Hugging Face's logo

Sorour
/
cls_alldata_phi3_v1

PEFT
TensorBoard
Safetensors
trl
sft
Generated from Trainer
Model card Files
xet
Metrics Community
Sorour commited on
Commit
fa323db
·
verified ·
1 Parent(s): 9c3b229

Model save

Browse files
Files changed (1) hide show
  1. README.md +101 -0
README.md ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: peft
4
+ tags:
5
+ - trl
6
+ - sft
7
+ - generated_from_trainer
8
+ base_model: microsoft/Phi-3-mini-4k-instruct
9
+ datasets:
10
+ - generator
11
+ model-index:
12
+ - name: cls_alldata_phi3_v1
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # cls_alldata_phi3_v1
20
+
21
+ This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.4956
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 2
44
+ - eval_batch_size: 8
45
+ - seed: 42
46
+ - gradient_accumulation_steps: 4
47
+ - total_train_batch_size: 8
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: constant
50
+ - lr_scheduler_warmup_ratio: 0.03
51
+ - num_epochs: 2
52
+ - mixed_precision_training: Native AMP
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss |
57
+ |:-------------:|:------:|:----:|:---------------:|
58
+ | 0.7566 | 0.0559 | 20 | 0.7643 |
59
+ | 0.6863 | 0.1117 | 40 | 0.7089 |
60
+ | 0.6538 | 0.1676 | 60 | 0.6706 |
61
+ | 0.6261 | 0.2235 | 80 | 0.6499 |
62
+ | 0.6402 | 0.2793 | 100 | 0.6321 |
63
+ | 0.594 | 0.3352 | 120 | 0.6226 |
64
+ | 0.5956 | 0.3911 | 140 | 0.6121 |
65
+ | 0.5743 | 0.4469 | 160 | 0.6016 |
66
+ | 0.5494 | 0.5028 | 180 | 0.5903 |
67
+ | 0.5861 | 0.5587 | 200 | 0.5887 |
68
+ | 0.5431 | 0.6145 | 220 | 0.5801 |
69
+ | 0.5404 | 0.6704 | 240 | 0.5746 |
70
+ | 0.5401 | 0.7263 | 260 | 0.5695 |
71
+ | 0.5363 | 0.7821 | 280 | 0.5644 |
72
+ | 0.5534 | 0.8380 | 300 | 0.5608 |
73
+ | 0.5936 | 0.8939 | 320 | 0.5552 |
74
+ | 0.5139 | 0.9497 | 340 | 0.5496 |
75
+ | 0.5096 | 1.0056 | 360 | 0.5468 |
76
+ | 0.4891 | 1.0615 | 380 | 0.5468 |
77
+ | 0.4524 | 1.1173 | 400 | 0.5433 |
78
+ | 0.4568 | 1.1732 | 420 | 0.5397 |
79
+ | 0.4462 | 1.2291 | 440 | 0.5374 |
80
+ | 0.4605 | 1.2849 | 460 | 0.5337 |
81
+ | 0.4469 | 1.3408 | 480 | 0.5328 |
82
+ | 0.458 | 1.3966 | 500 | 0.5313 |
83
+ | 0.4378 | 1.4525 | 520 | 0.5250 |
84
+ | 0.4654 | 1.5084 | 540 | 0.5232 |
85
+ | 0.4563 | 1.5642 | 560 | 0.5200 |
86
+ | 0.4664 | 1.6201 | 580 | 0.5155 |
87
+ | 0.4308 | 1.6760 | 600 | 0.5128 |
88
+ | 0.443 | 1.7318 | 620 | 0.5082 |
89
+ | 0.4508 | 1.7877 | 640 | 0.5070 |
90
+ | 0.4511 | 1.8436 | 660 | 0.4999 |
91
+ | 0.4467 | 1.8994 | 680 | 0.4996 |
92
+ | 0.4723 | 1.9553 | 700 | 0.4956 |
93
+
94
+
95
+ ### Framework versions
96
+
97
+ - PEFT 0.11.1
98
+ - Transformers 4.41.1
99
+ - Pytorch 2.3.0+cu121
100
+ - Datasets 2.19.1
101
+ - Tokenizers 0.19.1