Model save
Browse files- README.md +73 -0
- added_tokens.json +5 -0
- all_results.json +9 -0
- config.json +28 -0
- generation_config.json +14 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- runs/Jun21_14-04-01_131-186-38-174/events.out.tfevents.1718979292.131-186-38-174.21907.0 +3 -0
- special_tokens_map.json +21 -0
- tokenizer.json +0 -0
- tokenizer_config.json +43 -0
- train_results.json +9 -0
- trainer_state.json +1468 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: Qwen/Qwen2-0.5B-Instruct
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: qwen2-0.5b-instruct-simpo-lr-5e-07-gamma-1.5
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# qwen2-0.5b-instruct-simpo-lr-5e-07-gamma-1.5
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [Qwen/Qwen2-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2-0.5B-Instruct) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 1.6598
|
| 21 |
+
- Rewards/chosen: -3.3421
|
| 22 |
+
- Rewards/rejected: -3.4735
|
| 23 |
+
- Rewards/accuracies: 0.5282
|
| 24 |
+
- Rewards/margins: 0.1314
|
| 25 |
+
- Logps/rejected: -1.3894
|
| 26 |
+
- Logps/chosen: -1.3368
|
| 27 |
+
- Logits/rejected: -5.2590
|
| 28 |
+
- Logits/chosen: -5.3573
|
| 29 |
+
|
| 30 |
+
## Model description
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Intended uses & limitations
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training and evaluation data
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Training procedure
|
| 43 |
+
|
| 44 |
+
### Training hyperparameters
|
| 45 |
+
|
| 46 |
+
The following hyperparameters were used during training:
|
| 47 |
+
- learning_rate: 5e-07
|
| 48 |
+
- train_batch_size: 2
|
| 49 |
+
- eval_batch_size: 4
|
| 50 |
+
- seed: 42
|
| 51 |
+
- distributed_type: multi-GPU
|
| 52 |
+
- num_devices: 8
|
| 53 |
+
- gradient_accumulation_steps: 8
|
| 54 |
+
- total_train_batch_size: 128
|
| 55 |
+
- total_eval_batch_size: 32
|
| 56 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 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/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 64 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 65 |
+
| 1.6693 | 0.8549 | 400 | 1.6598 | -3.3421 | -3.4735 | 0.5282 | 0.1314 | -1.3894 | -1.3368 | -5.2590 | -5.3573 |
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
### Framework versions
|
| 69 |
+
|
| 70 |
+
- Transformers 4.41.2
|
| 71 |
+
- Pytorch 2.3.1+cu121
|
| 72 |
+
- Datasets 2.20.0
|
| 73 |
+
- Tokenizers 0.19.1
|
added_tokens.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<|endoftext|>": 151643,
|
| 3 |
+
"<|im_end|>": 151645,
|
| 4 |
+
"<|im_start|>": 151644
|
| 5 |
+
}
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9981298423724285,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.7099600828689476,
|
| 5 |
+
"train_runtime": 3104.7014,
|
| 6 |
+
"train_samples": 59876,
|
| 7 |
+
"train_samples_per_second": 19.286,
|
| 8 |
+
"train_steps_per_second": 0.15
|
| 9 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Qwen/Qwen2-0.5B-Instruct",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Qwen2ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 151643,
|
| 8 |
+
"eos_token_id": 151645,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 896,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 4864,
|
| 13 |
+
"max_position_embeddings": 32768,
|
| 14 |
+
"max_window_layers": 24,
|
| 15 |
+
"model_type": "qwen2",
|
| 16 |
+
"num_attention_heads": 14,
|
| 17 |
+
"num_hidden_layers": 24,
|
| 18 |
+
"num_key_value_heads": 2,
|
| 19 |
+
"rms_norm_eps": 1e-06,
|
| 20 |
+
"rope_theta": 1000000.0,
|
| 21 |
+
"sliding_window": 32768,
|
| 22 |
+
"tie_word_embeddings": true,
|
| 23 |
+
"torch_dtype": "bfloat16",
|
| 24 |
+
"transformers_version": "4.41.2",
|
| 25 |
+
"use_cache": false,
|
| 26 |
+
"use_sliding_window": false,
|
| 27 |
+
"vocab_size": 151936
|
| 28 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.1,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.41.2"
|
| 14 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9948210d1d9c80524b8794e5f06f8432d564886c86f013d2ed7094e400c6e3a0
|
| 3 |
+
size 988097824
|
runs/Jun21_14-04-01_131-186-38-174/events.out.tfevents.1718979292.131-186-38-174.21907.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:266711b2c7f36729c36f562721d3de29764389dcb7fa666a3b15de18d96c547a
|
| 3 |
+
size 70644
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|im_start|>",
|
| 4 |
+
"<|im_end|>"
|
| 5 |
+
],
|
| 6 |
+
"bos_token": "<|im_end|>",
|
| 7 |
+
"eos_token": {
|
| 8 |
+
"content": "<|im_end|>",
|
| 9 |
+
"lstrip": false,
|
| 10 |
+
"normalized": false,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"single_word": false
|
| 13 |
+
},
|
| 14 |
+
"pad_token": {
|
| 15 |
+
"content": "<|endoftext|>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false
|
| 20 |
+
}
|
| 21 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"151643": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"151644": {
|
| 13 |
+
"content": "<|im_start|>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"151645": {
|
| 21 |
+
"content": "<|im_end|>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
}
|
| 28 |
+
},
|
| 29 |
+
"additional_special_tokens": [
|
| 30 |
+
"<|im_start|>",
|
| 31 |
+
"<|im_end|>"
|
| 32 |
+
],
|
| 33 |
+
"bos_token": "<|im_end|>",
|
| 34 |
+
"chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
|
| 35 |
+
"clean_up_tokenization_spaces": false,
|
| 36 |
+
"eos_token": "<|im_end|>",
|
| 37 |
+
"errors": "replace",
|
| 38 |
+
"model_max_length": 32768,
|
| 39 |
+
"pad_token": "<|endoftext|>",
|
| 40 |
+
"split_special_tokens": false,
|
| 41 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 42 |
+
"unk_token": null
|
| 43 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9981298423724285,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.7099600828689476,
|
| 5 |
+
"train_runtime": 3104.7014,
|
| 6 |
+
"train_samples": 59876,
|
| 7 |
+
"train_samples_per_second": 19.286,
|
| 8 |
+
"train_steps_per_second": 0.15
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1468 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9981298423724285,
|
| 5 |
+
"eval_steps": 400,
|
| 6 |
+
"global_step": 467,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0021373230029388193,
|
| 13 |
+
"grad_norm": 8.463839594398154,
|
| 14 |
+
"learning_rate": 1.0638297872340425e-08,
|
| 15 |
+
"logits/chosen": -5.03189754486084,
|
| 16 |
+
"logits/rejected": -4.649786472320557,
|
| 17 |
+
"logps/chosen": -1.325859785079956,
|
| 18 |
+
"logps/rejected": -1.2691845893859863,
|
| 19 |
+
"loss": 1.6361,
|
| 20 |
+
"rewards/accuracies": 0.375,
|
| 21 |
+
"rewards/chosen": -3.3146491050720215,
|
| 22 |
+
"rewards/margins": -0.1416877955198288,
|
| 23 |
+
"rewards/rejected": -3.172961473464966,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.010686615014694095,
|
| 28 |
+
"grad_norm": 10.461652514268042,
|
| 29 |
+
"learning_rate": 5.3191489361702123e-08,
|
| 30 |
+
"logits/chosen": -4.613029479980469,
|
| 31 |
+
"logits/rejected": -4.432791709899902,
|
| 32 |
+
"logps/chosen": -1.3983737230300903,
|
| 33 |
+
"logps/rejected": -1.4233554601669312,
|
| 34 |
+
"loss": 1.8223,
|
| 35 |
+
"rewards/accuracies": 0.453125,
|
| 36 |
+
"rewards/chosen": -3.495934247970581,
|
| 37 |
+
"rewards/margins": 0.0624544657766819,
|
| 38 |
+
"rewards/rejected": -3.5583884716033936,
|
| 39 |
+
"step": 5
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.02137323002938819,
|
| 43 |
+
"grad_norm": 14.978344124531132,
|
| 44 |
+
"learning_rate": 1.0638297872340425e-07,
|
| 45 |
+
"logits/chosen": -4.634655952453613,
|
| 46 |
+
"logits/rejected": -4.519498348236084,
|
| 47 |
+
"logps/chosen": -1.5365169048309326,
|
| 48 |
+
"logps/rejected": -1.5281957387924194,
|
| 49 |
+
"loss": 1.8105,
|
| 50 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 51 |
+
"rewards/chosen": -3.8412926197052,
|
| 52 |
+
"rewards/margins": -0.020802771672606468,
|
| 53 |
+
"rewards/rejected": -3.8204894065856934,
|
| 54 |
+
"step": 10
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.03205984504408229,
|
| 58 |
+
"grad_norm": 9.265143341850845,
|
| 59 |
+
"learning_rate": 1.5957446808510638e-07,
|
| 60 |
+
"logits/chosen": -4.536217212677002,
|
| 61 |
+
"logits/rejected": -4.47308874130249,
|
| 62 |
+
"logps/chosen": -1.5140950679779053,
|
| 63 |
+
"logps/rejected": -1.4939788579940796,
|
| 64 |
+
"loss": 1.748,
|
| 65 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 66 |
+
"rewards/chosen": -3.785238265991211,
|
| 67 |
+
"rewards/margins": -0.05029074102640152,
|
| 68 |
+
"rewards/rejected": -3.7349472045898438,
|
| 69 |
+
"step": 15
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.04274646005877638,
|
| 73 |
+
"grad_norm": 7.20742555554014,
|
| 74 |
+
"learning_rate": 2.127659574468085e-07,
|
| 75 |
+
"logits/chosen": -4.679377555847168,
|
| 76 |
+
"logits/rejected": -4.679813385009766,
|
| 77 |
+
"logps/chosen": -1.4990894794464111,
|
| 78 |
+
"logps/rejected": -1.5615761280059814,
|
| 79 |
+
"loss": 1.7569,
|
| 80 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 81 |
+
"rewards/chosen": -3.7477238178253174,
|
| 82 |
+
"rewards/margins": 0.15621671080589294,
|
| 83 |
+
"rewards/rejected": -3.9039406776428223,
|
| 84 |
+
"step": 20
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.053433075073470476,
|
| 88 |
+
"grad_norm": 8.314479583512021,
|
| 89 |
+
"learning_rate": 2.659574468085106e-07,
|
| 90 |
+
"logits/chosen": -4.735424041748047,
|
| 91 |
+
"logits/rejected": -4.67797327041626,
|
| 92 |
+
"logps/chosen": -1.5032179355621338,
|
| 93 |
+
"logps/rejected": -1.5036437511444092,
|
| 94 |
+
"loss": 1.7663,
|
| 95 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 96 |
+
"rewards/chosen": -3.758044719696045,
|
| 97 |
+
"rewards/margins": 0.001064938260242343,
|
| 98 |
+
"rewards/rejected": -3.7591099739074707,
|
| 99 |
+
"step": 25
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.06411969008816458,
|
| 103 |
+
"grad_norm": 8.24895452301186,
|
| 104 |
+
"learning_rate": 3.1914893617021275e-07,
|
| 105 |
+
"logits/chosen": -4.979925632476807,
|
| 106 |
+
"logits/rejected": -4.7350873947143555,
|
| 107 |
+
"logps/chosen": -1.526336431503296,
|
| 108 |
+
"logps/rejected": -1.5037505626678467,
|
| 109 |
+
"loss": 1.7647,
|
| 110 |
+
"rewards/accuracies": 0.512499988079071,
|
| 111 |
+
"rewards/chosen": -3.81584095954895,
|
| 112 |
+
"rewards/margins": -0.05646459385752678,
|
| 113 |
+
"rewards/rejected": -3.7593765258789062,
|
| 114 |
+
"step": 30
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.07480630510285867,
|
| 118 |
+
"grad_norm": 8.218009051807567,
|
| 119 |
+
"learning_rate": 3.7234042553191484e-07,
|
| 120 |
+
"logits/chosen": -4.755112171173096,
|
| 121 |
+
"logits/rejected": -4.576828479766846,
|
| 122 |
+
"logps/chosen": -1.5506718158721924,
|
| 123 |
+
"logps/rejected": -1.5067684650421143,
|
| 124 |
+
"loss": 1.7863,
|
| 125 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 126 |
+
"rewards/chosen": -3.876680374145508,
|
| 127 |
+
"rewards/margins": -0.10975880920886993,
|
| 128 |
+
"rewards/rejected": -3.7669215202331543,
|
| 129 |
+
"step": 35
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.08549292011755276,
|
| 133 |
+
"grad_norm": 10.346763541049816,
|
| 134 |
+
"learning_rate": 4.25531914893617e-07,
|
| 135 |
+
"logits/chosen": -4.490562915802002,
|
| 136 |
+
"logits/rejected": -4.430506706237793,
|
| 137 |
+
"logps/chosen": -1.4821277856826782,
|
| 138 |
+
"logps/rejected": -1.5026350021362305,
|
| 139 |
+
"loss": 1.7765,
|
| 140 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 141 |
+
"rewards/chosen": -3.7053191661834717,
|
| 142 |
+
"rewards/margins": 0.05126795917749405,
|
| 143 |
+
"rewards/rejected": -3.756587266921997,
|
| 144 |
+
"step": 40
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.09617953513224686,
|
| 148 |
+
"grad_norm": 8.159954026995056,
|
| 149 |
+
"learning_rate": 4.787234042553192e-07,
|
| 150 |
+
"logits/chosen": -4.584813594818115,
|
| 151 |
+
"logits/rejected": -4.397616386413574,
|
| 152 |
+
"logps/chosen": -1.5396957397460938,
|
| 153 |
+
"logps/rejected": -1.5210858583450317,
|
| 154 |
+
"loss": 1.7353,
|
| 155 |
+
"rewards/accuracies": 0.4375,
|
| 156 |
+
"rewards/chosen": -3.8492393493652344,
|
| 157 |
+
"rewards/margins": -0.046524763107299805,
|
| 158 |
+
"rewards/rejected": -3.8027145862579346,
|
| 159 |
+
"step": 45
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.10686615014694095,
|
| 163 |
+
"grad_norm": 6.871700800549116,
|
| 164 |
+
"learning_rate": 4.999370587356267e-07,
|
| 165 |
+
"logits/chosen": -4.505709648132324,
|
| 166 |
+
"logits/rejected": -4.605370998382568,
|
| 167 |
+
"logps/chosen": -1.4193190336227417,
|
| 168 |
+
"logps/rejected": -1.5022988319396973,
|
| 169 |
+
"loss": 1.7108,
|
| 170 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 171 |
+
"rewards/chosen": -3.54829740524292,
|
| 172 |
+
"rewards/margins": 0.20744964480400085,
|
| 173 |
+
"rewards/rejected": -3.7557475566864014,
|
| 174 |
+
"step": 50
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.11755276516163506,
|
| 178 |
+
"grad_norm": 8.013382748882483,
|
| 179 |
+
"learning_rate": 4.995525324419337e-07,
|
| 180 |
+
"logits/chosen": -4.650801658630371,
|
| 181 |
+
"logits/rejected": -4.462812900543213,
|
| 182 |
+
"logps/chosen": -1.3039531707763672,
|
| 183 |
+
"logps/rejected": -1.3392069339752197,
|
| 184 |
+
"loss": 1.7299,
|
| 185 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 186 |
+
"rewards/chosen": -3.259882688522339,
|
| 187 |
+
"rewards/margins": 0.08813469856977463,
|
| 188 |
+
"rewards/rejected": -3.348017454147339,
|
| 189 |
+
"step": 55
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.12823938017632916,
|
| 193 |
+
"grad_norm": 8.303134763756917,
|
| 194 |
+
"learning_rate": 4.988189843662815e-07,
|
| 195 |
+
"logits/chosen": -4.751210689544678,
|
| 196 |
+
"logits/rejected": -4.686429023742676,
|
| 197 |
+
"logps/chosen": -1.2432037591934204,
|
| 198 |
+
"logps/rejected": -1.2768182754516602,
|
| 199 |
+
"loss": 1.7359,
|
| 200 |
+
"rewards/accuracies": 0.550000011920929,
|
| 201 |
+
"rewards/chosen": -3.1080093383789062,
|
| 202 |
+
"rewards/margins": 0.08403660356998444,
|
| 203 |
+
"rewards/rejected": -3.1920459270477295,
|
| 204 |
+
"step": 60
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.13892599519102325,
|
| 208 |
+
"grad_norm": 6.777024673739444,
|
| 209 |
+
"learning_rate": 4.977374404419837e-07,
|
| 210 |
+
"logits/chosen": -4.5330071449279785,
|
| 211 |
+
"logits/rejected": -4.71773099899292,
|
| 212 |
+
"logps/chosen": -1.3066389560699463,
|
| 213 |
+
"logps/rejected": -1.2977256774902344,
|
| 214 |
+
"loss": 1.7457,
|
| 215 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 216 |
+
"rewards/chosen": -3.266597270965576,
|
| 217 |
+
"rewards/margins": -0.02228318527340889,
|
| 218 |
+
"rewards/rejected": -3.244314193725586,
|
| 219 |
+
"step": 65
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.14961261020571734,
|
| 223 |
+
"grad_norm": 9.380383365435003,
|
| 224 |
+
"learning_rate": 4.963094133060148e-07,
|
| 225 |
+
"logits/chosen": -4.556839466094971,
|
| 226 |
+
"logits/rejected": -4.482637882232666,
|
| 227 |
+
"logps/chosen": -1.3503261804580688,
|
| 228 |
+
"logps/rejected": -1.309463620185852,
|
| 229 |
+
"loss": 1.7783,
|
| 230 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 231 |
+
"rewards/chosen": -3.3758158683776855,
|
| 232 |
+
"rewards/margins": -0.1021568551659584,
|
| 233 |
+
"rewards/rejected": -3.2736592292785645,
|
| 234 |
+
"step": 70
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.16029922522041143,
|
| 238 |
+
"grad_norm": 8.316182420752773,
|
| 239 |
+
"learning_rate": 4.945369001834514e-07,
|
| 240 |
+
"logits/chosen": -4.512010097503662,
|
| 241 |
+
"logits/rejected": -4.473996639251709,
|
| 242 |
+
"logps/chosen": -1.3320636749267578,
|
| 243 |
+
"logps/rejected": -1.3820534944534302,
|
| 244 |
+
"loss": 1.6619,
|
| 245 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 246 |
+
"rewards/chosen": -3.3301587104797363,
|
| 247 |
+
"rewards/margins": 0.12497501075267792,
|
| 248 |
+
"rewards/rejected": -3.455134153366089,
|
| 249 |
+
"step": 75
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.17098584023510552,
|
| 253 |
+
"grad_norm": 7.968870017741486,
|
| 254 |
+
"learning_rate": 4.924223800941717e-07,
|
| 255 |
+
"logits/chosen": -4.695282936096191,
|
| 256 |
+
"logits/rejected": -4.676292419433594,
|
| 257 |
+
"logps/chosen": -1.3696571588516235,
|
| 258 |
+
"logps/rejected": -1.341069221496582,
|
| 259 |
+
"loss": 1.7283,
|
| 260 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 261 |
+
"rewards/chosen": -3.4241433143615723,
|
| 262 |
+
"rewards/margins": -0.0714699923992157,
|
| 263 |
+
"rewards/rejected": -3.352673053741455,
|
| 264 |
+
"step": 80
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.18167245524979964,
|
| 268 |
+
"grad_norm": 6.985848218948992,
|
| 269 |
+
"learning_rate": 4.899688103857222e-07,
|
| 270 |
+
"logits/chosen": -4.516369342803955,
|
| 271 |
+
"logits/rejected": -4.401045799255371,
|
| 272 |
+
"logps/chosen": -1.3444452285766602,
|
| 273 |
+
"logps/rejected": -1.356410264968872,
|
| 274 |
+
"loss": 1.7358,
|
| 275 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 276 |
+
"rewards/chosen": -3.3611133098602295,
|
| 277 |
+
"rewards/margins": 0.02991255186498165,
|
| 278 |
+
"rewards/rejected": -3.3910255432128906,
|
| 279 |
+
"step": 85
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.19235907026449373,
|
| 283 |
+
"grad_norm": 7.509298977705249,
|
| 284 |
+
"learning_rate": 4.871796225971999e-07,
|
| 285 |
+
"logits/chosen": -4.371309280395508,
|
| 286 |
+
"logits/rejected": -4.230607509613037,
|
| 287 |
+
"logps/chosen": -1.4209774732589722,
|
| 288 |
+
"logps/rejected": -1.3977888822555542,
|
| 289 |
+
"loss": 1.7377,
|
| 290 |
+
"rewards/accuracies": 0.4375,
|
| 291 |
+
"rewards/chosen": -3.552443265914917,
|
| 292 |
+
"rewards/margins": -0.05797123908996582,
|
| 293 |
+
"rewards/rejected": -3.494471788406372,
|
| 294 |
+
"step": 90
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.20304568527918782,
|
| 298 |
+
"grad_norm": 7.494464457003994,
|
| 299 |
+
"learning_rate": 4.840587176599343e-07,
|
| 300 |
+
"logits/chosen": -4.4767680168151855,
|
| 301 |
+
"logits/rejected": -4.559601306915283,
|
| 302 |
+
"logps/chosen": -1.3669825792312622,
|
| 303 |
+
"logps/rejected": -1.4395772218704224,
|
| 304 |
+
"loss": 1.7103,
|
| 305 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 306 |
+
"rewards/chosen": -3.4174563884735107,
|
| 307 |
+
"rewards/margins": 0.181486576795578,
|
| 308 |
+
"rewards/rejected": -3.598942995071411,
|
| 309 |
+
"step": 95
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.2137323002938819,
|
| 313 |
+
"grad_norm": 9.417753270846694,
|
| 314 |
+
"learning_rate": 4.806104604416823e-07,
|
| 315 |
+
"logits/chosen": -4.715841770172119,
|
| 316 |
+
"logits/rejected": -4.59860897064209,
|
| 317 |
+
"logps/chosen": -1.341230869293213,
|
| 318 |
+
"logps/rejected": -1.2855507135391235,
|
| 319 |
+
"loss": 1.7501,
|
| 320 |
+
"rewards/accuracies": 0.4375,
|
| 321 |
+
"rewards/chosen": -3.3530774116516113,
|
| 322 |
+
"rewards/margins": -0.13920024037361145,
|
| 323 |
+
"rewards/rejected": -3.2138772010803223,
|
| 324 |
+
"step": 100
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.224418915308576,
|
| 328 |
+
"grad_norm": 8.25939695647074,
|
| 329 |
+
"learning_rate": 4.768396736419662e-07,
|
| 330 |
+
"logits/chosen": -4.579805850982666,
|
| 331 |
+
"logits/rejected": -4.472721576690674,
|
| 332 |
+
"logps/chosen": -1.2403936386108398,
|
| 333 |
+
"logps/rejected": -1.2722570896148682,
|
| 334 |
+
"loss": 1.7103,
|
| 335 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 336 |
+
"rewards/chosen": -3.1009843349456787,
|
| 337 |
+
"rewards/margins": 0.07965835183858871,
|
| 338 |
+
"rewards/rejected": -3.180642604827881,
|
| 339 |
+
"step": 105
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.2351055303232701,
|
| 343 |
+
"grad_norm": 7.0286538282806905,
|
| 344 |
+
"learning_rate": 4.7275163104709194e-07,
|
| 345 |
+
"logits/chosen": -4.310456275939941,
|
| 346 |
+
"logits/rejected": -4.180710315704346,
|
| 347 |
+
"logps/chosen": -1.3125991821289062,
|
| 348 |
+
"logps/rejected": -1.3054355382919312,
|
| 349 |
+
"loss": 1.7814,
|
| 350 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 351 |
+
"rewards/chosen": -3.2814979553222656,
|
| 352 |
+
"rewards/margins": -0.017909258604049683,
|
| 353 |
+
"rewards/rejected": -3.2635891437530518,
|
| 354 |
+
"step": 110
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.2457921453379642,
|
| 358 |
+
"grad_norm": 7.4251876749652315,
|
| 359 |
+
"learning_rate": 4.683520501542824e-07,
|
| 360 |
+
"logits/chosen": -4.345327854156494,
|
| 361 |
+
"logits/rejected": -4.324080944061279,
|
| 362 |
+
"logps/chosen": -1.269272804260254,
|
| 363 |
+
"logps/rejected": -1.255380392074585,
|
| 364 |
+
"loss": 1.7077,
|
| 365 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 366 |
+
"rewards/chosen": -3.1731820106506348,
|
| 367 |
+
"rewards/margins": -0.03473084047436714,
|
| 368 |
+
"rewards/rejected": -3.138451099395752,
|
| 369 |
+
"step": 115
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.2564787603526583,
|
| 373 |
+
"grad_norm": 7.68817663348518,
|
| 374 |
+
"learning_rate": 4.636470841752404e-07,
|
| 375 |
+
"logits/chosen": -4.619086265563965,
|
| 376 |
+
"logits/rejected": -4.4949049949646,
|
| 377 |
+
"logps/chosen": -1.2920095920562744,
|
| 378 |
+
"logps/rejected": -1.2791041135787964,
|
| 379 |
+
"loss": 1.7086,
|
| 380 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 381 |
+
"rewards/chosen": -3.2300238609313965,
|
| 382 |
+
"rewards/margins": -0.03226359933614731,
|
| 383 |
+
"rewards/rejected": -3.1977603435516357,
|
| 384 |
+
"step": 120
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.2671653753673524,
|
| 388 |
+
"grad_norm": 7.125331982786664,
|
| 389 |
+
"learning_rate": 4.5864331343032565e-07,
|
| 390 |
+
"logits/chosen": -4.59976053237915,
|
| 391 |
+
"logits/rejected": -4.594804286956787,
|
| 392 |
+
"logps/chosen": -1.3433685302734375,
|
| 393 |
+
"logps/rejected": -1.4002816677093506,
|
| 394 |
+
"loss": 1.729,
|
| 395 |
+
"rewards/accuracies": 0.5625,
|
| 396 |
+
"rewards/chosen": -3.3584213256835938,
|
| 397 |
+
"rewards/margins": 0.14228278398513794,
|
| 398 |
+
"rewards/rejected": -3.500704288482666,
|
| 399 |
+
"step": 125
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.2778519903820465,
|
| 403 |
+
"grad_norm": 8.312615727941836,
|
| 404 |
+
"learning_rate": 4.533477361453819e-07,
|
| 405 |
+
"logits/chosen": -4.355613708496094,
|
| 406 |
+
"logits/rejected": -4.472823619842529,
|
| 407 |
+
"logps/chosen": -1.3405156135559082,
|
| 408 |
+
"logps/rejected": -1.3726789951324463,
|
| 409 |
+
"loss": 1.7552,
|
| 410 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 411 |
+
"rewards/chosen": -3.3512885570526123,
|
| 412 |
+
"rewards/margins": 0.08040888607501984,
|
| 413 |
+
"rewards/rejected": -3.431697368621826,
|
| 414 |
+
"step": 130
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.2885386053967406,
|
| 418 |
+
"grad_norm": 8.624812004095546,
|
| 419 |
+
"learning_rate": 4.4776775866408533e-07,
|
| 420 |
+
"logits/chosen": -4.605002403259277,
|
| 421 |
+
"logits/rejected": -4.486802577972412,
|
| 422 |
+
"logps/chosen": -1.2775276899337769,
|
| 423 |
+
"logps/rejected": -1.2938398122787476,
|
| 424 |
+
"loss": 1.7193,
|
| 425 |
+
"rewards/accuracies": 0.5,
|
| 426 |
+
"rewards/chosen": -3.193819284439087,
|
| 427 |
+
"rewards/margins": 0.040780358016490936,
|
| 428 |
+
"rewards/rejected": -3.2345993518829346,
|
| 429 |
+
"step": 135
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.2992252204114347,
|
| 433 |
+
"grad_norm": 8.928849598835027,
|
| 434 |
+
"learning_rate": 4.4191118508950277e-07,
|
| 435 |
+
"logits/chosen": -4.492905616760254,
|
| 436 |
+
"logits/rejected": -4.466243743896484,
|
| 437 |
+
"logps/chosen": -1.4048876762390137,
|
| 438 |
+
"logps/rejected": -1.4387624263763428,
|
| 439 |
+
"loss": 1.7204,
|
| 440 |
+
"rewards/accuracies": 0.612500011920929,
|
| 441 |
+
"rewards/chosen": -3.512219190597534,
|
| 442 |
+
"rewards/margins": 0.08468694984912872,
|
| 443 |
+
"rewards/rejected": -3.5969061851501465,
|
| 444 |
+
"step": 140
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.30991183542612877,
|
| 448 |
+
"grad_norm": 8.071911039943062,
|
| 449 |
+
"learning_rate": 4.357862063693485e-07,
|
| 450 |
+
"logits/chosen": -4.527454376220703,
|
| 451 |
+
"logits/rejected": -4.5367279052734375,
|
| 452 |
+
"logps/chosen": -1.344947099685669,
|
| 453 |
+
"logps/rejected": -1.371901273727417,
|
| 454 |
+
"loss": 1.7418,
|
| 455 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 456 |
+
"rewards/chosen": -3.362368106842041,
|
| 457 |
+
"rewards/margins": 0.06738501042127609,
|
| 458 |
+
"rewards/rejected": -3.429752826690674,
|
| 459 |
+
"step": 145
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.32059845044082286,
|
| 463 |
+
"grad_norm": 8.671271998589708,
|
| 464 |
+
"learning_rate": 4.294013888402029e-07,
|
| 465 |
+
"logits/chosen": -4.329895496368408,
|
| 466 |
+
"logits/rejected": -4.347752571105957,
|
| 467 |
+
"logps/chosen": -1.2269880771636963,
|
| 468 |
+
"logps/rejected": -1.3332128524780273,
|
| 469 |
+
"loss": 1.7002,
|
| 470 |
+
"rewards/accuracies": 0.612500011920929,
|
| 471 |
+
"rewards/chosen": -3.0674703121185303,
|
| 472 |
+
"rewards/margins": 0.2655617594718933,
|
| 473 |
+
"rewards/rejected": -3.3330321311950684,
|
| 474 |
+
"step": 150
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.33128506545551695,
|
| 478 |
+
"grad_norm": 11.410050720477157,
|
| 479 |
+
"learning_rate": 4.227656622467162e-07,
|
| 480 |
+
"logits/chosen": -4.60593843460083,
|
| 481 |
+
"logits/rejected": -4.5206708908081055,
|
| 482 |
+
"logps/chosen": -1.2302855253219604,
|
| 483 |
+
"logps/rejected": -1.285362958908081,
|
| 484 |
+
"loss": 1.7276,
|
| 485 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 486 |
+
"rewards/chosen": -3.075713872909546,
|
| 487 |
+
"rewards/margins": 0.13769367337226868,
|
| 488 |
+
"rewards/rejected": -3.2134079933166504,
|
| 489 |
+
"step": 155
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.34197168047021104,
|
| 493 |
+
"grad_norm": 8.929425526594324,
|
| 494 |
+
"learning_rate": 4.158883072525528e-07,
|
| 495 |
+
"logits/chosen": -4.636877536773682,
|
| 496 |
+
"logits/rejected": -4.64215612411499,
|
| 497 |
+
"logps/chosen": -1.2857083082199097,
|
| 498 |
+
"logps/rejected": -1.2909152507781982,
|
| 499 |
+
"loss": 1.7261,
|
| 500 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 501 |
+
"rewards/chosen": -3.214270830154419,
|
| 502 |
+
"rewards/margins": 0.01301727257668972,
|
| 503 |
+
"rewards/rejected": -3.227288007736206,
|
| 504 |
+
"step": 160
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.3526582954849052,
|
| 508 |
+
"grad_norm": 8.999759770579269,
|
| 509 |
+
"learning_rate": 4.087789424605447e-07,
|
| 510 |
+
"logits/chosen": -4.477012634277344,
|
| 511 |
+
"logits/rejected": -4.342657089233398,
|
| 512 |
+
"logps/chosen": -1.2676000595092773,
|
| 513 |
+
"logps/rejected": -1.2652432918548584,
|
| 514 |
+
"loss": 1.7307,
|
| 515 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 516 |
+
"rewards/chosen": -3.168999671936035,
|
| 517 |
+
"rewards/margins": -0.005891731474548578,
|
| 518 |
+
"rewards/rejected": -3.1631083488464355,
|
| 519 |
+
"step": 165
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.36334491049959927,
|
| 523 |
+
"grad_norm": 7.38370964011058,
|
| 524 |
+
"learning_rate": 4.0144751096020497e-07,
|
| 525 |
+
"logits/chosen": -4.322202682495117,
|
| 526 |
+
"logits/rejected": -4.344858646392822,
|
| 527 |
+
"logps/chosen": -1.3334827423095703,
|
| 528 |
+
"logps/rejected": -1.3564600944519043,
|
| 529 |
+
"loss": 1.7042,
|
| 530 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 531 |
+
"rewards/chosen": -3.3337066173553467,
|
| 532 |
+
"rewards/margins": 0.05744355916976929,
|
| 533 |
+
"rewards/rejected": -3.3911502361297607,
|
| 534 |
+
"step": 170
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.37403152551429336,
|
| 538 |
+
"grad_norm": 7.609720560260437,
|
| 539 |
+
"learning_rate": 3.939042664214184e-07,
|
| 540 |
+
"logits/chosen": -4.588566780090332,
|
| 541 |
+
"logits/rejected": -4.577520847320557,
|
| 542 |
+
"logps/chosen": -1.3666011095046997,
|
| 543 |
+
"logps/rejected": -1.4189504384994507,
|
| 544 |
+
"loss": 1.6971,
|
| 545 |
+
"rewards/accuracies": 0.574999988079071,
|
| 546 |
+
"rewards/chosen": -3.4165027141571045,
|
| 547 |
+
"rewards/margins": 0.13087329268455505,
|
| 548 |
+
"rewards/rejected": -3.5473761558532715,
|
| 549 |
+
"step": 175
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.38471814052898745,
|
| 553 |
+
"grad_norm": 7.522883946304073,
|
| 554 |
+
"learning_rate": 3.8615975875375676e-07,
|
| 555 |
+
"logits/chosen": -4.611153602600098,
|
| 556 |
+
"logits/rejected": -4.639096736907959,
|
| 557 |
+
"logps/chosen": -1.3451780080795288,
|
| 558 |
+
"logps/rejected": -1.407920241355896,
|
| 559 |
+
"loss": 1.7286,
|
| 560 |
+
"rewards/accuracies": 0.550000011920929,
|
| 561 |
+
"rewards/chosen": -3.3629448413848877,
|
| 562 |
+
"rewards/margins": 0.15685562789440155,
|
| 563 |
+
"rewards/rejected": -3.5198006629943848,
|
| 564 |
+
"step": 180
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.39540475554368154,
|
| 568 |
+
"grad_norm": 8.15961865987409,
|
| 569 |
+
"learning_rate": 3.7822481935147655e-07,
|
| 570 |
+
"logits/chosen": -4.689408302307129,
|
| 571 |
+
"logits/rejected": -4.5546979904174805,
|
| 572 |
+
"logps/chosen": -1.3114018440246582,
|
| 573 |
+
"logps/rejected": -1.333992838859558,
|
| 574 |
+
"loss": 1.7275,
|
| 575 |
+
"rewards/accuracies": 0.512499988079071,
|
| 576 |
+
"rewards/chosen": -3.2785046100616455,
|
| 577 |
+
"rewards/margins": 0.05647749826312065,
|
| 578 |
+
"rewards/rejected": -3.334981918334961,
|
| 579 |
+
"step": 185
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.40609137055837563,
|
| 583 |
+
"grad_norm": 8.427441389080544,
|
| 584 |
+
"learning_rate": 3.7011054594483443e-07,
|
| 585 |
+
"logits/chosen": -4.588101863861084,
|
| 586 |
+
"logits/rejected": -4.52250862121582,
|
| 587 |
+
"logps/chosen": -1.3444761037826538,
|
| 588 |
+
"logps/rejected": -1.4100086688995361,
|
| 589 |
+
"loss": 1.6814,
|
| 590 |
+
"rewards/accuracies": 0.612500011920929,
|
| 591 |
+
"rewards/chosen": -3.3611907958984375,
|
| 592 |
+
"rewards/margins": 0.16383083164691925,
|
| 593 |
+
"rewards/rejected": -3.5250213146209717,
|
| 594 |
+
"step": 190
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.4167779855730697,
|
| 598 |
+
"grad_norm": 8.693634103691119,
|
| 599 |
+
"learning_rate": 3.618282870789081e-07,
|
| 600 |
+
"logits/chosen": -4.41694450378418,
|
| 601 |
+
"logits/rejected": -4.407210826873779,
|
| 602 |
+
"logps/chosen": -1.3954448699951172,
|
| 603 |
+
"logps/rejected": -1.3883302211761475,
|
| 604 |
+
"loss": 1.7438,
|
| 605 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 606 |
+
"rewards/chosen": -3.488612413406372,
|
| 607 |
+
"rewards/margins": -0.017786988988518715,
|
| 608 |
+
"rewards/rejected": -3.4708251953125,
|
| 609 |
+
"step": 195
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.4274646005877638,
|
| 613 |
+
"grad_norm": 7.817917194646963,
|
| 614 |
+
"learning_rate": 3.5338962624163016e-07,
|
| 615 |
+
"logits/chosen": -4.41127872467041,
|
| 616 |
+
"logits/rejected": -4.464591026306152,
|
| 617 |
+
"logps/chosen": -1.3187341690063477,
|
| 618 |
+
"logps/rejected": -1.3379795551300049,
|
| 619 |
+
"loss": 1.7493,
|
| 620 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 621 |
+
"rewards/chosen": -3.296835422515869,
|
| 622 |
+
"rewards/margins": 0.04811403155326843,
|
| 623 |
+
"rewards/rejected": -3.344949245452881,
|
| 624 |
+
"step": 200
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.4381512156024579,
|
| 628 |
+
"grad_norm": 7.87816681095113,
|
| 629 |
+
"learning_rate": 3.448063656632321e-07,
|
| 630 |
+
"logits/chosen": -4.585541248321533,
|
| 631 |
+
"logits/rejected": -4.440451145172119,
|
| 632 |
+
"logps/chosen": -1.3579752445220947,
|
| 633 |
+
"logps/rejected": -1.3495149612426758,
|
| 634 |
+
"loss": 1.6908,
|
| 635 |
+
"rewards/accuracies": 0.375,
|
| 636 |
+
"rewards/chosen": -3.3949382305145264,
|
| 637 |
+
"rewards/margins": -0.02115105651319027,
|
| 638 |
+
"rewards/rejected": -3.3737874031066895,
|
| 639 |
+
"step": 205
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.448837830617152,
|
| 643 |
+
"grad_norm": 8.192581867844957,
|
| 644 |
+
"learning_rate": 3.360905098097587e-07,
|
| 645 |
+
"logits/chosen": -4.702603340148926,
|
| 646 |
+
"logits/rejected": -4.5605058670043945,
|
| 647 |
+
"logps/chosen": -1.3157254457473755,
|
| 648 |
+
"logps/rejected": -1.326529860496521,
|
| 649 |
+
"loss": 1.7765,
|
| 650 |
+
"rewards/accuracies": 0.4375,
|
| 651 |
+
"rewards/chosen": -3.289313554763794,
|
| 652 |
+
"rewards/margins": 0.02701149880886078,
|
| 653 |
+
"rewards/rejected": -3.3163249492645264,
|
| 654 |
+
"step": 210
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.45952444563184613,
|
| 658 |
+
"grad_norm": 9.916408251919977,
|
| 659 |
+
"learning_rate": 3.272542485937368e-07,
|
| 660 |
+
"logits/chosen": -4.547693252563477,
|
| 661 |
+
"logits/rejected": -4.488272666931152,
|
| 662 |
+
"logps/chosen": -1.2768534421920776,
|
| 663 |
+
"logps/rejected": -1.3246023654937744,
|
| 664 |
+
"loss": 1.6999,
|
| 665 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 666 |
+
"rewards/chosen": -3.192133665084839,
|
| 667 |
+
"rewards/margins": 0.11937196552753448,
|
| 668 |
+
"rewards/rejected": -3.3115055561065674,
|
| 669 |
+
"step": 215
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.4702110606465402,
|
| 673 |
+
"grad_norm": 11.54297063214211,
|
| 674 |
+
"learning_rate": 3.1830994032548e-07,
|
| 675 |
+
"logits/chosen": -4.549879550933838,
|
| 676 |
+
"logits/rejected": -4.424410820007324,
|
| 677 |
+
"logps/chosen": -1.354941725730896,
|
| 678 |
+
"logps/rejected": -1.3328911066055298,
|
| 679 |
+
"loss": 1.7459,
|
| 680 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 681 |
+
"rewards/chosen": -3.387354612350464,
|
| 682 |
+
"rewards/margins": -0.05512649938464165,
|
| 683 |
+
"rewards/rejected": -3.332228183746338,
|
| 684 |
+
"step": 220
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.4808976756612343,
|
| 688 |
+
"grad_norm": 7.745936948551203,
|
| 689 |
+
"learning_rate": 3.0927009442887437e-07,
|
| 690 |
+
"logits/chosen": -4.522200584411621,
|
| 691 |
+
"logits/rejected": -4.433478355407715,
|
| 692 |
+
"logps/chosen": -1.3177942037582397,
|
| 693 |
+
"logps/rejected": -1.312595248222351,
|
| 694 |
+
"loss": 1.7151,
|
| 695 |
+
"rewards/accuracies": 0.5,
|
| 696 |
+
"rewards/chosen": -3.294485569000244,
|
| 697 |
+
"rewards/margins": -0.012997297570109367,
|
| 698 |
+
"rewards/rejected": -3.2814879417419434,
|
| 699 |
+
"step": 225
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.4915842906759284,
|
| 703 |
+
"grad_norm": 8.982993035215612,
|
| 704 |
+
"learning_rate": 3.001473539458182e-07,
|
| 705 |
+
"logits/chosen": -4.790830135345459,
|
| 706 |
+
"logits/rejected": -4.502416610717773,
|
| 707 |
+
"logps/chosen": -1.2848570346832275,
|
| 708 |
+
"logps/rejected": -1.3198983669281006,
|
| 709 |
+
"loss": 1.7007,
|
| 710 |
+
"rewards/accuracies": 0.512499988079071,
|
| 711 |
+
"rewards/chosen": -3.2121429443359375,
|
| 712 |
+
"rewards/margins": 0.08760317414999008,
|
| 713 |
+
"rewards/rejected": -3.299746036529541,
|
| 714 |
+
"step": 230
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.5022709056906225,
|
| 718 |
+
"grad_norm": 7.9015489492124225,
|
| 719 |
+
"learning_rate": 2.909544778537844e-07,
|
| 720 |
+
"logits/chosen": -4.630471229553223,
|
| 721 |
+
"logits/rejected": -4.5115509033203125,
|
| 722 |
+
"logps/chosen": -1.3522698879241943,
|
| 723 |
+
"logps/rejected": -1.337274193763733,
|
| 724 |
+
"loss": 1.7185,
|
| 725 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 726 |
+
"rewards/chosen": -3.3806750774383545,
|
| 727 |
+
"rewards/margins": -0.03748970106244087,
|
| 728 |
+
"rewards/rejected": -3.3431854248046875,
|
| 729 |
+
"step": 235
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.5129575207053166,
|
| 733 |
+
"grad_norm": 8.651538383776115,
|
| 734 |
+
"learning_rate": 2.817043232212371e-07,
|
| 735 |
+
"logits/chosen": -4.58451509475708,
|
| 736 |
+
"logits/rejected": -4.510278701782227,
|
| 737 |
+
"logps/chosen": -1.3096239566802979,
|
| 738 |
+
"logps/rejected": -1.3383967876434326,
|
| 739 |
+
"loss": 1.6684,
|
| 740 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 741 |
+
"rewards/chosen": -3.274059295654297,
|
| 742 |
+
"rewards/margins": 0.07193277776241302,
|
| 743 |
+
"rewards/rejected": -3.34599232673645,
|
| 744 |
+
"step": 240
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.5236441357200107,
|
| 748 |
+
"grad_norm": 8.276564878362981,
|
| 749 |
+
"learning_rate": 2.7240982722585837e-07,
|
| 750 |
+
"logits/chosen": -4.700469970703125,
|
| 751 |
+
"logits/rejected": -4.604641914367676,
|
| 752 |
+
"logps/chosen": -1.296589970588684,
|
| 753 |
+
"logps/rejected": -1.295531988143921,
|
| 754 |
+
"loss": 1.7276,
|
| 755 |
+
"rewards/accuracies": 0.4375,
|
| 756 |
+
"rewards/chosen": -3.2414746284484863,
|
| 757 |
+
"rewards/margins": -0.0026448846329003572,
|
| 758 |
+
"rewards/rejected": -3.2388298511505127,
|
| 759 |
+
"step": 245
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.5343307507347048,
|
| 763 |
+
"grad_norm": 8.739959615068186,
|
| 764 |
+
"learning_rate": 2.63083989060736e-07,
|
| 765 |
+
"logits/chosen": -4.688320636749268,
|
| 766 |
+
"logits/rejected": -4.622137069702148,
|
| 767 |
+
"logps/chosen": -1.420803427696228,
|
| 768 |
+
"logps/rejected": -1.5039843320846558,
|
| 769 |
+
"loss": 1.706,
|
| 770 |
+
"rewards/accuracies": 0.512499988079071,
|
| 771 |
+
"rewards/chosen": -3.552008867263794,
|
| 772 |
+
"rewards/margins": 0.20795249938964844,
|
| 773 |
+
"rewards/rejected": -3.7599613666534424,
|
| 774 |
+
"step": 250
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.5450173657493989,
|
| 778 |
+
"grad_norm": 8.205463032414816,
|
| 779 |
+
"learning_rate": 2.537398517538159e-07,
|
| 780 |
+
"logits/chosen": -4.498805046081543,
|
| 781 |
+
"logits/rejected": -4.545159816741943,
|
| 782 |
+
"logps/chosen": -1.3838533163070679,
|
| 783 |
+
"logps/rejected": -1.427234411239624,
|
| 784 |
+
"loss": 1.7113,
|
| 785 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 786 |
+
"rewards/chosen": -3.4596335887908936,
|
| 787 |
+
"rewards/margins": 0.10845281928777695,
|
| 788 |
+
"rewards/rejected": -3.5680861473083496,
|
| 789 |
+
"step": 255
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.555703980764093,
|
| 793 |
+
"grad_norm": 9.403793469995996,
|
| 794 |
+
"learning_rate": 2.4439048392604877e-07,
|
| 795 |
+
"logits/chosen": -4.465289115905762,
|
| 796 |
+
"logits/rejected": -4.415510654449463,
|
| 797 |
+
"logps/chosen": -1.3352388143539429,
|
| 798 |
+
"logps/rejected": -1.4621715545654297,
|
| 799 |
+
"loss": 1.6556,
|
| 800 |
+
"rewards/accuracies": 0.550000011920929,
|
| 801 |
+
"rewards/chosen": -3.338097095489502,
|
| 802 |
+
"rewards/margins": 0.3173316717147827,
|
| 803 |
+
"rewards/rejected": -3.655428647994995,
|
| 804 |
+
"step": 260
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.566390595778787,
|
| 808 |
+
"grad_norm": 9.87099892276829,
|
| 809 |
+
"learning_rate": 2.3504896151374144e-07,
|
| 810 |
+
"logits/chosen": -4.503221035003662,
|
| 811 |
+
"logits/rejected": -4.532763481140137,
|
| 812 |
+
"logps/chosen": -1.3822702169418335,
|
| 813 |
+
"logps/rejected": -1.3944904804229736,
|
| 814 |
+
"loss": 1.7068,
|
| 815 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 816 |
+
"rewards/chosen": -3.4556756019592285,
|
| 817 |
+
"rewards/margins": 0.03055066429078579,
|
| 818 |
+
"rewards/rejected": -3.4862258434295654,
|
| 819 |
+
"step": 265
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.5770772107934812,
|
| 823 |
+
"grad_norm": 9.709835641918701,
|
| 824 |
+
"learning_rate": 2.2572834948067795e-07,
|
| 825 |
+
"logits/chosen": -4.5645036697387695,
|
| 826 |
+
"logits/rejected": -4.5933027267456055,
|
| 827 |
+
"logps/chosen": -1.2755674123764038,
|
| 828 |
+
"logps/rejected": -1.3178540468215942,
|
| 829 |
+
"loss": 1.6978,
|
| 830 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 831 |
+
"rewards/chosen": -3.1889185905456543,
|
| 832 |
+
"rewards/margins": 0.10571658611297607,
|
| 833 |
+
"rewards/rejected": -3.294635057449341,
|
| 834 |
+
"step": 270
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.5877638258081752,
|
| 838 |
+
"grad_norm": 10.023847803695947,
|
| 839 |
+
"learning_rate": 2.164416835455862e-07,
|
| 840 |
+
"logits/chosen": -4.247835159301758,
|
| 841 |
+
"logits/rejected": -4.194937229156494,
|
| 842 |
+
"logps/chosen": -1.3332767486572266,
|
| 843 |
+
"logps/rejected": -1.3482983112335205,
|
| 844 |
+
"loss": 1.6989,
|
| 845 |
+
"rewards/accuracies": 0.4375,
|
| 846 |
+
"rewards/chosen": -3.3331921100616455,
|
| 847 |
+
"rewards/margins": 0.03755370154976845,
|
| 848 |
+
"rewards/rejected": -3.3707454204559326,
|
| 849 |
+
"step": 275
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.5984504408228694,
|
| 853 |
+
"grad_norm": 12.08817520180493,
|
| 854 |
+
"learning_rate": 2.072019519505062e-07,
|
| 855 |
+
"logits/chosen": -4.784656524658203,
|
| 856 |
+
"logits/rejected": -4.656561851501465,
|
| 857 |
+
"logps/chosen": -1.3093583583831787,
|
| 858 |
+
"logps/rejected": -1.3450324535369873,
|
| 859 |
+
"loss": 1.6805,
|
| 860 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 861 |
+
"rewards/chosen": -3.2733960151672363,
|
| 862 |
+
"rewards/margins": 0.08918508142232895,
|
| 863 |
+
"rewards/rejected": -3.362581253051758,
|
| 864 |
+
"step": 280
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.6091370558375635,
|
| 868 |
+
"grad_norm": 8.794437977214896,
|
| 869 |
+
"learning_rate": 1.980220772955602e-07,
|
| 870 |
+
"logits/chosen": -4.4688849449157715,
|
| 871 |
+
"logits/rejected": -4.370944499969482,
|
| 872 |
+
"logps/chosen": -1.335540533065796,
|
| 873 |
+
"logps/rejected": -1.3469078540802002,
|
| 874 |
+
"loss": 1.6982,
|
| 875 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 876 |
+
"rewards/chosen": -3.3388514518737793,
|
| 877 |
+
"rewards/margins": 0.028418350964784622,
|
| 878 |
+
"rewards/rejected": -3.367269992828369,
|
| 879 |
+
"step": 285
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.6198236708522575,
|
| 883 |
+
"grad_norm": 13.055246394354226,
|
| 884 |
+
"learning_rate": 1.8891489846552644e-07,
|
| 885 |
+
"logits/chosen": -4.446020603179932,
|
| 886 |
+
"logits/rejected": -4.3907389640808105,
|
| 887 |
+
"logps/chosen": -1.332921028137207,
|
| 888 |
+
"logps/rejected": -1.3605272769927979,
|
| 889 |
+
"loss": 1.646,
|
| 890 |
+
"rewards/accuracies": 0.512499988079071,
|
| 891 |
+
"rewards/chosen": -3.3323025703430176,
|
| 892 |
+
"rewards/margins": 0.06901548057794571,
|
| 893 |
+
"rewards/rejected": -3.401318311691284,
|
| 894 |
+
"step": 290
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.6305102858669517,
|
| 898 |
+
"grad_norm": 9.19761267983714,
|
| 899 |
+
"learning_rate": 1.7989315267349933e-07,
|
| 900 |
+
"logits/chosen": -4.640501022338867,
|
| 901 |
+
"logits/rejected": -4.582611083984375,
|
| 902 |
+
"logps/chosen": -1.345268964767456,
|
| 903 |
+
"logps/rejected": -1.3664015531539917,
|
| 904 |
+
"loss": 1.6828,
|
| 905 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 906 |
+
"rewards/chosen": -3.363173007965088,
|
| 907 |
+
"rewards/margins": 0.05283108353614807,
|
| 908 |
+
"rewards/rejected": -3.416003704071045,
|
| 909 |
+
"step": 295
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.6411969008816457,
|
| 913 |
+
"grad_norm": 8.163371854491718,
|
| 914 |
+
"learning_rate": 1.7096945764674398e-07,
|
| 915 |
+
"logits/chosen": -4.53090763092041,
|
| 916 |
+
"logits/rejected": -4.449822425842285,
|
| 917 |
+
"logps/chosen": -1.357412338256836,
|
| 918 |
+
"logps/rejected": -1.340632677078247,
|
| 919 |
+
"loss": 1.7361,
|
| 920 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 921 |
+
"rewards/chosen": -3.393531084060669,
|
| 922 |
+
"rewards/margins": -0.04194915294647217,
|
| 923 |
+
"rewards/rejected": -3.3515820503234863,
|
| 924 |
+
"step": 300
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 0.6518835158963399,
|
| 928 |
+
"grad_norm": 9.037247969057203,
|
| 929 |
+
"learning_rate": 1.621562939796643e-07,
|
| 930 |
+
"logits/chosen": -4.495213508605957,
|
| 931 |
+
"logits/rejected": -4.3966827392578125,
|
| 932 |
+
"logps/chosen": -1.4379794597625732,
|
| 933 |
+
"logps/rejected": -1.473179578781128,
|
| 934 |
+
"loss": 1.7018,
|
| 935 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 936 |
+
"rewards/chosen": -3.5949490070343018,
|
| 937 |
+
"rewards/margins": 0.08800011873245239,
|
| 938 |
+
"rewards/rejected": -3.682948589324951,
|
| 939 |
+
"step": 305
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.6625701309110339,
|
| 943 |
+
"grad_norm": 13.030910514110566,
|
| 944 |
+
"learning_rate": 1.5346598767856345e-07,
|
| 945 |
+
"logits/chosen": -4.379870414733887,
|
| 946 |
+
"logits/rejected": -4.560307025909424,
|
| 947 |
+
"logps/chosen": -1.398297905921936,
|
| 948 |
+
"logps/rejected": -1.4838316440582275,
|
| 949 |
+
"loss": 1.6849,
|
| 950 |
+
"rewards/accuracies": 0.5625,
|
| 951 |
+
"rewards/chosen": -3.4957451820373535,
|
| 952 |
+
"rewards/margins": 0.21383419632911682,
|
| 953 |
+
"rewards/rejected": -3.7095787525177,
|
| 954 |
+
"step": 310
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.673256745925728,
|
| 958 |
+
"grad_norm": 9.601996227444385,
|
| 959 |
+
"learning_rate": 1.4491069292260866e-07,
|
| 960 |
+
"logits/chosen": -4.930965423583984,
|
| 961 |
+
"logits/rejected": -4.876108646392822,
|
| 962 |
+
"logps/chosen": -1.430011510848999,
|
| 963 |
+
"logps/rejected": -1.4526395797729492,
|
| 964 |
+
"loss": 1.7084,
|
| 965 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 966 |
+
"rewards/chosen": -3.575028896331787,
|
| 967 |
+
"rewards/margins": 0.05657036229968071,
|
| 968 |
+
"rewards/rejected": -3.631598949432373,
|
| 969 |
+
"step": 315
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.6839433609404221,
|
| 973 |
+
"grad_norm": 8.370911393375984,
|
| 974 |
+
"learning_rate": 1.365023750651133e-07,
|
| 975 |
+
"logits/chosen": -4.7936224937438965,
|
| 976 |
+
"logits/rejected": -4.740353584289551,
|
| 977 |
+
"logps/chosen": -1.282147765159607,
|
| 978 |
+
"logps/rejected": -1.4656821489334106,
|
| 979 |
+
"loss": 1.6538,
|
| 980 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 981 |
+
"rewards/chosen": -3.205369472503662,
|
| 982 |
+
"rewards/margins": 0.4588359296321869,
|
| 983 |
+
"rewards/rejected": -3.6642050743103027,
|
| 984 |
+
"step": 320
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.6946299759551162,
|
| 988 |
+
"grad_norm": 10.059762786625233,
|
| 989 |
+
"learning_rate": 1.2825279389890818e-07,
|
| 990 |
+
"logits/chosen": -4.382018089294434,
|
| 991 |
+
"logits/rejected": -4.410761833190918,
|
| 992 |
+
"logps/chosen": -1.3325350284576416,
|
| 993 |
+
"logps/rejected": -1.4160031080245972,
|
| 994 |
+
"loss": 1.6493,
|
| 995 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 996 |
+
"rewards/chosen": -3.3313376903533936,
|
| 997 |
+
"rewards/margins": 0.20867028832435608,
|
| 998 |
+
"rewards/rejected": -3.5400078296661377,
|
| 999 |
+
"step": 325
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.7053165909698104,
|
| 1003 |
+
"grad_norm": 9.161029493778742,
|
| 1004 |
+
"learning_rate": 1.201734872092077e-07,
|
| 1005 |
+
"logits/chosen": -4.73400354385376,
|
| 1006 |
+
"logits/rejected": -4.58245849609375,
|
| 1007 |
+
"logps/chosen": -1.3593249320983887,
|
| 1008 |
+
"logps/rejected": -1.5017703771591187,
|
| 1009 |
+
"loss": 1.6762,
|
| 1010 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1011 |
+
"rewards/chosen": -3.3983123302459717,
|
| 1012 |
+
"rewards/margins": 0.3561137616634369,
|
| 1013 |
+
"rewards/rejected": -3.7544262409210205,
|
| 1014 |
+
"step": 330
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 0.7160032059845044,
|
| 1018 |
+
"grad_norm": 8.300761837819268,
|
| 1019 |
+
"learning_rate": 1.1227575463697439e-07,
|
| 1020 |
+
"logits/chosen": -4.566507816314697,
|
| 1021 |
+
"logits/rejected": -4.648960590362549,
|
| 1022 |
+
"logps/chosen": -1.3698114156723022,
|
| 1023 |
+
"logps/rejected": -1.4295297861099243,
|
| 1024 |
+
"loss": 1.683,
|
| 1025 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 1026 |
+
"rewards/chosen": -3.4245285987854004,
|
| 1027 |
+
"rewards/margins": 0.14929580688476562,
|
| 1028 |
+
"rewards/rejected": -3.573824405670166,
|
| 1029 |
+
"step": 335
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.7266898209991985,
|
| 1033 |
+
"grad_norm": 9.524750095557723,
|
| 1034 |
+
"learning_rate": 1.0457064187534861e-07,
|
| 1035 |
+
"logits/chosen": -4.421561241149902,
|
| 1036 |
+
"logits/rejected": -4.395345687866211,
|
| 1037 |
+
"logps/chosen": -1.2686647176742554,
|
| 1038 |
+
"logps/rejected": -1.3926507234573364,
|
| 1039 |
+
"loss": 1.7023,
|
| 1040 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1041 |
+
"rewards/chosen": -3.171661615371704,
|
| 1042 |
+
"rewards/margins": 0.3099651038646698,
|
| 1043 |
+
"rewards/rejected": -3.4816269874572754,
|
| 1044 |
+
"step": 340
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 0.7373764360138926,
|
| 1048 |
+
"grad_norm": 9.264218645867375,
|
| 1049 |
+
"learning_rate": 9.706892522124838e-08,
|
| 1050 |
+
"logits/chosen": -4.445697784423828,
|
| 1051 |
+
"logits/rejected": -4.289374351501465,
|
| 1052 |
+
"logps/chosen": -1.3950523138046265,
|
| 1053 |
+
"logps/rejected": -1.3890727758407593,
|
| 1054 |
+
"loss": 1.6517,
|
| 1055 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 1056 |
+
"rewards/chosen": -3.487631320953369,
|
| 1057 |
+
"rewards/margins": -0.014949033036828041,
|
| 1058 |
+
"rewards/rejected": -3.472681760787964,
|
| 1059 |
+
"step": 345
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7480630510285867,
|
| 1063 |
+
"grad_norm": 9.805135028816036,
|
| 1064 |
+
"learning_rate": 8.978109650374396e-08,
|
| 1065 |
+
"logits/chosen": -4.54650354385376,
|
| 1066 |
+
"logits/rejected": -4.534266471862793,
|
| 1067 |
+
"logps/chosen": -1.4010884761810303,
|
| 1068 |
+
"logps/rejected": -1.4561593532562256,
|
| 1069 |
+
"loss": 1.7054,
|
| 1070 |
+
"rewards/accuracies": 0.5,
|
| 1071 |
+
"rewards/chosen": -3.5027213096618652,
|
| 1072 |
+
"rewards/margins": 0.13767699897289276,
|
| 1073 |
+
"rewards/rejected": -3.6403985023498535,
|
| 1074 |
+
"step": 350
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.7587496660432808,
|
| 1078 |
+
"grad_norm": 13.674204223054398,
|
| 1079 |
+
"learning_rate": 8.271734841028552e-08,
|
| 1080 |
+
"logits/chosen": -4.590233325958252,
|
| 1081 |
+
"logits/rejected": -4.405337333679199,
|
| 1082 |
+
"logps/chosen": -1.3775599002838135,
|
| 1083 |
+
"logps/rejected": -1.3799619674682617,
|
| 1084 |
+
"loss": 1.6717,
|
| 1085 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 1086 |
+
"rewards/chosen": -3.4438998699188232,
|
| 1087 |
+
"rewards/margins": 0.0060052452608942986,
|
| 1088 |
+
"rewards/rejected": -3.4499053955078125,
|
| 1089 |
+
"step": 355
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.7694362810579749,
|
| 1093 |
+
"grad_norm": 8.839622500870057,
|
| 1094 |
+
"learning_rate": 7.588756023130833e-08,
|
| 1095 |
+
"logits/chosen": -4.494322776794434,
|
| 1096 |
+
"logits/rejected": -4.6520586013793945,
|
| 1097 |
+
"logps/chosen": -1.43705153465271,
|
| 1098 |
+
"logps/rejected": -1.5561714172363281,
|
| 1099 |
+
"loss": 1.6713,
|
| 1100 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1101 |
+
"rewards/chosen": -3.5926284790039062,
|
| 1102 |
+
"rewards/margins": 0.2977997660636902,
|
| 1103 |
+
"rewards/rejected": -3.890428066253662,
|
| 1104 |
+
"step": 360
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.7801228960726689,
|
| 1108 |
+
"grad_norm": 7.5612671070202895,
|
| 1109 |
+
"learning_rate": 6.930128404315214e-08,
|
| 1110 |
+
"logits/chosen": -4.652656555175781,
|
| 1111 |
+
"logits/rejected": -4.5095977783203125,
|
| 1112 |
+
"logps/chosen": -1.4330899715423584,
|
| 1113 |
+
"logps/rejected": -1.393972635269165,
|
| 1114 |
+
"loss": 1.6884,
|
| 1115 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 1116 |
+
"rewards/chosen": -3.5827248096466064,
|
| 1117 |
+
"rewards/margins": -0.09779311716556549,
|
| 1118 |
+
"rewards/rejected": -3.484931468963623,
|
| 1119 |
+
"step": 365
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.7908095110873631,
|
| 1123 |
+
"grad_norm": 10.689889310789306,
|
| 1124 |
+
"learning_rate": 6.296773134861824e-08,
|
| 1125 |
+
"logits/chosen": -4.595254421234131,
|
| 1126 |
+
"logits/rejected": -4.460488796234131,
|
| 1127 |
+
"logps/chosen": -1.455096960067749,
|
| 1128 |
+
"logps/rejected": -1.4411834478378296,
|
| 1129 |
+
"loss": 1.7402,
|
| 1130 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1131 |
+
"rewards/chosen": -3.637742280960083,
|
| 1132 |
+
"rewards/margins": -0.034783538430929184,
|
| 1133 |
+
"rewards/rejected": -3.602959156036377,
|
| 1134 |
+
"step": 370
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.8014961261020572,
|
| 1138 |
+
"grad_norm": 11.234917641139386,
|
| 1139 |
+
"learning_rate": 5.6895760193850145e-08,
|
| 1140 |
+
"logits/chosen": -4.608285903930664,
|
| 1141 |
+
"logits/rejected": -4.652877330780029,
|
| 1142 |
+
"logps/chosen": -1.3920633792877197,
|
| 1143 |
+
"logps/rejected": -1.4802567958831787,
|
| 1144 |
+
"loss": 1.6655,
|
| 1145 |
+
"rewards/accuracies": 0.5625,
|
| 1146 |
+
"rewards/chosen": -3.480158567428589,
|
| 1147 |
+
"rewards/margins": 0.22048339247703552,
|
| 1148 |
+
"rewards/rejected": -3.7006423473358154,
|
| 1149 |
+
"step": 375
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.8121827411167513,
|
| 1153 |
+
"grad_norm": 9.88912726576546,
|
| 1154 |
+
"learning_rate": 5.109386277955477e-08,
|
| 1155 |
+
"logits/chosen": -4.704692840576172,
|
| 1156 |
+
"logits/rejected": -4.580458641052246,
|
| 1157 |
+
"logps/chosen": -1.4294003248214722,
|
| 1158 |
+
"logps/rejected": -1.4911302328109741,
|
| 1159 |
+
"loss": 1.6673,
|
| 1160 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1161 |
+
"rewards/chosen": -3.573500871658325,
|
| 1162 |
+
"rewards/margins": 0.15432511270046234,
|
| 1163 |
+
"rewards/rejected": -3.72782564163208,
|
| 1164 |
+
"step": 380
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.8228693561314454,
|
| 1168 |
+
"grad_norm": 10.411305158642252,
|
| 1169 |
+
"learning_rate": 4.557015358389216e-08,
|
| 1170 |
+
"logits/chosen": -4.629119396209717,
|
| 1171 |
+
"logits/rejected": -4.591732978820801,
|
| 1172 |
+
"logps/chosen": -1.371460199356079,
|
| 1173 |
+
"logps/rejected": -1.4805608987808228,
|
| 1174 |
+
"loss": 1.6999,
|
| 1175 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 1176 |
+
"rewards/chosen": -3.4286506175994873,
|
| 1177 |
+
"rewards/margins": 0.27275168895721436,
|
| 1178 |
+
"rewards/rejected": -3.701402187347412,
|
| 1179 |
+
"step": 385
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.8335559711461394,
|
| 1183 |
+
"grad_norm": 10.400854396409033,
|
| 1184 |
+
"learning_rate": 4.0332358013644015e-08,
|
| 1185 |
+
"logits/chosen": -4.841821670532227,
|
| 1186 |
+
"logits/rejected": -4.738966941833496,
|
| 1187 |
+
"logps/chosen": -1.4363723993301392,
|
| 1188 |
+
"logps/rejected": -1.4949367046356201,
|
| 1189 |
+
"loss": 1.6777,
|
| 1190 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1191 |
+
"rewards/chosen": -3.590930938720703,
|
| 1192 |
+
"rewards/margins": 0.14641080796718597,
|
| 1193 |
+
"rewards/rejected": -3.7373416423797607,
|
| 1194 |
+
"step": 390
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.8442425861608336,
|
| 1198 |
+
"grad_norm": 7.772358738977712,
|
| 1199 |
+
"learning_rate": 3.538780159953347e-08,
|
| 1200 |
+
"logits/chosen": -4.330094337463379,
|
| 1201 |
+
"logits/rejected": -4.12509298324585,
|
| 1202 |
+
"logps/chosen": -1.3891313076019287,
|
| 1203 |
+
"logps/rejected": -1.3778597116470337,
|
| 1204 |
+
"loss": 1.7104,
|
| 1205 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 1206 |
+
"rewards/chosen": -3.4728283882141113,
|
| 1207 |
+
"rewards/margins": -0.02817906066775322,
|
| 1208 |
+
"rewards/rejected": -3.4446492195129395,
|
| 1209 |
+
"step": 395
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.8549292011755276,
|
| 1213 |
+
"grad_norm": 8.138330588799487,
|
| 1214 |
+
"learning_rate": 3.074339975080836e-08,
|
| 1215 |
+
"logits/chosen": -4.4906816482543945,
|
| 1216 |
+
"logits/rejected": -4.360891819000244,
|
| 1217 |
+
"logps/chosen": -1.4243663549423218,
|
| 1218 |
+
"logps/rejected": -1.5000258684158325,
|
| 1219 |
+
"loss": 1.6693,
|
| 1220 |
+
"rewards/accuracies": 0.5,
|
| 1221 |
+
"rewards/chosen": -3.5609164237976074,
|
| 1222 |
+
"rewards/margins": 0.18914875388145447,
|
| 1223 |
+
"rewards/rejected": -3.7500648498535156,
|
| 1224 |
+
"step": 400
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.8549292011755276,
|
| 1228 |
+
"eval_logits/chosen": -5.357321739196777,
|
| 1229 |
+
"eval_logits/rejected": -5.259011745452881,
|
| 1230 |
+
"eval_logps/chosen": -1.3368273973464966,
|
| 1231 |
+
"eval_logps/rejected": -1.3893804550170898,
|
| 1232 |
+
"eval_loss": 1.6597751379013062,
|
| 1233 |
+
"eval_rewards/accuracies": 0.5282257795333862,
|
| 1234 |
+
"eval_rewards/chosen": -3.342068672180176,
|
| 1235 |
+
"eval_rewards/margins": 0.13138233125209808,
|
| 1236 |
+
"eval_rewards/rejected": -3.4734511375427246,
|
| 1237 |
+
"eval_runtime": 31.809,
|
| 1238 |
+
"eval_samples_per_second": 61.649,
|
| 1239 |
+
"eval_steps_per_second": 1.949,
|
| 1240 |
+
"step": 400
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 0.8656158161902218,
|
| 1244 |
+
"grad_norm": 9.025721024350105,
|
| 1245 |
+
"learning_rate": 2.6405648083415833e-08,
|
| 1246 |
+
"logits/chosen": -4.550287246704102,
|
| 1247 |
+
"logits/rejected": -4.305168628692627,
|
| 1248 |
+
"logps/chosen": -1.3779646158218384,
|
| 1249 |
+
"logps/rejected": -1.3646600246429443,
|
| 1250 |
+
"loss": 1.7228,
|
| 1251 |
+
"rewards/accuracies": 0.4375,
|
| 1252 |
+
"rewards/chosen": -3.444911479949951,
|
| 1253 |
+
"rewards/margins": -0.033261336386203766,
|
| 1254 |
+
"rewards/rejected": -3.4116501808166504,
|
| 1255 |
+
"step": 405
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.8763024312049158,
|
| 1259 |
+
"grad_norm": 11.89923381726874,
|
| 1260 |
+
"learning_rate": 2.2380613335296033e-08,
|
| 1261 |
+
"logits/chosen": -4.46389102935791,
|
| 1262 |
+
"logits/rejected": -4.445181846618652,
|
| 1263 |
+
"logps/chosen": -1.3117091655731201,
|
| 1264 |
+
"logps/rejected": -1.3666512966156006,
|
| 1265 |
+
"loss": 1.6743,
|
| 1266 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1267 |
+
"rewards/chosen": -3.2792727947235107,
|
| 1268 |
+
"rewards/margins": 0.13735508918762207,
|
| 1269 |
+
"rewards/rejected": -3.416628360748291,
|
| 1270 |
+
"step": 410
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.88698904621961,
|
| 1274 |
+
"grad_norm": 9.427141264972793,
|
| 1275 |
+
"learning_rate": 1.8673924881500823e-08,
|
| 1276 |
+
"logits/chosen": -4.327395439147949,
|
| 1277 |
+
"logits/rejected": -4.3094940185546875,
|
| 1278 |
+
"logps/chosen": -1.3074382543563843,
|
| 1279 |
+
"logps/rejected": -1.3440717458724976,
|
| 1280 |
+
"loss": 1.6969,
|
| 1281 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 1282 |
+
"rewards/chosen": -3.2685959339141846,
|
| 1283 |
+
"rewards/margins": 0.09158400446176529,
|
| 1284 |
+
"rewards/rejected": -3.3601794242858887,
|
| 1285 |
+
"step": 415
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.897675661234304,
|
| 1289 |
+
"grad_norm": 8.90218297589139,
|
| 1290 |
+
"learning_rate": 1.5290766861003475e-08,
|
| 1291 |
+
"logits/chosen": -4.604608058929443,
|
| 1292 |
+
"logits/rejected": -4.5904130935668945,
|
| 1293 |
+
"logps/chosen": -1.3577730655670166,
|
| 1294 |
+
"logps/rejected": -1.4380416870117188,
|
| 1295 |
+
"loss": 1.6707,
|
| 1296 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1297 |
+
"rewards/chosen": -3.394432544708252,
|
| 1298 |
+
"rewards/margins": 0.20067182183265686,
|
| 1299 |
+
"rewards/rejected": -3.595104694366455,
|
| 1300 |
+
"step": 420
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 0.9083622762489981,
|
| 1304 |
+
"grad_norm": 10.88127447829711,
|
| 1305 |
+
"learning_rate": 1.2235870926211616e-08,
|
| 1306 |
+
"logits/chosen": -4.505583763122559,
|
| 1307 |
+
"logits/rejected": -4.377171516418457,
|
| 1308 |
+
"logps/chosen": -1.379010558128357,
|
| 1309 |
+
"logps/rejected": -1.4630334377288818,
|
| 1310 |
+
"loss": 1.6586,
|
| 1311 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1312 |
+
"rewards/chosen": -3.447526454925537,
|
| 1313 |
+
"rewards/margins": 0.21005713939666748,
|
| 1314 |
+
"rewards/rejected": -3.6575839519500732,
|
| 1315 |
+
"step": 425
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.9190488912636923,
|
| 1319 |
+
"grad_norm": 10.139040628163542,
|
| 1320 |
+
"learning_rate": 9.513509625323518e-09,
|
| 1321 |
+
"logits/chosen": -4.474516868591309,
|
| 1322 |
+
"logits/rejected": -4.397044658660889,
|
| 1323 |
+
"logps/chosen": -1.3109729290008545,
|
| 1324 |
+
"logps/rejected": -1.3905436992645264,
|
| 1325 |
+
"loss": 1.6431,
|
| 1326 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1327 |
+
"rewards/chosen": -3.2774322032928467,
|
| 1328 |
+
"rewards/margins": 0.19892734289169312,
|
| 1329 |
+
"rewards/rejected": -3.4763596057891846,
|
| 1330 |
+
"step": 430
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 0.9297355062783863,
|
| 1334 |
+
"grad_norm": 9.603243560454548,
|
| 1335 |
+
"learning_rate": 7.127490426783123e-09,
|
| 1336 |
+
"logits/chosen": -4.4349188804626465,
|
| 1337 |
+
"logits/rejected": -4.401902675628662,
|
| 1338 |
+
"logps/chosen": -1.382652997970581,
|
| 1339 |
+
"logps/rejected": -1.4352567195892334,
|
| 1340 |
+
"loss": 1.6374,
|
| 1341 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1342 |
+
"rewards/chosen": -3.456632137298584,
|
| 1343 |
+
"rewards/margins": 0.13150940835475922,
|
| 1344 |
+
"rewards/rejected": -3.588141918182373,
|
| 1345 |
+
"step": 435
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 0.9404221212930804,
|
| 1349 |
+
"grad_norm": 9.022313878003665,
|
| 1350 |
+
"learning_rate": 5.08115039419113e-09,
|
| 1351 |
+
"logits/chosen": -4.515431880950928,
|
| 1352 |
+
"logits/rejected": -4.496222496032715,
|
| 1353 |
+
"logps/chosen": -1.3486416339874268,
|
| 1354 |
+
"logps/rejected": -1.328399896621704,
|
| 1355 |
+
"loss": 1.6747,
|
| 1356 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1357 |
+
"rewards/chosen": -3.3716042041778564,
|
| 1358 |
+
"rewards/margins": -0.05060465261340141,
|
| 1359 |
+
"rewards/rejected": -3.3209996223449707,
|
| 1360 |
+
"step": 440
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.9511087363077745,
|
| 1364 |
+
"grad_norm": 10.793631217244918,
|
| 1365 |
+
"learning_rate": 3.3773515191196646e-09,
|
| 1366 |
+
"logits/chosen": -4.495880126953125,
|
| 1367 |
+
"logits/rejected": -4.391939640045166,
|
| 1368 |
+
"logps/chosen": -1.355960726737976,
|
| 1369 |
+
"logps/rejected": -1.3439215421676636,
|
| 1370 |
+
"loss": 1.6919,
|
| 1371 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1372 |
+
"rewards/chosen": -3.389902114868164,
|
| 1373 |
+
"rewards/margins": -0.03009791299700737,
|
| 1374 |
+
"rewards/rejected": -3.3598036766052246,
|
| 1375 |
+
"step": 445
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.9617953513224686,
|
| 1379 |
+
"grad_norm": 11.320376659575638,
|
| 1380 |
+
"learning_rate": 2.0184767183584474e-09,
|
| 1381 |
+
"logits/chosen": -4.227232456207275,
|
| 1382 |
+
"logits/rejected": -4.283637523651123,
|
| 1383 |
+
"logps/chosen": -1.4779750108718872,
|
| 1384 |
+
"logps/rejected": -1.5757644176483154,
|
| 1385 |
+
"loss": 1.6992,
|
| 1386 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1387 |
+
"rewards/chosen": -3.694937229156494,
|
| 1388 |
+
"rewards/margins": 0.24447353184223175,
|
| 1389 |
+
"rewards/rejected": -3.939411163330078,
|
| 1390 |
+
"step": 450
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 0.9724819663371627,
|
| 1394 |
+
"grad_norm": 9.391067862633497,
|
| 1395 |
+
"learning_rate": 1.0064265011902328e-09,
|
| 1396 |
+
"logits/chosen": -4.655524253845215,
|
| 1397 |
+
"logits/rejected": -4.6156415939331055,
|
| 1398 |
+
"logps/chosen": -1.4518479108810425,
|
| 1399 |
+
"logps/rejected": -1.3639124631881714,
|
| 1400 |
+
"loss": 1.7101,
|
| 1401 |
+
"rewards/accuracies": 0.3125,
|
| 1402 |
+
"rewards/chosen": -3.62962007522583,
|
| 1403 |
+
"rewards/margins": -0.21983873844146729,
|
| 1404 |
+
"rewards/rejected": -3.4097812175750732,
|
| 1405 |
+
"step": 455
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 0.9831685813518568,
|
| 1409 |
+
"grad_norm": 9.291539758676725,
|
| 1410 |
+
"learning_rate": 3.4261631135654167e-10,
|
| 1411 |
+
"logits/chosen": -4.43479585647583,
|
| 1412 |
+
"logits/rejected": -4.606427192687988,
|
| 1413 |
+
"logps/chosen": -1.3911080360412598,
|
| 1414 |
+
"logps/rejected": -1.5079371929168701,
|
| 1415 |
+
"loss": 1.6583,
|
| 1416 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1417 |
+
"rewards/chosen": -3.4777698516845703,
|
| 1418 |
+
"rewards/margins": 0.29207324981689453,
|
| 1419 |
+
"rewards/rejected": -3.769843339920044,
|
| 1420 |
+
"step": 460
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.9938551963665508,
|
| 1424 |
+
"grad_norm": 9.789963041329289,
|
| 1425 |
+
"learning_rate": 2.797454743164174e-11,
|
| 1426 |
+
"logits/chosen": -4.726589679718018,
|
| 1427 |
+
"logits/rejected": -4.504673957824707,
|
| 1428 |
+
"logps/chosen": -1.3526123762130737,
|
| 1429 |
+
"logps/rejected": -1.3742250204086304,
|
| 1430 |
+
"loss": 1.6691,
|
| 1431 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1432 |
+
"rewards/chosen": -3.38153076171875,
|
| 1433 |
+
"rewards/margins": 0.05403166264295578,
|
| 1434 |
+
"rewards/rejected": -3.4355628490448,
|
| 1435 |
+
"step": 465
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 0.9981298423724285,
|
| 1439 |
+
"step": 467,
|
| 1440 |
+
"total_flos": 0.0,
|
| 1441 |
+
"train_loss": 1.7099600828689476,
|
| 1442 |
+
"train_runtime": 3104.7014,
|
| 1443 |
+
"train_samples_per_second": 19.286,
|
| 1444 |
+
"train_steps_per_second": 0.15
|
| 1445 |
+
}
|
| 1446 |
+
],
|
| 1447 |
+
"logging_steps": 5,
|
| 1448 |
+
"max_steps": 467,
|
| 1449 |
+
"num_input_tokens_seen": 0,
|
| 1450 |
+
"num_train_epochs": 1,
|
| 1451 |
+
"save_steps": 1000000,
|
| 1452 |
+
"stateful_callbacks": {
|
| 1453 |
+
"TrainerControl": {
|
| 1454 |
+
"args": {
|
| 1455 |
+
"should_epoch_stop": false,
|
| 1456 |
+
"should_evaluate": false,
|
| 1457 |
+
"should_log": false,
|
| 1458 |
+
"should_save": false,
|
| 1459 |
+
"should_training_stop": false
|
| 1460 |
+
},
|
| 1461 |
+
"attributes": {}
|
| 1462 |
+
}
|
| 1463 |
+
},
|
| 1464 |
+
"total_flos": 0.0,
|
| 1465 |
+
"train_batch_size": 2,
|
| 1466 |
+
"trial_name": null,
|
| 1467 |
+
"trial_params": null
|
| 1468 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e96bdcdc0e35c2559ff01f27b96f1c338d4e3ca5a278f6d7ab1b8ff608207cb4
|
| 3 |
+
size 6584
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|