metadata

base_model: google/gemma-2-2b-it
datasets:
  - GaetanMichelet/chat-60_ft_task-2_auto
library_name: peft
license: gemma
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Gemma-2-2B_task-2_60-samples_config-2_full_auto
    results: []

Gemma-2-2B_task-2_60-samples_config-2_full_auto

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-2_auto dataset. It achieves the following results on the evaluation set:

Loss: 0.8896

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
distributed_type: multi-GPU
gradient_accumulation_steps: 16
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss
1.2776	0.6957	2	1.3061
1.2902	1.7391	5	1.2795
1.2269	2.7826	8	1.1875
1.1326	3.8261	11	1.1151
1.0639	4.8696	14	1.0711
1.016	5.9130	17	1.0248
0.9412	6.9565	20	0.9830
0.9191	8.0	23	0.9520
0.8743	8.6957	25	0.9389
0.8391	9.7391	28	0.9238
0.8374	10.7826	31	0.9124
0.8155	11.8261	34	0.9037
0.782	12.8696	37	0.8979
0.7796	13.9130	40	0.8937
0.7521	14.9565	43	0.8909
0.752	16.0	46	0.8899
0.715	16.6957	48	0.8896
0.726	17.7391	51	0.8901
0.7131	18.7826	54	0.8918
0.6912	19.8261	57	0.8945
0.6795	20.8696	60	0.8986
0.6741	21.9130	63	0.9028
0.654	22.9565	66	0.9070
0.6469	24.0	69	0.9121

Framework versions

PEFT 0.12.0
Transformers 4.44.0
Pytorch 2.1.2+cu121
Datasets 2.20.0
Tokenizers 0.19.1