falcon7binstruct_fine_tunning_PPF_english_conversations_2

This model is a fine-tuned version of vilsonrodrigues/falcon-7b-instruct-sharded on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3611

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• 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: 0.03
• training_steps: 250
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.4114	0.0667	10	0.8542
0.715	0.1333	20	0.6213
0.5797	0.2	30	0.5449
0.5136	0.2667	40	0.5064
0.4744	0.3333	50	0.4767
0.4719	0.4	60	0.4568
0.4437	0.4667	70	0.4366
0.4396	0.5333	80	0.4302
0.4078	0.6	90	0.4168
0.4571	0.6667	100	0.4099
0.3773	0.7333	110	0.4028
0.3879	0.8	120	0.3917
0.3965	0.8667	130	0.3832
0.3747	0.9333	140	0.3783
0.3626	1.0	150	0.3742
0.2861	1.0667	160	0.3721
0.2816	1.1333	170	0.3744
0.3017	1.2	180	0.3737
0.2946	1.2667	190	0.3687
0.3064	1.3333	200	0.3672
0.3031	1.4	210	0.3651
0.2959	1.4667	220	0.3628
0.2854	1.5333	230	0.3615
0.2972	1.6	240	0.3612
0.291	1.6667	250	0.3611

Framework versions

• PEFT 0.11.2.dev0
• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1