Hugging Face's logo

Grigorij
/
zephyr_instruct_generation

PEFT
Safetensors
trl
sft
Generated from Trainer
Model card Files Community
Grigorij commited on
Commit
3d48a30
·
1 Parent(s): 7dab921

Model save

Browse files
Files changed (1) hide show
  1. README.md +27 -17
README.md CHANGED
@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.5753
22
 
23
  ## Model description
24
 
@@ -44,27 +44,37 @@ The following hyperparameters were used during training:
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: constant
46
  - lr_scheduler_warmup_steps: 0.03
47
- - training_steps: 150
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:-----:|:----:|:---------------:|
53
- | 2.8455 | 0.4 | 10 | 2.2310 |
54
- | 1.8394 | 0.8 | 20 | 1.3360 |
55
- | 1.0736 | 1.2 | 30 | 0.9132 |
56
- | 0.8515 | 1.6 | 40 | 0.7353 |
57
- | 0.7499 | 2.0 | 50 | 0.6281 |
58
- | 0.5876 | 2.4 | 60 | 0.5897 |
59
- | 0.6041 | 2.8 | 70 | 0.5737 |
60
- | 0.5057 | 3.2 | 80 | 0.5581 |
61
- | 0.512 | 3.6 | 90 | 0.5704 |
62
- | 0.5419 | 4.0 | 100 | 0.5537 |
63
- | 0.4451 | 4.4 | 110 | 0.5777 |
64
- | 0.4467 | 4.8 | 120 | 0.5541 |
65
- | 0.4351 | 5.2 | 130 | 0.5612 |
66
- | 0.3963 | 5.6 | 140 | 0.5898 |
67
- | 0.4264 | 6.0 | 150 | 0.5753 |
 
 
 
 
 
 
 
 
 
 
68
 
69
 
70
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.7409
22
 
23
  ## Model description
24
 
 
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: constant
46
  - lr_scheduler_warmup_steps: 0.03
47
+ - training_steps: 250
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:-----:|:----:|:---------------:|
53
+ | 2.9344 | 0.4 | 10 | 2.3554 |
54
+ | 1.822 | 0.8 | 20 | 1.3051 |
55
+ | 1.0639 | 1.2 | 30 | 0.9542 |
56
+ | 0.9127 | 1.6 | 40 | 0.8368 |
57
+ | 0.8577 | 2.0 | 50 | 0.7642 |
58
+ | 0.7091 | 2.4 | 60 | 0.7175 |
59
+ | 0.7112 | 2.8 | 70 | 0.6682 |
60
+ | 0.5931 | 3.2 | 80 | 0.6508 |
61
+ | 0.5809 | 3.6 | 90 | 0.6226 |
62
+ | 0.5901 | 4.0 | 100 | 0.5975 |
63
+ | 0.4863 | 4.4 | 110 | 0.6240 |
64
+ | 0.5024 | 4.8 | 120 | 0.5896 |
65
+ | 0.4932 | 5.2 | 130 | 0.6103 |
66
+ | 0.44 | 5.6 | 140 | 0.6158 |
67
+ | 0.4737 | 6.0 | 150 | 0.6168 |
68
+ | 0.4191 | 6.4 | 160 | 0.6859 |
69
+ | 0.4179 | 6.8 | 170 | 0.6237 |
70
+ | 0.3921 | 7.2 | 180 | 0.6603 |
71
+ | 0.3694 | 7.6 | 190 | 0.6766 |
72
+ | 0.3822 | 8.0 | 200 | 0.6559 |
73
+ | 0.3495 | 8.4 | 210 | 0.6989 |
74
+ | 0.341 | 8.8 | 220 | 0.6622 |
75
+ | 0.333 | 9.2 | 230 | 0.7345 |
76
+ | 0.3357 | 9.6 | 240 | 0.7215 |
77
+ | 0.3284 | 10.0 | 250 | 0.7409 |
78
 
79
 
80
  ### Framework versions