Hugging Face's logo

YYYYYYibo
/
zephyr-7b-lora-64-no-quant

PEFT
Safetensors
mistral
alignment-handbook
Generated from Trainer
trl
dpo
Model card Files Community
YYYYYYibo commited on
Commit
8fb5279
·
verified ·
1 Parent(s): b3c9658

Model save

Browse files
Files changed (4) hide show
  1. README.md +62 -0
  2. all_results.json +8 -0
  3. train_results.json +8 -0
  4. trainer_state.json +86 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ library_name: peft
4
+ tags:
5
+ - trl
6
+ - dpo
7
+ - generated_from_trainer
8
+ base_model: alignment-handbook/zephyr-7b-sft-full
9
+ model-index:
10
+ - name: zephyr-7b-lora-64-no-quant
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
+ # zephyr-7b-lora-64-no-quant
18
+
19
+ This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 5e-06
39
+ - train_batch_size: 2
40
+ - eval_batch_size: 2
41
+ - seed: 42
42
+ - distributed_type: multi-GPU
43
+ - num_devices: 4
44
+ - gradient_accumulation_steps: 32
45
+ - total_train_batch_size: 256
46
+ - total_eval_batch_size: 8
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 1
51
+
52
+ ### Training results
53
+
54
+
55
+
56
+ ### Framework versions
57
+
58
+ - PEFT 0.7.1
59
+ - Transformers 4.36.2
60
+ - Pytorch 2.1.2+cu121
61
+ - Datasets 2.14.6
62
+ - Tokenizers 0.15.2
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.6793185824002975,
4
+ "train_runtime": 3271.4615,
5
+ "train_samples": 10000,
6
+ "train_samples_per_second": 3.057,
7
+ "train_steps_per_second": 0.012
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.6793185824002975,
4
+ "train_runtime": 3271.4615,
5
+ "train_samples": 10000,
6
+ "train_samples_per_second": 3.057,
7
+ "train_steps_per_second": 0.012
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9984,
5
+ "eval_steps": 100,
6
+ "global_step": 39,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.03,
13
+ "learning_rate": 1.25e-06,
14
+ "logits/chosen": -2.6728761196136475,
15
+ "logits/rejected": -2.6764090061187744,
16
+ "logps/chosen": -261.5552978515625,
17
+ "logps/rejected": -274.96466064453125,
18
+ "loss": 0.6931,
19
+ "rewards/accuracies": 0.0,
20
+ "rewards/chosen": 0.0,
21
+ "rewards/margins": 0.0,
22
+ "rewards/rejected": 0.0,
23
+ "step": 1
24
+ },
25
+ {
26
+ "epoch": 0.26,
27
+ "learning_rate": 4.646121984004666e-06,
28
+ "logits/chosen": -2.7910807132720947,
29
+ "logits/rejected": -2.7170495986938477,
30
+ "logps/chosen": -285.13818359375,
31
+ "logps/rejected": -261.8572998046875,
32
+ "loss": 0.6908,
33
+ "rewards/accuracies": 0.5572916865348816,
34
+ "rewards/chosen": 0.002809329889714718,
35
+ "rewards/margins": 0.004932739306241274,
36
+ "rewards/rejected": -0.002123409416526556,
37
+ "step": 10
38
+ },
39
+ {
40
+ "epoch": 0.51,
41
+ "learning_rate": 2.835583164544139e-06,
42
+ "logits/chosen": -2.7387890815734863,
43
+ "logits/rejected": -2.6834359169006348,
44
+ "logps/chosen": -278.64666748046875,
45
+ "logps/rejected": -251.53060913085938,
46
+ "loss": 0.6812,
47
+ "rewards/accuracies": 0.7250000238418579,
48
+ "rewards/chosen": 0.020782727748155594,
49
+ "rewards/margins": 0.027223270386457443,
50
+ "rewards/rejected": -0.006440544966608286,
51
+ "step": 20
52
+ },
53
+ {
54
+ "epoch": 0.77,
55
+ "learning_rate": 7.723433775328385e-07,
56
+ "logits/chosen": -2.7434194087982178,
57
+ "logits/rejected": -2.673462152481079,
58
+ "logps/chosen": -294.9715881347656,
59
+ "logps/rejected": -267.701904296875,
60
+ "loss": 0.6738,
61
+ "rewards/accuracies": 0.7124999761581421,
62
+ "rewards/chosen": 0.03247123584151268,
63
+ "rewards/margins": 0.04708124324679375,
64
+ "rewards/rejected": -0.014610009267926216,
65
+ "step": 30
66
+ },
67
+ {
68
+ "epoch": 1.0,
69
+ "step": 39,
70
+ "total_flos": 0.0,
71
+ "train_loss": 0.6793185824002975,
72
+ "train_runtime": 3271.4615,
73
+ "train_samples_per_second": 3.057,
74
+ "train_steps_per_second": 0.012
75
+ }
76
+ ],
77
+ "logging_steps": 10,
78
+ "max_steps": 39,
79
+ "num_input_tokens_seen": 0,
80
+ "num_train_epochs": 1,
81
+ "save_steps": 100,
82
+ "total_flos": 0.0,
83
+ "train_batch_size": 2,
84
+ "trial_name": null,
85
+ "trial_params": null
86
+ }