checkpoints – AI Model by jdi1009 | AlphaNeural AI

You can deploy this model and start earning money today!

jdi1009/
checkpoints

like

0

generated_from_trainer

beomi/gemma-ko-2b

Views

No views yet

checkpoints

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

Loss: 0.7898

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: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.05
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.6036	0.1058	50	1.4076
0.9044	0.2116	100	0.9465
0.8924	0.3175	150	0.9083
0.8364	0.4233	200	0.8880
0.8379	0.5291	250	0.8730
0.8329	0.6349	300	0.8551
0.8191	0.7407	350	0.8501
0.8438	0.8466	400	0.8427
0.8266	0.9524	450	0.8367
0.7792	1.0582	500	0.8265
0.7475	1.1640	550	0.8224
0.805	1.2698	600	0.8179
0.7423	1.3757	650	0.8101
0.7808	1.4815	700	0.8081
0.7228	1.5873	750	0.8023
0.7853	1.6931	800	0.7999
0.7779	1.7989	850	0.7954
0.7448	1.9048	900	0.7953
0.7341	2.0106	950	0.7924
0.8145	2.1164	1000	0.7915
0.7167	2.2222	1050	0.7938
0.7539	2.3280	1100	0.7906
0.7376	2.4339	1150	0.7901
0.773	2.5397	1200	0.7904
0.7874	2.6455	1250	0.7902
0.7395	2.7513	1300	0.7900
0.7056	2.8571	1350	0.7897
0.7157	2.9630	1400	0.7898

Framework versions

PEFT 0.12.0
Transformers 4.44.0
Pytorch 2.10.0+cu128
Datasets 2.20.0
Tokenizers 0.19.1