Downloads · 30 days
7
20% of all-time downloads
Phi-1_5-Summarization-QLoRa
pkbiswas/Phi-1_5-Summarization-QLoRa
Phi-1_5-Summarization-QLoRa is a summarization model from pkbiswas. Use it when you need a shorter version of a longer text. It is set up for peft. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
35
Public
Repo size
25.2 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors25.2 MB · 88%
At a glance
- Task
- Summarization
- Library
- peft
- License
- mit
- Access
- Public
- Created
- Mar 19, 2024
- Updated
- Nov 17, 2024
- SHA
- b5850fed
Try a prompt
Base models
Datasets
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. -->
Phi-1.5 Summarization (QLoRA)
This model is a fine-tuned version of microsoft/phi-1_5 on the scitldr dataset. It achieves the following results on the evaluation set:
- Loss: 2.5866
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: 0.0002
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.5683 | 0.25 | 500 | 2.6196 |
| 2.5308 | 0.5 | 1000 | 2.5992 |
| 2.558 | 0.75 | 1500 | 2.5886 |
| 2.4925 | 1.0 | 2000 | 2.5827 |
| 2.3252 | 1.26 | 2500 | 2.5948 |
| 2.3128 | 1.51 | 3000 | 2.5879 |
| 2.4622 | 1.76 | 3500 | 2.5866 |
Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Task
- Summarization
- Library
- peft
- License
- mit
- Created
- Mar 19, 2024
- Updated
- Nov 17, 2024