metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: ahmedabdelwahed/Mojiz-sft
model-index:
- name: test
results: []
test
This model is a fine-tuned version of ahmedabdelwahed/Mojiz-sft on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0004
- Rewards/chosen: 4.9321
- Rewards/rejected: -3.1907
- Rewards/accuracies: 1.0
- Rewards/margins: 8.1228
- Logps/rejected: -102.9468
- Logps/chosen: -276.9925
- Logits/rejected: -11.5352
- Logits/chosen: -12.8189
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 150
- training_steps: 4000
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.59 | 0.41 | 100 | 0.5689 | 0.2142 | -0.0529 | 1.0 | 0.2670 | -71.5682 | -324.1714 | -11.7019 | -12.9321 |
| 0.3343 | 0.82 | 200 | 0.2046 | 1.1503 | -0.4195 | 1.0 | 1.5698 | -75.2343 | -314.8096 | -11.5589 | -12.6981 |
| 0.1352 | 1.22 | 300 | 0.0544 | 2.1788 | -0.9611 | 1.0 | 3.1399 | -80.6501 | -304.5251 | -11.3938 | -12.4321 |
| 0.0394 | 1.63 | 400 | 0.0235 | 2.7263 | -1.2443 | 1.0 | 3.9706 | -83.4828 | -299.0500 | -11.3316 | -12.3395 |
| 0.0406 | 2.04 | 500 | 0.0129 | 3.0958 | -1.4814 | 1.0 | 4.5771 | -85.8533 | -295.3555 | -11.2949 | -12.2909 |
| 0.0288 | 2.45 | 600 | 0.0085 | 3.3507 | -1.6368 | 1.0 | 4.9875 | -87.4073 | -292.8057 | -11.2931 | -12.2987 |
| 0.0168 | 2.86 | 700 | 0.0058 | 3.5726 | -1.7858 | 1.0 | 5.3584 | -88.8972 | -290.5872 | -11.2839 | -12.2960 |
| 0.0124 | 3.27 | 800 | 0.0044 | 3.7498 | -1.8907 | 1.0 | 5.6405 | -89.9465 | -288.8147 | -11.2953 | -12.3275 |
| 0.0087 | 3.67 | 900 | 0.0035 | 3.8535 | -2.0148 | 1.0 | 5.8684 | -91.1878 | -287.7778 | -11.2906 | -12.3302 |
| 0.007 | 4.08 | 1000 | 0.0029 | 3.9744 | -2.0910 | 1.0 | 6.0654 | -91.9495 | -286.5689 | -11.3025 | -12.3602 |
| 0.0122 | 4.49 | 1100 | 0.0023 | 4.0733 | -2.1965 | 1.0 | 6.2698 | -93.0046 | -285.5799 | -11.3092 | -12.3815 |
| 0.0062 | 4.9 | 1200 | 0.0020 | 4.1431 | -2.2995 | 1.0 | 6.4426 | -94.0349 | -284.8822 | -11.3084 | -12.3870 |
| 0.0113 | 5.31 | 1300 | 0.0017 | 4.2353 | -2.3425 | 1.0 | 6.5778 | -94.4643 | -283.9596 | -11.3304 | -12.4328 |
| 0.0096 | 5.71 | 1400 | 0.0015 | 4.3008 | -2.4035 | 1.0 | 6.7044 | -95.0748 | -283.3046 | -11.3438 | -12.4598 |
| 0.0054 | 6.12 | 1500 | 0.0013 | 4.3515 | -2.4735 | 1.0 | 6.8250 | -95.7742 | -282.7977 | -11.3476 | -12.4711 |
| 0.0127 | 6.53 | 1600 | 0.0012 | 4.4101 | -2.5237 | 1.0 | 6.9338 | -96.2767 | -282.2122 | -11.3609 | -12.4973 |
| 0.0046 | 6.94 | 1700 | 0.0011 | 4.4473 | -2.5927 | 1.0 | 7.0400 | -96.9663 | -281.8404 | -11.3694 | -12.5148 |
| 0.0049 | 7.35 | 1800 | 0.0010 | 4.4882 | -2.6539 | 1.0 | 7.1421 | -97.5784 | -281.4307 | -11.3787 | -12.5361 |
| 0.0039 | 7.76 | 1900 | 0.0009 | 4.5274 | -2.7045 | 1.0 | 7.2319 | -98.0843 | -281.0386 | -11.3905 | -12.5571 |
| 0.0044 | 8.16 | 2000 | 0.0008 | 4.5889 | -2.7282 | 1.0 | 7.3171 | -98.3214 | -280.4241 | -11.4088 | -12.5904 |
| 0.0065 | 8.57 | 2100 | 0.0008 | 4.6306 | -2.7613 | 1.0 | 7.3919 | -98.6527 | -280.0075 | -11.4171 | -12.6086 |
| 0.0033 | 8.98 | 2200 | 0.0007 | 4.6508 | -2.8105 | 1.0 | 7.4613 | -99.1442 | -279.8048 | -11.4214 | -12.6158 |
| 0.0041 | 9.39 | 2300 | 0.0007 | 4.6876 | -2.8441 | 1.0 | 7.5316 | -99.4804 | -279.4375 | -11.4311 | -12.6356 |
| 0.0025 | 9.8 | 2400 | 0.0006 | 4.7079 | -2.8946 | 1.0 | 7.6026 | -99.9857 | -279.2337 | -11.4363 | -12.6466 |
| 0.0113 | 10.2 | 2500 | 0.0006 | 4.7531 | -2.9135 | 1.0 | 7.6665 | -100.1742 | -278.7823 | -11.4524 | -12.6747 |
| 0.004 | 10.61 | 2600 | 0.0006 | 4.7825 | -2.9382 | 1.0 | 7.7207 | -100.4218 | -278.4879 | -11.4644 | -12.6965 |
| 0.0035 | 11.02 | 2700 | 0.0005 | 4.7952 | -2.9793 | 1.0 | 7.7745 | -100.8324 | -278.3613 | -11.4722 | -12.7103 |
| 0.0018 | 11.43 | 2800 | 0.0005 | 4.8215 | -3.0007 | 1.0 | 7.8222 | -101.0460 | -278.0980 | -11.4836 | -12.7306 |
| 0.0036 | 11.84 | 2900 | 0.0005 | 4.8430 | -3.0254 | 1.0 | 7.8684 | -101.2935 | -277.8831 | -11.4928 | -12.7447 |
| 0.0015 | 12.24 | 3000 | 0.0005 | 4.8521 | -3.0583 | 1.0 | 7.9104 | -101.6220 | -277.7916 | -11.4961 | -12.7505 |
| 0.0052 | 12.65 | 3100 | 0.0005 | 4.8654 | -3.0809 | 1.0 | 7.9463 | -101.8486 | -277.6590 | -11.5056 | -12.7670 |
| 0.0022 | 13.06 | 3200 | 0.0005 | 4.8844 | -3.0964 | 1.0 | 7.9807 | -102.0031 | -277.4695 | -11.5118 | -12.7789 |
| 0.002 | 13.47 | 3300 | 0.0004 | 4.8951 | -3.1182 | 1.0 | 8.0133 | -102.2215 | -277.3618 | -11.5162 | -12.7874 |
| 0.0022 | 13.88 | 3400 | 0.0004 | 4.9070 | -3.1334 | 1.0 | 8.0404 | -102.3737 | -277.2430 | -11.5216 | -12.7971 |
| 0.0021 | 14.29 | 3500 | 0.0004 | 4.9110 | -3.1522 | 1.0 | 8.0632 | -102.5618 | -277.2032 | -11.5242 | -12.8018 |
| 0.0016 | 14.69 | 3600 | 0.0004 | 4.9225 | -3.1606 | 1.0 | 8.0831 | -102.6452 | -277.0878 | -11.5305 | -12.8112 |
| 0.0038 | 15.1 | 3700 | 0.0004 | 4.9251 | -3.1769 | 1.0 | 8.1020 | -102.8086 | -277.0620 | -11.5312 | -12.8121 |
| 0.0081 | 15.51 | 3800 | 0.0004 | 4.9290 | -3.1834 | 1.0 | 8.1124 | -102.8736 | -277.0229 | -11.5335 | -12.8159 |
| 0.0021 | 15.92 | 3900 | 0.0004 | 4.9308 | -3.1902 | 1.0 | 8.1210 | -102.9418 | -277.0052 | -11.5346 | -12.8179 |
| 0.0019 | 16.33 | 4000 | 0.0004 | 4.9321 | -3.1907 | 1.0 | 8.1228 | -102.9468 | -276.9925 | -11.5352 | -12.8189 |
Framework versions
- PEFT 0.7.1
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.15.0