bosch_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr5e-4_r16_2602201457

leobianco/bosch_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr5e-4_r16_2602201457

bosch_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr5e-4_r16_2602201457 is a machine learning model from leobianco. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as gemma.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

4% of all-time downloads

All-time downloads

116

Public

Repo size

57.9 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.json34.4 MB · 76%

At a glance

Library: peft
License: gemma
Access: Public
Created: Feb 20, 2026
Updated: Feb 20, 2026
SHA: 0f139a10

Base models

google/gemma-2-2b-it

Model card

From the Hugging Face model README

bosch_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr5e-4_r16_2602201457

This model is a fine-tuned version of google/gemma-2-2b-it on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 2.9420
Roc Auc: 0.9028
Best Threshold: 0.7937
Tpr At Best Threshold: 0.8592
Fpr At Best Threshold: 0.2045
Accuracy At Best Threshold: 0.8441
Avg Score True Positives: 0.9061
Avg Score True Negatives: 0.3819

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: 2
eval_batch_size: 2
seed: 130104
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 16
total_eval_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Roc Auc	Best Threshold	Tpr At Best Threshold	Fpr At Best Threshold	Accuracy At Best Threshold	Avg Score True Positives	Avg Score True Negatives
No log	0	0	5.5312	0.5	inf	0.0	0.0	0.2366	0.5	0.5
5.293	0.1316	5	5.6906	0.7298	0.9533	0.6056	0.2045	0.6505	0.9507	0.9398
4.1309	0.2632	10	4.1134	0.8603	0.6922	0.8169	0.2273	0.8065	0.7203	0.6691
2.1616	0.3947	15	6.0511	0.8752	0.9846	0.7817	0.1591	0.7957	0.9828	0.9406
3.959	0.5263	20	3.7485	0.9008	0.3789	0.8662	0.1591	0.8602	0.6621	0.3287
4.8967	0.6579	25	3.6489	0.8964	0.9679	0.8662	0.1591	0.8602	0.9574	0.6089
4.2324	0.7895	30	2.9326	0.8740	0.6280	0.8803	0.2273	0.8548	0.8366	0.3259
0.8539	0.9211	35	2.9398	0.8584	0.8948	0.9296	0.25	0.8871	0.9295	0.5168
9.9754	1.0526	40	4.2100	0.8412	0.9698	0.8169	0.2727	0.7957	0.9633	0.7276
3.0117	1.1842	45	2.8868	0.9008	0.6914	0.7817	0.1364	0.8011	0.7732	0.3292
3.2566	1.3158	50	3.0517	0.8976	0.8170	0.7606	0.0909	0.7957	0.8738	0.6509
3.207	1.4474	55	2.8641	0.9150	0.6352	0.7746	0.0682	0.8118	0.7926	0.2433
4.4536	1.5789	60	3.5796	0.9048	0.9804	0.8521	0.1364	0.8548	0.9605	0.5673
0.9561	1.7105	65	3.0011	0.8899	0.9236	0.7887	0.1364	0.8065	0.9249	0.5442
2.3199	1.8421	70	2.8339	0.8788	0.7886	0.7958	0.1364	0.8118	0.8444	0.4564
4.7998	1.9737	75	3.8323	0.8793	0.5446	0.7394	0.1136	0.7742	0.6762	0.2443
1.4698	2.1053	80	2.7076	0.8946	0.8592	0.8169	0.1364	0.8280	0.8738	0.3603
3.3485	2.2368	85	2.9366	0.9013	0.4484	0.9014	0.2273	0.8710	0.8672	0.3065
0.4092	2.3684	90	3.1499	0.9005	0.8832	0.7465	0.0682	0.7903	0.8243	0.2694
3.0081	2.5	95	2.9978	0.9016	0.8955	0.7817	0.1136	0.8065	0.8622	0.3159
2.7492	2.6316	100	2.9338	0.9004	0.6348	0.8662	0.2045	0.8495	0.8820	0.3459
0.327	2.7632	105	2.9305	0.9016	0.7578	0.8592	0.2045	0.8441	0.8984	0.3701
4.6086	2.8947	110	2.9420	0.9028	0.7937	0.8592	0.2045	0.8441	0.9061	0.3819

Framework versions

PEFT 0.18.1
Transformers 4.57.6
Pytorch 2.9.1+cu128
Datasets 4.5.0
Tokenizers 0.22.2

Library: peft
License: gemma
Created: Feb 20, 2026
Updated: Feb 20, 2026