Update README.md
Browse files
README.md
CHANGED
|
@@ -6,33 +6,25 @@ tags:
|
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
| 8 |
model-index:
|
| 9 |
-
- name:
|
| 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 |
-
#
|
| 17 |
|
| 18 |
-
|
| 19 |
-
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 3.9890
|
| 21 |
-
- Accuracy: 0.4213
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
## Intended uses & limitations
|
| 28 |
-
|
| 29 |
-
More information needed
|
| 30 |
|
| 31 |
## Training and evaluation data
|
| 32 |
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
## Training procedure
|
| 36 |
|
| 37 |
### Training hyperparameters
|
| 38 |
|
|
@@ -46,58 +38,10 @@ The following hyperparameters were used during training:
|
|
| 46 |
- num_epochs: 50.0
|
| 47 |
|
| 48 |
### Training results
|
|
|
|
| 49 |
|
| 50 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 51 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 52 |
-
| No log | 1.0 | 148 | 5.8378 | 0.2287 |
|
| 53 |
-
| No log | 2.0 | 296 | 5.5732 | 0.2491 |
|
| 54 |
-
| No log | 3.0 | 444 | 5.3800 | 0.2680 |
|
| 55 |
-
| 5.761 | 4.0 | 592 | 5.2259 | 0.2842 |
|
| 56 |
-
| 5.761 | 5.0 | 740 | 5.1006 | 0.2963 |
|
| 57 |
-
| 5.761 | 6.0 | 888 | 4.9997 | 0.3067 |
|
| 58 |
-
| 5.1725 | 7.0 | 1036 | 4.9103 | 0.3153 |
|
| 59 |
-
| 5.1725 | 8.0 | 1184 | 4.8287 | 0.3230 |
|
| 60 |
-
| 5.1725 | 9.0 | 1332 | 4.7578 | 0.3301 |
|
| 61 |
-
| 5.1725 | 10.0 | 1480 | 4.6942 | 0.3376 |
|
| 62 |
-
| 4.8482 | 11.0 | 1628 | 4.6364 | 0.3439 |
|
| 63 |
-
| 4.8482 | 12.0 | 1776 | 4.5813 | 0.3497 |
|
| 64 |
-
| 4.8482 | 13.0 | 1924 | 4.5328 | 0.3554 |
|
| 65 |
-
| 4.609 | 14.0 | 2072 | 4.4897 | 0.3610 |
|
| 66 |
-
| 4.609 | 15.0 | 2220 | 4.4454 | 0.3657 |
|
| 67 |
-
| 4.609 | 16.0 | 2368 | 4.4132 | 0.3702 |
|
| 68 |
-
| 4.4241 | 17.0 | 2516 | 4.3742 | 0.3738 |
|
| 69 |
-
| 4.4241 | 18.0 | 2664 | 4.3438 | 0.3782 |
|
| 70 |
-
| 4.4241 | 19.0 | 2812 | 4.3164 | 0.3817 |
|
| 71 |
-
| 4.4241 | 20.0 | 2960 | 4.2879 | 0.3848 |
|
| 72 |
-
| 4.283 | 21.0 | 3108 | 4.2602 | 0.3878 |
|
| 73 |
-
| 4.283 | 22.0 | 3256 | 4.2373 | 0.3902 |
|
| 74 |
-
| 4.283 | 23.0 | 3404 | 4.2160 | 0.3932 |
|
| 75 |
-
| 4.1606 | 24.0 | 3552 | 4.1964 | 0.3954 |
|
| 76 |
-
| 4.1606 | 25.0 | 3700 | 4.1816 | 0.3976 |
|
| 77 |
-
| 4.1606 | 26.0 | 3848 | 4.1605 | 0.3995 |
|
| 78 |
-
| 4.1606 | 27.0 | 3996 | 4.1443 | 0.4016 |
|
| 79 |
-
| 4.0685 | 28.0 | 4144 | 4.1290 | 0.4034 |
|
| 80 |
-
| 4.0685 | 29.0 | 4292 | 4.1146 | 0.4052 |
|
| 81 |
-
| 4.0685 | 30.0 | 4440 | 4.1037 | 0.4067 |
|
| 82 |
-
| 3.9859 | 31.0 | 4588 | 4.0874 | 0.4082 |
|
| 83 |
-
| 3.9859 | 32.0 | 4736 | 4.0784 | 0.4098 |
|
| 84 |
-
| 3.9859 | 33.0 | 4884 | 4.0669 | 0.4115 |
|
| 85 |
-
| 3.9275 | 34.0 | 5032 | 4.0581 | 0.4125 |
|
| 86 |
-
| 3.9275 | 35.0 | 5180 | 4.0479 | 0.4136 |
|
| 87 |
-
| 3.9275 | 36.0 | 5328 | 4.0384 | 0.4148 |
|
| 88 |
-
| 3.9275 | 37.0 | 5476 | 4.0330 | 0.4159 |
|
| 89 |
-
| 3.8799 | 38.0 | 5624 | 4.0262 | 0.4166 |
|
| 90 |
-
| 3.8799 | 39.0 | 5772 | 4.0212 | 0.4174 |
|
| 91 |
-
| 3.8799 | 40.0 | 5920 | 4.0136 | 0.4180 |
|
| 92 |
-
| 3.8348 | 41.0 | 6068 | 4.0111 | 0.4186 |
|
| 93 |
-
| 3.8348 | 42.0 | 6216 | 4.0048 | 0.4193 |
|
| 94 |
-
| 3.8348 | 43.0 | 6364 | 4.0004 | 0.4195 |
|
| 95 |
-
| 3.8151 | 44.0 | 6512 | 3.9978 | 0.4199 |
|
| 96 |
-
| 3.8151 | 45.0 | 6660 | 3.9951 | 0.4203 |
|
| 97 |
-
| 3.8151 | 46.0 | 6808 | 3.9934 | 0.4207 |
|
| 98 |
-
| 3.8151 | 47.0 | 6956 | 3.9904 | 0.4209 |
|
| 99 |
-
| 3.7948 | 48.0 | 7104 | 3.9911 | 0.4211 |
|
| 100 |
-
| 3.7948 | 49.0 | 7252 | 3.9890 | 0.4213 |
|
| 101 |
| 3.7948 | 50.0 | 7400 | 3.9890 | 0.4213 |
|
| 102 |
|
| 103 |
|
|
|
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
| 8 |
model-index:
|
| 9 |
+
- name: AmpGPT2
|
| 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 |
+
# AmpGPT2
|
| 17 |
|
| 18 |
+
AmpGPT2 is a language model capable of generating de novo antimicrobial peptides (AMPs). Generated sequences are predicted to be AMPs 95.83% of the time.
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
| 22 |
+
AmpGPT2 is a fine-tuned version of [nferruz/ProtGPT2](https://huggingface.co/nferruz/ProtGPT2) based on the GPT2 Transformer architecture.
|
| 23 |
+
To validate the results the Antimicrobial Peptide Scanner vr.2 (https://www.dveltri.com/ascan/v2/ascan.html) was used. It is a
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
## Training and evaluation data
|
| 26 |
|
| 27 |
+
run_clm.py
|
|
|
|
|
|
|
| 28 |
|
| 29 |
### Training hyperparameters
|
| 30 |
|
|
|
|
| 38 |
- num_epochs: 50.0
|
| 39 |
|
| 40 |
### Training results
|
| 41 |
+
tr
|
| 42 |
|
| 43 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 44 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
| 3.7948 | 50.0 | 7400 | 3.9890 | 0.4213 |
|
| 46 |
|
| 47 |
|