
+
"loss": 0.596,
|
| 943 |
+
"rewards/accuracies": 0.59375,
|
| 944 |
+
"rewards/chosen": -2.673713445663452,
|
| 945 |
+
"rewards/margins": 0.34852334856987,
|
| 946 |
+
"rewards/rejected": -3.0222370624542236,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.7,
|
| 951 |
+
"learning_rate": 1.2394572821496953e-06,
|
| 952 |
+
"logits/chosen": 0.07642360031604767,
|
| 953 |
+
"logits/rejected": 0.1105254516005516,
|
| 954 |
+
"logps/chosen": -521.8109130859375,
|
| 955 |
+
"logps/rejected": -533.8194580078125,
|
| 956 |
+
"loss": 0.5713,
|
| 957 |
+
"rewards/accuracies": 0.668749988079071,
|
| 958 |
+
"rewards/chosen": -2.5358338356018066,
|
| 959 |
+
"rewards/margins": 0.5434619188308716,
|
| 960 |
+
"rewards/rejected": -3.0792956352233887,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.71,
|
| 965 |
+
"learning_rate": 1.1613595214152713e-06,
|
| 966 |
+
"logits/chosen": 0.07207540422677994,
|
| 967 |
+
"logits/rejected": 0.1519361436367035,
|
| 968 |
+
"logps/chosen": -538.2649536132812,
|
| 969 |
+
"logps/rejected": -527.5178833007812,
|
| 970 |
+
"loss": 0.6009,
|
| 971 |
+
"rewards/accuracies": 0.550000011920929,
|
| 972 |
+
"rewards/chosen": -2.5930490493774414,
|
| 973 |
+
"rewards/margins": 0.24494171142578125,
|
| 974 |
+
"rewards/rejected": -2.8379909992218018,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.72,
|
| 979 |
+
"learning_rate": 1.0850520736699362e-06,
|
| 980 |
+
"logits/chosen": 0.09551219642162323,
|
| 981 |
+
"logits/rejected": 0.20561787486076355,
|
| 982 |
+
"logps/chosen": -522.4210815429688,
|
| 983 |
+
"logps/rejected": -531.2017822265625,
|
| 984 |
+
"loss": 0.5918,
|
| 985 |
+
"rewards/accuracies": 0.675000011920929,
|
| 986 |
+
"rewards/chosen": -2.5857510566711426,
|
| 987 |
+
"rewards/margins": 0.47004079818725586,
|
| 988 |
+
"rewards/rejected": -3.0557916164398193,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.73,
|
| 993 |
+
"learning_rate": 1.0106369933615043e-06,
|
| 994 |
+
"logits/chosen": 0.0834534615278244,
|
| 995 |
+
"logits/rejected": 0.27686116099357605,
|
| 996 |
+
"logps/chosen": -509.4612731933594,
|
| 997 |
+
"logps/rejected": -507.83258056640625,
|
| 998 |
+
"loss": 0.5759,
|
| 999 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1000 |
+
"rewards/chosen": -2.489625930786133,
|
| 1001 |
+
"rewards/margins": 0.5027724504470825,
|
| 1002 |
+
"rewards/rejected": -2.992398262023926,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.74,
|
| 1007 |
+
"learning_rate": 9.382138040640714e-07,
|
| 1008 |
+
"logits/chosen": 0.08181905746459961,
|
| 1009 |
+
"logits/rejected": 0.10011023283004761,
|
| 1010 |
+
"logps/chosen": -566.5830078125,
|
| 1011 |
+
"logps/rejected": -552.9130249023438,
|
| 1012 |
+
"loss": 0.5783,
|
| 1013 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1014 |
+
"rewards/chosen": -2.718501091003418,
|
| 1015 |
+
"rewards/margins": 0.40024662017822266,
|
| 1016 |
+
"rewards/rejected": -3.1187477111816406,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.75,
|
| 1021 |
+
"learning_rate": 8.678793653740633e-07,
|
| 1022 |
+
"logits/chosen": 0.0735187977552414,
|
| 1023 |
+
"logits/rejected": 0.13505037128925323,
|
| 1024 |
+
"logps/chosen": -513.2781372070312,
|
| 1025 |
+
"logps/rejected": -523.3275146484375,
|
| 1026 |
+
"loss": 0.5936,
|
| 1027 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1028 |
+
"rewards/chosen": -2.5609078407287598,
|
| 1029 |
+
"rewards/margins": 0.3891633450984955,
|
| 1030 |
+
"rewards/rejected": -2.9500718116760254,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.76,
|
| 1035 |
+
"learning_rate": 7.997277433690984e-07,
|
| 1036 |
+
"logits/chosen": 0.06112724542617798,
|
| 1037 |
+
"logits/rejected": 0.0807592123746872,
|
| 1038 |
+
"logps/chosen": -558.6268310546875,
|
| 1039 |
+
"logps/rejected": -531.7501220703125,
|
| 1040 |
+
"loss": 0.5879,
|
| 1041 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1042 |
+
"rewards/chosen": -2.575448989868164,
|
| 1043 |
+
"rewards/margins": 0.3756105601787567,
|
| 1044 |
+
"rewards/rejected": -2.951059341430664,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.77,
|
| 1049 |
+
"learning_rate": 7.338500848029603e-07,
|
| 1050 |
+
"logits/chosen": 0.005916979629546404,
|
| 1051 |
+
"logits/rejected": 0.17464521527290344,
|
| 1052 |
+
"logps/chosen": -569.9360961914062,
|
| 1053 |
+
"logps/rejected": -577.1893310546875,
|
| 1054 |
+
"loss": 0.5867,
|
| 1055 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1056 |
+
"rewards/chosen": -2.632016181945801,
|
| 1057 |
+
"rewards/margins": 0.4428483545780182,
|
| 1058 |
+
"rewards/rejected": -3.074864387512207,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.79,
|
| 1063 |
+
"learning_rate": 6.70334495204884e-07,
|
| 1064 |
+
"logits/chosen": 0.1314985156059265,
|
| 1065 |
+
"logits/rejected": 0.16822925209999084,
|
| 1066 |
+
"logps/chosen": -500.46636962890625,
|
| 1067 |
+
"logps/rejected": -552.1305541992188,
|
| 1068 |
+
"loss": 0.5901,
|
| 1069 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1070 |
+
"rewards/chosen": -2.497035503387451,
|
| 1071 |
+
"rewards/margins": 0.5148124694824219,
|
| 1072 |
+
"rewards/rejected": -3.011847972869873,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.8,
|
| 1077 |
+
"learning_rate": 6.092659210462232e-07,
|
| 1078 |
+
"logits/chosen": 0.052326373755931854,
|
| 1079 |
+
"logits/rejected": 0.16787569224834442,
|
| 1080 |
+
"logps/chosen": -557.7818603515625,
|
| 1081 |
+
"logps/rejected": -567.0888061523438,
|
| 1082 |
+
"loss": 0.5689,
|
| 1083 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1084 |
+
"rewards/chosen": -2.623483657836914,
|
| 1085 |
+
"rewards/margins": 0.46536707878112793,
|
| 1086 |
+
"rewards/rejected": -3.088850736618042,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.81,
|
| 1091 |
+
"learning_rate": 5.507260361320738e-07,
|
| 1092 |
+
"logits/chosen": 0.08203858882188797,
|
| 1093 |
+
"logits/rejected": 0.14716270565986633,
|
| 1094 |
+
"logps/chosen": -574.7208862304688,
|
| 1095 |
+
"logps/rejected": -586.5050659179688,
|
| 1096 |
+
"loss": 0.5999,
|
| 1097 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1098 |
+
"rewards/chosen": -2.7186100482940674,
|
| 1099 |
+
"rewards/margins": 0.42362919449806213,
|
| 1100 |
+
"rewards/rejected": -3.1422390937805176,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.82,
|
| 1105 |
+
"learning_rate": 4.947931323697983e-07,
|
| 1106 |
+
"logits/chosen": 0.10785499960184097,
|
| 1107 |
+
"logits/rejected": 0.18369507789611816,
|
| 1108 |
+
"logps/chosen": -542.8053588867188,
|
| 1109 |
+
"logps/rejected": -549.775146484375,
|
| 1110 |
+
"loss": 0.5772,
|
| 1111 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1112 |
+
"rewards/chosen": -2.621422529220581,
|
| 1113 |
+
"rewards/margins": 0.4870278239250183,
|
| 1114 |
+
"rewards/rejected": -3.108450412750244,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.83,
|
| 1119 |
+
"learning_rate": 4.4154201506053985e-07,
|
| 1120 |
+
"logits/chosen": 0.11180669069290161,
|
| 1121 |
+
"logits/rejected": 0.14754420518875122,
|
| 1122 |
+
"logps/chosen": -532.26123046875,
|
| 1123 |
+
"logps/rejected": -550.2486572265625,
|
| 1124 |
+
"loss": 0.6016,
|
| 1125 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1126 |
+
"rewards/chosen": -2.573822021484375,
|
| 1127 |
+
"rewards/margins": 0.3509921431541443,
|
| 1128 |
+
"rewards/rejected": -2.924813985824585,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.84,
|
| 1133 |
+
"learning_rate": 3.910439028537638e-07,
|
| 1134 |
+
"logits/chosen": 0.10717664659023285,
|
| 1135 |
+
"logits/rejected": 0.24534320831298828,
|
| 1136 |
+
"logps/chosen": -510.2649841308594,
|
| 1137 |
+
"logps/rejected": -519.7151489257812,
|
| 1138 |
+
"loss": 0.562,
|
| 1139 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1140 |
+
"rewards/chosen": -2.3766260147094727,
|
| 1141 |
+
"rewards/margins": 0.4747850000858307,
|
| 1142 |
+
"rewards/rejected": -2.8514111042022705,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.85,
|
| 1147 |
+
"learning_rate": 3.4336633249862084e-07,
|
| 1148 |
+
"logits/chosen": 0.07645310461521149,
|
| 1149 |
+
"logits/rejected": 0.17721445858478546,
|
| 1150 |
+
"logps/chosen": -552.8776245117188,
|
| 1151 |
+
"logps/rejected": -514.6980590820312,
|
| 1152 |
+
"loss": 0.5707,
|
| 1153 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1154 |
+
"rewards/chosen": -2.620849847793579,
|
| 1155 |
+
"rewards/margins": 0.33325469493865967,
|
| 1156 |
+
"rewards/rejected": -2.95410418510437,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.86,
|
| 1161 |
+
"learning_rate": 2.98573068519539e-07,
|
| 1162 |
+
"logits/chosen": 0.04426421970129013,
|
| 1163 |
+
"logits/rejected": 0.21165044605731964,
|
| 1164 |
+
"logps/chosen": -541.8455810546875,
|
| 1165 |
+
"logps/rejected": -530.4864501953125,
|
| 1166 |
+
"loss": 0.5871,
|
| 1167 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1168 |
+
"rewards/chosen": -2.5347847938537598,
|
| 1169 |
+
"rewards/margins": 0.558411180973053,
|
| 1170 |
+
"rewards/rejected": -3.093196153640747,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.87,
|
| 1175 |
+
"learning_rate": 2.5672401793681854e-07,
|
| 1176 |
+
"logits/chosen": 0.11617372184991837,
|
| 1177 |
+
"logits/rejected": 0.16498759388923645,
|
| 1178 |
+
"logps/chosen": -526.9368286132812,
|
| 1179 |
+
"logps/rejected": -577.3372802734375,
|
| 1180 |
+
"loss": 0.568,
|
| 1181 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 1182 |
+
"rewards/chosen": -2.5850746631622314,
|
| 1183 |
+
"rewards/margins": 0.4706156253814697,
|
| 1184 |
+
"rewards/rejected": -3.055690050125122,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.88,
|
| 1189 |
+
"learning_rate": 2.178751501463036e-07,
|
| 1190 |
+
"logits/chosen": 0.06383179128170013,
|
| 1191 |
+
"logits/rejected": 0.056464362889528275,
|
| 1192 |
+
"logps/chosen": -508.4523010253906,
|
| 1193 |
+
"logps/rejected": -521.2445068359375,
|
| 1194 |
+
"loss": 0.6203,
|
| 1195 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1196 |
+
"rewards/chosen": -2.663949728012085,
|
| 1197 |
+
"rewards/margins": 0.27986857295036316,
|
| 1198 |
+
"rewards/rejected": -2.9438185691833496,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.89,
|
| 1203 |
+
"learning_rate": 1.820784220652766e-07,
|
| 1204 |
+
"logits/chosen": 0.09964722394943237,
|
| 1205 |
+
"logits/rejected": 0.19454631209373474,
|
| 1206 |
+
"logps/chosen": -553.30517578125,
|
| 1207 |
+
"logps/rejected": -551.3551635742188,
|
| 1208 |
+
"loss": 0.5899,
|
| 1209 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1210 |
+
"rewards/chosen": -2.6559367179870605,
|
| 1211 |
+
"rewards/margins": 0.4735372066497803,
|
| 1212 |
+
"rewards/rejected": -3.12947416305542,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.9,
|
| 1217 |
+
"learning_rate": 1.4938170864468636e-07,
|
| 1218 |
+
"logits/chosen": 0.05580083280801773,
|
| 1219 |
+
"logits/rejected": 0.1297743022441864,
|
| 1220 |
+
"logps/chosen": -535.3272705078125,
|
| 1221 |
+
"logps/rejected": -541.6774291992188,
|
| 1222 |
+
"loss": 0.6136,
|
| 1223 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1224 |
+
"rewards/chosen": -2.696950912475586,
|
| 1225 |
+
"rewards/margins": 0.3545316159725189,
|
| 1226 |
+
"rewards/rejected": -3.0514824390411377,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.91,
|
| 1231 |
+
"learning_rate": 1.1982873884064466e-07,
|
| 1232 |
+
"logits/chosen": 0.04851340129971504,
|
| 1233 |
+
"logits/rejected": 0.13393761217594147,
|
| 1234 |
+
"logps/chosen": -480.47125244140625,
|
| 1235 |
+
"logps/rejected": -523.5387573242188,
|
| 1236 |
+
"loss": 0.592,
|
| 1237 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1238 |
+
"rewards/chosen": -2.486987590789795,
|
| 1239 |
+
"rewards/margins": 0.4138285219669342,
|
| 1240 |
+
"rewards/rejected": -2.900815963745117,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.92,
|
| 1245 |
+
"learning_rate": 9.345903713082305e-08,
|
| 1246 |
+
"logits/chosen": 0.04031291604042053,
|
| 1247 |
+
"logits/rejected": 0.15236200392246246,
|
| 1248 |
+
"logps/chosen": -546.6204833984375,
|
| 1249 |
+
"logps/rejected": -551.7069702148438,
|
| 1250 |
+
"loss": 0.6008,
|
| 1251 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1252 |
+
"rewards/chosen": -2.5852303504943848,
|
| 1253 |
+
"rewards/margins": 0.3579770624637604,
|
| 1254 |
+
"rewards/rejected": -2.943207263946533,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.93,
|
| 1259 |
+
"learning_rate": 7.030787065396866e-08,
|
| 1260 |
+
"logits/chosen": 0.13291895389556885,
|
| 1261 |
+
"logits/rejected": 0.15924690663814545,
|
| 1262 |
+
"logps/chosen": -508.44775390625,
|
| 1263 |
+
"logps/rejected": -528.5171508789062,
|
| 1264 |
+
"loss": 0.6118,
|
| 1265 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1266 |
+
"rewards/chosen": -2.3992390632629395,
|
| 1267 |
+
"rewards/margins": 0.3830375373363495,
|
| 1268 |
+
"rewards/rejected": -2.7822766304016113,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.94,
|
| 1273 |
+
"learning_rate": 5.0406202043228604e-08,
|
| 1274 |
+
"logits/chosen": 0.09626724570989609,
|
| 1275 |
+
"logits/rejected": 0.17059974372386932,
|
| 1276 |
+
"logps/chosen": -527.4398193359375,
|
| 1277 |
+
"logps/rejected": -531.8056640625,
|
| 1278 |
+
"loss": 0.6108,
|
| 1279 |
+
"rewards/accuracies": 0.625,
|
| 1280 |
+
"rewards/chosen": -2.479921817779541,
|
| 1281 |
+
"rewards/margins": 0.4598274827003479,
|
| 1282 |
+
"rewards/rejected": -2.9397494792938232,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.95,
|
| 1287 |
+
"learning_rate": 3.378064801637687e-08,
|
| 1288 |
+
"logits/chosen": 0.02929285541176796,
|
| 1289 |
+
"logits/rejected": 0.18183095753192902,
|
| 1290 |
+
"logps/chosen": -556.4517822265625,
|
| 1291 |
+
"logps/rejected": -552.9277954101562,
|
| 1292 |
+
"loss": 0.5773,
|
| 1293 |
+
"rewards/accuracies": 0.65625,
|
| 1294 |
+
"rewards/chosen": -2.6010327339172363,
|
| 1295 |
+
"rewards/margins": 0.4194592535495758,
|
| 1296 |
+
"rewards/rejected": -3.0204920768737793,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.96,
|
| 1301 |
+
"learning_rate": 2.0453443778310766e-08,
|
| 1302 |
+
"logits/chosen": 0.1071668416261673,
|
| 1303 |
+
"logits/rejected": 0.20206472277641296,
|
| 1304 |
+
"logps/chosen": -557.05419921875,
|
| 1305 |
+
"logps/rejected": -545.5590209960938,
|
| 1306 |
+
"loss": 0.5713,
|
| 1307 |
+
"rewards/accuracies": 0.65625,
|
| 1308 |
+
"rewards/chosen": -2.5374794006347656,
|
| 1309 |
+
"rewards/margins": 0.5199281573295593,
|
| 1310 |
+
"rewards/rejected": -3.057407855987549,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.97,
|
| 1315 |
+
"learning_rate": 1.0442413283435759e-08,
|
| 1316 |
+
"logits/chosen": 0.07009036839008331,
|
| 1317 |
+
"logits/rejected": 0.16552026569843292,
|
| 1318 |
+
"logps/chosen": -491.12237548828125,
|
| 1319 |
+
"logps/rejected": -496.7190856933594,
|
| 1320 |
+
"loss": 0.569,
|
| 1321 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1322 |
+
"rewards/chosen": -2.3404624462127686,
|
| 1323 |
+
"rewards/margins": 0.5322117209434509,
|
| 1324 |
+
"rewards/rejected": -2.8726742267608643,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.98,
|
| 1329 |
+
"learning_rate": 3.760945397705828e-09,
|
| 1330 |
+
"logits/chosen": 0.007874694652855396,
|
| 1331 |
+
"logits/rejected": 0.0997551903128624,
|
| 1332 |
+
"logps/chosen": -552.2734985351562,
|
| 1333 |
+
"logps/rejected": -549.2908935546875,
|
| 1334 |
+
"loss": 0.5796,
|
| 1335 |
+
"rewards/accuracies": 0.65625,
|
| 1336 |
+
"rewards/chosen": -2.5978238582611084,
|
| 1337 |
+
"rewards/margins": 0.399721622467041,
|
| 1338 |
+
"rewards/rejected": -2.9975457191467285,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.99,
|
| 1343 |
+
"learning_rate": 4.1797599220405605e-10,
|
| 1344 |
+
"logits/chosen": 0.1123199313879013,
|
| 1345 |
+
"logits/rejected": 0.08204374462366104,
|
| 1346 |
+
"logps/chosen": -510.1936950683594,
|
| 1347 |
+
"logps/rejected": -531.1867065429688,
|
| 1348 |
+
"loss": 0.5859,
|
| 1349 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1350 |
+
"rewards/chosen": -2.560084819793701,
|
| 1351 |
+
"rewards/margins": 0.4064672589302063,
|
| 1352 |
+
"rewards/rejected": -2.9665520191192627,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.0,
|
| 1357 |
+
"step": 955,
|
| 1358 |
+
"total_flos": 0.0,
|
| 1359 |
+
"train_loss": 0.0333975282639109,
|
| 1360 |
+
"train_runtime": 873.8284,
|
| 1361 |
+
"train_samples_per_second": 69.962,
|
| 1362 |
+
"train_steps_per_second": 1.093
|
| 1363 |
+
}
|
| 1364 |
+
],
|
| 1365 |
+
"logging_steps": 10,
|
| 1366 |
+
"max_steps": 955,
|
| 1367 |
+
"num_input_tokens_seen": 0,
|
| 1368 |
+
"num_train_epochs": 1,
|
| 1369 |
+
"save_steps": 100,
|
| 1370 |
+
"total_flos": 0.0,
|
| 1371 |
+
"train_batch_size": 4,
|
| 1372 |
+
"trial_name": null,
|
| 1373 |
+
"trial_params": null
|
| 1374 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5816
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73fd79cf0fa3e8e41b78f3efed62ad4156ff3350324403480e87b15cb8fe3669
|
| 3 |
size 5816
|