Downloads · 30 days
8
16% of all-time downloads
gLM2-1protonly-8epoch-finetune
ishanjmukherjee/gLM2-1protonly-8epoch-finetune
gLM2-1protonly-8epoch-finetune is a fill-mask model from ishanjmukherjee. Use it when you need the model to fill a missing word. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
50
Public
Parameters
671M
284 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors2.7 GB · 100%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- Model type
- gLM2
- Access
- Public
- Created
- May 8, 2025
- Updated
- May 12, 2025
- SHA
- 3570b1a8
Try a prompt
Base models
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
gLM2-1protonly-8epoch-finetune
This model is a fine-tuned version of tattabio/gLM2_650M on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.9922
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: 0.0001
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 8
- total_eval_batch_size: 8
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.231 | 0.0529 | 500 | 1.2229 |
| 1.1912 | 0.1059 | 1000 | 1.1914 |
| 1.1681 | 0.1588 | 1500 | 1.1674 |
| 1.1636 | 0.2118 | 2000 | 1.1496 |
| 1.1281 | 0.2647 | 2500 | 1.1372 |
| 1.1213 | 0.3177 | 3000 | 1.1247 |
| 1.1152 | 0.3706 | 3500 | 1.1150 |
| 1.1094 | 0.4235 | 4000 | 1.1050 |
| 1.0952 | 0.4765 | 4500 | 1.0973 |
| 1.0907 | 0.5294 | 5000 | 1.0903 |
| 1.0912 | 0.5824 | 5500 | 1.0847 |
| 1.0886 | 0.6353 | 6000 | 1.0795 |
| 1.0764 | 0.6883 | 6500 | 1.0742 |
| 1.0715 | 0.7412 | 7000 | 1.0686 |
| 1.0771 | 0.7942 | 7500 | 1.0637 |
| 1.0806 | 0.8471 | 8000 | 1.0604 |
| 1.0493 | 0.9000 | 8500 | 1.0563 |
| 1.0569 | 0.9530 | 9000 | 1.0522 |
| 1.0416 | 1.0059 | 9500 | 1.0504 |
| 1.0382 | 1.0589 | 10000 | 1.0479 |
| 1.0444 | 1.1118 | 10500 | 1.0446 |
| 1.0642 | 1.1648 | 11000 | 1.0427 |
| 1.025 | 1.2177 | 11500 | 1.0384 |
| 1.0265 | 1.2706 | 12000 | 1.0366 |
| 1.0307 | 1.3236 | 12500 | 1.0338 |
| 1.0289 | 1.3765 | 13000 | 1.0309 |
| 1.0071 | 1.4295 | 13500 | 1.0291 |
| 1.032 | 1.4824 | 14000 | 1.0276 |
| 1.0286 | 1.5354 | 14500 | 1.0241 |
| 1.0266 | 1.5883 | 15000 | 1.0222 |
| 1.0072 | 1.6413 | 15500 | 1.0206 |
| 1.0198 | 1.6942 | 16000 | 1.0194 |
| 1.0171 | 1.7471 | 16500 | 1.0172 |
| 1.007 | 1.8001 | 17000 | 1.0160 |
| 1.0175 | 1.8530 | 17500 | 1.0143 |
| 1.0265 | 1.9060 | 18000 | 1.0125 |
| 0.9966 | 1.9589 | 18500 | 1.0108 |
| 0.9973 | 2.0119 | 19000 | 1.0097 |
| 1.0099 | 2.0648 | 19500 | 1.0086 |
| 0.9914 | 2.1177 | 20000 | 1.0074 |
| 1.0189 | 2.1707 | 20500 | 1.0051 |
| 1.0053 | 2.2236 | 21000 | 1.0040 |
| 0.9951 | 2.2766 | 21500 | 1.0032 |
| 0.99 | 2.3295 | 22000 | 1.0014 |
| 0.9849 | 2.3825 | 22500 | 1.0007 |
| 0.9964 | 2.4354 | 23000 | 1.0000 |
| 0.9951 | 2.4884 | 23500 | 0.9986 |
| 0.9822 | 2.5413 | 24000 | 0.9972 |
| 0.988 | 2.5942 | 24500 | 0.9967 |
| 0.993 | 2.6472 | 25000 | 0.9955 |
| 0.9974 | 2.7001 | 25500 | 0.9942 |
| 0.9983 | 2.7531 | 26000 | 0.9937 |
| 0.9824 | 2.8060 | 26500 | 0.9938 |
| 0.9758 | 2.8590 | 27000 | 0.9927 |
| 0.9742 | 2.9119 | 27500 | 0.9928 |
| 0.9682 | 2.9648 | 28000 | 0.9922 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.7.0+cu126
- Datasets 3.5.1
- Tokenizers 0.21.1
- Task
- Fill-Mask
- Library
- transformers
- Type
- gLM2
- Created
- May 8, 2025
- Updated
- May 12, 2025