LikelySurf
commited on
Commit
•
dd309a4
1
Parent(s):
bfe170d
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
base_model: gpt2
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
model-index:
|
7 |
+
- name: MammoLLM_bz32_acc4_lr1e4_large_epoch20
|
8 |
+
results: []
|
9 |
+
---
|
10 |
+
|
11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
12 |
+
should probably proofread and complete it, then remove this comment. -->
|
13 |
+
|
14 |
+
# MammoLLM_bz32_acc4_lr1e4_large_epoch20
|
15 |
+
|
16 |
+
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the None dataset.
|
17 |
+
It achieves the following results on the evaluation set:
|
18 |
+
- Loss: 1.0667
|
19 |
+
|
20 |
+
## Model description
|
21 |
+
|
22 |
+
More information needed
|
23 |
+
|
24 |
+
## Intended uses & limitations
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Training and evaluation data
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training procedure
|
33 |
+
|
34 |
+
### Training hyperparameters
|
35 |
+
|
36 |
+
The following hyperparameters were used during training:
|
37 |
+
- learning_rate: 0.0001
|
38 |
+
- train_batch_size: 64
|
39 |
+
- eval_batch_size: 64
|
40 |
+
- seed: 42
|
41 |
+
- gradient_accumulation_steps: 4
|
42 |
+
- total_train_batch_size: 256
|
43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
44 |
+
- lr_scheduler_type: cosine
|
45 |
+
- lr_scheduler_warmup_steps: 1000
|
46 |
+
- num_epochs: 20
|
47 |
+
|
48 |
+
### Training results
|
49 |
+
|
50 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
51 |
+
|:-------------:|:-----:|:-----:|:---------------:|
|
52 |
+
| 3.5007 | 0.48 | 500 | 2.1723 |
|
53 |
+
| 1.9621 | 0.96 | 1000 | 1.7328 |
|
54 |
+
| 1.6708 | 1.44 | 1500 | 1.5700 |
|
55 |
+
| 1.5458 | 1.92 | 2000 | 1.4780 |
|
56 |
+
| 1.453 | 2.4 | 2500 | 1.4207 |
|
57 |
+
| 1.4098 | 2.87 | 3000 | 1.3808 |
|
58 |
+
| 1.3492 | 3.35 | 3500 | 1.3499 |
|
59 |
+
| 1.3205 | 3.83 | 4000 | 1.3220 |
|
60 |
+
| 1.2694 | 4.31 | 4500 | 1.2945 |
|
61 |
+
| 1.2459 | 4.79 | 5000 | 1.2724 |
|
62 |
+
| 1.2006 | 5.27 | 5500 | 1.2482 |
|
63 |
+
| 1.1749 | 5.75 | 6000 | 1.2302 |
|
64 |
+
| 1.139 | 6.23 | 6500 | 1.2146 |
|
65 |
+
| 1.105 | 6.71 | 7000 | 1.1944 |
|
66 |
+
| 1.0754 | 7.19 | 7500 | 1.1791 |
|
67 |
+
| 1.0407 | 7.66 | 8000 | 1.1618 |
|
68 |
+
| 1.0188 | 8.14 | 8500 | 1.1497 |
|
69 |
+
| 0.9787 | 8.62 | 9000 | 1.1374 |
|
70 |
+
| 0.9652 | 9.1 | 9500 | 1.1300 |
|
71 |
+
| 0.9177 | 9.58 | 10000 | 1.1139 |
|
72 |
+
| 0.9165 | 10.06 | 10500 | 1.1088 |
|
73 |
+
| 0.8636 | 10.54 | 11000 | 1.0979 |
|
74 |
+
| 0.8693 | 11.02 | 11500 | 1.0909 |
|
75 |
+
| 0.812 | 11.5 | 12000 | 1.0895 |
|
76 |
+
| 0.8243 | 11.98 | 12500 | 1.0779 |
|
77 |
+
| 0.7734 | 12.46 | 13000 | 1.0796 |
|
78 |
+
| 0.7792 | 12.93 | 13500 | 1.0717 |
|
79 |
+
| 0.74 | 13.41 | 14000 | 1.0763 |
|
80 |
+
| 0.7403 | 13.89 | 14500 | 1.0681 |
|
81 |
+
| 0.7088 | 14.37 | 15000 | 1.0699 |
|
82 |
+
| 0.708 | 14.85 | 15500 | 1.0650 |
|
83 |
+
| 0.6846 | 15.33 | 16000 | 1.0684 |
|
84 |
+
| 0.6811 | 15.81 | 16500 | 1.0652 |
|
85 |
+
| 0.6644 | 16.29 | 17000 | 1.0688 |
|
86 |
+
| 0.6582 | 16.77 | 17500 | 1.0665 |
|
87 |
+
| 0.6512 | 17.25 | 18000 | 1.0669 |
|
88 |
+
| 0.6433 | 17.72 | 18500 | 1.0663 |
|
89 |
+
| 0.6403 | 18.2 | 19000 | 1.0668 |
|
90 |
+
| 0.6347 | 18.68 | 19500 | 1.0666 |
|
91 |
+
| 0.6318 | 19.16 | 20000 | 1.0668 |
|
92 |
+
| 0.6304 | 19.64 | 20500 | 1.0667 |
|
93 |
+
|
94 |
+
|
95 |
+
### Framework versions
|
96 |
+
|
97 |
+
- Transformers 4.31.0
|
98 |
+
- Pytorch 2.0.1+cu117
|
99 |
+
- Datasets 2.14.3
|
100 |
+
- Tokenizers 0.13.3
|