Downloads · 30 days
14
44% of all-time downloads
damgomz/ft_32_2e6_base
ft_32_2e6_base is a fill-mask model from damgomz. Use it when you need the model to fill a missing word. It is set up for transformers.
Downloads · 30 days
14
44% of all-time downloads
All-time downloads
32
Public
Parameters
11.7M
1.4 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors46.7 MB · 95%
From the Hugging Face model README
| Metric | Value |
|---|---|
| Duration (in seconds) | 94566.8026907444 |
| Emissions (Co2eq in kg) | 0.0572237513240666 |
| CPU power (W) | 42.5 |
| GPU power (W) | [No GPU] |
| RAM power (W) | 3.75 |
| CPU energy (kWh) | 1.1164102884280973 |
| GPU energy (kWh) | [No GPU] |
| RAM energy (kWh) | 0.09850555888017 |
| Consumed energy (kWh) | 1.2149158473082664 |
| Country name | Switzerland |
| Cloud provider | nan |
| Cloud region | nan |
| CPU count | 2 |
| CPU model | Intel(R) Xeon(R) Platinum 8360Y CPU @ 2.40GHz |
| GPU count | nan |
| GPU model | nan |
| Metric | Value |
|---|---|
| CPU energy (kWh) | 0.18204109517968298 |
| Emissions (Co2eq in kg) | 0.03703866438720822 |
21 May 2024
| Config | Value |
|---|---|
| checkpoint | albert-base-v2 |
| model_name | ft_32_2e6_base |
| sequence_length | 400 |
| num_epoch | 6 |
| learning_rate | 2e-06 |
| batch_size | 32 |
| weight_decay | 0.0 |
| warm_up_prop | 0.0 |
| drop_out_prob | 0.1 |
| packing_length | 100 |
| train_test_split | 0.2 |
| num_steps | 32586 |
| Epoch | Train Loss | Test Loss | Accuracy | Recall |
|---|---|---|---|---|
| 0 | 0.566979 | 0.493775 | 0.769300 | 0.826475 |
| 1 | 0.433719 | 0.412018 | 0.817031 | 0.850808 |
| 2 | 0.357331 | 0.387126 | 0.825282 | 0.869627 |
| 3 | 0.293457 | 0.401614 | 0.822777 | 0.850921 |
| 4 | 0.237306 | 0.430223 | 0.810255 | 0.821977 |
| 5 | 0.185810 | 0.469887 | 0.799501 | 0.834173 |