gemma-2-9b_pct_reverse_r32

This model is a fine-tuned version of google/gemma-2-9b on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 9.9020

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.02
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
2.5978	0.0206	8	4.9799
11.0056	0.0412	16	11.4640
11.7215	0.0618	24	11.8765
11.8793	0.0824	32	11.9059
11.8739	0.1030	40	11.9781
11.8763	0.1236	48	11.9487
11.8231	0.1442	56	11.8282
11.7758	0.1648	64	11.7664
11.8011	0.1854	72	11.4372
11.6991	0.2060	80	11.7334
11.8108	0.2266	88	11.5005
11.6519	0.2472	96	11.5944
11.6905	0.2678	104	11.5944
11.6003	0.2885	112	11.5914
11.5813	0.3091	120	11.5684
11.5493	0.3297	128	11.7560
11.5458	0.3503	136	11.4566
11.5838	0.3709	144	11.4331
11.4815	0.3915	152	11.5174
11.5369	0.4121	160	11.5271
11.4617	0.4327	168	11.5392
11.4399	0.4533	176	11.3691
11.3199	0.4739	184	10.9823
10.6547	0.4945	192	10.0666
8.8163	0.5151	200	8.7638
9.5635	0.5357	208	8.5153
8.7862	0.5563	216	9.2186
10.2774	0.5769	224	10.3834
9.7932	0.5975	232	9.7972
9.5421	0.6181	240	9.6845
9.5401	0.6387	248	9.3438
10.9001	0.6593	256	10.6070
9.959	0.6799	264	9.6172
9.5409	0.7005	272	10.4762
10.8074	0.7211	280	10.4872
9.1645	0.7417	288	7.6567
8.1072	0.7623	296	8.7911
9.7069	0.7829	304	10.0043
10.0752	0.8035	312	10.1655
9.9734	0.8241	320	9.9644
9.6722	0.8447	328	9.7803
9.8279	0.8654	336	9.5826
9.6714	0.8860	344	9.5533
9.655	0.9066	352	9.6339
9.7184	0.9272	360	9.7651
9.6142	0.9478	368	9.8536
9.9249	0.9684	376	9.8906
9.8654	0.9890	384	9.9020

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.3.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1