Downloads · 30 days
14
6% of all-time downloads
hindi_gpt2
sushant-iit/hindi_gpt2
hindi_gpt2 is a text generation model from sushant-iit. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
223
Public
Parameters
128M
14.9 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors512 MB · 100%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- mit
- Model type
- gpt2
- Access
- Public
- Created
- Apr 11, 2024
- Updated
- Apr 15, 2024
- SHA
- d7e83709
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. -->
hindi_gpt2
This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9772
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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 500
- num_epochs: 8
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.9843 | 0.48 | 500 | 3.1384 |
| 2.7769 | 0.95 | 1000 | 2.5036 |
| 2.4479 | 1.43 | 1500 | 2.3352 |
| 2.3239 | 1.9 | 2000 | 2.2456 |
| 2.2272 | 2.38 | 2500 | 2.1854 |
| 2.1782 | 2.86 | 3000 | 2.1405 |
| 2.1102 | 3.33 | 3500 | 2.1053 |
| 2.0774 | 3.81 | 4000 | 2.0731 |
| 2.0276 | 4.29 | 4500 | 2.0551 |
| 1.9978 | 4.76 | 5000 | 2.0267 |
| 1.9624 | 5.24 | 5500 | 2.0143 |
| 1.9351 | 5.71 | 6000 | 1.9955 |
| 1.914 | 6.19 | 6500 | 1.9894 |
| 1.8886 | 6.67 | 7000 | 1.9807 |
| 1.8806 | 7.14 | 7500 | 1.9780 |
| 1.8692 | 7.62 | 8000 | 1.9772 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.4.0.dev20240415
- Datasets 2.18.0
- Tokenizers 0.15.2
- Task
- Text Generation
- Library
- transformers
- Type
- gpt2
- License
- mit
- Created
- Apr 11, 2024
- Updated
- Apr 15, 2024