Howard881010
commited on
Commit
•
ec84c3f
1
Parent(s):
8c77e6a
Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +4 -0
- README.md +91 -0
- adapter_config.json +34 -0
- adapter_model.safetensors +3 -0
- all_results.json +20 -0
- checkpoint-1000/README.md +202 -0
- checkpoint-1000/adapter_config.json +34 -0
- checkpoint-1000/adapter_model.safetensors +3 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state_0.pth +3 -0
- checkpoint-1000/rng_state_1.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/special_tokens_map.json +24 -0
- checkpoint-1000/tokenizer.json +3 -0
- checkpoint-1000/tokenizer_config.json +0 -0
- checkpoint-1000/trainer_state.json +1789 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-1125/README.md +202 -0
- checkpoint-1125/adapter_config.json +34 -0
- checkpoint-1125/adapter_model.safetensors +3 -0
- checkpoint-1125/optimizer.pt +3 -0
- checkpoint-1125/rng_state_0.pth +3 -0
- checkpoint-1125/rng_state_1.pth +3 -0
- checkpoint-1125/scheduler.pt +3 -0
- checkpoint-1125/special_tokens_map.json +24 -0
- checkpoint-1125/tokenizer.json +3 -0
- checkpoint-1125/tokenizer_config.json +0 -0
- checkpoint-1125/trainer_state.json +2001 -0
- checkpoint-1125/training_args.bin +3 -0
- checkpoint-500/README.md +202 -0
- checkpoint-500/adapter_config.json +34 -0
- checkpoint-500/adapter_model.safetensors +3 -0
- checkpoint-500/optimizer.pt +3 -0
- checkpoint-500/rng_state_0.pth +3 -0
- checkpoint-500/rng_state_1.pth +3 -0
- checkpoint-500/scheduler.pt +3 -0
- checkpoint-500/special_tokens_map.json +24 -0
- checkpoint-500/tokenizer.json +3 -0
- checkpoint-500/tokenizer_config.json +0 -0
- checkpoint-500/trainer_state.json +911 -0
- checkpoint-500/training_args.bin +3 -0
- eval_results.json +15 -0
- runs/Dec18_22-22-59_yadi/events.out.tfevents.1734560709.yadi.436386.0 +3 -0
- runs/Dec18_22-22-59_yadi/events.out.tfevents.1734574921.yadi.436386.1 +3 -0
- special_tokens_map.json +24 -0
- tokenizer.json +3 -0
- tokenizer_config.json +0 -0
- train_results.json +8 -0
- trainer_log.jsonl +131 -0
- trainer_state.json +2010 -0
.gitattributes
CHANGED
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
checkpoint-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
37 |
+
checkpoint-1125/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
38 |
+
checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
39 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: mistralai/Mistral-Nemo-Instruct-2407
|
3 |
+
library_name: peft
|
4 |
+
license: other
|
5 |
+
tags:
|
6 |
+
- llama-factory
|
7 |
+
- lora
|
8 |
+
- generated_from_trainer
|
9 |
+
model-index:
|
10 |
+
- name: sft_dpo_fs
|
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 |
+
# sft_dpo_fs
|
18 |
+
|
19 |
+
This model is a fine-tuned version of [mistralai/Mistral-Nemo-Instruct-2407](https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407) on the heat_transfer_dpo dataset.
|
20 |
+
It achieves the following results on the evaluation set:
|
21 |
+
- Loss: 0.1535
|
22 |
+
- Rewards/chosen: 17.2823
|
23 |
+
- Rewards/rejected: 11.3004
|
24 |
+
- Rewards/accuracies: 0.9610
|
25 |
+
- Rewards/margins: 5.9819
|
26 |
+
- Logps/chosen: -2.2063
|
27 |
+
- Logps/rejected: -60.6033
|
28 |
+
- Logits/chosen: 0.0035
|
29 |
+
- Logits/rejected: -0.0076
|
30 |
+
|
31 |
+
## Model description
|
32 |
+
|
33 |
+
More information needed
|
34 |
+
|
35 |
+
## Intended uses & limitations
|
36 |
+
|
37 |
+
More information needed
|
38 |
+
|
39 |
+
## Training and evaluation data
|
40 |
+
|
41 |
+
More information needed
|
42 |
+
|
43 |
+
## Training procedure
|
44 |
+
|
45 |
+
### Training hyperparameters
|
46 |
+
|
47 |
+
The following hyperparameters were used during training:
|
48 |
+
- learning_rate: 5e-06
|
49 |
+
- train_batch_size: 4
|
50 |
+
- eval_batch_size: 4
|
51 |
+
- seed: 42
|
52 |
+
- distributed_type: multi-GPU
|
53 |
+
- num_devices: 2
|
54 |
+
- total_train_batch_size: 8
|
55 |
+
- total_eval_batch_size: 8
|
56 |
+
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
57 |
+
- lr_scheduler_type: cosine
|
58 |
+
- lr_scheduler_warmup_ratio: 0.1
|
59 |
+
- num_epochs: 1
|
60 |
+
|
61 |
+
### Training results
|
62 |
+
|
63 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/chosen | Logps/rejected | Logits/chosen | Logits/rejected |
|
64 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:------------:|:--------------:|:-------------:|:---------------:|
|
65 |
+
| 0.3835 | 0.0533 | 60 | 0.3287 | 17.1482 | 15.7095 | 0.9280 | 1.4386 | -3.5472 | -16.5118 | -0.5827 | -0.5914 |
|
66 |
+
| 0.2552 | 0.1067 | 120 | 0.1900 | 17.1335 | 13.7535 | 0.9320 | 3.3799 | -3.6944 | -36.0722 | -0.2065 | -0.2218 |
|
67 |
+
| 0.2362 | 0.16 | 180 | 0.2024 | 17.0614 | 11.9722 | 0.9510 | 5.0892 | -4.4150 | -53.8850 | -0.1087 | -0.1222 |
|
68 |
+
| 0.1781 | 0.2133 | 240 | 0.1546 | 17.0620 | 12.2862 | 0.9500 | 4.7758 | -4.4089 | -50.7448 | -0.1243 | -0.1381 |
|
69 |
+
| 0.265 | 0.2667 | 300 | 0.1536 | 17.2493 | 12.6444 | 0.9440 | 4.6050 | -2.5355 | -47.1637 | -0.1744 | -0.1856 |
|
70 |
+
| 0.1605 | 0.32 | 360 | 0.3194 | 17.3612 | 12.2655 | 0.9210 | 5.0958 | -1.4165 | -50.9525 | -0.1062 | -0.1173 |
|
71 |
+
| 0.2894 | 0.3733 | 420 | 0.1679 | 17.3116 | 12.2496 | 0.9450 | 5.0620 | -1.9131 | -51.1113 | -0.0905 | -0.1026 |
|
72 |
+
| 0.1149 | 0.4267 | 480 | 0.2951 | 17.0540 | 11.9844 | 0.9230 | 5.0696 | -4.4890 | -53.7628 | -0.0770 | -0.0883 |
|
73 |
+
| 0.0384 | 0.48 | 540 | 0.1739 | 17.2042 | 12.1334 | 0.9490 | 5.0708 | -2.9873 | -52.2731 | -0.0512 | -0.0612 |
|
74 |
+
| 0.4008 | 0.5333 | 600 | 0.1706 | 17.2853 | 11.6981 | 0.9470 | 5.5872 | -2.1760 | -56.6266 | -0.0358 | -0.0469 |
|
75 |
+
| 0.1678 | 0.5867 | 660 | 0.2050 | 17.2021 | 11.5656 | 0.9450 | 5.6365 | -3.0082 | -57.9516 | -0.0160 | -0.0270 |
|
76 |
+
| 0.2272 | 0.64 | 720 | 0.1402 | 17.3928 | 11.7696 | 0.9520 | 5.6233 | -1.1005 | -55.9117 | -0.0229 | -0.0322 |
|
77 |
+
| 0.1915 | 0.6933 | 780 | 0.2441 | 17.3947 | 11.7656 | 0.9320 | 5.6290 | -1.0823 | -55.9507 | -0.0166 | -0.0266 |
|
78 |
+
| 0.0635 | 0.7467 | 840 | 0.1689 | 17.3812 | 11.5343 | 0.9450 | 5.8469 | -1.2169 | -58.2643 | -0.0111 | -0.0217 |
|
79 |
+
| 0.1703 | 0.8 | 900 | 0.1400 | 17.3271 | 11.3817 | 0.9610 | 5.9455 | -1.7577 | -59.7906 | 0.0002 | -0.0105 |
|
80 |
+
| 0.1138 | 0.8533 | 960 | 0.1441 | 17.3149 | 11.3432 | 0.9630 | 5.9718 | -1.8795 | -60.1756 | 0.0015 | -0.0094 |
|
81 |
+
| 0.0513 | 0.9067 | 1020 | 0.1412 | 17.3211 | 11.3263 | 0.9610 | 5.9948 | -1.8178 | -60.3445 | 0.0045 | -0.0065 |
|
82 |
+
| 0.1189 | 0.96 | 1080 | 0.1508 | 17.2887 | 11.3001 | 0.9610 | 5.9886 | -2.1420 | -60.6061 | 0.0074 | -0.0036 |
|
83 |
+
|
84 |
+
|
85 |
+
### Framework versions
|
86 |
+
|
87 |
+
- PEFT 0.12.0
|
88 |
+
- Transformers 4.46.0
|
89 |
+
- Pytorch 2.4.0+cu121
|
90 |
+
- Datasets 2.21.0
|
91 |
+
- Tokenizers 0.20.1
|
adapter_config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "mistralai/Mistral-Nemo-Instruct-2407",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layer_replication": null,
|
10 |
+
"layers_pattern": null,
|
11 |
+
"layers_to_transform": null,
|
12 |
+
"loftq_config": {},
|
13 |
+
"lora_alpha": 16,
|
14 |
+
"lora_dropout": 0.0,
|
15 |
+
"megatron_config": null,
|
16 |
+
"megatron_core": "megatron.core",
|
17 |
+
"modules_to_save": null,
|
18 |
+
"peft_type": "LORA",
|
19 |
+
"r": 8,
|
20 |
+
"rank_pattern": {},
|
21 |
+
"revision": null,
|
22 |
+
"target_modules": [
|
23 |
+
"q_proj",
|
24 |
+
"gate_proj",
|
25 |
+
"k_proj",
|
26 |
+
"o_proj",
|
27 |
+
"v_proj",
|
28 |
+
"down_proj",
|
29 |
+
"up_proj"
|
30 |
+
],
|
31 |
+
"task_type": "CAUSAL_LM",
|
32 |
+
"use_dora": false,
|
33 |
+
"use_rslora": false
|
34 |
+
}
|
adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6404a5e1c60aa26a99a4ccbb402f485483d5eea90c08441f0448e29c51e2b1d0
|
3 |
+
size 114106856
|
all_results.json
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"eval_logits/chosen": 0.0034948738757520914,
|
4 |
+
"eval_logits/rejected": -0.007565807551145554,
|
5 |
+
"eval_logps/chosen": -2.20627498626709,
|
6 |
+
"eval_logps/rejected": -60.60332489013672,
|
7 |
+
"eval_loss": 0.15350937843322754,
|
8 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
9 |
+
"eval_rewards/chosen": 17.28226661682129,
|
10 |
+
"eval_rewards/margins": 5.981877326965332,
|
11 |
+
"eval_rewards/rejected": 11.30038833618164,
|
12 |
+
"eval_runtime": 361.5652,
|
13 |
+
"eval_samples_per_second": 2.766,
|
14 |
+
"eval_steps_per_second": 0.346,
|
15 |
+
"total_flos": 1.4338459346927616e+18,
|
16 |
+
"train_loss": 0.23597783709896936,
|
17 |
+
"train_runtime": 13850.7044,
|
18 |
+
"train_samples_per_second": 0.65,
|
19 |
+
"train_steps_per_second": 0.081
|
20 |
+
}
|
checkpoint-1000/README.md
ADDED
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: mistralai/Mistral-Nemo-Instruct-2407
|
3 |
+
library_name: peft
|
4 |
+
---
|
5 |
+
|
6 |
+
# Model Card for Model ID
|
7 |
+
|
8 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
9 |
+
|
10 |
+
|
11 |
+
|
12 |
+
## Model Details
|
13 |
+
|
14 |
+
### Model Description
|
15 |
+
|
16 |
+
<!-- Provide a longer summary of what this model is. -->
|
17 |
+
|
18 |
+
|
19 |
+
|
20 |
+
- **Developed by:** [More Information Needed]
|
21 |
+
- **Funded by [optional]:** [More Information Needed]
|
22 |
+
- **Shared by [optional]:** [More Information Needed]
|
23 |
+
- **Model type:** [More Information Needed]
|
24 |
+
- **Language(s) (NLP):** [More Information Needed]
|
25 |
+
- **License:** [More Information Needed]
|
26 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
27 |
+
|
28 |
+
### Model Sources [optional]
|
29 |
+
|
30 |
+
<!-- Provide the basic links for the model. -->
|
31 |
+
|
32 |
+
- **Repository:** [More Information Needed]
|
33 |
+
- **Paper [optional]:** [More Information Needed]
|
34 |
+
- **Demo [optional]:** [More Information Needed]
|
35 |
+
|
36 |
+
## Uses
|
37 |
+
|
38 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
39 |
+
|
40 |
+
### Direct Use
|
41 |
+
|
42 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
43 |
+
|
44 |
+
[More Information Needed]
|
45 |
+
|
46 |
+
### Downstream Use [optional]
|
47 |
+
|
48 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
49 |
+
|
50 |
+
[More Information Needed]
|
51 |
+
|
52 |
+
### Out-of-Scope Use
|
53 |
+
|
54 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
55 |
+
|
56 |
+
[More Information Needed]
|
57 |
+
|
58 |
+
## Bias, Risks, and Limitations
|
59 |
+
|
60 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
61 |
+
|
62 |
+
[More Information Needed]
|
63 |
+
|
64 |
+
### Recommendations
|
65 |
+
|
66 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
67 |
+
|
68 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
69 |
+
|
70 |
+
## How to Get Started with the Model
|
71 |
+
|
72 |
+
Use the code below to get started with the model.
|
73 |
+
|
74 |
+
[More Information Needed]
|
75 |
+
|
76 |
+
## Training Details
|
77 |
+
|
78 |
+
### Training Data
|
79 |
+
|
80 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
81 |
+
|
82 |
+
[More Information Needed]
|
83 |
+
|
84 |
+
### Training Procedure
|
85 |
+
|
86 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
87 |
+
|
88 |
+
#### Preprocessing [optional]
|
89 |
+
|
90 |
+
[More Information Needed]
|
91 |
+
|
92 |
+
|
93 |
+
#### Training Hyperparameters
|
94 |
+
|
95 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
96 |
+
|
97 |
+
#### Speeds, Sizes, Times [optional]
|
98 |
+
|
99 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
100 |
+
|
101 |
+
[More Information Needed]
|
102 |
+
|
103 |
+
## Evaluation
|
104 |
+
|
105 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
106 |
+
|
107 |
+
### Testing Data, Factors & Metrics
|
108 |
+
|
109 |
+
#### Testing Data
|
110 |
+
|
111 |
+
<!-- This should link to a Dataset Card if possible. -->
|
112 |
+
|
113 |
+
[More Information Needed]
|
114 |
+
|
115 |
+
#### Factors
|
116 |
+
|
117 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
118 |
+
|
119 |
+
[More Information Needed]
|
120 |
+
|
121 |
+
#### Metrics
|
122 |
+
|
123 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
124 |
+
|
125 |
+
[More Information Needed]
|
126 |
+
|
127 |
+
### Results
|
128 |
+
|
129 |
+
[More Information Needed]
|
130 |
+
|
131 |
+
#### Summary
|
132 |
+
|
133 |
+
|
134 |
+
|
135 |
+
## Model Examination [optional]
|
136 |
+
|
137 |
+
<!-- Relevant interpretability work for the model goes here -->
|
138 |
+
|
139 |
+
[More Information Needed]
|
140 |
+
|
141 |
+
## Environmental Impact
|
142 |
+
|
143 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
144 |
+
|
145 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
146 |
+
|
147 |
+
- **Hardware Type:** [More Information Needed]
|
148 |
+
- **Hours used:** [More Information Needed]
|
149 |
+
- **Cloud Provider:** [More Information Needed]
|
150 |
+
- **Compute Region:** [More Information Needed]
|
151 |
+
- **Carbon Emitted:** [More Information Needed]
|
152 |
+
|
153 |
+
## Technical Specifications [optional]
|
154 |
+
|
155 |
+
### Model Architecture and Objective
|
156 |
+
|
157 |
+
[More Information Needed]
|
158 |
+
|
159 |
+
### Compute Infrastructure
|
160 |
+
|
161 |
+
[More Information Needed]
|
162 |
+
|
163 |
+
#### Hardware
|
164 |
+
|
165 |
+
[More Information Needed]
|
166 |
+
|
167 |
+
#### Software
|
168 |
+
|
169 |
+
[More Information Needed]
|
170 |
+
|
171 |
+
## Citation [optional]
|
172 |
+
|
173 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
174 |
+
|
175 |
+
**BibTeX:**
|
176 |
+
|
177 |
+
[More Information Needed]
|
178 |
+
|
179 |
+
**APA:**
|
180 |
+
|
181 |
+
[More Information Needed]
|
182 |
+
|
183 |
+
## Glossary [optional]
|
184 |
+
|
185 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
186 |
+
|
187 |
+
[More Information Needed]
|
188 |
+
|
189 |
+
## More Information [optional]
|
190 |
+
|
191 |
+
[More Information Needed]
|
192 |
+
|
193 |
+
## Model Card Authors [optional]
|
194 |
+
|
195 |
+
[More Information Needed]
|
196 |
+
|
197 |
+
## Model Card Contact
|
198 |
+
|
199 |
+
[More Information Needed]
|
200 |
+
### Framework versions
|
201 |
+
|
202 |
+
- PEFT 0.12.0
|
checkpoint-1000/adapter_config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "mistralai/Mistral-Nemo-Instruct-2407",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layer_replication": null,
|
10 |
+
"layers_pattern": null,
|
11 |
+
"layers_to_transform": null,
|
12 |
+
"loftq_config": {},
|
13 |
+
"lora_alpha": 16,
|
14 |
+
"lora_dropout": 0.0,
|
15 |
+
"megatron_config": null,
|
16 |
+
"megatron_core": "megatron.core",
|
17 |
+
"modules_to_save": null,
|
18 |
+
"peft_type": "LORA",
|
19 |
+
"r": 8,
|
20 |
+
"rank_pattern": {},
|
21 |
+
"revision": null,
|
22 |
+
"target_modules": [
|
23 |
+
"q_proj",
|
24 |
+
"gate_proj",
|
25 |
+
"k_proj",
|
26 |
+
"o_proj",
|
27 |
+
"v_proj",
|
28 |
+
"down_proj",
|
29 |
+
"up_proj"
|
30 |
+
],
|
31 |
+
"task_type": "CAUSAL_LM",
|
32 |
+
"use_dora": false,
|
33 |
+
"use_rslora": false
|
34 |
+
}
|
checkpoint-1000/adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7b661db279de1245ec057ccc9ea7b2a9d8e593601b17b234d71a8b65eea724cc
|
3 |
+
size 114106856
|
checkpoint-1000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b5df0d088d43a257985e1c6b5e7afebb17236c24d7243b77ad1477e87251acd3
|
3 |
+
size 228536930
|
checkpoint-1000/rng_state_0.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:41d62848396b464be71e7417b09981050d5f1d377e9e0c4894645772a871d81e
|
3 |
+
size 14512
|
checkpoint-1000/rng_state_1.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a408fa3a85250385243240a3c4f6e6dc4513da1d97421e8faae4c8d880e7d1a
|
3 |
+
size 14512
|
checkpoint-1000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4549dc1554a9cef5b3b89b84239d0624d2e4f7b2af98173873173c77ff786861
|
3 |
+
size 1064
|
checkpoint-1000/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "</s>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
checkpoint-1000/tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0240ce510f08e6c2041724e9043e33be9d251d1e4a4d94eb68cd47b954b61d2
|
3 |
+
size 17078292
|
checkpoint-1000/tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-1000/trainer_state.json
ADDED
@@ -0,0 +1,1789 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.8888888888888888,
|
5 |
+
"eval_steps": 60,
|
6 |
+
"global_step": 1000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.008888888888888889,
|
13 |
+
"grad_norm": 8.760091781616211,
|
14 |
+
"learning_rate": 4.4247787610619474e-07,
|
15 |
+
"logits/chosen": -0.8248252868652344,
|
16 |
+
"logits/rejected": -0.8263720273971558,
|
17 |
+
"logps/chosen": -0.36086463928222656,
|
18 |
+
"logps/rejected": -5.696224689483643,
|
19 |
+
"loss": 1.1038,
|
20 |
+
"rewards/accuracies": 0.5125000476837158,
|
21 |
+
"rewards/chosen": 17.43745994567871,
|
22 |
+
"rewards/margins": 0.5984855890274048,
|
23 |
+
"rewards/rejected": 16.838973999023438,
|
24 |
+
"step": 10
|
25 |
+
},
|
26 |
+
{
|
27 |
+
"epoch": 0.017777777777777778,
|
28 |
+
"grad_norm": 8.855981826782227,
|
29 |
+
"learning_rate": 8.849557522123895e-07,
|
30 |
+
"logits/chosen": -0.8169006109237671,
|
31 |
+
"logits/rejected": -0.819770872592926,
|
32 |
+
"logps/chosen": -0.12464660406112671,
|
33 |
+
"logps/rejected": -7.139842987060547,
|
34 |
+
"loss": 1.1887,
|
35 |
+
"rewards/accuracies": 0.4000000059604645,
|
36 |
+
"rewards/chosen": 17.17649269104004,
|
37 |
+
"rewards/margins": 0.19107049703598022,
|
38 |
+
"rewards/rejected": 16.98542022705078,
|
39 |
+
"step": 20
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 0.02666666666666667,
|
43 |
+
"grad_norm": 16.764184951782227,
|
44 |
+
"learning_rate": 1.3274336283185843e-06,
|
45 |
+
"logits/chosen": -0.8003113865852356,
|
46 |
+
"logits/rejected": -0.8030117750167847,
|
47 |
+
"logps/chosen": -0.34651467204093933,
|
48 |
+
"logps/rejected": -6.967917442321777,
|
49 |
+
"loss": 1.0563,
|
50 |
+
"rewards/accuracies": 0.44999998807907104,
|
51 |
+
"rewards/chosen": 17.280975341796875,
|
52 |
+
"rewards/margins": 0.40005987882614136,
|
53 |
+
"rewards/rejected": 16.88091468811035,
|
54 |
+
"step": 30
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 0.035555555555555556,
|
58 |
+
"grad_norm": 8.33682918548584,
|
59 |
+
"learning_rate": 1.769911504424779e-06,
|
60 |
+
"logits/chosen": -0.7695047855377197,
|
61 |
+
"logits/rejected": -0.7739207148551941,
|
62 |
+
"logps/chosen": -1.5993006229400635,
|
63 |
+
"logps/rejected": -8.504932403564453,
|
64 |
+
"loss": 0.7596,
|
65 |
+
"rewards/accuracies": 0.5,
|
66 |
+
"rewards/chosen": 17.283912658691406,
|
67 |
+
"rewards/margins": 0.6976072192192078,
|
68 |
+
"rewards/rejected": 16.5863037109375,
|
69 |
+
"step": 40
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 0.044444444444444446,
|
73 |
+
"grad_norm": 4.494723320007324,
|
74 |
+
"learning_rate": 2.212389380530974e-06,
|
75 |
+
"logits/chosen": -0.7154140472412109,
|
76 |
+
"logits/rejected": -0.7225576043128967,
|
77 |
+
"logps/chosen": -3.112199068069458,
|
78 |
+
"logps/rejected": -12.212080001831055,
|
79 |
+
"loss": 0.6083,
|
80 |
+
"rewards/accuracies": 0.4625000059604645,
|
81 |
+
"rewards/chosen": 17.03064727783203,
|
82 |
+
"rewards/margins": 0.7148451805114746,
|
83 |
+
"rewards/rejected": 16.3158016204834,
|
84 |
+
"step": 50
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.05333333333333334,
|
88 |
+
"grad_norm": 5.110287666320801,
|
89 |
+
"learning_rate": 2.6548672566371687e-06,
|
90 |
+
"logits/chosen": -0.6322453022003174,
|
91 |
+
"logits/rejected": -0.6387485265731812,
|
92 |
+
"logps/chosen": -5.650620460510254,
|
93 |
+
"logps/rejected": -12.759811401367188,
|
94 |
+
"loss": 0.3835,
|
95 |
+
"rewards/accuracies": 0.9125000238418579,
|
96 |
+
"rewards/chosen": 17.101289749145508,
|
97 |
+
"rewards/margins": 1.1824612617492676,
|
98 |
+
"rewards/rejected": 15.918828964233398,
|
99 |
+
"step": 60
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.05333333333333334,
|
103 |
+
"eval_logits/chosen": -0.5826543569564819,
|
104 |
+
"eval_logits/rejected": -0.5914276838302612,
|
105 |
+
"eval_logps/chosen": -3.5471787452697754,
|
106 |
+
"eval_logps/rejected": -16.51181983947754,
|
107 |
+
"eval_loss": 0.3286525011062622,
|
108 |
+
"eval_rewards/accuracies": 0.9280000925064087,
|
109 |
+
"eval_rewards/chosen": 17.148174285888672,
|
110 |
+
"eval_rewards/margins": 1.4386365413665771,
|
111 |
+
"eval_rewards/rejected": 15.709539413452148,
|
112 |
+
"eval_runtime": 372.0227,
|
113 |
+
"eval_samples_per_second": 2.688,
|
114 |
+
"eval_steps_per_second": 0.336,
|
115 |
+
"step": 60
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 0.06222222222222222,
|
119 |
+
"grad_norm": 5.098133563995361,
|
120 |
+
"learning_rate": 3.097345132743363e-06,
|
121 |
+
"logits/chosen": -0.5378152132034302,
|
122 |
+
"logits/rejected": -0.5494933724403381,
|
123 |
+
"logps/chosen": -1.5099802017211914,
|
124 |
+
"logps/rejected": -21.206321716308594,
|
125 |
+
"loss": 0.2931,
|
126 |
+
"rewards/accuracies": 0.9375,
|
127 |
+
"rewards/chosen": 17.083791732788086,
|
128 |
+
"rewards/margins": 1.5844331979751587,
|
129 |
+
"rewards/rejected": 15.499359130859375,
|
130 |
+
"step": 70
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 0.07111111111111111,
|
134 |
+
"grad_norm": 29.787437438964844,
|
135 |
+
"learning_rate": 3.539823008849558e-06,
|
136 |
+
"logits/chosen": -0.443774938583374,
|
137 |
+
"logits/rejected": -0.45571577548980713,
|
138 |
+
"logps/chosen": -1.5804342031478882,
|
139 |
+
"logps/rejected": -22.606929779052734,
|
140 |
+
"loss": 0.202,
|
141 |
+
"rewards/accuracies": 0.9750000238418579,
|
142 |
+
"rewards/chosen": 17.302125930786133,
|
143 |
+
"rewards/margins": 2.174014091491699,
|
144 |
+
"rewards/rejected": 15.128110885620117,
|
145 |
+
"step": 80
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"epoch": 0.08,
|
149 |
+
"grad_norm": 23.14398193359375,
|
150 |
+
"learning_rate": 3.982300884955752e-06,
|
151 |
+
"logits/chosen": -0.3626072406768799,
|
152 |
+
"logits/rejected": -0.3787815570831299,
|
153 |
+
"logps/chosen": -2.203828811645508,
|
154 |
+
"logps/rejected": -29.433551788330078,
|
155 |
+
"loss": 0.2123,
|
156 |
+
"rewards/accuracies": 0.925000011920929,
|
157 |
+
"rewards/chosen": 17.00284194946289,
|
158 |
+
"rewards/margins": 2.320391893386841,
|
159 |
+
"rewards/rejected": 14.682450294494629,
|
160 |
+
"step": 90
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 0.08888888888888889,
|
164 |
+
"grad_norm": 29.672739028930664,
|
165 |
+
"learning_rate": 4.424778761061948e-06,
|
166 |
+
"logits/chosen": -0.3035663962364197,
|
167 |
+
"logits/rejected": -0.31762221455574036,
|
168 |
+
"logps/chosen": -3.433589458465576,
|
169 |
+
"logps/rejected": -29.9322509765625,
|
170 |
+
"loss": 0.2592,
|
171 |
+
"rewards/accuracies": 0.9375,
|
172 |
+
"rewards/chosen": 16.929956436157227,
|
173 |
+
"rewards/margins": 2.31272029876709,
|
174 |
+
"rewards/rejected": 14.617237091064453,
|
175 |
+
"step": 100
|
176 |
+
},
|
177 |
+
{
|
178 |
+
"epoch": 0.09777777777777778,
|
179 |
+
"grad_norm": 1.873722791671753,
|
180 |
+
"learning_rate": 4.867256637168142e-06,
|
181 |
+
"logits/chosen": -0.2679600715637207,
|
182 |
+
"logits/rejected": -0.2826440930366516,
|
183 |
+
"logps/chosen": -0.9653514623641968,
|
184 |
+
"logps/rejected": -30.235322952270508,
|
185 |
+
"loss": 0.1336,
|
186 |
+
"rewards/accuracies": 0.949999988079071,
|
187 |
+
"rewards/chosen": 17.462385177612305,
|
188 |
+
"rewards/margins": 3.1994175910949707,
|
189 |
+
"rewards/rejected": 14.26296615600586,
|
190 |
+
"step": 110
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 0.10666666666666667,
|
194 |
+
"grad_norm": 1.6913721561431885,
|
195 |
+
"learning_rate": 4.999409761242696e-06,
|
196 |
+
"logits/chosen": -0.22222033143043518,
|
197 |
+
"logits/rejected": -0.23720571398735046,
|
198 |
+
"logps/chosen": -4.4953508377075195,
|
199 |
+
"logps/rejected": -34.074745178222656,
|
200 |
+
"loss": 0.2552,
|
201 |
+
"rewards/accuracies": 0.8999999761581421,
|
202 |
+
"rewards/chosen": 17.04866600036621,
|
203 |
+
"rewards/margins": 3.1014418601989746,
|
204 |
+
"rewards/rejected": 13.947224617004395,
|
205 |
+
"step": 120
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.10666666666666667,
|
209 |
+
"eval_logits/chosen": -0.206527978181839,
|
210 |
+
"eval_logits/rejected": -0.22178640961647034,
|
211 |
+
"eval_logps/chosen": -3.69442081451416,
|
212 |
+
"eval_logps/rejected": -36.072166442871094,
|
213 |
+
"eval_loss": 0.18996010720729828,
|
214 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
215 |
+
"eval_rewards/chosen": 17.133451461791992,
|
216 |
+
"eval_rewards/margins": 3.379946708679199,
|
217 |
+
"eval_rewards/rejected": 13.753504753112793,
|
218 |
+
"eval_runtime": 361.5279,
|
219 |
+
"eval_samples_per_second": 2.766,
|
220 |
+
"eval_steps_per_second": 0.346,
|
221 |
+
"step": 120
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 0.11555555555555555,
|
225 |
+
"grad_norm": 61.80262756347656,
|
226 |
+
"learning_rate": 4.996519466816778e-06,
|
227 |
+
"logits/chosen": -0.18473535776138306,
|
228 |
+
"logits/rejected": -0.1988501250743866,
|
229 |
+
"logps/chosen": -3.7009687423706055,
|
230 |
+
"logps/rejected": -39.289939880371094,
|
231 |
+
"loss": 0.1394,
|
232 |
+
"rewards/accuracies": 0.9624999761581421,
|
233 |
+
"rewards/chosen": 17.106964111328125,
|
234 |
+
"rewards/margins": 3.633338212966919,
|
235 |
+
"rewards/rejected": 13.473625183105469,
|
236 |
+
"step": 130
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.12444444444444444,
|
240 |
+
"grad_norm": 1.6732702255249023,
|
241 |
+
"learning_rate": 4.9912234871722805e-06,
|
242 |
+
"logits/chosen": -0.16134041547775269,
|
243 |
+
"logits/rejected": -0.17547868192195892,
|
244 |
+
"logps/chosen": -3.0637736320495605,
|
245 |
+
"logps/rejected": -40.07548522949219,
|
246 |
+
"loss": 0.1408,
|
247 |
+
"rewards/accuracies": 0.9750000238418579,
|
248 |
+
"rewards/chosen": 17.392223358154297,
|
249 |
+
"rewards/margins": 4.242353439331055,
|
250 |
+
"rewards/rejected": 13.149867057800293,
|
251 |
+
"step": 140
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 0.13333333333333333,
|
255 |
+
"grad_norm": 0.346453994512558,
|
256 |
+
"learning_rate": 4.98352692559805e-06,
|
257 |
+
"logits/chosen": -0.13797929883003235,
|
258 |
+
"logits/rejected": -0.15283086895942688,
|
259 |
+
"logps/chosen": -5.14492130279541,
|
260 |
+
"logps/rejected": -47.97212219238281,
|
261 |
+
"loss": 0.2153,
|
262 |
+
"rewards/accuracies": 0.9375,
|
263 |
+
"rewards/chosen": 16.896778106689453,
|
264 |
+
"rewards/margins": 4.227695465087891,
|
265 |
+
"rewards/rejected": 12.669081687927246,
|
266 |
+
"step": 150
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.14222222222222222,
|
270 |
+
"grad_norm": 0.21871662139892578,
|
271 |
+
"learning_rate": 4.973437198621237e-06,
|
272 |
+
"logits/chosen": -0.12396670132875443,
|
273 |
+
"logits/rejected": -0.13780555129051208,
|
274 |
+
"logps/chosen": -6.108860015869141,
|
275 |
+
"logps/rejected": -54.90739440917969,
|
276 |
+
"loss": 0.0388,
|
277 |
+
"rewards/accuracies": 0.9750000238418579,
|
278 |
+
"rewards/chosen": 16.75935935974121,
|
279 |
+
"rewards/margins": 4.755282878875732,
|
280 |
+
"rewards/rejected": 12.004077911376953,
|
281 |
+
"step": 160
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.1511111111111111,
|
285 |
+
"grad_norm": 235.12429809570312,
|
286 |
+
"learning_rate": 4.960964028860621e-06,
|
287 |
+
"logits/chosen": -0.1140839159488678,
|
288 |
+
"logits/rejected": -0.1263057291507721,
|
289 |
+
"logps/chosen": -12.605452537536621,
|
290 |
+
"logps/rejected": -53.81230926513672,
|
291 |
+
"loss": 0.4651,
|
292 |
+
"rewards/accuracies": 0.875,
|
293 |
+
"rewards/chosen": 16.101238250732422,
|
294 |
+
"rewards/margins": 3.9864249229431152,
|
295 |
+
"rewards/rejected": 12.114812850952148,
|
296 |
+
"step": 170
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.16,
|
300 |
+
"grad_norm": 190.97048950195312,
|
301 |
+
"learning_rate": 4.946119435657738e-06,
|
302 |
+
"logits/chosen": -0.10746976733207703,
|
303 |
+
"logits/rejected": -0.11878640949726105,
|
304 |
+
"logps/chosen": -8.5105562210083,
|
305 |
+
"logps/rejected": -51.314781188964844,
|
306 |
+
"loss": 0.2362,
|
307 |
+
"rewards/accuracies": 0.925000011920929,
|
308 |
+
"rewards/chosen": 16.719980239868164,
|
309 |
+
"rewards/margins": 4.549674034118652,
|
310 |
+
"rewards/rejected": 12.170306205749512,
|
311 |
+
"step": 180
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 0.16,
|
315 |
+
"eval_logits/chosen": -0.10870806127786636,
|
316 |
+
"eval_logits/rejected": -0.12223993986845016,
|
317 |
+
"eval_logps/chosen": -4.414996147155762,
|
318 |
+
"eval_logps/rejected": -53.885032653808594,
|
319 |
+
"eval_loss": 0.20236633718013763,
|
320 |
+
"eval_rewards/accuracies": 0.9510000944137573,
|
321 |
+
"eval_rewards/chosen": 17.06139373779297,
|
322 |
+
"eval_rewards/margins": 5.089176177978516,
|
323 |
+
"eval_rewards/rejected": 11.97221851348877,
|
324 |
+
"eval_runtime": 361.4355,
|
325 |
+
"eval_samples_per_second": 2.767,
|
326 |
+
"eval_steps_per_second": 0.346,
|
327 |
+
"step": 180
|
328 |
+
},
|
329 |
+
{
|
330 |
+
"epoch": 0.1688888888888889,
|
331 |
+
"grad_norm": 56.81266784667969,
|
332 |
+
"learning_rate": 4.928917723494854e-06,
|
333 |
+
"logits/chosen": -0.10682469606399536,
|
334 |
+
"logits/rejected": -0.12124393880367279,
|
335 |
+
"logps/chosen": -3.058413028717041,
|
336 |
+
"logps/rejected": -55.052528381347656,
|
337 |
+
"loss": 0.2442,
|
338 |
+
"rewards/accuracies": 0.9500000476837158,
|
339 |
+
"rewards/chosen": 17.058589935302734,
|
340 |
+
"rewards/margins": 5.056097984313965,
|
341 |
+
"rewards/rejected": 12.002490043640137,
|
342 |
+
"step": 190
|
343 |
+
},
|
344 |
+
{
|
345 |
+
"epoch": 0.17777777777777778,
|
346 |
+
"grad_norm": 175.06552124023438,
|
347 |
+
"learning_rate": 4.909375468210947e-06,
|
348 |
+
"logits/chosen": -0.10520349442958832,
|
349 |
+
"logits/rejected": -0.12018950283527374,
|
350 |
+
"logps/chosen": -4.114959716796875,
|
351 |
+
"logps/rejected": -55.9394645690918,
|
352 |
+
"loss": 0.1915,
|
353 |
+
"rewards/accuracies": 0.9500000476837158,
|
354 |
+
"rewards/chosen": 16.98603057861328,
|
355 |
+
"rewards/margins": 5.105838775634766,
|
356 |
+
"rewards/rejected": 11.880191802978516,
|
357 |
+
"step": 200
|
358 |
+
},
|
359 |
+
{
|
360 |
+
"epoch": 0.18666666666666668,
|
361 |
+
"grad_norm": 78.06558990478516,
|
362 |
+
"learning_rate": 4.8875115010289655e-06,
|
363 |
+
"logits/chosen": -0.10475558042526245,
|
364 |
+
"logits/rejected": -0.11949175596237183,
|
365 |
+
"logps/chosen": -6.760301113128662,
|
366 |
+
"logps/rejected": -53.91607666015625,
|
367 |
+
"loss": 0.2843,
|
368 |
+
"rewards/accuracies": 0.9375,
|
369 |
+
"rewards/chosen": 16.857545852661133,
|
370 |
+
"rewards/margins": 4.917357921600342,
|
371 |
+
"rewards/rejected": 11.94018840789795,
|
372 |
+
"step": 210
|
373 |
+
},
|
374 |
+
{
|
375 |
+
"epoch": 0.19555555555555557,
|
376 |
+
"grad_norm": 15.880486488342285,
|
377 |
+
"learning_rate": 4.863346890409768e-06,
|
378 |
+
"logits/chosen": -0.11213523149490356,
|
379 |
+
"logits/rejected": -0.12581588327884674,
|
380 |
+
"logps/chosen": -6.759585380554199,
|
381 |
+
"logps/rejected": -51.10936737060547,
|
382 |
+
"loss": 0.5104,
|
383 |
+
"rewards/accuracies": 0.875,
|
384 |
+
"rewards/chosen": 16.859071731567383,
|
385 |
+
"rewards/margins": 4.638372898101807,
|
386 |
+
"rewards/rejected": 12.220698356628418,
|
387 |
+
"step": 220
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.20444444444444446,
|
391 |
+
"grad_norm": 46.97845458984375,
|
392 |
+
"learning_rate": 4.836904921750224e-06,
|
393 |
+
"logits/chosen": -0.11947059631347656,
|
394 |
+
"logits/rejected": -0.1329912692308426,
|
395 |
+
"logps/chosen": -3.608184814453125,
|
396 |
+
"logps/rejected": -48.794761657714844,
|
397 |
+
"loss": 0.2134,
|
398 |
+
"rewards/accuracies": 0.925000011920929,
|
399 |
+
"rewards/chosen": 17.235904693603516,
|
400 |
+
"rewards/margins": 4.859888076782227,
|
401 |
+
"rewards/rejected": 12.376014709472656,
|
402 |
+
"step": 230
|
403 |
+
},
|
404 |
+
{
|
405 |
+
"epoch": 0.21333333333333335,
|
406 |
+
"grad_norm": 24.032859802246094,
|
407 |
+
"learning_rate": 4.808211074945042e-06,
|
408 |
+
"logits/chosen": -0.1200513243675232,
|
409 |
+
"logits/rejected": -0.1333036869764328,
|
410 |
+
"logps/chosen": -3.7552154064178467,
|
411 |
+
"logps/rejected": -49.87453079223633,
|
412 |
+
"loss": 0.1781,
|
413 |
+
"rewards/accuracies": 0.9500000476837158,
|
414 |
+
"rewards/chosen": 17.094650268554688,
|
415 |
+
"rewards/margins": 4.68077278137207,
|
416 |
+
"rewards/rejected": 12.41387939453125,
|
417 |
+
"step": 240
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 0.21333333333333335,
|
421 |
+
"eval_logits/chosen": -0.12433278560638428,
|
422 |
+
"eval_logits/rejected": -0.13808581233024597,
|
423 |
+
"eval_logps/chosen": -4.408891201019287,
|
424 |
+
"eval_logps/rejected": -50.744781494140625,
|
425 |
+
"eval_loss": 0.1546352356672287,
|
426 |
+
"eval_rewards/accuracies": 0.9500000476837158,
|
427 |
+
"eval_rewards/chosen": 17.06200408935547,
|
428 |
+
"eval_rewards/margins": 4.775761604309082,
|
429 |
+
"eval_rewards/rejected": 12.286243438720703,
|
430 |
+
"eval_runtime": 361.4974,
|
431 |
+
"eval_samples_per_second": 2.766,
|
432 |
+
"eval_steps_per_second": 0.346,
|
433 |
+
"step": 240
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 0.2222222222222222,
|
437 |
+
"grad_norm": 0.25737640261650085,
|
438 |
+
"learning_rate": 4.7772929998339485e-06,
|
439 |
+
"logits/chosen": -0.12348780035972595,
|
440 |
+
"logits/rejected": -0.13704943656921387,
|
441 |
+
"logps/chosen": -4.4299187660217285,
|
442 |
+
"logps/rejected": -53.074607849121094,
|
443 |
+
"loss": 0.1373,
|
444 |
+
"rewards/accuracies": 0.9375,
|
445 |
+
"rewards/chosen": 17.087068557739258,
|
446 |
+
"rewards/margins": 5.06691837310791,
|
447 |
+
"rewards/rejected": 12.020149230957031,
|
448 |
+
"step": 250
|
449 |
+
},
|
450 |
+
{
|
451 |
+
"epoch": 0.2311111111111111,
|
452 |
+
"grad_norm": 0.1839389204978943,
|
453 |
+
"learning_rate": 4.744180489557859e-06,
|
454 |
+
"logits/chosen": -0.12177034467458725,
|
455 |
+
"logits/rejected": -0.1342695653438568,
|
456 |
+
"logps/chosen": -3.775188446044922,
|
457 |
+
"logps/rejected": -53.98720932006836,
|
458 |
+
"loss": 0.1896,
|
459 |
+
"rewards/accuracies": 0.949999988079071,
|
460 |
+
"rewards/chosen": 17.12021255493164,
|
461 |
+
"rewards/margins": 5.148064613342285,
|
462 |
+
"rewards/rejected": 11.972146987915039,
|
463 |
+
"step": 260
|
464 |
+
},
|
465 |
+
{
|
466 |
+
"epoch": 0.24,
|
467 |
+
"grad_norm": 12.258485794067383,
|
468 |
+
"learning_rate": 4.708905451849754e-06,
|
469 |
+
"logits/chosen": -0.11067859083414078,
|
470 |
+
"logits/rejected": -0.12377731502056122,
|
471 |
+
"logps/chosen": -6.418317794799805,
|
472 |
+
"logps/rejected": -56.57402801513672,
|
473 |
+
"loss": 0.2315,
|
474 |
+
"rewards/accuracies": 0.9375,
|
475 |
+
"rewards/chosen": 16.738832473754883,
|
476 |
+
"rewards/margins": 4.884931564331055,
|
477 |
+
"rewards/rejected": 11.853900909423828,
|
478 |
+
"step": 270
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.24888888888888888,
|
482 |
+
"grad_norm": 77.56194305419922,
|
483 |
+
"learning_rate": 4.671501878287879e-06,
|
484 |
+
"logits/chosen": -0.1184445172548294,
|
485 |
+
"logits/rejected": -0.1339874565601349,
|
486 |
+
"logps/chosen": -10.12116527557373,
|
487 |
+
"logps/rejected": -53.403907775878906,
|
488 |
+
"loss": 0.5343,
|
489 |
+
"rewards/accuracies": 0.862500011920929,
|
490 |
+
"rewards/chosen": 16.458633422851562,
|
491 |
+
"rewards/margins": 4.402472496032715,
|
492 |
+
"rewards/rejected": 12.056160926818848,
|
493 |
+
"step": 280
|
494 |
+
},
|
495 |
+
{
|
496 |
+
"epoch": 0.2577777777777778,
|
497 |
+
"grad_norm": 67.53883361816406,
|
498 |
+
"learning_rate": 4.6320058115409295e-06,
|
499 |
+
"logits/chosen": -0.1448262631893158,
|
500 |
+
"logits/rejected": -0.15793387591838837,
|
501 |
+
"logps/chosen": -3.4666190147399902,
|
502 |
+
"logps/rejected": -48.79213333129883,
|
503 |
+
"loss": 0.5017,
|
504 |
+
"rewards/accuracies": 0.887499988079071,
|
505 |
+
"rewards/chosen": 16.945899963378906,
|
506 |
+
"rewards/margins": 4.2686333656311035,
|
507 |
+
"rewards/rejected": 12.677268028259277,
|
508 |
+
"step": 290
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 0.26666666666666666,
|
512 |
+
"grad_norm": 0.17521341145038605,
|
513 |
+
"learning_rate": 4.590455310636778e-06,
|
514 |
+
"logits/chosen": -0.16128253936767578,
|
515 |
+
"logits/rejected": -0.17375555634498596,
|
516 |
+
"logps/chosen": -2.9032950401306152,
|
517 |
+
"logps/rejected": -47.69734191894531,
|
518 |
+
"loss": 0.265,
|
519 |
+
"rewards/accuracies": 0.925000011920929,
|
520 |
+
"rewards/chosen": 17.18383026123047,
|
521 |
+
"rewards/margins": 4.541309356689453,
|
522 |
+
"rewards/rejected": 12.642518997192383,
|
523 |
+
"step": 300
|
524 |
+
},
|
525 |
+
{
|
526 |
+
"epoch": 0.26666666666666666,
|
527 |
+
"eval_logits/chosen": -0.17444846034049988,
|
528 |
+
"eval_logits/rejected": -0.18559777736663818,
|
529 |
+
"eval_logps/chosen": -2.535512924194336,
|
530 |
+
"eval_logps/rejected": -47.16367721557617,
|
531 |
+
"eval_loss": 0.15360687673091888,
|
532 |
+
"eval_rewards/accuracies": 0.9440000653266907,
|
533 |
+
"eval_rewards/chosen": 17.249343872070312,
|
534 |
+
"eval_rewards/margins": 4.604989051818848,
|
535 |
+
"eval_rewards/rejected": 12.644353866577148,
|
536 |
+
"eval_runtime": 361.4575,
|
537 |
+
"eval_samples_per_second": 2.767,
|
538 |
+
"eval_steps_per_second": 0.346,
|
539 |
+
"step": 300
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 0.27555555555555555,
|
543 |
+
"grad_norm": 0.5040452480316162,
|
544 |
+
"learning_rate": 4.54689041428819e-06,
|
545 |
+
"logits/chosen": -0.16974106431007385,
|
546 |
+
"logits/rejected": -0.1810058057308197,
|
547 |
+
"logps/chosen": -1.233938217163086,
|
548 |
+
"logps/rejected": -49.907745361328125,
|
549 |
+
"loss": 0.1132,
|
550 |
+
"rewards/accuracies": 0.9500000476837158,
|
551 |
+
"rewards/chosen": 17.34117889404297,
|
552 |
+
"rewards/margins": 4.934173583984375,
|
553 |
+
"rewards/rejected": 12.407005310058594,
|
554 |
+
"step": 310
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 0.28444444444444444,
|
558 |
+
"grad_norm": 100.02949523925781,
|
559 |
+
"learning_rate": 4.501353102310901e-06,
|
560 |
+
"logits/chosen": -0.15705889463424683,
|
561 |
+
"logits/rejected": -0.1695334017276764,
|
562 |
+
"logps/chosen": -1.0820492506027222,
|
563 |
+
"logps/rejected": -52.577110290527344,
|
564 |
+
"loss": 0.1194,
|
565 |
+
"rewards/accuracies": 0.9500000476837158,
|
566 |
+
"rewards/chosen": 17.33388900756836,
|
567 |
+
"rewards/margins": 5.154760837554932,
|
568 |
+
"rewards/rejected": 12.179126739501953,
|
569 |
+
"step": 320
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.29333333333333333,
|
573 |
+
"grad_norm": 0.2689219117164612,
|
574 |
+
"learning_rate": 4.453887255171206e-06,
|
575 |
+
"logits/chosen": -0.13849371671676636,
|
576 |
+
"logits/rejected": -0.14990833401679993,
|
577 |
+
"logps/chosen": -1.8435032367706299,
|
578 |
+
"logps/rejected": -54.79044723510742,
|
579 |
+
"loss": 0.0926,
|
580 |
+
"rewards/accuracies": 0.9500000476837158,
|
581 |
+
"rewards/chosen": 17.2423095703125,
|
582 |
+
"rewards/margins": 5.28987979888916,
|
583 |
+
"rewards/rejected": 11.952428817749023,
|
584 |
+
"step": 330
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 0.3022222222222222,
|
588 |
+
"grad_norm": 0.09305431693792343,
|
589 |
+
"learning_rate": 4.404538611702055e-06,
|
590 |
+
"logits/chosen": -0.12299702316522598,
|
591 |
+
"logits/rejected": -0.13453055918216705,
|
592 |
+
"logps/chosen": -2.9897143840789795,
|
593 |
+
"logps/rejected": -52.954498291015625,
|
594 |
+
"loss": 0.2873,
|
595 |
+
"rewards/accuracies": 0.925000011920929,
|
596 |
+
"rewards/chosen": 17.17474365234375,
|
597 |
+
"rewards/margins": 5.071004867553711,
|
598 |
+
"rewards/rejected": 12.103739738464355,
|
599 |
+
"step": 340
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 0.3111111111111111,
|
603 |
+
"grad_norm": 59.282073974609375,
|
604 |
+
"learning_rate": 4.3533547250284015e-06,
|
605 |
+
"logits/chosen": -0.11913029849529266,
|
606 |
+
"logits/rejected": -0.12785324454307556,
|
607 |
+
"logps/chosen": -3.9456872940063477,
|
608 |
+
"logps/rejected": -48.68487548828125,
|
609 |
+
"loss": 0.4332,
|
610 |
+
"rewards/accuracies": 0.875,
|
611 |
+
"rewards/chosen": 17.12805938720703,
|
612 |
+
"rewards/margins": 4.669450283050537,
|
613 |
+
"rewards/rejected": 12.458610534667969,
|
614 |
+
"step": 350
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.32,
|
618 |
+
"grad_norm": 0.31101909279823303,
|
619 |
+
"learning_rate": 4.300384916744261e-06,
|
620 |
+
"logits/chosen": -0.11280188709497452,
|
621 |
+
"logits/rejected": -0.12300585210323334,
|
622 |
+
"logps/chosen": -2.1714723110198975,
|
623 |
+
"logps/rejected": -54.74174118041992,
|
624 |
+
"loss": 0.1605,
|
625 |
+
"rewards/accuracies": 0.9500000476837158,
|
626 |
+
"rewards/chosen": 17.326162338256836,
|
627 |
+
"rewards/margins": 5.467062473297119,
|
628 |
+
"rewards/rejected": 11.859098434448242,
|
629 |
+
"step": 360
|
630 |
+
},
|
631 |
+
{
|
632 |
+
"epoch": 0.32,
|
633 |
+
"eval_logits/chosen": -0.10620756447315216,
|
634 |
+
"eval_logits/rejected": -0.11727114766836166,
|
635 |
+
"eval_logps/chosen": -1.4165427684783936,
|
636 |
+
"eval_logps/rejected": -50.9525146484375,
|
637 |
+
"eval_loss": 0.3194349706172943,
|
638 |
+
"eval_rewards/accuracies": 0.9210000038146973,
|
639 |
+
"eval_rewards/chosen": 17.36124038696289,
|
640 |
+
"eval_rewards/margins": 5.095769882202148,
|
641 |
+
"eval_rewards/rejected": 12.26546859741211,
|
642 |
+
"eval_runtime": 361.5072,
|
643 |
+
"eval_samples_per_second": 2.766,
|
644 |
+
"eval_steps_per_second": 0.346,
|
645 |
+
"step": 360
|
646 |
+
},
|
647 |
+
{
|
648 |
+
"epoch": 0.3288888888888889,
|
649 |
+
"grad_norm": 6.1126532554626465,
|
650 |
+
"learning_rate": 4.24568022938566e-06,
|
651 |
+
"logits/chosen": -0.10354311764240265,
|
652 |
+
"logits/rejected": -0.11526636779308319,
|
653 |
+
"logps/chosen": -1.2935255765914917,
|
654 |
+
"logps/rejected": -55.57566833496094,
|
655 |
+
"loss": 0.1711,
|
656 |
+
"rewards/accuracies": 0.9500000476837158,
|
657 |
+
"rewards/chosen": 17.439346313476562,
|
658 |
+
"rewards/margins": 5.700921058654785,
|
659 |
+
"rewards/rejected": 11.738424301147461,
|
660 |
+
"step": 370
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.3377777777777778,
|
664 |
+
"grad_norm": 34.15927505493164,
|
665 |
+
"learning_rate": 4.189293377245241e-06,
|
666 |
+
"logits/chosen": -0.1029932051897049,
|
667 |
+
"logits/rejected": -0.11382515728473663,
|
668 |
+
"logps/chosen": -2.5132687091827393,
|
669 |
+
"logps/rejected": -55.50346374511719,
|
670 |
+
"loss": 0.4359,
|
671 |
+
"rewards/accuracies": 0.8875000476837158,
|
672 |
+
"rewards/chosen": 16.731037139892578,
|
673 |
+
"rewards/margins": 4.368172645568848,
|
674 |
+
"rewards/rejected": 12.362865447998047,
|
675 |
+
"step": 380
|
676 |
+
},
|
677 |
+
{
|
678 |
+
"epoch": 0.3466666666666667,
|
679 |
+
"grad_norm": 2.8422904014587402,
|
680 |
+
"learning_rate": 4.131278695575952e-06,
|
681 |
+
"logits/chosen": -0.10793520510196686,
|
682 |
+
"logits/rejected": -0.12109285593032837,
|
683 |
+
"logps/chosen": -3.014652729034424,
|
684 |
+
"logps/rejected": -53.98411560058594,
|
685 |
+
"loss": 0.2161,
|
686 |
+
"rewards/accuracies": 0.949999988079071,
|
687 |
+
"rewards/chosen": 17.137393951416016,
|
688 |
+
"rewards/margins": 5.105995178222656,
|
689 |
+
"rewards/rejected": 12.03139877319336,
|
690 |
+
"step": 390
|
691 |
+
},
|
692 |
+
{
|
693 |
+
"epoch": 0.35555555555555557,
|
694 |
+
"grad_norm": 54.0329475402832,
|
695 |
+
"learning_rate": 4.071692088232743e-06,
|
696 |
+
"logits/chosen": -0.10393750667572021,
|
697 |
+
"logits/rejected": -0.11834606528282166,
|
698 |
+
"logps/chosen": -2.1508543491363525,
|
699 |
+
"logps/rejected": -45.60733413696289,
|
700 |
+
"loss": 0.2077,
|
701 |
+
"rewards/accuracies": 0.925000011920929,
|
702 |
+
"rewards/chosen": 17.586124420166016,
|
703 |
+
"rewards/margins": 5.077212333679199,
|
704 |
+
"rewards/rejected": 12.5089111328125,
|
705 |
+
"step": 400
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 0.36444444444444446,
|
709 |
+
"grad_norm": 81.61144256591797,
|
710 |
+
"learning_rate": 4.010590973802737e-06,
|
711 |
+
"logits/chosen": -0.09564584493637085,
|
712 |
+
"logits/rejected": -0.10617707669734955,
|
713 |
+
"logps/chosen": -3.4572842121124268,
|
714 |
+
"logps/rejected": -50.92162322998047,
|
715 |
+
"loss": 0.2478,
|
716 |
+
"rewards/accuracies": 0.8875000476837158,
|
717 |
+
"rewards/chosen": 17.010910034179688,
|
718 |
+
"rewards/margins": 4.556198596954346,
|
719 |
+
"rewards/rejected": 12.454713821411133,
|
720 |
+
"step": 410
|
721 |
+
},
|
722 |
+
{
|
723 |
+
"epoch": 0.37333333333333335,
|
724 |
+
"grad_norm": 0.30974289774894714,
|
725 |
+
"learning_rate": 3.948034230275781e-06,
|
726 |
+
"logits/chosen": -0.09134417027235031,
|
727 |
+
"logits/rejected": -0.1020016297698021,
|
728 |
+
"logps/chosen": -5.046698570251465,
|
729 |
+
"logps/rejected": -48.908958435058594,
|
730 |
+
"loss": 0.2894,
|
731 |
+
"rewards/accuracies": 0.8999999761581421,
|
732 |
+
"rewards/chosen": 17.007888793945312,
|
733 |
+
"rewards/margins": 4.53641414642334,
|
734 |
+
"rewards/rejected": 12.471475601196289,
|
735 |
+
"step": 420
|
736 |
+
},
|
737 |
+
{
|
738 |
+
"epoch": 0.37333333333333335,
|
739 |
+
"eval_logits/chosen": -0.09054450690746307,
|
740 |
+
"eval_logits/rejected": -0.10264354199171066,
|
741 |
+
"eval_logps/chosen": -1.913105845451355,
|
742 |
+
"eval_logps/rejected": -51.11127471923828,
|
743 |
+
"eval_loss": 0.16789735853672028,
|
744 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
745 |
+
"eval_rewards/chosen": 17.311582565307617,
|
746 |
+
"eval_rewards/margins": 5.061989784240723,
|
747 |
+
"eval_rewards/rejected": 12.249593734741211,
|
748 |
+
"eval_runtime": 361.5337,
|
749 |
+
"eval_samples_per_second": 2.766,
|
750 |
+
"eval_steps_per_second": 0.346,
|
751 |
+
"step": 420
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.38222222222222224,
|
755 |
+
"grad_norm": 12.824393272399902,
|
756 |
+
"learning_rate": 3.884082138308699e-06,
|
757 |
+
"logits/chosen": -0.08666776865720749,
|
758 |
+
"logits/rejected": -0.0997733399271965,
|
759 |
+
"logps/chosen": -1.7306327819824219,
|
760 |
+
"logps/rejected": -54.273292541503906,
|
761 |
+
"loss": 0.2298,
|
762 |
+
"rewards/accuracies": 0.9500000476837158,
|
763 |
+
"rewards/chosen": 17.167621612548828,
|
764 |
+
"rewards/margins": 5.065673351287842,
|
765 |
+
"rewards/rejected": 12.101947784423828,
|
766 |
+
"step": 430
|
767 |
+
},
|
768 |
+
{
|
769 |
+
"epoch": 0.39111111111111113,
|
770 |
+
"grad_norm": 0.30713599920272827,
|
771 |
+
"learning_rate": 3.818796323137896e-06,
|
772 |
+
"logits/chosen": -0.09174907952547073,
|
773 |
+
"logits/rejected": -0.10376611351966858,
|
774 |
+
"logps/chosen": -1.489154577255249,
|
775 |
+
"logps/rejected": -54.580726623535156,
|
776 |
+
"loss": 0.2513,
|
777 |
+
"rewards/accuracies": 0.9375,
|
778 |
+
"rewards/chosen": 17.22280502319336,
|
779 |
+
"rewards/margins": 5.175349235534668,
|
780 |
+
"rewards/rejected": 12.047454833984375,
|
781 |
+
"step": 440
|
782 |
+
},
|
783 |
+
{
|
784 |
+
"epoch": 0.4,
|
785 |
+
"grad_norm": 87.4791488647461,
|
786 |
+
"learning_rate": 3.7522396951963303e-06,
|
787 |
+
"logits/chosen": -0.09688778221607208,
|
788 |
+
"logits/rejected": -0.10897806286811829,
|
789 |
+
"logps/chosen": -3.157695770263672,
|
790 |
+
"logps/rejected": -50.96417236328125,
|
791 |
+
"loss": 0.1758,
|
792 |
+
"rewards/accuracies": 0.9500000476837158,
|
793 |
+
"rewards/chosen": 17.345651626586914,
|
794 |
+
"rewards/margins": 5.245656967163086,
|
795 |
+
"rewards/rejected": 12.099993705749512,
|
796 |
+
"step": 450
|
797 |
+
},
|
798 |
+
{
|
799 |
+
"epoch": 0.4088888888888889,
|
800 |
+
"grad_norm": 146.2008056640625,
|
801 |
+
"learning_rate": 3.684476389492026e-06,
|
802 |
+
"logits/chosen": -0.09378582239151001,
|
803 |
+
"logits/rejected": -0.10475654900074005,
|
804 |
+
"logps/chosen": -0.5611928701400757,
|
805 |
+
"logps/rejected": -56.518890380859375,
|
806 |
+
"loss": 0.1981,
|
807 |
+
"rewards/accuracies": 0.9500000476837158,
|
808 |
+
"rewards/chosen": 17.113712310791016,
|
809 |
+
"rewards/margins": 5.068872928619385,
|
810 |
+
"rewards/rejected": 12.044839859008789,
|
811 |
+
"step": 460
|
812 |
+
},
|
813 |
+
{
|
814 |
+
"epoch": 0.4177777777777778,
|
815 |
+
"grad_norm": 1.9137721061706543,
|
816 |
+
"learning_rate": 3.6155717038065783e-06,
|
817 |
+
"logits/chosen": -0.08695463836193085,
|
818 |
+
"logits/rejected": -0.09596743434667587,
|
819 |
+
"logps/chosen": -1.5298550128936768,
|
820 |
+
"logps/rejected": -50.27445983886719,
|
821 |
+
"loss": 0.2066,
|
822 |
+
"rewards/accuracies": 0.9375,
|
823 |
+
"rewards/chosen": 17.35186004638672,
|
824 |
+
"rewards/margins": 5.014693260192871,
|
825 |
+
"rewards/rejected": 12.337167739868164,
|
826 |
+
"step": 470
|
827 |
+
},
|
828 |
+
{
|
829 |
+
"epoch": 0.4266666666666667,
|
830 |
+
"grad_norm": 84.80391693115234,
|
831 |
+
"learning_rate": 3.545592035773192e-06,
|
832 |
+
"logits/chosen": -0.0746893435716629,
|
833 |
+
"logits/rejected": -0.08653923869132996,
|
834 |
+
"logps/chosen": -2.0052125453948975,
|
835 |
+
"logps/rejected": -57.502811431884766,
|
836 |
+
"loss": 0.1149,
|
837 |
+
"rewards/accuracies": 0.9500000476837158,
|
838 |
+
"rewards/chosen": 17.14373016357422,
|
839 |
+
"rewards/margins": 5.360415935516357,
|
840 |
+
"rewards/rejected": 11.783313751220703,
|
841 |
+
"step": 480
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 0.4266666666666667,
|
845 |
+
"eval_logits/chosen": -0.07700399309396744,
|
846 |
+
"eval_logits/rejected": -0.08828537166118622,
|
847 |
+
"eval_logps/chosen": -4.48896598815918,
|
848 |
+
"eval_logps/rejected": -53.76282501220703,
|
849 |
+
"eval_loss": 0.29511645436286926,
|
850 |
+
"eval_rewards/accuracies": 0.9230000376701355,
|
851 |
+
"eval_rewards/chosen": 17.053997039794922,
|
852 |
+
"eval_rewards/margins": 5.069558143615723,
|
853 |
+
"eval_rewards/rejected": 11.984437942504883,
|
854 |
+
"eval_runtime": 361.5035,
|
855 |
+
"eval_samples_per_second": 2.766,
|
856 |
+
"eval_steps_per_second": 0.346,
|
857 |
+
"step": 480
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 0.43555555555555553,
|
861 |
+
"grad_norm": 82.9616470336914,
|
862 |
+
"learning_rate": 3.4746048188948806e-06,
|
863 |
+
"logits/chosen": -0.06675051152706146,
|
864 |
+
"logits/rejected": -0.07860895991325378,
|
865 |
+
"logps/chosen": -4.162237167358398,
|
866 |
+
"logps/rejected": -54.77789306640625,
|
867 |
+
"loss": 0.2979,
|
868 |
+
"rewards/accuracies": 0.9125000238418579,
|
869 |
+
"rewards/chosen": 17.047603607177734,
|
870 |
+
"rewards/margins": 5.138361930847168,
|
871 |
+
"rewards/rejected": 11.909242630004883,
|
872 |
+
"step": 490
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.4444444444444444,
|
876 |
+
"grad_norm": 0.04293210059404373,
|
877 |
+
"learning_rate": 3.4026784575644887e-06,
|
878 |
+
"logits/chosen": -0.06424491107463837,
|
879 |
+
"logits/rejected": -0.07567107677459717,
|
880 |
+
"logps/chosen": -2.05729603767395,
|
881 |
+
"logps/rejected": -56.646087646484375,
|
882 |
+
"loss": 0.4378,
|
883 |
+
"rewards/accuracies": 0.8875000476837158,
|
884 |
+
"rewards/chosen": 16.947803497314453,
|
885 |
+
"rewards/margins": 4.919981956481934,
|
886 |
+
"rewards/rejected": 12.02782154083252,
|
887 |
+
"step": 500
|
888 |
+
},
|
889 |
+
{
|
890 |
+
"epoch": 0.4533333333333333,
|
891 |
+
"grad_norm": 0.07136644423007965,
|
892 |
+
"learning_rate": 3.329882261149148e-06,
|
893 |
+
"logits/chosen": -0.06423303484916687,
|
894 |
+
"logits/rejected": -0.07512776553630829,
|
895 |
+
"logps/chosen": -3.1519265174865723,
|
896 |
+
"logps/rejected": -53.53008270263672,
|
897 |
+
"loss": 0.2613,
|
898 |
+
"rewards/accuracies": 0.9375,
|
899 |
+
"rewards/chosen": 17.112144470214844,
|
900 |
+
"rewards/margins": 5.046430587768555,
|
901 |
+
"rewards/rejected": 12.065712928771973,
|
902 |
+
"step": 510
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 0.4622222222222222,
|
906 |
+
"grad_norm": 0.29279613494873047,
|
907 |
+
"learning_rate": 3.25628637720269e-06,
|
908 |
+
"logits/chosen": -0.060233693569898605,
|
909 |
+
"logits/rejected": -0.07076811790466309,
|
910 |
+
"logps/chosen": -1.2358124256134033,
|
911 |
+
"logps/rejected": -50.61806106567383,
|
912 |
+
"loss": 0.1517,
|
913 |
+
"rewards/accuracies": 0.9375,
|
914 |
+
"rewards/chosen": 17.457351684570312,
|
915 |
+
"rewards/margins": 5.2062835693359375,
|
916 |
+
"rewards/rejected": 12.251070022583008,
|
917 |
+
"step": 520
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 0.4711111111111111,
|
921 |
+
"grad_norm": 0.7946074604988098,
|
922 |
+
"learning_rate": 3.181961723870359e-06,
|
923 |
+
"logits/chosen": -0.054482050240039825,
|
924 |
+
"logits/rejected": -0.0657092034816742,
|
925 |
+
"logps/chosen": -0.7682158946990967,
|
926 |
+
"logps/rejected": -58.81409454345703,
|
927 |
+
"loss": 0.2304,
|
928 |
+
"rewards/accuracies": 0.9375,
|
929 |
+
"rewards/chosen": 17.075973510742188,
|
930 |
+
"rewards/margins": 5.222441673278809,
|
931 |
+
"rewards/rejected": 11.853530883789062,
|
932 |
+
"step": 530
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 0.48,
|
936 |
+
"grad_norm": 3.1405210494995117,
|
937 |
+
"learning_rate": 3.1069799215509847e-06,
|
938 |
+
"logits/chosen": -0.05050881579518318,
|
939 |
+
"logits/rejected": -0.061149902641773224,
|
940 |
+
"logps/chosen": -0.954046368598938,
|
941 |
+
"logps/rejected": -52.70227813720703,
|
942 |
+
"loss": 0.0384,
|
943 |
+
"rewards/accuracies": 0.987500011920929,
|
944 |
+
"rewards/chosen": 17.378849029541016,
|
945 |
+
"rewards/margins": 5.253483295440674,
|
946 |
+
"rewards/rejected": 12.125364303588867,
|
947 |
+
"step": 540
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.48,
|
951 |
+
"eval_logits/chosen": -0.05120665580034256,
|
952 |
+
"eval_logits/rejected": -0.06123337894678116,
|
953 |
+
"eval_logps/chosen": -2.9872913360595703,
|
954 |
+
"eval_logps/rejected": -52.27314758300781,
|
955 |
+
"eval_loss": 0.17387841641902924,
|
956 |
+
"eval_rewards/accuracies": 0.9490000605583191,
|
957 |
+
"eval_rewards/chosen": 17.204164505004883,
|
958 |
+
"eval_rewards/margins": 5.070757865905762,
|
959 |
+
"eval_rewards/rejected": 12.133406639099121,
|
960 |
+
"eval_runtime": 361.5449,
|
961 |
+
"eval_samples_per_second": 2.766,
|
962 |
+
"eval_steps_per_second": 0.346,
|
963 |
+
"step": 540
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 0.4888888888888889,
|
967 |
+
"grad_norm": 0.07270358502864838,
|
968 |
+
"learning_rate": 3.0314132238824416e-06,
|
969 |
+
"logits/chosen": -0.05125313252210617,
|
970 |
+
"logits/rejected": -0.06174170970916748,
|
971 |
+
"logps/chosen": -3.2163877487182617,
|
972 |
+
"logps/rejected": -47.79279327392578,
|
973 |
+
"loss": 0.2087,
|
974 |
+
"rewards/accuracies": 0.9125000238418579,
|
975 |
+
"rewards/chosen": 17.408517837524414,
|
976 |
+
"rewards/margins": 5.061524391174316,
|
977 |
+
"rewards/rejected": 12.346992492675781,
|
978 |
+
"step": 550
|
979 |
+
},
|
980 |
+
{
|
981 |
+
"epoch": 0.49777777777777776,
|
982 |
+
"grad_norm": 0.10005924850702286,
|
983 |
+
"learning_rate": 2.955334448116915e-06,
|
984 |
+
"logits/chosen": -0.041773442178964615,
|
985 |
+
"logits/rejected": -0.05364570394158363,
|
986 |
+
"logps/chosen": -0.363404780626297,
|
987 |
+
"logps/rejected": -56.32415008544922,
|
988 |
+
"loss": 0.0969,
|
989 |
+
"rewards/accuracies": 0.9624999761581421,
|
990 |
+
"rewards/chosen": 17.3824462890625,
|
991 |
+
"rewards/margins": 5.5404510498046875,
|
992 |
+
"rewards/rejected": 11.841995239257812,
|
993 |
+
"step": 560
|
994 |
+
},
|
995 |
+
{
|
996 |
+
"epoch": 0.5066666666666667,
|
997 |
+
"grad_norm": 7.818356990814209,
|
998 |
+
"learning_rate": 2.8788169049530533e-06,
|
999 |
+
"logits/chosen": -0.04309462010860443,
|
1000 |
+
"logits/rejected": -0.05494442582130432,
|
1001 |
+
"logps/chosen": -2.2242724895477295,
|
1002 |
+
"logps/rejected": -56.444740295410156,
|
1003 |
+
"loss": 0.1447,
|
1004 |
+
"rewards/accuracies": 0.949999988079071,
|
1005 |
+
"rewards/chosen": 17.275390625,
|
1006 |
+
"rewards/margins": 5.55007791519165,
|
1007 |
+
"rewards/rejected": 11.725313186645508,
|
1008 |
+
"step": 570
|
1009 |
+
},
|
1010 |
+
{
|
1011 |
+
"epoch": 0.5155555555555555,
|
1012 |
+
"grad_norm": 0.03519747406244278,
|
1013 |
+
"learning_rate": 2.8019343278926397e-06,
|
1014 |
+
"logits/chosen": -0.03506368771195412,
|
1015 |
+
"logits/rejected": -0.046854715794324875,
|
1016 |
+
"logps/chosen": -0.5200096964836121,
|
1017 |
+
"logps/rejected": -59.05330276489258,
|
1018 |
+
"loss": 0.0998,
|
1019 |
+
"rewards/accuracies": 0.949999988079071,
|
1020 |
+
"rewards/chosen": 17.320554733276367,
|
1021 |
+
"rewards/margins": 5.727260589599609,
|
1022 |
+
"rewards/rejected": 11.593294143676758,
|
1023 |
+
"step": 580
|
1024 |
+
},
|
1025 |
+
{
|
1026 |
+
"epoch": 0.5244444444444445,
|
1027 |
+
"grad_norm": 0.04519123584032059,
|
1028 |
+
"learning_rate": 2.7247608021898265e-06,
|
1029 |
+
"logits/chosen": -0.03204537555575371,
|
1030 |
+
"logits/rejected": -0.04383891448378563,
|
1031 |
+
"logps/chosen": -1.1271060705184937,
|
1032 |
+
"logps/rejected": -59.308895111083984,
|
1033 |
+
"loss": 0.1162,
|
1034 |
+
"rewards/accuracies": 0.9750000238418579,
|
1035 |
+
"rewards/chosen": 17.256351470947266,
|
1036 |
+
"rewards/margins": 5.6881890296936035,
|
1037 |
+
"rewards/rejected": 11.568161010742188,
|
1038 |
+
"step": 590
|
1039 |
+
},
|
1040 |
+
{
|
1041 |
+
"epoch": 0.5333333333333333,
|
1042 |
+
"grad_norm": 0.028689857572317123,
|
1043 |
+
"learning_rate": 2.647370693461432e-06,
|
1044 |
+
"logits/chosen": -0.02834726870059967,
|
1045 |
+
"logits/rejected": -0.03827046602964401,
|
1046 |
+
"logps/chosen": -5.673943519592285,
|
1047 |
+
"logps/rejected": -55.72624588012695,
|
1048 |
+
"loss": 0.4008,
|
1049 |
+
"rewards/accuracies": 0.8875000476837158,
|
1050 |
+
"rewards/chosen": 16.761056900024414,
|
1051 |
+
"rewards/margins": 4.804043769836426,
|
1052 |
+
"rewards/rejected": 11.957012176513672,
|
1053 |
+
"step": 600
|
1054 |
+
},
|
1055 |
+
{
|
1056 |
+
"epoch": 0.5333333333333333,
|
1057 |
+
"eval_logits/chosen": -0.035840023308992386,
|
1058 |
+
"eval_logits/rejected": -0.04687971621751785,
|
1059 |
+
"eval_logps/chosen": -2.1760435104370117,
|
1060 |
+
"eval_logps/rejected": -56.62664031982422,
|
1061 |
+
"eval_loss": 0.17061151564121246,
|
1062 |
+
"eval_rewards/accuracies": 0.9470000267028809,
|
1063 |
+
"eval_rewards/chosen": 17.285289764404297,
|
1064 |
+
"eval_rewards/margins": 5.587231636047363,
|
1065 |
+
"eval_rewards/rejected": 11.698057174682617,
|
1066 |
+
"eval_runtime": 361.5056,
|
1067 |
+
"eval_samples_per_second": 2.766,
|
1068 |
+
"eval_steps_per_second": 0.346,
|
1069 |
+
"step": 600
|
1070 |
+
},
|
1071 |
+
{
|
1072 |
+
"epoch": 0.5422222222222223,
|
1073 |
+
"grad_norm": 1.1994622945785522,
|
1074 |
+
"learning_rate": 2.569838576027068e-06,
|
1075 |
+
"logits/chosen": -0.03231767192482948,
|
1076 |
+
"logits/rejected": -0.04397805407643318,
|
1077 |
+
"logps/chosen": -1.5904741287231445,
|
1078 |
+
"logps/rejected": -62.10520553588867,
|
1079 |
+
"loss": 0.1265,
|
1080 |
+
"rewards/accuracies": 0.9624999761581421,
|
1081 |
+
"rewards/chosen": 17.11206817626953,
|
1082 |
+
"rewards/margins": 5.734784126281738,
|
1083 |
+
"rewards/rejected": 11.37728500366211,
|
1084 |
+
"step": 610
|
1085 |
+
},
|
1086 |
+
{
|
1087 |
+
"epoch": 0.5511111111111111,
|
1088 |
+
"grad_norm": 0.14163845777511597,
|
1089 |
+
"learning_rate": 2.4922391610481544e-06,
|
1090 |
+
"logits/chosen": -0.03293662518262863,
|
1091 |
+
"logits/rejected": -0.04335154965519905,
|
1092 |
+
"logps/chosen": -1.3568997383117676,
|
1093 |
+
"logps/rejected": -58.347434997558594,
|
1094 |
+
"loss": 0.097,
|
1095 |
+
"rewards/accuracies": 0.9624999761581421,
|
1096 |
+
"rewards/chosen": 17.345399856567383,
|
1097 |
+
"rewards/margins": 5.823373794555664,
|
1098 |
+
"rewards/rejected": 11.522026062011719,
|
1099 |
+
"step": 620
|
1100 |
+
},
|
1101 |
+
{
|
1102 |
+
"epoch": 0.56,
|
1103 |
+
"grad_norm": 1.125027060508728,
|
1104 |
+
"learning_rate": 2.4146472245350804e-06,
|
1105 |
+
"logits/chosen": -0.02864963933825493,
|
1106 |
+
"logits/rejected": -0.03835710883140564,
|
1107 |
+
"logps/chosen": -2.494175434112549,
|
1108 |
+
"logps/rejected": -55.33067321777344,
|
1109 |
+
"loss": 0.3115,
|
1110 |
+
"rewards/accuracies": 0.925000011920929,
|
1111 |
+
"rewards/chosen": 17.262205123901367,
|
1112 |
+
"rewards/margins": 5.4287261962890625,
|
1113 |
+
"rewards/rejected": 11.833479881286621,
|
1114 |
+
"step": 630
|
1115 |
+
},
|
1116 |
+
{
|
1117 |
+
"epoch": 0.5688888888888889,
|
1118 |
+
"grad_norm": 0.03514016419649124,
|
1119 |
+
"learning_rate": 2.337137535291868e-06,
|
1120 |
+
"logits/chosen": -0.02757749892771244,
|
1121 |
+
"logits/rejected": -0.0376611053943634,
|
1122 |
+
"logps/chosen": -2.3163387775421143,
|
1123 |
+
"logps/rejected": -53.54579162597656,
|
1124 |
+
"loss": 0.1268,
|
1125 |
+
"rewards/accuracies": 0.9500000476837158,
|
1126 |
+
"rewards/chosen": 17.542863845825195,
|
1127 |
+
"rewards/margins": 5.799897193908691,
|
1128 |
+
"rewards/rejected": 11.742965698242188,
|
1129 |
+
"step": 640
|
1130 |
+
},
|
1131 |
+
{
|
1132 |
+
"epoch": 0.5777777777777777,
|
1133 |
+
"grad_norm": 34.20791244506836,
|
1134 |
+
"learning_rate": 2.259784782867782e-06,
|
1135 |
+
"logits/chosen": -0.02032250165939331,
|
1136 |
+
"logits/rejected": -0.031542714685201645,
|
1137 |
+
"logps/chosen": -1.7910137176513672,
|
1138 |
+
"logps/rejected": -62.980018615722656,
|
1139 |
+
"loss": 0.1831,
|
1140 |
+
"rewards/accuracies": 0.9500000476837158,
|
1141 |
+
"rewards/chosen": 16.982879638671875,
|
1142 |
+
"rewards/margins": 5.589078903198242,
|
1143 |
+
"rewards/rejected": 11.393800735473633,
|
1144 |
+
"step": 650
|
1145 |
+
},
|
1146 |
+
{
|
1147 |
+
"epoch": 0.5866666666666667,
|
1148 |
+
"grad_norm": 32.356956481933594,
|
1149 |
+
"learning_rate": 2.182663505585314e-06,
|
1150 |
+
"logits/chosen": -0.01711965538561344,
|
1151 |
+
"logits/rejected": -0.028038471937179565,
|
1152 |
+
"logps/chosen": -2.662904739379883,
|
1153 |
+
"logps/rejected": -63.326297760009766,
|
1154 |
+
"loss": 0.1678,
|
1155 |
+
"rewards/accuracies": 0.9624999761581421,
|
1156 |
+
"rewards/chosen": 16.997777938842773,
|
1157 |
+
"rewards/margins": 5.741157531738281,
|
1158 |
+
"rewards/rejected": 11.256620407104492,
|
1159 |
+
"step": 660
|
1160 |
+
},
|
1161 |
+
{
|
1162 |
+
"epoch": 0.5866666666666667,
|
1163 |
+
"eval_logits/chosen": -0.016018809750676155,
|
1164 |
+
"eval_logits/rejected": -0.027020033448934555,
|
1165 |
+
"eval_logps/chosen": -3.00822114944458,
|
1166 |
+
"eval_logps/rejected": -57.951629638671875,
|
1167 |
+
"eval_loss": 0.2050127536058426,
|
1168 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1169 |
+
"eval_rewards/chosen": 17.202072143554688,
|
1170 |
+
"eval_rewards/margins": 5.636512756347656,
|
1171 |
+
"eval_rewards/rejected": 11.565557479858398,
|
1172 |
+
"eval_runtime": 361.5073,
|
1173 |
+
"eval_samples_per_second": 2.766,
|
1174 |
+
"eval_steps_per_second": 0.346,
|
1175 |
+
"step": 660
|
1176 |
+
},
|
1177 |
+
{
|
1178 |
+
"epoch": 0.5955555555555555,
|
1179 |
+
"grad_norm": 86.53874969482422,
|
1180 |
+
"learning_rate": 2.1058480187138863e-06,
|
1181 |
+
"logits/chosen": -0.012352555990219116,
|
1182 |
+
"logits/rejected": -0.023966707289218903,
|
1183 |
+
"logps/chosen": -2.129209041595459,
|
1184 |
+
"logps/rejected": -62.1393928527832,
|
1185 |
+
"loss": 0.1025,
|
1186 |
+
"rewards/accuracies": 0.9624999761581421,
|
1187 |
+
"rewards/chosen": 17.069538116455078,
|
1188 |
+
"rewards/margins": 5.678930759429932,
|
1189 |
+
"rewards/rejected": 11.390605926513672,
|
1190 |
+
"step": 670
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 0.6044444444444445,
|
1194 |
+
"grad_norm": 2.892672300338745,
|
1195 |
+
"learning_rate": 2.0294123428584985e-06,
|
1196 |
+
"logits/chosen": -0.011562807485461235,
|
1197 |
+
"logits/rejected": -0.020859256386756897,
|
1198 |
+
"logps/chosen": -3.2486608028411865,
|
1199 |
+
"logps/rejected": -59.56721496582031,
|
1200 |
+
"loss": 0.1961,
|
1201 |
+
"rewards/accuracies": 0.9500000476837158,
|
1202 |
+
"rewards/chosen": 17.051376342773438,
|
1203 |
+
"rewards/margins": 5.515361785888672,
|
1204 |
+
"rewards/rejected": 11.536016464233398,
|
1205 |
+
"step": 680
|
1206 |
+
},
|
1207 |
+
{
|
1208 |
+
"epoch": 0.6133333333333333,
|
1209 |
+
"grad_norm": 30.26588249206543,
|
1210 |
+
"learning_rate": 1.953430132632311e-06,
|
1211 |
+
"logits/chosen": -0.011488726362586021,
|
1212 |
+
"logits/rejected": -0.021527227014303207,
|
1213 |
+
"logps/chosen": -1.794357180595398,
|
1214 |
+
"logps/rejected": -60.649505615234375,
|
1215 |
+
"loss": 0.2821,
|
1216 |
+
"rewards/accuracies": 0.9375,
|
1217 |
+
"rewards/chosen": 17.114843368530273,
|
1218 |
+
"rewards/margins": 5.612217903137207,
|
1219 |
+
"rewards/rejected": 11.50262451171875,
|
1220 |
+
"step": 690
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 0.6222222222222222,
|
1224 |
+
"grad_norm": 0.030314341187477112,
|
1225 |
+
"learning_rate": 1.8779746056819104e-06,
|
1226 |
+
"logits/chosen": -0.014436552301049232,
|
1227 |
+
"logits/rejected": -0.026044374331831932,
|
1228 |
+
"logps/chosen": -3.1617178916931152,
|
1229 |
+
"logps/rejected": -60.94548797607422,
|
1230 |
+
"loss": 0.3141,
|
1231 |
+
"rewards/accuracies": 0.9375,
|
1232 |
+
"rewards/chosen": 17.08698272705078,
|
1233 |
+
"rewards/margins": 5.719264984130859,
|
1234 |
+
"rewards/rejected": 11.367716789245605,
|
1235 |
+
"step": 700
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 0.6311111111111111,
|
1239 |
+
"grad_norm": 0.06711317598819733,
|
1240 |
+
"learning_rate": 1.8031184721336364e-06,
|
1241 |
+
"logits/chosen": -0.017312290146946907,
|
1242 |
+
"logits/rejected": -0.026554957032203674,
|
1243 |
+
"logps/chosen": -5.120705604553223,
|
1244 |
+
"logps/rejected": -54.33483123779297,
|
1245 |
+
"loss": 0.3827,
|
1246 |
+
"rewards/accuracies": 0.887499988079071,
|
1247 |
+
"rewards/chosen": 17.1247501373291,
|
1248 |
+
"rewards/margins": 5.327882766723633,
|
1249 |
+
"rewards/rejected": 11.796867370605469,
|
1250 |
+
"step": 710
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 0.64,
|
1254 |
+
"grad_norm": 60.36637496948242,
|
1255 |
+
"learning_rate": 1.7289338645289711e-06,
|
1256 |
+
"logits/chosen": -0.01987219974398613,
|
1257 |
+
"logits/rejected": -0.02766304835677147,
|
1258 |
+
"logps/chosen": -2.1757986545562744,
|
1259 |
+
"logps/rejected": -54.02531433105469,
|
1260 |
+
"loss": 0.2272,
|
1261 |
+
"rewards/accuracies": 0.949999988079071,
|
1262 |
+
"rewards/chosen": 17.239776611328125,
|
1263 |
+
"rewards/margins": 5.250313758850098,
|
1264 |
+
"rewards/rejected": 11.989462852478027,
|
1265 |
+
"step": 720
|
1266 |
+
},
|
1267 |
+
{
|
1268 |
+
"epoch": 0.64,
|
1269 |
+
"eval_logits/chosen": -0.022905193269252777,
|
1270 |
+
"eval_logits/rejected": -0.0321992002427578,
|
1271 |
+
"eval_logps/chosen": -1.100506067276001,
|
1272 |
+
"eval_logps/rejected": -55.91169738769531,
|
1273 |
+
"eval_loss": 0.14018221199512482,
|
1274 |
+
"eval_rewards/accuracies": 0.9520000219345093,
|
1275 |
+
"eval_rewards/chosen": 17.392841339111328,
|
1276 |
+
"eval_rewards/margins": 5.623291492462158,
|
1277 |
+
"eval_rewards/rejected": 11.769551277160645,
|
1278 |
+
"eval_runtime": 361.416,
|
1279 |
+
"eval_samples_per_second": 2.767,
|
1280 |
+
"eval_steps_per_second": 0.346,
|
1281 |
+
"step": 720
|
1282 |
+
},
|
1283 |
+
{
|
1284 |
+
"epoch": 0.6488888888888888,
|
1285 |
+
"grad_norm": 0.029294608160853386,
|
1286 |
+
"learning_rate": 1.6554922683164875e-06,
|
1287 |
+
"logits/chosen": -0.016416028141975403,
|
1288 |
+
"logits/rejected": -0.025527067482471466,
|
1289 |
+
"logps/chosen": -0.906692385673523,
|
1290 |
+
"logps/rejected": -59.142173767089844,
|
1291 |
+
"loss": 0.1171,
|
1292 |
+
"rewards/accuracies": 0.9624999761581421,
|
1293 |
+
"rewards/chosen": 17.282169342041016,
|
1294 |
+
"rewards/margins": 5.6978583335876465,
|
1295 |
+
"rewards/rejected": 11.584310531616211,
|
1296 |
+
"step": 730
|
1297 |
+
},
|
1298 |
+
{
|
1299 |
+
"epoch": 0.6577777777777778,
|
1300 |
+
"grad_norm": 33.89070510864258,
|
1301 |
+
"learning_rate": 1.5828644529673592e-06,
|
1302 |
+
"logits/chosen": -0.016044551506638527,
|
1303 |
+
"logits/rejected": -0.025811903178691864,
|
1304 |
+
"logps/chosen": -1.1167538166046143,
|
1305 |
+
"logps/rejected": -59.84492492675781,
|
1306 |
+
"loss": 0.2001,
|
1307 |
+
"rewards/accuracies": 0.9624999761581421,
|
1308 |
+
"rewards/chosen": 17.266983032226562,
|
1309 |
+
"rewards/margins": 5.749438285827637,
|
1310 |
+
"rewards/rejected": 11.51754379272461,
|
1311 |
+
"step": 740
|
1312 |
+
},
|
1313 |
+
{
|
1314 |
+
"epoch": 0.6666666666666666,
|
1315 |
+
"grad_norm": 0.53726726770401,
|
1316 |
+
"learning_rate": 1.5111204037807844e-06,
|
1317 |
+
"logits/chosen": -0.012529855594038963,
|
1318 |
+
"logits/rejected": -0.023765765130519867,
|
1319 |
+
"logps/chosen": -0.3461765646934509,
|
1320 |
+
"logps/rejected": -55.841102600097656,
|
1321 |
+
"loss": 0.1692,
|
1322 |
+
"rewards/accuracies": 0.9375,
|
1323 |
+
"rewards/chosen": 17.61865997314453,
|
1324 |
+
"rewards/margins": 6.019055366516113,
|
1325 |
+
"rewards/rejected": 11.599604606628418,
|
1326 |
+
"step": 750
|
1327 |
+
},
|
1328 |
+
{
|
1329 |
+
"epoch": 0.6755555555555556,
|
1330 |
+
"grad_norm": 0.7173987627029419,
|
1331 |
+
"learning_rate": 1.4403292544450625e-06,
|
1332 |
+
"logits/chosen": -0.013380522839725018,
|
1333 |
+
"logits/rejected": -0.022107835859060287,
|
1334 |
+
"logps/chosen": -3.2759666442871094,
|
1335 |
+
"logps/rejected": -53.836822509765625,
|
1336 |
+
"loss": 0.2418,
|
1337 |
+
"rewards/accuracies": 0.949999988079071,
|
1338 |
+
"rewards/chosen": 17.383333206176758,
|
1339 |
+
"rewards/margins": 5.590358734130859,
|
1340 |
+
"rewards/rejected": 11.792974472045898,
|
1341 |
+
"step": 760
|
1342 |
+
},
|
1343 |
+
{
|
1344 |
+
"epoch": 0.6844444444444444,
|
1345 |
+
"grad_norm": 2.744900941848755,
|
1346 |
+
"learning_rate": 1.3705592204192853e-06,
|
1347 |
+
"logits/chosen": -0.011318420059978962,
|
1348 |
+
"logits/rejected": -0.021840626373887062,
|
1349 |
+
"logps/chosen": -2.592001438140869,
|
1350 |
+
"logps/rejected": -57.06926345825195,
|
1351 |
+
"loss": 0.1749,
|
1352 |
+
"rewards/accuracies": 0.9500000476837158,
|
1353 |
+
"rewards/chosen": 17.286128997802734,
|
1354 |
+
"rewards/margins": 5.662715911865234,
|
1355 |
+
"rewards/rejected": 11.623414993286133,
|
1356 |
+
"step": 770
|
1357 |
+
},
|
1358 |
+
{
|
1359 |
+
"epoch": 0.6933333333333334,
|
1360 |
+
"grad_norm": 70.76551818847656,
|
1361 |
+
"learning_rate": 1.301877533199859e-06,
|
1362 |
+
"logits/chosen": -0.013944407925009727,
|
1363 |
+
"logits/rejected": -0.025667501613497734,
|
1364 |
+
"logps/chosen": -2.1070234775543213,
|
1365 |
+
"logps/rejected": -57.72039794921875,
|
1366 |
+
"loss": 0.1915,
|
1367 |
+
"rewards/accuracies": 0.925000011920929,
|
1368 |
+
"rewards/chosen": 17.28545379638672,
|
1369 |
+
"rewards/margins": 5.708344459533691,
|
1370 |
+
"rewards/rejected": 11.577108383178711,
|
1371 |
+
"step": 780
|
1372 |
+
},
|
1373 |
+
{
|
1374 |
+
"epoch": 0.6933333333333334,
|
1375 |
+
"eval_logits/chosen": -0.01664295792579651,
|
1376 |
+
"eval_logits/rejected": -0.026626665145158768,
|
1377 |
+
"eval_logps/chosen": -1.082255244255066,
|
1378 |
+
"eval_logps/rejected": -55.95073699951172,
|
1379 |
+
"eval_loss": 0.24412688612937927,
|
1380 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
1381 |
+
"eval_rewards/chosen": 17.394668579101562,
|
1382 |
+
"eval_rewards/margins": 5.629020690917969,
|
1383 |
+
"eval_rewards/rejected": 11.765647888183594,
|
1384 |
+
"eval_runtime": 361.4602,
|
1385 |
+
"eval_samples_per_second": 2.767,
|
1386 |
+
"eval_steps_per_second": 0.346,
|
1387 |
+
"step": 780
|
1388 |
+
},
|
1389 |
+
{
|
1390 |
+
"epoch": 0.7022222222222222,
|
1391 |
+
"grad_norm": 0.41359376907348633,
|
1392 |
+
"learning_rate": 1.2343503755351729e-06,
|
1393 |
+
"logits/chosen": -0.012772129848599434,
|
1394 |
+
"logits/rejected": -0.023186586797237396,
|
1395 |
+
"logps/chosen": -0.8030359148979187,
|
1396 |
+
"logps/rejected": -57.84947967529297,
|
1397 |
+
"loss": 0.2143,
|
1398 |
+
"rewards/accuracies": 0.9375,
|
1399 |
+
"rewards/chosen": 17.32242202758789,
|
1400 |
+
"rewards/margins": 5.645486831665039,
|
1401 |
+
"rewards/rejected": 11.676933288574219,
|
1402 |
+
"step": 790
|
1403 |
+
},
|
1404 |
+
{
|
1405 |
+
"epoch": 0.7111111111111111,
|
1406 |
+
"grad_norm": 0.41748157143592834,
|
1407 |
+
"learning_rate": 1.168042817650881e-06,
|
1408 |
+
"logits/chosen": -0.012502101249992847,
|
1409 |
+
"logits/rejected": -0.023272844031453133,
|
1410 |
+
"logps/chosen": -1.0377256870269775,
|
1411 |
+
"logps/rejected": -57.05615997314453,
|
1412 |
+
"loss": 0.0792,
|
1413 |
+
"rewards/accuracies": 0.9750000238418579,
|
1414 |
+
"rewards/chosen": 17.60501480102539,
|
1415 |
+
"rewards/margins": 6.159370422363281,
|
1416 |
+
"rewards/rejected": 11.445646286010742,
|
1417 |
+
"step": 800
|
1418 |
+
},
|
1419 |
+
{
|
1420 |
+
"epoch": 0.72,
|
1421 |
+
"grad_norm": 0.039210401475429535,
|
1422 |
+
"learning_rate": 1.1030187545472012e-06,
|
1423 |
+
"logits/chosen": -0.008177272044122219,
|
1424 |
+
"logits/rejected": -0.01832464337348938,
|
1425 |
+
"logps/chosen": -3.8701748847961426,
|
1426 |
+
"logps/rejected": -57.661476135253906,
|
1427 |
+
"loss": 0.3177,
|
1428 |
+
"rewards/accuracies": 0.925000011920929,
|
1429 |
+
"rewards/chosen": 17.14336395263672,
|
1430 |
+
"rewards/margins": 5.583393573760986,
|
1431 |
+
"rewards/rejected": 11.55997085571289,
|
1432 |
+
"step": 810
|
1433 |
+
},
|
1434 |
+
{
|
1435 |
+
"epoch": 0.7288888888888889,
|
1436 |
+
"grad_norm": 1.6096951961517334,
|
1437 |
+
"learning_rate": 1.0393408444287048e-06,
|
1438 |
+
"logits/chosen": -0.006832236424088478,
|
1439 |
+
"logits/rejected": -0.01682097464799881,
|
1440 |
+
"logps/chosen": -2.1741790771484375,
|
1441 |
+
"logps/rejected": -57.273414611816406,
|
1442 |
+
"loss": 0.2128,
|
1443 |
+
"rewards/accuracies": 0.949999988079071,
|
1444 |
+
"rewards/chosen": 17.426712036132812,
|
1445 |
+
"rewards/margins": 5.911205291748047,
|
1446 |
+
"rewards/rejected": 11.515506744384766,
|
1447 |
+
"step": 820
|
1448 |
+
},
|
1449 |
+
{
|
1450 |
+
"epoch": 0.7377777777777778,
|
1451 |
+
"grad_norm": 3.7820959091186523,
|
1452 |
+
"learning_rate": 9.770704483258782e-07,
|
1453 |
+
"logits/chosen": -0.009998206980526447,
|
1454 |
+
"logits/rejected": -0.0204261876642704,
|
1455 |
+
"logps/chosen": -1.9802953004837036,
|
1456 |
+
"logps/rejected": -56.901512145996094,
|
1457 |
+
"loss": 0.1296,
|
1458 |
+
"rewards/accuracies": 0.9750000238418579,
|
1459 |
+
"rewards/chosen": 17.56071662902832,
|
1460 |
+
"rewards/margins": 6.169583320617676,
|
1461 |
+
"rewards/rejected": 11.391134262084961,
|
1462 |
+
"step": 830
|
1463 |
+
},
|
1464 |
+
{
|
1465 |
+
"epoch": 0.7466666666666667,
|
1466 |
+
"grad_norm": 1.126626968383789,
|
1467 |
+
"learning_rate": 9.162675709666865e-07,
|
1468 |
+
"logits/chosen": -0.00826224498450756,
|
1469 |
+
"logits/rejected": -0.018977787345647812,
|
1470 |
+
"logps/chosen": -1.4256607294082642,
|
1471 |
+
"logps/rejected": -61.65986633300781,
|
1472 |
+
"loss": 0.0635,
|
1473 |
+
"rewards/accuracies": 0.987500011920929,
|
1474 |
+
"rewards/chosen": 17.324975967407227,
|
1475 |
+
"rewards/margins": 6.09440803527832,
|
1476 |
+
"rewards/rejected": 11.230567932128906,
|
1477 |
+
"step": 840
|
1478 |
+
},
|
1479 |
+
{
|
1480 |
+
"epoch": 0.7466666666666667,
|
1481 |
+
"eval_logits/chosen": -0.01111944392323494,
|
1482 |
+
"eval_logits/rejected": -0.021697774529457092,
|
1483 |
+
"eval_logps/chosen": -1.2168633937835693,
|
1484 |
+
"eval_logps/rejected": -58.2642822265625,
|
1485 |
+
"eval_loss": 0.1689341962337494,
|
1486 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1487 |
+
"eval_rewards/chosen": 17.381206512451172,
|
1488 |
+
"eval_rewards/margins": 5.846914291381836,
|
1489 |
+
"eval_rewards/rejected": 11.534292221069336,
|
1490 |
+
"eval_runtime": 361.6192,
|
1491 |
+
"eval_samples_per_second": 2.765,
|
1492 |
+
"eval_steps_per_second": 0.346,
|
1493 |
+
"step": 840
|
1494 |
+
},
|
1495 |
+
{
|
1496 |
+
"epoch": 0.7555555555555555,
|
1497 |
+
"grad_norm": 5.250723838806152,
|
1498 |
+
"learning_rate": 8.569908029550686e-07,
|
1499 |
+
"logits/chosen": -0.006854387000203133,
|
1500 |
+
"logits/rejected": -0.018336206674575806,
|
1501 |
+
"logps/chosen": -0.6238930821418762,
|
1502 |
+
"logps/rejected": -60.925689697265625,
|
1503 |
+
"loss": 0.1157,
|
1504 |
+
"rewards/accuracies": 0.9750000238418579,
|
1505 |
+
"rewards/chosen": 17.443281173706055,
|
1506 |
+
"rewards/margins": 6.164813995361328,
|
1507 |
+
"rewards/rejected": 11.278467178344727,
|
1508 |
+
"step": 850
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 0.7644444444444445,
|
1512 |
+
"grad_norm": 3.1401162147521973,
|
1513 |
+
"learning_rate": 7.992972643121227e-07,
|
1514 |
+
"logits/chosen": -0.0037835021503269672,
|
1515 |
+
"logits/rejected": -0.013135241344571114,
|
1516 |
+
"logps/chosen": -0.8492221832275391,
|
1517 |
+
"logps/rejected": -55.516075134277344,
|
1518 |
+
"loss": 0.2252,
|
1519 |
+
"rewards/accuracies": 0.9375,
|
1520 |
+
"rewards/chosen": 17.488588333129883,
|
1521 |
+
"rewards/margins": 5.736725807189941,
|
1522 |
+
"rewards/rejected": 11.751862525939941,
|
1523 |
+
"step": 860
|
1524 |
+
},
|
1525 |
+
{
|
1526 |
+
"epoch": 0.7733333333333333,
|
1527 |
+
"grad_norm": 55.528812408447266,
|
1528 |
+
"learning_rate": 7.432425494343509e-07,
|
1529 |
+
"logits/chosen": -0.0033687639515846968,
|
1530 |
+
"logits/rejected": -0.013152632862329483,
|
1531 |
+
"logps/chosen": -1.3188884258270264,
|
1532 |
+
"logps/rejected": -57.9510498046875,
|
1533 |
+
"loss": 0.1398,
|
1534 |
+
"rewards/accuracies": 0.9624999761581421,
|
1535 |
+
"rewards/chosen": 17.412578582763672,
|
1536 |
+
"rewards/margins": 5.868515968322754,
|
1537 |
+
"rewards/rejected": 11.544061660766602,
|
1538 |
+
"step": 870
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 0.7822222222222223,
|
1542 |
+
"grad_norm": 0.039824869483709335,
|
1543 |
+
"learning_rate": 6.888806735220396e-07,
|
1544 |
+
"logits/chosen": -0.0010406378423795104,
|
1545 |
+
"logits/rejected": -0.012095071375370026,
|
1546 |
+
"logps/chosen": -2.0619027614593506,
|
1547 |
+
"logps/rejected": -59.65806579589844,
|
1548 |
+
"loss": 0.2966,
|
1549 |
+
"rewards/accuracies": 0.9375,
|
1550 |
+
"rewards/chosen": 17.192535400390625,
|
1551 |
+
"rewards/margins": 5.7006731033325195,
|
1552 |
+
"rewards/rejected": 11.491861343383789,
|
1553 |
+
"step": 880
|
1554 |
+
},
|
1555 |
+
{
|
1556 |
+
"epoch": 0.7911111111111111,
|
1557 |
+
"grad_norm": 0.7536466717720032,
|
1558 |
+
"learning_rate": 6.362640205293583e-07,
|
1559 |
+
"logits/chosen": -0.0016857212176546454,
|
1560 |
+
"logits/rejected": -0.010936147533357143,
|
1561 |
+
"logps/chosen": -2.1478958129882812,
|
1562 |
+
"logps/rejected": -58.20386505126953,
|
1563 |
+
"loss": 0.1641,
|
1564 |
+
"rewards/accuracies": 0.9375,
|
1565 |
+
"rewards/chosen": 17.243247985839844,
|
1566 |
+
"rewards/margins": 5.678771018981934,
|
1567 |
+
"rewards/rejected": 11.564477920532227,
|
1568 |
+
"step": 890
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 0.8,
|
1572 |
+
"grad_norm": 0.3857377767562866,
|
1573 |
+
"learning_rate": 5.854432926863684e-07,
|
1574 |
+
"logits/chosen": 0.00038508616853505373,
|
1575 |
+
"logits/rejected": -0.011034643277525902,
|
1576 |
+
"logps/chosen": -1.606274127960205,
|
1577 |
+
"logps/rejected": -62.58662414550781,
|
1578 |
+
"loss": 0.1703,
|
1579 |
+
"rewards/accuracies": 0.9624999761581421,
|
1580 |
+
"rewards/chosen": 17.211261749267578,
|
1581 |
+
"rewards/margins": 5.966868877410889,
|
1582 |
+
"rewards/rejected": 11.244392395019531,
|
1583 |
+
"step": 900
|
1584 |
+
},
|
1585 |
+
{
|
1586 |
+
"epoch": 0.8,
|
1587 |
+
"eval_logits/chosen": 0.00021816430671606213,
|
1588 |
+
"eval_logits/rejected": -0.010477552190423012,
|
1589 |
+
"eval_logps/chosen": -1.7576563358306885,
|
1590 |
+
"eval_logps/rejected": -59.79063415527344,
|
1591 |
+
"eval_loss": 0.13995186984539032,
|
1592 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1593 |
+
"eval_rewards/chosen": 17.327129364013672,
|
1594 |
+
"eval_rewards/margins": 5.945469856262207,
|
1595 |
+
"eval_rewards/rejected": 11.381658554077148,
|
1596 |
+
"eval_runtime": 361.4592,
|
1597 |
+
"eval_samples_per_second": 2.767,
|
1598 |
+
"eval_steps_per_second": 0.346,
|
1599 |
+
"step": 900
|
1600 |
+
},
|
1601 |
+
{
|
1602 |
+
"epoch": 0.8088888888888889,
|
1603 |
+
"grad_norm": 0.13143697381019592,
|
1604 |
+
"learning_rate": 5.364674616415547e-07,
|
1605 |
+
"logits/chosen": 0.0005570838693529367,
|
1606 |
+
"logits/rejected": -0.011198626831173897,
|
1607 |
+
"logps/chosen": -0.12791283428668976,
|
1608 |
+
"logps/rejected": -65.7052993774414,
|
1609 |
+
"loss": 0.0507,
|
1610 |
+
"rewards/accuracies": 0.9750000238418579,
|
1611 |
+
"rewards/chosen": 17.224443435668945,
|
1612 |
+
"rewards/margins": 6.163690567016602,
|
1613 |
+
"rewards/rejected": 11.060752868652344,
|
1614 |
+
"step": 910
|
1615 |
+
},
|
1616 |
+
{
|
1617 |
+
"epoch": 0.8177777777777778,
|
1618 |
+
"grad_norm": 102.9654541015625,
|
1619 |
+
"learning_rate": 4.893837212719859e-07,
|
1620 |
+
"logits/chosen": -0.0008557128603570163,
|
1621 |
+
"logits/rejected": -0.01163212489336729,
|
1622 |
+
"logps/chosen": -1.3292646408081055,
|
1623 |
+
"logps/rejected": -61.644893646240234,
|
1624 |
+
"loss": 0.0881,
|
1625 |
+
"rewards/accuracies": 0.9500000476837158,
|
1626 |
+
"rewards/chosen": 17.276538848876953,
|
1627 |
+
"rewards/margins": 5.96080207824707,
|
1628 |
+
"rewards/rejected": 11.3157377243042,
|
1629 |
+
"step": 920
|
1630 |
+
},
|
1631 |
+
{
|
1632 |
+
"epoch": 0.8266666666666667,
|
1633 |
+
"grad_norm": 146.04498291015625,
|
1634 |
+
"learning_rate": 4.442374422065493e-07,
|
1635 |
+
"logits/chosen": 0.002922601066529751,
|
1636 |
+
"logits/rejected": -0.007130052894353867,
|
1637 |
+
"logps/chosen": -1.6069023609161377,
|
1638 |
+
"logps/rejected": -61.52588653564453,
|
1639 |
+
"loss": 0.1373,
|
1640 |
+
"rewards/accuracies": 0.949999988079071,
|
1641 |
+
"rewards/chosen": 17.184974670410156,
|
1642 |
+
"rewards/margins": 5.7995734214782715,
|
1643 |
+
"rewards/rejected": 11.38540267944336,
|
1644 |
+
"step": 930
|
1645 |
+
},
|
1646 |
+
{
|
1647 |
+
"epoch": 0.8355555555555556,
|
1648 |
+
"grad_norm": 23.539485931396484,
|
1649 |
+
"learning_rate": 4.0107212810610974e-07,
|
1650 |
+
"logits/chosen": 0.0018056132830679417,
|
1651 |
+
"logits/rejected": -0.007847340777516365,
|
1652 |
+
"logps/chosen": -3.637047290802002,
|
1653 |
+
"logps/rejected": -61.21245574951172,
|
1654 |
+
"loss": 0.2763,
|
1655 |
+
"rewards/accuracies": 0.925000011920929,
|
1656 |
+
"rewards/chosen": 16.960758209228516,
|
1657 |
+
"rewards/margins": 5.551811695098877,
|
1658 |
+
"rewards/rejected": 11.40894603729248,
|
1659 |
+
"step": 940
|
1660 |
+
},
|
1661 |
+
{
|
1662 |
+
"epoch": 0.8444444444444444,
|
1663 |
+
"grad_norm": 31.52926254272461,
|
1664 |
+
"learning_rate": 3.599293737426932e-07,
|
1665 |
+
"logits/chosen": 0.0032081177923828363,
|
1666 |
+
"logits/rejected": -0.007756482809782028,
|
1667 |
+
"logps/chosen": -1.172515869140625,
|
1668 |
+
"logps/rejected": -66.5853271484375,
|
1669 |
+
"loss": 0.1758,
|
1670 |
+
"rewards/accuracies": 0.949999988079071,
|
1671 |
+
"rewards/chosen": 16.992876052856445,
|
1672 |
+
"rewards/margins": 5.886469841003418,
|
1673 |
+
"rewards/rejected": 11.106407165527344,
|
1674 |
+
"step": 950
|
1675 |
+
},
|
1676 |
+
{
|
1677 |
+
"epoch": 0.8533333333333334,
|
1678 |
+
"grad_norm": 0.07474468648433685,
|
1679 |
+
"learning_rate": 3.208488249181216e-07,
|
1680 |
+
"logits/chosen": 0.0022508346009999514,
|
1681 |
+
"logits/rejected": -0.009156409651041031,
|
1682 |
+
"logps/chosen": -0.7930470108985901,
|
1683 |
+
"logps/rejected": -60.068790435791016,
|
1684 |
+
"loss": 0.1138,
|
1685 |
+
"rewards/accuracies": 0.9624999761581421,
|
1686 |
+
"rewards/chosen": 17.401771545410156,
|
1687 |
+
"rewards/margins": 6.0026960372924805,
|
1688 |
+
"rewards/rejected": 11.399076461791992,
|
1689 |
+
"step": 960
|
1690 |
+
},
|
1691 |
+
{
|
1692 |
+
"epoch": 0.8533333333333334,
|
1693 |
+
"eval_logits/chosen": 0.001469604205340147,
|
1694 |
+
"eval_logits/rejected": -0.009397665038704872,
|
1695 |
+
"eval_logps/chosen": -1.8795456886291504,
|
1696 |
+
"eval_logps/rejected": -60.17564010620117,
|
1697 |
+
"eval_loss": 0.1441129744052887,
|
1698 |
+
"eval_rewards/accuracies": 0.9630000591278076,
|
1699 |
+
"eval_rewards/chosen": 17.314937591552734,
|
1700 |
+
"eval_rewards/margins": 5.9717817306518555,
|
1701 |
+
"eval_rewards/rejected": 11.343156814575195,
|
1702 |
+
"eval_runtime": 361.5344,
|
1703 |
+
"eval_samples_per_second": 2.766,
|
1704 |
+
"eval_steps_per_second": 0.346,
|
1705 |
+
"step": 960
|
1706 |
+
},
|
1707 |
+
{
|
1708 |
+
"epoch": 0.8622222222222222,
|
1709 |
+
"grad_norm": 68.90747833251953,
|
1710 |
+
"learning_rate": 2.838681402606952e-07,
|
1711 |
+
"logits/chosen": 0.004552370868623257,
|
1712 |
+
"logits/rejected": -0.005488495342433453,
|
1713 |
+
"logps/chosen": -3.7298974990844727,
|
1714 |
+
"logps/rejected": -64.72488403320312,
|
1715 |
+
"loss": 0.2425,
|
1716 |
+
"rewards/accuracies": 0.925000011920929,
|
1717 |
+
"rewards/chosen": 16.832754135131836,
|
1718 |
+
"rewards/margins": 5.644216537475586,
|
1719 |
+
"rewards/rejected": 11.18853759765625,
|
1720 |
+
"step": 970
|
1721 |
+
},
|
1722 |
+
{
|
1723 |
+
"epoch": 0.8711111111111111,
|
1724 |
+
"grad_norm": 0.028206102550029755,
|
1725 |
+
"learning_rate": 2.490229549367443e-07,
|
1726 |
+
"logits/chosen": 0.0025807656347751617,
|
1727 |
+
"logits/rejected": -0.008657123893499374,
|
1728 |
+
"logps/chosen": -1.1825838088989258,
|
1729 |
+
"logps/rejected": -63.57493591308594,
|
1730 |
+
"loss": 0.0609,
|
1731 |
+
"rewards/accuracies": 0.9750000238418579,
|
1732 |
+
"rewards/chosen": 17.227130889892578,
|
1733 |
+
"rewards/margins": 6.053717136383057,
|
1734 |
+
"rewards/rejected": 11.173412322998047,
|
1735 |
+
"step": 980
|
1736 |
+
},
|
1737 |
+
{
|
1738 |
+
"epoch": 0.88,
|
1739 |
+
"grad_norm": 0.39742231369018555,
|
1740 |
+
"learning_rate": 2.1634684631203412e-07,
|
1741 |
+
"logits/chosen": 0.0048486413434147835,
|
1742 |
+
"logits/rejected": -0.006056814920157194,
|
1743 |
+
"logps/chosen": -2.919680595397949,
|
1744 |
+
"logps/rejected": -57.8403205871582,
|
1745 |
+
"loss": 0.3464,
|
1746 |
+
"rewards/accuracies": 0.8999999761581421,
|
1747 |
+
"rewards/chosen": 17.175275802612305,
|
1748 |
+
"rewards/margins": 5.559727668762207,
|
1749 |
+
"rewards/rejected": 11.615548133850098,
|
1750 |
+
"step": 990
|
1751 |
+
},
|
1752 |
+
{
|
1753 |
+
"epoch": 0.8888888888888888,
|
1754 |
+
"grad_norm": 30.22509002685547,
|
1755 |
+
"learning_rate": 1.8587130159608196e-07,
|
1756 |
+
"logits/chosen": 0.0050649940967559814,
|
1757 |
+
"logits/rejected": -0.0071399761363863945,
|
1758 |
+
"logps/chosen": -0.16131475567817688,
|
1759 |
+
"logps/rejected": -66.25190734863281,
|
1760 |
+
"loss": 0.0063,
|
1761 |
+
"rewards/accuracies": 1.0,
|
1762 |
+
"rewards/chosen": 17.313472747802734,
|
1763 |
+
"rewards/margins": 6.412895679473877,
|
1764 |
+
"rewards/rejected": 10.9005765914917,
|
1765 |
+
"step": 1000
|
1766 |
+
}
|
1767 |
+
],
|
1768 |
+
"logging_steps": 10,
|
1769 |
+
"max_steps": 1125,
|
1770 |
+
"num_input_tokens_seen": 0,
|
1771 |
+
"num_train_epochs": 1,
|
1772 |
+
"save_steps": 500,
|
1773 |
+
"stateful_callbacks": {
|
1774 |
+
"TrainerControl": {
|
1775 |
+
"args": {
|
1776 |
+
"should_epoch_stop": false,
|
1777 |
+
"should_evaluate": false,
|
1778 |
+
"should_log": false,
|
1779 |
+
"should_save": true,
|
1780 |
+
"should_training_stop": false
|
1781 |
+
},
|
1782 |
+
"attributes": {}
|
1783 |
+
}
|
1784 |
+
},
|
1785 |
+
"total_flos": 1.2745297197268992e+18,
|
1786 |
+
"train_batch_size": 4,
|
1787 |
+
"trial_name": null,
|
1788 |
+
"trial_params": null
|
1789 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5286bf380083ef64133f25e161ddfb8549ff44790e432f478753f221ef89e695
|
3 |
+
size 5368
|
checkpoint-1125/README.md
ADDED
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: mistralai/Mistral-Nemo-Instruct-2407
|
3 |
+
library_name: peft
|
4 |
+
---
|
5 |
+
|
6 |
+
# Model Card for Model ID
|
7 |
+
|
8 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
9 |
+
|
10 |
+
|
11 |
+
|
12 |
+
## Model Details
|
13 |
+
|
14 |
+
### Model Description
|
15 |
+
|
16 |
+
<!-- Provide a longer summary of what this model is. -->
|
17 |
+
|
18 |
+
|
19 |
+
|
20 |
+
- **Developed by:** [More Information Needed]
|
21 |
+
- **Funded by [optional]:** [More Information Needed]
|
22 |
+
- **Shared by [optional]:** [More Information Needed]
|
23 |
+
- **Model type:** [More Information Needed]
|
24 |
+
- **Language(s) (NLP):** [More Information Needed]
|
25 |
+
- **License:** [More Information Needed]
|
26 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
27 |
+
|
28 |
+
### Model Sources [optional]
|
29 |
+
|
30 |
+
<!-- Provide the basic links for the model. -->
|
31 |
+
|
32 |
+
- **Repository:** [More Information Needed]
|
33 |
+
- **Paper [optional]:** [More Information Needed]
|
34 |
+
- **Demo [optional]:** [More Information Needed]
|
35 |
+
|
36 |
+
## Uses
|
37 |
+
|
38 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
39 |
+
|
40 |
+
### Direct Use
|
41 |
+
|
42 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
43 |
+
|
44 |
+
[More Information Needed]
|
45 |
+
|
46 |
+
### Downstream Use [optional]
|
47 |
+
|
48 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
49 |
+
|
50 |
+
[More Information Needed]
|
51 |
+
|
52 |
+
### Out-of-Scope Use
|
53 |
+
|
54 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
55 |
+
|
56 |
+
[More Information Needed]
|
57 |
+
|
58 |
+
## Bias, Risks, and Limitations
|
59 |
+
|
60 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
61 |
+
|
62 |
+
[More Information Needed]
|
63 |
+
|
64 |
+
### Recommendations
|
65 |
+
|
66 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
67 |
+
|
68 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
69 |
+
|
70 |
+
## How to Get Started with the Model
|
71 |
+
|
72 |
+
Use the code below to get started with the model.
|
73 |
+
|
74 |
+
[More Information Needed]
|
75 |
+
|
76 |
+
## Training Details
|
77 |
+
|
78 |
+
### Training Data
|
79 |
+
|
80 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
81 |
+
|
82 |
+
[More Information Needed]
|
83 |
+
|
84 |
+
### Training Procedure
|
85 |
+
|
86 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
87 |
+
|
88 |
+
#### Preprocessing [optional]
|
89 |
+
|
90 |
+
[More Information Needed]
|
91 |
+
|
92 |
+
|
93 |
+
#### Training Hyperparameters
|
94 |
+
|
95 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
96 |
+
|
97 |
+
#### Speeds, Sizes, Times [optional]
|
98 |
+
|
99 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
100 |
+
|
101 |
+
[More Information Needed]
|
102 |
+
|
103 |
+
## Evaluation
|
104 |
+
|
105 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
106 |
+
|
107 |
+
### Testing Data, Factors & Metrics
|
108 |
+
|
109 |
+
#### Testing Data
|
110 |
+
|
111 |
+
<!-- This should link to a Dataset Card if possible. -->
|
112 |
+
|
113 |
+
[More Information Needed]
|
114 |
+
|
115 |
+
#### Factors
|
116 |
+
|
117 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
118 |
+
|
119 |
+
[More Information Needed]
|
120 |
+
|
121 |
+
#### Metrics
|
122 |
+
|
123 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
124 |
+
|
125 |
+
[More Information Needed]
|
126 |
+
|
127 |
+
### Results
|
128 |
+
|
129 |
+
[More Information Needed]
|
130 |
+
|
131 |
+
#### Summary
|
132 |
+
|
133 |
+
|
134 |
+
|
135 |
+
## Model Examination [optional]
|
136 |
+
|
137 |
+
<!-- Relevant interpretability work for the model goes here -->
|
138 |
+
|
139 |
+
[More Information Needed]
|
140 |
+
|
141 |
+
## Environmental Impact
|
142 |
+
|
143 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
144 |
+
|
145 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
146 |
+
|
147 |
+
- **Hardware Type:** [More Information Needed]
|
148 |
+
- **Hours used:** [More Information Needed]
|
149 |
+
- **Cloud Provider:** [More Information Needed]
|
150 |
+
- **Compute Region:** [More Information Needed]
|
151 |
+
- **Carbon Emitted:** [More Information Needed]
|
152 |
+
|
153 |
+
## Technical Specifications [optional]
|
154 |
+
|
155 |
+
### Model Architecture and Objective
|
156 |
+
|
157 |
+
[More Information Needed]
|
158 |
+
|
159 |
+
### Compute Infrastructure
|
160 |
+
|
161 |
+
[More Information Needed]
|
162 |
+
|
163 |
+
#### Hardware
|
164 |
+
|
165 |
+
[More Information Needed]
|
166 |
+
|
167 |
+
#### Software
|
168 |
+
|
169 |
+
[More Information Needed]
|
170 |
+
|
171 |
+
## Citation [optional]
|
172 |
+
|
173 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
174 |
+
|
175 |
+
**BibTeX:**
|
176 |
+
|
177 |
+
[More Information Needed]
|
178 |
+
|
179 |
+
**APA:**
|
180 |
+
|
181 |
+
[More Information Needed]
|
182 |
+
|
183 |
+
## Glossary [optional]
|
184 |
+
|
185 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
186 |
+
|
187 |
+
[More Information Needed]
|
188 |
+
|
189 |
+
## More Information [optional]
|
190 |
+
|
191 |
+
[More Information Needed]
|
192 |
+
|
193 |
+
## Model Card Authors [optional]
|
194 |
+
|
195 |
+
[More Information Needed]
|
196 |
+
|
197 |
+
## Model Card Contact
|
198 |
+
|
199 |
+
[More Information Needed]
|
200 |
+
### Framework versions
|
201 |
+
|
202 |
+
- PEFT 0.12.0
|
checkpoint-1125/adapter_config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "mistralai/Mistral-Nemo-Instruct-2407",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layer_replication": null,
|
10 |
+
"layers_pattern": null,
|
11 |
+
"layers_to_transform": null,
|
12 |
+
"loftq_config": {},
|
13 |
+
"lora_alpha": 16,
|
14 |
+
"lora_dropout": 0.0,
|
15 |
+
"megatron_config": null,
|
16 |
+
"megatron_core": "megatron.core",
|
17 |
+
"modules_to_save": null,
|
18 |
+
"peft_type": "LORA",
|
19 |
+
"r": 8,
|
20 |
+
"rank_pattern": {},
|
21 |
+
"revision": null,
|
22 |
+
"target_modules": [
|
23 |
+
"q_proj",
|
24 |
+
"gate_proj",
|
25 |
+
"k_proj",
|
26 |
+
"o_proj",
|
27 |
+
"v_proj",
|
28 |
+
"down_proj",
|
29 |
+
"up_proj"
|
30 |
+
],
|
31 |
+
"task_type": "CAUSAL_LM",
|
32 |
+
"use_dora": false,
|
33 |
+
"use_rslora": false
|
34 |
+
}
|
checkpoint-1125/adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6404a5e1c60aa26a99a4ccbb402f485483d5eea90c08441f0448e29c51e2b1d0
|
3 |
+
size 114106856
|
checkpoint-1125/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:679b631e5cc31300cab94ec2968b5eafa50b74a27dc17dffbac7d227a1ae878e
|
3 |
+
size 228536930
|
checkpoint-1125/rng_state_0.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:55f6ef5fd81f87dc8a69f1b9a751ba37cc49c37318322e45ba4733ff23a92208
|
3 |
+
size 14512
|
checkpoint-1125/rng_state_1.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f510db96e40d7f66609c96cf485c13417fc4eaf253603d2b6591466c3fb5f63a
|
3 |
+
size 14512
|
checkpoint-1125/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb6eb7801fbbb5055fabd75421a32f0209bef58fe235404c9e167738db177888
|
3 |
+
size 1064
|
checkpoint-1125/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "</s>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
checkpoint-1125/tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0240ce510f08e6c2041724e9043e33be9d251d1e4a4d94eb68cd47b954b61d2
|
3 |
+
size 17078292
|
checkpoint-1125/tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-1125/trainer_state.json
ADDED
@@ -0,0 +1,2001 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.0,
|
5 |
+
"eval_steps": 60,
|
6 |
+
"global_step": 1125,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.008888888888888889,
|
13 |
+
"grad_norm": 8.760091781616211,
|
14 |
+
"learning_rate": 4.4247787610619474e-07,
|
15 |
+
"logits/chosen": -0.8248252868652344,
|
16 |
+
"logits/rejected": -0.8263720273971558,
|
17 |
+
"logps/chosen": -0.36086463928222656,
|
18 |
+
"logps/rejected": -5.696224689483643,
|
19 |
+
"loss": 1.1038,
|
20 |
+
"rewards/accuracies": 0.5125000476837158,
|
21 |
+
"rewards/chosen": 17.43745994567871,
|
22 |
+
"rewards/margins": 0.5984855890274048,
|
23 |
+
"rewards/rejected": 16.838973999023438,
|
24 |
+
"step": 10
|
25 |
+
},
|
26 |
+
{
|
27 |
+
"epoch": 0.017777777777777778,
|
28 |
+
"grad_norm": 8.855981826782227,
|
29 |
+
"learning_rate": 8.849557522123895e-07,
|
30 |
+
"logits/chosen": -0.8169006109237671,
|
31 |
+
"logits/rejected": -0.819770872592926,
|
32 |
+
"logps/chosen": -0.12464660406112671,
|
33 |
+
"logps/rejected": -7.139842987060547,
|
34 |
+
"loss": 1.1887,
|
35 |
+
"rewards/accuracies": 0.4000000059604645,
|
36 |
+
"rewards/chosen": 17.17649269104004,
|
37 |
+
"rewards/margins": 0.19107049703598022,
|
38 |
+
"rewards/rejected": 16.98542022705078,
|
39 |
+
"step": 20
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 0.02666666666666667,
|
43 |
+
"grad_norm": 16.764184951782227,
|
44 |
+
"learning_rate": 1.3274336283185843e-06,
|
45 |
+
"logits/chosen": -0.8003113865852356,
|
46 |
+
"logits/rejected": -0.8030117750167847,
|
47 |
+
"logps/chosen": -0.34651467204093933,
|
48 |
+
"logps/rejected": -6.967917442321777,
|
49 |
+
"loss": 1.0563,
|
50 |
+
"rewards/accuracies": 0.44999998807907104,
|
51 |
+
"rewards/chosen": 17.280975341796875,
|
52 |
+
"rewards/margins": 0.40005987882614136,
|
53 |
+
"rewards/rejected": 16.88091468811035,
|
54 |
+
"step": 30
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 0.035555555555555556,
|
58 |
+
"grad_norm": 8.33682918548584,
|
59 |
+
"learning_rate": 1.769911504424779e-06,
|
60 |
+
"logits/chosen": -0.7695047855377197,
|
61 |
+
"logits/rejected": -0.7739207148551941,
|
62 |
+
"logps/chosen": -1.5993006229400635,
|
63 |
+
"logps/rejected": -8.504932403564453,
|
64 |
+
"loss": 0.7596,
|
65 |
+
"rewards/accuracies": 0.5,
|
66 |
+
"rewards/chosen": 17.283912658691406,
|
67 |
+
"rewards/margins": 0.6976072192192078,
|
68 |
+
"rewards/rejected": 16.5863037109375,
|
69 |
+
"step": 40
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 0.044444444444444446,
|
73 |
+
"grad_norm": 4.494723320007324,
|
74 |
+
"learning_rate": 2.212389380530974e-06,
|
75 |
+
"logits/chosen": -0.7154140472412109,
|
76 |
+
"logits/rejected": -0.7225576043128967,
|
77 |
+
"logps/chosen": -3.112199068069458,
|
78 |
+
"logps/rejected": -12.212080001831055,
|
79 |
+
"loss": 0.6083,
|
80 |
+
"rewards/accuracies": 0.4625000059604645,
|
81 |
+
"rewards/chosen": 17.03064727783203,
|
82 |
+
"rewards/margins": 0.7148451805114746,
|
83 |
+
"rewards/rejected": 16.3158016204834,
|
84 |
+
"step": 50
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.05333333333333334,
|
88 |
+
"grad_norm": 5.110287666320801,
|
89 |
+
"learning_rate": 2.6548672566371687e-06,
|
90 |
+
"logits/chosen": -0.6322453022003174,
|
91 |
+
"logits/rejected": -0.6387485265731812,
|
92 |
+
"logps/chosen": -5.650620460510254,
|
93 |
+
"logps/rejected": -12.759811401367188,
|
94 |
+
"loss": 0.3835,
|
95 |
+
"rewards/accuracies": 0.9125000238418579,
|
96 |
+
"rewards/chosen": 17.101289749145508,
|
97 |
+
"rewards/margins": 1.1824612617492676,
|
98 |
+
"rewards/rejected": 15.918828964233398,
|
99 |
+
"step": 60
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.05333333333333334,
|
103 |
+
"eval_logits/chosen": -0.5826543569564819,
|
104 |
+
"eval_logits/rejected": -0.5914276838302612,
|
105 |
+
"eval_logps/chosen": -3.5471787452697754,
|
106 |
+
"eval_logps/rejected": -16.51181983947754,
|
107 |
+
"eval_loss": 0.3286525011062622,
|
108 |
+
"eval_rewards/accuracies": 0.9280000925064087,
|
109 |
+
"eval_rewards/chosen": 17.148174285888672,
|
110 |
+
"eval_rewards/margins": 1.4386365413665771,
|
111 |
+
"eval_rewards/rejected": 15.709539413452148,
|
112 |
+
"eval_runtime": 372.0227,
|
113 |
+
"eval_samples_per_second": 2.688,
|
114 |
+
"eval_steps_per_second": 0.336,
|
115 |
+
"step": 60
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 0.06222222222222222,
|
119 |
+
"grad_norm": 5.098133563995361,
|
120 |
+
"learning_rate": 3.097345132743363e-06,
|
121 |
+
"logits/chosen": -0.5378152132034302,
|
122 |
+
"logits/rejected": -0.5494933724403381,
|
123 |
+
"logps/chosen": -1.5099802017211914,
|
124 |
+
"logps/rejected": -21.206321716308594,
|
125 |
+
"loss": 0.2931,
|
126 |
+
"rewards/accuracies": 0.9375,
|
127 |
+
"rewards/chosen": 17.083791732788086,
|
128 |
+
"rewards/margins": 1.5844331979751587,
|
129 |
+
"rewards/rejected": 15.499359130859375,
|
130 |
+
"step": 70
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 0.07111111111111111,
|
134 |
+
"grad_norm": 29.787437438964844,
|
135 |
+
"learning_rate": 3.539823008849558e-06,
|
136 |
+
"logits/chosen": -0.443774938583374,
|
137 |
+
"logits/rejected": -0.45571577548980713,
|
138 |
+
"logps/chosen": -1.5804342031478882,
|
139 |
+
"logps/rejected": -22.606929779052734,
|
140 |
+
"loss": 0.202,
|
141 |
+
"rewards/accuracies": 0.9750000238418579,
|
142 |
+
"rewards/chosen": 17.302125930786133,
|
143 |
+
"rewards/margins": 2.174014091491699,
|
144 |
+
"rewards/rejected": 15.128110885620117,
|
145 |
+
"step": 80
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"epoch": 0.08,
|
149 |
+
"grad_norm": 23.14398193359375,
|
150 |
+
"learning_rate": 3.982300884955752e-06,
|
151 |
+
"logits/chosen": -0.3626072406768799,
|
152 |
+
"logits/rejected": -0.3787815570831299,
|
153 |
+
"logps/chosen": -2.203828811645508,
|
154 |
+
"logps/rejected": -29.433551788330078,
|
155 |
+
"loss": 0.2123,
|
156 |
+
"rewards/accuracies": 0.925000011920929,
|
157 |
+
"rewards/chosen": 17.00284194946289,
|
158 |
+
"rewards/margins": 2.320391893386841,
|
159 |
+
"rewards/rejected": 14.682450294494629,
|
160 |
+
"step": 90
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 0.08888888888888889,
|
164 |
+
"grad_norm": 29.672739028930664,
|
165 |
+
"learning_rate": 4.424778761061948e-06,
|
166 |
+
"logits/chosen": -0.3035663962364197,
|
167 |
+
"logits/rejected": -0.31762221455574036,
|
168 |
+
"logps/chosen": -3.433589458465576,
|
169 |
+
"logps/rejected": -29.9322509765625,
|
170 |
+
"loss": 0.2592,
|
171 |
+
"rewards/accuracies": 0.9375,
|
172 |
+
"rewards/chosen": 16.929956436157227,
|
173 |
+
"rewards/margins": 2.31272029876709,
|
174 |
+
"rewards/rejected": 14.617237091064453,
|
175 |
+
"step": 100
|
176 |
+
},
|
177 |
+
{
|
178 |
+
"epoch": 0.09777777777777778,
|
179 |
+
"grad_norm": 1.873722791671753,
|
180 |
+
"learning_rate": 4.867256637168142e-06,
|
181 |
+
"logits/chosen": -0.2679600715637207,
|
182 |
+
"logits/rejected": -0.2826440930366516,
|
183 |
+
"logps/chosen": -0.9653514623641968,
|
184 |
+
"logps/rejected": -30.235322952270508,
|
185 |
+
"loss": 0.1336,
|
186 |
+
"rewards/accuracies": 0.949999988079071,
|
187 |
+
"rewards/chosen": 17.462385177612305,
|
188 |
+
"rewards/margins": 3.1994175910949707,
|
189 |
+
"rewards/rejected": 14.26296615600586,
|
190 |
+
"step": 110
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 0.10666666666666667,
|
194 |
+
"grad_norm": 1.6913721561431885,
|
195 |
+
"learning_rate": 4.999409761242696e-06,
|
196 |
+
"logits/chosen": -0.22222033143043518,
|
197 |
+
"logits/rejected": -0.23720571398735046,
|
198 |
+
"logps/chosen": -4.4953508377075195,
|
199 |
+
"logps/rejected": -34.074745178222656,
|
200 |
+
"loss": 0.2552,
|
201 |
+
"rewards/accuracies": 0.8999999761581421,
|
202 |
+
"rewards/chosen": 17.04866600036621,
|
203 |
+
"rewards/margins": 3.1014418601989746,
|
204 |
+
"rewards/rejected": 13.947224617004395,
|
205 |
+
"step": 120
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.10666666666666667,
|
209 |
+
"eval_logits/chosen": -0.206527978181839,
|
210 |
+
"eval_logits/rejected": -0.22178640961647034,
|
211 |
+
"eval_logps/chosen": -3.69442081451416,
|
212 |
+
"eval_logps/rejected": -36.072166442871094,
|
213 |
+
"eval_loss": 0.18996010720729828,
|
214 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
215 |
+
"eval_rewards/chosen": 17.133451461791992,
|
216 |
+
"eval_rewards/margins": 3.379946708679199,
|
217 |
+
"eval_rewards/rejected": 13.753504753112793,
|
218 |
+
"eval_runtime": 361.5279,
|
219 |
+
"eval_samples_per_second": 2.766,
|
220 |
+
"eval_steps_per_second": 0.346,
|
221 |
+
"step": 120
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 0.11555555555555555,
|
225 |
+
"grad_norm": 61.80262756347656,
|
226 |
+
"learning_rate": 4.996519466816778e-06,
|
227 |
+
"logits/chosen": -0.18473535776138306,
|
228 |
+
"logits/rejected": -0.1988501250743866,
|
229 |
+
"logps/chosen": -3.7009687423706055,
|
230 |
+
"logps/rejected": -39.289939880371094,
|
231 |
+
"loss": 0.1394,
|
232 |
+
"rewards/accuracies": 0.9624999761581421,
|
233 |
+
"rewards/chosen": 17.106964111328125,
|
234 |
+
"rewards/margins": 3.633338212966919,
|
235 |
+
"rewards/rejected": 13.473625183105469,
|
236 |
+
"step": 130
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.12444444444444444,
|
240 |
+
"grad_norm": 1.6732702255249023,
|
241 |
+
"learning_rate": 4.9912234871722805e-06,
|
242 |
+
"logits/chosen": -0.16134041547775269,
|
243 |
+
"logits/rejected": -0.17547868192195892,
|
244 |
+
"logps/chosen": -3.0637736320495605,
|
245 |
+
"logps/rejected": -40.07548522949219,
|
246 |
+
"loss": 0.1408,
|
247 |
+
"rewards/accuracies": 0.9750000238418579,
|
248 |
+
"rewards/chosen": 17.392223358154297,
|
249 |
+
"rewards/margins": 4.242353439331055,
|
250 |
+
"rewards/rejected": 13.149867057800293,
|
251 |
+
"step": 140
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 0.13333333333333333,
|
255 |
+
"grad_norm": 0.346453994512558,
|
256 |
+
"learning_rate": 4.98352692559805e-06,
|
257 |
+
"logits/chosen": -0.13797929883003235,
|
258 |
+
"logits/rejected": -0.15283086895942688,
|
259 |
+
"logps/chosen": -5.14492130279541,
|
260 |
+
"logps/rejected": -47.97212219238281,
|
261 |
+
"loss": 0.2153,
|
262 |
+
"rewards/accuracies": 0.9375,
|
263 |
+
"rewards/chosen": 16.896778106689453,
|
264 |
+
"rewards/margins": 4.227695465087891,
|
265 |
+
"rewards/rejected": 12.669081687927246,
|
266 |
+
"step": 150
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.14222222222222222,
|
270 |
+
"grad_norm": 0.21871662139892578,
|
271 |
+
"learning_rate": 4.973437198621237e-06,
|
272 |
+
"logits/chosen": -0.12396670132875443,
|
273 |
+
"logits/rejected": -0.13780555129051208,
|
274 |
+
"logps/chosen": -6.108860015869141,
|
275 |
+
"logps/rejected": -54.90739440917969,
|
276 |
+
"loss": 0.0388,
|
277 |
+
"rewards/accuracies": 0.9750000238418579,
|
278 |
+
"rewards/chosen": 16.75935935974121,
|
279 |
+
"rewards/margins": 4.755282878875732,
|
280 |
+
"rewards/rejected": 12.004077911376953,
|
281 |
+
"step": 160
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.1511111111111111,
|
285 |
+
"grad_norm": 235.12429809570312,
|
286 |
+
"learning_rate": 4.960964028860621e-06,
|
287 |
+
"logits/chosen": -0.1140839159488678,
|
288 |
+
"logits/rejected": -0.1263057291507721,
|
289 |
+
"logps/chosen": -12.605452537536621,
|
290 |
+
"logps/rejected": -53.81230926513672,
|
291 |
+
"loss": 0.4651,
|
292 |
+
"rewards/accuracies": 0.875,
|
293 |
+
"rewards/chosen": 16.101238250732422,
|
294 |
+
"rewards/margins": 3.9864249229431152,
|
295 |
+
"rewards/rejected": 12.114812850952148,
|
296 |
+
"step": 170
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.16,
|
300 |
+
"grad_norm": 190.97048950195312,
|
301 |
+
"learning_rate": 4.946119435657738e-06,
|
302 |
+
"logits/chosen": -0.10746976733207703,
|
303 |
+
"logits/rejected": -0.11878640949726105,
|
304 |
+
"logps/chosen": -8.5105562210083,
|
305 |
+
"logps/rejected": -51.314781188964844,
|
306 |
+
"loss": 0.2362,
|
307 |
+
"rewards/accuracies": 0.925000011920929,
|
308 |
+
"rewards/chosen": 16.719980239868164,
|
309 |
+
"rewards/margins": 4.549674034118652,
|
310 |
+
"rewards/rejected": 12.170306205749512,
|
311 |
+
"step": 180
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 0.16,
|
315 |
+
"eval_logits/chosen": -0.10870806127786636,
|
316 |
+
"eval_logits/rejected": -0.12223993986845016,
|
317 |
+
"eval_logps/chosen": -4.414996147155762,
|
318 |
+
"eval_logps/rejected": -53.885032653808594,
|
319 |
+
"eval_loss": 0.20236633718013763,
|
320 |
+
"eval_rewards/accuracies": 0.9510000944137573,
|
321 |
+
"eval_rewards/chosen": 17.06139373779297,
|
322 |
+
"eval_rewards/margins": 5.089176177978516,
|
323 |
+
"eval_rewards/rejected": 11.97221851348877,
|
324 |
+
"eval_runtime": 361.4355,
|
325 |
+
"eval_samples_per_second": 2.767,
|
326 |
+
"eval_steps_per_second": 0.346,
|
327 |
+
"step": 180
|
328 |
+
},
|
329 |
+
{
|
330 |
+
"epoch": 0.1688888888888889,
|
331 |
+
"grad_norm": 56.81266784667969,
|
332 |
+
"learning_rate": 4.928917723494854e-06,
|
333 |
+
"logits/chosen": -0.10682469606399536,
|
334 |
+
"logits/rejected": -0.12124393880367279,
|
335 |
+
"logps/chosen": -3.058413028717041,
|
336 |
+
"logps/rejected": -55.052528381347656,
|
337 |
+
"loss": 0.2442,
|
338 |
+
"rewards/accuracies": 0.9500000476837158,
|
339 |
+
"rewards/chosen": 17.058589935302734,
|
340 |
+
"rewards/margins": 5.056097984313965,
|
341 |
+
"rewards/rejected": 12.002490043640137,
|
342 |
+
"step": 190
|
343 |
+
},
|
344 |
+
{
|
345 |
+
"epoch": 0.17777777777777778,
|
346 |
+
"grad_norm": 175.06552124023438,
|
347 |
+
"learning_rate": 4.909375468210947e-06,
|
348 |
+
"logits/chosen": -0.10520349442958832,
|
349 |
+
"logits/rejected": -0.12018950283527374,
|
350 |
+
"logps/chosen": -4.114959716796875,
|
351 |
+
"logps/rejected": -55.9394645690918,
|
352 |
+
"loss": 0.1915,
|
353 |
+
"rewards/accuracies": 0.9500000476837158,
|
354 |
+
"rewards/chosen": 16.98603057861328,
|
355 |
+
"rewards/margins": 5.105838775634766,
|
356 |
+
"rewards/rejected": 11.880191802978516,
|
357 |
+
"step": 200
|
358 |
+
},
|
359 |
+
{
|
360 |
+
"epoch": 0.18666666666666668,
|
361 |
+
"grad_norm": 78.06558990478516,
|
362 |
+
"learning_rate": 4.8875115010289655e-06,
|
363 |
+
"logits/chosen": -0.10475558042526245,
|
364 |
+
"logits/rejected": -0.11949175596237183,
|
365 |
+
"logps/chosen": -6.760301113128662,
|
366 |
+
"logps/rejected": -53.91607666015625,
|
367 |
+
"loss": 0.2843,
|
368 |
+
"rewards/accuracies": 0.9375,
|
369 |
+
"rewards/chosen": 16.857545852661133,
|
370 |
+
"rewards/margins": 4.917357921600342,
|
371 |
+
"rewards/rejected": 11.94018840789795,
|
372 |
+
"step": 210
|
373 |
+
},
|
374 |
+
{
|
375 |
+
"epoch": 0.19555555555555557,
|
376 |
+
"grad_norm": 15.880486488342285,
|
377 |
+
"learning_rate": 4.863346890409768e-06,
|
378 |
+
"logits/chosen": -0.11213523149490356,
|
379 |
+
"logits/rejected": -0.12581588327884674,
|
380 |
+
"logps/chosen": -6.759585380554199,
|
381 |
+
"logps/rejected": -51.10936737060547,
|
382 |
+
"loss": 0.5104,
|
383 |
+
"rewards/accuracies": 0.875,
|
384 |
+
"rewards/chosen": 16.859071731567383,
|
385 |
+
"rewards/margins": 4.638372898101807,
|
386 |
+
"rewards/rejected": 12.220698356628418,
|
387 |
+
"step": 220
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.20444444444444446,
|
391 |
+
"grad_norm": 46.97845458984375,
|
392 |
+
"learning_rate": 4.836904921750224e-06,
|
393 |
+
"logits/chosen": -0.11947059631347656,
|
394 |
+
"logits/rejected": -0.1329912692308426,
|
395 |
+
"logps/chosen": -3.608184814453125,
|
396 |
+
"logps/rejected": -48.794761657714844,
|
397 |
+
"loss": 0.2134,
|
398 |
+
"rewards/accuracies": 0.925000011920929,
|
399 |
+
"rewards/chosen": 17.235904693603516,
|
400 |
+
"rewards/margins": 4.859888076782227,
|
401 |
+
"rewards/rejected": 12.376014709472656,
|
402 |
+
"step": 230
|
403 |
+
},
|
404 |
+
{
|
405 |
+
"epoch": 0.21333333333333335,
|
406 |
+
"grad_norm": 24.032859802246094,
|
407 |
+
"learning_rate": 4.808211074945042e-06,
|
408 |
+
"logits/chosen": -0.1200513243675232,
|
409 |
+
"logits/rejected": -0.1333036869764328,
|
410 |
+
"logps/chosen": -3.7552154064178467,
|
411 |
+
"logps/rejected": -49.87453079223633,
|
412 |
+
"loss": 0.1781,
|
413 |
+
"rewards/accuracies": 0.9500000476837158,
|
414 |
+
"rewards/chosen": 17.094650268554688,
|
415 |
+
"rewards/margins": 4.68077278137207,
|
416 |
+
"rewards/rejected": 12.41387939453125,
|
417 |
+
"step": 240
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 0.21333333333333335,
|
421 |
+
"eval_logits/chosen": -0.12433278560638428,
|
422 |
+
"eval_logits/rejected": -0.13808581233024597,
|
423 |
+
"eval_logps/chosen": -4.408891201019287,
|
424 |
+
"eval_logps/rejected": -50.744781494140625,
|
425 |
+
"eval_loss": 0.1546352356672287,
|
426 |
+
"eval_rewards/accuracies": 0.9500000476837158,
|
427 |
+
"eval_rewards/chosen": 17.06200408935547,
|
428 |
+
"eval_rewards/margins": 4.775761604309082,
|
429 |
+
"eval_rewards/rejected": 12.286243438720703,
|
430 |
+
"eval_runtime": 361.4974,
|
431 |
+
"eval_samples_per_second": 2.766,
|
432 |
+
"eval_steps_per_second": 0.346,
|
433 |
+
"step": 240
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 0.2222222222222222,
|
437 |
+
"grad_norm": 0.25737640261650085,
|
438 |
+
"learning_rate": 4.7772929998339485e-06,
|
439 |
+
"logits/chosen": -0.12348780035972595,
|
440 |
+
"logits/rejected": -0.13704943656921387,
|
441 |
+
"logps/chosen": -4.4299187660217285,
|
442 |
+
"logps/rejected": -53.074607849121094,
|
443 |
+
"loss": 0.1373,
|
444 |
+
"rewards/accuracies": 0.9375,
|
445 |
+
"rewards/chosen": 17.087068557739258,
|
446 |
+
"rewards/margins": 5.06691837310791,
|
447 |
+
"rewards/rejected": 12.020149230957031,
|
448 |
+
"step": 250
|
449 |
+
},
|
450 |
+
{
|
451 |
+
"epoch": 0.2311111111111111,
|
452 |
+
"grad_norm": 0.1839389204978943,
|
453 |
+
"learning_rate": 4.744180489557859e-06,
|
454 |
+
"logits/chosen": -0.12177034467458725,
|
455 |
+
"logits/rejected": -0.1342695653438568,
|
456 |
+
"logps/chosen": -3.775188446044922,
|
457 |
+
"logps/rejected": -53.98720932006836,
|
458 |
+
"loss": 0.1896,
|
459 |
+
"rewards/accuracies": 0.949999988079071,
|
460 |
+
"rewards/chosen": 17.12021255493164,
|
461 |
+
"rewards/margins": 5.148064613342285,
|
462 |
+
"rewards/rejected": 11.972146987915039,
|
463 |
+
"step": 260
|
464 |
+
},
|
465 |
+
{
|
466 |
+
"epoch": 0.24,
|
467 |
+
"grad_norm": 12.258485794067383,
|
468 |
+
"learning_rate": 4.708905451849754e-06,
|
469 |
+
"logits/chosen": -0.11067859083414078,
|
470 |
+
"logits/rejected": -0.12377731502056122,
|
471 |
+
"logps/chosen": -6.418317794799805,
|
472 |
+
"logps/rejected": -56.57402801513672,
|
473 |
+
"loss": 0.2315,
|
474 |
+
"rewards/accuracies": 0.9375,
|
475 |
+
"rewards/chosen": 16.738832473754883,
|
476 |
+
"rewards/margins": 4.884931564331055,
|
477 |
+
"rewards/rejected": 11.853900909423828,
|
478 |
+
"step": 270
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.24888888888888888,
|
482 |
+
"grad_norm": 77.56194305419922,
|
483 |
+
"learning_rate": 4.671501878287879e-06,
|
484 |
+
"logits/chosen": -0.1184445172548294,
|
485 |
+
"logits/rejected": -0.1339874565601349,
|
486 |
+
"logps/chosen": -10.12116527557373,
|
487 |
+
"logps/rejected": -53.403907775878906,
|
488 |
+
"loss": 0.5343,
|
489 |
+
"rewards/accuracies": 0.862500011920929,
|
490 |
+
"rewards/chosen": 16.458633422851562,
|
491 |
+
"rewards/margins": 4.402472496032715,
|
492 |
+
"rewards/rejected": 12.056160926818848,
|
493 |
+
"step": 280
|
494 |
+
},
|
495 |
+
{
|
496 |
+
"epoch": 0.2577777777777778,
|
497 |
+
"grad_norm": 67.53883361816406,
|
498 |
+
"learning_rate": 4.6320058115409295e-06,
|
499 |
+
"logits/chosen": -0.1448262631893158,
|
500 |
+
"logits/rejected": -0.15793387591838837,
|
501 |
+
"logps/chosen": -3.4666190147399902,
|
502 |
+
"logps/rejected": -48.79213333129883,
|
503 |
+
"loss": 0.5017,
|
504 |
+
"rewards/accuracies": 0.887499988079071,
|
505 |
+
"rewards/chosen": 16.945899963378906,
|
506 |
+
"rewards/margins": 4.2686333656311035,
|
507 |
+
"rewards/rejected": 12.677268028259277,
|
508 |
+
"step": 290
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 0.26666666666666666,
|
512 |
+
"grad_norm": 0.17521341145038605,
|
513 |
+
"learning_rate": 4.590455310636778e-06,
|
514 |
+
"logits/chosen": -0.16128253936767578,
|
515 |
+
"logits/rejected": -0.17375555634498596,
|
516 |
+
"logps/chosen": -2.9032950401306152,
|
517 |
+
"logps/rejected": -47.69734191894531,
|
518 |
+
"loss": 0.265,
|
519 |
+
"rewards/accuracies": 0.925000011920929,
|
520 |
+
"rewards/chosen": 17.18383026123047,
|
521 |
+
"rewards/margins": 4.541309356689453,
|
522 |
+
"rewards/rejected": 12.642518997192383,
|
523 |
+
"step": 300
|
524 |
+
},
|
525 |
+
{
|
526 |
+
"epoch": 0.26666666666666666,
|
527 |
+
"eval_logits/chosen": -0.17444846034049988,
|
528 |
+
"eval_logits/rejected": -0.18559777736663818,
|
529 |
+
"eval_logps/chosen": -2.535512924194336,
|
530 |
+
"eval_logps/rejected": -47.16367721557617,
|
531 |
+
"eval_loss": 0.15360687673091888,
|
532 |
+
"eval_rewards/accuracies": 0.9440000653266907,
|
533 |
+
"eval_rewards/chosen": 17.249343872070312,
|
534 |
+
"eval_rewards/margins": 4.604989051818848,
|
535 |
+
"eval_rewards/rejected": 12.644353866577148,
|
536 |
+
"eval_runtime": 361.4575,
|
537 |
+
"eval_samples_per_second": 2.767,
|
538 |
+
"eval_steps_per_second": 0.346,
|
539 |
+
"step": 300
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 0.27555555555555555,
|
543 |
+
"grad_norm": 0.5040452480316162,
|
544 |
+
"learning_rate": 4.54689041428819e-06,
|
545 |
+
"logits/chosen": -0.16974106431007385,
|
546 |
+
"logits/rejected": -0.1810058057308197,
|
547 |
+
"logps/chosen": -1.233938217163086,
|
548 |
+
"logps/rejected": -49.907745361328125,
|
549 |
+
"loss": 0.1132,
|
550 |
+
"rewards/accuracies": 0.9500000476837158,
|
551 |
+
"rewards/chosen": 17.34117889404297,
|
552 |
+
"rewards/margins": 4.934173583984375,
|
553 |
+
"rewards/rejected": 12.407005310058594,
|
554 |
+
"step": 310
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 0.28444444444444444,
|
558 |
+
"grad_norm": 100.02949523925781,
|
559 |
+
"learning_rate": 4.501353102310901e-06,
|
560 |
+
"logits/chosen": -0.15705889463424683,
|
561 |
+
"logits/rejected": -0.1695334017276764,
|
562 |
+
"logps/chosen": -1.0820492506027222,
|
563 |
+
"logps/rejected": -52.577110290527344,
|
564 |
+
"loss": 0.1194,
|
565 |
+
"rewards/accuracies": 0.9500000476837158,
|
566 |
+
"rewards/chosen": 17.33388900756836,
|
567 |
+
"rewards/margins": 5.154760837554932,
|
568 |
+
"rewards/rejected": 12.179126739501953,
|
569 |
+
"step": 320
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.29333333333333333,
|
573 |
+
"grad_norm": 0.2689219117164612,
|
574 |
+
"learning_rate": 4.453887255171206e-06,
|
575 |
+
"logits/chosen": -0.13849371671676636,
|
576 |
+
"logits/rejected": -0.14990833401679993,
|
577 |
+
"logps/chosen": -1.8435032367706299,
|
578 |
+
"logps/rejected": -54.79044723510742,
|
579 |
+
"loss": 0.0926,
|
580 |
+
"rewards/accuracies": 0.9500000476837158,
|
581 |
+
"rewards/chosen": 17.2423095703125,
|
582 |
+
"rewards/margins": 5.28987979888916,
|
583 |
+
"rewards/rejected": 11.952428817749023,
|
584 |
+
"step": 330
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 0.3022222222222222,
|
588 |
+
"grad_norm": 0.09305431693792343,
|
589 |
+
"learning_rate": 4.404538611702055e-06,
|
590 |
+
"logits/chosen": -0.12299702316522598,
|
591 |
+
"logits/rejected": -0.13453055918216705,
|
592 |
+
"logps/chosen": -2.9897143840789795,
|
593 |
+
"logps/rejected": -52.954498291015625,
|
594 |
+
"loss": 0.2873,
|
595 |
+
"rewards/accuracies": 0.925000011920929,
|
596 |
+
"rewards/chosen": 17.17474365234375,
|
597 |
+
"rewards/margins": 5.071004867553711,
|
598 |
+
"rewards/rejected": 12.103739738464355,
|
599 |
+
"step": 340
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 0.3111111111111111,
|
603 |
+
"grad_norm": 59.282073974609375,
|
604 |
+
"learning_rate": 4.3533547250284015e-06,
|
605 |
+
"logits/chosen": -0.11913029849529266,
|
606 |
+
"logits/rejected": -0.12785324454307556,
|
607 |
+
"logps/chosen": -3.9456872940063477,
|
608 |
+
"logps/rejected": -48.68487548828125,
|
609 |
+
"loss": 0.4332,
|
610 |
+
"rewards/accuracies": 0.875,
|
611 |
+
"rewards/chosen": 17.12805938720703,
|
612 |
+
"rewards/margins": 4.669450283050537,
|
613 |
+
"rewards/rejected": 12.458610534667969,
|
614 |
+
"step": 350
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.32,
|
618 |
+
"grad_norm": 0.31101909279823303,
|
619 |
+
"learning_rate": 4.300384916744261e-06,
|
620 |
+
"logits/chosen": -0.11280188709497452,
|
621 |
+
"logits/rejected": -0.12300585210323334,
|
622 |
+
"logps/chosen": -2.1714723110198975,
|
623 |
+
"logps/rejected": -54.74174118041992,
|
624 |
+
"loss": 0.1605,
|
625 |
+
"rewards/accuracies": 0.9500000476837158,
|
626 |
+
"rewards/chosen": 17.326162338256836,
|
627 |
+
"rewards/margins": 5.467062473297119,
|
628 |
+
"rewards/rejected": 11.859098434448242,
|
629 |
+
"step": 360
|
630 |
+
},
|
631 |
+
{
|
632 |
+
"epoch": 0.32,
|
633 |
+
"eval_logits/chosen": -0.10620756447315216,
|
634 |
+
"eval_logits/rejected": -0.11727114766836166,
|
635 |
+
"eval_logps/chosen": -1.4165427684783936,
|
636 |
+
"eval_logps/rejected": -50.9525146484375,
|
637 |
+
"eval_loss": 0.3194349706172943,
|
638 |
+
"eval_rewards/accuracies": 0.9210000038146973,
|
639 |
+
"eval_rewards/chosen": 17.36124038696289,
|
640 |
+
"eval_rewards/margins": 5.095769882202148,
|
641 |
+
"eval_rewards/rejected": 12.26546859741211,
|
642 |
+
"eval_runtime": 361.5072,
|
643 |
+
"eval_samples_per_second": 2.766,
|
644 |
+
"eval_steps_per_second": 0.346,
|
645 |
+
"step": 360
|
646 |
+
},
|
647 |
+
{
|
648 |
+
"epoch": 0.3288888888888889,
|
649 |
+
"grad_norm": 6.1126532554626465,
|
650 |
+
"learning_rate": 4.24568022938566e-06,
|
651 |
+
"logits/chosen": -0.10354311764240265,
|
652 |
+
"logits/rejected": -0.11526636779308319,
|
653 |
+
"logps/chosen": -1.2935255765914917,
|
654 |
+
"logps/rejected": -55.57566833496094,
|
655 |
+
"loss": 0.1711,
|
656 |
+
"rewards/accuracies": 0.9500000476837158,
|
657 |
+
"rewards/chosen": 17.439346313476562,
|
658 |
+
"rewards/margins": 5.700921058654785,
|
659 |
+
"rewards/rejected": 11.738424301147461,
|
660 |
+
"step": 370
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.3377777777777778,
|
664 |
+
"grad_norm": 34.15927505493164,
|
665 |
+
"learning_rate": 4.189293377245241e-06,
|
666 |
+
"logits/chosen": -0.1029932051897049,
|
667 |
+
"logits/rejected": -0.11382515728473663,
|
668 |
+
"logps/chosen": -2.5132687091827393,
|
669 |
+
"logps/rejected": -55.50346374511719,
|
670 |
+
"loss": 0.4359,
|
671 |
+
"rewards/accuracies": 0.8875000476837158,
|
672 |
+
"rewards/chosen": 16.731037139892578,
|
673 |
+
"rewards/margins": 4.368172645568848,
|
674 |
+
"rewards/rejected": 12.362865447998047,
|
675 |
+
"step": 380
|
676 |
+
},
|
677 |
+
{
|
678 |
+
"epoch": 0.3466666666666667,
|
679 |
+
"grad_norm": 2.8422904014587402,
|
680 |
+
"learning_rate": 4.131278695575952e-06,
|
681 |
+
"logits/chosen": -0.10793520510196686,
|
682 |
+
"logits/rejected": -0.12109285593032837,
|
683 |
+
"logps/chosen": -3.014652729034424,
|
684 |
+
"logps/rejected": -53.98411560058594,
|
685 |
+
"loss": 0.2161,
|
686 |
+
"rewards/accuracies": 0.949999988079071,
|
687 |
+
"rewards/chosen": 17.137393951416016,
|
688 |
+
"rewards/margins": 5.105995178222656,
|
689 |
+
"rewards/rejected": 12.03139877319336,
|
690 |
+
"step": 390
|
691 |
+
},
|
692 |
+
{
|
693 |
+
"epoch": 0.35555555555555557,
|
694 |
+
"grad_norm": 54.0329475402832,
|
695 |
+
"learning_rate": 4.071692088232743e-06,
|
696 |
+
"logits/chosen": -0.10393750667572021,
|
697 |
+
"logits/rejected": -0.11834606528282166,
|
698 |
+
"logps/chosen": -2.1508543491363525,
|
699 |
+
"logps/rejected": -45.60733413696289,
|
700 |
+
"loss": 0.2077,
|
701 |
+
"rewards/accuracies": 0.925000011920929,
|
702 |
+
"rewards/chosen": 17.586124420166016,
|
703 |
+
"rewards/margins": 5.077212333679199,
|
704 |
+
"rewards/rejected": 12.5089111328125,
|
705 |
+
"step": 400
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 0.36444444444444446,
|
709 |
+
"grad_norm": 81.61144256591797,
|
710 |
+
"learning_rate": 4.010590973802737e-06,
|
711 |
+
"logits/chosen": -0.09564584493637085,
|
712 |
+
"logits/rejected": -0.10617707669734955,
|
713 |
+
"logps/chosen": -3.4572842121124268,
|
714 |
+
"logps/rejected": -50.92162322998047,
|
715 |
+
"loss": 0.2478,
|
716 |
+
"rewards/accuracies": 0.8875000476837158,
|
717 |
+
"rewards/chosen": 17.010910034179688,
|
718 |
+
"rewards/margins": 4.556198596954346,
|
719 |
+
"rewards/rejected": 12.454713821411133,
|
720 |
+
"step": 410
|
721 |
+
},
|
722 |
+
{
|
723 |
+
"epoch": 0.37333333333333335,
|
724 |
+
"grad_norm": 0.30974289774894714,
|
725 |
+
"learning_rate": 3.948034230275781e-06,
|
726 |
+
"logits/chosen": -0.09134417027235031,
|
727 |
+
"logits/rejected": -0.1020016297698021,
|
728 |
+
"logps/chosen": -5.046698570251465,
|
729 |
+
"logps/rejected": -48.908958435058594,
|
730 |
+
"loss": 0.2894,
|
731 |
+
"rewards/accuracies": 0.8999999761581421,
|
732 |
+
"rewards/chosen": 17.007888793945312,
|
733 |
+
"rewards/margins": 4.53641414642334,
|
734 |
+
"rewards/rejected": 12.471475601196289,
|
735 |
+
"step": 420
|
736 |
+
},
|
737 |
+
{
|
738 |
+
"epoch": 0.37333333333333335,
|
739 |
+
"eval_logits/chosen": -0.09054450690746307,
|
740 |
+
"eval_logits/rejected": -0.10264354199171066,
|
741 |
+
"eval_logps/chosen": -1.913105845451355,
|
742 |
+
"eval_logps/rejected": -51.11127471923828,
|
743 |
+
"eval_loss": 0.16789735853672028,
|
744 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
745 |
+
"eval_rewards/chosen": 17.311582565307617,
|
746 |
+
"eval_rewards/margins": 5.061989784240723,
|
747 |
+
"eval_rewards/rejected": 12.249593734741211,
|
748 |
+
"eval_runtime": 361.5337,
|
749 |
+
"eval_samples_per_second": 2.766,
|
750 |
+
"eval_steps_per_second": 0.346,
|
751 |
+
"step": 420
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.38222222222222224,
|
755 |
+
"grad_norm": 12.824393272399902,
|
756 |
+
"learning_rate": 3.884082138308699e-06,
|
757 |
+
"logits/chosen": -0.08666776865720749,
|
758 |
+
"logits/rejected": -0.0997733399271965,
|
759 |
+
"logps/chosen": -1.7306327819824219,
|
760 |
+
"logps/rejected": -54.273292541503906,
|
761 |
+
"loss": 0.2298,
|
762 |
+
"rewards/accuracies": 0.9500000476837158,
|
763 |
+
"rewards/chosen": 17.167621612548828,
|
764 |
+
"rewards/margins": 5.065673351287842,
|
765 |
+
"rewards/rejected": 12.101947784423828,
|
766 |
+
"step": 430
|
767 |
+
},
|
768 |
+
{
|
769 |
+
"epoch": 0.39111111111111113,
|
770 |
+
"grad_norm": 0.30713599920272827,
|
771 |
+
"learning_rate": 3.818796323137896e-06,
|
772 |
+
"logits/chosen": -0.09174907952547073,
|
773 |
+
"logits/rejected": -0.10376611351966858,
|
774 |
+
"logps/chosen": -1.489154577255249,
|
775 |
+
"logps/rejected": -54.580726623535156,
|
776 |
+
"loss": 0.2513,
|
777 |
+
"rewards/accuracies": 0.9375,
|
778 |
+
"rewards/chosen": 17.22280502319336,
|
779 |
+
"rewards/margins": 5.175349235534668,
|
780 |
+
"rewards/rejected": 12.047454833984375,
|
781 |
+
"step": 440
|
782 |
+
},
|
783 |
+
{
|
784 |
+
"epoch": 0.4,
|
785 |
+
"grad_norm": 87.4791488647461,
|
786 |
+
"learning_rate": 3.7522396951963303e-06,
|
787 |
+
"logits/chosen": -0.09688778221607208,
|
788 |
+
"logits/rejected": -0.10897806286811829,
|
789 |
+
"logps/chosen": -3.157695770263672,
|
790 |
+
"logps/rejected": -50.96417236328125,
|
791 |
+
"loss": 0.1758,
|
792 |
+
"rewards/accuracies": 0.9500000476837158,
|
793 |
+
"rewards/chosen": 17.345651626586914,
|
794 |
+
"rewards/margins": 5.245656967163086,
|
795 |
+
"rewards/rejected": 12.099993705749512,
|
796 |
+
"step": 450
|
797 |
+
},
|
798 |
+
{
|
799 |
+
"epoch": 0.4088888888888889,
|
800 |
+
"grad_norm": 146.2008056640625,
|
801 |
+
"learning_rate": 3.684476389492026e-06,
|
802 |
+
"logits/chosen": -0.09378582239151001,
|
803 |
+
"logits/rejected": -0.10475654900074005,
|
804 |
+
"logps/chosen": -0.5611928701400757,
|
805 |
+
"logps/rejected": -56.518890380859375,
|
806 |
+
"loss": 0.1981,
|
807 |
+
"rewards/accuracies": 0.9500000476837158,
|
808 |
+
"rewards/chosen": 17.113712310791016,
|
809 |
+
"rewards/margins": 5.068872928619385,
|
810 |
+
"rewards/rejected": 12.044839859008789,
|
811 |
+
"step": 460
|
812 |
+
},
|
813 |
+
{
|
814 |
+
"epoch": 0.4177777777777778,
|
815 |
+
"grad_norm": 1.9137721061706543,
|
816 |
+
"learning_rate": 3.6155717038065783e-06,
|
817 |
+
"logits/chosen": -0.08695463836193085,
|
818 |
+
"logits/rejected": -0.09596743434667587,
|
819 |
+
"logps/chosen": -1.5298550128936768,
|
820 |
+
"logps/rejected": -50.27445983886719,
|
821 |
+
"loss": 0.2066,
|
822 |
+
"rewards/accuracies": 0.9375,
|
823 |
+
"rewards/chosen": 17.35186004638672,
|
824 |
+
"rewards/margins": 5.014693260192871,
|
825 |
+
"rewards/rejected": 12.337167739868164,
|
826 |
+
"step": 470
|
827 |
+
},
|
828 |
+
{
|
829 |
+
"epoch": 0.4266666666666667,
|
830 |
+
"grad_norm": 84.80391693115234,
|
831 |
+
"learning_rate": 3.545592035773192e-06,
|
832 |
+
"logits/chosen": -0.0746893435716629,
|
833 |
+
"logits/rejected": -0.08653923869132996,
|
834 |
+
"logps/chosen": -2.0052125453948975,
|
835 |
+
"logps/rejected": -57.502811431884766,
|
836 |
+
"loss": 0.1149,
|
837 |
+
"rewards/accuracies": 0.9500000476837158,
|
838 |
+
"rewards/chosen": 17.14373016357422,
|
839 |
+
"rewards/margins": 5.360415935516357,
|
840 |
+
"rewards/rejected": 11.783313751220703,
|
841 |
+
"step": 480
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 0.4266666666666667,
|
845 |
+
"eval_logits/chosen": -0.07700399309396744,
|
846 |
+
"eval_logits/rejected": -0.08828537166118622,
|
847 |
+
"eval_logps/chosen": -4.48896598815918,
|
848 |
+
"eval_logps/rejected": -53.76282501220703,
|
849 |
+
"eval_loss": 0.29511645436286926,
|
850 |
+
"eval_rewards/accuracies": 0.9230000376701355,
|
851 |
+
"eval_rewards/chosen": 17.053997039794922,
|
852 |
+
"eval_rewards/margins": 5.069558143615723,
|
853 |
+
"eval_rewards/rejected": 11.984437942504883,
|
854 |
+
"eval_runtime": 361.5035,
|
855 |
+
"eval_samples_per_second": 2.766,
|
856 |
+
"eval_steps_per_second": 0.346,
|
857 |
+
"step": 480
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 0.43555555555555553,
|
861 |
+
"grad_norm": 82.9616470336914,
|
862 |
+
"learning_rate": 3.4746048188948806e-06,
|
863 |
+
"logits/chosen": -0.06675051152706146,
|
864 |
+
"logits/rejected": -0.07860895991325378,
|
865 |
+
"logps/chosen": -4.162237167358398,
|
866 |
+
"logps/rejected": -54.77789306640625,
|
867 |
+
"loss": 0.2979,
|
868 |
+
"rewards/accuracies": 0.9125000238418579,
|
869 |
+
"rewards/chosen": 17.047603607177734,
|
870 |
+
"rewards/margins": 5.138361930847168,
|
871 |
+
"rewards/rejected": 11.909242630004883,
|
872 |
+
"step": 490
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.4444444444444444,
|
876 |
+
"grad_norm": 0.04293210059404373,
|
877 |
+
"learning_rate": 3.4026784575644887e-06,
|
878 |
+
"logits/chosen": -0.06424491107463837,
|
879 |
+
"logits/rejected": -0.07567107677459717,
|
880 |
+
"logps/chosen": -2.05729603767395,
|
881 |
+
"logps/rejected": -56.646087646484375,
|
882 |
+
"loss": 0.4378,
|
883 |
+
"rewards/accuracies": 0.8875000476837158,
|
884 |
+
"rewards/chosen": 16.947803497314453,
|
885 |
+
"rewards/margins": 4.919981956481934,
|
886 |
+
"rewards/rejected": 12.02782154083252,
|
887 |
+
"step": 500
|
888 |
+
},
|
889 |
+
{
|
890 |
+
"epoch": 0.4533333333333333,
|
891 |
+
"grad_norm": 0.07136644423007965,
|
892 |
+
"learning_rate": 3.329882261149148e-06,
|
893 |
+
"logits/chosen": -0.06423303484916687,
|
894 |
+
"logits/rejected": -0.07512776553630829,
|
895 |
+
"logps/chosen": -3.1519265174865723,
|
896 |
+
"logps/rejected": -53.53008270263672,
|
897 |
+
"loss": 0.2613,
|
898 |
+
"rewards/accuracies": 0.9375,
|
899 |
+
"rewards/chosen": 17.112144470214844,
|
900 |
+
"rewards/margins": 5.046430587768555,
|
901 |
+
"rewards/rejected": 12.065712928771973,
|
902 |
+
"step": 510
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 0.4622222222222222,
|
906 |
+
"grad_norm": 0.29279613494873047,
|
907 |
+
"learning_rate": 3.25628637720269e-06,
|
908 |
+
"logits/chosen": -0.060233693569898605,
|
909 |
+
"logits/rejected": -0.07076811790466309,
|
910 |
+
"logps/chosen": -1.2358124256134033,
|
911 |
+
"logps/rejected": -50.61806106567383,
|
912 |
+
"loss": 0.1517,
|
913 |
+
"rewards/accuracies": 0.9375,
|
914 |
+
"rewards/chosen": 17.457351684570312,
|
915 |
+
"rewards/margins": 5.2062835693359375,
|
916 |
+
"rewards/rejected": 12.251070022583008,
|
917 |
+
"step": 520
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 0.4711111111111111,
|
921 |
+
"grad_norm": 0.7946074604988098,
|
922 |
+
"learning_rate": 3.181961723870359e-06,
|
923 |
+
"logits/chosen": -0.054482050240039825,
|
924 |
+
"logits/rejected": -0.0657092034816742,
|
925 |
+
"logps/chosen": -0.7682158946990967,
|
926 |
+
"logps/rejected": -58.81409454345703,
|
927 |
+
"loss": 0.2304,
|
928 |
+
"rewards/accuracies": 0.9375,
|
929 |
+
"rewards/chosen": 17.075973510742188,
|
930 |
+
"rewards/margins": 5.222441673278809,
|
931 |
+
"rewards/rejected": 11.853530883789062,
|
932 |
+
"step": 530
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 0.48,
|
936 |
+
"grad_norm": 3.1405210494995117,
|
937 |
+
"learning_rate": 3.1069799215509847e-06,
|
938 |
+
"logits/chosen": -0.05050881579518318,
|
939 |
+
"logits/rejected": -0.061149902641773224,
|
940 |
+
"logps/chosen": -0.954046368598938,
|
941 |
+
"logps/rejected": -52.70227813720703,
|
942 |
+
"loss": 0.0384,
|
943 |
+
"rewards/accuracies": 0.987500011920929,
|
944 |
+
"rewards/chosen": 17.378849029541016,
|
945 |
+
"rewards/margins": 5.253483295440674,
|
946 |
+
"rewards/rejected": 12.125364303588867,
|
947 |
+
"step": 540
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.48,
|
951 |
+
"eval_logits/chosen": -0.05120665580034256,
|
952 |
+
"eval_logits/rejected": -0.06123337894678116,
|
953 |
+
"eval_logps/chosen": -2.9872913360595703,
|
954 |
+
"eval_logps/rejected": -52.27314758300781,
|
955 |
+
"eval_loss": 0.17387841641902924,
|
956 |
+
"eval_rewards/accuracies": 0.9490000605583191,
|
957 |
+
"eval_rewards/chosen": 17.204164505004883,
|
958 |
+
"eval_rewards/margins": 5.070757865905762,
|
959 |
+
"eval_rewards/rejected": 12.133406639099121,
|
960 |
+
"eval_runtime": 361.5449,
|
961 |
+
"eval_samples_per_second": 2.766,
|
962 |
+
"eval_steps_per_second": 0.346,
|
963 |
+
"step": 540
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 0.4888888888888889,
|
967 |
+
"grad_norm": 0.07270358502864838,
|
968 |
+
"learning_rate": 3.0314132238824416e-06,
|
969 |
+
"logits/chosen": -0.05125313252210617,
|
970 |
+
"logits/rejected": -0.06174170970916748,
|
971 |
+
"logps/chosen": -3.2163877487182617,
|
972 |
+
"logps/rejected": -47.79279327392578,
|
973 |
+
"loss": 0.2087,
|
974 |
+
"rewards/accuracies": 0.9125000238418579,
|
975 |
+
"rewards/chosen": 17.408517837524414,
|
976 |
+
"rewards/margins": 5.061524391174316,
|
977 |
+
"rewards/rejected": 12.346992492675781,
|
978 |
+
"step": 550
|
979 |
+
},
|
980 |
+
{
|
981 |
+
"epoch": 0.49777777777777776,
|
982 |
+
"grad_norm": 0.10005924850702286,
|
983 |
+
"learning_rate": 2.955334448116915e-06,
|
984 |
+
"logits/chosen": -0.041773442178964615,
|
985 |
+
"logits/rejected": -0.05364570394158363,
|
986 |
+
"logps/chosen": -0.363404780626297,
|
987 |
+
"logps/rejected": -56.32415008544922,
|
988 |
+
"loss": 0.0969,
|
989 |
+
"rewards/accuracies": 0.9624999761581421,
|
990 |
+
"rewards/chosen": 17.3824462890625,
|
991 |
+
"rewards/margins": 5.5404510498046875,
|
992 |
+
"rewards/rejected": 11.841995239257812,
|
993 |
+
"step": 560
|
994 |
+
},
|
995 |
+
{
|
996 |
+
"epoch": 0.5066666666666667,
|
997 |
+
"grad_norm": 7.818356990814209,
|
998 |
+
"learning_rate": 2.8788169049530533e-06,
|
999 |
+
"logits/chosen": -0.04309462010860443,
|
1000 |
+
"logits/rejected": -0.05494442582130432,
|
1001 |
+
"logps/chosen": -2.2242724895477295,
|
1002 |
+
"logps/rejected": -56.444740295410156,
|
1003 |
+
"loss": 0.1447,
|
1004 |
+
"rewards/accuracies": 0.949999988079071,
|
1005 |
+
"rewards/chosen": 17.275390625,
|
1006 |
+
"rewards/margins": 5.55007791519165,
|
1007 |
+
"rewards/rejected": 11.725313186645508,
|
1008 |
+
"step": 570
|
1009 |
+
},
|
1010 |
+
{
|
1011 |
+
"epoch": 0.5155555555555555,
|
1012 |
+
"grad_norm": 0.03519747406244278,
|
1013 |
+
"learning_rate": 2.8019343278926397e-06,
|
1014 |
+
"logits/chosen": -0.03506368771195412,
|
1015 |
+
"logits/rejected": -0.046854715794324875,
|
1016 |
+
"logps/chosen": -0.5200096964836121,
|
1017 |
+
"logps/rejected": -59.05330276489258,
|
1018 |
+
"loss": 0.0998,
|
1019 |
+
"rewards/accuracies": 0.949999988079071,
|
1020 |
+
"rewards/chosen": 17.320554733276367,
|
1021 |
+
"rewards/margins": 5.727260589599609,
|
1022 |
+
"rewards/rejected": 11.593294143676758,
|
1023 |
+
"step": 580
|
1024 |
+
},
|
1025 |
+
{
|
1026 |
+
"epoch": 0.5244444444444445,
|
1027 |
+
"grad_norm": 0.04519123584032059,
|
1028 |
+
"learning_rate": 2.7247608021898265e-06,
|
1029 |
+
"logits/chosen": -0.03204537555575371,
|
1030 |
+
"logits/rejected": -0.04383891448378563,
|
1031 |
+
"logps/chosen": -1.1271060705184937,
|
1032 |
+
"logps/rejected": -59.308895111083984,
|
1033 |
+
"loss": 0.1162,
|
1034 |
+
"rewards/accuracies": 0.9750000238418579,
|
1035 |
+
"rewards/chosen": 17.256351470947266,
|
1036 |
+
"rewards/margins": 5.6881890296936035,
|
1037 |
+
"rewards/rejected": 11.568161010742188,
|
1038 |
+
"step": 590
|
1039 |
+
},
|
1040 |
+
{
|
1041 |
+
"epoch": 0.5333333333333333,
|
1042 |
+
"grad_norm": 0.028689857572317123,
|
1043 |
+
"learning_rate": 2.647370693461432e-06,
|
1044 |
+
"logits/chosen": -0.02834726870059967,
|
1045 |
+
"logits/rejected": -0.03827046602964401,
|
1046 |
+
"logps/chosen": -5.673943519592285,
|
1047 |
+
"logps/rejected": -55.72624588012695,
|
1048 |
+
"loss": 0.4008,
|
1049 |
+
"rewards/accuracies": 0.8875000476837158,
|
1050 |
+
"rewards/chosen": 16.761056900024414,
|
1051 |
+
"rewards/margins": 4.804043769836426,
|
1052 |
+
"rewards/rejected": 11.957012176513672,
|
1053 |
+
"step": 600
|
1054 |
+
},
|
1055 |
+
{
|
1056 |
+
"epoch": 0.5333333333333333,
|
1057 |
+
"eval_logits/chosen": -0.035840023308992386,
|
1058 |
+
"eval_logits/rejected": -0.04687971621751785,
|
1059 |
+
"eval_logps/chosen": -2.1760435104370117,
|
1060 |
+
"eval_logps/rejected": -56.62664031982422,
|
1061 |
+
"eval_loss": 0.17061151564121246,
|
1062 |
+
"eval_rewards/accuracies": 0.9470000267028809,
|
1063 |
+
"eval_rewards/chosen": 17.285289764404297,
|
1064 |
+
"eval_rewards/margins": 5.587231636047363,
|
1065 |
+
"eval_rewards/rejected": 11.698057174682617,
|
1066 |
+
"eval_runtime": 361.5056,
|
1067 |
+
"eval_samples_per_second": 2.766,
|
1068 |
+
"eval_steps_per_second": 0.346,
|
1069 |
+
"step": 600
|
1070 |
+
},
|
1071 |
+
{
|
1072 |
+
"epoch": 0.5422222222222223,
|
1073 |
+
"grad_norm": 1.1994622945785522,
|
1074 |
+
"learning_rate": 2.569838576027068e-06,
|
1075 |
+
"logits/chosen": -0.03231767192482948,
|
1076 |
+
"logits/rejected": -0.04397805407643318,
|
1077 |
+
"logps/chosen": -1.5904741287231445,
|
1078 |
+
"logps/rejected": -62.10520553588867,
|
1079 |
+
"loss": 0.1265,
|
1080 |
+
"rewards/accuracies": 0.9624999761581421,
|
1081 |
+
"rewards/chosen": 17.11206817626953,
|
1082 |
+
"rewards/margins": 5.734784126281738,
|
1083 |
+
"rewards/rejected": 11.37728500366211,
|
1084 |
+
"step": 610
|
1085 |
+
},
|
1086 |
+
{
|
1087 |
+
"epoch": 0.5511111111111111,
|
1088 |
+
"grad_norm": 0.14163845777511597,
|
1089 |
+
"learning_rate": 2.4922391610481544e-06,
|
1090 |
+
"logits/chosen": -0.03293662518262863,
|
1091 |
+
"logits/rejected": -0.04335154965519905,
|
1092 |
+
"logps/chosen": -1.3568997383117676,
|
1093 |
+
"logps/rejected": -58.347434997558594,
|
1094 |
+
"loss": 0.097,
|
1095 |
+
"rewards/accuracies": 0.9624999761581421,
|
1096 |
+
"rewards/chosen": 17.345399856567383,
|
1097 |
+
"rewards/margins": 5.823373794555664,
|
1098 |
+
"rewards/rejected": 11.522026062011719,
|
1099 |
+
"step": 620
|
1100 |
+
},
|
1101 |
+
{
|
1102 |
+
"epoch": 0.56,
|
1103 |
+
"grad_norm": 1.125027060508728,
|
1104 |
+
"learning_rate": 2.4146472245350804e-06,
|
1105 |
+
"logits/chosen": -0.02864963933825493,
|
1106 |
+
"logits/rejected": -0.03835710883140564,
|
1107 |
+
"logps/chosen": -2.494175434112549,
|
1108 |
+
"logps/rejected": -55.33067321777344,
|
1109 |
+
"loss": 0.3115,
|
1110 |
+
"rewards/accuracies": 0.925000011920929,
|
1111 |
+
"rewards/chosen": 17.262205123901367,
|
1112 |
+
"rewards/margins": 5.4287261962890625,
|
1113 |
+
"rewards/rejected": 11.833479881286621,
|
1114 |
+
"step": 630
|
1115 |
+
},
|
1116 |
+
{
|
1117 |
+
"epoch": 0.5688888888888889,
|
1118 |
+
"grad_norm": 0.03514016419649124,
|
1119 |
+
"learning_rate": 2.337137535291868e-06,
|
1120 |
+
"logits/chosen": -0.02757749892771244,
|
1121 |
+
"logits/rejected": -0.0376611053943634,
|
1122 |
+
"logps/chosen": -2.3163387775421143,
|
1123 |
+
"logps/rejected": -53.54579162597656,
|
1124 |
+
"loss": 0.1268,
|
1125 |
+
"rewards/accuracies": 0.9500000476837158,
|
1126 |
+
"rewards/chosen": 17.542863845825195,
|
1127 |
+
"rewards/margins": 5.799897193908691,
|
1128 |
+
"rewards/rejected": 11.742965698242188,
|
1129 |
+
"step": 640
|
1130 |
+
},
|
1131 |
+
{
|
1132 |
+
"epoch": 0.5777777777777777,
|
1133 |
+
"grad_norm": 34.20791244506836,
|
1134 |
+
"learning_rate": 2.259784782867782e-06,
|
1135 |
+
"logits/chosen": -0.02032250165939331,
|
1136 |
+
"logits/rejected": -0.031542714685201645,
|
1137 |
+
"logps/chosen": -1.7910137176513672,
|
1138 |
+
"logps/rejected": -62.980018615722656,
|
1139 |
+
"loss": 0.1831,
|
1140 |
+
"rewards/accuracies": 0.9500000476837158,
|
1141 |
+
"rewards/chosen": 16.982879638671875,
|
1142 |
+
"rewards/margins": 5.589078903198242,
|
1143 |
+
"rewards/rejected": 11.393800735473633,
|
1144 |
+
"step": 650
|
1145 |
+
},
|
1146 |
+
{
|
1147 |
+
"epoch": 0.5866666666666667,
|
1148 |
+
"grad_norm": 32.356956481933594,
|
1149 |
+
"learning_rate": 2.182663505585314e-06,
|
1150 |
+
"logits/chosen": -0.01711965538561344,
|
1151 |
+
"logits/rejected": -0.028038471937179565,
|
1152 |
+
"logps/chosen": -2.662904739379883,
|
1153 |
+
"logps/rejected": -63.326297760009766,
|
1154 |
+
"loss": 0.1678,
|
1155 |
+
"rewards/accuracies": 0.9624999761581421,
|
1156 |
+
"rewards/chosen": 16.997777938842773,
|
1157 |
+
"rewards/margins": 5.741157531738281,
|
1158 |
+
"rewards/rejected": 11.256620407104492,
|
1159 |
+
"step": 660
|
1160 |
+
},
|
1161 |
+
{
|
1162 |
+
"epoch": 0.5866666666666667,
|
1163 |
+
"eval_logits/chosen": -0.016018809750676155,
|
1164 |
+
"eval_logits/rejected": -0.027020033448934555,
|
1165 |
+
"eval_logps/chosen": -3.00822114944458,
|
1166 |
+
"eval_logps/rejected": -57.951629638671875,
|
1167 |
+
"eval_loss": 0.2050127536058426,
|
1168 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1169 |
+
"eval_rewards/chosen": 17.202072143554688,
|
1170 |
+
"eval_rewards/margins": 5.636512756347656,
|
1171 |
+
"eval_rewards/rejected": 11.565557479858398,
|
1172 |
+
"eval_runtime": 361.5073,
|
1173 |
+
"eval_samples_per_second": 2.766,
|
1174 |
+
"eval_steps_per_second": 0.346,
|
1175 |
+
"step": 660
|
1176 |
+
},
|
1177 |
+
{
|
1178 |
+
"epoch": 0.5955555555555555,
|
1179 |
+
"grad_norm": 86.53874969482422,
|
1180 |
+
"learning_rate": 2.1058480187138863e-06,
|
1181 |
+
"logits/chosen": -0.012352555990219116,
|
1182 |
+
"logits/rejected": -0.023966707289218903,
|
1183 |
+
"logps/chosen": -2.129209041595459,
|
1184 |
+
"logps/rejected": -62.1393928527832,
|
1185 |
+
"loss": 0.1025,
|
1186 |
+
"rewards/accuracies": 0.9624999761581421,
|
1187 |
+
"rewards/chosen": 17.069538116455078,
|
1188 |
+
"rewards/margins": 5.678930759429932,
|
1189 |
+
"rewards/rejected": 11.390605926513672,
|
1190 |
+
"step": 670
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 0.6044444444444445,
|
1194 |
+
"grad_norm": 2.892672300338745,
|
1195 |
+
"learning_rate": 2.0294123428584985e-06,
|
1196 |
+
"logits/chosen": -0.011562807485461235,
|
1197 |
+
"logits/rejected": -0.020859256386756897,
|
1198 |
+
"logps/chosen": -3.2486608028411865,
|
1199 |
+
"logps/rejected": -59.56721496582031,
|
1200 |
+
"loss": 0.1961,
|
1201 |
+
"rewards/accuracies": 0.9500000476837158,
|
1202 |
+
"rewards/chosen": 17.051376342773438,
|
1203 |
+
"rewards/margins": 5.515361785888672,
|
1204 |
+
"rewards/rejected": 11.536016464233398,
|
1205 |
+
"step": 680
|
1206 |
+
},
|
1207 |
+
{
|
1208 |
+
"epoch": 0.6133333333333333,
|
1209 |
+
"grad_norm": 30.26588249206543,
|
1210 |
+
"learning_rate": 1.953430132632311e-06,
|
1211 |
+
"logits/chosen": -0.011488726362586021,
|
1212 |
+
"logits/rejected": -0.021527227014303207,
|
1213 |
+
"logps/chosen": -1.794357180595398,
|
1214 |
+
"logps/rejected": -60.649505615234375,
|
1215 |
+
"loss": 0.2821,
|
1216 |
+
"rewards/accuracies": 0.9375,
|
1217 |
+
"rewards/chosen": 17.114843368530273,
|
1218 |
+
"rewards/margins": 5.612217903137207,
|
1219 |
+
"rewards/rejected": 11.50262451171875,
|
1220 |
+
"step": 690
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 0.6222222222222222,
|
1224 |
+
"grad_norm": 0.030314341187477112,
|
1225 |
+
"learning_rate": 1.8779746056819104e-06,
|
1226 |
+
"logits/chosen": -0.014436552301049232,
|
1227 |
+
"logits/rejected": -0.026044374331831932,
|
1228 |
+
"logps/chosen": -3.1617178916931152,
|
1229 |
+
"logps/rejected": -60.94548797607422,
|
1230 |
+
"loss": 0.3141,
|
1231 |
+
"rewards/accuracies": 0.9375,
|
1232 |
+
"rewards/chosen": 17.08698272705078,
|
1233 |
+
"rewards/margins": 5.719264984130859,
|
1234 |
+
"rewards/rejected": 11.367716789245605,
|
1235 |
+
"step": 700
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 0.6311111111111111,
|
1239 |
+
"grad_norm": 0.06711317598819733,
|
1240 |
+
"learning_rate": 1.8031184721336364e-06,
|
1241 |
+
"logits/chosen": -0.017312290146946907,
|
1242 |
+
"logits/rejected": -0.026554957032203674,
|
1243 |
+
"logps/chosen": -5.120705604553223,
|
1244 |
+
"logps/rejected": -54.33483123779297,
|
1245 |
+
"loss": 0.3827,
|
1246 |
+
"rewards/accuracies": 0.887499988079071,
|
1247 |
+
"rewards/chosen": 17.1247501373291,
|
1248 |
+
"rewards/margins": 5.327882766723633,
|
1249 |
+
"rewards/rejected": 11.796867370605469,
|
1250 |
+
"step": 710
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 0.64,
|
1254 |
+
"grad_norm": 60.36637496948242,
|
1255 |
+
"learning_rate": 1.7289338645289711e-06,
|
1256 |
+
"logits/chosen": -0.01987219974398613,
|
1257 |
+
"logits/rejected": -0.02766304835677147,
|
1258 |
+
"logps/chosen": -2.1757986545562744,
|
1259 |
+
"logps/rejected": -54.02531433105469,
|
1260 |
+
"loss": 0.2272,
|
1261 |
+
"rewards/accuracies": 0.949999988079071,
|
1262 |
+
"rewards/chosen": 17.239776611328125,
|
1263 |
+
"rewards/margins": 5.250313758850098,
|
1264 |
+
"rewards/rejected": 11.989462852478027,
|
1265 |
+
"step": 720
|
1266 |
+
},
|
1267 |
+
{
|
1268 |
+
"epoch": 0.64,
|
1269 |
+
"eval_logits/chosen": -0.022905193269252777,
|
1270 |
+
"eval_logits/rejected": -0.0321992002427578,
|
1271 |
+
"eval_logps/chosen": -1.100506067276001,
|
1272 |
+
"eval_logps/rejected": -55.91169738769531,
|
1273 |
+
"eval_loss": 0.14018221199512482,
|
1274 |
+
"eval_rewards/accuracies": 0.9520000219345093,
|
1275 |
+
"eval_rewards/chosen": 17.392841339111328,
|
1276 |
+
"eval_rewards/margins": 5.623291492462158,
|
1277 |
+
"eval_rewards/rejected": 11.769551277160645,
|
1278 |
+
"eval_runtime": 361.416,
|
1279 |
+
"eval_samples_per_second": 2.767,
|
1280 |
+
"eval_steps_per_second": 0.346,
|
1281 |
+
"step": 720
|
1282 |
+
},
|
1283 |
+
{
|
1284 |
+
"epoch": 0.6488888888888888,
|
1285 |
+
"grad_norm": 0.029294608160853386,
|
1286 |
+
"learning_rate": 1.6554922683164875e-06,
|
1287 |
+
"logits/chosen": -0.016416028141975403,
|
1288 |
+
"logits/rejected": -0.025527067482471466,
|
1289 |
+
"logps/chosen": -0.906692385673523,
|
1290 |
+
"logps/rejected": -59.142173767089844,
|
1291 |
+
"loss": 0.1171,
|
1292 |
+
"rewards/accuracies": 0.9624999761581421,
|
1293 |
+
"rewards/chosen": 17.282169342041016,
|
1294 |
+
"rewards/margins": 5.6978583335876465,
|
1295 |
+
"rewards/rejected": 11.584310531616211,
|
1296 |
+
"step": 730
|
1297 |
+
},
|
1298 |
+
{
|
1299 |
+
"epoch": 0.6577777777777778,
|
1300 |
+
"grad_norm": 33.89070510864258,
|
1301 |
+
"learning_rate": 1.5828644529673592e-06,
|
1302 |
+
"logits/chosen": -0.016044551506638527,
|
1303 |
+
"logits/rejected": -0.025811903178691864,
|
1304 |
+
"logps/chosen": -1.1167538166046143,
|
1305 |
+
"logps/rejected": -59.84492492675781,
|
1306 |
+
"loss": 0.2001,
|
1307 |
+
"rewards/accuracies": 0.9624999761581421,
|
1308 |
+
"rewards/chosen": 17.266983032226562,
|
1309 |
+
"rewards/margins": 5.749438285827637,
|
1310 |
+
"rewards/rejected": 11.51754379272461,
|
1311 |
+
"step": 740
|
1312 |
+
},
|
1313 |
+
{
|
1314 |
+
"epoch": 0.6666666666666666,
|
1315 |
+
"grad_norm": 0.53726726770401,
|
1316 |
+
"learning_rate": 1.5111204037807844e-06,
|
1317 |
+
"logits/chosen": -0.012529855594038963,
|
1318 |
+
"logits/rejected": -0.023765765130519867,
|
1319 |
+
"logps/chosen": -0.3461765646934509,
|
1320 |
+
"logps/rejected": -55.841102600097656,
|
1321 |
+
"loss": 0.1692,
|
1322 |
+
"rewards/accuracies": 0.9375,
|
1323 |
+
"rewards/chosen": 17.61865997314453,
|
1324 |
+
"rewards/margins": 6.019055366516113,
|
1325 |
+
"rewards/rejected": 11.599604606628418,
|
1326 |
+
"step": 750
|
1327 |
+
},
|
1328 |
+
{
|
1329 |
+
"epoch": 0.6755555555555556,
|
1330 |
+
"grad_norm": 0.7173987627029419,
|
1331 |
+
"learning_rate": 1.4403292544450625e-06,
|
1332 |
+
"logits/chosen": -0.013380522839725018,
|
1333 |
+
"logits/rejected": -0.022107835859060287,
|
1334 |
+
"logps/chosen": -3.2759666442871094,
|
1335 |
+
"logps/rejected": -53.836822509765625,
|
1336 |
+
"loss": 0.2418,
|
1337 |
+
"rewards/accuracies": 0.949999988079071,
|
1338 |
+
"rewards/chosen": 17.383333206176758,
|
1339 |
+
"rewards/margins": 5.590358734130859,
|
1340 |
+
"rewards/rejected": 11.792974472045898,
|
1341 |
+
"step": 760
|
1342 |
+
},
|
1343 |
+
{
|
1344 |
+
"epoch": 0.6844444444444444,
|
1345 |
+
"grad_norm": 2.744900941848755,
|
1346 |
+
"learning_rate": 1.3705592204192853e-06,
|
1347 |
+
"logits/chosen": -0.011318420059978962,
|
1348 |
+
"logits/rejected": -0.021840626373887062,
|
1349 |
+
"logps/chosen": -2.592001438140869,
|
1350 |
+
"logps/rejected": -57.06926345825195,
|
1351 |
+
"loss": 0.1749,
|
1352 |
+
"rewards/accuracies": 0.9500000476837158,
|
1353 |
+
"rewards/chosen": 17.286128997802734,
|
1354 |
+
"rewards/margins": 5.662715911865234,
|
1355 |
+
"rewards/rejected": 11.623414993286133,
|
1356 |
+
"step": 770
|
1357 |
+
},
|
1358 |
+
{
|
1359 |
+
"epoch": 0.6933333333333334,
|
1360 |
+
"grad_norm": 70.76551818847656,
|
1361 |
+
"learning_rate": 1.301877533199859e-06,
|
1362 |
+
"logits/chosen": -0.013944407925009727,
|
1363 |
+
"logits/rejected": -0.025667501613497734,
|
1364 |
+
"logps/chosen": -2.1070234775543213,
|
1365 |
+
"logps/rejected": -57.72039794921875,
|
1366 |
+
"loss": 0.1915,
|
1367 |
+
"rewards/accuracies": 0.925000011920929,
|
1368 |
+
"rewards/chosen": 17.28545379638672,
|
1369 |
+
"rewards/margins": 5.708344459533691,
|
1370 |
+
"rewards/rejected": 11.577108383178711,
|
1371 |
+
"step": 780
|
1372 |
+
},
|
1373 |
+
{
|
1374 |
+
"epoch": 0.6933333333333334,
|
1375 |
+
"eval_logits/chosen": -0.01664295792579651,
|
1376 |
+
"eval_logits/rejected": -0.026626665145158768,
|
1377 |
+
"eval_logps/chosen": -1.082255244255066,
|
1378 |
+
"eval_logps/rejected": -55.95073699951172,
|
1379 |
+
"eval_loss": 0.24412688612937927,
|
1380 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
1381 |
+
"eval_rewards/chosen": 17.394668579101562,
|
1382 |
+
"eval_rewards/margins": 5.629020690917969,
|
1383 |
+
"eval_rewards/rejected": 11.765647888183594,
|
1384 |
+
"eval_runtime": 361.4602,
|
1385 |
+
"eval_samples_per_second": 2.767,
|
1386 |
+
"eval_steps_per_second": 0.346,
|
1387 |
+
"step": 780
|
1388 |
+
},
|
1389 |
+
{
|
1390 |
+
"epoch": 0.7022222222222222,
|
1391 |
+
"grad_norm": 0.41359376907348633,
|
1392 |
+
"learning_rate": 1.2343503755351729e-06,
|
1393 |
+
"logits/chosen": -0.012772129848599434,
|
1394 |
+
"logits/rejected": -0.023186586797237396,
|
1395 |
+
"logps/chosen": -0.8030359148979187,
|
1396 |
+
"logps/rejected": -57.84947967529297,
|
1397 |
+
"loss": 0.2143,
|
1398 |
+
"rewards/accuracies": 0.9375,
|
1399 |
+
"rewards/chosen": 17.32242202758789,
|
1400 |
+
"rewards/margins": 5.645486831665039,
|
1401 |
+
"rewards/rejected": 11.676933288574219,
|
1402 |
+
"step": 790
|
1403 |
+
},
|
1404 |
+
{
|
1405 |
+
"epoch": 0.7111111111111111,
|
1406 |
+
"grad_norm": 0.41748157143592834,
|
1407 |
+
"learning_rate": 1.168042817650881e-06,
|
1408 |
+
"logits/chosen": -0.012502101249992847,
|
1409 |
+
"logits/rejected": -0.023272844031453133,
|
1410 |
+
"logps/chosen": -1.0377256870269775,
|
1411 |
+
"logps/rejected": -57.05615997314453,
|
1412 |
+
"loss": 0.0792,
|
1413 |
+
"rewards/accuracies": 0.9750000238418579,
|
1414 |
+
"rewards/chosen": 17.60501480102539,
|
1415 |
+
"rewards/margins": 6.159370422363281,
|
1416 |
+
"rewards/rejected": 11.445646286010742,
|
1417 |
+
"step": 800
|
1418 |
+
},
|
1419 |
+
{
|
1420 |
+
"epoch": 0.72,
|
1421 |
+
"grad_norm": 0.039210401475429535,
|
1422 |
+
"learning_rate": 1.1030187545472012e-06,
|
1423 |
+
"logits/chosen": -0.008177272044122219,
|
1424 |
+
"logits/rejected": -0.01832464337348938,
|
1425 |
+
"logps/chosen": -3.8701748847961426,
|
1426 |
+
"logps/rejected": -57.661476135253906,
|
1427 |
+
"loss": 0.3177,
|
1428 |
+
"rewards/accuracies": 0.925000011920929,
|
1429 |
+
"rewards/chosen": 17.14336395263672,
|
1430 |
+
"rewards/margins": 5.583393573760986,
|
1431 |
+
"rewards/rejected": 11.55997085571289,
|
1432 |
+
"step": 810
|
1433 |
+
},
|
1434 |
+
{
|
1435 |
+
"epoch": 0.7288888888888889,
|
1436 |
+
"grad_norm": 1.6096951961517334,
|
1437 |
+
"learning_rate": 1.0393408444287048e-06,
|
1438 |
+
"logits/chosen": -0.006832236424088478,
|
1439 |
+
"logits/rejected": -0.01682097464799881,
|
1440 |
+
"logps/chosen": -2.1741790771484375,
|
1441 |
+
"logps/rejected": -57.273414611816406,
|
1442 |
+
"loss": 0.2128,
|
1443 |
+
"rewards/accuracies": 0.949999988079071,
|
1444 |
+
"rewards/chosen": 17.426712036132812,
|
1445 |
+
"rewards/margins": 5.911205291748047,
|
1446 |
+
"rewards/rejected": 11.515506744384766,
|
1447 |
+
"step": 820
|
1448 |
+
},
|
1449 |
+
{
|
1450 |
+
"epoch": 0.7377777777777778,
|
1451 |
+
"grad_norm": 3.7820959091186523,
|
1452 |
+
"learning_rate": 9.770704483258782e-07,
|
1453 |
+
"logits/chosen": -0.009998206980526447,
|
1454 |
+
"logits/rejected": -0.0204261876642704,
|
1455 |
+
"logps/chosen": -1.9802953004837036,
|
1456 |
+
"logps/rejected": -56.901512145996094,
|
1457 |
+
"loss": 0.1296,
|
1458 |
+
"rewards/accuracies": 0.9750000238418579,
|
1459 |
+
"rewards/chosen": 17.56071662902832,
|
1460 |
+
"rewards/margins": 6.169583320617676,
|
1461 |
+
"rewards/rejected": 11.391134262084961,
|
1462 |
+
"step": 830
|
1463 |
+
},
|
1464 |
+
{
|
1465 |
+
"epoch": 0.7466666666666667,
|
1466 |
+
"grad_norm": 1.126626968383789,
|
1467 |
+
"learning_rate": 9.162675709666865e-07,
|
1468 |
+
"logits/chosen": -0.00826224498450756,
|
1469 |
+
"logits/rejected": -0.018977787345647812,
|
1470 |
+
"logps/chosen": -1.4256607294082642,
|
1471 |
+
"logps/rejected": -61.65986633300781,
|
1472 |
+
"loss": 0.0635,
|
1473 |
+
"rewards/accuracies": 0.987500011920929,
|
1474 |
+
"rewards/chosen": 17.324975967407227,
|
1475 |
+
"rewards/margins": 6.09440803527832,
|
1476 |
+
"rewards/rejected": 11.230567932128906,
|
1477 |
+
"step": 840
|
1478 |
+
},
|
1479 |
+
{
|
1480 |
+
"epoch": 0.7466666666666667,
|
1481 |
+
"eval_logits/chosen": -0.01111944392323494,
|
1482 |
+
"eval_logits/rejected": -0.021697774529457092,
|
1483 |
+
"eval_logps/chosen": -1.2168633937835693,
|
1484 |
+
"eval_logps/rejected": -58.2642822265625,
|
1485 |
+
"eval_loss": 0.1689341962337494,
|
1486 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1487 |
+
"eval_rewards/chosen": 17.381206512451172,
|
1488 |
+
"eval_rewards/margins": 5.846914291381836,
|
1489 |
+
"eval_rewards/rejected": 11.534292221069336,
|
1490 |
+
"eval_runtime": 361.6192,
|
1491 |
+
"eval_samples_per_second": 2.765,
|
1492 |
+
"eval_steps_per_second": 0.346,
|
1493 |
+
"step": 840
|
1494 |
+
},
|
1495 |
+
{
|
1496 |
+
"epoch": 0.7555555555555555,
|
1497 |
+
"grad_norm": 5.250723838806152,
|
1498 |
+
"learning_rate": 8.569908029550686e-07,
|
1499 |
+
"logits/chosen": -0.006854387000203133,
|
1500 |
+
"logits/rejected": -0.018336206674575806,
|
1501 |
+
"logps/chosen": -0.6238930821418762,
|
1502 |
+
"logps/rejected": -60.925689697265625,
|
1503 |
+
"loss": 0.1157,
|
1504 |
+
"rewards/accuracies": 0.9750000238418579,
|
1505 |
+
"rewards/chosen": 17.443281173706055,
|
1506 |
+
"rewards/margins": 6.164813995361328,
|
1507 |
+
"rewards/rejected": 11.278467178344727,
|
1508 |
+
"step": 850
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 0.7644444444444445,
|
1512 |
+
"grad_norm": 3.1401162147521973,
|
1513 |
+
"learning_rate": 7.992972643121227e-07,
|
1514 |
+
"logits/chosen": -0.0037835021503269672,
|
1515 |
+
"logits/rejected": -0.013135241344571114,
|
1516 |
+
"logps/chosen": -0.8492221832275391,
|
1517 |
+
"logps/rejected": -55.516075134277344,
|
1518 |
+
"loss": 0.2252,
|
1519 |
+
"rewards/accuracies": 0.9375,
|
1520 |
+
"rewards/chosen": 17.488588333129883,
|
1521 |
+
"rewards/margins": 5.736725807189941,
|
1522 |
+
"rewards/rejected": 11.751862525939941,
|
1523 |
+
"step": 860
|
1524 |
+
},
|
1525 |
+
{
|
1526 |
+
"epoch": 0.7733333333333333,
|
1527 |
+
"grad_norm": 55.528812408447266,
|
1528 |
+
"learning_rate": 7.432425494343509e-07,
|
1529 |
+
"logits/chosen": -0.0033687639515846968,
|
1530 |
+
"logits/rejected": -0.013152632862329483,
|
1531 |
+
"logps/chosen": -1.3188884258270264,
|
1532 |
+
"logps/rejected": -57.9510498046875,
|
1533 |
+
"loss": 0.1398,
|
1534 |
+
"rewards/accuracies": 0.9624999761581421,
|
1535 |
+
"rewards/chosen": 17.412578582763672,
|
1536 |
+
"rewards/margins": 5.868515968322754,
|
1537 |
+
"rewards/rejected": 11.544061660766602,
|
1538 |
+
"step": 870
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 0.7822222222222223,
|
1542 |
+
"grad_norm": 0.039824869483709335,
|
1543 |
+
"learning_rate": 6.888806735220396e-07,
|
1544 |
+
"logits/chosen": -0.0010406378423795104,
|
1545 |
+
"logits/rejected": -0.012095071375370026,
|
1546 |
+
"logps/chosen": -2.0619027614593506,
|
1547 |
+
"logps/rejected": -59.65806579589844,
|
1548 |
+
"loss": 0.2966,
|
1549 |
+
"rewards/accuracies": 0.9375,
|
1550 |
+
"rewards/chosen": 17.192535400390625,
|
1551 |
+
"rewards/margins": 5.7006731033325195,
|
1552 |
+
"rewards/rejected": 11.491861343383789,
|
1553 |
+
"step": 880
|
1554 |
+
},
|
1555 |
+
{
|
1556 |
+
"epoch": 0.7911111111111111,
|
1557 |
+
"grad_norm": 0.7536466717720032,
|
1558 |
+
"learning_rate": 6.362640205293583e-07,
|
1559 |
+
"logits/chosen": -0.0016857212176546454,
|
1560 |
+
"logits/rejected": -0.010936147533357143,
|
1561 |
+
"logps/chosen": -2.1478958129882812,
|
1562 |
+
"logps/rejected": -58.20386505126953,
|
1563 |
+
"loss": 0.1641,
|
1564 |
+
"rewards/accuracies": 0.9375,
|
1565 |
+
"rewards/chosen": 17.243247985839844,
|
1566 |
+
"rewards/margins": 5.678771018981934,
|
1567 |
+
"rewards/rejected": 11.564477920532227,
|
1568 |
+
"step": 890
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 0.8,
|
1572 |
+
"grad_norm": 0.3857377767562866,
|
1573 |
+
"learning_rate": 5.854432926863684e-07,
|
1574 |
+
"logits/chosen": 0.00038508616853505373,
|
1575 |
+
"logits/rejected": -0.011034643277525902,
|
1576 |
+
"logps/chosen": -1.606274127960205,
|
1577 |
+
"logps/rejected": -62.58662414550781,
|
1578 |
+
"loss": 0.1703,
|
1579 |
+
"rewards/accuracies": 0.9624999761581421,
|
1580 |
+
"rewards/chosen": 17.211261749267578,
|
1581 |
+
"rewards/margins": 5.966868877410889,
|
1582 |
+
"rewards/rejected": 11.244392395019531,
|
1583 |
+
"step": 900
|
1584 |
+
},
|
1585 |
+
{
|
1586 |
+
"epoch": 0.8,
|
1587 |
+
"eval_logits/chosen": 0.00021816430671606213,
|
1588 |
+
"eval_logits/rejected": -0.010477552190423012,
|
1589 |
+
"eval_logps/chosen": -1.7576563358306885,
|
1590 |
+
"eval_logps/rejected": -59.79063415527344,
|
1591 |
+
"eval_loss": 0.13995186984539032,
|
1592 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1593 |
+
"eval_rewards/chosen": 17.327129364013672,
|
1594 |
+
"eval_rewards/margins": 5.945469856262207,
|
1595 |
+
"eval_rewards/rejected": 11.381658554077148,
|
1596 |
+
"eval_runtime": 361.4592,
|
1597 |
+
"eval_samples_per_second": 2.767,
|
1598 |
+
"eval_steps_per_second": 0.346,
|
1599 |
+
"step": 900
|
1600 |
+
},
|
1601 |
+
{
|
1602 |
+
"epoch": 0.8088888888888889,
|
1603 |
+
"grad_norm": 0.13143697381019592,
|
1604 |
+
"learning_rate": 5.364674616415547e-07,
|
1605 |
+
"logits/chosen": 0.0005570838693529367,
|
1606 |
+
"logits/rejected": -0.011198626831173897,
|
1607 |
+
"logps/chosen": -0.12791283428668976,
|
1608 |
+
"logps/rejected": -65.7052993774414,
|
1609 |
+
"loss": 0.0507,
|
1610 |
+
"rewards/accuracies": 0.9750000238418579,
|
1611 |
+
"rewards/chosen": 17.224443435668945,
|
1612 |
+
"rewards/margins": 6.163690567016602,
|
1613 |
+
"rewards/rejected": 11.060752868652344,
|
1614 |
+
"step": 910
|
1615 |
+
},
|
1616 |
+
{
|
1617 |
+
"epoch": 0.8177777777777778,
|
1618 |
+
"grad_norm": 102.9654541015625,
|
1619 |
+
"learning_rate": 4.893837212719859e-07,
|
1620 |
+
"logits/chosen": -0.0008557128603570163,
|
1621 |
+
"logits/rejected": -0.01163212489336729,
|
1622 |
+
"logps/chosen": -1.3292646408081055,
|
1623 |
+
"logps/rejected": -61.644893646240234,
|
1624 |
+
"loss": 0.0881,
|
1625 |
+
"rewards/accuracies": 0.9500000476837158,
|
1626 |
+
"rewards/chosen": 17.276538848876953,
|
1627 |
+
"rewards/margins": 5.96080207824707,
|
1628 |
+
"rewards/rejected": 11.3157377243042,
|
1629 |
+
"step": 920
|
1630 |
+
},
|
1631 |
+
{
|
1632 |
+
"epoch": 0.8266666666666667,
|
1633 |
+
"grad_norm": 146.04498291015625,
|
1634 |
+
"learning_rate": 4.442374422065493e-07,
|
1635 |
+
"logits/chosen": 0.002922601066529751,
|
1636 |
+
"logits/rejected": -0.007130052894353867,
|
1637 |
+
"logps/chosen": -1.6069023609161377,
|
1638 |
+
"logps/rejected": -61.52588653564453,
|
1639 |
+
"loss": 0.1373,
|
1640 |
+
"rewards/accuracies": 0.949999988079071,
|
1641 |
+
"rewards/chosen": 17.184974670410156,
|
1642 |
+
"rewards/margins": 5.7995734214782715,
|
1643 |
+
"rewards/rejected": 11.38540267944336,
|
1644 |
+
"step": 930
|
1645 |
+
},
|
1646 |
+
{
|
1647 |
+
"epoch": 0.8355555555555556,
|
1648 |
+
"grad_norm": 23.539485931396484,
|
1649 |
+
"learning_rate": 4.0107212810610974e-07,
|
1650 |
+
"logits/chosen": 0.0018056132830679417,
|
1651 |
+
"logits/rejected": -0.007847340777516365,
|
1652 |
+
"logps/chosen": -3.637047290802002,
|
1653 |
+
"logps/rejected": -61.21245574951172,
|
1654 |
+
"loss": 0.2763,
|
1655 |
+
"rewards/accuracies": 0.925000011920929,
|
1656 |
+
"rewards/chosen": 16.960758209228516,
|
1657 |
+
"rewards/margins": 5.551811695098877,
|
1658 |
+
"rewards/rejected": 11.40894603729248,
|
1659 |
+
"step": 940
|
1660 |
+
},
|
1661 |
+
{
|
1662 |
+
"epoch": 0.8444444444444444,
|
1663 |
+
"grad_norm": 31.52926254272461,
|
1664 |
+
"learning_rate": 3.599293737426932e-07,
|
1665 |
+
"logits/chosen": 0.0032081177923828363,
|
1666 |
+
"logits/rejected": -0.007756482809782028,
|
1667 |
+
"logps/chosen": -1.172515869140625,
|
1668 |
+
"logps/rejected": -66.5853271484375,
|
1669 |
+
"loss": 0.1758,
|
1670 |
+
"rewards/accuracies": 0.949999988079071,
|
1671 |
+
"rewards/chosen": 16.992876052856445,
|
1672 |
+
"rewards/margins": 5.886469841003418,
|
1673 |
+
"rewards/rejected": 11.106407165527344,
|
1674 |
+
"step": 950
|
1675 |
+
},
|
1676 |
+
{
|
1677 |
+
"epoch": 0.8533333333333334,
|
1678 |
+
"grad_norm": 0.07474468648433685,
|
1679 |
+
"learning_rate": 3.208488249181216e-07,
|
1680 |
+
"logits/chosen": 0.0022508346009999514,
|
1681 |
+
"logits/rejected": -0.009156409651041031,
|
1682 |
+
"logps/chosen": -0.7930470108985901,
|
1683 |
+
"logps/rejected": -60.068790435791016,
|
1684 |
+
"loss": 0.1138,
|
1685 |
+
"rewards/accuracies": 0.9624999761581421,
|
1686 |
+
"rewards/chosen": 17.401771545410156,
|
1687 |
+
"rewards/margins": 6.0026960372924805,
|
1688 |
+
"rewards/rejected": 11.399076461791992,
|
1689 |
+
"step": 960
|
1690 |
+
},
|
1691 |
+
{
|
1692 |
+
"epoch": 0.8533333333333334,
|
1693 |
+
"eval_logits/chosen": 0.001469604205340147,
|
1694 |
+
"eval_logits/rejected": -0.009397665038704872,
|
1695 |
+
"eval_logps/chosen": -1.8795456886291504,
|
1696 |
+
"eval_logps/rejected": -60.17564010620117,
|
1697 |
+
"eval_loss": 0.1441129744052887,
|
1698 |
+
"eval_rewards/accuracies": 0.9630000591278076,
|
1699 |
+
"eval_rewards/chosen": 17.314937591552734,
|
1700 |
+
"eval_rewards/margins": 5.9717817306518555,
|
1701 |
+
"eval_rewards/rejected": 11.343156814575195,
|
1702 |
+
"eval_runtime": 361.5344,
|
1703 |
+
"eval_samples_per_second": 2.766,
|
1704 |
+
"eval_steps_per_second": 0.346,
|
1705 |
+
"step": 960
|
1706 |
+
},
|
1707 |
+
{
|
1708 |
+
"epoch": 0.8622222222222222,
|
1709 |
+
"grad_norm": 68.90747833251953,
|
1710 |
+
"learning_rate": 2.838681402606952e-07,
|
1711 |
+
"logits/chosen": 0.004552370868623257,
|
1712 |
+
"logits/rejected": -0.005488495342433453,
|
1713 |
+
"logps/chosen": -3.7298974990844727,
|
1714 |
+
"logps/rejected": -64.72488403320312,
|
1715 |
+
"loss": 0.2425,
|
1716 |
+
"rewards/accuracies": 0.925000011920929,
|
1717 |
+
"rewards/chosen": 16.832754135131836,
|
1718 |
+
"rewards/margins": 5.644216537475586,
|
1719 |
+
"rewards/rejected": 11.18853759765625,
|
1720 |
+
"step": 970
|
1721 |
+
},
|
1722 |
+
{
|
1723 |
+
"epoch": 0.8711111111111111,
|
1724 |
+
"grad_norm": 0.028206102550029755,
|
1725 |
+
"learning_rate": 2.490229549367443e-07,
|
1726 |
+
"logits/chosen": 0.0025807656347751617,
|
1727 |
+
"logits/rejected": -0.008657123893499374,
|
1728 |
+
"logps/chosen": -1.1825838088989258,
|
1729 |
+
"logps/rejected": -63.57493591308594,
|
1730 |
+
"loss": 0.0609,
|
1731 |
+
"rewards/accuracies": 0.9750000238418579,
|
1732 |
+
"rewards/chosen": 17.227130889892578,
|
1733 |
+
"rewards/margins": 6.053717136383057,
|
1734 |
+
"rewards/rejected": 11.173412322998047,
|
1735 |
+
"step": 980
|
1736 |
+
},
|
1737 |
+
{
|
1738 |
+
"epoch": 0.88,
|
1739 |
+
"grad_norm": 0.39742231369018555,
|
1740 |
+
"learning_rate": 2.1634684631203412e-07,
|
1741 |
+
"logits/chosen": 0.0048486413434147835,
|
1742 |
+
"logits/rejected": -0.006056814920157194,
|
1743 |
+
"logps/chosen": -2.919680595397949,
|
1744 |
+
"logps/rejected": -57.8403205871582,
|
1745 |
+
"loss": 0.3464,
|
1746 |
+
"rewards/accuracies": 0.8999999761581421,
|
1747 |
+
"rewards/chosen": 17.175275802612305,
|
1748 |
+
"rewards/margins": 5.559727668762207,
|
1749 |
+
"rewards/rejected": 11.615548133850098,
|
1750 |
+
"step": 990
|
1751 |
+
},
|
1752 |
+
{
|
1753 |
+
"epoch": 0.8888888888888888,
|
1754 |
+
"grad_norm": 30.22509002685547,
|
1755 |
+
"learning_rate": 1.8587130159608196e-07,
|
1756 |
+
"logits/chosen": 0.0050649940967559814,
|
1757 |
+
"logits/rejected": -0.0071399761363863945,
|
1758 |
+
"logps/chosen": -0.16131475567817688,
|
1759 |
+
"logps/rejected": -66.25190734863281,
|
1760 |
+
"loss": 0.0063,
|
1761 |
+
"rewards/accuracies": 1.0,
|
1762 |
+
"rewards/chosen": 17.313472747802734,
|
1763 |
+
"rewards/margins": 6.412895679473877,
|
1764 |
+
"rewards/rejected": 10.9005765914917,
|
1765 |
+
"step": 1000
|
1766 |
+
},
|
1767 |
+
{
|
1768 |
+
"epoch": 0.8977777777777778,
|
1769 |
+
"grad_norm": 53.02751922607422,
|
1770 |
+
"learning_rate": 1.5762568750059604e-07,
|
1771 |
+
"logits/chosen": 0.005312003195285797,
|
1772 |
+
"logits/rejected": -0.0038596936501562595,
|
1773 |
+
"logps/chosen": -4.00323486328125,
|
1774 |
+
"logps/rejected": -57.656890869140625,
|
1775 |
+
"loss": 0.2741,
|
1776 |
+
"rewards/accuracies": 0.925000011920929,
|
1777 |
+
"rewards/chosen": 17.167781829833984,
|
1778 |
+
"rewards/margins": 5.632095813751221,
|
1779 |
+
"rewards/rejected": 11.535685539245605,
|
1780 |
+
"step": 1010
|
1781 |
+
},
|
1782 |
+
{
|
1783 |
+
"epoch": 0.9066666666666666,
|
1784 |
+
"grad_norm": 119.56330108642578,
|
1785 |
+
"learning_rate": 1.316372219412454e-07,
|
1786 |
+
"logits/chosen": 0.004333779215812683,
|
1787 |
+
"logits/rejected": -0.007937717251479626,
|
1788 |
+
"logps/chosen": -0.10221245884895325,
|
1789 |
+
"logps/rejected": -67.09977722167969,
|
1790 |
+
"loss": 0.0513,
|
1791 |
+
"rewards/accuracies": 0.9750000238418579,
|
1792 |
+
"rewards/chosen": 17.151248931884766,
|
1793 |
+
"rewards/margins": 6.162137985229492,
|
1794 |
+
"rewards/rejected": 10.98911190032959,
|
1795 |
+
"step": 1020
|
1796 |
+
},
|
1797 |
+
{
|
1798 |
+
"epoch": 0.9066666666666666,
|
1799 |
+
"eval_logits/chosen": 0.004528110846877098,
|
1800 |
+
"eval_logits/rejected": -0.006501312367618084,
|
1801 |
+
"eval_logps/chosen": -1.817779302597046,
|
1802 |
+
"eval_logps/rejected": -60.34454345703125,
|
1803 |
+
"eval_loss": 0.14118175208568573,
|
1804 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1805 |
+
"eval_rewards/chosen": 17.321117401123047,
|
1806 |
+
"eval_rewards/margins": 5.994848251342773,
|
1807 |
+
"eval_rewards/rejected": 11.32626724243164,
|
1808 |
+
"eval_runtime": 361.545,
|
1809 |
+
"eval_samples_per_second": 2.766,
|
1810 |
+
"eval_steps_per_second": 0.346,
|
1811 |
+
"step": 1020
|
1812 |
+
},
|
1813 |
+
{
|
1814 |
+
"epoch": 0.9155555555555556,
|
1815 |
+
"grad_norm": 0.9030271768569946,
|
1816 |
+
"learning_rate": 1.0793094781005792e-07,
|
1817 |
+
"logits/chosen": 0.0036234352737665176,
|
1818 |
+
"logits/rejected": -0.008390933275222778,
|
1819 |
+
"logps/chosen": -0.09899584949016571,
|
1820 |
+
"logps/rejected": -60.80555725097656,
|
1821 |
+
"loss": 0.0349,
|
1822 |
+
"rewards/accuracies": 0.987500011920929,
|
1823 |
+
"rewards/chosen": 17.619266510009766,
|
1824 |
+
"rewards/margins": 6.4814043045043945,
|
1825 |
+
"rewards/rejected": 11.137863159179688,
|
1826 |
+
"step": 1030
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 0.9244444444444444,
|
1830 |
+
"grad_norm": 157.070068359375,
|
1831 |
+
"learning_rate": 8.652970884369255e-08,
|
1832 |
+
"logits/chosen": 0.005321115255355835,
|
1833 |
+
"logits/rejected": -0.004976513795554638,
|
1834 |
+
"logps/chosen": -1.3535833358764648,
|
1835 |
+
"logps/rejected": -61.2861328125,
|
1836 |
+
"loss": 0.1927,
|
1837 |
+
"rewards/accuracies": 0.9500000476837158,
|
1838 |
+
"rewards/chosen": 17.209800720214844,
|
1839 |
+
"rewards/margins": 5.795952796936035,
|
1840 |
+
"rewards/rejected": 11.413846015930176,
|
1841 |
+
"step": 1040
|
1842 |
+
},
|
1843 |
+
{
|
1844 |
+
"epoch": 0.9333333333333333,
|
1845 |
+
"grad_norm": 24.539953231811523,
|
1846 |
+
"learning_rate": 6.745412761086007e-08,
|
1847 |
+
"logits/chosen": 0.005845514126121998,
|
1848 |
+
"logits/rejected": -0.003921338357031345,
|
1849 |
+
"logps/chosen": -1.7223193645477295,
|
1850 |
+
"logps/rejected": -58.172142028808594,
|
1851 |
+
"loss": 0.1317,
|
1852 |
+
"rewards/accuracies": 0.9750000238418579,
|
1853 |
+
"rewards/chosen": 17.360435485839844,
|
1854 |
+
"rewards/margins": 5.838218688964844,
|
1855 |
+
"rewards/rejected": 11.522216796875,
|
1856 |
+
"step": 1050
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 0.9422222222222222,
|
1860 |
+
"grad_norm": 0.061458222568035126,
|
1861 |
+
"learning_rate": 5.0722585640090305e-08,
|
1862 |
+
"logits/chosen": 0.003759522922337055,
|
1863 |
+
"logits/rejected": -0.0068417866714298725,
|
1864 |
+
"logps/chosen": -1.351855754852295,
|
1865 |
+
"logps/rejected": -59.56566619873047,
|
1866 |
+
"loss": 0.2279,
|
1867 |
+
"rewards/accuracies": 0.9375,
|
1868 |
+
"rewards/chosen": 17.245838165283203,
|
1869 |
+
"rewards/margins": 5.7038679122924805,
|
1870 |
+
"rewards/rejected": 11.541970252990723,
|
1871 |
+
"step": 1060
|
1872 |
+
},
|
1873 |
+
{
|
1874 |
+
"epoch": 0.9511111111111111,
|
1875 |
+
"grad_norm": 101.87332916259766,
|
1876 |
+
"learning_rate": 3.635120570700784e-08,
|
1877 |
+
"logits/chosen": 0.007607857696712017,
|
1878 |
+
"logits/rejected": -0.0035320711322128773,
|
1879 |
+
"logps/chosen": -1.4004669189453125,
|
1880 |
+
"logps/rejected": -63.46752166748047,
|
1881 |
+
"loss": 0.1021,
|
1882 |
+
"rewards/accuracies": 0.9624999761581421,
|
1883 |
+
"rewards/chosen": 17.240036010742188,
|
1884 |
+
"rewards/margins": 6.078882694244385,
|
1885 |
+
"rewards/rejected": 11.161155700683594,
|
1886 |
+
"step": 1070
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 0.96,
|
1890 |
+
"grad_norm": 1.9229660034179688,
|
1891 |
+
"learning_rate": 2.4353836298169343e-08,
|
1892 |
+
"logits/chosen": 0.006209026090800762,
|
1893 |
+
"logits/rejected": -0.0033816141076385975,
|
1894 |
+
"logps/chosen": -2.0281822681427,
|
1895 |
+
"logps/rejected": -61.460670471191406,
|
1896 |
+
"loss": 0.1189,
|
1897 |
+
"rewards/accuracies": 0.9750000238418579,
|
1898 |
+
"rewards/chosen": 17.19972801208496,
|
1899 |
+
"rewards/margins": 5.881450653076172,
|
1900 |
+
"rewards/rejected": 11.318277359008789,
|
1901 |
+
"step": 1080
|
1902 |
+
},
|
1903 |
+
{
|
1904 |
+
"epoch": 0.96,
|
1905 |
+
"eval_logits/chosen": 0.00737445754930377,
|
1906 |
+
"eval_logits/rejected": -0.0035832570865750313,
|
1907 |
+
"eval_logps/chosen": -2.1419789791107178,
|
1908 |
+
"eval_logps/rejected": -60.60612106323242,
|
1909 |
+
"eval_loss": 0.15081512928009033,
|
1910 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1911 |
+
"eval_rewards/chosen": 17.288694381713867,
|
1912 |
+
"eval_rewards/margins": 5.98858642578125,
|
1913 |
+
"eval_rewards/rejected": 11.300108909606934,
|
1914 |
+
"eval_runtime": 361.4165,
|
1915 |
+
"eval_samples_per_second": 2.767,
|
1916 |
+
"eval_steps_per_second": 0.346,
|
1917 |
+
"step": 1080
|
1918 |
+
},
|
1919 |
+
{
|
1920 |
+
"epoch": 0.9688888888888889,
|
1921 |
+
"grad_norm": 0.11888863146305084,
|
1922 |
+
"learning_rate": 1.4742038266447046e-08,
|
1923 |
+
"logits/chosen": 0.006043245084583759,
|
1924 |
+
"logits/rejected": -0.004712546244263649,
|
1925 |
+
"logps/chosen": -1.4127472639083862,
|
1926 |
+
"logps/rejected": -64.81529998779297,
|
1927 |
+
"loss": 0.0728,
|
1928 |
+
"rewards/accuracies": 0.9750000238418579,
|
1929 |
+
"rewards/chosen": 17.181049346923828,
|
1930 |
+
"rewards/margins": 6.14100456237793,
|
1931 |
+
"rewards/rejected": 11.040044784545898,
|
1932 |
+
"step": 1090
|
1933 |
+
},
|
1934 |
+
{
|
1935 |
+
"epoch": 0.9777777777777777,
|
1936 |
+
"grad_norm": 2.001819372177124,
|
1937 |
+
"learning_rate": 7.525073690809737e-09,
|
1938 |
+
"logits/chosen": 0.0055408780463039875,
|
1939 |
+
"logits/rejected": -0.005598037503659725,
|
1940 |
+
"logps/chosen": -1.017110824584961,
|
1941 |
+
"logps/rejected": -60.28044891357422,
|
1942 |
+
"loss": 0.0357,
|
1943 |
+
"rewards/accuracies": 0.9750000238418579,
|
1944 |
+
"rewards/chosen": 17.4679012298584,
|
1945 |
+
"rewards/margins": 6.193048477172852,
|
1946 |
+
"rewards/rejected": 11.274852752685547,
|
1947 |
+
"step": 1100
|
1948 |
+
},
|
1949 |
+
{
|
1950 |
+
"epoch": 0.9866666666666667,
|
1951 |
+
"grad_norm": 0.027056939899921417,
|
1952 |
+
"learning_rate": 2.709896951238744e-09,
|
1953 |
+
"logits/chosen": 0.00636716466397047,
|
1954 |
+
"logits/rejected": -0.004360577557235956,
|
1955 |
+
"logps/chosen": -2.8028905391693115,
|
1956 |
+
"logps/rejected": -60.86518859863281,
|
1957 |
+
"loss": 0.1973,
|
1958 |
+
"rewards/accuracies": 0.949999988079071,
|
1959 |
+
"rewards/chosen": 17.210033416748047,
|
1960 |
+
"rewards/margins": 5.936794281005859,
|
1961 |
+
"rewards/rejected": 11.273238182067871,
|
1962 |
+
"step": 1110
|
1963 |
+
},
|
1964 |
+
{
|
1965 |
+
"epoch": 0.9955555555555555,
|
1966 |
+
"grad_norm": 0.15637506544589996,
|
1967 |
+
"learning_rate": 3.0114802737818415e-10,
|
1968 |
+
"logits/chosen": 0.0038183885626494884,
|
1969 |
+
"logits/rejected": -0.007296917960047722,
|
1970 |
+
"logps/chosen": -1.1477452516555786,
|
1971 |
+
"logps/rejected": -59.37739181518555,
|
1972 |
+
"loss": 0.0538,
|
1973 |
+
"rewards/accuracies": 0.987500011920929,
|
1974 |
+
"rewards/chosen": 17.553251266479492,
|
1975 |
+
"rewards/margins": 6.2790141105651855,
|
1976 |
+
"rewards/rejected": 11.274236679077148,
|
1977 |
+
"step": 1120
|
1978 |
+
}
|
1979 |
+
],
|
1980 |
+
"logging_steps": 10,
|
1981 |
+
"max_steps": 1125,
|
1982 |
+
"num_input_tokens_seen": 0,
|
1983 |
+
"num_train_epochs": 1,
|
1984 |
+
"save_steps": 500,
|
1985 |
+
"stateful_callbacks": {
|
1986 |
+
"TrainerControl": {
|
1987 |
+
"args": {
|
1988 |
+
"should_epoch_stop": false,
|
1989 |
+
"should_evaluate": false,
|
1990 |
+
"should_log": false,
|
1991 |
+
"should_save": true,
|
1992 |
+
"should_training_stop": true
|
1993 |
+
},
|
1994 |
+
"attributes": {}
|
1995 |
+
}
|
1996 |
+
},
|
1997 |
+
"total_flos": 1.4338459346927616e+18,
|
1998 |
+
"train_batch_size": 4,
|
1999 |
+
"trial_name": null,
|
2000 |
+
"trial_params": null
|
2001 |
+
}
|
checkpoint-1125/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5286bf380083ef64133f25e161ddfb8549ff44790e432f478753f221ef89e695
|
3 |
+
size 5368
|
checkpoint-500/README.md
ADDED
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: mistralai/Mistral-Nemo-Instruct-2407
|
3 |
+
library_name: peft
|
4 |
+
---
|
5 |
+
|
6 |
+
# Model Card for Model ID
|
7 |
+
|
8 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
9 |
+
|
10 |
+
|
11 |
+
|
12 |
+
## Model Details
|
13 |
+
|
14 |
+
### Model Description
|
15 |
+
|
16 |
+
<!-- Provide a longer summary of what this model is. -->
|
17 |
+
|
18 |
+
|
19 |
+
|
20 |
+
- **Developed by:** [More Information Needed]
|
21 |
+
- **Funded by [optional]:** [More Information Needed]
|
22 |
+
- **Shared by [optional]:** [More Information Needed]
|
23 |
+
- **Model type:** [More Information Needed]
|
24 |
+
- **Language(s) (NLP):** [More Information Needed]
|
25 |
+
- **License:** [More Information Needed]
|
26 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
27 |
+
|
28 |
+
### Model Sources [optional]
|
29 |
+
|
30 |
+
<!-- Provide the basic links for the model. -->
|
31 |
+
|
32 |
+
- **Repository:** [More Information Needed]
|
33 |
+
- **Paper [optional]:** [More Information Needed]
|
34 |
+
- **Demo [optional]:** [More Information Needed]
|
35 |
+
|
36 |
+
## Uses
|
37 |
+
|
38 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
39 |
+
|
40 |
+
### Direct Use
|
41 |
+
|
42 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
43 |
+
|
44 |
+
[More Information Needed]
|
45 |
+
|
46 |
+
### Downstream Use [optional]
|
47 |
+
|
48 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
49 |
+
|
50 |
+
[More Information Needed]
|
51 |
+
|
52 |
+
### Out-of-Scope Use
|
53 |
+
|
54 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
55 |
+
|
56 |
+
[More Information Needed]
|
57 |
+
|
58 |
+
## Bias, Risks, and Limitations
|
59 |
+
|
60 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
61 |
+
|
62 |
+
[More Information Needed]
|
63 |
+
|
64 |
+
### Recommendations
|
65 |
+
|
66 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
67 |
+
|
68 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
69 |
+
|
70 |
+
## How to Get Started with the Model
|
71 |
+
|
72 |
+
Use the code below to get started with the model.
|
73 |
+
|
74 |
+
[More Information Needed]
|
75 |
+
|
76 |
+
## Training Details
|
77 |
+
|
78 |
+
### Training Data
|
79 |
+
|
80 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
81 |
+
|
82 |
+
[More Information Needed]
|
83 |
+
|
84 |
+
### Training Procedure
|
85 |
+
|
86 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
87 |
+
|
88 |
+
#### Preprocessing [optional]
|
89 |
+
|
90 |
+
[More Information Needed]
|
91 |
+
|
92 |
+
|
93 |
+
#### Training Hyperparameters
|
94 |
+
|
95 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
96 |
+
|
97 |
+
#### Speeds, Sizes, Times [optional]
|
98 |
+
|
99 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
100 |
+
|
101 |
+
[More Information Needed]
|
102 |
+
|
103 |
+
## Evaluation
|
104 |
+
|
105 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
106 |
+
|
107 |
+
### Testing Data, Factors & Metrics
|
108 |
+
|
109 |
+
#### Testing Data
|
110 |
+
|
111 |
+
<!-- This should link to a Dataset Card if possible. -->
|
112 |
+
|
113 |
+
[More Information Needed]
|
114 |
+
|
115 |
+
#### Factors
|
116 |
+
|
117 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
118 |
+
|
119 |
+
[More Information Needed]
|
120 |
+
|
121 |
+
#### Metrics
|
122 |
+
|
123 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
124 |
+
|
125 |
+
[More Information Needed]
|
126 |
+
|
127 |
+
### Results
|
128 |
+
|
129 |
+
[More Information Needed]
|
130 |
+
|
131 |
+
#### Summary
|
132 |
+
|
133 |
+
|
134 |
+
|
135 |
+
## Model Examination [optional]
|
136 |
+
|
137 |
+
<!-- Relevant interpretability work for the model goes here -->
|
138 |
+
|
139 |
+
[More Information Needed]
|
140 |
+
|
141 |
+
## Environmental Impact
|
142 |
+
|
143 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
144 |
+
|
145 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
146 |
+
|
147 |
+
- **Hardware Type:** [More Information Needed]
|
148 |
+
- **Hours used:** [More Information Needed]
|
149 |
+
- **Cloud Provider:** [More Information Needed]
|
150 |
+
- **Compute Region:** [More Information Needed]
|
151 |
+
- **Carbon Emitted:** [More Information Needed]
|
152 |
+
|
153 |
+
## Technical Specifications [optional]
|
154 |
+
|
155 |
+
### Model Architecture and Objective
|
156 |
+
|
157 |
+
[More Information Needed]
|
158 |
+
|
159 |
+
### Compute Infrastructure
|
160 |
+
|
161 |
+
[More Information Needed]
|
162 |
+
|
163 |
+
#### Hardware
|
164 |
+
|
165 |
+
[More Information Needed]
|
166 |
+
|
167 |
+
#### Software
|
168 |
+
|
169 |
+
[More Information Needed]
|
170 |
+
|
171 |
+
## Citation [optional]
|
172 |
+
|
173 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
174 |
+
|
175 |
+
**BibTeX:**
|
176 |
+
|
177 |
+
[More Information Needed]
|
178 |
+
|
179 |
+
**APA:**
|
180 |
+
|
181 |
+
[More Information Needed]
|
182 |
+
|
183 |
+
## Glossary [optional]
|
184 |
+
|
185 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
186 |
+
|
187 |
+
[More Information Needed]
|
188 |
+
|
189 |
+
## More Information [optional]
|
190 |
+
|
191 |
+
[More Information Needed]
|
192 |
+
|
193 |
+
## Model Card Authors [optional]
|
194 |
+
|
195 |
+
[More Information Needed]
|
196 |
+
|
197 |
+
## Model Card Contact
|
198 |
+
|
199 |
+
[More Information Needed]
|
200 |
+
### Framework versions
|
201 |
+
|
202 |
+
- PEFT 0.12.0
|
checkpoint-500/adapter_config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "mistralai/Mistral-Nemo-Instruct-2407",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layer_replication": null,
|
10 |
+
"layers_pattern": null,
|
11 |
+
"layers_to_transform": null,
|
12 |
+
"loftq_config": {},
|
13 |
+
"lora_alpha": 16,
|
14 |
+
"lora_dropout": 0.0,
|
15 |
+
"megatron_config": null,
|
16 |
+
"megatron_core": "megatron.core",
|
17 |
+
"modules_to_save": null,
|
18 |
+
"peft_type": "LORA",
|
19 |
+
"r": 8,
|
20 |
+
"rank_pattern": {},
|
21 |
+
"revision": null,
|
22 |
+
"target_modules": [
|
23 |
+
"q_proj",
|
24 |
+
"gate_proj",
|
25 |
+
"k_proj",
|
26 |
+
"o_proj",
|
27 |
+
"v_proj",
|
28 |
+
"down_proj",
|
29 |
+
"up_proj"
|
30 |
+
],
|
31 |
+
"task_type": "CAUSAL_LM",
|
32 |
+
"use_dora": false,
|
33 |
+
"use_rslora": false
|
34 |
+
}
|
checkpoint-500/adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2405da5efa3fa58305eb789eb90dbe77348a677c94c781b71f5a7828902865ef
|
3 |
+
size 114106856
|
checkpoint-500/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:83c427609b64317bf25a71f520f1925991aac70912f290b39b6953fa9c67daa4
|
3 |
+
size 228536930
|
checkpoint-500/rng_state_0.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:91bd7f619e4cd37883f469c08e90105c4d218fd82ffc43ae58fa9fdbcc37fce5
|
3 |
+
size 14512
|
checkpoint-500/rng_state_1.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5b0a7593f9ab52bf47328c6d50954dce1fcd69866aa6f5f35851aef7f7af3899
|
3 |
+
size 14512
|
checkpoint-500/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0fd9d73796ba0c0da5d08a8e5aa955713fdd4d633eb58f1fea83f6389e8837c8
|
3 |
+
size 1064
|
checkpoint-500/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "</s>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
checkpoint-500/tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0240ce510f08e6c2041724e9043e33be9d251d1e4a4d94eb68cd47b954b61d2
|
3 |
+
size 17078292
|
checkpoint-500/tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
checkpoint-500/trainer_state.json
ADDED
@@ -0,0 +1,911 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.4444444444444444,
|
5 |
+
"eval_steps": 60,
|
6 |
+
"global_step": 500,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.008888888888888889,
|
13 |
+
"grad_norm": 8.760091781616211,
|
14 |
+
"learning_rate": 4.4247787610619474e-07,
|
15 |
+
"logits/chosen": -0.8248252868652344,
|
16 |
+
"logits/rejected": -0.8263720273971558,
|
17 |
+
"logps/chosen": -0.36086463928222656,
|
18 |
+
"logps/rejected": -5.696224689483643,
|
19 |
+
"loss": 1.1038,
|
20 |
+
"rewards/accuracies": 0.5125000476837158,
|
21 |
+
"rewards/chosen": 17.43745994567871,
|
22 |
+
"rewards/margins": 0.5984855890274048,
|
23 |
+
"rewards/rejected": 16.838973999023438,
|
24 |
+
"step": 10
|
25 |
+
},
|
26 |
+
{
|
27 |
+
"epoch": 0.017777777777777778,
|
28 |
+
"grad_norm": 8.855981826782227,
|
29 |
+
"learning_rate": 8.849557522123895e-07,
|
30 |
+
"logits/chosen": -0.8169006109237671,
|
31 |
+
"logits/rejected": -0.819770872592926,
|
32 |
+
"logps/chosen": -0.12464660406112671,
|
33 |
+
"logps/rejected": -7.139842987060547,
|
34 |
+
"loss": 1.1887,
|
35 |
+
"rewards/accuracies": 0.4000000059604645,
|
36 |
+
"rewards/chosen": 17.17649269104004,
|
37 |
+
"rewards/margins": 0.19107049703598022,
|
38 |
+
"rewards/rejected": 16.98542022705078,
|
39 |
+
"step": 20
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 0.02666666666666667,
|
43 |
+
"grad_norm": 16.764184951782227,
|
44 |
+
"learning_rate": 1.3274336283185843e-06,
|
45 |
+
"logits/chosen": -0.8003113865852356,
|
46 |
+
"logits/rejected": -0.8030117750167847,
|
47 |
+
"logps/chosen": -0.34651467204093933,
|
48 |
+
"logps/rejected": -6.967917442321777,
|
49 |
+
"loss": 1.0563,
|
50 |
+
"rewards/accuracies": 0.44999998807907104,
|
51 |
+
"rewards/chosen": 17.280975341796875,
|
52 |
+
"rewards/margins": 0.40005987882614136,
|
53 |
+
"rewards/rejected": 16.88091468811035,
|
54 |
+
"step": 30
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 0.035555555555555556,
|
58 |
+
"grad_norm": 8.33682918548584,
|
59 |
+
"learning_rate": 1.769911504424779e-06,
|
60 |
+
"logits/chosen": -0.7695047855377197,
|
61 |
+
"logits/rejected": -0.7739207148551941,
|
62 |
+
"logps/chosen": -1.5993006229400635,
|
63 |
+
"logps/rejected": -8.504932403564453,
|
64 |
+
"loss": 0.7596,
|
65 |
+
"rewards/accuracies": 0.5,
|
66 |
+
"rewards/chosen": 17.283912658691406,
|
67 |
+
"rewards/margins": 0.6976072192192078,
|
68 |
+
"rewards/rejected": 16.5863037109375,
|
69 |
+
"step": 40
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 0.044444444444444446,
|
73 |
+
"grad_norm": 4.494723320007324,
|
74 |
+
"learning_rate": 2.212389380530974e-06,
|
75 |
+
"logits/chosen": -0.7154140472412109,
|
76 |
+
"logits/rejected": -0.7225576043128967,
|
77 |
+
"logps/chosen": -3.112199068069458,
|
78 |
+
"logps/rejected": -12.212080001831055,
|
79 |
+
"loss": 0.6083,
|
80 |
+
"rewards/accuracies": 0.4625000059604645,
|
81 |
+
"rewards/chosen": 17.03064727783203,
|
82 |
+
"rewards/margins": 0.7148451805114746,
|
83 |
+
"rewards/rejected": 16.3158016204834,
|
84 |
+
"step": 50
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.05333333333333334,
|
88 |
+
"grad_norm": 5.110287666320801,
|
89 |
+
"learning_rate": 2.6548672566371687e-06,
|
90 |
+
"logits/chosen": -0.6322453022003174,
|
91 |
+
"logits/rejected": -0.6387485265731812,
|
92 |
+
"logps/chosen": -5.650620460510254,
|
93 |
+
"logps/rejected": -12.759811401367188,
|
94 |
+
"loss": 0.3835,
|
95 |
+
"rewards/accuracies": 0.9125000238418579,
|
96 |
+
"rewards/chosen": 17.101289749145508,
|
97 |
+
"rewards/margins": 1.1824612617492676,
|
98 |
+
"rewards/rejected": 15.918828964233398,
|
99 |
+
"step": 60
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.05333333333333334,
|
103 |
+
"eval_logits/chosen": -0.5826543569564819,
|
104 |
+
"eval_logits/rejected": -0.5914276838302612,
|
105 |
+
"eval_logps/chosen": -3.5471787452697754,
|
106 |
+
"eval_logps/rejected": -16.51181983947754,
|
107 |
+
"eval_loss": 0.3286525011062622,
|
108 |
+
"eval_rewards/accuracies": 0.9280000925064087,
|
109 |
+
"eval_rewards/chosen": 17.148174285888672,
|
110 |
+
"eval_rewards/margins": 1.4386365413665771,
|
111 |
+
"eval_rewards/rejected": 15.709539413452148,
|
112 |
+
"eval_runtime": 372.0227,
|
113 |
+
"eval_samples_per_second": 2.688,
|
114 |
+
"eval_steps_per_second": 0.336,
|
115 |
+
"step": 60
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 0.06222222222222222,
|
119 |
+
"grad_norm": 5.098133563995361,
|
120 |
+
"learning_rate": 3.097345132743363e-06,
|
121 |
+
"logits/chosen": -0.5378152132034302,
|
122 |
+
"logits/rejected": -0.5494933724403381,
|
123 |
+
"logps/chosen": -1.5099802017211914,
|
124 |
+
"logps/rejected": -21.206321716308594,
|
125 |
+
"loss": 0.2931,
|
126 |
+
"rewards/accuracies": 0.9375,
|
127 |
+
"rewards/chosen": 17.083791732788086,
|
128 |
+
"rewards/margins": 1.5844331979751587,
|
129 |
+
"rewards/rejected": 15.499359130859375,
|
130 |
+
"step": 70
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 0.07111111111111111,
|
134 |
+
"grad_norm": 29.787437438964844,
|
135 |
+
"learning_rate": 3.539823008849558e-06,
|
136 |
+
"logits/chosen": -0.443774938583374,
|
137 |
+
"logits/rejected": -0.45571577548980713,
|
138 |
+
"logps/chosen": -1.5804342031478882,
|
139 |
+
"logps/rejected": -22.606929779052734,
|
140 |
+
"loss": 0.202,
|
141 |
+
"rewards/accuracies": 0.9750000238418579,
|
142 |
+
"rewards/chosen": 17.302125930786133,
|
143 |
+
"rewards/margins": 2.174014091491699,
|
144 |
+
"rewards/rejected": 15.128110885620117,
|
145 |
+
"step": 80
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"epoch": 0.08,
|
149 |
+
"grad_norm": 23.14398193359375,
|
150 |
+
"learning_rate": 3.982300884955752e-06,
|
151 |
+
"logits/chosen": -0.3626072406768799,
|
152 |
+
"logits/rejected": -0.3787815570831299,
|
153 |
+
"logps/chosen": -2.203828811645508,
|
154 |
+
"logps/rejected": -29.433551788330078,
|
155 |
+
"loss": 0.2123,
|
156 |
+
"rewards/accuracies": 0.925000011920929,
|
157 |
+
"rewards/chosen": 17.00284194946289,
|
158 |
+
"rewards/margins": 2.320391893386841,
|
159 |
+
"rewards/rejected": 14.682450294494629,
|
160 |
+
"step": 90
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 0.08888888888888889,
|
164 |
+
"grad_norm": 29.672739028930664,
|
165 |
+
"learning_rate": 4.424778761061948e-06,
|
166 |
+
"logits/chosen": -0.3035663962364197,
|
167 |
+
"logits/rejected": -0.31762221455574036,
|
168 |
+
"logps/chosen": -3.433589458465576,
|
169 |
+
"logps/rejected": -29.9322509765625,
|
170 |
+
"loss": 0.2592,
|
171 |
+
"rewards/accuracies": 0.9375,
|
172 |
+
"rewards/chosen": 16.929956436157227,
|
173 |
+
"rewards/margins": 2.31272029876709,
|
174 |
+
"rewards/rejected": 14.617237091064453,
|
175 |
+
"step": 100
|
176 |
+
},
|
177 |
+
{
|
178 |
+
"epoch": 0.09777777777777778,
|
179 |
+
"grad_norm": 1.873722791671753,
|
180 |
+
"learning_rate": 4.867256637168142e-06,
|
181 |
+
"logits/chosen": -0.2679600715637207,
|
182 |
+
"logits/rejected": -0.2826440930366516,
|
183 |
+
"logps/chosen": -0.9653514623641968,
|
184 |
+
"logps/rejected": -30.235322952270508,
|
185 |
+
"loss": 0.1336,
|
186 |
+
"rewards/accuracies": 0.949999988079071,
|
187 |
+
"rewards/chosen": 17.462385177612305,
|
188 |
+
"rewards/margins": 3.1994175910949707,
|
189 |
+
"rewards/rejected": 14.26296615600586,
|
190 |
+
"step": 110
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 0.10666666666666667,
|
194 |
+
"grad_norm": 1.6913721561431885,
|
195 |
+
"learning_rate": 4.999409761242696e-06,
|
196 |
+
"logits/chosen": -0.22222033143043518,
|
197 |
+
"logits/rejected": -0.23720571398735046,
|
198 |
+
"logps/chosen": -4.4953508377075195,
|
199 |
+
"logps/rejected": -34.074745178222656,
|
200 |
+
"loss": 0.2552,
|
201 |
+
"rewards/accuracies": 0.8999999761581421,
|
202 |
+
"rewards/chosen": 17.04866600036621,
|
203 |
+
"rewards/margins": 3.1014418601989746,
|
204 |
+
"rewards/rejected": 13.947224617004395,
|
205 |
+
"step": 120
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.10666666666666667,
|
209 |
+
"eval_logits/chosen": -0.206527978181839,
|
210 |
+
"eval_logits/rejected": -0.22178640961647034,
|
211 |
+
"eval_logps/chosen": -3.69442081451416,
|
212 |
+
"eval_logps/rejected": -36.072166442871094,
|
213 |
+
"eval_loss": 0.18996010720729828,
|
214 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
215 |
+
"eval_rewards/chosen": 17.133451461791992,
|
216 |
+
"eval_rewards/margins": 3.379946708679199,
|
217 |
+
"eval_rewards/rejected": 13.753504753112793,
|
218 |
+
"eval_runtime": 361.5279,
|
219 |
+
"eval_samples_per_second": 2.766,
|
220 |
+
"eval_steps_per_second": 0.346,
|
221 |
+
"step": 120
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 0.11555555555555555,
|
225 |
+
"grad_norm": 61.80262756347656,
|
226 |
+
"learning_rate": 4.996519466816778e-06,
|
227 |
+
"logits/chosen": -0.18473535776138306,
|
228 |
+
"logits/rejected": -0.1988501250743866,
|
229 |
+
"logps/chosen": -3.7009687423706055,
|
230 |
+
"logps/rejected": -39.289939880371094,
|
231 |
+
"loss": 0.1394,
|
232 |
+
"rewards/accuracies": 0.9624999761581421,
|
233 |
+
"rewards/chosen": 17.106964111328125,
|
234 |
+
"rewards/margins": 3.633338212966919,
|
235 |
+
"rewards/rejected": 13.473625183105469,
|
236 |
+
"step": 130
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.12444444444444444,
|
240 |
+
"grad_norm": 1.6732702255249023,
|
241 |
+
"learning_rate": 4.9912234871722805e-06,
|
242 |
+
"logits/chosen": -0.16134041547775269,
|
243 |
+
"logits/rejected": -0.17547868192195892,
|
244 |
+
"logps/chosen": -3.0637736320495605,
|
245 |
+
"logps/rejected": -40.07548522949219,
|
246 |
+
"loss": 0.1408,
|
247 |
+
"rewards/accuracies": 0.9750000238418579,
|
248 |
+
"rewards/chosen": 17.392223358154297,
|
249 |
+
"rewards/margins": 4.242353439331055,
|
250 |
+
"rewards/rejected": 13.149867057800293,
|
251 |
+
"step": 140
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 0.13333333333333333,
|
255 |
+
"grad_norm": 0.346453994512558,
|
256 |
+
"learning_rate": 4.98352692559805e-06,
|
257 |
+
"logits/chosen": -0.13797929883003235,
|
258 |
+
"logits/rejected": -0.15283086895942688,
|
259 |
+
"logps/chosen": -5.14492130279541,
|
260 |
+
"logps/rejected": -47.97212219238281,
|
261 |
+
"loss": 0.2153,
|
262 |
+
"rewards/accuracies": 0.9375,
|
263 |
+
"rewards/chosen": 16.896778106689453,
|
264 |
+
"rewards/margins": 4.227695465087891,
|
265 |
+
"rewards/rejected": 12.669081687927246,
|
266 |
+
"step": 150
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.14222222222222222,
|
270 |
+
"grad_norm": 0.21871662139892578,
|
271 |
+
"learning_rate": 4.973437198621237e-06,
|
272 |
+
"logits/chosen": -0.12396670132875443,
|
273 |
+
"logits/rejected": -0.13780555129051208,
|
274 |
+
"logps/chosen": -6.108860015869141,
|
275 |
+
"logps/rejected": -54.90739440917969,
|
276 |
+
"loss": 0.0388,
|
277 |
+
"rewards/accuracies": 0.9750000238418579,
|
278 |
+
"rewards/chosen": 16.75935935974121,
|
279 |
+
"rewards/margins": 4.755282878875732,
|
280 |
+
"rewards/rejected": 12.004077911376953,
|
281 |
+
"step": 160
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.1511111111111111,
|
285 |
+
"grad_norm": 235.12429809570312,
|
286 |
+
"learning_rate": 4.960964028860621e-06,
|
287 |
+
"logits/chosen": -0.1140839159488678,
|
288 |
+
"logits/rejected": -0.1263057291507721,
|
289 |
+
"logps/chosen": -12.605452537536621,
|
290 |
+
"logps/rejected": -53.81230926513672,
|
291 |
+
"loss": 0.4651,
|
292 |
+
"rewards/accuracies": 0.875,
|
293 |
+
"rewards/chosen": 16.101238250732422,
|
294 |
+
"rewards/margins": 3.9864249229431152,
|
295 |
+
"rewards/rejected": 12.114812850952148,
|
296 |
+
"step": 170
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.16,
|
300 |
+
"grad_norm": 190.97048950195312,
|
301 |
+
"learning_rate": 4.946119435657738e-06,
|
302 |
+
"logits/chosen": -0.10746976733207703,
|
303 |
+
"logits/rejected": -0.11878640949726105,
|
304 |
+
"logps/chosen": -8.5105562210083,
|
305 |
+
"logps/rejected": -51.314781188964844,
|
306 |
+
"loss": 0.2362,
|
307 |
+
"rewards/accuracies": 0.925000011920929,
|
308 |
+
"rewards/chosen": 16.719980239868164,
|
309 |
+
"rewards/margins": 4.549674034118652,
|
310 |
+
"rewards/rejected": 12.170306205749512,
|
311 |
+
"step": 180
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 0.16,
|
315 |
+
"eval_logits/chosen": -0.10870806127786636,
|
316 |
+
"eval_logits/rejected": -0.12223993986845016,
|
317 |
+
"eval_logps/chosen": -4.414996147155762,
|
318 |
+
"eval_logps/rejected": -53.885032653808594,
|
319 |
+
"eval_loss": 0.20236633718013763,
|
320 |
+
"eval_rewards/accuracies": 0.9510000944137573,
|
321 |
+
"eval_rewards/chosen": 17.06139373779297,
|
322 |
+
"eval_rewards/margins": 5.089176177978516,
|
323 |
+
"eval_rewards/rejected": 11.97221851348877,
|
324 |
+
"eval_runtime": 361.4355,
|
325 |
+
"eval_samples_per_second": 2.767,
|
326 |
+
"eval_steps_per_second": 0.346,
|
327 |
+
"step": 180
|
328 |
+
},
|
329 |
+
{
|
330 |
+
"epoch": 0.1688888888888889,
|
331 |
+
"grad_norm": 56.81266784667969,
|
332 |
+
"learning_rate": 4.928917723494854e-06,
|
333 |
+
"logits/chosen": -0.10682469606399536,
|
334 |
+
"logits/rejected": -0.12124393880367279,
|
335 |
+
"logps/chosen": -3.058413028717041,
|
336 |
+
"logps/rejected": -55.052528381347656,
|
337 |
+
"loss": 0.2442,
|
338 |
+
"rewards/accuracies": 0.9500000476837158,
|
339 |
+
"rewards/chosen": 17.058589935302734,
|
340 |
+
"rewards/margins": 5.056097984313965,
|
341 |
+
"rewards/rejected": 12.002490043640137,
|
342 |
+
"step": 190
|
343 |
+
},
|
344 |
+
{
|
345 |
+
"epoch": 0.17777777777777778,
|
346 |
+
"grad_norm": 175.06552124023438,
|
347 |
+
"learning_rate": 4.909375468210947e-06,
|
348 |
+
"logits/chosen": -0.10520349442958832,
|
349 |
+
"logits/rejected": -0.12018950283527374,
|
350 |
+
"logps/chosen": -4.114959716796875,
|
351 |
+
"logps/rejected": -55.9394645690918,
|
352 |
+
"loss": 0.1915,
|
353 |
+
"rewards/accuracies": 0.9500000476837158,
|
354 |
+
"rewards/chosen": 16.98603057861328,
|
355 |
+
"rewards/margins": 5.105838775634766,
|
356 |
+
"rewards/rejected": 11.880191802978516,
|
357 |
+
"step": 200
|
358 |
+
},
|
359 |
+
{
|
360 |
+
"epoch": 0.18666666666666668,
|
361 |
+
"grad_norm": 78.06558990478516,
|
362 |
+
"learning_rate": 4.8875115010289655e-06,
|
363 |
+
"logits/chosen": -0.10475558042526245,
|
364 |
+
"logits/rejected": -0.11949175596237183,
|
365 |
+
"logps/chosen": -6.760301113128662,
|
366 |
+
"logps/rejected": -53.91607666015625,
|
367 |
+
"loss": 0.2843,
|
368 |
+
"rewards/accuracies": 0.9375,
|
369 |
+
"rewards/chosen": 16.857545852661133,
|
370 |
+
"rewards/margins": 4.917357921600342,
|
371 |
+
"rewards/rejected": 11.94018840789795,
|
372 |
+
"step": 210
|
373 |
+
},
|
374 |
+
{
|
375 |
+
"epoch": 0.19555555555555557,
|
376 |
+
"grad_norm": 15.880486488342285,
|
377 |
+
"learning_rate": 4.863346890409768e-06,
|
378 |
+
"logits/chosen": -0.11213523149490356,
|
379 |
+
"logits/rejected": -0.12581588327884674,
|
380 |
+
"logps/chosen": -6.759585380554199,
|
381 |
+
"logps/rejected": -51.10936737060547,
|
382 |
+
"loss": 0.5104,
|
383 |
+
"rewards/accuracies": 0.875,
|
384 |
+
"rewards/chosen": 16.859071731567383,
|
385 |
+
"rewards/margins": 4.638372898101807,
|
386 |
+
"rewards/rejected": 12.220698356628418,
|
387 |
+
"step": 220
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.20444444444444446,
|
391 |
+
"grad_norm": 46.97845458984375,
|
392 |
+
"learning_rate": 4.836904921750224e-06,
|
393 |
+
"logits/chosen": -0.11947059631347656,
|
394 |
+
"logits/rejected": -0.1329912692308426,
|
395 |
+
"logps/chosen": -3.608184814453125,
|
396 |
+
"logps/rejected": -48.794761657714844,
|
397 |
+
"loss": 0.2134,
|
398 |
+
"rewards/accuracies": 0.925000011920929,
|
399 |
+
"rewards/chosen": 17.235904693603516,
|
400 |
+
"rewards/margins": 4.859888076782227,
|
401 |
+
"rewards/rejected": 12.376014709472656,
|
402 |
+
"step": 230
|
403 |
+
},
|
404 |
+
{
|
405 |
+
"epoch": 0.21333333333333335,
|
406 |
+
"grad_norm": 24.032859802246094,
|
407 |
+
"learning_rate": 4.808211074945042e-06,
|
408 |
+
"logits/chosen": -0.1200513243675232,
|
409 |
+
"logits/rejected": -0.1333036869764328,
|
410 |
+
"logps/chosen": -3.7552154064178467,
|
411 |
+
"logps/rejected": -49.87453079223633,
|
412 |
+
"loss": 0.1781,
|
413 |
+
"rewards/accuracies": 0.9500000476837158,
|
414 |
+
"rewards/chosen": 17.094650268554688,
|
415 |
+
"rewards/margins": 4.68077278137207,
|
416 |
+
"rewards/rejected": 12.41387939453125,
|
417 |
+
"step": 240
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 0.21333333333333335,
|
421 |
+
"eval_logits/chosen": -0.12433278560638428,
|
422 |
+
"eval_logits/rejected": -0.13808581233024597,
|
423 |
+
"eval_logps/chosen": -4.408891201019287,
|
424 |
+
"eval_logps/rejected": -50.744781494140625,
|
425 |
+
"eval_loss": 0.1546352356672287,
|
426 |
+
"eval_rewards/accuracies": 0.9500000476837158,
|
427 |
+
"eval_rewards/chosen": 17.06200408935547,
|
428 |
+
"eval_rewards/margins": 4.775761604309082,
|
429 |
+
"eval_rewards/rejected": 12.286243438720703,
|
430 |
+
"eval_runtime": 361.4974,
|
431 |
+
"eval_samples_per_second": 2.766,
|
432 |
+
"eval_steps_per_second": 0.346,
|
433 |
+
"step": 240
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 0.2222222222222222,
|
437 |
+
"grad_norm": 0.25737640261650085,
|
438 |
+
"learning_rate": 4.7772929998339485e-06,
|
439 |
+
"logits/chosen": -0.12348780035972595,
|
440 |
+
"logits/rejected": -0.13704943656921387,
|
441 |
+
"logps/chosen": -4.4299187660217285,
|
442 |
+
"logps/rejected": -53.074607849121094,
|
443 |
+
"loss": 0.1373,
|
444 |
+
"rewards/accuracies": 0.9375,
|
445 |
+
"rewards/chosen": 17.087068557739258,
|
446 |
+
"rewards/margins": 5.06691837310791,
|
447 |
+
"rewards/rejected": 12.020149230957031,
|
448 |
+
"step": 250
|
449 |
+
},
|
450 |
+
{
|
451 |
+
"epoch": 0.2311111111111111,
|
452 |
+
"grad_norm": 0.1839389204978943,
|
453 |
+
"learning_rate": 4.744180489557859e-06,
|
454 |
+
"logits/chosen": -0.12177034467458725,
|
455 |
+
"logits/rejected": -0.1342695653438568,
|
456 |
+
"logps/chosen": -3.775188446044922,
|
457 |
+
"logps/rejected": -53.98720932006836,
|
458 |
+
"loss": 0.1896,
|
459 |
+
"rewards/accuracies": 0.949999988079071,
|
460 |
+
"rewards/chosen": 17.12021255493164,
|
461 |
+
"rewards/margins": 5.148064613342285,
|
462 |
+
"rewards/rejected": 11.972146987915039,
|
463 |
+
"step": 260
|
464 |
+
},
|
465 |
+
{
|
466 |
+
"epoch": 0.24,
|
467 |
+
"grad_norm": 12.258485794067383,
|
468 |
+
"learning_rate": 4.708905451849754e-06,
|
469 |
+
"logits/chosen": -0.11067859083414078,
|
470 |
+
"logits/rejected": -0.12377731502056122,
|
471 |
+
"logps/chosen": -6.418317794799805,
|
472 |
+
"logps/rejected": -56.57402801513672,
|
473 |
+
"loss": 0.2315,
|
474 |
+
"rewards/accuracies": 0.9375,
|
475 |
+
"rewards/chosen": 16.738832473754883,
|
476 |
+
"rewards/margins": 4.884931564331055,
|
477 |
+
"rewards/rejected": 11.853900909423828,
|
478 |
+
"step": 270
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.24888888888888888,
|
482 |
+
"grad_norm": 77.56194305419922,
|
483 |
+
"learning_rate": 4.671501878287879e-06,
|
484 |
+
"logits/chosen": -0.1184445172548294,
|
485 |
+
"logits/rejected": -0.1339874565601349,
|
486 |
+
"logps/chosen": -10.12116527557373,
|
487 |
+
"logps/rejected": -53.403907775878906,
|
488 |
+
"loss": 0.5343,
|
489 |
+
"rewards/accuracies": 0.862500011920929,
|
490 |
+
"rewards/chosen": 16.458633422851562,
|
491 |
+
"rewards/margins": 4.402472496032715,
|
492 |
+
"rewards/rejected": 12.056160926818848,
|
493 |
+
"step": 280
|
494 |
+
},
|
495 |
+
{
|
496 |
+
"epoch": 0.2577777777777778,
|
497 |
+
"grad_norm": 67.53883361816406,
|
498 |
+
"learning_rate": 4.6320058115409295e-06,
|
499 |
+
"logits/chosen": -0.1448262631893158,
|
500 |
+
"logits/rejected": -0.15793387591838837,
|
501 |
+
"logps/chosen": -3.4666190147399902,
|
502 |
+
"logps/rejected": -48.79213333129883,
|
503 |
+
"loss": 0.5017,
|
504 |
+
"rewards/accuracies": 0.887499988079071,
|
505 |
+
"rewards/chosen": 16.945899963378906,
|
506 |
+
"rewards/margins": 4.2686333656311035,
|
507 |
+
"rewards/rejected": 12.677268028259277,
|
508 |
+
"step": 290
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 0.26666666666666666,
|
512 |
+
"grad_norm": 0.17521341145038605,
|
513 |
+
"learning_rate": 4.590455310636778e-06,
|
514 |
+
"logits/chosen": -0.16128253936767578,
|
515 |
+
"logits/rejected": -0.17375555634498596,
|
516 |
+
"logps/chosen": -2.9032950401306152,
|
517 |
+
"logps/rejected": -47.69734191894531,
|
518 |
+
"loss": 0.265,
|
519 |
+
"rewards/accuracies": 0.925000011920929,
|
520 |
+
"rewards/chosen": 17.18383026123047,
|
521 |
+
"rewards/margins": 4.541309356689453,
|
522 |
+
"rewards/rejected": 12.642518997192383,
|
523 |
+
"step": 300
|
524 |
+
},
|
525 |
+
{
|
526 |
+
"epoch": 0.26666666666666666,
|
527 |
+
"eval_logits/chosen": -0.17444846034049988,
|
528 |
+
"eval_logits/rejected": -0.18559777736663818,
|
529 |
+
"eval_logps/chosen": -2.535512924194336,
|
530 |
+
"eval_logps/rejected": -47.16367721557617,
|
531 |
+
"eval_loss": 0.15360687673091888,
|
532 |
+
"eval_rewards/accuracies": 0.9440000653266907,
|
533 |
+
"eval_rewards/chosen": 17.249343872070312,
|
534 |
+
"eval_rewards/margins": 4.604989051818848,
|
535 |
+
"eval_rewards/rejected": 12.644353866577148,
|
536 |
+
"eval_runtime": 361.4575,
|
537 |
+
"eval_samples_per_second": 2.767,
|
538 |
+
"eval_steps_per_second": 0.346,
|
539 |
+
"step": 300
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 0.27555555555555555,
|
543 |
+
"grad_norm": 0.5040452480316162,
|
544 |
+
"learning_rate": 4.54689041428819e-06,
|
545 |
+
"logits/chosen": -0.16974106431007385,
|
546 |
+
"logits/rejected": -0.1810058057308197,
|
547 |
+
"logps/chosen": -1.233938217163086,
|
548 |
+
"logps/rejected": -49.907745361328125,
|
549 |
+
"loss": 0.1132,
|
550 |
+
"rewards/accuracies": 0.9500000476837158,
|
551 |
+
"rewards/chosen": 17.34117889404297,
|
552 |
+
"rewards/margins": 4.934173583984375,
|
553 |
+
"rewards/rejected": 12.407005310058594,
|
554 |
+
"step": 310
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 0.28444444444444444,
|
558 |
+
"grad_norm": 100.02949523925781,
|
559 |
+
"learning_rate": 4.501353102310901e-06,
|
560 |
+
"logits/chosen": -0.15705889463424683,
|
561 |
+
"logits/rejected": -0.1695334017276764,
|
562 |
+
"logps/chosen": -1.0820492506027222,
|
563 |
+
"logps/rejected": -52.577110290527344,
|
564 |
+
"loss": 0.1194,
|
565 |
+
"rewards/accuracies": 0.9500000476837158,
|
566 |
+
"rewards/chosen": 17.33388900756836,
|
567 |
+
"rewards/margins": 5.154760837554932,
|
568 |
+
"rewards/rejected": 12.179126739501953,
|
569 |
+
"step": 320
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.29333333333333333,
|
573 |
+
"grad_norm": 0.2689219117164612,
|
574 |
+
"learning_rate": 4.453887255171206e-06,
|
575 |
+
"logits/chosen": -0.13849371671676636,
|
576 |
+
"logits/rejected": -0.14990833401679993,
|
577 |
+
"logps/chosen": -1.8435032367706299,
|
578 |
+
"logps/rejected": -54.79044723510742,
|
579 |
+
"loss": 0.0926,
|
580 |
+
"rewards/accuracies": 0.9500000476837158,
|
581 |
+
"rewards/chosen": 17.2423095703125,
|
582 |
+
"rewards/margins": 5.28987979888916,
|
583 |
+
"rewards/rejected": 11.952428817749023,
|
584 |
+
"step": 330
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 0.3022222222222222,
|
588 |
+
"grad_norm": 0.09305431693792343,
|
589 |
+
"learning_rate": 4.404538611702055e-06,
|
590 |
+
"logits/chosen": -0.12299702316522598,
|
591 |
+
"logits/rejected": -0.13453055918216705,
|
592 |
+
"logps/chosen": -2.9897143840789795,
|
593 |
+
"logps/rejected": -52.954498291015625,
|
594 |
+
"loss": 0.2873,
|
595 |
+
"rewards/accuracies": 0.925000011920929,
|
596 |
+
"rewards/chosen": 17.17474365234375,
|
597 |
+
"rewards/margins": 5.071004867553711,
|
598 |
+
"rewards/rejected": 12.103739738464355,
|
599 |
+
"step": 340
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 0.3111111111111111,
|
603 |
+
"grad_norm": 59.282073974609375,
|
604 |
+
"learning_rate": 4.3533547250284015e-06,
|
605 |
+
"logits/chosen": -0.11913029849529266,
|
606 |
+
"logits/rejected": -0.12785324454307556,
|
607 |
+
"logps/chosen": -3.9456872940063477,
|
608 |
+
"logps/rejected": -48.68487548828125,
|
609 |
+
"loss": 0.4332,
|
610 |
+
"rewards/accuracies": 0.875,
|
611 |
+
"rewards/chosen": 17.12805938720703,
|
612 |
+
"rewards/margins": 4.669450283050537,
|
613 |
+
"rewards/rejected": 12.458610534667969,
|
614 |
+
"step": 350
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.32,
|
618 |
+
"grad_norm": 0.31101909279823303,
|
619 |
+
"learning_rate": 4.300384916744261e-06,
|
620 |
+
"logits/chosen": -0.11280188709497452,
|
621 |
+
"logits/rejected": -0.12300585210323334,
|
622 |
+
"logps/chosen": -2.1714723110198975,
|
623 |
+
"logps/rejected": -54.74174118041992,
|
624 |
+
"loss": 0.1605,
|
625 |
+
"rewards/accuracies": 0.9500000476837158,
|
626 |
+
"rewards/chosen": 17.326162338256836,
|
627 |
+
"rewards/margins": 5.467062473297119,
|
628 |
+
"rewards/rejected": 11.859098434448242,
|
629 |
+
"step": 360
|
630 |
+
},
|
631 |
+
{
|
632 |
+
"epoch": 0.32,
|
633 |
+
"eval_logits/chosen": -0.10620756447315216,
|
634 |
+
"eval_logits/rejected": -0.11727114766836166,
|
635 |
+
"eval_logps/chosen": -1.4165427684783936,
|
636 |
+
"eval_logps/rejected": -50.9525146484375,
|
637 |
+
"eval_loss": 0.3194349706172943,
|
638 |
+
"eval_rewards/accuracies": 0.9210000038146973,
|
639 |
+
"eval_rewards/chosen": 17.36124038696289,
|
640 |
+
"eval_rewards/margins": 5.095769882202148,
|
641 |
+
"eval_rewards/rejected": 12.26546859741211,
|
642 |
+
"eval_runtime": 361.5072,
|
643 |
+
"eval_samples_per_second": 2.766,
|
644 |
+
"eval_steps_per_second": 0.346,
|
645 |
+
"step": 360
|
646 |
+
},
|
647 |
+
{
|
648 |
+
"epoch": 0.3288888888888889,
|
649 |
+
"grad_norm": 6.1126532554626465,
|
650 |
+
"learning_rate": 4.24568022938566e-06,
|
651 |
+
"logits/chosen": -0.10354311764240265,
|
652 |
+
"logits/rejected": -0.11526636779308319,
|
653 |
+
"logps/chosen": -1.2935255765914917,
|
654 |
+
"logps/rejected": -55.57566833496094,
|
655 |
+
"loss": 0.1711,
|
656 |
+
"rewards/accuracies": 0.9500000476837158,
|
657 |
+
"rewards/chosen": 17.439346313476562,
|
658 |
+
"rewards/margins": 5.700921058654785,
|
659 |
+
"rewards/rejected": 11.738424301147461,
|
660 |
+
"step": 370
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.3377777777777778,
|
664 |
+
"grad_norm": 34.15927505493164,
|
665 |
+
"learning_rate": 4.189293377245241e-06,
|
666 |
+
"logits/chosen": -0.1029932051897049,
|
667 |
+
"logits/rejected": -0.11382515728473663,
|
668 |
+
"logps/chosen": -2.5132687091827393,
|
669 |
+
"logps/rejected": -55.50346374511719,
|
670 |
+
"loss": 0.4359,
|
671 |
+
"rewards/accuracies": 0.8875000476837158,
|
672 |
+
"rewards/chosen": 16.731037139892578,
|
673 |
+
"rewards/margins": 4.368172645568848,
|
674 |
+
"rewards/rejected": 12.362865447998047,
|
675 |
+
"step": 380
|
676 |
+
},
|
677 |
+
{
|
678 |
+
"epoch": 0.3466666666666667,
|
679 |
+
"grad_norm": 2.8422904014587402,
|
680 |
+
"learning_rate": 4.131278695575952e-06,
|
681 |
+
"logits/chosen": -0.10793520510196686,
|
682 |
+
"logits/rejected": -0.12109285593032837,
|
683 |
+
"logps/chosen": -3.014652729034424,
|
684 |
+
"logps/rejected": -53.98411560058594,
|
685 |
+
"loss": 0.2161,
|
686 |
+
"rewards/accuracies": 0.949999988079071,
|
687 |
+
"rewards/chosen": 17.137393951416016,
|
688 |
+
"rewards/margins": 5.105995178222656,
|
689 |
+
"rewards/rejected": 12.03139877319336,
|
690 |
+
"step": 390
|
691 |
+
},
|
692 |
+
{
|
693 |
+
"epoch": 0.35555555555555557,
|
694 |
+
"grad_norm": 54.0329475402832,
|
695 |
+
"learning_rate": 4.071692088232743e-06,
|
696 |
+
"logits/chosen": -0.10393750667572021,
|
697 |
+
"logits/rejected": -0.11834606528282166,
|
698 |
+
"logps/chosen": -2.1508543491363525,
|
699 |
+
"logps/rejected": -45.60733413696289,
|
700 |
+
"loss": 0.2077,
|
701 |
+
"rewards/accuracies": 0.925000011920929,
|
702 |
+
"rewards/chosen": 17.586124420166016,
|
703 |
+
"rewards/margins": 5.077212333679199,
|
704 |
+
"rewards/rejected": 12.5089111328125,
|
705 |
+
"step": 400
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 0.36444444444444446,
|
709 |
+
"grad_norm": 81.61144256591797,
|
710 |
+
"learning_rate": 4.010590973802737e-06,
|
711 |
+
"logits/chosen": -0.09564584493637085,
|
712 |
+
"logits/rejected": -0.10617707669734955,
|
713 |
+
"logps/chosen": -3.4572842121124268,
|
714 |
+
"logps/rejected": -50.92162322998047,
|
715 |
+
"loss": 0.2478,
|
716 |
+
"rewards/accuracies": 0.8875000476837158,
|
717 |
+
"rewards/chosen": 17.010910034179688,
|
718 |
+
"rewards/margins": 4.556198596954346,
|
719 |
+
"rewards/rejected": 12.454713821411133,
|
720 |
+
"step": 410
|
721 |
+
},
|
722 |
+
{
|
723 |
+
"epoch": 0.37333333333333335,
|
724 |
+
"grad_norm": 0.30974289774894714,
|
725 |
+
"learning_rate": 3.948034230275781e-06,
|
726 |
+
"logits/chosen": -0.09134417027235031,
|
727 |
+
"logits/rejected": -0.1020016297698021,
|
728 |
+
"logps/chosen": -5.046698570251465,
|
729 |
+
"logps/rejected": -48.908958435058594,
|
730 |
+
"loss": 0.2894,
|
731 |
+
"rewards/accuracies": 0.8999999761581421,
|
732 |
+
"rewards/chosen": 17.007888793945312,
|
733 |
+
"rewards/margins": 4.53641414642334,
|
734 |
+
"rewards/rejected": 12.471475601196289,
|
735 |
+
"step": 420
|
736 |
+
},
|
737 |
+
{
|
738 |
+
"epoch": 0.37333333333333335,
|
739 |
+
"eval_logits/chosen": -0.09054450690746307,
|
740 |
+
"eval_logits/rejected": -0.10264354199171066,
|
741 |
+
"eval_logps/chosen": -1.913105845451355,
|
742 |
+
"eval_logps/rejected": -51.11127471923828,
|
743 |
+
"eval_loss": 0.16789735853672028,
|
744 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
745 |
+
"eval_rewards/chosen": 17.311582565307617,
|
746 |
+
"eval_rewards/margins": 5.061989784240723,
|
747 |
+
"eval_rewards/rejected": 12.249593734741211,
|
748 |
+
"eval_runtime": 361.5337,
|
749 |
+
"eval_samples_per_second": 2.766,
|
750 |
+
"eval_steps_per_second": 0.346,
|
751 |
+
"step": 420
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.38222222222222224,
|
755 |
+
"grad_norm": 12.824393272399902,
|
756 |
+
"learning_rate": 3.884082138308699e-06,
|
757 |
+
"logits/chosen": -0.08666776865720749,
|
758 |
+
"logits/rejected": -0.0997733399271965,
|
759 |
+
"logps/chosen": -1.7306327819824219,
|
760 |
+
"logps/rejected": -54.273292541503906,
|
761 |
+
"loss": 0.2298,
|
762 |
+
"rewards/accuracies": 0.9500000476837158,
|
763 |
+
"rewards/chosen": 17.167621612548828,
|
764 |
+
"rewards/margins": 5.065673351287842,
|
765 |
+
"rewards/rejected": 12.101947784423828,
|
766 |
+
"step": 430
|
767 |
+
},
|
768 |
+
{
|
769 |
+
"epoch": 0.39111111111111113,
|
770 |
+
"grad_norm": 0.30713599920272827,
|
771 |
+
"learning_rate": 3.818796323137896e-06,
|
772 |
+
"logits/chosen": -0.09174907952547073,
|
773 |
+
"logits/rejected": -0.10376611351966858,
|
774 |
+
"logps/chosen": -1.489154577255249,
|
775 |
+
"logps/rejected": -54.580726623535156,
|
776 |
+
"loss": 0.2513,
|
777 |
+
"rewards/accuracies": 0.9375,
|
778 |
+
"rewards/chosen": 17.22280502319336,
|
779 |
+
"rewards/margins": 5.175349235534668,
|
780 |
+
"rewards/rejected": 12.047454833984375,
|
781 |
+
"step": 440
|
782 |
+
},
|
783 |
+
{
|
784 |
+
"epoch": 0.4,
|
785 |
+
"grad_norm": 87.4791488647461,
|
786 |
+
"learning_rate": 3.7522396951963303e-06,
|
787 |
+
"logits/chosen": -0.09688778221607208,
|
788 |
+
"logits/rejected": -0.10897806286811829,
|
789 |
+
"logps/chosen": -3.157695770263672,
|
790 |
+
"logps/rejected": -50.96417236328125,
|
791 |
+
"loss": 0.1758,
|
792 |
+
"rewards/accuracies": 0.9500000476837158,
|
793 |
+
"rewards/chosen": 17.345651626586914,
|
794 |
+
"rewards/margins": 5.245656967163086,
|
795 |
+
"rewards/rejected": 12.099993705749512,
|
796 |
+
"step": 450
|
797 |
+
},
|
798 |
+
{
|
799 |
+
"epoch": 0.4088888888888889,
|
800 |
+
"grad_norm": 146.2008056640625,
|
801 |
+
"learning_rate": 3.684476389492026e-06,
|
802 |
+
"logits/chosen": -0.09378582239151001,
|
803 |
+
"logits/rejected": -0.10475654900074005,
|
804 |
+
"logps/chosen": -0.5611928701400757,
|
805 |
+
"logps/rejected": -56.518890380859375,
|
806 |
+
"loss": 0.1981,
|
807 |
+
"rewards/accuracies": 0.9500000476837158,
|
808 |
+
"rewards/chosen": 17.113712310791016,
|
809 |
+
"rewards/margins": 5.068872928619385,
|
810 |
+
"rewards/rejected": 12.044839859008789,
|
811 |
+
"step": 460
|
812 |
+
},
|
813 |
+
{
|
814 |
+
"epoch": 0.4177777777777778,
|
815 |
+
"grad_norm": 1.9137721061706543,
|
816 |
+
"learning_rate": 3.6155717038065783e-06,
|
817 |
+
"logits/chosen": -0.08695463836193085,
|
818 |
+
"logits/rejected": -0.09596743434667587,
|
819 |
+
"logps/chosen": -1.5298550128936768,
|
820 |
+
"logps/rejected": -50.27445983886719,
|
821 |
+
"loss": 0.2066,
|
822 |
+
"rewards/accuracies": 0.9375,
|
823 |
+
"rewards/chosen": 17.35186004638672,
|
824 |
+
"rewards/margins": 5.014693260192871,
|
825 |
+
"rewards/rejected": 12.337167739868164,
|
826 |
+
"step": 470
|
827 |
+
},
|
828 |
+
{
|
829 |
+
"epoch": 0.4266666666666667,
|
830 |
+
"grad_norm": 84.80391693115234,
|
831 |
+
"learning_rate": 3.545592035773192e-06,
|
832 |
+
"logits/chosen": -0.0746893435716629,
|
833 |
+
"logits/rejected": -0.08653923869132996,
|
834 |
+
"logps/chosen": -2.0052125453948975,
|
835 |
+
"logps/rejected": -57.502811431884766,
|
836 |
+
"loss": 0.1149,
|
837 |
+
"rewards/accuracies": 0.9500000476837158,
|
838 |
+
"rewards/chosen": 17.14373016357422,
|
839 |
+
"rewards/margins": 5.360415935516357,
|
840 |
+
"rewards/rejected": 11.783313751220703,
|
841 |
+
"step": 480
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 0.4266666666666667,
|
845 |
+
"eval_logits/chosen": -0.07700399309396744,
|
846 |
+
"eval_logits/rejected": -0.08828537166118622,
|
847 |
+
"eval_logps/chosen": -4.48896598815918,
|
848 |
+
"eval_logps/rejected": -53.76282501220703,
|
849 |
+
"eval_loss": 0.29511645436286926,
|
850 |
+
"eval_rewards/accuracies": 0.9230000376701355,
|
851 |
+
"eval_rewards/chosen": 17.053997039794922,
|
852 |
+
"eval_rewards/margins": 5.069558143615723,
|
853 |
+
"eval_rewards/rejected": 11.984437942504883,
|
854 |
+
"eval_runtime": 361.5035,
|
855 |
+
"eval_samples_per_second": 2.766,
|
856 |
+
"eval_steps_per_second": 0.346,
|
857 |
+
"step": 480
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 0.43555555555555553,
|
861 |
+
"grad_norm": 82.9616470336914,
|
862 |
+
"learning_rate": 3.4746048188948806e-06,
|
863 |
+
"logits/chosen": -0.06675051152706146,
|
864 |
+
"logits/rejected": -0.07860895991325378,
|
865 |
+
"logps/chosen": -4.162237167358398,
|
866 |
+
"logps/rejected": -54.77789306640625,
|
867 |
+
"loss": 0.2979,
|
868 |
+
"rewards/accuracies": 0.9125000238418579,
|
869 |
+
"rewards/chosen": 17.047603607177734,
|
870 |
+
"rewards/margins": 5.138361930847168,
|
871 |
+
"rewards/rejected": 11.909242630004883,
|
872 |
+
"step": 490
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.4444444444444444,
|
876 |
+
"grad_norm": 0.04293210059404373,
|
877 |
+
"learning_rate": 3.4026784575644887e-06,
|
878 |
+
"logits/chosen": -0.06424491107463837,
|
879 |
+
"logits/rejected": -0.07567107677459717,
|
880 |
+
"logps/chosen": -2.05729603767395,
|
881 |
+
"logps/rejected": -56.646087646484375,
|
882 |
+
"loss": 0.4378,
|
883 |
+
"rewards/accuracies": 0.8875000476837158,
|
884 |
+
"rewards/chosen": 16.947803497314453,
|
885 |
+
"rewards/margins": 4.919981956481934,
|
886 |
+
"rewards/rejected": 12.02782154083252,
|
887 |
+
"step": 500
|
888 |
+
}
|
889 |
+
],
|
890 |
+
"logging_steps": 10,
|
891 |
+
"max_steps": 1125,
|
892 |
+
"num_input_tokens_seen": 0,
|
893 |
+
"num_train_epochs": 1,
|
894 |
+
"save_steps": 500,
|
895 |
+
"stateful_callbacks": {
|
896 |
+
"TrainerControl": {
|
897 |
+
"args": {
|
898 |
+
"should_epoch_stop": false,
|
899 |
+
"should_evaluate": false,
|
900 |
+
"should_log": false,
|
901 |
+
"should_save": true,
|
902 |
+
"should_training_stop": false
|
903 |
+
},
|
904 |
+
"attributes": {}
|
905 |
+
}
|
906 |
+
},
|
907 |
+
"total_flos": 6.372648598634496e+17,
|
908 |
+
"train_batch_size": 4,
|
909 |
+
"trial_name": null,
|
910 |
+
"trial_params": null
|
911 |
+
}
|
checkpoint-500/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5286bf380083ef64133f25e161ddfb8549ff44790e432f478753f221ef89e695
|
3 |
+
size 5368
|
eval_results.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"eval_logits/chosen": 0.0034948738757520914,
|
4 |
+
"eval_logits/rejected": -0.007565807551145554,
|
5 |
+
"eval_logps/chosen": -2.20627498626709,
|
6 |
+
"eval_logps/rejected": -60.60332489013672,
|
7 |
+
"eval_loss": 0.15350937843322754,
|
8 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
9 |
+
"eval_rewards/chosen": 17.28226661682129,
|
10 |
+
"eval_rewards/margins": 5.981877326965332,
|
11 |
+
"eval_rewards/rejected": 11.30038833618164,
|
12 |
+
"eval_runtime": 361.5652,
|
13 |
+
"eval_samples_per_second": 2.766,
|
14 |
+
"eval_steps_per_second": 0.346
|
15 |
+
}
|
runs/Dec18_22-22-59_yadi/events.out.tfevents.1734560709.yadi.436386.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:697250b13dc063f122f3cccdacb9ae4ca8a13cbea14f3d9f6c0083f67e62277a
|
3 |
+
size 95792
|
runs/Dec18_22-22-59_yadi/events.out.tfevents.1734574921.yadi.436386.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:77f0ff4dbe534e3fef2d0e8a554d703d5b9903029ef4e11f5a8c6830e79a5bfe
|
3 |
+
size 828
|
special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "</s>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0240ce510f08e6c2041724e9043e33be9d251d1e4a4d94eb68cd47b954b61d2
|
3 |
+
size 17078292
|
tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"total_flos": 1.4338459346927616e+18,
|
4 |
+
"train_loss": 0.23597783709896936,
|
5 |
+
"train_runtime": 13850.7044,
|
6 |
+
"train_samples_per_second": 0.65,
|
7 |
+
"train_steps_per_second": 0.081
|
8 |
+
}
|
trainer_log.jsonl
ADDED
@@ -0,0 +1,131 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{"current_steps": 10, "total_steps": 1125, "loss": 1.1038, "accuracy": 0.5125000476837158, "lr": 4.4247787610619474e-07, "epoch": 0.008888888888888889, "percentage": 0.89, "elapsed_time": "0:01:11", "remaining_time": "2:12:28"}
|
2 |
+
{"current_steps": 20, "total_steps": 1125, "loss": 1.1887, "accuracy": 0.4000000059604645, "lr": 8.849557522123895e-07, "epoch": 0.017777777777777778, "percentage": 1.78, "elapsed_time": "0:02:19", "remaining_time": "2:08:48"}
|
3 |
+
{"current_steps": 30, "total_steps": 1125, "loss": 1.0563, "accuracy": 0.44999998807907104, "lr": 1.3274336283185843e-06, "epoch": 0.02666666666666667, "percentage": 2.67, "elapsed_time": "0:03:27", "remaining_time": "2:06:15"}
|
4 |
+
{"current_steps": 40, "total_steps": 1125, "loss": 0.7596, "accuracy": 0.5, "lr": 1.769911504424779e-06, "epoch": 0.035555555555555556, "percentage": 3.56, "elapsed_time": "0:04:36", "remaining_time": "2:04:51"}
|
5 |
+
{"current_steps": 50, "total_steps": 1125, "loss": 0.6083, "accuracy": 0.4625000059604645, "lr": 2.212389380530974e-06, "epoch": 0.044444444444444446, "percentage": 4.44, "elapsed_time": "0:05:44", "remaining_time": "2:03:19"}
|
6 |
+
{"current_steps": 60, "total_steps": 1125, "loss": 0.3835, "accuracy": 0.9125000238418579, "lr": 2.6548672566371687e-06, "epoch": 0.05333333333333334, "percentage": 5.33, "elapsed_time": "0:06:51", "remaining_time": "2:01:46"}
|
7 |
+
{"current_steps": 60, "total_steps": 1125, "eval_loss": 0.3286525011062622, "epoch": 0.05333333333333334, "percentage": 5.33, "elapsed_time": "0:13:03", "remaining_time": "3:51:50"}
|
8 |
+
{"current_steps": 70, "total_steps": 1125, "loss": 0.2931, "accuracy": 0.9375, "lr": 3.097345132743363e-06, "epoch": 0.06222222222222222, "percentage": 6.22, "elapsed_time": "0:14:08", "remaining_time": "3:33:12"}
|
9 |
+
{"current_steps": 80, "total_steps": 1125, "loss": 0.202, "accuracy": 0.9750000238418579, "lr": 3.539823008849558e-06, "epoch": 0.07111111111111111, "percentage": 7.11, "elapsed_time": "0:15:13", "remaining_time": "3:18:57"}
|
10 |
+
{"current_steps": 90, "total_steps": 1125, "loss": 0.2123, "accuracy": 0.925000011920929, "lr": 3.982300884955752e-06, "epoch": 0.08, "percentage": 8.0, "elapsed_time": "0:16:19", "remaining_time": "3:07:39"}
|
11 |
+
{"current_steps": 100, "total_steps": 1125, "loss": 0.2592, "accuracy": 0.9375, "lr": 4.424778761061948e-06, "epoch": 0.08888888888888889, "percentage": 8.89, "elapsed_time": "0:17:24", "remaining_time": "2:58:22"}
|
12 |
+
{"current_steps": 110, "total_steps": 1125, "loss": 0.1336, "accuracy": 0.949999988079071, "lr": 4.867256637168142e-06, "epoch": 0.09777777777777778, "percentage": 9.78, "elapsed_time": "0:18:29", "remaining_time": "2:50:35"}
|
13 |
+
{"current_steps": 120, "total_steps": 1125, "loss": 0.2552, "accuracy": 0.8999999761581421, "lr": 4.999409761242696e-06, "epoch": 0.10666666666666667, "percentage": 10.67, "elapsed_time": "0:19:34", "remaining_time": "2:43:55"}
|
14 |
+
{"current_steps": 120, "total_steps": 1125, "eval_loss": 0.18996010720729828, "epoch": 0.10666666666666667, "percentage": 10.67, "elapsed_time": "0:25:35", "remaining_time": "3:34:23"}
|
15 |
+
{"current_steps": 130, "total_steps": 1125, "loss": 0.1394, "accuracy": 0.9624999761581421, "lr": 4.996519466816778e-06, "epoch": 0.11555555555555555, "percentage": 11.56, "elapsed_time": "0:26:41", "remaining_time": "3:24:16"}
|
16 |
+
{"current_steps": 140, "total_steps": 1125, "loss": 0.1408, "accuracy": 0.9750000238418579, "lr": 4.9912234871722805e-06, "epoch": 0.12444444444444444, "percentage": 12.44, "elapsed_time": "0:27:46", "remaining_time": "3:15:25"}
|
17 |
+
{"current_steps": 150, "total_steps": 1125, "loss": 0.2153, "accuracy": 0.9375, "lr": 4.98352692559805e-06, "epoch": 0.13333333333333333, "percentage": 13.33, "elapsed_time": "0:28:51", "remaining_time": "3:07:35"}
|
18 |
+
{"current_steps": 160, "total_steps": 1125, "loss": 0.0388, "accuracy": 0.9750000238418579, "lr": 4.973437198621237e-06, "epoch": 0.14222222222222222, "percentage": 14.22, "elapsed_time": "0:29:56", "remaining_time": "3:00:36"}
|
19 |
+
{"current_steps": 170, "total_steps": 1125, "loss": 0.4651, "accuracy": 0.875, "lr": 4.960964028860621e-06, "epoch": 0.1511111111111111, "percentage": 15.11, "elapsed_time": "0:31:01", "remaining_time": "2:54:18"}
|
20 |
+
{"current_steps": 180, "total_steps": 1125, "loss": 0.2362, "accuracy": 0.925000011920929, "lr": 4.946119435657738e-06, "epoch": 0.16, "percentage": 16.0, "elapsed_time": "0:32:06", "remaining_time": "2:48:34"}
|
21 |
+
{"current_steps": 180, "total_steps": 1125, "eval_loss": 0.20236633718013763, "epoch": 0.16, "percentage": 16.0, "elapsed_time": "0:38:08", "remaining_time": "3:20:12"}
|
22 |
+
{"current_steps": 190, "total_steps": 1125, "loss": 0.2442, "accuracy": 0.9500000476837158, "lr": 4.928917723494854e-06, "epoch": 0.1688888888888889, "percentage": 16.89, "elapsed_time": "0:39:13", "remaining_time": "3:12:59"}
|
23 |
+
{"current_steps": 200, "total_steps": 1125, "loss": 0.1915, "accuracy": 0.9500000476837158, "lr": 4.909375468210947e-06, "epoch": 0.17777777777777778, "percentage": 17.78, "elapsed_time": "0:40:18", "remaining_time": "3:06:24"}
|
24 |
+
{"current_steps": 210, "total_steps": 1125, "loss": 0.2843, "accuracy": 0.9375, "lr": 4.8875115010289655e-06, "epoch": 0.18666666666666668, "percentage": 18.67, "elapsed_time": "0:41:23", "remaining_time": "3:00:19"}
|
25 |
+
{"current_steps": 220, "total_steps": 1125, "loss": 0.5104, "accuracy": 0.875, "lr": 4.863346890409768e-06, "epoch": 0.19555555555555557, "percentage": 19.56, "elapsed_time": "0:42:28", "remaining_time": "2:54:41"}
|
26 |
+
{"current_steps": 230, "total_steps": 1125, "loss": 0.2134, "accuracy": 0.925000011920929, "lr": 4.836904921750224e-06, "epoch": 0.20444444444444446, "percentage": 20.44, "elapsed_time": "0:43:33", "remaining_time": "2:49:28"}
|
27 |
+
{"current_steps": 240, "total_steps": 1125, "loss": 0.1781, "accuracy": 0.9500000476837158, "lr": 4.808211074945042e-06, "epoch": 0.21333333333333335, "percentage": 21.33, "elapsed_time": "0:44:38", "remaining_time": "2:44:35"}
|
28 |
+
{"current_steps": 240, "total_steps": 1125, "eval_loss": 0.1546352356672287, "epoch": 0.21333333333333335, "percentage": 21.33, "elapsed_time": "0:50:39", "remaining_time": "3:06:48"}
|
29 |
+
{"current_steps": 250, "total_steps": 1125, "loss": 0.1373, "accuracy": 0.9375, "lr": 4.7772929998339485e-06, "epoch": 0.2222222222222222, "percentage": 22.22, "elapsed_time": "0:51:44", "remaining_time": "3:01:05"}
|
30 |
+
{"current_steps": 260, "total_steps": 1125, "loss": 0.1896, "accuracy": 0.949999988079071, "lr": 4.744180489557859e-06, "epoch": 0.2311111111111111, "percentage": 23.11, "elapsed_time": "0:52:49", "remaining_time": "2:55:44"}
|
31 |
+
{"current_steps": 270, "total_steps": 1125, "loss": 0.2315, "accuracy": 0.9375, "lr": 4.708905451849754e-06, "epoch": 0.24, "percentage": 24.0, "elapsed_time": "0:53:54", "remaining_time": "2:50:42"}
|
32 |
+
{"current_steps": 280, "total_steps": 1125, "loss": 0.5343, "accuracy": 0.862500011920929, "lr": 4.671501878287879e-06, "epoch": 0.24888888888888888, "percentage": 24.89, "elapsed_time": "0:54:59", "remaining_time": "2:45:56"}
|
33 |
+
{"current_steps": 290, "total_steps": 1125, "loss": 0.5017, "accuracy": 0.887499988079071, "lr": 4.6320058115409295e-06, "epoch": 0.2577777777777778, "percentage": 25.78, "elapsed_time": "0:56:04", "remaining_time": "2:41:26"}
|
34 |
+
{"current_steps": 300, "total_steps": 1125, "loss": 0.265, "accuracy": 0.925000011920929, "lr": 4.590455310636778e-06, "epoch": 0.26666666666666666, "percentage": 26.67, "elapsed_time": "0:57:09", "remaining_time": "2:37:09"}
|
35 |
+
{"current_steps": 300, "total_steps": 1125, "eval_loss": 0.15360687673091888, "epoch": 0.26666666666666666, "percentage": 26.67, "elapsed_time": "1:03:10", "remaining_time": "2:53:44"}
|
36 |
+
{"current_steps": 310, "total_steps": 1125, "loss": 0.1132, "accuracy": 0.9500000476837158, "lr": 4.54689041428819e-06, "epoch": 0.27555555555555555, "percentage": 27.56, "elapsed_time": "1:04:15", "remaining_time": "2:48:56"}
|
37 |
+
{"current_steps": 320, "total_steps": 1125, "loss": 0.1194, "accuracy": 0.9500000476837158, "lr": 4.501353102310901e-06, "epoch": 0.28444444444444444, "percentage": 28.44, "elapsed_time": "1:05:21", "remaining_time": "2:44:25"}
|
38 |
+
{"current_steps": 330, "total_steps": 1125, "loss": 0.0926, "accuracy": 0.9500000476837158, "lr": 4.453887255171206e-06, "epoch": 0.29333333333333333, "percentage": 29.33, "elapsed_time": "1:06:26", "remaining_time": "2:40:04"}
|
39 |
+
{"current_steps": 340, "total_steps": 1125, "loss": 0.2873, "accuracy": 0.925000011920929, "lr": 4.404538611702055e-06, "epoch": 0.3022222222222222, "percentage": 30.22, "elapsed_time": "1:07:31", "remaining_time": "2:35:54"}
|
40 |
+
{"current_steps": 350, "total_steps": 1125, "loss": 0.4332, "accuracy": 0.875, "lr": 4.3533547250284015e-06, "epoch": 0.3111111111111111, "percentage": 31.11, "elapsed_time": "1:08:36", "remaining_time": "2:31:54"}
|
41 |
+
{"current_steps": 360, "total_steps": 1125, "loss": 0.1605, "accuracy": 0.9500000476837158, "lr": 4.300384916744261e-06, "epoch": 0.32, "percentage": 32.0, "elapsed_time": "1:09:41", "remaining_time": "2:28:05"}
|
42 |
+
{"current_steps": 360, "total_steps": 1125, "eval_loss": 0.3194349706172943, "epoch": 0.32, "percentage": 32.0, "elapsed_time": "1:15:42", "remaining_time": "2:40:53"}
|
43 |
+
{"current_steps": 370, "total_steps": 1125, "loss": 0.1711, "accuracy": 0.9500000476837158, "lr": 4.24568022938566e-06, "epoch": 0.3288888888888889, "percentage": 32.89, "elapsed_time": "1:16:47", "remaining_time": "2:36:42"}
|
44 |
+
{"current_steps": 380, "total_steps": 1125, "loss": 0.4359, "accuracy": 0.8875000476837158, "lr": 4.189293377245241e-06, "epoch": 0.3377777777777778, "percentage": 33.78, "elapsed_time": "1:17:52", "remaining_time": "2:32:40"}
|
45 |
+
{"current_steps": 390, "total_steps": 1125, "loss": 0.2161, "accuracy": 0.949999988079071, "lr": 4.131278695575952e-06, "epoch": 0.3466666666666667, "percentage": 34.67, "elapsed_time": "1:18:57", "remaining_time": "2:28:48"}
|
46 |
+
{"current_steps": 400, "total_steps": 1125, "loss": 0.2077, "accuracy": 0.925000011920929, "lr": 4.071692088232743e-06, "epoch": 0.35555555555555557, "percentage": 35.56, "elapsed_time": "1:20:02", "remaining_time": "2:25:04"}
|
47 |
+
{"current_steps": 410, "total_steps": 1125, "loss": 0.2478, "accuracy": 0.8875000476837158, "lr": 4.010590973802737e-06, "epoch": 0.36444444444444446, "percentage": 36.44, "elapsed_time": "1:21:07", "remaining_time": "2:21:27"}
|
48 |
+
{"current_steps": 420, "total_steps": 1125, "loss": 0.2894, "accuracy": 0.8999999761581421, "lr": 3.948034230275781e-06, "epoch": 0.37333333333333335, "percentage": 37.33, "elapsed_time": "1:22:12", "remaining_time": "2:17:58"}
|
49 |
+
{"current_steps": 420, "total_steps": 1125, "eval_loss": 0.16789735853672028, "epoch": 0.37333333333333335, "percentage": 37.33, "elapsed_time": "1:28:13", "remaining_time": "2:28:05"}
|
50 |
+
{"current_steps": 430, "total_steps": 1125, "loss": 0.2298, "accuracy": 0.9500000476837158, "lr": 3.884082138308699e-06, "epoch": 0.38222222222222224, "percentage": 38.22, "elapsed_time": "1:29:18", "remaining_time": "2:24:20"}
|
51 |
+
{"current_steps": 440, "total_steps": 1125, "loss": 0.2513, "accuracy": 0.9375, "lr": 3.818796323137896e-06, "epoch": 0.39111111111111113, "percentage": 39.11, "elapsed_time": "1:30:23", "remaining_time": "2:20:43"}
|
52 |
+
{"current_steps": 450, "total_steps": 1125, "loss": 0.1758, "accuracy": 0.9500000476837158, "lr": 3.7522396951963303e-06, "epoch": 0.4, "percentage": 40.0, "elapsed_time": "1:31:28", "remaining_time": "2:17:12"}
|
53 |
+
{"current_steps": 460, "total_steps": 1125, "loss": 0.1981, "accuracy": 0.9500000476837158, "lr": 3.684476389492026e-06, "epoch": 0.4088888888888889, "percentage": 40.89, "elapsed_time": "1:32:33", "remaining_time": "2:13:48"}
|
54 |
+
{"current_steps": 470, "total_steps": 1125, "loss": 0.2066, "accuracy": 0.9375, "lr": 3.6155717038065783e-06, "epoch": 0.4177777777777778, "percentage": 41.78, "elapsed_time": "1:33:38", "remaining_time": "2:10:29"}
|
55 |
+
{"current_steps": 480, "total_steps": 1125, "loss": 0.1149, "accuracy": 0.9500000476837158, "lr": 3.545592035773192e-06, "epoch": 0.4266666666666667, "percentage": 42.67, "elapsed_time": "1:34:43", "remaining_time": "2:07:16"}
|
56 |
+
{"current_steps": 480, "total_steps": 1125, "eval_loss": 0.29511645436286926, "epoch": 0.4266666666666667, "percentage": 42.67, "elapsed_time": "1:40:44", "remaining_time": "2:15:22"}
|
57 |
+
{"current_steps": 490, "total_steps": 1125, "loss": 0.2979, "accuracy": 0.9125000238418579, "lr": 3.4746048188948806e-06, "epoch": 0.43555555555555553, "percentage": 43.56, "elapsed_time": "1:41:49", "remaining_time": "2:11:57"}
|
58 |
+
{"current_steps": 500, "total_steps": 1125, "loss": 0.4378, "accuracy": 0.8875000476837158, "lr": 3.4026784575644887e-06, "epoch": 0.4444444444444444, "percentage": 44.44, "elapsed_time": "1:42:54", "remaining_time": "2:08:37"}
|
59 |
+
{"current_steps": 510, "total_steps": 1125, "loss": 0.2613, "accuracy": 0.9375, "lr": 3.329882261149148e-06, "epoch": 0.4533333333333333, "percentage": 45.33, "elapsed_time": "1:44:00", "remaining_time": "2:05:25"}
|
60 |
+
{"current_steps": 520, "total_steps": 1125, "loss": 0.1517, "accuracy": 0.9375, "lr": 3.25628637720269e-06, "epoch": 0.4622222222222222, "percentage": 46.22, "elapsed_time": "1:45:05", "remaining_time": "2:02:16"}
|
61 |
+
{"current_steps": 530, "total_steps": 1125, "loss": 0.2304, "accuracy": 0.9375, "lr": 3.181961723870359e-06, "epoch": 0.4711111111111111, "percentage": 47.11, "elapsed_time": "1:46:10", "remaining_time": "1:59:11"}
|
62 |
+
{"current_steps": 540, "total_steps": 1125, "loss": 0.0384, "accuracy": 0.987500011920929, "lr": 3.1069799215509847e-06, "epoch": 0.48, "percentage": 48.0, "elapsed_time": "1:47:15", "remaining_time": "1:56:11"}
|
63 |
+
{"current_steps": 540, "total_steps": 1125, "eval_loss": 0.17387841641902924, "epoch": 0.48, "percentage": 48.0, "elapsed_time": "1:53:16", "remaining_time": "2:02:43"}
|
64 |
+
{"current_steps": 550, "total_steps": 1125, "loss": 0.2087, "accuracy": 0.9125000238418579, "lr": 3.0314132238824416e-06, "epoch": 0.4888888888888889, "percentage": 48.89, "elapsed_time": "1:54:21", "remaining_time": "1:59:33"}
|
65 |
+
{"current_steps": 560, "total_steps": 1125, "loss": 0.0969, "accuracy": 0.9624999761581421, "lr": 2.955334448116915e-06, "epoch": 0.49777777777777776, "percentage": 49.78, "elapsed_time": "1:55:26", "remaining_time": "1:56:28"}
|
66 |
+
{"current_steps": 570, "total_steps": 1125, "loss": 0.1447, "accuracy": 0.949999988079071, "lr": 2.8788169049530533e-06, "epoch": 0.5066666666666667, "percentage": 50.67, "elapsed_time": "1:56:31", "remaining_time": "1:53:27"}
|
67 |
+
{"current_steps": 580, "total_steps": 1125, "loss": 0.0998, "accuracy": 0.949999988079071, "lr": 2.8019343278926397e-06, "epoch": 0.5155555555555555, "percentage": 51.56, "elapsed_time": "1:57:36", "remaining_time": "1:50:30"}
|
68 |
+
{"current_steps": 590, "total_steps": 1125, "loss": 0.1162, "accuracy": 0.9750000238418579, "lr": 2.7247608021898265e-06, "epoch": 0.5244444444444445, "percentage": 52.44, "elapsed_time": "1:58:41", "remaining_time": "1:47:37"}
|
69 |
+
{"current_steps": 600, "total_steps": 1125, "loss": 0.4008, "accuracy": 0.8875000476837158, "lr": 2.647370693461432e-06, "epoch": 0.5333333333333333, "percentage": 53.33, "elapsed_time": "1:59:46", "remaining_time": "1:44:47"}
|
70 |
+
{"current_steps": 600, "total_steps": 1125, "eval_loss": 0.17061151564121246, "epoch": 0.5333333333333333, "percentage": 53.33, "elapsed_time": "2:05:47", "remaining_time": "1:50:04"}
|
71 |
+
{"current_steps": 610, "total_steps": 1125, "loss": 0.1265, "accuracy": 0.9624999761581421, "lr": 2.569838576027068e-06, "epoch": 0.5422222222222223, "percentage": 54.22, "elapsed_time": "2:06:52", "remaining_time": "1:47:06"}
|
72 |
+
{"current_steps": 620, "total_steps": 1125, "loss": 0.097, "accuracy": 0.9624999761581421, "lr": 2.4922391610481544e-06, "epoch": 0.5511111111111111, "percentage": 55.11, "elapsed_time": "2:07:57", "remaining_time": "1:44:13"}
|
73 |
+
{"current_steps": 630, "total_steps": 1125, "loss": 0.3115, "accuracy": 0.925000011920929, "lr": 2.4146472245350804e-06, "epoch": 0.56, "percentage": 56.0, "elapsed_time": "2:09:02", "remaining_time": "1:41:23"}
|
74 |
+
{"current_steps": 640, "total_steps": 1125, "loss": 0.1268, "accuracy": 0.9500000476837158, "lr": 2.337137535291868e-06, "epoch": 0.5688888888888889, "percentage": 56.89, "elapsed_time": "2:10:07", "remaining_time": "1:38:36"}
|
75 |
+
{"current_steps": 650, "total_steps": 1125, "loss": 0.1831, "accuracy": 0.9500000476837158, "lr": 2.259784782867782e-06, "epoch": 0.5777777777777777, "percentage": 57.78, "elapsed_time": "2:11:12", "remaining_time": "1:35:52"}
|
76 |
+
{"current_steps": 660, "total_steps": 1125, "loss": 0.1678, "accuracy": 0.9624999761581421, "lr": 2.182663505585314e-06, "epoch": 0.5866666666666667, "percentage": 58.67, "elapsed_time": "2:12:17", "remaining_time": "1:33:11"}
|
77 |
+
{"current_steps": 660, "total_steps": 1125, "eval_loss": 0.2050127536058426, "epoch": 0.5866666666666667, "percentage": 58.67, "elapsed_time": "2:18:18", "remaining_time": "1:37:26"}
|
78 |
+
{"current_steps": 670, "total_steps": 1125, "loss": 0.1025, "accuracy": 0.9624999761581421, "lr": 2.1058480187138863e-06, "epoch": 0.5955555555555555, "percentage": 59.56, "elapsed_time": "2:19:23", "remaining_time": "1:34:39"}
|
79 |
+
{"current_steps": 680, "total_steps": 1125, "loss": 0.1961, "accuracy": 0.9500000476837158, "lr": 2.0294123428584985e-06, "epoch": 0.6044444444444445, "percentage": 60.44, "elapsed_time": "2:20:28", "remaining_time": "1:31:55"}
|
80 |
+
{"current_steps": 690, "total_steps": 1125, "loss": 0.2821, "accuracy": 0.9375, "lr": 1.953430132632311e-06, "epoch": 0.6133333333333333, "percentage": 61.33, "elapsed_time": "2:21:33", "remaining_time": "1:29:14"}
|
81 |
+
{"current_steps": 700, "total_steps": 1125, "loss": 0.3141, "accuracy": 0.9375, "lr": 1.8779746056819104e-06, "epoch": 0.6222222222222222, "percentage": 62.22, "elapsed_time": "2:22:38", "remaining_time": "1:26:36"}
|
82 |
+
{"current_steps": 710, "total_steps": 1125, "loss": 0.3827, "accuracy": 0.887499988079071, "lr": 1.8031184721336364e-06, "epoch": 0.6311111111111111, "percentage": 63.11, "elapsed_time": "2:23:43", "remaining_time": "1:24:00"}
|
83 |
+
{"current_steps": 720, "total_steps": 1125, "loss": 0.2272, "accuracy": 0.949999988079071, "lr": 1.7289338645289711e-06, "epoch": 0.64, "percentage": 64.0, "elapsed_time": "2:24:48", "remaining_time": "1:21:27"}
|
84 |
+
{"current_steps": 720, "total_steps": 1125, "eval_loss": 0.14018221199512482, "epoch": 0.64, "percentage": 64.0, "elapsed_time": "2:30:49", "remaining_time": "1:24:50"}
|
85 |
+
{"current_steps": 730, "total_steps": 1125, "loss": 0.1171, "accuracy": 0.9624999761581421, "lr": 1.6554922683164875e-06, "epoch": 0.6488888888888888, "percentage": 64.89, "elapsed_time": "2:31:54", "remaining_time": "1:22:11"}
|
86 |
+
{"current_steps": 740, "total_steps": 1125, "loss": 0.2001, "accuracy": 0.9624999761581421, "lr": 1.5828644529673592e-06, "epoch": 0.6577777777777778, "percentage": 65.78, "elapsed_time": "2:32:59", "remaining_time": "1:19:35"}
|
87 |
+
{"current_steps": 750, "total_steps": 1125, "loss": 0.1692, "accuracy": 0.9375, "lr": 1.5111204037807844e-06, "epoch": 0.6666666666666666, "percentage": 66.67, "elapsed_time": "2:34:04", "remaining_time": "1:17:02"}
|
88 |
+
{"current_steps": 760, "total_steps": 1125, "loss": 0.2418, "accuracy": 0.949999988079071, "lr": 1.4403292544450625e-06, "epoch": 0.6755555555555556, "percentage": 67.56, "elapsed_time": "2:35:09", "remaining_time": "1:14:30"}
|
89 |
+
{"current_steps": 770, "total_steps": 1125, "loss": 0.1749, "accuracy": 0.9500000476837158, "lr": 1.3705592204192853e-06, "epoch": 0.6844444444444444, "percentage": 68.44, "elapsed_time": "2:36:14", "remaining_time": "1:12:01"}
|
90 |
+
{"current_steps": 780, "total_steps": 1125, "loss": 0.1915, "accuracy": 0.925000011920929, "lr": 1.301877533199859e-06, "epoch": 0.6933333333333334, "percentage": 69.33, "elapsed_time": "2:37:19", "remaining_time": "1:09:34"}
|
91 |
+
{"current_steps": 780, "total_steps": 1125, "eval_loss": 0.24412688612937927, "epoch": 0.6933333333333334, "percentage": 69.33, "elapsed_time": "2:43:20", "remaining_time": "1:12:14"}
|
92 |
+
{"current_steps": 790, "total_steps": 1125, "loss": 0.2143, "accuracy": 0.9375, "lr": 1.2343503755351729e-06, "epoch": 0.7022222222222222, "percentage": 70.22, "elapsed_time": "2:44:25", "remaining_time": "1:09:43"}
|
93 |
+
{"current_steps": 800, "total_steps": 1125, "loss": 0.0792, "accuracy": 0.9750000238418579, "lr": 1.168042817650881e-06, "epoch": 0.7111111111111111, "percentage": 71.11, "elapsed_time": "2:45:30", "remaining_time": "1:07:14"}
|
94 |
+
{"current_steps": 810, "total_steps": 1125, "loss": 0.3177, "accuracy": 0.925000011920929, "lr": 1.1030187545472012e-06, "epoch": 0.72, "percentage": 72.0, "elapsed_time": "2:46:35", "remaining_time": "1:04:47"}
|
95 |
+
{"current_steps": 820, "total_steps": 1125, "loss": 0.2128, "accuracy": 0.949999988079071, "lr": 1.0393408444287048e-06, "epoch": 0.7288888888888889, "percentage": 72.89, "elapsed_time": "2:47:40", "remaining_time": "1:02:21"}
|
96 |
+
{"current_steps": 830, "total_steps": 1125, "loss": 0.1296, "accuracy": 0.9750000238418579, "lr": 9.770704483258782e-07, "epoch": 0.7377777777777778, "percentage": 73.78, "elapsed_time": "2:48:44", "remaining_time": "0:59:58"}
|
97 |
+
{"current_steps": 840, "total_steps": 1125, "loss": 0.0635, "accuracy": 0.987500011920929, "lr": 9.162675709666865e-07, "epoch": 0.7466666666666667, "percentage": 74.67, "elapsed_time": "2:49:49", "remaining_time": "0:57:37"}
|
98 |
+
{"current_steps": 840, "total_steps": 1125, "eval_loss": 0.1689341962337494, "epoch": 0.7466666666666667, "percentage": 74.67, "elapsed_time": "2:55:51", "remaining_time": "0:59:39"}
|
99 |
+
{"current_steps": 850, "total_steps": 1125, "loss": 0.1157, "accuracy": 0.9750000238418579, "lr": 8.569908029550686e-07, "epoch": 0.7555555555555555, "percentage": 75.56, "elapsed_time": "2:56:56", "remaining_time": "0:57:14"}
|
100 |
+
{"current_steps": 860, "total_steps": 1125, "loss": 0.2252, "accuracy": 0.9375, "lr": 7.992972643121227e-07, "epoch": 0.7644444444444445, "percentage": 76.44, "elapsed_time": "2:58:01", "remaining_time": "0:54:51"}
|
101 |
+
{"current_steps": 870, "total_steps": 1125, "loss": 0.1398, "accuracy": 0.9624999761581421, "lr": 7.432425494343509e-07, "epoch": 0.7733333333333333, "percentage": 77.33, "elapsed_time": "2:59:06", "remaining_time": "0:52:29"}
|
102 |
+
{"current_steps": 880, "total_steps": 1125, "loss": 0.2966, "accuracy": 0.9375, "lr": 6.888806735220396e-07, "epoch": 0.7822222222222223, "percentage": 78.22, "elapsed_time": "3:00:11", "remaining_time": "0:50:09"}
|
103 |
+
{"current_steps": 890, "total_steps": 1125, "loss": 0.1641, "accuracy": 0.9375, "lr": 6.362640205293583e-07, "epoch": 0.7911111111111111, "percentage": 79.11, "elapsed_time": "3:01:15", "remaining_time": "0:47:51"}
|
104 |
+
{"current_steps": 900, "total_steps": 1125, "loss": 0.1703, "accuracy": 0.9624999761581421, "lr": 5.854432926863684e-07, "epoch": 0.8, "percentage": 80.0, "elapsed_time": "3:02:20", "remaining_time": "0:45:35"}
|
105 |
+
{"current_steps": 900, "total_steps": 1125, "eval_loss": 0.13995186984539032, "epoch": 0.8, "percentage": 80.0, "elapsed_time": "3:08:22", "remaining_time": "0:47:05"}
|
106 |
+
{"current_steps": 910, "total_steps": 1125, "loss": 0.0507, "accuracy": 0.9750000238418579, "lr": 5.364674616415547e-07, "epoch": 0.8088888888888889, "percentage": 80.89, "elapsed_time": "3:09:27", "remaining_time": "0:44:45"}
|
107 |
+
{"current_steps": 920, "total_steps": 1125, "loss": 0.0881, "accuracy": 0.9500000476837158, "lr": 4.893837212719859e-07, "epoch": 0.8177777777777778, "percentage": 81.78, "elapsed_time": "3:10:31", "remaining_time": "0:42:27"}
|
108 |
+
{"current_steps": 930, "total_steps": 1125, "loss": 0.1373, "accuracy": 0.949999988079071, "lr": 4.442374422065493e-07, "epoch": 0.8266666666666667, "percentage": 82.67, "elapsed_time": "3:11:36", "remaining_time": "0:40:10"}
|
109 |
+
{"current_steps": 940, "total_steps": 1125, "loss": 0.2763, "accuracy": 0.925000011920929, "lr": 4.0107212810610974e-07, "epoch": 0.8355555555555556, "percentage": 83.56, "elapsed_time": "3:12:41", "remaining_time": "0:37:55"}
|
110 |
+
{"current_steps": 950, "total_steps": 1125, "loss": 0.1758, "accuracy": 0.949999988079071, "lr": 3.599293737426932e-07, "epoch": 0.8444444444444444, "percentage": 84.44, "elapsed_time": "3:13:46", "remaining_time": "0:35:41"}
|
111 |
+
{"current_steps": 960, "total_steps": 1125, "loss": 0.1138, "accuracy": 0.9624999761581421, "lr": 3.208488249181216e-07, "epoch": 0.8533333333333334, "percentage": 85.33, "elapsed_time": "3:14:51", "remaining_time": "0:33:29"}
|
112 |
+
{"current_steps": 960, "total_steps": 1125, "eval_loss": 0.1441129744052887, "epoch": 0.8533333333333334, "percentage": 85.33, "elapsed_time": "3:20:53", "remaining_time": "0:34:31"}
|
113 |
+
{"current_steps": 970, "total_steps": 1125, "loss": 0.2425, "accuracy": 0.925000011920929, "lr": 2.838681402606952e-07, "epoch": 0.8622222222222222, "percentage": 86.22, "elapsed_time": "3:21:57", "remaining_time": "0:32:16"}
|
114 |
+
{"current_steps": 980, "total_steps": 1125, "loss": 0.0609, "accuracy": 0.9750000238418579, "lr": 2.490229549367443e-07, "epoch": 0.8711111111111111, "percentage": 87.11, "elapsed_time": "3:23:02", "remaining_time": "0:30:02"}
|
115 |
+
{"current_steps": 990, "total_steps": 1125, "loss": 0.3464, "accuracy": 0.8999999761581421, "lr": 2.1634684631203412e-07, "epoch": 0.88, "percentage": 88.0, "elapsed_time": "3:24:07", "remaining_time": "0:27:50"}
|
116 |
+
{"current_steps": 1000, "total_steps": 1125, "loss": 0.0063, "accuracy": 1.0, "lr": 1.8587130159608196e-07, "epoch": 0.8888888888888888, "percentage": 88.89, "elapsed_time": "3:25:12", "remaining_time": "0:25:39"}
|
117 |
+
{"current_steps": 1010, "total_steps": 1125, "loss": 0.2741, "accuracy": 0.925000011920929, "lr": 1.5762568750059604e-07, "epoch": 0.8977777777777778, "percentage": 89.78, "elapsed_time": "3:26:18", "remaining_time": "0:23:29"}
|
118 |
+
{"current_steps": 1020, "total_steps": 1125, "loss": 0.0513, "accuracy": 0.9750000238418579, "lr": 1.316372219412454e-07, "epoch": 0.9066666666666666, "percentage": 90.67, "elapsed_time": "3:27:23", "remaining_time": "0:21:20"}
|
119 |
+
{"current_steps": 1020, "total_steps": 1125, "eval_loss": 0.14118175208568573, "epoch": 0.9066666666666666, "percentage": 90.67, "elapsed_time": "3:33:25", "remaining_time": "0:21:58"}
|
120 |
+
{"current_steps": 1030, "total_steps": 1125, "loss": 0.0349, "accuracy": 0.987500011920929, "lr": 1.0793094781005792e-07, "epoch": 0.9155555555555556, "percentage": 91.56, "elapsed_time": "3:34:29", "remaining_time": "0:19:47"}
|
121 |
+
{"current_steps": 1040, "total_steps": 1125, "loss": 0.1927, "accuracy": 0.9500000476837158, "lr": 8.652970884369255e-08, "epoch": 0.9244444444444444, "percentage": 92.44, "elapsed_time": "3:35:34", "remaining_time": "0:17:37"}
|
122 |
+
{"current_steps": 1050, "total_steps": 1125, "loss": 0.1317, "accuracy": 0.9750000238418579, "lr": 6.745412761086007e-08, "epoch": 0.9333333333333333, "percentage": 93.33, "elapsed_time": "3:36:39", "remaining_time": "0:15:28"}
|
123 |
+
{"current_steps": 1060, "total_steps": 1125, "loss": 0.2279, "accuracy": 0.9375, "lr": 5.0722585640090305e-08, "epoch": 0.9422222222222222, "percentage": 94.22, "elapsed_time": "3:37:44", "remaining_time": "0:13:21"}
|
124 |
+
{"current_steps": 1070, "total_steps": 1125, "loss": 0.1021, "accuracy": 0.9624999761581421, "lr": 3.635120570700784e-08, "epoch": 0.9511111111111111, "percentage": 95.11, "elapsed_time": "3:38:49", "remaining_time": "0:11:14"}
|
125 |
+
{"current_steps": 1080, "total_steps": 1125, "loss": 0.1189, "accuracy": 0.9750000238418579, "lr": 2.4353836298169343e-08, "epoch": 0.96, "percentage": 96.0, "elapsed_time": "3:39:54", "remaining_time": "0:09:09"}
|
126 |
+
{"current_steps": 1080, "total_steps": 1125, "eval_loss": 0.15081512928009033, "epoch": 0.96, "percentage": 96.0, "elapsed_time": "3:45:55", "remaining_time": "0:09:24"}
|
127 |
+
{"current_steps": 1090, "total_steps": 1125, "loss": 0.0728, "accuracy": 0.9750000238418579, "lr": 1.4742038266447046e-08, "epoch": 0.9688888888888889, "percentage": 96.89, "elapsed_time": "3:47:00", "remaining_time": "0:07:17"}
|
128 |
+
{"current_steps": 1100, "total_steps": 1125, "loss": 0.0357, "accuracy": 0.9750000238418579, "lr": 7.525073690809737e-09, "epoch": 0.9777777777777777, "percentage": 97.78, "elapsed_time": "3:48:05", "remaining_time": "0:05:11"}
|
129 |
+
{"current_steps": 1110, "total_steps": 1125, "loss": 0.1973, "accuracy": 0.949999988079071, "lr": 2.709896951238744e-09, "epoch": 0.9866666666666667, "percentage": 98.67, "elapsed_time": "3:49:10", "remaining_time": "0:03:05"}
|
130 |
+
{"current_steps": 1120, "total_steps": 1125, "loss": 0.0538, "accuracy": 0.987500011920929, "lr": 3.0114802737818415e-10, "epoch": 0.9955555555555555, "percentage": 99.56, "elapsed_time": "3:50:15", "remaining_time": "0:01:01"}
|
131 |
+
{"current_steps": 1125, "total_steps": 1125, "epoch": 1.0, "percentage": 100.0, "elapsed_time": "3:50:48", "remaining_time": "0:00:00"}
|
trainer_state.json
ADDED
@@ -0,0 +1,2010 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.0,
|
5 |
+
"eval_steps": 60,
|
6 |
+
"global_step": 1125,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.008888888888888889,
|
13 |
+
"grad_norm": 8.760091781616211,
|
14 |
+
"learning_rate": 4.4247787610619474e-07,
|
15 |
+
"logits/chosen": -0.8248252868652344,
|
16 |
+
"logits/rejected": -0.8263720273971558,
|
17 |
+
"logps/chosen": -0.36086463928222656,
|
18 |
+
"logps/rejected": -5.696224689483643,
|
19 |
+
"loss": 1.1038,
|
20 |
+
"rewards/accuracies": 0.5125000476837158,
|
21 |
+
"rewards/chosen": 17.43745994567871,
|
22 |
+
"rewards/margins": 0.5984855890274048,
|
23 |
+
"rewards/rejected": 16.838973999023438,
|
24 |
+
"step": 10
|
25 |
+
},
|
26 |
+
{
|
27 |
+
"epoch": 0.017777777777777778,
|
28 |
+
"grad_norm": 8.855981826782227,
|
29 |
+
"learning_rate": 8.849557522123895e-07,
|
30 |
+
"logits/chosen": -0.8169006109237671,
|
31 |
+
"logits/rejected": -0.819770872592926,
|
32 |
+
"logps/chosen": -0.12464660406112671,
|
33 |
+
"logps/rejected": -7.139842987060547,
|
34 |
+
"loss": 1.1887,
|
35 |
+
"rewards/accuracies": 0.4000000059604645,
|
36 |
+
"rewards/chosen": 17.17649269104004,
|
37 |
+
"rewards/margins": 0.19107049703598022,
|
38 |
+
"rewards/rejected": 16.98542022705078,
|
39 |
+
"step": 20
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 0.02666666666666667,
|
43 |
+
"grad_norm": 16.764184951782227,
|
44 |
+
"learning_rate": 1.3274336283185843e-06,
|
45 |
+
"logits/chosen": -0.8003113865852356,
|
46 |
+
"logits/rejected": -0.8030117750167847,
|
47 |
+
"logps/chosen": -0.34651467204093933,
|
48 |
+
"logps/rejected": -6.967917442321777,
|
49 |
+
"loss": 1.0563,
|
50 |
+
"rewards/accuracies": 0.44999998807907104,
|
51 |
+
"rewards/chosen": 17.280975341796875,
|
52 |
+
"rewards/margins": 0.40005987882614136,
|
53 |
+
"rewards/rejected": 16.88091468811035,
|
54 |
+
"step": 30
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 0.035555555555555556,
|
58 |
+
"grad_norm": 8.33682918548584,
|
59 |
+
"learning_rate": 1.769911504424779e-06,
|
60 |
+
"logits/chosen": -0.7695047855377197,
|
61 |
+
"logits/rejected": -0.7739207148551941,
|
62 |
+
"logps/chosen": -1.5993006229400635,
|
63 |
+
"logps/rejected": -8.504932403564453,
|
64 |
+
"loss": 0.7596,
|
65 |
+
"rewards/accuracies": 0.5,
|
66 |
+
"rewards/chosen": 17.283912658691406,
|
67 |
+
"rewards/margins": 0.6976072192192078,
|
68 |
+
"rewards/rejected": 16.5863037109375,
|
69 |
+
"step": 40
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 0.044444444444444446,
|
73 |
+
"grad_norm": 4.494723320007324,
|
74 |
+
"learning_rate": 2.212389380530974e-06,
|
75 |
+
"logits/chosen": -0.7154140472412109,
|
76 |
+
"logits/rejected": -0.7225576043128967,
|
77 |
+
"logps/chosen": -3.112199068069458,
|
78 |
+
"logps/rejected": -12.212080001831055,
|
79 |
+
"loss": 0.6083,
|
80 |
+
"rewards/accuracies": 0.4625000059604645,
|
81 |
+
"rewards/chosen": 17.03064727783203,
|
82 |
+
"rewards/margins": 0.7148451805114746,
|
83 |
+
"rewards/rejected": 16.3158016204834,
|
84 |
+
"step": 50
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.05333333333333334,
|
88 |
+
"grad_norm": 5.110287666320801,
|
89 |
+
"learning_rate": 2.6548672566371687e-06,
|
90 |
+
"logits/chosen": -0.6322453022003174,
|
91 |
+
"logits/rejected": -0.6387485265731812,
|
92 |
+
"logps/chosen": -5.650620460510254,
|
93 |
+
"logps/rejected": -12.759811401367188,
|
94 |
+
"loss": 0.3835,
|
95 |
+
"rewards/accuracies": 0.9125000238418579,
|
96 |
+
"rewards/chosen": 17.101289749145508,
|
97 |
+
"rewards/margins": 1.1824612617492676,
|
98 |
+
"rewards/rejected": 15.918828964233398,
|
99 |
+
"step": 60
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.05333333333333334,
|
103 |
+
"eval_logits/chosen": -0.5826543569564819,
|
104 |
+
"eval_logits/rejected": -0.5914276838302612,
|
105 |
+
"eval_logps/chosen": -3.5471787452697754,
|
106 |
+
"eval_logps/rejected": -16.51181983947754,
|
107 |
+
"eval_loss": 0.3286525011062622,
|
108 |
+
"eval_rewards/accuracies": 0.9280000925064087,
|
109 |
+
"eval_rewards/chosen": 17.148174285888672,
|
110 |
+
"eval_rewards/margins": 1.4386365413665771,
|
111 |
+
"eval_rewards/rejected": 15.709539413452148,
|
112 |
+
"eval_runtime": 372.0227,
|
113 |
+
"eval_samples_per_second": 2.688,
|
114 |
+
"eval_steps_per_second": 0.336,
|
115 |
+
"step": 60
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 0.06222222222222222,
|
119 |
+
"grad_norm": 5.098133563995361,
|
120 |
+
"learning_rate": 3.097345132743363e-06,
|
121 |
+
"logits/chosen": -0.5378152132034302,
|
122 |
+
"logits/rejected": -0.5494933724403381,
|
123 |
+
"logps/chosen": -1.5099802017211914,
|
124 |
+
"logps/rejected": -21.206321716308594,
|
125 |
+
"loss": 0.2931,
|
126 |
+
"rewards/accuracies": 0.9375,
|
127 |
+
"rewards/chosen": 17.083791732788086,
|
128 |
+
"rewards/margins": 1.5844331979751587,
|
129 |
+
"rewards/rejected": 15.499359130859375,
|
130 |
+
"step": 70
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 0.07111111111111111,
|
134 |
+
"grad_norm": 29.787437438964844,
|
135 |
+
"learning_rate": 3.539823008849558e-06,
|
136 |
+
"logits/chosen": -0.443774938583374,
|
137 |
+
"logits/rejected": -0.45571577548980713,
|
138 |
+
"logps/chosen": -1.5804342031478882,
|
139 |
+
"logps/rejected": -22.606929779052734,
|
140 |
+
"loss": 0.202,
|
141 |
+
"rewards/accuracies": 0.9750000238418579,
|
142 |
+
"rewards/chosen": 17.302125930786133,
|
143 |
+
"rewards/margins": 2.174014091491699,
|
144 |
+
"rewards/rejected": 15.128110885620117,
|
145 |
+
"step": 80
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"epoch": 0.08,
|
149 |
+
"grad_norm": 23.14398193359375,
|
150 |
+
"learning_rate": 3.982300884955752e-06,
|
151 |
+
"logits/chosen": -0.3626072406768799,
|
152 |
+
"logits/rejected": -0.3787815570831299,
|
153 |
+
"logps/chosen": -2.203828811645508,
|
154 |
+
"logps/rejected": -29.433551788330078,
|
155 |
+
"loss": 0.2123,
|
156 |
+
"rewards/accuracies": 0.925000011920929,
|
157 |
+
"rewards/chosen": 17.00284194946289,
|
158 |
+
"rewards/margins": 2.320391893386841,
|
159 |
+
"rewards/rejected": 14.682450294494629,
|
160 |
+
"step": 90
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 0.08888888888888889,
|
164 |
+
"grad_norm": 29.672739028930664,
|
165 |
+
"learning_rate": 4.424778761061948e-06,
|
166 |
+
"logits/chosen": -0.3035663962364197,
|
167 |
+
"logits/rejected": -0.31762221455574036,
|
168 |
+
"logps/chosen": -3.433589458465576,
|
169 |
+
"logps/rejected": -29.9322509765625,
|
170 |
+
"loss": 0.2592,
|
171 |
+
"rewards/accuracies": 0.9375,
|
172 |
+
"rewards/chosen": 16.929956436157227,
|
173 |
+
"rewards/margins": 2.31272029876709,
|
174 |
+
"rewards/rejected": 14.617237091064453,
|
175 |
+
"step": 100
|
176 |
+
},
|
177 |
+
{
|
178 |
+
"epoch": 0.09777777777777778,
|
179 |
+
"grad_norm": 1.873722791671753,
|
180 |
+
"learning_rate": 4.867256637168142e-06,
|
181 |
+
"logits/chosen": -0.2679600715637207,
|
182 |
+
"logits/rejected": -0.2826440930366516,
|
183 |
+
"logps/chosen": -0.9653514623641968,
|
184 |
+
"logps/rejected": -30.235322952270508,
|
185 |
+
"loss": 0.1336,
|
186 |
+
"rewards/accuracies": 0.949999988079071,
|
187 |
+
"rewards/chosen": 17.462385177612305,
|
188 |
+
"rewards/margins": 3.1994175910949707,
|
189 |
+
"rewards/rejected": 14.26296615600586,
|
190 |
+
"step": 110
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 0.10666666666666667,
|
194 |
+
"grad_norm": 1.6913721561431885,
|
195 |
+
"learning_rate": 4.999409761242696e-06,
|
196 |
+
"logits/chosen": -0.22222033143043518,
|
197 |
+
"logits/rejected": -0.23720571398735046,
|
198 |
+
"logps/chosen": -4.4953508377075195,
|
199 |
+
"logps/rejected": -34.074745178222656,
|
200 |
+
"loss": 0.2552,
|
201 |
+
"rewards/accuracies": 0.8999999761581421,
|
202 |
+
"rewards/chosen": 17.04866600036621,
|
203 |
+
"rewards/margins": 3.1014418601989746,
|
204 |
+
"rewards/rejected": 13.947224617004395,
|
205 |
+
"step": 120
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.10666666666666667,
|
209 |
+
"eval_logits/chosen": -0.206527978181839,
|
210 |
+
"eval_logits/rejected": -0.22178640961647034,
|
211 |
+
"eval_logps/chosen": -3.69442081451416,
|
212 |
+
"eval_logps/rejected": -36.072166442871094,
|
213 |
+
"eval_loss": 0.18996010720729828,
|
214 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
215 |
+
"eval_rewards/chosen": 17.133451461791992,
|
216 |
+
"eval_rewards/margins": 3.379946708679199,
|
217 |
+
"eval_rewards/rejected": 13.753504753112793,
|
218 |
+
"eval_runtime": 361.5279,
|
219 |
+
"eval_samples_per_second": 2.766,
|
220 |
+
"eval_steps_per_second": 0.346,
|
221 |
+
"step": 120
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 0.11555555555555555,
|
225 |
+
"grad_norm": 61.80262756347656,
|
226 |
+
"learning_rate": 4.996519466816778e-06,
|
227 |
+
"logits/chosen": -0.18473535776138306,
|
228 |
+
"logits/rejected": -0.1988501250743866,
|
229 |
+
"logps/chosen": -3.7009687423706055,
|
230 |
+
"logps/rejected": -39.289939880371094,
|
231 |
+
"loss": 0.1394,
|
232 |
+
"rewards/accuracies": 0.9624999761581421,
|
233 |
+
"rewards/chosen": 17.106964111328125,
|
234 |
+
"rewards/margins": 3.633338212966919,
|
235 |
+
"rewards/rejected": 13.473625183105469,
|
236 |
+
"step": 130
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 0.12444444444444444,
|
240 |
+
"grad_norm": 1.6732702255249023,
|
241 |
+
"learning_rate": 4.9912234871722805e-06,
|
242 |
+
"logits/chosen": -0.16134041547775269,
|
243 |
+
"logits/rejected": -0.17547868192195892,
|
244 |
+
"logps/chosen": -3.0637736320495605,
|
245 |
+
"logps/rejected": -40.07548522949219,
|
246 |
+
"loss": 0.1408,
|
247 |
+
"rewards/accuracies": 0.9750000238418579,
|
248 |
+
"rewards/chosen": 17.392223358154297,
|
249 |
+
"rewards/margins": 4.242353439331055,
|
250 |
+
"rewards/rejected": 13.149867057800293,
|
251 |
+
"step": 140
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 0.13333333333333333,
|
255 |
+
"grad_norm": 0.346453994512558,
|
256 |
+
"learning_rate": 4.98352692559805e-06,
|
257 |
+
"logits/chosen": -0.13797929883003235,
|
258 |
+
"logits/rejected": -0.15283086895942688,
|
259 |
+
"logps/chosen": -5.14492130279541,
|
260 |
+
"logps/rejected": -47.97212219238281,
|
261 |
+
"loss": 0.2153,
|
262 |
+
"rewards/accuracies": 0.9375,
|
263 |
+
"rewards/chosen": 16.896778106689453,
|
264 |
+
"rewards/margins": 4.227695465087891,
|
265 |
+
"rewards/rejected": 12.669081687927246,
|
266 |
+
"step": 150
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 0.14222222222222222,
|
270 |
+
"grad_norm": 0.21871662139892578,
|
271 |
+
"learning_rate": 4.973437198621237e-06,
|
272 |
+
"logits/chosen": -0.12396670132875443,
|
273 |
+
"logits/rejected": -0.13780555129051208,
|
274 |
+
"logps/chosen": -6.108860015869141,
|
275 |
+
"logps/rejected": -54.90739440917969,
|
276 |
+
"loss": 0.0388,
|
277 |
+
"rewards/accuracies": 0.9750000238418579,
|
278 |
+
"rewards/chosen": 16.75935935974121,
|
279 |
+
"rewards/margins": 4.755282878875732,
|
280 |
+
"rewards/rejected": 12.004077911376953,
|
281 |
+
"step": 160
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.1511111111111111,
|
285 |
+
"grad_norm": 235.12429809570312,
|
286 |
+
"learning_rate": 4.960964028860621e-06,
|
287 |
+
"logits/chosen": -0.1140839159488678,
|
288 |
+
"logits/rejected": -0.1263057291507721,
|
289 |
+
"logps/chosen": -12.605452537536621,
|
290 |
+
"logps/rejected": -53.81230926513672,
|
291 |
+
"loss": 0.4651,
|
292 |
+
"rewards/accuracies": 0.875,
|
293 |
+
"rewards/chosen": 16.101238250732422,
|
294 |
+
"rewards/margins": 3.9864249229431152,
|
295 |
+
"rewards/rejected": 12.114812850952148,
|
296 |
+
"step": 170
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.16,
|
300 |
+
"grad_norm": 190.97048950195312,
|
301 |
+
"learning_rate": 4.946119435657738e-06,
|
302 |
+
"logits/chosen": -0.10746976733207703,
|
303 |
+
"logits/rejected": -0.11878640949726105,
|
304 |
+
"logps/chosen": -8.5105562210083,
|
305 |
+
"logps/rejected": -51.314781188964844,
|
306 |
+
"loss": 0.2362,
|
307 |
+
"rewards/accuracies": 0.925000011920929,
|
308 |
+
"rewards/chosen": 16.719980239868164,
|
309 |
+
"rewards/margins": 4.549674034118652,
|
310 |
+
"rewards/rejected": 12.170306205749512,
|
311 |
+
"step": 180
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 0.16,
|
315 |
+
"eval_logits/chosen": -0.10870806127786636,
|
316 |
+
"eval_logits/rejected": -0.12223993986845016,
|
317 |
+
"eval_logps/chosen": -4.414996147155762,
|
318 |
+
"eval_logps/rejected": -53.885032653808594,
|
319 |
+
"eval_loss": 0.20236633718013763,
|
320 |
+
"eval_rewards/accuracies": 0.9510000944137573,
|
321 |
+
"eval_rewards/chosen": 17.06139373779297,
|
322 |
+
"eval_rewards/margins": 5.089176177978516,
|
323 |
+
"eval_rewards/rejected": 11.97221851348877,
|
324 |
+
"eval_runtime": 361.4355,
|
325 |
+
"eval_samples_per_second": 2.767,
|
326 |
+
"eval_steps_per_second": 0.346,
|
327 |
+
"step": 180
|
328 |
+
},
|
329 |
+
{
|
330 |
+
"epoch": 0.1688888888888889,
|
331 |
+
"grad_norm": 56.81266784667969,
|
332 |
+
"learning_rate": 4.928917723494854e-06,
|
333 |
+
"logits/chosen": -0.10682469606399536,
|
334 |
+
"logits/rejected": -0.12124393880367279,
|
335 |
+
"logps/chosen": -3.058413028717041,
|
336 |
+
"logps/rejected": -55.052528381347656,
|
337 |
+
"loss": 0.2442,
|
338 |
+
"rewards/accuracies": 0.9500000476837158,
|
339 |
+
"rewards/chosen": 17.058589935302734,
|
340 |
+
"rewards/margins": 5.056097984313965,
|
341 |
+
"rewards/rejected": 12.002490043640137,
|
342 |
+
"step": 190
|
343 |
+
},
|
344 |
+
{
|
345 |
+
"epoch": 0.17777777777777778,
|
346 |
+
"grad_norm": 175.06552124023438,
|
347 |
+
"learning_rate": 4.909375468210947e-06,
|
348 |
+
"logits/chosen": -0.10520349442958832,
|
349 |
+
"logits/rejected": -0.12018950283527374,
|
350 |
+
"logps/chosen": -4.114959716796875,
|
351 |
+
"logps/rejected": -55.9394645690918,
|
352 |
+
"loss": 0.1915,
|
353 |
+
"rewards/accuracies": 0.9500000476837158,
|
354 |
+
"rewards/chosen": 16.98603057861328,
|
355 |
+
"rewards/margins": 5.105838775634766,
|
356 |
+
"rewards/rejected": 11.880191802978516,
|
357 |
+
"step": 200
|
358 |
+
},
|
359 |
+
{
|
360 |
+
"epoch": 0.18666666666666668,
|
361 |
+
"grad_norm": 78.06558990478516,
|
362 |
+
"learning_rate": 4.8875115010289655e-06,
|
363 |
+
"logits/chosen": -0.10475558042526245,
|
364 |
+
"logits/rejected": -0.11949175596237183,
|
365 |
+
"logps/chosen": -6.760301113128662,
|
366 |
+
"logps/rejected": -53.91607666015625,
|
367 |
+
"loss": 0.2843,
|
368 |
+
"rewards/accuracies": 0.9375,
|
369 |
+
"rewards/chosen": 16.857545852661133,
|
370 |
+
"rewards/margins": 4.917357921600342,
|
371 |
+
"rewards/rejected": 11.94018840789795,
|
372 |
+
"step": 210
|
373 |
+
},
|
374 |
+
{
|
375 |
+
"epoch": 0.19555555555555557,
|
376 |
+
"grad_norm": 15.880486488342285,
|
377 |
+
"learning_rate": 4.863346890409768e-06,
|
378 |
+
"logits/chosen": -0.11213523149490356,
|
379 |
+
"logits/rejected": -0.12581588327884674,
|
380 |
+
"logps/chosen": -6.759585380554199,
|
381 |
+
"logps/rejected": -51.10936737060547,
|
382 |
+
"loss": 0.5104,
|
383 |
+
"rewards/accuracies": 0.875,
|
384 |
+
"rewards/chosen": 16.859071731567383,
|
385 |
+
"rewards/margins": 4.638372898101807,
|
386 |
+
"rewards/rejected": 12.220698356628418,
|
387 |
+
"step": 220
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.20444444444444446,
|
391 |
+
"grad_norm": 46.97845458984375,
|
392 |
+
"learning_rate": 4.836904921750224e-06,
|
393 |
+
"logits/chosen": -0.11947059631347656,
|
394 |
+
"logits/rejected": -0.1329912692308426,
|
395 |
+
"logps/chosen": -3.608184814453125,
|
396 |
+
"logps/rejected": -48.794761657714844,
|
397 |
+
"loss": 0.2134,
|
398 |
+
"rewards/accuracies": 0.925000011920929,
|
399 |
+
"rewards/chosen": 17.235904693603516,
|
400 |
+
"rewards/margins": 4.859888076782227,
|
401 |
+
"rewards/rejected": 12.376014709472656,
|
402 |
+
"step": 230
|
403 |
+
},
|
404 |
+
{
|
405 |
+
"epoch": 0.21333333333333335,
|
406 |
+
"grad_norm": 24.032859802246094,
|
407 |
+
"learning_rate": 4.808211074945042e-06,
|
408 |
+
"logits/chosen": -0.1200513243675232,
|
409 |
+
"logits/rejected": -0.1333036869764328,
|
410 |
+
"logps/chosen": -3.7552154064178467,
|
411 |
+
"logps/rejected": -49.87453079223633,
|
412 |
+
"loss": 0.1781,
|
413 |
+
"rewards/accuracies": 0.9500000476837158,
|
414 |
+
"rewards/chosen": 17.094650268554688,
|
415 |
+
"rewards/margins": 4.68077278137207,
|
416 |
+
"rewards/rejected": 12.41387939453125,
|
417 |
+
"step": 240
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 0.21333333333333335,
|
421 |
+
"eval_logits/chosen": -0.12433278560638428,
|
422 |
+
"eval_logits/rejected": -0.13808581233024597,
|
423 |
+
"eval_logps/chosen": -4.408891201019287,
|
424 |
+
"eval_logps/rejected": -50.744781494140625,
|
425 |
+
"eval_loss": 0.1546352356672287,
|
426 |
+
"eval_rewards/accuracies": 0.9500000476837158,
|
427 |
+
"eval_rewards/chosen": 17.06200408935547,
|
428 |
+
"eval_rewards/margins": 4.775761604309082,
|
429 |
+
"eval_rewards/rejected": 12.286243438720703,
|
430 |
+
"eval_runtime": 361.4974,
|
431 |
+
"eval_samples_per_second": 2.766,
|
432 |
+
"eval_steps_per_second": 0.346,
|
433 |
+
"step": 240
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 0.2222222222222222,
|
437 |
+
"grad_norm": 0.25737640261650085,
|
438 |
+
"learning_rate": 4.7772929998339485e-06,
|
439 |
+
"logits/chosen": -0.12348780035972595,
|
440 |
+
"logits/rejected": -0.13704943656921387,
|
441 |
+
"logps/chosen": -4.4299187660217285,
|
442 |
+
"logps/rejected": -53.074607849121094,
|
443 |
+
"loss": 0.1373,
|
444 |
+
"rewards/accuracies": 0.9375,
|
445 |
+
"rewards/chosen": 17.087068557739258,
|
446 |
+
"rewards/margins": 5.06691837310791,
|
447 |
+
"rewards/rejected": 12.020149230957031,
|
448 |
+
"step": 250
|
449 |
+
},
|
450 |
+
{
|
451 |
+
"epoch": 0.2311111111111111,
|
452 |
+
"grad_norm": 0.1839389204978943,
|
453 |
+
"learning_rate": 4.744180489557859e-06,
|
454 |
+
"logits/chosen": -0.12177034467458725,
|
455 |
+
"logits/rejected": -0.1342695653438568,
|
456 |
+
"logps/chosen": -3.775188446044922,
|
457 |
+
"logps/rejected": -53.98720932006836,
|
458 |
+
"loss": 0.1896,
|
459 |
+
"rewards/accuracies": 0.949999988079071,
|
460 |
+
"rewards/chosen": 17.12021255493164,
|
461 |
+
"rewards/margins": 5.148064613342285,
|
462 |
+
"rewards/rejected": 11.972146987915039,
|
463 |
+
"step": 260
|
464 |
+
},
|
465 |
+
{
|
466 |
+
"epoch": 0.24,
|
467 |
+
"grad_norm": 12.258485794067383,
|
468 |
+
"learning_rate": 4.708905451849754e-06,
|
469 |
+
"logits/chosen": -0.11067859083414078,
|
470 |
+
"logits/rejected": -0.12377731502056122,
|
471 |
+
"logps/chosen": -6.418317794799805,
|
472 |
+
"logps/rejected": -56.57402801513672,
|
473 |
+
"loss": 0.2315,
|
474 |
+
"rewards/accuracies": 0.9375,
|
475 |
+
"rewards/chosen": 16.738832473754883,
|
476 |
+
"rewards/margins": 4.884931564331055,
|
477 |
+
"rewards/rejected": 11.853900909423828,
|
478 |
+
"step": 270
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.24888888888888888,
|
482 |
+
"grad_norm": 77.56194305419922,
|
483 |
+
"learning_rate": 4.671501878287879e-06,
|
484 |
+
"logits/chosen": -0.1184445172548294,
|
485 |
+
"logits/rejected": -0.1339874565601349,
|
486 |
+
"logps/chosen": -10.12116527557373,
|
487 |
+
"logps/rejected": -53.403907775878906,
|
488 |
+
"loss": 0.5343,
|
489 |
+
"rewards/accuracies": 0.862500011920929,
|
490 |
+
"rewards/chosen": 16.458633422851562,
|
491 |
+
"rewards/margins": 4.402472496032715,
|
492 |
+
"rewards/rejected": 12.056160926818848,
|
493 |
+
"step": 280
|
494 |
+
},
|
495 |
+
{
|
496 |
+
"epoch": 0.2577777777777778,
|
497 |
+
"grad_norm": 67.53883361816406,
|
498 |
+
"learning_rate": 4.6320058115409295e-06,
|
499 |
+
"logits/chosen": -0.1448262631893158,
|
500 |
+
"logits/rejected": -0.15793387591838837,
|
501 |
+
"logps/chosen": -3.4666190147399902,
|
502 |
+
"logps/rejected": -48.79213333129883,
|
503 |
+
"loss": 0.5017,
|
504 |
+
"rewards/accuracies": 0.887499988079071,
|
505 |
+
"rewards/chosen": 16.945899963378906,
|
506 |
+
"rewards/margins": 4.2686333656311035,
|
507 |
+
"rewards/rejected": 12.677268028259277,
|
508 |
+
"step": 290
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 0.26666666666666666,
|
512 |
+
"grad_norm": 0.17521341145038605,
|
513 |
+
"learning_rate": 4.590455310636778e-06,
|
514 |
+
"logits/chosen": -0.16128253936767578,
|
515 |
+
"logits/rejected": -0.17375555634498596,
|
516 |
+
"logps/chosen": -2.9032950401306152,
|
517 |
+
"logps/rejected": -47.69734191894531,
|
518 |
+
"loss": 0.265,
|
519 |
+
"rewards/accuracies": 0.925000011920929,
|
520 |
+
"rewards/chosen": 17.18383026123047,
|
521 |
+
"rewards/margins": 4.541309356689453,
|
522 |
+
"rewards/rejected": 12.642518997192383,
|
523 |
+
"step": 300
|
524 |
+
},
|
525 |
+
{
|
526 |
+
"epoch": 0.26666666666666666,
|
527 |
+
"eval_logits/chosen": -0.17444846034049988,
|
528 |
+
"eval_logits/rejected": -0.18559777736663818,
|
529 |
+
"eval_logps/chosen": -2.535512924194336,
|
530 |
+
"eval_logps/rejected": -47.16367721557617,
|
531 |
+
"eval_loss": 0.15360687673091888,
|
532 |
+
"eval_rewards/accuracies": 0.9440000653266907,
|
533 |
+
"eval_rewards/chosen": 17.249343872070312,
|
534 |
+
"eval_rewards/margins": 4.604989051818848,
|
535 |
+
"eval_rewards/rejected": 12.644353866577148,
|
536 |
+
"eval_runtime": 361.4575,
|
537 |
+
"eval_samples_per_second": 2.767,
|
538 |
+
"eval_steps_per_second": 0.346,
|
539 |
+
"step": 300
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 0.27555555555555555,
|
543 |
+
"grad_norm": 0.5040452480316162,
|
544 |
+
"learning_rate": 4.54689041428819e-06,
|
545 |
+
"logits/chosen": -0.16974106431007385,
|
546 |
+
"logits/rejected": -0.1810058057308197,
|
547 |
+
"logps/chosen": -1.233938217163086,
|
548 |
+
"logps/rejected": -49.907745361328125,
|
549 |
+
"loss": 0.1132,
|
550 |
+
"rewards/accuracies": 0.9500000476837158,
|
551 |
+
"rewards/chosen": 17.34117889404297,
|
552 |
+
"rewards/margins": 4.934173583984375,
|
553 |
+
"rewards/rejected": 12.407005310058594,
|
554 |
+
"step": 310
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 0.28444444444444444,
|
558 |
+
"grad_norm": 100.02949523925781,
|
559 |
+
"learning_rate": 4.501353102310901e-06,
|
560 |
+
"logits/chosen": -0.15705889463424683,
|
561 |
+
"logits/rejected": -0.1695334017276764,
|
562 |
+
"logps/chosen": -1.0820492506027222,
|
563 |
+
"logps/rejected": -52.577110290527344,
|
564 |
+
"loss": 0.1194,
|
565 |
+
"rewards/accuracies": 0.9500000476837158,
|
566 |
+
"rewards/chosen": 17.33388900756836,
|
567 |
+
"rewards/margins": 5.154760837554932,
|
568 |
+
"rewards/rejected": 12.179126739501953,
|
569 |
+
"step": 320
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.29333333333333333,
|
573 |
+
"grad_norm": 0.2689219117164612,
|
574 |
+
"learning_rate": 4.453887255171206e-06,
|
575 |
+
"logits/chosen": -0.13849371671676636,
|
576 |
+
"logits/rejected": -0.14990833401679993,
|
577 |
+
"logps/chosen": -1.8435032367706299,
|
578 |
+
"logps/rejected": -54.79044723510742,
|
579 |
+
"loss": 0.0926,
|
580 |
+
"rewards/accuracies": 0.9500000476837158,
|
581 |
+
"rewards/chosen": 17.2423095703125,
|
582 |
+
"rewards/margins": 5.28987979888916,
|
583 |
+
"rewards/rejected": 11.952428817749023,
|
584 |
+
"step": 330
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 0.3022222222222222,
|
588 |
+
"grad_norm": 0.09305431693792343,
|
589 |
+
"learning_rate": 4.404538611702055e-06,
|
590 |
+
"logits/chosen": -0.12299702316522598,
|
591 |
+
"logits/rejected": -0.13453055918216705,
|
592 |
+
"logps/chosen": -2.9897143840789795,
|
593 |
+
"logps/rejected": -52.954498291015625,
|
594 |
+
"loss": 0.2873,
|
595 |
+
"rewards/accuracies": 0.925000011920929,
|
596 |
+
"rewards/chosen": 17.17474365234375,
|
597 |
+
"rewards/margins": 5.071004867553711,
|
598 |
+
"rewards/rejected": 12.103739738464355,
|
599 |
+
"step": 340
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 0.3111111111111111,
|
603 |
+
"grad_norm": 59.282073974609375,
|
604 |
+
"learning_rate": 4.3533547250284015e-06,
|
605 |
+
"logits/chosen": -0.11913029849529266,
|
606 |
+
"logits/rejected": -0.12785324454307556,
|
607 |
+
"logps/chosen": -3.9456872940063477,
|
608 |
+
"logps/rejected": -48.68487548828125,
|
609 |
+
"loss": 0.4332,
|
610 |
+
"rewards/accuracies": 0.875,
|
611 |
+
"rewards/chosen": 17.12805938720703,
|
612 |
+
"rewards/margins": 4.669450283050537,
|
613 |
+
"rewards/rejected": 12.458610534667969,
|
614 |
+
"step": 350
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 0.32,
|
618 |
+
"grad_norm": 0.31101909279823303,
|
619 |
+
"learning_rate": 4.300384916744261e-06,
|
620 |
+
"logits/chosen": -0.11280188709497452,
|
621 |
+
"logits/rejected": -0.12300585210323334,
|
622 |
+
"logps/chosen": -2.1714723110198975,
|
623 |
+
"logps/rejected": -54.74174118041992,
|
624 |
+
"loss": 0.1605,
|
625 |
+
"rewards/accuracies": 0.9500000476837158,
|
626 |
+
"rewards/chosen": 17.326162338256836,
|
627 |
+
"rewards/margins": 5.467062473297119,
|
628 |
+
"rewards/rejected": 11.859098434448242,
|
629 |
+
"step": 360
|
630 |
+
},
|
631 |
+
{
|
632 |
+
"epoch": 0.32,
|
633 |
+
"eval_logits/chosen": -0.10620756447315216,
|
634 |
+
"eval_logits/rejected": -0.11727114766836166,
|
635 |
+
"eval_logps/chosen": -1.4165427684783936,
|
636 |
+
"eval_logps/rejected": -50.9525146484375,
|
637 |
+
"eval_loss": 0.3194349706172943,
|
638 |
+
"eval_rewards/accuracies": 0.9210000038146973,
|
639 |
+
"eval_rewards/chosen": 17.36124038696289,
|
640 |
+
"eval_rewards/margins": 5.095769882202148,
|
641 |
+
"eval_rewards/rejected": 12.26546859741211,
|
642 |
+
"eval_runtime": 361.5072,
|
643 |
+
"eval_samples_per_second": 2.766,
|
644 |
+
"eval_steps_per_second": 0.346,
|
645 |
+
"step": 360
|
646 |
+
},
|
647 |
+
{
|
648 |
+
"epoch": 0.3288888888888889,
|
649 |
+
"grad_norm": 6.1126532554626465,
|
650 |
+
"learning_rate": 4.24568022938566e-06,
|
651 |
+
"logits/chosen": -0.10354311764240265,
|
652 |
+
"logits/rejected": -0.11526636779308319,
|
653 |
+
"logps/chosen": -1.2935255765914917,
|
654 |
+
"logps/rejected": -55.57566833496094,
|
655 |
+
"loss": 0.1711,
|
656 |
+
"rewards/accuracies": 0.9500000476837158,
|
657 |
+
"rewards/chosen": 17.439346313476562,
|
658 |
+
"rewards/margins": 5.700921058654785,
|
659 |
+
"rewards/rejected": 11.738424301147461,
|
660 |
+
"step": 370
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.3377777777777778,
|
664 |
+
"grad_norm": 34.15927505493164,
|
665 |
+
"learning_rate": 4.189293377245241e-06,
|
666 |
+
"logits/chosen": -0.1029932051897049,
|
667 |
+
"logits/rejected": -0.11382515728473663,
|
668 |
+
"logps/chosen": -2.5132687091827393,
|
669 |
+
"logps/rejected": -55.50346374511719,
|
670 |
+
"loss": 0.4359,
|
671 |
+
"rewards/accuracies": 0.8875000476837158,
|
672 |
+
"rewards/chosen": 16.731037139892578,
|
673 |
+
"rewards/margins": 4.368172645568848,
|
674 |
+
"rewards/rejected": 12.362865447998047,
|
675 |
+
"step": 380
|
676 |
+
},
|
677 |
+
{
|
678 |
+
"epoch": 0.3466666666666667,
|
679 |
+
"grad_norm": 2.8422904014587402,
|
680 |
+
"learning_rate": 4.131278695575952e-06,
|
681 |
+
"logits/chosen": -0.10793520510196686,
|
682 |
+
"logits/rejected": -0.12109285593032837,
|
683 |
+
"logps/chosen": -3.014652729034424,
|
684 |
+
"logps/rejected": -53.98411560058594,
|
685 |
+
"loss": 0.2161,
|
686 |
+
"rewards/accuracies": 0.949999988079071,
|
687 |
+
"rewards/chosen": 17.137393951416016,
|
688 |
+
"rewards/margins": 5.105995178222656,
|
689 |
+
"rewards/rejected": 12.03139877319336,
|
690 |
+
"step": 390
|
691 |
+
},
|
692 |
+
{
|
693 |
+
"epoch": 0.35555555555555557,
|
694 |
+
"grad_norm": 54.0329475402832,
|
695 |
+
"learning_rate": 4.071692088232743e-06,
|
696 |
+
"logits/chosen": -0.10393750667572021,
|
697 |
+
"logits/rejected": -0.11834606528282166,
|
698 |
+
"logps/chosen": -2.1508543491363525,
|
699 |
+
"logps/rejected": -45.60733413696289,
|
700 |
+
"loss": 0.2077,
|
701 |
+
"rewards/accuracies": 0.925000011920929,
|
702 |
+
"rewards/chosen": 17.586124420166016,
|
703 |
+
"rewards/margins": 5.077212333679199,
|
704 |
+
"rewards/rejected": 12.5089111328125,
|
705 |
+
"step": 400
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 0.36444444444444446,
|
709 |
+
"grad_norm": 81.61144256591797,
|
710 |
+
"learning_rate": 4.010590973802737e-06,
|
711 |
+
"logits/chosen": -0.09564584493637085,
|
712 |
+
"logits/rejected": -0.10617707669734955,
|
713 |
+
"logps/chosen": -3.4572842121124268,
|
714 |
+
"logps/rejected": -50.92162322998047,
|
715 |
+
"loss": 0.2478,
|
716 |
+
"rewards/accuracies": 0.8875000476837158,
|
717 |
+
"rewards/chosen": 17.010910034179688,
|
718 |
+
"rewards/margins": 4.556198596954346,
|
719 |
+
"rewards/rejected": 12.454713821411133,
|
720 |
+
"step": 410
|
721 |
+
},
|
722 |
+
{
|
723 |
+
"epoch": 0.37333333333333335,
|
724 |
+
"grad_norm": 0.30974289774894714,
|
725 |
+
"learning_rate": 3.948034230275781e-06,
|
726 |
+
"logits/chosen": -0.09134417027235031,
|
727 |
+
"logits/rejected": -0.1020016297698021,
|
728 |
+
"logps/chosen": -5.046698570251465,
|
729 |
+
"logps/rejected": -48.908958435058594,
|
730 |
+
"loss": 0.2894,
|
731 |
+
"rewards/accuracies": 0.8999999761581421,
|
732 |
+
"rewards/chosen": 17.007888793945312,
|
733 |
+
"rewards/margins": 4.53641414642334,
|
734 |
+
"rewards/rejected": 12.471475601196289,
|
735 |
+
"step": 420
|
736 |
+
},
|
737 |
+
{
|
738 |
+
"epoch": 0.37333333333333335,
|
739 |
+
"eval_logits/chosen": -0.09054450690746307,
|
740 |
+
"eval_logits/rejected": -0.10264354199171066,
|
741 |
+
"eval_logps/chosen": -1.913105845451355,
|
742 |
+
"eval_logps/rejected": -51.11127471923828,
|
743 |
+
"eval_loss": 0.16789735853672028,
|
744 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
745 |
+
"eval_rewards/chosen": 17.311582565307617,
|
746 |
+
"eval_rewards/margins": 5.061989784240723,
|
747 |
+
"eval_rewards/rejected": 12.249593734741211,
|
748 |
+
"eval_runtime": 361.5337,
|
749 |
+
"eval_samples_per_second": 2.766,
|
750 |
+
"eval_steps_per_second": 0.346,
|
751 |
+
"step": 420
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.38222222222222224,
|
755 |
+
"grad_norm": 12.824393272399902,
|
756 |
+
"learning_rate": 3.884082138308699e-06,
|
757 |
+
"logits/chosen": -0.08666776865720749,
|
758 |
+
"logits/rejected": -0.0997733399271965,
|
759 |
+
"logps/chosen": -1.7306327819824219,
|
760 |
+
"logps/rejected": -54.273292541503906,
|
761 |
+
"loss": 0.2298,
|
762 |
+
"rewards/accuracies": 0.9500000476837158,
|
763 |
+
"rewards/chosen": 17.167621612548828,
|
764 |
+
"rewards/margins": 5.065673351287842,
|
765 |
+
"rewards/rejected": 12.101947784423828,
|
766 |
+
"step": 430
|
767 |
+
},
|
768 |
+
{
|
769 |
+
"epoch": 0.39111111111111113,
|
770 |
+
"grad_norm": 0.30713599920272827,
|
771 |
+
"learning_rate": 3.818796323137896e-06,
|
772 |
+
"logits/chosen": -0.09174907952547073,
|
773 |
+
"logits/rejected": -0.10376611351966858,
|
774 |
+
"logps/chosen": -1.489154577255249,
|
775 |
+
"logps/rejected": -54.580726623535156,
|
776 |
+
"loss": 0.2513,
|
777 |
+
"rewards/accuracies": 0.9375,
|
778 |
+
"rewards/chosen": 17.22280502319336,
|
779 |
+
"rewards/margins": 5.175349235534668,
|
780 |
+
"rewards/rejected": 12.047454833984375,
|
781 |
+
"step": 440
|
782 |
+
},
|
783 |
+
{
|
784 |
+
"epoch": 0.4,
|
785 |
+
"grad_norm": 87.4791488647461,
|
786 |
+
"learning_rate": 3.7522396951963303e-06,
|
787 |
+
"logits/chosen": -0.09688778221607208,
|
788 |
+
"logits/rejected": -0.10897806286811829,
|
789 |
+
"logps/chosen": -3.157695770263672,
|
790 |
+
"logps/rejected": -50.96417236328125,
|
791 |
+
"loss": 0.1758,
|
792 |
+
"rewards/accuracies": 0.9500000476837158,
|
793 |
+
"rewards/chosen": 17.345651626586914,
|
794 |
+
"rewards/margins": 5.245656967163086,
|
795 |
+
"rewards/rejected": 12.099993705749512,
|
796 |
+
"step": 450
|
797 |
+
},
|
798 |
+
{
|
799 |
+
"epoch": 0.4088888888888889,
|
800 |
+
"grad_norm": 146.2008056640625,
|
801 |
+
"learning_rate": 3.684476389492026e-06,
|
802 |
+
"logits/chosen": -0.09378582239151001,
|
803 |
+
"logits/rejected": -0.10475654900074005,
|
804 |
+
"logps/chosen": -0.5611928701400757,
|
805 |
+
"logps/rejected": -56.518890380859375,
|
806 |
+
"loss": 0.1981,
|
807 |
+
"rewards/accuracies": 0.9500000476837158,
|
808 |
+
"rewards/chosen": 17.113712310791016,
|
809 |
+
"rewards/margins": 5.068872928619385,
|
810 |
+
"rewards/rejected": 12.044839859008789,
|
811 |
+
"step": 460
|
812 |
+
},
|
813 |
+
{
|
814 |
+
"epoch": 0.4177777777777778,
|
815 |
+
"grad_norm": 1.9137721061706543,
|
816 |
+
"learning_rate": 3.6155717038065783e-06,
|
817 |
+
"logits/chosen": -0.08695463836193085,
|
818 |
+
"logits/rejected": -0.09596743434667587,
|
819 |
+
"logps/chosen": -1.5298550128936768,
|
820 |
+
"logps/rejected": -50.27445983886719,
|
821 |
+
"loss": 0.2066,
|
822 |
+
"rewards/accuracies": 0.9375,
|
823 |
+
"rewards/chosen": 17.35186004638672,
|
824 |
+
"rewards/margins": 5.014693260192871,
|
825 |
+
"rewards/rejected": 12.337167739868164,
|
826 |
+
"step": 470
|
827 |
+
},
|
828 |
+
{
|
829 |
+
"epoch": 0.4266666666666667,
|
830 |
+
"grad_norm": 84.80391693115234,
|
831 |
+
"learning_rate": 3.545592035773192e-06,
|
832 |
+
"logits/chosen": -0.0746893435716629,
|
833 |
+
"logits/rejected": -0.08653923869132996,
|
834 |
+
"logps/chosen": -2.0052125453948975,
|
835 |
+
"logps/rejected": -57.502811431884766,
|
836 |
+
"loss": 0.1149,
|
837 |
+
"rewards/accuracies": 0.9500000476837158,
|
838 |
+
"rewards/chosen": 17.14373016357422,
|
839 |
+
"rewards/margins": 5.360415935516357,
|
840 |
+
"rewards/rejected": 11.783313751220703,
|
841 |
+
"step": 480
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 0.4266666666666667,
|
845 |
+
"eval_logits/chosen": -0.07700399309396744,
|
846 |
+
"eval_logits/rejected": -0.08828537166118622,
|
847 |
+
"eval_logps/chosen": -4.48896598815918,
|
848 |
+
"eval_logps/rejected": -53.76282501220703,
|
849 |
+
"eval_loss": 0.29511645436286926,
|
850 |
+
"eval_rewards/accuracies": 0.9230000376701355,
|
851 |
+
"eval_rewards/chosen": 17.053997039794922,
|
852 |
+
"eval_rewards/margins": 5.069558143615723,
|
853 |
+
"eval_rewards/rejected": 11.984437942504883,
|
854 |
+
"eval_runtime": 361.5035,
|
855 |
+
"eval_samples_per_second": 2.766,
|
856 |
+
"eval_steps_per_second": 0.346,
|
857 |
+
"step": 480
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 0.43555555555555553,
|
861 |
+
"grad_norm": 82.9616470336914,
|
862 |
+
"learning_rate": 3.4746048188948806e-06,
|
863 |
+
"logits/chosen": -0.06675051152706146,
|
864 |
+
"logits/rejected": -0.07860895991325378,
|
865 |
+
"logps/chosen": -4.162237167358398,
|
866 |
+
"logps/rejected": -54.77789306640625,
|
867 |
+
"loss": 0.2979,
|
868 |
+
"rewards/accuracies": 0.9125000238418579,
|
869 |
+
"rewards/chosen": 17.047603607177734,
|
870 |
+
"rewards/margins": 5.138361930847168,
|
871 |
+
"rewards/rejected": 11.909242630004883,
|
872 |
+
"step": 490
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.4444444444444444,
|
876 |
+
"grad_norm": 0.04293210059404373,
|
877 |
+
"learning_rate": 3.4026784575644887e-06,
|
878 |
+
"logits/chosen": -0.06424491107463837,
|
879 |
+
"logits/rejected": -0.07567107677459717,
|
880 |
+
"logps/chosen": -2.05729603767395,
|
881 |
+
"logps/rejected": -56.646087646484375,
|
882 |
+
"loss": 0.4378,
|
883 |
+
"rewards/accuracies": 0.8875000476837158,
|
884 |
+
"rewards/chosen": 16.947803497314453,
|
885 |
+
"rewards/margins": 4.919981956481934,
|
886 |
+
"rewards/rejected": 12.02782154083252,
|
887 |
+
"step": 500
|
888 |
+
},
|
889 |
+
{
|
890 |
+
"epoch": 0.4533333333333333,
|
891 |
+
"grad_norm": 0.07136644423007965,
|
892 |
+
"learning_rate": 3.329882261149148e-06,
|
893 |
+
"logits/chosen": -0.06423303484916687,
|
894 |
+
"logits/rejected": -0.07512776553630829,
|
895 |
+
"logps/chosen": -3.1519265174865723,
|
896 |
+
"logps/rejected": -53.53008270263672,
|
897 |
+
"loss": 0.2613,
|
898 |
+
"rewards/accuracies": 0.9375,
|
899 |
+
"rewards/chosen": 17.112144470214844,
|
900 |
+
"rewards/margins": 5.046430587768555,
|
901 |
+
"rewards/rejected": 12.065712928771973,
|
902 |
+
"step": 510
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 0.4622222222222222,
|
906 |
+
"grad_norm": 0.29279613494873047,
|
907 |
+
"learning_rate": 3.25628637720269e-06,
|
908 |
+
"logits/chosen": -0.060233693569898605,
|
909 |
+
"logits/rejected": -0.07076811790466309,
|
910 |
+
"logps/chosen": -1.2358124256134033,
|
911 |
+
"logps/rejected": -50.61806106567383,
|
912 |
+
"loss": 0.1517,
|
913 |
+
"rewards/accuracies": 0.9375,
|
914 |
+
"rewards/chosen": 17.457351684570312,
|
915 |
+
"rewards/margins": 5.2062835693359375,
|
916 |
+
"rewards/rejected": 12.251070022583008,
|
917 |
+
"step": 520
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 0.4711111111111111,
|
921 |
+
"grad_norm": 0.7946074604988098,
|
922 |
+
"learning_rate": 3.181961723870359e-06,
|
923 |
+
"logits/chosen": -0.054482050240039825,
|
924 |
+
"logits/rejected": -0.0657092034816742,
|
925 |
+
"logps/chosen": -0.7682158946990967,
|
926 |
+
"logps/rejected": -58.81409454345703,
|
927 |
+
"loss": 0.2304,
|
928 |
+
"rewards/accuracies": 0.9375,
|
929 |
+
"rewards/chosen": 17.075973510742188,
|
930 |
+
"rewards/margins": 5.222441673278809,
|
931 |
+
"rewards/rejected": 11.853530883789062,
|
932 |
+
"step": 530
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 0.48,
|
936 |
+
"grad_norm": 3.1405210494995117,
|
937 |
+
"learning_rate": 3.1069799215509847e-06,
|
938 |
+
"logits/chosen": -0.05050881579518318,
|
939 |
+
"logits/rejected": -0.061149902641773224,
|
940 |
+
"logps/chosen": -0.954046368598938,
|
941 |
+
"logps/rejected": -52.70227813720703,
|
942 |
+
"loss": 0.0384,
|
943 |
+
"rewards/accuracies": 0.987500011920929,
|
944 |
+
"rewards/chosen": 17.378849029541016,
|
945 |
+
"rewards/margins": 5.253483295440674,
|
946 |
+
"rewards/rejected": 12.125364303588867,
|
947 |
+
"step": 540
|
948 |
+
},
|
949 |
+
{
|
950 |
+
"epoch": 0.48,
|
951 |
+
"eval_logits/chosen": -0.05120665580034256,
|
952 |
+
"eval_logits/rejected": -0.06123337894678116,
|
953 |
+
"eval_logps/chosen": -2.9872913360595703,
|
954 |
+
"eval_logps/rejected": -52.27314758300781,
|
955 |
+
"eval_loss": 0.17387841641902924,
|
956 |
+
"eval_rewards/accuracies": 0.9490000605583191,
|
957 |
+
"eval_rewards/chosen": 17.204164505004883,
|
958 |
+
"eval_rewards/margins": 5.070757865905762,
|
959 |
+
"eval_rewards/rejected": 12.133406639099121,
|
960 |
+
"eval_runtime": 361.5449,
|
961 |
+
"eval_samples_per_second": 2.766,
|
962 |
+
"eval_steps_per_second": 0.346,
|
963 |
+
"step": 540
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 0.4888888888888889,
|
967 |
+
"grad_norm": 0.07270358502864838,
|
968 |
+
"learning_rate": 3.0314132238824416e-06,
|
969 |
+
"logits/chosen": -0.05125313252210617,
|
970 |
+
"logits/rejected": -0.06174170970916748,
|
971 |
+
"logps/chosen": -3.2163877487182617,
|
972 |
+
"logps/rejected": -47.79279327392578,
|
973 |
+
"loss": 0.2087,
|
974 |
+
"rewards/accuracies": 0.9125000238418579,
|
975 |
+
"rewards/chosen": 17.408517837524414,
|
976 |
+
"rewards/margins": 5.061524391174316,
|
977 |
+
"rewards/rejected": 12.346992492675781,
|
978 |
+
"step": 550
|
979 |
+
},
|
980 |
+
{
|
981 |
+
"epoch": 0.49777777777777776,
|
982 |
+
"grad_norm": 0.10005924850702286,
|
983 |
+
"learning_rate": 2.955334448116915e-06,
|
984 |
+
"logits/chosen": -0.041773442178964615,
|
985 |
+
"logits/rejected": -0.05364570394158363,
|
986 |
+
"logps/chosen": -0.363404780626297,
|
987 |
+
"logps/rejected": -56.32415008544922,
|
988 |
+
"loss": 0.0969,
|
989 |
+
"rewards/accuracies": 0.9624999761581421,
|
990 |
+
"rewards/chosen": 17.3824462890625,
|
991 |
+
"rewards/margins": 5.5404510498046875,
|
992 |
+
"rewards/rejected": 11.841995239257812,
|
993 |
+
"step": 560
|
994 |
+
},
|
995 |
+
{
|
996 |
+
"epoch": 0.5066666666666667,
|
997 |
+
"grad_norm": 7.818356990814209,
|
998 |
+
"learning_rate": 2.8788169049530533e-06,
|
999 |
+
"logits/chosen": -0.04309462010860443,
|
1000 |
+
"logits/rejected": -0.05494442582130432,
|
1001 |
+
"logps/chosen": -2.2242724895477295,
|
1002 |
+
"logps/rejected": -56.444740295410156,
|
1003 |
+
"loss": 0.1447,
|
1004 |
+
"rewards/accuracies": 0.949999988079071,
|
1005 |
+
"rewards/chosen": 17.275390625,
|
1006 |
+
"rewards/margins": 5.55007791519165,
|
1007 |
+
"rewards/rejected": 11.725313186645508,
|
1008 |
+
"step": 570
|
1009 |
+
},
|
1010 |
+
{
|
1011 |
+
"epoch": 0.5155555555555555,
|
1012 |
+
"grad_norm": 0.03519747406244278,
|
1013 |
+
"learning_rate": 2.8019343278926397e-06,
|
1014 |
+
"logits/chosen": -0.03506368771195412,
|
1015 |
+
"logits/rejected": -0.046854715794324875,
|
1016 |
+
"logps/chosen": -0.5200096964836121,
|
1017 |
+
"logps/rejected": -59.05330276489258,
|
1018 |
+
"loss": 0.0998,
|
1019 |
+
"rewards/accuracies": 0.949999988079071,
|
1020 |
+
"rewards/chosen": 17.320554733276367,
|
1021 |
+
"rewards/margins": 5.727260589599609,
|
1022 |
+
"rewards/rejected": 11.593294143676758,
|
1023 |
+
"step": 580
|
1024 |
+
},
|
1025 |
+
{
|
1026 |
+
"epoch": 0.5244444444444445,
|
1027 |
+
"grad_norm": 0.04519123584032059,
|
1028 |
+
"learning_rate": 2.7247608021898265e-06,
|
1029 |
+
"logits/chosen": -0.03204537555575371,
|
1030 |
+
"logits/rejected": -0.04383891448378563,
|
1031 |
+
"logps/chosen": -1.1271060705184937,
|
1032 |
+
"logps/rejected": -59.308895111083984,
|
1033 |
+
"loss": 0.1162,
|
1034 |
+
"rewards/accuracies": 0.9750000238418579,
|
1035 |
+
"rewards/chosen": 17.256351470947266,
|
1036 |
+
"rewards/margins": 5.6881890296936035,
|
1037 |
+
"rewards/rejected": 11.568161010742188,
|
1038 |
+
"step": 590
|
1039 |
+
},
|
1040 |
+
{
|
1041 |
+
"epoch": 0.5333333333333333,
|
1042 |
+
"grad_norm": 0.028689857572317123,
|
1043 |
+
"learning_rate": 2.647370693461432e-06,
|
1044 |
+
"logits/chosen": -0.02834726870059967,
|
1045 |
+
"logits/rejected": -0.03827046602964401,
|
1046 |
+
"logps/chosen": -5.673943519592285,
|
1047 |
+
"logps/rejected": -55.72624588012695,
|
1048 |
+
"loss": 0.4008,
|
1049 |
+
"rewards/accuracies": 0.8875000476837158,
|
1050 |
+
"rewards/chosen": 16.761056900024414,
|
1051 |
+
"rewards/margins": 4.804043769836426,
|
1052 |
+
"rewards/rejected": 11.957012176513672,
|
1053 |
+
"step": 600
|
1054 |
+
},
|
1055 |
+
{
|
1056 |
+
"epoch": 0.5333333333333333,
|
1057 |
+
"eval_logits/chosen": -0.035840023308992386,
|
1058 |
+
"eval_logits/rejected": -0.04687971621751785,
|
1059 |
+
"eval_logps/chosen": -2.1760435104370117,
|
1060 |
+
"eval_logps/rejected": -56.62664031982422,
|
1061 |
+
"eval_loss": 0.17061151564121246,
|
1062 |
+
"eval_rewards/accuracies": 0.9470000267028809,
|
1063 |
+
"eval_rewards/chosen": 17.285289764404297,
|
1064 |
+
"eval_rewards/margins": 5.587231636047363,
|
1065 |
+
"eval_rewards/rejected": 11.698057174682617,
|
1066 |
+
"eval_runtime": 361.5056,
|
1067 |
+
"eval_samples_per_second": 2.766,
|
1068 |
+
"eval_steps_per_second": 0.346,
|
1069 |
+
"step": 600
|
1070 |
+
},
|
1071 |
+
{
|
1072 |
+
"epoch": 0.5422222222222223,
|
1073 |
+
"grad_norm": 1.1994622945785522,
|
1074 |
+
"learning_rate": 2.569838576027068e-06,
|
1075 |
+
"logits/chosen": -0.03231767192482948,
|
1076 |
+
"logits/rejected": -0.04397805407643318,
|
1077 |
+
"logps/chosen": -1.5904741287231445,
|
1078 |
+
"logps/rejected": -62.10520553588867,
|
1079 |
+
"loss": 0.1265,
|
1080 |
+
"rewards/accuracies": 0.9624999761581421,
|
1081 |
+
"rewards/chosen": 17.11206817626953,
|
1082 |
+
"rewards/margins": 5.734784126281738,
|
1083 |
+
"rewards/rejected": 11.37728500366211,
|
1084 |
+
"step": 610
|
1085 |
+
},
|
1086 |
+
{
|
1087 |
+
"epoch": 0.5511111111111111,
|
1088 |
+
"grad_norm": 0.14163845777511597,
|
1089 |
+
"learning_rate": 2.4922391610481544e-06,
|
1090 |
+
"logits/chosen": -0.03293662518262863,
|
1091 |
+
"logits/rejected": -0.04335154965519905,
|
1092 |
+
"logps/chosen": -1.3568997383117676,
|
1093 |
+
"logps/rejected": -58.347434997558594,
|
1094 |
+
"loss": 0.097,
|
1095 |
+
"rewards/accuracies": 0.9624999761581421,
|
1096 |
+
"rewards/chosen": 17.345399856567383,
|
1097 |
+
"rewards/margins": 5.823373794555664,
|
1098 |
+
"rewards/rejected": 11.522026062011719,
|
1099 |
+
"step": 620
|
1100 |
+
},
|
1101 |
+
{
|
1102 |
+
"epoch": 0.56,
|
1103 |
+
"grad_norm": 1.125027060508728,
|
1104 |
+
"learning_rate": 2.4146472245350804e-06,
|
1105 |
+
"logits/chosen": -0.02864963933825493,
|
1106 |
+
"logits/rejected": -0.03835710883140564,
|
1107 |
+
"logps/chosen": -2.494175434112549,
|
1108 |
+
"logps/rejected": -55.33067321777344,
|
1109 |
+
"loss": 0.3115,
|
1110 |
+
"rewards/accuracies": 0.925000011920929,
|
1111 |
+
"rewards/chosen": 17.262205123901367,
|
1112 |
+
"rewards/margins": 5.4287261962890625,
|
1113 |
+
"rewards/rejected": 11.833479881286621,
|
1114 |
+
"step": 630
|
1115 |
+
},
|
1116 |
+
{
|
1117 |
+
"epoch": 0.5688888888888889,
|
1118 |
+
"grad_norm": 0.03514016419649124,
|
1119 |
+
"learning_rate": 2.337137535291868e-06,
|
1120 |
+
"logits/chosen": -0.02757749892771244,
|
1121 |
+
"logits/rejected": -0.0376611053943634,
|
1122 |
+
"logps/chosen": -2.3163387775421143,
|
1123 |
+
"logps/rejected": -53.54579162597656,
|
1124 |
+
"loss": 0.1268,
|
1125 |
+
"rewards/accuracies": 0.9500000476837158,
|
1126 |
+
"rewards/chosen": 17.542863845825195,
|
1127 |
+
"rewards/margins": 5.799897193908691,
|
1128 |
+
"rewards/rejected": 11.742965698242188,
|
1129 |
+
"step": 640
|
1130 |
+
},
|
1131 |
+
{
|
1132 |
+
"epoch": 0.5777777777777777,
|
1133 |
+
"grad_norm": 34.20791244506836,
|
1134 |
+
"learning_rate": 2.259784782867782e-06,
|
1135 |
+
"logits/chosen": -0.02032250165939331,
|
1136 |
+
"logits/rejected": -0.031542714685201645,
|
1137 |
+
"logps/chosen": -1.7910137176513672,
|
1138 |
+
"logps/rejected": -62.980018615722656,
|
1139 |
+
"loss": 0.1831,
|
1140 |
+
"rewards/accuracies": 0.9500000476837158,
|
1141 |
+
"rewards/chosen": 16.982879638671875,
|
1142 |
+
"rewards/margins": 5.589078903198242,
|
1143 |
+
"rewards/rejected": 11.393800735473633,
|
1144 |
+
"step": 650
|
1145 |
+
},
|
1146 |
+
{
|
1147 |
+
"epoch": 0.5866666666666667,
|
1148 |
+
"grad_norm": 32.356956481933594,
|
1149 |
+
"learning_rate": 2.182663505585314e-06,
|
1150 |
+
"logits/chosen": -0.01711965538561344,
|
1151 |
+
"logits/rejected": -0.028038471937179565,
|
1152 |
+
"logps/chosen": -2.662904739379883,
|
1153 |
+
"logps/rejected": -63.326297760009766,
|
1154 |
+
"loss": 0.1678,
|
1155 |
+
"rewards/accuracies": 0.9624999761581421,
|
1156 |
+
"rewards/chosen": 16.997777938842773,
|
1157 |
+
"rewards/margins": 5.741157531738281,
|
1158 |
+
"rewards/rejected": 11.256620407104492,
|
1159 |
+
"step": 660
|
1160 |
+
},
|
1161 |
+
{
|
1162 |
+
"epoch": 0.5866666666666667,
|
1163 |
+
"eval_logits/chosen": -0.016018809750676155,
|
1164 |
+
"eval_logits/rejected": -0.027020033448934555,
|
1165 |
+
"eval_logps/chosen": -3.00822114944458,
|
1166 |
+
"eval_logps/rejected": -57.951629638671875,
|
1167 |
+
"eval_loss": 0.2050127536058426,
|
1168 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1169 |
+
"eval_rewards/chosen": 17.202072143554688,
|
1170 |
+
"eval_rewards/margins": 5.636512756347656,
|
1171 |
+
"eval_rewards/rejected": 11.565557479858398,
|
1172 |
+
"eval_runtime": 361.5073,
|
1173 |
+
"eval_samples_per_second": 2.766,
|
1174 |
+
"eval_steps_per_second": 0.346,
|
1175 |
+
"step": 660
|
1176 |
+
},
|
1177 |
+
{
|
1178 |
+
"epoch": 0.5955555555555555,
|
1179 |
+
"grad_norm": 86.53874969482422,
|
1180 |
+
"learning_rate": 2.1058480187138863e-06,
|
1181 |
+
"logits/chosen": -0.012352555990219116,
|
1182 |
+
"logits/rejected": -0.023966707289218903,
|
1183 |
+
"logps/chosen": -2.129209041595459,
|
1184 |
+
"logps/rejected": -62.1393928527832,
|
1185 |
+
"loss": 0.1025,
|
1186 |
+
"rewards/accuracies": 0.9624999761581421,
|
1187 |
+
"rewards/chosen": 17.069538116455078,
|
1188 |
+
"rewards/margins": 5.678930759429932,
|
1189 |
+
"rewards/rejected": 11.390605926513672,
|
1190 |
+
"step": 670
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 0.6044444444444445,
|
1194 |
+
"grad_norm": 2.892672300338745,
|
1195 |
+
"learning_rate": 2.0294123428584985e-06,
|
1196 |
+
"logits/chosen": -0.011562807485461235,
|
1197 |
+
"logits/rejected": -0.020859256386756897,
|
1198 |
+
"logps/chosen": -3.2486608028411865,
|
1199 |
+
"logps/rejected": -59.56721496582031,
|
1200 |
+
"loss": 0.1961,
|
1201 |
+
"rewards/accuracies": 0.9500000476837158,
|
1202 |
+
"rewards/chosen": 17.051376342773438,
|
1203 |
+
"rewards/margins": 5.515361785888672,
|
1204 |
+
"rewards/rejected": 11.536016464233398,
|
1205 |
+
"step": 680
|
1206 |
+
},
|
1207 |
+
{
|
1208 |
+
"epoch": 0.6133333333333333,
|
1209 |
+
"grad_norm": 30.26588249206543,
|
1210 |
+
"learning_rate": 1.953430132632311e-06,
|
1211 |
+
"logits/chosen": -0.011488726362586021,
|
1212 |
+
"logits/rejected": -0.021527227014303207,
|
1213 |
+
"logps/chosen": -1.794357180595398,
|
1214 |
+
"logps/rejected": -60.649505615234375,
|
1215 |
+
"loss": 0.2821,
|
1216 |
+
"rewards/accuracies": 0.9375,
|
1217 |
+
"rewards/chosen": 17.114843368530273,
|
1218 |
+
"rewards/margins": 5.612217903137207,
|
1219 |
+
"rewards/rejected": 11.50262451171875,
|
1220 |
+
"step": 690
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 0.6222222222222222,
|
1224 |
+
"grad_norm": 0.030314341187477112,
|
1225 |
+
"learning_rate": 1.8779746056819104e-06,
|
1226 |
+
"logits/chosen": -0.014436552301049232,
|
1227 |
+
"logits/rejected": -0.026044374331831932,
|
1228 |
+
"logps/chosen": -3.1617178916931152,
|
1229 |
+
"logps/rejected": -60.94548797607422,
|
1230 |
+
"loss": 0.3141,
|
1231 |
+
"rewards/accuracies": 0.9375,
|
1232 |
+
"rewards/chosen": 17.08698272705078,
|
1233 |
+
"rewards/margins": 5.719264984130859,
|
1234 |
+
"rewards/rejected": 11.367716789245605,
|
1235 |
+
"step": 700
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 0.6311111111111111,
|
1239 |
+
"grad_norm": 0.06711317598819733,
|
1240 |
+
"learning_rate": 1.8031184721336364e-06,
|
1241 |
+
"logits/chosen": -0.017312290146946907,
|
1242 |
+
"logits/rejected": -0.026554957032203674,
|
1243 |
+
"logps/chosen": -5.120705604553223,
|
1244 |
+
"logps/rejected": -54.33483123779297,
|
1245 |
+
"loss": 0.3827,
|
1246 |
+
"rewards/accuracies": 0.887499988079071,
|
1247 |
+
"rewards/chosen": 17.1247501373291,
|
1248 |
+
"rewards/margins": 5.327882766723633,
|
1249 |
+
"rewards/rejected": 11.796867370605469,
|
1250 |
+
"step": 710
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 0.64,
|
1254 |
+
"grad_norm": 60.36637496948242,
|
1255 |
+
"learning_rate": 1.7289338645289711e-06,
|
1256 |
+
"logits/chosen": -0.01987219974398613,
|
1257 |
+
"logits/rejected": -0.02766304835677147,
|
1258 |
+
"logps/chosen": -2.1757986545562744,
|
1259 |
+
"logps/rejected": -54.02531433105469,
|
1260 |
+
"loss": 0.2272,
|
1261 |
+
"rewards/accuracies": 0.949999988079071,
|
1262 |
+
"rewards/chosen": 17.239776611328125,
|
1263 |
+
"rewards/margins": 5.250313758850098,
|
1264 |
+
"rewards/rejected": 11.989462852478027,
|
1265 |
+
"step": 720
|
1266 |
+
},
|
1267 |
+
{
|
1268 |
+
"epoch": 0.64,
|
1269 |
+
"eval_logits/chosen": -0.022905193269252777,
|
1270 |
+
"eval_logits/rejected": -0.0321992002427578,
|
1271 |
+
"eval_logps/chosen": -1.100506067276001,
|
1272 |
+
"eval_logps/rejected": -55.91169738769531,
|
1273 |
+
"eval_loss": 0.14018221199512482,
|
1274 |
+
"eval_rewards/accuracies": 0.9520000219345093,
|
1275 |
+
"eval_rewards/chosen": 17.392841339111328,
|
1276 |
+
"eval_rewards/margins": 5.623291492462158,
|
1277 |
+
"eval_rewards/rejected": 11.769551277160645,
|
1278 |
+
"eval_runtime": 361.416,
|
1279 |
+
"eval_samples_per_second": 2.767,
|
1280 |
+
"eval_steps_per_second": 0.346,
|
1281 |
+
"step": 720
|
1282 |
+
},
|
1283 |
+
{
|
1284 |
+
"epoch": 0.6488888888888888,
|
1285 |
+
"grad_norm": 0.029294608160853386,
|
1286 |
+
"learning_rate": 1.6554922683164875e-06,
|
1287 |
+
"logits/chosen": -0.016416028141975403,
|
1288 |
+
"logits/rejected": -0.025527067482471466,
|
1289 |
+
"logps/chosen": -0.906692385673523,
|
1290 |
+
"logps/rejected": -59.142173767089844,
|
1291 |
+
"loss": 0.1171,
|
1292 |
+
"rewards/accuracies": 0.9624999761581421,
|
1293 |
+
"rewards/chosen": 17.282169342041016,
|
1294 |
+
"rewards/margins": 5.6978583335876465,
|
1295 |
+
"rewards/rejected": 11.584310531616211,
|
1296 |
+
"step": 730
|
1297 |
+
},
|
1298 |
+
{
|
1299 |
+
"epoch": 0.6577777777777778,
|
1300 |
+
"grad_norm": 33.89070510864258,
|
1301 |
+
"learning_rate": 1.5828644529673592e-06,
|
1302 |
+
"logits/chosen": -0.016044551506638527,
|
1303 |
+
"logits/rejected": -0.025811903178691864,
|
1304 |
+
"logps/chosen": -1.1167538166046143,
|
1305 |
+
"logps/rejected": -59.84492492675781,
|
1306 |
+
"loss": 0.2001,
|
1307 |
+
"rewards/accuracies": 0.9624999761581421,
|
1308 |
+
"rewards/chosen": 17.266983032226562,
|
1309 |
+
"rewards/margins": 5.749438285827637,
|
1310 |
+
"rewards/rejected": 11.51754379272461,
|
1311 |
+
"step": 740
|
1312 |
+
},
|
1313 |
+
{
|
1314 |
+
"epoch": 0.6666666666666666,
|
1315 |
+
"grad_norm": 0.53726726770401,
|
1316 |
+
"learning_rate": 1.5111204037807844e-06,
|
1317 |
+
"logits/chosen": -0.012529855594038963,
|
1318 |
+
"logits/rejected": -0.023765765130519867,
|
1319 |
+
"logps/chosen": -0.3461765646934509,
|
1320 |
+
"logps/rejected": -55.841102600097656,
|
1321 |
+
"loss": 0.1692,
|
1322 |
+
"rewards/accuracies": 0.9375,
|
1323 |
+
"rewards/chosen": 17.61865997314453,
|
1324 |
+
"rewards/margins": 6.019055366516113,
|
1325 |
+
"rewards/rejected": 11.599604606628418,
|
1326 |
+
"step": 750
|
1327 |
+
},
|
1328 |
+
{
|
1329 |
+
"epoch": 0.6755555555555556,
|
1330 |
+
"grad_norm": 0.7173987627029419,
|
1331 |
+
"learning_rate": 1.4403292544450625e-06,
|
1332 |
+
"logits/chosen": -0.013380522839725018,
|
1333 |
+
"logits/rejected": -0.022107835859060287,
|
1334 |
+
"logps/chosen": -3.2759666442871094,
|
1335 |
+
"logps/rejected": -53.836822509765625,
|
1336 |
+
"loss": 0.2418,
|
1337 |
+
"rewards/accuracies": 0.949999988079071,
|
1338 |
+
"rewards/chosen": 17.383333206176758,
|
1339 |
+
"rewards/margins": 5.590358734130859,
|
1340 |
+
"rewards/rejected": 11.792974472045898,
|
1341 |
+
"step": 760
|
1342 |
+
},
|
1343 |
+
{
|
1344 |
+
"epoch": 0.6844444444444444,
|
1345 |
+
"grad_norm": 2.744900941848755,
|
1346 |
+
"learning_rate": 1.3705592204192853e-06,
|
1347 |
+
"logits/chosen": -0.011318420059978962,
|
1348 |
+
"logits/rejected": -0.021840626373887062,
|
1349 |
+
"logps/chosen": -2.592001438140869,
|
1350 |
+
"logps/rejected": -57.06926345825195,
|
1351 |
+
"loss": 0.1749,
|
1352 |
+
"rewards/accuracies": 0.9500000476837158,
|
1353 |
+
"rewards/chosen": 17.286128997802734,
|
1354 |
+
"rewards/margins": 5.662715911865234,
|
1355 |
+
"rewards/rejected": 11.623414993286133,
|
1356 |
+
"step": 770
|
1357 |
+
},
|
1358 |
+
{
|
1359 |
+
"epoch": 0.6933333333333334,
|
1360 |
+
"grad_norm": 70.76551818847656,
|
1361 |
+
"learning_rate": 1.301877533199859e-06,
|
1362 |
+
"logits/chosen": -0.013944407925009727,
|
1363 |
+
"logits/rejected": -0.025667501613497734,
|
1364 |
+
"logps/chosen": -2.1070234775543213,
|
1365 |
+
"logps/rejected": -57.72039794921875,
|
1366 |
+
"loss": 0.1915,
|
1367 |
+
"rewards/accuracies": 0.925000011920929,
|
1368 |
+
"rewards/chosen": 17.28545379638672,
|
1369 |
+
"rewards/margins": 5.708344459533691,
|
1370 |
+
"rewards/rejected": 11.577108383178711,
|
1371 |
+
"step": 780
|
1372 |
+
},
|
1373 |
+
{
|
1374 |
+
"epoch": 0.6933333333333334,
|
1375 |
+
"eval_logits/chosen": -0.01664295792579651,
|
1376 |
+
"eval_logits/rejected": -0.026626665145158768,
|
1377 |
+
"eval_logps/chosen": -1.082255244255066,
|
1378 |
+
"eval_logps/rejected": -55.95073699951172,
|
1379 |
+
"eval_loss": 0.24412688612937927,
|
1380 |
+
"eval_rewards/accuracies": 0.9320000410079956,
|
1381 |
+
"eval_rewards/chosen": 17.394668579101562,
|
1382 |
+
"eval_rewards/margins": 5.629020690917969,
|
1383 |
+
"eval_rewards/rejected": 11.765647888183594,
|
1384 |
+
"eval_runtime": 361.4602,
|
1385 |
+
"eval_samples_per_second": 2.767,
|
1386 |
+
"eval_steps_per_second": 0.346,
|
1387 |
+
"step": 780
|
1388 |
+
},
|
1389 |
+
{
|
1390 |
+
"epoch": 0.7022222222222222,
|
1391 |
+
"grad_norm": 0.41359376907348633,
|
1392 |
+
"learning_rate": 1.2343503755351729e-06,
|
1393 |
+
"logits/chosen": -0.012772129848599434,
|
1394 |
+
"logits/rejected": -0.023186586797237396,
|
1395 |
+
"logps/chosen": -0.8030359148979187,
|
1396 |
+
"logps/rejected": -57.84947967529297,
|
1397 |
+
"loss": 0.2143,
|
1398 |
+
"rewards/accuracies": 0.9375,
|
1399 |
+
"rewards/chosen": 17.32242202758789,
|
1400 |
+
"rewards/margins": 5.645486831665039,
|
1401 |
+
"rewards/rejected": 11.676933288574219,
|
1402 |
+
"step": 790
|
1403 |
+
},
|
1404 |
+
{
|
1405 |
+
"epoch": 0.7111111111111111,
|
1406 |
+
"grad_norm": 0.41748157143592834,
|
1407 |
+
"learning_rate": 1.168042817650881e-06,
|
1408 |
+
"logits/chosen": -0.012502101249992847,
|
1409 |
+
"logits/rejected": -0.023272844031453133,
|
1410 |
+
"logps/chosen": -1.0377256870269775,
|
1411 |
+
"logps/rejected": -57.05615997314453,
|
1412 |
+
"loss": 0.0792,
|
1413 |
+
"rewards/accuracies": 0.9750000238418579,
|
1414 |
+
"rewards/chosen": 17.60501480102539,
|
1415 |
+
"rewards/margins": 6.159370422363281,
|
1416 |
+
"rewards/rejected": 11.445646286010742,
|
1417 |
+
"step": 800
|
1418 |
+
},
|
1419 |
+
{
|
1420 |
+
"epoch": 0.72,
|
1421 |
+
"grad_norm": 0.039210401475429535,
|
1422 |
+
"learning_rate": 1.1030187545472012e-06,
|
1423 |
+
"logits/chosen": -0.008177272044122219,
|
1424 |
+
"logits/rejected": -0.01832464337348938,
|
1425 |
+
"logps/chosen": -3.8701748847961426,
|
1426 |
+
"logps/rejected": -57.661476135253906,
|
1427 |
+
"loss": 0.3177,
|
1428 |
+
"rewards/accuracies": 0.925000011920929,
|
1429 |
+
"rewards/chosen": 17.14336395263672,
|
1430 |
+
"rewards/margins": 5.583393573760986,
|
1431 |
+
"rewards/rejected": 11.55997085571289,
|
1432 |
+
"step": 810
|
1433 |
+
},
|
1434 |
+
{
|
1435 |
+
"epoch": 0.7288888888888889,
|
1436 |
+
"grad_norm": 1.6096951961517334,
|
1437 |
+
"learning_rate": 1.0393408444287048e-06,
|
1438 |
+
"logits/chosen": -0.006832236424088478,
|
1439 |
+
"logits/rejected": -0.01682097464799881,
|
1440 |
+
"logps/chosen": -2.1741790771484375,
|
1441 |
+
"logps/rejected": -57.273414611816406,
|
1442 |
+
"loss": 0.2128,
|
1443 |
+
"rewards/accuracies": 0.949999988079071,
|
1444 |
+
"rewards/chosen": 17.426712036132812,
|
1445 |
+
"rewards/margins": 5.911205291748047,
|
1446 |
+
"rewards/rejected": 11.515506744384766,
|
1447 |
+
"step": 820
|
1448 |
+
},
|
1449 |
+
{
|
1450 |
+
"epoch": 0.7377777777777778,
|
1451 |
+
"grad_norm": 3.7820959091186523,
|
1452 |
+
"learning_rate": 9.770704483258782e-07,
|
1453 |
+
"logits/chosen": -0.009998206980526447,
|
1454 |
+
"logits/rejected": -0.0204261876642704,
|
1455 |
+
"logps/chosen": -1.9802953004837036,
|
1456 |
+
"logps/rejected": -56.901512145996094,
|
1457 |
+
"loss": 0.1296,
|
1458 |
+
"rewards/accuracies": 0.9750000238418579,
|
1459 |
+
"rewards/chosen": 17.56071662902832,
|
1460 |
+
"rewards/margins": 6.169583320617676,
|
1461 |
+
"rewards/rejected": 11.391134262084961,
|
1462 |
+
"step": 830
|
1463 |
+
},
|
1464 |
+
{
|
1465 |
+
"epoch": 0.7466666666666667,
|
1466 |
+
"grad_norm": 1.126626968383789,
|
1467 |
+
"learning_rate": 9.162675709666865e-07,
|
1468 |
+
"logits/chosen": -0.00826224498450756,
|
1469 |
+
"logits/rejected": -0.018977787345647812,
|
1470 |
+
"logps/chosen": -1.4256607294082642,
|
1471 |
+
"logps/rejected": -61.65986633300781,
|
1472 |
+
"loss": 0.0635,
|
1473 |
+
"rewards/accuracies": 0.987500011920929,
|
1474 |
+
"rewards/chosen": 17.324975967407227,
|
1475 |
+
"rewards/margins": 6.09440803527832,
|
1476 |
+
"rewards/rejected": 11.230567932128906,
|
1477 |
+
"step": 840
|
1478 |
+
},
|
1479 |
+
{
|
1480 |
+
"epoch": 0.7466666666666667,
|
1481 |
+
"eval_logits/chosen": -0.01111944392323494,
|
1482 |
+
"eval_logits/rejected": -0.021697774529457092,
|
1483 |
+
"eval_logps/chosen": -1.2168633937835693,
|
1484 |
+
"eval_logps/rejected": -58.2642822265625,
|
1485 |
+
"eval_loss": 0.1689341962337494,
|
1486 |
+
"eval_rewards/accuracies": 0.9450000524520874,
|
1487 |
+
"eval_rewards/chosen": 17.381206512451172,
|
1488 |
+
"eval_rewards/margins": 5.846914291381836,
|
1489 |
+
"eval_rewards/rejected": 11.534292221069336,
|
1490 |
+
"eval_runtime": 361.6192,
|
1491 |
+
"eval_samples_per_second": 2.765,
|
1492 |
+
"eval_steps_per_second": 0.346,
|
1493 |
+
"step": 840
|
1494 |
+
},
|
1495 |
+
{
|
1496 |
+
"epoch": 0.7555555555555555,
|
1497 |
+
"grad_norm": 5.250723838806152,
|
1498 |
+
"learning_rate": 8.569908029550686e-07,
|
1499 |
+
"logits/chosen": -0.006854387000203133,
|
1500 |
+
"logits/rejected": -0.018336206674575806,
|
1501 |
+
"logps/chosen": -0.6238930821418762,
|
1502 |
+
"logps/rejected": -60.925689697265625,
|
1503 |
+
"loss": 0.1157,
|
1504 |
+
"rewards/accuracies": 0.9750000238418579,
|
1505 |
+
"rewards/chosen": 17.443281173706055,
|
1506 |
+
"rewards/margins": 6.164813995361328,
|
1507 |
+
"rewards/rejected": 11.278467178344727,
|
1508 |
+
"step": 850
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 0.7644444444444445,
|
1512 |
+
"grad_norm": 3.1401162147521973,
|
1513 |
+
"learning_rate": 7.992972643121227e-07,
|
1514 |
+
"logits/chosen": -0.0037835021503269672,
|
1515 |
+
"logits/rejected": -0.013135241344571114,
|
1516 |
+
"logps/chosen": -0.8492221832275391,
|
1517 |
+
"logps/rejected": -55.516075134277344,
|
1518 |
+
"loss": 0.2252,
|
1519 |
+
"rewards/accuracies": 0.9375,
|
1520 |
+
"rewards/chosen": 17.488588333129883,
|
1521 |
+
"rewards/margins": 5.736725807189941,
|
1522 |
+
"rewards/rejected": 11.751862525939941,
|
1523 |
+
"step": 860
|
1524 |
+
},
|
1525 |
+
{
|
1526 |
+
"epoch": 0.7733333333333333,
|
1527 |
+
"grad_norm": 55.528812408447266,
|
1528 |
+
"learning_rate": 7.432425494343509e-07,
|
1529 |
+
"logits/chosen": -0.0033687639515846968,
|
1530 |
+
"logits/rejected": -0.013152632862329483,
|
1531 |
+
"logps/chosen": -1.3188884258270264,
|
1532 |
+
"logps/rejected": -57.9510498046875,
|
1533 |
+
"loss": 0.1398,
|
1534 |
+
"rewards/accuracies": 0.9624999761581421,
|
1535 |
+
"rewards/chosen": 17.412578582763672,
|
1536 |
+
"rewards/margins": 5.868515968322754,
|
1537 |
+
"rewards/rejected": 11.544061660766602,
|
1538 |
+
"step": 870
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 0.7822222222222223,
|
1542 |
+
"grad_norm": 0.039824869483709335,
|
1543 |
+
"learning_rate": 6.888806735220396e-07,
|
1544 |
+
"logits/chosen": -0.0010406378423795104,
|
1545 |
+
"logits/rejected": -0.012095071375370026,
|
1546 |
+
"logps/chosen": -2.0619027614593506,
|
1547 |
+
"logps/rejected": -59.65806579589844,
|
1548 |
+
"loss": 0.2966,
|
1549 |
+
"rewards/accuracies": 0.9375,
|
1550 |
+
"rewards/chosen": 17.192535400390625,
|
1551 |
+
"rewards/margins": 5.7006731033325195,
|
1552 |
+
"rewards/rejected": 11.491861343383789,
|
1553 |
+
"step": 880
|
1554 |
+
},
|
1555 |
+
{
|
1556 |
+
"epoch": 0.7911111111111111,
|
1557 |
+
"grad_norm": 0.7536466717720032,
|
1558 |
+
"learning_rate": 6.362640205293583e-07,
|
1559 |
+
"logits/chosen": -0.0016857212176546454,
|
1560 |
+
"logits/rejected": -0.010936147533357143,
|
1561 |
+
"logps/chosen": -2.1478958129882812,
|
1562 |
+
"logps/rejected": -58.20386505126953,
|
1563 |
+
"loss": 0.1641,
|
1564 |
+
"rewards/accuracies": 0.9375,
|
1565 |
+
"rewards/chosen": 17.243247985839844,
|
1566 |
+
"rewards/margins": 5.678771018981934,
|
1567 |
+
"rewards/rejected": 11.564477920532227,
|
1568 |
+
"step": 890
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 0.8,
|
1572 |
+
"grad_norm": 0.3857377767562866,
|
1573 |
+
"learning_rate": 5.854432926863684e-07,
|
1574 |
+
"logits/chosen": 0.00038508616853505373,
|
1575 |
+
"logits/rejected": -0.011034643277525902,
|
1576 |
+
"logps/chosen": -1.606274127960205,
|
1577 |
+
"logps/rejected": -62.58662414550781,
|
1578 |
+
"loss": 0.1703,
|
1579 |
+
"rewards/accuracies": 0.9624999761581421,
|
1580 |
+
"rewards/chosen": 17.211261749267578,
|
1581 |
+
"rewards/margins": 5.966868877410889,
|
1582 |
+
"rewards/rejected": 11.244392395019531,
|
1583 |
+
"step": 900
|
1584 |
+
},
|
1585 |
+
{
|
1586 |
+
"epoch": 0.8,
|
1587 |
+
"eval_logits/chosen": 0.00021816430671606213,
|
1588 |
+
"eval_logits/rejected": -0.010477552190423012,
|
1589 |
+
"eval_logps/chosen": -1.7576563358306885,
|
1590 |
+
"eval_logps/rejected": -59.79063415527344,
|
1591 |
+
"eval_loss": 0.13995186984539032,
|
1592 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1593 |
+
"eval_rewards/chosen": 17.327129364013672,
|
1594 |
+
"eval_rewards/margins": 5.945469856262207,
|
1595 |
+
"eval_rewards/rejected": 11.381658554077148,
|
1596 |
+
"eval_runtime": 361.4592,
|
1597 |
+
"eval_samples_per_second": 2.767,
|
1598 |
+
"eval_steps_per_second": 0.346,
|
1599 |
+
"step": 900
|
1600 |
+
},
|
1601 |
+
{
|
1602 |
+
"epoch": 0.8088888888888889,
|
1603 |
+
"grad_norm": 0.13143697381019592,
|
1604 |
+
"learning_rate": 5.364674616415547e-07,
|
1605 |
+
"logits/chosen": 0.0005570838693529367,
|
1606 |
+
"logits/rejected": -0.011198626831173897,
|
1607 |
+
"logps/chosen": -0.12791283428668976,
|
1608 |
+
"logps/rejected": -65.7052993774414,
|
1609 |
+
"loss": 0.0507,
|
1610 |
+
"rewards/accuracies": 0.9750000238418579,
|
1611 |
+
"rewards/chosen": 17.224443435668945,
|
1612 |
+
"rewards/margins": 6.163690567016602,
|
1613 |
+
"rewards/rejected": 11.060752868652344,
|
1614 |
+
"step": 910
|
1615 |
+
},
|
1616 |
+
{
|
1617 |
+
"epoch": 0.8177777777777778,
|
1618 |
+
"grad_norm": 102.9654541015625,
|
1619 |
+
"learning_rate": 4.893837212719859e-07,
|
1620 |
+
"logits/chosen": -0.0008557128603570163,
|
1621 |
+
"logits/rejected": -0.01163212489336729,
|
1622 |
+
"logps/chosen": -1.3292646408081055,
|
1623 |
+
"logps/rejected": -61.644893646240234,
|
1624 |
+
"loss": 0.0881,
|
1625 |
+
"rewards/accuracies": 0.9500000476837158,
|
1626 |
+
"rewards/chosen": 17.276538848876953,
|
1627 |
+
"rewards/margins": 5.96080207824707,
|
1628 |
+
"rewards/rejected": 11.3157377243042,
|
1629 |
+
"step": 920
|
1630 |
+
},
|
1631 |
+
{
|
1632 |
+
"epoch": 0.8266666666666667,
|
1633 |
+
"grad_norm": 146.04498291015625,
|
1634 |
+
"learning_rate": 4.442374422065493e-07,
|
1635 |
+
"logits/chosen": 0.002922601066529751,
|
1636 |
+
"logits/rejected": -0.007130052894353867,
|
1637 |
+
"logps/chosen": -1.6069023609161377,
|
1638 |
+
"logps/rejected": -61.52588653564453,
|
1639 |
+
"loss": 0.1373,
|
1640 |
+
"rewards/accuracies": 0.949999988079071,
|
1641 |
+
"rewards/chosen": 17.184974670410156,
|
1642 |
+
"rewards/margins": 5.7995734214782715,
|
1643 |
+
"rewards/rejected": 11.38540267944336,
|
1644 |
+
"step": 930
|
1645 |
+
},
|
1646 |
+
{
|
1647 |
+
"epoch": 0.8355555555555556,
|
1648 |
+
"grad_norm": 23.539485931396484,
|
1649 |
+
"learning_rate": 4.0107212810610974e-07,
|
1650 |
+
"logits/chosen": 0.0018056132830679417,
|
1651 |
+
"logits/rejected": -0.007847340777516365,
|
1652 |
+
"logps/chosen": -3.637047290802002,
|
1653 |
+
"logps/rejected": -61.21245574951172,
|
1654 |
+
"loss": 0.2763,
|
1655 |
+
"rewards/accuracies": 0.925000011920929,
|
1656 |
+
"rewards/chosen": 16.960758209228516,
|
1657 |
+
"rewards/margins": 5.551811695098877,
|
1658 |
+
"rewards/rejected": 11.40894603729248,
|
1659 |
+
"step": 940
|
1660 |
+
},
|
1661 |
+
{
|
1662 |
+
"epoch": 0.8444444444444444,
|
1663 |
+
"grad_norm": 31.52926254272461,
|
1664 |
+
"learning_rate": 3.599293737426932e-07,
|
1665 |
+
"logits/chosen": 0.0032081177923828363,
|
1666 |
+
"logits/rejected": -0.007756482809782028,
|
1667 |
+
"logps/chosen": -1.172515869140625,
|
1668 |
+
"logps/rejected": -66.5853271484375,
|
1669 |
+
"loss": 0.1758,
|
1670 |
+
"rewards/accuracies": 0.949999988079071,
|
1671 |
+
"rewards/chosen": 16.992876052856445,
|
1672 |
+
"rewards/margins": 5.886469841003418,
|
1673 |
+
"rewards/rejected": 11.106407165527344,
|
1674 |
+
"step": 950
|
1675 |
+
},
|
1676 |
+
{
|
1677 |
+
"epoch": 0.8533333333333334,
|
1678 |
+
"grad_norm": 0.07474468648433685,
|
1679 |
+
"learning_rate": 3.208488249181216e-07,
|
1680 |
+
"logits/chosen": 0.0022508346009999514,
|
1681 |
+
"logits/rejected": -0.009156409651041031,
|
1682 |
+
"logps/chosen": -0.7930470108985901,
|
1683 |
+
"logps/rejected": -60.068790435791016,
|
1684 |
+
"loss": 0.1138,
|
1685 |
+
"rewards/accuracies": 0.9624999761581421,
|
1686 |
+
"rewards/chosen": 17.401771545410156,
|
1687 |
+
"rewards/margins": 6.0026960372924805,
|
1688 |
+
"rewards/rejected": 11.399076461791992,
|
1689 |
+
"step": 960
|
1690 |
+
},
|
1691 |
+
{
|
1692 |
+
"epoch": 0.8533333333333334,
|
1693 |
+
"eval_logits/chosen": 0.001469604205340147,
|
1694 |
+
"eval_logits/rejected": -0.009397665038704872,
|
1695 |
+
"eval_logps/chosen": -1.8795456886291504,
|
1696 |
+
"eval_logps/rejected": -60.17564010620117,
|
1697 |
+
"eval_loss": 0.1441129744052887,
|
1698 |
+
"eval_rewards/accuracies": 0.9630000591278076,
|
1699 |
+
"eval_rewards/chosen": 17.314937591552734,
|
1700 |
+
"eval_rewards/margins": 5.9717817306518555,
|
1701 |
+
"eval_rewards/rejected": 11.343156814575195,
|
1702 |
+
"eval_runtime": 361.5344,
|
1703 |
+
"eval_samples_per_second": 2.766,
|
1704 |
+
"eval_steps_per_second": 0.346,
|
1705 |
+
"step": 960
|
1706 |
+
},
|
1707 |
+
{
|
1708 |
+
"epoch": 0.8622222222222222,
|
1709 |
+
"grad_norm": 68.90747833251953,
|
1710 |
+
"learning_rate": 2.838681402606952e-07,
|
1711 |
+
"logits/chosen": 0.004552370868623257,
|
1712 |
+
"logits/rejected": -0.005488495342433453,
|
1713 |
+
"logps/chosen": -3.7298974990844727,
|
1714 |
+
"logps/rejected": -64.72488403320312,
|
1715 |
+
"loss": 0.2425,
|
1716 |
+
"rewards/accuracies": 0.925000011920929,
|
1717 |
+
"rewards/chosen": 16.832754135131836,
|
1718 |
+
"rewards/margins": 5.644216537475586,
|
1719 |
+
"rewards/rejected": 11.18853759765625,
|
1720 |
+
"step": 970
|
1721 |
+
},
|
1722 |
+
{
|
1723 |
+
"epoch": 0.8711111111111111,
|
1724 |
+
"grad_norm": 0.028206102550029755,
|
1725 |
+
"learning_rate": 2.490229549367443e-07,
|
1726 |
+
"logits/chosen": 0.0025807656347751617,
|
1727 |
+
"logits/rejected": -0.008657123893499374,
|
1728 |
+
"logps/chosen": -1.1825838088989258,
|
1729 |
+
"logps/rejected": -63.57493591308594,
|
1730 |
+
"loss": 0.0609,
|
1731 |
+
"rewards/accuracies": 0.9750000238418579,
|
1732 |
+
"rewards/chosen": 17.227130889892578,
|
1733 |
+
"rewards/margins": 6.053717136383057,
|
1734 |
+
"rewards/rejected": 11.173412322998047,
|
1735 |
+
"step": 980
|
1736 |
+
},
|
1737 |
+
{
|
1738 |
+
"epoch": 0.88,
|
1739 |
+
"grad_norm": 0.39742231369018555,
|
1740 |
+
"learning_rate": 2.1634684631203412e-07,
|
1741 |
+
"logits/chosen": 0.0048486413434147835,
|
1742 |
+
"logits/rejected": -0.006056814920157194,
|
1743 |
+
"logps/chosen": -2.919680595397949,
|
1744 |
+
"logps/rejected": -57.8403205871582,
|
1745 |
+
"loss": 0.3464,
|
1746 |
+
"rewards/accuracies": 0.8999999761581421,
|
1747 |
+
"rewards/chosen": 17.175275802612305,
|
1748 |
+
"rewards/margins": 5.559727668762207,
|
1749 |
+
"rewards/rejected": 11.615548133850098,
|
1750 |
+
"step": 990
|
1751 |
+
},
|
1752 |
+
{
|
1753 |
+
"epoch": 0.8888888888888888,
|
1754 |
+
"grad_norm": 30.22509002685547,
|
1755 |
+
"learning_rate": 1.8587130159608196e-07,
|
1756 |
+
"logits/chosen": 0.0050649940967559814,
|
1757 |
+
"logits/rejected": -0.0071399761363863945,
|
1758 |
+
"logps/chosen": -0.16131475567817688,
|
1759 |
+
"logps/rejected": -66.25190734863281,
|
1760 |
+
"loss": 0.0063,
|
1761 |
+
"rewards/accuracies": 1.0,
|
1762 |
+
"rewards/chosen": 17.313472747802734,
|
1763 |
+
"rewards/margins": 6.412895679473877,
|
1764 |
+
"rewards/rejected": 10.9005765914917,
|
1765 |
+
"step": 1000
|
1766 |
+
},
|
1767 |
+
{
|
1768 |
+
"epoch": 0.8977777777777778,
|
1769 |
+
"grad_norm": 53.02751922607422,
|
1770 |
+
"learning_rate": 1.5762568750059604e-07,
|
1771 |
+
"logits/chosen": 0.005312003195285797,
|
1772 |
+
"logits/rejected": -0.0038596936501562595,
|
1773 |
+
"logps/chosen": -4.00323486328125,
|
1774 |
+
"logps/rejected": -57.656890869140625,
|
1775 |
+
"loss": 0.2741,
|
1776 |
+
"rewards/accuracies": 0.925000011920929,
|
1777 |
+
"rewards/chosen": 17.167781829833984,
|
1778 |
+
"rewards/margins": 5.632095813751221,
|
1779 |
+
"rewards/rejected": 11.535685539245605,
|
1780 |
+
"step": 1010
|
1781 |
+
},
|
1782 |
+
{
|
1783 |
+
"epoch": 0.9066666666666666,
|
1784 |
+
"grad_norm": 119.56330108642578,
|
1785 |
+
"learning_rate": 1.316372219412454e-07,
|
1786 |
+
"logits/chosen": 0.004333779215812683,
|
1787 |
+
"logits/rejected": -0.007937717251479626,
|
1788 |
+
"logps/chosen": -0.10221245884895325,
|
1789 |
+
"logps/rejected": -67.09977722167969,
|
1790 |
+
"loss": 0.0513,
|
1791 |
+
"rewards/accuracies": 0.9750000238418579,
|
1792 |
+
"rewards/chosen": 17.151248931884766,
|
1793 |
+
"rewards/margins": 6.162137985229492,
|
1794 |
+
"rewards/rejected": 10.98911190032959,
|
1795 |
+
"step": 1020
|
1796 |
+
},
|
1797 |
+
{
|
1798 |
+
"epoch": 0.9066666666666666,
|
1799 |
+
"eval_logits/chosen": 0.004528110846877098,
|
1800 |
+
"eval_logits/rejected": -0.006501312367618084,
|
1801 |
+
"eval_logps/chosen": -1.817779302597046,
|
1802 |
+
"eval_logps/rejected": -60.34454345703125,
|
1803 |
+
"eval_loss": 0.14118175208568573,
|
1804 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1805 |
+
"eval_rewards/chosen": 17.321117401123047,
|
1806 |
+
"eval_rewards/margins": 5.994848251342773,
|
1807 |
+
"eval_rewards/rejected": 11.32626724243164,
|
1808 |
+
"eval_runtime": 361.545,
|
1809 |
+
"eval_samples_per_second": 2.766,
|
1810 |
+
"eval_steps_per_second": 0.346,
|
1811 |
+
"step": 1020
|
1812 |
+
},
|
1813 |
+
{
|
1814 |
+
"epoch": 0.9155555555555556,
|
1815 |
+
"grad_norm": 0.9030271768569946,
|
1816 |
+
"learning_rate": 1.0793094781005792e-07,
|
1817 |
+
"logits/chosen": 0.0036234352737665176,
|
1818 |
+
"logits/rejected": -0.008390933275222778,
|
1819 |
+
"logps/chosen": -0.09899584949016571,
|
1820 |
+
"logps/rejected": -60.80555725097656,
|
1821 |
+
"loss": 0.0349,
|
1822 |
+
"rewards/accuracies": 0.987500011920929,
|
1823 |
+
"rewards/chosen": 17.619266510009766,
|
1824 |
+
"rewards/margins": 6.4814043045043945,
|
1825 |
+
"rewards/rejected": 11.137863159179688,
|
1826 |
+
"step": 1030
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 0.9244444444444444,
|
1830 |
+
"grad_norm": 157.070068359375,
|
1831 |
+
"learning_rate": 8.652970884369255e-08,
|
1832 |
+
"logits/chosen": 0.005321115255355835,
|
1833 |
+
"logits/rejected": -0.004976513795554638,
|
1834 |
+
"logps/chosen": -1.3535833358764648,
|
1835 |
+
"logps/rejected": -61.2861328125,
|
1836 |
+
"loss": 0.1927,
|
1837 |
+
"rewards/accuracies": 0.9500000476837158,
|
1838 |
+
"rewards/chosen": 17.209800720214844,
|
1839 |
+
"rewards/margins": 5.795952796936035,
|
1840 |
+
"rewards/rejected": 11.413846015930176,
|
1841 |
+
"step": 1040
|
1842 |
+
},
|
1843 |
+
{
|
1844 |
+
"epoch": 0.9333333333333333,
|
1845 |
+
"grad_norm": 24.539953231811523,
|
1846 |
+
"learning_rate": 6.745412761086007e-08,
|
1847 |
+
"logits/chosen": 0.005845514126121998,
|
1848 |
+
"logits/rejected": -0.003921338357031345,
|
1849 |
+
"logps/chosen": -1.7223193645477295,
|
1850 |
+
"logps/rejected": -58.172142028808594,
|
1851 |
+
"loss": 0.1317,
|
1852 |
+
"rewards/accuracies": 0.9750000238418579,
|
1853 |
+
"rewards/chosen": 17.360435485839844,
|
1854 |
+
"rewards/margins": 5.838218688964844,
|
1855 |
+
"rewards/rejected": 11.522216796875,
|
1856 |
+
"step": 1050
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 0.9422222222222222,
|
1860 |
+
"grad_norm": 0.061458222568035126,
|
1861 |
+
"learning_rate": 5.0722585640090305e-08,
|
1862 |
+
"logits/chosen": 0.003759522922337055,
|
1863 |
+
"logits/rejected": -0.0068417866714298725,
|
1864 |
+
"logps/chosen": -1.351855754852295,
|
1865 |
+
"logps/rejected": -59.56566619873047,
|
1866 |
+
"loss": 0.2279,
|
1867 |
+
"rewards/accuracies": 0.9375,
|
1868 |
+
"rewards/chosen": 17.245838165283203,
|
1869 |
+
"rewards/margins": 5.7038679122924805,
|
1870 |
+
"rewards/rejected": 11.541970252990723,
|
1871 |
+
"step": 1060
|
1872 |
+
},
|
1873 |
+
{
|
1874 |
+
"epoch": 0.9511111111111111,
|
1875 |
+
"grad_norm": 101.87332916259766,
|
1876 |
+
"learning_rate": 3.635120570700784e-08,
|
1877 |
+
"logits/chosen": 0.007607857696712017,
|
1878 |
+
"logits/rejected": -0.0035320711322128773,
|
1879 |
+
"logps/chosen": -1.4004669189453125,
|
1880 |
+
"logps/rejected": -63.46752166748047,
|
1881 |
+
"loss": 0.1021,
|
1882 |
+
"rewards/accuracies": 0.9624999761581421,
|
1883 |
+
"rewards/chosen": 17.240036010742188,
|
1884 |
+
"rewards/margins": 6.078882694244385,
|
1885 |
+
"rewards/rejected": 11.161155700683594,
|
1886 |
+
"step": 1070
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 0.96,
|
1890 |
+
"grad_norm": 1.9229660034179688,
|
1891 |
+
"learning_rate": 2.4353836298169343e-08,
|
1892 |
+
"logits/chosen": 0.006209026090800762,
|
1893 |
+
"logits/rejected": -0.0033816141076385975,
|
1894 |
+
"logps/chosen": -2.0281822681427,
|
1895 |
+
"logps/rejected": -61.460670471191406,
|
1896 |
+
"loss": 0.1189,
|
1897 |
+
"rewards/accuracies": 0.9750000238418579,
|
1898 |
+
"rewards/chosen": 17.19972801208496,
|
1899 |
+
"rewards/margins": 5.881450653076172,
|
1900 |
+
"rewards/rejected": 11.318277359008789,
|
1901 |
+
"step": 1080
|
1902 |
+
},
|
1903 |
+
{
|
1904 |
+
"epoch": 0.96,
|
1905 |
+
"eval_logits/chosen": 0.00737445754930377,
|
1906 |
+
"eval_logits/rejected": -0.0035832570865750313,
|
1907 |
+
"eval_logps/chosen": -2.1419789791107178,
|
1908 |
+
"eval_logps/rejected": -60.60612106323242,
|
1909 |
+
"eval_loss": 0.15081512928009033,
|
1910 |
+
"eval_rewards/accuracies": 0.9610000848770142,
|
1911 |
+
"eval_rewards/chosen": 17.288694381713867,
|
1912 |
+
"eval_rewards/margins": 5.98858642578125,
|
1913 |
+
"eval_rewards/rejected": 11.300108909606934,
|
1914 |
+
"eval_runtime": 361.4165,
|
1915 |
+
"eval_samples_per_second": 2.767,
|
1916 |
+
"eval_steps_per_second": 0.346,
|
1917 |
+
"step": 1080
|
1918 |
+
},
|
1919 |
+
{
|
1920 |
+
"epoch": 0.9688888888888889,
|
1921 |
+
"grad_norm": 0.11888863146305084,
|
1922 |
+
"learning_rate": 1.4742038266447046e-08,
|
1923 |
+
"logits/chosen": 0.006043245084583759,
|
1924 |
+
"logits/rejected": -0.004712546244263649,
|
1925 |
+
"logps/chosen": -1.4127472639083862,
|
1926 |
+
"logps/rejected": -64.81529998779297,
|
1927 |
+
"loss": 0.0728,
|
1928 |
+
"rewards/accuracies": 0.9750000238418579,
|
1929 |
+
"rewards/chosen": 17.181049346923828,
|
1930 |
+
"rewards/margins": 6.14100456237793,
|
1931 |
+
"rewards/rejected": 11.040044784545898,
|
1932 |
+
"step": 1090
|
1933 |
+
},
|
1934 |
+
{
|
1935 |
+
"epoch": 0.9777777777777777,
|
1936 |
+
"grad_norm": 2.001819372177124,
|
1937 |
+
"learning_rate": 7.525073690809737e-09,
|
1938 |
+
"logits/chosen": 0.0055408780463039875,
|
1939 |
+
"logits/rejected": -0.005598037503659725,
|
1940 |
+
"logps/chosen": -1.017110824584961,
|
1941 |
+
"logps/rejected": -60.28044891357422,
|
1942 |
+
"loss": 0.0357,
|
1943 |
+
"rewards/accuracies": 0.9750000238418579,
|
1944 |
+
"rewards/chosen": 17.4679012298584,
|
1945 |
+
"rewards/margins": 6.193048477172852,
|
1946 |
+
"rewards/rejected": 11.274852752685547,
|
1947 |
+
"step": 1100
|
1948 |
+
},
|
1949 |
+
{
|
1950 |
+
"epoch": 0.9866666666666667,
|
1951 |
+
"grad_norm": 0.027056939899921417,
|
1952 |
+
"learning_rate": 2.709896951238744e-09,
|
1953 |
+
"logits/chosen": 0.00636716466397047,
|
1954 |
+
"logits/rejected": -0.004360577557235956,
|
1955 |
+
"logps/chosen": -2.8028905391693115,
|
1956 |
+
"logps/rejected": -60.86518859863281,
|
1957 |
+
"loss": 0.1973,
|
1958 |
+
"rewards/accuracies": 0.949999988079071,
|
1959 |
+
"rewards/chosen": 17.210033416748047,
|
1960 |
+
"rewards/margins": 5.936794281005859,
|
1961 |
+
"rewards/rejected": 11.273238182067871,
|
1962 |
+
"step": 1110
|
1963 |
+
},
|
1964 |
+
{
|
1965 |
+
"epoch": 0.9955555555555555,
|
1966 |
+
"grad_norm": 0.15637506544589996,
|
1967 |
+
"learning_rate": 3.0114802737818415e-10,
|
1968 |
+
"logits/chosen": 0.0038183885626494884,
|
1969 |
+
"logits/rejected": -0.007296917960047722,
|
1970 |
+
"logps/chosen": -1.1477452516555786,
|
1971 |
+
"logps/rejected": -59.37739181518555,
|
1972 |
+
"loss": 0.0538,
|
1973 |
+
"rewards/accuracies": 0.987500011920929,
|
1974 |
+
"rewards/chosen": 17.553251266479492,
|
1975 |
+
"rewards/margins": 6.2790141105651855,
|
1976 |
+
"rewards/rejected": 11.274236679077148,
|
1977 |
+
"step": 1120
|
1978 |
+
},
|
1979 |
+
{
|
1980 |
+
"epoch": 1.0,
|
1981 |
+
"step": 1125,
|
1982 |
+
"total_flos": 1.4338459346927616e+18,
|
1983 |
+
"train_loss": 0.23597783709896936,
|
1984 |
+
"train_runtime": 13850.7044,
|
1985 |
+
"train_samples_per_second": 0.65,
|
1986 |
+
"train_steps_per_second": 0.081
|
1987 |
+
}
|
1988 |
+
],
|
1989 |
+
"logging_steps": 10,
|
1990 |
+
"max_steps": 1125,
|
1991 |
+
"num_input_tokens_seen": 0,
|
1992 |
+
"num_train_epochs": 1,
|
1993 |
+
"save_steps": 500,
|
1994 |
+
"stateful_callbacks": {
|
1995 |
+
"TrainerControl": {
|
1996 |
+
"args": {
|
1997 |
+
"should_epoch_stop": false,
|
1998 |
+
"should_evaluate": false,
|
1999 |
+
"should_log": false,
|
2000 |
+
"should_save": true,
|
2001 |
+
"should_training_stop": true
|
2002 |
+
},
|
2003 |
+
"attributes": {}
|
2004 |
+
}
|
2005 |
+
},
|
2006 |
+
"total_flos": 1.4338459346927616e+18,
|
2007 |
+
"train_batch_size": 4,
|
2008 |
+
"trial_name": null,
|
2009 |
+
"trial_params": null
|
2010 |
+
}
|