Downloads · 30 days
5
19% of all-time downloads
damgomz/fp_bs16_lr5_x4
fp_bs16_lr5_x4 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
5
19% of all-time downloads
All-time downloads
27
Public
Parameters
32.5M
1.7 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors130 MB · 98%
From the Hugging Face model README
| Metric | Value |
|---|---|
| Duration (in seconds) | 199750.60248947144 |
| Emissions (Co2eq in kg) | 0.1306726388971817 |
| CPU power (W) | 42.5 |
| GPU power (W) | [No GPU] |
| RAM power (W) | 7.5 |
| CPU energy (kWh) | 2.3581629528201304 |
| GPU energy (kWh) | [No GPU] |
| RAM energy (kWh) | 0.4161441510031621 |
| Consumed energy (kWh) | 2.77430710382331 |
| Country name | Switzerland |
| Cloud provider | nan |
| Cloud region | nan |
| CPU count | 4 |
| CPU model | Intel(R) Xeon(R) Platinum 8360Y CPU @ 2.40GHz |
| GPU count | nan |
| GPU model | nan |
| Metric | Value |
|---|---|
| CPU energy (kWh) | 0.3845199097922325 |
| Emissions (Co2eq in kg) | 0.07823565264170965 |
23 May 2024
| Config | Value |
|---|---|
| checkpoint | albert-base-v2 |
| model_name | fp_bs16_lr5_x4 |
| sequence_length | 400 |
| num_epoch | 6 |
| learning_rate | 5e-05 |
| batch_size | 16 |
| weight_decay | 0.0 |
| warm_up_prop | 0.0 |
| drop_out_prob | 0.1 |
| packing_length | 100 |
| train_test_split | 0.2 |
| num_steps | 40911 |
| Epoch | Train Loss | Test Loss |
|---|---|---|
| 0.0 | 16.038300 | 11.435867 |
| 0.5 | 4.387820 | 3.824438 |
| 1.0 | 3.580076 | 3.468719 |
| 1.5 | 3.337705 | 3.273031 |
| 2.0 | 3.178614 | 3.155824 |
| 2.5 | 3.066128 | 3.047407 |
| 3.0 | 2.977899 | 2.974327 |
| 3.5 | 2.898752 | 2.909501 |
| 4.0 | 2.834338 | 2.851398 |
| 4.5 | 2.787738 | 2.810708 |
| 5.0 | 2.723077 | 2.755004 |
| 5.5 | 2.680311 | 2.718634 |
| 6.0 | 2.655619 | 2.699163 |